Dopamine D1A directly interacts with otoferlin synaptic pathway proteins: Ca2+ and phosphorylation underlie an NSF-to-AP2mu1 molecular switch

Abstract

Dopamine receptors regulate exocytosis via protein-protein interactions (PPIs) as well as via adenylyl cyclase transduction pathways. Evidence has been obtained for PPIs in inner ear hair cells coupling D1A to soluble N-ethylmaleimide-sensitive factor (NSF) attachment protein receptor (SNARE)-related proteins snapin, otoferlin, N-ethylmaleimide-sensitive factor (NSF), and adaptor-related protein complex 2, mu 1 (AP2mu1), dependent on [Ca²⁺] and phosphorylation. Specifically, the carboxy terminus of dopamine D1A was found to directly bind t-SNARE-associated protein snapin in teleost and mammalian hair cell models by yeast two-hybrid (Y2H) and pull-down assays, and snapin directly interacts with hair cell calcium-sensor otoferlin. Surface plasmon resonance (SPR) analysis, competitive pull-downs, and co-immunoprecipitation indicated that these interactions were promoted by Ca²⁺ and occur together. D1A was also found to separately interact with NSF, but with an inverse dependence on Ca²⁺ Evidence was obtained, for the first time, that otoferlin domains C2A, C2B, C2D, and C2F interact with NSF and AP2mu1, whereas C2C or C2E do not bind to either protein, representing binding characteristics consistent with respective inclusion or omission in individual C2 domains of the tyrosine motif YXXΦ. In competitive pull-down assays, as predicted by K_D values from SPR (+Ca²⁺), C2F pulled down primarily NSF as opposed to AP2mu1. Phosphorylation of AP2mu1 gave rise to a reversal: an increase in binding by C2F to phosphorylated AP2mu1 was accompanied by a decrease in binding to NSF, consistent with a molecular switch for otoferlin from membrane fusion (NSF) to endocytosis (AP2mu1). An increase in phosphorylated AP2mu1 at the base of the cochlear inner hair cell was the observed response elicited by a dopamine D1A agonist, as predicted.

Keywords: AP2mu1; NSF; dopamine D1A; hair cell; otoferlin synaptic complex; surface plasmon resonance.

Figure 1.. Dopamine D1A carboxy sequence alignment and PCR amplification from rat OC.

(A) Alignment (with OMIGA 2.0) of the carboxy terminus aa sequences of trout hair cell (HC) D1A4 (ACA96732) with rat D1A (NP_036678) and human D1A (NP_000785). (B) PCR amplification of an ~286 bp cDNA product (including restriction sites; red arrow) corresponding to the carboxy terminus of dopamine D1A expressed in rat OC (NM_012546.2) (S, standards; OC, organ of Corti; B, blank).

Figure 2.. Dopamine D1A4 and snapin yeast co-transformation and protein localization.

(A) Co-transformation of trout HC dopamine D1A4 carboxy terminus in BD vector (Clontech) and snapin in AD vector (Clontech) in quadruple drop-out medium with antibiotics. First lane, interaction of trout HC D1A4 as bait and snapin-like protein as prey (multiple streakings), the latter originally identified as a binding partner for D1A4 carboxyl terminus in Y2H mating protocols. Second lane, negative control with D1A4-C as bait and empty prey vector. Third lane, negative control with empty bait vector and snapin prey sequence. (B) Reversal of bait and prey, snapin in BD vector, and D1A4-C in AD vector, in yeast co-transformation. First lane, snapin-like protein as bait and D1A4-C as prey. Second lane, negative control with snapin as bait and empty prey vector. Third lane, negative control with empty bait vector and D1A4-C prey sequence. (C) Snapin immunolocalization to trout saccular hair cells (thcs) detected with DAB. Snapin was specifically expressed in saccular hair cells, as would occur for a hair cell marker protein. Snapin immunoreactivity was observed close to the hair cell basolateral membrane (long arrows) and was concentrated at apical subcuticular plate sites (short arrow). Note the large unreactive afferent nerve passing through the connective tissue layer (asterisk), across the basal lamina, and into the sensory epithelium adjacent to unreactive supporting cells. Scale bar = 10 μm. (D) Dopamine D1A immunolocalization in cochlear outer hair cells of the adult rat. Immunoreactivity to D1A was localized to the base of a first row outer hair cell (short arrow), similar in position to dopaminergic efferent input. The Deiters’ cells were immunopositive (long arrow), observed at the level of the cell body and in phalangeal extensions wrapping around the cochlear outer hair cell. (E) Amino acid alignment of human snapin (NP_036569.1), rat snapin (NP_001164047.1), and thc snapin (AHB38897.1). (F) Alignment of aa for otoferlin C2F domains in human (NP_919224.1), rat (NP_001263649.1), and thc (AHB38896.1).

Figure 3.. Pull-down assays and SPR of dopamine D1A4 with snapin and snapin with otoferlin C2F in the thc.

(A) Interaction of trout hair cell (HC) D1A4 and snapin by pull-down analysis. Lane 1, purified trout HC snapin in pRSET-A fusion protein used in the present study; lane 2, protein standards (S); lane 3, PPI of trout snapin fusion protein (~50 ng) with GST–trout D1A4-C cleared lysate, pulled down with glutathione-Sepharose, and detected with anti-Xpress monoclonal antibody; lane 4, trout snapin fusion protein (~100 ng) with GST–trout D1A4-C lysate; lane 5, binding of trout snapin fusion protein (~100 ng) with GST–trout D1A4-C lysate and no beads, as a negative control; lane 6, trout snapin (~100 ng) incubated with GST-bacterial lysate as a negative control; lane 7, trout snapin fusion protein (~100 ng) mixed with Sepharose beads under the same conditions as above, representing another negative control. (B) Pull-down of C2F domain of trout saccular hair cell (HC) otoferlin with GST–trout HC snapin fusion protein. Lane 1, purified trout HC C2F fusion product used in this experiment, ~24 kDa; lane 2, protein standards; lane 3, interaction of trout HC C2F fusion protein (~250 ng) with GST–trout HC snapin cleared lysate, pulled down with glutathione-Sepharose, and detected with anti-Xpress monoclonal antibody; lane 4, ~50 ng of trout HC C2F fusion protein with GST–trout HC snapin lysate; lane 5, trout HC C2F fusion protein (~250 ng) with GST–trout HC snapin lysate and no beads as a negative control; lane 6, trout HC C2F (~250 ng) incubated with GST-bacterial lysate (no insert) as a negative control; lane 7, trout HC C2F fusion protein (~250 ng) mixed with Sepharose beads under the same conditions as above, representing another negative control. (C) Competitive pull-down assay for trout HC snapin with trout HC D1A4 and C2F domain of otoferlin. Lane 1, purified HC D1A4 in pRSET fusion protein at ~14 kDa; lane 2, purified trout HC C2F in pRSET fusion protein at ~24 kDa; lane 3, protein standards; lane 4, affinity-purified trout D1A4 and C2F fusion proteins (250 ng each) mixed with GST–trout snapin cleared lysate. The pull-down assay revealed that D1A4 and C2F were capable of binding snapin at the same time without steric hindrance, further confirmed in lane 5 with ~500 ng of affinity-purified D1A4 and C2F fusion proteins with detection with anti-Xpress monoclonal antibody. We validated the results with three different kinds of negative control reactions: lane 6, ~500 ng of both D1A4 and C2F mixed with GST–snapin and no beads; lane 7, ~500 ng of both the proteins with GST-bacterial lysate with unconjugated Sepharose beads, i.e. without snapin; lane 8, ~500 ng of both D1A and C2F fusion proteins with Sepharose beads alone under the same conditions. (D) Ca²⁺ dependence of the interaction of dopamine D1A4 as a ligand (L) with snapin as an analyte (A) (at 100 nM, dissolved in HBS-N buffer). RUs are indicated for binding with snapin at 68 μM Ca²⁺ (magenta), 26.5 μM Ca²⁺ (green), and 1 mM EGTA (cyan), compared with RU for HBS-N + 1 mM EGTA alone (blue) or HBS-N buffer alone (red). (E) Reversal of ligand and analyte: with snapin as a ligand and D1A4 as an analyte (100 nM), responses were obtained at 68 μM Ca²⁺ (blue), 26.5 μM Ca²⁺ (cyan), and 1 mM EGTA (green), compared with the response for HBS-N + 1 mM EGTA alone (blue) or HBS-N buffer alone (magenta). (F) Quantitative analysis of PPI between snapin and dopamine D1A4 as dependent on Ca²⁺ by SPR [15,16] with mean ± standard deviation, n = 3. ***P = 0.0095 for 26.5 vs. 68 μM Ca²⁺ for the unpaired, two-tailed t-test. (G) Kinetic series: purified D1A4 was immobilized on a CM5 sensor chip, and snapin was diluted in a series of concentrations in HBS-N buffer run at 68 μM Ca²⁺; 0 nM (gray), 10 nM (green), 20 nM (cyan), 40 nM (blue), 80 nM (red), and 160 nM (magenta). (H) Kinetic series for the reversal of ligand and analyte: snapin as a ligand and D1A4 as an analyte at 68 μM Ca²⁺ in HBS-N buffer; D1A4 at 0 nM (gray), 10 nM (yellow), 20 nM (red), 40 nM (cyan), 80 nM (green), 160 nM (magenta), and 320 nM (blue). (I) Ca²⁺ dependence of the interaction of trout HC snapin as a ligand with C2F as an analyte (100 nM dissolved in HBS-N buffer); 68 μM Ca²⁺ (blue), 26.5 μM Ca²⁺ (red), and 1 mM EGTA (magenta), compared with the response for HBS-N + 1 mM EGTA alone (green) or HBS-N buffer alone (cyan). (J) Reversal of ligand and analyte: Ca²⁺ dependence of binding between C2F as a ligand and snapin as an analyte(100 nM dissolved in HBS-N buffer); 68 μM Ca²⁺ (blue), 26.5 μM Ca²⁺ (green), and 1 mM EGTA (red), compared with the response for HBS-N + 1 mM EGTA alone (magenta) or HBS-N buffer alone (cyan). (K) Quantitative analysis of PPI between snapin and otoferlin C2F as dependent on Ca²⁺ by SPR [15,16] with mean ± standard deviation, n = 3. ****P = 0.0001 for 26.5 vs. 68 μM Ca²⁺ for the unpaired, two-tailed t-test; ***P = 0.008 for 26.5 vs. 68 μM Ca²⁺ for the unpaired two-tailed t-test.(L) Quantitative characterization of binding between snapin and C2F in a kinetic study. Purified snapin was immobilized as a ligand on a CM5 sensor chip, and C2F analyte was diluted in a series of concentrations with HBS-N buffer containing 68 μM Ca²⁺; 0 nM (red), 2.5 nM (magenta), 5 nM (cyan), 10 nM (red), 20 nM (cyan), 40 nM (green), 80 nM (magenta), and 160 nM (blue).(M) A kinetic series was performed for the reversal of C2F as a ligand and snapin as an analyte at 68 μM Ca²⁺ in HBS-N buffer; 0 nM (red), 5 nM (green), 10 nM (cyan), 20 nM (magenta), 40 nM (blue), 80 nM (red), and 160 nM (green).

Figure 4.. Interactions of rat D1A with snapin and snapin with otoferlin, positively dependent on Ca²⁺, and interaction of rat D1A with NSF, inhibited by Ca²⁺.

(A) Interactions between rat OC D1A and snapin by pull-down analysis. Lane 1, affinity-purified rat OC D1A in pRSET-A fusion protein at ~14 kDa; lane 2, protein standards; lane 3, binding interaction of rat OC D1A fusion protein (~250 ng) with GST–rat snapin cleared lysate, pulled down with glutathione-Sepharose, and detected with anti-Xpress monoclonal antibody; lane 4, rat D1A fusion protein (~500 ng) with GST–rat snapin cleared lysate; lane 5, binding of rat D1A fusion protein (~250 ng) with GST–rat snapin lysate and no beads as a negative control; lane 6, rat D1A (~250 ng) incubated with GST-bacterial lysate, negative control; lane 7, rat D1A fusion protein (~250 ng) mixed with Sepharose beads under the same conditions as above, representing another negative control. (B) Interaction of rat brain (BR) GST–snapin with otoferlin C2F domain. A hexahistidine–C2F fusion protein was produced from BR otoferlin cDNA amplified by PCR with C2F-2 primers (34 kDa), purified [14], and incubated with glutathione-Sepharose beads and GST–snapin fusion protein (from PCR amplification of rat brain snapin cDNA) in binding buffer (see Experimental). The interacting fusion protein was eluted with heat denaturation and detected on an SDS–PAGE gel blot with a mouse monoclonal anti-Xpress antibody (Invitrogen). Lane 1, GST negative control; lane 2, C2F(34 kDa, red arrow) with 1 mM EGTA; lane 3, only beads; lane 4, C2F with 100 μM added calcium; lane 5, molecular mass standards. (C) Competitive pull-down assays determining the interaction of rat OC D1A and otoferlin C2F with snapin in GST vector. Lane 1, affinity-purified rat OC D1A in pRSET fusion protein at ~14 kDa; lane 2, purified OC C2F domain of otoferlin in pRSET fusion proteins at ~24 kDa; lane 3, protein standards; lane 4, affinity-purified rat D1A and C2F fusion proteins, ~100 ng of each, were mixed with GST–OC snapin cleared lysate. The pull-down products detected with anti-Xpress monoclonal antibody showed that snapin was capable of binding to both D1A and C2F at the same time; lane 5, pull-down with ~300 ngof affinity-purified rat D1A and C2F fusion proteins, again indicating that both the interactions of D1A with snapin and C2F with snapin can occur together without steric hindrance, forming a three-protein complex. Negative control reactions included:lane 6, ~300 ng of both rat D1A and C2F were mixed with GST–rat snapin with no beads; lane 7, in another negative control ~300 ng of both the proteins were mixed with GST-bacterial lysate; lane 8, in a third negative control, ~300 ng of both D1A and C2F fusion proteins were mixed with Sepharose beads alone under the same conditions. (D) Interaction of D1A as an analyte (A) and affinity-purified rat OC NSF fusion protein as a ligand (L). Maximum responses are shown for the interaction of analyte D1A at 200 nM with ligand NSF at different concentrations of Ca²⁺. No interactions occurred for buffer alone (bar 1) or buffer + 1 mM EGTA (bar 2). The largest response for analyte was observed at 1 mM EGTA (bar 3) with a lesser response at26.5 μM Ca²⁺ (bar 4) and least at 68 μM Ca²⁺ (bar 5), overall indicating an inverse relationship between binding and [Ca²⁺]. The mean ± SD are presented (n = 3). ****P < 0.0001 for RU bar 3 vs. bar 4, unpaired, two-tailed Student’s t-test; ****P < 0.0001 for bar 4 vs. bar 5, unpaired, two-tailed Student’s t-test. (E) A kinetic series at 26.5 μM Ca²⁺ for binding between NSF (ligand) and D1A (analyte) at 160 nM (blue), 80 nM (green), 40 nM (red), 20 nM (magenta), 10 nM (cyan), 5 nM (gray), and 0 nM (yellow). (F) Western blot for full-length D1A from rat brain lysate detected with an affinity-purified rabbit polyclonal antibody raised against a peptide in the carboxyl terminus of D1A of rat origin (AB1765P, Millipore). Lane 1, Magic XP Western standards; lane 2, full-length D1A at 75 kDa (arrow). (G) Immunoprecipitation of NSF from rat brain lysate by antidopamine D1A antibody. Lane 1, Magic XP Western standards. Lane 2, anti-dopamine D1A immunoprecipitation of NSF (~80 kDa) from rat brain lysate in the absence of calcium (i.e. 1 mM EGTA) detected by specific anti-NSF antibody. Lane 3, antidopamine D1A immunoprecipitation of NSF at 100 μM Ca²⁺. Lane 4, western for full-length NSF in rat brain lysate. Lane 5, negative control using rabbit normal IgG immunoprecipitation with 1 mM EGTA. The blot is representative of at least three separate experiments. (H) Immunoprecipitation of full-length otoferlin (220 kDa) from rat brain lysate with dopamine D1A antibody (Millipore) with and without added calcium. Lane 1, western standards; lane 2, anti-D1A immunoprecipitation of otoferlin (OTO) (~220 kDa) at 100 μM Ca²⁺, detected with custom antiotoferlin antibody raised in rabbit. Lane 3, immunoprecipitation in the absence of calcium (i.e. 1 mM EGTA). A second smaller protein, ~140 kDa, in addition, was immunopositive with the otoferlin antibody for lanes 2 and 3. Lane 4, western blot of full-length otoferlin in rat brain lysate detected with anti-otoferlin antibody at ~220 kDa. Lane 5, negative control, rabbit normal IgG immunoprecipitation at 100 μM Ca²⁺. All experiments were repeated three times, and blots are representative of at least three separate experiments. (I) Quantitative analysis of immunoprecipitation by anti-D1A antibody of NSF illustrated in G. Band intensities were analyzed and averaged using the software Image J. The mean ± SE (error bars) (n = 3) were plotted using Microsoft Excel. Bar 1, immunoprecipitation of NSF at 0 μM Ca²⁺; bar 2, immunoprecipitation of otoferlin at 100 μM Ca²⁺. **P =0.0245, unpaired, two-tailed t-test for bar 1 vs. bar 2. (J) Quantitative analysis using the software Image J for immunoprecipitation illustrated in H (mean ± SE), n = 3. Bar 1, immunoprecipitation of otoferlin at 0 μM Ca²⁺; bar 2, immunoprecipitation of otoferlin at 100 μM Ca²⁺; bar 2 vs. bar 1, *P = 0.058, the unpaired, two-tailed t-test.

Figure 5.. Rat OC otoferlin C2 domains interact with AP2mu1 by pull-down and SPR.

(A) Interaction of rat OC otoferlin C2 domains with GST–AP2mu1 by pull-down assay. Lane 1, protein standards; lanes 2, 3, 5, and 7, pull-down by GST–AP2mu1 of C2A (20 kDa), C2B (24 kDa), C2D (29 kDa), and C2F (28 kDa from the use of C2F-3 primers), respectively, whereas lanes 4 and 6 show no protein bands at ~25 and ~28 kDa, representing C2C and C2E, respectively (300 ng for each C2 domain, anti-Xpress antibody was used for detection); lane 8, protein standards; lanes 9 and 10, negative controls for C2D and C2F with GST-bacterial lysate (no insert), respectively. (B) Binding interactions of OC AP2mu1 with otoferlin C2A and OC AP2mu1 with otoferlin C2F as a function of [Ca²⁺]: mean ± range, n = 2. (C) Binding interactions of OC otoferlin C2 domains with AP2mu1 (unphosphorylated) observed with pull-down assays were confirmed with SPR analysis, exhibiting both positive and negative dependence on Ca²⁺. A representative SPR analysis plot is illustrated with AP2 mu1 immobilized on a CM5 sensor chip as a ligand (L) and C2A domain as an analyte (A) (100 nM). An inverse relation of binding with [Ca²⁺] was observed (see also B). The binding response was enhanced at 1 mM EGTA (blue) and decreased with elevation ofCa²⁺: 26.5 μM Ca²⁺ (green), 68 μM Ca²⁺ (red), and 118 μM Ca²⁺ (magenta) with no binding for negative controls HBS-N + 1 mM EGTA (red) and HBS-N buffer alone (cyan). (D) The interaction of AP2mu1 with the C2B domain of otoferlin (100 nM) occurred in the absence of Ca²⁺ (1 mM EGTA, green) and was unchanged at 26.5 μM Ca²⁺ (blue) or 68 μM Ca²⁺ (red). No binding was found for negative controls HBS-N + 1 mM EGTA (magenta) or HBS-N SPR buffer alone (cyan). (E) The interaction of AP2mu1 with otoferlin C2D was positively dependent on Ca²⁺. Maximum binding was observed at 68 μM free Ca²⁺ (red), followed by intermediate levels of binding at 26.5 μM Ca²⁺ (magenta). Binding did occur in the absence of free Ca²⁺ (1 mM EGTA, green) compared with the response with HBS-N + 1 mM EGTA (cyan) or HBS-N buffer alone (blue). (F) The interaction of AP2mu1 (ligand) and otoferlin C2F (as an analyte at 100 nM) was also positively dependent on Ca²⁺ (also see B). A maximum response was found at 68 μM free Ca²⁺ (red) compared with that for 26.5 μM Ca²⁺ (blue). Again, there was a smaller but finite response for EGTA (1 mM, green). Negative controls included HBS-N + 1 mM EGTA (cyan) and HBS-N buffer alone (magenta). (G) A kinetic series is illustrated for AP2mu1 (unphosphorylated) as a ligand and otoferlin C2F as an analyte at 26.5 μM Ca²⁺. Affinity-purified AP2mu1 was immobilized on a CM5 sensor chip, and the C2F domain was diluted in a series of concentrations: 160 nM (green), 80 nM (magenta), 40 nM (red), 20 nM (cyan), 10 nM (blue), and 0 nM (green), yielding a K_D of 1.29 × 10⁻⁷ M.

Figure 6.. Alternatively, rat OC otoferlin C2 domains interact with NSF by pull-down and SPR.

(A) Interaction of rat OC otoferlin C2 domains with GST–NSF by pull-down assays. Lanes 1 and 10, western standards; lanes 2 and 4, GST negative controls (no NSF insert) with C2A and C2B domains, respectively; detection by anti-Xpress antibody. Lanes 3, 5, 7, and 9 indicate pull-down of C2A (20 kDa), C2B (24 kDa), C2D (29 kDa), and C2F (28 kDa) with GST–NSF, whereas lanes 6 and 8 indicate negative results for pull-down of C2C (25 kDa) and C2E (28 kDa) with GST–NSF; 300 ng of each C2 domain was used for each reaction in the representative experiment illustrated. (B) SPR-determined binding interactions of OC NSF with otoferlin C2A as a function of Ca²⁺ (mean ± range, n = 2) and also, OC NSF PPI with otoferlin C2F as a function of [Ca²⁺]: mean ± range, n = 3, P = 0.0538(*), the unpaired, two-tailed t-test for 1mM EGTA vs. 68 μM Ca²⁺.(C) SPR sensorgram analysis of interactions between rat NSF as ligand (L) and the C2 domains of otoferlin as analyte (A) with each C2 domain tested at 100 nM in buffer with Ca²⁺ at 26.5 μM. The otoferlin C2F (red) gives the maximum response, followed by C2D (green), C2A (magenta), and C2B (blue). Very little binding was observed for C2C (cyan) and none detectedfor C2E (gray), compared with the response to HBS-N buffer alone (red) as a negative control. (D) An inverse dependence on [Ca²⁺] was observed for binding between C2A (50 nM) and NSF. The maximum response (65 RU) was observed in the absence of Ca²⁺ (1 mM EGTA, green) relative to a lesser response at 26.5 μM Ca²⁺ (blue) and less yet at 68 μM Ca²⁺ (red). Negative controls included HBS-N + 1 mM EGTA (magenta) and HBS-N buffer alone (cyan). (E) Interaction of the C2B domain of otoferlin as an analyte at 50 nM with NSF as ligand was maximized in the absence of Ca²⁺ (1 mM EGTA, blue). The interaction occurred to a lesser extent at 68 mM free Ca²⁺ (red) and at 26.5 μM Ca²⁺ (green), although not following strict inverse concentration dependence. HBS-N buffer alone (cyan) and HBS-N buffer + 1 mM EGTA (magenta) served as negative controls. (F) The interaction of otoferlin C2D (50 nM) with NSF was positively dependent on Ca²⁺. The maximum response (43 RU) was obtained at 68 μM free Ca²⁺ (magenta), followed by 26.5 μM free Ca²⁺ (green). A measurable but low response was observed with no added Ca²⁺ (1 mM EGTA, blue) compared with negative controls HBS-N + 1 mM EGTA (red) and HBS-N buffer alone (cyan). (G) By SPR analysis, the interaction between NSF and the C2F domain was positively dependent on Ca²⁺ with maximum interaction recorded at 68 mM free Ca²⁺ (red), followed by the response at 26.5 μM free Ca²⁺ (blue). C2F did interact with NSF but to a lesser extent in the complete absence of Ca²⁺ (1 mM EGTA, green) compared with binding responses for negative controls, HBS-N + 1 mM EGTA (magenta), and HBS-N buffer alone (cyan) under the same conditions. (H) A kinetic series for interactions of NSF and C2F is illustrated. With purified NSF immobilized on a CM5 sensor chip, the responses are given for analyte otoferlin C2F (26.5 μM free Ca²⁺) at 320 nM (green), 160 nM (magenta), 80 nM (red), 40 nM (blue), 20 nM (cyan), 10 nM (green), and 0 nM (magenta). High-affinity interaction can be inferred from the calculated K_D of 2.3 × 10⁻⁹ M.

Figure 7.. Competitive protein interactions of AP2mu1 with NSF and AP2mu1 with otoferlin C2F as a function of AP2mu1 protein phosphorylation.

(A) Competitive binding interactions of rat OC NSF and AP2mu1 for the otoferlin C2F domain by GST pull-down assay and anti-Xpress detection. Lane 1, protein standards; lane 2, affinity-purified rat OC NSF and AP2mu1 fusion proteins (~500 ng each) mixed with rat OC otoferlin C2F in GST: C2F interacted with NSF (~82 kDa) to the exclusion of AP2mu1 (~32 kDa); lane 3, negative control for which NSF and AP2mu1 were mixed with GST-bacterial lysate (no insert). (B) Panel is similar to A, except that a specific anti-NSF monoclonal antibody (Experimental) was used for detection. (C) Western blot of rat OC adapter-associated kinase 1 (AAK1), aa 1–400, in pRSET-A fusion protein. Lane 1, protein standards; lane 2, affinity-purified fusion protein of AAK1 at (52 kDa) detected with anti-Xpress antibody; lane 3, pRSET-A bacterial lysate as a negative control. (D) Western blot of unphosphorylated and phosphorylated-enriched forms of AP2mu1, synthesized in vitro, and detected with antiphospho-AP2mu1 monoclonal antibody (Cell Signaling Technology). Lane 1, protein standards; lane 2, untreated form of AP2mu1 at ~54 kDa (full-length AP2mu1 + vector); lane 3, phosphorylated form of AP2mu1 by treatment with AAK1 at~54 kDa; lane 4, pRSET-A bacterial lysate as a negative control. (E) Interaction of otoferlin C2F domain in the GST construct with phosphorylated-enriched and unphosphorylated forms of AP2mu1 in pRSET-A by the pull-down assay for proteins in vitro. Lane 1, protein standards; lane 2, 250 ng of unphosphorylated form of AP2mu1 mixed with C2F-GST cleared lysate and the pull-down product detected by anti-Xpress, indicating limited binding between C2F and AP2mu1 at ~54 kDa; lane 3, 250 ng of phosphorylated-enriched form of AP2mu1 mixed with C2F in GST cleared lysate; results show that the binding interaction is up-regulated by using the phosphorylated-enriched form of AP2mu1 at ~54 kDa (detected with anti-Xpress monoclonal antibody). (F) Competitive binding interactions of phosphorylated-enriched and unphosphorylated forms of AP2 mu1 and rat OC NSF in pRSET-A with the C2F domain of otoferlin in the GST fusion construct by the pull-down assay. Lane 1, protein standards; lane 2, affinity-purified rat OC NSF and unphosphorylated AP2mu1 fusion proteins (each ~250 ng) mixed with GST-bacterial cleared lysate as a negative control; lane 3, affinity-purified rat NSF and unphosphorylated AP2mu1 fusion proteins (each ~250 ng) mixed with C2F-GST-bacterial cleared lysate. Pull-down with anti-Xpress detection indicates that C2F binds more strongly to the NSF (~82 kDa) than to unphosphorylated AP2mu1 (~54 kDa); lane 4 presents a negative control in which NSF and the unphosphorylated form of AP2mu1 were mixed with GST beads (without C2F) and no proteins were detected; lane 5, another negative control in which NSF and phosphorylated-enriched form of AP2mu1 were mixed with GST-bacterial lysate (no insert); lane 6, affinity-purified rat NSF and phosphorylated-enriched form of AP2mu1 fusion proteins (each ~250 ng) mixed with C2F-GST-bacterial cleared lysate. Pull-down indicated that C2F binding to AP2mu1 (~54 kDa), when phosphorylated, was enhanced (relative to that observed in lane 3) and C2F binding to NSF (~82 kDa) was reduced in a competitive binding assay, i.e. evidence for a molecular switch. Lane 7, another negative control in which NSF and the phosphorylated form of AP2 mu1 were mixed with GST beads (no C2F). No proteins were pulled down as detected with anti-Xpress antibody. (G) A second example of competitive binding interactions for phosphorylated-enriched and unphosphorylated AP2 mu1 and NSF in pRSET-A for the C2F domain of otoferlin in GST fusion protein. Lane 1, protein standards; lane 2, affinity-purified rat OC NSF and unphosphorylated AP2mu1 fusion proteins (~250 ng) mixed with GST-bacterial cleared lysate as a negative control; lane 3, affinity-purified rat NSF and unphosphorylated form of AP2mu1 fusion proteins (each ~250 ng) mixed with C2F-GST-bacterial cleared lysate. Again, pull-down indicated stronger binding of C2F with NSF (~82 kDa) than to AP2mu1 (at ~54 kDa); lane 4, affinity-purified rat NSF and the phosphorylated-enriched form of AP2mu1 fusion proteins (each ~250 ng) mixed with C2F-GST-bacterial cleared lysate. Pull-down again indicated that the phosphorylated form AP2mu1 (at ~54 kDa) binds more strongly to C2F than the unphosphorylated (lane 3), whereas in this competition with phosphorylated AP2mu1, NSF (~82 kDa) binds less strongly (compare with lane 3), again indicating that phosphorylation of AP2mu1 would constitute a molecular switch; lane 5, negative control with NSF and the phosphorylated form of AP2mu1 mixed with GST-bacterial lysate (without C2F), with anti-Xpress monoclonal antibody used for detection.

Figure 8.. Model for dopamine D1A protein interactions with otoferlin synaptic pathways suggesting that Ca²⁺ and phosphorylation are used as molecular switches to move from membrane fusion to endocytosis.

In the absence of Ca²⁺ influx, dopamine D1A receptor is hypothesized to bind via its carboxy terminus to NSF. In response to increases in [Ca²⁺]_i (a molecular switch), protein interactions of D1A to snapin and the otoferlin C2F domain (and full-length otoferlin) are promoted. Otoferlin C2F interacts with (unphosphorylated) NSF in preference to interacting with (unphosphorylated) clathrin AP2mu1. Influx of Ca²⁺ and consequent phosphorylation of AP2mu1 gives rise to a molecular switch and formation of an alternate protein complex with C2F realigning from its interaction with NSF to alternatively aligning with clathrin AP2 mu1, a cargo adaptor subunit for the plasma membrane that links the cargo protein to the clathrin-coated pit in endocytosis. (A) Plus D1A agonist: expression of phosphorylated AP2mu1 (red/pink) overlapping actin (phalloidin-blue) was enhanced at the base of cochlear IHCs relative to controls (C) after 5 min of exposure of rat OC to dopamine D1A receptor agonist SKF (10 μM). The IHCs, which alternate with supporting cells, were identified by cellular co-expression of the ribbon synaptic marker protein ribeye CtBP2 (see B), that is specifically expressed in the hair cells (~8 μm in Z-direction from IHC stereocilia) and not in supporting cells. (B) Plus D1A agonist: highlighting immunoreactivity for CtBP2 (green), a ribbon synapse hair cell marker (Santa Cruz primary antibody SC-5966). Clustering (in white) was observed of phosphorylated AP2mu1 (red), actin (blue), and CtBP2 (green) at the base of cochlear IHC occurring after 5 min incubation of rat OC with dopamine D1A receptor agonist SKF (10 μM). (C) Minus D1A agonist (control): immunofluorescence localization of phosphorylated AP2mu1 (red), actin (blue), and CtBP2 (green) at the base of IHCs in controls incubated in physiological saline without dopamine D1A receptor agonist SKF. (D) AP2mu1 (dots/IHC) (1-mm optical slices from z-stack confocal microscopy) is plotted for dopamine D1A agonist-treated OC (D1A) (n = 11) compared with control (C) OC incubated with saline (n = 8). The mean ± standard deviation for D1A and C were significantly different, P < 0.0001(****), by the GraphPad unpaired, two-tailed t-test.