Physiological factors Influencing Climate-Smart Agriculture: Daylength-Mediated Interaction Between Tillering and Flowering in Rice

1. Introduction

In the realm of agriculture, particularly in livestock and crop production, methane is a greenhouse gas that contributes significantly to global warming [1,2]. Methane production is primarily favored by a reducing state of the soil, as is the case during rice cultivation under waterlogged conditions [3]. Therefore, alternative cropping schemes, including alternate wetting and drying systems of rice intensification, and other such water management practices were tested for their potential to reduce soil methane emissions [4,5].

Soil methane emission into the atmosphere occurs primarily through the aerenchyma tissue of rice tillers [6] and can be reduced by shortening the growth period [3]. Therefore, controlling the extent of rice tillering and the tiller-growing period is imperative for reducing methane emissions and promoting climate-smart rice production.

Rice tillering exhibits considerable variation, attributed to both environmental and genetic factors [7,8]. Furthermore, the number of panicles associated with tillering significantly impacts annual yield because of the large variation among yield components [7,9]. Factors that affect rice tillering include crop management practices such as planting density, fertilization rates, and irrigation depth. In addition, environmental factors such as light intensity and quality, temperature, and daylength also exert marked effects of rice tillering [10-13]. Particularly, daylength is a critical environmental factor influencing the development of tillers and plays a major role in regulating growth duration through phenological responses [14,15]. Specifically, the optimum daylength for the promotion of tillering contrasts with that which accelerates flowering and shortens the vegetative growth period [16,17,18,19,20].

Floral induction is essential for flowering [21,22], and as it initiates a shift toward reproductive growth, it serves as a key regulator of the vegetative growth period [23,24,25]. Further, because tillering mainly occurs during the vegetative growth period, it is strongly influenced by developmental processes such as floral induction [13]. Specifically, FLOWERING LOCUS C (FLC) and FRIGIDA (FRI), two floral repressors in the vernalization pathway, reportedly regulate tillering in Arabidopsis thaliana [13,26], and similar related studies were reported in other species [27,28]; thus, in wheat, for example, tillering and spikelet number are influenced by the overexpression of five FLOWERING LOCUS T (FT)-like genes [29].

Since the initial discovery of the MOC1 gene as a crucial regulator of rice tillering, several studies were conducted at both physiological and molecular levels [30,31]. Thus, for example, the correlation between vegetative growth (e.g., tiller and leaf development depending on nitrogen supply) and flowering time was investigated [8,30]. Additionally, the involvement of plant hormones in shoot branching induction, specifically cytokinin and auxin, was reviewed [32]. Further, studies have highlighted the significance of strigolactone in suppressing the outgrowth of tiller buds [32,33,34].

At the molecular level, various studies confirmed the interaction between tillering and flowering responses using gene-edited mutants related to these processes [34,35,36,37,38,39]. Similar to the increase in tillering under conditions of delayed flowering [15,19,40], research has shown that tillering increases concomitantly with the overexpression or knockdown of genes, e.g., OsRFL, OsSOC1, and OsLUX [34,41,42]. Conversely, other studies have tillering to decrease along with delayed flowering [35,38]. However, to date, few studies have addressed these conflicting results.

We hypothesized that this lack of consistency among research reports is due to the differences in plant development (e.g., tillering, panicle formation, and flowering) that result from the specific daylength conditions used in different studies. Although the florigen activation complex (FAC) was found to affect growth phenomena such as tillering, in addition to flowering and floral induction [37], the interaction between tillering and flowering under specific daylength conditions has not been elucidated. Therefore, in the present study, we aimed to analyze the interaction between tillering and flowering responses in two rice cultivars by artificially restricting tillering and growing plants under controlled daylength.

2. Materials and Methods

2.1. Ethics Statement

This study was performed in accordance with Institute-approved guidelines and regulations. The test varieties were provided by SeoYeong Yang of the Rice Production and Physiology Division of the National Institute of Crop science (NICS). We obtained permission from the NICS to use these varieties (https://www.nics.go.kr/apo/breed.do?m=100000128&homepageSecod=nics).

2.2. Experimental Materials and Design

A pot experiment was conducted using a completely randomized design incorporating two factors of variation (daylength and tiller removal) at two levels of variation each (short and long day, and with and without tillers removed). Two experiments were conducted in a controlled environment facility (ENT Inc., Boocheon, Korea) at the National Institute of Crop Science in Jeonju, South Korea (35°49ʹ19ʺ N, 127°8ʹ56ʺ E), where light intensity, temperature, and humidity can be artificially controlled (Figure. S1). The lighting chamgers utilized in this facility are capable of controlling temperature, humidity, light length, and light intensity (Figure S1). Each chambers has a floor area of 3.8m2(2.4 in width and 1.6m in length) with a height of 2m. Lights were turned on at 7:30 h regardless of daylength treatment, such that daylength was adjusted by the lights-off time. Two rice cultivars representing ecotypes with different maturation times were used, namely, early maturing ‘Odae’ (Oryza sativa ssp. japonica, IT218242) and mid-late maturing ‘Saenuri’ (Oryza sativa ssp. japonica, IT235281).

Fifteen-day-old seedlings of both varieties were transplanted into 1/5000 a Wagner pots at a density of three plants per pot. A composite slow-release fertilizer was applied with 9, 4.5, and 5.7 kg nitrogen, phosphate, and potassium per 1000 m², respectively, at a rate based on the area used by three plants (i.e., 0.042 m²; planting distance: 30 × 14 cm) instead of the entire pot area. Tillers were removed every 2–3 d starting 7 d after transplanting, and water was continuously applied at a depth of 2–3 cm or more. A plant with tillers removed is shown in Figure S2. These procedures were performed as previously described [15,25].

2.3. Experiment 1: Preliminary test of Growth And Heading Responses To Tiller Removal Under Short-Day Conditions

Temperature was set to 22 °C (maximum 28 °C/minimum 18 °C) and daylength was fixed to relatively short conditions (12 h 30 min light/11 h 30 min dark) from sowing to the heading stage [22,43]. The light intensity was set at 700 µmol m^-2 s^-1 photosynthetically active radiation (PAR) and the relative humidity was set at 65%.

2.4. Experiment 2: Analysis of Growth And Heading Upon Tiller Removal Under Short- And Long-Day Conditions

Before tiller removal, the temperature was set to 25 °C (maximum 30 °C/minimum 20 °C) for 22 d (15 d after sowing and 7 d after transplanting), and to minimize the induction of photosensitivity, daylength was set to 15 h [22,43], which is slightly longer than that generally used as a long-day condition (14 h 30 m light/9 h 30 m dark). The long-day condition was set before tiller removal to differentiate the photosensitive response to daylength conditions after tiller removal. Daylength conditions after tiller removal were matched between short (12 h light/12 h dark) and long (14 h 30 m light/9 h 30 m dark) days [22,43], and the temperature was set to 28 °C (maximum 33 °C/minimum 23 °C).

After the heading stage date, conditions were adjusted to 25 °C (maximum 30 °C/minimum 20 °C) and long day (14 h light/10 h dark) during the ripening stage. Temperature and daylength were changed again after the ripening stage , as 28 °C is deemed excessively high for the ripening stage, could present challenges in accurately evaluating grain weight. In addition, different daylength conditions can affect grain weight; therefore, we set the same daylength condition again for the ripening stage. These conditions were set to observe the effect of temperature and daylength up to the heading stage. The light intensity was set at 700 µmol m^-2 s^-1 PAR and the relative humidity at 65%.

2.5. Growth and Development Measurements

Plant height was measured from the ground to the top of the apical leaf tip. Stem length was measured from the ground to the uppermost internode. Leaf age was expressed as leaf number on the main stem. In both experiments, the heading date was calculated as the number of days (growth period) after transplanting, before panicle emergence from the leaf sheath. Panicle emergence was examined daily from 1300 to 1400 h. Fifteen plants were used to analyze heading date and growth. These procedures were performed as previously described [15].

2.6. RNA Extraction and Gene Expression

Two plant organs were sampled, namely the 2^nd and 3^rd leaves of the main stem and tiller, immediately frozen under liquid nitrogen, and stored at −80 °C until RNA extraction. Nine rice plants were analyzed per treatment, three of which were sampled for one repetition and used for analysis in three repetitions for real-time polymerase chain reaction (RT-PCR). Plant material for RNA extraction was sampled at 1000 h (2.5 h after lights were turned on) because Hd3a, RFT1, Ehd1, Ghd7, and MFT1 reportedly maintain a high expression level for 0–4 h after plant exposure to light [44]. Total RNA was extracted according to the protocol of Chang et al. [45]. cDNA synthesis was performed using a Primescript RT reagent kit with gDNA eraser (TaKaRa Bio, Inc., Kusatsu, Japan). For RT-PCR, SYBR Green (SYBR Realtime PCR Master Mix, Toyobo, Japan) was used as a fluorescent dye. The analysis was conducted using a Roter-Gene 6000 (Corbett Research, Australia). Primer sequences are listed in Table S1. The procedures were performed as previously described [15].

2.7. Statistical Analysis

Data were analyzed using R software (v.2.2), in which analysis of variance (ANOVA) was conducted, followed by Duncan’s multiple range test to evaluate significant differences at P < 0.05. The following equation (1) [46] was used to model the development of plant height, leaf age, and tiller number based on the number of growing days from transplanting:

$$F={\displaystyle \frac{H_{max} or L_{max} or T_{max}}{1+e^{-\left(t-tm\right)\mathrm{*}rF}}}$$

(1)

where H_max is the final plant height; L_max is the final leaf age; T_max is the final number of tillers; _rF is the rate of development up to final plant height, leaf age, or tiller number; t is the number of days after transplanting; and tm is the timepoint at which half of the final plant height or leaf age or tiller number was achieved, each referring to the timepoint when plant height, leaves, or tillers reached the maximum rate of development; in turn, H_max, L_max, T_max, _rF, and tm are coefficients determined by nonlinear regression analysis performed using Sigmaplot v11.1. These procedures were performed as previously described [15].

3. Results

3.1. Daylength-Mediated Effects of Tiller Removal On Plant Growth and Development

Daylength is expected to vary during rice cultivation. In this study, we investigated the variations in growth and development of the main stem, with and without tillers, depending on daylength conditions. The data showed that under long days plant height (H_max) tended to increase in both cultivars in the tiller-removal treatment (TR) group compared to that in the control. However, under short days, which supposedly promote flowering [15], H_max decreased in the TR group for ‘Saenuri’ but did not change for ‘Odae’ (Figure 1a, b; Table 1). Under short days, H_max of ‘Odae’ was 89.8 and 89.9 cm in the control and TR groups, respectively, whereas those of ‘Saenuri’ were 75.3 and 72.4 cm in the control and TR groups, respectively. Furthermore, under long days, the rate of increase in plant height (_rF) was lower in the TR than that in the control group, although the period of increase (t_m) in plant height was prolonged in the TR group, compared to that recorded in the control group (Figure 1a,b; Table 1). However, under short days, _rF was higher in the TR than in the control group; concomitantly, t_m was shorter for ‘Saenuri,’ whereas in ‘Odae,’ it differed slightly compared with that observed under long days.

Meanwhile, L_max of the main stem was higher in the TR group than that in the control for both varieties across the experimental conditions (Figure 1c,d; Table 1). Evidently, L_max was more influenced by the period of leaf age (tm) than by the rate of leaf age (rF), which varied between daylength conditions. Specifically, under short-day conditions, it was slightly lower in the tiller removal treatment in both cultivars; however, it differed between the two cultivars under long days (Figure 1c,d; Table 1). In contrast, t_m was longer in the tiller-removal treatment groups under all conditions for both cultivars (Figure 1c,d; Table 1). Additionally, T_max was lower under short than under long days for both cultivars (Figure 1e,f; Table 1).

Differences in growth at the heading stage, i.e., after vegetative growth was completed, in response to tiller removal as per daylength are shown in Figure 2. Similar to the results described for plant height (Figure 1a,b), stem length was greater in the tiller removal than in the control group under long days, whereas no difference was observed under short days (Figure 2a,b). Panicle length was only slightly greater in the tiller removal treatment under short days but showed a substantial increase under long days (Figure 2c,d).

Overall, the TR treatment increased vegetative growth and biomass of the main stem under long days; however, under short days, biomass was either smaller or did not significantly differ between the TR and control treatments (Figure 1, 2; Table 1).

3.2. Effect of Tiller Removal on Heading Response and Yield Components

When other relevant factors, such as fertilization, are controlled, tiller development is generally enhanced as a result of the longer growth duration [16,19,20]. Under the short-day conditions used in Experiment 1, days to heading (DTH) was shortened in the TR compared to that in the control treatment group for both cultivars (Figure S3). Further, in Experiment 2, we evaluated DTH under two daylength conditions (Figure 3) and found that under short days, DTH was shortened in the TR compared to that in the control group, similar to the results of Experiment 1. Conversely, DTH was longer in the TR than in the control group under long days (Figure 3). The ANOVA revealed that DTH did not show any significant difference due to TR but showed a highly significant difference as a result of the interaction between daylength and TR (Table S3).

Among yield components, spikelet number per panicle (SPP) showed the largest change associated with TR and daylength treatments (Table 2). In particular, SPP was lower under short than under long days for both control and TR groups (Table 2). Additionally, SPP showed a greater daylength-mediated variation in the TR than in the control group (Table 2). Furthermore, in the ‘Saenuri’ cultivar, SPP was the lowest (62.2) and highest (235.0) in the TR group under short and long days, respectively, with a 3.78-fold difference between the two extreme values (Table 2). The proportion of ripened grain slightly decreased in the TR treatment, particularly under short days, in which case SPP actually increased significantly (Table 2). Further, 1000-grain weight did not significantly differ between the experimental groups for either cultivar (Table 2). Similar results were found for Experiment 1 (Table S2).

Florigen gene Hd3a-expression levels significantly increased under short than under long days at 3 and 7 d after treatment (DAT, Figure 4a). At 7 DAT under short days, Hd3a levels were 102.4, 79.0, and 75.8 in the main stem of the treatment (TMS), control (CMS), and in the tillers of the control (CT, Figure 4a) groups, respectively. In addition, the expression of Hd3a in the main stem of the TR group tended to increase under short days and increased over time after TR treatment. However, under long days, Hd3a expression decreased compared to its level before daylength treatment, with no discernible difference between TR groups.

The MOTHER OF FT AND TFL1 (MFT1) gene reportedly increases SPP and is associated with delayed flowering [47]. Similar to the heading and SPP responses induced by daylength and TR treatments, MFT1 expression differed considerably between daylength conditions. In particular, the MFT1 expression level was higher in the CMS and CT than in the TMS groups under short days (Figure 2c); however, the opposite trend was observed under long days (Figure 2c).

Other flowering-related genes were analyzed together (Figure S4). Thus, for example, RFT1, which is another florigen gene that controls flowering under short and long days [22] showed relative expression levels of 61.1, 55.5, and 55.1 in the TMS, CMS, and CT groups, respectively, under short days (Figure S4a). Similarly, the relative expression levels of Ehd1, which enhances the expression of Hd3a and RFT1 under short-day conditions [19], were 20.1, 19.2, and 14.7 in the TMS, CMS, and CT groups, respectively (Figure S4b). Meanwhile, the relative expression levels of Ghd7 under long days, which suppresses the expression of Hd3a and RFT1 under such conditions [22], were 1.9, 1.7, and 1.4 in the TMS, CMS, and CT groups, respectively (Figure S4c).

4. Discussion

Rice is grown over a wide range of latitudes globally; therefore, the plant grows under different daylength conditions depending on the specific location. Even within the same region, it will grow under different daylength conditions depending on climate change or planting season [48].

Tillering is greatly affected by temperature and daylength [7]. Generally, short days and high temperature lead to earlier flowering as a result of the restriction of tillering imposed by the shortened vegetative growth period [10,49]. However, to date, the interaction between tillering and flowering has not been fully unraveled and in fact remains a controversial issue.

Several studies reported a negative relationship between tillering and flowering 7,10,18,20]. The FAC is a structure in which Hd3a binds to 14-3-3 proteins that act as intracellular receptors, and OsFD, a transcription factor in the bzip region, is attached to this complex and promotes floral induction [37]. These FACs activate OsMADS 14 and 15, genes located downstream from Hd3a that promote floral formation and development [37]. However, other OsFD-like transcription factors bound to FAC participate in lateral branching in the axillary meristem [41] and leaf development [50]. Therefore, Hd3a promotes lateral branching in the axillary meristem and leaf development over floral induction [37]. In addition, OsLUX-overexpressing mutants show reduced photoperiod sensitivity and a prolonged juvenile phase, which in turn results in an increased number of tillers and delayed heading 42].

However, tillering and flowering were also reported as positively correlated [35,38]. The number of tillers reportedly increases with flowering induced by the overexpression of OsRFL, which regulates the flowering activator OsSOC1. Conversely, RFL knockdown results in the restriction of development of secondary tillers and panicle branches, and delay in flowering [35]. Similarly, when the expression of OsWDRa or OsTRx1 of the COMPASS-like complex was reduced by RNA interference under long and short days, secondary branches and grain number decreased concomitant with delayed heading [38].

Here, we found that these contrasting results may be influenced by daylength (Figure 5). The data showed that, when tillers were removed, the heading stage was reached earlier under short days, whereas it was delayed under long days (Figure 3, S2). Similar results were observed when tiller development was restricted [37]. It is likely that, in conjunction with the FAC complex, Hd3a may promote floral induction rather than lateral branching in the axillary meristem under conditions of restricted tiller development [37]. Based on this hypothesis, we analyzed the florigen Hd3a in the main stem and tiller under each of these treatment conditions (Figure 4a). We observed that under short days the relative expression levels of the florigen genes increased with treatment time. Furthermore, under short days the relative expression levels of both florigens (Hd3a and RFT1) in the main stem, from which the tillers were removed, increased compared to those in the control (Figure 4a, S3a). This change in gene expression may provide a novel insight into the early heading stage transition after tiller removal under short days.

Meanwhile, Hd3a had a less prominent role under long than under short days, whereas SPP increased markedly in the TR group under long days (Table 2). For cultivar ‘Saenuri’ under short days, SPP of the TR group (59.3) was less than that of the control group (62.2); however, under long days, SPP for the TR group (235) was 3.4 times greater than that of the control group (69.1) (Table 2). In general, the heading date is delayed in situations of expanding quantitative growth, such as an increase in SPP due to excessive nitrogen supply [51]. To interpret these results, we analyzed the expression of the MFT1 gene under each treatment, given its role in the formation of spikelets and branches while suppressing flowering [47]. In contrast to short-day conditions, the expression of MFT1 in CMS and CT groups decreased compared to that in the TMS groups under long days (Figure 4b). Further, the SPP of TMS groups increased under long days (Table 2). Therefore, the phenotype of rice plants, including heading date and SPP, corresponded to the expression of MFT1.

Summarily, under short-day conditions, in which case floral induction is optimal, Hd3a may play a more prominent role in floral induction of the main stem than that in tillering (i.e., lateral bud formation) in the TR treatment; thus, the time required to reach the heading stage may be shortened (Figure 2a). In contrast, the difference in SPP between TR and control groups under long days was significantly greater than that under short days. Additionally, the characteristics of the expression of the related gene, MFT1, also showed the same pattern; specifically, MFT1 expression significantly increased in the TR group than in the control group only under long days. Furthermore, the expression levels of MFT1 showed the opposite pattern to those of Hd3a and RFT1 under short days (Figure 2c). Meanwhile, the larger increase in SPP in the TR treatment was the primary factor contributing to the observed delay in flowering under long days, in contrast to that of the results obtained under short days.

5. Conclusions

It is essential to implement climate-smart crop production schemes to address the current challenges in meeting global food security requirements and effectively managing critical variables during the cropping season. Our study effectively identified the plant physiological traits required to optimize growth timing and shorten the rice tillering period and the time to flowering, while maintaining high crop productivity. In particular, our study revealed that tillering and flowering, whose study has frequently rendered conflicting results, showed daylength-dependent variability. Furthermore, our findings have the potential to contribute to a novel interpretation of these phenomena, ultimately enhancing the effectiveness of plant efforts to adapt to ongoing climate change by manipulating the daylength dependency of the interaction between rice tillering and flowering.