The taro cultivars of Xiangsha and Longxiang were planted in the experimental field of Yangzhou University from April 2019 to November 2019. We grew seed taro of Xiangsha taro on April 14, 2019 for 10 weeks in normal conditions, and then with 4 weeks of drought stress. The same-size and disease-free mother taro specimens were selected for the experiment.

Cross-sections of the taro corms were placed in iodine-potassium iodide (I${}_2$-KI) solution (1% potassium iodide and 0.3% iodine) in the dark for 10 s at 25 °C, observed, and photographed with the Leica DMLS microscope (Leica Microsystems, Wetzlar, Germany).

Structural observation of taro corm was conducted and starch was isolated, which was further used for morphological observation, granule size distribution, and X-ray diffractometer (XRD) analysis in accordance with the method described by Yu et al. [25]. A small amount of starch was dispersed and spotted on the sample table of a scanning electron microscope. After drying at room temperature, the surface was plated with gold in an etch coater (BAL-TEC SCD 500 Sputter Coater, Leica, Germany). The samples were observed under a field-emission Scanning Electron Microscope (S4800, Hitachi, Tokyo, Japan) and photographed. For the determination of size distribution of starch granules, a small amount of dried starch was transferred to a glass slide, mixed by 50% glycerol, and then covered with a coverslip. Photographs of the freely dispersible sample were captured using an optical microscope (200$\times$). The Image-Pro Plus image analysis software (Image-Pro Plus 7, Media Cybernetics, Rockville, MD, USA) was used to measure starch granule size (maximum axial length of the center of the granule) in the image. Starch granules from each starch sample of the four groups was randomly counted for 1500 times, and the measurement was repeated in triplicate. The starch samples were first absorbed by saturated sodium chloride for 7 days and then compressed with XRD (D8 Advance, Bruker, Mannheim, Germany) to obtain the starch sample XRD pattern.

Both drought stress team and control of Xiangsha taro were used for transcriptomic analysis. Total RNA was extracted from the inner tissues of the middle parts, cut into 1-cm-thick square samples of Xiangsha taro. The samples were separated, frozen immediately in liquid nitrogen, and stored at –80 °C before RNA isolation. Total RNA was isolated using the Ultrapure RNA Kit (CW0581S; CoWin Biosciences [CWBIO], Cambridge, MA, USA). The mRNA was sent to CWBIO for transcriptome analysis according to the method described by Huang et al. [26]. In this study, transcript expression was evaluated, and differentially expressed genes (DEGs) with q 0.05 and fold change $>$2 between the control and treated samples were considered significantly differentially expressed [26]. The functional annotation of DEGs was performed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis [26].

The results are represented as the main average of taro plants cormsin each of the three control vs three drought individual rows. All samples data were analyzed with Student’s t-test using Microsoft Excel 2010 (Microsoft, Redmond, WA, USA) and SPSS software (19.0, IBM Corp., Armonk, NY, USA) to determine any statistically significant differences among the samples. Student’s t-test was used to compare means at a significance level of p 0.05 or p 0.01.

Xiangsha taro and Longxiang taro are two different cultivars of multiseed taros, which were grown in a paddy field 2019 in April in the Jiangsu Province of China. The images of the cultivated taro at 8, 12, 16, and 20 weeks are shown in Fig. 1A,C. Both Xiangsha taro and Longxiang taro were formed at 12 weeks. The height and width of Xiangsha taro were increased to 166.12% and 131.03% at 8 and 12 weeks, respectively (Fig. 1B), whereas the height and width of Longxiang taro were increased to 171.87% and 168%, respectively (Fig. 1D). The width increased faster than height as the seed taro grew at 16 weeks. As growth continued to 20 weeks, there was recovery in the height of the mother taro. The rate of increase in height was faster than that in width. The fastest period of increase in total weight was at 12–16 weeks in Xiangsha taro and 16–20 weeks in Longxiang taro.

Fig. 1.

Agronomic characters of ‘Xiangsha’ taro and ‘Longxiang’ taro in different periods. (A,B) ‘Xiangsha’ taro. (C,D) ‘Longxiang’ taro.

To analyze the difference between Xiangsha taro and Longxiang taro at the starch accumulation level, we used cross-sectioning and I${}_2$ staining. In the vegetative stage of bulb development, the starch bodies in the cells were present as fine granules and distributed on one side of the cells (Fig. 2A,A1). After 4 weeks of development, the starch bodies increased in volume, but with no obvious changes in location, distribution, or quantity (Fig. 2B,B1). However, in the mature stage of the starch development of taro, the number of starch bodies was significantly increased. The distribution changed from one side of the cell, spreading almost to the entire inner portions of the cell, showing a trend of aggregation in morphology (Fig. 2C,C1). This suggests that both Xiangsha taro and Longxiang taro have similar agronomic characteristics in height, width, and weight during starch accumulation. However, the starch bodies of Xiangsha taro tend to combine into larger granule sizes in the later stage of development, whereas Longxiang taro is still dispersed in cells as multiple small sizes of starch.

Fig. 2.

Starch characteristics of ‘Longxiang’ taro and ‘Xiangsha’ taro in different periods. (A–C, A1–C1) Semi thin section staining observation of amyloid; (D–F,D1–F1) starch granules under scanning electron microscope; (G–I,G1–I1) starch particle size distribution map. (A,D,G,A1,D1,G1) Budding period (12 weeks). (B,E,H,B1,E1,H1) Bulb expansion stage (16 weeks). (C,F,I,C1,F1,I1) Maturity (20 weeks). Scale bar: (A–C, A1–C1)—10 µm.

To further prove this difference, we analyzed the starch granule size with a scanning electron microscope. Both taro starch granules were polygonal in shape with sharp angles and edges (Fig. 2D–F,D1–F1), but the size of starch in Xiangsha taro was bigger than that in Longxiang taro (Fig. 2G–I,G1–I1). The granule size distribution of starches in Xiangsha taro peaked at 1–2.5 µm and contained less large-sized starch granules ($>$6 µm) at 12 weeks. The granule size distribution of starches in Xiangsha taro peaked at 2–3.5 µm and 1.5–3 µm at 16 and 20 weeks, respectively, and also contained less large-sized starch granules ($>$6 µm). However, a bigger size ($>$6 µm) of starch granule size at both 16 and 20 weeks was observed. There was no difference in granule size distribution of starches among Longxiang taro at 12, 16, and 20 weeks, and all granule sizes were no more than 4 µm and peaked at 1.25 µm. These data indicated that the increase in starch particle size of Xiangsha taro was greater than that of Longxiang taro during the corm expansion period.

Crystallinity of starch is a parameter that reveals the properties of starch granules, and its numerical changes directly affect the application properties of starch. The starch crystallinity of Longxiang taro basically increased with the growth cycle, whereas the crystallinity of Xiangsha taro was highest in the bulbous expansion period. Even though the crystal structure of taro starch was similar, the relative crystallinity difference between different varieties in different periods was clear. The results of comparisons of different periods were as follows: Xiangsha taro (36.25%) in the bulbous expansion period $>$ Xiangsha taro (35.05%) in the mature period $>$ Xiangsha taro (31.30%) in the growing period; Longxiang taro in the mature stage (33.65%) $>$ Longxiang taro in the bulbous expansion stage (32.08%) $>$ Longxiang taro in the growing stage (31.01%) (Fig. 3). These data suggested that Xiangsha taro in the bulbous expansion period had a significantly higher transition degree than Longxiang taro.

Fig. 3.

X-ray diffractometer (XRD) patterns and relative crystallinity of ‘Xiangsha’ taro and ‘Longxiang’ taro starch grains in different periods. (A–C) XRD patterns of starch grains. (A) Budding stage. (B) Bulb expansion stage. (C) Maturity period. (D) Relative crystallinity of starch particles.

The period of 16 weeks taro is around July to August, which is very hot and short of water. As observed, starch accumulation in Xiangsha taro was more sensitive than that in Longxiang taro. So we chose Xiangsha taro for drought stress analysis. We found that the taro plant growth was higher under normal conditions than under drought stress (Fig. 4A), and there was more seed taro in the control (Fig. 4A). With I${}_2$-stained tests, both taros showed bluish purple color. These data demonstrated that drought stress inhibited taro extension and seed taro number.

Fig. 4.

Transcriptome sequencing analysis of related genes of ‘Xiangsha’ Taro under drought stress. (A) Agronomic characters of ‘Xiangsha’ taro after 4 weeks of drought. (B) Bulb iodine staining. (C,D) Transcriptome sequencing data compare with Wild type (WT) and drought stress. (E) Gene Ontology (GO) annotation analysis. (F) Top 20 of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment.

To investigate drought stress-related genes, we conducted transcriptome analysis using mRNAs prepared from the bulbs of Xiangsha taro in the expansion stage. The RNA sequencing data identified 2864 genes with at least twofold higher transcript levels, and 1605 genes with at least twofold lower levels in drought stress conditions compared with the control (Fig. 4B,D). These genes are involved in multiple signaling pathways, particularly endocytosis, plant hormone signal transduction, phosphatidylinositol signaling system, starch and sucrose metabolism, cutin and wax biosynthesis, nitrogen metabolism biosynthesis of secondary metabolites, among others (Fig. 4C–F). The top 20 enriched pathways are described in Fig. 4D. Drought stress may affect hormone and cell wall biosynthesis signaling. We found 16 genes in ABA, ethylene, jasmonate, cytokinin, and auxin signal transduction (Table 1). The ABA receptor PYR/PYL (comp9009_c0_seq1) and TF MYC2 (comp8652_c0_seq1) are two putative ABA downstream signaling genes, which also respond to drought stress [27]. We also identified four genes in cutin and wax biosynthesis pathways, and four genes involved in second messenger systems (Table 1). Homologue genes of CER1 in Oryza sativa L. and Arabidopsis Thaliana mediate cell wall synthesis, and respond to drought stress. CER1 in taro was obviously increased. Drought stress caused taro extension by starch accumulation. We further selected starch biosynthesis-related genes. We found that six homologue genes were induced more than six times, which encoded trehalose 6-phosphate synthase/phosphatase. The transcript level of 5 Uridine Diphosphate (UDP)-glucose 6-dehydrogenase and 3 UDP-glucuronate 4-epimerase related genes was also increased (Table 2).

As starch accumulates in crops, starch granules display diverse morphological characteristics, such as shape and size, which further affect the physical and chemical properties [28]. In Oryza sativa L., starch granules are large and polygonal with irregular shapes accompanied by some small spherical starch granules [28]. The diameter of rice starch granules is no more than 15 µm. In Triticum aestivum L., starch granules are circular or elliptical in shape, with a few irregular particles [25]. A-type starch has a diameter of 25–35 µm, and B-type starch has a diameter of only 2–8 µm. Both Xiangsha taro and Longxiang taro are multiseed taro, which have similar growth characteristics in height, width, and weight. The taro starch basically presents an irregular polyhedron shape, with a smooth surface and plump particles, and generally no gap. The diameter of taro starch granules is about 1–4 µm [29]. The main difference between the varieties is the size of the starch granules. The average starch granules size of Xiangsha taro is significantly larger than Longxiang taro starch in the same growth stage, and the grain size of Xiangsha taro changes with the growth period, showing a trend of gradual increase. Longxiang taro has no obvious increase trend at different times, and the grain size is always maintained at a constant level.

The development of the taro plant is seriously inhibited under drought stress, especially seed taro formation and extension. Drought induces the root: shoot ratio and reduces total plant biomass loss. Nutrient efficiency, water use efficiency, chlorophyll content index, and nitrogen content could improve the photosynthesis rate. Trehalose plays a crucial role in starch biosynthesis by affecting the main raw materials supply [30]. The application of exogenous sugar has shown that trehalose functions as a key sugar signal during rice grain filling. Trehalose regulates the expression of genes related to sucrose conversion and starch synthesis, thereby promoting the conversion of sucrose to starch [31]. Transcriptome sequencing of drought-stressed taro corms has shown that trehalose 6-phosphate synthase/phosphatase is upregulated, which might inhibit starch accumulation. The ratio of auxin and cytokinin is a key factor for shoot formation in rice and other crops [32]. Here, we also found that the transcript level of auxin and cytokinin-related genes were affected, as well as ABA. ABA is a stress-induced phytohormone, which accumulates in leaves to induce stomatal closure, preventing water loss through inhibition of transpiration [33]. Thus, a putative model is that drought stress inhibits starch accumulation and seed taro formation, as determined by the ratio of auxin and cytokinin. Further studies using transgenic taro are required to fully understand the mechanisms underlying taro growth.