A potent voltage-gated calcium channel inhibitor engineered from a nanobody targeted to auxiliary CaVβ subunits

Inhibiting high-voltage-activated calcium channels (HVACCs; CaV1/CaV2) is therapeutic for myriad cardiovascular and neurological diseases. For particular applications, genetically-encoded HVACC blockers may enable channel inhibition with greater tissue-specificity and versatility than is achievable with small molecules. Here, we engineered a genetically-encoded HVACC inhibitor by first isolating an immunized llama nanobody (nb.F3) that binds auxiliary HVACC CaVβ subunits. Nb.F3 by itself is functionally inert, providing a convenient vehicle to target active moieties to CaVβ-associated channels. Nb.F3 fused to the catalytic HECT domain of Nedd4L (CaV-aβlator), an E3 ubiquitin ligase, ablated currents from diverse HVACCs reconstituted in HEK293 cells, and from endogenous CaV1/CaV2 channels in mammalian cardiomyocytes, dorsal root ganglion neurons, and pancreatic β cells. In cardiomyocytes, CaV-aβlator redistributed CaV1.2 channels from dyads to Rab-7-positive late endosomes. This work introduces CaV-aβlator as a potent genetically-encoded HVACC inhibitor, and describes a general approach that can be broadly adapted to generate versatile modulators for macro-molecular membrane protein complexes.


Introduction
Inhibition of high-voltage-activated calcium channels (HVACCs) is an important prevailing or potential therapy for diverse cardiovascular (hypertension, cardiac arrhythmias, cerebral vasospasm) and neurological diseases (epilepsy, chronic pain, Parkinson's disease) (Zamponi et al., 2015). Small molecule HVACC inhibitors include Ca V 1 blockers (dihydropyridines, benzothiazepenes phenylalkylamines) and venom peptides that target Ca V 2.1 (!-agatoxin), Ca V 2.2 (!-conotoxin), and Ca V 2.3 (SNX-482) channels. When introduced into an organism, small-molecule HVACC blockers are typically widely distributed leading to off-target effects that can narrow the therapeutic window and, thereby, adversely impact therapy. Genetically-encoded HVACC inhibitors can circumvent off-target effects because they can be selectively expressed in target tissues or cells; thus, they may be useful alternatives or complements to small molecule therapy Murata et al., 2004).
There are seven distinct HVACCs (Ca V 1.1 -Ca V 1.4; Ca V 2.1 -Ca V 2.3) which exist in cells as multisubunit complexes comprising pore-forming a 1 -subunits assembled with auxiliary proteins which include b, a 2 -d, and g subunits (Zamponi et al., 2015;Buraei and Yang, 2010;Dolphin, 2012). HVACCs are named according to the identity of the component a 1 subunit (a 1A -a 1F ; a 1S ) which also contains the voltage sensor, selectivity filter, and channel pore. The various auxiliary subunits typically regulate HVACC trafficking, gating, and modulation, and are recognized as potential targets for developing HVACC-directed therapeutics. For example, gabapentin, which is clinically utilized for treating epilepsy and neuropathic pain, targets HVACC a 2 -d subunits (Gee et al., 1996). Based on the presumption that the association of a 1 with b is obligatory for the formation of surface-targeted functional HVACCs as indicated by heterologous expression experiments (Buraei and Yang, 2010), disruption of the a 1 -b interaction has been long pursued as a strategy to develop HVACC inhibitors (Young et al., 1998;Findeisen et al., 2017;Chen, 2018;Khanna et al., 2019). To this end, over-expression of peptides derived from the a 1 -interaction domain (AID) which contains the amino acid sequence responsible for high-affinity a 1 -b association (Pragnell et al., 1994;Van Petegem et al., 2004;Chen et al., 2004;Opatowsky et al., 2003), has been utilized by several groups as putative genetically-encoded HVACC inhibitors (Findeisen et al., 2017;Yang et al., 2019). However, the efficacy of this approach in vivo may be limited as recent data suggests that in some adult tissue the a 1 -b interaction is not absolutely essential for surface trafficking of HVACCs Meissner et al., 2011).
Rad/Rem/Rem2/Gem/Kir (RGK) proteins are endogenous small Ras-like G-proteins that profoundly inhibit all HVACCs when over-expressed in either heterologous cells or native tissue (Béguin et al., 2001;Finlin et al., 2003;Chen et al., 2005;Xu et al., 2010). They form ternary complexes with HVACCs via binding to constituent b subunits and inhibit currents via multiple mechanisms including removal of surface channels and impairing gating Yang et al., 2010). Despite their efficacy, utility of RGKs as genetically-encoded HVACC inhibitors is confounded by potential off-target effects since they interact with and regulate other binding partners such as cytoskeletal proteins, 14-3-3, calmodulin, and CaM kinase II Correll et al., 2008;Royer et al., 2018;Béguin et al., 2005;Ward et al., 2004). A critical unmet need is the development of genetically-encoded HVACC inhibitors that possess the high efficacy of RGKs but lack the problematic interactions with other signaling proteins. Here, we achieve this by fusing the homologous to the E6-AP carboxyl terminus (HECT) catalytic domain of the E3 ubiquitin ligase, neural precursor cell developmentally down-regulated protein 4 (Nedd4-2 or hereafter referred to as Nedd4L), to a Ca V b-targeted nanobody. The resulting construct, termed Ca V -ablator, eliminated diverse HVACCs both in both reconstituted systems and native excitable cells, providing a unique new tool for probing Ca V 1/Ca V 2 signaling and regulation in vivo, and potential development into a therapeutic.

Isolation and characterization of Ca V b-targeted nanobodies
We sought to develop a nanobody targeted to Ca V bs that would be incorporated into Ca V channel complexes but be functionally silent, to serve as a vehicle to potentially address distinct enzymatic moieties or sensors to endogenous channels. We expressed Ca V b 1b and Ca V b 3 in HEK293 cells using BacMam expression and purified the proteins using affinity purification, ion exchange, and size exclusion chromatography (Figure 1a). Purified b 1 and b 3 (1 mg each) were used for llama immunization, and successful serum conversion was confirmed by ELISA (not shown). Messenger RNA was extracted from isolated lymphocytes, PCR-amplified and cloned into a plasmid vector (pComb3XSS) to generate a V HHS phage library (Figure 1b). Putative nanobody binders were enriched from the phage library using three rounds of phage display and panning (Pardon et al., 2014). We performed a 96-well ELISA on enriched phage libraries and selected 14 positive clones for sequencing ( Figure 1c). We identified at least seven distinct classes of nanobody binders based on the unique sequences within complementarity determining regions (CDR1-3), the major determinants of antigen binding (Figure 1d,e).
We adopted a small-molecule-induced fluorescence co-translocation assay to simultaneously determine whether: (1) individual nanobodies were well-behaved when expressed in mammalian cells (i.e. do not aggregate), and (2) bound Ca V bs. A tripartite construct consisting of individual nanobodies fused to CFP and the C1 domain of PKCg was cloned into a CMV expression vector and transiently co-transfected with YFP-tagged Ca V bs into HEK293 cells. After pilot experiments, we chose one nanobody clone, nb.F3, for in-depth characterization and development. Both nb.F3-CFP-C1 and YFP-b 1 were uniformly expressed in the cytosol of transfected HEK293 cells (Figure 1f). Application of 1 mM phorbol-12,13-dibutyrate (PdBu) led to the rapid and dramatic redistribution of nb.F3-CFP-C1 from the cytosol to the plasma and nuclear membranes ( Figure 1f). Reassuringly, YFP-b 1 concomitantly redistributed to the plasma and nuclear membranes, providing a convenient visual confirmation that it associates with nb.F3 inside cells (Figure 1f). Similar experiments conducted with the other Ca V bs (b 2 -b 4 ) showed that they all bind with nb.F3-CFP-C1 in cells (Figure 1f; Figure 1-figure supplement 1), indicating the nanobody interacts with an epitope conserved among Ca V bs. Isothermal titration calorimetry using purified nb.F3 and Ca V b 2b indicated a high-affinity (K d = 13.2 ± 7.2 nM) interaction and a 1:1 stoichiometry (Figure 1g).
It was important to our overall strategy that nb.F3 incorporate into assembled HVACC complexes without impacting channel function or subunit stability. We utilized a flow cytometry assay to assess KPEDTAVYYCAA ASSYWS R-------SVDEYDY YWGQGTQVTVSS ******** *: : :*********** * (c) Phage ELISA using Ca V b 1 as bait and periplasmic extracts from single infected E. coli clones. Red bars represent clones that were selected for subsequent analyses; blue bar represents a negative control from an E. coli expressing an anti-GFP nanobody. (d) Cartoon showing conventional IgG antibody (left) and camelid heavy-chain antibody (center). Right, a schematic representation of the variable heavy chain (VHH or nanobody) of camelid heavy-chain antibodies. The three CDR loops which are the primary determinants of antigen-binding are shown in red, green, and blue. (e) Sequence alignment of CDR3 from selected clones. (f) Left, schematic of cotranslocation assay to determine nanobody/Ca V b interaction in HEK293 cells. Right, confocal images showing membrane co-translocation of Ca V b X -YFP and nb.F3-CFP-C1 PKCg in response to treatment with 1 uM phorbol 12,13-dibutyrate (PdBu). (g) Left, exemplar isothermal titration calorimetry trace using purified Ca V b 2b and nb.F3. Right, summary of ITC thermodynamic parameters. N, stoichiometry; K d , dissociation constant; K a , affinity constant; DH, enthalpic change; DS entropic change. T = 298K. DOI: https://doi.org/10.7554/eLife.49253.002 The following figure supplement is available for figure 1: the impact of nb.F3 on recombinant Ca V 2.2 trafficking, subunit expression levels, and whole-cell currents, all of which are known to be regulated by Ca V b ( Figure 2) (Waithe et al., 2011). We used an engineered a 1B harboring two tandem high-affinity bungarotoxin-binding sites (2XBBS) in the extracellular domain IV S5-S6 loop and a C-terminus YFP tag to enable simultaneous detection of surface (Alexa647-conjugated bungarotoxin) and total (YFP fluorescence) channel populations in non-permeabilized cells (Figure 2a). We co-expressed BBS-a 1B -YFP and Ca V b either with or without nb.F3- P2A-CFP and utilized flow cytometry to rapidly measure surface and total channel expression. In cells expressing BBS-a 1B -YFP and b 1b , nb.F3 had no impact on Alexa647 or YFP fluorescence compared to control (Figure 2b-d), indicating no disruption of channel trafficking or effect on a 1B expression. Similar results regarding the inertness of nb.F3 on a 1B trafficking and stability were obtained when Ca V 2.2 was reconstituted with the other Ca V b (b 2 -b 4 ) subunits ( Figure 2d).
To examine a potential direct impact of nb.F3 on Ca V b itself, we applied the flow cytometry assay to cells expressing BBS-a 1B + b-YFP ± nb .F3-P2A-CFP ( Figure 2e). Not surprisingly, nb.F3 did not impact the surface trafficking of BBS-a 1B co-expressed with any of the four Ca V b isoforms (Figure 2f-h, Figure 2-figure supplement 1). The expression levels of b 1 -YFP and b 4 -YFP were unaffected by nb.F3, whereas the levels of b 2 and b 3 were modestly reduced (although this effect did not reach statistical significance), suggesting a possible slightly increased vulnerability of these two isoforms to degradation when bound by the nanobody (Figure 2h). Similar observations regarding the lack of effect of nb.F3 on channel trafficking and subunit expression levels were made in cells expressing Ca V 1.2 channels reconstituted from BBS-a 1C + b-YFP ± nb . Finally, we used patch-clamp electrophysiology to evaluate the impact of nb.F3 on whole-cell currents through recombinant Ca V 2.2 channels reconstituted in HEK293 cells. Cells expressing a 1B + b 1b + a 2 d displayed robust whole-cell Ba 2+ currents that were completely unaffected by nb.F3 ( Figure 2i; I peak,0mV = À104.4 ± 22.0 pA/pF, n = 10 for CFP, and I peak,0mV = À103.5 ± 39.5 pA/pF, n = 10 for nb.F3). A similar lack of effect of nb.F3 was observed on currents from either Ca V 2. Overall, these results indicate that nb.F3 binds b 1 -b 4 subunits in cells, and is potentially assembled into Ca V channel complexes in a functionally silent manner, essentially acting as an unobtrusive passenger. However, it was also possible that the apparent functional inertness of nb.F3 on Ca V 2.2 and Ca V 1.2 channels had a more trivial explanation-that Ca V bs assembled with pore-forming a 1 -subunits are simply inaccessible to nb.F3. We could discriminate between these two possible scenarios by determining whether nb.F3 could be used to target bioactive molecules to regulate assembled channels, as we did next.
Potent functional effects of an F3-Nedd4L chimeric protein on Ca V 1/ Ca V 2 channels We hypothesized that fusing the catalytic domain of an E3 ubiquitin ligase to nb.F3 would generate a genetically-encoded molecule that inhibits Cav1/Cav2 channels by reducing their surface density (Kanner et al., 2017). Accordingly, we generated a chimeric construct (nb.F3-Nedd4L) by fusing the catalytic HECT domain of Nedd4L to the C-terminus of nb.F3. We also generated a catalytically dead mutant of the chimeric construct (nb.F3-Nedd4L[C942S]) to distinguish between ubiquitination-dependent and independent effects. Both constructs were generated in a P2A-CFP expression vector, enabling use of CFP fluorescence to confirm protein expression.
In experiments mimicking those described for nb.F3, we examined the impact of nb.F3-Nedd4L and nb.F3-Nedd4L[C942S] on reconstituted Ca V 2.2 channel trafficking, subunit expression levels, and whole-cell currents ( Figure 3). Given the classical role of E3 ubiquitin ligases in mediating degradation of target proteins, we first assessed if nb.F3-Nedd4L affected total Ca V b expression (Figure 3a,b). In cells expressing BBS-a 1B + b 1b -YFP + a 2 d, neither F3-Nedd4L nor F3-Nedd4L [C942S] had any significant impact on b 1b total expression as reported by the unchanged YFP fluorescence compared to negative control cells (Figure 3a,b). Similar results were obtained when BBSa 1B was reconstituted with YFP-tagged b 2 , b 3 , or b 4 subunits, though there was a trend towards lower fluorescence with b 2a and b 4 ( Figure 3b). By contrast, nb.F3-Nedd4L significantly suppressed surface density of BBS-a 1B irrespective of the identity of the co-expressed YFP-tagged Ca V b ( Figure 3c    (e-f) Same format as in (b-d) for cells expressing BBS-a 1B -YFP + Ca V b + a 2 d-1. (g) Exemplar traces (top) and population I-V curves (bottom) from whole-cell patch clamp measurements in HEK293 cells expressing a 1B + Ca V b 1b + a 2 d-1 and nb.F3 (black, I peak, 0mV = -103.5 ± 39.5 pA/pF, n=10), nb.F3-Nedd4L (red, I peak, 0mV = -3 ± 0.53 pA/pF, n=11), or nb.F3-Nedd4L[C942S] (green, I peak, 0mV = -117 ± 34.8 pA/pF, n=8). (h-j) Same format as (g) for The following figure supplements are available for figure 3: Finally, we examined the functional impact of nb.F3-Nedd4L on reconstituted Ca V 2.2 whole-cell currents. Remarkably, nb.F3-Nedd4L essentially eliminated Ca V 2.2 currents reconstituted from a 1B + a 2 d co-expressed with any of the four Ca V bs (Figure 3g-j). Further, nb.F3-Nedd4L was equally effective in ablating whole-cell currents in reconstituted Ca V 1.2, Ca V 1.3, Ca V 2.1, and Ca V 2.3 channels ( Figure 4).
Given its exceptional efficacy in ablating whole-cell HVACC currents via a functionalized Ca V b-targeted nanobody, we named nb.F3-Nedd4L as Ca V -ablator, and describe the process of HVACC current elimination by this molecule as Ca V -ablation.
Ca V -ablation of endogenous Ca V 1.2 channels in cardiomyocytes We next determined whether Ca V -ablator could effectively inhibit HVACC currents in native cells where the nano-environment around Ca V 1/Ca V 2 channels is typically more complex than in heterologous cells. Cultured adult guinea pig ventricular cardiomyocytes (CAGPVCs) provided an initial  exceptional challenge because they have an intricate cyto-architecture and express Ca V 1.2 channels that are predominantly targeted to specialized dyadic junctions. Moreover, as it has now been shown that in adult cardiomyocytes binding of a 1C to Ca V b is not obligatory for substantive Ca V 1.2 channel trafficking to the surface sarcolemma Meissner et al., 2011), the fraction of Ca V b-bound Ca V 1.2 channels contributing to the whole-cell L-type current (I Ca,L ) in ventricular myocytes is ambiguous. We used adenovirus to express Ca V -ablator or nb.F3-Nedd4L[C942S] in CAGPVCs which retain the rod-shaped phenotype and overall cyto-architecture of freshly isolated heart cells ( Figure 5a). Control (non-infected) cardiomyocytes expressed I Ca,L that peaked at a 0 mV test pulse (Figure 5a,b; I peak,0mV = À6.5 ± 0.2 pA/pF, n = 8). By contrast, in contemporaneous experiments, cardiomyocytes expressing Ca V -ablator via adenovirus-mediated infection displayed virtually no Ca V 1.2 currents, demonstrating an exceptional Ca V -ablation efficiency in this system (Figure 5a,b; I peak,0mV = À1.0 ± 0.3 pA/pF, n = 9). Cardiomyocytes expressing nb.F3-Nedd4L[C942S] displayed I Ca,L similar to control (I peak,0mV = À5.1 ± 0.6 pA/pF, n = 10), indicating that ubiquitination is necessary for Ca V -ablation in this system. What is the mechanism of Ca V -ablation in cardiomyocytes? We used immunofluorescence to probe how Ca V -ablator affected expression levels and sub-cellular localization of Ca V 1.2 a 1C and b 2 subunits, respectively, in cardiomyocytes. Ca V a 1C in uninfected cardiomyocytes presented with a characteristic striated punctate distribution pattern that co-localized with that of ryanodine (RyR2) receptors (Figure 5c), reflecting their well-known predominant localization at dyadic junctions (Scriven et al., 2000;Bers, 2002). A similar distribution pattern for a 1C was observed in cardiomyocytes expressing nb.F3-Nedd4L[C942S], consistent with the lack of effect of this protein on I Ca,L . In cardiomyocytes expressing Ca V -ablator, the signal intensity for punctate a 1C staining was unchanged from control cells (Figure 5-figure supplement 1), suggesting no impact of the presumed increase in ubiquitination on the stability of the protein. However, there was a redistribution of a 1C from dyadic junctions, as reported by a dramatic loss of co-localization between a 1C and RyR2 ( Figure 5c). Rather, the punctate a 1C signals in Ca V -ablator-expressing cardiomyocytes coincided with Rab7, but not Rab5 or LAMP1, immunofluorescence signals (Figure 5d; Figure 5-figure supplement 1). Thus, the mechanism of Ca V -ablator inhibition of I Ca,L is redistribution of a 1C from dyadic junctions to intracellular compartments, specifically Rab7-positive late endosomes (Figure 5h) (Rink et al., 2005).
Cardiomyocytes expressing Ca V -ablator also showed no difference in total Ca V b 2 levels as compared to either uninfected or nb.F3-Nedd4L[C942S]-expressing cells (Figure 5-figure supplement 1). Hence, Ca V -ablator-mediated redistribution of Ca V 1.2 in cardiomyocytes cannot be explained as simply due to an absence of Ca V b. An intriguing possibility was that though Ca V -ablator is specifically targeted to Ca V b in channel complexes, it is also able to directly catalyze ubiquitination of a 1 subunits within the macro-molecular complex. Indeed, in pulldown experiments of recombinant Ca V 1.2 channels, Ca V -ablator substantially increased ubiquitination of both a 1C (Figure 5e,f) and Ca V b 1b subunits ( Figure 5g). Nevertheless, the overall levels of a 1C expression was unchanged with Ca Vablator despite the increased ubiquitination ( Figure 5e). Taken together, our results suggest that direct ubiquitination of a 1C by Ca V -ablator may underlie the redistribution of Ca V 1.2 channels from dyads to Rab7-positive late endosomes (Figure 5h).

Ca V -ablation in dorsal root ganglion (DRG) neurons and pancreatic b cells
We next tested the efficacy of Ca V -ablator to suppress HVACCs in murine dorsal root ganglion (DRG) neurons which were of interest because they express multiple Ca V 1/Ca V 2 channel types (Murali et al., 2015;McCallum et al., 2011), and also play a key role in the processing of noxious signals including pain and itch (Han et al., 2013;Kim et al., 2016). We infected cultured DRG neurons with adenovirus expressing either GFP, Ca V -ablator, or nb.F3-Nedd4L[C942S]. Given their heterogeneous nature, we first used fura-2 to measure calcium influx into a population of DRG neurons in response to depolarization with 40 mM KCl (Figure 6a,b). Recordings were done in the presence of 5 mM mibefradil to block low-voltage-activated T-type calcium channels which are also prevalent in these cells (Puckerin et al., 2018;Jagodic et al., 2008). In neurons expressing GFP or nb.F3-Nedd4L[C942S], a substantial fraction of cells displayed large increases in fura-2-reported Ca 2+ transients in response to 40 mM KCl, indicating the opening of Ca V 1/Ca V 2 channels (Figure 6a,  (c) Left, exemplar confocal images of cardiomyocytes fixed and immunostained with a 1C (green) and ryanodine receptor (RyR2, magenta) antibodies. Figure 5 continued on next page contrast, depolarization-induced Ca 2+ influx was virtually eliminated in neurons expressing Ca V -ablator, demonstrating highly efficient Ca V -ablation in this system (Figure 6a,b).
We used whole-cell patch clamp to further characterize the impact of Ca V -ablator on calcium currents in DRG neurons. It was of particular interest to determine relative effects of Ca V -ablator on HVACCs and LVA T-type channels that are present in a subset of DRG neurons. We recorded families of whole-cell currents evoked by test pulses (from À40 mV to +60 mV in 10 mV increments) from a holding potential of either À90 mV or À50 mV to inactivate any T-type channel present (Figure 6c). Cells expressing GFP (control) or F3-Nedd4L[C942S] displayed large I Ba irrespective of the holding potential (Figure 6c,d; I peak,-10mV = À173.9 ± 28.2 pA/pF, n = 6 for GFP, I peak,-10mv = À206.7 ± 36.4 pA/pF, n = 5 for F3-Nedd4L[C942S]), though those recorded with a À50 mV holding potential had a lower amplitude reflecting inactivation of T-type channels and also a fraction of HVACCs. Cells expressing Ca V -ablator displayed essentially no HVACC currents (Figure 6c,d; I peak,-10mV = À14.3 ± 6.2 pA/pF), most evident as an absence of I Ba recorded from a À50 mV holding potential (Figure 6c, middle). Moreover, in these cells, when currents were recorded from a À90 mV holding potential, they displayed fast inactivation kinetics characteristic of T-type channels ( Figure 6c). Overall, these results indicate Ca V -ablator selectively eliminates HVACCs in DRG neurons without impacting LVA T-type channels.
Finally, we tested whether Ca V -ablator is also effective in murine pancreatic b-cells, which have multiple Ca V channel types (Ca V 1.2, Ca V 1.3, and Ca V 2.1) involved in insulin release (Yang and Berggren, 2006). We used adenovirus to infect digested islets isolated from transgenic mice expressing tdTomato in pancreatic b-cells. Control cells expressing GFP or nb.F3-Nedd4L[C942S] displayed robust glucose-or KCl-evoked fura-2-reported Ca 2+ transients that were essentially abolished in cells expressing Ca V -ablator (Figure 6e-g). Altogether, these results reveal the exceptional activity of Ca V -ablator as a genetically-encoded HVACC inhibitor that is effective across diverse cellular contexts.

Discussion
This work introduces Ca V -ablator as a novel genetically-encoded molecule that potently inhibits HVACCs by targeting auxiliary Ca V b subunits. Ca V -ablator combines the exquisite specificity of a Ca V b-targeted nanobody and the powerfully consequential catalytic activity of an E3 ubiquitin ligase. We discuss four distinct aspects of this work, based on viewing Ca V -ablator from different perspectives; 1) as a unique tool to selectively erase HVACCs in cells, 2) as a method to probe mechanisms of HVACC regulation and trafficking, 3) as a potential therapeutic, and 4) as a prototype engineered protein that enables probing new dimensions of macro-molecular membrane protein signaling.
Ca 2+ is a universal second messenger critical to the biology of virtually all cells. In excitable cells, both LVACCs and HVACCs transduce electrical signals encoded in action potentials into changes in intracellular Ca 2+ that then drive many biological responses. In cells expressing both classes of channels, the physiological effects mediated specifically through LVACCs versus HVACCs in vivo can be difficult to decipher. Ca V -ablator now presents as a tool that can be deployed in target cells to virtually erase all HVACCs while leaving LVACC actions intact. The closest existing proteins that can similarly eliminate HVACCs are RGK GTPases which are capable of potently inhibiting Ca V 1/Ca V 2 channels when over-expressed in target cells (Murata et al., 2004;Chen et al., 2005;Xu et al., 2010;Puckerin et al., 2018;Bannister et al., 2008). However, a distinct disadvantage of RGKs is their propensity for off-target effects due to their known interactions with, and regulation of, cytoskeletal proteins and other signaling molecules including 14-3-3, calmodulin, and CaM kinase II Correll et al., 2008;Royer et al., 2018;Béguin et al., 2005). Over the last two decades, several groups have sought to disrupt the a 1 -Ca V b interaction with either small molecules or by over-expressing the AID peptide as a Ca V b sponge (Findeisen et al., 2017;Chen, 2018;Khanna et al., 2019;Yang et al., 2019). While this approach has shown some efficacy in certain instances, the potency of HVACC inhibition falls well short of that achieved here with Ca Vablator. Indeed, over-expressing the AID peptide in adult cardiac myocytes is not effective in inhibiting Ca V 1.2 channels , because in this context a 1C binding to Ca V b is not absolutely required for channel trafficking to the surface Meissner et al., 2011). Nevertheless, the ability of Ca V -ablator to essentially eradicate I Ca,L in adult cardiomyocytes indicates that under normal physiological conditions essentially all a 1C subunits are associated with a Ca V b in ventricular heart cells. Ca V 1/Ca V 2 channels and other surface membrane proteins spend a significant portion of their life cycles in intracellular compartments reflecting their biogenesis, recycling, and ultimate destruction. The signals regulating HVACC degradation and trafficking among compartments are arcane and poorly understood, but likely prominently involve post-translational modifications of channel subunits. Here, we show that targeted ubiquitination of a 1C /b 2 complexes in cardiomyocytes with Ca Vablator specifically arrests Ca V 1.2 channels in Rab7-positive late endosomes. Ca V -ablator possesses the catalytic HECT domain of Nedd4L which is known to principally catalyze the addition of K63-linkage polyubiquitin chains to target proteins (Kim and Huibregtse, 2009;Scheffner and Kumar, 2014). Thus, our results suggest that K63-ubiquitin chains on a 1C /b 2 subunits may be a key signal directing Ca V 1.2 channels to late endosomes. We further found that targeted ubiquitination of HVACC a 1 subunits with Ca V -ablator did not lead to their enhanced degradation either in heterologous cells or cardiomyocytes. By contrast, using a GFP nanobody to target the Nedd4L HECT domain to YFP-tagged KCNQ1, a known substrate of endogenous Nedd4L, resulted in reduced expression of this K + channel pore-forming a 1 subunit (Kanner et al., 2017). Hence, the impact of Nedd4L HECT domain on the stability of membrane proteins is likely substrate-dependent. We speculate that arming nb.F3 with the catalytic domains of other types of E3 ligases that catalyze formation of different polyubiquitin chains will elucidate the precise signals dictating Ca V 1/Ca V 2 channel degradation and trafficking among distinct compartments. Beyond ubiquitination, the approach could also be potentially used to elucidate functional consequences and mechanisms of other posttranslational modifications such as phosphorylation/dephosphorylation on Ca V 1/Ca V 2 channels, as well as to localize sensors that report on signals within HVACC nano-domains in live cells.
Blocking the activity of specific HVACCs with small molecules is a prevailing or potential therapy for many cardiovascular and neurological diseases including; pain, hypertension, cardiac arrhythmias, epilepsy, and Parkinson's disease (Zamponi, 2016). A limitation of small molecule or toxin blockers for HVACCs is the propensity for off-target effects due to their inevitable widespread distribution when administered to a patient. In some circumstances such off-target effects may limit the therapeutic window sufficiently to adversely affect treatment efficacy. Genetically-encoded HVACC inhibitors have great potential to be useful therapeutics with the advantage that their expression can be restricted to target tissues/cell types, or even to spatially discrete channels within single cells (Murata et al., 2004;Makarewich et al., 2012). Given its potency in silencing HVACC activity, Ca Vablator could be a lead molecule for future development into a gene therapy for particular applications where a genetically-encoded HVACC inhibitor is warranted. For this purpose, it may be desirable to generate Ca V -ablator versions whose time course and extent of action could be tuned by either a small molecule or light. Indeed, this is a focus of ongoing work.
Finally, an exciting prospect is the potential of Ca V -ablator as a prototype that can be further developed to engineer proteins that regulate Ca V 1/Ca V 2 channel complexes with new dimensions of specificity. For example, a prevailing idea is that Ca V 1/Ca V 2 channels of a particular type (e.g. Ca V 1.2 channels in cardiomyocytes) may yet form discrete signaling units with different functional outputs in single cells based on their incorporation into divergent macro-molecular complexes (Shaw and Colecraft, 2013). There are tantalizing hints that different Ca V b isoforms could be a node of signal diversification by promoting formation of molecularly distinct HVACC macro-molecular complexes (McEnery et al., 1998;Brice and Dolphin, 1999;Campiglio and Flucher, 2015). Hence, the ability to inhibit specific Ca V channel macro-molecular complexes based on the identity of the constituent Ca V b is biologically important, yet not rigorously addressable with conventional knockout/knockdown approaches. However, this capability may be readily achieved with Ca Vablators directed towards particular Ca V b isoforms. A challenge to realize this possibility is the development of Ca V b isoform-specific nanobodies which should be feasible given that there is sequence divergence among Ca V bs outside the conserved src homology 3 (SH3) and guanylate kinase (GK) domains (Buraei and Yang, 2010). In a broader context, the phenomenon of ion channel pore-forming a 1 subunits assembled with diverse auxiliary subunits in individual cells is common throughout biology (O'Malley and Isom, 2015;Copits and Swanson, 2012;Trimmer, 2015). Hence, Ca V -ablator-inspired molecules and approaches might be expected to elucidate functional dimensions of ion channel macro-molecular complex signaling that, to date, have remained refractory to analyses.

Protein purification
We used the BacMam expression system to purify Ca V b 1B and Ca V b 3 (Goehring et al., 2014). Briefly, full-length Ca V b 1b and Ca V b 3 were cloned into a modified pEG BacMam vector with a C-terminal FLAG tag using BamHI and EcoRI sites. BacMam virus was subsequently generated in Sf9 cells and harvested after three rounds of amplification. 100 mL of BacMam virus was used to infect 1 L of HEK293 GnTIcells (N-acetylglucosaminyltransferase I-negative) and kept shaking at 37˚C. After 18 hrs the cells were stimulated with 10 mM sodium butyrate and harvested 72 hrs later. Cells were lysed using an Avestin Emulsiflex-C3 homogenizer in buffer containing 50 mM Tris, 150 mM KCl, 10% sucrose, 1 mM PMSF (phenylmethylsulfonyl fluoride), and EDTA-free Complete protease inhibitor cocktail (Roche), pH 7.4. Lysate was spun down at 35,000 g for 1 hr. Ca V b was subsequently isolated from supernatant with anti-FLAG antibody (M2) affinity chromatography, and eluted with 100 mg/mL FLAG peptide (Sigma Millipore) in 50 mM TrisHCl, 150 mM KCl, pH 7.4. The protein was then applied to an ion exchange column (MonoQ, GE) and eluted with a linear KCl gradient of 50 mM to 1M. Peak fractions were collected and subjected to size exclusion chromatography (Superdex 200, GE) in a buffer containing 20 mM Tris, 150 mM KCl, pH 7.4. Proteins were brought to 20% glycerol, flash frozen, and stored at À80˚C. For isothermal titration calorimetry experiments, both Ca V b 2b and nb.F3 were cloned via Gibson assembly (Gibson et al., 2009) into an IPTG (isopropyl b-D-1-thiogalactopyranoside) inducible, kanamycin-resistant pET derived plasmid (Novagen, Madison, Wisconsin), with an N-terminal deca-histidine tag (His10) and transformed into Rosetta DE3 E. coli (Millipore Sigma), following manufacturers' instructions. Cells were grown at 37˚C in 1L 2xTY media supplemented with 50 ug/mL carbenicillin and 35 mg/mL chloramphenicol and shook at 225 rpm. Protein expression was induced with 0.2 mM IPTG when the cells reached an OD of 0.6-0.8. The cells were then grown overnight at 22˚C.
Nb.F3 was purified as previously described (McMahon et al., 2018): briefly, cells were harvested and resuspended in 100 mL buffer containing (mM) 500 sucrose, 200 Tris (pH 8), 0.5 EDTA and osmotically shocked with the addition of 200 mL water with stirring. The lysate was brought to a concentration of (mM) 150 NaCl, 2 MgCl 2 , and 20 imidazole and centrifuged at 20,000 g, 4˚C for 30 min. The supernatant was combined with 2 mL Ni-NTA Sepharose resin (Qiagen) in batch, washed with 70 mM imidazole, and eluted with 350 mM imidazole. The eluant was dialyzed into a buffer containing 150 mM NaCl, 10 mM HEPES, pH 7.4 and purified with an S200 size exclusion column (GE Healthcare).
For the purification of Ca V b 2b , cells were pelleted and resuspended in a buffer containing (mM) 300 NaCl, 20 Tris HCl, 10% glycerol, pH 7.4, 0.5 PMSF, and EDTA-free Complete protease inhibitor cocktail (Roche). Cells were lysed using an Avestin Emulisflex-C3 homogenizer and spun at 35,000 g for 30'. The solubilized protein was applied to Ni-NTA Sepharose (Qiagen) and purified as nb.F3.

Nanobody generation
One llama was immunized with an initial injection of 600 mg purified Ca V b 1b and Ca V b 3 , with four boosters of 200 ug each protein administered every other week (Capralogics Inc, Hardwick, MA). 87 days after the first immunization, lymphocytes were isolated from blood and a cDNA library with ProtoScript II Reverse Transcriptase (New England Biolabs). Nanobodies were isolated as previously described (Pardon et al., 2014), using a two-step nested PCR. Amplified Vhh genes were cloned into the phagemid plasmid pComb3xSS, a gift from Carlos Barbas (Andris-Widhopf et al., 2000) (Addgene plasmid # 63890). A phage display library was created using electrocompetent TG1 E. coli cells (Lucigen). Three rounds of phage display were performed as previously described (Pardon et al., 2014), using 100 nM biotinylated Ca V b 3 as bait on neutravidin-coated Nunc-Immuno plates (Thermo Scientific). Clones of interest were subsequently cloned into mammalian expression systems for further study (see below).

Molecular biology and plasmid construction
Potential nbs were PCR amplified with primers flanking their conserved framework (FW) FW1 and FW4 regions and inserted into the mammalian expression plasmid pcDNA3 (Invitrogen) using HindIII and EcoRI sites. An additional GSG linker was included in the PCR and the insert was ligated upstream of an enhanced CFP and C1 domain of human PKCg (residues 51-180).
Rat Ca V b 1b , a kind gift from Dr. Jian Yang (Columbia University), was PCR amplified for subsequent overlap PCR with YFP, inserting a GSG linker between the two proteins. The resulting Ca V b 1b -GSG-YFP sequence was digested with BamHI and NotI and ligated into a PiggyBac CMV mammalian expression vector (System Biosciences). A similar cloning strategy was used for Ca V b 3 and Ca V b 4 . Rat Ca V b 2a was PCR amplified with an N-terminal YFP to prevent palmitoylation of the b 2a subunit (Chien et al., 1996) and inserted with a similar strategy.
A customized bicistronic vector (xx-P2A-CFP) was synthesized in the pUC57 vector, in which coding sequence for P2A peptide was sandwiched between an upstream multiple cloning site and enhanced cyan fluorescent protein (CFP) (Genewiz). The xx-P2A-CFP fragment was amplified by PCR and cloned into the PiggyBac CMV mammalian expression vector (System Biosciences) using NheI/ NotI sites. To generate nb.F3 -P2A-CFP, we PCR amplified the coding sequence for nb.F3 and cloned it into xx-P2A-CFP using NheI/AflII sites. A similar backbone was created in the PiggyBac CMV mammalian expression vector in which CFP-P2A-xx contained a multiple cloning site downstream of the P2A site (Genewiz). Nb.F3 was PCR amplified and ligated into the vector with BglII/ AscI sites. The HECT domain of human Nedd4L (Gao et al., 2009) (a gift from Joan Massague, Addgene plasmid # 27000) consisting of residues 594-974 was PCR amplified and inserted downstream of nb.F3 using AscI/AgeI sites. Mutagenesis of C942S was accomplished using site-directed mutagenesis.

Cell culture and transfection
Human embryonic kidney (HEK293) cells were a kind gift from the laboratory of Dr. Robert Kass (Columbia University). Cells were mycoplasma free, as determined by the MycoFluor Mycoplasma Detection Kit (Invitrogen, Carlsbad, CA). Low passage HEK293 cells were cultured at 37˚C in DMEM supplemented with 5% fetal bovine serum (FBS) and 100 mg/mL of penicillin-streptomycin. HEK293 cell transfection was accomplished using the calcium phosphate precipitation method. Briefly, plasmid DNA was mixed with 7.75 mL of 2 M CaCl 2 and sterile deionized water (to a final volume of 62 mL). The mixture was added dropwise, with constant tapping to 62 mL of 2x Hepes buffered saline containing (in mM): Hepes 50, NaCl 280, Na 2 HPO 4 1.5, pH 7.09. The resulting DNA-calcium phosphate mixture was incubated for 20 min at room temperature and then added dropwise to HEK293 cells (60-80% confluent). Cells were washed with Ca 2+ -free phosphate buffered saline after 4-6 hr and maintained in supplemented DMEM.

Adenoviral generation
Adenoviral vectors expressing GFP and CFP-P2A-nb.F3-Nedd4L[C942S] were generated using the pAdEasy system (Stratagene) according to manufacturer's instructions as previously described (Kanner et al., 2017;Subramanyam, 2013). Plasmid shuttle vectors (pShuttle CMV) containing cDNA for CFP-P2A-nb.F3-Nedd4L[C942S] were linearized with PmeI and electroporated into BJ5183-AD-1 electrocompetent cells pre-transformed with the pAdEasy-1 viral plasmid (Stratagene). PacI restriction digestion was used to identify transformants with successful recombination. Positive recombinants were amplified using XL-10-Gold bacteria, and the recombinant adenoviral plasmid DNA linearized with PacI digestion. HEK cells cultured in 60 mm diameter dishes at 70-80% confluency were transfected with PacI-digested linearized adenoviral DNA. Transfected plates were monitored for cytopathic effects (CPEs) and adenoviral plaques. Cells were harvested and subjected to three consecutive freeze-thaw cycles, followed by centrifugation (2,500 Â g) to remove cellular debris. The supernatant (2 mL) was used to infect a 10 cm dish of 90% confluent HEK293 cells. Following observation of CPEs after 2-3 d, cell supernatants were used to re-infect a new plate of HEK293 cells. Viral expansion and purification was carried out as previously described (Colecraft et al., 2002). Briefly, confluent HEK293 cells grown on 15 cm culture dishes (x8) were infected with viral supernatant (1 mL) obtained as described above. After 48 hr, cells from all of the plates were harvested, pelleted by centrifugation, and resuspended in 8 mL of buffer containing (in mM) 20 Tris HCl, 1 CaCl 2 , one and MgCl 2 (pH 8). Cells were lysed by four consecutive freeze-thaw cycles and cellular debris pelleted by centrifugation. The virus-laden supernatant was purified on a cesium chloride (CsCl) discontinuous gradient by layering three densities of CsCl (1.25, 1.33, and 1.45 g/mL). After centrifugation (50,000 rpm; SW41Ti Rotor, Beckman-Coulter Optima L-100K ultracentrifuge; 1 hr, 4˚C), a band of virus at the interface between the 1.33 and 1.45 g/mL layers was removed and dialyzed against PBS (12 hr, 4˚C). Adenoviral vector aliquots were frozen in 10% glycerol at À80˚C until use. Generation of CFP-P2A-nb.F3-Nedd4L was performed by Vector Biolabs (Malvern, PA).

Flow cytometry assay of total and surface calcium channels
Cell surface and total ion channel pools were assayed by flow cytometry in live, transfected HEK293 cells as previously described (Kanner et al., 2017;Aromolaran et al., 2014). Briefly, 48 hr posttransfection, cells cultured in 12-well plates were gently washed with ice cold PBS containing Ca 2+ and Mg 2+ (in mM: 0.9 CaCl 2 , 0.49 MgCl 2 , pH 7.4), and then incubated for 30 min in blocking medium (DMEM with 3% BSA) at 4˚C. HEK293 cells were then incubated with 1 mM Alexa Fluor 647 conjugated a-bungarotoxin (BTX 647 ; Life Technologies) in DMEM/3% BSA on a rocker at 4˚C for 1 hr, followed by washing three times with PBS (containing Ca 2+ and Mg 2+ ). Cells were gently harvested in Ca 2+ -free PBS, and assayed by flow cytometry using a BD Fortessa Cell Analyzer (BD Biosciences, San Jose, CA, USA). CFP-and YFP-tagged proteins were excited at 407 and 488 nm, respectively, and Alexa Fluor 647 was excited at 633 nm.

Immunofluorescence staining
Approximately 48 hr after adenoviral infection, guinea pig cardiomyocytes were fixed in 4% paraformaldehyde (wt/vol, in PBS) for 20 min at RT. Cells were washed twice with PBS and then incubated in 0.1M glycine (in PBS) for 10 min at RT to block free aldehyde groups. Fixed cells were then permeabilized with 0.2% Triton X-100 (in PBS) for 20 min at RT. Non-specific binding was blocked with a 1 hr incubation at RT in PBS solution containing 3% (vol vol À1 ) normal goat serum (NGS), 1% BSA, and 0.1% Triton X-100. Cells were then incubated with primary antibody in PBS containing 1% NGS, 1% BSA, and 0.1% BSA overnight at 4˚C. Cells were washed three times for 10 min each with PBS with 0.1% Triton X-100 and then stained with secondary antibody for 1 hr at RT. Antibody dilutions were prepared in PBS solution containing 1% NGS, 1% BSA, and 0.1% Triton X-100. The cells were then washed in PBS with 0.1% Triton X-100 and imaged in the same solution. Primary antibodies and working dilutions were as follows: a 1C : Alomone, 1:1000; UC Davis/NIH NeuroMab Facility, clone N263/31, 1:200. RyR: Sigma Aldrich, 1:1000. Ca V b 2 : Alomone, 1:200. Rab7: Cell Signaling Technology, 1:100. Rab5: Cell Signaling Technology, 1:200. Lamp1: Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242, 1:100. Secondary antibodies (Thermofisher) were used at a dilution of 1:1000.
For Ca V b 1b pulldowns, lysates were precleared with 10 mL of protein A/G sepharose beads (Rockland) for 1 hr at 4˚C and then incubated with 2 mg anti-Ca V b 1 antibody (UC Davis/NIH NeuroMab Facility, clone N7/18) for 1 hr at 4˚C. Equivalent amounts of protein were then added to spin columns with 25 mL equilibrated protein A/G sepharose beads and rotated overight at 4˚C. Immunoprecipitates were washed a total of five times with RIPA buffer and then eluted with 30 mL elution buffer (50 mM Tris, 10% (vol vol À1 ) glycerol, 2% SDS, 100 mM DTT, and 0.2 mg mL À1 bromophenol blue) at 55˚C for 15 min. For a 1C pulldowns, lysates were added to spin columns containing 10 mL of equilibrated RFP-trap agarose beads, rotated at 4˚C for 1 hr, and then washed/eluted as described above. Proteins were resolved on a 4-12% Bis Tris gradient precast gel (Life Technologies) in MOPS-SDS running buffer (Life Technologies) at 200 V constant for~1 hr. Protein bands were transferred by tank transfer onto a polyvinylidene difluoride (PVDF, EMD Millipore) membrane in transfer buffer (25 mM Tris pH 8.3, 192 mM glycine, 15% (vol/vol) methanol, and 0.1% SDS). The membranes were blocked with a solution of 5% nonfat milk (BioRad) in Tris-buffered saline-tween (TBS-T) (25 mM Tris pH 7.4, 150 mM NaCl, and 0.1% Tween-20) for 1 hr at RT and then incubated overnight at 4˚C with primary antibodies (Ca V b 1 , UC Davis/NIH NeuroMab Facility. Actin, Sigma Aldrich) in blocking solution. The blots were washed with TBS-T three times for 10 min each and then incubated with secondary horseradish peroxidase-conjugated antibody for 1 hr at RT. After washing in TBS-T, the blots were developed with a chemiluminiscent detection kit (Pierce Technologies) and then visualized on a gel imager. Membranes were then stripped with harsh stripping buffer (2% SDS, 62 mM Tris pH 6.8, 0.8% ß-mercaptoethanol) at 50˚C for 30 min, rinsed under running water for 2 min, and washed with TBST (3x, 10 min). Membranes were pre-treated with 0.5% glutaraldehyde and re-blotted with antiubiquitin (VU1, LifeSensors) as per the manufacturers' instructions.

Calcium imaging
DRG neurons were washed twice in basal solution containing (mM): 145 NaCl, 5 KCl, 2 CaCl 2 , 1 MgCl 2 , one sodium citrate, 10 HEPES, 10 D-glucose, pH 7.4, and incubated in the same solution containing 5 uM fura-2 with 0.05% Pluronic F-127 detergent (Life Technologies) for 1 hr at 37˚C, 5% CO 2 . Afterwards, cells were washed twice in same solution and placed on an inverted Nikon Tieclipse microscope with a Nikon Plan fluor 20x objective (0.45 N.A.). Fura-2 measurements were recorded at excitation wavelengths of 340 and 380 nm using EasyRatioPro (HORIBA Scientific). DRG neurons were depolarized with a solution in which NaCl was reduced to 110 mM and KCl increased to 40 mM.

Data and statistical analysis
Data were analyzed off-line using FloJo, PulseFit (HEKA), Microsoft Excel, Origin and GraphPad Prism software. Statistical analyses were performed in Origin or GraphPad Prism using built-in functions. Statistically significant differences between means (p<0.05) were determined using Student's t test for comparisons between two groups or one-way ANOVA for three groups, with Tukey's posthoc analysis. Data are presented as means ± s.e.m.