Cell wall assembly by Pneumocystis carinii. Evidence for a unique gsc-1 subunit mediating beta -1,3-glucan deposition.

Pneumocystis carinii remains a persistent cause of severe pneumonia in immune compromised patients. Recent studies indicate that P. carinii is a fungal species possessing a glucan-rich cyst wall. Pneumocandin antagonists of beta-1,3-glucan synthesis rapidly suppress infection in animal models of P. carinii pneumonia. We, therefore, sought to define the molecular mechanisms of beta-glucan cell wall assembly by P. carinii. Membrane extracts derived from freshly purified P. carinii incorporate uridine 5'-diphosphoglucose into insoluble carbohydrate, in a manner that was completely inhibited by the pneumocandin L733-560, an antagonist of Gsc-1-type beta-glucan synthetases. Using degenerative polymerase chain reaction and library screening, the P. carinii Gsc-1 catalytic subunit of beta-1,3-glucan synthetase was cloned and characterized. P. carinii gsc1 exhibited homology to phylogenetically related fungal beta-1,3-glucan synthetases, encoding a predicted 214-kDa integral membrane protein with 12 transmembrane domain structure. Immunoprecipitation of P. carinii extracts, with a synthetic peptide anti-Gsc-1 antibody, specifically yielded a protein of 219.4 kDa, which was also capable of incorporating 5'-diphosphoglucose into insoluble glucan carbohydrate. As opposed to other fungi, the expression of gsc-1 mRNA is uniquely regulated over P. carinii's life cycle, having minimal expression in trophic forms, but substantial expression in the thick-walled cystic form of the organism. These results indicate that P. carinii contains a unique catalytic subunit of beta-1,3-glucan synthetase utilized in cyst wall formation. Because synthesis of beta-1,3-glucan is absent in mammalian cells, inhibition of the P. carinii Gsc-1 represents an attractive molecular target for therapeutic exploitation.

Classification of Pneumocystis carinii as a fungus revolutionized study of this organism, which continues to cause lifethreatening pneumonia in immune-compromised patients with AIDS, malignancy, and organ transplantation (1)(2)(3)(4)(5). Studies of P. carinii in infected lung indicate its life cycle alternates between smaller trophic forms and thick-walled cysts (6 -8). The origin of P. carinii cysts remains controversial, but it has been postulated they arise from sexual conjugation, analogous to ascomycetous fungi (9,10). The mechanisms of cyst wall assembly by P. carinii are not well known, although recent studies reveal that these walls are largely composed of ␤-glucans, glycoprotein A, and chitins (11)(12)(13)(14)(15)(16).
Because mammalian cells do not possess glucan biosynthetic pathways, cell wall assembly represents an attractive target for the treatment of fungal infection. It is particularly noteworthy that pneumocandin inhibitors of ␤-glucan synthesis have been shown to rapidly inhibit P. carinii growth in rodent models (33)(34)(35)(36). Such studies provide evidence of the importance of ␤-glucan generation during life cycle progression of this organism. Despite the considerable importance of ␤-glucan assembly in life cycle expression of this organism, in immune recognition during infection, and as a potential therapeutic target for pneumonia, the mechanisms of ␤-1,3-glucan assembly by P. carinii are not yet fully understood.
The current investigation was undertaken to accomplish the following: 1) to establish whether P. carinii cell wall assembly occurs through action of a Gsc-1 protein mediating ␤-1,3-glucan synthesis; 2) to clone and characterize the respective gsc-1 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM (37). Nitrocellulose membranes containing P. carinii chromosomes separated by contour-clamped homogenous field electrophoresis were the gift of Dr. M. T. Cushion, University of Cincinnati (38).
P. carinii Cell Wall and Membrane Isolation and Assessment of Glucan Synthetase Activity-P. carinii pneumonia was induced in Harlan Sprague-Dawley rats by immune suppression with dexamethasone and transtracheal inoculation, as we previously reported (39,40). P. carinii were purified by homogenization and filtration through 10-m filters to remove lung cells (41,42). To determine whether P. carinii contained glucan synthetase activity, cell wall membrane fractions were isolated and assessed for incorporation of UDP-[ 14 C]glucose into carbohydrate (43). One gram of P. carinii was suspended in 50 mM Tris-HCl, 150 mM NaCl buffer with 1 mM EDTA and sonicated on ice. The homogenate was centrifuged (55,000 ϫ g for 45 min) to isolate membranes, and the pellet was washed once with 50 mM Tris-HCl buffer containing 1 mM EDTA and 1 mM mercaptoethanol (pH 7.5; Buffer A). The pellet was then resuspended in Buffer A containing 33% glycerol. A final assay mixture (40 l of total volume) was prepared with 5 mM UDP-[ 14 C]glucose (250,000 cpm/mol), 7.5 mM Tris-HCl (pH 7.5), bovine serum albumin (1 mg/ml), 25 mM KF, 1 mM EDTA, 20 M GTP␥S, and enzyme (20 -35 l of P. carinii membrane suspension). The reaction mixtures were incubated for 4 h at room temperature. Aliquots were spotted onto glass filter discs, washed twice with 20% trichloroacetic acid, washed twice with acetone, and counted. The background cpm (buffer alone) was subtracted from the data. Additional assays were performed in the presence of the pneumocandin B 0 analog L-733,560 (0 -10 M), an agent that inhibits ␤-1,3-glucan synthesis through Gsc-1-type proteins (34).
Cloning of P. carinii Genomic and cDNA Sequences Encoding Gsc-1-A degenerate PCR strategy was utilized to clone P. carinii sequences encoding the putative Gsc-1 synthetase (44 -46). PCR amplification of P. carinii genomic DNA was performed using degenerate primers designed from conserved amino acid sequences found in ␤-1,3glucan catalytic subunits from Aspergillus nidulans (fksA) and S. cerevisiae (FKS1/FKS2) (27,28,(47)(48)(49). The codon bias for P. carinii was used to limit degeneracy in the third codon position (46). . P. carinii genomic DNA served as the template. An initial 5-min hot start at 94°C was followed by 30 cycles of 94°C for 60 s, 60°C for 60 s, 72°C for 60 s, and a final 72°C 15-min extension. A single 324-bp amplicon was generated. Sequence comparisons to GenBank were performed using the Basic Local Alignment Search Tool (BLAST) genetic analysis program (National Center for Biotechnology Information) (50). The 324-bp amplicon was used to obtain genomic DNA sequence by probing a rat-derived P. carinii genomic DNA library in gt11 (37). The initial amplicon was represented near the 5Ј-region of Gsc-1. Sequences from the 3Ј-regions of these genomic sequences were used to further probe this library for four additional genomic clones. By identifying overlapping sequences from these four clones, the entire 6029-bp genomic region of the P. carinii gsc-1 gene was isolated. P. carinii gsc-1 cDNA sequences were subsequently isolated by reverse transcriptase-PCR of total RNA extracted from freshly purified P. carinii by guanidium isothiocyanate and using overlapping primer pairs covering the entire P. carinii gsc-1 genomic sequence. PCR amplification products were subcloned into pCRII and sequenced.
Southern and Chromosomal Hybridization of P. carinii Gsc-1-To verify that the PCR product was of P. carinii origin, the 324-bp amplicon was hybridized both to digested P. carinii genomic DNA and separated P. carinii chromosomes. The amplicon was labeled using [␣-32 P]dATP by the random primer method (Rediprime System, Amersham Pharmacia Biotech). 20 g of genomic DNA was digested with the restriction enzymes specified, separated on a 1% agarose gel, and transferred to nitrocellulose (51). The membranes were incubated with the probe (1.5 ϫ 10 6 cpm/ml) at 60°C over 1 h, washed three times at room temperature for 40 min in 2ϫ SSC buffer containing 0.05% SDS, washed twice at 50°C for 40 min in 2ϫ SSC buffer containing 0.1% SDS, and examined by autoradiography. In parallel, the 324-bp amplicon was hybridized to P. carinii chromosomes separated by contourclamped homogenous field (CHEF) electrophoreses blot as described previously (38).
Antibody Generation to the Predicted P. carinii Gsc-1 Protein and Immunoblotting of P. carinii Extracts-To determine whether a Gsc-1 protein was present in P. carinii membranes, a polyclonal antibody was generated to the predicted amino acid sequence of P. carinii Gsc-1. A 15-residue peptide (EEMTPTEESPYNPNE), corresponding to amino acids 1196 -1210, was chosen for its high specificity and favorable antigenic profile (GCG software, Oxford Molecular Co., Madison WI) (52). Polyclonal antisera against this peptide were generated in rabbits, reactivity was verified by enzyme-linked immunosorbent assay, and reactive IgG antibody was isolated by purification over protein A-Sepharose (45). To isolate the P. carinii Gsc-1 protein, P. carinii cell wall membrane isolates (ϳ50 g) were immunoprecipitated with either 100 g/ml of the anti-Gsc-1 antibody or non-immune IgG using protein A-Sepharose. Precipitated proteins were subsequently separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. The presence of P. carinii Gsc-1 was assessed by immunoblotting with anti-Gsc-1 or non-immune IgG (100 g/ml) for 1 h. To assess the abundance of Gsc-1 over the life cycle of P. carinii, isolated organisms were further separated into cystic and trophic forms by differential filtration through 3-m filters, which allow the passage of trophic forms, but retain cysts. This separation procedure yielded 99.5% pure trophic populations and preparations that were 40-fold enriched in P. carinii cysts (44). The separated populations were lysed and analyzed by immunoprecipitation and immunoblotting as described above. To further assess that the anti-Gsc-1 antibody was not reacting with any host cell contaminant, an identical concentration of uninfected rat protein was assayed in an identical fashion.
Determination of Glucan Synthetase Activity by P. carinii Gsc-1 Protein-To further verify that the cloned P. carinii gsc-encoded a protein with functional Gsc-1 glucan synthetase activity, cell wall membrane isolates of P. carinii (50 g) were immunoprecipitated with either anti-Gsc-1 or non-immune IgG. Precipitated proteins were resuspended in 50 mM Tris-HCl, 150 mM NaCl buffer with 1 mM EDTA, 1 mM mercaptoethanol, and 33% glycerol (pH 7.5), and UDP-[ 14 C]glucose incorporation into trichloroacetic acid-insoluble material assayed as described previously (43).
Assessment of P. carinii gsc-1 mRNA Expression over the Life Cycle of the Organisms-P. carinii were separated into cystic and trophic forms by differential filtration through 3-m filters, as detailed previously. RNA was extracted from the separated forms (Trizol Reagent, Life Technologies, Inc.), and equal RNA (5.0 g) was separated by electrophoresis through 1.2% agarose in the presence of 2.2 M formaldehyde. Separated RNA species were transferred to nitrocellulose. The 324-bp gsc-1 DNA partial clone was labeled with [␣-32 P]dCTP (Amersham Pharmacia Biotech) by a random primer method, added to the hybridization solution (1 ϫ 10 6 cpm/ml), and incubated with the membranes for 1 h at 68°C. After hybridization, the membranes were washed four times with 2ϫ SSC solution (salt sodium citrate, where 1ϫ solution contained 150 mM NaCl and 15 mM sodium citrate; pH 7.0) with 0.05% SDS at room temperature for 40 min followed by 0.1ϫ SSC with 0.1% SDS solution at 50°C for 40 min. The blots were visualized by autoradiography. For comparison to P. carinii gsc-1, membranes were rehybridized with a P. carinii actin probe (53).

P. carinii Cell Wall Membrane Isolates Incorporate UDP in a
Fashion Inhibitable by the Pneumocandin B 0 Analog L733-560 -Isolated P. carinii cell wall membrane preparations incorporated significant UDP-[ 14 C]glucose into insoluble carbohydrate over 4 h of incubation ( Fig. 1; *p ϭ 0.0392 comparing UDP-Glc incorporation in the presence and absence of P. carinii membrane isolates). To further address whether P. carinii cell wall membrane isolates contained a Gsc-1-type ␤-1,3-glucan synthetase, parallel assays were performed in the presence of the pneumocandin analog L-733,560 (Merck, Inc.) (34). In a concentration-dependent manner, L-733,560 suppressed glucan synthesis ( Fig. 1; **p Ͻ 0.05 comparing UDP-Glc incorporation in the presence of ϭ 1.0 mM of L-733,560 versus its absence). Pneumocandin L-733,560 exhibited an IC 50 of ϳ0.30 M against P. carinii cell wall membrane isolates, strongly indicating that isolated P. carinii membranes assemble glucan through a Gsc-1-type synthetase.
To further verify that the activity was indeed derived from P. carinii membranes rather than from some host cell contaminant, parallel experiments were also performed in which equal volumes of rat lung from uninfected control animals were processed in an identical fashion. An equal amount of protein from uninfected control lung was assayed in parallel to P. carinii cell wall membrane preparations for their ability to incorporate UDP-[ 14 C]glucose under these conditions. Normal rat lung protein (30 g) only incorporated 98.8 Ϯ 15.4 cpm, barely distinguishable above background, compared with a concurrent incorporation of 1327.5 Ϯ 122.2 cpm resulting from an equal concentration (30 g) of P. carinii-derived cell wall membrane isolates. Hence, the activity measured strongly represents the ability of P. carinii organisms, and not any host cell contaminant, to incorporate UDP-[ 14 C]glucose into trichloroacetic acidinsoluble glucans utilizing a Gsc-1-type glucan synthetase.
Molecular Cloning Reveals That P. carinii Contains a Unique gsc-1 ␤-1,3-Glucan Synthetase Gene-We utilized a degenerative PCR cloning strategy based on amino acid similarities in other fungal glucan synthetases to clone P. carinii gsc-1 (26 -28). Amplification of P. carinii genomic DNA with these degenerate primers yielded a single 324-bp product with homologies to related fungal gsc and fks genes. To verify that the PCR product was specifically represented within P. carinii and not an amplified contaminant, this amplicon was hybridized to digested P. carinii genomic DNA demonstrating strong localization as a single band following both EcoRI and HindIII digestions (Fig. 2). In addition, the 324-bp P. carinii gsc-1 sequence was hybridized to a CHEF blot of P. carinii chromosomes. The 324-bp probe consistently hybridized to a single P. carinii chromosomal location (Fig. 3). Together, these studies indicate that P. carinii gsc-1 is present as a single copy gene within the organism.
Subsequently, a P. carinii gt11 genomic library was screened, and four clones were isolated based on hybridization with the 324-bp gsc-1 fragment. Clone G6 of ϳ3.5-kb size was fully sequenced. Comparison using BLAST-X analyses revealed this unique sequence to contain homologies to the amino-terminal sequences of the ␤-1,3-glucan catalytic subunits of A. nidulans, A. fumigatus, and S. cerevisiae. Sequence from the 3Ј-coding region of clone G6 were used to further screen the library. Three additional clones were identified and found to contain additional sequence, including a putative stop codon.
Complete genomic DNA sequences of P. carinii gsc-1 have been deposited in GenBank (accession number AF191096). Translation suggested three separate open reading frames encoding peptide sequences with homology to A. nidulans FksAp and A. fumigatus Fksp. An analysis for the presence of putative P. carinii-type introns, using our recently reported acceptor and donor consensus criteria, revealed the likelihood of three introns positioned at nucleotides 334 -398, 2619 -2695, and 5495-5546 of P. carinii gsc-1 (54). Predicted excision of these introns resulted in a mature protein with overall sequence homology comparable to other fungal ␤-1,3-glucan synthetases. To further confirm the intron/exon splicing sites, complete cDNA sequences for P. carinii gsc-1 (GenBank accession number AF291999) were derived by reverse transcriptase-PCR FIG. 2. The 324-bp P. carinii gsc-1 gene fragment specifically hybridizes to P. carinii genomic DNA. P. carinii was freshly isolated and genomic DNA isolated and digested with the indicated restriction endonucleases. The digestion products were separated by electrophoresis and transferred to nitrocellulose. The 324-bp gsc-1 amplicon was labeled and hybridized to the membrane showing specific interaction as a single band on the EcoRI and HindIII digestions. Thus, gsc-1 appears to represented as a single locus within P. carinii genomic DNA.

FIG. 1. P. carinii cell wall membrane preparations incorporate UDP-[ 14 C]glucose into ␤-glucans, in a manner inhibited by the pneumocandin analog L-733,560.
To assess whether P. carinii actively incorporates UDP-Glc into insoluble carbohydrates, organisms were purified and cell wall membrane preparations isolated and reacted with radiolabeled UDP-Glc as described under "Experimental Procedures." Complex carbohydrates were precipitated with trichloroacetic acid, washed, and counted. P. carinii cell wall membrane isolates mediated significant incorporation of UDP-Glc (* denotes p Ͻ 0.05 comparing UDP-Glc incorporation in the presence and absence of PC cell wall membrane extract). To further determine the presence of a Gsc-1-type glucan synthetase in P. carinii, additional incorporation assays were conducted in the presence of the pneumocandin analog L-733,560, an agent which specifically inhibits Gsc-1-type proteins. In a concentration-dependent manner, P. carinii glucan synthetase activity was inhibited by this pneumocandin (** denotes p Ͻ 0.05 comparing P. carinii-mediated incorporation of UDP-Glc in the presence and absence of L-733,560). of total RNA extracted from freshly isolated P. carinii using overlapping primer pairs covering the entire P. carinii gsc-1 genomic sequence. Comparison of the genomic and cDNA sequences confirmed that the P. carinii gsc-1 gene is comprised of four exons and three introns spliced at the sites predicted using the acceptor and donor consensus criteria. Minor variations in sequences derived from different sources of P. carinii organ-isms have been previously reported by our group and others (44 -46). A 3.6% difference in nucleotide sequence was observed comparing the genomic clone, derived from the University of Cincinnati library, to the cDNA sequence generated from P. carinii RNA freshly obtained from our rat colony housed in Rochester, MN. This is comparable with previous reports for P. carinii cdc2, another similarly conserved single

FIG. 5. An antibody generated against the predicted P. carinii
Gsc-1 protein, recognizes a molecule of the appropriate mass from P. carinii cell wall membrane extracts. A, anti-Gsc-1 antibody was generated against a synthetic peptide predicted from the extracytoplasmic domain of P. carinii Gsc-1 and used to immunoprecipitate P. carinii cell wall membrane isolates and uninfected control rat lung. The anti-Gsc-1 antibody specifically precipitated a single protein of 219.4-kDa molecular mass, detected by silver staining (arrow). In contrast, uninfected control rat lung exhibited no specific reactivity. The material migrating at ϳ50 kDa is consistent with immunoglobulin heavy chain and was detected as expected in both the P. carinii and control uninfected lung precipitations. B, the anti-peptide antibody was further used to evaluate relative Gsc-1 abundance in separated cysts and trophic forms of P. carinii. P. carinii cell wall membrane isolates were sequentially immunoprecipitated and immunoblotted with the anti-Gsc-1 antibody. Abundant Gsc-1 was detected in P. carinii cysts, with much lesser amounts noted in trophic forms by immunoblot analysis. An equal amount of uninfected rat lung protein failed to exhibit any specific immune reactivity with anti-Gsc-1, thus further confirming specificity of this antibody.

FIG. 6. P. carinii gsc-1 expression is regulated over the life cycle of the organism.
To examine whether gsc-1 expression is differentially regulated over the life cycle of P. carinii, freshly isolated organisms were separated into cystic and trophic populations, and RNA was extracted and examined for gsc-1. The top panel shows hybridization of the nitrocellulose membrane with the P. carinii gsc-1 probe (ϳ9.5-kb mRNA transcript), whereas the bottom panel shows rehybridization with P. carinii actin to confirm equal RNA loading. P. carinii gsc-1 mRNA expression was predominantly restricted to the cystic form of the organisms. Also included is an identical amount of RNA derived from uninfected rat lung (negative control), which did not hybridize with either P. carinii gsc-1 or actin. 3. The P. carinii gsc-1 amplicon hybridizes to a single P. carinii chromosome. The left panel shows ethidium-stained P. carinii chromosomes resolved by contour-clamped homogenous field electrophoresis (CHEF). Each of the four lanes represents the chromosomes derived from an individual P. carinii isolate purified from a single infected rat (38). In the right panel, the separated chromosomes were transferred to nitrocellulose and hybridized with the radiolabeled 324-bp gsc-1 probe. The P. carinii gsc-1 amplicon interacts with a single P. carinii chromosome within the organism's genome. copy gene, which demonstrated a 6.3% difference comparing nucleotide sequences obtained from P. carinii strains obtained from Rochester, MN and Cincinnati, OH (44). These very minor differences between the genomic and the cDNA P. carinii gsc-1 sequences likely represent strain variations in the P. carinii sources.
Isolated P. carinii Cell Wall Membrane Preparations Contain an Appropriately Sized Gsc-1 Protein-An antibody to a unique extracytoplasmic region of P. carinii Gsc-1 was generated and used to immunoprecipitate P. carinii membrane isolates and uninfected control rat lung. The anti-Gsc-1 antibody reacted specifically with a single P. carinii protein of molecular mass 219.4 kDa on silver staining (arrow, Fig. 5A). In contrast, uninfected rat lung exhibited no specific reactivity. The additional material migrating at ϳ50 kDa is consistent with immunoglobulin heavy chain, and was present as expected in both the P. carinii and control lung precipitations. It is noteworthy that the observed molecular mass of 219.4 kDa correlated very closely to the 214-kDa size predicted from the gene sequences.
To further assess the presence of Gsc-1 over the life cycle of P. carinii, isolated organisms were next separated into cysts and trophic forms. Previous immunohistochemical studies have demonstrated that ␤-glucans are prominent components of P. carinii cysts and are relatively less prevalent in the walls of trophic forms (20). In contrast, recent work also suggests that pneumocandin Gsc-1 antagonists may exert some activity against trophic forms of the organism (36). To address this, trophic forms and cysts were separated by differential filtration, and their cell wall membrane isolates were analyzed by sequential immunoprecipitation and immunoblotting (Fig. 5B). Considerable Gsc-1 protein was present in P. carinii cysts. In contrast, detectable, but rather limited amounts, of Gsc-1 were observed in isolated trophic forms. Thus, the Gsc-1 glucan synthetase, putatively active in cell wall assembly, is present at markedly different levels over the life cycle of the organism, with greatest expression found in P. carinii cysts. As a further confirmation of the specificity of the anti-Gsc-1 antibody, an equal amount of uninfected rat lung protein was similarly processed as before and showed no reactivity with the anti-Gsc-1 antibody.
P. carinii gsc-1 mRNA Expression Is Largely Restricted to Cystic Forms-Having observed differential abundance of Gsc-1 protein in isolated cyst and trophic forms of the organism, we next investigated whether gsc-1 gene expression was regulated over P. carinii's life cycle. Cysts and trophic forms were separated and gsc-1 mRNA expression evaluated by Northern analysis (Fig. 6). Again, P. carinii gsc-1 mRNA was largely restricted to cysts. Repeat hybridization with P. cariniispecific actin revealed that the marked abundance of gsc-1 in cysts was not the consequence of RNA loading. (In fact, greater actin mRNA was present in the trophic lane.) Lastly, we observed no cross-hybridization of either the P. carinii gsc-1 or actin probes with RNA derived from uninfected rat lung, confirming the specificity of the P. carinii sequences. Thus, the expression of gsc-1 is uniquely regulated over the life cycle of P. carinii.
Antibody to the Predicted P. carinii Gsc-1 Protein Immunoprecipitates a Molecule Catalyzing UDP-Glc Incorporation into Glucans-We next sought to directly link enzymatic activity to the genetic sequences to confirm that we have indeed cloned and characterized a Gsc-1-type protein from P. carinii, capable of mediating glucan synthesis. This was investigated by performing immunoprecipitation with the specific anti-P. carinii Gsc-1 antibody followed by UDP-[ 14 C]glucose incorporation of the precipitated products (Fig. 7). Anti-P. carinii Gsc-1 antibody, but not non-immune IgG, precipitation of P. carinii cell wall membrane isolates yielded a product that strongly incorporates UDP-[ 14 C]glucose into trichloroacetic acid-insoluble material, thus providing strong independent evidence that we have identified a Gsc-1 glucan synthetic enzyme from P. carinii organisms. In contrast, uninfected rat lung proteins did not appreciably react with either the ani-Gsc-1 antibody or nonimmune IgG, again confirming that the activity identified is from the organisms and not the result of host cell contamination. DISCUSSION P. carinii membrane isolates possess the ability to incorporate UDP-[ 14 C]glucose into insoluble carbohydrate, which is FIG. 7. P. carinii Gsc-1 mediates UDP-[ 14 C]glucose incorporation into glucans. P. carinii or uninfected rat lung proteins were immunoprecipitated with either anti-Gsc-1 antibody or with nonimmune IgG, and the precipitates were reacted with radiolabeled UDP-Glc as described under "Experimental Procedures." Complex carbohydrates were precipitated with trichloroacetic acid, washed, and counted. P. carinii cell wall membrane isolates incubated with anti-Gsc-1 antibody, but not non-immune IgG, incorporated significant UDP-Glc into complex carbohydrate. In contrast, uninfected rat lung protein failed to immunoprecipitate any product capable of mediating UDP-Glc incorporation. Therefore, immunoreactive P. carinii Gsc-1 protein specifically mediates UDP-[ 14 C]glucose incorporation into complex glucan carbohydrate. (*, p ϭ 0.0001 compared with control.) inhibited by the pneumocandin L-733,560 antagonist of Gsc-1type ␤-1,3-glucan synthetases. Molecular cloning of the P. carinii gsc-1 gene predicts a mature protein with both similarities and unique differences to other fungal ␤-1,3-glucan synthetases. Specifically, the mRNA and protein expression of P. carinii gsc-1 are highly regulated over the life cycle of the organism being predominantly expressed by cystic forms of the organism. A protein corresponding to the predicted sequence of the cloned gsc-1 gene was present in P. carinii membrane isolates. Furthermore, immunoprecipitation of the putative P. carinii Gsc-1 protein with a synthetic peptide antibody yielded a product capable of mediating incorporation of UDP-Glc into trichloroacetic acid-insoluble material, consistent with glucan.
P. carinii Gsc-1 glucan synthetase exhibits several unique features. As discussed, structural differences were detected in the domain configuration of P. carinii Gsc-1 compared with S. cerevisiae and Aspergillus (26 -28). Of further contrast, is the significant restriction of gsc-1 expression predominantly to the cystic forms of the P. carinii. Most other fungi, including ascomycetous fungi, exhibit cell wall assembly constitutively throughout the life cycle (55). Our immunoblot and Northern analyses are complementary to previous immune localization studies, which also indicate that ␤-1,3-glucan is largely found within cysts (6,20).
Because mammalian hosts do not possess an equivalent to Gsc-1, inhibition of ␤-1,3-glucan synthesis represents an attractive target for treatment of fungal infections. Echinocandins and pneumocandins are selective lipopeptide inhibitors that may expand our armamentaria for fungal infections, including those organisms resistant to standard agents (32). Schmatz and colleagues (33,34) have shown rapid reduction of organisms in rat and mouse models of P. carinii pneumonia. One pneumocandin in particular, L-671,329 has shown remarkable activity in the P. carinii rat model with Ͼ98% of cysts being eliminated (33).
Concern had arisen that pneumocandins might only be effective against cystic forms of P. carinii, thereby limiting efficacy of such compounds during P. carinii pneumonia. The striking results of pneumocandins in animal models of P. carinii pneumonia strongly argue to the contrary (34). These findings do suggest, however, that progression of trophic forms into cysts represents an essential component of life cycle progression in P. carinii, rather than an elective form utilized only under hostile conditions (7). Other investigators have also found some effect of echinocandins on trophic structure after in vitro exposure (36). Although our study demonstrates low levels of gsc-1 mRNA expression in trophic forms, small residual amounts of Gsc-1 protein were detected by Western analysis to remain within trophic forms. In addition, echinocandin and pneumocandin compounds may also effect other targets within P. carinii.
Until recently, the lack of a reliable culture system has hindered studies of life cycle regulation by P. carinii (56). Recent studies implicate a cyclin-dependent kinase cell cycle control system, which exhibits regulated activity during progression of P. carinii trophic forms to cysts (44,57). Considerable questions remain as to how assembly of the thickened ␤-glucan-rich pellicle is specifically limited to the cystic form of the organism. A variety of potential environmental signals, including interaction with lung epithelium, availability of nitrogen and lipid substrates, and the presence of differential mating types might initiate progression to cyst formation. Recently, Merali and Clarkson (58) have reported continuous axenic culture of P. carinii. Further exploitation of this system using the molecular tools we report should provide essential insights into regulation of the P. carinii life cycle.
In summary, we have observed pneumocandin-inhibitable glucan synthetase activity within cell wall membrane isolates of P. carinii and have identified and characterized a gsc-1-type ␤-1,3-glucan synthetase gene from this organism. A corresponding protein of appropriate 219.5-kDa molecular mass was present in P. carinii using an antibody generated to the predicted Gsc-1 protein. The expression of P. carinii gsc-1 is regulated over the life cycle of the organisms. In view of its central role in assembly of the P. carinii cyst wall, the Gsc-1 ␤-1,3glucan synthetase is an attractive therapeutic target for the treatment of P. carinii pneumonia.