INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

PspA (pneumococcal surface protein A) plays a role in the ability of a pneumococcus to cause disease (17) and is present on all pneumococcal isolates. We have previously demonstrated that this protein is antigenically variant among different pneumococcal isolates (8). Despite this variation, we have observed that immunization with a limited number of variant PspAs can elicit cross-protection against a diverse number of pneumococcal isolates (15, 23). The cross-protection results indicate that while there is variation among PspAs, there must also be conserved PspA epitopes. Understanding the basis of the variation and conservation among PspAs is important in determining the mechanism of cross-protection elicited by PspAs.

PspA is attached to the surface to a pneumococcus by binding to choline in the pneumococcal lipoteichoic acids (27). While this attachment mechanism is novel when compared to that of most other gram-positive surface proteins, a number of other pneumococcal surface proteins have also been observed to bind choline (2, 10, 20).

To date the only complete nucleotide sequence of a pspA gene has been determined for pspA/Rx1 (Rx1 indicates the strain of origin). The deduced amino acid sequence of pspA/Rx1 reveals four distinct domains. The N-terminal half of the protein has a sequence expected to be an -helical coiled-coil structure similar to those of many surface proteins of other gram-positive bacteria (6, 25). The -helix of PspA/Rx1 is followed by a region of 81 amino acids, of which 23 are proline. Within the proline-rich domain, the prolines are clustered in Pro-Ala-Pro-Ala-Pro consensus repeats at either end of a 31-amino-acid nonproline region. This proline-rich domain is followed by 10 20-amino-acid repeats that interact with choline to anchor PspA to the pneumococcal surface. These repeats constitute the choline-binding domain (CBD) (25, 27). The CBD is followed by a slightly hydrophobic tail of 17 amino acids. The absence of these last 17 amino acids apparently does not affect the attachment of PspA to a cell surface (25, 27). Previous studies of PspA using antibodies to -helical-region epitopes and DNA probes for -helical-region sequences have indicated that -helical regions are especially varied (15, 22).

Here we report a comparison of the complete nucleotide sequence of a second pspA with that of pspA/Rx1. The pspA/EF5668 gene encoded a PspA that had a greater molecular size than that of PspA/Rx1. Our studies demonstrated that the greater size of PspA/EF5668, compared to that of PspA/Rx1, was due to additional nucleotides in the -helical coding region of the gene. This finding further confirmed the variability of the -helical regions of PspAs. Additionally, we observed that in spite of differences in the sequences encoding PspA/EF5668 and PspA/Rx1, both molecules were able to elicit cross-protection against several different pneumococcal isolates.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and transformation. Strains and plasmids used in this study are listed in Table 1. Pneumococcal and Escherichia coli strains were grown and stored as previously described (13, 17). E. coli was transformed by the method of Hanahan (11).

Hybridoma antibodies and immunoblot procedure. The anti-PspA hybridoma cell lines that secrete the monoclonal antibodies (MAb) Xi126 and XiR278 and the epitopes recognized by these MAb have been described previously (8, 12-14). Immunoblot analysis with MAb to detect PspA was carried out as previously described (15).

Amplification of pspA. PCR was carried out on genomic DNA isolated from Streptococcus pneumoniae EF5668 as previously described (13) with the oligonucleotide primers LSM2 (3') and LSM112 (5'), whose designs (22) are based on the sequence of pspA/Rx1 (25). LSM2 and LSM112 started at nucleotides 1990 and 47 of the pspA/Rx1 sequence. LSM2 and LSM112 contain SalI and BamHI restriction sites at their 5' ends.

DNA sequencing and analysis. Both strands of DNA encoding PspA/EF5668 were sequenced. The data were obtained by direct sequencing of plasmid pKSD2106, which contains the entire pspA gene from S. pneumoniae EF5668. Sequencing was done either with Sequenase (U.S. Biochemicals) or on an ABI 377 automated DNA sequencer (Perkin-Elmer, Foster City, Calif.). Sequencing primers were prepared as needed to facilitate sequencing of the cloned pneumococcal DNAs. In a few cases data were confirmed by sequencing of PCR-amplified fragments from the cloned pneumococcal DNAs. Sequence analyses were performed with the programs of the University of Wisconsin's Genetics Computer Group (GCG), MacVector 5.0 (Oxford Molecular), Sequencer 3.0 (GeneCodes, Inc.), and GeneJockey 1.5 (Biosoft, Cambridge, United Kingdom). The Matcher program was used to determine what portions of the sequence matched the 7-amino-acid motif characteristic of coiled-coil proteins (9). To provide direct comparison between the potential structural characteristics of PspA/EF5668 and PspA/Rx1 sequences, we analyzed both sequences using the Matcher program.

Purification of recombinant PspA/EF5668. E. coli KSD2106 was grown to mid-log phase as determined by optical density in 500 ml of Luria-Bertani medium. The cells were centrifuged and osmotically shocked (18) to release the periplasmic contents. NaCl was added to the solution to a final concentration of 0.25 M. This solution was passed over a choline-Sepharose column preequilibrated with 50 mM Tris acetate buffer (pH 6.9) containing 0.25 M NaCl (TAB). The column was subsequently washed with 10 bed volumes of TAB. The column was eluted with TAB containing 2% choline chloride, and 1-ml-volume fractions were collected. The presence of PspA/EF5668 was detected in the individual fractions by dot spotting 1 µl of 1/3 serial dilutions of each fraction onto nitrocellulose. The presence of PspA/EF5668 on the membranes was detected by anti-PspA MAb XiR278 followed by alkaline phosphatase-conjugated anti-mouse immunoglobulin. Those fractions containing recombinant PspA/EF5668 were pooled and further analyzed with silver stain (Silver Stain Plus; Bio-Rad, Hercules, Calif.) following sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Immunization and challenge. Immunization studies used CBA/N mice (Jackson Laboratory, Bar Harbor, Maine), which carry the X-linked immunodeficiency trait and fail to respond to polysaccharide antigens, making them very susceptible to pneumococcal challenge (3). Mice were injected subcutaneously with approximately 5 µg of isolated recombinant PspA (rPspA)/EF5668 in 0.2 ml of Freund's complete adjuvant. The mice were boosted at 2 weeks with an additional 5 µg of rPspA/EF5668 in incomplete Freund's adjuvant. Control mice were injected with adjuvant and a comparable volume of a comparable column fraction from an E. coli strain that did not express PspA. Approximately 7 days later, the mice were challenged intravenously with a minimum of 100 times the 50% lethal dose of the indicated virulent encapsulated pneumococcal isolate.

Nucleotide sequence accession number. The nucleotide sequence of pspA/EF5668 has been deposited in GenBank under the accession no. U89711.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning and expression of pspA/EF5668. LSM112, the 5' primer, contained an additional BamHI site at the 5' end, while LSM2, the 3' primer, contained an additional SalI site at the 5' end. These sites were used to clone the PCR product amplified from EF5668 genomic DNA into the BamHI/SalI site of pJY4163 (26). The resulting plasmid was designated pKSD2106 and carried a gene that expressed a 92-kDa protein that reacted with XiR278, an anti-PspA MAb (Fig. 1). The apparent molecular mass of PspA/EF5668 was approximately 10 kDa more than that observed for PspA/Rx1. Both rPspA/Rx1 and rPspA/EF5668 had several lower-molecular-weight bands than was observed for material derived from the strains from which the genes encoding PspA were cloned. Following the purification of the rPspA, the lower-molecular-weight bands were reduced in intensity or not detected at all by Western blotting (data not shown).

View larger version (40K): [in this window] [in a new window]   FIG. 1.   Immunoblot of PspA detected with the anti-PspA MAb XiR278. The lanes labeled Rx1 and EF5668 contain choline-eluted PspAs from those pneumococcal strains. The Rx1 lane was purposely overloaded to demonstrate the complex banding pattern that can be seen with PspA. The lanes labeled KSD1014 and KSD2106 contain rPspA purified from osmotically shocked E. coli expressing PspA/Rx1 and PspA/EF5668, respectively. Lane KSD1000 contains a similar E. coli preparation from a strain whose vector plasmid contains no insert.

Sequence analysis of pspA/EF5668. The complete nucleotide sequence of pspA/EF5668 was determined. The open reading frame of pspA/EF5668 includes 102 more nucleotides than were found for the open reading frame of pspA/Rx1. The nucleotides within the open reading frame were sufficient to encode a mature PspA with an expected molecular mass of 69,871.98 Da.

Amino acid sequence analysis. The deduced amino acid sequence of PspA/EF5668 is shown aligned with the previously published sequence of PspA/Rx1 (Fig. 2). The percentages of identity of the different regions of PspA from the two pneumococcal strains are shown in Table 2. The overall identity between the two molecules was 71%. Two blocks of 11 and 34 amino acids were present in the -helical region of PspA/EF5668 which were absent from PspA/Rx1. The larger of these two blocks fell within a region of PspA/Rx1 to which protective MAb had previously been mapped (13). Despite this divergence, analysis of amino acids 1 to 352 of the mature PspA/EF5668, like the corresponding region of PspA/Rx1, predicts the formation of a coiled-coil -helical structure (Fig. 3).

View larger version (56K): [in this window] [in a new window]   FIG. 2.   Alignment of the amino acid sequence of PspA/EF5668 with that of PspA/Rx1. Sequences are shown in standard one-letter amino acid code. The vertical lines indicate identity. Double and single dots between sequences indicate strong and weak similarity, respectively, based on Pam250 matrix alignment with the GCG program. Dots within a sequence indicate gaps in the alignment. The different domains of PspA are indicated above the sequences in boldface type. The words Leader, Mature PspA, Prolines, Repeats, and C-terminus each begin at the first codon that marks their respective domains.

View larger version (65K): [in this window] [in a new window]   FIG. 3.   Amino acid sequences of the N-terminal 352 and 288 amino acids of PspA/EF5668 and PspA/Rx1, respectively, arranged to illustrate the coiled-coil motif. In this motif the amino acids in positions a and d are generally nonhydrophilic, whereas amino acids at positions b, c, e, f, and g are generally hydrophilic and frequently charged. The most hydrophilic amino acids are indicated in boldface type. Amino acids compatible with hydrophilic environments are indicated in boldface, and amino acids compatible with nonsurface exposure are indicated in lightface. Regions of high identity between PspA/EF5668 and PspA/Rx1 are indicated by a single underline. Regions present in the PspA/EF5668 sequence that are absent from the PspA/Rx1 sequence, based on GCG Best Fit alignment (Fig. 3), are indicated by a double underline. Numbers to the left of each column denote the number of the first amino acid in each column. Numbers to the right of each column indicate the numbers of amino acids in contiguous blocks of the uninterrupted coiled-coil sequence motif.

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Comparison of the PspA/EF5668 coiled-coil structure with that of PspA/Rx1 and pooled data (7) from other known coiled-coil proteins. The charges of the amino acid side chains are listed to the right of amino acids Glu, Lys, Asp, Arg, and His. Those listed in boldface type include the most hydrophilic amino acids and include all charged amino acids. Those listed in lowercase letters and underlined are the most hydrophobic amino acids. Amino acid positions in the coiled-coil motif (a through g) are listed at the top of the figure. The observed amino acid frequencies for the pooled PspA data (PspA/EF5668 and PspA/Rx1) differed from those of the coiled-coiled regions of other proteins in several cases (indicated by shading) as follows: for the lysine at position e, P = 0.036; for the lysine at position f, P < 0.0001; for the arginine at position f, P = 0.0003; and for the alanine at position d, P = 0.0098.

Region of alternating charge. In the analysis of the PspA/Rx1 sequence it was observed that for 12 consecutive turns of the -helix (amino acids 125 to 166), the charge alternated between negative and positive. The positive charges were associated with lysines at many of the e, f, and g positions, and the negative charges were associated with glutamic acids at many of the b and c positions (25) (Fig. 4). The sequence of the coiled-coil region of PspA/EF5668 reveals a similar alternating charge motif for the 10 turns of the -helix between amino acids 123 and 157.

Cross-protection with PspA/EF5668. To assess the ability of PspA/EF5668 to elicit protective immune responses against pneumococci, mice were immunized with recombinant PspA/EF5668 or a comparably prepared column fraction from E. coli that contained the vector with no pneumococcal insert. Different groups of mice immunized with each antigen preparation were challenged by intravenous injection with one of five pneumococcal isolates representing four capsular types and five PspA serotypes. In all cases mice immunized with PspA/EF5668 showed significant survival compared to that of the controls. With two of the five strains, all immunized mice survived (Table 3).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

PspA, either by itself or in combination with additional pneumococcal proteins, offers the possibility of a protein-based pneumococcal vaccine. Such a vaccine could be important in targeting groups at risk for pneumococcal infection. Although a capsular conjugate vaccine will be effective in affording specific protection, there are possible limitations to this approach. If PspA is to serve as a component of a pneumococcal vaccine, it is important to define the antigenic diversity that is seen in PspAs. One approach to this problem is to obtain additional nucleotide sequence information of the genes encoding PspAs from many different pneumococcal isolates.

Previous studies have shown that all protective MAb recognize the -helical regions of PspAs, and most were found to recognize epitopes in the C-terminal 100 amino acids of the -helical region (13). Moreover, immunization with fragments of PspA containing portions of the -helical region has been shown to elicit protection (13, 19). Present data indicate that immunization with the C-terminal end of the -helical region shows similar cross-protection to immunization with the intact molecule (13). Earlier studies using fragments of pspA/Rx1 as DNA probes found the greatest degree of conservation among PspAs to be in their CBDs and proline-rich domains. The -helical domains were found to be the most diverse (16). This was confirmed in studies using specific oligonucleotides as hybridization probes (22). Studies of the expression of epitopes detected by MAb in different PspAs have also demonstrated high serological variability of the -helical regions (8, 24).

The present sequence provides the first insight into the diversity and conservation within PspAs at the sequence level. The encoded sequences displayed almost complete conservation in their leader regions, proline-rich regions, CDBs, and 17-amino-acid C termini. The only significant diversity was observed in the -helical domains, and even then there were some regions that were much more conserved than others. It was of particular interest that the region from amino acids 254 to 278 (numbered based on the numbering of PspA/Rx1) was the largest highly conserved portion in the -helical region and that it partially overlapped the region (amino acids 192 to 260) that has been identified as particularly cross-protective (13).

The most conserved regions of the -helix are all associated with major frameshifts or gaps in the coiled-coil motif. This finding may suggest that these regions are particularly important in the conformation of the protein. Alternatively, these regions may be unimportant in the elicitation of protection and thus not be subject to evolutionary selection for antigenic variation.

Although there were considerable differences in the -helical regions of the two PspAs, they both conformed to a largely coiled-coil motif. Both had similar amino acids at the several positions of the motif that differed significantly from amino acids of non-PspA coiled-coil proteins. One striking aspect of both sequences was the relatively high usage of lysine rather than of other positively charged amino acids such as arginine. It was also of note that the PspAs made more frequent use of alanine at nearly all positions than was observed in other coiled-coil proteins. We previously proposed that the long lysine side chains allow PspA to interact more effectively with the negative charges of most capsular polysaccharides.

Another aspect of both the -helical sequences that appeared to be relatively unique to PspA was that there were two regions in each PspA sequence where consecutive turns of the -helix alternated in charge. It seems likely that these unusual regions and the high frequencies of alanines and lysines in PspA may be important to the still unknown function of PspA.

In addition to being more varied than any other part of pspA, the portion of the gene encoding the -helical region is also more varied than the noncoding DNA 5' to the leader. This finding suggests that the greater diversity of the -helical region than of the rest of the molecule is because of selection favoring diversity rather than a lack of selection favoring function. As we have suggested in the past, it is likely that the diversity of this region may have been the result of selection favoring antigenic variability that helped pneumococci escape responses to prior pneumococcal infections. If this is the case, then it would be a strong argument for anti-PspA-mediated protection against pneumococci in humans, since humans are thought to be the primary reservoir of this pathogen.