S1: gene expression in all DNA binding proteins, auxin, and the circadian clock (DONE)
S2: the scatterplots in elf3-1 and phyABCDE in each of the GO term categories (DONE)
Table S1: dawn peak genes within each GO category (DONE)
---
There are multiple waves of gene expression after light exposure at dawn
While many regulatory genes have peak expression at the onset of light, we do not know the precise timing of this burst, so a high-resolution time series experiment was performed at the onset of light, with 25 time points sampled at 22oC (Table S2).
To visualise this data, we clustered the gene expression values from the 22oC time series, focussing on genes with GO terms associated with DNA binding-- see Methods and Table S3. The dawn burst consists of multiple coordinated waves of gene expression of genes associated with DNA binding. While there are genes whose expression levels decrease (clusters 1-2) or increase (clusters 3-4) over the time course, there are clearly three distinct bursts of expression at 16-24 minutes (clusters 7-8), 18-45 minutes (clusters 9-10), and 45-105 minutes (clusters 5-6), see Figure 2A.
These clusters contain many genes that are relevant to light signalling, temperature response, and immunity. Clusters 7-8 include many genes that are related to temperature, such as HEAT SHOCK FACTOR A1A (HSFA1A) and immunity genes such as the WRKY transcription factor WRKY33 and BASIC LEUCINE ZIPPER 10 (bZIP10). This is consistent with other research that suggests that plants may be less sensitive to certain types of infections at dawn \cite{Ingle2015,Wang2011}. Interestingly, clusters 5-6 include HEAT SHOCK FACTOR B2B (HSFB2B) that suppresses the heat shock response and inhibits HSFA1A \cite{Ikeda_2011}. The clusters of genes that are expressed at 45-105 minutes include genes that appear to be downstream of red (RVE7, HY5, JMJ22, PNT1) and blue (MYC2, MYC4, CIB2, CRY3) light signalling.
The dawn gene expression waves are sensitive to temperature and light
Since these clusters contained many genes associated with temperature and light signalling pathways, we investigated the effect of temperature and light perturbations on these expression waves. We repeated the high resolution RNA-seq experiment at an elevated but ambient temperature of 27oC (Table S2), and observed that genes continue to peak at the same time points, but the earlier expressed genes (clusters 1-2 and 7-8) have elevated expression and the later expressed genes (clusters 3-6) have lower expression than at 22oC, see Figure 2A. Please recall that the clusters were determined using the 22oC time series only.
Previously, \citealt{Rugnone_2013} found sets of genes that were induced or repressed by light, and moreover they identified genes whose sensitivity to light was dependent on whether the light treatment occurred at night or during the subjective day (after an extended night). We compare their gene lists with ours in Figure 2B, also see Table S4. There was significant enrichment for light repressed genes among early expressing genes in clusters 1-2 (p-value < 1e-14 for genes that were repressed by light either during the day or in a time-neutral manner, based on Fisher exact test with Bonferroni correction). In contrast, those genes that were expressed late in the time series or that peaked at 45-105 minutes or 18-45 minutes tended to be light induced. Previously, we noticed that there were a number of key light signalling genes in clusters 5-6 and we find that over a third of genes within this group are light induced. Intriguingly, the clusters of genes that includes HSFA1A (clusters 7-8) are not enriched for light induced genes, except for a slight enrichment for genes whose expression is induced by light at night (p-value <0.037).
These results suggests that there are multiple coordinated bursts of gene expression at dawn that are differentially regulated by temperature and light, and moreover many of the dawn expressed genes are regulated by light in a time-of-day dependent manner.