Effects of disrupting the polyketide synthase gene WdPKS1 in Wangiella [Exophiala] dermatitidis on melanin production and resistance to killing by antifungal compounds, enzymatic degradation, and extremes in temperature

Background Wangiella dermatitidis is a human pathogenic fungus that is an etiologic agent of phaeohyphomycosis. W. dermatitidis produces a black pigment that has been identified as a dihydroxynaphthalene melanin and the production of this pigment is associated with its virulence. Cell wall pigmentation in W. dermatitidis depends on the WdPKS1 gene, which encodes a polyketide synthase required for generating the key precursor for dihydroxynaphthalene melanin biosynthesis. Results We analyzed the effects of disrupting WdPKS1 on dihydroxynaphthalene melanin production and resistance to antifungal compounds. Transmission electron microscopy revealed that wdpks1Δ-1 yeast had thinner cell walls that lacked an electron-opaque layer compared to wild-type cells. However, digestion of the wdpks1Δ-1 yeast revealed small black particles that were consistent with a melanin-like compound, because they were acid-resistant, reacted with melanin-binding antibody, and demonstrated a free radical signature by electron spin resonance analysis. Despite lacking the WdPKS1 gene, the mutant yeast were capable of catalyzing the formation of melanin from L-3,4-dihyroxyphenylalanine. The wdpks1Δ-1 cells were significantly more susceptible to killing by voriconazole, amphotericin B, NP-1 [a microbicidal peptide], heat and cold, and lysing enzymes than the heavily melanized parental or complemented strains. Conclusion In summary, W. dermatitidis makes WdPKS-dependent and -independent melanins, and the WdPKS1-dependent deposition of melanin in the cell wall confers protection against antifungal agents and environmental stresses. The biological role of the WdPKS-independent melanin remains unclear.


Background
Wangiella [Exophiala] dermatitidis is a polymorphic, dematiaceous [darkly pigmented] fungal pathogen of humans that exists predominantly as a yeast form in vitro, but produces various morphological structures, such as pleomorphic yeast, pseudohyphae, true hyphae, and sclerotic bodies in tissues [1]. The dark pigmentation of W. dermatitidis is considered to be due to deposition of 1,8-dihydroxynaphthalene [1, melanin in the cell wall of the organism [2][3][4]. Melanins are multifunctional polymers that are found in species from all biological kingdoms [5]. Whereas human melanin synthesis is solely catalyzed by tyrosinase, microbes primarily synthesize melanin via phenoloxidases [such as tyrosinases, laccases or catacholases] and/or a polyketide synthase pathway. Pigments derived from acetate via the polyketide synthase pathway are generally black or brown and are generally referred to as dihydroxynaphthalene [DHN] melanins. Eumelanins, also called DOPA melanins, are characteristically black or brown and are formed by the oxidative polymerization of phenolic and/or indolic compounds catalyzed by phenoloxidases. Nonetheless, both types of melanins are negatively charged, hydrophobic pigments of high molecular weight [6][7][8]. The production of melanin is associated with W. dermatitidis virulence wherein the absence of the pigment in the cell wall results in increased killing of the fungus by host cells and decreased disease severity [2,[9][10][11][12][13]. Melanins are important for virulence of some other human pathogenic fungi, such as Cryptococcus neoformans, Aspergillus species, and Sporothrix schenckii [reviewed in [14]]. Melanins have also been implicated in the virulence of several fungal plant pathogens [6,15].
Diverse compounds have been shown to interact with melanins, including antifungal drugs [16][17][18]. Traditionally, amphotericin B has been the major antifungal drug used in the treatment of W. dermatitidis infections [19]. However, recently the in vitro efficacy of amphotericin B was shown to be reduced by melanization in C. neoformans [16][17][18], Histoplasma capsulatum [16], and Blastomyces dermatitidis [20]. Although fluconazole and caspofungin are not effective in vivo against W. dermatitidis, melanization can protect C. neoformans and H. capsulatum against these antifungal drugs in vitro [16,18]. Additionally, the activity of microbicidal peptides has been shown to be reduced in melanized C. neoformans yeast cells [21].
The gene WdPKS1, which has been cloned, sequenced and disrupted previously, encodes a polyketide synthase [WdPks1p] in W. dermatitidis that is considered to be the major enzyme controlling melanin production in this fungus [2]. Compared to their wild-type parent, wdpks1Δ mutants are more susceptible to neutrophil killing and less virulent in a mouse infection model, but the gross cel-lular morphology of the wdpks1Δ cells is not affected by the loss of WdPks1p. In this study, we further characterize the structural differences and production of melanins among the wdpks1Δ-1 mutant, the wdpks1Δ-1 mutant complemented with the WdPKS1 wild-type gene, and their intact W. dermatitidis parental strain. We report the surprising finding that W. dermatitidis produces melanin independent of WdPKS1 activity. Additionally, we evaluate the role of W. dermatitidis cell wall melanin in protection against antifungal compounds, heat and cold, and osmotic stress.

Results
Wangiella melanization and isolation of melanin-like particle Yeast cells of wild type strain 8656 and the wdpks1Δ-1-501 complemented strain became black after 3 days of growth at 37°C in YPD broth. In contrast, the wdpks1Δ-1 yeast similarly cultured did not turn black. Instead, after 5 to 7 days of growth in YPD, the wdpks1Δ-1 yeast began to develop a light brown coloration. However, when grown in minimal medium with L-3,4-dihyroxyphenylalanine [L-DOPA], the wdpks1Δ-1 yeast were grey after 3 to 4 days and black after 7 days whereas they were white in the absence of L-DOPA [ Figure 1]. Transmission electron microscopy revealed that the yeast cell walls of strains 8656 and wdpks1Δ-1-501 grown for 7 days in YPD were significantly thicker than those of the wdpks1Δ-1 strain [ Figure 2]. Measurement of twenty non-serial cell sections of each strain showed that the wall thicknesses for strains 8656, wdpks1Δ-1, and wdpks1Δ-1-501 were 227 ± 13 nm, 141 ± 13, and 229 ± 19 nm, respectively [p < 0.001, for strain 8656 compared to wdpks1Δ-1 by Student's t test]. In addition to being thicker, layers of the cell walls of strains 8656 and wdpks1Δ-1-501 were dramatically more electron dense than in the melanin deficient strain.
Treatment of cells from 10-day cultures of the black strains 8656 and wdpks1Δ-1-501 grown in YPD with enzymes, denaturant and hot acid produced cell residues that were similar in size and shape to unextracted yeast W. dermatitidis strain wdpks1Δ-1 after growth for 10 days in a defined chemical medium with [A] and without [B] L-DOPA [ Figure 3A and 3C]. In contrast, treatment of the wdpks1Δ-1 cells grown identically yielded black particles that were significantly smaller than the cells from which they were derived [ Figure 3B]. Instead of the yeast-like forms seen with the black strains, the particles derived from the melanin-deficient strain varied greatly in size and were globular and irregular in shape. However, when grown in defined chemical medium with L-DOPA, treatment of wdpks1Δ-1 yeast resulted in particles similar in shape to their propagules that had grayish outer spherical layer with dense black pigment within [ Figure 1A, inset]. The wdpks1Δ-1 yeast are completely solubilized when grown in the chemical medium without L-DOPA [data not shown]. Melanins from cells grown in YPD for 10 days were subjected to elemental analysis. Quantitative analysis revealed that melanin comprised 16.1 ± 0.4% of the dry mass of wild type strain 8656 yeast cells. The weight of wdpks1Δ-1-501 complemented strain similarly consisted of 15.1 ± 0.2% melanin, whereas the pigment constituted 4.6 ± 0.1% of the wdpks1Δ-1 yeasts' dry mass.

Immunofluorescence analysis of melanin-like particles
The melanin-like particles derived from the three strains were analyzed for their reactivity with a mAb specific for melanin by immunofluorescence. As expected, the particles from strains 8656 and wdpks1Δ-1-501 reacted with the mAb [ Figure 4A,B and 4E,F, respectively]. The smaller particles isolated from wdpks1Δ-1 cells also reacted with the mAb [ Figure 4C,D] suggesting that the particles are comprised of melanin-like compounds. No reactivity occurred when the control mAb or secondary antibody alone were used [data not shown].
Imaging of the 15-day-old cells disrupted with liquid nitrogen revealed that all cells reacted with the melaninbinding antibody [ Figure 5]. The disrupted wild-type cells demonstrated reactivity with the melanin-binding mAb both at the cell wall and within the cytoplasm [ Figure 5B], whereas the cytoplasm of the melanin-deficient wdpks1Δ-1 cells were labeled [ Figure 5D]. Additionally, light microscopy shows that the wild-type cells were significantly more resistant to damage induced by freezing in liquid nitrogen than the melanin-deficient mutant. In fact, the melanin-lacking cells collapsed after liquid nitrogen treatment, whereas those having melanin did not. As expected, cells from the WdPKS1-complemented strain wdpks1Δ-1-501 appeared similar to the wild-type yeast [data not shown]. No reactivity occurred when the control mAb or secondary alone were used [data not shown].

Electron spin resonance [ESR] spectroscopy
An ESR spectroscopy spectrum generated by the black particles collected from strain 8656 W. dermatitidis was consistent with that of a melanin pigment [ Figure 6A] [22], and was nearly identical to the signals generated with particles from C. neoformans [23], P. brasiliensis [24], H. capsulatum [25], S. schenckii [26], Pneumocystis spp. [27], and Scytalidium dimidiatum [28]. Although of lower amplitude, the signal generated with the particles isolated from the wdpks1Δ-1 yeast cells [ Figure 6B] was similar to that seen with the particles purified from the pigmented 8656 and wdpks1Δ-1-501 strains [ Figure 6A and 6C]. The lower amplitude from the melanin isolated from the wdpks1Δ-1 cells probably reflects the fact that there was a smaller quantity of particles available for ESR testing or structural differences of the material.

Elemental analysis of melanins
Elemental quantitative analyses were performed to determine the relative elemental composition of the melanins. A significant percentage of nitrogen was detected in all isolates. The C:N ratios of the particles exposed to Representative transmission electron micrographs of W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 7 days in YPD medium at 37°C Figure 2 Representative transmission electron micrographs of W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 7 days in YPD medium at 37°C. Scale bar, 2 μm.
Representative scanning electron microscopy of particles collected after treatment with enzymes, denaturant, and hot acid from W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 10 days in YPD medium at 37°C Figure 3 Representative scanning electron microscopy of particles collected after treatment with enzymes, denaturant, and hot acid from W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 10 days in YPD medium at 37°C. enzymes, denaturants, and acid were calculated and showed that the ratio for the wild type and complemented isolates were 18.6:1 and 19:1, respectively. In contrast, the ratio from the pigment isolated from the mutant wdpks1Δ-1 yeast cells was 26.2:1.

Quantitative phenoloxidase assay and PAGE analysis for phenoloxidase-like activity
The 2,2'-Azino-bis{3-ethylbenzothiazoline-6-sulfonic acid} [ABTS] assay results demonstrated the presence of phenoloxidase-like activity at the cell surface of the three Wangiella strains. Strain 8656 and the gene deleted strain produced similar results at an OD 420 nm, 0.131 and 0.133, whereas the complemented wdpks1Δ-1-501 strain revealed a lower activity, 0.087. The OD 420 nm determinations for the positive controls were 1.793 and 0.261 for commercial laccase and C. neoformans, respectively. When cytoplasmic yeast extract [CYE] were used, the activity was significantly higher with measurements of 0.527, 0.722, and 0.205 for strains 8656, wdpks1Δ-1, and wdpks1Δ-1-501, respectively. The experiments were repeated with similar results. Additional evidence for the presence of a phenoloxidase-like enzyme is demonstrated by the in situ polymerization of melanin in nondenaturing gels with Corresponding light and immunofluorescence microscopy images of W. dermatitidis yeast cells of strain 8656 [A, B] and wdpks1Δ-1 [C, D] grown for 15 days in YPD medium at 37°C that were exposed to liquid nitrogen for 10 minutes prior to incubation with the melanin-binding mAb 6D2 Figure 5 Corresponding light and immunofluorescence microscopy images of W. dermatitidis yeast cells of strain 8656 [A, B] and wdpks1Δ-1 [C, D] grown for 15 days in YPD medium at 37°C that were exposed to liquid nitrogen for 10 minutes prior to incubation with the melanin-binding mAb 6D2. Scale bar, 5 μm. Figure 7]. Similar results were observed with the commercially available R. vernificera laccase and the enzymatic activity of the CYE was abrogated with KCN.

Mean inhibitory concentration
There were no significant differences in the mean inhibitory concentration [MIC] with the three strains for any of the drugs tested with or without visible melanization of the yeast cells in culture. The MIC determinations for the three strains were 1 μg/ml for amphotericin B, 0.0625 μg/ ml for voriconazole, and 0.5 μg/ml for caspofungin. For W. dermatitidis strain 8656 the MIC with fluconazole was 0.5 μg/ml, whereas the MIC was 1 μg/ml for wdpks1Δ-1 and pks1Δ-1-501. The MICs were within values previously reported for strains of this fungus [29][30][31][32]. The MICs were performed three times with similar results.

Killing assays
Cells from the darkly pigmented strains 8656 and wdpks1Δ-1-501 were significantly less susceptible to amphotericin B than the wdpks1Δ-1 strain [ Figure 8A]. The differences were statistically significant for concentrations of 0.5 and 1 μg/ml, respectively. For voriconazole, the melanin deficient strain was more susceptible at higher [0.125 and 0.25 μg/ml] concentrations of the drug [ Figure  8B]. The wdpks1Δ-1 strain was also more susceptible to NP-1 defensin at concentrations of 0.5 and 1 μg/ml The lysing enzymes were significantly more toxic to the melanin-deficient mutant, having a reduction in survival of 50% or more than the wild-type or reconstituted strains [ Figure 10]. The wild type and the reconstituted strains were not significantly different in their relative susceptibility. Exposure of wdpks1Δ-1 yeast cells to -20°C or 42°C significantly reduced their survival [ Figure 11]. The advantage of the wild-type strain in surviving heat was significantly more pronounced than that measured for cold stress. Interestingly, the reconstituted strain wdpks1Δ-1-501 was more susceptible to both heat and cold compared to the wild-type strain. There was no difference between the survival of the melanin-deficient mutant and the reconstituted strain exposed to -20°C, p = 0.063. In con- Electron spin resonance spectroscopy on particles collected after treatment with enzymes, denaturant, and hot acid from W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 10 days in YPD medium at 37°C Figure 6 Electron spin resonance spectroscopy on particles collected after treatment with enzymes, denaturant, and hot acid from W. dermatitidis strains A] 8656, B] wdpks1Δ-1, and C] wdpks1Δ-1-501 grown for 10 days in YPD medium at 37°C. The spectra wdpks1Δ-1 is noisy because of the higher amplitude setting used to detect the weaker signal generated by the isolated melanin-like material.
Defensin NP-1 time-kill assay Figure 9 Defensin NP-1 time-kill assay. A] The rates of survival of strains 8656, wdpks1Δ-1, and wdpks1Δ-1-501 yeast cells grown for 7 days in YPD medium at 37°C after exposure to various concentrations of NP-1 for 30 min compared to those of fungal cells incubated in PBS. B] Survival of strain 8656 and wdpks1Δ-1 after exposure to 1 μg/ml of NP-1 and of wdpks1Δ-1 after exposure to 1 μg/ml NP-1 incubated with melanin particles derived from the gene deleted strain for 1 h prior to the use of the defensin solution in the assay. P value was calculated by comparing wdpks1Δ-1 cells treated with NP-1 with or without pre-incubation with melanin particles using the Student's t test. Values are averages of the standard errors of the means for three measurements. *, p < 0.01. trast, significantly more wdpks1Δ-1 yeast cells were killed than wdpks1Δ-1-501 yeast cells at 42°C, p < 0.001.

Discussion
Although melanin production occurs in diverse organisms, the mechanisms for melanization at specific locations, such as in the cell wall of fungi, are poorly understood. The regulation of melanin has proven complex in certain fungi, such as in C. neoformans where melanin formation can be catalyzed by at least 2 distinct laccases [33,34] and pathways like the cAMP cascade [33,35] and metal transporters [36] influence melanin synthesis. In W. dermatitidis, disruption of the polyketide synthase gene WdPKS1 results in an melanin-deficient phenotype [2]. However, transforming melanin-deficient strains generated by UV mutagenesis with WdPKS1 fails to restore melanin production in all mutants [2], indicating that there are additional genes or regulatory pathways involved in melanin production in W. dermatitidis. In fact, we have recently described the potential role for a hexapeptide in the synthesis of DHN melanin in W. dermatitidis [37], which suggests that the production of melanin in this pathogenic organism is more complicated than previously thought.
Disruption of WdPKS1 in W. dermatitidis results in a decrease in virulence in mice and cells deficient in the gene are more susceptible to killing by neutrophils [2]. Our current study begins to explore potential reasons for these effects. The cell wall of the late stationary phase wdpks1Δ-1 yeast is nearly 40% thinner than those of the wild type 8656 and the complemented wdpks1Δ-1-501 strain cultured identically [ Figure 3]. Additionally, the melanin isolated from the mutant strain cultured for 10 days in YPD comprised only 4.6% of the dry yeast cell mass compared to 16.1 and 15.1% for the pigmented wild-type and reconstituted strains, respectively. The values for the strains producing cell wall melanin is higher than previously reported [12], however in the prior work the analysis was an estimation by a spectrophotometric method and used 36 to 48-h-old cultures [late log phase]. For comparison, melanin makes up 15.7% of the weight of pigmented C. neoformans cells cultured for 10 days [38]. The cell wall measurement for the wild-type strain is in accordance with prior determinations from transmission electron microscopy studies on W. dermatitidis [39]. The appearance of the cell wall of the mutant wdpks1Δ-1 is similar to a melanin-deficient mutant [3,40] and to thinwalled less-melanized yeast from 1-to 3-day-old cultures [41,42]. studies have shown that MIC is not a sensitive measurement of the ability of melanization to affect susceptibility to antifungal agents and that finding was confirmed here for W. dermatitidis [43]. Nevertheless, killing assays revealed that cells expressing cell wall associated melanin where significantly less susceptible to antifungal agents. The discrepancy between MICs and killing-assays could result from the fact that young daughter cells produced in exponentially growing W. dermatitidis cultures have less melanin than older mother cells or stationary phase cells [41,42], which results in the daughter cells having a greater susceptibility to the antifungal drugs. It has been shown that C. neoformans buds must synthesize melanin de novo [44] and transmission electron micrographs suggest that this is also true for W. dermatitidis [41,45]. Hence, to determine the effect of antifungals on W. dermatitidis cells with and without melanin is best determined by killing assays.
The melanin-deficient wdpks1Δ-1 strain was significantly more susceptible to killing by exposure to lysing enzymes and to heat or cold than were the wild-type strain and the complemented strain [ Figure 10 and 11]. Also, the melanin-deficient cells were more susceptible to damage by rapid freezing in liquid nitrogen [ Figure 5]. These results were not particularly surprising, because we have previously shown that inhibition of melanization of W. dermatitidis yeast by culture with tricyclazole significantly increases killing by zymolyase or glusulase lysis compared to melanized cells [12]. Most likely the melanin binds and inactivates certain hydrolytic enzymes or combines with substrates to protect cells from the enzymatic degradation of their cell walls leading to death by lysis [reviewed in [46]]. Similarly, we have also shown that melanized C. neoformans have significantly greater survival during exposure to extremes in temperature than cells lacking melanin [47]. Thus, the reduced susceptibility of wild-type cells of W. dermatitidis to heat probably similarly reflects the capacity for melanin to absorb heat energy, dissipate it, and temporarily shield against heat damage [reviewed in [5]]. Melanin in the cell wall may also protect against freezing temperatures by providing mechanical stability to the cell against shearing stress produced by ice crystals.
In some fungal pathogens of plants, melanin is essential for the ability of the fungal cells to resist high internal pressures [reviewed in [14]]. Interestingly, invasive growth in W. dermatitidis is dependent upon melanin biosynthesis [48]. While it is not clear why the reconstituted wdpks1Δ-1-501 strain was more susceptible than the wild type to injury from heat and cold, we speculate that the insertion of 13 kb of extra genetic information during complementation may have resulted in a positional effect on downstream genes or in a chromosomal rearrangement that increased sensitivity to temperature shifts [2].
The residues isolated from the wild type and complemented strains after treatment with enzymes, denaturant and hot acid were similar in size and morphology to the cells from which they were derived, whereas the nondescript particles from the wdpks1Δ-1 yeast were notably smaller [ Figure 3]. Nonetheless the melanin-binding antibody reacted with the residues from all three types of cells [ Figure 4], indicating that even the small particles from wdpks1Δ-1 were comprised of a melanin-like material. When cell wall integrity was perturbed by liquid nitrogen and then incubated with the melanin-binding antibody, melanin-like compounds were only detected within the cytoplasm of the melanin-deficient mutant, whereas cells of the wild type and reconstituted strains similarly treated were labeled at the cell wall and in the cytoplasm [ Figure  5]. Support for our contention that the small debris and the cell-sized particles were melanin-like is provided by the ESR analysis that revealed the presence of stable free radicals in the material isolated from the three W. dermatitidis strains [ Figure 6]. ESR spectroscopy has been used previously to study and define melanins based on the properties of unpaired electrons present in melanin [22] and it has also been used to identify pigments as melanins in several fungi, including C. neoformans [38] and P. brasiliensis [24]. Our data strongly supports the conclusion that the small particles isolated from the melanin-deficient wdpks1Δ-1 strain are melanin-like compounds. One possibility is that precursors for melanin exist within the fungal cells and that they are polymerized during the isolation procedure. However, we determined that the CYE of the strains of W. dermatitidis used in our studies had phenoloxidase activity capable of catalyzing the formation of melanin from a phenolic precursor. We also show that strain wdpks1Δ-1 is capable of melanization in the presence of L-DOPA [ Figure 1]. The final step of DHN synthesis is purported to be an oxidation reaction, which could be accomplished by a phenoloxidase. In fact, based on biochemical studies n the 1980s, it has been suggested that a laccase was responsible for the polymerization and oxidation of DHN into melanin [6]. Interestingly, phenoloxidase activity has been demonstrated in A. fumigatus [51] and genes homologous to laccases have been indentified in A. fumigatus [52] and A. nidulans [53]. Our results suggest that a phenoloxidase that may be involved in DHN melanin synthesis can additionally produce WdPKS-independent melanin in W. dermatitidis.
There are numerous electron dense granules within the cytoplasm of W. dermatitidis cells [ Figure 2] and it is possible that these granules may be polymerized melanin precursors or shunt metabolites. These granules could then coalesce into larger spheres during the melanin isolation process or become compartmentalized within vacuoles in the cytoplasm. Several studies of W. dermatitidis by transmission electron microscopy have revealed the presence of larger "dense bodies" that are electron opaque vacuoles typically less than 0.5 μm in diameter [39,41,42,45,54,55]. The micrographs obtained during our study with 7-day-old cells do not have well represented vacuoles. However, the visualization of these structures varies greatly with the method of preparation [55] and the age of the cell [41,42,45]. It is noteworthy that deposition of melanin in electron dense intracellular organelles ["melanosomes"] has been demonstrated in Fonsecaea pedrosoi [56,57] and similar structures have also been reported in Cladosporium carrionii and Hormoconis resinae [58]. Abolition of DHN melanin formation via the polyketide synthase pathway in F. pedrosoi has recently allowed for the identification of a second type of melanin in this fungus [59]. Melanin production by the addition of phenolic substrate after abolition of the polyketide synthase pathway has been demonstrated in other fungi, such as Thielaviopsis basicola and Verticillium dahliae, resulting in pigment formation that differs from that produced in the absence of blocking the polyketide synthase [60,61]. Furthermore, although visually albino, spherical melaninlike particles have recently been isolated from C. albicans yeast [62]. These findings are consistent with the identification of small melanin-like particles in W. dermatitidis wdpks1Δ-1 yeast and suggest that the gene defect halts the polymerization of melanin in the fungal cell wall but not the synthesis of melanin within the cytoplasm or vacuoles of the cells. Consistent with this is the finding that the activity of the phenoloxidase-like enzyme is substantially greater in the cytoplasm of W. dermatitidis compared to the cell surface. The low level of the enzymatic activity on the cell surface may account for the light brown coloration achieved in the wdpks1Δ-1 cells grown in YPD, which contains small amounts of phenolic compounds. The light brown cells do not produce enough melanin to form polymerized shells in their walls, since cell shape is not maintained in the isolation of melanin from the melanindeficient yeast. When grown in the presence of L-DOPA, the gene deficient yeast cultures are visibly black and particles derived from them by the melanin isolation procedure maintain the size and shape of the yeast, though there is more pigment produced within the cell than at the cell surface. Possibly when the genome of this pathogen becomes deciphered, the precise nature of the additional pathways for the synthesis of melanin will become evident.
In summary, we have shown that W. dermatitidis produces more than one type of melanin. We have also documented that the deposition of the dominant WdPKSdependent DHN melanin polymer in the cell significantly improved the pathogen's capacity to resist diverse stressors. Although the role of intracellular melanin-like compound(s) remains uncertain, our results suggest that they continue to be an area ripe for future investigations.

Isolation of melanin particles
Melanin particles were isolated from yeast grown for 10 d by a modification of a described methodology [49]. Briefly, cells were collected by centrifugation, autoclaved, washed with PBS, air dried, and the pellets were weighed. The cells were also suspended in 1.0 M sorbitol-0.  [24,49]. Chitin is also solubilized by this treatment [24].

Electron spin resonance spectroscopy
Electron spin resonance [ESR] spectroscopy was performed on particles isolated from yeast cells grown in YPD media as described [23], except that a Gunn diode was used as the microwave source.

Elemental analysis
Carbon, nitrogen, and oxygen analyses on lyophilized melanin samples were performed by Quantitative Technologies Inc. [Whitehouse, NJ]. Briefly, melanin samples were converted into gases such as CO 2 , H 2 O, and N 2 by combustion. The product gases were separated under steady-state conditions and the percentage of each element in the samples was measured as a function of thermal conductivity. Ratios of C:N were calculated by dividing the percentage of each element in the samples by their respective atomic weights.

Polyacrylamide gel electrophoresis [PAGE] analysis for phenoloxidase-like activity
The phenoloxidase-like activity of cytoplasmic yeast extract [CYE] for catalyzing the polymerization of melanin from L-DOPA was determined as described previously [23]. Briefly, W. dermatitidis yeast cells were collected, suspended in 0.1 M Na 2 HPO 4 with protease inhibitor, and treated for 6 min in a bead beater at 2-min intervals alternating with 5 min on ice. Supernatants were separated from cellular debris by centrifugation at 8,000 rpm for 10 min. Supernatants and commercial laccase were separated by 10% PAGE electrophoresis run at 18 mA overnight under nondenaturing conditions. As controls, samples were treated with 0.

Time-kill assays
Yeast cells grown for 7 days were suspended in sterile normal saline at a density of 2.2 × 10 3 cells/ml. Cell counts were determined by hemacytometer. Microcentrifuge tubes containing 0.1 ml aliquots of an antifungal at 10 times the final concentration were inoculated with 0.9 ml of the yeast suspensions. Final drug concentrations ranged from 0.0625 to 0.25 μg/ml for voriconazole and 0.5 to 2 μg/ml for amphotericin B. After incubation at 35°C for 2 hours, aliquots were plated on YPD agar to determine their viabilities as measured by CFU. Survival was compared to that of fungal cells incubated in PBS.
The defensin NP-1 [gift of R. Lehrer, Los Angeles, CA] was also used in time-kill studies. NP-1 is a defensin derived from the neutrophils of rabbits with the sequence of VVCACR-RALCLPRERRAGFCRIRGRIHPLCCRR-NH 2 .
Final concentrations of peptide tested ranged from 0.5 to 10 μg/ml and the exposure period was shortened to 30 min. To evaluate whether the presence of melanin could influence the outcome of the interaction of the defensin with W. dermatitidis, the gene-deleted strain was also exposed to a solution of NP-1 pre-incubated with 20 μg of melanin particles derived from strain wdpks1Δ-1 yeast cells.
Enzymatic degradation assay W. dermatitidis yeast grown for 7 days were washed in PBS and incubated for 2 h at room temperature with 5 or 10 mg/ml of lysing enzyme [from Trichoderma harzianum; Sigma Chemical Corp., Cleveland, OH], which contains protease, cellulose and chitinase activites. The cells were washed twice in PBS and then plated for CFU. The survival was determined relative to CFUs obtained with cells incubated in PBS alone.

Heat and cold exposure
After 7 days of growth, yeast cells were washed in PBS and either incubated in a 42°C water bath for 1 h or frozen at -20°C for 24 hours. The frozen cells were thawed at room temperature. Aliquots were plated onto YPD agar plates subsequently incubated at 37°C. Survival rates were determined by counting the number of colonies relative to those of non-exposed cells [control]. Control cells were plated at the same time the experimental cells were heated or frozen to avoid changes in cell number.