7/28/2019 Hidrofobicidad de Aminoacidos en Proteinas Globulares

1/6

Hydrophobicity of Amino Acid Residues in Globular Proteins

Author(s): George D. Rose, Ari R. Geselowitz, Glenn J. Lesser, Richard H. Lee, Micheal H.

Zehfus

Source: Science, New Series, Vol. 229, No. 4716, (Aug. 30, 1985), pp. 834-838

Published by: American Association for the Advancement of Science

Stable URL: http://www.jstor.org/stable/1695273

Accessed: 22/04/2008 19:49

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=aaas.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

http://www.jstor.org/stable/1695273?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=aaashttp://www.jstor.org/action/showPublisher?publisherCode=aaashttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1695273?origin=JSTOR-pdf

7/28/2019 Hidrofobicidad de Aminoacidos en Proteinas Globulares

2/6

Hydrophobicity of Amino AcidResidues in Globular Proteins

George D. Rose, Ari R. Geselowitz, Glenn J. LesserRichard H. Lee, Micheal H. Zehfus

A hydrophobic substance is readilysoluble in nonpolar solvents but onlysparinglysoluble in water (1). The hy-drophobic effect is believed to play amajorrole in organizing he self-assem-bly of water-soluble, globularproteins(1-4). Upon folding, residues with non-polar side chains that are driven fromwaterwill comprisea molecular nteriorwhere they can be shieldedfrom solvent

tophan and tyrosine to be very hydro-phobic, while Wolfendenand co-work-ers (2) find these residues to be veryhydrophilic. Scales are compared andtheirdifferenceshave beendiscussed(6).We nowderivetwo new scales that arebasedon accessibilityto solvent forresi-dues withinproteinsof knownstructure.These scales measure wo quantities hatcan be distinguished:Abstract. During biosynthesis, a globular protein folds into a tight particle with aninterior core that is shielded from the surrounding solvent. The hydrophobic effect isthought to play a key role in mediating this process: nonpolar residues expelled fromwater engender a molecular interior where they can be buried. Paradoxically, resultsof earlier quantitative analyses have suggested that the tendency for nonpolarresidues to be buried withinproteins is weak. However, such analyses merely classifyresidues as either "exposed" or "buried." In the experiment reported in this articleproteins of known structure were used to measure the average area that each residueburies upon folding. This characteristic quantity, the average area buried, iscorrelated with residue hydrophobicity.

access. The effect is analogous to thesegregation of oil in water, with theimportant distinction that residues inproteins are covalently bound to theirchain neighborsandcannotpartition n-dependently.To quantitate his effect, many scalesof hydrophobicity or the amino acids,their residues, and their analogs havebeen proposed(1-5). Such scales can beclassifiedas solutionmeasurements, m-pirical calculations, or some combina-tion of the two. Solutionscales arebasedon distributioncoefficients between anaqueous phase and a suitably chosenorganicphase, whileempirical cales arebased on partitioningbetween the sol-vent accessible surface and the buriedinterior n proteinsof known structure.Significant differences exist amongscales. Residues that are stronglyhydro-phobic on one scale may appearto bestrongly hydrophilic on some otherscale. For example, in solutionmeasure-ments, Nozaki and Tanford 1)findtryp-834

1) The area lost when a residue istransferredrom a definedstandard tateto a folded protein. The area a residueburiesuponfoldingis proportionalo itshydrophobiccontribution o the confor-mational free energy, AGconf 7).2) Thefractionalaccessibilityof a res-idue, defined as its meanaccessible areain proteinmolecules dividedby the stan-dardstatearea. The fractionalaccessibil-ity is an intrinsic measureof hydropho-bicity.Althoughrelated, these quantitiesarenot equivalent. For example, a bulkyarginine esiduemakes a largehydropho-bic contribution to AGconf because itsarea loss upon folding is large, approxi-mating hat of leucine. Yet, thefractionalaccessibility of an arginine s compara-tively highbecause the remaining nbur-ied area is also large.In his influential review (8), Kauz-mann used model compoundsto showthat burial of hydrophobicgroups is asignificant ource of stabilization nergy

in proteins. More than a decade later,analysisof x-rayelucidatedproteinsdis-closed that many hydrophobic groupsremain unburied(3-4, 9). Summarizingthese findings,Richardswrote (7):Of the accessible areas of native structures,roughlyhalf representspolaratoms and halfnonpolaratoms. Thus the 'grease' is by nomeans all 'buried'. In the folding processthere areroughlyequivalentdecreases in theaccessibilityof both the polar and nonpolargroups. The relevant forces and the finalstructure equiremore carefuldefinition hanis impliedby the commonfeelingthat insideequals nonpolarand outsideequals polar.

Subsequent studies revealed furthercomplexity: hecorrelation f hydropho-bicity with total residue surface area isexcellent (10), but the correspondingcorrelation with residue area buriedupon foldingis poor (3, 4).Thus, it has been widely accepted(2-4, 11) that only a weak relation existsbetween the tendencyfor residuesto beburiedwithin proteinsand their hydro-phobicity, when measured as the freeenergyof transfer rom waterto organicsolvent (1). However, we now reportfindings hat leadto the oppositeconclu-sion, revealing a strong correlationbe-tween hydrophobicityand the surfacearea residuesbury uponfolding.Calculationof solvent-accessible sur-face areas. The solvent-accessible sur-face area (9) was calculated for 4410residues takenfrom monomersof 23 x-rayelucidatedproteins(12);atomicradiiused werethoseof Richards 9).Normal-ized distribution functions were thencompiled,as shown in Fig. 1. Each dis-tribution s, in effect, a histogram how-ing the percentageof residues of eachtype that are:fully accessible, 5 percentburied,10percentburied,and so on untilcompletelyburied.Accessibility must be measuredrela-tive to a defined standardstate. Twoways of specifyinga standard tate havebeen discussed (13, 14). The extendedstandard tate for a residue, X, is takento be the surface area of that residue inthe extended tripeptideGly-X-Gly(seeFig. 2), with dihedralangles (>= -140?,0 = 135, Xi = -120?, and X2...N=180? 13). In the present study we usea stochastic standardstate, defined asthe meanaccessibilityof an ensemble oftripeptideshavingdihedralangles takenfromthe observed distribution n actualproteins 14).Forexample,if there areN

RichardLee and MichealZehfusarepostdoctoralfellows, AriGeselowitzandGlennLesseraremedi-cal students,andGeorgeRoseis an associateprofes-sor in the Department f BiologicalChemistry,TheMiltonS. HersheyMedicalCenter,The Pennsylva-nia StateUniversity,Hershey17033.SCIENCE,VOL. 229

7/28/2019 Hidrofobicidad de Aminoacidos en Proteinas Globulares

3/6

phenylalanines in the database of pro-teins, then N tripeptides, Gly-Phe-Gly,are constructed with angles(i, 4i, Xi(1)where i = 1, 2, . . N

The mean accessibility for phenylalanineand its substituent groups is the ensem-ble average. Physically, the stochasticstandard state reflects the degree towhich residues are buried by backboneatoms from covalent neighbors in theensemble. Both extended and stochasticstandard states give similar values; indi-vidual residue differences range between1 and 15 percent.Surface area and hydrophobicity.Mean accessibilities, , were calcu-lated for the distributions shown in Fig.1. Using the mean and standard stateaccessibilities, we derived several char-acteristic quantities for each residue.These are listed in Table 1 and include:1) AO, the stochastic standard state

accessibility-that is, the solvent acces-sible surface area of a residue in thestandard state.2) , the mean solvent-accessiblesurface area-that is, the average sol-vent accessible surface area of a residuein folded proteins.3) A0 - , the mean area buriedon transfer from the standard state to thefolded protein (proportional to the hy-drophobic contribution to AGconf).4) (A? - )/IA, the mean fraction-al area loss, denoted f, where

f= 1 - (/A?)5) AGt the Nozaki-Tanford (1) freeenergy of transfer from water to organicsolvent, in kilocalories per mole (scaleincludes 11 residues only).The mean fractional area loss,f, is theaverage area a residue buries upon fold-ing, normalized by its standard statearea. The relation between area buried

upon folding and standard state area ispresented graphically in Fig. 3. The resi-dues can be divided into three groups: (i)hydrophobic, including (Gly), Ala, Cys,Fig. 1. Normalizeddistribution unctions ofaccessibility to solvent for the 20 residues,calculatedfrom proteinsof known structure(12). Each functionis, in effect, a histogramshowing the percentage of residues of thattype thatare:fully accessible, 5 percentbur-ied, 10percentburied ..., completelybur-ied. An integratedcurve is superimposedonthe histograms;hecurve is a plotof percent-ageof residuesthat are0 to 5 percentburied,0 to 10 percentburied . . . 0 to 95 percentburied,0 to 100 percent buried. (By defini-tion, the integratedcurve must rangefrom0to 100percentof total numberof residuesasthe proportionof buried residues is variedfrom 0 to 100percent.)30AUGUST 985

Table 1. Mean (designated) and standardstate surfaceareas (A?)for the amino acidresidues.A A -............ AGtResidue* (

7/28/2019 Hidrofobicidad de Aminoacidos en Proteinas Globulares

4/6

X2

C,93X, ~ HC Cx N

\ x ~~~~~~~~~~~glycagl N (0 V) CaIIH 0

IPhe

Fig. 2. Degrees of freedom in the tripeptideunit Gly-X-Gly.4 and i are backbonedihe-dral angles, and X1, X2, . . are side-chaindihedralangles.

Val, Ile, Leu, Met, Phe, and Trp; (ii)moderately polar, including (Gly), Ser,Thr, His, and Tyr; and (iii) very polar,including (Gly), Pro, Asp, Asn, Glu, Gln,and Arg.For each of the above groups, theobserved linear relationship extrapolatesback through glycine, the logical firstmember of any group. Equations for thestraight lines of best-fit (and standarderrors) for each group areHydrophobicA? = 1.0(?0.05) ?(A? - )+ 17.3(+7.9)Moderately polarA? = 1.2(+0.07) ?(A? - )+ 18.6(+9.3)Very polarA? = 1.7(+0.08) ?(A? - )- 9.6(+8.4)

'lieVal

0 80 160 240Average area buried upon folding (A2)

Fig. 3. (Left)Plot of averageareaburieduponfolding, A? - , against the standardstate area,A?, for all 20 residues. The meanfractionalarealoss, f, defined as the averageareaburieddividedby the standard tatearea,is an intrinsic measure of residuehydropho-bicity. The straight ines of best-fit are givenin Eqs. 1, 2, and 3. (Right)A computer-generatedpace-fillingmodel of the proteinhen egglysozyme, showing the largest cross section throughthe molecule (19). The color codingscheme,based on the data n the leftpanel,divides the residues ntothreegroups:hydrophobic(purple),moderatelypolar(blue),andvery polar (orangered). With the sequencesencodedinthis way, it can be seen that hydrophobicresidues tend to clusterin the interior,while verypolarresidues are dispersedto the exterior.

TrpbiPhe

* Tyr

Leu1600 -

Met

? Thr

i I? Ser I0 80 160 240

Area in standard state (,2)

* Met

* His Thr 0 His

i Ser I I

(1)

(2)

(3)The form of Eqs. 1 to 3 is:A? = m . (average area buried) + bThe reciprocal slope, l/m, reflects theincremental area buried, on average, foreach increment of standard state area.

Equations 1 to 3 are interpreted to meanthat beyond a threshold quantity, b, hy-drophobic residues are fully buried(m = 1), moderately polar residues are83 percent buried (m = 1.2), and verypolar residues are 59 percent buried(m = 1.7).Another interpretation of Fig. 3 is alsoconsistent with the data. If groups 1 to 3are extrapolated back through zero, in-stead of through Gly, the slopes aresimilar and the standard errors are com-parable. In this case, Gly is included ingroup 2 only. The resultant equationshave no residual threshold and are inter-preted to mean that, on average, hydro-phobic residues are still fully buried,moderately polar residues are 76 percentburied, and very polar residues are 63percent buried.In Fig. 3, left, proline appears to bemore polar than anticipated, a probableconsequence of its frequent occurrencein reverse turns which are usually situat-ed at the protein surface (6). Lysine is anoutlier, appearing more polar than anyother residue. In our more detailed anal-ysis lysine backbone and beta carbon(Cp) atoms were observed to behavenormally, while C,, Ca, and C, appear tohave anomalous values off. Such behav-

0 80 160 240Average area buried upon folding (A2)

Fig.4 (a)Plot of stochastic standard tatearea, AO, ersus the Nozaki-Tanfordreeenergyof transfer romwater o organic olvent,AG (1). Theline, given by Eq. 4, was best-fit o Gly andmembersof the hydrophobic eries:Ala, Val, Leu, Phe, andTrp.(b)Plot of the averageareaburiedupon folding,A? - , as a functionof the Nozaki-Tanfordree energyof transfer,AGO1). Theline, givenby Eq. 5, was best-fitto Glyandmembersof thehydrophobic eries. The arearesiduesburyupontransferrom hestochasticstandard tateto the mean oldedstate s linearly e-lated to the Nozaki-Tanford alues.

240cu

5)o