Tubulin binding blocks mitochondrial voltage-dependent anion channel and regulates respiration

Abstract

Regulation of mitochondrial outer membrane (MOM) permeability has dual importance: in normal metabolite and energy exchange between mitochondria and cytoplasm and thus in control of respiration, and in apoptosis by release of apoptogenic factors into the cytosol. However, the mechanism of this regulation, dependent on the voltage-dependent anion channel (VDAC), the major channel of MOM, remains controversial. A long-standing puzzle is that in permeabilized cells, adenine nucleotide translocase (ANT) is less accessible to cytosolic ADP than in isolated mitochondria. We solve this puzzle by finding a missing player in the regulation of MOM permeability: the cytoskeletal protein tubulin. We show that nanomolar concentrations of dimeric tubulin induce voltage-sensitive reversible closure of VDAC reconstituted into planar phospholipid membranes. Tubulin strikingly increases VDAC voltage sensitivity and at physiological salt conditions could induce VDAC closure at <10 mV transmembrane potentials. Experiments with isolated mitochondria confirm these findings. Tubulin added to isolated mitochondria decreases ADP availability to ANT, partially restoring the low MOM permeability (high apparent K(m) for ADP) found in permeabilized cells. Our findings suggest a previously unknown mechanism of regulation of mitochondrial energetics, governed by VDAC and tubulin at the mitochondria-cytosol interface. This tubulin-VDAC interaction requires tubulin anionic C-terminal tail (CTT) peptides. The significance of this interaction may be reflected in the evolutionary conservation of length and anionic charge in CTT throughout eukaryotes, despite wide changes in the exact sequence. Additionally, tubulins that have lost significant length or anionic character are only found in cells that do not have mitochondria.

Fig. 1.

Tubulin at nanomolar concentrations induces reversible partial blockage of VDAC. (A) Representative current record through a single channel before and after addition of 50 nM tubulin to the cis side at −25 mV applied voltage. Addition of tubulin induces time-resolved rapid events of partial channel blockages to a single well-defined conductance level, shown in the Inset at a finer scale. Here and elsewhere, dashed lines indicate zero-current level; current records were filtered by using averaging times of 10 ms (except for 1 ms in Inset). (B) Typical VDAC voltage gating in tubulin-free solution under periodically applied −50 mV voltage impulses. Under applied voltage, channel conductance moves from a single high-conducting open state to various low-conducting closed states. Relaxing the voltage to 0 mV reopens the channel. The dashed-and-dotted line indicates the conductance of the channel in its open state and dotted lines in its closed states. (C) Tubulin induces a single well-resolved closed state of VDAC, whereas in control voltage-induced gating a wide variety of closed states is observed. The distribution of conductances in control was collected from a series of −50 mV impulses of 30 s duration interrupted by 70 s periods of zero applied voltage to restore the open state. (D) Statistical analysis of the tubulin-induced closures of VDAC performed by logarithmic exponential fitting of the open times (τ_on) and closed times (τ_off). It is seen that open time is described by single exponent with characteristic time τ_on = 198.7 ± 3.5 ms. Closed-time histogram could be fitted by at least 2 exponents with characteristic times of τ_off⁽¹⁾ = 2.4 ± 0.1 ms and τ_off⁽²⁾ = 30.4 ± 5.7 ms. The time histograms were collected from the current traces presented in A. Bilayer membranes were formed from the mixture of asolectin and cholesterol (10:1 wt/wt). VDAC was isolated from rat liver mitochondria. The medium consisted of 1 M KCl buffered with 5 mM Hepes at pH 7.4.

Fig. 2.

The binding parameters for tubulin-induced VDAC closure depend on tubulin concentration, electrolyte concentration, and applied voltage. (A) (a and d) VDAC open time between successive blockages, τ_on, linearly decreases with tubulin concentration and depends on salt concentration. The medium consisted of 1 M KCl (a, filled circles) and 0.25 M KCl (d, filled diamonds). Both components of tubulin residence (closed) time, τ_off⁽¹⁾and τ_off⁽²⁾, are independent of tubulin or salt concentration (b, c, e, and f, open symbols) and are equal to 2.8 ± 0.5 ms and 23.7 ± 6.3 ms, respectively. The applied voltage was −20 mV. (B) Voltage dependence of VDAC residence times, τ_off⁽¹⁾ and τ_off⁽²⁾, in the presence of 50 nM tubulin in cis (open diamonds), trans (open circles), or both sides (filled squares) of the membrane. Residence time in extrapolation to 0 voltage does not depend on 1-side (cis or trans) or 2-side tubulin addition. (C) Voltage dependence of the on-rate, κ_on, with 50 nM tubulin added to the cis side. The line is an exponential fit to κ_on = κ₀exp(nVF/RT) with n = 4.97. Each time value presents the characteristic time of 9 different log probability fitting procedures ± SE. The medium consisted of 1 M KCl (B and C) and buffered with 5 mM Hepes at pH 7.4. VDAC was isolated from N. crassa mitochondria. Bilayer membranes were formed from DPhPC.

Fig. 3.

Tubulin interaction with VDAC requires the presence of C-terminal tails of tubulin. Tubulin (50 nM) with truncated CTT, tubulin-S (A), or a mixture of 10 μM 2 synthetic peptides of mammalian α- and β-brain tubulin CTT (B) does not induce channel blockage characteristic for intact tubulin (10 nM), with fast reversible blockage to 1 closed state (C). Representative current traces through single VDAC were obtained at ±25 mV of applied voltage in 1 M KCl solutions buffered with 5 mM Hepes at pH 7.4. Tubulin, tubulin-S, and CTT peptides were added to both sides of the membrane. Other experimental conditions were as in Fig. 2 B and C.

Fig. 4.

Model of tubulin–VDAC interaction. One tubulin CTT partially blocks channel conductance by entering VDAC pore. This process is voltage-dependent and could be described by the 1st-order reaction of one-to-one binding of tubulin to VDAC. Some additional interaction between tubulin globular body and VDAC may be involved. The model of a tubulin dimer was redrawn from ref. .

Fig. 5.

Tubulin dramatically increases apparent K_m for ADP in regulation of respiration of isolated brain mitochondria. Shown are double-reciprocal representations of the respiration kinetics of brain mitochondria activated by ADP in control (A) and in the presence of 1 μM tubulin (B). The 2 straight lines represent 2 different respiration kinetics in the presence of tubulin (B). (Insets) Enlargements of A and B. Each data point is a mean of 6–9 independent experiments ±SE.

Fig. 6.

Analysis of tubulin CTT sequences. (A) Sequence alignments of β-tubulin CTT. Full sequences were aligned, but only the CTT, defined as the residues from C-terminal to the last residue in the crystal structure, are shown. Acidic residues are shown in red, basic residues in blue. More extensive alignments of both α-and β-CTT are in Fig. S4. (B) Summary of length and charge values for α- and β-CTT, taken from our alignments and from the data in ref. . Each value is a mean ± SD; n, number of species.