INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Pneumocystis carinii remains an important life-threatening opportunistic pathogen of immunocompromised patients, especially those with human immunodeficiency virus (HIV) infection. The major surface glycoprotein (MSG; also called glycoprotein A) is the most abundant protein expressed on the surface of P. carinii, as assessed by Coomassie blue staining (22, 29, 36), and appears to play a critical role in the pathogenesis of pneumocystosis, possibly by acting as an attachment ligand to lung cells (7, 28, 45). MSG is also a target of both humoral and cellular immune responses by the host (8, 11, 22, 23, 37-39). Previously, we reported that multiple genes encode the MSG of rat P. carinii, and we demonstrated that different MSGs can be expressed in the lung of a rat infected with P. carinii (1, 19). Similarly, multiple genes encode the MSG of P. carinii infecting ferrets and mice (13, 14, 44). Additional studies have shown that there is a single genomic site for expression of rat MSG variants (5, 35, 42, 43). These studies suggest that P. carinii has developed an elaborate system for antigenic variation, presumably to evade host defense mechanisms.

Molecular and immunological studies have clearly demonstrated that P. carinii organisms isolated from different host species are distinct organisms and may in fact be separate species (10, 16, 17, 33). While animal models can provide important information about the biology of P. carinii, studies examining human interactions with P. carinii need to use human P. carinii-derived reagents. The cloning of human P. carinii (Pneumocystis carinii f. sp. hominis) MSG genes has recently been reported (9, 34). Since only one full-length sequence was reported, the present study was undertaken to further characterize the P. carinii f. sp. hominis-derived family of MSG genes. In addition, we undertook to express these genes, since P. carinii f. sp. hominis cannot be cultured and there is no reliable source of organisms for purifying large amount of antigens or other biologically relevant proteins.

(This work was presented in part at the 5th International Workshops on Opportunistic Protists and the 5th General Meeting of the European Concerted Action on Pneumocystis Research, September, 1997, Lille, France [31].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

DNA preparation. DNA was isolated from an autopsy lung sample from an HIV-infected patient with P. carinii pneumonia according to standard methods, by using sodium dodecyl sulfate (SDS) and proteinase K (0.5 µg/ml), followed by phenol-chloroform extraction and ethanol precipitation (3). A genomic library using the same DNA cloned into the XhoI site of the lambda GEM12 vector (Promega, Madison, Wis.) was commercially prepared (Lofstrand Labs Limited, Gaithersburg, Md.).

PCR and subcloning. Primers to amplify full-length human P. carinii genes were designed on the basis of published data (9). The sense primer, JK151 (5'-TTT CAT ATG GCG CGG GCG GTC AAG CGG CAG-3'), corresponds to nucleotides 153 to 175 of a published MSG sequence (GenBank accession no. L27092), and the antisense primer, JK152 (5'-CTA AAT CAT GAA CGA AAT AAC CAT TGC TAC-3'), is complementary to nucleotides 3215 to 3244. An NdeI site, which substitutes a methionine for the valine of the original sequence, was created at the beginning of JK151 in order to facilitate subcloning and expression. For amplification, 1 µg of genomic DNA was added to a 50-µl reaction mixture containing primers (25 pM each), deoxynucleoside triphosphates (0.2 mM), 5 U of AmpliTaq (Perkin-Elmer), and MgCl₂ (2.5 mM). DNA amplification was performed on a Perkin-Elmer Cetus DNA thermal cycler. An initial denaturation cycle (1 min at 96°C) was followed by 36 cycles of denaturation at 95°C for 1 min, annealing at 50°C for 2 min, and extension at 72°C for 2 min, followed by a final extension after the last cycle at 72°C for 10 min.

Southern hybridization and library screening. For Southern hybridization with a radioactive probe, DNA was treated with restriction enzymes, separated by agarose gel electrophoresis, and transferred to Hybond N+ membranes (Amersham Life Science, Arlington Heights, Ill.) with 0.4 M NaOH. DNA was probed with an approximately 600-bp XbaI fragment of the human P. carinii MSG III gene (9) (a gift from James R. Stringer, University of Cincinnati, Cincinnati, Ohio) that had been labeled with [-³²P]dATP or [-³²P]dCTP by using a random priming kit (Boehringer Mannheim). Filters were prehybridized for 4 h and then hybridized overnight at 55°C in 6× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7])-0.5% SDS and 5× Denhardt's solution. Blots were washed in 6× SSPE-0.5% SDS at room temperature for 10 min and then in 0.5× SSPE-0.5% SDS at 55°C twice for 30 min each time. The genomic library was screened by using a gel-purified full-length fragment of human P. carinii MSG 11 under the conditions described above. One clone that hybridized strongly to the probe was subcloned into the BamHI site of pBluescript II (Stratagene, La Jolla, Calif.).

Nucleotide sequencing. Sequencing was performed with an automated sequencer (model 373 or 377; Applied Biosystems, Perkin-Elmer, Foster City, Calif.) either in our laboratory or by contract (San Diego State University, San Diego, Calif.). The nucleotide sequence and deduced amino acid sequence data were analyzed by Factura and AutoAssembler (both from Applied Biosystems), Sequencher (Gene Codes Corp., Ann Arbor, Mich.), MacVector (Scientific Imaging Systems, New Haven, Conn.), ClustalW (40), and GeneWorks (IntelliGenetics, Mountain View, Calif.).

Construction and expression of recombinant human P. carinii MSG. The full-length human P. carinii MSG 32 gene was inserted into pBlueBacHis2A (Invitrogen) at the EcoRI site for expression in a baculovirus insect cell system. Correct insertion was confirmed by restriction mapping and sequencing. Isolation of recombinant virus, plaque purification, and amplification of high-titer virus stocks were performed according to the manufacturer's protocols (Invitrogen). PCR amplification with gene-specific primers was used to confirm the presence of the gene in the virus. Sf9 cells were grown at 27°C in SFII-900 medium (GIBCO BRL, Grand Island, N.Y.) with 5% fetal calf serum to a density of 2.0 × 10⁶ cells/ml. Cells were infected at a multiplicity of infection of 5. Seventy-two hours after infection, cells were harvested by centrifugation, washed with phosphate-buffered saline (PBS) supplemented with phenylmethylsulfonyl fluoride (PMSF) (1 mM), and then resuspended in 10 mM Tris-HCl, pH 8, with 1 mM PMSF and sonicated. The cell lysates were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting.

SDS-PAGE and Western blotting. SDS-PAGE and Western blotting were performed by standard techniques, as previously described (18). Electrophoresis was carried out in prepoured discontinuous 8 and 14% acrylamide-Tris-glycine gels (Novex, San Diego, Calif.). Proteins were stained with Coomassie blue or transferred to nitrocellulose membranes, following which Western blotting was performed with a variety of antisera by standard techniques (18). Recombinant rat P. carinii MSG GP3 (expressed in a baculovirus system) (24) and purified recombinant -galactosidase (expressed in the pET28-E. coli system) were used as controls in Western blotting.

Nucleotide sequence accession numbers. The sequences of the five PCR-generated human P. carinii MSG clones and the 12,792-bp clone from the genomic library have been deposited in GenBank under accession no. AF033208 through AF033212 and AF038556, respectively.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Although the human P. carinii MSG gene has recently been identified, only one full-length clone has been reported (9, 34). Since multiple variant copies of the MSG gene are present in the rat P. carinii genome, and available data support a similar organization in human P. carinii (9, 34), we wanted to obtain sequence information on additional full-length human P. carinii MSG clones in order to better characterize this family of genes. Since in rat P. carinii the 5' and 3' ends of the MSG genes are highly conserved, primers based on the published sequences of these regions of the human P. carinii MSG gene (with modifications to facilitate subcloning and expression) were used in an attempt to amplify additional full-length genes. By using DNA isolated from the lung of an AIDS patient with P. carinii pneumonia, a band of the correct size (approximately 3.1 kb) was amplified and subcloned. Five clones that differed in their restriction mapping and hybridization patterns were identified and sequenced (GenBank accession no. AF033208 to AF033212).

All clones encoded MSG variants that were clearly related but differed from each other. The coding regions of the clones varied in length from 3,054 to 3,087 bases, encoding proteins of 1,008 to 1,028 amino acids with predicted molecular sizes of 114 to 117 kDa. When pairs of clones are compared, they are 74 to 91% identical at the nucleotide level and 63 to 88% identical at the amino acid level. Overall, approximately 50% of the amino acids are conserved in all five clones. The clones are more closely related to each other than to rat P. carinii MSG genes. There is approximately 60% identity at the DNA level and 40% identity at the amino acid level between human P. carinii MSG and rat P. carinii MSG GP3.

To obtain additional P. carinii f. sp. hominis MSG sequences, a genomic library was screened with one of these clones, and one clone that hybridized strongly was subcloned and sequenced. This 12,792-bp clone (GenBank accession no. AF038556) contained three full-length MSG sequences and one partial MSG sequence in a head-to-tail tandem arrangement, similar to that previously reported (9, 34). One of the full-length MSG sequences did not have a complete open reading frame due to a frame shift between bases 6290 and 6347. The codon corresponding to a methionine at the beginning of rat P. carinii MSG clones encoded a valine in all the open reading frames, consistent with earlier observations (9, 34).

Figure 1 shows an alignment of the predicted proteins encoded by these genes, together with a rat P. carinii MSG sequence. Among the human P. carinii MSG sequences, there is substantial variability downstream of the amino terminus, while the region near the carboxyl terminus is highly conserved. For example, there is 63% identity in the last 100 amino acids among all the genes (excluding the region encoded by the PCR primer), which is about 5 times as high as the conservation among the first 100 amino acids (13% excluding the primer region). Like most known genes of P. carinii, all human P. carinii MSG genes show a strong AT bias, especially in the third position (approximately 70% A or T) (4, 9, 19, 41). As in other MSG molecules, cysteine residues of the human P. carinii MSG molecules are relatively numerous (5.7 to 5.9%) and are highly conserved: 96% of all the cysteine residues present in the human P. carinii MSG clones are conserved in all the clones. In a comparison between human MSG 11 and rat P. carinii MSG clone GP3, 94% of cysteine residues are conserved. The cysteine residues are unevenly distributed in four main regions and often show a pattern of two cysteines separated by six to seven amino acids, similar to the pattern seen in rat P. carinii (19). There is no predictable pattern to the intervening amino acids. All MSG proteins share a highly conserved amino acid domain rich in threonine and serine residues near the carboxyl terminus. Seven to 13 potential N-linked glycosylation sites [NX(S/T)] were observed in the MSGs. A premature stop codon was seen in MSG 32 after residue 1008, most probably due to a PCR artifact resulting in a point mutation; studies using the ligase chain reaction with primers specific for the mutation supported this conclusion (data not shown).

View larger version (152K): [in this window] [in a new window]   FIG. 1.   Alignment of the deduced amino acid sequences encoded by two of the human P. carinii MSG genes contained in the genomic clone (HMSGp1 and HMSGp3) and the five genes generated by PCR (HMSG11, HMSG14, HMSG32, HMSG33, and HMSG35), together with a published sequence (GBHMSG) and a rat P. carinii MSG sequence (RMSGGP3) (GenBank accession no. L05906). Alignment was performed with ClustalW. Amino acids that are conserved in at least six of the sequences are boxed. A methionine was substituted for valine at position 1 in the PCR clones during PCR to facilitate expression and thus is excluded from the alignment. The peptides that were synthesized and used to generate antipeptide antibodies are underscored with a solid box (conserved epitope) or a diagonally striped box (human MSG 32-specific epitope). The arrows flank the conserved region that was expressed in pET28a.

Because a major goal of these studies was to produce reagents derived from P. carinii f. sp. hominis that could be used to investigate immune responses to human P. carinii, a major effort was made to express a full-length P. carinii f. sp. hominis MSG gene at a high level. By using a baculovirus-insect cell system, a nearly full-length clone, MSG 32 (which contains the premature stop codon), was expressed. A time course showed that maximal expression occurred after 60 to 72 h of infection. The identity of the recombinant protein was confirmed by Western blotting using both an antibody against a peptide tag present in the vector and an antipeptide antibody raised against a peptide specific for MSG 32 (Fig. 2). No reactivity was seen when Sf9 cells alone or recombinant baculovirus-derived rat MSG GP3 was used as the target. The multiple bands seen in the Western blots, especially when the MSG-specific antipeptide antibody was used, likely represent protein degradation products, or possibly modification of the recombinant protein.

View larger version (74K): [in this window] [in a new window]   FIG. 2.   Immunoblots of recombinant nearly full-length human P. carinii MSG 32 expressed in a baculovirus system. Human P. carinii MSG 32 has a stop codon at amino acid 1010 that is likely a PCR artifact but results in elimination of the hydrophobic tail. For both blots, the lanes are as follows: lanes 1, recombinant baculovirus-derived rat P. carinii MSG GP3; lanes 2, recombinant baculovirus-derived human P. carinii MSG 32; lanes 3, Sf9 cells alone. (A) The reactivity of Xpress monoclonal antibody, which was generated against a peptide tag encoded by the vector, can be seen with a protein of approximate 116 kDa (arrow). (B) The reactivity of an antipeptide antibody generated against a peptide specific for human P. carinii MSG 32 can be seen with a protein of the same molecular size (arrow).

Although rat MSG GP3 could be produced at a high level in a baculovirus system and was easily purified by affinity chromatography using a nickel column (24), prolonged attempts to produce and purify high levels of human P. carinii MSG were unsuccessful. We then focused on expressing a highly conserved region at the carboxyl terminus of human P. carinii MSGs (Fig. 1). Preliminary data obtained by epitope mapping have demonstrated that this region is highly immunogenic for antibody production in both rats and humans (1a), suggesting that a peptide encompassing this region could be used in seroepidemiological studies.

PCR was used to amplify this conserved region without the carboxyl-terminal hydrophobic tail, since this hydrophobic tail could potentially interfere with expression and purification. A fragment of approximately 300 bp was obtained by PCR amplification using primers JK451 and JK452, with MSG 33 as a template. The PCR product was subcloned into pET28A, and expression was induced by culturing in 1 mM IPTG. High-level expression was observed within 2 h (Fig. 3A); no equivalent band was seen when pET28A was used without an insert under the same conditions (Fig. 3B). The presence of a six-histidine sequence in the expressed portion of the vector preceding the insert allowed rapid, one-step purification of the recombinant protein (Fig. 4). Although the yield was variable from experiment to experiment, during a typical study about 7 mg of purified protein was obtained from a 1-liter culture of E. coli. The identity of the protein was confirmed by immunoblotting using both the T7-tag monoclonal antibody and a polyclonal anti-epitope antibody generated in rabbits against an epitope (TITSTITSKITLTST) contained within the recombinant carboxyl-terminal fragment (Fig. 5A). No reactivity was seen with preimmune rabbit serum, with uninduced E. coli extracts (data not shown), or with the second antibody alone (Fig. 5C).

View larger version (89K): [in this window] [in a new window]   FIG. 3.   Time course of expression of a conserved region of human P. carinii MSG 33, as evaluated by SDS-PAGE (14% gel) and Coomassie blue staining. (A) Expression of recombinant protein (arrow) at different time points after induction with IPTG (1 mM). Maximal expression of the protein can be reached at 2 h (arrow). (B) The pET28A vector alone under the same conditions. No equivalent band is seen.

View larger version (72K): [in this window] [in a new window]   FIG. 4.   Purification of a conserved region of human P. carinii MSG 33 (arrow). Lane 1, whole E. coli extract following a 2-h induction with IPTG (1 mM) and solubilization in binding buffer containing 6 M urea; lane 2, flowthrough after binding of the recombinant protein to a nickel column; lane 3, eluate following elution without urea; lane 4, threefold-concentrated purified protein; lane 5, eluate following elution with 3 M urea. Elution with urea leads to a broader band on SDS-PAGE gels.

View larger version (36K): [in this window] [in a new window]   FIG. 5.   Immunoblots with a recombinant conserved human P. carinii MSG peptide. (A) Lane 1, reactivity of the T7-tag monoclonal antibody, which reacts with a pET28A vector-derived epitope that precedes the MSG peptide (arrow); lane 2, reactivity with a polyclonal anti-epitope antibody generated against a conserved epitope contained within the recombinant MSG fragment; lanes 3 and 4, reactivity with serum samples from healthy humans (diluted 1:100). (B) Reactivities of a variety of human serum samples (diluted 1:100). Lanes 1, 4 and 6, serum samples from healthy humans; lanes 2 and 7, serum samples from HIV-infected patients with a history of P. carinii pneumonia; lanes 3, 5, and 8, serum samples from patients with a history of P. carinii pneumonia but without HIV infection; lane 9, serum from an immunosuppressed patient without P. carinii pneumonia or HIV infection. All samples reacted by immunoblotting with the recombinant peptide (arrow). (C) Reactivity with no first antibody (lane 1) or reactivity of serum (diluted 1:100) from a cat, mouse, or rat (lanes 2 to 4, respectively). For lane 1, the second antibody was alkaline phosphatase-conjugated goat anti-human antibody; for each of the remaining lanes, the appropriate alkaline phosphatase-conjugated second antibody was used.

To evaluate the utility of this recombinant peptide as a tool for investigating the seroepidemiology of P. carinii infection, immunoblotting studies with a variety of human serum samples (diluted 1:100) were undertaken. Samples included those from 11 immunosuppressed patients with recent or acute P. carinii pneumonia but without HIV infection, from 5 patients with HIV infection and P. carinii pneumonia, from 17 patients with HIV infection but without P. carinii pneumonia, from 3 patients with neither HIV infection nor P. carinii pneumonia, and from 13 healthy laboratory workers. All 49 samples reacted by immunoblotting with the recombinant peptide (Fig. 5B). Because the recombinant peptide included a region that was vector derived, a subset of four samples was simultaneously evaluated by immunoblotting for reactivity with recombinant -galactosidase expressed in the same vector. None of the samples reacted with the recombinant -galactosidase (data not shown), demonstrating that the reactivity seen was against the P. carinii-derived peptide region. In addition, little or no reactivity was seen when rat, mouse, or cat serum was used (Fig. 5C).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In recent years, a great deal of interest has focused on the MSG of P. carinii, since it is both a likely virulence factor and a target of host immune responses (7, 8, 11, 22, 23, 28, 37-39, 45). Studies of human P. carinii infection ideally require P. carinii f. sp. hominis-derived reagents, given the high level of variation among homologous genes, including the MSG genes, isolated from different host-specific strains of the organism. We thus undertook to better characterize and express P. carinii f. sp. hominis MSG variants. A previous study reported that, like rat and ferret P. carinii MSG, human P. carinii MSG is also encoded by a family of related genes (1, 9, 19, 34, 44). The present study extends these observations by providing data on additional complete MSG genes.

This study has further confirmed that the structure of human P. carinii MSG is similar to those of the MSGs of P. carinii organisms infecting other hosts, with a high degree of variation near the amino terminus and a high level of conservation near the carboxyl terminus. The serine- and threonine-rich domains seen in rat P. carinii are also present in human P. carinii MSGs, but, as previously noted (9), a proline- and glycine-rich region present in rat P. carinii MSGs is absent from human P. carinii MSGs (Fig. 1). In addition, cysteine residues are highly conserved not only among human P. carinii MSGs but between human and rat P. carinii MSGs. It is probable that these cysteines play an important role in determining the tertiary structure of MSG; thus, conservation of these residues suggests that, despite substantial variation in the predicted primary amino acid sequences, all MSGs have similar tertiary structures. This would be consistent with a similar function for all MSGs despite the sequence variability.

The genomic clone confirms that valine rather than methionine is present at the beginning of the MSG open reading frame in a position homologous to that of the methionine in rat P. carinii MSGs. While this was initially interpreted as potentially representing an alternative translation initiation codon (9), recent studies with rat P. carinii have shown that the MSG variant genes are inserted in frame downstream of a leader sequence encoded in the single expression site for these genes (5, 35, 42, 43). Mouse P. carinii MSG cDNA clones also appear to encode a similar leader sequence (13). Thus, the translation initiation codon for P. carinii f. sp. hominis MSG genes is presumably a methionine present at the beginning of this leader sequence, which has not yet been characterized. The genomic sequences have confirmed that the Lys-Arg sequence located 5 positions downstream of the valine is conserved in all P. carinii f. sp. hominis MSGs identified to date. This is consistent with the postulated role of this sequence in all P. carinii strains as a target for cleavage by P. carinii kexin (20, 31), which would eliminate the conserved leader from the surface-expressed form of the MSG variants and maximize antigenic variation.

A second major focus of the present study was to express P. carinii f. sp. hominis MSG at a high level in order to provide reagents for studies of human-P. carinii interaction. While a nearly full-length MSG could be expressed in a baculovirus system, we were unable to induce high-level expression, perhaps because of the high proportion of cysteine residues. The presence of a hydrophobic tail may also interfere with expression, as the only full-length clone we were able to express had a premature stop codon prior to this hydrophobic tail.

However, an immunodominant region at the carboxyl terminus (minus the hydrophobic tail) could be expressed at a high level in a bacterial expression system and was easily purified via a six-histidine tag encoded in the vector. By immunoblotting, all 49 human serum samples tested were reactive with this peptide, regardless of their immune status or history of P. carinii pneumonia, supporting the hypothesis that most humans have been exposed to P. carinii at some point (22, 25). Reactivity was not seen with a control protein expressed in the same plasmid. Little or no reactivity was seen when serum from a rat, mouse, or cat was used (Fig. 5), suggesting that the responses are specific for human P. carinii and not a result of exposure to an irrelevant environmental antigen. The low level of reactivity seen with the rat serum may also represent cross-reactivity due to prior exposure to P. carinii f. sp. carinii, given that there is homology between rat and human P. carinii MSG sequences in this region (Fig. 1).

Previous studies evaluating the serological responses of humans to P. carinii have relied on a variety of assays, including fixed-tissue staining, immunofluorescence, enzyme-linked immunosorbent assays, and Western blotting using either rat or human P. carinii organisms or purified proteins as antigens. Results have been conflicting: some studies have shown that a high proportion of humans have serological reactivity with P. carinii antigens, while others have shown low response rates (2, 6, 15, 21-23, 26, 32). In previous studies using purified human P. carinii MSG, we noted response rates of 34 to 66% (23). Western blot studies have shown similar rates of reactivity with MSG (26). Differences between these previous results and the present results may be related to the sensitivity of the techniques, the integrity of the human P. carinii antigens used in prior studies (given that the antigens were derived from autopsy samples), or other methodological differences. Thus, this peptide is the first human P. carinii-specific antigen available in sufficient quantities for large-scale studies. The preliminary studies reported here suggest that this antigen may be a useful tool for seroepidemiologic investigation of P. carinii infection in humans, for example, to identify the period of seroconversion in humans, although more extensive studies are needed to verify this utility.

While antibodies appear to play a role in clearing P. carinii, T-cell response appears to be of primary importance in clearance (12, 30). Additional studies are needed to determine if this recombinant fragment contains T-cell epitopes in addition to B-cell epitopes. If so, evaluation of proliferative responses to this antigen may provide useful prognostic information, for example, about the risks of developing P. carinii pneumonia during HIV infection. In addition, given that it is a highly conserved region of the highly variable MSG, this recombinant peptide is a good candidate for evaluation as a vaccine for the prevention of P. carinii pneumonia.