Adaptive response to long-term high temperatures during the reproductive development in Arabidopsis thaliana

Abstract

Main conclusion

Flowering Arabidopsis plants adapt to long-term high temperature by shortening the flowering period and reducing their fertility. The study also demonstrated that the commonly used hot1-3 mutant is tetraploid.

Abstract

Increasing global temperatures and the rising frequency of heat waves pose a significant threat to plant reproduction. The reproductive phase is particularly sensitive to heat stress, yet the underlying mechanisms regulating thermotolerance during this stage remain insufficiently understood, despite significant advances in its understanding during vegetative growth. Heat stress responses are largely controlled by heat shock factors (HSFs) and their downstream targets, including heat shock proteins (HSPs). Among these, HSP101 is essential for acquired thermotolerance and recovery from stress, while HEAT SHOCK BINDING PROTEIN (HSBP) acts as a negative regulator of HSF activity, modulating the heat shock response. Here, we investigated the impact of elevated temperature regimes on the reproductive development of Arabidopsis thaliana, with a particular focus on pollen development and fertility. Our results show that heat stress negatively affects pollen development in a dose-dependent manner, leading to reduced reproductive success. We confirmed the critical role of HSP101 in reproductive thermotolerance using the hot1-3 mutant, deficient in HSP101. Furthermore, we provide evidence that the hot1-3 mutant is tetraploid. The origin of this event is unknown, but it is tempting to speculate that disruption of heat stress responses and interference with meiotic processes may lead to whole genome duplication. Overall, this study provides new insights into the regulation of plant reproductive development under heat stress and highlights the importance of HSP101 in maintaining fertility. These findings contribute to a better understanding of plant responses to rising temperatures and may inform strategies to enhance crop resilience under climate change.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00425-026-05108-6.

Introduction

In recent decades, average global temperatures have increased due to climate change, negatively affecting ecosystems and causing losses amounting to millions of dollars in decreased food and feed production. Climate change not only results in a steady rise in global temperatures, but also leads to an increase in the frequency of extreme meteorological events, having a dramatic and adverse impact on crop production. The production of staple crops, such as rice, wheat, and soybean, is estimated to decrease significantly in the upcoming decades due to heat-induced sterility (Khan et al. 2023). High temperatures alter cellular homeostasis in plants, triggering various tolerances. These include the expression of molecular chaperones, detoxifying enzymes, such as ascorbate peroxidases, and the production of antioxidants. As sessile organisms, plants cannot escape adverse environmental conditions, leading to complex developmental responses to stresses, including heat. These responses interconnect different hormonal routes, including abscisic acid (Suzuki et al. 2016), cytokinins (Yang et al. 2016; Prerostova et al. 2020), auxin (Ai et al. 2023), ethylene (Jegadeesan et al. 2018), and brassinosteroids (Luo et al. 2022). The hormonal pathways are linked with the expression of molecular chaperones (Shi et al. 1998; Hahn et al. 2011), transcription factors (Nover et al. 2001; Koini et al. 2009; Song et al. 2018; Mizoi et al. 2019), activation of the Unfolded Protein Response (UPR), secondary metabolism and cellular responses (Avin‐Wittenberg 2018). However, in severe cases, these adaptive responses may be insufficient to counteract the damage caused by stress conditions. This can result in damaged organs, reduced fertility (Sage et al. 2015; Zhang et al. 2018), and, occasionally, plant death (Vacca et al. 2004; Wahid et al. 2007). Most studies have focused on the vegetative stage of plant development, but the reproductive phase is considered the most sensitive to heat stress (Young et al. 2004; Barnabás et al. 2008; Zinn et al. 2010; Deng et al. 2016; RECROP 2025). High temperatures can impair the development of female and male gametophytes, leading to heat-induced sterility (Endo et al. 2009; Zhang et al. 2018; Khaitova et al. 2024) and reducing fertilization of ovules (Snider et al. 2011; Sage et al. 2015), causing smaller fruits after fertilization (Chu and Chang 2020).

The HEAT SHOCK FACTORs (HSFs) are the primary regulators of the heat shock response (HSR) in plants (Scharf et al. 2012). The most notable HSF is HSFA2, which mediates the HSR (Charng et al. 2007), thermomemory (Friedrich et al. 2021), or other stress responses (Ogawa et al. 2007; Lin et al. 2018b; Zupanska et al. 2019), acting as a super activator (Chan-Schaminet et al. 2009). During the HSR, various HSFs regulate the expression of chaperones (Jacob et al. 2017), with HEAT SHOCK PROTEIN 101 (HSP101) being one of the most studied due to its role in thermotolerance (Queitsch et al. 2000; McLoughlin et al. 2019) and mediating gene expression during the stress recovery phase (Merret et al. 2017). Notably, HSP101 also mediates different aspects of plant reproduction under non-stress conditions (Fu et al. 2002; Qin et al. 2021). The hot1-3 mutant, a knock-out mutant of the HSP101 gene (Hong and Vierling 2001), is highly sensitive to heat and often used as a positive control in heat stress experiments (Kim et al. 2012; Wu et al. 2013; Lin et al. 2018a; Tak et al. 2022).

During the heat stress recovery phase, the HEAT SHOCK BINDING PROTEIN (HSBP) acts as a master negative regulator of the HSF transcriptional activity in eukaryotes (Satyal et al. 1998). HSBP binds to the oligomerization domain of HSFs, resulting in the dissociation of active HSF trimers to inactive monomers. The HSF monomers are then coupled to various chaperones (Fu et al. 2005; Rana et al. 2012; Huang et al. 2024). This attenuates the HSF transcriptional activity and HSR (Morimoto 1998). Under non-stress conditions, HSBP is mainly expressed in Arabidopsis siliques (Hsu et al. 2010) and regulates seed development in Arabidopsis, rice, maize and Brassica rapa (Fu et al. 2005; Hsu et al. 2010; Rana et al. 2012; Muthusamy et al. 2023). The Arabidopsis hsbp-2 mutant exhibits a reduced fertilization rate accompanied by about 35% seed abortion (Hsu et al. 2010). HSBP represses HSFs by binding to the promoter of the HEI10 E3 ligase, limiting meiotic crossovers in gametes in normal and high temperatures and, thereby, influencing fertility (Kim et al. 2022). High temperatures alter meiotic crossover formation and induce the formation of unreduced gametes by interfering with cytokinesis during meiotic divisions, producing diploid and aneuploid spores (Storme and Geelen 2020).

This study aimed to provide a comprehensive analysis of the flowering phase of Arabidopsis under different high-temperature scenarios. The results offer new insights into the adaptive response to high temperatures during the plant reproductive phase. We show that high temperatures negatively affect pollen development in a dose-dependent manner and that the hot1-3 mutant is tetraploid.

Methods

Seed material

We used seeds from Col-0, the GCVC pollen marker line (N67829) (Borges et al. 2012), hsbp-2 (salk_046465) (Hsu et al. 2010), hot1-3 (N16284) (Hong and Vierling 2001), and hsp101 (N678620/SALK_066374C, N656804/SALK_099583C). All mutant and reporter lines are in the Col-0 accession. Primers used for genotyping hot1-3 and hsp101 mutant lines are provided in Supplementary Table S3.

Growth and stress conditions

Arabidopsis seeds were sterilized using chlorine gas for five hours. Seeds were ventilated and plated on Murashige and Skoog (MS) medium. After two days of cold stratification, plates were moved to a cultivation room (Photon Systems Instruments, Czech Republic): 21/18 ºC (light/dark) with a 16-h light/8-h dark condition (LED illumination, light intensity of 150 µmol.m⁻².s⁻¹) and 50% humidity. One week after germination, plants were transferred to soil in individual pots. Plants in soil are cultivated in closed cultivation banks in phytotrons with controlled light (combination of white/red/far-red LED illumination, light intensity of 150 µmol.m⁻².s⁻¹) and temperature regimes. Arabidopsis plants remained in this condition (long-day, 21/18) until activation of the inflorescence meristem, corresponding to growth stage 5 (Boyes et al. 2001), before the elongation of the flowering stem started. At this point, plants were moved to different stress conditions or kept in this control condition (CC, 21/18) for the experiments. For Stress Condition 1 (S1, 28/18), the temperature rises from the night temperature (18 ºC) to 28 ºC in one hour in the first hour of the light period, remains stable for 14 h, and decreases to 18 ºC in the last hour of the light period. For Stress Condition 2 (S2, 28/24), the night temperature is 24 ºC and rises to 28 ºC in 2 h at the start of the light period, remains stable for 12 h, and decreases to 24 ºC in the last 2 h of the light period. For Stress Condition 3 (S3, 34/18), the temperature rises from the night temperature (18 ºC) to 34 ºC in 5 h, remains stable for 6 h, and decreases to 18 ºC in 5 h. The plants stayed in these growth conditions until they stopped flowering.

Phenotyping of inflorescence

The length of the primary inflorescence of Col-0 plants was measured when the first flower opened in all growth conditions. Measurements were performed in intervals of five days until the end of flowering. Opened and defective flowers were counted daily during the flowering time in each growth condition. Fifteen mature siliques per plant (silique 6 to silique 21, silique 1 being from the first opened flower) from the primary inflorescence were measured to calculate the length of siliques in different conditions. A total of 20 plants were measured for all parameters in three biological replicates.

Pollen viability and measurements

Pollen grains of at least 10 plants from all four conditions were stained with Alexander staining (Alexander 1969). Samples were taken 5 and 12 days after the beginning of flowering (DAF, Day After the start of Flowering). One flower was dipped into 15 μL of Alexander staining and vortexed vigorously. Ten microliters of the supernatant were mounted on a slide. Visualization was immediately performed using a ZEISS Axioscope.A1 microscope equipped with DIC optics and an Axiocam 506 CCD camera. The area of 150 pollen grains per plant, ten plants per treatment, was measured using Image J to determine the size of the pollen grains. To quantify pollen viability, at least 200 pollen grains were evaluated. Whole anthers were stained with Alexander staining and kept at 50 ºC for 24 h prior to visualization. The experiment was performed in at least three biological replicates.

Pollen ploidy evaluation

Ploidy measurements were performed as described (Yang et al. 2021) with the following modifications. To determine the somatic ploidy levels, 1–2 young leaves from a 10 day-old seedling were chopped with a razor blade in 500 μL Otto I solution (0.1 M citric acid, 0.5% Tween 20 v/v). The nuclear suspension was filtered through 50 µm nylon mesh and stained with 1 mL of Otto II solution (0.4 M Na₂HPO₄·12H₂O) containing 2 µg DAPI (4’,6-diamidino-2-phenylindole). The ploidy was analyzed on a Partec PAS I flow cytometer with diploid WT plants used as an external standard.

To determine ploidy levels of male gametes, the samples were prepared as described (Borges et al. 2012). The nuclei-enriched pellet was resuspended in 1 mL of LB01 buffer (15 mM TRIS, 2 mM Na₂EDTA, 0.5 mM spermine, 80 mM KCl, 20 mM NaCl, 15 µM mercaptoethanol, 0.1% Triton X-100, pH 9) containing 2 µg DAPI. The ploidy was analyzed on FACSAria (Becton Dickinson) flow-sorter with diploid WT male gametes used as an external standard.

To visualize segregating chromosomes, mature pollen samples were prepared as described (Teng et al. 2008).

Fertilization rate assessment

Pistils from all growth conditions were emasculated and hand-pollinated the following day with pollen of plants from the same genotype grown in 21/18 ºC conditions to achieve the maximum pollination rate possible. Siliques were dissected at two, three, and four days after pollination, and developing seeds and unfertilized ovules were counted. The experiments were performed in three biological replicates with at least 20 siliques per genotype and condition. The data from one replicate is presented. The number of counted siliques is provided in Tables 2 and 3.

Table 2.

Fertilization rate. developed seeds, aborted seeds, and unfertilized ovules were counted in 25 siliques per growth condition

Condition	C	S1	S2	S3
Silique
n	25	25	25	25
Fertilized ovules
n	1741	1139	627	564
%	97.43	79.04	51.31	44.62
p-values (compared to CC)		0.0365	< 0.0001	< 0.0001
Unfertilized ovules
n	46	302	595	700
%	2.57	20.96	48.69	55.38
p-values (compared to CC)		< 0.0001	< 0.0001	< 0.0001
Aborted seeds
n	0	0	10	23
%	0.00	0.00	0.82	1.82
p-values (compared to CC)		1	5.72E-02	2.30E-06
Total n	1787	1441	1222	1264
n ovules/silique	71.48	57.64	48.88	50.56
p-values (compared to CC)		< 0.0001	< 0.0001	< 0.0001
n dev. seeds/silique	69.64	45.56	24.68	21.64
p-values (compared to CC)		0.0365	< 0.0001	< 0.0001

Statistical analysis was performed by Negative Binomial regression, except for seed abortion (Kruskal–Wallis and Dunn multiple comparison tests with p-values adjusted by the Holm method). Aborted seeds are also counted as fertilized ovules. Developed seeds are fertilized ovules that form seeds and do not abort

CC control conditions, S1 stress conditions 1, S2 stress conditions 2, S3 stress conditions 3

Table 3.

Fertilization rate in Col-0 and hsbp-2. Developed seeds, aborted seeds, and unfertilized ovules were counted in n siliques per growth condition

Genotypes	Col-0			hspb-2
Condition	CC	S1	S3	CC	S1	S3
Silique
n	25	24	25	54	22	50
Ovules
n	1806	1386	1096	2751	1152	1848
Ovules/siliques
n	72.24	57.75	43.84	50.94	52.36	36.96
p-values (compared to CC of the same genotype)		< 0.0001	< 0.0001		1	< 0.0001
p-values (compared to Col-0 from the same conditions)				< 0.0001	0.0805	< 0.0001
Fertilized ovules
n	1804	1386	1063	1670	350	82
%	99.89	100.00	96.99	60.71	30.38	4.44
Unfertilized ovules
n	2	0	33	1081	802	1766
%	0.11	0.00	3.01	39.29	69.62	95.56
p-values (compared to CC of the same genotype)		1	0.0376		< 0.0001	< 0.0001
p-values (compared to Col-0 from the same conditions)				< 0.0001	< 0.0001	< 0.0001
Developed seeds
n	1794	1134	674	1301	277	34
% of fertilized ovules	99.45	81.82	63.41	77.90	79.14	41.46
p-values (compared to CC of the same genotype)	< 0.0001	< 0.0001		< 0.0001	< 0.0001	< 0.0001
p-values (compared to Col-0 from the same conditions)				< 0.0001	< 0.0001	< 0.0001
ABORTED seeds
n	10	252	389	369	73	48
% of fertilized ovules	0.55	18.18	36.59	22.10	20.86	58.54
p-values (compared to CC of the same genotype)	< 0.001	0.0003		< 0.0001	< 0.0001	< 0.001
p-values (compared to Col-0 from the same conditions)				< 0.0001	0.9288	< 0.0001

Statistical analysis was performed on Kruskal–Wallis and Dunn multiple comparison tests with p-values adjusted by the Holm method. Aborted seeds are also counted as fertilized ovules. Developed seeds are fertilized ovules that form seeds that do not abort

CC control conditions, S1 stress conditions 1, S3 stress conditions 3

RNA extraction and cDNA synthesis

Pistils from the primary inflorescence at all growth conditions were emasculated the day before pollination. Forty unpollinated pistils were harvested at midday (at the warmest temperature), frozen on dry ice, and stored at −80 ºC. Samples were collected 5 and 12 days after flowering (DAF) began. Samples were ground in liquid nitrogen using ceramic beads and a Retsch Mixer Mill MM 400 (Qiagen). RNA was extracted following the manufacturer’s instructions using the NucleoSpin RNA Plant and Fungi Kit (Macherey–Nagel). After treatment with an RQ1 RNase-Free DNase (Promega), two micrograms of total RNA were used for cDNA synthesis using M-MLV Reverse transcriptase (Promega) following the manufacturer’s instructions.

RT-qPCR

The FastStart Essential DNA Green Master (Roche) with SYBR Green I as the fluorescent dye was used for qPCR on a LightCycler^®96 (Roche). Each reaction contained 5 µl FastStart Essential DNA Green Master (Roche), 2.5 µl of four times diluted cDNA sample, 500 nM gene-specific primer combinations, and PCR-grade water (Roche) to a final volume of 10 µL. The selection of the candidate reference genes, TMA7 (At1g15270) and PP2AA3 (At1g13320), for normalization was made using qBASE + (BioGazelle) and a combination of the Normfinder algorithm in R and BestKeeper Excel sheet (Vandesompele et al. 2002; Andersen et al. 2004; Taylor et al. 2019). The expression of the reference genes was not affected by the growth conditions. The primers targeting the tested transcripts are listed in Supplementary Table S3. Primers for target genes were validated by testing their efficiency and specificity. Three biological replicates were used for the experiments, each with three technical replicates. The PCR reactions consisted of a preincubation step at 95 ºC for 10 min, followed by 40 cycles of 95 °C for 10 s, 55 ºC for 10 s, and 72 ºC for 14 s. Products were checked on a 3% agarose gel. The calculation of log2-fold changes was made using the 2^DCq method according to the guidelines in Taylor et al. (2019). Statistical analysis was performed using a two-way ANOVA and post-hoc Tukey test in R Studio. A log2-fold change of 0.8 or higher was considered significant.

Statistical analysis

To compare flowering times in the different growth conditions, a negative binomial regression test was performed, followed by ANOVA and Tukey test to make multiple pairwise comparisons. A Kaplan–Meier Survival analysis and Cox proportional hazards regression model were used to study the flowering speed, the end of flowering, and the presence of sterile flowers. One-way or two-way ANOVA and Tukey HSD tests were performed to compare the length of siliques and pollen size. To study the viability of pollen grains, number of flowers, number of ovules, number of developed seeds, non-fertilized ovules, and aborted seeds, Kruskal–Wallis, Mann–Whitney U and Dunn tests were performed, followed by a non-parametric Spearman rank correlation to establish a correlation between the number of ovules and developed seeds.

Accession numbers

Arabidopsis AGI locus identifiers: APX2 (At3g09640), HSBP (At4g15802), HSFA1D (At1g32330), HSFA2 (At2g26150), HSFA4C (At5g45710), HSP70 (At3g12580), HSP101/HOT1 (At1g74310), PP2AA3 (At1g13320), TAA1 (At1g70560), TMA7 (At1g15270),

Results

Impact of long-term high temperatures on flowering rate and flower production

Various phenotypic changes in the flowering plants were notable when we analyze the effects of long-term high temperatures (Supplementary Figure S1). In all tested growth conditions, flowering commenced six to nine days after transferring the plants to the different stress and control conditions. The duration of flowering significantly shortened under high-temperature conditions, decreasing from approximately twenty-one days in the control condition to a range of twelve to fifteen days in all stress conditions (Fig. 1a). This reduction in flowering time affected the growth of the primary inflorescence branch, which we monitored by measuring its length at different time points (Fig. 1b). In the Control Condition (hereafter, referred to as CC), plants uniformly elongated until the end of flowering. Under Stress Condition 1 (28/18 °C, hereafter, referred to as S1) and Stress Condition 2 (28/24 °C, hereafter, referred to as S2), growth accelerated until ten days after the beginning of flowering (10 DAF), then slowed from 10 DAF to the end of flowering. Under Stress Condition 3 (34/18 °C, hereafter, referred to as S3), there was a significant reduction in the inflorescence growth rate compared to the control condition, indicating compromised inflorescence meristem activity.

Fig. 1.

Effect of high temperatures on flower production. a Daily production of flowers in each growth condition on the primary inflorescence branch. Statistical analysis of the flowering rate is presented in Supplementary Table S1. b Length of the primary branch in cm in each growth condition. The 95% CI and mean are shown in (a) and (b). c Total number of flowers produced in the primary branch. Significant difference with p-value of p < 0.001 is indicated by ***. d, e Representative pictures of normal flowers (d) and immature flowers (e). The scale bar represents 1 cm. f Percentage of immature flowers produced per plant (n = 20). Statistical analysis using a Kaplan–Meier survival estimator and Cos proportional hazard regression is presented in Supplementary Fig. S2. CC control condition (21/18 °C), S1 stress condition 1 (28/18 °C), S2 stress condition 2 (28/24 °C), S3 stress condition 3 (34/18 °C)

Reduced flowering duration led to the production of fewer flowers (Fig. 1c). Under CC, plants produced two to four flowers daily on the primary inflorescence branch. In contrast, under stress conditions, plants produced up to eight flowers daily, with a high variability day to day (Fig. 1a). High temperatures also induced the production of flower buds that stopped their development in a dose-dependent manner, up to 14% at Stress Conditions S2 and S3 (Fig. 1d-f). This indicates that high temperatures not only reduce the primary inflorescence branch growth rate, but also reduce flower production, with a significant proportion of flowers that will not reach maturation under strong stress conditions. Inflorescence development and flower production were most affected in S2 and S3. Our statistical analysis, utilizing an exponential growth model (Supplementary Figure S2), revealed a strong correlation between inflorescence growth and flowering duration across different conditions, suggesting that high temperatures directly impact the flower meristem activity.

Pollen grains are damaged by high-temperature stress

To better understand how high temperatures influence seed production, we investigated the effects on pollen grains in plants grown under various high-temperature conditions. Pollen is highly sensitive to heat (Rieu et al. 2017; Raja et al. 2019). Using Alexander staining on stamens, we observed that, in CC, approximately 0.10% (n = 2012) of pollen grains were aborted, while in stress conditions, the percentage of aborted pollen grains in 5 DAF flowers varies from 2.95% (n = 2308) in S1, 4.65% (n = 2364) in S2, to 7.27% (n = 2021) in S3. Interestingly, in 12 DAF flowers, pollen grains appeared to adapt to the stress conditions with reduced abortion rates of 1.23% (n = 2270, S1), 3.57% (n = 2131, S2), and 3.85% (n = 2283, S3) (Table 1, Fig. 2a). This slight decrease in pollen viability at high temperatures compared to control temperature is significant at these time points.

Table 1.

Pollen viability assessed by Alexander’s staining (n = 10 plants)

DAF	Condition	Aborted pollen (n)	% of aborted pollen	Statistical groups
5	CC	2 (2012)	0.10	a, b
12	CC	3 (2174)	0.14	a, a
5	S1	68 (2308)	2.95	a, a
12	S1	28 (2270)	1.23	b, a
5	S2	110 (2364)	4.65	a, a
12	S2	76 (2131)	3.57	a, b
5	S3	147 (2021)	7.27	a, a
12	S3	88 (2283)	3.85	b, c

Statistical analysis was performed by Kruskal–Wallis and Dunn multiple comparison tests with p-values adjusted by the Holm method. Non-significantly different samples were grouped by letter. The first letter for comparing the effect of time in each growth condition; and the second letter for comparing the effect of growth temperature at each time point

DAF days after flowering, CC control conditions, S1 stress conditions 1, S2 stress conditions 2, S3 stress conditions 3

Fig. 2.

Effect of long-term high temperatures on pollen development. a Pollen viability assessed by counting the number of viable pollen grains from 10 plants grown under the four growth conditions (CC control, S1 Stress 1, S2 Stress 2, S3 Stress 3) visualized by Alexander's stain at 5 and 12 Days After Flowering (DAF) (n = 1874—2254). Statistical analysis was performed by Kruskal–Wallis, Mann–Whitney U test and Dunn tests. Data are presented as boxplot with median value, interquartile range as box and minimum and maximum values as whiskers. Groups with the same letter are not significant. b Pollen size in µm² from 150 grains from 10 plants grown under the four growth conditions. Statistical analysis was performed by two-way ANOVA followed by Tukey HSD tests. c Pollen grains from a SCVC marker plant grown at CC. The sperm cell nuclei are marked in green and the vegetative cell nucleus is marked in magenta. The scale bar is 50 µm. d Pollen grains from a SCVC marker plant grown at S3. The sperm cell nuclei are marked green and the vegetative cell nucleus is marked in magenta. The scale bar is 50 µm. e Silique length in mm from self-pollinated flowers (10 siliques) from 15 plants grown under the four growth conditions. Statistical analysis was performed by one-way ANOVA followed by Tukey HSD tests, showing significant differences between groups with p < 0.0001. Groups with the same letter are not significant. CC control condition (21/18 °C), S1 stress condition 1 (28/18 °C), S2 stress condition 2 (28/24 °C), S3 stress condition 3 (34/18 °C)

Remarkably, the size of pollen grains also varied between stress conditions. In CC, S1 and S2, pollen grain size was similar, averaging around 400 µm². Notably, a higher number of small pollen grains (around 200 µm²) was only observed in stress conditions. Under S3, pollen grains in 5 DAF flowers were significantly larger, sometimes twice the size of those from the CC (Fig. 2b). These large pollen grains were not observed in 12 DAF flowers. Using the fluorescent GCVC reporter line marking distinctly the sperm cell nuclei and vegetative cell nucleus, we found no apparent changes in the number of nuclei in larger pollen grains despite their size (Fig. 2c, d).

The fertilization rate is reduced at high temperatures

Self-pollination revealed a dose-dependent reduction in fertilization rates. The fertilization rates significantly dropped from 97.43% in CC to 79.04%, 51.31%, and 44.62% under S1, S2, and S3, respectively (Table 2). Additionally, seed abortion rates increased to 0.82% and 1.82% at S2 and S3, respectively. The correlation between the number of ovules and developed seeds (number of ovules versus number of developed seeds) weakens as temperature rises (Spearman rank correlation rho 0.95 for CC, 0.70 for S1, −0.04 for S2, −0.07 for S3), with S2 and S3 showing the least significant effects, suggesting a negative effect of the long-term high temperature on fertilization rate.

The reduced fertilization rate and increased seed abortion led to significantly fewer developed seeds per silique (Table 2). Under CC, plants averaged 69.64 developed seeds per silique, whereas this amount was reduced to 45.56, 24.68, and 21.64 under S1, S2, and S3, respectively (Table 2). This reduction in seed number also significantly affected the length of the siliques measured at flower stage 17, from 16 mm in the CC to 9.5, 7.9, and 7.4 mm in S1, S2, and S3, respectively (Fig. 2e).

The HSR is active in temperature-stressed pistils

Heat shock and oxidative responses are crucial in managing temperature stress. To understand the role of the HSR components in response to high temperatures during fertilization, we collected unpollinated pistils at 5 and 12 DAF and analyzed the expression of the various heat-responsive genes across the different growth conditions. We focused on the expression of HSF genes known to be expressed in reproductive tissues (HSFA1d, HSFA2, HSFA4c), genes related to the HSR (HSP70, HSP101, APX2, HSBP), and auxin biosynthesis (TAA1) (Fig. 3). The duration of the growth at stress conditions had no impact on the expression of the tested genes, except for HSBP (Fig. 3d) for which a prolonged warmness led to a significantly more pronounced upregulation. The TAA1 expression was significantly downregulated in pistils from plants grown under all stress conditions compared to plants grown in CC (Fig. 3h). HSFA2, HSP70, HSP101, and APX2 were the most significantly overexpressed genes in pistils from plants grown at S3 compared to plants grown in CC and S1 and S2 (at 5 DAF, up to 6.3-, 7.0-, 5.1-, and 8.8-log₂-fold change, respectively, Fig. 3b, e–g). To a lesser extent, HSFA1d and HSFA4c genes were also significantly upregulated in pistils from plants grown under S2 and S3 compared to plants grown in CC (at 5 DAF, 1- and 1.1- log₂-fold change, respectively, Fig. 3a and c). HSFA1d, HSFA4c and HSBP were only mildly affected by the growth conditions. It is important to note that HSF transcripts can undergo post-transcriptional regulation, affecting their activity and stability in the cell (Cohen-Peer et al. 2010; Evrard et al. 2013; Carranco et al. 2017; Andrási et al. 2019). Therefore, even small upregulations in these transcription factors can significantly impact cellular responses, as they may already be present in an inactivated state. In addition, this gene expression analysis indicated that S3 appears to be the most stressful and that gene expression patterns under S1 and S2 were rather similar. Therefore, we decided to continue our analysis only with plants grown at S1 and S3.

Fig. 3.

Effect of long-term high temperatures on expression of HSR-related genes in Col-0 pistils. Normalized fold expression assessed by RT-qPCR of HSFA1d (a), HSFA2 (b), HSFA4c (c), HSBP (d), HSP70 (e), HSP101 (f), APX2 (g), and TAA1 (h) in pistils collected at 5 (blue) and 12 (green) DAF from plant grown under the four growth conditions (CC, S1, S2, S3). Expression levels are normalized to the expression in the 5 DAF pistils at CC. Statistical analysis was performed by two-way ANOVA followed by Tukey HSD tests, showing significant differences between groups with at least p < 0.05 to p < 0.0001. Groups with the same letter are not significant. CC control condition (21/18 °C), S1 stress condition 1 (28/18 °C), S2 stress condition 2 (28/24 °C), S3 stress condition 3 (34/18 °C)

The hot1-3 and hsbp-2 mutant pollen is defective

The expression analysis indicates that S3 most severely affected the development of reproductive organs. The hot1-3 and hspb-2 mutants, which are impaired in HSP101 and HSBP, respectively, are known for their heat-sensitivity (Hong and Vierling 2001; Hsu et al. 2010) and reproductive issues in non-stress conditions (Fu et al. 2002; Qin et al. 2021). We phenotyped the mature pollen of both mutants under CC, mild stress (S1), and severe stress (S3). Pollen development was compromised in both mutant lines. The pollen defects and viability frequencies were similar in the hsbp-2 mutant and the wild type in CC. Still, S3 caused complete sterility in the hsbp-2 pollen at 5 DAF (Fig. 4c). Partial recovery was observed at 12 DAF with 70% viable hsbp-2 pollen (Fig. 4d). On the other hand, 10% of hot1-3 pollen grains were aborted in CC, reaching 32% at 5 DAF and 57% at 12 DAF under S3, indicating an absence of acclimation (Fig. 4c, d).

Fig. 4.

Effect of long-term high temperatures on pollen development in hot1-3 and hsbp-2 mutant. a, b Pollen size in µm² from 500 grains from 10 plants (Col-0, hot1-3 and hsbp-2) grown under the three growth conditions (CC control (red), S1 Stress 1 (green), S3 Stress 3 (purple)), assessed at 5 and 12 Days After Flowering (DAF) (n = 1874–2254). Statistical analysis was performed by two-way ANOVA followed by Tukey HSD tests. c, d Pollen viability assessed by counting the number of viable pollen grains from 10 plants grown under the three growth conditions (CC control (red), S1 Stress 1 (green), S3 Stress 3 (purple)) visualized by Alexander’s stain at 5 and 12 Days After Flowering (DAF). Statistical analysis was performed by Kruskal–Wallis, Mann–Whitney U test and Dunn tests. Data are presented as boxplot with median value, interquartile range as box and minimum and maximum values as whiskers. Groups with the same letter are not significant. e–h Pollen grains of Col-0 (e, f) and hot1-3 (g, h) grown under CC (e, g) and S3 (f, h) growth conditions. Pollen nuclei were stained by DAPI. Scale bars are 20 μm. h Note that pollen grains of two sizes are shown (the arrowhead points to the large one). CC control condition (21/18 °C), S1 stress condition 1 (28/18 °C), S2 stress condition 2 (28/24 °C), S3 stress condition 3 (34/18 °C)

At 5 DAF, both Col-0 and hsbp-2 pollen grains reduced their size, with an increased number of outliers for larger and smaller grains in all stress conditions (Fig. 4a). Notably, at 12 DAF, the number of outliers decreased for Col-0, but not in the hsbp-2 mutant, which produced larger pollen grains at S3 compared to CC (Fig. 4b). The hot1-3 mutant produced larger pollen grains, twice the size of the Col-0 ones under CC. The overall size of the hot1-3 pollen grains was reduced at 5 DAF under both stress conditions when compared to CC (Fig. 4a). Under S3 at 12 DAF, while the average size of the hot1-3 pollen grains was comparable to their size under CC, outliers were twice or three times the standard size (Fig. 4b).

The defects in pollen grains in hsbp-2 translated into a decreased fertilization rate and seed production (Table 3). In the CC, 39.29% of ovules were unfertilized (n = 2751), and 22.10% of the developing seeds were aborted in hsbp-2 mutant, reproducing published data (Hsu et al. 2010). Under stress conditions 1 and 3, 69.62% and 95.56% of hsbp-2 ovules were unfertilized, producing only 30.38% and 4.44% seeds, respectively, among which 20.86% and 58.54% aborted. The hsbp-2 mutants exhibit severe reproductive defects under high-temperature stress. These findings underscore the importance of HSBP and the HSR pathway in maintaining reproductive development under stress conditions.

hot1-3 is tetraploid and can develop aneuploidy

Given the size of hot1-3 pollen grains, we investigated whether the mutant had issues with mitotic divisions during pollen development. None of the evaluated pollen grains, stained by DAPI, had an abnormal number of nuclei, regardless of the growth condition or size of the pollen grains (Fig. 4e-h). Next, we assessed the ploidy of Col-0 and hot1-3 sperm nuclei from CC and S3 using flow cytometry. Col-0 consistently showed a haploid DNA content cluster of the diploid parent at 20 DAPI signal units (Fig. 5a, b), regardless of the temperature regime. However, the hot1-3 sperm nuclei displayed two clusters corresponding to haploid DNA content (40–50 DAPI signal units) and likely diploid DNA content (100 DAPI signal units) under S3 of tetraploid parents. This result indicated two unexpected findings: (i) hot1-3 plants were tetraploid and (ii) produced unreduced microspores (Fig. 5c, d). We confirmed the tetraploidy of hot1-3 plants in leaf samples (Fig. 5e-h, Supplementary Fig. S3). Furthermore, staining anaphase chromosomes with DAPI identified ten segregating chromosomes in Col-0 sepals, but twenty chromosomes in hot1-3 sepals (Fig. 5k-m). Overlaying the chromatographs from leaves revealed a double peak at 200 DAPI signal units and two peaks at 350 and 400 DAPI signal units, identifying that some hot1-3 plants are also aneuploid (Fig. 5h). Also, more dispersed appearance of the flow cytometry clusters for hot1-3 sperm nuclei compared to wild-type ones indicates varying DNA contents within one ploidy (Fig. 5a-c), indirectly implying presence of extensive aneuploidy.

Fig. 5.

hot1-3 is tetraploid and aneuploid. a-d Bivariate scatter plot of sperm nuclei from Col-0 (a, b) and hot1-3 mutant (c, d) plants grown under the CC (a, c) and S3 (b, d) growth conditions. The x-axis shows relative DAPI intensity (DNA content) and the y-axis side scatter pulse height (SSC-h), indicating the running time of the nuclei through the sensor. e–h Flow cytometry histograms showing representative profiles of peaks from diploid Col-0 (e) and hot1-3 mutant f–h plants grown under the CC growth condition. The first peak corresponds to 2 C, the second to 4 C and the third to 8 C nuclei (8C peak not shown in hot1-3). The histograms show two types of ploidy in hot1-3 mutant: tetraploid (f) and aneuploid (g). The histogram in (h) displays the overlap when the material of the two plants is mixed prior analysis. i-m Representative pictures (i, k, l) and projection of 5–8 images (j, m) from sepals of Col-0 (i, j) and hot1-3 (k-m) flowers. The red circles points to cells where chromosomes are visible (5 is j and k compared to 10 in k-m). Nuclei are staining by DAPI. Scale bar is 5 μm

To confirm whether the ploidy change was due to the inactivation of HSP101, we assessed the ploidy from the original hot1-3 seeds obtained from NASC (N16284) and two other hsp101 mutant alleles (SALK_066374C and SALK_099583C) (Supplementary Fig. S4). The original hot1-3 mutant plants were tetraploid, but the two other hsp101 mutant lines were diploid (Supplementary Fig. S3), suggesting that the hot1-3 became tetraploid at some point during the selection and characterisation of the mutant allele (Mathur et al. 1998; Hong and Vierling 2001).

Discussion

In recent years, the global climate crisis has significantly reduced crop production (Sidhu 2023), highlighting the need to understand plant adaptive strategies in response to high temperatures. To address this, we studied different high-temperature scenarios using Arabidopsis thaliana (this work) and Brassica napus (Mácová et al. 2022) as model plants. Previous research primarily focused on plant reactions to short heat stress, which is less representative of natural conditions. In our study, we designed different nature-like scenarios that allow plants to grow and adapt to different heat stress conditions, providing comprehensive insights into plant thermo-adaptation to high-temperature environments.

Our study found that flowering in Arabidopsis at 34 ºC results in a reduced number of fertile flowers and decreased fertility. Previous studies have linked this reduction in crop production to heat-induced pollen sterility (Endo et al. 2009; Begcy et al. 2019). While high temperatures did affect pollen development in our growth setup, it adapted to long-term heat stress, likely due to cooler night temperatures. Unlike our approach, most Arabidopsis studies use intense heat shocks (Lei et al. 2020; Rutley et al. 2021) or constant high temperatures over long periods (Yoshida et al. 2008; Hedhly et al. 2020), conditions that are uncommon in nature and cause severe plant damages. High temperatures also affect the female gametophyte in other plant species (Wang et al. 2021; Shi et al. 2022; Jedličková et al. 2023), but this process is not understood in Arabidopsis thaliana. Consequently, damage to both gametophytes may result in reduced pollination and fertilization rates, significantly decreasing seed production (Jiang et al. 2019).

Expression profiling showed that 34 ºC severely affects plant metabolism. Stress-related genes, such as HSP101, HSP70, and APX2, were extremely upregulated in pistils of plants grown under stress conditions, while transcription factors (HSFA2, HSFA1d, or HSFA4c) regulating HSR were upregulated to a lesser extent. Chaperones, such as HSP101, are key to thermo-adaptation (Gurley 2000; Queitsch et al. 2000; Hong and Vierling 2001; Nieto-Sotelo et al. 2002; Wu et al. 2013). Adverse environments induce Reactive Oxygen Species (ROS) production, such as H₂O₂, which must be detoxified, notably by peroxidases, for example during abiotic stresses (Chang et al. 2009), as well as during developmental processes (Ye and Tran 2000; Jacobowitz et al. 2019).

The Arabidopsis hot1-3 mutant, characterized by a T-DNA insertion in the second exon of the HSP101 gene (Hong and Vierling 2001), displays severe defects during male gametophyte development when grown under high temperature conditions. This underscores the critical role of HSP101 in protecting gametes during thermal stress. Optimal temperatures are essential for meiotic crossovers (Lloyd et al. 2018) and chromosome segregation (Lei et al. 2020; Khaitova et al. 2024). Our study revealed that the hot1-3 mutant is tetraploid. This line was isolated from a T-DNA insertion population generated from Agrobacterium-mediated transformation of Arabidopsis cell suspensions (Mathur et al. 1998). Selected calli were regenerated using classical tissue culture techniques. The resulting plants may exhibit alteration in ploidy and chromosome number or structure due to genome instability inherent to the tissue culture environment (Fras and Maluszynska 2003; Bednarek and Orłowska 2020). Although it is unclear when the polyploidization occurred, it may likely be during its early generation as it is already present in the material donated to the stock centre by Suk-Whan Hong (Hong and Vierling 2001). Interestingly, our data showed that hot1-3 developed aneuploidy under high temperatures. However, because hot1-3 was the only line evaluated for this phenotype, it remains unclear whether this susceptibility is directly caused by the loss of HSP101 function or if it stems from the specific genomic background of this line. Heat stress is known to induce unreduced microspores (Storme and Geelen 2020), as observed in Arabidopsis Col-0 plants grown at 35/22 ºC (Nguyen et al. 2019). However, this effect was not seen under our Stress Condition 3, which could be attributed to the different temperature fluctuations between day and night (13 °C in Nguyen et al. (2019) versus 18 °C in our study).

HSFs coordinate both the HSR and plant development. HSBP, a negative regulator of HSF transcriptional activity, attenuates the HSR when temperatures decrease (Hsu et al. 2010). Using the hsbp-2 mutant, which cannot efficiently attenuate the HSR, we examined how Arabidopsis seeds develop under our stress conditions. HSBP regulates seed development (Hsu et al. 2010) and meiotic crossover during ovule development (Kim et al. 2022). Our findings show that hsbp-2 pollen development is susceptible to high temperatures. The analysis of a hypersensitive mutant to temperature stress (hot1-3) and a hyperreactive mutant (hsbp-2) provided insights into the reproductive phase under temperature stress.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.F.S.L., A.P., and H.S.R. designed the research; J.F.S.L., M.Š, O.G. and F.Y. performed the analysis and analyzed the data; J.F.S.L. and H.S.R. wrote the article. All authors read and approved the manuscript.

Funding

Open access publishing supported by the institutions participating in the CzechELib Transformative Agreement. Grantová Agentura České Republiky, GA22-29717S, Helene S Robert, GA24-10021S, Ales Pecinka, Ministerstvo Školství, Mládeže a Tělovýchovy, CZ.02.01.01/00/22_008/0004581, Helene S Robert

Data availability

Source data are provided in the Zenodo repository (10.5281/zenodo.19607439).

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.