Native UCP1 displays simple competitive kinetics between the regulators purine nucleotides and fatty acids.

Elucidation of the regulation of uncoupling protein 1 (UCP1) activity in its native environment, i.e. the inner membrane of brown-fat mitochondria, has been hampered by the presence of UCP1-independent, quantitatively unresolved effects of investigated regulators on the brown-fat mitochondria themselves. Here we have utilized the availability of UCP1-ablated mice to dissect UCP1-dependent and UCP1-independent effects of regulators. Using a complex-I-linked substrate (pyruvate), we found that UCP1 can mediate a 4-fold increase in thermogenesis when stimulated with the classical positive regulator fatty acids (oleate). After demonstrating that the fatty acids act in their free form, we found that UCP1 increased fatty acid sensitivity approximately 30-fold (as compared with the 1.5-fold increase reported earlier based on nominal fatty acid values). By identifying the UCP1-mediated fraction of the response, we could conclude that the interaction between purine nucleotides (GDP) and fatty acids (oleate) unexpectedly displayed simple competitive kinetics. In GDP-inhibited mitochondria, oleate apparently acted as an activator. However, only a model in which UCP1 is inherently active (i.e."activating" fatty acids cannot be included in the model), where GDP functions as an inhibitor with a K(m) of 0.05 mm, and where oleate functions as a competitive antagonist for the GDP effect (with a K(i) of 5 nm) can fit all of the experimental data. We conclude that, when examined in its native environment, UCP1 functions as a proton (equivalent) carrier in the absence of exogenous or endogenous fatty acids.

Uncoupling protein 1 (UCP1) 1 is essential for the phenomenon of classical nonshivering thermogenesis (1,2). When chronically exposed to cold, animals that can recruit UCP1 in their brown-adipose tissue will initially shiver but will then successively increase their capacity for nonshivering thermogenesis and ultimately cease shivering. Mice that lack UCP1 (3) never cease shivering in the cold (4). Thus, no other process or protein can replace UCP1 in its physiological thermogenic function.
Brown-fat cells from animals that can express UCP1 respond to stimulation with the physiological activator norepinephrine with at least a 5-fold increase in the rate of oxygen consumption (5,6) and thus in the rate of thermogenesis (7). In contrast, there is no detectable effect of norepinephrine on heat production in brown-fat cells that lack UCP1 (6). Thus, UCP1 is essential for the ability of brown-fat cells to produce heat. Exogenous fatty acids can also stimulate respiration to the same 5-fold extent as can norepinephrine in normal brown-fat cells (5,6,8). In brown-fat cells that lack UCP1, fatty acids are without effect (6). The responses to fatty acids are of particular interest, because fatty acids are classically considered as positive modulators of UCP1 activity (9 -11) and it is generally accepted that it is via an increase in endogenous fatty acids that norepinephrine stimulates thermogenesis (2).
In brown-fat mitochondria, the situation is less clear. Undoubtedly, the high "uncoupled" rate of respiration that is found in isolated brown-fat mitochondria that contain UCP1 is not found in brown-fat mitochondria without UCP1 (12,13). Fatty acids can induce uncoupling in all mitochondria (14) and not just UCP1-containing brown-fat mitochondria, but the presence of UCP1 should nonetheless be expected to induce dramatic effects in the fatty acid sensitivity of brown-fat mitochondria. However, the reported effects of UCP1 on the fatty acid sensitivity of brown-fat mitochondria have been minor (12,15,16), making it difficult to reconcile the qualitative and dramatic effects of fatty acids on brown-fat cells and the suggested mediatory role of fatty acids in norepinephrine stimulation of the cells with the apparent relative insensitivity of isolated brown-fat mitochondria to fatty acids. For example, when fatty acid-induced de-energization was investigated, the reported difference between brown-fat mitochondria with and without UCP1 was only ϳ1.5-fold (but with a 50-fold difference between brown-fat mitochondria without UCP1 and liver mitochondria) (12,15,16). Also, when thermogenesis was assessed as the fatty acid-induced increase in the rate of oxygen consumption, the total increase caused by fatty acid corresponded only to an increase of 25-75% compared with the 5-fold increase seen in brown-fat cells, and the difference in sensitivity to added fatty acids was also small (approximately a factor 2) (12,15,16). Thus, the data so far published comparing brown-fat mitochondria with and without UCP1 are not adequate to explain the role of fatty acids as a regulator of UCP1. In this context, it is noteworthy that recently a series of alternative UCP1 activators have been proposed, including retinoic acid (17), ubiquinone (18), superoxide (19), and 4-hydroxy-2-nonenal (20). Considering the relatively small effects of fatty acids on UCP1 reported so far, the possibility could be raised that these substances would better fill the role as physiological activators of UCP1.
However, in order to distinguish between regulatory and oxidative effects of fatty acids, all of the experiments published to date on effects of fatty acids on mitochondria from UCP1ablated animals have been performed with glycerol 3-phosphate as substrate. Because thermogenesis is dependent on brown-fat mitochondria respiring on complex I-linked sub-strates present in the matrix (mainly fatty acid derivatives), we have here examined whether our understanding of UCP1 function could be enhanced if a complex I-linked substrate was used in the investigations.

MATERIALS AND METHODS
Animals-UCP1-ablated mice, originally on a mixed 129/SvPas and C57Bl/6 background (progeny of those described previously (3)), were backcrossed to C57Bl/6 for 10 generations. These mice were then intercrossed to obtain two mouse strains: UCP1(Ϫ/Ϫ) and UCP1(ϩ/ϩ). Southern blot analysis of DNA with the short-arm probes (3) was used for genotyping of the mice. In the brown-adipose tissue of the UCP1(Ϫ/Ϫ) mouse strain finally obtained, no UCP1 was detected with polyclonal antibodies (data not shown). Because no effect of genetic background on parameters used for our analysis was evident, principally in agreement with results reported previously (15), some of the initial data obtained with the standard (UCP1(ϩ/ϩ)) C57Bl/6 strain and with UCP1(Ϫ/Ϫ) on the originally mixed (C57Bl/6 and 129/SvPas) genetic background were included in the data presented here. The mice were fed ad libitum (R70 standard diet, Lactamin), had free access to water, and were kept on a 12-h light/12-h dark cycle. No differences in normal growth were detected between the UCP1(Ϫ/Ϫ) and UCP1(ϩ/ϩ) strains at normal (24°C) animal house temperature. Adult (8 -12-week-old) male mice were used for the experiments. The experiments were approved by the Animal Ethics Committee of the North Stockholm region.
Mitochondrial Preparation-Brown-fat mitochondria were prepared principally as described previously (21) with some modifications. Routinely, on each experimental day, five wild type and/or 5 UCP1-ablated mice were anesthetized for 1-2 min by a mixture of 79% CO 2 and 21% O 2 and decapitated. The interscapular, periaortic, axillary, and cervical brown-adipose tissue depots were dissected out and cleaned from white adipose tissue, nerves, and connective tissue and pooled in ice-cold 250 mM sucrose solution. Throughout the isolation process, the tissue was kept at 0 -4°C. The tissue was minced with scissors, homogenized in 40 ml of 250 mM sucrose solution, filtered through gauze, and centrifuged at 8500 ϫ g for 10 min. The supernatant with the floating fat layer was discarded. The pellet was resuspended in 250 mM sucrose containing 0.2% fatty acid-free bovine serum albumin and centrifuged at 800 ϫ g for 10 min. The resulting supernatant was then centrifuged at 8500 ϫ g for 10 min, and the resulting mitochondrial pellet was resuspended in 125 mM sucrose. Mitochondrial protein concentration was measured using the fluorescamine method, and the suspensions were diluted to stock concentrations of 25 mg of mitochondrial protein/ml of 125 mM sucrose with 0.2% fatty acid-free bovine serum albumin. The mitochondrial suspensions were kept on ice and used for no more than 4 h.
Oxygen Consumption-Isolated mitochondria, at a final concentration of 0.5 mg of mitochondrial protein/ml, were added to 1.1 ml of a continuously stirred incubation medium consisting of 125 mM sucrose, 20 mM K ϩ -TES (pH 7.2), 2 mM MgCl 2 , 1 mM EDTA, 0.1% fatty acid-free bovine serum albumin, 4 mM KPi, 3 mM malate, and 1.3 g of oligomycin/ml. Oxygen consumption rates were monitored with a Clark-type oxygen electrode (Yellow Springs Instrument Co.) in a sealed chamber at 37°C. The output signal from the oxygen electrode amplifier was electronically time-differentiated and collected every 0.5 s by a Power-Lab/ADInstrument (application program Chart, version 4.1.1.). The Chart data files were transferred to the KaleidaGraph MacIntosh application and converted to absolute values based on an oxygen content of 217 nmol of O 2 /1 ml of water on the calibrated electronic differentiation constants and on the amount of mitochondrial protein used. For calculation of stable oxygen consumption rates, the mean values of ϳ1 min were obtained from these recordings.
Swelling-Mitochondrial swelling was monitored as the change in absorbance at 540 nm in an Aminco DW-2 UV-visible spectrophotometer. The mitochondrial incubation medium and the experimental conditions were those described for oxygen consumption experiments. The substrate was 5 mM pyruvate.
Data Analysis-The free concentration of oleate was calculated using the equation used by Richieri et al. (22) for the binding of oleate to bovine serum albumin at 37°C: [free fatty acid] (nM) ϭ 6.5n Ϫ 0.19 ϩ 0.13 exp(1.54n), where n is the molar ratio of oleate to albumin. Routinely, the albumin concentration in the incubation medium was 0.1% (i.e. 16 Because only the free purine nucleotide (ATP, ADP, GDP, or GTP) (i.e. not the Mg 2ϩ -nucleotide complex) is relevant for nucleotide binding to UCP1 (23)(24)(25), the values for free GDP were used in graphs and for calculation of kinetic parameters. Free GDP and Mg 2ϩ concentrations in the incubation medium at 37°C and pH 7.2 were calculated by use of a computer program for making buffers of a defined ion concentration, BAD 4 (26). The nominal levels of GDP added were 0.7, 1.0, 2.2, 3.5, and 4.8 mM. These levels transferred to the free levels of 0.24, 0.37, 0.78, 1.25, and 1.72 mM. The nominal concentration of Mg 2ϩ was 2 mM. This concentration transfers to a free concentration of 0.23 mM in the presence of all of the medium components and 1 mM GDP. The concentration of free Mg 2ϩ was adjusted to 0.2-0.3 mM in all of the experimental traces with different GDP concentrations by adding a calculated amount of MgCl 2 to the incubation medium.
Concentration-response curve data were analyzed with the general fit option of the KaleidaGraph application for MacIntosh for adherence to simple Michaelis-Menten kinetics, V(x) ϭ basal ϩ V max ϫ (x/(K m ϩ x)). To analyze the effect of albumin, Michaelis-Menten kinetics with a Hill coefficient of 2 were used for fitting, V(x) ϭ basal Ϫ V max ϫ (x 2 /(K m 2 ϩ x 2 )), in accordance with two of fatty acid-binding sites on albumin having nearly similar and high affinity (22).
Statistics-All of the data are expressed as the means Ϯ S.E. Statistical analysis for the comparison of two groups was performed using Student's t test.

RESULTS
Relative Significance of UCP1 Depends on Oxidative Substrate Investigated-In earlier investigations of the UCP1-dependent components of thermogenesis in isolated brown-fat mitochondria, glycerol 3-phosphate was used as substrate (12,13,15,16). Although the evident effects of the presence of UCP1 were seen, they were quantitatively relatively small, especially when the effects of fatty acids were investigated.
There were very significant UCP1-dependent effects when the complex I-linked substrate, pyruvate, was used instead (Fig. 1). 2 In Fig. 1A, it is seen that in brown-fat mitochondria without UCP1 (thin trace), the basal respiration was low and the addition of GDP was without effect but respiration could be markedly increased by FCCP. In brown-fat mitochondria with UCP1 (heavy trace), the initial rate of respiration was high and it was reduced 3-4-fold by GDP. However, the sensitivity of the mitochondria to FCCP was unaffected by the presence or absence of UCP1 (Fig. 1, A and B).
Significance of UCP1 for Oleate-induced Respiration-As seen in Fig. 1C, oleate could stimulate pyruvate respiration in brown-fat mitochondria without UCP1 (thin trace), but the effect of oleate occurred at lower oleate concentrations and was larger in brown-fat mitochondria with UCP1 (heavy trace). From the compilation in Fig. 1D, it is evident that the presence of UCP1 led to a higher sensitivity to added oleate: from an EC 50 of 91 Ϯ 5 M to an EC 50 of 56 Ϯ 2 M (p Յ 0.05), i.e. a 2-fold increase. However, it was also clear that, in the mitochondria with UCP1, fatty acids (oleate) could induce an increase in the rate of oxygen consumption that was almost 4-fold, i.e. a relative increase close to the fatty acid-induced increase in respiration seen in UCP1containing brown-fat cells (6). In comparison, the fatty acidinduced increase in oxygen consumption when glycerol 3-phosphate was used as substrate was only ϳ30% (Fig. 1E), in agreement with earlier observations (12,15,16).
No Inhibitory UCP1-dependent Effect of Fatty Acids-The maximum respiratory capacity of the mitochondria was clearly independent of the presence of UCP1 (Fig. 1B). However, the ability of oleate to stimulate pyruvate respiration was greater in UCP1-containing mitochondria than in mitochondria without UCP1 (Fig. 1D), which may be due to an inhibitory effect of high oleate concentrations. Therefore, we examined whether high oleate concentrations would affect the maximal rate of oxygen consumption observed after the addition of FCCP (as exemplified in Fig. 2A). As seen in mitochondria without UCP1 (Fig. 2B), the oleate concentrations necessary to induce full stimulation (Ͼ 80 M) had an inhibitory effect on pyruvate oxidation. However, in brown-fat mitochondria with UCP1, a full restimulation was observed before any detrimental effect of oleate on pyruvate oxidation was evident (Fig. 2C).
Oleate-induced Restimulation/Stimulation of Respiration Is Not Caused by Mitochondrial Disintegration-It may be envisaged that the restimulating/stimulating ("uncoupling") effect of oleate was due to damage to the mitochondrial membrane associated with swelling of the mitochondria. UCP1 could influence the sensitivity of the mitochondria in this respect. Therefore, oleate-induced swelling in brown-fat mitochondria was analyzed as shown on Fig. 3A. No effect on mitochondrial volume was observed by the addition of nominal oleate amounts of up to Ϸ100 M, i.e. the concentration range used in the experiments. The minimal concentration of oleate capable of inducing mitochondrial swelling was 120 M and was independent of UCP1 (Fig. 3B). Similar concentration-response curves were observed for liver mitochondria (data not shown). Thus, fatty acid-induced mitochondrial swelling occurred in brown-fat mitochondria independently of UCP1.
Free Fatty Acids Are the Functional Species-Although a higher oleate sensitivity was observed in UCP1-containing brown-fat mitochondria than in brown-fat mitochondria without UCP1 (Fig. 1D), the increased sensitivity was small (ϳ2fold) when expressed as the nominal addition of oleate. To examine whether it was the nominal ("added") amount of oleate, the oleate/mitochondrial protein ratio, or the "free" oleate concentration that restimulated UCP1, we examined conditions that allowed the distinction among these possibilities, i.e. we altered the amount of albumin in the incubation medium.
In brown-fat mitochondria without UCP1, the response to a given nominal oleate addition was markedly affected by the amount of albumin present (Fig. 4A). When recalculated to free oleate concentrations according to Richieri et al. (22), the curves approached each other but they did not fully coincide (Fig. 4B). Also, in brown-fat mitochondria with UCP1 ( Fig. 4C), albumin concentration affected the oleate-induced respiration, but when these oleate concentrations were recalculated to free oleate levels, a clear coincidence of the responses was observed. Thus, it is the free oleate concentration, rather than the amount added or the oleate/protein ratio, that affects UCP1 activity (a similar conclusion was recently reached by Rial et al. (27) from investigations of wild-type mitochondria).
Concentration-response curves for brown-fat mitochondria without or with UCP1 based on calculated free oleate concentrations are shown in Fig. 4E. When analyzed in this way, the EC 50 of oleate for stimulation of respiration in brown-fat mitochondria without UCP1 was as high as 533 Ϯ 35 nM (n ϭ 6). However, in brown-fat mitochondria with UCP1, the EC 50 for restimulation was as low as 39 Ϯ 6 nM (n ϭ 6). Thus, the presence of UCP1 increased the oleate sensitivity 13-fold. Even this value may be considered an underestimate. The EC 50 in the UCP1-deficient mitochondria is influenced by the respiratory inhibition occurring at high oleate concentrations (Fig.  2B). If instead the shift in the amount of oleate that causes an equal amount of stimulation (e.g. 50 nmol of O 2 ) is read from Fig. 4E, the increased sensitivity is found to be 30-fold.
It is evident from Fig. 4E that, at a given concentration of oleate, the observed respiration will be a composite of contributions from oleate-stimulated UCP1-dependent and oleatestimulated UCP1-independent processes. However, for each oleate concentration, it is possible by subtraction to estimate the fraction that is truly UCP1-dependent. This UCP1-dependent oleate-re-induced respiration is depicted in Fig. 4F. This possibility to quantitate the exact value of the UCP1-dependent oleate effect is essential for the analysis of the control of UCP1 activity.
Is the Activating Fatty Acid Trapped by UCP1?-Some models of UCP1 function imply that activating fatty acids in some way associate with the protein (28), implicitly in such a way that they are no longer in immediate equilibrium with the free fatty acids of the medium. To examine the nature of the fatty acid/UCP1 interaction, we performed experiments as those shown in Fig. 5, A and B. In brown-fat mitochondria without UCP1, an addition of 70 M oleate led, as expected, to an increase in respiration (Fig. 5A); however, upon a further addition of albumin, the induced oxygen consumption reverted to the basal level, i.e. the fatty acid effect was fully reversible. A similar experiment performed with UCP1-containing brownfat mitochondria appeared principally different (Fig. 5B). As expected from Fig. 1D, 70 M oleate induced a much larger increase in the rate of oxygen consumption in these mitochondria. However, the same amount of albumin that fully reverted respiration in brown-fat mitochondria without UCP1 was now unable to suppress the oleate-induced respiration more than marginally (Fig. 5B). Even successive additions of albumin were not able to fully revert the respiration to basal values.
In Fig. 5C, the results of this type are plotted as oleateinduced increases in respiration as a function of increasing albumin concentrations. From this plot, a different result is apparent in the two types of mitochondria. Irrespective of oleate concentration, the addition of albumin could fully revert respiration to basal in brown-fat mitochondria without UCP1; however, when UCP1 was present, the activating fatty acid would seem to have escaped the chelating effect of albumin. This gives the impression that oleate was associated in an irreversible way with UCP1, supporting the hypotheses of an integrated role of certain fatty acids in UCP1 function. However, a further analysis revealed that this explanation need not be evoked. In Fig. 5D, the induced increase in oxygen consumption is plotted as a function of the calculated free oleate concentration. Superimposed on these lines are curves based on experiments as those in Fig. 4E with different initial albumin concentrations. As seen, these data coincide reasonably. Thus, the apparent inability of added albumin to reverse oleate-induced respiration is rather a reflection of the increased fatty acid sensitivity of UCP1containing mitochondria, and the experiments of the type depicted in Fig. 5B do not indicate that (some of) the activating fatty acids associate in an irreversible way with UCP1.
Interaction between Fatty Acid and GDP-That both fatty acids and purine nucleotides affect brown-fat mitochondria is well established (1), but both of these agents may also have UCP1-independent effects. The present system permits the identification of the true UCP1-dependent component. In brown-fat mitochondria without UCP1, the dose-response curve for oleate stimulation of oxygen consumption was slightly left-shifted when the amount of GDP was increased (Fig. 6A); a similar effect was seen when oxygen consumption was stimulated with increasing amounts of FCCP (Fig. 6B). However, in brown-fat mitochondria with UCP1, the dose-response curve for oleate was markedly influenced by the GDP concentration. The apparent K m was increased 3-fold from 38 Ϯ 5 nM to 102 Ϯ 12 nM in the presence of 3 mM rather than 1 mM GDP. Therefore, the shift in concentration causing equal stimulation (e.g. 40 nmol of O 2 ) was even higher, ϳ12-fold (Fig. 6E). The interaction between oleate and GDP was nucleotide-specific in that CDP did not influence the oleate sensitivity at all (Fig. 6C), in agreement with the expectations of a specific purine nucleotide interaction with UCP1 but no interaction with pyrimidine nucleotides (29). The interaction was specific for oleate and UCP1, because the dose-response curve for FCCP in mitochondria with UCP1 was only affected by GDP (or CDP) (Fig. 6, D and F) to the same extent as were the curves for mitochondria without UCP1 (Fig. 6B).
Kinetics of the Interaction between Fatty Acid and GDP-The analysis above indicated that the true quantitative interaction between purine nucleotides and fatty acids could be revealed when the UCP1-dependent fraction of the responses was iden-tified. Therefore, we examined the interaction between fatty acids and purine nucleotides in detail in this system to increase the understanding of the regulation of UCP1 activity (and its mode of function).
Until now, at least three formulations have been forwarded, all indicating independent effects of purine nucleotides and fatty acids on UCP1 as reviewed by Nedergaard et al. (1). In one formulation (30), which may be referred to as the allosteric formulation, purine nucleotides and free fatty acids interact independently with UCP1, interacting negatively and positively, respectively, at independent sites. The topology of the purine nucleotide-binding site is fairly well established (25) but that of the fatty acids is not, although some attempts have been made (31).
In a second formulation, which may be referred to as the cofactor formulation (28), the negative effect of GDP is mediated through a regulatory site. However, the positive effect of free fatty acid is not considered to be that of an activator but rather a cofactor. The fatty acid is necessary for function, probably by allowing proton transfer, and may be located within the UCP1 itself (implicitly in a tightly bound state, cf. the analysis of Fig. 5). Here, the expected kinetics would show independent purine nucleotide and fatty acid effects. Until the cofactor sites are filled, the activity would be fatty acid-dependent, but then the inhibition would be only GDP-dependent.
In a third formulation, which may be referred to as the cycling formulation (32,33), the fatty acid in its protonated form passes passively over the mitochondrial membrane and is returned in its anionic form to the cytosol through UCP1 (leading de facto to H ϩ reentry into the mitochondria). In this formulation, the expected kinetics are not simple but it is not expected that purine nucleotides and fatty acids would interact directly. Thus, in none of these formulations is any simple kinetic interaction between fatty acids and purine nucleotides expected.
To examine whether the possibility to delineate UCP1 activity in the absence of effects on non-UCP1 components could reveal new information concerning the fatty acid/purine nucleotide interaction, oleate concentration-response curves were obtained in the presence of different GDP concentrations. Nom-  Fig. 1D)). F, the UCP1-dependent part of the oleate-induced stimulation of oxygen consumption. Points were obtained by subtracting the corresponding values in E. B and D-F, concentration-response curves were drawn for best fit to Michaelis-Menten kinetics with a Hill coefficient of 1.
inally added amounts were 0.7, 1.0, 2.2, 3.5, and 4.8 mM GDP. Concentrations of GDP of Ͻ0.7 mM could not be used, because the initial spontaneous respiration then was not inhibited (cf. Fig. 1). As it has been established that it is the free purine nucleotides (and not their Mg 2ϩ complexes) that interact with UCP1 (23)(24)(25), the corresponding free GDP concentrations were calculated. As Mg 2ϩ may interact, for example, with oleate (34), the solutions were adjusted so that free Mg 2ϩ concentrations remained similar (0.2-0.3 mM). The experiments were performed with brown-fat mitochondria with and without UCP1, and the difference in responses was calculated and considered as the UCP1-dependent fraction (cf. Figs. 4F  and 6E). In the resulting curves, the oleate-induced UCP1-dependent respiration is plotted for different concentrations of the inhibitor GDP in its free form (Fig. 7A). Since respiration is already fully stimulated in the absence of GDP (Fig. 1), a curve for zero-free GDP cannot be constructed.
It is evident from Fig. 7A that there was an interaction between GDP and oleate concentrations. Therefore, we undertook a classical kinetic analysis of the interaction. When the data were plotted in the form of a Lineweaver-Burk double reciprocal plot (Fig. 7B), it was evident that the interaction could be convincingly described as competitive. At an infinitely high oleate concentration, the same rate of oxygen consumption was reached irrespective of GDP concentration; this rate was 175 nmol of O 2 /min/mg. This apparent competitive interaction was very unexpected based on the prevailing formulations concerning fatty acid and purine nucleotide effects on UCP1, for- mulations that do not predict a competitive interaction as summarized above.
Because it is not possible to obtain a curve in the absence of inhibitory GDP, the K m value for oleate could not be directly determined from the double reciprocal plot. Therefore, a secondary replot of the slopes of the inhibitor curves versus inhibitor concentration was constructed. This yielded a straight line that extrapolated at zero GDP to K m /V max , giving a K m value for oleate of 5 nM (Fig. 7C). Thus, this extrapolation allows the determination of the kinetics of oleate stimulation in the absence of GDP, i.e. the experiment that cannot be performed with these mitochondria.
The K i value was obtained from a corresponding secondary replot of 1/V against inhibitor concentration (Fig. 7D). This secondary replot has an intercept of ϪK i on the axis of inhibitor concentration. Thus, the K i was determined to be to 0.05 mM GDP.
From this analysis, it is evident that the interaction among oleate, GDP, and UCP1 in brown-fat mitochondria under these conditions, i.e. in GDP-inhibited mitochondria, can be described by Equation 1, described by this equation. Therefore, the equation does not fully describe the respiratory rate of brown-fat mitochondria in experiments because that rate also includes the UCP1-independent effects of fatty acids and purine nucleotides not included in the present model. Fig. 7, yielding simple competitive kinetics, were all obtained from experiments with mitochondria where the initial UCP1 activity had been fully inhibited by GDP. We further examined whether a similar influence of GDP on oleate sensitivity would be observable when the amount of GDP added was not sufficient to fully inhibit the inherent respiration. In the experiment depicted in Fig. 8A, a comparison was made between the response to oleate in brown-fat mitochondria inhibited with 0.53 mM free GDP and with only 0.19 mM free GDP. As seen, at the lower GDP concentration, the mitochondria were more sensitive to oleate (as would be expected from the competitive kinetics established above). However, these mitochondria were partially uncoupled, i.e. maintained a lower membrane potential, and the possibility existed that the higher sensitivity could be secondary to the lower membrane potential. Therefore, we inhibited mitochondria with 0.53 mM GDP but then added a titrated amount of FCCP sufficient to increase respiration (and thus lower the membrane potential) to the level observed in Fig. 8A with 0.19 mM GDP. If it were the membrane potential that influenced the sensitivity to oleate in this situation, the mitochondria should have become more sensitive to oleate; however, the response to oleate was identical to that seen in Fig. 8A with this amount of GDP. Thus, the interaction does not have the mitochondrial membrane potential as an intermediate.

Interaction in Partially GDP-inhibited Mitochondria Is Not Membrane Potential-mediated-The data analyzed in
Are Free Fatty Acids Necessary for UCP1 Activation-The above analysis was performed to analyze the ability of oleate to overcome GDP-induced inhibition of respiration, i.e. it represents a restimulation of already inhibited UCP1 function. However (as is also evident from Fig. 1), brown-fat mitochondria when freshly isolated are inherently uncoupled and this uncoupling can be overcome by the addition of GDP. The kinetics of GDP inhibition of the initial UCP1 activity is examined in Fig.  9A. The curve in Fig. 9A is drawn for simple Michaelis-Menten kinetics of GDP inhibiting the respiration, yielding an EC 50 value for GDP of 0.056 Ϯ 0.003 mM (n ϭ 3) (a value remarkably close to the K i value calculated above (0.05 mM) from the interaction between GDP and oleate).
A reasonable question to ask is whether this "initial" inter-action between GDP and brown-fat mitochondria can be understood using the same kinetic model developed in Fig. 7 for the ability of oleate to overcome GDP inhibition or whether we have to postulate another model for the initial preinhibited state.
It is evident that, in a model of this type (which we will refer to as Model I) where UCP1 is considered inherently inactive and is only active in the presence of fatty acids, respiration is a function of oleate concentration. Thus, if we enter a value of 0 oleate into Equation 1, we necessarily obtain a straight line as depicted at the bottom of Fig. 9B. Without activator, no respiration is predicted.
This limitation could be overcome if it is assumed that a small amount of free fatty acid is associated with the brown-fat mitochondria, even in the presence of albumin. Indeed, the experiments in Fig. 5 imply that it may be difficult with albumin to diminish free fatty acid levels below ϳ5 nM. Therefore, we attempted to fit Equation 1 to the experimental points in Fig. 9A by assuming the presence of different amounts of endogenous fatty acids. By assuming the presence of ϳ6 nM of fatty acid (stipled line in Fig. 9B), we could approximate some of the fit. However, there was a clear problem. In the absence of GDP, the predicted respiration in the presence of 6 nM oleate was only a fraction of what it should have been (Fig. 9B). Of course, it is possible to assume the presence of such a high level of endogenous fatty acid that it would stimulate respiration to the level normally observed in native mitochondria. It turns out that 40 nM oleate would be sufficient for respiration. However, the predicted effect of GDP then fell far from the data observed (Fig. 9B). Thus, it was not possible to reconcile the model that explains the interaction between GDP and oleate after GDP addition with the observations concerning the addition of GDP in the absence of exogenous oleate.
We then tested a model, Model II, where UCP1 is assumed to be inherently in its fully active state. For this model, we then rearranged Equation 1 so that the roles of GDP and fatty acids were reversibly formulated, i.e. GDP is functioning as an inhibitor and oleate has no effect in itself but is a competitive antagonist for GDP action as shown in Equation 2. When this model was used to predict the outcome of the GDP titrations, it was found that a very good fit was obtained (Fig.  9C). As this model also predicts the behavior of UCP1 after full GDP inhibition (Fig. 7), it has the advantage that it can in itself describe all of the UCP1-dependent interactions between fatty acids and purine nucleotides in brown-fat mitochondria. Therefore, we favor this model. It will be understood that the ability to describe UCP1 activity in the simple kinetic form developed here based on the analysis of brown-fat mitochondria from UCP1-ablated and UCP1-containing mice also has implications concerning the mode of action of UCP1. DISCUSSION In the present investigation, we have utilized the availability of UCP1-ablated mice to examine the kinetic characteristics of UCP1 in its native environment, i.e. in the brown-fat mitochondria. This allowed a clear delineation between those properties of brown-fat mitochondria that are truly UCP1-related and those that are inherent to brown-fat mitochondria irrespective of whether UCP1 is present or not. With this system, we were unexpectedly able to demonstrate that the interaction between oleate and GDP (i.e. between fatty acids and purine nucleotides) displayed simple competitive kinetics. These studies may be helpful both for the understanding of the regulation of UCP1 activity and of the mode of action of UCP1.
The Significance of UCP1 for the Oleate Sensitivity of Brownfat Mitochondria-Certain clear effects of the presence of UCP1 in brown-fat mitochondria have been reported earlier. As anticipated, the presence of UCP1 makes isolated brown-fat mitochondria inherently uncoupled and causes them to display a GDP-sensitive "basal" proton conductance (12,13). However, studies of the effect of the presence of UCP1 in brown-fat mitochondria have to date indicated only minor effects both with respect to sensitivity and magnitude of stimulation by fatty acids (12,15,16). These studies utilized glycerol 3-phosphate as substrate to eliminate any risk for oxidation of the added fatty acids during the investigation. However, basal respiration is higher with this substrate, limiting the magnitude of stimulation, and its dehydrogenase is inhibited by fatty acids (35,36). Thus, here we show that provided an adequate substrate is utilized (pyruvate), the presence of UCP1 allows oleate to stimulate respiration as much as 4-fold. Since we also established that it is in their free form that the fatty acids are (in equilibrium with) the activating species, we found that the presence of UCP1 increased the sensitivity to (free) oleate ϳ30-fold. Thus, UCP1 can indeed be shown in isolated mitochondria to have properties allowing it to mediate fatty acidinduced uncoupling and thus thermogenesis.
Are Free Fatty Acids the Activating Species Physiologically?-Although the sensitivity of brown-fat mitochondria to fatty acid uncoupling is increased by the presence of UCP1, the sensitivity (inherently ϳ5 nM) is not of the order of magnitude that would be expected. Thus, under the conditions used here, as much as 100 nM free oleate outside the mitochondria (i.e. in situ in the cytosol) was needed for full stimulation of UCP1. What fatty acid concentration would really be needed in cells is difficult to estimate, simply because of the competitive interaction demonstrated here between purine nucleotides and fatty acids. The free GDP concentration used here was routinely 0.3 mM. We do not have data for the free concentrations of ATP ϩ ADP ϩ GTP ϩ GDP in brown adipocytes, but as total purine nucleotide concentrations are normally estimated to be several millimolars, the concentration of free purine nucleotides probably exceeds 0.3 mM. It would indeed be difficult to assume that the level of free purine nucleotides would be lower than ϳ0.5 mM, i.e. a concentration that fully inhibits inherent UCP1 activity (Fig. 9A), and in this case, the free fatty acid level required to enable full UCP1 activity must be well over 100 nM. In reality, the total Mg 2ϩ concentration in the cell would also determine the level of free purine nucleotides. It is a consequence of the kinetics described here that FIG. 9. Examination of two models for regulation of UCP1 activity. A, brown-fat mitochondria from UCP1(ϩ/ϩ) mice were examined as in Fig. 1C with the exception that the initial GDP addition was varied as indicated. Points are from three independent mitochondrial preparations. Curve is drawn for simple Michaelis-Menten kinetics, yielding an EC 50 for GDP inhibition of 0.06 mM. B, predicted effect of GDP if UCP1 activity is regulated as described in Equation 1, i.e. if UCP1 is inherently inactive and needs fatty acids for activation. Curves drawn are predicted values for no endogenous free fatty acids and for hypothetical endogenous levels of 6 and 40 nM. C, predicted effect of GDP if UCP1 activity is regulated as described in Equation 2, i.e. if UCP1 is inherently active and is inhibited by GDP with oleate as competitive antagonist to the inhibition. alterations in Mg 2ϩ levels in the brown adipocyte could determine UCP1 activity if all of the other factors were unaltered.
The question then arises as to whether such high free fatty acid concentrations are encountered within the brown adipocytes. There is no information on this, nor is such information easily obtained. General assumptions have indicated that free fatty acid levels in cells do not normally exceed ϳ5 nM (22), and if this is the case, the levels needed for UCP1 activation are unrealistically high. Even though UCP1-dependent thermogenesis can be induced by exogenous addition of fatty acids to brown adipocytes (6), the effects observed may represent an artificial situation where the high extracellular fatty acid levels saturate the cytosolic fatty acid buffers, leading to a fatty acid-induced UCP1 activation as that studied here; however, it is not certain whether this observation represents the physiological avenue for UCP1 activation.
Indeed, whereas fatty acid-induced activation of UCP1 would be favored if the levels of free fatty acids in brown adipocytes were "less" buffered than in most other cell types, the opposite appears to be the case. Thus, brown adipose tissue has a high expression of the adipocyte form of the fatty acid-binding protein (A-FABP or aP2) (37), and it is a distinguishing trait of brown versus white adipose tissue that brown adipose tissue also expresses the heart form of the fatty acid-binding protein (H-FABP) (37). The affinities of these cytosolic binding proteins for fatty acids are also very high, higher than the highest affinities of the sites on albumin (used here as a pragmatic substitute for the cytosolic fatty acid-binding proteins). The affinity of the heart fatty acid-binding protein for oleate is as high as 4 nM (38). However, the fatty acid-binding proteins may play more active roles in the cells than being just fatty acid buffers (e.g. see Ref. 39). Thus, if the fatty acid-binding proteins were in some way to present the fatty acids directly to UCP1 through some kind of direct interaction similar to that described for the interaction between fatty acid-binding proteins and hormone-sensitive lipase (39), the apparent free fatty acid level at the UCP1 site could be much higher than the bulk level and UCP1 would react to these apparent levels and become activated.
Recently, compounds other than fatty acids have been proposed or discussed as possible regulators of UCP1 activity. These include retinoic acid (17), ubiquinone (18), superoxide (19), and 4-hydroxy-2-nonenal (20). We are currently examining these compounds in the present system to revalidate their UCP1-specific activity. However, when fatty acids (or their metabolites) are compared with these novel activators regarding their possible role as physiological regulators of UCP1 activity, fatty acids (or their metabolites) retain a teleological advantage. Whereas it is easy to envisage the cellular mechanism that leads from norepinephrine stimulation of a cell to lipolysis and increases in fatty acids and thus to UCP1 activation (2), an equally simple cellular mechanism has not been formulated for the novel activators.
GDP Affinity-The kinetically determined K m for free GDP was found to be 0.05 mM at pH 7.2. Although this value is low compared with the values normally considered when the inhibition is expressed in terms of total GDP (21), the value is unexpectedly high compared with the binding affinity for GDP originally reported (ϳ0.005 mM) (23) and confirmed in the detailed analyses of Rafael et al. (24) (ϳ0.001 mM at pH 7.0) and Huang and Klingenberg (40) (0.003 mM at pH 7.0). Apparently, the interaction between GDP and UCP1 is different when it is studied in non-respiring mitochondria (as in binding studies) compared with when the brown-fat mitochondria are respiring (as studied here). The reason is unclear, but there may be an effect of the mitochondrial membrane potential as has been discussed earlier (41).

Simple Competitive Kinetics between Fatty Acids and Purine
Nucleotides in Regulation of UCP1 Activity-Although the ability of purine nucleotides and fatty acids to interact with UCP1 has been recognized for decades, their competitive interaction has not been established earlier. The main reason for this is probably that until now it has not been possible to distinguish between the UCP1-dependent and UCP1-independent effects of activators and inhibitors and thus the kinetics were not evident.
In this context, it may be remembered that it is a classical contention that both removal of fatty acids (by albumin or oxidation) and the presence of GDP (purine nucleotides) are necessary to fully couple brown-fat mitochondria (10,42). This contention may at first be considered to be at divergence with the outcome of the present investigation, i.e. that GDP alone is able to fully inhibit UCP1-dependent activity. However, any UCP1-independent uncoupling effect of endogenous fatty acids cannot be counteracted by GDP. Furthermore, it will be understood from the competitive kinetics described here that any free fatty acid present would competitively counteract the coupling effect of a given amount of GDP. Experimentally, it could appear as if fatty acids interacted in another way than GDP because coupling was only induced when the fatty acids were removed. Rather, it may be that the removal of fatty acids leads to an increase in the apparent affinity of GDP so that the amount of GDP added is then sufficiently high to inhibit UCP1.
Do Oleate and GDP Interact at One Site?-Simple competitive interaction is most easily understood as two agents interacting at one site. Therefore, a simple molecular explanation for the competitive interaction demonstrated here would be that fatty acids and purine nucleotides (GDP and oleate) compete for the same binding site. However, this explanation is not readily acceptable. Firstly, fatty acids and purine nucleotides have very different structures. Secondly, there have been a series of studies where the ability of fatty acids to compete with GDP for the GDP-binding site in brown-fat mitochondria has been investigated. Several studies report no competition at all or perhaps a low affinity component (e.g. (11)). Although it was demonstrated in a recent study (43) that fatty acids could inhibit the binding of a purine nucleotide analogue, the K i observed for oleate was 112 M. This should be compared with the K i of 5 nM of free oleate calculated here, i.e. competitive effects on binding were apparently seen at a 20,000 higher concentration (the free concentration used by Huang (43) is difficult to estimate because no fatty acid buffer was used). Thus, no relevant competition between fatty acid and GDP has been observed in UCP1-binding site studies.
Nonetheless, the present data demonstrate a competitive interaction. The molecular background for this cannot currently be described. It may not be the fatty acids that interact with UCP1 but rather a metabolite. Structurally, a metabolite such as acyl-CoA has the necessary properties to compete for binding (44,45), but it is unlikely that sufficient amounts of acyl-CoA (oleyl-CoA) could be formed under the conditions investigated here (no ATP or CoA are added). Alternatively, it may be possible to formulate dynamic models that show apparent competition, e.g. by development of the model suggested by Rial and Nicholls (46).
Are Free Fatty Acids Necessary for UCP1 Thermogenic Function-The kinetic analysis performed here strongly supports a model (Model II) in which UCP1 in its native form is inherently active, allowing (the equivalent of) protons to (re)enter the mitochondrial matrix. In this model, purine nucleotides have an inhibitory function. The role of free fatty acids then is to overcome this inhibition. They are not required per se for the thermogenic activity of UCP1. In the physiological situation, they would be required for activation, but in the hypothetical absence of nucleotides, they would not be necessary for UCP1 function.
This conclusion is not uncontroversial in relation to at least two of the prevailing formulations concerning UCP1 function, i.e. the cofactor hypothesis and the cycling hypothesis. Although in these formulations fatty acids are believed to be necessary for UCP1 function, it has been proposed that they do not need to be added to mitochondrial preparations because endogenous fatty acids are suggested to be present. However, we find that if fatty acids were present in sufficiently high amounts to activate UCP1 in the fatty acid activation model (Model I), they would compete so strongly with GDP that the apparent GDP affinity would be much lower than that observed (Fig. 9). Therefore, we think that based on the kinetic analysis, we can also exclude a requirement for (endogenous) activating fatty acids for UCP1 function.
This result is in contrast to studies where UCP1 has either been reconstituted or ectopically expressed (47)(48)(49)(50)(51)(52). In such conditions, the addition of fatty acids is apparently necessary to induce UCP1 activity and experiments are routinely performed in the presence of an activating fatty acid. The UCP1 studied here is found in its native environment and should represent UCP1 in its native conformation. Why then are fatty acids not required here? One suggestion could be that the brown-fat mitochondria contain a hitherto undefined activating factor. This factor would then have to interact with UCP1 in a way different from fatty acids, i.e. at another site, and it would clearly not be in equilibrium with external albumin, for example.
However, as it appears from the present study that fatty acids are not needed for UCP1 function, the question becomes: why are fatty acids needed for UCP1 activity in the other systems? One possibility could be that ectopically expressed or reconstituted UCP1 may not be in its normal configuration but may have retained some properties of general mitochondrial carrier proteins. Several mitochondrial carriers share the feature that, if provided with fatty acids, they will function as uncoupling proteins, probably in the cycling mode originally proposed by Skulachev (53) for the ATP/ADP carrier (and extended to include UCP1 (53,54). What is observed in the reconstituted or ectopic systems may thus be a reflection of this rather general mitochondrial carrier property and may not correspond to the mechanism of UCP1 functioning in situ.

CONCLUSIONS
The development of UCP1-ablated mice has provided a novel possibility to distinguish between UCP1-dependent and UCP1independent properties of brown-fat mitochondria. This has been used here to demonstrate that, in the presence of endogenous levels of purine nucleotides, UCP1 can be activated by fatty acids in a way that is compatible with thermogenesis regulated physiologically through fatty acids (or their metabolites). Particularly, the system has revealed a simple competitive interaction between fatty acids and purine nucleotides, an interaction until now masked by the UCP1-independent effects of fatty acids (and purine nucleotides). This observation implies that, whereas fatty acids may physiologically regulate UCP1 activity, they are not necessary for UCP1 function and that models for the mode of action that are based on such involvement of fatty acids may therefore require reconsideration.