Cooperative and acute inhibition by multiple C-terminal motifs of L-type Ca2+ channels
Peer review process
This article was accepted for publication as part of eLife's original publishing model.
History
- Version of Record published
- Accepted Manuscript published
- Accepted
- Received
Decision letter
-
Kenton J SwartzReviewing Editor; National Institutes of Health, United States
In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.
Thank you for submitting your article "Cooperative and acute inhibition by multiple C-terminal motifs of L-type Ca2+ channels" for consideration by eLife. Your article has been favorably evaluated by Richard Aldrich (Senior Editor) and three reviewers, one of whom, Kenton J Swartz (Reviewer #1), is a member of our Board of Reviewing Editors.
The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.
Summary:
This manuscript reports interactions among C-terminal structural components of L-type Ca channels IQ, PCRD and DCRD, and the effects of such interactions on voltage gated activation (VGA) and CaM mediated Ca dependent inactivation (CDI). The interactions among IQ, PCRD and DCRD can alter channel function (known as CMI) only when any two of them are tethered together by fusion protein or rapamycin-mediated heterodimerization. The spatial closeness of any two of these structural domains may strengthen their interactions enough to compete off the IQ-CaM interaction, which is suggested by experiments with reduced apoCaM and live cell FRET tw0-hybrid assays. This mechanism directly explains why the interactions among IQ, PCRD and DCRD reduce CDI. Previous studies suggest that CDI may actually close the channel by reducing activation and results in this study show that at the end of CDI and CMI the currents levels are the same, the overall results suggest that CMI may alter channel function similarly as CDI. The mechanism revealed by this manuscript explains well the different phenotypes of various L-type Ca channels. By computational modeling, the manuscript suggests physiological importance of CMI, and rapamycin-mediated heterodimerization experiments provided a novel rationale for pharmacological intervention of Ca channel function. The study is significant and the results in general provide a strong argument for the proposed mechanism.
Essential revisions:
1) An overarching and major weakness of the manuscript is the lack of clarity in the written presentation. The text requires extensive editing to be acceptable for publication in eLife. The authors need to provide clear and logical rationale for each of the experiments performed and for what the results mean. The current presentation assumes that the reader is familiar with the authors previous work, as well as an extensive body of work on CDI. The authors need to undertake a major revision of the text to make the work accessible to a broad audience, and for the authors to provide more details on how each experiment was performed. The Abstract is full of hyperbole and needs to be rewritten to convey the main findings of the study and why they are important. We emphasize that we think this is an interesting study, but that the written presentation needs a lot of work. The comments below give more specific guidance.
2) The manuscript uses a lot of jargon and abbreviations, but does not give clear definition for all of them. The authors need to clearly explain, using correct and simple English sentences, in the beginning of Results the important molecular processes and parameters, such as CMI, CDI, SCa, VGA, JCa and r50 etc., and use both text and figures to illustrate the definition, the experimental phenomena of these processes, and how they are measured.
3) Since CMI inhibits both voltage gated activation (VGA) and Ca dependent inactivation (CDI), it is too simplistic to measure CMI only based on CDI size (Figures 2, 4). It would be better to measure both VGA and CDI at various voltages and the variation in VGA and CDI changes observed with different experimental conditions need better explanation and discussion.
4) The authors tend to make conclusions by simply referring to figures without convincing analyses or clear explanation. The authors should describe the most relevant feature of the result in the text and point out how the result leads to the conclusion. The authors should also explain clearly if the conclusion is novel or seemingly at odds with previously established mechanisms.
5) Structural domains from α1F and α1D such as PCRD-F and PCRD-D are intermittently used in different experiments without justification or explanation. In the text or figure legend, prior to the description of results, briefly explain why D or F form is used; do not label D or F if the two forms do not make any difference to the result.
6) Figure 2. CMI is supposed to inhibit both activation and CDI, but in Figure 2C the right two panels only show a reduction of CDI but not activation. The authors should explain and discuss this.
7) Do the authors suggest that in Figure 2C, with IQ or IQ/PCRD, the channel can have CDI without CaM?
8) Figure 3. There is no experimental evidence that PCRD or DCRD were actually expressed separately.
9) There is a lack of clarity about the chimeric constructs used. The authors explain that they will "employ homologous DCTF from α1F with the strongest CMI among CaV1 family to construct chimeric CaV1.3 channels by fusing PCRDF-DCRDF onto IQ motif of α1DS". However, some of the experiments are conducted with PCRDD (Figures 1C, 2B, 4C, 4E, 6B) without giving an explanation or rationale.
10) The quantification of the FRET experiments is missing: what is the equation fitting FRET strength (as in Figure 4E, etc.)? Please report fitting parameters.
11) The rapamycin experiments are elegant and insightful. However, it is not clear why 300ms pulses were chosen, as it does not appear steady state for CMI (Figure 6A red current traces). This is also evident in Figure 1B, where the current does not seem to level off at 300ms. The end stage of CMI and CDI may not be the same with longer pulses (true steady-state). This point seems crucial to the authors' conclusion that CDI and CMI functionally result into indistinguishable gating at the ending stage. Experiments with longer depolarizing pulses or measurements of CDI and CMI time constants may help to discuss this point.
https://doi.org/10.7554/eLife.21989.021Author response
Essential revisions:
1) An overarching and major weakness of the manuscript is the lack of clarity in the written presentation. The text requires extensive editing to be acceptable for publication in eLife. The authors need to provide clear and logical rationale for each of the experiments performed and for what the results mean. The current presentation assumes that the reader is familiar with the authors previous work, as well as an extensive body of work on CDI. The authors need to undertake a major revision of the text to make the work accessible to a broad audience, and for the authors to provide more details on how each experiment was performed. The Abstract is full of hyperbole and needs to be rewritten to convey the main findings of the study and why they are important. We emphasize that we think this is an interesting study, but that the written presentation needs a lot of work. The comments below give more specific guidance.
We have extensively revised the text according to the suggestions and comments.
2) The manuscript uses a lot of jargon and abbreviations, but does not give clear definition for all of them. The authors need to clearly explain, using correct and simple English sentences, in the beginning of Results the important molecular processes and parameters, such as CMI, CDI, SCa, VGA, JCa and r50 etc., and use both text and figures to illustrate the definition, the experimental phenomena of these processes, and how they are measured.
We have revised the definitions by using text and figures, for the major processes of CDI, VGA and CMI, and key parameters of SCaand JCa. Also, we have provided the descriptions and illustrations for the amplitudes (I) and densities (J) of Ca2+ and/or Ba2+ currents measured at different time points, i.e., instantaneous peak, 50 ms, 300 ms and 1000 ms. Specifically,
A) Figure 1A and subsection “Provisional CMI concurrently attenuates both CDI and VGA”: descriptions on CDI and VGA;
B) The Introduction, subsection “Provisional CMI concurrently attenuates both CDI and VGA” and subsection “CMI is a unique type of inhibition but sharing similarities with CDI”: the introduction, basic characterization and eventual establishment for CMI;
C) Figure 1A and subsection “Provisional CMI concurrently attenuates both CDI and VGA”: ICaand IBa; SCaand JCa; Ipeakand Jpeak; I50and r50; D) Figure 6—figure supplement 1, Figure 6—figure supplement 2 and subsection “CMI is a unique type of inhibition but sharing similarities with CDI”: I300 and I1000.
3) Since CMI inhibits both voltage gated activation (VGA) and Ca dependent inactivation (CDI), it is too simplistic to measure CMI only based on CDI size (Figures 2, 4). It would be better to measure both VGA and CDI at various voltages and the variation in VGA and CDI changes observed with different experimental conditions need better explanation and discussion.
We have revised all related figures to include both VGA and CDI (with the indices of JCaand SCa), which are Figure 2D, Figure 2—figure supplement 1, Figure 3 and Figure 3—figure supplement 2, and Figure 4A-C. We agree with the reviewer about measuring both VGA and CDI. Technically, CDI and its index SCa are more favorable because for CaV1.3 channels CDI is independent of global Ca2+ so insensitive to buffer capacities, variations in expression levels, or the general health of the cells, thus more robust than VGA (JCa). In general, VGA and JCaare less stable even when experimental conditions are well controlled. Basically, any conclusion on JCaneeds to be carefully drawn from pronounced changes. In fact, this was one of the motivations underlying chemical-inducible CMI, for us to confidently count on each individual cell to quantify VGA and other properties. In some cases, the mild changes in JCamay lead to uncertainties which need to be confirmed by analyzing concurrent CDI changes. We have discussed these aspects with relevant data (subsection “Differential roles of DCRD and PCRD unveiled by low [apoCaM]” and Figure 2—figure supplement 1A).
4) The authors tend to make conclusions by simply referring to figures without convincing analyses or clear explanation. The authors should describe the most relevant feature of the result in the text and point out how the result leads to the conclusion. The authors should also explain clearly if the conclusion is novel or seemingly at odds with previously established mechanisms.
We have paid attention to these aspects in the revision, by providing more details of the experiments and data in the main text in addition to the figure captions, especially for Figures 4–6, which represent the major conclusions in this work (subsections “Cooperative scheme of CMI consists of three major combinations”, “Acute CMI based on rapamycin-inducible heterodimerization and cooperation” and “CMI is a unique type of inhibition but sharing similarities with CDI”). Also, we thank the reviewer for the suggestion about scrutinizing the novelty of particular results. It has been well taken to emphasize on the key findings of this work and we have revised the relevant text accordingly, to list a few:
A) Subsection “Differential roles of DCRD and PCRD unveiled by low [apoCaM]” and Figure 2C, D, surprising observations from α1DS-GX-DCRD that CDI and VGA were attenuated at low [apoCaM];
B) Subsection “Cooperative scheme of CMI consists of three major combinations” and Figure 4C, novel roles of PCRD unveiled by effective CMI (combination III);
C) Subsection “CMI is a unique type of inhibition but sharing similarities with CDI” and Figure 6A, findings about stable end-stage CDI (I300) from both versions of inducible CMI, distinct from known modalities such as run-down and small-molecule blockage;
D) Subsection “CMI is a unique type of inhibition but sharing similarities with CDI”and Figure 6C, new discovery that CMI attenuation on CDI is achieved by inhibiting the peak amplitude of Ca2+ current while maintaining its late-phase levels, which also reconciles the concurrent VGA/CDI attenuations apparently leading to contradictory effects on Ca2+ influx.
5) Structural domains from α1F and α1D such as PCRD-F and PCRD-D are intermittently used in different experiments without justification or explanation. In the text or figure legend, prior to the description of results, briefly explain why D or F form is used; do not label D or F if the two forms do not make any difference to the result.
We agree with the reviewer that it is unnecessary in the context of this work to distinguish between PCRDDand PCRDF(subsection “Provisional CMI concurrently attenuates both CDI and VGA”). We have updated the sequence alignment for homology among different CaV1 isoforms including α1D and α1F (Figure 1—figure supplement 1). And the functional roles of the two motifs in CMI are essentially very close, because PCRDD-DCRD and PCRDF-DCRD peptides attenuated α1DS similarly (Figure 4—figure supplement 1 and subsection “Cooperative scheme of CMI consists of three major combinations”). Following the suggestion, we have also simplified the nomenclature for DCRD and IQV, since in this work either only one isoform was used (DCRDF); or even different isoforms were used but functionally similar (PCRDD vs. PCDRF; IQV from CaV1.3 vs. CaV1.2). Meanwhile, details on various constructs including their original templates have been provided in the section of “Materials and methods” (“Molecular biology”).
6) Figure 2. CMI is supposed to inhibit both activation and CDI, but in Figure 2C the right two panels only show a reduction of CDI but not activation. The authors should explain and discuss this.
We thank the reviewer for the comment with specifics. As the reviewer pointed out, it is true that both CDI and activation (VGA) were inhibited in the mentioned cases. Due to the reasons as stated in item 3, we prefer to take advantages of CDI (SCa) as the major index to examine CMI effects and thus we did not show VGA data. In response to this comment and also for further assurance, we have provided full profiles of VGA (Figure 2—figure supplement 1), with JCavalues summarized in Figure 2D. These results have also been described accordingly in the text (subsection “Differential roles of DCRD and PCRD unveiled by low [apoCaM]”).
7) Do the authors suggest that in Figure 2C, with IQ or IQ/PCRD, the channel can have CDI without CaM?
BSCaMIQ, a specific apoCaM (Ca2+-free CaM) chelator, was to substantially reduce free apoCaM levels. However, in practice, BSCaMIQ cannot eliminate all the free apoCaM in cells. Since CaM is generally needed, supposedly no cells with complete elimination of apoCaM would be available for experimental assessment. We have modified the illustration (Figure 2B) accordingly. Under the residue apoCaM in realistic cells, without DCRD competition, α1DS or α1DS-PCRD is still able to produce pronounced CDI similarly as in control conditions, owing to the fact that under these conditions apoCaM binds the channels (IQV or IQV-PCRD) reasonably well (Liu et al. 2010). Although such low apoCaM would not affect α1DS or α1DS-PCRD, it does reduce the CDI of some variants besides α1D-GX-DCRD, such as the native α1DL variants containing full-length DCTs. The established principles for CaM-mediated CDI are still valid and applicable in all these cases, regardless of actual CaM levels. We have revised the text to clarify this notion (subsection “Differential roles of DCRD and PCRD unveiled by low [apoCaM]”).
8) Figure 3. There is no experimental evidence that PCRD or DCRD were actually expressed separately.
For the peptide experiments in Figure 3, both PCRD and DCRD were tagged with fluorescent proteins. And in the case of PCRD/DCRD coexpression, PCRD and DCRD were fused with YFP vs. CFP to make sure both were present in the same cell (Figure 3—figure supplement 1 and subsection “Cooperative scheme of CMI consists of three major combinations”). We have amended the construct names with explicit information of the tags in Figure 3and also other peptide experiments (mainly in Figure 4and Figure 5). In addition, subsequent inducible-CMI experiments (e.g., Figure 5) demonstrate that the ultrastrong CDI as in control was substantially attenuated upon induced PCRD-DCRD dimerization, which provides another proof that both peptides can be separately expressed very well in the same cell without affecting CDI of α1DS.
We assume this comment is also concerned with the potential binding/assembly between the two individual peptides of PCRD and DCRD, which should be unlikely according to this work and previous reports. For PCRD and DCRD peptides, no binding was detected with the FRET two-hybrid assay (Liu et al. 2010). Nevertheless, this would not completely exclude all potential interactions between PCRD and DCRD. For instance, electrostatic interactions might exist between PCRD and DCRD in the holo-channel of CaV1.2 (Hulme et al. 2005), provided that the major binding has been established in the first place, e.g., between IQV-PCRD and DCRD. The important consensus here is that even if the putative PCRD/DCRD interactions exist, they should be weaker than the binding between IQV and DCRD (Figure 4D); and the latter binding is further enhanced by the presence of PCRD in the close vicinity to IQV or DCRD, which might be partly contributed by PCRD/DCRD interaction. We have revised the text to incorporate some of the above facts and analyses (subsection “Cooperative scheme of CMI consists of three major combinations”).
9) There is a lack of clarity about the chimeric constructs used. The authors explain that they will "employ homologous DCTF from α1F with the strongest CMI among CaV1 family to construct chimeric CaV1.3 channels by fusing PCRDF-DCRDF onto IQ motif of α1DS". However, some of the experiments are conducted with PCRDD (Figures 1C, 2B, 4C, 4E, 6B) without giving an explanation or rationale.
Please see item 5.
10) The quantification of the FRET experiments is missing: what is the equation fitting FRET strength (as in Figure 4E, etc.)? Please report fitting parameters.
The fitting procedures were based on established protocols and algorithms (Erickson et al. 2001; Liuet al. 2010; Butz et al. 2016). In the “Materials and methods” section (“FRET optical imaging”) we have added a brief description of the FRET two-hybrid binding assay (3-cube FRET) used in this work. We appreciate the reviewer for this suggestion reminding us to provide the values for the key binding parameters (effective dissociation equilibrium constant Kd indicating affinity and also the maximum FRET ratio FRmax intrinsic to the bound pair, which we obtained from the iterative curve fitting procedures with a customized Matlab program. Accordingly, for all the pairs of FRET binding, Kd and FRmax values have been included for Figure 4D in the caption and the main text (subsection “Cooperative scheme of CMI consists of three major combinations”), and in Figure 4—figure supplement 2, and Figure 6—figure supplement 3.
11) The rapamycin experiments are elegant and insightful. However, it is not clear why 300ms pulses were chosen, as it does not appear steady state for CMI (Figure 6A red current traces). This is also evident in Figure 1B, where the current does not seem to level off at 300ms. The end stage of CMI and CDI may not be the same with longer pulses (true steady-state). This point seems crucial to the authors' conclusion that CDI and CMI functionally result into indistinguishable gating at the ending stage. Experiments with longer depolarizing pulses or measurements of CDI and CMI time constants may help to discuss this point.
We thank the reviewer for bringing up this important matter. We have conducted the experiments with longer depolarization pulses as suggested for both constitutive and rapamycin-inducible CMI (Figure 6—figure supplement 2). Briefly, for CaV1.3 channels, 300 ms is already very close to the end stage or the steady state of CDI; and with both 300 ms and 1000 ms pulses, the “end-stage” CMI (or full- strength CMI) is very close to the end-stage CDI (or steady-state CDI).
The decreasing trend or tendency observed from the Ca2+ current at 300 ms is largely due to VDI (voltage-dependent inactivation), as confirmed by the comparison between current traces in Ba2+ and Ca2+. VDI varies slightly from cell to cell, but generally much slower than CDI so often observable at 300 ms or even later phase during the depolarization step. Therefore, “unsteady” ICa is not necessarily a sign of incomplete CDI.
Meanwhile, we agree with the reviewer that longer pulse such as 1000 ms would be informative and potentially more confirmative. Consistent with J300-based results, J1000 (Ca2+ current density measured at 1000 ms) profiles from α1DS control vs. α1DS-PCRD-DCRD were nearly identical, both suggesting that end-stage CDI remained unaltered even with potent CMI attenuation (Figure 6— figure supplement 2A). For rapamycin-inducible CMI, both I1000 and I300 remained stable throughout the experiments (Figure 6—figure supplement 2B), as compared to decaying Ipeak and I100 (measured at 0 and 100 ms respectively) (Figure 6—figure supplement 2C). Taken together, the performance of I300 is comparable to I1000, so both are qualified indices of end-stage CDI (for channels with or without CMI) (Figure 6—figure supplement 2C).
We would like to clarify that “end-stage” of inducible CMI refers to its steady state in the time-dependent profile, and its actual level of attenuation is dependent on the potency of particular CMI competition. For constitutive CMI, “end-stage” CMI refers to the full-strength CMI or complete competition, which was closely approached but not yet fully reached in most experiments (but see very strong CMI effects in Figure 1B and Figure 4B). In contrast, end-stage CDI in this work always refers to the time domain, reaching its steady state at ~300 ms or later. As suggested by the whole-cell ICa from ensemble channels, we have concluded that the particular channel inhibited by CMI (with CDI abolished) should be functionally equivalent to the channel (e.g., with no CMI) into its end-stage CDI. Provided that VDI is negligible, ICa attenuated by full-strength CMI (evaluated at 0 ms, Ipeak) would be equal to ICa at the end stage of CDI (≥300 ms, e.g., I300), as our data of rapamycin-inducible CMI (version 2) strongly suggest (Figure 6 and Figure 6—figure supplement 2).
Some of these discussions have been included in the revision (Figure 6C, subsection “CMI is a unique type of inhibition but sharing similarities with CDI”).
https://doi.org/10.7554/eLife.21989.022