Introduction

The genetic regulation of plant development and flowering has been dissected in the model plant Arabidopsis to the largest extent, revealing the complex intricate system that enables the plants to percept the various environmental signals and to adjust their development accordingly. The large amount of information collected in Arabidopsis contributed greatly to the dissection and to the better understanding of this regulation network in cultivated plant species1. In temperate small grain cereals, it helped to identify various genes of the vernalization and photoperiod response pathways, several members of the circadian clock (input, core oscillation and output genes), photoreceptors, plant hormonal pathways, and central elements of flower initiation, with new findings published almost daily. These comparative studies have revealed similarities and basic differences between Arabidopsis and cereals in the structures and functionality of orthologous genes and in the compositions of the specific regulation pathways.

Due to these studies, the basic features of the two most important regulation pathways in cereals—vernalization and photoperiod response—have been fairly deeply dissected. The vernalization requirement is determined by the epistatic interactions between two major vernalization response gene families. These are the dominant flowering activators VRN1 (orthologue of the Arabidopsis gene AP1) and the dominant flowering repressors VRN2 (ZCCT, having no orthologue in Arabidopsis)2,3,4,5,6,7,8,9,10,11. According to the latest flowering regulation models of cereals, in autumn, after germination, when the days are still sufficiently long, the active VRN2 gene prevents the transcription of the VRN3 gene (orthologue of the Arabidopsis gene FT), which could be important in maintaining the activity of VRN1 at a low level12,13. In the course of vernalization, the activity of the VRN1 gene is gradually enhanced by the cold effect, leading to increasing expression of VRN1 transcription factors, which inhibit the functioning of VRN2 while stimulating that of VRN314,15,16. The latter, in turn, enhances the activity of the VRN1 gene, thus inducing heading14,16,17,18. The VRN2 gene is unaffected by low temperature as opposed to photoperiod. Short day inhibits while long day stimulates the expression of this gene6,12,19,20. In barley, both phenomic and gene expression analyses have confirmed that VRN2 appears to be controlled by both the CONSTANS (CO) and VRN1, suggesting that this gene is a joint element in photoperiod and vernalization regulatory pathways6,20. PPD1 (PRR37) is the major gene of photoperiod sensitivity21,22,23,24, and it acts with both the photoperiod regulating pathway and the circadian clock. In wheat, the dominant photoperiod insensitivity alleles in PPD1 result in rapid heading irrespective of the photoperiod, an effect that is more pronounced under short photoperiods25. VRN3 genes (FT1) are the central integrators of the vernalization and photoperiod pathways, and as such, these important genes are under the control of VRN1, VRN2 and PPD1 series26,27,28,29. The functional polymorphisms in these plant developmental genes have been thoroughly analysed, revealing the similarities and differences between the various cereal species11,30,31,32,33,34,35. In barley, it has already been described that different allele combinations of genes that are responsible for the regulation of vernalization requirement and photoperiod sensitivity lead to different plant development types6,20. However, the knowledge on the complex regulation system of plant development in cereals is still incomplete, necessitating further experiments.

One of the major differences between Arabidopsis and cereals appears in the relationship of floral differentiation and stem elongation. In Arabidopsis, these two processes normally occur in parallel; the first visible sign of the vegetative—generative turning point is the initiation of bolting from the vegetative rosette stage. Thus, the differentiation of the various floral organs takes place during the stem elongation. In the case of cereals, however, these two processes are separated in time. The vegetative—generative transition, followed by the floral differentiation and even the terminal spikelet formation in the case of wheat have all already taken place by the time the intensive stem elongation starts (the first visible sign for the latter is the appearance of the first node at the base of the stem). This phenomenon is well documented in the previous studies36,37, but it has not been considered in genetic models of cereal flowering regulation. In temperate cereals, photoperiod and low temperature vernalization are the two most decisive environmental factors determining the developmental processes of the plant14,27. In addition, several other factors fine-tune the heading or flowering, including the ambient temperature above vernalizing levels, the various characteristics of light, and water and nutrient availability that can show large variability within a short time frame, and it also varies by location and season.

The temperature above vernalizing levels has a more complex effect on plant developmental dynamics than the day length, as it may exhibit considerable deviation not only seasonally, but also at an annual or daily level36,38,39,40,41,42. Research on cereals proved that the ambient temperature influences not only the heading date, but also the number and development rate of leaves, tillers, and spikelets43,44,45. In addition to the light impulse, ambient temperature is also an important input signal in the entrainment of the circadian rhythm46. In barley seedlings germinating in the dark, the daily cycling in ambient temperature in itself is able to activate the circadian rhythm47. The ability to compensate for changes in external temperature is also characteristic of the circadian rhythm. The length of the period remains relatively constant in a wider range of ambient temperature, which is a fundamental element in keeping the cell metabolism at a relatively stable level, regardless of the external temperature.

Circadian rhythm with its incoming and outgoing regulation networks has been identified as the major component in continuously adjusting the various plant processes to the actual environmental conditions. As such, it must be also an important part of regulating the process of intensive stem elongation in cereals. The circadian rhythm is an internal regulatory mechanism (autonomous oscillator) by which plants synchronise their internal biological processes with daily changes in the various external environmental conditions including temperature and light47,48,49. Plant circadian clocks consist of multiple interconnected transcription / translation feedback loops1. The central oscillator mechanism of the clock is based on proteins providing feedback by directly or indirectly repressing their own activity or their own transcriptions from the corresponding genes. One of the central loops consists of three genes TOC1, and CCA1/LHY. Expression of genes encoding the transcription factors CIRCADIAN CLOCK-ASSOCIATED 1 (CCA1) and LATE-ELONGATED HYPOCOTYL (LHY) peaks in the early morning50,51. In the morning, CCA1 and LHY genes inhibit the expression of the TIMING OF CAB EXPRESSION 1 (TOC1) gene, then in the late afternoon, the expression of the TOC1 gene is increased, which reduces the expression of the CCA1 and LHY genes (negative feedback loop)52,53. Subsequently, the expression of PSEUDORESPONSE REGULATOR (PRR) genes also increases until the evening hours (PRR9, PRR7, PRR5, PRR3, PRR59, PRR73, and PRR95)51. The ARRHYTHMO (LUX), EARLY FLOWERING 3 (ELF3) and ELF4 circadian genes show maximum expression in the evening and night periods, respectively and they are part of the evening complex (EC)29,51. In Arabidopsis, the circadian clock regulates the induction of FT and its critical transcriptional inducer CONSTANS (CO) via an external coincidence mechanism in which light coincides with an inductive window that is gated by the circadian clock. In this process, the clock component GIGANTEA (GI) plays an important role in forming a transcriptional complex via cycling changes in their protein abundance1,54. The study of the molecular bases of circadian clock in monocots has lagged behind1. Orthologs of circadian clock genes have been identified in cereals; however, their functions can largely differ from those described in Arabidopsis55,56,57,58,59. In barley and in wheat, to a smaller extent, several orthologs have been identified as the candidate genes of various earliness per se loci, which contributed significantly to the latitudinal spreading of these species38,58,60,61,62,63.

The photoreceptors are important input genes of the circadian oscillation, acting as the major sensors of light, but new results demonstrate the possible participation of some of its members in ambient temperature sensing as well64,65. In Arabidopsis, two major photoreceptor families have been identified, the five members of phytochromes (from PHYA to PHYE) and the cryptochromes (CRY1 and CRY2)57,66,67,68,69. In cereals, the orthologues of PHYA, PHYB, PHYC, CRY1 and CRY2 have already been identified57,70,71,72. Their effects on flowering regulation have been proven mostly in association with photoperiodism and light spectral composition, but only limited information is available on their reaction to ambient temperature.

In summary, the regulatory mechanism of the plant circadian rhythm has been extensively explored in Arabidopsis, and several homologous genes have already been described in wheat, but few of these have been studied thoroughly. Much of this knowledge has been originated from diploid Triticum monococcum and tetraploid species with smaller genome sizes, or from specific crossing lines (RIL, NIL, mutant and transgenic lines). In addition, even less is known about the intrinsic relationship between circadian rhythm, major plant developmental genes, and photoreceptors in cultivated wheat varieties, and about the influence of secondary environmental elements such as ambient temperature. These environmental elements have a fundamental impact on the plant development. However, based on the results of our previous studies, we hypothesize that there may be significant differences in the molecular-genetic regulation of the intensive stem elongation phase between mono- and dicotyledonous plants.

Therefore, the aim of this research was to examine the possible associations between the major plant developmental, circadian, and photoreceptor genes in the diverse genetic backgrounds of three winter bread wheat cultivars as influenced by ambient temperature. For this purpose, the daily expression patterns of three vernalization response genes (VRN1, VRN2 and VRN3), the major photoperiod sensitivity gene (PPD1), seven circadian clock genes (CCA1, PRR95, TOC1, LUX, ELF3, GI and CO1), and five photoreceptor genes (PHYA, PHYB, PHYC, CRY1 and CRY2) were studied under controlled environmental conditions (phytotron). The application of long photoperiod (16 h) was common in all the environments, while the ambient temperature profile differed between the growth chambers. Two constant levels of temperature − 18 °C and 25 °C—were applied in the growth chambers. Vernalization, ambient temperature, daily timing, and genotype effects were all significant determinants of the variance in gene expressions, but their contributions varied remarkably with the different gene families. Unlike Arabidopsis, bread wheat has a very large and complex genome. Due to epistatic interactions between different allelic variants of plant developmental genes, it is often not possible to directly study the effect of a target gene. Consequently, here we can observe just indirect gene effect.

Materials and methods

Plant materials

Three hexaploid winter wheat cultivars (‘Mv Toborzó’ /AT1/, from Hungary, ‘Tommi’ /AT3/, from Germany and ‘Charger’ /AT20/, from United Kingdom) with different plant developmental and genetic diversities were included in the experiments. These plant materials originated from the GeneBank collection of the HUN-REN Centre for Agricultural Research, Hungary. They were selected on the basis of a previous study, where more details can be found on them36 (Fig. S1). In addition, the phenological responses of these genotypes to 18 °C and 25 °C have also been characterized in detail36. Based on the genotypic analysis with gene specific markers, all the three genotypes carry winter alleles of the major vernalization genes (VRN-A1, VRN-B1 and VRN-D1) and in terms of photoperiod sensitivity, ʻMv Toborzóʼ carries the insensitive allele of PPD-D1, and the other two contain the sensitive allele type (‘Tommi’ and ‘Charger’), while all three cultivars carry the same sensitive allele in PPD-B136 (Table S1).

Experimental conditions and design

The controlled tests consisted of combinations of one photoperiod (16 h light period in a 24 h day) and two constant ambient temperature levels in three treatments (18 °C vernalized /18 °Cvern/, 18 °C unvernalized /18°Cuv/ and 25 °C vernalized /25 °C vern/), while all the other environmental factors were kept uniform across the treatments, in the phytotron (Phy) facilities of the Agricultural Institute, HUN-REN Centre for Agricultural Research, Hungary. The two applied temperatures represented optimal (18 °C) and supra‐optimal but no heat‐stress (25 °C) temperature levels. The same two Conviron PGV‐36 growth chambers (Controlled Environments Limited, Winnipeg, Canada) were used for these purposes, each representing one temperature level. The 16-h photoperiod was set in all environments at the same period of the day, from 5:00 AM (ZT0) to 21:00 PM (ZT16). The light source was metal halide lamps with a broad light spectrum (to create a spectrum which closely resembles natural light), and the light intensity was set to 240 μmol m−2 s−1 photosynthetic photon flux density—PPFD (suitable for the seedling stage), being the same as in the previous, corresponding study36. The humidity was constantly close to 70%. For all the experiments (18°Cvern, 18°Cuv and 25°Cvern), the plants were germinated in Jiffy® pots (Zwijndrecht, Netherlands) at room temperature for one week. This was followed (with the exception of unvernalized plants – 18°Cuv) by the vernalization treatment at 3 °C for 60 d under short photoperiod (9 h light period in a 24 h day) and low PPFD (20 μmol m−2 s−1). At the end of the vernalization the plants were at 1–2 leaf stages. They were then transplanted into pots 12 cm in diameter and 18 cm in height, with a soil capacity of 1.5 kg filled with a 4:1 mixture of garden soil and sand, and they were placed into the growth chambers. For the purpose of RNA sampling and apex examinations four plants per pots of each genotype were planted (50 pots holding 200 plants per genotype). The plants were left growing for two weeks, and then the collection of leaf samples began. After the completion of leaf tissues collections in each treatment, from the remaining seedlings the main tillers of three, two-week old plants per genotypes and treatments were dissected in order to examine the apex structure by measuring the length of the apex (APL) and evaluating the apex’ developmental phase using the Waddington scale73. For this purpose, the visible external leaves (ELN) and then the non-visible internal leaf sheaths (ILN) were removed by scalpel and counted, then the apex was photographed under a stereomicroscope (Alpha STO-PL6, Elektro-Optika Ltd, Érd, Hungary) equipped with length measuring capacity (ScopePhoto 3.0., ScopeTek Ltd, Bedford, UK). Furthermore, plant height (PH), and reproductive tillers (RT), were also measured. For determining the later developmental phases, two plants per pot and two pots per genotypes were also placed in each environment, in addition to the 50 pots of plants used for RNA sampling. The latter plants were raised till maturity or—in the case of the unvernalized treatment—until the termination of the experiment (at 120 days). This group of plants was not sampled for RNA isolation. In the latter sets of plants, two phenophases were defined based on the Zadoks’ scale74: appearance of the first main stem node (ZD31) and heading (ZD59), both expressed in days.

Genetic characterisation of genotypes

Genomic DNA extraction was performed with the DNeasy® Plant Mini Kit (Qiagen Ltd, Hilden, Germany) from young leaves (100 mg) according to the manufacturer’s instructions. DNA samples were stored at -20 °C until use.

KASP marker system

KASP marker system was performed at the John Innes Centre (Norwich Research Park, Norwich, UK), according to the criteria provided by the company that developed the method. A detailed description of the method used and the reaction components and tools used can be found at https://www.lgcgenomics.com.

Determination of copy number variation of VRN-A1

The haploid copy number was determined using the multiplex TaqMan® assay method in collaboration with IDna Genetics Ltd (Norwich Research Park, Norwich, UK).

Studying the gene‐expression levels

For studying the daily patterns of gene expressions of the major plant developmental, circadian and photoreceptor genes, collection of leaf tissues started two weeks after transplanting. In each treatment, the leaf sampling started at Zeitgeber time 1 (ZT 0 = 5:00 am, lights on), 1 h after the start of the light period; it was carried out every three hours, for two consecutive days (48 h), and it was finished with the ZT1 sampling of the 3rd day. Thus, at each time point, two data for each genotype and treatment became available from the two consecutive days, making it possible to include the day as a factor into the variance analyses. In each cases, the last fully expanded leaf was collected. One—one leaf of three plants from three separate pots were pooled for one biological replicate into a 1.5 ml Eppendorf tube that was immediately frozen in liquid N2, with a total of three biological replicates for each genotype (that is, a total of nine leaves per genotype, per sampling time). This approach was intended to minimize potential physiological differences between individual plants. Leaf sample was collected only once from each plant. For nighttime sampling, a low intensity green wavelength LED bulb (Philips-low energy 1W, Philips Electronics Ltd. Amsterdam, Netherlands) was used, which did not interfere with the normal function of the genes under studied. The leaf samples were stored in a – 80 °C freezer until further used.

Total RNA was isolated using the Qiagen RNeasy Plant Mini Kit (Qiagen Ltd, Hilden, Germany) after Trizol extraction, with an extra step of DNase treatment (with Qiagen Rnase-Free Dnase Set, Hilden, Germany—used to remove some remained gDNA contaminations) programmed in the QIAcube equipment (Qiagen Ltd, Hilden, Germany). The cDNA transcription was performed from 1.0 μg of total RNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Inc, Waltham, USA) with the standard protocol provided by the company. Changes in the expression levels were analysed for the vernalization response genes, VRN1, VRN2 and VRN3, for the photoperiod sensitivity gene PPD1, for circadian genes, CCA1, PRR95, TOC1, LUX, ELF3, GI and CO1, and for photoreceptor genes PHYA, PHYB, PHYC, CRY1 and CRY2, using the generic gene specific primers listed in Table S2. As the primers were designed by other research teams (Table S2), we checked the correctness of the primer sequences using a genome database (https://plants.ensembl.org). Furthermore, in all cases, the qRT-PCR melting curve profile and the length of the amplified region were found to be as expected. The primer pairs amplify all homoeoalleles (A, B and D copy) of each studied genes. The quantitative real‐time PCR was carried out in three biological and two technical replicates in a Rotor‐Gene Q equipment (Qiagen Ltd, Hilden, Germany) applying the SYBER‐Green technology of the company. There were 350 PCR reactions, with a total of 25,200 samples analysed with 72 samples per reaction. Gene expression was calculated using the Rotor-Gene software, which also considers the amplification efficiency. Relative concentration was normalized against the geometric mean of three housekeeping genes (Actin, β-tubulin and Ta30797), according to Vandesompele et al.75. The determination of the geometric average of three housekeeping genes and of the relative expression level of the gene of interests were carried out based on the following formulas:

$${\text{GAHKG}} = \sqrt[3]{{\left( {\Delta {\text{CtHKG}}1~ \times ~\Delta {\text{CtHKG}}2 \times \Delta {\text{CtHKG}}3} \right)}}$$

RQ = Ampl^GAHKG-ΔCtTG; GA, Geometric average; Ct, Cycle threshold; HKG: Housekeeping gene; RQ, Relative concentration; TG, Target gene; Ampl, amplification coefficient [corresponds to reaction efficiency of each sample (values of 2 = 100% efficiency)].

Statistical analyses

The variance components were estimated by GenStat 18.0 (VSN International Ltd.) and SPSS 23.0 (IBM Data Science Community, Thomas J. Watson Research Center, Yorktown Heights, NY, USA) program packages using a restricted maximum likelihood method (REML). For the heatmap with the trees of hierarchical clustering, „gplots” and „dendextend” of the R program were used76. In the three different environments, the networks between all the examined genes were visualized with R qgraph 1.9.277 on the 16 time-point data matrices of the three cultivars combined. The correlation heatmaps and the Principal Component Analysis (PCA) were visualized by metan 1.1.0 R package78 and by factoextra 1.0.7 R package (ggplot2)79. The chord diagrams were created using the circlize 0.4.16 R package80.

Results

Phenology in the context of temperature

Based on the population structure analysis36, the allelic composition of the PPD-D1 gene impacted the phenotypic traits, including temperature response. On this basis, Kiss et al.36 distinguished three groups of cultivar response to temperature: photoperiod insensitive (PPD_ins), photoperiod sensitive (PPD_sens) and photoperiod sensitive with strong vernalization requirement (PPD_sens, VRN + +) (Fig. S1). The PPD_ins group was represented by AT1, while the PPD_sens, VRN +  + group by AT3 and AT20. Although the AT3 and AT20 varieties belong to the same temperature reaction group, their genetic distance is significant (Fig. S1). However, based on the gene-specific marker analysis, these selected varieties showed differences only in the copy number of VRN-A1, and in the allele types of PPD-D1, RHT-B1 and RHT-D1 among the main developmental genes that have been tested (Table S1). In the unvernalized treatment, none of the three varieties headed, and there was no significant difference in the developmental stage of their apices at two weeks of age. In vernalized treatments however, their development showed significant variation strongly influenced by temperature. AT1 (photoperiod insensitive) showed a significantly faster growth rate in 18°Cvern and 25°Cvern treatments compared to the two late cultivars (AT3 and AT20), illustrated by the average values of the parameters plant height (PH), apex length (APL), ZD31 and ZD59 (Fig. S2). There was no significant difference in the average heading of AT1 (ZD59) between 18 and 25 °Cvern treatments (Fig. S2). The two photoperiod sensitive cultivars (AT3 and AT20) suffered a significant delay in heading after the 25 °C treatment (106 and 118 days on average for replicates) compared to 18 °C, with the greatest delay in AT3.

Daily expression patterns of the various gene families

Analyses of variance on the complete dataset of three genotypes × three environments × eight daily time points × two days revealed all four main factors being highly significant, but explaining a different portion of the variance that strongly depended on the gene families (Table S3). In general, the circadian genes were mostly influenced by the daily timing, the photoreceptor genes by the environment, and the plant developmental genes by the genotype and the environment. Within the gene families however, there were remarkable variations between the individual genes in the importance of the main factors. The 18°Cuv vs 18°Cvern and 18 °Cvern vs 25 °Cvern subsets were subjected to further analysis of variance to determine the most influencing environmental factor of each gene. These results are hereinafter demonstrated along the separate gene families. Figure S3 demonstrates the daily expression patterns of all the individual genes across the various genotypes and treatments. As the difference between the same time points of the two days was completely negligible compared to the main effects of genotype, environment and daily timing (Table S3), the 8 average values across a 24-h period are presented in these graphs, with the original data from the two days utilised in the determination of error bars and the significance levels (Fig. S3).

Daily expression patterns of the major developmental genes

The expressions of all the three VRN genes were influenced to the largest extent both by the genotype and the environment (Table S3). Within this, VRN1 and VRN3 were more genotype dependent while the activity of VRN2 was more influenced by the environment. Of the environmental factors, VRN1 and VRN3 were mostly determined by the vernalization treatment, but not by the ambient temperature. The activity of VRN2 was, on the other hand, significantly modified both by vernalization and ambient temperature. In the case of PPD1, both environmental factors were the strongest compared to genotype. It is interesting to note that although all the four developmental genes showed daily fluctuations (Fig. S3), this factor explained more than 10% of the variance only in the context of uv – vern treatments for the VRN genes and in association with the genotype. For PPD1, the effect of daily timing, as a main factor, was above 10% in addition to its interaction with genotype.

In unvernalized plants, the expression profile of the flowering enhancing gene, VRN1, was very low, and it did not show any characteristic daily rythm irrespective to the genotype (Table S4, Fig. S3a). After vernalization, a daily rhythm in VRN1 gene expression could be detected, but this was highly specific both to the genotype and the environment. This finding is also well illustrated both by chord diagram and by heatmap of the daily average gene expression of the cultivars in the three treatments as compared to the percentage of average gene activities (Fig. 1a, Fig. S4a). The differences appeared not only in the daily timing of the peak expression, but also in its magnitude and the amplitude of the rhythm. In general, AT1, the photoperiod insensitive, early cultivar showed the highest peak expression followed by AT3, and then AT20, which showed a strong positive coincidence with the heading time (Fig. 1a, Fig. S3a, Fig. S4a).

Fig. 1
figure 1

Chord diagram of the daily average expressions of the plant developmental genes (a), the circadian clock genes (b) and the photoreceptor genes (c) measured in the three treatments (25°Cvern – 25 °C vernalized, 18°Cvern – 18 °C vernalized and 18°Cuv – 18 °C unvernalized) based on the daily average gene activities of the three wheat cultivars. The ratio of values to each other expressed as a percentage. CCA1, CIRCADIAN CLOCK-ASSOCIATED 1; CO1, CONSTANS 1; CRY1, CRYPTOCHROME 1; CRY2, CRYPTOCHROME 2; ELF3, EARLY FLOWERING 3; GI, GIGANTEA; LUX, ARRHYTHMO; PHYA, PHYTOCHROME A; PHYB, PHYTOCHROME B; PHYC, PHYTOCHROME C; PPD1, PHOTOPERIOD1; PRR95, PSEUDORESPONSE REGULATOR 95; TOC1, TIMING OF CAB EXPRESSION1; VRN1, VERNALIZATION1 (APETALA1) ; VRN2, VERNALIZATION2; VRN3, VERNALIZATION3 (FLOWERING LOCUS T).

Full size image

It is interesting to note that only AT1 showed similar magnitude of expression in both vernalized environments with an environmental dependent specific daily rythm. In this cultivar, the peak expression of VRN1 was at ZT4 in 18 °Cvern, which shifted to ZT13 in the 25 °Cvern (Table S4). In the case of the two late heading PPD1 sensitive cultivars, the low VRN1 expression in 25 °Cvern and only one peak at around ZT7 in the 18 °Cvern were the common features while they showed genotype specific differences in the daily rythm in both vern environments (Table S4, Fig. S3a).

Opposing to VRN1 and VRN3, the activity of the flowering repressor gene, VRN2, was the highest in the unvernalized treatment with a distinct daily rythm that was quite similar for all the three cultivars. A significant decrease occurred after the early morning peak ZT1–4 that was followed by a second peak in the late afternoon ZT13–16 environments (Table S4, Fig. S3b). In the vernalized treatments, the VRN2 activity and its daily pattern was more genotype specific. The level of decrease in the daily absolute values coincided with the earliness in all three vern treatments. Thus, they were the lowest in the early AT1 followed by AT3, and they remained the highest in AT20 where they were close to that measured in the unvernalized treatment (Fig. 1a, Fig. S3b, Fig. S4a). The daily fluctuation patterns were also genotype specific with the only exception of 25°Cvern. In the latter case, all three genotypes replicated at various extents the daily pattern detected in the unvernalized treatment.

In the case of VRN3, both the magnitude and the daily patterns were completely genotype specific during the various treatments. Only two tendencies were more general across the genotypes. The gene activity significantly increased due to vernalization, and the level of this increase coincided with the earliness of the genotype (Fig. 1a, Fig. S3c, Fig. S4a).

Environment impacted the expression levels of PPD1 the most. In the unvernalized treatment, all three genotypes showed similar magnitude of the activity, and they were similar in the aspect that the daily fluctuation of this gene started with an early morning peak ZT1, but then the further courses were more genotype specific (Table S4, Fig. S3c). There was one aspect that showed connection with the allele type of PPD-D1, in the insensitive genotype: there was a night peak (detectable in all environments) that was missing from both sensitive genotypes (Fig. S3c). During the 18°Cvern, the PPD1 activity was the highest for all three genotypes, and it was the only treatment where the magnitude of PPD1 activity showed negative coincidence with the earliness (Fig. 1a, Fig. S3d, Fig. S4a). At 18°Cvern, the daily rhythm of the insensitive vs. sensitive genotypes could also be characterised with distinctly different patterns from each other (Fig. S3d). In the case of the insensitive AT1, the highest peak of early morning ZT1 was followed by a smaller peak in early afternoon (ZT13). On the other hand, three peaks of similar magnitude were characteristic to the daily oscillation of PPD1 of the photoperiod sensitive genotypes: an early morning (ZT1), a midday (ZT7), and a late-night peak. These characteristics could not be detected during any of the remaining treatments.

Daily expression patterns of the major circadian genes

The activity of the circadian genes both in the morning (CCA1 and PRR95) and the evening (TOC1, LUX and ELF3) loops were mostly determined by the daily timing (Table S3). CCA1 of the morning loop however, was the only one demonstrating the most stable daily cycling pattern irrespective to the genotypes and environmental conditions. The daily pattern accounted for 62.9–78.7% of its expressional variance in the different environments. It was slightly influenced by the genotype (σ2 = 15.2%***) in the context of unvernalized vs. vernalised treatments and by the temperature (17.6%***) in the context of 18 °C vs 25 °C (Table S3). The activity of PRR95 was more influenced by the environment within which especially the strong effect of ambient temperature became evident in the vernalized treatments. For the genes in the evening loop, the effect of daily timing was more characteristic to TOC1 and LUX. This was, however, modified to a smaller but significant extent by vernalization in the case of TOC1 while by the ambient temperature for LUX. On the other hand, the environmental and genotypic plasticity of ELF3 was its most characteristic feature with significant interactions detectable between all the factors in various combinations. The daily timing patterns of the two central integrator genes were much less characteristic. The activity of CO1 was significantly influenced by genotype and environment (both vernalization and ambient temperature), while the activity of GI was more environment dependent. Of the environment factors, GI was mostly influenced by the ambient temperature but not by the vernalization treatment (Table S3).

The gene expression peaks of circadian genes with characteristic daily rhythms (CCA1, PRR95, TOC1 and LUX) were basically specific, irrespective to the genotype and treatment. Based on our results, the CCA1 gene expression peak was observed in the early morning hours. It was followed by PRR95 around midday, then by the TOC1 gene, together with LUX that were expressed in the late afternoon. The daily pattern of ELF3 from the evening loop was not obvious, but its peak expression generally could be detected during the night hours.

Comparing the individual expression profiles of the three wheat genotypes under the different environments, significant differences could be observed for all the circadian genes, the extent of which was gene specific (Fig. S3e–k). There were significant variations even in the case of CCA1 with the most conserved daily cycling pattern. However, it requires further studies to determine the phenotypic significance of these genotypic and environmental differences. This study highlights only the most interesting features including those that may show any possible associations with the allele phase of PPD-D1 and thus earlier heading (Fig. 1b, Fig. S3e). This phenomenon was not quite characteristic to the morning loop genes but rather to the genes belonging to the evening loop. In the case of TOC1, all the three wheat genotypes showed the same shift in the daily peak due to the vernalization treatment (Fig. S3f.). In unvernalized plants, the TOC1 expression peaked at ZT10, which shifted to the later hours ZT13–16 in the vernalized plants irrespective to the ambient temperature level. There was also an environmental dependent peak shift in the activity of LUX, but in this case, this shift did not quite distinguish the unvernalized treatment from the vernalized but rather the genotypes, and it coincided with the PPD-D1 allelic phase (Fig. S3g). The peak shift in LUX activity of the PPD-D1 insensitive genotype (AT1) was similar to that of TOC1. In the cases of the two PPD-D1 sensitive genotypes (AT3 and AT20) however, LUX peaked in the unvernalized plants at the end of the light period, which moved earlier to ZT13 in 25 °C treatment and to ZT10 in the 18 °C treatment. In the case of ELF3, there were again remarkable differences between the insensitive and sensitive wheat cultivars (Fig. S3i). The amplitude of ELF3 daily expression was narrow in the PPD-D1 sensitive genotypes, with one or more smaller peaks depending on the environment. In contrast, a distinct night peak was characteristic to the insensitive cultivar AT1 in two of the three environments. The only exception was the 25 °C treatment where all the three genotypes similarly demonstrated the flattest expression curve of ELF3. It is also interesting to note that the three genotypes behaved similarly in the 18°Cuv treatment where in addition to the night peak, a second peak was apparent in the middle of daylight (ZT7–10).

The central circadian genes (GI and CO1) were characterised with mostly genotype and environment specific patterns. The expression profile of GI was more or less similar in the unvernalized plants of all the three cultivars with its maximum expression occurring at ZT13; there was only a slight difference between the genotypes in the maximum value (Fig. S3j). The behaviour of the cultivars differed most in the 18 °C vernalized treatment where the expression level in itself and compared to the non-vernalized level negatively coincided with the heading date. Vernalization decreased the activity of GI to the largest extent in AT1, the early heading cultivar. It was similar in magnitude but with a peak shift between the unvernalized and vernalized treatments for AT3, while it was the highest in the vernalized plants of AT20 with an accentuated peak at ZT13. This phenomenon however, was not apparent under the 25 °C treatments. As for CO1, its expression level was low in the unvernalized plants of the insensitive AT1 compared to the sensitive AT3 and AT20 (Fig. S3k). In addition, the largest difference between the genotypes became apparent under the 25 °C treatment. While in AT3, CO1 showed similar activities in all the three vernalized treatments, AT1 was characterised with a significant night peak apparent only at 25 °C. However, it was AT20 that showed the most distinct and highest CO1 activity, especially at 25 °C that was strongly elevated in late afternoon and after a short decrease before the end of the light period returned to that elevated level during the night.

Daily expression patterns of the major photoreceptor genes

The expressions of all photoreceptor genes were influenced both by the genotype, environment, and daily timing (Table S3). Within this, PHYA, PHYB and CRY1 were more environment dependent, while the activities of PHYC and CRY2 were more influenced by the genotype, environment, and daily timing. In the case of PHYA, PHYB and CRY1, both environmental factors were the strongest compared to genotype. Of the environmental factors, PHYC and CRY2 were mostly determined by the vernalization treatment but not by the ambient temperature. For these two genes, the effect of daily timing as a main factor was above 15%.

In the case of PHYA and PHYB, a significant high level of daily peak expression was shown in the 18°Cvern treatment (Table S4, Fig. S3l–m, Fig. S4c). This phenomenon was nearly tenfold higher than in 25°Cvern treatment (Fig. 1c, Fig. S3l–m, Fig. S4c). Furthermore, the appearance of daily peak expression of these genes was different in the three varieties. In the case of photoperiod insensitive variety (AT1), the peak expression was at 24:00 (ZT22) (Table S4, Fig. S3l–m). In the photoperiod sensitive varieties, two peaks were visible at ZT1 and at ZT10; for AT3 there were three peaks at ZT1 0, at ZT7–13, and at ZT22 for AT20 (Table S4, Fig. S4c). In vernalized treatments, PHYC showed a maximum activity at night in 25°Cvern for all three cultivars (Fig. S3n, Fig. S4c). A daily rythm in CRY1 and CRY2 genes expression was detected, but this was highly specific both to the genotype and the environment (Table S3). The differences appeared not only during the daily timing of the peak expression but also in its magnitude of the rhythm. The daily peak expression of CRY1 was stronger in late heading varieties (AT3 and AT20) at 18 °C (regardless of the cold treatment) (Fig. S3o, Fig. S4c). Furthermore, a difference in the daily expression pattern was also observed between the three varieties. Under 25°Cvern condition, the transcriptional maximum of both genes was observed in the early morning (ZT1–4) and at night (ZT22) (Table S4, Fig. S3o).

Interactions of flowering pathway genes in hexaploid wheat

The network system of genes studied was significantly altered by the treatments. The fewest gene interactions were observed in 18°Cvern and 25°Cvern, while the most were visualized in 18°Cuv treatment (Figs. 2, 3, 4). In each treatment, the correlations between the expression levels of the 16 genes tested across the three wheat cultivars were also evaluated by principal component analysis (Figs. 2b, 3b, 4b).

Fig. 2
figure 2

Gene network systems (a), PCA analysis (b) and correlation matrix (c) carried out on the data matrix of daily gene expression levels of 3 wheat cultivars in 18°Cvern (18 °C vernalized) treatment. The PCA analysis was presented as a supplementary, even more relevant illustration of the correlations between the genes investigated. CCA1, CIRCADIAN CLOCK-ASSOCIATED 1; CO1, CONSTANS 1; CRY1, CRYPTOCHROME 1; CRY2, CRYPTOCHROME 2; ELF3, EARLY FLOWERING 3; GI, GIGANTEA; LUX, ARRHYTHMO; PHYA, PHYTOCHROME A; PHYB, PHYTOCHROME B; PHYC, PHYTOCHROME C; PPD1, PHOTOPERIOD1; PRR95, PSEUDORESPONSE REGULATOR 95; TOC1, TIMING OF CAB EXPRESSION1; VRN1, VERNALIZATION1 (APETALA1) ; VRN2, VERNALIZATION2; VRN3, VERNALIZATION3 (FLOWERING LOCUS T).

Full size image
Fig. 3
figure 3

Gene network systems (a), PCA analysis (b) and correlation matrix (c) carried out on the data matrix of daily gene expression levels of 3 wheat cultivars in 25°Cvern (25 °C vernalized) treatment. The PCA analysis was presented as a supplementary, even more relevant illustration of the correlations between the genes investigated. CCA1, CIRCADIAN CLOCK-ASSOCIATED 1; CO1, CONSTANS 1; CRY1, CRYPTOCHROME 1; CRY2, CRYPTOCHROME 2; ELF3, EARLY FLOWERING 3; GI, GIGANTEA; LUX, ARRHYTHMO; PHYA, PHYTOCHROME A; PHYB, PHYTOCHROME B; PHYC, PHYTOCHROME C; PPD1, PHOTOPERIOD1; PRR95, PSEUDORESPONSE REGULATOR 95; TOC1, TIMING OF CAB EXPRESSION1; VRN1, VERNALIZATION1 (APETALA1) ; VRN2, VERNALIZATION2; VRN3, VERNALIZATION3 (FLOWERING LOCUS T).

Full size image
Fig. 4
figure 4

Gene network systems (a), PCA analysis (b) and correlation matrix (c) carried out on the data matrix of daily gene expression levels of 3 wheat cultivars in 18°Cuv (18 °C unvernalized) treatment. The PCA analysis was presented as a supplementary, even more relevant illustration of the correlations between the genes investigated. CCA1, CIRCADIAN CLOCK-ASSOCIATED 1; CO1, CONSTANS 1; CRY1, CRYPTOCHROME 1; CRY2, CRYPTOCHROME 2; ELF3, EARLY FLOWERING 3; GI, GIGANTEA; LUX, ARRHYTHMO; PHYA, PHYTOCHROME A; PHYB, PHYTOCHROME B; PHYC, PHYTOCHROME C; PPD1, PHOTOPERIOD1; PRR95, PSEUDORESPONSE REGULATOR 95; TOC1, TIMING OF CAB EXPRESSION1; VRN1, VERNALIZATION1 (APETALA1) ; VRN2, VERNALIZATION2; VRN3, VERNALIZATION3 (FLOWERING LOCUS T).

Full size image

The results confirmed the efficiency of the method used to visualize gene interaction (qgraph). The relationship between the three main VRN genes differed significantly between treatments, but the positive connection between VRN1 and VRN3 genes was maintained at a significant level, being the strongest in 25°Cvern treatment (Fig. 2). A strong positive association between VRN1 and CCA1 was observed in the 18°Cvern, but this phenomenon was not visualized in all other treatments. It is interesting to note that after vernalization, a negative association between VRN2 and the other two VRN genes (VRN1 and VRN3) was observed, which became more expressive at 18 °C. In this treatment, the positive association between VRN2 and PPD1 genes became stronger irrespective of vernalization, but this phenomenon was not observed in 25°Cvern treatment. A similar trend was found between VRN2 vs GI and VRN2 vs CO1 genes. A strong positive correlation was found between VRN3 and GI only in 18°Cvern treatment. The interaction between PPD1 vs. PHYB and PPD1 vs CCA1 became more effective in 18°Cuv treatment. The interaction between PPD1 and VRN3 was expressed in most of the treatments with the only exception of 18°Cvern.

Three circadian genes (LUX, TOC1 and CCA1) formed an internal network system in all treatments (Fig. 2, 3, 4). A positive correlation at a significant level detected between TOC1 and LUX, while a negative relation at a significant level was observed to TOC1 vs CCA1 and to CCA1 vs LUX. A strong positive correlation was observed between GI and LUX in all three environments. Furthermore, a strong positive correlation between the GI vs LUX vs CO1 gene group was observed at 18 °C treatment, irrespective of vernalization. The interaction between CO1 and LUX genes became more effective in 18°Cuv treatment. A positive correlation was described between GI and CO1, which became more significant at 18 °C, irrespective of vernalization treatment. A negative correlation was found between ELF3 vs GI, ELF3 vs CO1 and ELF3 vs VRN2. This finding was however only observed during 18°Cuv treatment. Based on our results, during 25°Cvern treatment, CO1 showed a negative association with VRN3.

A positive correlation between one photoreceptor gene (PHYC) and one of EC (ELF3) was observed in all treatments with the strongest correlation in 25°Cvern treatment (Fig. 2, 3, 4). At 18 °C, the association between PHYA and PHYB genes became stronger irrespective of vernalization. Direct linkage of PHYA to PHYB and PHYC was abolished during 25°Cvern treatment. The relationships between the three phytochrome genes (PHYA, PHYB and PHYC) and PPD1 were significantly altered by the treatments. While a strong positive association was observed between these genes during 18°Cuv, only PHYB in 25°Cvern and none of the phytochromes in 18°Cvern showed any correlation with the PPD1 gene. A significant negative correlation was observed between CRY2 vs GI and CRY2 vs LUX in 25°Cvern and 18°Cuv treatment. A similar tendency was observed between CRY1 and PRR95. A positive correlation is observed between CRY1 and PHYC. This relationship was weakest during the 18°Cuv treatment.

Discussion

The regulatory mechanism of the plant circadian rhythm has been extensively explored in Arabidopsis, and several homologous genes have already been described in wheat, but few of these have been studied thoroughly. Much of this knowledge has been originated from diploid Triticum monococcum and tetraploid species with smaller genome sizes, or from specific crossing lines (RIL, NIL, mutant and transgenic lines). In contrast to most studies, our aim was not to determine the effect of a single gene on ambient temperature using specific lines (RIL or NIL). The focus of this study was to describe the effect of different ambient temperature levels on the diurnal rhythm of expression of the main developmental (VRN1, VRN2, VRN3 and PPD1), circadian (CCA1, PRR95, TOC1, LUX, ELF3, GI and CO1) and photoreceptor (PHYA, PHYB, PHYC, CRY1 and CRY2) genes in juvenile plants and to investigate the interrelationships between them. For this purpose, three hexaploid wheat cultivars were included with different genetic backgrounds, based on the previous results36. However, we would like to emphasize, that the conclusions from the correlation matrix we presented do not imply necessarily the existence of functional relationships between the genes investigated. They give additional information on the interconnectedness of the genes in the complex genetic background of the hexaploid wheat genome and demonstrate how the various environmental factors – vernalization saturation and the ambient temperature in the current research – may influence this network. In the future, it will be essential to examine the homoeolog-specific regulation of the VRN and PPD genes using a broader range of varieties, and the studies should also be extended to description of specific effects occurring in later developmental stages.

The results identify that temperature significantly influenced the daily expression patterns of main developmental, circadian, and photoreceptor genes in hexaploid wheat. In the case of VRN1 and VRN3, both the magnitude and the daily patterns were completely genotype and environment specific. The activity of these two genes significantly increased due to vernalization, and the level of this increased expression coincided with the earliness of the genotype. The positive relationship between VRN1 and VRN3 genes remained significant in all treatments (Fig. 5), which confirms the finding that the close relationship between these two genes is essential for the regulation of flowering17,81,82.

Fig. 5
figure 5

Summary of the major gene linkage systems (at least P ≤ 0.01) identified via comparing the daily gene expression patterns in three wheat cultivars. Boxes represent genes. Arrows indicate the positive (promotion—green) or negative (inhibition—red) associations between gene expression patterns. Arabidopsis orthologues of some key flowering genes in wheat are shown in parentheses. 18°Cvern/18 °C vernalized/, 18°Cuv/18 °C unvernalized/ and 25°Cvern/25 °C vernalized/. The presence of linkages between genes does not necessarily imply functional relationships among the investigated genes. CCA1, CIRCADIAN CLOCK-ASSOCIATED 1; CO1, CONSTANS 1; CRY1, CRYPTOCHROME 1; CRY2, CRYPTOCHROME 2; ELF3, EARLY FLOWERING 3; GI, GIGANTEA; LUX, ARRHYTHMO; PHYA, PHYTOCHROME A; PHYB, PHYTOCHROME B; PHYC, PHYTOCHROME C; PPD1, PHOTOPERIOD1; PRR95, PSEUDORESPONSE REGULATOR 95; TOC1, TIMING OF CAB EXPRESSION1; VRN1, VERNALIZATION1 (APETALA1) ; VRN2, VERNALIZATION2; VRN3, VERNALIZATION3 (FLOWERING LOCUS T).

Full size image

It has been demonstrated that during vernalization, cold gradually activates the VRN1 gene, which also stimulates the VRN3. This gene feeds back to the VRN1 to further increase its activity, thus inducing heading. Based on our results, the expression of VRN1 and VRN3 in the early cultivar (AT1) was 3–tenfold higher than in the late cultivars (AT3 and AT20), independent of temperature treatment, which showed a strong positive coincidence with the heading time (Fig. S2). The vernalized VRN2 gene showed an explicit diurnal activity in late heading cultivars. This effect became most significant at 18 °C and 25 °C for the AT20 causing late heading similarly to the unvernalized treatment. Furthermore, a strong positive association between VRN2 and CO1 was observed at 18 °C, irrespective vernalization (Fig. 5). Mulki and von Korff83 described that HvVRN2 was up-regulated by HvCO in barley. It can be assumed that VRN2 is an integrator between light and vernalization pathway in temperate cereals51, and this gene may still play an important role in the regulation of plant development even after vernalization extending its function already reported by Yan et al.4, Fjellheim et al.84 and Kiss et al.36. The diurnal rhythm of PPD1 gene was also significantly affected by genotype and temperature, which can be the consequence of using generic primers amplifying PPD-A1, PPD-B1 and PPD-D1 at the same time. These three genes participate directly in determining photoperiod sensitivity and they differ not only in the causal polymorphisms but also in the daily cycling pattern. The gene expression level of the photoperiod sensitive allele was defined as very low in the early morning, peaking in the morning, and declining again22,85,86. In contrast, photoperiod insensitive alleles do not show this diurnal fluctuation, and the gene activity level is consistently elevated11,22,26,29,85,86. The results of this study also support these findings in most treatments with the only exception of 18°Cvern. Here, the PPD1 activity was the highest for all three genotypes, and it was the only treatment where the magnitude of PPD1 activity showed negative coincidence with the earliness. This study also found a significant change in the strength and number of PPD1 vs. circadian gene interactions. This phenomenon may correspond to the finding that the PPD1 gene is closely related to the Arabidopsis PRR7 gene. This gene also plays a role in the perception of light and temperature in the circadian rhythm, so the PPD1 gene may have a similar role in cereals36,87. Furthermore, a strong positive correlation between PPD1 and VRN3 gene was observed in almost all treatments (excluding 18°Cvern) (Fig. 5), supporting previous studies26,29. Therefore, the wheat VRN3 gene can be a central gene in the vernalization and photoperiod regulatory pathways that integrates different environmental signals28.

The activity of both morning (CCA1 and PRR95) and afternoon (TOC1, LUX and ELF3) loop circadian genes were mostly determined by daily timing, irrespective of vernalization and temperature treatments. The central genes involved in daily circadian rhythm regulation, CO1 were significantly influenced by genotype, while the activity of GI was more environment dependent. Based on the previous experiments50,51,88,89,90,91,92,93,94 with Arabidopsis, the expression peak of CCA1 and PRR95 genes are observed in the early morning hours, followed by TOC1 around in late afternoon. Finally, the transcription peak of EC genes (LUX, ELF3 and ELF4) was detected at night. In the morning, the CCA1 and LHY genes inhibit the expression of TOC1 and EC genes respectively, then in the late afternoon, the TOC1 repress the transcription of the CCA1 and LHY genes (negative feedback loop) and also impacts its own function, reducing its own activity. The LUX, ELF3 and ELF4 genes inhibit the PRR5, PRR7, PRR9, TOC1 and GI genes, resulting in a decrease in the inhibition of CCA1 and LHY genes, thus initiating a new diurnal cycle. This study identified similar functions: CCA1 expression being the highest at early morning, while TOC1 expression at evening, and a significant negative association between them irrespective to the environment. The homologs of CCA1/LHY in temperate grasses are responsive to photoperiod and temperature regulation, but their effects on flowering time remain unknown47,60. This study found that, indeed, the level of expression of CCA1 was significantly influenced by the temperature and genotype. For AT1 (early heading type), it was 18 ºC highest throughout of the day, while for both AT3 and AT20 (late heading types) the morning peak was the highest at 25 ºC. The TOC1 gene was also identified in wheat, which has been characterized to have an evening expression peak95, matching the diurnal functional pattern of its Arabidopsis homologue gene96. However, the transcriptional mechanism and gene interaction of this gene are not yet known97. The PRR95 expression being the highest at the middle of the day while LUX and ELF3 expression at evening and night. This study found a significant negative association between CCA1 and LUX while a significant positive correlation between TOC1 and LUX, irrespective to the environment (Fig. 5). The close, temperature-independent relationship between these major circadian genes may also illustrate their fundamental role in the floral regulatory system even for varieties with different genetic backgrounds. In Arabidopsis (long day plant), the circadian clock regulates the induction of FT and its critical transcriptional inducer CO via an external coincidence mechanism in which light coincides with an inductive window that is gated by the circadian clock1. In this process, the clock component GI has an important role in forming a transcriptional complex with another factor (FLAVIN-BINDING KELCH REPEAT F-BOX 1 /FKF1/) via cycling changes in their protein abundance. In long days, the phase of peak GI accumulation coincides with that of FKF1 before dusk; thus, their complex is able to degrade the CYCLING DOF FACTOR 1 (CDF1) repressor of CO in late afternoon relieving the transcriptional repression of CO, leading to the accumulation of stable CO protein, which is able to activate FT. The positive effect of HvCO1 on HvFT1 expression has been described in barley98, but little information is available in wheat. In accordance with these schemes, this study did not find strong positive relationship between the CO1 and VRN3 (TaFT) genes. Furthermore, in the long-day experiment, GI expression peaked during the late afternoon in most of the cases mostly irrespective to the environments. At 18 °C (including 18°Cuv), the activity of this gene increased 3–fivefold compared to the 25 °C experiment, irrespective of genotype. A similar tendency was observed for the PRR95. In contrast, in barley, Ford et al.47 described, that increased expression of these genes was induced by increasing temperature (from 15 to 25 °C). The activation peak of CO1 however, significantly varied with the genotypes and treatments. In addition, a direct association between these two genes was only found in 18°Cuv and 18°Cvern treatments, but not in 25 °C environment (Fig. 5). Zhao et al.99 also described the positive correlation between the wheat GI and CO genes. Furthermore, a strong positive correlation was observed between GI and LUX gene irrespective of the environment (Fig. 5). A negative interaction between wheat ELF3 and GI, consistent with the function of the Arabidopsis homolog gene has been described55,58. Based on the results, this phenomenon was only confirmed in 18°Cuv treatment (Fig. 5). Furthermore, a similar tendency was also observed between ELF3 vs CO1 and ELF3 vs VRN2, which was eliminated by the cold treatment.

PHYA, PHYB and CRY1 genes were strongly affected by the environment, while PHYC and CRY2 genes were significantly influenced mainly by genotype, environment, and daily timing. PHYA and PHYB genes showed a nearly tenfold increase in average daily gene activity in 18°Cvern compared to 25°Cvern treatment. It has been described that the PHYB photoreceptor in Arabidopsis also functions as a temperature transmitter. However, this has not yet been tested in temperate cereals64,65. Based on the correlation data matrices, in bread wheat, the PHYA may play a similar role to PHYB. Furthermore, a strong positive correlation of PHYA was identified with VRN3 (in 18°Cvern and 18°Cuv) and VRN2 (in 25°Cvern and 18°Cuv) (Fig. 5). In their work on mutant tetraploid wheat, Alvarez et al.100 described the important role of light in the interaction of ELF3, PHYB and PHYC having a direct effect on PPD1 gene. Woods et al.107 demonstrated, that ELF3 express downstream from PHYC and acts as a repressor of PPD1 in the photoperiod flowering pathway of Brachypodium distachyon. This study demonstrates a significant level of positive correlation between PHYC and ELF3 that was observed in all treatments (Fig. 5). A strong positive association was also observed between the three phytochrome genes (PHYA, PHYB and PHYC) and PPD1 in 18°Cuv treatment (Fig. 5). In wheat, a positive co-expression role of PHYC and PHYB in flowering has already been described, but no information was available on the PHYA57,70,72. The results of this study highlighted the fact that not only PHYB and PHYC showed a strong association with plant development genes but also PHYA (Fig. 5), probably in a redundant manner. A critical element of the photoperiod pathway is the light-mediated stabilization of the CO protein, in which PHYA, CRY1 and CRY2 and also PRRs play important roles in Arabidopsis when abundantly present101. In this experiment, positive association between CO and CRY1 was detected in 25°Cvern only; between CO and PHYA in 25°Cvern, but no other connections reached the significance level. The relationship between the VRN1 vs VRN3, CCA1 vs TOC1 vs LUX vs GI and PHYC vs ELF3 genes play a fundamental role in the vegetative-generative transition of apex showing a strong correlation at a significant level irrespective of ambient temperature and vernalization treatment (Fig. 5).

This study highlights the complex relationships between gene families determining early development in hexaploid wheat cultivars, which were significantly influenced by different ambient temperatures. The genetic plasticity present in varieties in response to different environmental cues can be an important factor in the ecological adaptability. Therefore, a better understanding of the mechanisms of action of environmental factors can help to produce new varieties that can adapt to future changing environmental conditions.

Conclusions

The molecular genetic regulation of the transition between the vegetative and generative stages in the model plant Arabidopsis has been most extensively studied. In the contrast, the molecular genetic processes of plant development in wheat are much less well understood, and only the major components of the vernalization and photoperiod regulation pathways have been identified in detail. Studying the genes that determine flowering in Arabidopsis may also provide a good basis for investigating orthologous genes in wheat. In hexaploid wheat varieties with different developmental patterns we found a positive relationship between VRN1 and VRN3 genes irrespective of ambient temperature treatments. The vernalized VRN2 gene showed an explicit diurnal activity in late heading cultivars and this gene may still play an important role in the regulation of plant development after cold treatment. The diurnal rhythm of PPD1 gene was significantly affected by temperature. At 18 °C, the vernalized PPD1 activity showed negative coincidence with the earliness. The close temperature-independent relationship between the investigated major circadian genes (CCA1, PRR95, TOC1, LUX, ELF3, GI and CO1) may also illustrate their fundamental role in the floral regulatory system. PHYA, PHYB and CRY1 genes were strongly affected by environment, while PHYC and CRY2 genes were significantly influenced mainly by genotype, environment and daily timing. This study is one of the few works which investigate the diurnal expression patterns of key developmental, circadian, and photoreceptor genes in the diverse genetic background of bread wheat cultivars and under different ambient temperatures. This might open the floor for further research in this field which, in turn, could provide more detailed insights into the genetic regulation of cereal plant development across a broad genetic background.