Characterization of N-Acetylglucosamine Biosynthesis in Pneumocystis species. A New Potential Target for Therapy

Abstract

N-acetylglucosamine (GlcNAc) serves as an essential structural sugar on the cell surface of organisms. For example, GlcNAc is a major component of bacterial peptidoglycan, it is an important building block of fungal cell walls, including a major constituent of chitin and mannoproteins, and it is also required for extracellular matrix generation by animal cells. Herein, we provide evidence for a uridine diphospho (UDP)–GlcNAc pathway in Pneumocystis species. Using an in silico search of the Pneumocystis jirovecii and P. murina (Pm) genomic databases, we determined the presence of at least four proteins implicated in the Saccharomyces cerevisiae UDP-GlcNAc biosynthetic pathway. These genes, termed GFA1, GNA1, AGM1, and UDP-GlcNAc pyrophosphorylase (UAP1), were either confirmed to be present in the Pneumocystis genomes by PCR, or, in the case of Pm uap1 (Pmuap1), functionally confirmed by direct enzymatic activity assay. Expression analysis using quantitative PCR of Pneumocystis pneumonia in mice demonstrated abundant expression of the Pm uap1 transcript. A GlcNAc-binding recombinant protein and a novel GlcNAc-binding immune detection method both verified the presence of GlcNAc in P. carinii (Pc) lysates. Studies of Pc cell wall fractions using high-performance gas chromatography/mass spectrometry documented the presence of GlcNAc glycosyl residues. Pc was shown to synthesize GlcNAc in vitro. The competitive UDP-GlcNAc substrate synthetic inhibitor, nikkomycin Z, suppressed incorporation of GlcNAc by Pc preparations. Finally, treatment of rats with Pneumocystis pneumonia using nikkomycin Z significantly reduced organism burdens. Taken together, these data support an important role for GlcNAc generation in the cell surface of Pneumocystis organisms.

Clinical Relevance

Pneumocystis pneumonia remains an important infection in immune-compromised patients. These studies describe the pathways used to synthesize a specialized sugar, termed N-acetylglucosamine (GlcNAc), which is an important component of the organisms' cell surface. Inhibition of synthesis of GlcNAc reduces organism burden during infection in rodent models.

The amino sugar, N-acetylglucosamine (GlcNAc), is an abundant component of the eukaryotic and prokaryotic cell wall. In yeast, uridine diphospho (UDP)–GlcNAc, the active form of the amino sugar, is an essential precursor needed for the synthesis of glycosylphosphatidylinositol (GPI) anchors, mannoproteins, chitin, and chitosan (1). Members of the UDP-GlcNAc synthetic pathway in fungi have been shown to be essential in fungal growth, and therefore represent viable targets for antifungal therapy. For example, recently, analysis of the Aspergillus fumigatus UDP-GlcNAc pyrophosphorylase (UAP), the final enzyme in eukaryotic UDP-GlcNAc biosynthesis, which converts uridine triphosphate and GlcNAc-1-phosphate (GlcNAc-1P) to UDP-GlcNAc, was shown by crystal structure analysis to have enough variation in its composition from the human enzyme to be a potential target for antifungal therapy (2).

The cell surface of the pathogenic fungal species represented by Pneumocystis organisms have not been fully elucidated, but are known to contain β-1,3 and β-1,6 glucans, and the major surface mannose-rich glycoproteins, variously termed major surface glycoprotein (MSG) or glycoprotein A (gpA) (3–5). We and others have shown that the Pneumocystis species contain β-glucan synthetases that can be inhibited by specific β-glucan inhibitors, such as the echinocandins, pneumocandins, and newly generated compounds (3, 4, 6, 7). A number of these drugs are already implemented as antifungal therapies against pathogens (8). Similarly to β-glucans, mannoproteins on the surface of fungi have been demonstrated to be required for organism viability and to avoid host immune surveillance (9).

Although a number of studies have evaluated the presence and antigenic variation of MSG/gpA, the Pneumocystis equivalent to yeast mannoproteins (10, 11), information regarding the initial characterization of the UDP-GlcNAc pathway that most likely provides the chitobiose structure responsible for tethering the membrane-bound mannoprotein to the cell wall is lacking. Herein, we analyze biochemical and genetic evidence and demonstrate the presence of the GlcNAc biosynthesis pathway, as well as GlcNAc-containing molecules on the cell surface of Pneumocystis species.

Identification of GlcNAc-containing residues such as those present in chitobiose or GPI anchored molecules in Pneumocystis species, and their role in cell wall biology, have not been fully studied. For example, over the past 20 years, only a handful of studies have searched for GlcNAc residues in Pneumocystis organisms. These included classic biochemical methods for determination of this moiety in the organism via GlcNAc-binding/-labeling experiments using antibody staining techniques, as well as incorporation of ³H-glucosamine into the P. carinii (Pc) cell walls, supportive of GlcNAc deposition (12–14). More recently, an intriguing study by Rapaka and colleagues (15) suggested that mice infected with P. murina (Pm) could generate increased levels of IgM antibodies against fungal carbohydrates, including GlcNAc-containing carbohydrates. Recently, however, it has been shown that, although Pneumocystis species contain chitin chaperone proteins implicated in proper transportation of major chitin synthases to the cell membrane in yeast, the genomes of Pneumocystis species lack major chitin synthase or chitin-degrading chitinase enzymes (5). In this light, we initiated a comprehensive analysis of the various enzymes potentially leading to the synthesis of GlcNAc in Pneumocystis organisms. Using both detailed biochemical and genetic analytical strategies, we provide evidence that Pneumocystis species indeed possess GlcNAc residues. We further provide cumulative genetic information for a potential GlcNAc biosynthetic pathway in these organisms, serving as an essential precursor of important cell wall components, such as GPI anchors for Pneumocystis MSG/gpA. In addition, treatment of isolated Pc cell wall preparations and rats with Pneumocystis pneumonia (PCP) using nikkomycin Z, an analog of UDP-GlcNAc that acts as a competitive inhibitor of cell surface carbohydrate synthesis, verified significant inhibition of newly synthesized GlcNAc residues in vitro, as well as reduction in Pc organism burdens during infection.

Materials and Methods

Strains and Reagents

For these studies, Pc and Pm organisms were originally derived from American Type Culture Collection (Manassas, VA) stocks and propagated in corticosteroid-treated rats or mice, as reported previously (16, 17). Populations of Pc were isolated from chronically infected rat lungs by homogenization and filtration through 10-μm filters, as we previously described (3, 18). Unless otherwise noted, the remaining reagents were obtained from Sigma-Aldrich (St. Louis, MO).

Detailed Experimental Procedures

Detailed experimental procedures are found in the online supplement. Procedures in the supplement include: immunofluorescence detection of GlcNAc reactivity in Pc (19–26), Western analysis of Pc GlcNAc complexed to proteins (27, 28); verification of GlcNAc biosynthetic genes in the Pc and Pneumocystis jirovecii (Pj) genomes (16); expression of Pmuap1 mRNA in Pm-infected mice (29); expression, purification, and enzymatic activity of PmUap1 (2, 16, 30–32); determination of Pc synthesis of GlcNAc (33–35); glycosyl composition of Pc GlcNAc preparations using gas chromatography/mass spectrometry (GC/MS) (36–39); and nikkomycin Z treatment of Pc-infected rats (40–42). Furthermore, all nucleotide sequence accession numbers are included in the supplement. Specific PCR primers used in these studies are listed in Table E1 in the online supplement.

Statistical Analysis

Statistical analyses between various experimental conditions were first assessed using ANOVA and subsequently Student’s t tests, as indicated. Nonparametric statistics were used when data were distributed in a non-gaussian manner. Statistical testing was performed employing GraphPad Prism version 5.0b (GraphPad Software, Inc., La Jolla, CA), and statistical differences were considered to be significant with a P value of less than 0.05.

Results

The Pc Cell Wall Exhibits GlcNAc Reactivity by Immunofluorescence and GlcNAc Binding Analysis

As an initial characterization of the presence of GlcNAc residues in Pc, we assessed the presence of such molecules in freshly isolated Pc organisms using immunofluorescence (Figure 1). Pc strongly bound wheat-germ agglutinin (WGA), a lectin probe with reactivity to GlcNAc-containing structures (Figure 1C). Similar findings have been previously observed by other investigators (43, 44). Because WGA binds not only GlcNAc present in fungal cell wall membranes, but also N-acetylneuraminic acid, the major sialic acid present in mammalian cells, we further used an immunofluorescence assay employing a GlcNAc-recognizing chitin-binding domain (CBD) protein linked to the engineered O-6 methylquanine-DNA methyl transferase (SNAP)-tagged protein (45, 46). This protein-tagged probe exhibits high specificity for GlcNAc residues, including chitobiose (GlcNAc)₂, based on its selective binding domain isolated from the bacteria, Bacillus circulans WL-12 (47). After 15 minutes of zymolyase digestion, the SNAP-tagged CBD probe displayed strong binding to Pc (Figure 1D). Of note, however, some, though not all, Pc organisms in the preparations were detected using this approach. Thus, expression levels of GlcNAc residues may occur at specific times in the organism life cycle or under selected conditions (for example, when production of MSG/gpA mannoproteins are needed for host cell attachment or host evasion by antigenic variation). Taken together, these findings support the presence of GlcNAc residues in Pc.

Figure 1.

Lectin and affinity binding assays demonstrate N-acetylglucosamine (GlcNAc) reactivity in the Pneumocystis cell wall. (A and B) Phase microscopy images of Pneumocystis carinii (Pc) life forms (200× magnification). (C) Wheat-germ agglutinin (WGA)-FITC staining of Pc life forms. (D) Fluorescence microscopy staining of engineered O-6 methylquanine-DNA methyl transferase (SNAP)-tagged chitin-binding domain (CBD) binding protein of Pc life forms followed by addition of polyclonal anti-SNAP antibody displays specific staining of Pc organisms.

We next used the SNAP-tagged CBD for Western analysis adapting a method used for detection of β-1,6 glucans associated with Candida albicans proteins (27). Using this strategy, a prominent GlcNAc-containing protein conjugate band with a size of roughly 45 kD was detected (Figure 2A, left panel). Immune blotting with a nonimmune isotype control antibody yielded no specific reactivity (Figure 2A, right panel).

Figure 2.

Western analysis of Pc GlcNAc–bound complexes confirms GlcNAc reactivity in Pneumocystis extracts. (A) Pc cell wall preparations were separated by PAGE, and potential GlcNAc complexed to Pc proteins were assayed by SNAP-tagged CBD protein incubation followed by anti-SNAP antibody (Ab) detection. Concentrations of total Pc protein preparations were as follows: lane 1, 100 μg; lane 2, 50 μg; lane 3, 25 μg. (B) Total Pc lysate preparations were either treated with chitinase reaction buffer alone or with 1 U of chitinase from Streptomyces griseus for 2 hours at 37°C before analysis.

It has been shown that Streptomyces fungal species exert antifungal activities against pathogenic fungi (48). In Streptomyces griseus, this antifungal activity has been attributed to a major endochitinase, a member of a family of chitinases containing 19 such hydrolytic enzymes, first discovered in plants (28, 49). This enzyme cannot only digest chitin polymers, but can also digest chitobiose that is linked to mannoproteins (26). It is of interest that digestion of our Pc cell wall preparations with this endochitinase, followed by PAGE and probing with the SNAP-tagged CBD, resulted in almost complete elimination of the 45-kD band with the formation of a single band of less than 15-kD (Figure 2B). These observations further support the presence of GlcNAc-containing molecules on the Pc cell surface.

The Pj Genomes Contain GlcNAc Biosynthetic Genes

In Saccharomyces cerevisiae, the GlcNAc metabolism genes, GFA1, GNA1, PCM1/AGM1, and QRI1/UAP1 are essential for viability (see Figure 3 for the yeast UDP-GlcNAc biosynthetic pathway) (50). In an attempt to determine whether members of these GlcNAc pathway–related genes are present in human-derived Pj, we searched the completed Pj genome (http://genome.jgi.doe.gov/pneji1/pneji1.home.html) using keywords for these yeast genes. After computer-based analysis of the Pj genome, we further determined that, indeed, Pj genomic material from human bronchoalveolar lavage fluid samples contains all four members of the GlcNAc synthetic pathway. In addition, Pj homologs of the upstream metabolic protein members, Pgi1p, a glycolytic phosphoglucose isomerase catalyzing the interconversion of glucose-6-phosphate and fructose-6-phosphate (51), and Hxh2p, which catalyzes the phosphorylation of glucose in the yeast cell (52), were also detected (Table E2). Primers were designed based on the cDNA sequences of Pjgfa1, Pjgna1, and Pjagm1 (http://genome.jgi.doe.gov/pneji1/pneji1.home.html). These primers were added to either Pj genomic DNA or healthy human lung genomic DNA. To confirm the Pj Basic Local Alignment Search Tool X (BLASTX) database search, we performed PCR amplification of these three essential GlcNAc-regulatory genes using Pj DNA, and successfully amplified portions of all three GlcNAc synthase pathway genes (Figure 4). A primer set based on hGAPDH confirmed the integrity of the human DNA integrity. Furthermore, because the essential UAP, the final enzyme in the fungal UDP-GlcNAc synthetic pathway, has shown to be a potential antifungal target in A. fumigatus, we also wanted to determine whether the UAP Pj homologue, Pjuap1, was also present in the Pj material isolated from patients with PCP. Indeed, as shown in Figure 5A, a specific amplicon was noted, indicating the presence of the UAP homolog within the Pj genomic material. These results provide further evidence for the GlcNAc synthase pathway existing in Pneumocystis species, and of particular importance in Pj, which is responsible for disease in humans.

Figure 3.

Uridine diphospho (UDP)–GlcNAc synthetic pathway in Saccharomyces cerevisiae and other yeast. Proteins in bold indicate these genes are confirmed to be present in P. jirovecii (Pj). Of note, GFA1, GNA1, PCM1/AGM1, and QRI1/UDP-GlcNAc pyrophosphorylase (UAP1) are essential in S. cerevisiae.

Figure 4.

Pj GlcNAc biosynthetic pathway homolog genes are present in organisms derived from human bronchoalveolar lavage fluid (BALF). Pjgfa1-, Pjgna1-, and Pjagm1-specific primer sets amplified specific products from Pj genomic DNA derived from BALF but not human genomic DNA. Human glyceraldehyde-3-phosphate dehydrogenase (hGAPDH) serves as an appropriate amplification control. bp, base pairs.

Figure 5.

Pneumocystis species genomes contain a functional Uap1 final biosynthetic enzyme in the GlcNAc pathway, and this transcript is abundantly expressed in mice with Pneumocystis pneumonia (PCP). (A) Amplification of Pjuap1 from Pj genomic DNA confirms its presence in this genome. (B) Active expression of P. murina (Pm) uap1 (Pmuap1) during PCP in the mouse model suggests abundant expression of this GlcNAc synthesis gene during Pneumocystis infection (*P < 0.01 comparing expression in Pm-infected lung samples to uninfected healthy lung). (C) PmUap1 protein incorporation assay confirms enzymatic activity. Specificity of phosphosugar substrates for utilization by PmUap1p was assayed with varying concentrations of uridine triphosphate for 20 minutes at room temperature. Final synthesized products were analyzed by the Biomol green assay coupled with yeast pyrophosphatase. The results are expressed as the mean (±SD) for three experimental determinations (*P < 0.01 comparing GlcNAc-1-phosphate [GlcNac-1P] to glucose-1-phosphate (Glc-1P) activity in the assay at identical concentrations).

The Pm Pmuap1 Transcript Is Actively Expressed during PCP in Mice

To confirm whether the final GlcNAc biosynthetic gene Pmuap1 was not only present in the genomes of Pneumocystis species, but also expressed during active infection during PCP, we isolated total RNA from 10-week-old mice with active Pm infection, or healthy, uninfected mouse lungs. After isolation of total RNA and elimination of any potential contaminating genomic DNA, we analyzed the derived RT-amplified cDNA for the presence of actively transcribed mRNA for the GlcNAc biosynthetic gene. We observed abundant mRNA transcription by quantitative PCR (qPCR) of Pmuap1. No amplification occurred using the mRNA samples from healthy mouse control lungs (Figure 5B). These data indicate that, during PCP, the organisms actively transcribe Pmuap1 mRNA.

Pneumocystis Possesses a Functional Uap1p Enzyme

To further determine whether Pneumocystis species contain a functional Uap1p enzyme, essential for the yeast metabolic GlcNAc synthase pathway, the Pm-derived Pmuap1 cDNA was cloned into a bacterial vector for expressing fusion proteins with a thrombin cleavage site and expressed with a glutathione S-transferase (GST) fusion tag in Escherichia coli. After purification with glutathione sepharose, PmUap1p activity and substrate preference was assessed by incubation of the enzyme with uridine triphosphate and either the phosphosugar, GlcNAc-1P (the preferred substrate before GlcNAc in the GlcNAc synthetic pathway) (2, 32) or with glucose-1-phosphate (Glc-1P), a nonpreferred substrate for Uap1p in yeast (32). As demonstrated in Figure 5C, PmUap1p exhibits preference for GlcNAc-1P, although it can also cleave phosphate from the Glc-1P substrate, but to a much lesser degree than GlcNAc-1P. Figure E1 further demonstrates that, when another nonspecific substrate, mannose 1-phosphate, was used and incubated with PmUap1p, only negligible amounts of detectable free pyrophosphate were generated (see the online supplement). Cleavage of phosphate to inorganic pyrophosphate has also been previously demonstrated in S. cerevisiae with Uap1p acting on Glc-1P, which has been termed “dual substrate utility” (32). Thus, our results indicate that PmUap1 can function as a UAP similar to Uap1 proteins in the GlcNAc biosynthetic pathway of other fungi.

GlcNAc Incorporation into the Pc Cell Wall

We further measured the incorporation of UDP-N-GlcNAc into insoluble cell surface products in Pc. Total Pc cell preparations were incubated in reaction buffer as previously described at various temperatures and de novo UDP-GlcNAc incorporation into insoluble GlcNAc cell surface material detected using wheat germ agglutinin (WGA)-horse radish peroxidase (HRP) conjugate (33, 53). Because Pneumocystis spp. are mammalian lung pathogens, the organism’s preferred temperature of growth in vivo is approximately 37°C, conditions that have been confirmed in the laboratory for short-term maintenance of the organisms in vitro (54, 55). Therefore, as expected, newly generated GlcNAc, as measured by GlcNAc incorporation into insoluble cell surface products, was observed to increase as the reaction temperatures increased from room temperature (26°C) to 37°C (Figure 6A). These experiments document optimal incorporation of GlcNAc into the Pc cell wall under conditions relevant to infection in the mammalian host.

Figure 6.

GlcNAc is incorporated into Pc cell preparations and is inhibited by nikkomycin Z. (A) De novo incorporation of GlcNAc into Pc cell preparations was assessed in the presence of added UDP-GlcNAc (+UDP-GlcNAc) compared with basal incorporation of GlcNAc by the Pc cell preps (−UDP-GlcNAc). Shown are the mean (±SD) from three determinations performed in triplicate (*P < 0.05, significant GlcNAc incorporation at each temperature condition compared with the baseline levels of GlcNAc measured in reactions without added UDP-GlcNAc, **P < 0.05 comparing UDP-GlcNAc incorporation at 37°C to room temperature [RT]). (B) Nikkomycin Z inhibits GlcNAc incorporation into the Pc cell preparations. GlcNAc incorporation by Pc cell preparations was measured in the presence of UDP-GlcNAc and increasing concentrations of nikkomycin at 37°C (*P < 0.01, significant suppression of GlcNAc incorporation).

To further confirm our potential findings that Pc has the ability to incorporate GlcNAc into Pc cell surface components, we added the competitive UDP-GlcNAc analog, nikkomycin Z, to our enzymatic assays. In other fungi, nikkomycin Z typically suppresses the synthesis of chitin and other GlcNAc-containing cell surface components through competitive inhibition of GlcNAc incorporation (56). Recently, nikkomycin Z has also been demonstrated to have activity against the bacterial, a β-1,4-N-acetylglucosamine transferase (57). Even though Pneumocystis genomic analyses recently suggested the absence of chitin synthetic and degradative genes in Pneumocystis species, our data strongly support the present of other GlcNAc-containing molecules on the Pc cell surface. Accordingly, we tested the effects of nikkomycin Z on Pc incorporation of GlcNAc molecules by the organisms. Incubation with nikkomycin Z significantly reduced the generation of newly synthesized GlcNAc containing precipitable material by Pc preparations, as measured by the WGA-HRP binding assay (Figure 6B). Our results indicate that Pneumocystis species have the ability to incorporate GlcNAc into cell surface components, and that nikkomycin Z can inhibit this activity.

Determination of Glycosyl Residue Composition of Pc Cell Preparations by GC-MS

Biochemical support for the existence of GlcNAc in Pneumocystis species further required the isolation of the material from the organism. For this process, we implemented an extraction protocol as described by Wagener and colleagues (58). After extraction of the Pc cell surface material, the resulting samples were analyzed by GC-MS methods, as previously described (59). The results of the analysis are shown in Table E3. Of the total mole percent of the glycosyl residues analyzed, roughly 20% of the sample had GlcNAc. These results further support that Pneumocystis species possess GlcNAc containing material within their insoluble cell surface components.

Efficacy of the Inhibitor, Nikkomycin Z, in the Rat Model of PCP

Finally, we evaluated the use of nikkomycin Z, a competitive antagonist of GlcNAc incorporation, for treatment of active PCP. To accomplish this, we employed the rat Pc pneumonia model that has been used extensively to analyze drug activity against the organism (6, 60, 61). Due to limited drug supply, we treated the rats with nikkomycin Z by oral gavage after establishment of fulminate PCP by immune suppressing rats with corticosteroids and inoculating with Pc. The animals were maintained on the immune suppressive regimen for 8 weeks to develop active PCP. Thereafter, we provided twice-daily nikkomycin Z by oral gavage for 7 days. This regimen was selected due to the relatively short half-life of this drug and the success of this method in treating experimental pulmonary blastomycosis (62). After treatment with nikkomycin Z, rat lungs were harvested and RNA isolated from the lungs. Pc organism burden was determined by qPCR. Similar to what has been shown for other fungal pathogens (62, 63), we observed significant reductions in Pc organism burden in the infected lungs of the rats treated with nikkomycin Z compared with the saline control group (Figure 7E). Furthermore, histology of rodent lungs from animals treated with nikkomycin Z demonstrated greatly reduced organism burdens, as visualized by hematoxylin and eosin and Gomori methenamine silver staining (Figures 7A–7D), respectively.

Figure 7.

Nikkomycin Z reduces Pc organism burden during PCP. After the establishment of fulminate PCP in rats, nikkomycin Z treatment (20 mg/kg) was administered twice daily for 7 days. After this therapy, the rats were killed and histology and Pc organism burden determined, as described in the Materials and Methods. A total of six rats per treatment group was analyzed. (A) Pneumocystis-infected rats treated with saline, hematoxylin and eosin (H&E) staining; (B) Pneumocystis-infected rats treated with nikkomycin Z, H&E staining; (C) Pneumocystis-infected rats treated with saline, Gomori methenamine silver (GMS) staining; and (D) Pneumocystis-infected rats treated with nikkomycin Z, and stained with GMS. All images were photographed at ×100 magnification. (E) The Pc organism burdens were further determined by quantifying 16S mitochondrial RNA, which further documented significant reduction of Pc organism burdens in the rats with PCP treated with nikkomycin Z (*P < 0.05, comparing nikkomycin treatment to saline-treated controls with PCP).

We further used a novel qPCR technique to analyze Pc life form–associated mRNAs to determine whether nikkomycin Z affects both the trophic and cyst forms of the organism. To accomplish this, we used qPCR quantitation of Pcint1 RNA, a gene expressed predominantly by trophic forms (42), and Pcran1 RNA, a gene expressed largely by the cyst forms of the organisms (41) (Figure E2). Using this approach, we observed that nikkomycin Z treatment of Pc-infected rats significantly reduced both forms of the organism, with observed reduction of the Pcint1 RNA associated with trophic forms by 94.1 (±2.7)% (P = 0.0004), and reduction of the Pcran1 RNA associated with cyst forms by 85.3 (±3.9)% (P = 0013). Due to the reported absence of chitin synthetic genes from the Pneumocystis species genome (5), these results suggest that nikkomycin Z can also potentially inhibit incorporation of GlcNAc into other cellular components, thereby impairing viability of the organisms. As such, these results support extended activity for nikkomycin Z as a potential agent in the treatment of PCP affecting both life cycle forms of the organism.

Discussion

Pj causes life-threatening pneumonia in immunocompromised patients, with an estimated 400,000 cases of PCP occurring annually worldwide and resulting mortalities as high as 80% in certain patient groups (64). PCP is considered the second most important invasive fungal infection in the world, only behind Cryptococcosis, and having far greater numbers than Aspergillosis and Histoplasmosis combined (64). With these concerns, additional research is required to understand the organism’s life cycle and to develop new agents for therapy. One area for potential therapeutic exploitation is the cell surface of Pneumocystis. For instance, targeting the Pneumocystis β-1,3 and β-1,6 glucan synthases, which generate major structural carbohydrates on the cell surface, may represent a novel therapeutic approach for this infection (3, 4, 7, 61, 65). Unfortunately, a comprehensive understanding of the Pneumocystis cell surface has not yet been elucidated. A recent keyword search of PubMed for “yeast cell wall” yielded over 7,000 publications compared with only 115 for “Pneumocystis cell wall.” In addition to β-glucans, most fungi also contain cell wall mannoproteins (mannans), which are important cell wall polysaccharides, typically containing an α-(1–6)–linked backbone and α(1, 2) and α(1–3) –linked branch chains (66). Some fungi, such as C. albicans, have more elaborate and extensive side-chain branching (67). In pathogenic fungi, mannans have been shown to be important for host cell binding, inhibition of macrophage cell uptake, and for masking inflammatory β-glucans (9). GlcNAc is essential for fungi to synthesize mannan structures and to properly localize β-glucans in the cell wall.

Fungal mannoprotein attachment to the cell membrane generally requires the presence of the chitobiose (GlcNAc)₂ core (68). However, in some fungi, such as C. albicans, genomic analyses suggest that there are over 100 putative GPI-anchored proteins, twice that of S. cerevisiae (69). These GPI-anchored proteins are linked at the carboxyl terminus through a phosphodiester linkage of phosphoethanolamine to a trimannosyl-nonacetylated glucosamine core (70). Mutations in these GPI-anchored proteins in yeast lead to profound growth defects, loss of cell wall integrity, increased sensitivity to antifungal agents, and reduction in virulence (69, 71). In Pneumocystis species, the presence of GPI-linked cell wall proteins is also important to the life cycle of the organism. For example, the MSG/gpA superfamily of Pneumocystis mannoproteins is bound to the cell surface by a GlcNAc-containing chitobiose core. It is noteworthy that MSG/gpA represents an astounding 3–6% of the genomes of Pneumocystis species (5), further suggesting the importance of these mannoproteins in the organism’s life cycle. We have also shown that Pc contains an additional GPI-anchored protein, termed PcPhr1p, a putative glycosidase required for proper cross-linking of β-1,3 and β-1,6-glucans that provides structural integrity to the Pneumocystis cell wall (72). Deletions in these carbohydrate cross-linking proteins have profound morphological and virulence defects in other fungi (73, 74). Taken together, these observations indicate that inhibiting GlcNAc biosynthetic pathways might provide a novel strategy for PCP therapy.

With this background in mind, we sought to evaluate the available evidence supporting the presence of a functional UDP-GlcNAc pathway in Pneumocystis species. As an initial step to determine the potential presence of GlcNAc, immunofluorescence microscopy studies were performed with WGA-FITC, a fluorescent reagent that binds terminal GlcNAc, chitobiose, and chitooligosaccharides. In addition, we used a recombinant CBD protein known to bind chitobiose and chito-oligosaccharides. Both approaches demonstrated significant binding to Pc, supporting the presence of GlcNAc-containing molecules in the cell wall of Pneumocystis organisms.

We further searched for the presence of GlcNAc biosynthetic–related genes in genomes of Pneumocystis species. With the advent of the newly completed Pc, Pj, and Pm databases (5), we performed a comprehensive search for other genes related to GlcNAc biosynthesis (Figure 3). Our in silico search verified the presence of a number of new GlcNAc synthase pathway genes in these Pneumocystis species. These include sequences for GFA1, GNA1, AGM1, and UAP1 in the Pj genome, with their presence being verified using PCR on Pj isolated from the bronchoalveolar lavage fluid of patients suffering from PCP. We were further able to verify the activity of the PmUap1 protein. Uap1p is a fungal UAP required in the last step of GlcNAc biosynthesis to generate the substrate, UDP-GlcNAc, for the major chitin synthases in yeast, and to provide the GlcNAc substrate for GPI linkages for fungal mannoproteins or functional GPI-anchored proteins (32, 75). PmUap1 protein was expressed in E. coli and used to confirm the appropriate biochemical activity measured by release of inorganic phosphate. Interestingly, PmUap1p could also cleave glucose-1-phosphate (Glc-1P), a phenomenon that has also been observed with Uap1p form S. cerevisiae (32). Of note, Uap1p proteins are essential for viability in some fungi, as evidenced by a recent report that a conditional mutant in A. fumigatus resulted in significant defects in cell survival due to altered cell wall carbohydrate composition and structure. Of further interest, the crystal structure of A. fumigatus Uap1p was elucidated and shown to possess substantial differences from the human enzyme (2). These observations, along with the abundant expression profile of Pmuap1 during active PCP in mice, suggest that the PjUap1 protein may also represent another viable target for PCP therapy.

To confirm the active deposition of GlcNAc into the cell wall, Pc cell wall preparations were analyzed and shown to incorporate GlcNAc into cell wall carbohydrate, as measured by WGA-HRP binding (53). Furthermore, this activity was shown to be greatest at 37°C, the usual temperature present in the lung environment. These results confirm that not only does Pneumocystis contain a complete GlcNAc pathway in its genome, but also that the fungus can use this pathway to generate GlcNAc residues and deposit them in the cell wall. Finally, we demonstrated that treatment of Pc cell wall membrane preparations with nikkomycin Z, an analog of UDP-GlcNAc, significantly reduced GlcNAc incorporation by Pc cell wall preparations.

These observations encourage us to use nikkomycin as a potential therapy in our rat models of PCP. Twice-daily gavage treatment with the agent significantly reduced the organism burden in the rodent PCP model. Interestingly, we observed that nikkomycin Z reduced both cyst and trophic forms of the organism. Recent evidence suggests that Pneumocystis may not contain an extensive GlcNAc-containing chitin matrix found in many other fungi (5). However, GlcNAc-containing chitobiose dimers are also important core molecules needed for the generation of mannoproteins and mannans (5). We, therefore, postulate that GlcNAc residues are needed to establish the core structures needed for certain Pneumocystis surface glycoconjugates, including potential MSG/gpA complexes. Regardless, our initial proof-of-concept studies provide evidence that nikkomycin Z may represent a viable treatment for PCP that impacts both forms of the organism. Additional full-dose and time-ranging experiments will be required to determine whether higher doses and longer treatment schedules of nikkomycin Z may completely eliminate Pneumocystis infections.

Because recent data suggest the absence of synthetic and degradative chitin proteins in Pneumocystis species (5), our data suggest that this antifungal agent might be targeting other Pneumocystis cell surface biosynthetic pathways. Because chitobiose is a core molecule for mannoprotein synthesis, nikkomycin may potentially impact processing and generation of MSG/gpA. In addition, one interesting Pneumocystis gene that we have cloned in the past is termed Pcgcs1 (accession no. AF338415), and functions as a glucosylceramide synthase (Gcs) homolog. Mammalian Gcs proteins are known to also use UDP-GlcNAc as a substrate (76). Gcs proteins typically catalyze the formation of glucosylceramide from ceramide and UDP-glucose (76). In pathogenic fungi, such as Cryptococcus neoformans and Penicillium digitatum, such ceramide conjugates are also required for proper growth and virulence (77, 78). Nikkomycin Z may also inhibit such pathways.

In summary, we have characterized the presence of a UDP-GlcNAc biosynthetic pathway in Pneumocystis species, as well as the presence of GlcNAc in the organism. Furthermore, treatment with nikkomycin Z inhibited the synthetic ability of cell preparations to incorporate GlcNAc into insoluble cell components, and this agent further significantly decreased the Pc organism burden in the lung during active infection. Due to the overall importance of the UDP-GlcNAc pathway in fungi, this initial characterization of this pathway in Pneumocystis should lead to exploiting these enzymes as potential targets for the treatment of PCP.