Cho Wu,Kehui Cui*,Qiuqin Hu,Wencheng Wng,Lixio Nie,Jinling Hung,Shobing Peng

a National Key Laboratory of Crop Genetic Improvement,MOA Key Laboratory of Crop Ecophysiology and Farming System in the Middle Reaches of the Yangtze River,Huazhong Agricultural University,Wuhan 430070,Hubei,China

b College of Agronomy,Nanjing Agricultural University,Key Laboratory of Crop Physiology Ecology and Production Management,Ministry of Agriculture,Jiangsu Collaborative Innovation Center for Modern Crop Production,Nanjing 210095,Jiangsu,China

c Hubei Collaborative Innovation for Grain Industry,Yangtze University,Jingzhou 434023,Hubei,China

Keywords:Heat stress Rice(Oryza sativa L.)Spikelet fertility Stigm a exsertion Stigm a and pollen viability

A B S T R A C T With global warming,rice plants may be subjected to heat stress more regularly during the heatsensitive flowering stage,causing spikelet sterility and grain yield loss.Stigma exsertion is considered to increase pollen reception and promote fem ale reproductive success.The aim of this study was to investigate the role of stigma exsertion on spikelet fertility at high temperatures.Five rice cultivars(Liangyoupeijiu,Shanyou 63,Huanghuazhan,Nagina 22,and IR64)with differing degrees of stigma exsertion were cultivated and exposed to high temperature at anthesis.Heat-tolerant cultivars did not always show a high percentage of spikelets with exserted stigmas,and vice versa.Irrespective of the presence of more pollen grains on exserted stigmas,spikelets with exserted stigmas did not show greater spikelet fertility than spikelets with fewer exserted stigmas or hidden stigmas under heat stress.GA3 application augmented the percentage of spikelets with exserted stigmas;however,it did not increase spikelet fertility under heat stress.Spikelet fertility of whole panicles was negatively correlated with the percentage of spikelets with exserted stigmas,but positively with that with hidden stigmas.Viability of the hidden stigmas was less reduced than that of exserted stigmas under heat stress,suggesting that hidden stigmas have an advantage in maintaining viability.Heat stress delayed anther dehiscence and reduced the viabilities of both exserted stigmas and pollens,thereby causing low spikelet fertility.Together,these results suggest that high spikelet fertility does not depend on stigma exsertion and that enclosed stigma generally contributes to higher spikelet fertility and heat tolerance under high-temperature conditions during flowering in rice.

1.Introduction

Rice(Oryza sativa L.)is a staple food crop w orldw ide,especially in developing countries.Am bient tem perature w orldw ide is gradually increasing,w ith probable severe adverse effects on crop yields,including rice[1].Heat stress,especially at the flow ering stage,often leads to substantial yield loss in rice[2].Spikelet fertility is usually associated w ith both pollination and fertilization;how ever,heat-induced spikelet sterility in rice is due prim arily to poor pollination[3].Crop pollination requires pollen to land on a stigm atic surface,and proceeds through three stages:(i)the septum ruptures,the stom ium splits in the apical and basal part of the theca,and the anther dehisces[4];(ii)pollen grains are released from the dehisced thecae[5];and(iii)pollen grains from the anther are deposited on the stigma[6].Generally,there should be＞20 pollen grains deposited on a stigma to insure successful pollination[7].

Poor pollination caused by high-temperature stress at flow ering is due to insufficient pollen deposition on stigm as[8].Heat stress during flow ering hinders shedding of pollen grain by delaying or even blocking anther dehiscence[9].The abnorm al anther dehiscence m ay be attributed to the inhibition of pollen sw elling due to a disturbance of w ater balance caused by high tem perature[10].Heat stress also,by disturbing anther dehydration,results in sticky pollen grains that are retained inside the locules of dehiscent anthers[9].Heat stress affects ion(K+,Ca2+)balance[11],carbohydrate m etabolism[12],and levels of regulators(phytohorm ones)in pollen[13],w hich disturbs pollen germination and pollentube polarized grow th.Anther characters such as the percentage of dehisced thecae and the length of basal dehiscence on the thecae act as the m ajor determ inants of pollen release in rice under heat stress[14].

A higher density of pollen grains shed on a stigma m ay result in higher pollen germ ination,a consequence term ed the pollen population effect[15].The increased num ber of pollen grains deposited on exserted stigm as m ay com pensate for reduced pollen germ ination under heat-stress conditions[16].Although stigm a viability m ay also be affected by heat stress,stigm as show higher tolerance to heat stress than pollen[17,18].It is reported[19]that＞60%of stigmas are viable on the second day of heat treatment.To date,most observations have focused on the process of pollen release,and little attention has been paid to subsequent processes affecting individual pollen grains landing on stigm as under heat conditions.

Given that pollen lands and germ inates on the stigm a,stigm a exsertion is thought to affect fem ale reproductive success by influencing the quantity and quality of pollen that lands on the stigm atic surface.There are tw o types of stigma:(i)exserted stigm as,w hich rem ain outside the spikelet after floret opening,and(ii)hidden stigm as,w hich rem ain inside the spikelet after flow ering[20].Exserted stigm as often receive more pollen.In cytoplasmic male sterile rice lines,an increased percentage of spikelets w ith exserted stigm as m ay im prove pollination,enhance outcrossing,and increase seed setting rate[21],and spikelets w ith dual exserted stigm as show ed higher spikelet fertility than ones w ith a single exserted stigm a[22].These results suggest that increased stigma exsertion may be beneficial for high spikelet fertility ow ing to their advantage in pollen reception.

Stigm a exsertion show s varietal differences[23]and is affected by environm ental conditions such as heat stress[8,24].Generally,indica rice cultivars show higher stigm a exsertion than japonica cultivars[25],and are m ore tolerant to heat stress than japonica cultivars[8].Em pirically,hybrid rice cultivars have higher stigm a exsertion than rice inbred cultivars,but are m ore susceptible to heat stress than inbred cultivars.Although higher stigm a exsertion had positive effects on spikelet fertility[21,22],Wassm ann et al.[26]proposed that low stigm a exsertion is associated w ith low spikelet sterility in rice under high tem perature stress.Thus,the relationship betw een percentage of spikelets w ith exserted stigm as and spikelet fertility in rice under heat conditions rem ains unclear.The objective of this study w as to characterize the relationship betw een percentage of spikelets w ith exserted stigm as and spikelet fertility in rice under heat conditions.

2.Materials and m ethods

2.1.Plant cultivation

Pot experim ents w ere conducted during the 2014 rice-grow ing season at Huazhong Agricultural University,Wuhan,China(30°29′N,114°22′E).Five rice cultivars w ith differing heat tolerance w ere used in this study:Nagina 22 (N22),Huanghuazhan(HHZ),Liangyoupeijiu(LYPJ),Shanyou 63(SY63),and IR64.Am ong these five cultivars,N22 is com m only reported to be highly tolerant to heat stress during flow ering,HHZ and SY63 w ere moderately tolerant to heat stress,and IR64 and LYPJw ere extrem ely heat-susceptible[2,27].

Seeds w ere sow n in plastic seeding trays w ith loam y soil after breaking dormancy at 50°C for 5 days.At the three-leaf stage,four seedlings of each cultivar w ere transplanted into a 14-L plastic pot(height,28.5 cm,top diam eter,30 cm;base diam eter,25 cm)containing a m ixture of 17 kg soil(loam:sand,2:1)and 12.5 g compound fertilizer(N:P2O5:K2O,16%:16%:16%).The potted rice plants w ere random ly arranged w ith four replications under natural am bient conditions.Seedlings w ere thinned to three plants per pot(w ith each plant having three tillers)8 days after transplanting,and the m ain tillers w ere tagged.Each pot w as topdressed w ith 1.0 g urea 10 days after transplanting.The plants w ere irrigated w ith a w ater table depth of approxim ately 2 cm from transplanting to maturity.Each pot was manually rotated 90°clockw ise every seven days to avoid positional effects.Pests,diseases,birds,and w eeds w ere controlled.

2.2.High-temperature treatment

All plants w ere cultivated under natural am bient conditions and w ere then moved into tw o temperature-controlled greenhouses(4 m length×4 m w idth×4.5 m height)one day before anthesis(plants of N22 and SY63 w ere m oved on August 25 and those of LYPJ,HHZ,and IR64 on August 27).Each greenhouse w as equipped w ith a dehum idifier,an air conditioner,tw o ventilators,and tw o sensors for m onitoring relative humidity and air temperature,as described by Wu et al.[28].The plants w ere continuously exposed to high daytim e tem perature for seven days and then w ere rem oved and grow n under norm al and natural conditions until m aturity.

The tw o tem p erature treatm ents w ere high daytim e tem perature(HT),w hich im p osed high tem perature from 07:00 to 19:00,and the control(CK)treatm ent,w hich maintained a favorable tem perature for rice plant grow th during the entire day(Fig.1).For CK,tem p eratures w ere set at 24°C at 5:00-6:00,31°C at 10:00-11:00,32°C at 12:00-13:00,and 27°C at 20:00-21:00.For HT,tem p eratures w ere set to 24°C at 5:00-6:00,35°C at 10:00-11:00,36°C at 12:00-13:00,and 27°C at 20:00-21:00(Fig.1).The relative hum idity(%)in the greenhouses w as set at 80%.The air tem perature and relative hum idity in the greenhouses w ere controlled by a central auto-controller(Auto-Greenhouse Monitoring and Data Managem ent System,Version 3.00,Auto,Beijing,China),w ith tem p eratures inside the greenhouses gradually changed to approach the next set tem perature values.Air tem perature and relative hum idity w ere recorded 5 cm above the rice canop y using a standalone sensor(HOBO,H08-003-02,Onset Com puter Corporation,Bourne,MA,USA).The recorded mean,m aximum,and m inim um tem peratures w ere respectively 27.6,31.8,and 24.8°C und er CK and 29.2,35.9,and 24.1°C under HT.The recorded m ean relative hum id ity w as 81.2%und er CK and 81.7%und er HT.

2.3.Application of exogenous gibberellin A3(GA3)

A solution of GA3(60 mg L-1)w as prepared by dissolving 60 m g GA3(Sigm a-Aldrich,St.Louis,MO,USA)in 20 m L ethanol,follow ed by dilution to a final volum e of 1 L w ith double-distilled H2O.The solution w as mixed thoroughly after addition of tw o drops of 0.01%(v/v)Tw een 20 as a surfactant.The GA3solution w as sprayed onto the panicles of plants under the CK and HT treatm ents,w ith 20 m L per dose per plant.GA3application prom otes stigm a exsertion during 1%-50%heading stage[29].The GA3solution w as applied tw ice for ensuring stigm a exsertion,at 9:00 on the day before HT treatm ent and on the second day after HT treatm ent.When GA3w as applied,the treated plants w ere m oved out of the greenhouse to avoid GA3drifting onto the other plants,and then w ere m oved back into the greenhouse after GA3application.The sam e volum es of double-distilled H2O containing the sam e concentration of Tw een 20 w ere applied to the control plants.

2.4.Determination of degree of stigma exsertion and spikelet fertility

To determ ine the degree of stigm a exsertion,three m ain panicles w ere harvested 14 days after heading from three individual plants grow n in different pots for each replication.The spikelets w ere separated into three groups(Fig.2):sp ikelets w ith dual-exserted stigm as(w ith tw o stigm as exserted from both sides of a sp ikelet),single-exserted stigm a(w ith a single stigma exserted from one side of a sp ikelet),and non-exserted stigm a(w ith no stigm a exserted from a spikelet).Spikelets w ere tested by pressure betw een the forefinger and thum b,and spikelets containing a kernel were assigned as fertile and those w ithout as sterile.The num bers of fertile and sterile spikelets w ere recorded for each group.

The percentage of spikelets w ith single or dual exserted stigmas and the fertility of spikelets w ith single or dual exserted stigm as or non-exserted stigm as w ere calculated according to Yan et al.[23].The percentage of spikelets w ith a single exserted stigm a w as defined as 100×the ratio of spikelets w ith a single exserted stigma to the total number of spikelets.The fertility of spikelets w ith single exserted stigm a w as defined as 100×the ratio of fertile spikelets w ith single exserted stigm a to total spikelets w ith single exserted stigm a.Sim ilar calculations w ere perform ed for spikelets w ith dual exserted stigm as and w ith non-exserted stigm a.The spikelet fertility(%)of w hole panicles w as calculated on the basis of all spikelets per panicle.

Fig.1-Tem perature settings for high-tem perature treatm ent and control.

Fig.2-Stigm a exsertion typ es in rice spikelets.

2.5.Observation of anther dehiscence,stigma and pollen vigor,and pollen grain deposition on stigma

According to previous results[28,30],the rice cultivars LYPJ and SY63 w ith differing heat tolerance w ere selected for determ ination of pollen vigor,stigm a receptivity,and pollen germ ination.Anthers w ere collected at flow ering from opening florets at 9:30,11:30,and 16:00 under control and HT treatm ents.The tem peratures in the greenhouses at these tim es w ere 29.8,31.3,and 29.9°C under the control and 33.1,35.3,and 32.2°C under HT(Fig.1).Sampling was repeated three times for a given treatment.An anther w ith opened apical and/or basal pores w as recorded as dehisced.The percentage of anther dehiscence w as calculated as 100×the ratio of dehisced anthers to total anthers[28].

To assess stigm a vigor,stigm as of florets that had opened at the same time(about 08:00)w ere dissected at 9:30,11:30,and 16:00 on the sam e day and 30 exserted and non-exserted stigm as from each sam pling w ere im m ersed in benzidinehydrogen peroxide solution(1%benzidine:3%H2O2:H2O=4:11:22)for 5 m in.A viable stigm a w as indicated by the presence of peroxidase-generated bubbles around stigm as and the oxidation of benzidine,w hich produced a blue coloration[31].The stained stigmas w ere then examined at 40×magnification using a DP70 digital camera attached to an Axioplan 2 m icroscope(Carl Zeiss,Jena,Germ any),and colored stigmas(gradually changing from blue to red-brow n)w ere considered viable.The percentage of viable stigm as w as calculated as 100×the ratio of viable stigm as to total stigm as.

Pollen grains w ere scattered and stained on a glass slide w ith one drop of 1%2,3,5-triphenyl tetrazolium chloride by gently tapping the opening florets that had been sam pled at 9:30,11:30,and 16:00[28].The stained pollen grains w ere exam ined by m icroscopy as described above.Three observations w ere performed per slide,w ith each observation made on approxim ately 200 pollen grains.Pollen grains that w ere stained red w ere considered viable.The percentage of viable pollen grains w as calculated as 100×the ratio of viable pollen grains to total pollen grains.

For observations of pollen grain deposition and germ ination on stigm as,spikelets w ere collected in a 10-m Ltube filled w ith FAA solution(38%form aldehyde:acetic acid:70%ethanol=1:1:18)at 09:30 and 16:00,and w ashed in deionized w ater,follow ed by dissection under a stereom icroscope.Six exserted and non-exserted stigm as for each tim e point w ere stained with 0.1%aniline blue solution in 0.1 mol L-1K3PO4for 10 m in and then exam ined by m icroscopy as described above.The total num ber of pollen grains and germ inated pollen grains w ith a pollen tube em erging from the pollen coat on each stigma w ere recorded[3].The percentage of germinated pollen grains w as calculated as 100×the ratio of the num ber of germ inated pollen grains to the total num ber of pollen grains on the stigm as.

2.6.Data analysis

The four treatm ents differing in tem perature regime and GA3application w ere designated as CK-H2O,HT-H2O,CK-GA3,and HT-GA3.The m eans of four experim ental replicates w ere used for analysis.Analysis of variance(ANOVA)w as perform ed using Statistix 8.0(Analytical Softw are,Tallahassee,Florida,USA)and show ed that there w ere significant differences am ong the results from the different tem perature treatm ents and am ong cultivars.The differences am ong m eans from the four treatm ents w ere compared by the least significant difference(LSD)test at P＜0.05.The m eans over four replications w ere used to estim ate the relationships betw een the percentages of spikelets w ith exserted and hidden stigm as and spikelet fertility based on w hole panicles under CK and HT treatm ents,using Sigm aplot softw are version 12.5(Systat Softw are,Inc.,San Jose,CA,USA).

3.Results

3.1.Effects of high-temperature treatment and GA 3 application on stigma exsertion and the fertility of spikelets

Stigm a exsertion show ed varietal differences(Table 1).N22 show ed no exserted stigm as.LYPJ show ed the highest percentage of spikelets w ith single and dual exserted stigm as,respectively,68.8%and 17.6%across the four treatm ents.They w ere follow ed by HHZ(22.9%of spikelets w ith single exserted stigm as and 3.2%of spikelets w ith dual exserted stigmas).SY63 and IR64 showed similar percentages of spikelets w ith single and dual exserted stigm as across the four treatments:11.5%and 2.7%for SY63 and 11.2%and 3.8%for IR64(Table 1).

Regardless of GA3application,HT treatm ent generally show ed no substantial effect on the percentages of spikelets with single and dual exserted stigmas except in LYPJ,in w hich HT treatm ent reduced the num ber of spikelets w ith dual exserted stigm as(Table 1).Sim ilarly,regardless of tem perature treatm ent,GA3application generally show ed no substantial effect on the tw o percentages.How ever,GA3application increased the num bers of spikelets w ith dual exserted stigm as in LYPJ under the tw o tem perature treatm ents and in IR64 under HT treatm ent,and also increased the num bers of spikelets w ith single exserted stigm as in LYPJand HHZ under the tw o tem perature treatm ents(Table 1).

Among the five investigated cultivars,N22,w hich had no stigm a exsertion,show ed the highest w hole-panicle spikelet fertility,with the other four cultivars showing low spikelet fertility under both CK and HT treatments(Table 1),regardless of their various degrees of stigm a exsertion.There w ere varietal variations in the spikelet fertility of whole panicles in response to HT treatment,w hich reduced spikelet fertility significantly in LYPJ,IR64,HHZ,and SY63 by approximately 46%,53%,39%,and 20%,respectively.How ever,HT treatm ent did not significantly reduce spikelet fertility in N22(Table 1).Similarly,HT treatment resulted in reductions in the fertility of spikelets with single exserted,dual exserted,or non-exserted stigma,regardless of GA3application.Generally,GA3application had negligible effect on fertility under the CK and HT treatments,except for that of spikelets w ith dual exserted stigmas and of entire panicles in LYPJunder CK treatment.

Interactions betw een cultivar and tem perature treatm ent were observed for the six traits related to stigma exsertion and spikelet fertility(Table 1).Significant interactions betw een cultivar and GA3application w ere observed for percentage of spikelets w ith single exserted stigma and percentage of spikelets w ith dual exserted stigm as,and significant interaction betw een tem perature and GA3application w as found only for spikelet fertility of w hole panicles(Table 1).

3.2.Relationship betw een stigma exsertion and spikelet fertility

Spikelet fertility of whole panicles show ed a significant negative correlation w ith the percentage of spikelets w ith single exserted stigm as under the CK and HT treatm ents(Fig.3-A).There w ere significant negative correlations betw een spikelet fertility of w hole panicles and percentage of spikelets w ith dual exserted stigm as under HT treatm ent(Fig.3-B).Spikelet fertility of w hole panicles show ed negative associations w ith the percentage of all spikelets w ith exserted stigmas under the CK and HT treatments(Fig.3-C).How ever,spikelet fertility w as positively correlated w ith the percentage of spikelets w ith hidden stigm as w ith r=0.77(P＜0.05)under CK and 0.84(P＜0.01)under HT(Fig.3-D).

3.3.Effects of high temperatureand GA3 application on anther dehiscence and pollen viability

Fig.4-A-Eillustrates the effects of HT and GA3application on anther dehiscence.Generally,HT treatm ent had negligible effect on anther dehiscence in N22 at the three time points(Fig.4-E).Irrespective of GA3application,HT treatm ent generally reduced anther dehiscence percentages significantly in LYPJ,SY63,HHZ,and IR64 at 9:30 and 11:30.At 16:00,HT treatm ent significantly reduced anther dehiscence percentage only in LYPJand IR64 w ith/w ithout GA3application,and there w as a non-significant reduction in SY63 and HHZ.Differences in anther dehiscence betw een the CK and HT treatments dim inished from morning to afternoon.These results indicate that HT treatm ent delayed anther dehiscence.Generally,GA3application show ed little effect on anther dehiscence and pollen viability in the five cultivars under the CK and HT treatm ents but significantly increased percentage of anther dehiscence in SY63 under HT treatm ent at 9:30(Fig.4-B)and in LYPJ and IR64 under HT treatm ent at 11:30(Fig.4-A,C).

Table 1-Effects of high tem perature(HT)and ex ogenous GA 3 ap plication on stigm a exsertion(%)and spikelet fertility(%).

Fig.3-Relationships betw een the percentages of sp ikelets w ith ex serted and hidden stigm as(%)and the spikelet fertility of w hole p anicles(%).

Generally,HT treatment resulted in a low er percentage of viable pollen grains w hen com pared w ith CK treatm ent.There w as no significant difference betw een CK and HT in LYPJand SY63 at 9:30,regardless of GA3application(Fig.4-F-G).At 11:30,HT treatm ent reduced percentage of viable pollen grains in LYPJ and had no effect in SY63.At 16:00,HT treatm ent significantly reduced the percentage of viable pollen grains in both cultivars,especially in LYPJ,in which it fell by 78%(w ith GA3application)and 83%(w ithout GA3application).Generally,the gap betw een the values under the CK and HT treatments increased from morning to afternoon.GA3application slightly increased the percentage of viable pollen grains under the CK and HT treatm ents(Fig.4-F-G).

3.4.Effects of high temperature and GA 3 application on stigma viability

LYPJis used as an exam ple to illustrate the effects of hightemperature treatment on stigma viability(Fig.5).The exserted and hidden stigm as w ere deeply colored under CK treatm ent(Fig.5-A).How ever,exserted stigm as had a slightly red-brow n color(Fig.5-B,C),and hidden stigm as w ere deeply colored under HT treatm ent(Fig.5-B).The im ages clearly indicate the low viability of exserted stigm as under HT treatm ent.

For a given cultivar,the percentages of viable hidden stigmas at 9:30 w ere sim ilar across tem perature treatments and GA3application(Fig.6-A-E).At 11:30,HT treatm ent slightly reduced the percentage of viable stigm as.At 16:00,the difference between CK and HT treatment increased.Am ong the five cultivars,SY63 show ed the largest reduction in percentage of viable hidden stigm as at 11:30 and 16:00,w hereas the values from N22 and IR64 w ere relatively stable from morning to afternoon.

Fig.4-Effects of high-tem perature treatm ent and ex ogenous GA 3 ap plication on the percentage of anther dehiscence and viable pollen grains.Data are presented as m ean±SD(n=4).Vertical bars represent least significant d ifferences at the 5%level of significance for com paring treatment means betw een treatments of GA 3 application.CK-H2O,control w ithout GA3 app lication;CK-GA,GA 3 ap plication under control cond ition;HT-H2O,high-tem perature treatm ent w ithout GA 3 app lication;HT-GA,GA 3 application under high tem p erature treatm ent.

Sim ilar changes w ere observed for the percentage of viable exserted stigm as (Fig.6-F-I).Generally,the viability of exserted stigm as decreased from m orning to afternoon under the four treatm ents.At 9:30,the different tem perature treatm ents affected only IR64.How ever,HT treatm ent increased the reduction w ith tim e in the four cultivars,especially at 16:00.

Fig.5-Im ages show ing stigm a viability in LYPJ.CK,control;HT,high tem perature treatm ent.Scale bar=100μm.

Com pared w ith hid den stigm as,the p ercentage of viable exserted stigm as w as reduced by 4.4%in SY63,4.7%in LYPJ,3.7%in HHZ,and 9.4%in IR64 at 11:30 und er HT treatm ent,and by 9.1%in SY63,10.6%in LYPJ,9.5%in HHZ,and 14.6%in IR64 at 16:00 und er HT treatm ent(Fig.6);h ow ever,the percentage of viablely exserted stigm as w as red uced by 1%in SY63,1.8%in LYPJ,1.9%in HHZ,and 1.9%in IR64 at 11:30 und er CK,and by 1.8%in SY63,2.8%in LYPJ,2.8%in HHZ,and 1.9%in IR64 at 16:00 un der CK treatm ent.The p ercentage of viable exserted stigm as w as sim ilar to that of viable hid d en stigm as for a given cultivar und er both CK and HT,respectively at 9:30(w ith a red uction of＜1%,but 5.6%und er HT in IR64 at 9:30).Th ese data indicate that HT treatm ent decreased the viability of exserted stigm as.

Com pared w ith CK treatment,HT treatment reduced the percentage of viable exserted stigm as by 9%in SY63,11%in LYPJ,9%in HHZ,and 15%in IR64;how ever,HT reduced the percentage of viable hidden stigm as by 2%in SY63,2%in LYPJ,4%in HHZ,and 2%in IR64 at 16:00,irrespective of GA3application(Fig.6).Sim ilar responses w ere observed at 11:30.The viability of exserted stigm as w as m ore susceptible to heat stress than that of hidden stigm as.

GA3application did not result in substantial effects on the percentages of viable hidden stigmas at the three time points and viable exserted stigm as at 9:00 and 11:30(Fig.6).How ever,GA3application increased the percentage of viable exserted stigm as by 3%in SY63,3%in LYPJ,5%in HHZ,and 8%in IR64 under HT at 16:00,and 1%,2%,2%,and 2%under CK at 16:00.GA3application had little effect,even in the afternoon under HT treatm ent,and that GA3application w as m ore effective in increasing viability under HT treatment than under CK,especially at 16:00.

3.5.Effects of high temperature and GA 3 application on stigma receptivity

Results for pollen grain deposition on exserted and hidden stigm as in LYPJ and SY63 under CK and HT treatm ents are presented in Fig.7 and Table 2.In any given cultivar,there w as no significant difference in the num bers of pollen grains attached to exserted versus hidden stigm as under either CK or HT treatm ent at 9:30.Values at 16:00 w ere sim ilar to those at 09:30 for hidden stigmas;however,exserted stigmas received m ore pollen grains than hidden stigm as.HT treatm ent significantly reduced the num bers of pollen grains on both types of stigm as at 9:30 and on hidden stigmas at 16:00,how ever,the num ber of pollen grains on exserted stigm as increased at 16:00.

Fig.6-Effects of high tem perature treatm ent and exogenous GA3 application on stigm a viability.Data are presented as m ean±SD(n=4).Vertical bars represent least significant differences at 5%level of significance for com paring treatm ent m eans w ithin tim e scale of the day.CK-H2O,control w ithout GA 3 application;CK-GA,GA 3 application under control condition;HT-H 2O,high tem p erature treatm ent w ithout GA 3 ap plication;HT-GA,GA 3 ap plication under high-tem perature treatm ent.

Fig.7-Im ages show ing pollen recep tivity of stigm a and p ollen germ ination.CK,control;HT,high-tem perature treatm ent.(AD)LYPJund er the control condition,(E-H)LYPJund er high tem p erature,(I-L)SY63 under the control condition,and(M-P)SY63 und er high tem perature.Scale bar=100μm.

Table 2-Effects of high-tem perature treatm ent on stigm a recep tivity and pollen germ ination.

There w as no substantial variation in the num bers of germinated pollen grains per stigma betw een exserted and hidden stigm as under CK or HT treatm ent at 9:30(Table 2).The num bers of germ inated pollen grains on hidden stigm as of the tw o cultivars at 16:00 w ere sim ilar to those at 9:30 under the CK and HT treatments,respectively.How ever,the num bers of germ inated pollen grains on exserted stigm as increased at 16:00 under both CK and HT treatm ents.In addition,HT treatm ent reduced the num bers of germ inated pollen grains on both types of stigm as at 9:30 and 16:00.

Betw een 9:30 and 16:00,there w as no m arked difference in the percentage of germ inated pollen grains on hidden stigm as under the CK and HT treatm ents,and on exserted stigm as under CK treatm ent(Fig.7,Table 2).How ever,the percentage of germ inated pollen grains on exserted stigm as w as dram atically reduced at 16:00 under HT treatm ent in both cultivars,and this reduction was due m ainly to the greater total num ber of pollen grains on the stigm as.

4.Discussion

4.1.Varietal differences of spikelet fertility in response to high temperature treatment

HT treatment reduced rice spikelet fertility significantly during the flow ering stage.Heat-induced injury in LYPJ,IR64,and HHZ appeared m ore severe than that in SY63 and N22(Table 1).Similarly in previous studies[27,32],LYPJand IR64 w ere found to be highly heat-susceptible,and HHZ and SY63 w ere relatively tolerant to HT treatm ent during flow ering.The spikelet fertility of N22 w as not m arkedly reduced by HT treatment in the study(Table 1).Together,these results show that N22 is highly tolerant to heat stress during flow ering in term s of spikelet fertility,a finding consistent w ith previous results[8],SY63 and HHZ are m oderately tolerant to heat stress,and LYPJand IR64 are highly susceptible to heat stress.

It is notew orthy that IR64 show ed high susceptibility to heat stress(the recorded daily m ean tem perature of 35.9°C)in the study(Table 1),a finding consistent w ith that of Fahad et al.[32]at the same experimental site.However,Jagadish et al.[33]reported that IR64 w as highly tolerant at a high m ean am bient tem perature of 38°C.Thus,spikelet fertility response to heat stresses in a given cultivar may depend to some degree on heat intensity and the environm ental conditions of experim ental treatm ents w hen high tem perature is im posed.

4.2.Varietal differences in stigma exsertion and its relationship with heat tolerance

Varietal variations in the percentage of spikelets w ith exserted stigm as w ere observed in rice in our study(Table 1)and previously(Table 3).For the heat-susceptible cultivars under CK and HT,the percentage of spikelets w ith exserted stigm as w as＞80%in LYPJ and＜19%in IR64 in the study.However,the trait value w as relatively low,show ing variation am ong heat tolerant cultivars(Table 1).Thus,our findings show that spikelet fertility and heat tolerance are not alw ays coupled with a high percentage of spikelets with exserted stigm as,and vice versa.These results are consistent w ith those of previous studies(Table 3).They also show that enclosed stigm a is not alw ays associated w ith high spikelet fertility and heat tolerance during flow ering in rice(Tables 1 and 3).Thus,our observations,together w ith prior observations,indicate that the percentage of spikelets w ith exserted stigm as has no causal association w ith high spikelet fertility and tolerance to heat stress in rice.

Spikelets w ith exserted stigm as did not show higher spikelet fertility than spikelets w ithout exserted stigm as under heat stress(Table 1).GA3application augm ented the percentage of spikelets with exserted stigm as and stigma length in rice[34].An increased percentage of spikelets w ith exserted stigm as via application of exogenous GA3did not contribute to spikelet fertility;on the contrary,it resulted in decreases of spikelet fertility to some degree under the HT and CK conditions,especially in heat susceptible cultivars(Table 1).Our results are consistent w ith those of Huang and Zou[35],w ho found that an increased percentage of spikelets w ith exserted stigm as resulting from GA3application show ed no influence on yield and seed setting in LYPJ.These results together suggest that high stigm a exsertion m ay not be associated w ith high spikelet fertility under heat stress in rice.

Considering the opposite responses of spikelet fertility in IR64 to heat stress in our study and reported by Jagadish et al.[33],the other four cultivars w ere used to estim ate the relationship betw een the percentages of spikelets w ith exserted and hidden stigmas and spikelet fertility of whole panicles(Fig.3).Rem arkably,the percentages of spikelets w ith single,dual,and entire exserted stigm as w ere negatively correlated w ith the spikelet fertility of w hole panicles(Fig.3-A-C;Table 3).Combining the observations in the study(Table 1)and in previous reports(Table 3)under high tem peratures over nine cultivars,a sim ilarly negative correlation w as also observed betw een spikelet fertility and percentage of spikelets w ith stigm a exsertion under heat conditions(n=23,r=-0.58,P=0.004).How ever,spikelet fertility w as significantly and positively correlated w ith percentage of spikelets w ith hidden stigmas under CKand HT in our study(Fig.3-D).Weerakoon et al.[36]observed that the temperature inside unopened spikelets w as low er than the am bient air tem perature,ow ing to increased transpiration at low relative hum idity,especially under high-tem perature treatm ents.Collectively,these results indicate that enclosed stigma generally contributes to higher spikelet fertility and heat tolerance during flow ering in rice.

Table 3-Cultivar variation in heat tolerance and percentage of spikelets w ith stigm a ex sertion.

4.3.Stigma receptivity and pollen germination under high temperature and GA 3 application

In this study,anther dehiscence w as delayed to som e degree by HT treatm ent,especially in heat-susceptible cultivars such as LYPJand IR64(Fig.4).In addition to delaying dehiscence,HT treatm ent disturbed the coordinated processes of anther dehiscence,pollen shedding,and stigma elongation,and thus reduced the num ber of pollen grains landing on stigmas[37,38].Despite the delay of anther dehiscence under HT treatment,exserted stigmas showed superior pollen reception,especially under HT treatm ent(Fig.7,Table 2).Stigm a exsertion m ay ensure that sufficient pollen grains are deposited on the stigm a,avoiding poor pollination due to decreased anther dehiscence under heat stress during flow ering,especially later in the day.Exserted stigm as show ed a sim ilar percentage of germ inated pollen grains to that of hidden stigm as in heat-susceptible(LYPJ)and tolerant cultivars(SY63)under HT treatment in the morning(Table 2),but exserted stigm as show ed significantly low er percentages of pollen germ ination in the afternoon.How ever,the percentage of germ inated pollen on hidden stigm as w as not affected or decreased slightly under HT treatm ent(Table 2).Thus,regardless of its benefit for pollen receptivity,stigm a exsertion is not alw ays beneficial for pollen germ ination,and thus m akes no substantial contribution to spikelet fertility under HT conditions.In tomato,the use of stigma exsertion as a criterion for selecting against fruit set at high tem perature is also thought[39]to be m isleading.As discussed above,our results show that stigm a exsertion is not alw ays associated w ith heat tolerance in rice as expressed in stigm a receptivity and pollen germ ination.

A prior study[40]found that HT treatm ent did not substantially affect stigma viability in either heat-tolerant or heat-susceptible cultivars.How ever,rice plants displayed w ithered stigm as under extrem e heat stress[41,42].In the present study,w e exam ined the dynam ic changes in stigm a vigor and pollen viability over the course of one day(Figs.4 and 6).HT treatm ent show ed little effect on the viabilities of hidden stigm as,but it im paired the viabilities of exserted stigm as(Figs.5 and 6),regardless of the heat tolerance of the cultivars used in the study.These results indicate that stigma vigor could be reduced by HT and suggest that heat injury to stigm as m ight be m ore serious w hen high tem perature is increased or duration of high temperature treatment is prolonged.

Exserted stigm as are m ore vulnerable than hidden stigm as to environm ental factors,especially HT stress.The observed air tem perature outside spikelet w as about 2°C higher than that inside spikelets at 38°C[44],suggesting that exserted stigm as m ay be m ore readily dried by HT treatm ent than hidden ones.Drying exudates and absence of peroxidase in stigmas resulted in a relatively reduced number of pollen tubes on the stigm a,and pollen tubes could not be m ade to germ inate on w ilting stigm as[43].The w ilting of stigm as m ay be partially attributed to the increased transpiration rate of spikelets in HT-stressed rice plants[44].Physiologically,HT treatm ent also induced abnorm ality in the distribution of calcium and increased reactive oxygen species in the stigm a,w hich m ight result in disorder in the grow th of pollen tubes[44,45].

Generally,reductions in percentage of viable hidden stigm as w ere low er than reductions in percentage of viable exserted stigm as across four cultivars and high tem perature treatment w ith or w ithout GA3app lication(Fig.6),indicating that viability of the hidden stigm as w as stable relative to that of exserted stigm as,especially in the afternoon.This finding may be explained by the higher tem perature of am bient air than that inside spikelets under HT[36,44].Koike et al.[46]also reported that cleistogam y red uced the detrim ental effect of HT treatment during flow ering in rice.Although stigm a exsertion m ay be advantageous for pollen receptivity (Table 2),these results show the different responses of hid den and exserted stigm as to heat treatm ent,and exserted stigm as are prone to heat stress,suggesting that hidden stigm as have a potential ad vantage in m aintaining their viabilities.

In our study,heat stress reduced the percentage of viable pollen grains by＞33%at 16:00(Fig.4).The largest reductions in the percentage of viable stigm as w ere of about 10%in LYPJ and occurred in exserted stigm as at 16:00(Fig.6).The results are consistent w ith those of Dupuis and Dum as[47],in w hich fem ale m aize tissues appeared m ore resistant to heat stress than m ale organs.Pollen viability also w as red uced to a greater extent as the duration of HT treatm ent w as p rolonged (Fig.4).Pollen viability decreased gradually throughout the day even under natural conditions[48],and p ollen becom e inviable under HT treatm ent(35°C)for＞0.5 h[49].

GAs regulate pollen viability and lack of sufficient active GAs triggers pollen abortion[50].In rice cultivars,a gradual decrease in pollen viability w as accom panied by a decline in endogenous GAs in anthers in heat-stressed plants[51].Under low-temperature conditions,the reduced GAs induced abnorm al enlargem ent of tapetal cells,consequently disrupting pollen developm ent,and application of exogenous GA prom oted norm al pollen developm ent[52].GA signaling in the tapetum may regulate pollen viability by governing processes of tapetum secretory functions and entry into program m ed cell death under abiotic stress[53].How ever,in the present study,exogenous GA3treatm ent induced only an insignificant increase in the percentage of viable pollen grains(Fig.4).Exogenous GA3w as applied to prom ote stigm a exsertion and test w hether an increased percentage of spikelets w ith exserted stigm as prom otes spikelet fertility under heat conditions in this study.The optim al concentration of GA3for stigm a exsertion is 40-60 m g L-1[54].How ever,floral organs respond differently to various GA3concentration.The concentration(60 m g L-1)of GA3w as optim al for stigm a exsertion,but m ay not be optim al for pollen viability in rice plants,as suggested by our results(Fig.4)and previous observations[34].Our results show that increased stigm a exsertion is not beneficial to higher spikelet fertility,regardless of increase in percentage of spikelets w ith exserted stigm as(Table 1)and slight increases in both pollen viability(Fig.4)and the percentage of viable stigm as(Fig.6).It should be noted that spikelets w ith a developing kernel inside were recorded as fertile in this study,spikelet fertility is different from pollen and stigm a viabilities.In addition,even if sufficient pollen grains w ere deposited on exserted stigm as under HT treatment,most of the pollen grains w ere of reduced viability and did not germ inate(Table 2),and stigm a exsertion did not preserve pollen viability.It is thus possible that GA3application exerted different effects on spikelet fertility,pollen and stigm a viabilities.In fact,several factors m ay affect spikelet fertility,including pollen and stigm a viability,anther dehiscence,stigm a exsertion,pollen dissemination(pollination),stigmatic receptivity of pollens,and fertilization.Heat stress m ay influence each process and result in low spikelet fertility.

4.4.Application in rice breeding for improved heat tolerance

Many physiological and m orphological characteristics are often m easured for estim ation of heat tolerance[4,10,28,30,55]and these characteristics may be used to screen genotypes for heat tolerance.Our study has show n that enclosed stigm a generally contributes to higher spikelet fertility and heat tolerance during flow ering in rice and that a high percentage of spikelets w ith exserted stigm as may not be associated w ith high spikelet fertility under heat stress,as discussed above.Phenotypic variation is broadly observed in stigm a exsertion[20,56].Sim ilarly,both percentage of stigm a exsertion and heat tolerance show large differences am ong genotypes(Table 1,Table 3).Thus,screening germ plasm resources for high percentage of spikelets w ith hidden stigm as m ight be useful for im proving heat tolerance and identifying elite genes.Stigm a exsertion in rice species is thought to be controlled by polygenes[22],and a QTL for increased frequency of exserted stigm as has been used in QTL pyramiding for increasing seed production in hybrid rice[56].Thus,heat tolerance can be transferred to rice plants by increasing the percentage of spikelets w ith enclosed stigm as by m arker-assisted selection.

5.Conclusions

Exserted stigm as increased the num ber of pollen grains on stigm as,but spikelets w ith exserted stigm as did not display higher spikelet fertility under heat stress than did spikelets with hidden stigm as.High spikelet fertility w as not always coupled w ith a high percentage of spikelets w ith exserted stigm as in rice,and vice versa.Increasing the percentage of spikelets w ith exserted stigm as by GA3application did not increase spikelet fertility under heat stress.Spikelet fertility of w hole panicles w as negatively correlated w ith the percentage of spikelets w ith exserted stigm as,and positively w ith that of spikelets w ith hidden stigmas.Viability of hidden stigmas show ed less reduction than that of exserted stigm as,suggesting that exserted stigm as are prone to heat stress and that hidden stigm as have an advantage in m aintaining their viabilities.Heat stress treatm ent delayed anther dehiscence,and reduced the viability of exserted stigm as and pollens,leading to low spikelet fertility.In sum m ary,our results suggest that high spikelet fertility does not depend on stigm a exsertion and that enclosed stigma generally contributes to higher spikelet fertility and heat tolerance under hightem perature conditions during flow ering in rice.

Conflict of interest

Authors declare that there are no conflicts of interest.

Acknow ledgm ents

We thank Professor Pravat K.Mohapatra from School of Life Sciences,Sam balpur University,Sam balpur,Odisha,India for his assistance in improving English w riting.This w ork was supported by the National Natural Science Foundation of China(30971707,31361140368),the National Key Research and Developm ent Program of China(2017YFD0300100),the Natural Science Foundation of Jiangsu Province(BK20180537),the China Postdoctoral Science Foundation(2017M621757),and the Priority Academ ic Program Developm ent of Jiangsu Higher Education Institutions.