Identification of potential components in the filamentation pathway using an engineered strain that produces a homogeneous hyphal population

We previously demonstrated that Znf2 is the decision maker of filamentation in Cryptococcus and it is sufficient and necessary to direct hyphal formation [7], [11], [28]. Firstly, Znf2 is required for initiating filamentation and the deletion of ZNF2 locks cells in the yeast form [11]. Secondly, the expression level of ZNF2 is positively correlated with the amplitude of filamentation [7], [28]. Thirdly, overexpression of ZNF2 stimulates robust filamentation irrespective of environmental stimulation or pheromone signaling [7]. Lastly, attempts of UV mutagenesis in the znf2Δ mutant background failed to recover any suppressor mutations that could restore hyphal growth (data not shown). By comparison, from the 60,000 mutants that we screened (19 Mb genome, ∼7,000 genes), mutants that could form other cell types (e.g. pseudohypha or shmoo) were isolated. Given the central role of Znf2 in filamentation, we hypothesize that Znf2 coordinates the expression of multiple factors that are critical for hyphal development. Thus, characterization of Znf2 regulon is expected to provide a global insight into the molecular bases for hyphal development.

Filamentation in C. neoformans usually occurs in response to mating cues (Figure 1B) and the morphological state in a mating community is highly heterogeneous (Figure S1). Even in a mature mating colony, the yeast sub-population dominates over the hyphal sub-population (Figure S1). Such heterogeneity in morphotype presents a challenge to identify filamentation-specific molecules. To circumvent this complication, we took advantage of a strain with the ZNF2 gene under the control of the promoter of the copper transporter gene CTR4 (P_CTR4-2-ZNF2) [7]. The proportion of filaments generated by this P_CTR4-2-ZNF2 strain is positively correlated to the degree of copper-limitation, which induces its expression [7]. To eliminate potential noise caused by other cell types (e.g. shmoo cells) generated in response to mating-inducing cues, we cultured this engineered strain under conditions that are known to be extremely mating-suppressive [7]. Under such conditions, the P_CTR4-2-ZNF2 strain generated a nearly homogenous filamentous population (Figures 1B and S1).

We then compared the transcription profiles of the two populations with homogeneous morphotypes cultured under the above mentioned conditions: the wild type (yeasts) and the P_CTR4-2-ZNF2 strain (hyphae) (Figure S1). The genes that were differentially expressed between the two morphotypes (populations) were selected. A much larger number of differentially expressed genes were identified with a wider range of expression amplitudes by this current approach compared to the previous data generated from a comparison between the wildtype and the znf2Δ mutant in response to mating stimulation (Figure S1) [11]. This current approach of using the two morphologically homogenous populations allowed us to identify 184 genes that exhibited more than 3 fold differences in the expression level (Figure 1C and Table S1). In comparison, only 24 genes above that threshold were identified by our previous approach [11]. Although different culture conditions were used, we believe that the lower sensitivity of the previous approach was largely attributable to the morphological heterogeneity of the mating-colony (Figure S1). Thus, this current approach provides us with a more comprehensive regulon of Znf2.

Functional prediction of this expanded regulon indicates that genes related to secretion and cell wall organization are the two functional groups overrepresented in the filamentous state (Figure 1D). As expected, some genes that are highly induced by Znf2 were also identified and confirmed in our previous study [7]. A gene ontology analysis of the new data by the DAVID program [29] revealed additional biological processes associated with Znf2-stimulated filamentous growth, including polarized growth and stress responses (Figure 1D). Both are also implicated in filamentous growth in other dimorphic fungi [30], [31], [32]. “Response to pheromone” was also identified as a distinctive class, although this class was not found to be statistically significant (Figure 1D). However, given that the condition used for the transcriptome experiment was extremely suppressive to the pheromone sensing pathway or mating [7], we speculate that this identification is biologically significant even though the P-value for this category might be underwhelming. This finding likely reflects the unique feature of Cryptococcus among human fungal pathogens in that filamentation in this organism is typically a cellular response to mating signals. The indication that Znf2, a filamentation master regulator, may actually repress pheromone signaling genes is seemingly contradictory to the above notion. However, it is consistent with an inhibitory effect of Znf2 on the expression of early mating genes and early mating events noted in two earlier studies [11], [12]. Collectively, these observations suggest negative feedback from hyphal growth on the mating stimulation that initiates the hyphal formation. Such prohibition of cell fusion and pheromone-response during hyphal differentiation could be important to confine the cell lineage for the proper development progression.

Pum1 is an important regulator in the filamentation pathway

To assess the biological roles of Znf2 targets in filamentation, we chose to overexpress part of the Znf2 regulon (29 genes) that encode potential regulators or extracellular products (Figure 2A). To identify genes that are specifically involved in filamentation but not important in activating pheromone signaling, we screened this overexpression library for filamentation on YPD medium, a culture condition that suppresses mating behaviors [7]. We found that overexpression of CNB04870 (serotype D JEC21/XL280 background) resulted in a robust filamentation phenotype under the mating-suppressing condition (Figure 2B). Conversely, the deletion of this gene dramatically reduced filamentation in the hyper-filamentous XL280 background under filamentation-inducing condition (Figure 2B). Thus both the gain of function (overexpression) and the loss of function (gene knockout) demonstrated the significance of this gene in initiating filamentation. Motif searches using the Motif Scan and the InterProScan programs revealed a CLN-rich region at its N-terminus and a Pumilio-family RNA binding domain at its C-terminus (Figure S2A). The Pumilio domain is conserved throughout the eukaryotic domain and its family members are known for their regulatory roles in a wide range of processes, including mitochondrial biogenesis, cell division, development, and differentiation [33], [34]. We therefore named this gene PUM1 (Pumilio protein 1) (Figure S2A). Consistent with the idea that ZNF2 regulates PUM1, the expression of PUM1 was considerably increased by the overexpression of ZNF2 (Figure S2B). To examine the cell type that expresses Pum1 during the development of a mating community, we fused the fluorescence protein mCherry with Pum1 that is driven by the PUM1's native promoter. In agreement with the notion that Pum1 belongs to the Znf2-controled filamentation pathway, Pum1 is expressed in the hyphal population at a significantly higher level than that in the yeast population (Figure 2C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. The Znf2 downstream target Pum1 directs filamentation.

(A) The 29 Znf2 downstream targets were placed under the control of the constitutive P_GDP1 promoter to construct a Cryptococcus overexpression library. The screen for their effects on filamentation was conducted under filamentation-suppressing or filamentation-inducing conditions with a 5-day incubation period. CNAG_0‘number’ indicates the gene locus name used in Broad Institute C. neoformans H99 sequence annotation, and these locus names were used because they are previously uncharacterized genes. The gene locus number of PUM1 is CNAG_04003. (B) Pum1 positively controls filamentation. The overnight cultures of the wildtype strain XL280, the pum1Δ mutant, and the PUM1 overexpression strain of the same cell density were dropped onto the V8 agar (mating-inducing) or the YPD agar (mating-suppressing) and photographed after 5 days of growth. Scale bar: 1 mm. (C) The Pum1-mCherry shows a biased expression in the hyphal subpopulation. The images of the fluorescence labeled wildtype strain were taken at 72 hrs post mating stimulation. The localization patterns of Pum1 in the figure are representative for each given cell types (at least 40 cells for each cell type were examined). See Materials and Methods for the detailed description of the experimental condition for the sub-cellular localization of Pum1-mCherry. Scale bar: 10 µm.

https://doi.org/10.1371/journal.ppat.1004185.g002

It is known that Znf2 and its previously characterized downstream factor Cfl1 do not control filamentation through feedback activation of the upstream pheromone signaling circuit or early mating events [2], [11]. Rather, Znf2 and Cfl1 are dedicated morphogenesis/filamentation regulators [2], [11]. Pum1 is also a filamentation regulator, as overexpression of PUM1 initiated hyphal formation even under the condition that was extremely inhibitory to mating activation (Figure 2B and Figure 3D). The observation that the transcriptional induction of PUM1 was delayed compared to early mating genes (e.g. the pheromone gene MF1α) during mating development (Figure 3A), as we previously observed for ZNF2 and CFL1 [2], [7], led us to speculate that Pum1 is also a filamentation specific regulator and it is not critical for early mating events. Indeed, the overexpression of PUM1 did not increase the expression of MAT2, a key transcription factor that controls pheromone sensing and response [11](Figure 3B). On the contrary, we observed a reduction in expression of the early mating genes that are involved in the synthesis (MF1α), secretion (STE6), and response (STE3) of the pheromone signal in the PUM1 overexpression strain (Figure 3B). This finding echoes the inhibitory effect of Znf2 and Cfl1 on the early mating genes [2], [11] and suggests that Pum1 is likely involved in suppressing the pheromone circuit in an inhibitory feedback manner. Consistently, the deletion of PUM1 enhanced cell fusion efficiency, an early cellular response to mating induction (Figure 3C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Pum1 controls filamentation but not early mating processes.

(A) Transcriptional dynamics of MF1α and PUM1 during mating. The RNA from the a+α mixture co-cultured on V8 agar at 22°C in the dark at indicated time points was subjected to the transcriptional analysis by real time PCR. There was a delay in the transcriptional induction of PUM1 compared to that for the pheromone gene MF1α. (B) Overexpression of PUM1did not increase the transcript level of the genes in the pheromone pathway. The expression level of the corresponding genes in the wildtype strain XL280 grown under the same condition was arbitrarily set as 1 for comparison. (C) Pum1 is modestly inhibitory for cell fusion. The α and a partners from wildtype strains or the pum1Δ mutant marked with different auxotrophic markers were co-cultured on V8 agar or YPD liquid medium for 22 hours at 22°C in the dark. The cocultures were then collected, washed, and plated onto YNB minimal medium to select for fusion products. The cell fusion efficiency of the wildtype coculture was arbitrarily set as 1 for comparison. (D) Overexpression of PUM1 and CFL1 are sufficient to trigger filamentation even under mating-suppressive conditions. The wildtype strain, the P_CTR4-2-PUM1 strain, and the P_CTR4-2-CFL1 strain were grown in the YPD liquid medium. Filamentation and cell aggregation (cell-cell adhesion) were observed in the CFL1^oe and PUM1^oe strains but not in the wildtype strain. Scale bar: 20 µm. (E) The P_GPD1-PUM1 strain was significantly less virulent in a murine intranasal inhalation infection model compared with the wildtype strain XL280 and the pum1Δ mutant (P<0.01). (F) Pum1-mCherry is expressed at a relatively constant level throughout hyphal growth and basidial development (>40 hyphae for each stage were examined). Scale bar: 10 µm.

https://doi.org/10.1371/journal.ppat.1004185.g003

To study Pum1's spatiotemporal regulation of filamentation, we examined the localization of Pum1-mCherry driven by its native promoter throughout hyphal development (Figure 3E). Pum1 exhibited a relatively constant expression level during the progression of hyphal development, and the majority of Pum1 proteins were localized in the cytoplasm. Such subcellular localization of Pum1 is consistent with its predicted function as a mRNA-binding protein [33]. Intriguingly, Pum1 accumulated at the aerial hyphal apexes where basidia form upon the initiation of fruiting body development (Figure 3E). An intense patch of Pum1-mCherry was observed in the basidial heads and the signal gradually became dimmer after spore formation. The spatiotemporal expression pattern of Pum1 implicates its involvement in the regulation of fruiting body formation in addition to its role in filamentation.

Morphotype has long been associated with virulence in Cryptococcus, with the yeast form being pathogenic and the filamentous form attenuated in virulence [11], [28]. Our previous study demonstrated that Znf2 links cryptococcal morphotype and virulence potential. The deletion of ZNF2 locks cells in the yeast phase and enhances virulence while the overexpression of ZNF2 leads to filamentation in vivo and abolishes cryptococcal virulence in the clinical isolate H99 background and significantly attenuates the virulence in the XL280 background in a mouse model of cryptococcosis [7], [28]. Given that Pum1 regulates filamentation, we decided to test the impact of mutations of Pum1 on cryptococcal virulence. Here we examined the wildtype XL280, the pum1Δ mutant, and the P_GPD1-PUM1 strain in the inhalation infection model of murine cryptococcosis. The P_GPD1-PUM1 strain harbors the PUM1 gene driven by the promoter of the house keeping gene GPD1. This strain was considerably attenuated in pathogenicity (Figure 3F). The comparable level of virulence between the wildtype and the pum1Δ mutant was likely due to a low level of expression of Pum1 in the wildtype during infection and it is known that mammalian physiological conditions are extremely inhibitory to the filamentation program [7]. Consistently, the Pum1-mCherry was below detectable level under the conditions that mimic host environments.

Pum1 and Cfl1 represent two branches downstream of Znf2 in controlling aerial structure development

During the development of a mating colony, a portion of cryptococcal cells switch from the yeast to the hyphal form and subsequently generate two types of hyphae: invasive hyphae and aerial hyphae [7]. Invasive hyphae penetrate and grow beneath the solid substrate in search of nutrient while aerial hyphae expand into the air for reproduction and spore dispersal. Znf2 is required for the formation of both types of hyphae [7]. Surprisingly, the deletion of PUM1 considerably reduced the formation of aerial hyphae but exerted a much more modest effect on invasive filamentation under different conditions (Figures 4A, S3A and S3B). Interestingly, the previously characterized extracellular matrix signal protein Cfl1 [2] is also more critical for aerial hyphal development than for invasive hyphal growth (Figure S3A). The similar effect of Pum1 and Cfl1 on the three-dimensional architecture of a mating colony prompted us to further investigate their genetic interactions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Cfl1 and Pum1 represent two major circuits downstream of Znf2 in directing aerial hyphae formation.

(A) Transcriptional analysis by qPCR indicates that alterations of Cfl1 or Pum1 do not affect the transcript level of each other. The expression level of genes in the wildtype strain XL280 was arbitrarily set as 1 for comparison. (B) The production or the release of Cfl1 proteins was not drastically affected in the pum1Δ mutant. The expression of Cfl1 proteins was abolished in the znf2Δ mutant and enhanced when the compatible mating partner was present. The mCherry-labeled Cfl1 was used to measure the Cfl1 protein level by fluorescence microscopy and released Cfl1 proteins was detected by colony immunoblot. (C) Pum1 is dispensable for the expression or the processing of Cfl1. By comparison, the Cfl1 protein expression was below the detectable level in the znf2Δ mutant. This served as a negative control. The asterisk indicates the non-specific band, which is also shown in control strain XL280 without mCherry. (D) Double deletion of CFL1 and PUM1 resulted in a much more severe defect in aerial filamentation during bisexual mating compared to either of the single deletion. (E) Pum1 and Cfl1 represent two major downstream regulatory branches of Znf2 in initiating filamentation. (F) Pum1 is also involved in hyphal extension. The deletion of PUM1 resulted in shorter filaments during bisexual mating. Such a phenotype was not observed in the cfl1Δ mutant. Scale bar: 20 µm.

https://doi.org/10.1371/journal.ppat.1004185.g004

To interrogate the relationship between Pum1 and Cfl1, we first examined that the reciprocal effect of the disruption or the overexpression of one gene on the transcript level of the other. We did not observe any significant changes in expression of one gene when the other gene was altered (deletion or overexpression) (Figure 4B). Thus it appears that Cfl1 and Pum1 do not modulate each other at the transcript level in a major way. We next examined the effect of Pum1 on the abundance of Cfl1 protein within a community. We introduced Cfl1-mCherry under the control of CFL1's native promoter into the wildtype strain, the pum1Δ mutant, and the znf2Δ mutant. As expected, the Cfl1 protein expression was almost abolished in the znf2Δ mutant (Figure 4C). By comparison, the Cfl1 protein expression level in the pum1Δ mutant was only slightly affected, suggesting that Pum1 has no major impact on the protein level of Cfl1. Western blot analyses further supported this conclusion (Figure 4D). Thus, unlike Znf2, Pum1 is not required for the expression of the Cfl1 protein.

We previously established that Cfl1 is secreted extracellularly and that some of the Cfl1 proteins are cleaved and released into the surrounding milieu [2]. The released Cfl1 proteins can induce the expression of endogenous Cfl1 in nearby cells and stimulate their filamentation in a paracrine manner [2]. Thus an impairment of Cfl1's secretion or its release could lead to deficient filamentation, a phenotype observed in the pum1Δ mutant. To assess the impact of the deletion of PUM1 on the secretion and releasing of Cfl1 from the cell wall, we employed the colony immunoblot assay, which detects released proteins from an intact colony [2]. As shown in Figure 4C (bottom panel), released Cfl1 was not detected in the znf2Δ mutant while the deletion of PUM1 only modestly reduced the level of released Cfl1 proteins. Furthermore, some of the Cfl1 proteins were correctly cleaved into a smaller product in the pum1Δ mutant as in the wildtype strain (Figure 4D). Taken together, these observations suggest that Pum1 is not critical for the expression, secretion, or the processing of Cfl1. Because the pum1Δ cfl1Δ double mutant displayed a much more severe defect in aerial hypha development than either of the single mutant alone (Figure 4A), Pum1 and Cfl1 represent two parallel branches modulating Znf2-mediated filamentation program.

To examine if Pum1 and Cfl1 represent the major targets of Znf2 in eliciting filamentation, we overexpressed ZNF2 in the cfl1Δ, the pum1Δ, and the pum1Δ cfl1Δ mutants. The resulting overexpression strains were plated for single colonies under a mating-suppressive condition and were evaluated for the incidence of forming wrinkled colonies. The formation of wrinkled colonies reflects strong cell-cell adhesion and filamentation induced by the Znf2 controlled regulon [7]. The existence of a threshold level of Znf2 appears to control the colony morphological heterogeneity, with the whole population filamentous at a high expression level [7]. A population with a modest level of Znf2 displays a heterogeneous phenotype, likely reflecting the inherent fluctuation in the expression of Znf2's downstream genes in the population [7]. Similar phenomena are observed in another fungal pathogen Candida albicans in terms of the white-opaque switch controlled by the regulator Wor1 [35], [36], and in the bacterium Bacillus subtilis in terms of the motile-chained state switch controlled by the regulator SlrR [37]. In Cryptococcus, the overexpression of ZNF2 in the wildtype strain evoked homogeneous wrinkled colony morphology in the population (Figure 4E), consistent with our previous study [7]. However, when ZNF2 was overexpressed in the pum1Δ mutant or the cfl1Δ mutant, less than a third of the population showed wrinkled colony morphology. Double deletion of PUM1 and CFL1 almost abolished the ability of the population to form wrinkled colony morphology (<1%) in response to the ZNF2 overexpression (Figure 4E). Thus, the fluctuation of the expression level of Pum1 and Clf1 likely contributes to the phenotypic heterogeneity in the population. Taken together, Pum1 and Cfl1 represent the major targets of Znf2 in directing filamentation and cell-cell adhesion in C. neoformans.

Pum1 directs the initiation and the extension of hyphae partially through a secretory protein Fas1

The results presented earlier highlighted the major roles of Pum1 and Cfl1 in directing hyphal morphogenesis. Next we wanted to examine if Pum1 and Cfl1 are functionally redundant or they have overlapping but distinct roles. A closer examination of the hyphal morphology in the absence of CFL1 or PUM1 suggests that Pum1 and Cfl1 have overlapping but distinct roles in hyphal growth. Although both mutants showed a reduction in the abundance of aerial hyphal production, hyphae produced in the pum1Δ mutant were notably shorter than that in the wildtype strain or the cfl1Δ mutant (Figures 4F and S4) during both bisexual and unisexual reproductions. Conversely, overexpression of Pum1 elicited a phenotype of longer hyphae compared with the wildtype (Figure S4). Thus Pum1 is not only important for hyphal initiation, but also for hyphal extension. Interestingly, overexpression of ZNF2 could not bypass the defect of hyphal extension caused by the disruption of Pum1 (Figure S5), indicating that other Znf2 targets could not replace the role of Pum1 in controlling hyphal extension. To discover Pum1-mediated biological processes during hyphal development, we performed the whole-transcriptome microarray analysis of the hyper-filamentous PUM1^OE strain. There were 85 genes exhibiting more than a 3 fold difference in response to PUM1 overexpression (Table S2). In agreement with the RT-PCR result (Figure 4B), CFL1 expression is not affected by the overexpression of PUM1. Gene ontology analyses indicated enrichment of two functional groups among the Pum1 regulon: one involved in meiosis and the other in extracellular proteins. The latter functional group (GO: 0005576) was also identified in the Znf2 regulon (Figure 1D). A comparison between the regulons of these two regulators revealed 3 genes encoding extracellular products [CNAG_00925 (named as FAD1), CNAG_05729 (named as FAS1) and Dha1] that were upregulated by overexpression of ZNF2 and PUM1 (Figure 5A). To evaluate their biological role in filamentation, we deleted these three genes. Only the deletion of CNAG_05729/FAS1 impaired filamentation and recapitulated the phenotypes produced by the disruption of PUM1 (Figure 5B). Conversely, overexpression of CNAG_05729/FAS1 stimulated robust aerial morphogenesis and longer hyphal growth (Figure 5C), confirming its position as the major Pum1 target in quantitatively controlling various aspects of filamentation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Pum1 directs hyphal initiation and extension partially through a novel hypha-specific extracellular protein Fas1.

(A) FAS1, FAD1, and DHA1 are the genes encoding extracellular proteins overlapped in the regulons of Znf2 and Pum1. (B) Fas1 recapitulates the role of Pum1 in unisexual hyphal production. The fas1Δ mutant, like the pum1Δ mutant, showed reduced abundance of aerial hyphae. (C) The indicated strains were grown on V8 agar plate containing 200 µM BCS in the dark at 22°C. Images of the colonies were photographed after 5 days of incubation. Released Fas1 proteins were detected by colony immune-blot. Only overexpression of Fas1 with intact signal peptide led to a more robust filamentation and longer hyphae. Thus, the secretion of Fas1 is crucial for its biological function. (D) Fas1-mCherry is observed on hyphal surface and in vesicles at 72 hrs post unisexual mating stimulation. (E) The FAS1 expression level was positively regulated by Pum1 based on the transcriptional analysis of its transcript level in the pum1Δ mutant and the P_GPD1-PUM1 strain by qPCR. The deletion or the overexpression of CFL1 did not significantly affect the FAS1 expression level. (F and G) Pum1 increased the intensity of Fas1 expression and reduces its stochasticity. The average intensity of Fas1-mCherry fluorescence was dramatically reduced in the absence of Pum1, but not Cfl1. The deletion of PUM1 led to a much higher variability in the Fas1-mCherry level among hyphal cells during unisexual filamentation (Inset of Figure? 5G). The frequencies of Fas1-mCherry's mean intensity in the hyphal population were plotted for the wildtype strain, the cfl1Δ mutant, and the pum1Δ mutant (software: ZEN 2011).

https://doi.org/10.1371/journal.ppat.1004185.g005

CNAG_05729 is predicted to encode a small secreted protein (154 aa) with a signal peptide for secretion at its N-terminus. The mCherry-fused protein was almost exclusively expressed in the filamentous subpopulation during mating colony differentiation (Figure 5D and F), which corroborates its involvement in the filamentation pathway. Thus, we named CNAG_05729 FAS1 (Filament-Associated Secretory Protein 1). To investigate whether Fas1 is specially controlled by Pum1 or/and Cfl1, we assessed the impact on the expression of FAS1 when PUM1 or CFL1 was either absent or overexpressed. Neither the absence nor the overexpression of CFL1 altered the expression level of FAS1. By comparison, the transcript level of FAS1 was positively correlated with that of PUM1 (Figure 5E). At the protein level measured based on the fluorescence intensity of Fas1-mCherry, the deletion of PUM1 led to a marked reduction in Fas1 expression in the hyphal subpopulation (Figure 5F). However, the level of Fas1-mCherry in hyphae was not significantly different in the presence or absence of CFL1, suggesting that Fas1 belongs to the Pum1 pathway but not the Cfl1 pathway. Interestingly, the variation in Fas1 expression among individual hyphae in the absence of Pum1 was noticeably higher, and some hyphal cells in the pum1Δ mutant did not produce any measurable Fas1 proteins (Figure 5F and G). This suggests that Pum1 may mediate a stochastic buffering-like regulatory mechanism [38].

To gain the clue for the molecular functions of Fas1, we employed a variety of bioinformatic tools to identify potential functional domains or motifs harbored in this small protein. However, no annotated domain was identified in Fas1. A PSI-BLAST analysis with regions of low complexity masked suggested that there are Fas1 homologues present in other fungal species, nearly all of them belong to the family of tremellaceae from the phylum of Basidiomycota. This suggests that Fas1 proteins may have arisen relatively recently. Most of the Fas1 homologues contain a signal peptide for secretion and are predicted to be secreted proteins.

We previously showed that the Basidiomycota-specific adhesion protein Cfl1 displays two patterns of subcellular localization: cell surface and intracellular vesicles that are morphologically similar in appearance to known fungal secretory vesicles [39], [40]. Similar localization patterns were also observed for Fas1-mCherry during hyphal development (Figure 5D). Like Cfl1, Fas1 can be also released into extracellular matrix, as released Fas1 from the community was easily detected by the colony immunoblot assay (Figure 5C). The deletion of the N-terminal signal peptide predicted for secretion abolished its secretion (Figure 5C). As expected, the ability of Fas1 to be secreted is critical for its function in enhancing filamentation and extending filamentous growth (Figure 5C).

Pum1 bridges filamentation and meiosis during unisexual and bisexual mating

The aforementioned results demonstrate the major roles of Pum1 and Cfl1 in directing aerial hyphal morphogenesis, a process that could lead to the formation of fruiting bodies (sexual sporulation) critical for species survival and expansion into new niches. The enrichment of meiosis and sporulation genes in the Pum1 regulon in addition to the filamentation genes (e.g. FAS1) raises the possibility that Pum1 may bridge the progression from filamentation to meiosis and to post-meiotic sporulation during sexual reproduction. One meiosis-specific gene that was significantly upregulated in the PUM1^OE strain is DMC1. Dmc1 is the recombinase specifically involved in homologous recombination at an early stage of meiosis. Earlier investigations have demonstrated that DMC1 is important for the formation of tetrad spore chains in Cryptococcus [12], [16]. Overexpression of PUM1 resulted in a greater than 4 fold increase in the DMC1 transcript level whereas the deletion of PUM1 led to a two-fold reduction in its expression (Figure 6A). By contrast, we did not observe any significant impact on DMC1 expression by altering the expression level of CFL1 (Figure 6A). The results reinforce the notion that Pum1 and Cfl1 are functionally different and support our hypothesis that Pum1 may control subsequent development following filamentation through its regulation of factors such as DMC1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Unisexual meiotic progression is only observed in the Pum1-controlled aerial hyphal population.

(A) Pum1, but not Cfl1, affects the expression of DMC1 during unisexual reproduction. (B) Dmc1 expression is associated with a portion of hyphal cells but not yeast cells. Scale bar: 20 µm. (C) Fas1 is exclusively expressed in the hyphal population and it is used as a molecular indicator to visualize the hyphal population in a mating colony. A side-view of a single-cell derived colony (3 days old) shows the topology of Fas1-expressing hyphal population in the community. (D) Details of the morphology and Dmc1 expression for aerial hyphae and root hyphae (invasive hyphae) from the unisexual mating community are shown. Fungal cell wall was stained blue with calcofluor white. Dmc1 expression was exclusively observed in aerial hyphae but not in invasive hyphae (more than 200 invasive hyphae were examined). Scale bar: 10 µm. (E) The dynamics of Dmc1-mCherry monitored during unisexual development. Scale bar: 10 µm. The localization patterns of Dmc1 in the figure are representative for each given stage of development (>40 hyphae for each stage were examined).

https://doi.org/10.1371/journal.ppat.1004185.g006

The function of Dmc1 in the progression of meiosis is extremely conserved among evolutionally divergent eukaryotes [41] and it serves as a molecular marker for the meiotic process [42]. Therefore, we decided to use Dmc1-mCherry controlled by the DMC1's native promoter to identify the cell type associated with Dmc1 and monitor meiotic progression during the development of a unisexual mating community. Dmc1-mCherry was only detected in apical cells in a portion of hyphae and undetectable in yeast cells (Figure 6B and D). To further specify the subpopulation of hyphae associated with Dmc1 expression, we probed the topological pattern of the hypha-specific protein Fas1 in a unisexual community derived from a single cell using confocal microscopy (Figure 6C). The vertical transverse cross section of the colony indicated that the aerial hyphal population was enriched in the peripheral region of the upper colony while invasive filaments radially rooted the colony into the agar (Figure 6C). Interestingly, Dmc1-mCherry was detected only in some of the aerial hyphae, and not in any of the invasive hyphae (Figure 6B and D). A closer examination of aerial hyphae showed that Dmc1-mCherry was not detected in actively growing hyphae with no sign of differentiation into basidia. Rather, an intense signal was detected in the apical compartment of aerial hyphae when their apexes began to differentiate into basidia (Figure 6E). The signal gradually became focused in the basidial head during basidial maturation until four pre-spores started to emerge. The signal disappeared from the basidia when spores matured (Figure 6E). The dynamic expression of Dmc1 mirrors the progression from vegetative aerial hyphal growth to unisexual meiosis and sporulation. Given that mutations of PUM1 affect the transcript level of DMC1 (Figure 6A), we decided to examine the impact of the deletion of PUM1on Dmc1 protein expression. Remarkably, many aerial hyphal apexes differentiating into basidia lost the signal of Dmc1-mCherry in the pum1Δ mutant during unisexual colony development (Figure 7A). For those that had the Dmc1-mCherry signal, the fluorescent intensity was significantly lower (Figure 7B). Again, the loss of Pum1 increased the stochasticity of Dmc1 dynamics in basidia (Figure 7B) as observed for Fas1 in filaments (Figure 5G).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Pum1 coordinates filamentation and sexual reproduction.

(A and B) The deletion of PUM1 impairs the expression of Dmc1 in basidia in a stochastic manner. Some but not all differentiating basidial heads lost the expression of Dmc1-mCherry without Pum1. Those that did expressed Dmc1at lower levels. Scale bars: 20 µm. (C) The unisexual sporulation was perturbed in the absence of the meiosis-specific protein Dmc1 and the global regulator Pum1. The wildtype strain XL280, the dmc1Δ mutant, and the pum1Δ mutant were diluted on V8 agar to form single colonies. Typical mature basidia observed from these strains were shown. Scale bar: 10 µm. (D and E) The deletion of PUM1 and DMC1 also dramatically reduced the sporulation efficiency during bisexual mating. (F) The effect of the deletion of ZNF2, PUM1, or DMC1 on aerial hyphal morphogenesis at the colony level during both bisexual and unisexual development. Scale bar: 1 mm.

https://doi.org/10.1371/journal.ppat.1004185.g007

Previous studies indicated that mutations of meiosis-specific genes perturb sporulation [12], [16]. The effect of such mutations are highly specific, as the deletion of DMC1 or SPO11(an endonuclease creating double-stranded breaks to initiate meiosis) did not affect hyphal initiation, hyphal extension, or the terminal differentiation of some aerial hyphae into basidia (Figure 7F)[12]. The deletion of PUM1 likewise resulted in a severe defect in the formation of basidia and appropriate tetrads of spore chains during either unisexual or bisexual reproduction (Figure 7C, D, E and F). However, disruption of Pum1 also drastically reduced the ability of this fungus to initiate and to maintain hyphal growth (Figures 2B and 4F). These results suggest that Pum1 links sexual reproduction with filamentation, and meiosis per se does not influence earlier morphogenesis events. Although most of the studies were performed in a laboratory strain XL280 background, we confirmed that Pum1 plays a similar role in bridging sex and filamentation in a clinical isolate H99 (Figure S6). Our data presented here indicate that Pum1 plays a conserved role in genetically orchestrating the transition from yeasts to hyphae and the subsequent transition from aerial hyphae to sexual reproduction in this eukaryotic pathogen.

Ethics statement

All the animal experiments were performed according to the guidelines of NIH and Texas A&M University Institutional Animal Care and Use Committee (IACUC permit # 2011–22).

Statistics

The programs Originpro and GraphPad Prism were used for statistical analyses in this study. A two-tailed unpaired Student t test was performed to compare the mean florescence intensity or transcript levels from two groups. P values less than 0.05 were considered statistically significant.

Strains and growth conditions

Strains used in this study are listed in Table S3. For both unisexual and bisexual mating assays, strains were maintained on V8 agar unless indicated otherwise. The optical density of cell suspensions measured at 600 nm was used to represent cell density. When the inducible promoter of the copper transporter gene CTR4 (e.g. P_CTR4-2-FAS1) was used to drive the genes, the gene expression level was manipulated by the addition of 25 µM CuSO₄ (inhibitor) or 200 µM the copper chelator Bathocuproine disulphonate (BCS; inducer) to the media as described previously [7], [53], [54].

UV mutagenesis and suppressor screen for filamentation mutant

Overnight culture of the znf2Δ mutant in YPD liquid medium was plated on YPD agar. The plates were dried and then explored to UV light (300 J/m²) for 3 seconds by using the UV cross linker. The cells were then collected, washed by PBS and plated on V8 agar medium for 7 days in the dark at 22°C to form visible colonies. The colonies were screen under a stereoscope for the formation of hyphae. The mutants with hypha-like morphology were furthered examined with a compound microscope. No true hyphae were observed in any of the mutants screened but mutants with pseudohyphae and/or shmoo cells were observed.

Mating and self-filamentation assays

For bisexual mating assays, a and α cells of equal number (original optical density of 1.5 at OD₆₀₀) were cocultured on V8 juice agar medium (pH = 7.0 for serotype D strains, pH = 5.0 for serotype A strains) in the dark at 22°C [55]. Random basidiospores were isolated using a micromanipulator. For unisexual mating assay (serotype D XL280 strain background), cells were spotted on V8 medium alone. Both bisexual and unisexual matings were examined microscopically for formation of mating hyphae and chains of basidiospores.

Gene knockout and overexpression

Primers used in this study are listed in Table S4. To disrupt the PUM1, FAS1, FAD1, DHA1 or the ZNF2 genes, an overlap PCR products were generated with the NAT or the NEO dominant drug marker amplified from plasmid pAI1 or pJAF1 [11] and 5′ and 3′ flanking sequences (1∼1.5 kb) of the coding genes from strain JEC21α or H99α as we described previously [11]. The PCR product was directly introduced into strains JEC21α, XL280α, JEC20a, or H99α by biolistic or electronic transformation. The resulting deletion mutants generated via homologous recombination were confirmed by PCR and genetic crosses. For overexpression, an overlap PCR product with the NEO or NAT marker and the wild-type genes containing their native promoter (∼1.2 kb) and GPD1 terminator (∼0.5 kb) from strain JEC21 or H99 was generated. The PCR products were directly introduced into appropriate backgrounds as indicated in the text. Mutant strains in the MATa background were obtained by crossing the corresponding α mutants with the widltype congenic a strains. For overexpression, genes (ORF) were amplified by PCR and the amplified fragments were digested and inserted into pXL1 or pXC after the GPD1 or the CTR4-2 promoters respectively as we previously described [7]. Overexpression plasmids were linearized via restriction enzyme digestion before being introduced into the relevant Cryptococcus strains as we previously described [7].

Murine models of cryptococcosis

Animals were intranasally infected as previously described [7], [56], [57]. Nine A/J mice (6- to 8-week-old female; Jackson Labs) per group were used for survival studies. Statistical significance (P≤0.05) of the survival data between different groups was assessed by the Mantel-Cox log-rank test [58].

Cell fusion assay

The coculture of α and a cells with different auxotrophic markers (lys for α; ade2 for a) were inoculated on V8 agar or in YPD liquid medium. After 22 hrs of incubation, the coculture was collected, washed by cold water twice, and plated on minimum YNB agar to select fusion products at 37°C as described previously [11].

Construction of fluorescent proteins and microscopic examination

The mCherry was fused to the C-terminus of the interested protein used in this study. Overlap PCR was used to piece together the fragments including the coding region of the gene with 1∼1.2 kb upstream sequences and the mCherry. The resulting products were introduced into plasmid pXL1 as described previously [7]. The FAS1-mCherry without the promoter was amplified and introduced to pXC to generate the plasmid P_CTR4-2-FAS1-mCherry. To overexpress Fas1-mCherry that lacks the N-terminal signal peptide [Fas1 (sigPΔ)-mCherry], primers Linlab 1581 and Linlab864 were used to generate FAS1(sigPΔ)-mCherry allele. The P_CTR4-2-FAS1-mCherry construct was used as the template. The resulting PCR product was cloned into pXC to produce P_CTR4-2-FAS1(sigPΔ)-mCherry. To examine the sub-cellular localization of morphogenesis or meiosis-related proteins during filamentation, the fluorescence strains were grown on V8 agar medium at 22°C for 72 hrs before being examined microscopically. Images were acquired and processed with a Zeiss Axioplan 2 imaging system with the AxioCam MRm camera software Zen 11 (Carl Zeiss Microscopy). To assess the expression level of P_CTR4-2 -FAS1-mCherry and P_CTR4-2-FAS1(sigPΔ)-mCherry in a community, the overexpression strains were grown on V8 agar medium containing the copper chelator BCS before examined with the fluorescence stereoscope.

RNA purification, qPCR analyses, and microarray analyses

The purelink RNA kit (Invitrogen) and the Superscript III cDNA synthesis kit (Invitrogen) were used for total RNA purification and the first strand cDNA synthesis respectively according to the manufacture instruction. The house-keeping gene TEF1 was used as the endogenous control for normalizing gene expression levels. The relative transcript levels were determined using the comparative CT method as described previously [7].

For transcriptional profiling of different morphotypes, the yeast H99 wildtype strain and the filamentous P_CTR4-2-ZNF2 strain were grown in YPD liquid medium containing BCS at 22°C for 48 hrs. The cell morphology and the proportion of hyphal and yeast populations were measured before the total RNAs were extracted. RNAs were processed for microarray analyses as described previously [7]. For the transcript analysis to identify the Pum1 regulon using microarray, XL280 and the P_GPD1-PUM1 strain were grown on V8 agar medium at 22°C for 72 hrs. Unisexual filamentation and sporulation were confirmed microscopically before extraction of total RNAs. The arrays were analyzed similarly as described previously [11].

Protein extraction and western-blot analyses

For Western Blot analysis of the Cfl1 expression, overnight cultures of indicated strains in YPD liquid medium were spotted onto V8 agar for 3 days at 22°C. The cells were removed and suspended in cold PBS for protein extraction. Protein extraction and western analysis were performed as described previously [2]. Briefly, the cell suspension was subjected to centrifuging at 13,000xg for 10 min at 4°C. The pellet was freeze dried and disrupted. The disrupted cells were subsequently suspended for western immunoblotting analysis of whole-cell proteins. The house keeping protein β-actin was used as the endogenous control.

Colony immunoblot for the detection of secreted Fas1 and Cfl1 proteins

For colony immunoblot assays, 3 microliters of cell suspension with the density of 1.5×10⁷ cells/mL were spotted onto the same plate for the assay as described previously [2]. The strains were grown on V8 agar medium containing BCS at 22°C for 24 hrs to form visible colonies and the plate with the colonies was laid over by a sterile nitrocellulose membrane (Millipore, Billerica, MA). The membrane was removed from the plate after incubation for additional 3 days. The membrane was subsequently washed with 1× Tris-buffered saline (TBS) to remove attached cells. The blots were incubated with the anti-mCherry primary antibody (1/1000 dilution) and then a rabbit anti-mouse secondary antibody (1/10,000 dilution). Detection was performed using the ECL system according to the manufacture's instruction.

Confocal microscope analyses

P_DMC1-DMC1-mCherry or P_FAS1-FAS1-mCherry were grown on V8 (pH7.0) plates for 5 days at room temperature in dark to form visible colonies. For probing DMC1 expression in different hyphal subpopulations, 50 µg/ml calcofluor white (stain cell wall) was added into the medium to enable the visualization of all fungal cell types. For sample preparation, colonies were embedded in low-gelling agarose (sigma-Aldrich) on the plates [59]. Immediately after gelling, agarose-embedded colonies were cut vertically in the middle and transferred to the cover glass with the longitudinal section facing the cover glass for the side view of the colonies. Cover the sample with 1∶9 PBS-Glycerol. An Olympus Fv1000 confocal laser scanning microscope and an Olympus IX81 spinning disk confocal microscope were used for the image acquisition. For Olympus Fv1000 confocal laser scanning microscope, calcoflour white and mCherry were excited with 488 nm argon and 543 nm He-Ne laser lines respectively using a dry 40× objective. For Olympus IX81 spinning disk confocal microscope, calcoflour white and mCherry were excited with 405 nm argon and 561 nm laser lines respectively with dry 40× objective. The data analyses were carried out with MetaMorph Microscopy Automation & Image Analysis Software (Molecular Devices, PA).