THE USE OF AFFINITY CHROMATOGRAPHY IN THE STUDY OF PROTEIN FOLDING Christian B. Anfinsen x ‘ “4 Department of Biology SJ Johns Hopkins University 3 Wo Baltimore, Maryland qu Much of the current research in the field of synthetic vaccines is based on the assumption that fragments of antigens --for example, fragments oF viral coat proteins or toxins-— can, upon attachment to a nedtral carrier Structure and injec- tion into patients, induce the synthesis of antibodies that will recognize and immobilize the pathogenic agent. Anti- bodies recognize native conformations of protein determinants ("conformational determinants") and also primarily random peptide fragments (“sequential determinants") (1). It becomes important to determine the extent to which the latter, random structures can take on the "native format" that such fragments exhibit when part of the native protein structure. Only by such involvement in an equilibrium between the_ran and proper Ty-fekded-confGHhSLion WONG THE -TESGRELT oe as a meaningrul -aetgen tr-vaccine prd@ietiom— TT ar 3 we Showed that a fragment of hen's egg white lysozyme could be synthesized, attached to a carrier module, and be injected into animals with the produc- tion of antibody that would react with, and inactivate the original enzyme molecule. The synthetic material was, in a sense, a "vaccine" against lysozyme. As will be seen in Fig- ure 1, this antigenic lysozyme "loop" was stabilized in a suitable conformation by a disulfide bridge. Peptide frag— ments without such stabilization would tend to occupy a large set of random conformations, very few of which would be likely to be recognized by an antibody formed against the native, parent protein molecule. We have used affinity chromatography as a tool in the study of the interaction of peptides with antibodies (Anti- body.) raised against the native protein. By the attachment AFFINITY CHROMATOGRAPHY AND BIOLOGICAL RECOGNITION Gopyreg © 1983 bp Academic Press, Inc. 355 All rights of reproductfon in any form reserved. ISBN 0-12-166580-1 356 CHRISTIAN B. ANFINSEN ET AL. s8z o8z NO 082 oB: Cleavage from resin ond removal of blocking groups by anhydrous HF DEBLOCKED PEPTIDE Reduction by mercaptoethanol and reoxidation in air SYNTHETIC LOOP ” R f FIGURE 1. The solid phase synthesis, deprotection and ¢ ring closure of a peptide "loop" in hen's egg white lysozyme. \) of fragments of the protein chain to a Sepharose supporting Q backbone, subfractions of the mixed Antibody, population can bu be isolated. The isolation of such monospecific antibodies Y provides reagents that are no longer precipitating. Thus, one x can deal quantitatively with antigen-antibody reactions in % solution, and measurements are made more easily and more accu- rately. The binding to such columns of monospecific antibod- ies already suggests that fragments of the protein molecule, which appear to be totally lacking in structure as determined by the usual physical measurements, do indeed occupy ordered structure for at least a small portion of their lifetime. Thus, a peptide appears to "flicker" between the random form and the form that the peptide occupies when it is part of the D Fp myph 9 "mmm stonete biter” “f° PROTEIN FOLDING 357 FIGURE 2. The amino acid sequence of staphylococcal nuclease, FIGURE 3. Drawing of the three-dimensional structure of staphylococcal nuclease. - * SD CHRISTIAN B. ANFINSEN ET AL. F fegecenb., 67 99106 122: «133-149 L ! 000] a 6 438 49 149 L | L- 1@L044) [ FIGURE 4. Some fragments prepared from staphylococcal nuclease. Top: location of pleated sheet and helical sec- tions of nuclease. Fragments labelled 6-48 and 49-149 are produced by limited trypsin digestion. Peptide 1-126 (and the fragment 127-149, not shown) are obtained by trypsin digestion of trifluoroacetylated nuclease with subsequent removal of the TFA protecting groups. Peptide 99-149 is one of five fragments produced by cyanogen bromide digestion of the protein. native protein structure. Guided by its amino acid sequence, it can occasionally assume its conformation in the intact parent molecule. Much of this work has been done with the convenient pro- tein, staphylococcal nuclease (Fig. 2 and 3), which is devoid of SS bridges and undergoes very rapid denaturation and rena- turation. It may be cleaved into a variety of peptide frag- ments by enzymic or chemical methods (Fig. 4). The purified {~ nuclease was fractionated by sequential immunoabsorption and elution using columns of Sepharose to which one or another of the peptide fragments had been attached. The preparation of me non-precipitating, but inactivating antibody, is summar- ized in Figure 5 (3). Anti-(1-149) _, i.e., anti-native uclease, was first isolated on a sNpharose-nuclease column. his mixed antibody population was then successively fraction- ated on Sepharose-(99-149) and Sepharose-(127-149). The "pass through" on the Tas contained antibody directed against he sequence (99-126). This fraction was monodeterminant, and id not precipitate with nuclease, although it still combined ith the 99-126 portion of native nuclease. When anti-(99-126). was added to nuclease, the enzyme was inactivated (3,4). However, addition of increasing amounts PROTEIN FOLDING 359 ANTI-NUCLEASE ANTIBODIES pa , « © 2.0, Anti- Nuclease on Sepharose-(i-149) ‘vn ‘ , 7 [A } { [ é enh, . fa-iap/ noe Minne Sepharose -(99-149) Anti- (99-149), 4 pH, 727 22 2 Oo 10 20 30 40 50 FIGURE 5. Fractionation of antibody against native nuclease by sequential absorption on Sepharose columns to which either nuclease, peptide 99-149 or peptide 127-149 have been attached. of the randomly structured peptide 99-149 led to increasing release of the antibody-bound enzyme. Determinations of the successive activity levels permitted calculation of the equi- Librium constant relating random structure, and structure sufficiently similar to the native "format" of the peptide to permit competition with nuclease for antibody (99-126)_. These calculations indicated that (99-149), (i.e., random )occupied an antibody-compatible conformation, approx— inatel 0-022 08 the time. Other peptide fragments have been subjected-to similar analysis and show similar propensities for assuming their native format. For example, peptide (127- 149) and its monospecific antibody, were employed in an analy- sis similar to the above study of competition between intact nuclease and peptide fragment, and yielded the same estimate of folding frequency. Using similar methods, one can determine the frequency of unfolding of a native structure such as nuclease (5). How 360 CHRISTIAN B. ANFINSEN ET AL. COMPETITION ASSAY R I ; + Rabbit Anti-Goat Immunoglobulin 140 _ (99-149) Ab. +t 4¢-(99-149) - supernatant: 140_ (99-149), Ne? Ny 1 a . 4 precipitate: Aba» Abe: c-(99-149), Abp N FIGURE 6. Double antibody procedure for establishing the degree to which native nuclease can unfold into a family of denatured forms. often is a protein molecule sufficiently unwound to be recog- nized by an antibody directed against a random polypeptide structure? For these studies, antibody was prepared against the peptide (99-149). This antibody, (antizso=In97-> was irst purified from the IgG fraction by absorption Bn a column of Sepharose~ (99-149). The purified antibody was then reacted with ~'C-carbamoylated (99-149). It was found that nuclease could compete for 1s in ion to an extent consistent with a “conformation of 3.x 10° at 25°C. This indicates par- tial unfolding of nuclease. As indicated in Figure 6, the extent of competition could be measured by using rabbit anti- goat immunoglobulin to determine the fraction of the labelled peptide in the precipitate and supernatant components. When the experiments were carried out at a°C and 30_C, the confor- mational equilibrium constants were 4000 and_4Q0, respectively. Addition of stabilizing ligands for nuclease (thymidine di- phosphate and calcium tong) led to a constant of 40,000. The tendency for certain peptide fragments OL protein molecules to occupy their native "format" in solution may ecome a strong factor in the preparation of totally synthetic vaccines. Kren the total sequence of a protein associated with a pathogenic agent is known, organic synthesis of a rela- tively Small antigenic portion of the molecule may be adequate for the préparation of a useful vaccine. Such an approach has been taken “by Arnon and Sela-and-tneir colleagues (6) at the Weizmann Institute of Science and, more recently, by Lerner and the synthetic vaccine group at the Scripps Clinic of the University of California in San Diego (7). Att hog, aman Cbn Si Mberce ha wet baa Si PROTEIN FOLDING 361 REFERENCES 1. Sela, M., Schechter, A. Schechter, I., and Borek, F. Cold med Harbor Symp. Quant. Biol. 32, 537 (1967). 2. Arnon, R., Maron, E., Sela, M., and Anfinsen, Cc. B. Proc, Natl. Acad. Sci. U.S.A. 68, 1450 (1971). 3. Sachs, D. H., Schechter, A. N., Eastlake, A., and Anfin- sen, C. B., J. Immunol. 109, 1300 (1972). 4, Sachs, D. H., Schechter, A., Eastlake, A., and Anfinsen, C. B., Proc. Natl. Acad. Sci. U.S.A. 69, 3790 (1972). >. Furie, B., Schechter, A. N., Sachs, D. H., and Anfinsen, C. B., J. Mol. Biol. 92, 497 (1975). 6. Arnon, R., Sela, M., Parant, M., and Chedid, L., Proc. Natl. Acad. Sci. U.S.A, 77, 6769 (1980). 7. This volume. DN wet 6, “ Hoey... tl sagt bk alas, MY] te te © ed oy fog wile ta, ow D iis - Men OW in ant