EP0973876A1 - DNA ENCODING $i(PNEUMOCYSTIS CARINII) PROTEASE - Google Patents

Abstract

The invention relates to a novel Pneumocystis carinii protease with counterparts in P.carinii infecting various different species, including human, as well as nucleic acids encoding it.

Description

DNA ENCODING PNEUMOCYSTIS CARINII PROTEASE

This invention relates to a novel Pneumocystis carinii protease and to nucleic acids encoding it. The invention also relates to vectors containing the nucleic acids, to cells transformed with the vectors and to antibodies specific for the protease. In addition, the invention describes uses of all of the above.

The fungal pathogen Pneumocystis carinii causes potentially fatal pneumonia in the immunocompromised, including those receiving immunosuppressive therapy for organ transplantation, those with advanced malignancy and in particular those with HIV infection. The lack of an effective in vitro culture system still remains a major obstacle in the understanding of the biology of P. carinii and its interactions with its host. Molecular techniques have been employed in the study of the organism, and a number of genes have now been cloned. Among these is the multi- gene family encoding the major surface glycoprotein, (MSG or gpA) of the parasite.

The P.cannii major surface glycoprotein is highly mannosylated and is antigenically distinct in organisms isolated from different mammalian host species (Lundgren et al., 1991 ; Gigliotti, 1992). The MSG multi-gene family has been identified in the genome of P.carinii sp. f. carinii (rat-derived P.carinii) Kovacs et al., 1993; Wada et al., 1993; Sunkin et al., 1994), P.carinii sp. f. mustelae (ferret-derived P.carinii) (Haidaris et al., 1992; Wright et al., 1995), P.carinii sp. f. hominis (human- derived P.carinii) (Stringer et al., 1993) Garbe & Stringer, 1994) and

P.carinii sp. f. muris (mouse-derived P.carinii) (Wright et al., 1994). The different copies of P.carinii sp. f. carinii MSG genes are of similar size but heterogeneous in sequence. They have been found on multiple chromosomes and often organised in tandem arrays. The majority of MSG genes are located in the subtelomeric regions of the P.carinii sp. f. carinii chromosomes (Underwood et al, 1996; Sunkin & Stringer, 1996). The expression of MSG genes has been shown to be mediated by the upstream conserved sequence (UCS) which is found on a single chromosome situated in the subtelomeric region. Different copies of MSG have been shown to be linked to the UCS. It has been postulated that this differential expression of MSG may occur in a strategy to evade the immune response of the host by antigenic variation (Wada et al., 1995; Sunkin & Stringer, 1996).

Presently there are two standard treatments for Pneumocystis pneumonia, namely pentamidine or cotrimoxazole. These drugs were originally used because it was thought that Pneumocystis was a protozoan; only recently has genetic sequence analysis placed it in the fungal kingdom. Despite its classification as a fungus, Pneumocystis does not respond to the usual anti-fungal drugs and hence the drug regimes have remained all but unchanged. These regimes are particularly unpleasant with many patients reacting adversely, thus requiring a switch in treatment. Thus AIDS patients in particular would benefit from the development of new anti-PneL/mocys. s therapies since a high proportion of AIDS patients suffer adverse side effects, and many have multiple episodes of P. carinii pneumonia due to their decreasing CD4+ lymphocyte count and persistence of immune suppression.

Recently, a novel family genes from P.carinii sp. f. carinii has been described (Lugli and Wakefield 1996). The genes are found in the subtelomeric regions of the P.carinii sp. f. carinii genome, and show homology to protease genes from a number of fungi.

Wada and Nakumura (1994) describes the discovery of an open reading frame (designated ORF-3) encoding a protein of unknown function in P.carinii sp. f. carinii and located close to the MSG genes. The sequence given (DDBJ/EMBL/GenBank accession no. D31909 and D17441) corresponds to a portion of the genes discussed above (Lugli and Wakefield 1996).

It has now been discovered that there is a P.carinii sp. f. hominis counterpart to the family of genes in the rat-derived P. carinii species referred to above, the human-derived P. carinii species having at least 50% difference to the rat-derived P. carinii species in its nucleotide sequence. The novel multi-gene family is known as PPT7 (Protease 1); the genes show high levels of homology with the subtilisin-like serine proteases. The subtilisin-like serine proteases are a group of endoproteases which have been characterised from a wide variety of organisms including bacteria, fungi and higher eukaryotes. They have been found to function in the specific endoproteolytic processing of pro- proteins at cleavage sites of paired basic amino acid residues, to generate regulatory proteins in a mature and biologically active form. The pro- hormone processing enzyme kexin, encoded by the KEX2 gene of Saccharomyces cerevisiae has been characterised and found to cleave the precursors of the α-mating factor and the killer toxin (Fuller et al., 1989). Genes encoding a similar processing endoprotease have been identified in a number of other fungi, the KEX1 gene from the yeast Kluyveromyces lactis (Tanguy-Rougeau et al., 1988J, the gene encoding the EX2-. elated protease (krpj from Schizosaccharomyces pombe (Davey et al., 1994) and the XPR6 gene from Yarrowia lipolytica (Enderlin & Ogrydziak, 1994). Mammalian homologues have also been identified incluidng the human fur gene (fes upstream region) in the region upstream of the fes proto- oncogene, encoding the enzyme furin (van den Ouweland et al., 1990). The genes Dfurl and Dfur2 from the insect Drosophila melanogaster encoding furin-like proteins (Roebroek et al., 1992) and the bli-4 gene from the nematode Caenorhabditis elegans have also been studied. Other members of the subtilisin-like serine protease family have been identified and the specific endoproteolytic activity of some of them has been elucidated. However for many others, the precise biological function has not yet been determined.

The PRT1 gene product may be a specific endoproteolytic processing enzyme, such as is seen in other subtilisin-like serine proteases. Given that in genetic organisation some copies of PRT1 are generally found in the subtelomeric region, just downstream from the MSG gene, the PRT1 protein encoded by these genes may be involved in the processing of MSG to its mature form. The multicopy nature of the PRT1 gene may reflect the need for processing of enzymes of different specificity for the different types of MSG. Whatever its precise role, the activity of the PRT1 protein is undoubtedly essential to the viability and therefore the pathogenesis of P.carinii.

Recently, there has been considerable interest in targeting proteases, for the control of a number of different diseases and in particular HIV infection. Combination therapies for HIV treatment employ protease inhibitors; a large variety of protease inhibitors are therefore available for testing against new proteases. The Invention Part of the catalytic domain of a PRT1 gene has been cloned, sequenced and characterised from three types of the host specific fungal pathogen P.carinii, namely P.carinii sp. f. rattus (rat variant), P.carinii sp. f. muris (mouse) and P.carinii sp. f. hominis (human). The newly discovered human-infecting P.carinii PRT1 catalytic domain sequence is shown in figure 1 and nucleotide sequence alignments for rat P. carinii, rat variant P. carinii, mouse P. carinii and human-infecting P.carinii PRT1 clones are shown in figure 2. These will enable the sequencing of the remaining parts of a PRT1, using techniques known to those skilled in the art of molecular biology. The invention therefore provides in one aspect an isolated DNA comprising part or all of a PRT1 gene of a non-rat infecting species of Pneumocystis carinii.

The invention also provides an isolated DNA comprising a sequence shown in figure 1 , or a non-rat P. carinii sequence shown in figure 2, or a sequence which hybridises to either of these under stringent conditions.

In further aspects, the invention provides recombinant vectors containing PPT7 DNA sequences as described herein, and recombinant polypeptides which are part or all of a PRT1 gene product, encoded by the vectors.

In another aspect, the invention provides synthetic peptides corresponding to antigenic portions of a PRT1 gene product.

In further aspects, the invention provides a method of producing antibodies specifically immunoreactive with a P.carinii protease, which method comprises using a recombinant polypeptide or a synthetic peptide as described herein to generate an immune response; and antibodies produced by the method.

In another aspect, the invention provides a method of screening for anti-Pπet/mocys /s carinii compounds, which method comprises providing a source of a recombinant polypeptide expressed by part or all of a PRT7 gene or cDNA, and contacting the compound with the recombinant polypeptide.

In another aspect, the invention provides an engineered cell transfected with a recombinant vector containing PRT1 DNA sequences as described herein.

In another aspect, the invention provides an engineered cell line expressing a recombinant polypeptide from part or all of a PPT7 gene or cDNA, useful in a method of screening for anti-P carinii compounds such as protease inhibitors effective against P.carinii. In another aspect, the invention provides a P.carinii protease isolated using an antibody specifically immunoreactive with a P.carinii protease, as described herein.

In another aspect, the invention provides PPT7 clones for part or all of a human-infecting P.carinii PRT1 gene from the PRT1 multi- gene family.

A part of the PRT1 gene as referred to herein may be for example a fragment of the gene which codes for a specific domain such as the catalytic domain, or it may be a shorter sequence such as a sequence not less than 15 nucleotides in length or not less than 20 nucleotides in length. Sequences of about 15 or about 20 nucleotides in length are generally the shortest practical length of oligonucleotide useful as a sequence specific primer or probe. That is, these are generally the shortest lengths of sequence that will hybridise specifically to a gene sequence under stringent conditions.

Within the PRT1 multi-gene family will be related genes which will be easily identifiable as such by those skilled in the art, but which may nevertheless differ in location, function and sequence. It will be evident that all members of the PRT1 multi-gene family, which members may variously be described as different genes in the family or as different copies of the PRT1 gene, are included within the scope of the invention.

Known methods to mutate or modify nucleic acid sequences can be used in conjunction with this invention to generate useful PRT1 mutant sequences. Such methods include but are not limited to point mutations, site directed mutagenesis, deletion mutations, insertion mutations, mutations obtainable from homologous recombination, and mutations obtainable from chemical or radiation treatment.

Furthermore, recombinant DNA techniques are available to mutate the DNA sequences described herein, to link these DNA sequences to expression vectors and express the PRT1 protein or part of the protein eg. the catalytic domain or the P-domain. In the attached figures:

Figure 1 shows the genomic DNA sequence of part of the catalytic domain of PRT1 from P.carinii sp. f. hominis. (SEQ ID NO: 22)

Figure 2 shows DNA sequence alignments for part of the catalytic domain of PRT1 from P.carinii. (Found in GenBank AF001305, GenBank

AF001304, and SEQ ID NOS: 23 - 29, in the order in which they appear).

Figure 3 shows amino acid sequence alignments of part of the catalytic domain of PRT1, translated from the nucleotide sequences in figure 2.

(Found in GenBank and SEQ ID NOS: as for Figure 2).

Figure 4 shows alignment of P.carinii PRT1 derived amino acid sequences from P.carinii sp. f. carinii clones. (Found in GenBank AF001305,

GenBank AF001304 and SEQ ID NOS: 30, 31 , 33 - 47, 32, 48 - 50). Figure 5 shows DNA sequence alignments for P.carinii sp.f. carinii PRT1 clones. (Found in GenBank AF001305, GenBank AF001304 and SEQ ID

NOS: 30 - 32)

Figure 6 shows a schematic representation of the P.carinii sp. f. carinii

PRT1 gene. Figure 7 shows expressed recombinant PRT1 fragments.

By analogy to P.carinii sp. f. carinii there are expected to be many copies of the PRT1 gene within the P.carinii sp. f. hominis genome.

Some of these copies may be significantly different and form a number of different sub-types. They will all, however, be classed as members of the PRT1 multi-gene family by virtue of homology at some domains of the gene, for example the catalytic domain.

Seven different domains have been identified to date in the

P.carinii sp. f. carinii PRT1 amino acid sequence, namely: i) N-terminal hydrophobic domain ii) Pro-domain iii) Catalytic domain iv) P-domain v) Proline-rich domain vi) Serine-threonine rich domain vii) C-terminal hydrophobic domain

The P.carinii sp. f. hominis homologues may have fewer, the same number or more domains. Although some domains in some members of P.carinii sp. f. hominis PRT1 gene family may be absent or some extra domains may be present, these genes will still be considered to be members of the PR7^~7 multi-gene family.

The proteins encoded by different copies of this gene family may have a variety of different functions, including: i) as a constituent of the outer cell surface of the parasite, and attached to the cell membrane by a glycosyl- phosphatidylinositol (GPI) anchor ii) the proteolytic processing within a P. carinii sub-cellular organelle of the P.carinii major surface glycoprotein (MSG) to its mature form, possibly at a conserved dibasic amino acid site in the upstream conserved sequence of MSG iii) in the interaction of the parasite with its host, forming a specific ligand on the parasite cell surface which binds to a host receptor molecule

There may be other functions of the members of this gene family which have not yet been recognised. These may include functioning as a protease on as yet unidentified pro-proteins, or as a structural glycoprotein at some life-cycle stage of the parasite.

It has been demonstrated that the protease is a surface protease. Therapeutic intervention The PRT1 protein presents a target for a variety of different therapeutic interventions, which may include: i) Inhibitors of protease activity

It is postulated that the proteolytic activity of PRT1 is essential for the viability of the parasite. The predicted structure of the catalytic domain of the PRT1 protein suggests that there are subtle differences compared to other such proteases so far studied. These differences may be exploited in the design of specific drugs, with less toxic side- effects than seen in the present available treatments, ii) Vaccines

Available data indicates that some copies of PRT1 may comprise a major surface antigen and therefore provide a potential target for vaccine development.

iii) Immunotherapy

Passive immunisation with antibodies to PRT1 may be protective, iv) Analogues Analogues designed to imitate PRT1 may be active in blocking the adherence of P.carinii organisms to a receptor on the human cells.

Identification of a subtilisin-like serine protease in P.carinii so. f. carinii

METHODS

P.carinii DNA extraction

P.carinii infection was induced in Sprague Dawley rats by steroid immunosuppression. The organisms were isolated and purified from infected rat lung tissue by the method described by Peters et al., (1992). Genomic P.carinii DNA was extracted by digestion with proteinase K (1 mg/ml) in the presence of 0.5% SDS and 10mM EDTA, pH8.0, at 50°C for 16h, followed by phenokchloroform extraction and ethanol precipitation. P.carinii DNA for use in PFGE experiments was prepared in SeaPlaque GTG agarose as described by Banerji et al., (1993).

For oligonucleotide primers, see Table 1 and Lugli et al 1997. Isolation of copies of the PRT1 gene from P.carinii sp. f. carinii genomic and cDNA libraries

A copy of the PPT7 gene was isolated from an unamplified genomic library from P.carinii sp. f. carinii constructed in λEMBL3 (Banerji et al., 1993). The library was screened with a cDNA clone containing a region of a P.carinii sp. f. carinii MSG gene (GenBank Accession number GBPLN:PMCANTIA, donated by Dr C J Delves and Dr F Volpe). A relatively high number of recombinant plaques gave positive hybridization signals compared to the positive recombinant plaques when the library was screened with a probe derived from the single copy arom locus (Banerji et al., 1993). Five recombinant phages were isolated from the tertiary screen and the DNA was subcloned into the plasmid vector pBluescriptl I .

In order to isolate a full cDNA clone, a P. carinii sp. f. carinii cDNA library constructed in λZAPII (donated by Dr CJ Delves and Dr F Volpe, see Dyer et al., 1992), was screened with PCR products derived from amplification of the 5' end of the gene with oligonucleotide primer pair pcprot9 and prp4r (9/4r product), and of the 3' end of the gene with pcprot13/RI and pcprot12/RI (13/12 product). The primary screening was carried out using both probes, and the secondary and tertiary screens were carried out using only the 9/4r product. The number of positive clones when screening the cDNA library with the two probes appeared to be relatively high when compared to the number obtained using a single copy gene. Four recombinant phage isolated from the cDNA library were partially characterized. The recombinant DNA was recovered from the λ phage by in vivo excision as pBlueScript plasmid DNA. The size of the recombinant DNA ranged from 2.7kb to 2.9kb, and sequence analysis revealed that all four clones contained a polyA tail. One recombinant, 73j was selected for further analysis and the recombinant DNA was sequenced in full from both strands. DNA amplification

Oligonucleotide primers were designed to various regions of the P.carinii PRTI nucleotide sequences. Some oligonucleotides had an EcoRI restriction endonuclease site incorporated at the 5' end to facilitate cloning of the amplification products into EcoRI-digested plasmid vectors pBluescript SK(-) (Stratagene) or pUC18 (Pharmacia). The final concentration of the amplification reaction mix was 50mM KCI, 10mM Tris (pH8.0), 0.1 % Triton X-100, 3mM MgCI₂, 400μM (each) deoxynucleoside triphosphate, 1 μM oligonucleotide primer and 0.025 U Taq polymerase ml^"1 (Promega, UK). With primer pair pcprotθ and pcprotI O, forty cycles of amplification was performed at 94°C for 1.5 min., 53°C for 1.5 min., and 72°C for 2.0 min. With primer pair pcpro.9 and pcprot4r the same conditions were used, except an annealing temperature of 50°C was used. With all other primer pairs, ten cycles of amplification were carried out at 94°C for 1.5 min., 55°C for 1.5 min., and 72°C for 2.0 min, followed by 30 cycles of 94°C for 1.5 min., 63°C for 1.5 min., and 72°C for 2.0 min. Negative controls were included in each experiment.

The entire putative gene was amplified as three overlapping fragments, Prpδe (1626 bp), M14 (1279 bp) and Prp2g (251 bp). Oligonucleotide primer pairs pcprotθ with pcprotI O, followed by pcprotδ/RI with pcprot4/RI were used in a nested PCR to amplify the 5' fragment, designated Prp5e, of length 1626 base pairs (bp). The second portion, called M14, spanning 1279 bp of the central region oi PRTI, was amplified using a nested PCR with primer pairs pcprot2/RI with pcprotl4/RI, followed by pcprot7/RI with pcprot12/RI. The third fragment, Prp2g, encompassing the 3' end of the sequence (251 bp), was amplified using oligonucleotides primers pcprot13/RI and pcprot14/RI (Table 1 and Lugli et al 1997).

Five different overlapping regions of the PRTI gene were also amplified, cloned and the DNA sequences were determined. The first region amplified with primer pair pcprot1/RI and pcprot3/RI spanned approximately half of the subtilisin-like catalytic domain, the second region amplified with primer pair pcprot2/RI and pcprot4/RI spanned the end of the subtilisin-like catalytic domain and the start of the P-domain, the third region amplified with primer pair pcprot7/RI and pcprot8/RI spanned the P-domain, the fourth region amplified with primer pair 36ex/RI and P.3/RI spanned the proline-rich domain and the fifth region amplified with primer pair pcprot13/RI and pcprot 14/RI spanned the C-terminal hydrophobic domain. The sequences Prpla, Prp3a, Prp7a, Prp2c, Prp3c, Prp4c, Prptaf2, Prpf4, Prpδf, Prpg3 and Prpδg were amplified from the P. carinii cDNA library, and sequences Pcr-19, Pcr-14, Pcr-5, Pcr-3, Pcr-1 , Lam-1 and Prpg4 from the P.carinii genomic DNA (Figure 4). DNA sequence analysis

DNA sequence analysis was performed using the dideoxy chain termination method. Sequence data was obtained in full from both strands for all sequences. Analysis of the sequence data was carried out using the University of Wisconsin Genetics Computing Group (UWGCG) Sequence Analysis Software Package, Version 8, 1994, Genetics Computer Group, Madison, Wisconsin. Pulsed Field Gel Electrophoresis P. carinii sp. f. carinii organisms were isolated from an infected rat lung and the chromosomes were separated by pulsed field gel electrophoresis (PFGE), using a Contour Clamped Homogeneous Electric

Field (CHEF) DRII apparatus (Bio-Rad, UK) operated at 4°C. Electrophoretic separation was achieved using 0.9% Seakem agarose gel with initial switching time of 10 sec increasing to a final switching time of 60 sec at 180 V for 48 hours. A karyotype corresponding to P.carinii sp. f. carinii form 1 was observed (Cushion er a/., 1993). Southern hybridisation

Southern blotting and hybridization were carried out using standard techniques (Sambrook et al., 1989). PFGE blots were hybridised with three probes derived from different domains of the PP7^~7 gene. The product 9/4r was derived from amplification of the 5' end of the PRT1 gene with primer pair pcprotθ and pcprot4r/RI, product 2/4 from amplification of the central catalytic region with primer pair pcprot2/RI and pcprot4/RI, and product 13/12 from amplification of the 3' end of the gene with primer pair pcprot13/RI and pcprot12/RI. The amplification products were gel-purified (GeneCleanl l , BIO101) and labelled with [α-³²P]-dCTP by random priming (Megaprime, Amersham). Hybridisation was carried out at 45°C and stringency washing at 60°C in 0.2xSSC and 0.1 % SDS. Southern blots of genomic P.carinii DNA digested with restriction endonuclease Ps l or SamHI were probed with oligonucleotide probes pcprot3/RI, pcprot5/Rl, pctel2, and msgterm, labelled with [γ-³²P]- dATP using polynucleotide kinase. Hybridisation was carried out at 46°C and stringency washing at 52°C in 5xSSC and 0.5% SDS.

RESULTS

Analysis of DNA and deduced amino acid sequence of copies of the

PRT1 gene

We have identified a family of genes in the P.carinii sp. f. carinii genome which shows homology to the subtilisin-like serine proteases. We have named this gene family PPT7 (protease 1). A copy of the PRT1 gene (Paga) was isolated from a P.carinii genomic library, the open reading frame (3069bp) containing seven short putative intervening sequences. A copy of the PPT7 gene (73j) was also isolated from a cDNA library, of length 2370bp. Portions of the gene were amplified by PCR from the cDNA library as three overlapping fragments, at the 5' end (Prpδe), the central region (M14) and the 3' end (Prp2g). Five other regions of the gene were also amplified, from either the P.carinii cDNA or genomic libraries.

Analysis of the DNA sequence of the copy of the PRT1 gene from the genomic library, PRT1 (Paga), and of the copy from the cDNA library, PPT7 (73j), confirmed the presence of seven short introns in the genomic DNA sequence. The introns ranged in length from 38 bp to 45 bp, with a base composition ranging from 71% to 84% A+T. In all seven introns, the dinucleotide GT was present at the 5' splice donor site and AG at the 3' splice acceptor site. The sequence YTRAT, which has been identified as the putative lariat forming motif in other P.carinii sp. f. carinii introns (Zhang & Stringer, 1993), was present in the first, second, fourth, fifth and seventh intron. The eukaryotic lariat consensus sequence, YYRAY, was identified in the third and sixth intron. The sequence of the cDNA clone, PRT1 (73]), contained an open reading frame of 2370bp, which on translation resulted in a peptide of 790 amino acids (Figure 4). The deduced amino acid sequence was compared to sequences in the GenBank and EMBL databases and showed homology to fungal and other eukaryotic subtilisin-like serine proteases. The A+T content of the ORF was 64%, with a high A+T content at the third base position of the codons. The base composition of the 5' upstream sequence was 74% A+T, and the 3' downstream sequence was 75% A+T. A consensus polyadenylation signal, AATAAA, was observed 68bp downstream of the stop codon. The deduced amino acid sequence of the genomic clone

PPT7(Paga), the cDNA clone PRT1(73\), the three fragments obtained by PCR amplification of the cDNA library and the other recombinant clones generated by DNA amplification were compared (Figure 4). Several regions of homology were found and also a number of regions in which significant divergence was observed. These data suggested that the sequences were derived from different copies of the PPT7 gene. Comparison with other subtilisin-like serine proteases

The deduced amino acid sequence of the cDNA clone PRT1(73 ) was aligned with nine other subtilisin-like serine proteases including fungal, mammalian, insect and nematode sequences. The PRT1 sequences showed homology with all the other sequences, with a high level of homology in the subtilisin-like catalytic domain. The three essential residues of the catalytic active site, aspartic acid (Asp₂₁ ), histidine (His₂₅₂) and serine (Ser₄₂₃) were conserved in all the PRT1 sequences. The highest levels of homology between all the sequences were around these residues.

The structural organisation of the fungal sequences showed domains characteristic of this class of processing endoproteases, a hydrophobic signal sequence, a pro domain that may be cleaved by autoproteolysis, a subtilisin-like catalytic domain, a P-domain which is known as such because it is essential for proteolytic activity, a serine/threonine-rich domain which may potentially be modified by O-linked glycosylation, a carboxy-terminal hydrophobic trans-membrane domain and a C-terminal tail with acidic residues (Van de Ven et al., 1993) The P.carinii PRT1 sequences showed a putative similar structural organisation but unlike the nine other subtilisin-like serine proteases, they also had a proline-rich domain preceeding the serine-threonine rich domain and the C- terminal hydrophobic domain (Figure 6). The P.carinii PRT1 (73j) sequence had a hydrophobic signal sequence at the N-terminus, followed by a putative pro-domain, a subtilisin-like catalytic domain from Ser₁₇₁ to His₄₇₄, a P-domain from residue Tyr₄₇₅ to Ser₆₃₁, a proline-rich domain from residue Pro₆₄₁ to Pro₇₀₇, a serine-threonine rich domain from residues Thr₇₀₈ to Ser₇₆₅, and a carboxy-terminal hydrophobic domain from residues His₇₇₁ to Phe₇₉₀. Analysis of subtilisin-like catalytic domain

The three-dimensional structures of four subtilisin-like serine proteases have been determined, subtilisin BPNVNovo from Bacillus amyloliquefaciens (Hirono et al., 1984; Bott et al., 1988), subtilisin Carlsberg from β. licheniformis (McPhalen & James, 1988), thermitase from Thermoactinomyces vulgaris (Gros et al., 1989; Teplyakov et al., 1990) and proteinase K from Titirachium album (Betzel et al., 1988). The amino acid sequence of these four proteases has been compared to that of 31 other subtilisin-like serine proteases isolated from bacteria, fungi and higher eukaryotes and the essential core structure of the catalytic domain of this group of molecules has been identified (Siezen et al., 1991).

We have compared the deduced amino acid sequence of the P.carinii PRT1 (73j) gene with the multiple sequence alignment of the other subtilisin-like serine proteases and have identified the three essential residues of the catalytic active site aspartic acid, histidine and serine in the PRT1 sequence (Asp₂₁₄, His₂₅₂ and Ser₄₂₃). On the basis of the sequence alignment, the P.carinii PRT1 sequence could be assigned to the class 1 subtilases, within the subgroup l-E which contained the pro-hormone processing proteases from yeasts and higher eukaryotes (Siezen et al., 1991 ).

Eight α-helical domains and nine β-sheet regions have been defined as the structurally conserved regions within the essential core structure. The variable regions which connect the core segments have been found to differ both in length and in amino acid sequence (Siezen et al., 1991). High levels of homology were observed between the PRT1 sequences and the other sequences in the regions of the two conserved internal helices, helix C (residues 252 to 262) and helix F (residues 422 to 438). Eleven amino acid residues have previously been found to be totally conserved in all the characterized subtilisin-like serine proteases, and most but not all are conserved in the PRT1 sequences. These amino acid residues are at the active site, Asp₂₁₄, His₂₅₂ and Ser₄₂₃, [found in all the PRT1 sequences except PRT1 (Prp7a)] and in the internal helices at residues Gly₂₅₃, Gly₂₅₈, Pro₄₂₇. The residues Ser₃₁₀, Gly₃₁₂, Gly₃₅₁, Gly₄₂₁ and Thr_422] involved in substrate binding, were conserved in all the PRT1 sequences, except Thr₄₂₂ which was found only in two sequences generated by PCR, PRTI (Prpla) and PRT1 (Prp7a).

In addition to the totally conserved residues, seven other amino acid residues have been identified which are highly conserved, of these six were conserved in the P.carinii PRT1 sequences and included the oxyanion hole residue (Asn₃₅₂), residues near the active site, Gly₂₁₆, Thr₂₅₄, and also residues Gly₂₀₅, Gly₂₇₁ and Gly₃₄₃. Seven conserved cysteine residues were found in all the P.carinii PRT1 sequences, Cys₂₅₆, Cys₂₆₈, Cys₃₀₉, Cys₃₅₉, Cys₃₈₉, Cys₃₉₁ and Cys₄₁₅. Nineteen variable regions, generally located in loops on the surface of the molecule, have been identified in the subtilase family, of which 14 were found in the P.carinii PRT1 sequences. Three positions have been identified at which charge is totally conserved in all the subtilisin-like proteases examined, and these were also conserved in the P.carinii PRT1 sequences, the positive charge on Arg₂₈₂ and the negative charges on residue Asp₂₁₄ (active site) and Asp₂₂₃.

It has been proposed that the high specificity of the class l-E subtilisin-like serine proteases for paired basic residues Lys-Arg or Arg-Arg may be facilitated by a high density of negative charge at the substrate- binding face, provided by nine highly conserved Asp residues and one Glu residue (Siezen et al., 1991). Two of the Asp residues, Asp₃₅₃ and Asp₄₀₉ were found in all the P.carinii PRT1 sequences and also the Glu₂₉₃. In addition, four other Asp residues were found in some but not all of the copies of PRT1. Analysis of the domains flanking the subtilisin-like catalytic domain

The putative domains of the PRT1 (73j) polypeptide are summarised in Figure 6. A hydrophobicity plot of the PRT1 (73j) sequence revealed a hydrophobic region at the N-terminus suggesting that this may be a signal sequence. Residues 1 to 23 of the N-terminus of the sequence showed a high level of homology to the N-terminus of the P.carinii sp.f. carinii multifunctional folic acid synthesis fas gene which encodes dihydroneopterin aldolase, hydroxymethyldihydropterin pyrophosphokinase and dihydropteroate synthase (Volpe er a/., 1992, 1993). This region was followed by the presumptive pro-domain, which may be cleaved by autocatalysis. Potential autocatalytic sites of paired basic residues were identified in the PRTI (Paga) and PRT1 (Prp5e) sequences at Lys₁₁₅ - Arg₁₁₆ and Arg₁₃₆ - Arg₁₃₇, but were absent in the PRT1 (73j) sequence. Five other semi-conserved autocatalytic sites were found in some copies, but not all, of the P.carinii PRT1 sequences, two in the catalytic domain (Lys₄₀₀ - Arg₄₀₁, Arg₄₇₃ - Arg₄₇₄), three in the P-domain (Arg₅₂₁ - Arg₅₂₂, Arg₅₅₅ or Lys₅₅₅ - Arg₅₅₆, Arg₅₇₆ - Arg₅₇₇). One potential autocayalytic site at the start of the carboxy-terminal hydrophobic region (Lys₇₆₉ - Arg₇₇₀), which was found in all the sequences. The PRT1(73j) sequence contained two of the potential autocatalytic sites, Arg₅₇₆ - Arg₅₇₇ and Lys₇₆₉ - Arg₇₇₀.

The PRT1 sequences showed homology with the other subtilisin-like serine proteases in the region of the P-domain, the highest homology being with the derived amino acid sequence of the S. pombe krp gene. Four potential sites for N-linked glycosylation were observed in all the PRT1 sequences, three in the subtilisin-like catalytic domain (Asn₁₉₄, Asn₂₇₇, Asn₄₄₂), and one in the P-domain (Asn₆₀₃).

A serine-threonine rich region was also identified in the PRT1 (73j) sequence from residue Thr₇₀₈ to Ser₇₆₅, and the hydrophobicity plot of the PRT1 (73j) sequence revealed a hydrophobic region at the C- terminal end, residues His₇₇₁ to Phe₇₉₀, suggesting a membrane-associated domain. Unlike most other serine protease sequences, however, all the copies of the PRT1 polypeptide contained a proline-rich region downstream of the P-domain. Genetic organization of the PRT1 multi-gene family

Analysis of the alignments of the DNA and the deduced amino acid sequences of copies of the PRT7 gene from genomic DNA, the cDNA sequence and the three fragments obtained by PCR of the cDNA library revealed domains in the PRT1 gene which were highly conserved and also regions where significant divergence was observed, again suggesting that PP7^~7 comprises a multi-gene family (Figure 4). The subtilisin-like catalytic domain and the P-domain appeared to be conserved whereas high levels of heterogeneity were observed in the proline-rich domain and the C-terminal domain. The variation in this region was both in length and in sequence. A number of repeated DNA sequence motifs were found in the proline-rich region. Nucleotide sequences encoding polyproline were found in all the sequences, and also the dipeptides Pro- Glu and Pro-Gin and the tetrapeptides Pro-Glu-Pro-GIn and Pro-Glu-Thr- Gln. The order and number of tandem repeats varied in each sequence. The overall length of this region varied from approximately 67 amino acid residues in the shortest sequence, PRT1 (73j), to 233 residues in the longest sequence, PRT1(M14).

In order to further substantiate the presence within the P. carinii genome of multiple copies of the PP77 gene, P.carinii sp. f. carinii chromosomes, separated by pulsed field gel electrophoresis, were analysed by hybridisation with three probes derived from different domains of PRT1. All three probes showed similar patterns of hybridization, anealing at high stringency to all the chromosome bands except for one, the third smallest in size, approximatey 350Kbp. This provided further evidence that the P.carinii sp. f. carinii genome contained many copies of the PPT7 gene, which were present on most of the P.carinii sp. f. carinii chromosomes.

The sequences of the PPT7 gene family showed high levels of homology with ORF3, which has been demonstrated to be contiguous with a copy of the gene encoding the major surface glycoprotein MSG100 (Wada & Nakamura, 1994). This gene arrangement was reported in 15 other λ clones, in which a gene showing high homology to ORF3 was located downstream of a copy of MSG (Wada & Nakamura, 1994). Most copies of the MSG genes have been demonstrated to be located in the P.carinii sp. f. car//.// subtelomeric regions (Underwood et al., 1996; Sunkin & Stringer, 1996). The copy of the PRT1 gene encoded by the PRTI (Paga) sequence was cloned from a λ EMBL3 genomic library as a single 14kb fragment and was approximately 1150bp downstream of a copy of MSG. Four other λ clones isolated from the same library contained a copy of PPT7 contiguous with a copy of MSG.

P.carinii sp. f. carinii genomic DNA was digested with either restriction endonuclease Pstl or βamHI and probed sequentially with four oligonucleotide probes, derived from the 5' end of PPT7 gene (pcprot5/RI), from the catalytic domain of the gene (pcprot3/RI), an MSG probe (msgterm) and a subtelomeric probe (Pctel2). All probes hybridised to multiple bands. The hybridisation pattern of some of the bands, ranging in size from 7kb to greater than 12kb, were the same for all four probes. However, hybridisation to other fragments was not coincident, with the PRT7 probes alone hybridising to some high molecular weight fragments and also low molecular weight fragments of less than 7kb.

DISCUSSION

We describe the cloning and characterisation of copies of the PPT7 multi-gene family from P.carinii sp. f. carinii. A copy of the PRT7 gene was isolated from a P.carinii sp. f. carinii genomic library. A different copy was isolated from a cDNA library, indicating that this copy of the gene was transcribed, and also identifying the presence of seven short introns in the genomic sequence. Consistent with many other P.carinii genes, the coding region and the flanking sequences of the PRT1 sequences showed a strong bias for adenine or thymine, and in particular at the third base position of the codons. Similarly, the presence of short A+T rich introns has been reported in other P.carinii genes. In the PRT7 sequences, the introns were not distributed throughout the gene, but six of the seven introns were found in the subtilisin-like catalytic domain, and the seventh in the P-domain. The introns may play a role in restricting the variation in this region of the gene, whereas no introns were observed in the highly heterogeneous proline-rich region (Rogers, 1985).

The high level of homology of the P.carinii PRT1 sequences to the subtilisin-like serine proteases, and in particular in the region of the catalytic domain, strongly suggested that this gene encoded a protease of this type. The predicted P.carinii PRT1 polypeptide sequences possessed the three essential residues of the catalytic active site as well as many other highly conserved motifs. The domain organisation of the PPT7 gene strongly resembled that of the fungal prohormone processing proteases, with the exception of the proline-rich domain. This proline-rich region is very uncommon in the subtilisin-like serine protease superfamily, although the KRP6 gene from Y. lipolytica is reported to contain a short region of a tetrapeptide repeat, the consensus sequence of the four amino acids being Glu (Asp/Glu) Lys Pro (Enderlin and Ogrydziak, 1994). A proline-rich region has also been found in the carboxy-terminal tail domain of the mammalian serine protease acrosin, a proteolytic enzyme of sperm cells, located in the acrosome at the apical end of the spermatozoan (Klemm et al., 1991).

In the African trypanosome, Trypanosoma brucei, a proline- rich domain has been identified in the procyclic acidic repetitive proteins (PARPs). These proteins are found on the cell surface of the insect form of the parasite and are encoded by a family of polymorphic genes which contain a variable region with heterogeneity both in length and sequence. The variable region contains the proline-rich domain and is primarily composed of the dipeptide Glu-Pro (Roditi et al., 1989).

Unlike any of the other fungal prohormone processing proteases, which appear to be single copy genes, the data reported in this study suggest that the PPT7 sequence is present in many copies, which are similar but not identical, in the genome of P.carinii sp. f. carinii. The relatively large number of recombinants present in both the genomic and the cDNA libraries suggested a multi-copy gene and this was substantiated by PFGE data, revealing that at least one copy of a PRT1 gene was present on all but one of the P.carinii chromosomes. Southern hybridisation of restriction endonucleolytic digests of P.carinii sp. f. carinii DNA probed with PRT1 sequences also confirmed the presence of many copies of the gene. Analysis of sequence data generated by the amplification of the locus showed heterogeneity, suggesting that a variety of different copies of the gene were present in the P.carinii genome. Some domains, including the subtilisin-like catalytic domain and the P-domain, were highly conserved between gene copies, whereas the highest levels of divergence were observed in the proline-rich domain, which varied both in length and in sequence.

Of five genomic clones analyzed in this study, all possessed a copy of PRT1 contiguous with a MSG gene. It has been reported that 15 independent genomic clones which encoded MSG were contiguous with the ORF3 sequence, which from our analysis, appears to encode the proline-rich domain of PPT7 (Wada & Nakamura, 1994). It has been demonstrated that most copies of MSG are subtelomeric (Underwood et al., 1996, Sunkin & Stringer, 1996). It is therefore highly likely that many copies of the PRT1 multi-gene family are located in the subtelomeric regions of the P.carinii sp. f. carinii genome. However PFGE analysis has shown that not every P.carinii sp. f. carinii chromosome contained a copy of PRT1, and the preliminary characterisation of a clone of one of the subtelomeric regions of P.carinii sp. f. carinii has not revealed a copy of PRT1 (Underwood & Wakefield, unpublished results). Hybridisation of MSG and subtelomeric probes to endonuclease digested P.carinii sp. f. carinii DNA resulted in positive hybridisation to fragments greater than approximately 7 kb in size. Probes derived from the PPT7 sequence hybridised to these bands but also to low molecular weight fragments, again suggesting that not all copies of PPT7 are subtelomeric.

The P.carinii PRT1 gene family shows some striking similarities to that of MSG. Both are composed of many genes, copies of which are found on most P.carinii chromosomes and show sequence heterogeneity. Some copies of PPT7 are contiguous with MSG and are located in the subtelomeric regions of the P.carinii chromosomes.

It is interesting to note that one of the major components of the cell surface of Leishmania has proteolytic activity. The Leishmania major surface protease (msp or gp63), a zinc endoprotease, is found in all species of Leishmania and is encoded by a family of genes, some of which are tandemly arrayed (Bouvier et al., 1989; Webb et al., 1991). Expression of different copies of the gene is regulated during the development of the parasite and different isoforms of the protein are found in the promastigote stage in the gut of the sand fly and in the amastigote stage in the phagolysosomes of the macrophages (Frommel et al., 1990; Roberts et al., 1995; Ramamoorthy et al., 1995). The major surface protease is thought to play an important role in the virulence of Leishmania by involvement in the degredation of components of the extracellular matrix and by facilitating promastigote attachment to host macrophages (McMaster et al., 1994). Immunisation with MSP protein confers partial protection of mice against Leishmania infection (Abdelhak et al., 1995). The proteins encoded by the P.carinii PRT1 gene family show highest homology to the subtilisin-like serine proteases. A wide diversity of different types of precursor proteins are processed by this family of proteases to mature and active regulatory proteins, but the precise function of many of these proteases has not yet been determined. Some of the fungal homologues have been shown to function in the processing of several proteins, such as the S. cerevisiae KEX2 gene product which processes both the pheromone α-factor and the killer toxin (Fuller et al., 1989). The krp gene product from S.pombe, which cleaves the pheromone precursor pro-P-factor to its active form, is thought to also function in the processing of other regulatory proteins, since its activity is essential for cell viability (Davey et al., 1994). The XPR6 gene product from Y. lipolytica, although not essential for cell viability, when disrupted was found to cause aberrant growth and morphology (Enderlin and Ogrydziak, 1994). The function of the products of the P.carinii PRT1 gene family is not yet understood but it is likely to play an important role in the life cycle and possibly also the pathogenicity of the organism.

Identification and sequencing of a PRT1 gene from P.carinii sp. f hominis

PCR strategies using degenerate primers designed using P.carinii sp. f. carinii PRT1 sequence information failed to isolate any P.carinii sp. f. hominis PRT1 clones. The strategies employed included single round PCR and nested PCR, on post mortem samples from infected patients.

Given the failure of these approaches, it was decided to try to obtain additional sequence data from P. carinii derived from other organisms. MATERIALS AND METHODS Samples

Samples of Pneumocystis carinii sp. f. hominis were derived from HIV positive patients by fibreoptic bronchoscopy, an aliquot of this bronchoscopic alveolar lavage (BAL) sample being immediately frozen, stored at -20°C and transported to the Institute of Molecular Medicine for DNA extraction (samples D503B and D122B). One sample (C180) was derived from a post mortem lung from an HIV-negative patient; the parasites were first enriched by successive filtration through 70 μm, 12 μm and 8μm filters.

Samples of Pneumocystis from the infected lungs of four other mammalian hosts were used. These were Pneumocystis carinii sp. f. muris (mouse derived), Pneumocystis carinii sp. f. mustelae (ferret derived), Pneumocystis carinii sp. f. suis (pig derived), Pneumocystis carinii sp. f. carinii (rat-derived) and Pneumocystis carinii sp. f. rattus (rat derived). These were enriched for parasites prior to DNA extraction. DNA Extraction

DNA was extracted from an enriched parasite preparation by proteinase K digestion, followed by phenol-chloroform extraction. The DNA was purified and concentrated using a DNA binding resin (Promega Wizard DNA Clean-UP System). DNA Amplification

In general the following conditions were used in all PCR reactions. The final concentration of the reaction mix was 50mM KCI, 10mM Tris (pH 8.0), 0.1 % Triton X-100, 3mM MgCI₂, 400μM of each deoxynucleoside triphosphate, 1μM of each oligonucleotide primer and 0.025U of Taq polymerase (Promega) per ml. A total of forty cycles was used with 10 cycles at 94°C for 1.5 min (denaturation), annealing at a temperature between 48°C and 55°C dependant on primer Tm and required stringency of reaction for 1.5min and 72°C for 2min (extension), followed by 30 cycles at 94°C for 1.5min, 63°C for 1.5min and 72°C for 2min (the increased temperature at annealing now including the EcoRI site at the 5' end of the primers). Where there was no EcoRI site in the primer or where particularly low stringency was required all 40 cycles were carried out at the lower annealing temperature. A positive control of rat Pneumocyctis DNA (rat 1458 or rat 1189) was included in each PCR reaction. Negative controls of no added template DNA were included after each sample to monitor for cross contamination. In later PCR reactions, when degenerate primers were being used, a negative control of human DNA (Sigma), at a final concentration of 0.8ng/μl, was included to monitor for non-specific amplification of human DNA, which was unavoidably co- extracted with all human Pneumocystis DNA samples. The primers used are shown in Table 1 herein (and Table 1 of Lugli er a/ 1997)..

All PCR products were electrophoretically separated out on 1.2% or 1.5% agarose gels containing ethidium bromide, visualised under ultraviolet light.

Determination of the complete sequence of a copy of P.carinii sp. f. hominis PRT1 gene A number of different approaches are available for the isolation of the complete gene sequence of a P.carinii sp. f. hominis PRT1 gene. Some of the possible approaches are described below in detail.

DNA and RNA is prepared from P.carinii sp. f. hominis organisms, obtained from either bronchoalveolar lavage samples from P.carinii infected patients or from post-mortem lung samples, i) P.carinii sp. f. hominis genomic library

A P.carinii sp. f. carinii genomic library is constructed in λFIX and this is screened with the cloned fragment of PRT1. Positive recombinant phage are analysed by further rounds of screening, and full length clones selected for analysis. The arrangement of introns within the gene sequence is determined. The genomic organisation of copies of PRT1 is elucidated, and in particular the relationship with gene copies of MSG. The chromosomal organisation of different PRT7 copies is examined, including the analysis of copies which are in the subtelomeric regions and others which are at an internal location, i) Expressed copies of PPT7

Two different approaches can be used to examine transcribed copies of PRT7. In the first, Random

Amplification of cDNA Ends (RACE) is used to extend 5'- and 3'- of the cloned fragment of PPT7, using total RNA or poly A⁺ RNA from the enriched parasite preparation. Primers are designed to the sequence of the cloned fragment for use in this technique. The second approach is the construction of a cDNA library in λZAP from P.carinii sp. f. hominis, which is then screened with the cloned fragment. Different recombinant clones are compared for variation in sequence and used for expression studies. Expression i) Expression of cloned fragment of P.carinii sp. f. hominis

PR7^"7 (H13)

The known portion of the catalytic domain is subcloned into the pET32a expression vector and expressed in an E. coli expression system. Recombinant protein is purified and used to raise polyclonal antiserum in rabbits. In addition, synthetic peptides designed to the PRT1 derived amino acid sequence are used in the production of antibodies. ii) Expression of the complete gene sequence and fragments of the gene spanning different domains. Recombinant protein is expressed and purified from different domains and from the complete sequence, for use in the production of antibodies, and in biochemical and immunohistochemical studies. Biochemical studies

Biochemical studies are performed to determine the substrate specificity of the protease and the optimum conditions (e.g. pH, metal cofactors) for proteolytic activity. This provides an in vitro system for the testing of inhibitors to the PR77 protease. Crystallisation of the recombinant protein is carried out and the 3-D structure of the protein determined by X-ray crystallography and compared with the 3D structure of the four other subtilisin-like serine proteases whose structure has previously been determined. These structural data can used for purposes including the design of specific inhibitors of PRT7, and the prediction of antigenically important epitopes.

Immunohistochemistry

Antibodies raised to the recombinant PRT7 protein or to synthetic peptides can be used in the analysis of the subcellular localisation of PRT7 in P.carinii organisms, using both light microscopy and electron microscopy with immunogold.

Table 1

Oligonucleotide primers

Primer Sequence

Pcprot1d/R1 GGGAATTCTA^T _C ^T _A ^C _GNTGV_A ^C _GNTGGGGNCC Pcprot16d/RI GGGAATTCCA^c _τGgiACi^c _AGiTG^τ _cGCiGG Pcprot17d/RI GGGAATTCA^c _τ ^G _ATci^Tc^G _τCCAiGTiA^G _A ^G _AT^τ _ciGG Pcprot18d/RI GGGAATTCTAiGC^G _ATciAi^τ _cTTiCC^A _G ^A _TAiCC Pcprot24d/RI GGGAATTC^G _ACC^A _CGAATA^T _CGTAGAAGC Pcprot25d/RI GGGAATTCGTTTT^T _CGG^G _A ^A _T ^C _{G C}GAGG^A _TGG Pcprot26d/RI GGGAATTC^A _TGCAA^T _GAGGT^A _G ^T _C ^A _GGAAGCAGA Pcprot31/RI GGGAATTCGAAGATGTTGATATTGAGGAG Pcprot32/RI GGGAATTCATCGTCTCTTATCGCACCC Pcprot33/RI GGGAATTCTCAACTCAACTAATACC Pcprot39/RI GGGAATTCAGGAATGA I I I I I GTGGGCT 73jEx4/RI GGGAATTCTTATGGAACAGCTGTTTCC 73jEx5/RI GGGAATTCATCAATAGACTCTCCG PcprotH34/RI GGGAATTCTTGCGAATATTATCCGGGC PcprogH35/RI GGGAATTCGCACTTCCACCTGCATATG

Oligonucleotide Sequences. Note that I = inosine and N = any base in degenerate sequences.

The oligonucleotides above have SEQ ID NOS: 1-15, according to the order in which they appear in the above table.

Single round PCR on Rat Variant, Mouse, Ferret and Pig derived P.carinii Single round PCR on P.carinii sp. f. rattus and P.carinii sp.f. muήs samples gave strong amplification products at the same Mr as the rat P.carinii positive control. Primers used were Pcprot1/R1 and Pcprot3/R1. Sequence data is shown in Figure 2.

Single Round PCR on Human Post Mortem Sample using Redesigned Primer

New primers were designed based on regions of homology of the newly obtained rat variant P. carinii and mouse P. carinii PRT1 sequences with the rat prototype P. carinii sequence at both the DNA level and amino acid level. These were not fully degenerate, given that Pneumocystis DNA shows a high AT bias (60-70%); unless the sequence data suggested otherwise only A or T was used at potentially degenerate sites (as seen in the amino acid sequences). These new primers were used in reactions with one another and previously used primers. Of these reactions, only Pcprot16d/R1 and Pcprot26d/R1 gave a clear positive product at the expected Mr, close to that of the rat P. carinii positive control (-600 b.p.). The primers used were Pcprot25d/R1 + Pcprot26d/R1 ; Pcprot1d/R1 + Pcprot26d/R1 ; Pcprot16d/R1 + Pcprot26d/R1 ; Pcprot25d/R1 + Pcprotl 7d/R1 ; Pcprot25d/R1 + Pcprotl 8d/R1 ;

Pcprot25d/R1 + Pcprot24d/R1. The PCR products from the reactions were cloned and sequenced. Of the clones sequenced one contained an insert which showed homology to the PRT7 gene. Sequence data over the catalytic domain is shown in Figures 2 and 3.

Table showing percentage divergence of prototype rat-derived Pneumocystis (P.carinii sp. f. carinii). mt LSU rRNA - mitochondrial large subunit rRNA; mt SSU rRNA - mitochondrial small subunit rRNA. Values for Variant rat P. carinii from two clones; values for Mouse P. cahnii from three clones. DNA divergence calculated with Jukes-Cantor correction method. Protein divergence calculated using Kimura protein distance.

The above table shows that the PRT7 gene differs between P.carinii from different host organisms by far more than many other genes so far studied. Thus in P.carinii sp. f. hominis the PRT7 DNA sequence is around twice as divergent from P.carinii sp. f. carinii compared to other sequences and the amino acid sequence is over three times as divergent as the arom sequence. This is even more striking given that the PRT7 data are taken from the catalytic domain which should contain the highest level of conservation (catalytic, substrate binding, oxyanion hole and disulphide bridge residues). A similar level of divergence has previously been observed in the MSG (also called Glycoprotein A; gpA) genes. Indeed, early attempts to amplify some portions of gpA/MSG from P.carinii sp. f. hominis by PCR using primers based on the P.carinii sp. f. carinii sequence failed (Kovacs et al., 1993; Wright et al., 1994).

A high level of divergence is also seen in the PRT7 sequences from P.carinii sp. f. rattus and P.carinii sp. f. muris where the PRT7 DNA sequences are two to four times as divergent as the other sequences and the mouse P. carinii PRT1 amino acid sequence is over six times more divergent than that of arom.

The homology of the amino acid sequences from all three types of Pneumocystis to the subtilisin-like serine proteases is high. Of the known conserved residues, most can be seen to be conserved in the PRT7 sequences (where the data are available). Certainly in the P.carinii sp. f. hominis PRT1 amino acid sequence there is greater conservation of the negatively charged amino acids at the substrate-binding face than is seen in the P.carinii sp. f. carinii sequence. Although the homology to the subtilases is unmistakable, there is considerable variation to be seen between the PRT7 sequences. This presumably reflects differences in substrate specificity, whether the substrate is a host protein (or proteins) or a parasite protein (e.g. gpA MSG). The function of the subtilisin-like serine proteases so far studied is in the specific endoproteolytic processing of precursor proteins to their active form. Although the precise function of many subtilases is yet to be determined, some fungal homologues have been shown to be vital to cell viability or normal function. Thus krp in S. pombe has been shown to be vital to cell viability and disruption of XPR6 in Y. lipolytica causes aberrant growth and morphology. Parallels may also be drawn between Gp63 in Leishmania and PRT7 in Pneumocystis, as discussed in the introduction. The functions of the PRT1 proteins are not yet fully established, but it seems likely to be important to the life-cycle and/or the pathogenesis of the organism. The cloning of this gene, most especially from P.carinii sp.f. hominis, is thus a step towards the design of an effective anti-Pneumocysf/s drug. Generation of anti-PRT1 antibodies

Polyclonal antiserum was generated in rabbits to synthetic peptides, designed to the Pneumocystis carinii sp. f. carinii PRT1 sequence. Regions of the protein which were likely to be immunogenic were predicted using the appropriate software, and peptides (15 mers) to six different regions were synthesized. A mixture of six synthetic peptides was administered by subcutaneous injection to rabbits (New Zealand white). An antibody response was elicited by standard procedures, using Freunds complete adjuvant for the first injection and Freunds incomplete adjuvant for subsequent injections.

The resulting polyclonal antisera were tested against the peptides. The greatest cross-reactivity of the antisera was found with Peptide 7, designed to a region of the catalytic domain (amino acid residues 424 - 438 of the PRT1 (73j) sequence) and with Peptide 9, designed to the pro-domain (amino acid residues 64 - 78 of the PRT1(73j) sequence).

Peptide sequences

T RDVQALIVETAVP (2) (SEQ ID NO: 16)

ITSPSGVTSVLAHRR (4) (SEQ ID NO: 17)

ESEGVPPPSYPF SR (5) (SEQ ID NO: 18)

ASTPLAAGVIALLLS (7) (SEQ ID NO : 19) FRGESIVGN TIDVE (8) (SEQ ID NO : 20)

DNQHIFSIEKGVLED (9) (SEQ ID NO: 21)

EXAMPLES

Example 1

Expression of portions of the rat-derived P. carinii (P. carinii sp. f. carinii) PRT1(73j) gene.

The E. coli expression vector pET32a (Novagen, Madison, WI) was used. This vector contains an inducible T7lac promotor, a 6-His tag, a multiple cloning site and the recombinant protein is expressed as fusion protein with the Trx-tag thioredoxin protein (109 amino acids). Recombinant thioredoxin fusion proteins are generally more soluble and remain in the E. coli cytoplasmic fraction. Three different regions of the PRT1(73j) gene were cloned into pET32a: i) Cat2f1 , a portion of the catalytic domain, 585bp in length, from base 790 to base 1375; ii) F1 a1j, a portion of the pro-domain, 255bp in length, from base 120 to base 375; iii) G1 b1 c, a portion of the P domain, 384 bp in length, from base 1515 to base 1899.

The specific fragments were amplified by PCR from the PRT1 (73j) sequence as follows - i) Cat2f1 using primers Pcprot39/R1 and 73j Ex4; ii) F1a1j using primers Pcprot31/RI and Pcprot32/RI; iii) G1 b1c using primers Pcprot33/RI and 73jEx5/RI (see Table 1). All primers included an EcoRI site the 5' end to facilitate cloning. The fragments were initially cloned into the plasmid vector pUC, linearized with EcoRI and treated with alkaline phosphatase, to produce a stable, high copy number, recombinant plasmid. The recombinant DNA was then subcloned into the EcoRI site of the expression vector pET32a.

2. Transformation of E. co./^' with recombinant plasmids

E. coli DH5α competent cells were transformed with the recombinant plasmids. The cells were transformed with recombinant pUC plasmids, and also recombinant pET32a plasmids. The recombinant expression vector pET32a constructs were also transferred into E. coli DE3 (BL21) cells, for expression of the recombinant peptides.

3. Expression of recombinant PRT1 polypeptides

The recombinant pET32a constructs, transformed into E. coli DE3(BL21) were induced with IPTG, and the bacteria were grown for 3 to 4 hours. The cells were collected by centrifugation and disrupted by sonication. The bacterial proteins were separated by SDS-PAGE and electrophoretically transferred to nitrocellulose filter. The immobilised proteins were cross-reacted with anti-thioredoxin antibody (Sigma), and the bound antibody was visualised with a swine anti-rabbit immunoglobulins secondary antibody, conjugated to alkaline phosphatase. A band of the expected size (24kDa) was seen in the control vector pET32a, (lane 1) corresponding to the thioredoxin fusion protein and the His-tag. Bands corresponding to the expected sizes of the recombinant PRT1 protein fragments were observed (Figure 7, lanes 2 and 3).

4. Preparation of polyclonal mono-specific antibodies Polyclonal antisera raised against the six synthetic peptides were affinity purified. The peptide (Peptide 7 or Peptide 9) was covalently linked to an amine reactive support. Immunoglobulins which cross-reacted to the peptide were specifically retained by the column, and subsequently eluted. In this way, two polyclonal mono-specific antibodies were produced, anti-Peptide 7 and anti-Peptide 9.

5. Cross-reactivity of polyclonal, mono-specific antibodies with recombinant PRT1 polypeptides

Expressed proteins from transformation of E. coli DE3(BL21) with recombinant expression vector to the pro-domain (F1a1j) or to the catalytic domain (Cat2f1) were separated by SDS-PAGE and electrophoretically transferred to nitrocellulose membrane. The anti- Peptide 7 mono-specific antibody was shown to cross-react with the recombinant Cat2f1 polypeptide, but not to F1a1j or to the protein produced by the control plasmid pET32a. Likewise, the anti-Peptide 9 antibody specifically cross-reacted with the F1 a1j polypeptide. These results confirm the specificity of the mono-specific antisera to the two distinct domains of the PRT1 protein. 6. Identification of PRT1 protein in P.carinii sp. f. carinii organisms

P.carinii sp. f. carinii organisms were extracted and enriched from infected rat lungs. Organisms were disrupted by heating to 95°C in denaturing solution and the proteins separated by SDS-PAGE, followed by transfer to nitocellulose filters. The immolbilised proteins were cross- reacted with the anti-Peptide 7 and the anti-Peptide 9 antibody. Bound antibody was detected using an anti-rabbit secondary antibody, conjugated to alkaline phosphatase. A single, major band, at 40 kDa, was seen with each of the mono-specific antibodies. In addition, another major band at 38 kDa was seen with anti-Peptide 7 antibody and minor bands at 98 kDa and 16 kDa. With the anti-Peptide 9 antibody, minor bands at 200kDa, 98kDa and 43 kDa were observed. The predicted size of the full length PRT1 protein ranges from 87 to 102 kDa. The proteins detected with the mono-specific antibodies are assumed to be the products of autocatalysis at a number of dibasic residues found in the PRT1 sequence.

7. Sub-cellular localisation of the PRT1 protein in P.carinii sp. f. carinii organisms

Sections of P.carinii sp. f. carinii infected rat lungs, formalin fixed and embedded in paraffin, were prepared and incubated with anti- Peptide 7 antibody. Bound antibody was detected using a swine anti- rabbit immunoglobulin secondary antibody, conjugated to horse radish peroxidase, and the organisms viewed by light microscopy. The specific distribution of the antibody on the P.carinii sp. f. caπn/V organisms was characteristic of surface localisation of the PRT1 protein in the organisms.

Example 2

Expression of a portion of the human-derived P. carinii (P. carinii sp. f. hominis) PRT1 gene 1. Construction of recombinant vector containing a portion of the P.carinii sp. f. hominis PRT1 gene

The E.coli expression vector pET32a (Novagen, Madison, WI) was used. This vector contains an inducible T7lac promotor, a 6-His tag, a multiple cloning site and recombinant protein is expressed as fusion protein with the Trx-tag thioredoxin protein (109 amino acids). Thioredoxin fusion proteins are generally more soluble and remain in the E.coli cytoplasmic fraction. A 367bp portion of the cloned P. carinii sp. f. hominis

PRT1(H13) sequence was amplified using PCR with the primers PcprotH34/RI and PcprotH35/RI, corresponding to position 111 to position 478 on the PRT1 (H13) sequence, in the catalytic domain of the gene (see Table 1). The primers included an EcoRI site at the 5' end to facilitate cloning. The resulting fragment (H1a1a) was initially cloned into the EcoRI site of the plasmid vector pUC, and then subcloned into the EcoRI site of the expression vector pET32a.

2. Transformation of E. coli with recombinant plasmids E. coli DH5α competent cells were transformed with the recombinant plasmid. The cells were transformed with the recombinant pUC plasmid, and also the recombinant pET32a plasmid. The recombinant expression vector pET32a construct was also transferred into E. coli DE3 (BL21) cells, for expression of the recombinant peptide.

3. Expression of recombinant P.carinii sp. f. hominis PRT1 peptide

The recombinant pET32a construct (H1a1a), transformed into E. coli DE3(BL21) was induced with IPTG, and the bacteria were grown for 3 to 4 hours. The cells were collected by centrifugation and disrupted by sonication. The bacterial proteins were separated by SDS-PAGE and electrophoretically transferred to nitrocellulose filter. The immobilised proteins were cross-reacted with anti-thioredoxin antibody (Sigma), and the bound antibody was visualised with a swine anti-rabbit immunoglobulins secondary antibody, conjugated to alkaline phosphatase. A band of the expected size (24kDa) was seen in the vector pET32a control, (lane 1) corresponding to the thioredoxin fusion protein and the His-tag. A band corresponding to the expected size of the recombinant P.carinii sp. f. hominis PRT1 protein fragment was observed (Figure 7, lane 4).

4. Identification of PRT1 protein in P.carinii sp. f. hominis organisms

P.carinii sp. f. hominis organisms were extracted from bronchoalveolar lavage fluid from a patient with P. carinii pneumonia. The organisms were disrupted by heating to 95°C in denaturing solution and the proteins separated by SDS-PAGE, followed by transfer to nitrocellulose filters. The immobilised proteins were cross-reacted with the anti-Peptide 7 and the anti-Peptide 9 antibody. Bound antibody was detected using an anti-rabbit secondary antibody, conjugated to alkaline phosphatase. Two major bands, at 56 kDa and 49 kDa was seen with each of the mono- specific antibodies. In addition, minor bands at 116kDa, 95kDa, 86 kDa and 39 kDa were seen with the anti-Peptide 7 antibody, and at 200 kDa, 116kDa, 95kDa, 86 kDa and 29 kDa with the anti-Peptide 9 antibody. The proteins detected with the mono-specific antibodies are assumed to be the products of autocatalysis at a number of dibasic residues found in the P.carinii sp. f. hominis PRT1 sequence. REFERENCES

Abdelhak, S., Louzir, H., Timm, J., Blel, L., Banlasfar, Z.,

Lagranderie, M., Gheorghiu, M., Dellagi, K. & Gicquel, B. (1995). Recombinant BCG expressing the leishmania surface antigen Gp63 induces protective immunity against Leishmania major infection in

BALB/c mice. Microbiology 141 , 1585-1592.

Banerji, S., Wakefield, A.E., Allen, A.G., Maskell, D.J., Peters, S.E. and Hopkin, J.M. (1993). The cloning and characterization of the arom gene of Pneumocystis carinii. J Gen Microbiol 139, 2901-

2914.

Betzel, C, Pal G.P. and Saenger W. (1988). Three-dimensional structure of proteinase K at 0.15nm resolution. Eur J Biochem 178, 155-171.

Bott, R., Ultsch, M., Kossiakoff, A., Graycar, T., Katz, B. and Power, S. (1988). The three-dimensional structure of Bacillus amyloliquefaciens subtilisin at 1.8 A and an analysis of the structural consequences of peroxide inactivation. J Biol Chem 263, 7895-7906.

Bouvier, J., Bordier, C, Vogel, H., Reichelt, R. & Etges, R. (1989).

Characterization of the promastigote surface protease of Leishmania as a membrane-bound zinc endopeptidase. Mol & Biochem Paras

37, 235-246.

Cushion, M.T., Kaselis, M., Stringer, S.L and Stringer, J.R. (1993).

Genetic stability and diversity of Pneumocystis carinii infecting rat colonies.

Infect Immun 61 , 4801-4813. Davey, J., Davis, K., Imai, Y., Yamamoto, M., and Matthews, G. (1994).

Isolation and characterization of krp, a dibasic endopeptidase required for cell viability in the fission yeast Schizosaccharomyces pombe. EMBO

Journal 13, 5910-5921. Dyer, M., Volpe, F., Delves, C.J., Somia, N., Burns, S. and Scaife, J.G.

(1992). Cloning and sequence of a β-tubulin cDNA from Pneumocystis carinii: possible implications for drug therapy. Mol Microbiol 6, 991-1001.

Enderlin, C. S., and Ogrydziak, M. (1994). Cloning, nucleotide sequence and functions of XPR6, which codes for a dibasic processing endoprotease from the yeast Yarrowia lipolytica. Yeast

10, 67-79.

Frommel, T. O., Button, L. L., Fujikura, Y. & McMaster, W. R.

(1990). The major surface glycoprotein (GP63) is present in both life stages of Leishmania. Mol & Biochem Paras 38, 25-32.

Fuller, R. S., Brake, A., and Thorner, J. (1989). Yeast prohormone processing enzyme (KEX2 gene product) is a Ca²⁺-dependent serine protease. Proc. Natl. Acad. Sci. USA 86, 1434-1438.

Garbe, T. R. and Stringer, J. R. (1994). Molecular characterization of clustered variants of genes endocing major surface antigens of human Pneumocystis carinii. Infect Immun 62, 3092-3101.

Gigliotti F. (1992). Host species-specific antigenic variation of a mannosylated surface glycoprotein of Pneumocystis carinii. J Infect

Dis 165, 329-336. Gros, P., Betzel, C, Dauter, Z., Wilson, K. S and Hoi, W. G. J. (1989).

Molecular dynamics refinement of a thermitase-eglin-c-complex at 1.98 A resolution and comparison of two crystal forms that differ in calcium content. J Mol Biol 210, 347-367.

Haidaris, P. J., Wright, T. W., Gigliotti, F. and Haidaris, C. G. (1992). Expression and characterization of a cDNA clone encoding an immunodominant surfact glycoprotein of Pneumocystis carinii. J

Infect. Dis 166, 1113-1123.

Hirono, S., Akagawa, H., litaka, Y. and Mitsui, Y. (1984). Crystal structure at 2.6 A resolution of the complex of subtilisin BPN with Streptomyces subtilisin inhibitor. J Mol Biol 178, 389-414. Klemm, U., Mϋller-Esterl, W. and Engel, W. (1991). Acrosin, the peculiar sperm-specific serine protease. Human Genetics 87, 635-641. Kovacs, J. A., Powell, F., Edman, J. C, Lundgren, B., Martinez, A., Drew, B. and Angus, C. W. (1993). Multiple genes encode the major surface glycoprotein of Pneumocystis carinii. J Biol Chem 268, 6034-6040.

Lugli, E.B. and Wakefield, A.E. (1996). A novel subtelomeric multi- gene family in Pneumocystis carinii . 4th International Workshop on Opportunistic Protists, Tuscon, Arizona, USA, June 1996. Lugli, E.B..Allen, A.G. and Wakefield, A.E. (1997) A Pneumocystis carinii multi-gene family with homology to subtilisin-like serine proteases. Microbiology 143: 2223-2236.

Lundgren, B., Lipschik, G.Y. and Kovacs, J.A. (1991). Purification and characterization of a major human Pneumocystis carinii surface antigen. J Clin Invest 87, 163-170.

McMaster, R. R., Morrison, C. J., MacDonald, M.H. & Joshi, P. B.

(1994). Mutational and functional analysis of the Leishmania surface metalloproteinase GP63 : similarities to matrix metalloproteinases. Parasitology 108, S29-S36. McPhalen, CA. and James, M. N. G. (1988). Structural comparison of two serine proteinase-protein inhibitor complexes: eglin-c-subtilisin Carlsberg and CI-2-subtilisin Novo. Biochem 27, 6582-6598.

Peters, S.E., Wakefield, A.E., Banerji, S. and Hopkin, J.M. (1992).

Quantification of the detection of Pneumocystis carinii by DNA amplification. Molec Cell Probes 6, 115-117.

Ramamoorthy, R., Swihart, K. G., McCoy, J. J., Wilson, M. E. &

Donelson, J. E. (1995). Intergenic regions between Tandem gp63 genes influence the differential expression of gp63 RNAs in

Leishmania chagasi promastigotes. J Biol Chem 270(20), 12133- 12139. Roberts, S. C, Wilson, M. E. & Donelson, J. E. (1995).

Developmentally regulated expression of a novel 59-kDa product of the major surface protease (Msp or gp63) gene family of Leishmania

Chagasi. J Biol Chem 270(15), 8884-8892. Roditi, I., Schwarz, H., Pearson, T.W., Beecroft, R.P., Liu, M.K.,

Richardson, J.T., Buhring, H.J., Pleiss, J., Bulow, R., Williams, R.O. and

Overath, P. (1989). Procyclin gene expression and loss of the variant surface glycoprotein during differentiation of Trypanosoma brucei. J cell

Biol 108, 737-746. Roebroek, A. J. M., Creemers, J. W. M., Pauli, I. G. L, Kurzik-Dumke, U.,

Rentrop, M., Gateff, E. A. F., Leunissen, J. A. M. and Van de Ven, W. J. M.

(1992). Cloning and functional expression of Dfurin2, a subtilisin-like proprotein processing enzyme of Drosophila melanogaster with multiple repeats of a cysteine motif. J Biol Chem 267, 17208-17215. Rogers, J. (1985). Exon shuffling and intron insertion in serine protease genes. Nature 315, 458-459.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular cloning: a laboratory manual. Cold Spring Habour, New York: Cold Spring Habour

Laboratory Press. Siezen, R. J., de Vos, W. M., Leunissen, J. A. M. and Dijkstra, B. W.

(1991). Homology modelling and protein engineering strategy of subtilases, the family of subtilisin-like serine proteinases. Protein

Eng 4, 719-737.

Stringer, S. L., Garbe, T., Sunkin, S. M. and Stringer, J. R. (1993). Genes encoding antigenic surfact glycoproteins in Pneumocystis from humans. J Euk Microbiol 40, 821-826.

Sunkin, S. M., and Stringer, J. R. (1996). Translocation of surface antigen genes to a unique telomeric expression site in Pneumocystis carinii. Mol Microbiol 19, 283-295. Sunkin, S. M., Stringer, S. L. and Stringer, J. R. (1994). A tandem repeat of rat-derived Pneumocystis carinii genes encoding the major surface glycoprotein. J Euk Microbiol 41 , 292-300. Tanguy-Rougeau, C, Wesolowski-Louvel, M. and Fukuhara, H. (1988). The Kluyveromyces lactis KEX1 gene encodes a subtilisin- type serine proteinase. FEBS Lett. 234, 464-470.

Teplyakov, A. V., Kuranova, I. P., Harutyunyan, E. H. and Vainshtein, B. K. (1990). Crystal structure of thermitase at 1.4 A resolution. J Mol Biol 214, 261-279. Underwood, A. P., Louis, E.J., Borts, R. H., Stringer, J. R. and Wakefield, A. E. (1996). Pneumocystis carinii telomere repeats are composed of TTAGGG and the subtelomeric sequence contains a gene encoding the major surface glycoprotein. Mol. Microbiol 19, 273-281. van den Ouweland, A.M.W., van Duijnhove, H.L.P., Keizer, G.D., Dorssers, L.C.J. and Van de Ven, W.J.M. (1990). Structural homology between the human fur gene product and the subtilisin- like protease encoded by yeast KEX2. Nucl Acids Res 18, 664. Van de Ven, W. J. M. and Roebroek, A. J. M. (1993). Structure and function of eukaryotic proprotein processing enzymes of the subtilisin family of serine proteases. Critical Rev in Oncogenesis 4, 115-136. Volpe, F., Dyer, M., Scaife, J. G., Derby, G., Stammers, D.K. and Delves, C. J. (1992). The multifolate folic acid synthesis fas gene of Pneumocystis carinii appears to encode dihydropteroate synthase and hydroxymethyldihydropterin pyrophosphokinase. Gene 112, 213-218. Volpe, F., Ballantine, S. P., and Delves, C. J. (1993). The multifunctional folic acid synthesis fas gene of Pneumocystis carinii encodes dihydroneopterin aldolase, hydroxymethyldihydropterin pyrophosphokinase and dihydropteroate synthase. Eur J Biochem 216, 449-458. Wada, M., and Nakamura, Y. (1994). MSG gene cluster encoding major cell surface glycoproteins of rat Pneumocystis carinii. DNA Research 1 , 163-168.

Wada, M., Kitada, K., Saito, M., Egawa, K. and Nakamura, Y. (1993). cDNA sequence diversity and genomic clusters of major surface glycoprotein genes of Pneumocystis carinii. J Infect Dis 168, 979-985.

Wada, M., Sunkin, S.M., Stringer, J.R. and Nakamura, Y. (1995). Antigenic variation by positional control of major surface glycoprotein gene expression in Pneumocystis carinii. J Infect Dis

171 , 1563-1568.

Webb, J. R., Button, L. L. & McMaster, W. R. (1991). Heterogeneity of the genes encoding the major surface glycoprotein of Leishmania donovani. Mol & Biochem Paras 48, 173-184. Wright, T. W., Simpson-Haidaris, P. J., Gigliotti, F., Harmsen, A. G.

& Haidaris, C. G. (1994). Conserved sequence homology of cysteine-rich regions in genes encoding glycoprotein A in

Pneumocystis carinii derived from different host species. Inf &

Immun 62, 1513-1519. Wright, T. W., Bissoondial, T. Y., Hairdaris, C. G., Gigliotti, F. &

Simpson Haidaris, P. J. (1995). Isoform diversity and tandem duplication of the glycoprotein A gene in ferret Pneumocystis carinii.

DNA Research 2, 77-88.

Zhang, J. and Stringer, J.R. (1993). Cloning and characterization of an alpha-tubulin-encoding gene from rat-derived Pneumocystis carinii. Gene

123, 137-141. Figure Legends

Figure 2

Nucleotide sequence alignments of part of the catalytic domain of PRT7. 1-3 page, 11-3-73J andd 1-3prp5e from P. cahnii f.sp. carinii ^<8>; ratvδprti and ratv16prt1 from P. carinii f. sp. rattus; mouseel prt1 , mouse7prt1 and mouse13prt1 from P. carinii f. sp. muris; humanprtl from P. carinii f. sp.

Figure 3 Amino acid sequence alignments of part of the catalytic domain of PRT1, translated from the nucleotide sequences (Figure 2).

Pagaprtl , 73jpart1 and prpδeprti from P. carinii f. sp. P. carinii; ratvδprti and ratv16prt1 from P. carinii f. sp. rattus; mouse1 prt1 , mouse7prt1 and mouse 13part1 from P. carinii f. sp. muris; humanprtl from P. carinii f. sp. hominis. U marks conserved amino acids; numbering according to full amino acid sequence of cDNA clone 73j⁽⁸⁾; an asterisk marks positions of charge conservation in subtilases (see text).

Figure 4 Alignment of the P.carinii sp. f. carinii PRT1 deduced amino acid sequences from the genomic clone Paga, the cDNA clone 73j and the three overlapping PCR products amplified from a cDNA library corresponding to the 5' region (Prpδe), the central region (M14), and the 3' region (Prp2g). The deduced amino acid sequences of PCR products amplified from five different regions of the PRT7 gene family were also aligned; the catalytic domain: Prpla, Prp3a, Prp7a; the boundary of the catalytic domain and the P-domain: Prp2c, Prp3c, Prp4c; the P-domain: Prptaf2, Prpf4, Prpδf; the proline-rich region: Pcr-19, Pcr-14, Pcr-5, Pcr-3, Pcr-1 , Lam-1 ; the C-terminal region: Prpg4, Prpg3, Prpδg. Gaps were introduced to maximize homology; identical amino acids are boxed. Figure 6

Schematic representation of the P. carinii sp. f. carinii PRT1.

Patterned boxes represent different domains; small dots represent hydrophobic regions (HR), diagonal lines indicate the catalytic domain

(CAT), woven pattern indicates the P-domain (P), vertical lines indicate the proline-rich region, squares indicate the serine-threonine rich region (STR).

Boxes that are defined by a shaded line (PR and STR) indicate length and sequence variation in these regions. Diamonds indicate potential glycosylation sites; (t) catalytic active site residues D214, H252- S423; (|) conserved cysteine residues. Residues were numbered with reference to the PRT1 (73j) sequence.

Figure 7 Recombinant PRT1 polypeptides, expressed in E. coli as thioredoxin fusion proteins, separated by SDS-PAGE and cross-reacted with an anti-thioredoxin antibody. E. coli DE3(BL21) transformed with: lane 1 : control plasmid pET32a; lane 2: F1a1a (portion of pro-domain of P.carinii sp. f. carinii PRT1 gene); lane 3: G1 b1c (portion of P-domain of P.carinii sp. f. carinii PRT1 gene); lane 4: H1a1a (portion of catalytic domain of P.carinii sp. f. hominis PRT1 gene).

Claims

1. An isolated DNA comprising part or all of a PRT7 gene of a non-rat infecting species of Pneumocystis carinii.

2. The DNA according to claim 1 , comprising part or all of a

PRT7 gene of a human-infecting species of Pneumocystis carinii.

3. The DNA according to claim 1 or claim 2, wherein the PRT7 gene is in the form of cDNA.

4. An isolated DNA comprising a sequence shown in figure 1 , or a non-rat sequence shown in figure 2, or a sequence which hybridises to either of these under stringent conditions.

5. The DNA according to claim 1 or claim 4, wherein the PR77 gene has been mutated by point mutation, deletion, insertion, or other means.

6. A recombinant vector containing the DNA according to any one of claims 1 to 5.

7. A recombinant polypeptide which is part or all of a PRT7 gene product, expressed by a vector according to claim 6.

8. Synthetic peptides corresponding to antigenic portions of a PRT1 gene product.

9. A synthetic peptide chosen from:

T RDVQALIVETAVP (SEQ ID NO: 16)

ITSPSGVTSVLAHRR (SEQ ID NO: 17)

ESEGVPPPSYPFLSR (SEQ ID NO: 18) ASTPLAAGVIALLLS (SEQ ID NO: 19)

FRGESIVGNWTIDVE (SEQ ID NO: 20)

DNQHIFSIEKGVLED (SEQ ID NO: 21)

10. A method of producing antibodies specifically immunoreactive with a Pneumocystis carinii protease, which method comprises using a polypeptide according to claim 7 or a synthetic peptide according to claim 8 or claim 9 to generate an immune response.

11. Antibodies produced by the method according to claim 10.

12. Antibodies according to claim 11 , which are monoclonal.

13. A method of screening for anti-Pr/et/mocysf/^'s carinii compounds, which method comprises providing a source of a recombinant polypeptide expressed by part or all of a PRT7 gene or cDNA, and contacting the compound with the recombinant polypeptide.

14. The method according to claim 13, wherein the recombinant polypeptide is expressed at the surface of a cell.

15. The method according to claim 13 or claim 14, for screening for protease inhibitors effective against Pneumocystis carinii.

16. The method according to any one of claims 13 to 15, using a recombinant polypeptide corresponding to part or all of the catalytic domain of the protease.

17. A cell transfected with a vector according to claim 6 and expressing a polypeptide according to claim 7.

18. An engineered cell line expressing a recombinant polypeptide from part or all of a PRT7 gene or cDNA, which may be mutated by point mutation, deletion, insertion or other means, useful in the method according to any one of claims 13 to 16.

19. The cell line according to claim 18, wherein the PR77 gene or cDNA is from a human-infecting Pneumocystis carinii species.

20. The method according to any one of claims 13 to 16, wherein the PRT7 gene or cDNA has been mutated by point mutation, deletion, insertion or other means.

21. A Pneumocystis carinii protease isolated using an antibody according to claim 11 or claim 12.

22. A PRT7 clone for part or all of the human-infecting Pneumocystis carinii PRT1 gene.