4. DISCUSSION

“Early flowering” and “high yielding” are important goals for crop domestication and improvement, but they are contradictory. The duration and fate of floret primordia development in wheat have a significant impact on this contradiction; for example, the shorter the development duration, the fewer floret primordia survive to reproduce (Ghiglione et al. , 2008; Watson et al. , 2018). A thorough understanding of floret development can assist in overcoming the trade-off between “early flowering” and “high yielding” and achieving faster and more wheat production.
From the terminal spikelet to flowering, wheat floret primordia in SW treatment went through differentiation (Stage I), differentiation and morphology development concurrently (Stage II), morphology development (Stage III), and polarization (Stage IV), according to the high temporal-resolution investigation of floret primordia number, morphology, and spike transcriptome (Figure 2, Figure 4). During Stage IV, there were two morphologies of floret primordia, including fertile and infertile floret primordia (Figure S1b). Although the initiation of polarization varied between treatments (24 DAT in SW and 21 DAT in LW and LR; Figure 2e), all infertile floret primordia degenerated rapidly and finally became invisible (Figure 2d, 2g). This result was consistent with previous research, which proposed that initiating abortive floret primordia would inevitably degenerate and fail to develop into fertile florets even when the plant was grown under optimal conditions, such as adequate water and fertilizer (Zhang et al. , 2021). Meanwhile, we also found that there was a stable difference in the number of fertile floret primordia between treatments in the polarization, and almost all fertile floret primordia in the three light regimens continued their development and finally became fertile florets at flowering (Figure 2f). These results showed that infertile floret primordia could not be reversed to fertile floret primordia and that the number of fertile floret primordia could not be increased by light treatments in the polarization, suggesting that Stage IV had little effect on improving NFFs.
Fertile floret primordia in Stage IV, which originated from a portion of floret primordia in the first three stages, have greatly determined NFFs. Therefore, promoting the growth and development of floret primordia during Stage I-III was essential for increasing NFFs. Assimilate availability in the spike significantly affected NFFs; for example, the higher the dry spike weight at flowering, the more the NFFs (Figure 3), which was consistent with previous studies (González et al. , 2011; Zhang et al. , 2020). Hence, the accelerated growth and development of the most advanced floret primordium (i.e., F1; Figure 2d), as well as the increased dry spike weight and soluble sugar contents (Figure 3), observed in LW and LR treatments, were required for high NFFs (Ghiglione et al. , 2008; Ugarte et al. , 2010). However, comparisons between treatments (LW versus SW and LRversus LW) showed that NFFs and the number of fertile floret primordia were decreased in LW treatment but increased in LR treatment (Figure 2b, 2f). The maximum number of differentiable floret primordia did not differ between treatments (Figure 2e), supporting historical findings that the number of floret primordia was not a limiting factor for NFFs improvement (Guo & Schnurbusch, 2015; Zhang et al. , 2021). Some floret primordia, such as F6 (Figure 2d), were either interrupted by the advance of degradation initiation in LW treatment, resulting in a shorter growth duration and lower dry spike weight, or did not develop as quickly in LW treatment as in LR treatment due to inadequate assimilate supply, which prevented primordia from acquiring fertile potential comparable to SW and LR treatment. Increasing the available assimilates and promoting the development of more floret primordia to the scale of fertile floret primordia during the shortened Stage I-III became crucial for increasing wheat grains.
In comparison to SW, LW treatment accelerated flowering but sacrificed NFFs (Figure 2a, 2b), which was consistent with previous findings (Ghiglione et al. , 2008; Watson et al. , 2018). Furthermore, our findings suggested that the shortened Stage III in LW treatment was mainly responsible for flowering acceleration (Figure 4). In analogy with the conserved photoperiodic flowering pathway in Arabidopsis and rice (Fornara et al. , 2010; Zhou et al. , 2021), we also constructed a GI -CO -like-HD3Aframework during wheat floret development that played a pivotal role in accelerated flowering (Figure 5). First, LW treatment significantly regulated GI and its directly related IAA16 , which contributed to accelerated differentiation and growth of floret primordia. Then, LW treatment regulated the CO 3-centered gene network, which included several MYB transcription factor genes; this accelerated anther development, induced the up-regulation ofCOL2 , COL5 , and GHD7 and the down-regulation ofCOL3 , and caused Stage IV initiation to be advanced from 24 DAT to 21 DAT. Finally, HD3A synergized with its directly related genes involved in pollen tube growth to induce early flowering in LW treatment. Since our study found that wheat plants grown in LR treatment had earlier flowering and higher NFFs than those grown in LW treatment (Figure 2a, 2b), it would be worthwhile to explore how LR treatment regulated floret development at different stages.
Floret primordia under LR treatment mainly underwent Stage I, III, and IV, which corresponded to Stage I, II, and IV under LW treatment, respectively, based on the sampling times (Figure 4), indicating that LR treatment modulated gene expression at 15 DAT and 18 DAT, allowing spikes to reach the Stage III criterion more quickly. Compared to LW, LR treatment significantly up-regulated the expression of genes encoding ubiquitin and calcium signaling, which accelerated cell proliferation and histogenesis during Stage I (Figure 6e, Figure 7). Almost all floret primordia were generated in Stage I under LR treatment, whereas they required Stage I to II under LW treatment (Figure 2e). Hence, LR treatment mainly shortened the development time of floret primordia in Stage II, inducing early flowering. During Stage III, LR treatment significantly increased the response of the gene network centered on ubiquitin and ALDH s involved in primary and secondary metabolite production, cell wall biogenesis, and anther development (Figure 6e, Figure 7). This showed that floret primordia in LR treatment underwent adequate morphological development in Stage III, so that more floret primordia acquired fertile potential before polarization. During Stage IV, wheat spikes were directly exposed to light, requiring plants to coordinate growth, development, and defense (Lazzarin et al. , 2021). The gene network centered on ZF s and HSP s was strongly regulated by LR treatment, which not only accelerated environmental adaptation by promoting abscisic acid and jasmonic acid signaling pathways, but also improved stamen filament and pollen tube development in preparation for successful fertilization (Figure 6e, Figure 7). Furthermore, significant down-regulation of genes involved in carbohydrate, lipid, and fatty acid metabolism under LR treatment did not prevent the continued development of fertile floret primordia, indicating that LR treatment acted to maintain fertile floret primordia development during Stage IV until flowering (Figure 2f, Figure 6e). In summary, LR treatment accelerated floret primordia differentiation in Stage I, shortened Stage II, enhanced morphological development of floret primordia and increased the number of fertile floret primordia in Stage III, and maintained fertile floret primordia development in Stage IV through a gene regulatory network centered on ubiquitin, calcium signaling, ALDH s, ZF s, and HSP s, ultimately achieving the dual promoting effects of accelerating flowering and increasing NFFs.
Collectively, we proposed a modulated light regimen (i.e., long photoperiod supplemented with red light) and a dynamic regulatory model that could remove the barrier between “early flowering” and “high yielding” during wheat floret development. Early flowering could be achieved by shortening the development time of floret primordia in Stage II or Stage III. High yield could be obtained by increasing assimilate availability and promoting the morphological development of more floret primordia to acquire fertile potential during Stage III. These results provide important resources for future improvements in the “early flowering” and “high yielding” of wheat.