Review R71

Local versus nonlocal interactions in protein folding and stability— an experimentalist’s point of viewVictor Muñoz1 and Luis Serrano

One of the classic important issues in protein foldingand stability is the relative roles of noncovalent short-range (local) and long-range (nonlocal) interactions.Interest in this topic has been reinforced by recentdevelopments in the analytical theory of protein foldingand in lattice-based computer simulations. During thepast few years, a wealth of experimental informationrelevant to this issue has been accumulating. In thisreview, we focus specifically on experimental aspects,discussing some general ideas that arise from theresults obtained by many different groups using avariety of approaches. We also discuss a newexperimental strategy that allows us to engineer thecontribution of local interactions, and we discuss thefirst results obtained.

Address: European Molecular Biology Laboratory, Meyerhofstrasse 1,Heidelberg, D-69012, Germany. 1Present address: Laboratory ofChemical Physics, NIDDK, NIH, Building 5, Room 106, 9000 RockvillePike, Bethesda, Maryland 20892-0520, USA.

Correspondence: Luis Serranoe-mail: serrano@embl-heidelberg.de

Electronic identifier: 1359-0278-001-R0071

Folding & Design 01 Aug 1996, 1:R71–R77

© Current Biology Ltd ISSN 1359-0278

IntroductionUnderstanding protein folding and stability involvesunravelling the mechanisms by which natural proteinsattain their special physical properties, as compared torandom heteropolymers. To explain these properties, ithas been theoretically proposed that an energy gap existsbetween the native and any of the other possible confor-mations of a protein [1,2]. This gap requires the native setof noncovalent interactions between amino acids to bemore stabilizing than those in all other conformations.There are two types of noncovalent interaction in pro-teins: local and nonlocal, so defined in terms of the dis-tance in the sequence between the interacting residues[3]. Importantly, they are geometrically different; localinteractions participate in defining secondary structurewhile nonlocal interactions are involved in defining thetertiary structure. Both types of interaction can be eithersequence independent, those between atoms of the back-bone (i.e �-helix mainchain–mainchain hydrogen bond),or dependent on the chemical nature of the residuesinvolved. The latter group is obviously the more relevantto the protein folding problem.

The issue of the specific roles of local and nonlocal inter-actions is a very important one. Classically, local interac-tions have been considered less important for proteinstability and for attaining a definite three-dimensionalstructure [3]. They were, however, assigned an importantrole in driving the conformational search during foldingin a variety of phenomenological models of proteinfolding [4–10]. This view was sustained by the findingthat very short protein fragments could significantly pop-ulate native-like secondary structure in aqueous solution[11]. Nevertheless, other models put the emphasis onhydrophobic collapse as the guiding force in the foldingprocess [12]. More recently, another view of proteinfolding has emerged from statistical-mechanics theoryand computer simulations of folding (for recent reviews,see [1,2,13]). In these new approaches to the problem,however, the putative role of local interactions stillremains unclear. Some simplistic lattice simulations findthat secondary structure is mainly the product of proteincompaction [13]. Optimization of folding speed seems torequire a small contribution of local contacts to the sta-bility of the folded state [14,15]. A similar conclusion hasbeen theoretically drawn for proteins with a nucleation-condensation mechanism [16]. On the other hand, off-lattice simulations have shown that inclusion ofmainchain–mainchain hydrogen bonding is needed toobtain a secondary structure content akin to that foundin proteins [17]. Statistical-mechanical models [1] havealso been used to investigate the role of local helicalsignals in guiding the folding of helical proteins [18,19].In these studies a similar role is found for local and non-local interactions in defining the energy gap [18,19].These topics have been discussed in detail in recentreviews [1,2,13] and will, therefore, not be discussedfurther. In this review, we rather focus on the availableexperimental data that are relevant to the role of localinteractions in protein folding and stability, as well as onnew experiments specifically designed to directlyaddress this issue.

Model systems to experimentally study local interactionsSite-directed mutagenized proteins have been the stan-dard systems of choice to characterize and quantify thefree energy contribution of noncovalent interactions. Inthese studies, a pairwise interaction is analyzed by per-forming a thermodynamic cycle in which the changes infree energy of unfolding are measured for the single andthe double mutant proteins [20]. However, there are somedrawbacks in the use of proteins for this purpose. First, itis not always possible to carry out neat mutations to

monitor only a single interaction. Second, the effect ofmutations on the denatured state of the protein cannot bedetermined. The latter problem becomes more severe onthe study of local interactions because they fix fewdegrees of freedom in the polypeptide chain. They are,then, likely candidates to contribute to the stability of notonly the native state, but also the denatured ensemble. Analternative system to analyze local interactions is the useof synthetic model peptides. These peptides are short andmonomeric and present less context problems than pro-teins. Such systems have been used successfully to dissectthe contribution of local interactions to �-helix stability[21–23]. This has been possible due to two facts: shortpolyalanine-based peptides show important helical popu-lations in aqueous solution and there is a sufficientlyprecise method to quantify changes in helical content [24].On the other hand, short peptides populate large ensem-bles of conformations, which makes it necessary to utilizestatistical-mechanics treatments [21–23]. This usuallyinvolves, as a simplification, the assumption of the lack ofenergetic coupling between residues in the coil state. Bothmethods have been used extensively to study local inter-actions and have provided invaluable information on thethermodynamics of local interactions.

Thermodynamic analysis of local interactionsSecondary structure propensitiesIt is a well known fact that amino acids are found inprotein structure databases with different frequencies indifferent conformations [25]. Until recently, however,the thermodynamic relevance of these frequencies wasobscure. In the past few years, propensities to form �-helices [26–29], �-strand [30–32] and �-turn conforma-tions [33] have been experimentally studied in severalmodel systems. In general, it is observed that the 20amino acids have different propensities; these are nor-mally in the range of 1 kcal mol–1 between the best andthe worst ones for a given conformation (with the excep-tion of proline). In some cases context effects can be pre-dominant and partly mask these propensities [33–35].Overall, the results obtained in proteins and peptides arein good agreement for the relative secondary structurepropensity scales. There are, however, some discrepan-cies in the absolute free energy range [36]. Unfortu-nately, the origin of this discrepancy is not yet known.Interestingly, consistent correlations have been foundbetween the propensities to populate regions of �,�space, as observed in protein structure databases, and thepreviously mentioned thermodynamic scales for propen-sities [36–38]. This suggests that the changes in freeenergy observed in experiments are really related to dif-ferent propensities to populate regions of the �,� space.Different �,� propensities have also been identified asan important contribution to the changes in free energyproduced by Ala→Gly [39], Val→Ala and Val→Gly [40]mutations in proteins. These propensities seem to be

one of the factors determining the differences in NMRparameters measured for the different amino acids inshort random-coil peptides [41–43]. Such a conclusion issuggested by the correspondence between the experi-mental values and those derived from the �,� distribu-tions in protein structure databases [44–46]. The physicalbasis for the existence of different propensities in the 20amino acids is unknown. Maximization of hydrophobicsurface-buried [47] and different patterns of hydrogen-bonding with the solvent [48] have been used as possibleexplanations.

Interactions between residuesLocal interactions between residues placed close insequence (less than five amino acids away) have been sys-tematically analyzed in �-helices. Such studies have beenpossible using peptide model systems with polyalanine-based sequences. The results obtained so far show thatdifferent pairs of amino acids spaced in an �-helix bythree (i,i+3) and by four (i,i+4) residues have very differ-ent free energies of interaction. These energies can beattractive or repulsive, depending on the pair of residues,and have been described as hydrogen bonds, electrostaticinteractions between charged residues, charge–dipoleinteractions, dipole–dipole interactions, van der Waals’and hydrophobic interactions [21–23]. They seem to bestereochemically specific and thermodynamically impor-tant (some pairs have free energies of interaction higherthan 1 kcal mol–1 [49,50]). Other important local interac-tions related to �-helices are found in the ends of thesestructural elements. In this case, residues placed outsideor at the boundary of the �-helix interact with the main-chain (N-capping and C-capping [51–53]) or sidechain(local motifs [54–57]) of residues within the �-helix. Therange of free energies for this type of interaction seems tobe similar to that inside the �-helix [58]. An additionalinterest for these last interactions lies in their involve-ment in defining the �-helix limits. Due to this character-istic, they have been defined as helix start and stopsignals [59,60]. This term might be misleading, though, asit has been experimentally proven that other local interac-tions may be enough to bypass such ‘signals’ [61]. A rolein protein folding, guiding the coalescence between thehelix and the preceding or following chain segment, hasalso been suggested by several authors [56,62]. Localinteractions in other secondary structure elements havenot been so extensively analyzed. For �-turns, a randommutagenesis study suggested that �-turn-like local inter-actions are not thermodynamically relevant [63].However, a more recent analysis by Regan and colleagues[33] shows that mutating a single �-turn residue to each ofthe other 19 amino acids conveys important changes inthe free energy of unfolding of the protein. They alsofound a correlation between the changes in free energyupon mutation and the propensities observed in proteinstructure databases.

R72 Folding & Design Vol 1 No 4

How important are local interactions for natural proteins?We have illustrated above that local interactions can con-tribute to the thermodynamics of proteins. The questionis then, what is the specific role that local interactions playin the special properties of natural proteins? An indirectway to attain some information relevant to this pointcomes from the conformational analysis of protein frag-ments. Short fragments of proteins allow the study of localinteractions in isolation from the protein context. In thisregard, the entire sequence of several proteins has beensplit into short pieces that have been analyzed by circulardichroism (CD) and nuclear magnetic resonance (NMR):lysozyme [64], myohemerythrin [65], plastocyanin [66],thermolysin C-terminal domain [67], bovine pancreatictrypsin inhibitor (BPTI) [68], B1 domain of the IgG-binding protein [69], �-spectrin SH3 domain [70] and thebarley chymotrypsin inhibitor (CI-2) [71]. As a briefsummary, these studies show that in aqueous solutionsome fragments populate in small amounts native-likeconformations, while others do not deviate from random-coil behaviour (for review, see [11]). In a very few casesnon-native tendencies have been found [70]. Therefore,local interactions are more important in some regions ofproteins than in others. On the other hand, these patternsare specific to each protein and are not conserved inprotein families [72] or in proteins with the same three-dimensional structure [73]. Another interesting observa-tion is that the maximum amount of secondary structurefound so far in protein fragments is much less than thatshown by certain designed short heteropolypeptides.Moreover, it has been found that the amount of native-like secondary structure of protein fragments can beincreased greatly by designing two or three polar→polarmutations [74–76]. All of this suggests a negative selectionfor large contributions of local interactions in natural pro-teins. Nevertheless, it is worth mentioning that small pop-ulations of secondary structure do not imply that thecontribution of local interactions to the free energy ofnatural proteins is negligible. In �-helices, due to thecooperativity of the helix/coil transition, small changes inthe helical population of peptides with low helical contentinvolve large changes in free energy [74].

Engineering the ratio between local and nonlocalinteractions in natural proteinsA crucial experiment to be performed is to engineer, as incomputer simulations, the contribution of local interac-tions to the energy function of proteins. Once done, theeffect on protein folding and stability can be determinedwith physico-chemical techniques. Remarkably, suchexperiments are nowadays feasible for local interactionsinvolved in �-helix stability. The helical propensity ofpeptides can be increased and/or decreased by designingspecific mutations that introduce and/or delete local inter-actions [74–76]. These mutations are designed on thebasis of our current knowledge of the helix/coil transition

for heteropolymers [23]. The effective increases in helicalpropensity are, then, directly measured by far-UV CDanalysis of the peptides in solution. A similar strategy isapplied to the entire protein, requiring a previous target-ing of the positions that can be safely mutagenized. Muta-tions must involve only polar residues and the targetedsites must be fully exposed to the solvent and free of con-tacts with other regions of the protein [75]. Importantly,the same mutations are performed in peptides includingthe �-helices to serve as reference for the changesobserved in the protein. This experiment has been carriedout recently in our laboratory for three different proteins:Che Y [75], the activation domain of procarboxypeptidaseA (CPA) [76] and the SH3 domain of chicken spectrin(J Prieto, M Wilmmans, L Serrano, unpublished data).These proteins differ in their fold, size and folding behav-iour (Che Y is an �/� parallel protein of 129 amino acidsand has a folding intermediate [77]; CPA is an �+� proteinwith 80 residues and follows a two-state folding transition[78]; SH3 is an all �-protein with 60 residues and alsofollows a two-state folding transition [79]). In the two firstproteins, each of their native �-helices (five and two,respectively) have been enhanced in their helical propen-sity. In the third protein, non-native helical propensity hasbeen induced in the first �-strand. The following sectionsdiscuss the results obtained so far.

Local interactions and protein stabilityThe stabilization of native �-helices by local interactions(Che Y and CPA; Fig. 1a) produces important changes inthe thermodynamic properties of the protein. In all casesthe protein becomes significantly less sensitive to chemi-cal and temperature denaturation. There is a cleardecrease in the equilibrium m value of the urea-inducedunfolding transition, which is more pronounced in theChe Y mutants (Fig. 1b). The decrease in m has beenidentified in Che Y with the compaction of the dena-tured state and not with the stabilization of a foldingintermediate [77]. The compaction of the denaturedstate reflects its stabilization by local interactions. As aconsequence, the overall gain in protein stability due toenhancing native helical propensities is less than theincrease in free energy for �-helix formation in thepeptide [77,78] (Fig. 2). In some cases, there is even anet loss in protein stability, reflecting the preferentialstabilization of the denatured state. Induction of a non-native �-helical tendency in the SH3 domain of spectrinmainly results in a small change in the m value which isproportional to the helical propensity (J Prieto, M Wilm-mans, L Serrano, unpublished data). This suggests that,in natural proteins, the partition function of the dena-tured ensemble is not dominated by conformationshaving residual native secondary structure, but that con-formations with non-native secondary structure are alsopresent. Another important observation is that the com-paction of the denatured state appears more pronounced

Review Local versus nonlocal interactions in protein folding and stability Muñoz and Serrano R73

the larger the protein is. More proteins need to be ana-lyzed in this way to be able to generalize. Finally, theseresults indicate that there is an optimal local versus non-local ratio in order to maximize protein stability. As aconsequence, the concept of maintaining low local versusnonlocal ratios in de novo protein design strategies hasbeen proposed [77].

Local interactions and the rate of foldingThe study of the folding kinetics of the mutants withengineered ratios of local versus nonlocal interactions pro-vides direct information on the effect of local interactionsin the rate of protein folding. These results are still pre-liminary — Che Y ([77]; E López-Hernández, P Cronet,L Serrano, V Muñoz, unpublished data); CPA (A Viguera,V Villegas, X Aviles, L Serrano, unpublished data); SH3(J Prieto, V Muñoz, M Wilmmans, L Serrano, unpublisheddata) — but some general observations can be drawn. Thecompaction of the denatured state of Che Y, CPA andSH3 is further supported by the fact that the samedecreases in m observed at equilibrium are found kineti-cally (Fig. 1c). Furthermore, the changes appear in theurea dependence of the refolding rate constant. Generally,the folding transition state is not shifted in the reactioncoordinate in any of the mutants. However, the effect ofincreasing native helical propensities (Che Y and CPA) onthe rate of folding is complex, depending on the details ofthe folding mechanism of the protein. Whereas in Che Ythere is always a deceleration of the folding rate, in thesecond helix of CPA a strong acceleration of the foldingrate is observed. The critical point appears to be the struc-tural properties of the folding transition state as comparedto the structure of the ground states. In other words, if thestabilized helix is more structured in the transition statethan in the denatured state and/or the folding intermedi-ate (only for Che Y, as CPA follows a two-state model), therate of folding is accelerated. If the degree of structure issimilar the rate is not affected, but it is effectively reducedwhen the target helix is more structured in the groundstate than in the folding transition state. An interestingconsequence is that for proteins with a nucleation-conden-sation event as rate-limiting step (only the folding nucleusis formed in the transition state; Che Y in this analysis[80]), the general effect of enhancing native �-helices willbe a deceleration in the rate of folding, as has been postu-lated recently [16].

R74 Folding & Design Vol 1 No 4

Figure 1

Local native interactions produce less cooperative folding transitions.(a) CD spectra of a peptide corresponding to the wild-type �-helix A ofChe Y and a mutant peptide in which the helix propensity has beenenhanced by solvent-exposed polar→polar mutations. (b) Ureadenaturation profiles of wild-type Che Y and of a mutant protein inwhich the helical propensity has been enhanced by local interactions in�-helix A [77]. It can be appreciated that although the mutant protein ismore resistant to urea denaturation, the slope of the urea dependenceis smaller. (c) Kinetic analysis of unfolding and refolding reactions ofthe same two proteins as in (b). It can be appreciated that the slope ofthe natural logarithm of the urea dependence is smaller for the helix-enhanced mutant in the refolding reaction. There is no change in theunfolding reaction because �-helix 1 is made in the transition state ofChe Y [77,80]. For other helices of this protein, such as helices C, Dand E which are unfolded in the transition state [77,80], the unfoldingreaction slows down considerably when native-like helical propensitiesare enhanced. Qualitatively similar results are found for CPA.

-3

-2

-1

0

1

2

3

4

2 4 6 8 10

ln (λ

2)

[Urea] (M)

(c)

Helix stabilized

Wild-type

-4 104

-2 104

0

2 104

4 104

6 104

8 104

190 200 210 220 230 240 250

Mea

n re

sidu

e el

liptic

ity (d

eg.c

m2.d

mol

-1)

Wavelength (nm)

(a)

Wild-type peptide

Helix stabilized peptide

1.00

1.50

2.00

2.50

3.00

0 2 4 6 8

Fluo

resc

ence

(a.u

.)

[Urea] (M)

(b)

Helix stabilizedWild-type

ConclusionsAll these studies suggest that local interactions are ther-modynamically important. However, new experimentssuggest that they have low specificity for the native state,also stabilizing alternative conformations. This simpleinterpretation explains the effects observed in bothprotein stability and protein folding. Besides, it gives aclue as to why natural proteins have such low helicalpropensities, as compared to what is attainable with syn-thetic peptides. Finally, protein engineering and designprojects ought to be concerned with the balance betweenlocal and nonlocal interactions as well as minimizing non-native local interactions in order to optimize protein stabil-ity and folding speed.

References1. Bryngelson, J.D., Onuchic, J.N., Socci, N.D. & Wolynes, P.G. (1995).

Funnels, pathways and the energy landscape of protein folding: a syn-thesis. Proteins 21, 167–195.

2. Karplus, M. & Sali, A. (1995). Theoretical studies of protein folding andunfolding. Curr. Opin. Struct. Biol. 5, 58–73.

3. Dill, K.A. Dominant forces in protein folding. Biochemistry 29,7133–7155.

4. Anfinsen, C.B. & Scheraga, H.A. (1975). Experimental and theoreticalaspects of protein folding. Adv. Protein Chem. 29, 205–300.

5. Wetlaufer, D.B. (1973). Nucleation, rapid folding and globular intra-chain regions in proteins. Proc. Natl. Acad. Sci. USA 70, 697–701.

6. Karplus, M. & Weaver, D.L. (1976). Protein-folding dynamics. Nature260, 404–406.

7. Karplus, M. & Weaver, D.L. (1994). Protein folding dynamics. The dif-fusion-collision model and experimental data. Protein Sci. 3, 650–668.

8. Harrison, S.C. & Durbin, R. (1985). Is there a single pathway for thefolding of a polypeptide chain? Proc. Natl. Acad. Sci. USA 82,4028–4030.

9. Wright, P.E., Dyson, H.J & Lerner, R.A. (1988). Conformation ofpeptide fragments of proteins in aqueous solution: implications for initi-ation of protein folding. Biochemistry 27, 7167–7175

10. Kim, P.S. & Baldwin, R.L. (1990). Intermediates in the folding reactionsof small proteins. Annu. Rev. Biochem. 59, 631–660.

11. Dyson, H.J. & Wright, P.E. (1993). Peptide conformation and proteinfolding. Curr. Opin. Struct. Biol. 3, 60–65.

12. Ptitsyn, O.B. (1993). Stage mechanism of the self-organization ofprotein molecules. Dokl. Acad. Nauk. 210, 1213–1215.

13. Dill, K.A., et al., & Chan H.S. (1995). Principles of protein folding. Aperspective from simple exact models. Protein Sci. 4, 561–602.

14. Abkevich, V.I., Gutin, A.M. & Shakhnovich, E.I. (1995). Impact of localand non-local interactions on thermodynamics and kinetics or proteinfolding. J. Mol. Biol. 252, 460–471.

15. Govindarajan, S. & Goldstein, R.A. (1995). Optimal local propensitiesfor model proteins. Proteins 22, 413–418.

16. Fersht, A.R. (1995). Optimization of rates of protein folding: the nucle-ation-condensation mechanism and its implications. Proc. Natl. Acad.Sci. USA 92, 10869–10873.

17. Hunt, N.G., Gregoret, L.M. & Cohen, F.E. (1994). The origins ofprotein secondary structure: effects of packing density and hydrogenbonding studied by a fast conformational search. J. Mol. Biol. 241,214–225.

18. Luthey-Schulten, Z., Ramirez, B.E. & Wolynes, P.G. (1995). Helix-coil,liquid crystal and spin glass transitions of a collapsed heteropolymer.J. Phys. Chem. 99, 2177–2185.

19. Saven, J.G. & Wolynes, P.G. (1996). Local conformational signals andthe statistical thermodynamics of collapsed helical proteins. J. Mol.Biol. 257, 199–216.

20. Serrano, L. & Fersht, A.R. (1993). Principles of protein stability fromprotein engineering experiments. Curr. Opin. Struct. Biol. 3, 75–83.

21. Scholtz, J.M. & Baldwin, R.L. (1992). The mechanism of alpha-helix for-mation by peptides. Annu. Rev. Biophys. Biomol. Struct. 21, 95–118.

22. Chakrabartty, A. & Baldwin, R.L. (1995). Stability of �-helices. Adv.Protein Chem. 46, 141–176.

23. Muñoz, V. & Serrano, L. (1995). Helix design, prediction and stability.Curr. Opin. Biotechnol. 6, 382–386.

Review Local versus nonlocal interactions in protein folding and stability Muñoz and Serrano R75

Figure 2

Schematic diagram showing the effect of modifying local or nonlocalinteractions. (a) Residues that are establishing local contacts in thenative state could probably, because of entropic reasons, also bemaking the same interactions in some of the conformations of thedenatured ensemble of a protein. Therefore, introducing favourablelocal interactions will stabilize not only the native state, but also asignificant fraction of the conformations in the denatured ensemble.This results in a smaller increase in protein stability than expected. Ifthose conformations that have native-like secondary structures aremore compact on average than the others, then a compactization ofthe denatured ensemble should occur. (b) Residues that establishnonlocal interactions are likely to be separated in the denaturedensemble. Therefore, introducing favourable nonlocal interactionsshould mainly stabilize the native state. However, care must be takennot to introduce favourable non-native local contacts that stabilize non-native secondary structure conformations, as these can also produce acompactization of the denatured ensemble.

Native

Denatured ensemble(a) Local interactions

Native

Denatured ensemble(b) Nonlocal interactions

A

A

A

A

A

A

A

A

B

B

B

B

B

BB

B

24. Chen, Y.H., Yang, J.T. & Chow, K.H. (1974). Determination of the helixand � form of proteins in aqueous solution by circular dichroism. Bio-chemistry 13, 3350–3359.

25. Chou, P. & Fassman, G.D. (1978). Empirical predictions of proteinconformation. Annu. Rev. Biochem. 47, 251–276.

26. O’Neil, K.T. & DeGrado, W.F. (1990). A thermodynamic scale for thehelix-forming tendencies of the commonly occurring amino acids.Science 250, 646–650.

27. Horovitz, A., Matthews, J. & Fersht, A.R. (1992). �-Helix stability in pro-teins. II. Factors that influence stability at an internal position. J. Mol.Biol. 227, 560–568.

28. Blaber, M., Zang, X. & Matthews, B. (1993). Structural basis of aminoacid �-helix propensity. Science 260, 1637–1640.

29. Chakrabartty, A., Kortemme, T. & Baldwin, R.L. (1994). Helix propensi-ties of the amino-acids measured in alanine-based peptides withouthelix-stabilizing side chain interactions. Protein Sci. 3, 843–852.

30. Kim, C.A. & Berg, J.M. (1993). Thermodynamic �-sheet propensitiesmeasured using a zinc-finger host peptide. Nature 362, 267–270.

31. Minor, D.L. & Kim, P.S. (1993). Measurement of the �-sheet-formingpropensities of amino acids. Nature 367, 660–663.

32. Smith, C.K., Withka, J.M. & Regan, L.A. (1994). Thermodynamic scalefor the �-sheet forming tendencies of the amino acids. Biochemistry33, 5510–5517.

33. Predki, P.F., Agrawal, V., Brunger, A.T & Regan, L. (1996). Amino-acidsubstitutions in a surface turn modulate protein stability. Nat. Struct.Biol. 3, 54–58.

34. Minor, D.L. & Kim, P.S. (1994). Context is a major determinant of �-sheet propensity. Nature 371, 264–267.

35. Minor, D.L. & Kim, P.S. (1996). Context-dependent secondary struc-ture formation of a designed sequence. Nature 380, 730–734.

36. Muñoz, V. & Serrano, L. (1994). Intrinsic secondary structure propensi-ties of the amino acids, using statistical phi-psi matrices. Comparisonwith experimental scales. Proteins 20, 301–311.

37. Stites, W.E. & Pranata, J. (1995). Empirical evaluation of the influenceof side chains on the conformational entropy of the polypeptide back-bone. Proteins 22, 132–140.

38. Swindells, M.B., MacArthur, M.W. & Thornton, J.M. (1995). Intrinsic�,� propensities of amino acids, derived from the coil regions ofknown structures. Nat. Struct. Biol. 2, 596–603.

39. Lopez, E. & Serrano, L. (1995). Empirical correlation for the replace-ment of Ala by Gly: importance of amino acid secondary intrinsicpropensities. Proteins 22, 340–349.

40. Otzen, D.E & Fersht, A.R. (1995). Side-chain determinants of beta-sheet stability. Biochemistry 34, 5718–5724

41. Wüthrich, K. (1986). NMR of Proteins and Nucleic Acids. John Wiley& Sons, New York.

42. Wishart, D.S., Bigam, C.G., Holm, A., Hodges, R.S. & Sykes, B.D.(1995). 1H, 13C and 15N random coil chemical shifts of the commonamino acids. I Investigation of nearest-neighbor effects. J. Biomol.NMR 5, 67–81.

43. Merutka, G., Dyson, H.J. & Wright, P.E. (1995). ‘Random coil’ 1Hchemical shifts obtained as a function of temperature and trifluo-roethanol concentration for the peptide series GGXGG. J. Biomol.NMR 5, 14–24.

44. Serrano, L. (1995). Comparison between the � distribution of theamino acids in the protein database and NMR data indicates thatamino acids have various � propensities in the random-coil conforma-tion. J. Mol. Biol. 254, 322–333.

45. Smith, L.J., Bolin, K.A., Schwalbe, H., MacArthur, M.W., Thornton, J.M.& Dobson, C.M. (1996). Analysis of main chain torsion angles in pro-teins. Prediction of NMR coupling constants for native and random coilconformations. J. Mol. Biol. 255, 494–506.

46. Flebig, K.M., Schwalbe, H., Buck, M., Smith, L.J. & Dobson, C.M.(1996). Toward a description of the conformations of denatured statesof proteins. Comparison of a random coil model with NMR measure-ments. J. Phys. Chem. 100, 2661–2666.

47. Bundi, A. & Wüthrich, K. (1979). 1H-NMR parameters of the commonamino acid residues measured in aqueous solution of the lineartetrapeptides H-Gly-Gly-X-L-Ala-OH. Biopolymers 18, 285–298.

48. Bai, Y. & Englander, S.W. (1994). Hydrogen bond strength and �-sheet propensities: the role of a side chain blocking effect. Proteins18, 262–266.

49. Viguera, A.R. & Serrano, L. (1995). Side-chain interactions betweensulfur-containing amino acids and Phe in �-helices. Biochemistry 34,8771–8779.

50. Stapley, B.J., Rohl, C.A. & Doig, A.J. (1995). Addition of side chaininteractions to modified Lifson–Roig helix-coil transition theory: appli-

cations to energetics of Phe–Met interactions. Protein Sci. 4,2383–2391.

51. Serrano, L., Sancho, J., Hirshberg, J.M. & Fersht, A.R. (1992). �-Helixstability in proteins. I. Empirical correlations concerning substitution ofside chains at the N- and C-caps and the replacement of alanine byglycine or serine at solvent-exposed surfaces. J. Mol. Biol. 227,544–549.

52. Bell, J.A., Becktel, W.J., Sauer, U. Baase, W.A. & Matthews, B.W.(1992). Dissection of helix capping in T4 lysozyme by structural andthermodynamical analysis of six aminoacid substitutions at Thr59. Bio-chemistry 31, 3590–3596.

53. Doig, A.J. & Baldwin, R.L. (1994). N- and C-capping preferences for all20 amino acids in alpha-helical peptides. Protein Sci. 4, 1325–1336.

54. Zhou, H.X., Lyu, P., Wemmer, D.E. & Kallenbach, N.R. (1994). �-Helixcapping in synthetic model peptides by reciprocal side chain–mainchain interactions: evidence for an N-terminal “capping box”. Proteins18, 1–7.

55. Forood, B., Feliciano, E.J. & Nambiar, K.P. (1993). Stabilization of �-helical structures in short peptides via end capping. Proc. Natl. Acad.Sci. USA 90, 838–842.

56. Muñoz, V., Blanco, F. & Serrano, L. (1995). The “hydrophobic-staple”motif. A role for loop-residues in �-helix stability and protein folding.Nat. Struct. Biol. 2, 380–385.

57. Viguera, A.R. & Serrano, L. (1995). Experimental analysis of the Schell-man motif. J. Mol. Biol. 251, 150–160.

58. Muñoz, V. & Serrano, L. (1995). Analysis of i,i+5 and i,i+8 hydropho-bic interactions in a helical model peptide bearing the hydrophobicstaple motif. Biochemistry 34, 15301–15306.

59. Harper, E.T. & Rose, G.D. (1993). Helix stop signals in proteins andpeptides: the capping box. Biochemistry 32, 7605–7609.

60. Aurora, R., Srinivasan, R. & Rose, G.D. (1995). Rules for �-helix termi-nation by glycine. Science 264, 1126–1129.

61. Jimenez, M.A., Muñoz, V., Rico, M. & Serrano, L. (1994). Helix stop andstart signals in peptides and proteins: the capping box does not nec-essarily prevent helix elongation. J. Mol. Biol. 242, 487–496.

62. Seale, J.W., Srinivasan, R. & Rose, G.D. (1994). Sequence determi-nants of the capping box, a stabilizing motif at the N-termini of �-helices. Protein Sci. 3, 1741–1745.

63. Brunet, A.P., Huang, E.S., Huffine, M., Loeb, J.E., Weltman, R.J. &Hecht, M.H. (1993). The role of turns in the structure of an alpha-helical protein. Nature 364, 355–358.

64. Segawa, S.I., Fukuno, T., Fujiwara, K. & Noda, Y. (1991). Local struc-tures in unfolded lysozyme and correlation with secondary structuresin the native conformation: helix-forming or breaking propensity ofpeptide segments. Biopolymers 31, 497–509.

65. Dyson, H.J., Merutka, G., Waltho, J.P., Lerner, R.A. & Wright, P.E.(1993). Folding of peptide fragments comprising the completesequence of proteins. Models for initiation of protein folding. I Myohe-merythrin. J. Mol. Biol. 226, 795–817.

66. Dyson, H.J., Sayre, J.R., Merutka, G., Shin, H.-C., Lerner, R.A. &Wright, P. (1992). Folding of peptide fragments comprising the com-plete sequence of proteins. Models for initiation of protein folding II.Plastocyanin. J. Mol. Biol. 226, 819–835.

67. Jiménez, M.A., Bruix, M., González, C., Blanco, F.J., Nieto, J.L., Herranz,J. & Rico, M. (1993). CD and 1H-NMR studies on the conformationalproperties of peptides fragments from the C-terminal domain of ther-molysin. Eur. J. Biochem. 211, 569–581.

68. Kemmink, J. & Creighton, T.E. (1993). Local conformations of peptidesrepresenting the entire sequence of bovine pancreatic trypsin inhibitorand their roles in folding. J. Mol. Biol. 234, 861–878.

69. Blanco, F.J. & Serrano, L. (1995). Folding of protein G B1 domainstudied by the conformational characterization of fragments compris-ing its secondary structure elements. Eur. J. Biochem. 230, 634–649.

70. Viguera, A.R., Jimenez, M.A., Rico, M. & Serrano, L. (1995). Conforma-tional analysis of peptides corresponding to �-hairpins and a �-sheet,that represent the entire sequence of �-spectrin SH3-domain. J. Mol.Biol. 255, 507–521.

71. Itzhaki, L.S., Neira, J.L., De Prat Gay, G., Ruiz-Sanz, J. & Fersht, A.R.(1995). Search for nucleation sites in smaller fragments of chy-motrypsin inhibitor 2. J. Mol. Biol. 254, 250–259.

72. Muñoz, V., Blanco, F. & Serrano, L. (1995). The distribution of �-helixpropensity along the polypeptide chain is not conserved in proteinsfrom the same family. Protein Sci. 4, 1577–1586.

73. Muñoz, V., Jimenez, M.A. Rico, M. & Serrano, L. (1995). Structuralanalysis of peptides encompassing all the �-helices of three �/� paral-lel proteins: CheY, flavodoxin and P21-ras: implications for �-helix sta-bility and the folding of �/� parallel proteins. J. Mol. Biol. 247,

R76 Folding & Design Vol 1 No 4

648–669.74. Muñoz, V. & Serrano, L. (1995). Elucidating the folding problem of �-

helical peptides using empirical parameters, II. Helix macrodipoleeffects and rational modification of the helical content of natural pep-tides. J. Mol. Biol. 245, 275–296.

75. Muñoz, V., Cronet, P., López-Hernández, E. & Serrano, L. (1996).Analysis of the effect of local interactions on protein stability. Folding &Design 1, 167–178; 1359-0278-001-00167.

76. Villegas, V., Viguera, A.R., Aviles, F.X. & Serrano, L. (1996). Stabiliza-tion of proteins by rational design of �-helix stability using helix/coiltransition theory. Folding & Design 1, 29–34; 1359-0278-001-00029.

77. Muñoz, V., Lopez, E. & Serrano, L. (1994). Kinetic characterization ofthe chemotactic protein from E. coli CheY. Kinetic analysis of theinverse hydrophobic effect. Biochemistry 33, 5858–5866.

78. Villegas, V., et al., & Serrano, L. (1995). Evidence for a two-state tran-sition in the folding process of the activation domain of human procar-boxypeptidase A2. Biochemistry 34, 15105–15110.

79. Viguera, A.R., Martinez, J.C., Filimonov, V., Mateo, P. & Serrano, L.(1994). Thermodynamic and kinetic analysis of the SH3 domain ofspectrin. Biochemistry 33, 2142–2150.

80. López-Hernández, E. & Serrano, L. (1996). Structure of the transitionstate for folding of the 129 aa protein CheY resembles that of asmaller protein, CI-2. Folding & Design 1, 43–55; 1359-0278-001-00043.

Review Local versus nonlocal interactions in protein folding and stability Muñoz and Serrano R77