点击显示 收起
Department of Microbiology, National Taiwan University College of Medicine, and Departments of Internal Medicine and Forensic Medicine, National Taiwan University Hospital, Taipei, Taiwan
Background.
Primary pyogenic liver abscess (PLA) with septic complication by Klebsiella pneumoniae is an emerging infectious disease.
Methods and results.
Using DNA microarray hybridization, we identified a 20-kb chromosomal region that contained 15 open-reading frames (ORFs), including an iron-uptake system (kfu), a phosphoenolpyruvate sugar phosphotransferase system (PTS), and 6 unknown ORFs. The region was more prevalent among tissue-invasive strains (35/46) than among noninvasive strains (19/98) (P < .0001, 2 test). To test the role played by this region in pathogenesis, 3 different deletion mutants (NTUH-K2044 , K2044 [ORF79], and K2044 ) were constructed. Only the kfuABC mutants showed decreased virulence in mice, compared with the wild-type strain. An in vitro assay confirmed the involvement of kfu in iron acquisition. There was a high correlation rate (85%) between the kfu/PTS region and 2 tissue invasionassociated chromosomal regions (allS and magA). Moreover, all 3 regions were present in strains that caused PLA plus endophthalmitis or meningitis.
Conclusion.
Our results suggest that chromosomal heterogeneity is present in tissue-invasive K. pneumoniae strains. A genotype containing all 3 regions is strongly associated with PLA and metastatic infection. These regions may serve as convenient markers for the rapid diagnosis of emergent tissue-invasive strains.
Klebsiella pneumoniae is an important hospital-acquired pathogen that is a frequent cause of urinary tract infection, septicemia, and pneumonia in immunocompromised individuals; it is also an important pathogen with respect to community-acquired infectious diseases, such as community-acquired pneumonia [1]. In Taiwan, K. pneumoniaeassociated primary pyogenic liver abscess (PLA) has recently become an important emerging infectious disease [28]. This disease is also a global concern, as is attested by reports from North America [9], Europe [10, 11], and Asia [1214].
Fulminant tissue-invasive K. pneumoniae infections can attack healthy persons who have no history of hepatobiliary disease, and only 50% of patients have a predisposing condition, such as diabetes mellitus [28]. PLA is manifest with other septic metastatic lesions, including pyogenic meningitis and endophthalmitis, in 10%12% of patients [35, 7]. Despite aggressive antibacterial strategies, significant morbidity and mortality still exists, especially in those with diabetes mellitus [28].
Using transposon mutagenesis and full genome expression analysis, we recently identified genomic segments of 33 kb (magA) and 22 kb (allS), which were absent in the genome of MGH 78578 and present in most tissue-invasive strains from patients with PLA, meningitis, or endophthalmitis [15, 16]. Therefore, it is probable that these tissue-invasive strains might represent a specific genotype that harbors specific regions involved in pathogenesis and prevalence.
DNA microarray technology provides a useful tool for assessment of the differences and changes in bacterial genomes [17, 18]. Therefore, we used this technology to compare genomic variations between tissue-invasive strains and noninvasive strains.
PATIENTS, MATERIALS, AND METHODS
K. pneumoniae strains and culture conditions.
K. pneumoniae isolates were obtained consecutively from cultures of blood samples obtained from patients between 1997 and 2003. A total of 46 tissue-invasive K. pneumoniae strains were obtained from patients with PLA or meningitis at National Taiwan University Hospital (NTUH). Of these, 22 were isolated from patients with diabetes mellitus, and 24 were isolated from patients who had previously been healthy.
For comparison, 98 noninvasive strains were obtained from patients with sepsis but without any tissue-invasive disease (such as PLA, meningitis, and endophthalmitis). Of these strains, 33 were obtained from patients at NTUH, 32 were obtained from patients at Far Eastern Memorial Hospital (FEMH; Banciao, Taiwan), and 11 were obtained from patients at En Chu Kong Hospital (ECKH; Sansia, Taiwan). Also, 22 strains from the United States, including MGH 78578, were purchased from the American Type Culture Collection. MGH 78578 was chosen for full genome sequencing at Washington University (St. Louis, MO; available at: http://www.genome.wustl.edu/projects/bacterial/). Information pertaining to the 144 strains is listed in table 1. All strains were identified and cultured according to standard clinical microbiology methods [19].
Microarray construction and hybridization.
The genomic library was constructed from a clinical isolate, NTUH-K2044, obtained from a patient with PLA plus meningitis in a phagemid [20]. DNA fragments in phagemids of K. pneumoniae were amplified by polymerase chain reaction (PCR) with primers in vectors and were spotted onto a nylon membrane (Roche) by a computer-controlled XYZ translation system (PM500; Newport) [21].
Probe preparation and hybridization.
Genomic DNA from 4 PLA strains (NTUH-K2044, A1208, A3021, and A5011) and 3 noninvasive strains (N3423, N3529, and N5322) were extracted and were labeled with biotin-16-dUTP (Perkin Elmer) by a randomly primed polymerization reaction. The microarray membrane was prehybridized in 2 mL of hybridization buffer for 4 h at 65°C and was hybridized for 16 h at 68°C. The membrane was washed twice with 2× standard saline citrate (SSC) containing 0.1% SDS for 5 min at room temperature and was then washed 3 times with 0.1× SSC containing 0.1% SDS for 15 min at 65°C each time. Colorimetry detection and image analysis were then performed as described elsewhere [22].
Recombinant DNA techniques and plasmids.
K. pneumoniae deletion mutants were constructed by replacing the deletion region with a kanamycin (Km) cassette in a double-crossover integration of chromosomal DNA. All primers used in the present study are listed in table 2. To generate the NTUH-K2044 (kfu) mutant, a PCR fragment amplified by primers PVAR KO-1 and PVAR KO-2 was cloned into a pGEM-T Easy vector (Promega). PVAR KO-1(iPCR) and PVAR KO-2(iPCR) primers were used for inverse PCR with plaque-forming-unit polymerase (MBI Fermentas). A blunt-end Km gene was phosphorylated by use of a polynucleotide kinase (New England Biolabs) and was ligated to the inverse PCR product to generate the kfu disruption fragment. We cloned the fragment into a pUT suicide vector [23] containing an EcoRI site and added a second marker (spectionmycin; Spe) to the ApaLI site. pUT-(kfu) was transformed into wild-type NTUH-K2044 to generate a kfu deletion mutant by conjugation. Deletion clones were selected by Kmr and Spes. The same procedures were used for open-reading frames (ORFs) 79 and phosphoenolpyruvate sugar phosphotransferase system (PTS) deletion constructs. All of the deletion mutants were confirmed by PCR with multiple primer pairs and sequence determination.
Transcomplementation of Escherichia coli H1443.
An aroB- E. coli H1443 strain is deficient of siderophore for growth when cultured in medium containing iron chelator 2,2-dipyridyl (Sigma) [24, 25]. Transformed E. coli H1443 with plasmid pBR322::kfuABC, plasmid pSZ1 [25], or pBR322 were grown overnight at 37°C in Luria-Bertani (LB) broth. Each culture was diluted in fresh LB broth supplemented with 0.1, 0.2, or 0.5 mmol/L 2,2-dipyridyl. The growth rate was monitored spectrophotometrically at 620 nm.
RNA isolation and reverse-transcription (RT) PCR.
Total RNA was isolated from K. pneumoniae cultured at the exponential phase of growth, as described elsewhere [15]. For each RT, 5 g of RNA was used with 2 pmol of RT-PCR primer and 200 U of M-MLV reverse transcriptase (Invitrogen). Reaction mixtures without reverse transcriptase were included as negative controls. PCR was performed with 10% of each RT reaction volume under 30 cycles of amplification.
Murine experiments.
Initially, we infected BALB/cByl mice intraperitoneally (ip). However, later experiments showed that intragastric (ig) inoculation had a higher sensitivity to differential virulence [15, 16]. Mice were administered ig [26] either wild-type NTUH-K2044, MGH 78578, NTUH-K2044 (kfu), K2044 (ORF79), or K2044 (PTS) (103106 cfu; 4 mice for each dose). For ig inoculation, we carefully slipped a 1-mm polyethylene flexible tube (Becton Dickinson) past the pharynx into the stomach (5 cm of intubation) of each mouse and delivered 0.2 mL (103106 cfu) of bacterial suspension. Anesthesia was not used for this procedure. Mice were monitored for 4 weeks; upon death, the liver and brain were removed, and histopathological examination was conducted. Surviving mice were killed at the end of the 4 weeks. The LD50 was calculated as described elsewhere [27]. Survival was analyzed by Kaplan-Meier analysis with a log-rank test; P < .05 was considered to be statistically significant.
Slot-blot hybridization.
Ten micrograms of genomic DNA from K. pneumoniae strains were vacuum blotted onto nylon membranes. Hybridization was performed for 16 h at 68°C with each biotin-labeled probe generated by PCR. The gene encoding 23S rRNA was used as a positive control. Detection was performed by use of the Southern-Light Chemiluminescent detection system (Tropix), in accordance with the manufacturer's instructions.
RESULTS
DNA microarray hybridization.
A total of 3146 PCR clones were randomly selected for the microarray. The coverage rate was 88%, according to the formula N = ln(1 - P)/ ln(1 - f) [28]. To test the redundancy of the library, 798 of the 3146 clones were randomly selected for sequencing. These clones contained 678 distinct sequences, representing a redundancy rate of 15%.
Comparison of 4 tissue-invasive strains (A1208, A3021, A5011, and NTUH-K2044) and 3 noninvasive strains (N3423, N3529, and N5322) revealed 12 clones with significantly decreased hybridization signals (defined as 3 SDs of the mean ratio) in the noninvasive strains (figure 1 and table 3). These 12 clones were sequenced and then compared with the 10× shotgun sequences of K. pneumoniae MGH 78578 (available at: http://www.genome.wustl.edu/projects/bacterial/kpneumoniae/). Of them, there were 10 clones (clones 312) that bore no similarity in sequence to MGH 78578. Of these 10 clones, 8 overlapped and extended to encompass an 9-kb fragment, whereas the other 2 matched to a 200-kb plasmid (pLVPK) of K. pneumoniae CG43 [29] (hereafter, "the large plasmid").
Sequencing of the flanking regions of the 8 clones in NTUH-K2044.
The flanking regions of the 9-kb fragment were sequenced from the genomic library until both ends matched in MGH 78578. A 19,640-bp fragment (GenBank accession number AB115591) that was obtained replaced a 5292-bp fragment in the genome of MGH 78578 (figure 2). BLAST searches revealed that this fragment contained 15 ORFs (figure 2 and table 4). The common flanking regions of NTUH-K2044 and MGH 78578 respectively contained oppA and an ORF encoding a putative diogenase subunit. The overall GC content of this region was 56.9%, similar to the 57.7% GC content of the remainder of the genome.
Annotation of this fragment revealed a modular structure with 4 regions (figure 2 and table 4). Region 1 (2934704) carried ORF13. These ORFs showed homology to proteins involved in glycogen phosphorylase (ORF1), an antianti- factor (ORF2), and a putative protease (ORF3). Region 2 (51048743) harbored ORF46. These ORFs exhibited high homology to the bacterial ferric ironuptake system, which is a bacterial ABC iron transport system [30]. They include Sfu of Serratia marcescens [25, 31], Hit of Haemophilus influenzae [32], Yfu of Yersinia pestis [33], Afu of Actinobacillus pleuropneumoniae [34], and Fbp of Neisseria gonorrhoeae [35]. Therefore, these ORFs were putatively designated kfuA, kfuB, and kfuC, respectively ("kfu" stands for Klebsiella ferric iron uptake). Region 3 (899011,786) contained ORF79. ORF7 displayed no significant homology to any sequences in the database. However, ORF8 revealed similarity to the gene yijO, which encodes a putative ARAC-type regulatory protein in E. coli. ORF9 showed homology to a protein involved in the biosynthesis of mitomycin [36]. Region 4 (11,78619,553) extended from ORF10 to ORF15. These 6 ORFs showed high homology to the PTS. PTS catalyzes translocation with concomitant phosphorylation of sugars and hexitols and regulates metabolism in response to the availability of carbohydrates [37]. Some PTS proteins have been tentatively linked with bacterial virulence [3840]. This 20-kb region was designated "the kfu/PTS region."
Prevalences of the kfu/PTS region and the large plasmid among K. pneumoniae strains.
Because sequences of clones 1 and 2 were present in the noninvasive strain MGH 78578 (table 3), we studied the prevalences of the kfu/PTS region and the large plasmid among the clinical isolates to find regions that are specific to tissue-invasive strains. Genomic DNA extracted from 46 tissue-invasive and 98 noninvasive strains was used for PCR analysis. We used the inside and outside primers to detect the presence of the kfu/PTS region (figure 3), on the following basis: If a clinical strain contained the region being tested, then the primer pairs for the flanking regions (outside primers) should fail to amplify the fragments; however, the inside primers should amplify products with a predicted length. Conversely, if the clinical strain did not contain the region, then PCR with the outside primer pairs would be positive, whereas PCR with the inside primers would be negative. We also used the inside sequences of clones 11 and 12 to detect the existence of the large plasmid.
The prevalence of the kfu/PTS region was significantly higher in tissue-invasive strains than in noninvasive strains (35/46 vs. 19/98; P = .0001, 2 test) (table 1). However, no significant correlation was observed between the large plasmid and clinically invasive disease (42/46 vs. 76/98; P = .077, 2 test).
Analysis of deletion mutants.
Three deletion mutants (NTUH-K2044 , K2044 [ORF79], and K2044 ) were constructed by use of a suicide vector (figure 4A). The growth of the mutants was compared with that of the wild-type strain in nutrient-rich, undefined LB medium. However, no significant differences were found.
The effect on virulence of each mutation was investigated in a mouse model (figure 4B and 4C). All of the mice inoculated with 103 cfu survived. Half of the mice died when inoculated with 104 cfu of the wild-type strain, K2044 (ORF79), or K2044 (PTS); there were no significant differences in survival among these 3 groups (K2044 [ORF79] vs. wild-type strain, P = .76; K2044 vs. wild-type strain, P = .67; log-rank test) (figure 4C). In contrast, all of the mice inoculated with K2044 (kfu) at the doses of 104106 cfu survived and appeared to be healthy after 4 weeks. The survival of the mice inoculated with K2044 (kfu) differed significantly from that of the mice inoculated the wild-type strain (P = .0067, log-rank test) (figure 4C).
When inoculated with the wild-type strain, K2044 (ORF79), or K2044 (PTS) at the highest dose (106 cfu), most of the mice had died by 7 days after infection, with no obvious signs before death and with no pathological changes detected on histological examination; they were considered to have died of septic shock. At lower inoculation doses (104105 cfu), most mice died between day 12 and 21 after infection, with signs of lethargy, labored breathing, or trembling 12 days before death. Large liver and/or brain abscesses were observed in this group, and septic shock with possible organ failure was considered to be the cause of death. However, there was no histological change in the livers and brains obtained from the mice inoculated with K2044 (kfu) at doses of 104106 cfu (figure 5). Because the manifestations of K. pneumoniae infection observed in BALB/cByl mice were very similar to those observed in patients, we did not try other murine strains.
Functional analysis of the kfu system in K. pneumoniae.
K. pneumoniae kfuABC was found to be preceded by a putative 19-bp Fur box consensus sequence (figure 6A). The 3 genes of kfu were transcribed in the same direction and had short intergenic sequences (maximum 21 bp between kfuA and kfuB), suggesting that they were in a single transcriptional unit. RT-PCR analysis confirmed that the 3 genes were transcribed as a single operon (figure 6B).
Comparison of the growth rates of the wild-type strain, K2044 (kfu), MGH 78578, and a noninvasive strain, N3423, in an iron-chelated medium revealed slightly lower growth for K2044 (kfu) and MGH 78578 (figure 6C). The E. coli aroB mutant carrying kfuABC or sfuABC grew well under conditions of iron limitation. However, the growth of H1443 with vector only (pBR322) was significantly inhibited (figure 6D).
The distribution of the 3 chromosomal regions was further analyzed according to the clinical disease of the patients (table 5). Strains from the patients with PLA plus endophthalmitis or meningitis were positive for all 3 regions. However, when strains from the 4 patients with meningitis but without PLA were examined, 2 of the strains were found to be negative for all 3 regions, and 2 of the strains were found to be positive for the kfu/PTS and magA regions but not for the allS region.
DISCUSSION
In the present study, microarray comparison of the genomic DNA of 4 tissue-invasive K. pneumoniae strains and 3 noninvasive strains identified 12 clones with significantly different hybridization signals. Two clones that matched the sequence in MGH 78578 were probably the result of experimental variations, because both were present in the 2 groups of strains. Another 2 clones were located in a large plasmid. However, the prevalences of the large plasmid were invariant in the 2 groups of strains. The remaining 8 clones connected together and further extended to encompass a 20-kb fragment (table 3). Only the deletion mutant NTUH-K2044 (kfu) showed decreased virulence in vivo. Functional studies confirmed the operon's involvement in iron uptake. Because our array was estimated to cover only 88% of the entire K. pneumoniae genome, we could miss some regions specific to tissue-invasive strains; hence, the allS and magA regions were not identified by this method.
The 20-kb kfu/PTS region was markedly more prevalent among tissue-invasive strains. The kfu/PTS region encoded an iron-uptake system involved in virulence and fit the size range (10200 kb) of a pathogenicity island. However, the GC content of this region was similar to the remainder of the whole genome, and no mobile elements or insertion sequences were noted [42]. Thus, the fundamental nature of the heterogeneity of this specific region in K. pneumoniae awaits further study [43].
Acquisition of nutrients such as iron to sustain growth in the host environment is essential for bacterial pathogens to establish an infection. Iron uptake is also critical to pathogenesis as a vital cofactor for many components of microbial antioxidative stress defense, including such components as superoxide dismutase, catalase, and peroxidase [44]. We have shown here that the kfu operon was present in most of the genomes of the tissue-invasive K. pneumoniae strains we examined and was absent from most of the genomes of the noninvasive strains. It is possible that the presence of a functional kfu operon might principally or secondarily modulate virulence in vivo and so provide a strong competitive advantage to those strains that harbor it.
There was a significant difference in the prevalence of the kfu/PTS region between noninvasive control strains from NTUH and FEMH. Strains from FEMH were all isolated from patients with nosocomial infection, whereas strains from NTUH were all isolated from patients with community-acquired infection. This could be the reason for the genetic difference. In addition, there would be an epidemiological difference between strains from a medical center (NTUH) and a those from a community hospital (FEMH).
How K. pneumoniae enter into the bloodstream and liver has not yet been documented. However, bacterial cells would enter the bloodstream through M cells [45] or as a result of minor mucosal injuries of the gastrointestinal tract. Larger abscesses were found in the mice infected via ig inoculation, whereas, in the mice infected via ip inoculation, microabscesses were found. This finding supports the hypothesis that K. pneumoniae gains entry into the bloodstream via the gastrointestinal tract, because all venous returns in the gastrointestinal tract were collected via a portal vein into liver. Many bacteria were likely trapped in the liver by Kupffer cells; however, the tissue-invasive strains were resistant to phagocytosis and serum killing. This may result in the subsequent formation of liver abscesses.
The present study reinforces our awareness of the vital role that bacterial virulence factors play in the pathogenesis of PLA and metastatic infection. According to the prevalences of the 3 specific chromosomal regions in tissue-invasive strains, we were able to classify the 54 invasive strains into 3 groups: group 1 strains contained all 3 regions, group 2 strains contained 12 regions, and group 3 strains contained none of the regions. When all 3 regions were absent, the strains were only occasionally capable of causing PLA. The most interesting finding was that all of the strains from patients with PLA plus metastatic infection contained all 3 regions. Most of the group 3 strains were noninvasive (table 1), suggesting that they might be less virulent than the group 1 strains. That the group 1 strains had an increased ability to invade tissue may well relate to their enhanced survival and ability to compete for nutrients. The kfu/PTS region could enrich the ability of bacteria to secure iron, even in the relatively iron-deficient conditions of the human host. The allS region could help bacteria to compete for nitrogen sources via the allantoin-utilizing ability [15]. Moreover, the magA region could render bacteria resistant to phagocytosis by polymorphonuclear leukocytes and serum killing [16]. Therefore, the group 1 strains might be classified as a specific genotype that is associated with PLA and metastatic infection. Two strains from 4 patients with meningitis but without PLA were negative for all 3 regions, and the remaining 2 strains were negative for the allS region. These results suggest that different pathogenic mechanisms could be at work in patients with meningitis plus PLA and in those with meningitis only.
In conclusion, we have identified a 20-kb chromosomal kfu/PTS region in K. pneumoniae that is associated with iron acquisition and virulence. This region is more prevalent in tissue-invasive strains. Strains containing the kfu/PTS, allS, and magA regions are strongly associated with PLA and metastatic infection and may be classified as a specific genotype. However, a different virulence mechanism seems to be at work in patients with meningitis but without PLA. Therefore, the kfu/PTS region, as well as the allS and magA regions, can be exploited as a genetic marker for rapid molecular diagnosis and for tracing the source of these emergent tissue-invasive strains.
Acknowledgment
We thank Dr. Volkmar Braun, Universitat Tubingen (Tubingen, Germany), for providing Escherichia coli strain H1443 and pSZ1 plasmid.
References
1. Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11:589603. First citation in article
2. Chang SC, Fang CT, Hsueh PR, Chen YC, Luh KT. Klebsiella pneumoniae isolates causing liver abscess in Taiwan. Diagn Microbiol Infect Dis 2000; 37:27984. First citation in article
3. Cheng DL, Liu YC, Yen MY, Liu CY, Wang RS. Septic metastatic lesions of pyogenic liver abscess: their association with Klebsiella pneumoniae bacteremia in diabetic patients. Arch Intern Med 1991; 151:15579. First citation in article
4. Chiu CT, Lin DY, Liaw YF. Metastatic septic endophthalmitis in pyogenic liver abscess. J Clin Gastroenterol 1988; 10:5247. First citation in article
5. Fung CP, Chang FY, Lee SC, et al. A global emerging disease of Klebsiella pneumoniae liver abscess: is serotype K1 an important factor for complicated endophthalmitis Gut 2002; 50:4204. First citation in article
6. Ko WC, Paterson DL, Sagnimeni AJ, et al. Community-acquired Klebsiella pneumoniae bacteremia: global differences in clinical patterns. Emerg Infect Dis 2002; 8:1606. First citation in article
7. Liu YC, Cheng DL, Lin CL. Klebsiella pneumoniae liver abscess associated with septic endophthalmitis. Arch Intern Med 1986; 146:19136. First citation in article
8. Wang JH, Liu YC, Lee SS, et al. Primary liver abscess due to Klebsiella pneumoniae in Taiwan. Clin Infect Dis 1998; 26:14348. First citation in article
9. Saccente M. Klebsiella pneumoniae liver abscess, endophthalmitis, and meningitis in a man with newly recognized diabetes mellitus. Clin Infect Dis 1999; 29:15701. First citation in article
10. Cahill M, Chang B, Murray A. Bilateral endogenous bacterial endophthalmitis associated with pyogenic hepatic abscess . Br J Ophthalmol 2000; 84:1436. First citation in article
11. Casanova C, Lorente JA, Carrillo F, Perez-Rodriguez E, Nunez N. Klebsiella pneumoniae liver abscess associated with septic endophthalmitis . Arch Intern Med 1989; 149:1467. First citation in article
12. Lee KH, Hui KP, Tan WC, Lim TK. Klebsiella bacteraemia: a report of 101 cases from National University Hospital, Singapore. J Hosp Infect 1994; 27:299305. First citation in article
13. Ohmori S, Shiraki K, Ito K, et al. Septic endophthalmitis and meningitis associated with Klebsiella pneumoniae liver abscess. Hepatol Res 2002; 22:30712. First citation in article
14. Assantachai P, Luengrojanakul P, Phimolsarnti R, Charoenlarp K, Ratanarapee S. Liver abscess in polycystic liver disease. J Med Assoc Thai 1995; 78:2106. First citation in article
15. Chou HC, Lee CZ, Ma LC, Fang CT, Chang SC, Wang JT. Isolation of a chromosomal region of Klebsiella pneumoniae associated with allantoin metabolism and liver infection. Infect Immun 2004; 72:378392. First citation in article
16. Fang CT, Chuang YP, Shun CT, Chang SC, Wang JT. A novel virulence gene in Klebsiella pneumoniae strains causing primary liver abscess and septic metastatic complications. J Exp Med 2004; 199:697705. First citation in article
17. Fitzgerald JR, Musser JM. Evolutionary genomics of pathogenic bacteria. Trends Microbiol 2001; 9:54753. First citation in article
18. Kim CC, Joyce EA, Chan K, Falkow S. Improved analytical methods for microarray-based genome-composition analysis. Genome Biol 2002; 3:RESEARCH0065. First citation in article
19. Abbott SL. Klebisella, Enterobacter, Citrobacter, Serratia, Plesiomonas, and other Enterobacteriaceae. In: Murray PR, Baron EJ, Jorgensen JH, Pfaller M, Yolken RH, eds. Manual of clinical microbiology. 8th ed. Washington, DC: American Society for Microbiology, 2003:684700. First citation in article
20. Fang CT, Chen HC, Chuang YP, Chang SC, Wang JT. Cloning of a cation efflux pump gene associated with chlorhexidine resistance in Klebsiella pneumoniae. Antimicrob Agents Chemother 2002; 46:20248. First citation in article
21. Chen JJ, Wu R, Yang PC, et al. Profiling expression patterns and isolating differentially expressed genes by cDNA microarray system with colorimetry detection. Genomics 1998; 51:31324. First citation in article
22. Ang S, Lee CZ, Peck K, et al. Acid-induced gene expression in Helicobacter pylori: study in genomic scale by microarray. Infect Immun 2001; 69:167986. First citation in article
23. Herrero M, de Lorenzo V, Timmis KN. Transposon vectors containing non-antibiotic resistance selection markers for cloning and stable chromosomal insertion of foreign genes in gram-negative bacteria. J Bacteriol 1990; 172:655767. First citation in article
24. Zimmermann L, Angerer A, Braun V. Mechanistically novel iron (III) transport system in Serratia marcescens. J Bacteriol 1989; 171:23843. First citation in article
25. Angerer A, Klupp B, Braun V. Iron transport systems of Serratia marcescens. J Bacteriol 1992; 174:137887. First citation in article
26. Okamoto M, Nakane A, Minagawa T. Host resistance to an intragastric infection with Listeria monocytogenes in mice depends on cellular immunity and intestinal bacterial flora. Infect Immun 1994; 62:30805. First citation in article
27. Reed LJ, Muench H. A simple method of estimating fifty percent endpoints. Am J Hyg 1938; 27:4937. First citation in article
28. Winnacker E-L. Genomic libraries. In: Weller MG, ed. From genes to clones: introduction to gene technology. Weinheim, Germany: VCH Verlagsgesellschaft mbH, 1987:3835. First citation in article
29. Chen YT, Chang HY, Lai YC, Pan CC, Tsai SF, Peng HL. Sequencing and analysis of the large virulence plasmid pLVPK of Klebsiella pneumoniae CG43. Gene 2004; 337:18998. First citation in article
30. Tomii K, Kanehisa M. A comparative analysis of ABC transporters in complete microbial genomes. Genome Res 1998; 8:104859. First citation in article
31. Angerer A, Gaisser S, Braun V. Nucleotide sequences of the sfuA, sfuB, and sfuC genes of Serratia marcescens suggest a periplasmic-binding-protein-dependent iron transport mechanism. J Bacteriol 1990; 172:5728. First citation in article
32. Adhikari P, Kirby SD, Nowalk AJ, Veraldi KL, Schryvers AB, Mietzner TA. Biochemical characterization of a Haemophilus influenzae periplasmic iron transport operon. J Biol Chem 1995; 270:251429. First citation in article
33. Gong S, Bearden SW, Geoffroy VA, Fetherston JD, Perry RD. Characterization of the Yersinia pestis Yfu ABC inorganic iron transport system. Infect Immun 2001; 69:282937. First citation in article
34. Chin N, Frey J, Chang CF, Chang YF. Identification of a locus involved in the utilization of iron by Actinobacillus pleuropneumoniae. FEMS Microbiol Lett 1996; 143:16. First citation in article
35. Adhikari P, Berish SA, Nowalk AJ, Veraldi KL, Morse SA, Mietzner TA. The fbpABC locus of Neisseria gonorrhoeae functions in the periplasm-to-cytosol transport of iron. J Bacteriol 1996; 178:21459. First citation in article
36. Mao Y, Varoglu M, Sherman DH. Molecular characterization and analysis of the biosynthetic gene cluster for the antitumor antibiotic mitomycin C from Streptomyces lavendulae NRRL 2564. Chem Biol 1999; 6:25163. First citation in article
37. Postma PW, Lengeler JW, Jacobson GR. Phosphoenolpyruvate : carbohydrate phosphotransferase systems of bacteria. Microbiol Rev 1993; 57:54394. First citation in article
38. Kok M, Bron G, Erni B, Mukhija S. Effect of enzyme I of the bacterial phosphoenolpyruvate : sugar phosphotransferase system (PTS) on virulence in a murine model. Microbiology 2003; 149:264552. First citation in article
39. Tan MW, Rahme LG, Sternberg JA, Tompkins RG, Ausubel FM. Pseudomonas aeruginosa killing of Caenorhabditis elegans used to identify P. aeruginosa virulence factors. Proc Natl Acad Sci USA 1999; 96:240813. First citation in article
40. Turner AK, Lovell MA, Hulme SD, Zhang-Barber L, Barrow PA. Identification of Salmonella typhimurium genes required for colonization of the chicken alimentary tract and for virulence in newly hatched chicks. Infect Immun 1998; 66:2099106. First citation in article
41. Escolar L, Perez-Martin J, de Lorenzo V. Opening the iron box: transcriptional metalloregulation by the Fur protein. J Bacteriol 1999; 181:62239. First citation in article
42. Hacker J, Kaper JB. Pathogenicity islands and the evolution of microbes. Annu Rev Microbiol 2000; 54:64179. First citation in article
43. Lai YC, Yang SL, Peng HL, Chang HY. Identification of genes present specifically in a virulent strain of Klebsiella pneumoniae. Infect Immun 2000; 68:714951. First citation in article
44. Griffiths E. The iron-uptake systems of pathogenic bacteria. In: Bullen JJ, Griffiths E, eds. Iron and infection. New York: John Wiley & Sons, 1987:69137. First citation in article
45. Lu L, Walker WA. Pathologic and physiologic interactions of bacteria with the gastrointestinal epithelium. Am J Clin Nutr 2001; 73:S112430. First citation in article