Circadian clock components control daily growth activities by modulating cytokinin levels and cell division‐associated gene expression in Populus trees

Abstract Trees are carbon dioxide sinks and major producers of terrestrial biomass with distinct seasonal growth patterns. Circadian clocks enable the coordination of physiological and biochemical temporal activities, optimally regulating multiple traits including growth. To dissect the clock's role in growth, we analysed Populus tremula × P. tremuloides trees with impaired clock function due to down‐regulation of central clock components. late elongated hypocotyl (lhy‐10) trees, in which expression of LHY1 and LHY2 is reduced by RNAi, have a short free‐running period and show disrupted temporal regulation of gene expression and reduced growth, producing 30–40% less biomass than wild‐type trees. Genes important in growth regulation were expressed with an earlier phase in lhy‐10, and CYCLIN D3 expression was misaligned and arrhythmic. Levels of cytokinins were lower in lhy‐10 trees, which also showed a change in the time of peak expression of genes associated with cell division and growth. However, auxin levels were not altered in lhy‐10 trees, and the size of the lignification zone in the stem showed a relative increase. The reduced growth rate and anatomical features of lhy‐10 trees were mainly caused by misregulation of cell division, which may have resulted from impaired clock function.


| INTRODUCTION
Plants use an internal 24-hr (circadian) clock to synchronize their metabolism and growth with predictable changes in the environment.
The clock mechanism of Arabidopsis is composed of interlocked transcriptional-translational feedback loops (Millar, 2016). It resets to local time on a daily basis in response to light and temperature cues and by sensing sugar produced by photosynthesis (Haydon, Mielczarek, Robertson, Hubbard, & Webb, 2013;Shin et al., 2017).
Plant growth and development are coordinated by the circadian clock. In Arabidopsis, this results in maximal hypocotyl elongation towards the end of the night (Nozue et al., 2007;Nusinow et al., 2011), as well as in delayed flowering under short-day lengths (Seaton et al., 2015). Arabidopsis, however, is an annual species and far less is known about the regulation of growth in long-lived plant species such as deciduous trees. The Populus genome contains two LHY genes (LHY1 and LHY2), which appear to be orthologous with Arabidopsis LHY and CCA1. LHY1 and LHY2, together with TOC1, are the only proteins so far associated with clock function in Populus Takata et al., 2009). We previously showed that LHY1 and LHY2 are important in coordinating growth of Populus with the long days and warm temperatures of spring and early summer and in enabling the response to cold and the development of freezing tolerance during winter dormancy .
Temporal regulation of growth and development may be critical in maximizing trees' fitness at high latitudes, where growing seasons are short. To understand the role of the circadian clock in maximizing biomass production in a long-lived perennial plant, we investigated patterns of growth in trees with a faster circadian clock. We studied trees in which expression of the core clock genes LHY1 and LHY2 was reduced by RNAi, causing the clock period to shorten by 3-4 hr, to investigate the impact of the circadian clock in growth. To test the hypothesis that a functional clock is central for aligning daily growth processes in Populus trees, we carried out detailed investigations of gene expression and cell division and of metabolism of the growth regulators auxin and cytokinins, as well as of primary and secondary growth.

| Plant materials, growth, and sampling
All experiments were conducted using wild-type (WT) hybrid aspen (Populus tremula × P. tremuloides) T89 cv. and lhy-3, lhy-10, toc1-4, and toc1-5 RNAi lines, as indicated. In the RNAi lines, expression of either TOC1 or LHY1 and LHY2 is reduced by~40%, resulting in freerunning periods that are approximately 3 to 4 hr shorter than those of WT trees . Representative RNAi lines were selected from the 10 independently derived lines described previously .
Plants were propagated vegetatively and grown under long photoperiods (light:dark [LD] 18 hr:6 hr) at 18°C  or under indicated photoperiodic conditions. Nutrients (SuperbaS, Supra Hydro AB, Landskrona, Sweden) were supplied once weekly from Week 4. Plant height was measured weekly from approximately 21 days after potting. Once trees had reached approximately 20 cm in height, the stem diameters 10 cm above the soil were measured weekly.
Three biological pools of leaf blade samples were collected at 4-hr intervals from 28-day-old trees for microarray and metabolite analyses.
Leaf material was collected from Internodes 8-11 of WT and lhy-10 plantlets. The 28-day-old trees were sampled randomly, with respect to leaf position and plant, as biological pools of leaves (one leaf per plant) collected randomly from four individual plants every 4 hr, with at least 8 hr between resampling of individual trees.
RNA for microarray analysis was obtained from two biological pools (eight plants; each pool consisted of four leaves [two leaves per tree, from two independent trees]) sampled in parallel. Sample collection started 3 hr before dawn (ZT21) and ended 48 hr later. RNA was extracted using the cetyltrimethylammonium bromide (CTAB) method (Chang, Puryear, & Cairney, 1993) and purified by an RNeasy Plant Mini Kit (Qiagen, Hiden, Germany), including DNAse treatment as described in the manufacturer's protocol and hybridized to an Affymetrix Populus array (Affymetrix Inc., Santa Clara, CA, USA) at the Nottingham Arabidopsis Stock Centre (NASC) array facility (Craigon et al., 2004). Gene expression profiles were confirmed in an independent experiment using quantitative reverse transcription polymerase chain reaction (RT-qPCR). Leaves were sampled as described above; sampling began at dawn (ZT0) and ended 36 hr later. RNA was extracted and treated as described above.
Auxin measurements were made on three independent pools of four leaves (biological replicates), each with three technical replicates.
Material collected for cytokinin (CK) measurements consisted of a series of biological pooled samples, each with four technical replicates, collected at 4-hr time-points over 48 hr. The pools of leaf material collected for auxin and CK measurements overlapped with those collected for the microarray experiment.

| Microarray analysis
Microarray data were generated by the NASC array facility using the GeneChip Poplar Genome Array (Affymetrix), with RNA from the diurnal time course sampled from WT and lhy-10 (as described above).
Samples were processed according to NASC's standard procedure.
Briefly, RNA samples were quality controlled using the Agilent 2100 (RMA) in GeneSpring version 12.5 (Agilent Technologies), in which further statistical analysis was completed. RMA preprocessing was completed using a custom generated probe mask file specific for T89 hybrid trees, which was generated according to protocols described by NASC (Graham, Broadley, Hammond, White, & May, 2007;Hammond et al., 2005), using gDNA obtained by cetyl trimethylammonium bromide extraction (Eriksson et al., 2000), and a threshold signal level of >100 was applied.
Microarray data were preprocessed with RMA in GeneSpring ver-

| Circadian rhythmicity scoring using COSOPT
The cosine-wave fitting algorithm (COSOPT) analysis (without the linear regression option) was performed as described (Edwards et al., 2006) using median normalized Ln expression values exported from GeneSpring. The COSOPT method tests the fit of a single, modified cosine function with many parameters. Genes scored with a pMMC-ß threshold of <0.05, and periods 20-28 hr were considered rhythmic (Straume, 2004). Gene Ontology analysis was carried out on clusters formed by phase-binned COSOPT results (Edwards et al., 2006;Straume, 2004; Dataset S1). This analysis used singular enrichment analysis in AgriGO (Du, Zhou, Ling, Zhang, & Su, 2010), based on the Populus trichocarpa v3.0 annotations (PopARRAY database) and a custom background list. Genes were considered present for the analysis if at least one probe-set represented them for each individual cluster and for the background list.

| Bayesian Fourier clustering
Bayesian Fourier clustering analysis (Liverani, Anderson, Edwards, Millar, & Smith, 2009) was conducted using microarray data from WT trees (Dataset S2), as described previously (Edwards et al., 2006;Heard, Holmes, & Stephens, 2006). Bayesian Fourier clustering fits a wide range of waveforms, using up to five sines and cosines with a shared fundamental period.

| CK quantification
The concentrations of endogenous CK metabolites were determined in leaves from WT and lhy-10 trees, sampled as described above. Extraction and purification of metabolites from 100-mg leaf tissue or 40-mg stem tissue samples were as described previously (Novák et al., 2003;Novák, Hauserová, Amakorová, Doležal, & Strnad, 2008). The samples were purified by combining two ion-exchange chromatography steps (strong cation exchange, diethylaminoethyl-Sephadex combined with C18-cartridges) with immunoaffinity purification. CK levels were quantified using ultraperformance liquid chromatography electrospray tandem mass spectrometry (Novák et al., 2008). A mixed effects model was used to determine significant differences in levels of each metabolite between genotypes across all 13 time-points; p values were calculated in R using the lme4 package (Bates, Mächler, Bolker, & Walker, 2015) with "genotype" and "timepoint" included as fixed effects and "plant" and "leaf" included as random effects.

| In vivo assays of promoter CYCD3:LUCIFERASE and CCR2:LUCIFERASE activities
The CYCD3 promoter region (Potri.014G023000.1; corresponding to gene model Scaffold 961 P_tremuloides_ × _P_tremula_T89_v0001; http://popgenie.org/) was used for primer design. Nested PCR was performed to clone a 3034 BP promoter from a T89 cv. gDNA template using the following primers: First round PCR: forward 5'-ACATCTCAC- To test the dependence of CYCD3 expression on the Populus circadian clock, we cloned and fused its promoter to LUCIFERASE to enable real-time analysis in WT and lhy-10 backgrounds. The CYCD3 promoter sequence was ligated into pPZP221LUC+ to produce pCYCD3: LUC. pCYCD3:LUC was introduced into WT and lhy-10 trees using Agrobacterium-mediated transformation, as described previously (Eriksson et al., 2000), with gentamycin selection (50 μg/ml). to WT and lhy-10 trees has been described elsewhere . Levels of bioluminescence produced by the pCCR2:LUC and pCYCD3:LUC reporters were measured in detached leaves or apices of WT and lhy-10 plants from at least three independent lines per genotype, using one leaf from at least six different plants of each line.
We entrained leaves and apices (cut and trimmed of leaves and leaf primordia) from WT and lhy-10 plants carrying LUC reporter constructs as follows: Excised tissues were placed on plates containing 0.5 × Murashige-Skoog medium (plus vitamins but without additional sucrose) and entrained to LD 18:6 photoperiods for 7 days. Tissues were then grown under LD 18:6 (equal parts blue (470 nm) and red light (660 nm) from 40 μmolm −2 s −1 light-emitting diodes [MD Electronics]) during the light period at 22°C. After 1-3 days, the light regime was changed at dawn (ZT0) to LL (constant red plus constant blue light) at 22°C for recording of free-running bioluminescence rhythms. Plant imaging data were analysed using BRASS Fourier analysis software, as described previously . Analysis of phase was performed using data collected in LD 18:6; period length measurements were made using data collected 24-120 hr after the transfer to LL.

| Protoplast protein assays
Protoplasts were prepared from an Arabidopsis cell culture, transfected with each pRT104 construct, and treated as described (Johansson et al., 2011). For protein stability assays, protoplasts were cotransfected with BBX19 or BBX32 and LHY2 expression constructs.
After 18  Protein signals were detected following western blotting using West Femto Maximum sensitivity substrate (ThermoScientific, Rockford, IL, USA) and a FUJIFILM LAS-3000 Luminescent Image Analyser.

| Statistical analyses
Statistical significance was tested using one-way or two-way analysis of variance (ANOVA) followed by the multiple comparisons tests or unpaired Student's t-tests, as indicated, using GraphPad Prism version 6.0 for Windows (GraphPad Software, La Jolla California USA). In addition, specific statistical packages were used to analyse microarray studies, hormone measurements, and circadian rhythms, as described above.

| Perturbation of the circadian clock alters growth of Populus
To investigate the impact of clock perturbations on growth in Populus, tree height was measured in lines which had short circadian periods due to a reduction in clock gene expression caused by RNAi. WT trees were significantly taller than the RNAi lines ( Figure 1a). lhy-3 and lhy-10 had stronger growth defects ( Figure 1a) and shorter internal periods (approximately 20 hr) than toc1-4 and toc1-5 lines (approximately 21 hr; Ibáñez et al., 2010). Heights of clock mutant trees were significantly affected: one-way ANOVA (p = .0033; n = 8-9 per genotype) followed by Dunnett's multiple comparisons test showed that lhy-3 and lhy-10 (p < .01; n = 8-9) and toc1-5 (p < .05; n = 8-9) but not toc1-4 (ns; n = 9) differed significantly from WT. Because the clock and growth characteristics of the two lhy lines were similar ( Figure 1 a; Ibáñez et al., 2010), further investigations of height and diameter were made only in lhy-10 and WT trees grown under long-day photoperiods (LD 18:6).
WT trees were larger than lhy-10 trees, with increased stem height and diameters (Figure 1b,c). They showed consistently greater increases in stem volumes, and higher leaf, stem, and root biomasses, with growth of lhy-10 being 30-40% that of WT (Tables 1 and S1; Figure S1).
To investigate whether the perturbed growth of lhy-10 resulted from desynchronization between endogenous period and the environ-  Figure S3c,d). When growth was measured in LD 18:6 (T = 24 hr), WT trees grew significantly faster than lhy-10 (growth rates in LD: WT: 1.83 ± 0.03 cm day −1 ; lhy-10: 1.69 ± 0.04 cm day −1 [p = .0093; n = 9]). There was, however, no significant difference in growth rates between genotypes following a shift to constant light (growth rates in LL: WT: 1.33 ± 0.04 cm day −1 ; lhy-10: 1.28 ± 0.06 cm day −1 [p = .36; n = 9]). The growth rate of WT was reduced to the same level as lhy-10 in LL. All these results suggest the impaired growth of lhy-10 does not simply result from a mismatch between their endogenous period and the environmental LD cycle.

| lhy-10 trees show reduced levels and altered metabolite profiles of CK but not IAA
Assays of auxin and CK levels in expanding source leaves after 28-day growth in LD 18:6-before growth differences between genotypes became apparent (Figure 1c)-provided insight into the auxin status and CK metabolism of the trees. Relative to WT, CK metabolites in lhy-10 leaves showed substantial reductions in levels of the isoprenoid CKs trans-zeatin (tZ), cis-zeatin (cZ), dihydrozeatin (Sakakibara, 2006), and the aromatic ortho-topolin (oT; Sakakibara, 2006;Strnad, 1997

| Alteration in cytokinins and IAA timing and ratios separate auxin-driven xylem differentiation and increased wood formation in lhy-10
Changes in IAA levels are associated with, and required for, daily patterns of tree growth and, in particular, for cell elongation, cell division, and wood formation (Bhalerao & Fischer, 2014). Analyses of levels of IAA and its catabolite oxIAA (Pěnčík et al., 2013, Tuominen, Ostin, Sandberg, & Sundberg, 1994 in leaves showed no significant temporal or genotypic differences between lines (Figures 2b and S3), suggesting that IAA metabolism remained intact in lhy-10.
We investigated the zone of lignification and xylem differentiation and found it occupied a broader area of stems in lhy-10 than in WT, counted as lignified vessels per area (Figures 2c and S4a,b).
Phloroglucinol staining of the lignification zones in transverse sections of stem showed the extent of lignification and size distribution of fibres and vessels were similar in lhy-10 and WT ( Figure S4c,d) but

| Repression of LHY expression provides insights into circadian control of growth of Populus
To investigate the effect of repressing LHY1 and LHY2 on circadian regulation of gene expression, we performed a microarray time-course experiment using leaf tissue from WT and lhy-10 Populus trees grown under LD 18:6. In WT Populus, approximately 12% of genes represented on the microarray by at least one probe set (3,737 out of 31,561 genes) showed diurnal rhythms. This fell to 7% (2,320 genes) in lhy-10 trees. Times of peak gene expression in WT fell into two major clusters, one centred shortly after dawn (ZT2-4) and the other before dusk (ZT12-14; Figure 3a PRR7s was similar in both WT and lhy-10, suggesting they were less sensitive than PRR9s to LHY levels. As expected, both GI and ELF4 showed evening phases of expression in WT (Figure 3c; Table S2; Edwards et al., 2010). Strong dusk tracking, by ELF4 in particular, was observed, even in lhy-10, which may be important for photoperiodic regulation of growth Nozue et al., 2007;Nusinow et al., 2011).
In accordance with earlier findings , lhy-10 leaves exhibited an earlier phase of pCYCD3:LUC expression than WT at the point of transition from LD to LL (Figure 4b). WT leaves produced a rhythmic pattern of bioluminescence in LL, whereas lhy-10 leaves appeared arrhythmic (Figure 4b). Period analysis revealed a mean period length of 21.39 ± 0.08 hr in WT leaves (n = 12 rhythmic; one arrhythmic) and that all traces (n = 11) from lhy-10 leaves were indeed arrhythmic.
We used pCYCD3:LUC and an additional promoter:reporter construct, pCCR2:LUC, to investigate the clock's performance in apices and stem tissue. Plants were initially rhythmic, although lhy-10 tissues had earlier phases and shorter periods than WT ( Figure 5). The mean period lengths of pCCR2:LUC and pCYCD3:LUC observed in lhy-10 apices were 3-4 hr shorter than those of WT (Tables 2 and 3), consistent with previous observations . Thus, pCYCD3:LUC is clock-regulated, with an early phase and short period, in stem tissues of lhy-10 trees ( Figure 5; Table 2) and has a similar pattern of expression to the well-established evening reporter construct pCCR2:LUC ( Figure 5; Table 3). One-way ANOVA (p = .0001; n = 3, followed by Bonferroni's multiple comparisons test) found no significant differences between period lengths of pCYCD3:LUC and pCCR2:LUC in WT tissue (ns, n = 3); however, the period lengths of these reporters were significantly shorter in tissues from lhy-10 than in WT tissues (p < .0001; n = 3). Together, these data indicate that CYCD3 was clock-regulated in both apices and leaves, and dependent on LHY1 and LHY2 expression, consistent with the numerous CCA1-binding and circadian elements present in the promoter. Thus, the disruption of circadian clock function in lhy-10 probably affects CYCD3 expression directly, and this has an impact on cell division leading to diminished growth of lhy-10 trees.    processes generating reactive oxygen species (Wulund & Reddy, 2015). We found evidence for lower levels of radial cell division patterns in internodes of lhy-10 trees at night under LD cycles ( Figure 6). This is consistent with altered expression of CYCD3 in lhy-10 trees and interaction of CYCD3 with LHY2 (Figures 3-6).

| WT and lhy-10 plants show different patterns of cambial activity
CYCD3 availability is an important rate-limiting step for cell division in Arabidopsis, acting on populations of cells to maintain mitotic cycling and restrict endocycling (Dewitte et al., 2007;Menges, Samland, Planchais, & Murray, 2006). CYCD3 is receptive to mitogenic signals and functions at the G1 phase. Its expression is induced by CK in Arabidopsis (Riou-Khamlichi, Huntley, Jacqmard, & Murray, 1999); however, high CK levels are not required for high CYCD3 expression (Dewitte et al., 2007). CYCD3 expression also regulates growth in Populus (Karlberg et al., 2011). Our data suggest the clock affects both CK metabolism and CYCD3 expression, and, by generating a rhythm in CYCD3 expression and cell division, has an important impact on growth (Figures 4, 5, and 6). As the Arabidopsis CYCD3;1-3 promoters also contains putative CCA1-binding sites, this mechanism of regulation by the clock may be widespread across plant species.
Of the genes scored as rhythmic in lhy-10, 1,301 were specific to lhy-10 ( Figure 3a). showed an evening phase in lhy-10, the expression of EC components did not differ greatly between lhy-10 and WT trees (Figure 3c), suggesting the EC does not suppress expression of this gene subset at night. Together with earlier studies , this suggests an interaction between light and clock control, rather than conventional period resonance, underlies the timing of CK synthesis, cell division, and proliferation in Populus.   Table S1 and Figure S4. (HY5) function during light development (Holtan et al., 2011). Further, overexpression of BBX32 sets the hypocotyl elongation phase to the dark period in Arabidopsis (Holtan et al., 2011) and increases yield of soybean (Glycine max; Preuss et al., 2012).

Recent data from
Arabidopsis suggest BBX32 is part of a regulatory loop with CCA1 and/or LHY, because overexpression of BBX32 increases both their expression and circadian period length (Tripathi, Carvallo, Hamilton, Preuss, & Kay, 2017). Arabidopsis cca1;lhy mutants, moreover, show an earlier phase of BBX32 expression. In Populus, LHY1, LHY2, BBX19, and BBX32 are coexpressed in WT and have a significantly earlier phase of expression in lhy-10 (Figure 3d), mirroring BBX32 expression in Arabidopsis (Tripathi et al., 2017). Importantly, the phasing of BBX19 and BBX32 may be part of a clock-associated growth control mechanism across diverse species, including Arabidopsis, soybean and Populus (Figure 3).
Stem elongation, cell division, and wood formation in Populus are shaped by the interactions of multiple plant hormones, including auxins, CKs, ethylene, and gibberellins (Eriksson et al., 2000;Israelsson et al., 2003;Love et al., 2009;Nieminen et al., 2008;Tuominen, Puech, Fink, & Sundberg, 1997). Auxin and CKs are exported from actively growing "source" leaves and act as mobile signals regulating cell division in the meristems of the apices, cambium, and roots (Nieminen et al., 2008). Whereas the bioactivities of dihydrozeatin, immunoprecipitation (IP), and tZ are well-known (Sakakibara, 2006), those of cZ and oT are less well established. However, the Arabidopsis HISTIDINE KINASE 4 (AHK4/CRE/WOL) receptor responds to tZ and the aromatic CK, meta-topolin. Similarly, the Zea mays HISTIDINE KINASE 1 (ZmHK1) receptor responds to cZ and oT, which suggests all these species are active CKs (Mok et al., 2005). The reduction in levels of CK metabolites (and especially of tZ, cZ, and oT) in lhy-10 trees suggested their reduced growth rate might result from a change in CK metabolism. The growth reduction observed in lhy-10 phenocopies that of transgenic trees in which levels of active CKs are reduced by an increase in CYTOKININOXIDASE2 (CKX2) expression (Nieminen et al., 2008). Both CK levels and the diurnal pattern of metabolites are likely to be responsible for driving growth. Our data indicate CK metabolism in leaves is under both circadian and diurnal controls (Figures 2a and S4).
Auxin plays an important role in cell division, and a gradient of auxin over the cambium regulates secondary cell wall biosynthesis (Hertzberg et al., 2001;Tuominen et al., 1997;Uggla, Moritz, Sandberg, & Sundberg, 1996). Importantly, auxins and cytokinins function in different parts of the stem to regulate cell division and xylem differentiation; cytokinins have been specifically shown to affect diameter growth (Immanen et al., 2016). Although auxin biosynthesis, xylem formation, and lignin biosynthesis appeared normal in lhy-10, the relative amount of wood laid down was altered (Figure 2). The opposite effects seen on plant growth and secondary cell wall formation indicate CK biosynthesis is temporally and/or spatially separated from the clock's control of auxin in lhy-10 (Figures 1 and 2). Together with the observed patterns in hormone cycles, this suggests the clock differentially regulates the auxin and CK systems. Although the phases of LHY and PRR9 expression were strongly affected in lhy-10, expression of most EC members (including ELF3, ELF4, and LUX) was similar in WT and lhy-10 (Table S2), and ELF4 remained rhythmic in lhy-10 with only a small phase advance ( Figure 3; Table S2). This is consistent with regulation of diurnal auxin levels by the EC in Populus (Figure 7).
Biomass, defined as wood volume, is negatively correlated with lignin content (Novaes, Kirst, Chiang, Winter-Sederoff, & Sederoff, 2010). Rational manipulation of this trade-off requires biological regulators that dissect growth and wood development. The clock provides one such target regulator because of its distinctive, rhythmic control of plant metabolism and growth; for instance, PRR7a and ELF3 are candidate genes for Populus QTLs underlying diameter growth and internode count, explaining 4.42% and 6.69% of genetic variation, respectively (Novaes et al., 2009).

FIGURE 7
Temporal dissection of regulators of plant growth and development. Overview of growth coordination by the Populus oscillator. Cytokinin biosynthesis is controlled and sustained during the day. High levels of bioactive cytokinins are known to control cell division and proliferation. Environmental cues such as light and temperature reset the clock to local time. During the day, the clock, acting via LHYs, regulates CYCLIN D3 gene expression and protein function. LHY1 and LHY2 may promote expression of BBX19 and BBX32 and their proteins may be part of the timing mechanism regulating growth. Other components Takata et al., 2009), dependent on, for instance, TOC1, may maintain or respond to auxin and act to promote elongation growth similar to Arabidopsis hypocotyls , xylem differentiation, and lignification (Bhalerao & Fischer, 2014). Tentative interactions are indicated as dashed lines We propose that maximal expression of LHY1 and LHY2 during the morning promotes growth by activating genes such as BBX19 and BBX32 and that this affects CK biosynthesis and responses. These features may be linked; up-regulation of BBX32 in soybean gives a "stay green" phenotype consistent with high CK levels (Preuss et al., 2012), and CK perception and response proteins interact with Arabidopsis BBX32 (Tripathi et al., 2017). Our findings match reports of daily leaf growth rhythms in Populus deltoides, which shows high rates of growth during the evening and night (Matsubara et al., 2006). Reports from green algae (Chlorella sp.), flagellate algae (Euglena gracilis), and cyanobacteria (Synechococcus sp.) suggest the circadian clock gates cell division and growth (Bolige, Hagiwara, Zhang, & Goto, 2005;Mori, Binder, & Johnson, 1996;Stirk et al., 2011), with the level of active CKs affecting this gating (Stirk et al., 2011). Moreover, cyclins and Wee1 have been implicated in the circadian control of mammalian cell division (Feillet et al., 2014;Matsuo et al., 2003).
Our study is a first step towards understanding the clock's multilayered, temporal control of growth and biomass production in a perennial tree species. We show that the circadian clock acts to regulate cell division, probably through control of CYCD3 expression and physical interactions between proteins. This suggests a close interregulatory link, between the circadian clock and the cell cycle which is fundamental to the control of growth and production of biomass.