Acyl-CoA N-acyltransferase influences fertility by regulating lipid metabolism and jasmonic acid biogenesis in cotton

Cotton (Gossypium spp.) is an important economic crop and there is obvious heterosis in cotton, fertility has played an important role in this heterosis. However, the genes that exhibit critical roles in anther development and fertility are not well understood. Here, we report an acyl-CoA N-acyltransferase (EC2.3; GhACNAT) that plays a key role in anther development and fertility. Suppression of GhACNAT by virus-induced gene silencing in transgenic cotton (G. hirsutum L. cv. C312) resulted in indehiscent anthers that were full of pollen, diminished filaments and stamens, and plant sterility. We found GhACNAT was involved in lipid metabolism and jasmonic acid (JA) biosynthesis. The genes differentially expressed in GhACNAT-silenced plants and C312 were mainly involved in catalytic activity and transcription regulator activity in lipid metabolism. In GhACNAT-silenced plants, the expression levels of genes involved in lipid metabolism and jasmonic acid biosynthesis were significantly changed, the amount of JA in leaves and reproductive organs was significantly decreased compared with the amounts in C312. Treatments with exogenous methyl jasmonate rescued anther dehiscence and pollen release in GhACNAT-silenced plants and caused self-fertility. The GhACNAT gene may play an important role in controlling cotton fertility by regulating the pathways of lipid synthesis and JA biogenesis.

acyl-CoA N-acyltransferase (EC2.3) gene was significantly down-regulated with unknown function and designated GhACNAT. GhACNAT is homologous to genes in different species, including with At2g23390 (Fig. 1a), which is a member of the acyl-CoA N-acyltransferase superfamily in A. thaliana with a conserved domain that catalyzes N-acetyltransferase reactions in lipid synthesis and metabolism (Fig. 1b). We amplified the full-length 1410 bp cDNA of GhACNAT from C312. One highly homologous gene was found in the DD genome of G. raimondii.
The expression pattern of the GhACNAT gene analyzed by qRT-PCR. The GhACNAT gene was expressed at different levels in the tested tissues and organs, including roots, stems, leaves, bracts, petals, stamens stigmas, and fibers of different development stages (0-, 5-, 10-, 15-, 20-, and 25-DPA (days post anthesis)) in C312 (Fig. 2). The expression levels of the GhACNAT gene in the reproductive organs (bracts, petals, stigmas and stamens) were significantly higher than in the vegetative organs (roots, stems, and leaves) and fibers of different developmental stages. The GhACNAT gene maintained significantly high-level expression in the flower organs, including the sepals, petals, stamens and stigmas.
The construction of the pCLCrV-GhACNAT vector for VIGS. The pCLCrV-based virus-induced gene-silencing system produced an efficient and persistent gene silencing in cotton. To determine whether the pCLCrV-based silencing vector could silence endogenous gene expression in cotton, a 327 bp fragment of the phytoene desaturase gene (PDS), commonly used as a positive marker for VIGS, was cloned into the pCLCrVA vector to construct pCLCrV-PDS. Agrobacteria tumefaciens strain GV3101 containing different pCLCrVA and pCLCrVB vector was infiltrated by syringe into the cotyledons of 2-week-old seedlings of C312. By 20 days postinfiltration, cotton plants infiltrated with Agrobacteria harboring both pCLCrVA-PDS and pCLCrVB showed the typical PDS gene-silencing phenotype of photobleaching in newly emerging leaves.
The photobleaching phenotype in plants infiltrated with pCLCrV-PDS was observed on the newly emerging leaves by 3 weeks postinfiltration, first appearing in the nearby veins and gradually extending to the whole leaves, which became mosaic with green spots (Fig. S1). The efficiency of the photobleaching in the plants infiltrated with pCLCrV-PDS and pCLCrVB was greater than 75%. The transgenic pCLCrV-PDS plants were confirmed by PCR and showed the photobleaching mosaic with green spots in the leaves. Transgenic plants of the pCLCrVA-empty with pCLCrVB and pCLCrVA-GhACNAT with pCLCrVB showed a slight shrinking of the new leaves (Fig. S1 a, b, c). The transgenic plants with pCLCrVA-empty and pCLCrVA-GhACNAT were identified by PCR with specific primers (Table S3). The photobleaching in the transgenic pCLCrV-PDS plants appeared in different organs, including the stem, petiole, bracts, and bolls during the growth cycle (Fig. S1), indicating that the silencing caused by this vector persisted in the cotton plants. Pink stripes appeared on the stems in the transgenic pCLCrV-PDS plants (Fig. S1 e). To investigate the silencing efficiency, the levels of PDS mRNA in both pCLCrV-PDS-silenced and wild-type C312 plants were measured by quantitative real time PCR (qRT-PCR). In the transgenic pCLCrV-PDS plants with strong photobleaching of the new leaves, the expression level of the endogenous PDS gene was significantly reduced, 2-40-fold compared with wild-type C312 plants (Fig S1i). The control plants of C312 and pCLCrV-empty did not show the photobleaching phenotype at any time during the studies ( Fig S1c).
A 200 bp fragment of the GhACNAT gene was amplified to construct the pCLCrV-GhACNAT vector to silence the endogenous GhACNAT gene in C312 plants.
Silencing the GhACNAT gene influenced the development of the reproductive organs of transgenic plants. There were no obvious differences in the phenotypes of the control plants, the transgenic pCLCrV-PDS plants and the transgenic GhACNAT-silenced plants during vegetative growth except for the mosaic photobleaching spots in the transgenic pCLCrV-PDS plants. In the reproductive stage of the transgenic GhACNAT-silenced plants, the distinct phenotypic changes appeared in the reproductive organs (Fig. 3). The stigmas were forked and warped compared with the stigmas of C312 (Fig. 3a) and the transgenic control plants. The stamens of the transgenic GhACNAT-silenced plants appeared indehiscent but full of viable pollen (Fig. 3b,c,d and 4a,d,g). The filaments became shorter, and the number of stamens decreased significantly (Table S1). No pollen was released from the closed anthers in the transgenic GhACNAT-silenced plants. In the male sterile ms1 mutant, the anthers were closed and shrunken, and no pollen formed (Fig. 4a,f,h). The anthers of the C312 plants were dehiscent and full of active pollen grains, whereas the stamens of the transgenic GhACNAT-silenced plants were indehiscent but as full of viable pollen as the wild-type C312 (Fig. 4d,g). However, the pollen could not be released from the closed stamens. In the transgenic GhACNAT-silenced plants, mature bolls were obtained successfully by pollination with normal C312 pollen. The bolls had two to four loculi containing seeds covered with fibers ( Fig. 3e-i).
The analysis of genes differentially expressed in transgenic GhACNAT-silenced plants and control plants. To study differentially expressed genes related to the alteration of reproductive organs in the transgenic GhACNAT-silenced plants, 3 individual plants of wild type C312 and the transgenic pCLCrV-GhACNAT-silenced lines were selected for differential expression analysis by cDNA-AFLP. The stigmas and stamens of the flowers (on the day of anthesis) were picked and collected for total RNA isolation. The differentially expressed genes were selected, and 450 fragments (100-200 bp) were cloned and sequenced. The function annotations of the differentially expressed genes in the reproductive organs of the control plants and the transgenic GhACNAT-silenced plants were divided into several categories (Fig. 5). The main differentially expressed genes were focused on catalytic activity, translation regulator activity and metabolic processes involved in lipid metabolism. Acyl-CoA and acyl-CoA N-acyltransferase were involved in the lipid metabolic pathways.
The expression changes of genes related to the development of reproductive organs in the transgenic GhACNAT-silenced plants. Many mutants with anther indehiscence, including those with mutations in opr3 22 and dad1 23 , showed that jasmonic acid participates in the regulation of anther development, filament elongation and pollen viability 32 . AtAIF was identified as a repressor that controls anther dehiscence by regulating the jasmonic acid biosynthesis pathway in A. thaliana 24 . Linolenic acid (18:3) was released from membrane phospholipid by a lipolytic enzyme (DAD1), and the DAD1 gene participated in the regulation of anther development 23 .
JA plays a role in anther development. The expression levels of genes encoding key enzymes involved in the JA biosynthesis pathway were examined in transgenic GhACNAT-silenced plants and C312 plants. The genes (GhAIF, GhDAD1, GhLOX1 and GhOPR3) with high homology to AtAIF, AtDAD1, AtLOX1, and AtOPR3 in A. thaliana, which are genes related to reproductive organ development, were detected by qRT-PCR in stigmas and stamens of the C312 plants and the transgenic GhACNAT-silenced plants on the blossom day (Fig. 6). GhAIF was significantly up-regulated (29-fold) in the transgenic GhACNAT-silenced plants compared with the control plants. The GhLOX1 gene was up-regulated 3.5-fold compared with the gene in control plants. The corresponding enzyme catalyzes the conversion of linolenic acid to 13-hydroperoxylinoinic acid, which is the precursor for JA, and cis-3 hexenal+ traumatin using two different and competing pathways. The expression of GhAOC1 was not significantly different in C312 and transgenic GhACNAT-silenced plants. The expression of the GhDAD1 gene was significantly reduced in GhACNAT-silenced plants, and the expression of the GhOPR3 gene for JA biosynthesis was reduced approximately 4.5-fold in transgenic GhACNAT-silenced plants. This showed that JA biogenesis was influenced by changes in expression of these genes, which could cause anther indehiscence and short filaments. The GhSERK gene, which was shown to regulate anther dehiscence and pollen viability in cotton 30 , was identified from a differentially expressed fragment by cDNA-AFLP. The expression of GhSERK was significantly reduced in the stamens of the transgenic GhACNAT-silenced plants.
The GhACNAT silencing influenced fatty acid biosynthesis in the transgenic plants. GhACNAT belongs to the gene superfamily encoding acyl-CoA N-tranferase and has high homology with the At2g23390 gene. The acyl-CoA N-tranferase superfamily primarily participates in lipid metabolism to provide acyl-CoA. To investigate the function of GhACNAT in fatty acid biosynthesis, several key genes in the fatty acid biosynthetic pathway were measured by qRT-PCR. The single transcription factor WRI1, which has been shown to directly control the transcriptional activation of the fatty acid biosynthetic  The asterisks indicate statistically significant differences between the transgenic GhACNAT-silenced and WT C312 plants (Significant *P < 0.05, highly significant **P < 0.01, Student's t-test). pathway 33 , was significantly decreased in the transgenic GhACNAT-silenced plants. The glycolytic and late fatty acid biosynthetic genes, including GhPDH-E1a (homologous to At1g01090), GhBCCP2 (homologous to At5g15530), GhMAT (homologous to At2g30200), GhKASI (homologous to At5g46290), GhENR (homologous to At2g05990), and GhFATA (homologous to At3g25110), were measured in the stigmas and stamens of the control plants and the transgenic GhACNAT-silenced plants (Fig. 7a). In the transgenic GhACNAT-silenced plants, the expression levels of GhPDH-E1a, GhBCCP2, and GhMAT were reduced 2-fold compared with the control plants, and the expression levels of GhKASI and GhENR were significantly down-regulated. The relative mRNA levels of GhCYP86A2 (homologous to At4g00360), GhCYP86A8 (homologous to At2g45970), and GhCYP77A6 (homologous to At3g10570), which are involved in cuticle formation, were almost the same in the transgenic GhACNAT-silenced plants and the control plants. In the GhACNAT-silenced plants, the expression of the genes involved in lipid synthesis   was reduced significantly, and the tendency was consistent. The changes in gene expression influenced lipid synthesis and reduced JA biosynthesis in the stigmas and stamens of the GhACNAT-silenced plants.
The AtGPAT1 gene is associated with several fatty acid composition changes in flower tissues and seeds, including JA biosynthesis, catalyzed by lipoxygenase (LOX), allene oxide synthase (AOS), allene oxide cyclase (AOC), and 12-oxo-phytodienoic acid reductase (OPR). The genes GhGPAT1, GhGPAT3 and GhGPAT4 were detected by qRT-PCR and showed obvious differences in expression in C312 and the transgenic GhACNAT-silenced plants. The expression levels of GhGPAT1 and GhGPAT3 were significantly decreased, whereas GhGPAT4 expression increased in the transgenic GhACNAT-silenced plants. GhGPAT, GhLPAAT and GhFAD3 participate in the synthesis of phospholipids. In the transgenic GhACNAT-silenced plants, the expression of GhLPAAT did not change, but GhFAD3 expression decreased (Fig. 7b). In the JA biosynthesis pathway from the acyl-CoA pool to phospholipid to linolenic acid, the expression of genes encoding key enzymes significantly changed to influence JA biosynthesis.

Measurement of JA and rescue of transgenic
GhACNAT-silenced plants treated with exogenous MeJA. Endogenous GhACNAT expression was significantly decreased by VIGS in the reproductive organs of GhACNAT-silenced plants (Fig. 8a), and it significantly influenced the expression of genes related to JA biosynthesis (Fig. 6) and caused abnormal formation of reproductive organs. In transgenic GhACNAT-silenced plants, the genes for repressing JA biosynthesis were up-regulated, whereas the expression levels of genes involved in JA biosynthesis were down-regulated and decreased the JA levels in leaves and reproductive organs (Fig. 8b,c). The amounts of JA in fresh leaves of three transgenic GhACNAT-silenced lines were significantly lower than in C312, and in the reproductive organs, the amounts of JA were all higher than those in leaves. In the flowers of three GhACNAT-silenced lines, the amounts of JA were extremely low or even undetectable by GC/MS ( Fig S3) and significantly lower than in C312. However, the transcripts of genes for JA biosynthesis were apparently attenuated in transgenic GhACNAT-silenced lines compared with C312 plants, and the genes for repressing JA biosynthesis were up-regulated (Fig. 6). Consistent with the kinetics of gene expression in the JA pathway, JA levels in the leaves and reproductive organs in transgenic GhACNAT-silenced plants were dramatically decreased compared with C312 plants (Fig. 8b,c).
For the rescue experiment in transgenic GhACNAT-silenced plants to determine whether the defect in jasmonic acid synthesis caused the anther dehiscence, three concentrations (100 μ M, 500 μ M and 1000 μ M) of MeJA were used to treat the buds of transgenic GhACNAT-silenced plants before they opened. In the opening flowers of transgenic GhACNAT-silenced plants treated with MeJA, the morphologic changes in the reproductive organs were reversed, and the anthers were thoroughly or partially dehiscent, similar to the normal anthers in C312 plants (Fig. 9, Table S2). With 100 μ M MeJA, the anthers partially dehisced, and some anthers were still occlusive ( Fig. 9b; Table S2); with 500 μ M and 1000 μ M MeJA, all the anthers dehisced normally (Fig. 9c,d; Table S2). The bolls of transgenic GhACNAT-silenced plants were obtained by self-pollination after treatment with MeJA. The results showed that exogenous MeJA could reverse the anther phenotype and rescue the occlusive anthers caused by decreasing endogenous JA.

Discussion
VIGS provided an efficient method to analyze gene function in cotton during the entire growth cycle. VIGS has demonstrated to be an efficient method to analyze gene function in cotton 10,11,15,34 . By 3 weeks after the pCLCrV-Agrobacteria injection, transgenic pCLCrV-PDS plants showed obvious photobleaching in the new leaves. The efficiency of transformation reached 75%, and the effects were retained in leaves, stems, flowers, and even in bolls throughout the growth cycle. In the vegetative growth stage, there was no obvious phenotype except for slight wrinkling in the leaves of the transgenic GhACNAT-silenced plants. At the stage of reproductive growth, the phenotypes of the reproductive organs changed significantly, showing stigma wraped, anther indehiscence and no pollen release, which led to sterility in the transgenic GhACNAT-silenced plants. The sterility was mainly caused by the anther dehiscence and failure to release pollen. The changes of phenotypes in transgenic GhACNAT-silenced plants demonstrated that the effect of VIGS in cotton can be used for gene function analysis in vegetative growth, in reproductive growth and even in developing fibers.
The anther development was influenced by JA biosynthesis regulated by the GhACNAT gene. The GhACNAT gene belongs to the acyl-CoA N-acyltransferase gene superfamily and was selected from the genes differentially expressed in C312 and its male sterile mutant, ms1. The GhACNAT gene is homologous with At2g23390 in A. thaliana, which encodes acyl-CoA N-acyltransferase, an enzyme involved in the isopentenyl diphosphate biosynthetic pathway, maltose metabolism, regulation of proton transport, and starch biosynthesis and was found to be expressed in vegetative organs and in reproductive organs [35][36][37][38] .
Acyl-lipid metabolism in A. thaliana is a complex process and is organized into 12 sections based on the types of lipids produced and their subcellular localization 39 . Acyl lipids in plants have many diverse functions, involving more than 10 membrane lipid classes in the endoplasmic reticulum (ER), including the phospholipids, glycolipids, and sphingolipids [39][40][41][42][43] . The first committed step in fatty acid synthesis is the formation of malonyl-CoA from acetyl-CoA and bicarbonate by acetyl-CoA carboxylase (ACC) 16,44 . Before entering the fatty acid synthesis pathway, the malonyl-CoA group is transferred from CoA to ACP catalyzed by malonyl-CoA: acyl carrier protein malonyltransferase (MCMT). Malonyl-ACP provides two-carbon units at each step of elongation. The long-chain acyl groups hydrolyzed by acyl-ACP thioesterases release fatty acids. These fatty acids are ultimately activated to CoA esters by a long-chain acyl-CoA synthetase (LACS) and exported to the endoplasmic reticulum or possibly the PC at the plastid envelope by the action of lysophosphatidylcholine acyltransferase (LPCAT) [45][46][47] . Fatty acids are derived from plastid lipids desaturated in the ER and are precursors for potent signaling molecules, including jasmonates, leaf aldehydes or divinyl ethers 48,49 . The ER is the major site for phospholipid biosynthesis. Phosphatidic acid (PA) is the common precursor of phospholipids catalyzed by acyl-CoA: glycerol-3-phosphate acyltransferase (GPAT) and acyl-CoA: lysophosphatidic acid acyltransferase (LPAAT). PAs originating from the ER pathway exclusively contain C18 fatty acids in the sn-2 position in eukaryotic species, whereas PAs synthesized in plastids exclusively contain C16 fatty acids in the sn-2 position in prokaryotic species. The GPAT and LPAAT reactions of phospholipid synthesis or triacylglycerol synthesis utilize a mixed pool of acyl-CoA substrates that in many tissues is produced mostly from a PC acyl editing cycle. An AtGPAT1 deficiency had pleiotropic effects on glycerolipid biosynthesis and resulted in several fatty acid composition changes in flower tissues and seeds. The molar proportions of palmitic acid (16:0), stearic acid (18:0), linolenic acid (18:3), and the very-long-chain eicosenoic acid (20:1) decreased to varying degrees, whereas the proportions of oleic acid (18:1) and linoleic acid (18:2) increased 31 . Linolenic acid (18:3)-derived oxylipins are the biosynthetic C18 precursors of JA 17 . Acyl groups esterified to PC are the sites of extraplastidic FA desaturation 50 . The FAD2 51 and FAD3 52 enzymes convert PC-bound oleate to linoleate and then linolenate, respectively, which is the precursor of JA (Fig. 10).
The DAD1 gene encodes a lipase that hydrolyzes both TAGs (triacylglycerols) and glycerolipids 23 and releases linolenic acid for the biosynthesis of jasmonic acid. The precursor of JA originates from an 18-carbon fatty acid of linolenic acid (LA) released by the phospholipid in membranes, which is regulated by DAD1. The biosynthesis of JA is catalyzed by the key enzymes LOX, AOS, AOC and OPR3 in A. thaliana. The key genes of TAG synthesis were significantly down-regulated in the stigmas and stamens of transgenic GhACNAT-silenced plants. The gene expression of GhLOX and GhOPR3 displayed down-regulation in transgenic GhACNAT-silenced plants compared with the control plants. The GhAIF gene was identified as an inhibiting factor for JA biosynthesis 24 , and the expression level of GhAIF was significantly up-regulated in transgenic GhACNAT-silence plants.
The JA pathway was influenced because GhACNAT-silencing changed the expression levels of some key genes related to lipid and JA biological processes. The change in JA biosynthesis in transgenic GhACNAT-silenced plants resulted in the amount of JA decreasing, and the reduced JA caused anther indehiscence and other reproductive organ abnormalities. The expression changes of GhACNAT silenced by VIGS influenced many genes involved in biological processes of lipid and jasmonate acid. The  Gene cloning and construction of VIGS vectors. The candidate gene was obtained from the differentially expressed genes in C312 and its male sterility mutant (ms1). The mutant ms1 had the same background as C312 except for the mutation. The full-length 1410 bp cDNA of the GhACNAT gene was amplified from the cDNA library by PCR using the forward primer 5'-CTATGGCGGCAGCACTCA-3' and the reverse primer 5'-TGCACGATGGCTTTTCAA-3' . The PCR was incubated at 95 °C for 5 min and for 32 cycles at 95 °C for 30 s, 55 °C for 30 s, 72 °C for 100 s, and 72 °C for 10 min. The reaction was then kept at 4 °C. The PCR product was cloned into the pGEM-T Easy Vector (Promega, Madison, WI, USA), and the product was transformed into E.coli DH5α and sequenced.
The new CLCrV-based vector was modified from the CLCrV DNA-A and DNA-B components individually, which were inserted into the pCambia1300 vector to generate pCLCrVA and pCLCrVB, respectively 15 . The pCLCrVB is a geminivirus member and essential for the silence system in cooperation with pCLCrVA. The fragments of candidate genes were inserted into pCLCrVA to produce pCLCrVA-GhACNAT for VIGS in cotton plants as described by Gu et al., (2014) 15 .
A 200 bp fragment of the candidate gene GhACNAT was amplified by PCR (95 °C for 5 min and 30 cycles at 95 °C for 30 s, 59 °C for 30 s, 72 °C for 40 s, and 72 °C for 10 min. The reaction was then kept at 4 °C). The phytoene desaturase gene (PDS) causes loss of chlorophyll and carotenoids 53,54 and was used as a positive marker to visualize the timing and extent of endogenous gene silencing. A 327 bp fragment of the PDS gene isolated from C312 by PCR (95 °C for 5 min, 30 cycles at 95 °C for 30 s, 59 °C for 30 s, 72 °C for 40 s, and 72 °C for 10 min, then 4 °C). The fragments were inserted into the pGEM-T Easy Vector (Promega, Madison, WI, USA), and the construct was transformed into E.coli DH5α and sequenced. The recombinant plasmids were digested with SpeI and AscI (New England Biolabs, Beijing, China), and the products were individually inserted into the pCLCrVA vector to generate pCLCrV-GhACNAT and pCLCrV-GhPDS.
The four vectors with or without foreign genes were transformed individually into Agrobacterium tumefaciens strain GV3101 by electroporation using a Gene Pulser Apparatus (Bio-Rad, Hercules, CA, USA) as described by Huang et al., (2009) 55 . Three combinations, pCLCrVA-empty and pCLCrVB (for a negative control), pCLCrV-PDS and pCLCrVB (for a positive control), pCLCrV-GhACNAT and pCLCrVB (for target gene silencing) were used. Plants were transformed with pCLCrV-PDS and the pCLCrVA-empty vectors as controls. The gene sequences and primers for cloning and detection are listed in Table S1 of the supplementary file.
Plant growth and agroinfiltration. Cotton seedlings were grown in a growth chamber at 28 °C with a 14 h light and 10 h dark cycle. Healthy 2 week old seedlings were infiltrated with different Agrobacteria carrying pCLCrVA or one of its derivatives and pCLCrVB. Agrobacteria harboring pCLCrVA or one of its derivatives was mixed with an equal volume of Agrobacteria harboring pCLCrVB. The mixed Agrobacteria solutions were infiltrated into the abaxial side of the cotyledons of the 2-week-old cotton seedlings using syringes without needles. To facilitate infiltration, a few holes were gently punched into the underside of the cotyledons using a syring needle. The Agrobacteria preparation and the process of agroinfiltration used has been described by Gu et al., (2014) 15 . The growth chambers for the infected plants were kept at 23 °C with a 14/10 h light/dark cycle. The agroinfiltration was repeated at least three times with approximately 30 plants for each vector in the VIGS system.
Plants infiltrated with the pCLCrVA-empty vector and the wild-type C312 were used as controls and did not show any photobleaching during the experiment. Plants infiltrated with the pCLCrV-PDS vector showed the typical photobleaching phenotype in newly developed leaves. However, the efficiency of gene silencing was judged by the intensity of photobleaching, which varied in the different organs of the transgenic plants.
Anther observation and pollen staining. In our experiments, the transgenic GhACNAT-silenced plants and the control plants were compared to observe the change in phenotypes. Pollen activity was detected by staining with 13 mM of I 2 -KI. The solution was prepared by adding 12 mM KI to 5 ml ddH 2 O and then dissolving 4 mM of I 2 in the solution and adding ddH 2 O to 300 ml. The solution was stored in a brown glass bottle. Pollen grains were soaked in the I 2 -KI solution on glass slides for approximately 1 min, and then the shape and color of the pollen grains were observed with a fluorescence stereomicroscope (ZEISS SteREO LuMar.V12, Germany).

DNA extraction and PCR detection.
Total DNA was extracted from the leaves of pCLCrV-inoculated cotton plants using an extraction buffer containing 100 mM Tris-HCl, pH 8.0, 25 mM Na 2 EDTA, 2 M NaCl, 2% (w/v) CTAB, 2% (w/v) PVP and 2% (w/v) β -mercaptoethanol as described 56 . The presence of pCLCrV DNA in infected tissue was detected by PCR using primers specific for either pCLCrV DNA-A (CLCrVA F and CLCrVA R) or pCLCrV DNA-B (CLCrVB F and CLCrVB R). (Primers used in the experiment are listed in Table S3.) PCR was performed at 94 °C for 4 min, followed by 30 cycles of 94 °C for 30 s, 59 °C for 30 s and 72 °C for 1 min. The PCR products were visualized by electrophoresis on 1% (w/v) agarose gels.
The analysis by cDNA-AFLP of differentially expressed genes influenced by GhACNAT-silencing. The transgenic GhACNAT-silenced plants with significantly changed phenotypes were confirmed by PCR and expression analysis of the endogenous GhACNAT gene and then were selected for further analysis. The stigmas and stamens of C312 plants and transgenic GhACNAT-silenced plants were collected for RNA extraction using TRIzol ® Plus RNA Purification Kits (Invitrogen, Carlsbad, CA, USA). Double-stranded cDNA was synthesized from 200 ng RNA using an iScriptTM cDNA Synthesis Kit (Quanta BioSciences, Suite 1 Gaithersburg, MD, USA) according to a standard double-stranded cDNA synthesis protocol. The analysis of differentially expressed genes by cDNA-AFLP was performed to screen for the genes affected by GhACNAT-silencing. The cDNA-AFLP analysis was carried out as described by Hawkins et al., (2005) 57 . The cDNA was digested with the restriction enzymes MseI and EcoR I (New England Biolabs, Beijing, China). The digested products were then ligated to adapters with the following sequences. The MseI adapters were 5'-GACGATGAGTCCTGAG-3' and 3'-TACTCAGGACTCAT-5' . The EcoR I adapters were 5'-CTCGTATACTGCGTACC-3' and 3'-AATTGGTACGCAGTA-5' . Adapter-ligated cDNA was the template for pre-amplification, with PCR parameters of 30 cycles of 30 s at 94 °C, 60 s at 56 °C, and 60 s at 72 °C. The diluted (30-fold) amplified products were used as the templates for selective amplification. The differentially expressed bands between 100 and 200 bp were cloned into the pGEM-T Easy Vector (Promega, Madison, MI, USA), transformed into E.coli DH5α and sequenced.
The categorization of the transcript fragments on the basis of their presence or absence (qualitative variation) or differences in the amount (quantitative variation) in the control lines (wild-type C312 and transgenic pCLCrVA-empty plants) and transgenic GhACNAT-silenced lines showed that most of the variation was quantitative. The transcript fragments represented genes controlling different biological processes. The expression levels of GhACNAT and differentially expressed genes in the same biological pathway were measured in the control plants and the transgenic GhACNAT-silenced plants.

RNA extraction and quantitative real-time PCR (qRT-PCR) analysis.
Total RNA was extracted from fresh leaves or other tissues and organs (roots, stems, bracts, petals, stamens and stigmas, and fibers of different development stages) using a TRIzol ® Plus RNA Purification Kit according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA) and treated extensively with RNase-free DNase I. Double-stranded cDNA was synthesized from 100 ng RNA using an iScriptTM cDNA Synthesis Kit (Quanta BioSciences, Suite 1 Gaithersburg, MD, USA) according to a standard double-stranded cDNA synthesis protocol and then diluted 100-fold for qRT-PCR. The qRT-PCR included 5 μ l cDNA template (diluted 1/100), 5 μ l primers (2.4 μ M), and 10 μ l SYBR green mixture (GoTaq qPCR Master Mix, Promega, Madison, WI, USA) and was performed in an Eppendorf real-time PCR instrument (Mastercycler ep realplex, Hamburg, Germany). The incubation conditions were 95 °C for 1 min, followed by 40 cycles at 95 °C for 15 s, 60 °C for 30 s, 95 °C for 15 s and 60 °C for 15 s. The specificity of the amplified PCR product was determined by melting curve analysis. The primers for target genes were designed using Premier5 software (Premier Biosoft, Palo Alto, CA, USA) and are listed in Table S3 of the supplementary file.
The cotton Ubiquitin7 gene (GhUBQ7, Gen Bank accession number DQ116441) was used as an internal control for the assays. The expression levels of endogenous genes in cotton were obtained and standardized to the constitutive GhUBQ7 gene expression level. In each study, three independent experiments were conducted with three independent samples with the same phenotype per experiment.

JA Measurement. The new leaves and the whole opening flowers of C312 and transgenic
GhACNAT-silenced plants were collected and immediately frozen in liquid nitrogen, then stored at −80 °C. Based on the methods of JA extraction 58-60 , samples (approximately 150 mg) were ground and extracted with 1 ml pure cold methanol and stirred overnight at 4 °C in the dark. The suspension was centrifuged at 13.8 g for 15 min at 4 °C, and 900 μ l of the supernatant was carefully transferred to a new 1.5-ml tube. The pellet was mixed with 1 ml pure cold methanol, shaken for 4 h at 4 °C, and centrifuged. The two supernatants were combined and dried by a flow of nitrogen gas at room temperature. The residue was dissolved in 200 μ l pure methanol. Before GC-MS, derivatization was necessary for JA. The formation of MeJA using trimethylsilyldiazomethane was performed according to the methods in Schmelz et al., (2004) 58 . The GC-MS settings have been described 58,61-63 and were used with little modification.
An Agilent 7890 A gas chromatograph (Santa Clara, CA, USA) equipped with a split/splitless injector (splitless mode, 250 °C, injection volume 1 μ l) was interfaced to an Agilent 5973C mass spectrometer (Santa Clara, CA, USA) operated in chemical ionization (CI) mode with isobutene as the ionization gas. Compounds were separated on an HP-5 MS (30 m × 0.25 mm × 0.25 μ m) column held at 50 °C for 1 min after injection, and then the temperature was programmed at 15 °C (1 min) to 250 °C (10 min), with helium as the carrier gas (1.2 ml /min) (Fig S3).
Scientific RepoRts | 5:11790 | DOi: 10.1038/srep11790 The transgenic GhACNAT-silenced plant treated by exogenous MeJA. The anther indehiscence phenotype of the GhACNAT-silenced plants was rescued by exogenous MeJA. The buds of 3 to 5 days-before-opening in transgenic GhACNAT-silenced plants were sprayed with methyl jasmonate (MeJA solution). Different gradient concentrations with 100 μ M, 500 μ M, 1000 μ M dissolved in 0.05% (v/v) Tween-20 were sprayed on buds and leaves according to the description [22][23][24] . On the opening day of the flower, the productive organs were investigated and analyzed.