Comparative genomic analysis of Sanghuangporus sanghuang with other Hymenochaetaceae species

Abstract

Sanghuangporus sanghuang is a medicinal macrofungus with antioxidant and antitumor activities, and it is enriched with secondary metabolites such as polysaccharides, terpenes, polyphenols, and styrylpyrone compounds. To explore the putative core genes and gene clusters involved in sanghuang biosynthesis, we sequenced and assembled a 40.5-Mb genome of S. sanghuang (SH1 strain). Using antiSMASH, local BLAST, and NCBI comparison, 12 terpene synthases (TPSs), 1 non-ribosomal peptide synthase, and five polyketide synthases (PKSs) were identified in SH1. Combining the transcriptome analysis with liquid chromatography mass spectrometry-ion trap-time of flight analysis, we determined that ShPKS1, one phenylalanine aminolyase (ShPAL), and one P450 monooxygenase (ShC4H1) were associated with hispidin biosynthesis. Structural domain comparison indicated that ShPKS2 and ShPKS3 are involved in the biosynthesis of orsellinic acid and 2-hydroxy-6-methylbenzoic acid, respectively. Furthermore, comparative genomic analysis of SH1 with 14 other fungi from the Hymenochaetaceae family showed variation in the number of TPSs among different genomes, with Coniferiporia weirii exhibiting only 9 TPSs and Inonotus obliquus having 20. The number of TPSs also differed among the genomes of three strains of S. sanghuang, namely Kangneng (16), MS2 (9), and SH1 (12). The type and number of PKSs also varied among species and even strains, ranging from two PKSs in Pyrrhoderma noxium to five PKSs in S. sanghuang SH1. Among the three strains of S. sanghuang, both the structural domains and the number of PKSs in strains MS2 and SH1 were consistent, whereas strain Kangneng exhibited only four PKSs and lacked the PKS with the structural domain KS-AT-DH-KR-ACP. Additionally, Sanghuangporus species exhibited more similar PKSs to Inonotus, with higher gene similarity around five PKSs, while showing differences from those of other fungi in the same family, including Phellinus lamaoensis. This result supports the independent taxonomic significance of the genus Sanghuangporus to some extent.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42770-023-01212-x.

Introduction

Hymenochaetaceae, the largest family in the Basidiomycota phylum, is known for its wood-parasitic fungi that cause white rot disease [1]. Several well-known medicinal fungi, including Phellinus ignarius, I. obliquus, S. sanghuang, and S. baumii, are members of this family [2]. Sanghuang, previously known as Phellinus linteus and I. sanghuang [3, 4], is now classified as S. sanghuang. It commonly parasitizes mulberry trees (Morus) [5] and is rich in secondary metabolites such as polysaccharides, terpenes, and polyphenols [6]. Sanghuang exhibits various pharmacological activities, including antitumor, antioxidant [7], anti-inflammatory [8], immunomodulatory, antidiabetic [9], liver-protective [10], and neuroprotective [11]. Furthermore, Sanghuang mycelium has been found to have no mutagenic activity in bacterial reverse mutation tests (Ames test), and thus, it does not induce chromosomal aberrations [12]. Acute oral toxicity studies conducted on Institute of Cancer Research mice for a duration of 7 days revealed that the median lethal dose (LD₅₀) value was greater than 12 g/kg body weight. In the mouse micronucleus test, Sanghuang mycelium did not alter the abundance of reticulocytes or significantly increase the rate of micronucleated reticulocytes, which means Sanghuang does not cause chromosomal damage during the cell cycle. These findings indicate that Sanghuang possesses extremely low toxicity, thereby demonstrating its safety for human consumption [12].

Many fungi in the Hymenochaetaceae family, including S. sanghuang, produce styrylpyrone compounds. These compounds are believed to have functions similar to flavonoids in plants, including defense and signal transduction [13]. Hispidin, a widely distributed styrylpyrone in macrofungi of the genera Phellinus, Sanghuangporus, Inonotus, and Fomitiporia [14], exhibits cytotoxicity [15], potential hypoglycemic effects [16], antibacterial and anti-inflammatory properties [17–19], antiobesity effects [20], antiviral and neuroprotective activities, and antioxidant and antitumor properties [21, 22]. The synthesis pathway of hispidin in Sanghuang involves the deamination of phenylalanine to cinnamic acid by phenylalanine ammonia lyase (PAL), followed by the conversion of cinnamic acid to p-coumaric acid and caffeic acid by C4H. Caffeic acid is then converted to caffeoyl-CoA by 4-coumaroyl-CoA ligase (4CL). Hispidin biosynthesis likely occurs via chain extension and cyclization of caffeoyl-CoA, facilitated by polyketide synthase (PKS) and palmitoyl protein thioesterase genes [22–24].

As one of the primary sources of hispidin, the precise structural domain of the key PKS gene responsible for hispidin synthesis in Sanghuangporus remains uncertain. In this study, we employed various approaches, including antiSMASH, local BLAST, and NCBI comparison, to analyze a total of 15 genomes from the Hymenochaetaceae family. Our main focus was to identify and explore putative biosynthetic gene clusters in closely related species. Additionally, we conducted a comprehensive transcriptome analysis of strain SH1 under four distinct culture conditions and integrated the results with those of liquid chromatography mass spectrometry-ion trap-time of flight (LCMS-IT-TOF) analysis. This combined approach enabled us to make inferences about the specific PKS gene involved in hispidin biosynthesis in S. sanghuang. By uncovering the key genes responsible for hispidin synthesis, we aimed to facilitate further investigation and utilization of Sanghuang’s secondary metabolites through targeted gene mining strategies.

Materials and methods

Experimental materials

The S. sanghuang (SH1) strain was obtained from the Edible Fungi Research Institute of the Shanghai Academy of Agricultural Sciences.

Sample preparation

DNA sample preparation

Pure and uncontaminated mycelium was selected from yeast malt extract broth (MY, Qingdao Rishui Biotechnology Co., Ltd.) solid medium. The mycelium was ground using a tissue grinder and inoculated into a 250-mL conical flask containing 100 mL of MY liquid medium. The culture was shaken at 28 °C and 150 rpm. After 15 days, the mycelium was filtered, harvested, and frozen in liquid nitrogen, and it was stored at – 80 °C for DNA extraction.

RNA samples

Three replicates of RNA samples were prepared under four different culture conditions. Yeast malt extract broth (MY, Qingdao Rishui Biotechnology Co., Ltd.) served as the base medium, and lactose (RT) and sorbitol (SLC) were added at a concentration of 4 g/L to prepare liquid media. Maltose (1.8 g/L) and glucose (6 g/L) were used as the base medium, and yeast powder (JMF) and soya peptone (DDDB) were added at a concentration of 4 g/L to prepare liquid media. The sterilization, inoculation, and harvesting methods were the same as described above. The harvested mycelium samples from three replicates of each treatment were collected together, frozen in liquid nitrogen, and then stored at – 80 °C for total RNA extraction.

Crude extracts

SH1 was cultured in 100 mL of four different liquid media by shaking in a 250-mL Erlenmeyer flask at 28 °C and 150 rpm for 15 days. The culture broth and medium were collected and extracted with 100 mL of ethyl acetate. The extract was then concentrated by vacuum distillation. After centrifugation at 12,000 × g and filtration through a 0.22-μm filter membrane, the crude extract was dissolved in 2 mL of methanol for LCMS analysis.

Genome sequencing and assembly

The whole-genome shotgun strategy was employed to construct libraries with different insert fragments. These libraries were sequenced through paired-end sequencing on the Illumina NovaSeq sequencing platform by utilizing next-generation sequencing (NGS) technology. Raw data was checked for quality with FastQC (v. 0.11.7). AdapterRemoval [25] (v. 2.2.2) was used for trimming of adapter sequences and low-quality bases. And Soapec (V 2.0) software was used to perform quality correction on all reads based on the KMER frequency to obtain high-quality adaptor-free genome sequences. The filtered data were de novo assembled using A5-miseq and SPAdes software [26], generating contigs and scaffolds. The assembly results were corrected using pilon v. 1.18 software [27] and the integrity of the fungal genome assembly was assessed by BUSCO v. 3.0.2 software [28].

Gene prediction and annotation

A combination of homology prediction, de novo prediction, and RNA-Seq data annotation was used for the prediction of gene structures. The exonerate software (v. 2.2.0, http://www.ebi.ac.uk/about/vertebrate-genomics/software/) was utilized to obtain homologous gene prediction results using protein sequences from closely related species. The de novo prediction using Augustus (v. 3.03), glimmerHMM (v. 3.0.1), and GeneMark-ES (v. 4.35) to obtain corresponding gene prediction results. For gene structure annotation, RNA-seq data were subjected to gene structure annotation by PASA software (v. 2.2.3, http://pasapipeline.github.io). At last, EVidenceModeler software was utilized to integrate the prediction results [29]. The predicted genes were annotated by searching existing databases, including the NCBI non-redundant protein sequences, Swiss-Prot, Kyoto Encyclopedia of Genes and Genomes (KEGG), Evolutionary Genealogy of Genes: Non-supervised Orthologous Groups (EggNOG), Pathogen-Host Interactions database, and Carbohydrate-Active EnZymes (CAZy).

Secondary metabolite biosynthesis gene analysis

The secondary metabolite biosynthesis genes of C. sulphurascens, C. weirii, F. mediterranea, G. junonius, Pd. pouzari, Pn. lamaoensis, Pl. nigrolimitatus, Po. pini, Py. noxium, I. hispidus, I. obliquus, S. baumii, and three S. sanghuang strains (Table S1) were predicted using antiSMASH (https://antismash.secondarymetabolites.org/ (accessed in 2022)). FGENESH online program (www.softberry.com/) was employed to predict protein and cDNA sequences. The PKS/non-ribosomal peptide synthase (NRPS) online website (NRPS.igs.umaryland.edu/) was used to identify gene clusters containing genes with NRPS/PKS domains. NCBI online BLAST analysis (https://blast.ncbi.nlm.nih.gov/) was utilized to compare protein domains to identify overlapping groups carrying NRPS/PKS and TPS genes.

Phylogenetic analysis

The known PKS protein sequences downloaded from NCBI were aligned with the identified protein sequences from this study using One Step Bulid an ML Tree program from TBtools [30] (Muscle for multiple sequence alignment, trimAI for trimming the results, and IQ-tree for filtering the amino acid substitution model automatically and finally builds an ML tree) with default parameters, followed by visualization and optimization in iTOL [31].

LCMS analysis

LCMS analysis was performed using an LCMS-IT-TOF system (Shimadzu, Kyoto, Japan) with an Agilent Eclipse Plus C18 column (100 × 2.1 mm i.d., 1.8 μm, Agilent Technologies); column temperature was 30 °C. The mobile phase consisted of 0.05% formic acid–water solution (A) and acetonitrile (B). The flow rate was set to 0.2 mL·min-1, and the elution gradient was 5–100% linear gradient B for 12 min, followed by 100% B for 4 min, and then rapidly returned to the initial 5% B within 2 min. For each analysis, 2 μL of sample was injected.

The mass spectrometry conditions included positive and negative ion modes with a scan range of m/z 100–1000. The other parameters were set as follows: electrospray ionization source spray voltage, 4.50 kV/ − 3.50 kV; ionization temperature, 200 ℃; detector voltage, 1.65 kV; dry gas pressure, 110.0 kPa; nebulizer gas (N2) flow rate, 1.5 L/min; curved desolvation line temperature, 200 °C; and heat block temperature, 200 °C.

The Shimadzu component formula predictor was used to infer the molecular formula. The presence of hispidin was confirmed by comparison with a reference sample with the same retention time (tR = 9.1 min) and mass spectrum (positive, m/z 345.0944, [m + H] + , − 2.5 mD; negative, m/z 343.0812, [m-H], − 1.1 mD). Furthermore, the presence of hispidin in SH1 samples cultured on different media was confirmed.

Transcriptome sequencing

Total RNA was isolated using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany). After extraction, purification, and library construction, the library was sequenced using NGS technology based on the Illumina HiSeq sequencing platform with paired-end sequencing. The integrity, purity, and concentration of total RNA were checked, along with the library quality, using an Agilent 2100 Bioanalyzer. High-quality sequences were assembled de novo to obtain transcript sequences, which were then clustered, and the longest transcripts were selected for subsequent GO, KEGG, and EggNOG analyses (Transcriptome Accession: PRJNA961736).

Results

Genome sequencing and assembly

A total of 31,007,418 raw reads were obtained through Illumina sequencing (Illumina Inc., San Diego, CA, USA), resulting in 30,751,898 high-quality reads. The size of the whole genome of SH1 is 40.5 Mb, consisting of 106 overlapping groups and 106 scaffolds with a complete BUSCOs of 90.1%. The genome has an N50 of 944,366 bp and a GC content of 47.89%. Gene prediction identified a total of 12,307 protein-coding genes spanning 23.1 Mb, with an average sequence length of 1877.1 bp. The longest overlapping group is 5.6 Mb in size. For noncoding RNAs, tRNAscan, RNAmmer, and rfam_scan were used to predict the secondary structures of 111 tRNAs, 16 rRNAs, and 23 ncRNAs, respectively.

Annotation of the SH1 genome

The EggNOg database showed that except “Function unknown” (3025), the most of the predicted genes were functionally associated with “Replication, recombination and repair” (879), “Posttranslational modification, protein,” and “Carbohydrate transport and metabolism” in that order (Figure S1). Functional classification using KEGG revealed protein families involved in genetic information processing (2386), signaling and cellular processing (607), and signal transduction (615) (Figure S2). The diversity of genetic information and signaling proteins likely contributes to more efficient information exchange and secondary metabolism. GO annotation demonstrated a significant enrichment of genes associated with biological processes, including 2122 cellular nitrogen compound metabolic processes and 1830 biosynthetic processes. In addition, it revealed 3011 cellular components, 2910 intracellular components, and 2290 organelle-related components at different locations. Furthermore, it identified 5707 molecular functions, ion binding (2856), and oxidoreductase activity (854) (Fig. 1).

Fig. 1.

Functional annotation of SH1 gene-encoding proteins in the Gene Ontology (GO)

CAZy analysis indicated the presence of 399 genes encoding CAZy in SH1, including 212 glycoside hydrolases, 69 glycosyl transferase genes, 79 carbohydrate esterases, 79 auxiliary activities, 19 carbohydrate-binding modules, and 10 polysaccharide lyases (Fig. 2).

Fig. 2.

CAZy functional classification chart of SH1

Analysis of key genes for biosynthesis of secondary metabolites

In this study, 18 biosynthetic gene clusters were identified in the S. sanghuang (SH1) genome through antiSMASH and local BLAST analyses, including 12 TPSs, one complete NRPS, and five complete PKSs. The core structural domains of PKSs comprise β-ketoacyl synthase (KS), acyltransferase (AT), and acyl carrier protein (ACP). According to the degree of β-keto reduction, PKSs can be classified into three subgroups. Highly reducing PKSs possess a complete reduction ring, including β-ketoacyl reductase (KR), dehydratase (DH), and enoyl reductase (ER), as well as methyl transferase, in addition to the basic structural domains. Partially reducing PKSs lack a complete reduction ring, whereas non-reducing PKSs lack the KR, DH, and ER domains but typically possess starter acyltransferase (SAT) and product template (PT) domains, in addition to KS, AT, ACP, and thioesterase (TE) domains [32]. The SH1 strain comprises five polyketide synthesis-related structural domains: ShPKS1 (CaiC-ACP-KS-AT-DH-KR-ACP-ACP), non-reducing ShPKS2 (SAT-KS-AT-PT-ACP-ACP-TE), ShPKS3 (KS-AT-DH-KR-ACP-TE), ShPKS4 (CaiC-ACP-KS-AT), and the PKS-NRPS hybrid ShPKS5 (KS-ACP-TE-AT-A-ACP). Among these, PKS with the SAT-KS-AT-PT-ACP-ACP-TE structural domain is present in all 15 genomes. Comparative genomic analysis revealed differences in the number and types of PKSs within and between genera of species within the same family. For example, minimum PKSs (2) were identified in Pn. lamaoensis (FFPRI411162) and Py. noxium (FFPRI411160) genomes; three PKSs were identified in C. sulphurascens (FP133613(A)Sp), C. weirii (30910), and Pd. pouzarii (DSM 108285); four PKSs were identified in F. mediterranea (MF3/22), I. hispidus (NPCB_001), I. obliquus (CT5), Po. pini (BCRC 35384), S. baumii (821), and S. sanghuang (Kangneng); and five PKSs were identified in G. junonius (AH 44721), Pl. nigrolimitatus (SigPhenig9), S. sanghuang (MS2), and S. sanghuang (SH1). Compared to SH1, S. baumii (821) and S. sanghuang (Kangneng) lack the KS-AT-DH-KR-ACP domain, whereas F. mediterranea (MF3/22), G. junonius (AH 44721), and Po. pini (BCRC 35384) possess this PKS domain. Additionally, the number of TPSs varied significantly among the genomes. For instance, the genomes of the three S. sanghuang strains (Kangneng, MS2, and SH1) were found to have 16, 9, and 12 TPSs, respectively (Table 1).

Table 1.

Numbers of biosynthetic gene clusters identified by antiSMASH and local BLAST analyses in the 15 studied fungal genomes

Strains	PKS	NRPS	TPS
C. sulphurascens (FP133613(A)Sp)	3	1	10
F. mediterranea (MF3/22)	3	1	9
G. junonius (AH 44721)	4	1	15
C. weirii (30910)	5	0	11
I. hispidus (NPCB_001)	4	1	15
I. obliquus (CT5)	4	1	20
Pd. pouzarii (DSM 108285)	3	1	13
Pn. lamaoensis (FFPRI411162)	2	1	13
Pl. nigrolimitatus (SigPhenig9)	5	1	15
Po. pini (BCRC 35384)	4	1	19
Py. noxium (FFPRI411160)	2	1	16
S. baumii (821)	4	1	13
S. sanghuang (Kangneng)	4	1	16
S. sanghuang (MS2)	5	1	9
S. sanghuang (SH1)	5	1	12

ShPKS1 and other PKS1 domains (CaiC-ACP-KS-AT-DH-KR-ACP-ACP) show similarity to the hispidin synthase domain (CaiC-ACP-KS-AT-ACP), which is found in bioluminescent fungi such as Mycena sanguinolenta, M. kentingensis, and M. chlorophos, and they were found to cluster together in the phylogenetic analysis (Figure S3). Analysis of the surrounding genes in 14 genomes revealed high similarity between the genera Sanghuangporus and Inonotus, although Inonotus lacks glycoside hydrolases, peroxidases, and ribosomal genes but has an additional fungal transcription factor and protein kinase. Within Sanghuangporus, S. baumii (strain 821) showed notable differences in its surrounding genes compared with other strains, with significantly longer PKS and peroxidase genes and the absence of a protein kinase and WD40 gene (Fig. 3).

Fig. 3.

Comparison of PKS1s (putative core genes for hispidin biosynthesis) and their surrounding genes in 14 genomes

ShPKS2 shares the same structural domain (SAT-KS-AT-PT-ACP-ACP-TE) with the orsellinic acid synthesis-related PKS domains found in Armillaria mellea, Gloeophyllum trabeum, Moniliophthora roreri, Sparassis crispa, Stereum hirsutum, and Taiwanofungus camphoratus. The protein sequences of these PKSs clustered together in the phylogenetic analysis (Figure S3), suggesting that the product of ShPKS2 and the PKSs with the same domain may be orsellinic acid or its derivatives. Additionally, the analysis of this PKS and its surrounding genes in the 14 genomes revealed patterns similar to that of PKS1. Sanghuangporus and Inonotus show similarities, whereas within Sanghuangporus, the surrounding genes of S. baumii (strain 821) differ from Sanghuangporus strains (Fig. 4). Compared with the S. sanghuang strains MS2 and SH1, it has an additional dehydrogenase and lacks a zinc finger, while the reverse transcriptase and DUF undergo inversion. S. sanghuang (Kangneng) lacks a zinc finger but has a larger reverse transcriptase. Additionally, there are variations within Inonotus, with the FAD/NAD-binding protein and reverse transcriptase undergoing inversion. I. obliquus (CT5) has a longer PMCA and lacks a dehydrogenase. Moreover, the surrounding genes of this PKS in the genomes of C. sulphurascens, C. weirii, and Pd. pouzarii showed high similarity (Fig. 5).

Fig. 4.

Comparison of PKS2s (putative orsellinic acid synthesis-related PKS) and their surrounding genes

Fig. 5.

Comparison of PKS3s (putative core genes for 2-hydroxy-6-methylbenzoic acid biosynthesis) and the surrounding genes

The PKSs with the same domain as ShPKS4 (CaiC-ACP-KS-AT) were found only in the genomes of six species of Sanghuangporus and Inonotus. ShPKS4 and its surrounding genes showed fusion and separation of PKS, FAD/NAD-binding protein, and BEACH protein (Figs. 6 and 7).

Fig. 6.

Comparison of PKS4s and the surrounding genes, showing fusion and separation of genes

Fig. 7.

Comparison of PKS5s and the surrounding genes, showing fusion and separation of genes

Detection of hispidin in SH1 cultured on four different media using LCMS-IT-TOF

LCMS-IT-TOF analysis was used to detect hispidin in crude extracts from the four media. The chromatogram (Fig. 8) revealed that SH1 produced hispidin in the liquid media based on yeast malt infusion broth comprising 4 g/L of lactose (RT) and sorbitol (SLC), but not in the liquid media based on 1.8 g/L maltose and 6 g/L glucose, with the respective addition of 4 g/L of yeast powder (JMF) and soya peptone (DDDB).

Fig. 8.

The detection of hispidin in different SH1 culture extracts. The four media were based on yeast malt infusion broth, with the addition of 4 g/L lactose (RT) and sorbitol (SLC), and liquid media having 1.8 g/L maltose and 6 g/L glucose, with the addition of 4 g/L of yeast powder (JMF) and soya peptone (DDDB), respectively

Transcriptome data analysis of SH1 in different media

Transcriptome (Transcriptome Accession: PRJNA961736) data analysis (Table S2, S3) revealed the expression (FPKM values were used to normalize gene expression and then to compare the read count values of each gene) of the CaiC-ACP-KS-AT-DH-KR-ACP-ACP genes in different media in the following order: lactose (RT) > sorbitol (SLC) > soya protein peptone (DDDB) > and yeast powder (JMF); the highest expression was observed in lactose, which was 1.91 times that on sorbitol, 5.42 times that on soya protein peptone, and 10.58 times that on yeast powder. The difference is significant and this expression pattern was positively correlated with the yield of hispidin across the four media. Furthermore, other potentially involved enzymes were screened using local BLAST and NCBI annotation results, including four C4Hs (ShC4H1, ShC4H2, ShC4H3, ShC4H4), one PAL (ShPAL), and ShPKS1. These enzymes showed similar expression patterns and were consistent with hispidin production (Fig. 9). It is speculated that ShC4H1 and ShPAL, exhibiting the most consistent expression patterns, are involved in hispidin biosynthesis.

Fig. 9.

Heat map based on the expression of genes potentially involved in the hispidin synthesis pathway under four different culture conditions. Red color represent high expression, while blue color represent low expression

Discussion

In plants, flavonoid compounds are well-known for their presence in flowers, leaves, and seeds and their role in regulating plant responses to various environmental factors [33]. The synthesis of flavonoids involves several enzymatic steps, starting with the catalysis of three acetoacetyl-CoA decarboxylation additions to p-coumaroyl-CoA by type III PKS (chalcone synthase), resulting in the formation of chalcone [34]. Subsequently, chalcone is isomerized to flavanone by chalcone isomerase (CHI), and further modifications occur under the catalysis of various enzymes, leading to the production of a variety of flavonoid compounds [35]. Although flavonoids were initially considered as characteristic compounds of higher plants [36], recent studies have revealed that fungi can also synthesize a range of flavonoid compounds. For instance, the PKS8 gene cluster in Fusarium is involved in the biosynthesis of fusamarin derivatives [37]. While fungal genomes often lack many genes associated with flavonoid synthesis in plants [38], recent investigations into fungal PKSs have unveiled interesting findings. In the endophytic fungus Pestalotiopsis fici, an NRPS-PKS hybrid enzyme called FnsA with a structural domain of A-T-KS-AT-DH-KR-ACP-TE has been identified. FnsA can utilize p-coumaric acid or p-hydroxybenzoic acid as the starting unit and extend it with three or four acetoacetyl-CoA units to form a chalcone skeleton, ultimately synthesizing naringenin [39]. A similar core gene called cfoA, sharing the same structural domain as FnsA, has been discovered in A. candidus. The assembly mechanism of cfoA involves using benzoic acid or p-hydroxybenzoic acid as the starting unit and extending it with acetoacetyl-CoA for four rounds to form a chalcone skeleton, catalyzing the synthesis of chloroflavonin [40]. Additionally, CHI and flavone synthase responsible for converting chalcone into flavonoids in A. candidus differ from those reported in natural systems such as plants and bacteria [40]. These NRPS-PKS genes perform the function of synthesizing flavonoid-like compounds, and this unique fungal flavonoid biosynthesis system differs significantly from that found in plants. Styrylpyrones and flavonoids are both catalyzed by pks PKS, in Sanghuangporus, given the absence of CHI1 in the flavonoid biosynthesis pathway [17, 38], we propose that, similar to FnsA and cfoA, polyketide synthase involved in the biosynthesis pathway of hispidin, a styrylpyrone-like compound derived from the cinnamic acid pathway, is distinct from chalcone synthase.

Studies on bioluminescent fungi, such as Mycena chlorophos and Neonothopanus nambi, have revealed that the hisps gene in bioluminescent fungi possesses a structural domain of AMP-ACP-KS-AT-ACP, which lacks two structural domains, ketoreductase, and dehydratase, compared with closely related non-bioluminescent fungi of the order Agaricales. This finding confirms that in bioluminescent fungi, the biosynthesis of hispidin starting from caffeic acid is carried out by type I PKS [41], which supports the hypothesis presented in this study. Additionally, the CaiC-ACP-KS-AT-DH-KR-ACP-ACP structural domain of PKS gene required for hispidin synthesis in Sanghuangporus, as speculated in this study, is similar to that of non-bioluminescent hispidin synthase. Furthermore, PKS genes with the same or very similar structural domain as CaiC-ACP-KS-AT-DH-KR-ACP-ACP have been identified in several genera of fungi, including Fomitiporia, Gymnopilus, Inonotus, Phellinidium, Phellinus, Porodaedalea, and Sanghuangporus, where hispidin has been reported to exist [16, 42–47]. Additionally, PKS genes with the same or similar structural domain have also been identified in Coniferiporia and Py. noxium in which the presence of hispidin has not been reported yet. No PKS gene was identified in the strain SigPhenig9 of Pl. nigrolimitatus, which also lacks hispidin. The genomes of S. sanghuang’s other congeneric fungi, such as C. sulphurascens, C. weirii, F. mediterranea, Pd. pouzarii, Py. noxium, G. junonius, Po. pini, I. hispidus, I. obliquus, and S. baumii, which have been shown to produce hispidin, also contain PKS genes with the structural domain of CaiC-ACP-KS-AT-DH-KR-ACP-ACP. These findings align with the inferred structural domain of the PKS gene required for hispidin synthesis in Sanghuangporus proposed in this study.

While genome mining allows for the exploration of known and unknown biosynthetic gene clusters and compounds, a large number of genes or BGCs identified by this method may be silent and not produce desirable metabolites. Also, although the genome of a particular cell line is immutable, different growth conditions may lead to significant changes in gene expression profiles and metabolite production. At this point, combining the transcriptome with metabolite changes under different culture conditions can provide a basis for biosynthetic gene-compound correlation, in this study which means ShPKS1 was not only highly expressed in RT and SLC, but also corresponded to the presence of hispidin in both media. And in terms of secondary metabolite prediction methodology, although Jiang et al. applied a single-molecule real-time (SMRT) method to sequence the S. sanghuang strain MS2 [48], Shen et al. also re-annotated the MS2 genome by the newly sequenced transcriptomic data from the same strain [49]. These studies provide a reliable basis for the further application of S. sanghuang. However, their analyses both identified four T1PKS, which suggests that gene cluster analysis relying only on antismash may be incomplete and requires the assistance of local blast, NCBI database comparison, and other methods.

In the transcriptome data-based co-expression analysis of this study, the expression pattern of 4CL, which catalyzes the formation of caffeoyl-CoA from caffeic acid, can subsequently undergo cyclization and condensation to produce hispidin and lignin in plants [50], and does not correlate with the expression of PKS1 or hispidin production. Protein sequence analyses of plant-related genes (Figure S4) suggest that the CaiC structural domain (acyl-CoA synthetase (AMP forming)/AMP-acid ligase II) in the ShPKS1 gene of the SH1 strain may partially play a role similar to 4CL. Furthermore, in bacterial and fungal genomes, many secondary metabolites are synthesized through metabolic pathways encoded by the genes that are physically adjacent on chromosomes. This clustering phenomenon of genes is common in bacteria and filamentous fungi. However, in this study, the genes (10 upstream and 10 downstream) surrounding the PKS genes did not show co-expression with the PKS genes (Tables S4-8), and the genes related to the hispidin synthesis pathway were physically distant or even located on different scaffolds. Nonetheless, their expression patterns were similar or consistent with that of ShPKS1 (Fig. 9). Therefore, it is believed that in the family Hymenochaetaceae, the key genes involved in secondary metabolite synthesis are not tightly clustered but loosely dispersed over large regions of the genome. This observation may be attributed to the longer and more complex gene structure in basidiomycetes, resulting in a different definition of gene clusters compared with compact bacterial genomes and a more evolutionary resemblance to plants. These findings further support the notion that unique fungal biosynthesis mechanisms can lead to the synthesis of plant-like metabolites.

The taxonomy and systematic position of Sanghuangporus species have been a topic of debate due to their morphological similarity with closely related species from the genera Phellinus and Inonotus. Sanghuangporus has been established as a distinct genus in studies conducted by Wu et al. [4], Zhou et al. [51], and Han et al. [5]. In this study, the high conservation of PKSs’ surrounding genes among all members of Sanghuangporus, as well as their similarity with two species of Inonotus, supports the independent systematic position of the genus Sanghuangporus. Variations in the direction and position of sequences among closely related species could potentially be attributed to horizontal gene transfer events between species within the same genus [52]. It has also been suggested that the S. sanghuang (Kangneng) involved in this study is actually from S. vaninii [48, 53, 54], which may explain why strain Kangneng, like S. baumii, could not be identified to PKS5 as the other two S. sanghuang (SH1 and MS2) could be, as well as better correspond to the clustering tree based on the PKS protein sequences(Figure S3) in this study. Moreover, the difference between the number of glycoside hydrolases and glycosyl transferases in Sanghuang was more significant compared with that in other fungi like Ascomycota, Basidiomycota, Chytridiomycota, and Zygomycota [55], suggesting that the degradation of polysaccharides is more crucial for its growth and metabolism than their synthesis. The ability of Sanghuang to degrade wood is consistent with its reliance on lignocellulose degradation for survival [17]. Notably, certain families of CAZymes are encoded by more genes in S. sanghuang [48]. The nutritional strategy besides that shared in Sanghuangporus may lead to the host specificity of Sanghuangporus species.

In addition, the analysis of surrounding genes revealed fusion and separation phenomena of PKS, FAD/NAD-binding and BEACH genes around ShPKS4 (CaiC-ACP-KS-AT) and these phenomena of PKS, NRPS, and FAD/NAD-binding protein with the surrounding genes of ShPKS5 (KS-ACP-TE-AT-A-ACP). Furthermore, ShPKS1 (CaiC-ACP-KS-AT-DH-KR-ACP-ACP) can be considered as a chimera of Acyl-CoA synthetase and PKS. It is worth noting that gene fusion or enzyme fusion is a common phenomenon in catalyzing compound synthesis, particularly in fatty acid synthase, PKS, and NRPS systems [56]. Examples of enzyme fusion as dual/multi-functional biocatalysts can be found in various Aspergillus species, such as the NRPS-PKS hybrid enzymes FnsA and cfoA involved in naringenin synthesis or in the plant-derived endophytic fungus Xylaria arbuscula, which produces natural products such as pyranonigrin E and cytochalasins [57–59]. Enzyme fusion not only prevents the accumulation of intermediates that may inhibit enzyme activity but also improves the efficiency of enzyme-coupled reactions due to the proximity of active sites [60]. Therefore, in-depth studies on chimeric genes or fusion enzymes are warranted for discovering new biosynthetic pathways or even novel compounds [61]. However, regardless of the type of PKS identified in the present study, verifying their function through deletion mutations or heterologous expression with added promoters would be an essential next step to explore their role in hipsidin synthesis.

Conclusions

In this study, we reported a 40.5-Mb genome of with S. sanghuang (SH1). Combining transcriptome and LCMS-IT-TOF analysis, we concluded that ShPKS1, one phenylalanine aminolyase (ShPAL1), and one P450 monooxygenase (ShC4H1) were related with the biosynthesis of hispidin. According to domain analysis, we also inferred that ShPKS2 and ShPKS3 are involved in the biosynthesis of orsellinic acid and 2-hydroxy-6-methylbenzoic acid, respectively. Furthermore, comparative genomic analysis showed the number and type of PKSs and TPSs varied among different genomes. Different strain of S. sanghuang had different number and type of PKSs. According to domain structure and adjacent gene organization of PKS comparative analysis in Hymenochaetaceae family, Sanghuangporus genus is closed to Inonotus. Based on gene expression analysis, the adjacent genes of PKS were not coexpressed. It implied that biosynthesis genes of polyketide and terpene may not cluster together in S. sanghuang.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

C. sulphurascens

Coniferiporia sulphurascens

C. weirii

Coniferiporia weirii

F. mediterranea

Fomitiporia mediterranea

G. junonius

Gymnopilus junonius

I. hispidus

Inonotus hispidus

I. obliquus

Inonotus obliquus

Pd. pouzarii

Phellinidium pouzarii

Pn. lamaoensis

Phellinus lamaoensis

Pl. nigrolimitatus

Phellopilus nigrolimitatus

Po. pini

Porodaedalea pini

Py. noxium

Pyrrhoderma noxium

S. baumii

Sanghuangporus baumii

S. sanghuang

Sanghuangporus sanghuang

Author Contribution

Conceptualization, JZ and YW; methodology, YW, XYW, and ZWL; software, YW, XLY, and DW; formal analysis, ZWL; investigation, JMN and XLY; resources, YY; writing—original draft preparation, XYW; writing—review and editing, XYW, YW, and JZ.

Funding

This work was supported by a grant from the National Natural Science Foundation of China (31860177), General Project of Basic Research Program in Yunnan Province (202101AT070218), the Reserve Talents for Young and Middle-aged Academic and Technical Leaders of the Yunnan Province (202205AC160044).

Data Availability

This study project is available under NCBI BioProject accession number PRJNA960849 and BioSample accession number SAMN34331867. The complete genome assembly of Sanghuangporus sanghuang SH1 is available under GenBank accession number JASFXP000000000.

Declarations

Conflict of Interest

The authors declare no competing interests.