A second HD mating type sublocus of Flammulina velutipes is at least di-allelic and active: new primers for identification of HD-a and HD-b subloci

Maintenance and sequencing of F. velutipes strains

A total of 17 monokaryotic F. velutipes strains (Fv25-1, Fv-24, Fv25-3, Fv25-4, Fv27-3, Fv-27-1, Fv12-2, NJ6-21, NJ6-3, Fv10-4, Fv01-10, Fv10(Col)-4, Fv34-1, Fv34-18, L22, L11, and W23) were obtained from the Fujian Edible Fungi Germplasm Resource Collection Center of China and were maintained on 2% potato dextrose agar medium (PDA; 200 g/L potato; 20 g/L glucose; 20 g/L agar) at 25 °C. Genomic DNA (gDNA) was extracted from strains by using a 2% CTAB method (van Peer et al., 2011). Mycelium of each strain was grown separately on cellophane membranes and was used for gDNA extraction followed by sequencing. gDNA of seven strains (Fv25-1, Fv25-3, NJ6-21, NJ6-3, Fv34-1, Fv34-18, and L22) was sent for sequencing. Paired-end (90 bp) DNA libraries with 500 bp inserts were generated and sequenced (Illumina GAII platform, Zhejiang California International Nano Systems Institute, Hangzhou, China). Clean reads (six GB per genome, corresponding to ∼170 times coverage) were assembled into separate draft genomes by using SPAdes3.10.1 Linux with default parameters. Assembled draft genomes were converted for use in standalone blast using blast freeware (ncbi-blast-2.2.31, https://blast.ncbi.nlm.nih.gov/blast.cgi). Genome sequences of L11 and W23 strains were obtained from previous studies (Zeng et al., 2015; Wang et al., 2015).

Identification of Hd mating type genes and phylogenetic analysis of predicted proteins

Draft genome sequences of F. velutipes strains (Fv25-1, Fv25-3, NJ6-21, NJ6-3, Fv34-1, Fv34-18, and L22) were screened for Hd mating type genes with tBLASTn by using known F. velutipes HD proteins as query. Accession numbers of previously reported HD genes (van Peer et al., 2011; Wang et al., 2016) are mentioned in Table 1. Contigs/scaffolds of draft genomes that matched with HD proteins or MIP genes, were selected. Hd genes were annotated by following gene models of previously identified Hd genes and Hd proteins (van Peer et al., 2011; Wang et al., 2016). Fv_Hd_A_1-1 was predicted with the help of ZOOMlite software, and annotated in genomes of F. velutipes strains. Genes related to HD-a and HD-b subloci are mentioned in Table 2.

Nuclear localization signals (NLS), dimerization motifs (Di), and HD in HD mating type proteins were predicted by PSORT II Prediction (http://psort.hgc.jp/form2.html; Horton et al., 2007), COILS Server (http://embnet.vital-it.ch/software/COILS_form.html; Windows width = 14; Lupas, Van Dyke & Stock, 1991) and MOTIF Search (http://www.genome.jp/tools/motif/). Protein sequences of Hd genes were aligned using MUSCLE (Edgar, 2004). Phylogenetic analysis was performed with MEGA 6.0 by using default program settings and 1,000 Bootstrap value to construct a neighbor-joining tree (Tamura et al., 2013).

Crossing of strains with different HD subloci

A set of five F. velutipes strains (NJ6-3, NJ6-21, L11, Fv25-4, and Fv01-10) with compatible PR loci, different HD-a but similar HD-b subloci were crossed (Fig. 1). In each crossing, 0.8 cm diameter inoculum discs of PDA with actively growing mycelium (border of colony) were obtained from two compatible monokaryotic strains and placed at a distance of three cm apart from each other onto PDA plates, followed by incubation at 24 °C for 7 days. Similarly, compatibility of HD-b subloci was determined by crossing of six F. velutipes strains (L11, W23, Fv10-4, FV12-2, Fv-27-1, Fv-24) having different HD-b, but the same HD-a subloci (Fig. 2). Microscopic observations were done to examine the resulting colonies for the presence of clamp connections for confirmation of successful mating. For fruiting, dikaryotic mycelia were cultivated at 23 °C for 30 days in sterile plastic bags containing 52.5% cottonseed hulls, 25% wheat bran, 15% sawdust, 5% corn flour, 2% gypsum, and 0.5% ground limestone, with relative humidity of 61%. Induction of fruiting body formation to confirm the compatibility of crosses was performed at 16 °C with 90% relative humidity (Wang et al., 2016).

Design of primers for F. velutipes HD subloci

Primer sets for HD subloci were designed on stretches of conserved residues within and flanking the sequences of the HD-a and HD-b subloci. Conserved nucleotides were identified based on alignments of Hd genes and flanking regions with ClustalW (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The efficiency of each primer set was tested by PCR in 25 μL premix (Ex-Taq kit, Takara, China), with 35 cycles and an initial melting temperature of 94 °C for 5 min followed by 94 °C 1 min; 58/63 °C (according to TM of primer set used) 40 s; 72 °C (time according to length of fragment @1,000 bp/min). PCR products were sequenced at Sangon (Shanghai, China) to confirm amplification of the correct DNA fragments.

Identification of HD-a and HD-b subloci in F. velutipes

Flammulina velutipes strains Fv25-1, Fv25-3, Fv25-4, and Fv27-3 had been indicated to contain a HD-b sublocus but no FvHd-a2-1 gene, leaving the existence and composition of additional HD subloci in these strains undetermined (Wang et al., 2016). Monokaryons Fv25-1 and Fv25-3 that had been obtained from the same dikaryon (protoclones of dikaryotic stain Fv25) were selected for further analysis. In order to explore more about the HD-a sublocus, a collection of mating compatible monokaryons (protoclone pairs from different dikaryotic strains) was also screened by PCR to confirm the presence or absence of the conserved FvHd-a2-1 gene.

Monokaryon pairs with and without a FvHd-a2-1 gene were also obtained from dikaryon NJ6 (monokaryons NJ6-3 and NJ6-21) and Fv34 (monokaryon 34-1 and Fv34-18). Monokaryons NJ6-3 and Fv34-18 contained the FvHd-a2-1 gene while monokaryons NJ6-21 and Fv34-1 did not. gDNA of Fv25-1, Fv25-3, Fv34-1, Fv34-18, NJ6-3, and NJ6-21 was isolated and sent for whole genome sequencing. Cleaned, 90 bp paired-end reads (six GB, or ∼170× coverage/genome) were de novo assembled to generate six draft genomes. Scaffolds that contained HD subloci were selected from the respective draft genomes based on tBLASTn and BLASTn searches with known FvHD proteins, and the MIP gene (van Peer et al., 2011; Wang et al., 2016; Table 1). For all six monokaryons, the HD-b and an additional HD sublocus were identified (nine scaffolds in total, Table 2). Corresponding FvHd and MIP genes were annotated by using previously identified FvHd and MIP gene and their respective protein sequences (see also Table 1). In contrast to previous PCR based analyses, the draft genome of strain Fv25-1 revealed the presence of an HD-a sublocus with the conserved FvHd-a2-1 gene (Node-50, Table 2). Compared to other HD-a subloci with a FvHd-a2-1 gene, the HD-a sublocus of Fv25-1 differed in sequence through a 4,588 bp insert of noncoding and repeat containing DNA between the FvHd-a2-1 and the MIP gene (Fig. 3). Most probably this ∼4.6 kb insert prevented efficient amplification of FvHd-a2-1 using primers described in Wang et al. (2016), resulting in false negative test results for presence of a HD-a sublocus in this strain. No other strains with a 4.6 kb insert at the HD-a sublocus were found. Specific primers (Table 3) were designed for the region between MIP and FvHd-a2-1 or the FvHd-a1-1 gene. Amplification and sequencing of these regions between MIP and the Hd genes (FvHd-a2-1 or the FvHd-a1-1) showed that strain Fv25-1 indeed contained an extra 4,900 bp region (Fig. S5).The HD-a subloci of strain NJ6-3 and Fv34-18 were confirmed to contain the conserved FvHd-a2-1 gene in accordance with the PCR based screens, and were found to be flanked by a MIP gene at similar distances (respectively, 203 or 205 bp). Strains Fv25-3, NJ6-21, and Fv34-1 also contained an HD-a sublocus, yet without the conserved FvHd-a2-1 gene. Instead, all three strains contained the same, newly identified FvHd1 gene; FvHd-a1-1. The FvHd-a1-1 gene did not match with known FvHd1 gene models. This gene contained four predicted introns of which two were alternatives for intron 1. Protein predictions using the two alternate intron 1 varieties resulted in different proteins with varying amino acid sequences (Fig. S1). To obtain a correct gene model and predicted protein, total RNA was extracted from strain NJ6, converted to cDNA, amplified with primers HD-a FW and HD-a RV (Table 3), cloned and sequenced. The sequenced FvHd-a1-1 cDNA confirmed the presence of 3 introns and 4 exons, encoding a 510 amino acid protein (Fig. S1). The FvHd-a1-1 gene and its predicted protein, like those of FvHd-a2-1 on HD-a in other strains, was highly conserved between strains NJ6-21, Fv34-1, and Fv25-3 (Fig. 4). The FvHd-a1-1 was located at ∼130 bp distance from the MIP gene, in an opposite orientation of FvHd-a2-1 (Fig. 3). The novel FvHd-a1-1 gene was uploaded in GenBank with accession number KU852595.1 (Table 1). Blast and alignment results showed that protein FvHD_A_1-1 contained a homeobox transcription factor KN domain (a DNA-binding domain), that is, conserved from fungi to human and plants (Fig. 4). No new genes were identified for the HD-b subloci of the eight sequenced strains, and each contained FvHd1/FvHd2 gene pairs in the same combinations and orientations as reported before in strains W23, L11, and Fv10-4 (Table 1 in Wang et al., 2016).

Nuclear localization signal, Di, and HD analyses showed that FvHD_A_1-1 contains three predicted NLS however, no Di was found in this protein (Fig. S2). FvHD_A_1-1 is clearly different from the FvHd1 on the HD-b subloci (Fig. S2).

Functional analysis of the HD-a sublocus in mating

Interestingly, monokaryotic strains NJ6-3 and NJ6-21 contained the same set of FvHd genes (FvHd-b1-5/FvHd-b2-5) at their respective HD-b subloci (Table 2). Strains NJ6-3 and NJ6-21 have been obtained from dikaryon NJ6 that shows normal clamps and fruiting body development, and the monokaryons display normal HD mating compatibility in crosses with other strains.

In other crosses (NJ6-3 × Fv01-10) and (L11 × Fv25-4) between monokaryotic strains containing identical HD-b subloci but different HD-a subloci resulting dikaryons produced regular clamp connections and fruiting bodies (Table 4; Figs. 1 and 5). It is likely that the conserved FvHd-a2-1 gene at the HD-a sublocus is therefore compatible with the conserved FvHd-a1-1 gene. Crosses between strains with identical HD-a subloci and different HD-b subloci (HD-a ═; HD-b ≠) clamp connections were formed that demonstrated the activity of HD-b sublocus in mating in accordance with previous studies (Fig. 2).

Apparently, HD-a locus contains at least two “alleles,” one with a single unpaired FvHd2 gene, and one with a single unpaired FvHd1 gene. Relative grouping of the HD1 and HD2 proteins of the HD-a and HD-b subloci revealed two distinct clades, one containing the HD1 and HD2 proteins of the HD-a sublocus, and one containing the HD1 and HD2 proteins of the HD-b sublocus (Fig. 6). Clustering of proteins HD_A_1-1 and HD_A_2-1 in a separate clade from the HD-b sublocus, rather than clustering of all HD1 and HD2 proteins, further indicates the presence of two sets of FvHd1 and FvHd2 genes, belonging to two separate subloci.

Design of primers for the HD subloci of F. velutipes

To determine the HD mating type of a given strain of F. velutipes, the two distant subloci, HD-a and HD-b, will need to be identified. To provide a quick identification method, three specific primers sets were tested for HD-a, and two sets for HD-b (Table 3). Primer set “HD-a-locus-F/HD-a-locus-R” amplified the whole HD-a sublocus including a part of the MIP gene, and can be used for cloning of new HD-a subloci (Fig. S4) except for strains with a HD-a sublocus similar to Fv25-1. The two other HD-a primer sets were designed on conserved sequences within the FvHd-a1-1 and FvHd-a2-1 gene, respectively, to easily distinguish the different HD-a subloci (Fig. 4; Fig. S3). Similarly, two sets of primers were designed on the HD-b sublocus, one for amplification of the entire HD-b sublocus, and one on highly conserved sequences within the two genes at this sublocus (amplifying the dissimilar, mating type specific parts of the FvHd_B_1 genes and the FvHd_B_2 genes). Amplicons can easily be sequenced in a single run and allow identification of the two different FvHd genes on HD-b (Fig. S4), while amplicons covering the entire HD-b can be used for cloning of the complete subloci. To evaluate the efficiency of the different primer sets in random F. velutipes strains, five more monokaryons (L22, Fv01-1, Fv01-3, Fv01-10(Co1), and Fv01-11) and six dikaryons (Fv18, Fv0077 Fv50599, Fv25, FV26, and Fv29) were tested in addition to the strains on which they had been designed (Table 4). All five primer sets showed reproducible results for the different F. velutipes strains. Either FvHd-a2-1 or FvHd-a1-1 was amplified for the HD-a sublocus, also even in case for the case of Fv25-1 with the large insert between MIP and FvHd-a-2-1. Complete HD-a subloci were obtained for all strains but Fv25-1 (Fig. S4). For HD-b, both primer sets generated the expected products for all tested strains. Sequencing results from amplicons of the HD-a sublocus showed that Fv01-1, Fv01-3, Fv01-10(Co1), and Fv01-11 strains contained either the FvHd-a 2-1 or the FvHd-a1-1 gene, together with the MIP gene.