Identification and regulation of cold-inducible

Microbiology (2005), 151, 1895–1909

DOI 10.1099/mic.0.27785-0

Identification and regulation of cold-inducible factors of Bordetella bronchiseptica Dorothee Stu¨bs,1 Thilo M. Fuchs,2 Boris Schneider,13 Armin Bosserhoff3 and Roy Gross1 1

Lehrstuhl fu¨r Mikrobiologie, Biozentrum der Universita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany

Correspondence Roy Gross [email protected]

2

Zentralinstitut fu¨r Erna¨hrungs- und Lebensmittelforschung, Abteilung Mikrobiologie, D-85354 Freising, Germany

3

Zentrum fu¨r Molekulare Biologie, Universita¨t Heidelberg, D-69120 Heidelberg, Germany

Received 24 November 2004 Revised

22 February 2005

Accepted 7 March 2005

The expression of bacterial cold-shock proteins (CSPs) is highly induced in response to cold shock, and some CSPs are essential for cells to resume growth at low temperature. Bordetella bronchiseptica encodes five CSPs (named CspA to CspE) with significant amino acid homology to CspA of Escherichia coli. In contrast to E. coli, the insertional knock-out of a single csp gene (cspB) strongly affected growth of B. bronchiseptica independent of temperature. In the case of three of the csp genes (cspA, cspB, cspC) more than one specific transcript could be detected. The net amount of cspA, cspB and cspC transcripts increased strongly after cold shock, while no such effect could be observed for cspD and cspE. The exposure to other stress conditions, including translation inhibitors, heat shock, osmotic stress and nutrient deprivation in the stationary phase, indicated that the csp genes are also responsive to these conditions. The coding regions of all of the cold-shock genes are preceded by a long non-translated upstream region (59-UTR). In the case of the cspB gene, a deletion of parts of this region led to a significant reduction of translation of the resulting truncated transcript, indicating a role of the 59-UTR in translational control. The cold-shock stimulon was investigated by 2D-PAGE and mass spectrometric characterization, leading to the identification of additional cold-inducible proteins (CIPs). Interestingly, two cold-shock genes (cspC and cspD) were found to be under the negative control of the BvgAS system, the main transcriptional regulator of Bordetella virulence genes. Moreover, a negative effect of slight overexpression of CspB, but not of the other CSPs, on the transcription of the adenylate cyclase toxin CyaA of Bordetella pertussis was observed, suggesting cross-talk between the CSP-mediated stress response stimulon and the Bordetella virulence regulon.

INTRODUCTION Almost all single-celled organisms respond to a sudden decrease in temperature with a specific adaptive programme that results in a decrease of membrane fluidity, an increase in the negative supercoiling of DNA and an inhibition of translational initiation of bulk mRNA (Jones & Inouye, 1994). During this cold-adaptation phase, the transient induction of a subset of proteins termed cold-induced proteins (CIPs) is observed (Graumann & Marahiel, 1998; Gualerzi et al., 2003). Among these CIPs there is a family of 3Present address: Ingenium Pharmaceuticals AG, Fraunhofer Str. 13, D-82152 Martinsried, Germany. Abbreviations: CIP, cold-inducible protein; CSP, cold-shock protein; RT-qPCR, real-time quantitative PCR; 59-UTR, upstream untranslated region; VRG, virulence-repressed gene.

0002-7785 G 2005 SGM

Printed in Great Britain

small and mostly quite acidic proteins, the cold-shock proteins (CSPs). These CSPs mainly consist of the coldshock domain, which is found in many nucleic acid binding proteins of prokaryotic and eukaryotic organisms (Graumann & Marahiel, 1998). Escherichia coli K-12 encodes nine csp genes, but only four of them are strongly induced after cold shock (Wang et al., 1999). Accordingly, it is assumed that the other members of the CSP family are involved in the protection against stresses other than low temperature; e.g. transcription of the cspD gene is responsive to nutrient deprivation (Yamanaka et al., 1998). The response to cold shock has been extensively studied in E. coli, where a specific set of 26–27 CIPs is induced under these conditions (Gualerzi et al., 2003). The majority of these proteins bind nucleic acids and contribute to fundamental functions such as DNA packaging, transcription, RNA degradation, translation or ribosome assembly. 1895

D. Stu¨bs and others

The production of the major cold-shock protein CspA is transient and drops to a basal level after a few hours (Jones et al., 1987). CspA expression appears to be regulated mainly on the posttranscriptional level, e.g. by differential mRNA stability (Brandi et al., 1996). CspA acts as transcriptional antiterminator of certain genes that are upregulated in the cold (Bae et al., 2000) as well as an RNA chaperone, which prevents the formation of secondary structures in RNA, and thus ensures efficient translation of mRNA at low temperature (Jiang et al., 1997). Four hours after a cold shock, the cold-adaptation phase is completed, and regular protein synthesis and growth are resumed. Most members of the genus Bordetella are assumed to be obligatorily associated with host organisms (Gerlach et al., 2001; Weiss, 1992). Bordetella bronchiseptica causes chronic and often asymptomatic respiratory infections in a variety of animals, but only rarely in humans (Goodnow, 1980; Weiss, 1992). It is closely related to Bordetella pertussis and Bordetella parapertussis, the aetiological agents of whooping cough. Recently, comparative genomics revealed reductive genome evolution to occur in B. pertussis and the closely related pathogen B. parapertussis, which may have developed from a B. bronchiseptica-like ancestor (Parkhill et al., 2003). So far, only a single Bordetella species, Bordetella petrii, has been isolated from an environmental location (von Wintzingerode et al., 2001). Interestingly, B. bronchiseptica also has the capacity to survive harsh environmental conditions, and may therefore be able to exist outside a host organism (Porter et al., 1991). There appears to be an evolutionary trend in the pathogenic members of the genus Bordetella to lose factors possibly involved in survival outside of the host upon specialization to specific host organisms (Gerlach et al., 2001). Despite the reductive evolution during adaptation to an apparently obligate association with warm-blooded animals or man, the genes encoding cold-shock proteins are conserved among all Bordetella species. In fact, the family of CSPs consists of five members (CspA to CspE) and all five genes are found in the genomes of all Bordetella species sequenced so far (B. pertussis, B. parapertussis, B. bronchiseptica, Bordetella avium and B. petrii). In this study, the phenotype of csp mutants and the pattern of differential expression of the cold-shock genes in response to different stress conditions including cold shock, salt stress, heat shock and treatment with antibiotics were investigated. By proteome analysis, we show that B. bronchiseptica has a cold shock response similar to that of E. coli and other bacteria. In addition, we analysed the expression of cspB and obtained evidence for transcriptional control of its expression in B. bronchiseptica and for a posttranscriptional regulatory function of the long non-translated upstream region preceding the coding region of the cspB gene. 1896

METHODS Bacterial strains and growth conditions. The strains used in

this study are listed in Table 1. All Bordetella strains were grown at 37 uC either on Stainer–Scholte (SS) agar plates containing defibrinated horse blood [5 % (v/v)] or on charcoal (CC) agar plates (Difco). Liquid cultures of bordetellae were grown in SS broth (Stainer & Scholte, 1970), and E. coli strains were cultivated in Luria–Bertani (LB) broth. Antibiotics were used at the following concentrations: streptomycin sulfate, 100 mg ml21; ampicillin, 100 mg ml21; gentamicin, 15 mg ml21; kanamycin, 25 mg ml21; tetracycline, 30 mg ml21; chloramphenicol, 30 mg ml21. General techniques. DNA manipulations, isolation of chromosomal DNA and b-galactosidase assays were performed as described in standard protocols (Sambrook & Russell, 2001). DNA sequencing was done using the T7Sequencing Kit (Amersham Pharmacia Biotech) based on a variant of the dideoxynucleotide method (Sanger et al., 1977) as recommended by the manufacturer. For Western hybridizations, an ECL direct labelling and detection kit (Amersham) was used. Conjugational DNA transfer into B. pertussis and B. bronchiseptica was carried out as described previously (Gross & Rappuoli, 1988) with E. coli strain SM10 as a donor (Simon et al., 1983). Computer analysis. For the analysis of nucleotide and amino acid sequences, we used the HUSAR program package (http://genome. dkfz-heidelberg.de/), which is based on the GCG program package described by Devereux et al. (1984), the databases of the Sanger Institute (http://www.sanger.ac.uk/), and the pedant database of the Munich information center for protein sequences (mips; http:// pedant.gsf.de/). Construction of a B. bronchiseptica cspB deletion mutant.

Using the primer pairs cspB-KO-5-BamHI/cspB-KO-5-PstI and cspB-KO-3-PstI/cspB-KO-3-HindIII, two DNA fragments of 404 and 429 bp in size flanking the cspB gene were amplified from chromosomal B. bronchiseptica BB7865 DNA (Table 2). The amplified 404 and 429 bp DNA fragments were digested with BamHI/PstI and PstI/HindIII, respectively, and cloned into pBluescriptKS. Following linearization of the resulting construct with PstI, a kanamycinresistance cassette derived from pUC4K was cloned between the regions flanking the cspB gene on the chromosome. Both flanking sequences and the kanamycin cassette were then amplified by PCR using the primers cspB-KO-5-BamHI and cspB-KO-3-BamHI to introduce a second BamHI restriction site. The resulting BamHI fragment was then cloned into the allelic exchange vector pSS1129. The resulting construct, pSS1129-DcspB, was transferred by conjugation into B. bronchiseptica, and selection for double-crossover events was carried out as described elsewhere (Carbonetti et al., 1994; Stibitz & Yang, 1991). The resulting cspB deletion mutant was termed BB7865DcspB. Construction of cspB : : gfp fusions and determination of GFP fluorescence. A DNA fragment containing a promoterless gfp

gene was excised with XbaI and HindIII from plasmid pKEN2 and ligated into plasmid pBluescriptSK, giving rise to plasmid pSK-gfp. A 96 bp DNA fragment containing part of the upstream region of the cspB gene was amplified by PCR from chromosomal B. bronchiseptica DNA using the primer pair cspB-57-BamHI/cspB+39-XbaI. The resulting fragment was cloned into pSK-gfp to obtain plasmid pSK-I-gfp. The DNA fragment comprising the cspB upstream region fused to gfp was then excised with BamHI and HindIII, and subsequently ligated into plasmid pMMB206. E. coli SM10 was transformed with the resulting construct pMMB206-I-gfp and the construct was transferred into B. bronchiseptica wild-type strain BB7865 by conjugation. The plasmids pMMB206-II-gfp and Microbiology 151

Cold-inducible factors of B. bronchiseptica

Table 1. Bacterial strains and plasmids Strain or plasmid E. coli K-12 DH5a SM10 B. pertussis Tohama I Dtox B. bronchiseptica BB7865 BB7866 BB7865DcspB BB7865cspC : : Tn5 18323A4 Plasmids pBluescript pSL1180 pUC4K pKEN pMMB206 pSS1129 pLAFR2 pLAFR2-cspA pLAFR2-cspABC pLAFR2-cspB pLAFR2-cspC pLAFR2-cspD pLAFR2-cspE pSS1129-DcspB pMMB206-I-gfp pMMB206-II-gfp pMMB206-III-gfp pMMB206-IV-gfp pMMB206-V-gfp pMMB206-VI-gfp

Relevant feature

Source or reference

High-efficiency transformation Mobilizing strain

Gibco-BRL Simon et al. (1983)

Derivative of B. pertussis Tohama I, but rpsL and Dptx

Weiss & Falkow (1984)

Wild-type, but rpsL Phase variant of BB7865 with mutation in bvgS As for BB7865, but DcspB As for BB7865, but cspC : : Tn5 Derivative of B. pertussis 18323 carrying a fusion of lacZ with the PcyaA promoter

Arico et al. (1997) Monack et al. (1989) This study Schneider (2001) Goyard & Ullmann (1991)

High-copy-number cloning vector High-copy-number cloning vector Vector encoding a kanamycin cassette High-copy-number cloning vector containing a promoterless gfp gene Broad-host-range vector Vector for allelic exchange Broad-host-range vector pLAFR2 containing a 0?7 kb DNA fragment of B. bronchiseptica encoding cspA pLAFR2 containing a 0?4 kb DNA fragment of Bacillus cereus encoding cspA pLAFR2 containing a 0?5 kb DNA fragment of B. bronchiseptica encoding the cspB locus pLAFR2 containing a 0?9 kb DNA fragment of B. bronchiseptica encoding cspC pLAFR2 containing a 0?7 kb DNA fragment of B. bronchiseptica encoding cspD pLAFR2 containing a 0?5 kb DNA fragment of B. bronchiseptica encoding cspE pSS1129 containing the kanamycin gene flanked by the upstream and downstream sequences of cspB of B. bronchiseptica pMMB206 containing the cspB 59-UTR of B. bronchiseptica (257 to +39) fused to a promoterless gfp reporter gene pMMB206 containing the cspB 59-UTR of B. bronchiseptica (274 to +39) fused to a promoterless gfp reporter gene pMMB206 containing the cspB 59-UTR of B. bronchiseptica (2105 to +39) fused to a promoterless gfp reporter gene pMMB206 containing the cspB 59-UTR of B. bronchiseptica (2105 to +136) fused to a promoterless gfp reporter gene pMMB206 containing the cspB 59-UTR of B. bronchiseptica (2105 to +87) fused to a promoterless gfp reporter gene pMMB206 containing the cspB 59-UTR of B. bronchiseptica (2105 to +136) fused to a promoterless gfp reporter gene, but containing a mutated 9 bp box in the 59-UTR

Stratagene Pharmacia Amersham Cormack et al. (1996) Morales et al. (1991) Stibitz & Yang (1991) Friedman et al. (1982) This study This study; Mayr et al. (1996) This study

pMMB206-III-gfp were constructed in an analogous manner using the primer pairs cspB-74-BamHI/cspB+39-XbaI and cspB-105BamHI/cspB+39-XbaI, respectively. To investigate the role of the 59-UTR of the cspB gene, three additional plasmids, pMMB206-IV-gfp, pMMB206-V-gfp and pMMB206VI-gfp, were constructed. The construction of these plasmids was performed essentially as described above, using the primer pairs cspB105-BamHI/cspB+136-XbaI to generate construct IV and the primer pair cspB-105-BamHI/cspB+87-XbaI to obtain construct V. In the case of pMMB206-VI-gfp, the sequence of the conserved ‘9 bp box’ was changed to GGGGAATTC by site-directed mutagenesis. For this http://mic.sgmjournals.org

This This This This

study study study study

This study This study This study This study This study This study

purpose, a PCR was performed using the primer pair DelC/DelD containing the desired mutated sequence and the plasmid pMMB206IV-gfp as template. The oligonucleotides were extended by Pfu DNA polymerase (Promega) during temperature cycling (PCR programme: 95 uC/30 s; 18 cycles of 95 uC/30 s, 55 uC/1 min, 68 uC/10 min; 72 uC/ 10 min). Following PCR amplification the DNA fragment was treated with DpnI (New England Biolabs), and E. coli strain DH5a was then transformed with the nicked vector DNA. The mutated cspB gene fragment was digested with BamHI and HindIII, confirmed by DNA sequencing, and cloned into the pMMB206 vector for conjugation in B. bronchiseptica strain BB7865. FACS analysis was done as described previously (Schneider et al., 2002). 1897


Table 2. Oligonucleotides Name cspA-PE cspB-PE cspC-PE cspD-PE cspE-PE cyaA-PE bvgA-PE cspB-KO-5-BamHI cspB-KO-5-PstI cspB-KO-3-PstI cspB-KO-3-HindIII cspB-KO-3-BamHI cspB-5-KO-PCR cspB-3-KO-PCR cspA-I cspA-II cspB-I cspB-II cspC-I cspC-II cspD-I cspD-II cspE-I cspE-II tex for tex rev cspB for cspB rev gltA for gltA rev atpA for atpA rev sbp for sbp rev gmk for gmk rev groES for groES rev uspA for uspA rev pti for pti rev infB for infB rev cspB-105-BamHI cspB-74-BamHI cspB-57-BamHI cspB+39-XbaI cspB+136-XbaI cspB+87-XbaI DelC DelD Bc1 Bc2

1898

Sequence 59-TTC ACC AAG TTA CGG CTT TTC CC-39 59-GCA GCA ATT CAC CAA GTT GC-39 59-GTT CCA GTT CAC CAA GTT GC-39 59-GTC CCC TTG GAC AGC GGT TGC GG-39 59-CCC TTC TCG TCG TTG AAC CAC TTG-39 59-CGG CGG CGT TTG CGT AAC CAG CCT GAT GCG-39 59-ATA TGC TTT CAG TTT CG-39 59-GGC GGA TCC CAG CCA CCT GCT TGC CGA AC-39 59-GGC CTG CAG CTG CCG TGC CAC GAT GC-39 59-GGC CTG CAG GCA CTT CTC GGA AAT-39 59-GGC AAG CTT GAA TTG ATC CAG CAG-39 59-GGC GGA TCC GAA TTG ATC CAG CAG-39 59-GGT CAT GCG CAC ACC TTA TAA GC-39 59-CGT CGC CTG AGG ACC CTT CGG-39 59-GGC GGA TCC GCT GCA CGC CGA AAC CTT GC-39 59-GGC GGA TCC TCA AGC GAT GAC TTT TAT GT-39 59-GGC GGA TCC CCG ACA TTC ACG CGG CGG-39 59-GGC GGA TCC GCT TCG AAA CGC ATG G-39 59-GGC GGA TCC ACC TGT CGG ATC CAG GTG CCG-39 59-GGC GGA TCC GCA TGT AGG GAT CCT GGT TGT GCC-39 59-GGC GGA TCC GGT TCC ATA ATG CAC TGC AG-39 59-GGC GGA TCC TCA GGC AGA CGT AAT ATT GA-39 59-GGC GGA TCC CGA TCT GTT CGG GTG CAT GC-39 59-GGC GGA TCC TCA CAA CGG GGT GAT CTT GG-39 59-TGC AGG ATC CTT TGG CCG AAC T-39 59-GTT CAC GCA GTC CTC GAC CA-39 59-TTA TGG CTT CAT CAC CCC-39 59-TGG ACG CGC TCA GAT TCT TA-39 59-TGG AAT ACA CCA TCG TGG T-39 59-AGG CCT GCT TGG TCA GGT CGT-39 59-ACA AGT ATT CGC AAG GCC A-39 59-TGC AGG ATG AAG ATG CGA T-39 59-TCA TCA TGG CGA CCT TGA-39 59-TCG ATG GCC TCG AAA AGC T-39 59-ATG CGC AAC AGC TGC TCG AAT-39 59-AGG ACG AAT ATG CCG AT-39 59-AAG CCG GAT CAA GGT GAA GT-39 59-ATG ACG AGC AGC TCT T-39 59-ACA TCG ACA AGT CCA TGC T-39 59-TGC CGA TGA AGA TCA GGT-39 59-ACC TGG TGT CCG GAG AAT TTC GA-39 59-AAC TTG TTC AGG TCC GTC AGG A-39 59-TGA ATC TCC CAA GAG CGC CAA-39 59-TTG CGC GGA CGG CTG AGC ATC T-39 59-GGG GGA TCC GCG CAA AAA ATA AGC ATC GTG-39 59-GGG CCC GGA TCC GTG TGT TTT AAT ATT GGC ACG-39 59-GGG GGA TCC TTA TGT TGA CTT GTG AGG G-39 59-GGG TCT AGA GGC ACT GAA CAA ACA TTG CAA CG-39 59-GGG TCT AGA CTG TAT TTC CTA GAA ATT AAT GGG-39 59-GGG GAA TTC CCC GCC TCG CGC CAG AAC GC-39 59-CCG GGA CCA CGA ATC GAA TTC CCC CGG GGC A-39 59-TTC TAC TGC CCC GGG GGA ATT CGA TTC G-39 59-GGA TTG TGG ATC CTA TTT TAA G-39 59-AAA AAA GGA TCC CAT AAG CCA TC-39

Microbiology 151


Construction of pLAFR2 derivatives containing different csp genes. The coding regions of five cold-shock genes were amplified

together with their 59-UTR and their entire promoter regions from chromosomal B. bronchiseptica DNA by PCR using the following primer pairs: cspA-I/cspA-II, cspC-I/cspC-II, cspD-I/cspD-II, cspE-I/ cspE-II (Table 2). B. pertussis chromosomal DNA was used as a template to amplify cspB with the oligonucleotides cspB-I/cspB-II. The cspA gene of Bacillus cereus was amplified with the oligonucleotides Bc1 and Bc2. The PCR fragments were digested with BamHI and cloned into the BamHI-digested low-copy vector pLAFR2. The resulting constructs were conjugated into BB7865 as described above. RNA isolation. Whole-cell RNA of Bordetella strains was isolated

as described previously (Gross & Rappuoli, 1989). Briefly, bacteria grown under appropriate conditions were harvested in the midexponential phase (OD590 0?6) or in the stationary phase (OD590 1?2), and then lysed by the addition of a solution containing 100 mM Tris/HCl (pH 7?5), 1 % SDS and 2 mM EDTA. To test the response to several stress conditions, cells were grown to OD590 0?6 and then subjected to the following treatment: for heat shock the cultures were incubated at 48 uC for 20 min, for salt shock 2 M NaCl was added and the cultures were further incubated for 30 min. Finally, to test the response to translation inhibitors, tetracycline (50 mg ml21) or chloramphenicol (120 mg ml21) was added to the cultures, and cultivation was continued for 30 min. After boiling, the samples were kept on ice in the presence of 80 mM KCl for 5 min. This step was followed by the removal of cellular debris by centrifugation. To 3?5 ml of each sample, 4?56 g CsCl was added, and the RNA was pelleted by centrifugation in a Beckman SW50Ti rotor at 35 000 r.p.m. at room temperature for 20 h. Finally, the pellet was resuspended in 500 ml TE buffer [10 mM Tris/HCl (pH 8?0), 1 mM EDTA], extracted with an equal volume of a solution of 1 : 1 phenol/chloroform (v/v), precipitated once with ethanol, resuspended in 50 ml TE buffer and finally stored at 220 uC. Primer extension analysis. For the various csp genes, and the

cyaA and bvgA genes, this was done with the appropriate PE primers listed in Table 2. The oligonucleotides were labelled with [c-32P]ATP using T4 polynucleotide kinase, then hybridized with their respective templates. Primer extension reactions were performed with 25 mg total RNA for 45 min at 46 uC using avian myeloblastosis virus reverse transcriptase (Amersham). Primer extension products were resolved on polyacrylamide gels (6 %) containing 6 M urea. Sequencing reactions were done in parallel using the various pLAFR2 derivatives containing the respective csp genes as a template. Radiolabelled bands were quantified using a PhosphorImager system (Amersham Biosciences) and the corresponding software (Molecular Dynamics).

SS medium to OD590 0?6. Rifampicin was then added to a final concentration of 0?25 mg ml21, and the culture was divided into two equal samples, one of which was further incubated at 37 uC and the other was cold shocked in an ice bath and further incubated at 15 uC. At different time points, total RNA was isolated from 50 ml aliquots of the cultures as described above. Preparation of protein extracts and 2D protein gel electrophoresis. Cultures (50 ml) of B. bronchiseptica were grown to late

mid-exponential phase (OD578 1?0) and cold shocked as described above. At different time points after cold shock the cells were harvested by centrifugation. Equal amounts of up to 150 mg protein were separated by 2D-PAGE according to the method of O’Farrell (1975). Briefly, the proteins were separated in the first dimension by isoelectric focusing in a pH gradient ranging from 3 to 10; in the second dimension, the proteins were separated according to their molecular masses by SDS-PAGE, and visualized by Coomassie blue or silver staining. The cellular proteins of each tested strain were isolated in at least three separate experiments, and 2D-PAGE of each protein preparation was performed at least in triplicate. Only those protein spots that showed a reproducible and significant increase in their amount after cold shock were further analysed by mass spectrometry as described previously (Kimmel et al., 2000). Evaluation and quantification of the 2D gels were performed using the DELTA2D program from Decodon.

RESULTS The family of B. bronchiseptica CSPs comprises five members Within the genome sequence of B. bronchiseptica RB50 (Parkhill et al., 2003), five ORFs were identified exhibiting a highly significant amino acid similarity to the major coldshock protein CspA of E. coli (Fig. 1). The five putative CSPs show a sequence identity of 65–80 % at the amino acid level among each other and of 63–73 % with CspA of E. coli. According to their decreasing similarity with E. coli

RT-qPCR. For each reaction several dilutions of cDNA reverse transcribed from 5 mg total RNA with random primers and the

SuperScriptII RNaseH Reverse Transcriptase (Invitrogen) were used, with 1 pg to 2 ng of genomic B. bronchiseptica BB7865 DNA as a standard. RT-PCR reactions were performed using a DNA Engine Opticon System (MJ Research) and the qPCR Core Kit for Sybr Green I (Eurogentec) according to the manufacturer’s recommendations. From each of two independent RNA preparations two separate cDNA preparations were made. These cDNA preparations were characterized three times each by RT-qPCR, resulting in an overall replicate number of 12 experiments. The threshold cycle number was set to a fluorescence of 0?01. For this analysis the ‘for/ rev’ primers listed in Table 2 and named according to the respective genes were used. Rifampicin treatment. To investigate the half-life of transcripts,

the B. bronchiseptica wild-type strain BB7865 was grown at 37 uC in http://mic.sgmjournals.org

Fig. 1. Amino acid sequences of CspA, CspB, CspC, CspD and CspE. The CLUSTALW program was used to align the sequences. Black boxes, amino acids that are identical in at least four proteins; grey boxes, similar or identical amino acids in at least three of the Csp proteins. The groups of amino acids considered similar in this analysis were I, L, M and V; A, S, G, P and T; H, K and R; D, E, N and Q; F, W and Y; and C. Dashes mark the gaps placed in the sequences by the CLUSTALW program to optimize alignment. E.c., E. coli; B.b., B. bronchiseptica. 1899


CspA, the Bordetella CSPs were termed CspA, CspB, CspC, CspD and CspE. The calculated molecular masses of the B. bronchiseptica CSPs range from 7?3 (CspC) to 8?6 kDa (CspD). In comparison to the E. coli CspA protein (isoelectric point, pI=4?3), the pI values of the CSPs of B. bronchiseptia are significantly higher, with an unusually high pI value of 9?15 for CspD. The genomic organization of the CSP-encoding genes in B. bronchiseptica and the prediction of promoter and termination sites indicate that all of them are transcribed monocistronically (data not shown). The cspC, cspD and cspE genes were found to be scattered over the chromosome, while the cspA and cspB genes are clustered and separated only by a single ORF encoding a putative protein of unknown function (data not shown). Phenotypic characterization of mutants in the cspB and cspC genes In E. coli, the deletion of single csp genes did not result in an obvious phenotype, possibly due to back-up mechanisms based on the presence of closely related CSPs (Xia et al., 2001). In B. bronchiseptica, we investigated the phenotypic effects of mutations in the cspB and cspC genes. For this purpose we constructed a cspB deletion mutant by replacement of the gene with a kanamycin-resistance cassette, resulting in strain BB7865DcspB. In addition, we investigated a transposon-insertion mutant in the cspC gene described previously. The mutant, BB7865cspC : : Tn5, carries a transposon inserted at the beginning of the cspC coding region (Schneider, 2001). The wild-type strain and

the two isogenic mutants were cultivated at 37 and 15 uC. The generation time of each strain at the different growth temperatures was then calculated. The wild-type strain and the cspC mutant could not be distinguished from each other, and showed a generation time of 0?9 and 3?2 h at 37 and 15 uC, respectively. Since cspC is expressed maximally under modulating conditions inactivating the BvgAS twocomponent system (see below), the experiments were also performed in the presence of modulating agents such as MgSO4 and nicotinic acid, but again no difference between the wild-type and the cpsC mutant could be observed. In contrast, the cspB mutant grew much more slowly under all growth conditions tested and a generation time of 3?7 and 12?6 h could be calculated at 37 and 15 uC, respectively (data not shown). Thus, the generation time of the cspB mutant was prolonged about fourfold independently of the temperature conditions applied. Transcript analysis of cold-shock genes To characterize the promoters of the five B. bronchiseptica csp genes and to analyse the effect of cold stress on their transcriptional response, we conducted primer extension experiments with RNA preparations from wild-type strain BB7865 cultivated at 37 uC or cold shocked for 30 min at 15 uC. All five B. bronchiseptica cold-shock genes comprise extremely long 59-UTRs ranging from 109 bp (T1cspD) to 167 bp (T2cspB), a feature that appears to be characteristic for many of the cold-shock genes of eubacteria (Fig. 2) (Bae et al., 1997; Brandi et al., 1996; Fang et al., 1997; Jones &

Fig. 2. Nucleotide sequences of the promoter and upstream regions of the five csp genes of B. bronchiseptica. The start sites of the transcripts were determined by primer extension experiments and are marked by arrows, and in case of more than one transcript, the corresponding number. Putative ribosome-binding sites (Shine–Dalgarno sequences) are underlined and the conserved sequence motif present in the 59UTR of four of the five csp genes, the 9 bp box, is shaded. 1900

Microbiology 151


Inouye, 1994). Within the 59-UTRs of four of the csp genes (cspA, cspB, cspC and cspE), a highly conserved 9 bp sequence motif, the 9 bp box, with the consensus sequence 59-TCCTTGATT-39 was identified (Fig. 2). In all cases, this motif was found to be located approximately 51–57 bp upstream of the respective start codon. This sequence motif does not show any similarity with the ‘cold box’ present in the 59-UTR of transcripts of the csp genes of E. coli, for which a regulatory function was suggested previously (Jiang et al., 1996).

of a precursor mRNA generated under the various stress conditions. In contrast to the cspA, cspB and cspC genes, in the case of the cspD and cspE genes a single transcript could be detected, and the amount of these transcripts was not altered significantly after cold shock (Fig. 3 and data not shown). A fourfold to ten-fold increase of total transcript following cold shock was determined for each of the coldinduced csp genes by phosphorimager quantification of the primer extension products and by RT-qPCR (data not shown). To narrow down the temperature spectrum leading to induction of the cold-shock stimulon, we investigated the enhancement of the amount of the cspB transcript after a temperature downshift from 37 uC to 30, 26, 20 and 15 uC. A significant twofold cold-induced accumulation of the cspB T1 transcript could already be observed at a temperature of 26 uC but maximal transcript accumulation occurred only at a temperature of 20 uC or below (data not shown).

In the case of the cspA, cspB and cspC genes more than one transcript could be detected. In the case of the cspB gene, two transcripts were identified, which according to the deletion analysis of the promoter region of this gene are likely to be transcribed from two independent promoters (see below) (Fig. 3). The main transcript of the cspB gene at 37 uC starts 167 nucleotides upstream of the start codon (T2cspB), while a strong increase in the amount of the slightly smaller T1cspB transcript and, concomitantly, a weak decrease of the T2cspB transcript was observed after cold shock (Fig. 3). The amount of mRNA reached a maximum between 60 and 90 min after the cold-shock treatment and decreased after 150 min (data not shown). A temperature-dependent change in the amount of transcripts was also observed for the cspA gene, which revealed two strongly enhanced transcripts at 15 uC as compared to 37 uC (T3cspA and T1cspA), and a third transcript (T2cspA) that was not significantly affected by the temperature decrease (Fig. 3). In contrast to the cspA and cspB transcripts, only low amounts of cspC-specific transcripts (T1-3cspC) were detected in bacteria cultivated at 37 uC. A temperature downshift resulted in the accumulation of two of these transcripts, T1cspC and T2cspC. It is not clear so far whether these transcripts are transcribed from different promoters or whether some of them are specific degradation products

cspA CS A C GT+ _ + _

cspB CS A C G T + _+_

cspC CS A C GT + _ + _

T3

T2 T1

T3 T2 T1

cspD A C G T +

To examine whether stress conditions other than cold shock also affect csp transcription in B. bronchiseptica, whole-cell RNA was prepared from bacteria grown under salt stress caused by 2 M NaCl, after antibiotic treatment with the translation inhibitors chloramphenicol and tetracycline, and after heat shock to 48 uC as described in Methods. The previously identified cspA transcripts showed a complex response pattern to the different stress conditions. The T2cspA transcript was only slightly affected by all of the stress conditions applied, while the transcripts T1cspA and T3cspA strongly accumulated in the presence of the two translation inhibitors (Fig. 4). In addition, the amount of the T3cspA transcript strongly increased after heat shock. In the case of the cspB gene the transcript T2cspB was found to be independent of all tested stress conditions, while the transcript T1cspB accumulated in response not only to cold shock, but also to the translation inhibitors (Fig. 4). Two of the three cspC-specific transcripts, T1cpsC and T2cspC, were found to

_

CS _

cspE A C G T +

+

T2 T1

_

CS _

+

T1 T1

http://mic.sgmjournals.org

Fig. 3. Determination of the transcript 59ends of cspA, cspB, cspC, cspD and cspE in B. bronchiseptica by primer extension experiments. RNA was prepared from wildtype strain BB7865 (+) and the bvg mutant BB7866 (”) grown at 37 6C, and from the same culture after cold shock (CS) to 15 6C for 30 min. The corresponding DNA sequence reactions are shown on the left for each gene. The various transcripts (T) identified by primer extension were numbered according to their increasing distance to the ATG start codon of the respective gene and are indicated by arrows. 1901


accumulate under cold shock, while only the T2cspC transcript was also affected by the translation inhibitors (Fig. 4). With the exception of the T3cspC transcript, which showed a mild response to osmotic stress, none of the other transcripts of the various csp genes were affected by salt stress (Fig. 4). Moreover, the growth phase had little effect on transcription of the csp genes. In fact, a significant effect of the growth phase could be detected only in the case of the cspC transcript T3 and the respective signal was maximal in the stationary phase at an OD578 of 1?2 (data not shown).

Amount of transcript

We also investigated cold induction of cspB gene expression in B. pertussis by primer extension analysis. Since the results were virtually identical in the two species (data not shown), they suggest that an analogous cold-shock response is also occurring in B. pertussis despite its reductive genome evolution.

˚C Te t C m N aC l 37 ˚C 15 ˚C 48 ˚C Te t C m N aC 37 l ˚C 15 ˚C 48 ˚C Te t C m N aC l

˚C

48

15

37

˚C

cspA

T2

T3


T1

l aC N

m C

t Te

˚C

˚C

˚C 48

15

l

T2


T1

37

m

t

aC N

C

Te

˚C 48

˚C 15

37

˚C

cspB

Te t C m N aC 37 l ˚C 15 ˚C 48 ˚C Te t C m N aC 37 l ˚C 15 ˚C 48 ˚C Te t C m N aC l

˚C

48

˚C

15

37

˚C

cspC

T1

T2

T3

Fig. 4. Relative amounts of csp specific transcripts (T1–T3) of the cspA, cspB and cspC genes of strain BB7865 grown at different growth and stress conditions. The transcript numbers (T1–T3) are the same as in Fig. 3. The different growth conditions (see Methods) are indicated below the respective bars: normal growth at 37 6C; cold shock at 15 6C; heat shock at 48 6C, presence of translation inhibitors tetracycline (Tet) and chloramphenicol (Cm); osmostress caused by 2 M NaCl. The means of the amounts of transcripts detected in four independent primer extension experiments are shown; error bars indicate SD. 1902

Involvement of the upstream region of the cspB gene in transcriptional regulation of cspB expression To investigate the function of the upstream region of cspB in the transcriptional control of its expression, the wildtype sequence up to position 2105 as counted from T1, and a deletion derivative lacking parts of the upstream region between position 2105 and 274, were cloned in front of a promoterless gfp gene, resulting in the constructs pMMB206-I-gfp and pMMB206-II-gfp, respectively (Fig. 5a). Primer extension analysis using a gfp specific oligonucleotide demonstrated that both cspB specific transcripts, T1 and T2, were significantly reduced in the deletion variant. The further deletion of the upstream region up to sequence position 257 in the construct pMMB206III-gfp resulted in the complete disappearance of the T2 transcript, while the amount of the T1 transcript was not further diminished (Fig. 5b). Interestingly, the deletion of parts of the upstream region did not affect the coldinducibility of the T1 transcript, which still accumulated at the low temperature even in cells harbouring the construct pMMB206-III-gfp, although the general expression level was much lower as compared to that observed with the pMMB206-I-gfp construct (Fig. 5b). This suggests that synthesis of both transcripts of the cspB gene is under transcriptional control by an unknown factor interacting with the upstream region of the cspB gene; however, the upstream region is not essential for the cold induction of the T1 transcript. Moreover, the presence of the T1 transcript also in the strain harbouring pMMB206-III-gfp in which the T2 transcript disappeared indicates that cspB is transcribed from two independent promoters. Posttranscriptional control mechanisms affect the expression of the cspB gene The primer extension experiments showed that all of the csp genes are preceded by a long 59-UTR. Within this nontranslated region of the three cold-responsive genes cspA, cspB and cspC of B. bronchiseptica, a nucleotide sequence with the consensus 59-TCCTTGATT-39 is present, while the 59-UTRs of the other two cold-shock genes, which are not induced at 15 uC, carry a sequence with one mismatch within this 9 bp box (cspE) or lack the entire box (cspD) (Fig. 2). The conservation of this sequence and its similar location in the 59-UTR of the csp genes suggested that it might play a role in the differential regulation of expression of the Bordetella cold-shock genes. In order to determine regions of the transcripts possibly involved in the regulation of cspB expression at low temperature, various DNA fragments comprising different parts of the upstream region of the ORF encoding CspB were fused to a promoterless gfp reporter gene (Fig. 6a). The control fragment contains the entire 59-UTR up to position +136 with respect to the transcriptional start site T1cspB. In the second construct, the 59-UTR was shortened at its 39 side by 49 nt, resulting in a truncated 59-UTR that ends at position +87 and lacks the 9 bp box. The third construct comprised the entire 59-UTR, Microbiology 151

T2

pMMB206-I-gfp -105

T1

-30 +1 T2 T1

pMMB206-II-gfp -74

-30

+1 T1

-30

+1

pMMB206-III-gfp -57

fp I -g

37 15 37 15 37 15 ˚C

gfp +39

gf p

(b)

(a)

II-

III-

gf p


T2 gfp

+39 gfp +39

T1

Fig. 5. (a) Deletion analysis of the region upstream of the transcriptional start sites of the cspB gene. On the top, the fulllength fragment cloned in the vector pMMB206 resulting in the construct pMMB206-I-gfp is shown. The upstream and downstream regions of the cspB gene are counted from the start site of transcript T1. The promoterless gfp gene and its Shine–Dalgarno sequence is fused to the 59-UTR of the cspB gene at position +39. In the construct pMMB206-II-gfp the upstream region is shorter and ends at position ”74. In the third construct pMMB206-III-gfp the upstream region was further shortened and ends at position ”57. T1 and T2, transcriptional start sites. (b) Primer extension experiments carried out with whole-cell RNA prepared from strain BB7865 grown either at 37 6C or cold shocked at 15 6C carrying the constructs shown in (a). I-gfp, BB7865/pMMB206-I-gfp; II-gfp, BB7865/pMMB206-II-gfp; III-gfp, BB7865/pMMB206-III-gfp. T1 and T2 indicate the cspB-specific transcripts shown in Figs 3 and 5(a).

To address the question of whether additional posttranscriptional mechanisms contribute to the regulation of cspB expression, the half-life of the cspB transcripts was determined in B. bronchiseptica cultures grown at 37 uC and in cold-shocked cultures by addition of rifampicin and subsequent detection of the transcripts by RT-qPCR and primer extension analysis. RT-qPCR revealed an increase of about 10-fold in the half-life of the cspB transcripts in cultures exposed to low temperature as compared to the culture grown at elevated temperature. In agreement with this finding, the primer extension analysis also showed that the half-life of the T1 and T2 transcripts is increased about 10to 15-fold at 15 uC as compared to their half-life at 37 uC http://mic.sgmjournals.org

(Fig. 7). The increase in the half-life of these transcripts appears to be specific for the cspB gene, since the stability of other transcripts, for instance that of the guanylate kinase

(a) T2

pMMB206-IV-gfp -105

T1

9-bp box

-30 +1 T2 T1

pMMB206-V-gfp -105

-30 +1 T2 T1

pMMB206-VI-gfp -105

gfp

+136 gfp +87 9-bp box

-30 +1

+136

gfp

(b) Percentage GFP expression

but the sequence of the 9 bp box was changed to the unrelated sequence 59-GGGGAATTC-39 by site-directed mutagenesis. Transcripts deriving from the three different constructs were characterized by primer extension analysis of whole-cell RNA isolated from BB7865 derivatives harbouring these plasmids. No significant differences could be observed in the amount of the transcripts from the various constructs at low and high temperature (data not shown). However, the amount of GFP synthesized in the strains harbouring the different constructs showed interesting differences as determined by FACS and by Western blot analysis with a GFP-specific antiserum. A very significant reduction in GFP expression was observed in the strain carrying the deletion construct (pMMB206-V-gfp) as compared to the strain harbouring the wild-type construct (pMMB206-IV-gfp). Interestingly, a weaker but still significant reduction in GFP expression (P¡0?03) can also be observed in the strain carrying the mutated 9 bp box (pMMB206-VI-gfp) (Fig. 6b). This indicates that the 59-UTR is involved in the control of translation efficiency of GFP expression and that the 9 bp box may be a relevant sequence motif required for maximal translation efficiency.

100 80 60 40 20 2

4 6 Time (h)

8

10

Fig. 6. Deletion analysis of the 59-UTR of the cspB gene to investigate a possible role of the 59-UTR in translation efficiency. (a) Plasmid pMMB206-IV-gfp contains the entire promoter region and the entire 59-UTR region of the cspB gene cloned in front of the promoterless gfp gene. Construct pMMB206-V-gfp has a deletion of 49 bp from the 39 end of the 59-UTR destroying part of the 9 bp box. Construct pMMB206-VI-gfp is a derivative of pMMB206-IV-gfp made by site-directed mutagenesis with a changed 9 bp box sequence. (b) GFP expression at low temperature from the three constructs after conjugation in strain BB7865. GFP expression was determined by FACS analysis as described in Methods, and the means±SD of at least three independent experiments are shown. X, BB7865; &, BB7865(pMMB206-IV-gfp); m, BB7865(pMMB206-V-gfp); %, BB7865(pMMB206-VI-gfp). 1903


Proteome analysis of the cold-shock response

100 37 °C

Percentage of transcript

75 50 25 5

10

15

20

25

30

100 15 °C

75 50 25 5

10

15 20 Time (min)

25

30

Fig. 7. Half-life of the T1 (&) and T2(X) transcripts at 37 (top) and 15 6C (bottom) as determined by primer extension experiments of rifampicin-treated cultures (see Methods for details). The means±SD of three independent experiments are shown.

encoding gmk gene of B. bronchiseptica, was not altered significantly under the same experimental conditions (data not shown).

Protein expression during cold adaptation was investigated by the 2D-PAGE of proteins from B. bronchiseptica cultures harvested at various time points after cold shock. As expected, an elevated synthesis of distinct cytosolic proteins could be observed upon cold shock (Fig. 8). Eighteen different protein spots (CIPs) were excised from preparative gels loaded with cell extracts of bacteria that were cultivated at 15 uC for up to 3 h. The proteins were then identified by mass spectrometry (Table 3). Interesting examples of these cold-inducible proteins include a chaperone (Cpn10) with similarities to GroES and the putative translation inhibitor BB2940 (Table 3). Among these proteins we also found the cold-shock proteins CspA, CspB and CspC, as well as a low molecular mass protein with significant similarities to the universal stress protein UspA of E. coli, which is known to modify at least six other E. coli proteins in response to stress and starvation (Nystro¨m & Neidhardt, 1996). Other B. bronchiseptica CIPs apparently are involved in amino acid metabolism, such as the LysA diaminopimelate decarboxylase, the IlvC and TrpE proteins involved in isoleucine and tryptophan biosynthesis, and the 5-oxoprolinase OplaH. Two cold-inducible proteins, the citrate synthase GltA and the ATP synthase a-chain AtpA, are involved in respiration. Interestingly, the superoxide dismutase SodB

Fig. 8. Representative silver stained 2D-PAGE gels are shown. The gels were loaded with 75 mg protein of whole-cell lysates prepared from cells grown at 37 6C (0 h) and further incubated under cold-shock conditions for 1, 2 or 3 h. Thirteen protein spots with increased amount after cold shock are marked by arrows. The identities of the corresponding proteins are listed in Table 3. The gels were evaluated using the DELTA2D software (Decodon). 1904

Microbiology 151


Table 3. Factors of B. bronchiseptica induced by cold shock Gene product Universal stress protein A Cpn10 CspA CspB CspC RNA binding protein Tex Superoxide dismutase IF-2 Putative translation inhibitor Ketol-acid reductoisomerase Diaminopimelate decarboxylase Citrate synthase c-Glutamyl transpeptidase precursor 5-Oxoprolinase Anthranilate synthase component Probable carboxymuconolactone decarboxylase Phosphoglycerate mutase 1 Solute-binding protein Putative peptidase ATP synthase a-subunit

Synonym

Gene

Spot*

Detected byD

BB4260 BB0963 BB2251 BB2249 BB5013 BB3601 BB2012 BB3246 BB2940 BB0791 BB0068 BB3677 BB0978 BB2978 BB4625 BB0458 BB0398 BB0185 BB3096 BB4607

uspA cpn10 cspA cspB cspC tex sodB infB

2 3

RT-qPCR, 2D-PAGE RT-qPCR, 2D-PAGE PE, 2D-PAGE PE, RT-qPCR, 2D-PAGE PE, 2D-PAGE RT-qPCR PE, 2D-PAGE RT-qPCR RT-qPCR, 2D-PAGE 2D-PAGE 2D-PAGE RT-qPCR, 2D-PAGE 2D-PAGE 2D-PAGE 2D-PAGE 2D-PAGE 2D-PAGE RT-qPCR, 2D-PAGE 2D-PAGE RT-qPCR, 2D-PAGE

1

ilvC lysA gltA ggt oplaH trpE

5 8 7 9 13 11

gpmA

4 6 12 10

atpA

*The spot number refers to the protein spots in Fig. 8. DPE, primer extension.

In addition to the proteins identified by 2D-PAGE, we investigated the cold-inducibility of IF-2, which in E. coli is strongly induced after cold shock. The recently identified RNA-binding protein Tex was also included in this analysis because it contains an S1 domain (Fuchs et al., 1996; Ko¨nig et al., 2002). S1 domains are evolutionarily related to coldshock domains since both types of domains show a similar three-dimensional structure (Bycroft et al., 1997). As demonstrated by quantitative RT-PCR, the expression of the genes encoding IF-2 and Tex was found to be strongly induced upon cold shock (data not shown), classifying them as CIPs in B. bronchiseptica (Table 3). Interference of the cold-shock stimulon and the virulence regulon of B. bronchiseptica Transcriptional gfp fusions have been generated previously in B. bronchiseptica, by means of transposon mutagenesis, resulting in the identification of several gene loci that were negatively regulated by the virulence regulatory BvgAS two-component system, the so-called virulence-repressed genes (VRGs) (Knapp & Mekalanos, 1988; Schneider et al., 2002). Here we show that two of the CSPs also belong to the vrg regulon of B. bronchiseptica. In fact, primer extension http://mic.sgmjournals.org

experiments performed with mRNA isolated from the bvg-negative (phase variant) strain BB7866 revealed an accumulation of the T3cspC and T1cspD transcripts (Figs 3 and 9). The amount of these cspC- and cspD-derived transcripts was found to be significantly increased in the bvg mutant grown at 37 uC as compared to the wild-type strain, indicating that BvgAS is involved in the repression of CspC and CspD at high temperature. To confirm the impact of the two-component system in the regulation of these two


was found to be cold inducible in B. bronchiseptica, as well as the c-glutamyltranspeptidase precursor encoded by the ggt gene, which is involved in the control of glutathione metabolism (Table 3).

cspA

cspB

cspC

+_ + _ + _ T1 T2 T3

+ _ + _ T1 T2

+ _+ _ +_ T1 T2 T3

cspD cspE

300 200 100 +_ T1

+ _ T1

Fig. 9. BvgAS-dependent transcription of B. bronchiseptica csp genes investigated by primer extension analysis. The columns represent the relative amount of different transcripts of the five csp genes in the wild-type strain BB7865 (+) and in the bvg mutant strain BB7866 (”). At least three independent experiments were carried out for each csp gene, and the error bars indicate SD. 1905


loci, we investigated the influence of the modulating compounds nicotinic acid and MgSO4, known to be recognized by BvgS and leading to the inactivation of this histidine kinase. The amount of the transcripts increased significantly under these conditions in the wild-type strain BB7865, thus confirming the classification of cspC and cspD as members of the vrg regulon (data not shown). In the case of the BvgAS-repressed cspC-specific transcript T3, this effect is particularly prominent in bacteria harvested in stationary phase at an OD578 of 1?2 (data not shown). In contrast, as mentioned above, none of the other csp transcripts were found to be significantly affected by the transition to the stationary growth phase of the bacteria. These data suggested the possibility that the CSPs may also play a role in the regulation of Bordetella virulence genes. To investigate this question, we cloned four B. bronchisepticaderived cold-shock genes, cspA, cspC, cspD and cspE, the cspB gene of B. pertussis, and the cspA orthologue of Bacillus cereus into the low-copy vector pLAFR2. A non-haemolytic phenotype of transconjugants of both B. pertussis and B. bronchiseptica was observed with plasmid pLAFR2cspBBP, but not after introduction of the other csp genes. This demonstrates a negative regulatory effect of CspB overexpression on the expression of the adenylate cyclase toxin CyaA. In fact, a B. pertussis strain harbouring a fusion of the cyaA promoter to the lacZ reporter gene (Goyard & Ullmann, 1991) showed an approximately fourfold decrease in b-galactosidase activity after introduction of the additional cspB copy as compared to the wild-type (data not shown). In agreement with the lacZ reporter gene data, cyaA expression investigated by primer extension analysis was found to be significantly lower in the strain carrying the pLAFR2-cspB plasmid than in the control strain harbouring pLAFR2 only. This effect appears to be specific for cyaA expression since no alteration of the expression of the bvgAS two-component system could be observed by primer extension analysis in bacteria cultivated at 37 uC in the presence of the additional cspB gene copies (data not shown).

DISCUSSION Cold-inducible genes have been identified in a wide variety of psychrophilic and mesophilic bacteria (Graumann et al., 1997; Wouters et al., 1998), and are known to play an essential role in the adaptation to low temperature. Genomic sequencing revealed that the three mammalian pathogens B. bronchiseptica, B. parapertussis and B. pertussis encode five proteins homologous to CspA of E. coli (Parkhill et al., 2003). Orthologues of the same five proteins are also encoded by the bird pathogen B. avium (http://www. sanger.ac.uk/Projects/B_avium/) and by B. petrii, the only known environmental species of the genus Bordetella (R. Gross, C. Guzman, U. Go¨bel & H. Blo¨cker, unpublished results; von Wintzingerode et al., 2001). Therefore, despite the reductive genome evolution that has streamlined the genetic equipment of the specialized pathogens B. pertussis, 1906

B. parapertussis and B. avium, all bordetellae conserved the same set of CSPs, indicating a relevant role for these factors also during infection. Although CSPs may be endowed with activities required for growth under optimal environmental conditions, their specific function during stress adaptation is not well understood. Even the production of knock-out mutants did not contribute much to the understanding of their specific function, possibly because of compensatory effects by related cold-shock proteins. For example, in Bacillus subtilis all three csp genes must be deleted to obtain a lethal phenotype (Graumann et al., 1997), while four out of nine csp genes had to be inactivated before a coldsensitive phenotype was achieved in E. coli (Bae et al., 1997; Xia et al., 2001). In contrast to these data, the single cspB deletion in B. bronchiseptica led to a growth-impaired phenotype at low (15 uC) and elevated temperature (37 uC). Thus, CspB appears to play an important role in the normal growth of the bacteria independent of temperature. It has been observed that the cold-shock response can be induced by antibiotics like chloramphenicol and tetracycline, suggesting that ribosomes can act as temperature sensors in E. coli (VanBogelen & Neidhardt, 1990). The investigation of the effect of translation inhibitors in B. bronchiseptica revealed a complex pattern of csp gene expression. The cspA specific transcripts were found to be most responsive to the presence of such inhibitors (Fig. 4). Other stress conditions including heat shock and salt stress also resulted in the prevalence of certain CSP transcripts. For example, the amount of the T3cspA transcript is strongly increased after heat shock, while one CSP encoding gene, cspC, was found to be responsive to salt stress (Fig. 4). Similar to certain CSPs of E. coli and Lactobacillus plantarum (Derzelle et al., 2000; Yamanaka & Inouye, 1997), the growth phase can affect CSP expression in B. bronchiseptica, since the T3cspC transcript was found to accumulate strongly in stationary phase (data not shown). The regulation of expression of cold-shock proteins is complex, and transcriptional and posttranscriptional mechanisms appear to be involved (Fang et al., 1997; Gualerzi et al., 2003). Here we provide evidence for the contribution of both levels of regulation in the control of CSP expression in B. bronchiseptica. For instance, there is evidence that an unknown transcriptional activator is involved in the transcriptional control of the cspB gene. Moreover, the cspC and cspD genes are apparently under the negative transcriptional control of the BvgAS two-component system. The repression of transcription of these genes at 37 uC may occur either directly via the BvgA response regulator or indirectly by a repressor such as the BvgR protein, which is itself under the positive control of the BvgAS system (Bock & Gross, 2001; Cotter & DiRita, 2000). However, despite these indications for transcriptional control in the case of the cspB, cspC and cspD genes, it is not known so far whether the strong accumulation of several specific transcripts, in particular that observed for the cspA, cspB and Microbiology 151


cspC genes under various stress conditions, is due mainly to transcriptional induction or to posttranscriptional mechanisms. In fact, in the case of the cspB gene we provide evidence that additional control mechanisms contribute to its expression. The data indicate that the 59-UTR of the cspB gene is important for efficient translation of the mRNA at low temperature, since the deletion of part of the 59-UTR interfered negatively with translation of the respective transcript. Within this 59-UTR, the 9 bp box present in the transcripts of four of the five cold-shock genes may contribute to translational efficiency, since its mutation to an unrelated sequence interfered significantly with the translation of the transcript. Finally, the stability of cspB mRNA differs strongly under different growth conditions, since the half-life of the T1cspB and T2cspB transcripts was found to be increased more than ten-fold during cold shock. We analysed the changes in the pattern of proteins synthesized by B. bronchiseptica up to 3 h after cold shock by 2D-PAGE, and we determined the identity of 18 of these CIPs (Fig. 8, Table 3). Among them, several proteins were found that were also identified to be cold inducible in Listeria monocytogenes (Phan-Thanh & Gormon, 1997), in E. coli (Gualerzi et al., 2003; Jones et al., 1987) and in Bacillus subtilis (Graumann et al., 1996). Examples other than the CSPs are the general stress protein GroES, the translation initiation factor IF-2, and IlvC. The latter enzyme is cold inducible also in Bacillus subtilis and catalyses the second step in valine and isoleucine synthesis. An interesting difference between the cold-shock response of E. coli and B. bronchiseptica is the induction of the universal stress protein UspA, which is induced under various stress conditions, but not after cold shock, in E. coli (Ferianc et al., 1998). Finally, it is interesting to note that despite the reductive genome evolution in B. pertussis, all factors identified to be cold inducible in B. bronchiseptica have been retained in the streamlined genome of the highly specialized human pathogen, a fact that once more may indicate a role of the cold-shock factors during infection. Most of the known virulence factors expressed by B. pertussis and B. bronchiseptica are positively regulated by the BvgAS two-component system (Bock & Gross, 2001; Cotter & DiRita, 2000). In addition to the virulence factors that are activated by the bvg locus, a second class of factors encoded by the VRGs are directly or indirectly repressed by the bvg locus (Knapp & Mekalanos, 1988; Schneider et al., 2002; Stenson & Peppler, 1995). Environmental stimuli such as a temperature below 30 uC or the presence of compounds like nicotinic acid and sulfate ions cause a reversible inactivation of the system, and result in the lack of expression of virulence factors, but in the induction of VRGs, a phenomenon termed phenotypic modulation (Cotter & DiRita, 2000). B. bronchiseptica cells expressing VRGs are more resistant to toxic reagents and harmful growth conditions, e.g. the action of antimicrobial compounds, http://mic.sgmjournals.org

intracellular survival in phagocytes or survival in lake water (Banemann et al., 1997; Porter et al., 1991; Schneider et al., 2000). It has been suggested that the Bvg2 phase may have been relevant for an ancestor of the mammalian pathogens occupying environmental niches (Gerlach et al., 2001). On the other hand, the fine-tuned down-regulation of virulence gene expression, e.g. by a mild reduction of temperature, leads to the so-called intermediate phase, which is characterized by a lower level of VAG (virulence-activated gene) expression and by the appearance of novel antigens. This intermediate phase may contribute to aerosolmediated transmission of the bordetellae in late infection phases (Cotter & DiRita, 2000). Interestingly, the cold-shock stimulon of pathogenic bordetellae and the virulence regulon are somehow interconnected. In fact, cspC and cspD expression was found to be controlled by the BvgAS system in a negative manner (Fig. 9), indicating that CspC and CspD belong to the Bordetella VRGs. A second interference with the cold-shock-responsive system and the virulence regulon was found after a slight overexpression of CspB. In both B. pertussis and B. bronchiseptica, mild overexpression of CspB interferes negatively with the transcription of the adenylate cyclase toxin, leading to a non-haemolytic phenotype of the bacteria. Since some increase of CspB expression is already observed at a temperature leading to the expression of intermediate antigens of B. pertussis, the data may indicate an in vivo role for the CspB-mediated control of CyaA expression. This is reminiscent of the previously described negative effect on toxin expression by the Tex protein of B. pertussis (Fuchs et al., 1996; Ko¨nig et al., 2002). Interestingly, the RNAbinding protein Tex is also responsive to cold shock in B. bronchiseptica and therefore also belongs to the family of the CIP proteins. In conclusion, we show that the cold-shock response of B. bronchiseptica resembles that of E. coli but differs in several aspects, e.g. the composition of the CIP family of proteins and the requirement of an individual cold-shock protein, CspB, for normal growth of the bacteria. It appears that the cold-shock stimulon is conserved between B. bronchiseptica and its closely related human pathogen, B. pertussis, indicating that CSP-mediated responses are not an evolutionary relict, but part of a genetic programme that may still be relevant also for the human pathogen despite its reductive genome evolution. Since several of the CSPs were found to be induced by stress signals other than cold shock, and cold stress is not considered to be relevant for the obligate host-associated species, the CSP regulon may be engaged in defence against other stress conditions encountered during infection, and cold-shock factors may be involved in a differential fine-tuning of virulence gene expression.

ACKNOWLEDGEMENTS We thank Dagmar Beier and Geraldine Bresolin for critical reading of the manuscript. We are grateful to Klaus Neuhaus and Siegfried 1907

D. Stu¨bs and others Scherer for discussions, and for providing us the cspA gene of Bacillus cereus. This work was supported by grant SFB479/A2 and by a postdoctoral fellowship to T. M. F. from the Deutsche Forschungsgemeinschaft.

Goodnow, R. A. (1980). Biology of Bordetella bronchiseptica.

Microbiol Rev 44, 722–738. Goyard, S. & Ullmann, A. (1991). Analysis of Bordetella pertussis cya

operon regulation by use of cya-lac fusions. FEMS Microbiol Lett 61, 251–256. Graumann, P. L. & Marahiel, M. A. (1998). A superfamily of

REFERENCES Arico, B., Gross, R., Smida, J. & Rappuoli, R. (1987). Evolutionary

relationships in the genus Bordetella. Mol Microbiol 1, 301–308. Bae, W., Jones, P. G. & Inouye, M. (1997). CspA, the major cold

shock protein of Escherichia coli, negatively regulates its own gene expression. J Bacteriol 179, 7081–7088. Bae, W., Xia, B., Inouye, M. & Severinov, K. (2000). Escherichia coli

proteins that contain the cold-shock domain. Trends Biochem Sci 23, 286–290. Graumann, P., Schro¨der, K., Schmid, R. & Marahiel, M. A. (1996).

Cold shock stress-induced proteins in Bacillus subtilis. J Bacteriol 178, 4611–4619. Graumann, P., Wendrich, T. M., Weber, M. H., Schro¨der, K. & Marahiel, M. A. (1997). A family of cold shock proteins in Bacillus

CspA-family RNA chaperones are transcription antiterminators. Proc Natl Acad Sci U S A 97, 7784–7789.

subtilis is essential for cellular growth and for efficient protein synthesis at optimal and low temperatures. Mol Microbiol 25, 741–756.

Banemann, A., Deppisch, H. & Gross, R. (1997). The lipopoly-

Gross, R. & Rappuoli, R. (1988). Positive regulation of pertussis

saccharide of Bordetella bronchiseptica acts as a protective shield against antimicrobial peptides. Infect Immun 66, 5607–5612. Bock, A. & Gross, R. (2001). The BvgAS two-component system of

Bordetella spp. a versatile modulator of virulence gene expression. Int J Med Microbiol 291, 119–130. Brandi, A., Pietroni, P., Gualerzi, C. O. & Pon, C. L. (1996). Post-

transcriptional regulation of CspA expression in Escherichia coli. Mol Microbiol 19, 231–240. Bycroft, M., Hubbard, T. J., Proctor, M., Freund, S. M. & Murzin, A. G. (1997). The solution structure of the S1 RNA binding

domain: a member of an ancient nucleic acid-binding fold. Cell 88, 235–242. Carbonetti, N. H., Fuchs, T. M., Patamawenu, A. A., Irish, T. J., Deppisch, H. & Gross, R. (1994). Effect of mutations

toxin expression. Proc Natl Acad Sci U S A 85, 3913–3917. Gross, R. & Rappuoli, R. (1989). Pertussis toxin promoter sequences

involved in modulation. J Bacteriol 171, 4026–4030. Gualerzi, C. O., Giuliodori, A. M. & Pon, C. L. (2003). Transcriptional

and post-transcriptional control of cold-shock genes. J Mol Biol 331, 527–539. Jiang, W., Fang, L. & Inouye, M. (1996). The role of the 59-end untranslated region of the mRNA for CspA, the major cold-shock protein of Escherichia coli, in cold-shock adaptation. J Bacteriol 178, 4919–4925. Jiang, W., Hou, Y. & Inouye, M. (1997). CspA, the major cold-shock

protein of Escherichia coli, is an RNA chaperone. J Biol Chem 272, 196–202. Jones, P. G. & Inouye, M. (1994). The cold-shock response – a hot

causing overexpression of RNA polymerase alpha subunit on regulation of virulence factors in Bordetella pertussis. J Bacteriol 176, 7267–7273.

topic. Mol Microbiol 11, 811–818.

Cormack, B. P., Valdivia, R. H. & Falkow, S. (1996). FACS-optimized

of proteins in response to low temperature in Escherichia coli. J Bacteriol 169, 2092–2095.

mutants of the green fluorescent protein (GFP). Gene 173, 33–38. Cotter, P. A. & DiRita, V. J. (2000). Bacterial virulence gene

regulation: an evolutionary perspective. Annu Rev Microbiol 54, 519–565. Derzelle, S., Hallet, B., Francis, K. P., Ferain, T., Delcour, J. & Hols, P. (2000). Changes in cspL, cspP, and cspC mRNA abundance as a

function of cold shock and growth phase in Lactobacillus plantarum. J Bacteriol 182, 5105–5113. Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensive

set of sequence analysis programs for the VAX. Nucleic Acids Res 12, 387–395. Fang, L., Jiang, W., Bae, W. & Inouye, M. (1997). Promoter-

independent cold-shock induction of cspA and its derepression at 37 uC by mRNA stabilization. Mol Microbiol 23, 355–364. Ferianc, P., Farewell, A. & Nystro¨m, T. (1998). The cadmium-stress

stimulon of Escherichia coli K-12. Microbiology 144, 1045–1050. Friedman, A. M., Long, S. R., Brown, S. E., Buikema, W. J. & Ausubel, F. M. (1982). Construction of a broad host range cosmid

cloning vector and its use in the genetic analysis of Rhizobium mutants. Gene 18, 289–296. Fuchs, T. M., Deppisch, H., Scarlato, V. & Gross, R. (1996). A new

gene locus of Bordetella pertussis defines a novel family of prokaryotic transcriptional accessory proteins. J Bacteriol 178, 4445–4452.

Jones, P. G., VanBogelen, R. A. & Neidhardt, F. C. (1987). Induction

Kimmel, B., Bosserhoff, A., Frank, R., Gross, R., Goebel, W. & Beier, D. (2000). Identification of immunodominant antigens from

Helicobacter pylori and evaluation of their reactivities with sera from patients with different gastroduodenal pathologies. Infect Immun 68, 915–920. Knapp, S. & Mekalanos, J. J. (1988). Two trans-acting regulatory

genes (vir and mod) control antigenic modulation in Bordetella pertussis. J Bacteriol 170, 5059–5066. Ko¨nig, J., Bock, A., Perraud, A.-L., Fuchs, T. M., Beier, D. & Gross, R. (2002). Regulatory factors of Bordetella pertussis affecting virulence

gene expression. J Mol Microbiol Biotechnol 4, 197–203. Mayr, B., Kaplan, T., Lechner, S. & Scherer, S. (1996). Identification

and purification of a family of dimeric major cold shock protein homologs from the psychrotrophic Bacillus cereus WSBC 10201. J Bacteriol 178, 2916–2925. Monack, D. M., Arico, B., Rappuoli, R. & Falkow, S. (1989). Phase

variants of Bordetella bronchiseptica arise by spontaneous deletions in the vir locus. Mol Microbiol 3, 1719–1728. Morales, V. M., Backman, A. & Bagdasarian, M. (1991). A series of

wide-host-range low-copy-number vectors that allow direct screening for recombinants. Gene 97, 39–47. Nystro¨m, T. & Neidhardt, F. C. (1996). Effects of overproducing the

universal stress protein, UspA, in Escherichia coli K-12. J Bacteriol 178, 927–930.

Gerlach, G., von Wintzingerode, F., Middendorf, B. & Gross, R. (2001). Evolutionary trends in the genus Bordetella. Microbes Infect 3,

O’Farrell, P. H. (1975). High resolution two-dimensional electro-

61–72.

phoresis of proteins. J Biol Chem 250, 4007–4021.

1908

Microbiology 151


Parkhill, J., Sebaihia, M., Preston, A. & 50 other authors (2003).

Stibitz, S. & Yang, M. S. (1991). Subcellular localization and

Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat Genet 35, 32–40.

immunological detection of proteins encoded by the vir locus of Bordetella pertussis. J Bacteriol 173, 4288–4296.

Phan-Thanh, L. & Gormon, T. (1997). Stress proteins in Listeria

monocytogenes. Electrophoresis 18, 1464–1471. Porter, J. F., Parton, R. & Wardlaw, A. C. (1991). Growth and survival

of Bordetella bronchiseptica in natural waters and in buffered saline without added nutrients. Appl Environ Microbiol 57, 1202–1206. Sambrook, J. & Russell, D. W. (2001). Molecular Cloning: a Labora-

tory Manual, 3rd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

VanBogelen, R. A. & Neidhardt, F. C. (1990). Ribosomes as sensors

of heat and cold shock in Escherichia coli. Proc Natl Acad Sci U S A 87, 5589–5593. von Wintzingerode, F., Schattke, A., Siddiqui, R. A., Rosick, U., Gobel, U. B. & Gross, R. (2001). Bordetella petrii sp. nov., isolated

from an anaerobic bioreactor, and emended description of the genus Bordetella. Int J Syst Evol Microbiol 51, 1257–1265. Wang, N., Yamanaka, K. & Inouye, M. (1999). CspI, the ninth

Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with

member of the CspA family of Escherichia coli, is induced upon cold shock. J Bacteriol 181, 1603–1609.

chain-terminating inhibitors. Proc Natl Acad Sci U S A 74, 5463–5467.

Weiss, A. A. (1992). The genus Bordetella. In The Prokaryotes, 2nd

Schneider, B. (2001). Intracellular survival mechanisms of Bordetella

edn, pp. 2530–2543. Edited by A. Balows, H. G. Tru¨per, M. Dworkin, W. Harder & K.-H. Schleifer. New York: Springer.

species in professional phagocytes. Dissertation, University of Wu¨rzburg. Schneider, B., Gross, R. & Haas, A. (2000). Phagosome acidification

has opposite effects on intracellular survival of Bordetella pertussis and B. bronchiseptica. Infect Immun 68, 7039–7048. Schneider, B., Stu¨bs, D. & Gross, R. (2002). Identification and

genomic organization of gene loci negatively controlled by the virulence regulatory BvgAS two-component system in Bordetella bronchiseptica. Mol Genet Genomics 267, 526–535. Simon, R., Priefer, U. & Pu¨hler, A. (1983). A broad host range

mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Biotechnology 1, 784–791.

Weiss, A. A. & Falkow, S. (1984). Genetic analysis of phase change in Bordetella pertussis. Infect Immun 43, 263–269. Wouters, J. A., Sanders, J. W., Kok, J., de Vos, W. M., Kuipers, O. P. & Abee, T. (1998). Clustered organization and transcriptional analysis

of a family of five csp genes of Lactococcus lactis MG1363. Microbiology 144, 2885–2893. Xia, B., Ke, H. & Inouye, M. (2001). Acquirement of cold sensi-

tivity by quadruple deletion of the cspA family and its suppression by PNPase S1 domain in Escherichia coli. Mol Microbiol 40, 179–188.

medium for the production of phase I Bordetella pertussis. J Gen Microbiol 63, 211–220.

Yamanaka, K. & Inouye, M. (1997). Growth-phase-dependent expression of cspD, encoding a member of the CspA family in Escherichia coli. J Bacteriol 179, 5126–5130.

Stenson, T. H. & Peppler, M. S. (1995). Identification of two bvg-

Yamanaka, K., Fang, L. & Inouye, M. (1998). The CspA family in

repressed surface proteins of Bordetella pertussis. Infect Immun 63, 3780–3789.

Escherichia coli: multiple gene duplication for stress adaptation. Mol Microbiol 27, 247–255.

Stainer, D. W. & Scholte, M. J. (1970). A simple chemically defined

http://mic.sgmjournals.org

1909