Klebsiella pneumoniae is responsible for a wide range of clinical symptoms. How this bacterium adapts itself to ever-changing host milieu is still a mystery. Recently, small non-coding RNAs (sRNAs) have received considerable attention for their functions in fine-tuning gene expression at a post-transcriptional level to promote bacterial adaptation. Here we demonstrate that Hfq, an RNA-binding protein, which facilitates interactions between sRNAs and their mRNA targets, is critical for K. pneumoniae virulence. A K. pneumoniae mutant lacking hfq (Δhfq) failed to disseminate into extra-intestinal organs and was attenuated on induction of a systemic infection in a mouse model. The absence of Hfq was associated with alteration in composition of envelope proteins, increased production of capsular polysaccharides, and decreased resistance to H2O2, heat shock, and UV irradiation. Microarray-based transcriptome analyses revealed that 897 genes involved in numerous cellular processes were deregulated in the Δhfq strain. Interestingly, Hfq appeared to govern expression of many genes indirectly by affecting sigma factor RpoS and RpoE, since 19.5% (175/897) and 17.3% (155/897) of Hfq-dependent genes belong to the RpoE- and RpoS-regulon, respectively. These results indicate that Hfq regulates global gene expression at multiple levels to modulate the physiological fitness and virulence potential of K. pneumoniae.
Citation: Chiang M-K, Lu M-C, Liu L-C, Lin C-T, Lai Y-C (2011) Impact of Hfq on Global Gene Expression and Virulence in Klebsiella pneumoniae. PLoS ONE 6(7): e22248. https://doi.org/10.1371/journal.pone.0022248
Editor: Jean-Pierre Gorvel, Universite de la Mediterranee, France
Received: May 24, 2011; Accepted: June 17, 2011; Published: July 14, 2011
Copyright: © 2011 Chiang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the National Science Council and Chung-Shan Medical University of Taiwan R.O.C. (NSC 98-2311-B-194-001-MY3 to Ming-Ko Chiang, NSC98-2320-B-040-013 and CSMU-TSMH-099-005 to Yi-Chyi Lai). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Since Klebsiella was first identified as a pathogen of pneumonia in 1882, the remarkable ability of K. pneumoniae to cause a wild range of human diseases, from urinary tract infections to life-threatening systemic infections , has attracted increasing attention to the pathogenesis of this bacterium. Not solely confined inside the human host, K. pneumoniae has a great capacity for adaptation to diverse environments, including the surface water, sewage, soil, intestinal tracts of mammals , and even the interior of plants . How K. pneumoniae responds to environmental changes and thus adapts itself to a specific niche becomes an interesting question. Nevertheless, our knowledge with regard to the regulatory mechanisms which this bacterium utilizes to ensure its survival upon different conditions is very limited.
An ever-increasing number and variety of small non-coding RNAs (sRNAs) are being identified to serve regulatory functions in bacteria. Numerous cellular processes, such as iron homeostasis , outer membrane proteins (OMPs) biogenesis , sugar metabolism , quorum sensing  and various stress responses , are subject to the post-transcriptional control exerted by sRNAs. At present, most characterized sRNAs regulate gene expression by basepairing with mRNAs. While some sRNAs are cis-encoded having the potential to basepair mRNAs with long stretches, the majority of regulatory sRNAs in Gram-negative bacteria are trans-encoded and share limited complementarity with their target mRNAs . The trans-acting sRNAs are functionally analogous to eukaryotic miRNAs that usually exert negative regulation by repress protein levels through translation inhibition, mRNA degradation, or both . In many cases, because of the limited complementarity, the trans-acting sRNAs mediated regulation requires the chaperone protein Hfq to facilitate RNA-RNA interactions.
Hfq assembles into homohexameric rings which are structurally similar to those formed by Sm and Sm-like proteins in eukaryotic cells . Besides enhancing the formation of sRNA-mRNA duplex, Hfq contributes to RNA regulation through interacting with RNA turnover enzymes, including RNase E, polynucleotide phosphorylase, and poly (A) polymerase . Hfq has a broad and diverse impact on bacterial physiology and virulence beyond its original role as a host factor required for replication of Qβ RNA bacteriophage . Defects including reduced growth, impaired resistance to various stresses, and altered virulence are detected in E. coli lacking hfq . It has also been shown that virulence of several pathogenic bacteria, including Brucella abortus , Francisella tularensis , Vibrio cholera , Listeria monocytogenes , Legionella pneumophila , Pseudomonas aeruginosa , Yersinia , , Salmonella Typhimurium , and uropathogenic E. coli , were significantly attenuated by hfq mutations.
Recently, by sequence analysis, we have identified an hfq homologue and sRNAs from K. pneumoniae genomes. The presence of these homologues in various strains of K. pneumoniae, including NTUH-K2044 (NC_012731), MGH78578 (NC_009648), and 342 (NC_011283) suggests that this pathogen also utilizes the post-transcriptional regulation mediated by Hfq to control various cellular processes. However, the role of Hfq-sRNA mediated regulation in K. pneumoniae is yet to be defined. In this study, we aimed to understand how Hfq contributed to the control of gene expression and the pathogenesis in K. pneumoniae. An hfq deletion mutant was generated in K. pneumoniae CG43S. The loss of hfq attenuated K. pneumoniae virulence in a mouse model, as well as altered physiological characteristics of K. pneumoniae, including production of capsular polysaccharides, stress tolerance, and homeostasis of envelope. The microarray data demonstrated that the expression of almost a fifth of K. pneumoniae genes was drastically deregulated. It also suggested that beside directly regulating individual genes expression at the post-transcriptional level, by affecting sigma factor RpoS and RpoE, Hfq positioned itself at the upper level of the gene regulatory hierarchy that control the physiological fitness and virulence potential of K. pneumoniae.
Deletion of hfq attenuated K. pneumoniae virulence
The hfq gene is located in clockwise orientation at bps 446148–446456 in the genome of K. pneumoniae strain NTHU-K2044 . As in E. coli, it is located in the miaA-hfq-hflX cluster with three promoter regions as indicated in Fig. 1A. The nucleotide sequence of hfq region in K. pneumoniae CG43S is identical to that in strain NTHU-K2044. Additionally, K. pneumoniae Hfq protein shares 85.4% sequence identity with its homologue in E. coli, and just like Enterobacteriaceae Hfq protein, it has the conserved Sm1 and Sm2 motifs (Fig. 1B). To explore the physiological role of Hfq, an hfq deletion mutant, named Δhfq, was constructed in the genetic background of K. pneumoniae CG43S. Furthermore, two trans-complemented strains, Δhfq-C1 and Δhfq-C2, which carried the hfq gene under the control of pBAD promoter and its native promoters, respectively, were generated as described in Materials and Methods (Fig. 1A).
(A) Genomic organization of hfq gene in K. pneumoniae. In the Δhfq strain, the coding region of hfq was deleted and replaced by a chloramphenicol cassette amplified from pACYC184. The coding region of hfq, represented as a light grey bar, was cloned to pBAD202, under the control of arabinose-inducible promoter, to give the complementation plasmid pYC343. The grey bar indicates the region of hfq together with its native promoter cloned to pACYC184 yielding the complementation plasmid pYC457. Three promoter regions of hfq are indicated as P1, P2, and P3, respectively. (B) Alignment of hfq sequences of the Gram-negative pathogens with their Hfq functionally identified. The highly conserved SM1 and SM2 motifs are indicated as black and grey lines, respectively.
To examine whether Hfq was critical for virulence of K. pneumoniae, competitive assays were performed in 8-week-old male BALB/c mice by orally inoculating them with the bacterial mixture containing equal amount of Δhfq and the parental strain CG43S . At 48 hour post-inoculation (hpi), colonization of the small intestine by Δhfq was comparable to that by CG43S. However, in the colon Δhfq was out-competed by CG43S with a CI value of 0.36 (Fig. 2A). While all the CG43S-infected mice had a bacterial burden approaching 102 to 104 CFU in the liver, Δhfq was undetectable at the same time point (Fig. 2A). Diminishment of Δhfq in the liver suggested that the disseminating ability of K. pneumoniae to extra-intestinal organs was abolished by the absence of hfq. To further investigate whether Hfq was involved in a systemic infection of K. pneumoniae, groups of mice were intraperitoneally inoculated with Δhfq or CG43S and their survival was monitored for two weeks. The intraperitoneal inoculation method allowed the infection to bypass the colonization step of K. pneumoniae in the small intestine. When mice were inoculated with 104 CFU of CG43S, all of them died by day 2 (filled squares, Fig. 2B). Even when the inoculums was decreased to 103 CFU, 60% of the mice still succumbed to the infection of CG43S (filled diamonds, Fig. 2B). On the other hand, when mice were inoculated with 104 CFU of Δhfq, 80% of the mice survived the experimental period (open circles, Fig. 2B). An involvement of hfq in a systemic K. pneumoniae infection was further supported by in vivo competition results. As shown in Fig. 2C, Δhfq was out-competed by CG43S at 6 h after intraperitoneal inoculation with the average CI values of 0.11 and 0.01 in the liver and spleen, respectively (Fig. 2C). The significant decline in the in vivo CI values might not be due to an inability to compete under nutrient-limiting conditions, as the growth of Δhfq approximated to that of CG43S with the in vitro CI value of 0.82 in M9 medium.
(A) Bacterial loads in small intestine, colon, and liver determined at 48 h after oral inoculation with suspension containing equal amount of K. pneumoniae CG43S (5×108 CFU) and Δhfq (5×108 CFU). Filled and open circles represent CG43S and Δhfq retrieved from five BALB/c mice, respectively. (B) Survival of K. pneumoniae-infected mice. Groups of five mice were inoculated by intraperitoneal injection with 2×103 CFU of CG43S (filled diamonds), or with 1×104 CFU of CG43S (filled squares), Δhfq (open circles), ΔrpoS (open diamonds), or ΔrpoE (open squares), and monitored for 14 days. (C) Groups of six mice were inoculated intraperitoneally with bacterial suspension containing equal amount of K. pneumoniae CG43S (1×103 CFU) and Δhfq (1×103 CFU). Bacterial loads of CG43S (filled circles) and Δhfq (open circles) in spleen and liver were determined at 6 h post-inoculation. Horizontal bars indicate geometric means. The limit of detection was approximately 10 CFU. Samples which yielded no colonies were plotted having the value as 10 CFU g−1 tissues. Competitive index is defined as Δhfq output/CG43Soutput ÷ Δhfq input/CG43Sinput. The indicated P values were determined using the Student’s t-test.
Hfq affected global gene expression
Previous reports have shown that Hfq contributes to the regulation of numerous cellular pathways in E. coli , , . To gain insight into the genes whose expression was regulated by Hfq in K. pneumoniae, DNA microarray experiments were performed to compare the transcriptome of Δhfq with that of CG43S. Probes were made from the RNA of K. pneumoniae which were grown to log-phase in LB medium at 37oC. Of 5,024 genes in the K. pneumoniae genome, 897 genes (approximately 18%) showed a >1.5log2 fold change (2.83-fold) in transcript abundance in Δhfq when compared to that in CG43S. Among the 897 Hfq-dependent genes, down-regulated genes (n = 610) were more than twice as up-regulated genes (n = 287), suggesting that Hfq-mediated regulation in K. pneumoniae was more frequently positive than negative. Based on the genome annotation of NTHU-K2044 strain (NC012731;), the 897 Hfq-dependent genes belong to more than 19 functional categories (Fig. 3 and Table S1 and S2). Several categories of genes were more notably affected by the absence of hfq. 34.3% (36/105) of the genes in the signal transduction category were deregulated in the Δhfq strain, of which, 14 were up-regulated and 22 were down-regulated. In addition, Hfq-dependency accounts for 35.7%, 30.6%, 26.36%, and 18% of genes belonging to the categories of energy production and conversion, transport and metabolism of lipid, carbohydrate, and amino acid, respectively (Fig. 3).
Microarray-based transcriptome analyses revealed 897 ORFs that are significantly deregulated in Δhfq as compared to that in CG43S. Based on the genome project of K. pneumoniae NTHU-K2044 (NC012731;), the 897 Hfq-dependent genes were grouped into different functional classes. The values represent the percentage of genes affected by Hfq in Δhfq versus CG43S within the respective class. Black bars: up-regulated genes; grey bars: down-regulated genes.
Hfq modulated the production of K2 capsular polysaccharides
Capsule has been established as a virulence determinant in K. pneumoniae. Interestingly, Δhfq failed to induce a systemic infection in mice but exhibited an enhanced hypermucoviscosity phenotype with 4.4-fold more K2-specific capsular polysaccharides (CPS) production compared to CG43S. The quantity of K2-CPS was restored to a normal level in Δhfq by the introduction of hfq-complementing plasmid (Δhfq-C2) (Fig. 4A). As revealed by the microarray analysis, abundance of the major transcript orf3-orf15 of K2 cps gene cluster (D21242.1; Fig. 4B) increased in the range of 2–4 folds in the Δhfq strain as compared to that in CG43S (Fig. 4C). Furthermore, K2 cps genes have been shown to be regulated by transcriptional activators, RcsA  and RmpA , whose transcripts both significantly increased in the absence of hfq. These results suggested that Hfq may modulate the production of K2 CPS by negatively regulate the expression of RmpA and RcsA at the post-transcriptional level.
(A) Enhancement of K2 capsular polysaccharides by the deletion of hfq. Capsular polysaccharides were extracted from overnight-cultured K. pneumoniae CG43S, Δhfq, and the complementation strain Δhfq-C2 by the method described previously . The amount of K2 CPS was reflected by the uronic acid content that was determined by the method described  from a glucuronic acid standard curve and expressed as micrograms per 109 CFU. The indicated P values were determined using the Student’s t-test. (B) Genomic organization of K2 cps gene cluster which is depicted from that in K. pneumoniae Chedid (NCBI accession no. D21242.1). (C) Fold changes in transcript abundances of K2 cps genes detected by microarray (black bars) and QPCR (grey bars) in Δhfq relative to that in CG43S are indicated. (D) Fold changes in transcript abundances of CPS-regulating genes, rcsA, rcsB, rcsC, and rmpA detected by microarray (black bars) and QPCR (grey bars) in Δhfq relative to that in CG43S are indicated.
A loss of hfq altered expression profiles of envelope proteins
Previous studies indicated that lacking hfq caused accumulation of outer membrane proteins and thus induced an envelope stress in UPEC, Salmonella, and Vibrio , , . To examine whether a loss of hfq affected the expression profiles of K. pneumoniae envelope proteins, two-dimensional gel analysis was utilized to compare extracytoplasmic proteins which were purified from cultures of Δhfq and CG43S grown aerobically in LB medium. As shown in Fig. 5A, a significant difference in the expression of extracytoplasmic proteins between Δhfq and CG43S was noted. Among 94 proteins analyzed, 32 proteins were up-regulated and 10 were down-regulated in the absence of Hfq. Analyses of outer membrane proteins in one-dimensional gels revealed an increase in the protein levels of OmpF (OmpK36) and a decrease of OmpC (OmpK35) in Δhfq as compared to those in CG43S and in the trans-complemented strain Δhfq-C2 (Fig. 5B). These changes were consistent with the microarray data that the deletion of hfq caused the levels of ompF mRNA to rise by 23.9 fold and the ompC mRNA to drop by 3.3 fold (Fig. 5C).
Two hundred micrograms of extracytoplasmic proteins isolated from overnight-cultured K. pneumoniae CG43S (A) or Δhfq (B) were electrophoresized with two-dimension gels. Images of silver-stained gels analyzed with ImageMasterTM 2D Platinum Version 5.0 are shown. L1-3: landmarks for comparison. Proteins whose expression levels are significantly higher in CG43S or Δhfq are indicated with a red cross. (C) Outer membrane proteins were extracted, fractionated by 12% of SDS-PAGE gel, and silver-stained. The portion of the gel with bands corresponding to the major porins, OmpC, OmpF, and OmpA, is presented. (D) Fold changes in transcript abundances of ompK17, ompF, ompC, ompR, and ybfM detected by microarray (black bars) and QPCR (grey bars) in Δhfq relative to that in CG43S are indicated.
Δhfq lost its stress tolerance to H2O2, heat shock, and UV radiation
To determine whether Hfq was required for K. pneumoniae to cope with stresses that it may encounter inside the host, the growth of Δhfq was characterized in various stressful conditions. As shown in Fig. 6, while Δhfq grew normally in LB medium (Fig. 6A), the growth of Δhfq lagged behind that of CG43S by approximately one hour and reached a lower density at saturation when the strains were grown in minimal medium (Fig. 6B), suggesting that the loss of Hfq caused only minor effects on K. pneumoniae response to nutrient deficiency. On the other hand, the loss of hfq rendered K. pneumoniae incapable of surviving after 10 minutes of treatment with H2O2 (Fig. 6C) or heat shock (Fig. 6D), and also increased the sensitivity of K. pneumoniae to ultraviolet (UV) irradiation (Fig. 6E). The reduced tolerance of Δhfq to the stresses was restored to the wild-type level by the introduction of the hfq-complementing plasmid driven by its native promoters (Δhfq-C2) (Fig. 6C, D, and E).
Growth of K. pneumoniae CG43S (filled diamonds), Δhfq (open squares), and Δhfq-C2 (filled circles) in (A) LB and (B) M9 media is determined by CFU calculation at indicative time points. Survival of K. pneumoniae CG43S, Δhfq, Δhfq-C2, ΔrpoS, and ΔrpoE, upon the treatment of 200 mM H2O2 for 10 minutes (C), 50°C shock for 10 minutes (D), or UV irradiation (1 J/cm2) (E) are determined by CFU calculation and presented as (CFU after treatment/CFU before treatment) ×100%. * P<0.05, determined with the Student’s t-test.
Overlap among Hfq-, RpoS-, and RpoE-regulons
Our microarray data showed that the transcript level of rpoS was down-regulated by 3.6-fold in the absence of hfq (Fig. 7A). The decrease of rpoS transcripts in Δhfq was restored to the wild-type level by the complementation of hfq under the control of pBAD promoter and the restoration was confirmed by Northern blotting analysis (Fig. 7B). In accordance with this, the protein level of RpoS was down-regulated by the absence of Hfq (Fig. 7C). To determine whether the virulence attenuation and phenotypic alterations observed in the Δhfq strain was attributed to the downregulation of rpoS by the loss of Hfq, an rpoS deletion mutant (ΔrpoS) was generated in K. pneumoniae CG43S. Unlike Δhfq, whose virulence to mice was significantly attenuated, ΔrpoS behaved much like CG43S as it caused 80% mortality of mice within one week (open diamonds, Fig. 2B). ΔrpoS also displayed wild-type-level tolerance of K. pneumoniae in response to heat shock and UV irradiation (Fig. 6D and 6E). On the other hand, the loss of rpoS did abolish the ability of K. pneumoniae to conquer H2O2 stress (Fig. 6C). The results suggested that the downregulation of rpoS in the Δhfq strain contributed partially to the defects on stress tolerance resulted from the loss of hfq, but could not by itself attenuate the virulence of K. pneumoniae. Meanwhile, the expression of RpoE had also been examined. Although the transcript level of rpoE in the Δhfq strain was found similar to that in CG43S in the microarray analysis, Western blotting analysis revealed that the absence of hfq resulted in decreased protein level of RpoE at early- and mid-log phase (Fig. 7C). An rpoE deletion mutant (ΔrpoE) was generated. Unlike ΔrpoS, ΔrpoE was totally avirulent when given intraperitoneally with the same inoculums that caused 80% mortality in the ΔrpoS-infected group (open diamonds, Fig. 3B). ΔrpoE was as sensitive as Δhfq in its responses to heat shock and UV irradiation (Fig. 6D and 6E), whereas it exhibited a wild-type-level resistance to H2O2 (Fig. 6C). These results suggested that the virulence attenuation as well as the loss of tolerance to heat shock and UV irradiation in Δhfq may result from the decrease of RpoE protein by the lack of hfq.
(A) Fold changes in transcript abundances of rpoH, rpoD, rpoS, rpoE, and rpoN detected by microarray (black bars) in Δhfq relative to that in CG43S are indicated. (B) Transcripts of rpoS extracted from CG43S, Δhfq, and Δhfq-C1 were detected by Northern blotting with an rpoS-specific biotin-labeled riboprobe. (C) RpoS proteins isolated from the LB-grown stationary-phase cultures of CG43S, Δhfq, and Δhfq-C2 with or without the induction of 0.02% arabinose were detected by Western blotting with rabbit anti-RpoS antibody. The levels of OmpA were detected with rabbit anti-OmpA antibody as a loading control. (D) RpoE proteins isolated from the LB-grown cultures of CG43S, Δhfq, and Δhfq-C1 at indicated time points were detected by Western blotting with rabbit anti-RpoE antibody. (E) Transcripts of MicF extracted from the LB-grown cultures of CG43S and Δhfq at indicated time points were detected by Northern blotting with a MicF-specific biotin-labeled riboprobe.
To further determine whether the set of Hfq-dependent genes overlapped with those belonging to the RpoS- and/or RpoE-regulon, genes whose transcripts were deregulated significantly upon the overexpression of RpoS or RpoE were identified. As revealed by microarray analyses, among the 610 down-regulated genes identified in Δhfq, the transcript abundance of 90 and 109 genes were significantly affected by overexpression of RpoE and RpoS (Table S1), while 85 and 46 genes out of the 287 up-regulated genes were under control of RpoE and RpoS, respectively (Table S2). Overall, 19.5% (175/897) and 17.3% (155/897) of Hfq-dependent genes belong to the RpoE- and RpoS-regulon, respectively, and 4.8% (43/897) of Hfq-dependent genes were presented in both regulons. These results suggested that the Hfq-mediated downregulation of RpoS and RpoE was responsible for the Hfq-dependency for 32% of the 897 genes identified in the Δhfq strain.
Hfq-dependent changes in transcript abundances of sRNAs in K. pneumoniae
As an RNA chaperone which is critical in stabilizing sRNA-mRNA complexes, Hfq is required for the trans-acting sRNAs to regulate the expression of their target mRNAs . Most of the Hfq-dependent sRNAs are less stable and no longer regulate their target mRNAs in E. coli lacking hfq . Therefore, it was likely that the abundance of a number of Hfq-dependent genes were affected by the change of sRNA abundance in the absence of Hfq. To test this possibility, 33 K. pneumoniae sRNA genes which have homologues in E. coli K-12 were identified from the genome of K. pneumoniae NTHU-K2044 (Table 2) and their transcripts abundances was examined by microarray analyses. As shown in Table 2, 25 K. pneumoniae sRNAs were downregulated by the absence of Hfq, of which, 11 sRNAs (GcvB, MicF, RyeB, RyeE, RygB, RyhA, RyiA, SraB, MicM, SroC, and SroG) exhibited >3-fold lower transcript abundance in Δhfq than in CG43S. Among them, the diminishment in transcript abundance of MicF was confirmed by Northern blotting (Fig. 7D). In accordance with the finding in E. coli that MicF inhibited the translation of ompF mRNA , the decline of MicF in the Δhfq strain derepressed the negative control of ompF transcript and consequently enhanced the expression of K. pneumoniae OmpF protein (Fig. 7A and 7B). The identification of sRNAs that required Hfq for maintaining their abundance in K. pneumoniae suggested that a subset of K. pneumoniae genes whose expression changed in the absence of Hfq were controlled by the Hfq-dependent sRNAs.
Small RNAs, along with Hfq, are emerging as regulators that enable bacteria to modulate gene expression in response to changing environments. In this study, we demonstrate that Hfq modulates the expressions of a wide range of genes and thus regulates the physiology and virulence of K. pneumoniae. In the absence of Hfq, K. pneumoniae drastically lost its ability to disseminate into extra-intestinal organs, and was unable to induce a systemic infection. Comparison of the Δhfq strain with its parental strain revealed that several physiological characteristics were altered due to the Hfq deficiency, including production of capsular polysaccharides, envelope protein expression profiles, and tolerance to H2O2, heat shock, and UV irradiation. Genome-wide transcriptome analyses showed that 897 genes involved in numerous cellular processes had >2.83-fold change in their transcript abundance in Δhfq. The apparent influence of Hfq on almost a fifth of K. pneumoniae genes indicated that Hfq acted as a global regulator. Since RpoS and RpoE were found to be down-regulated in Δhfq, Hfq may also govern gene expression indirectly in K. pneumoniae through its positive effect on the availability of sigma factors. The idea was supported by the microarray data that 32% of Hfq-dependent genes belong to the RpoE- and/or RpoS-regulon. Moreover, reduction on the transcript levels of 25 K. pneumoniae sRNAs (out of 33 known sRNAs) in Δhfq indicated that a large proportion of genes that were affected in the absence of Hfq were mediated through the action of sRNAs.
The majority of the known sRNA-mediated regulation is negative , . As shown in Fig. 8A, with the aid of Hfq, sRNAs bind to the 5’-UTR of a single or multiple target mRNAs, occlude the ribosome-binding site, prevent ribosome association, and thus inhibit translation initiation. In many cases, the sRNA-mRNA duplex is then subject to degradation . Hfq recruits the RNA degradosome, consisting of RNase E, PNPase, RhlB helicase, and enolase, and RNA degradation is triggered by the RNase E cleavage at sites distal from the pairing region . Based on this model, we speculate that the absence of Hfq relieves the sRNAs-mediated negative regulation and elevates the transcript abundances of genes that are targeted by certain sRNAs. In line with this assumption, a significant number of 287 up-regulated genes identified in Δhfq (Table S2) might be subject to this kind of Hfq-sRNA-mediated negative regulation in K. pneumoniae. On the other hand, unlike eukaryotic miRNAs, there are three bacterial sRNAs, DsrA, RprA, and RyhA, have been identified to activate the expression of rpoS through anti-antisense mechanism whereby basepairing of the sRNAs disrupts an inhibitory secondary structure formed by the rpoS mRNA leader sequence , . As shown in Fig. 8B, derepression of rpoS translation increases the availability of RpoS (σS), and activates the expression of genes belonging to the rpoS regulon. Therefore, due to a loss of positive regulation by sRNAs, the absence of Hfq caused the reduced translation efficiency of rpoS (Fig. 7A–C). This result is consistent with some of the changes in the gene expression profile observed in Δhfq, as shown by microarray data that 17.3% of Hfq-dependent genes fall in the RpoS regulon. In addition to the influence on translational efficiency of rpoS, the decreased transcript level of rpoS suggested that Hfq-dependent regulation of rpoS in K. pneumoniae was more complicated than that in E. coli and Salmonella , . Although the mechanism requires further studies to clarify, the existence of Hfq-dependent transcriptional activators of rpoS or the possibility that a loss of hfq affects the stability of rpoS transcript may explain this result.
(A) sRNA-dependent negative regulation. Hfq-dependent sRNAs bind their target mRNAs with partial complementarity that occlude the ribosome binding sites, prevents ribosome association and thus represses translation. In many cases, translation inhibition can be coupled to RNA degradation. Hfq recruits RNA degradosome and RNA degradation will be achieved at distal sites from the paring region. (B) Positive regulation of rpoS translation. Hfq-dependent sRNAs, such as DsrA, RprA, and RyhA, relieve an inhibitory secondary structure formed by the rpoS mRNA leader sequence. Derepression of rpoS translation increases the availability of RpoS (σS) and thus activates the expression of genes belonging to the rpoS regulon. (C) Regulation through protein-protein interactions. Hfq may contact with proteins to affect cellular processes. Hfq can recruit RNA degradosome, consisting of RNaseE, RhlB helicase, enolase, and PNPase, to stimulate the RNA degradation of sRNA-mRNA duplex. Hfq may alter mRNA stability by promoting PAP I-mediated polyadenylation. In the presence of S1 protein, Hfq may interact with RNA polymerase to modulate transcriptional activities. A minor fraction of Hfq, which was found to associate with the nucleoid, may bind curved DNA, affect negative supercoiling, and then coordinate the transcription and translation activities.
Hfq acts as a global regulator in a variety of pathogens , however, the spectrum and severity of mutant phenotypes observed upon the deletion of hfq varied among the different pathogens so far analyzed. In S. Typhimurium and E. coli , , the involvement of Hfq in bacterial virulence was indicated by its requirement for translation derepression of RpoS. In K. pneumoniae, although the expression of rpoS (Fig. 7A and 7B) and the RpoS-mediated oxidative stress response (Fig. 6C) was affected by the absence of Hfq, Δhfq was far more attenuated in causing a systemic infection than was ΔrpoS (Fig. 2B). The dispensability of rpoS for K. pneumoniae virulence suggests that the Hfq-dependent control of virulence genes in K. pneumoniae is largely mediated through an RpoS-independent manner, which is in accordance with previous studies that the virulence defect of hfq mutants in V. cholera and B. abortus is independent of the effects on RpoS expression , , .
In addition to RpoS, the expression of envelope stress sigma factor, RpoE, was also affected by the absence of Hfq. While Δhfq lost its capacity to tolerate oxidative stress as ΔrpoS (Fig. 6C), the increased sensitivity to heat shock and UV irradiation in the absence of Hfq more closely resembled the phenotypes observed in ΔrpoE (Fig. 6D and 6E). The Hfq-associated regulation on RpoE expression varied among different bacteria. Previous studies indicated that Hfq negatively regulated the expression of RpoE in E. coli , S. Typhimurium , and V. cholerae . However, in Rhodobacter sphaeroides  and Sinorhizobium meliloti , Hfq exerted a positive effect on RpoE. Interestingly, in K. pneumoniae, despite the transcript abundance of rpoE was slightly increased in the absence of Hfq (Fig. 7A), the protein level of RpoE was decreased in the Δhfq strain and was elevated by the introduction of an Hfq-complementation plasmid (Fig. 7C). The result indicated that the expression of RpoE might be regulated by Hfq at multiple levels. The deletion of hfq in K. pneumoniae caused modification of the expression profiles of OMPs. As shown in the proteomic result, 10 extracellular proteins were down-regulated and 32 proteins were up-regulated (Fig. 5A and 5B). The disturbance of envelope homeostasis in Δhfq could induce the envelope stress and release RpoE from the anti-sigma factor RseA . Upon the overexpression of RpoE, as revealed by microarray data, the transcriptional level of rpoE-rseA-rseB operon and RybB were respectively elevated to >9-fold and 300-fold over the vector control, suggesting that the increased anti-sigma factor RseA and RseB as well as the RpoE-inhibiting sRNA, RybB, might exert a feedback control by down-regulating the elevated level of RpoE protein at either post-transcriptional or post-translational level. The decrease of RpoE protein in Δhfq might be an outcome of the net effect from the induction of RpoE by the envelope stress and the subsequent feedback control of RpoE by its negative regulators, RseA, RseB, and RybB. However, as the rybB transcript abundance was slightly reduced in Δhfq with a fold change of -1.39 (Table 2), there might be some other Hfq-dependent factors involved in the negative regulation of RpoE in K. pneumoniae. Although further studies are needed to unveil the underlying mechanism, the decreased RpoE level in Δhfq can partially explain the overlap of stress responsiveness and virulence attenuation observed among Δhfq and ΔrpoE (Fig. 2B).
In our oral infection mouse model , K. pneumoniae lacking hfq retained its intestine-colonization ability but failed to disseminate into extra-intestinal tissues that attenuated the induction of a systemic infection (Fig. 1). Why did not the loss of Hfq abrogate the ability of K. pneumoniae to establish intestinal colonization? Transcripts abundances of genes belonging to the type III fimbriae-encoding operon were significantly up-regulated in the absence of Hfq (Table S2). As demonstrated in our previous study that the type III fimbriae was a major factor for K. pneumoniae to establish its intestinal colonization , the increased production of type III fimbriae in Δhfq may compensate for the possible defects on intestinal colonization caused by the deletion of Hfq and can therefore render Δhfq with the wild-type-level colonizing ability. As the overproduction of RpoS caused 3-10 fold decrease in the transcripts abundances of type III fimbriae-encoding genes, the Hfq-dependent control on type III fimbriae expression might be attributed to the derepression of rpoS in Δhfq (Table S2). On the other hand, since capsule had been established as a virulence determinant for K. pneumoniae pathogenesis, it was interesting that Δhfq produced more K2 capsular polysaccharides, but failed to develop a systemic infection. It seems that the enhancement of capsular polysaccharides may not be advantageous in all cases. As our microarray data revealed that almost a fifth of genes (123/516; 23.8%) of the carbohydrate transport and metabolism category were significantly downregulated in Δhfq (Fig. 3), the overproduction of capsular polysaccharides in the absence of Hfq may unbalance the metabolic flow of carbon sources and therefore affects the bacterial physiology under a nutrient-deprivation surrounding, such as inside the host during infections. Since K. pneumoniae can survive upon different circumstances, according to the result, we speculate that Hfq may modulate the expression of capsular polysaccharides to set a balance between the carbon usage and the production of capsule, which benefits this bacterium with an adaptive potential to switch among a panel of growth programs.
Most sRNAs have negative impacts on gene expression in bacteria. Genes whose expression is subject to the control of sRNAs are believed to be up-regulated by the loss of hfq. Interestingly, in our study, the number of genes that were down-regulated (n = 610) was more than twice the number of up-regulated genes (n = 287) in Δhfq K. pneumoniae. The positive effects of Hfq on gene expression in K. pneumoniae may be mediated through upregulation of sigma factors, or repression of sRNA-controlled transcriptional repressors, or through an sRNA-independent manner. It has been shown that Hfq per se can bind with mRNAs, tRNAs , and various proteins. In addition to direct interacting with Qβ phage RNA , Hfq binds its own mRNA that inhibits the formation of translational initiation complex to autoregulate its own expression . A variety of proteins, including ribosomal proteins, RNases, helicases, Rho-factor, RNA polymerase, protein H-NS, polynucleotide phosphorylase (PNPase), and poly(A)polymerase (PAP I), exhibited a direct interaction with Hfq , , , , . We speculate that Hfq can contact with proteins to influence K. pneumoniae gene expression. As shown in Fig. 8C, Hfq can recruit the RNA degradosome to stimulate RNA degradation. Hfq can alter mRNA stability by promoting PAP I-mediated polyadenylation , , . In the presence of S1 protein, Hfq may interact with RNA polymerase to modulate transcriptional activities . Although the majority of Hfq is located in the cytoplasm, it has been found that a minor fraction of Hfq associates with the nucleoid that preferentially binds curved DNA. This observation suggests that with the dual binding capacity to both DNA and RNA, Hfq may have a role in functional coordination between transcription and translation . Taken together, the impact of Hfq on global gene expression that controls the physiological fitness and virulence potential of K. pneumoniae shown by this work emphasizes the role of Hfq as a pivotal coordinator for integrating a diversity of regulatory circuits and also the potential that Hfq may serve as a scaffold molecule for the design of novel antimicrobial drugs.
Materials and Methods
All animal experiments were performed in strict accordance with the recommendation in the Guide for the Care and Use of Laboratory Animals of the National Laboratory Animal Center (Taiwan), and the protocol was approved by the Animal Experimental Center of Chung-Shan Medical University (Permit number: 694). All surgery was performed under anesthesia, and all efforts were made to minimize suffering.
Strains, plasmids, and primers
Bacterial strains, plasmids, and primers used in this study are listed in Table 1. E. coli and K. pneumoniae CG43 and its derivatives were propagated in Luria-Bertani (LB) broth. Gene-specific deletion mutants were generated with an allelic exchange technique as described previously . In general, approximately 1,200-bp DNA fragments flanking the coding region of genes to be deleted (hfq or rpoE) were amplified with specific primer sets, p239/p240 and p241/p242 for hfq deletion; p251/p252 and p253/p254 for rpoE deletion and cloned into the suicide vector, pKAS46 . To facilitate the positive selection of transconjugants, a chloramphenicol-resistant or a tetracycline-resistant cassette which was amplified from pACYC184, cloned into the inserts on pKAS46 constructs to replace the deletion region, and the resulting plasmids for homologous recombination of hfq and rpoE were pYC324 (Cmr) and pYC445 (Tcr), respectively. After the occurrence of double crossover, the chloramphenicol-resistant or tetracycline-resistant colonies were selected, and the deletions of hfq or rpoE were verified by PCR and Southern blot analysis. The rpoS deletion mutant (ΔrpoS) which was generated as described  was kindly provided by Dr. Lin, G.T. (China Medical University, Taiwan).
The entire 309 bp hfq coding sequence and 745 bp of upstream sequence were amplified from CG43 chromosomal DNA with primer set p403/p285 and cloned into pACYC184 to generate pYC457. The coding sequences of hfq, rpoS, and rpoE were amplified with primer sets, p284/p285, p903/p904, and p909/p910, respectively, and cloned into pBAD202 (Invitrogen) to generate pYC343, pYC351, and pYC413. The protein of Hfq, RpoS, or RpoE with a C-terminally fused His6-tag was induced by addition of 0.02% arabinose and verified using Western blot analysis with anti-His antibodies (Santa-Cruz). All these constructs were introduced into K. pneumoniae CG43S and its mutants through electroporation.
Eight-week-old male BALB/c mice (National Laboratory Animal Center, Taiwan) were injected intraperitoneally with 100 µl of bacterial suspension containing 104 or 2×103 CFU of mid-log K. pneumoniae CG43S or mutants. The survival rate of the infected mice was monitored daily for 2 weeks. Mortality rate and mean number of days to death (MDD) were determined by Kaplan-Meier analysis using Prism4 for Windows (GraphPad); P values of <0.05 were considered statistically significant. To further examine virulence attenuation of the Δhfq strain, a competitive assay was performed as described previously . One hundred microliters of bacterial suspension containing equal amounts of K. pneumoniae CG43S and Δhfq was used to infect 8-week-old male BALB/c mice through an oral or an intraperitoneal route with inoculums of 1×109 or 2×103 CFU, respectively. At indicative time points after inoculation, viable counts of CG43S and Δhfq in a particular mouse tissue were determined using M9 agar supplemented with or without chloramphenicol (10 µg/ml). Competitive index (CI) values were calculated as described .
K. pneumoniae microarray was customized using Agilent eArray 5.0 program according to the manufacturer’s recommendations. The customized microarray (4×44K) contained spots in eight duplicates with 5,076 gene-specific oligonucleotides (45–60 mers in length) representing 5,024 genes in K. pneumoniae NTHU-K2044 genome (NCBI accession no. NC_012731), 33 sRNA-coding candidates in K. pneumoniae (as predicted by homologues of E. coli K12), and 19 K2 cps genes (NCBI accession no. D21242.1).
RNA isolation, labeling, hybridization, and scanning
Total RNA was isolated from 20 ml of log K. pneumoniae culture by using Tri-reagent (Molecular Research Center), and purified by QIAGEN RNeasy cleanup kit. The total RNA yield was quantified by nanodrop UV spectroscopy (Ocean Optics) and the quantity was verified by gel electrophoresis and analyzed on a RNA 6000 Nano LabChip (Agilent Technologies) using a 2100 bioanalyzer (Agilent Technologies). cDNA was synthesized from 1 µg of enriched mRNA with Cyscribe 1st-strand cDNA labeling kit, (GE Healthcare) and labeled with Cy3 or Cy5 (CyDye, PerkinElmer). The fluorescently labeled cDNA was purified with QIAGEN RNeasy cleanup kit and fragmented in fragmentation buffer (Agilent). The correspondingly labeled cDNA was mixed in GEx Hybridization Buffer HI-RPM (Agilent). Hybridization was performed in an Agilent microarray Hybridization Chamber for 17 h at 60°C. After hybridization, the slides were washed in Gene Expression Wash Buffer (Agilent) and dried by nitrogen gun blowing. Microarrays were scanned using an Agilent microarray scanner at 535 nm for Cy3 and 625 nm for Cy5. Feature extraction 9.5.3 and image analysis software (Agilent Technologies) was used to locate and delineate every spot in the array, to integrate each spot’s intensity, and to normalize data using the rank-consistency-filtering Lowess method. The data points which had flag value of non-zero or a signal-to-noise ratio smaller than 2.6 were masked. The remaining data were log2 transformed and averaged for each gene. All microarray data reported in the study is described in accordance with MIAME guidelines and has been deposited in NCBI's Gene Expression Omnibus (GSE 29448). For selection of genes which were significantly regulated by Hfq, RpoE, or RpoS, a 1.5×log2 fold change (approximately 2.83-fold) and pValueLogRatio <0.05 were used as thresholds.
Primer design and Real-time PCR
To prepare a cDNA pool from each RNA sample, total RNA (5 µg) was reverse transcribed using MMLV reverse transcriptase (Promega). Each cDNA pool was stored at −20°C until further real-time PCR analysis. Specific oligonucleotide primer pairs were selected from Roche Universal ProbeLibrary for real-time PCR assays. The specificity of each primer pair was validated by performing a RT-PCR reaction using pooled K. pneumoniae CG43S cDNA template, and the size of the PCR product was checked by a DNA 1000 chip (Agilent Technologies) run on Bioanalyzer 2100 (Agilent Technologies). Primer pairs of generating predicted product size and no other side-product were chosen to conduct the following real-time RT-PCR reaction.
Real-time PCR reactions were performed on the Roche LightCycler Instrument 1.5 using LightCycler® FastStart DNA MasterPLUS SYBR Green I kit (Roche). Briefly, 10 µl reactions contained 2 µl Master Mix, 2 µl of 3.75 µM or 2 µl of 2.5 µM with 5% DMSO, forward primer and reversed primer, and 6 µl cDNA sample. Each sample was run in triplicate. The RT-PCR program were 95°C for 10 min, 50 cycles of 95°C for 10 sec, 60°C for 15 sec, and 72°C for 10 sec. At the end of the program a melt curve analysis was done. At the end of each RT-PCR run, the data were automatically analyzed by the system and an amplification plot was generated for each cDNA sample. From each of these plots, the LightCycler3 Data analysis software automatically calculates CP value (crossing point, the turning point corresponds to the first maximum of the second derivative curve), which imply as the beginning of exponential amplification. The fold expression or repression of the target gene relative to internal control gene rfaH in each sample was calculated by the formula: 2−△△CP where △Cp = Cp target gene – Cp internal control and △△Cp = △Cp test sample - △Cp control sample.
Two-dimensional (2D) proteomic analysis
Extracytoplasmic proteins were extracted as described  with modifications. Briefly, 40 ml of K. pneumoniae CG43S or Δhfq cultures were harvested at late-log phase by centrifugation at 13,000 rpm for 10 min at 4°C. The bacterial pellets were lysed with 4 ml of B-PER reagent (Pierce) containing DNaseI (0.5 µg/ml) and lysozyme (15 mg/ml). After 10-min incubation at room temperature, unbroken cells and cellular debris were removed by centrifugation at 8,000 rpm for 10 min at 4°C. Supernatant was collected and was further centrifuged at 35,000 rpm for 60 min at 4°C. The pellet was resuspended in 100 µl of 3% (w/v) sodium lauryl sarcosinate (Sigma) and incubated at room temperature for 30 min. The concentration of resulting suspensions was quantified with the BCA protein assay kit (Thermal). Two hundred micrograms of extracytoplasmic proteins were dissolved in solution (8M urea; 2M thiourea; 4% CHAPS and 80 mM DTT) and subjected to isoelectric focusing (IEF) electrophoresis with pH 4–7 carrier ampholyte for 8000 Vh. After equilibration for 15 min, the gels were transferred to the second dimension electrophoresis using 12% polyacrylamide gel and stained with the Silver Stain Plus kit (BioRad). Gel image was analyzed with ImageMasterTM 2D Platinum version 5.0 (Amersham Biosciences).
Extraction and quantification of capsular polysaccharides (CPS)
CPS was extracted by the method described previously . Five hundred microliters of bacterial culture was mixed with 100 µl of 1% Zwittergent 3–14 detergent (Sigma-Aldrich) in 100 mM citric acid (pH 2.0), and then the mixture was incubated at 50°C for 20 min. After centrifugation, 250 µl of the supernatant was transferred to a new tube, and CPS was precipitated with 1 ml of absolute ethanol. The pellet was then dissolved in 200 µl of distilled water, and a 1,200-µl volume of 12.5 mM borax (Sigma-Aldrich) in H2SO4 was added. The mixture was vigorously vortexed, boiled for 5 min, and cooled, and then 20 µl of 0.15% 3-hydroxydiphenol (Sigma-Aldrich) was added and the absorbance at 520 nm was measured. The uronic acid content was determined by the method described  from a standard curve of glucuronic acid (Sigma-Aldrich) and expressed as micrograms per 109 CFU.
Stress resistance assays
The K. pneumoniae strains to be tested were grown in LB medium at 37°C for 16 hours (8×108 CFU). The stationary-phase-cultures containing 107 CFU of bacteria were inoculated into LB medium or M9–0.2% glucose medium (Difco). Bacterial growth at 37°C was monitored every 1 h by measuring CFU per ml of culture with a plate counting method. For heat, H2O2, and UV treatment, approximately 5×108 CFU of bacterial cells were transferred to 50°C for 10 min or to LB medium containing 200 mM H2O2 for 10 min, or exposed to UV irradiation (1 J/cm2), respectively. Dilutions of bacteria were spread on LB agar to determine the number of viable bacteria. Survival rates upon different stresses were expressed as the formula: (CFU after treatment/CFU before treatment) ×100%.
For Northern detection of the rpoS or the MicF transcript, 40 µg of total RNA isolated from K. pneumoniae CG43, Δhfq, or Δhfq-C1 was glyoxal denatured, separated on a 1% agarose gel, and then transferred onto a BrightStar Plus nylon membrane (Ambion). After UV cross-linking, the membrane was blotted with ULTRAhyb hybridization buffer (Ambion) overnight at 42°C against a rpoS- or MicF-specific biotin-labeled riboprobe, which was prepared using a BrightStar psoralen-biotin kit (Ambion). After a stringent wash, signals were detected with a BrightStar BioDetect kit (Ambion).
Thirty micrograms of K. pneumoniae total proteins which were extracted from CG43S, Δhfq, Δhfq-C1, or Δhfq-C2 with or without the induction of 0.02% arabinose were resolved by 12% of SDS-polyacrylamide gel and transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore). After blocking with 2% (w/v) skim milk at room temperature for 1 h and washes with 1×PBST, the membrane was hybridized with rabbit anti-RpoE antibody (1∶1000 dilution), rabbit anti-RpoS antibody (1∶1000 dilution), or rabbit anti-OmpA antibody (1∶1000 dilution) at 4°C for 16 hours and 5000-fold diluted HRP-conjugated anti-rabbit IgG secondary antibody was subsequently used. After stringent washes with 1×PBST, signals were detected with ECL reagent (Thermal).
K. pneumoniae genes down-regulated by the absence of hfq.
We thank Ting-Quan Yan for advice and assistance with microarray work, and Jia-Zhen Weng and Pei-Shan Wu for help with Northern and Western blotting experiments. Freeze ultracentrifuge was performed in the Instrument Center of Chung Shan Medical University, which is supported by National Science Council, Ministry of Education and Chung Shan Medical University.
Conceived and designed the experiments: Y-CL M-CL M-KC. Performed the experiments: L-CL. Analyzed the data: Y-CL. Contributed reagents/materials/analysis tools: C-TL. Wrote the paper: Y-CL M-KC.
- 1. Podschun R, Ullmann U (1998) Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 11: 589–603.
- 2. Fouts DE, Tyler HL, DeBoy RT, Daugherty S, Ren Q, et al. (2008) Complete genome sequence of the N2-fixing broad host range endophyte Klebsiella pneumoniae 342 and virulence predictions verified in mice. PLoS Genet 4: e1000141.
- 3. Masse E, Salvail H, Desnoyers G, Arguin M (2007) Small RNAs controlling iron metabolism. Curr Opin Microbiol 10: 140–145.
- 4. Guillier M, Gottesman S, Storz G (2006) Modulating the outer membrane with small RNAs. Genes Dev 20: 2338–2348.
- 5. Vanderpool CK (2007) Physiological consequences of small RNA-mediated regulation of glucose-phosphate stress. Curr Opin Microbiol 10: 146–151.
- 6. Lenz DH, Mok KC, Lilley BN, Kulkarni RV, Wingreen NS, et al. (2004) The small RNA chaperone Hfq and multiple small RNAs control quorum sensing in Vibrio harveyi and Vibrio cholerae. Cell 118: 69–82.
- 7. Repoila F, Majdalani N, Gottesman S (2003) Small non-coding RNAs, co-ordinators of adaptation processes in Escherichia coli: the RpoS paradigm. Mol Microbiol 48: 855–861.
- 8. Aiba H (2007) Mechanism of RNA silencing by Hfq-binding small RNAs. Curr Opin Microbiol 10: 134–139.
- 9. Waters LS, Storz G (2009) Regulatory RNAs in bacteria. Cell 136: 615–628.
- 10. Arluison V, Derreumaux P, Allemand F, Folichon M, Hajnsdorf E, et al. (2002) Structural Modelling of the Sm-like Protein Hfq from Escherichia coli. J Mol Biol 320: 705–712.
- 11. Morita T, Aiba H (2011) RNase E action at a distance: degradation of target mRNAs mediated by an Hfq-binding small RNA in bacteria. Genes Dev 25: 294–298.
- 12. Franze de Fernandez MT, Eoyang L, August JT (1968) Factor fraction required for the synthesis of bacteriophage Qbeta-RNA. Nature 219: 588–590.
- 13. Tsui HC, Leung HC, Winkler ME (1994) Characterization of broadly pleiotropic phenotypes caused by an hfq insertion mutation in Escherichia coli K-12. Mol Microbiol 13: 35–49.
- 14. Robertson GT, Roop RM Jr (1999) The Brucella abortus host factor I (HF-I) protein contributes to stress resistance during stationary phase and is a major determinant of virulence in mice. Mol Microbiol 34: 690–700.
- 15. Meibom KL, Forslund AL, Kuoppa K, Alkhuder K, Dubail I, et al. (2009) Hfq, a novel pleiotropic regulator of virulence-associated genes in Francisella tularensis. Infect Immun 77: 1866–1880.
- 16. Ding Y, Davis BM, Waldor MK (2004) Hfq is essential for Vibrio cholerae virulence and downregulates sigma expression. Mol Microbiol 53: 345–354.
- 17. Christiansen JK, Larsen MH, Ingmer H, Sogaard-Andersen L, Kallipolitis BH (2004) The RNA-binding protein Hfq of Listeria monocytogenes: role in stress tolerance and virulence. J Bacteriol 186: 3355–3362.
- 18. McNealy TL, Forsbach-Birk V, Shi C, Marre R (2005) The Hfq homolog in Legionella pneumophila demonstrates regulation by LetA and RpoS and interacts with the global regulator CsrA. J Bacteriol 187: 1527–1532.
- 19. Sonnleitner E, Hagens S, Rosenau F, Wilhelm S, Habel A, et al. (2003) Reduced virulence of a hfq mutant of Pseudomonas aeruginosa O1. Microb Pathog 35: 217–228.
- 20. Geng J, Song Y, Yang L, Feng Y, Qiu Y, et al. (2009) Involvement of the post-transcriptional regulator Hfq in Yersinia pestis virulence. PLoS One 4: e6213.
- 21. Schiano CA, Bellows LE, Lathem WW (2010) The small RNA chaperone Hfq is required for the virulence of Yersinia pseudotuberculosis. Infect Immun 78: 2034–2044.
- 22. Sittka A, Pfeiffer V, Tedin K, Vogel J (2007) The RNA chaperone Hfq is essential for the virulence of Salmonella typhimurium. Mol Microbiol 63: 193–217.
- 23. Kulesus RR, Diaz-Perez K, Slechta ES, Eto DS, Mulvey MA (2008) Impact of the RNA chaperone Hfq on the fitness and virulence potential of uropathogenic Escherichia coli. Infect Immun 76: 3019–3026.
- 24. Wu KM, Li LH, Yan JJ, Tsao N, Liao TL, et al. (2009) Genome sequencing and comparative analysis of Klebsiella pneumoniae NTUH-K2044, a strain causing liver abscess and meningitis. J Bacteriol 191: 4492–4501.
- 25. Lai YC, Peng HL, Chang HY (2003) RmpA2, an activator of capsule biosynthesis in Klebsiella pneumoniae CG43, regulates K2 cps gene expression at the transcriptional level. J Bacteriol 185: 788–800.
- 26. Guisbert E, Rhodius VA, Ahuja N, Witkin E, Gross CA (2007) Hfq modulates the sigmaE-mediated envelope stress response and the sigma32-mediated cytoplasmic stress response in Escherichia coli. J Bacteriol 189: 1963–1973.
- 27. Zhang Y, Sun S, Wu T, Wang J, Liu C, et al. (2006) Identifying Hfq-binding small RNA targets in Escherichia coli. Biochem Biophys Res Commun 343: 950–955.
- 28. Dierksen KP, Trempy JE (1996) Identification of a second RcsA protein, a positive regulator of colanic acid capsular polysaccharide genes, in Escherichia coli. J Bacteriol 178: 5053–5056.
- 29. Cheng HY, Chen YS, Wu CY, Chang HY, Lai YC, et al. (2010) RmpA regulation of capsular polysaccharide biosynthesis in Klebsiella pneumoniae CG43. J Bacteriol 192: 3144–3158.
- 30. Gottesman S (2005) Micros for microbes: non-coding regulatory RNAs in bacteria. Trends Genet 21: 399–404.
- 31. Urban JH, Vogel J (2007) Translational control and target recognition by Escherichia coli small RNAs in vivo. Nucleic Acids Res 35: 1018–1037.
- 32. Morita T, Maki K, Aiba H (2005) RNase E-based ribonucleoprotein complexes: mechanical basis of mRNA destabilization mediated by bacterial noncoding RNAs. Genes Dev 19: 2176–2186.
- 33. Brown L, Elliott T (1996) Efficient translation of the RpoS sigma factor in Salmonella typhimurium requires host factor I, an RNA-binding protein encoded by the hfq gene. J Bacteriol 178: 3763–3770.
- 34. Muffler A, Fischer D, Hengge-Aronis R (1996) The RNA-binding protein HF-I, known as a host factor for phage Qbeta RNA replication, is essential for rpoS translation in Escherichia coli. Genes Dev 10: 1143–1151.
- 35. Roop RM 2nd, Gee JM, Robertson GT, Richardson JM, Ng WL, et al. (2003) Brucella stationary-phase gene expression and virulence. Annu Rev Microbiol 57: 57–76.
- 36. Figueroa-Bossi N, Lemire S, Maloriol D, Balbontin R, Casadesus J, et al. (2006) Loss of Hfq activates the sigmaE-dependent envelope stress response in Salmonella enterica. Mol Microbiol 62: 838–852.
- 37. Berghoff BA, Glaeser J, Sharma CM, Zobawa M, Lottspeich F, et al. (2011) Contribution of Hfq to photooxidative stress resistance and global regulation in Rhodobacter sphaeroides. Mol Microbiol.
- 38. Barra-Bily L, Fontenelle C, Jan G, Flechard M, Trautwetter A, et al. (2011) Proteomic alterations explain phenotypic changes in Sinorhizobium meliloti lacking the RNA chaperone Hfq. J Bacteriol 192: 1719–1729.
- 39. Tu YC, Lu MC, Chiang MK, Huang SP, Peng HL, et al. (2009) Genetic requirements for Klebsiella pneumoniae-induced liver abscess in an oral infection model. Infect Immun 77: 2657–2671.
- 40. Lee T, Feig AL (2008) The RNA binding protein Hfq interacts specifically with tRNAs. RNA 14: 514–523.
- 41. Senear AW, Steitz JA (1976) Site-specific interaction of Qbeta host factor and ribosomal protein S1 with Qbeta and R17 bacteriophage RNAs. J Biol Chem 251: 1902–1912.
- 42. Vecerek B, Moll I, Blasi U (2005) Translational autocontrol of the Escherichia coli hfq RNA chaperone gene. RNA 11: 976–984.
- 43. Le Derout J, Boni IV, Regnier P, Hajnsdorf E (2011) Hfq affects mRNA levels independently of degradation. BMC Mol Biol 11: 17.
- 44. Butland G, Peregrin-Alvarez JM, Li J, Yang W, Yang X, et al. (2005) Interaction network containing conserved and essential protein complexes in Escherichia coli. Nature 433: 531–537.
- 45. Mohanty BK, Maples VF, Kushner SR (2004) The Sm-like protein Hfq regulates polyadenylation dependent mRNA decay in Escherichia coli. Mol Microbiol 54: 905–920.
- 46. Sukhodolets MV, Garges S (2003) Interaction of Escherichia coli RNA polymerase with the ribosomal protein S1 and the Sm-like ATPase Hfq. Biochemistry 42: 8022–8034.
- 47. Folichon M, Allemand F, Regnier P, Hajnsdorf E (2005) Stimulation of poly(A) synthesis by Escherichia coli poly(A)polymerase I is correlated with Hfq binding to poly(A) tails. FEBS J 272: 454–463.
- 48. Hajnsdorf E, Regnier P (2000) Host factor Hfq of Escherichia coli stimulates elongation of poly(A) tails by poly(A) polymerase I. Proc Natl Acad Sci U S A 97: 1501–1505.
- 49. Le Derout J, Folichon M, Briani F, Deho G, Regnier P, et al. (2003) Hfq affects the length and the frequency of short oligo(A) tails at the 3' end of Escherichia coli rpsO mRNAs. Nucleic Acids Res 31: 4017–4023.
- 50. Azam TA, Ishihama A (1999) Twelve species of the nucleoid-associated protein from Escherichia coli. Sequence recognition specificity and DNA binding affinity. J Biol Chem 274: 33105–33113.
- 51. Lai YC, Lin GT, Yang SL, Chang HY, Peng HL (2003) Identification and characterization of KvgAS, a two-component system in Klebsiella pneumoniae CG43. FEMS Microbiol Lett 218: 121–126.
- 52. Skorupski K, Taylor RK (1996) Positive selection vectors for allelic exchange. Gene 169: 47–52.
- 53. Lin CT, Peng HL (2006) Regulation of the homologous two-component systems KvgAS and KvhAS in Klebsiella pneumoniae CG43. J Biochem 140: 639–648.
- 54. Hensel M, Shea JE, Gleeson C, Jones MD, Dalton E, et al. (1995) Simultaneous identification of bacterial virulence genes by negative selection. Science 269: 400–403.
- 55. Nelson EC, Segal H, Elisha BG (2003) Outer membrane protein alterations and blaTEM-1 variants: their role in beta-lactam resistance in Klebsiella pneumoniae. J Antimicrob Chemother 52: 899–903.
- 56. Domenico P, Schwartz S, Cunha BA (1989) Reduction of capsular polysaccharide production in Klebsiella pneumoniae by sodium salicylate. Infect Immun 57: 3778–3782.
- 57. Blumenkrantz N, Asboe-Hansen G (1973) New method for quantitative determination of uronic acids. Anal Biochem 54: 484–489.
- 58. Chang HY, Lee JH, Deng WL, Fu TF, Peng HL (1996) Virulence and outer membrane properties of a galU mutant of Klebsiella pneumoniae CG43. Microb Pathog 20: 255–261.