DOI: 10.1002/cbic.

201402677 Communications

Novel Approaches in Selective Tryptophan Isotope

Labeling by Using Escherichia coli Overexpression Media
Julia Schrghuber,[a] Toms Sra,[b] Marilena Bisaccia,[a] Walther Schmid,[a] Robert Konrat,[b]
and Roman J. Lichtenecker*[a]

NMR-based investigations of large protein complexes require A major contribution to protein NMR methods arises from in-
optimized isotopic labeling schemes. We report new methods troducing 2H, 13C, and 15N at specific positions in the amino
to introduce stable isotopes into tryptophan residues; these acid sequences. These specifically positioned stable isotopes
are fine-tuned to the requirements of the particular protein lead to reduced spectral complexity and increased signal reso-
NMR experiment. Selective backbone labeling was performed lution, because of optimization of magnetization transfer path-
by using a new a-ketoacid precursor as an additive in cell- ways. Methods of protein isotope labeling are available for
based overexpression media. Additionally, we developed syn- both aliphatic and aromatic residues.[6] Despite the prominent
thetic routes to certain isotopologues of indole with 13C1H role of tryptophan in protein structure and function, reports
spin systems surrounded by 12C and 2H. The corresponding on selective labeling of this residue are scarce. Most describe
proteins, overexpressed in the presence of these precursor direct additions of the labeled amino acid in cell-free[7] or cell-
compounds, can be effectively analyzed for conformational based expression systems.[8] High cost and effort is associated
changes in tryptophan residues in response to external stimuli, with synthesized labeled chiral tryptophan, and partial degra-
such as interaction with other proteins or small molecules. dation by tryptophanase[9] or tryptophan transaminase[10] leads
to considerable isotope scrambling in the metabolic pathway
of an overexpressing organism (e.g., Escherichia coli).[11]
Nearly all cellular processes are guided and regulated
by the interactions of proteins with other proteins,
nucleic acids, or small ligands. However, only a small
subset of particular residues account for most of the
protein interfaces free energy of binding.[1] Alanine
scanning mutagenesis revealed tryptophan to be the
most frequent residue on the surface of these hot-
spots.[2] This high degree of conservation at protein
binding sites can be explained by the unique ability
of the tryptophan side chain to contribute to aromat-
ic interactions, act as a hydrogen donor, and provide
large hydrophobic areas to shield cavities from exter-
nal solvent molecules. The exact position and motion
of the bulky tryptophan side chain is thus crucial to
understand the molecular processes at protein inter-
action surfaces.
NMR spectroscopy is the method of choice not
only to elucidate the structure of proteins[3] but also
to analyze the dynamic motion of specific side chains
at atomic resolution.[4] Additionally, NMR-based
ligand-screening methods are becoming increasingly Scheme 1. Tailor-made tryptophan precursors for selective 15N, 13C, and 2H labeling in
important in drug identification and development.[5] cell-based protein overexpression systems.

In order to fine-tune tryptophan labeling patterns in protein

[a] J. Schrghuber, M. Bisaccia, Prof. Dr. W. Schmid, Dr. R. J. Lichtenecker
Institute of Organic Chemistry, University of Vienna
Whringerstrasse 38, 1090 Vienna (Austria)
E-mail: roman.lichtenecker@univie.ac.at
[b] Dr. T. Sra, Prof. Dr. R. Konrat
Department of Structural and Computational Biology
Max F. Perutz Laboratories, University of Vienna
Dr. Bohr-Gasse 9, 1030 Vienna (Austria)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201402677.
NMR applications, we envisioned developing indolepyruvate
into a tryptophan precursor compound for selective backbone
labeling, and using indole to install specific isotope arrange-
ments into tryptophan side chains (Scheme 1). a-Ketoacids
and their salts are frequently used as precursors in cell-based
labeling because of their effective uptake by expressing organ-
isms.[12, 13] Indolepyruvate is the product of the transaminase-
catalyzed first step of catabolic tryptophan degradation in

ChemBioChem 2015, 16, 746 751 746 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Scheme 2. Synthesis of [1-13C]indolepyruvate 1. a) KOCN, H2O, 100 8C, 3 h;
then HCl, 120 8C, 4 h, 99 %; b) indole-3-carboxaldehyde, ethanolamine, H2O/
EtOH, 120 8C, 75 %; c) NaOH (20 %), argon purging, 110 8C, 81 %.
E. coli. Given the high reversibility of transaminase-catalyzed
reactions, we prepared 13C-labeled sodium indolepyruvate 1
(Scheme 2) to test our method of using this compound as a
tryptophan precursor. Starting from commercially available [1-
13
C]glycine 4, we used a one-pot reaction where hydantoic
acid was formed with KOCN as the reagent, with subsequent
cyclization to the [13C]hydantoin 5 in acidic medium at elevat-
ed temperature.[13] The indole side chain was then installed by
condensation with indole-3-carboxaldehyde to yield the hy-
dantoin derivative 6. Finally, the target compound 1 was ob-
tained by hydrolysis of the hydantoin ring in the presence of
NaOH. It was essential to strictly exclude atmospheric oxygen,
as indolepyruvate 1 is prone to oxidative degradation in
a basic environment although it is bench-stable in its lyophi-
lized form as a sodium salt. This route (Scheme 2) can analo-
gously be applied to the preparation of [2-13C]- or [1,2-13C2]-
indolepyruvate, as the corresponding labeled glycine isotopo-
logues are commercially available.
Compound 1 was used in different concentrations to label
the polyhistidine-tagged B1 domain of protein G (His-tag GB1)
as described previously (see the Supporting Information).[13]
This model protein has one tryptophan residue but displays
high variability over its entire amino-acid sequence, and is
therefore ideally suited to scanning for tryptophan and poten-
tial cross-labeling to other residues (Figure 1 A). The normalized
signal intensities in HNCO spectra were analyzed to quantify
the concentration of indolepyruvate needed to achieve satura-
tion in tryptophan backbone labeling (Supporting Informa-
tion). The corresponding relative signal intensities were plotted
against the concentration of 1 in the growth medium by a
saturation curve with quantitative introduction of 13C at 70
100 mg L1 (Figure 1 C). It is worth mentioning that we ob-
served 80 % tryptophan labeling even at precursor concentra-
tions as low as 20 mg L1. No crosslabeling to other amino
acids was observed (Figure 1 B); this validates our use of this
labeling approach for ligand screening and optimization with
diverse target proteins in future studies.
Because of economic considerations in a synthetic effort, we
chose indole as a precursor to install specific isotope patterns
in the tryptophan side chain. Indole is converted into trypto-
phan in vivo by tryptophan synthase[14] and has been used in
protein overexpression.[15] As the compound is easily exported
ChemBioChem 2015, 16, 746 751 www.chembiochem.org
Communications
Figure 1. A) Amino acid sequence of His-tag GB1; B) HNCO of His-tag GB1
expressed in the presence of labeled indolepyruvate 1 (100 mg L1) and
15
NH4Cl (1 g L1); C) HNCO signal intensity against [1-13C]indolepyruvate
1 concentration in the medium.
from the producing bacterial cells,[16] a detailed study to corre-
late indole concentration in the growth medium with the
extent of tryptophan-residue labeling in the target protein was
performed. We synthesized simple [13C]- and [15N]indole isoto-
pologues (Scheme 3). [15N]indole 12 was prepared via [15N]-
aniline hydrochloride 10, which is accessible by a reported pro-
cedure with ammonium chloride 7 as the source for 15N.[17]
HoubenHoesch acylation with chloroacetonitrile and subse-
quent reductive cyclization with NaBH4 yielded [15N]indole
12.[18]
A similar route was developed to obtain [3-13C]indole 17:
[ C]potassium cyanide 13 was converted into [1-13C]hydroxy-
13
acetonitrile 14 by reaction with paraformaldehyde,[19] and sub-
sequent conversion with thionylchloride gave [1-13C]chloro-
acetonitrile 15. This compound was used for the acylation of
aniline hydrochloride to form the target precursor 17. Indoles
12 and 17 were used in different concentrations to label His-
tag GB1 (Figure 2 A and B). The normalized signal intensities in
15
N HSQC and HMBC spectra were analyzed in order to mea-
sure the amount of indole precursor needed to achieve quanti-
tative tryptophan labeling (Supporting Information). The result-
ing saturation curve reveals a steep rise in the relative
15
N HSQC signal intensity over a low concentration range (1
747 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Scheme 3. Synthesis of labeled indoles 12 and 17. Route A: a) benzoyl chlo-
ride, KOH, diethyl ether/H2O, 42 %; b) N-bromosuccinimide, 1,8-diazabicy-
clo[5.4.0]undec-7-ene, CH3OH, 94 %; c) conc. HCl. , reflux, 93 %; d) chloroace-
tonitrile, BCl3 in CH2Cl2, ZnCl2, ClCH2CH2Cl/CH2Cl2, 62 %; e) NaBH4, dioxane,
72 %. Route B: a) paraformaldehyde, H2O/H2SO4 ; b) SOCl2, pyridine, CH2Cl2 ;
c) aniline hydrochloride, BCl3 in CH2Cl2, ZnCl2, ClCH2CH2Cl/CH2Cl2, 48 % over
three steps; d) NaBH4, dioxane, 68 %.
10 mg L1), thus proving highly effective uptake of 12 by E. coli
(Figure 2 C).
Encouraged by these results, we aimed for the synthesis of
the more elaborate precursors 2 and 3 (Scheme 4), which fea-
ture an isolated side-chain 13C1H spin system in an otherwise
highly deuterated environment. Alternating 12C13C12C and/or
2
H1H2H isotope sequences have been shown to be of high
value for probing side-chain dynamics in the case of tyrosine
and phenylalanine residues, because of the significant reduc-
tion in scalar and dipolar couplings.[20] Several synthetic routes
have been reported for the synthesis of 13C- and 2H-labeled
indole or tryptophan,[21] but few describe multilabeling in the
benzene ring. We started our route to the 13C/2H-indole precur-
sor compounds from labeled aniline hydrochlorides 19 and 22,
which are accessible from labeled acetones 18 and 21 in five
and four steps, respectively, by a synthetic approach for the
synthesis of phenylalanine and tyrosine precursors
(Scheme 4).[22] The following acylations (19!20 and 22!23)
and subsequent reductive cyclizations to indoles 2 and 3 were
performed by following a known protocol.[18]
The 1H NMR spectra of the differently isotope-labeled in-
doles 2, 3, 12, and 17 in [D6]DMSO were compared (Figure 3).
The spectral data revealed accurate isotope distribution and
showed the expected heteronuclear coupling constants. These
compounds are highly economic Trp precursors, not only be-
cause of the low-cost isotope sources used in the synthesis,
ChemBioChem 2015, 16, 746 751 www.chembiochem.org
Communications
Figure 2. His-tag GB1 overexpression in the presence of A) [15N]indole 12 or
B) [13C]indole 17. C) Correlation of [15N]indole 12 in the culture medium with
15
N HSQC signal intensities.
but also because of the low concentrations required to achieve
high isotope incorporation.
In a first application, [4,6-13C2]-5,7-dideuterioindole 3 was
added to the growth medium of E. coli overexpressing CspA
(cold-shock protein). CspA functions as an RNA chaperone and
is produced by E. coli in response to cold stress. The interaction
of CspA with the 5-untranslated region (5-UTR) of its own
mRNA (cold-box RNA) is mainly guided by phenylalanines on
the surface of the b-barrel structure.[23] The corresponding
13
C HSQC spectrum of the purified protein (Figure 4 A; details
in the Supporting Information) again revealed selective label-
ing at the single tryptophan residue of CspA. Addition of an
RNA ligand that has been shown to bind to CspA[24] (5-UTR of
the anti-cold-box, partly complementary to the cold-box RNA)
resulted in significant differences in chemical shifts and a de-
crease in peak intensities (Figure 4 B). These preliminary results
confirm that the side chain of Trp11 is within the RNA binding
site on the surface of the protein and is clearly influenced by
ligand binding. The introduced stable isotope pattern sets the
stage for further kinetic analysis of the binding process, for
example, CarrPurcellMeiboomGill (CPMG) relaxation disper-
sion experiments, which have been performed at other resi-
dues to determine the sparsely populated states of CspA-RNA
complexation.[24]
In summary, we have developed two complementary trypto-
phan-labeling methods. The first uses [1-13C]indolepyruvate 1
748 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Communications

Figure 4. 13C HSQC of the cold-shock protein CspA, overexpressed in mini-

mal E. coli medium supplemented with [4,6-13C2]-5,7 dideuterioindole 3, A)
without and B) in the presence of an RNA ligand (10 % molar ratio).

Scheme 4. Synthesis of [5-13C]-4,6,7-trideuterioindole 2 and [4,6-13C2]-5,7-

dideuterioindole 3 (details in the Supporting Information). a) see ref. [22]; source. Our studies concerning the overexpression of the
b) chloroacetonitrile, BCl3 in CH2Cl2, ZnCl2, ClCH2CH2Cl/CH2Cl2, 70 % for 20, model protein His-tag GB1 showed effective 13C incorporation
65 % for 23; c) NaBH4, dioxane, 75 % for 2, 87 % for 3. without crosslabeling of other residues. The results indicate
that protein labeling based on indolepyruvate will highly sim-
plify NMR experiments that rely on backbone-shift perturba-
as a precursor for selective 13C backbone labeling; it is accessi- tions upon titration with ligand compounds. Such a restriction
ble by a straightforward route with glycine as a cheap isotope on tryptophan backbone spin systems does not only increase
the weight limit of accessible proteins, but also ena-
bles the analysis of proteinligand interactions in-
volving incompletely assigned proteins.
The second method aimed for tryptophan side-
chain labeling based on an indole precursor. A syn-
thetic approach to access various elaborate isotope
patterns of indole has been developed by using
cheap 13C-labeled acetone as a starting compound
and D2O as a source of deuterium. The routes com-
plement our previously reported synthetic concept[22]
to install isolated 13C1H spin systems into the side
chains of phenylalanine, tyrosine and tryptophan by
using independent precursor compounds 2, 3, 24
26 (Scheme 5). These compounds can be prepared
by diverging routes via labeled anilines 19 and 22 as
the common intermediates (Scheme 5). 13C/2H in-
doles 2 and 3 were applied to probe tryptophan
side-chain dynamics by spin-relaxation rate determi-
nation. Relaxation experiments are seriously affected
by coupling to adjacent spin systems. The isolated
13
C1H pairs of 2 and 3 eliminate unwanted relaxa-
tion pathways, thus resulting in straightforward and
direct NMR elucidation of tryptophan dynamic
Figure 3. 1H NMR spectra of A) commercially available indole, B) [15N]indole 12, C) [3-
13 13 13 motion. Investigations on phenylalanine and tyrosine
C]indole 17, D) [5- C]-4,6,7-trideuterioindole 2, and E) [4,6- C2]-5,7-dideuterioindole 3,
showing the following spin couplings: 1J N H (1), 2J N H (2), 1J C H (3), 2J C H (4) and 3J C H
15 1 15 1 13 1 13 1 13 1
conformational dynamics based on aromatic spin re-
(5). laxation, as well as on aromatic residual dipolar cou-

ChemBioChem 2015, 16, 746 751 www.chembiochem.org 749 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Scheme 5. Synthetic approach to prepare the phenylalanine precursors 24
and 25,[22] tyrosine precursor 26,[22] and tryptophan precursors 2 and 3 via
anilines 19 and 22.
pling, have been presented recently.[25] Our synthetic concept
will help to transfer these methods to applications concerning
the tryptophan side chain, and the corresponding studies con-
stitute an important part of our ongoing research.
Experimental Section
Spectroscopic characterization of all products and intermediates
and experimental details for protein overexpression and NMR spec-
troscopy are given in the Supporting Information.
Acknowledgements
This work was partly supported by grants from the Austrian Sci-
ence Foundation (FWF) (I844 to R.K.). We thank Christoph Kreutz
and Christoph H. Wunderlich (Institute of Organic Chemistry, Uni-
versity of Innsbruck) for providing the RNA oligonucleotide.
Keywords: indole isotope labeling NMR spectroscopy
protein expression tryptophan
ChemBioChem 2015, 16, 746 751 www.chembiochem.org
Communications
[1] I. S. Moreira, P. A. Fernandes, M. J. Ramos, Proteins Struct. Funct. Bioinf.
2007, 68, 803 812.
[2] A. A. Bogan, K. S. Thorn, J. Mol. Biol. 1998, 280, 1 9.
[3] a) L. E. Kay, J. Magn. Reson. 2005, 173, 193 207; b) L. E. Kay, K. H. Gard-
ner, Curr. Opin. Struct. Biol. 1997, 7, 722 731; c) V. Tugarinov, L. E. Kay,
ChemBioChem 2005, 6, 1567 1577; d) C. Gbl, T. Madl, B. Simon, M. Sat-
tler, Prog. Nucl. Magn. Reson. Spectrosc. 2014, 80, 26 63.
[4] a) A. J. Wand, Nat. Struct. Biol. 2001, 8, 926 931; b) L. E. Kay, Nat. Struct.
Biol. 1998, 5, 513 517; c) R. Rosenzweig, L. E. Kay, Annu. Rev. Biochem.
2014, 83, 291 315.
[5] a) B. Meyer, T. Peters, Angew. Chem. Int. Ed. 2003, 42, 864 890; Angew.
Chem. 2003, 115, 890 918; b) J. Weigelt, M. van Dongen, J. Uppenberg,
J. Schultz, M. Wikstrm, J. Am. Chem. Soc. 2002, 124, 2446 2447.
[6] a) K. H. Gardner, L. E. Kay, Annu. Rev. Biomol. Struct. 1998, 27, 357 406;
b) V. Tugarinov, P. M. Hwang, L. E. Kay, Annu. Rev. Biochem. 2004, 73,
107 146; c) V. Tugarinov, V. Kanelis, L. E. Kay, Nat. Protoc. 2006, 1, 749
754; d) C. G. Hoogstraten, J. E. Johnson, Jr., Concept Magn. Res. Part A
2008, 32A, 34 55; e) N. K. Goto, L. E. Kay, Curr. Opin. Struct. Biol. 2000,
10, 585 592.
[7] a) M. Kainosho, T. Torizawa, Y. Iwashita, T. Terauchi, A. M. Ono, P. Gntert,
Nature 2006, 440, 52 57; b) Y. Miyanoiri, M. Takeda, J. Jee, A. M. Ono, K.
Okuma, T. Terauchi, M. Kainosho, J. Biomol. NMR 2011, 51, 425 435.
[8] C. Berger, J. T. C. Ho, T. Kimura, S. Hess, K. Gawrisch, A. Yeliseev, Protein
Expression Purif. 2010, 70, 236 247.
[9] S.-Y. Ku, P. Yip, P. L. Howell, Acta Crystallogr. Sect. D 2006, 62, 814 823.
[10] C. Mavrides, W. Orr, J. Biol. Chem. 1975, 250, 4128 4133.
[11] C. Berger, S. Berndt, A. Pichert, S. Theisgen, D. Huster, Biotechnol.
Bioeng. 2013, 110, 1681 1690.
[12] a) N. K. Goto, K. H. Gardner, G. A. Mueller, R. C. Willis, L. E. Kay, J. Biomol.
NMR 1999, 13, 369 374; b) V. Tugarinov, L. E. Kay, J. Am. Chem. Soc.
2003, 125, 13868 13878; c) R. Lichtenecker, M. L. Ludwiczek, W.
Schmid, R. Konrat, J. Am. Chem. Soc. 2004, 126, 5348 5349; d) R. J. Lich-
tenecker, K. Weinhupl, W. Schmid, R. Konrat, J. Biomol. NMR 2013, 57,
327 331.
[13] R. J. Lichtenecker, N. Coudevylle, R. Konrat, W. Schmid, ChemBioChem
2013, 14, 818 821.
[14] M. F. Dunn, D. Niks, H. Ngo, T. R. M. Barends, I. Schlichting, Trends Bio-
chem. Sci. 2008, 33, 254 264.
[15] a) R. A. Rodriguez-Mias, M. Pellecchia, J. Am. Chem. Soc. 2003, 125,
2892 2893; b) T. Tarrag, B. Claasen, N. Kichik, R. A. Rodriguez-Mias, M.
Gair, E. Giralt, ChemBioChem 2009, 10, 2736 2739.
[16] a) K. Kawamura-Sato, K. Shibayama, T. Horii, Y. Iimuma, Y. Arakawa, M.
Ohta, FEMS Microbiol. Lett. 1999, 179, 345 352; b) S. Piero-Fernandez,
C. Chimerel, U. F. Keyser, D. K. Summers, J. Bacteriol. 2011, 193, 1793
1798.
[17] J. D. DeSousa, B. M. Novak, ACS Macro Lett. 2012, 1, 672 675.
[18] C. A. B. Wager, S. A. Miller, J. Labelled Compd. Radiopharm. 2006, 49,
615 622.
[19] C. S. Elmore, D. C. Dean, Y. Zhang, D. G. Mellilo, J. Labelled Compd. Radi-
opharm. 2004, 47, 837 846.
[20] V. Kasinath, K. G. Valentine, A. J. Wand, J. Am. Chem. Soc. 2013, 135,
9560 9563.
[21] a) E. M. M. van den Berg, A. U. Baldew, A. T. J. W. de Goede, J. Raap, J.
Lugtenburg, Recl. Trav. Chim. Pays-Bas 1988, 107, 73 81; b) E. M. M. van
den Berg, W. B. S. van Liemt, B. Heemskerk, J. Lugtenburg, Recl. Trav.
Chim. Pays-Bas 1989, 108, 304 313; c) E. M. M. van den Berg, F. J. H. M.
Jansen, A. T. J. W. de Goede, A. U. Baldew, J. Lugtenburg, Recl. Trav.
Chim. Pays-Bas 1990, 109, 287 297; d) B. R. Branchini, F. G. Prendergast,
G. A. Spencer, J. D. Hugdahl, B. D. Ray, M. D. Kemple, J. Labelled Compd.
Radiopharm. 1987, 24, 637 643; e) M. F. Lautie, M. Corval, J. Labelled
Compd. Radiopharm. 1982, 19, 1011 1020; f) S.-S. Yuan, A. M. Ajami, Tet-
rahedron 1982, 38, 2051 2053; g) C. J. Unkefer, S. N. Lodwig, L. A.
Silks III, J. L. Hanners, D. S. Ehler, R. Gibson, J. Labelled Compd. Radio-
pharm. 1991, 29, 1247 1256; h) N. Ilic, J. D. Cohen, J. Labelled Compd.
Radiopharm. 2004, 47, 635 646.
[22] R. J. Lichtenecker, Org. Biomol. Chem. 2014, 12, 7551 7560.
[23] a) B. Xia, H. Ke, W. Jiang, M. Inouye, J. Biol. Chem. 2002, 277, 6005
6011; b) K. Newkirk, W. Feng, W. Jiang, R. Tejero, S. D. Emerson, M.
Inouye, G. T. Montelione, Proc. Natl. Acad. Sci. USA 1994, 91, 5114 5118.
[24] T. Sra, T. C. Schwarz, D. Kurzbach, C. H. Wunderlich, C. Kreutz, R. Konrat,
ChemistryOpen 2014, 3, 115 123.
750 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Communications
[25] a) B. Sathyamoorthy, K. K. Singarapu, A. E. Garcia, T. Szyperski, ChemBio- chemistry 2014, 53, 4519 4525; e) U. Weininger, M. Respondek, C. Lw,
Chem 2013, 14, 684 688; b) U. Weininger, C. Diehl, M. Akke, J. Biomol. M. Akke, J. Phys. Chem. B 2013, 117, 9241 9247.
NMR 2012, 53, 181 190; c) U. Weininger, M. Respondek, M. Akke, J.
Biomol. NMR 2012, 54, 9 14; d) U. Weininger, K. Modig, M. Akke, Bio- Received: November 26, 2014
Published online on February 20, 2015

ChemBioChem 2015, 16, 746 751 www.chembiochem.org 751 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Copyright of ChemBioChem is the property of John Wiley & Sons, Inc. and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.