SCIENTIFIC CORRESPONDENCE

For these reasons, it is unlikely that the does not affect our analysis of unfolding as chaperones, whereas unfolding does not unfolding approach will provide new in­ it occurs after the rate-determining step depend on these factors. Far from the sight into the still enigmatic process of in the unfolding direction and accumu­ unfolding approach being unlikely to protein folding. lates only slightly in the transition region provide new insight into protein folding , it JOHANNES BUCHNER (unpublished data). should be clear from the above that un­ lnstitut fiir Biophysik und Physikalische Third, we stated clearly that micro­ folding studies are an essential element in Biochemie, scopic reversibility applies under identical the understanding of folding, and are just Universitat Regensburg, reaction conditions and for a reversible as important as refolding studies. Universitatstrasse 31< process (ref. 1; see also ref. 7, p. R9-90). It A.R. FERSHT D-8400 Regensburg, holds not just for "strongly denaturing J.T. KELLISJR FRG conditions" but for all concentrations of A.T.E.L. ~ATOUSCHEK urea, because folded and unfolded protein l. SERRANO THOMAS KIEFHABER always exist in equilibrium even though MRC Unit for Protein Function and Laboratorium fiir Biochemie, the equilibrium constant may be far from Design, Universitat Bayreuth, unity. At any one of these concentrations, University Chemical Laboratory, Postfach 10 12 51, the transition states for folding and un­ Lensfield Road, D-8580 Bayreuth, folding are the same. The question is Cambridge CB2 1EW, FRG whether or not the transition state changes UK with change of denaturant. Creighton" FERSHT ET AL. REPLY-Any novel analysis has summarized the evidence that the 1. Matouschek. M .. Kellis. J.T .. Serrano. L. & Fersht. A.R . Nature 340. 122-126 (1989). will, and should, attract intense scrutiny. nature of the transition state for folding­ 2. Udgoankar. J.B. & Baldwin, R.L. Nature 335. 694-699 All of the points that have been raised unfolding does not change with denatur­ (1988). 3. Roder. H. . Elove . G.A. & Englander. S.W. Nature 335. have precedents, but we welcome the ant concentration- the rate of unfolding 700--704 (1988). opportunity to discuss them. The analysis changes uniformly with changing condi­ 4. Schmid. F.X. & Baldwin. R.L. J. mo/ec. Bioi. 135, 199- 215 (1979). of a forward pathway by studying the tions but refolding changes non-uniformly 5. Kim . P.S. & Baldwin, R.L. A. Rev. Biochem. 51. 459-489 reverse, for example, is well-established because the nature of the unfolded state (1982) . in physical chemistry and enzymology' and changes'. (This is a principal reason why 6. Kuwajima. K. Proteins 6. 87-103 (1989). 7. Fersht. A.R. Enzyme Structure and Mechanism. 89-90, in protein folding". unfolding kinetics is easier to analyse.) 205-206.237.427 (Freeman. New York, 1985). First, the use of hypothetical thermo­ This laboratory has espoused the need 8. Segawa. S. ·I. & Sugihara. M. Biopo/ymers 23. 24 75- 2488. 2489-2498 (1984). dynamic cycles is perfectly valid and is for determining the full free-energy path­ 9. Rae. S.N .. Singh. U.C .. Bash. P.A. & Kollman. P.A. Nature frequently employed in thermodynamic ways of reactions by measuring rate con­ 328. 551-554 (1987). 10. Kuwajima. K.. Mitani. M. & Sugai. S. J. molec. Bioi. 206. analysis; for example, the classic Born­ stants in both the forward and reverse 54 7- 561 (1989) Haber cycle described in most elementary directions (see studies on the tyrosyl­ 11. Creighton. T.E. Proc. natn. Acad. Sci. U.S.A. 85. 5082- physical chemistry texts. This is allowed tRNA synthetase). We have measured the 5086 (1988). by the first law of thermodynamics, which rate constants for refolding ofbarnase and states that energy changes depend only on many of its mutants and combined these initial and final states and not on the path­ with the unfolding data to characterize the Who's who? way between them. Cycles equivalent to structure of the folding intermediate and SrR-The Haldane beetle story alluded to those we use' are also used by the theoreti­ the transition states by extension of our by Reg Passmore' and previous writers cians for their free-energy calculations protein engineering analysis (manuscript belongs to J. B.S., who made the remark (see, for example, ref. 9). in preparation). The combined data give in 1951 in a lecture on the biological prob­ Second, the data analysis is not based new information on the folding inter­ lems of space flight to the British Inter­ on the assumption that barnase exists in mediate but no more information on the planetary Society, whose journal then only two distinct conformational states. transition state than did the unfolding data reported ie (as discussed in my book For the equilibrium unfolding cycle, the alone. Comparative Social Recognition'). analysis is based simply on the energy dif­ Creighton has stated: "Instead of PATRICK COLGAN ference between initial (folded) and final searching for nucleation sites in unfolded Department of Biology, (unfolded) states and the first law of proteins, it might be more relevant to Queen's University, thermodynamics. Just as the law allows us search for unfolding nucleation events in Kingston, to add hypothetical intermediates, it the native conformation" (ref. 11). For Canada, allows real ones to be ignored. For the this reason , and for the necessity of con­ K7L3N6 kinetic cycle, as explained in the article', structing free-energy profiles, it is an transition-state theory when applied to e~sential element of folding studies to SIR-The Haldane beetle story' was an the unfolding kinetics avoids the problems measure unfolding rate data. Further­ invention by Hutchinson to tease Haldane; of intermediates accumulating, because more, because refolding kinetics is fraught J. B. S. Haldane was the Haldane referred this happens before the rate-determining with complications arising from the pre­ to and the story has no connection what­ step only on the refolding pathway. sence of folding intermediates, cis-trans ever with either Jowett or Oxford. The Incidentally, microcalorimetry is not the isomerization of prolines, and chemical apocrophyl conversation would have only way of detecting folding inter­ processes that occur when the denatured taken place, if it had happened, in a mediates. Kinetics can detect inter­ protein is incubated for many minutes, common room in University College, mediates that are not observed by equili­ unfolding kinetics is simpler to analyse London. brium microcalorimetry and, from and far less prone to artefact. Unfolding ~ARK WILLIAMSON measurements of both folding and unfold­ studies by themselves can be used to Department of Biology, ing rates, either show that a two-state define the last transition state on the re­ University of York, transition is obeyed or that an interme­ folding pathway. Unfolding studies might York Y015DD, diate accumulates (ref. 10, and our un­ also be more relevant than refolding for UK published work). We did indeed, detect relating to phenomena in vivo, because 1. Passmore. R. Nature342, 866 (1989). an intermediate on the refolding pathway the pathway of folding in vivo depends on 2. Slater, A. E. J. Br. interp/anet. Soc. 10. 154-158 (1951). 3. Colgan. P. Comparative Social Recognition. P.10 (Wiley, of barnase by kinetic methods and H-D the sequential formation of proteins NewYork.1983). exchange (see refs 2, 3), but its presence during translation and may be aided by 4. Hutchinson. G.E. Am.Nat. 93. 145 (1959). 602 NATURE · VOL 343 · 15 FEBRUARY 1990 © 1990 Nature Publishing Group