The Pheromone and Pheromone Receptor Mating-Type Locus Is Involved in Controlling Uniparental Mitochondrial Inheritance in Cryptococcus

Abstract

Mitochondria are inherited uniparentally during sexual reproduction in the majority of eukaryotic species studied, including humans, mice, and nematodes, as well as many fungal species. Mitochondrial uniparental inheritance (mito-UPI) could be beneficial in that it avoids possible genetic conflicts between organelles with different genetic backgrounds, as recently shown in mice, and it could prevent the spread of selfish genetic elements in the mitochondrial genome. Despite the prevalence of observed mito-UPI, the underlying mechanisms and the genes involved in controlling this non-Mendelian inheritance are poorly understood in many species. In Cryptococcus neoformans, a human pathogenic basidiomyceteous fungus, mating types (MATα and MATa) are defined by alternate alleles at the single MAT locus that evolved from fusion of the two MAT loci (P/R encoding pheromones and pheromone receptors, and HD encoding homeodomain transcription factors) that are the ancestral state in the basidiomycota. Mitochondria are inherited uniparentally from the MATa parent in C. neoformans, and this requires the SXI1α and SXI2a HD factors encoded by MAT. However, there is evidence that additional genes contribute to the control of mito-UPI in Cryptococcus. Here, we show that in C. amylolentus, a sibling species of C. neoformans with unlinked P/R and HD MAT loci, mito-UPI is controlled by the P/R locus and is independent of the HD locus. Consistently, by replacing the MATα alleles of the pheromones (MF) and pheromone receptor (STE3) with the MATa alleles, we show that these P/R locus-defining genes indeed affect mito-UPI in C. neoformans during sexual reproduction. Additionally, we show that during early stages of C. neoformans sexual reproduction, conjugation tubes are always produced by the MATα cells, resulting in unidirectional migration of the MATα nucleus into the MATa cell during zygote formation. This process is controlled by the P/R locus and could serve to physically restrict movement of MATα mitochondria in the zygotes, and thereby contribute to mito-UPI. We propose a model in which both physical and genetic mechanisms function in concert to prevent the coexistence of mitochondria from the two parents in the zygote, and subsequently in the meiotic progeny, thus ensuring mito-UPI in pathogenic Cryptococcus, as well as in closely related nonpathogenic species. The implications of these findings are discussed in the context of the evolution of mito-UPI in fungi and other more diverse eukaryotes.

MITOCHONDRIA are important eukaryotic organelles. In addition to providing cellular energy, they are involved in a variety of cellular processes, such as signaling, cellular differentiation, and cell death, as well as control of the cell cycle and cell growth (McBride et al. 2006). Furthermore, they are implicated in several human diseases, such as mitochondrial disorders and cardiac dysfunction, and may play a role in aging (Lesnefsky et al. 2001; Gardner and Boles 2005). In pathogenic fungi, such as Cryptococcus neoformans, mitochondria play critical roles in virulence, survival under low-oxygen conditions, and drug tolerance (Ingavale et al. 2008; Ma et al. 2009; Ma and May 2010; Shingu-Vazquez and Traven 2011; Kretschmer et al. 2012). Additionally, given the effects of mitochondrial dysfunction on fungal drug tolerance and virulence, these organelles are potential antifungal drug-development targets (Shingu-Vazquez and Traven 2011).

Unlike nuclear genes and genomes, the inheritance of mitochondrial (as well as chloroplast) genes and genomes does not follow Mendel’s laws. These organelle genomes are typically uniparentally inherited in the majority of species that have been studied. There are currently two hypotheses for the evolution of mitochondrial uniparental inheritance (mito-UPI). First, mito-UPI evolved as a mechanism to restrict the spread of mitochondrial selfish elements that enhance mitochondrial fitness to the detriment of their host. Second, mito-UPI evolved to avoid having two genetically different mitochondrial genomes in the zygote simultaneously. This prevents mitochondrial heteroplasmy and/or recombination, which are thought to either generate less-fit genomes, or cause nuclear–mitochondrial/mitochondrial–mitochondrial incompatibilities. Although it is not yet known why mitochondrial heteroplasmy is deleterious, mice that inherit both parental mitochondrial genomes exhibit signs of cellular or organismal imbalance associated with a variety of phenotypes, including behavioral and cognitive abnormalities (Lane 2012; Sharpley et al. 2012). This could be the result of Muller–Dobzhansky-type incompatibilities when one nuclear genome must work in concert with two distinct types of mitochondrial genomes (Maheshwari and Barbash 2011).

Among fungal species that have been examined, the majority exhibit mito-UPI, that is, the progeny all possess a mitochondrial genome inherited from one of the two mating parents. Examples of fungal mito-UPI include Aspergillus nidulans, Neurospora crassa, Candida albicans, Coprinopsis cinerea, Agaricus bisporus, C. neoformans, and Ustilago maydis (Xu 2005; Gyawali and Lin 2011; Ni et al. 2011; Goodenough and Heitman 2014; Shakya and Idnurm 2014). By contrast, in some fungal species such as Saccharomyces cerevisiae and Schizosaccharomyces pombe, inheritance of mitochondrial genomes is biparental. In these species, mating between isogametic sexual partners results in an equal contribution of organelles from the two gametes into the zygote, and the transient coexistence of two different mitochondria often results in recombination (Egal et al. 1980; Birky 1995, 2001). However, even in cases of biparental inheritance, homoplasy is rapidly reestablished after the initial heteroplasmic zygote, and each of the daughter cells possesses the mitochondrion of one genetic type, either from one of the two parents or a recombinant of the two (Birky 1995, 2001; Berger and Yaffe 2000; Barr et al. 2005).

Several mechanisms have been proposed to explain mito-UPI. In anisogametic species, such as animals and plants, the size differences between male and female gametes result in biased contributions of organelles in the zygotes, which could serve to restrict the transmission of mitochondria from the paternal parent. In addition, active mitochondrial marking and degradation during zygote formation have also been identified in several species (Al Rawi et al. 2011; Levine and Elazar 2011; Sato and Sato 2011; Zhou et al. 2011; Luo et al. 2013). For example, ubiquitin marking followed by mitophagy of unmarked sperm mitochondria shortly after zygote formation has recently been implicated in mito-UPI in Caenorhabditis elegans.

Most fungi are isogametic, and mating occurs between two gametes of the same or similar size, or between two compatible mycelia. Thus, unlike plants and animals, where size differences between the gametes contribute to unequal organelle contributions to the zygote and subsequently facilitate uniparental organelle inheritance, in fungi, several different mechanisms have evolved that could actively avoid mitochondrial heteroplasy during sexual reproduction (Wilson and Xu 2012). For example, in some fungal species (e.g., Co. cinerea), mating between two compatible mycelia is achieved by unidirectional migration of nuclei while all of the cytoplasm including the organelles is left behind; thus, the mixing of two different mitochondria is avoided (Hintz et al. 1988; May and Taylor 1988).

In unicellular fungi where mating occurs between two isogametic mating partners, there is evidence that organelles from one parent are actively degraded, thus ensuring homoplasy in the zygote. One example is the plant pathogenic fungus U. maydis (Fedler et al. 2009). U. maydis is a basidiomycete and has a tetrapolar mating system, constituted by the a and b loci. The biallelic a locus (a1 and a2) is involved in pheromone- and pheromone receptor-based cell recognition and fusion, while the multiallelic b locus encodes the homeodomain transcription factors. In U. maydis, mitochondrial inheritance is governed by the a2-specific genes LGA2 and RGA2. The Lga2 and Rga2 proteins both localize to mitochondria, and Lga2 interferes with mitochondrial dynamics and fusion (Bortfeld et al. 2004; Mahlert et al. 2009). Evidence also supports an active Lga2- and Rga2-mediated selective mitochondrial elimination process. Specifically, Lga2 is the destroyer, produced by the a2 locus and needed to destroy the mitochondria from the a1 parent, while Rga2 is the protector, produced by the a2 locus and needed to protect the a2 mitochondria from mitophagic destruction. However, the fact that deletion of RGA2 reverses the inheritance in favor of the a1-type mitochondria (rather than a biparental pattern) indicates that another RGA2-independent mechanism might also exist that is involved in the control of mitochondrial inheritance (Fedler et al. 2009), although it has been reported that the mitophagy-related gene ATG11 is not required for mito-UPI in U. maydis (Wagner-Vogel et al. 2015).

C. neoformans is a human pathogenic basidiomycete fungus that was classified into two varieties, var. grubii (serotype A) and var. neoformans (serotype D), that are now recognized as distinct species: C. neoformans (serotype A) and C. deneoformans (serotype D) (Hagen et al. 2015). C. neoformans has a defined bipolar mating system with two mating types (α and a), which are defined by the alleles present at the mating-type locus (MAT) (Kwon-Chung 1975). Compared to other basidiomycetes, the C. neoformans MAT locus is unusually large, spanning > 100 kb and containing > 20 genes. Studies have shown that the MAT locus in C. neoformans is a fusion product of the ancestral P/R and HD loci, and thus includes genes encoding both homeodomain transcription factors (HD), as well as pheromones and pheromone receptors (P/R) (Loftus et al. 2005; Hsueh et al. 2011; Sun et al. 2017; Passer et al. 2019). Mating in C. neoformans typically occurs between strains of opposite mating types (α and a), and mitochondrial inheritance during these opposite-sex matings is uniparental from the a parent (Xu et al. 2000; Yan and Xu 2003).

Opposite-sex mating can also occur between strains of different species (C. neoformans and C. deneoformans), and mito-UPI from the MATa parent remains intact during these interspecies opposite-sex matings (Yan et al. 2007a). Additionally, mating in C. neoformans can also occur between two α strains, and mitochondrial inheritance during these same-sex matings has been shown to be biparental from both parents (Yan et al. 2007a). Mito-UPI in C. neoformans is established rapidly after zygote formation, and several genes play important roles in this process. For example, the two homeodomain transcription factors SXI1α (in MATα) and SXI2a (in MATa) are both required to ensure mito-UPI, and deletion of either gene results in “leakage” of mitochondria from the MATα parent to the zygote (Yan et al. 2004, 2007a). Interestingly, when the SXI1α and SXI2a genes were exchanged between mating types, progeny still inherited mitochondria from the MATa parent, which then had a sxi2aΔ deletion and a transgenic copy of SXI1α, indicating that other genes in the MAT locus are required for mito-UPI (Hsueh et al. 2008). Recently, Mat2, which is a pheromone-activated transcription factor not encoded by MAT, was shown to be involved in mito-UPI in C. neoformans (Gyawali and Lin 2013). However, it is still not known what gene(s) in the C. neoformans MAT locus is the master regulator that initiates mito-UPI. Additionally, it is yet to be determined whether mito-UPI in C. neoformans also involves active mitochondrial marking and subsequent degradation via mitophagy or other processes, similar to mechanisms operating in C. elegans and U. maydis (Fedler et al. 2009; Al Rawi et al. 2011; Levine and Elazar 2011; Sato and Sato 2011; Zhou et al. 2011; Luo et al. 2013).

C. amylolentus, together with C. floricola and C. wingfieldii, are the most closely related known sibling species of the C. neoformans/C. gattii pathogenic species complex (Findley et al. 2009; Passer et al. 2019). C. amylolentus has a tetrapolar mating system with the two mating-type loci (A and B) located on different chromosomes. The A (P/R) locus is ∼100 kb in size and encodes the pheromones and pheromone receptors (and others), while the B (HD) locus is <10 kb in size and contains both SXI1α and SXI2a homologs (Findley et al. 2012). During C. amylolentus sexual reproduction, mitochondrial inheritance is uniparental and from the A2B2 parent (Findley et al. 2012). However, due to a lack of proper mitochondrial–nuclear marker combinations, we have not been able to further dissect which MAT locus—A, B or both—is responsible for mito-UPI in C. amylolentus.

In this study, we first obtained C. amylolentus isolates that had all eight possible mating type–mitochondrial combinations (HD–P/R–Mito). By analyzing mitochondrial inheritance in all possible pairwise crosses, we determined that mito-UPI in C. amylolentus is controlled by the P/R locus that encodes the pheromones and pheromone receptors, and is independent of the allele present at the HD locus. Additionally, we showed that in C. neoformans, the defining genes of the P/R locus in basidiomycetes (the pheromone and pheromone receptor genes) are indeed influencing mitochondrial inheritance during sexual reproduction. We also observed unidirectional conjugation tube formation from MATα cells, a process controlled by the P/R locus, and subsequent polarized hyphal formation from the zygote, which together could act as physical barriers that prevent mitochondria of MATα cells from spreading into the hyphae produced by the polarized zygote. Furthermore, we showed that deleting the CRG1 gene, which encodes a regulator of G-protein signaling (RGS) that negatively regulates the pheromone-signaling cascade and cyclic AMP (cAMP) pathways during sexual reproduction, resulted in MATα and MATa strains with elevated responses to pheromones during mating. However, this enhanced response to pheromone was significantly more pronounced in the MATα strains. Interestingly, progeny dissected from the crg1Δ bilateral crosses showed increased levels of mitochondrial leakage from the MATα cells, consistent with the dynamics of pheromone sensing and the initial stages of sexual development playing an important role in mito-UPI. Taken together, we propose an integrated model that involves both physical and genetic mechanisms operating coordinately to enforce mito-UPI in C. neoformans, which is consistent with previous studies and could serve as a foundation for future research on mitochondrial inheritance in C. neoformans, as well as in closely related basidiomycetes and beyond.

Materials and Methods

Strains employed in this study

Strains used in this study, as well as their genotypes, are listed in Table 1. All strains were maintained in −80° frozen stocks in 15% glycerol and subcultured from freezer stocks to YPD solid medium for study. Genotypic and phenotypic markers for the two laboratory-constructed C. neoformans strains are listed in Table 1, as well as in previous studies where these strains were originally published (Hsueh et al. 2008; Stanton et al. 2010).

Table 1. Strains analyzed in this study.

Strain	Species	Genotype	Mitochondrial type	Note
CBS6039	C. amylolentus	A1B1	a	Findley et al. (2012); Sun et al. (2017)
SSA033	C. amylolentus	A1B1	a	This study
SSA026	C. amylolentus	A1B2	a	This study
SSA028	C. amylolentus	A1B2	a	This study
SSB817	C. amylolentus	A1B1	a	This study
SSB821	C. amylolentus	A1B1	a	This study
SSA032	C. amylolentus	A1B2	a	This study
CBS6273	C. amylolentus	A2B2	b	Findley et al. (2012); Sun et al. (2017)
SSA067	C. amylolentus	A1B1	b	This study
SSA108	C. amylolentus	A1B1	b	This study
SSA790	C. amylolentus	A1B2	b	This study
SSA770	C. amylolentus	A1B2	b	This study
SSA052	C. amylolentus	A2B1	b	This study
SSA058	C. amylolentus	A2B1	b	This study
SSA089	C. amylolentus	A2B2	b	This study
SSA797	C. amylolentus	A2B2	b	This study
YPH716 (α → alf1)	C. neoformans	ura5 sxi1αΔ::NEO SXI2a-URA5 MFα1,2,3 STE3α	A	Hsueh et al. (2008)
CHY1517 (α → alf2)	C. deneoformans	ura5 mfα1Δ::ADE2 mfα2,3Δ::URA5 ste3αΔ::STE3a-NAT MFa1-NEO	D	Stanton et al. (2010)
JEC20 ura5	C. deneoformans	MATa ura5	D
M001	C. neoformans	MATα ade2 (UV irradiation)	A	Perfect et al. (1993b)
M049	C. neoformans	MATα ade2 (Gamma irradiation)	A	Perfect et al. (1993b)
H99	C. neoformans	MATα wild-type	H99	Perfect et al. (1993a)
KN99a	C. neoformans	MATa wild-type	KN99a	Nielsen et al. (2003)
YPH276	C. neoformans	MATα crg1Δ::NAT	H99	Hsueh et al. (2009)
YPH570	C. neoformans	MATa crg1Δ::NAT	KN99a	Hsueh et al. (2009)
SSH116	C. neoformans	MATα GFP-HEM15-NAT mCherry-CNA1-NEO	H99	This study
SSG269	C. neoformans	MATa GFP-HEM15-NAT	KN99a	This study

Construction of deletion strains for testing the effects of individual genes on mito-UPI

For the bilateral crosses labeled as “H-K” in the Supplemental Material, Table S2, deletion strains were constructed in the H99 (MATα) and KN99a (MATa) backgrounds. For unilateral crosses labeled as “H-A”, such as those for the MYO2 gene, the deletion strain was constructed by first generating a heterozygous deletion strain in the AI187 background. The AI187 heterozygous deletion strain was then induced to sporulate and the MATa progeny that inherited the deletion allele were recovered from randomly dissected basidiospores, as described in previous studies. Mito-UPI was subsequently assessed in crosses between the MATa progeny with the deletion mutation and the Bt63 wild-type strain. For the genes RCV1 and YPT7, the deletion strains were constructed in the KN99α background, and they were subsequently crossed with Bt63 and H99 wild-type strains, respectively, to assess the mito-UPI in these unilateral crosses.

Laboratory crosses for analyzing mito-UPI in C. amylolentus and C. neoformans

For the strains used for crosses in C. amylolentus, other than the two natural isolates, CBS6039 and CBS6273, all of the other strains were meiotic progeny (basidiospores) dissected from crosses between CBS6039 and CBS6273. Three pairs of meiotic progeny—SSA026 and SSA028, SSA032 and SSA033, and SSA052 and SSA058—were each derived from three different basidia, and thus three independent meiotic events. All of the other progeny employed for crosses were from different basidia, thus representing independent meiotic events. Because we did not have selectable markers to select for fusion products in C. amylolentus, it was not known when mito-UPI was established during sexual reproduction. Therefore, we chose to dissect multiple spore chains representing independent meiotic events from each cross to interpret the pattern of mitochondrial inheritance and the origin of the mitochondria in the meiotic progeny.

For C. neoformans, mitochondrial inheritance was analyzed in three crosses between serotype A and D isolates. First, mitochondrial inheritance was tested in a cross between strains CHY1517 and YPH716. CHY1517 has a serotype D MATα background. While the SXI1α gene is intact, the MATα genes encoding pheromones (MFα1, 2, and 3) and pheromone receptors (STE3α) have been deleted, and replaced with MATa copies (MFa and STE3a). In addition, strain CHY1517 also has two phenotypic markers: ura5 and NAT^R (Figure 1 and Table 1) (Stanton et al. 2010). Strain YPH716 has a serotype A MATα background, where the SXI1α gene has been deleted and a copy of the SXI2a gene from the MATa background has been transgenically introduced into its genome at the URA5 locus (Figure 1 and Table 1) (Hsueh et al. 2008). For comparison, mitochondrial inheritance was analyzed in two typical opposite sex crosses: JEC20 ura5 × M001 (H99 ade2) and JEC20 ura5 × M049 (H99 ade2). In these two crosses, the mitochondrial inheritance was expected to be uniparental from JEC20 ura5 (MATa), based on previous reports (Yan and Xu 2003; Yan et al. 2007a).

Figure 1.

Illustration of the pheromones, pheromone receptors, and homeodomain transcription factors of the trans-mating type-engineered C. neoformans strains. YPH716 (top) has a serotype A, MATα genetic background and serotype A mitochondria (red ovals). The SXI1α homeodomain transcription factor gene located in the MAT locus was deleted, and a copy of the SXI2a gene was transgenically inserted at the URA5 locus of the genome (Hsueh et al. 2008). Thus, during mating, YPH716 expresses α pheromones and the MFa pheromone receptor, but the a homeodomain transcription factor. CHY1517 (bottom) has a serotype D, MATα genetic background and serotype D mitochondria (blue ovals). The three α mating pheromone genes (MFα1, 2, and 3) and the α pheromone receptor (STE3α) were deleted from the MAT locus, and replaced with copies of the a mating pheromone (MFa1) and pheromone receptor (STE3a), respectively (Stanton et al. 2010). Thus, during mating, CHY1517 expresses the a pheromone and MFα pheromone receptor, but the α homeodomain transcription factor. We designate YPH716 as “α → alf1” and CHY1517 as “α → alf2” to indicate the genetic modifications of their mating types, in which “alf1” and “alf2” designate the strains as “a-like faker” type 1 and type 2, respectively.

We also analyzed mitochondrial inheritance in unilateral and bilateral crosses of strains in which a single gene had been deleted (see Table S2 for the list of genes). For these crosses, the parental strains were constructed in the H99 (MATα) and KN99a (MATa) backgrounds, and thus their mitochondrial types could be differentiated with PCR markers targeting the presence/absence of introns in the COX1 gene, as previously described (Toffaletti et al. 2004).

Mating, spore dissection, and mating-product screening

Mating and basidiospore dissection for C. amylolentus were carried out as previously described (Findley et al. 2012). Briefly, mating-compatible strains were mixed and spotted on V8 (pH = 5) medium. The mating plates were incubated in the dark at room temperature (agar side up with no parafilm) for 1–2 weeks until abundant hyphae, basidia, and basidiospore chains were visible under the microscope. The basidiospore chains were then transferred onto fresh YPD medium, and individual basidiospores were separated using a fiber optic needle spore dissecting system, as previously described (Idnurm 2010). Individual spores separated from the same spore chain were the products of one meiotic event (Findley et al. 2012).

For C. neoformans, the mating/fusion products from the cross between strains CHY1517 and YPH716 were obtained by first spotting the mixture of the two parental strains onto V8 (pH = 5) medium, and then after hyphae, basidia, and basidiospore chains were formed, the hyphal sectors at the edge areas of the mating spots were excised and suspended in 1 × PBS. The cell suspension was diluted, and spread onto synthetic dextrose (SD)-uracil plates to screen for Ura⁺ isolates. The Ura⁺ isolates were then transferred onto YPD + nourseothricin (NAT) plates to further screen for isolates that were also NAT-resistant, and thus represented recombination/fusion of the markers present in the two parental strains.

Mating/fusion products from crosses between strains JEC20 ura5 and M001 (H99 ade2), or JEC20 ura5 and M049 (H99 ade2), were recovered similarly to those from the cross between CHY1517 and YPH716. Here, after the hyphal sectors were cut out and suspended in 1 × PBS, the suspension was diluted and spread onto SD-uracil-adenine medium to screen for prototrophic isolates, thus representing either recombinants or fusion products of the two parental strains.

To test the effect of a specific gene on mito-UPI, the unilateral and bilateral crosses were set up on MS media, incubated at room temperature in the dark for 10 days, and random spores were then dissected and analyzed as previously described (Sun et al. 2019).

Genomic DNA, genetic markers, and genotyping

Germinated individual spores were transferred and patch-streaked onto fresh YPD medium, and genomic DNA was extracted from the biomass as described in a previous study (Sun et al. 2012). For C. amylolentus, mitochondrial genotyping was based on two PCR-RFLP markers targeting the NAD4 and NAD5 genes, respectively, while the mating types were determined using PCR-RFLP markers for ETF1 and SXI1 that are located within the A and B MAT locus, respectively (Figure S1 and Table S1). For C. neoformans, mitochondrial genotyping was based on the NAD2 and NAD5 genes, while serotype- and mating type-specific markers for the SXI1α, SXI2a, and STE20α/a genes were assayed to genotype the MAT locus (Figure S1 and Table S1). All of the markers are codominant (i.e., they can differentiate the two homozygous states, as well as the heterozygous one; Figure S1). All PCR reactions were carried out using Promega (Madison, WI) 2X Go Taq Master Mix according to the manufacturer’s instructions, and with the following PCR thermal cycles: first an initial denaturation at 94° for 6 min; then 36 cycles of 45 sec at 94°, 45 sec at 60°, and 90 sec at 72°; and a final extension at 72° for 7 min. All enzyme digestions were performed with enzymes purchased from New England Biolabs (Beverly, MA) and following the manufacturer’s instructions.

Microscopy imaging

For fluorescence imaging of yeast cells, conjugation tubes, and initial hyphal formation, the mating mixture was grown on MS solid medium. The cells were incubated for 12 hr for yeast cells and conjugation tubes, and 24 hr for initial hyphal formation. The cells were collected, washed, and the cellular structures were observed and imaged with a Zeiss ([Carl Zeiss], Thornwood, NY) Imager widefield fluorescence microscope.

Statistical analyses

Statistical tests of association of mito-UPI in C. amylolentus with specific mating-type loci/alleles were carried out with a binomial probability test (www.vassarstats.net). A P-value of < 0.05 was considered statistically significant, and was required to reject the null hypothesis that mitochondrial inheritance is not associated with a mating-type locus/allele, i.e., that the two mitochondrial types are inherited at equal frequency (50%) among all independent crosses, regardless of the mating types and mitochondrial types of the two parental strains. For mitochondrial inheritance in C. neoformans, Pearson’s χ² test was applied to test whether the observed distribution among the progeny is significantly different from the null hypothesis of an equal chance of inheritance of the two mitochondrial types. P-values of < 0.05 were considered statistically significant and used to reject the null hypothesis.

Data availability

Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and tables. Supplemental material available at figshare: https://doi.org/10.25386/genetics.11422590.

Results

Mito-UPI in C. amylolentus is controlled by the P/R MAT locus

In a previous study, we demonstrated that C. amylolentus has a tetrapolar mating system with the two mating-type loci, the P/R locus (A locus) and the HD locus (B locus), located on different chromosomes (Findley et al. 2012). Additionally, we found that among all of the basidiospores generated by mating between the two natural C. amylolentus isolates, CBS6039 (A1B1, mito-a) and CBS6273 (A2B2, mito-b), mitochondria were always inherited from CBS6273, the A2B2 parent (Findley et al. 2012). In the studies presented here, basidiospores produced by a CBS6039 × CBS6273 cross were dissected, and progeny with all four different mating types that have the mito-b mitochondrial type were recovered. Interestingly, during our efforts to dissect additional basidiospores from the same cross, we found that out of the 40 basidiospore chains dissected and analyzed, basidiospores from two different spore chains all inherited mitochondria from CBS6039, the A1B1 parent with mitochondrial type mito-a (basidia numbers 8 and 9 in Table 2). Thus, mitochondrial inheritance in C. amylolentus is indeed uniparental from the A2B2 parent (CBS6273), but with a low level of leakage (∼5%) from the other parent (CBS6039; A1B1). This low level of leakage allowed us to recover meiotic progeny of all four different mating types that possess the mito-a mitochondrial type.

Table 2. Summary of crosses and mitochondrial inheritance in C. amylolentus.

Cross typea	Crossa	Basidium	Number of spores dissected	Number of spores germinated	Germination rate (%)
A1B1 (mito-a) × A2B2 (mito-b)	CBS6039 × CBS6273b	1	16	6	38
		2	15	8	53
		3	13	6	46
		4	25	13	52
		5	17	8	47
		6	12	5	42
		7	18	16	89
		8c	30	21	70
		9c	26	18	69
	CBS6039 × SSA089	1	19	10	53
		2	14	10	71
	SSA033 × CBS6273	1	17	14	82
		2	15	12	80
	SSA033 × SSA089	1	13	9	69
		2	16	13	81
A1B2 (mito-a) × A2B1 (mito-b)	SSA026 × SSA052	1	20	5	25
	SSA026 × SSA058	1	12	9	75
		2	25	14	56
		3	29	21	72
	SSA028 × SSA052	1	12	7	58
	SSA028 × SSA058	1	17	13	76
		2	15	13	87
		3	13	10	77
A2B1 (mito-a) × A1B2 (mito-b)	SSB817 × SSA770	1	6	4	67
		2	8	2	25
		3	8	3	38
		4	14	9	64
	SSB817 × SSA790	1	29	17	59
	SSB821 × SSA797	1	11	5	45
A2B2 (mito-a) × A1B1 (mito-b)	SSA032 × SSA067	1	12	1	8
		2	9	3	33
	SSA032 × SSA108	1	7	4	57
		2	11	10	91
		3	12	4	33

^aThe parental type highlighted in bold font is the parental type from which the meiotic progeny inherited the mitochondria.

^bOnly nine of the >20 basidia that were dissected for crosses between CBS6039 and CBS6273 are shown here.

^cBasidia numbers 8 and 9 inherited mito-a from the A1B1 parental strain, CBS6039.

We next conducted pairwise matings between isolates with compatible mating types but different mitochondrial types, and dissected multiple basidiospore chains from most of these crosses (Table 2). In all of these crosses, basidiospores dissected from the same cross always inherited the same mitochondria, further supporting previous observations that mitochondrial inheritance in C. amylolentus is predominantly UPI. Additionally, the mitochondria in all of the basidiospores were inherited from the parental strain that possessed the A2 MAT allele, independent of the alleles present at the B locus (Table 2). The observed association between the A2 allele and mitochondrial inheritance is statistically significant (binomial probability test, P < 0.05), and thus provides evidence that mito-UPI in C. amylolentus is controlled by the A2 mating-type locus.

What gene(s) in the A2 locus is responsible for the mito-UPI in C. amylolentus then? As mentioned earlier, the A mating-type locus in C. amylolentus has undergone extensive expansion compared to typical pheromone/pheromone receptor mating-type loci in basidiomycetes and contains ∼20 genes (Findley et al. 2012; Sun et al. 2017). It is thus difficult to narrow down and define the genes that could be playing key roles in mito-UPI. However, we hypothesized that the pheromones and pheromone receptors might be involved in this process for the following reasons. First, pheromone and pheromone receptors are defining genes, and in some cases the only genes, of the P/R locus in basidiomycetes. If mito-UPI by the P/R locus is conserved in basidiomycetes, it is logical to hypothesize that the pheromones and pheromone receptors might be involved in this process. Also, pheromones and pheromone receptors are involved in the early stages of the mating process before zygote formation, including mating partner recognition and mating initiation. This coincides with the time when mitochondria from the two mating partners are thought to be differentially tagged for later protection or degradation in the zygotes. Additionally, in some basidiomycete species, the expression of pheromones and pheromone receptors is not symmetric between the mating partners (McClelland et al. 2004), which could provide opportunities for asymmetric sexual development between the opposite mating types (please see below), as well as for the mitochondria of different mating types to be differentially tagged for protection/degradation, which collectively contribute to mito-UPI.

Pheromone and pheromone receptors affect mito-UPI in C. neoformans

Because approaches for genetic manipulation of C. amylolentus are rudimentary, we decided to further investigate the role of pheromones and pheromone receptors in mito-UPI in its sister species, C. neoformans, based on the following considerations. First, C. neoformans is one of the most closely related known species to C. amylolentus. Additionally, it has been previously shown that mitochondria are uniparentally inherited in C. neoformans (Xu et al. 2000; Yan and Xu 2003), and several genes involved in mating have been shown to play important roles in ensuring proper mito-UPI during mating (Yan et al. 2004, 2007a; Gyawali and Lin 2013). Given their close relationship, the mechanisms of mito-UPI are likely to be conserved between these two species. Second, although C. neoformans has a bipolar mating system governed by one large MAT locus, it has been shown that this large contiguous MAT locus is the result of a fusion event between the ancestral A and B loci, at which point the A locus had already undergone a similar expansion as in C. amylolentus, with most of the genes that have been identified within the A locus in C. amylolentus also present in the MAT locus of C. neoformans (Hsueh et al. 2011; Findley et al. 2012; Sun et al. 2017).

To investigate the possible effects of pheromone and pheromone receptor genes on mito-UPI, two engineered MATα strains were analyzed. In one strain, YPH716 (serotype A, “α → a-like faker 1 (alf1)”; Figure 1 and Table 1), the original SXI1α gene in the MATα locus has been deleted, and the SXI2a from the MATa locus has been integrated at the URA5 locus. Thus, YPH716 lacks the homeodomain transcription factor Sxi1α, and expresses instead the a-specific homeodomain transcription factor Sxi2a, along with all other α-specific genes (Figure 1). In the second strain, CHY1517 (serotype D, “α → alf2”; Figure 1 and Table 1), the MATα genes encoding the pheromones (MFα1, MFα2, and MFα3) and pheromone receptor (STE3) have been deleted, and replaced with the alleles from the MATa locus. As the result, CHY1517 expresses the Ste3a pheromone receptor and MFa instead of Ste3α and MFα, and has all other α-specific genes (Figure 1). Thus, if the null hypothesis that pheromones and pheromone receptors are not involved in mito-UPI is correct, we should expect during the mating between CHY1517 and YPH716 that the two types of mitochondria (mito-A and mito-D; Table 1 and Figure S1) would be inherited at equal frequencies among meiotic progeny or fusion products of the two strains, as shown in a previous study (Yan et al. 2007a), because functional SXI1 and SXI2 genes are still present in the zygote. Alternatively, if pheromones and pheromone receptors are involved in mito-UPI, we would expect deviations of mitochondrial inheritance from biparental during the cross between CHY1517 and YPH716.

We recovered a total of 123 isolates from the cross between CHY1517 and YPH716 that have a combination of the phenotypic markers of the two parental strains (Table 3). Genotyping of these isolates using the four serotype- and mating type-specific markers (SXI1α, SXI2a, and STE20α/a; Table 3) revealed that they all possess all four alleles, indicating that they are either serotype AD diploid fusion products of the two parental strains, or meiotic products that are diploid or disomic for the MAT locus. Additionally, of these 123 mating/fusion products, 80 (65%) inherited mitochondria from CHY1517 (mito-D), 37 (30%) inherited mitochondria from YPH716 (mito-A), and 6 (5%) appeared to have recombinant mitochondria (Table 3). The distribution of mito-A and mito-D among these mating products was significantly different from the null hypothesis that both mitochondrial types have equal chances (50:50) of inheritance (χ² test, P < 0.0001). Importantly, the biased mito-inheritance was in favor of those from the strain CHY1517, which has the MFa and STE3a alleles, although the other mating partner, YPH716, possesses the SXI2a gene. Thus, our results suggest that the pheromones and pheromone receptors are indeed involved in controlling mito-UPI in C. neoformans. Specifically, replacing the MATα versions of the pheromones and pheromone receptor with those from MATa in an otherwise MATα strain significantly increased the chances of mitochondria being inherited during sexual reproduction.

Table 3. Summary of genotyping results for C. neoformans mating/fusion products.

	NAD2 (mito)a	NAD5 (mito)a	STE20- Aαb	STE20- Dαb	SXI2- Aab	SXI1- Dαb	Number of fusion products (%)
CHY1517 × YPH716
CHY1517	D	D	−	+	−	+	/
YPH716	A	A	+	−	+	−	/
Heterozygote	H	H	+	+	+	+	/
Progeny genotype 1	D	D	+	+	+	+	80 (65)
Progeny genotype 2	A	A	+	+	+	+	37 (30)
Progeny genotype 3	D	A	+	+	+	+	6 (5)
JEC20 × M001
JOHE50	D	D					/
M001	A	A					/
Heterozygote	H	H					/
Progeny genotype 1	D	D					45 (92)
Progeny genotype 2	D	A					4 (8)
JEC20 × M049
JOHE50	D	D					/
M049	A	A					/
Heterozygote	H	H					/
Progeny genotype 1	D	D					41 (82)
Progeny genotype 2	A	A					3 (6)
Progeny genotype 3	D	A					4 (8)
Progeny genotype 4	D	H					2 (4)

^aA indicates serotype A mitochondrial type, D indicates serotype D mitochondrial type, and H indicates heterozygous mitochondrial type.

^b“+” indicates the presence of the corresponding PCR product and “−” indicates the absence of the corresponding PCR product.

However, it should be noted that mito-UPI is not fully restored to a wild-type level in the cross between CHY1517 and YPH716. For comparison, we performed two interserotype opposite-sex crosses using strains with intact mating type loci [JEC20 ura5 × M001 (H99 ade2) and JEC20 ura5 × M049 (H99 ade2); Table 3]. In cross JEC20 ura5 × M001, we found that 45 out of 49 (92%) mating products inherited mitochondria from JEC20 ura5, the MATa parent, while the other four mating products (8%) inherited recombinant mitochondria. In cross JEC20 ura5 × M049, we found that 41 out of 50 (82%) mating products inherited mitochondria from JEC20 ura5, 3 mating products (6%) inherited mitochondria from M049, and the other 6 mating products (12%) inherited recombinant mitochondria. In both cases, mito-UPI was occurring at a significantly higher frequency compared to the cross between CHY1517 and YPH716 (Fisher’s exact test, P < 0.05). Taken together, these results suggest that pheromones and pheromone receptors contribute significantly to mito-UPI, but additional genes, which might also be located within the MATa locus, are also needed for wild-type levels of mito-UPI.

Conjugation tube formation and hyphae initiation are asymmetric in C. neoformans bisexual reproduction

In basidiomycetes, the interactions between pheromones and pheromone receptors are critical for the early stages of sexual reproduction, including mating partner sensing, conjugation formation, and zygote formation (Bandoni 1963; Spellig et al. 1994; Bölker 2001). To investigate the dynamics of sexual development and mitochondrial movement, we generated two MATa and MATα strains, SSG269 and SSH116, respectively, in which the mitochondrial protein Hem15 was tagged with GFP in both strains. Additionally, for the MATα strain (SSH116), the calcineurin A subunit Cna1 was tagged with mCherry. We set up crosses between these two strains and observed sexual development after 12 hr of mating on MS medium.

First, we found that the mitochondrial morphology was tubular when cells were grown on MS medium for 12 hr (Figure 2, A and B; >98%, n = 50). This is consistent with previous studies showing tubulization of C. neoformans mitochondria under stress conditions (Ma and May 2010; Chang and Doering 2018). Additionally, we observed that in cases where two cells were connected by a conjugation tube and the mCherry signal was not distributed universally throughout the “dumbbell”-shaped structures, the conjugation tubes always had the mCherry signal (Figure 2, C; >98%, n = 50). This suggests that during mating initiation, the conjugation tubes were always generated by the MATα strain (SSH116) that is marked by the mCherry signal. There also appeared to be separation between the mitochondria from the two parental cells and those in the conjugation tube (Figure 2, C; 80%, n = 15), suggesting that it is possible that during the initial stages of sexual development, the asymmetric development of the conjugation tubes from the MATα cells could generate a physical barrier that limits transmission of α mitochondria from the MATα parent to the a-cell side of the zygote.

Figure 2.

Microscope imaging of conjugation tube formation and hyphae initiation during C. neoformans sexual reproduction. (A–D) From left to right, DIC, GFP, mCherry, and merged images are shown. (A) Images of the parental strain SSH116 after incubation for 12 hr on MS solid medium. (B) Images of the parental strain SSG269 after incubation for 12 hr on MS solid medium. (C) Images of conjugation tube formation between strains SSH116 and SSG269 after 12 hr of coculturing on MS solid medium, in which the conjugation tubes are formed by the MATα strain that has the mCherry signal. (D) Images of hyphae initiation in crosses between strains SSH116 and SSG269 after 24 hr of coculturing on MS solid medium, in which the hyphae are formed from the MATa cells and on the opposite side of the conjugation tube. White arrowheads highlight the gaps between mitochondrial complexes from the zygote and the conjugation tubes. Bars, 10 μm.

We also observed that after zygote formation (24 hr of mating on MS solid medium), the hyphae were almost always initiated from the zygote at a position opposite from the conjugation site, and that the separation between the mitochondria from the zygote and the conjugation tube persisted (Figure 2, D; 90%, n = 10). This suggests that the mitochondria in the hyphae, and subsequently in the meiotic progeny, were mostly from the a-cell side of the zygote, whose mitochondria were contributed mostly by the MATa parent.

Deletion of the CRG1 gene disrupts mito-UPI in bilateral crosses

It has been shown that the CRG1 gene encodes an RGS that negatively regulates the pheromone-signaling cascade and cAMP pathways during sexual reproduction in C. neoformans, with crg1Δ deletion strains displaying enhanced hyphal development in response to mating pheromones (Fraser et al. 2003; Nielsen et al. 2003; Wang et al. 2004; Feretzaki and Heitman 2013). We hypothesized that deleting the CRG1 gene might also disrupt the dynamics of pheromone and pheromone receptor interaction during early stages of mating when conjugation occurs, and, consequently, compromise the fidelity of mito-UPI.

In confrontation assays involving the crg1Δ deletion strains, we found that both the MATα and MATa deletion strains showed elevated pheromone response and filamentation compared to their respective wild-type strains (Figure 3A). However, this enhanced pheromone response was not symmetric between the two mating types. Specifically, while the MATα crg1Δ strain produced hyphae when confronted with the wild-type MATa as well as the MATa crg1Δ strains (Figure 3A, II and IV), the MATa crg1Δ strain only displayed enhanced pheromone response and hyphal growth when confronted with the MATα crg1Δ strain, but not the MATα wild-type strain (Figure 3A, III and IV). Thus, it appears that the α mating type has an intrinsic higher sensitivity to pheromone compared to the a mating type.

Figure 3.

Effects of the CRG1 gene on mating and mito-UPI. (A) Wild-type (I) as well as unilateral (II and III) and bilateral (IV) confrontation assays of crg1Δ deletion strains. Bars, 20 μm. (B) Genotyping of the mating (STE20) and the mitochondrial types (COX1) of 40 random spores dissected from crg1Δ unilateral (I and II) and bilateral (III) crosses. The progeny highlighted with red stars are those that inherited mitochondria from the MATα parent. (C) Summary of spore dissection, germination, and mito-UPI frequencies from crg1Δ unilateral and bilateral crosses.

We next dissected random basidiospores from the crg1Δ unilateral and bilateral crosses, and analyzed their mating types and mitochondrial types (Figure 3B). We found that while both unilateral crosses showed complete mito-UPI from the MATa parent in their progeny, progeny from the crg1Δ × crg1Δ bilateral cross showed a significantly higher level of mitochondrial leakage, with 8 of the 40 (20%) random basidiospores analyzed inheriting mitochondria from the MATα parent (Figure 3C).

Discussion

Our data suggest that in addition to molecular mechanisms that actively degrade the mitochondria from the MATα parent, the dynamics of early stages of sexual development controlled by the pheromone and pheromone receptors may also provide an additional physical barrier for the α mitochondria to enter the zygote. Thus, consistent with previous hypotheses [e.g., in Wilson and Xu (2012)], multiple mechanisms, both physically and genetically, could act in concert to ensure faithful mito-UPI from the MATa cell during C. neoformans sexual reproduction (Figure 4).

Figure 4.

Model of mito-UPI regulation in Cryptococcus species. The mito-UPI process starts when compatible mating partners sense pheromones from each other [(A) and (B)]. The asymmetric nature of this process could lead to activation of transcription factors (e.g., MAT2) and target genes whose products differentially tag the mitochondria in the two mating partners so that they would be targeted for or protected from degradation by mitophagy, or other processes yet to be defined. (C) Successful pheromone and pheromone receptor interaction induces the formation of a conjugation tube from the MATα parent, through which the MATα nucleus migrates into the MATa parent to form the zygote. It is possible that the mitochondria remaining in the MATα parent start degradation at this point, a process that could involve mitophagy. (D) The zygote contains the two nuclei, as well as the mitochondria from the MATa parent. It is possible that trace amounts of mitochondria from the MATα parent also migrate into the zygote. The mitochondria remaining in the conjugation tube and the MATα parent cell continue degradation. (E) Hyphae are initiated from the zygote on the opposite side of the conjugation tube. This further increases the chances of mitochondria from the MATa parent being included in the hyphae, and consequently, the eventual mating progeny inherit mitochondria mostly from the MATa parent. Additionally, the trace amount of mitochondria from the MATα parent would be degraded through mechanisms such as mitophagy. mito-UPI, mitochondrial uniparental inheritance.

The mating process starts when compatible mating partners sense pheromones from each other. The asymmetric nature of this process could lead to divergent activation of transcription factors (e.g., Mat2) and pathways (e.g., mitophagy), resulting in asymmetric expression of genes critical for the initiation of sexual development (e.g., STE3 and MFs) and differential tagging of the mitochondria in the two mating partners. Successful pheromone and pheromone receptor interaction induces the formation of conjugation tubes from the MATα parent, through which the MATα nucleus migrates into the MATa cell to form the zygote, while the vast majority of the α mitochondria remain in the MATα cell and the conjugation tube, with only a small fraction migrating into the zygote accidentally. It is possible that the mitochondria remaining in the MATα cell start degradation at this point, a process that could involve mitophagy, which could also direct the active degradation of mitochondria from the MATα parent in the zygote. Subsequently, hyphae are initiated from the zygote on the opposite side of the conjugation tube, which could further limit the opportunities for mitochondria from the MATα parent being included in the hyphae, even in cases where small amounts remain in the zygote at this stage. Consequently, the eventual mating progeny inherit mitochondria mostly from the MATa parent.

Our proposed model is consistent with current understanding of mito-UPI in C. neoformans. Studies of zygotes formed at the beginning of the mating process suggest that mito-UPI in Cryptococcus is likely established at an early stage of sexual development (Sun and Xu 2007; Gyawali and Lin 2013). In our studies of C. amylolentus, all of the basidiospores from the same basidium have an identical mitochondrial type, and this is the case for all of the basidia, even the two anomalous basidia (numbers 8 and 9) from the CBS6039 × CBS6273 crosses that exhibited mito-UPI exclusively from the A1 parent. In most genetic studies of fungal UPI in basidiomycetes, this has involved random spore dissection rather than dissection from individual basidia. For example, the low level of UPI observed in C. neoformans from the α parent (generally <5%) has been termed leakage and has been thought to result from a low level of α mitochondria that survive in the hyphae. In this view, one might have expected to find that ∼5% of spores dissected from an individual basidium would have the mitochondrial genome from the less-favored parent. Instead, our observation suggests that mito-UPI may have been accomplished very early following zygote formation, such that ∼95% of the time the entire hyphal compartment has mitochondria derived from the A2 or the a parent. The remaining ∼5% of the time the hyphal compartment would have the mitochondria derived from the A1 or the α parent. If the cost of inheriting more than one mitochondrial genotype is exerted on cell types beyond the zygote, such as in the hyphae, this may be a reason for the control of mito-UPI being exerted early during sexual reproduction.

It has been shown that pheromones are asymmetrically expressed in the two mating types in C. neoformans under mating conditions, possibly through the regulation of Mat2, a key regulator of cell–cell fusion during zygote formation (Shen et al. 2002; McClelland et al. 2004; Kent et al. 2008; Lin et al. 2010; Gyawali and Lin 2013). Specifically, the expression of mating pheromones from the MATa cells are increased earlier and to a greater extent than those from the MATα cells (McClelland et al. 2004), suggesting that pheromone sensing by the pheromone receptors, as well as the initiation of the downstream genes (e.g., mitophagy-related mito-tagging), might also be asymmetric between the two mating partners. We hypothesize that the observed asymmetric gene expression underlies the observed unidirectional conjugation tube formation, as well as the polarized migration of the MATα nucleus through the conjugation tube into the zygote, and thus these patterns of gene expression could be integral to mito-UPI. This is also consistent with our findings with the crg1Δ strains, where the deletion in the MATα background exhibited elevated pheromone sensing and filamentation in a unilateral assay, while the MATa crg1Δ strain only showed an enhanced pheromone response when confronted with the MATα crg1Δ strain. This could be the reason why we failed to observe elevated mitochondrial leakage in the crg1Δ unilateral crosses, as the pheromone production and sensing dynamics in these cases were largely in accord with those present in the wild-type crosses. On the other hand, in crg1Δ bilateral crosses, because the MATa crg1Δ strain also showed elevated pheromone response and filamentation, the dynamics of pheromone production and response were thus disrupted and, possibly as a consequence, increased levels of mitochondrial leakage in the basidiospores were observed. It should be noted that deletion of the CRG1 gene appeared to have varied effects on mito-UPI in crosses between different species in the C. gattii species complex (Wang et al. 2015), suggesting that the regulation of mito-UPI involves multiple genes and that the process could be compromised during interspecies mating. This is also consistent with our analyses of the crosses between strains CHY1517 and YPH716. While CHY1517 possesses the STE3a and MFa genes, its mitochondria are not inherited at a level of 100% in the fusion products, suggesting that although the pheromone and pheromone receptor genes play important roles in controlling mito-UPI, other genes are required for faithful mito-UPI during sexual reproduction.

Functional homeodomain transcription factors SXI1α and SXI2a from the MATα and MATa parents, respectively, are both required for faithful mito-UPI in C. neoformans (Yan et al. 2004, 2007a), and it has been shown that deletion of the SXI1α gene enhances the spread of mobile mitochondrial introns in C. neoformans (Yan et al. 2018). The SXI1α/SXI2a heterodimer serves as a key regulator of sexual development after zygote formation, including hyphal growth. Thus, it may play a critical role in determining the location of hyphal initiation to ensure it occurs away from the conjugation site, where occasional inclusion of MATα mitochondria to the zygote might occur. For C. amylolentus, we showed that mito-UPI is controlled by the P/R locus. However, it should be noted that in this case, the HD locus might still play a similar role as in C. neoformans. That is, functional HD heterodimers may be required to ensure the completion of mito-UPI. Because the strains that we used in C. amylolentus crosses are all derived from wild-type strains and possess functional HD genes, it would be interesting to test C. amylolentus crosses between isolates with mutated HD genes to see if faithful mito-UPI is still occurring.

Studies have shown that mitophagy is involved in mito-UPI in a variety of species (Al Rawi et al. 2011; Levine and Elazar 2011; Sato and Sato 2011; Luo et al. 2013), and it is hypothesized that differential tagging of mitochondria from the MATa and MATα parents, and subsequent selective degradation of one group of mitochondria by mitophagy, is also playing an important a role in mito-UPI in C. neoformans (Gyawali and Lin 2013). To investigate this, we deleted several known genes in the S. cerevisiae mitophagy pathway, as well as several other genes that are located within the C. neoformans MAT locus, or that have been shown to be involved in mitochondrial movement and segregation, and studied their effects on mito-UPI in both uni- and bilateral crosses (Table S2). We did not find that deletion of any of these genes resulted in deviation from mito-UPI, suggesting that these genes are not likely involved in mito-UPI. However, it is possible that other genes in these pathways that have not been tested are playing the key roles in mito-UPI, or that these genes are redundant with others not as yet tested. It is also possible that the key genes involved in mito-UPI are pleiotropic and essential for cell survival, which would prevent them from being identified through gene deletion approaches.

Certain environmental factors, such as UV and temperature, influence mito-UPI in C. neoformans (Yan et al. 2007b), and natural strains of C. gattii with recombinant mitochondrial genomes have been isolated from the environment or following genetic crosses (Voelz et al. 2013), suggestion that mitochondrial leakage occurs during sexual reproduction in nature. It is possible that the expression of key genes involved in mito-UPI, such as the pheromone and pheromone receptor genes, as well as transcription factors including Mat2, Sxi1, and Sxi2, could be compromised by factors present in the natural environment or become incompatible between mating partners due to sequence divergence accumulated between different lineages during evolution. Such environmental perturbance or gene function incompatibility could compromise both nuclear and mitochondrial dynamics during the early stages of sexual development, as well as the regulation of the mitophagy pathway that is involved in mito-UPI, which collectively could lead to increased mitochondrial leakage from the MATα parent.

Based on the DNA sequence, previous studies have shown that the C. amylolentus A1 pheromone receptor is more similar to the STE3a allele in C. neoformans, and that the A2 pheromone receptor is more similar to the STE3α allele. It is then intriguing that in C. amylolentus the mito-UPI is from the A2 parent, while in C. neoformans it is from the STE3a parent. It should be noted that both the A1 and A2 STE3 alleles show significant divergence from the STE3a and STE3α alleles, respectively. Additionally, the MAT configurations in C. amylolentus and C. neoformans are highly different, with the former having a tetrapolar mating system with unlinked P/R and HD loci located on different chromosomes, and the latter having a bipolar mating system with a fused MAT locus that contains both P/R and HD genes, demonstrating that the mating system has undergone significant transitions between the two species. It is possible that at some intermediate stages of this transition the mito-UPI was compromised. Because of the detrimental effects of mito-heteroplasy, selection pressure would have favored the reestablishment of mito-UPI. The linkage of this phenomenon to the mating type is the result of natural selection as this is the best way to ensure mito-UPI. During this reestablishment of mito-UPI, it is possible that the alleles associated with the mating type whose mitochondria are preferentially inherited in the progeny could change. Thus, while mito-UPI is the rule, its regulation could be dynamic, maybe particularly during speciation events.