TEXT

Top Abstract Text References

Klebsiella pneumoniae is an important nosocomial pathogen that causes a wide range of infections, including pneumonia, bacteremia, urinary tract infections, and sometimes life-threatening septic shock. As an opportunistic pathogen, it primarily attacks immunocompromised individuals who are hospitalized and/or suffering from severe underlying diseases, such as diabetes mellitus, chronic alcoholism, or pulmonary obstruction (23). Many clinical strains of K. pneumoniae are highly resistant to antibiotics, indicating the relative ineffectiveness of current therapy.

During infections, bacterial pathogens must adapt to various changes in order to persist and proliferate in appropriate locations and to circumvent host defenses. It is reasonable to assume that the expression of many K. pneumoniae genes that participate in pathogenesis could be specifically induced within the host. Ideally, these in vivo-expressed (IVE) genes would serve as useful drug targets and vaccine candidates. Several approaches, including in vivo expression technology (IVET) (11, 15), comparative genomics (2), microarray DNA chips (7), signature-tagged mutagenesis (18, 26), differential display-PCR (1), and differential fluorescence detection (31), have allowed the identification of genes that are essential or specifically activated during infections. Nevertheless, none of these approaches has been applied to K. pneumoniae, primarily due to the limited number of mutants and genetic tools available for the bacterium.

IVE technology (IVET) is a powerful technique that has been used successfully for several important pathogens, including Salmonella enterica serovar Typhimurium (15, 16), Yersinia enterocolitica (33), Staphylococcus aureus (14), Pseudomonas aeruginosa (32), Escherichia coli (12), and Actinobacillus pleuropneumoniae (9). The original IVET, designed to identify promoters that turn on in vivo in S. enterica, used a tandem set of in vivo and in vitro promoterless genes as the reporter (15). There are now several different modifications of the IVET, such as the use of auxotrophic markers and antibiotic resistance genes, and induction of site-specific recombinase as the basis for the selection systems (12, 14, 32). In view of the limited genetic tools available for K. pneumoniae, we have designed a novel IVET selection system for this heavily encapsulated bacterium.

Rationale for the galU-based IVET selection system. The rationale for the IVET developed for K. pneumoniae is shown in Fig. 1. Instead of using a set of tandem reporter genes, our IVET system incorporates a copy of the promoterless galU gene as a dual-function reporter. The galU gene encodes the enzyme UDP glucose pyrophosphorylase, which regulates the supply of UDP galactose and UDP glucose, two major precursors for the biosynthesis of capsule polysaccharides (CPS) and lipopolysaccharides (LPS) in most enteric bacteria. A GalU mutant of K. pneumoniae produces defective forms of CPS and LPS and hence loses virulence and the mucoid colony phenotype (5). In addition, the K. pneumoniae GalU strain is incapable of fermenting galactose, a property that can be readily distinguished by using MacConkey-0.4% galactose agar (5). These unique properties make the galU gene an ideal reporter system for IVE gene selection. The bacterial strains which are able to survive in vivo selection while exhibiting a white nonmucoid colony phenotype on MacConkey-galactose agar would indicate that the DNA fragment upstream of the galU reporter contains an IVE promoter.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Overall selection strategy for the galU-based IVET system.

Construction of galU-based reporter gene system. Our IVET selection vector, designated pYC016 (Fig. 2), was constructed by using the mobilizable plasmid pSUP102 (27) as the backbone. Plasmid pYC016 was engineered sequentially in the order described as follows. Initially, a PCR-amplified copy of the Pseudomonas aeruginosa PAO1 galU gene coding sequence was inserted into the unique ClaI site of pSUP102. Subsequently, a nonessential BamHI site in the plasmid was eliminated by end-filling with Klenow fragment of DNA polymerase I, followed by religation. Finally, a cassette from pKK232-8 (Amersham-Pharmacia, Piscataway, N.J.) that contains a transcriptional terminator, multiple cloning sites, and a set of three-way translational stop codons was inserted into a ClaI site of the vector. The main features of pYC016 include (i) a unique BamHI site which is compatible with Sau3AI-digested chromosomal DNA fragments, (ii) use of the galU gene of P. aeruginosa PAO1, with the intention of minimizing the possibility of homologous recombination that could occur between the chromosomal and plasmid-borne galU genes in K. pneumoniae, (iii) the presence of translational stop codons in all three reading frames preceding the ribosome-binding site of the galU gene to ensure transcriptional fusions between inserted DNA fragments and the galU gene, and (iv) a transcriptional terminator, rrnB T₁ (4), upstream of the cloned chromosomal fragment to prevent transcription of the galU gene from a fortuitous plasmid promoter.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Diagram of K. pneumoniae IVET vector pYC016, showing the locations of the transcriptional terminator (rrnB T1), multiple cloning sites (MCS), three-way translational stop codons, and promoterless galU gene. Only the BamHI site in the multiple cloning site is unique in the vector.

Construction of K. pneumoniae IVET library. Two isogenic K. pneumoniae strains were used in this study. K. pneumoniae CG43 is a clinical isolate which is highly virulent to laboratory mice, with a 50% lethal dose (LD₅₀) of 10 CFU (22). K. pneumoniae CG43-17 is a GalU derivative of CG43, generated previously by Tn5 insertion mutagenesis. The virulence of K. pneumoniae CG43-17 was found to be significantly attenuated (LD₅₀, ~10⁶ CFU) (5). High-molecular-weight chromosomal DNA of K. pneumoniae CG43 was purified and partially digested with Sau3AI. Fragments ranging from 0.5 to 1 kb in size were purified from agarose gels and ligated into the unique BamHI site of pYC016. The ligation mixture was transformed into E. coli S17-1 (27) (hsdR recA pro RP4-2 [Tc::Mu; Km::Tn7]) and selected on Luria-Bertani (LB) agar supplemented with chloramphenicol (30 µg/ml). Approximately 5 × 10⁵ of the transformants, 90% of which contain a plasmid with an insert, were obtained. These plasmids were then mobilized into K. pneumoniae CG43-17 by conjugation. The transconjugants were selected on MacConkey-0.4% galactose plates supplemented with kanamycin (75 µg/ml), chloramphenicol (30 µg/ml), and ampicillin (50 µg/ml). After an 8-h incubation at 37°C, red mucoid colonies were scored as Gal⁺, whereas white nonmucoid colonies were scored as Gal. The red-white (Gal⁺:Gal) ratio of the transconjugants was about 1:3.

Screening of K. pneumoniae IVE genes. A pilot study was performed with about 10³ CFU of the IVE library. The study demonstrated that after infections in BALB/c mice and subsequent recovery from spleen, the ratio of red to white colonies shifted from the original 1:3 to 60:1, indicating that most of the Gal clones were eliminated effectively in vivo. To reduce the number of mice used in this study as well as to minimize undesired bacteria of the Gal⁺ background, approximately 5 × 10⁴ mostly white nonmucoid colonies were collected manually and separated into three batches for the experiments that followed.

Verification of inducibility of galU fusion constructs in vivo and in vitro. The plasmid DNA was extracted from the 20 IVE clones and retransformed into K. pneumoniae CG43-17, and the transformants were tested individually for virulence in BALB/c. Approximately 1 × 10⁶ to 5 × 10⁷ CFU could be recovered from 1 g of spleen from the sick mice. The number is comparable to that of the wild-type K. pneumoniae CG43. In contrast, less than 10 CFU of K. pneumoniae CG43-17(pYC016) was observed under these conditions. Moreover, the use of a standard assay method (5) indicated that all these clones did not exhibit detectable UDP glucose pyrophosphorylase activity, confirming that the promoters were indeed turned off under in vitro growth conditions.

Nucleotide sequence determination of IVE genes. DNA sequence determination was carried out by the PCR-mediated Taq DyeDeoxy Terminator Cycle sequencing kit on an Applied Biosystems model 373A DNA sequencer. The homology search of the GenBank/EMBL and SwissProt databases was performed using the BLAST programs provided by the National Center of Biotechnology Information through the Internet. The result of the sequence analysis is shown in Table 1.

Inducibility of IVE promoters under iron deprivation and oxidative stress. On entering the host, bacterial pathogens must circumvent iron deprivation and the attack of reactive oxygen species produced by the immune cells. Therefore, it is reasonable to assume that some of the IVE promoters identified in this study may encode a product that assists the bacteria in countering these stresses. To verify the possibility, the IVE clones were spotted individually on MacConkey-galactose plate containing either 200 µM 2',2'-dipyridyl, an iron chelator, or 10 µM paraquat, a superoxide generator. The phenotype of the colonies after overnight incubation at 37°C was then examined. Two IVE clones, AJ277397 and AJ292298, representing the promoters of two iron acquisition genes, iucA and fepA, respectively, were found to display the red mucoid colony phenotype in the presence of 2',2'-dipyridyl that reflects the activation of the promoters under iron deprivation conditions (Fig. 3).

View larger version (54K): [in this window] [in a new window]   FIG. 3.   Phenotypes of IVE clones on MacConkey-galactose agar containing 10 µM paraquat (A) or 100 µM 2',2'-dipyridyl (B). The bright red color of the colonies indicates expression of the galU reporter gene.