PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
 
J Bacteriol. 2013 June; 195(11): 2662–2674.
PMCID: PMC3676067

The Bartonella quintana Extracytoplasmic Function Sigma Factor RpoE Has a Role in Bacterial Adaptation to the Arthropod Vector Environment

Abstract

Bartonella quintana is a vector-borne bacterial pathogen that causes fatal disease in humans. During the infectious cycle, B. quintana transitions from the hemin-restricted human bloodstream to the hemin-rich body louse vector. Because extracytoplasmic function (ECF) sigma factors often regulate adaptation to environmental changes, we hypothesized that a previously unstudied B. quintana ECF sigma factor, RpoE, is involved in the transition from the human host to the body louse vector. The genomic context of B. quintana rpoE identified it as a member of the ECF15 family of sigma factors found only in alphaproteobacteria. ECF15 sigma factors are believed to be the master regulators of the general stress response in alphaproteobacteria. In this study, we examined the B. quintana RpoE response to two stressors that are encountered in the body louse vector environment, a decreased temperature and an increased hemin concentration. We determined that the expression of rpoE is significantly upregulated at the body louse (28°C) versus the human host (37°C) temperature. rpoE expression also was upregulated when B. quintana was exposed to high hemin concentrations. In vitro and in vivo analyses demonstrated that RpoE function is regulated by a mechanism involving the anti-sigma factor NepR and the response regulator PhyR. The ΔrpoE ΔnepR mutant strain of B. quintana established that RpoE-mediated transcription is important in mediating the tolerance of B. quintana to high hemin concentrations. We present the first analysis of an ECF15 sigma factor in a vector-borne human pathogen and conclude that RpoE has a role in the adaptation of B. quintana to the hemin-rich arthropod vector environment.

INTRODUCTION

The Gram-negative bacterial pathogen Bartonella quintana was first identified during World War I as the causative agent of trench fever, a 5-day relapsing fever (1). In the last 2 decades, there has been a resurgence of B. quintana infections, with the most severe illness occurring among immunocompromised individuals (2). B. quintana infection can cause relapsing fever, endocarditis, and vascular proliferative lesions (3). Vasculoproliferative B. quintana infection is usually progressive and can be fatal unless correctly diagnosed and treated with antibiotic therapy (4). Another manifestation of B. quintana infection is persistent bloodstream infection, which occurs in both immunocompromised and immunocompetent people (5, 6).

Bartonella is an arthropod vector-borne bacterium. The vector for B. quintana is the human body louse Pediculus humanus humanus. In a recent analysis, 33.3% of the body lice recovered from infested homeless individuals in California were PCR positive for B. quintana, underscoring the high prevalence of this potentially fatal bacterium in the human environment (7). B. quintana colonizes the louse alimentary tract and can attach to the apical surface of gut epithelial cells (8). The louse excretes B. quintana in its feces during feeding, and feces containing B. quintana are inoculated into the louse bite when the human scratches the bite. B. quintana forms a biofilm-like structure in the louse feces, allowing prolonged bacterial survival within the fecal environment (9).

During the infectious cycle, B. quintana alternates between two niches, the bloodstream of the human host (37°C) and the gut of the body louse vector (28°C) (10). To maintain the transmission cycle, B. quintana must survive and proliferate within these two different environments. The adaptive mechanisms utilized by Bartonella during the transition between the host and vector are unknown. In addition to temperature, a major environmental difference between the host and vector niches is the ambient hemin concentration; the bloodstream is severely hemin restricted, and the body louse gut is hemin rich. Hemin and hemoglobin are the only iron sources that Bartonella can utilize (11), making the acquisition and metabolism of these nutrients essential for Bartonella survival. However, hemin can produce reactive oxygen molecules that are potentially toxic (12). Bartonella is unique in its ability to survive exposure to hemin concentrations that are typically bactericidal (>1 mM) (11, 13, 14). For example, the growth of Staphylococcus aureus is severely limited in 10 to 20 μM hemin (15, 16) and the growth of Neisseria meningitidis is inhibited by ~0.2 mM hemin (17). Bartonella produces a family of hemin binding proteins (Hbp) that are responsive to temperature, hemin concentration, and oxidative stress (18, 19). Additionally, Bartonella contains a genomic locus that encodes hemin utilization (Hut) proteins involved in hemin sensing, degradation, and storage (20). It is believed that both Hbp and Hut proteins have a role in the unique ability of Bartonella to survive in a wide range of hemin concentrations (1821).

Many bacteria adapt to environmental changes by expressing different sigma factors. Bacterial sigma factors mediate shifts in gene expression by conferring promoter recognition specification on RNA polymerase (22). Thus, the expression of different sigma factors allows a bacterium to coordinate the transcription of specific sets of genes as a regulon. On the basis of sequence homology, the B. quintana genome encodes four sigma factors (23), none of which have been studied. The four B. quintana sigma factors belong to the σ70 family: RpoD (σ70, σD), RpoH1 (σ32, σH), RpoH2 (σ32, σF), and RpoE (σ24, σE). Bartonella RpoE was annotated as SigH in the original genome sequence (23); however, in an earlier publication it was referred to as RpoE (24). We have adopted the RpoE nomenclature because, in general, sigma factors from Gram-negative bacteria are given the designation rpo (for RNA polymerase subunit), whereas sigma factors from Gram-positive bacteria are typically given the sig designation (25).

The Bartonella alternative sigma factor RpoE is annotated as an extracytoplasmic function (ECF) sigma factor. ECF sigma factors typically sense and respond to changes in the extracellular environment, including oxidative stress, misfolded proteins, and changes in temperature, pressure, or nutrient concentration (26, 27). The ECF sigma factor family is the largest and most divergent of the bacterial sigma factor families (25). Over 40 distinct groups of ECF sigma factors have been identified (28). By sequence homology and genomic context, Bartonella RpoE appears to be a member of the sigma factor group designated ECF15, which is found only in alphaproteobacteria (28). The ECF15 sigma factor mediates the general stress response (GSR) in alphaproteobacteria (29). The GSR allows bacteria to respond to a wide variety of stress conditions in a nonspecific manner (29). The ECF15 group is characterized by the conserved genomic context of the sigma factor, with an anti-sigma factor and a unique response regulator found in close proximity (28). The involvement of the ECF15-associated response regulator in the alphaproteobacterial GSR was first described in 2006, for Methylobacterium extorquens (30). Members of this newly recognized ECF15 sigma factor group have been studied subsequently in the nonpathogenic alphaproteobacteria Bradyrhizobium (31), Sinorhizobium (32), Caulobacter (3335), and Sphingomonas (36, 37). In these organisms, the associated anti-sigma factor is named NepR, except for that of Sinorhizobium, which is called RsiA1 (32).

In the above-mentioned alphaproteobacteria, there is a response regulator found in close genomic proximity to the ECF15 sigma factor that functions as an anti-anti-sigma factor (28). The response regulator is named PhyR, corresponding to the name established in the initial studies in Methylobacterium (30, 38). PhyR response regulators contain an ECF sigma factor-like domain and function as an anti-anti-sigma factor by binding to the anti-sigma factor NepR, thus disrupting the NepR-RpoE interaction (33, 38). There is typically a sensor histidine kinase gene in close genomic proximity to the gene for either the ECF15 sigma factor or PhyR (29). It has been demonstrated in Caulobacter that the sensor histidine kinase (PhyK) phosphorylates the response regulator PhyR (35). Phosphorylation triggers the conformational change that allows PhyR to interact with NepR (33, 34). It also was shown that the putative histidine kinase in Sphingomonas, PhyP, is a phosphatase that functions to maintain PhyR in an unphosphorylated state (37).

The goals of this study were to identify the genomic components and functional interactions of the B. quintana ECF15 transcription regulatory system and to determine the role of RpoE-mediated transcription in the B. quintana infectious cycle. We demonstrated that the expression of the B. quintana sigma factor rpoE is significantly upregulated under low-temperature and high-hemin conditions that recapitulate the body louse environment. We further determined that RpoE-mediated transcription has a role in mediating B. quintana tolerance to toxic hemin concentrations. B. quintana RpoE is the first ECF15 group sigma factor found to be involved in adaptation to arthropod vector conditions; thus, our work has identified a novel adaptation of the alphaproteobacterial GSR ECF15 class of sigma factors. Additionally, our data provide the first detailed study of the role of an ECF15 sigma factor in a bacterium pathogenic to humans.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Wild-type B. quintana strain JK31 was isolated from a B. quintana-infected individual (39). Low-passage B. quintana strains were used for all experiments. B. quintana strains from frozen stock were streaked onto fresh chocolate agar plates (40) and grown for 6 to 7 days in candle extinction jars at 35°C prior to passage and use in experiments. The liquid Bartonella medium used in some experiments, M199S, consists of M199 medium supplemented with 20% fetal bovine serum, glutamine, and sodium pyruvate (40). Escherichia coli strains were grown at 37°C on Luria-Bertani medium. When required, kanamycin was used at a final concentration of 50 μg ml−1, ampicillin was used at 100 μg ml−1, cefazolin was used at 20 μg ml−1, and nalidixic acid was used at 2 μg ml−1. For sacB negative selection, a sterile filtered sucrose solution was added to chocolate agar to give a final concentration of 5%. The strains and plasmids used in this study are listed in Table 1.

Table 1
Bacterial strains and plasmids used in this study

Bartonella RNA, cDNA, and genomic DNA preparation.

For RNA isolation, B. quintana were harvested from confluent plates into 1 ml stop solution (M199, 45% ethanol, 5% water-saturated phenol) to prevent RNA degradation (41). Bacteria were then pelleted by centrifugation at 4,000 × g at 4°C. The bacterial pellet was stored at −80°C until RNA isolation. Bacterial cells were lysed by incubation in fresh lysozyme (0.4 mg ml−1) in 10 mM Tris and 1 mM EDTA for 5 min at room temperature. The RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA) in accordance with the manufacturer's instructions. Total RNA was RQ1 DNase (Promega, Madison, WI) treated for 2 h and then further purified with the RNeasy Mini Kit (Qiagen, Valencia, CA). A second RQ1 DNase treatment and RNA cleanup were performed to ensure complete DNA removal. cDNA was generated from 0.5 μg of total RNA with random hexamer primers (Invitrogen) and SuperScript III (Invitrogen) by following the manufacturer's instructions. Reverse transcription reactions without Superscript III were performed as negative controls and to evaluate DNase treatment efficiency. B. quintana genomic DNA was isolated with Qiagen Puregene Core Kit B by following the manufacturer's instructions.

Quantitative analysis of B. quintana gene expression.

The relative abundance of specific mRNA sequences was ascertained by reverse transcriptase quantitative PCR (RT-qPCR) with an MX3000P machine (Stratagene/Agilent Technologies, Santa Clara, CA). cDNA was diluted 1:10 for use in reaction mixtures. The reaction mixture included 10 μl of SYBR Fast qPCR master mix (Kapa Biosystems, Woburn, MA), 0.4 μl of ROX low (Kapa Biosystems), 7.6 μl of template, and 2 μl of 1 pmol μl−1 primer. The reaction conditions were 95°C for 10 min and 40 cycles of 95°C for 15 s and 60°C for 60 s, with the standard dissociation protocol. Threshold fluorescence was determined during the geometric phase of logarithmic gene amplification; from this the quantification cycle (Cq) was set. Standard curves for each primer set were generated by plotting log genomic DNA versus the Cq. These plots were used to ensure that equivalent reaction efficiency was obtained with all primer sets. The primers used are listed in Table S1 in the supplemental material. The relative levels of gene transcripts in samples were determined by converting the transcript level into a genomic copy number by using standard curves. This value was divided by the genomic copy number of the constitutively expressed B. quintana reference gene purA (adenylosuccinate synthetase) to obtain a relative level of transcription for each gene. Data from three independent experiments were used for statistical analysis by Student's t test and to determine average gene transcription values.

Mapping of the transcriptional start site (TSS) by 5′ random amplification of cDNA ends (RACE).

5′ RACE was performed as described by Frohman et al. (42). RNA was purified as described above. cDNA was generated from 5 μg of RNA, but gene-specific primers (GSP1) were used in place of random hexamers. All of the primers used for 5′ RACE are listed in Table S1 in the supplemental material. cDNA was treated with RNase H (Invitrogen) and then purified with Qiagen PCR purification columns. A poly(A) tail was added to the 3′ end of the purified cDNA with terminal transferase (Roche, Indianapolis, IN). Tailed cDNA was used as the template for PCR with the GSP2 primer and the oligo(dT) anchor primer. This PCR product was then used as the template for a nested PCR with the GSP3 primer and the PCR anchor primer. Products were separated on 2% agarose gels, and bands of interest were excised and cloned into the pCR2.1-TOPO vector (Invitrogen). Resultant bacterial colonies containing plasmids with the appropriate insert size were selected, and the insert was sequenced.

Overproduction and purification of B. quintana proteins from E. coli.

GFP with a C-terminal FLAG tag was PCR amplified from the pANT4 vector (43) with the primers listed in Table S1 in the supplemental material. The PCR product was ligated into the pCR2.1-TOPO-TA vector, which was then digested with XbaI and XhoI, and the insert was ligated into the pET-22b(+) vector (Novagen, Madison, WI). B. quintana wild-type strain JK31 genomic DNA was used as the template for PCR amplification of rpoE (BQ10960), nepR (BQ10970), and phyR (BQ10980); the primers used are listed in Table S1. The PCR products were ligated into the pCR2.1-TOPO-TA vector (Invitrogen). The rpoE-containing TOPO-TA vector was digested with NdeI and EcoRI. The digestion product was ligated into the pET-28a(+) vector (Novagen) to generate RpoE with an N-terminal 6×His tag. The nepR-containing TOPO-TA vector was digested with HindIII and EcoRI. The digestion product was ligated into the pMAL-c2x vector (New England BioLabs, Ipswich, MA) to generate NepR with an N-terminal maltose binding protein (MBP) tag. The phyR-containing TOPO-TA vector was digested with EcoRI and PstI. The digestion product was ligated into the pMAL-c2x vector to generate PhyR with an N-terminal MBP tag. All vectors were transformed into E. coli BL21(DE3).

For recombinant protein expression and purification, overnight cultures were diluted 1:40 and grown with shaking at 37°C for 1.5 h. Protein expression was induced by the addition of isopropyl-β-d-thiogalactopyranoside (IPTG) at a final concentration of 0.2 mM. Following induction, cells were grown at 30°C for 3 h, harvested, and stored at −20°C until purification.

To purify His-tagged RpoE and MBP-tagged NepR or PhyR, the cell pellet from 1 liter of bacteria was suspended in 20 ml of 4°C-chilled Novagen His-Bind buffer supplemented with 100 μg ml−1 lysozyme and protease inhibitor cocktail [PIC; 100× PIC is 20 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride, 1 mg ml−1 leupeptin, 0.36 mg ml−1 E-64 protease inhibitor, and 5.6 mg ml−1 benzamidine] or in 18 ml 4°C-chilled MBP buffer (20 mM Tris HCl [pH 7.4], 200 mM NaCl, 1 mM EDTA, 1 mM sodium azide) supplemented with 100 μg ml−1 lysozyme and PIC. The resuspended pellet was incubated for 15 min on ice, and 0.1% Sarkosyl was added. Cells were lysed by sonication and then centrifuged for 20 min at 9,000 × g at 4°C. The supernatant was passed through a 0.2-μm filter and then applied to a Novagen His-Bind column or a column containing Amylose Resin (New England BioLabs). Proteins were purified in accordance with the manufacturer's instructions. Following purification, elution buffer was exchanged by dialysis against 20 mM Tris HCl (pH 8.0) containing 10% glycerol.

To purify FLAG-tagged green fluorescent protein (GFP), the cell pellet from 600 ml of bacteria was suspended in 5 ml of 4°C-chilled CelLytic B solution (Sigma, St. Louis, MO) supplemented with 100 μg ml−1 lysozyme and PIC. Cells were incubated in this solution for 10 min at 4°C with vigorous shaking and then centrifuged for 10 min at 10,000 × g at 4°C. The supernatant was passed through a 0.2-μm filter and then loaded onto a column containing anti-FLAG M2 affinity gel resin (Sigma). FLAG-tagged protein was purified in accordance with the manufacturer's instructions. Following purification, elution buffer was exchanged by dialysis against 20 mM Tris HCl (pH 8.0) containing 10% glycerol.

In vitro transcription analysis.

Overlap PCR (43) was used to generate fusion products of gfp with the rpoE operon promoter and the phyR operon promoter. The promoters were amplified from B. quintana genomic DNA with the primers listed in Table S1 in the supplemental material. gfp was amplified from the pANT4 plasmid (43) with the primers listed in Table S1. PCR products were purified and ligated by overlap PCR. The resultant promoter-gfp fusion products were gel extracted and ligated into pCR2.1-TOPO-TA. The TOPO vectors containing the promoter-gfp fusion products were digested with NotI and BamHI, and the digestion products were ligated into a NotI- and BamHI-digested pANT3 vector (43). The resultant recombinant plasmids and pANT3, which contains promoterless gfp, were used as the templates for in vitro transcription assays.

Ten pmol of purified recombinant RpoE protein, with or without 20 pmol of purified recombinant NepR, was incubated for 30 min at 32°C in transcription buffer (40 mM Tris [pH 7.9], 10 mM MgCl2, 0.6 mM EDTA, 0.4 mM potassium phosphate, 1.5 mM dithiothreitol [DTT], 0.25 mg ml−1 bovine serum albumin, 20% glycerol) (44, 45). Next, 1.8 pmol of E. coli core RNA polymerase (Epicentre, Madison, WI) was added and the reaction mixtures were incubated for 10 min at 32°C. Following that incubation, 40 U of RNasin (Promega) and 50 fmol of template DNA (purified plasmid containing the promoter-gfp fusion or purified pANT3) were added to each reaction mixture. The reaction mixtures were incubated for an additional 10 min at 32°C to allow RNA polymerase binding to template DNA. Transcription was initiated by the addition of 300 μM nucleoside triphosphates. After 2 min at 32°C, 50 μg ml−1 heparin was added to prevent reinitiation. Reaction mixtures were incubated for a final 30 min at 32°C.

Following transcription, reaction mixtures were treated with 5 U of RQ1 DNase (Promega) for 4 h at 37°C. RNA was then purified with the Qiagen RNeasy Mini Kit and eluted in a final volume of 40 μl. Reverse transcription of 2 μl of RNA was carried out as described above. GFP transcription was quantified by RT-qPCR as described above, with the primers listed in Table S1 in the supplemental material. cDNA was diluted 1:200 prior to use in RT-qPCR. The relative level of transcription was determined by converting the transcript level into a copy number by using a standard curve. This value was then compared to reaction mixtures containing the plasmid encoding a promoterless gfp gene (pANT3) as the template. Data from three independent experiments were used for statistical analysis by Student's t test and to determine average gene transcription values.

Generation of B. quintana ΔrpoE ΔnepR null mutant and nepR overexpression strains.

The ΔrpoE ΔnepR mutant strain used in this study was generated by targeted allelic replacement. The genomic regions upstream and downstream of the rpoE and nepR genes were amplified with the primers listed in Table S1 in the supplemental material. These two products were then used as the template in an overlap PCR that generated a genomic fragment with an in-frame deletion of the coding sequence of rpoE and nepR. The ΔrpoE ΔnepR mutant deletion strain was constructed by a previously described two-step sacB-mediated mutagenesis method (46).

NepR was amplified from wild-type B. quintana genomic DNA to generate a NepR overexpression strain with the primers listed in Table S1 in the supplemental material. An N-terminal Shine-Dalgarno sequence and a C-terminal FLAG tag were added during PCR amplification. The PCR products were ligated into pCR2.1-TOPO-TA (Invitrogen). The TOPO-TA vector was then digested with SphI and BamHI, and the digestion product was ligated into the SphI- and BamHI-digested pANT4Δgfp vector. The resultant vector was transformed into E. coli S17-1, and biparental conjugation with wild-type B. quintana was performed. Kanr colonies were screened by immunoblotting with anti-FLAG antibody (Sigma) to confirm the expression of the recombinant protein.

B. quintana RpoE protein binding analyses.

For analysis of binding interactions, 250 pmol of purified 6×His-RpoE was incubated with 500 pmol of purified MBP-NepR or 500 pmol of purified GFP-FLAG for 10 min at room temperature and then for 20 min at 4°C to allow interaction. When PhyR was analyzed, 250 pmol of purified 6×His-RpoE was incubated with 500 pmol of purified MBP-NepR, with or without 500 pmol MBP-PhyR, for 30 min at room temperature to allow interaction. When included for phosphorylation, lithium potassium acetyl phosphate (Sigma) was present at a final concentration of 25 mM (47). Protein interactions were performed in interaction buffer (Novagen His-Bind buffer supplemented with 50 mM KCl, 5 mM MgCl2, and 1 mM DTT). Following incubation, 200 μl of charged His-Bind resin was added and the reaction mixtures were incubated for 30 min. Unbound protein was removed in accordance with the manufacturer's protocol for small-scale purification. Eluted proteins were assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by silver staining.

Quantification of B. quintana tolerance during exposure to high hemin concentrations.

To evaluate the ΔrpoE ΔnepR mutant strain for a growth defect, B. quintana wild-type or ΔrpoE ΔnepR mutant bacteria were grown on chocolate agar plates and the growth of the two strains was compared by plating serial dilutions. The B. quintana strains were harvested, pelleted, and washed three times in phosphate-buffered saline (PBS), and the bacterial pellet was resuspended in M199S medium. to a final optical density at 600 nm (OD600) of 1. Serial dilutions of the bacterial suspension were prepared, and 8 μl of each dilution were spotted onto chocolate agar plates. Duplicate sets of plates were prepared and grown in candle extinction jars, one set at 37°C and the other at 28°C.

To evaluate the tolerance of B. quintana to a high hemin concentration (10 mM), the B. quintana wild-type and isogenic mutant (ΔrpoE ΔnepR) strains were plated on minimal medium plates that do not support Bartonella growth, with hemin supplied via saturated disks placed on the agar. This assay was selected for our analysis of bacterial hemin tolerance on solid medium because the more restrictive growth conditions allowed us to observe hemin-specific growth phenotypes that could not be observed on permissive chocolate agar medium. Minimal medium plates were prepared with Brucella broth (BD Biosciences, San Diego, CA) and Bacto agar (BD Biosciences) (18). An 80-μl volume of a B. quintana suspension with an OD600 of 1 was spread onto each plate. Dilutions of each bacterial suspension also were plated on chocolate agar to enumerate CFU and to ensure that an equivalent number of each B. quintana strain was used.

For hemin solution preparation, hemin chloride powder (Sigma) was washed with 0.1 N HCl three times to remove any iron contamination. The hemin chloride was then suspended in 0.1 N NaOH and filtered to remove hemin chloride that had not solubilized. The concentration of the resultant solution was determined by measuring the OD572; a 1 mM hemin solution has an OD572 of 5.5 (48). The hemin solution was then adjusted to the appropriate concentration, and the pH was adjusted to the pH of M199S medium (~8.5). Hemin solutions were stored in the dark at −20°C and used within 1 week of preparation. A sterile 0.7-mm Whatman paper disk was placed on the center of each plate, and 8 μl of hemin solution or M199 control medium was applied to the disk. Plates were grown in candle extinction jars at 37°C for 14 days or at 28°C for 28 days. The distance from the edge of the disk to the leading edge of the zone of bacterial growth was measured. Three plates were analyzed per strain in each experiment, and statistical analysis by Student's t test was based on averaged data from three independent experiments.

A second method used to analyze B. quintana tolerance to exposure to high hemin concentrations was based on the hemin tolerance assay described by Roden et al. (21). For this assay, wild-type or ΔrpoE ΔnepR mutant B. quintana was grown on chocolate agar plates at 37 or 28°C. The bacteria were harvested, pelleted, and washed three times in PBS. The bacterial pellet was resuspended in M199S medium to a final OD600 of 1. A 20-μl volume of 500, 250, or 50 mM hemin was added to 1 ml of the bacterial suspension to generate a solution with a final hemin concentration of 10, 5, or 1 mM. All of the hemin solutions were pH balanced prior to addition to the bacterial suspension, and a pH-balanced control solution was added to bacteria for the 0 mM control cultures. Bacteria were grown for 24 h at 37 or 28°C. After incubation, the bacteria were pelleted and washed three times with M199S medium. Following the final wash, the bacteria were resuspended in 1 ml of M199S medium. Serial dilutions of the bacterial suspension were prepared and plated to enumerate the CFU. Percent survival was determined by comparing the B. quintana grown in hemin-containing medium with the B. quintana grown in the 0 mM hemin medium. Data from three independent experiments were used for statistical analysis by Student's t test.

Quantification and visualization of hemin binding by B. quintana.

To quantify hemin binding by B. quintana bacteria, a modification of the hemin binding assay described by Minnick et al. (49) and Roden et al. (21) was used. Bacteria were harvested from chocolate agar plates grown at 37 or 28°C and then washed three times with PBS. The pelleted bacteria were then resuspended in PBS and adjusted to an OD600 of 1. Hemin was added to the bacterial suspensions, and the mixture was incubated for 1 h at 37 or 28°C. During incubation, bacterial suspensions were mixed every 15 min. Following the 1 h of incubation, cells were pelleted by centrifugation for 2 min at 16,000 × g. The supernatant was removed and centrifuged again to remove any remaining bacteria. The amount of hemin present after centrifugation was quantified by measuring the absorbance of the supernatant at 400 nm (21, 49). The amount of hemin bound by B. quintana was calculated and compared to that in an identical control reaction mixture without bacteria. A saturation plot for hemin binding was generated. Data from three independent experiments were used for statistical analysis by Student's t test and for nonlinear regression analysis to determine the association constant (Ka) and maximum binding (Bmax). Nonlinear regression analysis was done with GraphPad Prism software.

To visually evaluate hemin binding, B. quintana strains were incubated in a 1 mM hemin solution for 1 h as described above. Following incubation, the bacteria were pelleted and washed three times with PBS. Control assays were performed without bacteria. The washed bacterial pellet was resuspended in 100 μl CelLytic B cell lysis solution (Sigma) and incubated for 15 min at room temperature. Serial 1:4 dilutions of the bacteria were prepared, and 4 μl of each dilution was spotted onto a nitrocellulose membrane. The membrane was dried, and hemin was then visualized by the addition of luminol-based chemiluminescence reagent and film exposure.

RESULTS

The homology and genomic context of B. quintana RpoE (BQ10960) define it as a member of the ECF15 sigma factor family.

From our results and on the basis of the genomic context, the B. quintana BQ10960 protein (RpoE) is an ECF sigma factor and a member of the ECF15 group (Fig. 1A) (28). ECF15 sigma factors are cotranscribed with a small, soluble anti-sigma factor that lacks sequence homology to other RpoE anti-sigma factors (28). BQ10970 (Fig. 1A) is a 67-amino-acid protein with no predicted signal sequence or transmembrane domains, and it lacks homology to any proteins of known function. The small size (7.8 kDa), amino acid sequence, and genomic placement of BQ10970 suggest that it is an anti-sigma factor with functional homology to the NepR proteins identified for other alphaproteobacteria. PhyR, a response regulator that functions as an anti-anti-sigma factor, is typically found in close proximity to ECF15 sigma factors (28). We determined that BQ10980 is a putative response regulator with a high degree of homology to the four previously studied PhyR proteins (see Fig. S1 in the supplemental material) (3133, 35, 38). BQ10980 has a C-terminal receiver domain and an N-terminal domain that has sequence similarity to ECF15 sigma factors; this domain structure is found in all of the PhyR response regulators studied (28). Additionally, previous genomic analysis of alphaproteobacteria performed by Starón and Mascher identified Bartonella species as containing a potential phyR locus with an associated ECF15 sigma factor and a NepR anti-sigma factor (29).

Fig 1
B. quintana RpoE (BQ10960) is an alphaproteobacterial ECF15 sigma factor. (A) The genomic context of the gene for RpoE (BQ10960) is shown. On the basis of homology, the gene name(s), predicted function, and predicted molecular mass are indicated below ...

Frequently, one or more histidine kinase-encoding genes are found in the direct vicinity of the gene for an ECF15 sigma factor (28). The sequences and structures of histidine kinase-encoding genes in genomic proximity to ECF15 sigma factor-encoding genes are highly variable (28). The B. quintana genome contains a gene that encodes a putative histidine kinase (BQ10950) upstream of rpoE (Fig. 1A). We also identified a histidine kinase (BQ10990) gene that is downstream of that for the B. quintana response regulator (BQ10980) (Fig. 1A). BQ10950 is predicted to be periplasmic, whereas BQ10990 is predicted to be cytoplasmic. We found that both B. quintana histidine kinases have structural homology to Caulobacter PhyK and Sphingomonas PhyP. Structural analysis was performed with the SMART database (http://smart.embl-heidelberg.de/). The presence of two histidine kinase genes in close genomic proximity to the gene for B. quintana RpoE is unique among the ECF15 family sigma factors studied (3133, 35, 37, 38). We used RT-PCR to define the structure of the rpoE operon and the BQ10980 (phyR) operon (Fig. 1B). As predicted, the genes for BQ10950, BQ10960 (rpoE), and BQ10970 (nepR) are cotranscribed as an operon; similarly, the genes for BQ10980 (phyR) and BQ10990 are cotranscribed as an operon (Fig. 1B).

Expression of rpoE is significantly upregulated in B. quintana after exposure to vector-specific conditions of low temperature (28°C) or high hemin concentrations.

Because B. quintana lives in two very different environments, the body louse vector and the human host, we sought to determine if RpoE is involved in directing bacterial transcription in these two environments. The expression of the alternative sigma factor gene rpoE and the housekeeping sigma factor gene rpoD was analyzed by RT-qPCR in bacteria grown at a temperature corresponding to the louse vector (28°C) or the human host (37°C), with two different growth media: solid chocolate agar and M199S liquid medium.

There was no significant difference in rpoD transcription at the two temperatures (Fig. 2A). However, the transcription of rpoE was significantly upregulated at 28°C compared with that at 37°C on solid chocolate agar (Fig. 2A). This suggested that RpoE is involved in directing transcription after exposure to the body louse temperature of 28°C. Interestingly, we observed a difference in the upregulation of rpoE at 28°C versus 37°C, depending on the medium type. Growth of B. quintana in M199S liquid medium without added hemin demonstrated no significant difference between the levels of rpoE expression at 28 and 37°C (Fig. 2B), in contrast to the significant upregulation of rpoE observed on solid chocolate agar medium at 28°C compared with 37°C (Fig. 2A). These two complex, undefined medium types are very different in composition and have different abilities to support B. quintana growth. This suggests that M199S medium (which has a bovine serum component) could present unique stressors that trigger the Bartonella GSR at 37°C, thus increasing the basal level of rpoE transcription and abrogating the difference between 28 and 37°C observed on solid agar.

Fig 2
Transcription of B. quintana rpoE is increased at 28°C and in response to high hemin concentrations. (A) Relative rpoD or rpoE transcript levels in bacteria grown on chocolate agar plates at 37°C (black) or 28°C (gray) normalized ...

When B. quintana colonizes the alimentary tract of the body louse, the bacterium is exposed to potentially toxic concentrations of hemin following the ingestion of a blood meal by the body louse (50). We therefore tested if rpoE transcription is responsive to changes in the hemin concentration. B. quintana was grown in M199S liquid medium supplemented with 0, 1, or 5 mM hemin for 24 h at 37 or 28°C, and rpoE transcription was quantified (Fig. 2B). In M199S medium, the transcription of rpoE was further upregulated in response to an increased hemin concentration (Fig. 2B), suggesting that RpoE has a role in the adaptation of B. quintana to life within the hemin-rich body louse gut. These data demonstrated that rpoE is highly transcribed under conditions mimicking those found in the body louse vector.

B. quintana RpoE positively regulates the transcription of the rpoE and phyR operons.

Many ECF sigma factors positively regulate their own transcription, which creates a positive feedback loop and amplification of the sigma factor-directed transcriptional changes (25). To determine if B. quintana RpoE positively regulates its own transcription, we first identified the TSS of the rpoE operon by 5′ RACE (Fig. 3A). We then performed in vitro transcription analysis by using the upstream region identified as the rpoE operon promoter fused to a promoterless GFP gene. When purified recombinant RpoE protein was added to E. coli RNA polymerase holoenzyme, significantly more gfp was transcribed than in the absence of RpoE (P < 0.05) (Fig. 3B). From these data, we concluded that RpoE directs the transcription of the rpoE operon. In Bradyrhizobium and Caulobacter, the transcription of phyR is positively regulated by the ECF15 sigma factor (31, 33). To test if the B. quintana phyR operon is similarly RpoE responsive, we first identified the TSS of the phyR operon by 5′ RACE (Fig. 3A). Next, in vitro transcription analysis was performed, which demonstrated that B. quintana RpoE directs the transcription of the phyR operon (Fig. 3C).

Fig 3
B. quintana RpoE positively regulates the transcription of the rpoE operon and the phyR operon. (A) The TSS of the rpoE operon and that of the phyR operon were determined by 5′ RACE. Each TSS is marked by an arrow. The −10 and −35 ...

A comparison of the identified B. quintana ECF15 promoter consensus with the three ECF15 promoters that have been experimentally verified in other alphaproteobacteria is shown in Fig. 3D (31, 38, 51). The canonical motif of ECF15 sigma factor promoters is AAC-N17-18-GTT (28). In our analysis of the B. quintana rpoE operon and the phyR operon, the −35 promoter elements we identified contained the AAC motif present in the canonical motif of ECF15 sigma factor promoters (28) (Fig. 3D). However, the −10 elements we identified lack a G residue associated with the TT motif. A secondary TSS was identified for the phyR operon. This secondary site was located 137 bp upstream of the ATG codon and yielded the −10 element CCATAT and the −35 element TTCATT. The identified −10 and −35 elements for this secondary TSS did not have any homology to either the −35 or the −10 consensus element of the ECF15 sigma factor (whereas the other phyR binding site had homology to the canonical −35 binding consensus), suggesting that this secondary phyR promoter functions independently of rpoE. To test for RpoE-independent transcription of the phyR operon, we sought to generate a strain of B. quintana in which rpoE is deleted. We were unable to delete only rpoE from the B. quintana genome, despite numerous attempts. However, simultaneous deletion of rpoE and nepR produced viable mutant bacteria. RT-qPCR was used to determine if phyR was transcribed in the ΔrpoE ΔnepR mutant strain. In the ΔrpoE ΔnepR mutant strain, phyR was transcribed; however, the level of transcription was significantly lower (P ≤ 0.01) than that observed in wild-type B. quintana (Fig. 4). This result confirmed the presence of an RpoE-independent promoter.

Fig 4
Transcription of phyR is significantly decreased in the B. quintana ΔrpoE ΔnepR mutant strain, but phyR transcript is still present. The presence of phyR transcript in the ΔrpoE ΔnepR mutant strain demonstrates that phyR ...

The protein BQ10970, a NepR homolog, is the anti-sigma factor for B. quintana RpoE.

RT-PCR analysis demonstrated that rpoE is in an operon with the genes for BQ10950 and BQ10970 (Fig. 1). The small size and genomic context of BQ10970 suggested that it is an anti-sigma factor. In other alphaproteobacteria, the ECF15 sigma factor directly binds an anti-sigma factor, NepR. This binding sterically inhibits the interaction between the RNA polymerase core and the sigma factor (31, 32, 35, 37, 38). We tested for a direct interaction between B. quintana RpoE and BQ10970 with purified, recombinant, tagged versions of BQ10970, the putative anti-sigma factor, and RpoE. Purified 6×His-tagged RpoE was incubated with purified MBP-BQ10970 or the control protein GFP-FLAG. MBP-BQ10970 coeluted with 6×His-RpoE (Fig. 5), which demonstrated that BQ10970 binds directly to RpoE and likely is a NepR anti-sigma factor for B. quintana RpoE.

Fig 5
B. quintana RpoE binds directly to the nepR homolog BQ10970. Purified 6×His-RpoE (250 pmol) was incubated with 500 pmol of purified MBP-BQ10970 or 500 pmol of purified GFP-FLAG. The protein mixture was then applied to Ni-nitrilotriacetic acid ...

To further establish that BQ10970 functions as an anti-sigma factor, in vitro transcription analysis was performed with the rpoE operon promoter-gfp fusion construct described above. The addition of purified MBP-BQ10970 significantly reduced the RpoE-mediated transcription of gfp (P < 0.05) (Fig. 6A). The phyR operon promoter-gfp fusion construct also was analyzed by this assay. Similar to what was observed for the rpoE operon, the addition of BQ10970 significantly reduced the RpoE-mediated transcription of gfp (P < 0.05) from the phyR promoter operon construct (Fig. 6B). Additionally, when BQ10970 was overexpressed in B. quintana (Fig. 7A), the transcription of rpoE and phyR was significantly decreased (P < 0.05) (Fig. 7B). From these in vitro and in vivo experiments, we concluded that BQ10970 inhibits RpoE function and thus functions as an anti-sigma factor. Because of the observed functional and genomic similarities of B. quintana BQ10970 and the Methylobacterium anti-sigma factor NepR (38), we named the protein BQ10970 NepR.

Fig 6
BQ10970 functions as the NepR anti-sigma factor for B. quintana RpoE. The relative gfp transcript levels produced in in vitro transcription assays with the rpoE operon promoter-gfp fusion construct (A) and the phyR operon promoter-gfp fusion construct ...
Fig 7
The B. quintana nepR* mutant strain overexpresses BQ10970 (nepR) and has decreased rpoE and phyR expression. (A) The relative BQ10970 transcript levels in wild-type B. quintana (gray) and a B. quintana mutant that overexpresses BQ10970 (black) grown at ...

RpoE function is positively regulated by the response regulator PhyR.

Directly downstream of the RpoE operon is the PhyR operon. In other alphaproteobacteria, PhyR functions as an anti-anti-sigma factor by disrupting the interaction between the ECF sigma factor and the anti-sigma factor NepR (3133, 38). PhyR function is phosphorylation dependent (33, 38). To test for a similar function of B. quintana PhyR, we constructed an E. coli strain expressing MBP-tagged B. quintana PhyR for use in interaction assays. Phosphorylation-dependent disruption of the RpoE-NepR interaction by B. quintana PhyR was analyzed with purified recombinant proteins. Addition of MBP-PhyR inhibited the binding interaction between 6×His-RpoE and MBP-NepR only when the PhyR protein was phosphorylated by the addition of acetyl phosphate (Fig. 8). From these data, we concluded that B. quintana PhyR functions as an anti-anti-sigma factor in a phosphorylation-dependent manner.

Fig 8
Phosphorylated BQ10980 (PhyR) disrupts the interaction between RpoE and NepR. Purified 6×His-RpoE (250 pmol) was incubated with 500 pmol of purified MBP-NepR with or without 500 pmol of purified MBP-PhyR. In the indicated reaction mixtures, MBP-PhyR ...

B. quintana RpoE confers increased tolerance to high hemin concentrations.

In initial experiments, we determined that rpoE transcription is increased when B. quintana is grown at a temperature that corresponds to the hemin-rich body louse vector niche, compared to the temperature that corresponds to the hemin-depleted human bloodstream (Fig. 2A). rpoE transcription also was upregulated when bacteria were exposed to high concentrations of hemin (Fig. 2B). These observations led us to hypothesize that RpoE-mediated transcription is important for bacterial adaptation and survival within the hemin-toxic body louse gut. To test this hypothesis, we used the isogenic B. quintana ΔrpoE ΔnepR mutant strain. We first tested the ΔrpoE ΔnepR mutant strain for a growth defect at 37 and 28°C by plating serial dilutions of the ΔrpoE ΔnepR mutant strain and the wild-type strain on chocolate agar plates. No growth defect of the ΔrpoE ΔnepR mutant strain was observed at either temperature on chocolate agar (Fig. 9).

Fig 9
Wild-type B. quintana and the ΔrpoE ΔnepR mutant strain display similar growth on chocolate agar. Wild-type and ΔrpoE ΔnepR mutant bacteria grown at 37 or 28°C were harvested into M199S medium and adjusted to an ...

The role of RpoE in B. quintana tolerance to a high-hemin environment was evaluated by plating B. quintana strains on plates with minimal medium agar lacking hemin; this agar does not support B. quintana growth. A suspension of wild-type or ΔrpoE ΔnepR mutant B. quintana was spread evenly over the surfaces of separate minimal medium agar plates. Then hemin, an essential iron source for B. quintana, was supplied by placing hemin-saturated disks on the B. quintana-inoculated agar. The hemin from each disk diffuses on the surface of the agar, establishing a hemin concentration gradient and rescuing bacterial growth when the hemin reaches a concentration sufficient to support B. quintana growth. As hemin continues to diffuse into the agar, the zone immediately around the disk gradually achieves higher concentrations of hemin, and if a B. quintana strain is more sensitive to these higher concentrations of hemin, growth will be abolished in this zone closest to the disk. This assay permits testing for inhibition of growth due to high hemin concentrations in different B. quintana mutants, by testing for the appearance of a zone of growth inhibition when toxic hemin concentrations accumulate immediately surrounding the disk. Minimal medium is used for this analysis because it allows a direct correlation between the bacterial growth observed and the hemin concentration provided from the disk. The plates were incubated at 37°C (for 7 days) or 28°C (for 21 days). At both temperatures, the growth of the wild-type strain was not inhibited by 10 mM hemin and no zone of growth inhibition was observed around the hemin-impregnated disk (Table 2). At 37°C, the ΔrpoE ΔnepR mutant strain displayed a decreased ability to tolerate high hemin concentrations, and a zone of growth inhibition was observed around the hemin-impregnated disk (Table 2). At 28°C, the ΔrpoE ΔnepR mutant strain grew very poorly on the minimal medium agar, with few or no colonies observed. These results demonstrated that RpoE- and/or NepR-mediated transcription is necessary for B. quintana adaptation to growth in high-hemin environments and that this requirement is greater at 28°C (body louse temperature).

Table 2
Inhibition of B. quintana growth by 10 mM hemin

A second method also was used to analyze the hemin sensitivity of the ΔrpoE ΔnepR mutant strain. Wild-type or ΔrpoE ΔnepR mutant B. quintana harvested from chocolate agar plates was cultivated in M199S liquid medium containing 0, 1, 5, or 10 mM hemin and grown for 24 h at 37 or 28°C. The ability of each strain to tolerate exposure to high hemin concentrations was evaluated by determining the percentage of B. quintana that survived following growth in M199S medium supplemented with hemin. At 37°C, the ΔrpoE ΔnepR mutant strain had a survival defect compared with wild-type B. quintana after cultivation in liquid medium supplemented with 5 mM hemin (Fig. 10A). Both the wild-type and ΔrpoE ΔnepR mutant strains tolerated high hemin concentrations better at 28°C than at 37°C (Fig. 10). At 28°C, the ΔrpoE ΔnepR mutant strain was significantly more sensitive than the wild-type strain to all three of the hemin concentrations tested (Fig. 10B). From these experiments, we concluded that the deletion of rpoE and nepR decreases the ability of B. quintana to survive exposure to high concentrations of hemin in liquid medium.

Fig 10
The B. quintana ΔrpoE ΔnepR mutant strain demonstrates increased sensitivity to high hemin concentrations. Wild-type (black) and ΔrpoE ΔnepR mutant (gray) bacteria were incubated in M199S medium supplemented with hemin ...

The B. quintana ΔrpoE ΔnepR mutant strain has a greater hemin binding capacity than the wild-type strain.

The decreased ability of the ΔrpoE ΔnepR mutant strain to tolerate high concentrations of hemin could be due to increased binding of potentially toxic hemin. To test this, we evaluated the hemin binding capacities of both the wild-type and ΔrpoE ΔnepR mutant strains at both 28 and 37°C (Table 3). A saturation plot for hemin binding was generated. Transformation of the data produced a linear plot, indicating that a single class of binding site was present (5255). The hemin binding data were subjected to nonlinear regression analysis to calculate the Ka and Bmax of the B. quintana strains for hemin (52, 54, 55). The Ka of the ΔrpoE ΔnepR mutant strain was greater than that of the wild-type strain at both 37 and 28°C (Table 3). The Bmax of the ΔrpoE ΔnepR mutant strain was greater than that of the wild-type strain at both 37 and 28°C; however, this difference was statistically significant only at 37°C (Table 3). From this analysis, we concluded that the ΔrpoE ΔnepR mutant strain is able to bind greater amounts of hemin than wild-type B. quintana.

Table 3
The ΔrpoE ΔnepR mutant strain demonstrates increased hemin binding at 37 and 28°C

Finally, the amount of hemin bound by each strain also was visualized by exploiting the natural electro-generated chemiluminescent (ECL) behavior of hemin (56, 57). The B. quintana wild-type and ΔrpoE ΔnepR mutant strains were incubated in medium containing 1 mM hemin for 1 h at 37 or 28°C. The bacteria were then washed, pelleted, and lysed. Dilutions of each strain were spotted onto nitrocellulose and visualized with a commercial ECL reagent. Similar to horseradish peroxidase, hemin serves as a catalyst in the chemiluminescence reaction between luminol and H2O2 (56, 57). The light produced was visualized by exposing film to the nitrocellulose. At both 37 and 28°C, more light was produced by the ΔrpoE ΔnepR mutant strain than by wild-type bacteria, indicating that a greater concentration of hemin was associated with the ΔrpoE ΔnepR mutant strain (Fig. 11). This increased hemin binding capacity and accumulation of hemin in the ΔrpoE ΔnepR mutant strain could contribute to the increased hemin sensitivity of this strain.

Fig 11
More hemin is associated with B. quintana ΔrpoE ΔnepR mutant bacteria than with wild-type bacteria. Wild-type and ΔrpoE ΔnepR mutant bacteria were incubated with 1 mM hemin for 1 h at 37 or 28°C. The bacteria were ...

DISCUSSION

The ECF sigma factors in the ECF15 family are found only in alphaproteobacteria, and they are believed to be responsible for the GSR in this class of organisms (28). To date, detailed functional and structural analyses of ECF15 sigma factors have been performed exclusively with nonhuman pathogens (3138, 51, 58). We present the first analysis of an ECF15 sigma factor in a human pathogen, B. quintana. We first identified the genomic components and functional interactions of the B. quintana ECF15 transcription regulatory system, rpoE (BQ10960; functions as a sigma factor), nepR (BQ10970; functions as an anti-sigma factor), and phyR (BQ10980; functions as an anti-anti-sigma factor). The subsequent focus of this study was to characterize the unique role of RpoE in two conditions encountered during B. quintana colonization of its human body louse vector: a decreased temperature and an increased hemin concentration. Specific environmental conditions under which ECF15 sigma factor-directed transcription has been demonstrated to be important for bacterial survival include heat, hyperosmotic conditions, nutrient deprivation, desiccation, and UV exposure (31, 33, 35, 37, 38, 51). For B. quintana, we determined that the transcription of the ECF15 sigma factor RpoE is increased in response to two stressors present in the body louse vector, a toxic hemin concentration and a decreased temperature. Additionally, we observed decreased survival of B. quintana bacteria lacking RpoE-directed transcription following exposure to toxic hemin concentrations (Fig. 10). Our data suggest that B. quintana employs the alphaproteobacteria GSR system to survive the unique stressors present in the body louse vector.

B. quintana RpoE is the first ECF15 sigma factor implicated in hemin tolerance. Non-ECF15 sigma factors from other Gram-negative bacteria are involved in iron sensing, acquisition, and metabolism. Examples of non-ECF15 sigma factors involved in iron sensing include E. coli FecI (59), Pseudomonas aeruginosa PvdS (60), Pseudomonas putida PupI (61), Bordetella avium RhuI (62), and Bordetella pertussis HurI (63). Activation of transcription directed by these ECF sigma factors involves an iron-sensing outer membrane protein that transmits a signal to an inner membrane protein, which then interacts with the cytoplasmic sigma factor (26). Bartonella RpoE and its regulators NepR and PhyR are predicted to be cytoplasmic on the basis of sequence analysis. B. quintana produces two putative histidine kinases (BQ10950 and BQ10990) in genomic proximity to phyR that could function to phosphorylate PhyR and activate the anti-anti-sigma factor function. Localization algorithms predict that one of these kinases is localized to the cytosol and the other is localized to the periplasm. Both of these putative B. quintana histidine kinases are structurally similar to Caulobacter PhyK, which phosphorylates PhyR (35). It is possible that one or both of these B. quintana kinases are phosphorylated in response to an outer membrane or cytoplasmic iron-sensing protein. Alternatively, these B. quintana kinases could directly sense hemin and autophosphorylate in response to changes in iron levels, as occurs with the sensor kinase ChrS of Corynebacterium diphtheriae (64).

Because we observed that the transcription of rpoE is induced by high concentrations of hemin (Fig. 2), we sought to determine if the ΔrpoE ΔnepR mutant strain displays perturbed sensitivity to high concentrations of hemin. The concentration of free hemin in the body louse gut following a blood meal has not been determined. However, it is known that in the adult dog tick (Dermacentor variabilis), the concentration of hemin in the arthropod gut lumen ranges from 3 to 12 mM over a period of 1 to 43 days after tick detachment from the mammalian host (65). Thus, the concentrations chosen for our hemin sensitivity assays (1, 5, and 10 mM) should encompass the range of hemin concentrations to which B. quintana is exposed in the body louse vector. The liquid culture-based hemin sensitivity assay we used in this study has previously been used to quantify the sensitivity of Bartonella henselae to hemin (21). Similar to the findings of Roden et al. for B. henselae, we observed significantly greater B. quintana survival in the presence of a high hemin concentration at 28°C than at 37°C (Fig. 10). In contrast, we found that wild-type B. quintana has greater tolerance of high levels of hemin than what was observed previously for B. henselae. The difference in hemin sensitivity between the two Bartonella species was most prominent at 28°C. The percent survival of B. quintana following exposure to 5 mM hemin at 28°C was 62% (Fig. 10), but that of B. henselae was less than 1% (21). This difference in hemin tolerance could be an evolutionary adaptation that allows each Bartonella species to adapt to its cognate arthropod vector. The vector for B. henselae is the cat flea, and the vector for B. quintana is the body louse. The cat flea can survive for long periods of time without a blood meal, and fleas do not always remain associated with the host after feeding (66). However, body lice live near the skin on the inner part of the host's clothing and typically require several blood meals per day (67). This vector feeding pattern results in persistently higher concentrations of hemin in the gastrointestinal tract of the body louse than in that of the cat flea; thus, B. quintana is required to withstand higher concentrations of hemin than B. henselae during vector colonization.

To explore potential mechanisms leading to the increased hemin sensitivity of the ΔrpoE ΔnepR mutant strain (Table 2; Fig. 10), bacterial hemin binding was analyzed. The ΔrpoE ΔnepR mutant strain had greater maximal hemin binding than wild-type B. quintana (Table 3). Additionally, we observed an increased amount of hemin associated with the ΔrpoE ΔnepR mutant strain at both 37 and 28°C (Fig. 11). The increased hemin binding by the ΔrpoE ΔnepR mutant strain could directly contribute to the increased hemin sensitivity of this strain. Other potential mechanisms that could result in increased hemin sensitivity include increased sequestration or decreased hemin export or degradation. It is also possible that RpoE indirectly affects the transcription of one or more of the five B. quintana hbp genes and/or the hut locus. The Hbp bind hemin (68) and are directly transcriptionally regulated by Irr and RirA transcription factors (69). Components of the hut locus are hemin responsive and also are primarily transcriptionally regulated by Irr (20). Future studies of the RpoE regulon will increase our understanding of hemin-regulatory and hemin-responsive networks in B. quintana.

In analyzing the functional importance of B. quintana RpoE, we determined that the deletion of rpoE alone is lethal to B. quintana but the simultaneous deletion of rpoE and the anti-sigma factor gene nepR produces viable bacteria. We also were unable to delete nepR alone from B. quintana. Similarly, in Sinorhizobium meliloti, the deletion of nepR alone (smc01505) is lethal (51). Sinorhizobium nepR can be disrupted only in the absence of its cognate sigma factor, rpoE2 (51). In both Sinorhizobium and B. quintana, nepR and rpoE/rpoE2 are cotranscribed and nepR directly precedes the sigma factor gene in the operon (51). Interestingly, unlike what we observed in B. quintana, in Sinorhizobium, deletion of the ECF15 sigma factor gene rpoE2 alone is not lethal (51). This difference could occur because Sinorhizobium encodes six ECF sigma factors that could have redundant functions (70). B. quintana has a small, compact 1.5-Mb genome (23) and produces a single ECF sigma factor; thus, it is not surprising that redundant mechanisms for RpoE and NepR function regulation are absent.

All of the phyR operons that have been studied have an RpoE-dependent promoter (31, 33, 37), and a second RpoE-independent promoter has been identified in only one alphaproteobacterial genus (37). We identified two distinct promoters for the B. quintana phyR operon, only one of which requires RpoE. The B. quintana RpoE-responsive promoter contained both −10 and −35 residues that correspond to the canonical motif for ECF15 sigma factors (28). We confirmed the presence of the second, RpoE-independent, promoter by analyzing the transcription of phyR in the ΔrpoE ΔnepR mutant strain (Fig. 4). There also are two promoters for the phyR operon in Sphingomonas, and similar to the situation in B. quintana, one of these promoters functions in the absence of the ECF15 sigma factor (37). In contrast, in Bradyrhizobium and Caulobacter, the ECF15 sigma factor is essential for the expression of phyR (31, 33), which is indicative of a single RpoE-responsive promoter for the phyR operon. The B. quintana RpoE-independent phyR operon TSS promoter elements we identified, −10 CCATAT and −35 TTCATT, lack strong homology to any characterized binding consensus sequence. This RpoE-independent consensus binding sequence and the RpoE binding consensus sequence are both found in the phyR promoter region of B. henselae, Bartonella grahamii, and Bartonella tribocorum. This finding suggests that the element that binds this consensus and allows RpoE-independent phyR operon transcription is conserved and functional in all four of these Bartonella species. It will be important to study this RpoE-independent phyR promoter to identify additional regulatory mechanisms for controlling the expression of the phyR operon in B. quintana. This likely will provide insight into regulation in Sphingomonas and other alphaproteobacteria, as well as in Bartonella species.

ECF15 sigma factors were identified relatively recently (30) and orchestrate the alphaproteobacterial GSR. Members of each alphaproteobacterial genus must be able to sense different stressors in their environment to launch the GSR, through PhyR regulation of RpoE. Both of the niches occupied by B. quintana are markedly different from the niches/environments occupied by the other alphaproteobacteria whose ECF15 sigma factors have been studied (Caulobacter, Sinorhizobium, Bradyrhizobium, and Methylobacterium). The stressors to which the latter genera respond include heat and osmotic stress, starvation for carbon or nitrogen, and desiccation (29). In contrast, B. quintana is exposed to alternating shifts in temperature (37 versus 28°C) and changes in ambient hemin concentrations (extremely limited versus toxic). The focus of our study was to evaluate the role of B. quintana RpoE in bacterial adaptation to two stressors that are associated with the body louse vector niche. Our work demonstrated that RpoE has a role in conferring the ability of B. quintana to tolerate the high hemin concentrations present in the vector gut. Transcriptional modification of hemin uptake and metabolism genes is likely essential for B. quintana survival during the transition from the hemin-restricted host bloodstream environment to the hemin-toxic body louse vector environment. Our observation that the ΔrpoE ΔnepR mutant strain exhibits decreased survival under the high hemin conditions found in the body louse vector suggests that members of the RpoE regulon are important for bacterial survival within the arthropod. Genes within the RpoE regulon likely include additional, vector-specific virulence genes and potential targets for the prevention of transmission to humans. Because vector colonization is essential for the infection of humans with B. quintana, targeting of the arthropod vector stage would be ideal for interruption of the body louse-human transmission cycle.

Supplementary Material

Supplemental material:

ACKNOWLEDGMENTS

J.E.K. received funding support from a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research, a California HIV Research Program Award, and NIH grant R01AI52813 from the NIAID. S.A. was supported by NIH grant T32A1007641.

Footnotes

Published ahead of print 5 April 2013

Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.01972-12.

REFERENCES

1. Vinson JW, Varela G, Molina-Pasquel C. 1969. Trench fever. 3. Induction of clinical disease in volunteers inoculated with Rickettsia quintana propagated on blood agar. Am. J. Trop. Med. Hyg. 18:713–722. [PubMed]
2. Greub G, Raoult D. 2004. Bartonella infections resurgence in the new century, p 35–68 In Fong IW, Drlica K, editors. (ed), Emerging infectious diseases of the 21st century, vol 2 Springer, New York, NY.
3. Pulliainen AT, Dehio C. 2012. Persistence of Bartonella spp. stealth pathogens: from subclinical infections to vasoproliferative tumor formation. FEMS Microbiol. Rev. 36:563–599. [PubMed]
4. Cockerell CJ, Whitlow MA, Webster GF, Friedman-Kien AE. 1987. Epithelioid angiomatosis: a distinct vascular disorder in patients with the acquired immunodeficiency syndrome or AIDS-related complex. Lancet ii:654–656. [PubMed]
5. Brouqui P, Lascola B, Roux V, Raoult D. 1999. Chronic Bartonella quintana bacteremia in homeless patients. N. Engl. J. Med. 340:184–189. [PubMed]
6. Koehler JE, Sanchez MA, Garrido CS, Whitfeld MJ, Chen FM, Berger TG, Rodriguez-Barradas MC, LeBoit PE, Tappero JW. 1997. Molecular epidemiology of Bartonella infections in patients with bacillary angiomatosis-peliosis. N. Engl. J. Med. 337:1876–1883. [PubMed]
7. Bonilla DL, Kabeya H, Henn J, Kramer VL, Kosoy MY. 2009. Bartonella quintana in body lice and head lice from homeless persons, San Francisco, California, USA. Emerg. Infect. Dis. 15:912–915. [PMC free article] [PubMed]
8. Ito S, Vinson JW. 1965. Fine structure of Rickettsia quintana cultivated in vitro and in the louse. J. Bacteriol. 89:481–495. [PMC free article] [PubMed]
9. Seki N, Kasai S, Saito N, Komagata O, Mihara M, Sasaki T, Tomita T, Kobayashi M. 2007. Quantitative analysis of proliferation and excretion of Bartonella quintana in body lice, Pediculus humanus. Am. J. Trop. Med. Hyg. 77:562–566. [PubMed]
10. Wigglesworth VB. 1941. The sensory physiology of the human louse Pediculus humanus corporis de Geer (Anoplura). Parasitology 33:67–109.
11. Sander A, Kretzer S, Bredt W, Oberle K, Bereswill S. 2000. Hemin-dependent growth and hemin binding of Bartonella henselae. FEMS Microbiol. Lett. 189:55–59. [PubMed]
12. Graça-Souza AV, Maya-Monteiro C, Paiva-Silva GO, Braz GR, Paes MC, Sorgine MH, Oliveira MF, Oliveira PL. 2006. Adaptations against heme toxicity in blood-feeding arthropods. Insect Biochem. Mol. Biol. 36:322–335. [PubMed]
13. Myers WF, Cutler LD, Wisseman CL., Jr 1969. Role of erythrocytes and serum in the nutrition of Rickettsia quintana. J. Bacteriol. 97:663–666. [PMC free article] [PubMed]
14. Myers WF, Osterman JV, Wisseman CL., Jr 1972. Nutritional studies of Rickettsia guintana: nature of the hematin requirement. J. Bacteriol. 109:89–95. [PMC free article] [PubMed]
15. Stauff DL, Bagaley D, Torres VJ, Joyce R, Anderson KL, Kuechenmeister L, Dunman PM, Skaar EP. 2008. Staphylococcus aureus HrtA is an ATPase required for protection against heme toxicity and prevention of a transcriptional heme stress response. J. Bacteriol. 190:3588–3596. [PMC free article] [PubMed]
16. Torres VJ, Stauff DL, Pishchany G, Bezbradica JS, Gordy LE, Iturregui J, Anderson KL, Dunman PM, Joyce S, Skaar EP. 2007. A Staphylococcus aureus regulatory system that responds to host heme and modulates virulence. Cell Host Microbe 1:109–119. [PMC free article] [PubMed]
17. Rasmussen AW, Alexander HL, Perkins-Balding D, Shafer WM, Stojiljkovic I. 2005. Resistance of Neisseria meningitidis to the toxic effects of heme iron and other hydrophobic agents requires expression of ght. J. Bacteriol. 187:5214–5223. [PMC free article] [PubMed]
18. Battisti JM, Sappington KN, Smitherman LS, Parrow NL, Minnick MF. 2006. Environmental signals generate a differential and coordinated expression of the heme receptor gene family of Bartonella quintana. Infect. Immun. 74:3251–3261. [PMC free article] [PubMed]
19. Liu M, Ferrandez Y, Bouhsira E, Monteil M, Franc M, Boulouis HJ, Biville F. 2012. Heme binding proteins of Bartonella henselae are required when undergoing oxidative stress during cell and flea invasion. PLoS One 7:e48408 doi:10.1371/journal.pone.0048408. [PMC free article] [PubMed]
20. Parrow NL, Abbott J, Lockwood AR, Battisti JM, Minnick MF. 2009. Function, regulation, and transcriptional organization of the hemin utilization locus of Bartonella quintana. Infect. Immun. 77:307–316. [PMC free article] [PubMed]
21. Roden JA, Wells DH, Chomel BB, Kasten RW, Koehler JE. 2012. Hemin binding protein C is found in outer membrane vesicles and protects Bartonella henselae against toxic concentrations of hemin. Infect. Immun. 80:929–942. [PMC free article] [PubMed]
22. Helmann JD. 1999. Anti-sigma factors. Curr. Opin. Microbiol. 2:135–141. [PubMed]
23. Alsmark CM, Frank AC, Karlberg EO, Legault BA, Ardell DH, Canback B, Eriksson AS, Naslund AK, Handley SA, Huvet M, La Scola B, Holmberg M, Andersson SG. 2004. The louse-borne human pathogen Bartonella quintana is a genomic derivative of the zoonotic agent Bartonella henselae. Proc. Natl. Acad. Sci. U. S. A. 101:9716–9721. [PubMed]
24. Resto-Ruiz SI, Sweger D, Widen RH, Valkov N, Anderson BE. 2000. Transcriptional activation of the htrA (high-temperature requirement A) gene from Bartonella henselae. Infect. Immun. 68:5970–5978. [PMC free article] [PubMed]
25. Helmann JD. 2002. The extracytoplasmic function (ECF) sigma factors. Adv. Microb. Physiol. 46:47–110. [PubMed]
26. Brooks BE, Buchanan SK. 2008. Signaling mechanisms for activation of extracytoplasmic function (ECF) sigma factors. Biochim. Biophys. Acta 1778:1930–1945. [PMC free article] [PubMed]
27. Raivio TL, Silhavy TJ. 2001. Periplasmic stress and ECF sigma factors. Annu. Rev. Microbiol. 55:591–624. [PubMed]
28. Starón A, Sofia HJ, Dietrich S, Ulrich LE, Liesegang H, Mascher T. 2009. The third pillar of bacterial signal transduction: classification of the extracytoplasmic function (ECF) sigma factor protein family. Mol. Microbiol. 74:557–581. [PubMed]
29. Starón A, Mascher T. 2010. General stress response in alpha-proteobacteria: PhyR and beyond. Mol. Microbiol. 78:271–277. [PubMed]
30. Gourion B, Rossignol M, Vorholt JA. 2006. A proteomic study of Methylobacterium extorquens reveals a response regulator essential for epiphytic growth. Proc. Natl. Acad. Sci. U. S. A. 103:13186–13191. [PubMed]
31. Gourion B, Sulser S, Frunzke J, Francez-Charlot A, Stiefel P, Pessi G, Vorholt JA, Fischer HM. 2009. The PhyR-sigma(EcfG) signalling cascade is involved in stress response and symbiotic efficiency in Bradyrhizobium japonicum. Mol. Microbiol. 73:291–305. [PubMed]
32. Bastiat B, Sauviac L, Bruand C. 2010. Dual control of Sinorhizobium meliloti RpoE2 sigma factor activity by two PhyR-type two-component response regulators. J. Bacteriol. 192:2255–2265. [PMC free article] [PubMed]
33. Herrou J, Foreman R, Fiebig A, Crosson S. 2010. A structural model of anti-anti-sigma inhibition by a two-component receiver domain: the PhyR stress response regulator. Mol. Microbiol. 78:290–304. [PMC free article] [PubMed]
34. Herrou J, Rotskoff G, Luo Y, Roux B, Crosson S. 2012. Structural basis of a protein partner switch that regulates the general stress response of alpha-proteobacteria. Proc. Natl. Acad. Sci. U. S. A. 109:E1415–E1423. [PubMed]
35. Lourenço RF, Kohler C, Gomes SL. 2011. A two-component system, an anti-sigma factor and two paralogous ECF sigma factors are involved in the control of general stress response in Caulobacter crescentus. Mol. Microbiol. 80:1598–1612. [PubMed]
36. Campagne S, Damberger FF, Kaczmarczyk A, Francez-Charlot A, Allain FH, Vorholt JA. 2012. Structural basis for sigma factor mimicry in the general stress response of Alphaproteobacteria. Proc. Natl. Acad. Sci. U. S. A. 109:E1405–E1414. [PubMed]
37. Kaczmarczyk A, Campagne S, Danza F, Metzger LC, Vorholt JA, Francez-Charlot A. 2011. The PhyR-NepR-σEcfG cascade in Sphingomonas sp. Fr1: role in the general stress response and identification of a negative regulator of PhyR. J. Bacteriol. 193:6629–6638. [PMC free article] [PubMed]
38. Francez-Charlot A, Frunzke J, Reichen C, Ebneter JZ, Gourion B, Vorholt JA. 2009. Sigma factor mimicry involved in regulation of general stress response. Proc. Natl. Acad. Sci. U. S. A. 106:3467–3472. [PubMed]
39. Zhang P, Chomel BB, Schau MK, Goo JS, Droz S, Kelminson KL, George SS, Lerche NW, Koehler JE. 2004. A family of variably expressed outer-membrane proteins (Vomp) mediates adhesion and autoaggregation in Bartonella quintana. Proc. Natl. Acad. Sci. U. S. A. 101:13630–13635. [PubMed]
40. Koehler JE, Quinn FD, Berger TG, LeBoit PE, Tappero JW. 1992. Isolation of Rochalimaea species from cutaneous and osseous lesions of bacillary angiomatosis. N. Engl. J. Med. 327:1625–1631. [PubMed]
41. Gaynor EC, Cawthraw S, Manning G, MacKichan JK, Falkow S, Newell DG. 2004. The genome-sequenced variant of Campylobacter jejuni NCTC 11168 and the original clonal clinical isolate differ markedly in colonization, gene expression, and virulence-associated phenotypes. J. Bacteriol. 186:503–517. [PMC free article] [PubMed]
42. Frohman MA, Dush MK, Martin GR. 1988. Rapid production of full-length cDNAs from rare transcripts: amplification using a single gene-specific oligonucleotide primer. Proc. Natl. Acad. Sci. U. S. A. 85:8998–9002. [PubMed]
43. Lee AK, Falkow S. 1998. Constitutive and inducible green fluorescent protein expression in Bartonella henselae. Infect. Immun. 66:3964–3967. [PMC free article] [PubMed]
44. Hefty PS, Stephens RS. 2007. Chlamydial type III secretion system is encoded on ten operons preceded by sigma 70-like promoter elements. J. Bacteriol. 189:198–206. [PMC free article] [PubMed]
45. Kang JG, Hahn MY, Ishihama A, Roe JH. 1997. Identification of sigma factors for growth phase-related promoter selectivity of RNA polymerases from Streptomyces coelicolor A3(2). Nucleic Acids Res. 25:2566–2573. [PMC free article] [PubMed]
46. MacKichan JK, Gerns HL, Chen YT, Zhang P, Koehler JE. 2008. A SacB mutagenesis strategy reveals that the Bartonella quintana variably expressed outer membrane proteins are required for bloodstream infection of the host. Infect. Immun. 76:788–795. [PMC free article] [PubMed]
47. Da Re SS, Deville-Bonne D, Tolstykh T, Véron M, Stock JB. 1999. Kinetics of CheY phosphorylation by small molecule phosphodonors. FEBS Lett. 457:323–326. [PubMed]
48. Windholz M. 1976. The Merck index: an encyclopedia of chemicals, drugs, and biologicals, 9th ed Merck & Co., Rahway, NJ.
49. Minnick MF, Sappington KN, Smitherman LS, Andersson SG, Karlberg O, Carroll JA. 2003. Five-member gene family of Bartonella quintana. Infect. Immun. 71:814–821. [PMC free article] [PubMed]
50. Paiva-Silva GO, Cruz-Oliveira C, Nakayasu ES, Maya-Monteiro CM, Dunkov BC, Masuda H, Almeida IC, Oliveira PL. 2006. A heme-degradation pathway in a blood-sucking insect. Proc. Natl. Acad. Sci. U. S. A. 103:8030–8035. [PubMed]
51. Sauviac L, Philippe H, Phok K, Bruand C. 2007. An extracytoplasmic function sigma factor acts as a general stress response regulator in Sinorhizobium meliloti. J. Bacteriol. 189:4204–4216. [PMC free article] [PubMed]
52. Amano A, Kuboniwa M, Kataoka K, Tazaki K, Inoshita E, Nagata H, Tamagawa H, Shizukuishi S. 1995. Binding of hemoglobin by Porphyromonas gingivalis. FEMS Microbiol. Lett. 134:63–67. [PubMed]
53. Rosenthal HE. 1967. A graphic method for the determination and presentation of binding parameters in a complex system. Anal. Biochem. 20:525–532. [PubMed]
54. Smalley JW, Birss AJ, McKee AS, Marsh PD. 1998. Hemin regulation of hemoglobin binding by Porphyromonas gingivalis. Curr. Microbiol. 36:102–106. [PubMed]
55. Tompkins GR, Wood DP, Birchmeier KR. 1997. Detection and comparison of specific hemin binding by Porphyromonas gingivalis and Prevotella intermedia. J. Bacteriol. 179:620–626. [PMC free article] [PubMed]
56. Chen GN, Zhang L, Lin RE, Yang ZC, Duan JP, Chen HQ, Hibbert DB. 2000. The electrogenerated chemiluminescent behavior of hemin and its catalytic activity for the electrogenerated chemiluminescence of lucigenin. Talanta 50:1275–1281. [PubMed]
57. Ikariyama Y, Suzuki S, Aizawa M. 1982. Luminescence immunoassay of human serum albumin with hemin as labeling catalyst. Anal. Chem. 54:1126–1129. [PubMed]
58. Gourion B, Francez-Charlot A, Vorholt JA. 2008. PhyR is involved in the general stress response of Methylobacterium extorquens AM1. J. Bacteriol. 190:1027–1035. [PMC free article] [PubMed]
59. Mahren S, Braun V. 2003. The FecI extracytoplasmic-function sigma factor of Escherichia coli interacts with the beta′ subunit of RNA polymerase. J. Bacteriol. 185:1796–1802. [PMC free article] [PubMed]
60. Tiburzi F, Imperi F, Visca P. 2008. Intracellular levels and activity of PvdS, the major iron starvation sigma factor of Pseudomonas aeruginosa. Mol. Microbiol. 67:213–227. [PubMed]
61. Koster M, van Klompenburg W, Bitter W, Leong J, Weisbeek P. 1994. Role for the outer membrane ferric siderophore receptor PupB in signal transduction across the bacterial cell envelope. EMBO J. 13:2805–2813. [PubMed]
62. Kirby AE, Metzger DJ, Murphy ER, Connell TD. 2001. Heme utilization in Bordetella avium is regulated by RhuI, a heme-responsive extracytoplasmic function sigma factor. Infect. Immun. 69:6951–6961. [PMC free article] [PubMed]
63. Brickman TJ, Vanderpool CK, Armstrong SK. 2006. Heme transport contributes to in vivo fitness of Bordetella pertussis during primary infection in mice. Infect. Immun. 74:1741–1744. [PMC free article] [PubMed]
64. Ito Y, Nakagawa S, Komagata A, Ikeda-Saito M, Shiro Y, Nakamura H. 2009. Heme-dependent autophosphorylation of a heme sensor kinase, ChrS, from Corynebacterium diphtheriae reconstituted in proteoliposomes. FEBS Lett. 583:2244–2248. [PubMed]
65. Tarnowski BI, Coons LB. 1989. Ultrastructure of the midgut and blood meal digestion in the adult tick Dermacentor variabilis. Exp. Appl. Acarol. 6:263–289. [PubMed]
66. Rust MK, Dryden MW. 1997. The biology, ecology, and management of the cat flea. Annu. Rev. Entomol. 42:451–473. [PubMed]
67. Kittler R, Kayser M, Stoneking M. 2003. Molecular evolution of Pediculus humanus and the origin of clothing. Curr. Biol. 13:1414–1417. [PubMed]
68. Carroll JA, Coleman SA, Smitherman LS, Minnick MF. 2000. Hemin-binding surface protein from Bartonella quintana. Infect. Immun. 68:6750–6757. [PMC free article] [PubMed]
69. Battisti JM, Smitherman LS, Sappington KN, Parrow NL, Raghavan R, Minnick MF. 2007. Transcriptional regulation of the heme binding protein gene family of Bartonella quintana is accomplished by a novel promoter element and iron response regulator. Infect. Immun. 75:4373–4385. [PMC free article] [PubMed]
70. Galibert F, Finan TM, Long SR, Puhler A, Abola P, Ampe F, Barloy-Hubler F, Barnett MJ, Becker A, Boistard P, Bothe G, Boutry M, Bowser L, Buhrmester J, Cadieu E, Capela D, Chain P, Cowie A, Davis RW, Dreano S, Federspiel NA, Fisher RF, Gloux S, Godrie T, Goffeau A, Golding B, Gouzy J, Gurjal M, Hernandez-Lucas I, Hong A, Huizar L, Hyman RW, Jones T, Kahn D, Kahn ML, Kalman S, Keating DH, Kiss E, Komp C, Lelaure V, Masuy D, Palm C, Peck MC, Pohl TM, Portetelle D, Purnelle B, Ramsperger U, Surzycki R, Thebault P, Vandenbol M, Vorholter FJ, Weidner S, Wells DH, Wong K, Yeh KC, Batut J. 2001. The composite genome of the legume symbiont Sinorhizobium meliloti. Science 293:668–672. [PubMed]
71. Simon R, Priefer U, Puhler A. 1983. A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in Gram negative bacteria. Nat. Biotechnol. 1:784–791.

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)