Beyond heading time: FT-like genes and spike development in cereals.

Development of the grain-bearing organ, or spike (inflorescence) is critical to cereal grain development and yield.

Flowering time is one of the most crucial target traits in crop breeding programmes due to its high correlation with final grain yield. Flowering plants respond to environmental cues to flower at a suitable time to maximize reproductive success, and so by optimizing this process breeders maximize grain yield (Jung and Muller, 2009). It is a highly complex trait that is determined by both environmental and endogenous factors. Integration of various environmental cues triggers the expression of florigen genes in leaves, in turn activating the expression of floral meristem identity genes in the shoot apical meristem and initiation of reproductive growth (Andres and Coupland, 2012).

Over the past two decades, numerous pleiotropic genes that regulate both flowering time and grain yield have been characterized in crops, such as the Grain number, plant height, and heading date (Ghd) gene series, Ghd7, Ghd8, Ghd7.1 and Ghd6 in rice and Photoperiod-1 (Ppd-H1) in barley (Yan ; Boden ; Zhang ). Interestingly, all these genes regulate heading time – the number of days from sowing to emergence of the grain-bearing organ – by suppressing the expression of FLOWERING LOCUS T-like (FT-like) florigen genes and increasing grain yield. The findings of Shaw add important new information to this body of work, helping show how FT-like genes regulate grain yield in cereals.

Diverse functions of FT-like genes

The phosphatidylethanolamine binding protein (PEBP) gene family is involved in regulation of flowering time, seed dormancy and panicle/spike development. It can be divided into three subfamilies: MOTHER OF FT AND TFL1-like (MFT-like), TERMINAL FLOWER1-like (TFL1-like) and FT-like. The MFT-like subfamily, the proposed ancestor of the other two subfamilies, regulates seed germination, flowering time and spikelets per panicle. The duplication and diversification of MFT-like genes eventually resulted in the FT-like and TFL1-like subfamilies (Wickland and Hanzawa, 2015). Comparison of the phenotypes donated by FT-like and TFL1-like genes in cereals (Table 1) shows that FT-like genes mainly induce flowering, while TFL1-like genes mainly possess anti-florigen activity which represses flowering.

Table 1.

FT2 and its homologue(s) in rice, maize, barley and wheat

Species	Clade	Genes	Flowering time	Floret number	Reference
Rice	FT-like	Hd3a/OsFTL2	Earlier flowering	Decrease	Kojima et al., 2002
	FT-like	RFT1/OsFTL3	Earlier flowering	Decrease	Zhao et al., 2015; Zhu et al., 2017
	FT-like	OsFTL1	Earlier flowering	Decrease	Izawa et al., 2002
	TFL1-like	RCN1-4	Later flowering	Increase	Kaneko-Suzuki et al., 2018
Maize	TFL1-like	ZCN1	Later flowering	Increase (in tassel)	Danilevskaya et al., 2010
	TFL1-like	ZCN2	Much later flowering	Increase (in tassel)	Danilevskaya et al., 2010
	TFL1-like	ZCN3	Unchanged	Increase (in tassel)	Danilevskaya et al., 2010
	TFL1-like	ZCN4	Much later flowering	Increase (in tassel)	Danilevskaya et al., 2010
	TFL1-like	ZCN5	Much later flowering	Increase (in tassel)	Danilevskaya et al., 2010
	TFL1-like	ZCN6	Unchanged	Increase (in tassel)	Danilevskaya et al., 2010
	FT-like	ZCN8	Earlier flowering	Decrease	Danilevskaya et al., 2011; Meng et al., 2011
Barley	FT-like	HvFT1/VRN3	Earlier flowering	Unknown	Kikuchi et al., 2009
	FT-like	HvFT2	Earlier flowering	Unknown	Kikuchi et al., 2009
	FT-like	HvFT3/Ppd-H2	Earlier flowering	Unchanged in LD but aborted in SD	Kikuchi et al., 2009; Mulki et al., 2018
	TFL1-like	HvCEN	Later flowering	Increased yield	Comadran et al., 2012
Wheat	FT-like	TaFT1/VRN3	Earlier flowering	Decrease	Lv et al., 2014
	FT-like	TaFT2	Slightly earlier flowering	Increase	Shaw et al., 2018
	FT-like	TaFT3	Earlier flowering	Unknown	Zikhali et al., 2017

Lv reported FT1 as a florigen candidate promoting flowering in wheat. Shaw add to this with their characterization of FT2, the closest paralogue of FT1, in the tetraploid wheat variety Kronos. The ft2 mutant shows slightly delayed heading time but greatly increased number of spikelets per spike and florets per spikelet. Although FT2 is highly expressed in leaves and under the control of VRN1, VRN2 and Ppd-1, the null mutant only causes a delay in heading time of 2–4 days; this compares with the null mutant of FT1 which causes a delay of more than 20 days. Therefore, FT2 is not primarily acting as a florigen despite functional FT2 being expressed in leaves. The double null mutant of ft1 and ft2 has a comparable flowering time to the ft1 single mutant, which indicates that the FT2 effect is independent of FT1, and other florigen genes exist besides FT1 in wheat.

Unlike other FT-like and TFL1-like genes, most of which have major effects on heading time, FT2 shows a major effect on spikelets per spike and fertility but minor effect on heading time. Besides being highly expressed in leaves, FT2 is also highly expressed in the distal part of the spike. The ft2 mutant greatly increases the floret number and dramatically reduces the fertility of florets on the spike (Shaw ). Thus, FT2 mainly appears to function in spike development. Notably, sterility is observed not only in the ft2 mutant in wheat, but also in transgenic knockdown B. distachyon and barley, indicating for the first time that FT2 control of fertility may be a conserved function in temperate cereals. It is clear that FT-like subfamily genes also exhibit some degree of sub-functionalization in cereals.

Spikelets per spike and heading time

Florigen interacts with the bZIP transcription factor FD and 14-3-3 proteins in the shoot apical meristem to form the florigen activation complex. This activates floral meristem identity genes such as OsMADS15 in rice or VRN1 in wheat (Li and Dubcovsky, 2008; Taoka ), which initiates the transition from the vegetative to the reproductive phase. Next, the shoot apical meristem begins to generate the inflorescence meristem. RICE CENTRORADIALIS (RCN) belongs to the TFL1-like subfamily. RCNs are expressed in the vasculature and move to the shoot apical meristem, where they compete with Hd3a for FD and 14-3-3 to form the florigen repression complex and regulate inflorescence development (Kaneko-Suzuki ). Members of the PEBP family seem to interact with FD and 14-3-3 to form a complex. Different complexes and their functions in heading time control and later panicle/spike development are still to be confirmed.

Overexpression of RCN greatly delays heading time and generates a larger inflorescence meristem, which produces a long and dense panicle in rice (Kyozuka ; Kaneko-Suzuki ). The heading date genes upstream of florigen genes such as Ghd7 in rice probably induce a larger inflorescence meristem by greatly delaying heading date, and finally significantly increase grain yield. Moreover, it seems that plants with a reasonable delay of heading time produce more spikelets per panicle and more grain yield (Zhang ).

FT2 represents another type of FT-like genes that greatly increase number of spikelets per spike (spike size) but with minor effects on flowering time. FT2 is mainly expressed in the inflorescence meristem to increase spikelets per spike after phase transition. This working model is similar to the scenario that OsMFT1 mainly functions in inflorescence meristem to down-regulate the expression of floral identity genes (e.g. FZP, OsMADS1) as floral organ determinants (Song ). To this end, the spikelets per panicle trait controlled by floral identity genes such as FZP is usually independent of flowering time (Bai ).

Potential breeding applications

Negative effects of FT2 on fertility limits its potential for breeding high yield wheat cultivars without significantly delaying heading date (Shaw ). Low fertility of both FT2 knockout mutants in wheat and RNAi mutants in barley indicates that non-functional or very low expression FT2 alleles are unlikely to be widely distributed. However, promoter variation could enable a moderate change in FT2 transcript level. Thus, investigation of natural variation of FT2 in wheat germplasm should be encouraged to seek regulatory elements which can generate a favourable FT2 allele to maximize grain yield by balancing the trade-off between spikelets per spike and fertility. This potential FT2 allele would be more valuable in developing high yield cultivars regardless of FT1 background. Alternatively, a series of FT2 RNAi plants with varied interference are worth generating by transformation: the new germplasm with optimized transcript level of FT2 would be screened in terms of yield performance from these RNAi plants. As suggested by the authors, the Ft-B1-Hope allele could probably compensate for the negative effect on fertility of FT2 (Shaw ). Identification of its downstream genes by ChIP sequencing and screening FT2-interaction proteins would help to separate the positive effects on spikelet per spike and negative effects on fertility.

To conclude, the features of FT2 provide us new opportunities to understand the mechanism of FT-like gene in flowering time and grain yield control. If its negative effects on fertility can be resolved, FT2 is a good candidate for manipulation in breeding high yield cultivars.