Biochemical Society Transactions

Insulin receptor tyrosine kinase structure and function

426

Leland Ellis,*§ Jeremy M. Tavare,*t and Barry A. Levine$ *Howard Hughes Medical Institute and Department of Biochemistry, University of Texas Southwestern Medical School, 5323 Harry Hines Blvd., Dallas, TX 75235-9050, U.S.A. and $Department of Inorganic Chemistry, University of Oxford, Oxford 0x1 3QR, U.K.

Introduction The insulin receptor The response of cells to the polypeptide hormone insulin initiates with the binding of insulin to a specific cell-surface receptor. The insulin receptor is an integral transmembrane glycoprotein comprised of two a-subunits (735 amino acids, M, 84214) which interact with the hormone, and two B-subunits (620 amino acids, M, 69 703) which each span the plasma membrane a single time and harbour a cytoplasmic protein-tyrosine kinase (PTK) domain [8, 411. Autophosphorylation of the B-subunit occurs at multiple tyrosine residues as a consequence of insulin binding. While the rapid phosphorylation of Tyr-1158, -1 162 and -1163 (which reside in the kinase homology region of the receptor primary sequence) seems to correlate with the activation of the enzyme towards exogenous substrates, the role of phosphorylation at a number of additional sites (Tyr-1328 and -1334 in the Cterminal tail, and one or more of Tyr-965, -972 and -984 in the juxtamembrane region of the cytoplasmic domain) is at present unknown [14, 33, 36-38, 451. Furthermore, the functional consequences of utilizing three closely spaced Tyr (i.e. 1158, 1162 and 1163) in the control of kinase activation is obscure.

Analysis of mutations of the human insulin receptor protein The availability of cDNAs encoding the human placental insulin receptor protein not only provided the deduced primary sequence and topology of the receptor protein, but also the means by which to establish heterologous cell expression systems in which to study the manifold aspects of receptor function. In particular, the use of recombinant DNA methods to introduce both site-directed point mutations and deletions into the receptor has provided an experimental venue by which to explore the Abbreviations used PTK, protein-tyrosine kinase; CHO, Chinese hamster ovary. tPresent address: Department of Biochemistry, School of Medical Sciences, University of Bristol, Bristol BS8 lTD, U.K. §To whom correspondence should be addressed.

Volume 19

structure and function of this receptor, and to test directly ideas concerning the functional significance of discrete receptor residues on particular aspects of receptor function (see references in the legends to Figs. 1 and 2). In Figs. 1 and 2, results derived from the study of a number of such mutations on insulin receptor autophosphorylation, exogenous substrate phosphorylation (PTK activity) and several signalling pathways in intact cells are summarized. At the outset, it should be noted that there are as yet unresolved controversies concerning the phenotypes observed for several mutations, which are likely to derive from experimental limitations inherent in the cell systems utilized thus far in these studies. First, most of the rodent cell lines used for transfection of the human receptor cDNA [e.g. Chinese hamster ovary (CHO), NIH 3T3 and rat1 cells] are not very responsive to insulin, e.g. CHO cells exhibit a 2-fold effect of insulin on glucose transport. Secondly, there is often clonal variation [i.e. between individual clones derived from a stable transfection with a particular cDNA construct (e.g. the basal activity of a particular metabolic response such as insulindependent thymidine uptake may vary between individual stably transfected clones); such elevations in basal activity complicate the interpretation of slight increases in sensitivity observed in insulin dose-response curves; see below]; thus, the study of a number of independently derived clones (or pools of transfected clones) for each construct is required. Thirdly, the analysis is always complicated by the presence of endogenous wild-type insulin receptors found in all the cell lines used to date (typically a few thousand endogenous receptors per cell in the commonly used fibroblast cell lines). The transfection of a heterologous compliment of human wild-type insulin receptors into these cells (levels of lo4to loh receptor per cell are readily obtainable) increases the sensitivity of the transfected cell to insulin (i.e. causes a leftward shift in the insulin dose-response). Thus, one is monitoring often subtle changes in insulin sensitivity above the endogenous response. Furthermore, mutations which compromise (but do not eliminate) receptor signalling, generally cause a loss of this increased insulin sensitivity towards that observed for mock-transfected cells, while cells expressing

-

-

Nucleotide Binding and Hydrolysis

Fig. i Analysis of human insulin receptor function in intact cells by domain deletion Summary of the qualitative effects of deletion of specific regions of the human insulin receptor primary amino acid sequence on various aspects of insulin receptor function in intact transfected cells. Key to symbols: autophosphorylation and PTK activity (either pp185 phosphorylation or activity towards synthetic substrates) are measured after isolation of receptors, in the presence of phosphatase inhibitors from insulin-treatedcells. Results are expressed as insulin stimulated, ; constitutively elevated and insulin independent, (+); little or no detectable activity, -; or not determined, nd. The insulin sensitivity of glucose transport, glycogen synthesis, thymidine uptake and 56 kinase are expressed as greater, upward pointing arrow; lower, downward pointing arrow, or unchanged, #, when compared with mock-transfected cells. Source of data: spBam (deletion of residues 13-599 [2, I I]); iBgl (deletion of residues 1-10 and 13-958: [2, I I]); Aex16 (deletion of exon 16= residues 956-977: [a]; similar results were obtained upon deletion of residues 966-977: [I]); ACT43 (deletion of C-terminal residues I3 13- 1355: [21, 23, 391) and T-t (deletion of C-terminal residues 1244- 1355: [ 10, 161).

+

Transmembrane

-kl

Tyrosine kinase Tail Wild- spBam iBgl Aex I 6 ACT43 T-t type Autophosphorylation

PTK activity Endocytosis

+ + ( + +)*( + +)* + + + + ++(++)*(++)*++ + + + nd nd - + + -

Glucose uptake

# # +

t

Glycogen synthesis

# # 4

nd

nd

nd

nd

nd

nd

#

#

nd

Thymidine uptake

56 kinare activity

n d # #

nd nd

nd

nd

*J.M. Tavare & L. Ellis, unpublished work. jConstitutively elevated in the absence of insulin.

receptors with point mutations or deletions that eliminate receptor signalling [e.g. mutation of Lys1030 in the kinase ATP-binding site, or C-terminal truncations which compromise PTK activity (the T-t mutant; Figs. 1 and 2) often have a lower sensitivity to insulin than mock-transfected cells (and thus exhibit a rightward shift in the insulin dose-response curve). Thus it is especially important to ensure that all cell lines within an experiment

possess near equal numbers of transfected wildtype or mutant insulin receptors. Despite the above limitations, there is now general agreement concerning several aspects of insulin receptor function that can be gleaned from the data summarized in Figs. 1 and 2, and these findings have now been observed in several laboratories (for references, see the legends to Figs. 1 and 2). (i) Lysine-1030, which is thought to interact with the phosphate moiety of ATP, is crucial for receptor PTK activity and almost all aspects of insulin receptor signalling including receptor internalization; however, see footnote 5 in the legend to Fig. 2. (ii) A stretch of 12 amino acids in the juxtamembrane region of the cytoplasmic domain (residues 966-977; [ 1, 401) appears to be involved in insulin receptor internalization. This sequence contains the NPXY motif which has been implicated in the internalization of the low-density lipoprotein receptor [3]. (iii) Replacement of two of the major sites of insulin-stimulated autophosphorylation, Tyr- 1162 and - 1163, substantially reduces the phosphorylation of exogenous substrates by receptors isolated from insulin-treated cells, as well as insulin-stimulated glucose uptake in intact cells [ 101. Thus both the insulin-stimulated autophosphorylation of the B-subunit of the insulin receptor on tyrosine residues, as well as PTK activity, are crucial aspects of insulin receptor transmembrane signalling. There are at present a number of actively debated aspects of insulin receptor function, particularly involving the individual roles of Tyr-1158, -1 162 and -1 163. The phosphorylation of all three tyrosines (to 3 mol of phosphate/mol of sites) has been proposed to correlate with activation of the kinase towards exogenous substrates [ 14, 451. The fact that replacement of Tyr-1162, or both Tyr1162 and -1163, with phenylalanine(s) compromises kinase activation appears, at first sight, to be in general agreement with this proposal. However, while Zhang et al. [5 l] report that replacement of Tyr-1158 with phenylalanine has no affect on insulin-stimulated receptor kinase activation, Wilden et al. [48] report a substantial reduction. The resolution of this discrepancy is imperative, as Wilden et al. [48] have proposed that the mutation of Tyr-1158 has opposing effects on insulin-stimulated glycogen synthesis and thymidine uptake compared with those observed in the Tyr-11621 1163 to phenylalanine double mutant (see Fig. 2).

Autonomous function by individual domains of the insulin receptor The first indication that the major functional

1991

427

Biochemical Society Transactions

Fig. 2 Analysis of human insulin receptor function in intact cells

428

Summary of the qualitative effects of substituting individual amino acids of the human insulin receptor (by site-directed mutagenesis) on various aspects of insulin receptor in function in intact cells. Key to symbols: see legend to Fig. I. Source of data: Y972 (Tyr-972 [46]); G I008 (Gly- I008 [26, 491); K I030 (Lys- I030 [4, 7, 16, 20, 22, 29, 301); A I I34 (Ala- I I 3 4 [24]); Y I I58 (Tyr- I I58 [48, 5 I]); Y I I62 (Tyr.I I62 (2, 10,341); Y I I62 (Tyr- I I62 and -I 163: [6, I0,27,34,5 I]) and W I200pr-l( 1200: [U]). Transmembrane

Tyrosine kinase

Tail

4 N

e Autophosphorylation PTK activity Receptor serine phosphorylation Receptor threonine phosphorylation Endocytosis Glucose uptake Glycogen synthesis Thymidine uptake 56 kinase activity/S6 phosphorylation

O D 0

88

Gz

0

wz

Wild-type

Y972*

Al134t

++ ++ ++ ++ ++

++

-

nd nd

++

nd nd nd

nd # f nd

nd nd nd nd

4

44 # #

-

*Recently, Backer et 01. [ I ] identified a second previously unidentified mutation in their construct which lies two codons away from tyrosine 972. This resulted in the substitution of Ser-974 with a threonine. At present it is not known whether this affects the interpretation of the data given in this Table. tMutations which were first identified in insulin receptors from human patients. $Data given in parentheses are from Zhang et 01. [5 I]. $This is disputed by Sbraccia et a/. [30]. 1IData in parentheses is from Yonezawa et ol. [50]. llData in parentheses is from Boulton eta/. [2].

domains of the insulin receptor were in fact capable of autonomous function derived from studies of receptors with a deletion within (or complete removal of) the extracellular ligand-binding domain. Such altered receptors exhibit insulin-independent autophosphorylation and activation of exogenous PTK activity, whether engineered and expressed as an externally truncated membrane-associated protein (the spBam mutant of Fig. 1) or as a truncated soluble receptor cytoplasmic domain (the iBgl mutants of Fig. 1). This agrees with the observations of Goren et al. [151, who found that truncation within the receptor ectodomain by limited tryptic proteolysis of the intact insulin receptor had an equivalent effect. Interestingly, only the spBam construct (and not the iBgl) appears to confer a constitutively elevated insulin-independent glucose uptake upon CHO cells [ll], while S6 kinase activity was apparently unaffected by either construct [Z].These observations suggest that the truncated molecule must express kinase activity in the proximity of a membrane compartment to signal elevated glucose uptake. To date, these observations have been made only in stably transfected mammal-

Volume 19

ian cell lines. The consequences of the chronic expression of such constructs in the context of an intact animal are unknown, but are now being explored in transgenic mice (J. M. Tavark, D. Clothier, B. Hammer & L. Ellis, unpublished work). After these initial observations of autonomous kinase domain function, and the observation that the cytoplasmic PTK could be expressed as a soluble cytoplasmic active enzyme (i.e. the iBgl construct of Fig. l), the deduced transmembrane topology of the human insulin receptor was formally tested and verified for the extracellular domain as well. Thus, truncation before the hydrophobic stretch of amino acids predicted to comprise the single transmembrane segment of each half-receptor, results in the secretion of a soluble heterotetramer (i.e. a2&, where /lo is the extracellular portion of the membrane-spanning /?-subunit) which binds insulin with high affinity [13, 18, 31-32, 471. Given the size (1355 residues per halfreceptor) and membrane-associated nature of the wild-type insulin receptor protein, it has subsequently proven experimentally advantageous to

Nucleotide Binding and Hydrolysis

exploit this relatively simple transmembrane topology of the receptor to engineer and study its two functional domains individually as soluble molecules, i.e. the extracellular domain as a soluble secreted insulin-binding protein, and the cytoplasmic domain as a soluble PTK. The use of these soluble receptor derivatives, together with p o w e h l heterologous cell-expression systems, has now rendered feasible the application of biophysical methods to the studies of these domains, which are still not possible for the native membrane-associated receptor.

The insulin receptor extracellular domain The extracellular ligand-binding domain of the insulin receptor is, at the outset, a very complex molecule to contemplate: each half-receptor is comprised of 929 residues derived from both a- (735 amino acids) and /3-subunits (194 amino acids), and includes 16 potential asparagine-linked glycosylation sites and 41 cysteines. Furthermore, there is little information at present as to how this domain folds during biosynthesis, or how it interacts with insulin. The study of an extensive series of deletion mutants has revealed sites within the receptor primary sequence at which truncation results in the generation of independently folded soluble subdomains [31]. As we now appreciate the stability, efficiency of secretion and interaction of these subdomains both with insulin and a panel of receptorspecific monoclonal antibodies, these sites now provide landmarks to guide further biochemical and molecular dissections of this complex domain. Furthermore, the establishment of a heterologous cell expression system, from which 10-20 mg of pure secreted heterotetrameric soluble ectodomain (each half-receptor is comprised of 923 amino acids, which includes all of the a-subunit and the extracellular 113 of the p-subunit) can be purified (E. Schaefer, L. James & L. Ellis, unpublished work), is expected to expedite direct biochemical dissection of the protein, as well as the mapping of insulin interaction sites by the use of covalent crosslinking and protein microsequencing (e.g. ref. [43]). Finally, these quantities now render feasible efforts to crystallize this domain, a prequisite for the elucidation of its three-dimensional structure by X-ray crystallography (in collaboration with Wayne Hendrickson, Columbia University).

-

The insulin receptor PTK domain To facilitate the biochemical study of the kinase domain of the receptor, a soluble derivative of the

cytoplasmic PTK domain of the human insulin receptor [401 amino acids, the cytoplasmic 213 of the /3-subunit (receptor residues 959-1355); M, 452971 has been expressed in insect Sf9 cells by the use of a recombinant Baculovirus vector [12, 171, whereby the enzyme can now be purified in 10 mg quantities from infected cells. The enzyme is recognized by a panel of conformation-sensitive anti-receptor monoclonal antibodies [ 121 and exhibits a low basal level of exogenous substrate activity which increases upon autophosphorylation [S]. Furthermore, the tyrosine sites of autophosphorylation are entirely typical of those observed for the native insulin receptor following insulin stimulation in intact cells (i.e. at sites corresponding to Tyr-1158, -1162, -1163, -1328 and -1334 of the native receptor; [35]). Thus, even though synthesized free of the plasma membrane and now insulin independent, this soluble enzyme displays a number of properties of the wild-type receptor kinase domain. As for studies considered above (see Figs. 1 and 2), the analysis of the soluble insulin receptor PTK domain is not without its unresolved questions, as both intermolecular [51 and intramolecular [42] mechanisms of soluble kinase autophosphorylation have been proposed. Furthermore, while the addition of Mn2+ to the enzyme causes a pronounced reduction in the a-helical contribution of the c.d. spectrum of the enzyme [44], such a change has not been observed with Mg2+ (B. Clack & L. Ellis, unpublished work). Finally, the two divalent cations clearly influence the efficacy of autophosphorylation to different degrees (i.e. Mn2+> Mg2+), whereas the addition of polybasic compounds such as protamine [5] or polylysine [28] enhance the degree of autophosphorylation observed with either Mn2+ or M$+ alone (J. M. Tavark, E. Alejos, B. A. Levine & L. Ellis, unpublished work). As different ionic conditions were employed in the studies reporting intermolecular (M$+ f protarnine) versus intramolecular (Mn2+ and M g + ) mechanisms for soluble kinase autophosphorylation, it is certainly possible that the disagreement results from these different conditions, and that in fact both intra- and inter-molecular modes of autophosphorylation can be utilized, depending upon the ionic conditions of the reactions examined (differences in the 'basal' state of phosphorylation of the purified enzymes are also possible; see [35] for discussion). Direct comparisons of these conditions are clearly needed 0. M. Tavari, E. Alejos, B. A. Levine & L. Ellis, unpublished work). With the quantities of purified enzyme now

-

1991

429

Biochemical Society Transactions

430

available, we have begun to utilize 'H-n.m.r. spectroscopy to study the functional properties of this soluble derivative of the cytoplasmic FTK domain of the receptor. These methods provide new experimental avenues by which to explore the interaction of small molecules (metal ions, ATP, peptide substrates) with the enzyme, and have provided the first look at catalysis by a PTK in real time. We can now follow the binding of peptide substrates to the enzyme, the phosphorylation of individual tyrosine residues of such peptides (especially the order of phosphorylation of multiple tyrosines), as well as study the role of individual amino acid residues of the peptide on its binding to the enzyme and its kinetics of phosphorylation [9, 191. For example, the proton resonances of each of the three tyrosines of the synthetic peptide RRDIYETDYYRK, which is derived from the receptor sequence around the three major sites of autophosphorylation [i.e. Tyr1162 (residue 5 of the peptide), Tyr-1162 (residue 9) and Tyr-1163 (residue lo)], are distinguishable in the aromatic region of 'H-n.m.r. spectra. Thus, it is feasible to follow directly, in solution and in real time, the phosphorylation at each site. Furthermore, not only is each of the three tyrosines phosphorylated, but phosphorylation of the peptide is highly ordered and progressive. First Tyr-9, then Tyr- 10 and finally Tyr-5 is phosphorylated, with phosphorylation proceeding to completion at each site before phosphorylation of the next site begins [19]. On the basis of these observations, it is clear that it is now feasible to explore in detail the stereochemical requirements and dynamics of exogenous peptide phosphorylation by this enzyme (B. A. Levine & L. Ellis, unpublished work). A logical extension of the n.m.r. studies of peptide phosphorylation by the enzyme is to assess the utility of this method to perhaps visualize autophosphorylation of the enzyme per se. This is a formidable technical undertaking at present: the 401 residue enzyme is significantly larger than other proteins for which two-, three- and now fourdimensional n.m.r. methods have been applied to date, and at least six tyrosines of the kinase can be phosphorylated. However, initial experiments demonstrate that changes are indeed observed in the aromatic region of 'H-n.m.r. spectra of the enzyme during autophosphorylation (B. A. Levine & L. Ellis, unpublished work). Thus, this experimental approach, when combined with the study Of appropriate mutant enzymes (both site-directed mutations of tyrosines as well as truncations are obviously of interest), can be expected to provide a starting point to begin to explore the dynamics of

Volume 19

PTK phosphorylation in solution in real time. These functional studies of the enzyme in solution complement efforts to obtain the three-dimensional structure of the enzyme by X-ray crystallography (in collaboration with Wayne Hendrickson, Columbia University).

Summary Six years have now elapsed since efforts to establish heterologous cell expression systems for studies of the human insulin receptor were begun. As is apparent from the results summarized in Figs. 1 and 2, a significant number of studies have been devoted to the analysis of receptor mutations, both experimentally derived (i.e. by mutagenesis) and those identified in human patients, as well as to the generation of soluble derivatives of the major functional domains of the receptor for use in biophysical studies. While it is certainly clear that these methods can be expected to yield an ever-increasing body of data concerning insulin receptor structure/function, it is equally apparent that attention to a number of basic experimental limitations inherent in these approaches (see above) will be required to resolve a number of fundamental questions and disagreements concerning particular receptor mutations. Given the level of interest in the insulin receptor that has persisted over the past several decades, one expects that these efforts will be forthcoming, and that our understanding of this complex transmembrane receptor will, with time, improve. We thank our colleagues (both past and present) in Dallas and Oxford for their experimental contributionsto the studies in our laboratories, and Richard Roth (Stanford) and Ken Siddle (Cambridge) for discussions of receptor structure/function. The work in our laboratories was supported by the N.I.H. (L.E.), the Howard Hughes Medical Institute (L.E.), the British Diabetic Association (J.M.T.) and the M.R.C. (J.M.T. and B.A.L.). J.M.T. is the recipient of a M.R.C. Travelling Fellowship. B.A.L. is an associate member of the Oxford Centre for Molecular Sciences.

1. Backer, J. M., Kahn, C. R., Cahill, D. A., Ullrich, A. &

White, M. F. (1990)J. Biol. Chem. 265, 16450-16454 2. Boulton, T. G., Gregory, J. S., Jong, S.-M., Wang, L.-H., Ellis, L. & Cobb, M. H. (1990) J. Biol. Chem. 265,2713-2719 3. Chen, W. J., Goldstein, J. L. & Brown, M. S. (1990) J. Biol. Chem. 265,3116-3123 4. Chou, C. K., D ~ UT., J., Russell, D. s., Gherzi, R., Lebwohl, D., Ullrich, A. & Rosen, 0. M. (1987) J. Biol. Chem. 262,1842- 1847

Nucleotide Binding and Hydrolysis

5. Cobb, M. H., Sang, B.-C., Gonzalez, R., Goldsmith, E. &Ellis, L. (1989)J. Biol. Chern. 264,18701-18706 6. Debant, A., Clauser, E., Ponzio, G., Filloux, C., Auzan, C., Contreres, J. 0. & Rossi, B. (1988) Proc. Natl. Acad. Sci. USA. 85,8032-8036 7. Ebina, Y., Araki, E., Taira, M., Shirnada, F., Mori, M., Craik, C. S., Siddle, K., Pierce, S. B., Roth, R. A. & Rutter, W. J. (1987) Proc. Natl. Acad. Sci. USA. 84, 704-708 8. Ebina, Y., Ellis, L., Jarnigan, K., Edery, M., Graf, L., Clauser, E., Ou, J., Masiarz, F., Kan, Y. W., Goldfine, I. D., Roth, R. A. & Rutter, W. J. (1985) Cell (Carnbridge, Mas.) 40,747-758 9. Ellis, L. & Levine, B. A. (1991) Methods Enzyrnol. 200,660-669 10. Ellis, L., Clauser, E., Morgan, D. O., Edery, M., Roth, R. A. & Rutter, W. J. (1986) Cell (Cambridge, Mass.) 45,721-732 11. Ellis, L., Morgan, D. O., Clauser, E., Roth, R. A. & Rutter, W. J. (1987) Mol. Endocrinol. 1, 15-24 12. Ellis, L., Levitan, A., Cobb, M. H. & Rarnos, P. (1988) J. Virol. 62, 1634-1639 13. Ellis, L., Sissorn, J. & Levitan, A. (1988) J. Mol. Recognit. 1,25-31 14. Flores-Riveros, J. R., Sibley, E., Kastelic, T. & Lane, M. D. (1989)J. Biol. Chem. 264,21557-21572 15. Goren, H. J., White, M. F. & Kahn, C. R. (1987) Biochemistry 26,2374-2382 16. Hari, J. & Roth, R. A. (1987) J. Biol. Chern. 262, 15341-15344 17. Herrera, R., Lebwohl, D., de Herreros, A. G., Kallen, R. G. & Rosen, 0. M. (1988) J. Biol. Chern. 263, 5560-5568 18. Johnson, J. D., Wong, M. L. & Rutter, W. J. (1988) Proc. Natl. Acad. Sci. U.S.A. 85,75 16-7520 19. Levine, B. A., Clack, B. & Ellis, L. (1991) J. Biol. Chern. 266,3565-3570 20. Maegawa, H., Olefsky, J. M., Theis, S., Boyd, D., Ullrich, A. & McClain, D. A. (1988) J. Biol. Chern. 263,12629-12637 21. Maegawa, H., McClain, D. A., Freidenberg, G., Olefsky, J. M., Napier, M., Lipari, T., Dull, T. J., Lee, J. & Ullrich, A. (1988)J. Biol. Chern. 263,8912-8917 22. McClain, D. A., Maegawa, H., Lee, J., Dull, T. J., Ullrich, A. & Olefsky, J. M. (1987) J. Biol. Chern. 262, 14663-1 4671 23. McClain, D. A., Maegawa, H., Levy, J., Huecksteadt, T., Dull, T. J., Lee, J., Ullrich, A. & Olefsky, J. M. (1988)J. Biol. Chern. 263,8904-891 1 24. Moller, D. E., Yokota, A., White, M. F., Pazianos, A. G. & Flier, J. S. (1990) J. Biol. Chern. 265, 14979-14985 25. Moller, D. E., Yokota, A., Ginsberg-Fellner, F. & Flier, J. S. ( 1990) Mol. Endocrinol. 4, 1 183- 1 19 1 26. Odawara, M., Kadowaki, T., Yarnarnoto, R., Shibasaki, Y., Tobe, K., Accili, D., Bevins, C., Mikarni, Y.,Matsuura, N., Akanurna, Y., Takaku, F., Taylor, S.I. & Kasuga, M. (1989) Science 245,66-68

27. Reynet, C., Caron, M., Magre, J., Cherqui, G., Clauser, E., Picard, J. & Capeau, J. (1990) Mol. Endocrinol. 4, 304-31 1 28. Rosen, 0. & Lebwohl, D. (1988) FEBS Lett. 231, 397-401 29. Russell, D. S., Gherzi, R., Johnson, E. L., Chou, C. K. & Rosen, 0. M. (1987) J. Biol. Chern. 262, 11833-11840 30. Sbraccia, P., Wong, K. Y., Brunetti, A., Rafaeloff, R., Trischitta, V., Hawley, D. M. & Goldfine, I. D. (1990) J. Biol. Chern. 265,4902-4907 3 1. Schaefer, E. M., Siddle, K. & Ellis, L. (1990) J. Biol. Chem. 265,13248-13253 32. Sissom, J. & Ellis, L. (1989) Biochern. J. 261, 119-126 33. Tavar6, J. M. & Denton, R. M. (1988) Biochern. J. 252, 607-61 5 34. Tavark, J. M. & Dickens, R. M. (1991) Biochern. J. 274,173-179 35. Tavar6, J. M., Clack, B. & Ellis, L. (1991) J. Biol. Chern. 266,1360-1395 36. Tavark, J. M., OBrien, R. M., Siddle, K. & Denton, R. M. (1988) Biochern. J. 253,783-788 37. Tornqvist, H. E., Pierce, M. W., Frackelton, A. R., Nernenoff, R. A. & Avruch, J. (1987) J. Biol. Chern. 262,102 12- 10219 38. Tornqvist, H. E., Gunsalus, J. R., Nernenoff, R. A., Frackelton, A. R., Pierce, M. W. & Avruch, J. (1988)J. Biol. Chern. 263,350-359 39. Theis, R. S., Ullrich, A. & McClain, D. A. (1989) J. Biol. Chern. 264,12820-12825 40. Theis, R S., Webster, N. J. & McClain, D. A. (1990)J. Biol. Chern. 265,10132-10137 41. Ullrich, A., Bell, J. R., Chen, R. Y., Herrera, R., Petruzzelli, L. M., Dull, T. J., Gray, A., Coussens, L., Liao, Y.-C., Tsubokawa, M., Mason, A., Seeburg, P. H., Grunfeld, C., Rosen, 0. M. & Rarnachandran, J. (1985) Nature (London) 313,756-761 42. Villalba, M., Wente, S. R., Russel, D. R., Ahn, J., Reichelderfer, C. F. & Rosen, 0. M. (1989) Proc. Natl. Acad. Sci. U.S.A. 86,7848-7852 43. Wedekind, F., Baer-Pontzen, K., Bala-Mohan, S., Choli, D., Zahn, H. & Brandenburg, D. (1989) Biol. Chern. Hoppe-Seyler 370,25 1-258 44. Wente, S. R., Villalba, M., Schramrn, V. L. & Rosen, 0. M. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 2805-2809 45. White, M. F., Shoelson, S. E., Keutmann, H. & Kahn, C. R. (1988)J. Biol. Chern. 263,2969-2980 46. White, M. F., Livingstone, J. N., Backer, J. M., Lauris, V., Dull, T. J., Ullrich, A. & Kahn, C. R. (1988) Cell (Cambridge, Mass.) 54,641-649 47. Whittaker, J. & Okarnoto, A. (1988) J. Biol. Chern. 263,3063-3066 48. Wilden, P. A., Backer, J. M., Kahn, C. R., Cahill, D. A., Schroeder, G. J. & White, M. F. (1990) Proc. Natl. Acad. Sci. USA. 87,3358-3362 49. Yarnarnoto-Honda, R., Koshio, O., Tobe, K., Shibasaki, Y., Monornura, K., Odawara, M.,

1991

43 I

Biochemical Society Transactions

Kadowaki, T., Takaku, F., Akanuma, Y. & Kasuga, M. (1990)J. Biol. Chem. 265,14777-14783

432

50. Yonezawa, K., Pierce, S., Stover, C., Aggerbeck, M., Rutter, W. J. & Roth, R. A. (1991) in Insulin, IGFs and Their Receptors (Razada, M. & Le-Roith, D.,

eds.),Plenum Publishing Co., New York, in the press 51. Zhang, B., Tavarb, J. M., Ellis, L. & Roth, R. A. (1991) J. Biol. Chem. 266,990-996 Received 21 December 1990

Fluorescence approaches to the study of the p21fa GTPase mechanism John F. Eccleston,* Keith J. M. Moore,* Geoffrey G. Brownbridge,* Martin R. Webb* and Peter N. Lowet *Division of Physical Biochemistry, National Institute for Medical Research, Mill Hill, London NW7 I AA, U.K., tDepartment of Molecular Sciences, Wellcome Research Laboratories, Langley Court, Beckenham, Kent BR3 385, U.K.

Introduction The protein products of the N-, K- and Ha-ras genes, termed p21", form part of a family of low molecular mass guanine-nucleotide-binding proteins which have high sequence homology [1-41. They bind GTP and GDP tightly and have low intrinsic GTPase activity. p21 rm proteins have been extensively studied because some single point mutants are found in a high proportion of cancers. In common with other GTPases, they appear to exist in a biologically active conformation when bound to GTP, and hydrolysis to the GDP complex causes deactivation of the biological effect. Their physiological role, however, is not known. By analogy with the heterotrimeric G-proteins, it has been generally assumed that they form part of a signal-transducing pathway, transmitting a signal from a cell-surface receptor to an intracellular effector molecule, although it has recently been suggested that they may mediate the assembly of macromolecular complexes in membranes [31. T o understand the mechanisms by which the concentrations of active and inactive complexes of p21" are regulated in the cell, we are investigating the elementary steps of the p2lr0"GTPase in solution. We aim to define the important intermediates in the process, measure their rates of interconversion and monitor at which stages structural changes occur which may be related to biological activity. This work complements the X-ray diffraction studies of crystals of p21 which have defined structural differences between the GTP- and GDPbound forms of p21" and between normal and transforming mutants [S, 61, and will form a basis for understanding how other cellular components Abbreviations used: GAP, GTPase-activating protein; mant, 2'(3')-0-( N-methy1)anthraniloyl; DTT, dithiothreitol.

Volume 19

such as the GTPase-activating protein (GAP) regulate the GTPase cycle. Most studies, including our own, have been on p21" expressed in Escherichziz coli without the post-translational modifications which occur in vivo [7]. Unless otherwise stated, all of the work described is on p21N-". The basis of our kinetic approach is to use well-defined nucleotide complexes of p21 in the absence of excess nucleotide, so that processes of interest occur under single-turnover conditions. The advantage of this approach is that observed rate constants are independent of the concentration of native p21 rm, a value usually based on filter-binding assays which are subject to systematic errors [8].

Basic GTPase mechanism The simplest scheme for the hydrolysis of GTP by p21rm is shown in scheme 1 where R is p2 I rm,

R + GDP

R $- GTP

(scheme 1) The rate constants, or limits on them, have been measured for the wild-type protein (Gly-12) and two oncogenic mutant proteins (Asp-12 and Val12). The experimental details and rate constants for the Gly-12 and Asp-12 proteins are given in Neal et al. [9], and rate constants for all three proteins are summarized in Table 1. The values of the rate constants show that it is not possible to form p21"eGTP in vitro by simply incubating p21"mGDP with excess GTP under these conditions, since the rate constants of GDP release and GTP cleavage are sufficiently close for significant hydrolysis to occur during the time course of the exchange reaction.