Vous êtes sur la page 1sur 10

REVIEWS

CHEMICAL GENETICS:
LIGAND-BASED DISCOVERY
OF GENE FUNCTION
Brent R. Stockwell
Chemical genetics is the study of gene-product function in a cellular or organismal context
using exogenous ligands. In this approach, small molecules that bind directly to proteins are
used to alter protein function, enabling a kinetic analysis of the in vivo consequences of these
changes. Recent advances have strongly enhanced the power of exogenous ligands such
that they can resemble genetic mutations in terms of their general applicability and target
specificity. The growing sophistication of this approach raises the possibility of its application
to any biological process.

FORWARD GENETICS
Classic FORWARD-GENETIC screens in model organisms have ly entails causing an environmental stress to the organ-
Genetic analysis that proceeds revealed the molecular basis for diverse biological ism, such as a temperature shift, which itself can have
from phenotype to genotype processes, including cell division in the yeast marked consequences9 and may confound the interpre-
through positional cloning or a Saccharomyces cerevisiae1, programmed cell-death in the tation of results. This drawback prevents the functions
candidate-gene approach.
nematode Caenorhabditis elegans2, and embryonic pat- of essential genes from being investigated because
tern formation in the fly Drosophila melanogaster3. Such organisms with mutations in these genes are not viable.
genetic screens use a three-step procedure (FIG. 1a) that In addition, organisms that carry constitutive mutations
entails: first, random mutagenesis of a large number of may have time to compensate for their effects by upreg-
organisms; second, screening the resulting mutants to ulating related genes, which can obscure the initial
identify those with a defect in the process of interest (for effects of a mutation.
example, cell division, apoptosis or embryonic pattern- Despite these limitations, the power of the genetic
ing); and last, identifying the mutations in specific genes approach stems from the ability of mutations to alter
that underlie the mutant phenotypes of interest. This is the function of a single gene product within the con-
a powerful method for identifying genes that regulate text of a complex cellular environment 10,11. In a
biological processes in lower organisms. chemical-genetic approach, exogenous ligands are
It is difficult, however, to use a genetic approach in likewise used to alter the function of a single gene
mammals because of their slow rate of reproduction, product within this complex cellular context.
large physical size, and large, diploid genome4. But Chemical-genetic screens use a three-step procedure
recent attempts at doing large-scale genetic screens in (FIG. 1b) that entails: first, assembling a set of mutation
vertebrates, such as mice and zebrafish, suggest that equivalents (ligands capable of altering protein func-
such efforts are warranted48. tion); second, doing a high-throughput screen for lig-
A second limitation of the genetic approach is that ands that affect a biological process of interest (such
Whitehead Institute for most mutations are not conditional they cannot be as cell division, programmed cell death or embryonic
Biomedical Research, turned on or off at will. Even when conditional muta- pattern formation); and last, identifying protein tar-
9 Cambridge Center,
Cambridge, Massachusetts tions can be found (for example, temperature-sensitive gets of these active ligands.
02142, USA. e-mail: mutations or mutant alleles under the control of an The chemical-genetic approach is conditional and
stockwell@wi.mit.edu inducible promoter), expressing the mutant allele usual- readily applicable to cells derived from complex organ-

116 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics


2000 Macmillan Magazines Ltd
REVIEWS

TARGET IDENTIFICATION isms such as vertebrates. This is because a ligand can This review summarizes the evolution of the con-
The process of finding cellular be added or removed at any time, enabling a kinetic ceptual framework and technologies necessary for
macromolecules that bind to a analysis of the consequences of changes in protein implementing chemical-genetic analyses. The current
small molecule or peptide. function in vivo. This permits an analysis of the cellu- techniques in generating chemical diversity, high-
lar or physiological events that occur immediately after throughput screening and TARGET IDENTIFICATION are sur-
altering the activity of a specific protein. In higher veyed, and recent successes with these technologies
organisms, such as mice and humans, the approach are highlighted.
can be used for genetic-like screens in cells and tissues
of these organisms, complementing genetic-screening Early history: pure substances and receptors
methods. Although this review focuses on the use of The two theoretical underpinnings of chemical genet-
protein-binding reagents, other exogenous ligands can ics are: first, that pure biologically active substances
be similarly used, such as RNA-binding antisense can be obtained; and second, that such substances act
reagents and DNA-binding reagents12,13. by binding to specific molecular targets within an

Figure 1 | Genetic and chemical-genetic approaches identify genes and proteins, respectively, that regulate
biological processes. a | Forward genetics entails introducing random mutations into cells, screening mutant cells for a
phenotype of interest and identifying mutated genes in affected cells. In the example shown, yeast cells are randomly mutated,
cells showing a large-bud phenotype are selected, and genes mutated in these cells are identified. Reverse genetics entails
introducing a mutation into a specific gene of interest and studying the phenotypic consequences of the mutation in a cellular or
organismal context. In the example shown, a single mutated gene is introduced into yeast cells and a large-bud phenotype is
observed. b | Forward chemical-genetics entails screening exogenous ligands in cells, selecting a ligand that induces a
phenotype of interest, and identifying the protein target of this ligand. In the example shown, one compound that induces a
large-bud phenotype is selected and the protein target of this ligand is subsequently identified. Reverse chemical-genetics
entails overexpressing a protein of interest, screening for a ligand for the protein, and using the ligand to determine the
phenotypic consequences of altering the function of this protein in a cellular context. In the example shown, a ligand for a
specific protein is found to induce a large-bud phenotype.

NATURE REVIEWS | GENETICS VOLUME 1 | NOVEMBER 2000 | 1 1 7


2000 Macmillan Magazines Ltd
REVIEWS

SMALL ORGANIC MOLECULE organism. The origin of these pivotal ideas can be Finally, the usefulness of SMALL ORGANIC MOLECULEs for
(also called small molecule) traced back to the eighteenth century, when many studying biological systems was discovered at the same
Carbon-containing compounds researchers began to suggest that plant extracts that time that the theoretical bases of chemical genetics
that usually have a molecular
affect animals each contain a single active ingredient were developed. Before the nineteenth century, unchar-
weight of less than 2,000 g mol-1.
that acts on a discrete part of the animal14. acterized, naturally occurring plant extracts were used
STEREOCENTRE Subsequently, Frederick Serturner succeeded in isolat- to perturb biological systems. Then Paul Ehrlich
A carbon atom that bears four
ing the first active compound (morphine) (FIG. 2e) from showed, somewhat surprisingly, that small organic
distinct functional groups.
a medicinal plant (opium)15. The first conceptual pillar molecules derived from an industrial waste product,
of chemical genetics was therefore solidified that coal tar, could selectively interact with cells and
biological activity resides within pure substances. tissues19. Ehrlich discovered that methylene blue (FIG. 2a)
The second premise of chemical genetics that could stain nerve cells selectively, and he promoted the
low-molecular-weight compounds act by binding to idea that low-molecular-weight organic compounds
specific protein receptors took another hundred could be of value for studying their receptors in biolog-
years to develop. In the early nineteenth century, ical systems14,18. Furthermore, in 1884, Hans Christian
Franois Magendie and Claude Bernard proposed that Gram found that a particular dye, crystal violet (FIG. 2b),
small molecules act within a specific body structure, and could selectively stain certain types of microorganism
Rudolf Buchheim formulated the idea that drug action in a process now referred to as the Gram stain20.
could be explained on the basis of a physicochemical Several useful small molecules, such as aspirin (FIG. 2c)
reaction between cell constituents and a particular and saccharin (FIG. 2d), were also discovered by studying
drug16. However, the crucial breakthrough came at the coal-tar extracts, showing that even very simple organic
beginning of the twentieth century, when Paul Ehrlich molecules could have a range of interesting and useful
developed both the term for, and the concept of, a biological effects21,22. Before long it was accepted that
receptor the single, specific protein target of a small simple organic molecules could profoundly affect cel-
molecule17,18. (Note that a small-molecule receptor may lular and organismal systems, paving the way for a
or may not function as a cell-surface receptor in the tra- comprehensive chemical-genetic approach to under-
ditional biological sense.) Today, Ehrlichs breakthrough standing biological systems.
has been extended with the vision that we will eventual- So, by the early twentieth century it was substantiat-
ly identify a small-molecule partner for every protein10. ed, first, that pure substances have biological activity,

Figure 2 | Small molecules have a large range of structural complexity. Simple small molecules, such as a | methylene
blue, b | crystal violet, c | aspirin and d | saccharin, lack STEREOCENTRES, whereas complex small molecules, such as e | morphine,
f | FK506, g | cyclosporin and h | peptides, have one or more stereocentres.

118 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics


2000 Macmillan Magazines Ltd
REVIEWS

Table 1 | Target-based ligand discovery peptide libraries has developed into a discipline in
itself the field of chemical diversity.
Ligand type Target Reference
Early efforts to generate chemical diversity focused
Peptide Thymidylate synthase 95
on chemically synthesized peptides because they are
Peptide Calmodulin 96 composed of the same building blocks as proteins and
Peptide Cyclin-dependent kinase 33 were therefore thought to be good candidates for lig-
Peptide Erythropoietin 97 ands that bind to proteins. These syntheses followed
Peptide Thrombopoietin 98 the pioneering work of Merrifield, who developed
highly efficient SOLID-PHASE SYNTHESES of peptides27.
Peptide Fibroblast growth factor 99
Chemical syntheses of such libraries have been done
Peptide Interleukin-1 100
using numerous methods: in solution using conven-
Small organic SH3 domains 101 tional chemistry28, on small plastic beads using solid-
molecule
phase peptide synthesis29,30, and on silicon wafers using
Small organic Kinases 102 light-directed synthesis31. Genetically encoded pep-
molecule
tide-aptamer libraries have also been developed, in
Small organic G-protein-coupled 103
molecule receptors
which short peptides are displayed in vitro on phage
particles32, or in vivo on scaffold proteins33. Peptide
Small organic Cytokine receptors 60
molecule libraries have yielded ligands for many different pro-
tein targets (TABLE 1).
Small organic Lectins 104
molecule Over the past decade, methods for the synthesis of
Small organic Proteases 105
chemical libraries of non-peptidic, low-molecular-
molecule weight organic compounds have been developed25,26.
Organic compounds have the advantage of often
being able to permeate the plasma membrane of tar-
and second, that these substances interact with specific get cells34. Although it is theoretically possible that
proteins within an organism. Furthermore, small mole- completely unbiased libraries of small molecules
cules were discovered to be generally useful tools for might lack any biological activity, this approach has,
probing biological systems because of their ability to in fact, been quite successful. Synthetic methods have
interact selectively with different cells, tissues and been developed for producing a wide range of chemi-
organisms. Nonetheless, the development of these prin- cal structures using COMBINATORIAL CHEMISTRY (TABLE 2).
ciples into chemical genetics, a general method for These libraries have yielded ligands for many proteins
determining protein function, took almost another cen- (TABLE 1), as well as compounds that are active in cell-
tury. It is only now that we are beginning to witness the based assays3538.
full power and extent of these simple ideas. However, despite the apparent success of combinato-
rial chemistry, the field is still in its infancy. It remains to
Chemical-genetic tools: chemical libraries be seen which types of combinatorial library will yield
In the first step of the chemical-genetic process, a collec- small molecules with exquisite protein-binding specifici-
tion of diverse ligands, capable of altering protein func- ty, which is required if such compounds are to be used as
tion, is assembled. These ligands can be small organic mutation equivalents. Most of the large combinatorial
molecules (FIG. 2) or PEPTIDE APTAMERs23,24. Both types of libraries synthesized so far contain simple compounds
ligand act as mutation equivalents because they alter the with few or no stereocentres26, in contrast to the stere-
function of their target protein. In either case, a LIBRARY ochemical complexity that has been found in many
of such ligands must be obtained. natural products (FIG. 2)39,40. Proteins are stereochemically
complex because they are composed of many -amino
PEPTIDE APTAMER
The origin of chemical libraries. Modern pharmaceuti- acids, most of which have at least one stereocentre. It is
A short oligomer of amino
acids, often fused to a protein cal companies have chemical libraries of small mole- likely that small molecules containing stereocentres will
scaffold. cules that collectively number in the millions. These show greater specificity in their protein-binding proper-
enormous collections have been assembled one-by- ties than small molecules lacking stereocentres. But
LIBRARY one during the course of synthesizing many variants of despite promising preliminary research41, compounds
A large collection of peptides,
small molecules or other
each drug candidate over the past century. rivalling the structural complexity, high potency and tar-
reagents. Furthermore, many tens of thousands of natural prod-
ucts, derived from marine sponges, fungi, bacteria and
Table 2 | Combinatorial libraries
SOLID-PHASE SYNTHESIS plants, have gradually been added to these libraries.
Chemical synthesis method
These collections have become a valuable resource for Small-molecule class Reference
using a solid support, such as a
screens related to therapeutic drug discovery and Benzodiazepines 105
plastic bead.
development. However, this screening approach was Peptidomimetic organic compounds 106
COMBINATORIAL CHEMISTRY largely unavailable to academic researchers until Non-peptidic amino-acid polymers 57
The synthesis of numerous recently, when small collections of organic molecules
organic compounds by Aminothiazoles 107
combining variations of each
became commercially available and, more important-
Shikimic acid derivatives 41
of the building blocks that ly, rapid syntheses of small molecule libraries were
comprise the compounds. invented25,26. The synthesis of organic-compound or Carbohydrates 103

NATURE REVIEWS | GENETICS VOLUME 1 | NOVEMBER 2000 | 1 1 9


2000 Macmillan Magazines Ltd
REVIEWS

Chemical-genetic tools: screens


Box 1 | Reverse chemical-genetics: a case study To test the effects of thousands or millions of small
The reverse chemical-genetic approach molecules, rapid methods need to be devised to con-
has recently been used to study the duct many assays in parallel. Such methods are
biology of MEK152. An in vitro high- referred to as high-throughput screens because of their
throughput kinase assay was used to ability to put numerous candidate compounds
select PD 184352 (see the figure), a through the screen in question. Testing small mole-
small organic molecule that potently cules in a high-throughput screen requires that each
inhibits the kinase activity of MEK1 in small molecule be spatially segregated to prevent
vitro . It was found to be a specific reagent mixing. This feature has stimulated efforts to
inhibitor of MEK1 in vitro because it minimize the volume of each assay and so maximize
could not inhibit the kinase activities of the number of assays done in a given time46.
several other kinases that were tested. As the number of compounds available for screening
To define the function of MEK1 in cell-
has increased, the throughput of assay technology has
cycle progression, cell growth and cell
kept pace. Assay volumes were reduced to 100300 l in
morphology, the effect of PD 184352 on
the 96-well plate format in the late 1970s, and to 2080 l
these processes was tested in colon
in the 384-well plate format by the 1990s. The past few
tumour cells. These studies found that
MEK1 activity is required for: the years have witnessed a proliferation of nanowell plates,
progression of cells from the G1 to S containing thousands of small wells that permit assay vol-
phase of the cell cycle, anchorage- umes of several hundred nanolitres4749. This fulfils two
independent growth, cell scattering and purposes: first, decreasing the length of time that is
the conversion of cells from a flattened required per assay (and therefore increasing the number
to a round morphology (as shown in of assays that can be completed in a given time), and sec-
the figure, PD 184352 causes the round ond, decreasing the cost of each assay. The combined
morphology of colon tumour cells effect has enabled the testing of hundreds of thousands of
(Adapted with permission from Nature Med. (upper panel) to revert to a flattened compounds each day in a fully automated high-through-
(REF. 52) (1999) Macmillan Magazines Ltd) morphology (lower panel))52 . Treating put screening laboratory46,49.
colon tumour-bearing mice with PD 184352 caused tumour size to shrink, So now the rate-limiting step in applying this
indicating that MEK1 function is also required for in vivo colon tumour cell approach to a biological question is the development of
growth. (PD 184352 had no effect in mice with leukaemia.) So in this study, a a robust high-throughput screen for the process of
small-molecule MEK1 inhibitor was discovered and used to define the function of interest. Towards this end, two types of screen are possi-
MEK1 in tumour processes. It is interesting to note that in the analogous reverse- ble target-based and phenotype-based screens.
genetic approach, Mek1-deficient mouse embryos died in early embryogenesis94
and, because it is an essential gene, its function could not be studied in tumour Target-based screens. These screens are used to identify
processes from adult mice by reverse-genetic methods. In such cases, reverse
ligands for a specific protein of interest. Subsequently,
chemical-genetic methods complement and extend reverse-genetic methods for
such ligands can be used to inhibit or to activate the
studying specific gene product functions in vivo.
function of this protein in vivo. This method is concep-
tually analogous to REVERSE-GENETIC methods, such as
get specificity of natural products are not available using gene targeting in mice, in that both methods search for
existing combinatorial synthetic methods. the phenotypic consequences of altering the activity of a
Surprisingly, the optimal types of chemical structure single protein (FIG. 1).
for perturbing biological systems are not known. Most Target-based screens are done by purifying the protein
synthetic efforts in the pharmaceutical industry have target of interest and developing binding in vitro or enzy-
focused on flat aromatic compounds, similar to the dyes matic assays. In either case, the assay must be miniatur-
synthesized in the nineteenth century (FIG. 2). The com- ized to 2100 l to allow screening in 96-, 384- or 1,536-
mon view is that such compounds will be more readily well plates. Typically, between 10,000 and 1,000,000
developed into human therapeutic agents42, although it compounds are tested in a target-based screen, yielding
is just as likely that many human therapeutic agents 10 to 100 candidate ligands. These candidates are re-test-
have been developed using these compounds simply ed several times at various concentrations and only con-
because they were the only compounds available for firmed hits are taken forward for tests of specificity and in
testing. So despite the progress in developing new syn- vivo functionality.
thetic reactions during this century, there has been a Ligands to a specific protein can be used to elucidate
lack of creativity in the selection of target chemical the phenotypic consequences of inhibiting or otherwise
structures until recently. We are beginning to witness an altering this protein50. For example, a small-molecule
explosion in creatively selected, novel, small organic inhibitor of p53 was identified in a target-based screen
molecules that will function as powerful tools for per- and was used to show that loss of p53 function reduced
turbing biological systems4345. In time, perhaps, we will the side effects of anti-tumour therapeutic regimens51. In
REVERSE GENETICS learn the fundamental principles that govern the rela- addition, a small-molecule inhibitor of the mitogen-acti-
Genetic analysis that proceeds
from genotype to phenotype
tionship between chemical structure and both potency vated protein kinase kinase, MEK1, has been used to
through gene-manipulation and selectivity. These will help guide the selection of tar- determine the function of this kinase in tumour cell
techniques. get molecules for combinatorial syntheses. growth in vivo (BOX 1) 52. Target-based screens have yielded

120 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics


2000 Macmillan Magazines Ltd
REVIEWS

as somatostatin receptor subtype-5 activation altered


insulin secretion from pancreatic -cells60. This study
also showed the feasibility of using small molecules to
activate, rather than inactivate, specific members of a
protein family. Such functional manipulation is difficult
to achieve using conventional reverse-genetic methods.
However, the difficulty of identifying specific agonists
for each member of a receptor family increases as the
number of proteins in the family increases. For example,
the kinase family, with hundreds of members, repre-
sents a formidable challenge. Identifying selective
inhibitors of each kinase will probably require larger
and more complex libraries of exogenous ligands than
have been synthesized so far.

Phenotype-based screens. These screens test the ability


of each member of a library to induce a specific pheno-
typic outcome in a cell or organism (FIG. 3). Such screens
are analogous to classic forward-genetic screens in
model organisms (FIG. 1)11.
Phenotype-based chemical-genetic screens have
been done in single-cell organisms, such as bacteria25,26
and fungi61, and in cells from multicellular organisms,
such as mice and humans62,63. No phenotype-based,
chemical-genetic screen has been done in whole ani-
mals because of the difficulty of maintaining, injecting
and observing thousands, or tens of thousands, of ani-
mals. However, recent progress in vertebrate pheno-
type-based genetic screening46,64,65 indicates that such
an approach may ultimately be feasible.
Detection methods for phenotype-based screens
Figure 3 | High-throughput cell-based assays. These assays measure the effect of a
collection of exogenous ligands on a specific biological process. a | To screen a large number
include the use of markers, functional assays and micro-
of small molecules, a microtitre plate is prepared with the cells of interest. b | A small volume of scopic imaging of cells. Assays using marker genes or
each compound to be tested is added to a well of the microtitre plate. ce | One of several proteins (for example, reporter-gene assays and anti-
possible methods is used to detect the effect of each compound on the cells. c | An antibody body-based cellular immunoassays) measure the abun-
detects post-translational or biosynthetic changes within the cell. d | A reporter-gene assay dance of a specific gene or protein epitope as a marker
detects changes in gene expression at a specific promoter. e | Morphological or other changes of a more broad cellular phenotype. For example, trans-
in the cells are visualised by microscopy. In each case, those reagents that cause the desired
cellular change are selected for further study (shown as the red well in b).
forming growth factor- (TGF-) induces numerous
transcriptional and post-translational changes in target
cells, but the expression level of a single gene (plasmino-
small-molecule or peptide ligands for many proteins, gen activator inhibitor type I) is often used as a marker
including kinases33, phosphatases53, proteases54, cell-sur- for active TGF- signalling66. Functional assays directly
face receptors55,56, SH3 (Src homology region 3) measure cellular activities such as cell division, metabo-
domains57, E3 ubiquitin ligases58 and steroid receptors59. lism, apoptosis, chemotaxis or adhesion. Finally, auto-
In general, small-molecule inhibitors of specific proteins mated microscope-based imaging systems can detect
can be enormously valuable for rapidly assessing whether changes in cellular morphology or changes in subcellu-
the protein is involved in a particular biological process. lar localization of marker proteins67.
The availability of whole-genome DNA sequences Small molecules or peptides discovered in pheno-
offers the possibility of doing target-based screens type-based screens can be used to identify proteins reg-
against all members of a protein family. Where whole- ulating a specific phenotype. The discovery of cal-
genome sequences are available, it is possible to screen cineurin as the target of the immunosuppressant FK506
all test compounds against each member and so identify (FIG. 2f) in 1987 represents the first successful application
specific inhibitors or activators of each family member. of the three steps of the chemical-genetic process. Before
An example of this strategy was recently reported in the discovery of this small molecule, the structurally
which a COMBINATORIAL LIBRARY was screened against five unrelated cyclic peptide cyclosporin A (CsA) (FIG. 2g)
subtypes of somatostatin receptors60. Specific agonists was used to treat solid-organ transplant rejection. CsA
were identified for each receptor subtype, and these ago- was known to inhibit the production of T-cell-derived
nists were used to identify downstream signalling events cytokines such as interleukin-2 (IL-2). Using a pheno-
COMBINATORIAL LIBRARY
A collection of small molecules
activated by each receptor60. For example, somatostatin type-based screen, Kino et al. tested fermentation broth
synthesized by combinatorial receptor subtype-2 activation led to inhibition of extracts from Streptomyces for their ability to pheno-
chemistry. glucagon release from mouse pancreatic -cells, where- copy CsA by blocking IL-2 production68,69. This screen

NATURE REVIEWS | GENETICS VOLUME 1 | NOVEMBER 2000 | 1 2 1


2000 Macmillan Magazines Ltd
REVIEWS

ALLELIC SCANNING resulted in the discovery of FK506, and subsequent a recent study, the induction of metal-ion-responsive
The rapid detection of DNA studies placed the molecular target of FK506, cal- genes by a small molecule revealed that the direct targets
sequence variations between cineurin, in the T-cell receptor-initiated signalling path- of the small molecule were specific metal ions66.
two strains by analysing total
way. These results led to the elucidation of the molecu- In a related approach, the pattern of gene expression
genomic DNA on high-density
DNA microarrays. lar events that govern this membrane-to-nucleus changes observed are treated as a molecular signature or
signalling pathway. More recent phenotype-based fingerprint of the compound. Pattern-matching algo-
screens have yielded reagents that block mitotic progres- rithms are then used to determine whether a small mol-
sion61,62,70, induce or suppress cell-cycle arrest11,66,71, pre- ecule has the same molecular signature as a gene dele-
vent mating-pheromone-induced cell-cycle arrest23,72, tion. For example, a recent study found that the pattern
prevent interleukin-6 (IL-6) secretion73, induce apopto- of gene expression changes that occurred in response to
sis74 and prevent endothelial cell activation75. treatment of yeast cells with the anaesthetic dyclonine
was the same as that which occurred in response to
Target identification. Since Paul Ehrlichs invention of deletion of the ERG2 gene. So it was revealed that the
the protein receptor concept, it has been understood protein target of dyclonine was Erg2 (REF. 88).
that small molecules exert their effects on biological sys- DNA arrays can also be used to determine those
tems by interacting with specific protein targets76. yeast strains that grow under selective pressure, such as
Despite this early insight, identifying the protein target in the presence of a small molecule in a culture contain-
of a specific small molecule was not feasible until meth- ing many different yeast strains. This is a particularly
ods were invented for purifying and sequencing pro- powerful approach if the pool of yeast strains consists of
teins and cloning their cognate genes. Perhaps the first a large number of strains, each with a different gene
example of protein target identification was the covalent deletion. A yeast strain bearing a deletion in the gene
labelling77, peptide sequencing78 and subsequent that encodes the protein target of a small molecule will
cloning79 of the acetylcholine (ACh) receptor. A similar not respond to it. Furthermore, although many genes in
strategy was used to identify the intracellular receptors yeast are haplosufficient, the gene encoding the protein
for various steroid hormones80. In these studies, ACh target of a small molecule may effectively be rendered
and steroid hormones were used to label their respective haploinsuffficient by the presence of the small molecule.
protein targets, allowing the proteins to be purified from Therefore, in a recent study, the known targets of six
cellular extracts. This process became a powerful small molecules were verified using DNA microarrays,
method for identifying protein targets of small mole- by determining those deletion strains that showed com-
cules, and was used successfully to identify the targets of pound-induced haploinsufficiency89.
many natural products, including FK506 (REF. 81), Finally, DNA arrays can be used for rapid ALLELIC
cyclosporin81, rapamycin82, trapoxin83, fumagillin84, SCANNING and for mapping loci of interest. In this for-
depudecin85 and lactacystin86. ward-genetic approach, genetic loci responsible for a
Although biochemical purification of protein targets specific phenotypic trait are identified. If the pheno-
has been successfully used for many natural products, typic trait is susceptibility or resistance to an exoge-
the method is laborious, difficult and often unreliable. nous ligand, such as a small molecule, then this genet-
As a result, new methods for rapid and efficient protein- ic-mapping strategy will identify the locus that
target identification are now being developed87. An encodes the target of the ligand. This strategy was
exciting new tool that can be applied towards this goal is recently applied in S. cerevisiae to map a drug resis-
DNA microarray technology. In one approach, the tance locus for cycloheximide by comparing a sensi-
changes in gene expression patterns that occur in tive strain with a resistant strain90.
response to treatment of cells with a small molecule are An alternative method of identifying ligand-interact-
determined. This pattern of gene expression may reveal ing proteins is the yeast two-hybrid system91 a yeast
specific gene expression changes that might reflect or transcription-based system that measures the interac-
explain the activity of a small molecule. For example, in tion of a test protein with many other proteins. In the
three-hybrid variant of this system, a DNA-binding
domain is fused to the protein target of a known ligand,
Box 2 | Forward chemical-genetics: a case study which is then covalently linked to a small molecule of
interest. The small molecule of interest is displayed
In an analysis of the spindle checkpoint61, a library of peptides containing 16 random
against a library of proteins fused to a transcriptional
amino acids was displayed on a scaffold protein (a staphylococcal nuclease) and was
expressed in yeast cells. In a screen for suppressors of the spindle checkpoint, three
activation domain. If the small molecule is capable of
peptides that suppressed the Mps1 (monopolar spindle 1)-induced activation of this binding to one such protein, the activation domain will
checkpoint were identified from 6.5 million screened peptides. One peptide also be recruited to the small molecule and the DNA-bind-
suppressed spindle checkpoint arrest induced by depolymerization of the mitotic ing domain, resulting in activation of a reporter
spindle and by the introduction of short linear minichromosomes, indicating that it gene47,92. The simpler two-hybrid approach has been
was a general inhibitor of the spindle checkpoint. Subsequently, the protein target of quite useful in determining interacting proteins for pep-
this peptide was determined. Using a two-hybrid analysis, the peptide suppressor of the tide aptamers. In one case, peptide aptamers that
spindle checkpoint was found to interact with the product of YDR517W, a yeast open blocked mating pheromone-induced arrest in S. cere-
reading frame of unknown function. Subsequent deletion of YDR517W revealed that visiae were used in a series of two-hybrid screens to
this gene is required for a functional spindle checkpoint, validating this forward identify interacting proteins23. Numerous targets were
chemical-genetic screen. identified that had not previously been linked to

122 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics


2000 Macmillan Magazines Ltd
REVIEWS

pheromone-induced cell-cycle arrest, including cell wall out of an entire genome that regulate a particular bio-
biosynthesis kinase 1 (Cbk1)23. So in this study, a chemi- logical process. Such screens have been enormously
cal-genetic approach was successfully used to identify powerful in dissecting the molecular events governing
proteins involved in the pheromone responsive sig- the behaviour of lower organisms. Much effort is now
nalling pathway. being devoted to developing technologies that allow
genetic-like screens to be done in mammalian systems.
Validation and specificity. Once a small molecule or The utility of mutations in a classical genetic
peptide is selected in either a forward or reverse chemi- approach derives from their specificity and their broad
cal-genetic screen, the specificity of this ligand for its generality. In other words, a mutation can specifically
protein target must be assessed and its ability to selec- alter the function of a single gene within a complex
tively perturb the protein target validated. Specificity genome, but because mutations can be made in any
can be assessed initially by testing whether the ligand gene, mutations are generally applicable and can be
binds to or inhibits the function of other related pro- used to study nearly any biological process. The use of
teins in vitro (BOX 1)60. With emerging genomic tech- exogenous ligands in the chemical-genetic approach in
nologies, it will probably become possible to test the place of mutations requires that such ligands be like-
ability of a ligand to bind to every protein in an organ- wise both general and specific. The former criterion
ism. However, the fact that a ligand does not bind to seems likely to be met with small-molecule and pep-
proteins other than a candidate protein, even in such tide ligands, as such ligands can be found for nearly
global surveys, cannot be used to conclusively prove its any protein. The latter criterion might also be met,
specificity because a ligand might bind to a because in some cases even very simple small mole-
proteinprotein interface or a conformational epitope cules can interact specifically with a single protein tar-
in vivo that is not found in vitro. get93. Therefore, it is possible that exogenous ligands
The ultimate test of the specificity of a ligand must can be both generally applicable and specific in their
occur in a cellular context, where all possible target mol- mechanism of action.
ecules are present in their native conformations. Recent It is, however, difficult to screen exogenous ligands in
advances in EXPRESSION PROFILING offer new avenues for whole animals, such as mice, because of the difficulty
determining the specificity of a ligand in a cellular con- and expense of testing numerous ligands in animals.
text. One test for specificity is done using reverse-genetic Furthermore, determining the target specificity of an
methods to delete the gene for a candidate target of a exogenous ligand is laborious, although essential to the
ligand. If a ligand has an effect in cells that do not chemical-genetic process. Future advances in this area
express a candidate target protein, there must be a sec- will probably involve improvements in the ease and
ondary target for this ligand. Norman et al.61 observed speed of target identification and specificity assessment.
such a result for a peptide ligand discovered in a forward In addition, future studies may reveal fundamental rela-
chemical-genetic screen that aimed at identifying sup- tionships between chemical structure and target speci-
pressors of the spindle assembly checkpoint (BOX 2). ficity. Knowledge of such relationships will greatly aid
Moreover, the global pattern of changes in gene the design of future chemical libraries.
expression occurring in response to a ligand can be used Finally, the chemical-genetic approach requires three
to create an expression profile for the ligand93. By com- critical technologies: first, diverse chemical or peptide
paring such an expression profile with the expression libraries; second, high-throughput screening; and last,
profiles obtained after deleting a candidate gene, it is protein-target identification. Early efforts using these
possible to determine whether a given ligand is targeting technologies focused on drug development, rather than
one or more proteins. For example, Marton et al. mea- on the dissection of biological systems. However, these
sured the expression profile of 3-aminotriazole, an technologies have become sufficiently robust and wide-
inhibitor of His3, in wild-type yeast cells and found spread that we can now conceive of using them to study
more than 1,000 changes in gene expression93. The virtually any biological process. Such an approach
expression profile of HIS3-deleted cells had a nearly promises to provide a generally useful, rapid and condi-
identical pattern of changes in gene expression, indicat- tional method of altering protein function in mam-
ing that His3 is probably the only protein target of 3- malian systems. It is possible that chemical genetics will
aminotriazole93. Furthermore, it is possible to measure prove to be as revealing in the century to come as genet-
the expression profile of a small molecule in cells delet- ics has been in the previous century.
ed for the primary target to determine whether there are
transcriptional changes caused as a result of binding to
other proteins. Using this logic, Marton et al. measured Links
the expression profile of 3-aminotriazole in HIS3-delet- DATABASE INFORMATION | p53 | MEK1 | somatostatin |
ed cells and found no changes in gene expression, indi- somatostatin receptor subtype-2 | somatostatin receptor
cating that 3-aminotriazole is a specific inhibitor of His3 subtype-5 | TGF- | plasminogen activator inhibitor type
in yeast cells93. 1 | interleukin-2 | interleukin-6
EXPRESSION PROFILING
FURTHER INFORMATION | The Gram stain | Paul Ehrlich |
The use of DNA microarrays to
determine the expression level Conclusion Animations of reporter-gene assays and antibody-based
of thousands of genes The power of classic genetic screens stems from their cellular immunoassays | Saccharomyces genome deletion
simultaneously. ability to select, in an unbiased manner, the few genes project | Field lab page | Stockwell lab page

NATURE REVIEWS | GENETICS VOLUME 1 | NOVEMBER 2000 | 1 2 3


2000 Macmillan Magazines Ltd
REVIEWS

1. Hartwell, L. H. Twenty-five years of cell cycle genetics. References 29 and 30 describe split-pool synthesis, convertases 1 and 2 using a peptide combinatorial library.
Genetics 4, 975980 (1991). a revolutionary method for synthesizing very large J. Biol. Chem. 273, 2658926595 (1998).
2. Metzstein, M. M., Stanfield, G. M. & Horvitz, H. R. chemical libraries. 55. Tian, S. S. et al. A small, nonpeptidyl mimic of granulocyte-
Genetics of programmed cell death in C. elegans: past, 31. Fodor, S. P. et al. Light-directed, spatially addressable colony-stimulating factor. Science 281, 257259 (1998).
present and future. Trends Genet. 14, 410416 (1998). parallel chemical synthesis. Science 251, 767773 (1991). 56. Zhang, B. et al. Discovery of a small molecule insulin
3. Nusslein-Volhard, C. & Wieschaus, E. Mutations affecting 32. Scott, J. K. & Smith, G. P. Searching for peptide ligands mimetic with antidiabetic activity in mice. Science 284,
segment number and polarity in Drosophila. Nature 287, with an epitope library. Science 249, 386390 (1990). 974977 (1999).
795801 (1980). 33. Colas, P. et al. Genetic selection of peptide aptamers that 57. Chen, J. K., Lane, W. S., Brauer, A. W., Tanaka, A. &
4. Justice, M. J., Noveroske, J. K., Weber, J. S., Zheng, B. & recognize and inhibit cyclin-dependent kinase 2. Nature Schreiber, S. L. Biased combinatorial libraries: Novel
Bradley, A. Mouse ENU mutagenesis. Hum. Mol. Genet. 8, 380, 548550 (1996). ligands for the SH3 domain of phosphatidylinositol 3-
19551963 (1999). 34. Lien, E. J. & Wang, P. H. Lipophilicity, molecular weight, kinase. J. Am. Chem. Soc. 115, 1259112592 (1993).
5. Mullins, M. C., Hammerschmidt, M., Haffter, P. & Nusslein- and drug action: reexamination of parabolic and bilinear 58. Bottger, V. et al. Identification of novel mdm2 binding
Volhard, C. Large-scale mutagenesis in the zebrafish: in models. J. Pharmacol. Sci. 69, 648650 (1980). peptides by phage display. Oncogene 13, 21412147
search of genes controlling development in a vertebrate. 35. Tan, D. S., Foley, M. A., Stockwell, B. R., Shair, M. D. & (1996).
Curr. Biol. 4, 189201 (1994). Schreiber, S. L. Synthesis and preliminary evaluation of a 59. Norris, J. D. et al. Peptide antagonists of the human
6. Vitaterna, M. H. et al. Mutagenesis and mapping of a library of polycyclic small molecules for use in chemical estrogen receptor. Science 285, 744746 (1999).
mouse gene, Clock, essential for circadian behavior. genetic assays. J. Am. Chem. Soc. 121, 90739087 60. Rohrer, S. P. et al. Rapid identification of subtype-selective
Science 264, 719725 (1994). (1999). agonists of the somatostatin receptor through
This study was the first to successfully use random 36. Sullivan, R. W. et al. 2-Chloro-4-(trifluoromethyl)pyrimidine- combinatorial chemistry. Science 282, 737740 (1998).
mutagenesis, phenotype-based screening and 5-N-(3,5-bis(trifluoromethyl)phenyl)-carboxamide: a These authors were the first to apply reverse
positional cloning (that is, a forward-genetic potent inhibitor of NF-B- and AP-1-mediated gene chemical-genetics to a protein family. They identify
approach) in mice. expression identified using solution-phase combinatorial specific small molecule partners of each member of
7. Reppert, S. M. & Weaver, D. R. Forward genetic approach chemistry. J. Med. Chem. 41, 413419 (1998). a family of somatostatin cell-surface receptors.
strikes gold: cloning of a mammalian Clock gene. Cell 89, 37. An, H., Haly, B. D. & Cook, P. D. Discovery of novel 61. Norman, T. C. et al. Genetic selection of peptide inhibitors
487490 (1997). pyridinopolyamines with potent antimicrobial activity: of biological pathways. Science 285, 591595 (1999).
8. Munroe, R. J. et al. Mouse mutants from chemically deconvolution of mixtures synthesized by solution-phase 62. Mayer, T. U. et al. Small molecule inhibitor of mitotic
mutagenized embryonic stem cells. Nature Genet. 24, combinatorial chemistry. J. Med. Chem. 41, 706716 spindle bipolarity identified in a phenotype-based screen.
318321 (2000). (1998). Science 286, 971974 (1999).
9. Schena, M. et al. Parallel human genome analysis: 38. Boger, D. L., Chai, W. & Jin, Q. Multistep convergent 63. Rosania, G. R. et al. Myoseverin, a microtubule-binding
microarray-based expression monitoring of 1000 genes. solution-phase combinatorial synthesis and deletion molecule with novel cellular effects. Nature Biotechnol. 18,
Proc. Natl Acad. Sci. USA 93, 1061410619 (1996). synthesis deconvolution. J. Am. Chem. Soc. 120, 304308 (2000).
10. Schreiber, S. L. Chemical genetics resulting from a passion 72207225 (1998). 64. Hrabe de Angelis, M. H. et al. Genome-wide, large-scale
for synthetic organic chemistry. Bioorg. Med. Chem. 6, 39. Hung, D. T., Jamison, T. F. & Schreiber, S. L. production of mutant mice by ENU mutagenesis. Nature
11271152 (1998). Understanding and controlling the cell cycle with natural Genet. 25, 444447 (2000).
11. Stockwell, B. R., Haggarty, S. J. & Schreiber, S. L. High- products. Chem. Biol. 3, 623640 (1996). 65. Nolan, P. M. et al. A systematic, genome-wide,
throughput screening of small molecules in miniaturized 40. Handelsman, J., Rondon, M. R., Brady, S. F., Clardy, J. & phenotype-driven mutagenesis programme for gene
mammalian cell-based assays involving post-translational Goodman, R. M. Molecular biological access to the function studies in the mouse. Nature Genet. 25, 440443
modifications. Chem. Biol. 6, 7183 (1999). chemistry of unknown soil microbes: a new frontier for (2000).
12. Agrawal, S. & Zhao, Q. Antisense therapeutics. Curr. Opin. natural products. Chem. Biol. 5, 245249 (1998). 66. Stockwell, B. R., Hardwick, J. S., Tong, J. K. & Schreiber,
Chem. Biol. 2, 519528 (1998). The authors describe a powerful method for S. L. Chemical genetic and genomic approaches reveal a
13. Dervan, P. B. & Burli, R. W. Sequence-specific DNA discovering new natural products using bacterial role for copper in specific gene activation. J. Am. Chem.
recognition by polyamides. Curr. Opin. Chem. Biol. 3, genetics. Soc. 121, 1066210663 (1999).
688693 (1999). 41. Tan, D. S., Foley, M. A., Shair, M. D. & Schreiber, S. L. 67. Boland, M. V., Markey, M. K. & Murphy, R. F. Automated
14. Parascandola, J. The theoretical basis of Paul Ehrlichs Stereoselective synthesis of over two million compounds recognition of patterns characteristic of subcellular
chemotherapy. J. Hist. Med. Allied Sci. 36, 1943 (1981). having structural features both reminiscent of natural structures in fluorescence microscopy images. Cytometry
15. Schmitz, R. Friedrich Wilhelm Serturner and the discovery products and compatible with miniaturized cell-based 33, 366375 (1998).
of morphine. Pharm. Hist. 27, 6174 (1985). assays. J. Am. Chem. Soc. 120, 85658566 (1998). 68. Kino, T. et al. FK-506, a novel immunosuppressant
16. Parascandola, J. Reflections on the history of This is the first demonstrated use of combinatorial isolated from a Streptomyces I. Fermentation, isolation,
pharmacology: the 1980 Kremers award address. Pharm. chemistry to synthesize a large library of structurally and physico-chemical and biological characteristics. J.
Hist. 22, 131140 (1980). complex small molecules. Antibiot. (Tokyo) XL, 12491255 (1987).
17. Ehrlich, P. Chemotherapeutics: Scientific principles, 42. Wang, J. & Ramnarayan, K. Toward designing drug-like 69. Kino, T. et al. FK-506, A novel immunosupressant isolated
methods and results. Lancet ii, 445451 (1913). libraries: a novel computational approach for prediction of from a Streptomyces II. Immunosuppressive effect of FK-
18. Travis, A. S. Science as a receptor of technology: Paul drug feasibility of compounds. J. Comb. Chem. 1, 506 in vitro. J. Antibiot. (Tokyo) XL, 12561265 (1987).
Ehrlich and the synthetic dyestuffs industry. Sci. Context 3, 524533 (1999). 70. Haggarty, S. J. et al. Dissecting cellular processes using
383408 (1989). 43. Gellman, S. H. Foldamers: a manifesto. Acc. Chem. Res. small molecules: identification of colchicine-like, taxol-like
19. Travis, A. S. Perkins mauve: ancestor of the organic 31, 173180 (1998). and other small molecules that perturb mitosis. Chem.
chemical industry. Technol. Cult. 31, 5182 (1990). 44. Schreiber, S. L. Target-oriented and diversity-oriented Biol. 7, 275286 (2000).
20. Popescu, A. & Doyle, R. J. The Gram stain after more than organic synthesis in drug discovery. Science 287, 71. Roberge, M. et al. High-throughput assay for G2
a century. Biotech. Histochem. 71, 145151 (1996). 19641969 (2000). checkpoint inhibitors and identification of the structurally
21. McTavish, J. R. Aspirin in Germany: the pharmaceutical 45. Vaino, A. R. & Janda, K. D. Euclidean shape-encoded novel compound isogranulatimide. Cancer Res. 58,
industry and the pharmaceutical profession. Pharm. Hist. combinatorial chemical libraries. Proc. Natl Acad. Sci. USA 57015706 (1998).
29, 103115 (1987). 97, 76927696 (2000). 72. Caponigro, G. et al. Transdominant genetic analysis of a
22. Kaufman, G. B. & Priebe, P. M. The discovery of saccharin: 46. Sundberg, S. A. High-throughput and ultra-high- growth control pathway. Proc. Natl Acad. Sci. USA 95,
a centennial retrospect. Ambix 25, 191207 (1978). throughput screening: solution- and cell-based 75087513 (1998).
23. Geyer, C. R., Colman-Lerner, A. & Brent, R. Mutagenesis approaches. Curr. Opin. Biotechnol. 11, 4753 (2000). 73. Yamaguchi, K. et al. 4-Phenylthiazole derivatives inhibit IL-
by peptide aptamers identifies genetic network members 47. Borchardt, A., Liberles, S. D., Biggar, S. R., Crabtree, G. R. 6 secretion in osteoblastic cells and suppress bone weight
and pathway connections. Proc. Natl Acad. Sci. USA 96, & Schreiber, S. L. Small molecule-dependent genetic loss in ovariectomized mice. Bioorg. Med. Chem. Lett. 9,
85678572 (1999). selection in stochastic nanodroplets as a means of 957960 (1999).
This is the first successful use of peptide aptamers detecting proteinligand interactions on a large scale. 74. Lu, X. P. et al. Novel retinoid-related molecules as
in a forward chemical-genetic screen. Chem. Biol. 4, 961968 (1997). apoptosis inducers and effective inhibitors of human lung
24. Colas, P. Combinatorial protein reagents to manipulate 48. You, A. J., Jackman, R. J., Whitesides, G. M. & Schreiber, cancer cells in vivo. Nature Med. 3, 686690 (1997).
protein function. Curr. Opin. Chem. Biol. 4, 5459 (2000). S. L. A miniaturized arrayed assay format for detecting 75. Rice, J. W., Davis, J. E., Crowl, R. M. & Johnston, P. A.
25. Dolle, R. E. Comprehensive survey of chemical libraries small moleculeprotein interactions in cells. Chem. Biol. 4, Development of a high volume screen to identify inhibitors
yielding enzyme inhibitors, receptor agonists and 969975 (1997). of endothelial cell activation. Anal. Biochem. 241, 254259
antagonists, and other biologically active agents: 1992 49. Mere, L. et al. Miniaturized FRET assays and microfluidics: (1996).
through 1997. Mol. Diversity 3, 199233 (1998). key components for ultra-high-throughput screening. Drug 76. Goldstein, A. The interactions of drugs and plasma
26. Dolle, R. E. & Nelson, K. H., Jr Comprehensive survey of Discov. Today 4, 363369 (1999). proteins. Pharmacol. Rev. 1, 102163 (1949).
combinatorial library synthesis: 1998. J. Comb. Chem. 1, 50. Wetterau, J. R. et al. An MTP inhibitor that normalizes 77. Karlin, A. & Winnik, M. Reduction and specific alkylation of
235282 (1999). atherogenic lipoprotein levels in WHHL rabbits. Science the receptor for acetycholine. Proc. Natl Acad. Sci. USA
27. Merrifield, B. Concept and early development of solid- 282, 751754 (1998). 60, 668674 (1968).
phase peptide synthesis. Methods Enzymol. 289, 313 51. Komarov, P. G. et al. A chemical inhibitor of p53 that 78. Raftery, M. A., Hunkapiller, M. W., Strader, C. D. & Hood,
(1997). protects mice from the side effects of cancer therapy. L. E. Acetylcholine receptor: complex of homologous
28. Schneider, C. H., Rolli, H. & Blaser, K. Liquidliquid Science 285, 17331737 (1999). subunits. Science 208, 14541457 (1980).
extraction in peptide synthesis. Int. J. Pept. Protein Res. 52. Sebolt-Leopold, J. S. et al. Blockade of the MAP kinase 79. Noda, M. et al. Primary structure of -subunit precursor of
15, 411419 (1980). pathway suppresses growth of colon tumors in vivo. Torpedo californica acetylcholine receptor deduced from
29. Lam, K. S. et al. A new type of synthetic peptide library for Nature Med. 5, 810816 (1999). cDNA sequence. Nature 299, 793797 (1982).
identifying ligand-binding activity. Nature 354, 8284 53. Aramburu, J. et al. Affinity-driven peptide selection of an 80. Kornel, L. On the effects and the mechanism of action of
(1991). NFAT inhibitor more selective than cyclosporin A. Science corticosteroids in normal and neoplastic target tissues:
30. Furka, A., Sebestyen, F., Asgedom, M. & Dido, G. General 285, 21292133 (1999). findings and hypotheses. Acta Endocrinol. Suppl.
method for rapid synthesis of multicomponent peptide 54. Apletalina, E., Appel, J., Lamango, N. S., Houghten, R. A. (Copenh.) 178, 145 (1973).
mixtures. Int. J. Pept. Protein Res. 37, 487493 (1991). & Lindberg, I. Identification of inhibitors of prohormone 81. Liu, J. et al. Calcineurin is a common target of

124 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics


2000 Macmillan Magazines Ltd
REVIEWS

cyclophilincyclosporin A and FKBPFK506 complexes. yeast genome. Science 281, 11941197 (1998). 90, 1064310647 (1993).
Cell 66, 807815 (1991). 91. Colas, P. & Brent, R. The impact of two-hybrid and related 100. Yanofsky, S. D. et al. High affinity type I interleukin 1
82. Brown, E. J. et al. & Schreiber, S. L. A mammalian protein methods on biotechnology. Trends Biotechnol. 16, receptor antagonists discovered by screening recombinant
targeted by G1-arresting rapamycinreceptor complex. 355363 (1998). peptide libraries. Proc. Natl Acad. Sci. USA 93,
Nature 369, 756758 (1994). 92. Licitra, E. J. & Liu, J. O. A three-hybrid system for 73817386 (1996).
83. Taunton, J., Hassig, C. A. & Schreiber, S. L. A mammalian detecting small ligandprotein receptor interactions. Proc. 101. Maly, D. J., Choong, I. C. & Ellman, J. A. Combinatorial
histone deacetylase related to the yeast transcriptional Natl Acad. Sci. USA 93, 1281712821 target-guided ligand assembly: identification of potent
regulator Rpd3p. Science 272, 408411 (1996). (1996). subtype-selective c-Src inhibitors. Proc. Natl Acad. Sci.
The authors demonstrate the power of affinity 93. Marton, M. J., et al. Drug target validation and USA 97, 24192924 (2000).
chromatography and biochemical purification to identification of secondary drug target effects using DNA 102. Richter, L. S. & Moos, W. H. Discovery of nanomolar
identify the protein target of the natural product microarrays. Nature Med. 4, 12931301 (1998). ligands for 7-transmembrane G-protein-coupled receptors
trapoxin. The authors demonstrate the utility of expression from a diverse N-(substituted)glycine peptoid library.
84. Sin, N. et al. The anti-angiogenic agent fumagillin profiling for identifying the target(s) of a small J. Med. Chem. 37, 26782685 (1994).
covalently binds and inhibits the methionine molecule. 103. Liang, R. et al. Parallel synthesis and screening of a solid
aminopeptidase MetAP-2. Proc. Natl Acad. Sci. USA 94, 94. Giroux, S. et al. Embryonic death of Mek1deficient mice phase carbohydrate library. Science 274, 15201522
60996103 (1997). reveals a role for this kinase in angiogenesis in the (1996).
85. Kwon, H. J., Owa, T., Hassig, C. A., Shimada, J. & labyrinthine region of the placenta. Curr. Biol. 9, 369372 104. Szardenings, A. K. et al. Identification of highly selective
Schreiber, S. L. Depudecin induces morphological (1999). inhibitors of collagenase-1 from combinatorial libraries of
reversion of transformed fibroblasts via the inhibition of 95. Blum, J. H., Dove, S. L., Hochschild, A. & Mekalanos, J. J. diketopiperazines. J. Med. Chem. 42, 13481357
histone deacetylase. Proc. Natl Acad. Sci. USA 95, Isolation of peptide aptamers that inhibit intracellular (1999).
33563361 (1998). processes. Proc. Natl Acad. Sci. USA 97, 22412246 105. Bunin, B. A. & Ellman, J. A. A general and expedient
86. Fenteany, G. et al. Inhibition of proteasome activities and (2000). method for the solid phase synthesis of 1,4-
subunit-specific amino-terminal threonine modification by 96. Adey, N. B. & Kay, B. K. Identification of calmodulin- benzodiazepine derivatives. J. Am. Chem. Soc. 114,
lactacystin. Science 268, 726731 (1995). binding peptide consensus sequences from a phage- 1099710998 (1992).
87. Sche, P. P., McKenzie, K. M., White, J. D. & Austin, D. J. displayed random peptide library. Gene 169, 133134 106. Peisach, E. et al. Interaction of a peptidomimetic
Display cloning: functional identification of natural product (1996). aminimide inhibitor with elastase. Science 269, 6669
receptors using cDNA-phage display. Chem. Biol. 6, 97. Wrighton, N. C. et al. Small peptides as potent mimetics of (1995).
707716 (1999). the protein hormone erythropoietin. Science 273, 458463 107. Bailey, N. et al. A convenient procedure for the solution
88. Hughes, T. R. et al. Functional discovery via a (1996). phase preparation of 2- aminothiazole combinatorial
compendium of expression profiles. Cell 102, 109126 98. Wrighton, N. C., Barrett, R. W. & Dower, W. J. Peptide libraries. Bioorg. Med. Chem. Lett. 6, 14091414
(2000). agonist of the thrombopoietin receptor as potent as the (1996).
89. Giaever, G. et al. Genomic profiling of drug sensitivities via natural cytokine. Science 276, 16961699 (1997).
induced haploinsufficiency. Nature Genet. 21, 278283 99. Yayon, A. et al. Isolation of peptides that inhibit binding of Acknowledgements
(1999). basic fibroblast growth factor to its receptor from a The author thanks J. M. Scharf and M. S. Stockwell for critical
90. Winzeler, E. A. et al. Direct allelic variation scanning of the random phage-epitope library. Proc. Natl Acad. Sci. USA reading of the manuscript.

NATURE REVIEWS | GENETICS VOLUME 1 | NOVEMBER 2000 | 1 2 5


2000 Macmillan Magazines Ltd