1. INTRODUCTION

Wheat production plays an important role in global food security, but it is limited to only one or two generations per year due to time-consuming field growth (Mueller et al. , 2012). To accelerate the developmental rate of plants, controlled environment agriculture independent of outdoor conditions has been proposed, which allows for multiple harvests throughout the year (van Delden et al. , 2021). Wheat is typically a long-day plant, and a long photoperiod can promote early flowering of wheat up to six generations per year in a controlled environment (Watson et al. , 2018), which is predicted to be a promising option for increasing future crop production (Asseng et al. , 2020). Grains per spike is one of the most important contributors to yield in wheat, but early flowering induced by a long photoperiod usually sacrifices grains (Ghiglione et al. , 2008; Rossi et al. , 2015). The contradiction between ”early flowering” and ”high yielding” is widespread, hampering further improvement of the annual wheat yield. Outside of the photoperiod, light spectral quality is also an important light variable that has a profound effect on plant growth and development (Lazzarin et al. , 2021). For example, red light has outstanding potential for increasing wheat grains (Dreccer et al. , 2022). Is it possible to attain “early flowering” and “high yielding” simultaneously in wheat cultivation using the long photoperiod supplemented with red light regimen?
Grains per spike in wheat is composed of spikelets per spike coupled with grains per spikelet (Gauley & Boden, 2019). Studies have found that adding more spikelets, like paired spikelets (Boden et al. , 2015) and supernumerary spikelets (Li K et al. , 2021; Wang et al. , 2022), can make more wheat grains. When spikelets remain constant, it is necessary to look into ways to increase grains even more. As a cleistogamous species, the number of wheat grains at maturity is largely determined by the number of fertile florets per spike (NFFs) at flowering (Figure S1a; Ferrante et al. , 2020). In general, up to 6-12 floret primordia are produced in a spikelet; however, only a small proportion of them develop into fertile florets, while others degenerate (Prieto et al. , 2018). Floret primordia are initiated and developed asynchronously in a spike and thus have uncertain fates (fertility or infertility at flowering), presenting a great opportunity to increase NFFs and grains by manipulating floret primordia development. Since stems and spikes are both growing rapidly during floret development, there is intensive intra-plant competition for available assimilates. In studies with prolonged photoperiod conditions or photoperiod-insensitive genes, NFFs and grains went down due to decreased assimilate accumulation in the spike (Ghiglione et al. , 2008; Prieto et al. , 2018); in contrast, sucrose-feeding (Ghiglione et al. , 2008) and well-watered conditions (Zhang et al. , 2020) suggested that increasing assimilate accumulation in the spike could increase floret survival rate and NFFs. Other studies, however, found that although low red/far red ratios reduced competition for assimilates in the uppermost stem (Casal, 1993) or improved assimilate availability in the spike delayed the death of abortive floret primordia (Zhang et al. , 2021), they did not increase NFFs. Focusing on the number of floret primordia and ignoring the morphological differences demonstrated by floret primordia may result in a different effect of assimilate availability on floret development and NFFs. Floret primordia that develop into or beyond the green anther and elongated stigma branch state (i.e., W7.5 according to the Waddington scale; Waddington et al. , 1983) usually become fertile florets at flowering (Zhang et al. , 2021). Similarly, we found that floret primordia with morphological scores greater than W7.5 at the onset of some floret primordia degeneration had a reliable potential to achieve fertility at flowering (R2 = 0.89, Figure S1b) and were therefore referred to as fertile floret primordia (Figure S1c). The development of all floret primordia in a wheat spike is a complex process that lasts about 30 days from the terminal spikelet stage to flowering stage. However, the variation in primordia quantity and morphology during this process, as well as their relationship to assimilate availability, is not well understood.
Flowering time is an important selection characteristic for crop improvement. Numerous genes involved in light-induced flowering have been identified. For example, overexpression of the circadian clock gene GIGANTEA (GI)can accelerate wheat flowering (Zhao et al. , 2005); the zinc finger transcription factor gene CONSTANS (CO ), which is controlled by GI , induces flowering when CO expression coincides with light (Valverde et al. , 2004); and the floral integrator gene FLOWERING LOCUS T (FT ) is activated by high CO expression in long-day plants (Nicolas Freytes et al. , 2021). Meanwhile, grains per spike is also a crucial selection feature for crop improvement and is influenced by the fate of each floret primordium. Many regulators responsible for floret fertility have been identified. For instance, defective expression of UNDEVELOPED TAPETUM1 (UDT1 ) andPERSISTENT TAPETAL CELL1 (PTC1 ) results in abnormal anther and pollen morphology, leading to male infertility (Li et al. , 2011; Yang et al. , 2019); aldehyde dehydrogenase (ALDH) activity provided by nuclear genes rf1 and rf2 is required to restore male fertility in male-sterile cytoplasm maize (Liu et al. , 2001); the 70 kDa heat shock protein (HSP) family has emerged as being indispensable for male fertility (Nixon et al. , 2017). The morphogenesis of male organs (anthers) seems to play a more important role than the ovaries in determining floret fertility (Ji et al. , 2010). Enhancing the morphological development of floral anthers may be a powerful way to increase NFFs. Although several genetic means for regulating flowering time or transforming floret fate have been proposed (Sakuma & Schnurbusch, 2020), flowering and yield cannot be simultaneously promoted through a single gene or biological process. High-throughput RNA sequencing (RNA-seq) is a powerful tool for comprehensively investigating developmental dynamics with high sensitivity and accuracy (Yi et al. , 2019; Thiel et al. , 2021). This provides an opportunity to clarify the gene regulatory network of light regimens on wheat floret development, flowering acceleration, and yield enhancement.
In this study, we aimed to dissect the complexity of assimilate partitioning, morphological development, and transcriptional dynamics during floret development and to clarify how the long photoperiod supplemented with red light regimen regulates flowering and yield in wheat. Our discoveries not only provided a guide for controlled wheat cultivation but also revealed critical stages and candidate genes for wheat genetics and breeding improvement during floret development.