Qin Hu, Shenghu Xio, Qinqin Gun,Lili Tu, Feng Sheng, Xuezhu Du,,Xinlong Zhng,

aState Key Laboratory of Biocatalysis and Enzyme Engineering,School of Life Sciences,Hubei University,Wuhan 430000,Hubei,China

bNational Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430000,Hubei,China

Keywords:Cotton GhLac1 Fiber development Jasmonic acid Flavonoids

ABSTRACT Cotton fibers are single cells originating in the epidermis of cotton ovules,and serve as the largest natural fiber source for the textile industry. In theory, all epidermal cells have the potential to develop into fibers, but only 15%-25% of epidermis cells develop into commercially viable lint fibers. We previously showed that GhLac1 participates in cotton defense against biotic stress. Here we report that GhLac1 also has a role in cotton fiber development. GhLac1 RNAi lines in cotton showed increased differentiation of fiber initials from epidermis and shortened fiber length, resulting in unchanged lint percentage.Suppression of GhLac1 expression led to constitutively hyperaccumulated jasmonic acid(JA)and flavonoids in ovules and fiber cells.In vitro ovule culture experiments confirmed the distinct roles of JA and flavonoids in fiber initiation and elongation, and showed that fiber development is spatially regulated by these chemicals: the increased fiber initiation in GhLac1 RNAi lines is caused by hyperaccumulated JA and rutin content during the fiber initiation stage and shortened fiber length is caused by constitutively increased JA and naringenin content during the fiber elongation stage.

1. Introduction

Cotton is the main natural fiber source for the textile industry.Cotton fibers are single-cell trichomes that originate in the epidermis of cotton ovules during the period from one day before anthesis (?1 DPA) to one day after anthesis (1 DPA),known as the fiber initiation period [1]. After initiation, fiber development proceeds through four overlapping but distinct stages: elongation, transition, secondary cell wall thickening,and maturation [2]. All epidermal cells have the potential to become fibers, but only 15%-25% of epidermis cells pass through all the stages and develop into commercially viable lint fibers[3,4].Thus,the yield of cotton fiber depends largely on the number of fibers initiated in the epidermis,suggesting a target for improving cotton fiber yield[5].

Cotton fibers share many similarities with Arabidopsis leaf trichomes[6,7].Studies[8-11]have shown a close relationship between these two types of cells using cotton fiber-associated genes,but the precise molecular mechanisms for cotton fiber initiation remain little known [12]. The homeostasis and regulation of cotton endogenous hormones play pivotal roles in fiber development [13,14]. Jasmonic acid (JA), a hormone biosynthesized from linolenic acid via the octadecanoid pathway, plays a key role in response to abiotic and biotic stress as well as in plant growth and development, including trichome development, leaf abscission, and senescence [15].In Arabidopsis, aos mutant plants exhibited blocked JA biosynthesis and leaves with significantly fewer trichomes,and the phenotype could be rescued by exogenous application of JA [16]. JA also induced the expression level of a key bHLH transcription factor GL3 involved in trichome formation that interacted with GL1 and TTG1 to promote trichome initiation[16]. JAZ proteins, the critical negative regulator of the JA signaling pathway, interacted with members of the R2R3-MYB/bHLH/WD40 transcription complex including GL1, GL3,and EGL3 to repress trichome initiation [17]. An increasing number of studies[18-20]have shown that JA also plays a role in cotton fiber development. An optimal JA concentration(0.001 μmol L?1) promoted fiber initiation in in vitro cultured ovules [20], whereas a higher JA concentration (2.5 μmol L?1)inhibited fiber initiation.

Flavonoids,a highly diverse class of low-molecular-weight secondary metabolites, play roles in pigmentation and UV light protection and act in several biological processes including transcriptional regulation and cell-to-cell communication [21-24]. The regulatory mechanism of flavonoid biosynthesis, root hair and trichome patterning has been intensively studied in Arabidopsis, where the core regulator is a ternary R2R3-MYB/bHLH/WD40 transcription complex[25-27].Several mutants with impaired flavonoid biosynthesis exhibit distinct phenotypes for trichome number and/or shape[28-30].In cotton,different flavonoids exerted different effects on fiber development during in vitro ovule culture,among which naringenin(NAR)and dihydrokaempferol(DHK)retarded fiber development [31]. Knocking down the expression level of flavanone 3-hydroxylase (F3H) by RNA interference in cotton increased NAR content in fiber cells and led to shortened fiber length [31]. However, whether flavonoids affect cotton fiber initiation has not been reported.

We have reported [32] that GhLac1, a laccase gene from upland cotton,is crucial for regulating tolerance against pests and pathogens of cotton, such as bollworm, aphid and Verticillium wilt, via manipulation of the phenylpropanoid pathway and JA synthesis. In the present study, we found that the expression level of GhLac1 was very high in developing fiber cells. The molecular and biochemical approaches were combined to investigate the bio-function of GhLac1 in fiber development. We verified that suppression of GhLac1 led to hyperaccumulated JA and flavonoid content which showed a cooperative effect on fiber initiation and elongation.These results depicted a new mechanism for fiber development regulation.

2. Materials and methods

2.1. Plant materials and growth conditions

Cotton plants of GhLac1 transgenic lines [32] and wild-type(WT, the transgene receptor cultivar YZ1) were grown in an experimental field under normal farming practices or grown in the greenhouse during the winter. The greenhouse was maintained at 25-28 °C under long-day conditions with an 8 h/16 h dark/light photoperiod and a relative humidity of 60%.Ovules and fibers at several development stages were collected from the plants grown in the experimental field.

2.2. Measurement of immature fiber length, mature fiber quality, lint percentage, and seed index

Samples for immature-fiber length measurement and mature-fiber quality analysis were collected from fieldgrown plants in 2015 and 2016. To track the fiber development of GhLac1 transgenic lines, cotton bolls were collected at 5, 10,15, 20, and 25 DPA, and fiber length measurement followed Hu et al. [20]. To test the fiber quality, lint percentage and seed index of the transgenic lines and WT, mature bolls were collected from the middle part of the cotton plants. After ginning, fiber (≥10 g sampled from each of at least 30 bolls)were kept at room temperature for two days and tested with a HFT9000 high volume fiber test system (Premier, Coimbatore,India). Lint percentage (fiber weight/seed cotton weight), seed cotton (seed + lint) and seed index (seed weight in grams per 100 seeds) were calculated. All the experiments were performed with at least three replicates.

2.3. RNA extraction and expression analysis

Total RNA of each tissue was isolated following Deng et al.[33],and the cDNA was reverse-transcribed using M-MLV reverse transcriptase (Promega, Beijing, China). Approxi-mately 3 μg of RNA was added into a 25-μL reaction mixture according to the manufacturer’s protocol. Quantitative real-time PCR (qRTPCR) was performed with PowerUp SYBR Green Master Mix(Applied Biosystems, Vilnius, Lithuania) on a 7500 Real-time PCR System (Applied Biosystems, Foster City, CA, USA)according to the manufacturer’s protocol. Relative gene expression levels were calculated by the 2-ΔΔCTmethod. Each tissue of each genotype was assayed in three biological and three or more technical replicates, with GhUB7 (GenBank accession number DQ116441) as an internal reference. Primers are listed in Table S1.

2.4. Measurement of phytohormones and flavonoids

Concentrations of endogenous JA and jasmonoyl-L-isoleucine (JA-Ile) in ovule or fiber were measured following Sun et al. [34]. Fresh samples of approximately 200 mg were ground in liquid nitrogen and incubated overnight at 4 °C in 300 μL extraction buffer containing 80% (v/v) methanol and 15 ng mL?1[+/-]9-,10-dihydro-JA. The mixture was then centrifugated for 20 min at 12,000 r min?1at 4 °C. The sediment was twice extracted with 200 μL of extraction buffer. The two supernatants were collected and mixed for quantifying JA and JA-Ile [35]. The internal standard was[+/-]9-,10-dihydro-JA (Olchemim, Olomouc, Czech Republic).

Total flavonoid content was measured by a previously reported [32] spectrophotometric method. A sample of approximately 200 mg of ovules or fibers was ground in liquid nitrogen and extracted with 1 mL 80% (v/v) methanol overnight at 4 °C. The mixture was then centrifuged for 30 min at 12,000 r min?1at 4 °C. The sediment was twice extracted with 1 mL 80% (v/v) methanol for 1 h at 4 °C. The two supernatants were collected and mixed for quantifying total flavonoids. A 100-μL aliquot of the extract or rutin standard solution was diluted with 0.5 mL distilled water in a test tube.The solution was stabilized for 6 min after addition of 30 μL NaNO2(5%, w/v) and for another 5 min after addition of 60 μL AlCl3(10%, w/v). Then 0.2 mL of NaOH (1 mol L?1) was used to terminate the reaction, and 1 mL distilled water was added and mixed well. The absorbance was measured immediately against a blank at 420 nm using a Multimode Plate Reader(Perkin Elmer, Waltham, MA, USA).

To determine the endogenous concentrations of NAR and rutin, samples of approximately 200 mg of ovules or fibers were ground in liquid nitrogen and extracted with 300 μL cold extraction buffer (80% methanol, v/v) overnight at 4 °C. After centrifugation for 20 min at 12,000 r min?1at 4 °C, the supernatant was collected and the residual pellet was reextracted with 300 μL of cold extraction buffer for 1 h at 4 °C.The two supernatants were collected and mixed for quantifying concentrations of NAR and rutin on a LC-MS instrument according to a previous study [31].

2.5. Ovule culture and microscopic observation

To observe the initiation of fiber cells, ovules collected from the center parts of the bolls at 0 DPA were fixed in 2.5% (v/v)glutaraldehyde and stored at 4 °C until use. After dehydration in an ethanol series, the samples were transferred into isoamyl acetate and dried to the critical point. Fiber initiation was observed and photographed with a JSM-6390/LV scanning electron microscope (SEM, Jeol, Tokyo, Japan).

Ovule culture was performed following a previous description [20]. To observe fiber initiation, ?2 DPA ovules from the same boll were divided into equal parts by treatment and cultured in liquid BT medium [36] to improve the reliability and consistency of the experiments. After 72 h of culture, the ovules were collected and fixed immediately in 2.5% (v/v)glutaraldehyde solution and then stored at 4 °C until use.After dehydration in an ethanol series, the samples were transferred into isoamyl acetate and dried to the critical point. Fiber initiation was observed and photographed by SEM.

To observe the effects of different chemicals on fiber elongation, the 0 DPA ovules from the same boll were divided into equal parts by treatment and cultured in flasks with equal volume of liquid BT medium for 15 days. They were then harvested for photography and fiber yield measurement.For fiber length measurement, at least 25 ovules were measured. The cultured ovules were soaked in 95 °C water for 5 min and measured with a ruler [37]. Fiber yield was expressed as total fiber units (TFU) as previously [31]described. Briefly, the cultured ovules were immersed in hot water to disperse the fibers, dried, stained for 30 s in 0.02% toluidine blue O,and then washed in running water for 2 min.The ovules were de-stained in glacial acetic acid:ethanol:water (10:95:5, v/v/v) for 2 h. The solvent absorbance was measured at 624 nm with the plate reader. More than three biological replicates were performed for each treatment, and each treatment with different chemicals was performed with at least 3 flasks from at least 6 bolls.

3. Results

3.1. GhLac1 is expressed predominantly in cotton fiber cells and associated with fiber traits

In addition to conferring tolerance to biotic stresses [32],GhLac1 may participate in fiber development, given that its expression in fiber cells was very high during fiber development (Fig. S1). Expression analysis in GhLac1 transgenic lines showed that GhLac1 was markedly upregulated and downregulated in respectively overexpression and RNAi lines in developing fiber cells(Fig.1-A).Accordingly,GhLac1 transgenic lines were used to investigate the role of GhLac1 in fiber development. Mature fiber length was not affected by overexpression of GhLac1 compared to the WT, but was significantly reduced in GhLac1 RNAi lines in two-year field experiments (Fig. 1-B, Table 1). Micronaire is a measure of the air permeability of cotton fiber and an indication of fiber fineness and maturity. The micronaire value increased by 6.9%-19.2% in GhLac1 overexpression lines but decreased by 3.2%-7.7% in RNAi lines as compared with that of WT. The fiber strength (the force required to break the fiber, and defined as grams per tex,“tex”is a unit equal to the weight in grams of 1000 m of fiber)of the RNAi lines was lower than that of the overexpression lines (P ＜0.05), which did not differ significantly from the WT (Table 1). Fiber elongation was markedly retarded from 5 to 25 DPA in the GhLac1 RNAi lines,but showed no evident difference from the WT in the overexpression lines(Fig.2-A-E).

3.2.Downregulating GhLac1 promoted cotton fiber initiation

Although the RNAi lines of GhLac1 showed shorter mature fiber length, the lint percent of the RNAi lines showed no significant difference from those of the WT and overexpression lines(Table 1).This phenotype could be attributed to altered seed weight and/or mature-fiber numbers. There was no significant difference in seed index and seed size between the transgenic lines and WT(Fig.S2).Nor was there any evident difference in fuzz (the short fiber on the seed after ginning)between the transgenic lines and WT(Fig.S2).Lint fibers are initiated in the epidermis of cotton ovules during the period ?2 to 0 DPA and ultimately reach 2.5-3.5 cm in length [38]. Ovules from the transgenic lines and WT at 0 DPA were used to determine the numbers of initial lint fibers. Numbers of initial lint fibers of the RNAi lines were much higher than that of the WT, but the overexpression lines showed no significant difference from the WT (Fig. 3-A, B). Thus, the unchanged lint percent of the RNAi lines was due mainly to increased lint fiber numbers.

Fig.1- Downregulation of GhLac1 results in shortened mature fiber.(A)Quantitative real-time PCR measurement of GhLac1 expression in WT and GhLac1 transgenic lines.WT,wild type(transgenic receptor cultivar YZ1);OL-12,OL-13 and OL-56,GhLac1 overexpression lines;iL-1 and iL-1,GhLac1 RNAi lines.Values are normalized to GhUB7 and expressed as means±SD;n=3.(B) Morphological differences in mature fiber between WT and GhLac1 transgenic lines.Mature fiber length was measured in T6 overexpression lines,RNAi lines and WT.Comparisons were performed with Student's t-test. *, P ＜0.05; **,P ＜0.01.Bar,1 cm.

3.3. Downregulating GhLac1 led to accumulated JA content in developing fiber cells

In a previous study [19], using lintless-fuzzless XinWX and linted-fuzzless XinFLM cotton,JA metabolism was associated with cotton fiber initiation. Appropriate JA application induced fiber initiation in in vitro ovule culture [20]. In our previous study [32], JA content was increased in roots and leaves of GhLac1 RNAi lines, but showed no significant difference between the overexpression line and WT under disease-free conditions. In agreement with this finding, JA was constitutively hyperaccumulated in the RNAi lines from?2 DPA ovules to 15 DPA fiber cells,and showed no significant difference between WT and GhLac1 over-expression lines(Fig. 4-A). Moreover, endogenous JA content was highest in ovules at ?1 DPA and dramatically decreased during fiber development,suggesting that fiber initiation requires a high concentration of JA but that fiber elongation requires a decrease in JA content (Fig. 4-A). Ovules at ?2 DPA ovules were collected from WT and treated with 0-0.1 μmol L?1JA to investigate the effects of JA on fiber initiation. The results showed that ovules treated with 0.0005 and 0.0010 μmol L?1JA exhibited significantly increased initiation numbers (Fig.4-B-E),and high (0.020 and 0.100 μmol L?1) concentrations of JA significantly inhibited fiber initiation (Fig. 4-F, G, S3). In 0 DPA ovules from WT low JA concentrations of 0.0005 and 0.0010 μmol L?1showed no effect on fiber elongation including fiber length and fiber yield, whereas higher concentrations of 0.020 and 0.100 μmol L?1significantly inhibited fiber elongation including fiber length and fiber yield(Figs.S3,S4).

3.4. Altered metabolites in GhLac1 RNAi lines are associated with fiber initiation and elongation

Suppression of GhLac1 expression leads to a redirection of metabolic flux in the phenylpropanoid pathway in the RNAi lines[32].Flavonoid metabolism is activated in early fiber cell development, and different flavonoid chemicals showed different effects on fiber development [31,39-42]. Accordingly, total flavonoid content of ovules from 0 DPA to 15 DPA fiber cells from GhLac1 transgenic lines and WT was measured. As shown in Fig. S5, total flavonoid content was significantly increased in the GhLac1 RNAi lines compared to WT,and slightly decreased in the overexpression lines.NAR is a negative regulator of fiber elongation [31]. Roots and young leaves of GhLac1 RNAi lines showed relatively high rutin content and rutin significantly inhibited the growth of V. dahliae mycelium and cotton bollworm in our previous study[32].For this reason,we investigated the NAR and rutin content in 0 DPA ovules to 15 DPA fiber cells from GhLac1 transgenic lines and WT. In accord with expectation, the contents of NAR and rutin were significantly greater in fiber cells of GhLac1 RNAi lines (Fig. 5-A, B). The ?2 DPA or 0 DPA ovules were collected and subjected to in vitro ovule culture in the presence of different flavonoids for 3 or 15 days to observe fiber initiation or elongation, respectively (Figs. 5, 6). The results showed that 10 μmol L?1NAR had no effects on fiber initiation but induced many abnormal cell protuberances(Fig.5-C),whereas 10 μmol L?1NAR was sufficient to suppress fiber elongation (Fig. 6-A, C) and 20 μmol L?1rutin significantly promoted fiber initiation (Fig. 5-D), but showed no effect on fiber elongation (Fig. 6-B, D). The fiber length and total fiber units measurements also support this conclusion(Fig.6-B,D).

Table 1-Fiber quality of field-grown transgenic lines and WT in 2015 and 2016.

4. Discussion

4.1. The role of JA in cotton fiber development

Phytohormones, are small molecules that are essential for the regulation of plant growth, development, reproduction and survival [13,43]. Studies [44,45] of in vitro cultured ovules with exogenous hormone application showed that phytohormones are indispensable for cotton fiber development. Cotton fiber cell development regulation is commonly compared with Arabidopsis leaf trichome development because of the similarities between these two cell types[8-11,46],but the regulatory pathway of cotton fiber initiation and elongation remains little known. Given that exogenous JA could be used to manipulate cotton fiber development and increase Arabidopsis leaf trichome density, the role of JA in fiber development has attracted research attention[18,20,47,48]. Appropriate low concentrations of JA promote fiber initiation and high concentrations of JA inhibit cotton fiber initiation and elongation [20,47,48], in agreement with our findings (Figs. 4-B-G and S4). Comparative transcriptomics of the lintless-fuzzless fiber mutant XinWX and the fuzzless-linted mutant XinFLM revealed [19] that JA metabolism participated in fiber initiation and that four members of the allene-oxide cyclase (AOC) family that functions in JA biosynthesis are upregulated in fiber initiation, especially at?1 DPA, while overproduction of JA disrupted normal fiber development.Thus,JA metabolism and JA-associated signaling pathway undisputedly participate in cotton fiber development and play complex roles at different fiber cell development stages.

The JAZ protein family are key repressors of the JA signaling pathway. In the presence of JA (JA-Ile), the JA-Ile-COI1-JAZ complex was recognized by 26S proteasome system,leading to the degradation of JAZ protein to initiate downstream responses [49,50]. GhJAZ2 was reported [20] to be a negative regulator of cotton fiber initiation with high expression level at ?1 DPA,and overexpression of GhJAZ2 resulted in reduced fiber initiation. GhJAZ2 interacted with GhMYB25-like, GhGL1, GhMYC2, and GhWD40 proteins, which are the core components of the WD-repeat/bHLH/MYB transcriptional complex, to inhibit cotton fiber initiation. In this study, we found that the GhLac1 RNAi lines showed increased fiber initiation with hyperaccumulated JA in ovules and fiber cells.The expression levels of genes important for cotton fiber initiation were markedly upregulated in GhLac1 RNAi lines,consistent with the downregulated GhJAZ2 transcriptional level (Fig. S6). Thus, a partial explanation for the increased fiber initiation in GhLac1 RNAi lines is the increased JA content at the fiber initiation stage, and the unchanged JA content accounts for the unchanged fiber initiation in GhLac1 overexpression lines.

Fig.2-Downregulation of GhLac1 leads to retarded fiber development.(A)-(E)Morphological differences and measurements of fiber length in WT and GhLac1 transgenic lines at several development stages.Values are means±SD;n =20. Comparisons were performed with Student's t-test. **,P ＜0.01. Bar,1 cm.

The peaking of JA content in ?1 DPA ovules, followed by its sharp decrease with flowering time, suggests that fiber initiation requires a relatively high concentration of JA, but that this concentration is not suitable for fiber cell elongation[20]. In in vitro ovule culture with exogenous JA supply [47],fiber cell elongation was quite sensitive to JA concentration.In the present study, low (0.0005 and 0.0010 μmol L?1) concentrations of JA promoted fiber initiation and showed no effect on fiber elongation, but a high (0.1 μmol L?1) concentration of JA significantly suppressed fiber development (Figs. 4-B-G,S4). Thus, the shorter fiber length of the GhLac1 RNAi lines may be caused by constitutively accumulated JA during the fiber elongation stage(Fig.4-A).

Fig.3-Downregulation of GhLac1 promoted fiber initiation.(A)Scanning electron microscope(SEM)images of ovules(at 0 DPA)of GhLac1 transgenic lines and WT.(B)Numbers of fiber initiations in the rectangular areas(as shown in(A))on the ovules of the transgenic lines and controls.More than 20 ovules were counted.Values are means±SD;n=20.Comparisons were performed with Student's t-test.**,P ＜0.01.

4.2.The flavonoid metabolism is a novel target for cotton fiber yield and quality improvement

Flavonoids, synthesized in one branch of phenylpropanoid metabolism, are involved in myriad biological processes in plants [23]. The presence of mutants in each step of the flavonoid biosynthesis pathway in Arabidopsis has provided solid evidence for the importance of flavonoids in modulating signaling transduction in plants [51]. Aberrant accumulation of flavonoids has also been linked to hyponastic cotyledons,altered shape of pavement cells,and deformed trichomes[52].The function of flavonoids in cotton fiber development has been little investigated owing to lack of cotton mutants with impaired flavonoid synthesis. The Arabidopsis rol1-2 mutant showed strongly reduced amounts of flavonols glycosylated with multiple rhamnose units compared with the wild type,and exhibited hyponastic growth,aberrant pavement cell and stomatal morphology in cotyledons, and defective trichome formation [30]. The defective trichome formation of rol1-2 is fully suppressed in the rol1-2 tt4 double mutant, confirming that the phenotype is flavonoid-dependent;exogenous application of naringenin to the rol1-2 tt4 double mutant rescued the defective trichome formation of rol1-2 and induced the irregular cell shapes characteristic of the rol1-2 mutation[30].In in vitro ovule culture, exogenous application of naringenin and dihydrokaempferol strongly retarded fiber development,whereas eriodictyol,dihydroquercetin,kaempferol,and quercetin showed little effect [31]. These findings indicate that different flavonoids exert different effects on trichome formation and/or fiber development.In the present study we showed that NAR and rutin play different roles in cotton fiber initiation and elongation. The exogenous application of NAR significantly suppressed fiber elongation but had no effect on fiber initiation, and NAR also induced abnormal cell protuberances, in agreement with the phenotype of the rol1-2 tt4 double mutant (Figs. 5-C, 6-A, C). Exogenous application of rutin promoted fiber initiation but showed no effect on fiber elongation (Figs. 5-D, 6-B, D). Thus, the increased fiber initiation and shorter fiber length of GhLac1 RNAi lines might result from the complex interaction of constitutively hyperaccumulated JA and flavonoids during fiber development (Figs. 4, 5, S4). In GhLac1 overexpression lines, the unchanged JA content and decreased flavonoid content may compensate each other and result in the unchanged fiber initiation and fiber length.

Fig.4- Increased fiber initiation depends partially on hyperaccumulated JA content in GhLac1 RNAi lines.(A)JA content in ovules or fiber cells from GhLac1 transgenic lines and WT at several development stages.Values are means± SD;n = 6. Comparisons were performed using Student's t-test.*,P ＜0.05; **,P ＜0.01.(B)-(G) SEM images of in vitro cultured ovules(?2 DPA)treated with several concentrations of JA for 72 h.

There are two potential mechanisms by which flavonoids may regulate growth and development. One target of flavonoids is auxin transport [52]. One flavonol, quercetin,inhibited IAA transport via the ATP-binding cassette protein B19 in vitro and inhibit shootward auxin transport and gravitropism in Arabidopsis [53,54]. In contrast, kaempferol is the active flavonoid regulating rootward auxin transport in Arabidopsis inflorescences [55,56]. Another target of flavonoids is maintenance of antioxidant activity. Flavonoids could act as antioxidants to reduce reactive oxygen species(ROS) levels and modulate the ROS-mediated signaling pathway [23,24,57]. The present study provides solid evidence for the function of NAR and rutin in cotton fiber development. However, several questions invite further research, including the in vivo optimal levels of NAR and rutin at different cotton fiber development stages and the mechanism by which NAR and rutin regulate cotton fiber initiation and elongation.

In our previous study [32], GhLac1 functioned as a lignin polymerization enzyme and transgenic manipulation of the expression of GhLac1 in cotton resulted in altered cell wall structure and components mimicking the cell wall damage associated with defense responses and led to changed JA and flavonoid content in roots and young leaves. Although the cotton fiber cell is considered[58]to have no lignin deposition in its secondary wall, several lignin-like phenolics were accumulated in cotton fibers, and genes encoding key enzymes in monolignol biosynthesis were highly expressed during the secondary wall synthesis stage [59,60]. Overexpression of a LIN-11,Isl1 and LIM-domain protein(WLIM1a)in cotton resulted in longer fiber length and increased lignin deposition in mature fiber cells compared to a control [61].These results support the hypothesis that lignin or lignin-like phenolics are synthesized in cotton fibers and play important roles in regulating fiber development. It is difficult to collect enough cell wall residue from ?2 to 15 DPA fiber to measure their lignin and polysaccharide content. Whether the altered JA and flavonoid content in the fiber cells of GhLac1 transgenic lines shared the same mechanism in roots and young leaves awaits further research.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2019.11.006.

Fig.5- NAR and rutin show distinct effects on fiber initiation.(A)NAR content in ovules or fiber cells from GhLac1 transgenic lines and WT at several development stages.Values are means±SD;n=6.Comparisons were performed with Student's t-test.*,P ＜0.05; **,P ＜0.01.(B)Rutin content in ovules or fiber cells from GhLac1 transgenic lines and WT at several development stages.Values are means±SD;n=6. (C)and (D)SEM images of in vitro cultured ovules(?2 DPA)treated with NAR or rutin for 72 h.White arrows in(C)indicate abnormal cell protuberances induced by NAR.

Declaration of competing interest

The authors declare no competing financial interests.

Acknowledgments

We are indebted to Dongqin Li and Hongbo Liu (National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, China) for phytohormone and flavonoid determination. This work was financially supported by the National Transgenic Plant Research Program of China(2016ZX08005-001), the Program of Introducing Talents of Discipline to Universities in China(B14032),the open funds of the National Key Laboratory of Crop Genetic Improvement(ZK201901) and the National Natural Science Foundation of China(31771837).

Fig.6-NAR and rutin show distinct effects on fiber elongation.(A)In vitro cultured ovules(0 DPA)from WT treated with several concentrations of NAR for 15 days.(B)In vitro cultured ovules(0 DPA)from WT treated with several concentrations of rutin for 15 days.(C)Fiber length of ovule cultured with several concentrations of NAR or rutin for 15 days.Values are means±SD;n=25.(D)Fiber yield(expressed as total fiber units)of ovules cultured with several concentrations of NAR or rutin for 15 days. Values are means±SD;n =25.Comparisons were performed with Student's t-test.**,P ＜0.01.Bar,1 cm.