Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Novel Antibacterials: A Genomics Approach to Drug Discovery
Document related concepts
Oncogenomics wikipedia , lookup
Human genome wikipedia , lookup
No-SCAR (Scarless Cas9 Assisted Recombineering) Genome Editing wikipedia , lookup
Gene expression programming wikipedia , lookup
Biology and consumer behaviour wikipedia , lookup
Genomic imprinting wikipedia , lookup
Vectors in gene therapy wikipedia , lookup
Ridge (biology) wikipedia , lookup
Essential gene wikipedia , lookup
Polycomb Group Proteins and Cancer wikipedia , lookup
Nutriepigenomics wikipedia , lookup
Genetic engineering wikipedia , lookup
Pharmacogenomics wikipedia , lookup
Genome (book) wikipedia , lookup
Metagenomics wikipedia , lookup
Microevolution wikipedia , lookup
Epigenetics of human development wikipedia , lookup
Designer baby wikipedia , lookup
Pathogenomics wikipedia , lookup
Site-specific recombinase technology wikipedia , lookup
Therapeutic gene modulation wikipedia , lookup
Public health genomics wikipedia , lookup
Gene expression profiling wikipedia , lookup
Genome evolution wikipedia , lookup
Artificial gene synthesis wikipedia , lookup
Transcript
Current Drug Targets - Infectious Disorders, 2002, 2, 291-308 291 Novel Antibacterials: A Genomics Approach to Drug Discovery Pan F. Chan1, Ricardo Macarron2, David J. Payne1, Magdalena Zalacain1 and David J. Holmes1* 1Department of Microbiology. Microbial, Musculoskeletal and Proliferative Diseases CEDD, GlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426-0989, USA, 2Department of Molecular Screening, GlaxoSmithKline, King of Prussia, PA, USA Abstract: The appearance of antibiotic resistant pathogens, including vancomycin resistant Staphylococcus aureus, in the clinic has necessitated the development of new antibiotics. The golden age of antibiotic discovery, in which potent selective compounds were readily extracted from natural product extracts is over and novel approaches need to be implemented to cover the therapeutic shortfall. The generation of huge quantities of bacterial sequence data has allowed the identification of all the possible targets for therapeutic intervention and allowed the development of screens to identify inhibitors. Here, we described a number of target classes in which genomics has contributed to its identification. As a result of analyzing sequence data, all of the tRNA synthetases and all of the two-component signal transduction systems were readily isolated; which would not have been easily identified if whole genome sequences were not available. Fatty acid biosynthesis is a known antibacterial target, but genomics showed which genes in that pathway had the appropriate spectrum to be considered as therapeutic targets. Genes of unknown function may seem untractable targets, but if those that are broad spectrum and essential are identified, it becomes valuable to invest time and effort to determine their cellular role. In addition, we discuss the role of genomics in developing technologies that assist in the discovery of new antibiotics including microarray gridding technology. Genomics can also increase the chemical diversity against which the novel targets can be screened. Key words: genomics, bioinformatics, target validation, tRNA synthetases, two-component signal transduction, genes of unknown function, enzyme based screening, whole cell screening, microarray gridding, inducible promoters, antisense. 1. INTRODUCTION In recent years there has been an explosion in the amount of available genome data in all of the kingdoms. These include sequences from the first complete genome, that of the bacterium Haemophilus influenzae, to the recent publication of the human genome sequence [1,2]. Moreover, there has been an explosion in the number of genome sequencing projects covering the three phylogenetic domains including those for the yeast chromosomes, and plant as well as bacterial and archael genomes. In fact over 100 complete bacterial genome sequences are readily available (Table 1) with hundreds more currently in progress. The data generated to date is only the beginning since bioinformatic, proteomic, genetic and, especially, biochemical approaches now need to be brought to bear. In this "post genomics era" efforts are now focused towards determining the function of many of these genes. Until recently, the bacterium was a black box of poorly understood physiology. With the availability of genome sequence data, researchers have access to all of the genes that are involved in the biochemical reactions of the bacterial cell, and which contribute to its physiological and structural components. Unraveling the *Address correspondence to this author at the Department of Microbiology. Microbial, Musculoskeletal and Proliferative Diseases CEDD, GlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426-0989, USA; TEL: 610 917 6464, FAX: 610 917 7901, email: [email protected] 1568-0053/02 $35.00+.00 specific function of each gene is not facile (see article by C. Volker & J. R. Brown) and in most cases at least one-third of the genes identified cannot be assigned even a putative function. The benefits of genome sequence data to pharmaceutical companies were obvious as it could provide targets for therapeutic intervention. The traditional method for discovering antibiotics involves screening for inhibition of bacterial cell growth by compound banks or, more commonly, natural product extracts. Subsequently, the antibacterial mode of action of the compounds had to be determined. This difficult task was simplified by the fact that most inhibitors discovered by this method affect relatively few cellular processes including transcription, translation, DNA replication and cell wall biosynthesis. However, while this approach was highly successful during the 1950's-1970's, very few novel antibiotics have been discovered in the last 25 years, Fig. (1). While antibiotic drug discovery has been on the decline, resistance to the therapeutic agents employed has steadily increased. The emergence of common pathogens that are resistant to multiple antibiotics coupled with the failure of traditional methods to yield novel anti-infective agents has required a creative new approach to drug discovery. Genome sequence data has led to the identification of new classes of broad spectrum targets, and through © 2002 Bentham Science Publishers Ltd. 292 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 combinatorial genomic techniques, novel sources of compounds to screen. In addition, novel technologies derived from genome sequence data have allowed improved Holmes et al. methods for determining the mode of action of compounds identified by their ability to inhibit the growth of whole cells. Table 1. Publically Available Genome Sequences Organism Domain Genes Total Bases Source Actinobacillus actinomycetemcomitans HK1651 Bacteria 1,988 2,095,439 University of Oklahoma Aeropyrum pernix K1 Archaea 1,840 1,669,695 NCBI Agrobacterium tumefaciens C58 Bacteria 5,299 5,673,563 NCBI Aquifex aeolicus VF5 Bacteria 1,553 1,590,791 NCBI Archaeoglobus fulgidus DSM4304 Archaea 2,420 2,178,400 NCBI Bacillus anthracis Ames Bacteria 5,287 5,227,297 TIGR Bacillus halodurans C-125 Bacteria 4,066 4,202,353 NCBI Bacillus stearothermophilus 10 Bacteria 3,342 3,269,999 University of Oklahoma Bacillus subtilis 168 Bacteria 4,112 4,214,814 NCBI Bordetella pertussis Tohama I Bacteria 3,892 4,086,186 Sanger Center Borrelia burgdorferi B31 Bacteria 1,637 1,519,856 NCBI Brucella melitensis 16M Bacteria 3,198 3,294,931 NCBI Buchnera sp. APS Bacteria 574 655,725 NCBI Campylobacter jejuni NCTC 11168 Bacteria 1,633 1,641,481 Sanger Center Caulobacter crescentus Bacteria 3,737 4,016,947 TIGR Chlamydia muridarum Bacteria 916 1,076,912 NCBI Chlamydia trachomatis D/UW-3/Cx Bacteria 921 1,042,519 NCBI Chlamydophila pneumoniae AR39 Bacteria 1,109 1,229,853 NCBI Chlamydophila pneumoniae CWL-029 Bacteria 1,052 1,230,230 Incyte Chlamydophila pneumoniae J138 Bacteria 1,069 1,226,565 NCBI Chlamydophila pneumoniae L2 Bacteria 1,042 1,234,390 TIGR Chlorobium tepidum TLS Bacteria 3,178 2,196,918 TIGR Clostridium acetobutylicum ATCC824 Bacteria 3,672 3,940,880 TIGR Clostridium perfringens 13 Bacteria 2,723 3,085,740 NCBI Corynebacterium diphtheriae NCTC13129 Bacteria 2,128 2,488,600 Sanger Center Corynebacterium glutamicum Bacteria 2,989 3,309,400 EBI Deinococcus radiodurans R1 Bacteria 3,103 3,284,156 TIGR Desulfovibrio vulgaris Bacteria 3,571,425 TIGR Enterococcus faecalis V583 Bacteria 3,148 3,359,973 TIGR Escherichia coli K-12 Bacteria 4,279 4,639,221 NCBI Escherichia coli O157:H7 Bacteria 5,361 5,498,450 NCBI Fusobacterium nucleatum ATCC 25586 Bacteria 2,067 2,174,500 NCBI Haemophilus influenzae KW20 Bacteria 1,714 1,830,138 NCBI Halobacterium sp. NRC-1 Archaea 2,605 2,571,010 NCBI Helicobacter pylori 26695 Bacteria 1,576 1,667,867 NCBI A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 293 (Table 1). contd..... Organism Domain Genes Total Bases Source Helicobacter pylori J99 Bacteria 1,491 1,643,831 NCBI Lactococcus lactis subsp. Bacteria 2,266 2,365,589 NCBI Listeria innocua Clip11262 Bacteria 2,968 3,011,208 NCBI Listeria monocytogenes EGD Bacteria 2,846 2,944,528 NCBI Mesorhizobium loti Bacteria 7,281 7,596,300 NCBI Methanobacterium thermoautotrophicum delta H Archaea 1,869 1,751,379 NCBI Methanococcus jannaschii DSM 2661 Archaea 1,770 1,739,927 NCBI Methanopyrus kandleri AV19 Archaea 1,687 1,694,969 NCBI Methanosarcina acetivorans Archaea 4,540 5,751,492 NCBI Mycobacterium leprae Bacteria 2,157 3,268,203 Sanger Center Mycobacterium tuberculosis CDC1551 Bacteria 4,187 4,403,836 NCBI Mycoplasma genitalium G-37 Bacteria 484 580,074 NCBI Mycoplasma pneumoniae M129 Bacteria 688 816,394 NCBI Mycoplasma pulmonis UAB CTIP Bacteria 782 963,879 NCBI Neisseria gonorrhoeae FA 1090 Bacteria 2,129 2,146,879 University of Oklahoma Neisseria meningitidis Z2491 Bacteria 2,065 2,184,406 Sanger Center Pasteurella multocida PM70 Bacteria 2,014 2,257,487 NCBI Porphyromonas gingivalis W83 Bacteria 1,777 2,343,478 TIGR Prochlorococcus marinus MED4 Bacteria 1,716 1,674,813 DOE Joint Genome Pseudomonas aeruginosa PAO1 Bacteria 5,565 6,264,403 University of Washington Pyrobaculum aerophilum Archaea 2,275 2,222,890 UCLA Dept. Micr Pyrococcus abyssi Archaea 1,765 1,765,118 NCBI Pyrococcus furiosus Archaea 2,208 1,908,253 Utah Pyrococcus horikoshii OT3 Archaea 2,058 1,738,505 NCBI Ralstonia solanacearum Bacteria 5,116 5,810,922 NCBI Rhizobium sp. NGR234 Bacteria 417 536,165 NCBI Rickettsia conorii Bacteria 1,374 1,268,755 NCBI Rickettsia prowazekii Madrid E Bacteria 834 1,111,523 NCBI Saccharomyces cerevisiae S288C Eukaryota 6,261 12,057,849 NCBI Salmonella typhi CT18 Bacteria 4,633 5,133,712 Sanger Center Salmonella typhimurium LT2(strain AZ1516) Bacteria 4,553 4,951,371 Washington Univ Shewanella putrefaciens Bacteria 4,221 5,131,063 TIGR Sinorhizobium meliloti 1021 Bacteria 6,205 6,691,694 NCBI Staphylococcus aureus EMRSA-16 Bacteria 2,679 2,902,619 Sanger Center Staphylococcus aureus MW2, Mu50, N315 Bacteria 2,632 2,820,462 NCBI Staphylococcus aureus NCTC 8325 Bacteria 2,631 2,836,373 University of Oklahoma Staphylococcus epidermidis RP62A Bacteria 2,444 2,646,310 TIGR Streptococcus mutans UA159 Bacteria 1,871 2,030,921 University of Oklahoma Streptococcus pneumoniae R6 hex Bacteria 2,043 2,038,615 NCBI Streptococcus pneumoniae type4 Bacteria 2,094 2,160,837 TIGR 294 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Holmes et al. (Table 1). contd..... Organism Domain Genes Total Bases Source Streptococcus pyogenes M1 GAS Bacteria 1,697 1,852,441 University of Oklahoma Streptococcus pyogenes MGAS8232 Bacteria 1,845 1,895,017 NCBI Sulfolobus solfataricus P2 Archaea 2,977 2,992,245 NCBI Sulfolobus tokodaii Archaea 2,826 2,694,765 NCBI Synechocystis sp. PCC6803 Bacteria 3,169 3,573,470 NCBI Thermoanaerobacter tengcongensis MB4T Bacteria 2,588 2,689,445 NCBI Thermoplasma volcanium Archaea 1,499 1,584,804 NCBI Thermotoga maritima MSB8 Bacteria 1,846 1,860,725 TIGR Treponema pallidum Nichols Bacteria 1,031 1,138,011 NCBI Vibrio cholerae N16961 Bacteria 3,835 4,033,464 TIGR Xanthomonas axonopodis citri str. 306 Bacteria 4,312 5,175,554 NCBI Xanthomonas campestris ATCC 33913 Bacteria 4,181 5,076,188 NCBI Xylella fastidiosa Bacteria 2,831 2,731,750 NCBI Yersinia pestis CO-92 Biovar Orientalis Bacteria 4,083 4,829,855 Sanger Center NCBI = National Center for Biotechnology Information, TIGR = The Institute for Genomic Research. In this article we will discuss the criteria that define an antibacterial target; their selection by bioinformatic analysis, validation by genetic methods and improvements in screening methods using specific examples. In addition, alternative sources of compound diversity will be reviewed along with methods for determining the cellular target of compounds derived from a bacterial whole cell screen. 2. CRITERIA FOR A VALID ANTIBACTERIAL TARGET 2.1 Novelty Current antimicrobial agents target a relatively small number of cellular processes. Fluoroquinolones and Fig. (1). Antibiotics under threat. While few new chemical classes of antibiotics have been discovered in the past 20 years, the number of resistant strains in the clinic continues to increase. A Genomics Approach to Drug Discovery novobiocin are examples of inhibitors of DNA replication; protein synthesis is prevented by aminoglycosides, macrolides, and tetracyclines while β -lactams and carbapenems affect cell wall biosynthesis. These compounds have been synthetically modified to overcome the resistance issues that have developed over the period of their clinical administration. In order to avoid being compromised by any of the current resistance mechanisms it is desirable to develop compounds that act against hitherto unexploited targets including new biosynthetic enzymes, transcription factors, or structural proteins etc. [3]. 2.2 Spectrum/Selectivity Since rapid, precise diagnosis of bacterial infection is still not extensively accessible, the present unmet clinical need is for a broad spectrum antibiotic to complement and/or replace current chemotherapies such as methicillin, amoxycillin and vancomycin which are compromised by resistance mechanisms found in clinical isolates. Comparison of bacterial genome sequences allows the identification of targets that are present in all clinically relevant pathogens and can be expected to be selective against humans due to their absence or significant difference in higher eukaryotes [4-6]. However, one can also imagine directed therapies that would target genes restricted to one or a few pathogens, thus affecting single or specific groups of bacteria. Consequently, genes specific to Gram positive or Gram negative bacteria could be targeted or in some cases, such as Helicobacter pylori it may be desirable to screen a species specific protein [7] (see article by Noonan B. and Alm R.A.). The result would be that specific disease states could be treated without necessarily affecting the entire bacterial flora. Bioinformatic analysis can show the anticipated spectrum of a compound aimed at any given target if the genome sequence is available. 2.3 Essential for Cell Viability Obviously, an antibacterial target must be essential for cell viability either under laboratory conditions or during infection. In order to focus the search for putative antibacterial targets the essential genes within a given bacterium should be identified. This could prove a daunting task given the number of genes (about 2000) in each pathogen that causes human disease. However, once bioinformatic analysis described above, identifies the broad spectrum genes in all organisms relevant to a particular disease, only a few hundred genes remain to be examined. Traditionally, essential genes were identified randomly using a variety of genetic techniques. For example, temperature sensitive mutants can be generated and the determinant identified following complementation with the corresponding wild-type gene. Transposon mutagenesis involves the random insertion of a mobile element into the chromosome. This method can be used to identify nonessential genes or genes essential under certain conditions. Lately, transposon mutagenesis has been revolutionized by the advent of genomics. Now, specific lengths of chromosome from naturally transformable bacteria can undergo saturation transposon mutagenesis in vitro and this Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 295 DNA can then be introduced back into the host organism. The bank of mutants generated can be analysed to determine which genes in the chromosome fragment contain insertions [8,9]. Those genes not containing insertions are probably essential for survival. However, it is not always possible to demonstrate that every gene in the DNA fragment suffered at least one insertion event. Systematic gene-by-gene analysis of essentiality is also possible. Methods are available for knocking out genes in a directed fashion. The simplest form is plasmid insertion mutagenesis [3]. Here a small internal fragment of the gene of interest is cloned in a plasmid tagged with an antibiotic resistance determinant that is expressed in the target organism. This plasmid, which cannot replicate in that bacterial species, is introduced into the organism and drugresistant colonies identified. The plasmid DNA is only maintained following homologous recombination into the target gene, and will therefore disrupt it. The method is efficient, but has caveats: in many cases significant portions of the "disrupted" gene remain intact resulting in partial gene activity being expressed. If partial activity is sufficient for cell survival, single crossover insertional inactivations may miss essential genes giving rise to false negatives. Since the insertion event introduces the entire vector into the chromosome, this will almost certainly effect the expression of genes downstream of the target (otherwise known as polarity effects). As a result, insertion in a non-essential gene may be lethal due to the effect on some distal gene vital for cell growth and would therefore be incorrectly assigned as essential. In order to minimise the effects of partial activity and polarity, allelic replacement can be employed. In this case, a gene is replaced by an antibiotic resistance marker following a double recombination event using homologous sequences that flank the target gene [10-13]. If the resistance determinant is carefully selected to ensure expression of downstream genes, polarity effects are avoided. Indeed there are examples where non-essential genes, upstream of known essential genes, have been successfully deleted using this method (Fig. (2)), Lunsford RD personal communication). The procedure only identifies genes that are likely to be essential, but very effectively defines those genes which are not, allowing attention to focus on potential antibacterial targets. In this manner, we have tested >400 Streptococcus pneumoniae and Staphylococcus aureus broad spectrum genes that were possible targets for antibiotic development. In over 70% of experiments it was possible to delete the targeted gene, thereby demonstrating that it was not essential for cell viability in vitro. The remaining genes, for which we were unable to isolate allelic-replacement mutants despite repeated attempts, are all potential targets for antibiotic action (Fig. (3a)), personal communication, D. Lunsford, GlaxoSmithKline). Over 100 mutants were tested for virulence attenuation in a relevant animal infection model and 72% were found to be attenuated by at least two logs, Fig. (3 b ). Whether this was due to disruption of the virulence process or simply effects on growth rate remain unclear. The remaining 28% of mutants were not significantly (< 2 logs) attenuated. 296 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Holmes et al. Fig. (2). Example where non-essential genes, upstream of a known essential gene, have been successfully deleted demonstrating that the allelic replacement technology used does not have polarity effects. The genes exoA and exoB encode subunits of an exonuclease and have been shown to be non-essential for cell viability. On the other hand, ispA encodes geranyltranstransferase, an essential gene involved isoprenyl pyrophosphate synthesis. Fig. (3). Results of essentiality testing. (a) Proportion of genes that could be disrupted by allelic replacement in vitro (non attenuated, dotted box; attenuated, oblique line box; in RTI model) and those that appeared essential; solid box. (b) Effects on pathogenicity of selected mutants in RTI model. Mutants were considered in vivo essential if attenuated by >5 logs in an RTI model, while those affected by <2 logs were considered non-attenuated. Perhaps the best approach to confirm that a gene is indeed essential for cell viability is by generating a directed conditional mutant by placing the target gene under the control of an inducible promoter. In contrast to knock-out methods that give an all or nothing response, promoter control systems allow the level of expression to be regulated and will result in reduced cell growth or death when expression is repressed. This method and the contribution of genomics to defining appropriate promoter systems will be discussed in more detail later. 2.4 Amenable to High-Throughput Screening The purpose of these efforts is to identify suitable targets which can be screened in order to discover new chemical entities that will inhibit bacterial growth. Therefore the novel target protein must have an activity that can be assayed in high throughput (although knowledge of the exact cellular function is not necessarily required) and reagents must be readily available. This will be discussed later in section 4. Below, four kinds of targets are described which exemplify the ways in which genomics has effected antiinfective drug discovery. 3. GENOMIC DERIVED TARGETS 3.1 Aminoacyl-tRNA Synthetases Translation of messenger RNA into protein involves precise recognition of the codons by the appropriately charged tRNA. Aminoacyl-tRNA synthetases are responsible for accurately combining the correct amino acid with its cognate tRNA [14]. While in Gram negative bacteria there is a specific enzyme for every tRNA molecule, Gram positive bacteria possess only 19 aminoacyl-tRNA synthetases since misacylated Glu-tRNAGln is the substrate for a transamidase which converts the glutamyl adduct tRNAGln [15]. Each of the tRNA synthetases are essential for cell viability. Until recently, little was known about these enzymes in Gram positive bacteria. Only two, isoleucyl-tRNA synthetase [16] and lysyl-tRNA synthetase [17] had been identified in S. aureus and none had been isolated from S. pneumoniae. Indeed one of them, isoleucyl-tRNA synthetase is the target of mupirocin, a valuable topical antibiotic with potent Gram positive activity, including against methicillin resistant S. aureus. Isolating each aminoacyl-tRNA synthetase by standard biochemical and genetic methods using appropriate DNA A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 297 probes is a plausible, albeit time-consuming and labor intensive, task since the number of enzymes is known. Clearly, this process is hugely simplified by the availability of entire genome sequences. Identification of all 19 of the aminoacyl-tRNA synthetase genes from both S. aureus and S. pneumoniae in silico by comparison with those from E. coli is facile. Moreover, this analysis can determine whether a specific enzyme may be more appropriate for broad spectrum or narrow spectrum utility as described above. For example, as there are distinct differences between the Gram positive and Gram negative methionyl-tRNA synthetases (MRS), this enzyme may be more appropriate if a selective Gram positive or Gram negative agent was required [18] and glutaminyl-tRNA synthetase would obviously be a Gram negative only target. Conversely, as the glutamyl-tRNA synthetases from Gram positive and Gram negatives are more similar this would be a more appropriate target if a broad spectrum agent was desired [18]. 3.2 Two-Component Signal Transduction Systems Two-component signal transduction systems (TCSTSs), more accurately defined as histidine-aspartate phosphorelay Table 2. systems, are perhaps the most widespread means of signal transduction in bacteria. They regulate many cellular responses, including osmoregulation, chemotaxis, sporulation, antibiotic production and pathogenicity, in a number of different bacteria [19]. TCSTSs are typically composed of two signaling proteins: a sensor kinase and its cognate response regulator. Specific environmental stimuli activate the sensor kinase protein, resulting in the autophosphorylation of a conserved histidine residue. This high-energy phosphate group is then transferred to a conserved aspartate residue in the cognate response regulator resulting in structural changes in the protein that mediates regulation of gene expression or protein function [20]. Histidine kinases are generally composed of a transmembrane-spanning, sensor kinase at the aminoterminal and a carboxyl-terminus containing the transmitter and kinase domains [21]. Since these systems are thought to be the bacteria's way of recognizing and responding to its environment, two-component signal transduction has been considered as a potential antimicrobial target [22,23]. Pathogenic bacteria are confronted with dramatic changes in environmental conditions (oxygen/pH/nutrient stress) when they infect a host, as well as being challenged in most cases by the host defence systems. Disruption of the TCST Analysis of S. pneumoniae Two-Component Signal Transduction Systems. Allelic Replacement Mutants were Generated (Y) in the Sensor Histidine Kinase (HK), the Response Regulator (RR), or the Gene Pair (HK & RR). 12 of these (Marked with an Asterisk) were Tested in a Murine RTI Model. ND: Not done. WT: No Attenuation Accession # Histidine Kinase Response Reg. HK RR HK & RR Attenuation in RTI model Homolog AAK99148 AAK99147 Y Y Y* WT YvqCE AAL00799 AAL00800 Y Y Y* 103-104 AAL00618 AAL00617 ND Y* ND 101 AAL00277 AAL00278 Y Y* Y 103-104 AAL00696 AAL00695 ND Y* ND 104 AAK98881 AAK98880 Y* ND ND 101 AAK99268 AAK99267 Y* ND ND 104 AAK99382 AAK99383 Y Y Y* 101-102 AAK99333 AAK99332 ND ND Y* 101 VncRS AAK99909 AAK99910 Y* Essential ND WT YycFG AAK98957 AAK98958 Y Y Y - AAK99140 - Y* - 104-105 AAL00844 AAL00843 ND ND ND AAK99512 AAK99511 ND ND Y* YvfTU PnpSR PhoPR ComDE 103 CiaRH 298 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 signalling systems should impair both the viability of the cell and its ability to establish and maintain infection. However, His-Asp phosphorelays have not been linked to the regulation of viability or pathogenicity. Prior to the advent of genomics, four probable TCSTSs had been identified in S. pneumoniae. The ciaRH, pnpSR and comDE gene pairs appear to play a role in the regulation of competence [24-28], and a fourth TCSTS encoded by the vncSR gene pair has been associated with the development of vancomycin tolerance [29]. In the absence of genome sequence data it would be impossible to know whether all of the TCSTSs had been identified from a particular bacterial species. Unlike aminoacyl-tRNA synthetases, it is difficult to estimate the number of TCSTS required by any particular microorganism. H. influenzae has a total of only 5 TCSTSs, Bacillus subtilis may have as many as 35 while Pseudomonas aeruginosa has the highest number of all bacterial genomes sequenced to date, with at least 63 TCSTSs [30]. Examination of the entire genome sequence of S. pneumoniae identified 14 TCSTSs. All have been investigated for their role for in vitro cell viability as well as their role in the establishment and maintenance of infection (Table 2). Systematic mutagenesis studies and complementation experiments demonstrated that one response regulator is essential for cell growth in S. pneumoniae. Surprisingly, inactivation of its cognate histidine kinase had little or no effect on growth either in vivo or in vitro [31]. This system is homologous to the YycFG TCSTS pair identified in S. aureus and B. subtilis [32,33]. Unlike the situation in S. pneumoniae, both the histidine kinase, YycG, and the response regulator, YycF, were found to be vital for growth in these bacteria. Characterization of the S. pneumoniae mutant deleted for YycG showed no apparent phenotypic alteration, indicating a functional YycF. This implies that the activity of the YycF response regulator protein is phosphorylation independent or is recognized by other apparently "non-cognate" sensor kinase proteins. Analysis of Streptococcus pyogenes, Enterococcus faecalis and Lactococcus lactis genome sequence data shows that they also contain YycFG homologues, suggesting that this TCSTS is a broad spectrum target. The staphylococcal yycFG locus may have a role in maintaining membrane integrity and, it remains possible that YycFG co-ordinates membrane growth with the cellular redox potential [33]. Of the 14 TCSTSs, identified in S. pneumoniae, seven gene pairs appear to be important for the establishment and/or maintenance of respiratory tract infections, while having little effect on in vitro bacterial growth under normal laboratory conditions [31]. Using a genomic-based approach, it has been possible to identify and systematically disrupt the entire TCSTS gene complement of S. pneumoniae and thereby define those important for both pathogenicity and viability in a respiratory tract infection. These studies provide support for the contention that two-component signal transduction systems are appropriate targets for the development of antibacterial drugs. Holmes et al. 3.3 Fatty Acid Biosynthesis Cerulenin, thiolactomycin, diazaborines and triclosan are antibiotics that selectively inhibit one or more of the enzymes involved in fatty acid biosynthesis (FAB) [34] validating this cyclic pathway as a therapeutic target. Access to bacterial genomes has enabled all the components of the FAB pathway to be identified in a variety of clinically important pathogens (Table 3). This particular example demonstrates the importance of genomics for predicting the potential spectrum of each of the target enzymes. Table 3. Occurrence of Fab Genes in Key Pathogens Target S. aureus S. pneumoniae E. faecalis H. influenzae FabA X X X FabB X X X FabD FabF X FabG FabH FabI X FabK X X FabZ = homologue present in genome; X = homologue absent from genome In the pre-genomic era FabI was thought to be the unique enoyl-acyl carrier protein (ACP) reductase in all bacteria as data obtained from E. coli was assumed to pertain to all prokaryotes. However, following the completion of several additional genomes, it was obvious that fabI was absent from a number of organisms, including S. pneumoniae and E. faecalis. Since fatty acid biosynthesis is a cyclic process, enoyl-ACP reductase activity had to exist in these bacteria. Once again, genome sequence data provided the answer. In S. pneumoniae the fatty acid biosynthetic genes are clustered in one region of the chromosome, Fig. (4). Sequence analysis identified many of the genes in the cluster and while fabI seemed to be absent, one of the genes in the cluster remained unannotated. Biochemical characterization showed that the remaining ORF encoded an alternative enoyl-ACP reductase (FabK), a homolog of which was later also identified in E. faecalis. Consequently, inhibitors of FabK could deliver selective streptococcal and enterococcal agents whereas compounds that inhibited both FabK and FabI could have broad spectrum utility. 3.4 Genes of Unknown Function Clearly, essential genes encoding proteins of unknown biochemical function constitute the most novel of all possible antibiotic targets. More than 30% of the genes in a bacterial genome are annotated as encoding proteins of unknown function and at least 20% of these are broad spectrum. The difficulty in working with this type of target in drug discovery is intrinsic in its nature. If their function and A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 299 Fig. (4). The fatty acid biosynthetic gene cluster in S. pneumoniae. The gene ncd-2 is similar to a 2-nitropropane dioxygenase gene from Neurospora crassa. In fact, it encodes an enoyl-ACP reductase (FabK) that is distinct from orthologs encoded by fabI in other bacteria. enzymatic activity are unknown, they cannot be used in high-throughput screening. A compromise can be achieved by identifying essential genes encoding proteins whose function can be predicted and here we discuss proteins involved in stable RNA modification. Table 4. All living cells contain modified nucleosides in their stable RNA (mostly ribosomal RNA and transfer RNA). To date, a total of 96 modified nucleosides for which structures have been assigned have been reported in RNA (http://medlib.med.utah.edu/RNAmods). The distribution of Stable RNA Modifying Enzymes. Genes Encoding tRNA or rRNA Modifying Enzymes were Tested for Essentiality in S. pneumoniae. Those Required for Cell Viability could be Targets for Antibiotics tRNA Modifying Enzymes Accession # Gene Name Function Essentiality data AAL00255.1 hisT/truA pseudouridine (ψ 38,39,40) synthetase Non-essential AAK99736.1 yerS tRNA methyltransferase Non-essential AAL00520.1 yfjO tRNA methyltransferase Non-essential AAK99491.1 trmD tRNA (m1G37) methyltransferase Essential AAL00077.1 queA SAM tRNA ribosyltransferase Non-essential AAL00671.1 tgt queuine tRNA ribosyltransferase NT AAK99547.2 asuE/ trmU tRNA (s2U34) thioltransferase Essential AAK99724.1 trmE tRNA (cmnm5H34) methyltransferase Essential AAK99895.1 truB pseudouridine (ψ 55) synthase Non-essential AAK99392.1 miaA tRNA isopentenylpyrophosphate transferase Non-essential AAK99587.1 iscS sulfurtransferase (s4U8 biosynthesis) Non-essential Accession # Gene Name Function Essentiality data AAL00382.1 fmu 16S rRNA methyltransferase (m5C907) Non-essential AAL00063.1 yloM putative rRNA methyltransferase Non-essential AAK99280.1 ytmQ putative rRNA methyltransferase Non-essential AAL00591.1 ysgA rRNA methyltransferase Non-essential AAL00115.1 yacO rRNA methyltransferase Non-essential AAK99236.1 cspR rRNA methyltransferase Non-essential AAK99634.1 ylyB (rluD) pseudouridine synthase Non-essential AAK99810.1 yjbO pseudouridine synthase Non-essential AAL00492.1 ypuL1 pseudouridine 516 synthase Non-essential AAK99060.1 ypuL2 putative pseudouridine synthase Non-essential AAK99401.1 ypuL3 pseudouridine synthase Essential AAL00627.1 yhcT putative pseudouridine synthase Essential rRNA Modifying Enzymes NT=Not tested 300 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 these modifications is varied but a subset of them are present in all three phylogenetic domains (Archaea, Bacteria and Eukarya) sometimes even at comparative positions within the RNA structure. On the other hand, there are also species specific modifications (for reviews see, [35,36]). In E. coli, there are 11 modified nucleosides in 16S rRNA, 23 in 23S rRNA and 31 in tRNA. The formation of these residues is catalyzed by highly specific enzymes which, with the exception of the synthesis of quenosine, operate during the maturation of the RNA after the primary transcript is synthesized. It can be estimated that nearly 2% of the genetic information in a bacteria is dedicated to the synthesis of these modified residues and about half of the enzymes involved in this process have not been identified. This is clearly an under exploited target. Bioinformatic analysis of completed genomes, have identified 23 broad spectrum RNA modifying enzymes (present in S. pneumoniae, S. aureus and H. influenzae), nine of which have orthologs characterized in other organisms (mostly E. coli). There is no biochemical evidence for the function that the other 14 enzymes performed, other than they modify RNA based on their similarity to proteins of known function (Table 4). Essentiality studies on 22 of these genes suggest that five of them, TrmD, TrmE, TrmU, YhcT and ypuL3, are indispensable for growth of S. pneumoniae. E. coli TrmD and TrmE modify positions 37 and 34 respectively in certain subsets of tRNAs and have been previously characterized [37-39]. Phylogenetic analysis indicates that YhcT and ypuL3 belong to the family of rRNA pseudouridine synthetases but the position they modify is unknown. TrmU has been annotated in the literature as a tRNA thioltransferase [40] although there is no biochemical evidence for that assertion and this needs to be confirmed. Further studies on the essentiality of these proteins in other organisms has shown that TrmE and ypuL3 are only essential in S. pneumoniae and therefore constitute novel, species specific targets. YhcT is also essential in S. aureus but it has three distinct paralogs in H. influenzae and none of them is essential individually. This would be a typical Gram positive only target. On the other hand, TrmD and TrmU are essential in both Gram positive and Gram negative organisms and are, therefore, broad spectrum targets. Consequently, genome sequence data can greatly expand the number of therapeutic targets within a given class by identifying putative orthologs. 4 TARGET SCREENING 4.1 Screening for Inhibitors Despite the myriad of targets for antibacterial intervention unveiled from genomic sequence analysis through the process described above, and the fact that many High-Throughput Screening (HTS) campaigns have been run against dozens of these targets both at GlaxoSmithKline and many other pharmaceutical companies, experience shows that finding leads to these antimicrobial targets is more difficult than for other therapeutic areas. Holmes et al. At GSK, a HTS success is defined as a screening campaign that affords at least one lead compound of sufficient quality to start a chemistry program, based on structure, potency, selectivity and initial SAR trends. These HTS campaigns are run according to well-established standards of quality [41]. Possible causes for the lower success rate of obtaining tractable hits in an antimicrobial HTS include the "drugability" of the target and/or the diversity and size of the screening collection. Target "drugability" refers to the feasibility of finding drugs to any given target (reviewed in [42]). In the last few years, the definition of a drug-like molecule has become increasingly restricted to synthetic small molecules that comply with the "rule-of-five" as defined by Lipinski et al. [43]. The latter are a compendium of guidelines derived from statistical observation of chemical properties (molecular weight, lipophilicity as measured by clogP, number of H bond donors and acceptors) of marketed drugs. Their intent is to guide lead optimization programs away from series or molecules with high probability of failure due to poor absorption and/or poor permeability. When comparing the success rate of in-house HTS campaigns across different target classes some trends are clearly observed. Most antibacterial targets screened are enzymes other than protein kinases and proteases. When this broad family of proteins is further analyzed, the success rates within this group vary considerably between subfamilies. Generally, HTS success correlates with the hydrophobicity of the substrate binding pocket of the target. For example, many chemically viable leads have been described for enzymes involved in fatty acid biosynthesis [44,45], whereas only a handful of less tractable ones have been discovered for sugar binding enzymes [46]. One simplistic explanation is that enzymes that utilize hydrophilic substrates (e.g. sugars) offer fewer opportunities for interaction with small synthetic compounds because the majority of these tend to be hydrophobic. This is due to the nature of organic building blocks and their synthetic routes, generally engineered in a water-free environment. On the other hand, there is some evidence to suggest that hydrophobic rather than hydrophilic ligands will form higher affinity interactions because the water solvation of a hydrophilic moiety imposes an energetic barrier to binding [47]. As a result according to Lipinski et al., "one of the most reliable methods in medicinal chemistry to improve in vitro activity is to incorporate properly positioned lipophilic groups". In addition to the biochemical perspective described above, most collections of synthetic molecules within pharmaceutical and biotechnology companies contain a significant proportion of compounds derived from specific chemistry programs around the initial leads obtained after successful screen campaigns. Accordingly, the majority of molecules in such compound collections interact with certain classes of G-protein coupled receptors, ion channels, proteases, and more recently kinases, [48]. Consequently, the probability of a new HTS target encountering high affinity ligands in a synthetic compound screening collection are higher if it is similar to a target that resulted in a A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 301 successful screening campaign and therefore contributed to the growth of the collection. Nevertheless, serendipity plays a role in HTS success, and novel targets can show high affinity for compounds designed to interact with unrelated proteins. For example, the series of methionyl-tRNA synthetase inhibitors described by [49], evolved from an HTS hit originally synthesized in an optimization program for antagonists of histamine H1 receptor. Given the low success of genomics-derived, isolated, molecular targets from bacteria in HTS, another approach [53] is to screen whole pathways or whole cells. The burden of these approaches is that if interesting leads are found, a labor intensive target identification effort must follow with uncertain results. However, the advantage remains that the intractable issue of cell penetrability is addressed up front [54,55]. As discussed above, synthetic molecules tend to be lipophilic. One might speculate on the degree to which the Lipinski rules are a result of this general hydrophobia – i.e. drug-like properties are so defined because molecules with different properties are not progressed and therefore not tested sufficiently – and how much they reflect the many characteristics a drug must display within a physiological environment: specific and potent binding, access to its target, low adsorption to cellular and extracellular milieu components, selectivity against many related and unrelated proteins, in summary good ADME (absorption, distribution, metabolism, excretion) and low toxicology. There are a number of avenues, that are just starting to be explored, which could potentially fill the current gap in the global antibacterials pipeline: The fact that many marketed antibiotics, and other natural products, are apparently the main exception to the rule-of-five seems to support this line of thought: a historically biased view of what constitutes a drug-like molecule may preclude finding antibiotics through the current Drug Discovery paradigm. In the past few years, HTS of natural products has been losing support in many pharmaceutical companies because of a cost-benefit analysis that some believe is short-sighted [50]. The fact remains that the return in value from traditional approaches to natural products screening has been in decline for the last two decades, mainly because of the labor intensive nature of this work and the fact that many of the metabolites found were either already known, and thus unexploitable from an intellectual property perspective, or too complex to enter a chemical optimization program. However, many marketed drugs are natural products that did not require any derivatization to be effective and safe (e.g. lovastatin, cyclosporin, clavulanic acid, erythromycin, caffeine, taxol). Many of these compounds are complex chiral molecules that contain several polar substituents – ideal for inhibiting less tractable targets such as protein-protein interactions and enzymes that act on polar substrates. These compounds may not be “drug-like” – but they are drugs! A biased view of the basic criteria in current drug discovery (a lead has to be amenable to chemical modification to drive optimization throughout drug development) may have diminished the chances of this proven source of effective antibiotics to add to our arsenal of weapons against bacterial infections. Interestingly, and despite the exception to the rule-of-five observed for some drugs of natural origin, recent studies have shown that collections of plant and microbial metabolites are, on average, more similar to synthetic compound collections than previously thought and most of them lie within the limits suggested by Lipinski et al. [51,52]. Thus, there is scope for the addition of these kinds of compounds to current collections to enrich the diversity, broaden the physico-chemical space and maximize the chances of finding leads for less tractable but highly validated targets. • exploitation of untapped sources of microbial metabolites (expression of biosynthetic pathways from unculturable microorganisms; see section 4.2) • reduce the risks and cost of the traditional extract screening by screening purified or semipurified samples [56]. • incorporate scaffolds from natural products into combinatorial chemistry approaches [57]. • derivatize natural products by biotransformation [58]. • structure-based screening and design [59,60]. Finally, when analyzing the success of HTS in this or any other area, it is fair to consider that the drug discovery process has been reshaped in the last decade and we are only now starting to see the fruits of this work entering the development pipelines. The load of new knowledge generated by high-throughput technologies applied in the last 10 years to molecular biology, biochemistry and chemistry and the increased help from computational tools will hopefully start to pay off in the next few years and truly new antibiotics – those with novel modes of action and enhanced properties - will flow to the market. 4.2 Combinatorial Genomics Identification of targets and development of robust highthroughput assays allows screening of millions of compounds for inhibitory activity. Most of the currently commercialized antibiotics are derived from natural product extracts and at least four, the macrolides azithromycin and clarithromycin, the cephalosporin ceftriaxone and the betalactam amoxicilin in conjunction with another natural product, the beta-lactamase inhibitor, clavulanic acid, each have annual sales in excess of $1B. In order to discover new antibacterial compounds it is imperative that the novel targets unraveled by genomics are screened against a wide diversity of compounds which will come from new organic syntheses or from the identification of novel natural products produced by soil and marine organisms as well as plants. Bacterial diversity is being exploited by searching extreme habitats for secondary metabolite producers. However, genomics is also being applied to the problem. Natural products are complex molecules largely (though not exclusively) produced by soil microorganisms such as 302 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Streptomyces. However, the vast majority of microbial species cannot be cultivated under laboratory conditions. Indeed, it is estimated that only 0.1-1% of microorganisms can be isolated [61]. Several biotechnology companies are now exploiting the ability of heterologous hosts to express DNA containing secondary metabolic pathways extracted from soil and thereby identify novel compounds [62]. In fact they have already been able to demonstrate the production of previously unidentified secondary metabolites using this methodology [63] thereby extending the spectrum of potential pharmaceuticals. This approach allows the rapid isolation of the biosynthetic gene cluster. Subsequently, it can be manipulated to provide intermediates and derivatives of complex chiral molecules that would be refractory to chemical synthesis and may be vital to determine SAR [64,65]. Finally, the streptomycetes are one of the major producers of secondary metabolites and the complete genome sequence of Streptomyces coelicolor was recently published [66]. Following decades of study, S. coelicolor was demonstrated to produce four secondary metabolites under a number of laboratory conditions. The genome sequence reveals that S. coelicolor may well be capable of producing as many as 20 secondary metabolites, all of which probably make specific interactions with macromolecules and some of which may prove to have some therapeutic use. 5. DETERMINATION OF MECHANISM OF ACTION OF ANTIBACTERIALS If an inhibitory compound has been identified by a target based screen, it is important to confirm that the antibacterial activity against whole bacteria is mediated via the suspected target. On the other hand, if the antimicrobial compound was isolated from a whole cell screen, then a concerted effort must be made to identify the molecular mode of action. Holmes et al. Genomics can contribute to both of these processes and examples are discussed below. 5.1 Regulated Gene Expression Tightly-regulated, inducible promoters have been used in antimicrobial drug discovery for demonstrating gene essentiality, and hence, the validation of drug targets [67,68]. Moreover, titratible promoter systems that are able to modulate the levels of the protein target have proven to be invaluable tools for tracking the mechanism of antibacterial activity of novel inhibitors. The Pspac promoter (previously characterized in B. subtilis) has been used in S. aureus to demonstrate that methionyl-tRNA synthetase (MRS) and peptide deformylase (PDF) are essential for cell viability [67]. Moreover, the MICs of inhibitors of these enzymes were directly related to their level of expression, correlating whole cell activity with inhibition of the specific target. Since such appropriate promoter systems were unavailable in S. pneumoniae, we have applied genomics to the problem (Chan P. et al., manuscript submitted). The entire genome of S. pneumoniae was scanned for putative novel sugar operons which are typically highly regulated. At least 14 putative carbohydrate utilization operons were identified, including those for putative cellobiose, fucose, fructose, galactose, glucose, lactose, maltose, mannitol, mannose, sucrose, trehalose and raffinose, and other unknown sugar utilization operons. Most gene clusters contained putative regulatory, uptake and sugar utilization determinants. Bioinformatic analysis of the regulatory regions in these operons identified a promoter within the fucose gene cluster that had the requisite properties for development of an inducible promoter system in S . pneumoniae. The fucose operon contains 10 genes, many of which show homology to the fucose catabolism genes of E. coli and H. influenzae. The first gene, fcsK, encodes a Fig. (5). (a) Genes under the control of an inducible promoter. By replacing the native promoter with an inducible promoter the level of expression can be controlled. (b) Inducer-dependent growth of S. pneumoniae (PfcsK::defI) regulated strain confirms PDF is essential for growth and a suitable antibacterial drug target. A Genomics Approach to Drug Discovery putative fuculose kinase enzyme in S. pneumoniae. fcsK is divergently transcribed from a gene whose product shows homology to the LacR family of transcriptional repressors and therefore indicating that the fcsK promoter is subject to down-regulation, and hence, an excellent candidate for further study. Transcript analysis revealed a near canonical bacterial promoter with a transcriptional start site located 24 bp upstream of fcsK. As predicted, the addition of fucose induced expression of fcsK (>23-fold). While galactose (a closely related sugar) was also able to induce f c s K expression; raffinose, lactose, trehalose and mannose had no significant effect on fcsK transcript levels. Moreover, transcription of fcsK was tightly repressed by sucrose or glucose (at least 10-fold). Hence the PfcsK promoter is inducible by fucose and repressible by sucrose giving a titratible dynamic range. Several S. pneumoniae genes of interest have been placed under the control of the fcsK promoter including def1 which encodes PDF. The strategy for construction of the regulated strain is outlined in Fig. (5a) and involves replacing the wild-type promoter with P f c s K by allelic exchange mutagenesis. The regulated S. pneumoniae PfcsK::def1 strain in Fig (5b) shows an absolute dependence on inducer fucose for normal growth confirming the essential nature of PDF for growth of S. pneumoniae. The availability of genomics data has allowed us to successfully identify a fucose-regulated, endogenous promoter from S. pneumoniae and its exploit it as a tool for target validation and antimicrobial mode of action studies. 5.2 Antisense In addition to the types of experiments described in the previous section, regulatable promoters can also be used to quantitatively express anti-sense RNA to specific transcripts for essential gene products [69]. Fragments of essential genes are inserted into a plasmid, downstream of the Pxyl-tet controllable promoter, in their reverse orientation. While cells are viable in the absence of inducer (since the wild-type gene is transcribed and translated), in the presence of inducer tetracycline, anti-sense RNA is generated, leading to degradation of the native mRNA and inhibition of bacterial growth. This system is particularly elaborate because by feeding animals tetracycline, it is possible to induce antisense in experimental animal models of infection and demonstrate that genes not required for growth on agar plates, are essential for pathogenesis [70]. 5.3 Microarray Gridding Microarray gridding is a powerful and quantitative genomic tool for analyzing the genome-wide expression of genes in bacteria. Since gene expression is primarily regulated at the level of transcription, the messenger RNA (mRNA) profile of completed genomes can provide a global, genomic snapshot of the transcriptome. Gene expression patterns generated are reflective of the physiological state of the cell [71]. Microarrays have been used to determine the mode of action of drugs by comparing the expressions in Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 303 drug-resistant and wild type strains, or in cells treated with and without drug. Using this approach the molecular target or cellular pathway being affected may be identified or predicted. Another powerful application of microarrays is the identification of the role of genes of unknown function in completed genomes, also called functional genomics. Details of DNA microarray technology, and its various applications in microbial systems have been extensively reviewed elsewhere [72-74]. Typically, a microarray consists of DNA representing every ORF of a genome, stamped onto a glass slide or nylon membrane. In mRNA expression profiling experiments, total RNA is commonly isolated from bacteria exposed to drugs, labeled, and hybridized to a microarray grid. Genes differentially expressed i.e. up- or down-regulated following compound treatment are identified. Microarray gridding technology has been applied to study bacterial regulatory networks, genetic alterations, host-microbe interactions, strain genotypes and validate drug targets, as well as determine drug-resistance mechanisms and antibacterial compound modes of action in many human pathogens. Microbial microarrays are available for many of the genomes sequenced to date including those for E. coli [75], B. subtilis [76], Saccharomyces cerevisiae [77,78], Mycobacterium tuberculosis [79,80], H. pylori [81], P. aeruginosa [82], H. influenzae [71], S. pneumoniae [83,84], and S. aureus1 [85,86]. Here, we focus on the impact of microarrays in antibiotic drug discovery and give our experiences of using microarrays to study the antibacterial mode of action of candidate lead compounds. A S. aureus array platform covering about 80% of the ORFs in the genome was constructed and used to profile1 four antibacterial compounds of known modes of action . Following two-color microarray hybridization, the response of each gene was compared in drug-treated and untreated samples, Fig. (6 ). mRNA profiling of gemifloxacin and ciprofloxacin, both quinolone class of DNA replication inhibitors, resulted in generally similar upregulation of genes associated with the SOS response and DNA repair mechanisms including recA, lexA, and uvrBA (excinuclease subunits). Triclosan, an inhibitor of FabI (see section 3.3) induced a strong up-regulation of heat shock proteins such as dnaK, ctsR, clpB, and groEL and some genes involved in fatty acid biosynthesis, and caused a down-regulation of purine biosynthetic genes and ribosomal proteins. In contrast, mupirocin an inhibitor of protein biosynthesis (see section 3.1) most strongly induced genes involved in branched chain amino acid biosynthesis1. Genes affected were generally consistent with inhibition of the metabolic pathway targeted by the different antibiotics. In addition, the expressions of many genes of unknown function were altered by drug treatments implying putative roles in the response of S. aureus to antibiotic stress and survival. The co-regulation of genes organized in the same operon in response to drug treatment further validated the data. Other examples of microarray drug profiling studies include isoniazid, a fatty acid pathway inhibitor in M . 1 Chan, P.F.; Lonetto, M.; Clark, S.M.; Gagnon, R.; Palmer, L.M.; Woodnutt, G.; Warren, P.; Jaworski, D.D. 2001, Abstract A-26. 101st American Society of Microbiology General Meeting, Orlando, FL 304 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Holmes et al. Fig. (6). Example of a microarray following gene expression profiling of an antibiotic in S. aureus. Test (drug treated or untreated) and reference RNA isolated from S. aureus were reverse transcribed to cDNA, labelled with fluorescence dyes Cy3 and Cy5 respectively, and hybridized onto a 1974 ORF microarray grid. Genes up- and down-regulated or unaffected in the test sample compared to the reference control, are shown by green, red and orange spots respectively. Antibiotic-responsive genes are identified by comparison of intensity ratios of each gene in the drug treated and the untreated samples. tuberculosis [79] and ciprofloxacin and novobiocin in H. influenzae [71]. In both cases, drug-specific gene expression signatures were identified consistent with drug mechanisms of action. We have profiled more than 12 known antibiotics of different modes of action in S. aureus including sub-classes of inhibitors, and in most cases, identified distinct 2drug signature expression patterns for these compounds . By establishing a data bank of expression profiles, the mode of action of antibacterial compounds identified from a bacterial whole cell screen may be classified according to the their transcriptome response. Microarray gene expression profiling though a powerful tool will not always identify the precise molecular target of the inhibitor but often predicts the metabolic pathway being affected. Genome-wide profiling using microarrays clearly has a role, in conjunction with other molecular mode of action tools, to support SAR efforts during lead optimization. Since expression of most of the genes on the grid is unaltered following drug treatment, once a cluster of the 300 most drug-responsive genes has been identified, a mini-array may then be constructed to facilitate throughput during a 2 Chan, P.F.; Gagnon, R.; Clark, S.M.; O'Brien, S.; Boyle, R.; Javed, R.; Au, J.; Lonetto, M.; Jaworski, D.D. 2002, Abstract A-43. 102nd American Society of Microbiology. lead validation and optimization project. Antibiotic gene signatures identified by microarray profiling may also be developed into a panel of cellular reporter strains as an alternative novel screening method for predicting mechanism of action or as a secondary assay for chemical optimization2. Though still a relatively new technology, microarray gridding has already become a useful tool in antimicrobial drug discovery 2 [71,85]. Furthermore, microarrays are invaluable for studying the function of genes of completed genomes. In functional genomics, the role of essential proteins of unknown biochemical function (see section 3.4) might be elucidated using microarrays. Expression profiling of regulated strains may lead to the identification of new drug targets and development of possible assays. Other current applications of microarrays include the study of bacterial host interactions whereby total RNA is isolated from infected tissue samples and hybridized to both microbial and human/rat arrays [87,88]. In summary, we have applied microarrays to both predict the mechanism of action of novel antibacterial agents identified from a whole cell screen and for validating the target of lead hits from a HTS and prioritize these compounds for development. Microarray gridding is a powerful tool for genome-wide gene expression profiling, which is becoming widely accepted while supplementing other established methods to accelerate the antibacterial drug discovery process. A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 305 Dew, I.; Fasulo, D.; Flanigan, M.; Florea, L.; Halpern, A.; Hannenhalli, S.; Kravitz, S.; Levy, S.; Mobarry, C.; Reinert, K.; Remington, K.; Abu-Threideh, J.; Beasley, E.; Biddick, K.; Bonazzi, V.; Brandon, R.; Cargill, M.; Chandramouliswaran, I.; Charlab, R.; Chaturvedi, K.; Deng, Z.; Di, F.V.; Dunn, P.; Eilbeck, K.; Evangelista, C.; Gabrielian, A.E.; Gan, W.; Ge, W.; Gong, F.; Gu, Z.; Guan, P.; Heiman, T.J.; Higgins, M.E.; Ji, R.R.; Ke, Z.; Ketchum, K.A.; Lai, Z.; Lei, Y.; Li, Z.; Li, J.; Liang, Y.; Lin, X.; Lu, F.; Merkulov, G.V.; Milshina, N.; Moore, H.M.; Naik, A.K.; Narayan, V.A.; Neelam, B.; Nusskern, D.; Rusch, D.B.; Salzberg, S.; Shao, W.; Shue, B.; Sun, J.; Wang, Z.; Wang, A.; Wang, X.; Wang, J.; Wei, M.; Wides, R.; Xiao, C.; Yan, C.; Yao, A.; Ye, J.; Zhan, M.; Zhang, W.; Zhang, H.; Zhao, Q.; Zheng, L.; Zhong, F.; Zhong, W.; Zhu, S.; Zhao, S.; Gilbert, D.; Baumhueter, S.; Spier, G.; Carter, C.; Cravchik, A.; Woodage, T.; Ali, F.; An, H.; Awe, A.; Baldwin, D.; Baden, H.; Barnstead, M.; Barrow, I.; Beeson, K.; Busam, D.; Carver, A.; Center, A.; Cheng, M.L.; Curry, L.; Danaher, S.; Davenport, L.; Desilets, R.; Dietz, S.; Dodson, K.; Doup, L.; Ferriera, S.; Garg, N.; Gluecksmann, A.; Hart, B.; Haynes, J.; Haynes, C.; Heiner, C.; Hladun, S.; Hostin, D.; Houck, J.; Howland, T.; Ibegwam, C.; Johnson, J.; Kalush, F.; Kline, L.; Koduru, S.; Love, A.; Mann, F.; May, D.; McCawley, S.; McIntosh, T.; McMullen, I.; Moy, M.; Moy, L.; Murphy, B.; Nelson, K.; Pfannkoch, C.; Pratts, E.; Puri, V.; Qureshi, H.; Reardon, M.; Rodriguez, R.; Rogers, Y.H.; Romblad, D.; Ruhfel, B.; Scott, R.; Sitter, C.; Smallwood, M.; Stewart, E.; Strong, R.; Suh, E.; Thomas, R.; Tint, N.N.; Tse, S.; Vech, C.; Wang, G.; Wetter, J.; Williams, S.; Williams, M.; Windsor, S.; Winn-Deen, E.; Wolfe, K.; Zaveri, J.; Zaveri, K.; Abril, J.F.; Guigo, R.; Campbell, M.J.; Sjolander, K.V.; Karlak, B.; Kejariwal, A.; Mi, H.; Lazareva, B.; Hatton, T.; Narechania, A.; Diemer, K.; Muruganujan, A.; Guo, N.; Sato, S.; Bafna, V.; Istrail, S.; Lippert, R.; Schwartz, R.; Walenz, B.; Yooseph, S.; Allen, D.; Basu, A.; Baxendale, J.; Blick, L.; Caminha, M.; Carnes-Stine, J.; Caulk, P.; Chiang, Y.H.; Coyne, M.; Dahlke, C.; Mays, A.; Dombroski, M.; Donnelly, M.; Ely, D.; Esparham, S.; Fosler, C.; Gire, H.; Glanowski, S.; Glasser, K.; Glodek, A.; Gorokhov, M.; Graham, K.; Gropman, B.; Harris, M.; Heil, J.; Henderson, S.; Hoover, J.; Jennings, D.; Jordan, C.; Jordan, J.; Kasha, J.; Kagan, L.; Kraft, C.; Levitsky, A.; Lewis, M.; Liu, X.; Lopez, J.; Ma, D.; Majoros, W.; McDaniel, J.; Murphy, S.; Newman, M.; Nguyen, T.; Nguyen, N.; Nodell, M. Science, 2001, 291, 1304. 6. CONCLUDING REMARKS Genomics has supplied us with every target in the bacterial cell. Imaginative science is now required to exploit this embarrassment of riches. It is now feasible to target therapy at particular pathogens, whether they be individual organisms such as H. pylori or specific groups such as Gram positive bacteria, leaving the remaining flora unaffected. Alternatively, inhibitors of broad spectrum targets, present in all bacteria can be used when rapid diagnosis is not possible. Certainly, these genes have been identified and, in many cases, their essentiality determined. Given this situation, why are we not inundated with novel antibacterial agents? It is likely that the targets have not been screened against sufficient chemical diversity to allow the identification of adequate lead compounds. However, novel technologies are now addressing this problem and genomics will again play a pivotal role. Without doubt, novel antibiotics that inhibit new bacterial targets are essential if the current resistance problems are to be overcome. ABBREVIATIONS RNA = Ribonucleic acid tRNA = Transfer ribonucleic acid mRNA = Messenger ribonucleic acid MRS = Methionyl-tRNA synthetase TCSTS = Two-component signal transduction system FAB = Fatty acid biosynthesis ACP = Acyl carrier protein HTS = High-throughput screening SAR = Structure activity relationship PDF = Peptide deformylasse RTI = Respiratory tract infection ORF = Open reading frame [3] Apfel, C.M.; Takacs, B.; Fountoulakis, M.; Stieger, M.; Keck, W. J. Bacteriol., 1999, 181, 483. [4] Tatusov, R.L.; Koonin, E.V.; Lipman, D.J. Science, 1997, 278, 631. REFERENCES [1] Fleischmann, R.D.; Adams, M.D.; White, O.; Clayton, R.A.; Kirkness, E.F.; Kerlavage, A.R.; Bult, C.J.; Tomb, J.F.; Dougherty, B.A.; Merrick, J.M. Science, 1995, 269, 496. [5] Tatusov, R.L.; Natale, D.A.; Garkavtsev, I.V.; Tatusova, T.A.; Shankavaram, U.T.; Rao, B.S.; Kiryutin, B.; Galperin, M.Y.; Fedorova, N.D.; Koonin, E.V. Nucleic Acids Res., 2001, 29, 22. [2] Venter, J.C.; Adams, M.D.; Myers, E.W.; Li, P.W.; Mural, R.J.; Sutton, G.G.; Smith, H.O.; Yandell, M.; Evans, C.A.; Holt, R.A.; Gocayne, J.D.; Amanatides, P.; Ballew, R.M.; Huson, D.H.; Wortman, J.R.; Zhang, Q.; Kodira, C.D.; Zheng, X.H.; Chen, L.; Skupski, M.; Subramanian, G.; Thomas, P.D.; Zhang, J.; Gabor Miklos, G.L.; Nelson, C.; Broder, S.; Clark, A.G.; Nadeau, J.; McKusick, V.A.; Zinder, N.; Levine, A.J.; Roberts, R.J.; Simon, M.; Slayman, C.; Hunkapiller, M.; Bolanos, R.; Delcher, A.; [6] Arigoni, F.; Talabot, F.; Peitsch, M.; Edgerton, M.D.; Meldrum, E.; Allet, E.; Fish, R.; Jamotte, T.; Curchod, M.L.; Loferer, H. Nat. Biotechnol., 1998, 16, 851. [7] Tomb, J.F.; White, O.; Kerlavage, A.R.; Clayton, R.A.; Sutton, G.G.; Fleischmann, R.D.; Ketchum, K.A.; Klenk, H.P.; Gill, S.; Dougherty, B.A.; Nelson, K.; Quackenbush, J.; Zhou, L.; Kirkness, E.F.; Peterson, S.; Loftus, B.; Richardson, D.; Dodson, R.; Khalak, H.G.; Glodek, A.; 306 Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Holmes et al. McKenney, K.; Fitzegerald, L.M.; Lee, N.; Adams, M.D.; Venter, J.C. .Nature, 1997, 388, 539. [27] Cheng, Q.; Campbell, E.A.; Naughton, A.M.; Johnson, S.; Masure, H.R. Mol. Microbiol., 1997, 23, 683. [8] Reich, K.A.; Chovan, L.; Hessler, P. J. Bacteriol., 1999, 181, 4961. [28] Novak, R.; Cauwels, A.; Charpentier, E.; Tuomanen, E. J. Bacteriol., 1999, 181, 1126. [9] Akerley, B.J.; Rubin, E.J.; Novick, V.L.; Amaya, K.; Judson, N.; Mekalanos, J.J. Proc. Natl. Acad. Sci. USA, 2002, 99, 966. [29] Novak, R.; Henriques, B.; Charpentier, E.; Normark, S.; Tuomanen, E. Nature, 1999, 399, 590. [30] [10] Norgren, M.; Caparon, M.G.; Scott, J.R. Infect. Immun., 1989, 57, 3846. Rodrigue, A.; Quentin, Y.; Lazdunski, A.; Mejean, V.; Foglino, M. Trends Microbiol., 2000, 8, 498. [31] [11] Hynes, W.L.; Hancock, L.; Ferretti, J.J. Infect. Immun., 1995, 63, 3015. Throup, J.P.; Koretke, K.K.; Bryant, A.P.; Ingraham, K.A.; Chalker, A.F.; Ge, Y.; Marra, A.; Wallis, N.G.; Brown, J.R.; Holmes, D.J.; Rosenberg, M.; Burnham, M.K. M o l . Microbiol., 2000, 35, 566. [12] Buckley, N.D.; Lee, L.N.; LeBlanc, D.J. J. Bacteriol., 1995, 177, 5028. [32] Fabret, C.; Hoch, J.A. J. Bacteriol., 1998, 180, 6375. [13] O'Connell, C.; Pattee, P.A.; Foster, T.J. J. Gen. Microbiol., 1993, 139, (Pt 7), 1449. [33] Martin, P.K.; Li, T.; Sun, D.; Biek, D.P.; Schmid, M.B. J. Bacteriol., 1999, 181, 3666. [14] Meinnel, T.; Mechulam, Y.; Blanquet, S. In t R N A : Structure, Biosynthesis, and Function; Soll, D.; RajBhandary, U. Eds. AMS Press, Washington, D.C. 1995, 251. [34] Payne, D.J.; Warren, P.V.; Holmes, D.J.; Ji, Y.; Lonsdale, J.T. Drug Discov. Today, 2001, 6, 537. [35] Bjork, G.R. In tRNA: Structure, Biosynthesis, and Function; Soll, D., RajBhandary, U. Eds. AMS Press, Washington, D.C. 1995, 165. [36] Bjork, G.R. In Escherichia coli and Salmonella. Cellular and molecular biology; Neidhardt, F.C. ed. AMS Press, Washington, D.C. 1996, 861. [37] Hjalmarsson, K.J.; Bystrom, A.S.; Bjork, G.R. J. Biol. Chem., 1983, 258, 1343. [38] Redlak, M.; Andraos-Selim, C.; Giege, R.; Florentz, C.; Holmes, W.M. Biochemistry, 1997, 36, 8699. [39] Cabedo, H.; Macian, F.; Villarroya, M.; Escudero, J.C.; Martinez-Vicente, M.; Knecht, E.; Armengod, M.E. EMBO J., 1999, 18, 7063. [40] Green, S.M.; Malik, T.; Giles, I.G.; Drabble, W.T. Microbiology, 1996, 142, 3219. [41] Macarron, R.; Hertzberg, R.P. In High Throughput Screening Methods and Protocols; Janzen, W.P. ed. Humana Press, Totowa, NJ. 2002, 1. [15] Schon, A.; Kannangara, C.G.; Gough, S.; Soll, D. Nature, 1988, 331, 187. [16] Grundy, F.J.; Haldeman, M.T.; Hornblow, G.M.; Ward, J.M.; Chalker, A.F.; Henkin, T.M. J. Bacteriol., 1997, 179, 3767. [17] Green, C.J.; Vold, B.S. GenBank Accession number L36472, 1994. [18] Payne, D.J.; Wallis, N.G.; Gentry, D.R.; Rosenberg, M. Current Opinion in Drug Discovery and Development, 2000, 3, 177. [19] Appleby, J.L.; Parkinson, J.S.; Bourret, R.B. Cell, 1996, 86, 845. [20] Egger, L.A.; Park, H.; Inouye, M. Genes Cells, 1997, 2, 167. [21] Stock, J.B.; Ninfa, A.J.; Stock, A.M. Microbiol. Rev., 1989, 53, 450. [22] Barrett, J.F.; Hoch, J.A. Antimicrob. Agents Chemother., 1998, 42, 1529. [42] Viswanadhan, V.N.; Balan, C.; Hulme, C.; Cheetham, J.C.; Sun, Y. Curr. Opin. Drug Discov. Devel., 2002, 5, 400. [23] Barrett, J.F.; Goldschmidt, R.M.; Lawrence, L.E.; Foleno, B.; Chen, R.; Demers, J.P.; Johnson, S.; Kanojia, R.; Fernandez, J.; Bernstein, J.; Licata, L.; Donetz, A.; Huang, S.; Hlasta, D.J.; Macielag, M.J.; Ohemeng, K.; Frechette, R.; Frosco, M.B.; Klaubert, D.H.; Whiteley, J.M.; Wang, L.; Hoch, J.A. Proc. Natl. Acad. Sci. USA, 1998, 95, 5317. [43] Lipinski, C.A.; Lombardo, F.; Dominy, B.W.; Feeney, P.J. Advanced Drug Delivery Reviews, 1997, 23, 3. [44] Miller, W.H.; Seefeld, M.A.; Newlander, K.A.; Uzinskas, I.N.; Burgess, W.J.; Heerding, D.A.; Yuan, C.C.; Head, M.S.; Payne, D.J.; Rittenhouse, S.F.; Moore, T.D.; Pearson, S.C.; Berry, V.; DeWolf, W.E. Jr.; Keller, P.M.; Polizzi, B.J.; Qiu, X.; Janson, C.A.; Huffman, W.F. J. Med. Chem., 2002, 45, 3246. [45] Payne, D.J.; Miller, W.H.; Berry, V.; Brosky, J.; Burgess, W.J.; Chen, E.; DeWolf, W.E. Jr.; Fosberry, A.P.; Greenwood, R.; Head, M.S.; Heerding, D.A.; Janson, C.A.; Jaworski, D.D.; Keller, P.M.; Manley, P.J.; Moore, T.D.; Newlander, K.A.; Pearson, S.; Polizzi, B.J.; Qiu, X.; Rittenhouse, S.F.; Slater-Radosti, C.; Salyers, K.L.; Seefeld, M.A.; Smyth, M.G.; Takata, D.T.; Uzinskas, I.N.; Vaidya, K.; Wallis, N.G.; Winram, S.B.; Yuan, C.C.; [24] Guenzi, E.; Gasc, A.M.; Sicard, M.A.; Hakenbeck, R. Mol. Microbiol., 1994, 12, 505. [25] Zahner, D.; Grebe, T.; Guenzi, E.; Krauss, J.; van der, L.M.; Terhune, K.; Stock, J.B.; Hakenbeck, R. Microb. Drug Resist., 1996, 2, 187. [26] Pestova, E.V.; Havarstein, L.S.; Morrison, D.A. Mol. Microbiol., 1996, 21, 853. A Genomics Approach to Drug Discovery Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 307 S.; Huang, C.H.; Kieser, T.; Larke, L.; Murphy, L.; Oliver, K.; O'Neil, S.; Rabbinowitsch, E.; Rajandream, M.A.; Rutherford, K.; Rutter, S.; Seeger, K.; Saunders, D.; Sharp, S.; Squares, R.; Squares, S.; Taylor, K.; Warren, T.; Wietzorrek, A.; Woodward, J.; Barrell, B.G.; Parkhill, J.; Hopwood, D.A. Nature, 2002, 417, 141. Huffman, W.F. Antimicrob. Agents Chemother., 2002, 46, 3118. [46] Baum, E.Z.; Montenegro, D.A.; Licata, L.; Turchi, I.; Webb, G.C.; Foleno, B.D.; Bush, K. Antimicrob. Agents Chemother., 2001, 45, 3182. [47] Morgan, B.P.; Scholtz, J.M.; Ballinger, M.D.; Zipkin, I.D.; Bartlett, P.A. Journal of the American Chemical Society, 1991, 113, 297. [67] Zhang, L.; Fan, F.; Palmer, L.M.; Lonetto, M.A.; Petit, C.; Voelker, L.L.; St John, A.; Bankosky, B.; Rosenberg, M.; McDevitt, D. Gene, 2000, 255, 297. [48] Hopkins, A.L.; Groom, C.R. Nat. Rev. Drug Discov., 2002, 1, 727. [68] [49] Jarvest, R.L.; Berge, J.M.; Berry, V.; Boyd, H.F.; Brown, M.J.; Elder, J.S.; Forrest, A.K.; Fosberry, A.P.; Gentry, D.R.; Hibbs, M.J.; Jaworski, D.D.; O'Hanlon, P.J.; Pope, A.J.; Rittenhouse, S.; Sheppard, R.J.; Slater-Radosti, C.; Worby, A. J. Med. Chem., 2002, 45, 1959. Fan, F.; Lunsford, R.D.; Sylvester, D.; Fan, J.; Celesnik, H.; Iordanescu, S.; Rosenberg, M.; McDevitt, D. Plasmid, 2001, 46, 71. [69] Ji, Y.; Zhang, B.; Van Horn, S.F.; Warren, P.; Woodnutt, G.; Burnham, M.K.; Rosenberg, M. Science, 2001, 293, 2266. [70] Ji, Y.; Marra, A.; Rosenberg, M.; Woodnutt, G. J . Bacteriol., 1999, 181 , 6585. [50] Demain, A.L. Nat. Biotechnol., 2002, 20, 331. [51] Henkel, T.; Brunne, R.M.; Muller, H.; Reichel, F. Angew. Chem. Int. Ed., 1999, 38, 643. [71] Gmuender, H.; Kuratli, K.; Di Padova, K.; Gray, C.P.; Keck, W.; Evers, S. Genome Res., 2001, 11, 28. [52] Lee, M.L.; Schneider, G. J. Comb. Chem., 2001, 3, 284. [72] [53] Wijkmans, J.C.; Beckett, R.P. Drug Discov. Today, 2002, 7, 126. Lucchini, S.; Thompson, A.; Hinton, J.C. Microbiology, 2001, 147, 1403. [73] Ye, R.W.; Wang, T.; Bedzyk, L.; Croker, K.M. J . Microbiol. Methods, 2001, 47, 257. [74] Freiberg, C.; Brunner, N.A. Targets, 2002, 1, 20. [75] Tao, H.; Bausch, C.; Richmond, C.; Blattner, F.R.; Conway, T. J. Bacteriol., 1999, 181, 6425. An, H.; Haly, B.D.; Cook, P.D. J. Med. Chem., 1998, 41, 706. [76] Ye, R.W.; Tao, W.; Bedzyk, L.; Young, T.; Chen, M.; Li, L. J. Bacteriol., 2000, 182, 4458. Manly, S.P.; Padmanabha, R.; Lowe, S.E. Methods Mol. Biol., 2002, 190, 153. [77] Bammert, G.F.; Fostel, J.M. Antimicrob. Chemother., 2000, 44, 1255. [78] Schena, M.; Shalon, D.; Davis, R.W.; Brown, P.O. Science, 1995, 270, 467. [79] Wilson, M.; DeRisi, J.; Kristensen, H.H.; Imboden, P.; Rane, S.; Brown, P.O.; Schoolnik, G.K. Proc. Natl. Acad. Sci. USA, 1999, 96, 12833. [54] [55] [56] Kung, P.P.; Casper, M.D.; Cook, K.L.; Wilson-Lingardo, L.; Risen, L.M.; Vickers, T.A.; Ranken, R.; Blyn, L.B.; Wyatt, J.R.; Cook, P.D.; Ecker, D.J. J. Med. Chem., 1999, 42, 4705. Agents [57] Ganesan, A. Drug Discov. Today, 2002, 7, 47. [58] Rich, J.O.; Michels, P.C.; Khmelnitsky, Y.L. Current Opinion in Chemical Biology, 2002, 6, 161. [59] van Dongen, M.; Weigelt, J.; Uppenberg, J.; Schultz, J.; Wikstrom, M. Drug Discov. Today, 2002, 7, 471. [60] Carr, R.; Jhoti, H. Drug Discov. Today, 2002, 7, 522. [80] [61] Goksoyr, T.V.J.; Daae, F.L. Appl. Environ. Microbiol., 1990, 56, 782. Betts, J.C.; Lukey, P.T.; Robb, L.C.; McAdam, R.A.; Duncan, K. Mol. Microbiol., 2002, 43, 717. [81] Salama, N.; Guillemin, K.; McDaniel, T.K.; Sherlock, G.; Tompkins, L.; Falkow, S. Proc. Natl. Acad. Sci. USA, 2000, 97, 14668. [82] Whiteley, M.; Bangera, M.G.; Bumgarner, R.E.; Parsek, M.R.; Teitzel, G.M.; Lory, S.; Greenberg, E.P. Nature, 2001, 413, 860. [83] de Saizieu, A.; Certa, U.; Warrington, J.; Gray, C.; Keck, W.; Mous, J. Nat. Biotechnol., 1998, 16, 45. [84] de Saizieu, A.; Gardes, C.; Flint, N.; Wagner, C.; Kamber, M.; Mitchell, T.J.; Keck, W.; Amrein, K.E.; Lange, R. J. Bacteriol., 2000, 182, 4696. [85] Dunman, P.M.; Murphy, E.; Haney, S.; Palacios, D.; Tucker-Kellogg, G.; Wu, S.; Brown, E.L.; Zagursky, R.J.; Shlaes, D.; Projan, S.J. J. Bacteriol., 2001, 183 , 7341. [62] Yap, W.H.; Li, X.; Soong, T.W.; Davies, J.E. J. Ind. Microbiol., 1996, 17, 179. [63] Wang, G.Y.; Graziani, E.; Waters, B.; Pan, W.; Li, X.; McDermott, J.; Meurer, G.; Saxena, G.; Andersen, R.J.; Davies, J. Org. Lett., 2000, 2, 2401. [64] Kao, C.M.; Katz, L.; Khosla, C. Science, 1994, 265, 509. [65] Gokhale, R.S.; Tsuji, S.Y.; Cane, D.E.; Khosla, C. Science, 1999, 284, 482. [66] Bentley, S.D.; Chater, K.F.; Cerdeno-Tarraga, A.M.; Challis, G.L.; Thomson, N.R.; James, K.D.; Harris, D.E.; Quail, M.A.; Kieser, H.; Harper, D.; Bateman, A.; Brown, S.; Chandra, G.; Chen, C.W.; Collins, M.; Cronin, A.; Fraser, A.; Goble, A.; Hidalgo, J.; Hornsby, T.; Howarth, 308 [86] Current Drug Targets - Infectious Disorders, 2002, Vol. 2, No. 4 Fitzgerald, J.R.; Sturdevant, D.E.; Mackie, S.M.; Gill, S.R.; Musser, J.M. Proc. Natl. Acad. Sci. USA, 2001, 98, 8821. Holmes et al. [87] Cummings, C.A.; Relman, D.A. Emerg. Infect. Dis., 2000, 6, 513. [88] Kato-Maeda, M.; Gao, Q.; Small, P.M. Cell Microbiol., 2001, 3, 713.
