Distinct and redundant roles of exonucleases in Cryptococcus neoformans: Implications for virulence and mating

Abstract

Opportunistic pathogens like Cryptococcus neoformans are constantly exposed to changing environments, in their natural habitat as well as when encountering a human host. This requires a coordinated program to regulate gene expression that can act at the levels of mRNA synthesis and also mRNA degradation. Here, we find that deletion of the gene encoding the major cytoplasmic 5’→3’ exonuclease Xrn1p in C. neoformans has important consequences for virulence associated phenotypes such as growth at 37°C, capsule and melanin. In an invertebrate model of cryptococcosis the alteration of these virulence properties corresponds to avirulence of the xrn1Δ mutant strains. Additionally, deletion of XRN1 impairs uni- and bisexual mating. On a molecular level, the absence of XRN1 is associated with the upregulation of other major exonuclease encoding genes (i.e. XRN2 and RRP44). Using inducible alleles of RRP44 and XRN2, we show that artificial overexpression of these genes alters LAC1 gene expression and mating. Our data thus suggest the existence of a complex interdependent regulation of exonuclease encoding genes that impact upon virulence and mating in C. neoformans.

1. Introduction

Common to all messenger RNAs, irrespective of the myriad of functions they fulfil, are their beginning by transcription and their ending by degradation. Whereas transcription occurs exclusively in the nucleus, degradation of mRNAs is a process found in the nucleus as well as in the cytoplasm, operated by compartment-specific machineries that are however similar from a mechanistic point of view (see (Garneau et al. 2007) for review). In the nucleus and the cytoplasm, 3’→5’ decay is carried out by the exosome (see (Chlebowski et al. 2013) for review). This multiprotein complex harbours two catalytic active subunits, Rrp44p and Rrp6p with the latter being specific to the nucleus. 5’→3’ decay requires a protein of the Xrn family (Xrn2p/Rat1p in the nucleus, Xrn1p in the cytoplasm; see (Nagarajan et al. 2013) for review). It could be assumed that mutations in one of these global players have comparable consequences for every transcript and deletions are presumably fatal, which is what is found across species for Rrp44p/Dis3p and Xrn2p/Rat1p. On the other hand, deletion of XRN1 is tolerated by the cell, yet accompanied by pleiotropic phenotypes, evident from the independent identification of XRN1 in various screens in Saccharomyces cerevisiae. Originally isolated as necessary for nuclear fusion in S. cerevisiae (Kim et al. 1990), Xrn1p also regulates a large number of processes including filamentation (Kim and Kim 2002) and resistance to different drugs among which is fluconazole (Kapitzky et al. 2010). Xrn1p was also identified as a regulator of filamentation in Candida albicans (An et al. 2004), assigning a potential role for Xrn1p in fungal pathogenesis because of the requirements for filaments in pathogenesis in this species.

Initially the phenotypes of mutants without XRN1 were attributed exclusively to secondary consequences that the absence of Xrn1p has (i.e. alteration of transcript levels). However, it was found in baker's yeast that these phenotypes are in part due to Xrn1p-specific functions that are independent of its exonucleolytic activity (Solinger and Pascolini 1999). Interestingly, Xrn1p has been recently shown to directly associate with chromatin (Haimovich et al. 2013), thus regulating transcription of gene expression.

The basidiomycetous yeast Cryptococcus neoformans is a major human pathogen responsible for more than an estimated 1,000,000 infections and about 600,000 deaths per year (Park et al. 2009). Like most fungal pathogens, its global importance is due mainly to its capacity to infect immunocompromised individuals such as HIV/AIDS patients or people receiving organ/bone marrow transplants. Three major virulence factors of C. neoformans have been established: 1. the ability to grow at 37°C (Perfect 2006; Vecchiarelli and Monari 2012), 2. the presence of a polysaccharide capsule (Vecchiarelli and Monari 2012) and 3. the production of the pigment melanin (Williamson 1997). The genomes of two varieties of C. neoformans have been sequenced and annotated (Loftus et al. 2005; Janbon et al. 2014). These studies revealed very complex transcriptomes being very intron-rich (99% of the genes contain introns) and in which alternative splicing is common (Grützmann et al. 2014; Janbon et al. 2014). Moreover, numbers of long non-coding RNAs (lncRNAs), mainly antisense, have been identified and a large set of proteins orthologous to metazoan serine/arginine-rich (SR) proteins has been identified (Janbon et al. 2014); (Warnecke et al. 2008). C. neoformans is an opportunistic pathogen and its natural habitat is outside an animal host e.g. in the soil or in association with certain tree species (Lin and Heitman 2006). As such it needs to cope with a large number of stresses. It has been hypothesized that its complex and plastic transcriptome provides an easy way to alter its metabolism in order to colonize successfully a large diversity of environmental niches.

Recently, we identified the two essential exonucleases Xrn2p and Rrp44p as being key partners in the intron-dependent regulation of gene expression in C. neoformans (Goebels et al. 2013). Here, we describe the characterisation of a xrn1Δ strain in C. neoformans. We find that the deletion of this gene is associated with the alteration of several virulence factors which manifests in avirulence of the mutant strain. Further, we find that Xrn1p is needed for the regular succession of the mating process. We also observed that the deletion of XRN1 is associated with an upregulation of RRP44 and XRN2. Finally, our experiments showed that artificial overexpression of RRP44 and XRN2 is sufficient to alter LAC1 expression and mating. Taken together, these results suggest that a fine-tuned, interdependent regulation of the major exonucleases controls virulence and mating in C. neoformans.

2. Material and Methods

2.1 Strains and culture conditions

C. neoformans strains used in this study are all serotype D strains and are listed in Table 1. The strains were routinely cultured on YPD medium at 30°C (Sherman 1992). Synthetic dextrose (SD) was prepared as described (Sherman 1992).

Table 1.

List of the C. neoformans var. neoformans strains used in this study.

Name	Genotype	Construction details	Reference
JEC33	MATα lys2		Wickes and Edman, 1995
XL280	MATα		Lin et al., 2005
NE292	MATa cas3 Δi ade2 ura5		Goebels et al., 2013
NE718	MATα rrp6Δ::NEO cas3 Δi ura5		Goebels et al., 2013
NE730	MATα NAT-PGAL7::RRP44 lys2		Goebels et al., 2013
NE787	MATa NAT-PGAL7::RRP44		Goebels et al., 2013
NE791	MATa cas3 Δi NAT-PGAL7::RRP44		Goebels et al., 2013
NE809	MATa xrn1Δ::NAT cas3 Δi ade2 ura5	Biolistic transformation NE292	This study
NE822	MAT a	Cross JEC33 × NE809	This study
NE824	MATα	Cross JEC33 × NE809	This study
NE826	MATa xrn1Δ::NAT	Cross JEC33 × NE809	This study
NE828	MATα xrn1Δ::NAT	Cross JEC33 × NE809	This study
NE867	MATα xrn1Δ::NEO lys2	Biolistic transformation JEC33	This study
NE856	MATα NEO-PGAL7::XRN2 lys2		Goebels et al., 2013
NE874	MATa cas3 Δi NEO-PGAL7::XRN2 ade2 ura5		Goebels et al., 2013
NE923	MATα xrn1Δ::NAT	Biolistic transformation XL280	This study
NE924	MATα xrn1Δ::NEO	Biolistic transformation XL280	This study

Strains were grown overnight at 30°C in liquid YPD, serially diluted (10⁴ – 10¹) and spotted onto different solid media to determine growth phenotypes. Likewise, melanin production was assessed after spotting serial dilutions of cells of each strain on Niger agar medium (Walton et al. 2005); all plates were read after 48h of incubation at 30°C, unless otherwise stated.

2.2 Immunoblotting with monoclonal antibodies (MAb)

The anticapsular MAb E1 (Dromer et al. 1988) (kindly provided by F. Dromer, Institut Pasteur, Paris, France), CRND-8 (Ikeda et al. 1996) (kindly provided by T. Shinoda, Tokyo, Japan), 4H3, 2H1, and 5E4 (Casadevall and Scharff 1991) (kindly provided by A. Casadevall, Albert Einstein College of Medicine, New York, N.Y.) were used in immunoblotting experiments, with techniques as previously described (Janbon et al. 2001).

2.3 RNA extraction and northern blot analysis

Cells were routinely harvested after being grown up to 5×10⁷ cells/mL in YPD. RNA was extracted with TRIZOL Reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. Total RNA (5 μg) was separated by denaturing agarose gel electrophoresis and transferred onto Hybond-N+ membrane (GE Healthcare, Piscataway, NJ) and probed with [³²P]dCTP-radiolabelled DNA fragments. β-particle emissions were quantified with a Typhoon 9200 imager (GE Healthcare) and Image Quantifier 5.2 software (Molecular Dynamics, Fairfield, CT).

For the analysis of LAC1 expression under carbon starvation conditions, 4×10⁶ cells/mL were grown in 50 mL of YPD or YPG for 3 hours. At 3 hours, cells were collected by centrifugation, washed with sterile distilled water, resuspended in 50 mL of Asparagine medium without glucose (1 g/L asparagine, 0.1 g/L MgSO₄, 10 mM NaH₂PO₄; pH 6.5) and incubated in this medium for another 3 hours.

2.4 Gene disruption

The genes described in this report have been deleted by biolistic transformation of a serotype D strain using a disruption cassette constructed by overlapping PCR as previously described (Moyrand et al. 2004). The primer sequences used are given in Table A.1. The transformants were then screened for homologous integration as previously described (Moyrand et al. 2004). The plasmid, pNAT used to amplify the NAT selective marker was kindly provided by Dr Jennifer Lodge (Saint Louis University School of Medicine). The plasmid pPZP-NEO1 was used to amplify the NEO selective marker. Multiple mutant strains were obtained through crosses of single mutant strains on V8 medium as previously described (Moyrand et al. 2004). Progenies were selected on minimum medium. Their genotypes were determined by PCR. The mating types of the strains were determined by crossing them on V8 medium to tester strains of known mating type.

2.5 Capsule growth

To induce capsule enlargement, cells were incubated overnight in liquid Sabouraud medium (Oxoid, England) at 30°C with moderate shaking, and then transferred to 10% Sabouraud medium at pH 7.3 buffered with 50 mM MOPS buffer (Zaragoza and Casadevall 2004) for 24 hours. The cells were then suspended in India Ink, and pictures were taken using a Leica DMI3000 microscope (Leica Microsystems, Germany). Capsule size was estimated using Adobe Photoshop 7.0 (Adobe, San Jose, CA). The diameter of the total cell size and of the cell body (delimited by the cell wall) was measured, and capsule size was determined as the difference between these two parameters. Statistically-significant differences between samples were determined using ANOVA and Student's t-Test (significance was considered when p<0.05).

2.6 Galleria mellonella infection

Galleria mellonella were infected as previously described (Mylonakis et al. 2005; García-Rodas et al. 2011). The larvae were obtained from Alcotan (Valencia, Spain). Larvae without any dark spots weighting between 0.2-0.3 g were selected and incubated at 30°C or 37°C the day before the infection. To prepare the inocula, yeast were grown overnight in liquid Sabouraud medium at 30°C with moderate shaking, washed with PBS plus ampicillin (50 μg/mL) and suspended in the same buffer. Cell density was estimated with an automatic cell counter TC10 (Bio-Rad, Hercules, CA). Then, suspensions at 10⁸ cells/mL were prepared in PBS+ampicillin. Larvae (20 per group) were injected with 10 μL of the yeast suspension (10⁶ yeast cells per larva) through the last right proleg, which was previously cleaned with 70% ethanol. A parallel group of larvae was injected with 10 μL of PBS+ampicillin as a control. The larvae were incubated at 30°C or 37°C, and death was daily monitored for 10 days.

Killing curves were adjusted using the Kaplan-Meier method and estimation of differences in survival were analysed with the log rank and Wilcoxon tests using GraphPad Prism 5 software (GraphPad, San Diego, CA). A p-value below 0.05 was considered significant.

2.7 In vivo phagocytosis assay

Yeasts were grown overnight in liquid Sabouraud medium, washed with PBS and stained with 10 μg/mL Calcofluor white (Sigma, St. Louis, MO) for 30 minutes at 37°C. Then, these cells were injected into G. mellonella (10⁶ cells/larva) as described above. After 3 hours of incubation at 30°C or 37°C, haemolymph was collected in 1.5 mL tubes in 50 μL cold PBS to avoid coagulation and melanization of the haemolymph. Haemocytes were placed on a slide and phagocytosis was visually quantified using a Leica DMI 3000B. 100 haemocytes from each larva were counted per experiment and results were expressed as the percentage of haemocytes that contained internalized cryptococcal cells.

2.8 Growth curves

Yeast strains were grown overnight in liquid Sabouraud medium. Then, cellular suspensions at 2×10⁵ cell/mL were prepared in fresh liquid Sabouraud medium and 170 μL were placed in 96-well microdilution plates (Costar, New York, NY). The plate was placed in an IEMS Reader MF spectrophotometer (Thermo Fisher Scientific, Germany) and incubated at 30°C or 37°C with shaking. Optical density (OD) was determined at 540 nm every hour for 72 hours. Data were processed with GraphPad Prism 5 software.

2.9 Mating and cell fusion assays

WT and xrn1Δ mutant cells of opposite mating type were mixed and co-cultured on 5 % V8 juice agar medium (pH 7) and incubated at room temperature in the dark for 1 week. The colonies were photographed following incubation.

To perform the cell fusion assays, wild-type strains JEC33 (MATα) and NE292 (MATa), and xrn1Δ mutant strains NE809 (MATa) and NE867 (MATα) were grown in YPD liquid medium overnight. Cells were washed and adjusted to 2×10⁷ cells/mL, mixed, and grown on V8 mating medium in the dark for 24 h. The colonies were removed with cell scrapers, resuspended in sterile water, and plated onto YNB medium to select for cell–cell fusion products that have both LYS2 and URA5 genes. Plates were incubated at 30°C for 5 days until colonies were observed. The experiment was conducted in three replicates and the average number of colonies was calculated. The fusion efficiency was determined by comparing the average number of the visible colonies from the xrn1Δ mutant cross to that from the wild-type cross.

To determine an effect of xrn1Δ on same sex mating, the gene was deleted in the hyperfilamentous strain XL280 background (Lin et al. 2005), and filamentation was assessed by incubation on V8 mating medium. Pictures were taken with a Leica stereomicroscope.

2.10 Ploidy determination by flow cytometry

Cells were grown to stationary phase in deep well 96-well plates. 10⁷ cells were harvested by centrifugation (5 min, 5000 rpm, room temperature) and fixed in 70 % ethanol for 1 hour at 4°C. Cells were then washed with water, resuspended in 200 μL TE buffer and incubated with 0.5 g/L RNaseA (Fermentas, Germany) overnight at 37°C. The next day, cells were centrifuged and resuspended in 500 μL TE buffer. 50 μL of this cell suspension were added to 450 μL Sytox Green (Invitrogen) staining solution (1 μM) and flow cytometry was performed on 40,000 cells on the FL1 channel of a MACSQuant flow cytometer (Miltenyi, Germany).

3. Results

3.1 Interdependent regulation of the different exonuclease genes

We recently reported reciprocal compensation between the major exonucleases Rrp44p and Xrn2p in C. neoformans (Goebels et al. 2013). In the same study, we also found that the deletion of RRP6 is associated with a slight up-regulation of RRP44. We here assayed mRNA accumulation of the major 5’→3’ cytoplasmic exonuclease gene XRN1 in three exonucleolytic mutants. As presented in Figure 1A, expression of XRN1 is slightly increased in all of them (rrp44, xrn2, rrp6Δ). Moreover, upon deletion of the XRN1 gene the transcript levels of RRP44 and XRN2 are significantly increased (Figure 1B). These data suggest a complex interdependent regulation and possible functional redundancies between the different exonucleases.

Figure 1. Widespread compensation between different exonucleases.

A Expression of XRN1 is increased in exonucleolytic mutants. Northern blot experiment results showing the transcript levels of XRN1 in different exonucleolytic mutants. Relative to wild-type level and normalised to ACT1, XRN1 mRNA abundances are 1.8x in the rrp44 strain, 1.7x in the xrn2 strain and 1.3x in the rrp6Δ strain B Deletion of XRN1 is compensated by overexpression of RRP44 and XRN2. Northern blot experiment results showing the transcript levels of RRP44 and XRN2 depending on the presence of XRN1. Normalised to ACT1, RRP44 and XRN2 mRNA abundances in the xrn1Δ mutant strain are 3.5x and 1.8x respectively, relative to wild-type level.

3.2 Deletion of XRN1 results in alteration of multiple virulence factors

We deleted the gene encoding Xrn1p (locus CNE03620) using a nourseothricin marker. The original deletant strain was then backcrossed to a wild-type strain to eliminate the possibility of secondary mutations obscuring the analysis.

As previously reported in S. cerevisiae (Larimer and Stevens 1990), the xrn1Δ mutants show decreased vegetative growth (Figure 2A). Of particular interest for a human pathogen, in C. neoformans this phenotype is exacerbated at 37°C (Figure 2A). Strikingly, we observed an aggravation of the xrn1Δ-related growth impairment at 37°C in MATa relative to MATα strains (Figure 2A, and data not shown). Furthermore, the mutants also show an increased sensitivity to SDS (Figure 2A).

Figure 2. XRN1 deletion alters virulence-associated phenotypes.

A Growth phenotypes associated with XRN1 deletion. Serial dilutions of cells were spotted onto different media. Pictures were taken after 4 (YPD) and 5 (Niger) days, respectively. B Capsule size differences associated with XRN1 deletion. Distribution of capsule diameter of wild-type and xrn1Δ cells at 30°C in Sabouraud medium and diluted Sabouraud medium. The line in each sample denotes the average of the distribution.

In addition to the ability to grow at 37°C, there are other well-defined virulence factors in C. neoformans, such as the presence of an antiphagocytic capsule and the production of melanin. As presented in Figure 2A, xrn1Δ cells lack the ability to produce melanin when grown on Niger seed agar. Wild-type Cryptococcus strains produce melanin under conditions of glucose starvation mainly via a pathway involving the two copper-oxidases Lac1p and Lac2p, with Lac1p being responsible for the majority of laccase activity (Zhu and Williamson 2004). To test whether the deficiency in melanin production observed in the xrn1Δ mutants was a direct consequence of reduced LAC1 expression, we performed northern analyses. The results presented in Figure 3A show that the induction of LAC1 transcription seen in a wild-type strain under glucose-depleted conditions (i.e. growth in Asparagine medium) hardly occurrs in a xrn1Δ mutant.

Figure 3. Expression of LAC1 requires XRN1.

Northern blot analysis of RNA from the indicated cells incubated for 3 hours in YPD (A) or YPG (B) and then shifted for an additional 3 hours to Asparagine medium (Asn) or grown for 6 hours in YPD (A) as control. A Induction of LAC1 by glucose starvation (Asn) in WT and xrn1Δ cells. Relative to wild-type level and normalised to ACT1, LAC1 mRNA abundance in the xrn1Δ strain is 0.1x B LAC1 expression in galactoseinducible mutants of RRP44 and XRN2. Relative to wild-type level and normalised to ACT1, LAC1 mRNA abundances are 0.7x and 0.9x in the P_GAL7::XRN2 and P_GAL7::RRP44 strains, respectively.

Arguably the most prominent virulence property of C. neoformans is the existence of an antiphagocytic capsule surrounding the yeast cell. Interestingly, an increase in capsule size under capsule inducing conditions (i.e. 10% Sabouraud medium) related to the deletion of XRN1 was found (Figure 2B). The alteration of size was not visible when the cells were cultivated under non-inducing conditions (i.e. standard Sabouraud medium). In order to test whether the deletion of XRN1 is also accompanied by any structural changes in the polysaccharide capsule, we tested the reactivity of wild-type and mutant cells to different anticapsule antibodies. Although wild-type strains of the D serotype do not normally react with the serotype A-specific antibody E1 (Dromer et al. 1987), we found a low but reproducible reactivity of the xrn1Δ strains with this antibody (Figure A.1), pointing to a different structural composition of the capsule (Moyrand et al. 2002) in the absence of XRN1. Also, an increased binding affinity of the anticapsule antibody 4H3 (Mukherjee et al. 1992) is seen for strains bearing the XRN1 deletion (data not shown). In contrast, the same assay using the anticapsular antibodies CRND-8 (Ikeda et al. 1996), 2H1 and 5E4 (Casadevall and Scharff 1991) revealed no difference between the xrn1Δ mutant strains and the wild-type.

3.3 xrn1Δ strains are avirulent in G. mellonella

In vitro determination of virulence factors can give hints as to whether a certain mutant is attenuated for virulence. However, it is not a sufficient substitute for virulence modelling. To confirm the relevance of the altered virulence properties for the overall virulence of xrn1Δ strains, we chose the G. mellonella model of cryptococcosis. Besides the general advantages of invertebrate infection models, such as lower costs and ethical concerns, this model also facilitates infection modelling at a temperature of 30°C where the xrn1Δ growth deficit is minor (Figure 2A and Figure 4A), allowing the assessment of the virulence of the mutants independently of its overall reduced fitness at elevated (mammalian body) temperature.

Figure 4. xrn1Δ cells are avirulent.

A Growth curves of wild-type and xrn1Δ cells at 30°C in Sabouraud medium. B Survival of G. mellonella infected with wild-type and xrn1Δ strains at 30°C. C In vivo phagocytosis of G. mellonella larvae infected with wild-type and xrn1Δ cells at 30°C.

In keeping with the physiological (growth) and biochemical (melanin, capsule structure) characteristics found in vitro, xrn1Δ strains also appeared to be completely avirulent in this invertebrate model of cryptococcosis (Figure 4B). As observed previously, phagocytosis (Figure 4C) does not seem to be an indicator for whether a strain is a successful pathogen or not, but rather a mere consequence of capsule size (García-Rodas et al. 2011) (Figure 2B).

3.4 Absence of XRN1 interferes with mating

The melanin and virulence phenotypes associated with the deletion of XRN1 were partially reminiscent of the phenotypes reported to be associated with the deletion of the VAD1 gene (Panepinto et al. 2005). Vad1p is the homologue of the S. cerevisiae protein Dhh1 which has been shown to activate decapping in response to impaired ribosome elongation thus triggering mRNA to degradation by Xrn1p (Fischer and Weis 2002; Sweet et al. 2012). Vad1p also negatively regulates mating in C. neoformans (Park et al. 2010). Given the role of Xrn1p in the regulation of mating in S. cerevisiae (Kim et al. 1990; Solinger and Pascolini 1999), we characterised the mating process in the absence of XRN1 in C. neoformans. All possible combinations of crosses, i.e. WT(JEC33) x WT(NE292), WT(JEC33) x xrn1Δ(NE809), xrn1Δ(NE867) x WT(NE292) and xrn1Δ(NE867) x xrn1Δ(NE809), were tested. Though no obvious mating defect was observed during unilateral mating between mutant and wild-type strains, a significant mating defect with reduced filamentation was apparent in a bilateral mutant cross (Figure 5A). In addition, less basidiospores were generated during the xrn1Δ x xrn1Δ cross. Furthermore, cell fusion assays (see 2.9) revealed that xrn1Δ cells were less efficient in cell fusion (10 % of wild-type level).

Figure 5. Deletion of XRN1 affects sexual reproduction.

A xrn1Δ mutants show a defect in bisexual mating. Mating between WT and xrn1Δ mutant cells of opposite mating type were mixed and co-cultured on 5% V8 juice agar medium (pH 7) and incubated at room temperature in the dark for 1 week. The colonies were photographed following incubation. Scale bar = 10 μm. B Deletion of XRN1 suppresses the hyperfilamentous phenotype of XL280. Strains were incubated on V8 juice agar medium (pH 7) for 15 days at room temperature in the dark. Pictures were taken with a Leica stereomicroscope. Scale bar = 5 mm.

Analysis of the phenotypes of 96 randomly chosen progenies revealed a non-Mendelian repartition of the nourseothricin resistance phenotype (8/96) associated with the XRN1 deletion cassette (Table A.2). We also analysed the ploidy of these isolates by flow cytometry (see 2.10) but we did not detect any diploid isolates among them (Table 2).

Table 2.

Results of ploidy analysis by flow cytometry.

Cross	Number of isolates analysed	Number of diploid strains
WT × WT	86	1
WT × xrnlΔ	92	0
RRP44OE × WT	92	4
WT × XRN2OE	93	3
RRP44OE × XRN2OE	94	6

Besides the well-defined a-α opposite sexual cycle, C. neoformans can also undergo α-α unisexual reproduction. While both these forms of sexual reproduction share common regulators, they are mechanistically similar but distinct (Wang and Lin 2011). To test whether the effect of XRN1 deletion is specific to only the bisexual cycle, we deleted the gene also in the hypersexual haploid strain XL280 (Lin et al. 2005). As demonstrated in Figure 5B, absence of XRN1 suppresses the filamentous phenotype of the XL280 strain when grown on mating-inducing media (V8). Based on this observation we conclude that Xrn1p most likely affects part of the regulatory pathway that is common to same and opposite sex mating.

3.5 Overexpression of XRN2 or RRP44 is enough to alter LAC1 expression in C. neoformans

Our data show that the melanin deficiency phenotype of the xrn1Δ strain is probably due to the near absence of LAC1 expression when the mutant strains are cultured in Asparagine medium. This absence of LAC1 expression might be due to no expression induction in the xrn1Δ mutant. Another possibility is that the LAC1 transcripts are rapidly degraded in a xrn1Δ context due to an increased activity of other exonucleases (Figure 1B and Figure 3A). Using regulatable alleles of RRP44 (P_GAL7::RRP44) and XRN2 (P_GAL7::XRN2), where the native promoters have been replaced with the GAL7 promoter, we assayed LAC1 transcript levels in these strains growing under inducing conditions (i.e. YP Galactose) prior to shifting them to Asparagine medium. The results presented in Figure 3B show that LAC1 induction is slightly less pronounced in the P_GAL7::RRP44 or P_GAL7::XRN2 strains compared to wild-type, though clearly superior to the level observed in xrn1Δ cells. We also looked at a different way to induce LAC1 expression by the addition of copper that has been reported for the serotype A (Jiang et al. 2009). However, its independence from the carbon source does not seem to be conserved between serotypes as LAC1 is almost undetectable in cells grown in Asparagine medium supplemented with glucose irrespective of the addition of copper (data not shown) and thus does not represent an alternative. Overall, although the precise mechanism by which XRN1 deletion leads to LAC1 repression remains to be defined, our data demonstrate that the sole overexpression of RRP44 or XRN2 is enough to modulate LAC1 expression.

3.6 Induced overexpression of RRP44 or XRN2 affects mating

We also tested whether induced overexpression of RRP44 or XRN2 would perturb the mating process using the galactose-inducible alleles. Therefore, we co-cultured the different mutant strains together with wild-type or mutant strains on V8 agar supplemented with 2 % galactose in order to induce transcription from the GAL7 promoter. Robust filamentation was observed in all possible combinations (data not shown).

To characterise further a potential perturbance of the mating process caused by overexpression of XRN2 (XRN2^OE) or RRP44 (RRP44^OE), we screened progeny from different crosses, i.e. RRP44^OE x WT, WT x XRN2^OE and RRP44^OE x XRN2^OE. First, by flow cytometrical analysis, we noticed an increased number of diploid strains coming from crosses of the overexpression mutants with WT, i.e. RRP44^OE x WT, WT x XRN2^OE (Table 2). Next, we screened the progeny with respect to segregation of different markers (Table A.2). In these single overexpression mutant x WT crosses, irrespective of a Mendelian-like segregation of the dominant selectable markers used to tag the regulatable alleles (47/96 for P_GAL7::RRP44::NAT, 41/96 for P_GAL7::XRN2::NEO), 15 % of P_GAL7::RRP44 and 100 % of P_GAL7::XRN2 strains retained the ability to grow on YPD which for haploid strains necessitates a wild-type copy of RRP44 or XRN2, respectively (Goebels et al. 2013). This result suggested that the strains could be heterozygous for these loci. Indeed, PCR analysis of randomly chosen clones confirmed the presence of the additional wild-type allele (Figure 6A and data not shown). Moreover, although detection of potential aneuploidy by multiplex PCR did not yield conclusive results (Figure A.2), a potential aneuploidy for chromosome 4 (harbouring RRP44 as well as the MAT locus) in some of the descendants from the RRP44^OE x WT cross seems likely, as selected P_GAL7::RRP44::NAT strains, that were haploid as assessed by flow cytometry (Figure 6C), filamented when cultured on V8 + 2 % galactose (Figure 6B).

Figure 6. Overexpression of RRP44 or XRN2 perturbs mating.

A Matings in which XRN2 is overexpressed (WT x XRN2^OE) yield haploid progeny that contain both, the wild-type XRN2 and the P_GAL7::XRN2 alleles. PCR analysis showing the simultaneous presence of wild-type and P_GAL7 alleles of XRN2 in progeny from the WT x XRN2^OE cross. PCR with primers flanking the XRN2 promoter region yield a fragment of about 2.7 kb for the wild-type allele and 4.5 kb for the P_GAL7 allele. B Progeny strains (III-17, III-51) from matings in which RRP44 is overexpressed (RRP44^OE x WT) filament on V8 + 2 % galactose. C Progeny strains (III-17, III-51) from matings in which RRP44 is overexpressed (RRP44^OE x WT) are haploid as determined by flow cytometry.

The double mutant cross (RRP44^OE x XRN2^OE ) further illustrated the intertwined, yet partially independent roles that these two exonucleases play during the mating process. The more frequent occurrence of diploid progeny, found in the single overexpression mutant x WT crosses (RRP44^OE x WT, WT x XRN2^OE) was also observed in progeny from the RRP44^OE x XRN2^OE cross (Table 2). Further screening of the progenies from this cross yielded an unexpected high portion of P_GAL7::RRP44 single mutants (59/96) whereas P_GAL7::XRN2 single mutant (16/96), P_GAL7::RRP44 P_GAL7::XRN2 double mutant (11/96) and wild-type (10/96) strains were present in equal, though much lower than expected quantities (Table A.2). As already observed in progeny from the WT x XRN2^OE cross, all P_GAL7::XRN2 descendants of the double mutant cross regained the ability to grow on YPD which coincided with the additional presence of the wild-type allele of XRN2 (Table A.2). Similarly, all 7 double mutant strains that were haploid by flow cytometry contained a wild-type copy of XRN2 (Table A.2). Overall, these data indicate that increased expression of either of these two exonucleases is sufficient to severely perturb the regular succession of the sexual reproduction process.

4. Discussion

This study identifies the cytoplasmic exonuclease Xrn1p as a regulator of multiple virulence factors in C. neoformans, extending previous studies that have found a central role for RNA metabolic processes in stress adaptation and virulence in this fungus (see (Bloom and Panepinto 2014) for a recent review).

One consequence of deletion of XRN1 is severely impaired cryptococcal growth at 37°C. Thermotolerance can be characterised as a virulence factor of “disproportionate importance” (Coelho et al. 2014) as it clearly delimits the comparatively few fungal species capable of causing systemic infections in humans (Robert and Casadevall 2009). Unlike capsule synthesis or melanin production, growth at elevated temperature requires adaptation of the whole cellular machinery and is consequently accompanied by extensive changes in gene expression (Steen et al. 2002; Kraus et al. 2004; Chow et al. 2007). These alterations involve mRNA synthesis as well as degradation. Recent work from the Panepinto lab explored the involvement of Ccr4p in the cellular response to host temperature (Havel et al. 2011; Bloom et al. 2013). Ccr4p-mediated deadenylation represents the initial step in cytoplasmic mRNA turnover that ultimately leads to 5’→3’ degradation by Xrn1p. Partially resembling phenotypes of strains with CCR4 and XRN1 mutations underline a functional connection. The heightened importance of Xrn1p in the coordination of mRNA synthesis and degradation was elegantly shown in two global studies in S. cerevisiae (Haimovich et al. 2013; Sun et al. 2013). Chromatin association of S. cerevisiae Xrn1p as reported by Haimovich et al. points to a role for Xrn1p in transcription. Assuming an involvement of Xrn1p in transcription also in C. neoformans could explain the observed impact that XRN1 deletion has on LAC1 expression and thus melanin production. Hence, the at first counter-intuitive observation of reduced LAC1 mRNA levels upon deletion of this exonuclease might be due to an involvement of Xrn1p in LAC1 transcription alongside with increased degradation of LAC1 by xrn1Δ-induced expression of other exonucleases.

The laccase defect that we found associated with deletion of XRN1 mimics that of a vad1Δ strain. A mutant in the DEAD-box RNA helicase encoding gene VAD1 was initially isolated in an insertional mutagenesis screen for laccase-deficiency (Panepinto et al. 2005). Localisation studies describe an accumulation of Vad1p in P-body-like structures (Panepinto et al. 2005). These “factories for mRNA decay” are clearly linked to Xrn1p activity in yeast (Sheth and Parker 2003; Kulkarni et al. 2010). A co-operative activity of Vad1p and Xrn1p could thus be at the origin of LAC1 repression. However, the interaction between Vad1p and Xrn1p does not seem to be readily defined when considering the xrn1Δ-related perturbance of the mating process described in this study that is reversed to that found for vad1Δ (Park et al. 2010). Further studies are obviously needed to elucidate their interconnection.

The pleiotropic roles of Xrn1p partially stem from its function as an exonuclease. Redundancy of this exonuclease activity in the cell obstructs a clear view of its functions. Comparison of the effect that induced overexpression of RRP44 and XRN2 has on the mating process helped to characterise the XRN1-related mating phenotype as being distinct from the exonucleolytic compensation, i.e. increased expression of RRP44 and XRN2 upon XRN1 deletion. At the same time, our data from mating experiments in which RRP44 or XRN2 were overexpressed clearly show the importance of equilibrated levels of exonucleases for successful sexual reproduction. Elucidation of the underlying molecular mechanisms promises to aid in understanding the regulation of the mating process. Overall, the here presented work adds Xrn1p to the growing number of RNA-related factors that orchestrate virulence in C. neoformans. It is apparent that a better understanding of the RNA metabolism present in this basidiomycete can give valuable insight into its multifaceted lifestyle, crucial to counter this pathogenic challenge.