Introduction

The early morning and late evening are critical time periods for Arabidopsis.  The circadian clock is entrained by changes in light  (Kinmonth-Schultz, Golembeski, and Imaizumi 2013; Covington et al. 2001; Edwards et al. 2010; Seo and Mas 2014), temperature (Michael, Salome, and McClung 2003; Gould 2006; McClung and Davis 2010; Salome, Weigel, and McClung 2010; Mizuno et al. 2014) and even humidity (Mwimba et al. 2018) that occur at dawn and dusk.    The evidence that these transition periods are important for entrainment comes from experiments that show that light or temperature pulses in different times of day cause phase shifts in the circadian clock (Covington et al. 2001; Michael, Salome, and McClung 2003).   However, dawn and dusk transitions are not exclusively responsible for entrainment– photosynthesis and sugar production can also gate the circadian clock, producing a second ‘metabolic dawn’ (Haydon et al. 2013).
There are several proposed mechanisms for circadian entrainment by light and temperature.  Red- and blue-light sensors– including phytochromeB (phyB) and cryptochrome1 (cry1)– may be involved in entrainment (Hall et al. 2002; Salomé et al. 2002; Devlin and Kay 2000) via interactions with ZEITLUPE (ZTL),  Flavin‐binding‐Kelch‐F‐box (FKF) and LOV‐Kelch protein 2 (LKP2) (Zoltowski and Imaizumi 2014; Somers, Kim, and Geng 2004; Kim 2005).  While phyB has also been implicated as a temperature sensor at night (Legris et al. 2016; Jung et al. 2016) and interacts with the Evening Complex (EC) that is part of the evening loop of the circadian clock (Ezer et al. 2017; Huang et al. 2016), it is unclear whether phyB plays a role in temperature entrainment of the circadian clock.  However, there is evidence that heat shock protein 90 (HSP90) may be responsible for temperature entrainment (Davis et al. 2018).
Environmental signals entrain the circadian clock, while at the same time the circadian clock controls a plant’s sensitivity to environmental stimuli (Grundy, Stoker, and Carré 2015; Cortijo et al. 2018).  This circadian pattern of response to abiotic and biotic factors enables the plant to be most responsive to stimuli when they are most likely to be present.  For instance, there are a larger number of genes that change their expression levels in response to light treatment in the middle of the subjective day than at night (Rugnone et al. 2013).  Specifically in the early morning, many researchers have observed bursts of gene expression in drought-response genes (Grundy, Stoker, and Carré 2015), temperature response genes such as heat shock factor 70 (HSP70) (Dickinson et al. 2018), and phytohormone genes (Michael et al. 2008).  In the latter case, (Michael et al. 2008) observed that dawn expressed phytohormone genes had a G-box motif in their promoters, and indeed there is a large set of genes with G-box promoter motifs that are expressed within one hour of dawn including many that are involved in response to metals (Ezer et al. 2017).    Moreover, Arabidopsis is less susceptible to certain fungal pathogens in the morning, linked to jasmonic acid signalling (Ingle et al. 2015).
Most of these studies of biotic and abiotic responses in the early morning were performed in artificial growth chambers, which do not have realistic diurnal light and temperature regimes for natural environments (Annunziata et al. 2018).   Therefore, these studies allow us to observe how the transcriptome responds to light stimulus in the morning, but not how it responds to a realistic dawn .  However, these kinds of experiments are critical for understanding the mechanisms that control light response in the morning, because we can learn the exact time delay between the environmental perturbation (light) and the transcriptional response.  Moreover, farming in artificial lights is becoming common, and it is important to understand how crops respond to lighting conditions that are reminiscent of growth chambers (Ibaraki 2016; Gupta and Agarwal 2017; Olvera-Gonzalez et al. 2013).
Dawn is an important time for plants, but the gene expression dynamics in this time period are poorly understood, because gene expression changes so rapidly.  The first main aims of this paper is to characterise with more detail than ever before the early morning transcriptome of Arabidopsis.  Although there have been a number of studies that demonstrate that there is a burst of gene expression after light stimulus in the morning, the dynamics of this burst have not been fully characterised because time points were not sampled frequently enough.  We uncover five coordinated waves of gene expression that are both light- and temperature-sensitive, and then infer a regulatory network that explains the observed expression patterns.
The second aim is to investigate how genes that are known to play a role in light and temperature entrainment behave in response to light in the early morning.  In particular, we are interested in the kinetics of the light response in the early morning and how this is perturbed under light sensing and arhythmic mutants.  We focus on _____
However, the mechanisms that control this burst of expression are poorly understood.  For instance, it is unclear whether this burst is driven by the circadian clock or light response, and it is unclear whether it is related to entrainment. 
Through a high resolution RNA-seq time series, we find five distinct transcriptional waves within the first two hours of the morning, among genes that encode transcription factors and DNA binding proteins.  We characterise how each wave of expression responds to temperature elevation and light signals during the night and subjective day, and how these waves are affected by a light signalling mutant (phyAphyBcry1cry2 ), a circadian clock mutant (prr5prr7prr9 ), and a temperature response mutant (hsfa1Q ).  Furthermore, we infer a gene regulatory network and validate a number of edges using DNA binding data.  Although phyAphyBcry1cry2  and prr5prr7prr9  have a similar global impact on the transcription of DNA binding proteins in response to light in the morning, these mutants have opposite affects on a number of specific temperature and ABA-sensing sub-networks.  This work provides unprecedented detail as to how light, temperature, and circadian genes are regulated at dawn.