Interaction between photoperiod and variation in circadian rhythms in tomato

Abstract

Background: Many biological processes follow circadian rhythmicity and are controlled by the circadian clock. Predictable environmental changes such as seasonal variation in photoperiod can modulate circadian rhythms, allowing organisms to adjust the timing of their biological processes to the time of the year. In some crops such as rice, barley or soybean, mutations in circadian clock genes have altered photoperiod sensitivity, enhancing their cultivability in specific seasons and latitudes. However, how changes in circadian rhythms interact with the perception of photoperiod in crops remain poorly studied. In tomato, the appearance during domestication of mutations in EMPFINDLICHER IM DUNKELROTEN LICHT 1 (EID1, Solyc09g075080) and NIGHT LIGHT-INDUCIBLE AND CLOCK-REGULATED GENE 2 (LNK2, Solyc01g068560) delayed both the phase and period of its circadian rhythms. The fact that variation in period and phase are separated in tomato provides an optimal tool to study how these factors affect the perception of photoperiod.

Results: Here we develop tomato near isogenic lines carrying combinations of wild alleles of EID1 and LNK2 and show that they recreate the changes in phase and period that occurred during its domestication. We perform transcriptomic profiling of these near isogenic lines under two different photoperiods, and observe that EID1, but not LNK2, has a large effect on how the tomato transcriptome responds to photoperiod. This large effect of EID1 is likely a consequence of the global phase shift elicited by this gene in tomato's circadian rhythms.

Conclusions: Our study shows that changes in phase that occurred during tomato domestication determine photoperiod perception in this species, while changes in period have little effect.

Keywords: Circadian rhythms; Photoperiod; RNA-seq; Tomato domestication.

Fig. 1

Circadian period (a) and relative phase (b) estimates in the near isogenic lines segregating for wild alleles of EID1 and LNK2. Data comes from three independent experiments shown in Figure S2. Relative phase values were obtained by subtracting each phase value from the average of MM in each experiment. MM stands for S. lycopersicum cv. MoneyMaker. Spim stands for S. pimpinellifolium. Different letters in each boxplot indicate significant differences (p < 0.05, one-way ANOVA and Tukey’s post hoc HSD test)

Fig. 2

Differentially expressed genes in response to wild alleles of LNK2 and EID1 in each photoperiod separately. A Horizontal bars represent the number of differentially expressed transcripts in each genotype and condition. Vertical bars represent the number of differentially expressed genes in each set, with a set being a combination of comparisons. Only sets with more than 20 genes are shown. B and C Phylogenetic tree with protein sequences from Arabidopsis and tomato for the RVE family (B) and the PRR family (C) of circadian clock genes. Arabidopsis and tomato protein names are highlighted in gray and red respectively. Tomato proteins whose transcripts oscillate during the diel cycle are marked with an orange dot. D Log2 fold change in expression (± standard error) between S. lycopersicum var MoneyMaker (MM) and each of the near isogenic lines in this work. SD and LD stand for short days and long days. Color of the bars indicate the significance of each log2 fold change, with non-significant values in gray, q < 0.05 in blue and q < 0.01 in black

Fig. 3

A Differentially expressed transcripts in response to photoperiod in each genotype separately. Horizontal bars represent the total number of differentially expressed transcripts in each genotype. Vertical bars represent the number of differentially expressed genes in each set, with a set being a combination of comparisons. B Number of cycling transcripts whose expression is affected by changes in photoperiod grouped by their phase. C Average expression of transcripts whose expression oscillated during the day grouped by their phase, genotype and condition. D Differences between the estimated phase in long days and short days. Error bars represent the standard deviation of the mean. Different letters in each dot indicate significant differences (p < 0.05, using two-way ANOVA and estimated marginal means). E Average log2 fold change in expression induced by photoperiod in cycling genes grouped by genotype and phase of expression. Error bars represent the standard error of the mean