Reprinted from the PIIOCC~DINOE or THE NATIONS ACADWBT or SCUNC*O Vol. 61. No. 2, pp. 836643. Ootobes, 19% SELECTIVE ENZYME PURIFICATION BY AFFINITY CHROMATOGRAPHY BY PEDRO CUATRECASAS, MEIR WILCHEK,* AND CHRISTIAN B. ANFINSEN IABORATORY OF CHEMICAL BIOLOOY, NATIONAL INSTITUTE OF ARTHRITIS AND MRTABOLlC DISXXSES, NATIONAL INSTITUTES OF HEALTH, BEmSDA, MARYLAND Commu&u.kd Aup& 9,1968 The purification of proteins by conventional procedures is frequently laborious and incomplete, and the yields are often low. Enzyme isolation based on a highly specific biological property-strong reversible association with specific substrates or inhibitors-has received only limited attention.`-' In affinity chromatography, the enzyme to be purified is passed through a column containing a cross-linked polymer or gel to which a specific competitive inhibitor of the enzyme has been covalently attached. All proteins without sub- stantial afIinity for the bound inhibitor will pass directly t,hrough the column, whereas one that recognizes the inhibitor will be retarded in proportion to its affinity constant. Elution of the bound enzyme is readily achieved by changing such parameters as salt concentration or pH, or by addition of a competitive inhibitor in solution. The successful application of the method requires tl&t the adsorbent have a number of favorable characteristics. Thus, the unsubstituted matrix or gel should show minimal interaction with proteins in general, both before and after coupling to the specific binding group. It must form a loose, porous network that permits easy entry and exit of macromolecules and which retains favorable flow properties during use. The chemical structure of the supporting material must permit the convenient and extensive attachment of the specific ligand under relatively mild conditions, and through chemical bonds that are stable to the conditions of adsorption and elution. Finally, the inhibitor groups critical in the interaction must be sufficiently distant from the solid matrix to minimize steric interference with the binding processes. In this report the general principles and potential application of affinity chro- matography are illustrated by results of its application to the purification of staphy- lococcal nucleaee, cr-chymotrypsin, and carboxypeptidase A. The solid matrix used in these studies wassepharose (a "beaded" form of the cross-linked dextran of highly porous structure, agaroses) which displays virtually all the desirable fea- tures listed above. Activation of the Sepharose by treatment with cyanogen bromidd* I results in a derivative that can be readily coupled to unprotonated amino groups of an inhibitory analog. The resultant Sepharose-inhibitor gel is a highly stable structure which has nearly ideal properties for selective column chromatography. Ezpesimenti Procedure.-M&tic&: Sepharose 4B wm obtained from Pharmacia, cyanogen bromide from Eastman, pdTp and benzoyl-Ltyrosine ethyl ester from Cal- biochem. Staphylococcal nuclease (Foggi strain) was obtained by modification* of techniques described by Fuchs et al.# The following purified enzymea were purchased from Worth- ington: a-chymotrypsin (CDS 7LC); a-chymotrypsin, DFP-treated (CD-DIP 204); chymotrypsinogen A (CCC SCC) ; subtilisin VIII (MB 3080) ; trypsin, DFP-treated Reprinted by the U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE National Institutes of Health VOL.& 1968 BIOCHEMISTRY: CUATRECASAS ET AL,. 637 (TDIP 7HA); pancreatic ribo- nuclease A (RAF 8BA) ; CM- boxypeptidase A (COA DFP 8FB) ; carboxypeptidase B (COB DFP 7GA). Beef pancreas ace- tone powder (26B-8890) was ob- tained from Sigma. 3'-(4Amino-phenylphosphoryl)- 1 1 3 fl deoxythymidine - 5' - phosphate (Fig. 1) was synthesized from .--i o'po+NH2 pdTp.`O The following com- I\" pounds were prepared by classical 0- 5 methods of peptide synthesis: FIQ. l.-Structure of nuclease inhibitor used for at- L-tyrosyl-D-tryptophan, D-trvp- tophan methyl ester, tam&o tschment to 8epharose. caproyl-D-tryptophan methyl ester, and fl-phenylpropionamide. Preparation of substituted Sepharoses: Cyanogen bromide activation of Sepharose ~8s based on procedures previously described.6. 7 Yepharose (decanted) is mixed with an equal volume of water, and cyanogen bromide (100 mg per ml of settled Sepharose) is added in an equH1 volume of water. The pH is immediately adjusted to, and maintained at, 11 by titration with 4 K NaOH. When the reaction has ended (about 8 min), the Sepharose is washed with about 20 vol of cold 0.1 M NaHCOa on a Buchner funnel under suction (about 2 min). The washed Sepharoee is suspended in cold 0.1 dl XaHCOz, pH 9.0, or another appropriate buffer, in a volume equal to that of the original Sepharose, and the inhibitor is quickly added in a soWon representing 5-15% of the final volume. This mixture is stirred gently at 4* for 24 hr, after which it is washed extensively with water and buffer. The quantity of inhibit,or coupled to the Pepharose can be controlled by the amount of inhibitor added to the act,ivated Sepharose, or by adding very large amounts of inhibitor to yield maximal coupling, followed by dilution of the final Sepharose-inhibitor with un- substituted Sepharose. Furthermore, to increase the amount of inhibitor coupled to the Sepharose, it is possible to repeat the activation and coupling procedures on the already substituted material, provided the inhibitor is stable at pH 11 (for 10 min). In the cases reported here the amount of inhibitor that was coupled was easily estimated by calculating the amount of inhibitor (spectroscopically measured) which was not recovered in the final washings. .Uternatively, acid hydrolysis of the Sepharose, followed by amino acid analysis, could be used for quantitetion in those Cases where amino acids or peptides are coupled to the Sepharose. The operational capacity ofm&omv 9 deter- mined by two methods. In the irst, an amount of enzyme in excess of the theoretical capacity was added to a sma.11 column (about 1 ml). After washing this column with buEer until nep;ligible quantities of protein emerged in the effluent, the enzyme was rapidly removed (i.e., by acetic acid washing) and its amount determined. In the second method, small aliquots of pure protein were added successively to t,he column until significant protein or enzymatic activity emerged; the total amount added, or that which was subsequently eluted, was conjlderetl the operational capacity. Results.--Siaphylococcal nucleasc: This extracellular enzyme of Staphylococcus aure~s, which is capable of hydrolyzing DNA and RNA, is inhibited, competi- tively, by pdTp (see ref. 11 for recent review). Thymidine 3',5'-di-p-nitro- phenylphosphate is a substrate for this enzyme, which rapidly releases p-nitro- phenylphosphate from the 5'-position and, slowly, p-nitrophenol from the 3'- position.12 The presence of a free 5'-phosphate group, however, endows various synthet.ic derivatives with strong inhibitory properties.`* 3'-(4-Amino-phenyl- 638 BIOCHEMISTRY: CUATRECASAS ET AL. Paoc. N. A. 8. phosphoryl)-deoxythymidine-5'-phosphate10 (Fig. 1) was selected as an ideal derivative for coupling to Sepharose for aflinity chromatography, since it has strong affinity for nuclease (K,, 10d M), its 3'-phosphodiester bond is not cleaved by the enzyme, it is stable at pH values of 5-10, the pK of the ammo group is low, and the ammo group is relatively distant from the basic structural unit (pTp-X) recognized by the enzymatic binding site.12 This inhibitor could be coupled to Sepharose with high efficiency, and the re- suiting inhibitor-Sepharose shows a high capacity for nuclease (Table 1). Col- umns containing this substituted Sepharose completely and strongly adsorbed samples of pure or partially purified nucleate (Fig. 2). If the amount of nucleate applied to such columns does not exceed half of the operational capacity of the Sepharose, virtually no enzyme activity escapes in the effluent, even after washing with a quantity of buffer more than 50 times the bed volume of the column. Elution of nuclease could be effected by washing with buffers having a pH inade quate for binding (less than 6). The yields of protein and activity were invari- ably greater than 90 per cent. Acetic acid (pH 3) was a convenient eluant since, with this solvent, the protein emerged sharply (in a few tubes) and the material could be directly lyophilized. The purity of the nuclease obtained in these studies was confirmed by its specific activity, immunodiffusion,g and disc gel electrophoresis.g Such columns can be used for rapid and effective large-scale purification of nuclease. For example, 110 mg of pure nuclease could be obtained from a crude nuclease concentrate (the same sample used in Fig. 2) with a 20-ml Sepharose column in an experiment completed in 1.5 hours. When the total concentration of protein in the sample exceeded 20-30 mg per ml, small amounts of nuclease appeared in the first peak of protein impurities, especially if very fast flow rates were used (400 ml/hr). However, with such flow rates, nuclease could still be completely extracted if more dilute samples were applied. The columns used in these experiments could be used repeatedly, and over protracted periods, without detectable loss of effectiveness. Nuclease treated with cyanogen bromide, which is enzymatically inactive, was not adsorbed or retarded by the nucleasespecific Sepharose columns. Spleen phosphodiesterase, which, like staphylococcal nuclease, hydrolyzes DNA and RNA to yield 3'-phnsphoryl derivatives but which is not inhibited by the 5'- phosphoryl nucleotides, passed unretarded through these columns. A dramatic illustration of the value of these techniques in enzyme purification TABLE 1. .EJicimcy of coupling of 3'-(4-amino-phenyIphosphar,y~)-deorl/thymidine-6'- phosphate (Fig. 1) to Sephurose, and capacity o$ the resulting adsorbent for staphylococcd ndease. ~Molee of Inhibitor/ml Sephaross Mg of Nucleaee/ml Sepharose Expt. Added Coupled* Theoreticdt Found A 4.1 2.3 44 . : 2.5 1.5 28 8 1.5 1.0 19 D 0.5 0.3 6 ; 1 .j o Determined by the procedurea described in the text. t Awuming equimolar binding. VOL. 61, 1!%8 BIOCHEMISTRY: CUATRECASAS ET AL. 639 Fro. 2.-Purification of staphy- lococcml nucleaee by aflinity ad- sorption chromatography on a nu- cleaeetlpecific Sepharoee column (0.8 X 5 cm)(aample B, Table 1). The column wan equilibrated with 0.05 M borate buffer, pH 8.0, con- taining 0.01 M CaCl:. Approxi- mately 40 mg of partially purified material (containing about 8 mg of nucleaee) WBB applied in 3.2 ml of the same buffer. After 50 ml of buffer had passed through the column, 0.1 M acetic acid was added to elute nucleaee. DNaae activity i ex- pressed ae the change in abeorbancy at 260 ~QI caused by 1 J of sample.`" 8.2 mg of pure nucleate and. all of the original activity was recovered. The flow rate WBB about 70 ml per hour. WM afforded by the one-step purification obtained by passing a crude culture of Staphylococms aureus through such a nuclease-specific Sepharose column after removal of cells by centrifugation." After adjustment of 500 ml to pH 8, addi- tion of 50 ml 1 M CnClz was then added to ensure a calcium ion concentration consistent with complex formation. The medium, containing about 6 pg of nuclease per ml, was passed through a l-ml column wit.h nearly complete adsorp- t.i& of nutllease activity, lvhich was subsequently eluted with acetic acid. Affinity columns can also be of use in the separation of active and inactive nuclease derivatives, as from samples aikylated with iodoacetic acid's or subjected t,o proteolvtic digestion.1" The specific Sepharose adsorbent can also be used effectively as an insoluble inhibitor to st,op enzymatic reactions (Fig. 3). a-Chymotrypsin: :I number of proteolytic enzymes, of which a-chymotrypsin and carboxypeptidase -1 are examples, are capable of binding, but not hydrolyz- ing, significantly, the &ntiomeric substrate analog. Therefore, the techniques described here should be applicable to a large number of enzymes of this class. D-tryptophan methyl emled-w~in experi- ments with a-chymotrypsin (Fig. 4). When this inhibitor is coupled directly to Sepharose, incomplete and unsat,;sfactory resolution of the enzyme results (Fig. 4B). However. dramatically stronger adsorption of enzyme occurs if a 6-carbon chain (t-amino caproic acid) is interposed between the Sephnrose matrix and the inhibitor (Fig. 4C). This illustrates the marked steric interference that results when the inhibitor is attached too closely to the supporting gel. The importance of specific afinity for the enzyme binding site is illustrated by the absence of adsorption of DFP-treated cr-chymotrypsin (E'ig. 40). Impurities in commercial ar-rhymotrypsin constituting 4-12 per cent, could be detected in different lots by thcsc techniques. Grentcr than 90 per cent of the activity and protein added to these affinity columns could be recovered, and the columns could be used re- peatedly without detectable loss of effectiveness. It is notable that significant retention was obtained with an atfinity adsorbent 640 BIOCHEMISTRY: CUATRECASAS ET AL. PFIOC. N. A. 9. Fm. S.-Stopping of nucleate- catalyzed reaction by addition of Sepharose inhibitor (sample A, Table 1). Ten pg of nuclease was added to each of three samples containing 1.5 ml of 0.05 M Tris buffer, pH 8.8, 0.01 M CaClt, and 0.1 mM synthetic substrate, thymichne 3'-5'-di-pnitrophenyl- phosphate.Q I* The change in absorbancy at 330 m&, represent- ing release of pnitrophenylphos-, phat.e, was recorded continuously in cuvette A. At 4.6 min, 0.5 ml of untreated Sepharoee or of Sepharose coupled with inhibit)or were added to B and C, respec- tively. After a Zmin centrifuga- tion to remove the Sepharose, the changes in absorbancy were recorded (8 min). The difference in act,ivity between A and B is due to dilution. that contained a relatively weak inhibitor, X-tnmino caproy!-D-trypt'ophnn methyl ester. N-acyl-D-t'ryptophan esters have k', values of about Cl.1 mM,17 some 100 times greater than that of the inhibitor used for purifiication of staphyl- ococcal nuclease. These results suggest that unusually strong affinity const'ants will not be an essential requirement for utilization of these techniques. The relatively unfavorable affinity constant of a-c.hymotrypsin for the D- tryptophan methyl ester derivatives was compensated for by coupling a very large amount of inhibitor to the Sephurose. About 6.5 ~.~moies of t,he compound were present per milliliter of Scpharose during the coupling procedure. and t,here was an uptake of 10 gmoles per ml, with a resuhiznt eft'cctive concetttr:Ltion of inhibitor. in the column, of about 10 mill. Such a high degree of substitution ~:unplr <2.5 mg) was applied in 0.5 ml of the eamc huger. One- nuliilit IX fractions were collected, performed at room ttmpersture. &2hvmotrypsin w,w &ted with 0.1 211 acetic acid, pit 3.0 (UTTOWS). Peaks preceding the arrowa in devoid of enzyme VOL. 61, 1968 BIOCHEMISTRY: CUATRECASAS ET AL. 641 occurred despite the relatively high pK of the amino group of the e-amino caproyl derivative. The coupling was done in 0.1 M NaHCOs buffer, pH 9.0; higherpH values could not be safely used because of the probability of hydrolysis of the ester bond. The highly substituted Sepharose derivatives retained very good flow properties. Figure 5 illustrates the effect of pH and ionic strength on the chromatographic patterns. Although stronger binding appears to occur with buffer of lower ionic strength (0.01 M), this should not be used since some proteins, such as pancreatic ribonuclease, will adsorb nonspecifically to unsubstituted or inhibitor-coupled Sepharose under those conditions. Figure 5 shows the patterns obtained with a number of other enzymes, emphasizing the specificity of the a- chymotrypsin-specific Sepharose columns. Very small chymotryptic impurities X0 - 005 Y TRIS. vH 8.0 2.0 - i 8 2.0 005 Y TRIS. In 14 P- Y s 5 1.0 2 IO 005YTRIS,pH68 1 t 001 w TRIS. OH 6 a :a< : ! l.OF 1 L 2 6 IO i4 a tz.15 M FIG. 5.-Effects of pH and ionic strength on affinity ad- sorption of a-chyrnotrypsin on a column (0.5 X 5 cm) of Seph- arose coupled wrth *amino caproyl-D-tryptophan methyl ester. X sample containing 2.5 mg of a-chymotr-ypsin in 0.j ml of buffer was applied to the column. Elution of a-chymotryp- sin WRB performed with 0.1 Jf acetic acid, pH 3.0 currow). Other conditions were as in Fig. 2. The first peak (tubea 24) was devoid of chymotrypsin activity," and the specific ac- tivity of the subsequently eluted protein was constant. GEN Imel I SUBTILISIN j0,05 FIG. 6.-Cbromatographic patterns obtained by passing several enzyme preparations through a column (0.5 X 5 cm) of Sepharose coupled with tamino caproyl-D- tryptophan methyl ester. The cohmm was equilibrated nnd run with 0.05 Jf T&Cl buffer, pH S.O. Approxi- mately 3 mg of protein (~1 through E), dissolved in 0.5 ml of the aame buffer, was applied to each column. Chy- motrypsin wss eluted with 0.1 Jf acetic acid (arrow). The sample used in F consisted of 1 ml of a supernatant (280 n+ absorbsncy, about 15) obt,ained by dissolving about 100 mg of an acetone powder of bovine pancrease in 3 nd of Tris-Cl, pH 8.0, followed by centrifugation for 20 min at 4000 X g. Enzyme activity is expressed as the change in absorbancy (at 256 mr) per minute per 5 4 of sample, with benzoyl-ltyrosine ethyl ester as substrate." Other conditions were m m Figs. 4 and 5. 642 BIOCHEMISTRY: CUATRECASAS ET AL. PROC. N. A. s. can be readily removed from other enzymes by these techniques. This may be a useful way of removing chymotryptic impurities from other proteases used for structural studies, where even small traces can lead to unexpected cleavages. It is of interest that chymotrypsinogen A is very slightly but significantly re- tarded by t*he chymotrypsin-specific column (Figs. 6.4 and B), suggesting that this precursor is capable of weakly recognizing this substrate analog. A proteo- lytic enzyme of broad substrate specificity, subtilisin, is not adsorbed (Fig. SD). A small amount of chymotryptic-like protein could be readily separated from a crude pancreatic digest (Fig. GF). This material contained all of the chymo- tryptic activityI of the crude digest, but the exact nafure of this material or the reason for the low specific act,ivity were not. determined. Unlike the results obtained with the nuclease-specific Sepharose system (Fig. 3), the chymotrypsin-specific Sepharose derivative ~3s relatively ineffective in stopping the hydrolysis of benzoyl-L-tyrosine et,hyl ester when added to a reaction mixture. Attempts to elute cu-rhymotrypsin from a coiumn, like that shown in Figure 4, with a 0.01 Jf solution of the above substrate were unsuccessful. However, elution did occur with a 0.018 &f solution of $phenylpropionamide, an inhibitor with a K, of 7 mXI.*g If the Sepharose column Fas equilibrated and 0.6 CARBOXYPEPTIDASE 6 I t-Tymsyl-0~TrypMpt.m 2 4 6 8 IO 12 14 16 EFFLUENT, YL FIQ. T.-Affinity chromatography of carboxypeptidase A on a column (0.5 x 6 cm) of Sepharoee coupled with L- tyrosine-D-tryptophan. The buff& used was 0.05 M T&Cl, pH 8.0, con- taining 0.3 N NaCL About 1 mg of pure cnrboxypeptidm A (11, B) and 1.8 mg of carboxypeptkk B (C), in 1 ml of t,he same buffer, were applied to the columns. Elution was accomplished with 0.1 M acetic acid (arrow). developed with 0.005S Jf &phenylpropion- amide solution, a-chymotrypsin was only moderately rebarded. These results indi- cate, again, t.hat the processes involved in t.he sepamtions are clearly related to the functional affinity of the enzymatic binding site for specific structural substances. Carbozypeptidasc -4 : .\ specific adsorbent for this enzyme was prepared by coupling the dipeptide, L-tvru~it!~-D-i.ryptophaIl, to Sepharose. In ihe ftoupling ijrocedure, about 60 pmoles oi rile rlipeptide irlhibitor were added per milliliter of Yephnrose, and approximntely iy grntih ti were coupied. Figure Y illustmtes that this enzyme was &on&y &orbed by a column containing such ;i substituted Sepharose. The yields obt:lined upon elution were again quantit:ltive. Discussion.-In recent years there has been considerable inIeresi, in the covalent attachment, of bioiogically active compounds (i.e., euzymc5, :~rl~ii~,dirs;, :~nd ;trliigens) to insoluble polymers.-" T!lese miteri& es- pecially the Jcrivstives of ctllulose, have found use in the puriticntion of nntibodies,2' VOL. 61, 1968 BIOCHEMISTRY: CUATRECASAS ET AL. 643 nucleotides,** complementary strands of nucleic acids,*' certain species of trans- fer RNA,*' and enzymes.1-3 The principles and procedures, as illustrated and outlined in this commun.ica- tion, should be of value in the purification and isolation of a great many bio- logically active proteins or polypeptides which can reversibly and specifically bind to small molecules. If the latter are not chemically altered during the reversible adsorption process (e.g., an inhibitor), and if an amino group can be introduced in a region of its structure in such a way that binding to the macro- molecule is unaffected, the procedures outlined here should be directly applicable. Summary.-Principles and techniques for selective enzyme purification by afhnity adsorption to inhibitor-Sepharose columns are presented and illustrated by experiments performed on staphylococcal nuclease, a-chymotrypsin, and carboxypeptidase A. Inhibitory substrate analogs linked to Sepharose provide adsorbents on which enzymes can be purified rapidly and completely in a single step. The authors acknowledge the collaborative assistance of Dr. Sheldon Schlaff in many of these studies, and we are grateful for hi efforts in the further extension of these meth- ods.15 We wish to express our gratitude to Mrs. Juanita Eldridge for technical assistance. * Fellow in the Visiting Program of the U.S. Public Health Service (1967-1968). On leave of absence from the Weizmann Institute of Science, Rehovoth, Israel. 1 Lerman, L. S., these PROCEEDINM, 39, 232 (1953). * Arsenis, C., and D. 3. McCormick, J. Biol. Chem., 239, 3093 (1964). 3 Ibid., 241, 330 (1966). 4 McCormick, D. B., And. Bit&em., 13, 194 (1965). 6 Hjerten, S., Biochim. Biophys. A&, 79, 393 (1964). 6 A.&n, R., J. Porath, andS. Ernback, Nature, 214, 1302 (1967). T Porath, J., R. Ax&, and S. Ernback, Ndure, 215, 1491 11967). * Moravek, I,., C. B. Anfinsen, J. Cone, H. Taniuchi, manuscript in preparation. p Fuchs, S., P. Cuatrecasss, and C. B. Anfinsen, J. Bid. Own., 242, 4768 (1967). IO Wilchek, M., P. Cuatrecasas, and C. B. An&en, unpublished. `I Cuatrecanas, P., H. Tnniuchi, and C. B. Aniinsen, in Brookhauen Symposia in Biology, in press. I* Cuatrecsass, P., M. Wichek, and C. B. Anfinsen, manuscript in preparation. Ii Cuatrecasss, P., S. Fuchs, and C. B. Anfuisen, .I. Btil. Ckm., 242,X41 (1967). 1' In collaboration with Dr. G. Omenn. I6 Schlaff, S., unpublished. ** Tan&hi, H., C. B. Anfinsen, and A. Sodja, these PROCEEDINGS, 58, 1235 (1967), and un- published data. I7 Huang, H. T., and C. Niemann, J. Am. Chem. Sot., 73,3228 (1951). I5 Hummel, B. C. W., Can. J. B&hem. Physid., 37, 1393 (1959). I@ Foster, R. J., and C. Niemann, J. Am. Chem. Sot., 77, 3365 (1955). m Silman, I., and E. Kstchaleki, Ann. Rev. B&hem., 35,873 (1966). *I Moudgal, N. R., and R. R. Porter, Biochim. Biophys. A&z, 71, 185 (1963). " Sander, E. G., D. B. McCormick, and L. D. Wright, J. Chtomcrtog., 21,419 (1966). z* Bautz, E. K. F., and B. D. Holt, these PROCEEDINOB, 48,400 (1962). 2` Erhan, S., L. G. Northrup, and F. R. Leach, these PROCEEDINQB, 53,646 (1965).