Circadian and diel regulation of photosynthesis in the bryophyte Marchantia polymorpha

Abstract Circadian rhythms are 24‐h biological cycles that align metabolism, physiology, and development with daily environmental fluctuations. Photosynthetic processes are governed by the circadian clock in both flowering plants and some cyanobacteria, but it is unclear how extensively this is conserved throughout the green lineage. We investigated the contribution of circadian regulation to aspects of photosynthesis in Marchantia polymorpha, a liverwort that diverged from flowering plants early in the evolution of land plants. First, we identified in M. polymorpha the circadian regulation of photosynthetic biochemistry, measured using two approaches (delayed fluorescence, pulse amplitude modulation fluorescence). Second, we identified that light‐dark cycles synchronize the phase of 24 h cycles of photosynthesis in M. polymorpha, whereas the phases of different thalli desynchronize under free‐running conditions. This might also be due to the masking of the underlying circadian rhythms of photosynthesis by light‐dark cycles. Finally, we used a pharmacological approach to identify that chloroplast translation might be necessary for clock control of light‐harvesting in M. polymorpha. We infer that the circadian regulation of photosynthesis is well‐conserved amongst terrestrial plants.


| INTRODUCTION
The rotation of the earth on its axis causes 24-h cycles in environmental conditions, such as light and temperature. These diel cycles are thought to have been selected for the evolution of circadian clocks within several Kingdoms of life (Dodd et al. 2005;Eelderink-Chen et al. 2021;Millar 2016;Ouyang et al. 1998;Spoelstra et al. 2016). Circadian rhythms are self-sustaining biological cycles that have a period of about 24 h, which are thought to provide a biological measure of the time of day. These rhythms coordinate and sequence processes around the diel cycle, while ensuring that responses to environmental cues are appropriate for the time of day.
Circadian rhythms in plants are generated by a molecular oscillator that is formed from a series of transcription-translation feedback loops, which has a predominance of negative feedback steps (Hsu & Harmer 2014). The oscillator is entrained to environmental cues such as light and temperature, which align the phase of the rhythm with the day/night cycle so that it provides an accurate biological measure of the time of day. There is circadian regulation of a variety of key aspects of the physiology of flowering plants, such as stomatal opening, light-harvesting, CO 2 fixation, and starch metabolism (Dakhiya & Green 2022;Dakhiya et al. 2017;Dodd et al. 2005;Graf et al. 2010). Furthermore, the circadian oscillator contributes to the photoperiodic control of flowering time (Hayama & Coupland 2003;Park et al. 1999) and the fitness of plants (Dodd et al. 2005; R. M. Green et al. 2002). This means that circadian regulation has a pervasive influence on physiology, metabolism, and development (Hotta et al. 2007;Millar 2016) and Mesembryanthemum crystallinum (Dakhiya et al. 2017;Davies & Griffiths 2012;Dodd et al. 2003Dodd et al. , 2004Dodd et al. , 2005Gould et al. 2009;Hennessey & Field 1991;Wyka et al. 2005). Aspects of photosynthetic physiology that are circadian regulated include the rates of CO 2 assimilation and oxygen evolution, and photosynthetic light-harvesting reported using chlorophyll fluorescence analysis (Dakhiya et al. 2017;Hennessey and Field 1991;Litthauer et al. 2015;Schweiger et al. 1964). There are also circadian rhythms in the stomatal opening, which are functionally separable from rhythms of CO 2 assimilation because rhythms of assimilation continue when the intercellular partial pressure of CO 2 is held at a constant level to remove the influence of rhythmic stomatal opening upon assimilation (Hennessey & Field 1991). In addition to flowering plants, aspects of metabolism associated with photosynthesis are circadian regulated in some cyanobacteria and algae (Cano-Ramirez et al. 2018;Mitsui et al. 1986;Samuelsson et al. 1983;Schneegurt et al. 1994;Schweiger et al. 1964). Despite this evidence for the circadian regulation of photosynthesis, the mechanisms and evolution of this process remain less well understood.
The pervasiveness of circadian regulation of photosynthesis within flowering plants and unicellular photosynthetic organisms might suggest that this process is conserved throughout multicellular plant life, with a potentially ancient evolutionary origin. One approach to examine this idea is to investigate plant species that diverged from flowering plants at a relatively early stage of land plant evolution, because this can establish whether the mechanism might originate from a common ancestor. As a model to investigate this question, we selected the bryophyte liverwort Marchantia polymorpha. M. polymorpha is thought to have diverged from flowering plants about 400 million years ago (Delwiche & Cooper 2015;Kohchi et al. 2021), with liverwort-like plants occurring earlier in the fossil record than other land plant forms such as vascular plants (Edwards et al. 1995). M. polymorpha is a useful model to investigate questions concerning the evolution of circadian regulation because it has a circadian oscillator that shares some components with flowering plants (Lagercrantz et al. 2021;Linde et al. 2017), there is circadian regulation of a subset of the transcriptome and the position of its thallus lobes (Lagercrantz et al. 2020;Lagercrantz et al. 2021), and it can be used in experimental designs comparable to models such as Arabidopsis. While circadian rhythms and some aspects of circadian clock structure have been identified in M. polymorpha, it is not yet known whether the regulation of photosynthesis represents a conserved output from its circadian system.  Figure S5a). For inhibitor experiments, lincomycin (Alfa Aesar) or rifampicin (Melford Laboratories Ltd.) was added for the pulse amplitude modulation (PAM) chlorophyll fluorescence experiments at concentrations of 25, 50, and 100 μg/ml, with dimethylsulfoxide vehicle control where necessary. In M. polymorpha, rifampicin is known to inhibit plastid ribosomal RNA transcription (Loiseax et al. 1975), but does not inhibit photosynthesis (Mache & Loiseax 1972). The effect of lincomycin upon M. polymorpha is less well characterized, although it is known to bind to 50S ribosomes (Chang & Weisblum 1967), which are present in chloroplasts of M. polymorpha (Ohyama et al. 1986

| Measurement of delayed fluorescence
Delayed fluorescence analysis was performed as described previously (Gould et al. 2009;Rees et al. 2019). This involved image capture were imported into FIJI (version 1.53c) and regions of interest were selected and extracted using the multimeasure plugin to measure the integrated density of each region. Raw data before filtration is provided in Supporting Information: Dataset S2. The DF data were baseline and amplitude (BAMP) detrended using Biodare2, and a moving average was used to smooth traces from individual thalli.

| Data analysis
The DF data were smoothed using a 5 h moving average (Gould et al. 2009) and BAMP detrended using Biodare2 (Zielinski et al. 2014). These data are provided in Figure 1b,d,f and Supporting Information: Figures S1-S3 to exemplify variation in the DF rhythms across all thalli. Data detrending was necessary because rapid thallus growth during the experiment produced a continuously increasing raw signal. The data were subsequently analysed using Metacycle (Wu et al. 2016) to identify rhythmic thalli (q < 0.001; period range 18-34 h). Data that passed this filtration step were used for the calculation of rhythmic features and presented in Figure 1a,c,e,g-k and Table 1.
Rhythmic features of the data were quantified using the fast Fourier transform-nonlinear least-squares method (FFT-NLLS) algorithm within Biodare2 (Zielinski et al. 2014), and also with Metacycle (Wu et al. 2016). Statistical analysis of derived data was conducted using IBM SPSS Statistics v24 and SigmaPlot v14 for PAM chlorophyll fluorescence and DF data. Graphs were prepared using R, with a combination of the ggplot2, scales, rhesape2, cowplot, png, grid, RColorBrewer, wesanderson, colorspace, ggpubr, ggrepel, and patchwork packages. Heat maps were created using the R package Pheatmap. A 5 h moving average was applied to DF and PAM fluorescence data before analysis (Gould et al. 2009).
the PAM method (Dakhiya & Green 2022;Gould et al. 2009;Maxwell & Johnson 2000). When illuminated leaves are transferred to darkness, the potential for electrons to move down the electron transport chain is removed, and energy within the system is instead re-emitted as light and heat. The light emitted within the nanosecond range after the lights turn off is referred to as PF (Dakhiya & Green 2022;Maxwell & Johnson 2000). Light emitted within the millisecond to second range following lights-off is referred to as DF (Gould et al. 2009;Rees et al. 2019;Strehler & Arnold 1951).
Both DF and certain features of PF are circadian-regulated in a range of flowering plants and some algae (Cano-Ramirez et al. 2018; Dakhiya & Green 2022;Gould et al. 2009;Gyllenstrand et al. 2014).
One exception appears to be spruce trees (Picea abies), where the DF rhythm damps rapidly in accordance with a rapidly damping circadian oscillator (Gyllenstrand et al. 2014    spectrum (Linde et al. 2017). Alternatively, this might reflect differences in the preparation of thalli for the two methods, since thalli were grown directly from gemmae for DF imaging, whereas for PAM experiments, the thalli were cut into squares to prevent hyponasty. Y(II) and NPQ oscillated with opposing phases (Figure 2a,c), which is consistent with other reports (Cano-Ramirez et al. 2018), and expected because these measures represent competing fates for light energy absorbed by leaves.
Together with the measurements of DF, we conclude that freerunning rhythms of several photosynthetic outputs occur in M.
polymorpha. The DF data suggest that free-running rhythms are less robust and more variable under FRD conditions, but remain detectable in a proportion of thalli (Table 1) This did not occur within our experiment (Figure 3e), where instead the phase followed the zeitgeber cycle length (Figure 3a-e). The dependency of the period and phase of DF upon the T cycle length suggests that any underlying circadian clock control was masked by the zeitgeber cycle. In circadian biology, masking is the process whereby the apparent coupling of an observable rhythm to a zeitgeber leads to a shared period and phase between the observed rhythm and the zeitgeber, independent of any clock control (Aschoff 1999;Roenneberg et al. 2005).
Therefore, it appears that the interaction between the cycle of light and dark and the biochemical processes of light-harvesting tend to conceal the underlying effect of circadian regulation under these experimental conditions.

| Abrogation of circadian rhythms of photosynthesis in M. polymorpha by a potential inhibitor of chloroplast translation
While the canonical circadian oscillator underlies the daily timing of photosynthesis (Dodd et al. 2004(Dodd et al. , 2005, the mechanisms by which the circadian oscillator regulates photosynthesis remain poorly understood. One approach to identify potential cellular processes associated with circadian rhythms of photosynthesis is to use a pharmacological strategy because several chemicals are known to inhibit specific processes within chloroplasts. We combined a pharmacological approach with the monitoring of circadian rhythms of chlorophyll fluorescence to identify candidate processes associated with the circadian regulation of photosynthesis.
First, we tested whether an inhibitor of chloroplast translation can attenuate circadian rhythms of chlorophyll fluorescence in M. polymorpha. Lincomycin is an antibiotic that binds to the 50S subunit of bacteria-like ribosomes (Chang & Weisblum 1967) and is thought to inhibit chloroplast but not mitochondrial protein synthesis (Sullivan & Gray 1999;Zhao et al. 2018). We supplemented the growth media with a range of lincomycin concentrations (   measures of photosynthesis are ratiometric, so the decreased amplitude is unlikely to be caused by a lower fluorescence signal in the presence of lincomycin. While we cannot exclude the possibility that lincomycin in M. polymorpha affects processes other than plastid translation, our data suggest that chloroplast translation might be required to maintain circadian rhythms of photosynthetic light- harvesting.
Next, we tested the hypothesis that an inhibitor of chloroplast transcription can attenuate circadian rhythms of photosynthesis in M.
polymorpha. In Arabidopsis, mutants of a circadian-regulated sigma submit of plastid-encoded plastid RNA polymerase (PEP) known as SIG5 can alter the phase of circadian rhythms of delayed fluorescence (Noordally et al. 2013). Therefore, we reasoned that chloroplast transcription by PEP might contribute to circadian rhythms of photosynthesis in M. polymorpha. To test this, we supplemented the growth media with rifampicin, which is thought to inhibit chloroplast transcription by one form of PEP (Pfannschmidt & Link 1997;Surzycki 1969). We measured rhythms of Y(II) and NPQ, under

| DISCUSSION
Here, we present evidence for the circadian regulation of two proxies for photosynthetic activity in M. polymorpha. Our experiments cannot prove that circadian rhythms of photosynthesis occurred in early land plants, but the presence of these rhythms in both bryophytes and angiosperms suggests that they might be well-conserved. Given that reporters of photosynthetic activity are also circadian regulated in aquatic photosynthetic organisms such as cyanobacteria, uni-and multicellular algae (Cano-Ramirez et al. 2018;Hastings et al. 1961;Okada et al. 1978;Schneegurt et al. 1994;Schweiger et al. 1964;Sorek & Levy 2012;Sorek et al. 2013;Sweeney and Haxo 1961), our study suggests that the circadian regulation of photosynthesis was conserved from aquatic ancestors during the terrestrialization of plants. One interpretation is that circadian regulation of photosynthesis confers a selective advantage to all photosynthetic life. The nature of the advantage and degree of selection pressure for this might depend upon the ecological niche occupied by the organism (Hellweger et al. 2020), particularly because certain species (e.g., Picea abies) appear to lack these rhythms (Gyllenstrand et al. 2014).
Our experiments identified that under FRL conditions, the period of the rhythm of delayed fluorescence was not uniform between replicate thalli, with this variation leading to phase desynchrony Therefore, circadian rhythms of the reporters of photosynthesis described here might be due entirely to the circadian regulation of the biochemical reactions of photosynthesis, rather than changes in the rate of CO 2 and O 2 exchange with the atmosphere. Nevertheless, we cannot rule out alterations in conductivity to gas exchange through diel or circadian changes in cell turgor or potential "motor cells" within liverwort air pores (Walker & Pennington 1939). Such changes might influence the measures of photosynthesis that we used because the presence of the pores and air spaces within the thallus appears to reduce its resistance to CO 2 diffusion (T. G. A. Green & Snelgar 1982).
The reduction in the robustness of rhythms of Y(II) caused by lincomycin could suggest that chloroplast translation is necessary for circadian rhythms of this reporter of photosynthesis. Lincomycin inhibits the synthesis of PSII D1, PSII D2, and RbcS proteins, and the expression of photosynthesis-associated nuclear-encoded genes in angiosperms (Bachmann et al. 2004;Karpinska et al. 2017;Mulo et al. 2003). Perhaps inhibiting the expression of these proteins prevents the replacement of PSII components, with ensuing impacts upon circadian cycles of lightharvesting. The absence of any effect of rifampicin on circadian rhythms of Y(II) or NPQ, contrasting the effect of lincomycin, could suggest that rhythms of these parameters involve genes transcribed by nuclear-encoded plastid RNA polymerase (NEP) rather than rifampicin-sensitive PEP, there is compensation by NEP for reduced PEP activity, or specific roles for the rifampicin sensitive-and insensitive forms of PEP (Pfannschmidt and Link 1994). While this suggests that protein synthesis in chloroplasts is necessary for the circadian regulation of photosynthesis, it does not identify that this is a rate-limiting regulatory step that underlies the clock control of the process. It also supports the idea that M. polymorpha represents an informative model for the investigation of mechanisms underlying the circadian regulation of physiology, particularly considering the range of gene-editing tools available. While