Unlocking allelic variation in circadian clock genes to develop environmentally robust and productive crops

In addition to TOC1, other PRR proteins are involved in circadian rhythm oscillations. PRR9, PRR7, PRR5 and TOC1 are consecutively expressed with moderate differences in their peak expression (Gendron et al. 2012; Liu et al. 2013; Matsushika et al. 2000; Michael et al. 2003; Nakamichi et al. 2010, 2012). PRR9, PRR7 and PRR5 also act as transcription repressors, and together with TOC1, repress the previously expressed PRRs and collectively restrict the transcription of CCA1 and LHY to a period during the dawn (Liu et al. 2013; Nakamichi et al. 2010, 2012). In turn, CCA1/LHY transcriptionally represses PRR9 and PRR7 during the dawn (Adams et al. 2015; Fogelmark and Troein 2014; Greenwood et al. 2022; Joanito et al. 2018; Kamioka et al. 2016; Shalit-Kaneh et al. 2018). Another key component of the clock that is transcriptionally repressed by CCA1/LHY and TOC1 is the evening complex (EC) (Gendron et al. 2012; Huang and Nusinow 2016). The EC is composed of the two plant-specific proteins Early Flowering 3 (ELF3) and Early Flowering 4 (ELF4) and the Lux Arrhythmo (LUX), a MYB-like TF that belongs to the GARP protein family or its close homolog Brother of Lux Arrhythmo (BOA) (Dai et al. 2011; Herrero et al. 2012; Nusinow et al. 2011). The EC complex is expressed after dusk and acts as a transcriptional repressor that binds to the promoter regions of PRR7, PRR9 and LUX (Chow et al. 2012; Dixon et al. 2011; Helfer et al. 2011). LUX (or BOA) is the EC component with DNA-binding capability, since ELF3 and ELF4 do not bind to DNA (Huang and Nusinow 2016). Other corepressors involved in the TTFL are COLD-Regulated 27 (COR27) and COR28, which are repressed by CAA1 and repress PRR5 and TOC1 (Li et al. 2016; Wang et al. 2017). Nevertheless, COR27 and COR28 cannot bind directly to DNA. Indeed, the TF that interacts with these to repress their molecular target(s) remains unknown (Li et al. 2016).

In contrast to the circadian components described above, the clock also possesses transcriptional activators playing key roles in the TTFL. Unlike CCA1 and LHY, other RVE proteins such as RVE8, RVE4 and RVE6 act as activators of the evening genes by binding to the EE motifs (Farinas and Mas 2011; Hsu et al. 2013; Rawat et al. 2011). These RVEs interact with the coactivators Night Light-Inducible and Clock-Regulated 1 (LNK1) and LNK2 to form a complex and activate the transcription of TOC1, PRR5 and ELF4 (Rugnone et al. 2013; Xie et al. 2014). Conversely, PRR9, PRR7, PRR5, and TOC1 repress the transcription of LNK1 and LNK2 from noon to early evening (Rugnone et al. 2013). Other transcription co-activators that play an early role in the circadian oscillations are the Light-Regulated WD1 (LWD1) and LWD2 (Wu et al. 2008). Dawn transcriptional activation of CCA1 requires the recruitment of LWD1 and LDW2 in conjunction with TCP20 and TCP22, two CHE-related transcriptional regulators that belong to the TCP family are able to bind to the TCP binding sites (TBS) in the promoter regions (Mockler et al. 2007; Wu et al. 2016). Several other clock genes have been proposed to recruit LWD1 and LWD2 to their promoter regions to be activated, including PRR9, PRR5, and TOC1 (Wu et al. 2008). Despite what has been elucidated to date, the TF(s) involved in recruitment of LWDs to the genes expressed at dusk remains to be determined.

Circadian clocks are drivers of biological rhythms that anticipate the light, dark, and temperature cycles of the rotating Earth. Thus, the circadian clock affects the timing and order of different biological processes and modulates signaling pathways in response to stimuli, which become dependent on the time of day they occur. Artificial selection has led to large changes in the circadian clock of plants through domestication (Siqueira et al. 2023). Indeed, domestication has induced variation in the peaks determining photoperiod perception, whereas changes in the time between peaks has smaller effects, as noted in tomato (Xiang et al. 2022).

Measuring circadian rhythms is a significant challenge in quantifying variation associated with circadian dynamics. The circadian rhythm (i.e., behavioral changes that follow a 24-h cycle) consists of a period (i.e., the time taken to complete one cycle, phase (i.e., the time of day when this peaks), and amplitude (i.e., the distance between the peak and the base line of oscillation). The circadian oscillator refers to networks of biochemical feedback loops that generate 24-h rhythms in living organisms (Saini et al. 2019). Equally important is the rhythm robustness (including resilience after disturbances); i.e., parameter (phase shifts) changes over time.

Circadian clocks help plants to buffer environmental changes through adjustments in their development and physiological characteristics (Sánchez and Kay 2016). Arabidopsis thaliana has become the plant model system for extensive investigation to understand how plants continuously adjust their circadian oscillator. The circadian oscillator has a dynamic plasticity (i.e., trait changes expressed by a genotype across different environments (Webb et al. 2019). Circadian clocks in plants are responsive to environmental signals via changes in period and phase. The genes regulating the circadian clock—e.g. transcriptional regulators acting as repressor and activators—control various processes during their development, particularly those related to primary metabolites in the life cycle (Kim et al. 2017). The clock transcriptional network of Arabidopsis is partly present in other angiosperms (Nakamichi 2020), which poses challenges to translate results from this model plant to crops.

Selection for early flowering during both domestication led to crop expansion related to photoperiodic adaptation (Lin et al. 2021). Selection of late flowering mutants allows longer growth periods to maximize yield in low latitudes. Short day plants, however, flower later when expanding to high latitudes due to their photosensitivity (Greenham et al. 2017). Various genes [phytochrome red light signaling; circadian clock ELF3, ELF4, LUX, and Pseudo-Response Regulator37 (PRR37), a legume-specific maturity (E1), and a short-day monocot-specific heading date (Ghd7)] contribute to adaptability of grain crops such as barley (Hordeum vulgare), maize (Zea mays), rice (Oryza sativa), pea (Pisum sativum), sorghum (Sorghum bicolor), soybean (Glycine max), wheat (Triticum aestivum) (McClung 2021 and references therein) and vegetables, e.g. tomato (Solanum lycopersicum) (Xiang et al. 2022). These are putative orthologs of genes found in Arabidopsis, which facilitated adaptation of these crops to new latitudes beyond their centers of origin. Osmotic stress, high temperature, light intensity, or total period length, amongst others, can change plant circadian oscillators.

During latitudinal migration, genes related to flowering time were selected in tomato (Blanca et al. 2022). Mutations in two genes (including phytochrome A-specific light signaling gene, EID1) were involved in this process. Moreover, deletion of clock gene(s) enhanced crop adaptation in tomato (Müller et al. 2016). Signatures of a selective sweep were detected in the genomic region harbouring EID1, which enhances performance under long day photoperiod in tomato. A partial deletion in the clock genes LNK2 is associated with increasing the photoperiod lengthening in this neutral-day crop (Müller et al. 2018). This example provides information about how humans selected for modulation of circadian rhythms when moving tomato cultivation away from the equator. It appears that mutant alleles in both EID1 and LNK2, which arose in early cultivated tomato and were further selected during its domestication, affected the light input to the circadian oscillator and led to the observed light-conditional deceleration when changing the day–night cycles in the high latitudes. Transcription profiling of near isogenic lines (NILs), homogeneous genetic stocks generated by repeated backcrossing with selection for the desired character at each round of crossing, combining wild alleles of EID1 and LNK2 grown under two different photoperiods further revealed that the former had a larger effect on tomato response to daylength (Xiang et al. 2022). This may be attributed to the shift elicited by EID1 in the circadian rhythm of this vegetable crop. Other clock-associated variants that facilitate crop latitudinal adaptation are related to the regulation of seasonal rhythms during developmental transitions, e.g. flowering and tuberization (Nakamichi 2015). These research findings are useful for breeding to increase crop productivity at high latitudes, particularly in crops with detrimental circadian asynchrony.

As the basis of an intracellular timekeeping system, circadian clock genes generate approximately 24-h rhythms in physiology and behavior. The detection of clock genes can be based on two complementary approaches. The forward-genetic approach is based on arhythmic mutants that produce rhythms deviating from 24-h cycles. Although this approach has been instrumental for revealing the detailed molecular mechanisms of circadian rhythms in model systems and various outcrossing species (King and Takahashi 1996; Miyoshi et al. 2019), it has some significant drawbacks. First, its translational relevance for a broad spectrum of species and biological processes is challenging since only a limited number of mutations can be identified. Also, many clock genes are not evolutionarily conserved; e.g., some clock genes are species-specific (Calixto et al. 2016). This suggests that ideally, detection of clock genes should be based on individual species. To overcome these drawbacks, a forward-genetic approach has been developed which takes advantage of genome-wide mapping and association genetics available for a range of species (Jones et al. 2019). Such an approach could identify a more complete set of clock genes. For example, using a mapping population, a number of QTL for circadian clock parameters, such as period, phase, and amplitude, were identified in Arabidopsis thaliana (Rubin et al. 2019). A review of progress in clock QTL mapping and its application to understanding the genetic architecture of plant adaptation is available (Anwer and Quint 2022).

Most clock mapping attempts have been based on genotype–phenotype association analysis using markers. However, the results from these analyses remain an approximation of reality as the circadian rhythm is complex, involving interactions of many genes with environment. A computational algorithm to map a complete set of clock genes using association genetics has been developed (Sun et al. 2021). This algorithm is powerful in terms of its capacity to reconstruct omnigenic interactome networks for rhythmic traits (Sun et al. 2021). Omnigenic theory predicts that all genes that an organism may carry can be involved in controlling a complex trait developed through gene–gene interaction networks (Boyle et al. 2017). Factoring epistasis interactions (i.e., epiGWAS) could resolve some of the missing heritability for complex traits often found in GWAS or biparental populations (Slim et al. 2020). The algorithm can coalesce all genes into bidirectional, signed, and weighted networks, allowing for the characterization of a roadmap of how each and every gene affects circadian rhythms through direct and indirect pathways. It can also discern how each gene functions. First, some clock genes are significant due to their own strong intrinsic capacity, thus implying that these genes govern circadian rhythms without relying on the promotion of other regulators. Second, the significance of some genes is due to favorable regulation involving other genes. The practical use of these two gene types to improve rhythmic traits should be based on different strategies. In addition, some genes are detected to be insignificant not because their effects are inherently small but because they are inhibited by negative regulators. Thus, the effects of these insignificant genes can be released by knocking down the negative regulators. In theory, this is also an approach for retrieving missing heritability, a common phenomenon detected in current genome-wide association studies. The statistical robustness of this approach has been confirmed (Sun et al. 2021) and also its biological value has been validated by analyzing periodic gene expression data of the malaria parasite Plasmodium falciparum (Smith et al. 2020) and growth trajectory data under contrast environmental conditions in a woody plant, Populus euphrates (Wang et al. 2021).

Lack of high throughput protocols has hindered progress in investigating circadian dynamics where such assays were either manually intensive, or of low throughput, technologically expensive, or required genetic modification (Ford et al. 2016; Nagano et al. 2012; Sugiyama et al. 2001; Xu et al. 2010). A non-invasive high-throughput assay based on chlorophyll fluorescence was deployed to measure variation in circadian rhythms among wild barley (Hordeum vulgare ssp. spontaneum) from diverse ecogeographical environments. Wild barley accessions showed variability for circadian traits. Circadian period lengths were correlated with temperature and site of origin of the plants, while the amplitudes of the rhythms were correlated with soil composition (Dakhiya et al. 2017). Delayed fluorescence (DF) image-based technology is a universal tool to measure circadian rhythms in higher plants (Gould et al. 2009). A further optimization of DF protocol (Rees et al. 2019) led to discovering significant differences in circadian periods between widely grown rapeseed (Brassica napus) and wheat cultivars. The latter displayed rhythm robustness under constant light vs constant darkness, thereby suggesting divergent networks underly circadian control in these species.

Plant growth rhythm in structural traits measured by terrestrial laser scanning (TLS) is crucial for better understanding of plant response to the ever-changing environment. TLS has been utilized for studying diurnal rhythms in maize structural traits at both the plant and the leaf levels from time-series data at four key growth periods (jointing, bell-mouthed, heading, maturity) under optimal and cold stressed field environments (Jin et al. 2021). Leaf inclination angle between the jointing stage and well-mouthed stage significantly decreased, with leaf azimuth remaining stable after the jointing stage. A few individual-level structural rhythms were consistent with leaf-level structural rhythms. The circadian rhythms of some traits were diurnal. Environmental factors (particularly temperature) led to better correlations with leaf traits under cold stress than under optimal conditions. This research highlighted the potential of time-series TLS in studying outdoor agricultural chronobiology (Hotta 2021; Steed et al. 2021).

Daily rhythmic movement of leaves has been extensively used to estimate circadian period in Arabidopsis. Manual estimates of circadian period by leaf movement is limited to analyzing one leaf or a cotyledon at a time. TRiP (Tracking Rhythms in Plants) estimates rhythm period using a motion estimation algorithm to measure whole plant images. It successfully tracked the movement of cotyledons and leaves of Arabidopsis RILs without the need to select individual leaves and could be applied to plant species with diverse leaf morphology (Greenham et al. 2015). A lightweight digital inertial measurement unit (IMU) sensor measures leaf movements and is easily attached to a leaf or plant organ to record angular traits in real time, both simple movement as well as complex lamina motions, for two dimensions with high resolution and detects small changes of mutations in the leaves of varying ages in diverse plant species (Geldhof et al. 2021).

Clearly, significant progress has been made toward dissecting the components of circadian rhythms, making it possible to study rhythmicity and apply this knowledge in plant breeding programs. Further advances are needed to develop protocols adapted to handling early generation breeding populations for selection of segregants with desired rhythms and expressing clock genes.

Phytohormones are signaling molecules present in plants that occur in extremely low concentrations and play a major role in plant growth, development, reproduction, and fitness. Abscisic acid (ABA), ethylene (ET), jasmonates (JA), and salicylic acid (SA) are some of the stress-responsive phytohormones, while auxin (AUX), brassinosteroids (BRs), cytokinins (CK), and gibberellins (GA) are growth promoting phytohormones (Gray 2004; Verma et al. 2016). Phytohormones control specific features of the plant–circadian system, though in distinct ways. CK delays circadian phase, AUX regulates circadian amplitude and clock precision, and BR and ABA modulate circadian periodicity. This indicates that plants have multiple input/output feedbacks to adjust the clock to external signals and accurately maintain the clock system (Hanano et al. 2006). The circadian clock interconnects with an intricate network of different pathways, including phytohormones (Singh and Mas 2018).

Arabidopsis has been extensively investigated to unlock the molecular basis of connecting the circadian rhythms and phytohormone pathways in response to endogenous and exogenous cues (Fig. 1C). A complex interaction between the circadian clock and ABA pathways in Arabidopsis regulates plant responses to abiotic stress tolerance. Late Elongated Hypocotyl (LHY) transcription factor (TF) modulates ABA biosynthesis and ABA responses, which drives the rhythmic accumulation of ABA to ensure peak accumulation of phytohormones when water deficit is most severe. LHY gene represses 9-cis-epoxycarotenoid dioxygenase enzymes, a rate-limiting step of ABA biosynthesis. Plants overexpressing LHY show reduced levels of ABA under drought, whereas loss-of-function mutants exhibited altered rhythm of ABA accumulation. LHY also binds the promoter of multiple components of ABA signaling pathways and regulates responses downstream of the phytohormone. LHY promotes the expression of ABA-responsive genes involved in stress tolerance while inhibiting the negative effect of ABA on seed germination and plant growth (Adams et al. 2018).

Circadian clock gates respond to environmental signals to maximize plant fitness. Clock gating is the clock’s regulation of signaling pathways in a manner that the same stimuli given at different time points result in different responses, usually different intensities, which help plants to reduce stress responses at certain times of the day to ensure optimized growth. The MYB96 TF connects with the clock oscillator to shape the circadian gating of ABA responses by directly binding to the TOC1 promoter to regulate its expression as evidenced in Arabidopsis myb96 and toc1-3 mutants. CCA1 may also regulate MYB96-TOC1 function to alter its circadian expression. Thus, a complex circuitry of CCA1-MYB96-TOC1 regulatory interactions provides the mechanistic basis to connect circadian clock and stress signaling to optimize plant defense and fitness (Lee et al. 2016). BRs control many physiological processes via extensive cross-talks with diverse signaling networks. The BR-activated TF bri1-EMS-Suppressor 1 (BES1) together with TPL represses CCA1 and LHY expression at night by binding to their promoters. This repression by BR treatment is compromised in bes1-ko and tpl8 mutants. Consistent BRs treatment for an extended periods shorten the circadian period, where BR-induced rhythmic shortening was impaired in bes1-ko and tpl8 single mutants as well as in the double mutant, cca1-1lhy-21, providing evidence that BES1/TPL-CCA1/LHY module shapes circadian gating of BR signaling in plants (Lee et al. 2020).

Leaf senescence is associated with plant ageing or induced by external stresses to relocate nutrients from aging tissues to juvenile and reproductive organs. JA, an endogenous signal, induces leaf senescence (Sobieszczuk-Nowicka et al. 2018). The evening complex (EC), a core component of the circadian oscillator, negatively regulates leaf senescence. EC is involved in JA signaling and response as evidenced by accelerated leaf senescence in EC mutants upon JA induction. EC binds MYC2 promoter to repress its expression. MYC2 is the key activator of JA-induced leaf senescence as in EC triple (myc2 myc3 myc4) mutants. The accelerated JA-induced leaf senescence in EC mutant is abrogated by myc2 myc3 myc4 triple mutation, providing direct evidence that a core component of the circadian clock gates JA signaling to regulate leaf senescence (Zhang et al. 2018b). JA signaling regulates multiple outputs of plant defense and growth to minimize adverse effects of growth-defense trade-offs in complex environments (Li et al. 2022).

Auxin plays a central role in plant growth and development in response to environmental cues. RVE1 is a Myb-like clock regulated TF homologous to CCA1 and LHY. Its inactivation has no effect on circadian rhythmicity but instead causes a growth phenotype, indicating it as a ‘clock output’ that affects plant development. CCA1 regulates growth via bHLH TFs PIF4 and PIF5. However, RVE1 acts independent of PIF4 and PIF5. RVE1 regulates auxin levels by promoting its production during the day but not at night. It positively regulates the expression of auxin biosynthesis gene YUC8 to promote growth. Thus, YUC8 connects two signaling networks to coordinate plant growth with rhythmic changes in the environment (Rawat et al. 2009).

OsCCA1, a key clock component in rice, confers tolerance to drought, osmotic, and salinity stresses. There are 692 direct transcriptional target genes of OsCCA1 where, genes involved in ABA signaling are substantially enriched. OsCCA1 binds the promoters of OsPP108 and OsbZIP46 to activate their expression. Oscca1 null mutants display enhanced sensitivity to ABA signaling. Thus, abiotic stress adaptation by OsCCA1 is orchestrated by ABA signaling, linking the circadian clock with ABA signaling in rice (Wei et al. 2022). The OsRCAR10-OsABI5-OsPRR95 feedback loop regulates ABA signaling to fine-tune seed germination and seedling growth, indicating molecular link between ABA signaling and circadian clock in rice (Wang et al. 2022a).

The circadian clock that is based on interlocked negative loops of transcriptional activators and repressors is crucial for adjusting crop growth. Genes that maintain the plant circadian clock “ticking” (e.g. morning genes: CCA1, LHY; mid-day genes: PRR9, PRR7, PRR5; evening genes: TOC1, ELF3, ELF4, LUX) are supported by evidence in model plant Arabidopsis and other crops. Clock genes impacting biological rhythms and associated with plant growth, development, reproduction, and stress tolerance, have been identified in staple crops. Likewise, the role of phytohormones in mediating clock and stress-responsive gene expression is known. Maintaining the intricate interactions between circadian rhythms and clock genes‒phytohormones‒trait/stress-responsive genes may open new paths to tailor crops adapted to external environmental cues. Crop adaptation by changing flowering time is a good example of the potential of clock genes, where loss of some segments of circadian clock genes facilitated adaptation of tomatoes to distant growing habitats far from its center of origin (Müller et al. 2016, 2018).