Deciphering the Biosynthetic Pathways and Regulatory Networks of the Active Components of Cibotium barometz by Transcriptomic Analysis

Abstract

Cibotium barometz (L.) J. Sm., a medicinally significant fern in traditional Chinese medicine, is little explored at the genomic level regarding its bioactive compounds. Using an integrated approach combining Illumina and PacBio sequencing technologies, we profiled its root, rachis, and pinna transcriptomes, identifying 12,718, 21,341, and 11,441 unigenes, respectively. Our analysis systematically characterized the transcriptional features of transcription factors (TFs), simple sequence repeats (SSRs), long non-coding RNAs (lncRNAs), and differentially expressed genes (DEGs). Enrichment analyses highlighted the roles of highly expressed unigenes in secondary metabolism. Seventeen key enzymes involved in polysaccharide biosynthesis showed tissue-specific expression patterns. Notably, total polysaccharide content correlated positively with UDP-arabinose 4-epimerase (UXE) expression but negatively with phosphoglucomutase (PGM) and 3,5-epimerase/4-reductase (UER1). Flavonoid accumulation inversely correlated with chalcone synthase (CHS) expression. Two lignin pathways (H-lignin and G-lignin) were characterized, with phenylalanine ammonia-lyase (PAL), cinnamate-4-hydroxylase (C4H), and cinnamyl alcohol dehydrogenase (CAD) as key genes. The absence of ferulate-5-hydroxylase (F5H) explains the undetected S-lignin pathway. Regulatory network analysis revealed positive correlations between PAL expression and NAC72/NAC78/WRKY35 and C4H expression and WRKY65/WRKY69/WRKY71, while a negative correlation was revealed between flavonoid 3′,5′-hydroxylase (F3′5′H) and MYB3R4. This study provides comprehensive transcriptomic insights into C. barometz bioactive compound biosynthesis, serving as a foundation for mechanistic research.

1. Introduction

Cibotium barometz (L.) J. Sm., also known as the golden chicken fern, is a member of the Dicksoniaceae family. This species grows erectly as a large tree fern, reaching up to 1 m in height, with fronds that can extend to 3 m in length and sori positioned on the margin of the pinnules. Native to certain regions in China and the Western Malay Peninsula, C. barometz is very common in the southern subtropical and tropical regions [1]. The fern’s rhizome resembles a golden-haired dog due to its thick, woody texture and is adorned with long, soft, golden-yellow hairs [1]. This characteristic has earned it the nickname “Jinmao Gouji” (golden-haired dog) in the Chinese Pharmacopoeia [2]. Cibotium barometz is a well-known traditional Chinese medicine, recognized for its restorative properties, particularly in strengthening the kidneys and addressing orthopedic ailments [3,4,5,6]. This fern is also an ornamental plant with high commercial value [7]. It stands out as one of China’s most economically important ferns for export [7,8].

The main bioactive components of C. barometz include flavonoids, terpenoids, and glycosides [7,9]. It is also a rich source of polysaccharides [5]. The polysaccharides play a dual role: they enhance anti-osteoporotic activity by promoting chondrocyte proliferation, osteoblast differentiation, and the mineralization process, and they also improve the immune system. Moreover, they demonstrate a particular effect on tumor and viral activity, with low toxicity [5,10,11,12]. Flavonoids are prevalent secondary metabolites in plants, serving various functions [13]. They are renowned for their antioxidant, anti-inflammatory, antimutagenic, and anticarcinogenic properties [14]. These compounds are particularly crucial in plant growth and development, as well as in the plants’ responses to environmental factors, including cell regulation, insect pollinator attraction, and the defense against biotic and abiotic stresses [14,15,16,17]. Currently, flavonoids are detected in various plant tissues [18,19]. Structural genes associated with flavonoid biosynthesis are under direct regulation by transcription factors (TFs) like MYB and bHLH [20]. Moreover, the plant-specific transcription factor WRKY can aid plants to cope with abiotic and biotic stresses by regulating MYC2-mediated flavonoid biosynthesis [21]. WD40-repeat proteins, one of the largest protein families, are particularly crucial in the precise regulation of flavonoid biosynthesis [22,23].

C. barometz has a tall and upright trunk, supported by the xylem. The xylem is rich in lignin [24], a key component of the plant cell wall and the most abundant natural phenolic polymer [25,26]. Lignin biosynthesis is a crucial process in plant growth and development, lodging resistance, and responses to both biotic and abiotic stresses [25]. For instance, studies have shown that gene expressions related to lignin biosynthesis, such as cinnamyl alcohol dehydrogenase and caffeate O-methyltransferase, are significantly and positively correlated with drought stress [27]. The biosynthesis of lignin involves the polymerization of phenylpropylene units (monolignols) [28,29]. The mechanisms underlying lignin biosynthesis in plants have been extensively investigated [30,31,32,33,34,35]. Several genes that govern the phenylpropanoid metabolic pathway have a substantial impact on both the lignin content and composition [25]. The use of RNAi technology to suppress the expression of cinnamoyl-CoA reductase (CCR1) and caffeic acid O-methyltransferase 1 (OMT1) has also been shown to significantly alter lignin content in Lolium perenne L. [36]. The AP2/ERF family, one of the largest transcription factor families in plants, plays a regulatory role in lignin accumulation [37]. However, there are relatively few reports focusing on the biosynthesis and regulation of lignin in tree ferns.

Transcriptome sequencing has emerged as a highly effective method for probing metabolic pathways and functional gene identifications associated with bioactive compounds [38]. This technique is particularly valuable in the absence of a reference genome, enabling the exploration of gene expression and the organ-specific regulatory networks that govern secondary metabolite synthesis [39]. It has facilitated the identification of genes linked to the polysaccharide synthesis across various plant species, including Panax ginseng C.A. Mey [40]. Similarly, the molecular mechanisms of flavonoid synthesis have been explored through transcriptome sequencing in a range of plants [19,41,42,43]. Furthermore, the technique has been instrumental in identifying genes involved in plant lignin biosynthesis [44,45,46,47,48].

In this study, we employed a combination approach using both Illumina and PacBio sequencing technologies to investigate the synthesis pathways and regulatory networks governing the active components of C. barometz. Our goals were: (1) to conduct transcriptome sequencing on the three distinct organs (the root, rachis, and pinna), annotate the genes identified, and assess their expression levels; (2) to identify differentially expressed genes across these organs and determine their functions and metabolic pathways; and (3) to uncover genes and metabolic pathways related to the biosynthesis of polysaccharide, flavonoid, and lignin. The results will reveal the interplay between polysaccharide and flavonoid biosynthesis and gene expression, offering new insights into the key bioactive compounds. Moreover, this study will also provide useful molecular resources for genomic, physiological, and pharmaceutical research into C. barometz.

2. Results

2.1. Content of Total Polysaccharides and Flavonoids in C. barometz Organs

To assess the concentration of total polysaccharides and flavonoids across the different organs of C. barometz, we quantified the two compounds in the root, rachis, and pinna. The polysaccharide content peaked at 14.98% in the root, with the rachis at 14.35% and the pinna at 11.85% (Figure 1A). Conversely, the pinna had the highest flavonoid content at 1.67%, surpassing the root’s 0.67% and the rachis’s 0.55% (Figure 1B). Based on these findings, we used the three organs to investigate the biosynthetic pathways and regulatory networks of the active components through a comparative transcriptomic analysis.

Figure 1.

The content of the active components in C. barometz (ns indicates no significant difference; *** indicates extremely significant difference of p < 0.001). (A) total polysaccharides; (B) total flavonoids.

2.2. De Novo Assembly of the Transcriptomes Using Illumina Data

Illumina sequencing generated 45,746,176 raw reads for the root, 47,488,634 for the rachis, and 53,363,062 for the pinna. After filtration, we derived 6.68 G, 6.64 G, and 8 G clean reads for each organ, respectively (Table S1). The average values of Q20 and Q30 were 97% and 93%, respectively, and the GC content was between 48% and 51%. Using the high-quality clean reads, we identified a total of 154,767 transcripts in the root, 72,911 in the rachis, and 87,567 in the pinna, with N50 lengths of 1312 bp, 1957 bp, and 2241 bp, respectively (Table S2). These transcripts were assembled into 98,888 unigenes in the root, 34,343 in the rachis, and 38,281 in the pinna. All unigenes were longer than 200 bp, and their length distribution is depicted in Figure S1A.

2.3. PacBio Data Processing

PacBio sequencing of C. barometz generated 210,752, 579,363, and 211,353 reads for the root, rachis, and pinna, with average read lengths of 58,802 bp, 29,684 bp, and 88,229 bp and N50 values of 119,775 bp, 49,528 bp, and 157,987 bp, respectively (Table S3). After quality filtering, we derived 12.0 G clean data for the root, 16.69 G for the rachis, and 18.14 G for the pinna. The data yielded 5,384,571 subreads for the root, 7,353,811 for the rachis, and 6,935,956 for the pinna, with average lengths of 2228 bp, 2270 bp, and 2616 bp and N50 values of 2471 bp, 3304 bp, and 3013 bp, respectively. In the process of self-correction, each organ produced 178,160 (root), 498,066 (rachis), and 191,003 (pinna) circular consensus sequences (CCS). Among these, the full-length non-chimeric reads (FLNC) were 162,144 for the root, 377,773 for the rachis, and 160,449 for the pinna. Ultimately, we obtained 19,129 (root), 38,661 (rachis), and 18,969 (pinna) polished consensus sequences.

The polished consensus sequences were refined using Illumina data to enhance accuracy. In the root, a total of 12,718 unigenes were identified, with 4756 falling within the 2–3 kb range and 4367 within the 1–2 kb range. In the rachis, we derived 21,341 unigenes, with 50.9% (10,856) exceeding 3 kb in length and 28.8% (6141) ranging from 2 to 3 kb. In the pinna, 11,441 unigenes were detected, consisting of 38.6% (4412) over 3 kb and 35.73% within the 2–3 kb range (Table 1).

Table 1.

Statistical results of SMRT transcript and unigene length in C. barometz.

Sample	Type	<500 bp	500–1k bp	1k–2k bp	2k–3k bp	>3k bp	Total
Root	Transcripts	51	465	6572	7289	4752	19,129
	Unigenes	19	311	4367	4756	3265	12,718
Rachis	Transcripts	1041	2549	3886	11,795	19,390	38,661
	Unigenes	271	1427	2646	6141	10,856	21,341
Pinna	Transcripts	13	306	4048	7042	7560	18,969
	Unigenes	6	228	2707	4088	4412	11,441

We identified a total of 1968 transcripts that were common across the root, rachis, and pinna. The number of transcripts unique to each organ was as follows: 7243 in the root, 15,162 in the rachis, and 5939 in the pinna. The highest number of transcripts shared between two organs was detected between the rachis and pinna (2076), followed by the root and rachis (2054). The least common transcripts were shared between the pinna and root (1386) (Figure S1B).

2.4. Functional Annotation of Genes

The unigenes of the three organs were annotated using seven databases: NR, Nt, Pfam, KOG, S—WISS-PROT, KEGG, and GO. In total, 11,119 genes (97.19%) were annotated in at least one database in the pinna transcriptome, whereas 12,231 (96.17%) and 20,310 (95.17%) were annotated in the root and rachis, respectively (Table S4). Out of the seven databases, the NR database demonstrated the highest annotation rates, achieving 96.47%, 93.95%, and 93.31% for the pinna, root, and rachis, respectively (Figure S2A).

Genes from the three organs were categorized into 26 distinct classes according to the annotation data obtained from the KOG database. In total, 565, 631, and 492 genes were annotated to “carbohydrate transport metabolism (G)” in the root, rachis, and pinna, respectively (Figure S2C). In this class, genes in the root, pinna, and rachis were mainly annotated as galactosyltransferases (KOG2288), glucose-6-phosphate isomerase (KOG2446), hexokinase (KOG1369), mannose-6-phosphate isomerase (KOG2757), sugar transporter/spinster transmembrane protein (KOG1330), and sucrose transporter and related proteins (KOG0637). Regarding the annotations derived from the GO database, genes were assigned into 51 (root), 54 (rachis), and 52 (pinna) subcategories. Within the biological process category, we individually annotated 213 (root), 295 (rachis), and 224 (pinna) unigenes to the carbohydrate metabolism process (Figure 2). Additionally, 8 (root), 24 (rachis), and 12 (pinna) unigenes were annotated to the 1,3-beta-D-glucan biosynthetic process (GO: 0006075).

Figure 2.

GO annotation classification of unigenes.

2.5. Detection of TF, SSR, and lncRNA

We predicted a total of 1130 TFs in the root, 1802 in the rachis, and 913 in the pinna (Figure 3A). In the root, the C3H family was the most abundant, followed by the bHLH and AP2/ERF families. In the rachis, the C3H, C2H2, and bHLH families were the three largest. In the pinna, the bHLH family was the most abundant, followed by the C3H and C2H2 families (Figure 3A). Furthermore, additional TFs, including GRAS, NAC, MYB, and WRKY, were annotated in the three organs.

Figure 3.

Distribution of transcription factors, lncRNAs, and SSRs in three C. barometz organs. (A) Distribution of the top 29 transcription factor families. (B) The number of lncRNAs in the three organs. (C) Distribution of SSR motifs. The X axis represents the SSR motif units (by number of repeating bases). The Y axis indicates the number of base repetitions, with different colors corresponding to specific repetition counts as shown in the inlet legend. The Z axis denotes the number of SSRs.

We identified 6796, 14,545, and 7395 simple sequence repeats (SSRs) in the root, rachis, and pinna transcriptomes, respectively. Dinucleotide repeats emerged as the most abundant SSRs, with 8142 in the rachis, 4112 in the pinna, and 1635 in the root. These were followed by single-nucleotide repeats and trinucleotide repeats, while pentanucleotide repeats were the least common (Table S5, Figure 3B).

We discovered 3626 lncRNAs in the root, 7115 in the rachis, and 2883 in the pinna. The rachis had the largest number (231) of shared lncRNAs, whereas the pinna had the lowest number (22) (Figure 3C).

2.6. Gene Expression Level

To investigate the expression profiles of unigenes, the clean Illumina reads were mapped to the PacBio non-redundant full-length transcripts. This allowed us to measure the expression level using the expected number of fragments per kilobase of transcript sequence per million base pairs sequenced (FPKM). The mapping rates for each organ were as follows: 67.01% for the root, 74.16% for the rachis, and 73.85% for the pinna (Table S6).

Across the three organs, unigenes with FPKM values ranging from 1 to 5 accounted for 30.86% of the root, 31.79% of the rachis, and 27.06% of the pinna (Table S7; Figure 4A). Unigenes with FPKM values between 5 and 15 represented 22.21%, 20.39%, and 24.37% of the transcripts in the root, rachis, and pinna, respectively. Unigenes with FPKM values between 15 and 60 made up 14.92% of the root, 13.23% of the rachis, and 17.43% of the pinna transcripts. Furthermore, the genes that were highly expressed with FPKM values greater than 60 were distributed as follows: 2363 in the root (6.60%), 2447 in the rachis (6.83%), and 2351 in the pinna (6.56%).

Figure 4.

Statistical plot of the gene number with different FPKM distribution (A) and venn diagram of differentially expressed genes (B).

2.7. Analysis of DEGs

We analyzed differentially expressed genes (DEGs) across the three organs, setting FPKM > 0.3 as the threshold. A total of 1996, 969, and 532 organ-specific DEGs were detected, with the highest number observed between the rachis and pinna, followed by the pinna and root, and the lowest between the root and rachis (Figure 4B). The DEGs between the pinna and root, pinna and rachis, and root and rachis amounted to 9077 (the highest), 8553, and 5207, respectively (Table S8). Of the DEGs comparing the root and pinna, 4318 (47.57%) were upregulated in the root, while 4759 (52.43%) were upregulated in the pinna. In contrast, for the DEGs comparing the root and rachis, 3081 (59.17%) were upregulated in the root, and 2126 (40.83%) were upregulated in the rachis.

We conducted further GO and KEGG enrichment analyses on the DEGs to gain a comprehensive understanding of their functions (Figure 5A–C). The KEGG enrichment analysis indicated that the upregulated genes in the transcriptome of C. barometz were mainly associated with the flavonoid biosynthesis, carbohydrate metabolism, and phenylpropanoid biosynthesis pathways (Figure 5A–C). Notably, the upregulated genes in comparisons between the pinna and root, as well as the pinna and rachis, were enriched in photosynthesis in the KEGG pathways (Figure 5C). The GO analysis of DEGs revealed that the pinna and root were enriched in 60 pathways, with a single biological process (GO:0044699) and an ion binding process (GO:0043167) being the most represented, containing 2667 and 2373 unigenes, respectively. The enrichment analysis of DEGs between the root and rachis generated 60 differentially expressed GO pathways, while the comparison between the pinna and rachis resulted in 55 differentially expressed GO pathways (Table S9, Figure S3A–C).

Figure 5.

KEGG annotation of up-regulated genes in comparison of the three organs (top 20). (A) root compared to rachis and pinna; (B) rachis compared to root and pinna; (C) pinna compared to root and rachis.

2.8. Polysaccharide Synthesis Pathway in C. barometz

A total of 971, 1138, and 896 unigenes were annotated to the “carbohydrate metabolism” subcategory in the KEGG metabolic pathway for the root, rachis, and pinna, respectively. This subcategory includes 14 metabolic pathways. Notably, the “starch and sucrose metabolism” pathway (ko00500) had the highest number of unigenes, with counts of 145, 238, and 137 in the root, rachis, and pinna, respectively. In addition, the “amino sugar and nucleoside sugar metabolism” pathway (ko00520) was represented by 104, 113, and 98 unigenes; the “fructose and mannose metabolism” pathway (ko00051) by 60, 57, and 51 unigenes; and the “galactose metabolism” pathway (ko00052) by 37, 33, and 28 unigenes in the root, rachis, and pinna, respectively.

According to the KEGG database, we have identified 98 unigenes that encode 17 key enzymes crucial for polysaccharide synthesis. Among these, the largest number of unigenes (14) were found to encode hexokinase (HK). Uridine diphosphate glucose pyrophosphorylase (galU) and β-fructofuranosidase were encoded by 11 and 10 unigenes, respectively (Table 2). DEGs were detected between the root and pinna, with nine β-fructofuranosidase (sacA), eight UDP-apiose/xylose synthase (AXS), and four mannose-1-phosphateguanylyltransferase (GMPP) encoding genes being particularly notable. Similarly, DEGs were identified between the pinna and rachis, highlighting six UDP-glucose 4,6-dehydratase (RHM), six UDP-apiose/xylose synthase (AXS), and five UDP-glucose 6-dehydrogenase (UGDH)-encoding genes.

Table 2.

The number of DEGs involved in the biosynthesis of starch and sucrose.

Enzyme Code	Enzyme Name	Unigenes	DEGs (Root vs. Pinna)	DEGs (Pinna vs. Rachis)	DEGs (Root vs. Rachis)
3.2.1.26	β-fructofuranosidase (sacA)	10	9	5	1
2.7.1.4	Fructokinase (scrK)	5	1	1	0
2.7.1.1	Hexokinase (HK)	14	3	1	1
5.3.1.9	glucose-6-phosphate isomerase (GPI)	4	2	2	1
5.3.1.8	Mannose-6-phosphate isomerase (MPI)	4	1	2	1
5.4.2.8	Phosphomannomutase (PMM)	1	0	0	0
2.7.7.13	mannose-1-phosphateguanylyltransferase (GMPP)	6	4	3	1
4.2.1.47	GDP-mannose 4,6-dehydratase (GMDS)	4	1	1	1
1.1.1.271	GDP-L-fucose synthase (TSTA3)	1	0	0	0
5.4.2.2	Phosphoglucomutase (PGM)	6	3	3	1
2.7.7.9	UTP-glucose-1-phosphate uridylyltransferase (galU)	11	4	3	3
5.1.3.2	UDP-glucose 4-epimerase (GALE)	1	0	1	1
4.2.1.76	UDP-glucose 4,6-dehydratase (RHM)	8	4	6	1
5.1.3.- 1.1.1.-	3,5-epimerase/4-reductase (UER1)	1	1	0	0
1.1.1.22	UDP-glucose 6-dehydrogenase (UGDH)	9	1	5	3
-	UDP-apiose/xylose synthase (AXS)	8	8	6	3
5.1.3.5	UDP-arabinose 4-epimerase (UXE)	5	0	0	0

In plants, the enzyme UDP-glucose pyrophosphorylase (UGPase) catalyzes the conversion of glucose-1-phosphate (Glc-1P) into uridine diphosphate glucose (UDP-Glu), a key precursor for nucleotide-diphospho-sugars (NDP-sugars). These NDP-sugar precursors are essential for the generation of plant polysaccharides, as they are activated by various glycosyltransferases (GTs), which facilitate the addition of NDP-sugars to both polysaccharides and glycoconjugate residues. The composition of cell walls is partially determined by the availability of NDP-sugars, which have effects on the formation of different types of wall polymers. Based on these metabolic processes, the biosynthetic pathway of polysaccharides in C. barometz was depicted (Figure 6A).

Figure 6.

Characterization of genes related to polysaccharide biosynthesis in C. barometz. (A) Proposed pathway for polysaccharide biosynthesis. Arrows with solid lines denote the identified enzymatic reactions, while arrows with dashed lines represent enzymatic reactions by multiple steps. Activated monosaccharide units, enzymes, and key intermediates are highlighted with yellow, blue, and red backgrounds, respectively. (B) Expression level of genes encoding the related enzymes.

The biosynthesis of polysaccharides in C. barometz can be delineated into three main stages. First, Glc-1P is transformed into UDP-Glc by the action of UDP-glucose pyrophosphorylase. Concurrently, Fru-6P is indirectly converted to GDP-Man by using a series of enzymes involving sacA, HK, and scrK. Second, GDP-Man is further converted to GDP-Fuc, while other NDP-sugars undergo conversion facilitated by NDP-sugar interconversion enzymes (NSEs), including GALE, UGE, UXE, and GMDS. Finally, polysaccharide chains are formed using diverse GTs, which help in attaching NDP-sugars to the sugar residues of various polysaccharides and glycoconjugates.

In this study, we conducted an analysis of the gene expression levels of enzymes crucial for polysaccharide synthesis. Although both fructokinase and hexokinase are implicated in the synthesis of fructose-6-phosphate, fructokinase exhibited a higher gene expression level compared to hexokinase (Figure 6B). Overall, the gene expression patterns of these enzymes were similar between the root and rachis, yet they were slightly reduced compared to those in the pinna. We further examined the correlation between the gene expression levels of these enzymes and the content of polysaccharides. The results showed that the polysaccharide content was significantly and positively correlated with the expression level of UXE (p < 0.05), but significantly and negatively correlated with the expression levels of PGM (p < 0.05) and UER1 (p < 0.01).

The CAZy is a comprehensive resource for enzymes that are involved in the synthesis or degradation of complex carbohydrates and sugar conjugates. The carbohydrate-active enzymes are classified into groups, including GTs, glycoside hydrolases (GHs), glycolipases (CEs), polysaccharide lyases (PLs), and carbohydrate-binding modules (CBMs). In our study, a total of 3180 sugar-related unigenes were annotated, which were distributed as: 822 (26%) were identified as glycosyltransferase genes, 1620 (51%) as glycoside hydrolase genes, 135 (4%) as CE genes, 41 (1%) as polysaccharide lyase genes, and 562 (18%) as carbohydrate-binding modules (Figure S4A). GTs constituted a large proportion of gene families engaged in carbohydrate synthesis. We observed that the pinna had a higher number of upregulated GT genes, with 149 and 202 more than the root and rachis, respectively (Figure S4B). We detected 121 GT genes that were upregulated in the root when compared with those in the rachis and pinna. However, the rachis had a relatively lower number of up-regulated GT genes, with 38 and 72 fewer than the root and pinna, respectively.

We identified 36 full-length protein sequences annotated as UDP-glycosyltransferase (UGT), members of the GT1 family within the GT superfamily. The UGTs have a characteristic PSPG box marked by a highly conserved sequence. Additionally, 18 unigenes were found to be upregulated in the root compared to both the pinna and rachis, while 15 unigenes were upregulated in the pinna, relative to the root and rachis (Figure 7A). We performed a MEMESuite motif analysis on the 36 UGT protein sequences and subsequently constructed a neighbor-joining tree (NJ tree), which branched into three groups. This analysis revealed three conserved protein motifs, exhibiting similar relative positions across the sequences (Figure 7C). We also predicted the theoretical isoelectric points and molecular weights for the 36 UGTs, which varied from 30 to 61 kDa and 5.13 to 6.44, respectively. Additionally, the UGT structure of transcript_HQ_CB3_Root_transcript9878/f8p0/2216 was characterized by using the UGT85H2 (PDBID: 2pq6) of Medicago truncatula as a reference. The 3D model of our UGT displayed similar Rossmann-type folds in both the N- and C-terminal domains (Figure 7B). The “HCGWNS” residues, which are highly conserved, were labeled as magenta dots. Based on the SWISS-PROT annotation, this sequence demonstrated a high degree of homology with the UGT of Catharanthus roseus (PDBID: 8wrj; Genbank accession No.: AHK60847).

Figure 7.

Feature analysis of UGT genes. (A) Expression level. (B) Three-dimensional (3D) model of UGT protein. The conserved HCGWNS residues are shown in magenta cartoons. (C) Phylogenetic tree and conserved motif structure.

The cellulose synthase superfamily encompasses the cellulose synthase (CesA) family and eight cellulose-like synthase (CslA-CslH) families. Their related genes are involved in the biosynthesis of mannan. In this study, we identified 41 CesA genes exhibiting differential expression among the three organs. Among these, six CesA genes were significantly upregulated in the pinna (Figure 8A). Our findings also showed that CesA gene expression was the highest in the pinna, followed by the root and rachis.

Figure 8.

(A) Gene expression level of cellulose synthase. (B) Phylogenetic tree of cellulose-like synthase.

Additionally, we identified 48 putative Csl genes that are members of the CslA, CslC and CslD families. We obtained 19 full-length protein sequences annotated as cellulose-like synthases, which were used to construct the NJ tree. The tree was grouped into three branches with >99% support (Figure 8B). Among these, nine sequences were found to be part of the CslA family cluster, whereas six others were grouped into a single class of the CslD family. The remaining four sequences were assigned to the CslC family cluster. We found the Glyco-tranf-GTA-type superfamily domain in the seven CslA protein sequences, with one of these sequences also featuring the PLN02190 superfamily domain. In the case of the CslC protein sequences, the CESA_CaSu_A2 domain was found in four of them, and one of them additionally contained the conserved PLN02190 superfamily domain. Among the six CslD protein sequences, five had lengths nearly identical to the PLN02248 (Cellulose synthase-like protein) domain. The sixth sequence was found in the PLN02248 superfamily domain (Figure S5A). The expression patterns of the 19 Csl genes are shown in Figure S5B. The nine CslA genes exhibited higher expression in the pinna than in the rachis and root. In contrast, the four CslC genes displayed higher expression in both the pinna and root than in the rachis. Among the six CslD genes, four demonstrated similar expression levels across the root, pinna, and rachis, while the remaining two genes showed the highest expression in the root.

There were 19 complete protein sequences annotated as sucrose synthase (SUS) and used to construct the NJ tree. CD-Search was used to identify conserved domains (Figure S6A). The PLN00142 domain, which is characteristic of sucrose synthase, was detected in 17 of these sequences. Over 90% of the amino acid residues were found to be highly conserved across the three identified protein motifs (Figure S6B). In addition, threonine (T) was observed at a high frequency in these conserved protein motifs, which is closely related to the hydrophilic properties of sucrose synthase.

As shown in Figure S6C, two genes displayed significantly high expression in the root, while another gene showed a similar trend in the pinna. Among the genes analyzed, the ten with the highest expression levels were predominantly located in the root, followed by the pinna and rachis.

2.9. Flavonoids Biosynthesis Pathway in C. barometz

The “Other secondary metabolite synthesis” subclass within the KEGG metabolic pathway was annotated with 294, 206, and 243 unigenes in the root, rachis, and pinna, respectively. This subclass contained several metabolic pathways, with the largest number of unigenes, 114 in the root, 86 in the rachis, and 102 in the pinna, being assigned to the phenylpropanoid biosynthesis pathway (ko00940). Subsequently, the flavonoid biosynthesis pathway (ko00941) accounted for 60 unigenes in the root, 26 in the rachis, and 37 in the pinna.

We specifically characterized the flavonoid biosynthesis pathway and flavonoid-related genes (Figure 9A). The pathway initiates with the metabolite phenylalanine. Phenylalanine is transformed into cinnamic acid by phenylalanine ammonia-lyase (PAL, 21 unigenes). Then cinnamic acid is converted into P-coumaryl-CoA by cinnamic acid 4-hydroxylase (C4H, 34 unigenes) and 4-coumaryl-CoA ligase (4CL, 14 unigenes). A pivotal step in flavonoid synthesis is the formation of chalcone, catalyzed by chalcone synthase (CHS, 18 unigenes). Chalcone serves as the precursor for various flavanols by the actions of enzymes, including chalketone isomerase (CHI, 2 unigenes), coumaryl quinine 3′-monooxygenase (C3′H, 7 unigenes), flavonoid 3′-monooxygenase (F3′H, 1 unigene), flavonoid 3′5′-hydroxylase (F3′5′H, 1 unigene), and caffeoyl coenzyme A-O-methyltransferase (CCoAOMT, 7 unigenes).

Figure 9.

(A) Proposed pathway for flavonoid biosynthesis. (B) Conserved motifs of PAL proteins.

In our analysis of gene expression levels associated with flavonoid synthesis, we observed that the two CHI genes were highly expressed across the root, rachis, and pinna. The root displayed notably higher expression for the majority of the C4H genes and a slightly elevated expression for the C3′H genes when compared to the pinna and rachis, respectively. Of the 14 HCTS identified, nine were significantly highly expressed in the pinna, while four were significantly highly expressed in the root, and one showed high expression across all three organs. Most of the CHS genes were highly expressed in either the root or pinna, with the rachis showing very low expression. Almost all the CCoAOMT genes displayed increased expression in the root and pinna compared to the rachis. A single F3′H gene was found with high expression specifically in the root. Notably, there was a single F3′5′H gene significantly more highly expressed in the rachis than in the root and pinna (Figure S7). In summary, most of the genes related to flavonoid synthesis were highly expressed in the root and pinna. A negative correlation was found between the flavonoid content and the expression of genes related to flavonoid synthesis (p < 0.05). Furthermore, we identified three conserved motifs in eight PAL (Figure 9B), eight C4H, and four 4CL intact protein sequences (Figures S8 and S9). Almost half of the amino acid residues within these conserved motifs showed a high degree of conservation.

The MYB-bHLH-WD40 complex plays a major transcriptional role in governing flavonoid biosynthesis in plants. In this study, we identified 34 full-length MYB and 54 full-length bHLH transcription factor genes. Additionally, we identified seven genes encoding WD40 repeat proteins. Using the 34 complete MYB protein sequences and their conserved domains, we constructed an NJ tree. This analysis led to the identification of 21 sequences containing the PLN03091 superfamily domain (Figure 10). Furthermore, we found six sequences possessing the Myb_DNA-bind_6 domain belonging to the DNA-binding sites in MYB transcription factors, and 12 sequences containing the REB1 superfamily domain, which is characteristic of the MYB superfamily.

Figure 10.

Phylogenetic tree and conserved domain of MYB transcription factors.

We successfully identified 54 complete protein sequences encoding bHLH transcription factors (Figure S10). Among these, one (transcript_HQ_CB3_Root_transcript12038/f2p0/2025) contained an N-terminal bHLH-MYC_N domain, which is crucial for the interaction between bHLH and MYB transcription factors. A typical bHLH_SF superfamily domain was present in 22 sequences. In addition, we identified several other domains: a bHLH_AtNAI1_like domain (in one sequence), bHLH_AtIND_like domain (four sequences), bHLH_AtbHLH_like domain (six sequences), bHLH_AtSAC51_like domain (seven sequences), bHLH_AtBPE_like domain (seven sequences), bHLH_AtBPE_like domain (five sequences), and bHLH_AtILR3_like domain (two sequences). One sequence was annotated as a DNA translocase and contained a PRK10263 superfamily domain.

We constructed an NJ tree using seven WD40 repeat protein sequences, which facilitated the identification of conserved domains (Figure S11A). Among these, six were found to contain the typical WD40 conserved domain. Four sequences contained the Katanin_con80 domain, which belongs to the microtubule shear protein subunit family. Additionally, two sequences were noted to have an N-terminal LisH domain. Using homology modeling with the template PDBID: 2ymu.1.A, we predicted a three-dimensional structure of the WD40 repeat proteins (transcript_HQ_CB3_Root_transcript10464/f2p0/2196 was selected as an example) (Figure S11B). The template was selected based on its high sequence similarity, structural conservation of the WD40 domain, and overall high quality.

2.10. Lignin Synthesis Pathway in C. barometz

We found the presence of two lignin synthesis pathways in C. barometz: the P-hydroxyphenyl lignin (H-lignin) and Guaiacyl lignin (G-lignin) pathways (Figure 11). However, the absence of ferulate-5-hydroxylase (F5H) precluded the occurrence of the syringyl lignin (S-lignin) pathway [49]. In the lignin synthesis pathway, we observed higher expression levels of CCR and CAD in the rachis and root of C. barometz. Meanwhile, COMT and CCoAOMT displayed the highest gene expression levels in the pinna and root, respectively (Figure 11). Phylogenetic analysis indicated that the two CCR protein sequences were closely related to those found in the tree fern A. spinulosa (Figure S12).

Figure 11.

Pathway of lignin synthesis of C. barometz (the colors in the square present the gene expression level of the pinna, root, and rachis). The pathways were adopted from the KEGG map and rendered by Pathview (version 1.12.0).

After filtering out unexpressed gene sequences and incomplete protein motifs, and deleting redundant genes with 95% similarity, we identified 101 NAC and 62 WRKY transcription factor gene sequences. Based on the complete protein sequences, we constructed NJ trees for NAC (40) and WRKY (12) transcription factors, and searched for their respective domains (Figure S13A,B). Among the 37 NAC sequences, conserved NAM domains were identified near the N-terminus, which are likely involved in DNA binding. Moreover, 11 NAC78 and 15 NAC53 protein sequences were found to form distinct clusters, grouping as sister clades. Of the 12 WRKY sequences, 11 contained the characteristic WRKY domain, while three of these also contained plant_zn_clust domains.

We explored the correlations between the expression levels of phenylpropanoid synthase genes and those of NAC, MYB, and WRKY transcription factors (Figure S14). The MYB3R4 transcription factor displayed a significant correlation with the enzyme-encoding gene F3′5′H, and WRKY1 showed a significant correlation with the enzyme-encoding gene F3′H (p < 0.01). The correlation between MYB3R4 and F3′5′H was negative, while that between WRKY1 and F3′H was positive. Furthermore, the expressions of MYB56 and the chalcone isomerase gene (CHI), MYB3R4 and 4CL, and MYB56 and CAD demonstrated significant positive correlations (p < 0.05). In addition, NAC72, NAC78, and WRKY35 were found to have positive correlations with the PAL expression (p < 0.05), and WRKY65, WRKY69, and WRKY71 showed positive correlations with the expression of C4H (p < 0.05). Lastly, WRKY46 and MYB transcription factor OD01 were positively correlated with the expression of F3′H (p < 0.05).

2.11. Real-Time Quantitative RT-PCR Verification

To accurately estimate gene expression levels, we randomly selected key enzyme-encoding genes (HK, PGM, CAD, and HCT) for qRT-PCR analysis. The qRT-PCR results corroborate the gene expression patterns derived from transcriptome sequencing (Figure 12).

Figure 12.

The expression levels of four genes in three organs (mean ± SD, n = 3). (A) Hexokinase, HK. (B) Phosphoglucomutase, PGM. (C) Hydroxycinnamoyl transferase, HCT. (D) Cinnamyl alcohol dehydrogenase, CAD.

3. Discussion

Cibotium barometz is a renowned traditional Chinese medicinal fern with a diverse range of pharmacological properties. However, the underlying biosynthetic pathways of its bioactive ingredients and the factors that influence stem development remain elusive, probably due to the lack of a reference genome. In this study, we first identified several potential genes implicated in the biosynthesis of key active components, including polysaccharides and flavonoids. Then we analyzed the genes associated with lignin biosynthesis, given its important role in stem growth.

Given the absence of a genome sequence for C. barometz, our study underscores the significance of leveraging transcriptome sequencing to uncover its genetic landscape.

This study represents the first to use a combined approach of PacBio and Illumina sequencing to explore the full-length transcripts and gene expression across three organs of C. barometz. The transcriptome data reveal the number and types of genes active in each organ, shedding light on potential metabolic pathways. Transcriptome sequencing yields an abundance of cDNA sequences, providing useful molecular data for genetic research [50]. PacBio sequencing stands out as a highly reliable and efficient method for obtaining full-length transcriptome data, especially for non-model plants that lack reference genome information [50]. Given the absence of a genome sequence for C. barometz, our study chooses to uncover its genetic information by transcriptome sequencing. We have identified 12,718, 21,341, and 11,441 high-quality unigenes in the root, rachis, and pinna, respectively. Illumina sequencing, known for its base-level accuracy, ensures the accuracy of ultra-long reads produced by PacBio sequencing.

3.1. Gene Annotation and Structure

We have uncovered a large number of transcription factors in the transcriptome of C. barometz, encompassing C3H, bHLH, AP2/ERF, C2H2, NAC, MYB, and WRKY families. These regulatory proteins are pivotal in orchestrating plant growth, development, and responses to stressors. C3H is involved in RNA processing, hormone regulation, and post-transcriptional modifications, thereby mediating plant resistance against abiotic stresses such as low temperature, drought, and salt stress [51]. The bHLH family mainly influences plant growth and development, metabolic pathways, signal transduction, abiotic stresses, photomorphogenesis, shade-avoidance reactions, and the synthesis of secondary metabolites [52]. The AP2/ERF superfamily, with AP2 at its core, is primarily engaged in developmental regulation [53]. Within the MYB family, members like MYC2, MYC3, and MYC4 regulate a range of responses, including root growth, leaf senescence, secondary metabolism, and defense responses [54]. The NAC family, comprising NAM, ATAF, and CUC2, is one of the largest transcription factor families unique to plants, playing a key role in growth, development, and responses to both biotic and abiotic stresses, with a significant correlation with flavonoid content as observed in Ginkgo biloba [55]. In this context, we have hypothesized that transcription factors may enhance the adaptability of C. barometz to adverse environmental conditions.

PacBio sequencing has advanced our understanding of lncRNAs in Cibotium barometz. In this study, we identified 6397, 6506, and 4818 lncRNAs in the root, rachis, and pinna, respectively. Although lncRNAs are not as abundant in plants as in other organisms, they are instrumental in a variety of biological processes and environmental responses, such as gene transcription, post-transcriptional modification, transcriptional activation, epigenetic regulation, transcriptional interference, and cell cycle control [35,56]. For instance, the lncRNA (SABC1) in Arabidopsis thaliana has been found crucial in balancing plant defense and growth by regulating the biosynthesis of salicylic acid [57]. In tomatoes, lncRNAs have been shown to respond to salt stress by engaging the abscisic acid signaling cascade [58]. In addition, drought-induced lncRNAs (DRIRs) in A. thaliana positively regulate the plant’s response to both drought and salt stress [59]. Thus, it is reasonable to postulate that the multitude of lncRNAs in C. barometz are critical in regulating its growth and adaptation to diverse environmental challenges.

SSRs serve as evolutionary tuning knobs and offer advantages in adaptive evolution [60]. In the present study, 6796, 14,545, and 7395 SSRs have been identified in the root, rachis, and pinna, respectively. The abundance of SSRs may enable C. barometz to adapt to harsh environments. It is hypothesized that eukaryotes with more DNA repeats can adapt more swiftly to environmental changes [61,62]. SSR can function as “tuning or control knobs” to progressively adjust gene expression through the variation in its number of repeats [63]. The presence of more repeats may lead to more refined tuning or controlling effects [64]. Moreover, SSR sequences are recognized as recombination hotspots [65], especially dinucleotide repeats, which have favorable recombination sites and display a strong affinity for recombination enzymes [66]. Some SSR sequences directly impact recombination rates by altering the structure of DNA [62]. In this study, dinucleotide repeats emerged as the predominant SSR type across the three organs, with 8142 in rachis, 4112 in pinna, and 1635 in root. Furthermore, SSRs are not only cost-effective but also easy to use as genetic markers. They have been widely used in molecular breeding, gene mapping, and phylogeographic studies. Our analysis of SSRs lays a solid foundation for future molecular biological research on C. barometz.

3.2. Differential Expression Between Organs

The highest number of DEGs, totaling 9077, was identified when comparing the pinna to the root. Subsequent enrichment analysis revealed distinct patterns among DEGs across the three organs. Notably, the upregulated DEGs in the pinna, as opposed to those in the rachis, were significantly enriched in genes associated with photosynthesis and the photosystem, consistent with the pinna’s role in photosynthetic function. Furthermore, the upregulated DEGs in the root compared to those in the pinna and rachis were predominantly involved in the metabolic pathways of secondary compounds, suggesting the root’s significance as a medicinal component.

3.3. Genes and Pathways Associated with Polysaccharide Biosynthesis

In the transcriptome analysis, a high percentage of genes were annotated, with 96.17% in the root, 95.17% in the rachis, and 97.19% in the pinna. Functional categorization using KEGG, GO, and KOG indicated that most of these genes were enriched in secondary metabolic processes. Specifically, 145 unigenes in the root, 238 in the rachis, and 137 in the pinna were enriched in the “starch and sucrose metabolism” pathway, whereas 104 in the root, 113 in the rachis, and 98 in the pinna were enriched in the “amino sugar and nucleoside sugar metabolism” pathway. These pathways are integral to the biosynthesis of polysaccharides in C. barometz, highlighting their biological significance.

Polysaccharide biosynthesis depends on a multitude of glycosyltransferases (GTs), enzymes that are ubiquitous in plants and can produce glycoside compounds by transferring the glycosyl groups from NDP-sugars to a variety of small-molecule compounds [66]. Acting as an essential downstream step in polysaccharide biosynthesis, these GTs catalyze the formation of polysaccharide chains from NDP-sugars [66]. According to the CAZy database, glycosyltransferases are classified into 105 families, with A. thaliana having 42 families encompassing 463 genes and Oryza sativa containing 43 families with 574 genes [67]. In this study, a total of 822 potential glycosyltransferase genes were identified, belonging to 40 families. The GT1 family, also known as the UDP-glycosyltransferase (UGT) family, comprises 112 members. These UGTs feature a highly conserved motif (PSPG) with 44 amino acids at their C-terminus, which functions as the binding site for NDP-sugars [68]. Compared to the conserved C-terminal domain, the N-terminal domain displayed greater variability, suggesting that it binds to acceptors, while the C-terminal domain accepts donor substrates [69]. UGT is also involved in glycosylation, which marks the final phase in the biosynthesis of many plant secondary metabolites, including terpenoids, flavonoids, saponins, and steroids, after cyclization and oxidation steps [70]. The interplay between glycosylation, methylation, and hydroxylation reactions enhances the solubility, diversity, and complexity of these plant-derived secondary compounds [71]. Glycosyltransferases may be involved in various metabolic pathways and responsive to diverse changes throughout the growth and development of C. barometz.

Mannose-1-phosphateguanylyltransferase (GMPP), featuring a nucleotidyltransferase domain at its N-terminus, is instrumental in the formation of glycosidic bonds through the catalysis of sugar moieties from activated donor molecules [72]. This enzyme is composed of two subunits: the α subunit (GMPPA) and the β subunit (GMPPB). GMPPA regulates the activity of GMPPB, facilitating the production of GDP-mannose, although it lacks catalytic activity itself [72]. In this study, the number of DEGs involved in polysaccharide synthesis was relatively low when comparing the root to the rachis, but it was relatively higher when comparing the root to the pinna and the rachis to the pinna. This is consistent with the findings that revealed no significant differences in polysaccharide content between the root and rachis of C. barometz, whereas a significant difference was found between the root and pinna. The pinna of C. barometz, which has the lowest polysaccharide content, exhibited the most upregulated enzyme genes related to polysaccharide synthesis compared to the root and rachis. This upregulation implies that the locations where polysaccharides are synthesized and where they accumulate may be decoupled. Additionally, our analysis revealed that the total polysaccharide content correlates positively with the expression levels of UDP-arabinose-4-epimerase (UXE), while it correlates negatively with the expression levels of phosphoglucose mutase (PGM) and 3,5 epimerase/4-reductase (UER1). These correlations deserve further exploration.

Genes from the cellulose synthase (CesA) superfamily, also known as the GT2 superfamily, have been implicated in the biosynthesis of mannan polysaccharides. This diverse group is categorized into nine subfamilies: CesA and CslA-CslH [73]. In this study, we identified 48 Csl genes across three families CslA, CslC, and CslD in C. barometz. While CesA genes encode enzymes that synthesize cellulose, Csl genes are involved in the synthesis of non-cellulosic polysaccharide backbones [73]. Specifically, CslA genes are known to encode the β-1,4-mannan synthases [74]; CslC genes are probably involved in the synthesis of the β-1,4-glucan backbone found in xyloglucan [75]; and the CslD, CslF, and CslH genes mainly encode the mixed-linkage (1,3;1,4)-β-glucans synthases [75]. The CesA and CslD/E/G/H subfamilies are highly conserved in plants, including orchids, characterized by 12–17 motifs and conserved cellulose synthase domains, whereas the CslA and CslC subfamilies possess only eight motifs with four conserved domains [76]. The 3D structural models of CslB/D/E/G/H show similarities to the CesA structure [77]. CslA and CslC, unlike the other six CesA/Csl members, contain four unique motifs [77]. Both CslA and CesA have been shown to positively mitigate the effects of drought stress, and CslD, along with CslA, can positively manage freezing stress [76]. Cibotium barometz, which thrives in warm and humid conditions, is sensitive to drought and low temperature stress. The study of CslA and CslD genes enhances our comprehension of their biological characteristics.

3.4. Genes and Pathways Associated with Flavonoid Biosynthesis

Flavonoids, as crucial secondary metabolites in plants, have a range of biological activities including the regulation of auxin transport, protection against ultraviolet damage, signaling in plant-pathogen interactions, and the promotion of pollen germination [78,79]. In our study, the total flavonoid content across organs was quantified, with the highest content in the pinna of C. barometz, followed by the roots, and the lowest in the rachis. A total of 60 unigenes in the root, 37 in the pinna, and 26 in the rachis were found to be enriched in the “flavonoid biosynthesis” pathway. Key genes in the flavonoid synthesis pathways, such as C4H, C3′H, HCT, F3′H, and CHS, were significantly upregulated in the root and pinna, potentially contributing to the flavonoid accumulation and accounting for the variations in total flavonoid content in these organs. Most of the flavonoid-synthesizing genes identified in our transcriptome data are consistent with those of other plant species, such as Dendrobium huoshanense [80], Carya illinoinensis [81], and Chrysanthemum morifolium [82]. Our findings indicate that the expression pattern of CHS is negatively correlated with flavonoid content, which is in line with previous research [81]. This negative correlation observed may be attributed to the following factors: (1) Negative feedback regulation of CHS expression by accumulating flavonoids. In many plant species, the end products of flavonoid biosynthesis can act as feedback inhibitors of early pathway genes, including CHS [83,84,85]. When flavonoid concentrations reach certain thresholds, they may suppress their own biosynthesis through post-translational modifications or transcriptional repression; (2) CHS genes typically exist as multigene families in plants, often exhibiting distinct spatial and temporal expression patterns [83]. In C. barometz, we identified 18 CHS unigenes across the three organs, with varying expression levels in root, pinna, and rachis. This complexity of the multigene family may result in functional specialization among different CHS isoforms, where certain family members might be highly expressed without necessarily correlating with overall flavonoid content due to regulatory feedback mechanisms; and (3) The observed negative correlation may reflect tissue-specific post-transcriptional and post-translational regulatory mechanisms.

The biosynthesis of flavonoids in plants is governed not only by the expression of key structural genes such as PAL, CHS, and CHI but also by transcription factors that modulate gene expression, including MYB, bHLH, and WRKY. Several studies have shown that the MBW complexes, consisting of MYB, bHLH, and WD40 proteins, play a regulatory role in flavonoid biosynthesis. For instance, two types of MBW complexes, MYB5 and MYB10, greatly enhance the accumulation of anthocyanins and proanthocyanidins in strawberries [86], while the AN2/PH4 (MYB)-AN1 (bHLH)-AN11 (WD40) complex facilitates the biosynthesis of anthocyanins in petunia petals [87]. In Arabidopsis, R2R3-MYB transcription factors like PAP1 and PAP2 are implicated in the regulation of flavonoid synthesis [88]. The MBW protein complex, which forms through the interaction of bHLH with R2R3-MYB transcription factors and TTG1, regulates the expression of genes in the anthocyanin biosynthesis pathway [88]. In addition, the expression of the R2R3MYB transcription factor correlates positively with flavonoid synthesis during grape development [89]. Transcription factors such as MYB1, MYB5, MYB9, MYB10, and MYB11 regulate strawberry flavonoid biosynthesis [90], while MYB10 and MYB41 affect anthocyanin accumulation in woodland strawberries [91]. In this study, we have identified 68 MYB-encoding genes, 114 bHLH-encoding genes, and seven WD40-encoding genes in C. barometz, all of which are useful for exploring its flavonoid biosynthesis.

3.5. Genes and Pathways Associated with Lignin Synthesis

Lignin synthesis is linked to the development of the erect stems of tree ferns, such as Cibotium barometz. This complex polymer is vital for providing structural integrity to plants, facilitating their growth and long-distance water transport [92]. Two lignin synthesis pathways, H-lignin and G-lignin, have been found fundamental to the vascular systems of plants [93]. Our analysis revealed that they are also vital in C. barometz. Well-lignification is essential for ensuring the vertical transport of water and for limiting lateral water movement, serving as a defense response against drought stress [94]. Additionally, lignin also functions as a barrier against pathogens, especially following injury or during pathogen attack [95]. A key step in lignin biosynthesis is the synthesis of phenylpropanoid, an important intermediate in the production of many plant metabolites [94].

This study reveals a unique characteristic of the lignin synthesis pathway in C. barometz as a fern species. Our findings demonstrate that C. barometz possesses only the H-lignin and G-lignin synthesis pathways, lacking the S-lignin pathway due to the absence of F5H. This characteristic may have significant implications for the structural properties of the plant’s medicinal supportive structures, particularly the rhizome. First, the H/G-type lignin provides adequate structural support for the fern’s erect growth form, contributing to the characteristic texture and resilience of its rhizome [96]. Second, as H/G-type lignin H/G-type lignin exhibits greater resistance to microbial degradation compared to S-lignin-enriched lignin [97], this property may help preserve the medicinal value of the rhizome.

This study enriched 114 unigenes for the root, 86 for the rachis, and 102 for the pinna to the “phenylpropanoid biosynthesis” pathway in C. barometz. We separately annotated key genes in the lignin biosynthesis, totaling 21, 34, 14, 14, 23, 7, and 2 genes for PAL, C4H, 4CL, HCT, CAD, CCoAMOT, and CCR, respectively. Previous studies have shown that the expression levels of PAL, C4H, 4CL, CCR, and COMT are positively correlated with the lignin content in Paeonia suffruticosa [47]. Furthermore, lignin biosynthesis has been found to be related to transcription factors, including NAC, MYB, and WRKY. Specifically, NAC and MYB transcription factors are known to function in controlling secondary cell wall formation in A. thaliana [98].

The NAC family of transcription factors, such as SND1 and NST3, and their functional homologs, NST1, NST2, VND6, and VND7, are important in the regulation of lignin biosynthesis and act switches in secondary wall formation [99]. These transcription factors have an extremely conserved N-terminal domain responsible for DNA-binding, which contrasts with their variable C-terminal regions that contain transcriptional regulatory domains [100]. NAC141 is a positive regulator of lignin biosynthesis in Eucalyptus and is predominantly expressed in tissues rich in lignin, such as stems and wood. When overexpressed in A. thaliana, NAC141 leads to enhanced lignification and increased lignin content [101]. The NAC transcription factor family members NST1, NST2, and NST3 are triggers of the secondary cell wall biosynthetic program, capable of regulating downstream transcription factors like MYB46 and MYB83, which are responsible for initializing the biosynthesis genes essential for secondary cell wall deposition [102].

WRKY transcription factors constitute one of the largest TF families in plants, functioning in regulatory signaling networks, immunity, lignification, and the formation of secondary cell walls [63,103]. Our study has identified a total of 101 NAC and 62 WRKY transcription factor genes in C. barometz. Further investigation on these genes is expected to offer new insights into the regulatory mechanisms governing lignin biosynthesis in tree ferns.

4. Materials and Methods

4.1. Plant Materials and RNA Extraction

Fresh tissue samples (root, rachis, and pinna) were collected from mature C. barometz individuals on 30 December 2018 at the South China Botanical Garden, Chinese Academy of Sciences (23°11′14.3″ N, 113°22′03.5″ E), Guangzhou, Guangdong Province, China. The formal identification of the plant material was performed by Dr. Qiongmei Lu of the South China Botanical Garden, Chinese Academy of Sciences. A voucher specimen (Zhangyuli2018123101) has been deposited in the herbarium of Sun Yat-sen University (SYS; Guangzhou, China). The samples were washed immediately, dried, and then immersed in an RNAlater solution (BioTeke, Shanghai, China). The samples were stored at −20 °C until RNA extraction. Total RNA was extracted using the RNeasy Plus Mini Kit (QIAGEN, Hilden, Germany). The RNA concentration and quantity were assessed using a Qubit 2.0 fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). The RNA integrity number (RIN) was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). High-quality RNA, with concentration ≥ 300 ng/μL, quantity ≥ 4 μg, and RIN ≥ 7.0, was used for subsequent analysis. Three biological replicates for each organ were sequenced for the transcriptome analysis.

4.2. Determination of Total Polysaccharide and Flavonoid Contents

Mature C. barometz individuals were sampled. Fresh tissues (root, rachis, and pinna) were collected in five replicates (100 g per replicate per tissue). Total polysaccharides were extracted from the freeze-dried tissues based on the Chinese Pharmacopoeia [104]. The phenol-sulfuric acid method was used to quantify the polysaccharide contents (%w/w) [105].

Total flavonoids were extracted using an ultrasonic method [106]. Flavonoid content (%w/w) was determined using the sodium nitrite-aluminum nitrate-sodium hydroxide colorimetric method [107].

4.3. Illumina Sequencing and De Novo Assembly

An Illumina library was prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA). Sequencing was performed on an Illumina NovaSeq platform (Illumina, San Diego, CA, USA), generating 300 bp paired-end reads. The adapter reads with more than 10% unknown bases, and reads with more than 50% low-quality bases (QPhred ≤ 20), were filtered from the raw data. The clean reads obtained were used for self-assembly using Trinity v2.4.0, with a minimum k-mer coverage value of 3 [108]. The de novo assembly sequences were clustered into unigenes using Corset v1.05 [109].

4.4. PacBio Sequencing

We constructed separate libraries for each organ type (the root, rachis, and pinna) following the PacBio Isoform Sequencing (Iso-Seq) experimental protocol [110]. These libraries were subsequently subjected to sequencing on the PacBio Sequel II platform (Pacific Biosciences, Menlo Park, CA, USA). Connectors and data with lengths of less than 50 bp were filtered to generate subreads. SMRTlink software (version 7.0) was used to process the subreads based on the Circular Consensus Sequence (CCS) algorithm with the following parameters: –min_length 50, –max_length 15,000, –min_passes 1. The consensus sequences of the subreads were calibrated using Arrow and further polished with Illumina data to guarantee sequencing accuracy and validation using the LoRDEC software (version 0.9) with the following parameters: -k 23, -s [111,112]. The corrected transcript sequences were clustered using CD-HIT v. 4.6.8 [113], based on 95% similarity, with the following parameters: -c 0.95, -T 6, -G 0, -aL 0.00, -aS 0.99, and -AS 30.

4.5. Function Annotation of Transcripts and Prediction of Coding Sequence (CDS), Transcription Factors (TFs), Long Non-Coding RNAs (lncRNAs), and Simple Sequence Repeats (SSRs)

The final acquired full-length transcripts were functionally annotated through a homology search against seven databases: Nucleotide sequences database (NT), Protein family (Pfam), Gene Ontology (GO), Non-redundant protein sequences database (NR), euKaryotic Orthologous Groups (KOG), SWISS-PROT (a manually annotated and reviewed protein sequence database), and Kyoto Encyclopedia of Genes and Genomes (KEGG).

We performed annotation using the software ncbi-blast-2.7.1+ with a cut-off E-value of 1 × 10⁻⁵ as the threshold [114]. The Hmmscan of the HMMER 3.1 package was used for NCBI Nt and Pfam database annotations [115]. Additional database annotation was performed using Diamond v0.8.36 software [116], with an E-value of 1 × 10⁻⁵.

ANGEL v2.4 software (https://github.com/PacificBiosciences/ANGEL; accessed on 5 September 2022) was used to predict the coding sequence (CDS), with the parameter set as -min_angel_aa_length 50. We used iTAK 1.7a software (https://github.com/kentnf/iTAK/; accessed on 2 October 2022) to predict the transcription factors [117]. The parameters were set as: -f 3F. Simple sequence repeats (SSRs) were identified using MISA v1.0 (http://pgrc.ipk-gatersleben.de/misa/misa.html; accessed on 11 October 2022) [118], with minimum repeat times: mono-10, di-6, tri-5, tetra-5, penta-5, and hexa-5. We used CNCI v2 [119], PLEK v1.2 [120], CPC v0.9 [121], and PfamScan v1.6 [122] to predict the lncRNA by screening the coding potential of the non-redundant genes.

4.6. Quantification of Gene Expression and Analysis of Differentially Expressed Genes

The read count value of each gene was acquired through RSEM v1.3.0 for each organ [123,124]. Expression levels were normalized and reported as FPKM values for each unigene [125].

To identify differentially expressed genes (DEGs) between organs, read counts were normalized using the trimmed mean of M value (TMM) method [126]. Differentially expressed genes (DEGs) were identified using DEGseq [127]. The False Discovery Rate (FDR) method was applied to adjust p-values for multiple testing. FDR < 0.05 and |log2 (fold change)| ≥ 1 were selected as the threshold to detect DEG. GO enrichment analysis was performed using GOSeq v1.10.0 [128]. KEGG pathway enrichment analysis was performed using KOBAS version 2.0.12 [129].

4.7. Analysis of Gene Family

Based on functional annotations, genes related to the synthesis pathway of polysaccharides and flavonoids in C. barometz were characterized using the pathview package (version 1. 12. 0; R software) [130]. Expression heatmaps for all identified pathway genes were generated. Pathway-related genes were identified through KEGG enrichment analysis. MEGA 11.0 was used to construct a neighbor-joining tree with 1000 bootstrap replicates [131]. MEME Suite v5.3 [132] was used to characterize the enzyme, and the theoretical isoelectric point (PI) and molecular weight (kDa) were predicted using ExPASy [133]. We used CD-HIT [113] to remove redundant sequences at a 95% threshold. HMMER (http://hmmer.org/ accessed on 16 February 2023) [134] was used to predict conserved protein domains with an E value of 1 × 10⁻⁵, and they were retrieved using the CD-Search (https://www.ncbi.nlm.nih.gov/cdd/; accessed on 15 January 2023) [135] and visualized using the TBtools V1.6 [136]. Tertiary structure prediction was performed using the Swiss model (https://swissmodel.expasy.org/; accessed on 20 January 2023) [137]. Enzyme family members were identified using the CAZy database. Similarly, NAC and WRKY transcription factor family members were screened based on the annotation results. The related gene sequences were obtained through comparison with OneKP database, including those of Alsophila spinulosa, Plagiogyria japonica, Culcita macrocarpa, Adiantum aleuticum, Cystopteris fragilis, Diplazium wichurae, Gymnocarpium Dryopteris, Woodsia ilvensis, Cyclosorus acuminatus, Polystichum acrostichoides, and Mickelopteris cordata. Pearson’s correlation analysis was performed on the transcription factors and related gene expressions.

4.8. Quantitative Real-Time PCR (qRT-PCR) Analysis

One of the HK, PGM, HCT, or CAD unigenes was selected for expression validation by qRT-PCR. Total RNAs from the root, rachis, and pinna were isolated as described above and then reverse-transcribed into cDNA using the HiScript III RT SuperMix for qPCR Kit (Vazyme, Nanjing, China). Primer3Plus (version 3.3.0) was used to design the primer pairs. qRT-PCR was conducted in triplicate using a ChamQ SYBR Color qPCR Master Mix Kit (Vazyme, Nanjing, China). The PCR procedures consisted of 40 cycles at 95 °C for 30 s, 95 °C for 10 s, and 60 °C for 30 s, verified by the standard melting curve. The β-actin FL unigene was used as the reference, and relative expression of the related genes was calculated using the 2^−ΔΔCt method [138].

4.9. Statistical Analysis

One-way analysis of variance (ANOVA) was employed to assess significant differences in polysaccharide and flavonoid content among the three organs (root, rachis, and pinna) of Cibotium barometz. Post hoc tests using Tukey’s Honest Significant Difference (HSD) procedure were conducted for pairwise comparisons when significant differences (p < 0.05) were detected. Pearson’s correlation analysis was performed to examine relationships between gene expression levels and metabolite contents. Correlation coefficients (r) and corresponding p-values were calculated, with p < 0.05 considered statistically significant. All statistical analyses were executed using R software (version 4.3.1) [139].

5. Conclusions

In this study, we used PacBio and Illumina sequencing technologies to explore the transcriptomes of three distinct organs (the root, rachis, and pinna) of C. barometz. Our analysis focused on uncovering genes related to polysaccharide, flavonoid, and lignin biosynthesis, deciphering their metabolic pathways, and the underlying regulatory networks. We generated a rich resource of unigenes from the three organs, specifically characterizing TFs, SSRs, and lncRNAs, as well as their gene expression patterns and DEGs. KEGG enrichment analysis revealed that highly expressed unigenes play special roles in plant secondary metabolism. Genes related to the polysaccharide synthesis pathway display intricate patterns of gene expression, reflecting both complexity and spatiotemporal specificity. The total polysaccharide content showed a positive correlation with the expression levels of UDP-arabinose-4-epimerase (UXE) but a negative correlation with the expression of phosphoglucose mutase (PGM) and 3,5 epimerase/4-reductase (UER1). The Csl family was found to be classified into three categories: CslA, CslC, and CslD. Additionally, a conserved glycosyltransferase motif, PSPG, was also identified.

Next, we examined the flavonoids and two lignin synthesis pathways (H-lignin and G-lignin) in C. barometz, with a particular focus on the synthesis of phenylpropanoid, a key step in lignin biosynthesis. The expression of the CHS gene was found to be negatively correlated with flavonoid content. The transcription factors NAC72, NAC78, and WRKY35 displayed a positive correlation with the expression of phenylalanine ammonia-lyase (PAL). Similarly, WRKY65, WRKY69, and WRKY71 showed a positive correlation with cinnamic acid 4-hydroxylase (C4H). In contrast, MYB3R4 was negatively correlated with flavonoids 3′,5′-hydroxylase (F3′5′H). These correlations underscore the intricate and specific regulatory interplay between transcription factors and the enzyme-encoding genes they regulate. This study establishes a foundation for elucidating the synthesis and regulation mechanisms of bioactive compounds in C. barometz.

Abbreviations

4CL	4-coumaryl-CoA ligase
AXS	Apiose/xylose synthase
C3H	Coumaryl quinine 3-monooxygenase
C4H	Cinnamic acid 4-hydroxylase
CBMs	Carbohydrate-binding modules
CCS	Circular consensus sequence
CCoAOMT	Cafeoyl coenzyme A-O-methyltransferase
CDS	Coding sequence
CesA	Cellulose synthase
CHI	Chalcone isomerase
CHS	Chalcone synthase
DEGs	Differentially expressed genes
DRIRs	Drought-induced lncRNAs
F3′5′H	Flavonoids 3′,5′-hydroxylase
F3′H	Flavonoid 3-monooxygenase
F5H	Ferulate-5-hydroxylase
FDR	False discovery rate
FLNC	Full-length non-chimeric reads
FPKM	The expected number of fragments per kilobase of transcript sequence per million base pairs sequenced
FRKs	Fructokinases
galU	Uridine diphosphate glucose pyrophosphorylase
GHs	Glycoside hydrolases
Glc-1P	Glucose-1-phosphate
GMPP	Mannose-1-phosphateguanylyltransferase
GTs	Glycosyltransferases
HXKs	Hexokinases
KEGG	Kyoto Encyclopedia of Genes and Genomes
KOG	Eukaryotic orthologous groups
Iso-Seq	Isoform sequencing
lncRNAs	Long non-coding RNAs
NSEs	NDP-sugar interconversion enzymes
OMT1	Caffeic acid O-methyltransferase 1
PAL	Phenylalanine ammonia-lyase
PI	Isoelectric point
PGM	Phosphoglucose mutase
PLs	Polysaccharide lyases
qRT-PCR	Quantitative real-time PCR
RIN	RNA integrity number
SSRs	Simple sequence repeats
SUS	Sucrose synthase
TFs	Transcription factors
TMM	The trimmed mean of M value
UGDH	UDP-glucose 6-dehydrogenase
UDP-Glu	Uridine diphosphate glucose
UER1	3,5 epimerase/4-reductase
UGPase	UDP-glucose pyrophosphorylase
UGT	UDP-glycosyltransferase
UXE	UDP-arabinose-4-epimerase

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms27042050/s1.

Author Contributions

Y.S. and T.W. were responsible for conceptualization, methodology, funding acquisition, and project administration. Y.Z. and Z.W. conducted validation, formal analysis, investigation, and writing. M.L. performed a qRT-PCR experiment. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Sequence data in this study have been deposited in the National Center for Biotechnology Information with the primary accession codes PRJNA838436, SAMN43754781, SAMN43754782, and SAMN43754783. All other data supporting this article are available from the corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This work was supported by the National Natural Science Foundation of China [31872670 and 32071781], Guangdong Basic and Applied Basic Research Foundation [2021A1515010911], Science and Technology Projects in Guangzhou [202206010107], Project of Department of Science and Technology of Shenzhen City, Guangdong, China [JCYJ20210324141000001, JCYJ20230807110359040, and JCYJ20240813150103005], and Research Project of the Reform about Teaching Method and Skills from Sun Yat-Sen University and Guangdong Province.