Characterization of Pneumocystis murina Bgl2, an Endo-β-1,3-Glucanase and Glucanosyltransferase

Abstract

Glucan is the major cell wall component of Pneumocystis cysts. In the current study, we have characterized Pneumocystis Bgl2 (EC 3.2.1.58), an enzyme with glucanosyltransferase and β-1,3 endoglucanase activity in other fungi. Pneumocystis murina, Pneumocystis carinii, and Pneumocystis jirovecii bgl2 complementary DNA sequences encode proteins of 437, 447, and 408 amino acids, respectively. Recombinant P. murina Bgl2 expressed in COS-1 cells demonstrated β-glucanase activity, as shown by degradation of the cell wall of Pneumocystis cysts. It also cleaved reduced laminaripentaose and transferred oligosaccharides, resulting in polymers of 6 and 7 glucan residues, demonstrating glucanosyltransferase activity. Surprisingly, confocal immunofluorescence analysis of P. murina–infected mouse lung sections using an antibody against recombinant Bgl2 showed that the native protein is localized primarily to the trophic form of Pneumocystis in both untreated mice and mice treated with caspofungin, an antifungal drug that inhibits β-1,3-glucan synthase. Thus, like other fungi, Bgl2 of Pneumocystis has both endoglucanase and glucanosyltransferase activities. Given that it is expressed primarily in trophic forms, further studies are needed to better understand its role in the biology of Pneumocystis.

We have demonstrated that Bgl2 of Pneumocystis murina is a β-1,3 endoglucanase that also has glucanosyltransferase activity. Surprisingly, Bgl2 localized to trophic forms, which lack β-1,3-glucan, but not to cysts, whose cell wall is composed primarily of β-1,3-glucan.

Pneumocystis is an opportunistic pathogen that causes life-threatening pneumonia in human immunodeficiency virus–infected as well as other immunocompromised patients [1–3]. Pneumocystis infecting different mammalian hosts is genetically divergent and represents unique species: P. jirovecii infects humans, P. carinii infects rats, and P. murina infects mice [4–7]. There are 2 major forms in the life cycle of Pneumocystis, trophic forms and cysts. The cyst cell wall has an electron lucent layer formed of β-1,3- and β-1,6-glucans, while β-glucans are not detected in trophic forms. Unlike virtually all other fungi, the Pneumocystis cyst wall does not contain chitin, and thus depends largely on β-glucans for maintaining cell wall integrity [8].

Fungal cell walls typically contain several β-glucanase enzymes that play a role in cell wall assembly and rearrangement, of which β-1,3 endoglucanases and glucanosyltransferases are among the most important [9, 10]. In Saccharomyces cerevisiae, Bgl2, a member of the glycoside hydrolase family 17, is a highly expressed cell wall protein [11–13] that has been shown to have both β-1,3 endoglucanase and glucanosyltransferase activities [12, 14]. Bgl2 is involved in cell wall metabolism in yeast, with a role in the branching of β-1,3-glucan through a β-1,6 linkage, as well as in the incorporation of mannoproteins into the cell wall [10, 15]. Homologs of Bgl2 have been characterized from other fungi, including Candida albicans and Aspergillus fumigatus [16–18].

Based on analysis of the recently published Pneumocystis genomes, all 3 Pneumocystis species noted above have genes encoding bgl2 homologues [8, 19]. The present study was thus undertaken to characterize the Bgl2 of Pneumocystis, given the importance of β-glucans in cell wall structure in Pneumocystis cysts.

MATERIALS AND METHODS

Pneumocystis Preparations

CD40L-knockout mice (B6;129S2-Tnfsf5^tm1Imx/J), which are highly susceptible to Pneumocystis infection [20], were obtained from the Jackson Laboratory and subsequently bred at the National Institutes of Health (NIH). Pneumocystis murina organisms were isolated from the lungs of heavily infected mice by Ficoll-Hypaque density gradient centrifugation [21]. All studies were performed under NIH Clinical Center Animal Care and Use Committee–approved protocols.

Recombinant Protein Expression

The complementary DNA (cDNA) sequence encoding 286 amino acids of P. murina Bgl2 (spanning amino acids 22–307; GenBank accession number XP_019613310) was optimized for bacterial expression, synthesized, and cloned into pET28b (+) expression vector, which includes a His tag (GenScript USA). The selected region corresponded to Bgl2 of S. cerevisiae, and excluded the additional carboxyl region found in Pneumocystis species. To facilitate expression, the signal peptide was excluded, and 2 amino acids, Met and Gly, were added to the N-terminal. For bacterial expression, the cDNA construct was transformed into Escherichia coli strain BL21 CodonPlus (DE3) RIL (Agilent Technologies). Recombinant protein was induced with 1 mM isopropyl-β-D-thiogalactopyranoside for 4 hours at 37°C and was purified using nickel-chelating resin (ProBond Purification System, Invitrogen). For mammalian expression, the bgl2 construct was cloned into pFLAG-CMV1 vector (Sigma-Aldrich) and transfected into COS-1 cells using Fugene 6 (Roche Applied Bioscience). After 48 hours, the cells were collected and lysed, followed by immunoblot analysis; vector alone with no insert served as the control.

A dectin-Fc construct encoding the extracellular domain of the murine dectin-1 receptor [22] and the murine immunoglobulin G1 (IgG1) Fc region in pSecTag 2 vector was obtained from Dr Chad Steele [23]. Recombinant dectin-Fc and recombinant Factor G, which is derived from the alpha subunit of clotting factor G of Limulus polyphemus, were expressed, and factor G was biotinylated, as previously described [24]; both proteins bind to β-1,3-glucan.

Anti-Bgl2 Antibodies

C57BL/6 mice were immunized with purified Bgl2 recombinant protein (~40 µg) in Freund adjuvant, complete (Sigma-Aldrich) followed by 4 booster injections at 2-week intervals with the same amount of Bgl2 in Freund adjuvant, incomplete (Sigma-Aldrich), after which time the animals were killed and serum was collected.

Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis and Immunoblot Analyses

Recombinant Bgl2 (rBgl2) protein preparations or partially purified P. murina organisms were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunoblotting. For recombinant protein, blots were probed with horseradish peroxidase–conjugated anti-His tag antibody (Santa Cruz Biotechnology); for P. murina–derived proteins, blots were probed with anti-Bgl2 antibody followed by horseradish peroxidase–conjugated antimouse immunoglobulin G (IgG) (Jackson ImmunoResearch Laboratories). Immunoreactive bands were visualized using one-step ultra TMB-blotting solution or SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific).

Assays for Bgl2 Functional Activity

rBgl2 was analyzed for potential endoglucanase and glucanosyltransferase activity. For the former, partially purified P. murina organisms were heat-fixed in 8-well slides, treated with trypsin (10 ng/µL, Promega) for 1.5 hours to expose the glucan of the cyst cell wall [24], and incubated with rBgl2 protein or negative control (vector alone, no insert) cell lysates for 3 hours at 37°C in 0.05 M sodium acetate buffer pH 5.0. Cysts were visualized by labeling their β-1,3-glucan with dectin-Fc followed by Alexa Fluor 488–conjugated anti-mouse IgG. Cysts in 20 random fields at ×400 magnification were counted as described previously [24]. Representative wells were tiled and z-stacked on a Zeiss LMS 700 with an EC-Plan-Neofluar 40x/NA 1.30 oil objective and 488 nm laser (Zeiss) using Zen 2012 version 14.0 SP5 software (Zeiss) for image capture and Imaris version 9.2.0 (Bitplane) for image processing.

For glucanosyltransferase activity, COS-1 cell lysates were incubated with 6 mM reduced laminaripentaose (1,3-β-D-laminaripentaitol; Megazyme) in 0.05 M sodium acetate buffer pH 5.0 at 50°C. Samples were collected every 20 minutes for 2 hours, and the reaction products in 1 µL were detected by Agilent quadrupole time of flight (Q-TOF, model 6540) liquid chromatography–mass spectrometry (LC-MS), which was coupled with an Agilent 1200 series LC system (Agilent). An Agilent ZORBAX Eclipse XDB-C18 rapid resolution HT 4.6 × 50 mm, 1.8 µm column was used with mobile phase A of 0.1% formic acid, 20 mM ammonium acetate, 2 mM ammonium fluoride, 2% acetonitrile, and 98% water, and mobile phase B of 0.1% formic acid, 100% methanol. The gradient was set to go from 0 to 15% B in 10 minutes.

Caspofungin Treatment Studies

Lungs from animals from a previously reported caspofungin treatment study were utilized to evaluate expression of Bgl2 by reverse-transcription quantitative polymerase chain reaction (RT-qPCR) [24]. Pneumocystis murina–infected mice were treated with 10 mg/kg caspofungin administered intraperitoneally daily for 9 days or for 21 days (5/7 days per week). Untreated infected mice served as controls.

Immunofluorescence and Confocal Microscopic Analysis

Pneumocystis murina–infected, fixed lung tissues from previous studies [24] were co-labeled with rabbit anti-Msg [25], anti-Bgl2, and recombinant factor G. Alexa Fluor 647–conjugated donkey antirabbit IgG, Alexa Fluor 594–conjugated donkey antimouse IgG, and Alexa Fluor 488–labeled streptavidin (Jackson ImmunoResearch) were used to detect anti-Msg, anti-Bgl2, and factor G, respectively; nuclei were stained with 4′,6-diamidino-2-phenylindole. Representative z-stacks were captured on the same Zeiss LMS 700 with an EC-Plan-Neofluar ×40/NA 1.30 oil objective and 405 nm, 488 nm, 555 nm, and 639 nm lasers using Zen for image capture and Imaris version 9.2.1 (Bitplane) for image processing.

RT-qPCR

Total RNA extracted from Pneumocystis-infected lungs of control or caspofungin-treated mice (RNeasy Mini Kit, Qiagen) was reverse transcribed using high-capacity RNA to cDNA kit (Applied Biosystems). cDNA was used as templates for real-time qPCR using power SYBR Green PCR Master mix (Bio-Rad) and ViiA 7 Real-Time PCR system according to the manufacturer’s instructions. Pneumocystis murina 18S ribosomal RNA (rRNA) was used as an endogenous control. The results were expressed as fold change in gene expression calculated using the relative quantification (ΔΔCT) method [26]. The primers used were GK10 pmurinabgl2 5′-CAGTAGGATCGGAGGTTCTTTATCG (corresponding to 374–398 of P. murina bgl2 mRNA, GenBank accession number XM_019757731) and GK10rev pmurinabgl2 5′- GTGTCCGCTGTTCCAAGTTTA (complementary to 480–500 of P. murina bgl2 messenger RNA [mRNA]) for bgl2, and GK3 pmurina18S 5′-GCTGTGGCCGAGTTAATAGCCCT (corresponding to 1948–1970 of P. murina 18S rRNA gene, GenBank accession number AY532651) and GK4 pmurina18S 5′- GCCATTTGGTCTGTGAACTGCATCCAT (complementary to 2053–2079 of P. murina 18S rRNA gene) for 18S rRNA.

Statistical Analysis

Endoglucanase activity and RT-qPCR results were compared using Student t test.

RESULTS

Characterization of bgl2 Gene From Pneumocystis

Given the importance of glucanosyltransferases in cell wall modifications in yeast, we undertook to characterize Bgl2, a potential glucanosyltransferase, from Pneumocystis. We identified bgl2 genomic and cDNA sequences of P. murina, P. carinii, and P. jirovecii from the Pneumocystis genome database [8, 19]. The alignment of genomic and cDNA sequences identified one intron in an identical location in all 3 species. The cDNA sequence of P. murina bgl2 showed 77% and 64% identity to that of P. carinii and P. jirovecii, respectively, while the latter 2 were 58% identical. Accession numbers for P. murina, P. carinii, and P. jirovecii bgl2 cDNA sequences in GenBank are XM_019757731, XM_018372126, and XM_018373049, respectively.

Deduced Amino Acid Sequences of Pneumocystis bgl2

The cDNA sequences of P. murina, P. carinii, and P. jirovecii bgl2 encode a protein containing 437, 447, and 408 amino acids, respectively, all of which include a signal peptide. Pneumocystis carinii Bgl2 shows 76% identity to P. murina Bgl2 and both show 51% identity to that of P. jirovecii.Figure 1 shows the alignment of the deduced amino acid sequences of Bgl2 from all 3 Pneumocystis species as well as those of Schizosaccharomyces pombe and S. cerevisiae. Pneumocystis Bgl2 exhibited 24%–27% identity to S. pombe Bgl2 and 23%–26% to that of S. cerevisiae. The previously identified catalytic sites as well as the sites required for transglycosylation are highly conserved across all species [10]. The Pneumocystis Bgl2 sequences contained 87–134 additional amino acid residues at the C-terminal when compared to Bgl2 from S. pombe and S. cerevisiae. This region contained 6 and 7 repeats of 12 amino acid residues for P. murina and P. carinii, respectively (Figure 1). Although similar repeats were not seen in the P. jirovecii protein, 3 amino acids (aspartic acid, asparagine, and serine) accounted for 59%–61% of the amino acid residues in the homologous, highly hydrophilic, regions in all 3 species. The presence of a signal peptide together with a hydrophobic tail region is consistent with a glycosylphosphatidylinositol-anchoring motif.

Figure 1.

Alignment of deduced amino acid sequences of Pneumocystis Bgl2 to that of yeasts. Bgl2 sequences of Pneumocystis carinii, Pneumocystis murina, and Pneumocystis jirovecii were aligned with Bgl2 of Saccharomyces cerevisiae and Schizosaccharomyces pombe using Clustal W. Identical or similar amino acid residues are boxed. The predicted signal peptide is double lined above the sequence. The catalytic active sites are denoted by asterisks, and the sites required for glucanosyltransferase activity are indicated by daggers. The 7 repeats in P. carinii sequence are marked by dashed lines above the sequences. GenBank accession numbers for P. murina, P. carinii, P. jirovecii, S. pombe, and S. cerevisiae bgl2 complementary DNA sequences are XM_019757731, XM_018372126, XM_018373049, CAB16200, and CAA97313, respectively.

Expression of Recombinant P. murina Bgl2 Protein

A region of P. murina Bgl2 (spanning amino acids 22–307) that included the catalytic sites and those needed for transglycosylation was expressed as a His tag fusion protein in bacteria. To facilitate expression, the signal peptide, hydrophilic carboxyl region, and hydrophobic tail were excluded. SDS-PAGE analysis of the bacterial extract expressing recombinant protein showed a prominent band of approximately 33 kDa protein when stained with Coomassie blue, which is the expected size of the recombinant protein; this band was not seen in the control (Figure 2A). The expressed protein showed immunoreactivity by immunoblot to anti-His tag antibody, while no immunoreactivity was seen with the control (Figure 2A). Because the protein from bacteria was largely insoluble, we also expressed it in COS cells, where the protein is expected to be conformationally correct and potentially functional (Figure 2B).

Figure 2.

Expression of Pneumocystis murina Bgl2 recombinant protein. A, Sodium dodecyl sulfate–polyacrylamide gel electrophoresis analysis of the bacterial extracts expressing recombinant protein showed a prominent Coomassie blue stained band of the expected size, ~33 kDa (indicated by the arrow, lane 2). This band was not seen when bacterial extracts of control vector with no insert were analyzed (lane 1).The bacterial extracts of the expressed protein showed immunoreactivity with this band when probed with anti-His tag antibody (lane 3), but no immunoreactivity was seen with the control vector (lane 4) when analyzed by immunoblotting. B, The COS-1 cell extracts expressing rBgl2 showed immunoreactivity by immunoblot with a ~33 kDa band when probed with anti-Bgl2 antibody (lane 1), whereas no immunoreactivity was observed in the cell extracts prepared from the control vector with no insert (lane 2).

Endoglucanase Activity of rBgl2 Protein

To determine if Bgl2 has endoglucanase activity as has been reported in yeast, we incubated COS-1 cell extracts containing rBgl2 with Pneumocystis cysts in which the cell wall glucans were exposed by trypsin treatment [24]. As seen in Figure 3A, the number of cysts as detected by labeling of β-1,3-glucan with dectin-Fc decreased following treatment with rBgl2 lysate compared to the negative control cell lysate. The number of cysts in 20 random microscopic fields was significantly lower in the rBgl2-treated slide (Bgl2 digest mean, 10.1 ± 1.1 standard deviation [SD]; control mean, 30.1 ± 1.9 SD; P < .001). Similar results were seen when staining with Gomori methenamine silver, suggesting that the cyst structure was disrupted with rBgl2 treatment.

Figure 3.

Bgl2 has both β-1,3-glucanase and glucanosyltransferase activity. A, Immunofluorescence labeling of β-1,3-glucan in Pneumocystis murina organisms following Bgl2 digestion demonstrates a decrease in cyst-associated β-1,3-glucan. Partially purified P. murina organisms were treated with trypsin to expose cell wall–associated glucans, followed by treatment with COS-1 cell lysate expressing rBgl2 (top panel) or control lysate (vector with no insert, bottom panel), prior to immunofluorescence labeling of β-1,3-glucan with dectin-Fc (to detect cyst forms). Fluorescence indicates binding to β-1,3-glucan. There are fewer cysts in Bgl2 digested sample compared to the control sample. B, Glucanosyltransferase activity of P. murina rBgl2. Extracts from COS-1 cells transfected with bgl2 construct were incubated with reduced laminaripentoase (rG5) oligosaccharide and the products formed at different time points were detected by quadrupole time of flight liquid chromatography–mass spectrometry. Extracts from COS-1 cells transfected with a no-insert vector served as negative control. rG2, rG3, rG6, and rG7 are the reduced sugars while G2, G3, G4 are nonreduced forms. The x-axis represents the time points (in minutes) that samples were tested; y-axis shows the peak area. The increase in the formation of rG6 and rG7 in the presence of rBgl2 when compared to control indicates that the expressed protein has glucanosyltransferase activity.

Glucanosyltransferase Activity of rBgl2 Protein

In yeasts, Bgl2 has also been shown to have glucanosyltransferase activity: It can cleave a disaccharide unit from a β-1,3-glucan polymer and transfer the remaining molecule to an acceptor (≥G4) via a β-1,6-linkage at the transfer site [9, 12, 14]. To determine if Pneumocystis Bgl2 has similar transferase activity, COS-1 cell lysates (Bgl2 or control) were incubated with reduced laminaripentaose (rG5) oligosaccharide and the samples were analyzed at different time points by Q-TOF LC-MS. With the Bgl2 lysate, the major products detected after 20 minutes were rG2 and G3, consistent with an endoglucanase function (Figure 3B); rG6 and rG7 increased with time, demonstrating transferase activity. Other products such as G2, rG3, and G4, were also detected in smaller amounts. None of these products was detected with the control lysate.

Immunofluorescent Analysis of Bgl2

Purified rBgl2 was used to immunize mice and generate hyperimmune serum, which reacted by immunoblot with rBgl2 (Figure 4A). This serum was used to examine the expression of Bgl2 in P. murina organisms. Immunoblot analysis of P. murina extracts showed immunoreactivity to an approximately 50-kDa band, which was blocked when the antibody was preincubated with rBgl2 (Figure 4B).

Figure 4.

Immunoreactivity of anti-rBgl2 antibody. A, rBgl2 expressed in bacteria was purified using nickel-chelating resin and used to immunize mice. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis analysis of the purified protein showed a band of ~33 kDa protein when stained with Coomassie blue (lane 1). This protein showed immunoreactivity to both anti-His tag antibody (lane 2) and mouse anti-Bgl2 antibody (lane 3) by immunoblot analysis. B, Protein extracts prepared from partially purified Pneumocystis murina organisms were subjected to immunoblot analysis using anti-Bgl2 antibody. An immunoreactive ~50-kDa band of expected size as indicated by the arrow was detected (lane 1). This immunoreactivity was blocked when the antibody was preincubated with recombinant protein (lane 2).

The localization of native Bgl2 in P. murina was analyzed by immunofluorescence microscopy. Pneumocystis murina–infected mouse lung sections were co-labeled with anti-Bgl2, anti-Msg, the most abundant surface protein of Pneumocystis, and factor G, which binds to β-1,3-glucan and is specific for the cyst form of Pneumocystis. Bgl2 immunoreactivity co-localized primarily to trophic forms and minimally to cysts (Figure 5).

Figure 5.

Immunofluorescence analysis of Bgl2 expression by Pneumocystis murina. Heavily infected lung sections from a CD40L knockout mouse were co-labeled with anti-Bgl2, anti-Msg, and factor G. Anti-Bgl2 (green, A), anti-Msg (red, B), factor G (magenta, C), and merge (D). There is co-localization of Bgl2 with Msg but not factor G, demonstrating that expression is associated with trophic forms but not cysts. The images represent maximum projection z-stacks obtained using a confocal microscope at ×400 magnification with zoom functionality. Bar is 10 microns.

Effects of In Vivo Caspofungin Treatment on the Expression of Bgl2

In C. albicans and Aspergillus niger, a subinhibitory dose of micafungin or caspofungin has been reported to increase the expression of Bgl2 [27, 28]. Thus, we looked at the expression of bgl2 mRNA in caspofungin-treated P. murina–infected mice. RT-qPCR analysis showed a significant increase (~2.3-fold; P = .001) in the expression of bgl2 mRNA in mice treated with caspofungin (10 mg/kg) for 9 days compared to untreated control mice (Figure 6A). A high dose of caspofungin (10 mg/kg) administered over a 21-day period nearly eliminated the cysts [24]; RT-qPCR analysis of these animals showed an approximately 45% decrease in the expression of bgl2 mRNA compared to controls (P = .004; Figure 6A). Immunolabeling a lung from the latter group with the anti-rBgl2 antibody demonstrated ongoing protein expression even in the setting of limited numbers of cysts, confirming that Bgl2 is expressed within trophic forms (Figure 6B).

Figure 6.

Expression of Bgl2 in caspofungin-treated mice. A, Reverse-transcription quantitative polymerase chain reaction (qPCR) analysis of bgl2 messenger RNA (mRNA) expression in Pneumocystis murina. RNA extracted from caspofungin-treated or untreated (control) mouse lung samples was reverse transcribed and bgl2 mRNA expression was analyzed by qPCR. Pneumocystis murina 18S ribosomal RNA was used as an endogenous control. The x-axis shows the drug treatment and the y-axis shows the fold change in bgl2 mRNA expression compared to control. The animals treated with caspofungin for 9 days showed an increase (~2.3-fold) in the expression of bgl2 mRNA compared to the control. Caspofungin treatment for 21 days resulted in an ~45% decrease in bgl2 mRNA expression compared to control. P values indicated above each pair were determined using Student t test. B, Immunofluorescence analysis of Bgl2 expression by P. murina. Infected lung sections from a CD40L knockout mouse treated with caspofungin for 21 days were co-labeled with anti-Bgl2, anti-Msg, and factor G. Anti-Bgl2 (green, A), anti-Msg (red, B), factor G (magenta, C), and merge (D). As seen in nontreated Pneumocystis-infected mouse lung, Bgl2 is associated with Msg but not factor G, demonstrating that expression is associated with trophic forms but not cysts. The images represent maximum projection z-stacks obtained using a confocal microscope at ×400 magnification with zoom functionality. Bar is 15 microns.

DISCUSSION

In this study we have shown that the bgl2 gene from Pneumocystis, which is present in the genomes of all 3 species that have been sequenced to date, encodes a protein with both endoglucanase and glucanosyltransferase activity. These proteins show homology to yeast Bgl2 protein, which belongs to glycoside hydrolase family 17. Surprisingly, in P. murina, Bgl2 protein is detected primarily in trophic forms by immunofluorescence analysis.

We identified the genomic and cDNA sequences of bgl2 of P. murina, P. carinii, and P. jirovecii from the genome sequence database [8, 19] based on high homology to yeast Bgl2. However, compared to S. cerevisiae, the Pneumocystis bgl2 sequences contain an extra 306–402 nucleotides at the 3′-end, which encode a hydrophilic region in all 3 Pneumocystis species. The function of this region, which is rich in aspartic acid, asparagine, and serine, is unknown. To verify that the 3′- end sequence is present in mRNA, we performed a Northern blot analysis using an oligonucleotide probe designed from the 3′-end region of P. carinii bgl2. A hybridization signal of approximately 1.5 kb was detected (data not shown), which is consistent with the predicted size of cDNA and is in agreement with whole genome RNA sequencing data [8, 19].

In yeast, bgl2 encodes an endo-β-1,3-glucanase, which is a highly expressed cell wall protein [15]. Yeast Bgl2 was first classified as an exoglucanase [29] and later designated as an endoglucanase [12]. It was subsequently demonstrated to also be a glucanosyltransferase that leads to branching of β-1,3-glucan chains through a β-1,6-linkage [12, 14, 16, 18]. Bgl2 is one of 2 enzymes critical for this branching activity. Double mutants with deletions in the genes encoding Bgl2 as well as Gas1, a β-1,3-glucanosyltransferase that belongs to the glycoside hydrolase family 72, led to loss of branched β-1,3-glucan, near total loss of β-1,6-glucans, and altered cell morphology [30].

In the current study, we were able to show that P. murina rBgl2 expressed in COS-1 cells has both β-1,3-glucanase and glucanosyltransferase activity. There was a decrease in the number of P. murina cysts following treatment with rBgl2, indicating it was able to degrade the cyst cell wall, which is largely composed of β-1,3-glucan. Yeast Bgl2 glucanosyltransferase activity has been described as involving 2 steps. In the first step, Bgl2 hydrolyzes β-1,3-glucan linkages of β-1,3-glucan oligosaccharides. The second step involves the transfer of the released oligosaccharide to the β-1,3-glucan acceptor molecules forming β-1,6-linkage at the transfer site [10, 14, 16]. We observed transferase activity as indicated by the formation of rG6 and rG7 from reduced laminaripentoase (rG5). It has previously been shown in yeast that the low concentration of substrate resulted only in the hydrolytic activity [14]. We also observed this effect when the concentration of the substrate was decreased 6-fold (data not shown).

By immunolabeling methods, we were able to demonstrate expression of Bgl2 by Pneumocystis. Immunoblot analysis of P. murina extract using an anti-rBgl2 antibody showed immunoreactivity toward an approximately 50-kDa protein band, which is the predicted size. Immunofluorescence analysis of P. murina–infected mouse lung tissue showed Bgl2 expression localized primarily to trophic forms, and only to a limited extent to cysts, as indicated by co-labeling with an anti-Msg antibody and factor G (Figure 5). This was surprising given that it plays a role in the metabolism of β-1,3-glucan, which is expressed exclusively in cysts but is absent in the trophic form. Pneumocystis may thus have adapted this enzyme to novel functions in the latter. Alternatively, it may be secreted and function extracellularly, perhaps to modulate biofilm formation, in which β-1,3-glucan plays a role in other fungi [31, 32].

In A. niger, Bgl2 has been reported to participate in spore release [33]. In C. albicans and A. niger, micafungin and caspofungin treatment resulted in an increase in Bgl2 expression [27, 28]. We similarly observed an increase in the expression of bgl2 mRNA when P. murina–infected mice were treated with caspofungin for 9 days, although 21-day treatment with caspofungin, which largely eliminates cysts, decreased the mRNA expression level by approximately 45%. Immunofluorescence analysis again demonstrated expression of Bgl2 protein within trophic forms of mice treated with caspofungin for 21 days (Figure 6B).

Thus, the current study has demonstrated the presence of a functional Bgl2 in Pneumocystis species, though our observation that it is expressed primarily by trophic forms does not fit with its predicted role in glucan metabolism given the absence of glucans in the wall of such forms. Further studies are needed to better understand the apparently unique role of Bgl2 in the biology and life-cycle of Pneumocystis.

Notes

Acknowledgments. We thank Rene Costello for providing animal care, and the staff of the Kansas State Veterinary Diagnostic Laboratory’s Histopathology Laboratory and Kansas State University College of Veterinary Medicine Confocal Core Facility for technical assistance.

Financial support. This project has been funded in whole or in part with federal funds from the Intramural Research Program of the US National Institutes of Health Clinical Center.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.