Long-Term Mild Heat Causes Post-Mitotic Pollen Abortion Through a Local Effect on Flowers

Abstract

Crop reproductive success is significantly challenged by heatwaves, which are increasing in frequency and severity globally. Heat-induced male sterility is mainly due to aborted pollen development, but it is not clear whether this is through direct or systemic effects. Here, long-term mild heat (LTMH) treatment, mimicking a heatwave, was applied locally to tomato flowers or whole plants and followed up by cytological, transcriptomic, and biochemical analyses. By analyzing pollen viability, LTMH was shown to act directly on the flowers and not via effects on other plant tissue. The meiosis to early microspore stage of pollen development was the most sensitive to LTMH and 3 days of exposure around this period was sufficient to significantly reduce pollen viability at the flower anthesis stage. Extensive cytological analysis showed that abnormalities in pollen development could first be observed after pollen mitosis I, while no deviations in tapetum development were observed. Transcriptomic and biochemical analyses suggested that pollen development suffered from tapetal ER stress and that there was a limited role for oxidative stress. Our results provide the first evidence that heat acts directly on flowers to induce pollen sterility, and that the molecular-physiological responses of developing anthers to the LTMH are different from those to severe heat shock.

Keywords: long-term mild heat; oxidative stress; pollen development; tapetum development; tomato (Solanum lycopersicum).

Figure 1

The effects of long-term mild heat location, timing, and duration on viability of mature pollen. (A) Local temperature treatments: plants were grown in control (CT) or long-term mild heat (LTMH) conditions, and flower trusses were either exposed to the ambient temperature directly (“ambient”), to the ambient temperature by flushing ambient air through a bag around the truss [“ambient (bag)”], or reduced or increased temperature by flushing air of modified temperature through a bag around the truss [“cooled (bag)” or “heated (bag)”]. Viability of mature pollen (i.e., from freshly opened flowers) was measured 10 to 20 days after the onset of the treatment. Values indicate the mean ± SE (n = 2–3 plants, with pools of data from 1 to 3 days, each with at least three flowers per plant). *, significantly different from both controls, one-way ANOVA with LSD P ≤ 0.05; **P ≤ 0.01. (B) A 4-week stress-and-release experiment, with plants growing in LTMH for 2 weeks, followed by growth in CT for 2 weeks. Viability of mature pollen was determined daily. Values indicate the mean ± SE (n = 4 plants, with pools of 2 to 9 flowers analyzed per plant per sampling point). Different letter indicate significant differences between treatments, one-way ANOVA with Tukey P ≤ 0.05. (C) Effect of LTMH treatments of variable durations and timing on viability of mature pollen. Samples are plotted according to their stage (expressed as days before anthesis; DBA) at the last day of the treatment. The 4-day LTMH treatment that ended at 10 DBA (i.e., flowers treated from 13 until 10 days before they would reach anthesis) was selected for subsequent cytological and gene expression experiments. Values indicate the mean ± SE (n = 5 plants, with pools of 2 to 3 flowers analyzed per plant per sampling point). *, significantly different from control at the same sampling point, one-way ANOVA with Tukey P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

Figure 2

Frequency analysis of pollen development in anthers from control and 4-day long-term mild heat treatment. Long-term mild heat (LTMH) or control temperature (CT) were applied for 4 days, and flowers that were at the 10-DBA stage on the last day of the treatment (i.e., they were treated from 13 DBA to 10 DBA) were labeled and then sampled on subsequent days (at stages of 9, 8, 6, 5, 3, and 1 DBA) for cytological analysis of pollen development (see Supplementary Table S1). For each flower stage per treatment, 100 to 140 locules were analyzed. Developing pollen were assigned to four categories: normal, aberrant shape, plasmolysed, or dead. *, Student's t-test, total percentage of normal, living pollen in LTMH treatment significantly less than in CT treatment on the same day P < 0.05; **P < 0.01; ***P < 0.001; n.s., not significant.

Figure 3

Comparison of pollen cytology after control and 4-day long-term mild heat treatment by light microscopy. (A,B) Early bicellular pollen from control (CT) (A) and 4-day long-term mild heat treatment (LTMH) (B). (C,D) Mature pollen from CT (C) and LTMH (D). (E,F) Anthesis stage pollen from CT (E) and LTMH (F). Bar = 20 μm.

Figure 4

Comparison of pollen ultrastructure between control and 4-day long-term mild heat treatment by electron microscopy. (A,B) Early bicellular from control temperature (CT) (A) and 4-day long-term mild heat treatment (LTMH) (B), with close-ups of starch granules. (C) Second vacuolisation stage, characterized by absence of starch granules, from LTMH. No apparent differences were found relevant to control. (D,E) Mature pollen stage from control (D) and LTMH (E). (F,G) Pollen at anthesis from control (F) and LTMH (G). s, starch granule (large, white spherically shaped); v, vacuole (grayish, more or less spherically shaped, with dark debris visible inside). Bar = 5 μm.

Figure 5

Comparison of tapetum development between control and 4-day long-term mild heat treatment. Long-term mild heat (LTMH) or control temperature (CT) were applied for 4 days, and flowers that were at the 10-DBA stage on the last day of the treatment (i.e., they were treated from 13 DBA to 10 DBA) were labeled and then sampled on subsequent days (at stages of 9, 8, 6, 5, 3, and 1 DBA) for cytological analysis. For each flower stage, average tapetum developmental stage (see Supplementary Table S2) was plotted against pollen developmental stages of the samples (see Supplementary Table S1). For each stage per treatment, 100 to 140 locules were analyzed. pMC, pollen mother cell.

Figure 6

Genes expressed significantly differently (FDR ≤ 0.05 and |fold change| > 1.5) between 4-day long-term mild heat (LTMH) and control temperature (CT) treatments. The first, second, third, and fourth columns are fold change (FC), FDR q value, Solyc gene code, and functional description, respectively. Positive FC value (red label), expressed higher after LTMH than CT; negative FC value (blue label), expressed lower after LTMH than CT.