https://orcid.org/0000-0003-1061-3639
Abstract
Mycorrhizae are one of the most fundamental symbioses between plants and fungi, with ectomycorrhizae being the most widespread in boreal forest ecosystems. Ectomycorrhizal fungi are hypothesized to have evolved convergently from saprotrophic ancestors in several fungal clades, especially members of the subdivision Agaricomycotina. Studies on fungal genomes have identified several typical characteristics of mycorrhizal fungi, such as genome size expansion and decreases in plant cell-wall degrading enzymes (PCWDEs). However, genomic changes concerning the evolutionary transition to the ectomycorrhizal lifestyle are largely unknown. In this study, we sequenced the genome of Lyophyllum shimeji, an ectomycorrhizal fungus that is phylogenetically related to saprotrophic species and retains some saprotroph-like traits. We found that the genome of Ly. shimeji strain AT787 lacks both incremental increases in genome size and reduced numbers of PCWDEs. Our findings suggest that the previously reported common genomic traits of mycorrhizal fungi are not essential for the ectomycorrhizal lifestyle, but are a result of abolishing saprotrophic activity. Since Ly. shimeji is commercially consumed as an edible mushroom, the newly available genomic information may also impact research designed to enhance the cultivation of this mushroom.
1. Introduction
Mycorrhizal symbiosis is one of the most abundant types of symbioses between fungi and plants, involving more than one hundred fungal genera and 80% of land plant species.1,2 This type of symbiosis is basically mutualistic in that fungi help plants take up soil-derived nutrients, such as phosphate and nitrate, in exchange for photoassimilates.3 Mycorrhizae are classified into several types by their morphology and function. Of them, ectomycorrhizae have the most impact on boreal forest biomass because dominant trees in the boreal forest harbour these types of mycorrhizal fungi.4,5
Ectomycorrhizal plants appear in several taxa, including the representative families Fagaceae and Pinaceae.2 These mycorrhizae are proposed to have changed their mycorrhizal style from ancestral arbuscular mycorrhizae and have succeeded in becoming adapted to boreal regions. In contrast, ectomycorrhizal fungi are thought to have evolved from saprotrophic ancestors.6 Accounting for their taxonomic appearance, the transition from the saprotrophic to the ectomycorrhizal lifestyle is proposed to have independently evolved dozens of times throughout the history of Basidiomycota and Ascomycota.1,6
Besides their ecological importance as forest components, many ectomycorrhizal fungi also impact our lives as foods. Many favoured mushrooms, such as truffles, boletes, and matsutake, are mycorrhizal.7 However, unlike saprotrophic species, most mycorrhizal mushrooms have difficulty artificially producing basidiocarps, probably because of their host-dependent lifestyle.7
Lyophyllum shimeji, an agaricoid fungus known as hon-shimeji or just shimeji in the older literature, has been one of the most favoured mushrooms in Japan for a long time.8,9Ly. shimeji is known to be an ectomycorrhizal fungus associated with plants in the Fagaceae and Pinaceae families.10,11 However, unlike many other mycorrhizal fungi, Ly. shimeji can metabolize botanical polysaccharides, such as starch.12 Moreover, the fruiting body of Ly. shimeji can be induced on a medium containing barley and sawdust without a host plant,13 and it is now commercially produced axenically. Thus, Ly. shimeji has physiological similarities to saprotrophic fungi though it has a mycorrhizal lifestyle. Also, Ly. shimeji has a systematically close saprotrophic relative, Ly. decastes.14,15 Morphological traits of Ly. shimeji, Ly. decastes and several other species highly resemble each other. Descriptions of the mycorrhizal ability of these relatives are also confusing. For example, Ly. decastes is often described as a mycorrhizal fungus,16,17 whereas other studies have reported its inability to form mycorrhizae,11,18 possibly because of improper species identification.19 Therefore, Ly. shimeji and related species include mycorrhizal and non-mycorrhizal fungi within a very close genetic distance. Since ancestral species of Lyophyllaceae are proposed to be saprotrophic,17Ly. shimeji may be one of the last groups of ectomycorrhizal fungi that has evolved from a saprotrophic ancestor.
Since the genome of Laccaria bicolor was sequenced as the first mycorrhizal genome in 2008,20 many other genomes of mycorrhizal fungi have been sequenced.21–26 Comparative studies between mycorrhizal and saprotrophic fungi have revealed that mycorrhizal fungi have several characteristic genomic features, such as genome size increments accompanied by the proliferation of transposable elements (TEs) and reductions in genes encoding plant cell-wall degrading enzymes (PCWDEs).23–26 These convergent genomic traits suggest that such features are adaptive to the mycorrhizal lifestyle. However, to what degree these features are essential for the mycorrhizal lifestyle is largely unknown. In this study, we sequenced the genome of Ly. shimeji as a representative of a putative primitive stage of mycorrhizal fungi in order to survey the initial genomic status in the ecological transition. The genome information of Ly. shimeji is also expected to be useful for research and cultivation of this praised mushroom.
2. Materials and methods
2.1. Fungal material
The strain of Ly. shimeji AT787 was initially isolated from Nagano and maintained at Shinshu University.27 This strain was confirmed to be mycorrhizal and has been used for research on mycorrhizal formation. Hyphae were cultured on a cellophane membrane placed on MMN (Modified Melin-Norkrans) medium28 and MMN medium without malt and sugar (hereafter referred to as MMN salt-only medium). After an 1-week incubation at 28°C, hyphae were peeled from the cellophane membrane and used for nucleic acid extraction.
2.2. Extraction of genomic DNA and RNA
Genomic DNA was isolated from hyphae cultured on MMN medium using the CTAB (cetyl trimethylammonium bromide) method29 with some modifications as described below. Collected hyphae were frozen with liquid nitrogen, ground with a pestle, and dispersed in extraction buffer (100 mM Tris–HCl pH 8.0, 20 mM EDTA, 0.75% sarkosyl, 0.1% PVP, 0.75% CTAB, 0.13 M sorbitol, 0.75 M NaCl, and 0.1 mg ml−1 proteinase K). After incubation at 37°C for 30 min, the aqueous phase was centrifuged at 12,000 rpm (rotations per minute) for 10 min at 4°C, and the pellet was discarded. An equal volume of phenol/chloroform (1:1, Vol:Vol) was added and centrifuged (12,000 rpm, 5 min at 4°C). The aqueous phase was collected, and an equal volume of chloroform was added to the sample and centrifuged (12,000 rpm, 5 min at 4°C). The aqueous phase was collected again, and 1:10 Vol of sodium acetate and 0.7 Vol of isopropanol were added, mixed, and centrifuged (12,000 rpm, 20 min at 4°C). The resulting pellet was washed twice with 70% EtOH and resuspended in TE buffer. Extracted DNA was purified with Genomic-tips (Qiagen, Netherlands). RNAs were extracted from hyphal samples cultured on MMN medium and MMN salt-only medium using an RNeasy Plant Mini Kit (Qiagen, Netherlands) with buffer RLC included in the kit.
2.3. Library construction and sequencing
PacBio libraries were constructed following the manufacturer’s protocol (Pacific Bioscience, USA) and sequenced with a PacBio RS II sequencer. Paired-end Illumina genomic libraries with insert sizes of 250 bp and 700 bp were constructed with a TruSeq DNA Sample Preparation v2 Kit (Illumina, USA) and were sequenced with a HiSeq 2500 System (Illumina, USA) for 151 bp from both ends. RNA-seq libraries were constructed with a TruSeq Stranded mRNA Library Prep Kit (Illumina, USA) and sequenced with a HiSeq 1500 System (Illumina, USA) for 126 bp from both ends.
2.4. Genome assembly
The genome size of Ly. shimeji was estimated with GenomeScope 2.030 using the Illumina reads. PacBio reads were primarily assembled with Falcon v2017.11.02-16.04-py2.7-ucs and Falcon-Unzip v0.4.0.31 Two allelic draft genomes were integrated into a single haploid-type genome with HaploMerger2.32 Repetitive elements were predicted using RepeatMasker v4.0.633 after modelling the repetitive sequences with RepeatModeler v1.0.833 with default settings. The completeness of the draft genome was assessed with BUSCO ver.4.1.4, using the database fungi_odb10.34
2.5. Gene prediction and annotation
Genes were predicted with mapped transcripts. Adapters and low-quality sequences were removed from Illumina RNA-seq reads with Cutadapt 2.5.35 Transcripts were assembled with Trinity v2.11.0.36 Genes were predicted with BRAKER v2.1.037 with the combination of GeneMark-ES v4.5738 and Augustus v3.3.1.39 The completeness of the gene set was assessed with BUSCO v4.1.4,34 using the database fungi_odb10 (n = 758). Genes were functionally annotated with eggNOG-mapper v2.40
2.6. Reference genome datasets
Reference genomes were downloaded from JGI Mycocosm.41,42 Assembly genomes of Agaricus bisporus var bisporus (H97) v2.0,43Amanita muscaria Koide v1.0,23Amanita thiersii Skay4041 v1.0,44Aspergillus oryzae RIB40,45Auricularia subgrabra v2.0,46Boletus edulis Prilba v1.0,25Cenococcum geophilum 1.58 v2.0,47Coprinopsis cinerea,48Galerina marginata v1.0,49Hebeloma cylindrosporum h7 v2.0,23Laccaria bicolor v2.0,20Lentinula edodes W1-26 v1.0,50Neurospora crassa OR74A v2.0,51Pisolithus tinctorius v1.0,23Pleurotus ostraeus PC15 v2.0,49Rhizophagus clarus HR1,52Rhizopus delemer 99-880,53Saccharomyces cerevisiae S288C,54Schizosaccharomyces pombe,55Tricholoma matsutake 945 v3.0,25 and Tuber melanosporum Mel28 v1.221 were used as references. Assembly genome and raw reads of Ly. shimeji JCM30591 were downloaded from the NCBI database.56
2.7. Comparative analyses
PCWDEs were searched with dbCAN2 web server.57,58 Only genes that were predicted as CAZymes by two or three of three methods in dbCAN2 were regarded as PCWDEs. Secretory proteins were predicted with SignalP v4.1.59 Small proteins were selected with Seqkit v0.5.2.60 Ortholog analyses were conducted with OrthoFinder v2.3.1161 with default parameters. Three-hundred fifty-one single-copy orthologs were aligned with MAFFT v7.49062 with default parameters to generate a phylogenetic tree constructed with IQ-tree v2.1.363 with 1,000 bootstrapping replicates and 1,000 SH-aLRT (Shimodaira-Hasegawa approximate likelihood ratio test) replicates. ‘Q.yeast+F+R6’ model was chosen by the ModelFinder in IQ-tree. PCA and correlation efficiency were computed, Welch’s t-test was executed, and figure drawings were created using R 4.0.3.64
3. Results and discussion
3.1. Genome sequencing and assembly
We sequenced the entire genome of Ly. shimeji AT787 with a short-read sequencer, Illumina HiSeq2500, and a long-read sequencer, PacBio RSII. A total of 13.6 Gbp of Illumina reads and 5.81 Gbp of PacBio reads were obtained. From the k-mer analysis of the Illumina short reads, the genome size was estimated to be approximately 30.5 Mbp with 2.3% heterozygosity (Fig. S1). We assembled a draft genome with Falcon-Unzip, which generates a ‘primary contig’ as a basal allele and a ‘haplotig’ as the opposite allele. A total of 36.3 Mbp of primary contigs and 23.7 Mbp of haplotigs were obtained. Since the primary contigs may contain both alleles partially, we merged primary contigs and haplotigs into single alleles, and 31.8 Mbp of the draft genome with an N50 of 1.61 Mbp was generated (Table 1). BUSCO assessment found 96.9% of the conserved single-copy genes were complete genes in the draft genome. From this draft genome, 12,904 genes were predicted (Table 1). The completeness of the BUSCO predicted proteome was estimated to be 97.8%, including 97.0% being single-copy sequences (Table 1). The contiguity of our assembly and BUSCO completeness of the genome are above or comparable to average for representative reference genomes (Table S1).
Basic information about the Ly. shimeji AT787 assembled genome and predicted genes
| Genome assembly (purged) | Predicted genes | ||
|---|---|---|---|
| Total bases | 31,834,823 bp | Number of genes | 12,904 |
| No. of scaffolds | 60 | Average of CDS length | 1375.9 bp |
| N50 | 1,610,882 bp | Average of protein length | 457.6 aa |
| L50 | 7 | ||
| GC% | 53.92% | ||
| BUSCO benchmarks of genome assembly | BUSCO benchmarks of predicted genes | ||
| Complete | 96.9 % | Complete | 97.8% |
| Complete single copy | 96.2 % | Complete single copy | 97.0% |
| Complete duplicated | 0.7 % | Complete duplicated | 0.8% |
| Fragmented | 0.3 % | Fragmented | 1.2% |
| Missing | 2.8 % | Missing | 1.0% |
| Genome assembly (purged) | Predicted genes | ||
|---|---|---|---|
| Total bases | 31,834,823 bp | Number of genes | 12,904 |
| No. of scaffolds | 60 | Average of CDS length | 1375.9 bp |
| N50 | 1,610,882 bp | Average of protein length | 457.6 aa |
| L50 | 7 | ||
| GC% | 53.92% | ||
| BUSCO benchmarks of genome assembly | BUSCO benchmarks of predicted genes | ||
| Complete | 96.9 % | Complete | 97.8% |
| Complete single copy | 96.2 % | Complete single copy | 97.0% |
| Complete duplicated | 0.7 % | Complete duplicated | 0.8% |
| Fragmented | 0.3 % | Fragmented | 1.2% |
| Missing | 2.8 % | Missing | 1.0% |
Basic information about the Ly. shimeji AT787 assembled genome and predicted genes
| Genome assembly (purged) | Predicted genes | ||
|---|---|---|---|
| Total bases | 31,834,823 bp | Number of genes | 12,904 |
| No. of scaffolds | 60 | Average of CDS length | 1375.9 bp |
| N50 | 1,610,882 bp | Average of protein length | 457.6 aa |
| L50 | 7 | ||
| GC% | 53.92% | ||
| BUSCO benchmarks of genome assembly | BUSCO benchmarks of predicted genes | ||
| Complete | 96.9 % | Complete | 97.8% |
| Complete single copy | 96.2 % | Complete single copy | 97.0% |
| Complete duplicated | 0.7 % | Complete duplicated | 0.8% |
| Fragmented | 0.3 % | Fragmented | 1.2% |
| Missing | 2.8 % | Missing | 1.0% |
| Genome assembly (purged) | Predicted genes | ||
|---|---|---|---|
| Total bases | 31,834,823 bp | Number of genes | 12,904 |
| No. of scaffolds | 60 | Average of CDS length | 1375.9 bp |
| N50 | 1,610,882 bp | Average of protein length | 457.6 aa |
| L50 | 7 | ||
| GC% | 53.92% | ||
| BUSCO benchmarks of genome assembly | BUSCO benchmarks of predicted genes | ||
| Complete | 96.9 % | Complete | 97.8% |
| Complete single copy | 96.2 % | Complete single copy | 97.0% |
| Complete duplicated | 0.7 % | Complete duplicated | 0.8% |
| Fragmented | 0.3 % | Fragmented | 1.2% |
| Missing | 2.8 % | Missing | 1.0% |
3.2. Comparison of genome size and content with other fungal genomes
Since mycorrhizal fungi tend to have large genomes with manyTEs, we compared the genome size, gene number, and repeat content, including TEs, of Ly. shimeji AT787 with the data registered in JGI Mycocosm (accessed Oct. 2020)41,42 and the data of another strain of Ly. shimeji, JCM30591 (NCBI GenBank GCA_001950515). The plot of genome size and gene number showed that genomes of mycorrhizal fungi are generally larger than those of other fungi while gene numbers are not; however, the genome size of Ly. shimeji AT787 is very small among mycorrhizal fungi yet the gene number is relatively similar to other mycorrhizal fungi, while that of Ly. shimeji JCM30591 is not (Fig. 1a). The low BUSCO duplication ratio (Table S1) suggest neither are mixture of multiple genomes or polyploid and genome size estimates from raw short reads (57.7 Mb for JCM30591, DRR037069) suggests the genome size of strain JCM30591 and AT787 differ by 1.9-fold. The difference in the genome size implies that JCM30591 and AT787 are different species in the so-called Lyophyllum decastes species complex; however, we cannot state which strain is strict ‘Ly. shimeji’ because the holotype sequence is unavailable. Since many strains including AT787 have been treated as Ly. shimeji in previous studies,27 we regarded both strains as Ly. shimeji in this paper. The repeat content of Ly. shimeji AT787 is 6.84% including 6.14% TEs (Fig. 1b). Most of the TEs in the Ly. shimeji AT787 genome were Gypsy-type LTR elements (Fig. 1b). No SINEs and LINEs were detected while certain amount of LINEs was found in the Ly. shimeji JCM30591 genome (Fig. 1b and c). When compared to representative fungal genomes, the repeat content in Ly. shimeji AT787 was smaller than that in most, if not all, ectomycorrhizal genomes (Fig. 1c). These results indicated that neither genome size enlargement nor TE proliferation occurred in the genome of Ly. shimeji AT787, unlike typical ectomycorrhizal fungi. Therefore, a large genome size is not an essential trait for the mycorrhizal lifestyle but, rather, a result of mycorrhizal history. Generally, TEs are not adaptive to organisms but provide disadvantages for cell proliferation.65 Mycorrhizal fungi tend to have a slower cell cycle than saprotrophic species; thus, mycorrhizal fungi may have experienced a weaker selection pressure to eliminate such junk sequences. We hypothesize that such a relaxed selection pressure might be the driving force for large genomes in mycorrhizal species.
Comparison of genome size and gene number among sequenced fungi Genome size and components of Ly. shimeji AT787 compared with other fungi. (A) Genome size and gene numbers of fungi. Red, yellow, and blue markers correspond to Ly. shimeji AT787, other mycorrhizal fungi, and non-mycorrhizal fungi, respectively. Circles, squares, and triangles correspond to Basidiomycota, Ascomycota, and other taxa, respectively. (B) Repetitive sequences, including TEs in Ly. shimeji AT787. C: Genome size and repeat content of fungi. The size of circles corresponds to the absolute number of repeated sequences. The fungal genomes registered in JGI Mycocosm and the genome of Ly. shimeji JCM30591 registered in NCBI were used as references.
3.3. PCWDE gene content
We next compared the repertoire of PCWDE genes in the Ly. shimeji AT787 genome compared with other fungal genomes because the number of PCWDEs is known to be lower in mycorrhizal genomes. We found 384 Carbohydrate-Active enZymes (CAZymes), including 178 secretory CAZymes, in Ly. shimeji AT787 (Fig. 2a). Genomes of reference species showed that saprotrophic fungi, except yeasts, harbour more CAZymes, especially secretory enzymes, than mycorrhizal fungi (Fig. 2a and b), as reported in previous studies.23,25 However, the number of CAZymes in Ly. shimeji AT787 was much higher than in reference mycorrhizal fungi and even some saprotrophic fungi (Fig. 2) despite its mycorrhizal ecology. Interestingly, the genome of Ly. shimeji JCM30591 also harbours high amount of CAZymes while genome size increment was observed like other mycorrhizal species (Fig. 2). Thus, depletion of PCWDEs is not an essential trait of ectomycorrhizal fungi. Therefore, unlike other mycorrhizal fungi, Ly. shimeji may have the capacity to decay plant-derived materials like saprotrophic species, a finding consistent with its ability to live on plant-derived substances. Mycorrhizal fungi have been hypothesized to avoid hurting plant cell walls to prevent the activation of plant defence systems.47 However, it is more likely that mycorrhizal fungi just abandoned the use of PCWDEs after their ecological change because mycorrhizal fungi can directly derive monosaccharides from their host plants, as also hypothesized by Miyauchi et al.25
Comparison of PCWDEs with other representative fungi numbers of PCWDEs in the genome of Ly. shimeji AT787 in comparison with other representative fungi that have been reported. (A) Numbers of CAZymes found in each genome. The phylogenetic tree was constructed with jointed sequences of 351 single-copy orthologs. GH, GT, PL, CE, AA, and CBM correspond to glycoside hydrolases, glycosyltransferases, polysaccharide lyases, carbohydrate esterases, auxiliary activities, and cellulose-binding motifs, respectively. Strains of Ly. shimeji are bordered in red. Asterisks after the species name indicate yeasts with a small number of total genes. Open and closed circles after the species name indicate mycorrhizal and saprotrophic references, respectively. (B) Comparison of numbers between Ly. shimeji strains, other mycorrhizal species, and saprotrophic species. Plots of Ly. shimeji AT787 are shown in red. Asterisks indicate significant differences (an asterisk for Welch’s T-test P < 0.05 and two asterisks for Welch’s T-test P < 0.01).
3.4. Small-secreted proteins
Mycorrhizal fungi harbour many small-secreted proteins (SSPs).23,66 Since SSPs include mycorrhiza-induced genes, and some of them are essential for the establishment of symbiosis, SSPs are thought to communicate with or modify mycorrhizal host plants.20,23,67,68 We identified the SSPs from the gene repertoire of Ly. Shimeji AT787 and compared them with those from reference genomes. However, when we selected proteins with predicted secretory signal peptides and sizes of ≦300 aa, there was no apparent difference in gene number between mycorrhizal and saprotrophic fungi even in reference species (Fig. S2). We next checked the functional annotation of predicted SSPs in Ly. Shimeji AT787 (Table S2). Some of the predicted SSPs were annotated as hydrophobins, which include proteins having a function in mycorrhizal symbiosis.69,70 Some annotated SSPs are secretory CAZymes (Table S2). The lack of apparent difference in the total number of SSPs between mycorrhizal and saprotrophic species may be because SSPs include effectors that are considered to be abundant in mycorrhizal fungi and CAZymes that tend to be reduced in mycorrhizal fungi. Approximately 60% of the predicted SSPs were functionally unannotated (Table S2). Since some mycorrhiza-related SSPs are lineage-specific,23 some of these unannotated SSPs could function in mycorrhizal symbiosis.
3.5. Ortholog analysis
We compared the gene repertoire of Ly. Shimeji AT787 with several other fungal genomes. When compared with two mycorrhizal fungi (La. bicolor and H. cylindrosporum) and two saprotrophic fungi (Co. cinerea and G. marginata) in the Agaricales, Ly. shimeji AT787 and the two saprotrophic species share 126 orthologs that are not shared with the two mycorrhizal species (Fig. 3a, bordered in red). In contrast, Ly. shimeji AT787 and the other two mycorrhizal species share only 32 orthologs that are not shared with saprotrophic species (Fig. 3a, bordered in white). Thus, Ly. shimeji AT787 shares more orthologs with saprotrophic species than with other mycorrhizal species. Genes shared by Ly. shimeji AT787 and two saprotrophic species include 20 glycoside hydrolases as the most enriched common orthologs (Table S3). This result confirmed the absence of a mycorrhiza-specific reduction in Ly. shimeji AT787 CAZYmes, found by CAZYme-targeted analysis as described above, by non-targeted orthologous analysis. In contrast, most orthologs shared by only Ly. Shimeji AT787 and two mycorrhizal fungi were functionally unannotated and did not show enrichment in specific functions (Table S4).
Shared orthologs with other fungi in the Agaricales. Venn diagrams of shared orthologs. (A) Orthologs shared with Ly. shimeji AT787, two mycorrhizal reference species (La. bicolor and H. cylindrosporum), and two saprotrophic reference species (Co. cinerea and G. marginata). Red, yellow, and blue ellipses correspond to Ly. shimeji AT787, other mycorrhizal fungi, and saprotrophic fungi, respectively. Regions with red or white borders correspond to orthologs that Ly. shimeji AT787 shares only with saprotrophic or mycorrhizal species, respectively. Open and closed circles after the species name indicate mycorrhizal and saprotrophic references, respectively. (B) Orthologs shared with Ly. shimeji AT787, two consumed reference species (Ag. bisporus and Le. edodes), two non-consumed reference species (Co. cinerea and La. bicolor). Red, orange, and green colours correspond to Ly. shimeji AT787, other consumed fungi, and non-consumed fungi, respectively. The region with a red border corresponds to orthologs that Ly. shimeji AT787 shares only with widely consumed species. Open and closed stars after the species name indicate consumed and non-consumed references, respectively.
We next searched for other genes that differ between mycorrhizal and saprotrophic fungi. We conducted PCA analysis of 19 dikaryon fungi, including 9 saprotrophic species, 9 mycorrhizal reference species/strains and Ly. shimeji AT787 (Fig. S3). However, there was no evidence of an axis reflecting ecology, whereas PC1 clearly reflected taxonomic differences (Fig. S3).
We also searched ortholog groups (orthogroups) whose member numbers are correlated with ecological types and phylogeny (Fig. S4). Although a higher concordance with phylogeny was observed, the correlation to ecology was limited and only eight orthogroups showed a high correlation (correlation coefficient > 0.6 or < −0.6) with ecological type (Fig. S4). In comparison, 2,778 orthogroups showed a high correlation with phylum (Fig. S4). Among orthogroups showing a high correlation with ecological type, an orthogroup encoding PCWDE, GH11, was absent in all mycorrhizal species, including two Ly. shimeji strains (Table S5). No other orthogroup was annotated as having any apparent function related to the mycorrhizal lifestyle. Three genes encoding COX2, COX3, and NADH5 are usually placed in the mitochondrial genome and could result from having mitochondrial data included in the genome dataset.
3.6. Metabolic processes for compounds related to taste
In addition to its primitive mycorrhizal trait, Ly. shimeji is also famous as an edible mushroom. We next surveyed genes involved with taste by comparing Ly. shimeji AT787 genes with genes from two of the most consumed mushroom species (Ag. bisporus and Le. edodes) and two species that are non-toxic but are not generally eaten (Co. cinerea and La. bicolor). Previous studies showed that major compounds providing the umami taste of mushrooms are glutamate and nucleotide monophosphates, such as inosinic acid and guanylic acid.71 We surveyed genes in the biosynthetic pathways for these umami molecules, but none of these genes showed a difference in gene numbers between consumed and non-consumed species (Table S6). Therefore, the accumulation of umami molecules might be regulated by other processes such as gene expression or degradation. Alternatively, the preference for edible fungi can be caused by other compounds, such as odorant compounds that are well-known for truffles and matsutake mushrooms. We also conducted an orthologous analysis of these fungi and found that 113 genes are shared between Ly. shimeji AT787 and widely consumed mushroom species but are not present in the primarily non-consumed species (Fig. 3b, bordered in red). We found that 14 of these genes were annotated as major facilitator superfamily (or abbreviated as ‘MFS’) transporters (Table S7). This family may be a candidate gene family that has expanded in consumed mushroom species. MFS transporters include sugar transporters.72 Some MFS sugar transporters have been mentioned as being involved with strawberry taste via sugar accumulation.73 Therefore, MFS transporters might affect mushroom taste.
3.7. Concluding remarks
In recent years, accumulating information on the fungal genome has revealed common traits of mycorrhizal genomes such as genome size expansion accompanied by TE proliferation and degradation of the PCWDE repertoire.23–26 The Ly. shimeji AT787 genome, however, did not show either of these genomic traits even though the organism is unquestionably mycorrhizal. Our results suggest that increments of genome size and the depletion of PCWDEs may not have been the initial genomic alterations during mycorrhizal evolution but, rather, a consequence of the mycorrhizal lifestyle.
Ly. shimeji is also a good edible mushroom, especially in Japan.8,9 Exceptional among mycorrhizal fungi, Ly. shimeji can develop basidiocarps in axenic culture and, therefore, can be cultivated commercially.13,27 Furthermore, research methods, including transformation, have also been developed.74,75 The genomic information is helpful for agriculture such as the establishment of adequate culture conditions based on metabolic gene repertoire, exploration of genetic markers to select strains with beneficial traits, and application of genetic engineering.76,77 The genome data reported here will aid future research efforts to improve the commercial uses of Ly. shimeji.
Acknowledgements
The authors thank the NIBB Trans-Omics Facility for technical support. The authors thank the NIBB Data Integration and Analysis Facility for supporting data analyses. This work was supported by the Kazusa DNA Research Institute Foundation and a Grant-in-Aid for Scientific Research from the MEXT (no. 22128006 to M.K.).
Conflict of interest
None declared.
Data availability
The sequencing reads, draft genome assembly, and annotations have been deposited in the DDBJ database under BioProject accession PRJDB13731 with some adjustment. The original data including genome assembly, gene prediction, annotation and ortholog tables are available through FigShare (10.6084/m9.figshare.20281272, 10.6084/m9.figshare.21641057).
