Nonsex Genes in the Mating Type Locus of Candida albicans Play Roles in a/α Biofilm Formation, Including Impermeability and Fluconazole Resistance

The mating type locus (MTL) of Candida albicans contains the mating type genes and has, therefore, been assumed to play an exclusive role in the mating process. In mating-incompetent a/α cells, two of the mating type genes, MTL a1 and MTLα2, encode components of the a1-α2 corepressor that suppresses mating and switching. But the MTL locus of C. albicans also contains three apparently unrelated “nonsex” genes (NSGs), PIK, PAP and OBP, the first two essential for growth. Since it had been previously demonstrated that deleting either the a/α copy of the entire MTL locus, or either MTLa1 or MTLα2, affected virulence, we hypothesized that the NSGs in the MTL locus may also play a role in pathogenesis. Here by mutational analysis, it is demonstrated that both the mating type and nonsex genes in the MTL locus play roles in a/α biofilm formation, and that OBP is essential for impermeability and fluconazole resistance.


Introduction
The a and a copies of the mating type locus (MTL) of Candida albicans contain several genes that play highly specialized roles in the mating process, MATa1 and MATa2 in the a copy, and MATa1 and MATa2 in the a copy [1]. This locus also contains three genes apparently unrelated to mating, the nonsex genes (NSGs) PIK, PAP and OBP [1][2][3]. Other members of the Candida clade of the hemiascomycetes also contain these NSGs in their MTLs [2], but members of the Saccharomyces clade, including Saccaromyces cerevisiae, do not [4]. Interestingly, the mating type loci of several other fungi contain NSGs [5][6][7][8]. The mating type locus of Cryptococcus neoformans, for instance, contains approximately 20 genes, including mating type genes, mating-related genes and NSGs, both essential and nonessential [5,6]. However, there is no evidence in these other fungi that the NSGs serve related functions or play roles in mating. In C. albicans, the NSG PIK is an essential gene that encodes a phophatidylinositol kinase involved in signal transduction [9][10][11][12]. The NSG PAP is an essential gene encoding the poly(A) polymerase, which polymerizes adenosine (A) at the 39 ends of mRNAs [13][14][15][16]. The NSG OBP is the only nonessential gene, encoding an oxysterol binding protein involved in sterol sensing in the cytoplasm [17][18][19][20][21]. In C. albicans, PIK, PAP, and OBP are not only located in the MTL locus [3], but the a and a alleles of each have diverged at rates far higher than those of other genes bordering the MTL locus, a characteristic of NSGs in the mating type loci of other fungi [6][7][8].
Until recently it had been assumed that in mating-incompetent a/a cells, the mating-related genes in the MTL locus play exclusive roles in repressing white-opaque switching and mating [22]. No attention had been given to the possibility that the three NSGs played a role in mating or in a common function unrelated to mating. Their known functions predicted no apparent direct role in the mating process. Two of the three genes, PIK and PAP, were demonstrated to be essential for growth in S. cerevisiae [9,13] and, therefore, presumably essential in MTL heterozygous and homozygous configurations of C. albicans. The mating-type genes in MTL, on the other hand, were all nonessential for growth, but MTLa1 and MTLa2 did function to suppress mating in a/a cells [1,23]. In S. cerevisiae, there were indications that the orthologs of C. albicans MTLa1 and MTLa2 also played non-mating roles in wall maintenance during the stress response [24] and that MATa2 played a role in invasive growth [25]. In C. albicans there were also indications that genes in the MTL locus played nonmating roles. The configuration of the MTL locus was implicated in virulence in a mouse model for systemic infection. First, it was observed that natural a/a strains were on average more virulent than natural a/a or a/a strains [26]. Second, it was observed that in every case in which an a/a or a/a derivative spontaneously emerged from a natural a/a strain during in vitro culturing, the parental a/a strain was more virulent than the MTL-homozygous offspring [26]. Finally, it was observed that deleting either MTLa1 or MTLa2 in unrelated a/a strains resulted in decreased virulence [26]. These results indicated for the first time that the mating type locus and mating type genes might play nonsex roles in the virulence of mating-incompetent a/a cells. Given that the three NSGs were also located in the MTL locus, we entertained the possibility that they may also contribute to the increased virulence and pathogenesis of a/a strains.
Here we have investigated this hypothesis by analyzing the effects of independently deleting, in two unrelated natural a/a strains, the a or a copy of the MTL locus, MTLa1 or MTLa2, all three a alleles of the NSG simultaneously, all three a alleles of NSGs simultaneously, each a allele of the three NSGs individually, each a allele of the three NSGs individually and both OBPa and OBPa simultaneously. The results reveal that the sex genes MTLa1 and MTLa2, and the NSGs, located in the mating type locus, play roles in biofilm formation and in the virulence of mating-incompetent a/a cell. In the case of the only nonessential NSG, OBP, analysis of the homozygous deletion derivative revealed that it played no discernible role in the pheromone response or fusion in the mating process, but it did play a role in the formation of a complete a/a biofilm that was both impermeable and drug-resistant, two of the most important traits of an a/a biofilm. Together, these results suggest that in mating-incompetent a/a cells all of the tested genes in the MTL locus play roles in common pathogenic processes that are unrelated to the mating process.

Nonsex genes in the MTL locus
An a/a strain of C. albicans harbors an a copy of the mating type locus, MTL, on one homolog of chromosome 5 and an a copy on the other [1]. The a copy of the locus contains the mating genes MTLa1 and MTLa2, as well as the non-sex genes (NSGas) PIKa, PAPa, and OBPa, and the a locus contains MTLa1 and MTLa2, as well as the non-sex genes (NSGas) PIKa, PAPa and OBPa ( Figure 1A) [1,2]. Three of four diploid species of the Candida clade (C. albicans, Candida dubliniensis, Candida tropicalis) similarly possess the three NSGas in the MTLa locus and the three NSGas in the MTLa locus [2,3] (Figure 1B). In C. albicans mating type and non-mating type genes differ both in the direction of transcription and in their relative position in the opposing a and a copies of the MTL locus ( Figure 1A). A fourth diploid species, Candida parapsilosis, contains the NSGas in MTLa, but the composition of MTLa is unknown [27] ( Figure 1A). Candida orthopsilosis, a close relative of C. parapsilosis, has both a and a loci with NSGas and NSGas [27]. Two additional Candida species (Candida guilliermondii, Candida lusitaniae), for which only haploid strains have been identified, contain either the NSGas in an MTLa locus or the NSGas in an MTLa locus [3] ( Figure 1B).
The DNA homologies of the open reading frames of the a and a copies of PIK, PAP, and OBP in C. albicans are 58, 66 and 66%, respectively, which is in marked contrast to the mean DNA homology of 99.5% computed for the open reading frames of the alleles of 50 genes neighboring the MTL locus (Figure 1 C). The identity of the deduced a and a proteins encoded by PIK, PAP, and OBP are 81, 89 and 91%, respectively, which again is in marked contrast to the mean identity of 99.6% computed for the deduced proteins encoded by a and a proteins of the same 50 genes neighboring the MTL locus ( Figure 1C).
The homology between the allelic promoter regions of nine genes located upstream of MTL and nine genes located downstream were compared with the putative promoter regions of the NSGs. These 18 genes were chosen from the same 50 genes neighboring the MTL locus, because they exhibited the most polymorphic open reading frames of the set. The average DNA homology of the promoter regions (700628 bp) of the alleles of the 18 genes was 96.3%63.1%. The promoter regions of alleles of all 18 genes revealed unambiguous alignment. However, comparison of the 1 kb or smaller upstream intergenic region of the a and a alleles of the NSGs revealed that sequence divergence was too high to support bona fide sequence alignments, and hence, did not allow a measure of homology. These differences indicate that the upstream regulatory regions as well as the open reading frames of the a and a copies of the three NSGs have diverged at far higher rates than genes immediately outside the MTL locus on chromosome 5.

Growth rates of all mutants are normal
To test whether MTLa1, MTLa2 and the a and a alleles of the NSGs played roles in biofilm formation and virulence, we generated the following deletion mutants in each of two unrelated natural a/a strains, P37039 and P37037 [28]: the entire MTLa locus (Dmtla); the entire MTLa locus (Dmtla) [26]; MTLa1 (Dmtla1) [26]; MTLa2 (Dmtla2) [26]; all three of the NSGa alleles simultaneously, Dpika Dpapa Dobpa (Dnsga); all three of the NSGa alleles simultaneously, Dpika Dpapa, Dobpa (Dnsga); PIKa (Dpika); PIKa (Dpika); PAPa (Dpapa); PAPa; (Dpapa); OBPa(Dobpa); OBPa (Dobpa); and both the a and a OBP alleles, (Dobpa/Dobpa) ( Figure 1D). All of the mutants formed colonies on nutrient agar after five days with average diameters similar to those of their respective parent strains (data not shown). All of the mutants of both strains generated in strain P37039 exhibited growth kinetics in suspension cultures containing modified Lee's medium [29] that were indistinguishable from those of the parent strains (Figure 1 E). All of the mutants generated in strain P37039 also exhibited growth kinetics in suspension cultures containing the richer medium YPD [30] that were indistinguishable from those of the parent strain ( Figure 1F). Changes in biofilm formation or virulence by any of the mutants generated for this study could, therefore, not be attributed to a major growth defect.

Author Summary
Most natural strains of the yeast pathogen Candida albicans are diploid and heterozygous (a/a) at the mating type locus (MTL). The MTL locus contains mating type genes and has been assumed to play roles exclusively in the mating process of a/a and a/a cells. In C. albicans, however, the MTL locus also contains three nonsex genes (NSGs), the essential phosphatidyl inositol kinase gene, PIK, the essential poly(A) polymerase gene, PAP, and the nonessential oxysterol binding protein gene, OBP. We demonstrate for the first time that both the mating type genes MTLa1 and MTLa2, and the three NSGs play nonmating roles in a/a biofilm formation and virulence. In addition, we show that the NSG OBP is necessary for impermeability and fluconazole resistance of a/a biofilms. These results demonstrate that nonsex genes as well as two mating type genes embedded in the mating type locus, play related roles in pathogenic processes unrelated to mating in a/a cells. Genotypes of the parent strains and basic mutants used in this study. E. Growth kinetics in supplemented Lee's medium and YPD medium revealed that none of the mutants used in this study had a growth defect. *These strains were previously described [28]; the remaining mutants were generated for this study. doi:10.1371/journal.ppat.1002476.g001 the level of b-glucan released from biofilms after 48 hour [31,32] ( Figure 2E, F). The decreases in thickness were significant (all p values were,2610 4 ) for the Dmtla and Dmtla mutants generated in each of the two strains. The decreases for adhesion were not significant (p values ranged from 0.18 to 0.068), and those for the release of b-glucan close to significant or significant (p values ranged from 0.081 to 0.043). However, the decreases were reproducible within a mutant strain and between mutant strains generated in different natural strains. The differences were consistent with those previously reported between the biofilms formed by parental strains and spontaneous a/a or a/a derivatives [33]. Deletion of either the a or a copy of MTL, however, had no discernible effect on the general architecture of biofilms (data not shown). Laser scanning confocal microscopy (LSCM) and scanning electron (SEM) revealed that after 48 hours, mutant biofilms, like parental P37039 and P37037 biofilms, possessed a basal layer of yeast cells at the substratum and an upper region of vertically oriented hyphae embedded in an acellular matrix (data not shown).

Deletion of MTLa1 or MTLa2 affects only biofilm thickness
We next tested whether deletion of either of the two genes encoding the a1-a2 corepressor of switching and mating, MTLa1 or MTLa2, affected biofilm formation. Deleting either of these genes in both strains P37039 and P37037 had no apparent effect on either adhesion (Figure 2A, B) or the release of b-glucan ( Figure 2E, F). The differences between wild type and mutants were statistically undistinguished (p values were close to 1.0). However, deleting either caused decreases in biofilm thickness ( Figure 2C, D) that were significant (p values ranged from 6610 24 to 10 24 ). These differences were obtained in three independent experiments performed for each of the two parental strains and mutant derivatives, measuring three locations in each of three biofilms of each strain. This provided 27 individual measurements per mutant or parental strain. In every experiment, the thickness of the biofilms formed by the Dmtla1 mutant derivative was 10 to 20% less than that of the parental strain and the biofilm of the Dmtla2 mutant 20 to 40% less than that of the parental strain (data not shown). LSCM and SEM revealed that the thinner biofilms formed by the Dmtla1 and Dmtla2 derivatives still exhibited the basic cell architecture of parental a/a biofilms, including a basal yeast cell layer and a larger upper portion containing vertically oriented hyphae interspersed with an acellular matrix (data not shown).

Simultaneous deletion of all three NSGa or NSGa alleles affects biofilm formation
Deletion of MTLa1 and MTLa2, therefore, did not affect adhesion or release of b-glucan, and the decreases in the thickness of the biofilms formed by Dmtla1 and Dmtla2 cells, although significant, were not as great as those that occurred when the entire a or a locus was deleted in the mutants Dmtla or Dmtla, respectively ( Figure 2C, D). These results suggested that it may have been the deletion of the NSGas or the NSGas in the full MTL deletion derivatives Dmtla and Dmtla that was responsible for the decreases in adhesion and b-glucan release, and a portion of the decreases in thickness observed in the full locus deletions. To test this directly, we generated the triple deletion mutant Dnsga, in which PIKa, PAPa and OBPa were deleted (Dpika Dpapa Dobpa) and the triple deletion mutant Dnsga, in which PIKa, PAPa, and OBPa were deleted (Dpika Dpapa Dobpa), in both the natural strain P37039 and P37037 ( Figure 1D). The homozygous triple deletion of the NSGas and the NSGas was not possible because both PIK and PAP are essential genes [13,19]. The deletion mutants Dnsga and Dnsga generated from both test strains exhibited reductions in adhesion ( Figure 3A, B), biofilm thickness ( Figure 3C, D), and the release of b-glucan ( Figure 3E, F). The reductions in adhesion and in the release of b-glucan were similar to those observed for the deletion mutants of the entire a or a copy of the MTL locus, Dmtla and Dmtla (Figure 2A, B and E, F, respectively). As was the case for the mutants Dmtla and Dmtla, the p values for reductions in adhesion and the release of b-glucan for the nsga and nsga mutants were close to significant or significant (p values ranged from 0.05 to 0.23 for adhesion and 0.04 to 0.13 for the release of b-glucan. However, the decreases were reproducible within a mutant and between mutants generated in different natural strains. The decreases in thickness, however, for the two nsga and nsga mutants were highly significant (p values ranged from 3610 23 to 2610 24 ). The reductions in thickness were, however, highly significant with p values ranging from 10 24 to 6610 24 . The decrcreases in thickness for the nsga and nsga mutants were less pronounced than those for the Dmtla or Dmtla mutants ( Figure 3C, D). That was consistent with the conclusion that Mtla1 and Mtla2 were each partially responsible for biofilm thickness (Figure 2A and B). LSCM and SEM revealed normal general architecture for both Dnsga and Dnsga biofilm (data not shown).

Deletions of individual NSGs affects biofilm formation
To assess which of the three NSGs played a role in biofilm formation, we deleted either the a or a copy of each NSG individually in the two strains P37039 and P37037. In both strains, the deletion mutant of either the a or a allele of each of the three NSGs resulted in a small or negligible decrease in adhesion ( Figure 3A, B) and a small or negligible decrease in the release of b-glucan ( Figure 3E, F). Although the consistency of the decreases in both strains suggested a trend, all of the decreases in adhesion and release of b-glucan for the individual NSG allelic mutants were insignificant (p value.0.05). All of the mutants, however, exhibited decreases in thickness ( Figure 3C, D) that were significant (p values ranged between 10 23 and , 10 24 ). Together, these results demonstrate that each of the a and a alleles of the NSGs play roles in a/a biofilm formation, most notably in biofilm thickness. Because the two mutants for each NSG allele were generated independently in two unrelated strains and exhibited similar defects, we considered complementation tests on all 26 mutants to be unnecessary. We therefore performed complementation tests on only two, Dpika and Dpapa of strain P37039. This was accomplished by adding back wild type PIKa or PAPa, respectively, under the regulation of the TET promoter, to generate strains Dpika-TETp-PIKa and Dpapa-TETp-PAPa. In both cases, addition, of 50 mg per ml of doxycycline, an inducer of the tetracycline promoter, completely, rescued the thickness defect (supplemental Figure S1).

The Dobpa/Dobpa mutant exhibits severe defects in biofilm formation
Both PIK and PAP are essential genes [13,19], so simultaneously deleting both the a and a copy of each to produce a homozygous deletion mutant was unattainable. Because OBP is not an essential gene [17,21], we were able to generate the homozygous deletion mutant Dobpa/Dobpa in both strain P37039 and P37037. The homozygous deletion mutant Dobpa/Dobpa of both strains exhibited far more dramatic decreases in adhesion ( Figure 3A, B), biofilm thickness ( Figure 3C, D) and release of b-glucan ( Figure 3E, F) than either of the heterozygous OBP deletion mutants, Dobpa or Dobpa, the heterozygous deletion mutants of the other two NSGs (Dpika, Dpika, Dpapa and Dpapa), the deletion mutants for all three a or a copies of the NSGs (Dnsga and Dnsga) and the deletion mutants of the a or a copy of the MTL locus, (Dmtla and Dmtla). The decreases in adhesion were clost to significant or significant (p values were 0.06 and 0.01, respectively) and decreases in the release of b-glucan significant (p value were 0.01 and 0.002) The decreases in biofilm thickness were highly significant (both p values were ,10 24 ).
The biofilms that were formed by the mutant Dobpa/Dobpa, however, still retained the general architecture of parental wild type biofilms. LSCM of Dobpa/Dobpa biofilms revealed that they possessed a basal layer of yeast cells just above the substratum ( Figure 3H) and vertically oriented hyphae 20 mm above the substrate ( Figure 3H), similar to that of the parental strains ( Figure 3G). The cell density of the mutants at the substratum and 20 mm above it, however, was lower than that of wild type biofilms (compare Figure 3G and H).
SEM revealed no differences between biofilms formed by the two parental strains and the deletion mutants Dmtla, Dmtla, Dnsga, Dnsga, Dpika, Dpika, Dpapa, Dpapa, Dobpa and Dobpa (data not shown). There was, however, a clear difference between the biofilms of the parental strain and that of the mutant Dobpa/ Dobpa. Dobpa/Dobpa mutant biofilms had far less densely packed hyphae resulting in noticeable gaps in the upper hyphal region ( Figure 3J). Biofilms of the parental strains lacked these large gaps ( Figure 3I). Together, these results demonstrate that although deletion of both alleles of OBP had no effect on growth in liquid medium ( Figure 1E) or on the general architecture of the biofilms they produced, it did have dramatic effects on early adhesion, thickness, the release of b-glucan ( Figure 3A through F), and the density of hyphae in the upper regions of the biofilm ( Figure 3G, H and I, J).

Biofilm permeability and fluconazole susceptibility
Recently we demonstrated that biofilms formed by a/a strains of C. albicans were relatively impermeable to low and high molecular weight molecules, impenetrable by polymorphonuclear leukocytes and resistant to fluconazole [33]. In marked contrast, a/a and a/a biofilms were permeable to low and high molecular weight molecules, readily penetrated by polymorphonuclear leukocytes and susceptible to fluconazole [33]. We therefore first tested whether the biofilms formed by any of the NSG mutants generated in the a/a strain P37037 became more permeable to Sypro Ruby, which has a molecular weight of 1.6 kDa [34,35]. The biofilms formed by the mutant derivatives Dnsga, Dpika, Dpapa and Dobpa were all slightly more permeable to Sypro Ruby than biofilms of the parental strains ( Figure 4A, B). Percent penetration into the biofilm of the parental wild type strain P37037 was approximately 12%, while that of Dpika, Dpapa and were approximately 28% ( Figure 4A, B). The percent penetration of Dobpa/Dobpa biofilms, however, averaged approximately 80% ( Figure 4A, B). We then tested resistance to fluconazole. While approximately 5% of cells in parental (P37037) biofilms were susceptible to fluconazole, approximately 20% of cells in biofilms of the mutants Dpika, Dpapa and Dobpa were susceptible ( Figure 4C). In marked contrast approximately 40% of cells in biofilms of Dobpa/Dobpa were susceptible ( Figure 4C). These results indicate that PIK, PAP and OBP play roles in a/a biofilm impermeability and fluconazole resistance.
The Dobpa/Dobpa mutant mates normally When NSGs are embedded in a mating type locus, there is always the question of whether they play a role in the mating process. Because only OBP was nonessential for growth, we were only able to test whether an a/a derivative of the homozygous OBP deletion mutant Dobpa/Dobpa could mate with an a/a strain, in this case strain WO-1 [36,37]. a/a derivatives were generated by screening colonies of the a/a homozygous mutant Dobpa/Dobpa grown on sorbose medium for opaque colonies [38,39], which reflect MTL homozygosis, and then by testing for the a/a genotype by polymerase chain reaction [38,39]. Opaque cells of the a/a derivatives of Dobpa/Dobpa mated with opaque cells of the a/a strain WO-1 ( Figure 5). Both the formation of evaginated cells (shmoos) and fusion occurred normally ( Figure 5). These results demonstrate that at least one of the NSGs, the nonessential gene OBPis not required for shmoo formation or fusion in the mating process.

Overexpression of PIK and PAP partially rescues Dobpa/ Dobpa defects
The decreases in adhesion and the release of b-glucan, the dramatic decrease in thickness, the decreases in hyphal density,  and the loss of impermeability and drug resistance by the Dobpa/ Dobpa mutant, demonstrate that OBP plays a major role in a/a biofilm development. But because we could not generate homozygous PIK and PAP deletion mutants due to the essential roles they play in growth [9,13], we were unable to distinguish if OBP alone or if all three of the NSGs were necessary for the above biofilm characteristics. To explore this question, we used an overexpression strategy. If overexpression of PIK or PAP rescued the Dobpa/Dobpa defects in biofilm formation, it would answer two questions, first whether PIK or PAP contributed to a/a biofilm permeability and fluconazole resistance, and second, whether over-expression of one NSG could rescue the defects of the homozygous deletion of another. We therefore placed wild type PIK or PAP under the regulation of the TET promoter, and inserted the construct into one allele of the gene ADH1 [40] in the P37039 mutant Dobpa/Dobpa. Overexpression of the inserted gene was achieved by adding 50 mg per ml of doxycycline at 0 hr followed after 24 hr by addition of 25 mg per ml during biofilm development. Overexpression wild type PIKa returned biofilm thickness to that of wild type, and overexpression of PAPa to a thickness greater than wild type (supplemental Figure S1A). Overexpression, however, only partially rescued impermeability to Sypro Ruby and fluconazole resistance (supplemental Figure S1B and C, respectively). Sypro Ruby penetrated into 5% in P37039 biofilms, but into approximately 80% of Dobpa/Dobpa biofilms (supplemental Figure S1B). It penetrated into approximately 55% of Dobpa/Dobpa-TETp-PIKa biofilms and 50% of Dobpa/Dobpa-TETp-PAPa biofilms (supplementary Figure S1B). Cell death due to fluconazole treatment was approximately 4% in P37039 biofilms, but approximately 43% in Dobpa/Dobpa biofilms. It was approximately 25% in Dobpa/Dobpa-TETp-PIKa biofilms and 20% in Dobpa/Dobpa-TETp-PAPa biofilms (supplemental S1C). Therefore, overexpression of either PIKa or PAPa in a Dobpa/ Dobpa background completely rescued the thickness defect, but only partially rescued the permeability and fluconazole susceptibility defects.

MTLa1, MTLa2, and the three NSGs are constitutively expressed
Using RT-PCR, we compared the expression levels in the natural strains P37037 and P37039, and the laboratory strain SC5314, of MTLa1, MTLa2, and each allele of the three NSGs. Measurements were made after 14 hours of planktonic growth or biofilm formation. MTLa1 and MTLa2, each of the two alleles of the three NSGs were expressed both during planktonic growth and biofilm formation ( Figure 6). The only difference consistent among the three test strains was expression of the PIKa transcript, which showed higher levels in biofilms ( Figure 6). Control experiments using RT-PCR to assess expression in mutants verified that the probes used were specific for the mating type genes and NSG alleles.

Virulence in a mouse systemic model
Finally, we compared the virulence of the select mutants Dmtla, Dmtla, Dnsga, Dnsga and Dobpa/Dobpa with the parental strain P37039, using the mouse tail injection model for systemic infection [26,41]. Host survival increased for every tested deletion mutant ( Figure 7A, B, C). The differences in survival, computed by the log rank test for survival, were significant between P37039 and the derivative mutants Dnsga (p value 0.008) and Dnsga (p value 0.005) ( Figure 7A). The differences in survival were significant, for the pooled data of the repeat experiments ( Figure 7B and C), between P37039 and Dmtla (p value 0.003), between P37039 and Dmtla (p value 0.007) and between P37039 and Dobpa/Dobpa (p value 0.006). These results indicate that in addition to a/a biofilm formation, the MTL locus and at least one specific NSG embedded in it, OBP, play roles in virulence in the mouse model for systemic infection.

Discussion
The majority of C. albicans biofilms that form on catheters, prosthetics and tissues are due to strains that are heterozygous (a/ a) at the mating type locus MTL [42][43][44]. Such biofilms are thick, impermeable to molecules ranging in molecular weight from 300 to 140,000 daltons, impenetrable by leukocytes and resistant to fluconazole [33]. The developmental program for the formation of  these biofilms is complex, resulting in a multicellular structure containing multiple phenotypes embedded in a multimolecular, extracellular matrix. The film is more than 50 times thicker than a yeast cell monolayer. These a/a biofilms are composed of a thin basal layer of highly adherent and cohesive yeast cells, and a thick upper layer of vertically oriented hyphae, originally formed from the yeast cells in the basal layer. These intertwined hyphae are embedded in a multimolecular matrix produced by the cells in the developing biofilm [45][46][47][48]. At the upper edge of a biofilm, yeast cells are formed by the hyphae, which appear to function as a dispersal mechanism [45,49,50]. a/a biofilms are regulated by the Ras1/cAMP pathway, which targets a transcription factor cascade consisting of Efg1RTec1RBcr1 [45,49,50]. Because the majority of biofilms formed in nature must be a/a, elucidating the molecular mechanisms that mediate their resistance to antifungals [33,50,51], their impermeability to molecules in the molecular weight range of antibodies, and their capacity to exclude polymorphonuclear leukocytes, will facilitate therapeutic strategies for this pervasive human pathogen.
Here we have investigated whether the nonsex genes embedded in the mating type locus play common or related nonmating roles in pathogenesis, in particular in the formation of a/a biofilms. We had previously presented evidence that natural strains that were homozygous at the MTL locus at the time of isolation, derivatives of natural a/a strains in which the a or a copy of MTL was specifically deleted and derivatives of natural a/a strains that had undergone spontaneous homozygosis in vitro at the MTL locus, were on average less virulent in the mouse tail-injection model for systemic infection than a/a strains [26]. In addition, we had demonstrated that although a/a or a/a cells biofilms were architecturally similar to a/a biofilms, they were thinner, far more permeable to low and high molecular weight molecules, far more susceptible to fluconazole, and readily penetrated by polymorphonuclear leukocytes. Therefore, only a/a biofilms exhibit those traits one would expect of a pathogenic biofilm. Moreover, MTLhomozygous biofilms were found to be regulated by the MAP kinase pathway, in contrast to a/a biofilms, which are regulated by the Ras1/cAMP pathway [33]. Together, these results suggested to us that the genes in the MTL locus, including the NSGs, could play non-mating roles in the formation of pathogenic a/a biofilms and in virulence in the mouse model for systemic infection.
The roles of the MTL genes in adhesion, thickness and bglucan release One might assume that one or more of the NSGs may play roles in mating, as was suggested as a possibility by Fraser et al. [8] for nonessential genes embedded in the MAT locus of Cryptococcus. In C. albicans, we have, however, found that this is not the case for the single nonessential NSG, OBP. The homozygous deletion of OBP in an a/a background had no effect on a-pheromone-induced evagination (''shmoo'' formation) or on fusion with cells of the opposite mating type.
Deleting simultaneously all of the a or the a alleles of the NSGs did, however, had an effect on six a/a biofilm parameters. Individual allelic deletions of NSGs caused significant decreases in biofilm thickness, but had very limited effects on adhesion and bglucan synthesis. Decreases in these two parameters were not statistically significant within an experiment, but were reproducible between experiments for a mutant of a single strain and among experiments for the two mutants of the unrelated parental strains.
Deleting either of the two mating type genes, MTLa1 or MTLa2, also caused a significant decrease in biofilm thickness, but had no discernible effect on adhesion or the release of b-glucan. The differential effects of the mating type genes and NSGs on biofilm parameters was not surprising given that each of the three parameters represented a very different and inherently complex aspect of biofilm formation. First, the increase in adhesion to a silicon elastomer surface, measured after sixteen hours of biofilm development, is a response by the initial yeast cell inoculum and their immediate yeast cell offsprings [31]. No hyphae have formed at the time of that assay. Thickness on the other hand is measured after 48 hours of biofilm development and reflects the extent of vertical hyphal elongation supported by extracellular matrix [52,53]. The release of b-glucan is also measured after 48 hours and has been interpreted as a reflection of matrix deposition [54][55][56]. Permeability is assessed after 48 hours of biofilm development by adding a solution of Sypro Ruby to the top of a living biofilm and then measuring how deep the dye has penetrated after 30 additional minutes [33]. The cell death assay is a similar measure of fluconazole penetration [33]. And finally, we assessed the presence of a basal yeast cell layer, hyphal orientation and hyphal density in 48 hour biofilms using LSCM and SEM. Our results demonstrate that MTLa1 and MTLa2 play selective roles in biofilm thickness, but not in adhesion or the release of b-glucan, whereas the NSGs may play roles in all of these biofilm traits. The observation that overexpression of PIK or PAP fully rescued the thickness defect of Dobpa/Dobpa biofilms, but only partially rescued the permeability and fluconazole susceptibility defects again dissociates biofilm parameters. These results underscore the complexity of biofilm formation and the dissociability of traits, and therefore the need to use multiple and independent measures of the different aspects of biofilm development in analyzing each mutant phenotype. Incomplete rescue by overexpressing PIK or PAP suggests again that the roles of the NSGs are not simply additive, a reasonable expectation given the disperate functions of the three NSGs.
The Dobpa/Dobpa mutants were defective in all biofilm parameters, and the decreases in these parameters were approximately double those measured for any of the heterozygous MTL deletion mutants or for the MTLa1 and MTLa2 deletion mutants. Moreover, the gaps in the hyphal network in the upper regions of Dobpa/Dobpa biofilms were not observed in biofilms formed by the Dmtla1, Dmtla2 or individual allelic NSG deletion mutants. Yet the qualitative aspects of biofilm architecture, which included the general organization of cell types and matrix, was maintained in the Dobpa/Dobpa mutant biofilm. It therefore appears that the mating type gene MTLa1 and MTLa2, and the NSGs PIK, PAP and OBP affect the quality of a biofilm rather than the basic structure.

The role of NSGs in biofilm permeability and drug susceptibility
Two of the most important characteristics of a pathogenic biofilm are impermeability to such molecular challenges as antibodies [33,57] and antifungals (i.e., drug resistance) [33,[58][59][60][61][62], and resistance to white blood cell penetration [33]. The a/a biofilm exhibits all of these characteristics, and for that reason has been distinguished as a ''pathogenic'' biofilm [33]. Here, we have shown that deleting one of the alleles of any of the three NSGs in an a/a strain has a small but reproducible effect on permeability and fluconazole susceptibility, and that deleting both alleles of OBP results in a major increase in both parameters. These results suggest that OBP, as well as PIK and PAP, play major roles in establishing the pathogenic traits of an a/a biofilm.
Total rescue of thickness and partial rescue of impermeability and fluconazole resistance in Dobpa/Dobpa a/a biofilms in which either PIKa or PAPa are overexpressed suggests at least partial additivity or roles in a dependent pathway. In the latter case, PIK and PAP may play roles downstream of OBP. Although the disperate functions of the mating type genes and the three NSGs do not provide any obvious explanation for the related roles the MTL genes play in biofilm formation, it is worth considering the known functions and regulations of NSG orthologs in other systems. OBPs encode a family of proteins involved in the transport and metabolism of sterols [17][18][19][20][21]63]. They have also been shown to serve as regulators of genes involved in lipid metabolism and intracellular signaling [63]. OBP gene products not only bind oxysterols, but also phosphoinositides, presumably an interaction involved in the regulation of sterol binding [21]. PIK encodes a kinase involved in the first step in the production of inositol-1, 4, 5triphosphate [64]. In S. cerevisiae, fractionation studies suggest that the PIK ortholog is involved in a nuclear phosphoinositide regulatory process [64]. The poly(A)polymerase, encoded by PAP, catalyzes 39 polyadenylation of mRNAs, a necessary step in transport and stability [65,66], in the nucleus [66,67]. Hence, there are loose or indirect connections between the three NSGs in other organisms ranging in complexity from yeast to man. There have been, however, no reported experiments performed to date in C. albicans that could be used to explain the mutant and overexpression results functionally linking these three colocalized genes, and the two mating type genes MTLa1 and MTLa2, to the formation of a/a biofilms. Given the importance of a/a biofilms in commensalism and pathogenesis, elucidating such relationships would be highly relevant in the future.

The role of NSGs in virulence in a mouse model
Finally, we tested whether the NSGs contributed to the role the MTL played in virulence in the mouse model for systemic infection [26]. Our results demonstrate that deleting MTLa or MTLa locus, the NSGas or the NSGas, or both copies of OBP results in similar decreases in virulence. Although it is not at all clear that biofilm formation plays any role in this model for systemic infection, it is clear that the MTL locus and the resident NSGs play a role not only in biofilm formation, but also in the virulence of a/a cells, two pathogenic traits apparently unrelated to mating.

Materials and Methods
Animal care and usage was in accordance with the National Institutes of Health guidelines for the use of laboratory animals in biological research. Experimental protocols were reviewed by the Institutional Animal Care and use Committee (IACUC) at the University of Iowa (ACURF no. 1101003).

Strains and media
The names, genotypes and origins of the basic strains and mutant derivatives used in this study are listed in Figure 1D, and the basic strains and mutants used for complementation, mating and overexpression experiments are listed in supplemental Table  S1. Growth conditions were previously described [31,36]. Supplemented Lee's medium [68] was employed for basic growth experiments. MTL genotypes of the parent strains were verified by PCR prior to use in experiments.

Generating hemizygous mutants
All mutants were generated for this study using the pop-out flipper cassette from pSF52A (40) containing the dominant nourseothricin resistance marker CaSAT1. To increase the efficiency of deletion, the Sma1 sites flanking the SAT-flipper cassette were subcloned into the pGEM-T Easy, to generate pGEM2A. All of the primers used are listed in supplemental Table  S1. The genotypes of the mutants employed in this study are presented in Figure 1D. For each hemizygous deletion mutant, the cassettes were constructed in two steps. First, the allele-specific 59 and 39 flanking fusion fragment was generated by PCR. Second, the SmaI-digested SAT-flipper cassette from pGEM2A was insented in between the flanking fragment. The deletion of the alleles of all three non-sex genes of MTLa was created in one-step due to their contiguous arrangement in the locus. Deletion of the NSG alleles of MTLa was carried out in two steps. First, two deletion cassettes were created, one for the PAPa allele and second, for the OBPa-PIKa allelic pair. Deletion cassettes were isolated from the plasmids by digestion with SacI and SphI, then used for transformation of parental strains. The putative transformants were verified by PCR. These transformants were also verified for a/a heterozygosity using the primers listed in supplemental Table S2.

Strains for complementation and overexpression
To test for complementation of the mutants Dpika and Dpapa of strain P37039, PIKa and PAPa, respectively were reintroduced under regulation of the TET promoter (40), using methods previously described in detail [33]. For overexpression of PIKa and PAPa in the mutant Dobpa/Dobpa, the same strategy was employed, using ADH1 (40) as the insertion locus. The primers used to create the constructs are listed in supplemental Table S2.

Generation of mating-competent strains
To test whether the Dobpa/Dobpa strain could mate, homozygous a/a derivatives of the P37039 mutant were generated using sorbose induction [38,39]. The a/a configuration was verified by PCR. Shmoo formation and fusion in suspension cultures of the a/a derivative of Dobpa/Dobpa and the a/a strains WO-1 was performed according to methods previously described [37]. The a/a cells were stained with Alexa 488 (Invitrogen) and the a/a cells with TR/TC Con-A (Invitrogen) according to methods previously described [37].

Measuring biofilm parameters
In brief, adhesion was measured by incubating cells in the wells of a 12 well plates for 16 hour, washing gently, then trypsinizing to release cells for counting [31]. Biofilm thickness was measured after 48 hours of biofilm development by Laser scanning confocal microscopy [31]. b-glucan measurements were done according to the methods of Mitchell and coworkers [32,69] with supernatants of 48 hour biofilms. Measurements of Sypro Ruby permeability and fluconazole susceptibility have recently been described in detail [33]. In brief, 48 hour biofilms were overlaid with Biofilm Tracer Sypro Ruby (Invitrogen) for 30 minutes prior to laser scanning confocal microscopy, as previously described [33], using a Bio-Rad Radiance 2100 MP laser scanning confocal microscope (Bio-Rad, Hermel Hampstead, United Kingdom) equipped with a Mai-Tai IR laser (Spectra-Physics Lasers, Mountain View, CA). Calcofluor was imaged using the Mai-Tai IR titaniumsapphire laser tuned to 818nm. Fluconazole susceptibility was measured as previously described in detail [33]. In brief, 25 mg per ml of fluconazole was added to the top of 48 hr biofilms and incubated an additional 24 hours. Biofilms were disassociated in 20 mM EDTA and cells assessed for viability by Dead Red staining. Scanning electron microscopy was performed by fixing biofilms in 2.5% glutaraldehyde in 0.1M cacodylate buffer. After post-fixation in 1% osmium, the biofilms were dehydrated through a graded ethanol series and further dehydrated by 50% hexamethydisilazane (HDMS) followed by two rinses in 100% HDMS. The biofilms were allowed to air-dry prior to sputter-coating with a 60:40 mixture of gold and palladium. The biofilms were viewed using a Hitachi S-4800 scanning electron microscope.

RT-PCR
The methods used for RT-PCR have been described previously in detail [33,69].

Virulence in a systemic mouse model
The method used to assess the virulence has been described previously in detail [26]. Mice were examined every 12 hours. Mice exhibiting the first sign of morbidity were euthanized by inhalation of saturated CO 2 .