We studied non-esterified fatty acid-induced uncoupling of heterologously expressed rat UCP1 (uncoupling protein 1) in yeast mitochondria, as well as UCP1 in rat BAT (brown adipose tissue) mitochondria. The proton-conductance curves and the relationship between the ubiquinone reduction level and membrane potential were determined in non-phosphorylating BAT and yeast mitochondria. The ADP/O method was applied to determine the ADP phosphorylation rate and the relationship between the ubiquinone reduction level and respiration rate in yeast mitochondria. Our studies of the membranous ubiquinone reduction level in mitochondria demonstrate that activation of UCP1 leads to a purine nucleotide-sensitive decrease in the ubiquinone redox state. Results obtained for non-phosphorylating and phosphorylating mitochondria, as the endogenous ubiquinone redox state was gradually varied by a lowering rate of the ubiquinone-reducing or ubiquinol-oxidizing pathways, indicate that the endogenous ubiquinone redox state has no effect on non-esterified fatty acid-induced UCP1 activity in the absence of GTP, and can only regulate this activity through sensitivity to inhibition by the purine nucleotide. At a given oleic acid concentration, inhibition by GTP diminishes when ubiquinone is reduced sufficiently. The ubiquinone redox state-dependent alleviation of UCP1 inhibition by the purine nucleotide was observed at a high ubiquinone reduction level, when it exceeded 85–88%.
UCP1 (uncoupling protein 1) is a well-characterized member of the UCP subfamily [1–5]. UCP1 short-circuits the proton electrochemical gradient generated by the respiratory chain in BAT (brown adipose tissue) mitochondria. It is responsible for adaptive thermogenesis by conducting protons across the mitochondrial inner membrane that uncouples ATP production from respiration and leads to heat production. NEFA (non-esterified fatty acid)-activated PN (purine nucleotide)-inhibited proton conductance sustained by UCP1 greatly increases the BAT respiration rate in response to hormonal stimuli. In non-thermogenic animal and plant tissues, as well as in unicellular organisms, the physiological role of UCP1 homologues has not yet been well established [1–4,6,7].
Since UCPs are specialized proteins for proton electrochemical gradient dissipation, their activity must be finely regulated. Inhibition of proton conductance by PNs is considered diagnostic of UCP activity. In the case of mammalian UCPs (UCP1, UCP2 and UCP3), ubiquinone (coenzyme Q, Q) has been shown to be an obligatory cofactor for their action in liposomes and isolated mitochondria [8–11]. It has been postulated that ubiquinone (probably the reduced form) activates proton conductance in mitochondria through the production of superoxide . Products of lipid peroxidation by ROS (reactive oxygen species) such as 4-hydroxy-2-nonenal are proposed to be direct activators of UCPs [12,13]. However, in a reconstituted system with heterologously expressed mammalian UCPs, no superoxide activation has been required to demonstrate the NEFA-activated PN-sensitive proton translocation, and ubiquinone has no significant activating effect or any effect on inhibition by PNs [14,15]. Studies with mitochondria isolated from yeast mutants expressing mouse UCP1, but lacking ubiquinone, have also shown that Q is not required for proton conductance by UCP1 . Taking into account the apparent affinity of reconstituted UCPs for PNs [9,15] and the concentration of nucleotides in vivo (at millimolar concentrations inside the cells), UCPs should be permanently inhibited under in vivo conditions, even in the presence of NEFAs, unless a regulatory factor or mechanism could alleviate the inhibition by PNs . Therefore it has been proposed that the membranous Q redox state could be a metabolic sensor that modulates PN inhibition of NEFA-activated UCP1 homologues; such is the case in isolated skeletal muscle (UCP3), potato tuber (plant UCP) and Acanthamoeba castellanii (protozoan UCP) mitochondria respiring under phosphorylating conditions in the absence of endogenous superoxide production [17–19]. In the case of protozoan UCP, this regulatory mechanism has also recently been observed in isolated non-phosphorylating mitochondria . Activation of UCP1 in vivo requires interactions with NEFAs and their ability to overcome the inhibition of UCP1 by PNs. Therefore the question arises as to whether UCP1 sensitivity to PNs could be regulated by the redox state of membranous Q as observed in the case of UCP1 homologues.
The aim of the present study was to examine the influence of the endogenous Q redox state on PN inhibition of the NEFA-induced activity of heterologously expressed UCP1 in yeast mitochondria, as well as UCP1 in rat BAT mitochondria. We show, for the first time, that UCP1 inhibition by PN is under control of the endogenous Q redox state.
Expression of UCP1 in Saccharomyces cerevisiae
S. cerevisiae (INVSc1 strain) cells were transformed with either a rat UCP1 low-expressing construct or a plasmid containing an empty vector (pYES2.1) (Invitrogen) . The preculture was grown at 28 °C under vigorous aeration in SC (S. cerevisiae) minimal medium containing 0.67% (w/v) yeast nitrogen base, 0.5% (w/v) (NH4)2SO4, 0.005–0.01% (w/v) amino acids and 2% (w/v) glucose. Glucose repressed GAL1-promoted gene expression. Afterwards, cultures were grown in SC medium with 3% (v/v) glycerol for 48 h. Protein expression was induced in medium identical with that used for preculture, but glucose was replaced with 2% (w/v) galactose and 5% glycerol. For both wild-type (empty vector) and UCP1-expressing yeasts, the generation time in continuously agitated cultures was ~2.5 h. Cells cultured for 20 h in 1.8 litre flasks were collected in the exponential phase (with an absorbance of 1.2–1.8 at 650 nm).
UCP1 expression in mitochondria isolated from UCP1-containing cells was determined by Western blot analysis. A concentration of ~1 μg of UCP1 per mg of mitochondrial protein was detected, confirming previous results using similar methods .
Isolation of yeast mitochondria
Yeast cells (wild-type with empty vector and UCP1-expressing cells) were harvested by centrifugation at 1000 g for 10 min, re-suspended in deionized water, and centrifuged a second time. Pellets were re-suspended in a buffer containing 0.1 M Tris/H2SO4 (pH 9.4) and 10 mM DTT (dithiothreitol) and incubated at 28 °C in an orbital shaker (160 rev./min) for 15 min. Pellets were again centrifuged and washed in 1.2 M sorbitol. The cells were re-suspended in 6 ml of a buffer containing 1.2 M sorbitol and 20 mM K/K phosphate buffer (pH 7.4) per 1 g of cells. Zymolyase was added (at 1 mg/g of wet weight of cells), and the suspension was incubated at 28 °C under gentle agitation until approx. 90% of cells had converted into spheroplasts (~30–50 min). The digestion was stopped by the addition of an equal volume of ice-cold buffer containing 1.2 M sorbitol, 10 mM Tris/HCl (pH 6.8), 2 mM EGTA and 0.2% BSA. All subsequent steps were performed at 4 °C. Spheroplasts were pelleted, washed twice in the latter buffer, and re-suspended in an isolation buffer [0.65 mM mannitol, 10 mM Tris/HCl (pH 6.8), 0.5 mM EDTA, 0.1 mM EGTA and 0.2% BSA] at a ratio of 9 ml of buffer per 1 g of spheroplasts. After homogenization by ten passes with a tight Dounce homogenizer, homogenates were centrifuged at 1000 g for 10 min. The pellets were re-suspended and centrifuged again to collect mitochondria remaining in the pellet. The supernatants were combined and centrifuged at 1000 g for 10 min. The resultant supernatants were centrifuged at 10000 g for 10 min. The mitochondrial pellets were washed with a buffer containing 0.65 M mannitol and 10 mM Tris/HCl (pH 6.8) and centrifuged at 10000 g for 10 min. The final pellet was re-suspended in a small volume of the same buffer.
Isolation of rat BAT mitochondria
BAT mitochondria were isolated essentially as described in . Interscapular and subscapular BATs were taken from four 9–11-week-old male rats (250–350 g), housed at 24 °C and fed ad libitum. Animal experimentation was carried out according to appropriate Institutional guidelines. Subsequent steps were carried out at 4 °C. Tissue was placed in STE buffer [0.25 M sucrose, 1 mM EGTA and 5 mM Tris/HCl (pH 7.2)] supplemented with 5% defatted BSA, minced, and homogenized in STE+BSA buffer with six strokes of an electric glass homogenizer (at middle rotation, Teflon C-type piston). Then, the homogenate was filtered through two layers of gauze. The filtrate was centrifuged at 8500 g for 10 min and the pellet was re-suspended in STE buffer (supplemented with 1% BSA) and centrifuged at 700 g for 10 min, and the supernatant was then centrifuged at 8500 g for 10 min. The mitochondrial pellet was suspended in STE buffer, re-centrifuged and finally re-suspended in ~500 μl of STE buffer.
Mitochondrial oxygen consumption and membrane potential measurements
Oxygen uptake was measured polarographically using a Rank Bros oxygen electrode or a Hansatech oxygen electrode in 1.4 ml or 2.8 ml respectively, of standard incubation medium [100 mM KCl, 10 mM Tris/HCl (pH 7.0), 1 mM EGTA, 2 mM KH2PO4, 1 mM MgCl2 and 0.1% (w/v) defatted BSA (BAT mitochondria); or 0.65 M mannitol, 10 mM Tris/HCl (pH 6.8), 0.5 mM EGTA, 2 mM MgCl2, 10 mM with KH2PO4 and 0.05% (w/v) defatted BSA (yeast mitochondria)] with 0.85–1 mg of mitochondrial protein at 28 °C. Values of oxygen uptake are in nmol of oxygen/min per mg of protein. The membrane potential (ΔΨ) of mitochondria was measured simultaneously with oxygen uptake using a TPP+ (tetraphenylphosphonium)-specific electrode. After each run, 0.5 μM FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) was added to release TPP+ for baseline correction. For calculation of the ΔΨ value, the matrix volume of mitochondria was assumed to be 2.0 μl per mg of protein. The calculation assumes that TPP+ distribution between mitochondria and medium followed the Nernst equation. Corrections were made for the binding of TPP+ to mitochondrial membranes. Values of ΔΨ are presented in mV. Succinate was used at 7.5 mM succinate as a respiratory substrate in the presence of 4 μM rotenone (an inhibitor of Complex I). To induce UCP1 activity-mediated respiration, measurements were made in the presence of 40–43 μM OA (oleic acid). To inhibit UCP1 activity, 2 mM GTP was added.
The proton-conductance response to a driving force can be expressed as the relationship between the oxygen consumption rate and the ΔΨ (flux/force relationship) when varying the potential by titrating with respiratory-chain inhibitors. Respiration rate and the ΔΨ were measured simultaneously using electrodes sensitive to oxygen and TPP+ with isolated rat BAT mitochondria or yeast mitochondria. Proton-leak rates can be calculated from respiration rates by multiplying by an proton/oxygen ratio of 6. Measurements were performed in the absence of added ADP, i.e. in a non-phosphorylating state (resting state, State 4). To exclude the activity of an ATP/ADP antiporter, 1.8 μM carboxyatractylozide was used.
Respiratory rate, ΔΨ and the Q redox state were varied by modulating the Q-reducing or QH2 [reduced ubiquinone (ubiquinol)]oxidizing pathways. To decrease the rate of the Q-reducing pathway during State 4 respiration (thereby decreasing a steady-state resting respiration), titration of succinate dehydrogenase activity via increasing the concentration of malonate (a competitive inhibitor; up to 7 mM) was carried out. To decrease the rate of the QH2-oxidizing pathway during State 4 respiration, Complex IV was inhibited with cyanide (up to 20 or 50 μM in yeast or BAT mitochondria respectively). In order to avoid possible errors due to non-steady-state conditions (when sequential addition of inhibitors is applied), as well as to assess the Q redox state for a given steady state, data from separate measurements with different inhibitor concentrations (as in Figures 3 and 7) and data from measurements with sequential inhibitor additions were afterwards combined to generate common curves (Figures 1, 2 and 6). To assess the statistical significance of the shifts in leak curves caused by GTP, we generally compared respiration rates at the lowest common ΔΨ (Figures 2 and 6) for pairs of curves from three to seven independent experiments using a Student's t test for paired data.
ADP phosphorylation rate measurements
Phosphorylating state (State 3) measurements were performed with yeast mitochondria. The ADP/O ratio was determined by an ADP pulse method with 250–450 nmol of ADP. The measured ADP/O ratio and State 3 respiration rate (V3) were used to calculate the rate of ADP phosphorylation (Jp=V3×ADP/O). ATP (80 μM) was applied to activate succinate dehydrogenase.
To decrease the rate of the Q-reducing pathway during State 3 respiration (thereby decreasing steady-State 3 respiration), titration of succinate dehydrogenase activity with an increasing concentration of n-butylmalonate (0–5 mM), a competitive inhibitor of succinate uptake, was performed. To decrease the rate of the QH2-oxidizing pathway during State 3 respiration, the bc1 complex (Complex III) was inhibited by antimycin A (up to 30 ng/ml).
Measurements of the ubiquinone reduction level
The redox state of Q in steady-state respiration was analysed by an extraction technique followed by HPLC detection . A LiChrosorb RP-18 (10 μm) HPLC column or Polaris C18-A (5 μm) HPLC column were used for separation of Q9 (rat BAT mitochondria) or Q6 (yeast mitochondria) respectively. For calibration and quantification of the Q peaks, commercial coenzymes (Sigma) were used. The presented values of the redox state of Q (Qr/Qt) deal with the active ubiquinone pool (the difference between the total Q pool minus the inactive Q pool) in a given mitochondrial preparation .
Increasing inhibitory effect of GTP on NEFA-induced and non-NEFA-induced UCP1-sustained proton leak when the Q redox state is decreased in non-phosphorylating yeast mitochondria
Proton-leak kinetics of yeast mitochondria containing UCP1, defined as the respiration rate needed to drive proton leak as a function of the driving force (ΔΨ), has been described previously [16,22,23–25]. Figure 1 shows the leak kinetics of yeast mitochondria with expression of modest amounts of rat UCP1 from plasmid pYES2.1. UCP1 expression did not greatly change the proton conductance, which is reflected by the relatively small increase in proton-leak rate at any given ΔΨ (no OA, no GTP and no Q-reducing pathway inhibitor) in mitochondria from yeast expressing UCP1 (Figure 1A) compared with mitochondria from yeast transformed with empty vector (Figure 1B). UCP1 was almost inactive in yeast mitochondria under conditions of no added OA (○), since GTP did not decrease the proton-leak rate during progressive inhibition of succinate oxidation by malonate (●) leading to a decrease in Q redox state from 81% to ~46% (Figure 1A). In contrast, GTP had a large inhibitory effect on the proton-leak rate through OA-activated UCP1 during malonate titration (leading to a decrease in the Q redox state up to ~43%) (▲). The above experiments indicate that rat UCP1 in our yeast mitochondria was competent and required addition of NEFAs before large GTP-sensitive proton-leak rates could be observed.
Kinetics of proton leak in non-phosphorylating yeast mitochondria: effect of GTP when the Q-reducing pathway is decreased
In the presence of OA, mitochondria from yeast transformed with empty vector revealed a slight proton conductance (lower proton leak at any ΔΨ) compared with those from UCP1-transformed cells (Figure 1B). Among tested NEFAs (linoleic, myristic, oleic and palmitic acids), OA led to the smallest non-UCP1-mediated mitochondrial uncoupling (results not shown). A NEFA-induced mitochondrial uncoupling effect has been previously observed in yeast mitochondria [21,23]. However, in empty vector mitochondria, OA-induced uncoupling is not sensitive to GTP during titration with malonate (decreasing the Q reduction level up to ~57%) (Figure 1B).
Decreasing inhibitory effect of GTP on NEFA-induced UCP1sustained proton leak when the Q redox state is increased in non-phosphorylating yeast mitochondria
To elucidate the role of respiratory rate and ΔΨ, we gradually decreased these with inhibitors of the QH2-oxidizing pathway (the cytochrome pathway) leading to an increase in the Q reduction level. Figure 2 shows proton-conductance curves (Figures 2A and 2C) and the relationship between Q reduction level compared with ΔΨ (Figures 2B and 2D) in the presence or absence of 40 μM OA and/or 2 mM GTP, during titration with increasing concentrations of cyanide. Comparison of Figures 1(A) and 1(B) and 2(A) and 2(C) indicates that yeast mitochondria (lacking or containing UCP1) energized with succinate and titrated with inhibitors of the Q-reducing or QH2-oxidizing pathways exhibited the same proton-leak curves in the absence of GTP, with or without NEFAs. This clearly indicates that the redox state of endogenous Q does not affect the basal or NEFA-induced proton conductance in both types of yeast mitochondria.
The effect of GTP when the QH2-oxidizing pathway is decreased in non-phosphorylating yeast mitochondria
In empty vector mitochondria during titration with cyanide, OA-induced uncoupling, which caused a slight upward replacement of the proton-leak curve (Figure 2C) and a slight downward replacement of the Q redox state versus ΔΨ (Figure 2D), was not sensitive to GTP.
In UCP1-containing yeast mitochondria, when the rate of the QH2-oxidizing pathway was gradually decreased with cyanide, the flux/force relationship established in the presence of 40 μM OA and in the presence or absence of 2 mM GTP indicated that inhibition of the OA-induced proton conductance by the nucleotide progressively diminished below a ΔΨ value of approx. 180 mV (Figure 2A). Below this value, points obtained in the presence of OA and GTP (▼) progressively came forward to the points obtained in the presence of OA only (∇). For mitochondria titrated by cyanide (Figure 2A) at the lowest common ΔΨ (168 mV), the respiration rate in the presence of GTP and OA [26.2±1.9 nmol of oxygen/min per mg of protein (mean±S.D.)] was significantly higher (P<0.01, Student's paired t test) from the control (no GTP, no OA) rate [14.5±2.4 nmol of oxygen/min per mg of protein (mean±S.D.)] in contrast with mitochondria titrated by malonate (Figure 1A). Transition of the GTP inhibitory effect is also revealed by the relationship between the Q reduction level and ΔΨ (Figure 2B). When the Q reduction level was gradually increased, the inhibition by GTP was progressively relieved when points obtained with OA and GTP reached ~88% of the Q redox state (at ~180 mV). A further decrease in succinate oxidation with a higher cyanide concentration led to a transition of points obtained with GTP and OA (▼) from the control relationship (no OA, no GTP) (□) closer to an OA-induced linear relationship (∇). Although titration of succinate oxidation with cyanide (Figures 2A and 2B) comprised of the respiratory rate and ΔΨ ranges at which GTP-sensitivity did not change during titration of the Q-reducing pathway (with malonate, Figures 1A), progressive cancellation of the inhibitory effect of GTP was revealed under these conditions. These results clearly indicate that the transition (attenuation) of the GTP inhibitory effect on OA-induced proton leak cannot be attributed to changes in the ΔΨ or respiratory rate.
Figure 3(A) shows an example experiment in which the respiratory rate, ΔΨ and the Q redox state were measured in UCP1-containing yeast mitochondria, concomitant with inhibition of succinate oxidation in the presence of OA by malonate or cyanide to ~60%, leading to the same level of ΔΨ (166–165 mV). The corresponding Q redox state was decreased from 72% to 46% by malonate or increased from 72% to 91% by cyanide. After subsequent addition of GTP, inhibition of the OA-induced proton leak was revealed by inhibition of the respiratory rate as well as restoration of ΔΨ and the Q redox state. However, the inhibitory effect of GTP was much stronger when the nucleotide was added to mitochondria that had been inhibited by malonate (leading to 55% of the Q redox state) compared with those inhibited by cyanide (leading to 94% of the Q redox state). This stronger inhibition by GTP is revealed by a larger effect of the nucleotide on ΔΨ and the Q redox state in the presence of malonate. These results indicate that the efficiency of GTP to inhibit UCP1-sustained uncoupling in non-phosphorylating yeast mitochondria depends on the endogenous Q redox state. Inhibition by GTP can be diminished when Q is reduced sufficiently.
The effect of GTP on the change in respiration, membrane potential, and Q redox state caused by OA-induced proton leak (UCP1 activity) when the Q reduction level is varied in UCP1-containing yeast mitochondria
Effect of GTP in phosphorylating yeast mitochondria when the QH2-oxidizing or Q-reducing pathways are decreased
In order to check whether the dependence of UCP1 sensitivity to GTP on the endogenous Q redox state also occurs in phosphorylating yeast mitochondria containing UCP1, we measured State 3 respiratory rates as well as the ADP/O ratio in the absence or presence of 40 μM OA, with or without 2 mM GTP (Figure 4). As shown in Figure 4, State 4 respiration and ΔΨ were perturbed (increased and decreased respectively) by OA, whereas State 3 respiration and ΔΨ remained unaffected. The respiratory control and ADP/O ratio were lowered by OA, indicating an induction of the proton leak. Both coupling parameters were recovered to control values when the measurement was performed in the presence of GTP. These results suggest that the OA-induced GTP-inhibited proton leak mediated by UCP1 expressed in yeast diverts energy from ATP synthesis during phosphorylating respiration.
The effect of GTP on coupling parameters of UCP1-containing yeast mitochondria
The NEFA-induced proton leak can be analysed in State 3 by applying the ADP/O method to calculate the rate of ADP phosphorylation when State 3 respiration, within the range where the ΔΨ remains constant, is titrated with various inhibitors [17–19]. In order to study the GTP sensitivity of the OA-induced proton leak, the rate of Q-reducing and QH2-oxidizing pathways were gradually lessened by n-butylmalonate or antimycin A respectively (Figure 5). Pair measurements of ADP/O ratios and State 3 respiration in the presence or absence of 40 μM OA, with or without 2 mM GTP, were performed for decreasing State 3 respiration. The titration range of State 3 was such that the ΔΨ remained constant [165±2 mV (mean±S.D., n=12)] in the absence or presence of OA (with or without GTP) for UCP1-containing yeast mitochondria. A constant ADP/O ratio [1.32±0.04 (mean±S.D., n=36)] was observed under all control conditions, i.e. when the Q-reducing pathway was titrated with n-butylmalonate either in the absence of OA (with or without GTP) or in the presence of both OA and GTP, as well as when the QH2-oxidizing pathway was titrated with antimycin A in the absence of OA (with or without GTP) (Figure 5A). In UCP1-containing yeast mitochondria, in the presence of OA alone, the measured ADP/O ratio was lowered to 1.1±0.07 (mean±S.D., n=9) and then declined with a decreasing State 3. Thus lowering the electron flux amplified the OA-induced decrease in the ADP/O ratio.
The effect of GTP in phosphorylating yeast mitochondria when the QH2-oxidizing or Q-reducing pathways are decreased
Titration of State 3 respiration led to a linear relationship between the rate of ADP phosphorylation (Jp=V3×ADP/O) and corresponding State 3 respiration (V3) (Figure 5B) in all control conditions (see above) where the ADP/O constancy was found (Figure 5A). Under these conditions, the straight line intersects the x-axis close to the origin, indicating that no basal proton leak occurs during State 3 respiration in UCP1-expressing yeast mitochondria. However, addition of OA during n-butylmalonate or antimycin A titration (no GTP) provoked a shift of the linear relationship between ADP phosphorylation rate and State 3 to the right. The respiration sustained by the OA-induced proton leak (no GTP) was 31.5±3.4 nmol of oxygen/min per mg of protein (mean±S.D., n=10).
As shown in Figure 5(C), the Q redox state followed a linear relationship with decreasing State 3 respiration. Titration with n-butylmalonate decreased the Q reduction level (from ~65% to ~28%), whereas antimycin A increased the Q reduction level (from ~65% to ~95%). When the Q-reducing pathway activity was inhibited (with n-butylmalonate), the ADP/O ratios (Figure 5A) and the relationship between the ADP phosphorylation rate versus State 3 respiration (Figure 5B), measured in the presence of OA and GTP (when compared with plus OA, no GTP conditions) revealed an inhibitory effect of GTP on the OA-induced proton leak (points are on the control lines), which thereby can be attributed to the UCP1 activity. Thus at 65% of the Q redox state (not titrated State 3) and below this value (State 3 titrated with n-butylmalonate), the full inhibitory effect of GTP is observed. On the other hand, with the decreasing QH2-oxidizing pathway activity (i.e. with the increasing Q reduction level), a transition of GTP inhibitory effect was observed. Namely, with increasing concentration of antimycin A (in the presence of OA and GTP), a progressive abatement of GTP inhibitory effect was revealed by (i) a non-linear relationship between the rate of ADP phosphorylation and State 3 respiration (Figure 5B, ▼) that transitioned from a control relationship (the straight line coming from the origin) to an OA-induced linear relationship (∇), and (ii) the ADP/O ratios (Figure 5A, ▼) that progressively dropped to the ADP/O values obtained in the presence of OA only (no GTP) (∇). A weakening of the GTP inhibitory effect was observed for State 3 respiration below ~132 nmol of oxygen/min per mg of protein, corresponding to a Q redox state higher than ~85%. This set of observations suggests that in phosphorylating UCP1-expressing yeast mitochondria, OA induces a UCP1-mediated proton leak and inhibition of this leak by GTP is controlled by the endogenous Q redox state. Taking into account results obtained for non-phosphorylating (Figure 2) and phosphorylating (Figure 5) UCP1-expressing yeast mitochondria, it can be concluded that changes in the GTP inhibitory effect on OA-induced UCP1-mediated uncoupling are observed for the same range of Q reduction levels. For both mitochondrial energetic states, full inhibition is observed below a Q redox state of 85–88%, whereas a progressive abatement of GTP inhibition occurs above this value.
In phosphorylating yeast mitochondria containing an empty vector, no OA-induced proton leak was observed, as neither the ADP/O ratio nor the linear relationship between ADP phosphorylation rate and State 3 respiration were changed in the presence of fatty acid (no GTP) (Figures 5D and 5E respectively). This confirms that the OA-induced GTP-sensitive proton leak observed in UCP1-expressing yeast mitochondria can be attributed to the UCP1 activity.
The Q redox state influences UCP1 sensitivity to GTP in mitochondria isolated from rat BAT
To investigate whether the ability of GTP to inhibit UCP1-mediated proton conductance depends on the membranous Q redox state in its native environment, the effects of a varied Q reduction level on the endogenous (observed under applied conditions) or external OA-induced proton leaks' sensitivity to GTP were assessed in non-phosphorylating mitochondria isolated from rat BAT. It is well known that mitochondria isolated from BAT tissue are essentially uncoupled, but on addition of PNs, respiration rates due to the UCP1-mediated proton leak are inhibited. Therefore BAT mitochondria are not a good material with which to study UCP1 action under phosphorylating conditions; for these studies we applied a heterologous yeast expression system (Figures 4 and 5) with modest amounts of rat UCP1 .
Figure 6 shows proton-conductance curves (Figure 6A) and the relationship between Q reduction level versus ΔΨ (Figure 6B) in the presence or absence of 43 μM OA and/or 2 mM GTP, when the QH2-oxidizing or Q-reducing pathways were gradually decreased by increasing concentrations of cyanide or malonate respectively. Rat BAT mitochondria energized with succinate and titrated with both inhibitors yielded the same proton-leak curves (no GTP, with or without OA) (Figure 6A). Thus the redox state of membranous Q does not affect the endogenous or OA-induced proton conductance of BAT mitochondria. In the absence of OA (no GTP), the initial Q reduction level (~65%) progressively increased up to 83% with cyanide (□), but decreased up to 54% with malonate (○) (Figure 6B). In the presence of 43 μM OA (no GTP), the initial Q reduction level (~60%) progressively increased up to 80% with cyanide (∇), yet progressively decreased up to 56% with malonate (Δ) (Figure 6B).
The effect of GTP when the QH2-oxidizing or Q-reducing pathways are decreased in non-phosphorylating rat BAT mitochondria
Figure 6(A) shows that the proton conductance of BAT mitochondria, both endogenous and external OA-induced, exhibits a different sensitivity to GTP when the QH2-oxidizing or Q-reducing pathways were inhibited. During titration with cyanide, the flux/force relationship established in the presence of GTP (with or without OA) indicated that inhibition of proton conductance by the nucleotide progressively diminished at a ΔΨ value below 183–184 mV (a lower slope is observed), in contrast with titration with malonate. At the lowest common ΔΨ value in the presence of GTP (178 mV), the respiration rate was significantly higher (P<0.05, Student's paired t test) for mitochondria titrated by cyanide [22.5±2.5 or 20.3±2 nmol of oxygen/min per mg of protein (mean±S.D.), without or with OA respectively] than by malonate [9.4±1.5 or 12±2 nmol of oxygen/min per mg of protein (mean±S.D.), without or with OA respectively]. Transition of the GTP inhibitory effect is also disclosed by the relationship between the Q reduction level and ΔΨ (Figure 6B). In the presence of GTP (with or without OA), increasing the concentration of malonate decreased the Q redox state from ~81% to ~76–74%. However, when the QH2-oxidizing pathway was gradually inhibited with cyanide, the Q reduction level increased to a maximal value of ~86% (corresponding to a ΔΨ value below 183 mV). A further increase in cyanide concentration led to no further increase in the Q reduction level which remained above 82%. These results indicate that the ability of GTP to inhibit UCP1-sustained uncoupling in non-phosphorylating BAT mitochondria is under control of the endogenous Q redox state.
In order to avoid possible errors due to non-steady-state conditions in the case of results from measurements with sequential inhibitor additions that are afterwards pooled together on common curves, measurements with a single inhibitor addition were performed as shown in Figure 7. This experimental approach clearly validates the observed phenomenon. Figures 7(A) and 7(C) show example experiments in which the respiratory rate, ΔΨ and the Q redox state were measured in rat BAT mitochondria under conditions where succinate oxidation in the absence (Figure 7A) or presence (Figure 7C) of 43 μM OA was inhibited by a single addition of malonate or cyanide resulting in the same value of ΔΨ (129–130 mV). The corresponding Q redox state was decreased to 56% by malonate (both in the absence or presence of OA) or increased to 82% or 75% by cyanide (in the absence or presence of OA respectively). After a subsequent addition of GTP, inhibition of the non-OA-induced or OA-induced proton leak was revealed by inhibition of the respiratory rate, as well as restoration of ΔΨ and the Q redox state. However, for both endogenous and OA-induced proton leaks, the inhibitory effect of GTP was much stronger when the nucleotide was added to mitochondria inhibited by malonate compared with those inhibited by cyanide, although the respiratory rate and ΔΨ before GTP addition were the same. This is revealed by the greater GTP effect on ΔΨ and the Q redox state in the presence of malonate. Thus the ability of GTP to inhibit UCP1-mediated proton conductance seems to depend on the endogenous Q redox state in BAT mitochondria, just as it does in yeast mitochondria. Inhibition by GTP can be diminished when Q is reduced sufficiently.
The inhibition of endogenous or OA-induced UCP1 activity when the Q reduction level is varied in non-phosphorylating rat BAT mitochondria: effect of GTP on respiration, membrane potential, and the Q redox state
The results of the present study obtained with non-phosphorylating and phosphorylating UCP1-containing yeast mitochondria, as well as with non-phosphorylating rat BAT mitochondria clearly indicate that UCP1 sensitivity to GTP can be regulated by the redox state of membranous Q, as has been observed previously for UCP1 homologues in isolated mitochondria of rat skeletal muscle, potato tubers and A. castellanii [17–20]. This indicates a likely universal mechanism through which UCPs are regulated. It must be emphasized that, in the present study, in non-phosphorylating UCP1-expressing yeast mitochondria and BAT mitochondria, at a given concentration of OA (or at no addition of external NEFA), the Q redox state-dependent alleviation of PN-mediated inhibition of UCP1 is observed for a similar range of Q reduction levels as that observed for phosphorylating yeast mitochondria. Thus for both mitochondrial energetic states and both types of mitochondria, full inhibition is observed below a Q redox state of ~85–88%. This observation, which was drawn from studies with the application of two different experimental approaches, strongly indicates that ubiquinone (probably QH2) affects the affinity of UCP1 for GTP.
Studies with mammalian UCPs have yielded conflicting results concerning the possibility that Q may be an obligatory cofactor for their action. Namely, oxidized ubiquinone has been shown to activate PN-sensitive NEFA-dependent proton transport through reconstituted UCP1-3 [8,9]. Photoaffinity labelling of purified UCP1 with retinoic acid has indicated that Q increases binding of an activator . On the other hand, other studies have shown that Q has no significant activating effect on NEFA-dependent proton translocation, or any effect on the inhibition by PN in reconstituted UCP1–3 . In isolated kidney mitochondria energized with succinate, titration with myxothiazol or cyanide (but not malonate) results in increased proton conductance only when external oxidized Q is added . From these flux–force relationship studies, it has been proposed that external Q activates proton conductance in mitochondria through the production of superoxide, as superoxide dismutase inhibits Q-induced mitochondrial uncoupling. However, it has been concluded that the redox state of endogenous Q does not affect mitochondrial conductance. Therefore the amount of endogenous ubiquinone reduced during titration with myxothiazol or cyanide is not enough to induce superoxide-stimulated proton leak until exogenous Q is added . Our results suggest that the endogenous Q redox state has no effect on the basal and NEFA-induced UCP1-catalysed proton conductance in the absence of PNs, but affects its sensitivity to inhibition by nucleotides. This conclusion can be made as determination of membranous Q reduction level was performed when proton-leak curves (non-phosphorylating mitochondria) or relationships between ADP phosphorylation rate and State 3 respiration rate (phosphorylating mitochondria) were established with inhibitors of the Q-reducing or QH2-oxidizing pathways.
It should be emphasized that the described regulation of UCPs by the endogenous Q redox state has so far been observed only for the NEFA-induced UCP-mediated uncoupling. Indeed, from proton-conductance curves established in different mitochondria under conditions when UCPs are activated by superoxide during non-phosphorylating respiration sustained by succinate oxidation [9,10,12,26–29], it is difficult to unequivocally estimate whether sensitivity to PN depends on the endogenous Q redox state during titrations of the Q-reducing pathway (with malonate) or the QH2-oxidizing pathway (with cyanide or myxothaizol). However, this cannot be excluded since determinations of Q reduction level have not been performed in these experiments. Therefore further studies are needed to explain whether regulation of the sensitivity of UCP1 and its homologues to PNs through the Q redox state can also be observed under conditions when uncoupling is activated by superoxide.
The current model for the activation of UCPs by superoxide through initiation of lipid peroxidation [12,13] assumes that superoxide generated within mitochondria and a high membranous Q reduction level (as required for superoxide formation) work indirectly as UCP activators by generating carbon-centred radicals of polyunsaturated fatty acid chains of phospholipids in the mitochondrial inner membrane. However, in our opinion, these indirect effects could be a late response of UCPs, as flux–force studies with isolated mitochondria do not reveal any effect of endogenously generated superoxide or the endogenous Q redox state on basal or NEFA-induced activity of UCP1 or its homologues in the absence of PNs. To observe indirect activation of UCPs by superoxide (or a high Q redox state), lipid peroxidation products or an exogenous system that generates superoxide (xanthine plus xanthine oxidase) must be applied. Our results obtained with UCP1-containing yeast and BAT mitochondria, as well as those previously described for UCP1 homologues [17–20], indicate that the quick response through the endogenous Q redox state could directly regulate UCP activity. Thus, in our model for activation of UCPs, a high endogenous Q reduction level activates UCP by relieving inhibition from PNs, and this quick response does not involve superoxide formation resulting in lipid peroxidation products.
Taking into account the concentration of nucleotides in cells (in millimolar concentrations), UCPs should be permanently inhibited in vivo, even in the presence of NEFAs, unless a regulatory factor could overcome the inhibition by PN [2,17]. In the case of UCP1, it has been proposed that NEFAs do this in a kinetically simple competitive manner . Our present results suggest that, at a given NEFA concentration, alleviation of PN inhibition is dependent on the endogenous Q redox state. Studies with mitochondria isolated from yeast mutant cells lacking Q and expressing mouse UCP1 have shown that NEFA-induced GDP-sensitive proton conductance by UCP1 expressed in yeast mitochondria is not dependent on the presence of Q in the mitochondrial membrane . However, there are no published results (to our knowledge) that conflicts with the observations described in the present study that a highly reduced membranous Q leads to relief of PN inhibition of UCP1 activity. The present study suggests that, in BAT mitochondria, the endogenous Q redox state is an additional metabolic sensor (beside NEFA content) which modulates PN inhibition of UCP1 activity. Certainly, further studies are necessary to elucidate the kinetic mechanism of such regulation. However, a model describing function of reduced ubiquinone (QH2) in modulating PN inhibition of UCP1 can be proposed (Figure 8). At a given NEFA concentration, an increased amount of QH2 could lead to decreased binding affinity of PN thereby alleviating inhibition of UCP1. QH2 may play a role of a negative regulator of PN binding to UCP1. Conversely, at a lower QH2 amount, PN may be bound to UCP1 and proton conductance through UCP1 is inhibited.
A tentative model for the regulation of UCP1 by the membranous Q redox state
We thank Gregory Mathy (University of Liege, Liege, Belgium) for providing UCP1-expressing yeast cells.
brown adipose tissue
carbonyl cyanide p-trifluoromethoxyphenylhydrazone
non-esterified fatty acid
- QH2 (or Qr)
total endogenous pool of Q in the inner mitochondrial membrane
reduction level of Q
- SC medium
Saccharomyces cerevisiae medium
Aleksandra Swida-Barteczka performed the experiments. Andrzej Woyda-Ploszczyca performed the experiments. Francis Sluse provided scientific expertise. Wieslawa Jarmuszkiewicz provided scientific guidance and edited the manuscript prior to submission.
This work was supported by the Polish Ministry of Education and Science [grant numbers 3382/B/P01/2007/33, 0505/B/P01/2009/36].