The Circadian-clock Regulates the Arabidopsis Gravitropic Response

Long-term missions in space to the moon and beyond will require efficient plant growth and food production in orbit. Several factors, such as light intensity, exposure to pathogens, and air quality, will vary on crewed missions. However, perhaps the most extreme difference is the lack of gravity. On Earth, gravity regulates the direction of plant growth and is important for nutrient, water, and light acquisition.

Life is well-adapted to the 1 g environment of the Earth. Plants use the gravity vector as a cue to establish their orientation; roots grow in the same direction as the gravity vector, while shoots grow in the opposite direction (Kiss, 2000). Deviation from a 1 g environment will have cascading effects on plant physiology. Circadian rhythms are regular oscillations of an organism's physiology with a period of approximately 24 h. This coordination with the Earth's 24 h rotation is regulated by an endogenous timekeeper, the circadian clock. In the model plant, Arabidopsis thaliana, circadian rhythms regulate a suite of physiological processes, including transcription, photosynthesis, growth, and flowering (McClung, 2019). In this work, we investigated if this control extended to the regulation of gravitropic responses.

Circadian systems function through a series of negative feedback loops that are temperature compensated. The circadian clock functions by modulating signaling pathways, and maintaining a period of about 24 h in the absence of external cues such as light and temperature cycles. However, an organism's circadian period is plastic and can be entrained via external signals. For example, circadian-regulated phenotypes in Arabidopsis persist in constant light (CL) and temperature conditions yet the circadian period decreases with increased light intensity, while reduced sugar has the opposite effect (Somers et al., 1998; Haydon et al., 2013).

The plant core circadian clock consists of a series of negative transcription-translation feedback loops. Components of the core clock include MYB-like transcription factors Circadian-clock Associated 1 (CCA1) and Late elongated HYpocotyl (LHY) that peak in abundance near dawn. CCA1 and LHY function as transcriptional repressors of Timing Of CAB2 expression1 (TOC1) and members of the Evening Complex (EC). The EC includes the clock genes LUX arrythmo (LUX) and EarLy Flowering 3 & 4 (ELF3, ELF4) (McClung, 2019). CCA1 and LHY bind the evening element motif of several clock-controlled genes repressing their transcription in the morning. Targets of CCA1 and LHY include the EC members: ELF3 (Lu et al., 2012), LUX (Zhang et al., 2019a), BOA/NOX (Dai et al., 2011).

CCA1 and LHY expression is tightly regulated by the combined activity of the Pseudo Response Regulator (PRRs) family, ensuring that expression of CCA1/LHY occurs in the early morning (Nakamichi et al., 2012, 2010). The expression of each PRR family member peaks sequentially throughout the day in the order PRR9, PRR7, PRR5, PRR1/TOC1. The latter reaches an expression peak near dusk. The EC transcription factors, LUX and BOA regulate the expression of PRR7 and PRR9. LUX and BOA bind the promoters of PRR7 and PRR9 and other circadian-regulated genes.

Three key circadian-controlled pathways that may be relevant to root gravitropism are starch metabolism, starch catabolism, and auxin signaling. At the level of the whole leaf, starch metabolism and catabolism are gated by the circadian clock. Two models have been proposed for the regulation of starch homeostasis by the clock. The first suggests that the clock operates as a timekeeper, calculating the appropriate starch degradation rate based on the available starch at dusk and the time until dawn (Seki et al., 2017). The second model proposes that a starch monitoring signal adjusts the phase of starch degradation activity (Webb and Satake, 2015).

The root gravitropic response has been extensively studied. Critical steps in gravity sensing and signaling are well characterized (for recent reviews, see (Sato et al., 2015; Su et al., 2017; Vandenbrink and Kiss, 2019; Muthert et al., 2019; Telewski, 2006). In plant root, gravity perception primarily occurs in specialized columella cells located in the root cap (Blancaflor et al., 1998). The sedimentation of dense, starch-filled organelles called statoliths located within the columella cells triggers a signaling cascade that leads to root curvature in response to gravity. Starch-deficient mutants, which in turn have smaller statoliths, exhibit reduced sensitivity to gravity, while mutants with larger statoliths are hypersensitive to gravity (Kiss et al., 1989, 1996; Fitzelle and Kiss, 2001; Vitha et al., 2007). Although circadian and diel variation controls leaf starch levels, it is unknown if such rhythmic cycling of starch levels also occurs in the statoliths (Lu et al., 2005).

The plant growth regulator auxin is an important player in the gravitropic response. When a plant is placed on its side horizontally, lateral gradients of auxin develop at the root tip and move to the elongation zone resulting in the downward curvature of the root to enable the plant root to reorient to vertical (Band et al., 2012). The formation of the asymmetric auxin gradient is facilitated by proteins involved in auxin influx (AUX and LAX) and efflux (PIN-formed, PINs) (Friml et al., 2002; Band et al., 2012) reviewed in (Sato et al., 2015).

In addition to mediating the differential growth response, an intriguing recent report (Zhang et al., 2019b) suggests that auxin may also play a role in gravity sensing by regulating starch synthesis within the columella cells. Auxin levels at the root tip were correlated with starch accumulation; furthermore, auxin was implicated in the expression of starch synthesis genes via the TIR1/AFB auxin receptor AXR3/IAA17 auxin signaling pathway.

As early as the 1930s, it was recognized that plant sensitivity to auxin varied according to the time of day (Went and Thimann, 1937). While many transcripts are under circadian control, genes involved in auxin signaling are disproportionately regulated by the clock (Covington and Harmer, 2007). Furthermore, in both diel and constant environmental conditions, transcriptional responses to auxin treatment are more robust at dawn. This differential response culminates in altered growth responses to auxin that depend on the time of treatment. Finally, there is evidence that auxin may feedback into clock outputs, indicating a two-way relationship between auxin and the clock (Covington and Harmer, 2007).

The gating of starch metabolism, starch catabolism, and auxin signaling raises the possibility that the time of the day that a change in the direction or magnitude of the gravity vector occurs could impact the response. Beyond the potential for the circadian influence, in diel conditions, daily changes in light and temperature could also affect the gravitropic response differently throughout the day. The interaction between the timing of the gravitropic stimulus and the response could have important consequences for experimental design and understanding the mechanisms of gravitropic signaling. Therefore, determining how temporal factors such as time of day and the circadian clock influence plant gravity sensing and responses will be necessary for cultivation outside of the typical Earth environment.

The first goal of this work was to characterize the role of the circadian clock on the root gravitropic response. Root bending assays were carried out which involved turning vertically grown seedlings by 90° over a 24-hour time course and recording time-course images.

We analyzed the root phenotype of seedlings that experienced gravistimulation at different times of day in diel and constant environmental conditions and examined genotypes with disrupted circadian clocks.

The second goal of this study was to determine if root growth responses were affected by simulated microgravity. Ground-based microgravity simulators provide cost-effective means to study how microgravity affects plant growth and development. Several simulators are available for use including 2-dimensional (2D) clinostats and Random Positioning Machines (RPM). A 2D clinostat rotates samples around one axis. The RPM is a specialized type of clinostat that rotates samples around two axes and the arms are capable of rotating at different speeds (Kiss et al., 2019). We designed a lighting system that allowed for two separate lighting regimes on the RPM so that direct comparisons can be made between different light conditions on the same RPM run. Our results indicate that root gravitropism is affected by the time of day the experiment is performed and is controlled, at least in part, by the circadian clock.

Root Reorientation Response Differs based on the Time of Day the Gravistimulus Occurs Gravistimulation experiments were performed by turning vertically grown Wild Type (WT) Arabidopsis seedlings by 90°. A series of plates were utilized and each plate was rotated horizontally only one time every 2 h after dawn, Zeitgeber time (ZT) 0. Seedlings were imaged every 15 min from 2 h before turning until 26 h after reorientation. From the image data, root angle and root growth rate were recorded. Root growth rate varied throughout the 24 h time course, increasing throughout the night and decreasing throughout the day. These findings are similar to previous studies investigating Arabidopsis growth kinetics (Yazdanbakhsh and Fisahn, 2009). Even though the growth rates over 2 h varied at specific times of day, when the growth is examined over the entire 24 h period, the cumulative root growth in that 24 h period remained similar for all plates independent of the time of day they were rotated horizontally (Figure S1A). Therefore, we measured the bending angle after a 24 h period. This avoids the confounding factor of different growth rates for each ZT at times <24 h (Yazdanbakhsh et al., 2011) (Figure S1B). This detailed analysis of root bending revealed that the degree of root reorientation after 24 h depended on the time of day the gravistimulation was initiated (Figure 1). Plants reoriented at dusk (ZT12) showed the largest response to gravity, while plants rotated horizontally at midday (ZT2, 4, and 6), just before dawn (ZT18, 20, and 22), and after dawn showed a reduced response (Figure S1C). These results indicate that the response to gravity is affected based on the time of day the plants were reoriented (Figure 1).

Figure 1

ZT12 and ZT22 Exhibited the Largest Difference in their Bending Response

We selected two time points with significant differences in their root bending angle for further study, ZT12 and ZT22. These time points had different bending angles 24 h post gravistimulation (Tukey posthoc, p-value < 0.05) and showed consistently different responses across multiple replicates (Figure 2A). Although the total root growth over a 24 h period was similar across all time points (Figure S1A), the root growth rate relative to the time of gravistimulation varied for each time point (Figure S1B). To visualize the connection between root growth rate and the rate of bending, we first calculated the progression of the bending angle every 4 h across the 24 h after the plate was rotated horizontally for plants reoriented 90° at ZT12 and ZT22 (Figure 2A).). Plates rotated horizontally at ZT12 appeared to have a slow change in the bending angle until about 12 h post-turn (Figure 2A). In contrast, plates rotated horizontally at ZT22 responded with an increasing bending angle until about 16 h post-turn, but after that there was almost no additional change in the angle. We next examined change in the growth rate throughout the 24 h period post-turn (Figure 2B). The growth rate was measured in 4-hour increments as this is the smallest unit of time where we could consistently detect a change in growth across all 4 h windows in the 24 h day. Plants rotated horizontally at ZT12 were just entering their period of maximal growth, whereas plants rotated horizontally at ZT22 entered a period of slower root growth right after turning. From the bending angle and the 4 h growth rate we determined the adjusted bending angle by calculating the change in the root angle/growth rate (Figure 2C). The rate of bending/growth for plants rotated horizontally at ZT12 and ZT22 was different 8 h and 24 h post-turn (Figure 2C). The early growth rate immediately after turning is steady at ZT22. When the growth rate increases about 16 h later, there is limited reorientation during this period of rapid growth. The opposite is true for ZT12 (Figure S1A). Immediately after turning, the growth of plants reoriented at ZT12 is slow but rapidly increases. The consistent bending rate throughout this period of rapid growth results in a greater overall bending response in ZT12 compared with ZT22 (Figures S1B, 2).

Figure 2

Variation in Responses Persists in CL and Clock Mutants

The time of day-dependent variation in the root bending response to gravity could be due to either the effects of light and dark conditions the plants were in at the time of root reorientation or the endogenous circadian clock. To evaluate the potential for circadian control of the root-bending response, we repeated the root bending experiment in clock mutants and after WT plants were shifted to growth in constant environmental conditions. Clock mutants were grown in the same conditions as WT plants above. For constant condition experiments, plants were grown in 12 h light:12 h dark for 5 days. Two days before the reorientation event, the lights were left on at night until the completion of the experiment. Thus, the plants were in CL for 48 h before the reorientation and remained in continuous light after turning. Starting at ZT48, reorientation was performed every 2 h on a set of plates. Each plate of vertically grown Arabidopsis seedlings was turned by 90°. An image was taken 1 h before turning and every 15 min after turning for 24 h. Root angles and growth rate were measured after 24 h. The bending angle was compared between the different ZT time points when reorientation occurred (Figure 3). The difference in reorientation response at ZT12 and 22 observed in 12 h light:12 h dark cycles persisted in constant environmental conditions suggesting that the circadian clock mediates this differential response.

Figure 3

Root Growth on a Two-Dimensional Clinostat is Altered in Some Genotypes with Disrupted Circadian Clocks

To further evaluate the role of the circadian clock in the gravitropic response, we compared WT plants and plants with an altered circadian clock using a two-dimensional (2D) clinostat (Wang et al., 2016). Root tip positions were compared between pre- and post-clinorotation to capture the differences in root response. The final root positions relative to the position before clinorotation were plotted (Figure 4A, B). Generally, plants grown on a clinostat exhibit random skewed root growth with significant deviation from vertical. We observed different root growth patterns post-clinorotation according to genotype. In 12 h light:12 h dark conditions, differences between genotypes were primarily noted between plants with constitutively high CCA1 expression (CCA1 OX) and all other genotypes (Figure 4A). In most genotypes, individual roots deviated on either side of the vertical axis with an even distribution. CCA1 OX and elf3-2 roots showed less deviation from vertical than WT, with CCA1 OX exhibiting the least variation under 12 h light: 12 h dark. Two measures were adopted to calculate this deviation, namely the vertical growth index (VGI) (Grabov et al., 2005a), and straightness, Lc/L (Vaughn and Masson, 2011) (Figure 5). In 12 h light:12 h dark conditions, CCA1 OX plants had significantly straighter growth as indicated by a larger Lc/L than WT plants. The CCA1 OX plants grew almost straight down during 2D clinorotation as indicated in the scatter plot (Figure 4A) resulting in a VGI value approaching one (Figure 5). However, CCA1 OX plants had less overall root growth in 12 h light:12 h dark conditions (Figure 6), as discussed below.

Figure 4

Figure 5

Figure 6

When placed on the 2D clinostat for 24 h in CL conditions (Figure 4B), VGI and Lc/L were not significantly different to 12 h light: 12 h dark conditions for WT roots (p-value < 0.05) indicating the deviation from vertical was similar to 12 h light:12 h dark conditions. However. CL conditions significantly affected how the elf3-2 genotype responded to growth on the 2D clinostat. The elf3-2 genotype had more random or skewed growth in CL compared to 12 h light:12 h dark as indicated by a decreasing Lc/L (12:12 = 0.94; CL = 0.86, p-value < 0.05). CCA1 OX seedlings also exhibited significantly more deviations from vertical growth in CL compared to 12 h light: 12 h dark (Lc/L in 12:12 = 0.98, in CL = 0.90) and a significantly smaller VGI (12:12 = 0.92, CL = 0.71). prr7 had a similar Lc/L in 12 h light:12 h dark and CL conditions but had significantly larger VGI in CL (12:12 = 0.73, in CL = 0.90). Overall, these differences indicate that growth in CL conditions did not significantly affect the response of WT plants but did disrupt CCA1 OX and elf3-2 genotypes which are arrhythmic plants in constant conditions.

In fact, when comparing genotypes, CCA1 OX went from growing straighter in 12 h light:12 h dark on the 2D clinostat (Lc/L 0.98) with a larger VGI (0.92) than WT (Lc/L 0.94; VGI 0.82) to being less straight (Lc/L 0.90) with a smaller VGI (0.71). elf3-2 also went from being similar to WT in 12 h light:12 h dark to lower Lc/L and VGI values indicating it was deviating from vertical growth more than WT in CL conditions (Figures 4 and 5). WT and prr7 genotypes were similar, with growth skewed to either side more evenly than CCA1 OX or elf3-2 (Figures 4, 5).

Observed Differences in Gravitropic Responses cannot be Explained by Amyloplast Differences Based on Starch Staining

Amyloplasts are important for gravity sensing. Starch levels in aerial tissues (leaves) vary throughout the day (Li et al., 1992). However, it is not known if this diel variation in total starch levels impacts the amyloplast. Therefore, we asked whether daily fluctuations in amyloplast starch occur and could explain the observed variation in gravity response throughout the day (Figures 1–3). Roots were harvested from WT, CCA1 OX, and elf 3-2 plants and stained for starch at ZT12 and ZT22 in 12 h light: 12 h dark to evaluate gross changes in starch levels (Figure 7A). If the amyloplast starch levels showed similar diel and circadian variation as leaf starch levels we would anticipate that ZT12, at the end of the day, would have significantly higher levels than ZT22, at the end of the dark period. However, as seen in the representative images (Figure 7A) WT roots at ZT12 and ZT22 showed comparable staining. Quantification of the area stained also showed there was no significant difference between ZT12 and ZT122 in the percent area of starch staining (p-value < 0.05, Figure 7B).

Figure 7

In contrast to WT, elf3-2 shows a significant variation in starch staining between the ZT12 and ZT22 timepoints, with reduced staining at ZT 22 (t-test p-value < 0.001). CCA1 OX showed a similar trend (Figure 7A), however, the small size of CCA1 OX roots in 12 h light: 12 h dark (e.g., Figure 6) makes an accurate comparison difficult. These results suggest that there are fluctuations in root amyloplast starch in mutants with disrupted circadian clocks that are not observed in WT plants which warrant further investigation with more quantitative detection methods over a finer time course.

Plants Entrained to Different Circadian Time Points Exhibit Subtle Root Phenotype Differences in a Simulated Microgravity Environment

Seedlings were transferred at two different time points (ZT12 and ZT22) to a RPM, Airbus Defense & Space, to examine if the time of day affected the response to simulated microgravity. Two sets of seedlings were entrained to time points shifted by 10 h so that the RPM treatment was carried out simultaneously on the same RPM for both time points. After 48 h, seedlings were removed from the RPM and imaged. Controls were grown vertically under identical conditions. The roots of seedlings grown on the RPM exhibited disorganized growth patterns compared to the vertical controls, growing in various directions, including out of the plate and into the media (Figure 8A). Our single plane images could not capture the overall growth and response in three dimensions (Figure 8A). Therefore, we calculated the variation only for roots that grew laterally in a single plane on the agar surface. Despite this limitation, we observed that seedlings grown on the RPM starting at ZT12 exhibited more dramatic skewing or slanted growth phenotypes than plants started at ZT22 (Figures 8B, S2). This trend was consistent between four independent RPM experiments.

Figure 8

This study shows that the plant response to gravistimulation depends on the time of day the stimulation occurs. There were substantial differences between the reorientation response at the end of the day and the end of the night. This altered response persists in CL conditions suggesting a role for the circadian clock in the perception of the gravistimulus or downstream signaling pathways. These findings suggest that attention to time can be a critical component in experimental design for gravitropic experiments. Moreover, the variation between different time points could be exploited to understand how the same stimulus in the same genotype manifests in altered responses. These findings encourage further investigation to evaluate if the time of day exacerbates or mitigates the effects of a change in the magnitude of the gravity vector in other tissues and other species, including humans during spaceflight.

We observed the significant influence of the time of day and the circadian clock on the plant response to a change in gravity using three different methods of altering gravity. First, root reorientation in response to gravistimulation was greatest 12 h after dawn and weakest 2 h before dawn (Figure 1). This pattern persisted in the absence of day/night cues and was lost in CCA1 OX plants suggesting that this variation is circadian regulated.

Root growth responses on a 2D clinostat were also disrupted in plants with an altered circadian clock. CCA1 OX plants showed less deviation from vertical growth than wild-type plants in entraining condition but was more responsive in constant environmental conditions (Figure 3). The elf3-2 mutant showed unidirectional skewing on the 2D clinostat in constant environmental conditions, unlike wild-type plants and the other circadian-disrupted genotypes, which had a more random orientation (Figure 4). The observed root-orientation phenotype for elf3-2 only in constant environmental conditions is consistent with other phenotypes in this mutant. In these free-run conditions, the elf3-2 mutant has an arrhythmic phenotype that is masked in entraining conditions such as 12 h light: 12 h dark cycles (Covington et al., 2001). The prr7 mutant did not show detectable differences in root reorientation consistent with its circadian phenotype, which manifests as a phase shift and not a complete disruption of rhythmicity like the CCA1 OX and elf3-2 genotypes (Farré and Kay, 2007).

Root length also varied between these genotypes. For example, when the plants were grown in light-dark cycles, CCA1 OX plants had reduced root growth compared to wild-type (12:12, Fig 6). At first glance, this seems to explain the difference in root orientation; with less growth in CCA1 OX plants, there is less variation in their growth on the 2D clinostat. However, when grown in CL, CCA1 OX root growth remained similar to 12 h light:12 h dark levels, at about 55% WT root growth in 12 h light: 12 h dark (Figure 6). Even with this reduced growth, the CCA1 OX gravitropic responses seemed to surpass WT levels in either 12 h light: 12 h dark or CL (Figure 5). WT root growth in CL was significantly reduced (Figure 6), but there was no significant difference in the growth responses in the two conditions on the 2D clinostat as measured by VGI or Lc/L (Figure 5). These results suggest interactions between light cycles, root growth, and the gravitropic response. The connection between auxin signaling, which is gated by the circadian clock (Covington and Harmer, 2007; Rawat et al., 2009), and growth could be one mechanism controlling this interaction.

Growth on a RPM indicated that the time of day the plants first perceive a loss of 1 g gravity vector affects their response even 48 h later (Figure 8). While a full, detailed quantification of the changes on the RPM will require monitoring the root growth in three dimensions, in this first evaluation, we observe differences in the root orientation after 48 h depending on what time of day the plants were transferred to the RPM. The plants that first sensed simulated microgravity at ZT12 bent more than those that first perceived it at ZT22.

Root growth and orientation are fundamentally overlapping responses. Roots can only change in orientation through new growth. In the root reorientation experiments, the total root growth after 24 h was the same for all plants, independent of when they were rotated horizontally (Figure S1A). However, previous work has shown that this growth is not even across the 24 h-cycle (Yazdanbakhsh and Fisahn, 2009, 2011; Yazdanbakhsh et al., 2011). Root growth rate peaks in the night and slows just after dawn. This timing coincides with starch utilization with the roots growing as the plant breaks down starch and growth cessation during the day as starch is accumulated. The observed reorientation responses were consistent with this growth pattern. Plants rotated horizontally at ZT12, just before the maximal growth rate, had the largest response, while those at ZT22, just before the period of reduced growth, had a significantly reduced response. This suggests that the timing between the window of graviperception and the growth cycle is important. The plants only respond at their maximum if the change in gravity vector occurs in the window before the period of rapid growth. If the change in gravity happens outside that window, the response is not as robust even though the plants still go through the same rapid growth period the following night.

Plant roots perceive a change in gravity, at least in part through the amyloplasts (Hashiguchi et al., 2013). Gross differences in starch accumulation have been observed previously in gravitropic mutants, such as the extreme starch reduction in the pgm mutant which is detectable by staining (Caspar and Pickard, 1989; Tanimoto et al., 2008). Prior research has shown that leaf starch accumulation is diel and circadian-regulated (Li et al., 1992). Starch levels accumulate during the day and fall at night (Seki et al., 2017; Li et al., 1992). This pattern is disrupted in circadian mutants (Graf et al., 2010). Thus, if amyloplast starch levels varied throughout the day, this could explain the temporal variation in gravitropic responses we observe. However, in WT plants we do not observe a corresponding daily alteration in amyloplast starch. Starch staining at ZT12 and ZT22 in 12 h light: 12 h dark showed no significant difference in WT plants. Previous analysis by Kiss and colleagues saw no differences in the gravitropic response in mutants with 50 or 60% starch reduction (Kiss et al., 1996). Although we did observe some variability in starch staining between individual WT seedlings, these differences are minimal compared to a 60% reduction in starch content. Therefore, it is unlikely that a difference in root amyloplast starch is a major factor regulating the time of day differences in gravitropic response. Since our approach does not quantify the total number of cells containing amyloplasts, quantitative detection methods would be required to determine if there are subtle changes in amyloplast density or number at different times of the day.

The elf3-2 genotype showed a significant variation in starch staining between ZT12 and ZT22, with lower levels at ZT22 (Figure 7). Thus, although amyloplasts do not show the same circadian regulation observed in leaf starch levels, there may be a connection between the circadian clock and starch levels in the amyloplast. Such differences would require a more quantitative evaluation of the amyloplast starch levels than the starch staining method employed in this study.

We have identified time of day-dependent differences in the root gravitropic response that persists through three different methods of measuring the plants’ responses to gravity. These results highlight the importance of paying attention to the time of day experiments are performed. The results presented here have implications for spaceflight and suggest that the transition to different gravity levels could be optimally timed.

This work also indicates that the circadian clock is an upstream regulator of the gravitropic response. Such clock-based control has also been observed in Neurospora crassa and Drosophila melanogaster suggesting a conserved interaction between clock-regulation and gravity across species (Ma et al., 2015; Ferraro et al., 1989; Sulzman et al., 1984). Moreover, these studies in other species also indicate that there is a reciprocal interaction between gravity and the clock; altered gravity can affect the circadian rhythms in Neurospora crassa (Ferraro et al., 1989; Sulzman et al., 1984), Drosophila melanogaster (Ma et al., 2015), and Rattus norvegicus (Fuller et al., 2003). The observations presented here provide the first step to ensure correct experimental design for efforts to examine the effects of altered gravity on the plant circadian clock, which has not previously been reported.