2 Physical Methods Part (iv) Chromatography (b) Affinity Chromatography By H. GUILFORD The Radiochemical Centre Amersham Bucks 1 Introduction Affinity -or more properly bioaffinity -chromatography is a separation technique depending on biospecific interaction between an immobilized ligand and a species for which this ligand is in the broadest sense a substrate. Introduced into general use cu. 1968 its wide acceptance as a useful biochemical tool may be adjudged from the appearance in 1974 of nearly two hundred pertinent papers a chemical review,‘ and an excellent book covering advances to mid-1973,2 and the commercial availability of many reagents. Biospecific adsorbents are now being used not only for chromatography but as mechanistic probes and matrix- bound reactants.Recently after the initial euphoria there has been time for reflection leading to some rigorous re-investigations which show that the term ‘bioaffinity’ is not strictly applicable to some well-known oft-quoted cases. Many biochemists have one reservation about affinity chromatography. They doubt their ability to undertake the synthesis of some of the more sophisticated affinity systems. Indeed at a recent international conference biochemists were described as scientists who cheerfully isolate minute amounts of fragile proteins but baulk at manipulating ten grams of simple chemicals. Whatever the truth of this cynical view (voiced by a biochemist it should be added) clearly there is a need for chemists and engineers to contribute to this field of biological study.Hopefully this chapter in Annual Reports will stimulate more interest among this section of the scientific fraternity. 2 Matrix Chemistry Cyanogen bromide (CNBr)-activated agarose continues to be the most widely reported support matrix. Two new simplifying modifications of the activation procedure eliminate the need for careful pH control. The cyanogen halide can be added either in aqueous solution to a suspension of gel in concentrated phosphate buffer at pH 11-12 or in acetonitrile solution to a suspension in ~arbonate.~ The influence of concentrations temperature and duration of H. GuiIford Chem. SOC.Rev. 1973 2 249. C. Lowe and P. D. G. Dean ‘Affinity Chromatography’ Wiley New York,1974. J. Porath K. Aspberg H.Drevin and R. Axen J. Chromatog. 1973 86 53. S. C. March I. Parikh and P. Cuatrecasas Analyr. Biochem. 1974,60 149. 56 Physical Methods-Par t (ivb) Afinity Chromatography 57 reaction on coupling is discussed as it is also in a study of immobilizing glycine and IgG.’ Prehydrolysis of the activated gel before coupling as a means of controlling yields has been investigated.6 The coupling of amino-acids to activated polysaccharides has been shown to involve isourea rather than carbon- ate or cyclic imidocarbonate linkage^.^ Such ammonia as is liberated is indepen- dent of the coupling reaction. The monomeric model compound methyl-46- 0-benzylidene-a-D-glucopyranoside, when activated reacts with glycine to form the isourea (1) only. A linear relationship exists between the optimum pH of /o-CH2 0-C-NHR I + NH (1) R = CH,CO,-or CH,CONHCH,CO,-coupling and the pK of the conjugate acid of the coupled amine.8 The slight pH-dependent leakage of ligands from CNBr-activated polysaccharides and from polyacrylamide supports is now well established.’ The enhancement of stability by multivalently coupled ligands demonstrated by Wilchek with poly- (L-lysyl)agarose has been quantified.lo It appears that increasing the number of linkages from 1 to 2 gives a much more worthwhile increase in stability than an increase from say 5 to 10. However this leakage problem is significant only when extremely small amounts of material are involved (e.g.in receptor localiza- tion). More details are available of the preparation and properties of gels derivatized by the relatively hydrophilic bifunctional oxirans suitable for attaching ligands uia hydroxy-groups.Affinity chromatography of a p-amylase on cyclohexa- amylose linked to agarose through 1,4-bis-(2,3-epoxypropoxy)butanehas been demonstrated.” Ovine hormones have been coupled to a new matrix divinyl- sulphonylagarose for isolation of antibodies. 3 Spacer-arms It is a widely accepted tenet of affinity chromatography that for a tailored adsorbent to function as intended its biospecific ligand must be separated from 5 D. F. Stage and M. Mannik Biochim. Biophys. Acta 1974 343 282. 6 Separation News 1974 March. 7 K. Brostrom S. Ekman L. KAgedal and S. Akerstrom Acta Chem. Scand. 1974 B28 102.8 L. Kagerdal Biochem. SOC.Trans. 1974 2 1328. 9 G. I. Tesser H.-U. Fisch and R. Schwyzer Helv. Chim. Acta 1974 57 1718. 10 T. C. J. Gribnau and G. I. Tesser Experientia 1974 in press. 11 L. Sundberg and J. Porath J. Chromatog. 1974 90 87. 12 P. Vretblad F.E.B.S. Letters 1974 47. 86. 13 M. R. Sairam W. C. Clarke D. Chung J. Porath and C. H Li Biuchem. Biophys. Res. Comm. 1974 61 355. H. Guilford the carrier matrix so as to render it sterically available to the (often much larger) moiety for which it has affinity. This effect is achieved by inserting a spacer-arm between the ligand and its support. Matrices which are recognized as capable of introducing non-specific effects by being either ionic e.g. ion-exchangers or hydrophobic e.g.polystyrenes have been avoided in affinity chromatography. Agarose glass and polyacrylamide have been carefully chosen as supports which minimize both types of interaction.' The choice of spacers however has usually been much less rigorous being dictated mainly by convenience of chemical modification commercial accessibility and the procedure for immobilization on a matrix. Typical examples are &-amino-caproic acid am-diamines and small peptides (e.g.glycylglycyltyrosine). It has now become apparent that the nature of the spacer can be highly significant since some so-called 'sticky' enzymes have such an affinity for the spacer moiety that biospecific interaction with the ligand is almost eclipsed. The involvement of spacers has been particularly intensively investigated by O'Carra and his colleagues at Galway some of their earlier results being sum- marized in an excellent review.14 A number of enzymes have been chromato- graphed not only on a supposed biospecific assembly but also on the matrix- spacer system either alone (i.e.lacking a ligand) or coupled to a residue having structural analogy with the ligand but lacking affinity for the enzyme.A per- suasive example of the involvement of a spacer is illustrated in Figure 1. Cuatre-casas had earlier shown that when p-aminophenyl-p-D-thiogalactopyranoside (2),a weak inhibitor of P-galactosidase is attached directly to agarose it shows no affinity for the enzyme. If a 10A spacer N-(2-aminoethyl)acetamide,is inserted there is retardation of fl-galactosidase and if system A (ca.21 A) is used the enzyme is strongly adsorbed. These observations have been widely quoted as proof of the importance of spacer length and ligand accessibility in bioaffinity chromato- graphy. However not only can the enzyme not be displaced from the long-arm gel by j3-thiogalactose but replacing the inhibitor ligand by an a-glycoside having no affinity for the enzyme does not reduce the binding of P-galactosidase to the gel. Rood and Wilkinson have attempted to purify influenza virus sialidase (muco- polysaccharide N-acetyl neuraminylhydrolase) by the method of Cuatrecasas and Ilhairo on columns prepared by coupling the diazonium derivative of the inhibitor N-(p-aminopheny1)-oxamicacid (3) to agarose-glycyl-tyrosine.Several * P.O'Carra F.E.B.S. Symposium 'Industrial Aspects of Biochemistry' ed. B. Spencer 1974 p. 107. Physical Methods-Part (ivb)Afinity Chromatography 0 I M Borate h pn 10 APPLICATION 1 2 3 4 I L I -EFFLUENT VOLUME (in COLUMN-VOLUME UNITS) E NH-CHI-CH,-CHI-NH-CHr~HI-CHI-NH-M-CH,-CH,-C~NH -Q-lde 41 i--06 0 1 M Borate -04 APPLICATION -OF SAMPLE (Reproduced by permission from F.E.B.S. Symposium ‘Industrial Aspects of Bio-chemistry’ 1974 p. 124) other protein components of the crude mixture were adsorbed on this column and omitting the inhibitor ligand had little effect on the chromatography which was concluded to be predominantly ion-exchange.” Clearly in both cases a non-biological interaction with the spacer was operating.It is usual to distinguish only between biospecificity and ‘non-specific’ effects due to gross electrostatic hydrophobic or other interactions. O’Carra points out l4 that this is inaccurate since the latter may be highly specific as in the former case discussed above which indeed resulted in a useful purification of the enzyme but in a way that has little to do with bioaffinity. J. I. Rood and R.G. Wilkinson Biochim. Biophys. Acta 1974 334 168. H. Guilford $-NH-CH2-CH,-CH,-CH,-CH,-CH2-NH-CO-CH,-NH 0 Sepharose 4B Figure 2 Some typical matrix-spacer arm assemblies bearing no biospecific ligand It is hydrophobic interactions between enzymes and spacers that have been particularly ignored in many studies of affinity chromatography.Systems such as those illustrated in Figure 2 should have minimal biological or ionic affinity for proteins yet a number of so-called ‘sticky’ enzymes e.g P-galactosidase yeast alcohol dehydrogenase xanthine oxidase xanthine dehydrogenase and AMP aminohydrolase are strongly adsorbed on a variety of such gels. The thesis that the hydrophobicity of the spacers is the dominant effect has recently been examined.I6 A number of hydrophilic spacers have been attached toagarose a typical synthesis being summarized in Figure 3. Each was constructed so that I o Agarose 1.3-DAP $-NHCH,CHCH,NH OH 0 I 1.3-DAP -NHCH,CHCH,NHCOCH,Br N~HCOI OH OH I I 3-NHCH2CHCH,NHCOCH,NHCH,CWCHCHzNHz Figure 3 Synthesis of a typical hydrophilic spacer-arm based on 1,3-diaminopropan-2-oI (I,3-DAP) at least every alternate atom forms part of relatively hydrophilic groups such as amido secondary amino or carbinol.The behaviour of several NAD +-dependent dehydrogenases was studied using the systems illustrated in Figure 4. Lactate dehydrogenase recognizes the immobilized oxamate derivative as a substrate analogue. Owing to its ordered kinetic mechanism it can only bind biospecifically in the presence of NADH and is released when the coenzyme is omitted from the eluant. With the hydrophobic system A it is necessary to use buffers of high ionic strength to overcome non-biospecific adsorption. With gels B and C this pre- caution is unnecessary. Lactate dehydrogenase binds to either B or C in the presence of NADH and elutes in its absence at low ionic strength.Similar effects P. O’Carra S.Barry and T. Griffin,F.E.B.S. Letters 1974 43 169. Physical Met hods-Pa rt (iub) Afinity Chromatography +9 z I z I1 0-u I I u Iz3 0-uI u I3:z I I O=U I 3:z I V I xz 0-uI u I z7 z7 $ z7 z? " z UI xx 0-0I u I zz 77777 3? I 3:z I V I zx 0-0 3 u I zz I uIO=U I 3:z I u I X3 0-u 3:u I 3? IN z? z7 1- I xz 'SIuI xz 0-uI u I zz I rn " 3? I 3 z Is 1-z? 3? IN IN 0 3:u I I V u x L) 3:u I Xz I m I 3 z I I3:z 0-u I u I3 z I 8 s IT7n zz I mmr zz 'I T7ml 4c m u W LLO H. Guilford are observed when comparing adsorbents D E and F. The 'sticky' enzyme alcohol dehydrogenase is strongly adsorbed on D from which it cannot be eluted with NAD' or NADH yet it is not significantly retarded by gels E or F implying that this enzyme has little bioaffinity for the C-8-linked azo-substituted NAD' residue.In contrast the chromatographic behaviour of this enzyme on gels G and H is similar suggesting that biospecific adsorption predominates on the 6-thiol-linked NAD'. These and other results show that much interference can be eliminated by using hydrophilic spacer-arms. It is hoped that commercial suppliers of functional matrices will take note. The effect of increasing spacer-arm length on the efficacy of a homologous series of N6-(o-aminoalkyl)-AMP's (4) as bioaffinity ligands for dehydrogenases HO OH (4) has been reported.' The binding of the lactate dehydrogenase isoenzymes H and M increases with values of n from 2 to 5 and increasing n above 7gives no better binding of any enzyme studied.It appears that above a certain spacer- length no greater steric availability can be attained as predicted.' 4 Bioafflnity Chromatographic Purifications The examples given in this section are intended to illustrate the range of appli- cations of affinity chromatography having been selected with a bias towards chemical novelty. The selection is far from exhaustive. 4-Phospho[ 14C]pantetheine-labelled pigeon liver fatty acid synthetase has been separated into two half-molecular weight subunits by affinity chroma- tography on &-aminocaproylpantetheinecoupled to CNBr-activated agarose. Acetyl CoA transacylase is retained by this adsorbent whereas the p-keto-acyl thioester reductase subunit containing 4-phospho[ 14C]pantetheine is not." The human serum cortisol-binding protein transcortin has been purified on an immobilized steroid analogue synthesized from corticosterone (5) as shown in Figure 5 from which it is eluted by a pulse of corti~ol.'~ M.C. Hipwell M. J. Harvey and P. D. G. Dean F.E.B.S. Letters 1974 42 355. F. A. Lornitzo A. A. Qureshi and J. W. Porter J. Biol. Chem. 1974 249 1654. l9 F. Le Gaillard A. Racadot N. Racadot-Leray and M. Dautrevaux Biochimie 1974 56 99. Physical Methods-Pa rt (ivb) Afinit y Chroma tograp h y COCH,OH CO,H HCIO, + H,N(CH,),NH(CH,),NH-E & { fi (5) &NH(CH,I,UH(CH&NH Figure 5 Synthesis of an immobilized cortisol analogue The troponin complex of muscles involved in regulating ATP-ase activity has at least three components troponins I C and T.Troponin I is particularly difficult to purify by conventional techniques because of its susceptibility to breakdown by endogenous cathepsins (proteolytic enzymes). This component forms a urea-stable complex with troponin C in the presence of Ca2+. This observation forms the basis-of a biospecific purification of troponin I from a crude rabbit muscle homogenate in Ca2+-containing buffered urea. A column of troponin C immobilized a CNBr-activated agarose adsorbs only troponin I from the mixture. Elution is effected with ethanedeoxybis(ethy1amine)tetra-acetate known to dissociate the complex.20 Bacterial luciferase has been chromatographed using the immobilized sub- strate flavin mononucleotide as bioadsorbent.2 ' Thymidylate synthetase cata- lyses an interesting reaction in DNA metabolism in which a single carbon unit is transferred from a folate coenzyme with concomitant reduction according to 2'-deoxyuridylate + 5,10-methylene-5,6,7,8-tetrahydrofolate*5-methyl-2'-de-oxyuridylate(thymidy1ate)+ 7,8-dihydrofolate.0 OH (6) '* H. Syska S. V. Perry and I. P. Trayer F.E.B.S. Leiiers 1974 40,253. '' C. A. Waters J. R. Murphy and J. W. Hastings Biochem. Biophys. Res. Comm. 1974 57 1152. 64 H. Guilford The inhibitor 5-fluoro-2’-deoxyuridine 5’-(p-aminophenylphosphate) (6) attached to CNBr-activated agarose through a spacer-arm provided a bioaffinity ligand on which thymidylate synthetase could be isolated from a cell-free extract of amethopterin-resistant Lactobacillus ~asei.~’ An adsorbent designed for fractionation of nucleic acids has been prepared by synthesizing an immobilized actinomycin analogue.3-Hydroxy-4-methyl-2-nitrobenzoic acid is coupled to aminohexylagarose. Reduction of the nitro-group followed by oxidative coupling with 2-amino-3-hydroxy-4-methylbenzoyl-/l-(N~-diethylamino)ethylamid~ affords the adsorbent (7) which shows a prefer- ential affinity for DNA over RNA.23 6-Phosphogluconic acid (8)linked through gNH(CH,),NHCO CONHCH ,CH ,NEt I I Me Me OH an amide bond to aminohexylagarose provides an adsorbent for glycerol-3- phosphate dehydrogenase which can be retrieved by elution with the natural substrate @-glycerolphosphate.The novelty of this system is that in solution 6-phosphogIuconic acid exhibits no affinity for this enzyme.24 Bilirubin has been purified on an albumin-agarose gel ;conversely albumin can be removed from serum with a bilirubin-agarose ad~orbent.’~ Cyclic AMPdependent Enzymes.-.The activity of a number of protein kinases which catalyse the phosphorylation of proteins according to Protein + ATP S phosphorylated protein + ADP is stimulated by adenosine 3’,5’-cyclic phosphate cyclic AMP. In several cases these enzymes have been shown to dissociate reversibly into catalytic and inactive cyclic AMP-binding components in the presence of cyclic AMP. In the one previous report on chromatography of a protein kinase on an immobilized cyclic AMP analogue [N6-(.+aminocaproyl)-cyclicAMP-agarose] the enzyme did dissociate the catalytic fraction now fully active i.e.no longer stimulated by cyclic AMP eluting unpurified in the void volume while the receptor unit defied all attempts to liberate it from the adsorbent. More recently three of these enzymes have been subjected to affinity chroma- tography on immobilized cyclic AMP analogues with spacer-arms attached at 22 J. M. Whiteley I. Jerkunica and T. Deits Biochemistry 1974 13 2044. ’’ F. Seela 2.Nururforsch. 1974 29 521. 24 J. F. McGinnis and J. de Vellis Biochem. Biophys. Res. Comm. 1974 60,186. 25 P. H. Plotz P. D. Berk B. F. Scharschmidt J. K. Gordon and J. Vergalla J. Clin. Invest. 1974 53 778; M. Hierowski and R. Brodersen Biochim.Biophys. Acta 1974 354 121. Physical Methods-Part (iub) Afinity Chromatography I NH I C-8 of the adenine ring instead of to the 6-amino-group. 8-(y-Carboxypropytthio)-cyclic AMP coupled to poly(1ysyl)agarose using a di-imide yields the adsorbent (9) on which pig brain histone kinase is adsorbed. In contrast to the above example an increase in ionic strength is required to release a fully active catalytic component along with much non-specifically adsorbed protein. A further increase released an eiectrophoretically homogeneous cyclic AMP-binding component thus achieving an apparently biospecific purification of the receptor component of a histone kinase." A partial purification of the receptor unit of pig muscle protein kinase has also been reported27 using 8-(6-aminohexyl)- amino-cyclic AMP (lo),' synthesized from 8-bromo-cyclic AMP and 1,6-diamino- hexane coupled to CNBr-activated agarose.The free ligand was shown to function as a competitive inhibitor of cyclic AMP binding to the receptor site but with much lower affinity. After incubating the adsorbent with the kinase a denatured protein was eluted by urea removal by dialysis of which regenerated cyclic AMP binding capacity. The catalytic part of this enzyme was not adsorbed or purified on this adsorbent. 8-(6-Aminohexyl)amino-cycIicAMP is not only an activator of protamine kinase from trout testes but at higher concentrations it inhibits competitively with respect to ATP as do both cyclic AMP and AMP. The catalytic part not 26 E.S. Severin S. N. Kochetkov M. V. Nesterova and N. N. Gulyaey F.E.B.S. Letters 1974 49 61. 27 J. Ramseyer H.R. Kaslow and G. N. Gill Biochem. Biophys. Res. Comm. 1974,59,813. 66 H. Guilford further stimulated by cyclic AMP is again eluted at increased ionic strength but also by a pulse of AMP with up to 100-fold purification. It seems possible that the enzyme dissociates on the immobilized ligand (which acts as an effector causing a conformational change and binding the receptor unit) and that the catalytic component is subsequently adsorbed on to other vacant ligands acting at the ATP binding site from which it can be eluted by the competitive inhibitor AMP. If this is the case then an interesting dual biospecificity is operative.8-(6-Aminohexyl)amino-AMPwas also synthesized in an attempt to exploit the inhibition by AMP in affinity chromatography. This failed because neither this analogue nor N6-(6-aminohexyl)-AMP inhibits protamine kinase and this enzyme has no affinity for these immobilized ligands2' These studies illustrate an advantage of using complete ligand-spacer molecules in affinity chroma- tography :the behaviour of the whole unit can be tested in solution often demon- strating differences in affinity between an inhibitor and the analogue for immobil- ization. 5 General Ligands The concept of general ligands for group-specific bioaffinity chromatography continues to be the focus of much attention. Purification of lectins by this method is particularly attractive since each of these carbohydrate-binding proteins can be eluted specifically with an appropriate saccharide.Barondes and co-workers have exploited the definitive ability of lectins to agglutinate erythrocytes by using formalinized erythrocytes as the affinity adsorbent on which lectins from jack bean meal wheat germ lipase Ulex europeus seeds and lima beans have been fra~tionated.~' Immobilized lectins have been used to fractionate b-galactosi- da~es.~' Immobilized streptomycin and gentamicin provide useful adsorbents for ribosome^.^' Adenyl Ligands.-Many enzymes function only in conjunction with a cofactor which is generally a chemically well-defined compound of low molecular weight. Some part of its structure can be identified with its catalytic involvement e.g.the dihydropyridino-moiety of NAD(P)H the cobalt-alkyl of 5-deoxyadenosyl-cobalamin the y-phosphate of ATP etc. Immobilization of these compounds has been the focus of much effort because of their potential as bioaffinity ligands and in the search for matrix-bound coenzymes retaining biological activity and readily retrievable for re-use. With the latter consideration in mind synthetic operations have been concentrated at sites remote for that of catalytic function. A component of many coenzymes [e.g.ATP NAD(P)' FAD and CoA] is the adenyl moiety itself rarely involved in chemical reaction but often having a separate binding site on the enzyme sometimes acting as inhibitor or effector. 28 B. Jergil H. Guilford and K. Mosbach Biochem.J. 1974 139 441. 29 R. W. Reitherman S. D. Rosen and S. H. Barondes Nature 1974 2411 599. 30 A. G. W. Norden and J. S. O'Brien Biochem. Biophys. Res. Comm. 1974 56 193; Separation News 1974 September. 31 F. Le Goffic,B. Baca and N. Moreau F.E.B.S. Letters 1974 41 69. Physical Methods-Part (ivb)Afinity Chromatography NH -(CH,),-NH2 NH-(CHz),-NH 0 OH 0 II IOH CH,-0-P-OR OH OH OH OH A B 00 II II 00 OH OH OH OH C D Figure 6 Adenosine phosphate and pyrophospha~e Iigands (R= H monophosphate diphosphate or imidodiphosphate) Hence much elegant chemistry has been evolved in the search for AMP- analogues coupled to supports. Several points of attachment have been con- sidered allowing comparison of affinities of one enzyme for the same basic ligand presented with different topographies (cJ Figures 4 and 6) sometimes providing insight into the necessary features for interaction in a particular case.Several significant advances have been made during 1974 including an excellent summary of the status quo of matrix-bound adenine n~cleotides.’~ Trayer’s group at Birmingham has compared and contrasted three series of adenosine 5’-mOnO- 5’-di- and 5’-imidodi-phosphates (and in addition simple pyrophosphate) (Figure 6). In each case the ligand-spacer has a terminal amino- group for attachment to CNBr-activated agarose. Affinity chromatography of several ATP-utilizing enzymes has been studied using the ADP-agarose gels. The muscle protein myosin and its biologically active subfragments are adsorbed and can be eluted with ATP.Glucokinase is specifically adsorbed on to the C8-ADP-agarose [cf. Figure 6 A R = -PO(OH),]. Hexokinase however 32 P.-0. Larsson ‘The Synthesis of Polymer-Bound Adenine Nucleotides and Their Use as Affinity Adsorbents and Active Cofactors’ University of Lund 1974. H. Guilford is not retained on this system but is adsorbed on to N6-(ti-aminohexyl)-ADP- agarose illustrating how apparently minor differences in ligand construction can drastically affect bi~specificity.~’ Several groups have synthesized adenyl analogues in which the 6-amino-group is replaced by a thiol. 6-Mercaptopurine riboside riboside-5’-phosphate riboside-5’-triphosphate and nicotinamide-6-mercaptopurinedinucleotide have been coupled to the matrix-spacer N ‘-agarosyl-N6-bromoacetyl-1,Bdiamino-hexane.Anderton and co-workers prepared the ATP analogue (11) by treating 2’,3’- 0-isopropylidene-6-mercaptopurineriboside with 2,4-dinitrofluorobenzene,then sequential phosphorylation. Coupling to ‘thiolated agarsse’ was accomplished under mild conditions affording adsorbent (12) used in chromatography of HNAc I 02NDNo2 S S(CH,),CHCONH(CH,)3NH(CHz)3NH I R R (1 1) (12) R = ribosyl 5‘-triphosphate Na* K+ATpa~e.~~ An alternative synthesis of N6-(o-aminoalky1)-AMP’s (4) differs from the original method only in that the am-diaminoalkane is treated with 6-mercaptopurineriboside 5’-phosphate instead of the 6-chloro-derivative.’ Mosbach’s group in Lund has further extended its studies of general ligands tailored to dehydrogenases.N6-(6-Aminohexy1)adenosine2,Sdiphosphate and N6-(6-aminohexy1)adenosine 3’,5’-diphosphate have been synthesized in a manner analogous to the corresponding 5‘-monophosphate (4;n = 6) except in that phosphorus trichloride is substituted for phosphoryl chloride to provide 2(3’),5’-diphosphorylation.Several NADP+-dependent enzymes which show little or no affinity for the 5’-monophosphate immobilized or otherwise are retained on the matrix-bound 2’,5’-diphosphate (cf. Figure 7 for structural relationship to the coenzyme). The 3’,5’-diphosphate (more formally related to CoA)is a relatively ineffective affinity ligand as expected from its lack of inhibition of most dehydrogenases in s~lution.’~ I.P. Trayer H. R. Trayer D. A. P. Small and R. C. Bottomley Biochem. J. 1974 139 609; I. P. Trayer and H. R. Trayer Biochem. J. 1974 141 775. B. H. Anderton F. W. Hulla H. Fasold and H. A. White F.E.E.S. Letters 1973 37 338. D. B. Craven M. J. Harvey C. R. Lowe and P. D. G. Dean European J. Biochem. 1974 41 329. P. Brodelius P.-0. Larsson and K. Mosbach Europcwn J. Biochem. 1974 47 8 1 Physical Methods-Part (ivb) Afinity Chromatography 69 Figure 7 summarizes the first synthesis of a chemically defined immobilized NADP+ analogue. The key step is the quaternization of the N' position of the adenine ring by iodoacetic acid which provides a mild method for introducing a functionality into NADP' (and NAD')32 suitable for attaching a spacer and subsequent immobilization.This NADP+-analogue retains a considerable measure of its coenzyme function when coupled to soluble dextran and appears to be useful as a bioaffinity ligand for NADP+-dependent dehydrogenases when insolubilized on agar~se.~~ Applications of General Ligands.-The above studies have come from laboratories cancerned with developing affinity chromatography per se. The acceptability of the concept of general ligands is apparent from its application to various purifi- cations. Myo-inositol-1-phosphate synthetase from rat testes and anaemic chicken blood3' has been purified using NAD+-agarose which has also been effective in purifying diphtheria toxin39 and 8-hydroxybutyrate dehydrogena~e.~' Human erythrocyte glucose-6-phosphate dehydrogenase has been partially purified on NADP+-agaro~e.~' The isoenzymes of horse alcohol dehydrogenase have been separated on AMP-agaro~e.~~ 6 NewTechniques Assay of Coupling Yields.-Proteins coupled to polysaccharides or collagen have been determined by measuring the tryptophan content by the acid-ninhydrin reaction.43 A high blank was reported with CNBr-activated agarose.A fluoro-metric assay has been developed in which fluorescamine is attached to amino- ligands liberated from the agarose matrix by basic hydroly~is.~~ Affinity Electrophoresis.-The technique of crossed immuno-affinity electro- phoresis has been designed to predict the efficacy of biospecific affinity chroma- tography systems. In the pilot study a concanavalin A-agarose gel component was incorporated on to the electrophoretic plate and the behaviour of serum proteins on this modified system was compared with that on a standard immuno- electrophoretic plate.45 The migration of glycoprotein components with o[-D-glucosidic or a-D-mannosidic termini with which concanavalin A binds was clearly affected in a way which paralleled the behaviour of serum protein on a conventional biospecific column.Concanavalin A is an example of a phyto- haemagglutinin or lectin. These proteins reminiscent of antibodies interact with 37 C. R. Lowe and K. Mosbach European J. Biochem. 1974 49 51 1. 38 F. Pittner W. Fried and 0. Hoffman-Ostenhof 2.physiof. Chem. 1974 355 222; R. Schwarcz W. Fried F. Pittner and 0.Hoffmann-Ostenhof Monatsh. 1974,105,445.39 C. Cukor J. D. Readio and R. J. Kuchler Biotechnof. and Bioeng. 1974 16 925. 40 A. K. Grover and G. G. Hammes Biochim. Biophys. Acta 1974 356 309. 4' A. De Flora F. Giuliano and A. Morelli Ital. J. Biochem. 1973 22 258. O2 L. Anderson H. Jornvall A. Akeson and K. Mosbach Eiochim. Biophys. Acta 1974 364 1. 43 A. Eshkami T. Chase jun. J. Freudenberger and S. G. Gilbert Anafyt. Biochem. 1974 57 421. 44 M. Naoi and Y.C. Lee Analyt. Biochem.. 1974.57 640. 45 T. C. Beg-Hansen Analy!. Biochem. 1973,56,480. H. Guilford X 2 2" V O=Y (=y I a I ci P d r" 5 z 8 G2-f X I 2 a v I o=v X -2 2" Y +$ Physical Methods-Part (iub)Afinity Chromatography 71 cell membrane components agglutinating embryonic and tumour cells and erythrocytes.Many have a sugar-binding specificity an observation exploited in affinity chromatographic purification A range of sugar-containing neutral hydrophilic gels has recently been prepared by copolymerization of alkenyl-0- glycosides with acrylamide and NN'-methylene-bisacrylamide. An example immobilized 0-a-L-fucopyranoside is tentatively assigned structure (13).46 \ / NH,-CO-CH / CH \2 /CH-CO-NH CH-CO-NH CH, \ CH\,CH /CH NH NH CH / \ /\ /\co/\FH2 CH co cHz CH-CO-NH CH-CO-NH, / \ /CH2 CH\, oO-CH,-(CH ) -CH NH,-CO-CH "/ \ 10 CH /CH, (q / CH-CO-NH CH-CO-NH OH \ CH /CH, \,CH NH NH CH / \ /\ /\ /\ CH\ CO CH CO /CH2 NH,-CO-CH NH,-CO-CH / \ CH /CH OH /\,CH-CO-NH CH-(CH,),-CH, \ p&& CH /CH, \2 NH,-CO-CH NH,-CO-CH / \ CH (1 3) Partial tentative structure of an 0-a-L-fucopyranosyl derivative; when ally1 CL-L-fucopyranoside is used for copolymerization n = 0 Subsequently these preparations have been used in conjunction with normal acrylamide gel electrophoresis.For instance the crude haemagglutinating protein fraction from pea seeds was applied to polyacrylamide and to gels consisting of normal polyacrylamide covered by a layer of 0-a-D-galactosyl- and 0-a-D-mannosyl-polyacrylamide, respectively. The resolutions were identical on the first two gels but on the third gel one band had lost its mobility owing to its specific affinity for mar~nose.~' 46 V.Horejsi and J. Kocourek Biochim. Biophys. Acta 1973 297 346. " V. Horejsi and J. Koeourek Biochim. Biophys. Acta 1974 336 338. 72 H. Guilford Resolution of Optical Isomers.-The differential binding of optical isomers by certain proteins has been exploited in a resolution of DL-tryptophan on immobilized bovine serum albumin which has affinity for the L-isomer only. D-Tryptophan eluted in the void volume whereas L-tryptophan was adsorbed but could be eluted with acetic acid.48 Adsorbent with Digestible Spacer.-Most immunoadsorbents used for fractiona- tion of lymphocyte populations according to cell-surface antigens suffer from the limitation that the antigen-specific cells cannot be recovered. This problem has been overcome by using an enzymically digestible bridge between matrix and ligand.Cells are applied to antiglobulin coupled with glutaraldehyde to gelatin insolubilized on agarose. The B lymphocytes [i.e.those concerned with antibody production) adsorbed on the column were released by collagenase digestion of the gelatin spacer. More than 90 % of the eluted lymphocyte population carried surface immunoglobulins with full cell viability.49 Magnetic Matrices.-An aspect of affinity chromatography on which little has been published is that of its application to completely crude homogenates. A special problem is that even should the component of interest be taken up specifically on a tailor-made affinity gel its physical separation from cell debris and viscous cellular components can be prohibitively difficult.An ingenious technique has now been introduced which may facilitate such separations in batch adsorption-desorption processes. Magnetic polyacrylamide gels can be prepared by polymerizing the usual monomeric ingredients in the presence of iron oxide. The biospecific ligand is then attached using conventional techniques e.g. the L-asparaginase inhibitor D-asparagine was coupled to magnetic amino- ethylpolyacrylamide using butane- 1,4-diol diglycidyl ether. This adsorbent was incubated with a crude cell homogenate of Escherichia coli from which the gel particles could subsequently be separated by magnetic sedimentation. That the preparation is an effective biospecific adsorbent was shown by releasing L-asparaginase with a pulse of ~-asparagine.” This technique may well simplify in particular large-scale biospecific purifications.7 Molecular Probes Use of immobilized biospecific ligands as ‘molecular probes’ provides one of the most fascinating aspects of this field of investigation. Ligands are usually spaced from the matrix with arms long enough to ensure optimum steric availability. In a study of the binding of cobalamin to hog intrinsic factor (the protein con- cerned with B, adsorption and transport across the ileum wall) a short spacer was chosen deliberately. The reaction of cobfr)alamin with bromoethylamino- agarose affords a BI2analogue immobilized so that the cobalt atom is no more than 5 %t from the matrik Intrinsic factor is strongly bound tothe cobalamin gel 48 K.K. Stewart and R. F. Doherty 9th F.E.B.S. Internat. Biochem. Symposium Stock- holm 1973 Abs. la 7. 49 D. B. Thomas and B. Phillips European J. Immunol. 1973 3 740. P. Dunnill and M. D. Lilly Biotechnol. and Bioeng. 1974 16 987. Physical Methods-Part (ivb)A&%ity Chromatography 73 indicating that since the Stokes radius of the intrinsic factor-B, complex is 32.8 A,the binding site is at ordose to the protein’s surface.” Much interest has been focused on the glycoprotein interferon because of its involvement in inducing antiviral activity against a broad spectrum of viruses. Studies on human interferon have been hindered by the crudeness of available material-a litre of Ieucocyte preparation contains about ten micrograms of interferon.Anfinsen and co-workers have developed a biospecific purification leading to much purer enriched material.52 Some insight has been gained into the mode of induction of resistance to viral infection by examining the effect of mouse interferon coupled to agarose on mouse L-cells which are several times smaller than the gel beads. Retention of biological activity was observed suggesting that interferon operates without entering the cell probably by binding to a specific receptor on the cell ~urface.’~ The rabbit muscle enzyme phosphorylase b which catalyses the reaction glycogen + glucose-1 -phosphate * glycogen-glucose + Pi is totally inactive in the absence of the effector AMP (or structural analogues). Mosbach and Gestrelius have capitalized on this fact by succeeding in ‘freezing’ the enzyme in its active c~nformation.’~ It was first shown that when phosphory- lase b is coupled to CNBr-activated agarose or to glass ca.30 % of its activity is retained on adding AMFto a suspension of the immobilized enzyme. Also N6-(6-aminohexyl)-AMP (4 ;n = 6)can replace AMP as activator either free in solution (ca.80% as effective as AMP) or when coupled to soluble dextran (ca. 30 % as effective). This analogue is also ca. 20 % as effective as AMP when immobilized on agarose. Continuous cycling of a mixture of enzyme and sub- strates led to a linear release of phosphate over thirty minutes. Finally the enzyme was applied to a column of N6-(6-aminohexyl)-AMP-agarose,unbound enzyme was washed away and the substrates alone were cycled through the column.Continuous liberation of phosphate proved that phosphorylase b was ‘frozen’ in its active conformation on its immobilized activator AMP. A similar effect was observed when the enzyme and N6-(6-aminohexyl)-AMP were simul- taneously coupled to agarose such that enzyme and activator were fixed in sufficiently close proximity to one another to provide a permanently active phosphorylase b. The activity is not optimal since it can be increased by addition of AMP showing that only a proportion of the enzyme activator sites are orien- tated towards the immobilized AMP. Unfortunately when phosphorylase b was coupled in the presence of AMP which was then washed away no activity was observed unless more AMP was added.5‘ E. L. Lien L. Ellenbogen P. Y. Law and J. M. Wood J. Biol. Chem. 1974 249 890. 52 C. B. Anfinsen S. Bose L. Corley and D. Gurari-Rotman Proc. Nut. Acud. Sci. U.S.A. 1974 71 3139. 53 E. Knight jun. Biochem. Biophys. Res. Comm. 1974,56 860. 54 K. Mosbach and S. Gestrelius F.E.B.S. Lerters 1974 42 200. H. Guivord DNA coexists as chromatin in the cellular nucleoplasm with histones and other proteins. Although nucleotide-protein interactions have evoked much interest protein-protein interactions are much less investigated. Histone-histone binding has now been studied by fractionating calf thymus histones on a column of the arginine-rich histone H,-agarose. The most lysine-rich histone H,is not retarded histones H, and H, are eluted at increased ionic strength but histone H4 could only be retrieved with HCl confirming earlier speculation that self-interaction of histones is the strongest.” 8 Physical Parameters in Afinity Chromatography Wankat has produceds6 a theoretical analysis of affinity chromatography (c$ ref.14). Studies of the effects of other physical and kinetic parameters in affinity chromatography have shown that column geometry ligand concentration and total ligand must be balanced carefully to achieve optimum res~lts.’~ The affinity of glycerokinase and yeast alcohol dehydrogenase for immobilized N6-(6-aminohexyl)-AMP decreases with increasing temperature. ’* Glutamate synthetase however has affinity for AMP only at 50°C or above at which temperature it sticks to immobilized (4;n = 6).Elution is achieved by cooling the column. J. Mizon M. Mayne-D’Haultfoeuille C. Mizon-Capron P. Sautiere and G. Biserte F.E.B.S. Letters 1974 47 125. 56 P. C. Wankat Analyr. Chem. 1974 46 1400. 57 C. R. Lowe M. J. Harvey and P. D. G. Dean European J. Biochem. 1974 41 341. 58 C. R. Lowe M.J. Harvey and P. D. G. Dean European J. Biochem. 1974 41 353. 59 R. P. Bywater Biochem. Soc. Symposium on Affinity Chromatography Galway 1974.