Down-regulation of GhADF1 in cotton (Gossypium hirsutum) improves plant drought tolerance and increases fiber yield

Lixia Qin, Huanyang Zhang, Jing Li, Yonghong Zhu, Gaili Jiao, Chuangyun Wang, Shenjie Wu

Shanxi Agricultural University, Taiyuan 030031, Shanxi, China

Keywords:Cotton (Gossypium hirsutum)Actin depolymerizing factors Drought stress Fiber yield Transcriptomic analysis

A B S T R A C T Fiber productivity and quality of cotton are severely affected by abiotic stresses. In this study, we identified the role of GhADF1, an actin depolymerizing factor, in cotton response to drought stress.GhADF1 expression in cotton could be induced by PEG6000. GhADF1-RNAi transgenic cotton showed increased tolerance to drought stress during seed germination and seedling development as well as at the reproductive stage. In contrast, overexpression of GhADF1 led to a drought-sensitive phenotype in transgenic plants. GhADF1-RNAi plants produced an enlarged root system with longer primary roots,more lateral roots, increased root dry biomass, and increased cell size. In leaves of GhADF1-RNAi cotton,proline content and activities of reactive oxygen species-scavenging enzymes were increased following drought stress compared with those in wild type.GhADF1-RNAi lines showed higher water-use efficiency than the wild type,accompanied by reduced leaf stomatal density and conductance.GhADF1-RNAi cotton produced higher fiber yield in the field under both normal and drought conditions.Transcriptomic analyses identified 124 differentially expressed genes in leaves of GhADF1-RNAi lines compared with the wild type following drought treatment. Upregulated genes included those encoding transcription factors,protein kinases, heat shock proteins, and other proteins known to be involved in stress responses. We conclude that GhADF1 reduces the expression of abiotic stress-associated genes in cotton response to drought stress and may be a promising candidate gene for crop improvement by genetic manipulation.

1. Introduction

Drought is a major limiting factor in crop yield and quality [1].As a major fiber crop, upland cotton (Gossypium hirsutum) is often confronted with drought stress that severely impairs its growth,productivity, and quality [2]. It is thus desirable to improve its drought resistance.Although conventional breeding methods have been successfully used for increasing plant abiotic stress tolerance,they are cost- and labor-intensive and time-consuming [3]. To solve this problem, creating stress tolerant crops by genetic engineering may be a useful strategy [4,5].

To adapt and survive under drought stress, plants have developed complex mechanisms involving regulating the expression of a specific set of genes [4,5]. Some of them, such asAtLOS5,ABI3,AtABI5,EDT1/HDG11,OsSIZ1, andAREB/ABFs, have been identified and used as candidate genes for genetic engineering [6-10].However, most such genetic modifications have been tested inArabidopsis, not in crop plants [11]. Plant organs such as roots and leaves coordinate defense mechanisms in response to drought stress [5,12]. The morphological and physiological characteristics of leaves and roots influence the growth, development, and total yield of plants [13,14]. Measurement of root systems in crops under drought stress revealed a positive correlation of root diameter, depth and density with plant vigor [13]. Growing crops with high root density and length in medium and deep soil layers may preserve yield under drought stress [14]. Stomatal density was positively correlated with the drought resistance coefficient[5,15].It is thus desirable to identify stress-resistance genes in cotton for molecular breeding, and it is of great significance for the cultivation of drought-resistant cotton varieties to change structures of roots and leaves by transgenic technology.

The actin cytoskeleton in plants participates in diverse cellular processes including guard cell movement,cell expansion,division,motility, organelle trafficking, and signal transduction [16]. Actin filament dynamics is precisely controlled by specific factors such as actin-binding proteins (ABPs). As members of ABPs,actin-depolymerizing factors (ADFs) are crucial for modulating the balance between actin depolymerization and polymerization,which is required for normal cell growth [17,18]. ADF has been implicated in plant cell division, cell movement, apical growth,and other physiological processes [19-26]. Overexpression of tobaccoNtADF1in elongating pollen tubes reduced normal actin filaments and inhibited the growth of pollen tube [21]. The mossPhyscomitrella patenscontains a singlePpADFgene that is essential for tip growth [22].AtADF9is functionally expressed in apical meristems ofArabidopsis,and its mutation resulted in fewer lateral branches and early flowering [23]. Overexpression ofGhADF7inArabidopsisled to a reduction in viable pollen grains,so that transgenic plants were mostly male-sterile [24]. Overexpression ofAtADF1inArabidopsisled to the disappearance of thick actin cables in multiple cell types,resulting in irregular cellular morphogenesis and decreased growth of cells, tissues and organs, whereas downregulated expression ofAtADF1increased the formation of actin cables, delayed flowering, and promoted cell expansion and plant growth [25]. ReducedGhADF1expression increased cotton fiber length and strength [26].

Recently, ADF has also been implicated in response to biotic stresses.AtADF2,AtADF4,andTaADF7are involved in plant response to diverse pathogens[27-29].TaADF3positively regulates controls wheat resistance to abiotic stresses and acts as a negative regulator responding toPuccinia striiformisin a reactive oxygen species(ROS)-dependent manner[30].Moreover,ADFs also act in response to abiotic stresses [31-33].TaADFwas specifically induced by low temperature,and expression ofTaADFwas higher in cold-resistant than in cold-sensitive cultivars under cold treatment [31].OsADF3andAtADF5expression increased under salt, drought, and exogenous ABA stresses and increased plant tolerance to drought or osmotic stress [32,33]. Overexpression of theADF5gene inArabidopsisinduced stomatal closure by regulating actin remodeling in response to ABA and drought stress,while a loss-of-function mutation ofADF5conferred a drought-sensitive phenotype with increased leaf water loss of leaves, decreased survival rates under drought treatment,and delayed stomatal closure by affecting actin cytoskeleton remodeling due to altered F-actin-bundling activity[33].

Although it has been reported [31-33] thatADFgenes participate in response to abiotic stresses in some plant species, little is known about the role ofADFgenes in cotton response to abiotic stresses, in particular drought stress. The objective of the present study was to generateGhADF1-overexpressing and RNAi transgenic cotton and to evaluate the drought tolerance of these transgenic plants in the laboratory, greenhouse and field.

2. Materials and methods

2.1. Plant material and growth conditions

Cotton(Gossypium hirsutum,cv.Coker 312)seeds were sterilized with 75%(v/v)ethanol for 1 min and 30%(v/v)H2O2for 2 h,followed by washing 3-4 times with sterile water.The sterilized seeds were germinated on half-strength(1/2)Murashige and Skoog(MS)medium under 16 h light/8 h dark cycles at 28°C for 6 days.

For salt and phytohormone treatments, 5 day-old sterile early seedlings were treated with 1/2 MS liquid medium (as control)or supplemented with 200 mmol L-1NaCl, l0 μmol L-1gibberellic acid(GA3),indoleacetic acid(IAA),6-benzylaminopurine(6-BA)or abscisic acid (ABA) for 12 h. In other treatments, 5-day-old sterile seedlings were placed in 1/2 MS liquid medium at 4°C(simulating cold treatment) or at normal temperature (as control) for 12 h.

For osmotic treatment,5 day-old seedlings were cultured in 1/2 MS liquid medium with or without 15% polyethylene glycol (PEG)6000 (m/v) (simulating cold treatment) for 1, 3, 6, 12, and 24 h.Total RNA was isolated from the treated seedlings. Cotton plants grown in soil were also subjected to total RNA extraction.

2.2. Real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis

Total RNAs from several tissues were extracted using a Spectrum plant total RNA Kit(Sigma-Aldrich,http://www.sigmaaldrich.com) and then purified with a RNeasy kit (Qiagen, Hilden,Germany) as described previously [34]. First-strand synthesis of cDNAs was performed using M-MLV reverse transcriptase (Promega, Madison, WI, USA) following the manufacturer’s instructions. Expression ofGhADF1and other drought stress-associated genes was measured by qRT-PCR as previously described [34].GhUBI1gene (GenBank accession ID EU604080) was used as an internal control. All gene-specific primers used for qRT-PCR are listed in Table S1.

2.3. Construction of GhADF1 recombinant vector and genetic transformation

Construction methods for theGhADF1overexpression vector andGhADF1-RNAi vector refer to Qin et al. [35]. In detail, a 130-bp sequence ofGhADF1open reading frame (ORF) was used for encoding the inverted-repeat RNA. All gene-specific primers used for these constructs are listed in Table S1. The recombinant plasmid was transferred into hypocotyl explants of Coker 312 byAgrobacterium(strain LBA4404) -mediated transformation as described previously [26] Transgenic plants were selected with 1/2 MS-agar plates supplemented with 100 mg L-1kanamycin and then transferred to soil in an open glasshouse. Positive transgenic plants (generations T1-T3) were confirmed by PCR.

2.4. Drought-tolerance assay

For seed-germination assay,seeds of wild type and independentGhADF1-RNAi transgenic lines (L6 and L7) were sown on 10%PEG6000-saturated or sterilized water-saturated filter papers (as controls) in a plant growth incubator (16 h light/8 h dark cycles,28°C). Germination rate was calculated as described previously[38].

Drought-tolerance tests of cotton seedlings from transgenic and wild-type lines in the greenhouse were performed as described by Yu et al.[8].Drought tolerance of transgenic cotton in the field was evaluated as described by the same author.

Seedling growth status was recorded at days 10,15,18,20,and 30 and lateral roots were counted. The length of primary roots from early seedlings was recorded after drought stress for 35 days.Primary roots from 15-day-old plants and leaves from 40-day-old plants were sampled after drought stress for 15 days. A root and leaf surface imprint method was used as previously described[36]. Root cells and stomata were counted and their sizes were determined following Yu et al. [36].

All experiments were performed with three technical replications.

2.5.Measurement of photosynthetic efficiency,transpiration rate, and water use efficiency

Photosynthetic efficiency,transpiration rate,and water use efficiency(WUE)in leaves of both transgenic lines and wild type from 40-day-old cotton plants were determined as previously described[37].

2.6. Measurement of proline, chlorophyll, malondialdehyde (MDA)contents,peroxidase(POD)and catalase(CAT)activity,and electrolyte leakage assay

Leaves at several developmental stages collected from wild type and transgenic plants grown under drought stress were used for measurement of proline, chlorophyll, MDA content, POD and CAT activity, and electrolyte leakage. Proline and chlorophyll contents were measured according to Qin et al.[38].MDA content was determined as described [39]. POD activity was measured as described previously [40]. A 50-μL enzyme extract was added to a reaction mixture consisting of 1.85 mL 0.1 mol L-1acetic acid-natrium aceticum(HAC-NaAC)buffer(pH 5.0),0.25%guaiacol,and 0.1 mL 0.75%H2O2. CAT activity was assayed as previously described [41]. Electrolyte leakage was assayed as previously described[42].

2.7. Transcriptomic analysis

Samples were collected from wild-type and transgenic cotton plants under drought stress conditions in which water was withheld for 30 days. Total RNA of fourth true leaf tissues was extracted and purified following Qin et al.[35]The samples were used for constructing RNA-seq libraries, which were sequenced by Nanjing Personal Gene Technology Co.,Ltd.(Nanjing,Jiangsu,China,http://www.personalbio.cn). The facility/location where the sequencing was performed and read length and type were described previously[43].

Cutadapt (v2.7) software were used to filter the sequence data with the connector at the 3′end and reads with a mean quality score ＞Q20 to yield high-quality sequence for further functional enrichment analysis, including gene ontology (GO) and KEGG.The software and parameters used for GO, heat map, and KEGG analysis were as described by Zhang et al. [44]. Statistical analysis of RNA-seq was conducted using a combination of two fold changes between the means of biological replicates and false discovery rate below 0.05 (P＜0.05).

3. Results

3.1. GhADF1 was induced by osmotic stress

Transcription ofGhADF1was up-regulated at least threefold in early seedlings treated with PEG6000,whereas no differences were observed between seedlings under the other treatments and controls (Fig. 1A).

BecauseGhADF1expression was up-regulated by PEG6000 treatment, expression profiles ofGhADF1were investigated in 5-day-old cotton seedlings under PEG6000 treatment over several time courses.GhADF1expression was gradually increased in cotton seedlings for 1 to 12 h, but then decreased from 12 to 24 h after PEG6000 treatment(Fig.1B),suggesting thatGhADF1was induced by osmotic stress.

3.2. Suppression of GhADF1 in cotton increased drought tolerance during seed germination

TheCaMV35S promoter-drivenGhADF1overexpression and RNAi vectors were constructed and introduced into cotton(Gossypium hirsutum, cv. Coker 312) byAgrobacterium-mediated transformation method, respectively. Over 100 plants of 10 independentGhADF1-RNAi transgenic cotton lines(T0),and 30 plants of 4 independentGhADF1-overexpressing cotton lines (T0) were obtained and planted into soil to grow to maturation. Determination of the transgenic progeny (generations T1-T3) was conducted by kanamycin selection and PCR detection.To investigate the function ofGhADF1under drought stress, we selected six T3RNAi homozygous lines and four overexpressing homozygous lines whose expression levels were confirmed by quantitative RT-PCR(Figs. 2A, S1A). Overexpressing plants developed twisted leaves,shortened and clustered internodes, shrunken growing points,fewer fruiting branches, and rare bolls with few seeds and shorter fibers.They yielded insufficient seeds for field observation of agronomic traits and cotton fiber yields. We accordingly focused on detailed analysis ofGhADF1-RNAi transgenic cotton lines.

Seeds of wild type andGhADF1-RNAi transgenic lines 6 and 7(L6 and L7) were sown on 10% PEG6000-saturated or sterilized water-saturated filter papers (Fig. 2B). Seed germination rates ofGhADF1-RNAi lines and wild type under PEG6000 stress are shown(Fig. 2C). No difference in the germination of wild-type andGhADF1-RNAi lines on sterilized water-saturated filter papers was observed, and all seeds germinated almost completely (≤6 days,P≥0.35, Fig. 2C). However, in the presence of 10%PEG6000,seeds ofGhADF1-RNAi lines germinated earlier and faster than those of the wild type (Fig. 2B). After 6 days, around 60% ofGhADF1-RNAi seeds germinated, while only approximately 20% of wild-type seeds germinated (P≤1.95E-9, Fig. 2C).Thus, suppression ofGhADF1expression increased drought tolerance in germinated seeds.

Fig. 1. GhADF1 expression was induced by drought stress. (A) Quantitative RT-PCR analysis of GhADF1 expression in cotton under abiotic stress treatments. Total RNA was extracted from 5 day-old seedlings treated with or without 250 mmol L-1 NaCl, 4°C, 15%PEG6000, l0 μmol L-1 GA3, ABA, IAA, and 6-BA for 10 h, respectively. Control (CK)was untreated 5-day-old seedlings.(B)GhADF1 expression under PEG6000 treatment.Total RNA was extracted from 5-day-old seedlings with 15%PEG6000 treatment for 0(CK), 1, 3, 6, 12, and 24 h. Relative expression of GhADF1 in cotton is shown as percentage of GhUBI1 (cotton ubiquitin 1 gene) expression level. Values are means±SDs of independent triplicate assays.Independent t-tests revealed a significant(*,P ＜0.05)or very significant(**,P ＜0.01)difference in GhADF1 expression levels between PEG6000-treated seedlings and untreated controls.

Fig. 2. Assay of seed germination of independent T3 GhADF1-RNAi transgenic cotton lines under PEG6000 treatment. (A) GhADF1 expression in 5-day-old seedlings from independent T3 GhADF1-RNAi transgenic lines and wild-type plants by quantitative RT-PCR.GhUBI1 was used as an internal control.Values are means±SDs of independent triplicate assays. (B)Growth status of wild type and transgenic seedlings germinated in 10% PEG6000 for 2,4, and 6 days. (C) Germination rate after 6 days of germination.Mean values and SE (bar) of three independent experiments (n ＞50 seeds per line) are shown. Independent t-tests revealed a very significant (**, P ＜0.01) difference in germination rate between transgenic lines and wild type under PEG6000 treatment. WT, wild-type cotton (Coker 312). Ri-L1, -L4, -L6 (-6), -L7 (-7), -L9, and -L11, GhADF1-RNAi transgenic cotton lines 1, 4, 6, 7, 9, and 11.

3.3. Suppression of GhADF1 in cotton enhances drought tolerance of seedlings

To investigate the roles ofGhADF1in plant drought tolerance,GhADF1-RNAi cotton plants were grown in the greenhouse under normal irrigation conditions as controls. After seed germination,the seedlings of theGhADF1-RNAi lines (L6 and L7) and wild type were grown in the greenhouse without watering. No differences betweenGhADF1-RNAi lines and wild type were observed during seedling development under well-watered conditions (Fig. 3A).However, as drought stress continued, on days 25 and 35, the transgenic plants displayed delayed leaf-wilting symptoms and decreased wilting rates compared with the wild type. Compared to 20% wilting in the wild type, only 4% (P≤8.95E-5) wilting ofGhADF1-RNAi seedlings were observed after 25 days of withholding water and increased to 45%(P≤9.94E-7)wilting after 35 days,in contrast to 97% in wild-type seedlings (Fig. 3B). Dry biomass of theGhADF1-RNAi plants was increased by 1.52-1.78-fold under drought conditions(P≤2.04E-6,Fig.3C).Thus,GhADF1-RNAi lines showed increased drought tolerance relative to the wild type. In contrast,GhADF1-overexpressing seedlings showed a droughtsensitive phenotype (Fig. S1B). On days 15 and 18 after drought stress,GhADF1-overexpressing lines showed earlier leaf-wilting symptoms than the wild type.After drought treatment for 20 days and recovery for 2 days, nearly all the overexpressing transgenic plants (L2 and L4) died, whereas all wild-type plants remained healthy(Fig.S1B).These results suggested thatGhADF1is involved in plant response to drought stress.

3.4. GhADF1-RNAi cotton showed increased drought tolerance,improved agronomic traits, and increased yield in the field

To assay drought tolerance at the reproductive phase, water was withheld for 40 days from 35-day-old well-watered greenhouse-grown plants. Wild-type plants showed an apparent drought-sensitive phenotype with severely wilted leaves, whereasGhADF1-RNAi lines (L6 and L7) displayed a relatively normal growth phenotype(Fig.3D).These results showed that suppression ofGhADF1increased cotton plant tolerance to drought stress at reproductive stage.

Under a rain shelter and under natural drought conditions,two transgenic lines (L6 and L7) showed clear growth advantages and increased drought tolerance compared with the wild type(Fig. 3E, F).

Fig.3. Phenotypes of GhADF1-RNAi transgenic cotton under drought stress.(A)Phenotypes of GhADF1-RNAi transgenic cotton at seedling stage in greenhouse under drought stress.Seeds were sown in soil pots with sufficient water,and seedlings after germination were grown for 25,30 and 35 days with(control,left panel)or without(drought,right panel) irrigation. Drought for 35 days and then recovery for 2 days. Scale bars, 10 cm. (B) Statistical analysis of plant wilting rate. Wild-type and transgenic seedlings after germination were grown for 25, 30 and 35 days without (drought, right panel, A) irrigation in greenhouse. (C) Shoot dry weight of plants treated with drought for 35 days in the greenhouse.Values are means±SD of 60 plants(**,P ＜0.01).(D)Phenotypes of GhADF1-RNAi transgenic plants at reproductive stage in the greenhouse under drought stress.Plants were grown normally in planting bags for 35 days,and watering was then withheld for 40 days.Scale bars,15 cm.(E)Phenotypes of transgenic lines and wild type in the field under a rainproof shed under drought stress.Plants grew normally for one month under well-watered conditions and then were treated with drought stress.Photographs were taken after drought stress treatment for 65 days.(F)Phenotypes of GhADF1-RNAi transgenic lines in the field in Yuncheng,Shanxi province,China.Plants were photographed under natural drought conditions for 80 days. WT, wild type; Ri-6, -7, GhADF1-RNAi transgenic lines 6 and 7.

In addition to better performance under drought conditions,the agronomic performance and cotton fiber yields in the field under natural drought and well-watered conditions were compared among theGhADF1-RNAi lines and wild type.As shown in Table 1,under normal irrigation conditions, in comparison with wild-type plants, theGhADF1-RNAi lines (L6 and L7) displayed sharply increased plant height (P≤0.03), boll number (P≤0.03), fruit branch number (P≤0.02), boll shedding (P≤0.04), boll fresh weight (P≤0.03), and seed fiber yield per plant (increased by 18.24% and 13.48%, respectively,P≤0.03). Under drought conditions, the seed fiber yields per plant from theGhADF1-RNAi lines increased by respectively 24.35%and 21%relative to the wild type(P≤0.02). Plant height (P≤0.03), fruit branch number (P≤0.03),boll fresh weight (P≤0.02, boll number (P≤0.02) and boll shedding per plant(P≤0.03)were also increased inGhADF1-RNAi lines relative to the wild type(Table 1). Thus, suppression ofGhADF1in cotton increased field drought resistance as well as agronomic traits and fiber yield under both well-watered and drought conditions.

TheGhADF1-RNAi plants also developed enlarged root systems(Fig. 4A) with longer primary roots (Fig. 4B, E;P≤3.32E-7) and more lateral roots (Fig. 4C,P≤0.009), increased root dry biomass(Fig. 5F,P≤5.25E-7), and increased cell size of roots, in both length and width (Fig. 4H,P≤4.92E-5). Consequently, the mean numbers of root cells per view area in transgenic lines 6 and 7 were lower than that of the wild type(Fig.4G,P≤2.16E-12).This welldeveloped root system would be beneficial for plant growth and drought resistance.

3.5. GhADF1-RNAi cotton showed reduced leaf stomatal density,increased water-use efficiency, increased photosynthetic rate, and decreased transpiration rate

Stomatal sizes and numbers in leaves of bothGhADF1-RNAi lines and wild type were assayed. The lengths of stomatal guard cells increased (P≤4.73E-36) while stomatal width decreased(P≤2.14E-28) inGhADF1-RNAi cotton plants (Fig. 5A-B). The mean stomatal densities of L6 and L7 lines were respectively 26.85% and 22.08% lower than those of the wild type(P≤1.03E-17, Fig. 5C). The reduction of stomatal density from the transgenic plants was due mainly to the enlarged size of epidermal cells relative to the wild type (P≤1.92E-10, Fig. 5D). TheGhADF1-RNAi lines showed higher water-use efficiency (WUE,P≤1.22E-7,Fig.5F),increased photosynthetic rates(P≤4.04E-6,Fig.5G),reduced stomatal conductance(P≤1.36E-6,Fig.5E),and decreased transpiration rates (P≤0.008, Fig. 5H), relative to the wild type.

3.6. GhADF1-RNAi cotton showed higher proline content and ROSscavenging enzyme activities but lower malondialdehyde (MDA)content under drought stress conditions

We compared several physiological parameters between wild type andGhADF1-RNAi lines in response to drought stress. In leaves of the transgenic cotton, contents of proline and MDA and activities of ROS-scavenging enzymes (including POD and CAT)were gradually increased over time under drought conditions(Fig. 6A, C-E). Compared with the control, MDA content was decreased whereas proline content was increased in theGhADF1-RNAi lines under drought treatment for 4 days,and this difference became progressively more apparent after 6, 8, and 10 days of water deprivation (Fig. 6A, C). CAT and POD activities increased gradually over 4 to 6 days under drought treatment (Fig. 6D, E).The electrolyte leakage and total chlorophyll content of bothGhADF1-RNAi transgenic lines and wild type under drought treatment are shown in Fig. 6F. Electrolyte leakage of transgenic lines and wild type increased gradually during the drought treatment(Fig. 6B). However, electrolyte leakage was markedly lower inGhADF1-RNAi lines than in the wild type under drought stress treatment. Under drought stress conditions, the total chlorophyll content of both transgenic lines and wild type gradually decreased over time, but much less in the transgenic plants than in the wild type (Fig. 6F). Thus, the transgenic plants showed more tolerance to drought stress than the wild type, owing to greater protection from oxidative damage under drought stress.

3.7.GhADF1 affects expressions of many drought-responsive genes and transcription factors

To further understand the regulating effect ofGhADF1in cotton response to drought stress,transcriptomic analysis was conducted to identify differentially expressed genes (DEGs) in theGhADF1-RNAi lines with or without irrigation. A Venn diagram (Fig. 7A)shows the DEGs in the wild type,GhADF1-RNAi lines 6 and 7(Control-vs-Drought).In total, 5878 DEGs,including 2965 upregulated genes and 2913 downregulated genes, were identified in the wild type following drought treatment (≥2-fold changes,FDR ＜0.01, Table S2). A total of 1732 DEGs,including 624 upregulated and 1108 downregulated genes, were identified inGhADF1-RNAi lines 6 and 7 following drought treatment(Fig.7A;Appendix S1; Table S2). Under normal watering conditions, 435 DEGs (including 225 upregulated and 210 downregulated genes) were identified inGhADF1-RNAi lines and the wild type. Under drought treatment,124 DEGs(including 64 upregulated and 60 downregulated genes) were identified in leaves ofGhADF1-RNAi lines compared with those of the wild type (Fig. 7A; Table S2; Appendixes S1 and S2). Functional classification of these 124 DEGs using GO enrichment revealed that the differentially regulated transcripts participate in diverse molecular functions, cellular compartments,and biological processes(Fig.7B).The transcript levels of these 124 DEGs in wild type, andGhADF1-RNAi lines 6 and 7, under drought stress conditions, are shown in heat maps (Fig. 7C).

Among 64 upregulated genes of the 124 DEGs, major classes of functional genes included a 14-3-3 protein(GH_A01G0120,associated with stomatal conductance [45]), a calcium-binding protein(GH_A05G2383), a calmodulin protein (GH_D02G0933), an ascorbate peroxidase (GH_D01G1910, associated with antioxidation[46,47]), two leucine-rich repeat receptor protein kinases(GH_D02G0263 and GH_D09G0554, associated with regulation of the antioxidative system [48]) two Ser/Thr-protein kinases(GH_D01G0255 and GH_D01G0463), two MAP kinases(GH_D13G1683 and GH_D13G1684, associated with regulating stomatal response and root growth [49,50]), and six transcription factors, including two AP2/ERF transcription factors(GH_D12G1116 and GH_D10G1928, associated with root development [51,52]), two zinc finger proteins (GH_A08G2575 and GH_A02G0151),a MADS-box protein(GH_D11G0898),a NAC transcription factor (GH_A02G2027,associated with root development[53]) and 15 heat shock proteins (GH_A07G1119, GH_D02G0370,GH_D02G0617, GH_D08G1208, GH_D03G1817, GH_A13G1429,etc. associated with scavenging reactive oxygen species (ROS)[2,54,55]), and these genes were strongly up-regulated in theGhADF1-RNAi lines under drought stress (Fig. 7D; Table S3).

Table 1Agronomic traits and fiber yields of transgenic lines in the field under natural drought and well-watered conditions.

Fig.4. Determination and analysis of root system of GhADF1-RNAi transgenic cotton.(A)The phenotype of wild-type and GhADF1-RNAi transgenic seedlings.Growth status of wild-type and GhADF1-RNAi transgenic seedlings grown under normal conditions for 10,15,20,and 30 days.Scale bars,3 cm.(B)Roots of the GhADF1-RNAi transgenic and wild-type cotton plants treated with drought for 35 days.Scale bars,3 cm.(C)Statistical analysis of the number of lateral roots from 10,15,20,and 30-day-old seedlings of wild-type and GhADF1-RNAi transgenic lines.(D)Micrographs of root cells from the wild type and a GhADF1-RNAi transgenic plant.Scale bars,50 μm.(E-F)Statistical analysis of primary root length and root dry weight of plants treated with drought for 35 days. (G-H) Mean number and size of root cells from the wild type and GhADF1-RNAi transgenic plant. WT, wild type; Ri-6, -7, GhADF1-RNAi transgenic lines 6 and 7. Values are means±SD of three replicates (**, P ＜0.01).

To confirm the reliability of the drought-responsive gene expression profiles for DEGs,the expressions of 18 representatively upregulated genes in theGhADF1-RNAi lines were verified by qRTPCR. The expression levels of the selected drought-responsive genes measured by qRT-PCR correlated with RNA-seq transcript abundances (Fig. 7E), reflecting the accuracy of transcriptomic analysis.

Fig.5. Reduced stomatal density, stomatal conductance, and transpiration rate,enlarged epidermal cell size,and increased water use efficiency(WUE),and photosynthetic rate in GhADF1-RNAi transgenic cotton.(A)Micrographs of adaxial epidermal cells from wild-type and GhADF1-RNAi transgenic plants.Scale bars,50 μm.(B)Stomatal size of wild-type and GhADF1-RNAi transgenic plants. Stomata were counted and measured by microscope. Mean values and SE (bar) are shown from three independent experiments(n=50 stomata per line).(C,D)Stomatal density and epidermal cell number of the wild-type and GhADF1-RNAi transgenic plants(n=50 images).(E-F)Stomatal conductance(E)and WUE(F)of wild-type and GhADF1-RNAi transgenic plants.(G-H)Photosynthetic rate(Pn,G)and transpiration rate(Tr,H)of wild-type and GhADF1-RNAi transgenic plants. (n=10 plants per line). WT, wild type; Ri-6, -7, GhADF1-RNAi transgenic lines 6 and 7. Values are means±SD (**, P ＜0.01).

4. Discussion

Crop productivity and quality are generally the ultimate aim of breeding. However, drought greatly affects crop growth and yield.Therefore,it is very urgent to develop crop cultivars with increased drought tolerance for improving yields. Although transgenic technology has become a rapid strategy for crop breeding, ectopic expression of many stress-response genes often gives rise to abnormal development, leading to yield loss [56,57]. Overexpression ofAtHDG11in cotton has shown great potential to improve agricultural productivity under drought or water-shortage conditions [8]. In the present study, similar results were found inGhADF1-RNAi transgenic plants, which showed no apparent negative or abnormal growth or development. On the contrary, the RNAi plants showed enlarged root system, increased leaf stomatal and epidermal cell size,and reduced stomatal density.Suppression ofGhADF1expression in cotton conferred drought resistance,improved agronomic traits, and increased fiber yield under both well-watered and natural drought conditions in field tests.GhADF1-RNAi transgenic cotton grown in the field showed an apparent growth advantage with increased height, boll number,fruit branch number, and fiber yield under both normal and drought conditions compared with the wild type(Table 1).In particular,the seed yield ofGhADF1-RNAi lines 6 and 7 was increased by 24.35% and 21% (*,P＜0.05) under drought conditions and by 18.24% and 13.48% (*,P＜0.05), respectively, under normal conditions.Compared with normal well-watered conditions,the relative cotton fiber yield of the transgenic lines increased more dramatically under drought conditions in the field (Table 1). The results suggest that downregulatingGhADF1benefits cotton under drought conditions more than under normal conditions, and thus our work offers a promising candidate geneGhADF1to overcome a severe problem facing the production of cotton.GhADF1as a potential target gene may be used for increasing crop drought tolerance in areas where water is a limiting factor for agricultural productivity, and the desirable agronomic performance observed in theGhADF1-RNAi cotton plants is beneficial to crops under certain water-deficit conditions.

In previous studies [8], increases in drought tolerance associated withAtHDG11in cotton have been attributed to developmental alterations in root system and stomatal density, increased photosynthesis and WUE, and increased tolerance to oxidative stress. Similarly, our results suggest that the drought tolerance ofGhADF1-silenced cotton is contributed by changes at both the morphological and physiological levels. First, the transgenic plants have enlarged root systems, maximizing water absorption and nutrient intake under drought stress. One possible reason for the changed root system of the transgenic plant is thatGhADF1affects the expression of cell wall-associated genes, which are known to alter root system architecture [58]. Second, the decrease in stomatal density, conductance, and transpiration rate in the transgenic cotton likely resulted in higher WUE and better water retention.Furthermore, theGhADF1-RNAi transgenic plants were better protected from oxidative and osmotic damage by reduced MDA levels and increased proline content and POD and CAT activities (Fig. 5).

Fig. 6. Malondialdehyde (MDA) and electrolyte leakage were decreased and proline and chlorophyll contents and activities of reactive oxygen species-scavenging enzymes(POD, CAT) were increased in GhADF1-RNAi transgenic cotton after drought stress compared with wild type. MDA content (A), electrolyte leakage (B), proline contents (C),POD(D)and CAT(E)activities and chlorophyll contents(F)of transgenic and wild-type plants after drought stress.Values are means±SD of three replicates(**,P ＜0.01).20-day-old plants(approximately 20 plants of each line)were treated for 2,4,6,8,or 10 days without irrigation.Mean values and SE(bar)from three independent experiments(n=20 fully expanded leaves per each line)are shown.Independent t-tests for comparison of means revealed highly significant difference between wild type and transgenic plants (**, P ＜0.01). WT, wild type. Ri-6, -7, GhADF1-RNAi transgenic lines 6 and 7. Assays were repeated three times along with three independent repetitions of the biological experiments.

Suppression ofAtADF1resulted in development of elongated hypocotyl cells inArabidopsis[25],and down-regulation ofGhADF1increased cotton fiber length [26]. In agreement with those observations,we found thatGhADF1-RNAi transgenic cotton plants had larger leaves, longer primary roots, more lateral roots, and increased height compared to the wild type (Fig. 4; Table 1), and thatGhADF1negatively influences cotton cell development. However,the roles of ADFs in stress tolerance vary greatly.Overexpression of riceOsADF3inArabidopsisincreased drought-stress tolerance by increasing germination rate of seeds, primary root length,and survival rate of seedlings[31].Tholl et al.[59]reported thatAtADF1andAtADF9regulate actin dynamics in a reverse way and compete with each other,suggesting that plants have evolved neofunctionalized ADFs to modulate actin dynamics synergistically. The loss of AtADF4 function led to stomatal closure in response to drought stress [20], whereas theAtADF5loss-offunction mutation increased water loss,reduced the vigor of plants under drought stress, and delayed stomatal closure by affecting actin cytoskeleton remodeling. It was thus speculated [33] that AtADF5 functions in concert with AtADF4 in response to drought stress by controlling stomatal movement. In contrast, we showed that down-regulation ofGhADF1in cotton increased drought tolerance, delayed leaf wilting, and increased WUE.

Fig. 7. Major classes of functional genes differentially upregulated in GhADF1-RNAi transgenic cotton under drought stress. (A) Venn diagram of unigenes, identified as differentially expressed in wild-type and transgenic lines(Control-vs-Drought).WT,wild type;Ri-6,-7,GhADF1 RNAi transgenic lines 6 and 7;M,mock,under normal growth conditions; D, under drought treatment. (B)Gene ontology(GO)functional classification of 124 differentially expressed genes(DEGs). BP,biological process;MF, molecular function; CC, cellular component. (C) Transcript levels of 124 DEGs in heat maps. Columns and rows in heat maps represent samples and DEGs, respectively. Red color indicates genes with high expression levels and green color indicates genes with low expression levels. WT, wild type; Ri-6, -7, GhADF1 RNAi transgenic lines 6 and 7; M,mock, under normal growth conditions; D, under drought treatment. (D) Major classes of the upregulated functional genes in transgenic lines from RNA-Seq data. (E)Quantitative RT-PCR analysis of the 18 selected major genes.Gh14-3-3-L(14-3-3 protein,GH_A01G0120),GhAPX1(ascorbate peroxidase,GH_D01G1910),GhLRRPK1(leucinerich repeat protein kinase 1, GH_D09G0554), GhEMS1 (leucine-rich repeat receptor protein kinase EMS1, GH_D02G0263), GhMAPK683, GhMAPK684 (MAP kinase,GH_D13G1683 and GH_D13G1684), GhRBK2i (serine/threonine-protein kinase RBK2 isoform, GH_D01G0255), GH_D01G0463 (serine/threonine-protein kinase), GhCAMP5(calmodulin-like protein 5, GH_D02G0933), GhCML44 (calcium-binding protein CML44, GH_A05G2383), GhERF2, GhERF038-L (AP2/ERF transcription factor, GH_D10G1928 and GH_D12G1116), GhZFN-L, GhC1H1ZN (zinc finger protein, GH_A08G2575 and GH_A02G0151), GhSOC1-L (MADS-box protein SOC1-like, GH_D11G0898), GhB-2a (heat stress transcription factor B-2a,GH_A10G0269),GhNAC2-L(NAC domain-containing protein 2-like,GH_A02G2027),GhHSP8-L(heat shock protein 8,GH_A09G2360).Valuesshown are means±SDs of three biological replicates. WT, wild type; Ri-6, -7, GhADF1-RNAi transgenic lines 6 and 7; Student’s t-tests revealed significant differences (**,P ＜0.01) among transcript levels.

Mounting evidence [21,32,33,60,61] indicates that rearrangement of the cytoskeleton alters signal cascades in response to drought,salt,cold,osmotic pressure,and pathogen attacks.Several genes involved in drought response were upregulated inOsADF3-overexpressingArabidopsisunder drought stress [32]. TaADF acts as a substrate for a wheat kinase, a component in low temperature-induced signaling. TaADF3 may be involved in response to cold tolerance via interaction with other proteins to regulate cell cytoskeleton dynamics[30].Interaction ofArabidopsis14-3-3 protein with AtADF1 possibly inhibits AtADF1 phosphorylation,thereby affecting F-actin stability and dynamics[61].AtADF5 regulates stomatal closure via the ABA signaling pathway under the control of ABF/AREB-family transcription factors [33]. In the present study, transcriptomic analysis revealed increased expression levels ofGh14-3-3-L,GhLRRPK1,GhEMS1,GhMAPK683,GhMAPK684,GhRBK2i,andGH_D01G0463in transgenic plants after drought treatment. It is thus tempting to suggest that GhADF1 is modulated at the post-translational level.

Arabidopsis14-3-3 protein GF14λ overexpressed in cotton increased drought-stress tolerance, and stomatal conductance may have been the principal factor in the observed higher photosynthetic rates under water-deficit conditions[45].Overexpression ofGhMKK3inNicotiana benthamianaincreased plant tolerance to drought stress. In contrast,GhMKK3-silenced transgenic cotton showed the opposite phenotype [50]. We accordingly speculate that GhMAPK683 and GhMAPK684 phosphorylate GhADF1 and that Gh14-3-3-L regulates the phosphorylation of GhMAPK683 and GhMAPK684 to GhADF1 and participates in drought stress response. The upregulation of stress-related genesGhAPX1,GhCAMP5,GhCML44, andGhHSPsinGhADF1-RNAi lines suggests that GhADF1 acts as a negative regulator of abiotic stressresponse genes in response to drought stress. Our RNA-seq and qRT-PCR analyses showed differential regulation of various families of transcription factors, includingGhERF2,GhERF038-L,GhZFN-L,GhC1H1ZN,GhSOC1-L,GhB-2a, andGhNAC2-L, inGhADF1-RNAi transgenic lines, although the position of GhADF1 in the transcriptional network remains to be verified. RiceSNAC1is predominantly induced in guard cells by drought, andSNAC1overexpression in cotton increased tolerance to drought by enhancing root-system development and reducing transpiration rates [53].Likewise,GhADF1-RNAi transgenic cotton displayed increased tolerance to drought stress and showed an enlarged root system and reduced decreased transpiration rates relative to the wild type.In view of these findings, we hypothesize that down-regulated expression ofGhADF1regulated byGhNAC2-Lpromotes root development and reduces transpiration rate, thus improving drought tolerance of cotton.

In summary,down-regulation ofGhADF1increased drought tolerance in cotton.GhADF1-RNAi transgenic plants showed multiple characteristics associated with drought tolerance, such as welldeveloped root system and increased WUE and photosynthetic rate, as well as increased activities of CAT and POD and increased concentrations of chlorophyll and proline, but decreased stomatal density, transpiration rate, and MDA level.GhADF1-RNAi transgenic cotton showed increased drought tolerance, improved agronomic traits, and cotton fiber yield in the field. We identified potential drought-responsive genes and transcription factors associated with GhADF1.GhADF1-RNAi transgenic cotton has agricultural potential, and is also an excellent system for studying actin dynamics and elucidating molecular and cellular mechanisms of response to drought stress.Whatever the mechanism,the desirable agronomic performance observed inGhADF1-RNAi cotton plants will be beneficial to crops on marginal land under certain waterdeficit conditions.

CRediT authorship contribution statement

Lixia Qin:Conceptualization, Methodology, Funding acquisition, Writing - original draft, Writing - review & editing.Huanyang Zhang:Data curation, Formal analysis.Jing Li:Data curation, Formal analysis.Yonghong Zhu:Investigation,Resources, Validation.Gaili Jiao:Investigation, Resources, Validation, Writing - review & editing.Chuangyun Wang:Conceptualization, Methodology, Project administration, Supervision.Shenjie Wu:Conceptualization,Methodology,Funding acquisition,Project administration, Supervision, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank Prof.Ling Yuan(University of Kentucky),Prof.Chengbin Xiang and Prof.Ge Shan(University of Science and Technology of China) for critically reading and revising the manuscript. This work was supported by the National Natural Science Foundation of China (31601350), the Project of Transgenic Research from the Ministry of Science and Technology of China (2016ZX08005-004-007), the Fundamental Research Project of Shanxi Province(20210302123381) and the Science and Technology Innovation Project of Higher Education Institutions of Shanxi Province(2021L115).

Appendix A. Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.12.012.