Contraselectable Streptomycin Susceptibility Determinant for Genetic Manipulation and Analysis of Helicobacter pylori

Autor: Dangeruta Kersulyte, Daiva Dailidiene, Giedrius Dailide, Douglas E. Berg
Rok vydání: 2006
Předmět:
Zdroj: Applied and Environmental Microbiology. 72:5908-5914
ISSN: 1098-5336
0099-2240
Popis: Genetic studies of the gastric pathogen Helicobacter pylori (21) often entail (i) the construction of cloned DNAs or PCR products containing a drug resistance determinant inserted into or near a gene of interest, (ii) DNA transformation, and (iii) selection for resistant transformants, which arise by homologous recombination and replacement of recipient DNA sequences by corresponding sequences from donor DNA. Although this strategy has been used in hundreds of H. pylori studies, successfully overcoming the rarity of natural transformation events, it becomes seriously limiting or flawed in at least four interesting situations, in each case because transformants retain resistance determinants at targeted loci: (i) if changes at numerous loci in the same strain are needed, e.g., in studies of phenotypes determined by multiple genes with additive or redundant effects (because only a few selectable resistance markers are available for H. pylori); (ii) if donor resistance determinants might affect downstream gene expression, or cellular physiology more generally (e.g., in many cases the resistance enzymes use cellular metabolites to modify antibiotics and thereby confer resistance to them [9]); (iii) if alleles with subtle (e.g., point mutation) differences are to be compared; or (iv) if alleles from many strains are to be studied in a common genetic background. Such limitations can be overcome by using a recipient strain that contains a contraselectable marker at the locus of interest. This allows the desired transformants to be selected by the loss of this recipient marker, thereby bypassing the need for a conventional resistance determinant in donor DNA. Three genes that have been used for contraselection in other systems are thyA (thymidylate requirement; trimethoprim sensitivity) (4, 34), sacB (sucrose sensitivity) (5, 27, 33), and rpsL (streptomycin sensitivity) (25, 33). H. pylori is naturally trimethoprim resistant. This could be ascribed to (i) its apparent lack of a thyA gene (based on BLASTP homolog searches [31]); (ii) an intrinsic resistance of its enzyme for reduced folate synthesis, an apparent dihydrofolate reductase-dihydropteroate synthase chimaera (18); and/or (iii) other factors. An early report had indicated that sacB could serve as a contraselectable marker in H. pylori but did not describe the details of efficiency or complications that may have been encountered (8). In our H. pylori sacB experiments, however, selection for resistance to a range of sucrose concentrations gave far more background growth of nominally sensitive cells than was expected based on Escherichia coli experiences (33, 34), and conditions for sucrose-based differential killing of sacB-containing H. pylori cells varied among trials (unpublished data). Other groups had also found contraselection of sacB to be difficult in H. pylori although better than having no such marker at all (T. L. Cover, unpublished data; R. Haas, unpublished data). A third contraselection strategy, implemented here, uses the rpsL (ribosomal protein S12) gene and is based on the dominance of wild-type streptomycin-sensitive (Strs) alleles to resistance-conferring mutant alleles (16, 25). Historically, streptomycin killing has been associated with diverse effects, including membrane damage and irreversible streptomycin uptake. Paradoxically, these were all blocked by the bacteriostatic translation inhibitor chloramphenicol. This implicated the capacity to synthesize proteins, even though streptomycin also inhibited protein synthesis. Sublethal streptomycin concentrations suppressed nonsense mutations, increased mutation, and caused membrane fragility. Biochemical studies showed that streptomycin allowed the continuation of translation once begun, but with errors in translation (misreading), and that it also allowed ribosomes to bind to mRNAs but blocked them from initiating translation (10). Davis' early unifying explanation for streptomycin's lethality (10) focused on errors it induced during translation and a resultant accumulation of defective proteins. He focused in particular on membrane proteins, interpreting that defects in them were responsible for the disruption of cell membrane integrity, massive leakage of ions and molecules, and cell death. This could explain the dominance of streptomycin sensitivity over resistance, as could the binding of Strs ribosomes to mRNAs, blocking the access of resistant ribosomes to these mRNAs, failure to translate them, and their ensuing degradation (10). Recent exquisitely detailed molecular and genetic analyses provide mechanistic understanding. Streptomycin binds several specific 16S rRNA loops and thereby diminishes ribosome flexibility and the changes in conformation that charged tRNA binding normally induces. It stabilizes a “ribosomal ambiguity” (ram) state in which there is little if any of the proofreading needed for error-free protein synthesis (7, 15, 22, 24). Protein S12, which is mutated in Strr strains, binds rRNA sequences near sites of streptomycin binding. Most positions in which change confers resistance are in S12 domains that interact with the rRNA target. In addition, most Strr mutant S12 proteins increase the accuracy of translation, apparently by destabilizing the ram state that streptomycin itself induces; the most extreme of such rpsL (S12) alleles make growth streptomycin dependent. This can be overcome by mutations in ribosomal proteins S4 and S5, that in turn stabilize the ram state in this dynamic and finely tuned ribonucleoprotein machine (7, 14). The two S12 residues that are most frequently changed in streptomycin-resistant H. pylori are Lys43 (Lys42 in E. coli numbering), which contacts the rRNA-bound streptomycin directly, and Lys88 (Lys87 in E. coli numbering), which is nearby in the structure (7, 14, 32). Although the Lys-to-Arg changes at these codons seem to be quite innocuous (14), other amino acid replacements can markedly diminish bacterial fitness. In particular, a large fitness cost and a resulting selection for compensatory mutations in other genes have been documented in E. coli and Salmonella enterica serovar Typhimurium using strains with Lys42 replaced with Asn (19, 26) (not Arg). More critically, the Lys88Arg replacement used here, and also the Lys43Arg allele, have each been incorporated into H. pylori strains used for mouse infections, without obvious negative effects on fitness (12, 13, 23). Streptomycin contraselection for chromosomal gene replacement had been developed for H. pylori previously (13) but has not been much used, perhaps because very few of the Strr isolates recovered after transformation had sustained the desired replacement. This, we suspected, was due to the near identity of the strs and strr H. pylori rpsL alleles used: frequent gene conversion between them would result in unwanted Strr gene convertants vastly outnumbering the desired Strr transformants. If correct, it seemed that the rpsL gene might still be useful for H. pylori genetic engineering if the rpsL gene conversion could be reduced. Generalized recombination (gene conversion included) depends on close matches between participating DNA molecules and is reduced by sequence divergence, even in species that, like H. pylori, lack the ability to cleave mismatch-containing (heteroduplex) DNAs (11, 20, 31). The rpsL genes of H. pylori and Campylobacter jejuni differ by some 18% in overall DNA sequence, but the encoded S12 proteins are 95% similar in amino acid sequence. The present study was initiated with an expectation or hope that H. pylori ribosomes containing the C. jejuni S12 protein would be functional and Strs and that DNA sequence divergence between the rpsL genes of C. jejuni and H. pylori would diminish gene conversion sufficiently for the effective recovery of transformants by streptomycin contraselection. The rpsL,erm (streptomycin susceptibility, erythromycin resistance) cassette that we constructed and tested in these studies is diagrammed in Fig. ​Fig.11. FIG. 1. Structure of the 1.5-kb rpsL,erm cassette, which confers dominant streptomycin susceptibility and selectable erythromycin resistance. Open boxes designate open reading frames; solid line indicates noncoding sequences. The cassette contains 272 bp of ...
Databáze: OpenAIRE
Externí odkaz: