Gibberellin Overproduction Promotes Sucrose Synthase Expression and Secondary Cell Wall Deposition in Cotton Fibers

Bioactive gibberellins (GAs) comprise an important class of natural plant growth regulators and play essential roles in cotton fiber development. To date, the molecular base of GAs' functions in fiber development is largely unclear. To address this question, the endogenous bioactive GA levels in cotton developing fibers were elevated by specifically up-regulating GA 20-oxidase and suppressing GA 2-oxidase via transgenic methods. Higher GA levels in transgenic cotton fibers significantly increased micronaire values, 1000-fiber weight, cell wall thickness and cellulose contents of mature fibers. Quantitative RT-PCR and biochemical analysis revealed that the transcription of sucrose synthase gene GhSusA1 and sucrose synthase activities were significantly enhanced in GA overproducing transgenic fibers, compared to the wild-type cotton. In addition, exogenous application of bioactive GA could promote GhSusA1 expression in cultured fibers, as well as in cotton hypocotyls. Our results suggested that bioactive GAs promoted secondary cell wall deposition in cotton fibers by enhancing sucrose synthase expression.


Introduction
Cotton is the leading natural fiber for textile industry worldwide. Biologically, cotton fibers are extremely elongated single-celled trichomes originating from outermost layer of ovule epidermis [1][2][3][4]. The development of cotton fiber may be divided into 4 stages, i.e. initiation, elongation, secondary cell wall deposition and maturation. Secondary cell wall deposition starts at around 14-17 days post anthesis (dpa) and lasts for over 30d [1,3,5]. In this stage, cellulose is intensely deposited to form a thick secondary cell wall. At maturation, cotton fiber consists primarily of secondary cell wall and over 90% dry weight of fiber may exist as cellulose. Therefore, carbon partitioning to cellulose biosynthesis is a key determinant of fiber weight and qualities, such as fiber strength and fineness [1,3,[5][6][7][8]. Many efforts have been taken to reveal the role of genes involved in the regulation of secondary cell wall deposition and to manipulate them by genetically modification for improvement of cotton yield and quality [1,6,7,9,10]. Recently, Jiang and coworkers showed that over-expressing a cotton sucrose synthase gene, GhsusA1, enhanced thickening of secondary cell wall and fiber qualities, suggesting an important role of sucrose synthase in controlling carbon partitioning to cellulose biosynthesis in cotton fibers [6].
Gibberellins (GA) are a class of important plant hormones involved in many physiological and developmental processes, including seed germination, cell elongation, photomorphogenesis, flowering and seed development [11]. In the last two decades, the molecular base of GA biosynthesis pathway and its regulation have been largely clarified in model plants [12][13][14]. Endogenous bioactive GA contents are regulated mainly through three 2oxoglutarate-dependent dioxygenases, i.e. GA 20-oxidase (GA20ox), GA 3-oxidase (GA3ox) and GA 2-oxidase (GA2ox) [13,14]. GA20ox and GA3ox catalyze the last two steps to synthesize bioactive GAs, while GA2ox convert bioactive GAs and their precursors to inactive 2-hydroxylated forms. A wealth of evidence demonstrated that both up-regulating GA20ox and suppressing GA2ox could significantly increase endogenous bioactive GA levels and lead to GA overproduction phenotypes in plants [15][16][17][18][19][20].
Physiological and molecular studies have revealed that GAs played important roles in fiber development. Exogenous application of GAs in vitro and in planta promoted fiber initiation and elongation [21,22]. Recently, we showed that over-expression of GhGA20ox1 in cotton significantly increased bioactive GA level and promote fiber initiation and elongation at early stage [17]. However, global up-regulation of GAs leaded to overgrowth of plant and somewhat negatively affected fiber development, especially at the late developmental stage. Instead, it is reasonable to elucidate GA roles in fiber development by tissue specific regulation of GA levels in developing fibers. To this end, we elevated the endogenous active GA levels in cotton fibers by tissuespecific up-regulation of GA20ox gene and down-regulation of GA2ox gene. We found that enhancement of GA production in fibers promoted sucrose synthase expression and secondary cell wall deposition. Our results implied that GAs might enhance carbon partitioning to cellulose and secondary cell wall synthesis via up-regulating sucrose synthase expression in cotton fibers.

Plant material and growth condition
Upland cotton (Gossypium hirsutum L. cv. Jimian No. 14) was used for cotton transformation and GA treatment. Cotton seedlings were grown in a greenhouse with a 16h/8h (light/dark) schedule and temperature kept at 26-30uC. Fibers and ovules were collected from field-grown cotton plants at growing season in Chongqing, China.
For GA 3 treatment of hypocotyls, 4-day-old seedlings were immersed in distilled water (pH6.0) or GA 3 solutions of various concentrations (0.05 mM, 0.1 mM and 0.5 mM, pH6.0) for 48 h. Then the hypocotyls were measured and collected for expression analyses and cellulose determination.
For GA treatment of in vitro fibers, cotton ovules were cultured as described by Beasley [23]. Cotton bolls were harvested at 0-dpa and surface-sterilized in 75% (v/v) ethanol for 1 min, rinsed in sterile water, then soaked in 0.1% w/v HgCl 2 solution for 12 min for sterilization, followed by rinsing with sterile water for six times. Ovules were separated, floated on BT media containing 5 mM IAA and GA 3 of various concentrations (0.5, 2.5, 10 and 25 mM), and then incubated in darkness at 32uC for 20d. Fibers were striped from ovules and used for RNA extraction.

Vector construction and plant transformation
To construct specific expression vector of cotton GA20ox (SCFP::GhGA20ox1), the CaMV 35S promoter in an overexpression vector (35S::GhGA20ox1) [17] was replaced with SCFP promoter [24]. The BAN promoter was amplified from Arabidopsis with a forward primer (59-TCTAGATAACAGAACCTTAC TGTAACACTATT-39) and a reverse primer (59-ACTAGT-GATTGTACTTTTGAAATTACAGAG AT-39) and cloned into TA cloning vector pMD19-T (TaKaRa, Dalian, China). After sequencing, the BAN promoter was digested from the cloning vector by HindIII and BamHI and inserted into a basic expression vector p5 vector [25] digested with the same enzymes to generate the vector p5-BAN. An intron-containing hairpin RNA construct of cotton GA20ox gene (GhGA2ox2RNAi) was amplified from cotton genomic DNA as previously described [26]. The 25-ml PCR mixture included 100 ng cotton genomic DNA, 106Ex Taq buffer (TaKaRa), 200 mM each dNTPs, 2 mM MgCl 2 , 400 nM flanking primer (59-GTATTGGTCTGGTGGGACTG-39), 40 nM bridge primer (59-CAAGTATCTCACATGCC AAGACCC-GAATTCTCCTTG-39), 1.5U Ex Taq DNA polymerase. The PCR thermo cycling parameters were as follows: 94uC for 5 min, followed by 35 cycles of 94uC for 30 s, 56uC for 30 s and 72uC for 30 s, and a final extension of 10 min at 72uC. The GhGA2ox2RNAi fragment was cloned and sequenced, then inserted into p5-BAN using BamHI and KpnI. Transgenic plants were generated using Agrobacterium-mediated transformation as described [25]. Based on expression analysis of target genes in transgenic cottons, two homologous transgenic lines were obtained by self-crossing, and their performances were documented at T3 and T4 generations in comparison with untransformed acceptor line (Jimian No. 14) grown in parallel in the field.

RNA extraction and qRT-PCR analyses
Total RNA was extracted from roots, hypocotyls, leaves, petals, anthers, ovules and fibers using a rapid plant RNA extraction kit (Aidlab, Beijing, China). The single-stranded cDNAs were synthesized from total RNA using a cDNA synthesis kit (TaKaRa, Dalian, China). The gene-specific primers used for real-time PCR amplification were list in table S1. Cotton histon3 gene (AF024716) was amplified as internal standard [27]. Real-time PCRs were performed on a CFX96 real-time PCR detection system with SYBR Green supermix (Bio-Rad, CA, USA). The thermocycling parameters were as follows: 95uC for 2 min, followed by 40 cycles of 95uC for 30 s, 56uC for 30 s and 72uC for 30 s, followed by a standard melting curve to monitor the specificity of PCR products. The reactions were duplicated for 3 times and data were analyzed using the software Bio-Rad CFX Manager 2.0 provided by the manufacturer.

Determination of endogenous GA contents
Cotton fibers (200 mg FW) were ground to fine powder in liquid N 2 , extracted overnight in 5 ml 80% methanol at 220uC and deuterium-labeled [17, 17-2 H 2 ] GA 1 and [17, 17-2 H 2 ] GA 4 (each 10 ng) from Prof. L. Mander (Australian National University) were added as internal standards. After centrifugation, supernatants were collected, dried in a rotavapor (BUCHI, Switzerland) at 40uC, and re-suspended in 3 ml 10% methanol. The extracts were applied on Oasis HLB extraction cartridges (60 mg, Waters) pretreated with 3 ml methanol and 3 ml water. After washing with 1 ml 10% methanol, GAs were eluted with 1 ml 90% methanol. The eluates were evaporated to dry, dissolved in 100 mL 10% methanol, and subjected to LC-MS assay. The procedures for LC-MS quantification of GAs were described previously [17].

Measurement of fiber quality
Mature fibers were harvested from the field-grown cotton in the same period (Aug. 20 to Sep. 10). After ginning, fibers were mixed well and 6 repeats of 10 g fibers were randomly sampled for each material. Fiber sample were tested independently for fiber quality traits (fiber length, fiber strength, micronarie value) using at a HVI system (HFT 9000, Uster Technologies, Swiss) in Cotton Fiber Quality Inspection and Testing Center, Ministry of Agriculture of China (Anyang, Henan, China).

Microscopic measurement of fiber cell wall thickness
Statistical analysis of cell wall thickness was performed according to Wang et al. [7]. After fixing in FAA (37% formaldehyde: acetic acid: ethanol: water, 10:5:50:35) at 25uC for 12 h, mature cotton fibers were dehydrated gradually in alcohol and tert-butyl alcohol series,and then infiltrated in tertbutyl alcohol/paraffin at 65uC and embedded in paraffin. The samples were sliced into 7-mm sections. The slices were mounted, stained with the Fast Green dye and photographed by a BX41TF light microscope (Olympus, Japan). Image-pro Plus program (Olympus) was employed to measure the thickness of cell wall and 1000 sections were measured for each sample.

Determination of fiber weight
Fibers on seeds were combed straight and striped manually. Approximately 1.5 mg fibers were randomly bundled and weighed precisely (W1). The fiber number of each fiber bundle (N) was counted as described [28]. The weight of 1000 fibers (W2) was calculated from the following equation: W2 = 1000W1/N. For each material, the average 1000-fiber weight was calculated on the basis of 60 fiber bundles from different seeds.

Sucrose synthase activity assays
The sucrose synthase was extracted according to Jiang et al. [6]. Fresh fibers (around 0.5 g) were ground to fine powder in liquid N 2 . The grinding continued for 5 min in cold extraction buffer (25 mM Hepes-KOH (pH 7.3), 5 mM EDTA, 1 mM DTT, 0.1% soluble PVP, 20 mM b-mercaptoethanol, 1 mM PMSF and 0.01 mM leupeptin). The homogenate was separated by centrifugation (10000 g, 5 min, 4uC) and the supernatant was used as the crude extracts for assays. Protein concentrations were determined via Bradford method [29]and sucrose sythase activities were assayed as previously described [30,31].

Analyses of soluble sugar contents
Fresh fibers (around 50 mg) were separated from developing bolls and ground to fine powder in liquid N2. The powder was extracted in 2 ml 80% (v/v) ethanol at 80uC?for 15 min. After centrifugation (3000 g, 10 min), supernatants were collected. The pellets were further extracted twice, and supernatants were combined and used for soluble sugar assays. The contents of glucose, fructose and sucrose were measured at 340 nm with a Synergy HT microplate reader (BioTek, Vermont, WS) as described [32].

Determination of cellulose content
Cellulose contents were determined according to Wang et al. [7]. Around 0.1 g fiber samples were extracted in 10 ml boiling acetic/nitric reagent (80% acetic/nitric, 10:1) for 1 h, then rinsed three times with distilled water and once with ethanol. Residuals were dried at 105uC for 2 h. The weight ratio of residual to initial samples was regarded as cellulose content. Hypocotyls were excised from seeding (10 hypocotyls per sample) and ground to fine powder in liquid N 2 . Samples were dried at 105uC for 2 h and extracted in 10 ml boiling acetic/nitric reagent (80% acetic/nitric, 10:1) for 1 h.Determination of cellulose content of per hypocotyls was carried out as described [33].

Statistical analyses
Performances of transgenic materials were compared to wildtype control and statistical significance of divergence between averages was determined by t test. All statistical calculations were performed using Microsoft Excel.

Enhancement of GA production in cotton fiber
We used two strategies to tissue-specially enhance GA production in fibers, i.e. to promote GA biosynthesis by upregulation GA 20-oxidase (GA20ox) and suppressing GA deactivation by down-regulation of GA2-oxidase (GA2ox). To this end, we used a fiber-specific promoter (SCFP) [24] and a seed coat-and fiber-specific promoter BAN [28,34] to direct the expression of GA20ox (GhGA20ox1) [17], and GA2ox, respectively. Among SCFP::GhGA20ox1 transgenic cottons (SG20), SG20-1 showed dramatically increase in GhGA20ox1 expression level in developing fibers ( Figure 1A and 1B).
We compared the expression pattern of six cotton GA 2-oxidase genes (GhGA2ox1-6, Figure S1,3). Among them, GhGA2ox2 showed predominant expression in fibers. Thus we selected GhGA2ox2 as RNAi target to suppress GA deactivation in fibers, and generated GhGA2oxRNAi transgenic cottons. Real-time RT-PCR revealed that the expression level of GhGA2ox2 was reduced in BAN::GhGA2oxRNAi (BG2i) transgenic cottons ( Figure 1C), in which transformant BG2i-2 showed most significant suppression of the target gene in fibers ( Figure 1C and 1D).
To detect the effect of GhGA20ox1 up-regulation and GhGA2ox2 down-regulation on GA homeostasis in fibers, we determined the contents of endogenous bioactive GAs (GA 1 and GA 4 ) in 8-and-20 dpa fibers of SG20-1 and BG2i-2 by LC-MS ( Figure 1E and 1F). Compared to the wild-type control, GA 4 contents in the 8and 20-dpa fibers of SG20-1 fibers increased 83.1% and 178.6%, respectively, while GA 1 content was moderately increased (24.0%) in the 20-dpa fibers( Figure 1E and 1F). In BG2i-2 fibers, GA 1 level was 21.6% and 65.9% higher than the control at 8-and 20-dpa respectively, whereas GA 4 contents remained almost unchanged compared to the control ( Figure 1E and 1F).

Effects of elevated GA levels on secondary cell wall thickening of cotton fiber
To clarify the effect of elevated GA levels on fiber development and fiber quality, we compared the agronomy performances of transgenic lines SG20-1 and BG2i-2 with the wild-type control in consecutive two-year field trails. No significant change in plant growth, yield traits and fiber length and strength (Figure 2A and S4; Table S2 and S3) was found between the transgenic lines and the wild type, except micronaire value. The micronaire values of SG20-1 and BG2i-2 fibers were significantly higher than that of the wild type ( Figure 2C). Micronaire value is a composite measure of fiber maturity and fineness. To clarify whether the fiber fineness was increased in transgenic cotton, we measured the weight per 1000 fibers. The weights of SG20-1 and BG2i-2 mature fibers significantly enhanced in comparison with the control (2.0% and 5.7%, respectively; Figure 2E). Microscopic observation further confirmed that the cell walls of SG20-1 and BG2i-2 mature fibers were thicker (5.4% and 6.6%, respectively) than the wild-type control ( Figure 2B and D). Considered that most of cell walls of mature cotton fibers consisted of secondary cell wall [2], it was reasonable that the fineness increase in SG20-1 and BG2i-2 fibers might be mainly attributed to promotion of secondary cell wall deposition. To prove this hypothesis, we determined the cellulose contents in fibers, and found that the contents of SG20-1 and BG2i-2 fibers were significantly higher than the control ( Figure 2F). Taken together, these results suggested that elevating bioactive GA levels in cotton fibers promoted secondary cell wall deposition.

Sucrose synthase expression in response to elevated GA levels in fibers and hypocotyls
To reveal the possible mechanism for GAs to control secondary cell wall deposition, we investigated transcript levels of six genes related to secondary cell wall biosynthesis, including GhCesA1, GhCesA2, GhRac13, GhSusA1, GhADF1 and GhCTL1 [6,7,35], in 20dpa fibers. Only sucrose synthase gene (GhSusA1) showed significant increase in transgenic cottons ( Figure 3A). The relative transcript levels of GhSusA1 in SG20-1 and BG2i-2 fibers were 53% and 50% higher than the control, respectively. Biochemical analysis demonstrated that the sucrose synthase activities in 20-dpa fibers of SG20-1 and BG2i-2 increased 8.3% and 10.7%, respectively, compared to the control ( Figure 3B). Meanwhile, the concentration of fructose, a direct product of sucrose synthase, was significantly higher in SG20-1 and BG2i-2 fibers ( Figure 3C). Furthermore, we found GhsusA1 transcript in cultured fibers was increased with GA 3 concentrations in ovule culture media ( Figure 3D). The result of fiber culture, along with the observations on the mature fibers, implied that GA may promote cellulose biosynthesis and secondary cell wall deposition through upregulation of the expression of sucrose synthase.
Like cotton fibers, hypocotyls that undergo rapid cell elongation require high-speed formation of cellulose. To investigate if same response takes place in hypocotyls, we detected the expression of sucrose synthase in the GA-treated hypocotyls. GhSusA1 transcript levels in hypocotyls were significantly enhanced along with increase of GA 3 ( Figure 4A). Meanwhile, elongation and cellulose deposition were accelerated ( Figure 4B-D). These results further supported that GA promoted cellulose biosynthesis and secondary cell wall deposition through up-regulation of sucrose expression.

Discussion
Genetically manipulating enzymes involved in GA biosynthesis and catabolism provided an effective strategy to regulate GA homeostasis in plants [15][16][17][18][19][20]. Our work demonstrated that in addition to up-regulation of GA20ox, down-regulation of GA2ox is another effective way to increase the active GA levels in plant tissues ( Figure 1E and F). However, there may be a subtle difference between the two strategies. Up-regulation of GA20ox not only increased the GA level, but also changed the composition of some active GAs by promoting the non-13-hydroxylation pathway for producing 13-H GA 4 rather than 13-OH GA 1 (Figure 1E and F) [17,19,20]. On the contrary, in our study, downregulation of GA2ox elevated the GA 1 instead of GA 4 ( Figure 1E and F). The exact physiological effects of different bioactive GAs on cotton fiber development are still to be elucidated.
Cotton fibers that undergo rapid elongation and intense cellulose synthesis represent a strong sink that competes with the developing embryos and endosperms in a single ovule [1,3,5,6,8,30,36,37]. It was revealed that sucrose synthase played important roles in carbon partition during fiber development. Suppression of sucrose synthase gene (SS3) inhibited fiber initiation and elongation [30], while over-expression of a potato sucrose synthase gene in cotton enhanced leaf expansion, early seed development and fiber elongation [5]. Recently, Jiang and coworkers cloned a novel cotton sucrose synthase gene (GhSusA1), which might be a key regulator of sink strength in cotton. Overexpression of GhSusA1 significantly enhanced cell wall thickening during secondary wall formation stage, and improved fiber length and strength [6]. In this study, we revealed that enhancement of GA level in cotton fibers led to an increase of sucrose expression (GhSusA1) gene expression, and promoted cellulose biosynthesis and secondary cell wall deposition (Figure 2 and 3). Moreover, GA-induced GhSusA1 up-regulation also found in cultured fibers and hypocotyls (Figure 4 and S5). GA has long been considered as an important regulator of sink strength in plants, but the molecular basis of how GA enhances the partitioning of carbon assimilates to sink tissues is still unknown [38,39].Our data offer an experimental evidence for the relationship between GAs and sucrose synthase.
Previous studies showed that exogenous application of GAs or constitutively increased endogenous GA levels promoted fiber initiation and elongation [17,21]. However, in this study we did not find significant improvement in fiber length of the transgenic GA-elevated fibers (TableS2). A possible explanation for this phenomenon is that the enhanced secondary cell wall deposition may fix the morphology of the fiber and, in turn, limit further elongation of the GA-enhanced fibers at the late stage of fiber elongation. Nevertheless, the finding that GA related regulation of sucrose synthase gives useful information to reveal the mechanism of cotton fiber development and to improve fiber yield and quality for cotton breeders.