INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Coccidioides immitis is the causative agent of a fungal respiratory disease known as San Joaquin Valley fever or coccidioidomycosis. The fungus is considered to be a primary pathogen of humans and dogs that is capable of establishing infections in an immunocompetent host (16, 23). The saprobic phase of C. immitis occurs in alkaline soil of southwestern desert regions of the United States, which extend from California to western Texas (26). Host infection typically occurs by inhalation of dry airborne spores (arthroconidia) that are small enough to reach the alveoli (5). The fungus is highly virulent: BALB/c mice inoculated intranasally with as few as 100 arthroconidia of a strain of the fungus known to be virulent die within 2 to 3 weeks as a result of necrotic pulmonary lesions and respiratory failure (6). The pathogenicity of C. immitis is further underscored by the fact that coccidioidomycosis is the most frequently diagnosed mycosis of laboratory personnel who work with medically important fungi (13).

The saprobic and parasitic phases are considered to be haploid (24), although this has yet to be confirmed. C. immitis has no known sexual state, but molecular evidence indicates that recombination occurs within the fungal population in nature (2). Morphogenesis of the asexual, parasitic phase of C. immitis is unlike that of any of the other human fungal pathogens (5). The parasitic cycle can be reproduced in vitro (21) and is initiated by conversion of the cylindrical arthroconidium into a multinucleate round cell (spherule). The latter undergoes isotropic growth and segmentation and ultimately gives rise to a multiplicity of endospores. The high fecundity of the parasitic phase of C. immitis may contribute to the ability of the pathogen to overcome innate cellular immune defenses of the host, particularly when a large inoculum of arthroconidia has been inhaled. Morphogenetic factors that control pivotal events of endosporulation represent potential anti-Coccidioides drug targets.

Our knowledge of the molecular biology of C. immitis is still in its infancy, largely because few laboratories have focused their efforts on this microorganism. A major emphasis in C. immitis gene cloning studies has been antigen identification, expression, and characterization. One immunoreactive macromolecule that has been the focus of multiple investigations is the complement fixation (CF) antigen, which has been used clinically for decades in the serodiagnosis of coccidioidal infections (27). The molecular size of the native CF antigen, estimated under reducing conditions by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), is 48 kDa (42). Biochemical characterization of the purified 48-kDa polypeptide revealed that the antigen functions as a chitinase (15) and is secreted by C. immitis in both the saprobic and parasitic phases. It was suggested that the active chitinase associates with the segmentation wall of parasitic cells during early endosporulation (3) and participates in an essential process of cell wall modification as endospores undergo differentiation and subsequent release from the maternal spherule (4). The gene which encodes the CF antigen (CTS1) has been cloned (29), and temporal expression of the chitinase gene was evaluated in vitro during parasitic cell development (P. W. Thomas and G. T. Cole, Abstr. 99th Gen. Meet. Am. Soc. Microbiol. 1999, abstr. F-52, p. 306, 1999). The results of the latter study suggested that the maximum level of expression of the CTS1 gene occurs during the endosporulation phase of C. immitis. In this paper, we report the first targeted disruption of a C. immitis gene and further evaluate whether the CTS1 chitinase plays a morphogenetic role in the parasitic cycle of the pathogen.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, media, and growth conditions. To obtain arthroconidia, C. immitis strain C735 in the saprobic phase was grown on GYE agar (1% glucose, 0.5% yeast extract, 1.5% agar) at 30°C for 3 to 4 weeks. For DNA extraction, parental and transformant strains in the saprobic phase were grown in GYE liquid medium at 30°C without or with hygromycin (75 µg/ml; Sigma, St. Louis, Mo.). For parasitic-phase growth, the same strains were grown in Converse medium as previously described (21).

Isolation of a BamHI genomic fragment which contains the CTS1 gene. An 8-kb genomic fragment was obtained by screening a genomic library of C. immitis (38) with a radioisotope-labeled 515-bp PCR product derived from amplication of a fragment of the CTS1 gene as described by Pishko et al. (29). The 8-kb genomic fragment was subcloned into pBluescript SK⁺ and restriction mapped using the known sequence of the CTS1 gene (29) and a standard digestion protocol (34).

Construction of a transformation plasmid (pcts1) for disruption of the CTS1 gene. The pcts1 plasmid was designed as a gene disruption vector and was constructed using pAN7.1 (32) and a 1.1-kb internal fragment of the CTS1 gene amplified by PCR (bp 68 to 1177 of the reported CTS1 gene [29]). To facilitate cloning, BglII and BstEII sites were added to the 5' end of the upstream and downstream primers, respectively (primers A and B in Table 1). These same two restriction sites are present in pAN7.1 upstream from the functional GPD promoter sequence (32). Both pAN7.1 and the 1.1-kb CTS1 PCR product were digested with BglII and BstEII. The CTS1 PCR product and the larger fragment of digested pAN7.1 (6.5 kb) were purified by agarose gel separation and extraction using a QIAEX II Gel Extraction Kit (Qiagen, Chatsworth, Calif.). The two fragments were then ligated, and the product was used to transform Escherichia coli strain DH5. The pcts1 plasmid isolated from E. coli was used for subsequent gene-targeted disruption experiments. Prior to transformation, the plasmid was linearized with ClaI and then purified by phenol-chloroform extraction and ethanol precipitation in accordance with standard protocols (34).

Transformation procedure. Transformation of C. immitis was performed using a modified protocol which has been successfully employed for Aspergillus nidulans and Aspergillus fumigatus (28, 33). Arthroconidia from 8 to 10 agar plates with dense sporulating mycelia were harvested by flooding each plate with 5 ml of GYE medium plus 0.1% Tween 80 (Sigma), followed by gentle disruption of the mycelia with a flamed wire inoculation loop. Suspensions of arthroconidia from each plate were pooled in a 50-ml tube, vortexed vigorously, and used to inoculate a 1-liter flask that contained 150 ml of GYE medium. In a typical experiment, the flask contained approximately 1.0 × 10⁶ to 3.0 × 10⁶ arthroconidia per ml of GYE medium. The cell suspension was incubated in a gyratory shaker (100 rpm, 30°C, 12 h) to obtain germinated arthroconidia, resulting in germ tubes that were approximately two to five times the length of the conidia. Germ tube formation was monitored microscopically. The germlings were harvested by centrifugation (2,000 × g, 4°C, 10 min) and transferred to a sterile, transparent 50-ml polypropylene tube. The germinated arthroconidia were resuspended in 20 ml of osmotic medium (OM; 10 mM Na-phosphate buffer [pH 5.8], 1.2 M MgSO₄), centrifuged (2000 × g, 4°C, 10 min), and again resuspended in 20 ml of OM. This washing step was repeated twice and followed by suspension of the germinated conidia in 5 ml of OM contained in the 50-ml polypropylene tube.

Isolation of homokaryons. Since the protoplasts of C. immitis prepared for these and earlier investigations (24) were shown to be multinuclear, it was reasonable to assume that at least some of the transformants produced would be heterokaryons. To obtain homokaryons, the isolated transformants were grown on GYE agar plates containing 100 µg of hygromycin/ml for 10 days, followed by three repeated streak-outs from single colonies of sporulating cultures. The final isolated colonies were transferred to separate hygromycin-containing GYE agar plates as described above. Transformants obtained by this procedure were then subjected to PCR, Southern, and immunoblot analyses.

PCR screening procedures. To obtain DNA for PCR screening, approximately two inoculating loops of fungal mycelia were isolated from plate cultures of each of the putative transformants or the parental strain and transferred separately to 1.5-ml microcentrifuge tubes containing 100 mg of glass beads (0.45 to 0.55 mm in diameter) and 250 µl of TE buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA). The fungal cells were homogenized using a Mini-Beadbeater (Biospec, Bartlesville, Okla.) at 3,000 rpm for 60 s, incubated at 100°C for 5 min, and then centrifuged (16,000 × g, 5 min). The DNA present in the supernatant was extracted with phenol-chloroform and precipitated with ethanol in accordance with standard protocols (34). The transformants were screened for the desired integration event by PCR using primers which amplified a specific 1.5-kb product only if the plasmid construct integrated into the C. immitis chromosomal DNA by homologous recombination. For this purpose, the upstream primer (primer C in Table 1) was located within the pAN7.1 vector sequence (see Fig. 1), whereas the downstream primer (primer D in Table 1) was located in the 5' region of the CTS1 gene, which was not present in the pcts1 construct. A second primer pair (E and F in Table 1 and Fig. 1) was used to confirm homologous integration based on the same rationale. PCR amplification of genomic DNA isolated from putative transformants in the latter case yielded a predicted amplicon of 1.4 kb. Initial screening for homokaryons among the putative transformants was accomplished by PCR using a primer pair (G and H in Table 1 and Fig. 1) which yielded a 1.3-kb product only if the nondisrupted CTS1 gene was present in the template DNA. Transformant template DNA which displayed the two PCR products (1.5 and 1.4 kb) specific for the desired homologous integration event and the absence of the 1.3-kb PCR product provided initial evidence for the isolation of a homokaryotic disruptant.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Plasmid construct and predicted outcome of pcts1 integration event. A 1.1-kb internal fragment of CTS1 was cloned into the pAN7.1 plasmid. Homologous recombination of the plasmid construct with CTS1 by a single crossover event results in the generation of two incomplete copies of the C. immitis gene (labeled cts1*) separated by the linearized sequence of pAN7.1. The primers used for amplification of the parental CTS1 gene are G and H. The primers used to screen for the incomplete cts1 genes are C and D (3' fragment) and E and F (5' fragment). The probes used for Southern hybridization are Pb1 and Pb2. UTR, untranslated region.

Southern hybridization and selection of probes. Southern hybridizations were performed in accordance with standard procedures (34) using restriction enzyme-digested genomic DNAs of parental and transformant strains. Probes were derived by PCR amplification using the specific primers described below. Template DNA included either pBluescript SK⁺, which contained the full-length CTS1 gene, or pAN7.1 plus the 1.1-kb insert of the CTS1 gene. The PCR-derived probes were conjugated with digoxigenin-11-dUTP (Boehringer, Mannheim, Germany) during the amplification reaction, and digoxigenin was detected with specific peroxidase-labeled antibody as recommended by the supplier. Insertion of the HPH gene present in pcts1 into the C. immitis genome as a single integration event was confirmed by Southern hybridization. The probe was a 392-bp PCR product (Pb1) amplified with primers I and J (Table 1 and Fig. 1), which were derived from the GPD promoter sequence. Confirmation of disruption of the CTS1 gene was performed by Southern hybridization using a 584-bp PCR product (Pb2). This was amplified with primers K and L (Table 1 and Fig. 1), which had been selected from sequences in the 5' region of the CTS1 gene and were not present in the pcts1 construct. A hypothetical restriction map was constructed, assuming homologous integration of pcts1, to predict the size of hybridized fragments in the Southern blot analysis. Restriction enzymes were chosen on the basis of this map and used to digest parental and transformant DNA preparations prior to hybridization with Pb1 or Pb2.

Recombinant CTS1 production and generation of anti-CTS1 polyclonal antibody. A 1.2-kb PCR product which encodes part of the CTS1 protein (amino acids 1 to 411 [29]) was amplified using template cDNA obtained by reverse transcription of total RNA of C. immitis. The RNA was isolated from mycelial-phase C. immitis grown in GYE liquid medium at 30°C for 6 days. High-fidelity Taq DNA polymerase (PCR SuperMix High Fidelity DNA Taq polymerase; Gibco BRL, Grand Island, N.Y.) was used for reverse transcription (RT)-PCR. Isolation of total RNA and the RT reaction were performed as previously described (9). PCR amplification of the CTS1 cDNA was conducted using a sense primer (nucleotides [nt] 427 to 447) derived from the genomic sequence of CTS1 (29). The sense primer included the start codon (AUG) and contained an engineered NdeI restriction site constructed as previously described (9). The antisense primer (nt 1973 to 1953) was also designed on the basis of the CTS1 genomic sequence. The PCR product was isolated as previously described (40) and digested with NdeI and SstI to yield a 1.2-kb cDNA fragment. The SstI restriction site (nt 1973) was previously determined to be upstream from the stop codon of the CTS1 gene (29). The cDNA fragment was subcloned into the E. coli expression vector pET28b (Novagen, Madison, Wis.) by ligation at the NdeI and HindIII restriction sites as previously reported (41). Confirmation of the predicted nucleotide sequence of the pET28b-CTS1 construct was performed by standard sequence analysis as previously described (9). The plasmid construct was used to transform E. coli strain BL21 (DE3) (Novagen), and expression was induced in the presence of isopropyl--D-thiogalactopyranoside (IPTG) as previously reported (40). The recombinant CTS1 (rCTS1) expressed by E. coli contained 20 amino acids at its N terminus and 13 amino acids at its C terminus which represented vector-translated peptides. The predicted molecular size of the rCTS1 fusion protein was 49.3 kDa. The rCTS1 was purified from homogenates of the IPTG-induced bacterial cells by nickel affinity chromatography. The column eluate was separated by SDS-PAGE and electrotransferred to an Immobilon P membrane (Millipore Corp., Bedford, Mass.), and the 49-kDa rCTS1 was excised for amino acid sequence analysis as previously described (9).

Immunoblot analyses. Saprobic-phase C. immitis was grown in modified GYE medium containing 0.1% glucose, 1% yeast extract, and 1% chitin in suspension (purified powder from crab shells; Sigma). Chitin was added to enhance secretion of the native CTS1 protein. After 5 days of incubation on a gyratory shaker (120 rpm, 30°C), the culture supernatant was harvested under aseptic conditions by filtration through a Whatman no. 1 filter. Residual fungal debris was removed from the filtrate by centrifugation (15,000 × g, 20 min, 4°C), and proteins were concentrated 200-fold with a Centriprep YM-10 filtration device in accordance with the protocol of the manufacturer (Millipore). The concentrated proteins were subjected to SDS-PAGE, and immunoblots were prepared essentially as previously described (12) using either mouse or human serum. The murine anti-rCTS1 serum described above and serum from a patient with confirmed coccidioidomycosis were diluted 1:500 in phosphate buffer (pH 7.4). Preimmune mouse serum and normal human serum were used as controls.

ID-CF assay. The immunodiffusion (ID) assay was performed by a method previously described (17) for detection of the CF antigen (CF-Ag). The reference C. immitis CF-Ag and goat anti-CF reference serum were obtained from Meridian Diagnostics (Cincinnati, Ohio).

Chitinase assay. Approximately 10⁸ arthroconidia isolated from plate cultures of the parental strain and the strain with the CTS1 gene disrupted (cts1) were inoculated separately into 500-ml flasks which contained 100 ml of GYE medium plus chitin as described above. The cultures were grown with constant shaking (100 rpm) at 30°C for 5 to 6 days. The mycelia were removed by filtration (0.2-µm pore size). The culture filtrate was dialyzed overnight against sterile distilled water (3,500 molecular-weight [MW] cutoff; BioDesign Dialysis Inc., Carmel, N.Y.) and then concentrated 10-fold against PEG (Sigma; 15,000 to 20,000 average MW). Each sample was then further concentrated to a final volume of 1 ml using a centrifugal filter unit (Ultrafree-4; Millipore; MW cutoff of 10,000). The protein concentration of each sample was determined by the Bio-Rad assay. The same amount of protein (20 µg) for each sample was separated by nondenaturing PAGE (8% polyacrylamide without SDS and reducing agent). Substrate gel electrophoresis was used to detect chitinase activity by the method reported by Tronsmo and Harman (36). Briefly, after electrophoresis, the gel was washed three times (10 min, room temperature) in substrate buffer containing 50 mM NaPO₄ (pH 6.1). The washed gel was immediately transferred to the substrate solution containing 0.1 mM 4-methylumbelliferyl-(-N-N¹-N¹¹-triacetylchitotriose) and incubated for 10 min at 37°C. The gel was irradiated with UV light to visualize the bands of chitinase activity. The same gel was either stained with Coomassie brilliant blue to detect protein bands or subjected to immunoblot analysis as described above.

Evaluation of genetic stability, virulence, and morphogenesis of the cts1 mutant. Arthroconidia (10³ cells) obtained from GYE plate cultures (30 days at 30°C) of the parental and cts1 strains were used to inoculate BALB/c mice (two mice each; females, 8 weeks old) by the intraperitoneal (i.p.) route. Two weeks after challenge, the spleen and lungs of each mouse were separately homogenized, dilution plated on GYE agar plus chloramphenicol (50 µg/ml; Sigma)but lacking hygromycin, and incubated for 3 days (30°C). Dilution plating was performed to determine the relative numbers of CFU derived from organ homogenates of the two groups of mice. Equal numbers of CFU from the parental-strain- and cts1 strain-challenged mice (3 × 10³) were transferred to separate, fresh GYE agar plates in the absence of hygromycin and incubated for 2 weeks for arthroconidium production. Conidia from the two sets of cultures were separately pooled and used to inoculate liquid GYE medium lacking hygromycin. After 3 days of incubation (30°C) on a gyratory shaker, separate pools of total genomic DNA from the parental and cts1 mutant strains were extracted, subjected to restriction enzyme digestion, and examined by Southern hybridization using the 584-bp probe (Pb2) as described above.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Transformation efficiency and identification of transformants. The transformation plasmid construct pcts1 was designed to integrate into C. immitis chromosomal DNA by a single crossover event involving the homologous fragment of the CTS1 gene as shown in Fig. 1. Transformation of 10⁷ protoplasts with 2 µg of ClaI-digested pcts1 plasmid DNA typically yielded 30 to 50 hygromycin-resistant colonies. After initial regeneration of mycelia from protoplasts plated on GYE soft agar plus hygromycin (75 µg/ml), colonies were isolated and transferred to GYE agar plates containing hygromycin at 100 µg/ml. Surviving colonies were then transferred to fresh plates, reisolated, and transferred again in an attempt to obtain homokaryons. The transformants were screened for disruption of the CTS1 gene by PCR as described above. Figure 2 shows the results of a typical transformation experiment. Template genomic DNA from 48 hygromycin-resistant isolates was screened by PCR using two sets of primers (C plus D and G plus H; Table 1). The use of one pair of primers (C plus D) resulted in the amplification of a 1.5-kb fragment of the pcts1 construct if it integrated into the CTS1 gene. The use of the other primer pair (G plus H) resulted in amplification of the 1.3-kb parental CTS1 gene. Four of the 48 transformants were thus identified as candidates which had apparently undergone the desired homologous integration. Results of PCR amplification of template DNAs from three of these putative transformants (10.2, 2, and 3) are compared to that of the parental strain (P) in Fig. 2. The transformants, but not the parental strain, revealed a 1.5-kb amplicon if primers C and D were used. Primers G and H amplified the 1.3-kb CTS1 gene product in the presence of parental genomic DNA only. Absence of the amplified full-length CTS1 gene in chromosomal DNA preparations of the transformants suggested that they were homokaryons. One candidate transformant (10.2), which is referred to below as the cts1 strain, was selected for further analysis.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Results of PCR screening of genomic DNAs of the parental (P) strain and three putative transformant (10.2, 2, and 3) strains. Two separate primer pairs were used (G and H or C and D; Fig. 1 and Table 1) to amplify either the 1.3-kb parental CTS1 gene or the 1.5-kb 3' incomplete cts1 fragment plus part of the plasmid construct, respectively. std., standards.

Confirmation of CTS1 gene disruption by Southern hybridization. To confirm that integration of pcts1 into the C. immitis genome had occurred at a single, homologous site and that the transformant was homokaryotic, genomic DNA of the cts1 strain was extracted, digested with selected restriction enzymes, and hybridized with one of two nucleotide probes. A restriction map was constructed for the 8.0-kb genomic fragment which included the encoding region of the CTS1 gene plus the 5' and 3' flanking regions (Fig. 3A). A restriction map was also constructed for the hypothetical form of this fragment with pcts1 integrated into the CTS1 gene (Fig. 3B). Hybridization of PstI- or BamHI-digested genomic DNA of the cts1 strain with a probe derived from the GPD promoter region of pcts1 (Pb1) revealed single 4.0- and 6.6-kb bands, respectively (Fig. 3C). The size of each band was predicted by the restriction map in Fig. 3B. The Pb1 probe failed to hybridize with the same restriction enzyme digests of genomic DNA derived from the parental strain. These results provide evidence that homologous integration had taken place at a single locus in the chromosomal DNA. The cts1 and parental strain DNA preparations were also digested with ApaI, XhoI, XbaI, or PstI and subjected to Southern hybridization to further test if integration had occurred at the homologous CTS1 site and whether the selected transformant was homokaryotic (Fig. 3C). Southern hybridization in this case was performed with a probe (Pb2) derived from the 3' end of the parental CTS1 gene, a region outside the pcts1 construct (Fig. 3A and B). The results of Southern hybridization of Pb2 with the cts1 (T) and parental (P) DNA digests showed single bands which were each of predicted size. These data provide additional evidence that homologous integration had occurred at a single locus, resulted in disruption of the CTS1 gene, and yielded the homokaryotic cts1 transformant.

View larger version (42K): [in this window] [in a new window]   FIG. 3.   Restriction map of the 8-kb BamHI genomic fragment which contains the CTS1 gene (A), predicted map of the corresponding genomic fragment after homologous integration of the pcts1 plasmid (B), and results of Southern hybridization of genomic DNA using either the Pb1 probe (left) with restriction enzyme digests of the cts1 transformant (T) strain or the Pb2 probe (right) with digests of both the parental (P) and cts1 (T) strains (C). The size of each band in panel C correlates with the predicted size in panels A and B, which provides evidence for homologous integration. The single bands revealed by hybridization of both Pb1 and Pb2 with genomic digests of the cts1 transformant indicate that integration occurred at a single chromosomal locus. Std., standards.

Specificity of anti-rCTS1 serum and immunoblot analyses of cts1 culture filtrate. The 1.2-kb cDNA which encodes part of the CTS1 protein was subcloned into pET28b and expressed by E. coli. As a control, the same E. coli strain was separately transformed with the pET28b plasmid which lacked the C. immitis gene insert. SDS-PAGE separations of total cytosolic preparations derived from bacterial cells transformed with pET28b-CTS1 and induced or not induced with IPTG during growth are shown in Fig. 4. A prominent Coomassie brilliant blue-stained band with a molecular size of approximately 49 kDa was visible in the cytosolic fraction of induced cells. The result of purification of rCTS1 by nickel affinity chromatography is also shown in Fig. 4. The gel-isolated protein was electrotransferred to an Immobilon P membrane, stained, excised, and digested with Lys-C in preparation for amino acid sequence analysis. The Lys-C digest was then fractionated by high-performance liquid chromatography (9), and the selected peptide which was purified to homogeneity was subjected to N-terminal sequence analysis. The sequence obtained was (K)FLLTIASPAGPQNY, which exactly matches the reported sequence of the translated CTS1 gene (amino acids 205 to 219 [29]). The SDS-PAGE-electroeluted rCTS1 was used to immunize mice, and the antiserum was evaluated for specificity by immunoblot analysis (Fig. 4). A single, 49-kDa band was identified in the cytosolic fraction of the IPTG-induced bacterial transformant. Preimmune mouse serum failed to recognize the rCTS1 protein (data not shown).

View larger version (74K): [in this window] [in a new window]   FIG. 4.   SDS-PAGE and corresponding immunoblot (Ib.) analysis of expression of C. immitis rCTS1 by E. coli strain BL21 (DE3) transformed with pET28b-CTS1. The 49-kDa recombinant protein was visualized only in lysates derived from transformed, IPTG-induced bacterial cells grown in vitro. The purified rCTS1 in lane 4, obtained by nickel affinity column chromatography, was used to immunize mice for production of antisera. The specificity of the antisera was confirmed by immunoblot assay of the lysate of transformed bacteria. Std., standards.

View larger version (65K): [in this window] [in a new window]   FIG. 5.   SDS-PAGE and corresponding immunoblot (Ib.) analysis of expression of the native 48-kDa CTS1 in culture filtrates of the cts1 transformant and parental (P) strains. The murine antiserum raised against the purified rCTS1 recognized the native CTS1 only in the filtrate of the parental strain (left). A commercial preparation of the CTS1 CF-Ag was included in the immunoblot analysis as a positive control. Also shown are SDS-PAGE and a corresponding immunoblot (Ib.) analysis of culture filtrates (parental [P] strain versus the cts1 transformant) reacted with serum from a CF antibody-positive patient with confirmed coccidioidal infection (right). The patient antibody (Pt.Ab) recognized a 48-kDa band only in the filtrate of the parental strain. Antiserum raised against purified CTS1 (anti-rCTS1) confirmed the presence of native CTS 1 in the parental strain culture filtrate only.

Loss of CF-Ag from culture filtrates of the cts1 transformant. The ID-CF assay was used to compare the immunoreactivities of culture filtrates derived from the parental and cts1 strains (Fig. 6). Fusion of the precipitin produced by the reference anti-CF serum and CF-Ag reaction with the precipitin produced between the wells which contained the reference antiserum and culture filtrate of the parental strain confirmed the presence of the CF-Ag in the filtrate preparation. The same volume and protein concentration of culture filtrate derived from the cts1 strain and the parental strain were added to separate wells of the ID plate (T and P, respectively, in Fig. 6). No precipitin band was formed between the T and reference wells, suggesting loss of the CF-Ag from the transformed strain. Saline was added to a single well (S) as a negative control.

View larger version (161K): [in this window] [in a new window]   FIG. 6.   Comparison of immunoreactivity by ID of culture filtrates derived from the parental (P) and cts1 transformant (T) strains. The reference wells contained the CF-Ag and goat anti-CF serum. One well contained saline (S) and was used as a negative control.

Loss of CTS1 chitinase activity by the cts1 transformant. Approximately equal amounts of protein derived from 5- and 6-day-old culture filtrates of the parental and cts1 strains were separated electrophoretically under nondenaturing conditions and subjected to substrate gel analysis (Fig. 7). Four fluorescent bands, each indicative of chitinase activity, were visible in the filtrate of the parental strain, with the strongest enzyme activity clearly associated with one of the upper bands in the 6-day culture filtrate. The culture filtrate of the transformant, on the other hand, showed only three fluorescent bands in the substrate gel. The prominent band present in the filtrate of the parental strain was absent from both the 5-day and 6-day culture filtrates of the cts1 strain. Immunoblot analysis of the nondenaturing gel separations of the culture filtrates was conducted using the rCTS1-specific antiserum. The antiserum recognized the prominent band in both the 5-day and 6-day filtrates of the parental strain, which identifies it as the CTS1 chitinase. No bands were recognized by the specific antiserum in filtrate preparations of the cts1 strain.

View larger version (76K): [in this window] [in a new window]   FIG. 7.   Nonreducing (nr) PAGE separation, substrate (Subst.) gel electrophoresis, and immunoblot (Ib.) analysis of filtrates derived from 5-day (5d) and 6-day (6d) mycelial cultures of the parental (P) strain and the cts1 transformant. Two reported chitinases (CTS1 and CTS2 [29]) of C. immitis were identified in the substrate gel by immunoblot analysis (only CTS1 identification using anti-rCTS1 serum is shown). Note the apparent increased CTS2 activity in the enzyme assay of the cts1 filtrate. Additional low-MW bands in the substrate gel have not been identified.

Mitotic stability of the cts1 transformant after animal passage and serial transfer in vitro. Since the cts1 strain was produced by disruption rather than replacement of the CTS1 gene (Fig. 3B), it is possible that the integrated plasmid could be excised and result in reversion of the mutant to the parental genotype. For this reason, we examined the stability of the mutation after animal passage, reisolation, and sequential culture transfer of the cts1 strain to GYE agar plates and liquid GYE medium, both in the absence of the selective pressure of hygromycin. Genomic DNA digests of the parental strain and the serially transferred cts1 strain were subjected to Southern hybridization with the Pb2 probe as previously described (Fig. 8). As in Fig. 3C, the blot was designed to detect both integration of the pcts1 construct at the single homologous locus and the presence of the intact parental CTS1 gene. The sizes of the single bands of the cts1 transformant and the parental strain were identical to those reported for the original cts1 transformant examined in Fig. 3C. No reversion to the wild-type CTS1 gene was detected.

View larger version (87K): [in this window] [in a new window]   FIG. 8.   Southern hybridization of the Pb2 probe with restriction enzyme-digested genomic DNAs obtained from the parental (P) and cts1 (T) strains after mouse passage, reisolation, and serial transfer in vitro in the absence of hygromycin. The sizes of the single bands identified in the cts1 strain are identical to those shown in Fig. 3C, which provides evidence for mitotic stability of the single-crossover integration of the pcts1 construct. Std., standards.

Comparative analyses of virulence and morphogenesis of the parental and cts1 strains. Two groups of 10 BALB/c mice each were challenged by the i.p. route with arthroconidia isolated from GYE agar plate cultures of either the parental or the cts1 strain. Survival plots for the two groups of mice are shown in Fig. 9A. The mean survival times of mice infected with the parental and mutant strains were 14 and 16.5 days, respectively. These values are not significantly different, based on a Kaplan-Meier survival analysis, which suggests that the parental and transformant strains demonstrate comparable levels of virulence. Histological sections of abscesses associated with mesenteric tissues from mice infected with each stain were compared at 12 days after challenge (Fig. 9B and C). Normal spherule segmentation and endosporulation appeared to be present in mice infected with either the parental strain (Fig. 9B) or the cts1 transformant (Fig. 9C).

View larger version (68K): [in this window] [in a new window]   FIG. 9.   (A) Survival plot of mice challenged i.p. with either the parental strain or the cts1 transformant. The mean survival times of the two groups of mice were not significantly different. (B and C) Histological sections of C. immitis-infected tissue from mice at 12 days post-i.p. challenge with the parental strain (B) or the cts1 strain (C). Typical segmented spherules (SS) and endosporulating spherules (ES) are shown in both preparations. Bar, 20 µm.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Disruption of genes in fungal pathogens, such as Cryptococcus neoformans, Candida albicans, and two close relatives of C. immitis, Histoplasma capsulatum and Blastomyces dermatitidis (25), has proved to be a powerful genetic tool for the identification of virulence factors and potential antifungal drug targets (19). Factors which contribute to the asexual reproductive capacity of these microbes in vitro and in vivo are relevant to our understanding of the pathogenicity of medically important fungi. Disruption of a gene that encodes an endochitinase of Saccharomyces cerevisiae resulted in a defect in separation of the daughter from the mother yeast cell at the culmination of the budding process (18). The authors have suggested that the chitin binding region of the enzyme functions in localizing it to the cell wall chitin, where it hydrolyzes chitin fibers that join the daughter and mother cells. We have suggested that certain fungal cell wall-associated hydrolases, including chitinases and -glucanases, can digest structural components of the chitin-glucan wall of C. immitis and thereby play key roles in the morphogenesis of this systemic pathogen (4, 17). Endosporulation in C. immitis is apparently accompanied by digestion of a network of cross walls (segmentation apparatus) that compartmentalizes the cytoplasm of the spherule in preparation for endospore differentiation (5). Chitin is a major component of the segmentation apparatus (5, 10). A secreted 48-kDa chitinase, which has been shown to be an immunodominant, serodiagnostic complement fixation antigen (CF-Ag) of C. immitis (14), was suggested to play a pivotal role in the digestion of the segmentation wall during endospore formation (3, 15). The CF-Ag-chitinase has been cloned (CTS1 [29]), and the recombinant protein expressed in E. coli has been crystallized. The latter undertaking was motivated by speculation that CTS1 is a rational molecular target for a chitinase inhibitor-derived antibiotic (11). The present study was conducted to further test these hypotheses.

Gene disruption in medically important fungi has progressed at a considerably slower pace than the development of gene knockout strategies in S. cerevisiae (30). In fact, transformation systems among the human fungal pathogens have so far been developed only for C. albicans, C. neoformans, A. fumigatus, Wangiella dermatitidis, H. capsulatum, B. dermatitidis, and C. immitis (20). With the exception of A. fumigatus (8), transformation of filamentous fungal pathogens by use of a protoplast method developed for A. nidulans has not been productive. This has necessitated the use of alternative approaches, including electroporation, biolistic transformation (20), and DNA transfer from Agrobacterium to recipient cells (1, 7). There are major problems with these methods of transformation of human fungal pathogens. Chromosomal integration of a gene knockout construct at a single site and its homologous recombination that typically results in disruption of the targeted gene either have not yet been demonstrated (1, 39) or are rare (37). However, here a procedure was developed for transformation of C. immitis which was based on one that had been used successfully to generate mutants of A. fumigatus (28). Protoplasts of C. immitis were prepared from germinated arthroconidia, and the uptake of exogenous DNA was promoted in the presence of PEG-CaCl₂. The pAN7.1 plasmid construct that was introduced into the recipient cells contained the E. coli-derived hygromycin resistance gene (HPH), whose expression is under the control of 5' and 3' signal elements of A. nidulans (32). Protoplast regeneration was conducted in an osmotically stabilized selective medium that contained hygromycin at a concentration which is known to inhibit the growth of the wild-type strain of C. immitis. Transformation efficiency was 24 hygromycin-resistant transformants/µg of transforming DNA. This is comparable to the efficiency reported for Aspergillus spp. (8). Approximately 10% of the hygromycin-resistant colonies (i.e., two to three transformants per microgram of DNA) revealed single, chromosomal locus-specific integration of the plasmid construct that resulted from homologous recombination.

Our transformation strategy involved disruption of the wild-type CTS1 gene by a single crossover event with a 1.1-kb internal fragment of CTS1 carried by pAN7.1. We predicted that the plasmid construct would integrate into the chromosomal DNA by homologous recombination at the site of the CTS1 gene. We linearized the plasmid construct as shown in Fig. 1, based on the assumption that uptake of exogenous DNA across the protoplast membrane would be more efficient in this configuration than in the native, circular plasmid form. In fact, transformation efficiencies in this study using the two forms of the plasmid construct were not significantly different (data not shown). The result of homologous integration of the pcts1 plasmid was predicted to be the generation of two incomplete copies of the CTS1 gene separated by the pAN7.1 vector. We further predicted that the two mutated forms of the CTS1 gene would be nonfunctional, since one is truncated and lacks 369 bp at the 3' end of the open reading frame, as well as the poly(A) tail, while the other lacks the promoter plus 69 bp of the 5' coding region (29). This same transformation strategy has been used successfully for disruption of genes in A. fumigatus (8). However, potential problems associated with this gene disruption approach have been identified, including the possibilities that a truncated peptide is still functional and that the gene disruption construct is unstable under nonselective conditions. Both problems can be resolved by the use of a gene replacement approach (8) in which a selectable marker is inserted within a targeted gene and an essential part of the coding region of that gene is simultaneously deleted (33, 35). We have recently compared the transformation efficiency of the single-crossover gene disruption strategy to the gene replacement approach in knockout experiments with a C. immitis gene that encodes a parasitic cell surface glycoprotein (unpublished data). The transformation efficiency for generation of mutants which occurred by single-locus, homologous recombination was 2.6-fold higher using the gene replacement method than using the gene disruption approach. It is not known whether this improvement in transformation efficiency applies to other genes of C. immitis.

In spite of these potential reservations about the use of the gene disruption strategy, the CTS1 gene was successfully disrupted as predicted, loss of gene function was confirmed, and the mutated gene remained stable after animal passage, reisolation, and serial transfer in vitro in the absence of hygromycin. Results of Southern hybridization confirmed the predicted homologous integration of the plasmid construct and provided definitive evidence that this diphasic fungus is haploid. Immunoblot analyses of the mycelial culture filtrate of the putative cts1 transformant, using either antibody raised against the recombinant CTS1 protein or serum from a CF antibody-positive patient with a known coccidioidal infection, failed to detect the 48-kDa band. Patient sera also failed to recognize the CF-Ag in ID assays of the cts1 culture filtrate. Examination of chitinase activity present in the mycelial culture filtrate of the cts1 strain by substrate gel electrophoresis revealed the absence of a major enzyme band which was prominent in the parental strain and identified by immunoblot analysis as the CTS1 protein. Finally, Southern hybridization studies of digests of genomic DNA isolated from the cts1 strain after animal passage and reisolation in vitro revealed the same pattern of bands as the original, selected cts1 transformant. The mutated gene, therefore, appears to be mitotically stable. On the basis of these data obtained from genotypic and phenotypic studies of cts1, we argue that use of protoplasts for introduction of exogenous DNA into C. immitis and the gene disruption method for homologous integration of the plasmid construct into chromosomal DNA are viable approaches to the transformation of this pathogen. This is the first successful targeted gene disruption in C. immitis.

The virulence of the parental and mutant strains were compared in our murine model of coccidioidomycosis to determine whether disruption of the CTS1 gene resulted in loss or significant reduction of the pathogenicity of the fungus. Our results suggested that the cts1 strain showed no significant change in virulence compared to the parental strain. Histological sections of infected mouse tissues suggested that morphological features of spherule formation, endospore differentiation, and endospore release from ruptured spherules were comparable after challenge with either the parental or cts1 strain. Loss of CTS1 gene function had no apparent effect on endosporulation. A possible explanation for the absence of a morphogenetic phenotype which defines the cts1 strain is that one or more of the other C. immitis chitinases compensate for loss of function of the CTS1 gene. In fact, substrate gel electrophoresis suggested that expression of a high-MW chitinase, identified as CTS2 (29) in Fig. 7, was distinctly elevated in the mycelial culture filtrate mutant compared to the parental strain. The deduced molecular size of the CTS2 chitinase of C. immitis is 91.4 kDa, and it shows 47% amino acid sequence similarity to the S. cerevisiae chitinase reported to play a role in cell separation (18). In both CTS2 and the yeast chitinase, the deduced catalytic domain lies upstream of an amino acid variable region which is adjacent to a serine-threonine-rich domain. The latter has been suggested to serve as a potential site for O-mannosylation in both cases. The Ser-Thr-rich domain is followed by a cysteine-rich, high-affinity chitin-binding region (18, 22, 29). The similarity between the structures of the S. cerevisiae and C. immitis CTS2 chitinases suggests that CTS2 plays a similar role in cell separation during the reproductive cycle. However, preliminary data from RT-PCR analyses of transcript levels of CTS2 during spherule-endospore formation suggested that the gene is constitutively expressed (unpublished data). In contrast, CTS1 expression is sharply increased at the onset of endosporulation. Thus, it is possible that CTS1 and CTS2 function synergistically in C. immitis to modify the structure of the segmentation wall complex of spherules in preparation for endospore differentiation and release from the parasitic cell. In order to test this hypothesis, our strategy will be to disrupt the CTS2 gene in the cts1 host. A second selective marker (i.e., sensitivity of C. immitis to phleomycin) will be used together with the E. coli plasmid (pAN8.1) which carries the phleomycin resistance gene (31). The results of these studies will be the basis of a future report.