The Ca2+-activated chloride channel anoctamin-2 mediates spike-frequency adaptation and regulates sensory transmission in thalamocortical neurons

Neuronal firing patterns, which are crucial for determining the nature of encoded information, have been widely studied; however, the molecular identity and cellular mechanisms of spike-frequency adaptation are still not fully understood. Here we show that spike-frequency adaptation in thalamocortical (TC) neurons is mediated by the Ca2+-activated Cl− channel (CACC) anoctamin-2 (ANO2). Knockdown of ANO2 in TC neurons results in significantly reduced spike-frequency adaptation along with increased tonic spiking. Moreover, thalamus-specific knockdown of ANO2 increases visceral pain responses. These results indicate that ANO2 contributes to reductions in spike generation in highly activated TC neurons and thereby restricts persistent information transmission.

In the present paper Ha and co-worker present a new and unexpected mechanistic basis for the phenomenon of spike frequency adaptation in neurons. The authors combine electrophysiological recordings, specific pharmacology, gene knock-down strategies, expression analysis and behavioral tests to identify the Ca2+-activated Cl-channel anoctamin-2 (ANO2) as the major basis for spike frequency adaptation in thalamocortical (TC) neurons of the ventrobasal thalamic complex (VB). This finding is novel and of high interest for the field of neuroscience. In general the paper is technically sound and well written. The data presented appear to be of high quality. Nevertheless there are some shortcomings which dampen my enthusiasm.
1) It would be important to evaluate to what extend the ANO2-dependent mechanism shown here is a general mechanism of spike frequency adaptation in neurons. Is it specific for VB or thalamus? Is it present in all adapting neuronal cell types? Since TC neurons in different nuclei and species usually show only moderate spike frequency adaptation [1][2][3], it would be essential to see whether the results are representative for TC neurons from other thalamic nuclei and more importantly for neurons known for prominent spike frequency adaptation, like hippocampal pyramidal cells (which even stop firing during long depolarizing steps).
2) It has been shown that the postnatal action potential firing maturation has a slow time course with TC neurons from 3-7 month old mice revealing firing rates up to 200 Hz for strong (700 pA) depolarizations [3]. Therefore complete input / output curves should be shown and older animals should be tested.
3) The regulation of tonic firing by ANO2 in response to continuous current injection has been analyzed here. This form of depolarization may recruit Ca2+ sources in a different way compared to synaptic stimulation. It should be therefore determined whether ANO2 is also influencing the response to repeated phasic excitatory postsynaptic potential (EPSP)-like stimuli [3]. 4) In the mammalian brain, external Ca2+ concentrations range from 1.5 to 2.0 mM 4. In the present paper an elevated extracellular Ca2+ concentration was used. It should be tested whether a more physiological concentration (like 1.8 mM) sustains this new mechanism. This is interesting since the authors point out several times that TC neurons have to be highly activated to see the effect. This may therefore be a mechanism which is also relevant for pathophysiological conditions in the brain. 5) The mechanisms of tonic firing and spike frequency adaptation in TC neurons are complex and have been addressed to some degree before. It has been shown that Kv3.2, Kv1, SK, BK, KCNQ and T-type Ca2+ channels play an important role thereby pointing to the contribution of additional mechanisms [3,[5][6][7]. While this not fully reflected in the discussion of the paper, the results presented here may also point in this direction. Knockdown of ANO2 increased the adaptation index from 50% to 77% (difference of 27%) thereby leaving another 23% open for additional (partially Ca2+-independent) mechanisms. Therefore other mechanisms like BK channels may be tested. Although iberiotoxin is not influencing the Ca2+-dependent tail current, spike frequency adaptation may be decreased. In addition the experiments shown in Fig. 3e may indicate that a single experiment in which apamin decreased the adaptation index may have prevented significant results. The number of experiments should be increased to address this possibility. 6) For the behavioral tests the injection sites should be documented for all recordings. Were the effects different for VPL (ventral posterior lateral nucleus) and VPM (ventral posterior medial nucleus) hits since both regions of the VB serve slightly different functions? It should be noted that also a reduction of specific high-frequency bursting in TC neurons of the VB may be associated with increased pain responses [7][8][9]. In this respect it would be interesting to see whether the intra-burst frequency of spikes (cf., Fig. 7j) was altered by ANO2 knock-down. 7) The membrane properties of TC neurons should be described in more detail (e.g., resting membrane potential, input resistance, membrane capacitance). What were the criteria for including cells into the statistical analysis? Was there any delayed onset of firing, since this is a characteristic feature of TC neurons in different species? In addition more methodological details should be stated. How was liquid junction potential correction done? Was this correction different for different solutions? What is the expected intracellular Ca2+ concentration for the used recording solutions? What was the recording temperature? Which test was used to determine normal distribution of data points? Which tests were performed to assess antibody specificity?
The manuscript by Ha et al describes some generally well conducted experiments that conclude that the modulation of action potential firing frequency is due to the activation of the calcium activated chloride channel ANO2. This finding is of significant interest as cation channels (e.g. calcium-activated potassium and HCN channels) typically modulate firing frequency, and this is the first demonstration that a chloride channels plays such a role in thalamic neurons. The manuscript is well written and it usually easy to follow the logic of each experiment. In general the data appear to be of good quality and the conclusions sound. I have some, mostly minor, comments and suggestions. 1. Details should be added to indicate the extent of the region infected with AAV. This will be important for interpretation of the behavioral experiments, as spread of the virus could weaken the conclusion that the behavioral differences resulted from knockdown of ANO2 in thalamic neurons. The inset in Figure 8a showing mCherry staining is not described or discussed in the results text. The percentage of thalamic neurons expressing mCherry should be given with an indication of the numbers/percentage of ANO2-expressing cells infected by the virus. 2. It would be instructive to determine the behavioral effect of ANO2 knockdown on an acute pain stimulus (e.g. hot plate or tail flick) that has a short duration where repetitive neuronal firing may be less evident. 3. Standard controls should be used for the immunohistochemistry -staining should be shown for no ANO2 primary antibody (in e.g. Figure 6c). How specific is the antibody staining for neurons? Are there data with this antibody from ANO2-null mice? 4. Some of the IHC staining is difficult to see (e.g. Best1). Can the images be improved? 5. Was the effect of removing calcium from the external solution on spike frequency reversible (page5/6, Fig 1a) or is this an irreversible effect? 6. As niflumic acid can act on some cation channels the conclusion that the NFA inhibited current is an anion current (page 7) is premature. This is only confirmed when the reversal potentials are measured (page 8) 7. The statement (page 17) that ANO2 is localized to the soma of thalamic neurons, unlike SK channels, needs clarification. Does this mean that ANO2 is not expressed on dendrites? ANO2 has been reported to be targeted to dendrites in Purkinje neurons (Zhang et al 2015 PLoS One 10: e0142160). 8. The hyperpolarizing protocol used for Figure 7k should be described in the methods (hyperpolarized to what membrane potential for what time period). At present there is a comment in the figure legend about hyperpolarizing to ~-80mV but the results section mentions 'the hyperpolarizing protocol). 9. What is the free calcium concentration in the internal solution with 5Ca-EGTA-NMDG? 10. State the supplier for the Mini Analysis software. 11. A GFAP antibody was used but details are not given in the methods section. 12. Define mANO1 and mANO2 as mouse ANO1 and ANO2 13. Define APA when first introduced (page 8) 14. Page 18. "...actively emits Cl<sup>-</sup>." 15. Extrudes or effluxes would be a more appropriate word than 'emit'

Reponses to Reviews
Reviewers' comments:

Reviewer #1 (Remarks to the Author):
This is an interesting and potentially important study in its field. The authors show that spikefrequency adaptation in thalamocortical neurons is mediated by the Ca 2+ -activated Clchannel (CACC) anoctamin-2 (ANO2). This finding is (to the best knowledge of the present referee) a novel one.
However, the paper has shortcomings which should be carefully addressed by the authors. Most of these can be overcome by text revisions, but perhaps one experiment could be added as explained in the first point in the list below.
A: We thank the reviewer for positive and insightful comments on our manuscript and for constructive suggestions to strengthen the manuscript. We tried our best to answer all inquiries and concerns that were raised by the reviewer and edited the figures and manuscript to address the suggestions. The edited text in the revised manuscript is outlined in blue for clarity. Our specific responses to each point that was raised by the reviewers are also outlined below in blue for clarity.
Major comments 1) Given the key role of intracellular Cl (Cli), it would be interesting if the authors would provide data on spike-frequency adaptation (SFA) under conditions where Cli would be intact and prone to activitydependent variation. The gramicidin patch technique has been used only in experiments using GABA to probe the steady-state level of Cli and the resulting Ecl. For studying SFA under conditions with intact Cli, on-cell patching might also be an effective approach. Anyway, it is entirely possible that even under in vitro conditions, somatic Cli would show activity-dependent (spike-firing dependent) fluctuations and thereby modulate both the ANO2-mediated current and, consequently, SFA. I do not insist on this experiment, but this is a point that should at least be taken up in the Discussion.
A. We thank the reviewer for the suggestions to strengthen the main idea of our manuscript by confirming the ANO2-mediated spike frequency adaptation under conditions with intact [Cl -]in. In response to the reviewer's suggestion, we performed the gramicidin-perforated patch clamp to measure the ANO2-mediated spike frequency adaptation, although the reviewer did not insist on the experiment. This was done on TC neurons that were infected with either AAV-shANO2 or AAV-Scr.
Then we compared the spike frequency adaptation of those TC neurons in which [Cl -]in could be intact. 2) Surprisingly, the authors do not clearly state the major, general novelty of their work in the Abstract.
On p. 9 they state that "no previous study has described a CACC that is involved in mAHP currents and neuronal spike patterns in the central nervous system". This important point should be included into the Abstract.
A. We thank the reviewer for this positive comment on the significance of our manuscript. In response to the reviewer's comment, we modified the abstract to state the novelty of our work as suggested by the reviewer. (Page 2 line 11 in the revised text) 3) The adaptation index used presently will be extremely sensitive to minor changes in the first and last interspike interval. I would suggest using an average of at least two if not three ISIs (i.e., the initial 2 or 3 and the last 2 or 3 ISIs) two make the parameter more robust.
A. We agree with the reviewer on the point that calculation of the adaptation index could be sensitive to the small changes in the first and last ISIs. In response to the reviewer's suggestion, we adjusted the analysis method of the adaptation index to use an average of the first two and the last two ISIs and newly analyzed results were added to the revised manuscript. We also revised the text as "The adaptation index (AI) was obtained by dividing the average of the first 2 ISIs by average of the last 2 ISIs." and replaced. (Revised Fig. 1d, 3d, 7f. Page 6 line 3 in the revised text) 4) The authors should explain why pain responses only have been studied (section starting on p. 14).
In thalamocortical sensory functions, other sensory modalities take a much more central place in the literature.
A. We apologize for the insufficient explanation on why we conducted visceral pain test. The reason is that it was previously reported that the altered firing properties of VB thalamic functions could influence the visceral pain responses 1 . Based on this, we predicted that the pain response could be a good tool to test the thalamic function affected by the changes in spiking frequency of TC neurons (increased AHP and decreased spike frequency adaptation). Here we measured pain responses to prove our idea that changes in the firing frequency adaptation in TC neurons can modulate the thalamic function as ANO2 is highly expressed in VB nuclei in TC region where the pain signals are known to be relayed. In general, the sensory information from pain and temperature receptors on peripheral body is transmitted through the spinothalamic tract. The spinothalamic tract is a representative sensory pathway from the periphery to the thalamus, specifically ventroposterior medial (VPM) and lateral (VPL) nucleus. Although VB thalamic function is closely related with pain behavior, we agree with the reviewer's comments that other sensory modalities could be affected, too.
In the future, it would be very interesting to study the function of ANO2 in modulating other sensory modalities. In response to the reviewer's suggestion, this rationale was added in the discussion part (Page 20 line 10 in the revised text). 5) Discussion, start of: the data in Fig. 2d do not indicate that this is an outwardly rectifying current as stated in the text. Because somatic Ca2+ is a key player, the authors should provide information on the possible effects (or attempts to minimize them) of the exogenous buffers. This would be relevant for the discussion on p. 19, 2nd paragraph.
A. As the reviewer mentioned, it was important not to disturb the intracellular Ca 2+ while we measured Ca 2+ -activated current and Ca 2+ -mediated firing frequency adaptation. Based on this, we used very low levels of EGTA to minimally chelate the Ca 2+ contamination from other chemicals and maintain the same level of free Ca 2+ level in the intrapipette solution. We used 0.02 mM EGTA in the experiments to measure spike firing rates and used 0.1 mM EGTA to measure the Ca 2+ activated AHP current. This low EGTA concentration in intrapipette solution was often used by many studies to study the Ca 2+current-mediated changes in AHP current or firing pattern changes with minimal disturbance onto the intracellular Ca 2+ level. 2, 3, 4 7) p. 18: Outward rectification has nothing to do with a channel's capability to hyperpolarize the plasma membrane. While this kind of a misconception is frequent, it is the reversal potential only which dictates the polarity of a channel-mediated response.
A. Although we agree with the reviewer that outward rectification has nothing to do with a channel's capability to hyperpolarize the plasma membrane, we never stated that outward rectification of ANO2 contributes to hyperpolarization. Instead, we stated that outward rectification limits the ability for neurons to depolarize ("Therefore, ANO2 would generate relatively large outward currents in TC neurons with an endogenous ionic content at the depolarized membrane potential.", Page 18 line 15, Discussion). We believe that outward rectification can act as a ceiling for neurons not to go beyond a certain depolarized membrane potential due to ANO2. We also agree with the reviewer on the comment that the reversal potential with endogenous intracellular chloride ion concentration would be the most important factor to determine the hyperpolarizing function of ANO2. Therefore, we first discussed the reversal potential with endogenous intracellular chloride ion concentration first and then addressed the outward rectification in the revised Discussion part. (Page 18 line 22 in the revised text) 8) There's a major confusion with regard to Fig. 7 where, for instance, panels j and k are not even 9) The style of writing is verbose and sometimes the flow is somewhat clumsy. I would urge the authors to aim at a much more compact style of writing throughout the ms (especially the Results section). This would also make the paper shorter.
A. We took the reviewer's comments into account and modified the word choice, notably the result section. We revised this section explaining the rationale to do each experiment in a more concise manner and brought our focus more to the meaning of the results. We also substantially revised the description of the behavioral test. b) The running title is not well-formulated. I would suggest "ANO2 and spike frequency adaptation" or something along these lines ("dampening information transfer" is vague/ambiguous) A: As per the reviewer's suggestion, we modified the running title to "ANO2-mediated spike-frequency adaptation regulates information transfer from thalamus to cortex". A. We agree with the reviewer's suggestion. We additionally supplemented a reference article published recently (2014). (Page 4 line 2) d) p. 6, line from below: "holding current" should be "holding potential" A. We apologize to the reviewer for the incorrect expression and corrected the sentence as suggested.
(Page 6 line 22) e) p. 6, last line: these were surely not "0 mV prepulse steps" but rather "prepulse steps to 0 mV" A. We corrected the sentence as suggested by the reviewer in the revised manuscript. ( The authors might also note that ANO2 is a very unspecific anion channel which may carry substantial A. We thank the reviewer for the critical comments to allow us to strengthen the main findings of our manuscript. We corrected the endogenous [Cl -]in to 5.4 mM with a correction for the bicarbonate permeability. In following, we corrected the endogenous [Cl -]in in the results addressing the bicarbonate permeability issue with references (Page 10 line 13).
As the reviewer states, it was known that Ca 2+ -activated Clchannels have the HCO3permeability. A. We corrected the grammar in the sentence as suggested by the reviewer. (Page 9 line 1) -p. 9, line 10 from below, "our previous data" should be replaced by "the data shown above" A. We revised the sentence as suggested by the reviewer. (Page 9 line 18) -p. 9, line 7 from below: "intact" rather than "endogenous" A. We edited the sentence as suggested by the reviewer. -p. 14: the abbreviation LTS is used only once after this and is therefore obsolete A. We removed the abbreviation from the text as suggested by the reviewer (Page 14 line 16).

Responses to Reviewer2
Reviewers' comments: Reviewer #2 (Remarks to the Author): In the present paper Ha and co-worker present a new and unexpected mechanistic basis for the phenomenon of spike frequency adaptation in neurons. The authors combine electrophysiological recordings, specific pharmacology, gene knock-down strategies, expression analysis and behavioral tests to identify the Ca 2+ -activated Cl-channel anoctamin-2 (ANO2) as the major basis for spike frequency adaptation in thalamocortical (TC) neurons of the ventrobasal thalamic complex (VB). This finding is novel and of high interest for the field of neuroscience. In general the paper is technically sound and well written. The data presented appear to be of high quality. Nevertheless there are some shortcomings which dampen my enthusiasm.
A: We thank the reviewer for the kind and insightful comments on our manuscript. We are especially grateful to the reviewer for many constructive suggestions to allow us to strengthen the manuscript.
We have tried our best to answer all inquiries and concerns raised by the reviewer and edited the figures and manuscript to address these suggestions. Here, we have done all the experiments suggested by the reviewer. The edited text in the revised manuscript is outlined in blue for clarity. Our specific responses to each point raised by the reviewers are also outlined below in blue for clarity.
1) It would be important to evaluate to what extend the ANO2-dependent mechanism shown here is a general mechanism of spike frequency adaptation in neurons. Is it specific for VB or thalamus? Is it present in all adapting neuronal cell types? Since TC neurons in different nuclei and species usually show only moderate spike frequency adaptation 1, 2, 3 it would be essential to see whether the results are representative for TC neurons from other thalamic nuclei and more importantly for neurons known for prominent spike frequency adaptation, like hippocampal pyramidal cells (which even stop firing during long depolarizing steps).
A. We appreciate the reviewer for the comments concerning the scope of our manuscript. In this study, we examined how the spike frequency adaptation is generated in thalamic neurons in the VB nuclei (included VPM and VPL), that relay somatosensory information. Unfortunately, we have not specified sub-region of thalamus in the current study (it might be interesting to the future work). Nevertheless, as the reviewer suggested, we tested the spike frequency adaptation in hippocampal neurons to see if ANO2-mediated spike frequency adaptation could be a general mechanism. As shown in the figure below, CA1 pyramidal neurons infected with AAV-shANO2 displayed the reduced spike frequency adaptation compared to CA1 neurons infected with AAV-Scr, though a larger portion of spike frequency adaptation remained. These results suggest that other channels such as potassium channels as known in many previous studies contribute a lot to the spike adaptation in CA1 neurons 4,5 . We hope that this preliminary data would answer some of the questions raised by the reviewer. We also hope that the reviewer would generously understand that we will not include these results in this manuscript because this is not the scope of our current work.
Because we studied the spike frequency adaptation of the thalamocortical neurons, not other regions of brain, we limited the mechanism of spike frequency adaptation to thalamic neurons. This result was 2) It has been shown that the postnatal action potential firing maturation has a slow time course with TC neurons from 3-7 month old mice revealing firing rates up to 200 Hz for strong (700 pA) depolarizations 3 . Therefore complete input / output curves should be shown and older animals should be tested.
A. We absolutely agree with the reviewer on the comment that the input-output curve of firing rate in TC neurons should be shown to strengthen our data. In response to the reviewer's comment, the complete input-output curves measured from TC neurons in 5-6 weeks old mice were added to the revised Fig. 1e and also in the revised text (Page 6 line 7).
We also accepted the reviewer's suggestion to measure the firing rate in older animals and measured spike frequency adaptation of TC neurons from 4-month-old mice. Input-output curves of spike firing rates, comparing between 5~6-week-old and 4-month-old mice, have been added to the revised supplementary Fig. S2. TC neurons from older mice also showed firing frequency adaptation pattern similar to 5-6 week old TC neurons (Suppl. Fig. S2a-b) with similar adaptation index (Suppl. Fig. S2c).
Input-output curve showed that there is no significant difference in spike firing rates of TC neurons from 5-week-old and 4-month-old mice with depolarizing current steps up to 400 pA. The stronger depolarizing current steps ranging from 500 to 700 pA induced a further increase in firing frequency in TC neurons from 4-month-old mice whereas TC neurons from 5-week old mice could not increase the firing frequency substantially. This tendency agrees with the previous report 3 . We included this reference in the revised manuscript (revised Suppl. Fig. S2).
3) The regulation of tonic firing by ANO2 in response to continuous current injection has been analyzed here. This form of depolarization may recruit Ca 2+ sources in a different way compared to synaptic stimulation. It should be therefore determined whether ANO2 is also influencing the response to repeated phasic excitatory postsynaptic potential (EPSP)-like stimuli 3 .
A. We appreciate the reviewer for suggesting an experiment to apply repeated phasic EPSP-like stimuli. To apply EPSP-like stimuli, we generated a model EPSC using double exponential equation as shown in equation (1) and (2) ln .
Rising and decay tau were 0.5 and 4.07, respectively 6 , adopted from as shown below (Suppl. Fig.   S3a). EPSP-like stimuli were delivered to TC neurons by applying various frequencies (20, 50 and 100 Hz) and amplitudes (ranging from 200 to 1000 pA) of the modeled EPSC. We quantified the adaptation index injecting EPSP-like stimulation of 100 Hz in supplementary figure S3. A more detailed method is written for EPSP-like stimulation in the method section. TC neurons infected with AAV-shANO2 displayed a tendency of an increase of adaptation index value compared to those infected with AAV-Scr with an insignificant p-value. (Suppl. Fig. S3c, p=0.07). Also in 20 Hz and 50 Hz, the spike adaptation was prominent as shown below and the firing pattern was similar to previous work 3 , implying that firing frequency adaptation may occur responding to physiological excitatory inputs.
100 Hz stimulation (Suppl. Fig. 3) 20 and 50 Hz stimulation 4) In the mammalian brain, external Ca 2+ concentrations range from 1.5 to 2.0 mM 7 . In the present paper an elevated extracellular Ca 2+ concentration was used. It should be tested whether a more physiological concentration (like 1.8 mM) sustains this new mechanism. This is interesting since the authors point out several times that TC neurons have to be highly activated to see the effect. This may therefore be a mechanism which is also relevant for pathophysiological conditions in the brain.
A. In response to the reviewer's comment, we executed the experiments to measure spike frequency adaptation under more physiological Ca 2+ concentration (1.5 -2.0mM in the brain). Under the extracellular buffer containing 1.8mM Ca 2+ concentration, we obtained a similar value of adaptation index from TC neurons shown in revised Suppl. Fig. S1. This result suggested that the phenomenon of spike adaption is also shown under physiological conditions (Revised Suppl. Fig. S1 and Page 5 line 16 in the revised text).
"TC neurons have to be highly activated to see the effect" means that the Ca 2+ mediated spike frequency adaptation is more distinctively visible when TC neurons make multiple spikes. Therefore, we believe that this spike frequency adaptation works as a type of suppression of spike generation at high frequency when there are many of excitatory inputs, which could be persistent pain stimuli, to TC neurons in the brain.

5) The mechanisms of tonic firing and spike frequency adaptation in TC neurons are complex and
have been addressed to some degree before. It has been shown that Kv3.2, Kv1, SK, BK, KCNQ and T-type Ca 2+ channels play an important role thereby pointing to the contribution of additional mechanisms 3,8,9,10 . While this not fully reflected in the discussion of the paper, the results presented here may also point in this direction. Knockdown of ANO2 increased the adaptation index from 50% to 77% (difference of 27%) thereby leaving another 23% open for additional (partially Ca 2+ -independent) mechanisms. Therefore other mechanisms like BK channels may be tested. Although iberiotoxin is not influencing the Ca 2+ -dependent tail current, spike frequency adaptation may be decreased.
A. We absolutely agree with the reviewer and were aware of these previous studies. We examined the possibility that many kinds of potassium channels and T-type Ca 2+ channels could affect this phenomenon. Firstly, as we showed in Fig 7i-  A. In response to the reviewer's comment, we provided schematic diagram of injection sites of the AAV virus from the brains of all mice used for the visceral pain test (Suppl. Fig. S7a). Furthermore, we displayed an example for the extent of the AAV-infected region for each animal measured by the posthoc staining of the AAV-infected area (mCherry-expressing area) to total VB region (Supple Fig. S7b).
We confirmed that the broad VB region, including both the VPL and VPM, was covered by 70.2  3.9 %. We excluded the mice whose AAV-infected VB region is below 50% because, as you could see in the viral infection area (S7b), single injection of virus infected both the VPL and VPM areas. We could not distinguish the different effect of VPL vs. VPM. The reviewer's comment on sub-regional effect on behavior would be very interesting for future work.
It should be noted that also a reduction of specific high-frequency bursting in TC neurons of the VB may be associated with increased pain responses 8,11,12 . In this respect it would be interesting to see whether the intra-burst frequency of spikes (cf., Fig. 7j) was altered by ANO2 knock-down.
A. In the response to the reviewer's concern whether or not intra-burst frequency of spikes was altered by ANO2 knock-down, we measured the intra-burst spike numbers and the intra-burst frequency under the condition of viral infection (AAV-Scr and AAV-shANO2). Because we were already aware of previous literatures 8,11,12 indicated by the reviewer, we showed the burst firing in TC neurons in pain response. As a result, there was no difference between the two groups. These points were added in the revised manuscript (Fig. 7i-j).
7) The membrane properties of TC neurons should be described in more detail (e.g., resting membrane potential, input resistance, membrane capacitance). What were the criteria for including cells into the statistical analysis?
A. We agree with the reviewer on the comment and added Supplementary A. We did not observe any delayed onset of firing in mouse TC neurons.
In addition more methodological details should be stated. How was liquid junction potential correction done? Was this correction different for different solutions?
A. We thank the reviewer for the comment on junction potential. Through experimentation, we corrected the junction potential for each intrapipette (i.p.) solution and extracelluar buffer pair.
Therefore, the corrected junction potential values are different in each i.p. solution which can be found in the Fig. 2d  Which test was used to determine normal distribution of data points?
A. We used Ryan-Joiner normality test to check whether data points have normal distribution to be suitable for t-test. This test assesses normality by calculating the correlation between original data and the normal scores of data, and the data points are seemed to be normal when the correlation coefficient is near 1.
Which tests were performed to assess antibody specificity?
A. We confirmed specificity of ANO2 antibody by performing western blot and immunohistochemistry.
In the western blot, we observed positive ANO2 signal in olfactory epithelium which is a known region of ANO2 expression, and ANO2-overexpressing NIH/3T3 cells, but not in normal NIH/3T3 cells (Figure 6a and 6b). In immunohistochemistry, we observed an ANO2 signal in AAV-Scr-infected TC neurons, but the signal was significantly reduced in AAV-shANO2-infected TC neurons (Figure 6d and   6e) indicating that the ANO2 antibody shows high specificity for the ANO2 protein. In fact, this antibody was provided by the research group which made ANO2 knockout mice and tested this antibody to the knockout mice 16 . Also, we examined the specificity of the secondary antibody to test its non-specific binding and obtained the supplementary figure S6 which indicates that no signal was observed in immunostaining data when the primary antibody was not used (Suppl. Fig. S6). We wish that the reviewer would agree with the antibody specificity from the results of several approaches and added results in the revised manuscript.
Additional points P. 11: What parts of the "TC region" is referred to here?
A. We are sorry for the confusion. In fact, 'TC region' implied VB nuclei of the thalamus where we recorded electrophysiological data. To prevent the misreading, we replaced "TC region" to "VB nuclei in TC region" on page 11 line 7.
P. 17: The lack of recurrent oscillatory bursting in TC neurons in the present study may be explained by the orientation (which is not stated in the method section) of the slices used for recordings.
Typically only horizontal thalamic slices reveal recurrent oscillatory bursting 17 .
A. We apologize for the lack of information in the method section.
1) We obtained horizontal thalamic slices for patch clamp recording and described this procedure in the methods section. (Page 22 line 5).
2) This is a misunderstanding in our data and the data in 17 . The recurrent oscillatory activity (Fig. 3 in 17 ) is different from the bursting in TC neurons. This type of recurrent intra-thalamic oscillatory activity can be induced by electrically stimulating the internal capsule (IC) in horizontal slices as the reviewer commented. This recurrent activity can be induced only when there are intact reciprocal synaptic inputs between TRN and TC neurons in horizontal or thalamocortical (with 45-50 degree angular slice cutting) slices with simultaneous stimulation of a large population of thalamic neurons. Then, the synchronized inhibitory inputs from TRN neurons would silence TC neurons followed by a synchronized spiking activity which probably contains many burst spikes in TC neurons. But, we cannot differentiate the clustered spikes which look like bursting from real burst spikes because they were multiunit recordings using electrodes with low impedance.
3) Here we recorded the firing pattern from individual TC neuron with stimulation onto a single TC neuron in either whole-cell or gramicidine-perforated patch configurations. Therefore, we could not observe this recurrent activity, although we recorded in horizontal slices. P. 18: How is the Ca 2+ -sensitivity of ANO2 compared to SK and BK channels?
A.. ANO2 channel is known to need Ca 2+ concentration higher than 1 M for opening 18 whereas ANO1 channels are activated at 0.1-0.3 M Ca 2+ concentration 19 . The Ca 2+ sensitivity of ANO2 is known to be almost 10-fold lower than in ANO1 channel 20 . BK and SK channels are activated at 10 M and 0.3 M Ca 2+ concentration, respectively 21 . In summary, SK channels have the highest Ca 2+ sensitivity.
ANO2, and BK channels are next as an order (ANO1 = SK > ANO2 > BK). P. 20: Is there a contribution of ANO2 to the resting membrane potential?
A. No, there is minimal contribution of ANO2 to the resting membrane potential. In fact, AAV-shAno2 had no effect on the resting membrane potential as shown in the supplementary table 1 added to the revised version of the manuscript. We think that the activity of ANO2 in the resting membrane potential may be low because intracellular Ca 2+ concentration is low at resting state and the reversal potential of Cl-channels in TC neurons with endogenous [Cl -]in was not far from the resting membrane potential as shown in Fig. 4.

Responses to Reviewer3
Reviewers' comments: Reviewer #3 (Remarks to the Author): The manuscript by Ha et al describes some generally well conducted experiments that conclude that the modulation of action potential firing frequency is due to the activation of the calcium activated chloride channel ANO2. This finding is of significant interest as cation channels (e.g. calcium-activated potassium and HCN channels) typically modulate firing frequency, and this is the first demonstration that a chloride channels plays such a role in thalamic neurons. The manuscript is well written and it usually easy to follow the logic of each experiment. In general the data appear to be of good quality and the conclusions sound.
I have some, mostly minor, comments and suggestions.
A: We appreciate the reviewer for the kind and insightful comments on our manuscript and constructive suggestions to strengthen the manuscript. We tried our best to answer all inquiries and concerns raised by the reviewer and edited the figures and manuscript to address these suggestions.
The edited text in the revised manuscript is outlined in blue for clarity. Our specific responses to each point raised by the reviewers are also outlined below in blue for clarity. total VB region. We confirmed that the viral infection area was covered by 70.2 ± 3.9 % of VB region, including both the VPL and VPM. We excluded mice whose AAV-infected VB region is below 50%. We added this result to the revised manuscript (Page 15 line 18).
2. It would be instructive to determine the behavioral effect of ANO2 knockdown on an acute pain stimulus (e.g. hot plate or tail flick) that has a short duration where repetitive neuronal firing may be less evident.
A. We respectfully agree with the reviewer on the point that it would be instructive to compare the effect of thalamic ANO2 in the acute type of pain with visceral pain which is categorized to persistent pain. Therefore, in response to the reviewer, we conducted a hot plate test, one of the acute pain tests, in AAV-Scr and AAV-shANO2 mice. In the results, there was no difference in the withdrawal latency of paws from hot plate (shown below, added to Supplementary Fig. S8). We added this point to the main text that the firing frequency adaption by ANO2 channels would function in the persistent type of pain, but be less evident in the acute type of pain. (Page 16 line 6) We thank the reviewer for the constructive suggestion to let us strengthen our manuscript.
3. Standard controls should be used for the immunohistochemistry -staining should be shown for no ANO2 primary antibody (in e.g. Figure 6c). How specific is the antibody staining for neurons? Are there data with this antibody from ANO2-null mice?
A. In response to the reviewer, we added the immunohistochemistry result by using no primary antibody as standard control in the revised supplementary Fig.S5.
In addition, we checked the antibody specificity through western blot by using ANO2-overexpressing and normal NIH/3T3 cells (main Fig.6). Also, the immunohistochemistry data on figure 6d-e indicated that AAV-shANO2-infected TC neurons have lower intensity of ANO2 than AAV-Scr-infected TC neurons, confirming the specificity of ANO2 antibody. This antibody was provided from the research group which made the knockout mice and tested this antibody on ANO2 knockout mice 1 .
4. Some of the IHC staining is difficult to see (e.g. Best1). Can the images be improved?
A. In response to the reviewer's suggestion, we replaced the BEST1 IHC staining images with new improved images. These new images were added to the revised manuscript (revised figure 5a).
5. Was the effect of removing calcium from the external solution on spike frequency reversible (page5/6, Fig 1a) or is this an irreversible effect?
A. Yes, it was fully reversible (data not shown).
6. As niflumic acid can act on some cation channels the conclusion that the NFA inhibited current is an anion current (page 7) is premature. This is only confirmed when the reversal potentials are measured (page 8) A. We agree with the reviewer on the reviewer's comment that NFA is not selective blocker. To respond to the reviewer's comment, we revised the text in that the NFA-sensitive anion current was replaced with NFA-sensitive current. We also mentioned that this current was identified as an anion current because of the reversal potential. A. We acknowledged the reviewer's comment on the location of ANO2 in TC neurons. Unlike our report prior to revision, at least, our immunohistochemical data showed the expression of ANO2 strongly localized to the soma of TC neurons. Fig. 6d shows that strong ANO2 signals are localized to the soma of TC neurons and a little bit in the very proximal part of dendrites, but not strong dendritic staining. However, we cannot exclude the possibility that we could not see the strong ANO2 signals in dendritic shape in these images due to the spreading morphology of dendrites of TC neurons compared to Purkinje neurons. We fully agree that it is difficult to conclude that ANO2 is not expressed on dendrites with regard to our data. Therefore, we toned down the statement to "appears to be" in the revised text as the reviewer concerned.  11. A GFAP antibody was used but details are not given in the methods section.
A. We apologize to the reviewer for lack of details and providing inaccurate information. To identify astrocytes, in fact, we used GFAP-GFP mice for immunohistochemistry data confirming mBEST1 expression (revised figure 5a). All the labels as GFAP were actually GFP signals. We corrected this confusion in the revised manuscript and the information about GFAP-GFP mice is added to the methods section (Page 21 line 17).
12. Define mANO1 and mANO2 as mouse ANO1 and ANO2 A. In the revised manuscript, we have included a description of mANO1 and mANO2 on page 11 line 18.

Define APA when first introduced (page 8)
A. In response to the reviewer, we decided not to use APA for the abbreviation of apamin to avoid confusion. We replaced APA to apamin in the overall text.
14. Page 18. "...actively emits Cl -." Extrudes or effluxes would be a more appropriate word than 'emit' A. In response to the reviewer, we changed the word to "effluxes" on page 18 line 21.
In the revised ms, the authors have addressed all my concerns, including the addition of novel data obtained by means of gramicidin patch-clamp which I suggested as an option to further strengthen their study.
I have only one comment: Unfortunately, I completely missed the fact that the experiments have been carried out at room temperature (see Referee 2 comments), probably because the recording temperature was not mentioned at all in the original ms. Now, the authors state that this information has been provided on p. 22, line 13. However, I cannot find this statement. I would strongly suggest that the authors would repeat some of their key experiments at a physiologically relevant temperature in vitro, around 33-34 C. The phenomena described in the present work involve a number of temperature-dependent cellular /molecular mechanisms. Moreover, a solution of the present kind, buffered with 26 HCO3/5% CO2, has a significantly lower pH at room temperature than at a more physiological one, which may have an effect in itself. I am confident that the authors are able to identify a suitable set of experiments to look at the possible effects of recording temperature.
Reviewer #2 (Remarks to the Author): The manuscript was adequately revised. I have no more concerns.
Reviewer #3 (Remarks to the Author): The authors have responded positively and modified the manuscript in response to the reviewers comments.
The questions and points that I initially raised have been addressed to my satisfaction. In my view the manuscript has been significantly strengthened by the additional experiments and revisions made by the authors.

Reviewer #1 (Remarks to the Author):
In the revised ms, the authors have addressed all my concerns, including the addition of novel data obtained by means of gramicidin patch-clamp which I suggested as an option to further strengthen their study .
A: We thank the reviewer for positive comments on our manuscript and for constructive suggestions to strengthen the manuscript. The edited text in the revised manuscript is outlined in blue for clarity. We tried our best to answer all inquiries and concerns that were raised by the reviewer and added new data to address the suggestions. Our specific responses to each point that was raised by the reviewers are also outlined below in blue for clarity.

Responses to reviewer #1's comments
I have only one comment: Unfortunately, I completely missed the fact that the experiments have been carried out at room temperature (see Referee 2 comments), probably because the recording temperature was not mentioned at all in the original ms. Now, the authors state that this information has been provided on p.
22, line 13. However, I cannot find this statement." A: We apologize for our mistake of missing the temperature information. The recording temperature is now added in p21, line 21 in the revised version.
I would strongly suggest that the authors would repeat some of their key experiments at a physiologically relevant temperature in vitro, around 33-34 C. The phenomena described in the present work involve a number of temperature-dependent cellular /molecular mechanisms." A: We understand the reviewer's concern and performed an additional experiment to confirm that Ca 2+ dependence of spike-frequency adaptation also occurs at 32C and added to Supplementary Fig. S1c,d in the revised manuscript. We also mentioned this in the main text (Page5 line16). As shown below, spike-frequency adaptation was decreased with replacement of Ca 2+ -free extracellular buffer at 32C which is almost identical to the pattern recorded at 25C. We wish that these experiments would answer the concerns raised by the reviewer.
However, to answer the concern raised by the reviewer, we performed the experiments to confirm this idea. We appreciate the reviewer for raising a critical question and wish that these experiments would answer the concerns raised by the reviewer.
Moreover, a solution of the present kind, buffered with 26 HCO3/5% CO2, has a significantly lower pH at room temperature than at a more physiological one, which may have an effect in itself. I am confident that the authors are able to identify a suitable set of experiments to look at the possible effects of recording temperature." A: In response to reviewer's comment, we immediately measured the pH of our aCSF (saturated with at 95% O2/5% CO2) at 25C and 32C. We found that the pH of our aCSF was 7.30-7.45 at both 25C and 32C. We did not find a significantly lower pH at room temperature. In addition, ANO2 is reported to be pH insensitive. Moreover, the additional experiments to Ca 2+ dependence of frequency adaptation at 32C confirmed that firing pattern of TC neuron was not temperature-dependent.
"-page 21, line 18: Subheading with "acutely dissociated brain slices": please delete the words "acutely dissociated"." A:In response to the reviewer's comment, we will delete the words "acutely dissociated."