Identification and function of the RNA chaperone Hfq in the Lyme disease spirochete Borrelia burgdorferi

Summary

Hfq is a global regulatory RNA-binding protein. We have identified and characterized an atypical Hfq required for gene regulation and infectivity in the Lyme disease spirochete Borrelia burgdorferi. Sequence analyses of the putative B. burgdorferi Hfq protein revealed only a modest level of similarity with the Hfq from Escherichia coli, although a few key residues are retained and the predicted tertiary structure is similar. Several lines of evidence suggest that the B. burgdorferi bb0268 gene encodes a functional Hfq homolog. First, the hfq_Bb gene (bb0268) restores the efficient translation of an rpoS::lacZ fusion in an E. coli hfq null mutant. Second, the Hfq from B. burgdorferi binds to the small RNA DsrA_Bb and the rpoS mRNA. Third, a B. burgdorferi hfq null mutant was generated and has a pleiotropic phenotype that includes increased cell length and decreased growth rate, as found in hfq mutants in other bacteria. The hfq_Bb mutant phenotype is complemented in trans with the hfq gene from either B. burgdorferi or, surprisingly, E. coli. This is the first example of a heterologous bacterial gene complementing a B. burgdorferi mutant. The alternative sigma factor RpoS and the outer membrane lipoprotein OspC, which are induced by increased temperature and required for mammalian infection, are not upregulated in the hfq mutant. Consequently, the hfq mutant is not infectious by needle inoculation in the murine model. These data suggest that Hfq plays a key role in the regulation of pathogenicity factors in B. burgdorferi and we hypothesize that the spirochete has a complex Hfq-dependent sRNA network.

Introduction

Borrelia burgdorferi is the causative agent of Lyme disease (Burgdorfer et al., 1982; Benach et al., 1983; Steere et al., 1983; Radolf et al., 2010). The spirochete cycles between a tick vector and a vertebrate host (Lane et al., 1991; Spielman, 1994; Piesman and Schwan, 2010), altering its gene expression to transition between and survive within these two vastly different environments (Singh and Girschick, 2004; Samuels and Radolf, 2009; Skare et al., 2010). The increase in temperature associated with transmission from the tick to the mammal is one signal known to increase synthesis of outer surface lipoprotein C (OspC) and other virulence factors (Schwan et al., 1995; Stevenson et al., 1995; Fingerle et al., 2000; Yang et al., 2000; Revel et al., 2002; Alverson et al., 2003; Ojaimi et al., 2003). OspC synthesis depends on the alternative sigma factor RpoS (σ^S or σ³⁸), which has emerged as a global regulator of the enzootic cycle (Hübner et al., 2001; Caimano et al., 2004; Burtnick et al., 2007; Caimano et al., 2007) and is required for infection in the murine model of Lyme disease (Caimano et al., 2004; Blevins et al., 2009). The response regulator Rrp2 and the alternative sigma factor RpoN regulate the transcription of rpoS in response to an unknown environmental signal(s) (Hübner et al., 2001; Yang et al., 2003), while post-transcriptional temperature-dependent RpoS synthesis is regulated by the small RNA (sRNA) DsrA_Bb (Lybecker and Samuels, 2007), the only sRNA identified to date in B. burgdorferi.

Recently, sRNAs have emerged as major regulators in the expression of many bacterial genes, including those involved in virulence and the stress response (Lease and Belfort, 2000; Hengge-Aronis, 2002; Repoila et al., 2003; Gottesman, 2004; Storz et al., 2004; Majdalani et al., 2005; Narberhaus et al., 2006; Serganov and Patel, 2007; Fröhlich and Vogel, 2009; Klinkert and Narberhaus, 2009). Many sRNAs post-transcriptionally regulate gene expression by base pairing with target trans-encoded mRNAs, affecting either their translation or stability (Repoila et al., 2003; Gottesman, 2004; Majdalani et al., 2005). In Escherichia coli, the sRNA DsrA base pairs with the rpoS mRNA upstream region, releasing the Shine-Dalgarno sequence and start site from a stem-loop that inhibits translation. In B. burgdorferi, DsrA_Bb and the upstream region of the rpoS transcript have extensive complementarity, suggesting a similar mechanism of translational regulation (Lybecker and Samuels, 2007). Most trans-acting antisense sRNAs require the RNA chaperone Hfq (Massé et al., 2003b; Gottesman, 2004; Storz et al., 2004; Valentin-Hansen et al., 2004; Majdalani et al., 2005; Romby et al., 2006; Brennan and Link, 2007; Waters and Storz, 2009). Hfq is a highly conserved RNA-binding protein that was first identified in E. coli as a host factor required for Qβ bacteriophage replication (Franze de Fernandez et al., 1968). Hfq is the bacterial homolog of the eukaryotic and archaeal Sm and LSm proteins (Arluison et al., 2002; Möller et al., 2002; Schumacher et al., 2002; Sun et al., 2002; Zhang et al., 2002; Sauter et al., 2003). Sm and LSm proteins are involved in various aspects of RNA metabolism, including processing of nuclear RNA, RNA localization, and mRNA decay. The Sm and LSm proteins are characterized by two highly conserved motifs, Sm1 and Sm2, and form cyclic ring-shaped oligomers (Achsel et al., 1999; Kambach et al., 1999; Achsel et al., 2001). The Hfq protein is an oligomeric torus, like Sm and LSm (Törö et al., 2001; Törö et al., 2002). Whereas Sm and LSm proteins form heptameric structures that are homomeric in archaea and heteromeric in eukaryotes, most Hfq homologs are found as homohexamers (reviewed in Brennan and Link, 2007).

Hfq hexamers have two independent RNA-binding surfaces that have different specificities. The proximal surface was defined based on the Staphylococcus aureus crystal structure and includes the Sm2 motif (Schumacher et al., 2002). RNAs that bind this surface typically have AU-rich single-stranded sequences either just 3′ or just 5′ of a stem-loop (Brescia et al., 2003; Moll et al., 2003b; Geissmann and Touati, 2004; Večerek et al., 2005). These RNAs bind by circumnavigating the proximal face of the central cavity. The distal face of Hfq, on the opposite side of the torus, also binds RNAs (Mikulecky et al., 2004). This surface, which has been characterized in a recent X-ray crystal structure, is typically associated with poly(A) and A-rich structures (Link et al., 2009). The crystal structure predicts that the distal surface has specificity for purine-rich repeats with the sequence (ARN)_2–5. However, several other secondary structural regions have been identified as Hfq-binding sites, including hairpin motifs, a loop region connecting two stems in a pseudoknot structure, and the T- and D-stems of tRNA (Antal et al., 2005; Lee and Feig, 2008).

Hfq functions as an RNA chaperone by facilitating the interaction of an sRNA with its target mRNA via colocalizing the RNAs and/or altering the structure of the RNAs (Sledjeski et al., 2001; Massé and Gottesman, 2002; Møller et al., 2002; Zhang et al., 2002; Kawamoto et al., 2006; Večerek et al., 2007). Hfq protects sRNAs from degradation likely by blocking the RNase E binding site on the sRNA (Massé et al., 2003a; Moll et al., 2003a). Hfq also functions in mRNA turnover by stabilizing an mRNA, stimulating turnover of the mRNA by promoting polyadenylation, or facilitating base pairing between an sRNA and its target RNA resulting in cleavage of both RNAs by RNase E (Tsui et al., 1997; Vytvytska et al., 2000; Folichon et al., 2003).

An E. coli hfq null mutant has a pleiotropic phenotype that includes decreased growth rate, increased cell length, and increased sensitivity to stress conditions (Tsui et al., 1994; Muffler et al., 1997). Hfq is required for post-transcriptional regulation of rpoS and, therefore, the syntheses of more than 50 proteins are affected by hfq mutations (Muffler et al., 1996; Brown and Elliott, 1997). Moreover, Hfq plays an important role in several pathogenic bacteria: hfq mutations either attenuate or abolish virulence (reviewed in Chao and Vogel, 2010). Hfq is largely conserved in bacteria (Sun et al., 2002); however, B. burgdorferi lacks an annotated hfq gene in its genome (Fraser et al., 1997) and BLAST searches using sequences of conserved hfq genes do not reveal a homolog. We now identify and characterize an unusual Hfq in B. burgdorferi that regulates RpoS and is required for murine infection by needle inoculation.

Results

Identification of Hfq in B. burgdorferi

The RNA chaperone Hfq is typically required for RNA-RNA interactions between small regulatory RNAs and their cognate target mRNAs. After identifying and characterizing the first sRNA, DsrA_Bb, in B. burgdorferi, we hypothesized that the spirochete has an Hfq despite the lack of an annotated, or readily identifiable, homolog in its genome. This was supported by an MFOLD (Zuker, 2003) secondary structure prediction of DsrA_Bb that revealed a putative Hfq-binding site, an AU stretch of residues situated between two stable stem loops (Fig. 1). BLAST searches of the B. burdorferi genome were performed using as queries the Hfq sequences from E. coli and S. aureus, as well as just the highly conserved N-terminal 68 amino acids of E. coli Hfq. No homologous sequences were found in B. burgdorferi (data not shown). We next searched the genome for small ORFs that had the potential to form Sm1 and Sm2 motifs, and we identified a conserved hypothetical ORF encoded by the bb0268 gene, which was predicted to have 33% similarity and 12% identity with E. coli Hfq. A ClustalW alignment revealed that several important amino acids are conserved and a Phyre prediction suggested that this ORF could fold into an Hfq-like structure (Fig. 2). The residues comprising the nucleotide-binding pocket of Hfq, in the S. aureus Hfq-RNA complex (Schumacher et al., 2002), are conserved or similar in the predicted BB0268 protein. Other Hfq motifs (Sm1 NG and Sm2 F/YKHA) are partially conserved in BB0268. However, the BB0268 protein (159 amino acids) is considerably larger than most Hfq proteins (typically 77 to 110 amino acids). The N-terminal 68 amino acids of known Hfq homologs are highly conserved and include the Sm motifs, while the C-terminal portions of the proteins vary in length and sequence (Sauter et al., 2003). The ClustalW alignment demonstrated that the BB0268 sequence is most similar to the N-terminal domain of the canonical E. coli Hfq protein, but BB0268 has a longer C-terminal tail with a divergent sequence (Fig. 2A).

Fig. 1.

Predicted secondary structure of DsrA_Bb. The arrows flank the rpoS complementary region (34 nucleotides). The sequence predicted to be the Hfq-binding site is in bold. The open circle represents a loop of 32 nucleotides.

Fig. 2.

Hfq structural homology. (A) ClustalW alignment of the BB0268 conserved hypothetical protein with the E. coli (Ec) Hfq protein. Amino acids denoted by asterisks (*) are identical, colons (:) strongly similar and dots (.) weakly similar. The dashed boxes indicate the Sm1 and Sm2 motifs and the secondary structure of the E. coli Hfq is indicated above the sequence. Two of the signature motifs of Hfq, the Sm1 NG and the Sm2 F(Y)KHA, are underlined. The residues that are important in RNA binding by the S. aureus Hfq are shown in gray boxes. This figure panel is modified from Nielsen et al. (2007). (B) Modeling of Hfq structures. The structure of Hfq_Bb (gray) was predicted using Phyre and superimposed onto the experimentally determined S. aureus Hfq (blue) and E. coli Hfq (cyan) structures using Chimera. Hfq homologs from many organisms have a variable C-terminal extension that has not been structurally characterized; therefore, we have no information on the folded structure of the C-terminal region of Hfq_Bb and it is not shown.

B. burgdorferi bb0268 complements an E. coli hfq mutant

In E. coli, Hfq is required for DsrA to stimulate translation of rpoS in response to a decrease in temperature and entrance into stationary phase (Sledjeski et al., 2001). Evidence suggests that Hfq colocalizes rpoS and DsrA, presenting the two RNAs in favorable positions for complex formation (Brescia et al., 2003; Lease and Woodson, 2004). To determine if BB0268 functions as an RNA chaperone, similar to Hfq in E. coli, we assayed its ability to trans-complement an hfq null mutant of E. coli (Sledjeski et al., 2001). The bb0268 gene was fused to an arabinose-inducible promoter in the vector pBAD24. The empty pBAD24 vector and conserved hypothetical ORF bb0260, which shares sequence similarity with bb0268, were used as negative controls; E. coli hfq was used as a positive control. Hfq is needed for growth-phase related regulation of rpoS::lacZ at 30°C in E. coli. The efficient translation of an rpoS::lacZ fusion, in response to changes in cell density, was measured with a β-galactosidase assay. bb0268 partially restored the efficient translation of rpoS::lacZ at 30°C (Fig. 3), suggesting that it has an Hfq-like RNA chaperone activity in E. coli. In contrast, neither bb0260 nor the empty vector complemented the hfq null mutant in E. coli. Five independent experiments with bb0268 yielded similar results. Therefore, we now refer to BB0268 as Hfq_Bb.

Fig. 3.

BB0268 has Hfq-like activity in E. coli. hfq-1 mutant strains containing an rpoS::lacZ translational fusion and carrying the indicated plasmids were assayed for total β-galactosidase activity. pBADHfq expresses E. coli Hfq under control of the araBAD promoter. pBAD24 is the empty araBAD vector. pBAD268 and pBAD260 have the B. burgdorferi bb0268 and bb0260 ORFs under the control of the araBAD promoter. Cultures were grown at 30 C either with (black bars) or without (gray bars) 150 μM arabinose and samples were taken at various times during logarithmic growth. Total β-galactosidase activity was determined as described by Miller (1972). Three independent experiments with BB0260 and five independent experiments with BB0268 were performed and yielded similar results.

Recombinant Hfq_Bb binds rpoS mRNA and DsrA

The Hfq and Sm/Lsm family of proteins form homohexameric and heteroheptameric oligomers (Tsui et al., 1994; Sun et al., 2002; Valentin-Hansen et al., 2004). We purified recombinant Hfq_Bb using an N-terminal His-tag. The predicted molecular mass of Hfq_Bb is ~20 kDa (with the His-tag) and gel filtration chromatography indicated that it has a molecular mass of ~34 kDa in solution (Fig. 4), suggesting that Hfq_Bb forms a dimer or a monomer, rather than the characteristic hexamer, under these in vitro conditions. In contrast, E. coli Hfq was primarily a hexamer as analyzed by gel filtration chromatography under the same conditions, although monomer was also observed in this experiment (Fig. 4C).

Fig. 4.

Hfq_Bb is not a stable hexamer in solution. (A) Standard curve of the Sigma gel filtration molecular weight markers resolved on a Superdex 75 HR 10/30 column. Regression analysis yielded the equation: y = 869.58 × e^(−2.1536x), which was used to calculate molecular mass. Recombinant Hfq_Bb (B) and E. coli Hfq (C) were resolved on a Superdex 75 HR 10/30 column and fractions analyzed using a HisProbe-HRP.

Electrophoretic mobility shift assays (EMSAs) were performed using recombinant Hfq_Bb protein and in vitro-transcribed rpoS and DsrA_Bb RNAs (Fig. 5). The RNAs were heated to 65°C for 10 min and slow-cooled to room temperature. The RNAs were then incubated with increasing amounts of Hfq_Bb protein. The Hfq_Bb protein bound both rpoS and DsrA_Bb as seen by the decreased migration of the RNAs in the EMSA (Fig. 5).

Fig. 5.

The Hfq_Bb protein binds RNA. Gel mobility shift analyses of Hfq_Bb binding to DsrA_Bb (A) and rpoS (B). The RNAs were labeled with [α-³²P]UTP during in vitro transcription, gel-purified, and incubated with increasing concentrations of purified recombinant Hfq_Bb. The binding reactions were fractionated on a native Novex^® Pre-Cast 6% DNA retardation gel.

Construction and complementation of the hfq_Bb mutant

To determine the function of Hfq_Bb in B. burgdorferi, we deleted the hfq_Bb gene in a low-passage infectious clone of strain 297 (Fig. 6A). Two different plasmids were constructed to complement the hfq_Bb mutation. The hfq_Bb (bb0268) and the E. coli hfq ORFs were fused to the inducible flac promoter and inserted into the shuttle vector pKFSS1-lacI, which carries the lacI gene fused to the constitutive flgB promoter (Fig. 6B). The two complementing plasmids and the parental shuttle vector were transformed into the hfq_Bb mutant strain. RT-PCR analysis demonstrated that the wild-type and hfq_Bb-complemented strains, but not the hfq_Bb mutant strain, expressed hfq_Bb transcript (Fig. 6C).

Fig. 6.

Mutation and complementation of hfq_Bb. (A) Strategy for disrupting the hfq_Bb gene (bb0268). The hfq_Bb gene was replaced with the kanamycin resistant cassette (flgBp-aphI). Transformation into wild-type (WT) B. burgdorferi and homologous recombination generated an hfq_Bb mutant strain (hfq⁻). (B) Plasmid for trans-complementing the hfq_Bb mutant strain. The B. burgdorferi or E. coli hfq genes were fused to the inducible promoter flacp and inserted into the modified shuttle vector pKFSS1-lacI, which carries the lacI repressor gene fused to the constitutive flgB promoter in the multiple cloning site. The plasmid also carries both a B. burgdorferi replication origin (cp9 rep) from the 9-kb circular plasmid and an E. coli replication origin (ColE1) from pCR^®-XL-TOPO. (C) RT-PCR of hfq mRNA in the wild type (WT), the hfq_Bb mutant (hfq⁻), and the hfq_Bb mutant complemented with hfq_Bb (hfq⁻ pKFSS1-lacI flacphfq_Bb). The RT-PCR (+RT), the PCR on samples that were not treated with reverse transcriptase (−RT), and the PCR without template (−) were fractionated on a 2% agarose gel and stained with ethidium bromide.

The hfq_Bb mutant has a decreased growth rate and increased cell length

We assayed growth rate using a wild-type low-passage infectious clone of 297, the hfq_Bb mutant, and the two complemented strains. B. burgdorferi cultures were grown at 23°C to a low cell density (1–3 × 10⁷ cells ml⁻¹) and then inoculated into a culture at 37°C at a cell density of 1 × 10³ cells ml⁻¹. Cell densities were determined daily until cultures reached stationary phase. The hfq_Bb mutant grew slower than the wild type; both the B. burgdorferi and E. coli hfq genes complemented the growth phenotype of the hfq_Bb mutant (Fig. 7). The hfq_Bb mutant strain also was significantly longer than wild type. The cell length phenotype was fully complemented by the hfq_Bb gene expressed in trans and, somewhat unexpectedly, partially complemented by the E. coli hfq gene (Fig. 8).

Fig. 7.

Growth phenotype of the hfq_Bb mutant. Growth curve of the wild type (WT), hfq_Bb mutant (hfq⁻), and hfq_Bb mutant trans-complemented with either hfq from B. burgdorferi (hfq⁻pKFSS1-lacI flacphfq_Bb) or hfq from E. coli (hfq⁻ pKFSS1-lacI flacphfq_Ec) following a temperature shift to 37°C. The means of the cell densities from three independent experiments are graphed with error bars representing the standard error of the means.

Fig. 8.

Cell morphology phenotype of the hfq_Bb mutant. Cultures were grown at 37°C to mid-log phase and the cells were visualized using Nomarski microscopy and analyzed using ImageJ. (A) Microscopic images of the wild-type (WT), hfq_Bb mutant (hfq⁻), and complementing strains (pKFSS1-lacI/flacphfq_{Bb or Ec}). (B) ImageJ was used to measure the length of fifty spirochetes per strain and the means of the lengths are graphed with error bars representing the standard error of the means. Differences between wild-type and hfq_Bb mutant strains and between hfq_Bb mutant and pKFSS1-lacI/flacphfq_Bb-complemented strains are statistically significant (P < 0.01) by both a one-way ANOVA with a Tukey’s post-hoc test and a Kruskal–Wallis test; differences between hfq_Bb mutant and pKFSS1-lacI/flacphfq_Ec-complemented strains (P < 0.01) and between wild-type and pKFSS1-lacI/flacphfq_Ec-complemented strains (P < 0.001) are significant by both a two-tailed unpaired t-test and a Mann Whitney U test.

RpoS and OspC are regulated by Hfq_Bb

DsrA_Bb post-transcriptionally regulates RpoS levels in response to increased temperature at low cell density (Lybecker and Samuels, 2007). This sRNA has extensive complementarity to the upstream region of the rpoS mRNA and our data suggest that DsrA stimulates efficient translation of the rpoS mRNA through base pairing with the upstream region of the mRNA. Hfq in other bacteria is required for regulating trans-encoded sRNAs and often post-transcriptionally regulates RpoS levels, so we expected that the hfq_Bb mutant strain would show reduced levels of RpoS and OspC proteins in response to a temperature shift at a low cell density, as seen in the dsrA_Bb mutant strains. B. burgdorferi low-passage wild-type, hfq_Bb mutant, and complemented strains were temperature-shifted to 37°C and grown to low and high cell densities (1–3 × 10⁷ and 1× 10⁸ cells ml⁻¹, respectively). Western blot analyses of whole cell lysates demonstrated that the wild-type and complemented strains displayed an increase in RpoS and OspC levels after a temperature shift, while the hfq_Bb mutant strain showed reduced levels of RpoS and OspC (Fig. 9A). Surprisingly, the E. coli hfq also restored the temperature-dependent increase in RpoS and OspC. The inducible promoter fusions are leaky in these trans-acting configurations (Fig. 6C and data not shown), so IPTG was not required for transcription of the complementing hfq genes. Adding high levels of IPTG caused decreases in RpoS and OspC levels, presumably due to the overabundance of Hfq inhibiting translation (data not shown).

Fig. 9.

Hfq_Bb regulates RpoS levels post-transcriptionally in response to a temperature shift. Immunoblot analyses of whole-cell lysates of wild-type, hfq_Bb mutant, and trans-complemented strains at 23°C and after a temperature shift to 37°C at low (L) and high (H) cell densities (1–3 × 10⁷ and 1 × 10⁸ cells ml⁻¹, respectively(A). The top panel is a Coomassie brilliant blue-stained SDS-PAGE gel and the bottom panels are immunoblots probed with either anti-RpoS or anti-OspC antibodies. Three independent experiments were performed and representative data are shown. Northern blot analyses of total RNA (15 μg) fractionated on a 6% polyacrylamide 7 M urea gel, blotted to a nylon membrane, and hybridized with a rpoS, rpoN, rrp2, or flaA single-stranded RNA probe (B). Two independent experiments were performed and representative data are shown; the signal was quantified on a FujiFilm LAS-3000.

Transcription of rpoS requires RpoN and Rrp2 (Hübner et al., 2001; Yang et al., 2003). To further understand the mechanism by which Hfq_Bb regulates RpoS, the steady-state levels of rpoS, rpoN, and rrp2 mRNA were analyzed by Northern blot analyses (Fig. 9B). The constitutively expressed flaA transcript was used as a gel loading control. The amounts of the rpoN and rrp2 transcripts were similar between the wild-type, mutant and complemented strains demonstrating that Hfq does not regulate the levels of rpoN or rrp2 mRNA. However, the levels of rpoS transcript were 1.5-fold higher in the hfq_Bb mutant strain compared to the wild-type and complemented strains; this pattern of regulation was also observed in the dsrA_Bb mutant strain (Lybecker and Samuels, 2007). These data, taken together, suggest that Hfq_Bb regulates RpoS synthesis post-transcriptionally and does not stabilize the rpoS mRNA.

As previously seen with the dsrA_Bb mutant (Lybecker and Samuels, 2007), serial passage of the hfq_Bb mutant results in a reversal of the phenotype. B. burgdorferi cultures were serially passaged at 23°C and then temperature-shifted to 37°C and grown to a low cell density. Western blot analyses demonstrated that both RpoS and OspC syntheses were increased, rather than decreased, in the mutant compared to the wild type after as few as six passages (data not shown). The mechanism by which the phenotype switches is not yet understood.

Hfq_Bb is required for mouse infection by needle inoculation

Several bacterial species require Hfq for virulence (Chao and Vogel, 2010). We hypothesized that the hfq_Bb mutant strain would not be infectious in mice due to the absence of RpoS and OspC. We assayed the infectivity of the hfq_Bb mutant in the mouse model of Lyme disease by needle inoculation (Barthold et al., 2010). B. burgdorferi cultures were temperature-shifted to 37°C and grown to a low cell density (1–3 × 10⁷ cells ml⁻¹); C3H-HeJ female mice were needle inoculated with 5 × 10³ cells i.p. Tissue samples were cultured and assayed for the presence of B. burgdorferi by dark-field microscopy; cultures that were negative by this method were also screened by PCR (Wang et al., 2010). The hfq_Bb mutant was unable to infect mice and complementation with hfq from B. burgdorferi, but not E. coli, restored infectivity (Table 1). Thus, hfq_Bb is essential for B. burgdorferi infection of mice by needle inoculation.

Table 1.

Hfq_Bb is required for mouse infection by needle inoculation

	# of positive mice by culture
B. burgdorferi strain	Ear	Joint	Bladder
297 (BbAH130) WT	6/6	6/6	6/6
297 hfq−	0/6	0/6	0/6
297 hfq− high passage	0/3	0/3	0/3
297 hfq−/flacp-hfqBb	3/3	3/3	3/3
297 hfq−/flacp-hfqEc	0/3	0/3	0/3

The high-passage hfq_Bb mutant retained the plasmids essential for infectivity and was used to needle inoculate mice. RpoS and OspC are highly expressed in the high-passage hfq_Bb mutant, so we expected it to be infectious, but the strain did not infect the mice (Table 1). These data and the finding that the hfq_Bb mutant complemented with hfq from E. coli was avirulent, even though it expressed wild-type levels of RpoS, suggest Hfq_Bb may play additional roles in mammalian infection besides regulation of RpoS.

Discussion

The RNA chaperone Hfq has emerged as an important global post-transcriptional regulator in mRNA translational control by sRNAs (Valentin-Hansen et al., 2004). In this study, we identified and characterized an unusual Hfq in B. burgdorferi. The Hfq_Bb primary sequence does not have a high degree of similarity or identity to annotated Hfq homologs, but does retain a few key residues and was predicted to have a superimposable tertiary structure (Fig. 2). A BLAST search using full-length Hfq_Bb, or the more highly conserved N-terminal region of Hfq_Bb, as the query only identified hypothetical open reading frames within the spirochete phylum (data not shown). Hfq is characterized as a part of the Sm/Lsm protein family and these proteins form ring-shaped oligomeric structures that are homohexameric in bacteria and heteroheptameric in eukaryotes. Despite these differences, the Hfq and Sm/Lsm monomers fold into a similar structure, consisting of an N-terminal α-helix followed by a curved five-stranded β-sheet (Schumacher et al., 2002; Sauter et al., 2003). Our gel filtration data suggested that recombinant Hfq_Bb does not form a stable hexamer in solution (Fig. 4), which is unlike most other characterized Hfq homologs. However, we hypothesize that the functional quaternary structure of Hfq_Bb in vivo is a homohexamer, which is unstable under our experimental conditions. Possibly, our inability to observe a hexamer in vitro is due to the presence of a lysine residue rather than the conserved histidine at position 57 (in the E. coli sequence), which is known to be involved in stabilizing the hexameric structure (Moskaleva et al., 2010). Despite the differences in primary and quaternary structure between Hfq_Bb and the canonical Hfq, our results clearly establish that Hfq_Bb functions as an RNA chaperone as demonstrated by hfq_Bb complementing an rpoS translational defect in an hfq null mutant of E. coli (Fig. 3). Moreover, E. coli hfq complemented the hfq_Bb mutant phenotype with regard to growth rate (Fig. 7), cell length (Fig. 8), and RpoS and OspC induction (Fig. 9). This is the first time, to our knowledge, that a heterologous bacterial gene has complemented a B. burgdorferi mutant.

The hfq_Bb mutant has a phenotype that includes decreased growth rate, increased cell length, altered RpoS regulation, and loss of infectivity. The B. burgdorferi and E. coli hfq genes complemented the in vitro mutant phenotype demonstrating that the phenotype is due to the loss of hfq_Bb; these phenotypes have been observed in several hfq mutants of other bacteria (Brennan and Link, 2007; Chao and Vogel, 2010). RpoS is required for mammalian infection by B. burgdorferi, and, therefore, the hfq_Bb mutant was not expected to be infectious in the mouse model. Our findings established that Hfq_Bb is required for mammalian infectivity, but that the mechanism is not solely due to the loss of RpoS induction. A high-passage hfq_Bb mutant strain that produces RpoS and OspC was still not infectious in mice (Table 1). These data demonstrated that Hfq_Bb is required for infectivity in mice independent of RpoS and OspC expression. We hypothesize that there are other Hfq-dependent sRNAs in B. burgdorferi that regulate expression of virulence factors. In E. coli and Salmonella typhimurium, Hfq has other regulatory functions independent of its effects on RpoS expression (Muffler et al., 1997; Sittka et al., 2007). Hfq in the pathogenic bacteria S. typhimurium, Vibrio cholerae, and Brucella abortus is required for full virulence, but the mechanism is independent of the requirement of Hfq for RpoS expression (Robertson and Roop, 1999; Ding et al., 2004; Sittka et al., 2007).

We hypothesized that Hfq_Bb regulates RpoS synthesis via the sRNA DsrA_Bb and, thus, we expected the RpoS defect would be similar in the dsrA_Bb mutant and hfq_Bb mutant strains. Our data demonstrated that Hfq_Bb post-transcriptionally regulates the temperature-dependent increase in RpoS. Surprisingly, unlike the dsrA_Bb mutant strain, which only shows reduced levels of RpoS at a low cell density (log-phase growth), the hfq_Bb mutant strain had reduced levels of RpoS at both low and high cell density (log and stationary growth phase, respectively), suggesting that Hfq_Bb also plays a DsrA_Bb-independent role in RpoS regulation.

In conclusion, we have identified and characterized an unusual Hfq protein in the Lyme disease spirochete B. burgdorferi that is essential for infectivity. Although only one sRNA has been identified to date in B. burgdorferi, the pleiotropic phenotype of the hfq_Bb mutant suggests other sRNAs exist and likely play a role in the enzootic life cycle.

Experimental procedures

B. burgdorferi strains and culture conditions

B. burgdorferi low-passage virulent strain 297 (clone BbAH130) (Hübner et al., 2001) and all derivatives used in this study were cultivated in Barbour-Stoenner-Kelly II (BSK-II) complete medium (Barbour, 1984) at either 23°C or 37°C. Cultures were passaged to 1 × 10⁵ cells ml⁻¹ from 23°C to 37°C and grown to either 1–3 × 10⁷ cells ml⁻¹ (low density), which corresponds to mid-log phase, or about 1 × 10⁸ cells ml⁻¹ (high density), which corresponds to stationary phase. For growth curves, cultures were grown at 23°C to 1–3 × 10⁷ cells ml⁻¹ and passaged to 1 × 10³ cells ml⁻¹ at 37°C. Cell density was determined using a Petroff Hausser counting chamber (Hausser Scientific Partnership) as previously described (Samuels, 1995). All cultures were counted in triplicate and at least three independent experiments were performed.

Sequence analysis and structural modeling

The ClustalW sequence alignment was performed with NPS@: Network Protein Sequence Analysis (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_clustalw.html) (Combet et al., 2000). A computational prediction of the Hfq_Bb fold was performed using Protein Homology/analogY Recognition Engine (Phyre) (http://www.sbg.bio.ic.ac.uk/~phyre/) (Kelley and Sternberg, 2009). The resulting PDB coordinate file covering residues 6–71 was then visualized and overlaid with the structures of Hfq from E. coli (PDB 1HK9) (Sauter et al., 2003) and Hfq from S. aureus (PDB 1KQ2) (Schumacher et al., 2002) using Chimera (http://www.cgl.ucsf.edu/chimera/) (Pettersen et al., 2004).

E. coli strains and plasmid construction for hfq complementation

E. coli strain DDS1631 has an hfq null mutation and an rpoS::lacZ translational fusion (Sledjeski et al., 2001). The pBADHfq plasmid (pDDS400) carries the E. coli hfq gene cloned into the EcoRI and PstI sites of the arabinose-inducible vector pBAD24 (Sledjeski et al., 2001). The bb0268 and bb0260 genes were PCR-amplified using the proofreading enzyme KOD with engineered MfeI and PstI sites at the 5′ and 3′ ends, respectively (Table 2). After poly(A) tailing with Taq, the PCR products were cloned into pCR^®2.1-TOPO. Positive transformants were screened by PCR and confirmed by DNA sequencing. The pBAD260 and pBAD268 plasmids were generated by directionally cloning the MfeI-PstI excised bb0268 and bb0260 fragments into the EcoRI and PstI sites of pBAD24.

Table 2.

Oligonucleotides used.

Name	Sequence (5′-3′)	Function
BB268 1F+MfeI	ACAATTGACCATGTTTATAAGCAGGGAATTG	Complementation in E. coli
BB268 480R+PstI	ACTGCAGTTATTCCTTCTTGCTCATTAAAG
BB260 1F+MfeI	ACAATTGACCATGAAAGGGTTTTTAGCG
BB260 1014R+PstI	ACTGCAGTTAAACATATTGATCTTTTTC
BB268 1F+NdeI	CATATGTTTATAAGCAGGGAATTG	Complementation, expression, and RT-PCR
BB268 480R+AatII	GACGTCTTATTCCTTCTTGCTCATTAAAG
BB268 249R+AatII	GACGTCTTAACTCTTCTTATCATCTTG	RT-PCR
Echfq 1F+NdeI	CATATGGCTAAGGGGCAATCTTTAC	Complementation in B. burgdorferi
Echfq 310R+AatII	GACGTCTTATTCGGTTTCTTCGCTGTC
BB268 U1048F	TTTCTTATGTTACGGATGGGC	Mutant construction
BB268 U4R+Aat+Age	ACCGGTCATGACGTCTTATTCCACACCAAAAAATC
BB268 474F+AatII	GACGTCGGAATAAATATGGATGACAGGGC
BB268 D1578+AgeI	ACCGGTTACTATATCTCCAGTAGCAGGTCC
dsrA L1F+T7	TAATACGACTCACTATAGGGAAAGCTAAATTAGAAGAGTTTATTG	Template for in vitro transcription
dsrA L275/352R	CATAAAAACTTTTTTTTGAATAG
rpoS U449F+T7	TAATACGACTCACTATAGGGTCACAAAATCTAAAATTTAAAAATC
rpoS 128R	GTTTTTTGCTTTTGCATTGC
rpoS 705R+T7	TAATACGACTCACTATAGGCGGTGCTTTTTTTGGGACTATTG	Northern blot probes
rpoS 443F	GATTCAACCTATCTCCTGCTCAG
rrp2 185F	GAGAAAAATTGCTCAAAATA
rrp2 394R+T7	TAATACGACTCACTATAGGCTTTTGAGTTTTTTCAAGAAG
rpoN 285F	GCAATTAAGAATTCAAAGAA
rpoN 572R+T7	TAATACGACTCACTATAGGCTTTTGAGTTTTTTCAAGAAG
flaA 64F	GCTCAAGAGACTGATGGATTAGC
flaA 284R+T7	TAATACGACTCACTATAGGCGCAGAAGGAGTAAGTAAAACGCTC

E. coli Hfq complementation

β-galactosidase units were assayed using the Zhang and Bremer (1995) modification of the Miller assay (1972). Briefly, cells were grown with shaking at 30°C in lysogeny broth medium supplied with kanamycin and carbomycin in the presence or absence of 150 μM arabinose. The cultures were inoculated 1:100 from overnight cultures and samples were taken at an OD₆₀₀ ranging from 0.4 to 1.0 (approximately log phase). The OD₆₀₀ of each culture was measured and 30μl aliquots were removed and mixed with 70 μl of permeabilization solution [100 mM Na₂HPO₄, 20 mM KCl, 2 mM MgSO₄, 0.8 mg ml⁻¹ CTAB (hexadecyltrimethylammonium), 0.4 mg ml⁻¹ sodium deoxycholate and 5.4 μl ml⁻¹ 2- mercaptoethanol] and stored at room temperature. The samples were incubated at 30°C for 30 min. The substrate solution [60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 1 mg ml⁻¹ o-nitrophenyl-β-D-galactoside (ONPG) and 2.7 μl ml⁻¹ 2-mercaptoethanol] was also incubated at 30°C for 30 min and 600 μl of substrate solution was added to each sample. After sufficient color developed, 700 μl of stop solution (1 M Na₂CO₃) was added to each sample and the time was recorded. The samples were centrifuged at 14,000 rpm in an Eppendorf 5417R for 5 min to remove cell particulates. The OD₄₂₀ (between 0.05 and 1.0) was measured for each sample and the total β-galactosidase units were calculated from the formula:

$$1000\times ({\text{OD}}_{420})÷[({\text{OD}}_{600}\text{of}\text{culture})\times (\text{volume}\text{of}\text{culture})\times (\text{reaction}\text{time})]$$

The volumes of permeabilization solution and cell culture varied between experiments and were appropriately changed in the calculation of total β-galactosidase units.

Cloning and protein purification of Hfq

The hfq_Bb gene was amplified by PCR with the proofreading enzyme KOD, poly(A) tailed, and cloned into pCR^®2.1-TOPO (Table 2). The hfq_Bb gene was directionally cloned into a derivative of pET28b with an N-terminal, tobacco etch virus protease-cleavable His₆ tag (Corbett et al., 2004) using an engineered NdeI site at the 5′ end of the gene and the pCR^®2. 1-TOPO vector BamHI site. The resulting plasmid pET28b-TEV::Hfq was sequenced and transformed into E. coli Rosetta (DE3) cells (Novagen). Cell cultures were grown at 37°C to an OD₆₀₀ of 0.4–0.6, induced with 1 mM of IPTG and grown overnight at 11°C and again overnight at 18°C. Cells were lysed using Bugbuster (Novagen), filtered through a 0.45 μm filter and loaded directly onto 1 ml Ni²⁺ column (HisTrap HP, Amersham Biosciences). The resin was washed with ten column volumes of 20 mM NaH₂PO₄, 500 mM NaCl and 5 mM imidazole, pH 7.4, five column volumes of 1M NH₄Cl, and five column volumes of 1M urea, and finally eluted with a 10-ml linear gradient of 10 to 500 mM imidazole, pH 7.4, in 20 mM NaH₂PO₄ and 500 mM NaCl. The fractions containing Hfq were dialyzed into a storage buffer (20 mM NaH₂PO₄ and 125 mM NaCl, pH 7.4) and concentrated in Amicon Ultra-4 spin columns. The concentration of the purified recombinant protein was determined with the BCA protein assay (Pierce). The E. coli Hfq protein was purified as previously described (Mikulecky et al., 2004).

Gel filtration

Recombinant purified Hfq_Bb and E. coli Hfq were dialyzed overnight in 50 mM Tris-HCl, pH 7.5 and 100 mM KCl. Sigma gel filtration molecular weight markers were run on a Superdex 75 HR 10/30 column and a standard curve was plotted. The Hfq proteins were run on the column and the molecular mass was calculated using the standard curve.

Template preparation for in vitro transcription

B. burgdorferi strain 297 total DNA was used as a template in PCR with the T7 bacteriophage promoter added to the 5′ end of either the forward primer to generate sense RNAs for the EMSA or the reverse primer to generate antisense RNAs for probes in Northern blots (Table 2). The PCR products were visualized on an agarose gel stained with ethidium bromide. If a single product was visualized, the PCR product was precipitated with ammonium acetate and resuspended in DEPC-treated water. If more than one product was visible, the appropriate band was purified with the Qiagen gel purification kit following the manufacturer’s protocol and resuspended in DEPC-treated water.

In vitro synthesized RNA

DsrA_Bb and rpoS RNAs for EMSAs were synthesized and radioactively labeled by incorporation of [α-³²P]UTP during in vitro transcription using the T7-Megascript In Vitro Transcription kit (Ambion) according to the protocol suggested by the manufacturer. PCR products (described above) were used as templates for the in vitro transcription reactions. Reaction time was increased to overnight at 37°C to increase the yield of short products. Following in vitro transcription, the full-length RNA transcripts were resolved on a 6% polyacrylamide-7M urea gel and eluted from gel slices by incubating in elution buffer (20 mM Tris-HCL pH 7.5, 0.5 M ammonium acetate, 10 mM EDTA, 0.1% SDS) at 37°C overnight or 65°C for 2 h. The RNAs were purified and concentrated by an ammonium acetate precipitation and resuspended in 60 μl of DEPC-treated water. OD₂₆₀ was used to determine the concentration of RNAs.

Electrophoretic mobility shift assay (EMSA)

The purified in vitro synthesized DsrA_Bb and rpoS RNAs were incubated with increasing amounts of purified recombinant Hfq_Bb in binding buffer (10 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 50 mM NaCl, and 50 mM KCl). 0.17 pmoles of DsrA_Bb or rpoS RNA was mixed with 7 μl of the binding buffer, heated to 65°C for 10 min, and slowly cooled to room temperature. Two μl of serially diluted Hfq_Bb was added to the RNA and incubated for 5 min at room temperature. The samples were mixed with 10 μl gel loading buffer II (Ambion), fractionated on a Novex^® Pre-Cast 6% DNA retardation gel (Invitrogen) in 1X TBE at 100V for 2 h. The gel was dried, exposed to a phosphorimager screen, and visualized using a Fujifilm FLA-3000G PhosphorImager.

Gene disruption and complementation of hfq_Bb

The hfq_Bb (bb0268) gene was deleted and replaced with a kanamycin resistance cassette (Bono et al., 2000). Regions flanking the hfq_Bb ORF were amplified by PCR (Table 2). The 3′ end of the upstream flanking sequence and the 5′ end of the downstream flanking sequence were engineered with AatII sites for insertion of the kanamycin resistance cassette. The two flanking regions were cloned into pCR^®2. 1-TOPO and ligated together to form a 2.1 kb target DNA fragment. The antibiotic resistance cassette was cloned into the synthetic AatII site. The plasmid was linearized with AhdI and electroporated into competent B. burgdorferi as previously described (Samuels, 1995). B. burgdorferi transformants were cloned in liquid BSK-II medium (Yang et al., 2004) containing kanamycin (200 μg ml⁻¹) at 34°C and a 1.5% CO₂ atmosphere. Transformants were screened by PCR with primers that flank the insertion site.

Two different plasmids were constructed to complement the hfq_Bb mutant strain. The E. coli or B. burgdorferi hfq genes were fused to the inducible flac promoter (Gilbert et al., 2007) and inserted into a modified pKFSS1 vector (Frank et al., 2003) carrying the lacI repressor gene fused to the constitutive flgB promoter in the multiple cloning site. The recombinant plasmids were electroporated into the kanamycin-resistant hfq_Bb mutant and selected in both streptomycin and kanamycin. The resulting transformants were assayed for the hfq_Bb expression plasmid by both PCR and xenodiagnosis (transformation of plasmid DNA from B. burgdorferi into E. coli).

RT-PCR

Total RNA was isolated from 100-ml cultures using TRIzol^™ (Gibco BRL) as previously described (Lybecker and Samuels, 2007). RNA samples were treated with TurboDNase (Ambion) and RT-PCR was performed using the RETROscript^™ kit (Ambion) according to manufacturer’s instructions, except a gene-specific primer was used in the reverse transcription reaction instead of random decamers (Table 2).

Microscopic analysis of B. burgdorferi

Cell cultures were grown at 37°C to a low cell density (1–3× 10⁷ cells ml⁻¹) and centrifuged at 5800 × g. Cell pellets were resuspended in 100 μl of water. Samples were visualized by differential interference contrast (DIC) microscopy using a Nikon E800 microscope at 100× . B. burgdorferi length was measured using ImageJ software. The length of hfq_Bb mutant cells do not display a normal distribution, so significance was evaluated using both parametric (one-way ANOVA with a Tukey’s post-hoc test and a two-tailed unpaired t-test) and non-parametric (Kruskal–Wallis one-way ANOVA and a Mann Whitney U test) methods.

SDS-PAGE and immunoblotting

B. burgdorferi protein extracts were prepared as previously described (Lybecker and Samuels, 2007). Protein extracts were fractionated in Novex^® Pre-Cast 4–20% Tris-Glycine gels (Invitrogen) and transferred to Immobilon-P PVDF membranes (Millipore) by electroblotting. Proteins were detected with either an anti-RpoS antiserum (Yang et al., 2000) or anti-OspC antiserum (Yang et al., 2005). The membranes were developed by chemiluminescence using the ECL Plus Western Blotting detection system (Amersham Biosciences) on a Fujifilm LAS-3000 (Yang et al., 2005; Gilbert et al., 2007).

Northern blot analysis

Total RNA was isolated from 100-ml cultures using TRIzol^™ (Gibco BRL) as previously described (Lybecker and Samuels, 2007). RNA samples were treated with TurboDNase (Ambion) and fractionated in a Novex^® Pre-Cast 6% polyacrylamide 7 M urea gel (Invitrogen) according to the manufacturers’ instructions. The RNA was transferred to a BrightStar-Plus^™ membrane by electroblotting using the Xcell SureLock^™ Mini-Cell (Invitrogen) following the manufacturer’s specifications. Northern blot hybridizations were performed using the NorthernMax^™ kit (Ambion) according to the manufacturer’s instructions. Non-isotopic RNA probes were generated from in vitro transcription of PCR-generated DNA templates using the MAXIscript^® in vitro transcription kit (Ambion). Biotin-16-UTP (Roche) was used in the in vitro transcription reaction. Northern blot membranes were developed by chemiluminescence using the BrightStar^® Biodetect^™ nonisotopic detection kit on a Fujifilm LAS-3000.

Mouse infectivity studies

C3H-HeJ female mice were needle inoculated intraperitoneally (i.p) with 5 × 10³ cells of wild-type 297 (BbAh130), hfq_Bb mutant, or complemented strains. All strains used in mouse infectivity studies were screened for the presence of plasmids essential for infectivity (lp28-1, lp25, and lp54) (Purser and Norris, 2000; Labandeira-Rey and Skare, 2001). Three weeks post injection, ear punches were taken and cultured in BSK-II. Five weeks post injection, mice were sacrificed and ear punches, tibiotarsal joints and bladders were collected and cultivated in BSK-II. Cultures were screened for the presence and growth of B. burgdorferi by darkfield microscopy. In addition, PCR was used to screen the cultures that appeared negative by dark-field microscopy. The University of Montana Institutional Animal Care and Use Committee approved all experimental procedures involving mice (protocol number 038-08SSDBS-081208).