Interactions of Rabconnectin-3 with Cav2 calcium channels

This study describes the interaction between Cav2 calcium channels and Rabconnectin-3, a di-subunit protein that is associated with synaptic vesicles. Immunostaining reveals that both Rabconnectin-3α (RB-3α) and Rabconnectin-3β (RB-3β) are colocalized in mouse hippocampal neurons. Co-immunoprecipitations from brain tissue is consistent with the formation of a protein complex between RB-3α and RB-3β and both Cav2.2 and the related Cav2.1 calcium channel. The coexpression of either RB-3α or RB-3β with Cav2.2 calcium channels in tsA-201 cells led to a reduction in Cav2.2 current density without any effects on the voltage-dependence of activation or inactivation. Coexpression of both Rabconnectin-3 subunits did not cause an additive effect on current densities. Finally, the presence of Rabconnectin-3 did not interfere with μ-opioid receptor mediated Gβγ modulation of Cav2.2 channels. Altogether, our findings show that Rabconnectin-3 has the propensity to regulate calcium entry mediated by Cav2.2 channels.


Introduction
Cav2 channels are essential mediators of synaptic communication in the mammalian and invertebrate nervous systems [1][2][3][4][5]. Calcium entry via these channels in response to membrane depolarization leads to the release of neurotransmitters, and this process is facilitated through physical interactions of the channels with the vesicle release machinery [6][7][8][9][10]. Indeed, numerous synaptic proteins such as SNAP25, syntaxin 1 and synaptotagmin have been shown to bind to presynaptic calcium channels in a dynamic and often calcium-dependent manner as part of the vesicle release process [11][12][13][14]. Conversely, calcium channel activity can be substantially altered by these interactions which have been attributed to a synaptic protein interaction (a.k.a. synprint) site that is localized within the intracellular region connecting domains II and III of the pore forming Cavα1 subunit [15][16][17][18].
Rabconnectin-3 is a poorly characterized synaptic protein that is formed by α and β subunits, encoded respectively by DMXL2 and WDR7 [19][20][21]. Rabconnectin-3α and-3β (RB-3α and RB-3β) belong to WD40 family of proteins that contain repeats consisting of two internal dipeptide sequences, glycine-histidine and tryptophanaspartic acid [22,23]. RB-3α contains 12 WD40 domains and it is mainly recovered in the synaptic soluble fraction [19]. On the other hand RB-3β has 7 WD40 domains (hence the name WDR7) [20]. Very little is known about their key functional or catalytic domains, or key sites for protein-protein interactions, but it is known that WD40 domains lead to the formation of a β-propeller structure that serves as an interaction platform for a variety of proteins involved in diverse cellular processes [22,23]. RB-3α and RB-3β are abundantly expressed in the brain where they are associated with synaptic vesicles [20]. They have been shown to interact with Rab3A GDP/GTP exchange proteins, and with members of the Rab3 small G protein family that are involved in the control of calcium dependent exocytosis [19,20]. Although the precise role of Rabconnectin-3 in modulating exocytosis has remained enigmatic, the observation that DMXL2 haplo-insufficiency or WDR7 gene deletion cause developmental delays and mental retardation suggests a critical neurophysiological function of Rabconnectin-3 [24,25].
Here, we present evidence that RB-3α and RB-3β form a molecular complex with Cav2.1 and Cav2.2 channels in mouse brain. Furthermore, we show that both subunits affect Cav2.2 current density, suggesting that Rabconnectin-3 may contribute to regulating brain function via alteration of presynaptic calcium influx.

Results and discussion
We first examined the co-localization of Rabcconnectin-3 subunits with Cav2.2 calcium channels expressed endogenously in cultured mouse hippocampal neurons (10 days in vitro). Fig. 1a shows confocal images obtained from hippocampal neurons labelled with anti-RB-3β (green) and anti-Cav2.2α1 (red), revealing that both proteins are diffusely distributed in soma, axons and dendrites from neighboring cells with numerous overlapping puncta. RB-3α (red) and Cav2.2 (green) were also found to be expressed in the same neurons (Fig. 1b). Next, we performed co-immunoprecipitations between Cav2.2 channels, RB-3β and RB-3α using mouse brain homogenate. As shown in Fig. 1c, immunoprecipitation with Cav2.1 and Cav2.2 antibodies and Western blot analysis with an RB-3β antibody resulted in a band near the expected size of 163 kDa. Immunoprecipitation with an RB-3β antibody allowed us to detect RB-3α (Fig. 1d, 340 kDa band) and the reverse experiment also confirmed that the two Rabconnectin-3 subunits are in a complex (Fig. 1e).
Consequently, using an RB-3α antibody, Cav2.2 channels could be immunoprecipitated (Fig. 1f). These data provide evidence that Cav2.2 channels and Rabconnectin-3 form a macromolecular protein complex, and are consistent with the immunostaining experiment.
To ascertain whether Rabconnectin-3 can alter Cav2.2 channel function, we recorded whole cell barium currents through recombinant Cav2.2 (+Cavβ 1b + Cavα 2 δ-1) channels expressed in tsA-201 cells in the absence or presence of Rabconnectin-3. Representative whole-cell Cav2.2 current recordings are shown in Fig. 2a. Figure 2b shows the average current density-voltage relationships (peak current amplitude normalized by Cm) in response to membrane depolarizations from a holding potential (Vh) of − 80 mV. The expression of either RB-3α or RB-3β significantly reduced the current density across a wide range of voltages and this is also reflected in a decrease in Gmax (Fig. 2c) We did not observe changes in voltage for half-maximal activation (Fig. 2b inset) and the slope factor (not shown). Neither Rabconectin-3 subunit had any significant effect on half-inactivation potential (Fig. 2d). Rabconnectin-3 is a cytoplasmic protein that, as we show here, appears to interact with Cav2.2 channels. When we transfected just the pore forming Cav2.2 α 1 subunit with RB-3β and performed a co-immunoprecipitation we found an interaction between these proteins (Fig. 2e). This means that the ancillary Cavβ subunit is not necessary for forming the molecular complex. Interestingly, biotinylation experiments from cells expressing RB-3β and Cav2.2 channels did not reveal any difference in cell surface expression (Fig. 2f ), suggesting that the decrease in current density is not due a reduction in the number of channels at the plasma membrane. Coexpression of both Rabconnectin-3 subunits produced effects that were similar to those observed in the presence of RB-3α ( Fig. 2g and h), indicating that there are no synergistic effects (current density at 10 mV in pA/pF -Cav2.2/β 1 /α 2 δ-1: control − 58.06 ± 5.6; +RB-3α: 22.02 ± 5.0; +RB-3β: -42.35 ± 6.1; Cav2.2/β 1 /α 2 δ-1: control − 56.27 ± 9.0; +RB-3(α + β): -30.95 ± 8.1; P > 0.05 compared to RB-3α), and that the formation of a complete Rabconnectin-3 complex is still capable of modulating Cav2.2 current density.
Given the bulky nature of Rabconnectin-3, we wondered if its presence might affect regulation of the channels by G protein βγ subunits which are known to target the N-terminus and domain I-II linker regions of Cav2.2 [26][27][28][29][30]. This was tested by coexpressing μ-opioid receptors with the channel along with RB-3α plus RB-3β, and then activating the receptors via DAMGO application (Fig. 3a). We found no difference in receptor mediated Cav2.2 channel inhibition (Fig. 3b), nor in the relief of this inhibition via strong voltage pulses (Fig. 3c). Hence, G proteins retain access to their binding site on the Cav2.2 channels, suggesting that the Rabconnectin-3 interaction site is distinct from that targeted by Gβγ.
At this point we do not know where on the channel the Rabconnectin-3 interaction site is located, and whether the interaction is direct, or mediated via an adaptor protein. It is interesting to note that both Rabconnectin-3 subunits independently produced a reduction in current density, which may suggest that there are two separate interaction sites for these two subunits, or that they both target the same region, and in such a manner that the functional effects are preserved when both subunits are present. It is interesting to note that RB-3β was less effective in reducing current densities compared to RB-3α, but perhaps not unexpected given that these two proteins show considerable differences in their numbers of WD40 repeats. We also do not know whether the reduction in current density is due a reduction in single channel amplitude or in maximum open probability. It will also be interesting to determine whether Rabconnectin-3 can interfere with the modulation of channel activity by synaptic proteins targeting the synprint site [7,8,31,32].
We reiterate that the importance of a fully functional Rabconnectin-3 protein is important for brain health. Tata and colleagues [25] described a case of three brothers with an in-frame deletion in the RB-3α encoding gene (DMXL2). This resulted in reduced RB-3α mRNA levels and manifested itself as a series of neurodevelopmental disorders such as mental retardation, dysregulation of the hypothalamus-pituitary hormonal axis and metabolic disorders such as diabetes and hypoglycemia. On the other hand, it has been shown that low DMXL2 expression in mice causes abnormal dendritic development. DMXL2+/− mice show differences in corpus callosum neuroanatomy compared to wild type animals [25]. Along these lines, Chen and co-workers [33] linked an RB-3α missense variant to nonsyndromic hearing loss. Altogether, this strongly suggests that RB-3α is a key regulator in neuronal and homeostatic processes. RB-3β is deleted along with the TXNL1 gene in a condition termed 18q syndrome, a condition characterized by mental retardation that varies in severity from patient to patient, along with a wide array of anatomical and physiological abnormalities, including dystonia [24,34]. It is unlikely that all these phenotypic abnormalities are mediated through alterations of Cav2.2 channel activity. That being said, our experiments in tsA-201 cells would be consistent with a gain of Cav2.2 channel function. In this context, it is interesting to note that a gain of Cav2.2 channel function through a missense mutation has been associated with dystonia [35].
In summary, we present the first description of a physical and functional interaction between Rabconnectin-3 and Cav2.2 calcium channels. This interaction adds to a list of many synaptic proteins that associate with this important presynaptic calcium channel.

Materials and methods
cDNAs Wild type rat calcium channel subunit cDNAs were donated by Dr. Terry Snutch (University of British Columbia, Vancouver, BC) and subcloned into pcDNA3.1 vectors. Human WDR7 cDNA was purchased from Thermo Scientific and human DMXL2 cDNA was obtained from Origene™.

cDNA transfection
After splitting and seeding to 50% confluence, tsA-201 cells were transfected with 3 μg of each plasmid encoding Cav2.2α1, Cavβ1 and Cavα2δ-1, respectively, in the presence or absence of WDR7, DMXL2 or both cDNAs. In addition, 0.5 μg of cDNA encoding green fluorescent protein was added to the transfection mixture to identify and select transfected cells. For G protein studies 3 μg of the μ-opioid receptor cDNA were added. Cells used for electrophysiology experiments were moved to 30°C after transfection, whereas those used for Western blotting were maintained at 37°C.

Patch clamp recordings
Electrophysiological recordings were performed using whole cell patch-clamp at room temperature (22-24°C). Currents were recorded using an Axopatch 200B amplifier linked to a computer with pCLAMP9.2 software. The external recording solution contained (in mM): 5 BaCl 2 , 10 TEA-Cl, 1 MgCl, 128 NaCl, 5 KCl, 10 HEPES and 10 glucose (pH 7.4). Patch pipettes were filled with a solution containing (in mM) 110 CsCl, 2.5 MgCl, 10 EGTA, 10 HEPES, 3 ATP, and 0.5 GTP (pH 7.4). Current densities were obtained by dividing peak current by the whole cell capacitance. Current density-voltage relationships were generated from the peak current obtained during 250 ms pulses between − 50 to + 40 in 5 mV increments from a holding potential of − 80 mV. Steady-state inactivation was measured by applying 2 s conditioning pulses from − 100 to + 20 mV in 10 mV increments followed by a 100ms test pulse to + 10 mV. I-V relationships were fitted with a Boltzmann equation of the form: I = Gmax*(Vm − Vr)/(1 + exp.(−(Vm − V1/2act)/k)), where I is the peak current, Vm is the membrane voltage, V1/2act is the half activation potential, Vr is the reversal potential, and k is the slope factor. Steady-state inactivation curves were fitted with the equation: I/Imax = 1/(1 + exp.((Vp − V1/ 2inac)/k)), where I/I max is the normalized peak current, Vp is the conditioning pre-pulse, V1/2inac is the halfinactivation potential and k is the slope factor. G protein modulation induced by μ-opioid receptor activation was assessed as described by us previously [36]. In brief, the extent of total G protein inhibition was assessed by a multi pulse protocol in which cells are held at − 80 mV. Currents are evoked by a 25 ms test depolarization to + 10 mV (P1), followed by a 500 ms repolarization to − 80 mV. Subsequently, a strong 50 ms depolarizing pulse to + 100 mV is applied (which causes a reversal of Gβγ mediated voltage-dependent inhibition, see [37]). This is followed by 5 ms repolarization and then a second test pulse to + 10 mV (P2). Total G protein inhibition is determined by monitoring the DAMGO-induced reduction in peak current amplitude during P1. Voltage-dependent G protein modulation is determined by the extent of prepulse relief of inhibition as assessed by calculating the ratio of current amplitudes observed during P2 and P1 in the presence of receptor agonist.

Cell surface biotinylation
Transfected cells were washed with ice-cold HEPESbased saline solution (HBSS) and incubated on ice for 15 min to stop trafficking of proteins. Surface proteins were biotinylated for 1 h on ice with 1 mg/ml of EZ-Link Sulfo-NHS-SS-Biotin (Thermo Scientific). The reaction was quenched with 100 mM glycine for 15 min, and cells were lysed in modified RIPA buffer (in mM: 50 Tris, 150 NaCl, 5 EDTA, 1% Triton X-100, 1% NP-40, 0.2% SDS, pH 7.4) for 45 min. Protein quantification was performed using a Bio-Rad protein assay dye, and 2 mg of lysates was incubated with 100 μl of Neutravidin beads (Thermo Scientific) for 1.5 h at 4°C. Beads were washed three times with lysis buffer and proteins eluted with 2× Laemmli sample buffer. Biotinylated proteins and lysates were resolved by SDS-PAGE and analyzed by western blot using anti-Cav2.2 (1:500, ACC-002 Alomone) and 1/5000 anti-Na/K-ATPase (1:5000, AB 7671 Abcam) antibodies. For each experiment, background was subtracted and integrated density of bands was measured and normalized to the Na/K-ATPase signal as loading control.