https://orcid.org/0009-0009-6768-3372
Figures
Abstract
Cotton fiber initiation is a critical determinant of yield and quality. To decipher the genetic networks underlying this process, we conducted a time-series transcriptome analysis of a wild-type cotton (142-WT) and its fuzzless-lintless mutant (142-fl) at 0, 1, and 2 days post-anthesis (DPA). Phenotypic characterization confirmed the absence of fiber protrusion in the mutant. Through integrated analysis of differentially expressed genes (DEGs) and weighted gene co-expression network analysis (WGCNA), we identified a key module highly correlated with fiber development. This module is overwhelmingly enriched for lipid metabolism pathways, notably fatty acid elongation. Furthermore, we identified 83 high-connectivity hub genes within this network, including key enzymes like 3-ketoacyl-CoA synthases (KCS) and regulatory transcription factors. Our findings propose a novel model in which the coordinated activation of fatty acid metabolism, particularly the biosynthesis of very-long-chain fatty acids, is essential for the initiation and early development of cotton fibers. This study provides crucial insights and valuable genetic resources for the molecular breeding of cotton with improved fiber traits.
Citation: Ma X, Wang G, Dong C, Liu F (2026) Integrative transcriptomics unravels the central role of fatty acid elongation in early cotton fiber development. PLoS One 21(3): e0343644. https://doi.org/10.1371/journal.pone.0343644
Editor: Shailender Kumar Verma, University of Delhi, INDIA
Received: October 7, 2025; Accepted: February 9, 2026; Published: March 12, 2026
Copyright: © 2026 Ma et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: National Center for Biotechnology Information(NCBI) SRA database (BioProject accession number:PRJNA1033041).
Funding: This work was supported by the Natural Science Support Program of XPCC (No. 2024DA017) and the National Natural Science Foundation of China (No. 31960439). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Cotton (Gossypium hirsutum L.) is a globally significant cash crop and the primary source of natural fiber for the textile industry. Among cultivated cotton species, upland cotton (G. hirsutum) dominates commercial production due to its high yield and adaptability. Cotton fibers are singular trichomes that differentiate from the outer epidermal cells of the ovule. Their development progresses through four overlapping stages: initiation, elongation (primary wall synthesis), secondary wall thickening, and maturation [1–4]. The initiation phase, occurring around anthesis (0 days post-anthesis, DPA), is particularly crucial, as it determines the final fiber density and yield. During this period, only 25–30% of the ovule epidermal cells are fated to initiate and develop into spinnable lint fibers, while the remainder either differentiate into short fuzz fibers or remain non-fiber cells [5,6]. A comprehensive understanding of the molecular mechanisms governing this fate determination and initial protrusion is therefore fundamental for cotton improvement.
Numerous recent review papers have summarized that lint fiber initiation in upland cotton is collectively regulated by the complex cross-talk among MYB MIXTA-like TFs, sugar metabolic signaling, and plant hormones [7–9]. A model has been proposed in which no lint fibre will initiate if the combined expression levels of MYB25-like_At and MYB25-like_Dt are below a critical threshold level at 0 DPA [10]. Among the identified transcription factor (TF) genes, a few transcription factor (TF) genes have been validated to positively regulate lint fiber initiation in upland cotton, with core members including R2R3-type MYB family transcription factors GhMYB109 [11] and GhMYB25-like [12]. Despite the preliminary establishment of the core regulatory network underlying fiber initiation through these studies, the dynamic expression patterns of key genes during the initiation and development of cotton fibers remain to be fully elucidated.
With the continuous development of sequencing technology, a large number of genes related to fiber development have been discovered using transcriptome sequencing technology [ 2,13–15]. These studies revealed the mechanism of fiber cell initiation, explained the process of fiber development, and provided strong evidence. One study used transcriptome analysis of the ovule epidermis of XZ-142 wild type and 142-fl mutants and found DEGs related to the regulation of cell membrane formation during the initial stage of fibers [16]. In addition, other study used iTRAQ technology to sequence the transcriptome of XZ-142 and its downy mutant, and identified a total of 5045 proteins, 202 of which were differential proteins between wild type and mutants. Further analysis found that these differential proteins were mainly involved in fatty acid metabolism, ribose metabolism, flavonoid synthesis, signal transduction and other pathways [17]. The development of cotton fibers is divided into 4 stages, among which the initiation of fiber cell differentiation is particularly important [18]. Current experimental methods used in some previous studies have limitations. For some differential genes identified in transcriptome sequencing experiments, long fibers and short fibers cannot be distinguished.
The advent of high-throughput sequencing has significantly advanced our understanding of fiber development. Transcriptome analyses have identified numerous genes associated with various fiber developmental stages [2,14]. Specifically, studies utilizing the fuzzless-lintless Xu142-fl mutant and its wild-type counterpart Xu142 have been instrumental in pinpointing genes involved in the initiation process. For instance, Wu et al. [16] identified differentially expressed genes (DEGs) related to cell membrane formation, while a subsequent proteomic study by Ma et al. [17] highlighted the involvement of differential proteins in fatty acid metabolism and signal transduction. However, many of these prior investigations were limited to single time points or lacked the systems-level analysis required to resolve the dynamic regulatory networks and identify key drivers.
To address these gaps, we conducted a time-series transcriptome analysis of the Xu142 wild-type and Xu142-fl mutant ovules across the critical 0–2 DPA window. By integrating differential expression analysis with Weighted Gene Co-expression Network Analysis (WGCNA), we moved beyond descriptive gene lists to delineate key regulatory modules and high-confidence hub genes. This study specifically found that lipid metabolism, particularly the fatty acid elongation pathway, plays a previously underappreciated role in the earliest stages of fiber morphogenesis. Our integrated approach provides novel insights into the functional gene networks that orchestrate the transition from fiber cell fate determination to the onset of elongation.
Materials and methods
Plant materials and treatments
The materials selected long fiber and short fluff land cotton Xuzhou 142 (142-WT) as the test group, without fluff and lint Xuzhou 142 mutant (142-fl) as the control group, and three biological replicates were set up in each group. The materials were obtained from the experimental base of Xinjiang Academy of Agricultural Sciences. On the day of flowering, cotton bolls were repeatedly labeled (0 DPA). Cotton bolls of both genotypes were harvested at 0, 1, and 2 DPA, respectively, and the ovules were removed and immediately placed in liquid nitrogen, and then stored in a −70 °C ultra-low temperature refrigerator.
Scanning electron microscopy (SEM) analysis
The samples of 142-WT and 142-fl at three stages (0, 1, and 2 DPA) were observed using a Phenom SEM (Scanning Electron Microscope). The collected ovule samples were fixed in 2.5% glutaraldehyde solution at 4 °C for 3 hours, followed by immersion in 0.5–1.0% OsO4 for 1 hour. Subsequently, the samples were further dehydrated through an ethanol gradient dehydration process, with each concentration step lasting 30 minutes, and then subjected to freeze-drying. The samples were then mounted on conductive rubber sample holders. Using a metal coating method, the sample holders were placed in an ion sputtering instrument for low-vacuum ion sputtering to form a conductive film on the sample surfaces. Finally, the samples were inserted into the sample holder of the Phenom SEM and placed in the microscope chamber for observation.
RNA extraction and RNA sequencing
Place the freshly collected fiber samples into centrifuge tubes, immerse them in dry ice, securely package the tubes, and ship them to Novogene Bio-Technology Co., Ltd. in Beijing for RNA extraction and subsequent sequencing using the Illumina platform. Three biological replicates were set for each sample during RNA-sequencing. Data quality results for each treatment are presented in S1 Table. Total RNA was extracted from fiber samples of wild-type (WT) and fl mutants using the Tiangen RNA Extraction Kit (Tiangen Biotech, Beijing, China). Residual genomic DNA was removed by treatment with DNase I. RNA integrity and potential contamination were assessed by agarose gel electrophoresis, while RNA purity and concentration were quantified using a NanoDrop 2000 microplate spectrophotometer (reflected by the OD₂₆₀/OD₂₈₀ ratio). The extent of RNA degradation was accurately evaluated with an Agilent 2100 Bioanalyzer. After passing quality control, RNA sequencing was conducted on the Illumina HiSeq 4000 platform. Raw data were processed using a Perl script to calculate GC content, Q20, and Q30 scores, yielding high-quality clean data for downstream analyses. Paired-end clean reads were aligned to the TM-1 reference genome using HISAT2 software [19]. Gene-level read counts were quantified with FeatureCounts v1.5.0-p3. Gene expression levels were estimated based on gene length and the number of mapped reads, and were reported as fragments per kilobase of transcript per million mapped reads (FPKM) [20].
Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR)
The same RNA samples used for RNA-seq were also utilized for RT-qPCR. A total of 500ng RNA was used to synthesize cDNA with oligo(dT) primers and M-MLV Reverse Transcriptase (Takara, Dalian, China) following the manufacturer’s instructions. The cDNA was then aliquoted into a 96-well plate for RT-qPCR analysis on a Light Cycler® 480 II system (Roche, Switzerland) using Power SYBR Green PCR Master Mix (Roche, Switzerland). Each 10 µL reaction contained 1 µL of cDNA, 4 µL of each target-specific primer pair, and 5 µL of SYBR Green PCR Master Mix. The PCR conditions were as follows: 95 °C for 5 min; 40 cycles of 94 °C for 10 s, 60 °C for 10 s, and 72 °C for 10 s. Relative gene expression levels were calculated using the 2^–ΔΔCt method, with GhHistone3 as the internal reference gene. Primers were designed using the NCBI Primer–BLAST tool (available online: http://www.ncbi.nlm.nih.gov/tools/primer-blast/; accessed on 27 March 2018). The specificity of amplification products and amplification efficiencies were verified for each primer pair. Primer sequences are listed in S2 Table and the results are shown in S3 Table.
Differential expression of genes, gene functional annotation, enrichment analysis
We performed differential expression analysis using the DESeq2 R package (v1.20.0). The resulting P-values were adjusted using the Benjamini–Hochberg method to control the false discovery rate [21]. Genes with an adjusted P-value of less than 0.05 and an absolute log₂ fold change ≥ 2 were classified as differentially expressed. For Gene Ontology (GO) enrichment analysis of these DEGs, we used the cluster Profiler R package. The analysis included correction for gene length bias. GO terms with an adjusted P-value below 0.05 were considered significantly enriched. We also evaluated enrichment of the DEGs in KEGG pathways using cluster Profiler, which applies a hypergeometric test for statistical assessment.
Gene network construction and visualization
Co-expression networks were built with the WGCNA R package (v1.29; [22]). We used the automatic block-wise network construction method with default parameters, except for several customized settings: soft-thresholding power was set to 14, minimum module size to 30, and merge cut height to 0.25. TOMType was also specified. The eigengene value for each module was calculated and used to evaluate correlations across samples. We also determined total connectivity, intramodular connectivity (via the Soft Connectivity function), and module membership (KME, based on eigengene connectivity), along with their corresponding p-values. The genes assigned to each module are listed in S1 Table. Finally, the resulting networks were visualized using Cytoscape [23].
Results
Phenotypic characterization of cotton fiber mutants
There was no difference in plant morphology during the growth of 142-WT and 142-fl plants. During reproductive growth, there are no significant differences in flowering stage, petal color, and stamen size between two types plants, but significant differences in ovule appearance and development. To identify developmental differences between 142-WT and 142-fl fibers, we observed ovaries under electron microscopy and compared changes in wild-type and mutant ovary surfaces at 0, 1, and 2 DPA developmental stages. When 142-WT is at 0 DPA, a large number of convex spherical cells are formed. At 1 DPA, the bulging cells gradually enlarged and became oval, villus appear on the raised cells, and these raised cells may form long fibers. In contrast, 142-fl had no fiber production at 0, 1 and 2 DPA before flowering (Fig 1).
SEM images of ovules at 0, 1, and 2 DPA in 142-WT (left) and 142-fl (right) were captured. All ovules were collected from the same location on comparably located cotton bolls on each plant. Scale bar graphs are shown at magnifications of 300 μm and 80 μm.
Quality control analysis of the transcriptome during cotton fiber lint development
In order to explore the molecular mechanism of upland cotton fiber initiation and development, we compared the ovule transcriptomes in the three stages mentioned above, sequenced a total of 18 RNA libraries, and obtained a total of high-quality clean reads ranging from 39421086 to 47226244. The obtained high-quality clean reads were compared with the upland cotton genome database TM-1 WHU (https://yanglab.hzau.edu.cn/CottonMD/). Total mapped was 89.12% ~ 96.51%, the proportion of Multiple Mapped was 3.82% ~ 4.85%, and the proportion of Uniquely Mapped was 85.03% ~ 92.32% S1 Table. The high-quality RNA-seq data can therefore be used confidently for further analysis.
Differential gene expression analysis during cotton fiber elongation and development
Using |log₂ FC| ≥ 2 and P-value < 0.05, the DEGs between 142-WT and 142-fl were analyzed and compared. All 708 DEGs were found in WT0 vs. fl0, of which 360 genes were up-regulated and 348 genes were down-regulated. A total of 1327 DEGs were discovered in WT1 vs. fl1, of which 955 genes were up-regulated and 372 genes were down-regulated; all 2324 DEGs were discovered in WT2 vs. fl2, of which 1078 genes were up-regulated and 1246 genes were down-regulated. At the same time, all 758 DEGs were found in WT1 vs. WT0, of which 691 genes were up-regulated and 67 genes were down-regulated. A total of 1121 DEGs were found in WT2 vs. WT0, including 853 genes. Up-regulated expression, and 268 genes down-regulated expression. A total of 212 DEGs were found in WT2 vs. WT1, of which 29 genes were up-regulated and 183 genes were down-regulated (Fig 2A).
(A) Difference comparison determines the number of up-regulation and down-regulation of DEGs. (B) Venn diagram showing DEGs during the 142-WT1 period. (C) Venn diagram showing DEGs during the 142-WT2 period. (D) Total number of non-redundant DEGs in 142-WT0, WT1, and WT2 periods.
In order to better understand the differences in each stage of the initial development of 142-WT fibers, the results showed that there were 1706 DEGs in the 142-WT1 stage, including 1356 DEGs in WT1 vs. fl1 (949 unique ones) and 758 DEGs in WT1 vs. WT0 (350 unique) and WT1 vs. fl1 and WT1 vs. WT0 totaled 408 (Fig 2B). There were 2608 DEGs in the 142-WT2 period, including 2324 of WT2 vs. fl2 (1618 unique) and 1121 of WT2 vs. WT0 (370 unique), 212 for WT2 vs. WT1 (120 unique), 640 for W2 vs. fl2 and WT2 vs. WT0, 41 for WT2 vs. fl2 and WT2 vs. W1, 28 for WT2 vs. WT1 and WT2 vs. WT0, and 23 for the three comparison groups (Fig 2C). Overall, there were a total of 3397 non-redundant DEGs across the three periods of 142-WT (Fig 2D). Among them, a total of 433 genes were differentially expressed in the three periods of 142-WT S4 Table.
DEGs were confirmed by qRT-PCR analysis
To validate the RNA-seq results, the expression levels and patterns of 12 genes, including four involved in fatty acid biosynthesis (S1A, S1B Fig), were examined by RT-qPCR. The RT-qPCR results showed strong agreement with the RNA-seq data, confirming the high reliability of the transcriptomic analysis.
Gene ontology analysis of DEGs
To better understand the functional roles of DEGs during fiber initiation and development, gene ontology (GO) analysis was performed. Using a significance threshold of P ≤ 0.05, we identified 3397 DEGs that were enriched in 100 GO terms S5 Table. We further determined that 31, 65, and 93 GO terms were enriched at 0 DPA, 1 DPA, and 2 DPA, respectively, in 142-WT fibers. Specifically, at each corresponding stage, there were 11, 16, and 42 GO terms that were uniquely enriched at 0 DPA, 1 DPA, and 2 DPA, respectively S5 Table, S2A Fig). Additionally, 2 GO terms were enriched at both 0 DPA and 1 DPA, 4 GO terms were enriched at both 0 DPA and 2 DPA, and 33 GO terms were enriched at both 1 DPA and 2 DPA. Overall, a total of 14 GO terms were enriched across all three stages in 142-WT fibers (S2B Fig).
During the WT0, WT1, and WT2 stages of 142-WT fiber development, the biosynthesis and metabolism of fatty acids and the biosynthesis and metabolism of monocarboxylic acids in the BP category were significantly enriched, and the xyloglucosyl transferase in the MF category was significantly enriched. Pathways are significantly enriched. It shows that the fiber initiation process may be related to the fatty acid pathway (Fig 3).
GO enrichment analysis of 0DPA, 1DPA, and 2DPA stage DEGs initiated by 142-W fiber. WT0, WT1, and WT2 represent the top 15 GO terms in the initial 0DPA, 1DPA, and 2DPA periods of the 142-WT optical fiber, respectively.
KEGG pathways of DEGs
We conducted a pathway analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) using 3397 non-redundant DEGs and found that these DEGs were associated with 106 pathways S6 Table. Many significantly altered pathways were related to fatty acid elongation, phenylpropanoid biosynthesis, unsaturated fatty acid biosynthesis, keratin, suberin, wax biosynthesis, cysteine and methionine metabolism. The 708 DEGs enriched in KEGG pathways at WT0 period were mapped to 74 pathways, among which 9 were significantly enriched (P ≤ 0.05). These included unsaturated fatty acid biosynthesis, fatty acid elongation, fatty acid metabolism, indole alkaloid biosynthesis, betalain biosynthesis, fructose and mannose metabolism, fatty acid biosynthesis, phenylpropanoid biosynthesis, and propanoate metabolism. The 1706 DEGs enriched in KEGG pathways at W1 period were mapped to 97 pathways, among which 8 were significantly enriched (P ≤ 0.05), including fatty acid elongation, keratin, suberin, and wax biosynthesis, unsaturated fatty acid biosynthesis, phenylpropanoid biosynthesis, cysteine and methionine metabolism, ubiquinone and other terpenoid-quinone biosynthesis, flavonoid and flavonol biosynthesis, and isoflavonoid biosynthesis. The 2608 DEGs enriched in KEGG pathways at WT2 period were mapped to 105 pathways, among which 18 were significantly enriched (P ≤ 0.05), including fatty acid elongation, phenylpropanoid biosynthesis, keratin, suberin, and wax biosynthesis, unsaturated fatty acid biosynthesis, cysteine and methionine metabolism, fructose and mannose metabolism, ubiquinone and other terpenoid-quinone biosynthesis, beta-alanine metabolism, monoterpenoid biosynthesis, starch and sucrose metabolism, leucine, isoleucine, and valine biosynthesis, sesquiterpenoid and triterpenoid biosynthesis, ABC transporters, nitrogen metabolism, leucine, isoleucine, and valine degradation, galactose metabolism, tyrosine metabolism, and alanine, aspartate, and glutamate metabolism. The number of annotated DEGs, P-values, and enriched pathways are listed in the tables.
Comprehensive analysis showed that among the DEGs in 0 DPA, 1 DPA and 2 DPA fibers, the fatty acid elongation pathway, the biosynthetic pathway of unsaturated fatty acids, and the phenylpropanoid biosynthetic pathway all showed significant enrichment in the comparison groups at three time points. According to previous studies, gene expression in pathways related to fatty acids may promote fiber development and may be involved in the synthesis of waxes in fibers. The expression of genes related to energy metabolism provides energy for the initial development of fibers, such as fructose and mannose metabolism, cysteine and methionine metabolism, propionic acid metabolism, etc. These results indicate that the initial development of cotton fibers is regulated by a complex gene network of multiple biosynthetic and metabolic pathways.
Co-expression network analysis identifies fiber initiation-related differentially expressed genes (DEGs)
WGCNA is an algorithm used to mine modules from large datasets, where each module contains a cluster of highly interconnected genes. Genes within the same cluster have high correlation coefficients. In this study, we used WGCNA to group the 3397 DEGs into 13 modules, with each module labeled with a distinctive color. The modules were defined by the main branches of the cluster dendrogram (Fig 4A), and the correlation coefficients between each module and sample features were indicated in the figure. The red module had the highest correlation with fl0 (R = 0.95, P = 9e-10); the brown module had the highest correlation with fl1 (R = 0.46, P = 0.05); the pink module had the highest correlation with fl2 (R = 0.95, P = 8e-06); the chartreuse module had the highest correlation with WT0 (R = 0.96, P = 6e-10); the salmon module had the highest correlation with WT1 (R = 0.81, P = 4e-05); and the turquoise module had the highest correlation with WT2 (R = 0.8, P = 7e-05) (Fig 4B).
(A) Hierarchical cluster tree showing co-expression modules identified by WGCNA. Each leaf on the tree represents a gene. Each major branch represents a module, with a total of 16 modules marked with different colors. (B) Module-sample association relationship. Each row corresponds to a module, labeled with the same color as in Figure (A). The number of genes in each module is shown next to the module name. Each column corresponds to a specific organization. The correlation coefficient and P value between the module and the sample or tissue are displayed at the intersection of rows and columns.
According to scanning electron microscopy observation. Cotton fibers begin to bulge around 1 DPA, so the module with the highest WT2 correlation the turquoise module is what we are most concerned about. There are a total of 1181 genes S7 Table in the turquoise module. The heat map analysis results show that the first category is the expression level of genes that gradually increases with fiber development, is significantly up-regulated in WT2, and has low expression in fl2. This type of genes accounts for the majority and the expression pattern of the second type of genes is opposite to that of the first type. As the expression level of fiber development genes gradually decreases, it is significantly down-regulated in WT2, but the expression level is higher in fl2 (Fig 5A). KEGG pathway enrichment S7 Table analysis showed that the turquoise module was significantly enriched in the following 14 pathways: fatty acid elongation; biosynthesis of cutin, suberin and wax; biosynthesis of unsaturated fatty acids; ubiquinone and other terpenoid quinone biosynthesis Synthesis; ABC transport; fatty acid metabolism; beta-alanine metabolism; cysteine and methionine metabolism; starch and sucrose metabolism; phenylpropanoid biosynthesis; stilbenes, diarylheptanes, and gingerol biosynthesis; galactose Metabolism Limonene and pinene were degraded (Fig 5B).
(A) Gene expression trend of Turquoise module. WT0, WT1, and WT2 represent the 0 DPA, 1 DPA, and 2 DPA developmental stages of 142-WT fibers, respectively. Similarly, fl0, fl1, and fl2 denote the 0 DPA, 1 DPA, and 2 DPA developmental stages of 142-fl fibers, respectively. (B) KEGG enrichment analysis of Turquoise module. (C) Co-expression network analysis of Turquoise modules. Red circles represent hub genes and transcription factors, and light blue circles represent hub genes.
Hub genes perform central roles in biological processes, often acting as pivotal regulators that influence the expression of other genes within associated pathways. According to the screening criteria KME (eigengene connectivity) ≥ 0.98 and edge weight value ≥ 0.6, a total of 83 key genes were identified S7 Table. Cytoscape software was used to visually analyze the interconnection relationships of core genes, and the results showed that all genes have high connectivity (Fig 5C). Studies have reported that the initial development of upland cotton fiber is mainly related to fatty acid elongation. In addition, transcription factors play a key role in the initial fiber development process. Therefore, the key genes include MYB Transcription factor (Ghi_D13G09836), OFP gene (Ghi_A07G06336), fatty acid elongation Very-long-chain 3-oxoacyl-CoA reductase 1 (Ghi_A03G08906), 3-ketoacyl-CoA synthase 9 (Ghi_D01G00221), 3-ketoacyl-CoA synthase 19 (Ghi_A10G13266), 3-ketoacyl-CoA synthase 20 (Ghi_D09G03516) were selected as target core genes S8 Table.
Analysis of fatty acid elongation pathways during fiber initiation and development
Combining differential gene enrichment analysis and WGCNA key module enrichment analysis, we further analyzed the role of the fatty acid elongation pathway in the initial development of cotton fibers (Fig 6A). The results show that in the fatty acid elongation pathway, long-chain fatty acids are combined with coenzyme A, and then undergo a series of enzyme-catalyzed reactions in the fatty acid β-oxidation pathway to form long-chain 3-oxoacyl-CoA, followed by fatty acid β-oxidation. During the process, the intermediate product long-chain 3-hydroxyacyl-CoA is formed, which participates in the hydroxylation reaction, and continues to be oxidized to form long-chain trans-2,3-dehydroacyl-CoA, which undergoes a dehydrogenation reaction in the process. In short, the oxidation process of long-chain fatty acids is accompanied by the release of a large amount of energy, which plays a vital role in maintaining normal energy metabolism starting from fiber. During the Greek reaction, 23 GhiKCS, 3 GhiKCR and 1 GhiPAS differential genes participated in the reactions and were highly expressed in the 142-WT2 period, thereby increasing the accumulation of fatty acids and thus affecting the initial formation of cotton fibers (Fig 6B).
(A) Fatty acid elongation pathway. (B) Expression pattern of DEGs during fatty acid elongation. The rectangles represent the expression changes of DEGs.WT0, WT1, and WT2 represent the 0 DPA, 1 DPA, and 2 DPA developmental stages of 142-WT fibers, respectively. Similarly, fl0, fl1, and fl2 denote the 0 DPA, 1 DPA, and 2 DPA developmental stages of 142-fl fibers, respectively.
Discussion
This study provides a comprehensive analysis of the molecular events governing the critical transition from fiber cell initiation to early elongation in upland cotton. By integrating time-series transcriptome data from the wild-type 142-WT and its isogenic fuzzless-lintless mutant 142-fl with WGCNA, we have moved beyond descriptive gene lists to propose a coherent model centered on lipid metabolism. Our findings consistently highlight the fatty acid elongation pathway as a cornerstone of early fiber development, a focus that provides both mechanistic insights and a refined perspective on previous observations.
The identification of four KCS genes (Ghi_A03G08906, Ghi_D01G00221, Ghi_A10G13266, Ghi_D09G03516) as high-connectivity hub genes within the key turquoise module underscores their potential role as rate-limiting regulators in this process. In higher plants, KCS is widely regarded as the rate-limiting enzyme and serves as a key determinant of both substrate and tissue specificity in the fatty acid elongation process [24,25]. VLCFAs are found in sphingolipids, seed oils, and cuticular waxes of plants. In the Arabidopsis genome, a total of 23 KCS genes of the FAE1-type and ELO-type have been identified, with the function of the ELO-type genes remaining unclear [26–29]. Studies have elucidated the biological functions and substrate specificities regulated by multiple KCS genes due to differences in C-chain length and unsaturation. For example, FATTY ACID ELONGATION 1 (FAE1), isolated from Arabidopsis fae1 mutants, encodes a β-ketoacyl-CoA synthase (FAE1 KCS) expressed exclusively in seeds and associated with lipid accumulation [30]. Xiao et al. identified 58 GhKCS genes, among which 19 genes exhibited transcriptional activation at different fiber developmental stages [31]. GBKCS3, GBKCS8, GBKCS20, and GBKCS34 exhibited consistently high expression levels across all developmental stages and overlapped with key genes identified in the Turquoise project, indicating that the KCS gene family is associated with cotton fiber elongation during the initial stages of fiber development [32]. The protein encoded by the FDH1 (FIDDLEHEAD1) gene may participate in the synthesis of long-chain lipids found in the cuticle, playing a crucial role in organ nutritional development [33]. In our study, the concerted upregulation of GhKCSs during the 1–2 DPA window in the wild-type suggests a massive demand for VLCFAs. This is not merely for the deposition of a protective cuticle on the emerging fiber, as traditionally assumed, but potentially for more dynamic roles in the initial protrusion itself. VLCFAs are critical components of membrane lipids and sphingolipid-based signaling molecules. We therefore posit that the KCS-driven elongation pathway is essential for rapid membrane biosynthesis and remodeling required for the polar expansion of the fiber cell, and possibly for generating lipid-derived signaling cues that reinforce the fiber differentiation program. This expands upon the findings of [Ma et al. [17]], who noted the involvement of fatty acid metabolism in this mutant, by pinpointing the specific enzymatic step and its central position in the co-expression network.
Beyond the KCS genes, the turquoise module harbors a suite of co-expressed genes that paint a picture of a coordinated lipid-based machinery. The presence of hub genes from the LTP (e.g., Ghi_D01G08176) and GDSL esterase (e.g., Ghi_A04G03676, Ghi_A10G03821) families supports this model. LTPs are implicated in the extracellular transfer of hydrophobic compounds, including cutin monomers, while GDSL esterases often function in cutin polymer assembly. Their co-expression with KCS genes suggests a tightly coupled system where VLCFAs are synthesized, transported, and assembled into the specialized extracellular matrix of the nascent fiber. This integrated activity would be crucial for building a robust yet flexible cell wall capable of withstanding turgor pressure during the rapid elongation phase that immediately follows initiation. The concurrent enrichment of phenylpropanoid biosynthesis, another pathway dedicated to cell wall components, further underscores the importance of building a structurally sound fiber cell from its inception.
The identification of specific transcription factors, notably the MYB protein (Ghi_D13G09836) and the OFP protein (Ghi_A07G06336), as hub genes provides tantalizing clues to the upstream regulation of this lipid-oriented program. MYB transcription factors are well-established master regulators of trichome development and the phenylpropanoid pathway in Arabidopsis, and several, such as GhMYB25, have been shown to be critical for cotton fiber initiation [ 12,34]. OFP proteins often act as transcriptional repressors and can interact with MYBs to fine-tune their activity. Their high connectivity within the turquoise module suggests they may sit atop a regulatory hierarchy that orchestrates the metabolic shift toward lipid and cell wall biosynthesis, directly activating or repressing the suite of structural genes we identified. Future work, such as Yeast One-Hybrid assays and CRISPR-Cas9 mutagenesis, will be essential to validate these TFs as direct regulators of the KCS and other hub genes, thereby closing the loop between transcriptional regulation and metabolic execution.
Conclusions
In conclusion, our multi-time-point transcriptome analysis of the 142-WT and 142-fl mutants, complemented by WGCNA, has provided a more dynamic and network-oriented view of early fiber development. We have pinpointed a key module and its associated hub genes, with strong implications for the role of fatty acid metabolism in the fiber initiation-elongation continuum. These findings shift the focus from static gene lists to dynamic network interactions and lay a solid foundation for future functional studies aimed at manipulating these pathways for cotton fiber improvement.
Supporting information
S1 Fig. Determination of the relative expression levels of 9 selected genes in two cotton varieties.
A: Heatmap showing the results of RNA-seq analysis; B: Results of RT-qPCR.
https://doi.org/10.1371/journal.pone.0343644.s010
(TIF)
S2 Fig. GO enrichment analysis of DEGs at the 0DPA, 1DPA, and 2DPA stages of 142-WT fiber initiation.
A. Venn diagram showing the number of GO terms for the 3 fiber onset periods. B. GO terms enriched in 142-WT0, W1T, and WT2. Y0 represents 142-WT 0DPA, W1 represents 142-WT 1DPA, and WT2 represents 142-WT 2DPA.
https://doi.org/10.1371/journal.pone.0343644.s001
(TIF)
S1 Table. Sample sequencing data evaluation statistics table.
https://doi.org/10.1371/journal.pone.0343644.s002
(XLSX)
S4 Table. Functional annotation of 433 genes.
https://doi.org/10.1371/journal.pone.0343644.s005
(XLSX)
S5 Table. Analysis of GO enrichment for 3397 common DEGs.
https://doi.org/10.1371/journal.pone.0343644.s006
(XLSX)
S6 Table. Analysis of KEGG pathway for 708 142-WT0 DEGs.
https://doi.org/10.1371/journal.pone.0343644.s007
(XLSX)
S7 Table. Analysis of KEGG pathway for turquoise module.
https://doi.org/10.1371/journal.pone.0343644.s008
(XLSX)
S8 Table. The FPKM of DEGs in fatty acid biosynthesis.
https://doi.org/10.1371/journal.pone.0343644.s009
(XLSX)
Acknowledgments
We thank all individuals and institutions involved in this study for their technical support, including equipment and related experimental materials, assistance with essential experimental tasks, provision of useful software, efficient guidance and review, and assistance with auxiliary tasks.
References
- 1. Graves DA, Stewart JM. Chronology of the differentiation of cotton (Gossypium hirsutum L.) fiber cells. Planta. 1988;175(2):254–8. pmid:24221720
- 2. Haigler CH, Betancur L, Stiff MR, Tuttle JR. Cotton fiber: a powerful single-cell model for cell wall and cellulose research. Front Plant Sci. 2012;3:104.
- 3. Wen X, Chen Z, Yang Z, Wang M, Jin S, Wang G, et al. A comprehensive overview of cotton genomics, biotechnology and molecular biological studies. Sci China Life Sci. 2023;66(10):2214–56. pmid:36899210
- 4. Zhang J, Mei H, Lu H, Chen R, Hu Y, Zhang T. Transcriptome time-course analysis in the whole period of cotton fiber development. Front Plant Sci. 2022;13:864529. pmid:35463423
- 5. Ruan YL, Llewellyn DJ, Furbank RT. Suppression of sucrose synthase gene expression represses cotton fiber cell initiation, elongation, and seed development. Plant Cell. 2003;15:952–64.
- 6. Wilkins TA, Arpat AB. The cotton fiber transcriptome. Physiologia Plantarum. 2005;124(3):295–300.
- 7. Huang G, Huang JQ, Chen XY, Zhu YX. Recent advances and future perspectives in cotton research. Annu Rev Plant Biol. 2021;72:437–62.
- 8. Watanabe R, Soga N, Fujita D, Tabata KV, Yamauchi L, Hyeon Kim S, et al. Arrayed lipid bilayer chambers allow single-molecule analysis of membrane transporter activity. Nature Communications. 2014;5(1):null.
- 9. Huang Y, Liu X, Tang K, Zuo K. Functional analysis of the seed coat-specific gene GbMYB2 from cotton. Plant Physiology and Biochemistry. 2013;73:16–22.
- 10. Zhu Q-H, Yuan Y, Stiller W, Jia Y, Wang P, Pan Z, et al. Genetic dissection of the fuzzless seed trait in Gossypium barbadense. J Exp Bot. 2018;69(5):997–1009. pmid:29351643
- 11. Pu L, Li Q, Fan X, Yang W, Xue Y. The R2R3 MYB transcription factor GhMYB109 is required for cotton fiber development. Genetics. 2008;180(2):811–20. pmid:18780729
- 12. Walford S-A, Wu Y, Llewellyn DJ, Dennis ES. GhMYB25-like: a key factor in early cotton fibre development. Plant J. 2011;65(5):785–97. pmid:21235650
- 13. Kim HJ, Triplett BA. Cotton fiber growth in planta and in vitro. Models for plant cell elongation and cell wall biogenesis. Plant Physiol. 2001;127:1361–6.
- 14. Li Y, Tu L, Pettolino FA, Ji S, Hao J, Yuan D, et al. GbEXPATR, a species-specific expansin, enhances cotton fibre elongation through cell wall restructuring. Plant Biotechnol J. 2016;14(3):951–63. pmid:26269378
- 15. Sampedro J, Carey RE, Cosgrove DJ. Genome histories clarify evolution of the expansin superfamily: new insights from the poplar genome and pine ESTs. J Plant Res. 2006;119(1):11–21. pmid:16411016
- 16. Wu Y, Machado AC, White RG, Llewellyn DJ, Dennis ES. Expression profiling identifies genes expressed early during lint fibre initiation in cotton. Plant Cell Physiol. 2006;47(1):107–27. pmid:16278222
- 17. Ma Q-F, Wu C-H, Wu M, Pei W-F, Li X-L, Wang W-K, et al. Integrative transcriptome, proteome, phosphoproteome and genetic mapping reveals new aspects in a fiberless mutant of cotton. Sci Rep. 2016;6:24485. pmid:27075604
- 18. Jan M, Liu Z, Guo C, Sun X. Molecular Regulation of Cotton Fiber Development: A Review. Int J Mol Sci. 2022;23(9):5004. pmid:35563394
- 19. Huang G, Wu Z, Percy RG, Bai M, Li Y, Frelichowski JE, et al. Genome sequence of Gossypium herbaceum and genome updates of Gossypium arboreum and Gossypium hirsutum provide insights into cotton A-genome evolution. Nat Genet. 2020;52(5):516–24. pmid:32284579
- 20. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010;28(5):511–5. pmid:20436464
- 21. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Stat Soc Series B: Stat Methodol. 1995;57(1):289–300.
- 22. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinform. 2008;9:559. pmid:19114008
- 23. Kohl M, Wiese S, Warscheid B. Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol. 2011;696:291–303. pmid:21063955
- 24. Denic V, Weissman JS. A molecular caliper mechanism for determining very long-chain fatty acid length. Cell. 2007;130(4):663–77. pmid:17719544
- 25. Lassner MW, Lardizabal K, Metz JG. A jojoba beta-Ketoacyl-CoA synthase cDNA complements the canola fatty acid elongation mutation in transgenic plants. Plant Cell. 1996;8(2):281–92. pmid:8742713
- 26. Joubès J, Raffaele S, Bourdenx B, Garcia C, Laroche-Traineau J, Moreau P, et al. The VLCFA elongase gene family in Arabidopsis thaliana: phylogenetic analysis, 3D modelling and expression profiling. Plant Mol Biol. 2008;67(5):547–66. pmid:18465198
- 27. Puyaubert J, Dieryck W, Costaglioli P, Chevalier S, Breton A, Lessire R. Temporal gene expression of 3-ketoacyl-CoA reductase is different in high and in low erucic acid Brassica napus cultivars during seed development. Biochim Biophys Acta. 2005;1687(1–3):152–63. pmid:15708363
- 28. Rossak M, Smith M, Kunst L. Expression of the FAE1 gene and FAE1 promoter activity in developing seeds of Arabidopsis thaliana. Plant Mol Biol. 2001;46(6):717–25. pmid:11575726
- 29. Huang G-Q, Gong S-Y, Xu W-L, Li W, Li P, Zhang C-J, et al. A Fasciclin-Like Arabinogalactan Protein, GhFLA1, Is Involved in Fiber Initiation and Elongation of Cotton. Plant Physiology. 2013;161(3):1278–90.
- 30. Ghanevati M, Jaworski JG. Active-site residues of a plant membrane-bound fatty acid elongase beta-ketoacyl-CoA synthase, FAE1 KCS. Biochim Biophys Acta. 2001;1530(1):77–85. pmid:11341960
- 31. Xiao GH, Wang K, Huang G, Zhu YX. Genome-scale analysis of the cotton KCS gene family revealed a binary mode of action for gibberellin A regulated fiber growth. J Integr Plant Biol. 2016;58:577–89.
- 32. Rui C, Chen X, Xu N, Wang J, Zhang H, Li S, et al. Identification and structure analysis of KCS family genes suggest their reponding to regulate fiber development in long-staple cotton under salt-alkaline stress. Front Genet. 2022;13:812449. pmid:35186036
- 33. Reina-Pinto JJ, Voisin D, Kurdyukov S, Faust A, Haslam RP, Michaelson LV, et al. Misexpression of FATTY ACID ELONGATION1 in the Arabidopsis epidermis induces cell death and suggests a critical role for phospholipase A2 in this process. Plant Cell. 2009;21(4):1252–72. pmid:19376931
- 34. Machado A, Wu Y, Yang Y, Llewellyn DJ, Dennis ES. The MYB transcription factor GhMYB25 regulates early fibre and trichome development. Plant J. 2009;59(1):52–62. pmid:19309462