GhFAD2–3 is required for anther development in Gossypium hirsutum
In higher plants, the FAD2 gene encodes the microsomal oleate Δ12-desaturase, one of the key enzymes essential for the biosynthesis of the polyunsaturated lipids that serve many important functions in plant development and stress responses. FAD2 catalyzes the first step, in the biosynthesis of the polyunsaturated fatty acids (PUFAs) found in the cell membrane and cell wall, and it is thus of great importance to investigate the regulatory role of FAD2 in anther development.
We reported the molecular characterization of the cotton (Gossypium hirsutum) GhFAD2 gene family and the essential role of GhFAD2–3 in cotton anther development. G. hirsutum contains four pairs of homoeologous FAD2 genes (GhFAD2–1 to GhFAD2–4). GhFAD2–3 is ubiquitously and relatively highly expressed in all analyzed tissues, particularly in anthers. Specific inhibition of GhFAD2–3 using the RNA interference approach resulted in male sterility due to impaired anther development at the stages from meiosis to maturation. The cellular phenotypic abnormality observed at the meiosis stage of the GhFAD2–3 silenced plant (fad2–3) coincides with the significant reduction of C18:2 in anthers at the same stage. Compared with that of the wild type (WT), the content of C18:1 was 41.48%, which increased by 5 fold in the fad2–3 anther at the pollen maturation stage. Moreover, the ratio of monounsaturated to polyunsaturated fatty acid was 5.43 in fad2–3 anther, which was much higher than that of the WT (only 0.39). Through compositional analysis of anthers cuticle and transcriptome data, we demonstrated it was unfavorable to the development of anther by regulating GhFAD2–3 expression level to increase the oleic acid content.
Our work demonstrated the importance of C18:2 and/or C18:3 in the development of the pollen exine and anther cuticle in cotton and provided clue for further investigation of the physiological significance of the fatty acid composition for plant growth and development.
KeywordsGossypium hirsutum Fatty acid composition GhFAD2 Male sterile Linoleic acid Oleic acid
- C16:0 DCA
- C18:1 DCA
α, ω-octadecenoic acid
- C18:2 DCA
α, ω-octadecadiendioic acid
Days post anthesis
Fatty acid desaturase 2
Fatty acid desaturase 3
Fatty acid epoxide hydrolase
Fatty acyl-coenzyme A Reductase
Fatty acid synthase
- NPT II
Neomycin phosphotransferase II gene
Cytochrome P450 monooxygenase
Polyunsaturated fatty acids
- triOH C18:1 FA
9,10,18-trihydroxy octadecenoic acid
In higher plants, the microsomal oleate Δ12-desaturase (fatty acid desaturase 2, FAD2) is a hydrophobic endoplasmic reticulum protein, catalyzing the reaction from monounsaturated oleic acid (C18:1) to polyunsaturated linoleic acid (C18:2) by introducing a double bond between the 12th and 13th carbon atoms of C18:1 . FAD2 was first identified in the model plant species Arabidopsis thaliana . Although only a single copy of FAD2 was found in A. thaliana, multiple copies of FAD2 were identified in many other plants, such as canola (Brassica napus), cotton (Gossypium hirsutum), soybean (Glycine max), olive (Olea europaea), sesame (Sesamum indicum) and sunflower (Helianthus annuus) [3, 4, 5, 6, 7, 8].
As the key gene controlling the conversion of C18:1 to C18:2, the FAD2 transcriptional level directly determines the relative content and proportion of polyunsaturated fatty acids (PUFAs) in plants . FAD2 could be significantly induced by abiotic stresses [10, 11], suggesting that PUFAs could play a key role in plant stress responses and adaptation to environmental change. Due to the oxidative instability of PUFAs, studies on the regulation of FAD2 were mainly focused on reducing its expression level to decrease the C18:2 content in oil seeds, and thus to improve oil quality [4, 12, 13, 14, 15, 16].
In flowering plants, the anther cuticle and pollen wall are rich in lipids, mainly fatty acids and their derivatives [17, 18, 19, 20, 21]. Genetic analyses of Arabidopsis and rice have found that many genes with an important role in the development of the anther cuticle and pollen wall are involved in lipid metabolism, such as MALE STERILITY 2 (MS2), FACELESS POLLEN 1 (FLP1), CYP703, CYP704B1, CYP704B2, Acyl-CoA Synthetase 5 (ACOS5), NO EXINE FORMATION1 (NEF1), Wax-Deficient Anther1 (WDA1), β-ketoacyl-coenzyme A synthase ECERIFERUM6 (CER6), Defective Pollen Wall (DPW), Fatty acyl-coenzyme A Reductase (FAR), 3-ketoacyl-CoA Synthase 9 (KCS9), and OsC6 [18, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34].
The upland cotton genome is large and complex allotetraploid (AADD; 2n = 52) and several rounds of genome duplication events have been identified in cotton genome , which makes it more difficult to analyze gene expression and regulation. In this study, based on the characterization of the cotton (G. hirsutum) FAD2 family genes, we investigated their expression patterns in various vegetative and reproductive tissues and found that GhFAD2–3 is the gene highly expressed in most tissues analyzed, particularly in anthers. We demonstrated that silencing of GhFAD2–3 resulted in male sterility, due to nonviable pollen grains and abnormal anther development resulted from significantly reduced levels of PUFAs at the meiosis and tetrad stages. This is also the first report on the dynamic changes of fatty acid constituents during cotton anther development, which were exactly opposite to those in developing and mature cottonseeds. Our work showed the effect of changes in fatty acid constituents on the physiological activity of anthers and revealed the essentialness of primary PUFAs in the development of the pollen wall and anther cuticle in cotton.
Genome-wide identification of GhFAD2 in cotton
The GhFAD2 genes in Gossypium hirsutum
Also known as
Chromosome and coordinates
GenBank accession no.
CDS length (bp)
No. of amino acid
Expression pattern of different members of the GhFAD2 gene family
Molecular characterization of transgenic plants
A 517-bp fragment targeting both homoeologous GhFAD2–3 sequences was used in making the hairpin construct. In total, 12 independent transgenic cotton lines were obtained by Agrobacterium-mediated transformation (Additional file 1: Figure S2). The transgenic cotton lines with the fusion fragment (1113 bp) from GhFAD2–3 and the gus linker were considered as positive transgenic events. To investigate the effect of downregulation of GhFAD2–3 on changes in the anther transcriptome, we compared transcriptomes of wild-type and fad2–3 using RNAs from anthers of the meiosis and tetrad stages, considering that the differences in cellular phenotypes between the wild-type and fad2–3 anthers began to be observed at the meiotic stage by transmission electron microscopy (see below). As expected, both homoeologous GhFAD2–3 (Gh_D11G3169 and Gh_A11G2814) were downregulated at both stages, although downregulation at the tetrad stage was more statistically significant (Additional file 1: Table S1). Among the other GhFAD2 genes, GhFAD2–1 (Gh_A13G1850 and Gh_D13G2238) was significantly downregulated at both stages in fad2–3 due to its closest sequence similarity with GhFAD2–3. However, compared to GhFAD2–3, GhFAD2–2D and GhFAD2–4A had a negligent expression level in anthers, and could thus hardly have function in anther development.
Significantly changed genes encoding key enzymes of these pathways were shown in Additional file 1: Table S1. These results revealed that many significantly changed pathways are related to the metabolism of lipids and their derivers; among them were those involved in biosynthesis of unsaturated fatty acids, alpha-linolenic acid metabolism, and the biosynthesis of cutin, suberin and wax. Apart from GhFAD2, DEGs related to biosynthesis of unsaturated fatty acids also included GhFAD3 (such as Gh_A09G0848 and Gh_A07G0946). The expression levels of genes involved in linoleic acid and alpha-linolenic acid metabolism was also statistically significant at the meiosis stage or tetrad stage in the fad2–3 anthers. In addition, fatty acyl-CoA reductase (FAR) catalyzes the reduction of fatty acyl-CoA to fatty alcohols, which are essential components of wax and cutin monomers. GhFAR2 genes (Gh_A09G1215), orthologous to Arabidopsis MS2 and rice DPW that are related to wax biosynthesis, were found to be downregulated at the tetrad stage in fad2–3. GhCYP86B1 (Gh_D04G1447, Gh_A04G0930, Gh_A03G2129 and Gh_D02G1587), a very long chain fatty acid hydroxylase specifically involved in cutin and suberin biosynthesis, was significantly upregulated in the fad2–3 anther. Furthermore, Cytochrome P450 CYP704B1, which participates in catalyzing omega-hydroxylation of long-chain fatty acids was also differentially expressed. Significant differences were also observed in the expression levels of some genes, such as those encoding peroxygenase and aldehyde dehydrogenase.
Phenotypic analysis of the RNAi plants
Silencing of GhFAD2–3 decreased C18:2 content in anthers
In the fad2–3 anthers, due to the significantly reduced level of GhFAD2–3, conversion of C18:1 into C18:2 was compromised after the microsporocyte stage. Consequently, a significant reduction in C18:2 was observed at the meiosis and tetrad stages, while the relative content of C18:1 was significantly increased at these two stages (Fig. 5b). Significant reduction in C18:2 at around the meiosis stage coincides with the developmental abnormalities observed at this stage of anther development in fad2–3. A reduction (49.18%) in C18:2 was detected in the mature anthers of fad2–3 compared to that of wild-type, but C18:1 increased almost five-fold in fad2–3 compared to wild-type, suggesting that the biosynthesis steps from C16:0 to C18:1 were not greatly affected by silencing of GhFAD2–3, resulting in accumulation of C18:1.
The ratio of C18:2/C18:1 was analyzed during anther development. The results showed that there was no significant difference between fad2–3 and WT during early anther development (Fig. 5c). At the meiosis stage, the ratio of C18:2/C18:1 was 1.49 in WT anthers. However, the corresponding ratio was only 0.48 in fad2–3 anthers. Compared with the WT anther, there was a more significant decrease in the ratio of C18:2/C18:1 in the fad2–3 anther at the pollen maturation stage (Fig. 5c).
Cuticular wax constituents and cutin monomer of cotton anthers
Interestingly, in wild-type anthers, wax constituents contain wax esters at high content (over 50%), prominently monopalmitin and monostearate, which are products of the acyl reduction pathway. This is different from waxy components in anthers of rice and maize [31, 39], perhaps because cotton is a woody perennial plant, and it seems to be similar to that in wax constituents of Jojoba . The content of alkanes was approximately 40%, mainly with chain lengths ranging from 23 to 35 carbons. Compared with the wild-type, there was no significant difference in the major components of wax in fad2–3. Although the content of fatty acid components was relatively low in wax constituents, fad2–3 had significantly lower levels of C18:2 and C18:3, but a relatively higher level of C18:1. The decrease in C16:0 and C18:0 was also observed in the anther wax of fad2–3. These results indicated that C18:2 and C18:3 were also important components of plant epidermis and downregulation of GhFAD2–3 affected the relative content of components in the cuticular layer.
Cutin monomers in the cuticular layer were methylated by methanolic HCl and reanalyzed by GC-MS. The major monomer was α, ω-octadecadiendioic acid (C18:2 DCA), which is a usual constituent of cutin. 9,10,18-trihydroxy octadecenoic acid (9,10,18-triOH C18:1 FA), hexadecane-1,16-dioic acid (C16:2 DCA) and 9,10-epoxy-18-OH-C18:1 (9,10 Epoxy 18-OH Acid) at relatively high levels had also been detected by GC-MS in wild-type anthers (Fig. 6b). Cutin monomers also include fatty acid components such as C18:2 and C18:3. The strongest effects of the fad2–3 anther on cutin monomers were apparent in the unsaturated C18:2 DCA and 9,10,18-triOH C18:1FA. Especially, the C18:2 DCA content decreased by 82.16% in fad2–3. Our results were also consistent with previous reports that the Arabidopsis fad2 mutant showed a decrease in double unsaturated C18 α, ω-diacids in leaf polyester . Nevertheless, C18:2 and C18:3 are important substrates for biosynthesis of many other lipids that are essential structural components of anthers and the pollen wall.
Functional specificity and redundancy of GhFAD2 genes
Each individual GhFAD2 gene is expected to function as a desaturase to convert C18:1 into C18:2 in different cotton organs and/or tissues, a process that is presumably determined by the expression specificity and level of each GhFAD2. Among the four pairs of GhFAD2 genes, GhFAD2–2 and GhFAD2–4 were expressed at very low levels in the tissues analyzed in this study, suggesting a limited role or a specific role in the tissues not analyzed in this study of these genes in cotton development. GhFAD2–1 seems to be specifically expressed in developing seeds, with the highest expression level detected in the 40 DPA seeds (Fig. 1), consistent with its role in the accumulation of C18:2 in seeds . In cotton seeds, the relative content of C18:2 could reach over 50% of the total fatty acid content . In contrast, GhFAD2–3 seems to be ubiquitously expressed in all tissues analyzed, but its expression level was remarkably lower than that of GhFAD2–1 in the 40 DPA seeds although its expression levels in the 5 DPA and 20 DPA seeds were significantly higher than or similar to that of GhFAD2–1, respectively (Fig. 1). The highest expression level of GhFAD2–3, particularly GhFAD2–3D, was observed in anthers, whereas the other three pairs of GhFAD2 genes were not or expressed at very low levels in anther. This expression pattern suggests that GhFAD2–3 is the major, if not the sole, gene responsible for the synthesis of C18:2 in anthers and is important for anther development. The male sterile phenotype observed in fad2–3 supported this conclusion. GhFAD2–3, particularly GhFAD2–3D, was also relatively highly expressed in stigma and leaves, but we did not observe obvious phenotypic changes in these two organs in fad2–3, probably due to the presence of a functional GhFAD2–2 that was not a target of the 517-bp fragment used in generation of fad2–3. This observation suggests that different GhFAD2 genes may be functionally redundant.
One interesting observation was the significantly biased expression levels of the two homoeologous GhFAD2–3 in anthers, stigma and leaves (Fig. 1). The significantly higher expression level of GhFAD2–3D would suggest it is the major functional gene. This speculation could not be tested by using fad2–3 generated in this study because both gene homeologs are targets of the RNAi construct but can be tested by gene homolog-specific knock-out using the gene editing approach.
A role for GhFAD2–3 in anther and pollen development
Fatty acid metabolism is an essential physiological process throughout the plant life cycle. In higher plants, acetyl-CoA carboxylase carboxylates acetyl-CoA to form malonyl-CoA, which is further converted by fatty-acid synthase (FAS) to long-chain fatty acids via the six recurring reactions, until the C16:0 is produced. When the 16:0 carbon fatty acids (FAs) are formed, it then undergoes some modifications leading to desaturation and/or elongation. The elongation begins with stearate (C18:0) and is mainly performed by several membrane-bound enzymes in the endoplasmic reticulum (ER). C18:0 was also further dehydrogenated by Δ9-stearyl-ACP desaturase (SAD) to form monounsaturated C18:1. After that, most PUFAs are synthesized by desaturases located in the ER, namely, FAD2 (C18:1 to C18:2 desaturation) and FAD3 (C18:2 to C18:3 desaturation). Most cuticular wax and cutin are derived from fatty acid precursors and play important roles in developmental events and physiological functions. Our results also showed that the major wax constituents in the cotton anthers were wax esters and alkanes. Importantly, it had been reported that fatty acid desaturases, including FAD2, are responsible for the biosynthesis of 30–35% of the cutin monomers from unsaturated C18 aliphatics . Similar to the observation of a decrease in the double unsaturated C18 diacids in leaf polyester of the Arabidopsis fad2 mutant , we showed that the contents of polyunsaturated C18 dioic acid were significantly reduced in the anther cutin of fad2–3, suggesting that maintaining a certain level of polyunsaturated C18 is important for proper development of the cuticular structure of cotton anthers.
In our model, GhFAD2 could play an important role in the biosynthesis of the cutin and suberin monomers. The fad2–3 is deficient in its ability to catalyze C18:1 to C18:2, which finally results in anther polyesters with less α, ω-octadecadiendioic acid (C18:2) and polyhydroxy-octadecenoic acid (9,10,18-triOH C18:1 FA). Our qRT-PCR results showed that the expression of GhFAD2 was relatively low at the sporogenous cell stage and microsporocyte stage, significantly increased at the tetrad stage, and peaked at the pollen maturation stage in the control plants (Fig. 7). In fad2–3, GhFAD2 had a very similar dynamic expression change at different anther developmental stages; however, its expression level was significantly lower than that of the wild-type at the time points investigated, particularly from the tetrad stage to the pollen maturation stage (Fig. 7). GhFAD3 also had a similar expression change in the wild-type and fad2–3 during anther development. Correspondingly, from the sporogenous cell stage to the microsporocyte stage, there was no difference in the relative content of C18:2 and C18:3 between the wild-type and fad2–3 anthers, but from the microsporocyte stage until the mature pollen stage, the relative content of C18:2 and C18:3 was significantly lower in fad2–3 than in wild-type (Fig. 5).
Overall, in fad2–3, the expression of GhFAD2 and GhFAD3 genes was significantly inhibited; however, the expression levels of the above determined related genes in the pathway, such as GhSAD, GhCYP86B1, GhCYP94C1, GhHTH and GhALDH, were apparently upregulated, especially at the meiosis stage and/or tetrad stage. Thus, it is interesting that the upregulation of the expression of these genes in fad2–3 was accompanied by the increase in C18:1 content relative to the control anther. Similar phenomena have also been previously observed with transgene expression of FAD2, which produces unusual FAs, including epoxidation, hydroxylation and double bonding conjugation [40, 41]. One hypothesis explaining this phenotype is that unusual FA products inhibit the activity of FAD2, thereby effectively preventing the conversion of C18:1 to C18:2 [43, 44, 45, 46]. In cottonseed, the relative content of C18:2 peaked in the mature seeds, accounting for over 50% of the total fatty acid content of seed . However, the accumulated unusual FA products, such as epoxy fatty acid, hydroxyl acid, and dioic acid, may act as inhibitors of GhFAD2 expression in anthers. Thus, this regulatory mechanism may account for part of why the relative proportion of C18:2 is not very high in anthers. In fad2–3, the expression of GhSAD was significantly upregulated at the meiosis stage and the tetrad stage, which would further lead to a much higher proportion of C18:1 content. Under this metabolic scenario, GhFAD2 would be further inhibited by the increased expression of related genes in the pathway, such as GhCYP86B1, GhALDH and GhCYP94C1, resulting in relatively low C18:2 content at the meiosis stage and the tetrad stage in fad2–3. In contrast, the expression of related genes in the pathway to synthesize waxy and cutin monomers could be induced by a low level of C18:2 content. This may be a self-protective mechanism of plant cells. Compared to wild type, GhHTH and GhCYP704B1 shows significantly higher expression level at the meiosis stage. However, this does not mean that it could result in an increase in the content of hydroxy and epoxy fatty acids. At the tetrad stage, the expression of GhHTH and GhCYP704B1 was very low in both wild type and fad2–3, and there was no significant difference between them. The anthers of fad2–3 were smaller than those of wild-type, and the fad2–3 pollen grains appeared to lack the exine layer. These results suggest that the lack of C18:2 and C18:3 have adverse effects on the establishment of functional anther cuticles and pollen exine in fad2–3, probably a result of insufficient biosynthesis and/or deposits of sporopollenin in these protective walls, as suggested by the observation that the fad2–3 pollen grains were unresistant to acetolysis treatment. The high C18:1 content in cottonseeds by specific inhibition of the expression of GhFAD2–1 had disadvantageous effects on seed vigor . As in cottonseed, it seemed to suggest that the too high accumulation of C18:1 was also unfavorable to the development of anthers.
In plants, the synthesis of various fatty acid components has a complex interrelationship, and the synthesis of other fatty acids will be affected when the synthesis of a particular fatty acid component is regulated. In general, with the increase in C18:1 content, the content of C16:1 decreased in this study. The decrease in C16:0 was also observed in the anther wax of fad2–3. Accordingly, the content of C16:0 DCA was reduced in the anther cutin of fad2–3. The relative proportion of fatty acid components accumulated in plant tissues is very important for ensuring that they could preserve the basic physiological activities. Changing the relative content and proportion of one or some fatty acid components in plant tissues may adversely affect vegetative or reproductive growth of plants. It had been reported that the Arabidopsis fad2 mutants showed dwarf phenotypes and the levels of PUFAs in phospholipids were relatively low. The limited membrane fluidity of the fad2 mutant resulted in its inability to survive at a low temperature . We had also previously found that the decreased C18:2 content in cottonseeds by specific inhibition of the expression of GhFAD2–1 had disadvantageous effects on seed vigor . C18:2 and C18:3 could serve as the essential structural components of anthers and pollen walls and/or as the substrates for biosynthesis of signaling molecules and/or hormones, such as Jasmonic acid (JA), essential for anther and pollen development. Jasmonic acid is a carbocyclic fatty acid and plays an essential role in pollen and anther development. Several Arabidopsis male sterile mutants have been shown to be caused by mutation in genes involved in JA biosynthesis and/or signaling pathways, such as coi1  and opr3/dde1 [49, 50]. It has been reported that the crucial requirement for C18:3 in A. thaliana pollen development and anther dehiscence seems to be as a substrate for JA biosynthesis through the octadecanoid pathway ; however, the threshold requirement for C18:3 is very low, and anthers with only 1–2% of C18:3 could remain fertile. The wild-type cotton anthers contain a relatively low level of C18:3 compared to C18:1 and C18:2 but could have a level higher than the low threshold reported in Arabidopsis. The C18:3 content was significantly reduced in the fad2–3 anthers and became as low as 0.3% (~ 10-fold reduction) at the mature pollen stage (Fig. 5). Taken together, C18:2 and C18:3 are essential structural components of anthers and are also important substrates for biosynthesis of many other lipids in anther development.
In this study, we show functional specificity and redundancy of GhFAD2 genes that would be helpful to understand gene expression and regulation in allopolyploid crops. We report for the first time on the dynamic changes in fatty acid constituents during anther development and showed that silencing of GhFAD2–3 reduced PUFAs (mainly C18:2 and C18:3) in anthers, severely hampered pollen development and significantly reduced the number and viability of pollen grains. Our results demonstrated that the relative proportions of fatty acid components accumulated in plant tissues is very important to ensure that they could preserve the basic physiological activities. In the future, a comprehensive and in-depth study of unsaturated fatty acids and their derivatives in plant cells would help us to better control plant growth and development on the basis of understanding the physiological significance of the formation of plant fatty acid components.
Characterization of the GhFAD2 gene family
The genome sequences and annotation files of G. hirsutum (AD1_NBI), G. raimondii (D5_JGI) and G. arboreum (A2_BGI) were downloaded from CottonGen (https://www.cottongen.org). To identify the members of the GhFAD2 family genes, the amino acid sequence of the previously reported GhFAD2 (GenBank accession no. X97016) was used to search for its homologs in the annotated proteins of the three genomes using BLASTP (E-value ≤10− 50). The hits with all three conserved histidine-clusters observed in all reported plant FAD2 sequences were considered as GhFAD2. The relationship of the identified GhFAD2 genes was investigated by sequence alignment and phylogenetic analysis.
We did two transcriptome analyses. One aimed to investigate the expression profiles of individual GhFAD2 in various tissues and the other compared the difference in anther transcriptomes between the GhFAD2–3 silenced transgenics and wild-type. For the first experiment, total RNA was isolated from root, leaf, stem, anther, stigma, ovary, developing and mature seed (at 5, 20, 40 and 60 days post anthesis) and fiber (at 12 and 24 days post anthesis). Each sample had three biological replicates. For each sample, a total of 3 μg RNA was used in preparing the RNA-seq library. Barcoded multiplexed RNA-seq libraries were created using the NEBNext® Ultra™ RNA Library Prep Kit for Illumina® (NEB, USA) according to the manufacturer’s protocol. Clean paired-end reads were aligned to the TM-1 reference genome , and the number of reads aligned to each gene was measured using HTSeq v0.6.1. The expression levels of individual genes were quantified using FPKM (fragments per kilobase of transcript per million mapped reads).
For the anther transcriptome experiment, total RNA was isolated from anthers of wild-type (control) and transgenic plants with silenced GhFAD2–3 (fad2–3). Anthers from two developmental stages, meiotic and tetrad, were investigated. Three samples were collected from each stage for both wild-type and fad2–3. A total amount of 3 μg RNA from each sample was used in generating index-coded RNA-seq libraries using the Illumina TruSeq RNA Library Prep Kit. The clustering of the index-coded libraries was performed using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina) on a cBot Cluster Generation System following the manufacturer’s manual. The libraries were then sequenced using the Illumina HiSeq 2500 platform. Read mapping and FPKM calculations were performed as described previously. The model based on negative binomial distribution was used to determine differentially expressed genes (DEG) with an adjusted p-value < 0.05 . The KOBAS software was used to test statistically the enrichment of DEGs in KEGG pathways.
Plant material and generation of RNAi plants
The cotton variety ‘Xinluzao 33’ provided by the Cotton Research Institute of Shihezi University was used in this study. A 517-bp fragment (Additional file 1: Figure S1) was amplified by PCR using PrimeSTAR™ HS DNA polymerase. The primers used were 5′-CACCCGCTCACTTATCCGTTCA-3′ (CACC was added at the 5′ end for directional cloning of the amplified fragment into the RNAi construct) and 5′-CGTTGTAGATAGGACCGTAT-3′. The PCR cycles were performed as follows: 95 °C for 5 min, followed by 29 cycles at 94 °C for 45 s, 55 °C for 45 s, 72 °C for 45 s, and a final extension at 72 °C for 10 min. The amplified DNA fragment was subcloned into pENTR/D-TOPO (Invitrogen) to generate the entry vector pENTR/D-GhFAD2 that was confirmed by sequencing. The LR recombination reaction between pENTR/D-GhFAD2 and the gateway vector pANDA35HK was then used to create pANDA35HK-dsGhFAD2 using the Gateway™ LR Clonase™ plus enzyme mix (Invitrogen). The native promoter of GhFAD2–3D isolated from cotton genomic DNA (Additional file 1: Figure S2a) was cloned into the Hin dIII-Bam HI site of pBI121 to generate pBIAP. Finally, the fragment with the 517-bp fragment inserted forwardly and reversely at two sides of the intron was excised from pANDA35HK-dsGhFAD2 and used to replace the GUS gene in pBIAP to create pBIAP-dsGhFAD2 (Additional file 1: Figure S2b), which was then electroporated into the Agrobacterium tumefaciens strain LBA4404. Cotton transformation was carried out using hypocotyl explants from G. hirsutum cultivar Xinluzao 33 as described by Jin et al. .
A fragment (1,113 bp) containing part of the 517-bp GhFAD2–3 segment and part of the gus linker was amplified by PCR to identify positive transgenic plants. Genomic DNA was isolated from cotton leaf tissues and used in PCR amplification using the primers, 5′-CTGTACAGCGAAGAGGCAGTC-3′ and 5′-CGTTGTAGATAGGACCGTAT-3′. The pair of primers 5′-GAGTCTGGTAATTGGAATGAG-3′ and 5′-TTCGCAGTTGTTCGTCTT-3′ was used to amplify the 18S rDNA gene as a control. Cultivar ‘Xinluzao 33’ and transgenic cotton plants were grown in soil in a growth room with a 16 L:8D photoperiod and 50% relative humidity.
Phenotypic analysis and pollen staining
Plants or flowers were photographed with a digital camera (Canon, Japan) or a SteREO Discovery microscope (Carl Zeiss). In all experiments, phenotypes of the RNAi plants (fad2–3) were analyzed together with wild-type. For testing pollen viability, pollen grains were stained with 1% I2-KI solution and photographed using a SteREO Discovery microscope (Carl Zeiss). For the control plants, sufficient mature pollens were obtained by placing 3–4 open flowers in a microfuge tube. For the fad2–3 plants, mature anthers were dissected from flowers and gently squashed in staining solution using dissecting needles. Acetolysis treatment was carried out according to Aarts et al. , and pollen grains were treated with a mixture of sulfuric acid and acetic anhydride at 100 °C. The pollen pellet was then transferred to a microscope slide and viewed with a SteREO Discovery microscope.
Scanning and transmission electron microscopy
For scanning electron microscopy, anthers at different developmental stages were fixed overnight in 2.5% glutaraldehyde, and then washed and postfixed in 1% osmium tetraoxide in 0.1 M sodium phosphate buffer (PBS, pH 7.2). Samples were then dehydrated in a graded ethanol concentration (30, 50, 70, 95 and 100%) and dried with liquid CO2. Before examination with the scanning electron microscope, dried anthers were sputtered with gold palladium for 300 s at 25 mA. Samples were finally visualized using a Hitachi S4500 microscope.
Transmission electron microscopy was performed using a Hitachi H7600 transmission electron microscope. Anthers were fixed in 2.5% glutaraldehyde (stored overnight at 4 °C), washed three times (5 min for each) with 0.1 M PBS, postfixed in 1% OsO4 for 2 h, and washed with PBS (three times, 5 min for each). Samples were then dehydrated as described above, treated with propylene oxide, and embedded in Spurr’s resin. Thin sections (70 nm) were taken using the Leica UC6 cryo ultramicrotome. Sliced sections were placed on 100-mesh copper grids and sequentially stained with uranyl acetate (30 min) and lead citrate (Sato’s Lead; 15 min).
Anther collection and fatty acid analysis
To analyze the composition of fatty acids in developing anthers, different sizes of flower buds (without bracts) with a diameter < 9 mm were collected from wild-type (fertile) and fad2–3 (sterile) and used in isolation of anthers with pollen grains at the following five developmental stages: sporogenous cells, microsporocyte, meiosis, tetrad and pollen maturation. Identification of these developmental stages was done using optical microscopy.
The whole anther (including both anther wall and pollen grains) of the five stages mentioned above was used in fatty acid analysis. The fatty acid methyl esters were prepared by alkaline transmethylation. Briefly, 0.5 g freeze-dried anthers were transferred into glass tubes, and used in oil extraction by using a Soxtherm apparatus (Gerhadt). Then, 5 ml 0.4 M KOH-methanol and 5 ml hexane were added and mixed. The solution was transferred into vials and shaken for 30 min at 40 °C. After adding ~ 1 g of anhydrous sodium sulfate to remove water, the upper hexane layer was used in GC-MS analysis. The quantitative standard curve was established by mixing 37 fatty acid methyl esters (Sigma) determined by GC-MS, and the external standard method was used for quantitative determination. The analyses were performed using GCMS-QP2020 at an electron ionization of 70 eV with an HP-88 capillary column (100 m × 0.2 mm) and film thickness of 0.2 μm. The column program used was: the injection temperature 250 °C, oven temperature kept at 40 °C for 2 min, then increased to 240 °C at a rate of 4 °C/min, and kept constant at 240 °C for 15 min. Operating conditions: helium carrier gas 2 mL/min, split ratio 10:1.
Analysis of anther cuticular waxes and cutin-like polyester
The wax of anther at mature pollen stage was analyzed using a published protocol  with some modifications. Briefly, 100 mg of freeze-dried anthers was submersed in 10 ml of chloroform containing 100 μg of tetracosane (Fluka; serving as an internal standard) for 1 min. The solution was transferred to a new vial, and then the solvent was evaporated under a nitrogen gas stream. The remaining compounds were incubated with 200 μl bis-N,N-(trimethylsilyl)-trifluoroacetamide (Sigma-Aldrich) in 200 μl pyridine for 60 min at 70 °C before GC-MS analysis. The constituent analyses were performed using GCMS-QP2020 with a DB-1 column of 30 m × 0.32 mm and film thickness of 0.1 μm. GC-MS analyses were performed as described by Jung et al. . Each compound was quantified against the internal standard by automatic integration of the peak areas.
The protocol for lipid polyester analysis was performed according to Li-Beisson et al. . First, 100 mg of freeze-dried anthers was delipidated. After that, depolymerization was performed by acid catalysis as described by Li-Beisson et al. . The resulting cutin monomer fraction was derivatized with BFTSA/pyridine (1:1) for 60 min at 70 °C, and then samples were analyzed using GCMS-QP2020 with a DB-1 column of 30 m × 0.32 mm and film thickness of 0.1 μm. The GC-MS was conducted according to Li-Beisson et al.  with helium carrier gas at 2 ml/min. Each compound was quantified on the basis of their total ion current as described by Li-Beisson et al. .
Quantitative real-time PCR (qRT-PCR) analysis
Total RNA was extracted from anthers at different developmental stages using RNAiso Plus (Takara), and then the RNA was reverse transcribed to obtain first-strand cDNA using a PrimeScript™ 1st Strand cDNA Synthesis Kit (Takara, China). The transcript levels of genes were analyzed by qRT-PCR using the LightCycler® 480 II (Roche, Germany). Each reaction was performed in 10 μl volumes using SYBR Green Master Mix (Takara, China) under the following PCR conditions: 94 °C for 3 min followed by 40 cycles of 94 °C for 15 s, 56 °C for 15 s, and 72 °C for 15 s. All gene specific primers for qRT-PCR were designed using the Primer 6.0 program (Additional file 1: Table S2). The cotton poly-ubiquitin gene (GhUBQ14, accession number in GenBank: DW505546) was used as an internal control. The primers for GhUBQ14 were 5′-CAACGCTCCATCTTGTCCTT-3′ and 5′-TGATCGT CTTTCCCGTAAGC-3′. All qRT–PCR reactions were performed in triplicate. The relative expression levels of target genes were calculated with the 2–ΔΔCt method .
JS, FL and QZ planned and designed the research. FL, LM, YL, FX, XZ, XN and YW performed experiments, conducted fieldwork, analysed data etc. FL, QZ and JS wrote the manuscript. JS and QZ contributed equally. All authors read and approved the final version of the manuscript.
This work was supported by National Key Research and Development Program of China [2016YFD0101900] and National Natural Science Foundation of China . The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
- 4.Jung JH, Kim H, Go YS, Lee SB, Hur CG, Kim HU, Suh MC. Identification of functional BrFAD2-1 gene encoding microsomal delta-12 fatty acid desaturase from Brassica rapa and development of Brassica napus containing high oleic acid contents. Plant Cell Rep. 2011; https://doi.org/10.1007/s00299-011-1095-x.PubMedGoogle Scholar
- 9.Dar AA, Choudhury AR, Kancharla PK, Arumugam N. The FAD2 gene in plants: occurrence, regulation, and role. Front Plant Sci. 2017; https://doi.org/10.3389/fpls.2017.01789.
- 10.Matteucci M, D’Angeli S, Errico S, Lamanna R, Perrotta G, Altamura MM. Cold affects the transcription of fatty acid desaturases and oil quality in the fruit of Olea europaea L. genotypes with different cold hardiness. J Exp Bot. 2011; https://doi.org/10.1093/jxb/err013.PubMedPubMedCentralGoogle Scholar
- 17.Piffanelli P, Ross JH, Murphy DJ. Intra- and extracellular lipid composition and associated gene expression patterns during pollen development in Brassica napus. Plant J. 1997; https://doi.org/10.1046/j.1365-313X.1997.11030549.x.PubMedGoogle Scholar
- 23.Chen W, Yu XH, Zhang K, Shi J, De Oliveira S, Schreiber L, Shanklin J, Zhang D. Male Sterile2 encodes a plastid-localized fatty acyl carrier protein reductase required for pollen exine development in Arabidopsis. Plant Physiol. 2011; https://doi.org/10.1104/pp.111.181693.PubMedPubMedCentralGoogle Scholar
- 24.Ariizumi T, Hatakeyama K, Hinata K, Inatsugi R, Nishida I, Sato S, Kato T, Tabata S, Toriyama K. Disruption of the novel plant protein NEF1 affects lipid accumulation in the plastids of the tapetum and exine formation of pollen, resulting in male sterility in Arabidopsis thaliana. Plant J. 2004; https://doi.org/10.1111/j.1365-313X.2004.02118.x.PubMedGoogle Scholar
- 25.Morant M, Jørgensen K, Schaller H, Pinot F, Møller BL, Werck-Reichhart D, Bak S. CYP703 is an ancient cytochrome P450 in land plants catalyzing in-chain hydroxylation of lauric acid to provide building blocks for sporopollenin synthesis in pollen. Plant Cell. 2007; https://doi.org/10.1105/tpc.106.045948.PubMedPubMedCentralGoogle Scholar
- 26.Li H, Pinot F, Sauveplane V, Werck-Reichhart D, Diehl P, Schreiber L, Franke R, Zhang P, Chen L, Gao YW, et al. Cytochrome P450 family member CYP704B2 catalyzes the omega-hydroxylation of fatty acids and is required for anther cutin biosynthesis and pollen exine formation in rice. Plant Cell. 2010; https://doi.org/10.1105/tpc.109.070326.PubMedPubMedCentralGoogle Scholar
- 27.de Azevedo SC, Kim SS, Koch S, Kienow L, Schneider K, McKim SM, Haughn GW, Kombrink E, Douglas CJ. A novel fatty acyl-CoA Synthetase is required for pollen development and sporopollenin biosynthesis in Arabidopsis. Plant Cell. 2009; https://doi.org/10.1105/tpc.108.062513.PubMedPubMedCentralGoogle Scholar
- 28.Ariizumi T, Hatakeyama K, Hinata K, Sato S, Kato T, Tabata S, Toriyama K. A novel male-sterile mutant of Arabidopsis thaliana, faceless pollen-1, produces pollen with a smooth surface and an acetolysis-sensitive exine. Plant Mol Biol. 2003; https://doi.org/10.1023/B:PLAN.0000009269.97773.70.PubMedGoogle Scholar
- 29.Jung KH, Han MJ, Lee DY, Lee YS, Schreiber L, Franke R, Faust A, Yephremov A, Saedler H, Kim YW, Hwang I, An G. Wax-deficient anther1 is involved in cuticle and wax production in rice anther walls and is required for pollen development. Plant Cell. 2006; https://doi.org/10.1105/tpc.106.042044.PubMedPubMedCentralGoogle Scholar
- 31.Shi J, Tan H, Yu XH, Liu Y, Liang W, Ranathunge K, Franke RB, Schreiber L, Wang Y, Kai G, et al. Defective pollen wall is required for anther and microspore development in rice and encodes a fatty acyl carrier protein reductase. Plant Cell. 2011; https://doi.org/10.1105/tpc.111.087528.PubMedPubMedCentralGoogle Scholar
- 32.Domergue F, Vishwanath SJ, Joubès J, Ono J, Lee JA, Bourdon M, Alhattab R, Lowe C, Pascal S, Lessire R, Rowland O. Three Arabidopsis fatty acyl-coenzyme a reductases, FAR1, FAR4, and FAR5, generate primary fatty alcohols associated with suberin deposition. Plant Physiol. 2010; https://doi.org/10.1104/pp.110.158238.PubMedPubMedCentralGoogle Scholar
- 33.Kim J, Jung JH, Lee SB, Go YS, Kim HJ, Cahoon R, Markham JE, Cahoon EB, Suh MC. Arabidopsis 3-ketoacyl-coenzyme a synthase9 is involved in the synthesis of tetracosanoic acids as precursors of cuticular waxes, suberins, sphingolipids, and phospholipids. Plant Physiol. 2013; https://doi.org/10.1104/pp.112.210450.PubMedPubMedCentralGoogle Scholar
- 41.Bonaventure G, Beisson F, Ohlrogge J, Pollard M. Analysis of the aliphatic monomer composition of polyesters associated with Arabidopsis epidermis: occurrence of octadeca-cis-6,cis-9-diene-1,18-dioate as the major component. Plant J. 2004; https://doi.org/10.1111/j.1365-313X.2004.02258.x.PubMedGoogle Scholar
- 52.Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. Transcript assembly and quantification by RNA-seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010; https://doi.org/10.1038/nbt.1621.PubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.