beta 3: an additional auxiliary subunit of the voltage-sensitive sodium channel that modulates channel gating with distinct kinetics

Abstract

The voltage-sensitive sodium channel confers electrical excitability on neurons, a fundamental property required for higher processes including cognition. The ion-conducting alpha-subunit of the channel is regulated by two known auxiliary subunits, beta1 and beta2. We have identified rat and human forms of an additional subunit, beta3. It is most closely related to beta1 and is the product of a separate gene localized to human chromosome 11q23.3. When expressed in Xenopus oocytes, beta3 inactivates sodium channel opening more slowly than beta1 does. Structural modeling has identified an amino acid residue in the putative alpha-subunit binding site of beta3 that may play a role in this difference. The expression of beta3 within the central nervous system differs significantly from beta1. Our results strongly suggest that beta3 performs a distinct neurophysiological function.

Figure 1

β3 expression in cells and tissues. (a) Northern blot of β3 expression in rat tissues and PC12 cell lines. (b) Expression of β3 and β1 in rat tissues, PC12, and A35C variant cell lines detected by PCR at each of two cDNA concentrations.

Figure 2

In situ distribution of sodium channel subunits in adult rat brain. X-ray autoradiographs of separate sagittal sections of rat brain (taken from the same animal) showing the distribution of rat αIIA (a–c), rat β1 (d–f), and rat β3 (g–i) mRNA transcripts as revealed by in situ hybridization with specific oligonucleotide probes. Control reactions with 100-fold excess unlabeled probes are shown for αIIA (c), β1 (f), and β3 (i). Slides were exposed to x-ray film for 10 days. Dark areas indicate high expression levels. Cb, cerebellum; Ctx, cortex; CP, caudate putamen.

Figure 3

Sequence comparison and three-dimensional modeling of β3. (a) Amino acid sequences of rat and human β3, aligned with the sequences of rat β1 (SWISS-PROT Q00954; ref. 8) and the extracellular domain of rat myelin P₀ (SWISS-PROT P06907; ref. 29). The sequence numbering is based on rat β3, starting from the predicted N terminus of the mature protein. Amino acid identities with rat β3 are indicated by shading. The putative signal sequence and internalization signal are underlined and labeled. The putative transmembrane domain (TM) is boxed. Three negatively charged amino acid residues, previously identified as part of the α-subunit binding site of β1, are boxed. Invariant residues and the position of amino acids characteristic of the IgV domain are indicated beneath the sequence of myelin P₀: h, hydrophobic; l, aliphatic; %, neutral or hydrophobic; +, base; =, hydrophobic or Ser or Thr; #, Gly or Ala (rarely Asp; ref. 17). Secondary structure elements in the crystal structure of myelin P₀ (19) used to model β3 are also shown: arrow, β-strand; cylinder, α- or 3₁₀-helix. The multiple alignment was generated with clustalw (30) and formatted with alscript (31). (b) The model for the three-dimensional structure of the mature extracellular domain (residues 1–123) of rat β3. The model was generated with modeller (32) by using the crystal structure of rat myelin P₀ (PDB 1neu; ref. 19) as a template and the alignment shown in a. The side chains of acidic residues in the putative α-subunit binding site are shown in ball-and-stick representation. Two predicted disulfide bonds are labeled in black. N-linked glycosylation sites (NXT and NXS; ref. 33) are indicated by asterisks. The potential glycosylation site on the F strand (N97) points away from the viewer and is below the plane of the paper. Fig. 3a was drawn with molscript (34) and raster3d (35). Note that in this model the B strand is broken into two parts labeled B and B′, respectively. This secondary structure assignment is based on the definition of Kabsch and Sander (36) for the PDB entry 1neu and is different from the assignment described in the original paper (19).

Figure 4

Coexpression of rat αIIA with rat β3-subunit modifies inactivation kinetics. (a) Na⁺ currents recorded from oocytes expressing rat αIIA, rat αIIA + rat β1, and rat αIIA + rat β3-subunits. Inward Na⁺ currents were evoked by applying depolarizing pulses in 5-mV increments from a holding potential of −100 mV, from −80 mV to +30 mV. Duration of the pulses was 50 ms. (b) Normalized Na⁺ currents from oocytes expressing rat αIIA, rat αIIA + rat β1, and rat αIIA + rat β3-subunits. Currents evoked by a voltage pulse to −10 mV were normalized to peak amplitudes. Inactivation of Na⁺ currents at −10 mV were fitted with a double exponential decay: I = A1 exp(−t/τ1) + A2 exp(−t/τ2) + C, where A1 and A2 are the relative amplitudes of fast and slow components, τ1 and τ2 are the inactivation time constants, and C is the steady-state asymptote. See Table 2 for fit parameters. (c) Recovery from inactivation of αIIA coexpressed with β1 or β3. The recovery pulse protocol was a 1-s inactivating pulse to −10 mV followed by conditioning pulses to −100 mV for increasing periods of time (from 1–1,000 ms), followed by a test pulse to −10 mV. Points were sampled every 1 ms from 1 to 20 ms and then every 50 ms from 50 to 1,000 ms. Peak current amplitudes measured during the test pulse were normalized to the peak currents evoked during the inactivating pulse and were plotted as function of conditioning pulse duration. ▵, αIIA; ●, αIIA + β1; ○, αIIA + β3. Data were fitted with a double exponential equation: I = 1 − [A1 exp(−t/τ1) + A2 exp(−t/τ2)], where A1 and A2 are the relative amplitudes of recovery and τ1 and τ2 are the recovery time constants. See Table 2 for fit parameters. (d) Voltage-dependence of inactivation of αIIA coexpressed with β1 or β3. A two-step protocol was applied with a conditioning pulse of 500-ms duration from −110 mV to +10 mV in 5-mV increments, followed by a test pulse to −10 mV. Peak current amplitudes evoked by the test pulse were normalized to the maximum peak current amplitude and plotted as a function of the conditioning pulse potential. ▵, αIIA; ●, αIIA + β1; ○, αIIA + β3. Data were fitted with a two-state Boltzman equation: g = 1/{1 + exp[(V − V_1/2)/k]}, where g is conductance, V_1/2 is the voltage of half-maximal inactivation, and k is the slope factor. See Table 2 for fit parameters.