Patent Grant
7910337
The present invention relates generally to bacterial diseases and more particularly to a methodology for the generation of bacterial strains useful in detecting the mechanism of action of antibacterial compounds and to determine genes and gene products that interact with existing antibiotics.
Antibiotics are useful agents in curing human and animal diseases caused by bacterial infections. Currently there are numerous antibiotics that control bacterial disease by interfering with a specific bacterial gene product, disrupting the biochemistry of the organism, and either preventing the organism from growing or killing the organism. However, antibiotic resistant strains of bacterial pathogens constantly emerge. Thus, there is a continual need to discover and develop new antibiotics effective against the new strains of pathogens. One way to find new agents is to discover and develop chemical compounds that act at a bacterial gene product not targeted by any current antibiotics. These potential targets for new antibiotics have been characterized as ‘essential’ by various methods that demonstrate the gene or gene product is required for bacterial growth. With the complete sequencing of the genomes of pathogens and various biochemical and microbiological analyses, numerous methodologies are available to define the set of ‘essential’ proteins in bacteria. Such proteins and the genes that encode them are prime targets for antibiotic development, as small molecule inhibitors of these functions can thereby serve to either stop the growth or to kill pathogenic bacteria in an infection.
Prior to genomic-based approaches, discovery and development of antimicrobial drugs relied upon the vast literature of bacterial physiology and biochemistry. For example, it has long been known that beta-lactam antibiotics inhibit and kill bacterial pathogens by interfering with the formation of the bacterial cell wall. The specific process inhibited by the beta-lactam antibiotics are the transpeptidation reactions that link peptidoglycan units in the final stages of cell wall biosynthesis. Transpeptidation is catalyzed by penicillin binding proteins (PBPs) encoded by bacterial genes. Thus, PBPs are validated targets for antibiotics and remain the subject of antibiotic discovery and development.
The peptidoglycan monomer unit is made by bacteria in a conserved biochemical pathway (
Another set of known antibacterial targets are those enzymes used for bacterial protein synthesis. Many different useful antibiotics are known to act through inhibition of various steps of protein synthesis. One step requires the activity of tRNA synthetase enzymes that “charge” by means of covalent acylation a specific amino acid to its cognate tRNA molecule. These charged tRNAs are basic building blocks in protein synthesis. Inhibition of the charging process can severely limit protein synthesis. Small molecule inhibition of tRNA synthetases has been demonstrated, and certain inhibitors have been commercialized for use in antibacterial chemotherapy. For example, the commercial antibiotic mupirocin inhibits bacterial isoleucyl-tRNA synthetase. Many patents have been filed and granted on other chemical inhibitors of tRNA synthetase, including methionyl-tRNA synthetase.
Still another set of known antibacterial targets are those enzymes used for bacterial DNA synthesis. One biochemical required for DNA synthesis is tetrahydrofolate, a metabolite produced by the enzyme dihydrofolate reductase (DHFR). Thus, inhibition of DHFR can result in an antibiotic effect. The widely used antibiotic trimethoprim acts through selective inhibition of bacterial DHFR. Many patents concern the discovery and development of antibiotics having DHFR as the mechanism of action.
The present invention concerns the development of strains of bacteria that can be used to detect specific mechanisms of antibiotic action and thereby aid in antibiotic discovery and development. The method of creating the strains of bacteria involves generating and expressing antisense RNA that confers a hypersensitive phenotype to the bacteria that is specific for any particular antibiotic mechanism of action.
Experimental means have long been sought to modulate gene expression in bacteria. Conditional reduction in expression of target essential genes can make the host bacterium more sensitive to small molecules that inhibit the products of these genes. One way of doing this is to conditionally express RNA that is complementary to the mRNA transcribed from target genes. The formation of a double-stranded RNA species can result in blockage of translation or to degradation of the targeted mRNA. A substantial body of research has described the incidence of naturally occurring antisense post-transcriptional regulation in bacteria. Antisense regulation has been shown to be involved in the regulation of plasmid copy number, global regulation of cellular physiology, and to post-transcriptional regulation of certain cellular functions. These findings have inspired a number of efforts to experimentally create and exploit antisense regulation systems that can attenuate regulation of particular genes of interest. One experimental system was demonstrated as early as 1984 when researchers reported that portions of the lacZ gene, when expressed from an inducible promoter in the antisense direction, could cause specific and substantial attenuation of beta-glactosidase expressed from the lacZ gene. (Pestka, S. et al., (1984). Subsequently, others generated a histidine auxotroph phenotype by transforming Mycobacterium smegmatis with a plasmid containing an antisense fragment to its hisD gene (“Development and use of a conditional antisense mutagenesis system in mycobacterium.” Parish T, and Stoker N G (1997), FEMS Microbiology Letters 154, 151-157). In another work (“Expression of an Antisense hla Fragment in Staphylococcus aureus Reduces Alpha-Toxin Production In Vitro and Attenuates Lethal Activity in a Murine Model.” Kernodle et al., (1997) Infection and Immunity, 65, 179-184) a 600 base pair fragment of the hla gene was cloned in antisense orientation into a plasmid and transformed into S. aureus. The resulting strain made 16-fold less alpha-toxin thereby significantly blocking lethality of these cells in a murine infection model. A method described for the protist Dictyostelium discoideum used an un-regulated promoter to drive expression of random cDNAs in the antisense orientation in order to globally survey the resulting phenotypes of affected genes. (“Mutagenesis and gene identification in Dictyostelium by shotgun antisense” Spann et al. (1996) Proc. Natl. Acad. Sci. USA, 93, 5003-5007). By this method, numerous genes were functionally cataloged by the phenotypic effects of specific antisense-expressed cDNAs. These and other works demonstrate that certain antisense fragments cloned into suitable vectors can be used to alter bacterial strain phenotype by lowering the expression of specific genes in essential and non-essential metabolic pathways. The resulting inducer-dependent phenotype of the bacteria engineered with antisense genes includes results from altered metabolism, altered virulence, and reduced growth.
The “shotgun antisense” method described above for Dictyoselium could also be used to globally to survey for essential genes if an inducible promoter is used. In this way, resulting cells would not exhibit the resulting growth sensitivity until expression of the antisense DNA is turned on. Two similar but independently conducted antisense-based essential gene surveys were recently described (“Identification of critical Staphylococcal genes using conditional phenotypes generated by antisense RNA” Ji et al., (2001), Science 293:2266-2269; “A genome-wide strategy for the identification of essential genes in Staphylocccus aureus” Forsyth et al., (2002) Molecular Microbiology 43:1387-1400). In both experiments, genomic DNA of S. aureus was purified, fragmented, and then “shotgun cloned” behind a plasmid-borne inducible promoter. After transformation of this shotgun library into S. aureus, resulting colonies were surveyed for growth in the presence or absence of inducer. Those that failed to grow due to expression of a gene fragment in the antisense orientation were collected and analyzed by DNA sequencing, and it was found that this pool of clones was highly enriched for known essential genes involved in such cellular functions as cell-wall biosynthesis, DNA replication, protein translation, RNA transcription, and fatty acid biosynthesis.
Other publications focused on antisense as a technique useful in identifying essential bacterial genes from an entire genome. Two patents were issued to Elitra Pharmaceuticals based on processes that were used in the Forsyth et al. publication above: U.S. Pat. No. 6,228,579 and U.S. Pat. No. 6,924,101, describe a process of surveying fragments of microbial genomic DNA for their capacity to reduce or block proliferation when expressed in the microorganism in the antisense orientation. The methods in the patent include fragmentation of genomic DNA from the organism of interest, cloning the fragments adjacent to an inducible promoter sequence on a plasmid, introducing the plasmid into the microorganism of interest, and comparing growth of the organism in the presence and absence of an inducer compound or stimulus that results in a dependent presence or absence of expression of the cloned fragment. Clones containing an antisense fragment that showed growth inhibition in the presence of inducer compound are selected, and the ‘essential’ gene for the clone is deduced by sequencing the antisense fragment. The presumption of these methods is that those fragments that cause reduction or blockage of proliferation of the microorganism do so by a mechanism that is specific to the function of the gene from which the fragment was derived, implying that this gene is one whose encoded cellular function is required for normal proliferation. Both patents state in their abstract and describe in their claims, “Antisense fragments that result in lethality when expressed indicate that the endogenous gene is a proliferation gene.” Additional claims in the antisense patent literature include the method of using antisense-expression for screening compounds for the purpose of identifying specific inhibitors of the protein-target of the antisense, especially U.S. Pat. No. 6,924,101.
While growth inhibition or lethality resulting in expression of an antisense fragment may be due to specific mRNA attenuation as described above, these phenotypes may also be due to unspecific mechanisms. Even though a gene fragment is expressed in the antisense orientation, it could still be translated to produce a toxic “cryptic peptide” that can inhibit growth (Lopes J M, Soliman N, Smith P K, Lawther R P, Mol Microbiol. 1989 August; 3(8): 1039-51. “Transcriptional polarity enhances the contribution of the internal promoter, ilvEp, in the expression of the ilvGMEDA operon in wild-type Escherichia coli K12”). Indeed, one analysis showed that substantial open reading frames can occur in the antisense orientation in many E. coli genes (Merino et al., (1994), “Antisense overlapping open reading frames in genes from bacteria to humans.” Nucleic Acids Res. 22, 1903-1908). The prevalence of antisense-oriented coding regions could result in identification of ostensibly antisense-oriented gene fragments that cause unspecific growth sensitivity phenotypes. Another way in which RNA fragments can inhibit growth in ways other than by specific antisense-based mRNA attenuation is demonstrated by aptamers. (Ellington A D, Szostak J W, “In vitro selection of RNA molecules that bind specific ligands.” Nature, 1990 Aug. 30; 346(6287):818-22; Blum, J H, Dove, S L, Hocschild, A, and Melalanos, J J (2000) “Isolation of peptide aptamers that inhibit intracellular processes” PNAS 97, 2241-2246). Such RNA fragments, which may be either in the sense or antisense orientation to the genes that they originated from, may actually form small molecules that may inhibit cellular functions entirely unrelated to the gene of origin
In any instance of antisense growth inhibition, expression of antisense RNA fragments can result in bacterial strains that are hypersensitive to many different classes of chemical compounds unrelated to the target gene. Therefore, a limitation of current antisense methodology is that the resulting strains may not be attenuated for a specific metabolic pathway. A new method is needed in order to demonstrate a direct cause-and-effect mechanism for any particular growth-inhibiting antisense fragment.
Another limitation of current antisense methods involves the specificity of generating antisense to any given gene. This limitation is related to artifacts discussed above. Specifically, it is not possible at present to predict the required sequence and length of antisense RNA for any given gene that will produce the desired specific antisense effect. Hasan et al., (“Antisense RNA does not significantly affect expression of the galK gene of Escherichia coli or the N gene of coliphage lambda.” Gene 72, 247-252) reported numerous failed attempts at producing a bona fide antisense effect despite three years of effort with many configurations of the genes and promoters. Hasan concludes that “ . . . clear-cut regulation is more an exception than a rule (with antisense), and requires the use of a suitable gene and careful design, combined with strong conviction and good luck.” This report demonstrates that simply cloning and expressing the inverted sequence of a gene will not reliably result in an antisense effect.
Another limitation of existing methodology concerns the use of genomic DNA as the source of the fragments for generating an antisense-based strains. First and foremost, existing methods define proliferation genes as those that generate a growth inhibited phenotype upon generation of an antisense strain to that gene. As indicated in the research cited, many antisense fragments will produce growth inhibition phenotype without connection to the gene from which it was derived. More importantly, targets for antibiotics that include these genes have in fact been defined by the integrated knowledge of bacterial genetics, physiology, and biochemistry. There have not been reports in the literature of an essential biochemical process identified using antisense that was not known through other means. Thus, it is inefficient to use genomic DNA fragmentation as the starting point for a method of constructing antisense-based sensitized strains for antibiotic discovery.
Second, the use of genomic DNA does not ensure that any particular gene will be adequately represented. The use of genomic DNA is biased towards genes with sequences that are more amenable to inhibition by antisense. Genes of equal value as antibiotic targets are each different in their susceptibility to antisense inactivation because of their different sequences and length. Because the gene fragment required for generating a specific antisense strain for a given gene cannot be specified, there is a significant risk that a useful antisense fragment will not be generated for any particular gene. To generate the number and diversity of fragments required to ensure a specific antisense strain will be represented for any particular gene, the use of genomic DNA is inappropriate. A bacterial genome may consist of several thousand genes, thus requiring millions of clones to be generated and examined.
Another limitation when using genomic DNA as a source for generating the fragments is that bacterial genomes often contain two or more copies of a gene encoding an essential function. Antisense to only one of these genes may not be sufficient to demonstrate a phenotype. For example, MurA in peptidoglycan biosynthesis is encoded by two different genes in many Gram-positive pathogens. The mode of action of the antibiotic fosfomycin is through inhibiting the MurA protein, demonstrating that MurA is an essential process. Antisense methodology as described in the literature on the whole genome would erroneously exclude MurA as an essential process.
One important process of antibiotic discovery is the determination of the mode of action for any given antibacterial compound. Mode of action determination can help differentiate those compounds with new and specific mechanisms of action from those compounds with non-specific or old modes of action. The utility of antisense in creating strains depleted in a particular target protein and thus sensitized to inhibitors of that target have been published. However, given that a particular target has been pre-selected for antibiotic discovery, it is inconvenient, inefficient, and incomplete to create any particular antisense-based strain using DNA from the entire genome. A new process is needed for specific work.
The identity of “proliferation” genes and other “essential” genes in bacteria has been accomplished a number of different ways and precede the use of antisense for this purpose.
Prior to genomics-based methods, microbial processes such as cell wall biosynthesis, protein synthesis, isoprenoid biosynthesis, and tetrahydrofolate synthesis were known to be useful sites of antibiotic action. Tools to track mode of action for antibiotics targeting these processes are useful in antibiotic discovery and development.
The presence of a gene fragment in antisense orientation can cause inducer-dependent growth-inhibited phenotype by mechanisms other than post-transcriptional reduction in mRNA. The potential for artifacts makes the selection of growth-inhibited phenotype an insufficient criterion for selecting specific antisense strains.
Many limitations exist for the methods of creating antisense-based hypersensitive strains for antibiotic discovery, including gene-specific requirements for antisense structure (sequence and length), an inconvenient cloning requirement to ensure adequate representation for any given gene, and multiple genes encoding essential processes.
Current published methods for generating antisense-based strains for detecting mode of action are subject to artifact, and are inconvenient, inefficient, and incomplete for use in antibiotic drug discovery.
The invention provides a method for generating and selecting drug-sensitizing antisense DNA fragments. In one embodiment, the method includes identifying a gene of interest using knowledge of bacterial physiology, biochemistry, genetics, genomics, and other means. The method includes PCR amplification of a gene of interest using genomic DNA as a template; fragmentation of the DNA by sonication or other means; selecting DNA fragments no longer than 400 base pairs; ligating the DNA fragments into a suitable expression plasmid with a selectable marker; transforming the plasmids containing the DNA fragments into the organism from which the gene of interest originated; and selecting clones from transformed cells that show a phenotypic difference of the clone grown in the presence of the inducer relative to the phenotype in the absence of inducer.
In another embodiment, the invention provides that the discernible phenotype is growth rate; relative sensitivity to a growth inhibiting compound; requirement for addition of a nutrient to the growth medium; morphology such as shape; relative sensitivity to osmotic stress, or colony size, for example.
In another embodiment, the invention provides a method for selecting bacterial strains containing specific drug-sensitizing DNA fragments including selecting those strains wherein the inserted DNA fragment is oriented antisense to some portion of the gene of interest selecting those strains wherein the mRNA level of the organism in the presence of inducer is less than 25% of the levels in the absence of inducer; selecting those strains according to the method of the invention wherein the magnitude of the phenotype is dependent on the concentration of the inducer.
“Beta-lactam” A beta-lactam (p-lactam) is a lactam with a heteroatomic ring structure, consisting of three carbon atoms and one nitrogen atom. The beta-lactam ring is part of several antibiotics, such as penicillin, which are therefore also called beta-lactam antibiotics. These antibiotics work by inhibiting the bacterial cell wall synthesis. Examples of beta-lactam antibiotic classes include penicillins, cephalosporins, and cabapenems.
“Antibiotic” is a compound used to control infections. Examples of antibiotics include
Cell wall inhibitors include: Beta-lactams include (penicillin G, penicillin V, nafcillin, methicillin, oxacillin, cloxacillin, dicloxacillin, flucloxacillin, ampicillin, amoxacillin, carbenicillin, ticarcillin, azocillin, mezlocillin, pipericillin, Cephalosporins, cephalothin, cefazolin, cefalexin, cefuroxime, cefamandole, cefoxitin, cefaclor, moxalactam, cefaperazone, ceftazidime, ceftriaxone, clavulanic acid, sulbactam, imipenem, aztreonam, Cefoxitin, cefozolin, Cefaclor).
Inhibitors of peptidoglycan biosynthesis, Inhibitors of GlmU, MurA (Fosfomycin) Inhibitors of MurB (dihydropyrrolones), Inhibitors of Undecaprenyl pyrophosphate Synthetase (hydantoins, sulfonamides), Inhibitors of LpxC (UDP-3-O-acyl N-acetylglycosamine deacetylase), Lipopeptides (daptomycin), Isoprenoid biosynthesis inhibitors (bacitracin, phosmidomycin), Fatty acid biosynthesis inhibitors (cerulenin, triclosan, isoniazid) Protein synthesis inhibitors, aminoglycosides (streptomycin, gentamicin and kanamycin), tetracyclines, chloramphenicol, macrolides (erythromycin, azithromycin, clarithromycin), lincosamides (lincomycin and clindamycin), and oxazolidinones (linezolid) tRNA synthetase inhibitors such as mupirocin, MetRS inhibitors (catechols, prolines, quinolones), Peptide deformylase inhibitors (hydroxamates), RNA synthesis inhibitors such as rifampicin, DNA gyrase inhibitors such as quinolones (nalidixic acid, ciprofloxacin) Antifolates (trimethoprim and sulfamethoxazole), “EC50” and “IC50” mean the concentration of compound resulting in 50% of the growth rate of the organism compared to the untreated control.
A general protocol for generating and selecting antisense-encoding DNA fragments is shown in
Our interest in antisense has been in development of specific antisense strains of Bacillus anthracis and of Staphylococcus aureus for use in tracking the mechanism of action of new antibiotic candidates. To overcome limitations in the current methods antisense methods, we developed the method outlined in
It is not obvious that such a method should be required for producing such useful mode of action tools, because previous publications indicate that many antisense fragments should work and that the only criterion necessary for selection is an inducer-dependent growth inhibition. Thus, the frequency of artifact in antisense is not appreciated. This method is novel because it begins with the target gene identified in advance and requires the generation of thousands of fragments from which to begin selection of those that will generate the desired drug sensitized strain.
We have built specifically-sensitized antisense strains using validated antibiotic targets. Four example targets are MurB (UDP-N-acetylenolpyruvoylglucosamine reductase [EC: 1.1.1.15 8]) in peptidoglycan biosynthesis (
Each of these four genes was subjected to the process outlined in
Table 1. B. anthracis strain performance featuring fragments from B. anthracis genes subjected to process in
Table 1 includes results for methionyl-tRNA synthetase (MetRS) encoded by the metRSl gene. Subjecting this gene to the protocols in
Table 1 also shows results for the MurB target, the second step of peptidoglycan biosynthesis (
The method described in
In addition to genes corresponding to cell wall biosynthesis, protein synthesis, and DNA synthesis, we have built antisense strains using the method described in
Examples 1 & 2 are genetic maps of the regions of the metRSl and metRS genes that generated xylose-responsive growth inhibitory antisense fragments from B. anthracis and S. aureus. These examples illustrate that the region of the gene and the fragment length producing useful xylose responsive growth inhibitory fragments are efficiently generated using the method described in
Example 3 is a genetic map of the regions of the murB2 gene from B. anthracis that generated xylose-responsive growth inhibitory antisense fragments. This example illustrates how certain genes have regions that are more or less likely to generate useful fragments.
Examples 4 & 5 are genetic maps of the regions of the glmU gene and murAl gene from B. anthracis that generated xylose-responsive growth inhibitory antisense fragments. These examples illustrate that each gene in the peptidoglycan biosynthesis pathway can be used in the method described in
Examples 6, 7, and 8 are genetic maps of the regions of the dxs, dxr-2, and gcpE genes from B. anthracis that generated xylose-responsive growth inhibitory antisense fragments. These examples illustrate that genes from the non-mevalonate isoprenoid biosynthesis pathway can be used in the method described in
Example 9 are genetic maps of the regions of the gyrA and gyrB genes from B. anthracis that generated xylose-responsive growth inhibitory antisense fragments. These examples illustrate that genes involved in DNA replication can be used in the method described in
Example 10 is the genetic map of the regions of the fabF gene from B. anthracis that generated xylose-responsive growth inhibitory antisense fragments. This examples illustrates that genes involved in fatty acid biosynthesis can be used in the method described in
Although the invention has been described with reference to the above examples, it will be understood that modifications and variations are encompassed within the spirit and scope of the invention. Accordingly, the invention is limited only by the following claims.
This application claims priority under §35 U.S.C. 119(e) from provisional application Ser. No. 60/782,961 filed Mar. 15, 2006, which is incorporated herein by reference in its entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 20040029129 | Wang et al. | Feb 2004 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| 20070218481 A1 | Sep 2007 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60782961 | Mar 2006 | US |
