The process of wheat domestication involved several hybridization events between tetraploid progenitor species Triticum turgidum L. (AABB genomes) and diploid Aegilops tauschii (DD genome) that gave rise to hexaploid bread wheat (AABBDD genomes) about 8,000 years ago (Wang et al., 2013). As a hexaploid, bread wheat displays a high level of gene redundancy (one copy of the homoeologous gene in each genome), and therefore the precise function of many flowering-related genes and the current understanding of the complexity of the gene networks controlling flowering is still incomplete. Table 1 provides a summary of the current understanding of wheat phenology genes which are either functional orthologs or homologs of Arabidopsis thaliana genes.

Rapid genomic changes in polyploid wheat facilitated wheat adaptation to different growing environments. In particular, wheat crop expansion is associated with the exploitation of natural variation in the photoperiod (PPD), vernalization (VRN), and EARLINESS PER SE (EPS) genes. In bread wheat, PPD genes are relatively well understood and molecular markers have been developed to aid selection for breeders (Beales et al., 2007; Cane et al., 2013). Three copies of the gene (homeologs) that control photoperiod response in wheat, namely TaPPD-A1, TaPPD-B1 and TaPPD-D1, are located on the short arm of the homoeologous group 2 chromosomes. The region containing TaPPD-D1 in wheat is colinear with the barley photoperiod gene HvPPD-H1 on chromosome 2HS (Laurie, 1997). TaPPD-D1 encodes a pseudo-response regulator (PRR) family protein gene orthologous to the Arabidopsis PRR7 gene, and is more distantly related to AtCO and AtVRN2 (Yan et al., 2004; Turner et al., 2005).

The wild ancestors of wheat were photoperiod-sensitive, and photoperiod insensitivity in many cultivated wheat varieties is the result of mutations in PPD genes (Thomas and Vince-Prue, 1997). Dominant PPD alleles induce constitutive activation of the photoperiod pathway irrespective of day length and significantly reduce sensitivity to photoperiod. This leads to an early flowering phenotype in both SDs and LDs, which has been associated with pleiotropic effects in certain agricultural environments including in southern Europe (Worland, 1996) and Australia (Richards et al., 2014) resulting in increased grain yields. The potencies of the homoeologous group-2 PPD genes for insensitivity are ranked in the order TaPPD-D1 > TaPPD-B1 > TaPPD-A1 (Worland et al., 1998).

Pseudo-response regulator proteins are commonly found in plants and involved in circadian clock-associated pathways, including PRR1 (TOC1) (Mizuno and Nakamichi, 2005). The mRNA of TOC1 starts to accumulate at the beginning and reaching its maximum level at the end of the light period, where it directly represses the transcription of the myeloblastosis (MYB) transcription factors LATE ELONGATED HYPOCOTYL (LHY) and CCA1. In wheat, Zhao et al. (2016) recently demonstrated that the microRNA tae-miR408 targets and down-regulates expression levels of all three wheat paralogs of TaTOC1 (TaTOC-A1, TaTOC-B1, and TaTOC-D1) under both LDs and SDs, thereby regulating heading time and associated agronomic traits.

The extensive natural variation for both the requirement and responsiveness to cold temperature has played a major role in the adaptation of many species over a wide latitudinal range. Unlike the model monocot crop rice, wheat and barley flower in response to LDs. The photoperiod and vernalization pathways intersect at the regulation of TaVRN3, which is a gene encoding a phosphatidylethanolamine-binding protein (PEBP) homologous to the Arabidopsis flowering time gene AtFT1 (Yan et al., 2006). Similarly, VRN3 also encodes a mobile protein that is transported from the leaves to the SAM, where it becomes part of a flowering complex that induces TaVRN1 transcription by binding to its promoter. TaVRN1 is an AP1 clade MADS-box transcription factor, homologous to the Arabidopsis meristem identity gene AtAP1, and present on the long arms of chromosomes 5A, 5B, and 5D, respectively (TaVRN-A1, TaVRN-B1, and TaVRN-D1) (Yan et al., 2003). While vernalization in Arabidopsis results in silencing of the flowering repressor AtFLC, it induces transcription of the flowering activator TaVRN1 in wheat and other monocots. TaVRN1 is expressed in the leaves where it represses transcription of TaVRN2, which encodes a protein with a putative zinc finger and a CCT protein–protein interaction domain (Yan et al., 2004; Kippes et al., 2015). TaVRN2 represses TaVRN3 expression in LDs to prevent flowering during unfavorable conditions in autumn before vernalization. In the absence of TaVRN2, the increase in day length during spring results in the up-regulation of TaVRN3 expression, which leads to a further increase of TaVRN1 transcription trough a positive feedback loop, and ultimately to an irreversible acceleration of flowering.

A fourth vernalization locus, TaVRN-D4, was recently identified on the short arm of chromosome 5D (Kippes et al., 2015). It was shown to be paralogous to TaVRN-A1, as it originated from an insertion of a large segment on chromosome 5AL containing the TaVRN-A1 gene into chromosome 5DS. The first intron of TaVRN-A1 contains binding sites for the RNA-binding protein GLYCINE-RICH RNA-BINDING PROTEIN 2 (TaGRP2), which binds to TaVRN-A1 pre-mRNA and represses TaVRN-A1 expression (Xiao et al., 2014). The inserted copy of TaVRN-A1 carries mutations in its coding and regulatory regions including single nucleotide polymorphisms (SNPs) in the first intron that impedes binding of TaGRP2. Protein modification of TaGRP2 by O-linked β-N-acetyl glucosamine (O-GlcNAcylation) mediates interaction with VERNALIZATION-RELATED 2 (TaVER2), a carbohydrate-binding jacalin lectin, which relieves repression of TaVRN1 expression via TaGRP2 when vernalization occurs.

The genetic regulation of vernalization response in wheat (and other monocots) has some distinct differences from Arabidopsis. Comparative genomics studies detected Arabidopsis FLC-like genes in monocots, but they have yet to be shown to produce proteins of similar function in wheat or other temperate cereals (Ruelens et al., 2013). AtFLC is positively regulated by AtFRI, but to date also no functional AtFRI homologs were detected in wheat. The AtFT1 and AtFRI promoters and first introns were shown to contain cis-regulatory sites important for the transcriptional regulation of these genes in Arabidopsis (Tiwari et al., 2010; Sanchez-Bermejo and Balasubramanian, 2015).

Only comparatively little is currently known about genetic sequence variations and non-gene elements that regulate phenology gene expression in wheat. Autumn-sown winter cultivars that contain the ancestral VRN1 allele require vernalization during winter. By contrast, spring wheat varieties carrying mutations in regulatory regions including in the TaVRN1 promoter, first intron of TaVRN1, or in the TaVRN4 region (Fu et al., 2005; Kippes et al., 2015; Muterko et al., 2015), do not require vernalization and are sown in spring. Very recently, Kippes et al. (2016) developed a triple TaVRN2 mutant that contained three non-functional TaVRN2 alleles. This mutant flowered early, had a limited vernalization response, and spring growth habit. In a different study, an insertion in the promoter of TaVRN3 led to increased TaVRN3, bypassing the delay of flowering by TaVRN2 (Yan et al., 2006). Sequence variations in TaVRN3 genes were shown to cause early flowering in wheat (Ó’Maoiléidigh et al., 2014). As another recent example, copy number variation (CNV) was shown to play a major role in global wheat adaptation. Wheat lines with an increased copy number of TaVRN-A1 showed an increased requirement for vernalization, and lines with an increased copy number of TaPPD-B1 led to early flowering (Würschum et al., 2015).

The third class of genes responsible for fine-tuning of wheat flowering time are EPS genes, which can be defined as the least number of days to reproductive growth independent of vernalization and photoperiod. Also known as earliness in the narrow sense, intrinsic earliness, or basic development rate, EPS genes are hypothesized to function as a fine tune adjustment of flowering time (Zikhali and Griffiths, 2015). They are often considered as polygenic and of small genetic effect, and can be detected in both winter and spring wheats. Variation in EPS genes was shown to lead to minor or major differences in flowering time under field conditions, reported in the range of a few days up to a few weeks (Appendino and Slafer, 2003; Zikhali et al., 2014). To date, no EPS genes have been cloned, there is only a limited understanding of gene function, and furthermore, no known homologs in Arabidopsis exist that could help to fill the knowledge gaps. Instead, several studies used near isogenic lines (NILs) to narrow down the location of EPS loci on the wheat genome, and to identify several potential candidate genes (Zikhali et al., 2014, 2015). The genes MOLYBDENUM TRANSPORTER 1 (TaMOT1), FILAMENTATION TEMPERATURE SENSITIVE H (TaFtsH4) (Faricelli et al., 2010), and EARLY FLOWERING 3 (TaELF3) (Zikhali et al., 2014) were proposed as candidate genes underlying the EPS -Am1 locus. Gawroński et al. (2014) suggested a cereal ortholog of Arabidopsis circadian clock gene LUX ARRHYTHMO/PHYTOCLOCK 1 (AtLUX/PCL1) as the candidate gene underlying the EPS -3A^m locus in einkorn wheat (Triticum monococcum L.). These recent findings question the previous assumption that EPS acts independently of the environment, but the exact nature of the environmental impact on EPS gene function remains unclear.

Previously, Dubcovsky et al. (2006) showed that an inductive LD photoperiod is needed to induce additional genes for normal spike development as the induction of TaVRN1 alone proved insufficient. Pearce et al. (2013) hypothesized that GAs may be involved in wheat spike development during LDs, which points to a functionally different role compared with Arabidopsis where GAs function in non-inductive SD conditions (Hyun et al., 2016). Pearce et al. (2013) applied exogenous GA to in wheat lines grown under non-inductive SDs, which resulted in an accelerated spike development only wheat lines expressing TaVRN1. Both GA and TaVRN1 are required for the up-regulation of the floral meristem identity genes TaSOC1 and TaLFY and the development of the wheat spike. Furthermore, GA biosynthetic gene expression was found to be elevated in the apices of plants transferred from SDs to LDs as well as in photoperiod-insensitive and transgenic wheat plants with increased TaFT transcription under SDs. Wheat genes for components of the GA biosynthetic pathway were identified based on homology with Arabidopsis, rice, and Brachypodium distachyon (Pearce et al., 2015), but due to the lack of a fully sequenced wheat genome, it is currently not possible to accurately determine the final number of genes in each family.

GA-sensitive and insensitive dwarfing genes have had an impact on all the main cereal crop species (Jia et al., 2009). Semi-dwarfing wheat varieties containing alleles of the REDUCED HEIGHT (RHT) DELLA genes were key factors to increasing yield during the “Green Revolution” (Flintham et al., 1997; Peng et al., 1999). Reduced crop height protected against lodging and improved harvest index, high spikelet fertility, and grain numbers per ear. RHT encodes a negative regulator of the GA signaling pathway (Pearce et al., 2013; Boden et al., 2014).

It is estimated that the common ancestor of barley and Arabidopsis shared about two-thirds of the circadian clock and clock-associated genes (Calixto et al., 2015) (Table 2). After separation into monocot and dicot classes, these genes further diversified due to extensive gene duplication and deletion events (Campoli et al., 2012a). Although some core clock genes were identified in barley based on map-based cloning (Yan et al., 2003), positional cloning (Yan et al., 2004), or homology with Arabidopsis (Yan et al., 2006), only a few are well characterized. Some of these genes were found to perform similar functions as their orthologs, whereas there is currently little evidence for other genes that would support their involvement in flowering.

The major genes controlling flowering time in barley in response to environmental cues are HvVRN-H1, HvVRN-H2, HvVRN-H3 (HvFT1), HvPPD-H1, and HvPPD-H2 (HvFT3) (Figure 2). The epistatic relationship of three genes involved in flowering regulation is well established: HvVRN-H1 and HvVRN-H2, with roles in vernalization, as well as the floral pathway integrator gene VRN-H3, which is synonymous to HvFT1 (Kikuchi et al., 2009).

Similarly as for Arabidopsis and wheat, FT-like genes in barley such as HvFT1 trigger flowering in response to LDs. HvFT1 plays a key role in promoting flowering and integrates the photoperiod and vernalization pathways: Under LDs, HvFT1 expression is low until induced by low temperatures in winter varieties that have a vernalization requirement (Yan et al., 2006). HvFT1 expression is primarily regulated by the major photoperiod response genes HvPPD-H1 and HvPPD -H2 (Laurie et al., 1994). Photoperiodic flowering under LDs is up-regulated by HvPPD-H1, which is a homolog of the Arabidopsis clock gene PRR7 (Turner et al., 2005). A mutation in the CCT domain of HvPPD-H1 is associated with lower transcript levels of HvFT1 and delayed flowering under LDs compared with the wild-type HvPPD-H1 allele, but it is not related to flowering variation under SDs (Turner et al., 2005; Hemming et al., 2008). HvPPD-H2 (with its proposed candidate gene HvFT3) enhances the expression of HvFT1 and promotes flowering under non-inductive SDs, whereas recessive mutant alleles confer delayed flowering under SDs, aiding the flowering repression over winter (Kikuchi et al., 2009).

In contrast to Arabidopsis, HvFT1 expression is induced via HvCO1 (the closest ortholog of Arabidopsis CO in barley) in both SDs and LDs (Campoli et al., 2012a). There is a high level of redundancy in both CO-like and FT-like genes, with currently nine and five members in barley, respectively (Faure et al., 2007). HvCO1 and HvCO2 are believed to be paralogs that exist due to a duplication event in temperate cereals. This functional diversification is important for the modulating of flowering responses and adaptation to different growing environments. Several studies demonstrated that functional polymorphisms in the HvFT1 gene alter its regulation, including mutations in the first intron that differentiate dominant from recessive alleles (Yan et al., 2006), and polymorphisms in the HvFT1 promoter region that lead to distinct phenotypic effects (Casas et al., 2011).

In barley, the VRN-H2 locus consists of three homologous CO-like genes, HvZCCT-Ha, HvZCCT-Hb, and HvZCCT-Hc (Dubcovsky et al., 2005; Karsai et al., 2005). VRN-H2 functions as a floral repressor of HvFT1 (VRN-H3) to delay flowering in plants that have not been vernalized (Trevaskis et al., 2006; Hemming et al., 2008). VRN-H2 is only expressed under LD conditions controlled by components of the circadian clock. Mutations in HvELF3 resulted in the expression of VRN-H2 under SD conditions due to an increased expression of PPD-H1 and, consequently, HvFT1 (Faure et al., 2012; Turner et al., 2013).

In winter barley cultivars, vernalization induces VRN-H1 to repress VRN-H2 which promotes the transition from vegetative to reproductive development (Trevaskis et al., 2006; Hemming et al., 2008). Sequence variations in the first intron of VRN-H1 alter the vernalization requirement for activation and, consequently, the repression of VRN-H2 and overall flowering behavior (Hemming et al., 2008, 2009). A recent survey showed that the down-regulation of VRN2 by cold is exclusively found in Pooid grasses including wheat and barley (Woods et al., 2016).

By contrast, spring barley cultivars are characterized by natural mutations at PPD-H1 and by deletions of the VRN-H2 locus and thus do not require vernalization (Dubcovsky et al., 2005). HvCO2 overexpression in spring barley was shown to induce flowering due to an increased expression of PPD-H1 and, as a direct result, HvFT1 (Mulki and von Korff, 2016). By contrast, overexpression of HvCO1/CO2 increased HvVRN-H2 expression in winter barley carrying the VRN-H2 locus, which led to lower HvFT1 levels and delayed flowering independently of photoperiod. Putative additional epistatic interactions of HvVRN-H2 with HvGI and HvCO1 point toward a role of HvVRN-H2 as an integrator of photoperiod and vernalization signals (Maurer et al., 2015).

In addition to the photoperiod and vernalization pathway genes, flowering time in barley is controlled by the EARLY MATURITY (EAM) loci referred to as EARLINESS PER SE (EPS) in wheat. The red/far-red light photoreceptor HvPHYC, an ortholog of Arabidopsis PHYC, was identified to underlie the EARLY MATURITY 5 (EAM5) QTL (Pankin et al., 2014). HvPHYC was found to promote light signal transmission to the circadian clock, modulating photoperiodic regulation of floral transition. It was also shown to interact with the PPD-H1 pathway and to increase HvFT1 expression not associated with the circadian clock (Nishida et al., 2013; Pankin et al., 2014). Similar mechanisms were discovered in wheat, where TaPHYC was found to activate TaPPD1 and TaVRN3 in inductive LDs (Chen et al., 2014).

The circadian clock gene HvLUX1, an ortholog of the Arabidopsis circadian gene LUX ARRHYTHMO, was detected as a candidate underlying the barley EAM10 locus (Campoli et al., 2013). As shown for barley plants carrying a mutation in EAM5 (HvPHYC), mutations in EAM10 (HvLUX1) altered the expression of HvPPD-H1 (as well as CCA1 in barley) and accelerated flowering under both LDs and SDs independently of the circadian clock. Furthermore, as shown for barley EAM8 mutants (Faure et al., 2012), early flowering of EAM10 mutants was linked with an up-regulation of HvFT1 transcription under SDs. In both cases this led to an induction of the LD photoperiod pathway under non-inductive (SD) conditions, suggesting that EAM10 acts as a repressor of HvPPD-H1.

In barley, HvCEN, a homolog of Antirrhinum majus CENTRORADIALIS, was identified at the EARLINESS PER SE 2 (EPS2) locus on chromosome 2H (Comadran et al., 2012). Antirrhinum cen mutants produce terminal flowers instead of indeterminate inflorescences common for wild-type plants without, affecting flowering time (Bradley et al., 1996). Analysis of the HvCEN alleles of a diverse set of spring and winter barley cultivars revealed that HvCEN facilitated the geographic range extension as well as the gradual separation between spring and winter cultivars (Comadran et al., 2012). A collection of flowering mutants was used to confirm sequence variations within HvCEN to be responsible for the observed variation in flowering time. While CEN homologs have been identified and characterized in rice (Nakagawa et al., 2002) and maize (Danilevskaya et al., 2008a), none have been identified to date in hexaploid wheat.

The circadian clock gene HvELF3, an ortholog of Arabidopsis ELF3, was identified to underlie the EARLY MATURITY 8 (EAM8) QTL (Faure et al., 2012; Zakhrabekova et al., 2012). Mutations in EAM8, also known as praematurum.a-8 (mat-a-8), were used since 1961 to facilitate short growing season adaptation and expansion of the geographic range in commercial barley varieties (Zakhrabekova et al., 2012). A loss-of-function mutation of HvELF3 led to an increased expression of HvFT1 compared with wild-type plants, causing a day-neutral early flowering phenotype (Faure et al., 2012). Boden et al. (2014) investigated barley plants carrying mutations in the HvELF3 gene and characterized them as early flowering under non-inductive SD photoperiods. The enhanced expression of HvGA20OX2 increased GA biosynthesis and expression of HvFT1. In spring barley, HvELF3 is necessary to maintain photoperiodic sensitivity through repression of HvFT1 and production of active GA via HvGA20OX2, a link not previously shown for AtELF3.

The phytohormone GA promotes flowering in barley and is essential for normal flowering of spring barley under inductive photoperiods (Boden et al., 2014). One of the gibberellin 20-oxidase genes, HvGA20OX2, was identified as the candidate gene underlying the allelic dwarfing gene sdw1/denso. sdw1/denso is an important locus selected for in barley breeding programs as the resulting semi-dwarf phenotype has reduced plant height and improved lodging resistance associated with higher yields and quality traits. The sdw1 mutant was distinguished from the denso mutants based on the more strongly reduced expression of HvGA20OX2 which was also linked to reduced plant height, enhanced grain yield, and lower grain quality (Jia et al., 2011).

Two independent pathways act in response to different photoperiods to control flowering time in rice. Day-light periods of less than 13.5 h accelerate flowering of rice (Itoh et al., 2010). Under inductive SD conditions, a HEADING DATE 1 (HD1)-dependent pathway is activated in which the AtCO homolog HD1 accelerates flowering via activation of the AtFT ortholog HEADING DATE 3A (HD3A) (Kojima et al., 2002). This pathway is conserved between Arabidopsis and rice (Table 3). For heading initiation under non-inductive LD conditions, an HD1-independent pathway is activated (Yano et al., 2000). The activity of genes unique to rice play a major role in conferring the different seasonal flowering responses when compared to LD plants such as Arabidopsis and are highlighted in the following sections.

GA signaling affects floral organ development and pathways are largely conserved between rice and Arabidopsis. The casein kinase I (CKI) gene EARLY FLOWERING1 (OsEL1) is associated with the negative regulation of GA-responsive signaling via post-translational modification of SLENDER RICE 1 (OsSLR1) protein. Mutations in OsEL1 lead to an enhanced response to GA signaling, and as a consequence to early flowering, lower spikelet fertility and thus lower grain yield under non-inductive LDs. However, a few extremely early flowering rice cultivars contain the mutation in OsEL1 but can maintain normal GA signaling through a currently unknown mechanism.

The rice gene SEMI-DWARF1 (OsSD1) contributed to the major increases in rice yield during the “Green Revolution” and is involved in GA signaling and biosynthesis (Peng et al., 1999). A similar reduced-height phenotype is the result of a loss-of-function mutation in OsGA20OX2, orthologous to the GA biosynthetic gene HvGA20OX2 in barley (Spielmeyer et al., 2002). HvGA20OX2 was proposed as the candidate gene underlying sdw1/denso locus and confers a reduced plant height phenotype in barley (Jia et al., 2011).

Maize was domesticated about 9,000 years ago from the wild grass teosinte (Zea mays ssp. parviglumis), a native to tropical Central America (Matsuoka et al., 2002). Teosinte requires SD photoperiods to flower, and domesticated maize has been adapted since then to cooler temperate regions of North America and Europe (Chardon et al., 2004). Maize post-domestication breeding was particularly driven by selection for genes and loci to adapt flowering time to new growth environments (Mascheretti et al., 2015). Photoperiod insensitive (day-neutral) maize varieties are cultivated in temperate climates, whereas photoperiod sensitive maize varieties are grown in tropical regions.

Flowering time in maize can range from only 35 up to 120 days, highlighting the high level of genetic diversity in maize phenology genes (Colasanti and Muszynski, 2009). Only a few mutations in flowering time genes have been identified so far, which contributes to the lack of current understanding of genetic and regulatory factors that determine maize phenology. Our current knowledge about maize flowering time is mainly based on results from QTL meta-analysis, (transposon) mutagenesis, and comparative genomics studies with the Arabidopsis and rice genomes (Buckler et al., 2009; Wei et al., 2015). Key phenology genes are listed in Table 4.

Maize contains a large family of Zea CENTRORADIALIS (ZCN) genes and related to both AtFT and AtTFL (Danilevskaya et al., 2008a). The most likely candidate for the FT ortholog in maize is Zea CENTRORADIALIS8 (ZmZCN8), a gene encoding a PEBP with high sequence similarity to the Arabidopsis flowering time gene AtFT1. In addition, a second maize florigen, the ZmZCN8 paralog ZmZCN7, was proposed (Mascheretti et al., 2015). Displaying florigen-like characteristics, ZmZCN8 is expressed only in the leaf phloem of mature leaves and interacts with the FD-like bZIP protein DELAYED FLOWERING1 (ZmDLF1) in the SAM (Figure 4). ZmZCN8 expression is elevated following a diurnal cycle under SDs in photoperiod-sensitive tropical maize, whereas ZmZCN8 levels remained unchanged in day-neutral temperate maize (Muszynski et al., 2006; Meng et al., 2011). Ectopic expression of ZmDLF1 resulted in an early flowering phenotype whereas silencing led to a severe delay flowering, as found for its functional equivalent OsID1 in rice (Matsubara et al., 2008). A putative AtCO ortholog in maize, ZmCONZ1, was located at a chromosome region syntenic with the rice ortholog HD1 and exhibits a diurnal expression pattern dependent on photoperiod (Miller et al., 2008). However, and there is currently no evidence that ZmCONZ1 can directly activate ZmZCN8 expression in SDs.

Other genes in the maize flowering time network include Zea FLORICAULA/LEAFY genes (ZmZFL1 and ZmZFL2) (Bomblies et al., 2003). ZmZFL1 and ZmZFL2 are orthologous to the meristem identity genes Antirrhinum majus FLORICAULA (AmFLO) and AtLFY and share conserved roles in controlling inflorescence architecture and patterning (Bomblies et al., 2003). Among other genes with likely conserved functions between maize and Arabidopsis are phytochromes ZmPHYA, ZmPHYB, and ZmPHYC (Sheehan et al., 2004). Unique for maize are Miniature Inverted-Repeat Transposable Element (MITE) sequence insertions detected in ZmPHYB2 that showed that the absence of the MITE correlated with enhanced ZmPHYB2 mRNA expression compared to ZmPHYB2 with MITE (Zhao et al., 2014). This evidence points toward a role of MITEs to inhibit ZmPHYB2 mRNA expression which ultimately leads to delayed flowering. Similarly, a MITE insertion into the VEGETATIVE TO GENERATIVE TRANSITION (VGT1) locus was associated with early flowering (Castelletti et al., 2014). The CCT domain-containing protein ZmCCT was found to be associated with photoperiod sensitivity and to regulate maize flowering time under LD conditions (Yang et al., 2013). The insertion of a CACTA-like transposon into its promoter represses ZmCCT gene transcription and thus reduced maize photoperiod sensitivity under LDs. Yang et al. (2013) provided evidence that the CACTA-like transposon was originally inserted in a tropical maize plant sensitive to SDs and was then accumulated selectively post-domestication as maize adapted to a range of LD environments.

Compared to flowering sensitive to vernalization as found in the temperate grasses wheat and barley or photoperiod-induced pathways as found in tropical rice, endogenous cues are more crucial for flowering than environmental signals in day-neutral temperate maize. The floral promoter INDETERMINATE1 (ZmID1) is a transcriptional regulator of the maize autonomous flowering pathway that functions in developing leaves to promote flowering (Wong and Colasanti, 2007). ZmID1 encodes a monocot-specific C2H2-type zinc finger protein without any known Arabidopsis ortholog and is hypothesized to act independently of the photoperiod pathway (Colasanti et al., 2006). Recent reports show that ZmID1 expression levels are high in developing leaves, and transcript and protein levels remain stable throughout the light cycle (Wong and Colasanti, 2007; Coneva et al., 2007). Mutations in ZmID1 lead to an extended vegetative phase (Colasanti et al., 2006). Genetic and expression data has shown that ZmID1 activates DELAYED FLOWERING1 (ZmDLF1) in the SAM, most likely indirectly (Muszynski et al., 2006; Meng et al., 2011), which is necessary for flowering through its interaction with ZmZCN8. Subsequent expression of floral identity genes, such as the floral transition MADS box gene ZmZMM4, then initiate reproductive development (Danilevskaya et al., 2008b).

Very recently, transcription and chromatin modifications of ZmZCN8 and its paralog ZmZCN7, a potential second maize florigen, were analyzed during floral transition in day-neutral maize and tropical teosinte (Mascheretti et al., 2015). This study led to the proposal of an alternative epigenetic mechanism of ZmID1-mediated regulation of ZmZCN8 expression in which ZmID1 establishes chromatin modifications in developing leaves of day-neutral maize to enable the expression of florigen genes ZmZCN8 and ZmZCN7 in the mature leaf later in development (Mascheretti et al., 2015). By contrast, a different set of ZmZCN8 chromatin modification patterns were detected in teosinte in response to inductive SDs, which highlights both conserved and unique features of epigenetically controlled flowering time mechanisms between autonomous and photoperiod-dependent pathways in maize.

GA accumulation and signaling has a direct positive affect on flowering time in maize (Thornsberry et al., 2001). The maize gene DWARF8 (ZmD8) encodes a DELLA protein orthologous to both Arabidopsis GIBBERELLIN INSENSITIVE (AtGAI) and the wheat REDUCED HEIGHT mutations TaRHT-B1 and TaRHT-D1, which were used to develop the high-yielding semidwarf varieties of the “Green Revolution” (Peng et al., 1999). Polymorphisms in ZmD8 were associated with differences in flowering time (Thornsberry et al., 2001). Very recently, ZEA MAYS GA REGULATORY FACTOR (ZmGRF) was identified as a new member of the bZIP protein transcription factor family in maize (Xu et al., 2015). ZmGRF transgenic Arabidopsis plants had enhanced GA levels indicating that ZmGRF has a function in plant morphology and development, and Arabidopsis has a currently unknown ortholog of ZmGRF.