Sex-Induced Silencing Operates During Opposite-Sex and Unisexual Reproduction in Cryptococcus neoformans

Abstract

Cryptococcus neoformans is a human fungal pathogen that undergoes a dimorphic transition from yeast to hyphae during a-α opposite-sex mating and α-α unisexual reproduction (same-sex mating). Infectious spores are generated during both processes. We previously identified a sex-induced silencing (SIS) pathway in the C. neoformans serotype A var. grubii lineage, in which tandem transgene arrays trigger RNAi-dependent gene silencing at a high frequency during a-α opposite-sex mating, but at an ∼250-fold lower frequency during asexual mitotic vegetative growth. Here we report that SIS also operates during α-α unisexual reproduction. A self-fertile strain containing either SXI2a-URA5 or NEO-URA5 transgene arrays exhibited an elevated silencing frequency during solo and unisexual mating compared with mitotic vegetative growth. We also found that SIS operates at a similar efficiency on transgene arrays of the same copy number during either α-α unisexual reproduction or a-α opposite-sex mating. URA5-derived small RNAs were detected in the silenced progeny of α-α unisexual reproduction and RNAi core components were required, providing evidence that SIS induced by same-sex mating is also mediated by RNAi via sequence-specific small RNAs. In addition, our data show that the SIS RNAi pathway also operates to defend the genome via squelching transposon activity during same-sex mating as it does during opposite-sex mating. Taken together, our results confirm that SIS is conserved between the divergent C. neoformans serotype A and serotype D cryptic sibling species.

COMPLEX genomes have evolved sophisticated mechanisms to sense the presence, and to control the behavior of genomic invaders such as transposable elements and viruses. Among a panoply of diverse strategies, homology-dependent gene silencing (HDGS) is a ubiquitous phenomenon among fungi, plants, and animals (Cogoni 2001). HDGS, also known as cosuppression, was first found to be triggered by transgenes, viruses, or unusual duplications of endogenous genes in both plants and fungi (Cogoni et al. 1996; English et al. 1996; Bingham 1997; Cogoni and Macino 1999a). Although a broad variety of HDGS phenomena are known, all of these gene silencing processes are based on the recognition of nucleic acid sequence homology, followed by inactivation of homologous sequences. Gene silencing can occur at a transcriptional or post-transcriptional level involving sequence-specific mRNA degradation. The most studied and best-characterized HDGS phenomenon in fungi is a process known as quelling, which occurs in Neurospora crassa vegetative tissue (Cogoni et al. 1996). Quelling is an RNAi-mediated silencing mechanism that post-transcriptionally inactivates duplicated sequences, in many cases resulting from introduction of transgenes. In quelling, the silenced loci can act in trans, leading to silencing of all homologous genes throughout the genome. The core RNAi components, including Argonaute, Dicer-like proteins, and RNA-dependent RNA polymerase (RdRP), are all required for quelling (Cogoni and Macino 1999b; Catalanotto et al. 2002; Catalanotto 2004; Fulci and Macino 2007).

Sexual reproduction is associated with a high frequency of recombination, which enables species to adapt to diverse environments. However, sexual reproduction can also be beneficial to transposon propagation because meiosis can enable transposon exchange between genomes. In addition, meiosis can lead to heterologous chromosome alignment and recombination or translocation due to the presence of ectopic copies of preexisting transposons (Jinks-Robertson and Petes 1986; Bourc’his and Bestor 2004; Kelly and Aramayo 2007; Ni et al. 2011). Thus, transposon control is especially critical during sexual reproduction. Many organisms have evolved specific sex-related host defense mechanisms that are activated during the sexual cycle to promote genome integrity.

The filamentous fungus Neurospora crassa is an exemplary model that has developed a number of complex defenses to preserve its genome integrity at different stages of the sexual cycle, including repeat-induced point mutation (RIP) and meiotic silencing of unpaired DNA (MSUD) (Ruiz et al. 1998; Borkovich et al. 2004; Galagan and Selker 2004; Kelly and Aramayo 2007). Like quelling, RIP is also a homology-based process that acts at the premeiotic stage of sexual development (Selker et al. 1987). RIP efficiently detects and inactivates repetitive DNA elements present in haploid genomes destined to participate in meiosis and operates by introducing C:G to T:A mutations into both copies of the duplicated sequences (Cambareri et al. 1989; Selker 2002). Up to 30% of the GC base pairs in duplicated sequences can be altered to AT pairs via RIP after a single sexual cycle (Cambareri et al. 1991). In contrast to RIP, MSUD occurs during meiosis and detects the unsuccessful pairing of discrete DNA regions during homolog pairing (Shiu et al. 2001). An RNAi-mediated silencing mechanism is triggered and post-transcriptionally silences all genes contained in the loop of unpaired DNA region (Lee et al. 2003; Fulci and Macino 2007). In this case, MSUD can silence ectopic sequences and newly transposed elements that are present as only a single copy during meiosis. Thus, MSUD serves as a complementary mechanism to RIP, and functions to protect the organism against any transposons that escape RIP. The presence of this powerful genome defense system in N. crassa has shaped its genome into one with a very low content of repetitive sequences (Galagan et al. 2003; Borkovich et al. 2004).

In addition to N. crassa, other fungi also exhibit different meiotic silencing mechanisms. Methylation induced premeiotically (MIP) in Ascobolus immersus and Coprinopsis cinerea is another HDGS phenomenon that is closely related to RIP (Rhounim et al. 1992; Barry et al. 1993; Freedman and Pukkila 1993). Like RIP, MIP also takes place premeiotically and silences repeated DNA sequences. MIP leads to methylation of C residues within both repeated sequences, but in contrast to RIP, no sequence changes occur (Faugeron 2000; Cogoni 2001). Furthermore, a novel sex induced silencing (SIS) phenomenon has been reported in the human fungal pathogen Cryptococcus neoformans (Wang et al. 2010). A tandem multicopy transgene insertion triggered a homology-based gene silencing process termed SIS that occurs at ∼250-fold higher frequency during the sexual cycle compared to vegetative mitotic growth. SIS is mediated by an RNAi pathway and the increased silencing frequency during sex appears to be linked to an increased abundance of the RNAi pathway elements during sexual reproduction. The RNAi-mediated SIS pathway also functions to squelch transposon activity during the sexual cycle, thus serving as a genome defense mechanism during meiosis (Wang et al. 2010). Transgenes are also repressed via a mitotic-induced silencing pathway (MIS) similar to quelling (Wang et al. 2012).

C. neoformans grows as a haploid budding yeast with a bipolar mating system that involves a and α mating types (Idnurm et al. 2005). Spores can be produced via two distinct pathways: a-α opposite sex mating and α-α unisexual reproduction. Monokaryotic fruiting, once thought to be mitotic and asexual, has been demonstrated to be a modified form of sexual reproduction occurring between strains of the same mating type and is most commonly observed in α strains but can also occur with a isolates (Fraser et al. 2005; Lin et al. 2005, 2007). Because α isolates predominate (>95%) in most naturally occurring C. neoformans populations, and the α mating type is linked to virulence (Kwon-Chung and Bennett 1978; Nielsen et al. 2005; Lin et al. 2006; Lin 2009), the discovery of α-α unisexual reproduction (same-sex mating) suggests how infectious spores can be produced by this almost unisexual pathogen. Although many similarities are shared between opposite-sex and same-sex mating, ongoing studies comparing the underlying mechanisms provide insight into how same-sex mating can contribute in unique ways to the fitness of these pathogenic fungi (Lee et al. 2010; Wang and Lin 2011).

SIS was first identified during the a-α opposite-sex mating cycle (Wang et al. 2010). In this study, we found that a transgene-induced RNAi-dependent sex-induced silencing pathway also operates during α-α unisexual reproduction and that the silencing frequency during opposite-sex mating and same-sex mating is comparable. More importantly, we observed that transposons were more active during the α-α unisexual cycle in the rdp1 mutants compared to wild type, leading to elevated transposition and mutation similar to our findings during a-α mating of RNAi mutants. Our results suggest that during α-α same-sex mating as well as a-α opposite-sex mating, SIS is deployed as a genome defense mechanism to protect the genome during meiosis.

Materials and Methods

Strains and media

Strains and plasmids used in this study are listed in Supporting Information, Table S1. Yeast cells were grown and maintained on yeast extract-peptone-dextrose (YPD) medium. Synthetic dextrose (SD) medium lacking uracil or containing 5-FOA (1 g/liter) were used to test whether isolates of interest are auxotrophic for uracil. Mating of C. neoformans was conducted on 5% V8 juice agar medium (pH = 7) in the dark as previously described (Hull et al. 2002). Spores generated by either unisexual or opposite-sex mating were isolated by micromanipulation as described (Hsueh et al. 2006). The transformed strains containing the SXI2a-URA5 or the NEO-URA5 transgene were obtained by biolistic transformation with plasmid DNA (Toffaletti et al. 1993). PCR with primers JOHE23492 (CCTGCTTCGTTTAACTCCATGGCAAA), JOHE23493 (CCAACGAGCTTCGCTTCAGGAGAG), JOHE26870 (GGAAGCCGGTCTTGTCGA TCAGG), and JOHE26871 (CTCGATGGGACCGGTGCTTAAGGC) were used to screen for the presence of the SXI2a-URA5 or the NEO-URA5 transgene in the genome. The transformed strains used in this study were subjected to mitotic stability examination as described before (Varma et al. 1992; Toffaletti et al. 1993) to confirm integration into the genome.

Real-time PCR

Quantitative real-time PCR was performed with primers specific to the actin gene ACT1 and URA5 (Wang et al. 2010) to determine the copy number of the SXI2a-URA5 and the NEO-URA5 transgenes. DNA isolated from the wild-type strain JEC20 or XL280, the transformed strains, and the progeny were used as templates. PCR was conducted with Brilliant SYBR Green QPCR Master mix (Stratagene) and the relative quantity of the URA5 gene determined by the ΔΔCt method according to the following equations: (Ct value means the threshold cycle) ΔCt = Ct (target) − Ct (normalize) and ΔΔCt = ΔCt (experimental) − ΔCt (control). Comparative expression level = 2^–ΔΔCt.

Gene disruption

A standard overlap PCR approach (Fraser et al. 2003) was used to disrupt the RDP1 gene in strains XW73 and XW139 with the NAT dominant selectable marker. The overlap PCR products were introduced into the genome by biolistic transformation (Davidson et al. 2000). Primers that were used for amplification of the ∼1 kb of 5′ and 3′ flanking regions of the RDP1 disruption cassette are described in Janbon et al. 2010. Transformants were initially screened by PCR and Southern blot analyses were then conducted to identify a single integration at the desired locus.

Detection of siRNAs by Northern blot

To detect siRNAs, total RNA from strains of interest was extracted with TRIzol and RNAs of low molecular weight were prepared by precipitating high molecular weight RNAs in the presence of 0.5 M NaCl and 5% polyethylene glycol (MW8000) and recovering the supernatant (Catalanotto et al. 2002). Small RNA-enriched fractions were then separated in 15% TBE-urea gels, followed by Northern blot. DNA oligonucleotides of 21 nt and 25 nt served as size markers. After electrophoresis, RNA was transferred to Hybond-N⁺ nylon membrane, and UV cross-linked. Single-stranded RNA probes were in vitro transcribed in the presence of ³²P-labeled UTP with the MAXIscript T7 kit (Ambion), and hydrolyzed to smaller size (∼30–50 nt) with 80 mM sodium bicarbonate and 120 mM sodium carbonate (Catalanotto et al. 2002). After hydrolysis, the RNA probes were added into the hybridization solution and hybridized overnight at 50°.

Measurement of mitotic silencing frequency

To measure the mitotic silencing frequency in a given strain, 10 independent colonies were inoculated into 5-ml YPD overnight cultures. Cells were then harvested, washed, and plated on both YPD and 5-FOA media at various dilutions. Colony numbers formed on 5-FOA medium and on YPD medium were counted from each individual culture and the silencing frequency was determined based on the counted colony numbers. The final reported silencing frequency is the mean of the 10 individual silencing frequencies.

Transposon trapping and determining the frequency of FK506-resistant mutants

The FKBP12 encoding gene, FRR1, was used as a transposon trap as described in previous studies (Cruz et al. 1999; Wang et al. 2010). Ten independent colonies from wild type (XL280) and from two independent rdp1 mutants (XW150 and XW151) were inoculated into 5-ml overnight cultures in YPD medium. Cells were then harvested, washed, and plated on both YPD and YPD with FK506 (1 μg/ml) medium (37°) at various dilutions. To measure spontaneous FK506 resistance rate due to trapped transposons or mutations, the number of colonies that formed on YPD medium were counted and compared with those formed on YPD with FK506 (37°). The final reported FK506 resistance rates are the mean of the 10 individual resistance rates.

To measure the spontaneous FK506 resistance rate during same-sex mating, wild type (XL280) and two independent rdp1 mutants (XW150 and XW151) were solo cultured on V8 pH = 7 medium for 2 weeks. Spores were isolated from the edges of mating cultures and plated on both YPD and YPD with FK506 (1 μg/ml) medium (37°) at various dilutions. The FK506 resistance rate was determined as described above. The same procedure was repeated five times and the final reported FK506 resistance rates from progeny are the mean of the five individual resistance rates.

To analyze the transposons inserted into the FRR1 gene in the FK506-resistant isolates, sequence analysis was done as follows. First, genomic DNA was isolated and used as a template for PCR amplifications for the FRR1 gene with primers JOHE38889 (ACGGACAAAAACGACAGACC) and JOHE38990 (AGAATGCCGGTATCAACGAC). The resulting PCR products larger than the wild-type FRR1 gene were subjected to DNA sequencing. Primers JOHE38893 (TATCCTCACCACCTCTCGTCC), JOHE38894 (GTCGAGAGCTTCACTAAG), JOHE39000 (CCGTTCGTCAGTGCCATCAC), and JOHE39005 (TTGACGAAGAGAACATCGTC) were used for sequencing. The identified transposon sequences were submitted to GenBank with accession nos. KC440172, KC440173, KC440174, KC440175, KC440176, and KC469619.

Results

Tandem copies of the SXI2a-URA5 allele are silenced during mating in the serotype D strain

Previously, we observed a sex-induced silencing phenomenon that is triggered by a tandem multicopy insertion of a SXI2a-URA5 transgene during a-α opposite-sex mating in the C. neoformans serotype A var. grubii lineage (Wang et al. 2010). Same-sex mating that has been observed under laboratory conditions is mostly found in the C. neoformans serotype D lineage and rarely in serotype A (Lin et al. 2005, 2006; Bui et al. 2008). To test whether SIS operates in the C. neoformans serotype D var. neoformans lineage, especially during same-sex reproduction, we first utilized a plasmid containing the serotype A SXI2a-URA5 transgene that had successfully triggered silencing in the serotype A strains (Wang et al. 2010). This plasmid was introduced into a recipient strain XL280 ura5 (var. neoformans, serotype D) that undergoes robust α-α mating during solo culture on V8 medium (Figure 1A). This approach generated isolates with multiple copies of the transgene. Integration of the transgene at the ura5 locus was first confirmed by PCR and Southern analysis. The copy number of the SXI2a-URA5 transgene was then analyzed by real-time PCR with primers specific to URA5 and the ACT1 actin gene as a control. One isolate (XW73, ura5 SXI2a-URA5) contained ∼12 copies of the URA5 gene, indicating that 11 transgene copies and one endogenous copy are present in this strain (Figure 1B). This transgene strain was then solo cultured on V8 medium in the dark for 2 weeks at room temperature to induce sexual reproduction. As shown in Figure 1C, abundant hyphal growth with production of basidia and spores was observed. Basidiospores were isolated by microdissection and examined for silencing of the URA5 gene on SD-uracil and 5-FOA media. Among 19 progeny tested, 5 of them (∼25%) were found to be uracil auxotrophic. All of the progeny, including both those that were Ura⁺ and those that were ura⁻, inherited the SXI2a-URA5 transgene based on PCR analysis (Figure S1 and Figure 1D).

Figure 1.

Silencing of the SXI2a-URA5 transgene during solo-sexual mating. (A) Schematic illustration of the SXI2a-URA5 transgene introduced into the hyperfilamentous recipient α strain XL280 ura5. (B) Quantitative real-time PCR revealed that ∼12 copies of the URA5 gene (1-copy endogenous URA5 gene and 11-copy SXI2a-URA5 transgene) are present in the genome of the transformed strain XW73. (C) Hyphal growth of strain XW73 bearing the SXI2a-URA5 transgene after 2 weeks culture on V8 medium in the dark (left). A basidium with four long basidiospore chains produced by strain XW73 (right). Bar, 10 μm. (D) Basidiospores were isolated from sexual reproduction, and 5 among a total of 19 f1 progeny were auxotrophic for uracil (see Figure S1). XL280 ura5, ura5 SXI2a-URA5 (XW73), one Ura⁺ progeny, and ura⁻ progeny strains were grown on YPD and SD medium lacking uracil. PCR analysis showed that the SXI2a-URA5 transgene is still present in the uracil-auxotrophic progeny.

Transgene copy number changes or genomic rearrangements were excluded as causes of the loss of transgene expression as follows. First, to examine whether the transgene copy number had been altered in the ura⁻ progeny, quantitative real-time PCR was conducted to compare the copy number of the SXI2a-URA5 transgene in the ura⁻ and Ura⁺ progeny. Both the ura⁻ and Ura⁺ progeny contained ∼11–13 copies of the transgene (Figure 2A), indicating that the transgenes are stably inherited during solo-sexual reproduction. Second, Southern analysis was performed with a URA5 probe to define the genomic configuration of the ura⁻ and Ura+ strains at the SXI2a-URA5 locus. As shown in Figure 2, B and C, identical Southern patterns were observed in parental strain XW73 and its ura⁻ progeny. Thus, no rearrangement has occurred at the transgene locus during solo-sexual reproduction, thus excluding this as another possible explanation for extinction of expression of the transgene. Furthermore, when compared with the signal derived from the endogenous URA5 locus (∼7.5 kb), the signal from the introduced SXI2a-URA5 transgene allele (∼3.8 kb) was much more intense. This finding is consistent with the observation that multiple copies of the transgene were inserted at ura5, resulting in a tandem repeat configuration (Figure 2B).

Figure 2.

URA5 is silenced in uracil-auxotrophic progeny. (A) Quantitative real-time PCR revealed that the URA5 gene is present at a similar copy number in the parental strain XW73 bearing the SXI2a-URA5 transgene and both the Ura⁺ and the ura⁻ progeny. (B) The deduced arrangement of the tandem integration of the SXI2a-URA5 transgene is depicted. (C) DNA isolated from WT, ura5 SXI2a-URA5 (XW73), and the ura⁻ progeny was cleaved with PvuII and analyzed by Southern hybridization with a URA5 gene-specific probe. The endogenous URA5 gene resulted in an ∼7.5-kb signal, whereas the ∼3.8-kb signal represents the cleavage products of the integrated SXI2a-URA5 allele in the transformed strain and its progeny. (D) URA5 siRNAs are present in the ura⁻ progeny. Total RNA was isolated from strains XL280 ura5, ura5 SXI2a-URA5 (XW73), and the ura⁻ progeny strains. siRNAs were enriched and resolved in a 15% denaturing polyacrylamide TBE-urea gel and probed with a sense URA5 probe. Northern analysis revealed that URA5 siRNAs are highly abundant in ura⁻ progeny, but not in the parental strain ura5 SXI2a-URA5 (XW73) or the ura5 mutant strain (XL280 ura5).

To investigate whether efficient silencing of the URA5 transgene array is linked to the solo-sexual mating cycle, we examined the silencing frequency of the transgene during vegetative mitotic growth. Spontaneous 5-FOA resistance was measured after the parental transgene array strain XW73 was grown on rich YPD medium. Approximately 3–7 in every 10,000 isolates exhibited a 5-FOA resistant (ura⁻) phenotype, demonstrating a much lower mitotic silencing frequency (∼0.05%) compared to meiotic progeny (∼25%). Multiple copies of the transgene are required for silencing, as no silencing was observed in a strain containing only one copy of the transgene (Table 1). Both meiotic and mitotic silencing of the transgene were found to be unstable, and reversion of ura⁻ to Ura⁺ occurred frequently after serial passage on nonselective YPD medium. Taken together, our results show that multiple copies of the SXI2a-URA5 transgene are silenced during both vegetative mitotic growth and meiotic growth, while the frequency is much higher in the latter process, suggesting SIS also operates in the C. neoformans serotype D var. neoformans lineage and during solo-sexual mating.

Table 1. Multiple copies of the SXI2a-URA5 transgenes are silenced by unisexual reproduction.

	ura5 SXI2a-URA5	ura5 SXI2a-URA5	ura5 SXI2a-URA5 rdp1Δ	ura5 SXI2a-URA5 rdp1Δ	ura5 SXI2a-URA5
Transgene copy number	∼11	∼11	∼11	∼11	1
Growth condition	YPD	V8	YPD	V8	YPD
Silencing frequency (%)	0.05	∼25	<0.00001	<2.5	<0.00001

YPD, yeast colonies generated by mitotic growth; V8, dissected spores produced by meiotic sexual reproduction.

Previously, we have established a link between the RNAi pathway and SIS in the C. neoformans serotype A var. grubii lineage. In this study, XW73 is derived from the C. neoformans serotype D var. neoformans strain, which also contains genes encoding Argonaute, Dicer, and RdRP. Whole genome and genetic studies have identified two Argonaute genes (AGO1 and AGO2), two Dicer genes (DCR1 and DCR2), and one RDP1 gene encoding an RdRP in the serotype D strains JEC21 and B3501A (Loftus et al. 2005; Janbon et al. 2010; Wang et al. 2010). To verify that SIS in solo-sexual mating is also dependent on an RNAi pathway, we deleted the only RDP1 gene in the JEC21 derived strain XW73. Two independent deletion mutants were obtained and analyzed. We first examined the silencing frequency of the progeny produced from strain XW73 bearing the rdp1 deletion. Among 39 meiotic progeny (from mutant 1) and 41 meiotic progeny (from mutant 2) that were analyzed, no ura⁻ isolate was observed (0/39 and 0/41, <2.5%) (Figure S2), indicating that Rdp1 is required for SIS in solo-sexual mating. Moreover, we found that the rdp1 deletion also abolished the silencing during vegetative mitotic growth (<0.00001%) (Table 1).

Because another distinguishing feature of RNAi is the generation of siRNA from dsRNA, we next examined the presence of siRNA in the URA5 silenced strains. RNA was extracted from ura⁻ silenced progeny and analyzed by Northern hybridization. When probed with a ³²P-labeled sense URA5 transcript, abundant siRNAs ∼22 nt in size were detected in the silenced progeny, but not in the original XL280 ura5 strain or the parental XW73 strain bearing the SXI2a-URA5 transgene (Figure 2D). The small RNAs are antisense transcripts of URA5, suggesting that they are derived from a dsRNA precursor and function as a trans-acting factor during silencing. We also noted a difference in abundance of the accumulated small RNAs among different ura⁻ progeny. We hypothesize that those isolates with less abundant siRNAs are attributable to some cells in the population that have reverted to a Ura⁺ phenotype during growth in nonselective YPD medium and thus lack siRNAs. This can be reflected by our observation that some of the ura⁻ progeny exhibited a modest slow growth phenotype on YPD medium at 30° similar to strain XL280 ura5, which can be overcome by adding uracil to the YPD medium, indicating these progeny may have a stronger silencing phenotype than other ura⁻ progeny. In addition, to examine whether the silencing effect is restricted to URA5 or spreads into its neighboring gene RPS8 that encodes a ribosomal protein, we also probed the blot with a RPS8 sense probe but failed to detect any RPS8 siRNAs.

Construction of a and α strains with a NEO-URA5 transgene array

Unisexual mating and opposite-sex mating share similar inducing stimuli and elements of the pheromone response pathway (Hsueh et al. 2009; Lin et al. 2010; Wang and Lin 2011). However, a key difference is that the homeodomain cell identity determinants Sxi1α and Sxi2a are only required for postfusion steps during opposite-sex mating and are dispensable for unisexual reproduction (Hull et al. 2002, 2005; Lin et al. 2005). We have observed that a serotype A SXI2a-URA5 transgene array triggered silencing during solo-sexual mating of a serotype D strain (XW73). We also found that a wild-type serotype A SXI2a gene is able to complement mating defects of a serotype D sxi2aΔ mutant strain (Figure S3). Thus, a key question is whether the Sxi2a protein produced from the transgene in strain XW73 leads to a signaling cascade more like that operating during opposite-sex mating and possibly thereby enabling a-α sex-induced silencing. To test this possibility, we utilized a plasmid bearing a NEO-URA5 transgene to generate strains containing multiple copies of this transgene.

As shown in Figure 3A, we first introduced the NEO-URA5 plasmid into the serotype D strain JEC34 (JEC20 ura5). Integration of the NEO-URA5 transgene at the ura5 locus was confirmed by PCR and Southern analysis. As described above, the hybridization signals derived from the NEO-URA5 transgene and the endogenous ura5 allele were detected individually by Southern blot (Figure 3B). The transformed strain (XW105) exhibited two hybridization signals, one at ∼8.5 kb and the other more intense one at ∼3.8 kb, suggesting that multiple copies of the transgenes have been integrated into the genome. Compared with multiple hybridization bands from other transformants reflecting additional ectopic integration events, the simpler Southern pattern from strain XW105 indicated that the transgenes in this strain are only located at the ura5 locus and are arranged as tandem repeats (Figure 3B). We then crossed the a strain XW105 bearing the NEO-URA5 transgene with the α strain XL280 (hyperfilamentous) to obtain a hyperfilamentous f1 α progeny (XW139) carrying the same transgene array (Figure 3C). Quantitative PCR confirmed that the URA5-NEO trangene is present in the same copy number in the XW105 a and XW139 α strains. In total, four copies of the URA5 gene reflect a three-copy transgene and the one endogenous ura5 gene (Figure 3D). Additionally, based on stability assays conducted following repeated mass transfer of cells grown under nonselective growth conditions, the transgenes in all of the strains used in this study were stably inherited during mitotic growth.

Figure 3.

Construction of a and α strains with a NEO-URA5 transgene array. (A) Plasmid containing the NEO-URA5 transgene was introduced into the a strain JEC34, which is JEC20 ura5 by transformation. (B) DNA obtained from individual transformed strains was cleaved with XbaI and subjected to Southern analysis. When hybridized with a probe specific for URA5, an 8.5-kb product was derived from the endogenous URA5 and a 3.8-kb product reflects the cleavage product from the NEO-URA5 transgene. Strain XW105 (indicated by the red arrow), exhibiting the expected hybridization pattern with no ectopic integration and an intense signal from the transgene locus, was selected for further study. (C) a strain XW105 bearing the NEO-URA5 transgene was crossed with α strain XL280 to obtain an α progeny (XW139) that carries the same transgene array and is able to undergo robust unisexual reproduction. (D) Quantitative real-time PCR revealed that four copies of the URA5 gene are present in strains XW105a and XW139α.

SIS frequency is comparable during α-α unisexual and a-α opposite-sex mating

We have generated a and α strains containing the same copy number of the NEO-URA5 transgene, which allowed us to examine SIS in unisexual mating and also to compare the silencing frequency between α-α unisexual and a-α opposite-sex mating. Progeny were isolated by microdissection from both unisexual and opposite-sex mating and analyzed for URA5 silencing based on growth or absence on SD −uracil and 5-FOA media. When the α transgene array strain XW139 was solo cultured on V8 medium, 4 of a total of 28 progeny (14%) were auxotrophic for uracil (Figure S4 and Figure 4A). In comparison, among 31 progeny isolated from an a-α cross between strains XW105 a and XW139 α on V8 medium, 4 exhibited a ura⁻ phenotype (13%) (Figure S5 and Figure 4A). We also noted that the α strain XW139 might undergo some α-α unisexual mating even when cocultured with an a strain in the a-α cross. To establish that the ura⁻ progeny obtained are bona fide progeny of an a-α opposite-sex cross, a mating type-specific PCR amplification assay was performed to determine the mating type of these ura⁻ progeny using STE20a/α primers established in previous studies (Fraser et al. 2004; Li et al. 2012). An excess of α progeny would have been interpreted as indicative of a higher level of α-α mating occurring during culture with the a partner. As shown in Figure 4B, an equal proportion of a and α progeny (1:1) were recovered, providing evidence that they are indeed derived from a-α mating. In addition, the ura⁻ progeny also exhibited a neomycin sensitive (neo^s) phenotype, indicating that the NEO gene is also silenced. However, the neo^s phenotype is much more unstable compared with the ura⁻ phenotype, which is reflected by the fact that >50% of cells reverted to neomycin resistance after nonselective growth on YPD medium. In comparison, the reversion rate to Ura⁺ of different ura⁻ isolates was within the range of 1 × 10⁻⁴–1 × 10⁻⁶.

Figure 4.

SIS frequency is comparable during unisexual mating and opposite-sex mating. (A) ura⁻ progeny from α-α unisexual mating (4 of 28, 14%, see Figure S4) and from a-α opposite-sex mating (4 of 31, 13%, see Figure S5) were grown on YPD medium with or without neomycin and SD medium without uracil. Strains XW139, XW105, JEC20 ura5 (JEC34), and XL280 ura5 were included as controls. (B) Mating type of ura⁻ progeny derived from a cross between the α and a strains XW139 and XW105 was determined by a mating type-specific PCR amplification assay using the STE20a/α gene primers (top). PCR analysis confirmed the presence of the NEO-URA5 transgene in the uracil-auxotrophic progeny (bottom). (C) ura⁻ progeny derived from either α-α unisexual or a-α opposite-sex mating were subjected to Southern analysis with a URA5 probe. The same hybridization pattern was observed in ura⁻ progeny and in the parental strains XW105 and XW139. (D) Quantitative real-time PCR revealed that the URA5 gene is present at the same copy number in all of the ura⁻ progeny. (E) URA5 siRNAs are present in the ura⁻ progeny based on hybridization with a URA5 sense probe.

PCR analysis showed that the NEO-URA5 transgenes were still present in all of the progeny with a silenced URA5 gene (Figure 4B). We also examined the genomic structures at the NEO-URA5 locus in these ura⁻ progeny by Southern analysis with a probe to the URA5 gene. As shown in Figure 4C, all of the ura⁻ progeny displayed an identical hybridization pattern compared to the parental transgene array strains XW105 and XW139. The intensity of the signal derived from the transgene indicated that multiple copies of the NEO-URA5 transgene have been integrated into the genome. Furthermore, quantitative real-time PCR results confirmed that the ura⁻ progeny still contain the same copy number of the NEO-URA5 transgene (Figure 4D). In addition, as we observed before in strain XW73 (Figure S2), the silencing occurring during unisexual mating was also abolished when RDP1 was deleted in the α strain XW139. We examined 29 progeny and 27 progeny from two rdp1Δ independent mutants, and no isolate was ura⁻ (0/29 and 0/27) (Figure 5 and Figure S6). Northern blot analysis was performed to detect siRNAs in the ura⁻ progeny. When probed with a sense ³²P-labeled URA5 transcript, abundant siRNAs were observed in the majority of the ura⁻ strains (Figure 4E). The ura⁻ isolates showed various levels of siRNAs, and the ones exhibiting less abundant siRNAs correspond to those with a higher reversion rate to Ura⁺. We also probed the same blot with a NEO probe, but failed to detect any siRNAs. This could be related to the high reversion rate of the NEO silencing phenotype and rapid loss of less abundant siRNAs may lead to the rapid phenotypic reversion.

Figure 5.

rdp1 deletion abolishes silencing of the NEO-URA5 transgene during α-α unisexual mating. The RDP1 gene was deleted in the α strain XW139. Twenty-nine progeny were isolated from α-α unisexual mating of the rdp1Δ mutant strain (XW211) and examined for silencing of the URA5 gene by growth on 5-FOA, SD-uracil, and YPD medium (streaked in that order). Strains JEC20 and JEC20 ura5 (JEC34) were included as controls. All of the plates were incubated at room temperature for 3 days and photographed. No isolates showed a ura⁻- or 5-FOA-resistant phenotype, other than the ura5 control strain. See Figure S6 for a second independent rdp1Δ mutant with an identical phenotype.

Taken together, our results showed that the silencing rates during a-α opposite-sex mating and α-α unisexual mating are similar (13 vs. 14%). Additionally, both the a strain XW105 and the α strain XW139 exhibited a similar low silencing rate (∼0.04%) during vegetative mitotic growth on YPD medium (Table 2), demonstrating that the silencing rate increased ∼350-fold during either α-α unisexual or a-α heterosexual reproduction.

Table 2. SIS frequency is comparable during unisexual mating and opposite-sex mating.

	NEO-URA5 MATa	NEO-URA5 MATα	NEO-URA5 MATa x MATα	NEO-URA5 MATα
Growth condition	YPD	YPD	V8	V8
Silencing frequency (%)	0.04	0.05	13	14

YPD, yeast colonies generated by mitotic growth; V8, dissected spores produced by meiotic sexual reproduction.

Transposons are more active in the rdp1 mutant strains during α-α same-sex mating

Previously, we have observed that transposons are hyperactive during mating of serotype A Cryptococcus strains and that the SIS RNAi pathway plays a critical role in squelching transposon activity during a-α opposite-sex mating. We hypothesized that transposon activity would also be suppressed during α-α same-sex mating by SIS.

To address this, the FKBP12 encoding gene, FRR1, was used as a transposon trap, as described in previous studies (Cruz et al. 1999; Wang et al. 2010). FKBP12 is the intracellular receptor for the antifungal drug FK506, which is toxic to C. neoformans at 37° when bound to FKBP12 via inhibition of calcineurin (Odom et al. 1997). Spores resulting from α-α unisexual reproduction of WT strain XL280 and two independent rdp1 mutants were germinated on YPD medium with or without FK506 and the rate of spontaneous FK506 resistance was examined. The rate of FK506 resistance during mitotic growth of XL280 and the rdp1 mutant strains was also assessed. As shown in Table 3, both rdp1 mutants generated more FK506 resistant isolates than their wild-type parental strain XL280 under both mating and mitotic growth conditions. The progeny obtained from two rdp1 mutants yielded the highest rate of FK506 resistance (4 × 10⁻⁶ or 3 × 10⁻⁶), which is ∼15–20 times higher than that of meiotic spores produced from the WT strain (2 × 10⁻⁷) and ∼50–60 times higher than the frequency observed during mitotic growth of the WT strain (6 × 10⁻⁸).

Table 3. Higher FK506-resistance rate in the rdp1 mutant strains during α-α unisexual mating.

	WT	rdp1	rdp1
Progeny produced by meiotic reproduction	2 × 10−7	4 × 10−6	3 × 10−6
	∼3	∼67	∼50
Yeast colonies generated by mitotic growth	6 × 10−8	6 × 10−7	9 × 10−7
	1a	∼10	∼15

^aFK506-resistance rate from the WT strain during mitotic growth was set as 1, and the resistance rates obtained from the rdp1 mutant strains were normalized to this.

To determine whether the FK506-resistance phenotype is attributable to spontaneous frr1 point mutations or transposon insertions into this gene, the FRR1 gene was PCR amplified from genomic DNA isolated from the FK506-resistant progeny derived from the rdp1 mutants. Among a total of 50 progeny examined, a PCR product of the expected size was obtained from 15 progeny, whereas increased size PCR products were observed from the remaining 35 progeny. A PCR product derived from the FRR1 locus was obtained from all 50 isolates, excluding deletions across either primer site or translocations. The PCR products larger than wild type from 10 isolates were subjected to DNA sequencing. The results suggested six different transposon trapping events occurred: three DNA transposons, T1, T2, and T3 had been inserted into the FRR1 open reading frame at different locations or in different orientations (details in Figure 6A). These DNA transposons have been identified in Cryptococcus from two independent labs, and each contains a transposase gene and terminal inverted repeats (Cruz et al. 1999; M. C. Cruz and J. Heitman, unpublished data; Janbon et al. 2010). The T2 transposon exists in the genome as a longer (∼3.5 kb) and a shorter (∼0.8 kb) version, and we found both versions integrated into the FRR1 gene in different FK506-resistant rdp1 isolates (Figure 6A).

Figure 6.

Transposons are hyperactive in the rdp1 mutant strains during α-α unisexual reproduction. (A) Schematic illustration of DNA transposon insertions into the FRR1 open reading frame detected by transposon trapping and DNA sequencing. The wild type and the rdp1 mutant strains are sensitive to FK506 at 37°. The FK506-resistant progeny resulting from α-α unisexual reproduction of the rdp1 mutant strain contain a mutant frr1 gene. Six different transposon trapping events were found: DNA transposon T1 (in XW238 and XW240), T2 (shorter version in XW235 and longer version in XW231 and XW233), and T3 (in XW237) had integrated into the FRR1 open reading frame at different locations or in different orientations. The arrows above the frr1 gene represent the orientation of the inserted transposons. (B) Total RNAs were isolated from the indicated strains grown on YPD medium or V8 medium (pH = 7). Northern blot analyses were conducted to analyze the expression levels of the T2 and T3 transposase genes. GPD1 served as a loading control.

In accord with the higher transposition rate observed in the progeny from the rdp1 mutant strains, Northern blot analysis also showed that the genes encoding the T2 and T3 transposases were highly expressed during α-α same-sex mating of the rdp1 mutants (Figure 6B). The expression levels of these genes were either barely detectable (transposon T2) or reduced (transposon T3) when the rdp1 mutants were mitotically grown on YPD medium. The WT strain XL280 consistently showed lower transcript levels of the transposase genes compared with the rdp1 mutants. Taken together, our results suggest that SIS functions during α-α unisexual reproduction to repress transposon activity and thus guards integrity of the genome.

Discussion

SIS was first identified as operating during opposite-sex a-α mating of C. neoformans serotype A var. grubii strains (Wang et al. 2010). Beyond silencing exogenous transgene arrays, SIS also plays a critical role in transposon control during sexual reproduction. This is reflected in our previous observations that progeny derived from RNAi mutants exhibited an increased transposition/mutation rate and abundant Rdp1-dependent siRNAs that mapped to repetitive transposable elements (Wang et al. 2010). Thus, SIS serves as a meiotic genome defense mechanism in a-α opposite sex mating of C. neoformans.

However, the natural Cryptococcus population is almost exclusively of the α mating type (Hull and Heitman 2002), and thus unisexual reproduction may be the predominant route to produce meiotic progeny and spores in nature. In fact, evidence from population genetic analyses has shown that unisexual reproduction occurs naturally and contributes to the population structure (Lin et al. 2007, 2009; Bui et al. 2008). Thus, uncovering whether SIS occurs during α-α unisexual reproduction is a key question to understand how the Cryptococcus genome has been shaped through this unique life cycle.

The results presented in this work demonstrate that SIS also operates during solo and α-α unisexual mating. A sexually precocious strain bearing a tandem multicopy NEO-URA5 transgene, when undergoing unisexual reproduction, exhibited ∼350-fold higher silencing frequency (∼14%) than during mitotic vegetative growth (∼0.04%). Additionally, when this strain was crossed to an a strain containing the same transgene array, a comparable silencing rate was found (Table 2). These findings indicate that SIS operates at a similar efficiency during unisexual reproduction and opposite-sex mating.

The role of RNAi as a mechanism to defend genomes against mobile transposons is broadly conserved among plants, animals, fungi, and ciliates (Ketting et al. 1999; Llave et al. 2002; Aravin et al. 2003; Drinnenberg et al. 2009). It has been shown that the RNAi pathways function to control the mobility of transposons in serotype D C. neoformans during vegetative mitotic growth (Janbon et al. 2010; Wang et al. 2010). In our study, not only did we observe an increased transposition rate and highly expressed transposase genes in the rdp1 mutant strains lacking the RNA-dependent RNA polymerase required for SIS, we also found that these transposons are more active during same-sex mating than during vegetative growth (Table 3 and Figure 6). This observation is similar to our finding with respect to SIS in opposite-sex mating (Wang et al. 2010). These findings lead us to propose that the potential high mutational burden due to hyperactive transposons during mating necessitates a mechanism like SIS, which is much more robust than silencing occurring during vegetative mitotic growth, to defend against transposition and thereby guard genomic integrity of the progeny.

Unisexual and opposite-sex mating involve many shared components, including the Cpk1 MAPK signal transduction cascade that responds to pheromones and governs the dimorphic transition in C. neoformans (Wang and Lin 2011; Hsueh et al. 2009; Lin et al. 2010). However, the homeodomain cell identity protein Sxi2a encoded by the a mating type locus is present and functions only during a-α mating but not α-α mating. Sxi2a forms a heterodimer with Sxi1α and functions to control postfusion hyphal development in a-α opposite-sex mating (Hull et al. 2002, 2005). In this study, we have introduced a heterologous serotype A SXI2a-URA5 transgene into a self-fertile α serotype D strain. This haploid strain XW73, thus may express the Sxi1α/Sxi2a complex similar to a-α mating and also exhibited RNAi-dependent SIS (Figure 2 and Table 1) when subjected to self-mating. SIS frequency in this case is higher (25 vs. 14%) than what we observed in the α strain (XW139) containing a NEO-URA5 transgene array. Given the previous findings that transgene copy number is linked to the silencing efficiency (Wang et al. 2010, 2012) together with the observation that similar SIS rates were found in a-α opposite-sex mating and α-α unisexual mating (Table 2), we propose that the higher SIS rate occurring in strain XW73 is attributable to the presence of a higher copy number of the SXI2a-URA5 transgene (11 copies of SXI2a-URA5 vs. 3 copies of NEO-URA5). Given that silencing occurred in α strains with transgenes both containing or lacking SXI2a, we conclude that SXI2a is not essential for the SIS pathway, and the shared pheromone-sensing Cpk1 MAPK signal transduction cascade in both unisexual and opposite-sex mating may be involved in triggering SIS.

Our results not only demonstrate that SIS operates during α-α unisexual reproduction, but demonstrate that SIS is conserved among different Cryptococcus varieties. The first SIS phenomenon was found in the C. neoformans serotype A variety grubii lineage. In this study, we observed that SIS occurs in both opposite-sex and unisexual mating of the serotype D variety neoformans strains. We also noted that SIS took place at an ∼50% frequency in the serotype A mating as reported before (Wang et al. 2010), which is somewhat higher than that in the serotype D mating (14%) induced by similar copies of the URA5 transgene. We first suspected that the lower efficiency might be due to use of the heterologous serotype A URA5 promoter and coding sequence in the transgene, resulting in less efficient silencing in serotype D strains. However, after replacing these elements with the serotype D derived sequences, a similar silencing frequency (∼13%) was observed in the serotype D mating progeny (data not shown). In addition, the silencing triggered by the SXI2a-URA5 transgene during vegetative growth also showed a lower frequency in the serotype D strain (0.05%, Table 1) compared with that in the serotype A strain (0.2%) (Wang et al. 2010). Furthermore, a similar phenomenon was reported previously that serotype A and serotype D strains exhibited different RNAi silencing efficiencies when responding to the same hairpin RNA construct (Skowyra and Doering 2012). Taken together, we submit that the SIS/RNAi rate in C. neoformans may be serotype specific and occurs at an approximately fourfold higher level in serotype A compared to D.

The facts that (1) mutation of RDP1 blocked SIS and (2) abundant antisense siRNAs mapping to repetitive transgene clusters were identified in the silenced progeny indicate that SIS is an RNAi-dependent silencing pathway. RNAi components are conserved in most of the Cryptococcus species, including C. neoformans (serotype A and D) and its sibling species C. gattii (Loftus et al. 2005; Janbon et al. 2010; Wang et al. 2010). C. gattii can be divided into four molecular types (VGI, VGII, VGIII, and VGIV) (Fraser et al. 2005; Bovers et al. 2008). Strikingly, C. gattii VGII strains are the only ones that lack RNAi pathway elements. In the genome of the C. gattii VGII strain R265, both Argonaute genes are absent and RDP1 and DCR1 are pseudogenes (D'Souza et al. 2011; B. Billmyre and J. Heitman, unpublished results). Notably, this RNAi deficient C. gattii lineage is the same one responsible for an ongoing expanding disease outbreak on Vancouver Island and the Pacific Northwest. This coincidence brings up a possibility that loss of RNAi could be beneficial under certain circumstances during evolution. Data have been shown to support that same-sexual mating occurring between two closely related VGIIα strains, the major Vancouver Island genotype isolates, may contribute to the production of infectious spores (Fraser et al. 2005). In this case, the absence of the SIS-RNAi silencing pathway in the same-sex mating between VGII strains may lead to more rapid genetic changes and result in hypervirulent progeny, thus providing a mechanism by which genetic diversity is increased during same sex reproduction within populations. Ongoing studies are being conducted to examine this mode of rapid adaptation in the C. gattii VGII strains.