Wenqing Zho, Horn Dong, Rizwn Zhoor,f, Zhiguo Zhou,b, John L. Snier,Yinglong Chen, Kmbot H.M.Siique, Youhu Wng,b,*

aKey Laboratory of Crop Ecophysiology and Management of Ministry of Agriculture,Nanjing Agricultural University,Nanjing 210095,Jiangsu,China

bJiangsu Collaborative Innovation Center for Modern Crop Production(JCIC-MCP),Nanjing Agricultural University,Nanjing 210095,Jiangsu,China

cThe UWA Institute of Agriculture,The University of Western Australia,Perth,WA 6001,Australia

dThe Department of Crop and Soil Sciences,University of Georgia,Tifton,GA 31794,USA

eState Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau,Northwest A&F University,Yangling 712100,Shannxi,China

fUniversity of the Agriculture,Faisalabad Sub-campus Depalpur,Okara,Pakistan

Keywords:Cotton (Gossypium hirsutum)Fiber length Drought stress Potassium application Osmolyte

ABSTRACT Fiber length of cotton(Gossypium hirsutum L.)decreases under drought stress,potassium(K)could diminish the decreased caused by drought, but the mechanism associated with this alleviation effect is not clear. We evaluated the effect of K on fiber elongation using two cotton cultivars, Simian 3 and Siza 3, grown in well-watered and drought-stressed conditions. Potassium fertilizer (K2O) was applied 0, 150, or 300 kg ha-1 in each growing condition. Drought stress reduced the final fiber length due to a decline in the maximum rate of rapid elongation (Vmax, mm day-1). The application of K alleviated the droughtinduced fiber length reduction by increasing Vmax. At 10 and 15 days post-anthesis (DPA),drought significantly reduced osmotic potential(OP)and increased K+and malate contents at all K rates, relative to well-watered conditions, which was associated with increased activities of phosphoenolpyruvate carboxylase(PEPC),V-ATPase, PPase,and PM H+-ATPase in cotton fiber. However, the relative contribution of K+ and malate to OP declined under drought in comparison with well-watered condition. Compared with control without K, K application decreased OP and increased the accumulation of osmolytes (K+, malate and soluble sugar) as well as the activities of related enzymes in fiber irrespective of water treatments. Moreover, K application increased osmotic adjustment during drought, and improved the contribution of K+ and malate to OP, especially under drought stress. This study showed that drought decreased fiber length by reducing Vmax, and K application ameliorates the decline in fiber elongation due to drought by enhancing osmolytes accumulation and their contribution to OP in fiber cells.

1. Introduction

Drought is a major restrictive factor for agricultural production worldwide. It is predictable that episodic summer drought events will increase in frequency and severity due to global climate change [1]. Drought may inhibit crop productivity especially under hot conditions. Cotton(Gossypium hirsutum L.)is one of the most important industrial crops of hot areas that provide raw fiber for the textile industry [2]. Fiber yield is essential for cotton growers to maximize profitability while fiber quality is a key trait for textile industry because it has a significant effect on yarn quality [3]; therefore, fiber quality determines the economic value per unit of fiber weight.Of the fiber quality parameters,fiber length is the most important determination of yarn quality and the marketability of textile properties [4]. Hence,increasing cotton fiber length is an essential goal for breeders[5], but is negatively affected under low-yielding conditions[6]. The predicted increase in the severity and frequency of yield-limiting drought events will have negative impacts on growers and fiber processors,especially in production regions such as the Yangtze Valley in China where it may be more affected by climate change than other cotton production regions[7].

Cotton fiber is an elongated unicellular trichome of the seed, and its development occurs in four stages: initiation,elongation, secondary cell wall deposition, and maturation[8]. Cotton fiber length is determined during the elongation stage when the fiber cell elongates rapidly from 0 to 21-26 days post-anthesis (DPA) to reach its final length [9].Previous studies have shown that two inner factors-the force and duration of cell turgor-regulate fiber elongation [10-12].In the developing fiber, cell turgor is largely influenced by an influx of water driven by osmolyte accumulation, where soluble sugars, K+, and malate account for 80% of the total osmotic potential (OP) [10,13]. Soluble sugars and K+are imported to the epidermal layer of the seed via phloem in the cotton seed coat. Active transport and accumulation of ions in the vacuole occurs through two types of electrogenic proton-translocating pumps,one of which hydrolyzes ATP(VATPase) and the other use PPi (PPase) as the primary energy sources [10]. In this process, plasma membrane (PM) H+-ATPase, structurally different from V-ATPase, generates the membrane potential to move ions across the PM and into the epidermal cell. Malate is synthesized locally by re-fixing CO2in the cytoplasm of the cotton fiber cell via phosphoenolpyruvate carboxylase (PEPC) activity. The osmolytes and the activities of the abovementioned enzymes have been correlated with fiber elongation rate and final fiber length [14].However,these factors might be influenced by environmental stresses,especially by drought[15].

Drought stress shortly after flowering and during fiber elongation could reduce final fiber length by directly inhibiting the mechanical and physiological processes of cell expansion[16-18].A recent study demonstrated that glucose,fructose, sucrose, and starch concentrations declined in water-stressed fibers [19]. Conversely, sucrose and K+contents increased in cotton leaves under drought stress due to osmotic adjustment(OA)[20,21].PEPC activity in the leaves of tobacco [22] and rice [23] and H+-ATPase activity in Crassula argentea leaves [24] increased in plants under water scarcity.However, the relationship between osmolyte accumulation and cotton fiber elongation under drought stress remained unknown. It is likely that drought stress decrease final fiber length by altering osmolyte concentrations in cotton fiber cells[25].

Potassium(K),the most abundant cation in plants,plays a critical role in OA.When plant water potential declines during drought, K can lead plants to accumulate solutes, lower OP,and increase water inflow to maintain turgor pressure[26].As drought restrains K uptake from the soil, the absorption of sufficient K before drought helps plants to maintain growth during a water-deficit period[27]. In drought-stressed cotton,optimized K application could promote OA capacity and mitigate the decline of seed cotton yield and fiber length[18,28,29]. Augmented K availability could increase cotton fiber K+content and the activities of V-ATPase,PPase,PM H+-ATPase,and PEPC,which subsequently increase the maximal fiber elongation rate (Vmax) and final fiber length [30].Drought-stressed cotton plants supplied with foliar K (1%)produced longer cotton fibers than non-K-treated plants [31].However,few reports have addressed the interactive effects of drought stress and K application regarding osmolyte accumulation, enzyme activity, and fiber elongation. Thus, we investigated the role of K in regulating the enzymes and osmolytes involved in fiber cell elongation in cotton plants exposed to drought.Our objectives were to clearify,1)whether drought stress restrains fiber cell elongation by limiting osmolyte accumulation and key enzymes’ activities involved in fiber cell elongation, and 2) whether elevated K levels ameliorate the negative effect of water deficit on fiber length by promoting the accumulation of compatible solutes, increasing fiber cell elongation rate and duration.

2. Materials and methods

2.1. Experimental set up and weather data

A pot experiment was conducted in 2015 and 2016 in a rainout shelter at Pailou Experimental Station(118°50′E,32°02′N,at an elevation of 20 m), Nanjing Agriculture University, Jiangsu province,China.Seeds of two cotton cultivars,Simian 3(low-K-tolerant) and Siza 3 (low-K-sensitive) [32], were sown in a nursery and transplanted at the three-leaf stage into pots(37 cm diameter,32 cm high).Each pot was filled with 25 kg of soil collected from the 0-30 cm soil layer at the experimental station, with each pot considered a replicate. The soil was a clay, mixed, thermic, typical Alfisols (Udalfs; FAO Luvisol),containing 16.5 g kg-1organic matter,1.1 g kg-1total nitrogen(N),70.8 mg kg-1available N,23.6 mg kg-1available phosphorus(P),and 97.3 mg kg-1exchangeable potassium.

The experiment was conducted as a completely randomized design with two cotton cultivars, three K levels (0 [K0],150[K1],and 300[K2],kg ha-1K2O),and two water treatments(WT),well-watered(WW),and short-term drought stress(DS).All pots were maintained at approximately 75% ± 5% relative soil water content (RSWC) before the water treatments commenced. When flowering had initiated on the fruiting branches of the 6th main-stem nodes of cotton plants,irrigation was withheld on half the pots for DS plants until the RSWC reached 40% ± 5% and leaf water potential was below -2.0 MPa, which was a duration of appoximately eight days,before re-watering to 75% ± 5%RSWC[29](Fig.1).For the K treatments, 40% of the potassium sulfate (K2SO4) dose was applied before transplanting and the remaining 60% at first flower. All of the phosphorous was applied before transplanting (120 kg ha-1P2O5), while 40% of the 240 kg ha--1N was applied before transplanting and 60% at flowering.During the drought treatment, soil moisture in all the pots was determined using the method described in Liu et al. [33];and maintained daily by collecting soil samples between 18:00 and 19:00 h, using an auger (2 cm diameter), and comparing the fresh and dry weights of the samples. Soil water content was expressed as g water g-1dried soil.

Weather data were collected from the National Meteorological Information Center (Nanjing Weather Station) in 2015 and 2016. Daily temperature parameters and relative humidity during the drought treatment in 2015 and 2016 are presented in Fig. 2. During the drought treatment in 2015 and 2016, mean daily temperatures were 28.09 °C and 32.99 °C, mean daily maximum temperatures were 31.54 °C and 37.24 °C, mean daily minimum temperatures were 25.44 °C and 28.58 °C, and mean relative humidities were 82.00% and 62.63%, respectively, indicating cooler and wetter conditions in 2015 than 2016.

2.2. Plant sampling

Flowers on the 6-8th fruiting branches were marked with plastic tags noting the flowering date.During 10-30 DPA,three to five tagged bolls of the same age were collected between 08:30 AM and 09:30 AM every five days. Samples were placed on ice and immediately transferred to the laboratory. Cotton seeds from one locule of each boll were stored for fiber length and OP measurements.Fibers from the remaining portions of the cotton bolls were removed with a scalpel and divided into two parts.One part was frozen in liquid nitrogen and stored at-40 °C for the enzyme activity assays,and the other was dried at 60 °C for 30 min and then 40 °C to constant weight for the osmolyte content assay. At maturation, tagged bolls from each treatment were harvested and ginned individually.

2.3. Measurements of fiber length and OP

Fiber length of cotton bolls collected from 10 to 25 DPA was evaluated using the water stream method of Thaker et al.[34],and those collected at 30 DPA and maturation were measured with a photoelectric stapler (Y-146 Photoelectric Stapler,Taicang Electron Apparatus Co., Ltd., Taicang, Jiangsu,China) on dried/mature fibers [35]. Changes in fiber length with DPA were quantified using a logistic function. The two main factors determining final fiber length-Vmaxand T(cotton fiber rapid elongation duration)-were calculated using the methods described in Zhao et al. [36] and Chen et al.[35].

Fig.1-Temporal trend in soil relative water content during the drought treatment(10 days from the onset of flowering on the fruiting branches of the 6-7th main-stem node)in a typical Alfisol(Udalfs;FAO Luvisol)soil in Nanjing in 2015 and 2016.K0,K1,and K2 are 0,150,and 300 kg K2O ha-1,respectively;WW,well-watered treatment;DS,drought treatment.Values are means of three replicates.Vertical bars represent standard errors.

Fig.2- Daily weather data at the experimental station in Nanjing during the experiment in 2015 and 2016.All data were collected from the National Meteorological Information Center(Nanjing Weather Station)located about 6 km from the experimental site.

Fiber OP was recorded using a Vapor Pressure Osmometer(Vapro 5600,WESCOR Co.,Ltd.,USA).The cell sap of fibers was obtained by extrusion and 10 μL of sap was used to measure OP. For each K treatments, OA was calculated by subtracting the OP recorded in DS plants from OP recorded in WW plants.The contributions of osmolyte(K+,malate,and soluble sugars)to OP and OA were then determined using the equation mentioned by Hessini et al. [37].

2.4. Measurement of osmolyte (K+, soluble sugars, malate)

Dried fiber samples were digested in H2SO4-H2O2[38]. The K+content in the sample solution was measured using an atomic absorption spectrophotometer (TAS-986 Atomic Absorption Spectrophotometer, Beijing PERSEE Co.,Ltd., Beijing, China).

The method to extract malate and soluble sugar from cotton fiber is detailed in Yang et al.[30],which combines the methods for measuring malate content described by Famiani et al. [39] and soluble sugar determination by Jones et al. [40]and Famiani et al. [39]. After extraction, soluble sugars(including glucose,fructose, and sucrose)were assayed using the enzyme-coupled method[39,40].

2.5. Measurements of PEPC, V-ATPase, and PPase activity

The tonoplast solution of cotton fibers was extracted using differential centrifugation according to Tang et al. [41]. The extraction buffer contained 250 mmol L-1mannitol,30 mmol L-1Hepes-Tris(pH 7.5),5 mmol L-1EGTA,1 mmol L--1PMSF,2 mmol L-1DTT,1.5%(w/v)PVP 4000,0.5%(w/v)BSA,10% glycerin, and 10% PVPK-30. Ground fresh fiber samples(0.5 g) were homogenized in 5 mL extraction buffer and centrifuged at 480 ×g for 10 min. The supernatant was then collected and centrifuged at 60,000 ×g for 30 min. The sediment was suspended in buffer containing 250 mmol L-1mannitol, 6 mmol L-1Hepes-Tris (pH 7.5), 2 mmol L-1DTT,0.1 mmol L-1PMSF, and 10% glycerin before adding 6%Dextran T-70 for gradient centrifugation at 70,000 ×g for 2 h at 4 °C; and the top 0-6% of supernatant was collected as the tonoplast solution.

The methods of Tang et al.[41]and Liu et al.[42]were used to determine V-ATPase and PPase activities. The reaction buffer contained 30 mmol L-1Hepes-Tris(pH 7.5),3 mmol L-1MgSO4, 50 mmol L-1KCl, 0.5 mmol L-1NaN3, 0.125 mmol L-1(NH4)2MoO4, and 0.125 mmol L-1Na3VO4. To determine VATPase activity,400 μL of reaction buffer,50 μL of 20 mmol L--1ATP-Tris,and 50 μL of tonoplast solution were combined at 37 °C.To determine PPase activity,400 μL reaction buffer and 50 μL of 20 mmol L-1PPi-Tris were added to 50 μL of tonoplast solution.After 20 min of incubation,1 mL of stop solution(5%(NH4)2MoO4:5 mol L-1H2SO4:H2O = 1:1:3) was added to terminate the reaction before adding 200 μL of the chromogenic agent (0.25 g aminophenol sulfonic acid dissolved in 100 mL of 1.5% Na2SO3at pH 5.5 and 0.5 g Na2SO3). After 40 min,enzyme (V-ATPase and PPase) activity was determined by measuring the absorbance of solutions at 660 nm.A standard curve was established using KH2PO4concentrations ranging from 0 to 10 μmol L-1.

Quantification of PEPC activity was carried out according to Smart et al.[10]and Chen et al.[35].The reaction solution was added to 100 μL enzyme extract, and the change in absorbance at 340 nm in duration of 1 min at 24 °C was recorded.During the reaction, malate dehydrogenase was enzymatically coupled, and the NADH oxidation rate monitored in the presence of 25 mmol L-1BTP-MES (pH 8.0), 10 mmol L-1MgCl2,10 mmol L-1NaHCO3,0.2 mmol L-1NADH,5 mmol L-1DTT, 3 mmol L-1PEP, and 10 units of malate dehydrogenase[35].

2.6. Measurement of PM H+-ATPase activity

Extraction of PM H+-ATPase was done according to De Michelis and Spanswick[43]and Dalir et al.[44].Cotton fibers were homogenized using a mortar and pestle in ice-cold buffer containing 0.25 mol L-1sucrose, 2 mmol L-1EGTA,2 mmol L-1ATP, 2 mmol L-1MgSO4, 1 mmol L-1PMSF, 10%glycerol, 2 mmol L-1DTT, 0.5% BSA, and 25 mmol L-1BTP(pH 7.6). The resulting homogenate was filtered through gauze, and centrifuged at 13,000 ×g for 10 min. The supernatant was collected and centrifuged at 80,000 ×g for 30 min.The supernatant was removed and crude microsomes were resuspended in the same buffer plus 0.25 mol L-1KI(1 mL for 1 g fresh weight),incubated on ice for 15 min,and centrifuged at 80,000 ×g for 30 min. The pellet was obtained and resuspended in buffer containing 0.25 mol L-1sucrose, 10% glycerol, 1 mmol L-1DTT, 0.2% BSA, and 2 mmol L-1BTP, and pH was adjusted to 7.0 with MES.The solution was separated into four aliquots and stored at-80 °C.

Activity of PM H+-ATPase was assayed using a modified method of Liu et al. [42].A reaction solution (450 μL)containing 1 mmol L-1NaN3, 1 mmol L-1Na2MoO4, 3 mmol L-1MgSO4,30 mmol L-1ATP-Na2, 36 mmol L-1Tris-MES, 50 mmol L-1KNO3, and either the presence or absence of 2.5 mmol L-1Na3VO4. Then 50 μL of plasma membrane vesicles was added to start the reaction, incubated at 37 °C for 20 min and then quenched by adding 1 mL of a solution containing 5% (NH4)2-MoO4:5 mol L-1H2SO4:H2O = 1:1:3.After 40 min,enzyme activity was determined by measuring the absorbance of the solution at 660 nm. A standard curve was established using KH2PO4concentrations ranging from 0 to 10 μmol L-1.

2.7. Data analysis

A three-way factorial analysis of variance (ANOVA) was performed to evaluate the effects of cultivars, water treatment, K rates, and their interactive effects using SPSS 21.0(SPSS Inc.,Chicago,IL,USA).Graphs were plotted using Sigma Plot 10.0 (Systat Software Inc.,Chicago,IL, USA).

3. Results

3.1. Fiber elongation

Cultivar (C), water treatment (WT), and potassium (K) significantly affected final cotton fiber length(Fig.3).WT and K had interactive trends (Fig. 4) even their interaction did not pass significance test(P ＞0.05).Drought significantly reduced fiber length in both years,more so in Siza 3 than Simian 3.The DS treatment reduced final cotton fiber length by respective averages of 8.74%(K0),7.11%(K1),and 6.71%(K2)in Siza 3 and 6.75% (K0), 5.44% (K1), and 4.98% (K2) in Simian 3, relative to the WW treatment. The application of K significantly increased fiber length and reduced the magnitude of fiber length reductions due to drought. Fiber length in K1 and K2 respectively increased by 3.99%and 8.24%in Siza 3 and 3.54%and 8.08%in Simian 3 under WW,and respectively increased by 5.85%and 10.67%in Siza 3 and 4.99%and 10.14%in Simian 3 under DS,relative to K0.

To determine T and Vmax, a logarithmic function was fit to the fiber length data. Drought reduced Vmaxand increased T.Increasing K levels increased Vmax,but had variable effects on T,depending on cultivar, water treatment and year (Fig. 5). The trends for Vmaxin response to K and water treatment were comparable to the trends of final fiber length.For example,Vmaxof DS fibers decreased by respective averages of 13.33% (K0),11.57%(K1),and 10.45%(K2)in Siza 3 and 7.32%(K0),6.01%(K1),and 3.02%(K2)in Simian 3,relative to WW fibers.In addition,the Vmaxrespectively increased by 7.99%(K1)and 17.20%(K2)in Siza 3 and 6.29%(K1)and 12.04%(K2)in Simian 3 in WW fibers,and by 10.22% (K1) and 21.24% (K2) in Siza 3 and 7.79% (K1) and 17.25%(K2) in Simian 3 in DS fibers,relative to K0. Importantly,data plots of Vmaxand T against final fiber length for all treatments, cultivars, and years revealed a slight negative relationship between final fiber length and T, and a strong positive relationship between final fiber length and Vmax(Fig.5).Thus,Vmaxwas the dominant driver of fiber length responses in both K and water treatments.

3.2. Osmotic potential

The OP of cotton fiber decreased at 10 and 15 DPA and then increased as fibers continued to elongate in all treatments(Fig.6).Fiber OP at 10 and 15 DPA decreased significantly with drought or increased K rate (P ＜0.05); WT × K effects significantly increased OP at 10 DPA in 2016. Because the drought treatment decreased cotton fiber OP significantly at 10 and 15 DPA, other parameters, such as osmolyte content and the activity of related enzymes, that were sampled at 20 and 25 DPA were not analyzed further.

Fig. 3 - Cotton fiber length of matured cotton bolls at different potassium levels (K), water treatments (WT), cultivars (C), and years. K0, K1, and K2 are 0, 150, and 300 kg K2O -1, respectively; WW, well-watered treatment; DS, soil drought treatment. Values are means of three replicates. Vertical bars represent standard errors. * and ** indicate significant differences at 0.05 and 0.01 probability levels, respectively. Ns, not significant. Values followed by different letters within the same cultivar differ significantly at the 0.05 probability level.

The OP of DS cotton fibers at 10 and 15 DPA decreased by 48.98%-40.61% (K0), 55.21%-46.93% (K1), and 56.81%-44.22%(K2)in Siza 3 and by 54.85%-49.22% (K0),70.12%-54.56%(K1),and 59.65%-50.73% (K2) in Simian 3, relative to the WW treatment. Increasing K application significantly reduced fiber OP. The OP of WW cotton fiber at 10 and 15 DPA declined by 4.26%-2.38% and 4.34%-6.98% in Siza 3 and 2.04%-0.30% and 7.48%-5.04% in Simian 3 under K1 and K2,respectively, relative to K0. In DS cotton fibers at 10 and 15 DPA, the OP decreased by 8.65%-6.98% and 9.86%-9.72% in Siza 3,and 12.14%-3.98%and 10.89%-6.20%in Simian 3 under K1 and K2, respectively, relative to K0. Thus, K application reduced fiber OP more in the DS treatment than WW treatment.

Fig. 4 - Changes in cotton fiber length between K levels in different water treatments (WT). K0, K1, and K2 are 0, 150, and 300 kg K2O h a -1, respectively; K1-K0 represent the changes in cotton fiber length between K1 and K0, calculated by 100%× (cotton fiber length under K1 - cotton fiber length under K0) / cotton fiber length under K0; K2-K1 represent the changes in cotton fiber length between K2 and K1, calculated by 100%× (cotton fiber length under K2 - cotton fiber length under K1) /cotton fiber length under K1. WW, well-watered treatment, DS, drought treatment.

Fig.5-Changes in maximum rate of fiber elongation(Vmax),duration of rapid development of fiber length(T),and relationship between Vmax/T and fiber length of matured cotton bolls for all the potassium and water treatments,cultivars,and years.*and**indicate significant differences at the 0.05 and 0.01 probability levels,respectively(n = 12,R20.05 = 0.3316,R20.01 = 0.5000),coefficients of stress(CS,Vmax) = 100% × (Vmax(K2) - Vmax(K0)) / Vmax(K2),CS(T) = 100% × (T(K2) - T(K0)) / T(K2).

Fig.6-Effect of drought on cotton fiber osmotic potential at three K levels in two cotton cultivars in Nanjing in 2015 and 2016.K0,K1,and K2 are 0,150,and 300 kg K2O ha-1,respectively;WW,well-watered treatment,DS,drought treatment.Values are means of three replicates. Vertical bars represent standard errors.*indicates significant differences between WW and DS on each sampling day at the 0.05 probability level.

Fig.7- Effect of drought on cotton fiber K+,malate,and soluble sugar contents at three K levels in two cotton cultivars in Nanjing in 2015 and 2016.K0, K1,and K2 are 0,150,and 300 kg K2O ha-1,respectively. WW,well-watered treatment; DS,drought treatment.Values are means of three replicates.Vertical bars represent standard errors.Values followed by different letters within the same cultivar differ significantly at the 0.05 probability level.C,cultivar;WT,water treatment;K,potassium level.* and**indicate significant differences at the 0.05 and 0.01 probability levels,respectively;Ns, not significant.

3.3. Osmolytes content and contributions to OP

3.3.1. K+content

Drought significantly increased fiber K+content (P ＜0.05) in both cultivars and years (Fig. 7). Fiber K+content under DS at 10 and 15 DPA increased by 11.82%-8.74% (K0), 12.61%-9.64%(K1) and 13.73%-13.44% (K2) in Siza 3, respectively, and by 10.36%-10.27% (K0), 16.97%-13.61% (K1), and 10.92%-10.44%(K2) in Simian 3, relative to WW. The highest K+content occurred in the DS K2 treatment.Compared to K0,cotton fiber K+contents under WW(at 10 and 15 DPA)increased by 7.41%-11.67% (K1) and 26.74%-28.49% (K2) in Siza 3, and by 6.83%-12.03% (K1) and 18.10%-24.99% (K2) in Simian 3. In DS plants,cotton fiber K+contents at 10 and 15 DPA increased by 8.19%-12.58% (K1) and 28.92%-34.04% (K2) in Siza 3, and 13.25%-15.42%(K1)and 18.75%-25.22%(K2)in Simian 3,relative to K0.

3.3.2. Malate content

Drought significantly increased the malate content in cotton fiber (P ＜0.05) at 10 and 15 DPA in both years (Fig. 7).Specifically, the malate content of DS cotton fibers at 10 and 15 DPA respectively increased by 9.35%and 8.58%(K0),12.10%and 8.12% (K1) and 13.13% and 7.19% (K2) in Siza 3, and by 12.24%and 10.66%(K0),14.95%and 9.97%(K1)and 15.42%and 9.43% (K2) in Simian 3, relative to the WW. Compared to K0,cotton fiber malate content increased in response to K application, ranging from 18.33% to 31.58% in Siza 3 and from 11.74% to 41.41% in Simian 3 under WW, and ranging from 21.32%to 36.14%in Siza 3 and 11.04%to 45.40%in Simian 3 under DS. Thus, K application significantly augmented malate content, more so in the DS treatment than WW treatment.

3.3.3. Soluble sugar content

Drought significantly reduced the soluble sugar content at 10 and 15 DPA in both years, relative to the WW treatment. The differences in soluble sugar content between WW and DS decreased with K application. The soluble sugar contents of DS cotton fibers at 10 and 15 DPA respectively declined by 24.04%and 21.55%(K0),16.47%and 12.36%(K1),and 8.76%and 14.12%(K2)in Siza 3 and by 17.30%and 20.14%(K0),13.04 and 19.29%(K1),and 12.60%and 16.08%(K2)in Simian 3,relative to the WW treatment (Fig. 7). Compared to K0 under WW condition, soluble sugar content at 10 (and 15) DPA under K1 and K2 increased respectively by 7.15% (13.69%) and 8.42%(18.15%) in Siza 3, and respectively by 3.84% (20.87%) and 14.73% (44.27%) in Simian 3; whereas under DS, it increased respectively by 17.94% (27.07%) and 30.21% (29.35%) in Siza 3,and respectively by 9.18% (22.09%) and 21.18% (50.20%) in Simian 3. Therefore, K application not only increased the soluble sugar content of cotton fiber but reduced the magnitude of the decline caused by drought stress.

3.3.4. Contribution of osmolyte to OP and OA

The contribution of cotton fiber osmolyte at 10 and 15 DPA to OP (Table 1) and OA (Table 2) were calculated. For OP,osmolytes contributed ～80% in the WW treatment and 70%in the DS treatment, and K+and malate were the dominant contributors to OP in both cultivars (Table 1). K application increased the total osmolyte, K+, and malate contributions to OP in both water treatments but had variable effects on the contribution of soluble sugars(Table 1).Compared to K0 under WW, the total osmolyte contribution increased with increasing K application from 2.82%to 16.67%in Siza 3,and 2.66%to 10.27% in Simian 3; the respective contribution increases for K+were from 0.86%to 20.40%in Siza 3,and 0.69%to 10.71%in Simian 3,and for malate were from 5.40%to 25.17%in Siza 3,and 1.57%to 26.53%in Simian 3.In the DS treatment,relative to K0, the total contribution increased with increasing K application from 6.85%to 29.90%in Siza 3,and 4.77%to 13.37%in Simian 3;the respective contribution increases for K+were from 14.81%to 23.21%in Siza 3,and 1.36%to 15.10%in Simian 3, and for malate were from 7.96% to 34.51% in Siza 3, and 2.21%to 30.01%in Simian 3.

OA under DS increased with increasing K application(P ＜0.05) at 10 and 15 DPA in both years. The contribution of osmolyte to OA increased with increasing K application,more so in Simian 3 than Siza 3 under the same conditions(Table 2).Take 2016 as an example,the OA of DS cotton fibers at 10 and 15 DPA were, respectively, 0.50 and 0.48 MPa (K0), 0.57 and 0.57 MPa (K1), 0.64 and 0.56 MPa (K2) in Siza 3 and 0.51 and 0.51 MPa(K0),0.61 and 0.58 MPa(K1),0.64 and 0.59 MPa(K2)in Simian 3.While using WWK0 as CK,OA under DSK1 at 10 and 15 DPA were respectively 0.64 and 0.60 MPa in Siza 3,and 0.62 and 0.62 MPa in Simian 3; and OA under DSK2 at 10 and 15 DPA were respectively 0.70 and 0.66 MPa in Siza 3, and 0.73 and 0.64 MPa in Simian 3.Correspondingly,the total osmolyte contribution and individual osmolyte contribution also increased with K application.

3.4. Activity of enzymes involved in osmolyte accumulation(PM H+-ATPase, V-ATPase, PPase, and PEPC)

3.4.1.PM H+-ATPase, V-ATPase, and PPase

The variables C, WT, and K significantly influenced the activities of PM H+-ATPase,V-ATPase,and PPase at 10 and 15 DPA; only the interaction of WT and K application had a significant effect on PM H+-ATPase activity (P ＜0.05, Fig. 8).Drought stress increased PM H+-ATPase activity by 11.21% to 28.93% in Siza 3 and 10.78% to 30.89% in Simian 3, V-ATPase activity by 10.66% to 58.97% in Siza 3 and 10.14% to 46.83%Simian 3, and PPase activity by 8.88% to 60.07% in Siza 3 and 6.78%to 49.23%in Simian 3,relative to the WW treatment.

PM H+-ATPase activity of WW cotton fibers increased in response to K application(K1 and K2)from 27.94%to 76.64%in Siza 3 and 11.09% to 89.13% in Simian 3; V-ATPase activity from 14.87%to 52.83%in Siza 3 and 5.99%to 28.04%in Simian 3;and PPase activity from 24.42%to 62.93%in Siza 3 and 8.41%to 57.33% in Simian 3. When compared with DS treatment at K0, K application (K1 and K2) under DS increased PM H+-ATPase activity from 47.68%to 93.21%in Siza 3 and 10.88%to 86.31%in Simian 3,V-ATPase activity from 15.32%to 38.74%in Siza 3 and 6.20%to 19.19%in Simian 3,and PPase activity from 18.81%to 62.44%in Siza 3 and 15.08%to 32.11%in Simian 3.

3.4.2.PEPC

The main effects of C,WT,K and interaction effect of K × WT significantly affected PEPC activity in both cultivars at 10 and 15 DPA while C × K × WT had a significant effect at 10 DPA in both years(P ＜0.05; Fig. 9). Under DS,PEPC activity increasedby 8.82%to 19.69%in Siza 3 and 11.88%to 35.28%in Simian 3,relative to the WW treatment. In the WW treatment, K application (K1 and K2) increased PEPC activity from 40.51%to 79.42%in Siza 3 and 21.52%to 48.54%in Simian 3,relative to K0.In the DS treatment,K application increased PEPC activity from 33.24%to 84.82%in Siza 3 and 26.35%to 62.74%in Simian 3, relative to K0.

Table 1-Effect of drought on the contribution of cotton fiber osmolytes to osmotic potential (OP) at 10 and 15 days postanthesis(DPA)in two cotton cultivars grown at three K levels in 2015 and 2016.

4. Discussion

Intermittent drought events, round the globe, are now occurring more frequently and with greater severity [1].Drought stress not only reduces crop yield but also negatively affects crop quality [17,31]. As a result, enhancing crop tolerance to drought is essential for maintaining yield stability and quality in future climates. Potassium is an important electrolyte in plants for many osmotic functions because it is readily translocated across cell membranes and between organs. Furthermore, plants can absorb excess K under conditions of high K availability and accumulate a ready supply of K reserves to remobilize when nutrient uptake is restrained by stress [45]. Therefore, high K availability prior to drought exposure has the potential to relieve the negative effects of drought stress[18,21,28].

Consistent with previous studies[16,17],drought significantly reduced cotton fiber length in both cultivars in the present study.However, the magnitude of the decline diminished as K application level increased, as reported by Ahmad et al. [31]and Yagmur et al. [18]. Fiber length is principally a function of Vmaxand T[36].Which one is the primary factor depends on the type of stress that the plant is exposed to [35]. For example,waterlogging-stress reduced fiber length primarily by reducing Vmax, while high-temperature stress reduced fiber length by shortening T[35].In the current study,drought reduced Vmaxbut increased T, the increased K application diminished the reduction of Vmaxeven its effects on T were not certain.Therefore, when considering characteristic values (Fig. 5), Vmaxmay be the dominant driver of fiber length responses to either K or water treatment[30].

Malate, K+, and soluble sugars are the most important osmolyte in cotton fiber. They combinedly determine the OP in fiber cell,the fundamental driving force of fiber elongation[30].In accordance with earlier studies[9,11,46],the measured OP and osmolyte contents reached their maxima at 10 and 15 DPA, when fiber elongation rates also reached their maxima.During this stage, drought significantly increased K+and malate contents but substantially decreased soluble sugar content and fiber OP. It is well-established that drought increases osmolyte concentration in plant organs due to dehydration and decomposition, and synchronouslydecreases OP [47]. In cotton leaves, K+, soluble sugar and organic acid contents all increased under drought condition[28], while in cotton fiber, soluble sugar content prominently declined during drought stress in this study and an earlier study[19].Sucrose functions as an important osmolyte during OA; its concentration depends on the availability of photosynthate in the developing boll. Since drought inhibits photosynthesis capbiltiy [29,48,49] and sucrose transferring from leaf to boll [50], it is likely that the decline in sucrose concentration in drought-stressed cotton fiber is an important factor affecting fiber elongation rate.

Table 2-Osmotic adjustment(OA)between well-watered and drought-stressed cotton fiber and the contribution of cotton fiber osmolyte to OA at 10 and 15 days post-anthesis(DPA)in two cotton cultivars grown at three K levels in 2015 and 2016.

Potassium is an important cofactor of many enzymes; it can increase enzyme activity and help to maintain osmotic balance under drought conditions [51]. Increases in osmolyte concentrations in cotton fiber were likely due to an increase in enzyme activity.Activities of V-ATPase,PPase,PM H+-ATPase,and PEPC increased(Figs.8-9)with corresponding increases in K+and malate contents (Fig. 7). Our results and other studies suggest that malate content is highly regulated by K concentration mediated by related enzymes[10].

In both well-watered and drought conditions, K application(K1, K2) significantly increased the activities of the abovementioned enzymes,K+and malate content in plants.Moreover,these may lead to higher osmotic pressure(-OP)to benefit fiber elongation. Similar findings have been reported [27]. The accumulation of solutes can increase the osmotic pressure of drought-stressed tissues to maintain an acceptable turgor pressure for growth [26]. In our data, drought stress increased solutes concentration, decreased OP and the Vmaxof fiber cell elongation; however, K application further increased osmolyte concentration and decreased OP, but increased Vmax, which suggests that K application could mitigate drought-induced inhibition of cotton fiber elongation.

The question arises as to why a decline in OP caused by drought stress would reduce the Vmaxand final fiber length while a decline in OP caused by higher K application would increase the Vmaxand final fiber length (more so at higher K than low K)? Is that contradictory? To address this question,we need to go back to see the essential driving force of fiber cell elongation. It is well known that turgor pressure is the essential force for fiber cell elongation, and that turgor pressure (TP) is the water potential (WP) minus OP: TP =WP - OP [46]. Under drought conditions, the increase in osmolytes is caused by dehydration, which is a passive cell response to drought and directly caused TP decrease and cell elongation decreased. At the same time, WP remained the same but OP decreased as osmolyte (K+, malate and sucrose contents) increased due to higher K applicaiton, thereby increasing TP and fiber cell elongation. As a result, the inference that K could mitigate drought-imposed reductions in fiber length by decreasing OP is reasonable.

Under drought stress, the benefit of K application to fiber elongation may also be related to fiber cell wall loosening. In the literature, cotton fiber elongation starts with cell loosening and is terminated by an increase in wall rigidity and loss of higher turgor[11].A future study on the mitigating effects of K on fiber length reduction in cotton should focus on changes in fiber cell wall loosening with K application.

Fig.8-Effect of drought on cotton fiber PM H+-ATPase,V-ATPase,and PPase activities at three K levels in two cotton cultivars in Nanjing in 2015 and 2016.K0, K1,and K2 are 0,150,and 300 kg K2O ha-1,respectively.WW,well-watered treatment;DS,drought treatment.Values are means of three replicates.Vertical bars represent standard errors.Values followed by different letters within the same cultivar differ significantly at the 0.05 probability level.C,cultivar;WT,water treatment;K,potassium level.* and**indicate significant differences at the 0.05 and 0.01 probability levels,respectively.Ns, not significant.

5. Conclusions

Fig.9- Effect of drought on cotton fiber phosphoenolpyruvate carboxylase(PEPC)activity at three K levels in two cotton cultivars in Nanjing in 2015 and 2016.K0,K1, and K2 are 0,150,and 300 kg K2O ha-1,respectively.WW,well-watered treatment;DS,drought treatment.Values are means of three replicates. Vertical bars represent standard errors.Values followed by different letters within the same cultivar differ significantly at the 0.05 probability level.C,cultivar;WT,water treatment;K,potassium level.* and**indicate significant differences at the 0.05 and 0.01 probability levels,respectively.Ns,not significant.

Drought decreased Vmaxin cotton fiber and,consequently,led to a significant decline in final fiber length.The application of K mitigated the drought-induced reductions in fiber length by increasing Vmax. Under drought, increasing K+and malate contents caused the decline in OP at 10 and 15 DPA,which was strongly associated with increases in PEPC, V-ATPase, PPase,and PM H+-ATPase activities. Physiologically, the application of K before the onset of drought could ameliorate reductions in fiber elongation by enhancing osmolyte (K+, malate and soluble sugars) accumulation. In addition, the higher K application level increased osmolyte (K+, malate and soluble sugars)contribution to OP in fiber cells more than the lower K application, suggesting that K could increase the osmotic adjustment capability of cotton, not only by regulating the osmolyte content but the composition of osmolytes.

Declaration of Competing Interest

Authors declare that there are no conflicts of interest.

Acknowledgments

We acknowledge financial support from the National Key Research and Development Program of China(2018YFD1000900), Jiangsu Collaborative Innovation Center for Modern Crop Production (JCIC-MCP) and Jiangsu Overseas Research and Training Program for University Prominent Young and Middle-aged Teachers and President(2016),China.