doi:10.1016/j.jmb.2011.02.053 J. Mol. Biol. (2011) 408, 514–528

Contents lists available at www.sciencedirect.com

Journal of Molecular Biologyj ourna l homepage: ht tp : / /ees .e lsev ie r.com. jmb

Contribution of Hydrophobic Interactions toProtein Stability

C. Nick Pace1,2⁎, Hailong Fu1, Katrina Lee Fryar2,John Landua2, Saul R. Trevino2, Bret A. Shirley1,Marsha McNutt Hendricks1, Satoshi Iimura2, Ketan Gajiwala1,J. Martin Scholtz1,2 and Gerald R. Grimsley2

1Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843, USA2Department of Molecular and Cellular Medicine, Texas A&M Health Science Center, College Station,TX 77843-1114, USA

Received 14 December 2010;received in revised form21 February 2011;accepted 23 February 2011Available online4 March 2011

Edited by K. Kuwajima

Keywords:hydrophobic interactions;hydrogen bonds;conformational entropy;protein stability;large proteins;small proteins

*Corresponding author. DepartmenCellular Medicine, Texas A&M HealReynolds Medical Building, College1114, USA. E-mail address: nickpacAbbreviations used: VHP, villin h

DSE, denatured state ensemble; WT

0022-2836/$ - see front matter © 2011 E

Our goal was to gain a better understanding of the contribution ofhydrophobic interactions to protein stability. We measured the change inconformational stability, Δ(ΔG), for hydrophobic mutants of four proteins:villin headpiece subdomain (VHP) with 36 residues, a surface protein fromBorrelia burgdorferi (VlsE) with 341 residues, and two proteins previouslystudied in our laboratory, ribonucleases Sa and T1. We compared ourresults with those of previous studies and reached the followingconclusions: (1) Hydrophobic interactions contribute less to the stabilityof a small protein, VHP (0.6±0.3 kcal/mol per –CH2– group), than to thestability of a large protein, VlsE (1.6±0.3 kcal/mol per –CH2– group). (2)Hydrophobic interactions make the major contribution to the stability ofVHP (40 kcal/mol) and the major contributors are (in kilocalories permole) Phe18 (3.9), Met13 (3.1), Phe7 (2.9), Phe11 (2.7), and Leu21 (2.7). (3)Based on the Δ(ΔG) values for 148 hydrophobic mutants in 13 proteins,burying a –CH2– group on folding contributes, on average, 1.1±0.5 kcal/mol to protein stability. (4) The experimental Δ(ΔG) values for aliphaticside chains (Ala, Val, Ile, and Leu) are in good agreement with their ΔGtrvalues from water to cyclohexane. (5) For 22 proteins with 36 to 534residues, hydrophobic interactions contribute 60±4% and hydrogen bondscontribute 40±4% to protein stability. (6) Conformational entropycontributes about 2.4 kcal/mol per residue to protein instability. Theglobular conformation of proteins is stabilized predominantly by hydro-phobic interactions.

© 2011 Elsevier Ltd. All rights reserved.

t of Molecular andth Science Center, 440Station, TX 77843-e@tamu.edu.eadpiece subdomain;, wild type.

lsevier Ltd. All rights reserve

Introduction

By the late 1930s, the structure of globularproteins was beginning to be understood, andBernal1 concluded,1 “Ionic bonds are plainly outof the question… the hydrophobe groups of theproteins must hold it together.… the proteinmolecule in solution must have its hydrophobegroups out of contact with water, that is, in contactwith each other,… In this way a force of association

d.

mailto:nickpace@tamu.eduhttp://dx.doi.org/10.1016/j.jmb.2011.02.053

515Hydrophobic Interactions to Protein Stability

is provided which is not so much that of attractionbetween hydrophobe groups, which is alwaysweak, but that of repulsion of the groups out ofthe water medium.” The importance of thissuggestion was not widely appreciated until1959 when Kauzmann's influential review waspublished.2 He presented convincing evidencethat “the hydrophobic bond is probably one of themore important factors involved in stabilizing thefolded configuration in many native proteins.” Thiskindled Tanford's long-term interest in the hydro-phobic effect, and he used the limited modelcompound data available in 1962 to show that3

“…the stability of the native conformation in watercan be explained… entirely on the basis of thehydrophobic interactions of the nonpolar parts ofthe molecule.” Tanford has written an interestinghistory of the hydrophobic effect.4

In 1971, Tanford and Nozaki published the firsthydrophobicity scale, using ethanol as amodel for theinterior of the protein.5 Other hydrophobicity scalesfollowed using octanol,6,7 N-methylacetamide,8 andcyclohexane9 as models of the protein interior. Theapplications and reliability of these scales have beendiscussed.10–13

In 1987, Yutani's groupwas the first to gain a betterunderstanding of the contribution of the hydropho-bic effect to protein stability, by studying hydropho-bicmutants of tryptophan synthase α subunit.14 Thiswas followed by similar studies of other proteins:barnase,15–17 staph nuclease,18,19 gene V protein,20

T4 lysozyme,21–23 chymotrypsin inhibitor 2,24

FK506-binding protein,25 human lysozyme,26 fibro-nectin type III domains,27 ubiquitin,28 Sac7d andSso7d,29 and apoflavodoxin.30 These studies showedthat the contribution of a buried nonpolar side chain

Fig. 1. Ribbon diagrams of theVHP (a) and VlsE (b) structuresshowing the mutated residues. Thefigures were generated using theYASARA program103 and the crystalstructures ofVHP (PDB code: 1YRI)35

and VlsE (PDB code: 1L8W).36

516 Hydrophobic Interactions to Protein Stability

to protein stability depends on two factors: first, thehydrophobicity of the side chain and the amount ofside-chain surface area removed from contact withwater when the protein folds, and second, the vander Waals interactions of the side chain when theprotein is folded and unfolded. The importanceof hydrophobicity had long been recognized,2,31 butit became clear only later that van der Waalsinteractions may make an even larger contributionbecause of the tight packing in the interior of foldedproteins.19,28,32,33

It seems likely that the stabilizing and destabiliz-ing forces that contribute to protein stability willdepend on protein size. As globular proteinsbecome smaller, it will not be possible to completelybury hydrophobic side chains, and the charged sidechains will be closer together and potentially indifferent environments than in a larger protein. Inaddition, the denatured state ensembles (DSEs)might depend on the size of the protein, and thiscould influence protein stability.In this study, we examine the contribution of

hydrophobic interactions to the stability of a smallprotein, villin headpiece subdomain (VHP) with 36residues,34,35 and of a large protein, Borrelia burg-dorferi protein (VlsE) with 341 residues.36,37 Themutants studied are superimposed on the structuresin Fig. 1. In addition, we have studied thecontribution of hydrophobic interactions to thestability of two proteins previously studied in ourlaboratory, ribonuclease Sa (RNase Sa) and ribonu-clease T1 (RNase T1). Aided by previous results,these new data allow us to gain an improvedunderstanding of the contribution of hydrophobicinteractions to protein stability.

0.0 0.5 1.0 1.5 2.0 2.5

0

20

40

60

80

100

% F

olde

d

Urea (M)

Fig. 2. Percent folded protein as a function of ureaconcentration for WT VlsE (▲) and the hydrophobicvariants I128V (♦), I306V (■), I154V (○), I210V (●), andI214V (Δ) in 5 mM sodium phosphate (pH 7.0), 25 °C.

Results

VlsE

Jones and Wittung-Stafshede previously studiedthe denaturation of VlsE by urea and guanidinehydrochloride (GuHCl), and found it to be com-pletely reversible, as did we for urea denaturation.37

They followed the unfolding with both fluorescenceand circular dichroism and got identical results thatwere consistent with a two-state folding mecha-nism. They also reported kinetic data that wereconsistent with a two-state mechanism. Based ontheir results and conclusions, we have assumed atwo-state folding mechanism for VlsE for theanalysis of our results.37

To study the contribution of the hydrophobiceffect to the stability of VlsE, we prepared five Ile-to-Val mutants (Fig. 1b). All of these Ile residues arecompletely buried, and the mutants contain one less–CH2– group than the wild type (WT). Urea

denaturation curves were determined by measuringthe circular dichroism at 220 nm as a function of ureaconcentration and analyzed using the linear extrap-olation method. Typical results are shown inSupplementary Data. The unfolding curves for WTVlsE and the five mutants are shown in Fig. 2, andthe results are given in Table 1. The mutants were allless stable than theWT, and the decreases in stabilityranged from 1.1 to 1.9 kcal/mol, with an average of1.6±0.3 kcal/mol.

VHP

McKnight et al. were the first to report studies onthe thermal and GuHCl denaturation of the VHPsubdomain.34 Since then, many studies on differentaspects of the folding of VHP and related variantshave been published (for recent examples, seeRefs. 38 and 39). Our VHP has an N-terminal Metand is identical to the protein designated HP-36 byMcKnight et al.34 We also found the thermal, urea,and GuHCl denaturation to be completely revers-ible, and our Tm=74.4 °C was in good agreementwith their value of 72 °C determined under differentconditions. They found a stability of 3.3±0.4 kcal/mol from GuHCl denaturation at 4 °C, and wefound a stability of 2.7 kcal/mol from both urea andGuHCl denaturation at 25 °C.34

To study the contribution of the hydrophobic effectto the stability of VHP, we prepared the followingnine mutants: L2A, F7A, V10A, F11A, M13A, F18A,L21A, L29A, and L35A (Fig. 1a). Urea denaturationcurves were determined by measuring circulardichroism of VHP at 222 nm as a function of ureaconcentration, and analyzed using the linear extrap-olation method. The typical results are shownin Supplementary Data. The results from the studies

image of Fig. 2

Table 1. Parameters characterizing the urea unfolding of VlsE and hydrophobic variants in 5 mM sodium phosphate(pH 7.0), 25 °C

VariantUrea1/2

a

(M)ΔUrea1/2

b

(M)mc

(cal mol−1 M−1)ΔG (H2O)

d

(kcal mol−1)ΔΔGe

(kcal mol−1)

WT 1.19 — 3865 4.6 —I128V 0.88 −0.31 3360 3.0 −1.1I154V 0.70 −0.49 3350 2.3 −1.8I210V 0.67 −0.52 3770 2.5 −1.9I214V 0.65 −0.54 3480 2.3 −1.9I306V 0.77 −0.42 3700 2.8 −1.5

The urea denaturation curves were analyzed as described.98a Midpoint of the unfolding curve. The error is ±2%.b ΔUrea1/2=Urea1/2 (variant)−Urea1/2 (WT).c The slope of plots of ΔG versus [Urea]. The error is ±10%.d ΔG (H2O)=the intercept of plots of ΔG versus [Urea] at 0 M urea.e From ΔUrea1/2×the average m value of WT and the variants (3590 cal mol

−1 M−1). The negative values indicate a decreasein stability.

517Hydrophobic Interactions to Protein Stability

of the five mutants using urea denaturation aresummarized in Table 2. The stabilities of all ninemutants were measured using thermal denaturation,and the results are given in Table 3. TheΔ(ΔG) valuesfrom urea and thermal denaturation curves were inreasonable agreement, and we used the average ofthe two for the mutants where both are available.

RNase Sa

We have previously used RNase Sa to study manydifferent aspects of protein stability and folding.40–45

To study the contribution of the hydrophobic effectto the stability of RNase Sa, we prepared five Ile-to-Val and five Leu-to-Ala mutants. Three thermaldenaturation curves were determined for eachmutant, and the average results are summarized inTable 4.

RNase T1

Like RNase Sa, RNase T1 is a member of themicrobial ribonuclease family that we have used

Table 2. Parameters characterizing the urea unfolding of VH(pH 7.0), 25 °C

VariantUrea1/2

a

(M)ΔUrea1/2

b

(M) (ca

WT 6.17 —L2A 3.22 −2.95V10A 4.61 −1.56F11A 1.30 −4.87L29A 2.76 −3.41L35A 5.19 −0.98

The urea denaturation curves were analyzed as described.98a Midpoint of the unfolding curve. The error is ±1%.b ΔUrea1/2=Urea1/2 (variant)−Urea1/2 (WT).c The slope of plots of ΔG versus [Urea]. The error is ±10%.d ΔG (H2O)=the intercept of plots of ΔG versus [Urea] at 0 M ureae From ΔUrea1/2×the average m value of WT and the variants (466

in many previous studies.46–48 To investigate thecontribution of the hydrophobic effect to thestability of RNase T1, we prepared two Ile-to-Val,two Val-to-Ala, and one Leu-to-Ala mutants. Twourea denaturation curves were determined for eachof the mutants, and the average results aresummarized in Table 5. Devos et al.49 previouslystudied the V16A and V78A mutants of RNase T1and determined their structures. Our Δ(ΔG) valuesin Table 5 are in excellent agreement with theirvalues of 2.49 kcal/mol for V16A and 4.08 kcal/molfor V78A, which were also obtained from ananalysis of urea denaturation experiments. Thethermodynamics of folding of these hydrophobicmutants of RNase T1 were also studied usingdifferential scanning calorimetry, and the resultsare given in Supplementary Data.

Other proteins

We have also compiled the Δ(ΔG) values for 138hydrophobic mutants in 11 proteins, and these aregiven in Supplementary Data. The compilation

P and hydrophobic variants in 50 mM sodium phosphate

mc

l mol−1 M−1)ΔG (H2O)

d

(kcal mol−1)ΔΔGe

(kcal mol−1)

450 2.7 —470 1.5 −1.4460 2.1 −0.7494 0.6 −2.3450 1.2 −1.6470 2.4 −0.5

.cal mol−1 M−1).

Table 3. Parameters characterizing the thermal unfolding of VHP and hydrophobic variants in 50 mM sodium phosphate(pH 7.0)

VariantΔHm

a

(kcal mol−1)ΔSm

b

(cal mol−1 K−1)Tm

c

(°C)ΔTm

d

(°C)Δ(ΔG)e

(kcal mol−1)

WT 30 88 74.4 — —L2A 23 70 60.4 −14.0 −1.0F7A 18 56 41.7 −32.7 −2.3V10A 30 89 67.0 −7.4 −0.5F11A 22 70 48.8 −25.6 −1.8M13A 15 47 34.7 −39.7 −2.8F18A 20 65 28.3 −46.1 −3.3L21A 22 68 45.3 −29.1 −2.1L29A 24 74 55.3 −19.1 −1.4L35A 29 86 67.2 −7.2 −0.5

The thermal denaturation curves were analyzed as described.98a Enthalpy of unfolding at Tm. The error is ±2 kcal mol

−1.b ΔHm/Tm. The error is ±5 cal mol

−1 K−1.c Midpoint of the unfolding curve. The error is ±0.5 °C.d ΔTm=Tm (variant)−Tm (WT).e Δ(ΔG)=ΔTm×the average ΔSm value of WT and the variants (71 cal mol

−1 K−1).

518 Hydrophobic Interactions to Protein Stability

includes 33 I→V, 39 V→A, 26 I→A, and 40 L→Amutants.

Discussion

Previous studies have shown that the Δ(ΔG)values for hydrophobic mutants are determinedmainly by two factors: first, a constant term thatdepends on the difference in hydrophobicitybetween the WT and mutant side chains, andsecond, a variable term that depends on thedifference in van der Waals interactions of theside chains. It is the second term that leads to the

Table 4. Parameters characterizing the thermal unfoldingof RNase Sa and hydrophobic variants in 30 mM Mops(pH 7.0)

VariantΔHm

a

(kcal mol−1)ΔSm

b

(cal mol−1 K−1)ΔTm

c

(°C)Δ(ΔG)d

(kcal mol−1)

WT 92 286 — —L8A 89 283 −5.0 −1.34L11A 78 252 −11.9 −3.08L19A 73 236 −13.8 −3.65L21A 85 265 −1.6 −0.41I22V 74 234 −5.7 −1.48I58V 89 277 0.1 0.00I70V 87 274 −3.4 − .94I71V 80 249 −0.3 −0.13L91A 91 288 −3.2 −0.84I92V 84 266 −3.9 −1.00

The thermal denaturation curves were analyzed as described.98

The data are the average value of three experiments.a Enthalpy of unfolding at Tm. The error is ±5 kcal mol

−1.b ΔHm/Tm. The error is ±15 cal mol

−1 K−1.c ΔTm=Tm (variant)−Tm (WT). The Tm for WT is 48.4 °C.40d Δ(ΔG)=ΔTm×the averageΔSm value of WT and the variants

(264 cal mol−1 K−1).

range of values (see Table 7) that are observed formutants of the same type in different proteins andat different sites in the same protein. It isdetermined by the distance to neighboring atomsand their polarizability; thus, the local environ-ment is important.19,28,32,33 In mutants of knownstructure, a cavity is generally observed in themutant, and Matthew's group showed that largercavities result in larger Δ(ΔG) values because theloss of van der Waals interactions is greater.22,23 Asimilar observation was made by Bowie's groupfor the membrane protein bacterial rhodopsin.50

These and other studies suggest that the van derWaals interactions of the nonpolar side chains canmake a larger contribution to the Δ(ΔG) valuesthan the differences in hydrophobicity.19,28,32,33

When the Δ(ΔG) values from these two studiesare extrapolated to zero cavity volume, the Δ(ΔG)values for T4 lysozyme are greater than those forrhodopsin.22,23,50 This probably results because therhodopsin side chains are in a more nonpolarenvironment in the unfolded protein, and thus thedifference in hydrophobicity is smaller than thatfor T4 lysozyme. We will later make use of theseideas in interpreting our results.

Hydrophobic interactions contribute more to thestability of a large protein than a small protein

The results for VlsE, VHP, RNase Sa, and RNaseT1 from Tables 1 to 5 were combined andsummarized in Table 6. For VHP, none of the sidechains were 100% buried, but we studied all of thehydrophobic side chains that were over 70% buried.In contrast, over 40% of the aliphatic side chainswere 100% buried in VlsE because of its larger size.The hydrophobicity term discussed above dependson the accessible hydrophobic surface area that is

Table 5. Parameters characterizing the urea unfolding of RNase T1 and hydrophobic variants

VariantUrea1/2

a

(M)ΔUrea1/2

b

(M)mc

(cal mol−1 M−1)ΔG (H2O)

d

(kcal mol−1)ΔΔGe

(kcal mol−1)

pH 7.0, 25 °C, 30 mM Mops bufferWT 5.30 — 1210 6.4 —V16A 3.12 −2.19 1265 3.9 −2.7I61V 3.64 −1.66 1255 4.6 −2.1V78A 2.05 −3.25 1280 2.6 −4.0L86A 1.70 −3.61 1211 2.1 −4.4I90V 4.49 −0.81 1150 5.2 −1.0

pH 5.0, 25 °C, 30 mM formate or acetate bufferWT 6.89 — 1384 9.5 —V78A 3.78 −3.12 1321 5.0 −4.2

The urea denaturation curves were analyzed as previously described.98 The data are the average value of two experiments.a Midpoint of the unfolding curve. The error is ±1%.b ΔUrea1/2=Urea1/2 (variant)−Urea1/2 (WT).c The slope of plots of ΔG versus [Urea]. The error is ±10%.d ΔG (H2O)=the intercept of plots of ΔG versus [Urea] at 0 M urea.e From ΔUrea1/2×the average m value of WT and the variants at pH 7 (1230 cal mol

−1 M−1) and pH 3 (1342 cal mol−1 M−1).

519Hydrophobic Interactions to Protein Stability

removed from contact with water.2,23 Consequently,this term will be reduced for side chains that are notcompletely buried in the folded protein. An approx-imate correction was made by dividing the observedΔ(ΔG) values by the fraction buried for the sidechain, so that the Δ(ΔG) values were compared atthe same accessibility, 100% buried.51,52 The ob-served and corrected Δ(ΔG) values are given inTable 6. This correction will have no effect on thevan der Waals term that also contributes to theΔ(ΔG) values. The same correction was made for allof the mutations analyzed in Table 7 when the WTside chain was less than 100% buried. The averagecorrected Δ(ΔG) values were never over 6% greaterthan the observed values, so the corrections wouldnot change any of the conclusions we reach.One goal of our studies was to see if the burial of

nonpolar groups makes the same contribution to thestability of a small protein, VHP (36 residues), as itdoes to the stability of a large protein, VlsE (341residues). To consider this, we analyzed the resultsof previous studies on the contribution of hydro-phobic interactions to protein stability. For themutants considered, a larger aliphatic side chainwas replaced by a smaller one to minimize stericeffects. The proteins ranged in size from 64 residues(chymotrypsin inhibitor 2) to 168 residues (apo-flavodoxin). (Supplementary Table S3 gives theproteins and mutations used in the analysis, andTables S4 to S7 list the results for individual proteinsfor 33 Ile-to-Val, 39 Val-to-Ala, 26 Ile-to-Ala, and 40Leu-to-Ala variants. Only the side chains that wereat least 70% buried in theWT protein were included.The results for RNases Sa and T1 reported here wereincluded in the analysis.) The analysis included 138mutations that removed a total of 309 –CH2– groupsin 11 proteins. The range and average of the Δ(ΔG)values for each type of mutation are given in Table 7.In addition, the differences in side-chain size are

given along with the Δ(ΔG) values expected whenoctanol or cyclohexane are used to model theinterior of the protein.The range of Δ(ΔG) values is large, and this gives

rise to large standard deviations from the averagevalues. This shows that individual side chains are inquite different environments in folded proteins. Theaverage experimental Δ(ΔG) values are in goodagreement with the Δ(ΔG) values based on cyclo-hexane. We will make use of this observation in thefollowing discussion. For the four types of mutantsin Table 7, the Δ(ΔG) values per –CH2– group aresimilar, suggesting that the average numbers of vander Waals contacts lost are similar even though thepotential differences in volume between the sidechains are large.The results for VHP and VlsE are compared with

those for the other proteins in Table 8. The averageΔ(ΔG) value for VHP is lower than that for anyother protein. This was expected. None of thehydrophobic side chains in VHP are completelyburied, so they will have less favorable van derWaals interactions than a completely buried sidechain. Since VHP is small, it should be possible fortheoreticians to use these results to achieve a betterunderstanding of the contribution of the van derWaals interactions of the hydrophobic side chains tothe stability of globular proteins.The average Δ(ΔG) values for RNase T1 and VlsE

are considerably higher than those for any of theother proteins. The results for VlsE are interestingand allow us to draw some conclusions. In 2000,Fleming and Richards examined the packing densityin a large sample of proteins and concluded,53 “Onewould expect increased overall packing in largeproteins, proteins with a high percentage of helixand proteins with a high percentage of aromatic andsmall residues and a low percentage of aliphaticresidues.” They also found that some proteins had

Table 6. Observed and accessibility corrected Δ(ΔG)values for the hydrophobic mutants of VlsE, VHP,RNase Sa, and RNase T1

Protein VariantSide-chain% burieda

Δ(ΔG)obsb

Δ(ΔG) acc.correctedc

Per–CH2–

d

VlsE I128V 100 1.1 1.1 1.1I154V 100 1.8 1.8 1.8I210V 100 1.9 1.9 1.9I214V 100 1.9 1.9 1.9I306V 100 1.5 1.5 1.5

VHP V10A 70 0.6 0.9 0.4L2A 70 1.2 1.7 0.6L21A 74 2.1 2.8 0.9L29A 73 1.5 2.1 0.7L35A 79 0.5 0.6 0.2F7A 98 2.3 2.3 —F11A 98 2.1 2.1 —F18A 98 3.3 3.4 —M13A 72 2.8 3.9 —

RNase Sa I22V 81 1.5 1.8 1.9I58V 50 0 0 0I70V 100 0.9 0.9 0.9I71V 100 0.1 0.1 0.1I92V 100 1.0 1.0 1.0L8A 84 1.3 1.6 0.5L11A 97 3.1 3.2 1.1L19A 95 3.7 3.8 1.3L21A 67 0.4 0.6 0.2L91A 64 0.8 1.3 0.4

RNase T1 I61V 97 2.1 2.2 2.2I90V 100 1.0 1.0 1.0V16A 98 2.7 2.8 1.4V78A 100 4.1 4.1 2.1L86A 74 4.4 5.9 2.0

a The percent buried for the side chain in the WT protein wascalculated using PFIS.99

b The VlsE results are from Table 1. The VHP results are basedon the data in Tables 2 and 3, and the average was used when twoΔ(ΔG) values were available. The RNase Sa results are theaverage of three values for each mutant and are from Table 4. TheRNase T1 results are the average of two values for each mutantand are from Table 5.

c Δ(ΔG) accessibility corrected=Δ(ΔG) observed/(side-chainfraction buried).

d The per –CH2– values=Δ(ΔG) accessibility corrected valuesfor the I to V mutants, divided by 2 for the V-to-A mutants anddivided by 3 for the I-to-A and L-to-A mutants.

Table 8. Number of residues and average Δ(ΔG) valuesper –CH2– group for 148 mutations in 13 proteins

Protein (# mutations) # ResiduesaΔ(ΔG)

(kcal/mol)b

VlsE (5) 341 1.6±0.3ApoFlavo (14) 168 0.9±0.4T4 Lyso (25) 164 1.0±0.4SN (28) 141 1.2±0.5hLyso (11) 130 0.7±0.3Ba (15) 110 1.2±0.4FKBP (15) 107 1.2±0.3RNase T1 (5) 104 1.7±0.5RNase Sa (7) 96 1.0±0.5AcCoABP (5) 86 0.9±0.5Ubiquitin (4) 76 1.3±0.4CI2 (9) 64 1.1±0.7VHP (5) 36 0.6±0.3Average 125 1.1±0.5

Ba, barnase; SN, staph nuclease; T4 Lyso, T4 lysozyme; CI2,chymotrypsin inhibitor 2; FKBP, FK506-binding protein; hLyso,human lysozyme; ApoFlavo, apoflavodoxin; AcCoABP, acyl-CoA-binding protein.

a Number of residues in the protein.b AverageΔ(ΔG) values for all of the aliphatic mutants studied

for these proteins. The individual Δ(ΔG) values for all of theproteins are given in Supplementary Data.

520 Hydrophobic Interactions to Protein Stability

packing densities that differed significantly from theaverage. In contrast, Liang and Dill examined thepacking density in the interiors of globular proteins,and concluded, in part,54 “We find that larger

Table 7. Average Δ(ΔG) values for 138 hydrophobic mutants

SubstitutionVolumea

(Å3)ΔGtr tooctanola

ΔGcylohe

Ile→Val (33) 25.8 −0.80 −0Val→Ala (39) 49.0 −1.24 −2Ile→Ala (26) 74.8 −2.04 −3Leu→Ala (40) 74.5 −1.90 −3

a From Table 1 of Ref. 52.b From Supplementary Data.

proteins are packed more loosely than smallerproteins. And we find that the enthalpies of folding(per amino acid) are independent of the packingdensity of a protein, indicating that van der Waalsinteractions are not a dominant component of thefolding forces.” Both of these studies and othershave pointed out that the packing of proteininteriors is heterogeneous, and this is one of thereasons for the differences between proteins seen inTable 8.55–57 VlsE was not included in theseanalyses, but our results suggest that it may bemore tightly packed than the other proteins thathave been studied, all of which are considerablysmaller than VlsE.It seemed possible that the DSEs of larger proteins

might have more structure under physiologicalconditions than those of smaller proteins becausethe percentage of nonpolar residues increases withsize.58 If so, the hydrophobic effect might make asmaller contribution to the stability of proteins asthey become larger. The results with VlsE suggestthat this is not the case.

from 11 proteins

tr toxanea Range

Δ(ΔG)b

averagePer

–CH2–

.88 −0.1 to 2.3 1.1±0.6 1.1±0.6

.23 0 to 4.9 2.4±1.1 1.2±0.5

.11 0.7 to 5.4 3.1±1.2 1.0±0.4

.11 0.2 to 6.2 3.2±1.2 1.1±0.4

521Hydrophobic Interactions to Protein Stability

The results for RNase T1 are interesting. Theaverage Δ(ΔG) is larger than for any other protein.This suggests that the cavities left in the RNase T1mutants might be larger than in the other proteins.This is supported by the results of De Vos et al.49

who determined the structures of the V16A andV78A mutants and concluded, “Little or no adapta-tion of the immediately surrounding residues isobserved that reduces the volume of the inducedcavity….” The atomic temperature factors of the WTprotein also suggest that the mutations are in a rigidpart of the molecule. This might result because thisprotein has two disulfide bonds so that it is moredifficult for the structure to relax and fill the cavity.However, this is not consistent with the results forhuman lysozyme. After VHP, human lysozyme hasthe lowest average Δ(ΔG) per –CH2– group. Thissuggests that this protein does not have largecavities in the mutants, despite the fact that thisprotein has four disulfide bonds, more than any ofthe other proteins in Table 8.

Forces contributing to protein stability

The average observed Δ(ΔG) values in Table 7 arein good agreement with the values predicted frommodel compound studies using cyclohexane (CH) asa model for the protein interior. This agreementprobably results because CH, on average, is morepolarizable than the interior of the protein but is notas tightly packed.9,52 This suggests that the ΔGtrvalues from water to CH can be used to estimate theaverage contribution of the hydrophobic effect toprotein stability. The aliphatic residues and the –CH2– groups from the polar and charged side chainsmake the largest contribution to the hydrophobiceffect, ranging from 78% for VHP to 96% for VlsEwith an average of 86±5%. For the nonpolar sidechains that can also form hydrogen bonds, Trp, Tyr,and Cys, it is preferable to use the ΔGtr values foroctanol so that hydrogen bonding does not contrib-ute to the ΔGtr value.

6,52 In Table 9, we estimate thecontribution of the hydrophobic effect to thestability of 22 proteins with the number of residuesranging from 36 to 534. These are all single-domainproteins, and one (bacterial rhodopsin) is a mem-brane protein.50,59,60 The details of how thesecalculations are done and the ΔGtr values used aregiven in the footnotes to Table 9.The hydrophobic interaction contribution depends

on how many nonpolar groups are buried onfolding, and this depends on the model used for theDSE. Table 9 gives the results for two different DSEmodels. For the least accessible DSE, the lowerboundary model of Creamer et al.61 is used, whichmodels the DSE using 17-residue fragments of theproteins with the same conformation they have inthe folded proteins. This gives hydrophobic inter-action contributions of 34 kcal/mol for VHP and

569 kcal/mol for Candida rugosa (CR) lipase(column 7). These estimates are probably too low.For our analysis, we will use the upper boundmodel of Creamer et al.,61 which models theproteins in an extended β-sheet conformation.Now the contribution of hydrophobic interactionsvaries from 45 for VHP to 741 for CR lipase(column 8). For the individual proteins, thecontribution of hydrophobic interactions per resi-due is lowest for RNase T1 at 1.0 kcal/mol andhighest for barstar at 1.7 kcal/mol with an averageof 1.4±0.2 kcal/mol per residue (column 9). Incomparison, the average is 1.1±0.1 for the lowerbound model. The upper bound model gives theestimate expected if the unfolded protein iscompletely accessible to water. More likely, thecorrect extent of exposure is intermediate betweenthe upper and lower bound models.62,63

Until 1990, the prevailing idea was that intramo-lecular hydrogen bonds were necessary for main-taining the structure of a protein but made only asmall contribution to protein stability.64,65 Morerecent evidence shows that hydrogen bonds make alarge contribution to protein stability.41,46,60,66–71 For35 Tyr-to-Phe mutations with the Tyr –OH grouphydrogen bonded, Δ(ΔG)=−1.4±0.9 kcal/mol, andfor 17 mutations with the –OH group not hydrogenbonded, Δ(ΔG)=−0.2±0.4 kcal/mol.41,68 Bowie hasshown that the difference between these two Δ(ΔG)values, −1.2 kcal/mol, is similar to the contributionof a hydrogen bond to protein stability determinedby double mutant cycle analyses.72 A similaranalysis of Thr-to-Val mutations leads to an estimateof 0.9 kcal/mol per hydrogen bond.68 A study ofAsn-to-Ala mutants in which the Asn amide groupis hydrogen bonded to a peptide group gives anestimate of 1.1 ± 0.6 kcal/mol per hydrogenbond.69,70,73,74 Finally, studies of amide to estermutations of backbone peptide groups suggest thatthese hydrogen bonds contribute 1.1±1.6 kcal/molper hydrogen bond to protein stability. These resultscover hydrogen bonds between side chains, be-tween backbone groups, and between side chainsand backbone groups. Consequently, we will makethe conservative assumption that each intramolecu-lar hydrogen bond in a folded protein contributes1 kcal/mol to the stability.In Table 9, the number of hydrogen bonds formed

by each protein is shown, and this provides anestimate of the amount they contribute to thestability. The contribution from hydrogen bondsranges from 28 kcal/mol for VHP to 548 kcal/molfor CR lipase. When the hydrogen bond contribu-tion per residue is considered, they range from0.8 kcal/mol for barstar to 1.2 kcal/mol for T4lysozyme. The average number of hydrogen bondsformed in these proteins is 0.95±0.10 per residue,with 65% between peptide groups, 23% betweenpeptide groups and side chains, and 12% between

Table 9. Forces contributing to the conformational stability of 22 proteins of increasing size

Proteina PDBb #Resc S-Sd HBeHB/Resf HΦLB

g HΦUBg

HΦ/resh

Totalstabilityi

CE(calculated)j

Pred stability(CE=2.41)k

% Diff(total vs. pred)l

VHP 1yrf 35 0 28 0.8 34.45 44.86 1.28 72.86 2.08 84.35 13.62Protein G 1pgb 56 0 54 0.96 51.24 66.72 1.19 120.72 2.16 134.96 10.55Ubiquitin 1ubq 76 0 69 0.91 92.63 120.63 1.59 189.63 2.50 183.16 −3.53Barstar 1bta 89 0 69 0.78 115.35 150.21 1.69 219.21 2.46 214.49 −2.20RNase Sa 1rgg 96 5 (1) 90 0.94 83.92 109.29 1.14 204.29 2.13 231.36 11.70RNase T1 9rnt 104 7 (2) 99 0.95 77.05 100.33 0.96 206.33 1.98 250.64 17.68Barnase 1bni 108 0 108 1.00 105.31 137.14 1.27 245.14 2.27 260.28 5.82RNase A 9rsa 124 15 (4) 110 0.89 109.42 142.48 1.15 267.48 2.16 298.84 10.49Che Y 3chy 128 0 122 0.95 154.72 201.48 1.57 323.48 2.53 308.48 −4.86hu Lyso 1lz1 130 18.2 (4) 137 1.05 139.89 182.17 1.40 337.37 2.60 313.30 −7.68Staph Nuc 1ey0 136 0 120 0.88 154.66 201.40 1.48 321.40 2.36 327.76 1.94T4 Lyso 1l63 162 0 188 1.16 186.02 242.24 1.50 430.24 2.66 390.42 −10.20Adn Kinase 3adk 195 0 160 0.82 219.71 286.12 1.47 446.12 2.29 469.95 5.07b Rhod 1c3w 222 0 243 1.09 269.12 350.46 1.58 593.46 2.67 535.02 −10.92Carb Anhy 1cah 258 0 238 0.92 288.43 375.60 1.46 613.60 2.38 621.78 1.32Trp Syn α 1wq5 258 0 225 0.87 290.95 378.89 1.47 603.89 2.34 621.78 2.88VlsE 1l8w 271 0 251 0.93 290.07 377.74 1.39 628.74 2.32 653.11 3.73b Lipase 1cvl 316 4.1 (1) 330 1.04 333.61 434.43 1.37 768.53 2.43 761.56 −0.92Pepsinogen 3psg 365 7.3 (3) 319 0.87 388.89 506.41 1.39 832.71 2.28 879.65 5.34MBP 1omp 370 0 337 0.91 423.53 551.53 1.49 888.53 2.40 891.70 0.36Glu Amyl 1glm 470 10 (3) 503 1.07 505.75 658.59 1.40 1171.59 2.49 1132.70 −3.43CR lipase 1crl 534 5.9 (2) 548 1.03 568.97 740.93 1.39 1294.83 2.42 1286.94 −0.61

a Adn Kinase, adenylate kinase; b Rhod, bacterial rhodopsin; Carb Anhy, carbonic anhydrase; Trp Syn α, tryptophan synthase αsubunit; MBP, maltose binding protein; Glu Amyl, glucoamylase; hu Lyso, human lysozyme; Staph Nuc, staphylococcal nuclease; bLipase, bacterial lipase.

b http://www.pdb.org/ and Ref. 100.c Number of residues observed in the crystal structure. For some of the proteins, this is less than the number of residues in the protein:

VHP, 36; barnase, 110; Staph Nuc, 149; T4 Lyso, 164; Carb Anhy, 259; Trp Syn α, 268; VlsE, 341; lipase, 319; pepsinogen, 370.d Eight of the proteins contain disulfide bonds and the number is shown in parentheses. The contribution of the disulfide bonds to the

stability is given in kilocalories per mole by the number in the column. The values for RNase Sa40 and RNase T1101 are measured valuesand the values for the other proteins were calculated as described.101

e This column gives the number of hydrogen bonds in the protein and is also an estimate of the contribution of hydrogen bonds to thestability in kilocalories per mole, since we assume that each intramolecular hydrogen bond contributes 1 kcal/mol to the stability.41,68 Thenumber of hydrogen bonds was calculated with PFIS. The criteria used to establish a hydrogen bond are the same as those used inprevious studies,75,102 except we count hydrogen bonds when the distance between the two electronegative atoms is 3.6 Å or less.

f This gives the hydrogen-bonding contribution per residue in kilocalories per mole.g This gives the hydrophobic interaction contribution in kilocalories per mole when the denatured state is modeled using the lower

boundary model (column 7) or the upper boundary model (column 8) of Creamer et al.61 For five proteins ranging from 110 to 275residues, they found that the denatured-state ASA of the upper boundary model was 17.8±0.6% less and that of the lower boundarymodel was 36.9±0.5% less than the ASA based on the Gly-X-Gly tripeptide model.

h This gives the hydrophobic interaction contribution per residue in kilocalories per mole when the denatured state is modeled usingthe upper boundary model of Creamer et al. 61 (column 8).

i Total stability=S-S+HB+HΦUB.j CE (calculated)=(total stability)/(# residues). CE is the conformational entropy contribution per residue, in kilocalories per mole,

needed to give an instability equal to the total stability.k Predicted stability=2.41×(# residues). When total stability is plotted versus the number of residues, the least-squares slope is

2.41 kcal/mol per residue. This is the average conformational entropy contribution per residue that gives the best agreement with totalstability.

l This is the percent difference between total stability (column 10) and predicted stability (column 12).

522 Hydrophobic Interactions to Protein Stability

side chains. In a larger sample of proteins, the Rosegroup found an average of 1.1 hydrogen bonds perresidue.75

The total stability is the sum of the contribution ofdisulfide bonds, hydrogen bonds, and hydrophobicinteractions (column 10). It ranges from 73 kcal/molfor VHP to 1295 kcal/mol for CR lipase. There is agood correlation between the stability estimates andthe number of residues: HB=0.98±0.02 (# residues)with a correlation coefficient of 0.989; HΦ=1.72±0.02 (# residues) with a correlation coefficient of0.994; and total stability=2.41±0.02 (# residues)with a correlation coefficient of 0.997. (These plots

are shown in Supplementary Fig. S2.) This analysissuggests that both hydrogen bonds and hydropho-bic interactions make large contributions to theconformational stability of globular proteins.In a previous study, a different approach sug-

gested that polar group burial contributes more toprotein stability than nonpolar group burial.74 Thiswas based, in part, on a paper by Harpaz et al.,76

who analyzed the completely buried surface in asample of 108 proteins. They showed that the totalvolume of the buried residues was 298,100 Å3. Ofthis, 123,653 Å3 came from nine nonpolar sidechains (Table 9) and 14,242 Å3 from the –CH2–

523Hydrophobic Interactions to Protein Stability

groups of the polar and charged side chains, for atotal of 137,895 Å3 (peptide groups contributed92,000 Å3). Based on the data available at that time,the contribution of hydrophobic group burial wasestimated using 49 cal/mol/Å3 of nonpolar volumeburied. Using the more recent data presented here,the contribution is 44±3 cal/mol/Å3, and this leadsto (137,895 Å3×044 kcal/mol/Å3)=6067 kcal/molas the contribution of the hydrophobic effect to thestability of these proteins. If the approach used inTable 9 is applied to the same data, it leads to a5954 kcal/mol estimate of the contribution of thehydrophobic effect. It is reassuring that these twoestimates differ by less than 2%. In the same paper,we estimated that burying the 92,000 Å3 of peptidegroups contributed 7200 kcal/mol to the stability.The major destabilizing force in protein folding is

conformational entropy. Studies by the Freire77 andRecord78 groups suggested that 1.7 kcal/mol perresidue can be used to estimate the contribution ofconformational entropy to protein stability. How-ever, this estimate is too small to compensate for thetotal stabilities shown in Table 9. This suggests thatthese conformational entropy estimates are too low,or that our estimates of the stabilizing forces are toolarge, or that there are other forces that make amajorcontribution to protein instability. We consider thesetopics next.

Forces contributing to protein instability

When the prevailing thought was that hydrogenbonds make only a small contribution to stability, itwas assumed that polar groups that were nothydrogen bonded would make a large unfavorablecontribution to the stability because they wouldform hydrogen bonds to water when the proteinunfolds. This was first analyzed in a paper fromFinney's group.79 They showed for a large sample ofproteins that there is an excellent correlation (0.99)between “lost hydrogen bonds” and three stabiliz-ing forces: hydrophobic effect, disulfide bonds, andion pairs. (The hydrophobic effect contributed over95% to their stabilizing forces.) They found that over90% of the polar atoms form intramolecular hydro-gen bonds or hydrogen bonds to water molecules.However, about 24% of the carbonyl oxygen atomsand 6% of the amide hydrogens of the peptide bondsin proteins do not form hydrogen bonds, and this isthe major source of the lost hydrogen bonds.More recently, analyses by Fleming and Rose80 led

them to conclude, “Unsatisfiedbackbonepolargroupsare energetically expensive, to the degree that theyalmost never occur.” For the Tyr-to-Phe mutationsdiscussed above, Δ(ΔG)=−0.2±0.4 kcal/mol for theTyr residues that were not hydrogen bonded.68 Thissuggests that side chains that are not hydrogenbonded do not make a large unfavorable contribu-tion to the stability. Even for 40Val-to-Thrmutations,

Δ(ΔG)=1.8±1.1 kcal/mol,41,68 showing that whena polar –OH group is placed at a site designed for a–CH3 group, the penalty is far less than the cost of 5to 6 kcal/mol proposed by Fleming and Rose.80

Consequently, we doubt that non-hydrogen-bondedpolar groups make a large contribution to proteininstability.Another interesting suggestion was made by

Kajander et al.58 They analyzed the structures of alarge sample of proteins of different molecularweights and showed that 65% more charged groupsare buried in proteins containing 700 amino acidsthan in proteins containing 100 amino acids. Fromthis they concluded, “Nature may use charge burialto reduce protein stability; not all buried charges arefully stabilized by a prearranged protein environ-ment.” If so, removing them may be a way ofincreasing the stability of proteins. The Garcia-Moreno group has shown that burying a charge ina nonpolar environment in staph nuclease candecrease the stability by over 5 kcal/mol,81and wehave observed that removing a naturally occurringburied charge in RNase Sa can increase the stabilityby over 3 kcal/mol.42 However, if buried chargedgroups are hydrogen bonded, they can make a largefavorable contribution to protein stability.82

Kajander et al.58 included the backbone atoms intheir analysis so that the peptide C=O and NHgroups were classed as polar uncharged and the α-carbon was classed as aliphatic. For a differentperspective, we have analyzed the proteins in Table9 and considered just the side chains. The fractionburial of nonpolar side chains (Ala, Val, Ile, Leu,Met, Phe, Tyr, Trp, Cys), polar uncharged sidechains (Asn, Gln, Ser, Thr, His), and polar chargedside chains (Asp, Glu, Lys, Arg) was determined as afunction of the number of residues. (The plots areshown in Supplementary Fig. S3.) For each group,the fraction of buried side chains increases linearlywith the number of residues: for nonpolar, fractionburied=0.363±0.007 (# residues) with a correlationcoefficient of 0.99; for polar uncharged, fractionburied=0.155±0.009 (# residues) with a correlationcoefficient of 0.93; and for polar charged, fractionburied=0.117±.005 (# residues) with a correlationcoefficient of 0.95. For the polar charged side chains,if we consider only the O atoms for Asp and Glu, theN for Lys, and the two outermost N for Arg, thefraction buried is similar: fraction buried=0.112±0.004 (# residues) with a correlation coefficient of0.95. The percentage of the charged groups that isburied is lowest for barstar (22%) and greatest forglucose amylase (70%). For the small and largeproteins studied here, 28% of the charged groupsare buried in VHP and 50% are buried in VlsE.Thus, we see a trend similar to that observed byKajander et al.58 This subject deserves further study.The destabilizing force of most interest is confor-

mational entropy. As noted above, two different

524 Hydrophobic Interactions to Protein Stability

approaches suggest that the entropic cost of foldinga protein is about 1.7 kcal/mol per residue.77,78 Wehave used this estimate in previous papers.40,83 Wenow think that this estimate is too low. Theestimates of the conformational entropy needed tobalance the total stability for each protein are givenin Table 9. The smallest estimate is 2.0 kcal/mol perresidue for RNase T1, and the largest estimate is 2.7for T4 lysozyme and rhodopsin. A value of 2.41±0.02 kcal/mol per residue gives the best fit for all 22proteins. Using this value gives the destabilizingestimates (column 12) shown in Table 9. In only twocases do the stabilizing and destabilizing estimatesdiffer by over 12%. The predicted stabilities are 17%too low for RNase T1 and 14% too low for VHP. Thisreflects the fact that the contribution of hydrophobicinteractions is exceptionally low for RNase T1, andthe contribution of hydrogen bonds is exceptionallylow for VHP. For rhodopsin, the predicted stabilityis 11% too high. This is expected since rhodopsin is amembrane protein. The measured values for thecontribution of hydrogen bonds and hydrophobicinteractions to rhodopsin are both lower than thevalues found with soluble proteins.50,60 This resultsbecause the unfolded protein will be in a morenonpolar environment and this will tend to lowerthe contribution that both hydrogen bonds andhydrophobic interactions make to the stability. Thestability estimates used in Table 9 are based onnonmembrane proteins, and they are expected tolead to an overestimate of the stability of amembrane protein such as rhodopsin.On the basis of experimental data, Makhatadze

and Privalov have pointed out why previousconformational entropy estimates are too low andsuggest an average value of 3.6 kcal/mol perresidue.84 Using an extended all-atom simulationof the folding of a three-helix bundle protein,Boczko and Brooks85 estimated a conformationalentropy cost of 2.6 kcal/mol per residue. Thesearticles and others were reviewed by Brady andSharp,86 who suggested that the conformationalentropy cost would be between 2.4 and 3.7 kcal/mol per residue. Thus, our estimate of 2.41 kcal/mol per residue is in line with other experimentaland theoretical estimates.Our estimates of the total stability of the proteins

given in Table 9 should be considered a lower limitfor the following reasons. First, the ΔGtr values forCH used to calculate the contribution of the burial ofthe nine nonpolar side chains and the –CH2– groupsto protein stability are more likely too small than toolarge. Second, the contribution of individual hydro-gen bonds to stability is more likely to be greaterthan 1 kcal/mol than less.67 Third, other forces thatcontribute to protein stability such as charge–chargeinteractions, ion pairs, and π–cation interactions willgenerally make a small favorable contribution.87,88 Ifthe estimates of the total stability were higher, the

estimates for the contribution of conformationalentropy would also be higher, but they would stillbe in the range suggested by Brady and Sharp.86

Stability of VHP

Because of its small size, the thermodynamics andkinetics of folding of VHP are currently beingstudied by a variety of experimental35,39,89 andtheoretical90–92 approaches. Consequently, there isconsiderable interest in determining which residuesmake important contributions to the stability andthe mechanism of folding of VHP. In 2002, Frank etal.93 replaced Met13 and the three Phe residues withLeu, andwere able to conclude that Phe18 makes thelargest contribution to the stability. Our resultssupport this conclusion. Raleigh's group has shownthat the double mutant with Ala at position 28 andMet at residue 30 raises the Tm by over 20 °C.

38 Theyhave also shown that a π–cation interaction betweenPhe7 and Arg15 contributes to the stability.38 TheKelly laboratory has studied five backbone and twoside-chain hydrogen bonds in VHP.39 They estimatethat the backbone hydrogen bonds contribute1.5 kcal/mol per hydrogen bond and the two side-chain hydrogen bonds contribute about 1 kcal/molper hydrogen bond. The nine VHP mutants studiedhere (Table 6) have not been studied previously. Wenext consider what these mutants reveal about thestability of VHP.The results in Table 6 show that the aromatic rings

of Phe7, Phe11, and Phe18 contribute 2.3, 2.1, and3.3 kcal/mol for a total contribution of 7.7 kcal/molto the stability of VHP. In addition, the 14 –CH2–group equivalents removed by the five aliphaticmutations contribute 5.9 kcal/mol, and the –CH2–S–CH3 group of Met13 contributes 2.8 kcal/mol. Thisamounts to 16.4 kcal/mol of stability based on ourmeasured Δ(ΔG) values. As we did in Table 9, wecan estimate the contribution of the other sixnonpolar side chains (Ala9, Ala17, Ala19, Leu23,Trp24, and Phe36) that are only partially buried as5 kcal/mol. In addition, another 34 –CH2– groupsare buried when the protein folds, and using0.6 kcal/mol per –CH2– group, these wouldcontribute another 20 kcal/mol to the stability.These are from the polar and charged side chainsand from the groups not removed in the mutants inTable 8. Based on this, we estimate that the burial ofnonpolar groups contributes about 41 kcal/mol tothe stability of VHP. This compares with an estimateof 45 kcal/mol in Table 9, which is higher because1 kcal/mol was used for the contribution of the –CH2– groups rather than the observed value forVHP of 0.6 kcal/mol. In comparison, VHP forms 28hydrogen bonds on folding: 23 between peptidegroups, 4 between peptide groups and side chains,and only 1 between two side chains. If we assign1 kcal/mol per hydrogen bond, then hydrogen

525Hydrophobic Interactions to Protein Stability

bonds contribute considerably less to the stabilitythan hydrophobic interactions. However, if we use1.5 kcal/mol per peptide hydrogen bond, as foundfor 5 of the 23 backbone hydrogen bonds,39 thenhydrogen bonds (40 kcal/mol) and hydrophobicinteractions (41 kcal/mol) make about the samecontribution to the stability. Thus, even for a smallprotein where the nonpolar side chains cannot becompletely buried, hydrophobic interactions con-tribute more to the stability than hydrogen bonds.

Concluding remarks

For the 22 proteins in Table 9, the averagecontribution of hydrophobic interactions to thestability is 60±4%. The contribution is smallest forRNase T1 (54%) and greatest for barstar (73%). Theaverage contribution of hydrogen bonds to thestability is 40±4%. The contribution is smallest forbarstar (27%) and largest for RNase T1 (43%). Thelargest contribution of disulfide bonds to thestability is for RNase A (5%). This analysis suggeststhat hydrophobic interactions make the dominantcontribution to protein stability always greater than50%. Hydrogen bonds also make a large contribu-tion to stability, but always less than that byhydrophobic interactions even for the smallestglobular proteins. The loss in conformational entro-py when a protein folds is the major destabilizingforce, but the burial of non-hydrogen-bonded polaror charged groups may also contribute instability infine-tuning protein stability.

Materials and Methods

All buffers and chemicals were of reagent grade. Ureawas from Amresco or Nacalai Tesque (Kyoto, Japan) andwas used without further purification. The plasmids forVHP and VlsE and their variants were derived from pETvectors (Novagen) and have been described previously.94

The plasmids for RNase Sa and RNase T1 and theirvariants were derived from the pEH100 plasmid asdescribed previously.40,95 The expression hosts for allproteins and variants were Escherichia coli strain RY1988(MQ), DS2000, or C41(DE3).94,96 Oligonucleotide primersfor mutagenesis were from Integrated DNA Technologies(Coralville, IA). Site-directed mutagenesis was performedusing a QuikChange™ Site-Directed Mutagenesis Kitfrom Stratagene (La Jolla, CA). Mutant plasmids weresequenced by the Gene Technologies Laboratory, TexasA&M University.VHP and VlsE and their variants were expressed and

purified as described previously.63,94 RNase Sa and RNaseT1 and their variants were expressed and purified asdescribed previously.82,95,96 The purity of all proteins wasconfirmed by SDS-PAGE and matrix-assisted laser de-sorption/ionization time-of-flight mass spectrometry.Urea and thermal denaturation curves were determined

using an AVIV 62DS or 202SF spectropolarimeter (AvivInstruments, Lakewood, NJ) to follow unfolding. The

methods for VHP and VlsE have been describedpreviously.63 The methods for RNase Sa and RNase T1and their variants have been previously described.97,98

The analysis of urea and thermal denaturation curvesassumed a two-state unfolding model and was done asdescribed.97,98

Acknowledgements

This work was supported by NIH grants GM37039 and GM 52483, Robert A. Welch Foundationgrants BE-1060 and BE-1281, and the Tom andJean McMullin Professorship. We thank GeorgeMakhatadze, Evan Powers, Jeff Kelly, and JamesBowie for providing important information, and JeffMyers and Doug Laurents for providing sugges-tions for the manuscript. We also thank MichaelPerham and Pernilla Wittung-Stafshede for helpwith the expression of VlsE.

Supplementary Data

Supplementary materials related to this article canbe found online at doi:10.1016/j.jmb.2011.02.053

References

1. Bernal, J. D. (1939). Structure of proteins. Nature, 143,663–667.

2. Kauzmann, W. (1959). Some factors in the interpre-tation of protein denaturation. Adv. Protein Chem. 14,1–63.

3. Tanford, C. (1962). Contribution of hydrophobicinteractions to the stability of the globular conforma-tion of proteins. J. Am. Chem. Soc. 84, 4240–4247.

4. Tanford, C. (1997). How protein chemists learnedabout the hydrophobic factor.Protein Sci. 6, 1358–1366.

5. Nozaki, Y. & Tanford, C. (1971). The solubility ofamino acids and two glycine peptides in aqueousethanol and dioxane solutions. establishment of ahydrophobicity scale. J. Biol. Chem. 246, 2211–2217.

6. Fauchere, J. L. & Pliska, V. E. (1983). Hydrophobicparameters of amino-acid side-chains from thepartitioning of N-acetyl-amino-acid amides. Eur. J.Med. Chem. 18, 369–375.

7. Wimley, W. C., Creamer, T. P. & White, S. H. (1996).Solvation energies of amino acid side chains andbackbone in a family of host-guest pentapeptides.Biochemistry, 35, 5109–5124.

8. Damodaran, S. & Song, K. B. (1986). The role ofsolvent polarity in the free energy of transfer ofamino acid side chains from water to organicsolvents. J. Biol. Chem. 261, 7220–7222.

9. Radzicka, A. & Wolfenden, R. (1988). Comparing thepolarites of the amino acids; side-chain distributioncoefficients between the vapor phase, cyclohexane,1-octanol, and neutral squeous solution. Biochemistry,27, 1644–1670.

http://dx.doi.org/10.1016/j.jmb.2011.02.053

526 Hydrophobic Interactions to Protein Stability

10. Wesson, L. & Eisenberg, D. (1992). Atomic solvationparameters applied to molecular dynamics of pro-teins in solution. Protein Sci. 1, 227–235.

11. Rose, G. D. & Wolfenden, R. (1993). Hydrogenbonding, hydrophobicity, packing, and protein fold-ing. Annu. Rev. Biophys. Biomol. Struct. 22, 381–415.

12. Karplus, P. A. (1997). Hydrophobicity regained.Protein. Sci. 6, 1302–1307.

13. Chan, H. S. & Dill, K. A. (1997). Solvation: how toobtain microscopic energies from partitioning andsolvation experiments. Annu. Rev. Biophys. Biomol.Struct. 26, 425–459.

14. Yutani, K., Ogasahara, K., Tsujita, T. & Sugino, Y.(1987). Dependence of conformational stability onhydrophobicity of the amino acid residue in a seriesof variant proteins aubstituted at a unique position oftryptophan synthase α subunit. Proc. Natl Acad. Sci.USA, 84, 4441–4444.

15. Kellis, J. T., Nyberg, K., Sail, D. & Fersht, A. R. (1988).Contribution of hydrophobic interactions to proteinstability. Nature, 333, 784–786.

16. Kellis, J. T., Nyberg, K. & Fersht, A. R. (1989).Energetics of complementary side chain packingin a protein hydrophobic core. Biochemistry, 28,4914–4922.

17. Serrano, L., Kellis, J. T., Cann, P., Matouschek, A. &Fersht, A. R. (1992). The folding of an enzyme: II.Substructure of barnase and the contribution ofdifferent interactions to protein stability. J. Mol.Biol. 224, 783–804.

18. Shortle, D., Stites, W. E. & Meeker, A. K. (1990).Contributions of the large hydrophobic amino acidsto the stability of staphylococcal nuclease. Biochem-istry, 29, 8033–8041.

19. Chen, J. & Stites, W. E. (2001). Packing is a keyselection factor in the evolution of protein hydro-phobic cores. Biochemistry, 40, 15280–15289.

20. Sandberg, W. S. & Terwilliger, T. C. (1991). Energet-ics of repacking a protein interior. Proc. Natl Acad. Sci.USA, 88, 1706–1710.

21. Eriksson,A.E., Baase,W.A., Zhang, X. J., Heinz,D.W.,Blaber, M., Baldwin, E. P. & Matthews, B. W. (1992).Response of a protein structure to cavity-creatingmutations and its relation to the hydrophobic effect.Science, 255, 178–183.

22. Xu, J., Baase, W. A., Baldwin, E. & Matthews, B. W.(1998). The response of T4 lysozyme to large-to-smallsubstitutions within the core and its relation to thehydrophobic effect. Protein Sci. 7, 158–177.

23. Baase, W. A., Liu, A. D., Tronrud, D. E. & Matthews,B. W. (2010). Lessons from the lysozyme of phage T4.Protein Sci. 19, 631–641.

24. Jackson, S. E.,Moracci,M., elMasry,N., Johnson,C.M.& Fersht, A. R. (1993). Effect of cavity-creatingmutations in the hydrophobic core of chymotrypsininhibitor 2. Biochemistry, 32, 11259–11269.

25. Main, E. R. G., Fulton, K. F. & Jackson, S. E. (1998).Context-dependent nature of destabilizing muta-tions on the stability of FKBP12. Biochemistry, 37,6145–6153.

26. Takano, K., Yamagata, Y. & Yutani, K. (1998). Ageneral rule for the relationship between hydropho-bic effect and conformational stability of a protein:stability and structure of a series of hydrophobic

mutants of human lysozyme. J. Mol. Biol. 280,749–761.

27. Cota, E., Hamill, S. J., Fowler, S. B. & Clarke, J. (2000).Two proteins with the same structure respond verydifferently to mutation: the role of plasticity inprotein stability. J. Mol. Biol. 302, 713–725.

28. Loladze, V. V., Ermolenko, D. N. &Makhatadze, G. I.(2002). Thermodynamic consequences of burial ofpolar and non-polar amino acid residues in theprotein interior. J. Mol. Biol. 320, 343–357.

29. Clark, A. T., McCrary, B. S., Edmondson, S. P. &Shriver, J. W. (2004). Thermodynamics of corehydrophobicity and packing in the hyperthermo-phile proteins Sac7d and Sso7d. Biochemistry, 43,2840–2853.

30. Bueno, M., Campos, L. A., Estrada, J. & Sancho, J.(2006). Energetics of aliphatic deletions in proteincores. Protein Sci. 15, 1858–1872.

31. Dill, K. A. (1990). Dominant forces in protein folding.Biochemistry, 29, 7133–7155.

32. Ratnaparkhi, G. S. & Varadarajan, R. (2000). Ther-modynamic and structural studies of cavity forma-tion in proteins suggest that loss of packinginteractions rather than the hydrophobic effectdominates the observed energetics. Biochemistry, 39,12365–12374.

33. Prevost, M., Wodak, S. J., Tidor, B. & Karplus, M.(1991). Contribution of the hydrophobic effect toprotein stability: analysis based on simulations of theIle-96-Ala mutation in barnase. Proc. Natl Acad. Sci.USA, 88, 10880–10884.

34. McKnight, J. C., Doering, D. S., Matsudaira, P. T. &Kim, P. S. (1996). A thermostable 35-residue sub-domain within villin headpiece. J. Mol. Biol. 260,126–134.

35. Chiu, T. K., Kubelka, J., Herbst-Irmer, R., Eaton,W. A., Hofrichter, J. & Davies, D. R. (2005). High-resolution x-ray crystal structures of the villinheadpiece subdomain, an ultrafast folding protein.Proc. Natl Acad. Sci. USA, 102, 7517–7522.

36. Eicken, C., Sharma, V., Klabunde, T., Lawrenz, M. B.,Hardham, J. M., Norris, S. J. & Sacchettini, J. C.(2002). Crystal structure of lyme disease variablesurface antigen VlsE of Borrelia burgdorferi. J. Biol.Chem. 277, 21691–21696.

37. Jones, K. & Wittung-Stafshede, P. (2003). The largestprotein observed to fold by two-state kinetic mech-anism does not obey contact-order correlation. J. Am.Chem. Soc. 125, 9606–9607.

38. Xiao, S., Bi, Y., Shan, B. & Raleigh, D. P. (2009).Analysis of core packing in a cooperatively foldedminiature protein: the ultrafast folding villin head-piece helical subdomain. Biochemistry, 48, 4607–4616.

39. Bunagan, M. R., Gao, J., Kelly, J. W. & Gai, F.(2009). Probing the folding transition state structureof the villin headpiece subdomain via side chainand backbone mutagenesis. J. Am. Chem. Soc. 131,7470–7476.

40. Pace, C. N., Hebert, E. J., Shaw, K. L., Schell, D., Both,V., Krajcikova, D. et al. (1998). Conformationalstability and thermodynamics of folding of ribonu-cleases Sa, Sa2 and Sa3. J. Mol. Biol. 279, 271–286.

41. Takano, K., Scholtz, J. M., Sacchettini, J. C. & Pace,C. N. (2003). The contribution of polar group burial

527Hydrophobic Interactions to Protein Stability

to protein stability is strongly context-dependent.J. Biol. Chem. 278, 31790–31795.

42. Trevino, S. R., Gokulan, K., Newsom, S., Thurlkill,R. L., Shaw, K. L., Mitkevich, V. A. et al. (2007).Asp79 makes a large, unfavorable contribution tothe stability of RNase Sa. J. Mol. Biol. 354, 967–978.

43. Trevino, S. R., Schaefer, S., Scholtz, J. M. & Pace, C. N.(2007). Increasing protein conformational stabilityby optimizing β-turn sequence. J. Mol. Biol. 373,211–218.

44. Fu, H., Grimsley, G. R., Razvi, A., Scholtz, J. M. &Pace, C. N. (2009). Increasing protein stability byimproving β-turns. Proteins, 77, 491–498.

45. Fu, H., Grimsley, G. R., Scholtz, J. M. & Pace, C. N.(2010). Increasing protein stability: importance ofDelta Cp and the denatured state. Protein Sci. 19,1044–1052.

46. Shirley, B. A., Stanssens, P., Hahn, U. & Pace, C. N.(1992). Contribution of hydrogen bonding to theconformational stability of ribonuclease T1. Biochem-istry, 31, 725–732.

47. Myers, J. K., Pace, C. N. & Scholtz, J. M. (1997). Helixpropensities are identical in proteins and peptides.Biochemistry, 36, 10923–10929.

48. Grimsley, G. R., Shaw, K. L., Fee, L. R., Alston, R. W.,Huyghues-Despointes, B. M. P., Thurlkill, R. L. et al.(1999). Increasing protein stability by alteringlong-range coulombic interactions. Protein Sci. 8,1843–1849.

49. De Vos, S., Backmann, J., Prevost, M., Steyaert, J. &Loris, R. (2001). Hydrophobic core manipulations inribonuclease T1. Biochemistry, 40, 10140–10149.

50. Joh, N. H., Oberai, A., Yang, D., Whitelegge, J. P. &Bowie, J. U. (2009). Similar energetic contributions ofpacking in the core of membrane and water-solubleproteins. J. Am. Chem. Soc. 131, 10846–10847.

51. Pace, C. N. (1992). Contribution of the hydrophobiceffect to globular protein stability. J. Mol. Biol. 226,29–35.

52. Pace, C. N. (1995). Evaluating contribution ofhydrogen bonding and hydrophobic bonding toprotein folding. Methods Enzymol. 259, 538–554.

53. Fleming, P. J. & Richards, F. M. (2000). Proteinpacking: dependence on protein size, secondarystructure and amino acid composition. J. Mol. Biol.299, 487–498.

54. Liang, J. & Dill, K. A. (2001). Are proteins well-packed? Biophys. J. 81, 751–766.

55. Kuntz, I. D. & Crippen, G. M. (1979). Proteindensities. Int. J. Pept. Protein Res. 13, 223–228.

56. Kurochkina, N. & Privalov, G. (1998). Heterogeneityof packing: structural approach. Protein Sci. 7,897–905.

57. Schell, D., Tsai, J., Scholtz, J. M. & Pace, C. N. (2006).Hydrogen bonding increases packing density in theprotein interior. Proteins: Struct. Funct. Genet. 63,278–282.

58. Kajander, T., Kahn, P. C., Passila, S. H., Cohen, D. C.,Lehtio, L., Adolfsen, W. et al. (2000). Buried chargedsurface in proteins. Structure, 8, 1203–1214.

59. Faham, S., Yang, D., Bare, E., Yohannan, S.,Whitelegge, J. P. & Bowie, J. U. (2004). Side-chaincontributions to membrane protein structure andstability. J. Mol. Biol. 335, 297–305.

60. Joh, N. H., Min, A., Faham, S., Whitelegge, J. P.,Yang, D., Woods, V. L. & Bowie, J. U. (2008). Modeststabilization by most hydrogen-bonded side-chaininteractions in membrane proteins. Nature, 453,1266–1270.

61. Creamer, T. P., Srinivasan, R. & Rose, G. D. (1997).Modeling unfolded states of proteins and peptides.II. Backbone solvent accessibility. Biochemistry, 36,2832–2835.

62. Auton, M., Holthauzen, L. M. F. & Bolen, D. W.(2007). Anatomy of energetic changes accompanyingurea-induced protein denaturation. Proc. Natl Acad.Sci. USA, 104, 15317–15322.

63. Pace, C. N., Huyghues-Despointes, B. M., Fu, H.,Takano, K., Scholtz, J. M. & Grimsley, G. R. (2010).Urea denatured state ensembles contain extensivesecondary structure that is increased in hydrophobicproteins. Protein Sci. 19, 929–943.

64. Tanford, C. (1968). Protein denaturation. Adv. ProteinChem. 23, 121–282.

65. Tanford, C. (1970). Protein denaturation Part C. Adv.Protein Chem. 24, 1–95.

66. Fersht, A. R. (1987). The hydrogen bond in molecularrecogition. Trends Biochem. Sci. 12, 301–304.

67. Myers, J. K. & Pace, C. N. (1996). Hydrogenbonding stabilizes globular proteins. Biophys. J. 71,2033–2039.

68. Pace, C. N., Trevino, S., Prabhakaran, E. & Scholtz,J. M. (2004). Protein structure, stability, and solu-bility in water and other solvents. Philos. Trans. R.Soc. London, Ser. B, 359, 1225–1235.

69. Powers, E. T., Deechongkit, S. & Kelly, J. W. (2006).Backbone-backbone H-bonds make context-depen-dent contributions to protein folding kinetics andthermodynamics: lessons from amide-to-ester muta-tions. Adv. Protein Chem. 72, 39–78.

70. Gao, J., Bosco, D. A., Powers, E. T. & Kelley, J. W.(2009). Localized thermodynamic coupling betweenhydrogen bonding and the microenvironment polar-ity sustantially stabilizes proteins. Nat. Struct. Mol.Biol. 16, 684–690.

71. Pace, C. N. (2009). Energetics of protein hydrogenbonds. Nat. Struct. Mol. Biol. 16, 681–682.

72. Bowie, J. U. (2010). Membrane protein folding: howimportant are hydrogen bonds? Curr. Opin. Struct.Biol, 21, 42–49.

73. Hebert, E. J., Giletto, A., Sevcik, J., Urbanikova, L.,Wilson, K. S., Dauter, Z. & Pace, C. N. (1998).Contribution of a conserved asparagine to theconformational stability of ribonucleases Sa, Ba,and T1. Biochemistry, 37, 16192–16200.

74. Pace, C. N. (2001). Polar group burial contributesmore to protein stability than nonpolar group burial.Biochemistry, 40, 310–313.

75. Stickle, D. F., Presta, L. G., Dill, K. A. & Rose, G. D.(1992). Hydrogen bonding in globular proteins.J. Mol. Biol. 226, 1143–1159.

76. Harpaz, Y., Gerstein, M. & Chothia, C. (1994).Volume changes on protein folding. Structure, 2,641–649.

77. D'Aquino, J. A., Gomez, J., Hilser, V. J., Lee, K. H.,Amzel, L. M. & Freire, E. (1996). The magnitude ofthe backbone conformational entropy change inprotein folding. Proteins, 25, 143–156.

528 Hydrophobic Interactions to Protein Stability

78. Spolar, R. S. & Record, M. T. (1994). Coupling of localfolding to site-specific binding of proteins to DNA.Science, 263, 777–784.

79. Savage, H. J., Elliott, C. J., Freeman , C. M. & Finney,J. L. (1993). Lost hydrogen bonds and buried surfacearea: rationalising stability in globular proteins.J. Chem. Soc., Faraday Trans. 89, 2609–2617.

80. Fleming, P. J. & Rose, G. D. (2005). Do all backbonepolar groups in proteins form hydrogen bonds?Protein Sci. 14, 1911–1917.

81. Karp, D. A., Gittis, A. G., Stahley, M. R., Fitch, C. A.,Stites, W. E. & Garcia-Moreno, E. B. (2007).High apparent dielectric constant inside a proteinreflects structural reorganization coupled to theionization of an internal Asp. Biophys. J. 92,2041–2053.

82. Thurlkill, R. L., Grimsley, G. R., Scholtz, J. M. & Pace,C. N. (2006). Hydrogen bonding markedly reducesthe pK of buried carboxyl groups in proteins. J. Mol.Biol. 362, 594–604.

83. Pace, C. N., Shirley, B. A., McNutt, M. & Gajiwala, K.(1996). Forces contributing to the conformationstability of proteins. FASEB J. 76, 75–83.

84. Makhatadze, G. I. & Privalov, P. L. (1996). On theentropy of protein folding. Protein Sci. 5, 507–510.

85. Boczko, E. M. & Brooks, C. L., 3rd (1995). First-principles calculation of the folding free energyof a three-helix bundle protein. Science, 269,393–396.

86. Brady, G. P. & Sharp, K. A. (1997). Entropy in proteinfolding and in protein–protein interactions. Curr.Opin. Struct. Biol. 7, 215–221.

87. Pace, C. N., Laurents, D. V. & Thomson, J. A. (1990).pH dependence of the urea and guanidine hydro-chloride denaturation of ribonuclease A and ribonu-clease T1. Biochemistry, 29, 2564–2572.

88. Pace, C. N. (2000). Single surface stabilizer. Nat.Struct. Biol. 7, 345–346.

89. Reiner, A., Henklein, P. & Kiefhaber, T. (2010). Anunlocking/relocking barrier in conformational fluc-tuations of villin headpiece subdomain. Proc. NatlAcad. Sci. USA, 107, 4955–4960.

90. Lei, H., Wu, C., Liu, H. & Duan, Y. (2007). Foldingfree-energy landscape of villin headpiece subdomainfrom molecular dynamics simulations. Proc. NatlAcad. Sci. USA, 104, 4925–4930.

91. Kubelka, J., Henry, E. R., Cellmer, T., Hofrichter, J. &Eaton, W. A. (2008). Chemical, physical, and theo-

retical kinetics of an ultrafast folding protein. Proc.Natl Acad. Sci. USA, 105, 18655–18662.

92. Yang, J. S., Wallin, S. & Shakhnovich, E. I. (2008).Universality and diversity of folding mechanics forthree-helix bundle proteins. Proc. Natl Acad. Sci. USA,105, 895–900.

93. Frank, B. S., Vardar, D., Buckley, D. A. & McKnight,C. J. (2002). The role of aromatic residues in thehydrophobic core of the villin headpiece subdomain.Protein Sci. 11, 680–687.

94. Fu, H. (2009). Understanding forces that contributeto protein stability: applications for increasingprotein stability. Ph.D. dissertation., Texas A&MUniversity.

95. Hebert, E. J., Grimsley, G. R., Hartley, R. W., Horn,G., Schell, D., Garcia, S. et al. (1997). Purification ofribonucleases Sa, Sa2, and Sa3 after expression inEscherichia coli. Protein Expr. Purif. 11, 162–168.

96. Shaw, K. L. (2000). Reversing the net charge ofribonuclease sa, Texas A&M University, Ph.D.Dissertation. College Station, TX.

97. Pace, C. N. (1986). Determination and analysis ofurea and guanidine hydrochloride denaturationcurves. Methods Enzymol. 131, 266–280.

98. Grimsley, G. R., Huyghues-Despointes, B. M. P., Pace,C. N. & Scholtz, J. M. (2004). Measuring the confor-mational stability of a protein. In Purifying Proteins forProteomics (Simpson, R. J., ed.), pp. 535–566, ColdSpring Harbor Laboratory Press, Cold Spring Harbor,New York.

99. Pace, C. N., Horn, G., Hebert, E. J., Bechert, J., Shaw,K., Urbanikova, L. et al. (2001). Tyrosine hydrogenbonds make a large contribution to protein stability.J. Mol. Biol. 312, 393–404.

100. Berman, H., Henrick, K. & Nakamura, H. (2003).Announcing the worldwide Protein Data Bank. Nat.Struct. Biol. 10, 980.

101. Pace, C. N., Grimsley, G. R., Thomson, J. A. & Barnett,B. J. (1988). Conformational stability and activity ofribonuclease T1 with zero, one, and two intactdisulfide bonds. J. Biol. Chem. 263, 11820–11825.

102. McDonald, I. K. & Thornton, J. M. (1994). Satisfyinghydrogen bonding in globular proteins. J. Mol. Biol.238, 777–793.

103. Krieger, E., Koraimann, G. & Vriend, G. (2002).Increasing the precision of comparative models withYASARA NOVA—a self-parameterizing force field.Proteins, 47, 393–402.

Supplementary Data

Fig. 1. Urea unfolding curves for VHP (a) and VlsE (b) monitored by measuring circular

dichroism at 222 nm (VHP) or 220 nm (VlsE). The conditions were 50 mM Sodium Phosphate,

pH 7.0, 25 oC for VHP and 5 mM Sodium Phosphate, pH 7.0, 25 oC for VlsE. The solid lines are

curves calculated as described in Ref. (98) in the text, giving the urea 1/2 and m values in Tables I

and II. ∆G for unfolding as a function of urea concentration is also shown for VHP (c) and VlsE

(d). These data were analyzed using the linear extrapolation method (see also Ref. (98) in the

text), giving urea1/2 = 6.19 M, m = 0.452 kcal mol-1 and ΔG (H2O) = 2.8 kcal mol-1 for VHP, and

urea1/2 = 1.21 M, m = 3.88 kcal mol-1 and ΔG (H2O) = 4.7 kcal mol-1 for VlsE.

Fig. 2. Contribution of hydrogen bonds (a), and hydrophobic interactions (b) to the total stability

(c), plotted versus the number of residues for the proteins in Table 9..

Fig. 3. Number of buried nonpolar (a), polar uncharged (b), and polar charged (c) side chains

plotted versus the number of residues for the proteins in Table 9.

Fig. 1.

0 2 4 6 8 10

-30

-25

-20

-15

-10

-5

0

θ 222

(mde

g)

Urea (M)

(a)

Fig. 1

0.0 0.5 1.0 1.5 2.0 2.5 3.0

-80

-70

-60

-50

-40

-30

-20

-10

0

θ 220

(mde

g)

Urea (M)

(b)

Fig. 1

0 1 2 3 4 5 6 7 8-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Δ

G (k

cal m

ol-1)

Urea (M)

(c)

Fig. 1

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

-2

-1

0

1

2

3

4

5

Δ

G (k

cal m

ol-1)

Urea (M)

(d)

Fig. 2.

0 100 200 300 400 500 6000

100

200

300

400

500

600[11/9/2010 09:29 "/Graph5" (2455509)]Linear Regression through origin for moreprot119_Hbonds:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 0.98007 0.01856------------------------------------------------------------

R SD N P------------------------------------------------------------0.98919 21.26412 22

Fig. 2

0 100 200 300 400 500 6000

200

400

600

800

1000[11/9/2010 09:32 "/Graph6" (2455509)]Linear Regression through origin for moreprot119_HPHB:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 1.7233 0.02186------------------------------------------------------------

R SD N P------------------------------------------------------------0.99447 25.04889 22

Fig. 2.

0 100 200 300 400 500 6000

200

400

600

800

1000

1200

1400

[12/1/2010 08:19 "/Graph6" (2455531)]Linear Regression through origin for NewPlots121_TotStab:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 2.40924 0.02374------------------------------------------------------------

R SD N P------------------------------------------------------------0.99685 27.20127 22

Fig. 3.

0 100 200 300 400 500 6000

50

100

150

200[11/12/2010 07:17 "/Graph11" (2455512)]Linear Regression through origin for Kajander_BurNP:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 0.3628 0.00654------------------------------------------------------------

R SD N P------------------------------------------------------------0.99071 7.49188 22

Fig. 3.

0 100 200 300 400 500 6000

20

40

60

80

100[1/14/2011 09:45 "/Fig4bcor" (2455575)]Linear Regression through origin for Fig4bcorr_BPNC:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 0.15457 0.00925------------------------------------------------------------

R SD N P------------------------------------------------------------0.92699 10.60267 22

Fig. 3.

0 100 200 300 400 500 6000

10

20

30

40

50

60 [11/12/2010 07:28 "/Graph11" (2455512)]Linear Regression through origin for Kajander_BurPC:Y = B * X

Parameter Value Error------------------------------------------------------------A 0 --B 0.11743 0.0045------------------------------------------------------------

R SD N P------------------------------------------------------------0.94886 5.15192 22

Table 1. Parameters characterizing the thermal unfolding of RNase T1 and hydrophobic

variants in 30 mM MOPS, pH 7.0 a,b

Variant

Tm ( oC)

ΔHcal

(kcal mol-1)

ΔHvH (kcal mol-1)

ΔHvH ⁄ ΔHcal

WT 51.92 104.5 109.0 1.04

V16A 43.36 85.9 101.4 1.18

I61V 45.84 101.5 98.3 0.97

V78A

38.50

38.8

99.7

2.57

V78I

50.30

102.0

89.5

0.88

V79A

38.06

76.7

105.5

1.38

L86A 39.20 - - -

L86I 44.96 94.6 94.7 1.00

V89A

39.86

66.6

98.3

1.48

I90V

50.26

99.6

104.6

1.05

a From Gajiwala, K.S. Structure and Stability of Ribonuclease T1: A Crystallographic

and Calorimetric Study. Ph.D. Dissertation, Texas A&M University, December 1994.

b The thermal denaturation curves were determined from DSC scans. The data in the

table are the average values of two or more experiments when possible.

Table 2. Observed and accessibility corrected Δ(ΔG) values for the hydrophobic variants of

RNase T1 obtained from DSC experiments.a

Variant

Δ(ΔG)obsb

(kcal mol-1)

Side Chain % Buried c

Δ(ΔG)corr (kcal mol-1) per −CH2−

V16A − 2.75 97.9 − 2.81 − 1.4

I61V − 1.95 97.3 − 2.00 − 2.0

V78A

− 4.31

100.0

− 4.31

− 2.2

V78I

− 0.52

-

− 0.52

-

V79A

− 4.45

100.0

− 4.45

− 2.2

L86A − 4.09 74.5 − 5.49 − 1.8

L86I − 2.23 - − 2.23 -

V89A

− 3.88

100.0

− 3.88

− 1.9

I90V

− 0.53

99.9

− 0.53

− 0.5

a From Gajiwala, K.S. Structure and Stability of Ribonuclease T1: A Crystallographic and

Calorimetric Study. Ph.D. Dissertation, Texas A&M University, December 1994.

b ∆(ΔG) = ∆ Tm (from Table 1) x the Δ Sm value of WT. ∆ Tm was a reference temperature

between the respective Tm values.

c Calculated using the Lee and Richards program (Lee& Richards (1971) J Mol Biol 55:

379-400).

Table 3. Eleven proteins and their 138 variants used in the analysis of aliphatic hydrophobic

mutants a

Protein

Ile → Val

Val → Ala

Ile → Ala

Leu → Ala

Staph Nuc

(28)

5

8

5

10

T4 Lysozyme

(25)

4

4

7

10

Barnase

(15)

6

2

6

1

FKBP (15)

1

8

3

3

apoflavo

(14)

3

4

1

6

hLysozyme

(11)

5

6

0

0

CI2 (9)

3

2

2

2

RNase Sa

(7)

4

0

0

3

RNase T1

(5)

2

2

0

1

AcCoA BP

(5)

0

0

2

3

Ubiquitin

(4)

0

3

0

1

Total

33

39

26

40

a The abbreviations used for the proteins in Tables 1 to 5 and the primary reference in the

text for the experimental data are as follows: staphylococcal nuclease (Staph Nuc),

References 18,19; T4 Lysozyme, References 21-23; barnase, References 15-17; FK506-

binding protein (FKBP), Reference 25; apoflavodoxin (apoflavo) Reference 30; human

lysozyme (hLyso) Reference 26; chymotrypsin inhibitor 2 (CI2), Reference 24;

ribonuclease T1 (RNase T1), This Paper and Reference 49; ribonuclease Sa (RNase Sa),

This paper; acyl-coenzyme A binding protein (AcCoA BP), Reference (Kragelund, B.B.,

Osmark, P., Neergaard, K., Schiodt, J., Kristiansen, K., Knudsen, J., Poulsen, F.M.

(1999) Nature Structural Biology, 6, 594-601); Ubiquitin, Reference 28.

Table 4. Data used to calculate ∆(∆G)values in Table 7 for the Ile → Val substitutions.

Protein

Residue

∆(∆G)obs

(kcal mol-1)

Side Chain % Buried

∆(∆G)corr

(kcal mol-1)

Staph Nuc

15

0.80

71

1.13

18 1.10 84 1.31 72 1.80 95 1.89 92 0.50 100 0.50 139 1.50 84 1.79

Barnase 25 1.12 89 1.26 51 1.80 100 1.80 76 0.82 100 0.82 88 1.34 100 1.34 96 0.88 100 0.88 109 0.76 86 0.88

T4 Lysozyme 3 0.40 99 0.40 78 0.80 100 0.80 100 0.40 94 0.43 150 2.10 91 2.31

hLysozyme 23 0.39 93.5 0.42 56 1.27 100 1.27 59 1.02 100 1.02 89 0.56 98.3 0.57 106 0.85 95.2 0.89

apoFlavo 22 1.37 100 1.37 51 1.83 100 1.83 52 1.40 100 1.40

CI2 39 1.20 100 1.20 48 1.00 95 1.05 76 -0.10 100 -0.10

FKBP 76 0.76 100 0.76 RNase T1 61 2.05 97 2.11

90 0.95 100 0.95 RNase Sa 22 1.48 81 1.83

70 0.94 100 0.94 71 0.13 100 0.13 92 1.00 100 1.00

Average

1.04

1.10

St. Dev. 0.53

0.57

Table 5. Data used to calculate ∆(∆G)values in Table 7 for the Val → Ala substitutions.

Protein

Residue

∆(∆G)obs

(kcal mol-1)

Side Chain %

Buried

∆(∆G)corr

(kcal mol-1)

Staph Nuc

23

2.90

100

2.90

39 2.20 100 2.20 66 2.20 100 2.20 74 3.10 100 3.10 99 3.20 100 3.20 104 2.90 100 2.90 111 4.70 100 4.70 114 0.00 92 0.00

Barnase 10 3.39 100 3.39 45 1.75 85 2.06

T4 Lysozyme 87 1.70 98 1.73 103 2.20 96 2.29 111 1.30 100 1.30 149 3.20 100 3.20

CI2 66 4.93 100 4.93 70 2.09 95 2.20

FKBP 2 2.43 97 2.51 4 2.78 90 3.09 23 2.97 100 2.97 24 3.19 100 3.19 55 2.13 97 2.20 63 2.97 100 2.97 98 2.16 85 2.54 101 2.75 100 2.75

hLysozyme 93 0.87 99 0.88 99 0.93 99 0.94 100 0.42 100 0.42 121 1.59 88 1.81 125 1.46 83 1.76 130 0.98 97 1.01

apoflavo 31 1.59 100 1.59 117 2.01 93 2.16 139 0.88 100 0.88 160 1.86 100 1.86

Ubiquitin 5 2.99 100 2.99 17 1.98 100 1.98 26 3.30 100 3.30

RNase T1 16 2.50 98 2.55 78 4.10 100 4.10

Average St. Dev

2.32 1.08

2.38 1.07

Table 6. Data used to calculate ∆(∆G)values in Table 7 for the Ile → Ala substitutions.

Protein

Residue

∆(∆G)obs

(kcal mol-1)

Side Chain % Buried

∆(∆G)corr

(kcal mol-1)

Staph Nuc

15

2.70

71

3.80

18 2.50 84 2.98 72 5.10 95 5.37 92 4.00 100 4.00 139 3.50 84 4.17

Barnase 25 3.52 89 3.96 51 4,71 100 4.71 76 1.89 100 1.89 88 4.01 100 4.01 96 3.17 100 3.17 109 2.07 86 2.41

T4 Lysozyme 3 0.70 99 0.71 27 3.10 100 3.10 29 2.70 99 2.73 50 2.00 97 2.06 58 3.20 100 3.20 78 1.20 100 1.20 100 3.40 94 3.62

apoflavo 109 3.56 100 3.56 CI2 48 3.81 95 4.01

76 4.16 100 4.16 FKBP 56 2.48 90 2.76

76 3.81 100 3.81 91 1.54 99 1.56

AcCoABP 39 1.06 100 1.06 74 1.34 91 1.47

Average

2.89

3.06

St. Dev. 1.16

1.21

Table 7. Data used to calculate ∆(∆G)values in Table 7 for the Leu → Ala substitutions.

Protein

Residue

∆(∆G)obs

(kcal mol-1)

Side

Chain % Buried

∆(∆G)corr

(kcal mol-1)

Staph Nuc

7

1.60

89

1.80

14 2.30 86 2.67 25 2.70 100 2.70 36 3.50 100 3.50 37 1.70 93 1.83 38 1.70 97 1.75 89 2.60 83 3.13 103 4.60 100 4.60 108 5.80 94 6.17 125 4.90 100 4.90

Barnase 14 4.32 100 4.32 T4 Lysozyme 7 2.30 100 2.30

33 2.90 100 2.90 46 2.60 100 2.60 66 3.30 98 3.37 84 3.80 100 3.80 91 2.60 100 2.60 99 5.00 100 5.00 118 3.50 94 3.72 121 2.70 100 2.70 133 3.60 100 3.60

CI2 27 2.69 98 2.74 68 3.76 100 3.76

FKBP 50 2.57 92 2.79 97 3.56 100 3.56 106 2.32 93 2.49

apoflavo 6 2.90 100 2.90 44 1.69 100 1.69 50 2.83 100 2.83 105 2.48 100 2.48 143 0.18 100 0.18 105 2.61 100 2.61

AcCoABP 15 3.57 98 3.64 25 2.02 85 2.38 80 4.57 97 4.71

RNase T1 86 4.40 74 5.95 RNase Sa 8 1.34 84 1.60

11 3.10 97 3.20 19 3.60 95 3.79

Ubiquitin 67 2.82 100 2.82

Average St. Dev

3.03 1.12

3.15 1.19

Contribution of Hydrophobic Interactions to Protein StabilityIntroductionResultsVlsEVHPRNase SaRNase T1Other proteins

DiscussionHydrophobic interactions contribute more to the stability of a large protein than a small proteinForces contributing to protein stabilityForces contributing to protein instabilityStability of VHPConcluding remarks

Materials and MethodsAcknowledgementsSupplementary DataReferences