WdChs2p, a Class I Chitin Synthase, Together with WdChs3p (Class III) Contributes to Virulence in Wangiella (Exophiala) dermatitidis

Abstract

The chitin synthase structural gene WdCHS2 was isolated by screening a subgenomic DNA library of Wangiella dermatitidis by using a 0.6-kb PCR product of the gene as a probe. The nucleotide sequence revealed a 2,784-bp open reading frame, which encoded 928 amino acids, with a 59-bp intron near its 5′ end. Derived protein sequences showed highest amino acid identities with those derived from the CiCHS1 gene of Coccidioides immitis and the AnCHSC gene of Aspergillus nidulans. The derived sequence also indicated that WdChs2p is an orthologous enzyme of Chs1p of Saccharomyces cerevisiae, which defines the class I chitin synthases. Disruptions of WdCHS2 produced strains that showed no obvious morphological defects in yeast vegetative growth or in ability to carry out polymorphic transitions from yeast cells to hyphae or to isotropic forms. However, assays showed that membranes of wdchs2Δ mutants were drastically reduced in chitin synthase activity. Other assays of membranes from a wdchs1Δwdchs3Δwdchs4Δ triple mutant showed that their residual chitin synthase activity was extremely sensitive to trypsin activation and was responsible for the majority of zymogenic activity. Although no loss of virulence was detected when wdchs2Δ strains were tested in a mouse model of acute infection, wdchs2Δwdchs3Δ disruptants were considerably less virulent in the same model, even though wdchs3Δ strains also had previously shown no loss of virulence. This virulence attenuation in the wdchs2Δwdchs3Δ mutants was similarly documented in a limited fashion in more-sensitive cyclophosphamide-induced immunocompromised mice. The importance of WdChs2p and WdChs3p to the virulence of W. dermatitidis was then confirmed by reconstituting virulence in the double mutant by the reintroduction of either WdCHS2 or WdCHS3 into the wdchs2Δwdchs3Δ mutant background.

Wangiella dermatitidis is an asexual, dematiaceous (melanized) human pathogen that can cause superficial, cutaneous, subcutaneous, and visceral or systemic pheohyphomycosis (26). This fungus is also a valuable model for discovering biologically and medically relevant information about the more than 100 other dematiaceous agents of mycosis, due to its vegetative polymorphism and ease of cultural manipulation (16, 49). Furthermore, the haploid nature and rapidly improving molecular genetic tractability of W. dermatitidis allows efficient gene disruption and site-specific, integrative gene expression studies (17, 26 27, 50, 51, 55, 56, 57, 58). At the simplest level, polymorphism in W. dermatitidis is expressed as three well-characterized modes of vegetative growth, (e.g., blastic, apical, and isotropic), which are primarily associated with the development of yeast, hyphal, and multicellular (sclerotic) morphologies (15, 19, 56). Transitions among these phenotypes in vitro are readily monitored and are easily induced in the wild type by extreme acidity or by calcium or nitrogen limitation or, in certain temperature-sensitive mutants, by the shift of cells to 37°C (19, 24, 47, 56). Collectively, these systems allow detailed evaluations of gene disruption and gene overexpression effects in all the growth forms of W. dermatitidis and make this fungus extremely attractive for studies of cell wall-related virulence factors at the molecular level (17).

Chitin, one of the major structural components of the fungal cell wall, is produced by a variety of chitin synthases (Chsp) that are encoded by a number of chitin synthase genes (CHS) (9, 20). These membrane-associated proteins are currently distributed among three to six isozyme classes depending on the fungus (4, 5, 33, 46). Chitin biosynthesis, chitin deposition patterns, and the functions of the three Chsp of Saccharomyces cerevisiae have been reviewed in great detail (6, 13, 38). Before bud emergence, Chs3p (class IV) lays down a chitin ring through which a new bud will emerge (41, 43). Chitin synthesis mediated by Chs2p (class II) then forms a primary septum after the daughter cell nears complete development (43, 44). Finally, Chs1p (class I) acts as a repair enzyme at cytokinesis, counterbalancing any chitin degradation caused by the chitinase that augments cell separation (7, 10). Although Chs1p contributes 90% of the chitin synthase activity when stimulated by trypsin in vitro, it actually adds very little chitin to the cell walls of yeast (37).

Genes encoding a class I chitin synthase have been identified in many fungi other than S. cerevisiae, but studies of the functional importance of their products have mainly involved Candida albicans, Aspergillus nidulans, and Aspergillus fumigatus (7, 12, 18, 21, 32, 35). In these fungi, as in S. cerevisiae, the genes that encode the class I isozymes are not essential and the functions of the class I enzymes are not obvious, because with the exception of reduced chitin synthase activities, mutants with these genes disrupted have no or only very minor defects. In contrast, chs mutants of some fungi with defective class III, IV, or V chitin synthases are reported to have significantly altered growth (1, 23, 32, 53, 54). By analogy with the roles played by the chitin synthases in S. cerevisiae, it is suspected that the chitin synthases of fungal pathogens of humans should also play different roles in growth and development and therefore should be important to their pathogenicity and virulence. However, with the exception of results with one of the two class III chitin synthases (AfChsGp) of A. fumigatus and the class II chitin synthase (CaChs1p) of C. albicans, few other published reports directly support this hypothesis (32, 36).

In W. dermatitidis, chitin is found primarily in the septal regions in yeast cells, whereas in hyphal and isotopic forms it is also localized throughout the cell wall (14, 22, 30). Previous research with the chitin synthase inhibitor, polyoxin, strongly suggests that chitin plays an important role in yeast-to-isotropic form transitions (14) and may also play a role in yeast-to-hypha transitions (30). Four different chitin synthase structural genes (WdCHS), which encode class I, II, III, and IV chitin synthases (WdChsp), were initially identified in W. dermatitidis by cloning and sequencing PCR products, prior to cloning and characterizing each full gene (5, 34, 48, 50, 51, 57). Very recently, evidence has also documented that W. dermatitidis has a fifth WdCHS gene (WdCHS5, class V), which encodes a protein with a myosin motor-like domain, as well as a chitin synthase domain (27). These genes were then disrupted with the ultimate aim of determining in a systematic way whether the chitin synthases of this pathogen contribute to its virulence (26, 27, 50, 51). To date, the results document that no single WdChsp is essential for viability and for yeast budding growth at 25°C and show that only WdChs5p is required for viability at 37°C and therefore virulence (27, 50, 51; unpublished data). These findings suggested that one or more of some chitin synthases of W. dermatitidis were redundant for function with the products of other WdCHS genes. Our finding that double mutants with both WdCHS1 and WdCHS2 disrupted are incapable of growth at 37°C but are capable of abnormal growth at 25°C strongly supported this hypothesis and further suggested that the products of at least these two genes were redundant for function at 25°C and also for viability at 37°C (27, 50, 57; unpublished data). These findings also strongly suggested that additional investigations of these latter two genes and the relationship of their products with those of WdCHS3 and WdCHS4 might help confirm this hypothesis.

In this report, we describe how WdCHS2, which encodes the class I chitin synthase in W. dermatitidis, was cloned, together with results about its characterization, expression, and disruption in the wild-type strain and in a wdchs3Δ disruption mutant. We also estimate the relative levels of residual enzymatic activities remaining in membranes of a number of single and double chitin synthase mutants with either WdCHS2,WdCHS3, or both disrupted and then compare these activities with those of the wild type and of a wdchs1Δwdchs3Δwdchs4Δ triple mutant. In addition, we report that, although the disruption of WdCHS2, like that of WdCHS3 (51), did not affect virulence, the disruption of both WdCHS2 and WdCHS3 in the same background caused a loss of virulence in mutants tested in two mouse models of acute infection. This virulence loss was then restored by the reintroduction of either WdCHS2 or WdCHS3 into the double-mutant background. This is the first report that directly implicates a class I chitin synthase with virulence among fungal pathogens of humans.

MATERIALS AND METHODS

Strains and media.

The W. dermatitidis strains used in this work are listed in Table 1. The wild-type strain 8656 (ATCC 34100 [Exophiala dermatitidis CBS 525.76]), which has been extensively described previously (see, for example, reference 24), was used as the original parental stock for the derivation of all other strains. Strains INVαF′ (Invitrogen, San Diego, Calif.), JM109 (Promega, Madison, Wis.), and XL1-Blue (Stratagene, La Jolla, Calif.) were used as Escherichia coli plasmid hosts. W. dermatitidis cells were grown in the rich broth medium YPD (50, 58) or the minimal medium SD (15). Drug selection plates for isolating W. dermatitidis transformants were made by adding agar (1.5%) and hygromycin B (HmB; Sigma, St. Louis, Mo.) or phleomycin (Sigma) to YPD at the final concentration of 50 μg/ml or chlorimuron ethyl (provided by J. Sweigard, DuPont Co., Wilmington, Del.) to SD for detection of resistance conferred by the sulfonyl urea resistance (sur) gene at the final concentration of 20 μg/ml. E. coli strains were grown in Luria-Bertani medium supplemented with 100 μg of ampillicin or 20 μg of chloramphenicol per ml.

TABLE 1.

Strains used in this study

Straina	Parent strain	Genotype	Reference or source
Wd8656		Wild type	ATCC 34100
wdchs2Δ-1	Wd8656	wdchs2::hph	This work
wdchs3Δ-1	Wd8656	wdchs3::hph	ATCC MYA-886 (51)
wdchs1Δwdchs3Δ-1	wdchs3Δ-1	wdchs1::blewdchs3::hph	This work
wdchs2Δwdchs3Δ-1	wdchs3Δ-1	wdchs2::surwdchs3::hph	This work
wdchs2Δwdchs3Δ-7	wdchs3Δ-1	wdchs2::surwdchs3::hph	This work
wdchs2Δwdchs3Δ-E	wdchs3Δ-1	wdchs2::surwdchs3::hph	This work
wdchs2Δwdchs3Δ-Z	wdchs3Δ-1	wdchs2::surwdchs3::hph	This work
wdchs1Δwdchs3Δwdchs4Δ-1	wdchs1Δwdchs3Δ-1	wdchs1::blewdchs3::hphwdchs4::sur	This work
wdchs2Δwdchs3Δ-E2-1	wdchs2Δwdchs3Δ-E	wdchs2::surwdchs3::hph-WdCHS2-ble	This work
wdchs2Δwdchs3Δ-E2-2	wdchs2Δwdchs3Δ-E	wdchs2::surwdchs3::hph-WdCHS2-ble	This work
wdchs2Δwdchs3Δ-E3-3	wdchs2Δwdchs3Δ-E	wdchs2::surwdchs3::hph-WdCHS3-ble	This work
wdchs2Δwdchs3Δ-E3-5	wdchs2Δwdchs3Δ-E	wdchs2::surwdchs3::hph-WdCHS3-ble	This work

^aAll of the strains used in this study are isogenic. 

Plasmids and transformations.

The WdCHS2 replacement vector pLZ41 was constructed by using pLZ38, which contains a 4.9-kb KpnI/XbaI fragment from the original WdCHS2 clone. After being cut with PstI, the cohesive ends were blunted by fill in by using a Klenow fragment, and the resulting fragment was cut by BamHI and then ligated with the 2.9-kb BamHI/StuI fragment of pAN7-1 (39), which contains the promoter region of the gpd gene and the hph gene for selection of HmB-resistant transformants. The WdCHS2 integrative disruption vector pLZ58 was constructed by using a 1.0-kb EcoRI fragment of WdCHS2 from pLZ13 (this work) that was inserted into the multiple cloning site of pCB1029 (provided by J. Sweigard), which contains the sur gene marker (50). The resulting plasmid pLZ52 was then used to produce pLZ58 by digestion with XbaI, fill in with a Klenow fragment, digestion with SmaI, and then allowing self ligation. The WdCHS1 gene disruption plasmid, pLZ56, was constructed by cloning a 3.3-kb BglII fragment from pUT737 containing the phleomycin resistance gene marker (ble) into the BglII site of pWdCHS1.KS (34). The WdCHS4 disruption plasmid, pHY1-1, was constructed by inserting the 0.8-kb BclI/EcoRI fragment from pCHS4-5 (50) into pCB1029. These latter two disruption plasmids were used sequentially to derive the triple mutant wdchs1Δwdchs3Δwdchs4Δ. For complementation of the wdchs2Δwdchs3Δ-E double mutant, plasmids pZW9911 that contained WdCHS2 and pZW9912 that contained WdCHS3 were constructed by cloning a 5.8-kb XbaI fragment from pLZ14 (this work) and a 3.4-kb KpnI-XbaI fragment from pZW122 (51) into the multiple cloning sites of pBF36, which is a pBluescript KS(+) vector that also contained the ble marker. Transformations of W. dermatitidis were performed by electroporation of yeast cells as described previously (50).

Preparation and analysis of DNA and construction of the partial genomic library.

Genomic DNA was isolated from W. dermatitidis as previously described (50, 58). Southern blotting was performed by standard methods (2). DNA fragments (25 ng) used for probes in Southern analysis and colony hybridizations were labeled with [α-³²P]dATP by using the Prime-a-Gene kit (Promega). To clone WdCHS2, a partial genomic DNA library was constructed by completely digesting genomic DNA with XbaI and then excising and eluting the DNA fragments that ranged from 5 to 7 kb from a gel after electrophoresis. After ligation of the DNA into the pBluescript KS(+) vector that was cut by XbaI and dephosphorylated, the ligation product was transformed into XL1-Blue competent cells. The resulting XbaI library contained about 7,000 independent clones that were divided into several small pools. One pool from about 1,400 independent colonies was shown to contain the WdCHS2 gene fragment by PCR amplification and then subjected to library screening by colony hybridization (2). Methods for nucleic acid analysis have also been described (50, 58), except for the analysis of the WdCHS2 promoter region, which was performed with MatInspector (40).

Primers and RT-PCR amplifications.

The primers used to detect expression and to confirm the putative intron in WdCHS2 were CHS2.RT (5′-GACATGGCTACAACCGTCTA-3′) and CHS2.GSP2 (5′-CTATGTAATCGTTTCTGCCGTAACC-3′). The first strand of cDNA synthesis was performed according to the protocol supplied with SuperScript II reverse transcriptase (Life Technologies, Inc., Rockville, Md.) by using 1 μg of mRNA isolated from wild-type cells grown at 25 or 37°C. The first strand synthesized was subject to PCR amplification. Amplification was for 35 cycles as follows: premelt, 94°C for 5 min; denaturation, 94°C for 1 min; hybridization, 50°C for 2 min; and extension, 72°C for 3 min (10 min on the last cycle). The primers for the semiquantitative reverse transcription-PCR (RT-PCR) were designed to ensure the exclusive amplification of WdCHS2 by identifying a region of low conservation (bases 1047 to 1462) in the gene after alignment of all the known WdCHS sequences. The resulting specific primers had the following sequences: CHS2F (5′-GCCTACGGTCAGCAGTATGGGCAGCAT-3′) and CHS2R (5′-GGAGGTTTCGCGTGAGGAACGTTGGC-3′). The RNA for the semiquantitative evaluations of WdCHS2 expression was obtained from 10⁹ wild-type cells cultured in liquid YPD at 25 or 37°C. The inoculum used for these experimental cultures was from yeast cultures grown at 25°C to mid-log phase through four successive growth cycles, after which time the final subculture was split into two equal parts for incubation at the two temperatures for 3 h prior to RNA extraction with acid phenol. After DNase treatment (1 h) at 37°C, DNA-free RNA samples were dissolved and diluted in RNase-free water and diluted to various concentrations. Primer pairs for the 18S rRNA gene and for their 18S PCR Competimers (Ambion, Inc., Austin, Tex.) were used to titrate the two RNA samples to the same total RNA level. The actual RT-PCR for each sample was then done with one tube by using the Access RT-PCR System (Promega). The RT-PCR mixture for 25 μl consisted of 1 μl of MgSO₄ (25 mM), 0.5 μl of deoxynucleoside triphosphate (10 mM), 2.5 μl of buffer (10×), 1 μl of forward and reverse primers (2.5 μM), 0.5 μl of avian myeloblastosis virus (AMV) reverse transcriptase, 0.5 μl of Tfl DNA polymerase, 1 to 8 μl of total RNA in proper dilution, and RNase-free water supplemented to 25 μl. The RT-PCR was run in a GeneAmp 2400 PCR system (Perkin-Elmer, Wellesley, Mass.) with the following cycling conditions: 48°C for 1 h and 94°C for 2 min; followed by different cycles of 94°C for 1 min, 58°C for 1 min, and 70°C for 1 min; and then an extra step for elongation at 68°C for 5 min. Both the cycling number and the template amount were carefully calibrated to ensure that the RT-PCR was done within the exponential phase of amplification. For each sample, a parallel negative control made up with the same components used for normal RT-PCR except AMV reverse transcriptase was used to ensure there was no trace DNA contamination. Products of the RT-PCR were run on 1.2% agarose gels and then viewed under UV light and analyzed by densitometry by using the AlphaImager 1220 Documentation and Analysis system (Carnock, Staffordshire, United Kingdom).

Chitin content and chitin synthase activity assays.

Chitin contents were measured by the procedure described previously (50, 58). Cell membranes were prepared and activities of chitin synthase were determined by a modification of the method of Orlean (37) as described previously (51, 58).

Virulence studies of mice.

Tests for virulence in an immunocompetent mouse model system were done as described previously (17, 50). Survival fractions were calculated by using the Kaplan-Meier method, and survival curves were tested for significant difference (P < 0.01) by the Mantel-Haenszel test with GraphPad Prism software, version 3.00, for Windows. Tests for virulence in an immunocompromised mouse model were carried out with mice injected with cyclophosphamide at day 4, and one was done prior to fungal infection. This treatment induces severe neutropenia resulting in a 10- to 100-fold-greater sensitivity in the mice to infectious fungi (31, 42, 45). This is normally reflected in a 10- to 100-fold less infectious dose to produce a 50% lethal dose (LD₅₀) comparable to the LD₅₀ with immunocompetent mice. Because the usual inoculum of W. dermatitidis used in mice with functioning immune systems was 9 × 10⁶ cells, this inoculum size and three other smaller ones (3 × 10⁶, 1 × 10⁶, and 3.3 × 10⁵) were used in these experiments, which involved comparisons of only the wild-type strain and one wdchs2Δwdchs3Δ mutant.

Nucleotide sequence accession number.

The nucleotide sequence of the WdCHS2 gene was assigned GenBank accession number AF 052606.

RESULTS

WdCHS2 encodes a class I chitin synthase.

A partial genomic DNA library was screened by using a 0.6-kb PCR fragment of WdCHS2 as a probe, and two clones were obtained. Restriction enzyme mapping showed that both (pLZ13 and pLZ14) contained identical 5.6-kb inserts (see Fig. 3A), but with opposite orientations. Several subclones were then used to sequence a 4.2-kb region on both strands. The resulting sequence (Fig. 1) revealed a 2,784-bp open reading frame with a 59-bp putative intron sequence found between bp 20 and 78. The putative intron began with GTAAAC, had the invariant CTGATT sequence in the middle, and ended with TAG. Therefore, it was similar to those reported for other eukaryotic introns (3) and also matched the consensus of most introns found in other genes of W. dermatitidis (the boldface nucleotides are conserved in most introns identified to date in W. dermatitidis [11]). The predicted amino acid sequence of WdChs2p derived from these nucleotide sequences was 928 amino acids, with a molecular mass of 106 kDa and a putative pI of 7.28. Comparisons of the deduced WdChs2p protein with other deduced chitin synthases indicated the highest amino acid identity with CiChs1p (72%) of Coccidioides immitis (GenBank submission no. AF276826; M. Mandel, J. N. Galgiani, and M. J. Orbach). Progressively lower amino acid identities were found with AnChsC (66%) of A. nidulans (35), NcChs3p (62%) of N. crassa (GenBank submission no. AF127086; A. Beth-Din and O. Yarden), CaChs1p (44%) of C. albicans (12), and Chs1p (44%) of S. cerevisiae (7). Thus, WdChs2p represents a class I chitin synthase as originally defined by Bowen et al. (5) and then extended by others (4, 46). Interestingly, analysis of the WdCHS2 5′ upsteam sequence identified a number of sequences for putative cis-acting elements for AbaAp and StrEp/BrlAp, which is similar to the finding with AnCHSC of A. nidulans, where such elements have been identified and suggested to be important in the transcriptional regulation of chitin biosynthesis during sporulation (18). A number of other stress-related and sporulation-related putative cis-acting elements were also identified for StuAp and for Nit2p. These putative regulatory elements have also been identified in the upstream regions of WdCHS3 and WdCHS5 (unpublished data).

FIG. 3.

Disruption of WdCHS2 in W. dermatitidis and complementation of the wdchs2Δwdchs3Δ-E double mutant with either WdCHS2 or WdCHS3. (A) Restriction enzyme map of the 5.8-kb XbaI fragment containing the WdCHS2 gene. The arrow indicates the position and length of WdCHS2 in the insert and the direction of its transcription. Abbreviations: B, BamHI; G, BglII; HindIII; K, KpnI; P, PstI; S, SalI; Sa, SacII. (B) The one-step gene disruption strategy for generating the disruption strain wdchs2Δ-1. The solid line indicates the position of the PCR-generated probe used in the results shown in panel C. (C) Southern blot analysis of the wild type (lane 1 and lane 7), the single disruption strains wdchs2Δ-1 (lanes 2 and 8) and wdchs3Δ-1 (lanes 3 and 9), the double disruption strain wdchs2Δwdchs3Δ-E (lanes 4 and 10), and the complemented strains wdchs2Δwdchs3Δ–WdCHS2-E2-1 (lanes 5 and 11) and wdchs2Δwdchs3Δ–WdCHS3-E3-3 (lanes 6 and 12). Genomic DNA from each strain was digested with XbaI and hybridized with a 417-bp PCR probe of WdCHS2 (lanes 1 to 6) and a 3.2-kb KpnI-XbaI fragment (lanes 7 to 12).

FIG. 1.

Nucleotide and derived amino acid sequences of the WdCHS2 gene of W. dermatitidis. The putative intron sequence is indicated in lowercase. The conserved intron splicing signals are underlined. The sequences corresponding to the CHS primer 1 and primer 2 binding sites are also underlined. The putative start site is at position 1.

Expression of WdCHS2 occurs at both 25 and 37°C.

Northern blot analyses of total RNA were not sensitive enough to detect the expression of WdCHS2 (data not shown). Therefore, more-sensitive assays involving RT-PCR were carried out, first to detect the expression of WdCHS genes and simultaneously to confirm the putative intron identified by sequencing WdCHS2 and then in an attempt to determine whether this gene, like WdCHS3 (51), exhibits significant enhanced expression in cells shifted from 25 to 37°C. For the former, the RT-PCR amplifications were performed with primers that bind sequences adjacent to the putative intron sequence. The results documented that WdCH2 was expressed at both 25 and 37°C (Fig. 2). Furthermore, the RT-PCR product derived was determined to be about 300 bp in size (Fig. 2A and B, lanes 3), which was about 50 to 60 bp shorter than the product amplified from the cloned WdCHS2 gene in pLZ13 (Fig. 2A and B, lanes 2). Sequencing of this RT-PCR product showed that the intron of WdCHS2 had been spliced from its mRNA (data not shown). For the latter, our more-sensitive, semiquantitative RT-PCR methodology confirmed that there was detectable expression of WdCHS2 at both 25 and 37°C (Fig. 2C). This methodology also indicated that, although WdCHS2 expression increased in cells shifted from 25 to 37°C (Fig. 2Ca), its estimated increase was only about 32% when the bands were standardized against the RT-PCR products generated with primers specific for 18S rRNA amplifications (Fig. 2Cb) and then compared by densitometric analysis (data not shown).

FIG. 2.

RT-PCR amplifications of WdCHS2 to confirm the putative intron sequence and to document and semiquantitatively estimate the levels of WdCHS2 gene expression at both 25 and 37°C. (A) The RT-PCR amplification was with mRNA isolated from wild-type cells grown to mid-log phase at 25°C in YPD medium (lane 3) and primers WdCHS2.RT and WdCHS2.GSP. A PCR amplification was also performed as a control by using a genomic clone of WdCHS2 as a temple (lane 2). The DNA standard (1-kb ladder) is in lane 1. (B) Similar to the analysis shown in panel A except the RT-PCR amplification was performed with mRNA isolated from wild-type cells grown to mid-log phase at 37°C in YPD medium. (C) Total RNA samples of 0.5 ng from 25°C (lane a1) and 37°C (lane a2) were reverse transcribed and further amplified by PCR with primer CHS2F and CHS2R, which were standardized against internal controls (lanes b1 and b2) represented by RT-PCR products from the same RNA samples used in lanes a1 and b1 but generated by using primers exclusive for the amplification of the 18S rRNA gene.

WdCHS2 is not an essential gene, but WdChs2p produces much of the chitin synthase zymogenic activity detected in vitro.

To study the function of WdChs2p in W. dermatitidis, a WdCHS2 gene disruption was performed by a one-step gene replacement method (Fig. 3B). Yeast cells were transformed with the KpnI and XbaI fragment released from pLZ41 at an efficiency of eight transformants/μg of DNA. Southern analysis of genomic DNA from 14 HmB-resistant transformants showed that 7 resulted from site-specific integrations and indicated that they were WdCHS2 disruption (wdchs2Δ) strains (Fig. 3C; data shown only for wdchs2Δ-1). However, neither growth rate determinations (Fig. 4) nor microscopic observations (data not shown) revealed that any of the wdchs2Δ disruptants had observable defects in terms of yeast vegetative growth at 25 or 37°C or of phenotype transition potential during the formation of hyphae or isotropic forms from yeast cells (data not shown). Chitin content and Calcofluor white staining patterns indicative of chitin localization in wdchs2Δ-1 cells were also identical to those of the wild type (data not shown).

FIG. 4.

Comparison of growth rate at 25°C and 37°C of W. dermatitidis 8656 (wild type [wt]) and selected wdchs2Δ and wdchs3Δ single and double mutants. Mid-log-phase yeast cells were used to inoculate YPD medium at an initial concentration of about 10⁵ cells/ml. Cultures were incubated with shaking at 37°C in a Psycrotherm Incubator-Shaker (New Brunswick Scientific Co., Edison, N.J.) or at 25°C in a New Brunswick G-76 Water Bath Shaker at gyrotory speed settings of 200 rpm. Optical densities were measured by using a DU Spectrophotometer (Beckman Instruments, Inc., Irvine, Calif.).

It has been shown that Chs1p (class I) in S. cerevisiae contributes most of the measurable zymogenic activity detected under standard assay conditions, even though it contributes little cell wall chitin (8). To test whether this was also true for the WdChs2p (class I) isozyme of W. dermatitidis, chitin synthase activity assays were performed with membrane preparations from the wild type and the wdchs2Δ-1 strain. The results showed that the WdChsp activities of membranes pretreated with trypsin (the so-called zymogenic activity) were reduced by about 85% in membrane preparations of strain wdchs2Δ-1 compared to those of the wild-type strain (Fig. 5). This confirmed similar findings with membrane preparations from this strain and others having WdCHS2 disrupted with a hisG cassette, such as the wdchs2Δ-3 (wdchs2::hisG) strain (58).

FIG. 5.

Comparisons of relative chitin synthase activities. The wild-type strain (wt) and the wdchs2Δ-1, wdchs3Δ-1, wdchs2Δwdchs3Δ-E, wdchs2Δwdchs3Δ–WdCHS2-E2-1, wdchs2Δwdchs3Δ–WdCHS3-E3-3, and wdchs1Δwdchs3Δwdchs4Δ-1 mutant strains were grown in YPD liquid medium at 37°C for 20 h prior to membrane harvest. The membrane protein (30 μg) from each strain was assayed after a 40-μg/ml trypsin treatment. The results are derived from at least two independent experiments. Standard deviations are shown.

WdChs2p is extremely sensitive to chitin synthase activation by trypsin.

Because about 85% of chitin synthase activity in membranes of cells grown at 37°C was estimated to result from WdChs2p (Fig. 5) and because a quadruple mutant with only WdChs2p has not been derived and WdChs5p seems to contribute only minimally to total WdChsp activity (unpublished data), the chitin synthase activity present in a wdchs1Δwdchs3Δwdchs4Δ triple mutant was measured. To derive this strain, the WdCHS1 gene was disrupted in the wdchs3Δ-1 mutant by transforming with an 8.5-kb StuI fragment from pLZ56 (57) by using a one-step replacement method. Of the resulting phleomycin-resistant transformants, 5% were shown to be wdchs1Δwdchs3Δ double mutants by Southern blotting (data not shown). The vector pHY1 was then linearized with BstXI and used to disrupt WdCHS4 in the wdchs1Δwdchs3Δ double mutant. Among the resulting sulfonyl urea-resistant transformants, 50% were confirmed by Southern blotting to be wdchs1Δwdchs3Δwdchs4Δ triple mutants (data not shown). Surprisingly, this triple mutant grew well at 25 and 37°C and showed only a short yeast chain-clumpy phenotype (data not shown), which is the combined phenotype of wdchs1Δ (chains) and the wdchs4Δ (clumpy) single mutants (50, 57), and was as virulent as the wild-type strain in our mouse model (see Fig. 7A). However, unexpectedly low levels of WdChs2p activity were detected after trypsin activation of membranes from the triple mutant grown at 37°C (Fig. 5), when tested by using our standard concentration of trypsin (40 μg/ml). Therefore, the WdChs2p activity was retested with lower trypsin concentrations, as was done previously during the characterization of the residual activity of a wdchs1Δwdchs2Δwdchs3Δ mutant, with only WdChs4p and WdChs5p (50). As was shown with that mutant, the results showed that our standard concentration of trypsin was in excess of the concentration required for demonstrating maximal activity of the residual membrane activity of the wdchs1Δwdchs3Δwdchs4Δ mutant (Fig. 6A). The results also substantiated that the enzyme was indeed zymogenic. In addition, the results showed that the residual activity in membranes of wdchs1Δwdchs3Δwdchs4Δ could be activated to 10 times higher activity than the residual membrane activity of the wdchs1Δwdchs2Δwdchs3Δ mutant (50) by trypsin concentrations as low as 1.5 μg/ml (Fig. 6A) in just 5 min (Fig. 6B).

FIG. 7.

Mouse survival analyses after injection of immunocompetent (A) and cyclophosphamide-induced immunocompromised (B and C) mice with wild-type and wdchsΔ mutant strains. For each experiment with immunocompetent mice (A), groups usually numbering 10 mice were injected with log-phase yeast cells of W. dermatitidis wild type (8656), wdchs2Δ-1, wdchs3Δ-1, wdchs2Δ wdchs3Δ-E or with wdchs2Δwdchs3Δ-E complemented with either WdCHS2 (strains E2-1) or WdCHS3 (strains E3-3). The mice were injected with 9 × 10⁶ cells per mouse and monitored for 10 to 15 days to determine the survival rate. The data presented are in all but one case (the triple mutant) the average data from at least two independent experiments with each strain. Virtually identical data were obtained for independently derived strains wdchs2Δwdchs3Δ-7 and wdchs2Δwdchs3Δ-Z and complemented strains wdchs2Δwdchs3Δ–WdCHS3-E2-2 and wdchs2Δwdchs3Δ–WdCHS3-E3-5. Survival fractions were calculated by the Kaplan-Meier method, and the survival curves were tested for significant difference (P < 0.01) by the Mantel-Haenszel test by using GraphPad Prism software, version 3.00, for Windows. The tests for virulence in the immunocompromised mouse model were carried out with groups of mice injected with cyclophosphamide at day 4 and at day 1 prior to fungal infection. For each strain, groups of five mice were inoculated at four concentrations of yeast cells, with the highest concentration being equivalent to that usually used with W. dermatitidis in mice with a functioning immune system, such as those used to derive the data shown in panel A.

FIG. 6.

Zymogenic characteristic of the residual activity in the triple mutant wdchs1Δwdchs3Δwdchs4Δ-1. Membrane proteins were isolated from mutant cells grown in YPD liquid medium at 37°C for 20 h. (A) Trypsin titration of 30 μg of membrane proteins. (B) Optimal activation time of WdChs2p in the presence of 1.5 μg of trypsin/ml. The results are derived from two independent experiments, and each sample was duplicated. Standard deviations are shown.

The wdchs2Δwdchs3Δ double mutant, but not the wdchs2Δ single mutant, is less virulent.

Because all of the single wdchsΔ mutants of W. dermatitidis, except those with newly discovered WdCHS5 disrupted, are as virulent as the wild type in animal tests (27, 50, 51; unpublished data), numerous combinations of double and triple wdchsΔ mutants have been constructed and tested similarly. Of those available and capable of growth at 37°C, which includes the wdchs1Δwdchs3Δwdchs4Δ-1 triple mutant but not those with WdCHS1 and WdCHS2 disrupted in the same background, only the wdchs2Δwdchs3Δ double mutants constructed by integrating the BamHI-linearized plasmid pLZ58 into the WdCHS2 locus of wdchs3Δ-1 have so far been found to be less virulent in our immunocompetent mouse model (Fig. 7A). Although about 12% of these sulfonyl-resistant transformants were proved to be wdchs2Δwdchs3Δ double disruptants by Southern blot analysis (Fig. 3; data shown only for wdchs2Δwdchs3Δ-E, lanes 4 and 10), none exhibited reduced growth rates or abnormal phenotypes compared to the wild type, even though all had extremely low chitin synthase activities (Fig. 5). Other tests designed to assess the extent of the attenuation of one of these double mutants (wdchs2Δwdchs3Δ-1) showed that even in cyclophosphamide-induced immunoincompetent (neutropenic) mice, an inoculum of these mutant cells at least threefold higher than that of the wild type was required in order to bring about lethal outcomes (Fig. 7B and C).

In order to test whether either the WdCHS2 or the WdCHS3 gene could restore the virulence of a wdchs2Δwdchs3Δ mutant, putatively reconstituted strains complemented with either WdCHS2 or WdCHS3 were constructed by directly transforming pZW9911 or pZW9912 into the double mutant wdchs2Δwdchs3Δ-E, respectively. Five sulfonyl urea-resistant transformants from each construction were randomly chosen and assayed for chitin synthase activities (Fig. 5; data shown only for wdchs2Δwdchs3Δ–WdCHS2-E2-1 and wdchs2Δwdchs3Δ–WdCHS3-E3-3). Of these, two of each were found to have significantly higher activities compared with that of the wdchs2Δwdchs3Δ-E parental strain, which indicated that the added chitin synthase activities came from the expression of exogenous WdCHS2 or WdCHS3. Southern blotting demonstrated that both pZW9911 and pZW9912, the plasmids carrying these genes, respectively, were ectopically integrated into the genomes of the two sets of strains (Fig. 3). Tests of these strains with our immunocompetent mouse model showed that the introduction of either gene into the double disruption strain reestablished virulence comparable to the results seen in the wild-type strain and the wdchs2Δ-1 and wdchs3Δ-1 single mutants (Fig. 7A).

DISCUSSION

We previously reported that single wdchsΔ mutants of W. dermatitidis with either WdCHS3 or WdCHS4 disrupted showed no loss of virulence when tested with a mouse model of acute infection (50, 51). This same result was also found for mutants with only WdCHS2 disrupted, as reported in this study, and for single disruption mutants of WdCHS1 (unpublished data). Therefore, among the single chitin synthase mutants at hand, only those with newly discovered WdCHS5 disrupted, which lose viability with time at 37°C, showed a loss of virulence in our acute model of infection, even though these strains grow normally at 25°C (27). Thus, wdchs5Δ mutants, in the same manner as wdchs1Δwdchs2Δ double and wdchs1Δwdchs2Δwdchs3Δ triple disruption mutants, presumably are rendered avirulent because their temperature sensitivity precludes survival at temperatures of infection (27, 50). In contrast, other tests with the same mouse model showed that loss of virulence was very pronounced in strains derived for this study, which had both WdCHS2 and WdCHS3 disrupted in the same background, even though these strains grew normally at wild-type rates and showed no morphological abnormalities when cultured at 25 or 37°C. The extent of this attenuation was documented further with wdchs2Δwdchs3Δ-1 in a more sensitive mouse model that is being developed for W. dermatitidis, because of the increasing numbers of phaeohyphomycosis cases that are being diagnosed among immunocompromised patients (25, 28, 29, 52). This surprising finding, coupled with results which showed that mutants with both WdCHS2 and WdCHS1 disrupted are unable to grow at 37°C but can grow weakly in an abnormal morphology at 25°C (50, 57) and that a mutant with only WdChs2p and WdChs5p is fully virulent, suggests that WdChs2p, unlike other class I chitin synthases, plays a number of critically important roles in growth and virulence because of its commonality to both processes. Although the basis for the critical roles suggested for WdChs2p is not totally clear at this time, we speculate that it resides in the possibility that WdChs2p functions mainly in an auxiliary or a redundant capacity, which compensates for the loss of function of one or more of the other chitin synthase isozymes of W. dermatitidis.

Support for the hypothesis that WdChs2p serves in a redundant or auxiliary capacity comes from a number of findings. The first was the determination that WdChs2p is a class I isozyme with a number of other attributes similar to those of its ortholog, Chs1p, in S. cerevisiae, which has become variously known as an auxiliary or repair enzyme in that fungus. The second was the finding previously (50, 57) that the disruption of both WdCHS1 and WdCHS2 in the same background produces strains that grow poorly at 25°C and not at all at 37°C (50, 57). We suggest this means that the products of these two genes are overlapping for function at 25°C and for viability at 37°C (50, 57). A similar functionally redundant role has been suggested recently for AnChsCp (class I) of A. nidulans, which when disrupted together with AnChsAp (class II), causes marked hyphal and conidial defects (18). The third was the observation that, like Chs1p of S. cerevisiae, WdChs2p is produced in excess of the collective activities of their counterparts and thus, in terms of activity, could easily serve in a redundant capacity capable of substituting for the other isozymes and particularly WdChs1p. The fourth was our finding that mutants with only functional WdChs2p and WdChs5p isozymes grow normally and are fully virulent, meaning that these two enzymes together or separately are also redundant for functions with one or more of the others. However, of these two enzymes we know that WdChs5p cannot substitute for the loss of both WdChs1p and WdChs2p, because the wdchs1Δwdchs2Δ and wdchs1Δwdchs2Δwdchs3Δ mutants cannot survive at 37°C and thus are not rescued by any or all of WdChs3p, WdChs4p, or WdChs5p. Taken together with the unexpected finding presented in this report that the wdchs2Δwdchs3Δ mutants grew well at 37°C but did not cause lethal infections in mice, we speculate that WdChs1p may not function well under the many stress conditions, in addition to temperature, which W. dermatitidis encounters during these infections. Under this scenario, we speculate that, during infections, either WdChs2p or WdChs3p must be present as an auxiliary chitin synthase with the capacity to substitute for poorly functioning WdChs1p. Studies of WdChs1p and newly discovered WdChs5p are in progress to address this hypothesis.