The VrrA sRNA controls a stationary phase survival factor Vrp of Vibrio cholerae

Abstract

Small non-coding RNAs (sRNAs) are emerging regulatory elements in bacteria. The Vibrio cholerae sRNA VrrA has previously been shown to down-regulate outer membrane proteins (OmpA and OmpT) and biofilm matrix protein (RbmC) by base-pairing with the 5′ region of the corresponding mRNAs. In this study, we present an additional target of VrrA in V. cholerae, the mRNA coding for the ribosome binding protein Vrp. Vrp is homologous to ribosome-associated inhibitor A (RaiA) of Escherichia coli which facilitates stationary phase survival through ribosome hibernation. We show that VrrA down-regulates Vrp protein synthesis by base-pairing to the 5′ region of vrp mRNA and that the regulation requires the RNA chaperone protein, Hfq. We further demonstrate that Vrp is highly expressed during stationary phase growth and associates with the ribosome of V. cholerae. The effect of the Vrp protein in starvation survival is synergistic with that of the VC2530 protein, a homolog of the E. coli hibernation promoting factor HPF, suggesting a combined role for these proteins in ribosome hibernation in V. cholerae. Vrp and VC2530 are important for V. cholerae starvation survival under nutrient deficient conditions. While VC2530 is down-regulated in cells lacking vrrA, mutation of vrp results in VC2530 activation. This is the first report indicating a regulatory role for an sRNA, modulating stationary factors involved in bacterial ribosome hibernation.

Abbreviations

HPF

hibernation promoting factor

IF

initiation factor

RaiA

ribosome-associated inhibitor A

RMF

ribosome modulation factor

SRA

stationary-phase-induced-ribosome associated protein

Vrp

VrrA-regulated ribosome binding protein

Introduction

In bacteria, ribosomal fidelity plays a central role in the adaptation to environmental stresses and acts as a checkpoint for sensing shifts in temperature or nutrient levels. Ribosomes in all organisms are composed of a small and a large subunit (30S and 50S, respectively, in bacteria) that cycle through stages of association and dissociation during protein synthesis. When bacteria enter stationary phase, ribosome structure can be modified facilitating a state of hibernation.¹ The ribosome binding proteins RMF (ribosome modulation factor), HPF (hibernation promoting factor), RaiA (ribosome-associated inhibitor A) and SRA (stationary-phase-induced ribosome associated protein) are expressed in stationary phase and involved in ribosome hibernation.² In E. coli, RMF and HPF play a role in formation of 100S dimer ribosome particles. The binding of RMF causes dimerization of 70S ribosomes into 90S particles, which are further stabilized as 100S dimers upon HPF binding thus leads to “ribosome hibernation."² RaiA or YfiA, on the other hand is known to promote the formation of translationally inactive monomeric 70S ribosomes which consequently prevents the recycling of ribosomes for translation initiation.³ Although RaiA and HPF share 40% amino acid sequence similarity, HPF converts 90S into 100S particles, whereas RaiA prevents RMF-dependent 90S formation.⁴ Both the monomeric 70S ribosomes subunits promoted by RaiA and the 100S dimer particles promoted by HPF have been shown to be translationally inactive ^3,5 that could aid longer cell survival under nutrient limited conditions.

The ribosome hibernation helps bacteria to survive under nutrient-limited conditions.^1,2 The crystal structure of the Thermus thermophilus 70S ribosome reveals that RaiA and HPF share a common binding site in the ribosome. In addition, the same binding site overlaps with those of all tRNA and the initiation factors IF1 and IF3, suggesting that RaiA and HPF can interfere with protein synthesis.⁵ Recent studies in Lactococcus lactis which lacks orthologues of rmf and hpf genes, showed that RaiA (also known as YfiA) is essential for ribosome dimerization and found in 100S ribosome particles.⁶

In the past few years, it has become increasingly clear that bacterial small non-coding RNAs (sRNAs) regulate many diverse cellular processes. sRNAs mostly function as antisense regulators, affecting translation, degrading target mRNA, or binding and sequestering proteins.⁷ To date, ∼150 sRNAs have been predicted in Escherichia coli K12 by bioinformatics and experimental approaches.^7,8 Although more than 500 sRNAs were predicted for V. cholerae,⁹ only a few of the sRNAs have been experimentally studied, including RyhB, which is involved in iron utilization;¹⁰ the sRNAs (Qrr1-Qrr4, CsrB, CsrC, CsrD) that are involved in quorum sensing regulation;^11,12 the TarA and TarB sRNAs which are regulated by ToxT and involved in V. cholerae virulence;¹³ the MicX sRNA, which regulates outer membrane proteins;¹⁴ the IGR7 sRNA, which is involved in modulating carbon metabolism;⁹ and the sRNA TfoR, which regulates natural competence in response to chitin.¹⁵

Previously we demonstrated that VrrA sRNA inhibited the expression of the outer membrane protein OmpA causing increased release of outer membrane vesicles.^16,17 In addition, VrrA also reduced the expression of another major outer membrane protein, OmpT.¹⁸ Recently we have shown that VrrA plays a role in repressing the biofilm matrix protein, RbmC by base-paring to the 5′ region of rbmC mRNA.¹⁹ Analogous to its RybB and MicA counterparts from E. coli and Salmonella,^20,21 transcription of vrrA is controlled by the alternative sigma factor E (σ^E) and V. cholerae cells lacking vrrA display increased colonization of the intestine in an infant mouse model.¹⁶

Many studies have been conducted to elucidate the role of RaiA in ribosome hibernation during stationary phase^6,22. However, little is known about the regulatory factors involved in RaiA expression and the role of RaiA in bacterial physiology.

In this study we show that VrrA is the first sRNA that directly regulates Vrp (a homolog of RaiA) in an Hfq-dependent manner by base-pairing with the 5′ region of the vrp mRNA in V. cholerae. We denote the protein Vrp for VrrA-regulated ribosome binding protein of V. cholerae which is encoded by the vc0706 gene locus in the NCBI data base (http://www.ncbi.nlm.nih.gov/gene/?term=vc0706). In addition to VrrA, mutation of vc2530 (a homolog of HPF) results in upregulation of Vrp. Ribosome profile analysis of the wild-type strain of V. cholerae reveals that Vrp is a ribosome-associated protein present in the 30S, 70S, and 100S ribosome fraction. Under nutrient deficient conditions, cells lacking both vrp and vc2530 genes exhibited reduced starvation survival compared to wild-type, however, single deletion of vrp or vc2530 has no significant effect. Interestingly, while VC2530 is downregulated in cells lacking vrrA, mutation of vrp resulted in VC2530 activation. Our data provides the first evidence suggesting a role for sRNA in ribosome modulation.

Results

In silico prediction of vrp as a target of VrrA

In our earlier studies, we showed that VrrA targeted the translation of the outer membrane proteins OmpA, OmpT and the biofilm matrix -protein RbmC by direct binding to the 5′UTR of these target mRNAs. In order to find new targets of VrrA, we performed an in silico analysis using the Target RNA program.^23,24 The target predicted by the TargetRNA program gives vrp as a 4^th hit in the list with score value of −73. For the TargetRNA program the complete sequence of VrrA was used for target search. The pairing region between VrrA and vrp was further authenticated using the RNA hybrid program ²⁵ where we limit the sequence length for sRNA and target RNA by eliminating sequences corresponding to stem loops and other secondary structures predicted in VrrA. In analysis using the RNA hybrid program, we limited the VrrA sequence to an open loop single stranded stretch region from nucleotide +70 to +106 (as shown in Fig. S1) to pair with a target sequence in the vrp mRNA region from the transcriptional start site to 6 nt into the vrp coding region. The result obtained by the RNA hybrid program predicted that residues 72–84 of VrrA forms a 13-bp duplex with the −3 to −15 region (numbers relative to AUG start codon) of the vrp mRNA with a calculated free energy of −28.2 kcal/mol. This interaction would partially mask the ribosome binding site of vrp mRNA required for translation initiation (Fig. 1A).

Figure 1.

In silco prediction and detection of Vrp as a VrrA target. (A) Graphical representation of the proposed seed pair region of VrrA with the target vrp mRNA. VrrA binding sites were predicted using the RNAhybrid program (in blue) and shown to overlap with the ribosome binding site of vrp. The start codon of the vrp is shown in green. Compensatory base pair changes at the 5′UTR of vrp resulted in a new mutant, M3*-vrp::gfp (pDS18). (B) VrrA down-regulates expression of the Vrp protein. Western blot analyses of whole cell lysates collected after overnight growth from the V. cholerae strains: lane 1, A1552 (wild type); lane 2, DNY7 (vrrA mutant); lane 3, DNY11 (ΔvrrA/pvrrA); lane 4, DNY12 (ΔvrrA/vector); lane 5, DNY8 (Δhfq); lane 6, DNY9 (ΔhfqΔvrrA); lane 7, DNY16 (ΔhfqΔvrrA/pvrrA); lane 8, DNY17 (ΔhfqΔvrrA/pMMB66HE); lane 9, DHS380 (Δvrp) detecting Vrp and OmpU as a loading control (lower panel). (C) Semi-quantitative analyses of Vrp levels obtained with the same samples as in Fig. 1B, after SDS-PAGE and Western blot analyses of protein lysates from different derivatives of the V. cholerae strain A1552. The plotted data show relative levels of Vrp, with the level of the wild-type strain set to 1.0.

VrrA down- regulates the expression of Vrp in an Hfq-dependent manner

To investigate the role of VrrA in Vrp expression, we determined the Vrp levels in whole cells of wild-type V. cholerae O1 strain A1552 and the ΔvrrA mutant by Western-blot analysis using a polyclonal anti-Vrp antiserum (Fig. 1B). The levels of Vrp in different strains were analyzed semi-quantitatively using a Fluor-S Multi-Imager (Bio-Rad) (Fig. 1C). Vrp expression was increased to 2.2 fold in the vrrA mutant (DNY7) in comparison with the wild-type V. cholerae strain A1552 (Fig. 1B, compare lanes 1 and 2; Fig. 1C lanes 1 and 2). Elevated Vrp expression could be restored to the wild-type level by expressing VrrA from a plasmid harboring the wild-type allele of vrrA (Fig. 1B, lane 3; Fig. 1C lane 3) compared to the vector control (Fig. 1B, lane 4; Fig. 1C lane 4). Further the repression of vrp by VrrA was also determined by Northern blot analysis (Fig. S2A; Fig. S2B) and result obtained is in consistent with Western blot analysis (Fig. 1B, lanes 1 to 4; Fig. 1C, lanes 1 to 4). In the strain lacking Hfq, the Vrp level are unaffected by VrrA over-expression, (Fig. 1B, lane 7 and 8; Fig. 1C, lane 7 and 8). In our earlier studies ¹⁶, we observed that the total level of VrrA was higher in the hfq mutant than in the wild-type V. cholerae strain A1552 suggesting that the Hfq protein somehow might reduce the stability, and thereby the level, of VrrA or indirectly might affect its expression. In the present study, the increased level of VrrA was unable to regulate the expression of Vrp in the hfq mutant, suggesting that Hfq is involved in Vrp repression by VrrA. We also observed that in the hfq mutant the basal level of Vrp was higher (compare lane 1 with lane 5 in Fig. 1B). The apparent role in Vrp repression by Hfq might not only dependent on VrrA but could also involve some other sRNA.

Vrp is a stationary phase protein

We analyzed Vrp expression during different bacterial growth phases by Western blot analysis using anti-Vrp polyclonal antiserum (Fig. 2A). The higher expression of Vrp was detected when the bacterial cells were harvested from an overnight culture (Fig. 2A, upper panel, lane 4) although only a trace amount of Vrp was detectable at an OD₆₀₀ of 2.0 (Fig. 2A, upper panel, lane 3). Vrp was not detectable at an OD₆₀₀ of 0.5 and 1 (Fig. 2A, upper panel, lanes 1 and 2). The quantification of Western blot data was shown in Fig. 2B.

Figure 2.

Vrp is ribosome binding stationary phase protein. (A) Growth phase dependent expression of Vrp. Bacterial whole-cell lysate samples were taken at OD 0.5, OD 1.0, OD 2.0 and overnight culture (OVN). Western blot analysis was performed for detecting Vrp and the level of OmpU was used as a loading control. (B) Semi-quantitative analyses of Vrp levels obtained with the Western blot analyses of same samples as in Fig. 2A. (C) Ribosome profile of overnight grown wild-type V. cholerae O1 strain A1552 after sucrose gradient centrifugation as described in material and methods. The Y axis represents OD260nm and the X axis represents different fractions from top to bottom (low to high density) of the sucrose gradient. Peaks containing ribosome fractions 30S, 50S, 70S and 100S are indicated. (D) Western blot analysis of 30S, 50S, 70S and 100S fractions of wild-type V. cholerae O1 strain A1552 confirms the presence of Vrp in all fractions except 50S (lane 2 compared to lanes 1,3,4). Whole-cell lysate from the V. cholerae strains A1552 (wild-type) was used as a positive control (lane 5); DHS380 (vrp mutant) was used as a negative control (lane 6). RpoS protein, a non-ribosomal binding protein, was used as an internal control (lower panel).

Vrp, a homolog of RaiA is a ribosome binding protein in V. cholerae

We analyzed the ribosome profile of overnight grown wild-type V. cholerae O1 strain A1552 on a sucrose density gradient as described in the Materials and Methods. After ultracentrifugation, different fractions were collected and optical densities at 260nm were measured using a nano-drop spectrophotometer. Ribosome profiles based on OD260nm were plotted for wild-type V. cholerae O1 strain A1552 (Fig. 2C). To determine if Vrp is a ribosome-associated protein, we used the 30S, 50S, 70S and 100S fractions for Western blot analysis. We could detect the presence of Vrp in the all the tested fractions except for 50S of the wild-type V. cholerae O1 strain A1552 using anti-Vrp polyclonal antiserum (Fig. 2D, upper panel, lanes 1, 2, 3, and 4). In this experiment, whole-cell lysate of wild-type V. cholerae O1 strain A1552 and the Δvrp mutant were used as a positive and a negative control respectively (Fig. 2D, upper panel, lanes 5 and 6). As an internal control, we tested for the presence of the non-ribosome binding protein RpoS. As shown in Fig. 2D (lower panel, lanes 1, 2, 3, and 4) there was no RpoS protein band detected in the 30S, 50S, 70S and 100S fractions of the wild-type strain. The whole-cell lysate of wild-type V. cholerae O1 strain A1552 and the ΔrpoS mutant (Fig 2D, lower panel, lanes 5 and 6) were used as positive and negative controls, respectively. These findings confirm that Vrp is a ribosome binding protein in V. cholerae.

Expression of Vrp is up-regulated in the Δvc2530 mutant

In E. coli, there are 2 steps involved in formation of 100S particles through the formation of a 90S particle in the stationary phase of bacterial growth. YhbH (a.k.a. HPF), a homolog of V. cholerae VC2530 protein, converts immature 90S particles into mature 100S ribosomal particles by promoting both particle formation and stabilization. In contrast, YfiA (a.k.a. RaiA, a homolog of the Vrp) prevents 70S dimer formation. Thus, although YfiA and YhbH are highly homologous to each other, they have opposite functions in 70S dimer formation.^2,3,26 To examine the correlation between VC2530 and Vrp in V. cholerae, Vrp expression levels were examined in the presence or absence of VC2530 by Western blot analyses (Fig. 3A). The level of the Vrp protein was analyzed semi-quantitatively as shown in Fig. 3B. We observed an increased level of Vrp expression in the Δvc2530 mutant compared to the wild-type although it remains unclear whether Vrp down-regulate the expression of VC2530 directly or indirectly.

Figure 3.

Vrp and VC2530 are important factors for starvation survival of V. cholerae.(A) Western blot analyses of overnight grown whole-cell lysates from the V. cholerae strains. Lane 1, A1552; lane 2, Δvrp; lane 3, Δvc2530; and lane 4, Δvrp Δvc2530 detecting Vrp and OmpU as a loading control. (B) The levels of the Vrp protein in the same strains as in Fig. 3A were analyzed semi-quantitatively. (C) Western blot analyses of overnight grown whole cell lysates from the V. cholerae strains: lanes 1, A1552-flag-vc2530; lane 2, ΔvrrA-flag-vc2530; lane 3, Δvrp-flag-vc2530; and lane 4, Δvc2530 detecting FLAG and OmpU as a loading control. (D) The levels of the Flag-VC2530 protein in the same strains as in Fig. 3C were analyzed semi-quantitatively. (E) Starvation survival assay: The survival of V. cholerae wild-type strain A1552 and its mutant derivatives Δvrp, Δvc2530, ΔvrpΔvc2530, and ΔvrrA, were examined by measuring CFU/ml at 37°C in minimal media containing 0.4% glycerol. The data shown is representative of 3 independent experiments.

Expression of VC2530 is up-regulated in the Δvrp mutant and down-regulated in the ΔvrrA mutant

We analyzed the expression of Flag tagged VC2530 in wild-type V. cholerae strain flag-vc2530 (DHS458), ΔvrrA-flag-vc2530 (DHS474), Δvrp-flag-vc2530 (DHS475) and Δvc2530. Western blot analysis was performed using anti-Flag monoclonal antiserum, which detected the expected ∼12-kDa protein (Fig. 3C) and the levels of the Flag-VC2530 protein were analyzed semi-quantitatively (Fig. 3D). As shown in Fig. 3C and D, Flag-VC2530 expression was downregulated in ΔvrrA and up-regulated in Δvrp (1.4-fold increase compared to wild-type).

Impaired starvation survival of the V. cholerae double deletion mutant ΔvrpΔvc2530

To determine if Vrp and VC2530 modulate the survival of V. cholerae under nutrient limited conditions, we performed starvation survival assays for the wild-type V. cholerae strain A1552, Δvrp, Δvc2530, ΔvrpΔvc2530 and ΔvrrA strains.

As shown in Fig. 3E, at day 0, all V. cholerae strains started with equal cell numbers, as determined by CFU/ml. However, at day 3 and day 4, the starvation survival of the double deletion mutant ΔvrpΔvc2530 was reduced 3.3 fold compared to the wild-type. The starvation survival of Δvrp and the Δvc2530 single mutant were similar to that of the wild-type throughout the experiment. In addition, starvation survival of the ΔvrrA mutant was similar to that of the wild-type. It might be due to up-regulation of Vrp and down-regulation of VC2530 production.

VrrA targets the vrp mRNA at the 5′UTR

In a previous study, plasmids expressing wild-type VrrA or mutant variants of VrrA were constructed and transformed into V. cholerae strain ΔvrrA (DNY7).¹⁸ The mutant variants M1 and M3 carry nucleotide substitutions at 73–78 and 73–74 with respect to VrrA (Fig. 4A), overlapping with the region predicted to form a VrrA/vrp-mRNA duplex. Whole-cell lysates from these strains were analyzed by Western blot to compare the production of Vrp (Fig. 4B). The levels of the Vrp protein were semi- quantified as shown in Fig. 4C. We found that the sRNA variants expressed by plasmids pTS2-M1 and pTS2-M3 were impaired in their ability to repress the expression of Vrp (Fig. 4B, compare lane 2, with lanes 3 and 4; Fig. 4C, compare lane 2, with lanes 3 and 4) while the wild-type VrrA still maintained the downregulation of Vrp expression (Fig. 4B, lane 2; Fig. 4C, lane 2).

Figure 4.

VrrA interact at the 5′ region of vrp mRNA. (A) The sequence of WT-vrrA and its mutant derivates carrying nucleotide substitutions (M1-vrrA and M3-vrrA). Substituted nucleotides are underlined. (B) Wild-type VrrA and its substituted nucleotide mutants derivatives (M1 and M3) were introduced into the ΔvrrA strain (DNY7). Western blot analyses of overnight grown strains were performed to detect the levels of Vrp and OmpU (as a loading control). (C) The levels of the Vrp protein in the same strains as in Fig. 4B were analyzed semiquantitatively. (D) GFP-based reporter assay of cells carrying compensatory mutations in the predicted location of base pairing with VrrA. Western blot analysis of Vrp::GFP in Top10 cells collected at OD₆₀₀ 1.0, with plasmids carrying different vrrA derivatives. Lane 1, WT-vrp::gfp & pJV300(Vector control); lane 2, WT-vrp::gfp & WT-vrrA; lane 3, WT-vrp::gfp & M3-vrrA; lane 4, M3*-vrp::gfp & pJV300(VC); lane 5, M3*-vrp::gfp & WT-vrrA; lane 6, M3*-vrp::gfp & M3-vrrA. GroEL expression was used to normalize the loading amount. (E) Semi-quantitative Western blot analysis showing expression of Vrp::GFP for the same samples as shown in Fig. 4D. GroEL expression was used as reference to normalize the loaded amount. Results are representative of data obtained with 3 independent cultures. (F) Agar plate-based colony fluorescence imaging analysis of the same strains as in Fig. 1D and E. Top10 cells carrying both plasmids were grown on LB agar. The upper panel image was obtained in the fluorescence excitation at 460 nm, and light emission was recorded using a 510 nm-filter. The lower panel image shows the same plate under visible light mode.

Validation of the interaction between the VrrA and its target vrp by compensatory base pair changes

We used a previously developed gfp-based translational fusion system that allows rapid validation of a VrrA target.²⁷ This reporter system consists of 2 plasmids; a high-copy plasmid (pJV300) carrying the vrrA clone ( or VrrA mutant derivatives) is co-expressed with a low copy plasmid (pXG20) carrying the 5′ UTR of wild-type vrp fused to gfp (green fluorescent protein) or M3*-vrp::gfp. In order to determine the transcriptional start site at the 5′UTR of vrp 5′ RACE was performed (Fig. S3). To study regulation at the seed pairing region of the M3-vrrA with the vrp mRNA, we created a mutant derivative of vrp at its 5′UTR termed M3*-vrp::gfp (Fig. 1A) by changing nucleotides at location −4GC to CG. Even in the absence of VrrA, the mutant M3*-vrp::gfp (pDS18) exhibited a reduced GFP signal compared to the WT-vrp::gfp (pTS32) (Fig. S4C, lanes 1 and 2, upper panel; Fig. S4D, lanes 1 and 2). It is possible that in m3*-vrp::gfp changing nucleotides at location −4GC to CG might affect ribosome binding efficiency. Nonetheless, the substituted nucleotide exchange mutants were used to test the regulatory effect by co-expression of the WT-vrrA (wild-type) or the M3-vrrA with that of WT-vrp::gfp (pTS32) or the mutated M3*-vrp::gfp (pDS18).

The compensatory M3* allele of the vrp::gfp can restore base pairing with M3-vrrA comparable to that of the WT-vrrA (Fig. 4D, lanes 5 and 6). As a loading control, GroEL was detected by Western blot analysis using anti-GroEL polyclonal antiserum (Fig. 4D, lower panel). Figure 4E shows semi-quantitative Western blot analysis of Vrp::GFP expression (normalized to GroEL) in the same samples as in Fig. 4D. These results demonstrate that compensatory mutations restore regulation by the corresponding variant of VrrA and disrupt regulation by WT-vrrA. To support the Western blot analysis of GFP in E. coli strains carrying different gfp fusion plasmids, we monitored the colony fluorescence using an agar plate-based colony fluorescence imaging as described in the Materials and Methods. As shown in Fig. 4F (upper panel), the results obtained by the agar plate-based colony fluorescence imaging was in agreement with the Western-blot results in Fig. 4D and E. Fig. 4F (lower panel) shows the imaging of the same whole-cell colony plate under the normal light as a control.

Discussion

Bacteria can reduce the efficiency of protein synthesis during nutrient starvation or in stationary phase by converting ribosomes into translationally inactive 100S dimers¹ or 70S monomers,²⁶ thereby inhibiting bacterial growth. The major players in the stationary phase conversion of ribosome conformation are RMF, HPF, and RaiA (in different bacterial species),⁵ but few known regulators of RMF and RaiA have been described.

In earlier studies, ppGpp was shown to be required for rmf transcription and it was suggested that intracellular ppGpp levels not only regulate ribosomal RNA synthesis and translational factor synthesis, but also modulate the translational activity of ribosomes during stationary phase by controlling formation of the inactive form (100S) of the 70S ribosome.²⁸ Recently, it was demonstrated that the bacterial metabolic regulator protein, cyclic AMP receptor protein (CRP), regulates transcriptional activation of the rmf gene for formation of 100S ribosome dimers in E. coli.²⁹ However, post transcriptional regulation of ribosome modulation factors had not been studied in bacteria.

In this study, we show that vrp is the first ribosome binding protein in Vibrio cholerae that is a direct target of sRNA, VrrA of V. cholerae. VrrA targets vrp mRNA by base-pairing at its ribosome binding site, thus decreasing vrp expression. The secondary structure of VrrA predicted by RNA fold software (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi) contains the open loop sequence from nucleotides 70 to 106 (Fig. S1), and part of this sequence from nucleotides 72 to 84 interacts with vrp mRNA. Based on this interaction, VrrA and its substituted mutant derivatives (M1 and M3) were used. Interestingly, M1 and M3 mutant derivatives of VrrA lost the ability to regulate Vrp expression.

V. cholerae can survive the extended adverse conditions of nutrient starvation, elevated salinity and decreased temperature by entering into a dormant, viable but nonculturable (VBNC) state.³⁰ Bacteria fail to grow on routine bacteriological media on which they would normally form colonies, but are alive and capable of renewed metabolic activity.³¹ The molecular mechanism of starvation survival of V. cholerae is not yet largely understood although the phosphoporin protein of V. cholerae has been suggested to play a role in the life cycle and survival of the bacterium in the natural environment.³²

In E. coli, RMF and HPF (homolog of VC2530) results in formation of 100S dimer ribosome particles while RaiA (homolog of Vrp) or YfiA, promotes the formation of translationally inactive monomeric 70S ribosomes.^2,5 The homologue of Vrp, RaiA was shown to associate preferentially with 70S ribosomes and the copy number of RaiA in the HPF deletion mutant was approximately 1.6 times higher than that of the wild-type strain.² In the wild-type V. cholerae strain A1552, we found that Vrp was associated mainly with 70S ribosomes and a lesser amount was associated with 30S and 100S ribosomes. In addition, the Vrp expression was higher in the Δvc2530 mutant in comparison with the wild-type strain. However, as for the observed elevated synthesis of Vrp in the Δvc2530 mutant, it remains likewise enigmatic how VrrA and Vrp "down- and up-regulate" VC2530, respectively. The significance of these reciprocal changes in levels of Vrp and VC2530 in the bacterial cells is not clear as yet. The mechanism(s) behind this regulation remained to be determined. In E. coli, inactivation of YfiA (Vrp homolog) stimulates ribosome dimerization as the YfiA interferes with the binding of RMF and HPF to the ribosomes. The E. coli yfiA::Km mutant appears to be somewhat more viable in the stationary phase. The enhanced viability of this strain has been explained by increased protection of 100S ribosomes against degradation by RNA hydrolases.^33,34

In this study, we demonstrated the role of ribosome binding proteins Vrp (homolog of RaiA) and VC2530 (homolog of HPF) in survival of V. cholerae under nutrient limited conditions. In V. cholerae, the vrp and vc2530 single-deletion mutants were able to survive to an extent similar to that of the wild-type strain, probably by maintaining stabilized 70S monomeric and 100S dimeric forms of ribosomes, respectively, during stationary phase. The previous studies in E. coli showed that RaiA and HPF have the same binding site on the 30S subunit of the ribosome indicating that both proteins might have similar function in stabilizing ribosomes during stationary phase.⁵ The expression level of Vrp was higher in the vc2530 deletion mutant than in the wild-type; excess Vrp protein may therefore be involved in stabilization of ribosomal subunits, potentially allowing the vc2530 deletion mutant to live as long as the wild-type. The vrp mutant also showed the same survival time as the wild-type under nutrient-limited conditions. In the vrp mutant the level of VC2530 expression is also higher than in the wild-type; excess VC2530 protein may stabilize the ribosome, thereby protecting the ribosome dimers from degradation and allowing the same survival time as for the wild-type. Deletion of both genes, vrp and vc2530, might result in a less stable ribosome, which leads to a reduction in starvation survival ability. However, the results obtained in our studies in V. cholerae appeared to be rather different from the findings by Ueta et al. in E. coli ² where viability of the ΔraiAΔhpf mutant in EP media was similar to that of the wild-type strain. Recently, it was shown that in Lactobacillus lactis, a ΔyfiA mutant did not display altered survival phenotype in rapidly growing cells, but the viability of the mutant was reduced when the bacterial strain was starved for carbon and energy sources for 10–20 d⁶ These differences might be due to the different bacterial species and the usage of different growth conditions. From the results of the present study, we suggest that Vrp, which is under the direct regulation of VrrA, and VC2530, which is indirectly regulated by VrrA, are important for V. cholerae starvation survival under nutrient-limited conditions.

Materials and Methods

Bacterial strains and growth conditions

E. coli and V. cholerae strains used in this study are listed in Table 1. Bacterial strains were grown at 37°C in Luria Bertani (LB) broth or LB plates supplemented with antibiotics (where appropriate) at the following concentrations: 100 μg/ml carbenicillin, 20 μg/ml chloramphenicol, 50 μg/ml rifampcin.

Table 1.

Bacterial strains used in this study

Bacterial strain	Description / Relevant Genotype	Source or reference
E. coli
DH5α	λ− φ80dlacZΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17(rK− mK−) supE44 thi-1 gyrA relA1	38
SM10λpir	thi thr leu tonA lacY supE recA::RP4–2 Tc::Mu Km λpir	39
TOP10	F− mcrA Δ(mrr-hsdRMS- mcrBC) φ80lacZΔM15 ΔlacX74nupG recA1 araD139Δ(ara-leu)7697 galE15 galK16rpsL(StrR) endA1 λ	Invitrogen
V. cholerae
A1552	O1 ElTor, Inaba, Rif R	40
DHS 380	Δvrp derivative of A1552	This study
DHS 389	Δvrp/pBAD18	This study
DHS 391	Δvrp/pBAD18:vrp	This study
DHS 458	A1552-flag-vc2530	This study
DHS 474	A1552ΔvrrA-flag-vc2530	This study
DHS 475	A1552Δvrp-flag-vc2530	This study
A1552Δvc2530	Δvc2530	This study
DHS 415	ΔvrpΔvc2530 derivative of A1552	This study
A1552Δrpos	Δrpos	41
DNY7	A1552ΔvrrA	16
DNY8	A1552Δhfq	16
DNY9	A1552ΔvrrA Δhfq	16
DNY11	A1552ΔvrrA + pvrrA	16
DNY12	A1552ΔvrrA + pMMB66HE	16
DNY16	A1552ΔvrrA Δhfq + pvrrA	16
DNY17	A1552ΔvrrA Δhfq + pMMB66HE	16
DNY34	A1552ΔvrrA + pJV300	18
DNY35	A1552ΔvrrA + pTS2	18
DNY44	A1552ΔvrrA + pTS2-M1	18
DNY64	A1552ΔvrrA + pTS2-M3	18
DNY 88	Top10 + pJV300 + pTS32	This study
DNY 89	Top10 + pTS2 + pTS32	This study
DNY 90	Top10 + pTS2-M1 + pTS32	This study
DHS-324	Top10 + pTS2-M3 + pTS32	This study
DHS-325	Top10 + pJV300 + pTS32-M3*	This study
DHS-326	Top10 + pTS2 + pTS32-M3*	This study
DHS-327	Top10 + pTS2-M3 + pTS32-M3*	This study

Primers and Plasmids

Primers and Plasmids used in this study are listed in Tables 2 and 3 respectively.

Table 2.

Primers used in this study

Primer	Sequence in 5′ to 3′ direction	Restriction site	Used for construction of
vrp-2	GTTTTTGCTAGCGATTGCAGAGGTGACTTC	NheI	pTS32 and 5′ Race for vrp
JVO-367	ACTGACATGGAGGAGGGA		pTS32 and 5′ Race for vrp
JVO-8239	AAAACGATTATGAAAATCAACAT		pTS32-M3*
JVO-8240	TAATCGTTTTTCCTCTGTGTC		pTS32-M3*
DS125	CGCTCTAGA AACTCGAGGC TTATCAGCAG	Xba1	vrp deletion mutant
DS126	CCCATCCACTATAAACTAACA AATGCTTTTTCCTCTGTGTCA		vrp deletion mutant
DS127	TGTTAGTTTATAGTGGATGGG ATAGAATGATGG GAGATAGCGC		vrp deletion mutant
DS128	AATTCTAGA TGTATTGAGGATCGAGCTGA		vrp deletion mutant
DS96	GC GAATTCCATAAGGGATGACACAGAG G	EcoR1	pBAD18:vrp
DS77	GCTCTAGATTATTCCACTTCTTCGCTCAG	Xba1	pBAD18:vrp
DS166	CGCTCTAGA GAAATACATG CATACCCCGC	Xba1	FLAG insertion vc2530
DS167	CTTGTCGTCATCGTCTTTGTAGTC CATAGACTTTCCTTCTCTAGTTTAGG		FLAG insertion vc2530
DS168	GACTACAAAGACGATGACGACAAG CAAATCAAC ATTCAAGGCCA		FLAG insertion vc2530
DS169	CGCTCTAGA AGTACTTCGCTCAATTGCAT	Xba1	FLAG insertion vc2530
DS170	CGAGACACTG CTCAAAGTTG		FLAG insertion vc2530
DS171	CTTGTCGTCATCGTCTTTGTAGTC		FLAG insertion vc2530
VC2530-A	CGCTCTAGA CGTACCAATTTGCAAGAGGCA		vc2530 deletion mutant
VC2530-B	CCCATCCACTATAAACTAACA GCCTTGAATGTTGATTTGCATAG		vc2530 deletion mutant
VC2530-C	TGTTAGTTTATAGTGGATGGG CATTAATCATGCAATTGAGCG		vc2530 deletion mutant
VC2530-D	CGCTCTAGAGGGCAGGTTATCGACACAGTA	Xba1	vc2530 deletion mutant
JVO-8168	AAGGGATGACACAGAGGAA		vrp riboprobe
JVO-8169	GTTTTTTTAATACGACTCACTATAGGGAGGTTGAGTTTACCCTCGATATGA		vrp riboprobe
JVO-8106	CTGTTTCGTTTCACTTCTGAGTTC		5S rRNA Probe

Table 3.

Plasmid used in this study

Original name	Plasmid Trival name	Relevant genotype/phenotype	Reference/source
pGEM-T Easy			Promega
pCR4-TOPO		TA cloning vector plasmid; Apr	Invitrogen
pBAD18		Cloning vector plasmid; Apr	42
pCVD442		Apr positive-selection suicide vector plasmid	43
pDS35	pBAD18:vrp	pBAD18 carrying vrp; Apr	This study
pXG20		Control plasmid for gfp fusion assays	27
pMMB66HE		Control plasmid	45
pvrrA		vrrA complementation plasmid, based on pMMB66HE	18
pJV300		ColE1 plasmid expressing a ∼50-nt nonsense transcript derived from rrnB terminator	27
pTS2	WT-vrrA	ColE1 plasmid expressing vrrA from its own promoter	18
pTS2-M1	M1-vrrA	pTS2 carrying a 6-nt substitution in putative vrp interaction sequence, as shown in Fig. 4A	18
pTS2-M3	M3-vrrA	pTS2 carrying a 2-nt substitution in putative vrp interaction sequence,as shown in Fig. 4A	18
pTS32	WT-vrp::gfp	Vibrio vrp translational GFP fusion Plasmid(pXG20) to 16th amino acid	This study
pDS18	M3*-vrp::gfp	Vibrio vrp translational GFP fusion Plasmid(pXG20) to 16th amino acid, Point mutation in vrp with change at the putative VrrA interaction sequence position from GC to CG as shown in Fig. 1A	This study

Construction of deletion mutants

In-frame deletions were constructed using procedures that have been described previously by Vaitkevicius et al.³⁵. Primer sequences are summarized in Table 2. Deletion of the vrp, vc2530, and double deletion ΔvrpΔvc2530 loci in V. cholerae strain A1552 resulted in DHS380, A1552Δvc2530 and DHS415, respectively.

Construction of chromosomal FLAG-tagged VC2530 strains

The V. cholerae chromosomal FLAG-tag insertion into vc2530 locus was carried out using a slight modification of the method described earlier by Skorupski and Taylor ³⁶. The FLAG tag sequence (CTTGTCGTCATCGTCTTTGTAGTC) was inserted after the start codon of the vc2530 gene, using primers DS166 and DS167 to generate a 326-bp product and, primers DS168 and DS169 to generate a 331-bp product . Both PCR products containing FLAG over-hangs were excised from an agrose gel and purified. Both fragments were mixed in equal concentrations at a 1:1 ratio used as a template for another PCR with primers DS166 and DS169, resulting in amplification of the vc2530 gene with a FLAG insertion (633-bp). Further, the resulting 633-bp PCR product containing the FLAG insertion was ligated into the vector pCVD442. The resulting plasmids were then integrated into the chromosome at the vc2530 loci by homologous recombination. Following sucrose selection, positive colonies containing the vc2530 loci with the FLAG insertion were verified by PCR using primers DS170 and DS171.

Construction of plasmids pTS32 and pDS18

WT-Vrp::gfp (pTS32) was constructed by following the protocol for “5′ RACE product cloning” as described by Urban and Vogel ²⁷. This protocol includes a 5′ RACE to determine the transcriptional start site of vrp and subsequent cloning into gfp-vector pXG20. In brief, total RNA was isolated from strain V. cholerae A1552 and treated with tobacco acid pyrophosphatase to enrich primary transcripts. After ligation of a 5′ end RNA adaptor, the RNA was converted to cDNA by reverse transcription with a random hexamer primer mix. Subsequently, a PCR reaction was performed using primers JVO-0367 (binding specifically to the RNA adaptor) and Vrp-2 (binding specifically to the vrp gene). The enriched PCR amplicon was gel extracted, digested with BseRI/NheI and subsequently cloned into a BsgI/NheI digested pXG20 backbone. The resulting plasmid pTS32 served as a template for creating and M3*-vrp::gfp (pDS18) by introducing nucleotide changes at location -4GC to CG using primer pair JVO-8239 and JVO-8240.

Anti-Vrp polyclonal antiserum preparation

The vrp gene was amplified by PCR using primers DS96 and DS77 and cloned into pBAD18 using EcoR1 and Xba1 restriction enzyme sites. Vrp expression was induced using 0.02% arabinose. The Vrp protein band was excised from an SDS gel, eluted from the gel, and used to generate polyclonal rabbit antiserum (AgriSera AB, Sweden).

Western blot analysis

Bacterial cultures were collected at OD₆₀₀ 1.0 for preparing GFP protein samples from whole-cell fractions. Overnight cultures (23 h) were used to prepare the Vrp samples. The protein samples were re-suspended in 1X sample loading buffer (50 mM Tris-HCl, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100mM β-mercaptoethanol), heated at 100°C for 15 min and separated by SDS–PAGE ³⁷. GFP, GroEL and Vrp were detected using anti-GFP monoclonal (Roche #11814460001), anti-GroEL (Sigma #A8705) and anti-Vrp polyclonal antiserum (in this study), respectively. OmpU, a loading control, and the non ribosomal binding protein RpoS, an internal control, were detected using anti-OmpU and anti-rpoS polyclonal antiserum, respectively. Western blot detection was done using the ECL+ chemiluminescence system (GE Healthcare, United Kingdom). The levels of Vrp or GFP were analyzed semiquantitatively using a Fluor-S Multi- Imager (Bio-Rad).

Isolation of ribosome fractions by sucrose density gradient

Ribosome analysis was performed as described previously by Maki et al.²² with minor modifications. The bacterial cells were harvested from culture an overnight culture (23 h) and centrifuged at 15000 x g for 15 min at 4°C. The pellet was suspended in buffer A (100 mM ammonium acetate, 15 mM magnesium acetate, 20 mM Tris-HCl at pH 7.4, and 6 mM 2-mercaptoethanol), and vigorously vortexed with glass beads (0.1 mm diameter, Biospec products, inc) for 5 min at 4°C. The resulting suspension was centrifuged at 15000 x g for 10 min at 4°C. The supernatant obtained was kept on ice and the pellet was re-suspended in buffer A, vortexed and centrifuged again under the same conditions. This procedure was done twice. The combined supernatants were layered on 10–40% sucrose gradients and centrifuged at 35000 r.p.m (154,693 x g) for 4 hours at 4°C in a Beckman SW 40Ti rotor. The fractions obtained after ultracentrifugation were collected and the optical densities were measured at 260 nm using a nano drop spectrophotometer (Thermo Scientific). The OD260nm values were used to plot a graph on the Y axis and the fractions from the top to bottom (low to high density) were used to plot on the X axis. The fraction containing the 30S, 50S, 70S and 100S peak was analyzed by Western blot.

5′ RACE analysis

5′ RACE was performed as previously described by Urban and Vogel²⁷ to determine the transcription start sites of the vrp gene. Total RNA isolated from the wild-type V. cholera strain A1552 was used to generate cDNA. Oligo Vrp-2 and JVO-0367 were used as specific primers in PCR. The PCR products were separated on a 2% agarose gel, gel-eluted and used as template for sequencing.

Gfp-based reporter assays

A gfp-based translational fusion system was constructed as previously described by Urban and Vogel²⁷. The reporter system consists of 2 plasmids a high-copy plasmid, pJV300 carrying the vrrA clone (pTS2), is co-expressed with a low-copy plasmid, pXG20 carrying the 5′ UTR of vrp fused to gfp (pTS32) or pKS18. E. coli Top10 cells carrying both plasmids pTS2 and pTS32 variants were grown overnight and samples were prepared for Western blot analysis.

Agar plate-based colony fluorescence imaging

E. coli Top10 cells carrying gfp fusion plasmids were grown overnight in LB broth. The overnight grown bacterial culture was diluted 100 times and 10 μl of diluted sample was dropped onto a LB plate supplemented with the appropriate antibiotics. After overnight incubation at 37°C, the plates were photographed in a FUJI LAS-4000 image analyzer using a CCD camera with a 510 nm emission filter and excitation at 460 nm.

sRNA target analysis

The TargetRNA analysis tool (http://snowwhite.wellesley.edu/targetRNA/index2.html) was used to predict the in-silico analysis of VrrA targets. The program was used with the following parameters: removed terminator of sRNA, target confined to mRNA region from −30 to +20, hybridization seed region value of 9, and with no G-U seed pairing.

The RNA hybrid program (RNAhybrid version 2.2)²⁵ was used to confirm the accuracy of the predicted interaction between the VrrA and vrp mRNA. The query sequences used for the vrp mRNA included the region from the transcriptional start site to 6 nt of the vrp coding regions and for VrrA we used the open loop stretch nucleotide sequences (+70 to +106) as shown in Fig. S1.

RNA extraction and Northern blotting

RNA was isolated using Trizol extraction as previously described Song et al.¹⁶. The bacterial cultures were centrifuged and the supernatant was decanted. The pellet was dissolved in 1 ml Trizol reagent. The mixture was transferred to 2 ml PhaseLock Tubes containing 400 μl chloroform, the samples were gently mixed by shaking and centrifuged (16,000 x g) for 15 min at room temperature. The supernatant containing RNA was transferred to a new tube and RNA was precipitated by addition of 2.5 volumes of isopropanol. The RNA was treated with DNase to remove any DNA contamination as described by Song et al.¹⁶. For Northern blot analysis, 10 μg RNA sample were resolved in a polyacrylamide or an agarose gel and transferred to a Hybond-XL membrane (GE Healthcare) by electro-blotting (1 h, 50 V, 4°C) in a tank blotter (Peqlab). Radiolabeled probes were used to visualize the required mRNA or sRNA. Northern blots were exposed to a phosphorimager screen and scanned on a StormTM phosphorimager (Molecular Dynamics, USA). Quantification was performed using Quantityone software (Roche). For vrp mRNA detection, Ribo-probes were prepared by PCR using primers JVO-8168 and JVO-8169, labeled with 5′P32 (α-P32-UTP), resulting in detection of 402 bp corresponding to vrp, and radio labeled (γ-P32-ATP) oligo probe (JV0–8106) was used to detect the 120-bp fragment corresponding to 5S rRNA.

Bacterial starvation survival assay

Bacterial cells grown overnight in LB media were isolated and washed 3 times with a minimal medium containing 1X M9 salt (For 100 ml of 10X M9 salt: 12.8 g Na₂HPO₄⋅7H₂O, 3 g KH₂PO₄, 0.5 g NaCl and 1 g NH₄Cl ), 2 mM MgSO₄, 0.1 mM CaCl₂, 0.4% glycerol. After washing, the bacterial cells were re-suspended in the same minimal media and the colony forming units of the bacterial suspension was adjusted to be 1 × 10⁷ CFU/ml. The samples were incubated at 37°C. The survival of bacteria in this minimal media was assessed by daily CFU/ml determination.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This work was performed within the Umeå Centre for Microbial Research (UCMR) Linnaeus Program and was supported by grants from the Swedish Research Council, the Faculty of Medicine at Umeå University and the National Institutes of Health (NIH). KP was supported by a Post-Doctoral fellowship from the Human Frontiers Science Program (HFSP).

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.