Flower development under drought stress: morphological and transcriptomic analyses reveal acute responses and long-term acclimation in Arabidopsis

Abstract

Drought dramatically affects plant growth and crop yield, but previous studies primarily examined responses to drought during vegetative development. Here, to study responses to drought during reproductive development, we grew Arabidopsis thaliana plants with limited water, under conditions that allowed the plants to initiate and complete reproduction. Drought treatment from just after the onset of flowering to seed maturation caused an early arrest of floral development and sterility. After acclimation, plants showed reduced fertility that persisted throughout reproductive development. Floral defects included abnormal anther development, lower pollen viability, reduced filament elongation, ovule abortion, and failure of flowers to open. Drought also caused differential expression of 4153 genes, including flowering time genes flowering locus t, suppressor of overexpression of CO1, and leafy, genes regulating anther and pistil development, and stress-related transcription factors. Mutant phenotypes of hypersensitivity to drought and fewer differentially expressed genes suggest that dehydration response element B1A may have an important function in drought response in flowers. A more severe filament elongation defect under drought in myb21 plants demonstrated that appropriate stamen development requires MYB domain protein 21 under drought conditions. Our study reveals a regulatory cascade in reproductive responses and acclimation under drought.

Figure 1.

Soil Water Moisture and Plant Growth during the Treatment Period. (A) Soil moisture at d 20 to 40 from when the seeds were planted; the arrow points to control day 0 (C0) immediately before drought treatment. The data points and error bars represent the mean and se (n = 5). (B) to (F) WW plants (left) and DT plants (right) at C0 (B), C3/T3 (C), C5/T5 (D), C7/T7 (E), and C10/T10 (F). (G) RWC over the course of the treatment. The values for DT plants on days 5 to 20 are significantly different from those of the wild type (P < 0.01 by a Student’s t test.). The data points and error bars represent the mean and se (n = 9). Bar = 3 cm. (H) The cumulative numbers of opened flowers on the main stem of WW plants and DT plants during the 20-d drought treatment. The DT plants differ significantly from the WW plants on days 6 to 20 (P < 0.01 by a Student’s t test.). The data points and error bars represent the mean and se (n = 6). (I) The number of floral buds (from stages 8 to 12) on the main inflorescence stem, with significant difference between the DT and WW plants on days 5 to 12 (P < 0.01 by a Student’s t test.). The data points and error bars represent the mean and se (n = 6). [See online article for color version of this figure.]

Figure 2.

Drought Affects Development of Reproductive Organs. (A) An example of WW inflorescences at C10, with one opened flower and many floral buds. (B) At T10, growth of DT inflorescences was retarded and flower buds were much smaller than those of WW inflorescences. (C) The newest opened flower (one petal was removed) from the WW inflorescence of (A). The filaments were sufficiently long for pollen grains to be deposited onto the stigma. (D) The newest opened flower (one petal was removed) from the DT inflorescence of (B). The filaments were apparently too short to deliver the pollen to the stigma. (E) Floral buds from the WW inflorescence, shown from the oldest (left, floral stage 12) to the youngest (right, stage 8). (F) Floral buds from a DT plant at stage 12 to stage 8. (G) A stage 12 floral bud from a WW inflorescence. (H) A stage 12 floral bud from a DT inflorescence. (I) Sepals and petals were removed from the floral bud in (G). (J) Sepals and petals were removed from the floral bud in (H). (K) to (N) Anther cross-sections of a C10 WW inflorescence from the second floral bud (K), fifth bud (L), eighth bud (M), and fourteenth bud (N). (O) to (R) Anther sections of a T10 DT inflorescence from the first bud (O), third bud (P), fifth bud (Q), and eighth bud (R). (S) to (W) Lengths of sepals (S), petals (T), anthers (U), filaments (V), and pistils (W). The data points and error bars represent the mean and se (n = 6). (X) The ratio lengths of stamens to pistils. *Significantly different (P < 0.05). **Highly significant (P < 0.01) by a Student’s t test. The data points and error bars represent the mean and se (n = 6). Bar in (A) to (J) = 1 mm; bar in (K) to (R) = 50 µm.

Figure 3.

Flower Development after 20 d of Drought Treatment. (A) The main inflorescence stem of a DT plant photographed at T20. The fruits and flowers are labeled from the oldest to the youngest (B) to (G), (J), (K), (N), (O), (R), and (S) were from (A). (H), (I), (L), (M), (P), (Q), (T), and (U) were from another inflorescence at the same positions as indicated by the blue lines. (B) The fourth fruit from (A) after dissection. (C) A portion of the fourth fruit with higher magnification. (D) The seventh fruit with fresh ovules and seeds shown. (E) A portion of (D) with higher magnification, showing normal and small ovules. (F) The eleventh bud. (G) The eleventh bud after dissection, showing reduced filament length (arrow). (H) The eleventh bud after maturation, showing the detachment at the base of sepals that were not separated at the top. (I) The dissected bud of (H), showing detached short stamens. (J) The twelfth bud. (K) The dissected twelfth bud with short filaments. (L) The twelfth bud after maturation with sepals detached at the base. (M) The dissected bud from (L). (N) The fourteenth bud. (O) The dissected fourteenth bud. (P) The fourteenth bud after maturation. (Q) The dissected bud from (P). (R) The fifteenth bud. (S) The dissected fifteenth bud. (T) The fifteenth bud after maturation. (U) The bud from (T) with two sepals and a petal removed. (V) The comparison of seed count per silique on the main stem from WW and DT plants. P < 0.01 by a Student’s t test. The data points and error bars represent the mean and se (n = 5). (W) The average seed number and silique length of different crossing plants: WW plants (WPi×WP), WW pistil pollinated with DT pollen (WPi×DP), DT pistil pollinated with WW pollen (DPi×WP), and DT plants (DPi×DP). The data points and error bars represent the mean and se (n = 3). Bar in (A) to (T) = 1 mm.

Figure 4.

Analysis of Transcriptomes from Inflorescences of WW and DT Plants. (A) Samples were collected at C0, C3, C4, C5, and C10 for WW plants and T3, T4, T5, and T10 for DT plants. (B) The numbers of the differentially expressed genes for T3/C3, T4/C4, T5/C5, and T10/C10. (C) Venn diagram analysis of all upregulated genes of each group. (D) Venn diagram analysis of all downregulated genes of each group. (E) K-means clustering of differentially expressed genes. Expression distribution of genes activated or repressed 3 d after drought treatment. The number indicates log2 ratio of the fold change between the DT groups with the control group before start of the drought treatment (C0). Yellow color represents genes that have higher expression levels after drought stress and blue indicates reduced expression. (F) The pink lines represent the mean expression profiles for each cluster. The y-axis is in log2 scale and the x-axis shows log2 (T0/C0), log2 (T3/C0), log2 (T4/C0), log2 (T5/C0), and log2 (T10/C0) in sequence. (G) The GO enrichment analysis of genes in each cluster, with P values.

Figure 5.

The Expression of Floral Genes Was Affected by Drought. (A) Expression heat maps for relative expression of flowering time related genes. (B) Expression heat maps for relative expression of ABC model genes. (C) Expression heat maps for relative expression of anther development related genes. (D) Expression heat maps for relative expression of pistil development related genes. The relative expression levels (log2) above 0 represent upregulation, whereas those below 0 represent downregulation.

Figure 6.

CRE Analysis of Differentially Expressed Genes and Prediction of New CREs. (A) CRE analysis within the 500 bp promoter region of all upregulated genes. (B) CRE analysis within the 500 bp promoter region of all downregulated genes. (C) CRE analysis within the 500 bp promoter region of all upregulated transcription factors. (D) CRE analysis within the 500 bp promoter region of all downregulated genes encoding transcription factors. (E) Sequence logo of CRE prediction for the differentially expressed genes at T4. (F) Sequence logo of CRE prediction for the differentially expressed genes at T5. (G) Sequence logo of CRE prediction for the differentially expressed genes at T10.

Figure 7.

Phenotypes and Gene Expression of the dreb1a Mutant. (A) Leaf RWC of dreb1a compared with the wild type. Asterisks indicate values that are significantly different with P < 0.01 by a Student’s t test. The data points and error bars represent the mean and se (n = 6). WT, Wild type. (B) Number of newly opened flowers each day during the DT from C0 to T20. The data points and error bars represent the mean and se (n = 8). WT, Wild type. (C) Phenotype of the dreb1a inflorescence compared with that of the wild type under drought. WT, Wild type. (D) DREB1A was not expressed in the wild type under WW condition at day 0 but was expressed at a high level in the wild-type ovule under DT at days 3 and 5 (arrows). (E) Numbers of differentially expressed genes at T4/T3 and T5/T3 in dreb1a and the wild type. Warm colors indicate upregulated genes and cool colors indicate downregulated genes. WT, Wild type. (F) A heat map of IXa and IXb subfamilies of ERF genes. WT, Wild type. Bar in (C) = 2 mm; bar in (D) = 200 µm.

Figure 8.

Molecular and Morphological Analyses of myb21 Mutants in Response to Severe Drought Stress. (A) myb21 main stem under the WW condition (left) and under DT (right). (B) Total fruit number and the sterile silique number on the main stem under the WW condition and DT. Asterisks indicate values that are significantly different with P < 0.01 by a Student’s t test. The data points and error bars represent the mean and se (n = 6). (C) Wild type (top left) and myb21 (top right) stage 13 floral buds after 20-d drought treatment and the phenotypes after removing sepals and petals (bottom panel). (D) Quantitative RT-PCR and microarray results of MYB21 in the wild type under drought at days 0, 3, 4, 5, and 10. Mean and sd of two biological replicates are presented in each case. (E) In situ hybridization results showed MYB21 expression was at a very low level in the wild type under WW conditions at day 5. (F) MYB21 expression was at a high level in wild type ovule and filament under drought treatment at day 5 (arrows). Bar in (A) = 1 cm; bar in (C) = 2 mm; bar in (F) = 200 µm.

Figure 9.

A Proposed Regulatory Network of Drought Response in Arabidopsis Flowers. Ovals represent genes with known functions in either stress response or developmental process; boxes show different proposed functions. Interactions are represented by dashed lines with arrows representing positive regulatory relationships based on past studies. Interactions with hammer-ended dashed lines represent inhibition. Red symbols indicate stress related, whereas cyan symbols indicate developmental regulation. [See online article for color version of this figure.]