Orco regulates the circadian activity of pheromone-sensitive olfactory receptor neurons in hawkmoths

Joint Public Review:

The revised manuscript is much clearer, and the additional analyses address several of the original concerns. RAIN analyses (Rhythmicity Analysis Incorporating Nonparametric methods) now detects circadian rhythmicity in 7/11 recordings under light-dark conditions and 8/12 recordings in constant darkness, compared with 2/12 following treatment with the Orco antagonist. This supports circadian modulation of spontaneous firing and a role for Orco in its normal expression. The expanded qPCR analysis of Orco also supports the conclusion that Orco transcript abundance is not circadian, and the cAMP experiment shows that cAMP can modulate Orco-dependent activity.

The remaining issue concerns the mechanistic interpretation. The lack of rhythmic Orco transcript abundance does not distinguish an autonomous post-translational feedback-loop (PTFL) clock from a model in which the canonical transcriptional-translational (TTFL) clock acts upstream through cAMP, calcium, kinases, phosphatases, channel trafficking, or related pathways to regulate Orco.

Similarly, the new Figure 10 provides a useful representation of the authors' hypothesis, but the proposed delayed feedback and coupling mechanisms are not experimentally demonstrated.

We do not think any further experiments are necessary for the present study. Instead, we recommend that the manuscript should clearly distinguish between what the data show and what remains proposed. The data support circadian modulation of ORN firing a role for Orco in its normal expression, non-circadian Orco transcript abundance, and cAMP-sensitive modulation of Orco-dependent activity. The proposal that an Orco-centred membrane feedback loop generates the rhythm is intriguing and may remain a hypothesis generated through this study that needs formal testing in the future. This should be explicitly stated. While this has been done in the discussion section, elsewhere, including in the abstract and elsewhere, the original claim remains.

Author response:

The following is the authors’ response to the previous reviews

eLife Assessment

This valuable study uses technically compelling long-term in vivo recordings and computational modeling to investigate whether hawkmoth olfactory receptor neurons show circadian modulation of spontaneous firing. The authors further propose the provocative model that post-translational mechanisms, rather than the transcriptional-translational processes, may contribute to circadian regulation of neuronal excitability.

We are pleased that our study is recognized as valuable and that our technically very challenging long-term in vivo recordings and computational modeling are appreciated. We agree that we are proposing a provocative model that opposes the current hypothesis in chronobiology, which suggests that all observed biological circadian rhythms are outputs of a transcriptional-translational feedback loop (TTFL) clock. Instead, we suggest that a cell comprises, in addition to the TTFL clock, other posttranslational feedback loop (PTFL) clocks without the need of daily transcription and daily degradation of its core elements. While the circadian TTFL clock is entrained to the daily light-dark cycle, the circadian PTFL clocks are suggested to be entrained to other daily cues such as to the availability of pheromone, or the daily changing levels of hormones and second messenger levels. Our novel hypothesis proposes that the TTFL and PTFL clocks are coupled and linked, constituting an adaptive, plastic network that can tune and phase-lock to different external and internal Zeitgeber signals. However, we certainly do not claim that the TTFL circadian clock is not at all involved in the circadian control of the ORN’s circadian membrane potential rhythms. We clarified our manuscript accordingly.

However, the evidence for circadian firing in these neurons […] remains incomplete.

As requested by the reviewers during the previous round of review, we had provided the results of RAIN analysis (Thaben and Westermark, 2014) of individual animals in our first revision (Fig. 4A), which clearly shows that two-thirds of the population express circadian rhythms in key attributes that we used to characterize the spontaneous spiking activity. In the initial review it was assumed by the reviewers that phase alignment of dispersed rhythms would bias interpretations of rhythmicity. After having established with RAIN that individuals show circadian rhythms, albeit dispersed across the population due to the lack of a zeitgeber in DD conditions, phase-alignment of recordings from DD animals is a valid next step to prepare the data for statistical analysis across the population. Phase-alignment of desynchronized rhythms is a generally accepted and necessary method employed in chronobiology (e.g., for insect ORNs: Gosh et al., 2024). It is proven as prerequisite to find and statistically analyze rhythmicity in complex, desynchronized data.

As we explained in the previous rebuttal and in our first revision, rhythms in electrical activity of insect ORNs cannot be easily synchronized by the light-dark cycle alone, but appear to require daily cycles of pheromone, as shown in other moth species (Gosh et al., 2024). Please be aware that the animals that we used here have never been exposed to pheromone, as we state in the Methods. Therefore, this lack of pheromone exposure can explain why about one third of our experimental population is not expressing any daily or circadian rhythmicity in spiking attributes. This is an important result of our manuscript, reported for the first time for Manduca sexta, providing evidence for our hypothesis that it is not the LD-entrained circadian TTFL clock that governs electrical activity rhythms in ORNs. As we explained in the first revision, and clarified further here in the second revision, we cannot phase-synchronize our animals with cycles of pheromone application in our experimental paradigm because we are researching circadian rhythms in spontaneous spiking activity and not pheromone responses. We failed to obtain phase-alignment with a single pheromone pulse the night before the experiments started. These data were added as supplementary Figure to Fig. 3 in the first revision. Here, we further revised our manuscript to clarify this important finding.

Thus, as requested by the reviewers in the initial review, we could successfully confirm our previous results of circadian firing in ORNs and the disruption of these circadian rhythms with Orco antagonist OLC15 with RAIN. In the current review, the reviewers raise no further specific critical points or comments that would doubt our careful rhythm analysis of our long-term recordings. Thus, we conclude that we provided clear evidence for our central, exciting new finding. For the first time we demonstrated an unexpected new task for Orco: Orco controls the circadian firing pattern in the spontaneous activity, and thus, of the ORN’s membrane potential, via its property as leak/pacemaker channel.

However, the evidence […] for post-translational modification of Orco as the underlying mechanism remains incomplete.

We agree with the reviewers that there are many more experiments and combined efforts of biochemists, structural biologists, and electrophysiologists required to provide complete evidence for post-translational modification of Orco and to reveal the underlying mechanism of its circadian control. It is beyond the scope of the current manuscript to provide all details of post-translational control of Orco.

The reviewers asked previously for additional evidence that Orco transcription is not controlled via the TTFL clock. As requested, we provided extended qPCR–based evidence that Orco, in contrast to timeless, is not controlled by the TTFL clock on the transcriptional level (Fig. 6 in Revision 1). Furthermore, we added a new result in Revision 1 to demonstrate cAMP-dependent post-translational modulation of Orco open-time probability (Fig 9 in Revision 1) at a ZT at which antennal cAMP levels are low (Schendzielorz et al., 2015). We already showed in Flecke et al., 2010, that the addition of cAMP at different ZTs increases the spontaneous spiking activity only at specific ZTs. Here, we show that the effect of cAMP depends on Orco. Since Orco’s circadian role is not mediated via TTFL control, it can be concluded that post-translational mechanisms provide daily/circadian temporal control. In this additional Figure we provide statistically significant proof that, in agreement with our model-prediction, Orco’s circadian control of the ORN spontaneous activity could be mediated via the second messenger cAMP. ZT-dependent input for Orco would be provided via daily changes in cAMP levels (Schendzielorz et al., 2015).

In contrast, the study does provide strong evidence that the application of cyclic nucleotides can modulate Orco-dependent activity at a single time point, and reports that the temporal pattern of Orco transcript abundance is not circadian.

We appreciate that the reviewer confirms that our revised manuscript with additional experiments now provides strong evidence that cAMP modulates Orco-dependent spontaneous activity of M. sexta ORNs. Since we already published that cAMP levels expresses daily rhythms in M. sexta antennae (Schendzielorz et al., 2015), and in vivo cAMP infusion increases spontaneous activity and sensitizes pheromone detection (Flecke et al., 2010), and our computational model here proves that circadian modulation of open time probability of Orco is sufficient to explain our experimental data, it is sufficient for our conclusions to test just the one specific zeitgeber time when endogenous cAMP levels are low and pharmacological cAMP increase has the strongest impact. To further reveal complete ZT-dependence of cAMP modulation of Orco´s control of spontaneous activity is beyond the scope of the current manuscript and not part of the current research question. The structure of Orco is extraordinarily conserved during evolution, thus, the cited experimental results from other laboratories and other species showing that Orco is a hub for posttranslational modification are very likely generalizable to different insect species. We clarified the manuscript accordingly. In Drosophila, Orco has at least 5 phosphorylation sites for protein kinase C (PKC), is cAMP-dependently sensitized, and has a Ca²⁺/calmodulin binding site that orchestrates the localization of the OR-Orco heteromer to the cilia. However, in fruit fly and other insects, so far, it can only be speculated how circadian control is provided for Orco, since there are no other publications that examine the circadian regulation of Orco in detail. We clarified our manuscript accordingly.

To summarize, the logical conclusion based on our newly provided data is that the current hierarchical hypothesis in chronobiology based solely on a circadian TTFL clock that controls Orco transcription does not explain our findings in hawkmoth ORNs. Therefore, we suggest a new systemic hypothesis based upon coupled TTFL and PTFL circadian clocks that can also reconcile otherwise inconsistent data published for insect and mammalian circadian clocks (please see reviewed data in: Stengl and Schneider, 2024). We clarified our manuscript in the second revision and added a new Figure 10 to illustrate our novel hypothesis.

However, the findings are incomplete to exclude a role for transcriptional-translational mechanisms and their associated multi-layered controls in circadian regulation.

We certainly do not imply excluding a role for the TTFL clock in (indirectly) affecting circadian control of the membrane potential of ORNs. The new qPCR experiments added in the first revision clearly show that the circadian control mediated via Orco is not an output of the TTFL clock via transcriptional control of Orco. Instead, we predict links between a PTFL membrane clock comprising Orco as hub to integrate posttranslational control and the TTFL nuclear clock. We clarified our manuscript accordingly, adding a new Figure 10 to further illustrate and visualize our hypothesis. The predictions of this systemic hypothesis will be challenged in further experiments that, however, are beyond the scope of the current manuscript.

Joint Public Review:

This manuscript puts forward the provocative idea that a posttranslational feedback loop regulates daily and ultradian rhythms in neuronal excitability. The authors used in vivo long-term tip recordings of the long trichoid sensilla of male hawkmoths to analyze spontaneous spiking activity indicative of the ORNs' endogenous membrane potential oscillations. This firing pattern was disrupted by pharmacological blockade of the Orco receptor. They then use these recordings together with computational modeling to predict that Orco receptor neuron (ORN) activity is required for circadian, not ultradian, firing patterns. Orco did not show a circadian expression pattern in a qPCR experiment, and its conductance was proposed to be regulated by cyclic nucleotide levels. This evidence led the authors to conclude that a post-translational feedback loop (PTFL) clockwork, associated with the ORN plasma membrane, allows for temporal control of pheromone detection via the generation of multi-scale endogenous membrane potential oscillations. The findings will interest researchers in neurophysiology, circadian rhythms, and sensory biology. However, the manuscript has limited experimental evidence to support its central hypothesis and is undermined by several assumptions that underlie their data analysis and model builds, as well as insufficient biological data including critical controls to validate and/or fully justify the model the authors are proposing.

We want to remind our reviewers that we used “ORN” as abbreviation for olfactory receptor neuron (= sensory receptor neurons, a.k.a. olfactory sensory neuron (OSN)) and not for Orco receptor neuron, although we focus on the function of Orco. Accordingly, our central finding is that Orco as ion channel is required for the daily/circadian modulation of spontaneous action potential activity generated by the olfactory receptor neurons in the absence of pheromone stimulation.

We do not understand the specific basis for the conclusions of the reviewers. Therefore, we ask to please specify what experimental evidence is missing to support our central hypothesis that Orco is not directly TTFL- but PTFL clock-controlled, and to name specifically what the “several assumptions” are that undermine our careful data analysis and model builds. Which specific argument in our previous rebuttal was wrong, was not conclusive? Furthermore, please specify your claim that “critical controls are missing”. Which controls are missing for which experiments? We did add a new figure panel in the first revision to demonstrate that neither the addition of DMSO (the OLC15 solvent) nor the repeated attachment of the recording electrode, which was necessary to obtain paired datasets, altered the spontaneous spiking activity (Fig 1B in Revision 1). Furthermore, we expanded the time series of qPCR data and added tim as positive control to Orco (Fig 7), strengthening our argument that Orco expression is not under TTFL control. Dose-response curves of various Orco agonists and antagonists have been published before (see our references in Revision 1) and are therefore not repeated here.

Our newly added data confirm what our modeling predicted: cAMP increases spontaneous ORN activity dependent on Orco. Previous publications provide evidence for daily rhythms in cAMP concentrations in hawkmoth antennae (Schendzielorz et al., 2012).

As is true for any other hypothesis, a hypothesis can only be falsified but not validated and needs to be tested by many experiments from many laboratories over a long time until it will be replaced by the next hypothesis that better explains accumulating contradicting evidence. We are very much looking forward to experimental challenges of our provocative new hypothesis by colleagues in the field of olfaction and of chronobiology. We are convinced that our manuscript will greatly stimulate the field, possibly provoking a paradigm switch in chronobiology and in olfactory research.

Strengths:

The authors raise several intriguing model-based hypotheses regarding the mechanisms that underlie the generation of olfactory rhythms. The electrophysiological approach and the long-term recording paradigm are elegant and technically impressive. In the revised version, the authors have added additional qPCR data supporting the lack of rhythmic Orco transcript expression and included a new figure suggesting that cAMP can modulate Orco conductance.

We thank the reviewers for their acknowledgement of our careful work and hope that our further revisions and clarifications help to argue our case.

Major weaknesses:

(1) The cAMP experiment was only conducted at one time-point, which is insufficient to support the central claim that "AMP and cGMP may have ZT-dependent effects on Orco conductivity".

We agree with the reviewers and revised our discussion accordingly to clarify that in this manuscript it is not our central claim that cAMP and cGMP may have ZT-dependent effects on Orco conductivity. Instead, our data show for the first time that Orco controls circadian rhythms of spontaneous activity of ORNs and that the circadian rhythmicity of spontaneous activity is lost when Orco is blocked. Therefore, we provide novel experimental evidence that Orco is a prerequisite to the circadian rhythmicity of spontaneous activity and thus, to circadian rhythms in the membrane potential of ORNs. Furthermore, as requested by the reviewers we provided clear evidence in the first revision that Orco is not controlled at the transcriptional level by the TTFL clock, in contrast to the TTFL clock protein TIMELESS. Thus, it follows logically that Orco is under post-transcriptional control. Since cAMP levels show circadian oscillations and Orco is gated by cAMP (Fig 9 in Revision 1), we used our computational model to show that a cAMP-dependent increase in Orco conductance alone, via daily oscillating concentrations of cAMP, is sufficient to explain our findings. Therefore, we propose here that daily/circadian oscillations of cAMP modify spontaneous spiking activity via Orco on a posttranscriptional level. But it certainly does not provide all evidence for respective mechanisms of how this cAMP modulation of Orco is obtained, since this is beyond the scope of the current manuscript.

Since we realized that it is difficult for our readers to visualize a circadian PTFL membrane clock we added a new hypothesis-Figure (Figure 10) and considerably focused and clarified especially the discussion of our manuscript. We pointed out that a membrane-associated signalosome that comprises delayed negative feedback mechanisms, and, thus, constitutes an oscillator, a “membrane clock” that generates oscillations. Based upon our data we propose a membrane-associated signalosome constituting a PTFL circadian clock with Orco as central element. This PTFL membrane clock generates superimposed ultradian and circadian rhythms in its outputs: rhythms in the membrane potential, Ca²⁺, and cAMP levels. The PTFL clock comprises positive feedforward elements that upregulate its outputs, resulting in more depolarization, higher Ca²⁺- and higher cAMP levels. Via the clock’s delayed negative feedback mechanisms these outputs are downregulated, again, resulting in hyperpolarization, decreasing Ca²⁺- and cAMP levels. This signalosome comprises the pacemaker channel Orco as a central hub that is controlled via changes in voltage, Ca²⁺, and cAMP levels. Nevertheless, we predict coupling between the multiscale PTFL membrane clock and the TTFL circadian clock in the nucleus to obtain stable circadian rhythms. As likely mechanism of coupling we predict that Ca²⁺- and cAMP-dependent kinases interlink both types of clocks, thereby obtaining robust and at the same time flexible interlinked cellular rhythms.

We hope to now successfully clarify and to visualize our central hypothesis of our manuscript that Orco is not directly controlled by a TTFL circadian clock but is a central element of a membrane-associated posttranslational feedback loop clock (PTFL) clock that is linked to but not forced by the TTFL clock which is predicted to control intracellular Ca²⁺ homeostasis in a circadian rhythm.

(2) The revised manuscript continues to rely heavily on prior publications or defers key mechanistic questions (or important manipulations) to future studies. In its current form, the evidence presented remains insufficient to support the central claim that a PTFL constitutes the primary underlying circadian clock mechanism. The proposed model is intriguing, but the data provided do not yet directly demonstrate the novel mechanism.

We do not understand why the reviewers considers it to be problematic that we “continue to rely heavily on prior publications”. Certainly, we built upon previous publications of our lab as well as on manifold experimental data published by other laboratories in the field of insect olfaction. Our ample citations demonstrate that we have an overview both of the current state of literature and relevant previous literature, dating back to the very first experiments that pioneered pheromone transduction in insects. Based upon our extensive knowledge and experimental data collected in insect olfaction and based on very careful, critical, rigorous analysis of our data and data published by others, we were able to come up with a novel interpretation of the current literature about insect olfaction that differs considerably from the current main views. We consider this to be our strength and judge it as good scientific practice and not a flaw of our work. However, since we do not focus on OR-Orco heteromers and their function in pheromone/odor transduction in the cilia in the current manuscript, we considerably shortened this part of the discussion, avoiding pointing out that highly sensitive moth pheromone transduction greatly differs from less sensitive general odor transduction in Drosophila. Furthermore, since here we focus on cAMP, but not on cGMP-dependent modulation of Orco, we also deleted/considerably shortened this part of our discussion.

We certainly agree with the reviewers that, while the data provided in the current manuscript are a logical basis for developing our novel hypothesis, they are not a direct and sufficient demonstration of proof and we are not able yet to directly demonstrate and explain the novel mechanism predicted. We would like to point out that if we provided this final proof, it would not be any more a novel hypothesis, but only a novel finding.

We agree with the reviewers that our provocative hypothesis requires rigorous testing by many further experiments, hopefully not only by our laboratory, but hopefully stimulating new experimental challenges by other laboratories employing different species. But certainly, these experiments with proof-of-principle will take many years and are beyond the scope of our current research paper.

As per eLife’s assessment system we would like to ask the reviewers to provide detailed feedback as to which experiments/results within the scope of this manuscript would complete this work, or how they think this study should be framed in the light of the results that we obtained. Nevertheless, we hope that with our current careful review the reviewers will be more convinced by our arguments and experiments as valid basis for our provocative new hypothesis.