Fig. 2. Schematic model for the regulation of the 32000-Mr uncoupling protein from brown-fat mitochondria. (a) In the absence of purine nucleotides the gate (G) ...
BIOCHEMICAL SOCIETY TRANSACTIONS
OH or Cl
Fig. 2. Schematic model for the regulation of the 32000-Mr uncoupling protein from brown-fat mitochondria ( a ) In the absence of purine nucleotides the gate (G) is open while O H - and CIcompete to pass through the ohmic anion channel (AC). (b) Addition of purine nucleotide closes the gate : conductance is low until bulk-phase potential rises sufficiently for the field across the gate to induce ions to force their way through the gate. (c) Further addition of fatty acid allows anion channel to be bypassed by protons (but not by Cl- ). Proportion of bulk-phase potential concentrated across gate increases; thus protons can force their way through the gate at a lower bulkphase potential. Graphs give schematic proton current (I)proton electrochemical potential ( V ) relationships found experimentally (Nicholls, 1977; E. Rial & D. G. Nicholls, unpublished work).
(4) fatty acids de-couple the competition between protons and chloride; (5) fatty acids, however, have no effect on the chloride permeability by themselves. The model depicted in Fig. 2 reconciles these data. It is proposed that the 32000-M, protein has two components, a nucleotide-sensitive gate and an anion-selective transmembrane channel. The channel would not allow the simultaneous passage of protons and chloride ions in opposite directions, and so would not allow swelling to occur in KCl plus nigericin. Both chloride and protons (or hydroxyl ions) could be regulated by the gate, which on binding nucleotide would convert its conductance characteristics from ohmic to non-ohmic. Fatty acids would enable protons (but not chloride) to bypass the anion-selective channel (but not the gate). This increased permeability of the transmembrane portion of the protein would increase the field experienced by the gate, which would then require less bulk-phase proton electrochemical potential to achieve the local field required to increase its conductance non-ohmically.
Crompton, M., Capano, M. & Carafoli, E. (1976) Eur. J. Biochem.
Akerman, K . E. 0. (1978) Biochim. Biopbys. Acra 502, 359-366
Zoccarato, F. & Nicholls, D. G. (1982) Eur. J . Biochem. 127.333-338
Heaton, J . M. & Nicholls, D. G. (1976) Eur. J . Biochem. 67, 51 I517
Heaton, J. M.,Wagenvoord, R. J . , Kemp, A. & Nicholls, D. G. (1978) Eur. J . Biochem. 82, 515-521
Locke, R. M., Rial, E., Scott. I. D. & Nicholls, D. G. (1982a) Eur. J . Biochem. 129, 373-380
Locke, R. M., Rial, E. & Nicholls, D. G. (1982b) Eur. J. Biochem. 129, 381-387
Mitchell, P. (1961) Nature (London) 191, 423427 Nicholls, D. G. (1974) Eur. J. Biochem. 49, 585-593 Nicholls, D. G. (1977) Eur. J . Biochem. 77, 349-356 Nicholls, D. G. (1979) Biochim. Biophys. Acra 549, 1-22 Nicholls, D. G. (1983) Bimci. Rep. 3, 431-440 Nicholls, D. G. & Akerman, K. E. 0.(1982) Biochim. Biophys. Acra 683,57-88 Nicholls, D. G. & Crompton, M. (1980) FEBS L P r f . 111, 261-268 Nicholls, D. G. & Lindberg, 0.(1973) Eur. J . Biochem. 37,523-530 Nicholls, D. G. & Locke, R. M. (1984) Physiol. Rev. in the press Siliprandi, D., Toninello, A,, Zoccarato, F., Rugolo, M. & Siliprandi, N . (1978) J. Bioenerg. Biomembr. 10, 1-11 Snelling, R. & Nicholls, D. G. (1982) Shorr Rep. Eur. Bioenerg. Con$ 2nd 495496
Characteristics of the uncoupling protein from brown-fat mitochondria M. KLINGENBERG lnstitute tor Physical Biochemistry, University of Munich. Goethestrasse 33, 8000 Munich 2, Federal Republic of Germany
The inner membrane from brown-fat mitochondria contains in large abundance an integral membrane protein Abbreviations used: DAN, dirnethylaminonaphthoyl: DCCD, dicyclohexylcarbodi-imide;SDS, sodium dodecyl sulphate.
which is thought to be instrumental for the thermogenesis of brown adipose tissue (Nicholls, 1979). It has been suggested that this protein forms a channel for H +,thus short-circuiting the H + efflux generated by the respiratory chain. As a result H +would bypass ATP synthesis and instead all the oxidative energy would be converted into heat. The main argument for the identitification of this protein with a H + channel comes comes from the inhibition in mitochondria of the uncoupling action by purine nucleotides, and the presumption, based on photoaffinity labelling, that an SDS gel 1984
606th MEETING, CORK band at M , 32000 is the nucleotide-binding one (Heaton et al., 1978).
Three years ago in our laboratory the nucleotide-binding protein was isolated and purified from brown-fat mitochondria by an efficient and simplified method which is a modification of the procedure for the isolation of the ADP/ATP carrier (Lin & Klingenberg, 1980, 1982). The major prerequisite was the survival of the nucleotidebinding capability after solubilization of the membranes in certain detergents such as Triton X-100. Indeed, on fractionation the nucleotide-binding capacity was found to increase parallel in with the 32000-M, component content. Thus two facts were established, that the 32000-M, component is the nucleotide-binding protein and that it remains intact, at least with regard to the nucleotide-binding function, after solubilization and isolation. With the the isolation and purification of this protein, the stage was set for investigating the actual function and mechanism, which so far has only been based on suggestive evidence from mitochondria. Furthermore, the studies on the structural characteristics of this protein are now possible as a longrange project which finally should facilitate elucidation of its molecular mechanism. Some chemical and physical characteristics of the isolated uncoupling protein
The binding of the purified uncoupling protein is approx. 16pmol of nucleotide bound/g of protein, which corresponds to a functional M , of 63000 (Lin & Klingenberg, 1982). In view of the fact that the protein in SDS/gel electrophoresis exhibits an apparent M ,of 32000, it is suggested that the protein forms a dimer with only one binding centre, similar to the ADP/ATP carrier (Lin & Klingenberg, 1980). Hydrodynamic studies of the uncoupling protein by gel filtration and sedimentation velocity and sedimentation equilibrium runs showed that indeed the protein has a M, of 62000 (Lin el af., 1980). This has been calculated from the M , of I71 000 of the mixed Triton protein micelle after subtracting the Triton portion. The micelles contain a high excess of Triton as the result of their large hydrophobic surfaces, similar to the ADP/ATP carrier. The amino acid composition of the uncoupling protein is dissimilar to that of the ADP/ATP carrier. The protein is less basic and slightly less hydrophilic. The U.V. absorption and the fluorescent excitation and emission spectra are characteristic primarily of tyrosine and also of phenylalanine contributions. This indicates that the protein may have a much lower tryptophan content than the ADP/ATP carrier. The U.V. and c.d. spectra also are dominated by tyrosine signals, whereas in the ADP/ATP carrier the tryptophan signal is prominant. The c.d. spectra in the far U.V. region permit evaluation of a 40% a-helical content, which is rather close to that of the ADP/ATP carrier. The binding of purine nucleotides
Purine-nucleotide binding is a major function of the uncoupling protein which can be studied. For this reason the term ‘nucleotide-binding protein’ has been used (Lin & Klingenberg, 1982), especially since the uncoupling function of the protein is not yet established. The purine di- and tri-nucleotides ATP, ADP, G T P and G D P are all quite effective ligands with K , values of approx. 1 0 - 6 (Table ~ 1). In mitochondria only the binding of G D P and G T P could be measured since ATP and A D P are converted and mixed with the endogenous ATP/ADP pool (Rafael & Heldt, 1976; Nicholls, 1976). Therefore G T P and G D P were mostly considered as the main ligands of the nucleotidebinding protein. With the isolated protein ATP is shown to be the most tightly binding ligand (Lin & Klingenberg, 1982). This would agree with the probable physiological role VOl.
39 1 Table 1. Dissociation constunts and p H diyrndmcr Binding of purine nucleotide to isolated uncoupling protein as determined by equilibrium dialysis. Ligand ADP ATP GDP GTP ITP AMPPNP
K , (PM) ( < pH*) 5 0.8
4.8 4.2 6.2 8.0
Break pH* 6.3
p ~ * )
6.3 7.2 6.4 6.25
6.2 < 5.8
- 2 1 0 -3 -1 -1
of ATP as an inhibitor of the uncoupling action since ATP is the major nucleotide in the cytosol. Nucleotides with other bases, such as CTP, UDP, etc., do not bind markedly. The affinities of the various nucleotides in the pH-independent range below pH6.2 range between 2 x lo-’ and 5 x 10-6M. The affinity shown by ATP is a remarkably tight binding for a nucleotide. It is considerably higher than that for the ADP/ATP carrier which has a K d of 1 0 - 5 ~in the membranes (Weidemann et al., 1970) and > 1 0 - 4 ~ after isolation. The much higher binding in the uncoupling protein may be rationalized by considering that the nucleotide binds as an inhibitor whereas in the nucleotide translocator it binds as a substrate. We have in this contrast a beautiful example illustrating the importance of binding energy for function (Jencks, 1975). Whereas in the uncoupling protein almost the full binding energy seems to be unleashed, in the translocator it is considerably tamed by compensatory debinding energy used in the activation process for translocation. This latter aspect has formerly been discussed for the translocator (Klingenberg, 1975). However, the nucleotide binding in the uncoupling protein appears not to be a simple process since it is relatively slow, as will be shown below. p H dependency of nucleotide binding
The pH dependence of nucleotide binding is a remarkable feature and possibly related to the function of the putative H translocating site in the uncoupling protein. Using elaborate methods for determining the binding of nucleotides over a wide pH range, we were able to obtain a broad picture of the pH dependency of most of the various nucleotide-binding components. The data are presented as ‘Dixon’ plots of the pK, vs pH (Fig. 1). For the four natural purine nucleotides a biphasic pH dependency emerges as a common feature with an essentially pH independence below pH 6.3-6.8. Above that range the affinity drastically decreases. In the pKd/pH plot this dependence can be described by a straight line with a slope of - 1. Whereas for GTP, G D P and ADP this simple relation holds up to pH7.6, a much steeper decline is observed for the affinity of ATP. Here the pK, decreases above pH 7.3 with a slope of - 2 and more, such that in a very small range of pH increase ATP becomes non-binding. The pH dependence may indicate that an ionizing group of the uncoupling protein regulates the binding of the nucleotides. For example, the protein binds nucleotides only when a certain group is protonated with a pK of around 6.6: +
U + H + e U H + , pK,-6.5; U H + + N + U H + . N , pK,=6.8-6.3 The protonating group could be histidine. The observed pH break can, however, also reflect disorder at the y-phosphate of the nucleotides which happens
BIOCHEMICAL SOCIETY TRANSACTIONS
Fig. 1. The p H dependence ofpurine-nucleotide binding to the uncoupling protein The protein was isolated from brown-adipose-tissue mitochondria obtained from cold-adapted hamsters. Evaluation of p Kd was by determining the concentration dependence of nucleotide binding.
to be in that range. In this instance we can consider two cases, one where ionization of the uncoupling protein is not involved and one where it is involved, however, with an H + dissociation group having a pK < 5. In the first case only the protonated nucleotides are in the binding form, such as ADPH*-, ATPH’ ~,etc :
U+NeUN; U + N- : reaction not possible where N = ADPHZ- and ATPH3- etc. and N- = ADP3and ATP4In the second case we have an opposite situation, only the ionized nucleotides, such as ADP3-, ADP4- etc., are binding. The uncoupling protein can bind only in the protonated form which with the pK