INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Wall-associated teichoic acids are a heterogeneous class of phosphate-rich polymers that are covalently linked to the cell wall peptidoglycan of gram-positive bacteria (2). They consist of a main chain of phosphodiester-linked polyols and/or sugar moieties attached to peptidoglycan via a linkage unit (1). Glycerol and ribitol are the most commonly occurring polyols and are often substituted with D-alanine or various sugar residues. Glycosylated glycerol teichoic acids are present in coagulase-negative staphylococci (13, 14), while glycosylated ribitol teichoic acids have been found in Staphylococcus aureus and Staphylococcus saprophyticus (13). Teichoic acids containing glycosylpolyol phosphates or sugar phosphates alone as components of their main chains occur in Streptococcus pneumoniae (40) and some species of staphylococci (3), respectively. The physiological function of teichoic acids is still not clear. However, they have been implicated in the control of autolysin activity (21), cation assimilation (7, 12, 22), and the provision of a phosphate reserve (15).

Thus far, all of the studies concerning the genetics of teichoic acid biosynthesis have been performed in B. subtilis, particularly B. subtilis 168. The major wall teichoic acid of B. subtilis 168 is poly(glycerophosphate) [poly(groP)], and the genes involved in the biosynthesis and translocation (tag genes) are organized into three operons: tagAB, tagDEF, and tagGH (30).

There is substantial evidence to support the conclusion that the poly(groP) teichoic acid is essential for cell viability. First of all, there is the isolation of conditionally lethal temperature-sensitive mutants defective in the synthesis of poly(groP). Such tag mutants exhibit a reduction in growth rate and a pronounced disturbance in cell morphology at the nonpermissive temperature (8, 37, 38). Second, there is the failure to disrupt the tagAB, tagDEF, and tagGH operons by insertion mutagenesis (29, 31, 32). Finally, there is the controlled reduction in the expression of the tagGH operon, which resulted in a rod-to-sphere transition in cell morphology characteristic of the conditional lethal tag mutants grown under nonpermissive conditions (29). Several reports indicate that wall teichoic acid also plays an important role in cell wall integrity of gram-positive cocci (9, 23, 35), but there have been no genetic studies to determine whether these polymers are essential for survival.

Here we report the cloning and sequencing of the S. epidermidis tagF gene and show that it can complement the temperature-sensitive tagF1 (rodC1) mutation of B. subtilis 1A486. tagF encodes the CDP-glycerol:poly(groP) glycerophosphotransferase, which is responsible for the polymerization of the main chain of the teichoic acid by sequential transfer of glycerol-phosphate units from CDP-glycerol to the linkage unit lipid (38). We also present evidence which suggests that the tagF gene of S. epidermidis is essential.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains and plasmids. Bacterial strains and plasmids used in this study are listed in Tables 1 and 2, respectively.

Bacterial growth media and antibiotics. The media used for culture of both Escherichia coli and B. subtilis strains was L broth (LB) or L agar (LA), supplemented, when required, with ampicillin (Ap) (100 µg ml¹) or chloramphenicol (Cm) (5 µg ml¹). S. aureus and S. epidermidis strains were grown in Trypticase soy broth (TSB) or on Trypticase soy agar (TSA) containing, when appropriate, Cm (5 µg ml¹) or tetracycline (Tc) (8 µg ml¹).

Manipulation of DNA. Standard procedures were used for DNA manipulation (5, 42). DNA modifying enzymes were purchased from New England Biolabs and Promega.

Degenerate oligonucleotide PCR. Degenerate oligonucleotide primers were designed by back-translation of the amino acid sequences ILYAPT (TP3 primer) and ITDYSSV (TP4 primer) shared between the aligned sequences of the TagB and TagF proteins of B. subtilis (31). The sequences of the TP3 and TP4 primers are 5'-AGCGAATTCATHYTNTAYGCNCCNAC-3' and 5'-AGCGAATTCACNSWNSWRTARTCNGTDAT-3', respectively (sequences incorporated into the primers for cloning of PCR fragments are underlined). The codes for degenerate positions are as follows: R, A+G; Y, C+T; S, G+C; W, A+T; H, A+T+C; D, G+A+T; and N, A+G+C+T. Reaction mixtures contained 100 ng of chromosomal DNA from S. epidermidis ATCC 14990, 2 µM concentrations of each oligonucleotide primer, 1.5 mM MgCl₂ and 2.5 U of Taq polymerase) in a final volume of 100 µl. Thermal cycling parameters began with an initial denaturing step at 94°C for 4 min, followed by 30 cycles at 94°C for 1 min, 40°C for 2 min, and 72°C for 1 min.

Southern hybridization. Transfer of DNA from agarose to Magna NT nytran membranes (Micron Separations, Inc.) and Southern hybridization with ³²P-labeled probes were performed by standard procedures (5). Probes were prepared by random primer labeling of purified DNA fragments with [-³²P]dATP by using the Prime-A-Gene kit (Promega). Autoradiography was performed by using X-Omat S film (Eastman Kodak Co.).

Colony hybridization. Colony hybridization was performed by the method of Hanahan and Meselson (18).

DNA sequencing and analysis. Progressive unidirectional deletions of the tagF locus were constructed by using the Erase-a-Base Kit (Promega). Nested deletions were made in both directions in two overlapping clones, pBN23 and pBH17, which spanned the tagF locus. The sequencing reactions were carried out by the cycle sequencing method with the Flash Dye Primer Sequencing Kit (Genpak) and analyzed on a model 373A sequencer (Applied Biosystems). Homology searches were performed by using the various BLAST algorithms available at the National Center for Biotechnology Information (NCBI) site. Protein sequence alignments were performed by using the subprogram PALIGN of PC/GENE (Intelligenetics) and the CLUSTAL W algorithm accessible at the Baylor College of Medicine Human Genome Sequencing Center web site.

Transformation. S. aureus and S. epidermidis cells were transformed by electroporation by the procedures of Oskouian and Stewart (33) and Augustin and Götz (4), respectively. Electroporation was performed with a Bio-Rad Gene Pulser equipped with a pulse controller (Bio-Rad Laboratories). Competent E. coli XL1-Blue (Stratagene) and B. subtilis cells were prepared and transformed by the procedures of Chung and Miller (10) and Karamata and Gross (24), respectively.

Construction of a site-directed mutation in the S. epidermidis tagF gene. A 436-bp fragment was amplified by PCR from pFC10 template DNA by using the oligonucleotide primers 5'-ATAATGACGTCTCTGAATTATTTTTAATAttTGATTGTTTAATTAC-3' (forward) and 5'-AATTTCAATTAAAATATTAAAAAGGGC-3' (reverse), as well as VENT DNA polymerase. The forward primer contained altered bases (lowercase lettering) and an AatII restriction site close to the 5' terminus (underlined).

Overexpression and purification of the GST-TagF fusion protein. In order to overexpress and purify a region of the S. epidermidis TagF protein, primers 5'-TACGGATCCAAAGTTAATCAATTTAG-3' (GF1) and 5'-TACAAGCTTTTATCATTGTTCCTTG-3' (GF2) were used to PCR amplify the region of the tagF gene encoding the 407 carboxyl-terminal amino acids of TagF and the two termination codons. Plasmid pFC10 was used as template DNA with VENT DNA polymerase and 30 temperature cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min 30 s. Primers GF1 and GF2 have 9-bp extensions at their 5' ends, which include restriction sites (underlined) to facilitate cloning. The PCR product was purified with the Wizard PCR Purification System (Promega), digested with HindIII and BamHI and ligated in frame with the glutathione S-transferase (GST) coding sequence of pGEX-KG to form pGEX-2 (GST-'tagF). pGEX-2 expressed a fusion protein in E. coli XL1-Blue of the predicted molecular mass following induction with isopropyl--D-thiogalactoside (IPTG).

1

1

Antibody generation and purification. Next, 500-µl samples of the purified GST-'TagF fusion protein (60 µg of protein ml¹) were emulsified in equal volumes of Freund complete adjuvant and injected subcutaneously into two New Zealand White rabbits, from which preimmune serum had previously been taken. Following two booster injections of 20 µg of fusion protein emulsified in Freund incomplete adjuvant, at 14-day intervals, the rabbits were sacrificed and bled out. Antibodies were purified from the serum by the procedure described by Owen (34).

SDS-PAGE and Western immunoblotting. SDS-PAGE was performed by standard procedures (28). The stacking and separating gels consisted of 4.5% (wt/vol) and 10% (wt/vol) 19:1 acrylamide-bisacrylamide, respectively. Following SDS-PAGE, gels were stained with Coomassie blue stain or transferred to nitrocellulose membranes (Millipore) by using a semidry blotter (Bio-Rad Transblot SD). Membranes were blocked overnight at 4°C in PBS containing 5% (wt/vol) skimmed milk (Marvel). After incubation with anti-GST-'TagF antibodies (1:2,000 in 5% skimmed milk) followed by incubation with horseradish peroxidase-conjugated anti-rabbit immunoglobulin G antibodies (Sigma) (1:1,000 in 5% skimmed milk), the TagF mutant and wild-type proteins were detected by using the enhanced chemiluminescence Western blotting reagent kit (Amersham) according to the manufacturer's instructions.

Detection of plasmid integration by PCR. Oligonucleotide primers, TQ1 and TQ2, were designed to detect chromosomal integration by homologous recombination of plasmids pTSTAG and pTSH17. The nucleotide sequences of primers TQ1 and TQ2 are 5'-CGTTTAAGTGCTAAAGAAGTTGTAGG-3' and 5'-GGAAATACAACGCATTTAC-3', respectively. Primer HD1 hybridizes at a region 67 bp downstream from the last codon of tagF and was used in combination with primer TQ2 as a positive control. The nucleotide sequence of primer HD1 is 5'-AATTTCAATTAAAATATTAAAAAG-3'.

Efficiency of plating. Strains of S. epidermidis TU3298 carrying the plasmids pTS2T, pTSH17, and pTSTAG were grown overnight at 30°C in TSB with Cm selection (5 µg ml¹) and then plated on TSA containing Tc (8 µg ml¹) at 45°C (restrictive temperature) and 30°C (permissive temperature). The efficiency of plating was determined as the proportion of colonies growing at 45°C compared to that growing at 30°C. At the restrictive temperature the plates needed to be incubated for 36 to 40 h before colonies were visible.

Nucleotide sequence accession numbers. The GenBank accession numbers of the S. epidermidis tagF locus and the partial sequence of the lctP gene are AF162863 and AF162862, respectively.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Identification of the S. epidermidis tagF gene. S. epidermidis ATCC 14990 produces a wall-associated glycerol teichoic acid which is chemically identical to the major wall-associated teichoic acid produced by B. subtilis 168 (J. Lonsdale [SmithKline Beecham], personal communication). It is reasonable to assume that the biosynthetic pathway leading to the production of this teichoic acid is identical or very similar in both organisms. For this reason, strain ATCC 14990 was chosen for isolating genes involved in glycerol teichoic acid synthesis. Using degenerate primers corresponding to shared amino acid sequences in the B. subtilis TagB and TagF proteins, genomic DNA from S. epidermidis strain ATCC 14990 was amplified by PCR, and a 260-bp PCR product was obtained. Sequence analysis of translated open reading frames (ORFs) revealed significant similarity with parts of the ORFs of both the TagF protein (P = 6.4 × e³¹) and to a lesser extent the TagB protein (P = 8.4 × e⁸) of B. subtilis 168. It was therefore thought likely that this fragment represented part of the S. epidermidis tagF gene.

Cloning the S. epidermidis tagF gene. The S. epidermidis tagF PCR product contains a single NsiI restriction site. When genomic DNA from S. epidermidis ATCC 14990 was cut with NsiI and hybridized with the ³²P-labeled tagF PCR product, two reactive bands appeared. The larger was ca. 2.3 kb in length, and the smaller was ca. 330 bp in length. Both fragments were isolated from a plasmid gene bank which had been constructed in pBluescript from ATCC 14990 genomic DNA cut to completion with NsiI. Reactive clones were identified by colony hybridization, and plasmids pGN23 and pGN333 containing the 2.3-kb NsiI fragment and the 330-bp NsiI fragment, respectively, were isolated.

Mapping the tagF gene. The orientation of the two NsiI fragments was determined by sequencing both ends of each fragment. The fragment cloned in pGN333 was 333 bp long and lay downstream from the 2.3-kb fragment, relative to the direction of tagF transcription. However, pGN333 did not contain the 3' end of the gene. The fragment in pGN23 contained the 5' part of tagF and was contiguous to the fragment in pGN333 (Fig. 1).

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Physical map of the S. epidermidis chromosomal region containing the tagF locus. The tagF gene is represented by the thick black arrow. Sequence analysis of the 3' end of the 5.5-kb fragment in pBCP55 revealed the presence of an ORF which shares significant homology (P = 4 × e23) with the E. coli lctP gene encoding L-lactate permease (11). The partially sequenced genes pal1, tpgX, and lctP are represented by unshaded bars; the thin arrows above indicate the direction of their transcription. Regions of DNA cloned into pGN333, pGN23, pBCP55, pGDH3, pBH17, and pBN23 are represented by black bars. The position of the 260-bp tagF PCR product generated with the TP3 and TP4 primers is indicated with an asterisk. Reference restriction sites are indicated as follows: P, PstI; N, NsiI; H, HindIII; E, EcoRV; C, ClaI; A, AatII. Dotted lines mark the unsequenced region, which is not drawn to scale.

22

Sequence analysis of the tagF locus. Sequence analysis of the 3,263-bp region of DNA encompassed by the overlapping clones in pBH17 and pBN23 revealed the entire tagF gene, the 3' end of the pal1 gene and the 5' end of a third gene, which will be referred to as tpgX. The tagF ORF is 2,163 bp long and encodes a predicted protein of 721 amino acids. The distance between the 3' end of the pal1 ORF and the 5' end of the tagF ORF is 276 bp. The distance between the 3' end of the tagF ORF and the first codon of the tpgX gene is 132 bp. The partial sequence of the tpgX ORF encodes the amino-terminal 140 amino acids of the predicted protein product and shares no significant similarity with any GenBank database sequences. The partial 5' sequence of the pal1 gene encodes the carboxyl-terminal 91 amino acids of the predicted Pal1 protein. All three genes are transcribed in the same direction (Fig. 1). Twenty-four bases downstream from the UAA termination codon of tagF is a 37-base sequence capable of forming a potential hairpin loop structure in mRNA with a G value of 21.4 kcal mol¹. Immediately following this hairpin loop is a 12-base sequence rich in rU residues (data not shown).

Analysis of the S. epidermidis tagF gene product. The protein encoded by the tagF gene from B. subtilis is 746 amino acids in length, which is 25 residues larger than the predicted product from the S. epidermidis TagF protein. Alignment of the two protein sequences shows 32.7% identity over their entire lengths. However, the carboxyl-terminal 410 amino acids of the proteins show higher levels of identity (45.9%; Fig. 2), while the remaining amino-terminal alignment possesses only 15.3% identity. The carboxyl-terminal region of the S. epidermidis TagF protein also shows sequence similarities with several other proteins (see legend to Fig. 2). It is reasonable to suggest that the more conserved C-terminal domain contains the catalytic activity and that conserved residues identified in Fig. 2 could be involved in catalyzing phosphodiester bonds during polymerization of polyol phosphate compounds. Honeyman and Stewart (19) suggested that the TagF protein in B. subtilis is cytoplasmic, and the same suggestion can be made for the TagF protein of S. epidermidis. However, this does not preclude association with the membrane by interaction with peripheral or integral membrane proteins involved with teichoic acid synthesis or translocation, for example, the TagG and TagH proteins (29).

View larger version (67K): [in this window] [in a new window]   FIG. 2.   Amino acid sequence alignment of the COOH-terminal regions of the TagF proteins of B. subtilis (BS) and S. epidermidis (SE). Identical amino acids are marked by bars, and conservative substitutions are indicated with dots. The BLASTP algorithm was used to conduct homology searches of the protein databases available at the NCBI site. Including the B. subtilis TagF, six gene products show sequence alignment scores of greater than or equal to 80 with the COOH-terminal region of the S. epidermidis TagF protein. In order of greatest similarity, the names, species of origin, accession numbers, and references for these gene products are as follows: TagF, B. subtilis, P13485 (19); Cps23fK, S. pneumoniae, AAC69534 (41); TasA, S. pneumoniae, CAA59773 (26); TagB, B. subtilis, P27621 (31); teichoic acid biosynthesis protein RodC related protein, M. thermoautotrophicum, AAB84867 (43); and hypothetical protein 3, H. influenzae, S49240 (44). Amino acid sequence alignment of these gene products with the S. epidermidis TagF protein was performed by using the CLUSTAL W algorithm. The asterisks correspond to amino acid residues common to the S. epidermidis and B. subtilis TagF proteins and at least three of the other five gene products examined.

Complementation of the tagF1 mutation in B. subtilis 1A486 with the S. epidermidis tagF gene. B. subtilis 1A486 (tagF1) was transformed with the vector pHPS9 as a control and with the pHPS9-tagF⁺ plasmid pFC10. Thus, strain 1A486(pFC10) contained a chromosomally located mutant copy of the B. subtilis tagF gene and multiple copies of the pFC10-located S. epidermidis tagF gene. In contrast, strain 1A486(pHPS9) contained only the chromosomally located mutant B. subtilis tagF gene.

View larger version (106K): [in this window] [in a new window]   FIG. 3.   Complementation analysis of B. subtilis 1A486 with the S. epidermidis tagF gene. Strains were grown for 16 h at 42°C on LA plates containing chloramphenicol (5 µg ml1). (A) Photomicrograph of strain 1A486(pHPS9) exhibiting the coccoidal morphology characteristic of the TagF1 (RodC1) phenotype. (B) Photomicrograph of strain 1A486(pFC10) showing the restoration of the wild-type rod-shaped morphology by complementation with the S. epidermidis tagF gene. (C) Comparison of growth of strains 1A486(pFCTH3) and 1A486(pFC10). (D) Comparison of growth of strains BSS40 and 1A486(pFC10).

Transformation of B. subtilis 1A486 with a mutated S. epidermidis tagF gene. A site-directed mutation in the S. epidermidis tagF gene was constructed. The nucleotide sequence of the tagF gene in pFCTH3 is identical to that of pFC10, except for a 2-bp mutation at positions 1846 and 1847 of the ORF. The mutation results in a single amino acid change, from serine to phenylalanine, in the TagF protein. This amino acid substitution is analogous to the tagF1 mutation in B. subtilis 1A486 which is responsible for the temperature-sensitive phenotype (19, 36, 38).

Integration of a single copy of the S. epidermidis tagF gene into the amyE locus of B. subtilis 1A486. It was of interest to determine if a single copy of the S. epidermidis tagF gene, integrated into the chromosome of B. subtilis 1A486, would also complement the tagF1 mutation. For this purpose plasmid pSC5, which harbored S. epidermidis tagF on a 3.26-kb fragment, was linearized before being transformed into strain 1A486. A transformant possessing the desired Amy Cm^r phenotype was isolated and named BSS40. The vector pDG268 was also transformed into strain 1A486 as a control to form strain BSS30. Thus, strain BSS40 contained a single copy of the S. epidermidis tagF gene and a single copy of the cat gene integrated into the chromosomal amy locus, whereas strain BSS30 contained only a single copy of the cat gene integrated into this locus.

Western immunoblotting analysis of B. subtilis 1A486 expressing the wild-type and mutant S. epidermidis TagF proteins. A protein band with an apparent molecular weight of 85 kDa was observed by SDS-PAGE analysis of a whole-cell lysate of B. subtilis 1A486(pFC10). The 85-kDa protein band corresponded to the size of the predicted TagF protein of S. epidermidis (85851 Da). This protein was present in relatively large amounts in the lysate of strain 1A486(pFC10) (Fig. 4A, lane 1). A whole-cell lysate of B. subtilis 1A486(pFCTH3) revealed the presence of a band that migrated slightly faster than that in 1A486(pFC10). This protein was present in relatively large amounts (Fig. 4A, lane 2). Both of these bands were absent from the lysate of B. subtilis 1A486(pHPS9) (Fig. 4A, lane 3). Western immunoblotting with anti-GST-'TagF antibody (Fig. 4B) indicates that the 85-kDa protein is S. epidermidis TagF.

View larger version (39K): [in this window] [in a new window]   FIG. 4.   SDS-PAGE and Western immunoblot analysis of whole-cell lysates of B. subtilis strains 1A486(pFC10), 1A486(pFCTH3), and 1A486(pHPS9). (A) Whole-cell lysates of B. subtilis strains 1A486(pFC10), 1A486(pFCTH3), and 1A486(pHPS9), separated by SDS-PAGE and stained with Coomassie blue stain, are shown in lanes 1, 2, and 3, respectively. A protein band of approximately 85 kDa (indicated by the arrowhead), present in lane 1 but absent from lanes 2 and 3, corresponds to the predicted molecular mass of the S. epidermidis TagF protein (85.9 kDa). A protein band with a molecular mass slightly less than 85 kDa is present in lane 2 (indicated by the arrowhead), corresponding to the mutated S. epidermidis TagF protein. (B) Western immunoblot analysis of whole-cell lysates of B. subtilis strains 1A486(pFC10), 1A486(pFCTH3), and 1A486(pHPS9). The polyclonal antiserum raised against the S. epidermidis GST-'TagF fusion protein is reactive against the wild-type and mutant S. epidermidis TagF proteins present in lanes 1 and 2, respectively. The lysate of strain 1A486(pHPS9) (lane 3) is nonreactive.

Attempted disruption of the S. epidermidis tagF gene by plasmid integration. The S. epidermidis tagF gene was cloned from strain ATCC 14990, but it proved impossible to introduce plasmid DNA into this strain by electroporation. Therefore, S. epidermidis TU3298, a strain capable of being transformed by electroporation (4), was used in the gene disruption experiments. Southern hybridization of the tagF locus from strain TU3298 indicated that it is very similar to that of strain ATCC 14990 (data not shown).

6

6

7

Conclusions. The tagF gene of S. epidermidis has been identified by sequence similarities of the encoded TagF protein with that of B. subtilis 168 and because it can complement a temperature-sensitive B. subtilis tagF mutant.