Ca²⁺-binding proteins tune Ca²⁺-feedback to Ca_v1.3 channels in mouse auditory hair cells

Abstract

Sound coding at the auditory inner hair cell synapse requires graded changes in neurotransmitter release, triggered by sustained activation of presynaptic Ca_v1.3 voltage-gated Ca²⁺ channels. Central to their role in this regard, Ca_v1.3 channels in inner hair cells show little Ca²⁺-dependent inactivation, a fast negative feedback regulation by incoming Ca²⁺ ions, which depends on calmodulin association with the Ca²⁺ channel α₁ subunit. Ca²⁺-dependent inactivation characterizes nearly all voltage-gated Ca²⁺ channels including Ca_v1.3 in other excitable cells. The mechanism underlying the limited autoregulation of Ca_v1.3 in inner hair cells remains a mystery. Previously, we established calmodulin-like Ca²⁺-binding proteins in the brain and retina (CaBPs) as essential modulators of voltage-gated Ca²⁺ channels. Here, we demonstrate that CaBPs differentially modify Ca²⁺ feedback to Ca_v1.3 channels in transfected cells and explore their significance for Ca_v1.3 regulation in inner hair cells. Of multiple CaBPs detected in inner hair cells (CaBP1, CaBP2, CaBP4 and CaBP5), CaBP1 most efficiently blunts Ca²⁺-dependent inactivation of Ca_v1.3. CaBP1 and CaBP4 both interact with calmodulin-binding sequences in Ca_v1.3, but CaBP4 more weakly inhibits Ca²⁺-dependent inactivation than CaBP1. Ca²⁺-dependent inactivation is marginally greater in inner hair cells from CaBP4^−/− than from wild-type mice, yet CaBP4^−/− mice are not hearing-impaired. In contrast to CaBP4, CaBP1 is strongly localized at the presynaptic ribbon synapse of adult inner hair cells both in wild-type and CaBP4^−/− mice and therefore is positioned to modulate native Ca_v1.3 channels. Our results reveal unexpected diversity in the strengths of CaBPs as Ca²⁺ channel modulators, and implicate CaBP1 rather than CaBP4 in conferring the anomalous slow inactivation of Ca_v1.3 Ca²⁺ currents required for auditory transmission.

In the cochlea, Ca_v1.3 voltage-gated Ca²⁺ channels are essential for the proper development and function of inner hair cells (IHCs) as mechanosensory transducers of acoustic stimuli. Ca_v1.3 channels mediate L-type Ca²⁺ currents that trigger exocytosis of neurotransmitter from IHCs onto auditory nerve afferents (Platzer et al. 2000; Brandt et al. 2003, 2005) and Ca²⁺ action potentials in IHCs before the onset of hearing (Brandt et al. 2003; Marcotti et al. 2003). Mice lacking Ca_v1.3 are congenitally deaf (Platzer et al. 2000; Dou et al. 2004), illustrating how factors that regulate these channels can affect sound coding at this first synapse in the auditory pathway.

Most features of IHC Ca_v1.3 channels, including their fast activation kinetics and relatively negative thresholds for activation, are reproduced in recombinant systems (Xu & Lipscombe, 2001; Koschak et al. 2001). However, Ca_v1.3 Ca²⁺ currents in IHCs show considerably less Ca²⁺-dependent inactivation (CDI) than in transfected cells (Platzer et al. 2000; Xu & Lipscombe, 2001; Koschak et al. 2001). CDI is manifest as faster inactivation of Ca²⁺ currents (I_Ca) than Ba²⁺ currents (I_Ba) and depends on calmodulin binding to the IQ domain in the C-terminal domain of the pore-forming Ca_v1 α₁-subunit (Qin et al. 1999; Peterson et al. 1999; Zühlke et al. 1999). Mutations of the IQ domain in the Ca_v1.3 α₁-subunit (α₁1.3) inhibit CDI, and a α₁1.3 splice variant lacking the IQ domain was identified in the cochlea (Shen et al. 2006). However, this α₁1.3 variant is expressed mainly in outer hair cells (OHCs) and not IHCs, suggesting an alternative mechanism for the limited CDI of native Ca_v1.3 channels in IHCs.

We have shown previously that Ca²⁺ binding proteins related to calmodulin (CaBP1 and CaBP4) can physically displace calmodulin from Ca_v channels, but with distinct functional consequences (Lee et al. 2002; Zhou et al. 2004; Haeseleer et al. 2004; Zhou et al. 2005). CaBP1 is expressed primarily in the brain and retina and abolishes CDI of Ca_v1.2 channels (Zhou et al. 2004, 2005). CaBP4 is localized in photoreceptor nerve terminals, where its association with Ca_v1.4 causes the channels to activate at more hyperpolarized membrane potentials required for normal visual transduction (Haeseleer et al. 2004). Considering the analogous functions of Ca_v1.3 in IHCs and Ca_v1.4 in photoreceptors, we hypothesized that CaBP4 might be expressed in IHCs and confer the native phenotype of Ca_v1.3 Ca²⁺ current in these cells. Evidence favouring this hypothesis was provided in a recent study (Yang et al. 2006). Here, we further investigated the role of CaBPs in the regulation of Ca_v1.3 Ca²⁺ currents in IHCs. Our results show that CaBP4 has relatively minor effects on Ca_v1.3 CDI, which alone cannot account for the slowly inactivating properties of Ca_v1.3 in IHCs. Instead, the data implicate CaBP1 as a key candidate for regulating CDI in IHCs.

Methods

Ethical approval

All procedures involving animals were approved by the Institutional Animal Care and Use Committee at Emory University in accordance with National Institutes of Health guidelines and animal welfare guidelines at the University of Göttingen and the State of Lower Saxony. After killing the mice by cervical dislocation the skull was opened, and the cochlea was removed and opened at the apex so that the apical coil could be harvested for immuncytochemical or electrophysiological experiments.

Constructs and molecular biology

The following Ca_v subunit cDNAs were used: α₁1.3 containing exon 42 (GenBank no. AF370009 and AF370010 for additional sequence encoded by exon 42), β_1b (GenBank no. NM017346), α₁1.2 (GenBank no. M67515), β_2A (GenBank no. NM053581), β₃ (GenBank no. NM012838), β₄ (GenBank no. L02315), and α₂δ (GenBank no. M21948). FLAG-α₁1.3 and GST (glutathione-S-transferase)-tagged α₁1.3 C-terminal construct (GST-α₁1.3 CT: nucleotides 5886–6465) were previously described (Calin-Jageman et al. 2007). The I-E mutation was introduced at position 1608 in α₁1.3 and CT by quick-change mutagenesis with specific oligonucleotide primers. GST-NT corresponds to amino acids 1–117 of α₁1.3 subcloned into BamHI/NotI sites of pGEX4T-1. CaBP1, CaBP2, CaBP4 and CaBP5 constructs used for mammalian cell expression were previously described (Haeseleer et al. 2004; Zhou et al. 2004).

Cell culture, transfection, and lysate preparation

Culture and transfection of HEK293T cells by Geneporter transfection reagent (Genlantis, San Diego, CA, USA) was as previously described (Calin-Jageman et al. 2007). For immunoprecipitation, HEK293T cells were transfected with cDNAs encoding Ca_v1.3 (FLAG-α₁1.3 (6 μg), β_2A, and α₂δ (2 μg each) with or without CaBP4 (2 μg). For electrophysiological experiments, HEK293T cells plated on 35 mm dishes were transfected with ∼3 μg total DNA (FLAG-α₁1.3 or FLAG-α₁1.2 (1.5 μg), β (0.8 μg), α₂δ (0.8 μg) ± GFP-tagged CaBP1 (0.1 μg), GFP-tagged or untagged CaBP4 (0.1 μg) or GFP expression plasmid (0.01 μg)). These transfection conditions have been optimized to compensate for the inhibitory effect of CaBPs on Ca_v current density, and yield measurable whole-cell currents (0.1−0.5 nA) that are easily amenable to biophysical analysis. For electrophysiological recording, GFP fluorescence was used to identify cells cotransfected with Ca_v1 subunits and CaBPs. Due to the inherently weak fluorescence associated with the GFP-tagged CaBP4 construct, we generally used untagged CaBP4 in combination with GFP expression plasmid (0.01 μg) for detection of cotransfected cells. Since there was no significant difference in the effects of GFP-CaBP4 or untagged CaBP4, data resulting from both transfection methods were pooled.

Binding and coimmunoprecipitation assays

Methods for pull-down of CaBP1 or CaBP4 from transfected HEK293T cell lysates by GST-α₁1.3 proteins immobilized on beads were published previously (Zhou et al. 2005; Zhou et al. 2004). Co-immunoprecipitation of Ca_v1.3 and CaBP4 from transfected HEK293T cells was also as previously described (Calin-Jageman et al. 2007). Western blotting was performed with appropriate antibodies (CaBP1 1: 1000, CaBP4 1: 2000) followed by HRP-conjugated anti-rabbit antibodies (1: 2000) and enhanced chemiluminescent detection reagents (Amersham Biosciences, Piscataway, NJ, USA).

Immunohistochemistry

For Figs 1A and B and 5B, C and E and online supplemental material Supplemental Fig. 1b (CaBP2 and CaBP5), immunostaining and confocal analysis of whole-mount mouse organ of Corti (P14–15, P28, or 2 months old as indicated) were performed essentially as described (Sendin et al. 2007). The following antibodies were used: mouse anti-CtBP2 IgG1 (1: 200; BD Biosciences, San Jose, CA, USA), anti-CaBP2, 4 and 5 (1: 400 (Haeseleer et al. 2000) and AlexaFluor488- and AlexaFluor568-labelled secondary antibodies (1: 200; Invitrogen, Carlsbad, CA, USA). For Fig. 5A and Supplemental Fig. 1a and b (CaBP1), cochleae were fixed with 4% paraformaldehyde/phosphate-buffered saline (PBS) and decalcified with 10% EDTA/PBS for 24 h prior to immunostaining with rabbit anti-CaBP1 or CaBP4 antibodies (Haeseleer et al. 2000; 1: 500) and FITC donkey anti-rabbit IgG (1: 200; Jackson ImmunoResearch, West Grove, PA, USA). Before mounting, most samples were incubated in Hoechst 34580 (1: 1000, Invitrogen) for 5–10 min. Confocal images were acquired using a Leica SP2 or SP5 (Leica Microsystems, Mannheim, Germany) or Zeiss LSM 510 laser scanning confocal microscope. Processed images were assembled for display in Adobe Photoshop and Illustrator or Macromedia Freehand software.

Figure 1. CaBP4 is localized in IHCs and functionally interacts with Ca_v1.3.

A, immunofluorescence of CaBP4 in the organ of Corti from WT (A) and CaBP4^−/− (B) mice. Projections of confocal sections show immunofluorescence for CaBP4 (green) and CtBP2/RIBEYE (red) in IHCs and OHCs. Hoechst staining (blue) marks cell nuclei. Scale bars, 10 μm. Results shown are representative of at least 2 experiments performed with 6 organs of Corti. C, schematic diagram of Ca_v1.3 α₁ subunit indicating CT and NT fragments used for binding assays with CaBP4. D, CaBP4 pull-down with GST-tagged CT or NT. GST or GST-CT or NT was immobilized on glutathione beads and incubated with lysates from HEK293T cells transfected with CaBP4 in the presence (+) or absence of (−) of Ca²⁺. Bound CaBP4 was detected by Western blot with CaBP4 antibodies (top). Input represents ∼5% of protein used in assay. Ponceau staining shows levels of GST proteins used in assay. Results shown are representative of 3 experiments. E, co-immunoprecipitation of CaBP4 with Ca_v1.3. HEK293T cells were transfected with Ca_v1.3 + CaBP4 (lanes 1, 2, 5) or with CaBP4 alone (lanes 3, 4, 6), lysed, and subjected to immunoprecipitation with Ca_v1.3 antibodies with (+) or without (−) Ca²⁺. Input represents ∼5% of lysate used for assay. Co-immunoprecipitated CaBP4 was detected by Western blot with CaBP4 antibodies. Results shown are representative of three experiments.

Figure 5. CaBP1, 2, and 5 are localized in IHCs.

A–C, confocal micrographs of whole-mount preparations of mouse organ of Corti (P14–P15) labelled with antibodies against CaBP1 (A), CaBP2 (B), or CaBP5 (C) and FITC-secondary antibodies (green) followed by Hoechst staining of nuclei (blue). Images represent single (A), or projections of, confocal section(s) (B and C). CaBP1 staining was detected in IHC soma and overlying pillar cells (PC). Scale bars, 10 μm. Results shown are representative of experiments performed with at least 6 organs of Corti. D, RT-PCR analysis of CaBP variants in mouse cochlear extracts. Reactions were performed with (+RT) or without reverse transcriptase (−RT) as a negative control. E, presynaptic localization of CaBP1 in IHCs from 2-month-old WT and CaBP4^−/− (KO) IHCs. Confocal micrographs show whole-mounts of organs of Corti double labelled with antibodies against CaBP1 (green, E1, 2, 5, 6) and CtBP2 (red, E3, E7). Examples of colocalization are indicated with arrows and appear yellow in the merged images (E4, 8). Scale bars, 10 μm in E1, E2; 5 μm in E2−4, E6 −8. Results shown are representative of experiments performed with 4 organs of Corti.

RT-PCR

RNA was isolated from mouse (P14) cochlea with Trizol reagent (Invitrogen) and subject to RT-PCR with primers specific for CaBP1, 2, 4, or 5. Amplification of CaBP1 was performed with standard primers, while that for CaBP2, 4, and 5 required nested primers. Amplicons were analysed on agarose gels and were detected alongside negative control reactions that lacked reverse transcriptase. The following primer sets were used: CaBP1 forward: GACAAAGACAAGGATGGTTAC, reverse: CTCTTCAAAGTCCACTCGTCC; CaBP2 forward 1: ATGGTTCAGAGACCCATGG, reverse 1: CCGAGACATCATCCGAACAAA, forward 2: GCATGCATCTTCCTTCGACCC, reverse 2: GGTCAATGTCTTGGAGTATC; CaBP4 forward 1: GAGTTTGACACTGACCAGGAC, reverse 1: CTCGTCAAAGTCTATGGTGC, forward 2: CAGCACGTGAAGATGCGCATG, reverse 2: CAACATCTCATCCAGTTCAGTG; CaBP5 forward 1: ATGCAGTTTCCAATGGGTCCTG, reverse 1: CGAGACATCATCTTCACAAACA, forward 2: CAGGATGAGCTTGATGAGCTC, reverse 2: TGCTGCAG TTCT GCCAGTGT.

Electrophysiology

Electrophysiological data were acquired with EPC-9 patch-clamp amplifiers driven by Pulse software (HEKA Elektronik, Lambrecht/Pfalz, Germany) and analysed with Igor Pro software (Wavemetrics, Lake Oswego, OR, USA). For transfected HEK293T cells, extracellular recording solutions contained (mm): 150 Tris, 1 MgCl₂, and 10 BaCl₂ or 10 CaCl₂. Intracellular solutions consisted of (mm): 140 N-methyl-d-glucamine, 10 Hepes, 2 MgCl₂, 2 Mg-ATP, and 5 EGTA. The pH of intracellular and extracellular recording solutions was adjusted to 7.3 with methanesulphonic acid. Electrode resistances were typically 1–2 MΩ in the bath solution, and series resistance was ∼2–4 MΩ, compensated up to 80%. Current density was measured by dividing the amplitude of the Ca²⁺ current by the cell capacitance. For this purpose, the Ca²⁺ current was evoked by −90 to −20 mV (for Ca_v1.3), −80 to −10 mV (for Ca_v1.2), and −90 to +10 mV (for Ca_v1.3I-E). For analysis of I–V curves, currents were evoked by 20 ms pulses to varying test voltages from a holding voltage of −90 mV. Current amplitudes were plotted against test voltage and fitted with the Boltzmann function:

where g is the maximum conductance, V is the test potential, E is the apparent reversal potential, V_1/2 is the potential of half-activation, k is the slope factor, and b is the baseline. Statistical comparisons were by t test.

For IHCs, perforated patch-clamp recordings were made from the apical coil of freshly dissected organs of Corti from NMRI, C57Bl/6, or CaBP4^−/− mice (2–4 weeks of age) at the basolateral face of IHCs at room temperature. The intracellular recording solutions contained (mm), for exocytosis: 140 caesium gluconate, 13 TEA-Cl, 10 CsOH-Hepes; for CDI: 130 caesium gluconate, 10 TEA-Cl, 10 4-AP, 10 NaOH-Hepes, 1 MgCl₂. Both solutions contained amphotericin B (250 μg ml⁻¹). The extracellular solutions contained (mm), for exocytosis: 105 NaCl, 35 TEA-Cl, 2.8 KCl, 2 CaCl₂, 1 MgCl₂, 10 NaOH-Hepes, 10 d-glucose; for CDI: 104 NaCl, 35 TEA-Cl, 2.8 KCl, 1 MgCl₂, 1 CsCl, 5 4-AP, 10 NaOH-Hepes, 10 d-glucose and 5 Ca²⁺ or 5 Ba²⁺. All solutions were adjusted to pH 7.2. With these solutions, electrode resistances were typically 4–6 MΩ. All voltages were corrected for liquid junction potentials. C_m was measured as previously described (Moser & Beutner, 2000) and ΔC_m was estimated as the difference between the mean C_m over 60–400 ms after the end of the depolarization and the mean prepulse capacitance (400 ms). Integrals of the Ca²⁺ current were calculated from the total depolarization-evoked inward current, including Ca²⁺ tail currents after leak subtraction. Cells with a membrane current exceeding −50 pA at the standard holding potential of −86 mV were discarded from the analysis.

All averaged data are presented as the mean ± s.e.m. Statistical significance of differences between two groups was determined by Student's t test or ANOVA as indicated (SigmaStat, Systat Software Inc., San Jose, CA, USA).

In vivo analysis of auditory function in CaBP^4+/− and CaBP4^−/− mice

The development and characterization of CaBP4^−/− mice were previously described (Haeseleer et al. 2004). For auditory brainstem responses (ABR) and distortion product otoacoustic emissions (DPOAE), mice (4 weeks old) were measured essentially as previously described (Khimich et al. 2005). Briefly, for ABR, tone bursts (4/6/8/12/16/24/32 kHz, 10 ms plateau, 1 ms rise/fall) or clicks (of 0.03 ms) were applied in the free field at 20 Hz (System 2, Tucker-Davis Technology, Gainesville, FL, USA and high frequency speaker: Monacor International, Bremen, Germany). Intensities are displayed as sound pressure level (dB root mean square for tone bursts, dB peak equivalent for clicks). Thresholds were estimated with 10 dB precision by visual inspection of average ABR traces.

Results

CaBP4 interactions with Ca_v1.3

By immunohistochemistry, CaBP4 was expressed in both IHCs and OHCs after the onset of hearing (Fig. 1A), while it was absent early in development (P0, data not shown). Within P15 IHCs, CaBP4 labelling was distributed throughout the soma (Fig. 1A), which was also the case at later ages (P28; online supplemental material, Supplemental Fig. 1a). CaBP4 immunofluorescence was not restricted to presynaptic sites, identified by double-labelling with antibodies against CtBP2 (Fig. 1A). The absence of CaBP4 immunofluorescence in IHCs from CaBP4^−/− mice (Fig. 1B and Supplemental Fig. 1a) confirmed the specificity of the antibodies for detecting CaBP4 in cochlear tissue.

Next, we tested if CaBP4 directly interacts with Ca_v1.3 in pull-down assays with GST-tagged fragments of α₁1.3. We focused on the cytoplasmic N-terminal (NT) and C-terminal (CT) domains of α₁1.3 (Fig. 1C), since they serve as the primary CaBP interaction sites in other Ca_vα₁ subunits (Lee et al. 2002; Haeseleer et al. 2004; Zhou et al. 2004, 2005). In these experiments, CaBP4 associated with NT and CT, but not the GST control both in the presence and absence of Ca²⁺ (Fig. 1D). Ca²⁺-independent association of CaBP4 with the intact channel was also shown by coimmunoprecipitation from lysates of HEK293T cells (Fig. 1E). Together with previous findings that CaBP1 and CaBP4 do not require Ca²⁺ for binding to Ca_v channels (Lee et al. 2002; Haeseleer et al. 2004; Zhou et al. 2004; Zhou et al. 2005), these results suggest that CaBPs are constitutive subunits of Ca_v channels under basal Ca²⁺ conditions.

Since the NT and CT are modulatory sites for CDI in Ca_v1.2 (Ivanina et al. 2000; Zhou et al. 2005), we tested if CaBP4 binding to these regions affects CDI of Ca_v1.3. Inactivation was measured as I_res/I_pk, which was the current amplitude at the end of a 300 ms test pulse normalized to the peak current amplitude. CDI was calculated as the difference between I_res/I_pk for I_Ca and I_Ba. CaBP4 caused a small but significant inhibition of CDI (∼13% compared to Ca_v1.3 alone, P < 0.0001; Fig. 2A). CaBP4 inhibited inactivation of I_Ca (I_res/I_pk= 0.21 ± 0.02 for Ca_v1.3 + CaBP4 versus 0.14 ± 0.01 for Ca_v1.3 alone; P < 0.001) but not I_Ba (I_res/I_pk= 0.91 ± 0.02 for Ca_v1.3 + CaBP4 versus 0.95 ± 0.01 for Ca_v1.3 alone; P = 0.06), indicating a selective effect on CDI. However, CDI was still quite prominent in cells cotransfected with CaBP4 and Ca_v1.3 (Fig. 2A). CaBP4 also caused a modest inhibition of CDI for Ca_v1.2 (Fig. 2A), but this was in contrast to the complete blockade of Ca_v1.2 CDI caused by CaBP1 (Zhou et al. 2005; Zhou et al. 2004).

Figure 2. aBP4 is a weak modulator of Ca_v CDI.

A, inhibitory effect of CaBP4 on CDI of Ca_v1.3 and Ca_v1.2. Ca²⁺ or Ba²⁺ currents (I_Ca, I_Ba) were recorded in HEK293T cells transfected with Ca_v1.3 (α₁1.3, β_2a, α₂δ) or Ca_v1.2 (α₁1.2, β_2a, α₂δ) alone or cotransfected with CaBP4. Test pulses were 300 ms steps from −90 to −20 mV for Ca_v1.3 or 1 s steps from −80 to −10 mV for Ca_v1.2. Inactivation was measured as I_res/I_pk, which was the residual current amplitude at the end of the pulse normalized to the peak current amplitude. CDI is the difference in I_res/I_pk for I_Ca and I_Ba. Representative current traces for I_Ca (black or red) and I_Ba (grey) are shown. *P < 0.01 compared to I_Ba by t test; **P < 0.005 compared to Ca_v1.3 or Ca_v1.2 alone, by ANOVA and post hoc Bonferroni t test. Number of cells in each group indicated in parentheses. Scale bar, 0.2 s for Ca_v1.3 and Ca_v1.3 + CaBP4; 0.65 s for Ca_v1.2. Representative current traces are shown above. Scale bar, 800 ms for Ca_v1.2, 250 ms for Ca_v1.3. B, current independence of Ca_v1.3 inactivation. I_res/I_pk for cells obtained in A was plotted against peak current amplitude for I_Ca and I_Ba in cells transfected with Ca_v1.3 alone (left) or cotransfected with CaBP4 (right). Each point represents I_res/I_pk for a single cell. Data were linear as determined by Run's test, with regression line indicated. For both Ca_v1.3 alone and Ca_v1.3 + CaBP4, the slopes of the regression line did not differ significantly from zero (for Ca_v1.3 alone, P = 0.08 for I_Ca, P = 0.28 for I_Ba; for Ca_v1.3 + CaBP4, P = 0.58 for I_Ca, P = 0.29 for I_Ba; by ANOVA).

In these experiments and as noted previously (Yang et al. 2006), CaBP4 also significantly inhibited Ca_v1.3 current density (Table 1). If CDI of Ca_v1.3 depended on global Ca²⁺ signals resulting from multiple open channels, it was possible that CaBP4 inhibited CDI indirectly via secondary effects on channel density. If so, then as shown for Ca_v2 channels (Kreiner & Lee, 2006; Soong et al. 2002), inactivation of Ca_v1.3 I_Ca, but not I_Ba, should increase nonlinearly with whole-cell current amplitude. In addition, CaBP4 would not change the parameters of this relationship, but rather the range of current amplitudes sampled. In this case, the mean value for I_res/I_pk for I_Ca in cells with Ca_v1.3 + CaBP4 would be greater than that in cells transfected with Ca_v1.3 alone since smaller currents show less CDI. However, in cells transfected with Ca_v1.3 alone or cotransfected with CaBP4, inactivation of I_Ca and I_Ba did not vary significantly with current amplitude (Fig. 2B). For both groups, the relationship between current amplitude and I_res/I_pk was linear, with a slope not significantly different from zero. Thus, like Ca_v1.2 channels (Neely et al. 1994), CDI of Ca_v1.3 is independent of channel expression levels, perhaps due to a reliance on very local increases in Ca²⁺ through individual channels. This result argues against a role for the CaBP4-mediated decrease in current density in the mechanism by which CaBP4 inhibits CDI.

Table 1.

Current-densities measured in cells transfected with Ca_v1.3 ± CaBPs

Transfection	Current-density (pA pF−1)	P value	n
Cav1.3 β1	48.3 ± 10.1	—	9
+ CaBP4	36.8 ± 10.5	0.44	9
+ CaBP1	23.9 ± 6.3	0.05	10
+ CaBP2	28.9 ± 7.3	0.13	12
+ CaBP5	25.2 ± 8.6	0.08	12
Cav1.3 β2	30.0 ± 6.9	—	20
+ CaBP4	14.7 ± 2.3	0.03	20
Cav1.3 β3	28.8 ± 4.3	—	9
+ CaBP4	16.9 ± 4.9	0.09	9
Cav1.3 β4	39.8 ± 7.9	—	12
+ CaBP4	32.2 ± 6.9	0.48	12

Because modulation of CDI can vary with the identity of the auxiliary Ca_vβ subunit (Lee et al. 2000), we tested if CaBP4 might have stronger effects on CDI of Ca_v1.3 channels comprised of different β subunits. CaBP4 caused a significant but still modest level of inhibition of I_Ca inactivation in channels with β₁, β₂ and β₄ without any effects on I_Ba (Fig. 3). Interestingly, CaBP4 did not influence inactivation of I_Ca for Ca_v1.3 channels containing β₃. In contrast to the stimulatory effects of CaBP4 on voltage-dependent activation of Ca_v1.4 (Haeseleer et al. 2004), CaBP4 did not similarly influence Ca_v1.3 channels, regardless of Ca_vβ (Table 2). In addition, the negative effects of CaBP4 on current density were significant only for Ca_v1.3 channels containing Ca_vβ₂ (Table 1). These data indicate that CaBP4 differentially modulates Ca_v1.3 compared to Ca_v1.4, and has weak modulatory effects on CDI for Ca_v1.3 containing most Ca_vβ subunits.

Figure 3. Weak modulation by CaBP4 is independent of Ca_vβ subunit.

I_res/I_pk was determined for I_Ca as in Fig. 2A in cells cotransfected with Ca_v1.3 (α₁1.3, β, α₂δ) alone or cotransfected with CaBP4. Inhibition of inactivation by CaBP4 was measured as ΔI_CaBP4=[(I_res/I_pk for Ca_v1.3) − (I_res/I_pk for Ca_v1.3 + CaBP4)]/(I_res/I_pk for Ca_v1.3). *P < 0.02 compared to Ca_v1.3 alone by t test. Representative current traces are shown above. Scale bar, 250 ms. Number of cells indicated in parentheses.

Table 2.

Parameters for activation of Ca_v1.3 from I–V relationships

Transfection	V1/2 (mV)	P value	k	P value	n
IBa
Cav1.3 β1	−15.35 ± 1.86		−6.77 ± 0.32		10
+ CaBP4	−17.60 ± 1.87	0.40	−6.97 ± 0.47	0.73	8
Cav1.3 β2	−13.55 ± 1.35		−8.52 ± 0.18		10
+ CaBP4	−10.02 ± 1.09	0.058	−8.71 ± 0.23	0.53	10
Cav1.3 β3	−14.05 ± 2.25		−9.90 ± 0.63		10
+ CaBP4	−13.36 ± 2.07	0.83	−7.53 ± 0.50	0.01	7
Cav1.3 β4	−17.28 ± 1.87		−7.23 ± 0.29		9
+ CaBP4	−14.58 ± 1.77	0.78	−7.09 ± 0.24	0.72	9
ICa
Cav1.3 β1	−7.31 ± 0.66		−7.44 ± 0.37		10
+ CaBP4	−7.32 ± 0.86	0.99	−7.80 ± 0.23	0.42	12
Cav1.3 β2	−6.86 ± 1.09		−8.43 ± 0.31		10
+ CaBP4	−6.55 ± 1.04	0.84	−8.84 ± 0.28	0.33	12
Cav1.3 β3	−5.44 ± 0.73		−8.73 ± 0.21		9
+ CaBP4	−7.29 ± 1.63	0.33	−8.39 ± 0.35	0.41	7
Cav1.3 β4	−8.68 ± 1.44		−7.39 ± 0.39		11
+ CaBP4	−8.23 ± 0.75	0.31	−8.31 ± 0.33	0.085	12

Normal hair cell synaptic transmission in mice lacking CaBP4

To address the significance of CaBP4 as a modulator of auditory Ca_v1.3 channels, we tested if the absence of CaBP4 affected CDI of IHCs and sound coding at the IHC afferent synapses. Patch-clamp recordings of IHCs in the apical coil of the organ of Corti monitored I_Ca and exocytic membrane capacitance changes (ΔC_m) reflecting neurotransmitter release (Moser & Beutner, 2000) from wild-type (WT, C57Bl/6) and CaBP4^−/− mice. As predicted from our results in transfected cells (Figs 2 and 3), CDI was marginally, but significantly enhanced (from −0.08 in WT to −0.13 in CaBP4^−/− IHCs, P = 0.036), while inactivation of I_Ba was not affected in CaBP4^−/− IHCs (P = 0.553; Fig. 4A–C). In addition, voltage-dependent activation and peak amplitudes of I_Ca were not different in WT and CaBP4^−/− IHCs (data not shown). As expected, analyses of I_Ca integrals and ΔC_m (Fig. 4D and E) revealed that Ca²⁺ influx and exocytosis were comparable in CaBP4^−/− and WT IHCs for depolarizations up to 200 ms (Fig. 4D and E).

Figure 4. Hearing is not impaired in CaBP4 knockout mice.

A and B, mean I_Ca (A) and I_Ba (B) of WT (black, n = 12) and CaBP4^−/− (grey, n = 14) IHCs evoked by 500 ms depolarizations to −16 mV. Currents were normalized to their peak amplitude and displayed with standard errors. C, I_res/I_pk and CDI were determined as in Fig. 2A except that currents were evoked as in (A and B) and residual current amplitude was measured at 300 ms. D, representative I_Ca and baseline-subtracted C_m traces (200 ms step to −16 mV). E, grand average of ΔC_m (top) and I_Ca integrals (bottom) plotted against pulse duration for WT (n = 25) and CaBP4 ^−/− (n = 7) IHCs. F, grand average of ABR evoked by suprathreshold clicks (80 dB) in WT (black) and CaBP4^−/− mice (grey). Roman numerals denominate the individual ABR peaks arising from the neurons of the ascending auditory pathway. G, ABR audiograms of WT (black) and CaBP4^−/− mice (grey) obtained by tone burst and click stimulation.

Consistent with these findings, we found normal auditory brainstem responses (ABR) to click or short tone burst stimuli in CaBP4^−/− mice. ABR waveforms (Fig. 4F), latencies (data not shown) and thresholds (Fig. 4G) were not different from WT mice. Distortion product otoacoustic emissions were also unaltered in CaBP4^−/− mice, indicating intact cochlear amplification (data not shown). These findings confirm that CaBP4 is a relatively modest suppressor of Ca_v1.3 CDI in IHCs, but is not essential for normal IHC development and IHC function in auditory transduction.

CaBP1 is a candidate modulator of Ca_v1.3 CDI in IHCs

To explore alternative mechanisms for inhibiting CDI in IHCs, we assessed the expression other CaBP isoforms by immunohistochemistry in the mouse organ of Corti. In addition to that for CaBP4, immunolabelling for CaBP1, CaBP2 and CaBP5 was detected in IHCs (Fig. 5A–C). The specificity of the immunostaining was supported by the substantial reduction of signal upon preadsorption of the CaBP antibodies with the corresponding immunogen (Supplemental Fig. 1b). The expression of CaBP1, CaBP2 and CaBP4 in cochlear extracts was confirmed by RT-PCR analysis (Fig. 5D). However, CaBP5 was not detected by this method perhaps due to the low expression levels of CaBP5 relative to other CaBPs in IHCs. Interestingly, CaBP1 showed a punctate distribution at the base of IHCs from older mice (P28, data not shown, and 2 months old; Fig. 5E) in contrast to the diffuse localization of CaBP1 in IHCs from younger mice (P14; Fig. 5A). The CaBP1-labelled puncta strongly colocalized with labelling for CtBP2. These data suggested the likely colocalization of CaBP1 with presynaptic Ca_v1.3 channels, which show a similar subcellular distribution (Brandt et al. 2005). The presynaptic localization of CaBP1 was qualitatively similar in CaBP4^−/− IHCs (Fig. 5E, lower panels). If CaBP1 more prominently suppressed CDI of Ca_v1.3 in IHCs, CaBP1/Ca_v1.3 interactions could explain the relatively normal auditory phenotype in CaBP4^−/− mice.

To test this, we compared the capacities of CaBP1, 2 and 5 as modulators of Ca_v1.3 CDI in transfected HEK293T cells. While CaBP5 had modest effects on CDI that were not significantly different from those of CaBP4 (P = 0.94), CaBP2 had no impact in this regard (I_res/I_pk= 0.14 ± 0.01 for + CaBP2 versus 0.12 ± 0.01 for Ca_v1.3 alone; P = 0.42; Fig. 6). However, CaBP1 caused a significantly stronger suppression of CDI than CaBP4 (P < 0.001; Fig. 6). Similar to CaBP4, CaBP1 interacted with the NT and CT of α₁1.3 in a Ca²⁺-independent manner in pull-down assays (Fig. 7A). Thus, while CaBP1 and CaBP4 may bind to the same sites in α₁1.3, they exhibit distinct modulatory strengths with respect to inactivation of I_Ca.

Figure 6. CaBP1 causes robust suppression of I_Ca inactivation compared to other CaBP variants.

Distinct effects of CaBP1, 2, 4 and 5 on inactivation of Ca_v1.3 in transfected HEK293T cells. Measurements of I_Ca, I_res/I_pk, and ΔI_CaBP were as in Fig. 3 for cells transfected with Ca_v1.3 (α₁1.3, β_2a, α₂δ) alone or cotransfected with CaBP1, 2, 4, or 5. *P < 0.02 compared to Ca_v1.3 alone by t test. **P < 0.001 compared to CaBP4 by ANOVA and post hoc t test. Number of cells indicated in parentheses.

Figure 7. CaBP1 binds to same sites in α₁1.3, but shows stronger expression levels and cell-surface targeting compared to CaBP4.

A, CaBP1 pull-down with GST-tagged CT or NT performed in the presence (+) or absence (−) of Ca²⁺ as in Fig. 1D. Results shown are representative of 3 experiments. B, Western blot of lysates from cells used for electrophysiological recording shows decreased levels of CaBP4 compared to CaBP1 in cells cotransfected with Cav1.3. Lanes 1 and 2 represent lysates of cells transfected with Ca_v1.3 alone (lane 1) or cotransfected with CaBP (lane 2). Lanes 3 and 4 contain purified GST-tagged CaBP1 (left panel) or His-tagged CaBP4 (right panel) at the indicated quantities. Western blotting was with CaBP1 or CaBP4 antibodies. Results shown are representative of 3 experiments. C, confocal micrographs of GFP-tagged CaBP1 or CaBP4 cotransfected with Ca_v1.3 as for electrophysiological experiments. Top panels show fluorescence images overlaying DIC image (scale bars, 50 μm). Higher magnification views in lower panels reveal GFP fluorescence restricted to the plasma membrane of cells cotransfected with GFP-CaBP1 (left) as compared to diffuse localization of GFP-CaBP4 (scale bars, 20 μm). Results shown are representative of 3 experiments.

Our electrophysiological results could also be explained by differences in the relative expression levels and/or transfection efficiency of CaBP1 and CaBP4. To address these possibilities, we performed Western blot analyses of lysates of cells transfected with Ca_v1.3 alone or cotransfected with CaBP1 or CaBP4. To control for differences in immunoblotting efficiencies of CaBP1 and CaBP4 antibodies, signals were qualitatively compared to those corresponding to known quantities of the purified proteins. With this approach, we consistently observed lower levels of CaBP4 than CaBP1 in transfected cell lysates (Fig. 7B). However, when GFP-tagged CaBP1 or CaBP4 constructs were cotransfected with Ca_v1.3 as for electrophysiological experiments, we did not find significantly weaker transfection efficiency of CaBP4 compared to CaBP1 (Fig. 7C upper panels). We did note that GFP-CaBP4 fluorescence was generally weaker and diffusely localized, in contrast to the strong localization of GFP-CaBP1 at the plasma membrane (Fig. 7C, lower panels). These results suggest that the weak effects of CaBP4 in modulating Ca_v1.3 could be influenced by limited cell-surface targeting and/or lower expression levels of CaBP4 compared to CaBP1 in transfected cells.

Differential modulation of Ca_v1.3 CDI by CaBP4 and CaBP1

The distinct effects of CaBP1 and CaBP4 could also be mediated by separate molecular determinants for each in the NT or CT of α₁1.3. If so, then alterations in the NT or CT should differentially affect modulation by CaBP1 and CaBP4. To test this, we initially deleted portions of the NT from α₁1.3, since similar deletions in α₁1.2 abolished modulation of Ca_v1.2 by CaBP1 (Zhou et al. 2005). Small and large deletions of the α₁1.3 NT resulted in non-functional channels, which precluded functional analyses of this region. Therefore, we focused our efforts on the α₁1.3 CT and substituted the isoleucine with glutamate in the IQ domain in the CT, since similar mutations in α₁1.2 inhibit binding and modulation of Ca_v1.2 by CaBP1 (Zhou et al. 2004, 2005). Consistent with the essential role of the IQ domain in transducing CDI of Ca_v1 channels (Peterson et al. 1999; Zühlke et al. 1999; Qin et al. 1999), the I-E mutation disrupted, although not completely, CDI of Ca_v1.3 (Ca_v1.3I-E, Fig. 8A). In addition, GST-tagged CT fusion proteins containing the I-E mutation showed weaker binding to CaBP4 and CaBP1 in pull-down assays (Fig. 8B), which suggested that Ca_v1.3I-E would show limited regulation by these CaBPs. While Ca_v1.3I-E channels were completely insensitive to modulation by CaBP4 (Fig. 9A and B), there were residual effects of CaBP1 in that I_Ca inactivated ∼2 times less in cells transfected with Ca_v1.3I-E + CaBP1 than with Ca_v1.3 alone (P < 0.001; Fig. 9C).

Figure 8. Mutation of the IQ domain in α₁1.3 inhibits CDI and binding of CaBP1 and CaBP4.

A, effect of I-E mutation on CDI. HEK293T cells were transfected with wild-type Ca_v1.3 or channels containing α₁1.3 with I-E substitution in IQ domain (+β_1b, α₂δ). Same analysis as in Fig. 2A for CDI. *P < 0.001 compared to I_Ba; **P < 0.001 compared to Ca_v1.3; by t test. Number of cells indicated in parentheses. B, effect of I-E mutation on CaBP binding. Pull-down of CaBP1 or CaBP4 by GST, CT, or CT containing I-E mutation was performed as in Fig. 7A. Results shown are representative of 3 experiments.

Figure 9. I-E mutation in α₁1.3 blocks effects of CaBP4 but partially spares modulation by CaBP1.

A and B, loss of CaBP4 effects on inactivation of Ca_v1.3I-E. HEK293T cells were cotransfected with CaBP4 and Ca_v1.3 or Ca_v1.3I-E (α₁1.3 or α₁1.3I-E, β_1b, α₂δ). Same analysis as in Fig. 3 for inactivation of I_Ca (A) and I_Ba (B). By t test, *P < 0.02 compared to Ca_v1.3 alone; **P < 0.001 compared to Ca_v1.3. C and D, effects of CaBP1 on inactivation are partially spared in Ca_v1.3I-E. Cells were cotransfected with CaBP1 and Ca_v1.3 or Ca_v1.3I-E (α₁1.3 or α₁1.3I-E, β_1b, α₂δ). Same analysis as in C. *P < 0.01, **P < 0.0001 compared to Ca_v1.3, by t test. In A–D, scale bars, 200 ms. E, decreased VDI of I_Ba in cells cotransfected with CaBP1 but not CaBP1. I_Ba was evoked by 1 s pulses from −90 mV to −20 mV in cells transfected with Ca_v1.3 alone, +CaBP4, or +CaBP1. *P < 0.001 by ANOVA and post hoc Bonferroni test. Vertical scale bars, 0.2 nA. In A–E, number of cells is indicated in parentheses.

Unexpectedly, and in contrast to CaBP4, CaBP1 significantly inhibited voltage-dependent inactivation (VDI) exhibited by I_Ba (∼36% compared to Ca_v1.3 alone, P < 0.001). In these experiments, CaBP regulation of VDI was probed with channels containing the Ca_vβ₁ subunit, which show more prominent VDI than those with Ca_vβ₂ (compare Fig. 9B to Fig. 2A). The effect of CaBP1 on VDI was completely spared by the I-E mutation (P = 0.23 for Ca_v1.3 versus Ca_v1.3I-E; Fig. 9D). During longer depolarizations (1 s), the inhibitory effect of CaBP1 on VDI was particularly apparent. While CaBP4 had no effect on inactivation of I_Ba (I_res/I_pk= 0.29 ± 0.03 for Ca_v1.3 + CaBP4 versus 0.34 ± 0.03 for Ca_v1.3 alone; P = 0.86), CaBP1 caused I_Ba to inactivate significantly less than in cells transfected with Ca_v1.3 alone or cotransfected with CaBP4 (Fig. 9E). These data reveal distinct molecular mechanisms by which CaBP1 and CaBP4 regulate Ca_v1.3 inactivation: while required for effects of CaBP4 on I_Ca, the IQ domain is necessary but not sufficient for the full modulation of I_Ca and I_Ba by CaBP1.

Discussion

Our study provides new insights into the physiological roles of CaBPs as integral subunits and modulators of Ca_v channels. First, CaBPs are not equally effective modulators of CDI. Second, CaBP4 and CaBP1 differ in how they suppress inactivation of Ca_v1.3: CaBP4 moderately limits inactivation of I_Ca through interactions with the α₁1.3 IQ domain, while CaBP1 strongly inhibits inactivation of I_Ca and I_Ba through interactions with the IQ domain and some other region, perhaps the NT, of α₁1.3. Third, CaBP4 contributes to, but alone does not account for, the unusually slow inactivation of Ca_v1.3 I_Ca in IHCs. The dispensability of CaBP4 for auditory function most likely results from additional modulation of Ca_v1.3 by CaBP1 and potentially other CaBPs, which are localized with CaBP4 in IHCs. Differential modulation of Ca_v1.3 by CaBPs may endow IHCs with a diverse repertoire of Ca²⁺ signals that may fine-tune transmission of acoustic stimuli.

Our results extend but also differ from those of Yang et al. (2006), who found that CaBP4 and CaBP1 both abolished CDI of Ca_v1.3 channels in transfected HEK293 cells. We employed the long variant of α₁1.3 with exon 42, which contains an additional ∼500 amino acids downstream of the IQ domain compared to the shorter α₁1.3 variant with exon 42A (Xu & Lipscombe, 2001) used by Yang et al. However, we did not find significant differences in CaBP modulation of Ca_v1.3 containing these α₁1.3 variants (data not shown). The discrepancy may arise from the use of an inducible CaBP4 expression system in which high levels of Ca_v1.3 expression by transient transfection were achieved prior to induction of CaBP4 in the previous study. This method, and higher permeant ion concentrations (20 mmversus 10 mm in our study) during electrophysiological recordings, overcomes the inhibitory effects we have noted of CaBPs on Ca_v expression levels (Lee et al. 2002; Zhou et al. 2004). At the same time, this approach may also mask functional differences in CaBPs, which might prevail under non-saturating CaBP expression levels, such as in our experiments where Ca_v1.3 was transiently cotransfected with CaBP1 or CaBP4.

In transfected cells, the decreased plasma membrane targeting of CaBP4 compared to CaBP1 may limit the modulatory potential of CaBP4. Unlike CaBP4, CaBP1 is N-terminally myristoylated (Haeseleer et al. 2000), which may aid in its cell-surface targeting and coassembly with Ca_v1.3 channels. Lack of this modification may weaken CaBP4 interactions with functional Ca_v1.3 channels, which may negatively influence the stability and net expression levels of CaBP4 in cotransfected cells. While electrophysiological analysis with Ca_v1.3I-E mutants suggest fundamental distinctions in how CaBP1 and CaBP4 interact with Ca_v1.3 channels, we cannot rule out that differential expression/targeting of these CaBPs also contributes to their strengths as Ca_v1.3 modulators.

Regardless of how differences in CaBP4 and CaBP1 in IHCs influence their roles in Ca_v1.3 regulation, our results argue against CaBP4 as the primary regulator of Ca_v1.3 in IHCs. Mouse IHCs lacking CaBP4 showed only a minor increase in CDI, indicating that CaBP4 is not essential for, but may contribute to, the peculiar slow CDI of Ca_v1.3 channels in IHCs. This may reflect the relatively weak effects of CaBP4 as a Ca_v1 modulator and/or limited presynaptic targeting/expression levels compared to other CaBPs. The dispensability of CaBP4 for hair cell sound coding is further demonstrated by the normal auditory brainstem responses in mice lacking CaBP4. The auditory physiology of CaBP4^−/− mice was conducted at an age when CaBP1 is strongly localized presynaptically both in WT and CaBP4^−/− IHCs. The less intense immunolabelling of CaBP1 in adult IHCs may have been missed by Yang et al. leading to their conclusion that CaBP4 was the critical modulator of Ca_v1.3 function at later ages (Yang et al. 2006). The use of cryosections in the previous study may have impaired detection of punctate labelling in adult IHCs, since only a few puncta might be visible in a single IHC with this method. In our experiments, presynaptic labelling was readily apparent in an entire row of IHCs in projections of confocal micrographs. Why CaBP1 exhibits a diffuse localization in P14 IHCs and becomes presynaptically clustered at later ages is not clear, but clearly allows CaBP1 to be well-situated to modulate Ca_v1.3 channels.

We also note that in transfected cells, CaBP1 not only strongly suppresses CDI (inactivation of I_Ca) but also inhibits VDI (inactivation of I_Ba), which was not seen with CaBP4 (Fig. 9). This latter point is relevant since Ca_v1.3 currents in IHCs show dramatically slower CDI and VDI than in recombinant systems (Platzer et al. 2000; Koschak et al., 2001). The fact that cotransfection of Ca_v1.3 with CaBP1 fully reproduces the native properties of Ca_v1.3 channels further implicates CaBP1 as a modulator of Ca_v1.3 in IHCs, which may support the crucial roles of Ca_v1.3 channels in hearing.

Supplemental material

Online supplemental material for this paper can be accessed at: http://jp.physoc.org/cgi/content/full/jphysiol.2007.142307/DC1 and http://www.blackwell-synergy.com/doi/suppl/10.1113/jphysiol.2007.142307