Maintenance of Sex-Related Genes and the Co-Occurrence of Both Mating Types in Verticillium dahliae

Verticillium dahliae is a cosmopolitan, soilborne fungus that causes a significant wilt disease on a wide variety of plant hosts including economically important crops, ornamentals, and timber species. Clonal expansion through asexual reproduction plays a vital role in recurring plant epidemics caused by this pathogen. The recent discovery of recombination between clonal lineages and preliminary investigations of the meiotic gene inventory of V. dahliae suggest that cryptic sex appears to be rare in this species. Here we expanded on previous findings on the sexual nature of V. dahliae. Only 1% of isolates in a global collection of 1120 phytopathogenic V. dahliae isolates contained the MAT1-1 idiomorph, whereas 99% contained MAT1-2. Nine unique multilocus microsatellite types comprised isolates of both mating types, eight of which were collected from the same substrate at the same time. Orthologs of 88 previously characterized sex-related genes from fungal model systems in the Ascoymycota were identified in the genome of V. dahliae, out of 93 genes investigated. Results of RT-PCR experiments using both mating types revealed that 10 arbitrarily chosen sex-related genes, including MAT1-1-1 and MAT1-2-1, were constitutively expressed in V. dahliae cultures grown under laboratory conditions. Ratios of non-synonymous (amino-acid altering) to synonymous (silent) substitutions in V. dahliae MAT1-1-1 and MAT1-2-1 sequences were indistinguishable from the ratios observed in the MAT genes of sexual fungi in the Pezizomycotina. Patterns consistent with strong purifying selection were also observed in 18 other arbitrarily chosen V. dahliae sex-related genes, relative to the patterns in orthologs from fungi with known sexual stages. This study builds upon recent findings from other laboratories and mounts further evidence for an ancestral or cryptic sexual stage in V. dahliae.


Introduction
Sexual reproduction is thought [1] to act as a mechanism to combine fit alleles from different individuals, and to break apart locally disadvantageous allele combinations under dynamic selection pressures [2]. While sexual reproduction may in theory be costly and disrupt favorable gene combinations, experimental evidence has suggested that sex in fungi increases the rate of adaptation to new environments [3]. Prior to molecular techniques, the formation of sexual structures and spores was the primary evidence of sex in fungi. It is now evident that sex in many taxa is rare, unpredictable and elusive. For many fungi, the only documented sexual structures are formed on certain media and/or growth conditions in vitro [4,5]. Some putatively asexual plant pathogens have been found to sexually reproduce in nature only in specific ecological conditions and geographic locales, such as near the center of origin of the species [6].
Advances in genetic markers and population biology have led to significant advances in the discovery of rare or cryptic sexual stages in fungi [7]. Populations of many species that lack obvious sexual stages in nature nevertheless have been found to harbor molecular patterns of sexuality based on investigations of mating type frequencies, population structure, multilocus linkage disequilibrium [8,9] and computer simulations [6]. Additionally, bioinformatic surveys of complete genomes, have been used to infer sexuality based on the meiotic gene inventory [10][11][12][13]. Advances in genomics have enabled the unprecedented implementation of these approaches to investigate sexuality in fungi. Many seemingly asexual fungi have retained the genes required for the sexual ''machinery'', including many that are important to the fields of agriculture and medicine [12,[14][15][16].
Verticillium is a small genus of phytopathogenic fungi that causes billions of dollars in agricultural losses annually [17]. Verticillium dahliae is a cosmopolitan, soilborne plant pathogen that causes an economically significant wilt disease. It is known for its extremely wide host range [18] and its ability to survive in soils as dormant resting structures for many years [17,19]. Historically, V. dahliae has been considered strictly asexual because it has failed to form sexual structures under the laboratory conditions tested. Vegetative anastomosis, the fusion of growing hyphae under laboratory conditions, has been reported [20,21], and several vegetative compatibility groups (VCGs) have been classified. Deep sequencing of all known VCGs of V. dahliae has revealed that VCGs are strongly correlated to clonal lineages [22], but has also revealed that putative sexual recombination between clonal lineages has occurred rarely [23].
Sexual compatibility and fruiting body formation in heterothallic fungi in the Ascoymycota is determined by a variety of sexrelated gene pathways. Of primary importance are the two idiomorphs of the MAT locus, which differ in gene content and are the master regulators of sexual recombination in the Ascoymycota [24]. One idiomorph contains a critical gene that encodes an a domain (MAT1-1-1), while the other contains a critical gene that encodes a DNA-binding domain of the high-mobility group (HMG) type (MAT1-2-1) [25]. Isolates with either of the idiomorphs are referred to as MAT1-1 or MAT1-2 [26]. Verticillium dahliae is considered heterothallic because both idiomorphs are known to exist [27], and only one idiomorph has been observed in any one isolate.
Previous sequences of the a and HMG domains of V. dahliae MAT genes showed high amino acid conservation with other fungi in the subphylum Pezizomycotina [27,28]. Mating type frequencies in V. dahliae have been reported in multiple studies as skewed [23,28,29]. Although previous studies have reported mating type distributions skewed towards MAT1-2 in V. dahliae, they have not clearly stated whether both mating types are sympatric in nature, that is, whether isolates of opposite mating coexist in nature. It is also unknown whether genetically identical multilocus microsatellite types contain both MAT idiomorphs, a condition which has previously been interpreted as unequivocal evidence for sexual recombination [11].
In addition to the presence of both mating types, other molecular signatures suggestive of sex have been reported in V. dahliae. Multilocus linkage equilibrium has been reported in collections of V. dahliae [30], although clonal expansion is of primary importance in pathogen reproduction and dissemination within regions where this pathogen is a severe problem in agriculture [31]. However, even in species with known sexual stages, signatures of clonality can predominate in multilocus data sets [32]. Gene trees with incongruent topologies may be a robust indicator of meiotic recombination when they occur within a strongly supported phylogenetic species [33,34]. Gene trees with incongruent topologies were previously reported in V. dahliae based on sequences of the protein coding genes actin (ACT), elongation factor 1-alpha (EF), glyceraldehyde-3-phosphate dehydrogenase (GPD), and tryptophan synthase (TS) [35]. The strongest evidence yet of recombination between lineages of V. dahliae was based on over 20,000 single nucleotide polymorphisms (SNPs) [23].
Genomic investigations of V. dahliae have also provided some evidence of sexuality in V. dahliae. For example, a single homolog of the gene encoding the DNA methyltranferase (DMT) RID exists in V. dahliae reference strain Ls 17, a gene which was first characterized as part of the Repeat-Induced Point (RIP) machinery in N. crassa [36]. Patterns consistent with RIP-like mutation were subsequently discovered in the V. dahliae genome in multiple long interspersed element (LINE)-like and long terminal repeat (LTR) retroelement sequences [37] and other transposons [38]. Furthermore, preliminary explorations of the meiotic gene inventory have revealed the presence of genes known to function in sex-related pathways in other fungal systems [23].
Comparative population genomics of V. dahliae has significantly advanced the understanding of the molecular basis of races, as well as the existence of inter-Kingdom horizontal gene transfer [39], and has also led some researchers to posit chromosomal reshuffling (genomic rearrangements and chromosomal length polymorphisms, despite a high degree of sequence conservation) as the sole mechanism for generating the diversity observed within V. dahliae [40]. Significant chromosomal rearrangements are expected to interfere with meiosis [41], so it is reasonable to expect sex to be impossible between isolates with extreme karyotypic polymorphisms [40,42].
It has been postulated that a detailed understanding of the genes required for the initiation and completion of meiosis in sexual fungi, that it should be possible to understand the molecular mechanisms that control sexual compatibility and to determine which of these genes are missing or nonfunctional in asexual fungi [43]. In fact, imperfect functioning of mating type genes and other sexual factors such as pheromone receptors have been hypothesized in V. dahliae [27]. In the context of exploring the functionality of sex-related genes (and not merely the existence of pseudogenes), Reverse transcriptase-PCR has been used to show that both mating type genes are expressed in fungi, for which no known sexual stage has been documented [44], while other studies have demonstrated pheromone receptor and precursor gene expression in other putatively asexual fungi [45]. To date, the expression of MAT genes and other sex-related genes in V. dahliae has never been investigated.
Evolutionary theory predicts that if amino acid-altering genetic mutations occur in genes or domains of critical function and result in lower fitness, they will be purged from populations through purifying selection [46]. Conversely, selection acting on mutations in non-essential genes or domains is ''relaxed'', and thus accumulation of amino acid-altering mutations is more likely in such regions. Calculations of the Ka/Ks ratios in a set of amino acid sequences can thus be used to estimate an evolutionary history of both positive and purifying selection at each amino-acid site. Strong purifying selection in 9,471 core eukaryotic genes was previously reported in the genomes of several isolates of V. dahliae [40]. Whether sex-related genes in the V. dahliae genome are similarly conserved, compared to related sexual fungi, is currently unknown.
The goals of this study were to: 1) characterize the mating types of V. dahliae from a large collection of phytopathogenic isolates; 2) determine whether isolates of opposite mating types are present concurrently in the same habitat; 3) determine whether genetically identical multilocus microsatellite types contain both MAT idiomorphs; 3) determine if the complete genome sequence of V. dahliae strain Ls 17 contains orthologs of fungal sex-related genes; 4) test whether such genes are constitutively expressed in both mating types under laboratory conditions; and 5) estimate the extent of positive (relaxed) and purifying selection in a subset of sex-related genes in V. dahliae, relative to fungi with known sexual stages.

Results and Discussion
Molecular assays to identify Verticillium species, MAT type, and multilocus microsatellite types All isolates used in this study were identified as the phylogenetic species V. dahliae sensu stricto [35]. The frequency of MAT idiomorphs was extremely skewed towards an overabundance of MAT1-2 (Table S1). The MAT1-1 idiomorph was only observed in 1% (12/1120) of isolates characterized. The MAT1-1 isolates comprised eight isolates from commercial spinach seed lots from Washington State, USA, two isolates from a commercial artichoke field in California, and one isolate each from two commercial tomato field in CA (Table S1).
Complete multilocus microsatellite types were generated for 941 isolates; all 12 MAT1-1 isolates had different MLMTs, whereas 410 different MLMTs were observed for MAT1-2 isolates. Thus, after clone correction, 97% (410/422) isolates were MAT1-2. Nine of the MAT1-1 MLMTs were identical to MLMTs of one or more MAT1-2 isolates (Table 1). Of the nine MLMTs that comprised both mating types, three of them were found to have overlapping ecological niches. That is, they were collected at the same time from the same location and were isolated from the same substrate (artichoke, spinach seed, and tomato) ( Table S1). The presence of multilocus genotypes common to both mating types has been interpreted as evidence of sexual recombination [11,47]. However, this interpretation assumes no homoplasy, and assumes that isolates of opposite mating types did not acquire the same alleles at the thirteen loci independently through mutation.

Verticillium genome queries and ortholog searches
Out of 93 sex-related genes considered, 88 were found in the V. dahliae genome ( Table 2). The five genes not found in V. dahliae genome searches were the N. crassa accessions NCU09793, NCU04329, which are DNA helicase and repair proteins, respectively, and S. cerivisae accessions YIL072W, YGL033W, and YGL183C, which correspond to HOP1, HOP2 and MND1. Since no orthologs to HOP1, HOP2 or MND1 were found among any of the Sordariomycetes in the FUNGIPath database, including the sexual fungi Neurospora crassa, Podospora anserina, and Nectria haematococca, it is reasonable to speculate that these three genes are not required for a fully functional sexual cycle for taxa in this group.  Figure 1A, Figure 2A). However, using the MEC model, positive selection was detected in 12/20 genes investigated ( Figure 1B, Figure 2B, Figure S1). Likelihood ratio tests between the MEC and M8a models revealed that in all cases, the AIC score of the MEC model was lower than the M8a model. Using the MEC model, Verticillium dahliae MAT1-1-1 contained 12% of codons under positive selection and 34% of codons under strong purifying selection (Table 3); V. dahliae MAT1-2-1 contained 9% of codons under positive selection and 35% of codons under strong purifying selection (Table 4). Of the 21 codons under positive selection in V. dahliae MAT1-1-1, only 3 were within the highly conserved a domain ( Figure 1B); similarly of the 51 codons under positive selection in V. dahliae MAT1-2-1, only 1 was within the highly conserved HMG domain ( Figure 2B). When only sequences from sexual fungi were considered, MAT1-1-1 codons under positive and purifying selection ranged from 9-15% and 33-43% respectively (Table 3), whereas MAT1-2-1 codons under positive and purifying selection ranged from 12-22% and 21-30% respectively (Table 4). Thus, the extent and type of selection estimated for V. dahliae MAT genes were comparable to the estimates for MAT genes from sexual fungi. Interestingly, MAT1-1-1 from the putatively asexual P. fulva contained the highest relative numbers of codons under positive selection and the lowest under strong purifying selection (Table 3); however, P. fulva MAT1-2-1 Ka/Ks estimates were similar to sexual fungi (Table 4). Table 1. Ecological characteristics of multilocus microsatellite types that comprised isolates of both mating types.
MLMT Alleles for 13-locus MLMT a   In addition to the MAT genes, Ka/Ks patterns were investigated in 18 other sex-related genes ( Table 5, Table S2). The percentage of codons in V. dahliae genes under positive and strong purifying selection ranged from 0-5% and 35-62%, respectively. Six genes, KEX1, MEI3, RAD21, RAD54, STE23, and V-SNARE contained no codons under positive selection using either the M8 or MEC model (Table 5, Figure S1).

Expression of sex-related genes based on RT-PCR
RT-PCR using RNA from both mating types of V. dahliae successfully amplified all 10 sex-related genes investigated ( Figure 3). As expected, RNA from MAT1-1-1 and MAT1-2-1 only amplified from the strain that carried the respective MAT1-1 and MAT1-2 idiomorph (Figure 3). DNAse was used to treat extracted RNA, and no amplification was observed in reactions with reverse transcriptase omitted, indicating that DNA contamination was not present in the reactions (gels not shown). Since fungal isolates were cultured independently, it appears that V. dahliae expressed these genes during vegetative growth on PDA in the absence of a compatible culture of opposite mating type.

Conclusions
The overabundance of MAT1-2 in V. dahliae has been reported on multiple scales, from heavily sampled single agricultural fields to larger scales such as countries. This phenomenon may be partly explained by clonal expansion of certain successful, highly fit genotypes which do not require sexual reproduction to complete the disease cycle [23,31], unlike some other plant pathogens. Nevertheless, in two field sites in coastal California and in commercial spinach seed lots from WA, identical multilocus microsatellite types comprising both mating types were found, indicating at the very least, that both MAT1-1 and MAT1-2 cooccur in some niches currently.
The sample of V. dahliae characterized in the current study was biased toward virulent, phytopathogenic isolates collected from diseased plant tissue, because most were isolated from plants with visible wilt symptoms in agricultural settings. This raises the hypothesis that V. dahliae MAT1-2 may be associated with higher virulence on some, if not all hosts, which is a phenomenon that has been reported in other fungal systems [48][49][50][51]. Preliminary data on the virulence of isolates from both idiomorphs originally collected from tomato suggest that MAT1-2 isolates are significantly more virulent than MAT1-1 isolates (Subbarao, unpublished data). A more comprehensive analysis of the virulence of the two idiomorphs is required to confirm these results, however with more experiments and by investigating the mating-type structure in populations of non-pathogenic, endophytic V. dahliae [52,53].
Although the current study clearly documents patterns of purifying selective pressures in protein coding regions of the sexrelated V. dahliae genes investigated, it is possible that there are mutations in non-coding, regulatory regions of the genome that affect the level, timing or location of sex-related gene expression and therefore hinder the sexual cycle. Furthermore, it is possible that genes that were originally associated with sexual reproduction in ancestral populations have evolved new functions, and this is the reason they are being maintained under selection. Yet, it has been previously supposed that the presence of the majority, if not all, of the meiosis-specific genes in the genome of a microorganism is the ''strongest indicator'' that genes are maintained for meiosis and sex, even if it is rare [10]. The V. dahliae genome is clearly replete with orthologs to genes known for their roles in pathways associated with the sexual cycle. Further, the SELECTON analyses provide evidence that sex-related genes are not in the process of becoming pseudogenes.
The production of actual sexual structures in vitro currently remains a mystery in V. dahliae, possibly due to the lack of research into the growth medium content requirements, such as nutrient (i.e. carbon) content and pH, which are highly variable for sexual fungi in the Pezizomycotina [5,16]. Nevertheless, the genomic evidence presented in the current study, taken together with previous studies of population structure and recombination [23], is compelling and could be reasonably interpreted as evidence of an ancestral or rare sexual cycle in this predominantly asexual species.

Fungal culture maintenance and DNA extraction for MAT characterization
In this study, 1120 isolates of V. dahliae, collected from 10 different countries, were characterized for mating type (Table S1). No specific permissions were required for isolating Verticillium from any of the regions in the current study. The field isolations did not involve endangered or protected species. Importation of Verticillium cultures was performed under the appropriate USDA-APHIS permits (P526P-11-02218, P526P-11-02476, P526P-11-02806). Verticillium cultures were originally cultivated on semiselective NP-10 medium [54], and then single-conidium purified and transferred to potato dextrose agar (PDA). Cultures were stored long-term as spore suspensions in 25% glycerol at 220uC. Mycelia for DNA extraction were grown in 250 ml Erlenmeyer flasks containing 50 ml potato dextrose broth (PDB). Each flask was inoculated with a piece of PDA culture with an approximate surface area of 1 cm 2 . Mycelia from PDB were harvested after 10 days, washed with sterile distilled water, dried using paper towels, lyophilized, and ground to a fine powder using a highspeed mixer mill (Model MM301; Retsch Inc., Newtown, PA). Genomic DNA of each isolate was extracted using a FastDNA Kit (MP Biomedicals LLC, Solon, OH) following the manufacturer's instructions. A Nano Drop (Model ND-1000, Thermo Scientific Inc., Waltham, MA) was used to quantify DNA extractions, which were diluted to 10 ng/ml, and stored in a freezer at -20uC until needed for PCR assays.

Molecular assays to identify Verticillium species and mating type
All isolates used in this study were identified as V. dahliae using a Verticillium species-specific multiplex as previously described [35]. Mating types were determined for 1120 V. dahliae isolates PCR assay with the previously developed primers Alf3 (CGATCGCGA-TATCGGCAAGG), MAT11r (CAGTCAGATCCAACCTG-CTGGCC), HMG21f (CGGCCGCCCAATTCGTACATCC)  and MAT21r (CATGCCTTCCATGCCATTAGTAGCC). These primers amplify a ,600-bp fragment from MAT1-1-1 isolates and a ,300-bp fragment from MAT1-2-1 isolates, as previously described [29,35,37]. PCR assays to characterize mating types were performed in 25 ml reactions using GoTaq Green Mastermix (Promega, Madison, WI). All PCR assays in this study were performed in a PTC-100 Peltier Thermal cycler (MJ Research, Inc., Waterman, MA). For mating type multiplex PCR, the following thermal profile was used: 2 min initial denaturation at 94uC, 35 cycles of 10 sec at 94uC, 20 sec at 57uC, and 1 min at 72uC, followed by a final extension of 7 min at 72uC. PCR amplicons were stained with 5 ml SyberGold (Invitrogen Life Technologies, Carlsbad, CA), and aliquots were loaded in a 1.5% (wt/vol) agarose gel and run for 120 min at 75 V in 0.5% TBE buffer [55]. A 100bp DNA ladder (Invitrogen Life Technologies, Carlsbad, CA) was Table 3. Comparison of codons under positive (relaxed) and purifying selection in MAT1-1-1, in a variety of fungi in the subphylum Pezizomycotina using the MEC model. SELECTON results for the putatively asexual fungus V. dahliae were calculated by analyzing a MAT1-1-1 codon sequence alignment including sequences from all other fungi listed except P. fulva. Results for the ten species A. fumigatus -S. sclerotiorum were calculated using a codon alignment of only these ten species. 2 Homothallic fungus. 3 Results for the putatively asexual fungus P. fulva were calculated by analyzing a MAT1-2-1 codon sequence alignment including sequences from all other fungi listed except V. dahliae. doi:10.1371/journal.pone.0112145.t003 Table 4. Comparison of codons under positive (relaxed) and purifying selection in MAT1-2-1, in a variety of fungi in the subphylum Pezizomycotina using the MEC model.  1 SELECTON results for the putatively asexual fungus V. dahliae were calculated by analyzing a MAT1-2-1 codon sequence alignment including sequences from all other fungi listed except P. fulva. Results for the ten species C. globosum -T. reseei were calculated using a codon alignment of only these ten species. 2 Homothallic fungus. 3 Results for the putatively asexual fungus P. fulva were calculated by analyzing a MAT1-2-1 codon sequence alignment including sequences from all other fungi listed except V. dahliae. doi:10.1371/journal.pone.0112145.t004 included in each gel and a transilluminator (Ultra-Violet Products, Ltd., Upland, CA) was used to visualize PCR products.

Multilocus microsatellite genotyping
Thirteen previously developed microsatellite loci were used in this study: VD1, VD2, VD3, VD8, VD9, VD10, VD11, VD12, VD27, VD69, VD73, VD92 and VD97 [56] which were developed using the V. dahliae strain Ls 17 complete genome sequence [30,37]. For all microsatellite loci, PCR was performed in 20 ml total volumes containing 4 ml of sterile, distilled water, 2 ml of 10 ng/ml genomic DNA, 2 ml each of 10 mM reverse and forward primer, and 12.5 ml of GoTaq Green PCR mix (Promega Inc., Madison, WI). Published thermocycling parameters were used as previously described [30]. PCR amplicons labeled with up to four fluorophores FAM, HEX, ROX and TAMRA (Invitrogen, Carlsbad, CA) were pooled [57]. One ml of the pooled amplicons was then combined with Hi-Di formamide and 0.3 ml of LIZ-500 size standard and separated on an ABI 3100 capillary electrophoresis genetic analyzer (Applied Biosystems, Carlsbad, CA) at the University of California-Davis DNA Sequencing Facility, Davis CA. The peaks in were scored using the GeneMarker software (SoftGenetics, State College, PA).
To assess reliability of microsatellite allele calls using capillary electrophoresis [58], 192 microsatellite amplicons representative of all 13 loci were arbitrarily selected for DNA sequencing using unlabeled forward and reverse primers. Amplicons from V. dahliae strain Ls 17 were also generated and compared to the results reported from the same strain in previous studies [30,56]. Different amplicon sizes at each locus were considered unique. Alleles were compiled across loci into multilocus microsatellite types (MLMTs).

Verticillium genome queries and ortholog searches
The FUNGIpath ortholog database was queried using a panel of 93 genes that have been characterized for functions related to sexual reproduction in the fungal model systems Neurospora crassa, Saccharomyces cerevisiae, and Podospora anserina. The set of 93 genes comprised the two mating type genes MAT1-1-1 and MAT1-2-1, 81 previously described Neurospora crassa genes associated with meiosis [36,59,60] which were retrieved from the Neurospora Genome Database [61,62], four previously described Saccharomyces cerevisiae pheromone-related genes STE23, RCE1, AFC1, KEX1 [63], which were retrieved from the Saccharomyces Genome Database [24,64], and three Podospora anserina pheromone-related genes PRE2, RAM1, STE6 [63] which were retrieved from the Podospora anserina Genome Database [25,65]. Since V. dahliae is heterothallic and the sequenced strain contains only MAT1-2-1, a sequence of V. dahliae MAT1-1-1 was obtained through National Center of Bioinformatics (NCBI) GenBank, Accession AB505215 [27]. Finally, three additional Saccharomyces cerevisiae genes broadly associated with meiosis in eukaryotes (HOP1, HOP2, and MND1) [10] were queried against the FUNGIpath database. For FUNGIpath ortholog database searches, either gene accession ids. or amino acid sequences were used as input [41]. In this way, V. dahliae genes were verified as orthologous to genes from sexual fungi. Ortholog gene accession ids. from other fungi in the Pezizomycotina were noted and downloaded from the respective genome databases for subsequent analyses.

Primer design
After identifying orthologs to sex-related genes in the genome of V. dahliae, coding sequences of MAT genes and eight other genes associated with meiosis in other systems were arbitrarily chosen Fungal taxa and gene accessions used to estimate selective pressures in V. dahliae genes are provided in Table S2. Color-coded SELECTON results for each gene are provided in Figure S1. doi:10.1371/journal.pone.0112145.t005 Sex-Related Genes in Verticillium dahliae PLOS ONE | www.plosone.org and downloaded from the Broad Institute website [41]. Forward and reverse primers were designed to amplify ,500 to 1000-bp targets within coding sequences for 8 of the genes, whereas the previously described primers Alf3-MAT11r and HMG21f-MAT21r [29] were used to amplify MAT1-1-1 and MAT1-2-1, respectively (Table 6).

RNA Extraction and RT-PCR
The two V. dahliae isolates 58 (MAT1-1-1) and Ls 17 (MAT1-2-1) were grown on PDA. For each culture, after ten days, 3 ml of sterile distilled water was poured onto the culture surface and spread with a plate spreader. One ml of the resulting conidia and hyphal suspensions was transferred to a 47 mm nitrocellulose membrane (0.45 mm pore size; Whatman, Maidstone, England) overlaid on a PDA plate. Cultures were maintained in the dark at 25uC. After 10 days, the nitrocellulose membranes covered in fungal tissue were harvested with sterilized forceps and ground to a fine powder in liquid nitrogen using a mortar and pestle. Total RNA was extracted from 100 mg of the ground powder using TRIzol Reagent (Life Technologies, Carlsbad, CA) following the manufacturer's protocol. Total RNA extracts were treated with TURBO DNase (Life Technologies, Carlsbad, CA) following the manufacturer's protocol, in order to degrade genomic DNA.
Reverse-transcriptase PCR (RT-PCR) was performed using a SuperScript III OneStep RT-PCR system with Platinum Taq DNA polymerase (Life Technologies, Carlsbad, CA) following the manufacturer's protocol. For RT-PCR the following thermal profile was used: a cDNA synthesis cycle of 30 min at 55uC, an initial denaturation of 94uC for 2 minutes, 40 cycles of 94uC for 15 sec, 55uC for 30 sec, and 68uC for 1 min, followed by a final extension of 68uC for 5 min. Separate reactions including ten micromolar concentrations of forward and reverse primers for each and every locus described above were performed. For a positive control, RT-PCR was performed with the primers AaDTr (CTGGATGGAGACGTAGAAGGC) and Df (CTCGATGCT-CAAGCAGTACAT), which target ACT (VDAG_08445). Amplicons were visualized as above.
To verify the absence of genomic DNA in both of the RNA preparations, SuperScript III/RT Platinum Taq mix was omitted from PCR assays, and instead, two units of Platinum Taq DNA polymerase (Life Technologies, Carlsbad, CA) were used in reactions using the primers AaDTr and Df, in accordance with the manufacturer's instructions. To test the hypothesis that V. dahliae mating type and meiosisassociated genes are being maintained under strong purifying selection, ratios of non-synonymous (amino-acid altering) to synonymous (silent) substitutions in V. dahliae genes (relative to sexual fungi) were calculated through the SELECTON server [66,67]. All MAT genes used in this study were either identified directly through the FUNGIPath database, or were obtained through NCBI GenBank and verified as orthologs to either Additionally, ortholog search results from the FUNGIpath database from taxa within the subphylum Pezizomycotina were downloaded for 18 arbitrarily chosen, previously characterized genes associated with meiosis (Table S2), which represented a subset of the aforementioned 93 genes. Unaligned nucleotide sequences of V. dahliae orthologs and sequences identified through the FUNGIpath database [68] from at least nine other Pezizomycotina fungi were used as input to the SELECTON server, to provide the recommended number of sequences.
Codon alignments were generated by the SELECTON server and for each codon, the Ka/Ks ratio was estimated using a Bayesian approach. SELECTON results for each codon were reported on a scale of 1-7, with scores of one or two indicating positive selection, and scores of six or seven indicating strong purifying selection. For comparative purposes, two evolutionary models with positive selection enabled were used in the analyses, namely the M8 model [86,87] and the mechanistic-empirical combination (MEC) model [88]. SELECTON implements several codon models, each of which assumes different biological assumptions. The MEC model takes into account the differences between different amino-acid replacement probabilities. For analyses with the MEC model, eight categories for the distribution, a JTT empirical amino-acid matrix, and a high precision level were used. In cases where positive selection sites were detected using the MEC model, a likelihood ratio test between the results of the MEC model and the M8a (null) model was performed, by comparing Akaike Information Content (AIC) scores [89].
Estimates of selection in genes may be influenced by the choice of taxa used in the codon alignment. Therefore, for comparative purposes of the two mating type genes, Ka/Ks ratios within each of the other MAT1-1-1 and MAT1-2-1 sequences from other species were also calculated as above. For these analyses, the V. dahliae sequence was removed from the set of nucleotide sequences, and each sequence from every fungal species was considered independently as the query sequence. Thus, the Ka/Ks ratios of MAT loci were calculated for several sexual fungi, relative to the same set of taxa used to estimate selective pressures in V. dahliae mating type genes. Lastly, the Ka/Ks ratios in MAT1-1-1 and MAT1-2-1 from Passalora fulvum, a putatively asexual species, were calculated in comparison with the same set of sexual fungi used in the analyses of V. dahliae genes. Figure S1 Color-coded results of SELECTON analyses of 18 Verticillium dahliae sex-related genes, compared to sequences from nine different sexual fungi in the Pezizomycotina.

(PPTX)
Table S1 V. dahliae isolates used in this study along with country of origin, location, plant host, and mating types, as determined by PCR assays.