CPS1, a Homolog of the Streptococcus pneumoniae Type 3 ...

INFECTION AND IMMUNITY, July 2006, p. 3930–3938 0019-9567/06/$08.00⫹0 doi:10.1128/IAI.00089-06 Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 7

CPS1, a Homolog of the Streptococcus pneumoniae Type 3 Polysaccharide Synthase Gene, Is Important for the Pathobiology of Cryptococcus neoformans Y. C. Chang,1 A. Jong,2 S. Huang,2 P. Zerfas,3 and K. J. Kwon-Chung1* National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland1; Department of Pediatrics, Childrens Hospital Los Angeles, University of Southern California, Los Angeles, California2; and Division of Veterinary Resources, National Institutes of Health, Bethesda, Maryland3 Received 17 January 2006/Returned for modification 7 March 2006/Accepted 6 April 2006

The polysaccharide capsule is known to be the major factor required for the virulence of Cryptococcus neoformans. We have cloned and characterized a gene, designated CPS1, that encodes a protein containing a glycosyltransferase moiety and shares similarity with the type 3 polysaccharide synthase encoded by the cap3B gene of Streptococcus pneumoniae. Cps1p also shares similarity with hyaluronan synthase of higher eukaryotes. Deletion of the CPS1 gene from a serotype D strain of C. neoformans resulted in a slight reduction of the capsule size as observed by using an India ink preparation. The growth at 37°C was impaired, and the ability to associate with human brain endothelial cells in vitro was also significantly reduced by the deletion of CPS1. Using site-specific mutagenesis, we showed that the conserved glycosyltransferase domains are critical for the ability of the strain to grow at elevated temperatures. A hyaluronan enzyme-linked immunosorbent assay method demonstrated that CPS1 is important for the synthesis of hyaluronan or its related polysaccharides in C. neoformans. Comparisons between the wild-type and the cps1⌬ strains, using three different transmission electron microscopic methods, indicated that the CPS1 gene product is involved in the composition or maintenance of an electron-dense layer between the outer cell wall and the capsule. These and the virulence studies in a mouse model suggested that the CPS1 gene is important in the pathobiology of C. neoformans. Cryptococcus neoformans, an encapsulated yeast, causes life-threatening infections primarily in immunocompromised hosts, especially those with impaired cell-mediated immunity such as patients with human immunodeficiency virus infection (19). The most distinctive virulence factor in this fungus is the extracellular polysaccharide capsule composed mainly of glucuronoxylomannan (GXM) with galactoxylomannan and mannoproteins as the minor components (reviewed in reference 3). The GXM is a large polysaccharide with a molecular mass of over 1 MDa and consists of an unbranched mannose backbone substituted with various amounts of xylose, glucuronic acid, and O-acetyl residues. Recently, significant progress has been made in dissecting the biosynthetic pathways, as well as in characterization of the genes involved in capsule formation. The initial molecular approaches to prove the importance of capsule in cryptococcal virulence used complementational cloning and gene deletion studies. Four genes—CAP59, CAP64, CAP60, and CAP10— were identified by similar approaches to be required for the formation of capsule (6–8). The biochemical function and the role of these genes, however, have yet to be elucidated. One of these genes, CAP64, was found to have six homologues in the C. neoformans genome that play an important role in determining the position and linkage of xylose and/or O-acetyl residues on the mannose backbone of the GXM (25). In addition, several genes involved in the biosynthesis of polysaccharide

capsule precursors, as well as two genes potentially involved in its polymerization, have been characterized (1, 26, 27, 31). Recently completed genome sequencing of two serotype D strains of C. neoformans (22) has provided a wealth of information useful for dissecting the pathways of capsule formation. We have identified a gene, CPS1 (GenBank accession number AY063511) in the C. neoformans database, that shares similarity with the type 3 polysaccharide synthase encoding gene (cap3B) of Streptococcus pneumoniae (14). cap3B is reported to be essential for capsular lipopolysaccharide synthesis, as well as for S. pneumoniae virulence. The type 3 synthase from S. pneumoniae is a ␤-glycosyltransferase that functions in the assembly of the type 3 polysaccharide [3)-␤-D-glucuronic acid (GlcUA) (134)-␤-D-glucose (Glc)-(13] by a multicatalytic process (5). Polymer synthesis occurs via alternate additions of Glc and GlcUA onto the nonreducing end of the growing polysaccharide chain. The type 3 synthase also shares significant homology with other members of group 2 glycosyltransferase family, including the hyaluronan synthases from both prokaryotes and eukaryotes, the cellulose synthases from plants and bacteria, the chitin synthases from yeast, and the Nod factor synthases from Rhizobium (14, 17). Although the C. neoformans capsule is different from that of S. pneumoniae, it is possible that CPS1 is associated with the capsule formation and hence the virulence in C. neoformans. In the present study, we deleted CPS1 from serotype D strains of C. neoformans and studied its effect on pathobiology. Deletion of CPS1 resulted in alterations of several phenotypes important for virulence of C. neoformans, including temperature sensitivity, modification of ultrastructure between cell wall and capsule, and a reduced ability to associate with human brain

* Corresponding author. Mailing address: LCID, NIAID, Bldg. 10, Rm. 11C304, National Institutes of Health, Bethesda, MD 20892. Phone: (301) 496-1602. Fax: (301) 480-3240. E-mail: June_Kwon-Chung@nih .gov. 3930

IMPORTANCE OF CPS1 IN PATHOBIOLOGY OF CRYPTOCOCCUS

VOL. 74, 2006

TABLE 1. List of plasmids relevant to this study

3931

MATERIALS AND METHODS

cells) of each yeast strain. Survival of the mice was monitored to evaluate the virulence of each strain. TEM studies. The TEM method was similar to the protocol described by Kwon-Chung et al. with slight modification (20). After the sample was placed into 0.5% uranyl acetate in acetate Veronal buffer overnight at 4°C, the sample was washed and serially dehydrated in ethanol and then embedded in Eponate 12 resin (Ted Pella, Redding, CA). Sections (⬃80 nm thick) were obtained utilizing the Leica Ultracut-UCT ultramicrotome (Leica, Deerfield, IL), placed onto 400 mesh copper grids, and stained with saturated uranyl acetate in 50% methanol and then with lead citrate. The grids were viewed in the Philips 410 electron microscope (FEI, Hillsboro, OR) at 80 kV, and images were recorded on Kodak SO-163 film. Ruthenium red staining. C. neoformans cells from an fresh overnight culture were fixed with 2% glutaraldehyde and 1% osmium tetroxide, stained with 0.75% ruthenium red, embedded in 3% agar, and then cut into ultrathin sections (60 to 80 nm) for transmission electron microscopy (TEM) studies (12, 28), using an FEI Tecnai Twin 12 transmission electron microscope. Peroxidase staining of opsonins adhered to yeast cells. We have made slight modifications to the protocol described by Kozel et al. (18) to stain the opsonins that were adhered to C. neoformans. Yeast cells were fixed for 2 h at 37°C in a vacuum oven in a solution of 2% paraformaldehyde and 2% glutaraldehyde in Sorenson buffer (SB; 0.2 M sodium phosphate [pH 7.2]). After being washed with SB, the cells were incubated with normal human serum for 1 h at 37°C, washed,

Strains, media, and general methods. The strains used for the present study were as follows: B-4500FO2 (MAT␣ ura5), LP1 (MAT␣ ura5 ade2), JEC32 (MATa lys2), TAV14E (MAT␣ ura5 plb1), TYCC542 (MAT␣ ura5 cps1), TYCC645 (MAT␣ ura5 cps1⌬), C588 (MAT␣ ura5 cps1⌬ ⫹pYCC662), and C589 (MAT␣ ura5 cps1⌬ ⫹pCIP3). YEPD contained 1% yeast extract, 2% Bacto peptone and 2% dextrose. Minimal media (YNB) contained 6.7 g per liter yeast nitrogen base without amino acids (Difco) with 2% glucose (pH 7.0). To test possible defects in cell wall integrity, the yeast cells were grown on YEPD and YNB media containing Fluorescent Brightner 28 (calcofluor white, 0.5, 1.0, and 1.5 ␮g/ml), 0.01% sodium dodecyl sulfate, 0.5% Congo red, or caffeine (0.2, 0.5, and 1.0 mg/ml). The method used to determine cell fusion events during early stages of mating has been described previously (11). Briefly, the lys2 mutant, JEC32, was used as a tester strain to mate with cps1 and plb1 strains, as well as a wild-type strain, which carried the ura5 marker. After 6 h of incubation on V8 juice agar media, the cells were washed off, plated on minimal agar, and incubated for 3 to 4 days. To test the temperature-sensitive (TS) phenotype, the cultures were grown at 37°C or at 39°C, which reduced the level of background growth. Capsule size was measured in India ink preparations from YEPD-grown cells under oil immersion (magnification, ⫻100). Pictures were taken with a Zeiss Axioplan and AxioCam camera (Carl Zeiss, New York, NY) with Openlab software imaging system (Improvison, Lexington, MA). At least three different fields were randomly chosen and photographed. The capsule size (distance from the edge of the capsule to the cell wall) of 20 cells from each strain was measured and averaged. Cloning and characterization of the CPS1 gene. The plasmids used in the present study are listed in Table 1. Briefly, the genomic sequence of CPS1 was cloned by PCR with primers constructed on the basis of the sequence obtained from the database. To isolate cDNA clones, rapid amplification of cDNA ends (RACE) was performed in accordance with the protocol of the SMART RACE cDNA amplification kit (Clontech, Palo Alto, CA). A PCR-based site-directed mutagenesis method was used to introduce point mutations into the desired positions. The resulting clones were analyzed to confirm the accuracy of the sequence and subcloned into pPM8, a telomere-based plasmid (24). The plasmids were linearized before delivery into C. neoformans by biolistic transformation. Plasmid pYCC542 was a partial deletion construct (Fig. 1B) in which the SnaBI/BglI fragment of the CPS1 coding region was replaced with the 3.0-kb BamHI/XbaI fragment of the ADE2 gene from pYCC76 (6). The resulting construct retained 38 nucleotides at the 5⬘ coding region and 693 nucleotides at the 3⬘ coding region. Plasmid pYCC645 was a complete deletion construct (Fig. 1C) in which the Bsu36I/BssHII fragment of CPS1 was replaced with the 3.0-kb BamHI/EcoRI fragment of the ADE2 gene. The region between 34 nucleotides upstream of the first ATG and 14 nucleotides upstream of the stop codon of CPS1 was deleted in the resulting construct. The linearized disruption construct was then introduced into the LP1 strain by biolistic transformation (29). Transformants were screened to identify the cps1 deletant by colony PCR. Deletion of the CPS1 was confirmed by Southern blot hybridization. Isolation and analysis of genomic DNA were carried out as described previously (9). Animal model. For the animal experiments, 7-week-old female BALB/c mice were injected via the lateral tail vein with 0.2 ml of a suspension (3 ⫻ 106/ml

FIG. 1. Genomic map and Southern analysis. (A) Map of pYCC595, a wild-type CPS1. (B) Map of pYCC542, the cps1 partialdeletion construct. (C) Map of pYCC645, the cps1 complete-deletion construct. White box represents the coding region of CPS1. The dotted box represents the ADE2 gene, which is not drawn to scale for simplicity. The dashed line represents the PCR-amplified DNA sequence used as the CPS1 probe. Asterisks represent the restriction enzyme sites removed during plasmid construction. B, BglI; H, BssHII; N, NciI; R, EcoRV; S, SnaBI; U, Bsu36I. (D) DNA blot analysis. Genomic DNA was isolated, digested with EcoRV (I and II) or NciI (III), and fractionated. The DNA blot was hybridized with a probe containing the CPS1 gene (I and III) or a probe composed of the deleted cps1 sequence (II).

Plasmid

Gene and/or description

pYCC542 .....................................Partial deletion construct of CPS1 pYCC547 .....................................5⬘ RACE cDNA pYCC548 .....................................3⬘ RACE cDNA pYCC595 .....................................CPS1 genomic clone in pBluescript PYCC605 .....................................CPS1 genomic clone in pPM8 pYCC607 .....................................Q278R in pPM8 pYCC613 .....................................D135R in pPM8 pYCC616 .....................................D137R in pPM8 pYCC645 .....................................Complete deletion construct of CPS1 pYCC662 .....................................CPS1 genomic clone in pCIP3

microvascular endothelial cells (HBMEC) in vitro. These results suggest that CPS1 plays an important role in the pathobiology of C. neoformans.

3932

CHANG ET AL.

INFECT. IMMUN.

and incubated with peroxidase-labeled anti-human C3 antibody for an additional 2 h at 37°C (Cappel Laboratories, Cochranville, PA). Samples were washed and incubated with affinity-purified, peroxidase-labeled goat anti-mouse immunoglobulin G (H⫹L) for 30 min at 0°C. Cells were washed and suspended in a solution containing 0.1 ml of 3,3⬘-diaminobenzidine tetrahydrochloride (pH 7.0) per ml and 0.01% hydrogen peroxide in 0.05 M Tris-HCl (pH 7.0) for 15 min at 0°C. Cells were further washed in cold 0.1 M cacodylate, fixed with 1% osmium tetroxide for 30 min at 25°C, and incubated with 2% OsO4 in cacodylate buffer for 30 min at 25°C. Cells were then suspended in 10% warm gelatin and fixed in 0.8% glutaraldehyde for 5 min. Samples were dehydrated for 45 min each in 30, 50, and 70% dimethyl formamide, followed by two 30-min incubations in 100% dimethyl formamide. The gelatin block was sequentially embedded as follows: 30 min in a 1:1 dimethyl-propylene oxide, 30 min in 100% propylene oxide, overnight in a 1:1 propylene oxide-eponate and two 2-h incubations with 100% eponate at 25°C. The block was then polymerized for 24 h at 60°C. Ultrathin sections were cut with a glass knife on a Porter-Blum microtome and collected on Formvar-coated grids. Electron micrographs were obtained on Kodak electron image plates with a Hitachi model 125 electron microscope. Enzyme-linked immunosorbent assays (ELISAs) for hyaluronan. The Hyaluronan (HA)-ELISA Assay Kit obtained from Corgenix, Inc. (Denver, CO), was used to assay hyaluronan. The yeast cells (106 cells) in an exponential growth phase were incubated in individual wells at room temperature to trap the surface polysaccharide. After 30 min, the wells were washed with washing buffer according to the manufacturer’s instruction. A second solution containing hyaluronanbinding protein–horseradish peroxidase conjugate was added to the wells. After incubation, the wells were rinsed, and chromogenic substrate (TMB/H2O2) was added to develop a color reaction. The intensity of the resulting color was measured in optical density units with a spectrophotometer at 450 nm. The concentrations of hyaluronan were calculated by comparing the absorbance of the sample against a reference curve prepared from the reagent blank and hyaluronan reference solutions. Binding of yeast cells to HBMEC. Growth of HBMEC and the in vitro binding assay were as previously described (16) with minor modifications. Briefly, HBMEC were grown until confluence in collagen-coated 24-well tissue culture plates (Costar, Corning, NY). An inoculum of 106 yeast cells suspended in 0.5 ml of Ham’s F-12–M199 (1:1 [vol/vol])–5% heat-inactivated fetal bovine serum (experimental medium) was added to the HBMEC monolayer (multiplicity of infection of 10) for 2 h at 37°C. Yeast cells unattached to the HBMEC monolayer were then removed by washing with experimental medium four times. HBMEC were lysed with 0.5% Triton X-100, diluted, and plated onto blood or YEPD agar plates to determine the CFU that had been associated with HBMEC. The results are presented as the percent adhesion of inoculum: [(number of Cryptococcus recovered)/(number of Cryptococcus inoculated)] ⫻ 100%.

FIG. 2. Phenotypes of the cps1 deletant. (A) India ink staining showing a slight reduction of the capsule size in the cps1 strain (TYCC542) compared to the wild type (LP1). Cells were grown on YEPD. (B) Test of dominance of cps1. Cells of opposite mating types each carrying a different auxotrophic marker were crossed on V8 agar, washed off, and plated on minimal agar plates for 3 days at 30°C. 1, JEC32 ⫻ B-4500FO2 (wild type ⫻ wild type); 2, JEC32 ⫻ TAV14E (wild type ⫻ plb1); 3, JEC32 ⫻ TYCC542 (wild type ⫻ cps1).

RESULTS Characterization of the CPS1 gene. To further our understanding of the molecular basis of capsule formation in C. neoformans, we searched the C. neoformans genomic database to identify DNA sequences that share similarity to the genes involved in capsule synthesis of other microorganisms. We found a DNA sequence that shared homology with Streptococcus pneumoniae cap3B that encodes the type 3 polysaccharide synthase. The genomic and cDNA clones of this gene were obtained from C. neoformans by PCR, and their sequences were determined. The gene, designated CPS1 (GenBank accession number AY063511), contains six introns and encodes a putative protein with a calculated molecular mass of 51.5 kDa that contains a conserved glycosyltransferase motif. Like other members of the group 2 glycosyltransferase protein family, Cps1p contains the conserved DXD and QXXXRW domains at position D135 and Q278. Although the putative protein sequence of CPS1 showed only a 39.4% similarity to that of Cap3B, the two proteins showed similar patterns in the hydrophilicity plot (data not shown). To study the function of CPS1, the gene was disrupted in the strain LP1 by biolistic transformation method using the disruption construct, pYCC542 (Fig. 1B). Genomic DNAs from

the putative cps1 disruptants were isolated and analyzed by Southern hybridization to confirm the gene disruption. When the DNA containing most of the CPS1 sequence was used as a probe, the 4.1-kb fragment in the wild-type strain was found to be replaced by the 5.3-kb sequence in the transformant, TYCC542 (Fig. 1D, panel I). When the DNA sequence from the deleted portion of CPS1 was used as a probe, no signal was detected in TYCC542 (Fig. 1D, panel II), suggesting that the cps1 allele in TYCC542 was generated by replacing the wildtype CPS1 allele with the disruption construct pYCC542. Phenotypes of the cps1 disruptant. Since we hypothesized that CPS1 may be involved in the process of capsule formation in C. neoformans, we first examined the status of capsule in the cps1 disruptant by staining with India ink. The cells of TYCC542 tended to clump and showed a slight reduction in capsule size compared to the wild type (0.71 ⫾ 0.16 ␮m versus 0.61 ⫾ 0.09 ␮m) (Fig. 2A). Interestingly, the cps1 strains grew poorly at a temperature of 37°C or higher (see below). However, no difference between cps1 and wild-type cells was observed in the patterns of cell division, mode of nuclear migration from mother to daughter cells, as well as distribution of the actin patches and actin ring patterns (data not shown). In

VOL. 74, 2006


addition, cps1 mutants were not susceptible to a variety of inhibitors for cell wall synthesis or biogenesis such as caffeine, Congo red, sodium dodecyl sulfate, and calcofluor white. These observations suggested that the cell wall integrity was not affected by the deletion of CPS1. Furthermore, analysis of the F1 progeny derived from genetic crosses between a wildtype strain and TYCC542 confirmed that the phenotypes associated with cps1 resulted from the deletion of the CPS1 gene (data not shown). Unexpectedly, two lines of evidence suggested that the TS phenotype of TYCC542 was a dominant feature. First, it has been shown that when strains of MATa and MAT␣, each carrying a different auxotrophic marker, are mixed on a V8 plate, incubated for 6 h, and plated on minimal agar, colonies with short hyphal protrusions appear within 3 to 4 days (11). These colonies are believed to be the fusion product of MATa/MAT␣ in which the auxotrophic deficiency was complemented by each other, thereby initiating hyphal formation. Figure 2 shows that cells derived from a cross of two opposite mating type strains, each carrying a wild-type CPS1 but different auxotrophic markers (lys2 or ura5), were able to form colonies at both 30 and 37°C (Fig. 2B, left). Similarly, yeast cells from a cross between a MATa CPS1 lys2 strain and a MAT␣ ura5 strain carrying the deletion of an irrelevant gene such as plb1 produced numerous colonies on minimal agar at both 30 and 37°C (Fig. 2B middle). In contrast, yeast cells derived from a cross between TYCC542 (MAT␣ cps1 ura5) and a strain of MATa CPS1 lys2 produced only a few sizeable colonies on minimal agar plate at 37°C (Fig. 2B, right). At 30°C, however, yeast cells from the same batch of mating produced abundant colonies on minimal agar. The ratios of colonies formed at 37°C versus those formed at 30°C, expressed in percent values, were 46% (wild type ⫻ wild type), 38% (wild type ⫻ plb1), and 0.28% (wild type cps1), respectively. It is not clear why the wild-type cells show less than 50% of colony count at 37°C compared to that observed at 30°C, but it is clear that the cps1 ⫻ wild type mating produced a drastically reduced number of colonies under these conditions. Second, when TYCC542 was transformed with the wild-type CPS1 gene to complement the TS phenotype, the resulting transformants grew normally at 30°C but failed to grow at 39°C (Fig. 3A, left panels). Since the cps1 allele in TYCC542 contains 38 nucleotides at the 5⬘ coding region and 693 nucleotides at the 3⬘ coding region (Fig. 1B), it is possible that the dominant TS phenotype of TYCC542 is due to the ability of the mutant to synthesize an aberrant protein from the incomplete gene sequence of the construct that interferes with other cellular functions. To test this possibility, the partial deletion construct, pYCC542, was transformed into the wild-type stain, and the resulting transformants containing an ectopic copy of pYCC542 DNA were tested for temperature sensitivity. These transformants grew normally at 30°C but displayed the TS phenotype at 37°C (data not shown). To construct a complete cps1 deletant, we obtained a new cps1 strain, TYCC645, in which most of the CPS1 coding region was deleted (see Fig. 1C and Materials and Methods). Figure 1D, panel III, shows that the 2.3-kb fragment in B-4500 was replaced by a 3.6-kb fragment in TYCC645, indicating that the cps1 gene had been deleted. This new strain also grew poorly at elevated temperatures, similar to TYCC542. However, when TYCC645 was transformed with the wild-type

3933

FIG. 3. Temperature sensitivity of cps1 mutants. (A) The TS phenotype in TYCC542 appears to be dominant. The cps1 strains were transformed with either the wild-type CPS1 gene or vector, the resulting transformants were serially diluted, spotted on agar media, and incubated at either 30 or 39°C. (B) Functional test of the conserved domains. The conserved sequence of CPS1 was mutated as indicated and transformed into TYCC645. The resulting transformants were serially diluted, spotted on the media, and incubated at either 30 or 37°C.

CPS1 gene, the resulting transformants were able to grow at 39°C (Fig. 3B, right panels). In addition, when TYCC645 was mated with a MATa strain carrying a different auxotrophic marker as described above, the resulting mating mixture was able to form colonies at an elevated temperature on minimal medium similar to the wild-type controls (data not shown). These results suggested that deletion of CPS1 from wild-type strain resulted in a TS phenotype and that the dominant TS phenotype of TYCC542 was due to the sequences remaining in the partially deleted cps1 allele. Cps1p contains the conserved DXD and QXXXRW domains as other members of the group 2 glycosyltransferase protein family. To test the importance of these domains for the function of Cps1p, several plasmids each containing point mutations at the DXD or QXXXRW domain were generated. These constructs were transformed into strain TYCC645, and the resulting transformants were tested for their ability to grow at 30 and 37°C. Figure 3B shows that constructs containing a mutation at D135, D137, or Q278 were not able to complement the TS phenotype of TYCC645, whereas the wild-type CPS1 gene was able to rescue the TS phenotype. These results suggested that the conserved DXD and QXXXRW domains play an important role in the function of Cps1p to facilitate the growth of C. neoformans at an elevated temperature. Ultrastructural alterations in the cps1 deletants. Although Cps1p of C. neoformans shares similarity with the type 3 polysaccharide synthase of S. pneumoniae, the chemical structure

3934

CHANG ET AL.

INFECT. IMMUN.

FIG. 4. TEM focusing on the morphology of cell wall/capsule. C. neoformans cells from overnight culture were fixed and sectioned for transmission electron micrograph. A portion of the cell wall/capsule area is shown. A, B-4500FO2 (wild type); B, TYCC645 (cps1⌬); C, C588 (TYCC645⫹CPS1); D, C589 (TYCC645⫹vector). A bracket with an arrow indicates the electron-dense layer as described in the text. CAP, capsule; CW, cell wall. Bar, 0.1 ␮m.

of the capsule in the two organisms is considerably different. Therefore, we considered the possibility that the slight reduction in capsule size of the cps1 mutant observed with an India ink preparation could be due to modifications of cellular components other than those of the capsule. We used TEM methods to study the possible differences between the cps1⌬ strain and the wild type at the ultrastructural level. Cells were harvested from overnight cultures, and ultrathin sections were prepared for the visualization by TEM. No clear difference was observed between the wild type and the cps1 mutant in terms of cellular content and cytoplasmic organelles. However, the thickness of an electron-dense layer between the outer cell wall and capsule was clearly reduced in the cps1 deletant,

TYCC645, compared to the wild type (Fig. 4A and B, bracket with arrow). Complementation of the cps1 deletion with the wild-type CPS1 gene resulted in the reconstitution of the electron-dense layer, whereas the transformant containing only the vector did not (Fig. 4C and D, bracket with arrow). The differences between cps1 and wild-type strain was also clearly demonstrated when ruthenium red was used to stain the cells. Figure 5A shows that ruthenium red stain produced a thick, dense furry layer outside the cell wall in the wild-type strain (white arrow). This electron-dense layer was absent in TYCC645 and instead many spike-like filamentous structures, likely of capsule, were observed on its surface (Fig. 5B). Since ruthenium red was used to stain external polyanionic molecule,

VOL. 74, 2006


3935

FIG. 5. TEM studies with ruthenium red-stained cells (A and B) or cells treated with human serum and anti-C3 antibody-HRP conjugate (C and D). A and C, B-4500FO2 (wild type); B and D, TYCC645 (cps1⌬). The white arrow indicates the electron-dense layer, and the black arrow indicates the zone of exclusion as described in the text. CW, cell wall. Bar, 0.1 ␮m.

the modification of the electron dense layer in the cps1 deletant is likely due to the alteration of polyanionic molecules. To further explore the ultrastructural alterations in the cps1 strain, we opsonized yeast cells of both wild-type and cps1 strains with human serum. The location of the bound C3 fragments was determined by incubation of the serum-treated yeast cells with an anti-human C3 antibody-horseradish peroxidase (HRP) conjugate. In the wild-type cells, C3 was found to have been deposited mostly on the capsule, leaving a white rim between the cell wall and capsule as shown in Fig. 5C (black arrow). In the cps1 strain, however, the white rim indicative of lack of C3 binding was drastically reduced (Fig. 5D). When the cps1 deletant was complemented, its ultrastructure was similar to the wild type (data not shown). These ultrastructural modifications observed in the strain with cps1⌬ indicated that the CPS1 gene product is related to the ultrastructure in the area between the cell wall and capsule. CPS1 and hyaluronan. We have performed microarray experiments with C. neoformans treated and untreated HBMEC and found that the gene encoding CD44, a receptor of hyaluronan, and its related genes were differentially expressed (A. Jong, unpublished observation). Hyaluronan is a high-molecular-mass (1,000 to 5,000 kDa) anionic polysaccharide composed of ␤-1,4-linked repeating disaccharides of glucuronic

acid and ␤-1,3-linked N-acetylglucosamine (21). Since Cps1p shares similarity with the eukaryotic hyaluronan synthases, we used a hyaluronic acid-binding protein and ELISA to examine the possible link between hyaluronan or hyaluronan-like molecules with CPS1 in C. neoformans. Figure 6A shows that the amount of hyaluronan-like molecules was significantly reduced in the cps1⌬ strain compared to the parental wild-type strain. Complementation of the cps1⌬ strain with the wild-type CPS1 gene restored the amount of hyaluronan-like molecules to wild-type levels. CPS1 and association of yeast cells with HBMEC. It has been shown that exposure of HBMEC to C. neoformans triggered the formation of microvillus-like membrane protrusions within 15 to 30 min and that C. neoformans adhered to and was internalized by the HBMEC (10). Recent observations indicated that hyaluronan may interact with cell surface receptors and influence cellular proliferation, differentiation, migration, and adhesion of eukaryotic cells (30). In group A Streptococcus, the binding of hyaluronan to CD44 triggers a signaling cascade required for the formation of membrane projection that promotes penetration of the bacteria into the tissue (13). Although the mechanism for the ability of C. neoformans to associate with HBMEC is not clear, it is possible that modification of the hyaluronan-like molecule in C. neoformans may

3936

CHANG ET AL.

FIG. 6. (A) HA-ELISA test. The hyaluronan or related molecules from the surface of C. neoformans cells were detected by using the HA-ELISA test kit. (B) Association of yeast cells with HBMEC. An in vitro binding assay was performed to examine the ability of C. neoformans cells to associate with HBMEC. The CFU of C. neoformans cells that had been adhered to HBMEC in vitro were analyzed. Columns (A and B): 1, LP1 (wild type); 2, TYCC645 (cps1⌬); 3, C588 (TYCC645⫹CPS1); 4, C589 (TYCC645⫹vector). (C) Virulence studies. Mice were injected with yeast strains via the tail vein, and mortality was monitored. B-4500FO2, wild type; TYCC645, cps1⌬ mutant.

influence the ability of yeast cell to associate with HBMEC. An in vitro binding assay was performed to examine the effect of CPS1 deletion on the association of yeast cells with HBMEC at 30 and 37°C. At both temperatures, the binding activity of TYCC645 was about one-third compared to that of the parental strain, LP1 (Fig. 6B, bar 1 versus bar 2). The binding efficiency of TYCC645 was restored to wild-type levels when the mutant was complemented with CPS1, whereas the binding efficiency of TYCC645 remained low when the mutant was transformed with the vector alone (Fig. 6B, bars 3 and 4). These results suggested that CPS1 is required for C. neoformans yeast cells to effectively associate with HBMEC. CPS1 and virulence. We compared virulence of the cps1 deletant with the wild-type strain. Since the ability to grow at 37°C is a requirement for C. neoformans to establish a systemic infection, the cps1 deletant would predictably be avirulent. As expected, all mice challenged with the wild-type strain succumbed to the infection by 57 days postinfection, whereas the mice challenged with the cps1 deletant, TYCC645, all survived more than 100 days postinfection (Fig. 6C). DISCUSSION Glycosyltransferases are a group of enzymes that are involved in the biosynthesis of disaccharides, oligosaccharides, and polysaccharides. These enzymes catalyze the glycosidic bond formation by transferring the sugar moieties from acti-

INFECT. IMMUN.

vated donor molecules to specific acceptor molecules. According to the nucleotide diphospho-sugar, nucleotide monophospho-sugar, and sugar phosphates (EC:2.4.1), glycotransferases have been classified into distinct sequence-based families (4). Cps1p shares sequence similarity with a number of group 2 glycosyltransferase proteins. This family contains proteins from bacteria, fungi, plants, and animals that includes hyaluronan synthases, chitin synthases, cellulose synthases, dolicholphosphate mannosyltransferase, and type 3 capsule polysaccharide synthase. Our mutagenesis studies on the conserved DXD and QXXXRW motifs in Cps1p shows that these motifs, as with other glycosyltransferases, are important for Cps1p function. However, the substrates and enzymatic product of Cps1p are yet to be identified. The results of three TEM methods indicated that the product of the CPS1 gene is involved in the composition or maintenance of an electron-dense layer between the cell wall and capsule, although the exact nature of this layer is not clear. In TEM studies, ruthenium red-stained wild-type cells showed a positively stained layer surrounding the cell wall and the equivalent area in the other TEM studies appeared as the zone of exclusion when tested with anti-human C3-HRP staining. These two features were clearly altered in the cps1 deletant. It has been shown that incubation of encapsulated cryptococci in normal human serum leads to the deposition of C3 complement at the surface, as well as in the interior part of the capsule, leaving a prominent zone of exclusion between the cell wall and the capsule (18). The nature of the component that forms the exclusion zone is not clear. It has been suggested that molecular sieving by the cryptococcal capsule may have functional consequences for the interaction of the yeast with C3, and it might preclude penetration of the C3 to the interiormost part of the capsule (15). TEM data suggest that CPS1 function is associated with the preservation of the C3 exclusion zone. It is possible that the exclusion zone is not formed simply by the physical barrier of dense interior capsule as previously suggested but may be related directly or indirectly to the product of Cps1p. Alternatively, the ultrastructural changes may have resulted from a modification in the chemical structure of the capsule in the strains with cps1 deleted that caused the reduction of the physical barrier of capsule. Although the cps1 deletants still strongly reacted with the commercially available anticapsule factor sera used for serotyping (data not shown), it would be interesting to investigate its capsule chemotype in detail by using nuclear magnetic resonance imaging. HA-ELISA results suggested that Cps1p is involved in the synthesis of hyaluronan or its related molecules. Hyaluronan is not only an important structural component in vertebrates but also plays important roles in many biological processes. Recent observations indicated that the association of hyaluronan with cell surface receptors such as CD44 and RHAMM (named for receptor for hyaluronan-mediated motility) influences cellular proliferation, differentiation, migration, and adhesion of eukaryotic cells (30). The HA-ELISA kit used in the present study uses microwells coated with a highly specific hyaluronic acid-binding protein (HABP) that is derived from bovine cartilage to capture hyaluronan. Hyaluronan can then be detected and measured by an enzyme-conjugated version of HABP. Although it is not clear whether the positive signal detected by HA-ELISA kit in the samples prepared from C. neoformans is


VOL. 74, 2006

hyaluronan or compounds related to hyaluronan, the deletion of CPS1 resulted in the significant reduction in the signal intensity. Purified HAS1, a mouse hyaluronan synthase, is capable of synthesizing both hyaluronan and chitin oligomers in vitro (32). Lee and Spicer (21a) proposed that higher eukaryotic hyaluronan synthase evolved from chitin or cellulose synthase through the addition of the ␤-1,3 glycosyltransferase activity to a preexisting ␤-1,4 glycosyltransferase enzyme and that the ability to synthesize hyaluronan is a recent event in the evolution of matazoan organisms. Furthermore, the alternative synthesis of hyaluronan or chitin by chlorovirus suggests that hyaluronan and chitin are structurally and functionally compatible in chlorovirus infection (23). The HA-ELISA kit is specific for HA according to the manufacturer and does not react with other fungal oligomers such as chitin (data not shown). It is possible that Cps1p is responsible for synthesizing hyaluronan-like molecules that reacted positively with the HAELISA kit. In addition, it is also possible that the hyaluronanlike molecule may cause differential gene expression of CD44 and its related genes in HBMEC upon treatment with C. neoformans in the microarray experiments (data not shown). The TS phenotype for cps1 is an interesting observation, especially the dominant phenotype of the partial cps1 deletant. The cause of the dominance of cps1 disruptant is not clear. It is possible that multimerization of products from the partially deleted cps1 or other yet-to-be-identified mechanisms may cause the dominance. Aside from the minor morphological changes in cps1 observed by light microscopy and TEM, we did not detect any changes in the cell wall integrity, as judged by its susceptibility to cell wall biogenesis-inhibiting drugs or the cytoskeletal arrangement patterns. It is not clear why deletion of CPS1 affects cell growth at elevated temperatures. Although the cps1 deletants were avirulent as expected due to their poor growth at 37°C, it is intriguing to observe the significant reduction in the association of the mutant with HBMEC. Cryptococcal meningoencephalitis develops as a result of hematogenous dissemination of inhaled C. neoformans from the lung to the brain. Before entering the brain, the yeast has to cross the bloodbrain barrier. C. neoformans is able to bind to HBMEC, the major component of the blood-brain barrier, in vitro but the factors necessary for their binding are unknown. The HBMEC culture system offers a good in vitro model to study the mechanisms involved in this process (12, 28). The change in the cell surface charge in cps1 strains implicated by ruthenium staining patterns or the change in the detectable amount of hyaluronan-like molecules or other yet-to-be-identified functional aspect of Cps1p may contribute to the reduction in the efficiency of yeast cells to associate with HBMEC. Our results clearly show that CPS1 is important in the pathobiology of C. neoformans.

ACKNOWLEDGMENTS We thank T. McPadden for technical assistance and A. Varma for reading the manuscript. This study was supported by funds from the intramural program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

3937

REFERENCES 1. Bar-Peled, M., C. L. Griffith, and T. L. Doering. 2001. Functional cloning and characterization of a UDP-glucuronic acid decarboxylase: the pathogenic fungus Cryptococcus neoformans elucidates UDP-xylose synthesis. Proc. Natl. Acad. Sci. USA 98:12003–12008. 2. Reference deleted. 3. Bose, I., A. J. Reese, J. J. Ory, G. Janbon, and T. L. Doering. 2003. A yeast under cover: the capsule of Cryptococcus neoformans. Eukaryot. Cell 2:655– 663. 4. Campbell, J. A., G. J. Davies, V. Bulone, and B. Henrissat. 1997. A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem. J. 326:929–939. 5. Cartee, R. T., W. T. Forsee, J. S. Schutzbach, and J. Yother. 2000. Mechanism of type 3 capsular polysaccharide synthesis in Streptococcus pneumoniae. J. Biol. Chem. 275:3907–3914. 6. Chang, Y. C., and K. J. Kwon-Chung. 1998. Isolation of the third capsuleassociated gene, CAP60, required for virulence in Cryptococcus neoformans. Infect. Immun. 66:2230–2236. 7. Chang, Y. C., and K. J. Kwon-Chung. 1999. Isolation, characterization, and localization of a capsule-associated gene, CAP10, of Cryptococcus neoformans. J. Bacteriol. 181:5636–5643. 8. Chang, Y. C., L. A. Penoyer, and K. J. Kwon-Chung. 1996. The second capsule gene of Cryptococcus neoformans, CAP64, is essential for virulence. Infect. Immun. 64:1977–1983. 9. Chang, Y. C., L. A. Penoyer, and K. J. Kwon-Chung. 2001. The second STE12 homologue of Cryptococcus neoformans is MATa-specific and plays an important role in virulence. Proc. Natl. Acad. Sci. USA 98:3258–3263. 10. Chang, Y. C., M. F. Stins, M. J. McCaffery, G. F. Miller, D. R. Pare, T. Dam, M. Paul-Satyaseela, K. S. Kim, and K. J. Kwon-Chung. 2004. Cryptococcal yeast cells invade the central nervous system via transcellular penetration of the blood-brain barrier. Infect. Immun. 72:4985–4995. 11. Chang, Y. C., B. L. Wickes, G. F. Miller, L. A. Penoyer, and K. J. KwonChung. 2000. Cryptococcus neoformans STE12␣ regulates virulence but is not essential for mating. J. Exp. Med. 191:871–882. 12. Chen, S. H. M., M. F. Stins, S. H. Huang, Y. H. Chen, K. J. Kwon-Chung, Y. C. Chang, K. S. Kim, K. Suzuki, and A. Y. Jong. 2003. Cryptococcus neoformans induces alterations in cytoskeleton of human brain microvascular endothelial cells. J. Med. Microbiol. 52:961–970. 13. Cywes, C., and M. R. Wessels. 2001. Group A Streptococcus tissue invasion by CD44-mediated cell signalling. Nature 414:648–652. 14. Dillard, J. P., M. W. Vandersea, and J. Yother. 1995. Characterization of the cassette containing genes for type 3 capsular polysaccharide biosynthesis in Streptococcus pneumoniae. J. Exp. Med. 181:973–983. 15. Gates, M. A., P. Thorkildson, and T. R. Kozel. 2004. Molecular architecture of the Cryptococcus neoformans capsule. Mol. Microbiol. 52:13–24. 16. Jong, A. Y., M. F. Stins, S.-H. Huang, S. H. M. Chen, and K. S. Kim. 2001. Traversal of Candida albicans across human blood-brain barrier in vitro. Infect. Immun. 69:4536–4544. 17. Keenleyside, W. J., and C. Whitfield. 1996. A novel pathway for O-polysaccharide biosynthesis in Salmonella enterica serovar Borreze. J. Biol. Chem. 271:28581–28592. 18. Kozel, T. R., M. A. Wilson, G. S. Pfrommer, and A. M. Schlageter. 1989. Activation and binding of opsonic fragments of C3 on encapsulated Cryptococcus neoformans by using an alternative complement pathway reconstituted from six isolated proteins. Infect. Immun. 57:1922–1927. 19. Kwon-Chung, K. J., and J. E. Bennett. 1992. Medical mycology. Lea & Febiger, Philadelphia, Pa. 20. Kwon-Chung, K. J., I. Polacheck, and T. J. Popkin. 1982. Melanin-lacking mutants of Cryptococcus neoformans and their virulence for mice. J. Bacteriol. 150:1414–1421. 21. Laurent, T. C., and J. R. Fraser. 1992. Hyaluronan. FASEB J. 6:2397–2404. 21a.Lee, J. Y., and A. P. Spicer. 2000. Hyaluronan: a multifunctional, megaDalton, stealth molecule. Curr. Opin. Cell Biol. 12:581–586. 22. Loftus, B. J., E. Fung, P. Roncaglia, D. Rowley, P. Amedeo, D. Bruno, J. Vamathevan, M. Miranda, I. J. Anderson, J. A. Fraser, J. E. Allen, I. E. Bosdet, M. R. Brent, R. Chiu, T. L. Doering, M. J. Donlin, C. A. D’Souza, D. S. Fox, V. Grinberg, J. Fu, M. Fukushima, B. J. Haas, J. C. Huang, G. Janbon, S. J. Jones, H. L. Koo, M. I. Krzywinski, J. K. Kwon-Chung, K. B. Lengeler, R. Maiti, M. A. Marra, R. E. Marra, C. A. Mathewson, T. G. Mitchell, M. Pertea, F. R. Riggs, S. L. Salzberg, J. E. Schein, A. Shvartsbeyn, H. Shin, M. Shumway, C. A. Specht, B. B. Suh, A. Tenney, T. R. Utterback, B. L. Wickes, J. R. Wortman, N. H. Wye, J. W. Kronstad, J. K. Lodge, J. Heitman, R. W. Davis, C. M. Fraser, and R. W. Hyman. 2005. The genome of the basidiomycetous yeast and human pathogen Cryptococcus neoformans. Science 307:1321–1324. 23. Mohammed Ali, A. M., T. Kawasaki, and T. Yamada. 2005. Genetic rearrangements on the Chlorovirus genome that switch between hyaluronan synthesis and chitin synthesis. Virology 342:102–110. 24. Mondon, P., R. Petter, G. Amalfitano, R. Luzzati, E. Concia, I. Polacheck, and K. J. Kwon-Chung. 1999. Heteroresistance to fluconazole and voricon-

3938

25. 26. 27. 28.

CHANG ET AL.

azole in Cryptococcus neoformans. Antimicrob. Agents Chemother. 43:1856– 1861. Moyrand, F., Y. C. Chang, U. Himmelreich, K. J. Kwon-Chung, and G. Janbon. 2004. Cas3p belongs to a seven-member family of capsule structure designer proteins. Eukaryot. Cell 3:1513–1524. Moyrand, F., B. Klaproth, U. Himmelreich, F. Dromer, and G. Janbon. 2002. Isolation and characterization of capsule structure mutant strains of Cryptococcus neoformans. Mol. Microbiol. 45:837–849. Sommer, U., H. Liu, and T. L. Doering. 2003. An alpha-1,3-mannosyltransferase of Cryptococcus neoformans. J. Biol. Chem. 278:47724–47730. Stins, M. F., F. Gilles, and K. S. Kim. 1997. Selective expression of adhesion molecules on human brain microvascular endothelial cells. J. Neuroimmunol. 76:81–90.

Editor: A. Casadevall

INFECT. IMMUN. 29. Toffaletti, D. L., T. H. Rude, S. A. Johnston, D. T. Durack, and J. R. Perfect. 1993. Gene transfer in Cryptococcus neoformans by use of biolistic delivery of DNA. J. Bacteriol. 175:1405–1411. 30. Toole, B. P. 1990. Hyaluronan and its binding proteins, the hyaladherins. Curr. Opin. Cell Biol. 2:839–844. 31. Wills, E. A., I. S. Roberts, M. Del Poeta, J. Rivera, A. Casadevall, G. M. Cox, and J. R. Perfect. 2001. Identification and characterization of the Cryptococcus neoformans phosphomannose isomerase-encoding gene, MAN1, and its impact on pathogenicity. Mol. Microbiol. 40:610–620. 32. Yoshida, M., N. Itano, Y. Yamada, and K. Kimata. 2000. In vitro synthesis of hyaluronan by a single protein derived from mouse HAS1 gene and characterization of amino acid residues essential for the activity. J. Biol. Chem. 275:497–506.