Benaissa 2021 These

Engineering of chemogenetic tools for the imaging and
manipulation of biological systems

Hela Benaissa
To cite this version:

Hela Benaissa. Engineering of chemogenetic tools for the imaging and manipulation of biological
systems. Theoretical and/or physical chemistry. Université Paris sciences et lettres, 2021. English.
�NNT : 2021UPSLE024�. �tel-03815066�
HAL Id: tel-03815066

https://theses.hal.science/tel-03815066
Submitted on 14 Oct 2022
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Préparée à l’École Normale Supérieure
Composition du jury :
Soutenue par
Hela Benaissa M. Dr. Dominique BOURGEOIS
Institut de Biologie Structurale Rapporteur
Le 11 mai 2021
Mme. Dr. Claire DÉO
Dirigée par European Molecular Biology Laboratory Examinateur
Arnaud Gautier Mme. Prof. Marie ERARD

Université de Paris Saclay Rapporteur
Ecole doctorale n° 388 M. Prof. Arnaud GAUTIER

Sorbonne Université Directeur de thèse
Chimie Physique et Chimie
Analytique de Paris Centre M. Prof. Ludovic JULLIEN
École Normale Supérieure Invité
Spécialité M. Dr. Matthieu SAINLOS

Chimie-Physique Institut Interdisciplinaire de Neurosciences Président du jury
2
Présentée par
Hela Benaissa
Thèse de doctorat de Chimie-Physique
3
4
Acknowledgements (Remerciements)
Je voudrais tout d’abord remercier infiniment mon directeur de thèse le Professeur Arnaud Gautier,
sans qui, ce travail de thèse et toute cette aventure n’auraient pas été les mêmes. Je voudrais le
remercier du fond du cœur pour son éternel soutien, ses encouragements, sa bonne humeur, son
optimisme constant, pour avoir été toujours de bon conseil, que ça soit d’un point de vue professionnel
ou personnel, pour les différentes relectures et corrections de mon manuscrit de thèse et enfin pour
m’avoir transmis toutes les connaissances nécessaires et sa passion de travailler dans ce domaine
de recherche. Je suis à la fois extrêmement contente et très chanceuse d’avoir pu partager cette
incroyable aventure à ses côtés.
Je voudrais remercier tous les membres du jury, Dr. Matthieu Sainlos, Dr. Dominique Bourgeois, Prof.
Marie Erard et le Dr. Claire Déo qui ont accepté d’évaluer mon travail de thèse ainsi que pour la
lecture et les corrections de mon manuscrit de thèse et la discussion au cours de ma soutenance
orale et enfin pour m’avoir décerné le titre de docteure. Je voudrais également remercier le Professeur
Ludovic Jullien pour sa contribution majeure à mes projets de thèse, pour ses précieux conseils et
pour m’avoir fait l’honneur d’être présent en tant qu’invité à mon jury de thèse.
Mes remerciements s’adressent également à tous les collaborateurs avec qui j’ai pu travailler. Je
tiens à les remercier pour leur aide très précieuse, pour toutes les connaissances et tous les conseils
partagés et pour m’avoir formée à des domaines que je ne connaissais pas forcément. Je voudrais
notamment remercier Dr. Nicolas Pietrancosta, Dr. Xavier Morin et Evelyne Fisher, Dr. Lydia Danglot,
Julie Nguyen, Dr. Philippe Bun et David Geny et enfin Dr. Marc Tramier et Dr. Giulia Bertolin. Je
remercie tout particulièrement l’équipe du Professeur Ludovic Jullien et plus particulièrement le Dr.
Isabelle Aujard et le Dr. Thomas Le Saux pour leur aide très précieuse, pour m’avoir transmis leurs
connaissances et Karim Ounoughi et Cyprien Velmir pour leurs contributions dans mes projets de
thèse. Par la même occasion, je tenais à remercier Annie Munier pour son soutien et son aide pour
les expériences de cytométrie à la plateforme de Saint Antoine.
Je tenais également à remercier toutes les personnes de notre superbe équipe, pour tous les bons
moments partagés et pour leur immense soutien au cours de ma thèse. Je tenais à remercier
chaleureusement Marie-Aude Plamont pour tout son soutien, pour tout ce qu’elle m’a appris et puis
finalement pour avoir partagé ces quelques cours de Zumba dont on ne se souviendra très
certainement plus. Je voudrais également remercier Fanny Broch sans qui je n’y serai très
certainement pas arrivée, je tenais à la remercier pour son soutien infini, pour avoir pris le temps de
relire mon manuscrit et pour tout ce qu’on a pu traverser/apprendre ensemble, tout simplement
5
merci Fanny! Un remerciement tout particulier à Louise-Marie Rakotoarison pour son éternel soutien,
sa bonne humeur, pour m’avoir aidée à surmonter toute cette aventure, mais également pour être
respo Apéro/Restaurants et pour m’avoir appris la signification de l’expression en soum-soum!
Un grand merci également à Alison G. Tebo pour m’avoir appris énormément de choses et pour
m’avoir transmis toutes ses connaissances. Je tenais également à remercier Lina El Hajji pour son
éternelle bonne humeur et à qui je souhaite beaucoup de succès pour la suite, qui, je n’en doute pas,
ne sera que grandiose. Et enfin, un grand merci à Chenge Li, Tiphaine Péresse, Amandine Gontier,
Sara Bottone et Aurélien Brion pour leurs soutiens, leurs bons conseils et pour ces supers moments
partagés au laboratoire.
J’ai eu la chance d’avoir pu travailler dans deux laboratoires différents et donc la chance de pouvoir
rencontrer et côtoyer un grand nombre de personnes qui ont participé de différentes manières, de
près comme de loin, à la réussite de mon projet de thèse. Je remercie tous les membres du laboratoire
PASTEUR et plus particulièrement Mary Aubry, Audrey Cochard, Lucas Sixdenier, Marina Garcia-
Jove Navarro, Raja Chouket, Beatrice Adelizzi et Emmanuelle Marie. Je remercie également tous les
membres du Laboratoire des BioMolécules (LBM) et plus particulièrement Eliane Moulinie, Françoise
Illien, Delphine Ravault et Mathieu Letrou. Je tenais enfin à remercier toute l’équipe Vivant de la Cité
des Sciences et de l’industrie pour m’avoir accueillie au cours de cette mission de médiation et plus
particulièrement ma tutrice Graziela Burchard.
Je remercie également tous mes amis pour leurs soutiens et leurs encouragements au cours de ces
quelques dernières années notamment Skander, Youssef, Fahmi, Myriam, Mathilde, Myriam, Hella,
Mehdy, Yasmine, Inès, Juliette, Sophie, Wassim, Sara, Fatma et tous les autres.
Je tenais finalement à remercier toute ma famille qui m’a très fortement soutenue dans cette
incroyable aventure. Je voudrais remercier plus particulièrement mes parents qui m’ont donné la
chance de venir en France il y a maintenant dix ans de cela et d’accomplir toutes ces choses
incroyables. Je voudrais les remercier du fond du cœur pour avoir toujours été là pour moi, pour
m’avoir soutenu toutes ces années et pour avoir toujours été de mon côté et m’avoir encouragé dans
toutes les décisions que j’ai pu prendre. Je remercie également ma sœur, mon beau-frère, mes
grands-mères, mes tantes, mes cousins et tous les autres. Je remercie enfin mon fiancé Ahmed
Belhaouane qui a été d’un soutien infini et indescriptible. Malgré la distance et tous les obstacles
traversés, il a su être à mes côtés pas à pas et m’a toujours fait voir le bon côté des choses. Merci
du fond du cœur, grâce à toi à mes côtés je sais que je pourrais faire face à toute épreuve et je suis
extrêmement impatiente et chanceuse de commencer cette nouvelle étape de ma vie à tes côtés.
6
List of abbreviations HBR derivatives: hydroxybenzylidene
rhodanine derivatives
HBT derivatives: hydroxybenzylidene
2,4-thiazolidinedione derivatives
AA: amino acid
HE: dihydroethidium
ADPA: anthracene-9-10-dipropionic acid
HEK293: Human embryonic kidney
(A)FP: (Auto)fluorescent protein
IC: Internal conversion
Aga2p: α-agglutunin mating protein
IFP: Infrared fluorescent protein
BR: Bilirubin
ISC: Intersystem crossing
BSA: Bovine serum albumin
JF dyes: Janelia Fluor dyes
BV: Biliverdin
LED: Light-emitting diode
CALI: Chromophore-assisted light inactivation
LOV: Light, oxygen, voltage
CLEM: Correlative light-electron microscopy
MEM: Minimum essential media
CMV: Human cytomegalovirus
mFAP: mini Fluorogen-Activating protein
CRABPII: Cellular retinoic acid binding
MG: Malachite Green
protein II
miniSOG: mini Singlet Oxygen Generator
CyDye: Cyanine dye
NIR: Near infrared
DAB: 3,3’-diaminobenzidine
PAS-GAF domain: Per-ARNT-Sim repeats-
DFHBI: 3,5-difluoro-4-hydroxybenzylidene
cGMP phosphodiesterase/adenylate
imidazolinone
cyclase/FhlA transcriptional activator domain
DMEM: Dulbecco's Modified Eagle Medium
PBS: Phosphate buffered saline
DNA: Deoxyribonucleic acid
PCR: Polymerase chain reaction
DPBS: Dulbecco’s Phosphate-Buffered Saline
PDB: Protein data bank
EGFP: Enhanced Green fluorescent protein
PHY domain: phytochrome domain
E. coli: Escherichia coli
POI: Protein of interest
FACS: Fluorescence-activated cell sorting
PPIs: protein-protein interactions
FAP: Fluorogen Activating Protein
PS: Photosensitizer
FAST: Fluorescence Activating absorption
PYP: Photoactive yellow protein
Shifting Tag
p-CA: p-coumaric acid
FbFP: Flavin-based fluorescent protein
RB: Rose Bengal
FMN: Flavin mononucleotide
RFP: Red fluorescent protein
FLIM: Fluorescence lifetime imaging
RMSF: Root mean square fluctuation
microscopy
RNA: Ribonucleic acid
FRET: Förster resonance energy transfer
ROS: Reactive oxygen species
GFP: Green fluorescent protein
ScFv: Single-chain variable fragment
GRAP: RNA Aptamer-based Photosensitizer
SELEX: Systematic evolution of ligands by
HBI: hydroxybenzylidine imidazolinone
exponential enrichment
HBIR derivatives: hydroxybenzylidene
SMLM: Single-molecule localization
isorhodanine
microscopy
HBO derivatives: hydroxybenzylidene
SiR: Silicon rhodamine
2,4-oxazolidinedione derivatives
STED: Stimulated emission depletion
HBP derivatives: hydroxybenzylidene
S. cerevisiae: Saccharomyces cerevisiae
pseudothiohydantoin derivatives
TAP: Targeted and Activated photosensitizer
HBRAA derivatives: hydroxybenzylidene
TMR: Tetramethylrhodamine
rhodanine-3-acetic acid derivatives
UV: Ultraviolet
YFP: Yellow fluorescent protein
7
Contents
AKNOWLEGMENTS ......................................................................................................................... 5
LIST OF ABBREVIATIONS ............................................................................................................. 7
CHAPTER I: GENERAL INTRODUCTION ................................................................................... 11
I.1 Molecular fluorescence ......................................................................................................... 11

I.2 Fluorescence imaging ........................................................................................................... 12
I.2.1. Fluorescence microscopy ......................................................................................... 12
I.2.2. Fluorescence reporters ............................................................................................. 14
I.2.2.1. Genetically encoded fluorescent proteins ............................................................. 14
I.2.2.2. Organic dyes ......................................................................................................... 17
I.2.2.3. ‘Fluorogen-activating’ hybrid reporters: FAST-tagging system ............................. 20
I.3 Ph.D. goals ........................................................................................................................... 24
I.4 References............................................................................................................................ 27
CHAPTER II: FLUORESCENT CHEMICAL-GENETIC HYBRIDS .............................................. 31
II.1 Abstract ................................................................................................................................. 31

II.2 Introduction ........................................................................................................................... 32
II.3 Hybrid reporters .................................................................................................................... 33
II.3.1. Tag engineering ........................................................................................................ 33
II.3.1.1. Rational design ................................................................................................. 34
II.3.1.2. Directed evolution ............................................................................................. 37
II.3.1.3. In silico and de novo design .............................................................................. 40
II.3.2. Fluorogen engineering .............................................................................................. 40
II.3.2.1. Design ............................................................................................................... 41
II.3.2.2. Tuning ............................................................................................................... 43
II.4 Hybrid actuators and biosensors .......................................................................................... 45
II.4.1. Actuators ................................................................................................................... 45
II.4.2. Biosensors ................................................................................................................ 46
II.5 Conclusion ............................................................................................................................ 47
II.6 References............................................................................................................................ 49
8
CHAPTER III: DEVELOPMENT OF A MULTIFUNCTIONAL CHEMOGENETIC REPORTER . 55
III.1 Presentation of the article ..................................................................................................... 56

III.2 Article: Engineering of a fluorescent chemogenetic reporter with tunable color for advanced
live-cell imaging ............................................................................................................................ 57
III.2.1. Abstract ..................................................................................................................... 57
III.2.2. Introduction ............................................................................................................... 57
III.2.3. Results ...................................................................................................................... 60
III.2.3.1. Molecular spectral tuning .................................................................................. 60
III.2.3.2. Optimization by directed protein evolution ........................................................ 60
III.2.3.3. Structural model ................................................................................................ 63
III.2.3.4. Multicolor fluorescent labeling in mammalian cells ........................................... 65
III.2.3.5. Photostability ..................................................................................................... 66
III.2.3.6. Imaging fusion proteins in mammalian cells and cultured neurons ................... 68
III.2.3.7. Enhanced fluorescent labeling in multicellular organisms ................................. 68
III.2.3.8. Reversible labeling ............................................................................................ 71
III.2.3.9. Imaging fusion proteins below the diffraction limit ............................................. 71
III.2.4. Discussion ................................................................................................................. 74
III.2.5. Supplementary information ....................................................................................... 76
III.2.6. Materials and methods ............................................................................................ 110
III.3 Prospects: toward smaller version of pFAST ...................................................................... 121
III.4 References.......................................................................................................................... 125
CHAPTER IV: CHEMOGENETIC PHOTOSENSITIZERS FOR MANIPULATING BIOLOGICAL

SYSTEMS ...................................................................................................................................... 129
IV.1 General introduction............................................................................................................ 129

IV.2 Engineering of targeted and activated chemogenetic photosensitizers for manipulating
biological systems on demand.................................................................................................... 133
IV.2.1. Introduction ............................................................................................................. 133
IV.2.2. Results .................................................................................................................... 134
IV.2.2.1. Molecular engineering and in vitro characterization of ROSgens ................... 134
IV.2.2.2. Selective dual imaging and cell perturbation ................................................... 139
IV.2.3. Discussion ............................................................................................................... 142
IV.2.4. Supplementary information ..................................................................................... 143
IV.2.5. Materials and methods ............................................................................................ 151
IV.3 References.......................................................................................................................... 155
9
CHAPTER V: GENERAL DISCUSSION ..................................................................................... 157
V.1. Development of pFAST-based biosensors ......................................................................... 158

V.2. Development of pFAST-based fluorescent reporters for super-resolution microscopy ...... 159
V.3. Development of pFAST-based chemogenetic photosensitizers ......................................... 160
V.4. References.......................................................................................................................... 161
DIGEST : INGÉNIERIE D’OUTILS CHÉMOGÉNÉTIQUES POUR L’IMAGERIE ET LA

MANIPULATION DE SYSTÈMES BIOLOGIQUES .................................................................... 163
1. Introduction générale (Chapitre I) ....................................................................................... 163

1.1. Mise en contexte ......................................................................................................... 163
1.2. Objectifs de la thèse ................................................................................................... 165
2. Ingénierie de rapporteurs chémogénétiques hybrides (Chapitre II) .................................... 167
3. Développement d’un marqueur chémogénétique multifonctionnel (Chapitre III) ................ 169
4. Développement de photosensibilisateurs chémogénétiques (Chapitre IV) ........................ 174
5. Discussion générale (Chapitre V) ....................................................................................... 176
6. Références.......................................................................................................................... 177
10
Chapter I
General introduction
Nowadays, researchers benefit from a wide range of imaging modalities such as optical microscopy,
electron microscopy, or mass-spectrometry imaging to decipher the complexity of living organisms.
Optical microscopy including techniques such as phase contrast, differential interference contrast
(DIC) and polarized microscopy permit to enhance the contrast of living specimens to make them
more visible. However, fluorescence microscopy became these last decades the benchmark for
imaging living systems as it permits to visualize proteins and other biomolecules normally invisible
and not trackable by common optical microscopes1.
Last advances and innovations associated with fluorescence microscopy techniques allow us to study
complex biological processes with an unprecedented temporal and spatial resolution2. Based on the
expanding palette of fluorescent proteins, Lichtman and his coworkers developed for example the
remarkable Brainbow technology allowing to map neuronal connectivity3,4. Benefiting from advances
in fluorescent probes, targeting strategies, instrumentation, and data analysis, fluorescence
microscopy enables high-throughput screens, single-molecule detection, multiplexing, and live cell
imaging with high speed, sensitivity and resolution.
I.1. Molecular fluorescence
Fluorescence was first observed by Sir George G. Stokes when he noticed that mineral fluorspar
(containing calcium fluoride) exhibited emitted light with a longer wavelength than the light used to
excite the specimen; a phenomenon that became known as the Stokes shift. The term fluorescence
was introduced by George G. Stokes himself in 18525. The fluorescence process is characterized by
the emission of photons arising from the lowest excited electronic state after absorption of light.
Because of the loss of energy due to vibrational relaxation, the emission of light is always located at
a higher wavelength (lower energy) than absorbed light (higher energy)6. The Perrin-Jablonski
diagram - depicted in the Figure I.1 - is convenient for visualizing the possible processes following
11
CHAPTER I : GENERAL INTRODUCTION
the absorption of light by a molecule e.g. internal conversion (IC), intersystem crossing (ISC),
fluorescence, phosphorescence, delayed fluorescence, and triplet-triplet transitions. The
fluorescence is specifically characterized by the emission of photons accompanying the S0 ← S1
relaxation2.
Internal
conversion
S2
T2
Intersystem
crossing
S1
T1
Vibrational
relaxation Fluorescence
Absorption
Phosphorescence
S0 S0
Figure I.1. Perrin-Jablonski diagram (adapted from Valeur2).
I.2. Fluorescence imaging
Seeing inside cells and organisms with high spatiotemporal resolution became these last decades
increasingly effective as both physical optics and probe design permitted to develop innovative
fluorescent markers while gaining sensitivity and spatiotemporal resolution with the development of a
variety of imaging modalities.
I.2.1. Fluorescence microscopy
Fluorescence microscopy like widefield epi-fluorescence, laser-scanning confocal, spinning disk, total
internal reflection microscopy (TIRF), or super-resolution microscopy provide a wide range of imaging
tools to investigate living organisms and cells. A non-exhaustive overview of fluorescence microscopy
techniques is introduced below.
Fluorescence microscopy based on conventional widefield modalities is usually user-friendly, non-

expensive, and easy to implement. In a widefield microscope, all parts of the sample are illuminated
simultaneously, allowing for simple acquisition and fast imaging. However, this simplicity comes with
a cost as simultaneous illumination results in images of low contrast and spatial resolution7. To provide
higher imaging contrast, the goal of laser-scanning confocal microscopy is to reject out-of-focus light
12
from the image by the presence of a pinhole aperture in front of the detectors8. In such systems, the
sample is scanned to record fluorescence intensity pixel by pixel at each position of the image7.
Because of the diffraction of light, one-photon confocal microscopy generally yields a lateral resolution
of approximately 200 nm7. Because of light diffraction, each fluorescent object appears as a blurry
spot, which is bigger than the actual object’s size. Consequently, two fluorescent objects closer than
200 nm, will not be well resolved.
Recent super-resolution techniques such as saturated structured-illumination microscopy (SSIM),

single-molecule localization microscopy (SMLM) or stimulated emission depletion (STED) microscopy
permitted to circumvent this diffraction barrier. All these techniques improve spatial resolution, but
depend on a different set of modalities. Whereas structured-illumination microscopy technique (SSIM)
improves the resolution by a factor of two compared to confocal microscopy9,10, SMLM or STED
microscopy techniques permit to further improve lateral resolution allowing to acquire nano-scale
information. SMLM approach is based on the imaging of a sparse subset of the emissive fluorescent
molecules through time. By utilizing photoswitchable fluorescent dyes (as in STORM for stochastic
optical reconstruction microscopy)11 or photoswitchable fluorescent proteins (as in PALM for
photoactivated localization microscopy)12 that can switch reversibly between a dark and a fluorescent
state (before eventually photobleaching), localizations from thousands of on-off cycles are combined
to reconstruct the entire super-resolved image. However, using SMLM techniques might result in a
relatively poor time resolution as several thousands of images must be captured to reconstruct the
resolved image7. In addition, these approaches constantly require novel improved photoswitchable –
photobleaching resistant fluorophores allowing live single-molecule tracking.
STED super-resolution technique on the other hand is based on a different approach. The principle
of STED microscopy is to reduce the size of the excitation spot. Although the microscope optics
remain limited by the diffraction barrier, a second laser in the shape of a “doughnut” is immediately
applied after the excitation pulse allowing the generation of a depleted doughnut-shape spot around
the excitation spot13,14. The STED laser features zero intensity at the very center of the excitation spot,
depleting the excitation state of the fluorophores at the periphery by stimulated emission15. To gain in
resolution, high STED light intensities are required, which can be deleterious for some of the common
fluorophores. In addition, this technique suffers from a small number of available depletion laser lines
and from small Stokes shift fluorophores which can be re-excited by the depletion laser, therefore
only a few fluorescent probes are known to be suitable for efficient STED nanoscopy.
13
I.2.2. Fluorescence reporters
A large collection of small organic fluorescent dyes, nanocrystals “quantum dots”, autofluorescent
proteins and hybrid chemogenetic reporters can be used to target specific proteins or biomolecules
in biological samples. According to the microscope setup, the biological question and the sample e.g.
in vitro, fixed or live cells, tissues and whole-body animals, users should carefully consider the entire
collection of fluorescent reporters to choose the best alternative.
I.2.2.1. Genetically encoded fluorescent proteins
(Auto)fluorescent proteins (AFPs or FPs) have revolutionized the way of imaging cells, multicellular
organisms and animal tissues. The first green fluorescent protein (GFP) was discovered in Aequorea
victoria (av) jellyfish in 196216 but the initial demonstration in 1994 that the avGFP could function as
a genetically encoded fluorescent tag in living cells was a huge milestone17. By molecular fusion of
the gene of interest to the gene expressing the fluorescent tag, thus promoting labeling specificity,
these FPs became over the last years indispensable tools for biological imaging ranging from single
molecules to entire organisms18 (Figure I.2a). In the avGFP, the 4-(hydroxybenzylidene)imidazolidin-
5-one (HBI) chromophore is embedded in the center of the β-barrel structure. Spontaneous
cyclization, dehydration and oxidation of Ser65–Tyr66– Gly67 amino acids were found responsible
for the full maturation of the HBI chromophore, which is strictly dependent on the presence of
molecular oxygen as a cofactor (Figure I.2b)19.
14
Figure I.2. Specific labeling with GFP. (a) Genetic fusion of GFP to the protein of interest (POI).
(b) The maturation of the HBI chromophore in the binding barrier of avGFP [PBD: 3SG2] (adapted from et al.19).
The original GFP and its enhanced EGFP version20 were first described to fluoresce in green by blue
or UV light illumination and were further evolved to fluoresce from blue (BFP)21 to yellow (YFP)22.
Additional genomic exploration of marine organisms such as corals, sea anemones, hydrozoans
permitted to discover other GFP-like proteins with emission colors ranging from cyan to red23–25
(Figure I.3).
Exc. 380 433/452 488 516 487/504 540 548 554 568 574 587 595 596 605 590 nm
Em. 440 475/505 509 529 537/562 553 562 581 585 596 610 620 625 636 648 nm
ECFP
EGFP
mHoneydew
mOrange
tdTomato
mTangerine
mStrawberry
mCherry
mGrape1
mGrape2
EBFP
YFP (Citrine)
mBanana
mRaspberry
mPlum
Figure I.3. Color palette of FP toolbox (adapted from Shaner et al.25 and Wang et al.26).
Nathan Shaner et al (2004) Nature Biotech. 22: 1567-1572

Lei Wang et al (2004) Proc. Natl. Acad. Sci. USA 101: 16745-16749
15
Over the past 20 years, a combination of rational design, structure-based mutagenesis and
considerable cycles of directed evolution were conducted to fine-tune the photophysical and
biochemical properties and to expand the color palette of available FPs. The main challenge of
engineering fluorescent proteins was to improve their brightness27. The brighter a FP, the higher the
imaging contrast that can be obtained. Although tremendous engineering efforts were conducted in
the past to develop glowing FPs, substantial room for brightness optimization remains possible. The
brightest FPs are mNeonGreen28 or mClover329. Until recently, most of the engineering techniques
were based on Escherichia coli expression of a library of mutants and bacterial colonies were simply
picked for fluorescence brightness or color. Alternative approaches for high throughput screening
compatible with prokaryotic and eukaryotic cells have recently emerged and should give promising
opportunities to probe for more diverse photophysical properties like FusionRed-M screened by a
microfluidic sorter for both increased fluorescence lifetime and brightness30. In addition, the natural
oligomeric form of some FPs motivated the biologists to engineer fully monomeric fluorescent
proteins31. Even if most of the engineered FPs are monomeric by now, a high local concentration of
fluorescent proteins might result in aggregation and oligomerization which might perturb the function
and the localization of the target protein32.
The growing panoply of fluorescent proteins emitting and obscuring the entire visible spectral window,
has recently encouraged researchers to turn to far-red and near-infrared fluorescent labels. In addition
to avoiding high-dose of blue light, which can be deleterious for cells, and to prevent cells and tissues
autofluorescence, far-red fluorophores are preferred for deep-tissue imaging. Far-red and NIR
fluorescent proteins homologous to GFP were not discovered in nature. Thus, extensive laborious
protein engineering approaches would be required to shift their spectral properties27. Researchers
preferred to turn to bacterial phytochromes and other-related cyanobacteriochromes to engineer long
wavelength hybrid fluorescent reporters which fluoresce in NIR by recognizing biliverdin chromophore
(see section II.2.1).
Finally, because many cellular biological phenomena operate at the nanoscale, extensive efforts have
been devoted to using FPs in super resolution microscopy to image beyond the limits of the diffraction
barrier. As discussed in the section I.2.1, photoactivatable GFP-like proteins that switch between
different molecular states have shown great performances in subdiffraction microscopy techniques33–
35
. When irradiating with a specific wavelength light, photoconversion or photoswitching of the HBI
chromophore provides a unique way to isolate single-molecule fluorophores with high spatial
resolution.
16
I.2.2.1. Organic dyes
The continuous ongoing demand for modern and super-resolution microscopy techniques has
brought scientists to develop more sophisticated fluorescent probes. With their excellent
photostability, brightness and versatility, organic dyes have recently taken a new turn36. Synthetic
dyes offer a wide range of physicochemical and spectral properties outperforming fluorescent proteins
under certain conditions. In cells, these small organic fluorescent dyes were originally developed to
be specifically fused to antibodies targeting the protein of interest, as the Alexa Fluor dyes which are
still widely used37,38. Classic immunofluorescence methods using antibody-fluorophores conjugation
are only possible at the surface of living cells requiring otherwise cell fixation or permeabilization37,39.
Later, methods to label specific intracellular structures such as organelles, nucleic acids or lipids have
emerged and are commonly industrialized, such as the well-known Hoechst molecule to label DNA in
live cells40 (see Figure I.4 for Hoechst 33342 structure).
The main advantages of organic dyes are their small size - to not disturb the biological system - and
their large versatility. As a matter of fact, a benefit of synthetic dyes is the ability to employ chemical
approaches to simply control the spectral properties of organic fluorophores39. A good illustration of
how organic chemistry can be used to modify the spectral properties of synthetic dyes is nicely
illustrated by the cyanine dyes (CyDye) (see Figure I.4 for Cy3, Cy5 and Cy7 structures). By
elongating the cyanine core, the well-known CyDye fluorophores based on a sulfoindocyanine
structure exhibited important bathochromic shifts41,42. Exhibiting a panoply of colors, some of the
common fluorescent organic dyes are depicted in figure I.443.
17
Figure I.4. Common fluorescent organic dyes (adapted from Raines et al.43).
In addition, chemical approaches can also be employed to improve the photochemical properties of
synthetic dyes. Replacement of the N,N-dialkyl groups substituents in classic organic dyes e.g.
tetramethylrhodamine (TMR), coumarins, naphthalimides or carborhodamines with azetidine rings led
to the design of a palette of fluorophores with improved brightness and photostability. Minimal
structural changes of the TMR dye with the addition of a four-membered azetidine (Janelia Fluor 549)
showed even greater performances in single-molecule microscopy while preserving its high cellular
permeability44 (Figure I.5).
TMR Janelia Fluor 549
with
Figure I.5. Development of Janelia Fluor 549 by replacement of the N,N-dimethyl group with a four-membered
azetidine rings in the TMR dye (adapted from Grimm et al.44).
18
Despite this large collection of dyes, relatively few organic fluorophores are known to exhibit proper
cell permeability for intracellular labeling44. Classic, net-neutral fluorophores such as coumarins and
rhodamines are some of the best candidates for cell permeability and rapid labeling kinetics44,45. A
large panoply of Janelia Fluor (JF) based on rhodamines - e.g. JF519 to JF711 according to their
absorption properties – that span the visible to the NIR region for intracellular live-cell imaging and
super-resolution microscopy are now available46,47. Because of an equilibrium between a lipophilic,
colorless lactone and the polar, fluorescent zwitterion, some of these JF dyes are mainly non-
fluorescent in aqueous solutions and become highly fluorescent in a less polar environment (Figure
I.6). As a matter of fact, the constant need for significantly more complicated protocols for chemical
delivery (microinjection, electroporation) and for washing-out excess of organic dyes have motivated
chemists and biologists to develop cell-permeable fluorescent environment-sensitive probes that
switch on upon interaction/reaction with a given biomolecule of interest. Without proper recognition
by their target, these dyes, also known as fluorogens,
N remain
X mainly
N non-fluorescent. The
Zwitterion
+
ARTICLES
fluorescence response is dependent b
a on the probe microenvironment c or
e.g. modification in polarity
N X N+ CO2 +
molecular order.
Zwitterion
+
ARTICLES
log KL–Z
e
a
F
X=
Always fluorescent
F
Zwitterion
a b c
fluorescent CO2 non-fluorescent
KL–Z + +
N
N
O
9
log KL–Z
Lactone
+
N X N Zwitterion N X N
Improved permeabilityN
Zwitterion
+
+ A
KL–Z Always fluorescent
O
10
Fluorogenic
Lactone
CON X N + + O
KL–Z
2
O
O
log KL–Z
Lactone
Zwitterion
CO2
CO2
Always nonfluorescent
Zwitterion Improved
Lactonepermeability
O
O
11
eO f
N
+
N
N
N
KL–Z
+
Figure I.6. The lactone-zwitterion states of Janelia Fluor (JF) rhodamine dyesFluorogenic
determined by the equilibrium
Lactone
N X N
F
Lactone Zwitterion
1N 2
F
constant (KL-Z) (adapted from Grimm et al.47). X 11

=
3
eO F f Always nonfluorescent
F 0 1 g
f
log KL–Z
F F
N+
Zwitterion
O
–4
–2
N X 1
0
log KL–Z 2 500

log KL–Z
11 500
9
Recently,Lactone
bipartite labeling systemsF composed
Zwitterion
Lactone
Lactoneof a genetically encoded
Zwitterion
F 0
tag and 3a synthetic fluorogen
–2
11
F F
e
have gained high attention, as the fluorogen
N f N
X becomes fluorescent
CO only wheng
4 it is specifically
+
500
2
log KL–Z
9 10
1
Lactone
+
9
recognized by the binding tag45,48,49. The initial breakthrough in live-cell labeling came by the
Zwitterion
1 2 –2 F
2
11 3 –4 F 6
10
introduction of biarsenical
F
dye ligands whichF exhibited
CO 0
fluorescence upon specific recognition
N
by a
3
5
Always nonfluorescent
Improved permeability
2
X= O N 10
600
F F
Lactone
Always fluorescent
+
+
4
N X N
short genetically encoded tetracysteine peptide tag50. Although
9 9 10the
4
–4 small size of the peptide tag is
11
log KL–Z
7 8500
5
JF JF JF
Fluorogenic
525 585 479
convenient for not disturbing the function of proteins of interest, such systems suffered from a low
X= O –2 N
6
6
7 8
700
j JFself-labeling k
CO2 5 O
9 10 11 500 600 700
signal-to-noise ratio and low binding specificity.
JF Since,
10
tags such as SNAP, CLIP
1 or Halo
Lactone
JF
Zwitterion
(nm) 11 2
525 585 479
abs
HN 3
7 8
tags have emerged. By specifically recognizing and
–4
activating the fluorogen module,
0 these 13self-
g
j k l
X X
X= O N Cl
X X
Zwitterion
labeling tags provide the binding specificity of fluorescentNproteins

O and the
11N physicochemical
1 2
properties
F
+
log KL–Z
3 9 1
F
9 10 11
X=
500 600 700

JF525 JF585 JF479 JF479–Hal
0
of labeling dyes. A review of the useX of self-labeling tags Fin Xbiological (nm)
imaging is13presented
–2
in14Chapter abs
Cl
500
F
500
X X
N
4
JF
N
1
O
N+
log KL–Z
II (section II.2.2) of jthis manuscript. k l

N O CO2 9 10
HN
Lactone
Zwitterion
H l T li
1 2 –2
11HTL
11
O
3 14 –4 6
F F
N
5
2
0
Br Br
X X CO2 H F
13 10
X
+
X=
O
N
O
Lactone
550
X X 4
CO2
d
N+
log KL–Z
13 14
N O 9 7 8500
O
–4
O
JF552 JF526 ClX

S
3
H F –2
N
X= 6
+
F F 14
O
N
5
CO2 13 14 500 600 700
Fig. 1|General rubric relating lactone– N zw itter
10
F
Lactone
600
JF552 JF526
F
(nm)
4
exchange
THF
19 –4
Metal–Br
abs
a, The lactone–zwitterion
7 8
equilibrium of NH rho
JF
h
Normalized fluorescence Cl N 2
H F
different
Fig. 1|General rubric dyeslactone–
relating based on
zwKitterion
L–Z.c, Synthesis, struc
equilibrium
2 : X = H; JFt
0.5
1.0
X=
Si
5
0
STL 54
N
13 14 500 600 700

13 : X = F; JF
65
3
I.2.2.1. Fluorogen-activating’ hybrid reporters: FAST-tagging system
AFPs and organic dyes are undoubtedly powerful tools for cell and tissue fluorescence imaging. On
one hand, by their specific localization in cells and their panoply of applications, genetically encoded
fluorescent proteins have revolutionized the way of imaging living systems. On the other hand by their
small size and large versatility, organic dyes are still showing strong interest in the field. However
both classes have specific limitations which have brought scientists to develop alternative markers.
As discussed above, alternative self-labeling tags have expanded the overall fluorescence toolbox.
In addition, ingenious hybrid chemogenetic reporters based on specific protein tag recognition have
arisen. This class of hybrid reporters is composed of a genetically encoded tag (protein or RNA) that
forms a fluorescent complex with small fluorogens. The fluorescence activation of the fluorogen in the
binding pocket of the tag provides high imaging contrast in living systems. Combining the labeling
specificity of AFPs and the large tunability of synthetic fluorophores, hybrid reporters are showing
great promise for pushing the frontiers of bioimaging. The overall family of chemogenetic hybrid
reporters is reviewed in Chapter II of this manuscript. Among them, a fluorogenic reporter named
FAST for Fluorescence Activating and absorption Shifting Tag was recently developed in our lab.
FAST is a small monomeric protein of 14 kDa which was first reported to non-covalently bind
hydroxybenzylidene rhodanine (HBR) analogues51 (Figure I.7). The specific recognition of HBR
derivatives by the protein tag permitted to efficiently activate their fluorescence. Composed of an
electron-donating phenol ring conjugated with an electron-withdrawing rhodanine heterocycle, HBR
derivatives are known as push-pull compounds. In solution or in cells, they dissipate light energy
mainly through radiationless decay channels ; whereas conformationally locked in the cavity of FAST,
they exhibit a strong green-yellow fluorescence activation providing high imaging contrast. Initially,
changing the electron withdrawing ability of the phenol ring through adequate substituents conjugated
to the rhodanine head permitted to shift the fluorescence of FAST to the red edge of the visible
spectrum52 (Figure I.7). HBR derivatives are fully protonated at physiological pH and they undergo a
50 to 80 nm red-shift in absorption upon deprotonation. Locked in the cavity of FAST, they display a
red-shifted absorption, similar to their deprotonate state, demonstrating that FAST stabilizes their
phenolate forms. This spectroscopic feature enables first to easily discriminate the free and bound
states of the fluorogen and allows specific red-shifted excitation of the phenolate form51.
20
Figure I.7. Specific FAST labeling with HBR analogs (POI: protein of interest).
FAST was engineered through directed evolution from the photoactive yellow protein (PYP) of
Halorodhospira halophila. PYP is a known natural photoreceptor found in purple photosynthetic
bacteria which senses light through the photoisomerization of its natural p-coumaric acid (pCA)
chromophore. In the wild-type PYP (wtPYP), the chromophore is covalently attached via a thioester
bond to the Cys69 residue53. In addition to its small size, PYP was chosen as a protein scaffold for
engineering FAST because the natural pCA chromophore and HBR showed common structural
features51. By incorporating C69G residue and by randomizing the 94-101 loop of PYP, screening by
yeast surface display coupled to fluorescence-activated cell sorting (FACS) allowed to efficiently
select for the positive FAST variant (Figure I.8a). The variants were displayed on the surface of the
yeast through fusion with the α-agglutunin mating protein (Aga2p). Projecting the variants away from
the yeast cell wall allowed good ligand recognition and efficient selection by fluorescence sorting
device (Figure I.8b). Yeast surface display which was the first time described by Boder and
Wittrup54,55 displayed numerous assets for high throughput screening of proteins such as (i)
incorporation of post-translational modification, (ii) normalization by protein expression with the
inclusion of epitope tags and (iii) compatibility with flow cytometry analysis56. Flow cytometer equipped
with a sorting device (commonly called fluorescence-activated cell sorter or FACS) enabled high
throughput selection of positive clones over a million of PYP variants. Based on this strategy, the
resulted FAST protein nicely illustrates the power of directed evolution.
21
C69G +
94 - 101 loop engineering
94-101 loop
b
Supplementary Figure 2. The Supplementary
story of FAST and frFAST
Figure engineering.
2. The FAST and
story of FAST was frFAST engineer
engineering from the photoactive engineering
yellow protein (PYP)
from theby mutating the
photoactive Cys protein
yellow 69 (onto(PYP)
whichby mutating the Cy
the 4-hydroxycinnamoyl chromophore is covalently attached
the 4-hydroxycinnamoyl in PYP) to Gly
chromophore and the loop
is covalently attached in PYP) to G
94-101 (FAST loop) that gates the binding
94-101 site[4]loop)
(FAST . It is assumed
that gatesthat
the FAST stabilized
binding site[4]. Itthe
is assumed that FAS
phenolate state of HBR analogs via the same
phenolate network
state of hydrogen
of HBR bonds
analogs via that stabilizes
the same network of hydrogen bond
the phenolate state of the hydroxycinnamoyl
the phenolate chromophore within the parentalchromophore
state of the hydroxycinnamoyl PYP within the
(involving Tyr 42 and Glu 46) and (involving
that the loop 94-101 interacts with the bound fluorogenic
Tyr 42 and Glu 46) and that the loop 94-101 interacts with the bo
HBR analogs[1,4]. The mutations HBR
introduced to [1,4]
analogs generate frFAST (inintroduced
. The mutations magenta) to
aregenerate
in the frFAST (in mag
chromophore binding site (F62L chromophore
and V107I) or binding
in closesite
proximity
(F62L with
and the chromophore
V107I) or in close proximity with th
entry site (D71V, P73S, E74G). These latter(D71V,
entry site mutations (inor
P73S, particular
E74G). P73S)
These maymutations
latter lead to a (in particular P73S
Supplementary Figure Supplementary
2. The story Figure 2. The
ofbackbone
FAST and story
frFAST of FASTaccommodate
and frFAST
engineering. engineering. FAST was
reorganization of the that enables to of
reorganization the backbone FAST
better that the was
largertosize
enables of accommodate
better th
engineering
HPAR-3OM and
engineering from the photoactive from
yellow the
might photoactive
create
protein (PYP) yellow
a loop (P73S
by protein
loop)
mutating that (PYP)
directly
the Cys 69 by(onto
mutating
interacts with the Cys 69 (onto which
HPAR-3OM.
which
HPAR-3OM and might create a loop (P73S loop) that directly interacts w
All shown
the structures were drawn
4-hydroxycinnamoyl using
chromophore thestructures
iscrystal structure of PYP (PDBthe6P4I) andstructure
the 4-hydroxycinnamoyl chromophore is covalently Allattached
shown incovalently
PYP) wereto Glyattached
drawn in
and using
the PYP)
loop to Gly
crystal and theofloop
PYP (
PyMOL
94-101 software.
(FAST loop) that gates the binding site [4]
. It is assumed that FAST stabilized the
[4] PyMOL software.
94-101 (FAST loop) that gates the binding site . It is assumed that FAST stabilized the
Figure I.8. Directed evolution of FAST
phenolate from
state theanalogs
of HBR photoactive yellow
via the same protein
network PYP. (a)
of hydrogen FAST
bonds thatwas obtained
stabilizes
phenolate state of HBR analogs via the same network of hydrogen bonds that stabilizes
by incorporating the mutation
the phenolate C69G
the
state of and randomizing
phenolate
the the loop 94-101
state of thechromophore
hydroxycinnamoyl hydroxycinnamoyl in
thePYP (PDBPYP
within chromophore
parental 6P4I).
within theIt parental
is assumed
PYP that
FAST stabilized theTyr

phenolate (involvingofTyr 42 and Glu 46)via
andthe
that same
the loopnetwork
94-101 interacts with thebonds
bound fluorogenic
(involving 42 and Glustate HBR
46) and that theanalogs
loop 94-101 interacts with the boundoffluorogenic
hydrogen that stabilizes
[1,4]
HBR analogs . The mutations introduced to generate frFAST (in magenta) are in the
the phenolate
HBRstate of [1,4]
analogs the. The
hydroxycinnamoyl chromophore
mutations introduced to generate within
frFASTthe parental are
(in magenta) PYPin (involving
the Tyr 42 and Glu
chromophore binding site (F62L and V107I) or in close proximity with the chromophore
chromophore
46) and that binding site
the loop 94-101 (F62L and
interacts withV107I) or in close
the bound proximityHBR
fluorogenic with the chromophore
analogs. The structures were drawn
entry site (D71V, P73S, E74G). These latter mutations (in particular P73S) may lead to a
entry site
using the crystal (D71V, P73S,
structure E74G).
of PYP and These
PyMOLlatter mutations
software. (b)(inYeast-surface
particular P73S)display
may lead to afluorescence-activating
and
reorganization of the backbone that enables to better accommodate the larger size of
reorganization of the backbone that enables to better accommodate the larger size of
cell sorting strategy enabled FAST selection
HPAR-3OM from create
and might the library
a loopof(P73S
pyp mutants
loop) that (adapted from Li.
directly interacts al.57).
withetHPAR-3OM.
HPAR-3OM and might create a loop (P73S loop) that directly interacts with HPAR-3OM.
All shown structures were drawn using the crystal structure of PYP (PDB 6P4I) and
All shown structures were drawn using the crystal structure of PYP (PDB 6P4I) and
As discussed
PyMOL Chapter II PyMOL
in software. (sectionsoftware.
II.2.1.2) the monomeric state, small size, selective labeling and
molecular oxygen independence of FAST allowed to efficiently visualize FAST-tagged proteins in
multiple organisms such as aerobe and strict anaerobe bacteria, yeast, mammalian cells, multicellular
zebrafish embryo/larvae and chicken embryo using both conventional and super-resolution
microscopy51,57–60. Thanks to its reduced genetic footprint and function in absence of molecular
oxygen, recent studies demonstrated that FAST system outperformed the widely used GFP-like
proteins. FAST permitted to efficiently study bacterial biofilm, label cocultured Clostridium species
and remained fully functional after secretion of bacterial proteins where FPs were reported to be
ineffective61–64.
22
Thanks to the bipartite nature of fluorogenic reporters, tuning both fluorogen and protein moieties
opened new prospects for FAST-tagging imaging. Molecular engineering permitted the development
of membrane-impermeant fluorogens for the selective visualization of extracellular protein trafficking65
(Figure I.9a). Recently, protein engineering of orthogonal - green and red - FAST variants allowed
efficient multicolor imaging66 (Figure I.9b). Finally a recent coordinated strategy of fluorogen
engineering and protein engineering permitted to develop a far-red fluorogenic reporter based on an
engineered FAST variant (frFAST). As FAST, the same evolution strategy based on yeast-display
and fluorescence-activated cell sorting allowed to develop frFAST. Because of an extra double-bond
between the phenol and the rhodanine moieties, (4-hydroxy-3-methoxy-phenyl)allylidene rhodanine
(HPAR-3OM) fluorogen displayed red-shifted spectral properties when locked in the cavity of frFAST57
(Figure I.9c).
NATURE CHEMICAL BIOLOGY

membrane-impermeant
ARTIC
a HBR-3,5DOM 1. HBR-3,5DOM
HBR-3,5DOM
fluorogen
1. 1. HBR-3,5DOM
1. 1. 1. FAST HMBR FAST1.
ormalized luorescenc
HBR-3,5DOM HBR-3,5DOM
or anc
1. 1.
or anc
or anc
HMBR HMBR + HMBR
xcitati
HMBR HMBR
xcitati
xcitati
FAST O fluorogen
FAST O A
engineering
ormalize a
ormalize a
ormalize a
.POI POI.
ARTICLES
. NH .. .. . .
NH H
rmaliz
COO-
rmaliz
rmaliz
S S
HO HO
O POI – POI
O
R S
R S FAST FAST:fluorogen
HBR derivatives
HBR derivatives HBRAA derivatives
HBR derivatives HBR erivative
1. HBR-3,5DOM 1. 06 1. HBR-3,5DOM
07 0 0 1,00 1,10
0 6 0 7 1,00membrane 1,00 1,10 1, 0 membrane
ormalize a or anc
6 7 1,10 1, 0
HMBR avelength (nm) HMBR
xcitati
avelength (nm) avelength (nm) avelength (nm) avelength (nm)
A (nm)
avelength
greenFAST redFA T
greenFAST
redFA T redFA T greenFAST
b 1. greenFAST
ARTICLES
. greenFASTgreenFAST .1.The copyright
.1. greenFAST
. 1.
1.bioRxiv preprint doi: https://doi.org/10.1101/2020.04.04.022145 holder for this preprint (which was not peer-review
or anc
O H
greenFAST greenFAST 1. 1.
MICAL BIOLOGY
1. redFAST
rmaliz
1. redFAST
or anc
CELL
or anc
redFAST
redFAST under a CC-BY-NC-ND
author/funder.
M It isredFAST
made available 4.0 International license.
xcitati
redFAST
xcitati
M
xcitati
M NH H
FAST H HBR H
erivativeS
Normalize a
H H
O greenFAST redFAST
. greenFAST
. .
redFAST
Normalize a
or anc a
N rmaliz
. ..S . 6 . 7 0 0 1,00 1,10 1, 0
directed1. HBR-3,5DOM
HMBR
M
HBR-3,5DOM
N rmaliz
ormalize a Normalize
1. 1.
rmaliz
M M HMBR
HMBR HBR-3,5DOM
HMBR avelength (nm) + HMBR avelength (nm)
evolution
xcitati
FAST HBR-3,5DOM HMBR greenFAST HBR-3,5DOM redFA T

O
N
1. greenFAST 1. 6 1. greenFAST
7 0 0 1,00 1,10
+ HBR-3,5DOM redFAST
redFAST6
or anc
O
redFAST
NH . 7 . . 0 6 X0 7 1,00Y 1,10 1, 0Wavelength
0 (nm) 0 1,00
X 1,10 1, 0 Y Wavelength (nm)
xcitati
MeO
rmaliz
S H NH Wavelength (nm) Wavelength (nm)

O Wavelength (nm) Wavelength (nm)
R S
Normalize a
S O . EL
. .
HBR derivatives
N rmaliz
S
HBR derivatives
ARTICLES
EL OMe EL
MICAL BIOLOGY HMBR 6 HBR-3,5DOM
7
HBR-3,5DOM 0 0 1,00 1,10 1, 0 greenFAST redFAST
avelength (nm)
greenFAST redFAST greenFAST
avelength (nm) redFAST
c redFA T
greenFASTWavelength (nm)
6 7
HBR-3,5D M
0
CELL
0
Wavelength (nm)
1,00 1,10 1, 0
1.
1.
greenFAST 1.
HBR-3,5DOM 1. 1. HBR-3,5DOM far-red
ormalized
1. redFAST
anc
or anc
redFAST X
HMBR HBR-3,5D X
HMBR
M fluorescence
xcitati
HBR-3,5D M
xcitati
M
FAST frFAST
FAST
directed frFAST
EL
ormalize aa or
X
O H fluorogen X X X
engineering
.. .. evolution
..
+ greenFAST redFAST
+
luorescenc
NH
Normalize
rmaliz
N rmaliz
M S Mito-greenFAST H2B-redFAST H2B-greenFAST MAP4-redFAST

HO
O
HBR-3,5DOM
R S
H2B-greenFAST POI HPAR-3OM POI
AST HBR H2B-redFAST Mito-greenFAST
derivatives
HBR derivatives H2B-redFASTH2B-greenFAST
HPAR-3OM MAP4-redFAST
HBR-3,5D M
MAP4-redFAST
66 7
X7
00 00 1,00
1,00 X 1,10
1,10 1,
1, 00
Wavelength (nm) Wavelength (nm)
avelength (nm) avelength (nm)
Figure I.9. Protein and fluorogen engineering strategies based on FAST-tagging reporters. (a) Synthesis
of membrane-impermeant fluorogens (HBRAA derivatives) allowed to study proteins at the cell surface (adapted
from Li. et al.65). (b) Directed evolution and rational design allowed to design orthogonal green and redFAST
variants to bind selectively HMBR or HBR-3,5DOM (adapted from Tebo. et al.66). (c) Coordinated strategy of
fluorogen and protein engineering allowed to design frFAST variant encoding far-red fluorescent assembly with
HPAR-3OM (adapted from Li. et al.57).
23
I.3. Ph.D. goals
To decipher the complexity of living organisms, great opportunities have arisen from the available and
growing palette of fluorescent probes. Combined with sophisticated imaging modalities,
autofluorescent proteins, organic dyes and hybrid reporters allow imaging and multiplexed imaging
with unprecedent spatial and time resolution. By employing sophisticated protein and molecular
engineering strategies, the development of hybrid chemogenetic reporters is showing great promise
in bioimaging and optogenetic. The chapter II gives an overview on the recent advances in the design
of hybrid-fluorogenic reporters for imaging living specimens.
To provide biologists with a general tool that could be used for various experimental scenarios, the
first objective of my Ph.D. has consisted of engineering a robust multicolor and versatile hybrid
reporter based on the FAST system for various applications in cellular imaging. First, in collaboration
with the group of Ludovic Jullien, analogs of the HBR derivatives with minimal structural changes
were synthetized. By replacing the electron-withdrawing rhodanine (R) heterocycle with isorhodanine
(IR), pseudothiohydantoin (P), 2,4-thiazolidinedione (T) and 2,4-oxazolidinedione (O) and by
adjusting the electronic properties of the molecules with varying the phenol electron-donating moiety
with various substituents, we successfully identified a set of fluorogenic chromophores providing
optical properties spanning the entire visible spectrum. Tremendous engineering efforts conducted
by directed evolution and rational design allowed us to identify a FAST variant (named pFAST) with
extended chromophore promiscuity able to form blue, cyan, green, yellow, orange and red fluorescent
assemblies with our collection of fluorogenic chromophores (Figure I.10). As for FAST, frFAST, green
and redFAST51,57,66, the use of yeast-surface display coupled to fluorescence-activated cell sorting
showed great efficiency for the directed evolution of this variant. In the frame of this work, we
discovered that pFAST conjugated to some isorhodanine derivatives permitted to generate absorbing-
only dark assemblies. This singular property was ingeniously used for rapidly turning off pFAST
fluorescence by chromophore replacement. pFAST showed also superior performances with the HBR
derivatives and allowed to efficiently bind and activate membrane-impermeant fluorogens enabling
efficient labeling of cell-surface proteins. We demonstrated furthermore that pFAST could be used to
efficiently image fusion proteins in living cells, neurons and in multicellular organisms, and that its
performance and photostability enabled imaging proteins with subdiffraction resolution in live cells
and neurons by stimulated emission depletion (STED) nanoscopy (see Chapter III).
24
a HBO derivatives HBRAA derivatives
or an
or an
HM
FAST FAST
NATURE CHEMICAL
NATURE
BIOLOGY
CHEMICAL BIOLOGY
O O
or ancormalize a
or ancormalize a
NH . NH .
fluorogen O
O
S
O
S
COO-
-activating tag R O
S 1.
R S HBR-3,5D
1.
HBR derivatives HBR derivatives
HBT derivatives HBR derivatives HM
FAST O FAST O
NATURE CHEMICAL
NATURE
BIOLOGY
CHEMICAL BIOLOGYavelength (nm)
6 7
ormalize a
ormalize a
. .
ARTICLES ARTICLES
NH NH
MICALBIOLOGYb NATURE CHEMICAL BIOLOGY greenFAST
O
greenFAST S T
redFA
S
O
redFAS T
greenFA
FAST O pFAST O R O 1. R 1.
:HBX
anc
anc
S S HBR-3,5D
HBR derivatives 1. derivatives
HBR 1.redFA
Normalizea aor oranc
Normalizea aor oranc

R R M M
1.1 HBR-3,5DOM 1. 1. 1. HBR-3,5DOM
1
HBR-3,5DOM 1. HBP derivatives
HBR-3,5DOM HBIR derivatives HM
1.
NH
or anc
ormalize a or anc
NH HMBR HMBR HMBR H FAST HMBR O FAST
HH S
O H 6 7
S
xcitati
xcitati
FAST O O
FAST Z
O HO
HO
fluorogen 450 . avelength (nm) .
O O . .
ormalize a
. R2 NH NH
engineering .NH M MredFA
ormalize
ormalize
. R2 W . . .
NH greenFAST redFA
greenFAST S T T
rmaliz
rmaliz
S Y
S S S
HMBR O HBR-3,5DOM
HMBR O 1. HBR-3,5DOM greenFA
HBT-3M (R1= Me, R2= H) R
O O 1.
HBO-3M (R1= Me, R2= H)
or anc
or anc
R R
X R
S S NH
S O
S
POI HBR derivatives & directed(R= Me, HBR
HBO-3,5DM
derivatives
R2= Me) HBT-3,5DM (R = Me,MR2HBR
= Me)
redFA
evolution
1
6 7 b0 0 HBT-3,5DOM
1,006 1,10 7 1, 0 (R
1
= OMe, R2= OMe)
0
derivatives
0 1,00 1,10
HH 1, 0
M HBR derivatives
6 7
1 H H
6 (nm) 7
500 Wavelength
Normalize a
Normalize a
avelength (nm) avelength
avelength(nm)
(nm)
HBR derivatives avelength (nm)
. .
avelength (nm)
redFA T greenFAST redFA T
with1.heavy atoms Cell killing
O greenFAST 1. 1. greenFAST greenFAST 1. greenFASTM M
ST
1. 1.
greenFAST SredFA
greenFAST T redFAEL T
or anc
redFAST Normalize a or anc redFAST redFAST OredFAST 1 ISC

R1 HMBR HBR-3,5DOM
HMBR HBR-3,5DOM
ROS CALI
xcitati
xcitati
O greenFA
absorption
M M
Y 1. 1.
T1
or anc
or anc
H H NH H
redFA
S NH ROS6 signaling
Normalize a
. . . . . . M M greenFAS
NH HO S 7
-
N rmaliz
N rmaliz
550 3
X M NH M HO H HH O2 H
Wavelength (nm)
HMBR R2 S
Normalize a
Normalize a
Y photosensitizer X S0 . .
HBP-3M (R1= Me, 6 R = 7H) -activating tag 0 0 1,006 1,10 7 1, 0 M M HBR-3,5D M

HMBR (R(X= Me, R2= = H)
0 0 1,00 1,10 1, 0
2 HBR-3Br 1 = Br, Y
EL
HBP-3,5DM (R 1
= Me, R2= Me)
Wavelength (nm) Wavelength
Wavelength (nm) HMBR
(nm)
HBR-3,5DM (R= Me, R =
HBR-3,5DOM
XHMBR
Wavelength (nm)
Me)
HBR-3,5DOM
X X
HBP-3,5DOM (R = OMe, R2= OMe) HBR-3,5DBr (X 1 = Br, Y2= Br)
fluorescence 1 600 HBR-3,5DOM (R = OMe, R = OMe) greenFAS

EL EL
HBR-3,5DI 1(X = I, Y =2 I) 6 7
greenFAST redFAST greenFAST redFAST Wavelength (nm)
Mito-greenFAST Mito-greenFAST
H2B-redFAST H2B-redFAST H2B-gr
Figure I.10. A coordinated fluorogen/protein
l em(nm) engineering strategy allowed the development
HBR-3,5D M of (a) a
EL
HBR-3,5D M HBR-3,5D M X X X
versatile fluorescent chemogenetic
X
reporter with tunableXcolor for various applications in cellular imaging (see
Chapter III) and (b) an activatable chemogenetic photosensitizer for Reactive Oxygen Species (ROS)
generation with light to promote cell killing, chromophore-assisted light inactivation (CALI) and ROS signaling
(see Chapter IV).
Beyond offering a panoply of opportunities in biological imaging, fluorogenic chemogenetic systems

are also powerful tools to investigate and study biochemical events with high spatiotemporal
resolution. Recently, FAST allowed the design of allosteric-like fluorescent reporters for sensing
intracellular analytes and the generation of a fluorescent split reporter enabling to study protein-
protein interactions with unprecedented spatial and temporal resolution67,68. Trying to expand the
scope of possibilities, we wondered if FAST could be engineered to study biological processes in new
and ingenious ways. Reactive oxygen species (ROS) which are known to cause damage in cells
through oxidative reactions can be used to promote chromophore-assisted light inactivation (CALI),
cell killing, correlative-light electron microscopy (CLEM) or to induce a biological response by ROS
signaling in cells. As introduced in Chapter IV, some of the common fluorescent reporters
(autofluorescent proteins, organic dyes and hybrid reporters) were recently fine-tuned to generate
efficiently reactive oxygen species for various applications in live and multicellular organisms. In
collaboration with the group of Ludovic Jullien, HBR derivatives functionalized with heavy atoms were
synthesized. Such substitutions are known to increase intersystem crossing thus promoting the
photogeneration of ROS in presence of molecular oxygen. Just like regular HBR derivatives, the
‘activation’ of heavy-atom-substituted HBR analogs is only initiated once recognized by the protein
tag; they otherwise dissipate light energy non-radiatively in solution.
25
Both in vitro and in vivo analysis demonstrated that the generation of ROS was strictly controlled by
the addition of these HBR analogs forming chemogenetic photosensitizer assemblies with FAST
variants when irradiated with cyan light (488 nm). Finally, we were able to photoinduce selectively
cellular disorders in live cells, demonstrating the potential of such systems for manipulating biological
systems (see Chapter IV).
26
I.4. References
1. Sanderson, M. J., Smith, I., Parker, I. & Bootman, M. D. Fluorescence microscopy. Cold Spring
Harb. Protoc. 2014, (2014).
2. Valeur, B. & Berberan-Santos, M. N. Molecular Fluorescence: Principles and Applications, 2nd
Edition | Wiley. Wiley.com.
3. Livet, J. et al. Transgenic strategies for combinatorial expression of fluorescent proteins in the
nervous system. Nature 450, 56–62 (2007).
4. Cai, D., Cohen, K. B., Luo, T., Lichtman, J. W. & Sanes, J. R. Improved tools for the Brainbow
toolbox. Nat. Methods 10, 540–547 (2013).
5. Stokes, G. G. XXX. On the change of refrangibility of light. Philos. Trans. R. Soc. Lond. 142,
463–562 (1852).
6. Ettinger, A. & Wittmann, T. Chapter 5 - Fluorescence live cell imaging. in Methods in Cell
Biology (eds. Waters, J. C. & Wittman, T.) vol. 123 77–94 (Academic Press, 2014).
7. Sanderson, M. J., Smith, I., Parker, I. & Bootman, M. D. Fluorescence Microscopy. Cold Spring
Harb. Protoc. 2014, (2014).
8. Shotton, D. & White, N. Confocal scanning microscopy: three-dimensional biological imaging.
Trends Biochem. Sci. 14, 435–439 (1989).
9. Gustafsson, M. G. L. Surpassing the lateral resolution limit by a factor of two using structured
illumination microscopy. J. Microsc. 198, 82–87 (2000).
10. Gustafsson, M. G. L. Nonlinear structured-illumination microscopy: Wide-field fluorescence
imaging with theoretically unlimited resolution. Proc. Natl. Acad. Sci. 102, 13081–13086 (2005).
11. Rust, M. J., Bates, M. & Zhuang, X. Sub-diffraction-limit imaging by stochastic optical
reconstruction microscopy (STORM). Nat. Methods 3, 793–796 (2006).
12. Betzig, E. et al. Imaging Intracellular Fluorescent Proteins at Nanometer Resolution. Science
313, 1642–1645 (2006).
13. Hell, S. W. & Wichmann, J. Breaking the diffraction resolution limit by stimulated emission:
stimulated-emission-depletion fluorescence microscopy. Opt. Lett. 19, 780–782 (1994).
14. Klar, T. A., Jakobs, S., Dyba, M., Egner, A. & Hell, S. W. Fluorescence microscopy with
diffraction resolution barrier broken by stimulated emission. Proc. Natl. Acad. Sci. 97, 8206 (2000).
15. Sednev, M. V., Belov, V. N. & Hell, S. W. Fluorescent dyes with large Stokes shifts for super-
resolution optical microscopy of biological objects: a review. Methods Appl. Fluoresc. 3, 042004
(2015).
16. Shimomura, O., Johnson, F. H. & Saiga, Y. Extraction, Purification and Properties of Aequorin,
a Bioluminescent Protein from the Luminous Hydromedusan, Aequorea. J. Cell. Comp. Physiol. 59,
223–239 (1962).
17. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W. & Prasher, D. C. Green fluorescent protein
as a marker for gene expression. Science 263, 802–805 (1994).
18. Duwé, S. & Dedecker, P. Optimizing the fluorescent protein toolbox and its use. Curr. Opin.
Biotechnol. 58, 183–191 (2019).
19. Tsien, R. Y. The Green Fluorescent Protein. Annu. Rev. Biochem. 67, 509–544 (1998).
20. Patterson, G. H., Knobel, S. M., Sharif, W. D., Kain, S. R. & Piston, D. W. Use of the green
fluorescent protein and its mutants in quantitative fluorescence microscopy. Biophys. J. 73, 2782–
2790 (1997).
21. Heim, R. & Tsien, R. Y. Engineering green fluorescent protein for improved brightness, longer
wavelengths and fluorescence resonance energy transfer. Curr. Biol. 6, 178–182 (1996).
22. Ormö, M. et al. Crystal Structure of the Aequorea victoria Green Fluorescent Protein. Science
273, 1392–1395 (1996).
23. Matz, M. V. et al. Fluorescent proteins from nonbioluminescent Anthozoa species. Nat.
Biotechnol. 17, 969–973 (1999).
24. Wiedenmann, J., Ivanchenko, S., Oswald, F. & Nienhaus, G. Ulrich. Identification of GFP-like
Proteins in Nonbioluminescent, Azooxanthellate Anthozoa Opens New Perspectives for
Bioprospecting. Mar. Biotechnol. 6, 270–277 (2004).
25. Shaner, N. C. et al. Improved monomeric red, orange and yellow fluorescent proteins derived
from Discosoma sp. red fluorescent protein. Nat. Biotechnol. 22, 1567–1572 (2004).
27
26. Wang, L., Jackson, W. C., Steinbach, P. A. & Tsien, R. Y. Evolution of new nonantibody
proteins via iterative somatic hypermutation. Proc. Natl. Acad. Sci. 101, 16745–16749 (2004).
27. Rodriguez, E. A. et al. The Growing and Glowing Toolbox of Fluorescent and Photoactive
Proteins. Trends Biochem. Sci. 42, 111–129 (2017).
28. Shaner, N. C. et al. A bright monomeric green fluorescent protein derived from Branchiostoma
lanceolatum. Nat. Methods 10, 407–409 (2013).
29. Bajar, B. T. et al. Improving brightness and photostability of green and red fluorescent proteins
for live cell imaging and FRET reporting. Sci. Rep. 6, 20889 (2016).
30. Manna, P. et al. Directed evolution of excited state lifetime and brightness in FusionRed using
a microfluidic sorter. Integr. Biol. 10, 516–526 (2018).
31. Campbell, R. E. et al. A monomeric red fluorescent protein. Proc. Natl. Acad. Sci. 99, 7877–
7882 (2002).
32. Wiedenmann, J., Oswald, F. & Nienhaus, G. U. Fluorescent proteins for live cell imaging:
Opportunities, limitations, and challenges. IUBMB Life 61, 1029–1042 (2009).
33. Nienhaus, K. & Nienhaus, G. U. Fluorescent proteins for live-cell imaging with super-
resolution. Chem. Soc. Rev. 43, 1088–1106 (2014).
34. Wang, S., Moffitt, J. R., Dempsey, G. T., Xie, X. S. & Zhuang, X. Characterization and
development of photoactivatable fluorescent proteins for single-molecule–based superresolution
imaging. Proc. Natl. Acad. Sci. 111, 8452–8457 (2014).
35. Querard, J. et al. Photoswitching Kinetics and Phase-Sensitive Detection Add Discriminative
Dimensions for Selective Fluorescence Imaging. Angew. Chem. Int. Ed. 54, 2633–2637 (2015).
36. Lavis, L. D. Chemistry Is Dead. Long Live Chemistry! Biochemistry 56, 5165–5170 (2017).
37. Giepmans, B. N. G., Adams, S. R., Ellisman, M. H. & Tsien, R. Y. The Fluorescent Toolbox
for Assessing Protein Location and Function. Science 312, 217–224 (2006).
38. Panchuk-Voloshina, N. et al. Alexa dyes, a series of new fluorescent dyes that yield
exceptionally bright, photostable conjugates. J. Histochem. Cytochem. Off. J. Histochem. Soc. 47,
1179–1188 (1999).
39. Wysocki, L. M. & Lavis, L. D. Advances in the chemistry of small molecule fluorescent probes.
Curr. Opin. Chem. Biol. 15, 752–759 (2011).
40. Crissman, H. A. & Hirons, G. T. Chapter 13 Staining of DNA in Live and Fixed Cells. in Methods
in Cell Biology (eds. Darzynkiewicz, Z., Paul Robinson, J. & Crissman, H. A.) vol. 41 195–209
(Academic Press, 1994).
41. Mujumdar, R. B., Ernst, L. A., Mujumdar, S. R., Lewis, C. J. & Waggoner, A. S. Cyanine dye
labeling reagents: Sulfoindocyanine succinimidyl esters. Bioconjug. Chem. 4, 105–111 (1993).
42. Waggoner, A. [15] Covalent labeling of proteins and nucleic acids with fluorophores. in
Methods in Enzymology vol. 246 362–373 (Academic Press, 1995).
43. Lavis, L. D. & Raines, R. T. Bright Ideas for Chemical Biology. ACS Chem. Biol. 3, 142–155
(2008).
44. Grimm, J. B. et al. A general method to improve fluorophores for live-cell and single-molecule
microscopy. Nat. Methods 12, 244–250, 3 p following 250 (2015).
45. Los, G. V. et al. HaloTag: a novel protein labeling technology for cell imaging and protein
analysis. ACS Chem. Biol. 3, 373–382 (2008).
46. Grimm, J. B. et al. A general method to fine-tune fluorophores for live-cell and in vivo imaging.
Nat. Methods 14, 987–994 (2017).
47. Grimm, J. B. et al. A general method to optimize and functionalize red-shifted rhodamine dyes.
Nat. Methods 17, 815–821 (2020).
48. Keppler, A. et al. A general method for the covalent labeling of fusion proteins with small
molecules in vivo. Nat. Biotechnol. 21, 86–89 (2003).
49. Gautier, A. et al. An Engineered Protein Tag for Multiprotein Labeling in Living Cells. Chem.
Biol. 15, 128–136 (2008).
50. Griffin, B. A., Adams, S. R. & Tsien, R. Y. Specific covalent labeling of recombinant protein
molecules inside live cells. Science 281, 269–272 (1998).
51. Plamont, M.-A. et al. Small fluorescence-activating and absorption-shifting tag for tunable
protein imaging in vivo. Proc. Natl. Acad. Sci. 113, 497–502 (2016).
52. Li, C. et al. Dynamic multicolor protein labeling in living cells. Chem. Sci. 8, 5598–5605 (2017).
28
53. Brudler, R. et al. Coupling of Hydrogen Bonding to Chromophore Conformation and Function
in Photoactive Yellow Protein. Biochemistry 39, 13478–13486 (2000).
54. Boder, E. T. & Wittrup, K. D. Yeast surface display for screening combinatorial polypeptide
libraries. Nat. Biotechnol. 15, 553–557 (1997).
55. Boder, E. T. & Wittrup, K. D. Yeast surface display for directed evolution of protein expression,
affinity, and stability. in Methods in Enzymology (eds. Thorner, J., Emr, S. D. & Abelson, J. N.) vol.
328 430–444 (Academic Press, 2000).
56. Cherf, G. M. & Cochran, J. R. Applications of yeast surface display for protein engineering.
Methods Mol. Biol. Clifton NJ 1319, 155–175 (2015).
57. Li, C. et al. A Far-Red Emitting Fluorescent Chemogenetic Reporter for In Vivo Molecular
Imaging. Angew. Chem. Int. Ed. 59, 17917–17923 (2020).
58. Streett, H. E., Kalis, K. M. & Papoutsakis, E. T. A Strongly Fluorescing Anaerobic Reporter
and Protein-Tagging System for Clostridium Organisms Based on the Fluorescence-Activating and
Absorption-Shifting Tag Protein (FAST). Appl. Environ. Microbiol. 85, e00622-19 (2019).
59. Venkatachalapathy, M., Belapurkar, V., Jose, M., Gautier, A. & Nair, D. Live cell super
resolution imaging by radial fluctuations using fluorogen binding tags. Nanoscale 11, 3626–3632
(2019).
60. Smith, E. M., Gautier, A. & Puchner, E. M. Single-Molecule Localization Microscopy with the
Fluorescence-Activating and Absorption-Shifting Tag (FAST) System. ACS Chem. Biol. 14, 1115–
1120 (2019).
61. Monmeyran, A. et al. The inducible chemical-genetic fluorescent marker FAST outperforms
classical fluorescent proteins in the quantitative reporting of bacterial biofilm dynamics. Sci. Rep. 8,
10336 (2018).
62. Charubin, K., Modla, S., Caplan, J. L. & Papoutsakis, E. T. Interspecies Microbial Fusion and
Large-Scale Exchange of Cytoplasmic Proteins and RNA in a Syntrophic Clostridium Coculture. mBio
11, e02030-20 (2020).
63. Chekli, Y. et al. Visualizing the dynamics of exported bacterial proteins with the chemogenetic
fluorescent reporter FAST. Sci. Rep. 10, 15791 (2020).
64. Peron-Cane, C. et al. Fluorescent secreted bacterial effectors reveal active intravacuolar
proliferation of Listeria monocytogenes in epithelial cells. PLOS Pathog. 16, e1009001 (2020).
65. Li, C. et al. Fluorogenic Probing of Membrane Protein Trafficking. Bioconjug. Chem. 29, 1823–
1828 (2018).
66. Tebo, A. G. et al. Orthogonal fluorescent chemogenetic reporters for multicolor imaging. Nat.
Chem. Biol. 17, 30–38 (2020).
67. Tebo, A. G. & Gautier, A. A split fluorescent reporter with rapid and reversible
complementation. Nat. Commun. 10, 2822 (2019).
68. Tebo, A. G. et al. Circularly Permuted Fluorogenic Proteins for the Design of Modular
Biosensors. ACS Chem. Biol. 13, 2392–2397 (2018).
29
30
Chapter II
Fluorescent chemical-genetic hybrids
This bibliographic chapter has been published as a minireview in European Journal of Organic
Chemistry under the title:
Engineering glowing chemogenetic hybrids for spying on cells.

European Journal of Organic Chemistry 35, 5637–5646 (2020).
H. Ben Aissa & A. Gautier
This chapter aimed at highlighting the power of designing fluorogenic hybrid reporters for bioimaging
applications. Combining the precise localization of genetically-encoded tags in cells and the large
versatility of organic fluorophores, chemogenetic hybrid reporters offer new ways to image living
systems and sense biochemical events in cells, expanding the overall fluorescence toolbox.
II.1. Abstract
Seeing biochemical events is essential to understand the functioning of cells and organisms. Such
imaging depends on the selective targeting of fluorescent reporters and biosensors in living systems.
These reporters and biosensors can be entirely synthetic, entirely genetically encoded or hybrid.
Hybrid reporters and biosensors composed of a genetically encoded module that binds a fluorogenic
chromophore and activates its fluorescence have recently drawn attention as they offer new
opportunities for pushing the frontiers of bioimaging.
31
CHAPTER II: FLUORESCENT CHEMICAL-GENETIC HYBRIDS
II.2. Introduction
Biological imaging has become an essential tool in Biology. Fluorescent reporters enable to see
biochemical events inside living cells and organisms with high spatiotemporal resolution. The
revolution came with the development of genetically encoded reporters, which enables to easily and
specifically insert the instructions for their synthesis into cells in the form of DNA. Fused to a protein
of interest (or expressed in a particular cell type), genetically encoded fluorescent reporters enable
selective labeling, tracking and localization of proteins (or cells) in living systems1. The best-known
example of such reporters are fluorescent proteins derived from the green fluorescent protein (GFP)
discovered in Aequorea Victoria jellyfish and the red fluorescent protein (RFP) found in Discosoma
sp. mushroom anemone. Fluorescent proteins are structurally homologous and share the unique
property of forming in presence of molecular oxygen a fluorescent chromophore from three amino
acid residues within their own polypeptide sequence. Thanks to extensive engineering efforts, a broad
palette of fluorescent proteins with various colors and improved photophysical and photochemical
properties is now available2. The concomitant development of directed evolution techniques has been
essential for engineering fluorescent proteins that are more photostable and monomeric3,4.
The fluorescence toolbox has been recently expanded with the discovery and development of
alternative chemogenetic hybrid fluorescent labels. These systems open new prospects for biological
imaging as they circumvent some shortcomings of fluorescent proteins – such as their large size (25-
30 kDa), their slow fluorescence maturation and their need for molecular oxygen to fluoresce5. Among
these systems, some come from Nature such as the small oxygen-insensitive green fluorescent
protein UnaG, found in Japanese eel, which fluoresces thanks to the binding of bilirubin (BR), a
catabolic product of heme present in cells6. But most of them have been developed through
engineering. They are in general composed of a genetically encoded tag designed to form a
fluorescent complex with a small chromophore (Figure II.1). These chromophores are eventually
fluorogenic: essentially non-fluorescent when free in solution, they strongly fluoresce when bound to
their cognate tag. Fluorogenic chromophores are preferred as they provide higher contrast for imaging
than classical fluorophores. By engineering the tag through e.g. rational design, directed evolution or
de novo design, and/or the fluorogenic chromophores (also called fluorogens) by molecular
engineering, chemists and biologists have built and breed a large collection of hybrid chemogenetic
reporters allowing the efficient imaging of proteins – but also of other biomolecules such as RNA –
and the creation of actuators and biosensors for various applications in Biology.
32
Figure II.1. Tag and fluorogen engineering enable the generation of innovative fluorescent chemogenetic
reporters.
II.3. Hybrid reporters

II.3.1. Tag engineering
The design of fluorescent chemogenetic hybrid reporters has largely benefited from the development
of protein and nucleic acid engineering. Programming new functions in proteins or nucleic acids can
be done nowadays using various approaches based either on rational design, directed evolution or
de novo design. Rational design can be efficient in reprogramming the function of an existing protein
using structural information and site-directed mutagenesis. Directed evolution is useful for improving
properties or programing new functions in a given protein or nucleic acid scaffold by mimicking the
strategy employed by Nature to create new functions. De novo design of proteins on the other hand
is a powerful and attractive alternative, as it allows the building of proteins with predetermined
structures and functions through exploration of the full sequence space in silico.
33
II.3.1.1. Rational design
The engineering of alternative fluorescent proteins from natural photoreceptor nicely illustrates the
power of rational design (Figure II.1). Photoreceptors allow organisms to use light as source of energy
or as information to trigger a biological response. They include a prosthetic chromophore (e.g. flavins,
bilins, retinal) that initiates a photochemical reaction upon light absorption. Most photoreceptors have
evolved to maximize the efficiency of their photocycle, and thus barely fluoresce. However,
introduction of variations in the backbone of photoreceptors can impair their natural photocycle and
enhance their fluorescence emission.
This strategy allowed the development of fluorescent reporters from LOV (light, oxygen, voltage)
sensing domains found in plants, algae and bacteria. Upon blue light excitation, a covalent adduct is
formed in these flavoproteins between a flavin mononucleotide (FMN) chromophore and a conserved
cysteine inducing a conformational change and thus a biological response7. Mutation of the conserved
cysteine to alanine inhibits their natural photocycle and maximizes their fluorescence (Figure II.2).
This allowed the generation of flavin-based fluorescent proteins (FbFP) and iLOV from bacterial and
plant LOV domains8,9. FbFPs and iLOV fluoresce cyan-green light in response to blue light excitation
thanks to the non-covalently bound FMN. Using comparative photobleaching measurements in E. coli
colonies, a photostable iLOV variant was generated by directed evolution. Crystallographic analysis
revealed that mutations close to the FMN-binding pocket rigidify the chromophore structure inside the
cavity, providing greater resistance to photobleaching10. Thermostable FbFPs with superior
photostability were also evolved from thermophilic bacteria11. Interestingly, as FMN is endogenously
present and its biosynthesis does not require oxygen, FbFPs allowed the study of complex anaerobic
processes in bacteria, fungi and mammalian cells8,12–14. The shorter size of FbFPs and iLOV (ranging
typically from 10 to 15 kDa) is a clear advantage over GFP-like fluorescent proteins. The reduced
genetic footprint of iLOV allowed for instance the study of virus infection and spread in plants9.
34
Figure II.2. Flavin-based fluorescent proteins (FbFPs) were engineered from LOV domains. Mutation of a
conserved cysteine to alanine inhibits the natural photocycle and maximizes the fluorescence of FMN.
Rational design combined with random mutagenesis have been also essential in the engineering of
far-red and near-infrared fluorescent proteins from phytochromes and their analogs. Phytochromes
are present in algae, fungi, plant, bacteria and cyanobacteria. They sense far-red light through
photoisomerization of a covalently attached biliverdin (BV) chromophore and function as light driven
signal transducers 15. Biliverdin is a catabolic product of heme and is endogenously present in cells.
In a seminal paper, Tsien and co-workers showed that suppression of the PHY domain of the
bacteriophytochrome of Deinococcus radiodurans and introduction of key mutations in the immediate
environment of the chromophore in the remaining PAS-GAF domains prevented BV
photoisomerization, locking it into an emissive conformation (Figure II.3). The resulting fluorescent
protein, named IFP1.4, fluoresces above 700 nm16. Such system opened new exciting prospects for
deep-tissue and whole-body imaging because of the higher transparency, lower light scattering, and
lower autofluorescence of mammalian tissues in the near-infrared region. Massive engineering effort
using rational design (but also directed evolution) has been done, using phytochromes from different
organisms, to obtain far-red and near-infrared fluorescent proteins that are bright and photostable,
incorporate efficiently BV and function as monomer17–20. The monomeric GAF domain of
cyanobacteriochrome (CBRC) (found in cyanobacteria) was recently engineered to efficiently attach
biliverdin (BV) rather than its usual phycocyanobilin (PCB) chromophore, giving eventually
miRFP670nano, a small monomeric far-red fluorescent proteins of 17 kDa that efficiently incorporates
BV at endogenous levels21.
35
Figure II.3. Far-red and near-infrared fluorescent proteins were engineered from phytochromes and their
analogs. Introduction of key mutations prevents biliverdin photoisomerization, locking it into an emissive
conformation.
Rational design can also be used elegantly to incorporate synthetic fluorogenic chromophores into
natural protein scaffolds. The 15 kDa cellular retinoic acid binding protein II CRABPII was engineered
to incorporate a merocyanine dye precursor. This dye precursor exhibits no fluorescence when free
and becomes fluorescent upon formation of a covalent adduct between its aldehyde group and a
lysine residue rationally introduced in the chromophore binding site (Figure II.4). The resulting
iminium-containing merocyanine exhibits far-red emission with high quantum efficiency and
brightness. Introduction of a large hydrophobic residue (Trp) and a negatively charge residue (Glu or
Asp) close to the binding pocket improved the fluorescent quantum yield and brightness.
Crystallographic structures of the protein:merocyanine complex revealed a relatively well packed and
restricted chromophore within the binding cavity compared to the original protein22. For efficient use
in mammalian and yeast cells, human cellular retinol-binding protein II (hCRBPII) was recently
engineered in a similar fashion to covalently bind merocyanine aldehyde chromophore. Incorporation
of a catalytic acid residue in the chromophore-binding pocket activates the precursor’s carbonyl,
favoring the nucleophilic attack of the reactive lysine residue23.
36
Figure II.4. CRABPII was engineered to incorporate a merocyanine dye precursor, which exhibits no
fluorescence when free and becomes fluorescent upon formation of a covalent adduct between its aldehyde
group and a lysine residue rationally introduced in the chromophore binding site.
II.3.1.2. Directed evolution
When rational design cannot be used, directed evolution can be a very efficient way to create a
complementary binding scaffold able to activate the fluorescence of a fluorogenic chromophore
(Figure II.1). A landmark study is the development of Fluorogen Activating Proteins (FAPs). FAPs
are single chain antibodies (scFvs) able to bind non-covalently and activate the fluorescence of
thiazole orange (TO) and malachite green (MG) derivatives (Figure II.5), two molecular rotors known
to dissipate light energy radiatively when conformationally immobilized24,25. Human scFvs were
chosen as binding scaffolds because of their relatively small size (< 30kDa), their ability to recognize
a wide range of antigens and for their ability to act as genetic fusion tags. FAPs were identified
screening large yeast-displayed libraries of human scFvs by fluorescence activated cell sorting26. This
strategy was expanded to several fluorogenic dimethylindole red (DIR) derivatives to identifying FAPs
that cover most of the visible and near IR region of the spectrum27. The use of FAPs was originally
restricted to image proteins at the cell surface or in non-reducing environments such as the secretory
pathway, because scFv required the formation of conserved disulfide bonds for proper folding, stability
and function. Directed evolution and activity-based selection allowed the development of a FAP
variant containing a single point mutation that restored the activity lost in the disulfide-bond free
mutant and that showed intracellular activity28. FAPs display reduced photobleaching because of
fluorogen renewal, opening great prospects for live-cell super-resolution microscopy. Highly
photostable FAP has been successfully applied for tracking proteins in live bacterial cells using
stimulated emission depletion (STED) microscopy29. FAPs also allowed single molecule tracking in
live cells labeling only a subpopulation of proteins with low concentrations of fluorogen30.
Directed evolution using yeast display and fluorescence activated cell sorting is a general method for
engineering fluorogen-activating scaffolds. Recently, this approach was employed to engineer the
Fluorescence Activating and absorption Shifting Tag (FAST), a small protein of 14 kDa that forms
non-covalent fluorescent complexes with fluorogenic hydroxybenzylidene rhodanine (HBR)
analogues31 (Figure II.5). Composed of an electron-donating phenol ring conjugated with an electron-
withdrawing rhodanine heterocycle, these push-pull fluorogens dissipate light energy non-radiatively
37
in solution but strongly fluoresce when they are conformationally locked within the cavity of FAST.
FAST was evolved from the Halorhodospira halophila Photoactive Yellow Protein (PYP), a natural
photoreceptor sensing blue light through photoisomerization of a covalently attached
32
hydroxycinnamoyl chromophore . Taking advantage of the common structural features of PYP’s
natural chromophore and HBR analogs, PYP binding pocket was repurposed to bind HBR analogs.
FAST allowed the selective labeling of fusion proteins in various localizations in living cells such as
bacteria, yeast, mammalian cells and in multicellular organisms like zebrafish embryos31. Because of
its oxygen-independence, FAST allowed the efficient imaging of proteins in strict anaerobe
Clostridium species33 and the study of bacterial biofilm dynamics in low-oxygen environments34.
Moreover, as fluorogens bind non-covalently, labeling can be easily reversed by washing, and
fluorogen is continuously recycled reducing the apparent photobleaching rate35. As only a subset of
proteins can be labeled, FAST can be used to image proteins below the diffraction limit using super-
resolution imaging by radial fluctuations (SRRF)36, which is sensitive to fluorophore density, or by
single-molecule localization microscopy (SMLM)7, which necessitates the stochastic generation of
sparse subsets of emitters over time.
Figure II.5. Directed evolution can be used to generate (A) complementary binding scaffolds able to activate
the fluorescence of a fluorogen. (B) Examples of synthetic fluorogens.
38
Directed evolution was also essential for the development of RNA aptamers able to bind and activate
the fluorescence of fluorogenic chromophores. Such aptamers open new ways to image RNA in live
cells. The design of RNA aptamers binding selectively triphenylmethane dyes such as malachite
green opened this field of research. Although these systems showed intrinsic cytotoxicity, making
them unsuitable for in vivo imaging38, this first generation of RNA aptamers38–40 set up the basis for
the design of effective fluorogenic RNA aptamers. In 2011, an RNA aptamer named Spinach was
successfully designed to switch on the fluorescence of the 3,5-difluoro-4-hydroxybenzylidene
imidazolinone (DFHBI)41 (Figure II.5). Mimick of the chromophore found in GFP, DFHBI is barely
fluorescent in solution but exhibits fluorescence enhancement when encased into Spinach. Selected
by systematic evolution of ligands by exponential enrichment (SELEX), Spinach is a RNA aptamer in
which DFHBI chromophore is stacked, preventing its intramolecular motions, thus enhancing the
fluorescence. In the original Spinach aptamer, the anionic form of the chromophore is embedded on
top of a three-tetrad quadruplex forming the aptamer’s core, which is stabilized by the presence of
cations42–44. Crystallographic studies have permitted to engineer variants with improved folding and
thermal stability such as Spinach245, iSpinach46 and baby Spinach, a miniaturized Spinach aptamer43.
Whereas the Spinach family was engineered on the basis of the affinity for DFHBI, SELEX coupled
to fluorescence activated cell sorting (FACS) permitted to identify the Broccoli aptamer47. In presence
of DFHBI analogues, Broccoli exhibited improved thermostability, robust folding and reduced
magnesium dependence compared to Spinach, which overall increased its effective brightness48.
A similar strategy was employed for engineering a photostable small RNA aptamer called Corn that
binds the 3,5-difluoro-4-hydroxybenzylideneimidazolinone-2-oxime (DFHO) fluorophore, an analogue
of the red fluorescent protein (RFP) chromophore (Figure II.5). Corn enabled direct quantitative
measurement of RNA levels when used as a RNA-based reporter for transcription49. Corn was shown
however to form a dimer, which is the only form that binds DFHO, making it unsuitable for tagging
and imaging mRNAs in cells.
The analogs of GFP or RFP chromophores are not the only fluorogens that have been used for the
design of fluorogenic aptamers. Various other classes were used. Using stringent selection
conditions, the RNA aptamer Mango was selected to bind Thiazole Orange derivatives with
nanomolar affinities (Figure II.5). The visualization of Mango in Caenorhabditis elegans gonads
demonstrated its potential for live-cell RNA imaging50. Microfluidic droplet-based functional selection
was performed to select Mango variants with improved fluorescent brightness51. SELEX was also
recently used to select Pepper, an aptamer binding derivatives of HBC (4-((2-
hydroxyethyl)(methyl)amino)-benzylidene)-cyanophenylacetonitrile), synthetic dyes with structurally
rigid electron acceptors and strong electron donors that strongly fluoresce upon constraining
intramolecular motion52 (Figure II.5). Pepper allowed the study of intracellular noncoding RNAs and
39
mRNAs and to label endogenous chromosomal loci. Furthermore, displaying high brightness and
photostability, Pepper was shown to be well suited for live-cell imaging of RNA aptamers beyond the
diffraction limit.
II.3.1.3. In silico and de novo design
Protein engineering can benefit from the use of in silico methods. In silico mutagenesis of residues
within the ligand binding pocket of E. coli lipocalin (Blc) allowed computational screening of binders
for a library of aminated GFP-like chromophores. The dynamic exchange of the fluorogen dye was
shown to increase the photostability of the protein-fluorogen pairs. These systems allowed single
molecule localization microscopy using low dye concentrations53.
The design of fluorogen-activating proteins have recently taken a new turn with the use of de novo
protein design, which facilitates and reduces the long process of protein engineering based on natural
proteins54 (Figure II.1). De novo design was used to create a small β-barrel which binds DFHBI,
analog of the GFP chromophore (vide supra). Candidates were identified using computational
screening method that combined a fixed backbone design (surrounding the fluorogen binding site)
and a flexible-backbone design (for the rest of the protein). Hierarchical grid-based search method
allowed the screening of sequences, in which each single amino acid was randomly mutated, for
fluorescence activation and proper folding in presence of DFHBI. Crystallographic structures revealed
that stable monomeric variants of 110 amino acids could activate DFHBI fluorescence by stabilizing
it in the binding pocket through multiple hydrogen bonds. Although they display weak brightness,
these mini-fluorescence-activating proteins (mFAP) could be imaged in bacterial, yeast and
mammalian cells by conventional widefield and confocal microscopy. mFAPs constitute the first
generation of fluorogen-based reporters fully engineered by computational design, and considerable
space for improvement remains possible. The use of de novo design for the generation of fluorogen-
based reporters and biosensors is just in its infancy, and there is no doubt that computational
approaches will make significant contributions in the future for the design of efficient reporters and
biosensors55.
II.3.2. Fluorogen engineering
Semi-synthetic hybrid systems being composed of a synthetic fluorogenic part, it is possible to benefit
from the power of molecular engineering to create, tune or optimize hybrid fluorescent reporters. High
contrast imaging requires both high fluorogenicity and selective activation, two parameters that can
be engineered using modern organic chemistry (Figure II.1). Molecular engineering offers also new
possibilities for tuning various properties such as cellular uptake or spectral characteristics.
40
II.3.2.1. Design
In the systems described so far, fluorogenicity arose from immobilization of molecular rotors or push-
pull systems (Figures II.5). These fluorogens are essentially non-fluorescent in solution because they
dissipate light energy through efficient non-radiative processes. However, when locked in a planar
conformation in a well fitted cavity, these processes are slow down and fluorescence becomes an
efficient energy dissipation mechanism. These systems are however not the only ones that can be
used to generate fluorogenic hybrid systems. Other mechanisms can be used to switch on
fluorescence through molecular interaction.
The development of fluorogenic rhodamine derivatives nicely illustrates how molecular engineering
allows the fine tuning of fluorogenicity. Rhodamines can exist as a non-fluorescent spirolactone closed
form or as a fluorescent zwitterionic opened form (Figure II.6). Creatively exploited to design
photoactivatable and blinking fluorescent dyes for super-resolution imaging56,57,58, this unique property
can be used for the design of fluorogenic rhodamines. In particular silicon rhodamines were shown to
exist mainly in their non-fluorescent spirolactone closed form in aqueous solution, but to adopt the
fluorescent zwitterionic opened form (that displays excitation and emission maxima in the far-red and
near-infrared region) in less polar environments. This property allowed the development of a collection
of fluorogenic substrates for protein labeling, exploiting the fact that a protein environment is less
polar than aqueous solution. SiR derivatives were coupled to substrates for self-labeling tags such as
SNAP-tag, CLIP-tag and HaloTag59–61, allowing fluorogenic labeling of fusion proteins in live cells
(Figure II.6). SiR fluorophores have been shown to be well suited for protein labeling in cells and
multicellular organisms, multi-color imaging, and super-resolution imaging. RNA aptamers were
evolved to recognize SiR and photostable carborhodamine fluorophores with nanomolar affinities.
Thanks to their far-red shifted fluorescence, these RNA aptamers were successfully applied to
visualize mRNAs in live bacteria using STED microscopy62. The use of four-membered azetidine ring
allowed furthermore the generation of fluorogenic rhodamine analogs with higher photostability and
brightness63,64. The lactone-zwitterion equilibrium constant enables to predict the fluorogenicity of
rhodamines, providing a quantitative framework for the design of new fluorogenic dyes with optimal
fluorogenicity58. Recently, a general method has been developed to transform regular rhodamine and
related fluorophores into fluorogenic chromophores with an excellent cell-permeability. Converting
their carboxyl group into an electron-deficient amide enabled to tune the dynamic equilibrium between
the fluorescent zwitterion and the non-fluorescent, cell-permeable spirolactam, allowing the
generation of fluorogens that light up upon interaction with their target. Using this design, fluorogens
with various spectral properties could be designed for wash-free, multi-color, live cell nanoscopy65.
41
Figure II.6. Fluorogenic rhodamines. Rhodamines can exist mainly in their non-fluorescent spirolactone closed
form in aqueous solution, but adopt their fluorescent zwitterionic opened form when bound to their target tag.
Fluorogenic substrates were also designed relying on intramolecular quenching. This was originally
proposed for the labeling of the PYP-tag, which can react in a covalent fashion with hydroxycinnamoyl
and coumarin thioester derivatives. Fluorogenicity was obtained using substrates composed of a
fluorophore linked to a quencher. Reaction with PYP leads to fluorophore unquenching and restores
fluorescence66–68. The approach is versatile and allowed one to change the spectral properties by
changing the fluorophore. Multicolor imaging using this approach enabled to elucidate the role of the
GLUT4 glycosylation in living cells69. For efficient use in live cells, highly cell-permeant and reactive
substrates were designed using smaller polarity-sensitive coumarins70. In highly polar environment,
coumarin ligand adopts a twisted intramolecular charge transfer (TICT) non-fluorescent state.
Whereas in lower polar environment, as PYP’s binding pocket, the planar conformation of the probe
induced a high fluorescence exaltation. A combined mutational and chemical approach was used to
create electrostatic interactions between the substrate and the tag, further increasing the temporal
resolution of the labeling and the brightness71. Introduction of a hydroxyl group into coumarin ligand
enabled to increase the labeling reaction rate and the OFF-ON fluorescence ratio72. This probe
allowed the observation of the transient nuclear localization of the mitochondrial deacetylase sirtuin-
3.
Intramolecular quenching was also creatively used for the design of fluorogenic probes for RNA
labeling. The aptamer Riboglow was developed from a short riboswitch sequence that recognizes
cobalamin73. Cobalamin can efficiently quench an attached fluorophore. Binding to the riboswitch
shields the fluorophore from cobalamin, resulting thus in high fluorescence activation. Riboglow
allowed the efficient imaging of RNA Pol II transcripts. Similarly, was developed Gemini561, a
fluorogen composed of two molecules of sulforhodamine B that self-quench in solution but strongly
fluoresce when they interact with the RNA aptamer oCoral (Figure II.7). This ultrabright aptamer,
selected by SELEX and microfluidic-assisted in vitro compartmentalization, was shown to display
unprecedented photostability compared to Broccoli, MangoIII and Corn aptamers74.
42
Figure II.7. Example of fluorogens based on intramolecular unquenching. Gemini561 is a fluorogen composed
of two molecules of sulforhodamine B that self-quench in solution but strongly fluoresce when they interact with
the RNA aptamer oCoral.
II.3.2.2. Tuning
Molecular engineering can also creatively be used to tune fluorogen properties, such as cellular
uptake. Although highly membrane-permeant fluorogens are preferred for efficient intracellular protein
labeling, membrane-impermeant fluorogens can restrict labeling to proteins located at the cell surface.
The use of membrane-impermeant fluorogens allowed the monitoring of dynamic processes such as
membrane protein trafficking, internalization, recycling or endocytosis. Membrane impermeant
malachite green (MG) derivatives enabled for instance to selectively track FAP-tagged membrane
receptors at the cell surface and to investigate the role of G-protein coupled receptors trafficking in
live cells. Addition of amphiphilic polyethyleneglycol linkers or negatively charged sulfonate groups
reduces their cellular uptake (Figure II.8), limiting their action to the plasma membrane only75,76. A
sulfonated MG analog combined with a FAP-tagged anti-EGFR affibody was recently used for tumor
43
visualization in mice77. Similarly, membrane impermeant HBR derivatives containing a negatively

charged carboxymethyl group on the rhodanine head allowed the selective labeling of FAST-tagged
proteins at the cell surface (Figure II.8), enabling to easily quantify the levels of cell-surface proteins
by various fluorometric techniques78.
Figure II.8. Membrane-impermeant fluorogens for FAP and FAST were generated by addition of negatively
charged groups.
Molecular engineering is also efficient in tuning the spectral properties of fluorogens. Structural
modifications can induce significant changes in the electronic properties of fluorogens, and thus their
absorption and emission properties. The spectral properties of many fluorogenic reporters have been
successfully tuned by molecular engineering, allowing with a single tag to have several colors,
providing a versatility that is not encountered with fluorescent proteins. The use of electron-donating
thiophene aryl groups has permitted to extend the spectral properties of FAP:MG analogs to the red79.
Introduction of electron-donating groups on the aromatic ring of HBR analogs allowed the design of
FAST fluorogens with red-shifted absorption and emission properties for applications in multicolor
imaging80. The ability to change color by fluorogen exchange allowed dynamic color switching in cells,
providing a unique kinetic signature that could be advantageously exploited for highly multiplexed
imaging80. By extending the length of the conjugated p-system, in addition to the introduction of
electron-withdrawing or donating heteroatom groups, a panel of GFP-like aminated chromophores
derivatives enabled to red shift the spectral properties of the lipocalin-based system81. The spectral
properties of Spinach could be tuned by chemical modification of GFP-like chromophores.
Substitutions on the imidazolone ring of DFHBI permitted to optimize the fluorescence of the Spinach
RNA and its derivatives to the optical filters commonly used in fluorescence microscopy82. Molecular
engineering also enabled to tune the spectral properties of the RNA aptamer Pepper from cyan to red
by tuning the aromatic p-structure and adjusting the electron donor/acceptor capability of the dye52.
Similarly, the spectral properties of rhodamine-based fluorogens could be varied by changing the
64 65
bridge in the xanthene moiety . Finally, molecular engineering can extend the spectral and
44
photophysical properties of natural fluorogens. Computational studies based on FMN chromophore

predicted theoretically the structures of derivatives with red-shifted emission of fluorescence covering
spectral regions of 500 up to 860 nm83.
II.4. Hybrid actuators and biosensors

II.4.1. Actuators
Beyond offering new opportunities in biological imaging, the development of chemogenetic hybrid
systems opens also new prospects for the design of actuators enabling to study biological systems in
new ways. A nice illustration of the potential of chemogenetic hybrids is the development of reactive
oxygen species (ROS) producers. ROS are known to cause oxidative damage to DNA, lipids and
proteins in living cells. They are naturally involved in many signaling processes (e.g. autophagy,
apoptosis and aging). ROS generators can be used for photoinducible protein inactivation or cell
ablation, in photodynamic therapy for targeting cancer cell lines or in electron microscopy (EM) for
local synthesis of osmiophilic products resolvable by EM. A genetically encoded photosensitizer was
engineered from the LOV2 domain of Arabidopsis thaliana phototropin 2, which naturally binds FMN
(vide supra). Directed evolution using IFP1.4 photobleaching as an indicator for ROS generation
allowed the generation of the mini Singlet Oxygen Generator (miniSOG). Crystallographic studies
showed an increase in rigidity of the chromophore’s environment compared to wild-type LOV2
suggesting that oxygen diffusion can occur properly84. When irradiated with blue light, miniSOG
fluoresces and generates both superoxide anion (and its derivatives) and singlet oxygen (1O2) through
type I and II mechanisms85,86. This dual function allowed the imaging of miniSOG by correlative light
and electron microscopy (CLEM) in Caenorhabditis elegans and mice synaptic cells87. Protein
engineering enabled to enhance 1O2 generation by mutating residues responsible for hydrogen bonds
and electron-donating transfer between the chromophore and the protein scaffold88,89. Engineered
miniSOG variant with improved ROS generation has been successfully used to induce specific cell
ablation in Drosophila melanogaster adult fly90.
A far-red fluorogenic photosensitizer was recently developed from the FAP family for applications in
deep tissues. Malachite green was substituted with heavy iodine atoms to increase intersystem
crossing and thus ROS production (Figure 9). The resulted system FAP-TAPs (Targeted and
Activated Photosensitizer) fluoresces and produces singlet oxygen only when the exogenous dye is
associated to FAP. Due to the very short excited state lifetime of free fluorogen, reactive oxygen
species appear exclusively upon binding to the encasing protein which makes the system a unique
‘on-demand’ genetically encoded photosensitizer. This system allowed deep cell ablation in adult
zebrafish and subcellular protein inactivation in living cells91.
45
Figure II.9. Targeted and Activated Photosensitizer (TAP). Malachite Green with heavy iodine atoms forms
with FAP/TAP a complex producing ROS because of increased intersystem crossing.
II.4.2. Biosensors
As alternative to fluorescent proteins, fluorescent chemogenetic reporters can be creatively used for
the design of various types of biosensors enabling to visualize intracellular events in living cells with
high resolution in space and time. Biosensors can be generated by taking advantage of the intrinsic
advantages of fluorescent chemogenetic reporters.
As all GFP-like fluorescent proteins depend on the presence of O2 for chromophore maturation, O2-
insensitive fluorogenic systems allow the creation of genetically encoded biosensors measuring O2
levels in living cells. The oxygen-insensitive FbFP allowed the design of a FRET-based oxygen
biosensor named FluBO. Oxygen variation was monitored in real-time in E. coli by measuring the
FRET efficiency between FbFP donor and the hypoxia-sensitive yellow fluorescent protein (YFP)92.
Reoxygenation of cells was also visualized by controlling the expression of the oxygen-insensitive
UnaG with an hypoxia responsive promoter93. Addition of a disulfide bridge to the β-barrel surface of
UnaG allowed moreover the generation of a redox-sensitive fluorescent sensor with high dynamic
range, enabling to monitor redox changes under hypoxia in living cells94.
Modification of the fluorogenic chromophore through molecular engineering is another way to

generate biosensors. FAPs were for example used to create a surface-selective pH biosensor using
pH-sensitive cyanine analogues covalently attached to thiazole orange or malachite green. Change
in FRET efficiency induced by pH change allowed the study of receptor-mediated endocytosis or
receptor recycling95,96. A ratiometric pH sensor was also engineered from hCRBPII using a non-
permeant julolidine retinal analog. The presence of a titratable functional group close to the
chromophore binding pocket renders both the absorption and fluorescence emission sensitive to pH
change97.
Fluorescent chemogenetic reporters can also be used creatively to generate new types of biosensors.
By coupling chemogenetic reporters with a sensing domain, it is possible to condition the fluorogen
46
binding (and thus fluorescence) to the recognition of the analyte. Insertion of Calmodulin (CaM), a
well-established Ca2+ sensing domain, within UnaG permitted to develop BReleaCa, a calcium dual-
ligand fluorogenic biosensor. BReleaCa releases bilirubin (BR) upon calcium binding, leading to a
loss of fluorescence98. A Ca2+ biosensor with a more favorable positive response was developed
coupling a circularly permuted version of FAST to CaM and the CaM-binding peptide of myosin light-
chain kinase (M13). In this sensor, calcium binding promotes fluorogen binding, and thus fluorescence
increase, enabling to image Ca2+ levels in cells99. Sensors can also be built by conditioning the
fluorogen binding to a given signal. An infrared fluorescent protease reporter was engineered from a
circularly permuted IFP1.4 so that the biliverdin chromophore incorporation is regulated by protease
activity. The ‘inactivated’ sensor iProtease is not fluorescent because of a hidden cysteine residue
essential for BV attachment. Proteolytic activation frees the cysteine residue which returns to the
binding cavity, thus promoting fluorescence by proper BV incorporation. This system allowed the
development of iCasper, an apoptosis cellular sensor which fluoresces under the activation of
caspase-3100.
Fluorescent chemogenetic reporters offer also great opportunities for the design of split fluorescent
reporters for the detection of protein-protein interactions (PPIs). The protein module can be split into
two complementary modules, which can assemble when fused to two interacting proteins, promoting
fluorogen binding and thus fluorescence activation. Such complementation system was designed by
splitting FAST in two complementary fragments. The resulting splitFAST was shown to allow the
visualization of PPIs in living cells. The complementation of splitFAST was demonstrated to be rapid
and reversible, enabling to study protein complex association and dissociation almost in real-time,
which has not been achievable using splitGFP derivatives101. Other fluorogenic split systems working
on the same principle were engineered from UnaG102 and the biliverdin-based far-red fluorescent
proteins103–105. Recently, a split version of miniSOG was designed to track PPIs by electron
microscopy (EM). Protein-protein interactions allowed the generation of reactive oxygen species from
split-miniSOG which in combination with diaminobenzidine produced a localized polymeric precipitate
providing high contrast in EM in presence of osmium106.
II.5. Conclusion
Hybrid reporters have recently opened new ways to image living cells and sense biochemical events.
Benefiting from numerous assets, as their shorter size, their O2 independence, their on-demand
fluorescence, fluorogenic systems present advantages over the widely used autofluorescent proteins.
As hybrid reporters are composed of a protein or an RNA tag and a fluorogenic chromophore, the
development and improvement of such systems can be done through engineering both components.
47
Engineering the scaffold by rational design and directed evolution permit to generate scaffolds able
to activate the fluorescence of natural or synthetic fluorogens and significantly enhance their
performance. The use of de novo design opens more over great prospect for further development
and/or improvement of custom fluorogenic systems. By reducing the duration of the long-process of
directed evolution (and other related methods), de novo design represents a promising way to design
fluorogenic reporters from scratch. Moreover, de novo design should greatly facilitate the
development of chemogenetic biosensors by speeding up the initial design and reducing the duration
of the optimization steps.
The possibility to create, tune or optimize hybrid chemogenetic reporters through molecular
engineering of the fluorogenic chromophores represents moreover a major advantage over
fluorescent proteins. While fluorescent proteins cannot be tuned as desired, as their fluorescence is
restricted to their embedded genetically encoded chromophore, the use of synthetic fluorogens
permits to finely tune the properties of chemogenetic systems. Various fluorescence activation
mechanisms can be used, and a broad palette of colors can be obtained. Beyond creating fluorogens
with various photophysical properties, molecular engineering can also be used to increase or reduce
fluorogen cellular uptake or to turn fluorogens into ROS generators for various applications in cell
biology.
The development and use of fluorogenic reporters and biosensors is still in its infancy. We are
convinced that hybrid systems will allow new discoveries in Biology from fundamental mechanisms to
the causes of disease, and the development of novel diagnostic tools and therapeutics. There is no
doubt that the continuous development of innovative and sophisticated imaging techniques (e.g.
tissue clearing, expansion microscopy) and modalities (e.g. correlative light and electron microscopy,
photoacoustic imaging) will motivate developers to invent fully adapted chemogenetic reporters and
biosensors in order to make the most of these technologies and push forward the frontiers of biological
imaging.
Notes
The authors declare the following competing financial interest: A.G. is co-founder and hold equity in
Twinkle Bioscience / The Twinkle Factory, a company commercializing the FAST technology.
Acknowledgements
This work was supported by the European Research Council (ERC-2016-CoG-724705
FLUOSWITCH).
Keywords: Fluorescent labels and biosensors • chemogenetic reporters • fluorescence imaging
48
II.6. References
2. Tsien, R. Y. Building and breeding molecules to spy on cells and tumors. FEBS Lett. 579,
927–932 (2005).
3. Bajar, B. T. et al. Improving brightness and photostability of green and red fluorescent proteins
for live cell imaging and FRET reporting. Sci. Rep. 6, 20889 (2016).
4. Shaner, N. C. et al. A bright monomeric green fluorescent protein derived from Branchiostoma
lanceolatum. Nat. Methods 10, 407–409 (2013).
5. Chudakov, D. M., Matz, M. V., Lukyanov, S. & Lukyanov, K. A. Fluorescent Proteins and Their
Applications in Imaging Living Cells and Tissues. Physiol. Rev. 90, 1103–1163 (2010).
6. Kumagai, A. et al. A Bilirubin-Inducible Fluorescent Protein from Eel Muscle. Cell 153, 1602–
1611 (2013).
7. Losi, A., Polverini, E., Quest, B. & Gärtner, W. First Evidence for Phototropin-Related Blue-
Light Receptors in Prokaryotes. Biophys. J. 82, 2627–2634 (2002).
8. Drepper, T. et al. Reporter proteins for in vivo fluorescence without oxygen. Nat. Biotechnol.
25, 443–445 (2007).
9. Chapman, S. et al. The photoreversible fluorescent protein iLOV outperforms GFP as a
reporter of plant virus infection. Proc. Natl. Acad. Sci. U. S. A. 105, 20038–20043 (2008).
10. Christie, J. M. et al. Structural tuning of the fluorescent protein iLOV for improved
photostability. J. Biol. Chem. 287, 22295–22304 (2012).
11. Wingen, M., Jaeger, K.-E., Gensch, T. & Drepper, T. Novel Thermostable Flavin-binding
Fluorescent Proteins from Thermophilic Organisms. Photochem. Photobiol. 93, 849–856 (2017).
12. Drepper, T., Gensch, T. & Pohl, M. Advanced in vivo applications of blue light photoreceptors
as alternative fluorescent proteins. Photochem. Photobiol. Sci. 12, 1125–1134 (2013).
13. Mukherjee, A. & Schroeder, C. M. Flavin-based fluorescent proteins: emerging paradigms in
biological imaging. Curr. Opin. Biotechnol. 31, 16–23 (2015).
14. Wingen, M. et al. The photophysics of LOV-based fluorescent proteins – new tools for cell
biology. Photochem. Photobiol. Sci. 13, 875–883 (2014).
15. Baloban, M. et al. Designing brighter near-infrared fluorescent proteins: insights from structural
and biochemical studies. Chem. Sci. 8, 4546–4557 (2017).
16. Shu, X. et al. Mammalian expression of infrared fluorescent proteins engineered from a
bacterial phytochrome. Science 324, 804–807 (2009).
17. Filonov, G. S. et al. Bright and stable near-infrared fluorescent protein for in vivo imaging. Nat.
Biotechnol. 29, 757–761 (2011).
18. Shcherbakova, D. M. & Verkhusha, V. V. Near-infrared fluorescent proteins for multicolor in
vivo imaging. Nat. Methods 10, 751–754 (2013).
19. Yu, D. et al. An improved monomeric infrared fluorescent protein for neuronal and tumour
brain imaging. Nat. Commun. 5, 3626 (2014).
20. Rodriguez, E. A. et al. A far-red fluorescent protein evolved from a cyanobacterial
phycobiliprotein. Nat. Methods 13, 763–769 (2016).
21. Oliinyk, O. S., Shemetov, A. A., Pletnev, S., Shcherbakova, D. M. & Verkhusha, V. V. Smallest
near-infrared fluorescent protein evolved from cyanobacteriochrome as versatile tag for spectral
multiplexing. Nat. Commun. 10, 279 (2019).
22. Yapici, I. et al. “Turn-On” Protein Fluorescence: In Situ Formation of Cyanine Dyes. J. Am.
Chem. Soc. 137, 1073–1080 (2015).
23. Santos, E. M. et al. Engineering of a Red Fluorogenic Protein/Merocyanine Complex for Live-
Cell Imaging. ChemBioChem 21, 723–729 (2020).
24. Nygren, J., Svanvik, N. & Kubista, M. The interactions between the fluorescent dye thiazole
orange and DNA. Biopolymers 46, 39–51 (1998).
25. Babendure, J. R., Adams, S. R. & Tsien, R. Y. Aptamers Switch on Fluorescence of
Triphenylmethane Dyes. J. Am. Chem. Soc. 125, 14716–14717 (2003).
26. Szent-Gyorgyi, C. et al. Fluorogen-activating single-chain antibodies for imaging cell surface
proteins. Nat. Biotechnol. 26, 235–240 (2008).
49
27. Özhalici-Ünal, H. et al. A Rainbow of Fluoromodules: A Promiscuous scFv Protein Binds to

and Activates a Diverse Set of Fluorogenic Cyanine Dyes. J. Am. Chem. Soc. 130, 12620–12621
(2008).
28. Yates, B. P., Peck, M. A. & Berget, P. B. Directed evolution of a fluorogen-activating single
chain antibody for function and enhanced brightness in the cytoplasm. Mol. Biotechnol. 54, 829–841
(2013).
29. Saurabh, S., Perez, A. M., Comerci, C. J., Shapiro, L. & Moerner, W. E. Super-resolution
Imaging of Live Bacteria Cells Using a Genetically Directed, Highly Photostable Fluoromodule. J. Am.
Chem. Soc. 138, 10398–10401 (2016).
30. Yan, Q. et al. Localization Microscopy using Noncovalent Fluorogen Activation by Genetically
Encoded Fluorogen-Activating Proteins. ChemPhysChem 15, 687–695 (2014).
10336 (2018).
35. Pimenta, F. M. et al. Chromophore Renewal and Fluorogen-Binding Tags: A Match Made to
Last. Sci. Rep. 7, 12316 (2017).
1120 (2019).
(2019).
38. Babendure, J. R., Adams, S. R. & Tsien, R. Y. Aptamers Switch on Fluorescence of
Triphenylmethane Dyes. J. Am. Chem. Soc. 125, 14716–14717 (2003).
39. Stojanovic, M. N. & Kolpashchikov, D. M. Modular Aptameric Sensors. J. Am. Chem. Soc.
126, 9266–9270 (2004).
40. Sparano, B. A. & Koide, K. A Strategy for the Development of Small-Molecule-Based Sensors
That Strongly Fluoresce When Bound to a Specific RNA. J. Am. Chem. Soc. 127, 14954–14955
(2005).
41. Paige, J. S., Wu, K. Y. & Jaffrey, S. R. RNA Mimics of Green Fluorescent Protein. Science
333, 642–646 (2011).
42. Huang, H. et al. A G-quadruplex–containing RNA activates fluorescence in a GFP-like
fluorophore. Nat. Chem. Biol. 10, 686–691 (2014).
43. Warner, K. D. et al. Structural basis for activity of highly efficient RNA mimics of green
fluorescent protein. Nat. Struct. Mol. Biol. 21, 658–663 (2014).
44. Truong, L. & Ferré-D’Amaré, A. R. From fluorescent proteins to fluorogenic RNAs: Tools for
imaging cellular macromolecules. Protein Sci. 28, 1374–1386 (2019).
45. Strack, R. L., Disney, M. D. & Jaffrey, S. R. A superfolding Spinach2 reveals the dynamic
nature of trinucleotide repeat–containing RNA. Nat. Methods 10, 1219–1224 (2013).
46. Autour, A., Westhof, E. & Ryckelynck, M. iSpinach: a fluorogenic RNA aptamer optimized for
in vitro applications. Nucleic Acids Res. 44, 2491–2500 (2016).
47. Filonov, G. S., Moon, J. D., Svensen, N. & Jaffrey, S. R. Broccoli: Rapid Selection of an RNA
Mimic of Green Fluorescent Protein by Fluorescence-Based Selection and Directed Evolution. J. Am.
Chem. Soc. 136, 16299–16308 (2014).
48. Filonov, G. S., Kam, C. W., Song, W. & Jaffrey, S. R. In-Gel Imaging of RNA Processing Using
Broccoli Reveals Optimal Aptamer Expression Strategies. Chem. Biol. 22, 649–660 (2015).
49. Song, W. et al. Imaging RNA polymerase III transcription using a photostable RNA–
fluorophore complex. Nat. Chem. Biol. 13, 1187–1194 (2017).
50
50. Dolgosheina, E. V. et al. RNA Mango Aptamer-Fluorophore: A Bright, High-Affinity Complex

for RNA Labeling and Tracking. ACS Chem. Biol. 9, 2412–2420 (2014).
51. Trachman, R. J. et al. Structure and functional reselection of the Mango-III fluorogenic RNA
aptamer. Nat. Chem. Biol. 15, 472–479 (2019).
52. Chen, X. et al. Visualizing RNA dynamics in live cells with bright and stable fluorescent RNAs.
Nat. Biotechnol. 1287–1293 (2019) doi:10.1038/s41587-019-0249-1.
53. Bozhanova, N. G. et al. Protein labeling for live cell fluorescence microscopy with a highly
photostable renewable signal. Chem. Sci. 8, 7138–7142 (2017).
54. Huang, P.-S., Boyken, S. E. & Baker, D. The coming of age of de novo protein design. Nature
537, 320–327 (2016).
55. Dou, J. et al. De novo design of a fluorescence-activating β-barrel. Nature 561, 485–491
(2018).
56. Grimm, J. B. et al. Bright photoactivatable fluorophores for single-molecule imaging. Nat.
Methods 13, 985–988 (2016).
57. Uno, S. et al. A spontaneously blinking fluorophore based on intramolecular spirocyclization
for live-cell super-resolution imaging. Nat. Chem. 6, 681–689 (2014).
58. Zheng, Q. et al. Rational Design of Fluorogenic and Spontaneously Blinking Labels for Super-
Resolution Imaging. ACS Cent. Sci. 5, 1602–1613 (2019).
59. Lukinavičius, G. et al. A near-infrared fluorophore for live-cell super-resolution microscopy of
cellular proteins. Nat. Chem. 5, 132–139 (2013).
60. Lukinavičius, G. et al. Fluorogenic probes for live-cell imaging of the cytoskeleton. Nat.
Methods 11, 731–733 (2014).
61. Lukinavičius, G. et al. Fluorogenic Probes for Multicolor Imaging in Living Cells. J. Am. Chem.
Soc. 138, 9365–9368 (2016).
62. Wirth, R., Gao, P., Nienhaus, G. U., Sunbul, M. & Jäschke, A. SiRA: A Silicon Rhodamine-
Binding Aptamer for Live-Cell Super-Resolution RNA Imaging. J. Am. Chem. Soc. 141, 7562–7571
(2019).
microscopy. Nat. Methods 12, 244–250 (2015).
Nat. Methods 14, 987–994 (2017).
65. Wang, L. et al. A general strategy to develop cell permeable and fluorogenic probes for
multicolour nanoscopy. Nat. Chem. 12, 165–172 (2020).
66. Hori, Y., Ueno, H., Mizukami, S. & Kikuchi, K. Photoactive yellow protein-based protein
labeling system with turn-on fluorescence intensity. J. Am. Chem. Soc. 131, 16610–16611 (2009).
67. Hori, Y., Hirayama, S. & Kikuchi, K. Development of cyanine probes with dinitrobenzene
quencher for rapid fluorogenic protein labelling. Phil Trans R Soc A 375, 20170018 (2017).
68. Hori, Y., Nakaki, K., Sato, M., Mizukami, S. & Kikuchi, K. Development of Protein-Labeling
Probes with a Redesigned Fluorogenic Switch Based on Intramolecular Association for No-Wash
Live-Cell Imaging. Angew. Chem. Int. Ed. 51, 5611–5614 (2012).
69. Hirayama, S., Hori, Y., Benedek, Z., Suzuki, T. & Kikuchi, K. Fluorogenic probes reveal a role
of GLUT4 N-glycosylation in intracellular trafficking. Nat. Chem. Biol. 12, 853–859 (2016).
70. Hori, Y. et al. Development of Fluorogenic Probes for Quick No-Wash Live-Cell Imaging of
Intracellular Proteins. J. Am. Chem. Soc. 135, 12360–12365 (2013).
71. Hori, Y., Hirayama, S., Sato, M. & Kikuchi, K. Redesign of a Fluorogenic Labeling System To
Improve Surface Charge, Brightness, and Binding Kinetics for Imaging the Functional Localization of
Bromodomains. Angew. Chem. Int. Ed. 54, 14368–14371 (2015).
72. Gao, J. et al. Development of Fluorogenic Probes for Rapid High-Contrast Imaging of
Transient Nuclear Localization of Sirtuin 3. ChemBioChem 21, 656–662 (2020).
73. Braselmann, E. et al. A multicolor riboswitch-based platform for imaging of RNA in live
mammalian cells. Nat. Chem. Biol. 14, 964–971 (2018).
74. Bouhedda, F. et al. A dimerization-based fluorogenic dye-aptamer module for RNA imaging in
live cells. Nat. Chem. Biol. 16, 69–76 (2020).
75. Holleran, J. et al. Fluorogen-activating proteins as biosensors of cell-surface proteins in living
cells. Cytometry A 77A, 776–782 (2010).
51
76. Yan, Q. et al. Near-instant surface-selective fluorogenic protein quantification using sulfonated
triarylmethane dyes and fluorogen activating proteins. Org. Biomol. Chem. 13, 2078–2086 (2015).
77. Wang, Y. et al. Affibody-targeted fluorogen activating protein for in vivo tumor imaging. Chem.
Commun. Camb. Engl. 53, 2001–2004 (2017).
1828 (2018).
79. Stensrud, K. F., Zanotti, K. J., Waggoner, A. S. & Armitage, B. A. Spectral Properties of
Fluorogenic Thiophene-derived Triarylmethane Dyes. Photochem. Photobiol. 95, 406–410 (2019).
81. Bozhanova, N. G., Baranov, M. S., Baleeva, N. S., Gavrikov, A. S. & Mishin, A. S. Red-Shifted
Aminated Derivatives of GFP Chromophore for Live-Cell Protein Labeling with Lipocalins. Int. J. Mol.
Sci. 19, 3778 (2018).
82. Song, W., Strack, R. L., Svensen, N. & Jaffrey, S. R. Plug-and-Play Fluorophores Extend the
Spectral Properties of Spinach. J. Am. Chem. Soc. 136, 1198–1201 (2014).
83. Meteleshko, Y. I., Nemukhin, A. V. & Khrenova, M. G. Novel flavin-based fluorescent proteins
with red-shifted emission bands: a computational study. Photochem. Photobiol. Sci. Off. J. Eur.
Photochem. Assoc. Eur. Soc. Photobiol. 18, 177–189 (2019).
84. Torra, J. et al. Tailing miniSOG: structural bases of the complex photophysics of a flavin-
binding singlet oxygen photosensitizing protein. Sci. Rep. 9, 2428 (2019).
85. Pimenta, F. M., Jensen, R. L., Breitenbach, T., Etzerodt, M. & Ogilby, P. R. Oxygen-Dependent
Photochemistry and Photophysics of “MiniSOG,” a Protein-Encased Flavin. Photochem. Photobiol.
89, 1116–1126 (2013).
86. Ruiz-González, R. et al. Singlet Oxygen Generation by the Genetically Encoded Tag miniSOG.
J. Am. Chem. Soc. 135, 9564–9567 (2013).
87. Shu, X. et al. A genetically encoded tag for correlated light and electron microscopy of intact
cells, tissues, and organisms. PLoS Biol. 9, e1001041 (2011).
88. Westberg, M., Holmegaard, L., Pimenta, F. M., Etzerodt, M. & Ogilby, P. R. Rational Design
of an Efficient, Genetically Encodable, Protein-Encased Singlet Oxygen Photosensitizer. J. Am.
Chem. Soc. 137, 1632–1642 (2015).
89. Westberg, M., Bregnhøj, M., Etzerodt, M. & Ogilby, P. R. No Photon Wasted: An Efficient and
Selective Singlet Oxygen Photosensitizing Protein. J. Phys. Chem. B 121, 9366–9371 (2017).
90. Makhijani, K. et al. Precision Optogenetic Tool for Selective Single- and Multiple-Cell Ablation
in a Live Animal Model System. Cell Chem. Biol. 24, 110–119 (2017).
91. He, J. et al. A near-infrared genetically targetable and activatable photosensitizer. Nat.
Methods 13, 263–268 (2016).
92. Potzkei, J. et al. Real-time determination of intracellular oxygen in bacteria using a genetically
encoded FRET-based biosensor. BMC Biol. 10, 28 (2012).
93. Erapaneedi, R., Belousov, V. V., Schäfers, M. & Kiefer, F. A novel family of fluorescent hypoxia
sensors reveal strong heterogeneity in tumor hypoxia at the cellular level. EMBO J. 35, 102–113
(2016).
94. Hu, H. et al. Monitoring cellular redox state under hypoxia using a fluorescent sensor based
on eel fluorescent protein. Free Radic. Biol. Med. 120, 255–265 (2018).
95. Grover, A. et al. A Genetically Encoded pH Sensor for Tracking Surface Proteins through
Endocytosis**. Angew. Chem. Int. Ed Engl. 51, 4838–4842 (2012).
96. Perkins, L. A. et al. Genetically Targeted Ratiometric and Activated pH Indicator Complexes
(TRApHIC) for Receptor Trafficking. Biochemistry 57, 861–871 (2018).
97. Berbasova, T. et al. A Genetically Encoded Ratiometric pH Probe: Wavelength Regulation-
Inspired Design of pH Indicators. ChemBioChem 19, 1288–1295 (2018).
98. Shitashima, Y., Shimozawa, T., Asahi, T. & Miyawaki, A. A dual-ligand-modulable fluorescent
protein based on UnaG and calmodulin. Biochem. Biophys. Res. Commun. 496, 872–879 (2018).
100. To, T.-L. et al. Rationally designed fluorogenic protease reporter visualizes spatiotemporal
dynamics of apoptosis in vivo. Proc. Natl. Acad. Sci. 112, 3338–3343 (2015).
52

102. To, T.-L., Zhang, Q. & Shu, X. Structure-guided design of a reversible fluorogenic reporter of
protein-protein interactions. Protein Sci. Publ. Protein Soc. 25, 748–753 (2016).
103. Filonov, G. S. & Verkhusha, V. V. A Near-Infrared BiFC Reporter for In Vivo Imaging of
Protein-Protein Interactions. Chem. Biol. 20, 1078–1086 (2013).
104. Shcherbakova, D. M. et al. Bright monomeric near-infrared fluorescent proteins as tags and
biosensors for multiscale imaging. Nat. Commun. 7, 12405 (2016).
105. Tchekanda, E., Sivanesan, D. & Michnick, S. W. An infrared reporter to detect spatiotemporal
dynamics of protein-protein interactions. Nat. Methods 11, 641–644 (2014).
106. Boassa, D. et al. Split-miniSOG for Spatially Detecting Intracellular Protein-Protein
Interactions by Correlated Light and Electron Microscopy. Cell Chem. Biol. 26, 1407-1416.e5 (2019).
53
54
Chapter III
Development of a multifunctional
chemogenetic reporter
This chapter is based on the following article:
Engineering of a fluorescent chemogenetic reporter with tunable color for advanced live-cell
imaging
bioRxiv DOI: https://doi.org/10.1101/2021.01.29.428635a
H. Benaissa, K. Ounoughi, I. Aujard, E. Fischer, R. Goïame, J. Nguyen, A. G. Tebo, C. Li, T. Le Saux,

L. Danglot, N. Pietrancosta, X. Morin, L. Jullien & A Gautier.
a
Link for the online article and movies
55
CHAPTER III: DEVELOPMENT OF A MULTIFUNCTIONAL CHEMOGENETIC REPORTER
III.1. Presentation of the article
Deciphering the complexity of living organisms and cells requires robust and multicolor probes
adaptable to a variety of experimental scenarios and allowing the visualization of various biomolecules
in action. Here, we present the design of a multifunctional protein able to recognize a palette of
fluorogenic chromophores that cover the entire visible spectrum. Benefiting from the power of
molecular engineering offered by modern organic chemistry and the robustness of directed evolution
strategy, we were able to engineer a promiscuous single protein tag suitable for a large variety of
imaging scenarios. Beyond allowing the genetic encoding of blue, cyan, green, yellow, orange and
red fluorescence, it allowed the selective labeling and imaging of proteins in live cells, neurons, and
multicellular organisms, and is well adapted to the most advanced imaging techniques, including
super-resolution microscopy for observation at the nanoscale.
The molecules described in this chapter were synthetized in collaboration with the group of Prof.
Ludovic Jullien (Karim Ounoughi and Dr. Isabelle Aujard with the help of Dr. Thomas Le Saux) of the
PASTEUR laboratory (Chemistry department, École Normale Supérieure) and by Dr. Chenge Li, a
former PhD candidate of our group. The generation of the initial library of FAST mutants described in
this chapter was designed by Dr. Alison G. Tebo, a former postdoctoral member of our group. The
structure models described in this chapter were generated by homology modeling in collaboration
with Dr. Nicolas Pietrancosta of the Laboratoire des BioMolécules / Neurosciences Paris Seine
(Sorbonne University). The experiments in chicken embryos were conducted by the group of Dr.
Xavier Morin (Evelyne Fisher, Rosette Goïame and Xavier Morin) in the Biology department of the
École Normale Supérieure (IBENS). The experiments in primary hippocampal neuronal cells were
conducted in collaboration with the group of Dr. Lydia Danglot (Julie Nguyen and Lydia Danglot) at
the Institut de Psychiatrie et Neurosciences de Paris. Finally, airyscan and STED experiments were
carried out at NeurImag Imaging core facility with the help of Lydia Danglot.
56
III.2. Article: Engineering of a fluorescent chemogenetic reporter with tunable color for
advanced live-cell imaging
III.2.1. Abstract
Biocompatible fluorescent reporters with spectral properties spanning the entire visible spectrum are
indispensable tools for imaging the biochemistry of living cells and organisms in real time. Here, we
present the engineering of a fluorescent chemogenetic reporter with tunable optical and spectral
properties. A collection of live-cell compatible fluorogenic chromophores with various electronic
properties enables to generate bimolecular fluorescent assemblies that cover the visible spectrum
from blue to red using a single protein tag engineered and optimized by directed evolution and rational
design. We showed that the ability to tune the fluorescence color and properties through simple
molecular modulation provides an unprecedent experimental versatility for imaging proteins in live
cells, including delicate cultured hippocampal neurons, and in multicellular organisms. The ability to
tune the spectral properties and fluorescence performance enables to match the spectral
specifications and requirements of the most advanced imaging techniques, and allowed us to achieve
efficient stimulated emission depletion (STED) nanoscopy of fusion proteins in live cells and live
primary cultured neurons.
III.2.2. Introduction
Fluorescent labels and biosensors have revolutionized the way we study living systems. Targeted to
specific biomolecules or cells, they allow non-invasive imaging of the machinery that govern cells and
organisms in real time. The discovery and development of fluorescent proteins (FPs) have been a
transformative milestone in molecular and cell imaging1. Genetic fusions to FPs allowed cell biologists
to visualize the dynamic of proteins in living cells and organisms with unprecedented spatial and
temporal resolution. The multiplicity of optical properties of FPs allowed multicolor imaging and the
development of various biosensors. Recently, fluorescent chemogenetic reporters consisting in
synthetic organic dyes anchored covalently or non-covalently to genetically encoded protein tags2–6
have challenged the conventional paradigm of fluorescent proteins and opened new prospects for
innovative on-demand bioimaging and biosensing7–9. These hybrid systems combine the targeting
selectivity of genetic tags with the advantages of synthetic organic fluorophores, which can be very
bright and photostable, and provide spectral properties covering most of the visible spectrum10–12.
Here, we present the engineering of a fluorescent chemogenetic reporter with highly tunable spectral
and optical properties. Spectral tuning was achieved by extensive modulation of the electronic
properties of the embedded chromophore in the Fluorescence-Activating and absorption-Shifting Tag
(FAST). Prototypical FAST is a small protein tag (14 kDa) with reduced genetic footprint, which was
57
originally designed to fluoresce instantaneously upon non-covalent binding of fluorogenic 4-

hydroxybenzylidene rhodanine (HBR) derivatives that are inherently non-fluorescent unless bound to
FAST13. Recent studies showed that FAST enabled (i) to circumvent some limitations of conventional
fluorescent proteins — allowing for instance the imaging of proteins in strict anaerobic organisms14,15
or the visualization of protein secretion by bacteria in real time16,17 — and (ii) to bioengineer functional
biosensors for monitoring intracellular analytes18, and observing protein-protein interactions with
unprecedent spatial and temporal resolution19. Capable of binding various permeant and non-
permeant HBR derivatives13,20,21, prototypical FAST exhibited great potential for engineering a
multifunctional fluorescent chemogenetic reporter with tunable color through chromophore
modulation. The challenge lied in obtaining a rainbow of color from a single protein tag, as spectral
tuning necessitated modulation of the electronic properties of the embedded chromophore, and thus
structural change. We achieved this engineering feat by replacing the rhodanine found in HBR
derivatives by isosteric heterocycles exhibiting various electron-withdrawing abilities in order to obtain
absorption and emission wavelengths that cover the entire visible spectrum (Figure III.1). Cycles of
directed protein evolution allowed the generation of a promiscuous FAST variant — named pFAST
— with optimized chromophore binding affinity and fluorescence brightness. The ability to tune the
fluorescence color from blue to red by choosing a different live-cell compatible fluorogenic
chromophore provides an unprecedented experimental versatility for multicolor live-cell imaging. We
demonstrate that this tunable fluorescent chemogenetic reporter has spectroscopic attributes suitable
for the most advanced imaging techniques, including STED super-resolution microscopy, and allows
investigators to face a large variety of experimental scenarios.
58
Figure III.1. The engineered protein tag pFAST is a fluorescent chemogenetic reporter with tunable
color. (a) Structures of the chromophores of the HBO, HBT, HBP, HBR, HBIR and HBRAA series positioned
according to the wavelengths of maximal emission and maximal absorption of their assembly with pFAST. The
fluorescence quantum yield (f), molar absorptivity at maximal absorbance (e) and thermodynamic dissociation
constant (KD) values of their assembly with pFAST are indicated into brackets below the molecules. (b)
Normalized absorption (left) and fluorescence (right) spectra of the different fluorogenic chromophores bound
to pFAST. Spectra were recorded in pH 7.4 PBS at 25°C. Fluorescence spectra were measured by exciting at
the maximal absorption wavelength. (c) Solutions containing the different pFAST:chromophore assemblies
illuminated with 365 nm light.
59
III.2.3. Results
III.2.3.1. Molecular spectral tuning
Extensive spectral tuning necessitated to find a range of fluorogenic chromophores that displayed a
wide variety of electronic properties and could still be recognized by a single binding pocket. To keep
structural changes as minimal as possible, we replaced the electron-withdrawing rhodanine (R)
heterocycle found in the push-pull fluorogenic HBR derivatives with isosteric pseudothiohydantoin (P),
2,4-thiazolidinedione (T) and 2,4-oxazolidinedione (O). Finer tuning was achieved by secondly
varying the substituents on the electron-donating phenol, as previously achieved within the HBR
series21 (See Figure III.1a for structures). Just like HBR analogs13, HBX (X = P, T or O) compounds
are mainly protonated at physiological pH (their pKAs range from 8.15 to 9.05) (Figure III.S1), they
undergo a significant absorption red-shift upon deprotonation (Figure III.S1), and they are inherently
non-fluorescent in solution. Preliminary screening showed that FAST was able to bind all
chromophores, albeit with low to modest affinity (Table III.S1). Like HBR analogs, the HBX (X = P, T
or O) analogs displayed red-shifted absorption when bound to FAST at physiological pH (Table
III.S1), suggesting that the binding pocket of FAST stabilizes their deprotonated phenolate state in a
similar fashion13. The corresponding bimolecular assemblies were all fluorescent, although in average
dimmer than those formed with the HBR analogs, with absorption/emission peaks blue-shifted by
~35/40 nm (for X = P), ~65/65 nm (for X = T) and ~85/75 nm (for X = O) relative to their HBR-based
counterpart. In a given series, changing the number and/or nature of the substituents present on the
phenol moiety allowed us furthermore to finely tune the absorption/emission peaks of the FAST
assemblies (Figure III.1, Table III.S1).
III.2.3.2. Optimization by directed protein evolution
To obtain an optimized fluorescent chemogenetic reporter with tunable color, we engineered a FAST
variant with extended chromophore binding promiscuity. Because the chromophores of the HBR,
HBP, HBT and HBO series were isosteric, we hypothesized that mutations able to increase the
binding affinity and brightness for one chromophore could be sufficient to generate a promiscuous
variant displaying improved properties with all chromophores. Directed protein evolution coupled with
rational design have previously been an efficient method to generate FAST variants forming bright
assembly with new chromophore22 or having orthogonal chromophore specificity23. We thus used a
combinatorial library of variants generated by random mutagenesis and displayed on yeast cells.
Iterative rounds of fluorescence activated cell sorting (FACS) decreasing progressively chromophore
concentrations through rounds allowed us to identify variants forming tighter and brighter assemblies
with chromophores of the HBX series (Figure III.2a-d, Text III.S1, Figure III.S2-S4, Tables III.S2-
S9). Beneficial mutations were combined to generate oFAST, a variant with the mutations Q41L,
60
D71N, V83I, M109L and S117R, which forms a blue fluorescent assembly with HBO-3,5DM 11-fold
tighter and 3-fold brighter than FAST:HBO-3,5DM (Figure III.2b, Table III.S2). In parallel, the genes
of five variants displaying improved properties with HBP-3,5DM were recombined by DNA shuffling
to generate a new library (library B, Figure III.2a). FACS screening allowed us to identify (i) tFAST, a
variant with the mutations G25R, Q41K, S72T, A84S, M95A, M109L and S117R relative to FAST,
which forms a cyan fluorescent assembly with HBT-3,5DM 5-fold tighter and 1.5-fold brighter than
FAST:HBT-3,5DM (Figure III.2c, Table III.S3), and (ii) an improved variant for HBP-3,5DM bearing
the mutations K17N, G21E, G25R, A30V, Q41L, S72T, V83A, M95T, S117R, which gave, after
addition of the mutation M109L found in several other selected variants, the variant pFAST able to
form a green fluorescent assembly with HBP-3,5DM 25-fold tighter and 2.5-fold brighter than
FAST:HBP-3,5DM (Figure III.2d, Table III.S4).
The variants oFAST, tFAST and pFAST showed superior properties not only with the chromophore
used for their selection, but with all chromophores of the HBO, HBT and HBP series, demonstrating
that our strategy successfully generated variants with increased chromophore-binding promiscuity.
Unlike FAST, the three selected variants bind most of the chromophores of the HBO, HBT and HBP
series with submicromolar affinity (Figure III.2h, Tables III.S2-S4). This gain in affinity was
accompanied with an increase of the fluorescence quantum yield of almost all fluorescent assemblies
(Figure III.2h, Tables III.S2-S4). The three variants displayed also an improved affinity (up to one
order of magnitude) for the chromophores of the HBR series, in agreement with an increase of the
chromophore binding promiscuity (Figure III.2h, Tables III.S2-S4).
61
Figure III.2. Engineering of promiscuous FAST variants. (a) Directed evolution allowed the generation of
oFAST, tFAST and pFAST. Yeast-displayed libraries of variants of FAST were screened in presence of the
indicated fluorogenic chromophore by fluorescence-activating cell sorting (solid arrows). Additional mutations
were introduced by rational design (dotted arrows). (b-d) KDs (thermodynamic dissociation constants) and
fluorescent quantum yields (f) of the clones isolated from the selections with (b) HBO-3,5DM, (c) HBT-3,5DM
and (d) HBP-3,5DM. Filled triangles are variants from library A ; black squares are variants from library B ; grey
triangles are the variants used for the generation of the DNA shuffling library B and unfulfilled triangles and
squares are variants generated by rational design. Values are also given for FAST for comparison. (e-g)
Confocal micrographs of HeLa cells expressing cytoplasmic (e) oFAST labeled with 10 µM HBO-3,5DM, (f)
tFAST labeled with 5 µM HBT-3,5DM and (g) pFAST labeled with 5 µM HBP-3,5DM. Scale bars 10 µm. See
Table III.S12 for imaging settings. (h) KDs (thermodynamic dissociation constants) and emission wavelength of
chromophores of the HBX ( X = R, P, T and O) series of fluorogens in presence of FAST variants. Fluorescent
quantum yields are given in %.
62
Of the three selected variants showing superior performance, pFAST emerged as the best, and was
thus tested with other chromophores. pFAST was shown to form tighter assemblies with HBR analogs
bearing an additional carboxymethyl group on the rhodanine head (HBRAA series, Figure III.1 and
Table III.S4). These chromophores were previously shown to have a reduced cellular uptake because
of their negatively charged carboxylate at physiological pH, allowing the specific labeling of FAST-
tagged proteins located at the cell surface20. In particular, we discovered that pFAST could efficiently
interact with the unreported HBRAA-3,5DM, forming a bright orange fluorescent assembly that was
2.5-fold brighter than FAST:HBRAA-3E, the most efficient combination for selective labeling of cell
surface proteins reported so far20. Finally, we discovered that pFAST could also bind substituted 4-
hydroxybenzyldidene isorhodanine derivatives (HBIR series, Figure III.1) with single or double digit
nanomolar affinities (Table III.S4). The resulting assemblies exhibited red-shifted absorption and
emission compared to their HBR counterparts. However, only the red fluorescent pFAST:HBIR-
3,5DOM assembly exhibited a decent fluorescent quantum yield: the assemblies with HBIR-3M and
HBIR-3,5DM, although among the tightest observed in this study, almost did not fluoresce. Although
these properties would disqualify HBIR-3M and HBIR-3,5DM for fluorescence imaging applications,
they allowed the efficient generation of absorbing-only dark assemblies (vide infra).
Overall, this set of experiments demonstrated that the selected pFAST was a highly promiscuous
protein tag able to bind with high affinity a large variety of structurally related chromophores with
diverse electronic properties. Our expanded collection of chromophores allowed the efficient
generation of a palette of non-fluorescent and fluorescent assemblies emitting blue, cyan, green,
yellow, orange, and red light using a single protein tag (Figure III.1, Figure III.S5).
III.2.3.3. Structural model
Primary sequence alignments showed that the mutations K17N, G21E, G25R, A30V, Q41L, S72T,
V83A, M95T, M109L, S117R found in pFAST were mutation hotspots found in other variants (Figure
III.S3). To get insights into the functional role of these mutations, atomic-resolution models of FAST
and pFAST were generated by homology modeling using the tridimensional crystal structure of the
evolutionary related photoactive yellow protein (PYP) from Halorhodospira halophila as template
(Figure III.3a and Figure III.S6a). The analysis of the root mean square fluctuation (RMSF) of each
residue during molecular dynamic simulations showed an overall rigidification of the structure of
pFAST compared to FAST (Figure III.3a,b). In particular, the loop 94-101 originally mutated to
generate FAST from PYP, shows very little movement within pFAST compared to FAST. This overall
rigidification was accompanied by a significant reduction of the accessible binding site volume
compared to FAST and PYP (Figure III.S6a,b). All mutations found in pFAST participate to the
rigidification of the overall structure and/or in the stabilization of the protein-chromophore assembly.
63
The mutations G21E, G25R introduce two polar residues at the protein surface, which improves the
solvation of the protein and thus the compacity of the structure (Figure III.S6e,f). The mutations A30V
and Q41L enable strong hydrophobic interactions with residues in the N-terminal domain, known to
be flexible in PYP, and thus play a major role in compacting and rigidifying the entire three-
dimensional structure (Figure III.S6g,h). The mutation S72T creates interactions with the residues
Glu 74, Phe75, Tyr76, which have a direct impact on the binding site structure (Figure III.S6i). The
mutation V83A enables closer contacts with the residues Phe79, Ser85, Gly86, Arg87 and Tyr 118,
strengthening core interactions (Figure III.S6j). The mutation M95T strengthens the C-terminal b-
sheet through additional interactions with Lys104 and Pro102 (Figure III.S6k). The mutation M109L
directly impacts the size of the binding site by creating hydrophobic interactions with Phe75 and
Val120 (Figure III.S6l). Finally, the mutation S117R participates to the overall rigidification of the
structure by creating polar interactions with Asp34 in the N-terminal domain and apolar interactions
with Ala112, strengthening the b-sheet structure (Figure III.S6m). The overall rigidification of pFAST
structure creates a chromophore binding site that is smaller, but more open and better defined (Figure
III.S6). Chromophore docking showed that pFAST was able to bind all chromophores of the HBO,
HBT, HBP and HBR series with higher affinities than FAST does, as observed experimentally (Figure
III.S6c). Our model confirmed that chromophores adopt (i) a quasi-planar conformation and (ii) an
orientation of the phenolic ring enabling the hydrogen-bond network formed by the conserved
residues Tyr42, Glu46, Thr50 and Arg52 to stabilize the phenolate anion in the same way that PYP
does with its chromophore24–27, as originally proposed13 (Figure III.3c, Figure III.S6n). This binding
mode explain why chromophore binding induces both an increase of the fluorescence quantum yield
(as the planar conformation favors radiative relaxation), and a strong red shift in absorption (as the
protonation/deprotonation equilibrium is shifted towards the deprotonated state). Our analysis showed
that Glu46 forms a strong hydrogen bond with the phenolate, while Arg52 interacts with the electron-
withdrawing heterocycle through hydrogen bonding with the endocyclic heteroatom (Figure III.3c and
Figure III.S6n). In addition, Trp94 and Pro97, known to be essential for chromophore recognition and
activation13, form hydrogen bound and apolar interactions with the chromophore. Fine analysis with
HBP-3,5DM allowed us to map the entire network of polar and apolar interactions involved in
chromophore recognition and activation (Figure III.3c).
64
Figure III.3. Structural model of pFAST. (a) Structural model of pFAST and FAST generated by homology
modeling using the tridimensional crystal structure of the Halorhodospira halophila Photoactive Yellow Protein
PYP (PDB: 6P4I) (See also Figure III.S6a). Mutations of pFAST relative to FAST are indicated. (b) Root Mean
Square Fluctuation (RMSF) of the residues within pFAST and FAST during molecular dynamic simulations. (c)
Polar and apolar interactions network involved in HBP-3,5DM binding and recognition within pFAST.
III.2.3.4. Multicolor fluorescent labeling in mammalian cells
Preliminary experiments showed that incubation with the new fluorogenic chromophores for 24 hours
had no deleterious effects on mammalian cells at the concentrations used for labeling (Figure III.S7)
and that the expression of the selected variants in mammalian cells was homogenous, in agreement
with high intracellular stability (Figure III.2e-g). Imaging of live HeLa cells expressing pFAST fused to
the histone H2B and treated for few tens of seconds with the different membrane permeant
fluorogenic chromophores demonstrated that pFAST could be used to efficiently encode blue, cyan,
green, yellow, orange and red fluorescence in mammalian cells (Figure III.4a). All chromophores
showed good to excellent performances for high-contrast imaging, except HBO-3,5DM and HBO-3M
that formed dimmed blue fluorescent assemblies in agreement with their in vitro properties.
Fluorescence quantification after labeling in both live and fixed cells enabled to show that fixation with
paraformaldehyde did not alter significantly fluorescence, suggesting that pFAST remained functional
after fixation and could still be labeled (Figure III.S8). We observed in particular that signal to
background ratio in live and fixed cells did not significantly change no matter the chromophore used
(Figure III.S8).
We next showed that selective labeling of cell-surface pFAST-tagged proteins could be efficiently
achieved with the membrane impermeant HBRAA-3M, HBRAA-3E and HBRAA-3,5DM. In agreement
with the low cell uptake of these three chromophores, labeling of a secreted transmembrane fusion
65
with a pFAST domain facing the extracellular space led to a fluorescence signal restricted at the
plasma membrane (Figure III.4a, Figure III.S9). We observed that pFAST outperformed FAST in
terms of effective brightness in cells, and that HBRAA-3,5DM was the best candidate for the efficient
labeling of pFAST-tagged cell-surface proteins, in agreement with its superior in vitro brightness.
III.2.3.5. Photostability
Continuous chromophore exchange can reduce the apparent rate of photobleaching through renewal
of photodamaged chromophores. Photooxidation of specific methionine and tryptophan residues was
however previously shown to cause FAST photodamage28. Long-term irradiation of live HeLa cells
expressing pFAST showed that pFAST displayed enhanced photostability with HMBR, HBR-3,5DM
and HBR-3,5DOM, reaching photostability has high as the highly photostable enhanced GFP (EGFP)
(Figure III.S10a). The mutations M95T and M109L may explain the higher photoresistance of pFAST
compared to FAST. Prolonged imaging in presence of HBP-3,5DM, HBP-3,5DOM, HBP-3M and HBT-
3,5DOM showed that pFAST exhibited also good photostability with these new fluorogenic
chromophores, although a rapid transient fluorescence decay was observed before reaching steady
state with the last three (Figure III.S10). Initial fluorescence could be fully recovered in the dark,
demonstrating that this initial photobleaching was fully reversible. Decreasing light intensities or
reducing imaging frequency enabled furthermore to reduce the amplitude of the initial fluorescence
drop. This set of experiments suggested that the observed reversible photobleaching was likely due
to photoisomerization and/or photoejection of the chromophore, as previously proposed to explain the
reversible photobleaching observed when labeling FAST and its variants with low concentrations of
HBR derivatives28,23. This singular photochemical signature might be used advantageously for
selective imaging techniques based on kinetic discrimination29,30 or for single-molecule localization
microscopies31,32.
66
a HBO-3M HBT-3,5DM HBT-3,5DOM HBP-3,5DOM HBRAA-3E HBRAA-3,5DM HBR-3,5DOM
473 nm 499 nm 532 nm 554 nm 558 nm 578 nm 600 nm
λem
450
600
500
550
(nm)
HBO-3,5DM HBP-3M HBP-3,5DM HMBR HBRAA-3M HBR-3,5DM HBIR-3,5DOM
483 nm 503 nm 520 nm 542 nm 554 nm 561 nm 616 nm
b pFAST H2B-pFAST Lyn11-pFAST LifeAct-pFAST Mito-pFAST MAP4-pFAST

c
*
live
pFAST H2B-pFAST Lyn11-pFAST LifeAct-pFAST Mito-pFAST MAP4-pFAST

fixed
d
Lyn11-pFAST LifeAct-pFAST MAP4-pFAST
Figure III.4. Selective imaging of pFAST fusions in various cellular localizations in mammalian cells and
cultured neurons. (a) Confocal micrographs of HeLa cells expressing pFAST fusions (either with H2B or to a
transmembrane domain) in presence of the entire set of fluorogenic chromophores. Scale bars 10 µm. (b)
Confocal micrographs of live and fixed HeLa cells expressing pFAST fused to: histone H2B, lyn11 (inner
membrane-targeting motif), LifeAct (actin binding peptide domain), mito (mitochondrial targeting motif) and to
microtubule-associated protein (MAP) 4 and labeled with 5 µM HBP-3,5DM. Note that control experiments
showed that the mislocalization of MAP4-pFAST observed in fixed cells (*) was due to MAP4 fixation rather
than pFAST (see also Figure III.S12). Scale bars, 10 µm. (c) Depth color coded three-dimensional
reconstruction (from 81 optical sections) of live HeLa cells expressing lyn11-pFAST fusion and labeled with 5
µM HBR-3,5DOM. (d) Confocal micrographs of live dissociated hippocampal neurons expressing pFAST fused
to lyn11, LifeAct and MAP4 and labeled with 10 µM HBR-3,5DOM. Scale bars, 50 µm. The Lyn11-pFAST image
results from the maximum intensity projection of five optical sections. (a-d) See Table III.S13 for imaging
settings.
67
III.2.3.6. Imaging fusion proteins in mammalian cells and cultured neurons
To verify that pFAST was well suited for the selective imaging of proteins in various cellular
localizations, HeLa and HEK293T cells expressing pFAST fused to the histone H2B (H2B-pFAST),
the mitochondrial targeting sequence from the subunit VIII of human cytochrome C oxidase (Mito-
pFAST), the inner plasma membrane targeting sequence lyn11 (Lyn11-pFAST), the actin binding
peptide LifeAct (LifeAct-pFAST) and the microtubule-associated protein (MAP) 4 (MAP4-pFAST)
were imaged by confocal microscopy in the presence of HBP-3,5DM. pFAST fusions showed proper
cellular localization both in live and fixed cells, demonstrating that pFAST did not perturb protein
functions (Figure III.4b, Figure III.S11 and Figure III.S12). The high performance of pFAST further
allowed us to generate high resolution three-dimensional reconstructions of live HeLa cells expressing
lyn11-pFAST from 81 optical sections (Figure III.4c) showing the filopodia decorating plasma
membrane contours.
To verify the suitability of pFAST for imaging proteins in delicate and sensitive cells, we next
expressed FAST fusions in dissociated hippocampal neurons. Dissociated hippocampal neurons
were successfully transfected with plasmids encoding Lyn11-pFAST, LifeAct-pFAST or MAP4-
pFAST. Live-cell confocal imaging showed (i) that labeled neurons remained fully viable with an
impressive transfection rate (Figure III.S13), in agreement with neither pFAST fusions nor the
fluorogenic chromophore being toxic, and (ii) that fusion proteins were properly localized (Figure
III.4d). As observed for mammalian cells, pFAST allowed us to generate high-resolution confocal
micrographs of live neurons (Figure III.4d) showing tiny but resolved protrusions, further
demonstrating the high brightness and performance of pFAST for live imaging.
III.2.3.7. Enhanced fluorescent labeling in multicellular organisms
Mechanistic studies of most biological processes require imaging in multicellular organisms. Imaging
of semisynthetic chemogenetic reporters in multicellular organisms is however often hampered by the
low cell permeability of synthetic probes. HBR derivatives have been previously shown to allow the
labeling of FAST and its derivatives in zebrafish embryos thanks to their very high cell uptake and
superior ability to diffuse across multicellular systems13,22,23. Preliminary experiments in which we
quantified the labeling efficiency in live HeLa cells by flow cytometry showed that full labeling of
pFAST was achieved at lower chromophore concentrations than FAST, in agreement with its superior
binding affinity (Figure III.S14), thus facilitating intracellular labeling and minimizing the amount of
chromophore required for efficient detection. To verify that pFAST outperformed FAST in a similar
fashion for imaging proteins in multicellular organisms, we compared the relative labeling efficiency
of pFAST and FAST with various fluorogenic chromophores in chicken embryo. Comparison was
68
performed in the same embryo by expressing each protein in a different side of the neural tube. Time-
lapse imaging allowed us to monitor simultaneously the labeling of pFAST and FAST upon successive
addition of chromophore solutions with gradually increased concentrations (Figure III.S15, Movie
III.S1). With all tested chromophores, pFAST reached full labeling at lower concentrations than FAST.
Fluorescence saturation was furthermore reached within few tens of minutes, in agreement with the
high cell uptake of the tested chromophores. The effective brightness of pFAST was in general
comparable or higher than that of FAST, further demonstrating the superiority of pFAST over FAST
for labeling proteins into tissue (Figure III.S15).
Encouraged by its superior performance in multicellular organisms, we next compared pFAST with
the cyan fluorescent protein mCerulean, the green fluorescent protein EGFP, the yellow fluorescent
protein EYFP and the orange fluorescent protein mKO. Direct comparisons were performed within
the same embryo by expressing selectively each reporter in a different side of the neural tube. Apart
from pFAST:HBP-3M that displayed lower effective brightness than mCerulean, as expected from
their relative molecular brightness, all the other pFAST:chromophore combinations showed
performances comparable with those of their fluorescent protein spectral equivalent (Figure III.5a-g,
Figure III.S16).
Next, we demonstrated that pFAST was perfectly well suited for imaging proteins involved in dynamic
cellular processes within the neuroepithelium of proliferating chicken neural tube by en-face time-
lapse multi-color imaging. Tagging mitochondria with Mito-pFAST (labeled with HBR-3,5DOM) and
membranes with iRFP670 (memb-iRFP670) allowed us to successfully visualize symmetrical
mitochondrial segregation during cell division (Figure III.5h, Movie III.S2). Expression of (i) H2B-
pFAST (labeled with HBP-3,5DOM), and the PACT domain of pericentrin fused to mKO (pact-mKO)
and (ii) H2B-pFAST (labeled with HMBR), pact-mKO, and memb-iRFP670 enabled us on the other
hand to visualize chromosome segregation and centromere dynamics during cell division (Figure
III.5i,j, Movies III.S3-III.S4).
Beyond demonstrating that pFAST was well suited to image dynamic processes in living organisms,
this set of experiments also illustrates how the spectral properties of pFAST can be fine-tuned at will
by using the most appropriate chromophore for a given experiment. Overall, these experiments
demonstrated that pFAST allowed mechanistic studies of biological processes in multicellular
organisms.
69
a E3
bilateral electroporation open-book culture and
right side left side left side right side en-face live imaging
E2
2h 24 h pFAST
– + + –
reporter pFAST
reporter
+ control + control
neural tube pFAST reporter
b HBR-3,5DOM c HBR-3,5DM d HBP-3,5DOM e HMBR f HBP-3,5DM g HBP-3M

- + - + - + - + - + - +
mb-mCherry H2B-mCerulean H2B-pFAST

H2B-pFAST
H2B-pFAST
H2B-pFAST
H2B-pFAST
H2B-pFAST
H2B-EGFP
H2B-EGFP
H2B-EYFP
H2B-EYFP
H2B-mKO
mb-mCherry
mb-mCherry
mb-mCherry
mb-mCherry
mb-EGFP
mb-mCherry
mb-mCherry
mb-mCherry
mb-mCherry
mb-mCherry
mb-EGFP
h mito-pFAST / mb-iRFP670
0 min 9 min 15 min 18 min 21 min 24 min 27 min 42 min
HBR-3,5DOM
i H2B-pFA
F ST / pact-mKO
0 min 4 min 8 min 10 min 12 min 18 min 38 min
HBP-3,5DOM
j H2B-pFAST / pact-mKO / mb-iRFP670

6 min 24 min 32 min 34 min 38 min 54 min
HMBR
Figure III.5. Selective imaging of pFAST fusions in chicken embryo. (a-g) Plasmids encoding H2B-pFAST
and (b) mKO, (c-d) EYFP, (e-f) EGFP and (g) mCerulean fused to H2B were electroporated in each side of the
neural tube in ovo at embryonic day 2 (E2, HH stage 13-14). mCherry or EGFP reporters were co-injected with
each construct to monitor electroporation efficiency. 24 h later, embryos with homogeneous bilateral reporter
expression in the neural tube were dissected and imaged by (b) spinning-disk confocal microscopy or (c-g)
widefield fluorescence microscopy before and after addition of the indicated fluorogenic chromophore. (h-j)
Plasmids encoding (h) mito-pFAST (mitochondrial-targeting motif) and memb-iRFP670 (membrane-targeting
motif); (i) H2B-pFAST and pact-mKO (PACT domain of pericentrin) and (j) H2B-pFAST, pact-mKO and mb-
iRFP670 were electroporated in the neural tube in ovo, at embryonic day 2 (E2, HH stage 13-14). 24 h later,
embryos were dissected, and imaged in presence of (h) 1 µM HBR-3,5DOM , (i) 5 µM HBP-3,5DOM and (j) 1
µM HMBR using a spinning disk confocal microscope. Time-lapse showing cell division presented correct
localization of the proteins (see also Movies III.S2-S4). Scale bars, 10 µm. (a-j) See Table III.S13 for imaging
settings.
70
III.2.3.8. Reversible labeling
Some experiments require one to rapidly switch off fluorescence of a given fluorescent reporter. As
FAST labeling is non-covalent, it has been previously shown that washing away the fluorogenic
chromophore by perfusing chromophore-free medium can reverse the labeling of FAST-tagged
proteins in cells and switch off fluorescence with an unprecedented simplicity13. As HBIR-3M forms a
non-fluorescent assembly with pFAST, we wondered if it could be used as a dark competitor capable
of turning off pFAST fluorescence through simple addition to the cell culture medium. Addition of an
excess of HBIR-3M proved to efficiently reverse the labeling of pFAST in mammalian cells within just
few seconds without the need to wash away the fluorogenic chromophore beforehand (Figure III.S17,
Movie III.S5). Because of its single digit nanomolar binding affinity, HBIR-3M was able to efficiently
replace fluorogenic chromophores binding pFAST with low, intermediate and even high affinities.
Encouraged by these results, we next tested if HBIR-3M could allow one to switch off pFAST
fluorescence in multicellular systems, in which unlabeling by washing is more challenging. Chicken
neural tubes expressing H2B-pFAST were dissected and labeled with fluorogenic chromophores.
After removal of the excess of chromophore, and washing with phosphate buffer, samples were
treated with HBIR-3M and simply washed a second time. While simple washing was unsuccessful to
switch off fluorescence, addition of an excess of HBIR-3M efficiently switched off pFAST within few
tens of minutes (Figure III.S18, Movie III.S6). Compared to protocols based on sequential washing
steps, unlabeling through addition of a dark competitor can be a very simple and effective method to
switch off fluorescence with unprecedented time resolution, in particular when slow passive diffusion
prevents efficient washing in cells and multicellular organisms.
III.2.3.9. Imaging fusion proteins below the diffraction limit
Encouraged by the superior performance of pFAST, we further tested pFAST together with advanced
confocal microscopy techniques. Airyscan confocal imaging is a powerful tool allowing the imaging of
biological structures in live cells with enhanced signal-to-noise ratio and increased spatial resolution
(up to 140 nm). Airyscan confocal imaging of live HeLa cells expressing Lyn11-pFAST or MAP4-
pFAST labeled with various HBR or HBP derivatives allowed us to visualize details unresolved by
traditional confocal microscopy (Figure III.S19). Airyscan imaging in live cells allowed us to image
the rapid dynamics of microtubules and membranes with high temporal and spatial resolution (Movie
III.S7).
We extended our study to Stimulated Emission Depletion (STED) nanoscopy. STED microscopy is a
powerful technique allowing the imaging of biological structures in fixed and live cells with nanoscale
71
resolution. As many super-resolution microscopy techniques, the impact of STED nanoscopy is

however limited by the number of compatible fluorophores. In STED, a red-shifted doughnut-shaped
STED pulse featuring zero intensity at the very center is applied immediately after the excitation pulse,
resulting in depletion of the excitation state of the fluorophores within the doughnut area by stimulated
emission. Maximal resolution is obtained with bright and photostable fluorophores that are highly
sensitive to the STED laser for efficient depletion, and that are not re-excited by it33. Among our
extended palette of chromophores, HBR-3,5DOM appeared well suited for STED as it allowed the
generation of a bright and photostable red fluorescent reporter with a 80 nm Stokes shift (Excitation /
Emission peaks at 520 nm / 600nm) that allows effective depletion using the common 775 nm pulsed
depletion laser while minimizing reexcitation. Labeling with HBR-3,5DOM allowed the imaging of
Lyn11-pFAST and MAP4-pFAST in live HeLa cells with nanoscale resolution (Figure III.6a-e).
pFAST:HBR3,5DOM was bright and photostable enough to resist multiple average acquisition (16
average line). Figure III.6a-c show STED image (b) and deconvoluted STED image (c) compared to
confocal (a) and illustrate how STED microscopy allowed us to resolve different filopodia structures
(see zoom in) on live samples with the use of bright and photostable pFAST:HBR-3,5DOM. Having
at hand a chromophore with spectroscopic properties compatible for STED nanoscopy allowed us to
successfully image pFAST fusions below the diffraction barrier in live dissociated hippocampal
neurons. Labeling actin with LifeAct-pFAST enabled us to resolve details of axons and dendrites, as
well as visualize neurite growth cones with super-resolution (Figure III.6f-i). Similarly, targeting
microtubule with MAP4-pFAST allowed us to image the microtubule network within live astrocytes
with subdiffraction resolution (Figure III.6j-l). Overall this set of experiments highlighted the potential
of pFAST:HBR3,5DOM for imaging proteins in live cells, including fragile neurons, with sub-diffraction
resolution by STED nanoscopy.
72
a confocal d confocal e deconvolved STED j
b STED
f confocal g deconvolved STED
k confocal
c deconvolved STED h confocal i deconvolved STED l deconvolved STED
confocal
100 a,b,c 100 d,e 100 f,g 100 h,i 100 k,l
STED
deconvolved
Intensity (a.u.)
STED
50 50 50 50 50
0 0 0 0 0
0 1 2 0 1 2 3 4 0.0 0.5 1.0 0 1 2 1 2 3 4 5
distance (µm) distance (µm) distance (µm) distance (µm) distance (µm)
Figure III.6. STED nanoscopy of pFAST-tagged proteins in live mammalian cells, live cultured neurons
and live astrocytes. (a-c) Confocal and STED micrographs of live HeLa cells expressing Lyn11-pFAST. (d,e)
Confocal and STED micrographs of HeLa cells expressing MAP4-pFAST. (f,g) Confocal and STED micrographs
of live dissociated hippocampal neuron expressing LifeAct-pFAST. (h,i) Confocal and STED micrographs of a
neurite growth cone expressing LifeAct-pFAST. (j-l) Confocal and STED micrographs of a live astrocyte
expressing LifeAct-pFAST. Cells were treated with 10 µM of HBR-3,5DOM before imaging. All scale bars are 5
µm, except the one in (i) that is 20 µm. (a-l) See Table III.S13 for imaging settings. Graphs show gain in
resolution.
73
III.2.4. Discussion
Spectral tuning of chromophoric proteins can be achieved though modulation of the chemical structure
of the embedded chromophore or by mutating amino acid residues in the chromophore vicinity. Here
we report the successful engineering of pFAST, a fluorescent chemogenetic reporter with highly
tunable spectral and optical properties. pFAST is a promiscuous variant of the protein tag FAST
engineered to efficiently recognize a palette of chromophores exhibiting various electronic properties,
and to form tight bimolecular assemblies with optical properties spanning the entire visible spectrum.
The use of isosteric electron-withdrawing heterocycles with various electronic properties allowed
extensive spectral tuning with minimal structural changes. A massive engineering effort mostly
conducted by directed evolution further enabled us to develop a versatile tag showing good-to-
excellent binding/spectroscopic properties with the entire set of fluorogenic compounds. pFAST was
evolved via a rapid and easy-to-implement fluorescence screening strategy based on yeast-display
and FACS, further supporting the power of molecular evolution. In agreement with an increase of the
chromophore binding promiscuity, pFAST showed improved properties with the chromophores of the
prototypical FAST. In particular, efficient labeling in cells and in multicellular organisms can be
achieved at lower chromophore concentration because of a gain in binding affinity, while longer or
more power-demanding imaging experiments can be performed thanks to superior photostability.
The extended chromophore promiscuity of pFAST provides an unprecedented experimental

versatility, allowing investigators to face a large variety of experimental scenarios. The panoply of
spectrally distinguishable chromophores enables to fine tune pFAST spectral properties to optimize
multicolor imaging. Selective labeling of cell-surface pFAST-tagged proteins can be efficiently
achieved using bright membrane-impermeant fluorogenic chromophores. The discovery of
chromophores forming tight absorbing-only dark assemblies provides moreover new opportunities to
control the fluorescence state of pFAST. In particular, the possibility to replace bright chromophores
with dark ones by direct competition for the binding site allows one to switch off pFAST fluorescence
in cells with an unprecedented time resolution. The ability to form dark complexes absorbing in the
green region opens furthermore interesting prospects for the design of dark acceptors for Förster
resonance energy transfer (FRET) measurements based on fluorescence lifetime imaging
microscopy (FLIM). Like Shadow Green and Yellow, dark acceptors previously engineered from GFP-
like proteins34,35, pFAST labeled with a dark chromophore could find applications as dark acceptor to
modulate the fluorescence lifetime of a donor without blocking an imaging channel. The use of a
chemogenetic system such as pFAST has the additional advantage to allow measurements in
presence and absence of the acceptor by addition or not of the ligand, and this in a single experiment,
eliminating the need for acceptor-free control experiments.
74
Our study demonstrates that pFAST was perfectly suited to image fusion proteins in live and fixed
mammalian cells, including delicate cells such as primary cultured neurons, and in multicellular
organisms such as chicken embryo. We show moreover that pFAST can be used to follow dynamic
processes in real-time in cells and in embryo tissues.
Finally, we demonstrated that pFAST enabled the imaging of structures below the diffraction barrier
in live mammalian cells and live cultured neurons using 3D STED nanoscopy, whose broad use is
currently limited by the low numbers of fluorophores displaying compatible spectral properties and
photostability. The labeling of pFAST with HBR-3,5DOM allowed the successful imaging of
membranes, microtubules and actin filaments with sub-diffraction resolution. Beyond applications in
STED nanoscopy, we anticipate that a complete and systematic study of the on-off fluorescence
blinking or photoswitching behaviors of pFAST could open new prospects for other advanced imaging
techniques, such as single-molecule tracking and super-resolution microscopy by single-molecule
localization microscopy32,36,37 or selective imaging exploiting the photokinetic dynamics of fluorescent
probes29,30.
75
III.2.5. Supplementary information
This part includes:

Legends for Movies III.S1 to III.S7
Texts III.S1 to III.S2
Figures III.S1 to III.S19
Tables III.S1 to III.S13
All the movies presented in this study are available on BioRxiv website:
https://www.biorxiv.org/content/10.1101/2021.01.29.428635v2.supplementary-materialb
b
Link for the online movies
76
Legends for movies
Movie III.S1. Labeling efficiency of pFAST and FAST with HBR-3,5DOM in dissected neural
tube of chicken embryo. Plasmids encoding H2B-pFAST and H2B-FAST were each electroporated
in one of each side of the neural tube in ovo at embryonic day 2 (E2, HH stage 13-14). An EGFP
reporter fused to H2B was co-injected with each construct to monitor electroporation efficiency. 24 h
later, embryos with homogeneous bilateral EGFP expression in the neural tube were dissected, and
imaged upon successive additions of 0.1, 1 and 10 µM HBR-3,5DOM using a spinning-disk confocal
microscope (see Table III.S13 for imaging settings).
Movie III.S2. Two-color dynamic cellular processes in dissected neural tube of chicken
embryo. Plasmids encoding mito-pFAST (fused to a mitochondrial localization signal) and mb-
mIRFP670 (fused to a membrane localization signal) were electroporated in the neural tube in ovo,
at embryonic day 2 (E2, HH stage 13-14). 24 h later, embryos were dissected, and imaged in presence
of 1 µM HBR-3,5DOM using a spinning-disk confocal microscope (see Table III.S13 for imaging
settings). The time-lapse shows cell division. Scale bar, 10 µm.
Movie III.S3. Two-color imaging in dissected neural tube of chicken embryo. Plasmids encoding
H2B-pFAST (fused to histone H2B) and pact-mKO (targeting the PACT domain of pericentrin) were
electroporated in the neural tube in ovo, at embryonic day 2 (E2, HH stage 13-14). 24 h later, embryos
were dissected, and imaged in presence of 5 μM HBP-3,5DOM using a spinning-disk confocal
microscope (see Table III.S13 for imaging settings). The time-lapse shows cell division. Scale bar,
10 µm.
Movie III.S4. Three-color imaging in dissected neural tube of chicken embryo. Plasmids
encoding H2B-pFAST (fused to histone H2B), pact-mKO (targeting the PACT domain of pericentrin)
and mb- miRFP670 (fused to a membrane localization signal) were electroporated in the neural tube
in ovo at embryonic day 2 (E2, HH stage 13-14). 24 h later, embryos were dissected, and imaged in
presence of 1 μM HMBR using a spinning-disk confocal microscope. The time-lapse shows cell
division (see Table III.S13 for imaging settings). Scale bar, 10 µm.
Movie III.S5. Reversible labeling of pFAST in live mammalian cells by chromophore

replacement. Time-lapse imaging of HeLa cells expressing cytoplasmic pFAST initially labeled with
1 µM HBR-3,5DM, 10 µM HBP-3,5DOM, 1 µM HMBR and 5 µM HBP-3,5DM were switched off upon
addition of 10 µM HBIR-3M dark-competitor (the concentration of fluorogenic chromophore was kept
77
constant during the overall experiment). The transmitted channels allowed to visualize the cells (see
Table III.S13 for imaging settings). Scale bars, 30 µm.
Movie III.S6. Reversible labeling of pFAST in dissected neural tube of chicken embryo. Plasmid
encoding H2B-pFAST was electroporated in the neural tube in ovo at embryonic day 2 (E2, HH stage
13-14). An mRFP reporter fused to H2B was co-injected to monitor electroporation efficiency. 24 h
later, embryos with homogeneous bilateral mRFP expression in the neural tube were dissected, and
initially labeled with 1 µM HMBR for 40 minutes. Time-lapse imaging allowed to monitor the
fluorescence evolution after washing the embryos with PBS and addition of fresh medium
supplemented with 0 µM (“wash”), 1 µM and 10 µM of the dark competitor HBIR-3M prior to imaging
(see Table III.S13 for imaging settings).
Movie III.S7. Rapid dynamics of microtubules and membranes in live cells observed with high
temporal and spatial resolution by Airyscan confocal microscopy. Time-lapse imaging of HeLa
cells expressing lyn11-pFAST (fused to membrane localization signal) (left) and MAP4-pFAST
(microtubule associated protein) (right) and labeled with 5 µM HBR-3,5DOM allowed to visualize the
rapid dynamics of membrane and microtubules with high temporal and spatial resolution by Airyscan
confocal microscopy (see Table III.S13 for imaging settings). Scale bars, 10 µm.
78
Text III.S1: Full description of the directed protein evolution experiments leading to oFAST,
tFAST and pFAST.
This supplementary text complements the main text, and details the full directed evolution
experiments done during this study.
We used a combinatorial library of 106 variants of FAST generated by random mutagenesis and
displayed on yeast cells (library A). Six different screenings in presence of either HBO-3M, HBO-
3,5DM, HBT-3M, HBT-3,5DM, HBP-3M or HBP-3,5DM were performed by iterative rounds of
fluorescence activated cell sorting (FACS) decreasing progressively chromophore concentrations
through rounds (typically from 10 µM to 2.5 µM) in order to identify variants forming tighter and brighter
assemblies (Figures III.S2 and III.S3). The screenings with HBO-3M, HBO-3,5DM and HBP-3,5DM
showed an increase in cell population fluorescence indicating the selection of improved variants
(Tables III. S5-S7). For each positive screening, we systematically isolated and sequenced twenty-
four clones after the fifth, sixth and seventh round of FACS, and then further analyzed the
performances of single clones by analytical flow cytometry. Five to ten individual clones with the
highest fluorescence performances relative to FAST were expressed in bacteria and purified for in
vitro characterization. The variants isolated from the three screenings formed tighter and brighter
assemblies with the chromophore used for their selection (Figures III.2b,d, Figure III.S4a, Tables
III.S5-7), in agreement with a successful molecular evolution. By screening variants that combined
mutations beneficial for the binding of HBO-3M (Table III.S5) and HBO-3,5DM (Table III.S6), we
identified oFAST, an improved variant with the mutations Q41L, D71N, V83I, M109L and S117R,
showing improved properties with both HBO-3M and HBO-3,5DM. oFAST binds HBO-3,5DM with a
KD of 3.0 µM, forming a blue fluorescent complex with 411/482 nm abs/em peaks and a fluorescent
quantum yield f = 30%, and it binds HBO-3M more tightly with a KD of 0.73 µM, forming a blue
fluorescent complex with 394/470 nm abs/em peaks and a fluorescent quantum yield f = 11% (Table
III.S2).
The third positive selection allowed us to identify several variants able to bind HBP-3,5DM with 3 to 7
fold higher affinity, and giving complexes displaying higher fluorescence quantum yield. Beneficial
mutations isolated from the three brighter variants were combined, leading eventually to three new
variants binding HBP-3,5DM about 10-fold tighter than FAST and leading to higher fluorescent
quantum yields (from f = 22 to 26 %) (Table III.S7). The five more promising variants were further
characterized with the other members of the HBP series and with the chromophores of the HBT series.
We discovered that the five variants were also able to form tighter and brighter assemblies with HBT-
3,5DM (Table III.S8), HBT-3,5DOM, HBP-3M and HBP-3,5DOM (Table III.S9).
79
We thus used these five variant genes to construct a new library using DNA shuffling, which allows
random DNA fragments recombination to access positive combinations. The five variant genes were
randomly recombined to generate a library of 5.5 ´ 107 variants (library B), which was displayed on
yeast cells. This new library was screened in presence of HBT-3,5DM, HBP-3,5DM and HBP-3,5DOM
by iterative rounds of FACS as described above. The screening with HBP-3,5DOM did not allow us
to identify variants with significantly better properties (Figure III.S4b, Table III.S9). However, the
screening with HBT-3,5DM allowed us to identify variants able to form brighter complexes (Figure
III.2c, Table III.S8). The variant possessing the mutations G25R, Q41K, S72T, A84S, M95A, M109L
and S117R relative to FAST, which we ultimately named tFAST, showed the most advantageous
properties: it was shown to form a tight cyan fluorescent assembly with HBT-3,5DM (KD = 0.33 µM,
435/497 nm abs/em peaks, f = 11%) (Table III.S3). On the other hand, the screening of the library B
with HBP-3,5DM permitted us to identify an improved variant bearing the mutations K17N, G21E,
G25R, A30V, Q41L, S72T, V83A, M95T, S117R, which was further refined by introduction of the
mutation M109L found in several other improved variants. The resulting variant, named pFAST, binds
HBP-3,5DM tightly with a KD of 0.15 µM forming a bright green fluorescent assembly (f = 27%,
465/520 nm abs/em peaks) (Table III.S4, Table III.S7).
80
Text III.S2: Theoretical model for the determination of the thermodynamic dissociation
constants of the dark chromophores
The thermodynamic dissociation constants (KD) of the dark HBIR-3M and HBIR-3,5DM chromophores
were determined by determining the apparent dissociation constant of HBP-3,5DM in presence of
various concentrations of dark competitors. Considering that the fluorogen HBP-3,5DM and the
protein interact to provide a fluorescent complex whereas the dark chromophores HBIR-3M or HBIR-
3,5DM and the protein interact to form a non-fluorescent complex, we adopted the following three-
state model
F1 + R ⇆ F1 R
+
F2
↑↓
F2 R
where F1 denotes the fluorogen, F2 denotes the dark chromophore, R denotes the protein and F1R
and F2R denote the two possible complexes, characterized by the dissociation constants
[F1 ] [R]
KD,1 =
[F1 R]
and
[F2 ] [R]
KD,2 =
[F2 R]
where [X] is the concentration of the species X at equilibrium. At equilibrium, the fraction of protein R
binding F1 is given by
[F1 ] [R]
[F1 R] KD,1 [F1 ]
B= = =
[R]+ [F1 R] + [F2 R] [F1 ] [R] [F ] [R] [F ]
[R] + + 2 [F1 ] +KD,1 %1 + 2 &
KD,1 KD,2 KD,2
[F1 ]
=
[F1 ] + KD,app
with
[F2 ]
KD,app = KD,1 '1+ (
KD,2
By choosing [F1]0 and [F2]0 >> [R]0, one can rewrite this equation assuming that [F1] » [F1]0 and [F2] »
[F2]0. Knowing KD,1 , determination of the apparent thermodynamic dissociation constant KD,app at
various [F2]0 enabled to determine KD,2 .
81
Figure. III.S1. Absorption spectra of fluorogenic chromophore at various pH. Absorption spectra of (a)
HBP-3M, (b) HBP-3,5DM, (c) HBIR-3M, (d) HBT-3M, (e) HBT-3,5DM, (f) HBO-3M and (g) HBO-3,5DM in
solution in function of pH. The spectra were recorded in 0.01 M Britton-Robinson buffer38 (0.1 M ionic strength)
at 25°C. (h) Extracted pKA values.
82
Figure III.S2. Evolutionary tree of the selections with HBO-3M (in purple), HBO-3,5DM (in blue), HBT-
3,5DM (in cyan), HBP-3,5DM (in green) and HBP-3,5DOM (in orange). Selected clones were generated by
directed evolution from initial libraries of mutants (solid arrows). The first library was constructed by random
mutagenesis from the original FAST sequence (library A) and the second library was designed by DNA shuffling
from five mutants, initially selected and rationally designed from the selection for HBP-3,5DM binders (library
B). Additional mutations were introduced by rational design (dotted arrows). The crosses indicate unsuccessful
selections. The mutations relative to FAST are indicated for each clone.
83
Figure III.S3. Sequence alignment of FAST and the clones identified from the selections with (a) HBO-3M, (b)
HBO-3,5DM, (c) HBP-3,5DM, (d) HBT-3,5DM and (e) HBP-3,5DOM.
84
Figure III.S4. Characteristics of the selected clones. Thermodynamic dissociation constants (KD) and
fluorescent quantum yields (f) of the clones isolated from the selections with (a) HBO-3M and (b) HBP-3,5DOM.
Values are also given for FAST for comparison. (a) Mutants of HBO-3M selection were isolated from the library
A (black triangles). (b) Mutants of HBP-3,5DOM were isolated from the library B (black squares) and compared
to the shuffling mutants used to design the library B (gray triangles).
85
Figure III.S5. Absorption and emission properties. Absorption (dashed lines) and fluorescence (solid lines)
spectra of the fluorogenic chromophores when free in solution (dark line) and bound to pFAST (colored lines)
for (a) HBIR-3,5DOM, (b) HBR-3,5DOM, (c) HBRAA-3,5DM, (d) HBIR-3,5DM, (e) HBIR-3M, (f) HBR-3,5DM,
(g) HBP-3,5DOM, (h) HBRAA-3E, (i) HBRAA-3M, (j) HMBR, (k) HBT-3,5DOM, (l) HBP-3,5DM, (m) HBP-3M,
(n) HBT-3,5DM, (o) HBO-3,5DM and (p) HBO-3M. Spectra were recorded using chromophore concentrations
ranging from 3 – 15 µM and pFAST at 40 µM in pH 7.4 phosphate buffer saline at 25°C.
86
Figure III.S6. Structural homology models. (a) Structural models of FAST and pFAST generated from the
tridimensional crystal structure of the Halorhodospira halophila Photoactive Yellow Protein PYP (PDB: 6P4I).
(b) Volumes of the ligand binding sites of PYP, FAST and pFAST. (c) In silico binding affinities (ligscore 2) of
FAST and pFAST for the HBIR, HBR, HBR, HBT and HBO series. (d-m) Mapping of the residues involved in
polar (green dashed lines) and apolar interactions (purple dashed lines) with the residues found mutated in
pFAST. (n) Molecular docking of pFAST-chromophore assemblies.
87
Figure III.S7. Two-color fluorescence viability assay. HeLa cells were incubated for 24 h with solutions of
(a-b) 0.1 % of DMSO, (c) HBO-3M at 5 µM, (d) HBO-3,5DM at 10 µM, (e) HBP-3M at 5 µM, (f) HBP-3,5DM at
5 µM, (g) HBP-3,5DOM at 10 µM, (h) HBT-3M at 5 µM, (i) HBT-3,5DM at 5 µM, (j) HBT-3,5DOM at 5 µM , (k)
HBIR-3M at 5 µM, (l) HBIR-3,5DM at 5 µM and (m) HBIR-3,5DOM at 5 µM. Control experiments of HeLa cells
non-incubated with dye (a, live cells) or incubated for 30 min with 0.1 mg/ml digitonin (b, dead cells) are shown.
Cell viability was tested with calceinAM and EthD1 probes (LIVE/DEADâ viability/cytotoxicity assay kit).
CalceinAM is a cell-permeant profluorophore cleaved by intracellular esterases releasing a uniform green
fluorescence polyanionic calcein in live cells (green channel). EthD1 (Ethidium homodimer-1) is a non-permeant
88
nucleic acid fluorescent stain that enters only cells with damaged membranes and undergoes a fluorescence
enhancement upon binding to nucleic acids, thereby producing a bright red fluorescence in dead cells (red
channel). Cell fluorescence was evaluated by confocal microscopy. Identical imaging settings were used for the
different experiments. The experiment shows that none of the chromophores are toxic for HeLa cells at the
concentrations used for imaging.
a b c
HBR-3,5DOM HBR-3,5DM
1000 40000 1000 ns
Fluorescence (a.u.)
50000
80 ns
** live fixed live fixed
60 100 21500 100 26500
S/B
55000
40
10 3000 10
3000
20 29000
1 1
0 live fixed live fixed
live fixed
3000
d e
HMBR HBP-3,5DOM
1000 ns ** 40000
live fixed 40000
100 live fixed
live
100 21500 21500
S/B
10
10 3000 3000
1 1
live fixed live fixed
f g
HBP-3,5DM HBP-3M
100 ns 60000 100 ns
live fixed live fixed 25000
35000 21500
S/B
10 10
fixed 10000 3000
1 1
h HBT-3,5DOM i HBT-3,5DM
ns ns 15000
100 live fixed 40000 100 live fixed
21500 9000
S/B
10 10
3000 3000
1 1
Figure III.S8. Imaging of H2B-pFAST in live and fixed HeLa cells after addition of various fluorogenic
chromophores. (a) Nuclear fluorescence in live and fixed cells incubated with 5 µM of HMBR and imaged with
identical microscope settings (n = 64 cells, 2 experiments). Median values are reported and a Wilcoxon test was
conducted to compare live and fixed cell populations (ns = not significative). Scale bars, 40 µm. (b-i) Comparison
of signal (nucleus) to background (cytosol) (S/B) ratio between live and fixed cells incubated with (b) HBR-
3,5DOM at 5 µM, (c) HBR-3,5DM at 5 µM, (d) HMBR at 5 µM, (e) HBP-3,5DOM at 10 µM, (f) HBP-3,5DM at 5
µM, (g) HBP-3M at 5 µM, (h) HBT-3,5DOM at 5 µM and (i) HBT-3,5DM at 5 µM. For each experiment, live and
fixed cells were imaged with identical microscope settings (n = 24 cells, 2 experiments). Median of S/B ratios
are reported and a Wilcoxon test was conducted to compare live and fixed cell populations (ns = not significative,
* p < 0.05 and ** p < 0.01). (see Table III.S13 for imaging settings). Scale bars, 10 µm.
89
a HBRAA-3E HBRAA-3E b HBRAA-3M HBRAA-3M c HBRAA-3,5DM HBRAA-3,5DM

+ HMBR + HMBR + HBR-3,5DM
pFAST
FAST
Figure III.S9. Selective imaging of cell-surface proteins. Confocal micrographs of HeLa cells expressing a
secreted transmembrane domain fused to pFAST versus FAST labeled with impermeant fluorogens (a) HBRAA-
3E at 10 µM, (b) HBRAA-3M at 10 µM and (c) HBRAA-3,5DM at 10 µM. Subsequent addition of 5 μM of
membrane-permeant HMBR (a,b) or HBR-3,5DM (c) revealed the total pool of proteins expressed at the surface
and within the secretory pathway. For each experiment, FAST and pFAST were imaged under the same imaging
settings. The detection settings were adjusted to take into account the difference of brightness of impermeant
and membrane-permeant fluorogens for each experiment. (see Table III.S13 for imaging settings). Scale bars,
10 µm.
90
Figure III.S10. In-cell photostability of pFAST. (a-c) Photostability of pFAST compared to FAST and iFAST
in presence of 10 µM HBR-3,5DOM (a), 5 µM HBR-3,5DM (b) and 5 μM HMBR (c). Photostability of EGFP is
also given for comparison. Cells were illuminated with a 488 nm laser excitation (with a power of 4.4 kW/cm2 at
the specimen plane) and 500 images were acquired every 2 s, n = 3 cells per curve. (d) Photostability of pFAST
in presence of 10 µM HBP-3,5DOM and 5 μM HBP-3,5DM under 488 nm laser excitation (with a power of 4.4
kW/cm2 at the specimen plane). Photostability of EGFP is also given for comparison. 500 images were acquired
every 2 s, n = 3 cells per curve. (e) Photostability of pFAST in presence of 5 µM HBT-3,5DOM and 5 µM HBP-
3M under continuous 458 nm laser excitation (with a power of 3.5 kW/cm2 at the specimen plane). Photostability
of mTurquoise2 is also given for comparison. 500 images were acquired every 2 s, n = 3 cells per curve. (f-g)
Photostability of pFAST in presence of 10 µM HBP-3,5DOM (f) and 5 µM HBP-3,5DM (g) under 488 nm laser
excitation with a power of 4.7 kW/cm2 (black and green curves) and 2.4 kW/cm2 (pink and cyan curves) at the
specimen plan. 150 images were acquired every 2s (black and pink curves) and 5s (green and cyan curves)
followed by 60s in the dark before acquisition was restarted. (h-i) Photostability of pFAST in presence of 5 µM
HBT-3,5DOM (h) and 5 µM HBP-3M (i) under 458 nm laser excitation with a power of 3.7 kW/cm2 (black and
green curves) and 2.4 kW/cm2 (pink, cyan and gray curves) at the specimen plan. 150 images were acquired
every 2 s (black and pink curves), every 5 s (green and cyan curves) and every 10 s (gray curve only for HBT-
3,5DOM) followed by 60 s in the dark before acquisition was restarted. (a-i) The proteins were expressed in the
cytoplasm in HeLa cells and images were acquired using a scanning confocal microscope with a pixel dwell
time of 2.55 µs (see Table III.S13 for imaging settings).
91
pFAST H2B-pFAST Lyn11-pFAST
LifeAct-pFAST Mito-pFAST MAP4-pFAST
Figure III.S11. Selective imaging of pFAST in live mammalian cells. Confocal micrographs of live HEK293T
cells expressing pFAST fused to: histone H2B, lyn11 (inner membrane-targeting motif), LifeAct (actin binding
peptide domain), mito (mitochondrial targeting motif) and to microtubule-associated protein (MAP) 4 and labeled
with 5 µM HBP-3,5DM (see Table III.S13 for imaging settings). Scale bars, 10 µm.
a HaloTag-TMR b mVenus c pFAST-HBP-3,5DM

live
fixed
Figure III.S12. Confocal micrographs of live and fixed HeLa cells expressing MAP4 fused to (a) HaloTag labeled
with 2.5 µM of HaloTag® TMR Ligand, (b) mVenus and (c) pFAST labeled with HBP-3,5DM at 5 µM (see Table
III.S13 for imaging settings). Scale bars, 10 µm.
92
Figure III.S13. Selective imaging of hippocampal neuronal network. Tiles of confocal micrographs of
dissociated hippocampal neurons transfected with a plasmid encoding pFAST fused to lyn11 (inner membrane
targeting motif), and labeled with 10 µM HBR-3,5DOM (see Table III.S13 for imaging settings). The huge field
of view enables to illustrate the high transfection rate and viability of transfected fragile cells like neurons.
Artificial look up table ‘fire’ color is used to illustrate intensity dynamics. Scale bar, 200 µm.
93
Figure III.S14. Labeling efficiency in live cells. Fluorescence of HEK 293T cells expressing cytoplasmic
pFAST (colored lines) versus FAST (black lines) in the presence of increased concentrations of (a) HBR-
3,5DOM, (b) HBR-3,5DM, (c) HMBR, (d) HBP-3,5DOM and (e) HBP-3,5DM (n = 3 replicates). Fluorescence
was measured by flow cytometry.
94
Figure III.S15. Labeling efficiency in chicken embryos. (a-f) Plasmids encoding H2B-pFAST and H2B-FAST
were electroporated in each side of the neural tube in ovo at embryonic day 2 (E2, HH stage 13-14). (b) EGFP
or (c-f) mCherry reporters were co-injected with each construct as a transfection efficiency control. 24 h later,
embryos with homogeneous bilateral reporter expression in the neural tube were dissected and imaged. Time-
lapse imaging upon sequential addition of fluorogenic chromophore were acquired using (b) spinning-disk
confocal or (c-f) widefield fluorescence microscopy. Conditions: (b) 0.1, 1 and 10 µM HBR-3,5DOM (see also
Movie III.S1) ; (c) 0.2, 1 and 10 µM HBR-3,5DM ; (d) 0.2, 1 and 10 µM of HBP-3,5DOM ; (e) 0.1, 1 and 10 µM
of HMBR and (f) 0.2, 1 and 10 µM of HBP-3,5DM (see Table III.S13 for imaging settings). (g-k) Fluorescence
intensities of pFAST and FAST were analyzed over time and normalized by the maximal fluorescence value of
pFAST for each fluorogenic chromophore. The sequential addition of the fluorogenic chromophore at the
different concentrations are indicated by arrows.
95
Figure III.S16. Comparison of pFAST with various fluorescent proteins in chicken embryos. Embryos
expressing pFAST and the fluorescent proteins (a) mKO, (b,c) EYFP, (d,e) EGFP and (f) mCerulean in each
side of the neural tube were dissected and labeled by sequential additions of various concentrations of
fluorogenic chromophores (arrows indicate additions). (a) 0.1, 1, 10 µM of HBR-3,5DOM; (b) 0.2, 1, 10 µM of
HBR-3,5DM; (c) 0.2, 1, 10 µM of HBP-3,5DOM; (d) 0.2, 1, 10 μM of HMBR; (e) 0.2, 1, 10 µM of HBP-3,5DM
and (f) 0.2, 1, 10 µM of HBP-3M. Time-lapse (a) spinning-disk confocal or (b-f) widefield fluorescence imaging
allowed to monitor the temporal evolution of the fluorescence intensity (Figure III.5 in main text shows the initial
and final images). Fluorescence intensities were normalized by the maximal fluorescence value of pFAST for
each experiment.
96
Figure III.S17. Reversible labeling of pFAST in live cells by chromophore replacement. (a-d) Confocal
micrographs of HeLa cells expressing cytoplasmic pFAST - initially labeled with (a) 1 µM HBR-3,5DM, (b) 10
µM HBP-3,5DOM, (c) 1 µM HMBR and (d) 5 µM HBP-3,5DM - acquired before and after addition of 10 µM
HBIR-3M dark-competitor (the concentration of fluorogenic chromophore was kept constant during the overall
experiment) (see Table III.S13 for imaging settings). Scale bars, 10 µm. (e-h) Temporal evolution of
fluorescence intensities upon addition of HBIR-3M (or mock solutions) as indicated by arrows (n = 3 cells) (see
also Movie III.S5). Note that the artefactual jump of fluorescence on (h) (grey area) was due to a change of
focus upon addition of the HBIR-3M solution or the mock solution.
97
Figure III.S18. Reversible labeling of pFAST in chicken embryos by chromophore replacement. Embryos
expressing H2B-pFAST in the neural tube were dissected and labeled with (a) 5 µM of HBP-3,5DOM and with
(b) 1 µM of HMBR for 40 minutes (see also Movie III.S6). The labeling solution was then removed, the samples
were washed once with PBS and fresh medium supplemented with 0 µM, 1 µM and 10 µM of the dark competitor
HBIR-3M was added. Time-lapse spinning-disk confocal imaging allowed to monitor the temporal evolution of
the fluorescence intensity after addition of the dark competitor (see Table III.S13 for imaging settings). The
graphs show the temporal evolution of the fluorescence intensities.
98
a HMBR d HMBR
confocal
confocal
airyscan
airyscan
b HBR-3,5DOM e HBP-3,5DOM
confocal
confocal
airyscan
airyscan
c HBR-3,5DM f HBP-3,5DM
confocal
confocal
airyscan
airyscan
Figure III.S19. Airyscan confocal imaging of live mammalian cells expressing (a-c) MAP4-pFAST
(microtubule associated protein) fusion or (d-f) Lyn11-pFAST (membrane targeting motif) fusion and labeled
with (a,d) 5 µM HMBR, (b) 5 µM HBR-3,5DOM, (c) 5 µM HBR-3,5DM, (e) 10 µM HBP-3,5DOM and (f) 5 µM
HBP-3,5DM (see Table III.S13 for imaging settings). Scale bars, 10 µm.
99
Table III.S1. Properties of FAST with various chromophores in PBS pH 7.4.
labs e Molecular
Chromophore Dlabs (nm) lem (nm) f KD (µM)
(nm) (mM-1cm-1) brightness
HBR-3,5DOM (ref 21) 520 114 600 39 0.31 12,000 0.97

HBR-3,5DM (ref 13) 499 97 562 48 0.49 24,000 0.08
HMBR (ref 13) 481 79 540 44 0.23 10,000 0.13
HBP-3,5DOM 488 121 556 31 0.29 9,000 ~100
HBP-3,5DM 464 105 522 30 0.16 4,000 3.7
HBP-3M 447 87 506 21 0.044 800 6.9
HBT-3,5DOM 449 92 532 54 0.22 12,000 3.3
HBT-3,5DM 434 82 500 25 0.072 2,000 1.7
HBT-3M 415 - 480 32 0.006 200 0.24
HBO-3,5DM 409 80 481 4 0.20 700 32
HBO-3M 394 68 471 11 0.056 600 6.4
Abbreviations are as follows: labs wavelength of maximal absorption; Dlabs = labs,bound - labs,unbound absorption red-shift upon chromophore
binding; lem wavelength of maximal emission; e, molar absorptivity at labs (standard error is typically 10%); f, fluorescence quantum yield;
molecular brightness = f ´ e; KD thermodynamic dissociation constant.
Table III.S2. Properties of oFAST with various chromophores in PBS pH 7.4
e Molecular
Chromophore labs (nm) Dlabs (nm) lem (nm) f KD (µ M)
(mM-1cm-1) brightness
HBR-3,5DOM 519 115 600 44 0.33 14,000 0.1

HBR-3,5DM 500 98 560 53 0.38 20,000 0.01
HMBR 481 79 540 47 0.24 11,000 0.19
HBP-3,5DOM 489 122 556 28 0.30 8,000 3.0
HBP-3,5DM 465 104 519 33 0.21 7,000 0.20
HBP-3M 445 85 505 40 0.054 2,000 0.37
HBT-3,5DOM 449 92 533 68 0.21 15,000 0.30
HBT-3,5DM 435 83 501 31 0.090 3,000 0.16
HBO-3,5DM 411 82 482 6 0.30 2,000 3.0
HBO-3M 394 68 470 15 0.11 1,600 0.73
binding; lem wavelength of maximal emission; e molar absorptivity at labs (standard error is typically 10%); f fluorescence quantum yield;
100
Table III.S3. Properties of tFAST with various chromophores in PBS pH 7.4.
e Molecular
Chromophore labs (nm) Dlabs (nm) lem (nm) f KD (µM)
HBR-3,5DOM 520 115 600 46 0.33 15,000 0.07

HBR-3,5DM 501 99 561 56 0.44 25,000 0.01
HMBR 481 79 541 52 0.24 12,000 0.19
HBP-3,5DOM 488 121 555 28 0.30 8,000 4.0
HBP-3,5DM 464 104 522 33 0.25 8,000 0.36
HBP-3M 447 87 505 39 0.075 3,000 0.56
HBT-3,5DOM 449 92 532 60 0.24 15,000 0.44
HBT-3,5DM 435 83 497 30 0.11 3,000 0.33
HBO-3,5DM 410 81 485 6 0.22 1,000 5.0
HBO-3M 393 67 471 13 0.069 1,000 0.8
binding; lem wavelength of maximal emission; e molar absorptivity at labs (standard error is typically 10%); f fluorescence quantum yield;
Table III.S4. Properties of pFAST with various chromophores in PBS pH 7.4.
e Molecular
Chromophore labs (nm) Dlabs (nm) lem (nm) f K D ( µ M)
HBIR-3,5DOM 562 128 616 40 0.10 3,000 0.04

HBIR-3,5DM 536 106 578 25 0.015 400 0.07
HBIR-3M 514 84 567 12 0.003 50 0.005
HBR-3,5DOM 520 115 600 44 0.35 15,000 0.06
HBR-3,5DM 501 99 561 49 0.44 22,000 0.01
HMBR 481 79 542 54 0.23 13,000 0.01
HBRAA-3,5DM 524 118 578 58 0.22 12,000 1.8
HBRAA-3E 506 96 558 53 0.05 3,000 0.05
HBRAA-3M 502 96 554 64 0.08 5,000 0.23
HBP-3,5DOM 487 120 554 37 0.33 12,000 1.9
HBP-3,5DM 465 105 520 37 0.27 10,000 0.15
HBP-3M 447 87 503 35 0.10 4,000 0.22
HBT-3,5DOM 449 92 532 60 0.27 16,000 0.20
HBT-3,5DM 433 81 499 33 0.10 3,000 0.17
HBO-3,5DM 411 82 483 6 0.23 1,000 2.4
HBO-3M 392 66 473 14 0.076 1,000 0.43
binding; lem wavelength of maximal emission; e, molar absorptivity at labs (standard error is typically 10%); f fluorescence quantum yield;
101
Table III.S5. Properties of the mutants isolated from the selection with HBO-3M compared to FAST in PBS pH
7.4.
labs lem e Molecular

Clone Mutations (relative to FAST) f KD (µM)
(nm) (nm) (mM-1cm-1) brightness
FAST 394 471 11 0.056 600 6.4

A-R6.8 D65V / E93D / M109L / S117R 395 468 15 0.090 1,300 2.6
A-R6.11 V83I / M109L 394 468 15 0.074 1,100 1.4
A-R7.1 M109L 394 470 13 0.081 1,000 4.0
A-R7.2 Q41L / M109L 394 468 12 0.092 1,000 1.1
A-R7.10 V83I / T103I 394 468 12 0.088 1,000 4.4
A-R7.12 K17R / A30V / E74K / E93V 386 466 13 0.097 1,200 2.0
A-R7.14 T50S / E93Q / M95I 382 468 13 0.077 1,000 1.4
Abbreviations are as follows: labs wavelength of maximal absorption; lem wavelength of maximal emission; e molar absorptivity at labs
(standard error is typically 10%); f fluorescence quantum yield; molecular brightness = f ´ e; KD thermodynamic dissociation constant.
Table III.S6. Properties of the mutants isolated from the selection with HBO-3,5DM compared to FAST in PBS
pH 7.4.

FAST 409 481 4 0.20 800 32

A-R6.6 V83I / M109L 402 481 5 0.22 1,200 11
A-R6.18 H3Q / D48A / T103I / M109L 408 479 5 0.22 1,000 6.2
A-R7.1 N13I / D48E / D71N / V83G / M95V 408 483 5 0.24 1,100 5.0
A-R7.2 Q41L / M95I / M109L 406 481 6 0.25 1,500 6.1
A-R7.5 K60R / V83I / K104R / S117R 406 481 5 0.22 1,000 7.1
A-R7.7 D71N / M109L / S117R 407 482 5 0.25 1,200 4.4
A-R7.7-1 D71N / M109L / S117R + Q41L 413 480 5 0.26 1,300 2.5
A-R7.7-2 D71N / M109L / S117R + Q41L + D48E 411 480 6 0.27 1,500 2.1
A-R7.7-3 D71N / M109L / S117R + Q41L + D65V 411 483 7 0.27 1,800 2.4
A-R7.7-4 = oFAST D71N / M109L / S117R + Q41L + V83I 411 482 6 0.30 2,000 3.0
A-R7.7-5 D71N / M109L / S117R + Q41L + M95I 411 482 5 0.26 1,400 3.0
102
Table III.S7. Properties of the mutants isolated from the selections with HBP-3,5DM compared to FAST in PBS
pH 7.4.

FAST 464 522 30 0.16 4,000 3.7

A-R5.3 Q41H / V83L / M95I 463 521 34 0.19 6,400 0.90
A-R5.7 Q41K / S72T / V83A / M95I 463 525 27 0.23 6,100 0.53
A-R5.12 K17N / Q41K / M95T 463 522 31 0.24 7,400 0.76
A-R5.13
Q41L / S117I 463 523 32 0.21 6,600 0.71
=shuffling mutant 4
A-R5.21
V83E / S117R 463 522 33 0.20 6,400 0.69
=shuffling mutant 5
A-R6.2 E93G / M95V / M109L 464 525 30 0.20 5,800 0.94
A-R7.1 D20N / E81Q / V83I / M109L 464 524 26 0.16 4,300 2.6
A-R7.6 G25R / A84S 464 521 32 0.23 7,100 1.2
A-R7.10 Q32K / Y76N 464 524 32 0.20 6,200 1.1
A-R5.7-1
Q41K / S72T / V83A / M95I + M109L 463 519 33 0.25 8,000 0.25
=shuffling mutant 1
A-R5.7-2 Q41K / S72T / V83A / M95I + S117I 463 521 36 0.26 9,000 0.36
A-R5.7-3 Q41K / S72T / V83A / M95I + M109L + S117I 464 521 34 0.24 8,200 0.20
A-R5.12-1
K17N / Q41K / M95T + S72T + V83A 464 520 34 0.22 7,600 0.30
=shuffling mutant 2
A-R7.6-1
G25R / A84S + Q41K + M95T 465 522 34 0.26 8,700 0.98
=shuffling mutant 3
B-R5.2 K17N / Q41K / S72T / M95T 463 521 30 0.24 7,000 0.35
B-R5.11 K17N / G25E / S72T / V83A / M95T 463 521 32 0.27 8,700 0.31
B-R5.19 Q41K / K80M / A84S / N89D / M95T / M109L 464 519 30 0.24 7,400 0.19
K17N / G21E / G25R / A30V / Q41L / S72T /
B-R6.1 464 520 36 0.27 9,800 0.17
V83A / M95T / S117R
B-R6.10 Q41K / S72T / V83A / S117R 463 520 32 0.22 7,100 0.46
B-R6.15 N13S / Q41K / S72T / V83A / M95T 464 519 29 0.24 7,000 0.32
K17I / G25R / Q41K / A44T / K60R / V83A /
B-R6.19 464 520 35 0.28 9,900 0.14
N89D / E93K / M95T /M109L / S117R
K17N / Q41K / D65E / S72T / V83A / K106M /
B-R7.3 464 520 35 0.23 8,000 0.30
M109L
K17N / G21E / G25R / A30V / Q41L / S72T /
B-R6.1-1 = pFAST 465 520 37 0.27 10,000 0.15
V83A / M95T / S117R + M109L
K17N / G21E / G25E / A30V / Q41L / S72T /
B-R6.1-2 464 521 30 0.31 9,300 0.19
V83A / M95T / S117R
K17N / G21E / G25R / A30V / Q4IK / S72T /
B-R6.1-3 464 521 36 0.27 9,700 0.14
V83A / M95T / S117R
103
Table III.S8. Properties of the mutants isolated from the selection with HBT-3,5DM compared to FAST in PBS
pH 7.4.

FAST 434 500 25 0.072 2,000 1.7

shuffling mutant 1 Q41K / S72T / V83A / M95I + M109L 432 500 28 0.090 2,500 0.26
shuffling mutant 2 K17N / Q41K / M95T + S72T + V83A 434 505 35 0.081 2,800 0.52
shuffling mutant 3 G25R / A84S + Q41K + M95T 433 501 32 0.085 2,700 0.58
shuffling mutant 4 Q41L / S117I 432 501 36 0.078 2,800 0.78
shuffling mutant 5 V83E / S117R 433 500 31 0.075 2,300 0.62
A16D / Q41K / Y76H / V83A / M95T / M109L
B-R6.11 435 497 27 0.10 2,700 0.23
/ S117R
G25R/ Q41K / S72T / A84S / M95A / M109L /
B-R6.13 = tFAST 435 497 30 0.11 3,000 0.33
S117R
B-R6.17 A27V / Q41K/ V83A / M95T / M109L / S117I 435 499 27 0.098 2,700 0.24
K17N / G21E / G25R / A30V / Q41L / S72T /
B-R6.24 434 497 29 0.094 2,700 0.26
V83A / M95T / S117R
B-R7.2 Q41K / K80R / V83A / M95T / M109L 432 496 24 0.095 2,300 0.34
B-R7.4 Q41K / M95T / S117I 432 496 26 0.095 2,500 0.53
Table III.S9. Properties of the mutants isolated from the selection with HBP-3,5DOM compared to FAST in PBS
pH 7.4.

FAST 488 556 31 0.29 9,000 100

shuffling mutant 1 Q41K / S72T / V83A / M95I + M109L 488 555 34 0.36 12,000 2.0
shuffling mutant 2 K17N / Q41K / M95T + S72T + V83A 488 556 30 0.36 11,000 2.1
shuffling mutant 3 G25R / A84S + Q41K + M95T 487 555 30 0.33 9,800 4.4
shuffling mutant 4 Q41L / S117I 485 556 31 0.33 10,000 7.2
shuffling mutant 5 V83E / S117R 488 556 32 0.30 9,500 5.9
G7D / Q41K / S72T / A84S / M95I / M109L /
B-R6.7 488 555 35 0.35 12,000 4.7
S117R / R124L
B-R7.3 K17N / Q41K / A84S / M95T / M109L / S117R 487 555 33 0.37 13,000 2.8
Q41K/ D48G / S72T / V83A / M95T / M109L /
B-R7.7 485 555 37 0.30 11,000 2.7
S117R
G21R / G25R / Q41K / Q60E / M95T / M109L
B-R7.22 490 557 33 0.34 12,000 5.9
/ A112G
104
Table III.S10. Plasmids used in this study

Plasmid Expression host Open reading frame Ref.
pCTCON-FAST Yeast FAST 13

pAG681 Yeast shuffling mutant 1 (libraryA-R5.7-1 [HBP-3,5DM]) this study
pAG684 Yeast shuffling mutant 4 (libraryA-R5.13 [HBP-3,5DM]) this study
pAG685 Yeast shuffling mutant 5 (libraryA-R5.21 [HBP-3,5DM]) this study
pAG686 Yeast pFAST this study
pAG687 Yeast tFAST this study
pAG688 Yeast oFAST this study
pAG104 Mammalian FAST 13
pAG29 Mammalian EGFP 13
pAG490 Mammalian CMV-FRB-NFAST-IRES-mTurquoise2 19
pAG244 Mammalian iFAST 39
pAG654 Mammalian pFAST this study
pAG655 Mammalian tFAST this study
pAG656 Mammalian oFAST this study
pAG657 Mammalian H2B-pFAST this study
pAG660 Mammalian lyn11-pFAST this study
pAG665 Mammalian MAP4-pFAST this study
pAG668 Mammalian LifeAct-pFAST this study
pAG671 Mammalian mito-pFAST this study
pAG897 Mammalian pDisplay-FAST this study
pAG876 Mammalian pDisplay-pFAST this study
pAG719 Mammalian MAP4-mVenus this study
pAG849 Mammalian MAP4-Halotag this study
X-888 Mammalian/bird (CAG promoter) Mito-pFAST this study
X-858 Mammalian/bird (CAG promoter) H2B-pFAST this study
X-889 Mammalian/bird (CAG promoter) H2B-FAST this study
X-892 Mammalian/bird (CAG promoter) H2B-mCerulean this study
pCX-H2B-EGFP Mammalian/bird (CAG promoter) H2B-EGFP 40
X-893 Mammalian/bird (CAG promoter) H2B-EYFP this study
pCX-H2B-mKO Mammalian/bird (CAG promoter) H2B-mKO 41
X-158 Mammalian/bird (CAG promoter) Mb-EGFP 42
X-159 Mammalian/bird (CAG promoter) Mb-mCherry 43
X-736 Mammalian/bird (CAG promoter) Mb-iRFP670 44
pCX-PACT-mKO Mammalian/bird (CAG promoter) PACT-mKO 41
105
Table III.S11. Clones isolated in this study (pET28 plasmids for E. coli expression)
Plasmid Chromophore Library Clone Mutations (relative to FAST)
used for
selection
pAG569 HBO-3M A R6.8 D65V / E93D / M109L / S117R
pAG562 HBO-3M A R6.11 V83I / M109L
pAG564 HBO-3M A R7.1 M109L
pAG565 HBO-3M A R7.2 Q41L / M109L
pAG566 HBO-3M A R7.10 V83I / T103I
pAG567 HBO-3M A R7.12 K17R / A30V / E74K / E93V
pAG568 HBO-3M A R7.14 T50S / E93Q / M95I
pAG562 HBO-3,5DM A R6.6 V83I / M109L
pAG563 HBO-3,5DM A R6.18 H3Q / D48A / T103I / M109L
pAG558 HBO-3,5DM A R7.1 N13I / D48E / D71N / V83G / M95V
pAG559 HBO-3,5DM A R7.2 Q41L / M95I / M109L
pAG560 HBO-3,5DM A R7.5 K60R / V83I / K104R / S117R
pAG561 HBO-3,5DM A R7.7 D71N / M109L / S117R
pAG795 HBO-3,5DM A-rational design R7.7-1 D71N / M109L / S117R + Q41L
pAG796 HBO-3,5DM A-rational design R7.7-2 D71N / M109L / S117R + Q41L + D48E
pAG644 HBO-3,5DM A-rational design R7.7-3 D71N / M109L / S117R + Q41L + D65V
pAG645 HBO-3,5DM A-rational design R7.7-4 D71N / M109L / S117R + Q41L + V83I
pAG797 HBO-3,5DM A-rational design R7.7-5 D71N / M109L / S117R + Q41L + M95I
pAG771 HBP-3,5DM A R5.3 Q41H / V83L / M95I
pAG772 HBP-3,5DM A R5.7 Q41K / S72T / V83A / M95I
pAG773 HBP-3,5DM A R5.12 K17N / Q41K / M95T
pAG556 HBP-3,5DM A R5.13 Q41L / S117I
pAG557 HBP-3,5DM A R5.21 V83E / S117R
pAG774 HBP-3,5DM A R6.2 E93G / M95V / M109L
pAG775 HBP-3,5DM A R7.1 D20N / E81Q / V83I / M109L
pAG776 HBP-3,5DM A R7.6 G25R / A84S
pAG777 HBP-3,5DM A R7.10 Q32K / Y76N
pAG553 HBP-3,5DM A-rational design R5.7-1 Q41K / S72T / V83A / M95I + M109L
pAG778 HBP-3,5DM A-rational design R5.7-2 Q41K / S72T / V83A / M95I + S117I
pAG779 HBP-3,5DM A-rational design R5.7-3 Q41K / S72T / V83A / M95I + M109L + S117I
pAG554 HBP-3,5DM A-rational design R5.12-1 K17N / Q41K / M95T + S72T + V83A
pAG555 HBP-3,5DM A-rational design R7.6-1 G25R / A84S + Q41K + M95T
pAG780 HBP-3,5DM B R5.2 K17N / Q41K / S72T / M95T
pAG639 HBP-3,5DM B R5.11 K17N / G25E / S72T / V83A / M95T
pAG781 HBP-3,5DM B R5.19 Q41K / K80M / A84S / N89D / M95T / M109L
pAG640 HBP-3,5DM B R6.1 K17N / G21E / G25R / A30V / Q41L / S72T / V83A / M95T / S117R
pAG782 HBP-3,5DM B R6.10 Q41K / S72T / V83A / S117R
pAG783 HBP-3,5DM B R6.15 N13S / Q41K / S72T / V83A / M95T
K17I / G25R / Q41K / A44T / K60R / V83A / N89D / E93K / M95T /M109L
pAG642 HBP-3,5DM B R6.19
/ S117R
pAG784 HBP-3,5DM B R7.3 K17N / Q41K / D65E / S72T / V83A / K106M / M109L
K17N / G21E / G25R / A30V / Q41L / S72T / V83A / M95T / S117R +
pAG641 HBP-3,5DM B-rational design R6.1-1
M109L
pAG785 HBP-3,5DM B-rational design R6.1-2 K17N / G21E / G25E / A30V / Q41L / S72T / V83A / M95T / S117R
pAG786 HBP-3,5DM B-rational design R6.1-3 K17N / G21E / G25R / A30V / Q4IK / S72T / V83A / M95T / S117R
pAG791 HBT-3,5DM B R6.11 A16D / Q41K / Y76H / V83A / M95T / M109L / S117R
pAG643 HBT-3,5DM B R6.13 G25R/ Q41K / S72T / A84S / M95A / M109L / S117R
pAG792 HBT-3,5DM B R6.17 A27V / Q41K/ V83A / M95T / M109L / S117I
pAG640 HBT-3,5DM B R6.24 K17N / G21E / G25R / A30V / Q41L / S72T / V83A / M95T / S117R
pAG793 HBT-3,5DM B R7.2 Q41K / K80R / V83A / M95T / M109L
pAG794 HBT-3,5DM B R7.4 Q41K / M95T / S117I
pAG787 HBP-3,5DOM B R6.7 G7D / Q41K / S72T / A84S / M95I / M109L / S117R / R124L
pAG788 HBP-3,5DOM B R7.3 K17N / Q41K / A84S / M95T / M109L / S117R
pAG789 HBP-3,5DOM B R7.7 Q41K/ D48G / S72T / V83A / M95T / M109L / S117R
pAG790 HBP-3,5DOM B R7.22 G21R / G25R / Q41K / Q60E / M95T / M109L / A112G
106
Table III.S12. Sequences of FAST and pFAST

FAST (125 amino acids, MW = 13,706 Da)
MEHVAFGSEDIENTLAKMDDGQLDGLAFGAIQLDGDGNILQYNAAEGDITGRDPKQVIGKNFFKDVAPGTDSPEFYGKFKEGVASGNLNT
MFEWMIPTSRGPTKVKVHMKKALSGDSYWVFVKRV
DNA sequence coding for FAST (375 bp)
atggagcatgttgcctttggcagtgaggacatcgagaacactctggccaaaatggacgacggacaactggatgggttggcctttggcgcaattcagctcgatggtgacgggaatatcctgcagtac
aatgctgctgaaggagacatcacaggcagagatcccaaacaggtgattgggaagaacttcttcaaggatgttgcacctggaacggattctcccgagttttacggcaaattcaaggaaggcgtag
cgtcagggaatctgaacaccatgttcgaatggatgataccgacaagcaggggaccaaccaaggtcaaggtgcacatgaagaaagccctttccggtgacagctattgggtctttgtgaaacgggt
g
pFAST (125 amino acids, MW = 13,884 Da)
MEHVAFGSEDIENTLANMDDEQLDRLAFGVIQLDGDGNILLYNAAEGDITGRDPKQVIGKNFFKDVAPGTDTPEFYGKFKEGAASGNLNTM
FEWTIPTSRGPTKVKVHLKKALSGDRYWVFVKRV
DNA sequence coding for pFAST (375 bp)
atggagcatgttgcctttggcagtgaggacatcgagaacactctggccaatatggacgacgaacaactggataggttggcctttggcgtaattcagctcgatggtgacgggaatatcctgctgtaca
atgctgctgaaggggacatcactggcagagatcccaaacaggtgattgggaagaacttcttcaaggatgttgcacctggaacggatactcccgagttttacggcaaattcaaggaaggcgcagc
gtcagggaatctgaacaccatgttcgaatggacgataccgacaagcaggggaccaaccaaggtcaaggtgcacttgaagaaagccctttccggtgacagatattgggtctttgtgaaacgggtg
107
Table III.S13. Imaging settings used in this study

Figure Panel Fluorescent reporter Excitation Emission Fluorescence Comments
settings (nm) settings (nm) microscopy
Figure Panel e oFAST:HBO-3,5DM 405 450 - 550
III.2 Panel f tFAST:HBT-3,5DM 405 450 - 550 Confocal
Panel g pFAST:HBP-3,5DM 488 500 - 600
Figure
Panel a pFAST:HBO-3M 405 420 - 550
III.4
pFAST:HBO-3,5DM 405 450 - 550
pFAST:HBT-3,5DM 405 450 - 600
pFAST:HBP-3M 405 450 - 600
pFAST:HBT-3,5DOM 488 500 - 600
pFAST:HBP-3,5DM 488 493 - 600
pFAST:HBP-3,5DOM 488 495 - 680
Confocal
pFAST:HMBR 488 495 - 600
pFAST:HBRAA-3E 488 490 - 700
pFAST:HBRAA-3M 488 490 - 700
pFAST:HBRAA-3,5DM 514 550 - 700
pFAST:HBR-3,5DM 488 500 - 630
pFAST:HBR-3,5DOM 514 550 - 700
pFAST:HBIR-3,5DOM 561 570 - 700
Panel b pFAST:HBP3,5DM 488 493 - 600
Confocal 81 optical sections
Panel c pFAST:HBR-3,5DOM 520 550 - 650
( z step 164 nm)
Maximum intensity projection of
Panel d pFAST:HBR-3,5DOM 520 550 - 650 Confocal 5 optical sections
(for lyn11-pfast)
Figure
Panel b pFAST:HBR-3,5DOM Spinning-disk
III.5 561 580 - 654
confocal
mKO
Panel c pFAST:HBR-3,5DM
510/25 540/30 Widefield LED light source
EYFP
Panel d pFAST:HBP-3,5DOM
EYFP
Panel e pFAST:HMBR
EGFP
Panel f pFAST:HBP-3,5DOM
EGFP
Panel g pFAST:HBP-3M
mCerulean
Panel h pFAST:HBR-3,5DOM 561 580 - 654
iRFP670 642 665 - 705
Panel i pFAST:HBP-3,5DOM 488 500 - 550
Spinning-disk
mKO 561 580 - 654
confocal
pFAST:HMBR 488 500 - 550
Panel j mKO 561 580 - 654
iRFP670 642 665 - 705
Figure 3D STED 775 nm pulsed depletion laser ;
III.6 motorized collar 93× glycerol NA
Panel
pFAST:HBR-3,5DOM 520 550 - 650 1.3 objective; optimized pixel
a-l
size(from 25 to 45 nm) and an
average line acquisition of 16
Figure Panel
pFAST:HMBR 488 495 - 600
III.S8 a,d
Panel b pFAST:HBR-3,5DOM 514 550 - 700
Panel c pFAST:HBR-3,5DM 488 500 - 630
Panel e pFAST:HBP-3,5DOM 488 495 - 680 Confocal
Panel f pFAST:HBP-3,5DM 488 493 - 600
Panel g pFAST:HBP-3M 458 470 - 580
Panel h pFAST:HBT-3,5DOM 488 495 - 600
Panel i pFAST:HBT-3,5DM 458 495 - 600
Figure
Panel a pFAST:HBRAA-3E 488 490 - 700
III.S9
pFAST:HMBR 488 490 - 700
Panel b pFAST:HBRAA-3M 488 490 - 700 Confocal
pFAST:HMBR 488 490 - 700
Panel c pFAST:HBRAA-3,5DM 514 530 - 700
pFAST:HMBR 514 530 - 700
108
Figure
Panel a pFAST:HBR-3,5DOM
III.S10
iFAST:HBR-3,5DOM 488 550 - 700 Confocal
FAST:HBR-3,5DOM
EGFP
Panel b pFAST:HBR-3,5DM
iFAST:HBR-3,5DM
488 500 - 650 Confocal
FAST:HBR-3,5DM
EGFP
Panel c pFAST:HMBR
iFAST:HMBR
488 495 - 600 Confocal
FAST:HMBR
EGFP
Panel
pFAST:HBP-3,5DOM
d,f,g
488 495 - 600 Confocal
pFAST:HBP-3,5DM
EGFP
Panel
pFAST:HBT-3,5DOM
e,h,i
458 470 - 600 Confocal
pFAST:HBP-3M
mTurquoise2
Figure Confocal
pFAST:HBP-3,5DM 488 493 - 600
III.S11
Figure
Panel a Halotag:TMR 561 568 - 700
III.S12
Panel b mVenus 514 520 - 600
Confocal
Panel c pFAST:HBP-3,5DM 488 493 - 600
Figure
pFAST:HBR-3,5DOM 520 550 - 680
III.S13
Figure Spinning-disk
Panel a pFAST:HBR-3,5DOM 561 580 - 654
III.S15 confocal
Panel b pFAST:HBR-3,5DM 510/25 540/30 Widefield LED light source
Panel c pFAST:HBP-3,5DOM 470/24 540/30 Widefield LED light source
Panel d pFAST:HMBR 470/24 525/50 Widefield LED light source
Panel e pFAST:HBP-3,5DM 470/24 525/50 Widefield LED light source
Figure
Panel a pFAST:HBR-3,5DM 488 500 - 630
III.S17
Panel b pFAST:HBP-3,5DOM 488 495 - 680 Confocal
Panel c pFAST:HMBR 488 495 - 600
Panel d pFAST:HBP-3,5DM 488 495 - 680
Figure Spinning-disk
Panel a pFAST:HBP-3,5DOM 488 500 - 550
III.S18 confocal
Panel b pFAST:HMBR 488 500 - 550
Figure Panel
pFAST:HMBR 488 495 - 550
III.S19 a,d
Panel b pFAST:HBR-3,5DOM 561 570 - 750 Equipped with a spectral detector
Confocal GaAsp of 32 channels
Panel c pFAST:HBR-3,5DM 488 525 - 620
(Airyscan module)
Panel e pFAST:HBP-3,5DOM 488 525 - 620
Panel f pFAST:HBP-3,5DM 488 495 - 650
Movie
pFAST:HBR-3,5DOM
III.S1 561 580 - 654 Spinning-disk
FAST:HBR-3,5DOM confocal
EGFP 488 500 - 550
Movie
pFAST:HBR-3,5DOM 561 580 - 654 Spinning-disk
III.S2
confocal
iRFP670 642 665 - 705
Movie
pFAST:HBP-3,5DOM 488 500 - 550 Spinning-disk
III.S3
confocal
mKO 561 580 - 654
Movie
pFAST:HMBR 488 500 - 550
III.S4 Spinning-disk
mKO 561 580 - 654 confocal
iRFP670 642 665 - 705
Movie
pFAST:HBR-3,5DM 488 500 - 630
III.S5
pFAST:HBP-3,5DOM 488 495 - 680 Confocal
pFAST:HMBR 488 495 - 600
pFAST:HBP-3,5DM 488 495 - 680
Movie Spinning-disk
pFAST:HMBR 488 500 - 550
III.S6 confocal
109
Movie Equipped with a spectral detector

III.S7 GaAsp of 32 channels
pFAST:HBR-3,5DOM 561 570 - 750 Confocal (Airyscan module)
Lyn11 : 1 image / 932 ms
MAP4 : 1 image /4.99 s
III.2.6. Materials and methods
Organic synthesis
General. Commercially available reagents were used as obtained. 1H and 13

C NMR spectra were
recorded at 300 K on a Bruker AM 300 spectrometer; chemical shifts are reported in ppm with
protonated solvent as internal reference 1H, CHCl3 in CDCl3 7.26 ppm, CHD2SOCD3 in CD3SOCD3
13 13 13
2.49 ppm, CHD2COCD3 in CD3COCD3 2.05 ppm; C, CDCl3 in CDCl3 77.0 ppm, CD3SOCD3 in
13
CD3SOCD3 39.7 ppm, CD3COCD3 in CD3COCD3 29.9 ppm; coupling constants J are given in Hz.
Mass spectra (chemical ionization and electronic impact with NH3 or CH4) were performed by the
Service de Spectrométrie de Masse de Chimie ParisTech and mass spectra high resolution were
performed by the Service de Spectrométrie de Masse de l’Institut de Chimie Organique et Analytique
(Orléans). The preparation of HMBR (4-hydroxy-3-methylbenzylidene rhodanine), HBR-3,5DM (4-
hydroxy-3,5-dimethylbenzylidene rhodanine), HBR-3,5DOM (4-hydroxy-3,5-dimethoxybenzylidene
rhodanine), HBRAA-3M ((5-(4-hydroxy-3-methylbenzylidene)-4-oxo-2-thioxothiazolidin-3-yl)acetic
acid) and HBRAA-3E (5-(4-hydroxy-3-ethylbenzylidene)-4-oxo-2-thioxothiazolidin-3-yl)acetic acid)
were previously described13,21,20. The starting compound oxazolidinedione was obtained according to
previously described methods45.
HBP-3M ((4-hydroxy-3-methylbenzylidene)-2-iminothiazolidin-4-one). Solid 4-hydroxy-3-

methylbenzaldehyde (130 mg, 1.0 mmol) and pseudothiohydantoin (69 mg, 0.6 mmol) were mixed
and the mixture was brought to 170°C. It first fused and then solidified. After cooling to 50-60°C,
ethanol was added and the mixture was heated at reflux. The solid partially disintegrated and formed
a suspension. It was filtered and the remaining powder was dissolved in 1 M sodium hydroxide. The
resulting solution was stirred for 5-10 min and was then slowly acidified with 1 M hydrochloric acid.
After cooling to 0°C, the precipitate was filtered, washed with cold water and dried over P2O5. HBP-
3M was obtained as an orange powder (42 %, 58 mg). 1H NMR (300 MHz, DMSO-d6, d): 10.03 (s,
1H), 9.28 (s,1H), 9.00 (s,1H), 7.74 (s, 1H), 7.30 (d, 1H, J = 2.0 Hz), 7.24 (dd, 1H, J = 8.3 Hz, 2.0 Hz),
6.92 (d, 1H, J = 8.3 Hz), 2.16 (s, 3H); 13C (75 MHz, DMSO-d6, d): 180.7, 175.4, 157.2, 132.2, 129.6,
128.9, 125.0, 124.8, 124.7, 115.3, 16.0; HRMS (ESI): m/z 235.0537 (calcd mass for C11H11N2O2S
[M+H]– = 235.0536.).
110
HBP-3,5DM ((4-hydroxy-3,5-dimethylbenzylidene)-2-iminothiazolidin-4-one). As HBP-3M using 4-

hydroxy-3,5-dimethylbenzaldehyde (207 mg, 1.4 mmol) and pseudothiohydantoin (100 mg, 0.86
mmol). HBP-3,5DM was obtained as a dark orange powder (83 %, 177 mg). 1H NMR (300 MHz,
DMSO-d6, d): 9.27(s,1H), 8.97(s,1H) ,8.93 (s, 1H),7.42 (s, 1H), 7.16(s, 2H), 2.20 (s, 6H); 13
C NMR
(75 MHz, DMSO-d6, d): 180.6, 175.4, 155.1, 130.1 (2C), 129.7, 125.2, 124.9 (2C), 16.7; HRMS (ESI):
m/z 249.0694 (calcd mass for C12H13N2O2S [M+H]– = 249.0692.).
HBP-3,5DOM ((4-hydroxy-3,5-dimethoxybenzylidene)-2-iminothiazolidin-4-one). As HBP-3M using

4-hydroxy-3,5-dimethoxybenzaldehyde (251 mg, 1.4 mmol) and pseudothiohydantoin (100 mg, 0.86
mmol). HBP-3,5DOM was obtained as an orange powder (61 %, 146mg). 1H NMR (300 MHz, DMSO-
d6, d): 9.31 (s,1H), 9.10 (s,1H), 9.03 (s,1H), 7.53 (s,1H), 6.88 (s,2H), 3.82 (s,6H); HRMS (ESI): m/z
281.0590 (calcd mass for C12H13N2O4S [M+H]- = 281.0591 (in accordance with previously described
analysis46).
HBT-3M ((4-hydroxy-3-methylbenzylidene)thiazolidine-2,4-dione). Solid 4-hydroxy-3-

methylbenzaldehyde (187 mg, 1.4 mmol) and 2,4-thiazolidinedione (100 mg, 0.85 mmol) were mixed
and the mixture was brought to 170°C. It first fused and then solidified. After cooling to 50-60°C,
ethanol was added to the solid and the suspension was stirred at reflux until complete dissolution.
Then an excess of water was slowly added. The resulting precipitate was filtered, washed with water,
and dried over P2O5. HBT-3M was obtained as a green powder (55 %, 110 mg). 1H NMR (300 MHz,
DMSO-d6, d): 12.43 (s, 1H), 10.23 (s, 1H), 7.65 (s, 1H), 7.33(d, 1H, J = 2.0 Hz), 7.29 (dd, J = 8.3, 2.0
Hz, 1H), 6.92 (d, J = 8.3 Hz, 1H), 2.16 (s, 3H); 13C NMR (75 MHz, DMSO-d6, d): 168.2, 167.6, 158.2,
133.2, 132.5, 129.8, 125.2, 123.8, 118.7, 115.4, 15.9; HRMS (ESI): m/z 236.0376 (calcd mass for
C11H10NO3S [M+H]–: 236.0376).
HBT-3,5DM ((4-hydroxy-3,5-dimethylbenzylidene)thiazolidine-2,4-dione). As HBT-3M using 4-

hydroxy-3,5-dimethylbenzaldehyde (205 mg, 1.4 mmol) and 2,4-thiazolidinedione (100 mg, 0.85
mmol). HBT-3,5DM was obtained as a yellow powder (75 %, 154 mg). 1H NMR (300 MHz, DMSO-d6,
d): 12.43 (s, 1H), 9.14 (s, 1H), 7.61 (s, 1H), 7.18 (s, 2H), 2.20 (s, 6H); HRMS (ESI): m/z 250.0533
(calcd mass for C12H12NO3S [M+H]–: 250.0532.) (in accordance with previously described
analysis47,48).
HBT-3,5DOM ((4-hydroxy-3,5-dimethoxybenzylidene)thiazolidine-2,4-one). As HBT-3M using 4-

hydroxy-3,5-dimethoxybenzaldehyde (258 mg, 1.4 mmol) and 2,4-thiazolidinedione (100 mg, 0.85
mmol). HBT-3,5DOM was obtained as a green powder (61 %, 146mg). 1H NMR (300 MHz, DMSO-
d6, d): 12.48 (s, 1H), 9.33 (s, 1H), 7.71 (s, 1H), 6.89 (s, 2H), 3.81 (s, 6H); 13C (75 MHz, DMSO-d6, d):
111
168.5, 167.9, 148.7, 139.1, 133.3, 123.7, 120.1, 108.4, 56.5; HRMS (ESI): m/z 282.0431 (calcd mass
for C12H12NO5S [M+H]–: 282.0431) (in accordance with previously described analysis47,49).
HBO-3M ((4-hydroxy-3-methylbenzylidene)oxazolidine-2,4-dione). A mixture of 4-hydroxy-3-

methylbenzaldehyde (100 mg, 0.74 mmol), 2,4-oxazolidinedione (148 mg, 1.5 mmol), and pyrrolidine
(60 µL, 0.74 mmol) in ethanol (2 mL) was stirred at reflux for 18 h. After cooling, the mixture was
diluted with water. The precipitate was collected by filtration, washed with cold water and dried over
P2O5. HBO-3M was obtained as an orange powder (48 %, 78mg). 1H NMR (300 MHz, DMSO-d6, d):
12.16 (s, 1H), 10.04 (s, 1H), 7.51 (d, 1H, J= 3 Hz), 7.47 (dd, 1H, J= 3, 9 Hz), 6.86 (d, 1H, J= 9 Hz),
6.58 (s, 1H), 2.14 (s, 3H); 13C NMR (75 MHz, DMSO-d6, d): 165.9, 157.2, 154.3, 137.8, 133.1, 129.9,
124.6, 122.4, 115.1, 109.7, 15.9; HRMS (ESI): m/z 220.060512 (calcd mass for C11H10NO4 [M+H] – =
220.060434).
HBO-3,5DM ((4-hydroxy-3,5-dimethylbenzylidene)oxazolidine-2,4-dione). As HBO-3M using 4-

hydroxy-3,5-dimethylbenzaldehyde (100 mg, 0.66 mmol) , 2,4-oxazolidinedione (134 mg, 1.3 mmol),
pyrrolidine (55 µL, 0.66 mmol) in ethanol (2 mL). HBO-3,5DM was obtained as an orange powder
(43mg, 28 %). 1H NMR (300 MHz, DMSO-d6, d): 12.10 (s, 1H), 8.95 (s,1H), 7.37 (s, 2H), 6.53 (s, 1H),
2.19 (s, 6H); HRMS (ESI): m/z 234.076560 (calcd mass for C12H12NO4 [M+H]– = 234.076084). (in
accordance with previously described analysis50)
HBIR-3M ((4-hydroxy-3-methylbenzylidene)-4-thioxothiazolidin-2-one). As HBP-3M using 4-hydroxy-

3-methylbenzaldehyde (164 mg, 1.2 mmol) and isorhodanine (100 mg, 0.75 mmol). HBIR-3M was
obtained as an orange powder (49%, 93 mg). 1H NMR (300 MHz, DMSO-d6, d): 10.53 (s, 1H), 8.01
(s, 1H), 7.43 (d, 1H, J = 2.0 Hz), 7.38 (dd, 1H, J = 8.3 Hz, 2.0 Hz), 6.96 (d, 1H, J = 8.3Hz), 2.16 (s,
3H); 13C NMR (75 MHz, DMSO-d6, d): 194.9, 170.8, 159.2, 137.3, 134.1, 130.8, 125.7, 125.6, 124.4,
115.7, 15.9; HRMS (ESI): m/z 252.014671 (calcd mass for C11H10NO2S2 [M+H]– = 252.014747).
HBIR-3,5DM ((4-hydroxy-3,5-dimethylbenzylidene)-4-thioxothiazolidin-2-one). As HBP-3M using 4-

hydroxy-3,5-dimethylbenzaldehyde (180 mg, 1.2 mmol) and isorhodanine (100 mg, 0.75 mmol).
HBIR-3,5DM was obtained as a brown powder (69%, 137 mg). 1H NMR (300 MHz, DMSO-d6, d):
13.66 (s, 1H), 9.39 (s, 1H), 7.97 (s, 1H), 7.28 (s, 2H), 2.21 (s, 6H); 13C NMR (75 MHz, DMSO-d6, d):
194.9, 170.8, 157.1, 137.2, 131.8, 125.7, 125.5, 124.5, 16.6; HRMS (ESI): m/z 266.030241 (calcd
mass for C12H12NO2S2 [M+H]– = 266.030397).
HBIR-3,5DOM ((4-hydroxy-3,5-dimethoxylbenzylidene)-4-thioxothiazolidin-2-one). As HBP-3M using

4-hydroxy-3,5-dimethoxybenzaldehyde (220 mg, 1.2 mmol) and isorhodanine (100 mg, 0.75 mmol).
112
HBIR-3,5DOM was obtained as a brown powder (48%, 110 mg). 1H NMR (300 MHz, DMSO-d6, d):
13.72 (s, 1H), 9.60 (s, 1H), 8.05 (s, 1H), 6.97 (s, 2H), 3.83 (s, 6H); 13C NMR (75 MHz, DMSO-d6, d):
194.9, 170.8, 148.3, 139.3, 137.5, 126.5, 123.8, 108.7, 56.1; HRMS (ESI): m/z 298.020265 (calcd
mass for C12H12NO4S2 [M+H]– = 298.020226) (in accordance with previously described analysis)51.
HBRAA-3,5DM 5-(4-hydroxy-3,5-dimethylbenzylidene)-4-oxo-2-thioxothiazolidin-3-yl) acetic acid. A

solution of rhodanine-3-acetic acid (191 mg, 1.0 mmol) and 4-hydroxy-3,5-dimethylbenzaldehyde
(150 mg, 1.0 mmol) in water (70 mL) was stirred at 90°C for 7 days. After cooling to 4°C and standing
overnight, the precipitate was filtered through a glass filter and the crude solid was washed with water,
ethanol and dried over P2O5, to give the desired product as an orange powder (160 mg, 50 %). 1H
NMR (300 MHz, CD3SOCD3, d): 9.40 (s, 1H), 7.70 (s, 1H), 7.27 (s, 2H), 4.72 (s, 2H), 2.23 (s, 6H); 13C
NMR (75 MHz, CD3SOCD3, d): 193.6, 167.8, 166.9, 157.7, 135.3, 132.4(2C), 126.0(2C), 124.3, 117.4,
45.4, 17.1(2C); MS (ESI): m/z 322.2 (calcd mass for C14H12NO4S2 [M-H]– = 322.0).
Biology
General. Synthetic oligonucleotides used for cloning were purchased from Integrated DNA
technology. PCR reactions were performed with Q5 polymerase (New England Biolabs) in the buffer
provided. PCR products were purified using QIAquick PCR purification kit (Qiagen). The products of
restriction enzyme digests were extracted and purified by preparative gel electrophoresis followed by
QIAquick gel extraction kit (Qiagen). Restriction endonucleases, DNAse I, T4 ligase, Fusion
polymerase, Tag ligase and Tag exonuclease were purchased from New England Biolabs and used
with accompanying buffers and according to manufacturer protocols. Isothermal assemblies (Gibson
assembly) were performed using homemade mix prepared according to previously described
protocols52. Small-scale isolation of plasmid DNA was done using QIAprep miniprep kit (Qiagen) from
2 mL overnight bacterial culture supplemented with appropriate antibiotics. Large-scale isolation of
plasmid DNA was done using the QIAprep maxiprep kit (Qiagen) from 150 mL of overnight bacterial
culture supplemented with appropriate antibiotics. All plasmid sequences were confirmed by Sanger
sequencing with appropriate sequencing primers (GATC Biotech). All the plasmids used in this study
are listed in Tables III.S10 and III.S11. The protein and DNA sequences of FAST and pFAST are
given in Table III.S12.
Yeast display
Library construction. The first yeast library (called library A in this study) of FAST constructed by error-
prone PCR using Genemorph II kit (Agilent) was previously described23. The second library (called
113
library B in this study) was constructed by DNA shuffling according to different DNA shuffling library
protocols53–55. Five gene variants (with more than 70% of sequences homology) previously screened,
selected and rationally designed from the library A were first amplified by PCR. They were randomly
digested by DNAse I (20 U/mL). The digested products were purified and mixed together. Two steps
of PCR amplification (one without primers and the second one with two primers to amplify the full-
length genes) were then performed enabling random recombination and amplification of the full-length
shuffled product. In addition, a mutation rate of 4 nt/gene and a deletion rate of 0.25 nt/gene were
obtained (as shown by Sanger sequencing of 17 individual clones). The shuffled product was cloned
into pCTCON2 using NheI and BamHI restriction sites. Large scale transformation into DH10B was
performed, yielding 5.5 ´ 107 transformants. The DNA was maxiprepped and transformed into yeast
strain EBY100 using a large-scale, high-efficiency protocol56 to yield 6.7 ´ 107 transformants.
Selection. The yeast library (about 1-5 x 109 cells) was grown over night at 30°C in 1L SD medium
(20 g/L dextrose, 6.7 g/L yeast nitrogen base, 1.92 g/L yeast synthetic dropout without tryptophan,
7.44 g/L NaH2PO4, 10.2 g/L Na2HPO4-7H2O, 1% penicillin-streptomycin 10,000 U/mL). The following
morning yeast culture was diluted to OD600nm 1 in 1L of SD and grown at 30°C until OD600nm 2-5. Next,
5 ´ 106 cells were pelleted and resuspended in 1 L of SG medium (20 g/L galactose, 2 g/L dextrose
6.7 g/L yeast nitrogen base, 1.92 g/L yeast synthetic dropout without tryptophan, 7.44 g/L NaH2PO4,
10.2 g/L Na2HPO4-7H2O, 1% penicillin-streptomycin 10,000 U/mL) to grow for 36 hours at 23°C. 5 ´
108 induced yeast cells were collected (2500 ´ g - 2min) , washed with 10 mL PBS (0.05 M phosphate
buffer, 0.150 M NaCl, pH 7.4) + BSA (bovine serum albumin 1 g/L) and incubated for 30 min at room
temperature with 1:250 dilution of primary antibody chicken anti-c-myc IgY (Life technologies) in 200
μL of PBS. Cells were then centrifuged, washed with PBS-BSA and incubated for 20min on ice with
1:100 dilution of secondary goat-anti chicken antibody coupled to AlexaFluor 647 in 200 μL of PBS.
After centrifugation and washing with 10 mL PBS + BSA, the cells were resuspended in 5 mL of PBS
supplemented with appropriate fluorogen concentration. The cells were sorted on a MoFloTM Astrios
Cell sorter (Beckman Coulter) equipped with 405 nm, 488 nm and 640 nm lasers. The fluorescence
of HBO-3,5DM and HBO-3M binders were detected using the following parameters: Ex 405 nm, Em
458 ± 30 nm. For the selection using library A, the fluorescence of HBT-3,5DM and HBT-3M binders
were detected using the following parameters: Ex 405 nm, Em 488 ± 5 nm. For the selection using
library B, the fluorescence of HBT-3,5DM binders was detected using the following parameters: Ex
405 nm, Em 513 ± 13 nm. The fluorescence of HBP-3,5DM and HBP-3M binders were detected using
the following parameters: Ex 488nm, Em 526 ± 26 nm. Finally, the fluorescence of HBP-3,5DOM
binders was detected using the following parameters: Ex 488 nm, Em 546 ± 10 nm. Sorted cells were
collected in SD, grown over night at 30°C and plates on SD plates (SD supplemented with 182 g/L D-
sorbitol and 15 g/L agar) for approximately 60 hours at 30°C. The resulting lawn was collected in SD
114
supplemented with 30% glycerol, aliquoted and frozen or directly used for the next round of selection.
In total, 5 to 8 rounds of selection were performed for each fluorogen. After each selection, 24 clones
per final rounds were isolated, screened by flow cytometry and their plasmidic DNA was isolated using
miniprep kit (Qiagen), transformed in DH10B and re-isolated for sequencing.
Cloning. Variants obtained by rational design were cloned in pCTCON2 plasmids driving EBY100
yeast surface expression using NheI and BamHI restriction sites as for library construction (vide
supra). All the plasmids used in this study are listed in Table III.S10.
Flow cytometry analysis. Flow cytometry was performed on a Gallios analyzer (Beckman Coulter)
equipped with 405 nm, 488 nm and 638 nm lasers and ten filters and channels. To prepare samples
for flow cytometry, small scale cultures were grown as for library expression (vide supra). Briefly, 5
mL of SD were inoculated with a single colony and grown overnight at 30°C. The following day, the
cultures were diluted in 5 mL of SD to a final OD600nm 1 and grown until OD600nm 2-5. These cultures
were used to inoculate 5 mL of either SD (non-induced cultures) or SG (induced cultures) to an
OD600nm of 0.5 and the cultures were grown for 36 h at 23°C. The cultures were collected (2500 ´ g -
2min) to reach a final concentration of 1 ´ 108 cells/mL. For analysis, the monoclonal yeast cells were
prepared and incubated with the set of antibodies as for library expression (vide supra). The cultures
were finally resuspended in PBS supplemented with appropriate fluorogens concentrations. Data
were analyzed using Kaluza Analysis software (Beckman Coulter).
Characterization of variants
Cloning. Plasmids driving E. coli expression of the variants with an N-terminal His-tag under the
control of a T7 promoter were constructed by replacing the sequence coding for FAST in the plasmid
pAG8713 using isothermal Gibson assembly. Site-directed mutagenesis was performed by isothermal
Gibson assembly using primers with the mutations of interest. All the characterized variants are listed
in Table III.S11.
Expression. The plasmids were transformed in Rosetta (DE3) pLys E. coli (New England Biolabs).
Cells were grown at 37°C in lysogen Broth (LB) medium supplemented with 50 μg/ml kanamycin and
34 μg/mL chloramphenicol to OD600nm 0.6. Expression was induced overnight at 16°C by adding
isopropyl-β-D-1-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. Cells were harvested
by centrifugation (4,300 ´ g for 20min at 4°C) and frozen.
115
Purification. The cell pellet was resuspended in lysis buffer (PBS supplemented with 2.5 mM MgCl2,
1mM of protease inhibitor PhenylMethaneSulfonyl Fluoride PMSF, 0.025 mg/mL of DNAse, pH 7.4)
and sonicated (5 min, 20 % of amplitude) on ice. The lysate was incubated for 2-3 hours on ice to
allow DNA digestion by DNAse. Cellular fragments were removed by centrifugation (9,000 ´ g for 1 h
at 4°C). The supernatant was incubated overnight at 4°C by gentle agitation with pre-washed Ni-NTA
agarose beads in PBS buffer complemented with 20 mM of imidazole. Beads were washed with 10
volumes of PBS complemented with 20 mM of imidazole and with 5 volumes of PBS complemented
with 40mM of imidazole. His-tagged proteins were eluted with 5 volumes of PBS complemented with
0.5 M of imidazole. The buffer was exchanged to PBS (0.05 M phosphate buffer, 0.150 M NaCl) using
PD-10 or PD-MidiTrap G-25 desalting columns (GE Healthcare). Purity of the proteins were evaluated
using SDS-PAGE electrophoresis stained with Coomassie blue.
Physico-chemical measurements. Steady state UV-Vis absorption spectra were recorded using a
Cary 300 UV-Vis spectrometer (Agilent Technologies), equipped with a Versa20 Peltier-based
temperature-controlled cuvette chamber (Quantum Northwest) and fluorescence spectra were
recorded using a LPS 220 spectrofluorometer (PTI, Monmouth Junction, NJ), equipped with a
TLC50TM Legacy/PTI Peltier-based temperature-controlled cuvette chamber (Quantum Northwest).
Fluorescence quantum yield measurements were determined as previously described using either
FAST:HMBR or quinine sulfate as a reference13,21,39,23. Reciprocal dilution with protein solution was
used so as to keep the protein concentration constant at 40 µM while diluting the fluorogen solution.
Absorption coefficients were determined by forward titration of fluorogens into a 40 µM protein
solution. Thermodynamic dissociation constants were determined as previously described13 using a
Spark 10M plate reader (Tecan) and fitting data in Prism 6 to a one-site specific binding model. The
thermodynamic dissociation constants (KD) of the dark HBIR-3M and HBIR-3,5DM chromophores
were determined by determining the apparent dissociation constant of HBP-3,5DM in presence of
various concentrations of dark competitors (See Text III.S2).
Modeling
Homology modeling. The homology models of FAST and pFAST were generated according to models
previously described57,58. Briefly, sequence alignments between FAST and pFAST and the ultra-high
resolution structure of the Halorhodospira halophila Photoactive Yellow Protein (PYP) (Protein Data
Bank (PDB) ID: 6P4I) were generated with Clustal W59. Alignments were manually refined to avoid
gaps in predicted (FAST and pFAST) and known (PYP) secondary structure elements. Three-
dimensional FAST and pFAST models were built from these alignments and from crystallographic
atomic coordinates of PYP using the automated comparative modeling tool MODELER (Sali and
116
Blundell) implemented in Discovery Studio. The best model according to DOPE score (Discrete
Optimized Protein Energy) and potential energy calculated by modeler were solvated (10 Å water box
and 0.145M NaCl) and minimized using Adopted Basis NR algorithm to a final gradient of 0.001. The
resulting structure were submitted to a 10ns NAMD dynamic.
Molecular Docking. Flexible ligand-rigid protein docking was performed using CDOCKER
implemented in Discovery Studio 201960. Random ligand conformations were generated from the
initial ligand structure through high-temperature molecular dynamics. The best poses according to
their ligscore261 were retained and clustered according to their binding mode. The most significant
poses were solvated and minimized using Adopted Basis NR algorithm to a final gradient of 0.001.
Experiments in mammalian cell
General. HeLa cells were cultured in Minimal Essential Media (MEM) supplemented with phenol red,
Glutamax I, 1 mM of sodium pyruvate, 1% (vol/vol) of non-essential amino-acids and 10% (vol/vol)
fetal calf serum (FCS), at 37 °C in a 5% CO2 atmosphere. HEK 293T cells were cultured in Dulbecco's
Modified Eagle Medium (DMEM) supplemented with phenol red and 10% (vol/vol) fetal calf serum at
37 °C in a 5% CO2 atmosphere. For imaging, cells were seeded in µDish IBIDI (Biovalley) coated
with poly-L-lysine. Cells were transiently transfected using Genejuice (Merck) according to the
manufacturer’s protocol for 24 hours prior to imaging. Live cells were washed with DPBS (Dulbecco’s
Phosphate-Buffered Saline), and treated with DMEM media (without serum and phenol red)
containing the fluorogens at the indicated concentration. The cells were imaged directly without
washing. Fixation of cells was performed using formaldehyde solution at 3.7 % for 30 min. Fixed cells
were then washed three times with DPBS and treated with DPBS containing the fluorogens at the
indicated concentration prior to imaging.
Cloning. The plasmids allowing the mammalian expression of pFAST, tFAST and oFAST variants
under the control of a CMV promoter were constructed by replacing the sequence coding for FAST13
or iFAST39 in the previously described plasmids pAG104 (FAST), pAG106 (lyn11-FAST), pAG109
(H2B-FAST), pAG156 (mito-FAST), pAG470 (lifeAct-iFAST) and pAG498 (MAP4-FAST) using
isothermal Gibson assembly. All the plasmids used in this study are listed in Table III.S10.
Cell viability assay. HeLa cells were treated with MEM media containing the appropriate fluorogen
(HBP-3,5DOM, HBP-3,5DM, HBP-3M, HBT-3,5DOM, HBT-3,5DM, HBT-3M, HBO-3,5DM, HBO-3M,
HBIR-3,5DOM, HBIR-3,5DM and HBIR-3M) at the indicated concentrations for the indicated times.
The cell viability was evaluated by fluorescence microscopy using the LIVE/DEAD®
viability/cytotoxicity assay kit (Molecular Probes, Life Technologies) following the manufacturer’s
117
protocol.
Fluorescence microscopy. The confocal micrographs of mammalian cells were acquired on a Zeiss
LSM 710 Laser Scanning Microscope equipped with a Plan Apochromat 63´ /1.4 NA oil immersion
objective and on a Leica TCS SP5 confocal laser scanning microscope equipped with a 63´/ 1.4 NA
oil immersion objective. ZEN and Leica LAS AF softwares were used to collect the data. Zeiss LSM
880 confocal laser scanning microscope equipped with a 63´/1.4 NA oil immersion objective and
equipped with photomultiplier modules for confocal imaging and a dedicated highly sensitive spectral
detector using the Airyscan module was used to acquire improved spatial resolution images of HeLa
cells. Airyscan uses a 32-channel gallium arsenide phosphide photomultiplier tube (GaAsP-PMT)
area detector that collects a pinhole-plane image at every scan position. Each detector element
functions as a single, very small pinhole62. ZEN blue and black softwares were used to collect the
data and processing of Airyscan images were performed on the software. The images were analyzed
with Fiji (Image J). Photobleaching measurements for HBR-3,5DOM, HBR-3,5DM, HMBR, HBP-
3,5DOM, HBP-3,5DM and HBP-3M were acquired using 405, 458 and 488 nm excitation in different
illumination conditions. At 488 nm excitation, EGFP was used as a control. At 458 nm excitation,
mTurquoise2 was used as a control. Samples were acquired continuously for 150 images or 500
images using different time lapse between two images. In all cases the pixel dwell was 2.55 µs. The
images were analyzed with Fiji (Image J).
Super resolution microscopy. STimulated Emission Depletion (STED) images were acquired on
NeurImag facility with a confocal laser scanning microscope LEICA SP8 STED 3DX equipped with a
93´/1.3 NA glycerol immersion objective and with two hybrid detectors (HyDs). The specimens were
imaged with a white-light laser and a pulsed 775 nm depletion laser to acquire nanoscale imaging.
Typically, images of 1024´1024 pixels were acquired with a magnification above 3 resulting in a pixel
size in the range of 25-45 nm. Deconvolution processing were performed on STED images using
CMLE analysis with Huyguens software. Iterative processes (up to 40 cycles) were used with a quality
criteria ranging from 1 to 5 percent.
Flow cytometry analysis. Flow cytometry on HEK 293T cells was performed on a Gallios analyzer
(Beckman Coulter) equipped with 405 nm, 488 nm and 638 nm lasers and ten filters and channels.
To prepare samples, cells were first grown in cell culture flasks, then transiently co-transfected with
pAG104 (cmv-FAST) and pAG753 (cmv-iRFP670) or pAG654 (cmv-pFAST) and pAG753 (cmv-
iRFP670) 24 hours after seeding, using Genejuice (Merck) according to the manufacturer’s protocol
for 24 hours. Simple positive controls were also prepared by transiently transfected cells with either
pAG753 or pAG654 plasmids. After 24 hours, cells were centrifuged in PBS with BSA (1mg/ml) and
118
resuspend in PBS-BSA supplemented with the appropriate amounts of fluorogens. For each
experiments, 20,000 cells positively transfected with iRFP670 (Ex 638 nm / Em 660 ± 10 nm) were
analyzed with the following parameters: Ex 488nm, Em 525 ± 20 nm for cells expressing FAST and
pFAST labeled with HMBR and HBP-3,5DM ; Ex 488 nm, Em 575 ± 15 nm for cells expressing FAST
and pFAST labeled with HBR-3,5DM and HBP-3,5DOM and Ex 488 nm, Em 620 ± 15 nm for cells
expressing FAST and pFAST labeled with HBR-3,5DOM. Data were analyzed using Kaluza Analysis
software (Beckman Coulter).
Experiments in primary hippocampal neuronal cells
All experiments involving rats were performed in accordance with the directive 2010/63/EU of the
European Parliament and of the Council of 22 September 2010 on the protection of animals used for
scientific purposes. Hippocampal neurons from embryonic rats (E18) were prepared as described
previously63. Cells were grown on onto poly-L-lysine-coated 18-mm coverslips (1 mg/ml) at a density
of 25,000 cells /cm2 in Neurobasal-B27 medium previously conditioned by a confluent glial feeder
layer [Neurobasal medium (ThermoFisher 21103049) containing a 2% B27 supplement
(ThermoFisher A3582801), and 500 μM L-glutamine (ThermoFisher 25030024)]. Neurons were
transfected 2 days before imaging using Lipofectamin 2000 (ThermoFisher). After 10 to 21 days in
vitro, neurons were incubated with the fluorogenic chromophores and imaged immediately at 37°C.
Experiments in chicken embryos
JA57 chicken fertilized eggs were provided by EARL Morizeau (8 rue du Moulin, 28190 Dangers,
France) and incubated at 38°C in a Sanyo MIR-253 incubator. Embryos used in this study were
between E2 (HH14) and E3 (HH14 + 24 h). The sex of the embryos was not determined.
Cloning. For expression of H2B-FAST and H2B-pFAST in the chick neural tube, the CMV promoters
in pAG109 and pAG657 and pAG671 were converted to a CAGGS promoter64 by replacing the
NdeI/BglII fragment with a NdeI/BglII CAGGS fragment from pCAGGS-MCS2 (X. Morin, unpublished).
For expression of mito-pFAST, the NdeI/EcoRI fragment of the CMV promoter in pAG671 was
replaced with the NdeI/EcoRI CAGGS fragment from pCAGGS-MCS2. pCAGGS-H2B-mCerulean
was created by removing an SphI fragment from Tol2-CAG::Nucbow (Addgene #158992), and
pCAGGS-H2B-EYFP was obtained by removing an AgeI fragment from PB-CAG::CytBow (Addgene
#158995). Construction details, complete sequences and plasmids are available upon request.
119
Electroporation in the chick neural tube was performed at embryonic day 2 (E2, HH stage 14), by
applying five pulses of 50 ms at 25 V with 100 ms in between, using a square-wave electroporator
(Nepa Gene, CUY21SC) and a pair of 5-mm gold-plated electrodes (BTX Genetrode model 512)
separated by a 4-mm interval. The DNA solution was injected directly into the lumen of the neural
tube via glass capillaries. Bilateral electroporation was achieved by switching the electrodes polarity
and repeating the procedure after 2 hours. DNA constructs were used at 0.5 μg/μl each, except pCX-
mbCherry which was used at a concentration of 0.3µg/µl. En-face culture of the embryonic
neuroepithelium was performed at E3 (24 h after electroporation). After extraction from the egg and
removal of extraembryonic membranes in PBS, embryos were transferred to 37°C F12 medium and
pinned down with dissection needles at the level of the hindbrain and hindlimbs in a 35mm Sylgard
dissection dish. A dissection needle was used to separate the neural tube from the somites from
hindbrain to caudal end on both sides of the embryo, and the roof-plate was then slit with the needle.
The neural tube and notochord were then “peeled off” from the remaining tissues and equilibrated 2
min in 1% low meting point agarose/F12 medium at 38°C. The tissue was then transferred in a drop
of agarose medium to a glass-bottom culture dish (MatTek, P35G-0-14-C) and excess medium was
removed so that the neural tube would flatten with its apical surface facing the bottom of the dish, in
an inverted open book conformation. After 30 s of polymerization on ice, an extra layer of agarose
medium (100µl) was added to cover the whole tissue and hardened on ice for 1 min. 1.9 mL of 37°C
culture medium was added (F12/Penicillin Streptomycin/Sodium pyruvate) and the culture dish was
transferred to the 37°C chamber of a spinning disk confocal microscope. To image pFAST and FAST,
intermediate dilutions (2 to 200µM) of ligands HBR-3,5DOM, HBR-3,5DM, HMBR, HBP-3,5DOM,
HBP-3,5DM, HBP-3M and HBIR-3M were prepared by diluting the original 20mM stocks in F12
medium, and appropriate volumes were added to the dish to reach the desired final concentrations.
Live imaging was performed on an inverted microscope (Nikon Ti Eclipse) equipped with a heating
enclosure (DigitalPixel, UK), a spinning disk head (Yokogawa CSUW1) and Borealis system (Andor)
for confocal imaging, a Spectra-X LED light engine (Lumencor) for widefield fluorescence illumination
and an sCMOS Camera (Orca Flash4LT, Hamamatsu) driven by MicroManager software65. Image
stacks were obtained at 2- or 3-min intervals either with a 10´ objective (CFI Plan APO LBDA, NA
0.45, Nikon; z-step = 4 µm) or a 100´ oil immersion objective (APO VC, NA 1.4, Nikon; z-step = 1
µm).
120
III.3. Prospects: toward smaller version of pFAST
The group of Dr. Mikhail S. Baranov at the Institute of Bioorganic Chemistry of the Russian Academy
of Sciences has described in a bioRxiv preprint early 2021 the structure of FAST using both NMR
spectroscopy and X-ray crystallography66. In this study, they showed that their previously described
N871b fluorogen67 is deprotonated in the cavity of FAST and that FAST forms a stabilizing H-bond
network in the same way that PYP does with it natural chromophore, as predicted by our models with
the chromophores of the HBR, HBIR, HBP, HBT and HBO series (see section III.2.3.3 and Figure
III.S6).
a b
Figure III.7. Structures of (a) N871b fluorogen and (b) FAST predicted by NMR spectroscopy in both free (FAST
apo) and bound states (FAST:N871b) (adapted from Povarova et al. 67 and Mineev et al. 66).
In the frame of this work, they found that protein rearrangement within the N-terminal domain seemed
to occur after fluorogen binding, and thus proposed that fluorogen binding could be possible in the
absence of the N-terminal region66 (Figure III.7). They generated a small version of FAST by
truncating the first 27 amino acids (AA) of the protein. They found that the resulting nanoFAST of 99
AA (98 AA plus the first Met residue) lost the ability to activate the classical chromophores of the HBR
series or the N871b fluorogen. However, they showed that nanoFAST could bind the 4-hydroxy-2,5-
dimethoxybenzylidene rhodanine (HBR-2,5DOM, called HBRDOM2 in their study) forming an orange
fluorescence assembly with a f = 55 % and a KD close to 1µM and remained suitable for cellular
imaging66.
As pFAST forms tighter and brighter fluorescent assemblies with our entire collection of fluorogenic
chromophores thank to the presence of a smaller and well-resolved chromophore binding site, we
wondered if a truncated version of pFAST would allow the generation of better defined chromophore
assemblies. Here, we present very preliminary results showing the potential of pFAST for generating
a shorter protein tag. We successfully generated the truncated proteins by removing the first 27
residues (Figure III.8) and purified correct amounts of both truncated proteins, suggesting that both
FAST and pFAST were soluble and stable in solution even without the N-terminal domain (see
section III.2.6 for materials and methods) (Figure III.9). We obtained however lower amount (about
121
10-fold less) of truncated pFAST (hereafter called nano-pFAST) than nanoFAST, suggesting that
truncated pFAST was less soluble than nanoFAST. This was not fully surprising given that many
mutations found in pFAST allow to form strong hydrophobic interactions between the N-terminal
domain and the core domain and play a major role in compacting and rigidifying the entire three-
dimensional structure (see section III.2.3.3). Removing the N-terminal domains therefore exposed
these hydrophobic residues reducing presumably the protein solubility.
72
Figure III.8. Sequence alignment of FAST, pFAST, nanoFAST and nano-pFAST.
6His-nanoFAST 6His-nano-pFAST 6His-

pFAST
FT W1 W2 E FT W1 W2 E E
70 kDa
25 kDa
15 kDa
10 kDa
Figure III.9. SDS-PAGE electrophoresis gel stained with Coomassie blue of nanoFAST and nano-pFAST
purification (FT, flowthrough; W1 wash 1; W2, wash 2 and E, elution) and compared to the elution of pFAST.
The arrows denote the expected eluted bands according to the protein sizes.
Nevertheless, both proteins could be successfully purified and seemed correctly folded allowing
further in vitro characterization. As only 2,5-disubstituted HBR analogs seemed to bind nanoFAST,
we first decided to characterize nanoFAST and nano-pFAST in combination with HBR-2,5DM (4-
hydroxy-2,5-dimethylbenzylidene rhodanine) (see Figure III.10a for structure), which was previously
described by our lab to form a bright and very tight fluorescence assembly with FAST21. Both
nanoFAST and nano-pFAST successfully activated the fluorescence of HBR-2,5DM. However, nano-
pFAST formed an assembly with HBR-2,5DM 12-fold tighter than nanoFAST (Figure III.10 and Table
III.1). In addition, we observed a well-defined red-shifted absorption spectrum for nano-pFAST:HBR-
2,5DM, suggesting that HBR-2,5DM was stabilized in its deprotonated state within nano-pFAST as in
122
FAST and pFAST. On the other hand, the absorption spectrum of nanoFAST:HBR-2,5DM revealed a
significant amount of protonated HBR-2,5DM (although less than 4 % of unbound chromophore
should have been present according to the binding affinity), suggesting that the chromophore can be
both protonated and deprotonated within nano-FAST (Figure III.10c,d).
b c
- FAST 1.0 70000 - pFAST 1.0
40000
Fluorescence (a.u.)
Fluorescence (a.u.)
60000
+ FAST + pFAST
50000
(M-1.cm-1)
(M-1.cm-1)
30000
40000
20000 0.5 0.5
30000
a
20000
10000
R1 O 10000
0 0.0 0 0.0
400 500 600 700 400 500 600 700
NH
S Wavelength (nm) Wavelength (nm)
HO d 40000 e
S - nanoFAST 1.0 40000 - nanopFAST 1.0
R2
Fluorescence (a.u.)
+ nanopFAST
Fluorescence (a.u.)
30000 + nanoFAST
(M-1.cm-1)
30000
(M-1.cm-1)
HBR-2,5DM
20000 20000 0.5
(R1 = Me, R2 = Me) 0.5
10000 10000
0 0.0 0 0.0
400 500 600 700 400 500 600 700
Figure III.10. Absorption and emission properties. (a) Structure of HBR-2,5DM and (b-e) absorption (dashed
lines) and fluorescence (solid lines) spectra of HBR-2,5DM when free in solution (dark line) and bound (colored
lines) to (b) FAST, (c) pFAST, (d) nanoFAST and (e) nano-pFAST. Spectra were recorded using chromophore
concentration ranging from 4 to 6 µM and proteins at 40 µM in pH 7.4 phosphate buffer saline at 25°C.
Table III.1. Properties of HBR-2,5DM in presence of FAST, pFAST, nanoFAST and nano-pFAST proteins in
PBS pH 7.4.
e Molecular
Mutant labs (nm) Dlabs (nm) lem (nm) f KD (µM)
(M-1cm-1) brightness
FAST (ref 21) 494 88 552 50,000 0.29 14,500 0.008

pFAST 498 94 551 63,000 0.34 21,500 0.010
nanoFAST 492 90 551 33,000 0.34 11,500 1.3
nano-pFAST 495 87 552 43,000 0.25 11,000 0.100
We next tested nanoFAST and nano-pFAST with other chromophores of the HBR derivatives. In
agreement with the report of Baranov et al.66, nanoFAST interacted poorly with the 3-substituted or
3,5-disubstitued HBR analogs with KDs ranging from 4.6 to 25 µM. However, we observed that nano-
pFAST formed tighter assemblies than nanoFAST with KDs ranging from 0.50 to 1.80 µM, leading to
higher fluorescence activation (Figure III.11g). Although nano-pFAST showed affinities that should
ensure almost complete complex formation in the conditions used for recording the spectra, the
123
absorption spectra revealed the presence of protonated chromophores, suggesting that bound
chromophores can be both protonated and deprotonated within nano-pFAST, presumably because
of an altered H-bond network in the binding cavity of the truncated protein (Figure III11a-f).
a HBR-3,5DOM b HBR-3,5DOM
g
30000 30000
- nanoFAST 1.0 - nanopFAST 1.0
KD ( M)
Fluorescence (a.u.)
Fluorescence (a.u.)
+ nanoFAST + nanopFAST
20000 20000
(M-1.cm-1)
(M-1.cm-1)
0.5 0.5 nanoFAST nano-pFAST
10000 10000
HBR-3,5DOM 25 1.80
0 0.0 0 0.0
400 500 600 700 400 500 600 700
Wavelength (nm) Wavelength (nm) HBR-3,5DM 4.6 0.51
HBR-3,5DM HBR-3,5DM
c d
50000 - nanoFAST 1.0 50000 - nanopFAST 1.0
HMBR 9.0 0.57
Fluorescence (a.u.)
Fluorescence (a.u.)
40000 + nanoFAST 40000 + nanopFAST
(M-1.cm-1)
(M-1.cm-1)
30000 30000
0.5 0.5
20000 20000
10000 10000
0 0.0 0 0.0
400 500 600 700 400 500 600 700
HMBR HMBR
e 50000
1.0
f 50000
- nanopFAST 1.0
- nanoFAST
40000 40000
Fluorescence (a.u.)
Fluorescence (a.u.)
+ nanoFAST + nanopFAST
(M-1.cm-1)
(M-1.cm-1)
30000 30000
0.5 0.5
20000 20000
10000 10000
0 0.0 0 0.0
400 500 600 700 400 500 600 700
Figure III.11. Absorption and emission properties. Absorption (dashed lines) and fluorescence (solid lines)
spectra of HBR-3,5DOM, HBR-3,5DM and HMBR when free in solution (dark lines) and bound (colored lines)
to (a,c,e) nanoFAST and (b,d,f) nano-pFAST. Spectra were recorded using chromophore concentration at 4
µM and proteins at 20 µM in pH 7.4 phosphate buffer saline at 25°C. (g) Affinities of HBR-3,5DOM, HBR-3,5DM
and HMBR with nanoFAST and nano-pFAST in PBS pH 7.4.
In conclusion, truncating the 27 first residues of pFAST resulted in a properly folded protein, nano-
pFAST, that maintain the ability to bind and activate most of the HBR derivatives and showed
remarkable properties with HBR-2,5DM, outperforming nanoFAST. Its observed lower solubility and
moderate performances with HBR-3,5DOM, HBR-3,5DM and HMBR could be optimized by rational
design and directed evolution. Improving the performances of nano-pFAST could allow the generation
of a multifunctional, small and robust protein tag with tunable properties and could be a benefit for
specific applications where the size of the protein tag is critical.
124
III.4. References
2. Keppler, A. et al. A general method for the covalent labeling of fusion proteins with small
molecules in vivo. Nat. Biotechnol. 21, 86–89 (2003).
Biol. 15, 128–136 (2008).
5. Hori, Y. et al. Development of Fluorogenic Probes for Quick No-Wash Live-Cell Imaging of
Intracellular Proteins. J. Am. Chem. Soc. 135, 12360–12365 (2013).
7. Gautier, A. & Tebo, A. G. Sensing cellular biochemistry with fluorescent chemical–genetic
hybrids. Curr. Opin. Chem. Biol. 57, 58–64 (2020).
8. Ben Aissa, H. & Gautier, A. Engineering Glowing Chemogenetic Hybrids for Spying on Cells.
Eur. J. Org. Chem. 2020, 5637–5646 (2020).
9. Broch, F. & Gautier, A. Illuminating Cellular Biochemistry: Fluorogenic Chemogenetic
Biosensors for Biological Imaging. ChemPlusChem 85, 1487–1497 (2020).
microscopy. Nat. Methods 12, 244–250 (2015).
Nat. Methods 14, 987–994 (2017).
12. Wang, L. et al. A general strategy to develop cell permeable and fluorogenic probes for
multicolour nanoscopy. Nat. Chem. 12, 165–172 (2020).
11, e02030-20 (2020).
1828 (2018).
Chem. Biol. 17, 30–38 (2020).
24. Meyer, T. E. et al. Site-Specific Mutations Provide New Insights into the Origin of pH Effects
and Alternative Spectral Forms in the Photoactive Yellow Protein from Halorhodospira halophila.
Biochemistry 42, 3319–3325 (2003).
25. Imamoto, Y. & Kataoka, M. Structure and Photoreaction of Photoactive Yellow Protein, a
Structural Prototype of the PAS Domain Superfamily†. Photochem. Photobiol. 83, 40–49 (2007).
26. Wei, L., Wang, H., Chen, X., Fang, W. & Wang, H. A comprehensive study of isomerization
125
and protonation reactions in the photocycle of the photoactive yellow protein. Phys. Chem. Chem.
Phys. PCCP 16, 25263–25272 (2014).
Last. Sci. Rep. 7, (2017).
29. Marriott, G. et al. Optical lock-in detection imaging microscopy for contrast-enhanced imaging
in living cells. Proc. Natl. Acad. Sci. 105, 17789–17794 (2008).
30. Quérard, J. et al. Resonant out-of-phase fluorescence microscopy and remote imaging
overcome spectral limitations | Nature Communications. Nat. Commun. 8, 969 (2017).
31. Kwon, J. et al. Bright ligand-activatable fluorescent protein for high-quality multicolor live-cell
super-resolution microscopy. Nat. Commun. 11, 273 (2020).
32. Turkowyd, B., Virant, D. & Endesfelder, U. From single molecules to life: microscopy at the
nanoscale. Anal. Bioanal. Chem. 408, 6885–6911 (2016).
33. Sednev, M. V., Belov, V. N. & Hell, S. W. Fluorescent dyes with large Stokes shifts for super-
resolution optical microscopy of biological objects: a review. Methods Appl. Fluoresc. 3, 042004
(2015).
34. Murakoshi, H., Shibata, A. C. E., Nakahata, Y. & Nabekura, J. A dark green fluorescent protein
as an acceptor for measurement of Förster resonance energy transfer. Sci. Rep. 5, 15334 (2015).
35. Murakoshi, H. & Shibata, A. C. E. ShadowY: a dark yellow fluorescent protein for FLIM-based
FRET measurement. Sci. Rep. 7, 6791 (2017).
(2019).
1120 (2019).
38. Britton, H. T. S. & Robinson, R. A. CXCVIII.—Universal buffer solutions and the dissociation
constant of veronal. J. Chem. Soc. Resumed 1456–1462 (1931).
39. Tebo, A. G., Pimenta, F. M., Zhang, Y. & Gautier, A. Improved Chemical-Genetic Fluorescent
Markers for Live Cell Microscopy. Biochemistry 57, 5648–5653 (2018).
40. Hadjantonakis, A.-K. & Papaioannou, V. E. Dynamic in vivo imaging and cell tracking using a
histone fluorescent protein fusion in mice. BMC Biotechnol. 4, 33 (2004).
41. Konno, D. et al. Neuroepithelial progenitors undergo LGN-dependent planar divisions to
maintain self-renewability during mammalian neurogenesis. Nat. Cell Biol. 10, 93–101 (2008).
42. Tozer, S., Baek, C., Fischer, E., Goiame, R. & Morin, X. Differential Routing of Mindbomb1 via
Centriolar Satellites Regulates Asymmetric Divisions of Neural Progenitors. Neuron 93, 542-551.e4
(2017).
43. Peyre, E. et al. A lateral belt of cortical LGN and NuMA guides mitotic spindle movements and
planar division in neuroepithelial cells. J. Cell Biol. 193, 141–154 (2011).
44. Baek, C. et al. Mib1 prevents Notch Cis-inhibition to defer differentiation and preserve
neuroepithelial integrity during neural delamination. PLOS Biol. 16, e2004162 (2018).
45. Vicente-García, E., Ramón, R. & Lavilla, R. New Heterocyclic Inputs for the Povarov
Multicomponent Reaction. Synthesis 2011, 2237–2246 (2011).
46. Zheng, Y., Zhou, J., A. Sayre, D. & O. Sintim, H. Identification of bromophenol thiohydantoin
as an inhibitor of DisA, a c-di-AMP synthase, from a 1000 compound library, using the coralyne assay.
Chem. Commun. 50, 11234–11237 (2014).
47. Ts, J. et al. Novel 3,5-diaryl pyrazolines and pyrazole as low-density lipoprotein (LDL)
oxidation inhibitor. Bioorg. Med. Chem. Lett. 14, 2719–2723 (2004).
48. Nakamura, S. et al. An Arylidene-Thiazolidinedione Derivative, GPU-4, without PPARγ
Activation, Reduces Retinal Neovascularization. Curr. Neurovasc. Res. 8, 25–34 (2011).
49. Ha, Y. M. et al. Design and synthesis of 5-(substituted benzylidene)thiazolidine-2,4-dione
derivatives as novel tyrosinase inhibitors. Eur. J. Med. Chem. 49, 245–252 (2012).
50. Cetenko, W. A., Connor, D. T., Sorenson, R. J., Unangst, P. C. & Stabler, S. S. Known and
selected novel arylmethylenyl derivatives of thiazolidinones, imidazolidinones and oxazolidinones
126
useful as antiallergy agents and anti-inflammatory agents. Patent US5208250 (May 4, 1993).
51. Sirivolu, V. R. et al. 5-Arylidenethioxothiazolidinones as Inhibitors of Tyrosyl–DNA
Phosphodiesterase I. J. Med. Chem. 55, 8671–8684 (2012).
52. Gibson, D. G. et al. Enzymatic assembly of DNA molecules up to several hundred kilobases.
Nat. Methods 6, 343–345 (2009).
53. Stemmer, W. P. Rapid evolution of a protein in vitro by DNA shuffling. Nature 370, 389–391
(1994).
54. Crameri, A., Raillard, S.-A., Bermudez, E. & Stemmer, W. P. C. DNA shuffling of a family of
genes from diverse species accelerates directed evolution. Nature 391, 288–291 (1998).
55. Coco, W. M. et al. DNA shuffling method for generating highly recombined genes and evolved
enzymes. Nat. Biotechnol. 19, 354–359 (2001).
56. Gietz, R. D. & Schiestl, R. H. Large-scale high-efficiency yeast transformation using the
LiAc/SS carrier DNA/PEG method. Nat. Protoc. 2, 38–41 (2007).
57. Dubois, L. et al. Amino Acids Bearing Aromatic or Heteroaromatic Substituents as a New Class
of Ligands for the Lysosomal Sialic Acid Transporter Sialin. J. Med. Chem. 63, 8231–8249 (2020).
58. Poirel, O. et al. LSP5-2157 a new inhibitor of vesicular glutamate transporters.
Neuropharmacology 164, 107902 (2020).
59. Thompson, J. D., Higgins, D. G. & Gibson, T. J. CLUSTAL W: improving the sensitivity of
progressive multiple sequence alignment through sequence weighting, position-specific gap penalties
and weight matrix choice. Nucleic Acids Res. 22, 4673–4680 (1994).
60. Wu, G., Robertson, D. H., Brooks, C. L. & Vieth, M. Detailed analysis of grid-based molecular
docking: A case study of CDOCKER—A CHARMm-based MD docking algorithm. J. Comput. Chem.
24, 1549–1562 (2003).
61. Krammer, A., Kirchhoff, P. D., Jiang, X., Venkatachalam, C. M. & Waldman, M. LigScore: a
novel scoring function for predicting binding affinities. J. Mol. Graph. Model. 23, 395–407 (2005).
62. Huff, J. The Airyscan detector from ZEISS: confocal imaging with improved signal-to-noise
ratio and super-resolution. Nat. Methods 12, i–ii (2015).
63. Danglot, L. et al. Vezatin Is Essential for Dendritic Spine Morphogenesis and Functional
Synaptic Maturation. J. Neurosci. 32, 9007–9022 (2012).
64. Niwa, H., Yamamura, K. & Miyazaki, J. Efficient selection for high-expression transfectants
with a novel eukaryotic vector. Gene 108, 193–199 (1991).
65. Edelstein, A., Amodaj, N., Hoover, K., Vale, R. & Stuurman, N. Computer Control of
Microscopes Using µManager. Curr. Protoc. Mol. Biol. 92, 14.20.1-14.20.17 (2010).
66. Mineev, K. S. et al. NanoFAST: Structure-based design of a small fluorogen-activating protein
with only 98 amino acids. bioRxiv 2020.12.29.424647 (2020)
67. Povarova, N. V. et al. Red-Shifted Substrates for FAST Fluorogen-Activating Protein Based
on the GFP-Like Chromophores. Chem. – Eur. J. 25, 9592–9596 (2019).
127
128
Chapter IV
Chemogenetic photosensitizers for

manipulating biological systems
IV.1 General introduction
Reactive oxygen species (ROS), which are known to cause damage in cells through oxidative
reactions, can allow the study of biological processes in ingenious ways. The selective generation of
ROS can be used for perturbing cell signaling, inducing cellular death, or protein inactivation (by
chromophore-assisted light inactivation (CALI)) (Figure IV.1). In addition, the selective generation of
ROS can also be used to locally convert 3,3’-diaminobenzidine (DAB) into a highly localized polymeric
precipitate, which can be stained with osmium tetroxide (OsO4) and identified by electron microscopy1
(Figure IV.1). Initially designed from genetically-encoded fluorescent proteins and chemogenetic
reporters, targeted ROS generators allowed specific photoinduced reactions in cells for various
biological applications.
129
CHAPTER IV: CHEMOGENETIC PHOTOSENSITIZERS FOR MANUPILATING BIOLOGICAL SYSTEMS
ROS Targeted photosensitizer
Chromophore-assisted Cellular ablation DAB polymerization ROS signaling

light inactivation (CALI) (Correlative light electron microscopy)
fused protein DAB
osmiophilic
polymer
imaging contrast in stress immune hypoxic ...

electron microscopy resistance signaling signaling
Figure IV.1. Schematic representation of the photoinduced reactions resulting from the generation of ROS by
genetically-encoded and chemogenetic photosensitizers upon light irradiation.
After absorption of light by a fluorophore, a part of the energy can be non-radiatively transferred to
the triplet state by intersystem crossing. This energy can be then transferred to a variety of oxygen
(and nitrogen) species within cells to generate free radicals and other highly reactive products before
eventually returning to the ground state by non-radiative or radiative (phosphorescence) deexcitation.
Upon light excitation, interaction of local molecular oxygen (3O2) with triplet state mostly generates
two types of reactive oxygen species: singlet oxygen 1O2 (via type II mechanisms) and superoxide
O2•– (via type I mechanisms)2 (Figure IV.2). While the generation of superoxide can result in a
cascade of different ROS such as hydrogen peroxide (H2O2) or hydroxyl radical (HO•), singlet oxygen
cannot directly convert to other ROS. Because of its high reactivity, singlet oxygen is believed to
cause universal damage in cells.
superoxide (type I)
S1 Intersystem
crossing O2•- H2O2 HO•
Non radiative decay
Fluorescence
T1
Absorption
Non radiative decay

Phosphorescence
1
O2 singlet oxygen (type II)
3
O2
S0
Figure IV.2. Simplified Jablonski diagram for classical production of reactive oxygen species (ROS) by a
photosensitizer through type I and type II mechanisms (adapted from Bevernaegie et al.3).
130
Chemicals such as malachite green, fluorescein, eosin, methylene blue and Rose Bengal are known
to efficiently produce reactive oxygen products. Although demonstrated to be excellent
photosensitizers, targeting and delivering chemical photosensitizers in cells proved to be ineffective.
Conjugating malachite green to an antibody4 or promoting fluorescein to bind a tetracysteine motif5
were not efficient enough to inactivate targeted proteins as they still suffer from low binding selectivity
and non-desirable accumulation in cells, overall generating phototoxicity6. Recently, incorporation of
halogen or sulfur atoms on rhodamine-based Janelia Fluor molecules allowed to design an efficient
photosensitizer (Figure IV.3a), which, combined with Halo-tag, allowed light-mediated destruction of
proteins, ablation of cells and was used in correlated light and electron microscopy7. As a matter of
fact, incorporation of heavy atoms favors the rate of intersystem crossing by increasing spin-orbit
coupling8. Therefore, modified fluorophores with relatively high atomic number enhance the
generation of reactive oxygen species9.
Figure IV.3. Photosensitizers based on (a) organic dyes (adapted from Binns et al.7), (b) genetically-encoded
fluorescent protein KillerRed (adapted from Wojtovich et al.6) and (c) fluorescent chemical-genetic hybrids:
miniSOG (adapted from Shu et al.10) and FAP-TAP (targeted and activated photosensitizer) (adapted from He
et al.9).
Genetically encoded photosensitizers on the other hand have emerged as excellent tools for targeted
reactive oxygen species generation. Firstly demonstrated in GFP11, most GFP-like proteins did not
allow significant ROS production, likely due to the inaccessibility of the HBI chromophore – trapped
in the β-barrel structure – for local molecular oxygen12. The first efficient genetically encoded
photosensitizer named KillerRed was evolved from a homolog of the GFP. KillerRed produces
superoxide and presumably other reactive oxygen species upon illumination with green light13. Its
tendency to oligomerize led to the design of its monomeric SuperNova variant14. Finally, as discussed
131
in Chapter II (section II.3.1) of this manuscript, actuators based on chemogenetic fluorescent

reporters benefited from chemical approaches by the development of FAP-TAP (Targeted and
Activated Photosensitizer)9 or from protein engineering strategies with the development of miniSOG
and its variants10,15–18 (Figure IV.3). In addition, a new Genetically encoded RNA Aptamer-based
Photosensitizer (GRAP) has been recently reported. Conjugation of known chemical photosensitizers
to a specific quenching molecule recognized by the RNA aptamer enabled efficient noninvasive target
cell ablation and stimuli-regulated ROS generation19.
To expand the photosensitizer toolkit, we wondered if the FAST-tagging system could be engineered
for targeting and activating the generation of reactive oxygen species. HBR derivatives with heavy
atoms were synthesized for enhancing the rate of intersystem crossing, thus increasing the transfer
from the singlet excited state to the triplet state. This triplet state then interacts with molecular oxygen
(3O2) to generate reactive oxygen species presumably through both electron transfer (type I) and
energy transfer (type II) photoreactions before relaxing to the ground state. The production of ROS is
only initiated when the chromophore binds to the protein tag, allowing the generation of on-demand
targeted and activated photosensitizers.
The molecules described in this chapter were synthetized in collaboration with the group of Prof.
Ludovic Jullien (Karim Ounoughi, Cyprien Velmir and Dr. Isabelle Aujard) of the PASTEUR laboratory
(Chemistry department, École Normale Supérieure).
132
IV.2 Engineering of targeted and activated chemogenetic photosensitizers for

manipulating biological systems on demand
IV.2.1. Introduction
Fluorescent labels and biosensors are powerful reporters for deciphering the complexity of living
systems. They allow biologists to label and study biomolecules and biological events in real time
without disturbing the cellular machinery. Fluorescent reporters can also be applied to study biological
processes in new ways through the design of photosensitizers able to generate reactive oxygen
species. In cells, ROS are naturally produced and regulated for the maintenance of physiological
functions. However, an abnormal increase of intracellular ROS levels can cause multiple damages in
cells through oxidative reactions20. Photosensitizers able to generate ROS can be used ingeniously
for functional manipulation of biological systems, including for instance selective signaling
perturbation, cellular ablation and protein inactivation (by chromophore-assisted light
inactivation)4,7,9,13,18,21,22. Fluorescent ROS generators have also been used as probes for correlative-
light and electron microscopy using their ability to locally generate ROS to convert 3,3’-
diaminobenzidine (DAB) into a highly localized polymeric precipitate, which can be stained with
osmium tetroxide (OsO4) and identified by electron microscopy7,10. Targeted and local ROS
generation in living cells was achieved using genetically-encoded photosensitizers such as
KillerRed13, miniSOG10 or on-demand chemogenetic photosensitizers7,9,19. Genetic targeting allows
the selective ablation of specific cells in a population or the selective inactivation of a proximal target.
Here, we present the development of a ROS-generating chemogenetic photosensitizer based on the

FAST technology. Prototypical FAST (Fluorescence-Activating and absorption-Shifting Tag) is a small
14kDa protein tag, able to form with exquisite specificity non-covalent bright fluorescent assemblies
with fluorogenic 4-hydroxybenzylidene rhodanine (HBR) derivatives and other analogs23–27. The
functionalization of HBR with heavy atoms such as bromine and iodine allowed us to enhance the
generation of reactive oxygen species through heavy atom effect28. Like regular HBR derivatives,
these chromophores dissipate light energy through rapid non-radiative processes when free in
solution and in cells, and thus are unable to produce any ROS unless they are ‘activated’ by FAST,
ensuring highly selective ROS generation. We called such activatable chromophores “ROSgens” to
highlight their ability to generate ROS only when bound to FAST variants. The possibility to generate
ROS on-demand by the addition of the ROSgenic chromophores allows for strict control of the
photoinduced reactions. We demonstrated that the resulting chemogenetic photosensitizers are able
to generate fluorescence and reactive oxygen species when excited with 488 nm light and allowed
one to photoinduce mitochondrial disruption when expressed in live cells.
133
IV.2.2. Results
IV.2.2.1. Molecular engineering and in vitro characterization of ROSgens
To develop an on-demand ROS-generating photosensitizer, we exploited the molecular tunability of

HBR derivatives. Push-pull chromophores composed of an electron-donating phenol conjugated to
an electron-withdrawing rhodanine, HBR derivatives exhibit high fluorescence efficiency when they
adopt a planar conformation in the cavity of prototypical FAST, and are otherwise dark in solution.
Their high fluorogenicity and high fluorescence performance make them well suited for wash-free
imaging of FAST-tagged proteins in live cells23,25,27. The introduction of heavy atoms on common
fluorescent and fluorogenic chromophores9,19 is known to increase the generation of ROS by
enhancing the rate of intersystem crossing. We thus synthesized HBR analogs substituted with
bromine or iodine atoms in ortho positions of the electron-donating phenol (Figure IV.4).
fluorescent
chemogenetic reporter
S1 O
O
S1
non radiative
R2
R2
Fluorescence
NH Fluorogen
decay
S NH
HO S
S + O
R1 S
pFAST R1
HMBR (R1 = Me, R2 = H)
S0 –
HBR-3,5DM (R1 = Me, R2 = Me)
S0
HBR-3,5DOM (R1 = OMe, R2 = OMe)
heavy-atom
effect chemogenetic
photosensitizer
O
ROSgen O S1
S1 ISC Cell killing
+
Fluorescence
Y Y ROS
non radiative
NH NH T1 ROS signaling
S S
decay
HO – O 3
S S O2
X X
HBR-3Br (X = Br, Y = H) S0
S0 HBR-3,5DBr (X = Br, Y = Br)
HBR-3,5DI (X = I, Y = I)
Figure IV.4. Chemogenetic photosensitizers for manipulating biological systems. pFAST can form ROS-
generating assemblies with heavy-atom-substituted HBR analogs. The targeted and activated generation of
reactive oxygen species (ROS) can be used for inducing selective ROS signaling and cell killing.
Just like HBR analogs, these heavy-atom-substituted HBR derivatives are non-fluorescent in solution,
but form tight (KD from 5 to 100 nM) fluorescent assemblies with FAST, iFAST29 and pFAST27 (Figure
IV.5 and Table 1). The wavelengths of absorption and emission, and molar absorptivity of the
bimolecular assemblies were comparable to those formed with the prototypical fluorogen HMBR
(Figure IV.5, Table 1). However we noticed a significant reduction of fluorescence quantum yield of
approximately 40% for HBR-3,5DI and about 70% for both HBR-3,5DBr and HBR-3Br (Figure IV.5e,
Table 1), suggesting that a part of the absorbed light energy is dissipated through other deexcitation
pathways. Note that unlike prototypical HMBR, which undergoes a strong red shift in absorption upon
134
binding because the protonation/deprotonation equilibrium is shifted from the protonated phenol in
solution to the deprotonated phenolate within FAST, the heavy-atom-substituted HBR derivatives
undergo a modest shift in absorption (Figure IV.5). As a matter of fact, they are already (at least
partially) deprotonated in solution because of the presence of electron-withdrawing halogen atoms in
ortho position of the phenol (the pKA of HBR-3Br is for instance 6.8, Figure IV.S1 in supporting
information IV.2.4).
a b e
HMBR HBR-3,5DI
60000
Fluorescence (a.u.)
- pFAST 1.0 60000 - pFAST 1.0
50000
50000
(M-1.cm-1)
+ pFAST + pFAST
40000 40000
30000 0.5 30000 0.5
20000 20000
10000 10000
0.0 0 0.0
400 500 600 700 400 500 600 700
c HBR-3,5DBr d HBR-3Br
- pFAST 1.0 60000
Fluorescence (a.u.)
50000 - pFAST 1.0
50000
(M-1.cm-1)
40000 + pFAST + pFAST

40000
30000
0.5 30000 0.5
20000 20000
10000 10000
0 0.0 0 0.0
400 500 600 700 400 500 600 700
Figure IV.5. Absorption and emission properties of heavy-atom-substituted HBR analogs bound to
pFAST. (a-d) Absorption (dashed lines) and fluorescence (solid lines) spectra of (a) HMBR (b) HBR-3,5DI, (c)
HBR-3,5DBr and (d) HBR-3Br when free in solution (dark lines) and bound to pFAST (colored lines). Spectra
were recorded using chromophore concentrations at 3 μM and pFAST at 40 μM in pH 7.4 phosphate buffer
saline at 25°C. (e) Relative fluorescent quantum yields of the different fluorescent assemblies.
Table IV.1. Properties of FAST, iFAST and pFAST with heavy-atom-substituted HBR analogs
Dlabs e Molecular
Mutant Fluorogen labs (nm) lem (nm) Φ KD (µM)
(nm) (mM-1cm-1) brightness
HMBR (ref 23) 481 79 540 44 0.23 10,000 0.13

HBR-3,5DI 484 33 547 51 0.15 7,600 0.11
FAST
HBR-3,5DBr 474 20 536 45 0.07 3,100 0.04
HBR-3Br 469 30 529 51 0.07 3,500 0.07
HMBR (ref 29) 480 78 541 41 0.22 9,000 0.07
HBR-3,5DI 481 30 544 55 0.13 7,200 0.07
iFAST
HBR-3,5DBr 473 20 536 54 0.06 3,200 0.02
HBR-3Br 469 29 530 48 0.06 3,000 0.02
HMBR (ref 27) 481 79 542 54 0.23 13,000 0.01
HBR-3,5DI 486 34 548 56 0.14 7,900 0.01
pFAST
HBR-3,5DBr 475 30 536 50 0.06 3,000 0.02
HBR-3Br 470 23 530 54 0.07 3,800 0.005
binding; lem wavelength of maximal emission; e, molar absorptivity at labs (standard error is typically 10%); Φ, fluorescence quantum yield;
molecular brightness = Φ x e ; KD thermodynamic dissociation constant.
135
To investigate the ability of heavy-atom-substituted HBR analogs to generate ROS when bound to
FAST variants, we first performed experiments in presence of anthracene-9-10-dipropionic acid
(ADPA), a commonly used singlet oxygen scavenger30. Singlet oxygen is known to react with
anthracene aromatic compounds to form an oxidized product31 (Figure IV.6.a). Thus, estimation of
ADPA photooxidation gives an indirect evaluation of the generation of singlet oxygen9,10. This
photobleaching reaction is however not fully specific to the presence of 1O2, as it has been reported
that anthracenes can also be oxidized by electron transfer processes, leading to overestimation of
the singlet oxygen quantum yield32–34. With this limitation in mind, we used ADPA photobleaching to
characterize the behavior of various assemblies between FAST variants and heavy-atom-substituted
HBR analogs. Samples were illuminated continuously with a 488 nm light (90 mW/cm2 intensity) and
the fluorescence intensity of ADPA was measured over time in order to estimate the efficiency of 1O2
generation (Figure IV.6, Figures IV.S3 and S4 in supporting information IV.2.4). Rose Bengal (RB),
a well-known chemical photosensitizer35 (see Figure IV.3 for structure), was used as reference. As
ROS generation can lead to photodestruction of the photosensitizer, we also monitored the
fluorescence intensity of the assemblies under the same illumination conditions to verify that
photodestruction was minimal during the duration of the experiment (Figure IV.S2). Comparison of
FAST assemblies with heavy-atom substituted HBR analogs with the FAST:HMBR assembly showed
that heavy atom substitution increase the efficiency of 1O2 generation, by 240% for FAST:HBR-3,5DI,
270% for FAST:HBR-3Br and 360% for FAST:HBR-3,5DBr. Furthermore, we showed that
iFAST:HBR-3,5DBr and pFAST:HBR-3,5DBr assemblies were 50% and 60% more efficient than
FAST:HBR3,5DBr, respectively, suggesting that the protein environment can tune the efficiency of
1
O2 generation for a given chromophore. The best assembly, pFAST:HBR-3,5DBr, was 13.5% and
21% as efficient as RB and miniSOG, respectively.
136
a ADPA were dissolved inADPA 20:80were

(v/v)dissolved
methanol:
"5
in deionised
20:80 (v/v) water,
methano
with a f nal ADPA concentration
with a f nal
of 5.22
ADPA ! 10 mol/dm3.of
concentration A 5.22 ! 10" 5
control
10" 5 mol/dm
solution of ADPA (5.22 !solution of ADPA3
(5.22
) in 20:80! 10 "5
(v/v) mol/dm 3
methanol:) in 20
deionised water, containing deionised
no TMPyP
water,was containing
also tested.
no TMPyP
Solutions
was al
2
were exposed to white were
light
exposed
(8.91 mW/cm
to white light (8.91from
) generated mW/cm
25 W halogen lights (Halolux25 W halogen
Ceram, Osram,
lights (Halolux
Germany), Ceram,
at a fOsram,
xed G
ADPA photooxidation distance of 11 cm for time distance
periods of ranging
11 cm for from
time5 to
periods
20 min,
ranging
with from
1
O2 + the use of a fan to maintainthe use
temperature
of a fan toatmaintain
27 !C, or temperature
maintained in at 27
dark conditions. Absorption dark spectra
conditions.
were Absorption
recorded between
spectra were
320 reco
and 550 nm, allowing visualisation
and 550 nm,ofallowing
the main visualisation
absorption of peaks
the mai
of ADPA, and the Soret band of ADPA,
of TMPyP.
and the Soret band of TMPyP.
To quantify the 1O2 generated
To quantify the 1O2ongenerated
by TMPyP irradiation,
by the
TMPyP
reactions occurring in solution
reactionsrequire
occurring consideration.
in solution Scheme
require 2 consid
describes the processes thatdescribes
occurthe
when processes
ADPA and thatTMPyP
occur are
whenirra-
ADPA
b Scheme 1. The conversion of Scheme
ADPA to 1.
anThed
endoperoxide
conversionon
of reaction
ADPA to with 1
an endoperoxide
O2. diated together
on reaction with 1Oin
2. solution
diated
(modif
together
ed from
in solution
[19,21]): (modif ed from [19
100
FAST FAST: FAST: FAST: FAST: RB
ADPA norm. fluo. (a.u.)
80 HMBR HBR-3,5DI HBR-3Br HBR-3,5DBr

FAST:HMBR
60 FAST:HBR-3Br k (rate constant)
0.030 0.102 0.111 0.140 1.564
FAST:HBR-3,5DI min-1
40
FAST:HBR-3,5DBr
20 Relative (%) 1.9 6.5 7.1 8.9 100
SO
RB
0
0 1 2 3 4 5
time (min)
c e
100
HBR-3,5DBr
ADPA norm. fluo. (a.u.)
miniSOG pFAST: iFAST: FAST: RB

80 FAST:HBR-3,5DBr HBR-3,5DBr HBR-3,5DBr HBR-3,5DBr
60 iFAST:HBR-3,5DBr
k (rate constant)
pFAST:HBR-3,5DBr 1.000 0.207 0.194 0.125 1.543
40 min-1
miniSOG
20
RB Relative SO (%) 65 13.5 12.5 8.1 100
0
0 1 2 3 4 5
time (min)
Figure IV.6. Evaluation of reactive oxygen species generation through photooxidation of anthracene-
9,10-dipronoic (ADPA). (a) The conversion of ADPA to an endoperoxide on reaction with singlet oxygen (1O2).
(adapted from Craig et al.36). (b,c) Solutions of ADPA and various photosensitizers were irradiated with 488 nm
light (90 mW.cm-2). ADPA fluorescence was measured after different exposure times (Excitation 374 nm and
Emission 384 – 550 nm) (see also Figures IV.S3 and IV.S4). Rose Bengal (RB) and miniSOG were used as a
standard and control of singlet oxygen generation. MiniSOG was 64% as efficient as RB, in good agreement
with initial results obtained with ADPA photooxidation6. Data were fitted using one phase exponential decay
equation. The absorption at 488 nm was identical in all samples (0.055 ± 10%), enabling direct comparison of
the kinetics and estimation of the relative singlet oxygen quantum yields. (d,e) Overview tables of the rate
constants k (min-1) and relative singlet oxygen quantum yields (ΦSO).
To test the generation of other ROS, we then performed experiments using the dihydroethidium (HE)
probe which is oxidized by a number of reactive oxygen species including superoxide radicals and
hydrogen peroxide37. HE fluoresces in blue until oxidized to ethidium, a fluorescent product with an
emission maximum at 610 – 630 nm. As 1O2 does not oxidize HE33,38, this probe provides an index of
ROS generation mostly produced through electron transfer processes (type I ROS). Solutions of
photosensitizers and HE were illuminated with 488 nm light (90 mW/cm2 intensity). The generation
of ethidium was measured over time, allowing us to estimate the efficiency of type I ROS generation
(Figure IV.7, Figures IV.S5 and IV.S6). We used miniSOG, reported to generate efficiently O2•– and
H2O233,34, as a reference. We observed that none of the ROSgenic complexes resulted in efficient HE
137
oxidation compared to miniSOG (Figure IV.7). Only about 30% of HE was oxidized after 5 min of
irradiation using pFAST:HBR-3,5DBr and FAST:HBR-3Br, while the same oxidation level was
reached within 5 s with miniSOG. Although the generation of type I ROS was very modest, we
nonetheless observed that heavy atom substitution increased type I ROS generation and that pFAST
was a slightly better scaffold compared to FAST and iFAST (Figure IV.7). In accordance with the
results obtained using both ADPA and HE photooxidation, it appeared that miniSOG and our targeted
chemogenetic photosensitizers were not producing the same reactive oxygen species (Figures IV.6
and IV.7).
a
H2N NH2 H2N H2N NH2 NH2 H2N NH2
HE photooxidation
+ +
H N N H N N
CH2CH3 CH2CH3 CH2CH3 CH2CH3
b HE dE+ HE E+
miniSOG
Norm. fluorescence (a.u.)
100
FAST: FAST: FAST: FAST: miniSOG
FAST:HBR-3Br = 520 nm HMBR HBR-3,5DI excitation
HBR-3Br
= 520 nmHBR-3,5DBr
80 excitation
= 610 nm emission = 610 nm
60 FAST:HBR-3,5DBr emission
k (rate constant)
0.022 of hydroethidine
0.049 (HE) and 0.083 0.067
ethidium (E+). 4.226
of hydroethidine (HE) and Fig.
Fig. 1. Chemical structureFAST:HBR-3,5DI 1. -1
ethidium
min Chem ical structure
(E+).
40
FAST:HMBR
20
Relative ROS I (%) 0.5 1.0 2.0 1.6 100
FAST
0
0 1 2 3 4 5
time (min)
c e
100
Norm. fluorescence (a.u.)
miniSOG pFAST: iFAST: FAST: miniSOG

80 HBR-3,5DBr HBR-3,5DBr HBR-3,5DBr
pFAST:HBR-3,5DBr
60 k (rate constant)
iFAST:HBR-3,5DBr 0.103 0.088 0.071 4.892
40 min-1
FAST:HBR-3,5DBr
20 Relative (%)
ROS I 2.1 1.8 1.5 100
HBR-3,5DBr
0
0 1 2 3 4 5
time (min)
Figure IV.7. Evaluation of reactive oxygen species generation through photooxidation of

dihydroethidium. (a) The conversion of dihydroethidium (HE) to ethidium on reaction with type I ROS (adapted
from Gomes et al.37). (b,c) Solutions of HE and various photosensitizers were irradiated with 488 nm light (90
mW.cm-2). Ethidium fluorescence was measured after different exposure times (Excitation 525 nm and emission
600 – 750 nm) (see also Figure IV.S5 and IV.S6). Values were normalized by the maximal fluorescent
intensities obtained with miniSOG and data were fitted using one phase exponential association equation. The
absorption at 488 nm was identical in all samples (0.055 ± 10%), enabling direct comparison of the kinetics and
estimation of the relative Type I ROS quantum yields. (d,e) Overview tables of the rate constants k (min-1) and
relative Type I ROS quantum yields (ΦROS I).
138
IV.2.2.2. Selective dual imaging and cell perturbation
Preliminary experiments showed that incubation with the ROSgenic dyes alone for 24 hours had no
deleterious effects on mammalian cells at the concentrations used for labeling (Figure IV.S7).
Although less bright than with regular HBR derivatives, pFAST assemblies with heavy-atom
substituted HBR derivatives were bright enough for the selective imaging of proteins in living cells by
confocal microscopy. The use of low excitation intensities and short exposure times allowed us to
image HeLa cells expressing pFAST in the cytoplasm, in mitochondria or at the inner plasma
membrane without affecting their morphology (Figure IV.8).
pFAST mito-pFAST lyn11-pFAST

HBR-3,5DI

HBR-3,5DBr

HBR-3Br
Figure IV.8. Selective imaging of pFAST in live mammalian cells. Confocal micrographs of live HeLa cells
expressing pFAST, or pFAST fused to the mitochondrial targeting sequence from the subunit VIII of human
cytochrome C oxidase (Mito-pFAST) and to the inner plasma membrane targeting sequence lyn11 (Lyn11-
pFAST) and labeled with 10 µM HBR-3,5DI, HBR-3,5DBr and HBR-3Br. Imaging conditions: Exc 488 nm, Em
500 – 590 nm. Scale bars, 10 µm.
Targeting the generation of ROS in cellular compartments as mitochondria, cell membranes or in the
nucleus can mediate strong cellular perturbations. ROS can cause damage to DNA, can lead to lipid
oxidation within the cellular membrane and can initiate different cellular death pathways in
mitochondria6. Mitochondrial dysfunction caused by a high level of ROS can induce cellular death
either by causing ATP depletion and energetic collapse or by promoting the release of the
139
intermembrane space proteins by the breakage of mitochondria outer membrane39. To evaluate our
ability to photogenerate ROS locally in cells and perturb cellular functions, HeLa cells expressing
mito-pFAST and treated with the ROSgenic chromophore HBR-3,5DBr were irradiated with 488 nm
light at low intensity for 10 minutes. Note that light intensity was chosen to avoid any non-specific
apparent cellular toxicity and was lower than the intensity usually used for imaging. Co-expression of
mito-mCherry allowed us to monitor changes in mitochondria morphology by imaging with 560 nm
light excitation. In cells treated with HBR-3,5DBr, 488 nm irradiation led to the specific formation of
globular and donut-shaped (toroidal) mitochondria (Figure IV.9a and Figure IV.S8a,b). Alterations of
mitochondria morphology are thought to be related to osmotic stress and oxidative damage. The
formation of globular and donut-shaped mitochondria was previously reported as a unique
morphological feature of mitochondrial dysfunction linked to a loss of mitochondrial membrane
potential40,41. Control experiments using HMBR instead of HBR-3,5DBr showed no alteration of
mitochondria morphology (Figure IV.9b and Figure IV.S8c,d), suggesting that the formation of
globular and donut-shaped mitochondria was induced by the specific photogeneration of ROS by
pFAST:HBR3,5DBr. Additional control experiments were performed irradiating HeLa cells expressing
mito-miniSOG using the same illumination conditions. We observed in that case more drastic cell
death features such as mitochondria fragmentation and cell shrinkage (Figure IV.9c and Figure
IV.S9). Unlike pFAST:HBR-3,5DBr that did not photobleach too much during the time course of the
experiment (Figure IV.9a and Figure IV.S8a,b), we noticed that illumination caused significant
miniSOG fluorescence photobleaching, suggesting that miniSOG could self-inactivate under these
conditions. Accordingly, the use of lower light intensity reduced miniSOG fluorescence
photobleaching and led to a stronger cellular response (Figure IV.S9c-e).
Overall, these preliminary results suggested that pFAST:HBR-3,5DBr can generate ROS upon light
illumination, and promote cellular dysfunction in a highly specific manner. The cellular perturbation
was less pronounced than in the case of miniSOG, in agreement with the lower ROS generation
efficiency observed in vitro. Such milder ROS generator could be used advantageously to induce in
a specific manner cellular dysfunctions more subtle than cell death, such as the formation of globular
and donut-shaped mitochondria, which was previously observed in cells under hypoxia-reoxygenation
conditions40.
140
a mito-pFAST:HBR-3,5DBr / mito-mCherry
0 min 4 min 6 min 10 min
10 min
b mito-pFAST:HMBR / mito-mCherry
10 min
c mito-miniSOG / mito-mCherry
10 min
Figure IV.9. The photoinduced reaction of mitochondrial disruption in HeLa cells by the generation of
reactive oxygen species. Confocal micrographs of live HeLa cells co-expressing (a,b) mito-pFAST or (c) mito-
miniSOG and (a,b,c) mito-mCherry. HeLa cells were irradiated with 488 nm laser excitation (light intensity 13.7
kW/cm2, pixel dwell time 4.9 µs)) for 10 minutes (1 scan / 5 sec). The fluorescence of (a,b) mito-pFAST in
presence of (a) 10 µM HBR-3,5DBr and (b) 10 µM HMBR, and of (c) mito-miniSOG were recorded before and
after irradiation. Mitochondria were monitored by the fluorescence of the co-expressed mito-mCherry (n = 3
replicates, see Figures IV.S9 and IV.S10). Scale bars, 10 µm.
141
IV.2.3. Discussion
Photosensitizers can be used in a variety of biological applications such as cellular perturbation,

targeted inactivation of proteins, cellular ablation, or labeling individual proteins by imaging contrast
in situ. We demonstrated that functionalizing the HBR fluorogenic chromophores with heavy atoms
allowed the generation of pFAST-based chemogenetic photosensitizer. Monitoring the photooxidation
of specific chemical fluorescent probes, we demonstrated that incorporation of bromine or iodine
atoms in HBR derivatives allowed to enhance the generation of ROS when they are bound to FAST
variants. Photosensitization is initiated by the strict addition of the ROSgenic chromophores, allowing
for on-demand activated and targeted photo-manipulation. This method allows to strictly control the
photoinduction reaction, eliminating the undesirable activation of the photosensitizers under ambient
light.
We demonstrated that pFAST:HBR-3,5DBr could act as a fluorescent reporter and a photosensitizer

able to induce cellular dysfunction with light. When targeted to mitochondria, this chemogenetic
photosensitizer allowed the selective photoinduction of morphological changes typical of
dysfunctional mitochondria. Although it was demonstrated that pFAST:HBR-3,5DBr produced lower
levels of ROS than miniSOG, such milder ROS generator could be used as ROS signaling to study
essential biomolecules in live cells, without irreversibly affecting their activities. In the near future, we
will further study our ability to photoinduce cellular dysfunction by targeting the expression of pFAST
to the plasma membrane or in the nucleus to photoinduce cellular dysfunction and/or alternative
cellular death pathways. Finally, we will evaluate the use of pFAST:HBR-3,5DBr for protein
inactivation by chromophore-assisted light inactivation (CALI) strategy. For a proof-of-concept, we
have already generated constructions for expressing δ1 Pleckstrin Homology (PH) fused to pFAST to
follow the membrane release of the PH domain upon light-induced ROS generation9,21.
142
IV.2.4. Supplementary information
This part includes:

Figures IV.S1 to IV.S9
a b
40000 pH pKa = 6.8
(395nm) (M–1.cm–1)
10.7 30000 40000
(447 nm) (M–1.cm–1)

(M–1.cm–1)
30000
25000 30000
20000 5.7
20000
20000
10000
15000
0 10000
300 350 400 450 500 550 10000
5 6 7 8 9 10 11
Wavelength (nm)
pH
Figure IV.S1. Absorption properties of HBR-3Br in solution in function of pH. (a) Absorption spectra at
various pH. (b) Evolution of the absorption at 395 nm and at 447 nm as a function of pH. The spectra were
recorded in 0.05 M Britton–Robinson buffer42 (0.28 M ionic strength) at 25°C.
a b
100 100 HBR-3,5DBr
FAST
Fluorescence (a.u.)
Fluorescence (a.u.)
80 80
60 60
40 HMBR 40
HBR-3,5DI FAST
20 HBR-3,5DBr 20 iFAST
HBR-3Br pFAST
0 0
0 1 2 3 4 5 0 1 2 3 4 5
time (min) time (min)
Figure IV.S2. Photobleaching measurement of (a) FAST (15 μM) in presence of HMBR, HBR-3,5DI, HBR-
3,5DBr and HBR-3Br and (b) HBR-3,5DBr in presence of FAST, iFAST and pFAST at 15 μM. The concentration
of chromophores (around 4 to 6 μM) was adjusted to achieve optically matched solutions at 488 nm (abs0 =
0.055 ± 10%). Photobleaching was monitored under the same illumination conditions as for ADPA
photobleaching and HE oxidation experiments (488 nm light exposure, 90 mW.cm-2) (Ex 488 nm; Em 498 – 700
nm). Data were fitted using one phase exponential distribution.
143
a 0.20 5min b 0.20 2.5×106

5min
4.5min
2.5×106
0min FAST FAST:HMBR 4min
Fluorescence (a.u.)
Fluorescence (a.u.)
FAST 4.5min 0min FAST:HMBR
2.0×106 4min 2.0×106 3.5min
0.15
Absorbance
0.15
Absorbance
3.5min 3min
1.5×106 3min 1.5×106 2.5min
0.10 2min 0.10 2min
1.0×106 1.5min 1.0×106 1.5min
0.05 1min 0.05 5.0×105 1min
5.0×105 0.5min 0.5min
0.00 0.0 0min 0.00 0.0 0min
300 400 500 600 400 450 500 550 300 400 500 600 400 450 500 550
Wavelength (nm) Wavelength (nm) Wavelength (nm) Wavelength (nm)
c 0.20 2.5×10 6
5min
4.5min d 0.20 2.5×106
5min
4.5min
0min FAST:HBR-3,5DI FAST:HBR-3,5DI 4min 0min FAST:HBR-3,5DBr FAST:HBR-3,5DBr 4min
Fluorescence (a.u.)
Fluorescence (a.u.)
2.0×10 6 3.5min 2.0×106 3.5min
0.15 0.15
Absorbance
Absorbance
3min 3min
1.5×106 2.5min 1.5×106 2.5min
0.10 2min 0.10 2min
1.0×106 1.5min 1.0×106 1.5min
0.05 1min 0.05 1min
5.0×105 5.0×105
0.5min 0.5min
0.00 0.0 0min 0.00 0.0 0min
300 400 500 600 400 450 500 550 300 400 500 600 400 450 500 550
e 5min f
0.20 2.5×106 4.5min 0.5 2.5×106
0min RB
Fluorescence (a.u.)
0min FAST:HBR-3Br FAST:HBR-3Br 4min RB
Fluorescence (a.u.)
2.0×10 6 3.5min 0.4 2.0×106 2.5min
Absorbance
0.15
Absorbance
3min 2min
1.5×106 2.5min 0.3 1.5×106 1.5min
0.10 2min 1min
1.0×106 0.2 1.0×106
1.5min 0.67min
0.05 1min 0.1 5.0×105 0.33min
5.0×105
0.5min 0min
0.00 0.0 0min 0.0 0.0
300 400 500 600 400 450 500 550 300 400 500 600 400 450 500 550
Figure IV.S3. Anthracene-9-10-dipronoic acid (ADPA) photooxidation measurements in presence of (a)

FAST (b) FAST:HMBR, (c) FAST:HBR-3,5DI, (d) FAST:HBR-3,5DBr, (e) FAST:HBR-3Br and (f) Rose Bengal
(RB). Absorption (left) and fluorescence (right) spectra were recorded and ADPA photooxidation was monitored
as a function of the duration of 488 nm light exposure (90 mW.cm-2) (Ex 374 nm; Em 384-550 nm). Experiments
were performed in presence of 50 μM ADPA and 15 μM of FAST; and the concentrations of HBR-3,5DI, HBR-
3,5DBr, HBR-3Br (around 4 to 6 μM) and RB were adjusted to achieve optically matched solutions at 488 nm
(abs0 = 0.055 ± 10%).
144
a 5min
4.5min
b 5min
4.5min
0.20 5×106 5×106
0min HBR-3,5DBr HBR-3,5DBr 4min FAST:HBR-3,5DBr 4min
Fluorescence (a.u.)
Fluorescence (a.u.)
4×106 3.5min 4×106 3.5min
0.15
Absorbance
3min 3min
3×106 2.5min 3×106 2.5min
0.10 2min 2min
2×106 1.5min 2×106 1.5min
0.05 1min 1min
1×106 1×106
0.5min 0.5min
0.00 0 0min 0 0min
300 400 500 600 400 450 500 550 400 450 500 550
Wavelength (nm) Wavelength (nm) Wavelength (nm)
c d 5min
4.5min 0.20 5×106 4.5min
0.20 5×106 pFAST:HBR-3,5DBr 4min
Fluorescence (a.u.)
iFAST:HBR-3,5DBr 4min 0min pFAST:HBR-3,5DBr
Fluorescence (a.u.)
0min iFAST:HBR-3,5DBr
3.5min 4×106 3.5min
4×106 0.15
Absorbance
0.15
Absorbance
3min 3min
3×106 2.5min 3×106 2.5min
0.10 0.10 2min
2min 2×106
2×106 1.5min 1.5min
0.05 0.05 1min
1×106 1min 1×106
0.5min 0.5min
0.00 0min 0.00 0 0min
300 400 500 600 0 300 400 500 600 400 450 500 550
400 450 500 550
e Wavelength (nm) f
0.25 5×106
0min miniSOG 2.5min 5min 0.5 5×106 2.5min
Fluorescence (a.u.)
miniSOG 2min 0min RB
Fluorescence (a.u.)
4.5min RB 2min
0.20 4×106 0.4 4×106
Absorbance
1.5min 4min 1.5min
Absorbance
0.15 3×106 1min 3.5min 1min
0.67min 3min 0.3 3×106
0.67min
0.10 2×106 0.5min 0.5min
0.33min 0.2 2×106
0.05 0.33min
1×106 0.17min 0.1 1×106 0.17min
0min 0min
0.00 0
300 400 500 600 400 450 500 550 0.0 0
300 400 500 600 400 450 500 550
Figure IV.S4. Anthracene-9-10-dipronoic acid (ADPA) photooxidation measurements in presence of (a)

HBR-3,5DBr, (b) FAST:HBR-3,5DBr, (c) iFAST:HBR-3,5DBr, (d) pFAST:HBR-3,5DBr, (e) miniSOG and (f)
Rose Bengal (RB). Absorption (left) and fluorescence (right) spectra were recorded and ADPA photooxidation
was monitored as a function of the duration of 488 nm light exposure (90 mW.cm-2) (Ex 374 nm; Em 384-550
nm emission). Experiments were performed in presence of 50 μM ADPA; FAST, iFAST and pFAST at 15 μM;
and the concentrations of HBR-3,5DBr (4 – 5 μM), RB and miniSOG were adjusted to achieve optically matched
solutions at 488 nm (abs0 = 0.055 ± 10%).
145
a 0.20 b 0.20
FAST:HMBR 5min
FAST 5min
Fluorescence (a.u.)
Fluorescence (a.u.)
FAST FAST:HMBR 6×105
6×105 4.5min 4.5min
0.15 0.15 4min
Absorbance
Absorbance
4min
3.5min 3.5min
4×105 3min 4×105 3min
0.10 0.10 2.5min
2.5min
2min 2min
0.05 2×105 0.05 2×105 1.5min
1.5min
1min 1min
0min 0.5min 0min 0.5min
0.00 0 0.00 0
300 400 500 600 600 650 700 750 300 400 500 600 600 650 700 750
c 0.20 d 0.20
FAST:HBR-3,5DBr 5min
Fluorescence (a.u.)
FAST:HBR-3,5DI 5min FAST:HBR-3,5DBr
Fluorescence (a.u.)
FAST:HBR-3,5DI 6×105 6×105 4.5min
4.5min 0.15 4min
Absorbance
0.15 4min
Absorbance
3.5min 3.5min
4×105 4×105 3min
0.10 3min 0.10 2.5min
2.5min 2min
2×105
2min
0.05 2×105 1.5min
0.05 1.5min
1min 1min
0min 0min 0.5min
0 0.5min 0.00 0
0.00 300 400 500 600 600 650 700 750
300 400 500 600 600 650 700 750
Wavelength (nm)
e 0.20 f 0.6 2.0×106 5min
FAST:HBR-3Br 5min
Fluorescence (a.u.)
miniSOG
Fluorescence (a.u.)
FAST:HBR-3Br miniSOG 4min

6×105 4.5min
1.5×106
Absorbance
0.15 3min
Absorbance
4min
3.5min 0.4 2min
0.10 4×105 3min 1.0×106 1min
2.5min 0.67min
5 2min 0.2 5.0×105 0.5min
0.05 2×10 1.5min 0.33min
0min 1min 0.17min
0.5min 0min 0.0
0.00 0 0.0 600 650 700 750
300 400 500 600 600 650 700 750 300 400 500 600
Wavelength (nm)
Figure IV.S5. Dihydroethidium (HE) photooxidation measurements in presence of (a) FAST (b)
FAST:HMBR, (c) FAST:HBR-3,5DI, (d) FAST:HBR-3,5DBr, (e) FAST:HBR-3Br and (f) miniSOG. Absorption
(left) and fluorescence (right) spectra were recorded and HE oxidation was monitored as a function of the
duration of 488 nm light exposure (90 mW.cm-2) (Ex 525 nm; Em 600-750 nm). Experiments were performed in
presence of 50 μM HE, 15 μM of FAST; and the concentrations of HBR-3,5DI, HBR-3,5DBr, HBR-3Br (around
4 to 6 μM) and miniSOG were adjusted to achieve optically matched solutions at 488 nm (abs0 = 0.055 ± 10%).
146
0.20 0.20
a HBR-3,5DBr 5min b FAST:HBR-3,5DBr 5min
Fluorescence (a.u.)
HBR-3,5DBr FAST:HBR-3,5DBr
Fluorescence (a.u.)
6×105 4.5min 6×105 4.5min
0.15 4min 0.15
Absorbance
Absorbance
3.5min 4min
3min 3.5min
0.10 4×105 0.10 4×105 3min
2.5min
2min 2.5min
2×105 1.5min 2min
0.05 1min 0.05 2×105 1.5min
0min 0.5min 0min 1min
0.00 0 0min 0.00 0.5min
300 400 500 600 300 400 500 600 0
600 650 700 750 600 650 700 750
c 0.20 d 0.20
iFAST:HBR-3,5DBr 5min pFAST:HBR-3,5DBr 5min
Fluorescence (a.u.)
Fluorescence (a.u.)
iFAST:HBR-3,5DBr pFAST:HBR-3,5DBr 6×105
6×105 4.5min 4.5min
0.15 0.15 4min
Absorbance
Absorbance
4min
3.5min 3.5min
4×105 3min 4×105 3min
0.10 0.10 2.5min
2.5min
2min 2min
0.05 2×105 0.05 2×105 1.5min
1.5min
1min 1min
0min 0min 0.5min
0.00 0 0.5min 0.00 0
300 400 500 600 600 650 700 750 300 400 500 600 600 650 700 750
e
0.6 2.0×106
miniSOG 5min
Fluorescence (a.u.)
miniSOG 4min
1.5×106 3min
Absorbance
0.4 2min
1.0×106 1min
0.67min
0.2 0.5min
5.0×105 0.33min
0.17min
0min 0.083min
0.0 0.0
300 400 500 600 600 650 700 750
Figure IV.S6. Dihydroethidium (HE) photooxidation measurements in presence of (a) HBR-3,5DBr (b)
FAST:HBR-3,5DBr, (c) iFAST:HBR-3,5DBr, (d) pFAST:HBR-3,5DBr and (e) miniSOG. Absorption (left) and
fluorescence (right) spectra were recorded and HE oxidation was monitored as a function of the duration of 488
nm light exposure (90 mW.cm-2) (Ex 525 nm; Em 600-750 nm). Experiments were performed in presence of 50
μM HE; FAST, iFAST and pFAST at 15 μM and the concentrations of HBR-3,5DBr (4 - 5 μM) and miniSOG
were adjusted to achieve optically matched solutions at 488 nm (abs0 = 0.055 ± 10%).
147
a + DMSO b +0.1 mg/mL digitonin c HBR-3Br 5 M d HBR-3,5DBr 5 M e HBR-3,5DI 5 M

CalceinAM
EthD1
Figure IV.S7. Two-color fluorescence viability assay. HeLa cells were incubated for 24 h with solutions of
(a-b) 0.1 % of DMSO, (c) HBR-3Br at 5 μM, (d) HBR-3,5DBr at 5 μM and (e) HBR-3,5DI at 5 μM. Control
experiments of HeLa cells non-incubated with dye (a, live cells) or incubated for 30 min with 0.1 mg/ml digitonin
(b, dead cells) are shown. Cell viability was tested with calceinAM and EthD1 probes (LIVE/DEAD®
viability/cytotoxicity assay kit). CalceinAM is a cell-permeant profluorophore cleaved by intracellular esterases
releasing a uniform green fluorescence polyanionic calcein in live cells (green channel). EthD1 (Ethidium
homodimer-1) is a non-permeant nucleic acid fluorescent stain that enters only cells with damaged membranes
and undergoes a fluorescence enhancement upon binding to nucleic acids, thereby producing a bright red
fluorescence in dead cells (pink channel). Cell fluorescence was evaluated by confocal microscopy. Identical
imaging settings were used for the different experiments. The experiment shows that none of the chromophores
are toxic for HeLa cells at the concentrations used for labeling. Scale bars, 30 µm.
148
a mito-pFAST:HBR-3,5DBr / mito-mCherry (n = 2)
10 min
b mito-pFAST:HBR-3,5DBr / mito-mCherry (n = 3)
10 min
c mito-pFAST:HMBR / mito-mCherry (n =2)

10 min
d mito-pFAST:HMBR / mito-mCherry (n =3)

10 min
Figure IV.S8. The photoinduced reaction of mitochondrial disruption in HeLa cells expressing mito-
pFAST. (a-d) Confocal micrographs of live HeLa cells co-expressing mito-pFAST and mito-mCherry in presence
of (a,b) 10 µM HBR-3,5DBr and (c,d) 10 µM HMBR. Cells were irradiated with 488 nm laser excitation at (light
intensity 13.7 kW/cm2, pixel dwell time 4.9 µs µs)) for 10 minutes (1 scan / 5 s). The fluorescence of mito-pFAST
in presence (a,b) HBR-3,5DBr or (c,d) HMBR were recorded before and after irradiation. Mitochondria were
monitored by the fluorescence of the co-expressed mito-mCherry (n = 3 replicates, see Figures IV.8). Scale
bars, 10 µm.
149
a 488 nm irradiation (13.7 kW.cm2) (n=2)

10 min
b 488 nm irradiation (13.7 kW.cm2) (n=3)


10 min
c d
488 nm irradiation (1.9 kW.cm2) (n=1) 0 min 10 min 488 nm irradiation (1.9 kW.cm2) (n=2) 0 min 10 min
0 min 0 min
10 min 0 min 10 min 10 min 0 min 10 min
e 488 nm irradiation (1.9 kW.cm2) (n=3) 0 min 10 min

0 min
10 min 0 min 10 min
Figure IV.S9. The photoinduced reaction of mitochondrial disruption in HeLa cells expressing mito-
miniSOG. (a-d) Confocal micrographs of live HeLa cells co-expressing mito-miniSOG and mito-mCherry were
irradiated with 488 nm laser excitation at (a,b) 13.7 kW/cm2 and (c-e) 1.9 kW/cm2 (pixel dwell time 4.9 µs) for
10 minutes (1 scan / 5 s). The fluorescence of mito-miniSOG was recorded before and after irradiation.
Mitochondria were monitored by the fluorescence of the co-expressed mito-mCherry (n = 3 replicates, see
Figures IV.8). Scale bars, 10 µm.
150
IV.2.5. Materials and methods
Organic synthesis
General. Commercially available reagents were used as obtained. 1H and 13

C NMR spectra were
recorded at 300 K on a Bruker AM 300 spectrometer; chemical shifts are reported in ppm with
protonated solvent as internal reference 1H, CHD2SOCD3 in CD3SOCD3 2.49 ppm; coupling constants
J are given in Hz. Mass spectra high resolution were performed by the Service de Spectrométrie de
Masse de l’Institut de Chimie Organique et Analytique (Orléans). The preparation of HMBR (4-
hydroxy-3-methylbenzylidene rhodanine) was previously described23.
HBR-3Br (4-hydroxy-3-bromobenzylidene)-3-phenyl-2-thioxothiazolidin-4-one). Solid 4-hydroxy-3-

bromobenzaldehyde (241 mg, 1.2 mmol ) and rhodanine (100 mg, 0.75 mmol) were mixed and the
mixture was brought to 170°C. It first fused and then solidified. After cooling to 50-60°C, ethanol was
added and the mixture was heated at reflux. The solid partially disintegrated and formed a suspension.
It was filtered and the remaining powder was dissolved in 1 M sodium hydroxide. HBR-3Br was
obtained as a yellow powder (60 %, 142 mg). 1H NMR (300 MHz, DMSO-d6, δ): 13.73 (s, 1H), 11.26
(s, 1H), 7.79 (s, 1H), 7.54 (s, 1H), 7.43 (d, J = 8 Hz, 1H), 7.10 (d, J = 8 Hz, 1H); 13C NMR (75 MHz,
DMSO-d6, δ): 195.3, 169.4, 156.6, 136.0, 130.9, 130.7, 125.6, 122.7, 112.1, 110.3 ; HRMS (ESI):
m/z 315.9096 (calcd mass for C10H7BrNO2S2 [M+H]-: 315.9096) (in accordance with previously
described analysis43–45).
HBR-3,5DBr (4-hydroxy-3,5-dibromobenzylidene)-3-phenyl-2-thioxothiazolidin-4-one). Solid 3,5-

dibromo-4-hydroxybenzaldehyde (252 mg, 0.9 mmol), rhodanine (100 mg, 0.75 mmol) and piperidine
(22 μL, 0.2 mmol) were mixed in ethanol (3 mL) and stirred at reflux temperature for 24h. After cooling,
the precipitate formed during the reaction was collected by filtration, washed with cold ethanol and
dried under reduced pressure. HBR-3,5DBr was obtained as an orange powder (78%, 232 mg). 1H
NMR (300 MHz, DMSO-d6, δ): 7.77 (s, 2H), 7.55 (s, 1H); HRMS (ESI): m/z 393.820175 (calcd mass
for C10H6Br2NO2S2 [M+H]-: 393.8201) (in accordance with previously described analysis46).
HBR-3,5DI (4-hydroxy-3,5-diiodobenzylidene)-3-phenyl-2-thioxothiazolidin-4-one). Solid 3,5-diiodo-

4-hydroxybenzaldehyde (342 mg, 0.9 mmol), rhodanine (100mg, 0.75mmol), and piperidine (22 μL,
0.2 mmol) were mixed in ethanol (3 mL) and stirred at reflux temperature for 24h. After cooling, the
precipitate formed during the reaction was collected by filtration, washed with cold ethanol and dried
under reduced pressure. HBR-3,5DI was obtained as a red powder (57%, 209 mg). 1H NMR (300
MHz, DMSO-d6, δ): 7.95 (s, 2H), 7.49 (s, 1H); HRMS (ESI): m/z 489.7917 (calcd mass for
C10H6I2NO2S2 [M+H]-: 489.7923) (in accordance with previously described analysis46).
151
Characterization of the chemogenetic photosensitizer complexes
Chemicals. Anthracene-9,10-dipropionic acid disodium salt (ADPA) was obtained from Coger,
dihydroethidium (HE) (≥ 95%) and Rose Bengal (RB) (95%) were obtained from Sigma-Aldrich and
used as received.
Expression. The plasmid for bacterial expression of miniSOG with an N-Terminal His-tag under the
control of a T7 promoter (pAG675) was constructed by replacing the sequence coding for FAST in
the plasmid pAG8723 using isothermal Gibson assembly. pAG675 and the plasmids driving E. coli
expression of FAST, iFAST, pFAST and miniSOG with an N-terminal His-tag (previously
described23,27,29) were transformed in Rosetta (DE3) pLys E. coli (New England Biolabs). Cells were
grown at 37°C in Lysogen Broth (LB) medium supplemented with 50 μg/mL kanamycin and 34 μg/mL
chloramphenicol to OD600nm 0.6. Expression was induced overnight at 16°C by adding isopropyl-β-D-
1-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. Cells were harvested by
centrifugation (4,300 ´ g for 20min at 4°C) and frozen.
Purification. The cell pellet was resuspended in lysis buffer (PBS supplemented with 2.5 mM MgCl2,
1mM of protease inhibitor PhenylMethaneSulfonyl Fluoride PMSF, 0.025 mg/mL of DNAse, pH 7.4)
and sonicated (5 min, 20 % of amplitude) on ice. The lysate was incubated for 2-3 hours on ice to
allow DNA digestion by DNAse. Cellular fragments were removed by centrifugation (9,000 x g for 1h
at 4°C). The supernatant was incubated overnight at 4°C by gentle agitation with pre-washed Ni-NTA
agarose beads in PBS buffer complemented with 20 mM of imidazole. Beads were washed with 10
volumes of PBS complemented with 20 mM of imidazole and with 5 volumes of PBS complemented
with 40 mM of imidazole. His-tagged proteins were eluted with 5 volumes of PBS complemented with
0.5 M of imidazole. The buffer was exchanged to PBS (0.05 M phosphate buffer, 0.150 M NaCl) using
PD-10 or PD-MidiTrap G-25 desalting columns (GE Healthcare). Purity of the proteins was evaluated
using SDS-PAGE electrophoresis stained with Coomassie blue.
Physico-chemical measurements. Steady state UV-Vis absorption spectra were recorded using a
Cary 300 UV-Vis spectrometer (Agilent Technologies), equipped with a Versa20 Peltier-based
Fluorescence quantum yield measurements were determined as previously described using
FAST:HMBR as a reference23,47. Reciprocal dilution with protein solution was used so as to keep the
protein concentration constant at 40 μM while diluting the fluorogen solution. Absorption coefficients
were determined by forward titration of fluorogens into a 40 μM protein solution. Thermodynamic
152
dissociation constants were determined as previously described23 using a Spark 10M plate reader
(Tecan) and fitting data in Prism 6 to a one-site specific binding model.
Reactive Oxygen species quantification. Steady state UV-Vis absorption spectra were recorded using
a Cary 300 UV-Vis spectrometer (Agilent Technologies), equipped with a Versa20 Peltier-based
For ADPA photobleaching, HE photooxidation and photobleaching experiments, the samples were
irradiated with a current-controlled blue color Light Emitting Diodes LED (M470L4, Thorlabs, NJ)
(filtered at λ = 488 ± 20 nm) aligned in a 90° degree angle with respect to the photomultiplier detector
of the spectrofluorometer. We relied on the photoswitching kinetics of a photochemically well-
characterized RSFP (Dronpa2 at pH 7.4) for measuring light intensities of the LED light (as previously
described48). All experiments were performed in 3 mm2 ultra-micro cells quartz cuvettes to avoid any
effect from diffusion. For ADPA photobleaching measurements, all samples were incubated with 50
μM of ADPA and fluorescence was monitored at 374 nm excitation and 384-550 nm emission as a
function of the duration of 488 nm light exposure (90 mW.cm-2). For HE oxidation measurements, all
samples were incubated with 50 μM of HE and an increase of fluorescent products was monitored at
525 nm excitation and 600 – 750 nm emission as a function of the duration of 488nm light (90 mW.cm-
2
).
Experiments in mammalian cells
General. HeLa cells were cultured in Minimal Essential Media (MEM) supplemented with phenol red,
Glutamax I, 1 mM of sodium pyruvate, 1% (vol/vol) of non-essential amino-acids and 10% (vol/vol)
fetal calf serum (FCS), at 37 °C in a 5% CO2 atmosphere. For imaging, cells were seeded in µDish
IBIDI (Biovalley) coated with poly-L-lysine. Cells were transiently transfected using Genejuice (Merck)
according to the manufacturer’s protocol for 24 hours prior to imaging. Live cells were washed with
DPBS (Dulbecco’s Phosphate-Buffered Saline), and treated with DMEM media (without serum and
phenol red) containing the fluorogens and the ROSgens at the indicated concentration prior to
imaging.
Cloning. The plasmids allowing the mammalian expression of pFAST, or pFAST fused to the
mitochondrial targeting sequence from the subunit VIII of human cytochrome C oxidase (Mito-pFAST)
and to the inner plasma membrane targeting sequence lyn11 (Lyn11-pFAST) under the control of a
CMV promoter are described in Chapter III of the manuscript (see section III.2.6).
153
Cell viability assay. HeLa cells were treated with MEM media containing the appropriate ROSgen
(HBR-3,5DI, HBR-3,5DBr and HBR-3Br) at the indicated concentrations for the indicated times. The
cell viability was evaluated by fluorescence microscopy using the LIVE/DEAD® viability/cytotoxicity
assay kit (Molecular Probes, Life Technologies) following the manufacturer’s protocol.
Fluorescence microscopy. The confocal micrographs of mammalian cells were acquired on a Leica
TCS SP5 confocal laser scanning microscope equipped with a 63´/ 1.4 NA oil immersion objective.
Leica LAS AF software was used to collect the data. The images were analyzed with Fiji (Image J).
154
IV.3 References
1. Rodríguez-Pulido, A. et al. Assessing the potential of photosensitizing flavoproteins as tags

for correlative microscopy. Chem. Commun. 52, 8405–8408 (2016).
2. Baptista, M. S. et al. Type I and Type II Photosensitized Oxidation Reactions: Guidelines and
Mechanistic Pathways. Photochem. Photobiol. 93, 912–919 (2017).
3. Bevernaegie, R. et al. Exploring the Phototoxicity of Hypoxic Active Iridium(III)-Based
Sensitizers in 3D Tumor Spheroids. J. Am. Chem. Soc. 141, 18486–18491 (2019).
4. Liao, J. C., Roider, J. & Jay, D. G. Chromophore-assisted laser inactivation of proteins is
mediated by the photogeneration of free radicals. Proc. Natl. Acad. Sci. 91, 2659–2663 (1994).
5. Adams, S. R. et al. New Biarsenical Ligands and Tetracysteine Motifs for Protein Labeling in
Vitro and in Vivo: Synthesis and Biological Applications. J. Am. Chem. Soc. 124, 6063–6076 (2002).
6. Wojtovich, A. P. & Foster, T. H. Optogenetic control of ROS production. Redox Biol. 2, 368–
376 (2014).
7. Binns, T. C. et al. Rational Design of Bioavailable Photosensitizers for Manipulation and
Imaging of Biological Systems. Cell Chem. Biol. 27, 1063–1072.e7 (2020).
Methods 13, 263–268 (2016).
11. Greenbaum, L., Rothmann, C., Lavie, R. & Malik, Z. Green fluorescent protein photobleaching:
a model for protein damage by endogenous and exogenous singlet oxygen. Biol. Chem. 381, 1251–
1258 (2000).
12. Jiménez-Banzo, A., Nonell, S., Hofkens, J. & Flors, C. Singlet Oxygen Photosensitization by
EGFP and its Chromophore HBDI. Biophys. J. 94, 168–172 (2008).
13. Bulina, M. E. et al. A genetically encoded photosensitizer. Nat. Biotechnol. 24, 95–99 (2006).
14. Takemoto, K. et al. SuperNova, a monomeric photosensitizing fluorescent protein for
chromophore-assisted light inactivation. Sci. Rep. 3, 2629 (2013).
15. Torra, J. et al. Tailing miniSOG: structural bases of the complex photophysics of a flavin-
binding singlet oxygen photosensitizing protein. Sci. Rep. 9, 2428 (2019).
16. Westberg, M., Holmegaard, L., Pimenta, F. M., Etzerodt, M. & Ogilby, P. R. Rational Design
of an Efficient, Genetically Encodable, Protein-Encased Singlet Oxygen Photosensitizer. J. Am.
Chem. Soc. 137, 1632–1642 (2015).
17. Westberg, M., Bregnhøj, M., Etzerodt, M. & Ogilby, P. R. No Photon Wasted: An Efficient and
Selective Singlet Oxygen Photosensitizing Protein. J. Phys. Chem. B 121, 9366–9371 (2017).
18. Makhijani, K. et al. Precision Optogenetic Tool for Selective Single- and Multiple-Cell Ablation
in a Live Animal Model System. Cell Chem. Biol. 24, 110–119 (2017).
19. Ren, K. et al. A Genetically Encoded RNA Photosensitizer for Targeted Cell Regulation.
Angew. Chem. 132, 22170–22174 (2020).
20. Schieber, M. & Chandel, N. S. ROS Function in Redox Signaling and Oxidative Stress. Curr.
Biol. 24, R453–R462 (2014).
21. Bulina, M. E. et al. Chromophore-assisted light inactivation (CALI) using the phototoxic
fluorescent protein KillerRed. Nat. Protoc. 1, 947–953 (2006).
22. Riani, Y. D., Matsuda, T., Takemoto, K. & Nagai, T. Green monomeric photosensitizing
fluorescent protein for photo-inducible protein inactivation and cell ablation. BMC Biol. 16, 50 (2018).
1828 (2018).
155
27. Benaissa, H. et al. Engineering of a fluorescent chemogenetic reporter with tunable color for
advanced live-cell imaging. bioRxiv 2021.01.29.428635 (2021).
28. Koziar, J. C. & Cowan, D. O. Photochemical heavy-atom effects. Acc. Chem. Res. 11, 334–
341 (1978).
Markers for Live Cell Microscopy. Biochemistry (Mosc.) 57, 5648–5653 (2018).
30. Hoebeke, M. & Damoiseau, X. Determination of the singlet oxygen quantum yield of
bacteriochlorin a: a comparative study in phosphate buffer and aqueous dispersion of dimiristoyl-L-
alpha-phosphatidylcholine liposomes. Photochem. Photobiol. Sci. Off. J. Eur. Photochem. Assoc. Eur.
Soc. Photobiol. 1, 283–287 (2002).
31. Lindig, B. A., Rodgers, M. A. J. & Schaap, A. P. Determination of the lifetime of singlet oxygen
in water-d2 using 9,10-anthracenedipropionic acid, a water-soluble probe. J. Am. Chem. Soc. 102,
5590–5593 (1980).
32. Kotani, H., Ohkubo, K. & Fukuzumi, S. Photocatalytic Oxygenation of Anthracenes and Olefins
with Dioxygen via Selective Radical Coupling Using 9-Mesityl-10-methylacridinium Ion as an Effective
Electron-Transfer Photocatalyst. J. Am. Chem. Soc. 126, 15999–16006 (2004).
33. Pimenta, F. M., Jensen, R. L., Breitenbach, T., Etzerodt, M. & Ogilby, P. R. Oxygen-Dependent
Photochemistry and Photophysics of “MiniSOG,” a Protein-Encased Flavin. Photochem. Photobiol.
89, 1116–1126 (2013).
34. Ruiz-González, R. et al. Singlet Oxygen Generation by the Genetically Encoded Tag miniSOG.
J. Am. Chem. Soc. 135, 9564–9567 (2013).
35. Riega, S. D. E., Rodríguez, H. B. & Román, E. S. Rose bengal in poly(2-hydroxyethyl
methacrylate) thin films: self-quenching by photoactive energy traps. Methods Appl. Fluoresc. 5,
014010 (2017).
36. Craig, R. A. et al. Quantification of singlet oxygen generation from photodynamic hydrogels.
React. Funct. Polym. 87, 1–6 (2015).
37. Gomes, A., Fernandes, E. & Lima, J. L. F. C. Fluorescence probes used for detection of
reactive oxygen species. J. Biochem. Biophys. Methods 65, 45–80 (2005).
38. Bilski, P. J., Karriker, B. & Chignell, C. F. Quenching and generation of singlet oxygen by
hydroethidine and related chromophores. Chem. Phys. Lett. 475, 116–119 (2009).
39. Simon, H.-U., Haj-Yehia, A. & Levi-Schaffer, F. Role of reactive oxygen species (ROS) in
apoptosis induction. Apoptosis 5, 415–418 (2000).
40. Liu, X. & Hajnóczky, G. Altered fusion dynamics underlie unique morphological changes in
mitochondria during hypoxia–reoxygenation stress. Cell Death Differ. 18, 1561–1572 (2011).
41. Miyazono, Y. et al. Uncoupled mitochondria quickly shorten along their long axis to form
indented spheroids, instead of rings, in a fission-independent manner. Sci. Rep. 8, 350 (2018).
42. Britton, H. T. S. & Robinson, R. A. CXCVIII.—Universal buffer solutions and the dissociation
constant of veronal. J. Chem. Soc. Resumed 1456–1462 (1931).
43. Cetenko, W. A., Connor, D. T., Sorenson, R. J., Unangst, P. C. & Stabler, S. S. Known and
selected novel arylmethylenyl derivatives of thiazolidinones, imidazolidinones and oxazolidinones
useful as antiallergy agents and anti-inflammatory agents. Patent US5208250 (May 4, 1993).
44. Harada, K. et al. Discovery of potent and orally bioavailable 17β-hydroxysteroid
dehydrogenase type 3 inhibitors. Bioorg. Med. Chem. 20, 3242–3254 (2012).
45. Geldenhuys, W. J. et al. Identification of small molecules that bind to the mitochondrial protein
mitoNEET. Bioorg. Med. Chem. Lett. 26, 5350–5353 (2016).
46. Song, Y. et al. Synthesis, Structure−Activity Relationships, and in Vivo Evaluations of
Substituted Di-tert-butylphenols as a Novel Class of Potent, Selective, and Orally Active
Cyclooxygenase-2 Inhibitors. 1. Thiazolone and Oxazolone Series. J. Med. Chem. 42, 1151–1160
(1999).
Last. Sci. Rep. 7, (2017).
48. Adelizzi, B., Gielen, V., Le Saux, T., Dedecker, P. & Jullien, L. Quantitative Model for
Reversibly Photoswitchable Sensors. ACS Sens. (2021).
156
Chapter V
General discussion
Fluorescent chemical-genetic hybrids are undoubtedly powerful tools for bioimaging and optogenetic.
Thanks to their bipartite nature, chemogenetic reporters expanded the fluorescent toolbox and
provided new ways to target, localize and study biological events in live cells with high spatial and
temporal resolution.
My main Ph.D. project has consisted of providing investigators with a general fluorescent chemical-
genetic tag applicable to a broad variety of biological applications, allowing selective labeling and
imaging of proteins in live cells, neurons and in a multicellular organism. We applied a coordinated
molecular and protein engineering strategy to develop a promiscuous and multifunctional protein tag
pFAST, able to recognize a palette of fluorogenic chromophores that cover the entire visible spectrum.
By employing molecular chemistry, minimal substitutions of few heteroatoms of the initial molecules
allowed to fine-tuned their electronic and optical properties. Furthermore, the use of directed evolution
strategy allowed us to engineer a single protein tag with suitable properties for the entire collection of
fluorogenic chromophores. In the frame of this work, we expanded the initial features of the
prototypical FAST protein by gaining improved functions by artificial selection. By combining
molecular biology techniques and high throughput screening, we were able to select for improved and
optimized variants on a relatively short laboratory time scale.
Recent protein engineering strategies allowed to achieve even more promising performances by the
use of in silico techniques and de novo design1,2. I am convinced that exploring the full-length
sequences and functions of proteins by computational design would allow for the generation of further
improved scaffolds for bioimaging applications. In this study, we analyzed by homology modeling how
pFAST interacts with the collection of fluorogens, giving us hints for further optimization of the protein
scaffold, which could reduce the timescale of further artificial selections. Our model allowed us to
explain how the removal of the N-terminal domain of pFAST reduced the solubility of the resulting
protein nano-pFAST, by exposing hydrophobic residues important for stabilizing the overall pFAST
157
CHAPTER V: GENERAL DISCUSSION
structure. Our model provided us with information for further optimization by targeting specific
residues predicted to destabilize the overall protein.
With the aim of providing a single chemogenetic reporter applicable to a variety of biological samples
and imaging techniques, the versatility of our engineered pFAST-based fluorescent reporter opens
great prospects for a diversity of bioimaging scenarios.
V.1. Development of pFAST-based biosensors
Beyond offering a panoply of opportunities in biological imaging, fluorogenic chemogenetic systems

can be used to investigate and study biochemical events with high spatiotemporal resolution. As FAST
was recently designed for sensing intracellular analytes3 and allowed the study of protein-protein
interactions by the design of split-FAST4, the superior performances of pFAST could enable the
design of improved biosensors for sensing specific biochemical and intracellular events in cells. In
particular, we could take advantage of its spectral tunability for the design of efficient Förster
resonance energy transfer (FRET)-based biosensors in combination with classical FPs5 (Figure
V.1a). In addition, the possibility to form dark complexes absorbing in the green region opens new
prospects for the design of FRET experiments based on fluorescence lifetime imaging microscopy
(FLIM) without blocking an imaging channel, which could be highly useful in multicolor FRET-pairs
with FLIM6 (Figure V.1b).
158
a low b In FLIM-FRET
high
1 FRET FRET mTurquoise2 pFAST: HBR-3,5DOM low high
Normalized Absorbance
Normalized Fluorescence
FRET FRET
D A 1.0 1.0
D A
CFP pFAST CFP pFAST

analyte A CFP pFAST CFP pFAST
0.5 0.5
+ analyte
+
X Y
– X Y
X Y
0.0 0.0 – X Y
400 500 600 700
low high
2 FRET FRET pFAST: HBP-3M sYFP2 mTurquoise2 pFAST:HBIR-3M
D A 1.0 1.0
1.0 1.0
pFAST YFP pFAST YFP

analyte B 0.5 0.5
0.5 0.5
+
Z W
– Z W
0.0 0.0 0.0 0.0
400 500 600 350 400 450 500 550 600
low high Wavelength (nm)
3 FRET FRET pFAST: HBP-3,5DOM mCherry
D A fluorescence lifetime of the donor
fluorescence intensity (a.u.)

1.0 1.0
100
pFAST RFP pFAST RFP

analyte C
0.5 0.5
+ 50 low
FRET
V T
– V T high
0.0 0.0 FRET
400 500 600 700 0
Wavelength (nm) 0 20 40 60 time
80 100 120 140
Figure V.1. Examples of compatible FRET-pairs for the design of biosensors based on pFAST. (a) 1)
CFP: cyan fluorescent protein (FRET donor, D) and pFAST:HBR-3,5DOM (FRET acceptor, A); 2) pFAST:HBP-
3M (FRET donor, D) and YFP: Yellow fluorescent protein (FRET acceptor, A) and 3) pFAST:HBP-3,5DOM
(FRET donor, D) and RFP: Red fluorescent protein (FRET acceptor, A). (b) Theoretical FLIM-FRET analysis of
CFP (FRET donor, D) – pFAST (FRET acceptor, A) FRET-pair in presence of the dark chromophore HBIR-3M
(adapted from Broch & Gautier7). (a,b) Normalized absorption (dashed lines) and fluorescence (solid lines)
spectra of the different fluorogenic chromophores bound to pFAST and the different FPs. The graphs show the
spectral overlap between the donor and the acceptor.
V.2. Development of pFAST-based fluorescent reporters for super-resolution

microscopy
In this study, we could also demonstrated that pFAST was suitable to label proteins in live cells and
in live dissociated hippocampal neurons with advanced confocal and super-resolution microscopy.
Benefiting from its high photostability, pFAST allowed to resolve details below the diffraction limit
within live cells and live neurons using Stimulated Emission Depletion (STED) nanoscopy in
combination with HBR-3,5DOM fluorogenic chromophore. Furthermore, we noticed that among our
extended palette of chromophores, some of them exhibited good photostability although a rapid
transient fluorescence decay was observed before reaching steady state. The ability to transiently
switch from fluorescent to dark states, presumably by photoswitching and/or photoejection of the
chromophores could be ingeniously used for other advanced imaging techniques. This singular
fluorescence signature could allow selective imaging exploiting the photokinetic dynamics of
159
fluorescent probes8,9. Finally, we could take advantage of their on-off switching fluorescent states by
the use of stochastic photomodulation of individual fluorescent molecules in combination with single-
molecule localization microscopy10.
V.3. Development of pFAST-based chemogenetic photosensitizers
In the frame of this work, we demonstrated that functionalizing HBR derivatives with heavy atoms
such as bromine and iodine allowed to enhance the generation of reactive oxygen species. Such
‘ROSgenic’ chromophores are only ‘activated’ when they are bound to FAST variants, ensuring on-
demand targeted ROS generation. We demonstrated that pFAST conjugated to a 3,5-dibrominated
HBR derivative (HBR-3,5DBr) could act as a fluorescent reporter and a photosensitizer able to
photoinduce cellular dysfunction. Both in vitro and in vivo analysis showed that this chemogenetic
photosensitizer produced lower levels of ROS, relative to chemical (Rose Bengal)11 or genetically-
encoded (miniSOG)12 photosensitizers, thus would probably disqualify it for cell photoablation.
However, the ability to target the expression of pFAST by fusing it to essential biomolecules (such as
the subunit VIII of human cytochrome C oxidase) allowed to photoinduce mitochondrial disruption in
a highly specific manner. Such milder ROS generator could be used advantageously to induce minor
cellular dysfunctions resulting in modified activities, whereas other photosensitizers were reported to
irreversibly photoinduce cellular death. We could also take advantage of pFAST tunability to modulate
its ROS generation by reversibly ‘turning-off’ the photoinduction reaction upon the addition of
fluorogenic chromophores. In addition, the possibility to switch from a fluorogenic to ROSgenic
chromophores can facilitate dual observation and photoinduction strategy, for the manipulation of
sensitive living systems such as multicellular organisms. Finally, we could take benefit from the
bipartite nature of our system to enhance the generation of ROS by synthesizing new ROSgenic
chromophores based on the extended palette of chromophores that we developed and/or by applying
protein engineering strategies.
160
V.4. References

(2018).
2. Bozhanova, N. G., Baranov, M. S., Baleeva, N. S., Gavrikov, A. S. & Mishin, A. S. Red-Shifted
Aminated Derivatives of GFP Chromophore for Live-Cell Protein Labeling with Lipocalins. Int. J. Mol.
Sci. 19, 3778 (2018).
5. Bajar, B. T., Wang, E. S., Zhang, S., Lin, M. Z. & Chu, J. A Guide to Fluorescent Protein FRET
Pairs. Sensors 16, 1488 (2016).
6. Bertolin, G. et al. Optimized FRET Pairs and Quantification Approaches To Detect the
Activation of Aurora Kinase A at Mitosis. ACS Sens. 4, 2018–2027 (2019).
7. Broch, F. & Gautier, A. Illuminating Cellular Biochemistry: Fluorogenic Chemogenetic
Biosensors for Biological Imaging. ChemPlusChem 85, 1487–1497 (2020).
014010 (2017).
161
162
Digest :
Ingénierie d’outils chémogénétiques pour
l’imagerie et la manipulation de systèmes
biologiques
Thèse de doctorat de l’université Paris Sciences et Lettres (Université PSL) préparée à l’École
Normale Supérieure et Sorbonne Université sous la direction du Professeur Arnaud Gautier.
1. Introduction générale (Chapitre I)

1.1. Mise en contexte
Les organismes vivants sont orchestrés par un ensemble d’événements et de processus cellulaires
dynamiques. Dans le but de les observer et de les étudier, plusieurs types de modalités d’imagerie
existent. En particulier, la microscopie de fluorescence est devenue une technique de référence en
imagerie si l’on désire observer spécifiquement une molécule biologique dans un échantillon donné.
La fluorescence est caractérisée spécifiquement par l’émission de lumière par une substance ayant
absorbé de la lumière. En couplant une molécule cible à une molécule fluorescente, jouant alors le
rôle de sonde moléculaire, celle-ci peut être spécifiquement marquée au sein d’une cellule ou d’un
organisme. Le signal de fluorescence donne alors des informations sur la distribution, la localisation
et la quantité de la molécule au sein de la cellule. Ainsi, la microscopie de fluorescence permet
d’imager avec une haute sensibilité et une grande résolution spatio-temporelle. Plusieurs types de
microscopes à fluorescence existent et permettent d’acquérir des informations de plus en plus fines
et sensibles en fonction des réglages et dispositifs utilisés. En plus d’importants progrès liés à
l’amélioration des composants optiques, le développement de sondes fluorescentes efficaces et
robustes joue un rôle primordial si l’on désire détecter de façon efficace une molécule cible au sein
de la cellule. La plupart des marqueurs fluorescents existants sont génétiquement encodés (cas des
protéines fluorescentes), synthétiques (fluorophores) ou semi-synthétiques.
163
DIGEST : INGÉNIERIE D’OUTILS CHEMOGENETIQUES POUR L’IMAGERIE ET LA MANIPULATION DE SYSTEMES BIOLOGIQUES
La découverte de la protéine fluorescente verte issue de la méduse bioluminescente Aequorea

victoria, (GFP pour Green Fluorescent Protein) a joué sans conteste un rôle très important dans le
développement des marqueurs fluorescents1. La possibilité de fusionner génétiquement le gène
exprimant la GFP à n’importe quel gène d’intérêt assure une spécificité de marquage absolue : seule
la protéine fusionnée à la GFP fluoresce. Depuis cette découverte, l’étude d’autres organismes
marins et l’ingénierie protéique ont permis le développement d’une collection de protéines
fluorescentes bleues, cyans, vertes, jaunes, oranges et rouges, permettant de marquer plusieurs
cibles dans un même échantillon et les distinguer grâce à leurs propriétés spectrales distinctes2–6.
Néanmoins la plupart de ces protéines fluorescentes ont des limitations communes comme par
exemple leur absence d’activité en absence d’oxygène, leur maturation lente ou encore leur taille (25-
30 kDa) et leur tendance naturelle à oligomériser, qui peut dans certains cas conduire à des protéines
fusionnées dysfonctionnelles. Ainsi l’utilisation de fluorophores chimiques peut dans certains cas être
plus adaptée. Leur performance de fluorescence, leur large versatilité et leur taille avantageuse
représentent des avantages notables par rapport aux protéines fluorescentes. Bien que les premières
expériences utilisant des fluorophores synthétiques nécessitaient la perméabilisation ou la fixation
des cellules, des progrès par synthèse chimique ont permis de développer un grand nombre de
molécules fluorescentes compatibles avec les systèmes biologiques vivants7. De plus, leurs
propriétés spectrales peuvent être facilement modulées par synthèse chimique, ainsi il existe une
large palette de fluorophores couvrant la totalité du spectre visible8,9. Néanmoins leur manque de
spécificité a poussé les biologistes et les chimistes à développer des systèmes semi-synthétiques
hybrides, qui comme leur nom l’indique, sont composés d’un tag génétiquement encodé (protéine ou
ARN) et d’un chromophore synthétique. Afin d’éviter tout signal de fluorescence non-spécifique, des
chromophores fluorogéniques qui ne fluorescent que lorsqu’ils sont associés au tag peuvent être
utilisés.
Récemment, ce principe générique a permis de développer un petit tag protéique de 14kDa appelé
FAST (pour Fluorescence Activating and absorption Shifting Tag) au sein de notre laboratoire10.
Initialement, FAST fluoresce dans le canal jaune-orange-rouge en s’associant spécifiquement à des
dérivés fluorogéniques du 4-hydroxybenzylidène rhodanine (HBR)10,11. Leur fluorogénicité provient
de leur structure « push-pull » composée d’un phénol donneur d’électrons conjugué à un hétérocycle
rhodanine électroattracteur. En solution, ces fluorogènes dissipent la lumière absorbée de manière
non radiative par rotation interne ou isomérisation cis-trans. L’immobilisation du fluorogène dans la
cavité de FAST augmente le rendement quantique de fluorescence permettant ainsi une forte
exaltation de fluorescence.
164
Figure 1. Le marquage sélectif de la protéine FAST couplée aux dérivés HBR.
FAST a été développé par évolution dirigée à partir de la protéine PYP (pour Photoactive Yellow
Protein), photorécepteur issu de la bactérie photosynthétique, Halorhodospira halophila. En plus de
sa petite taille, PYP a été choisie comme cible car elle lie de façon covalente l’acide para-
hydroxycinamique dont la structure moléculaire est proche de celle des dérivés d’HBR. Dans le but
de développer FAST à partir de PYP, des techniques d’ingénierie protéique par évolution dirigée ont
été utilisées. Le principe de l’évolution dirigée repose sur le criblage et la sélection de protéines
d’intérêts à partir d’un grand nombre de variants protéiques. Pour cela, une large bibliothèque de
mutants de la protéine initiale est construite par des techniques de biologie moléculaire puis une
stratégie de criblage et de sélection est adaptée en fonction des propriétés recherchées. Dans le cas
de FAST, la fonction recherchée étant la reconnaissance et l’activation de la fluorescence des
molécules HBR, une stratégie couplant la présentation des variants à la surface de levures (yeast-
surface display) à la sélection des variants les plus fluorescents (après incubation avec les
fluorogènes) par cytométrie en flux a été utilisée. De cette façon, FAST a été positivement isolé à
partir d’une banque de variants de l’apo-PYP-C69G générée par mutagénèse à saturation, illustrant
le pouvoir de l’évolution dirigée10. De nombreuses applications utilisant le système de marquage avec
FAST ont depuis vu le jour permettant d’imager des protéines cibles dans divers organismes comme
les bactéries (dont des bactéries anaérobies), les levures, les cellules mammifères ou encore des
organismes multicellulaires tels que l’embryon de poulet ou de poisson-zèbre10,12–19. Grâce à la nature
hybride du système, des techniques d’ingénierie moléculaire20, d’ingénierie protéique21 ou les deux12
ont récemment permis de développer des nouveaux systèmes basés sur FAST.
1.2. Objectifs de la thèse
Afin de fournir aux biologistes une méthode généralisable, multifonctionnelle et robuste à partir du
système de marquage basé sur FAST, nous avons dans un premier temps développé un nouveau
variant de FAST qui est capable de lier un grand nombre de molécules fluorogéniques couvrant la
totalité du spectre visible pour de multiples applications en imagerie cellulaire. En collaboration avec
165
l’équipe du Prof. Ludovic Jullien du Laboratoire PASTEUR à l’École Normale Supérieure, nous avons
dans un premier temps développé des analogues d’HBR. En remplaçant l’hétérocycle
électroattracteur rhodanine (R) par l’isorhodanine (IR), la pseudothiohydantoine (P), la 2,4-
thiazolidinedione (T) et la 2,4-oxazolidinedione (O) et en ajustant la partie donneuse d’électrons par
diverses substitutions sur le phénolate, nous avons identifié une palette de nouvelles molécules aux
propriétés spectrales variées. Par la suite, des efforts d’ingénierie protéique couplant le design
rationnel à l’évolution dirigée nous ont permis d’identifier un nouveau variant de FAST, nommé
pFAST. En s’associant à toute la collection de molécules fluorogéniques, pFAST permet de générer
des complexes bleu, cyan, vert, jaune, orange et rouge. Tout comme FAST et les autres mutants
développés par ingénierie protéique10,12,21, pFAST a été développé par la stratégie de « yeast-
display » et les variants d’intérêt ont été triés par cytométrie en flux. Nous avons découvert
qu’associés aux dérivés d’isorhodanines, les complexes formés étaient non fluorescents. Cette
propriété singulière a été ingénieusement utilisée afin d’éteindre très rapidement la fluorescence de
pFAST dans des systèmes vivants. En plus de s’associer aux molécules initialement synthétisées
(HBR) avec de meilleurs affinités et de meilleurs performances de fluorescence, nous avons
découvert que pFAST s’associait également aux dérivés rhodanine non-perméants de type (HBRAA)
avec des propriétés surpassant celles initialement évaluées avec FAST20 (Figure 2a). Grâce à la
grande variété de propriétés spectrales générées, aux grandes performances et à la photostabilité
de certaines des molécules fluorogéniques associées à pFAST, nous avons finalement montré que
ce petit tag protéique était adapté aux techniques d’imagerie les plus avancées en combinaison aux
techniques super-résolutives de type STED (Stimulated Emission Depletion).
En plus d’avoir des propriétés intéressantes pour l’imagerie cellulaire, les systèmes fluorogéniques
peuvent être utilisé pour étudier des évènements biologiques dynamiques et spécifiques au sein de
la cellule par le biais de biosenseurs et d’actuateurs. Des biosenseurs basés sur le système FAST
ont été récemment développés au sein du laboratoire22,23. Dans le but d’élargir le champ des
possibles, nous nous sommes demandé si FAST pouvait être, de la même façon, réarrangé pour le
développement de photosensibilisateurs fluorogéniques. En produisant des espèces réactives de
l’oxygène (ou ROS), de tels systèmes permettent d’étudier ingénieusement certains processus
biologiques. En effet, les espèces réactives de l’oxygène, connues pour endommager les cellules,
peuvent être détournées pour : (i) induire la mort cellulaire, (ii) inactiver spécifiquement une protéine
cible et/ou (iii) localiser une protéine cible par microscopie corrélative (CLEM pour correlative-light
and electron microscopy). En collaboration avec l’équipe du Prof. Ludovic Jullien, des dérivés d’HBR
possédant des substitutions par des atomes lourds ont été synthétisés. En effet, de telles substitutions
sont connues pour augmenter la conversion intersystème et ainsi favoriser la génération de ROS en
présence d’oxygène (Figure 2b). In vitro, il a été mis en évidence qu’associés à FAST ou pFAST,
ces dérivés étaient capables de produire certains types d’espèces réactives de l’oxygène. Par la suite,
166
des expériences en cellules mammifères ont permis de mettre en évidence la perturbation ciblée des
mitochondries.
1. 1.
HBR-3,5DOM 1.
or anc
or anc
dérivés HBO dérivés HBT dérivés
HMBR HBP
xcitati
FAST FAST O FAST O
O
NATURE CHEMI NATURE

NH CAL BI.OLOGYCHEMICAL.BIOLOGY
ormalize a
ormalize a
NH . NH
rmaliz
S
S
O S
S
HO
S
HO
O HO
O
b
O R
R O NH
S
ARTICLES ARTICLES
R O
S S
MICALBIOLOGY NATURE CHEMI
O
CAL BIOLOGY tag protéique HBR derivatives O
HBR derivatives HBR derivatives
FAST pFAST fluorogénique 1. HBR-3,5DOM
1. 1. 1H
or anc
or anc
R1 dérivés HBRAA
R1
dérivés HBR 6 dérivés
7
HMBR HBIR
xcitati
NH HBR-3,5DOM FAST O
NH
FAST O FAST S
O
avelengt
1. 1. HBR-3,5DOM HBR-3,5DOM HBR-3,5DOM avelength (nm)
1.
O 1.
S
1. 1.
or anc
ormalize a or anc
HO HMBR 450 HMBR HO HMBR HMBR
xcitati
xcitati
greenFAST redFA T
ormalize a
greenFAST redFA T
ormalize a
FAST FAST greenFAST redFA T O
ingénierieO
O Z
O . NH
. .
R2 HO
2 NH NH
greenFAST
1.
rmaliz
1. 1. S
or anc
ormalize a
S
or anc
moléculaire
. .NH . . HO R = Me, . HO
O. COO HO -
O redFAST
NH
SHBO-3M (R= Me, R = H)
W
HBT-3M R(R 1 O R
S 2= H)
rmaliz
rmaliz
R M
xcitati
1 2 Y
S M M R S O
S
HBT-3,5DM 1= Me, R2=HHMe)HBR derivatives
(R
O O
R HBO-3,5DM (R = Me, R2= Me) R HBR derivatives HH HBR derivatives H
S
1 X
S H
HBR derivatives & évolution HBR derivatives
HBT-3,5DOM (R
Normalize a
Normalize a
500 . . . 6
dirigée 1,10 6 1, 0 7
1= OMe,
dérivés HBR R2= OMe) 0
6 7 0 0 1,006 1,10 7 1, 0 0 1,00
N rmaliz
M M
avelength (nm) avelength (nm) M
avelength(nm) avelength avelength (nm)
(nm) Mortavelength (n
cellulaire
actuateur avec des atomes lourds HMBR HBR-3,5DOM
redFA T greenFAST O redFA T HMBR HMBR
greenFAST redFA T
fluorogénique
greenFAST greenFAST
greenFAST redFA T O S 1 redFA T ISC
R1 1. 1. 1. 1. greenFAST greenFAST 1. 1. greenFAST
ROS Inactivation
absorption
or anc
Normalize a or anc
redFAST Y redFAST
1. greenFAST
1. 1. 1
NH
redFAST redFAST
T1
or anc
protéique
or anc
xcitati
xcitati
M M 6 7
redFAST
S NH M M
xcitati
H H H M
HO S Wavelength
H 3 (nm) signalisation
Wavelengt
550 HO
Normalize a
. NH . . . H . H H . H O2 H
R2 S par des ROS
N rmaliz
N rmaliz
Normalize a
Normalize a
M M
X S0 . . .
HBR-3,5DOM HMBR HBR-3,5DOM EL
EL
N rmaliz
HBP-3M (R1= Me, R2= H) HMBR (X
HBR-3Br M = Me, R = H)
(X
M M
1 = Br, Y 2 =
HBP-3,5DM (R 1
= Me, R2= Me) 6 7 HMBR 0 0 HMBR
1,006 HBR-3,5DOM
HBR-3,5DM
1,10 7 1, 0
(R
0 0 HMBR
= Me, R = Me) HBR-3,5DOM
1,00 1,10 1, 0 HBR-3,5DOM
HBP-3,5DOM (R = OMe, R2= OMe)
Wavelength (nm) HBR-3,5DBr
Wavelength
Wavelength(nm)
(nm) (X1 = Br, Y2 = Br) Wavelength (nm)
greenFAST redFAST
ence 1 600 HBR-3,5DOM (R 1
= OMe, R 2
= OMe)
HBR-3,5DI (X = I, Y = I) 6 7 6
EL EL
Wavelength (nm) Wavelength (n
HBR-3,5D M HBR-3,5D M
Figure 2. Stratégie coordonnée pour
l emle développement de pFAST, (a) un X tag protéique fluorogénique
greenFAST redFAST greenFAST redFAST
(nm) X X X
EL EL
multicouleur et polyvalent et (b)
HBR-3,5D M un actuateur fluorogénique
HBR-3,5D M pour la génération d’espèces réactives de l’oxygène
X X greenFAST redFAST
(ROS).
2. Ingénierie de rapporteurs chémogénétiques hybrides (Chapitre II)
De nombreux autres systèmes hybrides tel que FAST existent et complètent la grande famille de
marqueurs fluorescents. En employant des techniques avancées en ingénierie moléculaire et/ou
protéique, le développement de systèmes fluorogéniques hybrides (ou chémogénétiques) est très
prometteur en imagerie cellulaire et en optogénétique. Ces systèmes hybrides combinent la
spécificité de marquage des protéines fluorescentes à la large flexibilité des fluorophores
synthétiques. Ce chapitre bibliographique basé sur la minirevue que nous avons publiée dans le
European Journal of Organic Chemistry « Engineering glowing chemogenetic hybrids for spying on
cells » porte sur l’ingénierie récente de tels systèmes pour l’imagerie biologique. La possibilité de
modifier le chromophore (naturel ou synthétique) - par ingénierie moléculaire – et/ou de modifier le
tag (protéique ou d’ARN) – par ingénierie protéique – donne aux chimistes et aux biologistes une
grande diversité de scénarios possibles pour développer, ajuster ou optimiser la large collection de
systèmes fluorogéniques.
À ce jour, seul UnaG, issu de l’anguille japonaise Unagi, possède naturellement des propriétés
fluorogéniques24. La plupart des systèmes exposés dans ce chapitre ont été conçus ou modifiés par
167
ingénierie. De nombreux exemples de systèmes fluorogéniques basés sur la reconnaissance de

chromophores naturels ou semi-synthétiques ont été développés par des techniques d’ingénierie
protéique e.g. design rationnel, évolution dirigée ou design de novo. Le cas des rapporteurs
fluorescents issus des photorécepteurs naturels illustre assez bien le pouvoir du design rationnel. De
tels photorécepteurs portent des chromophores prosthétiques e.g. flavine, biline ou rétinal qui sont
connus pour initier des réponses photochimiques25–27. Des informations structurales sur ces
photorécepteurs ont permis de cibler et modifier par mutagénèse dirigée (mais également par
évolution dirigée) la partie protéique permettant ainsi d’inhiber le cycle photochimique naturel de ces
chromophores et ainsi activer leur fluorescence. Lorsque le design rationnel ne suffit pas, des
techniques d’évolution dirigée peuvent être employées, comme dans le cas de FAST ou des
systèmes FAPs28–30 (Fluorogen-Activating Proteins). Ces techniques imitent le processus de sélection
naturelle à l’échelle du laboratoire. Finalement, des systèmes fluorogéniques issus de modélisation
in silico31 ou entièrement concues de novo ont été récemment développés. Le système mFAP (pour
mini-Fluorogen Activating Protein), par exemple, a été entièrement conçu par design computationnel
à partir de la structure de la GFP32. Il s’agit à ce jour du plus petit tag protéique reconnaissant le
chromophore naturellement présent dans la GFP. Diminuant considérablement le long processus des
techniques classiques d’ingénierie protéique et en ciblant la totalité de la séquence protéique, les
techniques in silico et de novo n’en sont qu’à leurs débuts dans le développement de rapporteurs et
de biosenseurs robustes et efficaces.
D’un autre côté, bénéficiant de la nature en deux parties des marqueurs fluorogéniques, des
techniques de chimie organique moderne peuvent être employées pour développer, ajuster ou
optimiser certains des systèmes hybrides. En utilisant l’ingénierie moléculaire, des chromophores
fluorogéniques synthétiques ont été exclusivement développés pour s’hybrider à des tags protéiques
appelés « self-labeling tags » comme les systèmes SNAP, CLIP, PYP ou Halo tags33–36. Ces
systèmes de marquage reconnaissent spécifiquement le chromophore synthétique qui émet alors un
signal de fluorescence. Le cas des dérivés de la rhodamine illustre assez bien le pouvoir de
l’ingénierie moléculaire. Ces composés adoptent une forme zwitterionique fluorescente dans des
environnements moins polaires, comme dans le cas des cavités protéiques, qui sont autrement non-
fluorescents en solution37. Des tags issus d’aptamères d’ARN ont également été récemment
développés pour reconnaître de tels chromophores synthétiques38. Il est également possible d’ajuster
ou modifier certaines des propriétés des fluorogènes par ingénierie moléculaire. Par synthèse
chimique, des dérivés fluorogéniques incapables de traverser la membrane cellulaire ont été
développés pour certaines applications biologiques20,39. De la même façon, des modifications
structurales ont permis d’adapter assez simplement les propriétés spectrales de certains des
fluorogènes. Cette propriété unique aux systèmes hybrides a permis de développer une grande
168
palette de marqueurs fluorogéniques couvrant la totalité du spectre visible et du proche infrarouge

pour l’imagerie multicouleur11,40,41.
Le développement de marqueurs fluorogéniques pour l’imagerie ouvre également de nouvelles

perspectives pour le développement de biosenseurs et d’actuateurs fluorogéniques pour l’étude de
processus biologiques au sein de la cellule. Une nouvelle fois, la possibilité de modifier le tag
protéique (ou d’ARN), de modifier le chromophore par synthèse chimique ou les deux ont donné aux
chimistes et aux biologistes une grande diversité de scénarios possibles pour faire des marqueurs
fluorogéniques des sondes hybrides capables de détecter un grand nombre de processus biologiques
et biochimiques dans la cellule. Bénéficiant d’un grand nombre d’avantages par rapport aux protéines
fluorescentes et aux marqueurs entièrement synthétiques, le développement et l’ingénierie des
systèmes fluorogéniques hybrides ouvrent de nouvelles possibilités aux utilisateurs afin d’imager ou
de suivre des évènements biochimiques au sein de systèmes vivants. Combinant à la fois leur grande
diversité de fonction par ingénierie moléculaire à leur grande modularité par ingénierie protéique, il
n’y a aucun doute que le développement de rapporteurs hybrides dans le domaine des biosenseurs,
de l’imagerie multiplexée et de l’imagerie haute-résolution n’en est qu’à ses débuts.
3. Développement d’un marqueur chémogénétique multifonctionnel (Chapitre III)
Ce chapitre présente le développement d’un tag protéique multifonctionnel pFAST conçu par
évolution dirigée. Ce petit tag protéique unique permet de reconnaître un grand nombre de molécules
fluorogéniques permettant de former des complexes fluorescents couvrant la totalité du spectre
visible. Grâce à la grande diversité de couleurs générées, ce système est adaptable à un grand
nombre de scénarios expérimentaux permettant aussi bien d’imager des cellules vivantes que des
organismes multicellulaires par microscopie de fluorescence traditionnelle et de super-résolution.
Afin de développer un tag protéique multifonctionnel, le défi a été de trouver un ensemble de

fluorogènes aux propriétés spectrales suffisamment différentes pour générer une grande palette de
couleur tout en gardant la spécificité de reconnaissance par un même tag protéique. Ainsi, des
molécules analogues aux molécules initialement conçues pour FAST (série des HBR) ont été
développées au sein du laboratoire. Les molécules HBR étant des molécules « push-pull »
composées d’une partie phénol donneuse d’électrons et d’un hétérocycle rhodanine électroattracteur,
nous avons généré une collection de fluorogènes analogues en ajustant leurs propriétés
électroniques (et donc spectrales) par modification de la partie électroattractrice. Cette première série
d’expériences nous a permis d’obtenir une large collection de fluorogènes couvrant la totalité du
spectre avec le minimum de modifications structurales (Figure 3).
169
promiscuous O Y O Me OHBP-3,5DM
HBIR 3,5DM NH 3,5DOM
HBIR OMe
a ; (mM 1.cmrecognition
1 R HBT-3,5DM
) ; K2D (µM)] [
HBO series HBT series
2
HO
HBP-3M
O R HBIR-3M
HBT-3,5DOM
HBIR-3M HBP X series HBIR-3,5DM
Me HBRMe seriesO MeO
O HBIR-3,5DOM
HBRAA Xseries
HBP-3,5DOM
O
HBIR series
NH HBT-3,5DM
NH Me HBT-3,5DOM
NH
600 S1 N NS R2 NY
S N R N Y
N O
S
HBIR-3,5DOM S
S HBRAA-3E
S HBRAA-3M
HOCOO SO O O HOS COO SO O O HO
COO 2
POI + HBP-3M O HO O HO
O non-radiative HO COO O S S COO SO O O COO Me
O O NH
fluorescence
HO
S O
MeMe Me NH HO HO NH
S Me Me S S1Et S MeMeO
Me OMe
SO MeO NH
MeMe Me
deexcitation Me EtO N MeNH NH S O
HBX analogs POI
NH N N MeO NH NH NH
non-radiative in NpFAST Me N NHNHO
S0 NH NH NH
S
NH
S COO
O S NH S
S HOHBP-3,5DM
S Me S COO
COO S O fluorescence HO N
S S S COO O
MeO HO
(X = O, T, P, R)SHOO O
HBRAA-3M HOHO HBRAA-3M
S HO HO COO
OHBP-3M O HBRAA-3E HO HO
HBRAA-3E
HO HBP-3,5DOM
HOHO HO S COO
HO O
HO OHO HO
R1 S O deexcitation
NH O S NH S NH O HO NH Me NH Me
HO
O Me Me NHMe O S Me HO SMeO S Et R1Me S Me HBRAA-3,5DM S HBRAA-3,5DM OMeS S NH
O
550 Me Me
promiscuous NHMe Me Me 0HONH Et NH ; 53 ; 0.05]
[0.05 HO NH Me Me OMe HO HBO-3M S HBO-3
OMe
Longueur d'onde d'absorption maximale (nm)
NH [0.08 ; 64 HBO-3,5DM
; 0.23] XNH NH
HBRAA-3E XMe NH [0.10 ; 40 ; 0.04] NH
35 ; 0.22] HBO-3M
[0.10 ;HBO-3M
recognition O HBRAA-3M
HBO-3,5DM
HO S
OHBP-3M Me
HBIR-3M HBRAA-3E
SHBIR-3,5DM
HBP-3,5DM
O S
HBP-3,5DOM
HBIR-3,5DOM OMe
HO
HBP-3M HO
HBP-3,5DM YHBP-3,5DOM
Me MeO Y HBRAA-3,5DM
HBRAA-3,5DM O
O HMBR R2
O R 2
+ HBT-3,5DMMe O O Me O NH O OHBP-3M O NH HBP-3,5DM
O O NH
HBR-3,5DOM
Me HBP-3,5DOM
O MeO
OS O S OMeO S
Me Me HBO-3M MeO HO MeOS
HBO-3,5DM
1
MeS HO Me S HO MeO S O NH
HBX analogs NH NH NH NHO S
500 S NHPOI NH S
Me S NH O non-radiative
SNH O S Me
Me S
O NH OMe
fluorescence
N O
MeOMe
S HO O
NH
HO
N S S
S HBIR-3,5DM
S S
Me HO SHO MeO S HONH O S O S NHON OS
(X = O, T, P, R)HO HO
NHCOO
S S deexcitation HO HO
HO
NHO
O HO HO HO NH OS COOS HOMe Me
NH S MeS MeO
S HBIR-3,5DOM
MeO S MeOS Me S S COO
OM
O Me SHBIR-3MO Me O Me
NHHBIR-3,5DM
HO Me OMe HO HO OMe
MeMeS Me MeO
S
OMe
[0.23
HO0 OMe S
; 54Me; 0.01] NH S NHEt S
O NH
NH NH O S NHNH NH
HBT-3,5DM HBT
HO
[0.10 ; 33 ; 0.17] HONH HOMe
HO S S
NH O
S
O HOHO Me S S S OMe
HO
[0.35
Me ; 44S;S 0.06]S
O HBT-3,5DM O HOHBT-3,5DOM HO
Me HBT-3,5DM S Me HBT-3,5DOM S HOOMe S HO
S
HO S
HO HO HBRAA-3M
HMBR
O Me MeHMBR Me S HBR-3,5DM OO HBRAA-3E
HBR-3,5DM
Me S HBR-3,5DOM
Me
O O HBR-3,5DOM
OMeOMe OMeS O
O HBIR-3MO HBIR-3,5DM
Me HBIR-3,5DOM
HBRAA-3,5DM
O450 HBO-3,5DM O Me HBIR-3,5DM OMe HBIR-3,5DOM [0.015
Me ; 25 ; 0.07] O O
HBIR-3M N
Me O O HBT-3,5DM O O N HBT-3,5DOM O HBIR-3M
O HBIR-3,5DM N Me
HBIR-3,5DOM M
S COO HBP-3,5DOM
HMBRHMBR S HBR-3,5DM
COO
HBR-3,5DM HBR-3,5DOM
S HBR-3,5DOM
COO
NH NH
NH MeNH HO O
Me O MeO O HO O HO HBRAA-3,5DM O
MeO S
HO
ONH O NH NH O MeNH
Me
S
NH Et OO NH
S
HO OMe SS HO O
S O Me
O S S O HO
S NH NMe S NH O S Me N Me MeONH N O Me NH
HO HO Me
N O HO
NH
HBP-3,5DM
O HBRAA-3M NO NH HO S HO
O COO O NH
HBRAA-3E
N NH S COO MeS COO NH
NH HO NH
HO HOCOO NHN HO NH
Me S [0.23
Me ; 6
COO; 2.4] MeMe Me S MeO
OMe S OMe S S N HO
HBRAA-3,5DM
HO400 NH HO
Me
O NH
Me Me SHO HO 1 OS Et
COO
HO
S
S
S
HBP-3M
COO HO S
S N
HBP-3,5DM
S SS HO [0.33 ; 37 ;COO
1.9]
NH
SS HO S
Me S S COO
Me HBP-3MS Et HBP-3,5DM Me S
HBP-3,5DOM S HO
HBP-3M
HO HBP-3,5DM
HBO-3M
HO HBP-3,5DOM
HBO-3,5DM
HO
Me S HBRAA-3E Et
Me S 1 [0.22 ; 58 ;1.8]
OMe
S
O NH HBRAA-3M
NH Me Sfree Me
HBRAA-3M
POI
Me Me
HBRAA-3E 0
OMe
NH HBRAA-3,5DM
O HBIR-3M S O S
STNH HBO-3M O pFAST:HBX [0.27
O ; 37 ; 0.15] HBT-3,5DOM
O HBRAA-3M O HMBRS HBR-3,5DM O
HBRAA-3,5DM HBRAA-3E
HBR-3,5DMS HBR-3,5DOM
Me MeO
S
HBP-3M S S
HBP-3,5DM
Me S S
HBP-3,5DOM Me MeO 0 HBRAA-3,5DM
Me MeO NH NH NH
O 350
Me NH Me NH
MeO NHMeO NH
NH NH
S S S
NH NHO NH NHO HBRAA-3M S NH O NH S HO HO
S HO S O HO HO O
S HO S HO O S HO S HO S S S HO O S O
HO O HO O S HO O O HO S O S Me OMe
MeO Me
HO MeS HO O Me Me Me O
Me
OMe O
O O MeMe
O
Me
O
Me Me MeO
Me MeO
OMe MeO
OMe NH [0.44 ; 49 ; 0.01] NH [0.003 ; O12 ; 0.005] NH O
[0.076 ;NH
HBO-3M 14 ; 0.43]
NH NH [0.27
HBO-3,5DM NH ; 60 ; 0.20]S NH NH SMe
1 S HBIR-3M MeOS HBIR-3,5DM
POI S S HBT-3,5DM S S
HBT-3,5DOM
HO
S S HO
SHBR-3,5DM
NH HO
NH NH
HBIR-3M
O HO HBIR-3M HBIR-3,5DM HBIR-3,5DM
HO HMBR
HBIR-3,5DOM
HO HBIR-3,5DOM
HO S HBR-3,5DOM S
HO
300 HO O O S O SO
Me
HO O
S
S Me S OMe
HO
S
Me Me Me Me Longueur d'onde
OMe d'émission
OMe maximale
S S(nm)
HO O O S
NHMe O MeO O Me O
0 S OMe
O Me
O S450 O NH OMe 500
HBIR-3,5DM NH O HMBR
HBIR-3,5DOM550 O HBR-3,5DM 600 HBR-3,5DOM 650 Me
HBIR-3M S MeO N N
Me NH S Me
b NHHO HMBR NH HBR-3,5DM
HO N
HBR-3,5DOM
NH S COO S COO
Fluorescence normalisée
Me N N N
Absorbance normalisée
S OS HO S N SNHMBR HO HBR-3,5DM HO HBR-3,5DOM HO

S S COO O pFAST HBIR-3,5DOM HBRAA-3M S
S HO
COO
1.0 COO
Me
HO
COO pFAST
OMe 1.0SHO S
COO COO S Et
HBR-3,5DOM HMBR
HBRAA-3,5DM HBT-3,5DOM
HBIR-3,5DM HBP-3,5DM
0.5 0.5 HBIR-3M HBP-3M
HBO-3,5DM HBR-3,5DM HBT-3,5DM
HBP-3,5DOM HBO-3,5DM
O
MeO HBRAA-3E HBO-3M
0.0 0.0
350 400 450 NH
500 550 600 400 450 500 550 600 650 700
S
HO Longueur d'onde (nm)
Longueur d'onde
O (nm)
OMe
-3,5DM HBT-3,5DOM
Figure 3. (a,b) pFAST génère une palette de complexes fluorescents associé aux molécules des séries HBO,
HBT, HBP, HBR, HBRAA et HBIR. O
MeO
Par la suite, la construction de deux bibliothèques de mutants par des techniques de mutagénèse
aléatoire et de « DNA shuffling », l’expression des variants à la surface des levures et la collecte en
cytométrie en flux après plusieurs cycles de sélection avec les fluorogènes (par Fluorescence
activated cell sorting FACS) nous ont permis d’identifier des variants avec de meilleures affinités et
de meilleures performances de fluorescence pour les fluorogènes en question. Finalement, trois
variants ont été retenus : oFAST issu d’une sélection avec la molécule HBO-3,5DM (série HBO) ;
tFAST issu d’une sélection avec la molécule HBT-3,5DM (série HBT) et enfin pFAST issu d’une
sélection avec la molécule HBP-3,5DM (série HBP) (Figure 4). Ce dernier forme avec HBP-3,5DM,
un complexe très affin et très brillant et fluoresce dans le canal vert. Ce variant possède également
de meilleures propriétés avec la totalité des fluorogènes nouvellement synthétisés. Enfin, pFAST a
montré d’excellentes performances avec les molécules de la série HBR et de la série HBRAA, non-
170
perméantes à la membrane cellulaire (Figure 3). Finalement, caractérisé avec la série HBIR, pFAST
forme des complexes très affins mais non-fluorescents. Cette propriété singulière nous a permis
d’éteindre très rapidement la fluorescence de pFAST. Ainsi la possibilité d’adapter la couleur de
fluorescence en choisissant un fluorogène parmi toute la collection générée permet de former des
complexes fluorescents couvrant la totalité du spectre en combinaison à un tag protéique polyvalent
et unique, pFAST.
Figure 4. (a) Stratégie d’ingénierie des variants protéiques de FAST par évolution dirigée et les principales
caractéristiques physico-chimiques des variants issus de la sélection avec (b) HBO-3,5DM, (c) HBT-3,5DM et
(d) HBP3,5DM. (e-g) Cellules HeLa exprimant oFAST, tFAST et pFAST dans le cytoplasme. Barres d’échelle,
10 μm.
Des études de modélisation menées avec le Dr. Nicolas Pietrancosta au sein du Laboratoire des
Biomolécules nous ont permis de comprendre le rôle des mutations introduites dans pFAST. Pour
cela, des modèles des structures de FAST et de pFAST ont été générés par homologie de séquence
à partir de la structure cristallographique de PYP d’Halorhodospira halophila. L’analyse plus
approfondie de chacune des mutations introduites ainsi que l’étude des mouvements moyens de
chacun des résidus montrent la compaction et la rigidification de pFAST par rapport à FAST réduisant
le volume du site de liaison (Figure 5). Des études de « docking » de l’ensemble des molécules
fluorogéniques dans le modèle généré de pFAST ont également permis de montrer que l’ensemble
des fluorogènes adoptaient une conformation quasi-planaire une fois associés à pFAST, en accord
avec la formation de complexes brillants ayant de très fortes affinités. Finalement, il a été montré que
les principaux acides aminés impliqués dans le réseau de liaisons hydrogènes stabilisant le
chromophore naturel dans PYP étaient également impliqués dans la stabilisation des formes
déprotonées des fluorogènes dans la cavité de pFAST.
171
Figure 5. Modélisation structurale de pFAST. (a) Modèles de FAST et de pFAST générés par homologie de
séquence à partir de la structure cristallographique tridimensionnelle de PYP (Photoactive Yellow Protein) issue
d’Halorhodospira halophila (PDB : 6P4I). Les mutations de FAST versus pFAST sont indiquées. (b) Volume
des sites de liaison de FAST et de pFAST généré par modélisation. (c) Fluctuations moyennes quadratiques
(Root Mean Square Fluctuation RMSF) des résidus de FAST et de pFAST (les cercles représentent les résidus
mutés dans pFAST).
L’utilisation de pFAST comme tag protéique nous a permis de marquer sélectivement plusieurs
protéines de fusion en cellules vivantes et fixées (Figure 6), en neurones et également dans des
systèmes multicellulaires. L’utilisation du modèle en embryons de poulet nous a permis de mener en
collaboration avec l’équipe du Dr. Xavier Morin à l’Institut de Biologie de l’École Normale Supérieure
des études comparatives. En électroporant de chaque côté du tube neural, un rapporteur fluorescent
(FAST ou des protéines fluorescentes classiques) et le marqueur exprimant pFAST nous avons pu
comparer systématiquement la fluorescence in vivo de notre système de marquage associé à
différents fluorogènes. De plus, la possibilité de faire de l’imagerie multicouleur en adaptant la couleur
en fonction du choix du fluorogène nous a permis de suivre des processus dynamiques de cellules
en division dans ce même modèle.
pFAST H2B-pFAST Lyn11-pFAST LifeAct-pFAST mito-pFAST MAP4-pFAST

live
pFAST H2B-pFAST Lyn11-pFAST LifeAct-pFAST mito-pFAST MAP4-pFAST

fixed
Figure 6. Images en microscopie confocale de cellules HeLa exprimant des fusions pFAST dans
différentes localisations cellulaires : dans le cytoplasme, dans le noyau, à la membrane, lié aux filaments
d’actines, dans les mitochondries et lié aux microtubules (de gauche à droite). Barres d’échelle, 10 μm.
172
Finalement, les performances de marquage et de photostabilité de pFAST en cellules, nous ont

permis de montrer que ce petit tag protéique était adapté aux techniques de microscopie les plus
avancées. En effet, la possibilité de choisir un fluorogène parmi une large collection de chromophores
fluorogéniques nous a permis de mettre en évidence que pFAST pouvait être compatible à plusieurs
modalités de microscopie de fluorescence. Associé au fluorogène HBR-3,5DOM, ce système de
marquage nous a permis de visualiser des structures en dessous de la limite de la diffraction de la
lumière en cellules vivantes et dans des cultures de neurones et astrocytes vivants en utilisant des
techniques super-résolutives STED.
a confocal b STED déconvolué
confocal STED déconvolué

b
Figure 7. Images en microscopie confocale et de super-résolution STED (après déconvolution) de (a)

cellules HeLa vivantes exprimant lyn11-pFAST (motif membranaire) en présence de 10 μM d’HBR-3,5DOM et
(b) astrocytes vivants exprimant MAP4-pFAST (protéine s’associant aux microtubules) en présence de 10 μM
d’HBR-3,5DOM. Barres d’échelle, 5 μm.
173
4. Développement de photosensibilisateurs chémogénétiques (Chapitre IV)
Bien que la présence d’espèces réactives de l’oxygène (reactive oxygen species ROS) puisse être
létale pour les systèmes vivants, celle-ci peut être ingénieusement détournée pour visualiser certains
processus biologiques au sein de la cellule. La génération de ROS peut être utilisée pour induire la
mort cellulaire, inactiver une protéine cible ou pour marquer une protéine de fusion en microscopie
corrélative. Dans le but de développer un actuateur chémogénétique basé sur le système FAST, nous
avons synthétisé, en collaboration avec l’équipe du Prof. Ludovic Jullien, des dérivés HBR portant
sur le noyau aromatique des atomes lourds de type halogènes. La présence d’atomes lourds est
connue pour augmenter la conversion intersystème (ISC) par couplage spin-orbite et ainsi favoriser
la génération de ROS42,43 par excitation lumineuse.
marqueur
chémogénétique fluorescent
S1 O
O
R2 S1
non radiative
deexcitation
R2
Fluorescence
NH Fluorogène
S NH
HO S
S + O
R1 S
pFAST R1
HMBR (R1 = Me, R2 = H)
S0 –
HBR-3,5DM (R1 = Me, R2 = Me)
S0
HBR-3,5DOM (R1 = OMe, R2 = OMe)
effet d'atomes
lourds photosensibilisateur
chémogénétique
O
ROSgène O S1
S1 ISC Mort cellulaire
+
Fluorescence
Y Y ROS
non radiative
deexcitation
NH NH T1 signalisation
S S
HO – O 3 ROS
S S O2
X X
HBR-3Br (X = Br, Y = H) S0
S0 HBR-3,5DBr (X = Br, Y = Br)
HBR-3,5DI (X = I, Y = I)
Figure 8. Principe de la photogénération d’espèces réactives de l’oxygène (ROS) : pFAST couplé à des dérivés
HBR iodés ou bromés permet de photogénérer des ROS par effet d’atomes lourds.
Dans un premier temps, les propriétés physico-chimiques de ces composés en présence de FAST,
pFAST mais également d’un autre variant iFAST44 (initialement développé à partir de FAST) ont été
caractérisées. Bien qu’HBR-3Br, HBR-3,5DBr et HBR-3,5DI gardent leur capacité à émettre de la
fluorescence une fois liés aux protéines, nous avons observé que le rendement quantique de
fluorescence des assemblages biomoléculaires était plus faible. Cette diminution du rendement
quantique de fluorescence (par rapport à la série des HBR) suggère qu’une partie de l’énergie
lumineuse absorbée est dissipée sous d’autres formes. Dans un second temps, nous avons donc
caractérisé la capacité de ces composés à générer des ROS sous l’action de la lumière une fois liés
à pFAST. La génération photoinduite de ROS a tout d’abord été mise en évidence en présence
d’ADPA (anthracene-9,10-dipropionic acid disodium). En présence d’espèces réactives de l’oxygène
174
(notamment d’oxygène singulet, de type I) cette molécule forme un dérivé non fluorescent. Une
deuxième série d’expériences a permis de mettre en évidence la présence de ROS générés par
transfert d’électrons de type II (majoritairement de type superoxyde et de peroxyde d’hydrogène). En
présence de molécules de dihydroethidium (HE) non fluorescent, la génération de ces espèces
réactives de l’oxygène conduit à la formation d’un dérivé oxydé HE* fluorescent (2-hydroxyethidium).
Finalement, des premières expériences en cellules vivantes exprimant pFAST fusionné à un motif
mitochondrial et en présence du « ROSgène » HBR-3,5DBr ont montré la perturbation ciblée des
mitochondries par action de la lumière.
175
5. Discussion générale (Chapitre V)
En conclusion, nous avons montré qu’une stratégie coordonnée d’ingénierie moléculaire et protéique
a permis de développer pFAST, un marqueur chémogénétique multifonctionel. pFAST permet de
former des complexes fluorescents avec une large collection de chromophores fluorogéniques
permettant de l’adapter à un grand nombres de scénarios expérimentaux. Ce petit tag protéique
multifonctionnel ouvre de plus de nouvelles perspectives pour le développement de biosenseurs
performants pour la détection d’analytes intracellulaires ou la caractérisation d’interactions
protéiques22,23. De plus, la large gamme spectrale accessible avec pFAST pourrait permettre le
développement de paires de FRET (Föster resonance energy transfer) en combinaison avec des
protéines fluorescentes classiques45. La possibilité de former des complexes non-fluorescents
permettrait également de développer des biosenseurs FRET avec un accepteur non-fluorescent, utile
pour des mesures en FLIM (fluorescence lifetime imaging microscopy) lors desquelles seul le temps
de vie de la molécule fluorescente donneuse est mesuré46–48. Finalement, nous avons montré que
pFAST pouvait permettre d’acquérir des informations à l’échelle nanométrique en combinaison avec
des techniques de microscopies super-résolues tel que le STED. La nature réversible de la liaison
des fluorogènes au tag protéique et leur capacité à passer d’un état fluorescent à un état non-
fluorescent, une fois excité, pourrait de la même façon être ingénieusement utilisée pour adapter
pFAST à d’autres techniques d’imagerie et de super-résolution avancées13,19,49–51.
Lors de ce travail, nous avons également montré que pFAST combiné à des molécules
« ROSgéniques » permettait la génération photoinduite d’espèces réactives de l’oxygène, ouvrant de
nouvelles perspectives pour la manipulation de processus cellulaires. Seule la liaison du ROSgène
(molécule générant des ROS) au tag protéique permet la génération photoinduite de ROS permettant
ainsi une manipulation à la demande. Bien que la caractérisation in vitro semble montrer que pFAST
générait moins de ROS que dans le cas de photosensibilisateurs de type chémogénétique (comme
miniSOG)52 ou chimique (comme Rose Bengal)53 (voir Chapitre IV), nous avons pu photo-induire une
perturbation ciblée des mitochondries au sein de cellules vivantes. De plus, nous pourrions envisager
d’induire l’inactivation de protéines cibles. Nous pourrions de la même façon envisager d’utiliser la
génération de ROS comme transducteur de signal (ou molécule sonde) pour l’étude de processus
biologiques cibles au sein de la cellule ou pour marquer spécifiquement une protéine cible en
microscopie corrélative. Finalement, dans le but de développer des systèmes plus performants et
robustes, nous pourrions tirer profit de la nature bipartite (protéine/molécule) du système FAST pour
développer par ingénierie moléculaire et/ou protéique de meilleurs générateurs d’espèces réactives
de l’oxygène.
176
6. Références
2. Patterson, G. H., Knobel, S. M., Sharif, W. D., Kain, S. R. & Piston, D. W. Use of the green
fluorescent protein and its mutants in quantitative fluorescence microscopy. Biophys. J. 73, 2782–
2790 (1997).
3. Heim, R. & Tsien, R. Y. Engineering green fluorescent protein for improved brightness, longer
wavelengths and fluorescence resonance energy transfer. Curr. Biol. 6, 178–182 (1996).
4. Matz, M. V. et al. Fluorescent proteins from nonbioluminescent Anthozoa species. Nat.
Biotechnol. 17, 969–973 (1999).
5. Wiedenmann, J., Ivanchenko, S., Oswald, F. & Nienhaus, G. Ulrich. Identification of GFP-like
Proteins in Nonbioluminescent, Azooxanthellate Anthozoa Opens New Perspectives for
Bioprospecting. Mar. Biotechnol. 6, 270–277 (2004).
6. Shaner, N. C. et al. Improved monomeric red, orange and yellow fluorescent proteins derived
from Discosoma sp. red fluorescent protein. Nat. Biotechnol. 22, 1567–1572 (2004).
7. Lavis, L. D. & Raines, R. T. Bright Ideas for Chemical Biology. ACS Chem. Biol. 3, 142–155
(2008).
Nat. Methods 14, 987–994 (2017).
9. Grimm, J. B. et al. A general method to optimize and functionalize red-shifted rhodamine dyes.
Nat. Methods 17, 815–821 (2020).
1120 (2019).
10336 (2018).
11, e02030-20 (2020).
(2019).
1828 (2018).
Chem. Biol. 17, 30–38 (2020).
24. Kumagai, A. et al. A Bilirubin-Inducible Fluorescent Protein from Eel Muscle. Cell 153, 1602–
1611 (2013).
177
25. Chapman, S. et al. The photoreversible fluorescent protein iLOV outperforms GFP as a
reporter of plant virus infection. Proc. Natl. Acad. Sci. U. S. A. 105, 20038–20043 (2008).
26. Drepper, T. et al. Reporter proteins for in vivo fluorescence without oxygen. Nat. Biotechnol.
25, 443 (2007).
27. Yapici, I. et al. “Turn-On” Protein Fluorescence: In Situ Formation of Cyanine Dyes. J. Am.
Chem. Soc. 137, 1073–1080 (2015).
29. Ozhalici-Unal, H. et al. A rainbow of fluoromodules: a promiscuous scFv protein binds to and
activates a diverse set of fluorogenic cyanine dyes. J. Am. Chem. Soc. 130, 12620–12621 (2008).
30. Yan, Q. et al. Localization Microscopy using Noncovalent Fluorogen Activation by Genetically
Encoded Fluorogen-Activating Proteins. ChemPhysChem 15, 687–695 (2014).
31. Bozhanova, N. G. et al. Protein labeling for live cell fluorescence microscopy with a highly
photostable renewable signal. Chem. Sci. 8, 7138–7142 (2017).
(2018).
33. Keppler, A., Pick, H., Arrivoli, C., Vogel, H. & Johnsson, K. Labeling of fusion proteins with
synthetic fluorophores in live cells. Proc. Natl. Acad. Sci. 101, 9955–9959 (2004).
Biol. 15, 128–136 (2008).
36. Hori, Y., Ueno, H., Mizukami, S. & Kikuchi, K. Photoactive yellow protein-based protein
labeling system with turn-on fluorescence intensity. J. Am. Chem. Soc. 131, 16610–16611 (2009).
37. Myochin, T. et al. Development of a Series of Near-Infrared Dark Quenchers Based on Si-
rhodamines and Their Application to Fluorescent Probes. J. Am. Chem. Soc. 137, 4759–4765 (2015).
38. Wirth, R., Gao, P., Nienhaus, G. U., Sunbul, M. & Jäschke, A. SiRA: A Silicon Rhodamine-
Binding Aptamer for Live-Cell Super-Resolution RNA Imaging. J. Am. Chem. Soc. 141, 7562–7571
(2019).
39. Holleran, J. et al. Fluorogen-activating proteins as biosensors of cell-surface proteins in living
cells. Cytometry A 77A, 776–782 (2010).
40. Stensrud, K. F., Zanotti, K. J., Waggoner, A. S. & Armitage, B. A. Spectral Properties of
Fluorogenic Thiophene-derived Triarylmethane Dyes. Photochem. Photobiol. 95, 406–410 (2019).
41. Song, W., Strack, R. L., Svensen, N. & Jaffrey, S. R. Plug-and-Play Fluorophores Extend the
Spectral Properties of Spinach. J. Am. Chem. Soc. 136, 1198–1201 (2014).
Methods 13, 263–268 (2016).
Markers for Live Cell Microscopy. Biochemistry 57, 5648–5653 (2018).
45. Bajar, B. T., Wang, E. S., Zhang, S., Lin, M. Z. & Chu, J. A Guide to Fluorescent Protein FRET
Pairs. Sensors 16, 1488 (2016).
46. Murakoshi, H., Shibata, A. C. E., Nakahata, Y. & Nabekura, J. A dark green fluorescent protein
as an acceptor for measurement of Förster resonance energy transfer. Sci. Rep. 5, 15334 (2015).
47. Murakoshi, H. & Shibata, A. C. E. ShadowY: a dark yellow fluorescent protein for FLIM-based
FRET measurement. Sci. Rep. 7, 6791 (2017).
48. Bertolin, G. et al. Optimized FRET Pairs and Quantification Approaches To Detect the
Activation of Aurora Kinase A at Mitosis. ACS Sens. 4, 2018–2027 (2019).
178
014010 (2017).
179
180
RÉSUMÉ
Les sondes fluorescentes jouent incontestablement un rôle primordial dans l’étude de systèmes
biologiques vivants. Nous présentons dans ce travail le développement d’un marqueur
chémogénétique hybride basé sur le système FAST (« Fluorescence-Activating and absorption
Shifting Tag »). Nous avons conçu par ingénierie moléculaire et protéique, un tag protéique unique,
pFAST, permettant de reconnaître une large collection de chromophores fluorogéniques, couvrant
la totalité du spectre visible. Grâce à la grande diversité de couleurs générées, ce système a été
adapté à un grand nombre de scénarios expérimentaux permettant aussi bien d’imager des cellules
vivantes que des organismes multicellulaires, par microscopie de fluorescence traditionnelle et de
super-résolution. Finalement, dans le but d’adapter pFAST à la manipulation de systèmes
biologiques, nous avons conçu des systèmes chémogénétiques photosensibilisateurs pouvant être
utilisés pour manipuler de façon ingénieuse les organismes vivants. De tels photosensibilisateurs
sont connus pour générer des espèces réactives de l’oxygène (ROS) permettant d’induire par
action de la lumière un grand nombre d’événements biologiques au sein de la cellule.
MOTS CLÉS
outils chémogénétiques, évolution dirigée, imagerie biologique
ABSTRACT
Fluorescent probes are undoubtedly powerful tools for the study of living organisms. Here, we
present the design of a hybrid chemogenetic reporter based on the FAST (Fluorescence-Activating
and absorption Shifting Tag) system. We applied a coordinated molecular and protein engineering
strategy to develop a small protein tag, pFAST, with suitable properties for an entire collection of
fluorogenic chromophores, spanning the entire visible spectrum. The ability to adapt the
fluorescence of a single tag by choosing a different live-cell compatible fluorogenic chromophore
provides a large experimental versatility, allowing to image live cells and multicellular organisms
using both conventional and super-resolution microscopy. Finally, we tailored pFAST for the
manipulation of biological systems by the design of chemogenetic photosensitizers. Such
photosensitizers are known to generate reactive oxygen species (ROS) which induce various
biological processes in live cells upon light illumination.
KEYWORDS
Chemogenetic tools, directed evolution, biological imaging.

Benaissa 2021 These

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Benaissa 2021 These

Transféré par

Droits d'auteur :

Formats disponibles

Engineering of chemogenetic tools for the imaging and

manipulation of biological systems

To cite this version:

HAL Id: tel-03815066

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

Engineering of chemogenetic tools for the imaging and

manipulation of biological systems

Arnaud Gautier Mme. Prof. Marie ERARD

Ecole doctorale n° 388 M. Prof. Arnaud GAUTIER

Spécialité M. Dr. Matthieu SAINLOS

manipulation of biological systems

Thèse de doctorat de Chimie-Physique

LIST OF ABBREVIATIONS ............................................................................................................. 7

CHAPTER I: GENERAL INTRODUCTION ................................................................................... 11

I.1 Molecular fluorescence ......................................................................................................... 11

CHAPTER II: FLUORESCENT CHEMICAL-GENETIC HYBRIDS .............................................. 31

II.1 Abstract ................................................................................................................................. 31

III.1 Presentation of the article ..................................................................................................... 56

CHAPTER IV: CHEMOGENETIC PHOTOSENSITIZERS FOR MANIPULATING BIOLOGICAL

IV.1 General introduction............................................................................................................ 129

V.1. Development of pFAST-based biosensors ......................................................................... 158

DIGEST : INGÉNIERIE D’OUTILS CHÉMOGÉNÉTIQUES POUR L’IMAGERIE ET LA

1. Introduction générale (Chapitre I) ....................................................................................... 163

I.1. Molecular fluorescence

Figure I.1. Perrin-Jablonski diagram (adapted from Valeur2).

I.2. Fluorescence imaging

I.2.1. Fluorescence microscopy

Fluorescence microscopy based on conventional widefield modalities is usually user-friendly, non-

Recent super-resolution techniques such as saturated structured-illumination microscopy (SSIM),

I.2.2. Fluorescence reporters

I.2.2.1. Genetically encoded fluorescent proteins

Nathan Shaner et al (2004) Nature Biotech. 22: 1567-1572

I.2.2.1. Organic dyes

TMR Janelia Fluor 549

constant (KL-Z) (adapted from Grimm et al.47). X 11

log KL–Z 2 500

525 585 479

labeling tags provide the binding specificity of fluorescentNproteins

500 600 700

II (section II.2.2) of jthis manuscript. k l

JF552 JF526 ClX

13 14 500 600 700

I.2.2.1. Fluorogen-activating’ hybrid reporters: FAST-tagging system

FAST stabilized theTyr

NATURE CHEMICAL BIOLOGY

FAST HBR-3,5DOM HMBR greenFAST HBR-3,5DOM redFA T

S H NH Wavelength (nm) Wavelength (nm)

M S Mito-greenFAST H2B-redFAST H2B-greenFAST MAP4-redFAST

I.3. Ph.D. goals

a HBO derivatives HBRAA derivatives

Normalizea aor oranc

Normalizea aor oranc

redFAST Normalize a or anc redFAST redFAST OredFAST 1 ISC

HBP-3M (R1= Me, 6 R = 7H) -activating tag 0 0 1,006 1,10 7 1, 0 M M HBR-3,5D M

fluorescence 1 600 HBR-3,5DOM (R = OMe, R = OMe) greenFAS

Beyond offering a panoply of opportunities in biological imaging, fluorogenic chemogenetic systems

Fluorescent chemical-genetic hybrids

Engineering glowing chemogenetic hybrids for spying on cells.

II.3. Hybrid reporters

II.3.1.1. Rational design

II.3.1.2. Directed evolution

II.3.1.3. In silico and de novo design

II.3.2. Fluorogen engineering

visualization in mice77. Similarly, membrane impermeant HBR derivatives containing a negatively