Inactivation of BK channels by the NH2 terminus of the beta2 auxiliary subunit: an essential role of a terminal peptide segment of three hydrophobic residues

Abstract

An auxiliary beta2 subunit, when coexpressed with Slo alpha subunits, produces inactivation of the resulting large-conductance, Ca(2+) and voltage-dependent K(+) (BK-type) channels. Inactivation is mediated by the cytosolic NH(2) terminus of the beta2 subunit. To understand the structural requirements for inactivation, we have done a mutational analysis of the role of the NH(2) terminus in the inactivation process. The beta2 NH(2) terminus contains 46 residues thought to be cytosolic to the first transmembrane segment (TM1). Here, we address two issues. First, we define the key segment of residues that mediates inactivation. Second, we examine the role of the linker between the inactivation segment and TM1. The results show that the critical determinant for inactivation is an initial segment of three amino acids (residues 2-4: FIW) after the initiation methionine. Deletions that scan positions from residue 5 through residue 36 alter inactivation, but do not abolish it. In contrast, deletion of FIW or combinations of point mutations within the FIW triplet abolish inactivation. Mutational analysis of the three initial residues argues that inactivation does not result from a well-defined structure formed by this epitope. Inactivation may be better explained by linear entry of the NH(2)-terminal peptide segment into the permeation pathway with residue hydrophobicity and size influencing the onset and recovery from inactivation. Examination of the ability of artificial, polymeric linkers to support inactivation suggests that a variety of amino acid sequences can serve as adequate linkers as long as they contain a minimum of 12 residues between the first transmembrane segment and the FIW triplet. Thus, neither a specific distribution of charge on the linker nor a specific structure in the linker is required to support the inactivation process.

SCHEME I

SCHEME II

Figure 1.

Sequence of NH₂ termini of the BK auxiliary β subunit family. (A) NH₂ termini of the four known auxiliary β subunits are shown. For the β3 subunit for which four alternatively spliced NH₂ termini have been identified (Uebele et al., 2000), the rapidly inactivating β3b variant is shown. TM1 designates the proposed beginning of the first TM segment. Positive and negative residues in the β2 subunit are in blue and red, respectively. Boxed sets of β2 residues (11–17 and 20–30) are thought to adopt relatively helical structures in an isolated peptide, while other portions of the NH₂ terminus are relatively disordered (Bentrop et al., 2001). (B) The initial 20 residues of several inactivating NH₂ termini are compared, showing the common theme of a hydrophobic segment at the NH₂ terminus and the downstream hydrophilic region.

Figure 2.

Deletions spanning positions 5–36 do not abolish inactivation. In A1, currents resulting from α subunits coexpressed with wild-type β2 subunits were activated by the indicated voltage protocol. In A2, currents were activated by a paired pulse protocol (activation steps to 100 mV) separated by steps of different duration to −140 mV. Currents during the initial activation step were truncated to allow better visualization of the recovery time course. In B, removal of Phe, Ile, and Trp in positions 2–4 (ΔFIW) results in removal of inactivation. In C1 and C2, currents arising from a β2 subunit with amino acids in positions 5–16 deleted (Δ5–16) are shown. The first 10 amino acids in this construct are MFIWEKRNIY. Note the steady-state current in this construct that may arise from the influence of charged residues in positions 5–7. In D1 and D2, currents arising from a construct with residues 16–25 deleted (Δ16–25) are shown. In E1 and E2, currents are shown for a construct with residues 27–36 deleted (Δ27–36). In F, the currents show that deletion of residues 5 through 35 (Δ5–35) results in removal of inactivation. In Δ5–35, the total length of the cytosolic portion of the NH₂ terminus is 14. In G, inactivation time constants (τ_on) for β2 (•, 4 patches), Δ5–16 (⋄, 3 patches), Δ16–25 (○, 4 patches), and Δ27–36 (♦, 4 patches) are plotted as a function of activation potential showing a similar weak voltage-dependence of τ_on for each construct. Each point is the mean and SD of 4–7 patches. In H, the recovery time course at −140 mV defined from the paired pulse protocol is shown for a set of patches for each construct. For β2 (•, 4 patches), the fitted τ_off is 23.4 ± 2.3 ms; for Δ5–16 (⋄, 3 patches), τ_off is 5.13 ± 0.19 ms; for Δ16–25 (○, 4 patches), τ_off is 9.30 ± 0.55 ms; for Δ27–36 (♦, 3 patches), τ_off is 6.19 ± 0.33 ms. Vertical calibration bar corresponds to: A1, 3 nA; A2, 2 nA; B, 6 nA; C1, 4 nA; C2, 3 nA; D1, 6 nA; D2, 4 nA; E1, 1.5 nA; E2, 1.2 nA; F, 8 nA.

Figure 3.

Deletions within positions 2–4 of the NH₂ terminus reduce and abolish inactivation. In A1, activation of wild-type α + β2 currents are illustrated while, in A2, the time course of recovery from inactivation at −140 mV is shown. In B1 and B2, currents resulting from a construct with deletion of Phe in position 2 (ΔF) are shown, along with the time course of recovery from inactivation for that construct. Note the appearance of some steady-state current at all activation potentials. In C1 and C2, currents resulting from a construct with deletion of Phe and Ile in positions 2 and 3 (ΔFI) are shown. More substantial steady-state current is observed along with more rapid recovery from inactivation. In D, τ_on is plotted as a function of activation potential for each of the three inactivating constructs (β2: •, 4 patches; ΔF: ○, 4 patches; ΔFI, ♦, 8 patches). In E, the time course of recovery from inactivation determined at −140 mV is illustrated for the three constructs. For β2, τ_off is given in Fig. 2; for ΔF (3 patches), τ_off is 4.5 ± 0.3 ms; for ΔFI (4 patches), τ_off is 2.99 ± 0.33 ms.

Figure 4.

Consequences of replacement or displacement of FIW residues with GGG. In A, currents arising from a construct in which residues FIW were replaced with GGG are shown. No inactivation is observed, and trypsin application resulted in no increase in outward current. In B, currents are shown for a construct in which GGG was appended to the initial FIW sequence. The apparent stability of inactivation is reduced, but inactivation still occurs. In C, currents are shown for a construct in which GGG was inserted between FIW and the remainder of the NH₂ terminus. In this case, steady-state inactivation is than for wild-type, but still substantial.

Figure 5.

Consequences of replacement of one or two residues in the FIW epitope. In A, inactivating currents resulting from wild-type β2 subunits are shown. In B–D, glycine was individually substituted for each residue in the FIW epitope. In each case, this resulted in a small weakening of inactivation, with the strongest effect arising from the F2G substitution. In E–G, two glycines were substituted for a pair of residues in the FIW epitope. In F, replacement of both F and W with G abolished inactivation, while the presence of a single F (E) or W (G) appears sufficient to maintain some fast inactivation. In H–K, the consequences of increasing the separation between F and W are illustrated. In H, the presence of two glycines between F and W results in currents similar to those with an FGW epitope, suggesting that W can still contribute to the stability of the inactivated state when there are two glycines interposed. In I and J, three and four glycines are interposed between F and W, in both cases resulting in currents in which steady-state inactivation is comparable to that resulting from FGG (E). This suggests that, in FGGGW (I) and FGGGGW (J), W may not substantially participate in defining the stability of the inactivated state. In K, the introduction of two negative charges between F and W abolishes inactivation.

Figure 6.

Introduction of single charges in the inactivation segment reduces but does not abolish rapid inactivation. In A–C, each residue in the inactivation segment was replaced individually with glutamate. Replacement of F with E (A) produced the most marked disruption of inactivation, with substantial steady-state current observed at all potentials. In D–F, the consequences of replacing each residue with arginine are illustrated. Arginine is less effective in each case at disrupting inactivation than glutamate, although at each position arginine produces some reduction in the stability of the inactivated state. Similar to the action of glutamate, replacement of F with R (D) had the strongest effects in disrupting inactivation. Vertical calibration: A, 4 nA; B, 1.5 nA; C, 5 nA; D, 6 nA; E, 5 nA; F, 4 nA.

Figure 7.

Relationship of inactivation parameters to alterations in the FIW triplet. In A, changes in τ_off relative to inactivation mediated by wild-type β2 subunits is expressed kT units. Mutations are grouped into those at position 2 (F2), those at position 3 (I3), those at position 4 (W4), constructs with deletions or multiple glycines in the NH₂ terminus, and then a set of repeated residues in the initial triplet (FFF, III, WWW). Error bars reflect standard errors for measurement of the mutant construct expressed relative to the mean β2 estimate. In B, changes in τ_on are shown for each construct. Except for the slowing in inactivation resulting from the WWW mutation, most mutations have minimal effects on inactivation onset. In C, effects of mutations are compared in terms of ln(K_mt/K_β2), which is calculated based on the steady-state current at 100 mV (f _ss) and τ_on (see materials and methods).

Figure 8.

Dependence of inactivation onset and recovery on properties of residues in the initial triplet. In A1, τ_on (on the left) and τ_off (on the right) is plotted as a function of the mean surface area of the residue in position 2 that would be buried on transfer from solvent to a folded protein (Rose et al., 1985). In A2, τ_off is plotted as a function of area what would be buried. The lines correspond to linear regressions [τ(area) = B*exp(C*area)] fit through only the uncharged residues. In B1 and B2, τ_on and τ_off are plotted with respect to area buried on transfer of a residue from solvent to a folded protein in regard to the residue in position 3 in the inactivation epitope. In C1 and C2, τ_on and τ_off are plotted with respect to area buried on transfer of a residue from solvent to a folded protein in regard to the residue in position 4 in the inactivation epitope.

Figure 9.

Dependence of inactivation parameters on bulk hydrophobicity within the initial triplet. In A, τ_off for constructs with mutations within the FIW triplet is plotted as a function of the mean surface area (for the three residues in positions 2–4) that would be buried on transfer from solvent to a folded protein (Rose et al., 1985). The solid line is a linear regression [τ_off(A) = 0.018 * exp(0.012*A)] for all constructs involving uncharged residues. Error bars are SD for a least three determinations. ♦, F2G, F2A, F2L; ▪, I3T, I3A, I3G; ▾, W4G, W4A, W4L; ▵, FGG, GGW; •, FFF, III, WWW; red ○, F2R, I3R, W4R; blue ⋄, F2E, I3E, W4E; red ▴, FWI, WIF, IWF. In B, τ_on is plotted as a function of area of residue buried on transfer to a folded protein. Symbols are as in A. The solid line was a fit to constructs with no charges in the initial triplet [τ_on(A) = 0.002 exp(0.016A) + 8.0].

Figure 10.

NH₂ termini with artificial polymeric linkers support inactivation of BK channels. In A and H, wild-type β2 currents are shown at two different time bases for comparison to mutant constructs. In B, an NH₂ terminus with a polymeric chain length of 30 glutamine residues separating FIW from R46 results in currents that exhibit inactivation. In C, an NH₂ terminus consisting solely of 30 glutamine residues (30Q) does not inactivate. In D, an NH₂ terminus with a polymeric chain length with 10 glutamine residues separating FIW from R46 results in currents that do not inactivate. In E, when the chain length reaches 12 residues, fast time-dependent block is observed at potentials positive to 140 mV, while at more moderate potentials the fast kinetics of block result in an apparent increase in current activation rate. In F and G, traces show inactivating currents resulting from linkers of 14 and 20 glutamine residues, as indicated. In H, wild-type β2 currents are shown on a different time base. In I, the NH₂ terminus contained a linker with 10 proline residues. In J, the linker contained 12 proline residues. In K, the linker contained 14 proline residues. In L, the linker contained 14 residues, an alternating sequence of 7 alanine-arginine pairs.

Figure 11.

Dependence of inactivation properties on lengths of artificial NH₂ termini and altered β2 NH₂ termini. In A, τ_on for poly-glutamine (Poly-Q) and poly-proline (poly-P) linkers is plotted as a function of linker length. In B, τ_off for poly-Q and poly-P linkers is plotted as a function of linker length. In C, ln(K*_mt/K*_β2) is plotted as a function of linker length. In D, τ_off is plotted as a function of the length of the β2 NH₂-terminal linker for various deletion constructs, as indicated. In E, τ_off is plotted as a function of β2 linker length for various deletion constructs. In F, ln(K*_mt/K*_β2) is plotted as a function of linker length for deletion constructs. Dotted lines correspond to values for the wild-type β2 construct.

Figure 12.

Mutations of charged residues have little impact on inactivation mediated by the β2 auxiliary subunit. For A–J, currents were activated by the indicated voltage-protocol, although in D longer activation steps were employed. In A, wild-type β2 currents are illustrated. In B, currents resulted from construct R8QR14QK18QR19Q. In C, currents are from construct R8Q R14QK18QR19QK24QR26QK35QK41Q; in D, D16RE17K; in E, neutralization of all charge in first 26 amino acids, R8QR14QK18QR19QK24QR26QD16NE17Q. In F, currents resulted from a construct with deletion of two of the residues in the inactivation epitope, FI (ΔFI). In G, currents resulted from mutation of R8QR14QK18QR19Q in a background of ΔFI; in H, ΔFI-R8QR14QK18QR19QK24QR26QK35QK41Q; in I, ΔFI-D16RE17K; in J, ΔFI-R8QR14QK18QR19QK24QR26QD16NE17Q. Vertical calibration: A, 3 nA; B, 0.6 nA; C, 1.5 nA; D, 1.5 nA; E, 2 nA; F, 2.5 nA; G, 1.3 nA; H, 5 nA; I, 2.5 nA; J, 1.5 nA.

Figure 13.

Dependence of inactivation properties on net charge in first 30 amino acids of NH₂ terminus. In A, τ_off is plotted as a function of net charge in the first 30 amino acids of the NH₂ terminus. Constructs in which residues K33, R34K35, and K41 were mutated were not included in these plots, since shifts in activation V_0.5 in these constructs resulted in shifts in τ_off because of coupling of inactivation to activation. The filled circle corresponds to the wild-type β2 NH₂ terminus. In B, τ_on is plotted as a function of net charge. With decreases in net charge, τ_on shows little change while increases in net charge result in a slowing of τ_on. In C, τ_off for constructs with a ΔFI background is plotted as a function of net charge, revealing little dependence of recovery from inactivation on net charge in the linker. In D, τ_on is plotted as a function of net charge for constructs with a ΔFI background.

Figure 14.

Effects of insertions in the β2 NH₂ terminus. A fourteen residue insert (6QSG6Q) was introduced into the β2 NH₂ terminus beginning at positions 9, 16, 27, 36, and 46. In A, inactivation onset for wild-type β2 currents is shown on the left for activation potentials from −100 through 140 mV. On the right, wild-type β2 currents resulting from a paired pulse protocol to define the time course of recovery from inactivation are shown. The duration of the initial inactivation pulse varied for different constructs to ensure that inactivation was essentially complete before the onset of a recovery interval (τ_on ∼22 ms; τ_off ∼24 ms). In B, inactivation onset (on the left) and recovery from inactivation for an NH₂ terminus with the 14 amino acid insert at position 9 (INS@9) are shown. Inactivation onset is slowed about fourfold (τ_on ∼183 ms), while recovery from inactivation (τ_off ∼34 ms) is only slightly affected. In C, currents resulted from an NH₂ terminus with the insert at position 16 (INS@16). Both inactivation onset (τ_on ∼174 ms) and recovery (τ_off ∼36.7 ms) are slowed relative to wild-type currents. In D, currents resulted from an insert at position 27 (INS@27), with both the onset (τ_on ∼26.2 ms) and recovery (τ_off ∼31 ms) from inactivation being similar to wild-type currents. In E, in construct INS@36, onset (τ_on ∼52 ms) and recovery (τ_off ∼17 ms) from inactivation are similar to wild-type currents. In F, in construct INS@46, a shift in the V_0.5 of activation is observed, along with an increase in recovery rate (τ_on ∼49 ms; τ_off ∼13.4 ms).