Seasons and shape: inflorescences from autumn to summer

Abstract

Flowering plants organize their reproductive organs within specialized structures named inflorescences. Plasticity in the architecture of these inflorescences allows adaptation to the environment during flowering, ultimately determining reproductive output and yield. Inflorescence development relies on meristems, hubs of pluripotent cells that direct organogenesis. In recent years, laboratory studies have uncovered the response of meristems and their resulting inflorescences to environmental cues such as temperature, which is subject to both unpredictable and seasonal fluctuations. In this review, we explore the mechanisms through which temperature regulates inflorescence development in both model and crop species, principally from the Brassicaceae family. We follow the trajectory of the apical meristem through the seasons, from acquisition of reproductive identity to arrest and including branch outgrowth, highlighting the current understanding of the mechanisms through which temperature influences development. While the role of temperature in regulating the floral transition has been well established, we emphasize significant gaps in our understanding of how subsequent developmental transitions are controlled. Furthermore, we find that many key gene networks underpinning inflorescence development were initially characterized in Arabidopsis under controlled laboratory conditions. However, recent studies in other Brassicaceae species and crops have revealed that our understanding of gene function in the field context remains limited, posing a challenge for breeding efforts aimed at climate resilience.

We review temperature regulation of reproductive development in seasonal contexts, focusing on the molecular factors controlling change at meristems from vegetative growth, via production of the inflorescence and branches, to the end of flowering.

Introduction

Angiosperms are defined by their flowers, and these highly varied and complex structures are largely credited with the rapid expansion of angiosperm species. However, the plant architectures that support these flowers are also highly varied in their forms. Typically, flowers are arranged in specialized structures named inflorescences, whose shape and architecture vary greatly between and within species. Inflorescences can make up the majority of the plant body after flowering, especially in monocarpic species such as Arabidopsis thaliana, despite representing only the reproductive phase of a plant’s life. As such, their size, rate of progression through developmental transitions, and resultant architecture have important effects on fitness in the wild and yield on the farm.

Most plant development is environmentally responsive, but developmental phase transitions, such as seed germination or the floral transition, are under particularly tight environmental control. In temperate regions, strong seasonal change provides both drivers and cues for changes in development. This has made initiation of the floral transition, particularly in Arabidopsis, a paradigm for studying the genetic underpinning of temperature and daylength responses. However, for a long time it has been recognized that the creation of inflorescences involves a sequence of distinct developmental transitions, from cessation of vegetative leaf production to final floral meristem (FM) formation, whereby different structures are formed over time in response to specific environmental and endogenous cues (Suh et al., 2003; Prusinkiewicz et al., 2007). Thus, temporal and spatial variation in the rate at which progression through these transitions occurs is known to have critical effects on the shape of the inflorescence (Azpeitia et al., 2021). Most of these effects rely on developmental decisions taken at the inflorescence meristems (IMs), niches of undifferentiated cells that govern the production of flowers and, ultimately, the final architecture of the inflorescence (Benlloch et al., 2007).

In this review, we will discuss what is known about temperature response throughout the development of the inflorescence, picking up as the meristem acquires inflorescence identity and following it through flowering branch production to cessation of flower production, although we will not discuss flower development itself which is covered in detail elsewhere (Heisler et al., 2022). We will use an Arabidopsis framework to discuss the conserved transitions between meristem and organ identities, but draw on examples from different dicot species, primarily from the Brassicaceae family.

Reproductive structures and identities

Inflorescences are produced by the shoot apical meristem (SAM). Vegetative SAMs are indeterminate and iteratively produce phytomers, structural units of the plant containing leaves and axillary meristems (AMs), usually separated by internodes. The floral transition results in an IM that initiates determinate FMs, each one of which holds the potential to become a flower and, later, a seed-bearing fruit. The rate of progression from a vegetative meristem to a determinate FM therefore influences the number of branch nodes available to the plant. However, as discussed already, this is a multistage transition with different environmental and developmental requirements at each stage (King, 2015). Moreover, AMs can acquire inflorescence identity as well, producing secondary and higher order inflorescence branches. When and how these multiple reproductive meristems arrest their growth further controls the shape of the inflorescence. Additionally, for polycarpic perennials which flower multiple times during their lifetime, this shape is not final, and growth arrest is replaced with developmental reprogramming to resume vegetative growth until the next reproductive season. Thus, there are many processes that influence the architecture of inflorescences over the seasons.

In dicots from temperate regions, the raceme is a common inflorescence type (Benlloch et al., 2007; Prusinkiewicz et al., 2007). In the Brassicaceae family, this raceme is composed of a main axis stemming from the vegetative zone (V) which bears either flowers or lateral axes (i.e. axillary inflorescences), which in turn reiterate this morphology (Prusinkiewicz et al., 2007). Arabidopsis and other annual species from the Brassica genus form one of the simplest racemes, with two zones characterized by the development of different organs (Fig. 1A). The inflorescence zone produced after the initial reproductive transition (I1) contains leaf phytomers with AMs which can grow out into inflorescence branches. The I2 zone holds the primary inflorescence, where solitary flowers are initiated on the periphery of the meristem and leaf production is suppressed (Ratcliffe et al., 1998; Lazaro et al., 2018). Inflorescence branches from the I1 follow a similar developmental trajectory to the main stem, whereby they themselves can produce subordinate AMs (also I1) before starting flower production (I2). However, these higher order AMs are more likely to be quiescent (Azpeitia et al., 2021).

Fig. 1.

Variety of architectures in the flowering Brassicaceae family. (A, B) Typical inflorescence zonation in annuals, weeds (A), and crops (B). Most vegetative organs formed prior to flowering are in the V zone. In the inflorescence, the basal I1 zone holds axillary buds and branches, and the apical I2 zone sustains solitary fruits, flowers, and floral buds. (C) Typical inflorescence zonation in perennials. The vegetative region can be divided into a V1 with axillary inflorescences, a V2 with dormant axillary buds, and a V3 with subapical vegetative branches. As in annuals, I1 and I2 are found in the inflorescence.

The zonation of the inflorescence has important implications in agriculture. This is best illustrated by the diverse crop B. oleracea, whose subspecies can take on a wide array of different vegetable forms depending on which inflorescence architecture is selected for (Fig. 1B). For example, cabbages (B. oleracea ssp. capitata) are a head of packed vegetative leaves enclosing the meristem, which are ultimately derived from the V region (Alemán-Báez et al., 2022). On the other hand, cauliflowers (B. oleracea ssp. botrytis) are a dense cluster of arrested reproductive meristems whose floral primordia do not develop into flowers, suggesting that the distinctive curd is analogous to the I1 zone (Sadik, 1962). Finally, broccoli (B. oleracea ssp. italica), which is made up of packed florets, arises from the I2 zone (Gray, 1982).

In addition to the typical V/I1/I2 zonation of A. thaliana, longer-lived relatives such as the polycarpic Arabis alpina contain additional V zones bearing secondary structures with different identities, all of which are initiated prior to the appearance of the apical inflorescence (R. Wang et al., 2009; Lazaro et al., 2018; Vayssières et al., 2020) (Fig. 1C). The V1, typically composed of the earliest developed meristems and often formed before winter, gives rise to axillary flowering branches. In contrast, the V3 zone, usually initiated during the cold period, holds subapical vegetative branches. Plants exposed to a sufficiently long period of cold form a V2 zone in between, containing a series of buds that are typically kept dormant during flowering (Vayssières et al., 2020). The difference in developmental trajectories among these different structures stems from the asynchronous development of their source AMs, with those from I1, I2, and V1 being the only ones having experienced enough cold, and being of the right age, to acquire inflorescence identity through the process of vernalization (Lazaro et al., 2018; Hyun et al., 2019; Vayssières et al., 2020). In turn, AMs of zones V2 and V3 are critical to perenniality, as they do not senesce after reproduction and allow for new vegetative growth the following season (Vayssières et al., 2020).

The making of inflorescence meristems

The molecular bases of inflorescence formation have been studied in detail in model plants, leading to the identification of a complex regulatory network which converges on to a subset of conserved regulators and markers of floral transitions (Fig. 2A). In Arabidopsis these are principally two modules of homologous MADS-box transcription factors, plus the unique transcription factor LEAFY (LFY). Expression of the ‘inflorescence integrators’ SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1, AGAMOUS-LIKE24, and XAANTAL2 (SOC1, AGL24, and XAL2; the ‘SAX’ MADS-box module) leads to up-regulation of the second module, formed of FM markers APETALA1 and its homologue CAULIFLOWER (AP1/CAL), which negatively regulates SAX expression in a feedback loop (Liu et al., 2007; Immink et al., 2012; Pajoro et al., 2014; Azpeitia et al., 2021) (Fig. 2B, C). Expression of these regulators at the meristem or in the subtending primordia are hallmarks of inflorescence identity. Thus, their interactions define the spatial and temporal developmental progress of the IM and its floral primordia through to determinacy and cessation.

Fig. 2.

Several temperature inputs converge at the floral integrators. (A) Diagram depicting morphological changes during the conversion of a vegetative SAM (VM) to an inflorescence meristem (IM), and the subsequent initiation of floral meristems (FMs). (B, C) Network summarizing the major molecular interactions discussed throughout the text which occur under cold (B) and warm (C) temperatures during this transition. Arrows and blunt arrows indicate activation and repression, respectively. Thermometers highlight temperature inputs into the system. Genes in grey indicate components that are inactive at the respective temperatures, and colours indicate factors relevant to a specific meristem identity.

Many elements of the floral transition network upstream of these factors are temperature sensitive, but two main pathways primarily relay temperature information: the vernalization pathway and the ambient temperature or thermosensory pathway (Srikanth and Schmid, 2011; Song et al., 2015). The vernalization pathway acts to prevent plants that germinate in the warmer months from opening flowers until the following spring, being most active in Arabidopsis accessions with a winter annual (sometimes ‘biennial’) life cycle. This is believed to be the ancestral life history, with most rapid cycling accessions (including the commonly used Col-0 and Ler-1) having low, though functional, expression levels of the main vernalization regulator FLOWERING LOCUS C (FLC) (Johanson et al., 2000; Gazzani et al., 2003; Whittaker and Dean, 2017). As a result, much knowledge of the floral transition network has been established in these low-FLC backgrounds, in warm and long-lighting conditions and so in a ‘spring’-like genetic and environmental context.

Spring and summer pathways

Two key integrators of environmental signals throughout the floral transition are the ‘florigen’ phosphatidylethanolamine-binding proteins (PEBPs) FLOWERING LOCUS T (FT) and TWIN SISTER OF FT (TSF). For FT in particular, the key environmental cue for its expression is photoperiod, the mechanisms of which have been extensively and well reviewed elsewhere (Song et al., 2015). Under promotive conditions, FT and TSF are expressed in the phloem companion cells in leaves (Kobayashi et al., 1999; Wigge et al., 2005; Yamaguchi et al., 2005; Chen et al., 2018). The resulting proteins are loaded into the phloem and moved towards the apex (Corbesier et al., 2007; Jaeger and Wigge, 2007; Mathieu et al., 2007). At the apex, they form a regulatory complex with the bZIP transcription factor FD which transcriptionally activates genes that specify the identity of the developing IM as well as of the floral primordia that it will form in its periphery (Abe et al., 2005; Wigge et al., 2005; Yoo et al., 2005). Within this complex molecular architecture, the bulk of the temperature signal integration into the flowering programme involves regulation of FT at various levels.

The thermosensory pathway dominates the transcriptional control of FT by temperature as characterized within a range of ∼10–27 °C, which has been reviewed in detail previously (Susila et al., 2018). The MADS-box transcription factor SHORT VEGETATIVE PHASE (SVP) has a key role in this process (Blázquez et al., 2003; Lee et al., 2013). In colder temperatures, SVP represses FT transcription both directly (Lee et al., 2007; Li et al., 2008) and indirectly, for example via activation of the TEMPRANILLO (TEM) genes (Castillejo and Pelaz, 2008; Marín-González et al., 2015). When acting directly, SVP interacts with itself and other MADS-box transcription factors including the MADS AFFECTING FLOWERING (MAF1–5) family to form homo- and heterotetramers that bind to the promoter of FT and TSF to repress their expression (Lee et al., 2007, 2013; Gu et al., 2013; Posé et al., 2013). Expression of SVP is only moderately affected by temperature (Lee et al., 2007; Posé et al., 2013), but MAF1, also commonly known as FLOWERING LOCUS M (FLM), is strongly up-regulated in the cold (Lee et al., 2013; Sureshkumar et al., 2016). Additionally, cold promotes the production of a specific spliceform, FLM-β, which seems to favour the binding of SVP to its targets (Balasubramanian et al., 2006; Lee et al., 2013; Posé et al., 2013; Sureshkumar et al., 2016). Indeed, complexation with FLM-β maintains SVP in the nucleus, while down-regulation of FLM-β in the warmth triggers its translocation to the cytosol, where it can be marked for proteasomal degradation (Jin et al., 2022). When present, a further MADS-box transcription factor controls the repression of FT; namely FLC. Like its MAF relatives, FLC dimerizes with SVP (Li et al., 2008) and binds to CArG box motifs in the genomic regions of FT and TSF to inhibit their transcription (Yamaguchi et al., 2005; Searle et al., 2006; Lee et al., 2007). Besides its transcriptional control, FT is post-translationally repressed in the cold through sequestration in intracellular membranes, where it interacts with phospholipids (Liu et al., 2020; Susila et al., 2021) or proteins (Susila et al., 2024). This compromises the long-range mobility of FT, further contributing to the delay of the floral transition at lower temperatures.

In addition to the repressors that govern the control of FT in the cold, several activators promote its transcription upon warming. Such is the case of the basic helix–loop–helix (bHLH) transcription factors PHYTOCHROME-INTERACTING FACTOR 4 (PIF4) and PIF5 (Kumar et al., 2012; Thines et al., 2014). Different mechanisms have been proposed to explain PIF4 induction in the warmth, including direct activation by TEOSINTE BRANCHED1/CYCLOIDEA/PCF TRANSCRIPTION FACTOR 5 (TCP5) (Han et al., 2019) and BRASSINAZOLE-RESISTANT 1 (BZR1), whose nuclear localization is temperature dependent (Ibañez et al., 2018). The evening complex of the circadian clock, composed of LUX ARRHYTHMO (LUX), EARLY FLOWERING 3 (ELF3), and ELF4, also controls PIF4 expression directly (Nusinow et al., 2011; Ezer et al., 2017), with a specific prion domain of ELF3 acting as a temperature sensor (Jung et al., 2020). Lastly, histone deacetylation has a role in the regulation of PIF4 and its targets (Tasset et al., 2018). The histone variant H2A.Z exhibits increased chromatin eviction at high ambient temperatures in Arabidopsis, increasing DNA accessibility and therefore expression of enclosed genes (Kumar and Wigge, 2010; Shen et al., 2019; van der Woude et al., 2019). This is mediated by temperature-dependent deacetylation of H2A.Z by POWERDRESS (PWR) and HISTONE DEACETYLASE9 (HDA9), ultimately leading to increased transcriptional competence of the PIF4 locus at warm ambient temperatures. Consequently, cold repression and promotion in warm long photoperiods ensure that, in the field, FT is a signature of flowering in spring and summer (Fig. 3).

Fig. 3.

Seasonal fluctuations in the transcription of key floral integrators and their upstream regulators. (A) Cartoon depicting development of a generic perennial from the Brassicaceae family exposed to seasonal temperature fluctuations. Bolting is assumed to occur in spring, and flowering to cease in summer. (B) Line graphs illustrating seasonal changes in expression of key genes involved in inflorescence development. Data taken from Satake et al. (2013) and Nagano et al. (2019).

A flower for all seasons: other temperature-sensitive pathways

Although the major temperature inputs to IM formation are usually conceptualized as coming through the vernalization and thermosensory pathways, other flowering pathways are temperature sensitive. One such example is the age pathway, which ensures that plants only flower after reaching maturity. Briefly, this involves the opposing activity of two miRNAs, miR156 and miR172. In juvenile plants, the highly abundant miR156 translationally inhibits SQUAMOSA PROMOTER BINDING-LIKE (SPL) transcription factors, which themselves can promote expression of miR172 (Wu and Poethig, 2006; Gandikota et al., 2007; J.-W. Wang et al., 2009; Wu et al., 2009) (Fig. 2B, C). As the plant ages into maturity, miR156 levels decline via histone-mediated repression linked to cell division, itself inherently sensitive to temperature-responsive changes in growth rates (Cheng et al., 2021). The decline in miR156 leads to derepression of SPL genes and subsequent induction of their targets including miR172, which in turn promote flowering by inhibiting the activity of reproductively repressive APETALA2 (AP2)-like genes such as TARGET OF EAT1 (TOE1) and TOE2 (Aukerman and Sakai, 2003; Chen, 2004; Wu and Poethig, 2006; Wu et al., 2009). SPL15 promotes flowering in the absence of both FT and TSF (Hyun et al., 2019), but SPL3 also directly up-regulates FT (Kim et al., 2012), such that the age pathway integrates within the floral network at multiple levels. Intriguingly, miR156 and miR172 exhibit opposing thermosensory responses: miR156 is up-regulated at cool ambient temperatures, probably through regulation at the RNA processing level (Kim et al., 2016), whilst miR172 is up-regulated at warm ambient temperatures (H. Lee et al., 2010), at least partly through the activity of the RNA-binding protein FLOWERING CONTROL LOCUS A (FCA) which promotes processing of primary-miR172 to mature miR172 (Jung et al., 2012). In Arabidopsis, variability in flowering responses to warm temperatures is dependent on regulation of the miR172 level, which is sensitive to mutations in SVP (H. Lee et al., 2010). In fact, both SVP and FLM bind to the promoter of miR172a (Cho et al., 2012; Tao et al., 2012; Posé et al., 2013), generating crosstalk between the thermosensory and age pathways in the regulation of reproductive development.

Another essential facet to IM development is the potent promotive activity of gibberellin (GA). GAs are a family of growth-inductive plant hormones with wide-ranging functions in plants which have long been established in promoting flowering (Rieu et al., 2008; Mutasa-Göttgens and Hedden, 2009). In brief, GAs regulate development by modulating the activity of the DELLA family of growth repressors. In Arabidopsis, there are five DELLAs, GA INSENSITIVE (GAI), REPRESSOR OF GAI (RGA), RGA-LIKE1 (RGL1), RGL2, and RGL3, all of which negatively regulate the activity of transcription factors including PIF4 and the SPLs (Peng et al., 1997; Lee et al., 2002; K.P. Lee et al., 2010; Tyler et al., 2004; de Lucas et al., 2008; Yu et al., 2012). In this pathway, bioactive GAs bind to the receptor GIBBERELLIN INSENSITIVE DWARF1 (GID1), leading to a conformational change which permits interaction with DELLA proteins, culminating in proteasomal degradation of the complex (Silverstone et al., 2001; Ueguchi-Tanaka et al., 2005; Griffiths et al., 2006; Willige et al., 2007; Murase et al., 2008; Shimada et al., 2008). Through degradation of DELLAs, which otherwise have inhibitory effects on development, GAs act as global regulators of many processes, for example floral transition and stem elongation (King et al., 2001; Cheng et al., 2004; Tyler et al., 2004; Rieu et al., 2008; Yu et al., 2012), and thus regulation of GA synthesis in response to environmental cues is critical for appropriate inflorescence development. Sensitivity to temperature is in part achieved through FLC and SVP, which can act independently and coordinately to repress GA biosynthesis genes (Andrés et al., 2014; Mateos et al., 2015). These include the GIBBERELLIN 3-OXIDASE (GA3OX) and GIBBERELLIN 20-OXIDASE (GA20OX) family, which function in the final stages of GA biosynthesis (Hedden, 2020). Additionally, TEM1 also exerts photoperiodic control over GA3OX1 and GA3OX2 by direct interaction with cis-regulatory elements (Osnato et al., 2012). That being said, the expression dynamics of GA metabolism genes in winter annuals under field conditions remain poorly characterized. In contrast, a study on the overwintering perennial Arabidopsis halleri revealed clear seasonal patterns of gene expression; for instance, GA2OX8 and GA3OX1 are strongly up-regulated during winter (Fig. 3B), coinciding with the repression of FLC (Komoto et al., 2024). In an annual Arabidopsis, it has been shown that GA only has minor contributions to flowering time under long days, but under short days it becomes the primary, essential, component (Wilson et al., 1992; Hisamatsu and King, 2008; Rieu et al., 2008). Taken together, these data suggest an important function for GAs in the early development of the inflorescence as floral initiation in overwintering plants often occurs during or before the winter (see ‘Floral buds are winter structures’). Furthermore, GAs positively regulate the floral integrators SOC1 (Moon et al., 2003; Porri et al., 2012) and LFY (Blázquez et al., 1998; Eriksson et al., 2006), which may provide a mechanism for the initiation of flowering in short days in the absence of FT and TSF.

Taking the slow road to flowering

For many monocarpic plants, life begins not in spring, but in summer or autumn, a life history that may allow plants to make full use of the year by growing through late autumn and winter. Gene variants promoting this life history show signatures of selection linking it to higher latitudes and altitudes, where growing seasons may be shorter (Lempe et al., 2005; Ågren et al., 2017). High FLC expression delays flowering in a quantitative manner in autumn, ensuring that winter annuals do not flower until they have experienced a prolonged period of cold (Lempe et al., 2005; Shindo et al., 2005; Hepworth et al., 2020). In Arabidopsis, this pathway is mainly active in accessions with functional alleles of FRIGIDA (FRI), a transcriptional activator that up-regulates expression of FLC early in development (Michaels and Amasino, 1999; Johanson et al., 2000). During autumn and winter, cold represses FLC expression through a network of temperature-sensitive upstream interactions (Antoniou-Kourounioti et al., 2018). These include various mechanisms such a sequestration of FRI into nuclear condensates, particularly during cold nights (Zhu et al., 2021), freezing-triggered up-regulation of the antisense long non-coding RNA COOLAIR (Zhao et al., 2021), and particularly the nucleation of repressive chromatin-remodelling complexes at the FLC locus triggered by the cold-induced PHD protein VERNALIZATION INSENSITIVE 3 (VIN3) (Sung and Amasino, 2004; De Lucia et al., 2008; Bond et al., 2009). The chromatin environment triggered during winter at the FLC locus maintains repression of FLC even when plants are returned to the warm, releasing repression of FT, SOC1, SPL15, and other floral integrators until the process is reset in the following generation. The regulation of FLC is highly complex, and is reviewed in more detail elsewhere (Whittaker and Dean, 2017), but it is an unusual case in that the range and response speed of the principal temperature inputs have been mapped in field-relevant temperatures (Antoniou-Kourounioti et al., 2018; Zhu et al., 2022). At least three temperature pathways control VIN3 expression alone, including clock-mediated pathways (Antoniou-Kourounioti et al., 2018; Hepworth et al., 2018; Kyung et al., 2022), rapid heat-sensitive repression pathways that prevent accumulation of VIN3 during warm autumn days (Sung and Amasino, 2004; Hepworth et al., 2018), and the very slow promotion of VIN3 via increased concentration of NTL8 protein due to reduced cell division (Zhao et al., 2020). Thus, slower growth in the cold is directly linked to reproductive development in a manner analogous to cell division-dependent miR156 repression in the warmth.

Floral buds are winter structures

Although vernalization is often referred to as being a ‘winter’ pathway (usually studied under constant 5 °C in the lab), in both dicots and monocots vernalization pathways function in the mid to high teens (Wollenberg and Amasino, 2012; Duncan et al., 2015; Dixon et al., 2019). In Arabidopsis, cool nights below 14 °C are sufficient for transient transcriptional repression of FLC, with epigenetic repression pathways triggered quantitatively when daily highs fall below ∼15 °C (Antoniou-Kourounioti et al., 2018; Hepworth et al., 2018). As shown in Fig. 4, this autumnal vernalization permits the floral transition to occur before winter in field-grown Arabidopsis and related winter annual crops such as B. napus, despite the absence of FT, due to the release from FLC repression of floral integrators such as SOC1 and SPL5 (O’Neill et al., 2019; Matar et al., 2021; Lu et al., 2022). The perennial A. alpina will also initiate a set of IMs in the cold during autumn following extended vernalization thanks to silencing of the FLC orthologue PERPETUAL FLOWERING 1 (PEP1) and up-regulation of orthologues of SOC1, FRUITFUL (FUL), and SPL15 (R. Wang et al., 2009; Lazaro et al., 2018; Hyun et al., 2019). These IMs produce floral buds that remain dormant until spring, while insufficiently vernalized AMs produced later in the season remain vegetative until the flowering season in the following year, a state reinforced by the reactivation of PEP1 in spring (R. Wang et al., 2009). In the apex, floral transition after vernalization is marked by up-regulation of LFY, FUL, and AP1 (Lazaro et al., 2018). Preceding this, SOC1 is up-regulated in both inflorescence and vegetative meristems (Lazaro et al., 2018; Matar et al., 2021). However, in perennial and winter annual Brassicaceae, bolting (stem extension) and anthesis do not occur until spring with the arrival of longer days and warmer temperatures. Therefore, there is a developmental window over the winter where the IM has initiated floral primordia but is maintained in a dormancy-like state, whereby elongation of reproductive tissues is inhibited (Fig. 4). This observation suggests that the winter annual life history in Brassicaceae more closely resembles bud dormancy in woody perennials (Penfield, 2024), where autumn cooling induces bud set and dormancy before further winter chilling permits a state of growth receptivity to spring conditions within buds. These mechanistic parallels in phenology are significant as they suggest that floral buds and the IM remain responsive to temperature signals over the winter.

Fig. 4.

Changes in meristem activity and identity across the seasons. Circular diagram illustrating the multiple developmental transitions that occur in aerial meristems of Arabidopsis during the year. The temperature profile and timing of events are based on a field trial conducted in Norwich, UK (Lu et al., 2022). The apical vegetative meristem (VM) acquires reproductive identity during late autumn and early winter, after which it initiates floral primordia and later becomes dormant. In spring, flowering-promotive conditions trigger bolting, and the VM from the primary inflorescence (PI), now fully transitioned into an inflorescence meristem (IM), resumes reproductive development in the I2 zone of the inflorescence. Although delayed in their floral transition, axillary meristems (AMs) follow a similar developmental fate in the I1 zone. Finally, IMs undergo proliferative arrest in summer.

Temperature sensitivity of initiated floral buds is clear in winter oilseed rape, whereby transient warming in short days results in a subsequent delay in flowering the following spring (Lu et al., 2022), as opposed to an acceleration of further reproductive development through the thermosensory pathway as previous models would have suggested (Blázquez et al., 2003; Kumar et al., 2012; Posé et al., 2013). This delay was attributed to the up-regulation of late-vernalizing copies of Bna.FLC and the activation of a BRANCHED1 (BRC1)-dependent and abscisic acid (ABA)-related dormancy programme within apical buds in winter accessions (Lu et al., 2022). In addition to delayed flowering, warming of quiescent floral primordia in the winter also has significant effects on floral organogenesis, with observations of precocious opening of apetalous floral buds and a reduction in seed number per pod in warmed plants (Lu et al., 2022). This is consistent with studies which correlated warm winter temperatures with low on-farm yields (He et al., 2017; Brown et al., 2019), highlighting the importance of temperature signal integration in floral primordia for modulating yield components of crop species, although further work is necessary to better understand the mechanisms downstream of warming-induced dormancy activation which leads to yield decline.

Unlike winter annuals, mechanisms of temperature signal integration in floral buds of woody perennials have been well characterized. In aspen (Populus sp.), shortening photoperiods and mild cooling in autumn promote ABA accumulation in buds and an up-regulation of SVP orthologues (Singh et al., 2018, 2019). SVP subsequently promotes dormancy through up-regulation of CALLOSE SYNTHASE 1 (CALS1), an established regulator of plasmodesmatal closure and symplastic isolation characteristic of dormant buds (Rinne et al., 2011; Tylewicz et al., 2018), and suppression of GA synthesis and signalling (Singh et al., 2018, 2019). As SVP also represses GA biosynthesis in Arabidopsis (Andrés et al., 2014), this suggests that mechanisms underpinning regulation of floral bud development in response to autumn and winter temperatures are conserved beyond the Brassicaceae.

In addition to SVP, further regulators of perennial bud dormancy are likely to have conserved roles in modulating development of floral primordia in winter annuals. DORMANCY ASSOCIATED MADS-BOX (DAM) transcription factors are master regulators of bud dormancy which are closely related to Arabidopsis SVP and AGL24 (Falavigna et al., 2019; Quesada-Traver et al., 2022). Their role in regulating development of floral buds was first studied by Bielenberg et al. (2008), who mapped the locus responsible for constant meristematic growth in the evergrowing peach mutant. DAM gene expression is highly seasonal, responding to both temperature and photoperiod signals to control chilling responses and dormancy release. Their expression peaks in autumn and winter, inducing a dormancy state within buds, before chilling-induced epigenetic silencing which permits budbreak in spring (Leida et al., 2012; H. Zhu et al., 2020). DAM gene expression has been shown to control flowering time in species such as peach, apple, and loquat (Leida et al., 2012; Wu et al., 2017; Quesada-Traver et al., 2022), suggesting that their homologues in winter annuals might be important additional factors for controlling timing of dormancy release and growth receptivity to spring conditions. In fact, Arabidopsis SVP and AGL24 directly interact with floral integrators such as SOC1, and also repress expression of floral homeotic genes to maintain FM identity in the early stages of floral primordia development (Liu et al., 2008; Gregis et al., 2009). In the field, several B. napus SVP and AGL24 genes follow expression profiles comparable with that of DAM genes (O’Neill et al., 2019), which further suggests a role for these transcription factors and their relatives in the integration of seasonal temperature cues to control development of floral buds across plant families.

Preventing vegetative reversion

Timing the floral transition to match suitable conditions is essential to ensure reproductive success, but merely initiating an IM is not enough. Rather, the appropriate genetic machinery must be activated to ensure that the meristem remains committed to a reproductive fate (Corbesier et al., 2007; R. Wang et al., 2009; Lazaro et al., 2018). Reproductive commitment relies on a pair of MADS-box transcription factors which are up-regulated in the inflorescence during floral transition, SOC1 and FUL (Melzer et al., 2008; Torti et al., 2012; Lazaro et al., 2018). Both are downstream targets of FT and therefore are under temperature control (Samach et al., 2000; Schmid et al., 2003; Teper-Bamnolker and Samach, 2005; Searle et al., 2006; Fig. 2). Besides their direct control by FT, both share common regulators from the thermosensory pathway such as FLM (Lee et al., 2013), SVP (Lee et al., 2007; Fujiwara et al., 2008; Li et al., 2008; Jang et al., 2009; Immink et al., 2012), and FLC (Hepworth et al., 2002; Helliwell et al., 2006; Searle et al., 2006; Li et al., 2008), which provides a local input for temperature regulation at the inflorescence level. However, unlike FUL, SOC1 expression is gradually down-regulated in the absence of a floral stimulus (Torti et al., 2012). SOC1 and FUL form homo- and heterodimers that bind to and activate downstream genes (de Folter et al., 2005; Immink et al., 2012; Balanzà et al., 2014). Members of the SPL family have been proposed as possible targets in the context of IM commitment, and both SOC1 and FUL are required for the activation of SPL4 in the IM (Torti et al., 2012). In agreement with this, SPL15 is essential for commitment in A. alpina (Hyun et al., 2019). SPL transcription factors can, in turn, promote SOC1 and FUL expression (Shikata et al., 2009; Yamaguchi et al., 2009; J.-W. Wang et al., 2009), involving them in a feedback loop. Indeed, the expression of SOC1 and FUL is controlled through an age-dependent pathway involving SPLs and miR156 (J.-W. Wang et al., 2009; Kim et al., 2012). In addition, SOC1 and FUL dimers also bind to the regulatory regions of SOC1, TEM1, TEM2, SVP, miR156, and AP2-like genes (Immink et al., 2012). However, one key area requiring better understanding is the environmental regulation of IM commitment, particularly given that known regulators are shared between SOC1 and FUL, despite evidence of their differing sensitivity to the environment (Torti et al., 2012).

Maintaining the inflorescence meristem

In contrast to floral primordia at the apex periphery, the inflorescence is kept undifferentiated; but how? At the genetic level, the identity of both vegetative and inflorescence meristems is marked by the homeobox transcription factor WUSCHEL (WUS), which promotes cell proliferation in the AM both before and during flowering by up-regulation of CLAVATA3 (CLV3) (Mayer et al., 1998). The activity of WUS ensures that a pool of pluripotent stem cells is maintained in the apex, and thus WUS expression is essential for meristem maintenance (Laux et al., 1996; Mayer et al., 1998). WUS itself is negatively regulated by CLV3, leading to a negative feedback loop which confines WUS expression within the organizing centre of the SAM (Brand et al., 2000; Schoof et al., 2000). Additional regulators of WUS include members of the FANTASTIC FOUR (FAF) family which are known to reduce meristem size through WUS inhibition (Wahl et al., 2010). Using modelling approaches, it has been demonstrated that changes in meristem size may affect inflorescence development by tuning the rate of organ production (Azpeitia et al., 2021), suggesting that mechanisms underpinning meristem maintenance may be critical for overall plant productivity.

Though WUS is pivotal for meristem maintenance, it is not exclusive to IMs, whose identity is specified by additional factors. A key example is TERMINAL FLOWER1 (TFL1), another member of the PEBP family along with FT and TSF. Similarly to FT, TFL1 forms a transcriptional regulator complex with FD; however, unlike the FT–FD complex, TFL1–FD acts predominantly by repressing the expression of downstream targets (Ahn et al., 2006; Hanano and Goto, 2011; Goretti et al., 2020). Like WUS, TFL1 is expressed in the IM where it inhibits the expression of FM identity genes such as AP1 and LFY, therefore maintaining the identity of the indeterminate inflorescence (Gustafson-Brown et al., 1994; Mandel and Yanofsky, 1995; Y. Zhu et al., 2020). Interestingly, AP1 and LFY themselves also regulate TFL1 expression (Goslin et al., 2017; Serrano-Mislata et al., 2017). Consequently, this crosstalk leads to spatially distinct zonation of cells belonging to either determinate FMs or indeterminate IMs. During the initial conversion of the SAM to an IM, a phenomenon of ‘doming’ is observed. Here, the IM increases in size, particularly in height, generating a dome-shaped morphology (Kinoshita et al., 2020; Cerise et al., 2023). This process is delayed by the activity of TFL1, which is achieved through repression of genes such as FUL, SEPALLATA 4 (SEP4), SPL3, and SPL8 (Goretti et al., 2020; Y. Zhu et al., 2020; Cerise et al., 2023).

It is also worth noting that in addition to maintaining inflorescence identity, TFL1 also functions as a negative regulator of the initial vegetative to floral transition (Hanano and Goto, 2011), a phenomenon genetically separable from its inflorescence maintenance role (Serrano-Mislata et al., 2017). In other species, TFL1 orthologues also repress flowering timing, and in a temperature-responsive manner (Périlleux et al., 2019). In the case of A. alpina, AaTLF1 expression contributes to the age dependency of meristem competence for floral transition, and its expression is down-regulated during vernalization in older plants (Wang et al., 2011).

Owing to the multi-faceted roles of TFL1, its activity within the IM is highly dynamic. This is thanks to complex temporal and spatial regulation of TFL1 expression and the ability of the transcribed protein to move between cells (Conti and Bradley, 2007; Goretti et al., 2020; Cerise et al., 2023). Early in development, TFL1 is weakly expressed below and within the lower portion of the vegetative meristem, with mRNA levels increasing as development progresses towards floral transition (Bradley et al., 1997; Conti and Bradley, 2007). The spatial overlap of TFL1 protein with its interacting partner FD is also stage dependent, with co-localization occurring below the SAM before floral transition and at the tips of the IM after transition, reflective of the dual role of TFL1 in the repression of floral transition in vegetative meristems and later maintenance of indeterminacy in the IM (Cerise et al., 2023). However, during the transitionary doming stage, TFL1 becomes more diffuse, a hypothesized requirement to transiently reduce TFL1 antagonism of FT to promote floral transition (Cerise et al., 2023). It is therefore clear that the activity of TFL1 is critical for inflorescence development and has a direct role in controlling overall inflorescence architecture by ensuring indeterminacy. This has been demonstrated beyond Arabidopsis in crops such as B. napus where tfl1 knockouts significantly reduce plant stature and fruit production (Sriboon et al., 2020). Additionally, targeting of the wild currant tomato (Solanum pimpinellifolium) TFL1 orthologue SELF PRUNING (SP) was used as a strategy to manipulate plant architecture for the ‘de novo domestication’ of an ancestral progenitor species to tomato (S. lycopersicum) (Zsögön et al., 2018), highlighting agronomic applications for the manipulation of the genetic framework underpinning inflorescence establishment and maintenance (Eshed and Lippman, 2019).

Similar to TFL1, GA signalling also has a dual and separable role in floral transitions (Yamaguchi et al., 2014). GAs promote the vegetative to inflorescence transition, including by inducing LFY expression (Wilson et al., 1992; Blázquez et al., 1998; Eriksson et al., 2006), but LFY itself up-regulates production of a GA-catabolizing enzyme, and the resultant local reduction in GA is required to allow SPL9 to up-regulate AP1 expression in floral primordia (Yamaguchi et al., 2014). As a result, GA mutants show an increased V zone but reduced I1 zone, due to delayed initial floral progression but accelerated flower formation (Yamaguchi et al., 2014).

However, our understanding of temperature signal integration in the control of inflorescence indeterminacy is extremely limited. It is unclear how the inflorescence-specific roles of factors such as TFL1 and GA are affected by temperature and whether this is separable from their role in initial floral transition; but this could be an interesting hypothesis to test considering the significance of temperature signals for driving development at the IM.

Axillary meristem identity, dormancy, and outgrowth

At one time, the floral transition in Arabidopsis was proposed to be a single event which travelled through the plant, with the different identities of rosette and cauline branches being due to apical and axillary meristems ‘catching the floral wave’ at different times. In commonly used rapid-cycling accessions of Arabidopsis (e.g. Col-0 and Ler-1) grown under standard long-day growth conditions, this may appear to be the case (Hempel and Feldman, 1994). Instead, studies working with late-flowering plants, such as rapid cyclers grown in non-inductive short days, or with winter annual accessions, challenged this paradigm in support of a two-step phase transition model (Suh et al., 2003). According to the latter, the trajectory of the reproductive meristem in slower flowering plants follows two stages. First, it initiates a series of lateral inflorescences which eventually develop to form the I1 region of the shoot system (Fig. 1A) and, later, it undergoes a second transition, becoming an IM, with primordia initiated thereafter acquiring FM identity, giving rise to the I2 region (Fig. 1A) (Schultz and Haughn, 1991).

In Arabidopsis, inflorescence identity acquisition is usually followed by the bolting process: the massive increase in internode length that produces the elongated inflorescence (for a review of outgrowth processes and their regulation, see McKim, 2020). However, not all AMs achieve inflorescence identity, and not all AMs bolt, with many instead remaining very slow-growing ‘axillary buds’, a state usually termed ‘dormant’, even though they may be producing many lateral organs. Evidently, AMs have different sensitivities to flowering-promotive signals (Niwa et al., 2013; Lazaro et al., 2018; Zhu and Wagner, 2020; Fig. 4). The I1 (‘cauline’) zone is a clear case; primordia initiated during the I1 phase usually grow out rapidly at bolting, producing further inflorescences (Lazaro et al., 2018). In contrast I2 zone AMs are pre-committed as FMs. Thus the rate of progression from the vegetative meristem to I1 then to I2 will set the pattern of axillary identities and subsequent architecture. Indeed, it has been proposed that the difference in inflorescence architectures between Arabidopsis and other species from the Brassicaceae (e.g. cauliflower) is likely to be due to differences in the balance of gene activity in floral networks (Prusinkiewicz et al., 2007; Azpeitia et al., 2021).

Interacting with this floral identity pattern are other processes that control relative activity and maintenance of AMs, particularly apical dominance and resource availability. These are largely processes in which local growth is coordinated with the whole-plant body plan by movement and local synthesis of photosynthates and hormones—particularly auxin, cytokinins (CKs), and strigolactones (SLs). For detailed reviews on the mode of action of these signals, see Zhu and Wagner (2020) and Barbier et al. (2023). In brief, auxin produced in existing inflorescences acts to competitively inhibit the outgrowth of other AMs (Ongaro et al., 2008; Prusinkiewicz et al., 2009; Bennett et al., 2016). Auxin effects at the AM itself are mediated indirectly as auxin itself does not enter the bud. Auxin enhances SL synthesis, which in turn enhances the competitive effect of auxin, concentrating growth in actively elongating branches (Crawford et al., 2010; Bennett et al., 2016). SL action in the bud is partly mediated by up-regulating expression of the repressive TCP transcription factor BRC1 (Aguilar-Martínez et al., 2007; Wang et al., 2015). Auxin also down-regulates CK production; CKs act locally in the bud to promote outgrowth at the cell cycle level (Tanaka et al., 2006). After dormancy release, commitment to sustained bud outgrowth is further supported by auxin export from the bud itself (Chabikwa et al., 2019), but also by GA signalling in a manner analogous to GA control of outgrowth of the main shoot, as reviewed elsewhere (McKim, 2020).

This hormonal network coordinates the effect of nutrient levels perceived in the roots (via modulation of SL and CK synthesis) and loss of resource sinks (auxin and apical dominance) with the local AM transcriptional programme. When dormant, AMs are characterized by a starvation-like transcriptional programme downstream of BRC1, which activates a triad of homeodomain leucine zipper (HD-Zip) transcription factors, HOMEOBOX PROTEIN21 (HB21), HB40, and HB53, to up-regulate local ABA signalling (González-Grandío et al., 2017; Martín-Fontecha et al., 2018). Photosynthate availability, communicated via trehalose-6-phosphate, counteracts this directly at the bud to promote outgrowth by promoting FT and TSF expression (Fichtner et al., 2021), much as it does at the whole-plant scale to promote flowering generally (Wahl et al., 2013). Sucrose and citrate also act to promote branching via the SL signalling pathway and CK levels (Bertheloot et al., 2020; Barbier et al., 2021; Salam et al., 2021; reviewed by Barbier et al., 2023).

Hormone levels, their respective signalling pathways, and photosynthate are all influenced by temperature (reviewed by Castroverde and Dina, 2021) although, with the exception of GA and ABA, the effects of these on flowering have largely not been studied, especially in field contexts. However, temperature-regulated floral genes such as FLC, FT, and the SPL family are closely interlinked to this network. FLC changes the pattern of branch outgrowth along the main axis, and vernalization promotes branch outgrowth in Arabidopsis and A. alpina (Huang et al., 2013; Lazaro et al., 2018; Vayssières et al., 2020). Indeed, FLC is one of the largest effect loci controlling branching in natural variants of Arabidopsis (Huang et al., 2013; Jong et al., 2019; Hepworth et al., 2020). To what extent this is a pre-patterning effect on meristem floral progression versus an effect later and local to buds is not clear, as FLC binds to the BRC1 locus in Arabidopsis while PEP1 does not in A. alpina (Mateos et al., 2015, 2017). However, A. alpina is a perennial, and reactivation of PEP1 in meristems is important to return plants to a non-flowering state over summer and autumn (R. Wang et al., 2009; Lazaro et al., 2018). In B. napus, up-regulation of FLC copies in warm winters is associated with up-regulation of BRC1 copies in floral buds (Lu et al., 2022). In Citrus trees, local patterning of the FLC family member CcMADS19 is associated with local control of FT expression, suppressing new flower formation adjacent to forming fruits (Agustí et al., 2020; Mesejo et al., 2022), suggesting that local FLC expression may be a conserved mechanism for patterning perennials over seasons.

Downstream of FLC, SPL15 and SPL9 directly repress BRC1 expression, promoting outgrowth of the bud, but this regulation is disrupted by the SL signalling components SUPPRESSOR OF MAX2-LIKE (SMXL6/7/8) proteins (Xie et al., 2020), a regulatory loop first identified in rice (Lu et al., 2013; Song et al., 2017). SLs have also been identified as delaying flowering via promotion of the protein activity of the repressive AP2 family member TOE1 (Bai et al., 2024). SLs communicate a number of abiotic inputs to plant architecture (Castroverde and Dina, 2021): whether temperature is an important factor for SL-mediated control of branch outgrowth is not yet well studied in the Brassicaceae.

The florigens FT and TSF have a direct local promotive effect on branch outgrowth, moving into buds via the vasculature from the subtending leaf to interact with BRC1 at the protein level and promote outgrowth (Huang et al., 2013; Niwa et al., 2013). BRC1 presence in AMs reduces expression of SOC1, AP1, and FUL, and, if misexpressed at the main SAM, delays flowering in Arabidopsis (Niwa et al., 2013). Thus the activities of BRC1 and FT are opposing for both bud outgrowth and floral identity, an effect conserved between Arabidopsis and wheat (Niwa et al., 2013; Dixon et al., 2018). Indeed, in poplar, the opposing action of BRC1 and FT orthologues is key to seasonal dormancy, with short day-induced growth cessation and bud set in autumn involving the down-regulation of PtFT1 and up-regulation of BRC1 (Böhlenius et al., 2006; Singh et al., 2018). As noted previously (‘Floral buds are winter structures’), SVP-like genes are also involved in poplar seasonal dormancy, further enhancing the links between seasonal cessation of growth and reproductive identity of meristems.

The end of the road: inflorescence arrest

Just like the initiation of inflorescences, and of organs therein, growth arrest of the reproductive structures is a key determinant of reproductive output. Indeed, plants do not flower indefinitely, rather inflorescences tend to interrupt their development after producing a variable number of fruits through a process termed proliferative arrest (Hensel et al., 1994). In Arabidopsis, two key phenomena are needed for this. On the one hand, the inflorescence must cease to initiate new floral primordia through IM arrest. On the other hand, not all of the initiated primordia complete floral development such that some of the youngest floral buds arrest their own development through the process of floral arrest.

At its core, IM arrest induces a reversible inactivation of the meristematic activity, coupled with a transcriptional down-regulation of the meristem maintenance integrator WUS (Balanzà et al., 2018; Merelo et al., 2022). This is under strong regulation by endogenous signals, particularly age. In older plants, higher levels of miR172 induce a transcriptional repression of AP2, a known activator of WUS. The environmental control of inflorescence arrest has been less investigated, but recent work has identified a role for temperature in regulating this process (Miryeganeh, 2020; González-Suárez et al., 2023), adding to a growing body of literature which supports that warming accelerates the end of flowering in various Brassicaceae (Lomas and Burd, 1983; Satake et al., 2013; Cao et al., 2016; Nagahama et al., 2018; Komoto et al., 2024). The molecular basis of this phenomenon is still not fully understood, but an attractive hypothesis is that temperature integration takes place upstream of FT, similar to what has been observed for the onset of flowering. After floral transition, FT is up-regulated over time in a temperature-dependent manner, and this seems necessary to trigger timely inflorescence arrest (González-Suárez et al., 2023). Within this model, how temperature feeds into FT expression is not well understood, but it is possible that this is through MADS box transcription factor complexes such as SVP–FLM and SVP–FLC, as knockout of these genes disrupts the response to temperature (González-Suárez et al., 2023). More intriguing is how a floral promotive signal such as FT can act to shut down the proliferative activity of the IM, which still remains an open question. In addition to upstream regulation of FT, another possibility is that temperature inputs directly into the age pathway, as miR172 accumulation is also induced in the warmth by known thermosensory regulators, such as FCA, SVP, and FLM (see ‘The making of inflorescence meristems’).

In the field, temperature control of inflorescence arrest is even less well characterized. However, perennial relatives of Arabidopsis show remarkable seasonal changes in their transcriptome, particularly in genes implicated in flowering (Nagano et al., 2019; Komoto et al., 2024). The end of the reproductive season has been associated with a peak of FT expression, which sharply declines thereafter and remains off until the next spring (Satake et al., 2013; Fig. 3B). In turn, FLC expression is gradually recovered during spring concomitant with an increase in temperatures, leading to its complete reactivation by the end of the flowering period. Both of these are highly temperature dependent. Interestingly, warming seems to accelerate the shutdown of FT to a greater degree than its activation and, as such, it shortens the duration of flowering (Satake et al., 2013). Several targets have been proposed to mediate this, including SOC1, GA synthesis genes, and members of the age pathway such as SPL15, all of which are negatively regulated by FLC and seasonally fluctuating (Komoto et al., 2024; Fig. 3B), but their specific function in inflorescence arrest is unclear. To better understand this, future research would benefit from extending lab studies to more realistic temperature scenarios using winter accessions of Arabidopsis, particularly considering the central role of FLC in promoting arrest in the Brassicaceae family (R. Wang et al., 2009; Satake et al., 2013).

Even after inflorescence arrest, already-initiated primordia can continue to develop and open into flowers. Thus, towards the end of the reproductive phase, plants must halt development of incipient floral primordia to focus their energy investment on pod filling, maximizing the reproductive potential of established fruits before the onset of autumn. It is important to distinguish floral arrest from FM termination, which marks the end of floral organogenesis, leading to a complete whorl of sepals, petals, stamens, and carpels (Min and Kramer, 2023). Instead, floral arrest takes place after all of the floral organs have been specified, and results in the distinctive cluster of undeveloped buds typical of arrested inflorescences (Hensel et al., 1994; Walker et al., 2023). Recently, this has been shown to be regulated by the same triad of HD-Zip TFs that are regulated by BRC1, and which promote ABA accumulation in floral buds during arrest in a manner similar to dormant axillary buds (Sánchez-Gerschon et al., 2024). ABA has also been implicated in delaying development of B. napus floral primordia in response to warm temperatures through up-regulation of BRC1, HB21, HB40, HB53, and ABA biosynthesis genes such as NINE-CIS-EPOXYCAROTENOID DIOXYGENASE-3 (NCED3) (Lu et al., 2022). This suggests that ABA might act as a general signal to block development of floral buds throughout the lifetime of the inflorescence, and that this inhibitory mechanism can be modulated by temperature signals (see ‘Floral buds are winter structures’). Nevertheless, additional signals are likely to be important for promoting floral bud arrest, as otherwise both newly initiated floral primordia and dormant axillary buds would arrest under certain conditions if ABA-related regulatory modules acted alone. Promising candidates for this are auxin and CK, which show promotive and inhibitory effects on floral bud arrest, respectively (Ware et al., 2020; Walker et al., 2023). Temperature sensitivity at the level of FT expression is key for timely IM arrest (González-Suárez et al., 2023), and this sensitivity has also been observed throughout inflorescence development in younger siliques (González-Suárez et al., 2024), posing the question of whether floral bud arrest may be under similar environmental control, though this remains to be investigated.

Conclusions

IMs are transient structures whose development is critical for the reproductive potential of all flowering plants. These tissues can be present throughout the reproductive phase and therefore experience a wide range of both environmental and endogenous signals, often interpreting these simultaneously to coordinate adaptive growth. Considering the significance of plant reproductive tissues for global agriculture and food security, understanding how the inflorescence responds to environmental signals at different developmental stages will be essential for unravelling the effects of climate change on crops. However, relatively little work has been done on the integration of temperature with later stages of meristem development, especially in the critical context of the field (Fig. 4). The baseline for our knowledge of the genetics of inflorescence formation mostly derives from the standard Arabidopsis growth conditions (long daylength >14 h, and warmer than 18 °C day and night); conditions which are at best loosely representative of late spring and high summer in most temperate regions. Within Arabidopsis, the genotypes more frequently studied are (with notable exceptions) rapid cyclers, such that flowering is highly accelerated in these conditions. It is perhaps unsurprising then that many recent advances in inflorescence progression and shoot architecture patterning derive from perennial relatives of Arabidopsis, and crops such as Brassica, poplar, and fruit trees, which have to be studied in longer lived contexts. Nevertheless, many important questions still remain, several of which regard the balance of effects at different temperatures and times of year. While research has uncovered many conserved pathways controlling seasonal inflorescence growth, for crop breeding details matter. Precise knowledge of temperature ranges at which the various genetic pathways act in the field, and the result of changing these by 2–3 °C, is currently lacking. Given current global warming trends, it is urgently needed.

Abbreviations

ABA

abscisic acid

AGL24

AGAMOUS-LIKE24

AM

axillary meristem

AP1

APETALA1

AP2

APETALA2

AP2L

APETALA2-LIKE

BRC1

BRANCHED1

CK

cytokinin

CLV3

CLAVATA3

DAM

DORMANCY ASSOCIATED MADS-BOX

ELF

EARLY FLOWERING

FCA

FLOWERING CONTROL LOCUS A

FLC

FLOWERING LOCUS C

FLM

FLOWERING LOCUS M

FM

floral meristem

FRI

FRIGIDA

FT

FLOWERING LOCUS T

FUL

FRUITFUL

GA

gibberellin

GA20OX

GIBBERELLIN 20-OXIDASE

GA3OX

GIBBERELLIN 3-OXIDASE

GAI

GA INSENSITIVE

HB

HOMEOBOX PROTEIN

HD-Zip

homeodomain leucine zipper

I1/2

inflorescence zone 1/2

IM

inflorescence meristem

LFY

LEAFY

MAF

MADS AFFECTING FLOWERING

PEBP

Phosphatidylethanolamine -binding protein

PEP1

PERPETUAL FLOWERING1

PIF

PHYTOCHROME INTERACTING FACTOR

PWR

POWERDRESS

RGA

REPRESSOR OF GAI

RGL

RGA-LIKE

SAM

shoot apical meristem

SL

strigolactone

SOC1

SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1

SPL

SQUAMOSA PROMOTER BINDING-LIKE

SVP

SHORT VEGETATIVE PHASE

TCP

TEOSINTE BRANCHED1/CYCLOIDEA/PCF

TEM

TEMPRANILLO

TFL1

TERMINAL FLOWER1

TOE

TARGET OF EAT

TSF

TWIN SISTER OF FT

V1/2/3

vegetative zone 1–3

VIN3

VERNALIZATION INSENSITIVE3

VM

vegetative SAM

WUS

WUSCHEL

Author contributions

PG-S and JH: conceptualization; PG-S: visualization and artwork; PG-S, TL, and JH: literature review and writing; PG-S, TL, JH, and SP: review and editing; JH and SP: supervision.

Funding

This work was supported by a Rank Prize New Lecturer Grant to JH. TL was supported by the UKRI Biotechnology and Biological Sciences Research Council Norwich Research Park Biosciences Doctoral Training Partnership (grant no. BB/T008717/1).

Data availability

No new data were generated or analysed in support of this research. Free to use, editable versions of all figures and diagrams can be found at http://www.pablidopsis.com/resources.