Organelle calcium-derived voltage oscillations in pacemaker neurons drive the motor program for food-seeking behavior in Aplysia

Article activity feed

1. 

   Version published to 10.7554/elife.68651 on eLife

   Jun 30, 2021
2. 

   eLife

   May 21, 2021

   Author Response:

   Reviewer #2:

   Motivated behaviors, such as food seeking when hungry, can also occur spontaneously at irregular intervals. Understanding how this irregular expression arises is important for understanding behavior and is relatively little investigated. The present work thus addresses an important and under-investigated area in neurobiology. Its demonstration of a potential cellular mechanism for irregular behavioral production has wide relevance, ranging from how cells make "decisions" to how whole organisms do so.

   Intact Aplysia occasionally produce bites even in the absence of food, and isolated buccal ganglia (which contain the biting central generator circuit) will occasionally spontaneously produce fictive bite motor patterns. The activity of central pattern generator networks has almost exclusively been …

   Author Response:

   Reviewer #2:

   Motivated behaviors, such as food seeking when hungry, can also occur spontaneously at irregular intervals. Understanding how this irregular expression arises is important for understanding behavior and is relatively little investigated. The present work thus addresses an important and under-investigated area in neurobiology. Its demonstration of a potential cellular mechanism for irregular behavioral production has wide relevance, ranging from how cells make "decisions" to how whole organisms do so.

   Intact Aplysia occasionally produce bites even in the absence of food, and isolated buccal ganglia (which contain the biting central generator circuit) will occasionally spontaneously produce fictive bite motor patterns. The activity of central pattern generator networks has almost exclusively been ascribed to the actions of the voltage-gated channels in the network neuron cell membranes and the synaptic connectivity among the network's neurons. Bédécarrats et al. show that a small, highly regular cell membrane voltage oscillation occurs in a neuron (B63) in the biting neural network, and that occasionally this oscillation becomes large enough to trigger a plateau potential in B63 and a single fictive bite from the entire circuit. They show that this oscillation is not due to cell membrane voltage dependent conductances, but instead from process involving the endoplasmic reticulum, mitochondria, or both. Although organelle-driven changes in cellular or tissue activity have been observed in other cell types, this is, to my knowledge, its first observation in a neural network. These data thus are potentially of great importance in understanding how neural networks function, most of which do not show the great regularity of central pattern generated behaviors.

   The presented data seem, to me, strong with respect to the small potential oscillations not being generated by voltage-dependent cell membrane conductances, and somehow involving the intracellular organelles. What is less clear to me is how local release of Ca from endoplasmic reticulum or mitochondria would result in changes in ion composition under the cell membrane, which is what gives rise to the cell membrane potential. Ca is highly buffered in the cytoplasm. It is thus unclear to me that free Ca would remain so for any length of time after release. It does, of course, in muscles, but these are evolved for this to occur. The authors themselves raise a variant of these concerns in the Discussion when considering how the B63 cell membrane voltage oscillations are transmitted to neurons electrically coupled to B63, invoking as a possibility Ca activation of second messengers, which would then themselves be responsible for the cell to cell communication. It seems to me that the same concerns arise with respect to how Ca release at sites distant from the cell membrane could charge the membrane's capacitance.

   A second remarkable observation is that B63 depolarization and firing does not reset the organelle-derived slow oscillation. B63 firing should result in substantial Ca concentration changes, at least in a shell under the cell membrane, so a possible feedback mechanism can be imagined. Most biological processes contain multiple feedback process that link cause and effect (e.g., the sequential current activations that return a cell to rest after an action potential, the interactions between sympathetic and parasympathetic system activity that maintain functionally proper body activation, the interactions that regulate hormonal levels). One possibility the authors mention is that the organelle-derived oscillation is used only for intermittent bite activities, and in feeding bites are instead generated solely by standard cell-membrane voltage-dependent processes. Regardless, it is a striking observation that merits additional investigation.

   These issues, however, do not change the data, which show a clear association of disruption of endoplasmic reticulum and mitochondrial function and cessation of the cell membrane voltage oscillation. Nor is it reasonable to expect an article like this, showing an organelle-driven cell membrane potential oscillation for the first time in a neuron, to describe every aspect of the mechanism by which it occurs. Indeed, it is a measure of the article's interest that it prompts such thinking. It will be very interesting to see the effects of similar organelle-disrupting treatment on the activity of other well-defined neural networks.

   The reviewer’s concern about how a local release of Ca from the endoplasmic reticulum or mitochondria could alter the ion composition under the cell membrane, and thereby cell membrane potential, raises an important issue in the context of our paper. However, we should point out that there is a considerable body of experimental evidence indicating that plasma membrane ionic conductances are altered by organelle-released Ca in many different non-muscle cells and tissues, including Aplysia and crustacean stomatogastric ganglion neurons (e.g., Hickey et al. 2010, J Neurophysiol 103:1543–1556; Knox et al., 1996, J Physiol 494:627-39; Kadiri et al., 2011, J Neurophysiol 106:1288–1298). Nonetheless, how intracellular Ca is able to alter plasma membrane conductance despite the presence of strong buffering mechanisms remains an open question. As the reviewer posits in a subsequent comment, this capability may indeed be due to a close proximity of the organelle and plasma membranes (see response to comment 8 below), and/or is mediated indirectly by a Ca-dependent activation of second messenger cascades (Lorenzetti et al. 2008 Neuron 59: 815–828).

   The reviewer also mentioned a striking observation that B63 depolarization and firing does not reset the organelle-derived slow oscillation. We can only speculate at this stage, but this lack of resetting may be related to cell compartmentalization. As now discussed in the manuscript, the Ca and associated voltage oscillations may be generated in a compartment (e.g., the neuropile) that is remote from the site where production of plateau potentials occurs (e.g., the soma). A propagation of the voltage oscillation from the first compartment (similar to postsynaptic potentials) could trigger the plateau at the second distant locus. Conversely, local calcium influxes induced by depolarization or plateau production at this latter site may be insufficient to alter the distant organelle-derived Ca oscillation. A paragraph in which this idea is expanded further has been added to the Discussion.

   Reviewer #3:

   In this report the authors characterize a mechanism that plays a role in inducing the rhythmic depolarizations that are observed in identified neurons that are part of the feeding CPG in Aplysia. The neurons studied (B63 neurons) are of interest because previous work has established that they play an important role in triggering cycles of motor activity. Further, previous work from this group has demonstrated that activity in the B63 neurons can be modified by operant conditioning.

   The authors present this study as though previous work had established that plateau potentials generated in the B63 neurons play an important role in driving network activity. For example, in line 102 they state "This essential role played by B63 is partly mediated by a bistable membrane property, which allows the sudden switching of the neuron's resting membrane potential to a depolarized plateau…" To support this statement, they reference Susswein et al. 2002, which does not support this statement. In the Susswein et al. study it is the B31/32 neurons that are modeled as having plateau properties.

   If previous work has not established the role of the B63 plateau potentials, the only data that speak to this issue are presumably in the current report. In this study the authors do provide data that indicate that the B63 neurons generate low amplitude oscillations that are not likely to depend on input from the electrically coupled neurons studied (notably B31). The authors also show that in some instances, these depolarizations do trigger plateau potentials in B63. It is, however, not clear that the B63 generated plateau potentials are then responsible for triggering network activity (e.g., as opposed to a situation where depolarizing input from B63 triggers plateau potentials in B31/32 and the depolarization in B31/32 drives the rest of the feeding circuit). For example, in Figs. 6A and Supplemental Fig. 4A it does not appear that the plateau depolarization in B63 is being transmitted to other electrically coupled neurons to any large extent.

   A clarification of this issue is important because it potentially impacts thinking concerning how 'decision making' is occurring. If decision making means induction of a motor program and this does not occur unless the depolarization in B63 is transmitted to B31/32, the process is more complicated than what the manuscript currently suggests.

   The title is misleading since there are no studies of behavior in this report.

   In part, interest in the mechanisms that drive spontaneous oscillatory activity in the B63 neurons stems from the overall context of this work. Namely the authors have previously established that oscillatory activity can be modified through associative learning. In the Sieling et al. 2014 study they demonstrate that two aspects of plasticity are accounted for by changes in synaptic properties and an effect on a leak current. For readers trying to understand this body of work as a whole, the Discussion should more clearly indicated how the results of the present study integrate with these previous findings.

   We agree with the reviewer that the present study is the first to establish that B63 is intrinsically capable of generating plateau potentials. We have therefore modified the manuscript to clarify this point:

   Lines 80-86 now state: “Thus, deciphering the mechanisms underlying the bursting activity of these key decision neurons is critical to understanding the process of radula motor pattern expression. Although earlier modeling evidence suggested that B63 bursting might rely on the cell’s electrical synapses with other circuit neurons that possess a plateau potential-generating capability (Susswein et al., 2002), the actual triggering process for spontaneous B63 bursts and consequently the irregular emission of buccal CPG output remains unknown.”

   Lines 98-106 now state: “This essential role played by B63 is partly mediated by sustained, large amplitude membrane depolarizations that activate high frequency bursts of action potentials (Figure 1C; see also Nargeot et al., 2009). Consistent with these underlying depolarizations arising from a bistable membrane property (Russell and Hartline, 1978), a brief intracellular injection of depolarizing current into an otherwise silent B63 neuron can trigger a depolarizing plateau and accompanying burst discharge that far outlasts the initiating stimulus (Figure 1D). The stimulated B63 in turn activates a similar burst-generating depolarization in the contralateral B63 cell and elicits a single BMP by the buccal CPG network.”

   Line 177-180: now explicitly draws conclusion on the endogenous origin of B63 plateau potentials: “Significantly, the continued expression of this burst-generating capability under functional synaptic isolation confirmed that the underlying plateau potentials, as suggested by evidence reported above (see Figure 1C), arose from an endogenous membrane property of the B63 neurons themselves.”

   The reviewer also raises the issue of the decision-making process leading to buccal motor pattern genesis. Our previous study found that experimental depolarization or hyperpolarization of B63 with intracellular current injection respectively either triggers or prevents buccal motor pattern genesis (Nargeot et al., 2009). Earlier modeling data indicated how experimentally elicited activity in a passive B63 neuron could trigger plateau potential generation in the electrically coupled B31/32 cells (Susswein et al., 2002). The present study, essentially conducted with chemical synapses blocked, including the strong excitatory synapse from B63 to B31/32, investigated the origin of the spontaneous intrinsic, rather than extrinsically elicited activity of B63 and did not address the cell’s dynamic relationship with other network neurons. Our data indicate that B63 generates a spontaneous pacemaker activity that can trigger endogenous plateau potentials without the involvement of any extrinsic influences. That said, however, due to our experimental conditions with the functional suppression of chemical synapses, we agree with the reviewer that the present study does not establish whether B63 is uniquely sufficient in the decision process for the induction of BMPs.

   This point is now made in the revised manuscript (Lines 577-583), which states: “However, because our experiments were mainly conducted with all the network’s chemical synapses blocked, we were unable to establish whether B63’s endogenous oscillatory and plateau properties are alone sufficient in the decision process for BMP genesis. Nevertheless, in normal saline conditions with the network remaining functionally intact, in contrast to all other identified circuit cells, the B63 neuron pair are the only elements found to be necessary and sufficient for triggering motor pattern expression and resultant food-seeking movement (Hurwitz et al., 1997; Nargeot et al., 2009)”.

   The title of the revised manuscript, now “Organelle calcium-derived voltage oscillations in pacemaker neurons drive the motor program for food-seeking behavior in Aplysia” hopefully satisfies the reviewer’s concern.

   We fully acknowledge the importance of ultimately placing our present findings in context with those reported previously, especially in relation to B63’s oscillatory behavior and associative learning. However, at this stage our data are insufficient to allow such an integrated assessment, although obviously this is a major goal of our future research.

   Read the original source
3. 

   eLife

   May 21, 2021

   Reviewer #3 (Public Review):

   In this report the authors characterize a mechanism that plays a role in inducing the rhythmic depolarizations that are observed in identified neurons that are part of the feeding CPG in Aplysia. The neurons studied (B63 neurons) are of interest because previous work has established that they play an important role in triggering cycles of motor activity. Further, previous work from this group has demonstrated that activity in the B63 neurons can be modified by operant conditioning.

   The authors present this study as though previous work had established that plateau potentials generated in the B63 neurons play an important role in driving network activity. For example, in line 102 they state "This essential role played by B63 is partly mediated by a bistable membrane property, which allows the sudden switching …

   Reviewer #3 (Public Review):

   In this report the authors characterize a mechanism that plays a role in inducing the rhythmic depolarizations that are observed in identified neurons that are part of the feeding CPG in Aplysia. The neurons studied (B63 neurons) are of interest because previous work has established that they play an important role in triggering cycles of motor activity. Further, previous work from this group has demonstrated that activity in the B63 neurons can be modified by operant conditioning.

   The authors present this study as though previous work had established that plateau potentials generated in the B63 neurons play an important role in driving network activity. For example, in line 102 they state "This essential role played by B63 is partly mediated by a bistable membrane property, which allows the sudden switching of the neuron's resting membrane potential to a depolarized plateau..." To support this statement, they reference Susswein et al. 2002, which does not support this statement. In the Susswein et al. study it is the B31/32 neurons that are modeled as having plateau properties.

   If previous work has not established the role of the B63 plateau potentials, the only data that speak to this issue are presumably in the current report. In this study the authors do provide data that indicate that the B63 neurons generate low amplitude oscillations that are not likely to depend on input from the electrically coupled neurons studied (notably B31). The authors also show that in some instances, these depolarizations do trigger plateau potentials in B63. It is, however, not clear that the B63 generated plateau potentials are then responsible for triggering network activity (e.g., as opposed to a situation where depolarizing input from B63 triggers plateau potentials in B31/32 and the depolarization in B31/32 drives the rest of the feeding circuit). For example, in Figs. 6A and Supplemental Fig. 4A it does not appear that the plateau depolarization in B63 is being transmitted to other electrically coupled neurons to any large extent.

   A clarification of this issue is important because it potentially impacts thinking concerning how 'decision making' is occurring. If decision making means induction of a motor program and this does not occur unless the depolarization in B63 is transmitted to B31/32, the process is more complicated than what the manuscript currently suggests.

   The title is misleading since there are no studies of behavior in this report.

   In part, interest in the mechanisms that drive spontaneous oscillatory activity in the B63 neurons stems from the overall context of this work. Namely the authors have previously established that oscillatory activity can be modified through associative learning. In the Sieling et al. 2014 study they demonstrate that two aspects of plasticity are accounted for by changes in synaptic properties and an effect on a leak current. For readers trying to understand this body of work as a whole, the Discussion should more clearly indicated how the results of the present study integrate with these previous findings.

   Read the original source
4. 

   eLife

   May 21, 2021

   Reviewer #2 (Public Review):

   Motivated behaviors, such as food seeking when hungry, can also occur spontaneously at irregular intervals. Understanding how this irregular expression arises is important for understanding behavior and is relatively little investigated. The present work thus addresses an important and under-investigated area in neurobiology. Its demonstration of a potential cellular mechanism for irregular behavioral production has wide relevance, ranging from how cells make "decisions" to how whole organisms do so.

   Intact Aplysia occasionally produce bites even in the absence of food, and isolated buccal ganglia (which contain the biting central generator circuit) will occasionally spontaneously produce fictive bite motor patterns. The activity of central pattern generator networks has almost exclusively been ascribed to …

   Reviewer #2 (Public Review):

   Motivated behaviors, such as food seeking when hungry, can also occur spontaneously at irregular intervals. Understanding how this irregular expression arises is important for understanding behavior and is relatively little investigated. The present work thus addresses an important and under-investigated area in neurobiology. Its demonstration of a potential cellular mechanism for irregular behavioral production has wide relevance, ranging from how cells make "decisions" to how whole organisms do so.

   Intact Aplysia occasionally produce bites even in the absence of food, and isolated buccal ganglia (which contain the biting central generator circuit) will occasionally spontaneously produce fictive bite motor patterns. The activity of central pattern generator networks has almost exclusively been ascribed to the actions of the voltage-gated channels in the network neuron cell membranes and the synaptic connectivity among the network's neurons. Bédécarrats et al. show that a small, highly regular cell membrane voltage oscillation occurs in a neuron (B63) in the biting neural network, and that occasionally this oscillation becomes large enough to trigger a plateau potential in B63 and a single fictive bite from the entire circuit. They show that this oscillation is not due to cell membrane voltage dependent conductances, but instead from process involving the endoplasmic reticulum, mitochondria, or both. Although organelle-driven changes in cellular or tissue activity have been observed in other cell types, this is, to my knowledge, its first observation in a neural network. These data thus are potentially of great importance in understanding how neural networks function, most of which do not show the great regularity of central pattern generated behaviors.

   The presented data seem, to me, strong with respect to the small potential oscillations not being generated by voltage-dependent cell membrane conductances, and somehow involving the intracellular organelles. What is less clear to me is how local release of Ca from endoplasmic reticulum or mitochondria would result in changes in ion composition under the cell membrane, which is what gives rise to the cell membrane potential. Ca is highly buffered in the cytoplasm. It is thus unclear to me that free Ca would remain so for any length of time after release. It does, of course, in muscles, but these are evolved for this to occur. The authors themselves raise a variant of these concerns in the Discussion when considering how the B63 cell membrane voltage oscillations are transmitted to neurons electrically coupled to B63, invoking as a possibility Ca activation of second messengers, which would then themselves be responsible for the cell to cell communication. It seems to me that the same concerns arise with respect to how Ca release at sites distant from the cell membrane could charge the membrane's capacitance.

   A second remarkable observation is that B63 depolarization and firing does not reset the organelle-derived slow oscillation. B63 firing should result in substantial Ca concentration changes, at least in a shell under the cell membrane, so a possible feedback mechanism can be imagined. Most biological processes contain multiple feedback process that link cause and effect (e.g., the sequential current activations that return a cell to rest after an action potential, the interactions between sympathetic and parasympathetic system activity that maintain functionally proper body activation, the interactions that regulate hormonal levels). One possibility the authors mention is that the organelle-derived oscillation is used only for intermittent bite activities, and in feeding bites are instead generated solely by standard cell-membrane voltage-dependent processes. Regardless, it is a striking observation that merits additional investigation.

   These issues, however, do not change the data, which show a clear association of disruption of endoplasmic reticulum and mitochondrial function and cessation of the cell membrane voltage oscillation. Nor is it reasonable to expect an article like this, showing an organelle-driven cell membrane potential oscillation for the first time in a neuron, to describe every aspect of the mechanism by which it occurs. Indeed, it is a measure of the article's interest that it prompts such thinking. It will be very interesting to see the effects of similar organelle-disrupting treatment on the activity of other well-defined neural networks.

   Read the original source
5. 

   eLife

   May 21, 2021

   Reviewer #1 (Public Review):

   This paper presents evidence that membrane potential excursions called plateau potentials are driven by subthreshold oscillation generated by calcium fluxes in mitochondria. Pharmacological and electrophysiological methods were used to deduce that calcium waves were generated in a bilateral pair of electrically-coupled neurons and spread to additional neurons that were coupled to that pair. The identified neurons in Aplysia allow for the detailed measurements needed to determine this mechanism.

   Read the original source
6. 

   eLife

   May 21, 2021

   Evaluation Summary:

   In this report the authors demonstrate convincingly that rhythmic activity in neurons that are part of the feeding CPG in Aplysia is generated via an unusual mechanism, organelle-derived intracellular calcium fluxes. The neurons that are studied (B63 neurons) play an important role in triggering cycles of motor activity and previous work from this group has demonstrated that activity in these neurons can be modified by operant conditioning. The paper was very well received by the reviewers who were impressed by the novelty of the mechanism uncovered as a driver of a fictive motor program and thus likely behavior.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer …

   Evaluation Summary:

   In this report the authors demonstrate convincingly that rhythmic activity in neurons that are part of the feeding CPG in Aplysia is generated via an unusual mechanism, organelle-derived intracellular calcium fluxes. The neurons that are studied (B63 neurons) play an important role in triggering cycles of motor activity and previous work from this group has demonstrated that activity in these neurons can be modified by operant conditioning. The paper was very well received by the reviewers who were impressed by the novelty of the mechanism uncovered as a driver of a fictive motor program and thus likely behavior.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their names with the authors.)

   Read the original source
7. 

   Version published to 10.1101/2021.03.30.437701v1 on bioRxiv

   Mar 31, 2021