The Sclerotinia sclerotiorum Mating Type Locus (MAT) Contains a 3.6-kb Region That Is Inverted in Every Meiotic Generation

Sclerotinia sclerotiorum is a fungal plant pathogen and the causal agent of lettuce drop, an economically important disease of California lettuce. The structure of the S. sclerotiorum mating type locus MAT has previously been reported and consists of two idiomorphs that are fused end-to-end as in other homothallics. We investigated the diversity of S. sclerotiorum MAT using a total of 283 isolates from multiple hosts and locations, and identified a novel MAT allele that differed by a 3.6-kb inversion and was designated Inv+, as opposed to the previously known S. sclerotiorum MAT that lacked the inversion and was Inv-. The inversion affected three of the four MAT genes: MAT1-2-1 and MAT1-2-4 were inverted and MAT1-1-1 was truncated at the 3’-end. Expression of MAT genes differed between Inv+ and Inv- isolates. In Inv+ isolates, only one of the three MAT1-2-1 transcript variants of Inv- isolates was detected, and the alpha1 domain of Inv+ MAT1-1-1 transcripts was truncated. Both Inv- and Inv+ isolates were self-fertile, and the inversion segregated in a 1∶1 ratio regardless of whether the parent was Inv- or Inv+. This suggested the involvement of a highly regulated process in maintaining equal proportions of Inv- and Inv+, likely associated with the sexual state. The MAT inversion region, defined as the 3.6-kb MAT inversion in Inv+ isolates and the homologous region of Inv- isolates, was flanked by a 250-bp inverted repeat on either side. The 250-bp inverted repeat was a partial MAT1-1-1 that through mediation of loop formation and crossing over, may be involved in the inversion process. Inv+ isolates were widespread, and in California and Nebraska constituted half of the isolates examined. We speculate that a similar inversion region may be involved in mating type switching in the filamentous ascomycetes Chromocrea spinulosa, Sclerotinia trifoliorum and in certain Ceratocystis species.


Introduction
Sclerotinia sclerotiorum is a filamentous ascomycete in the Sclerotiniaceae (Pezizomycotina) and a necrotrophic pathogen of more than 400 hosts worldwide, including many important agricultural crops [1,2]. In California, the biggest lettuce producer in the United States, S. sclerotiorum is a causal agent of lettuce drop [3] that reduces overall annual lettuce yield by 15%, and losses in individual fields commonly amount to up to 60% [4].
Sclerotinia sclerotiorum is very durable and survives in the soil in absence of a host for one or more years [5,6]. The survival structures are known as sclerotia that are clusters of cells surrounded by a melanized protective layer. New infections are initiated when sclerotia germinate to form hyphae, or when ascospores are released from fruiting bodies, known as apothecia that emerge from sclerotia [7]. Ascospores result from selfing [8] or outcrossing by heterokaryon formation and recombination [9]. Microconidia are formed and may play a role in fertilization [10].
In ascomycetes, mating system is determined by the mating type locus MAT that encodes transcription factors that regulate downstream gene expression [11]. There are two major mating systems which are heterothallism that is equivalent to obligate outcrossing, and homothallism that consists of selfing and outcrossing [12]. In heterothallic species, isolates generally carry one of two versions of MAT, and only isolates that differ at MAT are sexually compatible. In homothallic species, there is only one version of MAT that generally consists of all the MAT genes of heterothallic relatives, but assembled at a single locus [13]. The two versions of MAT in heterothallics are referred to as idiomorphs instead of alleles, because overall DNA sequence homology between idiomorphs is absent [14,15].
The role of MAT as master regulator of downstream gene expression has been well studied in some ascomycetes including S. cerevisiae [28], but little is known about MAT regulated genes in filamentous ascomycetes. MAT target genes in filamentous ascomycetes include pheromone and pheromone receptors that play important roles in recognition of mating partners [29,30,31], and possibly in nuclear pairing following fertilization [32,33,34,35], and genes that are involved in heterokaryon incompatibility [36]. But the majority of genes under MAT regulation appear to have functions that are not directly related to mating [31,37,38,39]. These pleiotropic capabilities of MAT may explain the morphological changes that correlate with mating system in species where mating type switching has been documented [21,23,40].
In this study, we examined the mating type loci of 283 S. sclerotiorum isolates from lettuce in California and other states and hosts, and found that MAT contained a 3.6-kb region that is inverted between generations which correlates with changes in MAT gene expression. We show that the fused S. sclerotiorum MAT arrangement most likely evolved from heterothallic ancestors, and speculate whether a MAT inversion similar to the one in S. sclerotiorum may be responsible for mating type switching in several other filamentous ascomycetes.

Sclerotinia sclerotiorum MAT locus
DNA sequencing coverage of three complete MAT regions was generated for S. sclerotiorum strains 44Ba1, 44Ba12 and 44Ba18 that are listed in Table S1. The DNA sequences measured 12,145 bp, 12,218 bp and 12,218 bp, respectively, and were submitted to GenBank (Accessions JQ815883, JQ815884, and JQ815885). The previously sequenced S. sclerotiorum strain 1980 MAT region from the recently completed genome sequencing project [17], was more than 98% similar overall across coding and non-coding regions to homologous regions of S. sclerotiorum strains 44Ba1, 44Ba12 and 44Ba18 ( Table 1). The S. sclerotiorum strain 44Ba1 MAT region contained the same ORFs as S. sclerotiorum strain 1980 [17], notably MAT1-1-5, MAT1-1-1, MAT1-2-4 and MAT1-2-1, flanked by APN2 and SLA2, whereas S. sclerotiorum strains 44Ba12 and 44Ba18 differed by a 3.6-kb inversion that extended from MAT1-1-1 to MAT1-2-1, and resulted in the inversion of MAT1-2-1 and MAT1-2-4, and the truncation of MAT1-1-1 at the 3'-end ( Figure 1; Alignment S1). Following the genetic nomenclature of plant pathogenic fungi on the designation of phenotypes [41], isolates with the MAT inversion were designated Inv+, and isolates without the inversion Inv-. The 3.6-kb MAT inversion in Inv+ isolates, and the homologous region in Inv-isolates, were referred to as the MAT inversion region.
There was little variation among the homologous MAT regions of the S. sclerotiorum isolates. Sclerotinia sclerotiorum strains 44Ba12 and 44Ba18 had identical sequences, and among the four MAT ORFs, MAT1-2-4 was most variable, and MAT1-2-1 was most conserved (Table S2). Phylogenetic analyses showed that irrespective of MAT gene, S. sclerotiorum strains 1980, 44Ba1, 44Ba12 and 44Ba18 were monophyletic ( Figure 2). Due to the low number of parsimony informative characters (Table S3), branch supports were not evaluated. More details on the phylogenetic analyses are given in Table S3.
Comparisons between the mating type genes of S. sclerotiorum generated in this study to the closely related Botrytis cinerea [17] revealed 52 to 74% nucleotide identity depending on the MAT gene (Table 1). Among the four mating type genes, MAT1-1-5 was the most divergent with the lowest identity (52%), the remaining three MAT genes were more than 71% identical to B. cinerea homologs. Except for MAT1-2-4 where B. cinerea had an extra intron, intron numbers were conserved between S. sclerotiorum and B. cinerea (Table 1).
Comparison between the flanks of the MAT inversion region in the S. sclerotiorum Inv+ strains 44Ba12 and 44Ba18, to the homologous regions of the S. sclerotiorum Inv-strain 44Ba1, revealed the presence of a 250-bp inverted repeat, referred to as the 250-bp motif, on either side of the MAT inversion region. The MAT1-1-5 proximal 250-bp motif was part of MAT1-1-1, and the MAT1-1-5 distal 250-bp motif was inverted, and overlapped with the 3'-end of MAT1-2-1 ( Figure 3A; Alignment S1). The DNA sequences of the two motifs were identical. Since the 250-bp motif was absent in B. cinerea MAT1-2-1, but present in MAT1-1-1 of both B. cinerea and S. homoeocarpa (Table S5; Alignment S2), this indicated that the 250-bp motif is a partial MAT1-1-1 that was integrated into MAT1-2-1 in an ancestor of S. sclerotiorum, possibly through a double crossing over ( Figure 4A). The 250-bp motif consisted of 60% AT which is similar to the S. sclerotiorum strain 1980 genome overall AT content [17], lacked repeats, and did not match any transposon sequences in GenBank.
In order to identify the S. sclerotiorum MAT1-1-1 -MAT1-2-1 fusion junction, we aligned S. sclerotiorum MAT with homologous B. cinerea MAT1-1 and MAT1-2 regions (Alignment S3), and found that the fusion of MAT1-1 and MAT1-2 most likely occurred in a region of 10 bp encompassing nucleotides 239 to 248 downstream of S. sclerotiorum MAT1-1-1. This is because DNA sequence homology between B. cinerea and S. sclerotiorum MAT1-1 idiomorphs extended 248 nucleotides downstream of S. sclerotiorum MAT1-1-1, and among the ten nucleotides in positions 239 to 248 downstream of S. sclerotiorum MAT1-1-1, five were identical between B. cinerea MAT gene expression analyses RT-PCR was performed for all four MAT genes in the eight S. sclerotiorum strains 1B331-1 -1B331-8 that represented an ordered tetrad. Our results agreed with previous studies with respect to gene boundaries and intron positions [17], with the following exceptions. MAT1-1-1 was extended by 280 bp at the 5'-end and contained an additional, 49-bp intron. This conclusion was based on the fact that we obtained an RT-PCR product ( Figure 6) with a forward primer situated approximately 200 bp upstream of the previously predicted start codon [17], and a MAT1-1-1 internal reverse primer ( Figure 6). Since an in-frame start codon was present 280 bp upstream of the earlier predicted start codon, conceptual translation indicated that MAT1-1-1 measured at least 1106 bp and contained an additional, 49-bp intron, 223 bp from the 5'-end. The 49-bp intron was spliced in Inv-isolates, but not in Inv+ isolates, resulting in a frame shift, a premature stop codon and a truncated MAT1-1-1 measuring 354 bp ( Figure 7). This is opposed to the 1106 bp-MAT1-1-1 in Inv-isolates ( Table 1). The MAT1-1-1 alpha1 domain in Inv+ isolates was thus shortened to 15 residues, from the 197 residues in Inv-isolates ( Figure 7).
A further difference between MAT loci was the presence of three MAT1-2-1 transcript variants in Inv-isolates, only one of the variants was detected in Inv+ isolates ( Figure 6). The first variant was unspliced, the second variant was spliced as MAT1-2-1 in S. sclerotiorum strain 1980 [17], and the third variant was spliced differently from the other two variants (Figure 7). Assuming identical start codons for all variants, conceptual translation indicated that variants 1 and 3 encoded identical proteins measuring 105 aa in length and lacking known functional domains, whereas variant 2, the only variant detected in both Inv-and Inv+ isolates, encoded a 395 aa protein containing an HMG domain.
The 3'-fragment of MAT1-1-1 was also expressed despite the absence of an in-frame start codon ( Figure 6) and did not contain any known functional domains.

MAT locus copy number
In order to assess whether the S. sclerotiorum MAT locus was single-or multi-copy, Southern hybridization was performed with a MAT1-2-1 probe and all eight S. sclerotiorum strains 1B331-1 -1B331-8 of the ordered tetrad. MAT1-2-1 was targeted due to the detection of three transcript variants differing in length, and thus potentially derived from more than one MAT region. The results obtained were consistent with the presence of a single MAT locus, as all isolates had only one Southern band, diagnostic of the presence or absence of the MAT inversion ( Figure 8).
Since the parent of the two tetrads was unknown, the two contending parental S. sclerotiorum strains were screened. Sclerotinia sclerotiorum strain BS001 was Inv-, and strain BS014 was Inv+ ( Figure 10F).
The orientation of the MAT inversion regions and the flanking 250-bp motifs was confirmed by DNA sequencing in two isolates of the complete tetrad, S. sclerotiorum strain 1B331-1 that was Inv+, and strain 1B331-3 that was Inv-. The DNA sequences of the two MAT inversion regions and the flanking 250-bp motifs were identical, but the orientation of the MAT inversion regions differed between Inv-and Inv+ isolates as expected (Alignment S1).
Proportions of Inv-and Inv+ isolates among random ascospore progeny Random ascospore progeny consisted of equal proportions of Inv-and Inv+ isolates ( Table 2). Progeny of four different parents were screened, including S. sclerotiorum strains BS011, BS013, BS017 and BS028 that were either Inv-or Inv+, and PCR analysis of 18 to 20 progeny for each parent showed that the ratio of Inv-to Inv+ isolates was 1:1 for the progeny of individual parents, and for all progeny combined ( Figure 10, Table S6).

Proportions of Inv-and Inv+ isolates among field populations
PCR screening for presence of the MAT inversion in S. sclerotiorum field populations in twelve different states showed that both types of MAT loci were present in California and Nebraska in equal proportions (Table 3). All isolates from Washington were Inv+, and for the remaining states only one or two isolates were screened (Table 3). Inv+ and Inv-isolates occurred on the same hosts that included lettuce, dry bean and canola, whereas in any given state, isolates from cauliflower, pepper, potato, soybean, sunflower and tobacco were either Inv-or Inv+. However, across states, soybean hosted both Inv+ and Inv-isolates (Table 3).

Comparison between inversion breakpoints of Inv-and Inv+ isolates
To confirm that the 250-bp motif was consistently present and intact on either side of the MAT inversion region, the inversion breakpoints were PCR amplified and sequenced in 32 different isolates of S. sclerotiorum, including 13 Inv-and 19 Inv+ isolates. We found that for each isolate, the inversion breakpoints contained the 250-bp motif as in the Inv-S. sclerotiorum strain 44Ba1 and the Inv+ strains 44Ba12 and 44Ba18 for which the entire MAT regions were sequenced (Alignment S4). All 250-bp motifs generated in this study and of S. sclerotiorum strain 1980 [17] were aligned after reverse complementing the MAT1-1-5 distal 250-bp motifs, and the DNA sequences of all 250-bp motifs were identical ( Figure 11, Alignment S5).

Impact of the MAT inversion on self fertility
We investigated the relationship between MAT inversion and self fertility, and found that among the 36 S. sclerotiorum Inv+ isolates tested, 35 formed fully expanded apothecia which are the sexual fruiting bodies, and of the 21 Inv-isolates, 20 formed fully expanded apothecia (Table 4). Ascospore viability was not assessed in detail, but the ascospores of S. sclerotiorum strains BS001 and BS014 that are Inv-and Inv+, respectively, were viable [44], as were the ascospore progeny used for inversion screening (Table  S6) and the 13 S. sclerotiorum strains used for assessment of mycelial compatibility groups (see below). In an apothecial stalk formation assay with another group of isolates, we found that of the 29 S. sclerotiorum Inv+ isolates, 18 produced apothecial stalks indicative of self-fertility (Table S7). All nine S. sclerotiorum Inv-isolates formed apothecial stalks (Table S7). However, the apothecial stalk assay may not necessarily reflect the ability to form mature apothecia [45].

Discussion
We investigated the Sclerotinia sclerotiorum mating type locus (MAT) in a collection of 283 isolates from lettuce in California and from other states and hosts, and found that S. sclerotiorum has two MAT alleles that differ in the orientation of a 3.6-kb region that is undergoing an inversion in every meiotic generation. The MAT inversion changes the orientation of two genes, truncates another gene and correlates with altered MAT gene expression. MAT inversions have also been reported in other ascomycetes including Cochliobolus kusanoi [24], Peyronellaea spp. [27] and Stemphylium spp. [25], but it is unknown whether in these fungi the inverted MAT regions are stable, or are inverted in every meiotic generation as in S. sclerotiorum.

The MAT inversion region is present in half the progeny
The process associated with the change in orientation of the 3.6kb MAT inversion region in S. sclerotiorum MAT may be highly regulated, because it results in a 1:1 ratio of ascospores that have the inversion and that are referred to as Inv+, to ascospores without the inversion designated Inv-. The 1:1 ratio was observed among random progeny and in ordered tetrads. Half of the 18 to 20 randomly selected ascospores of each of the four apothecia examined, were Inv+ and half were Inv-, regardless of whether the parental isolate was Inv+ or Inv-( Figure 12A). And the ratio of Inv+ to Inv-ascospores in a single ascus was 1:1. The two asci examined displayed inversion distribution patterns among the sibling ascospores indicative of segregation and recombination ( Figure 12B).
The 1:1 ratio of Inv+ to Inv-isolates was also found at the population level in California and Nebraska suggesting that the transition between Inv-and Inv+ that we documented in the laboratory, may also commonly occur in nature (Table 3). Inv+ isolates were widespread, and in addition to California and Nebraska, were also found in Georgia, Missouri, North Dakota, Ohio, South Dakota and Washington State, but with the exception of Washington where all 15 isolates tested were Inv+, no more than two isolates were included for each of these states (Table 3). Based on our dataset, both Inv-and Inv+ isolates were obtained from lettuce, canola and dry bean from California, Georgia and Nebraska, respectively. For all other states only single isolates were screened (Table S1, Table 3), with the exception of potato in Washington where all 15 isolates were Inv+ (Table 3).

MAT inversion does not occur during mycelial growth and sclerotia formation
We did not investigate in detail when during the S. sclerotiorum life cycle and under which conditions the MAT inversion occurs, but based on PCR results it appears that the rearrangement of the MAT inversion region does not occur during vegetative hyphal growth or during sclerotium formation. This is because the mycelia used for DNA extraction were grown from sclerotia that were generated in the laboratory, and all strains were either Invor Inv+ based on PCR assays, and none were both ( Figure 10). Also, the inversion phenotype of S. sclerotiorum strain 1B331-8 did not change following subculturing (data not shown). This is dissimilar to human blood cells where DNA regions flanked by inverted repeats similarly to the MAT inversion region, undergo recurrent inversions that result in cell populations that are mixed with regard to the orientations of inversion regions [46]. However, we expect a mixed population of nuclei within the same structure following rearrangement of the MAT inversion region, as S. sclerotiorum asci harbor both Inv-and Inv+ ascospores (Figures 8,  10, 12).
Under appropriate conditions, sclerotia in S. sclerotiorum give rise to apothecia that are sexual fruiting bodies where meiosis occurs within the asci, and since we did not detect rearrangement of the MAT inversion region during vegetative growth, inversion appears to occur during the sexual cycle. The arrangement of Inv-and Inv+ ascospores inside the asci is consistent with first and second  division segregation during meiosis ( Figure 12B), which suggests that Inv-and Inv+ nuclei were paired up prior to meiosis, and thus, the inversion precedes meiosis. Furthermore, pairing of Invand Inv+ nuclei may require nuclear recognition, a process that in heterothallic ascomycetes involves MAT and the production of nucleus-specific hormones and hormone receptors [32,33,34,35]. Sclerotinia sclerotiorum is homothallic, but Inv+ and Inv-MAT regions differ in gene expression ( Figure 6) and may be functionally equivalent to MAT1-1 and MAT1-2 of heterothallics, also because Inv+ MAT appears to lack a functional alpha1 box ( Figure 7). Thus, we hypothesize that rearrangement of the MAT inversion region occurs before meiosis but following sclerotium formation, and is involved in recognition between Inv-and Inv+ nuclei.
We do not know what triggers the MAT inversion, but in human blood cells, the recurrent inversion events are caused by non-allelic homologous recombination [46], a process that is initiated through DNA double strand breaks [47]. Non-allelic homologous recom-   bination that results in genome rearrangements is also known from Saccharomyces cerevisiae [48].

MAT inversion is facilitated by two 250-bp motifs
The trigger for the MAT inversion is unknown. But the inversion event is most likely caused by non-allelic homologous recombination [46,47] between the 250-bp motifs that flank the 3.6-kb MAT inversion region. The MAT1-1-5 distal 250-bp motif is inverted with respect to the MAT1-1-5 proximal 250-bp motif, and juxtaposition of the two motifs and recombination results in a 3.6-kb inversion without altering the DNA sequence composition of the 250-bp motifs (Figure 3). Since one 250-bp motif is located near the center of MAT1-1-1, and the other 250-bp motif overlapped with the 3'-end of MAT1-2-1, only MAT1-1-1 is truncated by the inversion (Figure 3; Alignment S1). In all 36 isolates examined, the two 250-bp motifs had identical DNA sequence composition ( Figure 11). We sequenced across the two 250-bp motifs and the 3.6-kb MAT inversion region in S. sclerotiorum strains 1B331-1 (Inv+) and 1B331-3 (Inv-), two progeny of the ordered tetrad, and did not find any differences between the two isolates apart from the orientation of the MAT inversion region (Alignment S1).

MAT locus is single copy
There is just one MAT locus in S. sclerotiorum, based on the analysis of strain 1980 [17]. And the results from Southern blotting with a MAT1-2-1 probe (Figure 8) were consistent with the presence of either Inv+ or Inv-MAT as documented by PCR screening (Figure 10). Thus, the transition between Inv-and Inv+ isolates is not due to the exchange of the MAT inversion region between two MAT loci, and thus differs from mating type switching in baker's yeast, Saccharomyces cerevisiae [49].
The S. sclerotiorum MAT arrangement resembles most closely the MAT arrangement of homothallic Stemphylium species where the MAT1-1 idiomorph is inverted with respect to heterothallic relatives, and is fused to MAT1-2 [25]. The fused MAT arrangement in Stemphylium evolved following the inversion of MAT1-1 that created a short stretch of DNA sequence identity between idiomorphs, facilitating a crossing over that resulted in the MAT1-1 -MAT1-2 fusion [25]. Similarly to Stemphylium, the orientation of MAT1-2 differs between S. sclerotiorum and B. cinerea [17]. To investigate whether in an ancestor of S. sclerotiorum the MAT1-2 inversion could have facilitated crossing over resulting in MAT fusion, we looked for a crossover site and aligned the MAT1-1 -MAT1-2 fusion junction in S. sclerotiorum Inv-isolates with homologous regions of B. cinerea, a proxy for the unknown ancestor. We found that the fusion of MAT1-1 and MAT1-2 most likely occurred in a region of 10 bp encompassing nucleotides 239 to 248 downstream of S. sclerotiorum MAT1-1-1. This is because DNA sequence homology between B. cinerea and S. sclerotiorum MAT1-1 idiomorphs extends 248 nucleotides downstream of S. sclerotiorum MAT1-1-1, and among the ten nucleotides in positions 239 to 248 downstream of S. sclerotiorum MAT1-1-1, five were identical between B. cinerea MAT1-1, S. sclerotiorum MAT1-1 and B. cinerea MAT1-2 ( Figure 5; Alignment S3). The relatively high DNA sequence identity between B. cinerea MAT1-1 and MAT1-2 idiomorphs at the S. sclerotiorum MAT1-1 -MAT1-2 fusion junction is consistent with the presence of an ancient crossover site. Unlike MAT1-1, the B. cinerea and S. sclerotiorum MAT1-2 regions were generally too divergent to be aligned with confidence outside of the presumptive crossover site (Alignment S3).
In other homothallic ascomycetes, MAT1-1 -MAT1-2 fusion junctions with similarities to homologous regions of related heterothallics are also known. Between the homothallic Cochliobolus luttrellii and its heterothallic relative, the fusion junction consisted of seven consecutive nucleotides shared between the two species [24]. Similarly, in Crivellia there were four shared consecutive nucleotides [26], in Didymella three [27] and in Stemphylium four [25].
An alternative explanation for the evolution of the S. sclerotiorum and B. cinerea MAT loci has been proposed. It was suggested that the fused idiomorphs of S. sclerotiorum were ancestral, and the separate idiomorphs of B. cinerea were derived [17]. This evolutionary scenario is less likely, since it requires at least three different steps to occur in an ancestor of B. cinerea, including a loss of MAT1-1, a loss of MAT1-2 and an inversion of MAT1-2. Our scenario is more parsimonious since it requires only two steps which are inversion of MAT1-2, and fusion of MAT1-2 to MAT1-1. The presence of partial MAT1-1-1 and MAT1-2-1 genes in B. cinerea MAT, proposed remnants of incompletely deleted MAT1-1-1 and MAT1-2-1 [17], likely resulted by crossing over between B. cinerea MAT1-1 and MAT1-2 after evolutionary separation from the S. sclerotiorum lineage.
A further difference between B. cinerea and S. sclerotiorum MAT is the presence of the 250-bp motifs delimiting the 3.6-kb MAT inversion region in S. sclerotiorum. Since a 250-bp motif is absent in B. cinerea MAT1-2-1, but present in MAT1-1-1 of both B. cinerea and S. homoeocarpa (Table S5; Alignment S2), the 250-bp motif is most likely a partial MAT1-1-1 that was integrated into MAT1-2-1 in an ancestor of S. sclerotiorum, possibly through a double crossing   (Table S1), and their progeny (Table S6), including the tetrads in Figure 10F. The top gel in each part figure shows the PCR bands obtained with primer pair Type-IIF / Type-IIR specific to Inv+, and the over ( Figure 4A). However, no potential crossover sites were found when comparing the flanks of the 250-bp motif in S. sclerotiorum to homologous regions of B. cinerea MAT1-1 and MAT1-2. Transposons are known to mediate inversions, for instance in bacteria [53] and in Drosophila [54,55], but the 250-bp motif did not contain any transposon sequences and lacked repeats.

MAT inversion correlates with MAT gene expression
Gene expression was examined in the eight S. sclerotiorum strains 1B331-1 -1B331-8 of the complete tetrad, and differences in expression of MAT1-1-1 and MAT1-2-1 were found between Invand Inv+ isolates (Figures 6, 7). In each of the four Inv-isolates, there were three MAT1-2-1 transcript variants, two of which did not encode any known proteins. The third variant encoded an HMG protein, and was the only MAT1-2-1 transcript present in the S. sclerotiorum strain 1980 transcriptome from the recent S. sclerotiorum genome sequencing project [17]. Inv-isolates contained a single MAT1-1-1 transcript variant encoding an alpha1 protein (Figures 6, 7), but only the 3'-end of the transcript was found in the S. sclerotiorum strain 1980 transcriptome. The four Inv+ isolates had only one of the three MAT1-2-1 transcript variants of Inv-isolates, and a different MAT1-1-1 variant. The Inv+ MAT1-2-1 variant encoded an HMG protein, and the Inv+ MAT1-1-1 variant encoded a truncated alpha1 protein, lacking all but the first Nterminal 15 alpha1 box residues of the 197 residues that are present in Inv-isolates. As expected, no full-length match to the Inv+ MAT1-1-1 variant was found in the S. sclerotiorum strain 1980 transcriptome that was Inv-. Reasons for the absence of MAT1-2-1 transcript variants 1 and 3 (Figure 7) from the S. sclerotiorum strain 1980 transcriptome may include incomplete transcriptome sequencing coverage, or alteration in MAT gene expression patterns due to differences in growth conditions. Multiple transcripts for MAT1-1-1 and MAT1-2-1 were also present in Cochliobolus heterostrophus [56], and alteration of gene expression upon gene inversion was also documented in Drosophila [54,55].
Based on our MAT gene expression data (Figure 7), the severely truncated MAT1-1-1 of S. sclerotiorum Inv+ isolates is expected to be non-functional, and thus, Inv+ isolates are expected to be functionally equivalent to heterothallic MAT1-2 isolates that generally are unable to self. However, we found that S. sclerotiorum Inv+ isolates were able to self, indicating that the truncated MAT1-1-1 may be functional, that a complete MAT1-1-1 may be expressed in a different part of the life cycle, or that Inv+ isolates are able to self without a functional MAT1-1-1. It is known that co-existence of MAT1-1-1 and MAT1-2-1 within one genome is not always required for selfing, as certain Stemphylium species and Neurospora africana are homothallic, but contain only a single MAT idiomorph [25,57].
Altered MAT gene expression did not have any influence on mycelial compatibility groups, as for all 13 parental strains tested, parent and progeny were of the same mycelial compatibility group, as were the S. sclerotiorum strains 1B331-1 -1B331-8 representing the ordered tetrad.

Sclerotinia sclerotiorum has two MAT alleles
The different versions of the MAT locus within one species are typically referred to as idiomorphs, due to their general lack of DNA sequence homology. The case of S. sclerotiorum is different, the Inv-and Inv+ versions of MAT differ mainly by an inversion (Figure 1), and it is readily apparent that they are 'related in structure and descent' unlike idiomorphs [14]. We thus propose to refer to the two versions of MAT in S. sclerotiorum as alleles. This may not be entirely accurate since alleles typically refer to different versions of a single gene and not groups of genes, but the term allele instead of idiomorph was previously used to refer to different versions of the mating type locus of Neurospora crassa [14].

Presence of a MAT inversion region could explain mating type switching in filamentous fungi
In several species of filamentous ascomycetes, only half the progeny derived from a selfing parent is able to self, and the other half is self-sterile but is able to outcross to self-fertile strains. Thus, in these filamentous ascomycetes, there is a transition in life style from homothallism to heterothallism that occurs within a single generation, a process known as mating type switching [58] that differs from mating type switching in baker's yeast, Saccharomyces cerevisiae [28]. Examples of filamentous ascomycetes where mating type switching has been demonstrated include Chromocrea spinulosa [21], S. trifoliorum [22] and certain Ceratocystis species [23].
In progeny of Ch. spinulosa and S. trifoliorum, mating type correlates with ascospore size dimorphism. Heterothallic progeny show a decrease in ascospore size with respect to homothallic progeny [21,40]. Similarly, heterothallic strains of Ce. fimbriata grow more slowly than homothallic strains [23]. Thus, altered ascospore size or growth rate may represent pleiotropic effects of MAT, similarly to Coniochaeta tetraspora where half of the eight ascospores in each ascus are aborted [59].
Molecular data on mating type switching does not contradict involvement of an inversion. In Ceratocystis spp., Witthuhn et al. [60] used HMG box specific PCR primers and found that PCR amplification failed in heterothallic isolates. This would be expected if the HMG box comprised an inversion breakpoint. In Ch. spinulosa, homothallic strains have a MAT1-2-1 adjacent to a MAT1-1 idiomorph, the MAT1-2-1 is missing in that position in heterothallic isolates [61]. There is evidence for inversion breakpoints in homothallic Ch. spinulosa isolates in the form of repeats situated on either side of MAT1-2-1. One of the repeats is present upstream of MAT1-2-1, the other downstream of MAT1-2-1, inside MAT1-1. Since the distance between upstream repeat and MAT1-2-1 is greater than the distance between downstream repeat and MAT1-2-1, inversion of the region between the repeats would increase the distance between MAT1-2-1 and MAT1-1. This could explain why MAT1-2-1 was not detected near MAT1-1 in heterothallic isolates. However, more research on the nature of mating type switching in Ch. spinulosa and other species should include comparisons of genomes and gene transcription profiles between parents and offspring.

Further research
Given the importance to agriculture, it is surprising how little is known about sexual processes in S. sclerotiorum. Future research should focus on elucidating when and how the inversion occurs, whether there are differences in target gene expression between Inv-and Inv+ strains, and whether the inversion phenotype plays a role in fertilization and outcrossing.

Fungal strains and culture conditions
This study was based on 283 Sclerotinia sclerotiorum isolates of which 214 were from lettuce in California. The remaining 69 isolates were from nine other substrates (canola, cauliflower, dry bean, field soil, pepper, potato, soybean, sunflower, and tobacco) from twelve states (California, Georgia, Illinois, Kansas, Minnesota, Missouri, North Dakota, Nebraska, Ohio, South Dakota, Washington and Wisconsin), or were part of ordered tetrads obtained under laboratory conditions (Table S1). Fungi were grown on potato dextrose agar medium (PDA) (Sigma-Aldrich, St. Louis, MO) and potato dextrose broth (PDB) (Sigma-Aldrich, St. Louis, MO) at room temperature. California isolates were derived from sclerotia collected from infected plants. One sclerotium per plant was surface sterilized in 10% bleach for 3 min, plated on PDA, and hyphal tips were transferred to fresh PDA plates to obtain pure cultures. Stocks were maintained as sclerotia at 4uC.
Additional strains used in this study were progeny obtained as random ascospore isolates from strains listed in Table S1. Ascospore suspensions were prepared as described by Wu et al. [44], diluted, spread out on PDA, and single germinated ascospores were transferred to new PDA plates by excising the Table 3. Occurrence of Sclerotinia sclerotiorum Inv+ isolates in twelve states and on different hosts in the United States based on PCR screening of all isolates in Table S1. Percentage is given when more than one isolate was screened. B Inv-to Inv+ ratio was 1:1 (x 2 test, P = 0.14). C Inv-to Inv+ ratio was 1:1 (x 2 test, P = 0.24). doi:10.1371/journal.pone.0056895.t003 agar block containing the ascospore using a forceps. Progeny are not included in Table S1, but parental strains are mentioned in the text.

Nucleic acid extraction
Sclerotinia sclerotiorum mycelia were grown at room temperature in 25 ml PDB on a lab bench, inoculated with three to four day old PDA cultures, and harvested after four days. For PCR, DNA was extracted by FastDNA extraction kit (MP Biomedicals, Solon, OH). For Southern blotting, a standard phenol-chloroform protocol was used [62]. Total RNA was extracted using TRIZOL (Invitrogen Life Technologies, Carlsbad, CA) following the manufacturer's protocol. The integrity of the RNA was verified by agarose gel electrophoresis. Nucleic acids were quantified using a NanoDrop system (NanoDrop Technologies, Wilmington, DE) and the DNA concentration was adjusted to 2-10 ng/ml in PCR grade water for PCR amplification and to 5-8 mg for Southern blotting.

Cloning of MAT loci, phylogenetic analyses and identification of repeated motifs
Complete MAT locus coverage of S. sclerotiorum strains 44Ba1, 44Ba12 and 44Ba18 was obtained by PCR using primers designed based on the S. sclerotiorum strain 1980 MAT locus sequences [17]. Three overlapping PCR reactions targeted the MAT locus and parts of the flanking APN2 and SLA2, using primer pairs MAT_1069F / MAT_6555R, MAT_6042F / MAT_11930R and MAT_11655F / MAT_14587R (Table S8). PCRs were performed in 50 ml reactions consisting of 25 ml GoTaq Green Master Mix (Promega Corp., Madison, WI), 1.5 ml of each primer (10 pmol/ml), 5 ml of genomic DNA (2 ng/ml) and 17 ml of PCR grade water. PCR amplifications were carried out on a Bio-Rad DNA Engine thermocycler (Bio-Rad Laboratories, Hercules, CA) with the following conditions: Initial denaturation at 95uC for 10 min, followed by 35 cycles of 95uC for 20 sec, 60uC for 30 sec, and 60uC for 5 min, with a 10 min final extension at 72uC [63].
Aliquots of PCR products (6ml) were separated on a 1% agarose gel by electrophoresis, gels were stained with ethidium bromide and visualized under a UV trans-illuminator (Ultra-Violet Products Ltd, Cambridge, UK). PCR products were sequenced with the PCR primers and additional internal primers were designed to generate complete sequencing coverage in both directions for all three isolates. Primer sequences are given in Tables S8, S9, and S10. DNA sequencing was performed at the UC Davis DNA Sequencing Facility, using ABI BigDye Terminator v3.1 Cycle Sequencing chemistry on an ABI 3730 Capillary Electrophoresis Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). Sequences were assembled by Geneious v4.8.5 [64]. The S. sclerotiorum strains 44Ba1, 44Ba12 and 44Ba18 MAT locus sequences were deposited in GenBank (JQ815883, JQ815884, and JQ815885, respectively).

PCR screening for MAT inversion and sequencing of inversion breakpoints
All S. sclerotiorum isolates in Table S1 and the progeny of S. sclerotiorum strains BS011, BS013, BS017 and BS028 listed in Table  S6 were screened for the presence and absence of the MAT inversion using two sets of PCR primers. The Inv+ specific primer pair Type-IIF / Type-IIR (Table S11) targeted a 1306 bp fragment spanning the MAT1-1-5 proximal inversion breakpoint ( Figure S1), and the primer pair MAT1-1-F / MAT1-1-R targeted a 671 bp fragment of the intact MAT1-1-1 of Inv-isolates [69]. Primer binding sites are illustrated in Figure S1. PCR conditions were as described for MAT locus cloning, except that the extension temperature and time were 72uC and 60 sec, respectively. PCR reactions were performed on a Bio-Rad DNA Engine thermocycler (Bio-Rad Laboratories, Hercules, CA). Aliquots of PCR products (6 ml) were separated on a 1% agarose gel by electrophoresis, stained with ethidium bromide and visualized under UV trans-illuminator (Ultra-Violet Products Ltd, Cambridge, UK). Selected PCR products were sequenced to confirm target-specific amplification.
The MAT inversion region was PCR amplified and sequenced for two of the progeny of the complete tetrad, S. sclerotiorum strain  Figure 12. The S. sclerotiorum MAT inversion region changes orientation in every meiotic generation and segregates at the first or second division of meiosis. Small ovals are ascospores, large ovals are asci, yellow ascospores are Inv+, red ascospores are Inv-. Single ascospores represent parental isolates, asci containing eight ascospores represent progeny. 12A) Regardless of whether the parent is Inv-or Inv+, 50% of the progeny is Inv+, suggesting a highly regulated process during the sexual cycle ensures the 1:1 ratio. The transition of Inv-to Inv+ involves a change in orientation of the MAT inversion region. 12B) Meiotic segregation pattern of Inv-and Inv+ among ascospores in S. sclerotiorum. The pattern that is characteristic for first division segregation in Neurospora crassa is shown on the left, the pattern for second division segregation following recombination is shown on the right [42]. Figure 12 is based on PCR screens for absence and presence of the MAT inversion region in tetrad and random progeny and their parents with DNA from the mycelial phase (Figures 9, 10), DNA sequencing of inversion breakpoints ( Figure 11) and the entire inversion regions and flanks in S. sclerotiorum strains 1B331-1 and 1B331-3 of the complete ordered tetrad (Alignment S1), and Southern blotting illustrating that S. sclerotiorum MAT is single copy (Figure 8). doi:10.1371/journal.pone.0056895.g012  [63]: Initial denaturation at 94uC for 3 min, followed by 35 cycles of 94uC for 10 sec, 60uC for 20 sec, and 60uC for 6 min, with a 7 min final extension at 60uC. The resulting PCR bands were cut out from gel, melted in 50 ml PCR quality water, diluted 10 and 100 times, and used as templates in PCR reactions with three overlapping primer pairs, SS1f / SSr5, SSf2 / SSr2, SSr4 / SS2r, for S. sclerotiorum strain 1B331-1, and primer pairs SS1f / SSr1, SSf2 / SSr2, SSf3 / SS2r, for S. sclerotiorum strain 1B331-3. The PCR primers, as well as primer SSr7 for S. sclerotiorum strain 1B331-1, and primers SSf5 and SSf6 for S. sclerotiorum strain 1B331-3 were used for DNA sequencing as described above (Table S11).

MAT gene expression analysis
The expression of all MAT genes was assessed using primer pairs specific to each gene (Figure 6), in S. sclerotiorum strains 1B331-1, 1B331-2, 1B331-5 and 1B331-6 that are Inv+, and strains 1B331-3, 1B331-4, 1B331-7 and 1B331-8 that are Inv-. The primer pairs for each gene and approximate expected amplicon sizes were as follows.  (Table S12). First-strand cDNA synthesis was done with DNAse treated total RNA using Clontech SMARTer cDNA synthesis kit following the manufacturer's instructions (Clontech Laboratories, Inc., CA). Double-stranded cDNA for all MAT genes was synthesized using the gene-specific primer pairs listed in Table S12.
When more than one transcript was obtained, PCR products were cloned into pCR2.1-TOPO following the manufacturer's protocol (Invitrogen Life Technologies, Carlsbad, CA) and sequenced using both vector (M13F / M13R) and gene specific primers (Table S5). Exon-intron boundaries were determined by comparison of transcript and gene sequences using CLUSTALW V2.1 [66].

Southern blotting of MAT region
To assess the numbers of MAT loci in S. sclerotiorum genomes, Southern blotting was performed for S. sclerotiorum isolates derived from the ordered tetrad. These were the S. sclerotiorum Inv+ strains 1B331-1, 1B331-2, 1B331-5 and 1B331-6, and the S. sclerotiorum Inv-strains 1B331-3, 1B331-4, 1B331-7 and 1B331-8. Genomic DNA (5 to 8 mg) was digested with BsaHI that had restriction sites upstream, downstream and on the MAT inversion ( Figure 8). Digests were run on 0.8% agarose gel and transferred to nylon membrane (Roche, Basel, Switzerland) following Selden [70]. A Southern probe specific to MAT1-2-1 (Figure 8) was generated by PCR using primers MAT1-2-F / MAT1-2-R [69]. The probe was labeled with digoxigenin (Roche, Basel, Switzerland) according to the manufacturer's instructions, and hybridization and detection was performed according to the manufacturer instructions (Roche, Basel, Switzerland). X-ray film (Kodak, Rochester, NY) was used and developed following a 30 min exposure. The probe was expected to hybridize to 1.16 and 1.5 kb fragments in Inv-and Inv+ isolates, respectively.

Generation of the sexual state in culture
Apothecium formation was evaluated following the method of Wu et al. [44]. Two different groups of isolates were used. The production of fully expanded apothecia was assessed for a group of 57 isolates based on 20 sclerotia per strain ( Table 4). The formation of apothecial stalks, an approximate measure of apothecia formation, was assessed for another group of 38 isolates (Table S7), and repeated once for isolates that failed to produce apothecial stalks in the first repetition.

Isolation of ordered tetrads
Ordered tetrads were obtained by culturing all ascospores of a single ascus. Asci were derived from a fully expanded apothecium grown in culture [44] and were released by placing apothecia in 1.5 ml centrifuge tubes, adding sterile water and gentle macerating by forceps. Random asci containing eight mature ascospores were squeezed from the distal end to release the ascospores using a scalpel, and ascospores were transferred in sequence to individual PDA plates and stored at 4uC. Two tetrads were obtained, a complete tetrad was represented by the eight S. sclerotiorum strains 1B331 -1B331-8, another tetrad was incomplete and consisted of the four S. sclerotiorum strains 1B321-2, 1B321-4, 1B321-6 and 1B321-8. The numbers following the dashes reflect the ascospore positions inside the ascus. For instance, S. sclerotiorum strain 1B331-1 was derived from the ascospore closest to the tip of the ascus. Tetrads were isolated from two different parental isolates, S. sclerotiorum strains BS001 and BS014. Our lab notes were inconclusive as to which of the two strains was the parent of the tetrads used in this study. Strains BS001 and BS014 were identical at the variable MAT regions that were sequenced and compared (data not shown).

Mycelial compatibility group assessment
Mycelial compatibility groups were assessed for 13 S. sclerotiorum isolates (BS001, BS002, BS003, BS011, BS013, BS014, BS017, BS028, BS047, BS058, BS071, BS095, BS096) (Table S1) and their progeny, at least 20 progeny were tested for each parent. Progeny were paired with their parent as described by Wu and Subbarao [3] following the methodology of Schafer and Kohn [71]. Sclerotinia sclerotiorum strains 1B331-1 -1B331-8 representing the complete ordered tetrad were paired in all possible combinations among one another. Figure S1 Positions of primers used for MAT inversion screening (MAT1-1F / MAT1-1R, Type-IIF / Type-IIR), and PCR amplification and sequencing of the inversion breakpoints in S. sclerotiorum isolates S1A) without (Inv-) and, S1B) with inversion (Inv+). Genes are boxes, white and dotted boxes correspond to alpha1 and HMG domains, respectively, directions of transcription are indicated by arrows, gene names are inside or by the boxes. Dashed box and arrow represent MAT1-1-1 3'-end fragment lacking an in frame start codon. Primer sites are indicated by half arrows. Diagrams are to scale. The Inv+ alpha1 box is truncated after 45 bp and is not illustrated, for details see text.

Supporting Information
(TIF) Table S1 Sclerotinia sclerotiorum isolates used in this study. Given are the isolate identifiers, the host's scientific and common names, the location and year of collection, the source, as well as the orientation of the MAT inversion. (DOC) Table S3 Phylogenetic analyses of Sclerotinia sclerotiorum and Botrytis cinerea MAT genes using maximum parsimony. For each analysis, the number of taxa, alignment length, the number of parsimony informative characters, the number and length of most parsimonious trees (MPT) and the consistency (CI) and retention indices (RI), are given. Respective alignments are available as Alignments S6, S7, S8 and S9. (DOC)     Table S10 Sequencing primers used in Sclerotinia sclerotiorum strain 44Ba1 targeting the MAT inversion region in S. sclerotiorum strains 44Ba12 and 44Ba18. The last letter in a primer name indicates the primer direction, forward and reverse, respectively. (DOC) Table S11 Primer pairs used for Sclerotinia sclerotiorum Inv+ PCR screening, and PCR amplification and DNA sequencing of inversion breakpoints and MAT inversion region. The 'f' or 'r' in a primer name indicates the primer direction, forward and reverse, respectively. (DOC) Alignment S1 FASTA text file with alignment of Sclerotinia sclerotiorum strains 1980, 44Ba1, 44Ba12 and 44Ba18 MAT regions, the MAT1-1-1 and MAT1-2-1 transcript variants, and the MAT inversion region sequences of the tetrad progeny S. sclerotiorum strains 1B331-1 (Inv+) and 1B331-3 (Inv-) obtained with primers SS1f / SS2r. Sequence accession numbers are given as part of sequence names. Indicated are genes, the MAT inversion region and the 250-bp motifs. Sequence accession numbers are given as part of sequence names for sequences in public databases. Alignment S4 FASTA text file with alignment of Sclerotinia sclerotiorum strains 44Ba1 (Inv-), 44Ba12 (Inv+), 44Ba18 (Inv+) and 1980 (Inv-) and inversion breakpoint sequences obtained with primers SS1f, SS1r2, SS2f2 and SS2r (Table S11, Figure S1) for 32 S. sclerotiorum isolates demonstrating that the 250-bp motifs spanning the inversion breakpoints were identical in all isolates. Positions of strain 44Ba1 MAT1-1-1 and MAT1-2-1 and the 250bp motifs are indicated. Sequence accession numbers, strain identifiers used in this study (Table S1, Table S6), and primers used to generate each sequence are part of the sequence names. None of the SS1f, SS1r2, SS2f2 or SS2r sequences was submitted to GenBank. (TXT) Alignment S5 FASTA text file with alignment of all 250bp motifs from Alignment S4, used to generate Figure 11. The MAT1-1-5 distal 250-bp motifs were reverse complemented before alignment. Taxon labels are as in Figure 11.