Biochem. J. (1974) 141, 555-565Printed in Great Britain

Affinity Chromatography of Bovine TrypsinA RAPID SEPARATION OF BOVINE a- AND ,B-TRYPSIN

By G. W. JAMESON and D. T. ELMOREDepartment ofBiochemistry, Medical Biology Centre, Queen's University,

Belfast BT9 7BL, N. Ireland, U.K.

(Received 19 November 1973)

Affinity adsorbents for bovine trypsin were prepared by covalently couplingp-(p'-amino-phenoxypropoxy)benzamidine to cellulose and to agarose. Trypsin binds to bothadsorbents at pH 6-8 and is released at low pH values or in the presence of n-butylaminehydrochloride. Pure fl-trypsin may be eluted from crude trypsin bound at pH 8.0to the cellulose adsorbent by stepwise elution with an acetate buffer, pH 5.0. Botha- and fl-trypsin may be isolated by chromatography of crude trypsin on the agarosederivative in an acetate buffer, pH4.0. These two methods for purifying the trypsin arespecific to the particular adsorbents. They are rapid and convenient in use. Both methodsleave a mixture of the two enzymes bound to the adsorbent and release occurs only atlow pH values. The effects of pH, composition and ionic strength of buffer and othervariables on both purification methods are described. Affinity adsorbents of soya-beantrypsin inhibitor and of N-a-(N'-methyl-N'-sulphanilyl)sulphanilylagmatine bound toagarose were prepared, but were found to be of limited usefulness in the purification oftrypsin.

An adsorbent for the affinity chromatographyof enzymes is prepared by covalently attaching anenzyme-specific ligand to an insoluble support.Separation ofan enzyme from a mixture is commonlyachieved by applying a crude enzyme preparation tothe column in a buffer chosen so that the desiredenzyme binds firmly to the adsorbent while otherproteins are unretarded on elution. The boundenzyme is then commonly eluted in a single sharppeak by an abrupt buffer change (Cuatrecasas &Anfinsen, 1971a,b; Friedberg, 1971). This techniquehas been applied to the separation of trypsin frominert protein by using both small and large ligands(Feinstein, 1970a,b; Ureni, 1971; Kasai & Ishii, 1972;Hixson & Nishikawa, 1973). Bovine trypsin,however, exists in two major forms, a-trypsinand IJ-trypsin, which have been isolated by chromato-graphy on sulphoethyl-Sephadex (Schroeder &Shaw, 1968; Berezin et al., 1969). ,B-Trypsin has asingle polypeptide chain, whereas a-trypsin has twochains resulting from an autocatalytic cleavage ofthe Lys(13 O-Ser('32) bond. The two enzymes exhibitdifferent kinetic characteristics in their reactionswith some synthetic substrates (Hruska et al., 1969;Roberts et al., 1971 ; Jameson et al., 1973; Roberts &Elmore, 1974). We report the separation ofa- and f-trypsin by two simple and rapid chromato-graphic procedures involving no gradients, and thefactors that affect the separations have been investi-gated. A preliminary account of this work has beenpublished (Jameson & Elmore, 1971). Since this work

Vol. 141

was completed, a separation of the two forms oftrypsin by gradient elution from chicken ovomucoidinhibitor linked to Sepharose has been reported(Robinson et al., 1971).

Experimental

Materials

Bovine trypsin (thrice-crystallized, salt-free, freeze-dried) and bovine a-chymotrypsin (thrice-crystal-lized, salt-free, freeze-dried) were purchased fromWorthington Biochemical Corp., Freehold, N.J.,U.S.A. Soya-bean trypsin inhibitor (five timescrystallized) was obtained from Mann ResearchLaboratories, New York, N.Y., U.S.A. p'-Nitro-phenyl p-guanidinobenzoate hydrochloride was syn-thesized by the method of Chase & Shaw (1967)and the corresponding 4-methylumbelliferyl esterhydrochloride was prepared as described by Jamesonet al. (1973). S-(2-Aminoethyl)-N-benzoyl-L-cysteinemethyl ester hydrochloride was synthesized by themethod of Elmore et al. (1967) and 4-methylumbelli-feryl p-(NNN-trimethylammonium)cinnamate chlor-ide by the method of Jameson et al. (1973). p-(p'-Aminophenoxypropoxy)benzamidine as the benzene-sulphonate was prepared by a method based on thatof Baker & Erickson (1968), but the yield of theintermediate p-hydroxybenzamidine (66%) wasalmost twice that reported by these workers.Dioxan (150ml for 0.1 mol of p-cyanophenol)

555

G. W. JAMESON AND D. T. ELMORE

replaced chloroform as solvent; reaction withethanol-HCl was allowed to proceed for 10 daysinstead of 18 h; unchanged p-cyanophenol wasremoved by distillation under reduced pressurebefore reaction with NH3. p-Hydroxybenzamidinewas isolated as the hydrochloride, m.p. 222-224°C[Partridge & Short (1947) quote m.p. 223-224°C],and the latter was then converted into the benzene-sulphonate, m.p. 183-186°C [Baker & Erickson(1968) quote m.p. 180-181°C; Partridge & Short(1947) quote m.p. 187°C]. The remainder of thepreparation of p-(p'-aminophenoxypropoxy)benz-amidine benzenesulphonate was carried out asdescribed by Baker & Erickson (1968) except thatthe hydrogenation of the p-nitro precursor wascarried out in the presence of 10% palladium oncharcoal at atmospheric pressure and the productwas crystallized twice from acetonitrile by treat-ment with activated charcoal. The product had m.p.180-182°C [Baker & Erickson (1968) quote m.p.173-174°C]. NN-Dimethylformamide was purifiedby fractional distillation from P205. Thionyl chloridewas purified by distillation from quinoline followedby distillation from linseed oil. Sepharose 4B andSepharose 2B were obtained from PharmaciaFine Chemicals, Uppsala, Sweden, and cellulosepowder (Whatman CC31) was purchased fromW. and R. Balston Ltd., Maidstone, Kent, U.K.Sodium N-benzyloxycarbonylsulphanilate. Sulphan-

ilic acid (44.3 g) was dissolved in water (I000ml) andallowed to react with benzyl chloroformate (47.0g)in ether (200ml) atpH 8 (in a pH-stat) with mechanicalstirring. The precipitated solid was filtered off andwashed with ether. Another crop was obtained byextracting the filtrate with ether and evaporating theaqueous phase under reduced pressure. The totalproduct (78%) was recrystallized from methanol;it did not melt below 300°C (Found: C, 50.8; H, 3.8.C14H12NNaO5S requires C, 51.1; H, 3.7%).

N-Benzyloxycarbonylsulphanilyl chloride. Drysodium N-benzyloxycarbonylsulphanilate (20.0g)was dissolved in dry NN-dimethylformamide (20ml),and dry benzene (200ml) followed by thionyl chloride(24.0g) were added. The mixture was stirred for 1 h,poured into an ice-water mixture (200ml) and thenneutralized with NaHCO3. The benzene layer waswashed with water (3 x 200ml), dried (with MgSO4)and evaporated to dryness under reduced pressure.The residual product (93%Y.) had m.p. 111-1 13°C,unchanged after recrystallization from benzene-light petroleum (b.p. 60-800C) (Found: C, 51.5;H, 3.9; S, 10.0. C14H12C1N04S requires C, 51.6;H, 3.7; S, 9.8 %).Sodium N-(N'-benzyloxycarbonylsulphanilyl)-N-

methylsulphanilate. N-Methylsulphanilic acid (10.Og)was dissolved in a mixture of acetonitrile andwater (1:1, v/v) by neutralizing it with NaOH. Asolution of N-benzyloxycarbonylsulphanilyl chloride

(17.4g) in acetonitrile (30ml) was added and the reac-tion mixture was maintained at pH6 (with a pH-stat)with vigorous mechanical stirring. When the re-action was complete (about 3h), the mixture wasevaporated to dryness under reduced pressure andthe resulting solid was recrystallized twice frommethanol-ether with one treatment with activatedcharcoal and removal of material insoluble inmethanol. The resulting sodium salt (20.6g) was apale-brown solid. One sample was converted intothe potassium salt through the free acid. Thepotassium salt was recrystallized from NN-dimethyl-formamide-ether and then from water. It was driedin vacuo, when it had m.p. 175-178°C (decomp.)and was homogeneous by t.l.c. on silica in methanolcontaining 1% (v/v) acetic acid (Found: C, 43.9;H, 4.5. C21H19KN207S2,3H2O requires C, 44.3;H, 4.4%). Another sample of the sodium salt wasconverted into the dicyclohexylammonium salt,which was recrystallized twice from methanol-etherand then had m.p. 187-188°C and was homogeneousby t.l.c. on silica in methanol containing I %(v/v) acetic acid (detected by 12 vapour) (Found:C, 59.9; H, 6.8. C33H43N307S2 requires C, 60.2;H, 6.6%).N-(N'- Benzyloxycarbonylsulphanilyl) - N- methyl-

sulphanilyl chloride. Sodium N-(N'-benzyloxycar-bonylsulphanilyl)-N-methylsulphanilate (10.Og) wassuspended in a mixture of dry NN-dimethylform-amide (20ml) and dry benzene (200ml) and treatedwith thionyl chloride (20.0g). The mixture wasstirred for 4h and poured into a slurry ofice and water(750ml). The aqueous solution was neutralized withNaHCO3, more benzene (100ml) was added and theupper phase was separated. The benzene solutionwas washed with water (3 x 300ml), dried (withMgSO4) and evaporated under reduced pressureto a small volume. Addition of light petroleum(b.p. 60-80'C) gave the product (82%), which hadm.p. 147-148°C after recrystallization from benzene-light petroleum (b.p. 60-80°C) and was shown to behomogeneous by t.l.c. on silica in benzene (Found:C, 51.9; H, 3.9; N, 5.6. C21H19CIN206S2 requiresC, 51.0; H, 3.9; N, 5.7%).N- a - [N'-(N- Benzyloxycarbonylsulphanilyl) -N'-

methylsulphanilylJagmatine hydrochloride. Agmatinesulphate was converted into the dihydrochloride bypassage through a column of Dowex 1 (X-4; Cl1form) ion-exchange resin (Bio-Rad Laboratories,Richmond, Calif., U.S.A.). N-(N'-Benzyloxy-carbonylsulphanilyl)-N-methylsulphanilyl chloride(1.48 g) dissolved in acetonitrile (30ml) was added to astirred solution of agmatine dihydrochloride (609mg)in water (30ml) and the apparent pH was kept at9.5 by addition of 2M-NaOH (in a pH-stat). Whenreaction ceased (30min), the solution was evaporatedunder reduced pressure, yielding a gum which wasthen extracted with warm propan-2-ol. Evaporation

1974

556

SEPARATION OF a- AND ,B-TRYPSIN BY AFFINITY CHROMATOGRAPHY

of the filtered extract under reduced pressure gavethe product (93%/) as an amorphous white solid.It had an ill-defined melting point and decomposedin the range 100-180°C depending on the rate ofheating. The homogeneity of the material wasdemonstrated by t.l.c. on silica in chloroform-methanol-water (11:8:2, by vol.) and in pyridine-2-methylbutan-2-ol-water (7:7:6, by vol.) (detectedby I2 vapour) (Found: C, 48.9; H, 5.3; N, 12.9.C26H33CIN606S2,H20 requires C, 48.6; H, 5.5;N, 13.0%).N- a-(N'-Methyl - N'-sulphanilyl)sulphanilylagma-

tine dihydrochloride. N-a-[N'-(N-Benzyloxycar-bonylsulphanilyl) - N'- methylsulphanilyl]agmatinehydrochloride was hydrogenolysed at atmosphericpressure in ethanol in the presence of 10% palladiumon charcoal. The filtered solution was acidified withconc. HCI and evaporated to dryness under reducedpressure, giving the product (96%) which wasdried in vacuo to an amorphous solid with apoorly defined m.p. of about 106°C (decomp.). Itresisted attempts at crystallization but was shownto be essentially homogeneous by t.l.c. on silicain methanol-chloroform-water (11:8:2, by vol.)(Found: C, 41.0; H, 5.5; S, 12.1. C18H28C12N604S2requires C, 41.0; H, 5.4; S, 12.2%). Addition of anexcess of ethanolic 1,4-dihydroxy-9,10-anthraquin-one-2-sulphonic acid (rhodizonic acid) to anethanolic solution of the dihydrochloride causedthe rhodizonate salt to crystallize. This was washedwith ethanol and recrystallized from 2-methoxy-ethanol-acetone, when it had an m.p. of about192°C (decomp.), which was dependent on the rateof heating (Found: C, 49.8; H, 4.6; S, 11.5.C32H3MN6011S3 requires C, 49.5; H, 4.6; S, 12.4 %).p -(p' - Aminophenoxypropoxy)benzamidine- Seph -

arose and N-a-(N'-methyl-N'-sulphanilyl)sulphanilyl-agmatine-Sepharose conjugates. p-(p'-Aminophen-oxypropoxy)benzamidine and N-a-(N'-methyl-N'-sulphanilyl)sulphanilylagmatine were coupled co-valently to Sepharose by a procedure based on that ofCuatrecasas et al. (1968). In a typical preparation,Sepharose 4B (50ml of settled gel) was stirred withwater (SOml) and CNBr (5g) was added. The pH waskept at 11 by addition of 4M-NaOH (in a pH-stat)and a little ice was added to prevent a markedrise in temperature. When alkali uptake ceased, thegel was rapidly filtered off, washed with coldO.lM-NaHCO3 (lOOOml) and suspended in O.1M-sodium acetate buffer (50ml) adjusted to pH 5.0 withacetic acid. A solution of p-(p'-aminophenoxy-propoxy)benzamidine benzenesulphonate (22.2mg) inaq. 50% (v/v) methanol (5ml) was added andthe niixture was gently agitated at room temperaturefor 24h. The gel was filtered off, washed with moreacetate buffer, and the content of p-(p'-aminophen-oxypropoxy)benzamidine in the filtrate was deter-mined spectrophotometrically at 263nm. The gelVol. 141

was washed slowly and successively in a shortcolumn with 1 M-sodium acetate-acetic acid, pH5.0(lOOOml), 0.2M-NH4Cl-NH3, pH9.5 (lOOOml), andwater (lOOOml). It was stored as a slurry in 0.02%(w/v) NaN3 at 4°C. Coupling of the benzamidinederivative to Sepharose 2B and to Sephadex G-200was carried out in a similar manner, but the lattershrank considerably and acquired poor flowcharacteristics. Coupling experiments were carriedout with smaller quantities of CNBr and ofp-(p'-aminophenoxypropoxy)benzamidine benzene-sulphonate. Normally, over 90% of the addedbenzamidine derivative coupled to the gel exceptwhen less than 20mg of CNBr/ml of gel was used inthe activation step (Table 1). Coupling of N-a-(N'-sulphanilyl)sulphanilylagmatine to Sepharose 4Bwas performed similarly. The acetate buffer used forthe reaction between N-a-(N'-methyl-N'-sulphanilyl)-sulphanilylagmatine(lOmg)andSepharose4B(lOml),which had been activated with CNBr (1 g), usually hadapH of4.5, but similarresults were obtained atpH 3.0.Unchanged agmatine derivative was determinedspectrophotometrically at 270nm and the resultsshowed that 94-95% of it had become attached co-valently to the gel. A test for the presence of primaryaromatic amino groups, which was carried out on theproducts by attempted diazotization and couplingwith 2-naphthol, was negative. When the reactionbetween activated Sepharose and the agmatinederivative was carried out in NaHCO3 buffer atpH9.0, however, a positive test for primary aromaticamino groups was obtained, indicating that in thiscasesome coupling had occurred on the sulphonamidoor guanidino N atoms.p - (p' - Aminophenoxypropoxy)benzamidine - cellu-

lose conjugate. Cellulose powder (Whatman CC31;30g) was stirred with 0.5M-NaOH (1200ml) for 1 h,washed with water until the washings were no longeralkaline, stirred with 0.5M-HCl for 1 h and rewashedwith water. Fines were removed and the remainingcellulose (66ml packed vol.) was activated withCNBr (2.5g), washed and allowed to react with thebenzamidine benzenesulphonate (22.2mg) in 0.1M-sodium acetate buffer, pH5.0. The product waswashed as described for the Sepharose derivatives;96% of the benzamidine derivative had becomeattached covalently to the cellulose.

Soya-bean trypsin inhibitor-Sepharose conjugate.Sepharose 4B (39ml of settled gel) was activated withCNBr (4.0g), washed with cold 0.1M-NaHCO3(800ml) and suspended in 0.1M-NaHCO3 buffer,pH9.0. A solution of soya-bean trypsin inhibitor(176mg) in water (4ml) was added and the mixturewas gently stirred overnight at 4°C. After the productwas washed, spectrophotometric assay of thewashings at 280nm showed that the gel contained2.9mg of trypsin inhibitor/ml (65% coupling effi-ciency).

557

G. W. JAMESON AND D. T. ELMORE

Methods

Solutions of trypsin were kept at 0-40C; stockenzyme solutions for kinetic measurements weremade up in 1 mM-HCl containing 20mM-CaCI2.Measurements ofpH weremade atroom temperature.Concentrations of buffer constituents are given asfinal concentrations.

Column chromatography. Columns were main-tained at 4'C. Eluent was delivered by means ofa peristaltic pump and eluent gradients were linear.The column effluent was monitored at 280nm with aUvicord u.v. analyser (LKB-Produkter AB, Stock-holm, Sweden) and checked manually in selectedcases on a spectrophotometer. The effluent wasdivided into timed fractions and delivered intotest tubes maintained at 4°C in a water bath.Fractions that were to be freeze-dried were con-centrated where necessary by ultrafiltration at 0°Cthrough a Diaflo UM-10 membrane (AmiconCorp., Lexington, Mass., U.S.A.) and desalted bygel-filtration chromatography in 1 mM-HCl by usinga short column of Sephadex G-25.Measurement of protein concentrations. Trypsin

concentrations were obtained from measurements ofthe extinction at 280nm; values of Ej,1,/ andmolecular weight were taken as 15.6 (Baines et al.,1964) and 23 800 (Cunningham, 1954) respectively.

Determination of the operational molarity oftrypsin solutions. In early experiments, the totalcontent of active (a- and fl-)trypsin in solutions wasdetermined by titration with p'-nitrophenyl p-guanidinobenzoate at pH8.3 (Chase & Shaw, 1967).In addition, titrations were performed at pH4.0(Hruska et al., 1969) in attempts to determinethe relative proportions of a- and f-trypsin. In mostcases, however, trypsin solutionswere titrated spectro-fluorimetrically with 4-methylumbelliferyl p-guani-dinobenzoate at pH8.3 and 25°C (Jameson et al.,1973). As previously reported, IJ-trypsin was acylatedby the latter reagent in a few seconds, but the reactionwith a-trypsin required over IOmin and this differencewas exploited for the determination of the amountsof each form of enzyme present. Although theacylation of a-trypsin by 4-methylumbelliferylp-guanidinobenzoate obeyed pseudo-first-order kin-etics fairly well under the conditions used, it waspreferred to calculate the amount of a-trypsinpresent by a curve-fitting procedure which requiredno assumptions about the order of the reaction.Fluorescence-intensity readings were taken at inter-vals varying from a few seconds to several minutesdepending on the rate of change of fluorescenceintensity. After the first few seconds, changes influorescence intensity (corrected for the non-enzymichydrolysis of 4-methylumbelliferyl p-guanidinobenz-oate) were due only to the acylation of ac-trypsin.A computed curve was fitted to these readings by

the method involving orthogonal polynomials byusing the program described by Elmore et al. (1963)and since rewritten in FORTRAN for the ICL 1907computer. The value of the fluorescence intensity atzero time obtained by back-extrapolation of thecomputed curve was taken as a measure of thecontent of,I-trypsin in the sample, and the content ofa-trypsin could then be obtained from the differencein fluorescence intensity at zero time and whenacylation was complete.

Determination of the operational molarity ofa-chymotrypsin solutions. Solutions of a-chymo-trypsin were titrated with 4-methylumbelliferylp-(NNN-trimethylammonium)cinnamateas describedby Jameson et al. (1973).Measurement of the trypsin-binding capacity of

affinity adsorbents. A known excess (2-4-fold) ofbovine trypsin was applied to a small column(about 0.9cmx3cm) in a pH8.0 buffer, usuallyof 0.05M-triethanolamine hydrochloride-triethanol-amine containing 10mM-CaCI2. The column waswashed with this buffer until elution of protein hadessentially ceased; the concentration of trypsin inthe pooled effluent was determined by measuringE280 and by spectrofluorimetric titration with4-methylumbelliferyl p-guanidinobenzoate. Theamount of trypsin bound by the affinity adsorbentcould then be calculated. The bound enzyme waseluted with a solution of 0.5M-n-butylamine hydro-chloride in the triethanolamine buffer or alternativelywith a solution of 0.1 M-citric acid containing 1 M-NaCl. The trypsin content of the pooled effluentwas determined as above giving the trypsin-bindingcapacity of the column directly. Four values of thetrypsin-binding capacity were thus obtained.Although serious discrepancies between the fourvalues were observed only when a very largeexcess of trypsin was used, it was consideredthat the spectrophotometric determination of theprotein content of the effluent obtained in thesecond step was most reliable because (a) the spectro-photometric method applied to the first effluentwould overestimate the amount of unboundtrypsin if inert protein were present and (b) titrationof either fraction with 4-methylumbelliferyl p-guanidinobenzoate would underestimate trypsinif any autolysis occurred during the binding andelution procedures.

Identification of N-terminal amino acids. Dansyla-tion of trypsin samples was carried out in0.1M-NaHCO3 and in 0.25M-NaHCO3 containing4M-urea as described by Gray (1967) and Hartley(1970). The solvent in both cases contained 50% (v/v)acetone. In the latter case, the solution of dansylatedprotein was desalted by gel filtration beforehydrolysis. Dansyl-amino acids in the acid hydro-lysates were separated by t.l.c. on polyamide layers asdescribed by Hartley (1970).

1974

S58

SEPARATION OF a- AND ,B-TRYPSIN BY AFFINITY CHROMATOGRAPHY

Steady-state kinetic measurements. These werecarried out with a recording pH-stat by using S-(2-aminoethyl)-N-benzoyl-L-cysteine methyl esterhydrochloride and a-N-benzoyl-L-arginine ethylester hydrochloride as substrates. Detailed descrip-tions of the experimental and computational methodshave been given elsewhere (Elmore et al., 1963;Baines et al., 1964).

ResultsAffinity adsorbentsThe insoluble support chosen for most of the

affinity adsorbents was agarose (Cuatrecasas et al.,1968). Cellulose has also been found to be satis-factory, an observation independently reported byLowe & Dean (1971) and Uren (1971). The structuresof synthetic inhibitors were devised so that theenzyme-specific moieties were held well away fromthe supporting matrix (Cuatrecasas, 1970). p-(p'-Aminophenoxypropoxy)benzamidine was selectedas a compound structurally related to trypsininhibitors with K1 values in the range 2-8 uM (Baker &Erickson, 1968). The spacing group comprises anaromatic ring and a flexible ether-linked aliphaticchain. The successful use of affinity adsorbents basedon p-(p'-aminophenoxypropoxy)benzamidine mustbe due to these structural features and/or their lowKg values. Table 1 lists the activation conditions,coupling yields, content of inhibitor and trypsin-binding capacity for affinity adsorbents containingthis inhibitor. N-a-(N'-Methyl-N'-sulphanilyl)sul-phanilylagmatine was chosen as an analogue of theknown trypsin inhibitor, a-N-toluene-p-sulphonyl-agmatine (Rule & Lorand, 1964) for which the Kgvalue is of the order of 1 mm. The two sulphanilylgroups provide a linear rigid arm by which theagmatine moiety is linked to the matrix. Soya-beantrypsin inhibitor was chosen as a representative of the

natural trypsin inhibitors and the addition of a spacerwas considered unnecessary.

Chromatographic behaviour of soya-bean trypsininhibitor-Sepharose conjugate

Bovine serum albumin was only slightly retardedwhen applied to a column of soya-bean trypsininhibitor-Sepharose conjugate in 50mM-Tris-HClbuffer containing 10mM-CaCI2, pH7.8, but bovinetrypsin was bound to the adsorbent in this bufferand was subsequently eluted with 1 M-n-butylaminehydrochloride, a competitive inhibitor of trypsin(K1 = 1.7mM; Inagami, 1964), at pH 3.0. A saturationexperiment showed that the trypsin-binding capacityof the adsorbent, as measured by titration of theeffluent with p'-nitrophenyl p-guanidinobenzoate,was 2.4mg of trypsin/ml of gel. Bovine a-chymo-trypsin bound to the adsorbent in pH8.2 boratebuffer (lOmM-Na2B407, 2mM-CaCl2, 50mM-NaCI,adjusted with HCI) and also in pH6.0 10mM-sodiumcacodylate-HCI buffer containing 2mm-CaCl2. Inboth cases, the a-chymotrypsin was eluted with0.5M-n-butylamine hydrochloride in 0.1 M-aceticacid. These results resemble those reported byFeinstein (1970b). When chromatography wascarried out in 0.1 M-sodium formate-formic acidbuffer, pH 3.5, some selectivity for trypsin emerged;a-chymotrypsin was virtually unretarded, whereasmost ofthe trypsin activity was retarded and appearedas a very broad peak immediately after a peak ofinert protein. No evidence of the resolution of a-and fl-trypsin was obtained by gradient elution oftrypsin bound to the affinity adsorbent in pH7.850mM-Tris buffer containing lOmM-CaCl2. In theseexperiments, pH gradients (obtained by mixing theabove Tris buffer with an equal volume of 0.05M-sodium citrate buffer, pH3.5), concentration gradi-ents of n-butylamine hydrochloride (to 1 M at pH7.8)

Table 1. Characteristics of the affinity adsorbents derivedfrom p-(p'-aminophenoxypropoxy)benzamidine: relative amountsof CNBr used in activation of supports, efficiency of the inhibitor coupling reactions, inhibitor contents of the adsorbents

and their actual trypsin-binding capacities

For details see the text. Available trypsin-binding capacity is calculated as a percentage of a theoretical trypsin-bindingcapacity equivalent to the inhibitor content of the adsorbent.

CNBr usedin activation

Yield ofinhibitor-coupling

reactionIn

Support (mg/ml of support) (Y. coupled)Sepharose 4B 100 94Sepharose 4B 100 44*Sepharose 4B 20 91Sepharose 2B 5 65Cellulose powder 38 96

* The reason for this unexpectedly low value is not known.

Vol. 141

ihibitor contentadsorbents(pmol/ml)

0.9400.4350.1820.6470.717

Trypsin-bindingcapacity ofadsorbents

(mg of enzyme/ml of adsorbent)

3.461.260.52

1.4, 1.10.49

Availabletrypsin-binding

capacity ofadsorbents

(D)15.412.212.0

8.9, 7.12.8

559

G. W. JAMESON AND D. T. ELMORE

and a combined gradient were all used. These resultsprompted us to investigate the possible use of insolu-bilized synthetic trypsin inhibitors, although othernatural trypsin inhibitors have been shown toprovide useful affinity adsorbents for trypsin(Robinson et al., 1971; Feinstein, 1970a,b).

Chromatographic behaviour of N-c-(N'-methyl-N'-sulphanilyl)sulphanilylagmatine-Sepharose conjugate

This agmatine derivative exhibited more selec-tivity for trypsin than did the soya-bean trypsininhibitor-Sepharose conjugate. Bovine a-chymo-trypsin (5mg) applied to a column (0.6cmx7cm)of N-a-(N'-methyl-N'-sulphanilyl)sulphanilylagma-tine-Sepharose conjugate and eluted with a pH8.0Tris buffer (lOmM-Tris-HCI containing 2mM-CaCI2,I= 0.022mol/1) was only very slightly retarded,but bovine trypsin (5mg), chromatographed underthe same conditions, yielded two peaks of protein.The first was largely inert protein and contained only0.4mg ofactive trypsin as determined by titration withp'-nitrophenyl p-guanidinobenzoate. The secondpeak was very broad, with a long training edge, andcontained 1.2mg of trypsin. Elution with 0.1 M-aceticacid then yielded a third peak containing 0.6mg oftrypsin. Titration with p'-nitrophenyl p-guanidino-benzoate at pH4.0 showed that the last two peakscontained about the same proportions of a- andfi-trypsins. Similar results were obtained when thetrypsin was applied to the column in a pH 6.0 buffer(10mM-sodium cacodylate-HCI containing 2mM-CaCl2, I= 0.022mol/l) and again finally eluted with0.1 M-acetic acid. Replacement of the acetic acidelution step with a NaCl concentration gradientyielded all the enzyme remaining bound in a singlepeak (1.2mg of trypsin) at I=0.18 mol/l. Ionicstrength also markedly influenced the binding oftrypsin to the affinity column in three buffers atpH8.0. These buffers (lOmM-Tris, lOmM-Na2B407and 10mM-sodium barbiturate, adjusted to pH 8.0with HCI, H3BO3 and HCI respectively and all con-taining 2mM-CaCl2) were adjusted to the requiredionic strength with NaCl. For I,0.07mol/l, most ofthe trypsin was eluted in a single unretarded peak,whereas at I= 0.02mol/l, two peaks were obtainedand some trypsin remained bound as described above.N- a - (N'- Methyl - N'- sulphanilyl)sulphanilylag -

matine-Sepharose conjugate clearly functions as aspecific affinity adsorbent for bovine trypsin, but onlyat low ionic strengths. This undesirable property maybe due to the high K, values of agmatine derivativesor to the rigid linear structure of the inhibitor, orboth. The elution of two main peaks, which bothcontain a- and f8-trypsin, after the enzyme has beenapplied to the column in pH 6.0 or 8.0 buffers ofI= 0.02mol/l and eluted at low pH or high ionicstrength is an additional unsatisfactory feature. This

behaviour may stem from the linkage of the agmatinederivative to Sepharose by more than one type ofcovalent bond, effectively forming multiple types ofinsolubilized inhibitor. Supporting this theory, itis known that N-substituted imidocarbonates,isourea derivatives and N-substituted carbamatesare all formed when amines are coupled to CNBr-activated polysaccharides (Axen & Ernback, 1971).The additional possible complication that couplingmight have involved the guanidino atoms isconsideredunlikely, since chemical tests showed that theaffinity adsorbent did not contain free amino groups.

Chromatographic behaviour of p-(p'-aminophenoxy-propoxy)benzamidine-Sepharose conjugate

When crude trypsin was applied to a column ofthis affinity adsorbent in 50mM-triethanolaminehydrochloride-triethanolamine buffer, pH8.0, con-taining lOmM-CaCI2, the enzyme was bound anda peak of inactive protein was soon eluted (Fig. la).Subsequent elution with 0.2M-KCI containinglmM-HCl yielded only an insignificant quantity ofprotein. Development with lOmM-HCI containing0.5M-n-butylamine hydrochloride eluted the boundtrypsin in a single peak, which contained 86% ofactive enzyme and represented 80% of the originalenzyme applied to the column. Subsequent experi-ments showed n-butylamine hydrochloride (>0.25M)to be an effective eluent in the pH range 2.0-8.0.When the inhibitor was applied in a concentrationgradient in a pH8.0 50mM-triethanolamine hydro-chloride-triethanolamine buffer containing 10mM-CaCl2, the bound trypsin was eluted in a singlebroad peak when the concentration of n-butylaminehydrochloride was approx. 0.13M. Various finaleluents of pH2.0-2.5 eluted the bound trypsin aseffectively as solutions of n-butylamine hydro-chloride. These included: 0.1 M-citric acid-IM-NaCI,pH2.0; 0.1 M-formic acid, pH2.4; 0.2M-formic acid-sodium formate buffer containing 0.5M-NaCI,pH 2.5. Less predictably, the bound trypsin was alsoeluted by water, although continuously rather thanas a discrete peak. The elution by water of proteinsbound to affinity adsorbents is uncommon but notunknown (Vahlquist et al., 1971; Blumberg et al.,1970).The specificity of the affinity column was

demonstrated by its failure to bind a-chymotrypsinsignificantly in a pH 6.0 buffer (50mM-sodiumsuccinate-HCl containing lOmM-CaCI2) (Fig. Ib),or in a pH 8.0 buffer (50mM-triethanolamine hydro-chloride-triethanolamine containing lOmM-CaCI2).The content of active a-chymotrypsin in the firstpeak was 94% of the total enzyme applied.The trypsin-binding capacity of the affinity column

was usually 1-3mg/ml of gel (Table 1), but only asmall proportion (7-16 %) of the insolubilized

1974

360

SEPARATION OF a- AND fi-TRYPSIN BY AFFINITY CHROMATOGRAPHY

(a)

0.9 I

0

0.5

O.

0

0.E0

0

20 40 60

Fraction no.

Ii80 100

on p-(p'-aminophenoxypropoxy)benzamidine-Seph-arose conjugate is a useful means of ridding trypsinof other proteins, including a-chymotrypsin, and itcould be used to free a-chymotrypsin ofcontaminatingtrypsin.

I.0

04

- 0.5 8

0120 E,

Fraction no.

Fig. 1. Chromatography of bovine trypsin and a-chymo-trypsin on p-(p'-aminophenoxypropoxy)benzamidine-Seph-

arose conjugate

(a) Trypsin (20mg) was appliedtoacolumn (1 .Scmx 25cm)of affinity adsorbent in 50mM-triethanolamine hydro-chloride-triethanolamine buffer containing l0mM-CaCI2,pH 8.0. The eluent was changed to 1 mM-HCI containing0.2M-KCI at the first arrow and then to lOmM-HCl con-taining 0.5M-n-butylamine hydrochloride at the secondarrow. , E280; o, active trypsin concentration asmeasured by titration with 4-methylumbelliferyl p-guanidinobenzoate. Flow rate, 24ml/h; fraction size,6.Oml. (b) Chymotrypsin (20mg) was applied to thecolumn in 50mM-sodium succinate-HCI buffer con-taining l0mM-CaC12, pH6.0. The eluent was changed to50mM-sodium succinate-HCI containing 0.5M-NaCI,pH6.0, at the first arrow and then to 0.2M-formic acid-sodium formate containing 0.5M-NaCl, pH2.5, at thesecond arrow. Flow rate, 20ml/h; fraction size, 4.9ml.Symbols are as in (a).

inhibitor was actually available for enzyme binding.This phenomenon appears to be unrelated to the rela-tive amount of CNBr used in the activation processand is therefore unlikely to be due to a substantialdecrease in the gel pore size, which could arise ifcross-linking of the gel occurred. Chromatography

Vol. 141

Chromatographic behaviour of p-(p'-aminophenoxy-propoxy)benzamidine-cellulose conjugate

Trypsin bound strongly to this adsorbent in bufferscomposed either of 50mM-triethanolamine hydro-chloride-triethanolamine, pH8.0, or of 0.2M-NH4HCO3-NH3, pH8.4, both buffers containingl0mM-CaCl2. The binding of the enzyme wasunaffected by the inclusion of0.5M-NaCl in the formerbuffer. Completeelution ofthe bound enzymecould beeffected by solutions of low pH including: 0.1M-citric acid-I M-NaCl, pH2.0; 0.1 M-citric acid-0.1 M-NaCl, pH2.1; 20mM-HCI; 50mM-citric acid-50mM-NaCl, pH2.4. In contrast, although dilute formicacid easily eluted trypsin from p-(p'-aminophenoxy-propoxy)benzamidine-Sepharose conjugate, 0.3 M-formic acid was only partially effective when theenzyme was bound to the corresponding inhibitor-cellulose conjugate.Chromatography of bovine trypsin on p-(p'-

aminophenoxypropoxy)benzamidine-cellulose con-jugate under conditions resembling those describedin Fig. 1 yielded a very similar elution profile, and themain difference lies in the binding capacities of thetwo affinity adsorbents. p-(p'-Aminophenoxyprop-oxy)benzamidine-cellulose conjugate contained atleast as much inhibitor as did the correspondinginhibitor-Sepharose conjugate, but the proportionof coupled inhibitor available for binding trypsin wasmuch lower in the former case (Table 1). It is possiblethat binding of enzyme may be subject to serioussteric hindrance as a result of the impermeablecrystalline structure of cellulose.

Isolation of j9-trypsin by using p-(p'-aminophenoxy-propoxy)benzamidine-cellulose conjugate

Experiments with a pH gradient showed that itwas possible to elute pure ,B-trypsin specifically aftercrude trypsinhad been bound by the affinity adsorbent.The effects of eluent composition are important andare discussed below. From these early observations,a stepwise elution scheme was developed whichoffers a very convenient and rapid methodfor isolating bovine ,B-trypsin (Fig. 2). Crude bovinetrypsin (20mg) was applied to a column of p-(p'-aminophenoxypropoxy)benzamidine-cellulose con-jugate in 50mM-triethanolamine hydrochloride-triethanolamine buffer containing l0mM-CaCI2 atpH8.0. Elution with this buffer yielded a peak ofinert protein. An abrupt change of eluent to 50mM-sodium acetate-acetic acid, pH5.0, gave a peak of

561

1.R

0.4

G. W. JAMESON AND D. T. ELMORE

f-trypsin (5.1-5.3mg). Assay with 4-methylumbelli-feryl p-guanidinobenzoate showed that a-trypsinwas absent and that the ,B-trypsin was 90-94% active.Its identity with /8-trypsin, isolated by the method ofSchroeder & Shaw (1968), was confirmed by measure-ment of the Km and kcat. values by using S-(2-amino-ethyl)-N-benzoyl-L-cysteine methyl ester as thesubstrate (Table 2) and by demonstrating the presenceof only one N-terminal residue, isoleucine. After/i-trypsin had been removed from the column, afurther quantity of enzyme (5.2-5.3mg), which wasshown to contain 18% a-trypsin by titration with4-methylumbelliferyl p-guanidinobenzoate, waseluted with a solution of 50mM-citric acid containing50mM-NaCI, pH2.3.

Isolation of a- and Il-trypsin by using p-(p'-amino-phenoxypropoxy)benzamidine-Sepharose conjugate

The behaviour of trypsin on this affinity adsorbentdepends strongly on the composition of the eluent,and this is discussed below. A successful separationof a- and fl-trypsin is shown in Fig. 3. In thisexperiment, crude bovine trypsin (50mg) was appliedto a column (1.5cm x 25cm) ofp-(p'-aminophenoxy-propoxy)benzamidine-Sepharose 4B conjugate in50mM-sodium acetate-acetic acid buffer, pH4.0,and eluted at 55ml/h. The three incompletelyresolved peaks successively eluted were shown bytitration with 4-methylumbelliferyl p-guanidino-benzoate to consist of inert protein, a-trypsin andfi-trypsin respectively. The fraction containing/8-trypsin was very dilute and it was necessary toconcentrate the solution by ultrafiltration at 0°Cand to desalt it on Sephadex G-25 before it wasfreeze-dried. The yields of a- and fl-trypsin isolatedwere 3.8 and 12.4mg respectively and the enzymeswere obtained in states of purity (>97%) at least asgood as those attainable by chromatography onsulphoethyl-Sephadex (Schroeder & Shaw, 1968).The identity of our preparations with those obtainedby Schroeder & Shaw (1968) was again confirmedby measurement of Km and kcat. with S-(2-amino-ethyl)-N-benzoyl-L-cysteine methyl ester as substrate(Table 2). A firmly bound fraction ofenzyme (3.4mg)

was eluted with 0.1 M-citric acid containing 1 M-NaCland contained 20% a-trypsin.

This procedure conveniently gives both a- and ,B-trypsin, but the latter is obtained as a very dilutesolution; the method was therefore used only whena-trypsin was specifically sought. Since this form ofthe enzyme lost activity significantly in a few hoursat pH4.0 and 0°C, a high flow rate and rapid proces-sing of the effluent were necessary to obtain goodyields. fl-Trypsin was much more stable in solution.The same relative stabilities were observed withfreeze-dried samples of a- and IJ-trypsin prepared bythe method of Schroeder & Shaw (1968) and storedat -20°C for approx. 2 years. Titration with 4-methyl-umbelliferyl p-guanidinobenzoate showed that IJ-trypsin had retained much of its activity, but a-trypsin was virtually inactive.

0.4

000LW 0.2

-I

ci0Ua

'U>H

Fraction no.

Fig. 2. Isolation of bovine fi-trypsin by affinity chromato-graphy of crude trypsin on p-(p'-aminophenoxypropoxy)-

benzamidine-cellulose conjugate

Crude trypsin (20mg) was applied to a column (2.2cmx14.3 cm) of affinity adsorbent in 5OmM-triethanolaminehydrochloride-triethanolamine buffer containing 10mM-CaCl2, pH8.0. The eluent was changed to 50mM-sodiumacetate-acetic acid, pH 5.0, at the first arrow and to 50mM-citric acid containing 50mM-NaCI, pH2.3, at the secondarrow. , E280; o, active trypsin concentration asmeasured by titration with 4-methylumbelliferyl p-guanidinobenzoate. Flow rate, 19ml/h; fraction size,5.Oml.

Table 2. kca,. and Km values for the hydrolysis ofS-(,8-aminoethyl)-N-benzoyl-L-cysteine methyl ester atpH8.0 and 25.00C bya- and fi-trypsin isolatedboth by affinity chromatography andby ion-exchange chromatography on sulphoethyl-Sephadex

Values are means ±S.D.Enzyme Source*Trypsin Affinity chromatography

Ion-exchange chromatography*-Trypsin Affinity chromatography

Ion-exchange chromatography** Values taken from Roberts & Elmore (1974).

ac-

'9-,

kcat. (s-')10.3 ± 0.28.8±0.1

32.8 + 0.535.0+ 0.6

I x 104Km (M)0.77 + 0.031.10±0.041.07 + 0.051.04+0.04

No. of runs910912

1974

562

SEPARATION OF a- AND fJ-TRYPSIN BY AFFINITY CHROMATOGRAPHY

Effect ofelution conditions on the resolution of a- andfi-trypsin onp-(p'-aminophenoxypropoxy)benzamidine-cellulose and -Sepharose conjugates

The development of the foregoing successfulmethods for the isolation of a- and 8-trypsin was theculmination of an extensive study of differentelution conditions which revealed that the resolutionof the mixture of enzymes depended on factors suchas pH, ionic strength, nature of the buffer speciesand flow rate. The choice of support for theinsolubilized inhibitor is another variable. Theresults ofsome of our experiments are summarized inTables 3 and 4.

aIN1\ 100 _

so

0 so _

1 0.5 r

0.30

Discussion

The results of some of our experiments could nothave been predicted in advance and cannot befully explained in retrospect. For example, the orderofelution of a- and fi-trypsin from thep-(p'-phenoxy-propoxy) benzamidine-cellulose complex depends onthe nature of the eluting buffer (Fig. 2, Table 3). Someof the experiments summarized here indicate thatsuccessful separation of closely related proteins byaffinity chromatography is largely serendipitous in thepresent state of our knowledge. This situation mightbe modified in the light of the critical appraisal(O'Carra et al., 1974) of the current state of theart of affinity chromatography. It is probable thatthe behaviour of a- and fl-trypsin on the variousinhibitor conjugates is not due to bioaffinity alonebut involves compound affinity as defined byO'Carra et al. (1974). The predominantly hydro-

0.I

0 40 60Fraction no.

I

a8*R

Fig. 3. Separation of bovine a- and fi-trypsin by affinitychromatography on p-(p'-aminophenoxypropoxy)benzami-

dine-Sepharose conjugateCrude trypsin (50mg) was applied to a column (1.5cmx25cm) of affinity adsorbent in 50mM-sodium acetate-acetic acid buffer, pH4.0. The eluent was changed to0.1 M-citric acid containing 1 M-NaCl, pH2.0, at thearrow. , E280; o, active trypsin concentration asmeasured by titration with 4-methylumbelliferyl p-guanidinobenzoate; 0, percentage of a-trypsin in the totaltrypsin present. Flow rate, 55ml/h; fraction size, 9.2ml.

Table 3. Effect of elution conditions on the resolution of a- and ,-trypsin on p-(p'-aminophenoxypropoxy)benzamidine-cellulose conjugate

Loading buffer Eluting buffer Chromatographic result50mM-Triethanolamine 50mM-Sodium acetate-acetic acid, pH4.0 f-Trypsin contaminated with a-trypsinhydrochloride-triethanol-amine containing 1OmM-CaCl2, pH 8.0

50mM-Sodium cacodylate-HCI, pH5.8 Similar to Fig. 2, but yield of /1-trypsin was low5OmM-Ammonium acetate-acetic acid, pH 5.0 Very broad peak of fi-trypsinpH gradient generated by adding 50mM- 8-Trypsin peak contained traces of a-trypsinsodium acetate, pH4.0, to 50mM-sodium in trailing edge. Citric acid-NaCl gave a peakcacodylate-HCI, pH7.0 then 0.1 M-citric containing both a- and fi-trypsin.acid-i M-NaCl.

0.2M-Ammonium formate- Two gradients generated by adding 5OmM- Two broad peaks only partially resolved.NH3, pH 8.4 ammonium formate-formic acid, pH4.0, The first contained both a- and f-trypsin

to the loading buffer then adding 20mM- (pH4.1), the second contained J-trypsinHCI to the formate buffer, pH4.0. (pH3.9).

Vol. 141

563

564 G. W. JAMESON AND D. T. ELMORE

Table 4. Effect of elution conditions on the resolution of a- and fi-trypsin on p-(p'-phenoxypropoxy)benzamidine-Sepharoseconjugate

Loading buffer Eluting buffer Chromatographic result50mM-Sodium acetate-acetic Loading buffer Unresolved peak containing inert protein and

acid, pH3.5 a-trypsin with f-trypsin in long trailing edge50mM-Sodiumacetate-acetic Loading buffer Similar to Fig. 3

acid, pH4.050mM-Sodium acetate-acetic Loading buffer Similar to Fig. 3 but peaks were broader

acid, pH4.550mM-Sodium acetate-acetic Loading buffer Peaks were broadened to give a continuum

acid, pH 5.050mM-Sodium acetate-acetic Loading buffer Similar to that obtained with pH 5.0 buffer

acid, pH4.0 containing0.5M-NaCl

50M-Ammoniumformate- Loading buffer No discrete peak of f-trypsinformic acid, pH4.0

Loading buffer+ 1 mM-cetyltrimethyl Elution volume of a-trypsin decreasedammonium bromide

50mM-Sodium succinate-HCl, pH gradient generated by adding0.2M- Two incompletely resolved peaks were elutedpH 6.0, containing 10mM- formic acid-NaOH, pH2.5, to loading at pH3.0and pH2.8; the first peak containedCaCl2 buffer; both buffers contained 0.SM- fi-trypsin, the other contained a mixture of

NaCl a- and fl-trypsin

phobic character of the spacer arm may causesome of the subsites of the enzyme to be occupiedand bring part of the surface of the enzyme intojuxtaposition with the support thereby facilitatingadditional interactions by hydrogen bonding. Ifso, the differences in chromatographic behaviourof the inhibitor conjugate derived from cellulose andSepharose can be partially explained. Nevertheless,binding of trypsin presumably involves the activesite, since the enzyme can be displaced by thereversible inhibitor, n-butylamine.

'Firmly-bound' trypsin

This description is applied to trypsin whichremained bound to conjugates derived from p-(p'-aminophenoxypropoxy)benzamidine after elutionwith acetate buffers, pH4.0 or 5.0, or with 20mM-butylamine hydrochloride. The retention of some'firmly bound' trypsin appeared to be independent ofthe relative amount of CNBr used to activate thesupport and of the nature of the support. The'firmly bound' trypsin was eluted, however, withsolutions of citric acid, formic acid or HCI at pH2or with solutions of n-butylamine hydrochloride(>0.2M). It always proved to be a mixture of the twoforms of trypsin, containing 15-20% a-trypsin.All attempts to resolve the 'firmly bound' a- and,B-trypsin by gradient elution with both pHgradients and n-butylamine concentration gradientswere unsuccessful.

We are indebted to the Medical Research Council fora grant to G. W. J.

References

Ax6n, R. & Ernback, S. (1971) Eur. J. Biochem. 18,351-360

Baines, N. J., Baird, J. B. & Elmore, D. T. (1964) Biochem.J. 90, 470-476

Baker, B. R. & Erickson, E. H. (1968) J. Med. Chem. 11,245-249

Berezin, I. V., Kazanskaya, N. F. & Larionova, N. I.(1969) Biokhimiya 35, 983-988

Blumberg, S., Schechter, I. & Berger, A. (1970) Eur. J.Biochem. 15, 97-102

Chase, T. & Shaw, E. (1967) Biochem. Biophys. Res.Commun. 29, 508-514

Cuatrecasas, P. (1970) J. Biol. Chem. 245, 3059-3065Cuatrecasas, P. & Anfinsen, C. B. (1971a) Methods

Enzymol. 22, 345-378Cuatrecasas, P. & Anfinsen, C. B. (1971b) Annu. Rev.

Biochem. 40, 259-278Cuatrecasas, P., Wilchek, M. & Anfinsen, C. B. (1968)

Proc. Nat. Acad. Sci. U.S. 61, 636-643Cunningham, L. W. (1954) J. Biol. Chem. 211, 13-19Elmore, D. T., Kingston, A. E. & Shields, D. B. (1963)

J. Chem. Soc. London 2070-2078Elmore, D. T., Roberts, D. V. & Smyth, J. J. (1967)

Biochem. J. 102, 728-734Feinstein, G. (1970a) FEBS Lett. 7, 353-355Feinstein, G. (1970b) Biochim. Biophys. Acta 214, 224-

227Friedberg, F. (1971) Chromatogr. Rev. 14, 121-131

1974

SEPARATION OF a- AND fi-TRYPSIN BY AFFINITY CHROMATOGRAPHY 565

Gray, W. R. (1967) Methods Enzymol. 11, 469-475Hartley, B. S. (1970) Biochem. J. 119, 805-822Hixson, H. F. & Nishikawa, A. H. (1973) Arch. Biochem.

Biophys. 154, 501-509Hruska, J. F., Law, J. H. & K6zdy, F. J. (1969)

Biochem. Biophys. Res. Commun. 36, 272-277Inagami, T. (1964) J. Biol. Chem. 239, 787-791Jameson, G. W. & Elmore, D. T. (1971) Biochem. J.

124, 66PJameson, G. W., Roberts, D. V., Adams, R. W., Kyle,W. S. A. & Elmore, D. T. (1973) Biochem. J. 131,107-117

Kasai, K. & Ishii, S. (1972) J. Biochem. (Tokyo) 71,363-366

Lowe, C. R. & Dean, P. D. G. (1971) FEBS Lett. 14,313-316

O'Carra, P., Barry, S. & Griffin, T. (1974) MethodsEnzymol. in the press

Partridge, M. W. & Short, W. F. (1947) J. Chem. Soc.London 390-394

Roberts, D. V. & Elmore, D. T. (1974) Biochem. J. 141,545-554

Roberts, D. V., Adams, R. W., Elmore, D. T., Jameson,G. W. & Kyle, W. S. A. (1971) Biochem. J. 123, 41P-42P

Robinson, N. C., Tye, R. W., Neurath, H. & Walsh, K. A.(1971) Biochemistry 10, 2743-2747

Rule, N. G. & Lorand, L. (1964) Biochim. Biophys. Acta81, 130-135

Schroeder, D. D. & Shaw, E. (1968) J. Biol. Chem. 243,2943-2949

Uren, J. R. (1971) Biochim. Biophys. Acta 236, 67-73Vahlquist, A., Nilsson, S. F. &Peterson, P. A. (1971)Eur.J.

Biochem. 20, 160-168

Vol. 141