Proc. Nati. Acad. Sci. USAVol. 86, pp. 2321-2325, April 1989Immunology

Comparison of the secondary structures of human class I and classII major histocompatibility complex antigens by Fourier transforminfrared and circular dichroism spectroscopy

(HLA antigens)

JOAN C. GORGA*t, AICHUN DONG*, MARK C. MANNINGt§, ROBERT W. WOODYf, WINSLOW S. CAUGHEYt,AND JACK L. STROMINGER**Department of Biochemistry and Molecular Biology, Harvard University, Cambridge, MA 02138; and tDepartment of Biochemistry, Colorado StateUniversity, Fort Collins, CO 80523

Contributed by Jack L. Strominger, December 21, 1988

ABSTRACT We have examined the secondary structuresof human class I and class IT histocompatibility antigens insolution by Fourier transform infrared spectroscopy andcircular dichroism in order to compare the relative amounts ofa-helix, fB-sheet, and other structures, which are crucialelements in the comparison of the protein structures. Quanti-tation of infrared spectra of papain-solubilized HLA-A2,HLA-B7, and DR1 in phosphate buffer gave a-helix contents of17%, 8%, and 10% and fl-sheet contents of 41%, 48%, and53%, respectively. By circular dichroism, papain-solubilizedHLA-A2, HLA-B7, and DR1 were also found to have compa-rable a-helix contents (e.g., 8%, 20%, and 17%, respectively).Circular dichroism analysis for fl-sheet gave 29% for papain-solubilized HLA-B7 and 42% for papain-solubilized DR1. Thevalue for papain-solubilized HLA-A2 (74%) was anomalous. Itis proposed that Trp-107 of HLA-A2, missing in both HLA-B7and DR1, may be responsible for much of the anomaly. Due tothe uncertainties inherent in quantitation of the amounts ofsecondary structures by both spectral methods, the differencesin the contents of a-helix and f-sheet in the three proteins arenot considered significant. However, differences in the natureof the fl-sheet structures are suggested by infrared spectros-copy. These results provide physical evidence for an overallstructure of class II antigens modeled on that of class I antigens.

Considerable evidence exists that the structures ofclass I andclass II histocompatibility antigens are similar. Most of theevidence (summarized in refs. 1 and 2) is based on sequencehomologies and similarities in domain structure at both theprotein and DNA levels. In addition, some T cells that arespecific for either class I or class II molecules use the samereceptor (3, 4). However, the secondary structures of purifiedclass I and class II antigens have not been directly compared.

In the crystal structure of the papain-solubilized class Iantigen HLA-A2 (A2pap) -42% of the amino acid residuesform antiparallel 8-pleated sheets (5). The membrane-proximal domains, a3 and 832-microglobulin (f32m), are f8-sandwich structures similar to the structure described forimmunoglobulin constant regions. The membrane-distal do-mains, al and a2, which do not show sequence homology toimmunoglobulin constant regions, consist of an antiparallelpB-pleated sheet under a long a-helical region. Approximately20% of the amino acid residues compose this set of a-helices,which form the sides of a peptide-binding groove at themembrane-distal surface of the molecule (6).The two membrane-proximal domains, a2 and p2, of class

II antigens also show strong sequence homology to immu-noglobulin constant regions, whereas the membrane-distal

domains, al and ,1, do not (1). A model for the membrane-distal domains of class II antigens based on the crystalstructure of A2pap has been proposed (2). In this manuscript,Fourier transform infrared (FTIR) spectroscopy and circulardichroism (CD) Etre used to compare the secondary structureof the papain-solubilized class II antigen DR1 (DRlpap) withthe secondary structures of the papain-solubilized class Iantigens HLA-A2 (A2pap) and HLA-B7 (B7pap) and with thoseof purified human p2m, bovine immunoglobulin G (IgG), andlysozyme, in order to address the question of whether a classII antigen has amounts of a-helix and 1-sheet consistent witha class II model based on the class I crystal structure.

MATERIALS AND METHODSProtein Sources. Purified human urinary f2m was obtained

from A. R. Sanderson (Serological Reagents Limited, En-gland); bovine IgG and lysozyme were purchased fromSigma; papain-solubilized HLA-A2 and HLA-B7 were puri-fied as described (7, 8); DR1 was purified by immunoaffinitychromatography and then solubilized by digestion with pa-pain (9).FTIR Spectroscopy. Spectra for solutions of A2pap, B7papq

DR1pap, p2m, IgG, and lysozyme at 20'C were measured incells with CaF2 windows (10) and a pathlength of 6 g.m usinga Perkin-Elmer model 1800 FTIR spectrophotometer at 2cm-1 resolution in the single beam mode. One thousand scanswere averaged for data recorded at 1-cm-1 intervals from4000 to 1000 cm-". Background spectra were recorded underidentical conditions with only buffer in the cell. The proteinspectrum was obtained by digital subtraction of the spectrumof the medium from the spectrum observed for the proteinsolution. Prior to any further manipulation, the data weresmoothed with a nine-point Savitsky-Golay function (11) toremove the possible white noise. The Perkin-Elmer Enhancefunction, which is analogous to the method developed byKauppinen et al. (12), with the parameters set to a half-bandwidth of 16 cm-1 and a K value of 2.3, was used toachieve spectral enhancement; second-derivative spectrawere obtained with Savitsky-Golay derivative function soft-ware for a five data point window. The method used todetermine protein secondary structure from second-derivative amide I spectra will be described in detail else-where (A.D., P. Huang, and W.S.C., unpublished data).

Abbreviations: A2 papain-solubilized HLA-A2; Wpap, papain-solubilized HLA-W5; 82m, 82-microglobulin; DR1 papain-solubilized DR1; FTIR, Fourier transform infrared; MlIC, majorhistocompatibility complex.tTo whom reprint requests should be addressed.§Present address: Department of Pharmaceutical Chemistry, Uni-versity of Kansas, Lawrence, KS 66045.

2321

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Janu

ary

13, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989)

CD Spectroscopy. CD spectra for A2pap, B7pap, DRipap, andl32m were obtained with each protein at =600,ug/ml in 5 mMsodium phosphate (pH 7.75) or 5 mM Tris (pH 7.75). Proteinconcentration was determined by amino acid analysis of thesamples used for CD spectroscopy. The spectra were ob-tained on a JASCO J41C circular dichrograph using a silicacell of 0.1- or 0.5-mm pathlength with a spectral bandwidth of1 nm and a time constant of 4 sec. The results of four or fivescans were averaged. The instrument was calibrated (13) with(+)-10-camphorsulfonic acid.

RESULTSFTIR spectroscopy has been shown to be a sensitive indi-cator of protein secondary structure (14). Because of thestrong absorption of water in the IR, FTIR of proteins hasnormally been done in 2H20. Recent technical advances inFTIR allow IR spectroscopy of proteins in aqueous solution(15) and therefore allow direct comparison with the results ofCD studies. With the use of very short pathlengths (in therange of 6-10 ,um) it is possible to measure the IR spectra ofproteins in aqueous solution with excellent signal-to-noiseratios. In addition, FTIR microscopy can be used to take IRspectra of very small samples (16) and of single cells (17) orcrystals.The amide I region of protein IR spectra contains strong

absorption bands due almost entirely to the C-O stretchvibration of the peptide linkages that constitute the backbonestructure (18). FTIR spectral enhancement (19) and second-derivative analysis have been shown to be a reliable indicatorof the component peak positions (20). The peak intensitiescan be used for a quantitative measure of the relative amountof each component. The amide I regions of FTIR spectra oflysozyme, IgG, and 182m in aqueous solution, along withFTIR spectral enhancement and second-derivative analysis(Fig. 1), revealed bands that could be assigned to a-helix,,3-sheet, turns, or unordered chain (A.D., P. Huang, andW.S.C., unpublished data). Amide I band frequencies aresensitive to secondary structure, as demonstrated in the

lysozyme and IgG spectra (Fig. 1). The lysozyme spectrumconsists of a major absorption centered at 1657 cm-' due toa-helix, a set of peaks between 1627 and 1642 cm-' due to/3-sheet, a set of peaks between 1666 and 1688 cm-' due toturns, and a peak at 1650 cm-' due to unordered structures.This secondary structure assignment is consistent with thecrystal structures of lysozyme (21, 22). In contrast, the IgGFTIR spectrum shows only a small amount of a-helix (at 1657cm-') and a large contribution from /-sheet (the secondderivative bands at 1642, 1638.5, 1632, and 1627 cm-') withsome unordered structure present (at 1650 cm-'). These IgGassignments are also consistent with x-ray crystal (23) andCD (24) data (see also Table 1).The 162m amide I spectrum is strikingly dissimilar to the

lysozyme spectrum and is nearly identical to the IgG spec-trum. As expected, and in agreement with x-ray (5, 25) andCD data (refs. 26-29; see Fig. 3), f2m is rich in /3-sheet withsome unordered structure and very little a-helix.The amide I spectra of the class I antigens A2pap and B7pap

and the class II antigen DR1pap are shown in Fig. 2. All threeproteins show peaks ascribable to a-helix (at 1657.5 cm-') and,/-sheet (at 1642-1643, 1638.5, 1632, and 1627 cm-') as well aspeaks ascribable to unordered chain and turns. The peaks at1657.5 cm-', due to a-helix, are considerably stronger thanthose due to a-helix at 1657 cm-' in the IgG and /32m spectra.The characters of the /3-sheet bands differ among the threeantigens, with those for A2pap and DR1pap very similar to eachother but different from the B7pap bands; however, the relativeproportions ofa-helix and /3-sheet are quite similar for all threeproteins. The relative proportions of secondary structureswere quantitatively estimated from the areas under the peaksin the second-derivative spectra (Table 1).The secondary structures of the MHC antigens were also

examined by CD spectroscopy. CD is a well-establishedtechnique for examining the secondary structure of proteins(30). Estimates of a-helix content are generally quite reliable,whereas /3-sheet contents are subject to greater uncertainty(31). The CD spectrum of/82m (Fig. 3) was in good agreementwith previously reported studies (26-29) and was consistent

1700 1650 1600 1700 1650 1600 1700 1650 1600

WAVENUMBER (cm 1)FIG. 1. IR spectra of lysozyme, IgG, and f32m in the amide I region. In each, the upper curve shows the difference spectrum-i.e., the

observed spectrum of the protein solution minus the spectrum of the medium in which the protein was dissolved; the middle curve shows theFTIR enhanced spectrum; the lower curve shows the second derivative of the FTIR enhanced spectrum. The spectra were obtained with theproteins at concentrations of between 10 and 15 mg/ml in 5 mM sodium phosphate (pH 7.4).

2322 Immunology: Gorga et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

13, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989) 2323

1 7 0 0 1650 1600 1700 1650 1600 1700 1 6 5 0 1 6 0 0

WAVENUMBER (cm-1)

FIG. 2. IR spectra of A2a, B7pap, and DR1pap in the amide I region, as described in the legend to Fig. 1.

with substantial amounts of 1-sheet and little or no a-helix(31). Quantitative analysis of the CD spectra (Table 1) gave0% a-helix and 59% 3-sheet, in reasonable agreement withthe x-ray data.

Fig. 3 also shows the CD spectra for two class I antigens,A2pap and B7pap, dissolved in phosphate buffer. The ampli-tude of the 217-nm band for B7pap agreed well with thatreported previously (28) and was somewhat higher than thatreported for a mixture of class I antigens (29). For bothproteins, the position of the negative band at 217 nm ischaracteristic of proteins in which ,1-sheet is the predominantsecondary structure, but the higher intensity and greaterbreadth of this band, as well as the position of the shorterwavelength positive band, indicate the presence of some

a-helix. Quantitative analysis bears this out (Table 1). ForA2pap, helix contents of 8-13% were obtained, depending onthe buffer used, whereas the 13-sheet content was estimatedat -75%. Comparison with the x-ray structure (5) indicatesthat the a-helix content was underestimated to some extent,whereas the 18-sheet content was greatly overestimated. Adifferent method of analysis (32) gave results in much betteragreement with the x-ray structure (,3-sheet content of-40%), but the fit to the CD spectrum was very poor.Comparison with the FTIR results showed that this discrep-ancy cannot be due to a genuine crystal-solution difference.The CD of B7pap gave 20% a-helix and 29-39% 13-sheet.

Contrary to the case of A2pap, the CD estimate of a-helix inB7pap was higher than that from FMIR, whereas that for the

Table 1. Secondary structures of human class I and class II major histocompatibility complex (MHC) antigens asmeasured by IR and CD spectra and x-ray methods

Secondary structure, %

Protein Buffer a-Helix ,-Sheet Turns Random Other* MethodstA2pap 17 41 28 14 IR

Phosphate 8 74 - 13 CDTris 13 77 10 CD

20 42 - X-ray (5)DR1p-p 10 53 24 13 IR

Phosphate 17 42 41 CDTris 23 38 - 39 CD

B7pap 8 48 32 12 IRPhosphate 20 29 - 51 CDTris 20 39 - 41 CD

,82m 6 52 33 9 IRPhosphate 0 59 41 CD

0 48 X-ray (5)IgG 3 64 28 5 IR

3 67 18 12 X-ray (21)Lysozyme 40 19 27 14 IR

45 19 23 13 X-ray (21)41 16 23 20 X-ray (22)

*Turns + random coil not calculated separately by CD.tCD spectra were analyzed by the method of Provencher and Glockner (22). CD data in the range of 240-190 nm were usedin the analyses for A2. and DR1pap; CD data in the range of 240-195 nm were used for B7pp and f82m. References forx-ray data are given in parentheses.

Immunology: Gorga et A

Dow

nloa

ded

by g

uest

on

Janu

ary

13, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989)

500Ca

U.

<0

-500

-1000

FIG. 3. C(A) in 5 mM

p-sheet waswas much I

Relativedisplayed asolved shoiindicative ogave an a-c38-42%. A

content andIn all thre

was observphosphate bcorrespond:(Table 1) a]cannot concDRipap- Neminor buffe

Analysis olamounts ofsuperfamilyeasy to assi

B7pap by C:that extendCD analysismall amotremain in 1

2000C

Do%

O. ..

=-0

0.-3--

of

FIG.4. C

sodium phos

D - recent advances in FTIR spectroscopy of proteins in aqueoussolution, it has become possible to assign peaks in the amide

o [ kI region to different forms of secondary structure and toquantitate the results (A.D., P. Huang, and W.S.C., unpub-lished data). Thus, it is possible to use both FTIR spectros-

o of / Ad \ copy and CD to determine the secondary structures ofproteins in solution.

Several factors limit the accuracy with which FT'IR and CDo s : can determine protein secondary structure. Both methods

assume that elements of secondary structure contributeindependently to the observed spectroscopic property. It is

0 - F also assumed that variations in size and distortions fromregularity are either intrinsically negligible or are made to beso by averaging over a protein basis set.

0 , Secondary structure analyses of the MHC antigens could180 200 220 240 260 also be affected by two unique features. Processed antigens

Wavelength (nm) retained in the antigen-binding site (6) can contribute to theobserved spectroscopic properties. It has recently been

D spectra of 182m (u), A2pap (.), B7pap (x), and DRlpap reported (33) that purified murine class II antigens appear tosodium phosphate (pH 7.75). have low molecular weight material associated with them. In

addition, the sites of papain cleavage of the DR1 a- andslower. However, the discrepancy for the p-sheet 3P-chains are not known. Since DRipap behaves like a soluble

less than in the case of A2pap. molecule and each chain is 4000-6000 daltons smaller thanto the class I proteins, the class II antigen DRipap the corresponding chain in the detergent-soluble form (9), iti roader minimum near 217 nm with an unre-broader minimum near 217 nm with an unre- is assumed that papain cleaves the DR1 chains somewhere in,ulder near 222 nm.(Fig . 3. ThIS pattern was* **-ldernear222nm (Fig.3).Thispatternwas the connecting peptide regions-that is, in the peptide seg-of increased a-helix content. Quantitative analysis ments connecting the a2 or P2 domains with the transmem-ielix content of 17-23% and a P-sheet content of etcontngheaor,2dmnsihtetase-ielicotenof17-3%nd p-hee coten of brane regions. Therefore, papain probably removes thes in the case of B7pap CD gave a higher a-helix cytoplasmic tail, the transmembrane segment, and part ofthe1alower p-sheet content than ETR._connecting peptide from each chain. Although papain cleavesZe of the MHC proteins studied, the CD spectrum the connecting peptide of HLA-A2 at the C terminus of the

'ed to be more intense in Tris buffer than in a3 domain, removing -9000 daltons from HLA-A2, it may)uffer, as shown for the caseof A2pap in Fig. 4. The cleave the connecting peptides of one or both of the DR1,ing differences in secondary structure contentingdifferences in secondary structure ct , ...chains closer to the transmembrane region. Thus, the re-re within the error limits of the method, so one maining segments of connecting peptide would contribute toclude that Tris Increases helix content In A2paor*-ueariipapr the secondary structure. If the additional connecting peptide-vertheless, the observed CD differences suggestvertheless, theobservedCDn differenes suggest in DR1 contributes to the contents of P-sheet, turns, or-r-induced conformational changes. random structure, then the proportion of a-helix in the

protein would be less than that in A2pap, and, conversely, ifDISCUSSION the connecting peptide contributes to the a-helix content, the

f the secondary structures of proteins with low proportion of a-helix would be greater.There are some specific limitations of FTIR spectroscopy

has in the past been difficult (30). Indeed it was (34). The very strong water band in the amide I regionm thas thepast beend t (30). Indeteditnwas requires very short pathlengths (6 um), highly sensitive_me that the small amount of a-helix detected in detectors, and extremely careful difference spectroscopy.D (28) was insignificant, especially with spectra Earlier studs have genrally used 2H20 solutions to avoidled to only 200 nm. Improvements in 'Methods of interference from the H20 band, thereby severely limiting theis now permit more accurate determination of H20 data available for reference. Second, a few amino acidants of a-helix. However, larger uncertainties residues (e.g., glutamine and asparagine) exhibit bands in thethe determination of p-sheet content. With the amide I region. However, such bands are too weak to

contribute very much to the overall spectrum, and theestimated contents of these residues are only slightly differ-ent for A2pap, B7pap, and DRipap. Thus the observed FTIRspectral differences between the MHC antigens are unlikelyto be due to differences in amounts of the amino acid residuesthat are IR active in the amide I region. Finally, there areproblems associated with resolving the various components,quantitation of the bands from second derivatives, andpossible variations in the intensities of amide I bands fromvarious types of secondary structure. The results for lyso-zyme, IgG, and A2pap, which are presented here, along withfurther details to be published elsewhere (unpublished data),of give us confidence that our results provide a satisfactorycharacterization ofthe unknown B7pap and DRipap secondary

___________,____.____,____.____,____.___ structures.180 200 220 240 260 There are also limitations of CD spectroscopy compared

Wavelength(nm) with FTIR for secondary structure determination. The re-solving power of CD is limited by the fact that the UV bands

'D spectra of A2pap in 5 mM Tris (pH 7.75) (n) and 5 mM of the a-helix, p-sheet, p-turns, and unordered structures-phate (pH 7.75) (.). strongly overlap. Contributions of chromophoric side chains

2324 Immunology: Gorga et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

13, 2

021

Proc. Natl. Acad. Sci. USA 86 (1989) 2325

pose another serious problem. Theoretical studies (ref. 35, ¶;M.C.M. and R.W.W., unpublished data) indicate that tyro-sine, tryptophan, and phenylalanine can make substantialcontributions to the far-UV CD of proteins. Experimentally,anomalous CD spectra for various proteins have been attrib-uted to aromatic (36, 37) or disulfide (38) contributions. Inaddition, in the present study, instrumental limits restrictmeasurements to wavelengths longer than about 190 nm.Hennessey and Johnson (39) have pointed out that CD datadown to -175 nm can greatly enhance the informationcontent and thus permit more reliable estimates of 8-sheetand even /-turn content.Both CD and FTIR show that all three MHC proteins

studied have small but significant amounts of a-helix. Thea-helix contents must be considered comparable for all threeproteins, in view of the differences between the CD and theFTIR results.The FTIR analysis also indicates that all three proteins

have comparable amounts of p-sheet. CD supports this in thecase of DRipap and B7pap but indicates a much higher 8-sheetcontent for A2pap. In view of the FTIR results and the strongsequence homology (2) between HLA-A2 and HLA-B7, theCD for A2pap must be considered anomalous.The anomaly in the CD spectrum of A2pap may be due in

large part to aromatic side-chain contributions unique to thisprotein. HLA-A2 has a tryptophan at position 107, which isreplaced by glycine in HLA-B7 and is also absent in DR1 (2).In addition, there are other points in the chain where anaromatic group in HLA-A2 is replaced by a nonchromophoricside chain in HLA-B7 (His-70 -- Gln, His-74 -* Asp, Phe-109-* Leu, His-114 -* Asp, His-145 - Arg, His-151 -* Arg) orby a different aromatic group (Phe-9 -+ Tyr, Tyr-113 -* His).In only one case does such a transition occur in the oppositedirection (Val-67 -- Tyr). If the entire difference in the CDspectrum were attributed to the replacement of a singletryptophan, that tryptophan would have [01220 106deg-cm2/dmol. Calculations of the interaction between atryptophan side chain and the two adjacent peptides¶ haveshown that such interactions can lead to [01220105deg-cm2/dmol. Avidin has a CD maximum at 228 nm, whichcorresponds to ca. +8 x 104 deg-cm2/dmol for each of thefive aromatic residues (one tyrosine, four tryptophans).Given that not all of the aromatic side chains contributeequally, that some may make negative contributions, and thatthe peptide backbone almost certainly makes a significantnegative contribution at this wavelength, individual tryp-tophan CD bands substantially in excess of 105 deg cm2/dmolare implied. Therefore, aromatic side-chain contributionsunique to A2pap, especially Trp-107, can account for a largepart of the anomaly.Although sequence homology and similarity in secondary

structure contents for HLA-A2 and HLA-B7 determined byFTIR spectroscopy argue for overall structural similarity, theFTIR spectrum indicates some differences in detailed struc-ture. The band at 1643 cm-1 in the second-derivative spec-trum for B7pap is much weaker than the corresponding bandin the A2pap or the DRipap spectra (Fig. 2). In addition, theB7pap spectrum has a much stronger band at 1676 cm-1 thando the A2pap and DRipap spectra. It has also been reported(40) that B7pap present as a contaminant of A2pap is excludedfrom A2pap crystals, indicating some differences in molecularshape.

It will be of interest to see whether other members of theimmunoglobulin superfamily, such as the T-cell receptor,CD4, CD8, and the class I-like molecules Qa and Tla, have

proportions of secondary structure similar to that of themajor histocompatibility antigens.We thank Don Wiley's laboratory for providing HLA-A2 and

HLA-B7, Todd Willis for performing the amino acid analyses, andRussell Middaugh (Univ. of Wyoming) for providing a copy of S. W.Provencher's CONTIN program. J.C.G. is a fellow of the Charles A.King Trust. This work was supported by U.S. Public Health ServiceGrants GM-22994 to R.W.W., HL-15980 to W.S.C., and AI-10736 toJ.L.S.

1. Kappes, D. & Strominger, J. L. (1988) Annu. Rev. Biochem. 57,991-1028.

2. Brown, J. H., Jardetsky, T., Saper, M. A., Samraoui, B., Bjork-man, P. J. & Wiley, D. C. (1988) Nature (London) 332, 845-850.

3. Rupp, F., Acha-Orbea, H., Hengartner, H., Zinkernagel, R. &Joho, R. (1985) Nature (London) 315, 425-427.

4. Marrack, P. & Kappler, J. (1986) Adv. Immunol. 38, 1-30.5. Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S.,

Strominger, J. L. & Wiley, D. C. (1987) Nature (London) 329, 506-512.

6. Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S.,Strominger, J. L. & Wiley, D. C. (1987) Nature (London) 329, 512-518.

7. Turner, M. J., Cresswell, P., Parham, P., Strominger, J. L., Mann,D. L. & Sanderson, A. R. (1975) J. Biol. Chem. 250, 4512-4519.

8. Parham, P., Alpert, B. N., Orr, H. T. & Strominger, J. L. (1977) J.Biol. Chem. 252, 7555-7567.

9. Gorga, J. C., Horejsf, V., Johnson, D. J., Raghupathy, R. &Strominger, J. L. (1987) J. Biol. Chem. 262, 16087-16094.

10. Gorga, J. C., Hazzard, J. H. & Caughey, W. S. (1985) Arch.Biochem. Biophys. 240, 734-746.

11. Savitsky, A. & Golay, M. J. E. (1964) Anal. Chem. 36, 1627-1639.12. Kauppinen, J. K., Moffat, D. J., Mantsch, H. H. & Cameron,

D. G. (1981) Appl. Spectrosc. 35, 271-276.13. Chen, G. C. & Yang, J. T. (1977) Anal. Lett. 10, 1195-1207.14. Krimm, S. & Bandekar, J. (1986) Adv. Protein Chem. 38, 181-364.15. Alvarez, J., Lee, D. C., Baldwin, S. A. & Chapman, D. (1987) J.

Biol. Chem. 262, 3502-3509.16. Kwiatkoski, J. M. & Reffner, J. A. (1987) Nature (London) 328,

837-838.17. Dong, A., Messerschmidt, R. G., Reffner, J. A. & Caughey, W. S.

(1988) Biochem. Biophys. Res. Commun. 156, 752-756.18. Susi, H. (1972) Methods Enzymol. 26, 455-472.19. Surewicz, W. K. & Mantsch, H. H. (1988) Biochim. Biophys. Acta

952, 115-130.20. Susi, H. & Byler, D. M. (1986) Methods Enzymol. 130, 290-311.21. Levitt, M. & Greer, J. (1977) J. Mol. Biol. 114, 181-293.22. Provencher, S. W. & Glockner, J. (1981) Biochemistry 20, 33-37.23. Amzel, L. M. & Poljak, R. J. (1979) Annu. Rev. Biochem. 48, 961-

997.24. Sears, D. W. & Beychok, S. (1973) in Physical Principles and

Techniques ofProtein Chemistry, ed. Leach, F. J. (Academic, NewYork), Part C, pp. 445-593.

25. Becker, J. W. & Reeke, G. N. (1985) Proc. Natl. Acad. Sci. USA82, 4225-4229.

26. Karlsson, F. A. (1974) Immunochemistry 11, 111-114.27. Isenman, D. E., Painter, R. H. & Dorrington, K. J. (1975) Proc.

Natl. Acad. Sci. USA 72, 548-552.28. Lancet, D., Parham, P. & Strominger, J. L. (1979) Proc. Natl.

Acad. Sci. USA 76, 3844-3848.29. TragArdh, L., Curman, B., Wiman, K., Rask, L. & Peterson, P. A.

(1979) Biochemistry 18, 2218-2226.30. Johnson, W. C., Jr. (1988) Annu. Rev. Biophys. Biophys. Chem. 17,

145-166.31. Yang, J. T., Wu, C.-S. C. & Martinez, H. M. (1986) Methods

Enzymol. 130, 208-269.32. Bolotina, I. A., Chekhov, V. O., Lugauskas, V. Yu. & Ptitsyn,

0. B. (1981) Mol. Biol. (Engl. Transl.) 14, 709-715.33. Buus, S., Sette, A., Colon, S. M. & Grey, H. M. (1988) Science 242,

1045-1047.34. Braiman, M. S. & Rothschild, K. J. (1988) Annu. Rev. Biophys.

Biophys. Chem. 17, 541-570.35. Woody, R. W. (1978) Biopolymers 17, 1451-1467.36. Green, N. M. & Melamed, M. D. (1966) Biochem. J. 100, 614-621.37. Day, L. A. (1973) Biochemistry 12, 5329-5339.38. Hider, R. C., Drake, A. F. & Tamiya, N. (1988) Biopolymers 27,

113-122.39. Hennessey, J. P., Jr., & Johnson, W. C., Jr. (1981) Biochemistry 20,

1085-1094.40. Bjorkman, P. J., Strominger, J. L. & Wiley, D. C. (1985) J. Mol.

Biol. 186, 205-210.

,Woody, R. W., Proceedings of the Second International Confer-ence on Circular Dichroism, ed. Kajtdr, M., Aug. 15-18, 1987,Budapest, pp. 38-56.

Immunology: Gorga et A

Dow

nloa

ded

by g

uest

on

Janu

ary

13, 2

021