Identification of two genes, cpsX and cpsY, with putative regulatory function on capsule expression in group B streptococci

Abstract

Two divergently transcribed open reading frames: cpsX and cpsY separated by a common regulatory region was identified upstream of the cpsA–D genes involved in polysaccharide capsule biosynthesis in group B streptococci (GBS). We suggest that these genes are involved in the regulation of capsule expression in GBS, since the CpsX protein shares sequence similarities with LytR of Bacillus subtilis, an attenuator of transcription while CpsY has similarity to a wide variety of members of the LysR family of transcriptional regulators. No deletions, insertions, DNA rearrangements, or apparent differences were discovered in the postulated regulatory genes when the gene region was compared in GBS with different capsule phenotypes. Thus, other yet unidentified gene loci may control capsule phase variation in GBS.

Group B streptococcus, Regulation, Capsule, Phase variation, lysR, lytR

1 Introduction

Group B streptococci (GBS) are currently the most common cause of neonatal sepsis and meningitis but may also cause serious infections in adults [1]. In neonatal early onset GBS disease, bacteria enter the blood through the lung after transfer to the baby from the maternal genital tract [2]. An important virulence factor of GBS is the polysaccharide capsule allowing the organism to avoid opsonophagocytic killing in the absence of type specific antibodies. The GBS capsule polysaccharide exists in at least 9 antigenically and structurally different types designated Ia, Ib, and II–VIII with type III being the predominant serotype isolated from infected infants [2]. GBS with aberrant expression of encapsulation may sometimes be encountered in the diagnostic laboratory. Non-typeable (NT) as well as GBS isolates with abundant encapsulation may be detected provided adequate methods are applied. Many clinical GBS isolates have the capability in vitro to change the capsule expression in a way which resembles phase variation [3]. The propensity to phase shift may differ 1000-fold between different isolates [4]. The phase variation is probably an important mechanism for sequential adaptation in GBS to changing host environment during different stages of infection [5–7]. How the polysaccharide capsule expression in GBS is regulated, and whether the phase-shift phenomenon is governed by a separate regulatory mechanism is yet unknown. Six genes involved in the capsule expression of GBS, cpsA–F, have been identified and sequenced [8, 9]. At some distance 5′ of cpsA, insertion sequence IS861 has been found to be located in strain COH-1 and it has been speculated to influence capsule gene expression [10]. Furthermore, the region directly upstream of cpsA has also been suggested to harbor regulatory elements [11].

1.1 Aims of study

The aim of the study was to map and compare the organization of the putative regulatory region between IS861 and the cpsA gene in two type III GBS isolates with different phase-shift frequency and that of an NT GBS isolate.

2 Materials and methods

2.1 Bacterial strains and culture method

M732, B4877, and COH-1 are invasive GBS type III isolates described previously [4, 9], 954 is a colonizing NT GBS isolate [12]. The bacteria were plated on blood agar plates (Columbia II agar base, BBL, supplemented with 5% horse blood) and incubated at 37°C overnight.

2.2 PCR techniques

Five to ten colonies, picked from a blood agar plate, were resuspended in 50 µl sterile distilled water in a micro test tube and heated for 5 min at 100°C. After centrifugation, 10 µl of the supernatant was used as template in each PCR reaction. Reaction conditions were 0.2 mM dNTPs, 20 pmol of each primer with 0.5 U Taq polymerase, buffer with 1.5 mM MgCl₂ (MBI Fermenta) and bovine serum albumin (0.17 mg ml⁻¹). The amplification was performed on a PTC-200 thermal cycler (MJ Research) with annealing temperatures depicted in Table 1. The PCR products were separated with electrophoresis on 0.7% agarose gel.

Table 1

Primers used in this study

Primer designation	Location	Primer sequence	Annealing temperature (°C)
cps28	5′ of cpsA	5′-CCATGACCAGTTAATGCTTG-3′	55
ISright	IS861	5′-CCTTGAGCAAGCTATCACAG-3′	55
cpsatu2	cpsX	5′-CATACAAACCTGCATGGG-3′	50
iselin2	cpsY	5′-GGGATTGGATTTGCTCAC-3′	50
cpsatu3	cpsX	5′-TGCGGCTAATTTGCTGACAG-3′	57
iselin3	cpsY	5′-TAGCGGCTTCATTCATGCTACC-3′	57
cpsatu4	cpsY	5′-GCACAGACATGACCCAATACG-3′	55
cpsatu5	cpsY	5′-CGAAATTGTCGCAATCCCTC-3′	57
A5′fo.	cpsX	5′-GAGGCGATAACGATAGAGGTAG-3′	52
cpsAfo	cpsA	5′-CGACATCACATAGAAGAAAAG-3′	55.5
cpsAre	cpsA	5′-CACTCGCTACAAAATGCAC-3	55.5
cpsBfo	cpsB	5′-TCATTAACCCTAATACCCC-3′	51
cpsBre	cpsB	5′-CCTGCACCAACAATAAAC-3′	51
cpsCfo	cpsC	5′-GGGGAAGGAAAATCCACTAC-3′	55
cpsCre	cpsC	5′-TATTGCGGCATCAACAAC-3′	55
cpsDfo	cpsD	5′-GCTAAGTTTTCACGAGATACC-3′	52
cpsDre	cpsD	5′-AATCCTACCATTACGACCTAC-3′	52
IS861‘275’	IS861	5′-AGAAAGATCGGGATGTCCTG-3′	55
IS861‘274’	IS861	5′-CTGTGATAGCTTGCTCAAGG-3′	55

Oligonucleotide primers were selected from nucleotide sequences with GenBank accession number; cpsA–D: L09116 [9]; IS861: M22449 [10]; and cpsX and cpsY: accession number Y17221, this study.

2.1 Southern blot

Chromosomal DNA, prepared as previously described [4], was digested with 15 U of the appropriate restriction endonuclease (MBI Fermenta) and separated on a 0.7% agarose gel by gel electrophoresis. DNA was transferred to Hybond-N⁺ filters (Amersham) with VacuGene™ XL, vacuum blotting system (Pharmacia Biotech AB) according to the manufacturer′s instructions. DNA-fragments were isolated by extraction from agarose gel with Gene Clean II kit (Bio 101), and labeled with digoxigenin (DIG) DNA labeling kit (Boehringer Mannheim). Probes used in Southern blot hybridization in this study were the 3.5-kb DNA fragment generated by PCR amplification with the primers cpsARE and isright, the 649-bp fragment generated with primers iselin3 and cpsatu3 (Table 1) and cut with EcoRI. Hybridization and detection was performed with DIG Nucleic Acid Detection Kit (Boehringer Mannheim) according to instructions from the manufacturer.

2.4 Sequencing technique

PCR amplification of the region to be sequenced was performed from two separate DNA preparations as described above. The PCR products were purified with ‘High Pure PCR Product Purification Kit’ (Boehringer Mannheim) and sequencing reactions were set up in doublets for each primer from both PCR products. The primers used for sequencing and their physical localization are shown in Table 1 and Fig. 2. Sequencing was performed with ABI PRISM™ products from Perkin-Elmer Applied Biosystems. Sequencing reactions were made according to the manufacturer's protocol: ‘ABI PRISM™ Dye Terminator Cycle Sequencing Ready Reaction Kit, With AmpliTaq DNA Polymerase, FS, Revision A, August 1995’. The sequencing reactions were run on a ABI 377 Sequencer, using the software ABI PRISM™ 377XL DNA Sequence Data Collection, version 2.0. The running parameters were as follows: electrophoresis voltage 1.68 kV, electrophoresis current 26.7 mA (max. 50 mA), electrophoresis power 44 W (max. 150 W), gel temperature 51°C and laser power 40 mW. Collected data were analyzed with ABI PRISM™ Sequencing Analysis software, version 3.0. For fragment analysis and assembly the Genetic Group Sequence analysis software package and programs available at ExPASy molecular biology server at Geneva Hospital and University of Geneva were used.

Figure 2

Schematic presentation of the cpsA-IS861 region of strain M732 showing the positions of the primers used for DNA sequencing.

3 Results

3.1 Physical mapping of the DNA region 5′ of cpsA gene

Amplification with primers from IS861 to cpsA could confirm the position of IS861 at approximately 2.6 kilo bases (kb) 5′ of cpsA in COH-1 and M732 but not in strains 954 or B4877. Southern blot hybridization with the 3.5-kb probe hybridized with the 2.8-kb and 1.4-kb EcoRI fragments in all isolates and to the 4.6-kb fragment in COH-1 and M732 and to a 3.0-kb fragment in the other isolates. The probe also hybridized to the 1.2-kb HindIII fragment in all isolates but only to the 3.7-kb HindIII fragment in COH-1. In all the other isolates the probe hybridized to a 2.8-kb fragment, in addition to the 1.2-kb fragment. Thus, the physical map of 954 and B4877 in this region corresponds to that of COH-1 and M732, except that the upstream EcoRI and HindIII sites were located closer to the 1.4-kb fragment since IS861 was missing at this site in these isolates (Fig. 1). Using PCR primers for the amplification of cpsA-cpsD, we could not detect indications of large DNA rearrangements in this region, consequently these genes seem to be organized as in COH-1.

Figure 1

Schematic presentation of the organization of the cpsA-IS861 region of GBS strains COH-1 and M732, and the corresponding region in strains B4877 and 954. Location of EcoRI sites (E) and HindIII sites (H) are indicated. Sizes (in kb) of the EcoRI fragments are depicted on the top of the map and the sizes of the HindIII fragments are indicated below the line.

3.2 Sequencing and analysis of the region 5′ of the cpsA gene

Sequencing was performed according to the strategy shown in Fig. 2. Computer analysis of the nucleotide sequence revealed the presence of two divergent operons: the operon cpsX in the same direction as cpsA and the operon cpsY in the opposite direction (Fig. 3). Search for possible transcription signals revealed that potential promoters were clustered in a 188-(M732 and B4877) and 187-(954) base pairs (bp) regulatory region which separates cpsX from cpsY. In the regulatory region we could identify two −10 and −35 regions for cpsX and one −10 and −35 region for cpsY overlapping the −10 to −35 region of cpsX. For cpsX a possible ribosome binding site (RBS) and a possible start of transcription +1 was deduced while for cpsY neither RBS nor start of transcription could be localized. Close to the RBS of cpsX a region with 87% A-T content was located (Fig. 3).

Figure 3

The nucleotide sequence of the region between cpsA and IS861 of strain M732. Both strands of DNA in the intergenic region between cpsX and cpsY are illustrated. Asterisks denote stop codons. Restriction sites for EcoRI are shown. Deduced amino acid sequences are given in single letter code. Ribosome binding sites are indicated by double underlining. −35 and −10 promoter regions, start of transcription +1 and start codons are marked. The A-T rich region close to cpsX is boxed. The target sequence for IS861 is underlined. GenBank accession no. Y17221. Accession number for the corresponding region in isolate B4877 is Y17241 and for 954, Y17218.

3.3 Nucleotide sequence comparison

The nucleotide sequence from M732 revealed 99.4% and 99.3% identity with the corresponding sequence in 954 and B4877, respectively. The start codon of cpsA gene is located 5 bp downstream of the cpsX stop codon in all isolates. The short sequence of cpsA that was sequenced shows 99% identity in all isolates when compared with the cpsA sequence of COH-1 [9]. The nucleotide sequence in M732 displayed minor differences as compared to the corresponding sequences in B4877 and 954. Within cpsY of B4877 3-bp changes were noted and in 954 2 bp differed as compared to the cpsY sequence in M732. In the cpsX-cpsY intergenic region 2-bp, and in 954 3-bp differences were found when compared to M732. In the cpsX gene 9- and 10-bp differences were detected in the sequence of B4877 and 954, respectively, compared to the corresponding sequence in M732. Despite the base pair substitutions present in 954 and B4877 compared to that of M732, both CpsX and CpsY open reading frames were intact. A deletion was found in the intergenic promoter region of 954 which did not affect the two reading frames of 954.

A comparison of the nucleotide sequence of 954 and B4877 with the corresponding sequence of M732 displays 99.7% and 99.4% identity, respectively, for cpsX, and 99.8% and 99.3%, respectively, for cpsY. Parts of cpsX from all isolates showed significant identity with capsule genes from different serotypes of Streptococcus pneumoniae: cps19fA (65%), cap1A (64–65%), and the dexB-orf1 space (62%) of type 3. Furthermore, 59% identity to the epsA gene, required for exopolysaccharide (EPS) synthesis, of S. thermophilus was observed. For the cpsY gene no significant identities at the nucleotide level could be found in the data bank.

3.4 Amino acid sequence comparison

The predicted products of ORFs cpsX and cpsY consist of 486 and 308 amino acid residues, respectively, in all isolates. The CpsX and CpsY protein of 954 and B4877 had 99.2 and 99.3% identity, respectively, with the corresponding proteins of M732. CpsY of 954 and B4877 were identical. CpsX is a basic protein and contains three hydrophobic segments in all three strains at amino acid residues 20–37, 53–68, and 77–92, according to hydrophobicity plots generated by the method of Kyte and Doolittle [13]. Furthermore, CpsX protein of 954, M732, and B4877 showed 56.4, 56.6 and 56.8% identities, respectively, to EpsA of S. thermophilus. CpsX from all three GBS isolates had 31.1% identity to LytR of B. subtilis, a protein which has 27% identity to EspA [14]. The identity of CpsX to LytR was seen predominantly in the N-terminal half of LytR (Fig. 4). The CpsY protein of 954, M732, and B4877 showed 33–44% identity to the N-terminal region of a wide variety of LysR family members (Fig. 5).

Figure 4

Multiple alignment comparisons of the N-terminal amino acid sequence of CpsX from GBS B4877, 954, and M732 with EpsA of Streptococcus thermophilus and LytR of Bacillus subtilis. Lower case letters indicate amino acid residues in CpsX that are non-identical between B4877, 954, and M732. Bold face letters show the amino acid residues that are identical in CpsX and EpsA, CpsX and LytR, or in all proteins. Hydrophobic regions are underlined [14, 19]. Sequences were obtained from the non-redundant GenBank CDS translations+PBD, Swiss-Prot+Spupdate+PIR and alignment was made separately between CpsX to EpsA and CPsX to LytR with Pileup by Genetic Computer Group systems software and compiled manually.

Figure 5

Multiple alignment comparisons of the N-terminal region of CpsY from M732, B4877, and 954 and representatives of the LysR family of proteins. A consensus sequence of a helix-turn-helix (HTH) motif of the LysR family of proteins, often positioned between residues 23–42, is depicted above the sequence comparison, where h indicates hydrophobic residues (V, I, L, M) and p represents hydrophilic residues (T, S, N, Q, D, E, K, R, H) [20]. Residues non-identical to CpsY are indicated by bold face letters and underlined residues match the consensus sequence of the HTH-motif. Sequences were obtained from the non-redundant GenBank CDS translations+PDB, Swiss-Prot+Spupdata+PIR and alignment was made manually.

4 Discussion

Phenotypic variation of surface components is a well known strategy for several pathogenic bacteria in their interaction with a changing environment [15–17]. The capability to modulate the polysaccharide expression may be such an important feature in order for GBS to be a successful invasive pathogen. The comparison between GBS isolates B4877, 954, and M732 did not reveal any deletions, insertions, or rearrangements within the cpsA–D gene cluster or any structural differences in the cpsX and cpsY genes to account for the phenotypic differences of the capsule expression that the isolates display. Furthermore, it seems that IS861 located upstream of cpsX is not involved in the positive or negative regulation of capsule expression. The nucleotide sequence analysis revealed the presence of two open reading frames, cpsX and cpsY, transcribed in opposite directions separated by a common regulatory region. The two proteins were intact in all sequenced isolates with highly different capsule expression phenotypes with only minor amino acid differences with no apparent influence on functional properties (Figs. 4 and 5).

The nucleotide sequence for cpsX reported in this paper shares high identity with the epsA gene of S. thermophilus required for exopolysaccharide (EPS) synthesis, and with genes essential for capsule formation in various serotypes of S. pneumoniae. Sequence similarities also exist at the amino acid residue level (data not shown). CpsX as well as cps19fA and epsA gene products exhibit significant amino acid sequence identity to LytR of B. subtilis (Fig. 4) [14, 18, 19]. LytR is a basic protein that acts as an attenuator of the lytABC operon, which determines autolytic activity in Bacillus, and is thought to be membrane bound via a single hydrophobic N-terminal anchoring domain. CpsX, EpsA, and Cps19fA are also basic proteins with hydrophobic segments. However, the positions of these segments do not correspond to those in LytR (Fig. 4). EpsA and Cps19fA have been suggested to be involved in the regulation of extracellular and capsule bound polysaccharides based on the above mentioned similarities to LytR [14, 18]. We therefore suggest that CpsX may also act as a regulator of capsule gene expression in GBS.

The presence of a LysR type of transcriptional regulator (LTTR) is usually considered if the amino acid sequence identity compared to the highly conserved N-terminus of known LTTR exceeds 20%[20]. Much of the similarity derives from a helix-turn-helix motif which is nearly 40% identical in all LTTRs. The CpsY gene revealed 30–44% identity to the N-terminus of a large number of LTTRs and also shared significant identity to the helix-turn-helix motif (Fig. 5). We therefore suggest that CpsY encodes a LTTR. It has been suggested that high Arg content in LTTRs might be related to the specific function of the protein as a transcriptional regulator. Some LTTRs have Lys-to-Arg ratios as low as 0.03–0.7, possibly as a result of a high G+C%[21]. However, the average streptococcal G+C content is low, 30–40%, and as a consequence the Lys-to-Arg ratio of CpsY is high, 1.5, as compared to other LTTRs [21].

A comparison of the intergenic region between cpsY-cpsX to that of lytR-lytABC and to the LysR family display several differences. At the lytR-lytABC intergenic region, two −10 to −35 regions in front of lytABC and one overlapping obvious −10 to −35 region for the regulator gene (lytR), which is transcribed in the opposite direction, were found. In the case of cpsY-cpsX the situation is the opposite, provided cpsY is the target gene. No correspondence to the inverted repeats, suggested to play a regulatory role, in the lytR-lytABC promoter can be found between cpsY-cpsX. However, similarities also exist: the two promoter regions cpsY-cpsX and lytR-lytABC are of approximately the same length. The promoter region of cpsY-cpsX is extended compared to the 25–70-bp promoter region present in most LTTRs and the partially dyadic binding sequence near −65 is missing [20]. An A+T rich region probably involved in regulation of DNA breathing by LysR family proteins could also be identified upstream of cpsX (Fig. 3) [21].

Numerous examples of mechanisms for phase variation have been published. In group A streptococci, phase variation of surface structures like C5a peptidase and the M-protein have been shown to be transcriptionally regulated while phase variation of capsule polysaccharide in Neisseria meningitidis has proved partially to be determined at the level of translation [22, 23]. Variation of encapsulation in Haemophilus influenzae is suggested to be due to gene amplification and repetitive intergenic elements influence the frequency of phase variation of colony opacity in S. pneumoniae[24, 25]. If the cpsX and cpsY gene products are involved in the regulation of the capsule gene cluster in GBS, the question still remains how this is brought about. The significant identity of CpsX to LytR and CpsY to LysR, and the presence of a complex intergenic promoter region, leaves several options open. Most LTTRs are coinducer-responsive transcriptional activators that activates transcription of a divergently located target gene while also repressing their own transcription [20]. Cell growth rate has been pointed out to regulate the expression of capsular polysaccharide in GBS [26], no further activating or repressing factors are known to influence regulation or phase variation of capsule expression or any other phase variation in GBS, e.g. that of colony opacity [27]. The question whether cpsX gene is the target gene of CpsY and if sensing of environmental changes occurs through coinducers of CpsY, remains open for further study.

5 Conclusions

The nucleotide sequence of the region upstream of cpsA in GBS contains two divergently transcribed open reading frames, cpsX and cpsY. The CpsX protein shares similarities with LytR, an attenuator of transcription in Bacillus subtilis. The CpsY protein has similarities to the LysR family of transcriptional regulators. No difference in nucleotide or amino acid sequence was found which could account for the phenotypic differences between GBS with different frequency of capsular phase shift (M732 and B4877) or a non-typeable isolate of GBS (954).

Acknowledgements

This work was supported by grants from the Swedish Medical Research Council (No. 08675), the Wiberg Foundation, Umeå University Medical Faculty, and the Västerbotten Läns landsting to M.N. and from the Swedish Society for Medical Research, and the J.C. Kempe Foundation to M.S.