Affinity chromatography * J. TURKOVA CONTENTS Principles of affinity chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Choiceofboundaffinant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General aspects of the affinant- sorbent bond .......................................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89 92 94 96
PRINCIPLES OF AFFINITY CHROMATOGRAPHY Affinity chromatography is a technique for the isolation of biologically active substances, making use of their exceptional property of selective and reversible binding of other substances, for which Reiner and Walch introduced the term “affinant”. An affinant can therefore be considered as a special case of a ligand. If an insoluble affinant is prepared, usually by covalently attaching to a solid support, and if the extract containing the biologically active component to be purified is passed through this column containing the above material, then all substances which possess no affinity for the given affinant pass directly through the column with the eluate. Substances that have an affinity for the bound affinant are retarded in proportion to their affinities under the experimental conditions used. The specifically adsorbed substances can then be eluted either by a soluble affinant or by changing the solvent composition in such a way as to cause the dissociation of the complex of the isolated substance with the bound affinant. Most commonly, changes in pH, ionic strength or temperature or the addition of reagents that cause the dissociation of the complex are used. As an example of affinity chromatography, the isolation of active chymotrypsin and trypsin from a pancreatic extract is shown in Fig. 7.1. The isolation was carried out on a column of 6%agarose with bound soya bean trypsin inhibitor (Porath). In alkaline medium, chymotrypsin and trypsin form a complex with the bound trypsin inhibitor, thus separating from the inactive material which passes through the column unretained. By making use of a pH gradient, it is possible to separate the complex of enzymes with the insoluble inhibitor. At pH 4.5, chymotrypsin is eluted first, while for the dissociation of the trypsin complex a lower pH should be used. For the release of chymotrypsin and trypsin from the complex with the insoluble inhibitor, instead of a change in pH, elution with specific inhibitors at a constant pH can also be employed, as *Obviously the term “affinity chromatography” is incorrect, as chromatographic separations in general are based o n differences in affinities. Therefore, the more recently introduced name for this technique, bioaffinity chromatography, seems t o be more suitable. However, as this term is not very common yet, “affinity chromatography” has been used throughout this book.
References p. 96
AFFINITY CHROMATOGRAPHY Inactive material
Fig. 7.1. Affinity chromatography of a pancreatic extract on a column of 6% agarose with bound soya bean trypsin inhibitor (Porath). Volume of column, 30 ml. The chromatogram was developed with a solvent system with decreasing pH, as indicated.
lnaciive material Chyrnotrypsin Trypsin pH 1.0
Fig. 7.2. Affinity chromatography of a pancreatic extract on a column of 6% agarose with bound soya bean trypsin inhibitor (Porath). Stepwise elution was accomplished with specific inhibitors.
PRINCIPLES OF AFFINITY CHROMATOGRAPHY
shown in Fig. 7.2 (Porath). A solution of tryptamine, which is a specific inhibitor for chymotrypsin, releases only chymotrypsin from the column of the insoluble trypsin inhibitor, while trypsin can be liberated by using a solution of benzamidine. The principle of affinity chromatography is often made use of in a very simplified manner when the biologically active substance to be isolated is sorbed on to an insoluble affinant, from which it can be desorbed in a single peak by a convenient procedure. As an example, the separation of active pancreatic trypsin inhibitor from inactive ballast on a column of hydroxyalkyl methacrylate gel with bound chymotrypsin is shown in Fig. 7.3 (Turkovi et al.). If the affinity of the isolated substance for the bound affinant is high, the batch process can be employed with advantage. The isolation procedure can then be considered to be a precipitation rather than chromatography. A 2 81.0 0[
0.6 0.4 ;I
Fig. 7.3. Affinity sorption of lung trypsin inhibitor on a column (10 X 1 cm) of hydroxyalkyl methacrylate gel with bound chymotrypsin (B), compared to chromatography on unmodified gel (A) (Turkovi et al.). Vertical arrow, elution buffer changes from pH 8.1 (0.05 M Tris-hydrochloric acid buffer) to pH 3.1 (approximately 0.1 Macetic acid). Solid line, absorbance at 280 nm; broken line, inhibitor activity; dotted line, pH.
The principle of affinity chromatography has been known for more than 20 y e a k In 195 1, Campbell et al. used it for the first time for the isolation of antibodies by means of a cellulose column with covalently bonded antigen. For the isolation of an enzyme, affinity chromatography was used for the first time in 1953 by Lerman, who isolated tyrosinase on a column of cellulose with ether-bonded resorcinol residues. Subsequently, this method has been only seldom used (Silman and Katchalski), evidently because of the nature of the insoluble carriers of affinants, which did not permit sufficient freedom for the formation of a complex between the isolated substance and the bonded affinant. Major developments of the method have occurred in recent years, the most significant step in the utilization of affinity chromatography being the introduction of the method of affinant binding on to cyanogen bromide-activated Sepharose (AxCn et al., Porath et d.). It was demonstrated by Cuatrecasas and Anfinsen (1971a) that Sepharose fulfils almost all of the requirements for an ideal carrier of the bound affinant. Thus, in 1968, using a Sepharose-bound affinant, Cuatrecasas et al. achieved excellent results in their first work on the affinity chromatographic isolation of nuclease, chymotrypsin and carboxypeptidase. That report, in which affinity chromatography was defined for the first time, led to the wide use of this method in the isolation of enzymes, their inhibitors, antibodies and antigens, nucleic acids, transfer and repressor proteins, hormones and their receptors, and many other substances. References p.96
CHOICE OF BOUND AFFINANT In general, any compound is suitable for the isolation of biologically active substances that bind it specifically, firmly and reversibly. Because of the widely varying character of biologically active substances, chemically affinants are also of very diverse types and their classification is therefore based on their biochemical function rather than on their chemical structure. For the isolation of enzymes, competitive inhibitors can be utilized as affinants (Baker and Siebeneick; Berman and Young; Blumberg et al.; Cuatrecasas, 1970a, b; Cuatrecasas et al. ; Steers et al.) or substrates (Baggio et al., Chua and Bushuk). In these instances, the affinant is bound to the active site of the enzyme and becomes more selective with increasing binding specificity. Further, effectors can alio be used as affinants (Chan and Takahashi, Sprossler and Lingens). In this instance, the bonding does not take place at the enzyme active site, and it is also possible to isolate enzymes of the same class and similar substrate specificity if the differences occur at the site for the binding of the effector. Enzymes, from which apoenzymes can be prepared by the elimination of coenzymes, can be isolated on carriers with bound coenzyme (Lowe and Dean, Weibel et al.). A disadvantage of this method is that in the presence of several apoenzymes that have the same coenzyme, a mixture is often obtained. For the isolation of inhibitors, the corresponding enzymes serve as affinants (Feinstein, Fritz et aZ.). For proteins binding vitamins, the corresponding vitamin serves as the affinant, for example biotin for the isolation of avidin (Cuatrecasas and Wilchek, McCormick). Thyroxine-binding globulin can be isolated using thyroxine as the affinant (Pensky and Marshall), and for estradiol receptor protein, estradiol serves the same purpose (Cuatrecasas, 1970a). For the isolation of polysaccharides or glycoproteins that contain glycosyl as the terminal group, concanavalin A is a suitable affinant (Aspberg and Porath, Edelman et al., Lloyd, Yariv et al.). For the isolation of SH-proteins, p-aminophenyl mercuriacetate is a useful affinant (Sluyterman and Wijdenes). Synthetic ribonuclease S-peptide was purified by means of ribonuclease S-protein bound to agarose (Kato and Anfinsen), the peptide of the active site from the tryptic hydrolyzate of the inhibited nuclease was purified using nuclease bound to agarose (Wilchek), and the peptides containing modified amino acid residues using insoluble antibodies (Wilchek et al.). Blasi and Goldberger (see Cuatrecasas and Anfinsen, 1971b) have shown that histidyl-tRNA synthetase is a suitable affinant for the isolation of histidyl-tRNA. DNA is a suitable affinant for the isolation of gene-specific mRNA (Bautz and Reilly, Nyggard and Hall). Insoluble oligonucleotides and polynucleotides of cellulose were employed for the isolation, fractionation and structure determination of various nucleic acids (Erhan et al., Gilham and Robinson, Sander et al.) and DNA polymerases (Jovin and Kornberg, Litman). For the isolation of DNA polymerases, adsorption on DNA-agarose was found to be more suitable (Poonian e t aZ.). For the isolation of antigen, the corresponding antibodies can be used as affinants, while for the isolation of antibodies, the corresponding antigens are employed (Akanuma et al.; Cuatrecasas, 1969; Goetzl and Metzger; Omenn et al.; Silman and Katchalski; Weintraub; Wofsy and Burr). Hapten can also serve as an affinant for the isolation of whole cells producing antibodies against the bound hapten (Truffa-Bachi and Wofsy).
CHOICE OF BOUND AFFINANT
The thermodynamic character of the bond between the affinant and the isolated biologically active substance (the enzyme) was discussed in a review by Reiner and Walch. The bond between the monomeric affinant A and the enzyme E is expressed by the equilibrium constant of the reaction, K,, on the supposition that this exists in a single tertiary structure form: k E + A & [E.A] k2
‘[E.A] After the binding of the affinant to the solid support, the equilibrium constant KA is affected to a certain extent. An increase in KA results in modification of the affinant by binding to the matrix, and the steric accessibility of the affinant is limited as a consequence of this binding. On the other hand, a decrease in K, causes non-specific adsorption of the enzyme to the solid support and to the molecules of the already adsorbed enzyme. On the assumption that a single enzyme of the crude protein has an affinity for the matrix, the equilibrium between the bound affinant and the isolated enzyme is given by the equation:
E + A’ Ki=
=$ [E.A’] k2
= - R T l n K‘
For the successful isolation of an enzyme by affinity chromatography, KA or K i should be very small for the desired enzyme and should be much smaller than any dissociation constant for adsorption between the protein and the matrix surface (i.e., non-specific adsorption). We can estimate the maximum KL value as follows. Starting mole/l concentration of inhibitor in the insoluble affinant and the requirefrom a ment of a 99% retention of the enzyme from the raw material, which contains about mole/l of enzyme in the three-fold volume of the insoluble affinant, we obtain l o 4 mole/l as the upper limit of the KA value of an effective affinant. In a 3% protein solution, where the active enzyme constitutes 10%of the total protein, which should have a molecular weight of 10 5 ,cu. 10%of the matrix capacity is utilized under the above conditions. From this estimate, it further follows that in view of the bond that can be formed between the inhibitor and the enzyme, the whole purification process should be considered to be precipitation rather than chromatography. This can also be shown by means of the adsorption isotherm for affinity chromatography shown in Fig. 7.4, from which it is evident that the gross adsorption isotherm (curve 3) can be defined as the sum of the specific (curve 1) and non-specific (curve 2 ) adsorption isotherms. The specific adsorption isotherm characterizes the ideal specific adsorption, when the adsorption energy, AG‘, for all adsorbed particles is constant and relatively large. Adsorption ceases when all accessible “affinant” sites are occupied. The non-specific adsorption isotherm characterizes the adsorption of proteins on non-specific sites of the matrix and on already adsorbed protein. References p . 96
ENZYME CONCENTRATION (mg/rnl)
Fig. 7.4. Adsorption isotherm for affinity chromatography (Reiner and Walch). Adsorption isotherms: 1, specific; 2, non-specific; 3, gross.
The AG; value is the sum of AG, and AGnonsp., where ACnonsp. is the reaction energy mole/l for K A of non-specific complexing and hindrance. Inserting a mean value of gives a value of ca. 7 kcal/mole for AG,. The adsorption energy for non-specific adsorption, AGnonsp., results from the hydrophobic, hydrophilic or even ionic interactions and is comparable with the adsorption energy in normal chromatography. It depends greatly on the nature of the solid carrier and the protein. AGnon-sp.should be as low as possible because it also comprises the adsorption of molecules forming non-specific complexes with the affinant. However, instances may occur in which the crude protein contains two or more enzymes that display an affinity for the bound affinant. If the equilibrium constant of the reaction of the second enzyme KA(II) is greater than 16-’ molell, then only minute amounts of the second enzyme will be retained together with the enzyme sought. If KA(II) is less than or equal to mole/l, then a mixture of both enzymes will be adsorbed, even though the KA value of the desired enzyme might be much less than KA(ll). This follows from the specific form of the adsorption isotherm for affinity chromatography, because the heat of adsorption is extremely high under chromatographic conditions. If KA(II) differs from KA by more than 50-100, a separation can still be achieved if differential elution is applied, for example, or if the isolation is carried out by the batch process, using an amount of the insoluble affinant that corresponds exactly to the more intimately binding enzyme, or if chromatography during which equilibrium between [E(II) * A] and [E(I) A] must be attained is very slow.
GENERAL ASPECTS OF THE AFFINANT-SORBENT BOND At the beginning of this chapter, it was indicated how important a role is played by the properties of the solid support in affinity chromatography. The nature of these properties and the extent to which they might affect the results of affinity chromatography are discussed in Chapter 10.
GENERAL ASPECTS OF THE AFFINANT-SORBENT BOND
The difference between AG, and AGA is due not only to the nature of the matrix and the modification of the affinant, but also to the change in its steric accessibility. In view of the different structures of the isolated substances, no general rule exists on the minimum distance between the affinant and the surface of the solid support. However, the affinant should be located at such a distance from the carrier surface that the bond would not require the deformation of the isolated substances. The effect of the distance of the affinant 3’-(4-aminophenylphosphoryl)deoxythymidine-5’-phosphate from the solid support surface (both Sepharose 4B and Bio-Gel P-300) on the capacity of the gel in the chromatography of staphylococcal nuclease (Cuatrecasas, 1970a) is shown in Table 7.1. In type A, the inhibitor is bound directly to the matrix, and in other types a chain of varying length is inserted between it and the carrier surface. TABLE 7.1 CAPACITY OF INSOLUBLE AFFINANTS PREPARED BY BINDING 3’-(4-AMINOPHENY LPHOSPHORYL) DEOXYTHYMIDINE-5’-PHOSPHATE ON SEPHAROSE 4B AND BIO-GEL P-300 DERIVATIVES IN THE AFFINITY CHROMATOGRAPHY OF STAPHYLOCOCCAL NUCLEASE (CUATRECASAS, 1970a) Type of Structure inhibitor bound on matrix
Capacity (mgof nucleaselml of gel) on derivative of Sepharose 4B
The course of affinity chromatography is also affected by the concentration of affinant on the matrix. When Steers et al. isolated 0-galactosidase on agarose containing a small amount of p-aminophenyl-0-D-thiogalactopyranoside, the adsorbed enzyme could References p . 96
be eluted with buffers containing the substrate. At a high concentration of the bound affinant, the adsorbed 0-galactosidase could not be eluted with substrate-containing buffers; on the contrary, even after binding, it remained active and the substrate present in the buffer was cleaved by it during its passage through the column. During the binding of chymotrypsin on various agarose derivatives, Axen and Ernback found that with larger amounts of the chymotrypsin bound per millilitre of the carrier the relative proteolytic activity decreased. Kalderon et al. found that on increasing the concentration of the bound affinant [N-(e-aminocaproy1)-p-aminophenyl]trimethylammonium bromide above 0.16 pmole per millilitre of the support, the adsorbent specificity for the binding of acetylcholinesterase decreased. This might be explained by the non-specific sorption on the adsorbent which acquired ion*exchanging properties through the increase in the content of ammonium groups. It had originally been hoped that the enzyme would be selectively adsorbed on to the insoluble affinant at high ionic strengths at which nonspecific electrostatic interactions would be avoided. However, the decreased affinity of acetylcholinesterase inhibitors for the enzyme at high ionic strength precluded the use of this approach. The capacity of the bound affinant is further influenced by preserving its original conformation with as little change as possible. Cuatrecasas (1970a) isolated insulin on columns of Sepharose with an antibody against hog insulin bound at both pH 6.5 and pH 9.5. As will be shown in Chapter 10, protein is bound on cyanogen bromide-activated Sepharose by its non-protonated forms of amino groups. On decreasing the pH, a decrease in the number of binding groups also takes place and the result of the pH difference was that the first derivative was able to bind almost 80%of the theoretically possible amount of insulin, while the second derivative, prepared by binding at pH 9.5, took only 7% of the capacity for insulin. As the total content of the bound affinant was identical in both instances, the second derivative must have contained immunoglobulin, which is unable t o bind antigen effectively. In the case of a large number of bound amino groups, disturbance of the native tertiary structure evidently occurred. Even at a low pH, adsorbents can be obtained that contain a large amount of active protein bound to Sepharose if the amount of cyanogen bromide is increased during the activation and the amount of protein during the binding. The importance of preservation of the original conformation of the affinant after binding on solid support was also shown by Lowe and Dean for the binding of the cofactor.
REFERENCES Akanurna, Y., Kuzuya, T., Hayashi, M.,Ide, T. and Kuzuya, N., Biochem. Biophys. Res. Commun., 38 (1970) 947. Aspberg, K. and Porath, J.,Acta Chem. Scand., 24 (1970) 1839. Axin, R . and Ernback, S., Eur. J. Biochem., 18 (1971) 351. Axdn, R., Porath, J . and Ernback, S., Nature (London), 214 (1967) 1302. Baggio, B., Pinna, L. A., Morel, V. and Siliprandi, N., Biochim. Biophys. Acta, 212 (1970) 515. Baker, B. R. and Siebeneick, H. U., J. Med. Chem., 14 (1971) 799. Bautz, E. K. F. and Reilly, E., Science, 151 (1966) 328. Berman, J. D. and Young, M.,Proc. Nat. Acad. Sci. US.,68 (1971) 395.
REFERENCES Blumberg, S., Schechter, 1. and Berger, A., Eur. J. Biochem., 15 (1 970) 97. Campbell, D. ti., Luescher, E. L. and Lerrnan, L. S., Proc. Nut. Acud. Sci. US.,37 (1951) 575. Chan. W. W. C. and Takahashi. M., Biochem. Biophys. Res. Commun.. 37 (1969) 272. Chua, G. K. and Bushuk, W., Biochem. Biophys. Res. Commun., 37 (1969) 545. Cuatrecasas, P., Biochem. Biophys. Res. Commun., 35 (1969) 53 1. Cuatrecasas, P., J. Biol. Chem., 245 (1970a) 3059. Cuatrecasas, P., Nature (London), 228 (1970b) 1327. Cuatrecasas, P. and Anfinsen, C. B., Methods Enzyrnol., 22 (197 la) 345. Cuatrecasas, P. and Anfinsen, C. B., Annu. Rev. Biochem., 40 (1971b) 259. Cuatrecasas, P. and Wilchek, M., Biocfzem. Biophys. Res Comnzun., 33 (1968) 235. Cuatrecasas, P., Wilchek, M: and Anfinsen, C . B., Proc. Nut. Acud. Sci. U.S., 61 (1968) 636. Edelman, G. M., Rutishauser, U. and Millettc, C. F.. Proc. Nut. Acud. Sci. US.,6 8 (1971) 2153. Erhan, S. L., Northrup, L. G. and Leach, F. R., Proc. Nut. Acad. Sci. US.,53 (1965) 646. Feinstein, G., Biochim. Biophys. Actu, 236 (1971) 73. Fritz, H., Gcbhardt, M., Mester, R., Illchrnann, K. and Hochstrasser, K., Hoppe-Seyler’s Z. Physiol. Chem., 351 (1970) 571. Gilham, P. T. and Robinson, W. E., J. Amer. Chem. Soc., 86 (1964) 4985. Goetzl. I-. J . and Metzgcr, H., Biochemistry, 9 (1970) 1267. Jovin, T. M. and Kornberg, A., J. Biol. Chem., 243 (1968) 250. Kalderon, N., Silman, 1.. Blumberg, S. and Dudai, Y., Biochim. Biophys. Acta, 207 (1970) 560. Kato, I . and Anfinsen, C. B., J. Biol. Chem., 244 (1969) 5849. Lernian, L. S., Proc. Nut. Acad. Sci. U.S., 39 (1953) 232. Litman, R. M.,J. Biol. Chem., 243 (1968) 6222. Lloyd, K. O., Arch. Biochem. Biophys., 137 (1970) 460. Lowe, C. R. and Dean, P. D. G . , FEBS Lett., 14 (1971) 313. McCormick, D. B.,Anul. Biochem., 13 (1965) 194. Nyggard. A. P. and Hall, B. D., Biochem. Biophys. Res. Commun., 12 (1963) 98. Omenn, G., Ontjes, D. A. and Anfinsen, C. B.. Nature (London), 225 (1970) 189. Pensky, J. and Marshall. J. S., Arch. Biochem. Biophys., 135 (1969) 304. Poonian, M. S., Schlabach, A. J . and Weissbach, A., Biochemisfiy, 10 (1971) 424. Porath, J., Biotechnol. Bioeng., Symp., No. 3 (1972) 145. Porath, J., Ax&&R. and Ernback, S.. Nufure (London)., 215 (1967) 1491. Reiner, R. H. and Walch, A.. Chromarogruphia. 4 (1971) 578. Sander, E. G., McCormick, D. B. and Wright, L. D.,J. Chromafogr., 21 (1966) 419. Silrnan, 1. H. and Katchalski, E., Annu. Rev. Biochem., 35 (1966) 873. Sluyterrnan, L. A. AE. and Wijdenes, J., Biochem. Biophys. Actu, 200 (1970) 593. Sprossler, B. and Lingens, F., FEBS Lett.. 6 (1970) 232. Steers, E., Cuatrecasas, P. and Pollard, H., J. Bioi. Chem., 246 (1971) 196. Truffa-Bachi, P. and Wofsy, L.. Proc. Nut. Acad. Sci. US.,66 (1970) 685. Turkovi, J., Hubilkovi, 0.. Kfivikova, M. and koupek, I., Biochim. Biophys. Actu, 322 (1973) 1. Weibel, M. K., Weetall, H. H. and Bright. A. J., Biochem. Biophys. Res. Commun., 44 (1971) 347. Weintraub, B. D., Biochem. Biophys. Rex Commun., 39 (1970) 8 3 . Wilchek, M., FEBS Lett., 7 (1970) 161. Wilchek, M., Bocchini, V., Becker, M. and Givol, D., Biochemistry, 10 (1971) 2828. Wofsy, L. and Burr, B., J. Immunol., 103 (1 969) 380. Yariv, J., Kalb, A. J. and Levitzki. A., Biochim. Biophys. Actu, 165 (1968) 303.