The Prospects for Microbial Genomics Providing Novel Exploitable Antibacterial Targets

Thomas J. Dougherty3 and John F. Barrett" "Pfizer Global Research & Development, Groton, CT, 06340 bMerck Research Laboratories, Rahway, NJ

Introduction - Antibiotics have been a major triumph in applied medical science since their introduction in the last century. The rapid improvement of patients afflicted with heretofore deadly infections led to these compounds being termed "miracle drugs". Unfortunately, it has become increasingly clear that bacterial resistance to these compounds is rising at a rate that threatens to undermine their future utility (1,2). In this sense, antimicrobials stand in a somewhat unique position among drug classes. Compounds that affect human physiology, such as lipid-lowering agents, ACE inhibitors, and anti-inflammatory agents, do not engender resistance to their pharmacological targets in the human population. However, in the case of infectious agents we are dealing with rapidly evolving populations of organisms placed under selective pressure by antibiotics (3). Exacerbating the situation is the fact that many of the existing antimicrobial classes are derived from natural products, and pre-existing resistance genes were present in the producer organisms (4). These genes have been mobilized on genetic elements such as transposons and plasmids, and disseminated widely among many pathogens. There is a clear need for new classes of antibiotics as part of the strategy to combat the emerging antibiotic resistance problem.

Genomics Development - In the mid-1990s, technologies were developed that permitted the rapid sequencing of the entire genetic complement of an organism (5,6). These involved random ("shotgun") sequencing of a gene library of an organism by the use of automated, high-throughput DNA sequencers. This process was coupled with sophisticated sequence assembly programs running on highspeed computers to link the multiple, redundant, relatively short sequences into a coherent genome. These methods were initially applied to microorganisms, and Haemophilus influenzae became the first organism whose genome was completed using what was termed "shotgun" sequencing and assembly in 1995 (7). This marked the beginning of the genomics era, and multiple microorganism genomes were rapidly sequenced, as were landmark organisms such as Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophilia melanogaster and the human genome (8-11). The avalanche of gene sequence information created opportunities to understand organisms in fundamentally new ways that are still very much in their nascent stages. In the human genome, the initial mining of the genomic information has revealed new members of important gene families, such as fibroblast growth factors, keratin, and TGFps (11). The extensive syntectic relationships between human and mouse genome permits the application of mouse mutants and genetics to solving physiological roles of human genes.

Microbial Genomics - With regard to microbial genomes, a fundamentally different approach is applied in mining genome information for the discovery of new antibacterial targets. In microbes, a small handful of essential, physiological processes are the targets of the existing classes of antibiotics. These include terminal stages in the synthesis of the bacterial cell wall peptidoglycan (e.g., p-lactams, glycopeptides), protein synthesis (macrolides, aminoglycosides, tetracyclines, oxazolidinones), and DNA topoisomerases (fluoroquinolones) (12). The advent of microbial genomics has suddenly thrust the complete "parts list" for multiple pathogens into the hands of biologists. By identifying additional essential gene products, it is believed that targets will become available to exploit in the identification of novel inhibitors. Development of some of these inhibitors into drugs will expand the classes of antimicrobial compounds. Most importantly, novel antimicrobials should circumvent resistance mechanisms to current antimicrobials.

Currently, the number of microbial genomes that have been sequenced since the initial H. influenzae genome is approximately fifty-nine, with many additional projects in progress. Limiting the examples to pathogens alone, organisms such as Streptococcus pyogenes, several strains of Streptococcus pneumoniae, Chlamydia trachomatis, and Chlamydia pneumoniae, Borrelia burgdorferi, Treponema pallidum, Mycobacterium tuberculosis and an enterohaemorrhagic strain of Escherichia coli (13-21) are among many additional pathogen genome sequences in the public domain. Two important organisms in which the bulk of microbial physiology has been performed, E. coli K-12 and Bacillus subtilis, are also available. These are key, as many of the putative function annotations of pathogen genomes are based on sequence similarities to gene products in these two organisms. With individual microbial genomes running between roughly 400,000 to over 4 million base pairs, the sheer volume of this information has made it imperative to use high-speed computer workstations to handle, sort, compare, and analyze the databases. In order to rationalize the information, programs have been devised to identify probable genes, compare multiple genomes, gather functional annotation evidence, and sort the gene products into biochemical pathways (22-26).

Applications of Genomics to Antimicrobial Target Identification - With the complete genome sequences of a large number of pathogenic bacteria now available, strategies to identify gene products that may be targets for inhibitors have been developed and deployed in a number of laboratories. These strategies can be based on a number of different suppositions, and the differences in approach reflect different models of the types of targets sought, as well as the methods to identify such targets. Figure 1 illustrates the general pathways and outcomes of these approaches to the common problem of novel target identification in bacteria. It is possible to take one of two differing general gene disruption approaches to identify novel antibacterial targets. It is also possible to test the disrupted genes on synthetic growth media (in vitro) or in the context of the infected host (in vivo animal infection model). In either case, the failure to recover a gene that has been disrupted on a plate, or in an infected animal, is considered evidence for the essential nature of the gene under the testing circumstances.

Identification of Novel Genomic Targets: Targeted Knockouts - As indicated above, the identification of novel antimicrobial targets can be approached by inactivation of genes to determine if their protein products are essential for bacterial survival. One way to identify these potential targets is by first examining and comparing genomes computationally to generate a list of likely target knockout candidates. Computational methods are available for comparing multiple microbial genomes to one another to identify, by sequence similarity, probable conserved gene functions (23,24,27-29). It is also possible to subtract eukaryotic sequences at different levels of similarity to reduce the likelihood of adverse events due to cross-inhibition of host functions.

This conservation of sequence similarity is used as a surrogate for spectrum of a potential antimicrobial agent against the target. Once a list of gene knockout candidates has been finalized, a decision on which bacterial species will

Gene Products

i

r -

r

Test for survival of organism

0 0

Post a comment