INTRODUCTION
Plant growth and biomass production are direct functions of the diurnal cellular carbon balance, which is regulated by a combination of light responses and the circadian clock. Light responses are triggered by a change in regime (i.e., presence or absence of light), while the circadian clock is comprised of transcriptional regulators that operate in anticipation of a change (e.g. transition from light to dark) and whose activities span the 24 hour (h) photoperiod (Nohales & Kay, 2016; Oakenfull & Davis, 2017; Seluzicki, Burko, & Chory, 2017). The clock transcriptional regulators include CCA1/LHY, PRR5, PRR7 and PRR9, which form the morning loop, and TOC1, ELF3, LF4 and LUX, which form the evening loop (Flis et al., 2015; Staiger, Shin, Johansson, & Davis, 2013). More than 30% of all Arabidopsis genes are regulated by the circadian clock at the transcript level (Blasing et al., 2005; Covington, Maloof, Straume, Kay, & Harmer, 2008). However, less is known about how the resulting diurnal transcriptome relates to protein abundance (Abraham et al., 2016; Choudhary, Nomura, Wang, Nakagami, & Somers, 2015; Graf et al., 2017) and post-translational protein modifications (Choudhary et al., 2015; Uhrig, Schlapfer, Roschitzki, Hirsch-Hoffmann, & Gruissem, 2019), both of which may also affect protein function at light-dark transitions and throughout the diurnal cycle. Transcript and protein abundance changes are often disconnected because changes in transcript levels show no corresponding change in protein abundance (Baerenfaller et al., 2012; Seaton et al., 2018). For example, this was found, in the circadian clock mutants CCA1/LHY, PRR7/PRR9, TOC1 and GI (Graf et al., 2017) or for the variability in the timing of peak protein levels relative to the cognate transcript expression (translational coincidence) as a function of the photoperiod-dependent coordination between transcriptome and proteome changes (Seaton et al., 2018). Variable delays between peak transcript and protein abundance have implicated post-transcriptional regulation (e.g. splicing), translational regulation (e.g. translation rate) as well as post-translational regulation (e.g. protein phosphorylation) as possible mechanisms to explain the temporal differences in RNA and protein abundance. Recent studies of plant protein-level regulation have found extensive variability in protein turnover (Li et al., 2017; Seaton et al., 2018). This adds further regulatory complexity because quantitative proteome workflows cannot easily account for protein turnover. Although RNA and protein synthesis, stability and turnover all contribute to the coordination of RNA and protein abundance, how these mechanisms are integrated is currently not well understood. Insights into this regulatory complexity requires both time-course experiments and multi-Omics analysis. We used a large-scale quantitative proteomics approach to determine the extent of diurnal abundance and/or phosphorylation changes of Arabidopsis leaf rosette proteins involved in cellular and metabolic processes and how these protein level changes relate to cognate transcript levels.
Reversible protein phosphorylation is the most abundant post-translational modification (PTM) in eukaryotes (Adam & Hunter, 2018; Rao, Thelen, & Miernyk, 2014). In non-photosynthetic eukaryotes phosphorylation is found to modulate more than 70% of all cellular processes (Olsen et al., 2006), including the circadian clock itself (Robles, Humphrey, & Mann, 2017). This is likely similar in land plants because they have a significantly larger kinome (Lehti-Shiu & Shiu, 2012) compared to humans, which encode 518 protein kinases (Manning, Whyte, Martinez, Hunter, & Sudarsanam, 2002). In contrast, both plants and humans have an equally comparable small number of protein phosphatases (Kerk, Templeton, & Moorhead, 2008). However, most protein phosphatases require association with regulatory subunits to achieve their specificity (Moorhead et al., 2008; Uhrig, Labandera, & Moorhead, 2013), suggesting that similar complexity in protein dephosphorylation across plants and humans has evolved through the expansion of protein phosphatase regulatory subunits.
In plants, diurnal protein phosphorylation is regulated either in response to light, by the circadian clock (Choudhary et al., 2015), or both (Uhrig et al., 2019), while the clock itself is regulated by phosphorylation (Kusakina & Dodd, 2012; Uehara et al., 2019). Recent studies of the circadian phosphoproteome combining the analysis of a free-running cycle and the circadian clock mutant elf4 (Choudhary et al., 2015) or CCA1-OX over-expression (Krahmer et al., 2019) have revealed temporally modified phosphorylation sites related to casein kinase II (CKII) and sucrose non-fermenting kinase 1 (SnRK1). SnRKs are likely involved in the regulation of the circadian phosphoproteome because the transcription of genes encoding multiple SnRK and calcinuerin B-like (CBL) interacting kinases (CIPK) was mis-regulated in the Arabidopsis circadian clock mutants cca1/lhy1 ,prr7prr9 , toc1 and gi201 mutants at end-of-day (ED) and end-of-night (EN) (Graf et al., 2017). Similarly, studies quantifying changes in the phosphoproteome at ED and EN in Arabidopsis rosette leaves, roots, flowers, siliques and seedlings have revealed a large number of diurnally changing phosphorylation events corresponding to diverse protein kinase motifs (Reiland et al., 2009; Uhrig et al., 2019).
Considering the physiological and metabolic changes at the light-dark (L-D) and dark-light (D-L) transitions (Annunziata et al., 2018; Gibon et al., 2009; Usadel et al., 2008), we performed a quantitative phosphoproteome analysis of proteins that are phosphorylated immediately before and after the L-D and D-L transitions during a 12 h light : 12 h dark photoperiod and asked how these phosphorylation events intersect with protein abundance changes. Together with the 24 h photoperiod time-course proteome data, our systems-level analysis provides new insights into diurnal protein and phosphorylation regulation in Arabidopsis rosette leaves.

MATERIALS AND METHODS

Arabidopsis Col-0 wild‑type plants were grown at the Forschungszentrum Jülich (Germany) in an environmentally controlled chamber (GrowScreen Chamber; https://eppn2020.plant-phenotyping.eu/EPPN2020_installations#/tool/30; Barboza-Barquero et al., 2015) under a 12 h light:12 h dark photoperiod and controlled conditions as described in Baerenfaller et al.(2012), including air temperature of 21ºC during the day and 20ºC during the night, air humidity of 70%, and an incident light intensity of ~220 mmol/m2/s at the plant level. Whole rosettes were harvested at 31 days after sowing (DAS) prior to flowering. Four whole rosettes were pooled into one sample and 4 biological replicates were taken at each time point. For total proteome analyses, samples were taken every 2 h during 24 h, starting at Zeitgeber time 1 (ZT1, i.e. 1 h after lights turned on). For protein phosphorylation analyses, samples were 30 min before, 10 min after and 30 min after the L-D and D-L transitions. Samples were snap-frozen in liquid N2 and stored at -80ºC until protein extraction.