Genome Assembly, Evaluation and Annotation of TPD11
SMRT sequencing of the endophytic fungus A. tamarii TPD11 using the Pacbio Sequel Ⅱ sequencing platform yielded 247,466 clean reads from 3,868,987,455 bp bases, and 10 contigs were assembled into a genome of 38.81 Mb (Fig. 1), with an N50 of 5.34 Mb and a GC content of 47.28% (Table 1). Sorting based on assembly-corrected genome sequence length information, the continuity of the assembled genome was good (Fig. S1). BUSCO assessment found about 96.70% complete gene elements, and the vast majority of conserved genes were assembled completely (Fig. S2, A). Prediction of the genome using CEGMA yielded 241 core genes (97.18%), of which 236 complete core genes (95.16%) were found, with good accuracy and completeness of the assembly (Fig. S2, B). The second-generation sequencing data was utilized to align the assembled genome, with an average sequencing depth of 183.64X, an alignment rate of 98.16%, and good sequence agreement. The third-generation genome reads were aligned to the corrected genome using minimap2, with a reads alignment rate of 99.99% and an average sequencing depth of 99.53X.
Table 1
Comparison of sequencing, assembly metrics and genome quality of A. tamarii.
Parameter
|
TPD11
|
VU91
|
UdeA_Ata2
|
UdeA_Afl2
|
Sequencing technology
|
PacBio Sequel II HiFi
|
Illumina NovaSeq
|
Illumina HiSeq
|
Illumina HiSeq
|
Coverage depth(X)
|
99.53
|
140
|
50
|
50
|
No. of scaffolds/contigs
|
10
|
302
|
910
|
592
|
Genomic size (Mb)
|
38.81
|
38.04
|
38.18
|
37.75
|
Largest length (Mb)
|
8.65
|
NA
|
0.53
|
0.80
|
Contig N50 (kb)
|
5335.9
|
273.9
|
89.6
|
108
|
GC content (%)
|
47.28
|
47
|
47.62
|
47.63
|
Complete BUSCOs (%)
|
96.7
|
99.5
|
NA
|
NA
|
No. of predicted genes
|
11 305
|
NA
|
13 771
|
13 630
|
GenBank accession no.
|
GCA_030556085.1
|
GCA_029962655.1
|
GCA_008973505.1
|
GCA_008973515.1
|
Reference
|
This study
|
[22]
|
[23]
|
[23]
|
The predicted protein-coding genes were 11,305 with an average gene length of 1,673.37 bp, each containing an average number of 3.2 exons, an average exon length of 463.59 bp and an average intron length of 86.19 bp (Table S1). There were 8,401 repetitive sequences with a total length of 1,234,306 bp, which accounted for 3.18% of the whole genome, of which TEs, simple repeats, and scattered repeats accounted for 2.51%, 0.04%, and 0.48%, respectively; 400 noncoding RNAs were predicted, among which there were 255 tRNAs, 105 rRNAs, 8 snRNAs, 37 sRNAs, and 3 homeotic regulatory elements (Table S2).
Functional annotation of protein-coding genes based on Nr, GO, KOG, KEGG, and SwissProt databases resulted in a total of 11,172 (98.82%) genes annotated (Fig. S3). Among the 9,670 genes annotated for functional classification in the GO database, the highest number of genes related to transmembrane transport in biological processes was found. The molecular functions constituted the major groups, with the main functions being protein binding, oxidoreductase activity, and transmembrane transport activity (Fig. S4). Functional annotation based on the KOG database identified 3716 genes with the main functions of R (general function prediction), Q (secondary metabolite biosynthesis, transport and metabolism), and I (lipid transport and metabolism) (Fig. S5). 4205 genes were annotated to the KEGG database to be involved in five pathways, and the largest number of genes were involved in metabolism, mainly in glucose metabolism, amino acid metabolism and lipid metabolism (Fig. S6). Among them, 130 genes were involved in the biosynthesis of secondary metabolites.
Comparative genomic analysis of TPD11
Phylogenetic analyses of ITS sequences of 24 Aspergillus species and 2900 single-copy homologous sequences shared by the species were carried out (Fig. 2). The ITS sequences grouped the 24 species of Aspergillus into three taxa, whereas the single-copy homologous sequences grouped A. candidus and A. campestris into a single taxon. Both the ITS sequences and the single-copy homologous sequences clustered A. aculeatus, A. brunneoviolaceus, A. homomorphus, and A. japonicus into one group. ITS sequences and single-copy homologous sequences clustered A. aculeatus, A. brunneoviolaceus, A. homomorphus, and A. japonicus into one taxon; A. thermomutatus, A. fumigatus, A. fischeri, and A. lentulus into one taxon; and A. thermomutatus, A. fischeri, and A. lentulus into one taxon. TPD11 clustered with A. tamarii and was identified as A. tamari TPD11, consistent with previous studies [15]. Phylogenetic analysis showed that TPD11 was most closely related to A. caelatus and A. pseudotamarii in terms of evolution.
Comparative analysis of A. tamari subspecies from different sources showed that the genome size of TPD11 was similar to that of VU91, UdeA_Ata2 and UdeA_Afl2 strains. Among them, TPD11 had the largest contig N50 value (5.3 Mb), indicating that the TPD11 genome was the best spliced (Table 1). A total of 1701 homologous groups were identified from the three A. tamarii species and 124 homologous groups were shared among the three species. TPD11 had the most unique groups (1198) compared to VU91 and UdeA_Afl2 (Fig. S7). TPD11 shared more homologous groups with UdeA_Afl2 (414) than with VU91 (185), and the similarity between TPD11 and UdeA_Afl2 was greater than that with VU91. Genomic covariance analysis showed some inversion of the VU91 and UdeA_Afl2 chromosomes without significant expansion and contraction compared with TPD11 (Fig. S8).
Phylogenomic and evolutionary analysis of TPD11
The phylogenetic relationships of 16 species were determined by comparing 12839 single-copy homologous genes from 13 Aspergillus species and three outgroup strains (Fig. 3). The mean time of divergence of the Emydomyces outgroup, represented by Emydomyces testavorans, was estimated at 180 million years ago (MYA), whereas the time of divergence of the Trichocomaceae and Thermoascaceae outgroups was estimated at 126.4 MYA and 116.3 MYA, respectively. In the Aspergillaceae classification, Monascus differentiated with Aspergillus and Penicillium at 95.2 MAY; the mean differentiation time for Aspergillus with Penicillium was estimated to be 88 MAY. The strain in this study, A. tamarii TPD11, differentiated with A. pseudocaelatus differentiated at 4.4 MYA.
Gene family contraction occurred more frequently than gene family expansion during the evolution of the 16 fungal species (Fig. 3). Among Aspergillus strains, A. tamarii TPD11, A. pseudocaelatus, A. oryzae, and A. flavus had 181, 723, 498 and 232 gene family expansions, respectively, and 1413, 533, 1124, and 556 gene family contractions, respectively. Of these gene families, 54 (8 expanded and 46 contracted), 113 (all expanded), 148 (104 expanded and 44 contracted), and 107 (55 expanded and 52 contracted) in A. tamarii TPD11, A. pseudocaelatus, A. oryzae, and A. flavus respectively) underwent significant evolution (Fig. 3).
Analysis of the enzyme-producing capacity of TPD11
CAZymes function to degrade, modify, and generate glycosidic bonds. The genomes of 16 strains were annotated using the CAZy database, and a total of 838 CAZy genes were annotated in the genome of TPD11 strain (Fig. 4). These included 154 auxiliary activities (AAs),13 sugar binding modules (CBMs), 160 glycoesterases (CEs), 335 glycoside hydrolases (GHs), 143 glycosyltransferases (GTs), and 33 polysaccharide lyases (PLs). Clustering of CAZy genes of 16 strains showed higher number of GHs genes in all strains except Emydomyces. Talaromyces and Asspergillus had more CAZy genes as compared to Penicillium, Monascus, and Emydomyces. CAZy genes of TPD11 has 29 unique functions, and the major enzymes are amylase, glucosidase, pectin cleavage enzyme, chitin synthase, chitinase, keratinase, and chitosanase (Table S3).
Of the 11,305 protein-coding genes, a total of 1,118 proteins (9.89%) were predicted to have a signal peptide, and 9,031 proteins (79.89%) did not have a transmembrane structural domain. The 925 proteins (8.18%) with signal peptides and no transmembrane structural domains were secreted proteins. Database annotations indicated that these secreted proteins were mainly pectin cleaving enzymes, chitinases, chitosanases, extracellular cellulases, secreted lipases, cell wall dextranases, and glycoside hydrolases (Table S5).
Analysis of the CYP450 family of TPD11
CYP450, a ubiquitous superfamily of monooxygenases, is an essential enzyme involved in primary and secondary metabolic processes in fungi and plays an important role in catalyzing the synthesis of natural products such as terpenes, sterols, and alkaloids [24].
A total of 231 CYP450s were screened in the genome of strain TPD11 according to Pfam based on domain characterization (Fig. 5). The 162 CYP450s in A. oryzae and 157 CYP450s in A. flavus were used as reference sequences to cluster the 231 CYP450s of TPD11 to obtain the classification of CYP450s. A total of 181 CYP450s were classified into 76 subfamilies by clustering, and these CYP450s had high homology with CYP450s in A. oryzae and A. flavus. The CYP65 family was the most abundant (11), followed by CYP5289 (9) and CYP620 (8) (Table S6).
Secondary metabolite biosynthesis analysis of TPD11
To assess the biosynthetic potential of TPD11 secondary metabolites, secondary metabolic BGCs of TPD11 were predicted. antiSMASH predicted 49 BGCs distributed on 8 contigs, and the core genes of these BGCs mainly included 30 non-ribosomal peptide synthases (NRPS and NRPS-like), 13 polyketide synthases (PKs) and 2 terpene synthases (Table S7). Among them, some BGCs were involved in eupenifeldin (ctg1), serinocyclin A / serinocyclin B (ctg3), naphthopyrone (ctg4), clavaric acid (ctg5.1) and squalestatin S1 (ctg6) biosynthesis synthesis (Fig. 6A and Table S7). CYP450, which is involved in NRPS biosynthesis, was found in two BGCs (ctg5.2 and ctg8).
PPⅠ, PPⅥ, PPⅦ and PPH were detected in TPD11 fermentation broth, with the highest content of PPⅥ up to 0.283 ± 0.012 mg/mL (Fig. 6B). The contents of PPI, PPH and PPⅦ were 0.048 ± 0.002, 0.035 ± 0.027 and 0.031 ± 0.021 mg/mL, respectively. The results indicated that TPD11 had a relative advantage in the biosynthesis of polyphyllins in the main production of PPⅥ.
Terpene skeleton and steroid biosynthesis are required for the synthesis of polyphyllin. A large number of terpene skeleton biosynthesis and steroid biosynthesis homologous genes were annotated in TPD11, with a total of 13 genes involved in farnesyl pyrophosphate biosynthesis (Fig. S9), and 29 genes involved in cholesterol biosynthesis (Fig. S10). The pathway for terpene skeleton biosynthesis of TPD11 is MVA. UGTs for glycosylation is necessary for TPD11 to convert steroidal saponin elements into various types of polyphyllins. Eleven UGTs were predicted from the TPD11 genome by constructing Hidden Markov Model (Fig. 6C). These 11 UGTs were categorized into five classes by constructing a phylogenetic tree, with class V and Ⅳ having the most 5 UGTs and 3 UGTs, respectively.