Doi:10.1016/j.cbpa.2006.06.016

Chemical biology and bacteria: not simply a matter oflife or deathDeborah T and Eric J Rubin Chemical biological approaches to understanding bacteria small molecules as tools to systematically dissect the have largely been confined to screening for antibiotics. More pathways involved in these complex phenotypes, culmi- complex phenotypes, such as virulence, have largely been nating in the generation of a new name for an old studied using bacterial genetics. However, it has recently become clear that these two methods are complementary andthat combining chemical biologic and genetic approaches to The field of chemical genetics has predominantly fallen studying bacteria brings new power to old problems.
within the domain of eukaryotic biology, in part for historical reasons. Some of the earliest demonstrations 1 Department of Microbiology and Molecular Genetics, Harvard of the power of chemical genetics resulted in important contributions to eukaryotic biology, using both specific 2 Department of Immunology and Infectious Disease, Harvard natural products as well as small molecules identified in high-throughput chemical screens. Examples include one Division of Infectious Disease, Brigham and Women’s Hospital, USA of the earliest applications of chemical biology, using colchicine to identify tubulin as well as more recentexamples. These include natural products FK506, cyclos-porine and rapamycin, which have helped to elucidate Current Opinion in Chemical Biology 2006, 10:321–326 immune signaling pathways [], and monasterol [] and blebbistatin found in high-throughput screens, which have helped to elucidate steps in mitosis and cytokinesis Edited by Clifton E Barry III and Alex Matter By contrast, in prokaryotic systems, much of chemical biology has been relegated to the arena of antibioticdevelopment, a life or death phenotype.
1367-5931/$ – see front matter# 2006 Elsevier Ltd. All rights reserved.
Chemical genetics has also been particularly applicable in eukaryotic systems because, until recently, targetedgenetic methods have been largely lacking. In contrast,prokaryotic systems have a rich tradition of classical and molecular genetics with diverse phenotypes often easily The concept of chemical biology is old, dating back generated through manipulation at the genetic level.
billions of years. Nature has long exploited the ability Despite the power of classical bacterial genetics however, of small organic molecules to regulate cellular processes.
we would argue that chemical approaches can be extre- From steroids acting as hormones in eukaryotic systems mely valuable in the study of diverse and complex bac- to quorum sensing molecules in prokaryotic commu- terial processes. In fact, the potential for chemical genetic nities these small-molecule effectors enter the cell to complement classical genetics (and genomics) may and modulate gene and protein expression and function.
well make prokaryotic biology an optimal domain forsmall-molecule approaches ().
This same phenomenon has also been artificiallyexploited by modern medicine with the recognition that small molecules can be used to control cell phenotypes Chemical genetics generally refers to the use of small that impact human disease. With the first use of nitrogen molecules to conditionally perturb gene function, result- mustards as anti-neoplastic, alkylating agents in 1942 and ing in the generation of phenotypic changes In con- Alexander Fleming’s discovery of penicillin as an anti- trast to classical genetics where manipulation occurs at biotic in 1928 came the recognition that the phenotype of the DNA level, small molecules more typically modulate cell death can be conditionally generated by small mole- protein function by inducing conformational changes or competing for naturally occurring protein–ligand or pro-tein–protein interaction sites, resulting in altered activity As our understanding of cell biology in both eukaryotes ). Although previously the repertoire of small and prokaryotes has grown more sophisticated, we have molecules that were known to function in this manner was come to realize that a multitude of phenotypes exist that relatively small, the advent of high-throughput screening are far more subtle than simply alive or dead. The past of large chemical libraries has provided a new opportunity decade has seen a marked acceleration of interest in using Current Opinion in Chemical Biology 2006, 10:321–326 Comparison of classical and chemical genetics Slow (except for conditionally stable proteins) Useful for studying proteins essential for in vitro survival By analogy to classical genetics, both forward and reverse expensive and more robust, having been performed in genetic strategies may be employed in high-throughput diverse organisms including Yersinia, Vibrio cholerae, Sta- chemical screens to find small molecules of interest phylococcus aureus and Escherichia coli [].
(). In the reverse approach, small-moleculelibraries are screened against purified protein targets.
In addition, the marriage of chemical and classical genet- Interesting small-molecule candidates can then be used ics in many bacterial systems can facilitate understanding to study mechanistic aspects of the protein or potentially of the mechanism of action of the small molecule. An to generate phenotypes related to that particular protein example of this is the use of classical genetics to under- target if it is sufficiently cell-permeable and potent. In stand the mechanisms of new vancomycin analogues by this approach, there is relatively little distinction between generating resistant mutants ]. This example shows eukaryotic and prokaryotic systems [].
that generation of random or transposon mutants, orcomplementation with cosmid or expression libraries to In contrast, a forward approach screens for phenotypes of generate resistance to a given small molecule are easily interest. The challenge, then, is to identify the protein attainable goals with the help of conventional genetics.
targets of the small-molecule candidates as a means of The inherently smaller size of bacterial genomes relative identifying regulators of a given phenotype. It is in this to eukaryotic ones (specifically, mammalian genomes) type of approach that the opportunities in eukaryotic and also allows straightforward systematic approaches to iden- tifying and understanding cellular targets of small mole-cules (For example, arrayed libraries of bacteria that knock-out knock down ] or overexpress There are several reasons why bacteria are well suited for every opening reading frame (ORF) can be used to assay chemical genetic approaches. The design of phenotypic compounds and quickly determine their effects. In eukar- screens in bacteria is typically easier and the assays less yotes, however, such systems are far more difficult to use Genetic approach to obtaining conditional phenotypes.
Current Opinion in Chemical Biology 2006, 10:321–326 Chemical biology and bacteria: not simply a matter of life or death Hung and Rubin proteins whose function is lost at high temperature. Suchmutants may be problematic as they grow only underrestricted conditions that can result in numerous unre-lated changes to the cell, and the mutant proteins oftenlack full activity under permissive conditions. Moreimportantly, temperature-sensitive mutants can only beisolated by a relatively laborious process that is onlysuccessful for about a third of known essential genes [ The limitation of these types of traditional mutants is thatthey are not truly conditional. One cannot turn the geneproducts on and off at will on a short time scale, in atunable manner that is also reversible. This limitation hasseveral implications. Essential genes are difficult to studybecause it is difficult to isolate mutations that result inlethality. Generation of mutations on the DNA leveloften results in compensatory changes such as up-regula- Forward and reverse approaches to using chemical genetics.
tion of other related genes that confound or distort thephenotype related to the single gene.
and, thus far, are limited to a small number of species].
Enter chemical genetics, a method that targets proteinson a rapid time scale with the addition of the small molecule, and in a reversible manner with washout of the small molecule, and can be fine-tuned to inhibit one Traditional bacterial genetics provides two distinct sets of but not another domain of a given protein. This approach tools: mutations that alter a gene product’s function, and is particularly useful in the study of bacteria for several mutations that affect the amount of a gene product.
reasons. It allows the study of wild-type bacteria that Mutations that alter function (i.e., point mutations) can have no deleterious mutations. Small molecules can be result in loss or gain of function or, in the case of enzymes, used to cause immediate changes in function, an attrac- changes in Kd, kcat, or substrate specificity.
tive feature in circumstances where protein or RNA half-lives are prolonged (for example, in non-dividing bacteria Mutations that affect the amount of a gene product fall such as Mycobacterium tuberculosis). Lastly, small mole- into two classes: those that affect RNA abundance and cules can be used to ‘conditionally’ probe the functions of translation, and those that affect protein stability. In many pathogens during infection within the context of the host, or bacterial systems, genetic techniques exist to make dele- of bacteria as they interact in complex communities in tions in genes, either in a directed fashion or randomly their natural ecological niche. Examples show how small through transposon mutagenesis. In what is essentially a molecules can be used to study the in vivo requirements chemical genetics approach, regulated promoters can be for quorum sensing in an S. aureus abscess model [] used to inducibly express genes at different levels ].
and for virulence regulators in a V. cholerae intestinal It is far more difficult to reliably construct proteins that colonization model []. With the increasing recogni- are conditionally stable. Most such mutations result in tion of the limitations of studying bacteria in vitro outsideof the context of their natural environment, the devel-opment of tools to study bacteria in their natural envir- onment will be vital to the next phase of microbiologicaldiscovery.
While the marriage between chemical genetics and pro- karyotic biology holds great potential, several challenges must be met to make these approaches widely applicable.
These challenges include systematic methods for identi- fying protein targets, specificity and potency of small molecule inhibitors, and development of appropriate chemical libraries suited for prokaryotic systems.
Target identification is a major challenge in forward genetic, phenotypic screens. While target identification Current Opinion in Chemical Biology 2006, 10:321–326 is not always critical, it is often important for further work.
products or ‘natural product-like’, with the fluoroquino- Methods identifying targets have been previously lones and linezolid as the marked synthetic exceptions.
reviewed [and include total sequencing of resistantmutants in M. tuberculosis ], transposon mutagenesis Is nature trying to teach us something? Unlike humans [in suppressor and enhancer screens, and affinity- who are perhaps the accidental tourist in the biosphere, based assays such as three-hybrid systems [], bacteria have existed millions of years in the environ- changes in 2-D electrophoretic mobility ], affinity ment, co-evolving with other organisms that produce purification ], and biotinylation with or without cross- antibacterial, natural products (for example, Streptomyces linking []. The rapidly expanding repertoire of genomic strains, marine organisms, plants, fungi, and other bac- and proteomic tools is likely to have a significant impact teria [This evolution of natural products most on target identification, with the ability to match gene certainly did not occur in the setting of selection pressure expression patterns ] or assay direct binding in protein from humans, but more likely from bacteria. In the words chips ]. However, even in the absence of direct target of Samuel Danishefsky, ‘‘There are major teachings in identification, small-molecule candidates can serve as these natural products that we would do well to consider.
tools to understand biology through indirect means.
They may be reflecting eons of wisdom and refinement’’ For example, because they cause conditional blocks, other regulators of phenotype can be identified by geneticsuppressor or enhancer screens, even in the absence of Because of the co-evolution of natural products with specific target identification. In the case of brefeldin, the bacterial targets, even more than eukaryotic ones, natural compound was used to study protein secretion long product or ‘natural product-like’ libraries may be optimal before its actual target was identified ].
for screening bacterial systems. Such libraries can beconstructed using strategies such as diversity-oriented Compounds that are used as drugs must generally be synthesis (DOS) ]. Whether these DOS libraries are specific and highly avid. However, even agents with less generated based on a natural product scaffold [or than optimal properties can be quite useful experimen- generating some configuration of ‘natural product-like’ tally. Many probes and ‘proof of concept’ inhibitors can be elements ], or a hybrid of the two [these libraries used to preliminarily define phenotypes and validate will probably be critical to the success of the marriage of targets without having a high binding affinity ].
chemical genetics and prokaryotic biology.
These can then serve as starting points for further experi-ments and, possibly, chemical optimization.
ConclusionIn the end, does chemical genetics have any relevance to The last major challenge is that existing compound therapeutics and drug development? Is it merely an libraries are not ideally suited for use in bacterial chemical academic exercise that, at best, generates powerful tools genetic screens. Much debate exists over the two types of for biological investigation and, at worst, results in a libraries that currently exist: ‘drug-like’ and ‘natural pro- collection of ‘hits’ relegated to some shelf in a laboratory? duct-like’ libraries. Which is better for screening prokar- We believe that one of the great strengths of chemical yotic systems? Large commercial compound libraries are genetics is exactly that it is based on small molecules and available that have been generated to fit a set of physi- thus a step closer to drug development than alternative cochemical criteria that describe existing drugs, creating a genetic approaches. This does not imply that every hit set of ‘drug-like’ molecules [The difficulty with these obtained in such chemical screens is a candidate drug.
collections in application to bacteria is twofold. First, Though all hits have the potential to be so, it would be although libraries contain enormous numbers of com- naı¨ve to have such expectations. Instead, we would argue pounds, many of these chemicals were originally synthe- that every hit can serve another important function.
sized with specific targets in mind (typically eukaryoticenzymes), and they are therefore biased toward a small Much effort is currently invested in identifying essential number of enzymes, such as protein kinase inhibitors.
genes in bacteria using classical genetics and genomics.
Thus, these libraries actually have limited diversity.
These genes are then designated as ‘potential drug tar-gets’, without any true sense of how ‘drug-targetable’ Secondly, current compound libraries consist largely of they are. Although essential gene products may appear to chemicals that have ‘drug-like properties’ as defined for be good targets, partial inhibition by a compound may not eukaryotic systems [In contrast, most known anti- result in death. In addition, the function of a putative biotics violate these properties, with higher molecular target might overlap enough with that of another protein weights, greater rigidity and fewer degrees of freedom, so that inhibition might produce little or no phenotypic more stereogenic centers, larger or more complex ring change. The not infrequent failure of candidates identi- scaffolds, fewer nitrogen, sulfur and halogen atoms with fied in reverse chemical genetic screens (which target a more oxygen atoms, and more hydrogen bonding ele- preselected gene product) to generate the desired phe- ments [In fact, most current antibiotics are natural notype is testament to the dangers of such an approach, Current Opinion in Chemical Biology 2006, 10:321–326 Chemical biology and bacteria: not simply a matter of life or death Hung and Rubin exacerbated by the uncertainty of the cell-permeability of 10. Kauppi AM, Nordfelth R, Uvell H, Wolf-Watz H, Elofsson M: Targeting bacterial virulence: inhibitors of type III secretion in yersinia. Chem Biol 2003, 10:241-249.
11. Hung DT, Shakhnovich EA, Pierson E, Mekalanos JJ: Small- In contrast, a forward chemical genetic approach guaran- molecule inhibitor of Vibrio cholerae virulence and intestinal tees that hits define good drug targets. This strategy colonization. Science 2005, 310:670-674.
Describes a chemical genetic approach to virulence expression in Vibrio allows the small molecule to identify the ‘Achilles heel’ cholerae through a high-throughput chemical screen to identify virstatin, a of the cellular pathway without relying on a priori small-molecule inhibitor of cholera toxin and TCP expression, by inhibi-tion of the transcriptional regulator ToxT, which has efficacy in an assumptions. Thus, this strategy allows us to identify intestinal mouse model for cholera, thus demonstrating the potential and focus on reasonable biological candidates for drug for chemical genetic in vivo studies.
development, rather than to be caught in random, a la 12. Nicolaou KC, Roecker AJ, Barluenga S, Pfefferkorn JA, Cao GQ: carte selection from long lists of essential genes.
Discovery of novel antibacterial agents active againstmethicillin-resistant Staphylococcus aureus fromcombinatorial benzopyran libraries. ChemBioChem 2001, The field of chemical genetics as applied to bacteria is currently wide open. We have the opportunity to reap the 13. Gauthier A, Robertson ML, Lowden M, Ibarra JA, Puente JL, benefits from decades of microbiological discovery that Finlay BB: Transcriptional inhibitor of virulence factors inenteropathogenic Escherichia coli. Antimicrob Agents have taken us far beyond the idea that the only phenotype of interest in bacteria is alive or dead. This approach 14. Nordfelth R, Kauppi AM, Norberg HA, Wolf-Watz H, Elofsson M: should allow us to considerably expand the limited reper- Small-molecule inhibitors specifically targeting type III toire of functions targeted by current antibiotics and secretion. Infect Immun 2005, 73:3104-3114.
provide important weapons to help turn the tide in the Describes small molecules (acetylated salicylates) that inhibit the type IIIsecretion system in Yersinia tuberculosis and Yop secretion, resulting in war of resistance that we are currently losing against 15. Eggert US, Ruiz N, Falcone BV, Branstrom AA, Goldman RC, Silhavy TJ, Kahne D: Genetic basis for activity differences between vancomycin and glycolipid derivatives ofvancomycin. Science 2001, 294:361-364.
DTH is supported by NIH grant AI060708. EJR is supported in part byNIH grant AI51929 and by the Investigators in Pathogenesis of 16. Jacobs MA, Alwood A, Thaipisuttikul I, Spencer D, Haugen E, Infectious Disease Award from the Burroughs Wellcome Fund.
Ernst S, Will O, Kaul R, Raymond C, Levy R et al.: Comprehensivetransposon mutant library of Pseudomonas aeruginosa.
Proc Natl Acad Sci USA 2003, 100:14339-14344.
References and recommended readingPapers of particular interest, published within the annual period of 17. Liberati NT, Urbach JM, Miyata S, Lee DG, Drenkard E, Wu G, Villanueva J, Wei T, Ausubel FM: An ordered, nonredundantlibrary of Pseudomonas aeruginosa strain PA14 transposon insertion mutants. Proc Natl Acad Sci USA 2006, 103:2833-2838.
18. Ji Y, Zhang B, Van SF et al.: Identification of critical staphylococcal genes using conditional phenotypesgenerated by antisense RNA. Science 2001, 293:2266-2269.
Freedman LP: Increasing the complexity of coactivation innuclear receptor signaling. Cell 1999, 97:5-8.
19. Lamesch P, Milstein S, Hao T, Rosenberg J, Li N, Sequerra R, Bosak S, Doucette-Stamm L, Vandenhaute J, Hill DE et al.: Camilli A, Bassler BL: Bacterial small-molecule signaling C. elegans ORFeome version 3.1: increasing the coverage pathways. Science 2006, 311:1113-1116.
of ORFeome resources with improved gene predictions.
Genome Res 2004, 14:2064-2069.
Mitchison TJ: Towards a pharmacological genetics. Chem Biol1994, 1:3-6.
20. Winzeler EA, Shoemaker DD, Astromoff A, Liang H, Anderson K, Andre B, Bangham R, Benito R, Boeke JD, Bussey H et al.: Weisenberg RC, Borisy GG, Taylor EW: The colchicine-binding Functional characterization of the S. cerevisiae genome by protein of mammalian brain and its relation to microtubules.
gene deletion and parallel analysis. Science 1999, 285:901-906.
Schreiber SL, Liu J, Albers MW, Karmacharya R, Koh E, Martin PK, 21. Gossen M, Bujard H: Tight control of gene expression in Rosen MK, Standaert RF, Wandless TJ: Immunophilin-ligand mammalian cells by tetracycline-responsive promoters.
complexes as probes of intracellular signaling pathways.
Proc Natl Acad Sci USA 1992, 89:5547-5551.
22. Guzman LM, Belin D, Carson MJ, Beckwith J: Tight regulation, Mayer TU, Kapoor TM, Haggarty SJ, King RW, Schreiber SL, modulation, and high-level expression by vectors containing Mitchison TJ: Small molecule inhibitor of mitotic spindle the arabinose PBAD promoter. J Bacteriol 1995, 177:4121-4130.
bipolarity identified in a phenotype-based screen. Science1999, 286:971-974.
23. Wong SM, Akerley BJ: Inducible expression system and marker-linked mutagenesis approach for functional genomics Straight AF, Cheung A, Limouze J, Chen I, Westwood NJ, of Haemophilus influenzae. Gene 2003, 316:177-186.
Sellers JR, Mitchison TJ: Dissecting temporal and spatialcontrol of cytokinesis with a myosin II Inhibitor. Science 2003, 24. Wright JS III, Jin R, Novick RP: Transient interference with staphylococcal quorum sensing blocks abscess formation.
Proc Natl Acad Sci USA 2005, 102:1691-1696.
Peterson JR, Mitchison TJ: Small molecules, big impact: a Describes the use of a small-molecule inhibitor (AIP variant) to inhibit history of chemical inhibitors and the cytoskeleton. Chem Biol quorum sensing in an in vivo model of staphylococcal abscess formation and shows the timing requirements for agr (global accessory generegulator) during infection.
Turk BE, Wong TY, Schwarzenbacher R, Jarrell ET, Leppla SH,Collier RJ, Liddington RC, Cantley LC: The structural basis for 25. Tochtrop GP, King RW: Target identification strategies in substrate and inhibitor selectivity of the anthrax lethal factor.
chemical genetics. Comb Chem High Throughput Screen 2004, Current Opinion in Chemical Biology 2006, 10:321–326 26. Andries K, Verhasselt P, Guillemont J, Gohlmann HW, Neefs JM, 35. Morinaga N, Tsai SC, Moss J, Vaughan M: Isolation of a brefeldin Winkler H, Van Gestel J, Timmerman P, Zhu M, Lee E et al.: A-inhibited guanine nucleotide-exchange protein for ADP A diarylquinoline drug active on the ATP synthase of ribosylation factor (ARF) 1 and ARF3 that contains a Sec7-like mycobacterium tuberculosis. Science 2005, 307:223-227.
domain. Proc Natl Acad Sci USA 1996, 93:12856-12860.
Describes the identification of a new drug against Mycobacterium tuber-culosis, with demonstration of rapid killing and clearing in mice, and the 36. Ajay A, Walters WP, Murcko MA: Can we learn to distinguish possible characterization of its target through total genome sequencing of between ‘drug-like’ and ‘nondrug-like’ molecules? J Med 27. Moore JM, Salama NR: Mutational analysis of metronidazole 37. Feher M, Schmidt JM: Property distributions: differences resistance in Helicobacter pylori. Antimicrob Agents Chemother between drugs, natural products, and molecules from combinatorial chemistry. J Chem Inf Comput Sci 2003,43:218-227.
28. Licitra EJ, Liu JO: A three-hybrid system for detecting small ligand-protein receptor interactions. Proc Natl Acad Sci USA 38. Gillespie DE, Brady SF, Bettermann AD, Cianciotto NP, Liles MR, Rondon MR, Clardy J, Goodman RM, Handelsman J: Isolation ofantibiotics turbomycin a and B from a metagenomic library of 29. Becker F, Murthi K, Smith C, Come J, Costa-Roldan N, soil microbial DNA. Appl Environ Microbiol 2002, 68:4301-4306.
Kaufmann C, Hanke U, Degenhart C, Baumann S, Wallner W et al.:A three-hybrid approach to scanning the proteome for targets 39. Brady SF, Chao CJ, Clardy J: New natural product families from of small molecule kinase inhibitors. Chem Biol 2004, an environmental DNA (eDNA) gene cluster. J Am Chem Soc 30. Towbin H, Bair KW, DeCaprio JA, Eck MJ, Kim S, Kinder FR, 40. Borman S: Interview with Samuel Danishefsky. Chem Eng News Morollo A, Mueller DR, Schindler P, Song HK et al.: Proteomics- based target identification: bengamides as a new class ofmethionine aminopeptidase inhibitors. J Biol Chem 2003, 41. Tan DS: Current progress in natural product-like libraries for discovery screening. Comb Chem High Throughput Screen2004, 7:631-643.
31. Harding MW, Galat A, Uehling DE, Schreiber SL: A receptor for the immunosuppressant FK506 is a cis-trans peptidyl-prolyl 42. Pelish HE, Westwood NJ, Feng Y, Kirchhausen T, Shair MD: isomerase. Nature 1989, 341:758-760.
Use of biomimetic diversity-oriented synthesis to discovergalanthamine-like molecules with biological properties 32. Sin N, Meng L, Wang MQ, Wen JJ, Bornmann WG, Crews CM: The beyond those of the natural product. J Am Chem Soc 2001, anti-angiogenic agent fumagillin covalently binds and inhibits the methionine aminopeptidase, MetAP-2. Proc Natl Acad SciUSA 1997, 94:6099-6103.
43. Tan DS, Foley MA, Shair MD, Schreiber SL: Stereoselective synthesis of over two million compounds having structural 33. Hughes TR, Marton MJ, Jones AR, Roberts CJ, Stoughton R, features both reminiscent of natural products and compatible Armour CD, Bennett HA, Coffey E, Dai H, He YD et al.: Functional with miniaturized cell-based assays. J Am Chem Soc 1998, discovery via a compendium of expression profiles. Cell 2000, 44. Nicolaou KC, Pfefferkorn JA, Cao GQ: Selenium-based solid- 34. MacBeath G, Schreiber SL: Printing proteins as microarrays for phase synthesis of benzopyrans I: applications to high-throughput function determination. Science 2000, combinatorial synthesis of natural products. Angew Chem Int Current Opinion in Chemical Biology 2006, 10:321–326

Source: http://micro.med.harvard.edu/faculty/pdf/DTHUNG1.pdf