Low-temperature-specific effects of PHYTOCHROME C on the circadian clock in Arabidopsis suggest that PHYC underlies natural variation in biological timing

The circadian clock is a fundamental feature of gene regulation and cell physiology in eukaryotes and some prokaryotes, and an exemplar gene regulatory network in Systems Biology. The circadian system in Arabidopsis thaliana is complex in part due to its photo-transduction pathways. Analysis of natural genetic variation between Arabidopsis accessions Cape Verde Islands (Cvi-0) and Landsberg erecta (Ler) identified a major, temperature-specific Quantitative Trait Locus (QTL) on chromosome V that altered the circadian period of leaf movement (Edwards et al., Genetics, 2005). We tested Near-Isogenic Lines (NILs) to confirm that Ler alleles at this PerCv5c QTL lengthened the circadian period at 12°C, with little effect at higher temperatures. The PHYTOCHROME C gene lies within the QTL interval, and contains multiple sequence variants. Plants carrying either a T-DNA-insertion into PHYC or a deletion of PHYC also lengthened circadian period under white light, except at 27°C. phyB and phyABE mutants lengthened period only at 12°C. These results extend recent data showing PhyC effects in red light, confirming the number of photoreceptor proteins implicated in the plant circadian system at eleven. The connection between light input mechanisms and temperature effects on the clock is reinforced. Natural genetic variation within PHYC is likely to underlie the PerCv5c QTL. Our results suggest that functional variation within the PHYC-Ler haplotype group might contribute to the evolution of the circadian system and possibly to clock-related phenotypes such as flowering time. These results have previously passed peer-review, so we provide them in this citable preprint.


INTRODUCTION
The circadian clock is a 24h endogenous timer that allows the correct temporal regulation of physiological, biochemical and developmental processes. The expression of over 30% of the Arabidopsis thaliana transcriptome is driven by the circadian clock (Covington et al. 2008;Michael et al. 2008;. Rhythmic transcriptional outputs control physiological processes such as daily rhythmic growth and photoperiodic flowering. In both cases, the mechanisms are sufficiently characterised to support mechanistic, mathematical models . The clock mechanism in all organisms includes interlocked transcriptional-translational feedback loops. The clock's rhythmic behaviour is thought to emerge from the dynamics of this clock gene circuit, which have been well characterised in Arabidopsis (Flis et al. 2015). The negative feedback loops in this model plant species incorporate two closely-related MYB transcription factors CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LONG ELONGATED HYPOCOTYL (LHY) that inhibit the expression of eveningexpressed genes such as a pseudo-response regulator (PRR) TIMING OF CAB EXPRESSION 1 (TOC1). The expression of CCA1 and LHY is tightly regulated by other clock components, including sequential inhibition by PRR9, PRR7 and PRR5. TOC1 and other PRR genes are repressed by an Evening Complex (Hsu and Harmer 2014).
For the clock to be useful, the endogenous period must be synchronised (entrained) to match the 24-hour environmental cycle (Johnson et al. 2003). The strongest entrainment signals are temperature and light. At least four families of photoreceptors have been identified as transducing light signals to reset the clock, the blue light sensing cryptochromes (CRY1 and CRY2), the red/far-red light (R/FR) sensing phytochromes (PHYA, PHYB, PHYD, PHYE), (Devlin and Kay 2000;Somers et al. 1998;Yanovsky et al. 2000), the UV-B photoreceptor UVR8 (Feher et al. 2011) and a family of three F-box proteins, including ZEITLUPE (ZTL) (Baudry et al. 2010). These ten photoreceptors transduce light signals to regulate clock genes and proteins (Fankhauser and Staiger 2002), with both specialised and overlapping roles. PHYC alone has little effect on other phenotypes in the absence of other phytochromes (Hu et al. 2013), and no role for PhyC in the Arabidopsis clock had been confirmed until, during the preparation of this paper, a long-period phenotype was reported under red light (Jones et al. 2015). In barley, some PHYC alleles have also shown circadian effects (Pankin et al. 2014;Nishida et al. 2013) and in Arabidopsis, natural genetic variation in related traits has been associated with PHYC (Balasubramanian et al. 2006).
Temperature effects on circadian clocks include entrainment by temperature cycles, whereas under constant temperatures, the circadian period is unusually constant across a physiological temperature range (termed 'temperature compensation'). Among multiple mechanisms implicated in how ambient thermocycles influence the clock are alternative RNA splicing, in Neurospora crassa, Drosophila melanogaster (Colot et al. 2005);Low et al. 2008) and Arabidopsis (James et al. 2012), and protein phosphorylation in N. crassa (Mehra et al. 2009). Mutants of some clock genes alter temperature compensation (Gould et al. 2006;Salome et al. 2010). A systems biology analysis of ambient temperature effects across the clock mechanism revealed a strong dependence on light quality, and suggested that light and temperature signalling converged upstream of the clock, such that photoreceptor pathways had a significant role in temperature responses (Gould et al. 2013). This work indicated multiple targets of temperature effects in the clock mechanism, including the morning genes implicated by Salome et al. 2010 andJames et al. 2012, as well as the Evening Complex implicated by Mizuno et al. (2014).
Quantitative genetic approaches have identified genetic variation in clock-affecting genes (Anwer and Davis 2013), including Quantitative Trait Loci (QTL) in Arabidopsis (de Montaigu et al. 2015;Swarup et al. 1999). Our earlier work tested circadian period at three temperatures, identifying multiple QTL (Edwards et al. 2005). Here we propose PHYC as a candidate gene for a clockaffecting QTL that is specific for low ambient temperature.

NILs recapitulate the temperature-specific QTL effect
The PerCv5c QTL was identified in the Cvi by Ler (CvL) recombinant inbred lines (RILs) and mapped to the middle of Chromosome 5. This QTL alone accounted for 44.6% of phenotypic variation in the period of rhythmic leaf movement at 12°C (Edwards et al. 2005), by far the largest single effect observed in that study. Near isogenic lines (NILs) carrying Cape Verde Islands (Cvi) alleles around in the putative circadian QTL were therefore used to identify Cvi loci that regulate the clock, in an isogenic Landsberg erecta (Ler) background (Edwards et al. 2005). Figure 1a shows the periods of NIL45a and NIL106 at 12C, 22C and 27C, with numerical values in Table 1. Both NILs had Cvi introgressions around the PerCv5c locus ( Figure 1b). Both NILs had shorter circadian period compared to the Ler parent especially at 12°C (Figure 1a), consistent with the original QTL identification. The larger introgression of NIL45a shortened the period more than NIL106. It is possible that multiple loci in the Cvi region of NIL45a affect the clock but the small period effects make this difficult to confirm. Natural allelic variation at PHYTOCHROME C PHYC maps within the confidence interval of PerCv5c, so was considered a possible candidate gene for the QTL. Quantitative RT-PCR showed that the PHYC alleles in both accessions were expressed under constant light conditions ( Figure 2). Expression was higher in subjective day (54h in constant light) than in subjective night (62h in constant light), though possibly with lower levels of mRNA in Ler ( Figure  2).
Sequence analyses for the PHYC gene and predicted PHYC protein revealed mutations between the Cvi and Ler backgrounds, leading to four changes in the protein sequence (Figure 3a, 3b). The AT5G35840 locus in sequences from the 1001 Genomes Project (Weigel and Mott 2009) confirmed these four changes ( Figure 3c). The 1001 Genomes resource also showed that the Cvi sequence is close to the Ler haplotype rather than the Col reference haplotype, which is shared by Ws-2. The changes from Ler to Cvi amino acid sequence are A27T, S230G, T352S and H470P, all of which fall in the N-terminal domains of phytochrome before the bipartite PAS motif. At position 230, G is the amino acid in the reference Col genome, whereas in the other positions, the Col sequence is the same as Ler (Figure 3c).

phyC mutants alter circadian period
To test a possible role for PHYC in the circadian clock, the phyC-1 mutant, in a PhyD-deficient background Wassilewskija-2 (Ws) (Franklin et al. 2003) was assayed for period at 12C, 22C and 27C (Figure 4a). phyC-1 displayed 1.6 and 1.4 hours period lengthening at 12C and 22C respectively, but was not significantly different from wild type at 27C. Numerical period values and statistical tests are shown in Table 1.
To confirm the phyC-1 phenotype, period was tested in a second null mutant for phyC. The gn7 line, first described as gne7 by Sorensen et al. (Sorensen et al. 2002), has a large genomic deletion removing the entire PHYC open reading frame in a C24 background ( Figure 5a). As with phyC-1, significant circadian period lengthening was observed in gn7 plants at 12C and 22C, but not 27C (Figure 4b), relative to the C24 parent line. Over expression of PHYC under the 35S promoter rescued this period phenotype, suggesting that the gn7 period effect was principally mediated by the deletion of PHYC (Figure 4b).
phyB-10 single mutants and phyABE triple mutants were tested in the same experiments. Neither mutant showed a period defect in white light at 22°C, as expected, but both lengthened period at 12°C (Figure 6; Table 2).

DISCUSSION
Circadian clock mechanisms include gene regulation by multiple, interlocking feedback loops, which can increase the flexibility of possible regulatory changes over evolutionary time and in the face of environmental variations. In lightgrown Arabidopsis seedlings, multiple photoreceptors contribute light input signals, adding further complexity to the clock network.
We showed that phyC contributes to control clock period in a temperature-specific manner under white light, adding to the four other phytochromes, two cryptochromes, UVR8 and the ZTL/FKF1/LKP2 family, and bringing the number of known circadian photoreceptor proteins in Arabidopsis to eleven. The rhythmic control of native PHYC expression (Toth et al. 2001) was not required for normal circadian rhythmicity, because the gn7 mutant phenotype was rescued by a 35S misexpression transgene (Figure 4b).
A similar phenotype was observed in the phyb-10 and phyabe triple mutant, suggesting that this effect was principally mediated by phyb in both lines. In contrast to the phyB mutants, both phyC mutants lengthened period at 22C as well as 12C. Taken together, these phenotypes suggested that the phyD mutation in the Ws background of phyc-1 and phyb-10 was not uniquely required for the phenotypes observed, though interactions among the phytochromes might well contribute. Hence, PHYB and PHYC are likely the major effectors of the clock regulation detected here.
No circadian period QTL were mapped to the PHYB locus in Edwards et al. (2005). However, an epistatic interaction was suggested between markers FD.222L-Col and CH.60C, which map near to PHYB and PHYC respectively (K.D. Edwards Ph.D. Thesis, University of Warwick). This interaction suggested the possibility of the Ler allele of PHYB enhancing the period difference between the Ler and Cvi alleles of PHYC.
The similarity of the phyC mutant phenotype to the PerCv5c QTL effect, together with the DNA sequences, strongly suggested that allelic variation at PHYC contributes to or solely causes this QTL. If so, then our results further suggested that the PHYC-Ler allele was less active than PHYC-Cvi, consistent with (Monte et al. 2003;Balasubramanian et al. 2006), because PHYC-Ler lengthened circadian period as phyC null mutants also do. The haplotypes of PHYC across multiple accessions were defined by Balasubramanian et al. (2006). Cvi is clearly among the Ler haplotypes, whereas Ws-2 is among the Col haplotypes ( Figure 3c). Our results suggest that PHYC function varies within the Ler haplotype group, at least at low temperature.
The variation within this group may be relevant, for example, to explain the association of PHYC alleles within the Ler haplotype group to particular habitats, such as areas of high precipitation in the Iberian peninsula (Mendez-Vigo et al. 2011).
Page 6, below, shows Figure 6. Circadian period of phyB single and multiple mutants.

Growth conditions
Growth conditions for samples used in leaf movement and RNA studies under constant light were essentially as described (Edwards et al. 2005;Edwards et al. 2010). Briefly, sterile seed were stored in 0.15% agar and stratified at 4° for 4-5 days prior to sowing on Murashige-Skoog 1.5% agar medium containing 3% sucrose. Seedlings were grown for 6 days under constant light of 55-60 μmol m−2 sec−1 cool white fluorescent light and then entrained for 4 days at 21°-22° under (12 hr/12 hr) light/dark cycles of 75 μmol m−2 sec−1 cool white fluorescent light.

Measurement of leaf movement
Individual period estimates were produced from leafmovement data as described (Edwards et al. 2005). In brief, leaf growth of Arabidopsis seedlings was imaged under constant light using a custom-built, 31-camera array. Y-axis centroid positions for each leaf were determined from the image stacks using Metamorph software. Circadian period of each leaf was estimated using the FFT-NLLS algorithm through the BRASS interface (Edwards et al., 2010). Mean period estimates for each genotype were based on data from two to four independent experiments at each temperature analysed using REML (Patterson and Thompson 1971) in the statistical package GENSTAT 5 (Payne et al. 1993) as described (Edwards et al. 2005). The significance of differences between pairs of genotypes was analysed via ttests using the SEM estimates derived from REML.

Molecular Assays
Quantitative RT-PCR was performed according to (Edwards et al. 2010). Phytochrome Western blotting was performed as described (Franklin et al. 2003).