Sensory conflict disrupts circadian rhythms in the sea anemone Nematostella vectensis
Curation statements for this article:-
Curated by eLife
eLife assessment
Understanding the integration and contribution of different combinations of environmental cues to the synchronization of the daily oscillator is important, because it provides insight into how organisms might be able to distinguish (and weight) between irregular (or in the tidal zone highly complex) versus regular individual daily changes of light and temperature. The study, which is thoroughly conducted and provides an impressive amount of experimental and analytical work, dissects the effects of sensory conflict on behavior and gene expression rhythms.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Circadian clocks infer time of day by integrating information from cyclic environmental factors called zeitgebers, including light and temperature. Single zeitgebers entrain circadian rhythms, but few studies have addressed how multiple, simultaneous zeitgeber cycles interact to affect clock behavior. Misalignment between zeitgebers (‘sensory conflict’) can disrupt circadian rhythms, or alternatively clocks may privilege information from one zeitgeber over another. Here, we show that temperature cycles modulate circadian locomotor rhythms in Nematostella vectensis , a model system for cnidarian circadian biology. We conduct behavioral experiments across a comprehensive range of light and temperature cycles and find that Nematostella ’s circadian behavior is disrupted by chronic misalignment between light and temperature, which involves disruption of the endogenous clock itself rather than a simple masking effect. Sensory conflict also disrupts the rhythmic transcriptome, with numerous genes losing rhythmic expression. However, many metabolic genes remained rhythmic and in-phase with temperature, and other genes even gained rhythmicity, implying that some rhythmic metabolic processes persist even when behavior is disrupted. Our results show that a cnidarian clock relies on information from light and temperature, rather than prioritizing one signal over the other. Although we identify limits to the clock’s ability to integrate conflicting sensory information, there is also a surprising robustness of behavioral and transcriptional rhythmicity.
Article activity feed
-
-
Author Response
Reviewer #2 (Public Review):
This is an interesting study investigating the effects of sensory conflict on rhythmic behaviour and gene expression in the sea anemone Nematostella vectensis. Sensory conflict can arise when two environmental inputs (Zeitgeber) that usually act cooperatively to synchronize circadian clocks and behaviour, are presented out of phase. The clock system then needs to somehow cope with this challenge, for example by prioritising one cue and ignoring the other. While the daily light dark cycle is usually considered the more reliable and potent Zeitgeber, under some conditions, daily temperature cycles appear to be more prominent, and a certain offset between light and temperature cycles can even lead to a breakdown of the circadian clock and normal daily behavioural rhythms. Understanding the …
Author Response
Reviewer #2 (Public Review):
This is an interesting study investigating the effects of sensory conflict on rhythmic behaviour and gene expression in the sea anemone Nematostella vectensis. Sensory conflict can arise when two environmental inputs (Zeitgeber) that usually act cooperatively to synchronize circadian clocks and behaviour, are presented out of phase. The clock system then needs to somehow cope with this challenge, for example by prioritising one cue and ignoring the other. While the daily light dark cycle is usually considered the more reliable and potent Zeitgeber, under some conditions, daily temperature cycles appear to be more prominent, and a certain offset between light and temperature cycles can even lead to a breakdown of the circadian clock and normal daily behavioural rhythms. Understanding the weighting and integration of different environmental cues is important for proper synchronization to daily environmental cycles, because organisms need to distinguish between 'environmental noise' (e.g., cloudy weather and/or sudden, within day/night temperature changes) and regular daily changes of light and temperature. In this study, a systematic analysis of different offsets between light and temperature cycles on behavioural activity was conducted. The results indicated that several degrees of chronic offset results in the disruption of rhythmic behaviour. In the 2nd part of the study the authors determine the effect of sensory conflict (12 hr offset that leads to robust disruption of rhythmic behaviour) on overall gene expression rhythms. They observe substantial differences between aligned and offset conditions and conclude a major role for temperature cycles in setting transcriptional phase. While the study is thoroughly conducted and represents and impressive amount of experimental and analytical work, there are several issues, which I think question the main conclusions. The main issue being that temperature cycles by themselves do not seem to fulfil the criteria for being considered a true Zeitgeber for the circadian clock of Nematostella.
Major points:
Line 53: 'However, many of these studies did not compare more than two possible phase relationships.....'. Harper et al. (2016) did perform a comprehensive comparison of different phase relationships between light and temperature Zeitgebers (1 hr steps between 2 and 10 hr offsets), similar to the one conducted here. I think this previous study is highly relevant for the current manuscript and -- although cited -- should be discussed in more detail. For example, Harper et al. show that during smaller offsets temperature is the dominant Zeitgeber, and during larger sensory conflict light becomes the dominant Zeitgeber for behavioural synchronization. Only during a small offset window (5-7 hr) behavioural synchronization becomes highly aberrant, presumably because of a near breakdown of the molecular clock, caused by sensory conflict. Do the authors see something similar in Nematostella? Figure 3 suggests otherwise, at least under entrainment conditions, where behaviour becomes desynchronized only at 10 and 12 hr offset conditions. But in free-run conditions behaviour appears largely AR already at 6 hr offset, but not so much at 4 and 8 hr offsets (Table 2). So there seems to be at least some similarity to the situation in Drosophila during sensory conflict, which I think is worth mentioning and discussing.
We have added a more detailed discussion of our results in the context of Harper et al. 2016 (L468-476).
Line 111: The authors state that 14-26C temperature cycle is 'well within the daily temperature range experienced by the source population'. Too me this is surprising, as I was not expecting that water temperature changes that much on a daily basis. Is this because Nematostella live near the water surface, and/or do they show vertical daily migration? Also, I do not understand what is meant by '...range of in situ diel variation (of temperature)'. I think a few explanatory words would be helpful here for the reader not familiar with this organism.
In fact, one of our motivations for studying temperature is that Nematostella naturally experience extreme temperature variation. The data we cite (Tarrant et al. 2019) are from in-situ water measurements. Nematostella live in extremely shallow water (in salt marshes), and the local population in Massachusetts experience wide swings in temperature due to the temperate latitude.
We have added this information to the Introduction (L88-90), and we also added a discussion of Nematostella’s ecology in the Discussion section (L591-654).
Lines 114-117: I was surprised that clock genes can basically not be synchronized by temperature cycles alone. Only cry2 cycled during temperature cycles but not in free-run, so the cry2 cycling during temperature cycles could just be masking (response to temperature). Later the authors show robust molecular cycling during combined LD and temperature cycles (both aligned and out of phase), indicating that LD cycles are required to synchronize the molecular clock. Moreover, a previous study has demonstrated that LD cycles alone (i.e., at constant temperature) are able to induce rhythmic molecular clock gene expression (Oren et al. 2015). Similarly, the free running behaviour after temperature cycles does not look rhythmic to me. In Figure 2A, 14-26C there is at best one peak visible on the first day of DD, and even that shows a ~6 phase delay compared to the entrained condition. After the larger amplitude temperature cycle (8:32C) behaviour looks completely AR and peak activity phases in free-run appear desynchronized as well (Fig. 2B). Overall, I think the authors present data demonstrating that temperature cycles alone are not sufficient to synchronize the circadian clock of Nematostella. One way to proof if the clock can be entrained is to perform T-cycle experiments, so changing the thermoperiod away from 24 hr (e.g., 10 h warm : 10 h cold). If in a series of different T-cycles the peak activity always matches the transition from warm to cold (as in 12:12 T-cycles shown in Fig. 1A) this would speak against entrainment and vice versa.
Thank you for these thoughtful comments and constructive suggestions. We have conducted an additional experiment, which provides further evidence that temperature cycles can, in fact, synchronize the circadian clock. To do this, we measured the behavior of animals entrained in cycles with a short (12h) period, half the length of a circadian period. This takes advantage of a phenomenon called “frequency demultiplication”, in which organisms in 12h environmental cycles display both 12h and 24h components--essentially, the clock perceives every other cycle as a “day” (Bruce, 1960; Merrow et al., 1999). The important thing is that the 24h behavioral component can only occur if the signal is entraining a circadian clock—otherwise, we would only observe a directly-driven 12h behavior pattern.
We first show that this phenomenon occurs with 6:6 LD cycles—which we expected, because we know light is a zeitgeber. We then show that animals entrained to a temperature cycle with a 12h period also display 24h behavioral rhythms—and in fact the 24h component is stronger than the 12h component. We believe this is strong evidence that temperature is a bona fide zeitgeber in this system. This experiment is now explained in the Results (L127-154) and in Figure 2–Figure supplement 1.
In terms of our original data, the reviewer is correct that the statistically-detectable free-running rhythms were weak and not visually obvious). Our confidence in thermal entrainment came from the fact that some individual animals had 24h rhythmicity in free-run, even if the signal was weak in the mean time series—this suggested that temperature must be at least capable of synchronizing internal clocks. It is also important to note that even light-entrained rhythms are “noisy” in cnidarians, which is why we were not surprised that the signal was weak. We have added a discussion of this observation in L601-612.
Lines 210-226: As mentioned above, I think it is not clear that temperature alone can synchronize the Nematostella clock and it is therefore problematic to call it a Zeitgeber. Nevertheless, Figure 3A, B, D show that certain offsets of the temperature cycle relative to the LD cycle do influence rhythmicity and phase in constant conditions. This is most likely due to a direct effect of temperature cycles on the endogenous circadian clock, which only becomes visible (measureable) when the animals are also exposed to certain offset LD cycles. My interpretation of the combined results would be that temperature cycles play only are very minor role in synchronizing the Nematostella clock (after all, LD and temperature cycles are not offset in nature), perhaps mainly supporting entrainment by the prominent LD cycles.
With our new data (see previous point), we believe we can safely say that temperature is a zeitgeber. We are not totally clear on what is meant by “a direct effect of temperature cycles on the endogenous circadian clock.” We argue that, because we see changes in free-running behavior during certain offsets, the timing of temperature cycles must affect the internal clock in a way that persists during constant conditions—it can’t just be a direct (clock-independent) effect of temperature.
Gene expression part: The authors performed an extensive temporal transcriptomic analysis and comparison of gene expression between animals kept in aligned LD and temperature cycles and those maintained in a 12 hr offset. While this was a tremendous amount of experimental work that was followed by sophisticated mathematical analysis, I think that the conclusions that can be drawn from the data are rather limited. First of all, it is known from other organisms that temperature cycles alone have drastic effects on overall gene expression and importantly in a clock independent manner (e.g., Boothroyd et al. 2007). Temperature therefore seems to have a substantially larger effect on gene expression levels compared to light (Boothroyd et al. 2007). In the current study, except for a few clock gene candidates (Figure 2C), the effects of temperature cycles alone on overall gene expression have not been determined. Instead the authors analysed gene expression during aligned and 12 h offset conditions making it difficult to judge which of the observed differences are due to clock independent and clock dependent temperature effects on gene expression. This is further complicated by the lack of expression data in constant conditions. I think the authors need to address these limitations of their study and tone down their interpretations of 'temperature being the most important driver of rhythmic gene expression' (e.g., line 401). At least they need to acknowledge that they cannot distinguish between clock independent, driven gene expression and potential influences of temperature on clock-dependent gene expression rhythms. Moreover, in their comparison between their own data and LD data obtained at constant temperature (taken from Oren et al. 2015), they show that temperature has only a very limited effect (if any) on core clock gene expression, further questioning the role of temperature cycles in synchronising the Nematostella clock. Nevertheless, I noted in Table 3 that there is a 1.5 to 3 hr delay when comparing the phase of eight potential key clock genes between the current study (temperature and LD cycles aligned) and LD constant temperature (determined by Oren et al.). To me, this is the strongest argument that temperature cycles at least affect the phase of clock gene expression, but the authors do not comment on this phase difference.
We agree with these points about the limitations of our study, and have revised the manuscript to phrase our conclusions more carefully. We still think it is reasonable to observe that temperature was a stronger drive of gene expression than light in our study, but this may not be true in other contexts.
In terms of the comparison with Oren et al. 2015, we didn’t want to over-interpret these results because there are other differences between the studies (L1181-1185), including the use of a different source population. In addition, we would prefer denser sampling (2h time points rather than 4h) and larger sample sizes to make claims about phase differences.
Network analysis: This last section of the results was very difficult to read and follow (at least for me). For example, do the colours in Figure 6A correspond to those in Figure 6B, C? A legend for each colour, i.e., which GO terms are included in each colour would perhaps be helpful. As mentioned above, I also do not think we can learn a lot from this analysis, since we do not know the effects of temperature cycles alone and we have no free-run data to judge potential influence on clock controlled gene expression. Under aligned conditions genes are expressed at a certain phase during the daily cycle (either morning to midday, or evening to midnight), which interestingly, is very similar to temperature cycle-only driven genes in Drosophila (Boothroyd et al. 2007). Inverting the temperature cycle has drastic effects on the peak phases of gene expression, but not so much on overall rhythmicity. But since no free-run data are available, we do not know to what extend these (expected) phase changes reflect temperature-driven responses, or are a result of alterations in the endogenous circadian clock.
We have revised and streamlined this section and Fig. 6, including removing panel 6C. The colors do correspond across panels in the figure. For space, GO terms of select modules are included in Fig. 6, and GO results for all modules are included in the Supplemental Data and discussed in the Results.
It is true that we can’t distinguish temperature-driven versus clock effects here, and it does seem like many modules simply follow the temperature cycle (which we say in this section). The most interesting finding from this section is probably that the co-expression structure (correlations between rhythmic genes) are substantially weakened during SC, and we do discuss certain modules of genes that lose or gain rhythmicity. We have revised this section to focus on the main points and have cut several of the less pertinent results.
Reviewer #3 (Public Review):
This article reflects a significant effort by the authors and the results are interesting.
For the third set of experiments, are temperature and light really out of synch? While peak in temperature no longer occurs along with lights on, we do still have two 24 hour cycles where changes in the environmental cues still occur simultaneously (lights on with peak in temperature, lights off with min in temperature). I wonder what would happen if light remained at a 24 hour cycle and temperature became either sporadic (randomly changing cycles) or was placed on a longer cycle altogether (temperature taking 20 hours to increase from min to max, and then another 20 hours to go from max to min).
Thank you for your interesting suggestions for future experiments. This point is addressed in our revisions responding to Reviewer #1, who requested a discussion of the phrase “sensory conflict.” We agree that the binary “in-sync vs. out-of-sync” may be too simplistic. Our original conception of sensory conflict was a situation in which light and temperature provide different phase information, as informed by experiments with only light (prior literature) or only temperature (this work).
In our revised manuscript, we discuss the idea that “sensory conflict” is not always a useful framework because there are many possible relationships between light and temperature. Although our 12h offset is certainly less “natural” than our aligned time series, it may be useful to think of them simply as 2 different possible light and temperature regimes in which the two signals interact, rather than abstract ideals of “aligned” or “misaligned.”
An area that could significantly benefit a broader readership would be to improve overall clarity of figures and rethink if all the results are necessary to convert the key findings of the paper. As written, the results sections is somewhat confusing.
We have revised Figs. 1 and 6 for clarity, and we have also shortened the network analysis portion of the Results.
-
eLife assessment
Understanding the integration and contribution of different combinations of environmental cues to the synchronization of the daily oscillator is important, because it provides insight into how organisms might be able to distinguish (and weight) between irregular (or in the tidal zone highly complex) versus regular individual daily changes of light and temperature. The study, which is thoroughly conducted and provides an impressive amount of experimental and analytical work, dissects the effects of sensory conflict on behavior and gene expression rhythms.
-
Reviewer #1 (Public Review):
The sea anemone Nematostella has been previously shown by the authors to exhibit diurnal patterns of movement around their culture dishes -- essentially they move around in darkness and not when in the light. This behaviour is entrained and continues to cycle when animals are kept in constant darkness. In this manuscript the authors test whether temperature cycles can substitute for light cycles in entraining this locomotory behaviour, and it turns out it can. They then test the effects of the two different entraining factors, light and temperature, when applied in phase (aligned exposure cycles) or out of phase (misaligned exposure cycles). The condition the authors call aligned (somewhat arbitrarily) is for the minimal temperature occurring at the beginning of the light cycle. They shift this alignment by …
Reviewer #1 (Public Review):
The sea anemone Nematostella has been previously shown by the authors to exhibit diurnal patterns of movement around their culture dishes -- essentially they move around in darkness and not when in the light. This behaviour is entrained and continues to cycle when animals are kept in constant darkness. In this manuscript the authors test whether temperature cycles can substitute for light cycles in entraining this locomotory behaviour, and it turns out it can. They then test the effects of the two different entraining factors, light and temperature, when applied in phase (aligned exposure cycles) or out of phase (misaligned exposure cycles). The condition the authors call aligned (somewhat arbitrarily) is for the minimal temperature occurring at the beginning of the light cycle. They shift this alignment by 2, 4, 6, 8, 10 and 12 hours; so for example the first altered set has lights on 2 hours after the minimal temperature, and the second set four hours after the minimal temperature. Animals are conditioned to these new entraining cycles for 'at least 2 weeks', then tests applied. The tests are either the behaviour of the animals at a constant temperature in constant darkness, or gene expression under one example of these altered out-of-phase conditions.
The authors view all data within a framework called 'sensory conflict' -- it is even the first two words of the title. This phrase is used in other papers to describe what seems to me an over-simplistic way to view the interactions between different entraining factors. Why is this 'conflict'? If they are two different environmental entraining signals and the cycle of one is shifted relative to the other you are simply changing the alignment of the cycles. By attempting to view the simple change in alignment within the formal framework of 'sensory conflict' the authors are limiting their (and the readers') ability to understand their results; all they see is different levels of conflict, which I would argue are not supported in any way by the results. If you were to shift the alignment of any two in-phase cycles so their peaks were then in anti-phase, you would wipe out any times in which neither of the two cycles was in the negative phase. It would not be 'sensory conflict' if in the anti-phase scenario no clear peak of the behaviour was evident. It would simply be the absence of peaks and troughs of conditions driving the behaviour. The framework used to discuss the data make it difficult to understand.
After exploring impacts on shifting the alignment of light and temperature cycles the authors also examine changes in gene expression patterns in animals entrained to the new regimen. This is a powerful approach as changes in gene expression underlie most of the changes to cellular responses to the environment. The very detailed analysis finds complex changes in transcription patterns. A number of genes associated with biological clocks, or daily cycles of light, have previously been identified in other animals, and a small field has explored them in cnidarians including Nematostella and their relatives, the corals. These cycling genes are found in the results, but temperature was in general a stronger influencer on changes in gene expression. In terms of light, the PAR-bZIP genes once again show up as major responders, and this is strongly supported by the authors' examination of regulatory regions near differentially expressed genes, which are enriched in PAR-bZIP binding sites.
Perhaps the most interesting set of genes identified are those that are only weakly rhythmic when the two entraining factor cycles are aligned, but become sharply rhythmic when they are in antiphase. The new sharp rhythms have an approximately 24 hour periodicity. As mentioned, temperature dominates changes in expression over light. Three core clock genes, Helt, PAR-bZIPa and Clock resist the shift to temperature and maintain their light driven cycles. The clear conclusion is that shifting the two entraining cycles results in large scale shifts in the underlying transcriptome for most rhythmic genes, with temperature dominating light except for core elements of the light responsive clock, and the major shifts are in metabolic processes.
The authors provide a gene level examination of the cellular response to shifting the alignment of two different entraining factors that allows us to view, if not completely understand, how interactions between environmental signals are integrated.
-
Reviewer #2 (Public Review):
This is an interesting study investigating the effects of sensory conflict on rhythmic behaviour and gene expression in the sea anemone Nematostella vectensis. Sensory conflict can arise when two environmental inputs (Zeitgeber) that usually act cooperatively to synchronize circadian clocks and behaviour, are presented out of phase. The clock system then needs to somehow cope with this challenge, for example by prioritising one cue and ignoring the other. While the daily light dark cycle is usually considered the more reliable and potent Zeitgeber, under some conditions, daily temperature cycles appear to be more prominent, and a certain offset between light and temperature cycles can even lead to a breakdown of the circadian clock and normal daily behavioural rhythms. Understanding the weighting and …
Reviewer #2 (Public Review):
This is an interesting study investigating the effects of sensory conflict on rhythmic behaviour and gene expression in the sea anemone Nematostella vectensis. Sensory conflict can arise when two environmental inputs (Zeitgeber) that usually act cooperatively to synchronize circadian clocks and behaviour, are presented out of phase. The clock system then needs to somehow cope with this challenge, for example by prioritising one cue and ignoring the other. While the daily light dark cycle is usually considered the more reliable and potent Zeitgeber, under some conditions, daily temperature cycles appear to be more prominent, and a certain offset between light and temperature cycles can even lead to a breakdown of the circadian clock and normal daily behavioural rhythms. Understanding the weighting and integration of different environmental cues is important for proper synchronization to daily environmental cycles, because organisms need to distinguish between 'environmental noise' (e.g., cloudy weather and/or sudden, within day/night temperature changes) and regular daily changes of light and temperature. In this study, a systematic analysis of different offsets between light and temperature cycles on behavioural activity was conducted. The results indicated that several degrees of chronic offset results in the disruption of rhythmic behaviour. In the 2nd part of the study the authors determine the effect of sensory conflict (12 hr offset that leads to robust disruption of rhythmic behaviour) on overall gene expression rhythms. They observe substantial differences between aligned and offset conditions and conclude a major role for temperature cycles in setting transcriptional phase. While the study is thoroughly conducted and represents and impressive amount of experimental and analytical work, there are several issues, which I think question the main conclusions. The main issue being that temperature cycles by themselves do not seem to fulfil the criteria for being considered a true Zeitgeber for the circadian clock of Nematostella.
Major points:
Line 53: 'However, many of these studies did not compare more than two possible phase relationships.....'. Harper et al. (2016) did perform a comprehensive comparison of different phase relationships between light and temperature Zeitgebers (1 hr steps between 2 and 10 hr offsets), similar to the one conducted here. I think this previous study is highly relevant for the current manuscript and -- although cited -- should be discussed in more detail. For example, Harper et al. show that during smaller offsets temperature is the dominant Zeitgeber, and during larger sensory conflict light becomes the dominant Zeitgeber for behavioural synchronization. Only during a small offset window (5-7 hr) behavioural synchronization becomes highly aberrant, presumably because of a near breakdown of the molecular clock, caused by sensory conflict. Do the authors see something similar in Nematostella? Figure 3 suggests otherwise, at least under entrainment conditions, where behaviour becomes desynchronized only at 10 and 12 hr offset conditions. But in free-run conditions behaviour appears largely AR already at 6 hr offset, but not so much at 4 and 8 hr offsets (Table 2). So there seems to be at least some similarity to the situation in Drosophila during sensory conflict, which I think is worth mentioning and discussing.
Line 111: The authors state that 14-26C temperature cycle is 'well within the daily temperature range experienced by the source population'. Too me this is surprising, as I was not expecting that water temperature changes that much on a daily basis. Is this because Nematostella live near the water surface, and/or do they show vertical daily migration? Also, I do not understand what is meant by '...range of in situ diel variation (of temperature)'. I think a few explanatory words would be helpful here for the reader not familiar with this organism.
Lines 114-117: I was surprised that clock genes can basically not be synchronized by temperature cycles alone. Only cry2 cycled during temperature cycles but not in free-run, so the cry2 cycling during temperature cycles could just be masking (response to temperature). Later the authors show robust molecular cycling during combined LD and temperature cycles (both aligned and out of phase), indicating that LD cycles are required to synchronize the molecular clock. Moreover, a previous study has demonstrated that LD cycles alone (i.e., at constant temperature) are able to induce rhythmic molecular clock gene expression (Oren et al. 2015). Similarly, the free running behaviour after temperature cycles does not look rhythmic to me. In Figure 2A, 14-26C there is at best one peak visible on the first day of DD, and even that shows a ~6 phase delay compared to the entrained condition. After the larger amplitude temperature cycle (8:32C) behaviour looks completely AR and peak activity phases in free-run appear desynchronized as well (Fig. 2B). Overall, I think the authors present data demonstrating that temperature cycles alone are not sufficient to synchronize the circadian clock of Nematostella. One way to proof if the clock can be entrained is to perform T-cycle experiments, so changing the thermoperiod away from 24 hr (e.g., 10 h warm : 10 h cold). If in a series of different T-cycles the peak activity always matches the transition from warm to cold (as in 12:12 T-cycles shown in Fig. 1A) this would speak against entrainment and vice versa.
Lines 210-226: As mentioned above, I think it is not clear that temperature alone can synchronize the Nematostella clock and it is therefore problematic to call it a Zeitgeber. Nevertheless, Figure 3A, B, D show that certain offsets of the temperature cycle relative to the LD cycle do influence rhythmicity and phase in constant conditions. This is most likely due to a direct effect of temperature cycles on the endogenous circadian clock, which only becomes visible (measureable) when the animals are also exposed to certain offset LD cycles. My interpretation of the combined results would be that temperature cycles play only are very minor role in synchronizing the Nematostella clock (after all, LD and temperature cycles are not offset in nature), perhaps mainly supporting entrainment by the prominent LD cycles.
Gene expression part: The authors performed an extensive temporal transcriptomic analysis and comparison of gene expression between animals kept in aligned LD and temperature cycles and those maintained in a 12 hr offset. While this was a tremendous amount of experimental work that was followed by sophisticated mathematical analysis, I think that the conclusions that can be drawn from the data are rather limited. First of all, it is known from other organisms that temperature cycles alone have drastic effects on overall gene expression and importantly in a clock independent manner (e.g., Boothroyd et al. 2007). Temperature therefore seems to have a substantially larger effect on gene expression levels compared to light (Boothroyd et al. 2007). In the current study, except for a few clock gene candidates (Figure 2C), the effects of temperature cycles alone on overall gene expression have not been determined. Instead the authors analysed gene expression during aligned and 12 h offset conditions making it difficult to judge which of the observed differences are due to clock independent and clock dependent temperature effects on gene expression. This is further complicated by the lack of expression data in constant conditions. I think the authors need to address these limitations of their study and tone down their interpretations of 'temperature being the most important driver of rhythmic gene expression' (e.g., line 401). At least they need to acknowledge that they cannot distinguish between clock independent, driven gene expression and potential influences of temperature on clock-dependent gene expression rhythms. Moreover, in their comparison between their own data and LD data obtained at constant temperature (taken from Oren et al. 2015), they show that temperature has only a very limited effect (if any) on core clock gene expression, further questioning the role of temperature cycles in synchronising the Nematostella clock. Nevertheless, I noted in Table 3 that there is a 1.5 to 3 hr delay when comparing the phase of eight potential key clock genes between the current study (temperature and LD cycles aligned) and LD constant temperature (determined by Oren et al.). To me, this is the strongest argument that temperature cycles at least affect the phase of clock gene expression, but the authors do not comment on this phase difference.
Network analysis: This last section of the results was very difficult to read and follow (at least for me). For example, do the colours in Figure 6A correspond to those in Figure 6B, C? A legend for each colour, i.e., which GO terms are included in each colour would perhaps be helpful. As mentioned above, I also do not think we can learn a lot from this analysis, since we do not know the effects of temperature cycles alone and we have no free-run data to judge potential influence on clock controlled gene expression. Under aligned conditions genes are expressed at a certain phase during the daily cycle (either morning to midday, or evening to midnight), which interestingly, is very similar to temperature cycle-only driven genes in Drosophila (Boothroyd et al. 2007). Inverting the temperature cycle has drastic effects on the peak phases of gene expression, but not so much on overall rhythmicity. But since no free-run data are available, we do not know to what extend these (expected) phase changes reflect temperature-driven responses, or are a result of alterations in the endogenous circadian clock.
-
Reviewer #3 (Public Review):
This article reflects a significant effort by the authors and the results are interesting.
For the third set of experiments, are temperature and light really out of synch? While peak in temperature no longer occurs along with lights on, we do still have two 24 hour cycles where changes in the environmental cues still occur simultaneously (lights on with peak in temperature, lights off with min in temperature). I wonder what would happen if light remained at a 24 hour cycle and temperature became either sporadic (randomly changing cycles) or was placed on a longer cycle altogether (temperature taking 20 hours to increase from min to max, and then another 20 hours to go from max to min).
An area that could significantly benefit a broader readership would be to improve overall clarity of figures and rethink if …
Reviewer #3 (Public Review):
This article reflects a significant effort by the authors and the results are interesting.
For the third set of experiments, are temperature and light really out of synch? While peak in temperature no longer occurs along with lights on, we do still have two 24 hour cycles where changes in the environmental cues still occur simultaneously (lights on with peak in temperature, lights off with min in temperature). I wonder what would happen if light remained at a 24 hour cycle and temperature became either sporadic (randomly changing cycles) or was placed on a longer cycle altogether (temperature taking 20 hours to increase from min to max, and then another 20 hours to go from max to min).
An area that could significantly benefit a broader readership would be to improve overall clarity of figures and rethink if all the results are necessary to convert the key findings of the paper. As written, the results sections is somewhat confusing.
-
-
-