Abstract
Torreya plants produce dry fruits with assorted functions. Here, we report the 19-Gb chromosome-level genome assembly of T. grandis. The genome is shaped by ancient whole-genome duplications and recurrent LTR retrotransposon bursts. Comparative genomic analyses reveal key genes involved in reproductive organ development, cell wall biosynthesis and seed storage. Two genes encoding a C18 Δ9-elongase and a C20 Δ5-desaturase are identified to be responsible for sciadonic acid biosynthesis and both are present in diverse plant lineages except angiosperms. We demonstrate that the histidine-rich boxes of the Δ5-desaturase are crucial for its catalytic activity. Methylome analysis reveals that methylation valleys of the T. grandis seed genome harbor genes associated with important seed activities, including cell wall and lipid biosynthesis. Moreover, seed development is accompanied by DNA methylation changes that possibly fuel energy production. This study provides important genomic resources and elucidates the evolutionary mechanism of sciadonic acid biosynthesis in land plants.
Similar content being viewed by others
Introduction
The emergence of seed plants consisting of angiosperms and gymnosperms marked a momentous event in the evolution of land plants and the change of earth environments. Angiosperms and gymnosperms diverged in the Lower Mississippian1, followed by rapid radiation of flowering plants resulting in approximately 352,000 extant species on Earth compared to only 1000 species of gymnosperms. There is a variety of morphological/anatomical diversity and metabolic versatility between angiosperms and gymnosperms, but the underlying genetic and biochemical mechanisms are largely elusive2,3.
Torreya grandis, a gymnosperm species belonging to a small genus of the yew family (Taxaceae), is a useful multipurpose tree, providing timber, medicine, edible seeds and oil4 (Fig. 1a). The first credible record of T. grandis as a medicinal source appears in Classic of the Materia Medica during the Three Kingdoms of China and dates back to the beginning of the 3rd century AD5. T. grandis is the only species in Taxaceae with edible seeds, which have been used as food for thousands of years in China due to their unique flavor and beneficial components5,6. Oils are enriched in seeds of T. grandis with an average content of 45.80–53.16%7. Sciadonic acid (SCA), a non-methylene-interrupted ω6 fatty acid, has been found as one of the major components in fatty acid composition of the kernel oil7. SCA has positive effects on human health, and functions in reducing inflammation, lowering triglycerides, preventing blood clots and regulating lipid metabolism8,9,10. Production of SCA has been detected in different lineages of gymnosperms and a handful of algae and ferns11. However, SCA is generally absent in flowering plants, with the exception of a few lower eudicots (e.g., Ranunculaceae)12, thus leaving a puzzle on its origin and evolution in green plants.
a Tree and fruit set of T. grandis. The lower panel shows the processed dry seed and its edible part (endosperm). b Circos plot of T. grandis genome and genomic features encoded by the chromosomes. Each feature was calculated based on a 10-Mb window across the chromosomes. Colored stars indicate the presence of telomeric sequences on 5’- (green) or 3’-end (orange) of the chromosome. c Distribution of LTR-RT insertion time. The left panel shows all members of Gypsy and Copia families and the top six most abundant subfamilies are shown on the right panel. d Ks distribution of orthologues among T. grandis, Sequoiadendron giganteum, Ginkgo biloba and Gnetum montanum. Ks of paralogues in T. grandis was fitted with Guassian mixture model and the putative ancient WGD is indicated. e Micro-collinearity between genomes of T. grandis, G. biloba and G. montanum. Source data are provided as a Source Data file.
Genome sequences are key to addressing critical questions of plant evolution. Analyses of genomes of representative gymnosperms have shown unique aspects of gene and genome evolution that distinguish themselves from flowering plants13,14,15,16,17,18,19. However, understanding the biological and evolutionary mechanisms of phenotypic diversity between flowering and non-flowering plants is still challenging, partly due to the limited availability of genomic resources of gymnosperms.
In this study, we assemble a chromosome-scale reference genome for T. grandis, accompanied by transcriptome and methylome profiling in multiple tissues. Our data, through comparative genomic analyses, unravel interesting footprints associated with morphological diversity of major land plant lineages, and discover and validate two key enzymes that are responsible for SCA biosynthesis. Information provided by this work will be useful for strategic design on improvement of SCA production, and promote the utilization of Torreya genetic resources.
Results
Genome assembly and annotation
We generated a total of 1.93 Tb of Illumina and 463.7 Gb of PacBio HiFi reads for T. grandis (Supplementary Data 1), representing 96.5× and 23.2× coverage, respectively, of the T. grandis genome that had an estimated size of ~20 Gb according to the k-mer analysis of the Illumina reads (Supplementary Fig. 1). The final assembly had a size of 19,050,820,213 bp, comprising 11,811 contigs with an N50 size of 2.82 Mb (Supplementary Table 1). Using Hi-C reads of approximately 106.2× coverage, 18.87 Gb (99.1%) of the assembled contigs were grouped into 11 chromosomes (Fig. 1b and Supplementary Fig. 2). All 11 chromosomes were found to be enriched with the 101-bp repetitive sequence unit that resembles the tandem centromeric satellite repeat known as the landmark of centromeres, while 9 chromosomes harbored telomeric sequences (5’-TTTAGGG-3’) in at least one end (Fig. 1b). Assessment of the T. grandis genome using Merqury20 revealed a consensus quality score of 46.9, equivalent to a base accuracy of 99.998%. BUSCO21 evaluation indicated that 1386 out of 1614 land plant conserved orthologues were successfully captured by the T. grandis assembly, which was comparable to that of other gymnosperm genome assemblies (Supplementary Table 2). The LTR assembly index (LAI) for the T. grandis genome was 10.7, which was higher than the proposed standard for a reference genome22. These together with the high DNA (99.48%) and RNA (up to 97.5%) read mapping rates suggested the high quality of the T. grandis genome assembly.
The T. grandis genome assembly harbored 11.4 Gb (59.8%) of repetitive sequences, of which LTR retrotransposons (LTR-RTs; 87.0%) were predominant, followed by DNA transposons (7.1%) and long interspersed nuclear elements (LINEs; 3.1%) (Supplementary Data 2). The proportion of Copia LTR-RTs (11.6%) was relatively higher in T. grandis than in other gymnosperms, possibly due to recent species-specific bursts occurring in multiple subfamilies of LTR-RTs (Fig. 1c). Most of the LTR-RT expansions in gymnosperms took place between 25-7 million years ago (mya; Supplementary Fig. 3a), overlapping with the geological time of Miocene epoch (23.03–5.33 mya) when the earth cooled down towards ice ages23, suggesting a potential environmental effect on the genome size evolution of gymnosperms.
A total of 47,089 protein-coding genes were predicted in the T. grandis genome, of which 46,338 were supported by homology and/or transcriptome evidence (Supplementary Table 1). Intron size is more variable in gymnosperms than in angiosperms (Supplementary Fig. 3b), which is attributed to the expansion of LTR-RTs. In plants, LTR-RTs can be eliminated through unequal recombination, creating solo-LTRs in the genome. The solo:intact LTR ratio is high in T. grandis (4.3) and other gymnosperms including Taxus wallichiana (5.5)18, Ginkgo biloba (4.26), Welwitschia mirabilis (3.87), and Gnetum montanum (2.07)16. Since gymnosperm genomes are enriched with ancient LTR-RTs (10–30 mya)14,15,16,17,18,19, we hypothesize that elimination of ancient LTR-RTs without recent expansion may have contributed to their high solo:intact LTR ratios. This is in contrast with the small-genome angiosperms, in which LTR-RT bursts are more recent (<4 mya)24. Epigenetic silencing of transposons and pericentromeric repeats is mediated by RNA-directed DNA methylation (RdDM) and 24-nt hetsiRNAs25. The T. grandis genome encoded homologs of key components of the RdDM pathway (Supplementary Data 3); however, small RNA profiling of seven tissues showed that 21-nt sRNAs were the most abundant in T. grandis, contrasting to the most abundant 24-nt sRNAs in angiosperms, while productions of 22-nt and 24-nt sRNAs were tissue-specific (Supplementary Fig. 4). This pattern is similar to that found in conifers13,26 and Welwitschia mirabilis16. Nonetheless, further extensive sampling from additional tissues and stages would provide deeper insights into the divergence of sRNA processing between gymnosperms and angiosperms.
Ancient whole-genome duplication
Whole-genome duplications (WGDs) have occurred across the breadth of eukaryote phylogeny27. In gymnosperms, several WGDs have been recognized although some of them remain in controversy16,17,18,19,28. The Ks distribution of 3859 paralogue groups within T. grandis indicated the absence of recent WGDs. However, we observed a peak of Ks ranging from 1 to 2 and a summit at 1.4, representing a potential ancient WGD that occurred in the common ancestor of conifers and ginkgophytes, a lineage diverged from gnetophytes (Fig. 1d). We then used a tree-based approach29, which calculates the frequency of gene duplication on every branch of a phylogeny by reconciliation of gene tree and species tree, to cross validate the WGD event. Analysis of 19,649 gene trees from eight selected species led to the discovery of three ancient WGD signals including two (zeta and omega) reported previously17,28 and one that was consistent with the Ks analysis (Supplementary Fig. 5). Whole-genome comparison showed high collinearity among genomes of T. grandis and two evolutionarily distant gymnosperms, Sequoiadendron giganteum and Ginkgo biloba (Supplementary Fig. 6), and also revealed traces of collinear blocks that were duplicated in both T. grandis and G. biloba but not in Gnetum montanum, agreeing with the timing at which the newly discovered WGD occurred (Fig. 1e).
Gene family evolution
We identified 19,362 orthologous groups (gene families) in 19 plant species comprising 7 gymnosperms and 12 representative species in major green plant lineages. Phylogeny and molecular dating using 219 low-copy gene families indicated that T. grandis separated from T. wallichiana around 68.5 mya (Fig. 2a). Expansion of gene families has been implicated in tight associations with morphological innovations30,31. Through reconstruction of gene family evolution, we found that bursts of gene family expansions coincided with major transitions of plant adaptation (Fig. 2a). Massive gene family expansion (n = 417, P < 0.05) was observed in the common ancestor of land plants, and subsequently in the extinct ancestors leading to seed plants (n = 575), angiosperms (n = 432) and various lineages of gymnosperms (n = 428–818). Functions of expanded gene families were mostly associated with plant organ development, response to biotic (e.g., bacteria and fungi) and abiotic (e.g., water deprivation, light, temperature, and salt) stresses, and biosynthesis and signaling of plant hormones (Supplementary Data 4). Many of the gene families were expanded continuously towards the evolution of higher plants, suggesting that gene duplication, possibly followed by sub/neo-functionalization, provides genetic foundation for morphological diversity and environmental adaptation in plants. Among the gene families that were significantly expanded in T. grandis, many of them encoded pfam domains associated with important biological functions, including lipid transfer (Oleosin and PF14368), biotic and abiotic stress responses (PF00201 and PF03018) and secondary metabolism (PF00067) (Supplementary Data 5). The T. grandis genome lacked the orthologues of taxadiene synthase, a core component of paclitaxel biosynthesis, explaining the absence of paclitaxel and relevant metabolites in this species.
a Gene family expansion and contraction during the evolution of green plants. The maximum likelihood phylogeny was built with 219 low-copy orthologous groups. Gene family analysis was started with 10,345 orthologous groups that were shared by the most recent common ancestor (MRCA) of green plants. Numbers on branches are the sizes of expanded (blue) and contracted (red) gene families at each node. Colored pies on the right represent the sizes of expanded/contracted gene families as well as gained/lost genes for each leaf node of the tree. b Expression of MIKCC type MADS-box genes in vegetative and reproductive tissues of T. grandis. c Proposed reproductive organ identity genes in T. grandis. The AP3/PI-like genes (TG7g01668 and TG7g01669) and TG8g01565 were predominantly expressed in male and female cones, respectively. The AG-like (TG10g01848), AGL6-like (TG2g00325) and AP1/SEP-like (TG4g01441) genes were expressed in both female and male cones, with the first two showing a pattern biased to female cones. d Maximum likelihood phylogeny showing the bacterial origin of T. grandis genes. Blue pies indicate bootstrap support greater than 80% at the corresponding branches. e Expression of putative horizontally transferred genes in different tissues. Source data are provided as a Source Data file.
Reproductive organ genes
Gymnosperms have unenclosed or naked seeds on the surface of scales or leaves, while flowers and fruits are angiosperm innovations. Phylogeny-based homolog search using well-studied flower development genes32 showed sporadic distribution of these homologs in gymnosperms and non-seed plants (Supplementary Data 6), indicating the stepwise emergence accompanied with secondary loss of flower development genes during the evolution of land plants, as exemplified by NOP10 (required for female gametophyte formation in flowers)33 and WUS (required for shoot and floral meristem integrity)34 genes that emerged early in land plants and were subsequently lost in both T. grandis and T. wallichiana (Supplementary Data 6).
The MADS-box family genes are a class of transcription factors involved in the regulation of floral organ specificity, flowering time, and fruit development. We identified 23 MIKCC MADS-box genes in T. grandis, including homologs of genes in the ABCE model of floral organ identity35. These included one AP1/SEP-like gene (A or E function), two AP3/PI-like genes (B function) and six AG-like genes (C function) (Supplementary Fig. 7). Transcriptome analysis of 18 samples from vegetative and reproductive organs revealed six MADS-box genes that were highly expressed in male and/or female cones of T. grandis, among which the two tandemly duplicated AP3/PI-like genes (TG7g01668 and TG7g01669) were predominantly expressed in the male cones, while one AG-like gene (TG10g01848) was expressed in male but upregulated by 6.6-fold in the female cones (Fig. 2b). Recent studies suggest that AGL6, member of an ancient subfamily of MADS-box genes, is involved in the E function of floral development in rice, maize and wheat36,37, while participates in A function in the basal angiosperm Nymphaea colorata38. In T. grandis, the AGL6-like gene (TG2g00325) showed an expression pattern similar to that of the C function genes, whereas the AP1/SEP-like gene (TG4g01441) was expressed at a moderately high level in both male and female cones, resembling an ancestral role of E function. Interestingly, the most highly expressed MADS-box gene (TG8g01565) was exclusively activated in the female cones (Fig. 2b). This gene was phylogenetically clustered with B clade genes comprising AP3, PI and B-sister genes TT16 and GOA (Supplementary Fig. 7); however, its expression pattern was opposite to that of the AP3/PI-like genes. In conclusion, our finding on the involvement of additional MADS-box genes in seed development of gymnosperms supports the basic “BC” model, where the C function genes are generally expressed in reproductive male and female organs and the B function genes are restricted to male reproductive organs39, and suggests a more sophisticated regulatory system for reproductive organ development in gymnosperms (Fig. 2c).
Seed storage proteins
The protein content of T. grandis seeds ranges from 10.34% to 16.43% depending on cultivars7. Genes encoding seed storage proteins (SSPs) including 2S albumins (n = 0–7), 7S globulins (n = 1–9) and 11S globulins (n = 2–14) were identified in T. grandis and other gymnosperms but not in the earlier forms of plants (Supplementary Data 7), suggesting their origin in seed plants. Transcriptome analysis showed that genes encoding 2S albumins and 7S globulins were expressed at an exceptionally high level (average transcripts per million (TPM) = 14,125) in the kernel of T. grandis seeds and the expression was increased during seed development (Supplementary Fig. 8a). In contrast, all SSP genes including 11S globulin genes, which were moderately expressed in the kernel, remained transcriptionally inactive in the vegetative tissues (Supplementary Fig. 8a). The 2S albumin proteins harbor numerous cysteine residues to form disulfide bridge within and between subunits40. We found that all these residues were conserved in T. grandis although the whole protein sequences were considerably divergent from the angiosperm counterparts (Supplementary Fig. 8b). Homology modeling revealed a high degree of protein structure conservation between the 2S albumin proteins from T. grandis (e.g., TG11g02972) and sunflower, particularly in the region where α-helices are formed (Supplementary Fig. 8c). Likewise, most of residues involved in trimer formation and stabilization as well as in correct globular folding of 11S globulins from flowering plants41 were conserved in T. grandis (Supplementary Fig. 9). Overall, the gene expression and structural analyses suggest a conservative role of the major SSPs in both gymnosperms and angiosperms.
Cell wall biosynthesis and horizontal gene transfer
Gymnosperms are mainly woody plants, and their genomes encode a large set of carbohydrate active enzymes (CAZymes) whose functions are closely associated with cell wall biosynthesis. Among the 19 selected representative plant species, T. grandis harbored more CAZymes than most others, particularly in the families of glycoside hydrolases (e.g., GH1, GH16, GH18, GH19, GH27, GH71, GH99, and GH152), GT61 glycosyltransferases, and PL1 polysaccharide lyases (Supplementary Data 7), many of which were also expanded in other gymnosperms. In contrast to most CAZyme families that were universally present in plants, we identified four families comprising 18 genes, GH71 (n = 7), GH99 (n = 9), GH103 (n = 1), and CE4 (n = 1), that were present only in gymnosperms and prior lineages but not in angiosperms (Supplementary Data 8). Phylogenetic analysis showed that these families were of possible bacterial origin (Fig. 2d and Supplementary Fig. 10). Through systematic analysis, we identified 14 additional T. grandis genes that were derived from horizontal gene transfers (HGTs; Supplementary Table 3). Most of these genes were expressed in different tissues of the plant (Fig. 2e), reinforcing the contribution of HGTs in the evolution of land plants42.
Lignin is a major component of plant secondary cell wall and is derived from p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S) monolignols. S-lignin is restricted to flowering plants and some lycophytes, whereas G- and H-lignin are fundamental to all vascular plants2. Consistently, two key genes for S-lignin biosynthesis, F5H and COMT, were found only in angiosperms but not in gymnosperms. Unlike angiosperms in which vessels comprise major water-conducting elements in xylem43, gymnosperm woods are mainly composed of tracheids2. Vessel differentiation is regulated by VASCULAR-RELATED MAC-DOMAIN (VND) proteins44, while fiber development is associated with NAC SECONDARY WALL THICKENING PROMOTING FACTOR (NST)/SECONDARY WALL-ASSOCIATED NAC DOMAIN (SND) proteins45. The T. grandis genome encoded genes homologous to VND4/5/6, but lacked homologs of VND1/2/3, NST and SND1 (Supplementary Fig. 11), which, combined with the finding of divergent regulatory networks of VND/NST homologs in conifer and flowering plants during wood formation46, suggests a close link between vessel formation and the emergence of master NAC transcription factors as well as their regulatory networks in angiosperms.
Discovery of key genes for sciadonic acid biosynthesis
Sciadonic acid (SCA) is a Δ5-olefinic fatty acid and its biosynthesis requires the activity of C18 Δ9-elongase and C20 Δ5-desaturase that uses 18:2-phosphatidylcholine (PC) as the initial substrate (Fig. 3a). Δ5-desaturases are known as the ‘front-end’ desaturases47, which usually encode a cytochrome b5-like heme/steroid binding domain (PF00173) and a fatty acid desaturase domain (PF00487), whereas the Δ9-elongases encode a GNS1/SUR4 family domain (PF01151) for long chain fatty acid elongation. The T. grandis genome encoded four desaturase genes and four elongase genes based on the domain search. However, only one desaturase (TgDES1) showed high similarity with the previously reported Δ5-desaturase in Anemone leveillei48, while two elongases were considered as putative Δ9-elongases but only one (TgELO1) was highly expressed in seed kernels (Supplementary Fig. 12). Since the unsaturated fatty acids are abundant components of seed oils, we investigated the expression of TgDES1 and TgELO1 during seed maturation. We found that SCA was accumulated in mature seeds, accompanied with the increased expression of TgDES1. Similar trend was observed for the expression of TgELO1 and the content of its putative product cis-11,14-eicosadienoic acid (Fig. 3b, c). Study of subcellular localization showed that both TgELO1 and TgDES1 were co-localized with the marker of endoplasmic reticulum (ER) in N. benthamiana leaves (Fig. 3d), suggesting that they were bound to ER membrane, consistent with the subcellular location of known desaturases and elongases49. To further verify their function in SCA biosynthesis, we overexpressed both TgELO1 and TgDES1 in A. thaliana, which neither encodes orthologues of TgELO1 and TgDES1 nor produces SCA or its precursor 20:2Δ11,14-PC. Gas chromatography analysis showed that SCA was successfully synthesized in seeds of the transgenic line expressing TgDES1 and TgELO1, demonstrating that TgELO1 and TgDES1 are capable of synthesizing SCA in T. grandis (Fig. 3e).
a Overview of the fatty acid biosynthetic pathway. PDH pyruvate dehydrogenase, CT carboxyltransferase, BC biotin carboxylase, BCCP biotin carboxyl carrier protein, MCMT malonyl-CoA:ACP malonyltransferase, ACP acyl carrier protein, KAS ketoacyl-ACP synthase, SAD stearoyl-ACP desaturase, FATA acyl-ACP thioesterase A, FATB acyl-ACP thioesterase B, LACS long-chain acyl-CoA synthetase, DGAT diacylglycerol acyltransferase, PDAT phospholipid:diacylglycerol acyltransferase, PAP phosphatidic acid phosphatase, LPAT lysophosphatidic acid acyltransferase, GPAT glycerol-3-phosphate acyltransferase, CPT cholinephosphotransferase, FAD2 oleate desaturase, FAD3 linoleate desaturase, PC phosphatidyl choline. b TgDES1 expression and SCA content in seeds from early development stage (May) to maturation stage (September). Different letters on the bars indicate statistical significance between samples at α = 0.05 (one-way ANOVA and Tukey’s test). Measurements were performed in three biological replicates and data are presented as mean + SD. c Expression of TgELO1 and the content of its product cis-11,14-Eicosadienoic acid in seeds. Different letters on the bars indicate statistical significance between samples at α = 0.05 (one-way ANOVA and Tukey’s test). Measurements were performed in three biological replicates and data are presented as mean + SD. d Subcellular localization of TgDES1 and TgELO1 in N. benthamiana leaves. e Detection of SCA and its precursor in Arabidopsis Col-0 and the transgenic line overexpressing both TgDES1 and TgELO1. Source data are provided as a Source Data file.
Origin and evolution of plant Δ5-desaturases and Δ9-elongases
Phylogenetic analysis of desaturases in green plants (Viridiplantae) showed that TgDES1 clustered with desaturases exclusively from non-angiosperm organisms, and this monophyletic clade was close to the family containing sphingolipid desaturases including AtSLDs from Arabidopsis (Fig. 4a). Interestingly, the TgDES1 clade clearly separated from the group harboring AL10 and AL21, two proteins that were found to be responsible for SCA biosynthesis in the basal eudicot Anemone leveillei48. Structure modeling of TgDES1, AtSLD2 and AL21 showed overall similar structures between TgDES1 and AtSLD2, particularly in the region where the active center was formed, whereas structure of AL21 was relatively diverged from TgDES1 (Fig. 4b). Since flowering plants rarely synthesize SCA, our phylogenetic and structural evidence suggested that this is possibly due to the loss of TgDES1 clade desaturases, while the ability of SCA biosynthesis in particular species of eudicots was largely attributed to the secondary gain of the Δ5-desaturase activity of evolutionarily independent counterparts. Similarly, close homologs of TgELO1 were not found in flowering plants but present in early land plants and algae, suggesting the co-evolution of Δ5-desaturase and Δ9-elongase in plants (Supplementary Fig. 13).
a Maximum likelihood phylogeny of plant desaturases. TgDES1 is clustered within a group (clade 1) close to a sister clade (clade 2) comprising Δ6- and Δ8- desaturases. b Structure modeling of TgDES1 and the desaturases from Arabidopsis (AtSLD2) and Anemone leveillei (AL21). Protein structures were modeled with AlphaFold2 and the bioactive center of each protein comprising three histidine-rich motifs is marked with yellow. c Comparison of conserved motifs in different desaturase groups showing in a. d Detection of SCA and its precursor in N. benthamiana leaves expressing Arabidopsis AtSLD2 with conserved histidine-rich motifs (motif2 and motif3) replaced by those from TgDES1 of T. grandis. AtSLD2, AtSLD2-Motif2, and AtSLD2-Motif3 are lines harboring Arabidopsis wild-type AtSLD2 gene, AtSLD2 with motif2 from TgDES1, and AtSLD2 with motif3 from TgDES1, respectively.
Characterization of protein sequences revealed the conservation of an N-terminal cytochrome b5-like domain and three histidine-rich boxes of TgDES1 clade desaturases (clade 1) and their two closely related groups (group 1 and group 2 of clade 2), whereas striking variation was observed in the first two histidine-rich boxes among different groups (Fig. 4c). A previous study reported that site-directed substitution of histidine boxes could influence substrate chain-length specificity and selectivity50. The single amino acid substitution likely directs the outcome of the desaturation reaction by modulating the distance between substrate fatty acyl carbon atoms and active center metal ions51. To test whether sequence variation of histidine-rich domains determined substrate specificity that led to the success of SCA biosynthesis, we replaced the histidine-rich domain of Arabidopsis desaturase AtSLD2 with that of TgDES1, and transiently expressed the construct in N. benthamiana leaves. We noted that TgELO1 was not coexpressed with the engineered desaturase gene because 20:2Δ11,14-PC, the product of Δ9-elongase catalysis, could be detected in leaves of the wild-type tobacco. SCA was undetectable in N. benthamiana leaves expressing wild-type AtSLD2; however, switch of either of the two histidine-rich boxes from TgDES1 was sufficient to synthesize SCA in N. benthamiana leaves (Fig. 4d). Taken together, our data suggest that mutations in these two histidine-rich motifs of desaturases have led to the alternation of substrate specificity and consequently the evolution of specific clade for SCA biosynthesis, loss of which marks the significant metabolic diversity between gymnosperms and angiosperms.
Dynamics of DNA methylation during seed development
Seed development in gymnosperms is a long process spanning multiple years3. To understand whether and how DNA methylation participates in seed development of T. grandis, as is evident in flowering plants52, we profiled seed methylomes at three developmental stages (Fig. 5a; Supplementary Data 9). Genes involved in DNA methylation of all three cytosine contexts (CG, CHG, CHH) were identified in the T. grandis genome (Supplementary Data 3). The global average methylation levels of mCG, mCHG, mCHH in T. grandis seed genome were 83%, 69% and 4%, respectively. Both mCG and mCHG methylation levels were higher than those in most of previously studied angiosperms53, coinciding with the proposal of positive correlation between genome sizes and mCG/mCHG methylation levels54. mC of all sequence contexts was enriched at centromeric and peri-centromeric regions, despite that both mCG and mCHG were also broadly distributed in chromosome arms (Fig. 1b). In flowering plants, exons of genes are sometimes enriched with mCG but depleted with both mCHG and mCHH, which is referred to as gene body methylation (gbM)55. We observed the enrichment of mCG and depletion of mCHH in T. grandis genes; however, enrichment of mCHG was also found in transcribed regions (Fig. 5b and Supplementary Fig. 14a, b), which is similar to the pattern found in conifers56. GbM has been proposed to regulate gene transcription55. We observed a clear enrichment of mCG instead of mCHG/mCHH on moderately expressed genes, for which the expression was positively correlated with methylation levels (Fig. 5c and Supplementary Fig. 14c), indicating functional conservation of gbM in the sister lineage of angiosperms. Evolution of gbM is hypothesized to be associated with DNA methylation silencing of TEs in proximity of genes55. Consistently, we found that LTR-RTs, which were the major component of TEs in gene regions, were highly methylated (Fig. 5d), and that genes with TE insertions had both higher expression and CG methylation than those without TEs (Supplementary Fig. 15).
a Sampled seeds for methylome profiling. Pictures show the outer side and inner side (through longitudinal section) of the seeds. b Methylation levels of exon, intron and gene flanking regions in seeds. c Methylation levels at three cytosine contexts on exonic region of genes. Genes are categorized into 20 groups based on ordered expression levels. For each group, the median value of gene expression and the average methylation level across all exonic regions of genes are recorded. d Methylation levels of intact LTR-RTs in the T. grandis genome. e GO terms enriched in genes overlapping with demethylation valleys shared by seeds of all three developmental stages. GO terms with adjusted P value < 0.05 (two-sided Fisher’s exact test with Benjamini–Hochberg correction) are plotted and sizes of GO terms in the word cloud figure correlate with their statistical significance. f View of methylation levels of selected genes overlapping with demethylation valleys. CES cellulose synthase, PE pectinesterase. g Comparison of mCG and mCHG methylation levels in different genomic regions of seeds at three stages. Source data are provided as a Source Data file.
Seed development and germination genes are frequently localized within demethylation valleys (DMVs), where methylation level was low (e.g., <5%) for any of the cytosine contexts57. We identified 5099 common DMVs in the seed genome of the three samples, which spanned 30 Mb including the largest interval extending to 144 kb. The DMVs intersected with 4200 protein-coding genes, many of which encoded important classes of seed proteins, such as storage proteins, transcriptional factors, and enzymes for cell wall modification, hormone homeostasis and fatty acid biosynthesis (Fig. 5e, f). T. grandis seed is coated with a specialized outgrowth, called aril (Fig. 5a). During development, seed coat develops heavily lignified secondary cell walls to reinforce the outer surface of the seed58, while preserves a soft inner one that directly surrounds the endosperm. Consistently, genes encoding laccases (n = 38), which function in cell wall lignification59, and expansins (n = 13) that are associated with cell wall loosening60, were frequently found in DMVs (Fig. 5f). Notably, 18% of T. grandis transcription factor (TF) genes (n = 370) were located within seed DMV regions, representing a significant enrichment (χ2 test; P < 0.0001; Supplementary Data 10). These TFs belonged to diverse gene families but were particularly abundant in MYB, NAC and AP2 families, which are known to regulate plant growth and development. The methylation of mCHH varied more remarkably than that of mCG and mCHG during seed development (Fig. 5g and Supplementary Fig. 16). We identified differentially methylated regions (DMRs) for each of the three cytosine contexts (Supplementary Data 11–13). Among genes overlapping with DMRs, 12% of them were differentially expressed, suggesting the translation of epigenetic variation to gene expression flexibility during seed development (Supplementary Fig. 17). GO enrichment analysis showed that DMR-associated genes were mainly enriched with those involved in photosynthesis and secondary metabolism (Supplementary Data 14), in line with the fact that photosynthesis fuels energy-generating biochemical pathways by contributing oxygen to seed tissues during the development of green seeds as developing seeds suffer from limited penetration of oxygen, particularly into inner tissues61.
Discussion
Gymnosperms are considered as a treasure trove of life history on the earth. Here, we assembled a chromosome-level reference genome for the gymnosperm species T. grandis. The genome size is huge and much larger than most plant species ever sequenced. Based on this assembly and analysis of multi-omics data, we conclude that (1) accumulation of ancient LTR-RTs contributes to the bloating of the T. grandis genome, whereas T. grandis counteracts TE expansion through unequal recombination and epigenetic silencing with a mechanism potentially different from angiosperms; (2) gain or loss of important gene families in T. grandis, e.g., those involved in cell wall activities and paclitaxel biosynthesis, underly its phenotypic diversity, and MADS-box genes associated with reproductive organ identity include not only classical B- and C-function genes that are already proposed in previous studies35,36,37,38,39 but also additional genes (e.g., TG8g01565) that show expression pattern different from B- and C-function genes; (3) the Δ9-elongase and Δ5-desaturase are capable of synthesizing SCA, and these two enzymes have co-evolved and have been lost in flowering plants; moreover, substrate specificity of Δ5-desaturase is determined by the two histidine-rich boxes, mutation on which may lead to the alternation of substrate recognition, and subsequently the change of its product; (4) the seed genome of T. grandis comprises both heavily methylated repeat sequences and demethylation valleys, latter of which intersect with genes exerting important seed functions such as cell wall modification and fatty acid biosynthesis, as well as regulation of gene expression and hormone homeostasis. Overall, our high-quality reference genome coupled with comparative and functional genomic analyses provide insights into the gymnosperm biology, particularly in the biosynthesis and evolution of SCA that features metabolic versatility between the major land plant lineages.
Methods
Plant materials and sequencing
Young leaves from a plant of T. grandis grown in Shaoxing, China, were collected in March 2018 and used for DNA extraction following the CTAB (2%) method62. A paired-end (PE) library with insertion size of 350 bp was constructed using the Illumina Genomic DNA Sample Preparation kit following the manufacturer’s instructions (Illumina), and sequenced on an Illumina NovaSeq system with a read length of 150 bp. A PacBio SMRTbell library was constructed using SMRTbell Express Template Prep Kit 2.0 and sequenced on a PacBio Sequel II platform. The circular consensus reads (HiFi reads) were generated using ccs software (https://github.com/pacificbiosciences/unanimity/) with parameter ‘-minPasses 3’. Hi-C library preparation and sequencing were performed by Novogene (Tianjin, China) following a protocol described elsewhere63. Briefly, libraries were prepared using leaf tissues fixed in 2% formaldehyde. Nuclei were extracted and permeabilized, and the chromatin was digested with DpnII restriction enzyme (NEB). The digested chromatin was blunt-ended and labeled with biotin. DNA ligation was performed using T4 DNA ligase (NEB), after which proteinase K was added for reverse crosslinking. The DNA fragments were then purified and sequenced on an Illumina NovaSeq platform with a read length of 2 × 150 bp.
To assist gene prediction, transcriptome sequencing was performed for samples collected from leaf, root, stem, young seed, aril, seed coat and kernel tissues of the same plant (Supplementary Data 1). Total RNA was extracted using TRIzol Reagent (Invitrogen) and quantified with NanoDrop ND-2000 spectrophotometer (NanoDrop Technologies). mRNA purified from total RNA with a RIN score ≥8 (Bioanalyzer 2100, Agilent Technologies) was used for library construction with the NEBNext Ultra II RNA Library Prep Kit for Illumina (NEB) following the manufacturer’s instructions. The non-stranded RNA-Seq libraries were sequenced on an Illumina NovaSeq platform under 2 × 150-bp mode. For PacBio Iso-seq, total RNA from leaf, root, stem, aril and kernel tissues were pooled equally and cDNA was synthesized using the SMARTer PCR cDNA Synthesis Kit (Clontech). Size fractionation and selection (1–2, 2–3 and 3–6 kb) were performed using the BluePippin Size Selection System (Sage Science). The SMRT libraries were generated using the SMRTbell Template Prep Kit 1.0 (Pacific Biosciences) and sequenced on the PacBio RSII platform.
Genome assembly and quality assessment
The HiFi reads were assembled using hifiasm64 (version 0.8-dirty-r280) with default parameters and the assembled contigs were further polished by Racon (https://github.com/lbcb-sci/racon; v1.4.13) with Illumina reads. Purge Haplotigs65 (version v1.1.0) was used to filter out redundant sequences in the assembly with parameters ‘-l 15 -m 70 -h 125’ for the ‘contigcov’ subcommand and ‘-a 55’ for the ‘purge’ subcommand. Illumina reads from Hi-C libraries were processed with Trimmomatic66 (v0.36) to remove adaptors and low-quality sequences. The cleaned reads were analyzed by HiCUP (https://www.bioinformatics.babraham.ac.uk/projects/hicup/) to identify non-duplicated valid alignments, which were then used for scaffolding with ALLHiC67 (version 0.9.8). The initial scaffolding was manually curated using Juicebox (https://github.com/aidenlab/Juicebox). Completeness of the assembly was evaluated using the Illumina sequencing reads, which were mapped to the genome assembly using BWA-MEM68.
Repeat annotation and LTR insertion time estimation
Repetitive sequences were identified using a combination of homology-based and de novo predictions. A species-specific TE library for T. grandis was constructed to include LTR retrotransposons (LTR-RTs) and other TE elements identified by LTR_Finder69 and RepeatModeler70, respectively. This library was then combined with the Repbase library71 for TE identification by RepeatMasker72 (v.4.0.7). Repetitive elements were also predicted by RepeatProteinMask and the tandem repetitive sequences were identified by the TRF program73. To estimate LTR-RT insertion times, intact LTR-RTs were searched by LTR_Finder and LTR-harvest74. MUSCLE75 was used to align LTR sequences of intact LTR-RTs, and the nucleotide distance (K) between them was calculated with the Kimura two-parameter criterion using the distmat program in the EMBOSS package (http://emboss.sourceforge.net). The insertion time (T) was calculated as
where the rate of nucleotide substitution (r) used for gymnosperm species was 2.2 × 10−9 per base per year11. The putative centromeric repeats were determined based on the copy number and chromosomal distribution of the tandem repeats identified by TRF.
Gene prediction and gene set assessment
Protein-coding genes were predicted using repeat-masked genome sequences. For homology-based prediction, protein sequences from one moss (Physcomitrella patens), one fern (Selaginella moellendorffii), seven angiosperms (Amborella trichopoda, Arabidopsis thaliana, Oryza sativa, Phalaenopsis equestris, Populus trichocarpa, Vitis vinifera and Zea mays), and four gymnosperms (Ginkgo biloba, Gnetum montanum, Picea abies, and Pinus taeda) were aligned to the T. grandis genome using TBLASTN76 with an e-value cutoff of 1E−5. GenBlastA77 was then applied to cluster adjacent high-scoring pairs from the same protein alignments, and the corresponding gene structures were identified with GeneWise78 (v.2.4.1). Raw RNA-Seq reads were cleaned with Trimmomatic66 (v0.36) and mapped to the T. grandis genome using TopHat279. Subsequently, Cufflinks80 (v.2.2.1) was employed to predict gene models. The cleaned RNA-Seq reads were also used to predict gene structures with Trinity81 (v2.0.13) and PASA82 (v2.2.0). All complete gene structures predicted by the PASA pipeline were used for gene model training for AUGUSTUS83, GlimmerHMM84, and SNAP85. These three predictors as well as geneid86 and GENSCAN87 were used for ab initio gene prediction with default parameters except that ‘-noInFrameStop=true -genemodel=complete’ was applied to AUGUSTUS. Finally, all gene models predicted with different approaches were integrated to generate a high-confidence gene set using EVidenceModeler88 with the following weight score matrix: PASA, 100; GeneWise, 20; Cufflinks, 20; AUGUSTUS, 5; other ab initio predictors, 1.
To evaluate the accuracy of predicted genes, we examined the coverage of highly conserved genes using BUSCO19. We further performed functional annotation of the T. grandis predicted gene models by searching against the databases Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/)89, SwissProt and TrEMBL (https://www.uniprot.org/) using BLASTP with an e-value cutoff of 1E-5, and the best alignment hits were used to assign homology-based gene functions. GO (http://geneontology.org/) categories and InterPro (https://www.ebi.ac.uk/interpro/) entries were obtained via InterProScan90.
Gene family evolution
The longest transcript of each of the protein-coding genes from 18 representative species (Taxus wallichiana, Amborella trichopoda, Arabidopsis thaliana, Ginkgo biloba, Gnetum montanum, Welwitschia mirabilis, Oryza sativa, Solanum lycopersicum, Physcomitrella patens, Pinus tabuliformis, Selaginella moellendorffii, Vitis vinifera, Sequoiadendron giganteum, Azolla filiculoides, Klebsormidium flaccidum, Chara braunii, Marchantia polymorpha and Penium margaritaceum) and T. grandis were selected to construct gene families based on all-against-all BLASTP alignments using OrthoFinder91. Phylogenetic analyses were conducted using IQ-TREE92 (v. 2.1.3). Based on MRCA analysis using CAFE93 (v.4.2.1), we determined the expansion and contraction of gene families between extant species and their last common ancestors.
Analysis of WGD events in the T. grandis genome
All-against-all BLASTP search was performed with an e-value cutoff of 1E-5. The top five alignments were selected for each gene and used to detect syntenic gene pairs located in collinear blocks with MCScanX94. Paralogous gene pairs were determined by the best reciprocal BLASTP alignments. Ks of each syntenic or paralogous gene pair was calculated using YN00 in the package PAML 4.8a95 with default parameters. Phylogeny based inference of WGD was carried out based on the reconciliation of each gene tree and the species tree.
Small RNA sequencing
Total RNA (3 μg) from leaves was isolated for small RNA library construction using NEB Next® Multiplex Small RNA Library Prep Set for Illumina® (NEB, USA) following manufacturer’s recommendations. DNA fragments in the constructed library within the range of 140~160 bp were recovered and the library was assessed on an Agilent Bioanalyzer 2100 system and subsequently sequenced on an Illumina HiSeq 2500 platform. Raw reads of small RNA library were processed with Trimmomatic66 (v0.36) to remove adapters and then aligned to the reference genome using Bowtie96 with no mismatch allowed.
Whole-genome bisulfite sequencing
T. grandis seeds were collected from a single tree on March 8 (stage 1), March 24 (stage 2) and April 8 (stage 3) of the year 2021 for bisulfite and transcriptome sequencing. About 100 ng high-quality genomic DNA spiked with 0.5 ng lambda DNA were sonicated with Covaris S220 (parameters: PIP, 50 W; duty factor, 20; cycles per burst, 200; treatment time, 110 s; temperature, 20 °C; sample volume, 52 μL). The fragmented DNA (200–300 bp) were treated with bisulfite using EZ DNA Methylation-GoldTM Kit (Zymo Research), and the library was quality assessed and sequenced on the Illumina NovaSeq platform with the paired-end mode.
Raw reads were cleaned with Trimmomatic66 (v0.36) to remove adaptors and low-quality sequences. To align the cleaned reads, both the reference genome and reads were transformed (C-to-T and G-to-A) and then aligned with Bismark97 (version 0.16.3) with parameters “-X 700 –dovetail”. Reads that produced a unique best alignment against both “Watson” and “Crick” strands of the genome were kept and the methylation state of all cytosine nucleotides were inferred. The sodium bisulfite conversion rate was estimated based on the read alignments to the lambda genome. Methylated sites were identified with a binomial test using the methylated counts (mC), total counts (mC+umC) and the conversion rate (r). Sites with FDR-corrected P value < 0.05 were considered as methylated sites. To calculate whole-genome methylation level, we divided the genome into 10-kb bins, and the methylation level of each window was calculated as count(mC)/(count(mC) + count(umC)). Differentially methylated regions (DMRs) were identified using the DSS software98 under the P value threshold of 0.05. DMRs were cataloged based on whether and how they overlapped with genes. Continuous cytosine sites across the T. grandis genome with methylation level <5% in any contexts were merged and defined as demethylation valleys.
Transcriptome sequencing and differential gene expression analysis
Male cones were collected from the T. grandis tree at eight different stages during February and April of 2021 with a time interval of 7 days, and female cones were collected at six different stages during January and April of 2021 with a time interval of 16 days. Other samples including kernel, leaf, root, and stem were collected from the same tree. Each sampling was performed with three biological replicates. Total RNA was extracted using TRIzol Reagent (Invitrogen). RNA-Seq libraries were constructed with the NEBNext Ultra II RNA Library Prep Kit for Illumina (NEB) following the manufacturer’s instructions and sequenced on an Illumina NovaSeq platform under 2 × 150-bp mode. Raw RNA-Seq reads were cleaned using Trimmomatic66 (v0.36). The cleaned reads were mapped to the genome using STAR aligner99 (v2.7.10a). The alignments were counted using HTSeq-count100 and differentially expressed genes were identified with DESeq2 (ref. 101) under the cutoff of adjusted P ≤ 0.01 and fold change ≥2.
HGT identification
Potential HGTs were identified based on homology scores and phylogeny signals102. Briefly, we created three customized databases, namely an out-group database comprising all protein sequences from archaea, bacteria and fungi, an in-group database including protein sequences from 10 published gymnosperm species, and a mid-group database consisting of sequences from all published plants excluding gymnosperms. Protein sequences of T. grandis were blasted against the three customized databases separately with an e-value cutoff of 1E-5. For each query protein sequence, we preserved no more than 100 blast hits (one hit per species) for each database and calculated the average bit-score value (ABV) of the alignments. Query proteins with the ABV of out-group larger than that of mid-group were retained. We performed rigorous phylogenetic analyses for each of remaining query proteins and manually inspected the topology of the tree. T. grandis genes supported by both ABV and phylogeny were considered as potential horizontally transferred genes.
Fatty acid determination
About 0.5 g dried samples were mixed with 9 mL 10% H2SO4-CH3OH solution at room temperature for 10 h. The fatty acid methyl esters were filtered and then extracted with 30 mL distilled water and 30 mL dichloromethane. The organic phase was dried with anhydrous sodium sulfate and concentrated to about 1 ml with a nitrogen blower. The concentrated extract was used for fatty acid analysis by gas chromatography (GC; Thermo Scientific TRACE-1300, Italy) with the methyl fatty acid used as an internal standard. GC separation was performed using an Agilent DB-WAX capillary GC column (30 m × 0.25 mm and 0.25 μm film thickness), and 1 μl of each sample was injected in split mode with a 1:20 ratio. Ultrapure helium was used as the carrier gas. The injection port and detector temperatures were set at 220 °C and 240 °C, respectively. The column temperature programming started at 140 °C held for 1 min and heated up to 250 °C at a rate of 4 °C/min. The column temperature was held for 2 min at 250 °C.
Subcellular localization
The CDS of each gene, without stop codon, was cloned and fused to the N-terminus of the GFP gene of the pCAMBIA1300-GFP vector. The resultant plasmid was introduced into Agrobacterium tumefaciens GV3101. Positive clones were incubated to an OD600 of 0.6, and then centrifuged at 8000 rpm for 6 min. Collected cells were resuspended with infiltration buffer (10 mM MgCl2, 0.2 mM acetosyringone, and 10 mM MES at pH 5.6), which was then injected into the leaves of Nicotiana benthamiana. After 3-d of culture, GFP fluorescence signal from the leaves was observed and captured using confocal laser scanning microscopy (LSM510: Karl Zeiss).
Quantitative real-time PCR analysis
Total RNA was extracted using the RNAprep pure Plant Kit (TIANGEN). First-strand cDNA was synthesized from 1 μg of total RNA using the PrimeScriptTM RT Master Mix Kit (Takara). SYBR Premix Ex Taq™ kit (Takara) was used to perform quantitative real-time PCR. Expression data of target genes were corrected with the expression of actin encoding gene. The reaction conditions were 95 °C for 10 s, 55 °C for 10 s, 72 °C for 20 s, 45 cycles. The relative expression was calculated using 2-ΔΔCt method.
Heterologous synthesis of SCA in Arabidopsis seeds and N. benthamiana leaves
Coding regions of TgEOL1, TgDES1, AtSLD2 (AT2G46210) and two recombinant genes (AtSLD2-Motif2 and AtSLD2-Motif3) were inserted into the downstream of the 35S promoter of the binary vector (pCAMBIA1300), respectively. Each of the resulting constructs was transformed into Agrobactrium tumefaciens strain GV3101, which was then grown at 28 °C in LB medium supplemented with kanamycin (50 mg/L) and rifampicillin (50 mg/L) until OD600 reaching 0.6. For transient expression of AtSLD2 and two recombinant genes in N. benthamiana leaves, cells were harvested and resuspended in 10 mM MES buffer (containing 10 mM MgCl2 and 0.1 mM acetosyringone) to a final OD600 of 1.0. The cells of each strain were infiltrated into the young leaves of five-week-old N. benthamiana plants using a needleless syringe, which were harvested 5 days later for measurement of SCA content. For generation of TgDES1 and TgELO1 overexpressed Arabidopsis, the pCAMBIA1300-TgELO1 and pCAMBIA1300-TgELO1 constructs were transformed into Arabidopsis via A. tumefaciens-mediated floral dip method. Hygromycin-resistant T1 plants were planted for seed harvesting, and T2 seeds with a hygromycin resistance ratio of 3:1 were selected to collect T3 seeds. T3 seeds with 100% resistance to hygromycin were used for determination of SCA content.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Genome assembly and raw reads for genome, transcriptome and methylome sequencing have been deposited in the National Center for Biotechnology Information BioProject database under accession PRJNA938254 and the CNGB Sequence Archive (CNSA) of China National GeneBank DataBase (CNGBdb) under accession CNP0003453. Genome assembly and annotation are also available at Figshare [https://doi.org/10.6084/m9.figshare.21089869]. Source data are provided with this paper.
References
Li, H. T. et al. Origin of angiosperms and the puzzle of the Jurassic gap. Nat. Plants 5, 461–470 (2019).
De La Torre, A. R. et al. Functional and morphological evolution in gymnosperms: A portrait of implicated gene families. Evol. Appl 13, 210–227 (2020).
Linkies, A., Graeber, K., Knight, C. & Leubner-Metzger, G. The evolution of seeds. N. Phytol. 186, 817–831 (2010).
Chen, X. & Jin, H. Review of cultivation and development of Chinese torreya in China. Trees Livelihoods 28, 68–78 (2019).
Wang, H., Guo, T. & Ying, G. Q. Advances in studies on active constituents and their pharmacological activities for plants of Torreya Arn. Chin. Tradit. Herb. Drugs 38, 1748–1750 (2007).
Wu, J. et al. De novo transcriptome sequencing of Torreya grandis reveals gene regulation in sciadonic acid biosynthesis pathway. Ind. Crops Prod. 120, 47–60 (2018).
He, Z. et al. Chemical components of cold pressed kernel oils from different Torreya grandis cultivars. Food Chem. 209, 196–202 (2016).
Berger, A. et al. Epidermal anti-inflammatory properties of 5,11,14 20:3: effects on mouse ear edema, PGE2 levels in cultured keratinocytes, and PPAR activation. Lipids Health Dis. 1, 5 (2002).
Pédrono, F. et al. Sciadonic acid derived from pine nuts as a food component to reduce plasma triglycerides by inhibiting the rat hepatic Δ9-desaturase. Sci. Rep. 10, 6223 (2020).
Endo, Y., Osada, Y., Kimura, F. & Fujimoto, K. Effects of Japanese torreya (Torreya nucifera) seed oil on lipid metabolism in rats. Nutrition 22, 553–558 (2006).
Song, L. et al. Advances on delta 5-unsaturated-polymethylene-interrupted fatty acids: resources, biosynthesis, and benefits. Crit. Rev. Food Sci. Nutr. 63, 767–789 (2023).
Aitzetmüller, K. An unusual fatty acid pattern in Eranthis seed oil. Lipids 31, 201–205 (1996).
Niu, S. et al. The Chinese pine genome and methylome unveil key features of conifer evolution. Cell 185, 204–217.e14 (2022).
Nystedt, B. et al. The Norway spruce genome sequence and conifer genome evolution. Nature 497, 579–584 (2013).
Liu, H. et al. The nearly complete genome of Ginkgo biloba illuminates gymnosperm evolution. Nat. Plants 7, 748–756 (2021).
Wan, T. et al. The Welwitschia genome reveals a unique biology underpinning extreme longevity in deserts. Nat. Commun. 12, 4247 (2021).
Liu, Y. et al. The Cycas genome and the early evolution of seed plants. Nat. Plants 8, 389–401 (2022).
Cheng, J. et al. Chromosome-level genome of Himalayan yew provides insights into the origin and evolution of the paclitaxel biosynthetic pathway. Mol. Plant 14, 1199–1209 (2021).
Xiong, X. et al. The Taxus genome provides insights into paclitaxel biosynthesis. Nat. Plants 7, 1026–1036 (2021).
Rhie, A., Walenz, B. P., Koren, S. & Phillippy, A. M. Merqury: reference-free quality, completeness, and phasing assessment for genome assemblies. Genome Biol. 21, 245 (2020).
Manni, M., Berkeley, M. R., Seppey, M., Simão, F. A. & Zdobnov, E. M. BUSCO update: novel and streamlined workflows along with broader and deeper phylogenetic coverage for scoring of eukaryotic, prokaryotic, and viral genomes. Mol. Biol. Evol. 38, 4647–4654 (2021).
Ou, S., Chen, J. & Jiang, N. Assessing genome assembly quality using the LTR assembly index (LAI). Nucleic Acids Res 46, e126 (2018).
Steinthorsdottir, M. et al. The Miocene: the future of the past. Paleoceanogr. Paleoclimatol. 36, e2020PA004037 (2021).
El Baidouri, M. & Panaud, O. Comparative genomic paleontology across plant kingdom reveals the dynamics of TE-driven genome evolution. Genome Biol. Evol. 5, 954–965 (2013).
Matzke, M. A. & Mosher, R. A. RNA-directed DNA methylation: an epigenetic pathway of increasing complexity. Nat. Rev. Genet. 15, 394–408 (2014).
Nakamura, M., Köhler, C. & Hennig, L. Tissue-specific transposon-associated small RNAs in the gymnosperm tree, Norway spruce. BMC Genom. 20, 997 (2019).
Van de Peer, Y., Mizrachi, E. & Marchal, K. The evolutionary significance of polyploidy. Nat. Rev. Genet. 18, 411–424 (2017).
Leebens-Mack, J. H. et al. One thousand plant transcriptomes and the phylogenomics of green plants. Nature 574, 679–685 (2019).
Sun, X. et al. Genome and evolution of the arbuscular mycorrhizal fungus Diversispora epigaea (formerly Glomus versiforme) and its bacterial endosymbionts. N. Phytol. 221, 1556–1573 (2019).
Jiao, C. et al. The Penium margaritaceum genome: hallmarks of the origins of land plants. Cell 181, 1097–1111.e12 (2020).
Bowman, J. L. et al. Insights into land plant evolution garnered from the Marchantia polymorpha genome. Cell 171, 287–304.e15 (2017).
Bouché, F., Lobet, G., Tocquin, P. & Périlleux, C. FLOR-ID: an interactive database of flowering-time gene networks in Arabidopsis thaliana. Nucleic Acids Res. 44, D1167–D1171 (2016).
Li, L. X. et al. Arabidopsis thaliana NOP10 is required for gametophyte formation. J. Integr. Plant Biol. 60, 723–736 (2018).
Schoof, H. et al. The stem cell population of Arabidopsis shoot meristems in maintained by a regulatory loop between the CLAVATA and WUSCHEL genes. Cell 100, 635–644 (2000).
Soltis, D. E., Chanderbali, A. S., Kim, S., Buzgo, M. & Soltis, P. S. The ABC model and its applicability to basal angiosperms. Ann. Bot. 100, 155–163 (2007).
Dreni, L. & Zhang, D. Flower development: the evolutionary history and functions of the AGL6 subfamily MADS-box genes. J. Exp. Bot. 67, 1625–1638 (2016).
Kong, X. et al. The wheat AGL6‐like MADS‐box gene is a master regulator for floral organ identity and a target for spikelet meristem development manipulation. Plant Biotechnol. J. 20, 75–88 (2022).
Zhang, L. et al. The water lily genome and the early evolution of flowering plants. Nature 577, 79–84 (2020).
Chanderbali, A. S. et al. Conservation and canalization of gene expression during angiosperm diversification accompany the origin and evolution of the flower. Proc. Natl Acad. Sci. USA 107, 22570–22575 (2010).
Souza, P. F. N. The forgotten 2S albumin proteins: importance, structure, and biotechnological application in agriculture and human health. Int. J. Biol. Macromol. 164, 4638–4649 (2020).
Tandang-Silvas, M. R. et al. Conservation and divergence on plant seed 11S globulins based on crystal structures. Biochim. Biophys. Acta 1804, 1432–1442 (2010).
Cheng, S. et al. Genomes of subaerial Zygnematophyceae provide insights into land plant evolution. Cell 179, 1057–1067.e14 (2019).
Weng, J. K. & Chapple, C. The origin and evolution of lignin biosynthesis. N. Phytol. 187, 273–285 (2010).
Yamaguchi, M. et al. VASCULAR-RELATED NAC-DOMAIN6 and VASCULAR-RELATED NAC-DOMAIN7 effectively induce transdifferentiation into xylem vessel elements under control of an induction system. Plant Physiol. 153, 906–914 (2010).
Zhong, R., Richardson, E. A. & Ye, Z. H. Two NAC domain transcription factors, SND1 and NST1, function redundantly in regulation of secondary wall synthesis in fibers of Arabidopsis. Planta 225, 1603–1611 (2007).
Jokipii-Lukkari, S. et al. NorWood: a gene expression resource for evo-devo studies of conifer wood development. N. Phytol. 216, 482–494 (2017).
Meesapyodsuk, D. & Qiu, X. The front-end desaturase: structure, function, evolution and biotechnological use. Lipids 47, 227–237 (2012).
Sayanova, O., Haslam, R., Venegas Caleron, M. & Napier, J. A. Cloning and characterization of unusual fatty acid desaturases from Anemone leveillei: identification of an acyl-coenzyme A C20 Δ5-desaturase responsible for the synthesis of sciadonic acid. Plant Physiol. 144, 455–467 (2007).
Xue, J. A. et al. Expression of yeast acyl-CoA-∆9 desaturase leads to accumulation of unusual monounsaturated fatty acids in soybean seeds. Biotechnol. Lett. 35, 951–959 (2013).
Lim, Z. L., Senger, T. & Vrinten, P. Four amino acid residues influence the substrate chain-length and regioselectivity of Siganus canaliculatus Δ4 and Δ5/6 desaturases. Lipids 49, 357–367 (2014).
Buček, A., Vazdar, M., Tupec, M., Svatoš, A. & Pichová, I. Desaturase specificity is controlled by the physicochemical properties of a single amino acid residue in the substrate binding tunnel. Comput. Struct. Biotechnol. J. 18, 1202–1209 (2020).
Chen, M. et al. Seed genome hypomethylated regions are enriched in transcription factor genes. Proc. Natl Acad. Sci. USA 115, E8315–E8322 (2018).
Niederhuth, C. E. et al. Widespread natural variation of DNA methylation within angiosperms. Genome Biol. 17, 194 (2016).
Ausin, I. et al. DNA methylome of the 20-gigabase Norway spruce genome. Proc. Natl Acad. Sci. USA 113, E8106–E8113 (2016).
Bewick, A. J. & Schmitz, R. J. Gene body DNA methylation in plants. Curr. Opin. Plant Biol. 36, 103–110 (2017).
Takuno, S., Ran, J. H. & Gaut, B. S. Evolutionary patterns of genic DNA methylation vary across land plants. Nat. Plants 2, 15222 (2016).
Lin, J. Y. et al. Similarity between soybean and Arabidopsis seed methylomes and loss of non-CG methylation does not affect seed development. Proc. Natl Acad. Sci. USA 114, E9730–E9739 (2017).
Chen, F., Tobimatsu, Y., Havkin-Frenkel, D., Dixon, R. A. & Ralph, J. A polymer of caffeyl alcohol in plant seeds. Proc. Natl Acad. Sci. USA 109, 1772–1777 (2012).
Hiraide, H. et al. Localised laccase activity modulates distribution of lignin polymers in gymnosperm compression wood. N. Phytol. 230, 2186–2199 (2021).
Cosgrove, D. J. Loosening of plant cell walls by expansins. Nature 407, 321–326 (2000).
Vigeolas, H., van Dongen, J. T., Waldeck, P., Huhn, D. & Geigenberger, P. Lipid storage metabolism is limited by the prevailing low oxygen concentrations within developing seeds of oilseed rape. Plant Physiol. 133, 2048–2060 (2003).
Murray, M. G. & Thompson, W. F. Rapid isolation of high molecular weight plant DNA. Nucleic Acids Res. 8, 4321–4325 (1980).
Zhu, W. et al. Altered chromatin compaction and histone methylation drive non-additive gene expression in an interspecific Arabidopsis hybrid. Genome Biol. 18, 157 (2017).
Cheng, H., Concepcion, G. T., Feng, X., Zhang, H. & Li, H. Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm. Nat. Methods 18, 170–175 (2021).
Roach, M. J., Schmidt, S. & Borneman, A. R. Purge Haplotigs: synteny reduction for third-gen diploid genome assemblies. BMC Bioinforma. 19, 460 (2018).
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Zhang, X., Zhang, S., Zhao, Q., Ming, R. & Tang, H. Assembly of allele-aware, chromosomal-scale autopolyploid genomes based on Hi-C data. Nat. Plants 5, 833–845 (2019).
Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
Xu, Z. & Wang, H. LTR_FINDER: an efficient tool for the prediction of full-length LTR retrotransposons. Nucleic Acids Res. 35, W265–W268 (2007).
Flynn, J. M. et al. RepeatModeler2 for automated genomic discovery of transposable element families. Proc. Natl Acad. Sci. USA 117, 9451–9457 (2020).
Bao, W., Kojima, K. K. & Kohany, O. Repbase Update, a database of repetitive elements in eukaryotic genomes. Mob. DNA 6, 11 (2015).
Tarailo-Graovac, M. & Chen, N. Using RepeatMasker to identify repetitive elements in genomic sequences. Curr. Protoc. Bioinforma. 25, Chapter 4:Unit 4.10 (2009).
Benson, G. Tandem repeats finder: a program to analyze DNA sequences. Nucleic Acids Res. 27, 573–580 (1999).
Ellinghaus, D., Kurtz, S. & Willhoeft, U. LTRharvest, an efficient and flexible software for de novo detection of LTR retrotransposons. BMC Bioinforma. 9, 18 (2008).
Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
She, R., Chu, J. S., Wang, K., Pei, J. & Chen, N. GenBlastA: enabling BLAST to identify homologous gene sequences. Genome Res. 19, 143–149 (2009).
Birney, E., Clamp, M. & Durbin, R. GeneWise and genomewise. Genome Res. 14, 988–995 (2004).
Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111 (2009).
Roberts, A., Pimentel, H., Trapnell, C. & Pachter, L. Identification of novel transcripts in annotated genomes using RNA-Seq. Bioinformatics 27, 2325–2329 (2011).
Grabherr, M. G. et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 29, 644–652 (2011).
Haas, B. J. et al. Improving the Arabidopsis genome annotation using maximal transcript alignment assemblies. Nucleic Acids Res. 31, 5654–5666 (2003).
Stanke, M. et al. AUGUSTUS: ab initio prediction of alternative transcripts. Nucleic Acids Res. 34, W435–W439 (2006).
Pertea, M., Salzberg, S. L. & Majoros, W. H. TigrScan and GlimmerHMM: two open source ab initio eukaryotic gene-finders. Bioinformatics 20, 2878–2879 (2004).
Korf, I. Gene finding in novel genomes. BMC Bioinforma. 5, 59 (2004).
Alioto, T., Blanco, E., Parra, G. & Guigó, R. Using geneid to identify genes. Curr. Protoc. Bioinforma. 64, e56 (2018).
Burge, C. & Karlin, S. Prediction of complete gene structures in human genomic DNA. J. Mol. Biol. 268, 78–94 (1997).
Haas, B. J. et al. Automated eukaryotic gene structure annotation using EVidenceModeler and the Program to Assemble Spliced Alignments. Genome Biol. 9, R7 (2008).
Morishima, K., Tanabe, M., Furumichi, M., Kanehisa, M. & Sato, Y. KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 45, 353–361 (2016).
Mitchell, A. et al. InterProScan 5: genome-scale protein function classification. Bioinformatics 30, 1236–1240 (2014).
Emms, D. M. & Kelly, S. OrthoFinder: phylogenetic orthology inference for comparative genomics. Genome Biol. 20, 238 (2019).
Nguyen, L. T., Schmidt, H. A., von Haeseler, A. & Minh, B. Q. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32, 268–274 (2015).
De Bie, T., Cristianini, N., Demuth, J. P. & Hahn, M. W. CAFE: a computational tool for the study of gene family evolution. Bioinformatics 22, 1269–1271 (2006).
Wang, Y. et al. MCScanX: a toolkit for detection and evolutionary analysis of gene synteny and collinearity. Nucleic Acids Res. 40, 49 (2012).
Yang, Z. PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol. Evol. 24, 1586–1591 (2007).
Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).
Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571–1572 (2011).
Park, Y. & Wu, H. Differential methylation analysis for BS-seq data under general experimental design. Bioinformatics 32, 1446–1453 (2016).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Anders, S., Pyl, P. T. & Huber, W. HTSeq-a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Li, Y. et al. HGT is widespread in insects and contributes to male courtship in lepidopterans. Cell 185, 2975–2987.e10 (2022).
Acknowledgements
This research was supported by grants from the National Natural Science Foundation of China (NSFC) to J.W. (grant no. U20A2049), L.S. (grant no. 31971699), X.S. (grant no. 32102318), the Key Research and Development Program of Zhejiang Province to H.L. (grant no. 2021C02001), the Scientific Research Startup Fund Project of Zhejiang A&F University to H.L. (grant no. 2018FR028), and the State Key Laboratory of Subtropical Silviculture grant to J.W. (grant no. ZY20180312 and ZY20180209). The authors thank Dr. Emily E.D. Coffey from Atlanta Botanical Garden (USA) and Prof. Mark W. Schwartz from University of California, Davis for providing plant samples.
Author information
Authors and Affiliations
Contributions
J.W., X.S., H.L. and L.S. conceived and supervised the project. W.C., Y.G. and S.Z. collected samples and performed transgenic experiments. X.S., X.L. and H.Z. constructed libraries and performed bioinformatics analysis. X.S. and H.L wrote the manuscript. Z.F. and J.W. revised the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Amanda De La Torre, Liang Guo, Nathaniel Street and Haifeng Wang for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Lou, H., Song, L., Li, X. et al. The Torreya grandis genome illuminates the origin and evolution of gymnosperm-specific sciadonic acid biosynthesis. Nat Commun 14, 1315 (2023). https://doi.org/10.1038/s41467-023-37038-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-023-37038-2
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
