Wencheng Wng,Kehui Cui,b,*,Qiuqin Hu,Cho Wu,1,Guohui Li,2,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,College of

b HubeiCollaborative InnovationforGrain Industry,Yangtze University,Jingzhou434023,Hubei,China

Keywords:

A B S T R A C T In rice,high-temperature stress(HT)during flowering results in decreased grain yield via a reduction in spikelet fertility;however,the effect of plant water status on spikelet fertility under HT remains unknown.To investigate the relationship between spikelet water status and spikelet fertility under HT,two experiments were performed under temperature-controlled conditions using four genotypes with varying tolerance to HT.Rice plants were exposed to HT for seven consecutive days during the flowering stage under three soil water treatments(soil water potential 0,-20,and-40 kPa),as well as under hydroponic conditions in a separate experiment.HT significantly decreased spikelet fertility,pollen fertility,and anther dehiscence under each of the three water treatments.HT significantly increased the spikelet transpiration rate,and this change was accompanied by a significant decrease in the internal temperature of the spikelets.HT decreased pollen grain diameter in heat-sensitive genotypes.HT had varying effects on the water potential of panicles and anthers but increased anther soluble-sugar concentration.Different aquaporin genes showed different expression profiles under HT,and the expression levels of PIPs for plasma membrane intrinsic proteins and TIPs for tonoplast intrinsic proteins increased in anthers but decreased in glumes.Correlation analyses showed that anther dehiscence and pollen(spikelet)fertility were tightly associated with anther water status,and the expression levels of almost all anther aquaporin genes were significantly correlated with anther dehiscence under HT.In summary,an increased spikelet transpiration rate and decreased internal spikelet temperature were associated with alleviation of the effects of HT in rice genotypes with varying degrees of heat tolerance,and the response of spikelet water status to HT,involving increased total expression of aquaporins and soluble sugar content,thereby improved pollen fertility,anther dehiscence,and spikelet fertility,especially in heat-resistant genotypes.The heat-resistant genotypes N22 and SY63 may adopt different approaches to reduce heat damage.

1.Introduction

Global climate change is increasing the global temperature and the occurrence of extreme high temperatures[1],both of which threaten global food security[2].Rice is a staple food for more than half of the world’s population[3],and global rice production must increase by approximately 1% annually to meet the growing demand for this vital crop[4].High-temperature stress(HT)sharply reduces the grain yield of rice[5],and a 1°C increase in average daily maximum or minimum temperatures during the rice growing season reduced grain yields by 6.2%–10%[6–8].

Spikelet fertility under heat stress is closely associated with ambient temperature[9]and negatively correlated with spikelet temperature[10,11].Plant transpiration is a key physiological process for lowering canopy and organ temperatures,and studies[12,13]have shown the role of panicle and spikelet transpirational cooling in relieving the injurious effects of HT on spikelet fertility.Large temperature differences between rice panicles and the surrounding air,perhaps caused by strong transpirational cooling,have been reported[14,15].These reports suggest that it is possible to relieve heat damage by developing rice varieties with a high transpiration rate.However,the direct relationship of the spikelet transpiration rate with spikelet temperature is not yet established.

Spikelet fertility is dependent primarily on anther dehiscence,pollination,pollen fertility and germination,and fertilization.HT causes abnormal anther dehiscence[16],poor pollen shedding[17],high pollen sterility[18],poor pollen germination,and slow pollen tube growth[19],as well as failed fertilization because there are fewer germinated pollen grains on each stigma[17,20]and pollen tubes do not reach the viable embryo sac[21].Anther dehiscence is thought to be a sequence of water allocation steps[22].Water flows into the anther,resulting in active hydration and inflation of pollen and endothecium[22,23].Cell inflation increases turgor pressure,which can increase tension in the anther wall and eventually rupture the anther stomium[23].Pollen swelling after water absorption is one of the major driving forces for anther dehiscence;in rice,HT inhibits pollen swelling,resulting in poor anther dehiscence[24,25].HT reduces pollen fertility by inducing abnormal pollen development[26–28].HT also disrupts carbohydrate metabolism in the male reproductive organs of crop plants.Insufficient sugar accumulation in pollen leads to reduced pollen fertility;high levels of sugar in anthers and pollen are required by heat-tolerant genotypes to maintain pollen fertility during HT[29–31].However,the underlying associations of anther and pollen water status with anther dehiscence and pollen fertility,especially under HT,await investigation.

Water status is involved in many physiological and biochemical processes involved in flower opening and anther dehiscence[22]and pollen fertility[32–34].Osmotic adjustment is a strategy by which plants tolerate water deficits by accumulating osmolytes and reducing osmotic potential.The accumulation of osmolytes,such as soluble sugars and ions,increases efficient absorption and transport of water and further increases desiccation tolerance[22].HT has been shown to increase concentrations of osmolytes and lower osmotic potential[35].Also,a high K concentration has been observed at the aperture area of mature pollens and the stomium area of mature anthers,where it may regulate rapid pollen swelling and anther dehiscence[36].Few studies have investigated the response of osmotic adjustment in reproductive organs to heat stress in rice.

Aquaporins are membrane channels that facilitate bidirectional water movement across cell membranes,which are involved in many physiological processes[37].Several aquaporins influence water flux in anther and pollen by adjusting the water potential of pollen grains and pollen tube turgor pressure,which is necessary for pollen hydration,anther dehiscence,and pollen tube tip growth[38–43].Information on the roles of aquaporins and aquaporin regulation in the water status of rice anther and pollen,especially under HT,remains limited.

The water status of the reproductive organs of rice is associated with anther development and behavior,as well as with anther dehiscence,pollen germination,and fertility.However,the physiological processes underlying these processes under HT are not well understood.Accordingly,two types of experiments were conducted to investigate water status and its association with anther dehiscence and pollen fertility in the context of water loss,osmotic adjustment,and aquaporin expression in reproductive organs,using heat-sensitive and heat-tolerant genotypes subjected to HT.

2.Materials and methods

2.1.Experiment 1(soil pot experiment with high temperature and water deficit treatments)

2.1.1.Crop cultivation

To establish the relationship between water status and spikelet fertility under different water levels,a soil pot experiment was conducted during the rice-growing season at Huazhong Agricultural University,Wuhan,China.Two heat-sensitive genotypes(LYP9 and TXZ25)and two heat-tolerant genotypes(SY63 and N22)were selected based on results of previous studies[17,18,44].Germinated seeds were sown in plastic trays containing sandy loam soil,and three uniform 23-day-old seedlings were transplanted in a triangular pattern in 12-L plastic pots(24.5 cm in diameter,25.5 cm in height)filled with 10 kg of sandy loam soil.The soil had the following properties:pH 6.6,10.5 g kg-1organic matter,0.1 g kg-1total N,8.1 mg kg-1Olsen P,and 113.8 mg kg-1exchangeable K.Nitrogen fertilizer was applied at a rate of 2 g N per pot in the form of urea;40% nitrogen fertilizer was applied as a basal dose,and the residual N was equally split as topdressing fertilizer at the middle tillering stage and the panicle initiation stage.In addition,1.33 g P and 1.67 g K per pot were applied as basal fertilizer in the form of KH2PO4.The basal fertilizers were thoroughly mixed with the soil in each pot,and the topdressing fertilizers were dissolved in water and applied with irrigation.Pests and diseases were controlled with chemical pesticides.

2.1.2.Experimental treatments

All potted rice plants were grown under natural conditions until exsertion of the panicle tip from the main stem,after which the plants were moved to two temperature-controlled greenhouses for seven consecutive days.The experiments were arranged in a split-split plot design with temperature treatments as the main plots,water irrigation treatments as the subplots,and genotypes as the sub-subplots,with three replicates.Each pot was manually rotated 90 degrees clockwise every two days to avoid positional effects.A central auto-controlled system(Auto-Greenhouse Monitoring and Data Management System,Version 3.00,Auto,Beijing,China)was used to control the temperature and relative humidity in each greenhouse,as described previously[18,45].The two temperature treatments were a high daytime temperature treatment(HT),in which plants were imposed to high temperature conditions from 7:00 to 19:00,and a control treatment with favorable temperatures for rice plant growth during the entire day.The temperature for the control greenhouse was set at 25 °C from 5:00–6:00,28°C from 07:00–09:00,32°C from 10:00–15:00,28°C from 15:00–19:00,and 27°C from 20:00–21:00.For the HT greenhouse,the temperature was set at 25 °C from 5:00–6:00,33 °C from 07:00–09:00,36 °C from 10:00–11:00,38 °C from 12:00–14:00,34°C from 15:00–18:00,and 27°C from 20:00–21:00.The air temperature in each greenhouse was controlled by a central autocontroller and gradually approached the next set temperature.Relative humidity(RH%)was set at 70% during the daytime for both temperature treatments.The actual air temperature and RH%inside the greenhouses were measured every 10 min at 5 cm above the canopy using standalone sensors(HOBO,H08-003-02,Onset Computer Corporation,Bourne,MA,USA).The mean daytime temperature of the HT greenhouse was 36.4±0.5 °C,and the mean daytime temperature of the control greenhouse was 30.2±0.8 °C.The mean daytime RH% of the HT greenhouse was 67.1%±2.8%,and the mean daytime RH% of the control greenhouse was 68.5%±5.9%.Details of the temperature and humidity inside the greenhouses are shown in Table S1.

Three water regimes were applied for each temperature treatment:well-watered treatment(W1),moderate drought treatment(W2),and severe drought treatment(W3).Soil tension meters were installed in the W2 and W3 pots at a depth of 15 cm to monitor soil water potential.For W2 and W3,pots were moved into the greenhouses after irrigation was stopped for 1 day or 2 days,respectively,when the soil water potential reached threshold values of approximately-20 and-40 kPa,respectively.After the pots were moved to the greenhouses at 7:00,the water depth at the soil surface was maintained at 1–2 cm(ψsoil=0 kPa)under W1,and the soil water potential was maintained at the threshold values of-20 kPa under W2 and-40 kPa under W3.To maintain soil water potential and avoid more severe drought,the W2 and W3 pots received 900 mL and 450 mL tap water,respectively,at 12:00,and an additional 450 mL and 275 mL,respectively,at 17:00.From 8:00 to 12:00,the soil water potential was approximately-20 kPa under W2 and-40 kPa under W3.The flag leaf potential and relative soil water content were measured after the water treatments to determine whether plants varied in their responses to the three water treatments described above(Fig.S1).

After seven days of exposure to the temperature and water treatments,the remaining potted plants were placed under natural conditions and irrigated every day(with an approximately 2-cm water layer above the soil surface)until they reached maturity,at which time experiments were conducted to assess spikelet fertility.

2.2.Experiment 2(hydroponic experiment with high-temperature treatment)

2.2.1.Crop cultivation

A hydroponic experiment was performed with the same cultivars used in experiment 1.Three fifteen-day-old seedlings were transplanted into each pot(24.5 cm in diameter,25.5 cm in height).Experiments were arranged in a split-plot design with temperature treatments as the main plots and genotypes as the subplots,with three replicates.Ten liters of Yoshida’s nutrient solution[46]was applied to each pot,in which the concentrations of N,P,and K were 1.43,0.32,and 0.51 mmol L-1,respectively.The nutrient solution was changed every 3 days,and pH was adjusted to 5.5 using a solution of 1 mol L-1HCl or NaOH.Pests and diseases were controlled with chemical pesticides.

2.2.2.Temperature treatment

Rice plants were grown under natural ambient temperature conditions until exsertion of the panicle tip from the main stem,at which point the pots were divided into the HT and control groups and moved to two temperature-controlled growth chambers(Model GR48,Conviron,Controlled Environments Limited,Winnipeg,MB,Canada).Daytime(from 7:00–19:00)/nighttime(19:00–7:00)temperatures were set at 38/28 °C for HT and 32/28 °C for the control,and relative humidity was set at 75%.The photosynthetic photon flux density was set to 600 μmol m-2s-1with 12/12 h daytime/nighttime periods in each growth chamber.HOBO sensors were used to record air temperature and RH%5 cm above the canopy.The recorded mean daytime temperature was 37.7±0.2 °C for HT and 32.2±1.3 °C for the control,and the mean nighttime temperature was 27.9±0.3 °C for HT and 27.6±1.5 °C for the control.The treatments were continued for seven consecutive days,after which the plants were placed under natural conditions until they reached maturity,at which time experiments were conducted to determine spikelet fertility.

2.3.Sampling and measurements

2.3.1.Anther dehiscence,pollen diameter,and pollen and spikelet fertility

Anther dehiscence was measured in experiment 1 following Wu et al.[18].Anthers were collected at flowering by scraping the opening spikelets with a glass slide,and all anthers and dehisced anthers were carefully counted.The collection process was repeated four times for each replicate.A dehisced anther was defined as an anther with open apical and/or basal pores,and the anther dehiscence rate(%)was calculated as the ratio of the number of dehisced anthers to the total number of anthers.

Pollen fertility was measured in experiment 1 by counting stained pollen grains using 1% iodine-potassium iodide solution(I2-KI)[47].Undehisced anthers were squeezed to release pollen grains onto a slide,following by staining with one drop of 1%I2-KI and examination at 10×magnification using an inverted fluorescence microscope(Nikon Corporation,Inc.,Tokyo,Japan).Three observations were performed for each slide,and five slides per replicate were measured.For each observation,a microscope image was photographed.Fully stained pollen grains were considered fertile and partly stained or unstained grains as unfertile.Pollen fertility(%)was calculated as the ratio of the number of fertile grains to the total number of grains.The diameters of 100 pollen grains were measured with Image J software(National Institutes of Health,Bethesda,MD,USA)based on the microscopy images,and the mean diameter(μm)was calculated.

To assess spikelet fertility in experiments 1 and 2,all grains were detached from harvested panicles from three plants grown in three pots,and then soaked in tap water to separate filled grains from other grains(empty and partially filled grains).Empty grains were separated from partially filled grains by winnowing and then re-checked carefully by manually pressing the grains between the forefinger and thumb.Spikelets with a kernel were classified as fertile and those without a kernel as sterile.Spikelet fertility(%)was calculated as the ratio of the total number of filled and partially filled grains to the total number of grains.Anther dehiscence,pollen fertility,and spikelet fertility were used to assess hightemperature resistance.

2.3.2.Water content and water potential of anthers and panicles

To measure anther water content and anther water potential in experiment 1,three panicles from the main stem and large stems were cut on the flowering day for each replicate during the period from 8:00 to 9:00.The anthers of spikelets that were unopened,but ready to open on the day of the experiment,were collected into a 2 mL Eppendorf tube on ice.A spikelet was expected to open soon when the anthers were at the middle or top position inside it.The fresh weight of the anther samples(～100 mg)was determined immediately,and dry weight was obtained after oven-drying the samples for 48 h at 80°C to constant weight.Anther water content(%)was calculated as the ratio of the difference between fresh and dry weights to the fresh weight.Additional anther samples were used immediately for measurement of water potential(MPa)with a dew point psychrometer(WP4C Dewpoint PotentiaMeter,Decagon Devices Inc.,Pullman,WA,USA).Panicle water potential(MPa)was measured using a pressure chamber(Model 3000,Soil Moisture Equipment Corp.,Santa Barbara,CA,USA)after panicles were cut 2 cm below the panicle node on the flowering day.The water content of the panicles in experiment 2 was also determined.Panicles were cut 2 cm below the panicle node on the flowering day,after which the fresh weight of the panicles was determined immediately,and dry weight was obtained after oven-drying the samples for 72 h at 80 °C to constant weight.Panicle water content(%)was calculated as the ratio of the difference between fresh and dry weights to the fresh weight.

2.3.3.Spikelet temperature

In experiment 2,four individual unopened spikelets in the middles of four panicles were used forin vivomeasurement of the air temperature outside and inside the spikelets for each replicate,using a thermocouple device(AZ Instrument Crop.Ltd.,Taichung,Taiwan,China)at 12:00 h on the flowering day.The thermocouple sensor was placed 2 cm from the side of each spikelet to measure the external temperature of the spikelet and then inserted into the spikelet to measure the internal temperature.The temperature difference(°C)between the outside and inside of the spikelet was recorded.

2.3.4.Spikelet transpiration

The spikelet transpiration rate of the plants in experiment 2 was measured.All leaves of each plant growing in each pot were cut on the flowering day,and the cut points,stems and pot surface were wrapped with plastic film to avoid water loss.The potted plant was weighed every 30 min for 3 h,after which all spikelets were collected for measurement of spikelet surface area using WinRhizo Ver 2003a(Regent Instruments Inc.,Quebec,Canada).The slope of the linear regression of water loss on time was determined,and the spikelet transpiration rate(mg cm-2min-1)was calculated as the ratio of the slope to the spikelet surface area.

2.3.5.Anther soluble-sugar concentration

The soluble-sugar concentration of the anthers(experiment 1)was determined by the anthrone colorimetry method[46].A sample consisting of approximately 50 mg of oven-dried anthers in powder form was subjected to soluble sugar extraction by incubation with 500 μL of 80% aqueous ethanol for 30 min at 80 °C.The mixture was centrifuged at 6000 r min-1for 5 min,after which the supernatant was transferred to another tube.This extraction process was repeated three times,and all supernatants were combined and mixed thoroughly.Next,400 μL of the extract was used for determination of soluble sugar content with 2 mL anthrone reagent by colorimetric assay.The optical density of each sample was measured at 620 nm using a microplate reader(SpectraMax i3x,Molecular Devices,San Jose,CA,USA).The concentration of soluble sugar was expressed as mg glucose g-1dry weight by comparison with a glucose standard curve.

2.3.6.Anther potassium concentration

A sample consisting of approximately 50 mg of oven-dried anthers from experiment 1 was digested with 5 mL of concentrated sulfuric acid,after which the digestion liquid was diluted to 25 mL with distilled water for measurement of anther potassium concentration using a flame photometer[46].The potassium concentration(mg g-1dry weight)was calculated by comparison with a potassium standard curve.

2.3.7.Relative expression of aquaporin genes in glumes and anthers

On the flowering day,panicles were detached from the main stem and large stems of three individual plants and immersed immediately in liquid nitrogen,after which they were stored at–80 °C.The anthers from four panicles(experiment 1)and glumes from another four panicles(experiment 2)were collected into separate centrifuge tubes floating on the surface of liquid nitrogen,after which they were ground into powder with liquid nitrogen to prepare them for extraction of total RNA using the Trizol Kit Reagent(Invitrogen,Carlsbad,USA)according to the manufacturer’s instructions.After RNA purification using a DNaseIKit(Invitrogen)and synthesis of first-strand cDNA using M-MLV reverse transcriptase(Invitrogen),quantitative real-time PCR(qRT-PCR)was performed using gene-specific primers and a FastStart Universal SYBR Green Master(ROX)kit(Roche,Basel,Switzerland)according to the manufacturer’s protocols on an Applied Biosystems QuantStudio 6 Flex Real-Time PCR System(Applied Biosystems,Life Technologies,Carlsbad,CA,USA).The transcription levels of 10PIPsfor plasma membrane intrinsic proteins and nineTIPsfor tonoplast intrinsic proteins,and 18 srRNA(internal control)were determined.The primer sequences are listed in Supplementary Table S2.The relative quantification method(2-ΔΔCT)[48]was used to calculate the relative expression levels of genes.

2.4.Data analysis

Statistix 9.0(Analytical Software,Tallahassee,FL,USA)was used for statistical analysis.The least significant difference test was used to determine the significance of differences among treatments at the 5% level.ANOVA was used to estimate the interactions among temperature treatment,water treatment,and genotype.

3.Results

3.1.Spikelet internal temperature

The air temperature outside the spikelets was approximately 30 °C under the control treatment and 38 °C under the HT treatment(Table 1).The air temperature inside was lower than that outside the spikelets.For LYP9,TXZ25,SY63,and N22 plants under the HT treatment,the temperature differences between the outside and inside of the spikelets were 1.1,1.1,1.5,and 2.3 °C,respectively;however,the temperature difference was only 0.1–0.5 °C for these four genotypes under the control treatment.The differences between the spikelet internal and external temperatures for SY63 and N22 plants were significantly greater than those for LYP9 and TXZ25 plants(Table 1).

3.2.Spikelet fertility,anther dehiscence,and pollen fertility

Under well-watered conditions,HT significantly decreased spikelet fertility by 46.1% in LYP9,28.3% in TXZ25,6.1% in SY63,and 9.5% in N22,compared with the control plants(Table 2).Similar reductions in spikelet fertility under HT were observed under moderate drought conditions(43.0% in LYP9,27.7% in TXZ25,19.3% in SY63,and 18.5% in N22)and severe drought conditions(45.0% in LYP9,37.1% in TXZ25,27.8% in SY63,and 16.3% in N22).Under HT,the decreases in spikelet fertility across the three water treatments in LYP9 and TXZ25 were much greater than those observed in SY63 and N22.

HT treatment reduced anther dehiscence by 54.4% in LYP9,56.1% in TXZ25,44.9% in SY63,and 30.8% in N22 under wellwatered conditions,by 50.1%,49.4%,49.5%,and 36.9%,respectively,under moderate drought conditions,and by 46.6%,41.6%,61.1%,and 64.8%,respectively,under severe drought conditions.

HT treatment reduced pollen fertility by 51.9%in LYP9,19.8%in TXZ25,17.0%in SY63,and 15.3%in N22 under well-watered conditions.Similarly,significant reductions of 31.4% in LYP9,23.2% in TXZ25,16.8%in SY63,and 32.5%in N22 were observed under moderate drought conditions,and significant reductions of 56.3%,40.5%,19.6%,and 32.3%,respectively,were observed under severe drought conditions.

Generally,the percentage reductions in spikelet fertility,anther dehiscence,and pollen fertility in LYP9 and TXZ25 were greater than those observed in SY63 and N22 under HT,especially under well-watered conditions(Table 2).Significant interactions were observed among genotype,water regime,and temperature treatment(except between water and temperature treatment),for pollen and spikelet fertility.

Table 1Air temperature outside and inside the spikelets of four genotypes under high-temperature treatment and well-watered treatment.

Table 2Anther dehiscence,pollen fertility,and spikelet fertility of four genotypes under different temperature and water conditions.

3.3.Pollen grain diameter and anther water content

Compared with the control treatment,HT significantly reduced pollen grain diameter by 9.0% in LYP9 and 4.9% in TXZ25,but did not affect the pollen grain diameter of SY63 and N22,under well-watered conditions(Fig.1A).Under moderate drought conditions,HT significantly reduced pollen grain diameter in LYP9(6.4%),TXZ25(4.1%),and SY63(4.9%),but did not affect the pollen grain diameter of N22(Fig.1B).Under severe drought conditions,HT significantly reduced the pollen grain diameter of all four tested genotypes(Fig.1C).

For the four tested genotypes,there was no significant difference in anther water content between the control and HTtreatments under either well-watered or moderate drought conditions(Fig.1D,E);HT significantly reduced anther water content in LYP9,TXZ25,and N22 under severe drought treatment,while that of SY63 was not significantly affected(Fig.1F).

3.4.Anther and panicle water potential

In comparison with the control treatment,HT reduced anther water potential by 8.6% in LYP9,10.9% in TXZ25,10.3% in SY63,and 18.2% in N22 under well-watered conditions(Fig.2A),and by 15.8%in LYP9,11.9%in SY63,and 34.0%in N22 under moderate drought conditions,under which the potential of TXZ25 remained unchanged(Fig.2B).Under severe drought conditions,HT significantly reduced the anther water potential in TXZ25(21.6%),SY63(29.5%),and N22(26.2%),but not in LYP9(7.3%)(Fig.2C).Under both control and HT treatments,anther water potential gradually decreased as the water deficit became more severe.

In comparison with the control treatment,HT reduced panicle water potential by 7.2% in LYP9,11.0% in TXZ25,4.6% SY63 and 31.6% in N22 under well-watered conditions(Fig.2D),by 11.5%in LYP9,20.7% in TXZ25,11.0% in SY63,and 27.1% in N22 under moderate drought conditions(Fig.2E),and by 28.7% in LYP9,59.0%in TXZ25,8.3%in SY63 and 7.7%in N22(Fig.2F)under severe drought conditions.The panicle water potential gradually decreased as the water deficit became more severe under both temperature treatments.

3.5.Spikelet transpiration and panicle water content

Fig.1.Pollen grain diameter(A,B,C)and anther water content(D,E,F)of four genotypes under different temperature and water conditions.Values are shown as mean±SEM;different lowercase letters indicate significant differences among various combinations of genotype and temperature treatment at P＜0.05 under identical water conditions.HT represents high-temperature treatment.W1,W2,and W3 are well-watered(ψsoil=0 kPa),moderate drought(ψsoil=-20 kPa),and severe drought(ψsoil=-40 kPa)conditions,respectively.

In comparison with the control treatment,HT significantly increased the spikelet transpiration rate in LYP9(100%),TXZ25(139%),SY63(128%),and N22(65%)(Fig.3A).The absolute values of the spikelet transpiration rates of N22 and SY63 were higher than those of LYP9 and TXZ25 under HT.Panicle water content significantly increased in LYP9 and SY63 under HT,but did not significantly increase in TXZ25 and N22(Fig.3B)in comparison with the control treatment.Under HT,the panicle water content of N22 was higher than that of LYP9,TXZ25 and SY63(Fig.3B).

3.6.Anther soluble-sugar and potassium concentrations

In comparison with the control treatment,HT significantly increased the anther soluble sugar concentration by 39% in SY63 and 86% in N22,but did not significantly increase it in LYP9 and TXZ25(Fig.3C).The absolute values of anther soluble sugar concentration in SY63 and N22 were greater than these in LYP9 and TXZ25 under HT(Fig.3C).HT did not substantially change the anther potassium concentration in any of the four genotypes(Fig.3D).

3.7.Gene expressions of aquaporins in anthers and glumes

Overall,HT increased the relative expression levels ofPIPsandTIPsin the anthers(Fig.4A).The responses of the relative expression levels ofPIPsandTIPsin the anthers to HT were genotypedependent.There was no substantial difference in the relative expression levels of all genes in LYP9 between the HT and control treatments.In comparison with the relative expression levels measured in the control plants,TXZ25,SY63,and N22 showed significant increases in the levels of four,four,and 11 genes,respectively,in plants subjected to HT treatment.The highest expression levels of most aquaporin genes in SY63 and N22 were greater than those observed in LYP9 and TXZ25 under HT(Fig.4A).

High temperature reduced the relative expression levels of individualPIPsandTIPsin the glumes(Fig.4B).Compared with expression under the control treatment,the expression levels of nine genes were lower under HT in N22,whereas four genes showed higher expression levels in LYP9.Generally,under HT,greater declines in the relative expression levels of individualPIPsandTIPswere found in N22 in comparison with those in LYP9,TXZ25 and SY63(Fig.4B).

In comparison with the control treatment,HT significantly increased the total expression levels of all antherPIPsin SY63 and N22,but not in LYP9 and TXZ25(Fig.S2A).Similarly,HT significantly increased the total expression levels of all antherTIPsin N22,but not in LYP9,TXZ25 or SY63(Fig.S2B).There were no marked differences in the total expression levels of glumePIPsbetween LYP9 and SY63 plants subjected to the control and HT treatments;but HT significantly reduced the total expression levels of allPIPsin TXZ25 and N22(Fig.S2C).High temperature significantly reduced the total expression levels of all glumeTIPsin SY63 and N22,but not in LYP9 and TXZ25(Fig.S2D).

3.8.Relationships among temperature difference,spikelet transpiration rate,panicle water content,and fertility

Fig.2.Anther water potential(A,B,C)and panicle water potential(D,E,F)of four genotypes under different temperature and water conditions.Values are shown as mean±SEM;different lowercase letters indicate significant differences among various combinations of genotype and temperature treatment at P＜0.05 under identical water conditions.HT represents high-temperature treatment.W1,W2,and W3 are well-watered(ψsoil=0 kPa),moderate drought(ψsoil=-20 kPa),and severe drought(ψsoil=-40 kPa)conditions,respectively.

The temperature difference between the outside and inside of spikelets was significantly and positively correlated with the spikelet transpiration rate(r=0.71,P＜0.01,n=12)and panicle water content(r=0.87,P＜0.01,n=12)under the HT treatment,and no significant correlation of the temperature difference with spikelet transpiration rate(r=0.55,P＞0.05,n=12)or panicle water content(r=0.47,P＞0.05,n=12)was observed under the control treatment.Spikelet fertility was significantly and positively correlated with the difference between the internal and external spikelet temperatures,spikelet transpiration rate,and panicle water content under the HT treatment(Fig.5).Similarly,spikelet fertility was positively and significantly correlated with spikelet transpiration rate and panicle water content under the control treatment,but not with the temperature difference(Fig.5).

3.9.Relationships among pollen-and water status-related traits

Pollen grain diameter was positively and significantly correlated with anther water content,anther water potential,and panicle water potential(Table 3).Anther dehiscence and pollen fertility were positively and significantly correlated with pollen grain diameter,anther water content,anther water potential,and panicle water potential,respectively(Table 3).

Spikelet fertility was positively and significantly correlated with anther dehiscence under the control(r=0.85,P＜0.01,n=36 across 4 genotypes,3 water treatments,and 3 replicates)and HT treatments(r=0.80,P＜0.01,n=36).A similar correlation was found between pollen fertility and spikelet fertility under the control(r=0.93,P＜0.01,n=36)and HT treatments(r=0.91,P＜0.01,n=36).

4.Discussion

4.1.Genotypic differences in spikelet fertility under high temperature

The finding that HT decreased spikelet fertility under different water conditions(Table 2)is consistent with those of previous studies[13,17,21].In our study,pollen fertility and anther dehiscence were strongly disturbed by HT(Table 2).The reductions in spikelet fertility caused by HT were in agreement with observed declines in pollen fertility and anther dehiscence under different water conditions(Table 2),and also supported by the high correlations of spikelet fertility with these two traits.Our results,together with the previous studies[17,25,49],suggest that HT increased spikelet sterility by reducing pollen fertility and anther dehiscence.

Genotypic variation in spikelet fertility was observed among the four genotypes,and SY63 and N22 had more fertile spikelets under HT,regardless of water conditions(Table 2).These findings are consistent with those of Jagadish et al.[50]and Wu et al.[18].Both genotypes showed small reductions in spikelet fertility due to HT(Table 2).Thus,HT slightly reduced the spikelet fertility of SY63 and N22,which are considered to be heat-resistant genotypes.

4.2.Spikelet water status is responsible for reduced anther dehiscence and pollen fertility

In the present study,HT significantly reduced pollen grain diameter,especially in LYP9 and TXZ25(Fig.1A–C),a finding consistent with reduction in anther dehiscence under HT(Table 2).Similarity,Matsui et al.[24]observed that the time of anther dehiscence coincided well with the time that pollen grains reached their maximum diameter in excised anthers.In our study,anther dehiscence showed positive correlations with pollen grain diameter,anther water content,anther water potential,and panicle water potential(Table 3).The previous studies[39,51]showed that anther water status was closely associated with pollen size and anther dehiscence under water-stressed conditions.We also observed positive relationships between anther dehiscence and anther water potential under W2(r=0.59,P＜0.05)and W3(r=0.77,P＜0.01)under the control conditions.These and previous findings suggest that decreased anther dehiscence may be attributed partly to poor anther water status and poor pollen swelling induced by HT and by water deficit treatment.

Fig.3.Spikelet transpiration rate(A),panicle water content(B),anther soluble sugar concentration(C),and anther potassium concentration(D)of four genotypes under hightemperature treatment and well-watered treatment(W1).Values are shown as mean±SEM;different lowercase letters indicate significant differences among various combinations of genotype and temperature treatment at P＜0.05 under identical water conditions.HT represents high-temperature treatment.

Pollen dehydration resulted in pollen sterility[32,34];for example,the pollen fertility ofCucurbita pepodeclined sharply from 85% to 13% when the pollen water content changed from 43% to 8.5%[52].In our study,HT reduced pollen grain diameter(Fig.1A–C),suggesting that HT inhibited pollen swelling.In addition,pollen fertility was positively correlated with pollen grain diameter,anther water content,and water potential(Table 3).Thus,decreased pollen fertility may also be attributed partly to poor pollen water status and poor pollen swelling induced by HT.

4.3.The relationship between spikelet transpiration and spikelet fertility

The internal spikelet temperature was reduced dramatically by 1.1 to 2.3 °C in comparison with the ambient temperature under HT,but by only 0.1 to 0.5°C under the control treatment(Table 1).Transpirational cooling is a key strategy for reducing organ temperatures[53].In contrast to previous studies focused on leaf or canopy transpiration[10,11,14],we directly measured the spikelet transpiration rate and found that HT increased the spikelet transpiration rate and panicle water content,especially in heat-tolerant genotypes(Fig.3A and B).Simulation modeling[12]also showed that the temperature differences between the ambient air and inside the spikelet can increase up to 12°C as the spikelet transpiration rate increases under higher temperatures and higher vapor pressure deficits.In contrast,we found that the spikelet transpiration rate was tightly correlated with the temperature difference between the outside and inside of the spikelet and panicle water content,especially under HT.High panicle transpiration conductance can increase heat avoidance capacity by lowering panicle temperature[15];a decrease in panicle conductance led to an increase in panicle temperature under combined heat and water stresses[54].These results show that a high spikelet transpiration rate can effectively lower the internal spikelet temperature under HT,especially when plants are well watered.As shown in Fig.5,spikelet fertility was tightly correlated with the internal/external spikelet temperature difference,spikelet transpiration rate,and panicle water content,especially under HT.Thus,transpirational cooling due to high spikelet transpiration may alleviate heat injury to spikelet fertility.

4.4.An increased soluble sugar concentration is beneficial for heat tolerance

Fig.4.Relative expression of PIPs and TIPs in the anthers(A)and glumes(B)of four genotypes under high-temperature treatment and well-watered treatment(W1).Each grid with different colors indicates the relative expression of a given gene in a given genotype under control or high-temperature treatment(HT);different lowercase letters indicate significant differences among various combinations of genotype and temperature treatment at P＜0.05 under well-watered conditions.

Fig.5.Correlations of the temperature difference between outside and inside spikelets with(A),spikelet transpiration rate(B)and panicle water content(C)with the spikelet fertility of four genotypes under high-temperature treatment and well-watered treatment(W1).ns,not significant;*,P＜0.05,**,P＜0.01(n=12 across 4 genotypes,and 3 replicates).

Fertile pollen grains contain high levels of carbohydrates and other solutes such as K+,which are useful for increasing water absorption and retention capacity[36,55].In our study,HT increased the soluble-sugar concentration in the anthers of rice plants(Fig.3C)and this change was accompanied by reduced panicle and anther water potentials(Fig.2).This finding is consistent with the observation of Das et al.[56]that high-temperature stress increased the accumulation of soluble sugars in pollen.Wahid andClose[35]also found that the relative water content of sugarcane leaves showed an initial loss under heat stress and then recovered to nearly the control level over time in a process that was attributed to increased soluble-sugar abundance and reduced water potential.These findings suggest that HT may change tissue water status via osmotic adjustment.Under the control treatment,although we observed significant correlations of anther soluble sugar concentration with anther water content and pollen grain diameter only under the control treatment,the anther solublesugar concentration was correlated with three heat-resistance traits under HT(Table S3).These findings are in accord with the report of Li et al.[30]that high anther carbohydrate content was associated with high anther dehiscence percentage,pollen germination rate,and spikelet fertility in rice.These findings suggest that enriched carbohydrates should be beneficial for heat resistance by increasing anther dehiscence and pollen fertility under HT.

Table 3Correlations among pollen-and water status-related traits under control and high-temperature(HT)treatments.

As a vital resource for structural components and energy,sugars are critical for pollen production and fertility[29].We found increased anther soluble sugar content(Fig.3C)and positive correlations between anther sugar and fertility under HT(Table S3).Ensuring sufficient carbohydrate accumulation in anthers and pollen might be an approach for improving thermotolerance under HT,also in accord with the results of previous studies[30,31].

The potassium(K+)concentration at the aperture area of mature pollens and the stomium area of mature anthers was shown[36]to be involved in rapid pollen swelling and anther dehiscence,indicating that anther dehiscence is partly due to K+-driven pollen water absorption and swelling.Our study did not identify an effect of HT on anther K+concentration(Fig.3D)or a relationship between K+accumulation and heat resistance under HT(Table S3).However,anther K+concertation was positively and significantly correlated with anther water content and pollen grain diameter under HT(Table S3).Thus,potassium may play a role in the regulation of anther and pollen water content.The physiological relationship between K concentration,pollen swelling,fertility,and anther dehiscence awaits further investigation.

4.5.The relationship between aquaporin expression and heat tolerance

Although researchers[57,58]have investigated aquaporin expression and function in root,leaf,anther and developing grain in rice under natural or heat-stressed conditions,there is limited information about the relationship between aquaporin expression and water status of anthers under HT.In the present study,the expression levels of most of the 19 investigated aquaporin genes were increased by HT in the anthers,resulting in increased total expression of allPIPs(51% in LYP9,36% in TXZ25,243% in SY63,and 69% in N22)and of allTIPsin each of the genotypes except LYP9(–14%,28%,25%,and 66%),especially in SY63 and N22(Figs.4,S2).These results suggest that HT induced expression of aquaporins,which may facilitate water transport to and from the anthers and pollen under HT.In contrast,HT reduced the expression of mostPIPsandTIPsin the glumes,resulting in reduced total expressions of allPIPsandTIPsin each of the tested genotypes with the exception of LYP9(Figs.4,S2).A high transpiration may result in feedback inhibition of aquaporin expression[59].When the spikelet transpiration rate was greatly increased by HT(Fig.3A),it is possible that expression of aquaporins was down-regulated to avoid excessive water loss from glumes exposed to high temperature.The aquaporin expression profiles of rice leaves,roots,spikelets,anthers,and other organs differ[57,60].Aquaporin genes in the anthers and glumes displayed nearly opposite responses to HT in our study,especially in SY63 and N22.The functions of aquaporins in anthers and glumes under heat stress merit further investigation.

Positive relationships were observed[61,62]between aquaporin expression and leaf transpiration rate/stomatal conductance.We observed that the expression levels of some individual aquaporin genes and the total expression levels of allPIPsandTIPsin glumes were significantly and positively correlated with panicle water content and spikelet transpiration rate(Table S4),suggesting that increased expression of aquaporins is required for efficient water transport of glumes,which may be useful for cooling spikelets and increasing fertility by increasing the spikelet transpiration rate(Fig.3A)and panicle water content(Fig.3B).Anther dehiscence was tightly associated with pollen swelling driven by water absorption,and was delayed or prevented by the absence of anther PIP2 protein[38,39].Interestingly,the expression levels of almost all anther aquaporin genes were significantly correlated with anther dehiscence under HT in our study(Table S5).These findings indicate that anther aquaporins play a role in anther dehiscence.

The expression of most aquaporinPIPgenes was not consistent with changes in plant transpiration in transformed maize lines[63].A similar situation also occurred in our study,as shown by the inconsistent expression patterns of various genes,inconsistent changes between gene expressions and water content in response to HT(Figs.1,4),and different correlations among aquaporin expression and water status in glumes and anthers(Tables S4,S5).These findings show that aquaporin gene expression has a complex relationship with water flow.For this reason,it is difficult to identify the functions of individual aquaporins in a given physiological process such as anther dehiscence or transmembrane water transport[64].

4.6.Genotype responses to high temperature stress

As described previously,compared with the two heat-sensitive genotypes(LYP9 and TXZ25),the two heat-tolerant genotypes(SY63 and N22)had higher values of and smaller reductions in the three traits selected for estimation of heat resistance(on average by 38%in tolerant genotypes vs.55%in sensitive genotypes for anther dehiscence,16% vs.36% for pollen fertility,and 8% vs.37%for spikelet fertility)under HT(Table 2).However,SY63 and N22 had greater spikelet transpiration rates and higher panicle water contents(Fig.3A and B).It is thus possible that strong transpirational cooling in SY63 and N22 increased the difference between the internal and external spikelet temperatures(Table 1).Differences among cultivars with respect to the difference between internal and external spikelet temperatures were observed previously[65].Jung et al.[9]also reported that a lower spikelet temperature increased spikelet fertility.A lower internal spikelet temperature might mitigate the negative effects of heat stress on spikelet fertility in SY63 and N22 under HT(Tables 1,2).Thus,a large temperature difference achieved by strong spikelet transpirational cooling may be partially responsible for the heat resistance of SY63 and N22 plants(Fig.6).

Although HT reduced the expression levels of glume aquaporins in the four tested genotypes,the expression levels of the heattolerant genotypes,especially N22,were higher than those of the heat-sensitive plants(Figs.4,S2).These high aquaporin expression levels may facilitate spikelet transpiration(Table S4),lower temperature inside spikelets(Table 1),and greater heat resistance(Fig.6).In contrast,the heat-tolerant genotypes,especially N22,showed higher values and large increases in the gene expression levels of anther aquaporins than the heat-sensitive genotypes(Figs.4,S2).Similarly,increased expression of aquaporin genes under HT was reported in heat-resistant cultivars of raspberry[66].The total expression levels of antherPIPswere increased by a greater magnitude in the heat-tolerant genotypes than in heatsensitive genotypes(Fig.S2A),and this difference was also associated with stable pollen diameter(Fig.1A,B).These findings suggest that increased expression of anther aquaporins facilitates water absorption and may support anther dehiscence,pollen fertility,and spikelet fertility in heat-tolerant plants(Fig.6).

The two heat-tolerant genotypes showed higher performance and larger increases in soluble-sugar concentration(mean of 63%)than the two heat-sensitive genotypes(average of 16%)under HT(Fig.3C).Although the soluble-sugar concentration was not correlated with water content and water potential,it was significantly correlated with the three heat-tolerance indicator traits under HT(Table S3).Thus,increased sugar content partially explained the heat tolerance of the tested heat-tolerant genotypes(Fig.6).

Compared with SY63 plants,N22 plants exposed to HT showed large increases in soluble-sugar concentration(Fig.3C)and large reductions in water potential(18.2%in N22 vs.10.3%in SY63 under W1,34.0% in N22 vs.11.9% in SY63 under W2)(Fig.2A,B);however,SY63 had higher absolute anther water content(Fig.1D,E),higher panicle water potential(Fig.2D–F),stable and smaller reductions in anther water potential(Fig.2A,B),and smaller increases in soluble sugar concentration(Fig.3C)in response to HT.Similarly,Wu et al.[45]reported that SY63 had a more stable CTK concentration than N22 under HT,and this difference was associated with high heat resistance.These findings suggest that SY63 may maintain relatively stable physiological processes under heat stress,whereas N22 may increase heat resistance by modifying physiological processes,such as enriching sugar content and lowering water potential.Thus,N22 and SY63 may adopt different approaches to alleviate heat damage.

The finding that the fertility and anther dehiscence of N22 plants grown in well-watered conditions was significantly higher under the control and HT treatment is consistent with those in previous reports[17,49,67].Because anther dehiscence begins soon after spikelet opening and is completed soon in N22,developmental escape may be a strategy for heat tolerance[17,68].These inherent advantages contributed to the true heat tolerance of N22.However,N22 also showed significant changes in water potential(Fig.2A,B),spikelet transpiration,and soluble sugar concentration(Fig.3A,C)under HT compared with the control,indicating that N22 had adaptive responses to HT.Li et al.[30]found that tolerant N22 responded with high expression of genes for a sugar transporter(MST8)and a cell wall invertase(INV4)under heat stress,and maintained sucrose supply to avoid carbohydrate starvation.In N22 under HT,heat shock proteins were highly upregulated,which was beneficial for the high spikelet fertility[17,50].N22 also maintained higher relative water content and turgor potential,lower H2O2levels and higher activity of superoxide dismutase and ascorbate peroxidase in panicles under water stress in comparison with the control conditions[69].These responses may contribute to stress tolerance in N22.Thus,the heat tolerance of N22 is due not only to its inherent advantages but also to adaptation.

5.Conclusions

HT during the flowering stage reduced anther dehiscence,pollen fertility,and spikelet fertility,and larger decreases were observed in heat-sensitive genotypes(LYP9 and TXZ25)than in heat-tolerant genotypes(SY63 and N22).HT increased the spikelet transpiration rate,resulting in strong transpirational cooling and lower temperature inside spikelets,especially in heat-tolerant genotypes.HT increased the anther soluble-sugar concentration,and this change was accompanied by reduced panicle and anther water potential,especially in heat-tolerant genotypes,suggesting the involvement of osmotic adjustment in their heat resistance.The expression levels of individual aquaporin genes showed different responses to heat stress;the total expression levels of all 19 aquaporin genes increased in anthers and decreased in glumes under HT.The positive correlations between the total expression levels of glumePIPsandTIPswith the spikelet transpiration rate,as well as between the total expression levels of antherPIPsandTIPswith anther dehiscence,suggest that aquaporins are involved in heat tolerance.The heat-resistant genotypes N22 and SY63 may adopt different approaches to relieve heat damage.

CRediT authorship contribution statement

Wencheng Wang:Conceptualization,Data curation,Formal analysis,Investigation,Methodology,Writing-original draft,Writing-review & editing.Kehui Cui:Conceptualization,Data curation,Formal analysis,Funding acquisition,Methodology,Project administration,Supervision,Validation,Writing-original draft,Writing-review & editing.Qiuqian Hu:Investigation,Methodology.Chao Wu:Methodology,Investigation.Guohui Li:Methodology.Jianliang Huang:Supervision,Writing-review & editing.Shaobing Peng: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

This work was supported by the National Natural Science Foundation of China(31871541).We thanks the anonymous reviewers for their helpful comments and suggestions.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.11.010.