J. Biochena., 74, 497-503 (1973)

Limited Proteolysis of Human IgG by Fruit *

Glorietta P. VIDAL and Makoto SASAKI Department of Biochemistry, Nagoya City University School of Medicine, Nagoya

Received for publication, February 9, 1973

Fruit bromelain [EC 3.4.4.24] cleaved human IgG into two kinds of fragments separable by CM and DEAF-cellulose column chromatography. The resulting frag ments, Fab(F) and Fc(F), apparently resembled those obtained using [EC 3.4.4. 10] as regards antigenicity, electrophoretic mobility, sedimentation constant, molecular weight, and amino acid composition. Fab(F) and Fc(F) had sedimentation constants of 3.7S and 3.5S, respectively. Fab(F) was shown by gel filtration to have a molecular weight of 51,000 and Fc(F) a molecular weight of 48,000. These results suggest that the possible site of fruit bromelain cleavage of human IgG could be very close to that of papain. Studies of the cleavage by fruit bromelain of human IgG at different substrate ratios and digestion times showed that human IgG has a stronger resistance to fruit bromelain than to papain. By experiments on subclass specificity, it was shown that this difference in resistance to enzymatic attack was mainly due to the difference in the nature of the enzyme-resistant subclass. Thus, the fruit bromelain resistant component involves IgG 1 and IgG 2, whereas it has been shown that the papain-resistant component involves only IgG 2.

Using proteolytic in structural studies teases such as papain [EC 3.4.4.10] (4) and of immunoglobulins, some major aspects of D [EC 3.4. 4.231 (5) have been in IgG have been investigated. Most studies vestigated. Among these, papain and pepsin revealed that several proteolytic enzymes at played an important role for their well-defined tacked at a unique sensitive site on the heavy fragmentation patterns, which led to the chain. Enzymatic fragments of serine prote schematic structure of human IgG. ases such as trypsin [EC 3.4.4.4] (1) and Fruit bromelain [EC 3.4.4.24] is an acidic plasmin [EC 3.4.4.14] (2), acidic such protein belonging to the cysteine as pepsin [EC 3.4.4.1] (3), and cysteine pro isolated from pineapple fruit. The present report describes in detail the limited proteolysis * This work was presented at the 2nd Annual Meeting of human IgG by fruit bromelain at various of the Japanese Society of Immunology, December, 1972 in Tokyo. enzyme-substrate ratios and digestion times. Our point of interset was to determine how Abbreviations: Fab(P), Fc(P) , papain enzymatic f rag ments; Fab(F), Fab-like fragment obtained from fruit closely the fruit bromelain enzymatic frag bromelain digest ; Fc(F) , Fc-like fragment from fruit ments resembled those of papain. For this bromelain digest; FK.BRL, fruit bromelain. purpose, a comparative study of the cleavage

Vol. 74, No. 3, 1973 497 498 G.P. VIDAL and M. SASAKI

patterns of fruit bromelain and papain was Fc to purify Fab fragments. Anti-Fab and carried on throughout the analysis. anti-Fc antibodies used for the insoluble col umn were produced in our laboratory by im munizing rabbits with papain enzymatic Fab MATERIALS AND METHODS and Fc fragments. The insoluble columns Enzymes and Chemicals-Twice-crystal were prepared according to the method of lized papain was obtained from Worthington Axen et al. (8, 9). Biochemical Corp. Fruit bromelain, purified Immunological Studies Immunological by the method of Ota et al. (6), was kindly studies were carried out by double diffusion supplied by Dr. S. Ota. EDTA and cysteine in agar by the technique of Ouchterlony, and were purchased from Katayama Kagaku Kogyo immunoelectrophoresis according to the micro Co. Ltd. DEAE and CM-cellulose were from method of Scheidegger. Rabbit antiserum to Seikagaku Kogyo Company. Sephadex G-150 human IgG was prepared in our laboratory. was obtained from Pharmacia Fine Chemicals. Some anti-Fab and anti-Fc for identification

Anti-human IgG subclass antisera (swine) were purposes were obtained commercially from purchased from Nordic Diagnostics and some Behringwerke. (rabbit) were generously supplied by Dr.. T. Ultracentrifugal Analyses Ultracentrifu Tada of Chiba University. Standard samples gal analyses were carried out in a Hitachi for molecular weight determination were ob ultracentrifuge model UCA-1A, at 20•Ž at a tained from Boehringer Mannheim. speed of 55,430 rpm. Photographs were taken Purification of Human IgG-Human IgG at intervals of 12 min. The solvent was 0.2 M was purified from pooled serum by ammonium NaCl. sulfate precipitation at 40% saturation, fol Molecular Weight Determination-Molecu lowed by 33% saturation, repeated twice. lar weights were determined by means of The IgG fraction was applied to the DEAE gel filtration (Sephadex G-150, column ; 1 cm •~ - cellulose column (0.015 M Tris phosphate buffer 100 cm) following Andrew's method (10). pH 8.3). The eluate was dialyzed against The flow rate was regulated at 4 ml/hr, using 2 mM NaCl, then centrifuged and the super a peristaltic pump for the maintenance of a natant was lyophilyzed. This sample showed steady flow rate. Authentic samples of a single homogeneous precipitin line in im known molecular weight, such as egg-white munoelectrophoresis when tested against anti lysozyme, chymotrypsinogen A, egg and whole human serum. serum albumin, aldolase [EC 4.1.2.13], cata Enzymatic Digestion and Isolation of Cleav lase [EC 1.11.1.6], and human IgG were age Products-Human IgG was dissolved in chromatographed. Fab and Fc fragments were 0.1 M sodium phosphate buffer pH 7.6. With also chromatographed under the same condi continuous stirring and flushing of nitrogen tions. One milliliter sample solution (2-5%) gas, 2 mm of EDTA was added to the solution, was applied to the column. followed by 10 mm of cysteine (final concentra Amino Acid Analyses-Samples of 2 mg tion) and lastly the enzyme (4). The reaction were hydrolyzed in 0.5 ml of 6 N HCl (special mixture was incubated at 37•Ž for 5-7 hr, then grade) in a sealed tube at 110•Ž for 24 hr. The chromatographed on Sephadex G-150 (0.02 M hydrolysate was evaporated to dryness at 45•Ž borate-buffered saline, pH 8.0), which sepa in a rotary evaporator. Amino acid analyses rated the digested part from the undigested were carried out according to the method of material. The former was applied to a CM Spackman et al. (11) with a Hitachi amino - cellulose column followed by DEAE-cellulose acid analyzer, model KLA-3B. column chromatography to isolate each frag ment (7 ). Further purification was then per RESULTS formed by the use of immunoadsorbent col umns. These consist of insolubilized anti-Fab When human IgG was incubated with fruit to purify Fc fragments and insolubilized anti bromelain for 7 hr at 1 : 50 enzyme-substrate

J. Biochem. PROTEOLYSIS OF IgG BY FRUIT BROMELAIN 499

(E/S) ratio, about 50% of the whole IgG was resolved into two peaks. The first peak molecule was digested, while when human consisted of mainly Fab(F) but contained a IgG was incubated with papain for just 3 hr trace of Fc(F). The former was designated as at 1 : 100 E/S ratio, about 80% was digested, Fab II(F), according to its order of separation. as shown in Fig. 1. The second peak contained one homogeneous After elimination of the undigested part Fc(F) component. Further purification was on a Sephadex G-150 column, the digested part carried out using immunoadsorbent for com

was subjected to chromatography on a CM plete elimination of small amounts of con -cellulose column at pH 7,6 with stepwise elu taminants (for details, see " MATERIALS AND tion, and two major peaks were obtained (Fig. METHODS"). The amount of Fab II(F) finally

2). purified through the anti-Fc immunoadsorbent The first peak came out with the starting was so small that analyses were centered buffer (0.01 M sodium phosphate buffer pH 7.6), mainly on Fab I(F) and Fc(F). and the second peak was eluted shortly after Digestion Time-In order to determine the the start of the second buffer (0.01 M sodium best conditions for limited proteolysis of human phosphate buffer, pH 7.6, containing 0.3M IgG with fruit bromelain, digestion for dif NaCl). The results of immunoelectrophoretic ferent periods of time was performed. The analyses showed that the first peak contains reaction mixture at 1 : 100 E/S ratio was in two components ; Fab(F) and Fc(F), while the cubated for 7, 17, 30 hr at 37•Ž, resulting in 5 second peak contains one homogeneous Fab 50% yield of digested material, as shown in component, which was designated as Fab I(F). Fig. 3. Even after 30 hr, 50% remained re The first peak was then chromatographed on sistant to enzymatic attack. a DEAE-cellulose column with stepwise elution (starting buffer, 0.01 M sodium phosphate buf fer, pH 8.0; second buffer, 0.01 M sodium phos phate buffer, pH 8.0, containing 0.3 M NaCI). It

Fig. 2. Isolation of Fab and Fc fragments from

fruit bromelain digest by ion exchange column chromatography. The digested part was chromato

Fig. 1. Gel filtration pattern of fruit bromelain and graphed on a CM-cellulose column (2 •~ 26 cm) equi librated with 0.01M sodium phosphate buffer, pH papain digests on Sephadex G-150. A Sephadex 7.6. Fab 1 (F) comes out with the second buffer G-150 column (3.5 •~ 140 cm) was equilibrated with 0.02M borate-buffered saline pH 8.0. Fruit bro (0.01M sodium phosphate buffer, pH 7.6, containing melain digest (1 : 50 E/S ratio, 7 hr incubation at 0.3m NaCl). Peak 1 was pooled and chromato 37•Ž). Papain digest (1 : 100 E/S ratio, 3 hr incuba graphed on a DEAE-cellulose column (2 cm •~ 26 cm)

tion at 37•Ž). 20 ml of reaction mixture was applied. equilibrated with 0,01M sodium phosphate buffer, Flow rate ; 15 ml/hr. Fraction volume ; 20 ml each. pH 8.0. Fc(F) comes out with the second buffer

Peak 1 contains the undigested part and peak 2 (0.01 M sodium phosphate buffer, pH 8.0, containing contains the digested part. 0.3 M NaCl).

Vol. 74, No. 3, 1973 500 G.P. VIDAL and M. SASAKI

Fig. 3. Gel filtration pattern of fruit bromelain digest on Sephadex G-150 at different digestion times. Fruit bromelain digest (1 : 100 E/S ratio, incubated

for different periods of time at 37•Ž) was applied to Fig. 5. Repeated hydrolysis of the undigested part a Sephadex G-150 column (3.5 •~ 140 cm) equilibrated after fruit bromelain treatment. The fruit bro with 0.02 M borate-buffered saline pH 8.0. Gel filtra melain-resistant portion was digested under the same tion was carried out under exactly the same condi conditions as the first digestion (1 : 50 E/S ratio, 7 hr). tions as in Fig. 1. Chromatography was carried out on Sephadex G-150

(3.5 •~ 140 cm) equilibrated with 0.02 M borate-buffered saline pH 8.0. Gel filtration was performed under

the same conditions as in Fig. 1.

at 1 : 33 and 1 : 20 E/S ratios were found to show a tendency for further fragmentation, on the basis of their ion exchange patterns and electrophoretic mobilities. These results in dicated that the best conditions for limited proteolysis of human IgG by fruit bromelain involve a 1 : 50 E/S ratio with a digestion time of 7 hr. Digestion of the Fruit Bromelain-Resistant Fig. 4. Gel filtration pattern of fruit bromelain Portion-Treatment of the undigested portion digest on Sephadex G-150 at different E/S ratios. with fruit bromelain under the same conditions Fruit bromelain digest (1: 100, 1:50, 1:33, and as in the previous digestion resulted in about 1 : 20 E/S ratio, 7 hr incubation at 37•Ž) was applied 30% digestion of the material: around 70% to a Sephadex G-150 column (3.5 •~ 140 cm) equili resisted enzymatic attack, as shown in Fig. 5. brated with 0.02m borate-buffered saline pH 8.0. Gel The difference between the second diges filtration conditions are the same as in Fig. 1. tion and the first, led us to suspect that the digestion might involve subclass specificity. Enzyme-Substrate Ratio When human This point was clarified after testing with IgG was incubated with fruit bromelain at anti-subclass antiserum (Fig. 6). The fruit 1 : 100 E/S ratio for 7 hr, 5% of the material bromelain-resistant portion of the second digest was digested, but at 1 : 50 E/S ratio, about was tested with anti-subclass antisera (anti 50% was digested. No significiant increase in IgG 1, anti-IgG 2, anti-IgG 3, and anti-IgG 4), digestion was observed even when the E/S comparing its reaction with normal IgG and ratio was at 1 : 33 or 1 : 20, as shown in Fig. 4. myeloma IgG. The fruit bromelain-resistant Fc fragments from fruit bromelain digest portion gave positive results with anti-IgG 1

J. Biochem. PROTEOLYSIS OF IgG BY FRUIT BROMELAIN 501

Fig. 6. Immunodiffusion analysis of the fruit bro melain-resistant portion. The center well contains anti-subclass antisera (anti-IgG 1, anti-IgG 2 , anti -IgG 4). Outer wells (I, III) ; normal IgG. Outer well (II) ; fruit bromelain-resistant portion after the Fig. 7. Electrophoretic mobilities of fruit bromelain second digestion. Outer well (IV); myeloma IgG . digest. Electrophoresis was carried out with a vol The concentration of antigen was 5 mg/ml and the tage gradient of 6-8V/cm for 2 hr in 1.2% pure volume was 3 gel each. The amount of antisera agar, 0.03 M barbital buffer, pH 8.6. The plate was ranged from 10 to 15 ƒÊl. incubated for 16 hr at 37°C after application of the antisera.

and 2, but no precipitin line with anti-IgG 3 and 4 was noted. Normal IgG reacted with all the antisera, as expected. Myeloma IgG re acted only with anti-IgG 1. This result shows that two kinds of minor subclass components, IgG 3 and IgG 4, appear to be sensitive to fruit bromelain digestion, while IgG 1 and IgG 2 seem to be fruit bromelain resistant. Another experiment showed that when IgG 1 myeloma protein was digested with fruit bromelain, about 20-30% digestion was observed, whereas with papain about 90-95% digestion occurred. Electrophoretic Mobilities and Immuological Properties of the Fragments-Figure 7(A) shows

the electrophoretic mobilities of fruit bromelain Fig. 8. Immunological properties of fruit brome digest. The undigested part shows a mobility lain and papain cleavage fragments. Double im identical with that of native IgG, while the munodiffusion ; 1.2% pure agar, 0.02M borate buff digested part is composed of a fast and a ered saline pH 8.0. Incubation time ; 16 hr at 37•Ž. slow-moving components. The fast-moving Electrophoresis ; 2 hr, 6-8V/cm. Incubation time ; component is Fc(F), since it reacted with au 16 hr at 37•Ž. For sharper resolution 1.2% Agarose thentic anti-Fc(P) antiserum (Fig. 7(B)) and dissolved in 0.03M barbital buffer pH 8.6 was used the slow component is Fab(F), since it reacted as the supporting media. with authentic anti-Fab(P) antiserum. The antigenic properties of these frag identity. Fc from both kinds of digests also ments are shown in Fig . 8 (double immuno showed identical reactions. The mobilities of diffusion). Precipitin lines of Fab fragments Fab and Fc from both kinds of digests were from fruit bromelain and papain digests fused also found to be identical as shown in Fig. 8 completely, demonstrating immunological (immunoelectrophoresis).

Vol. 74, No. 3, 1973 502 G.P. VIDAL and M. SASAKI

TABLE I. Amino acid composition of Fab and Fc fragments. Samples of 2 mg were dissolved in 0.5 ml of 6 N HCl. After complete evacuation under 10-4

mmHg freezing and thawing to maintain a semi solid state), test tubes were sealed and hydrolyzed

for 24 hr at 110•Ž. The hydrolysate was dried rapidly

(within 15 min) by means of a rotary evaporator at 45•Ž. The amino acid compositions were calculated as values relative to aspartic acid (including aspar

gine). Tryptophan content was not determined.

Fig. 9. Estimation of the sedimentation constant of fruit bromelain and papain cleavage products. Sample solutions of 2-6 mg/ml in 0.2M NaCl were analyzed using an analytical centrifuge apparatus, Hitachi model UCA-1A, at a speed of 55,430 rpm at constant temperature (20°C). The sedimentation constants (S20, W) of the samples at different con centrations were plotted after correction for the buffer effect.

Fig. 10. Molecular weight determination on a A sedimentation constant of 3.5 S was obtained Sephadex G-150 column. Column size, 1 x 100 cm ; for both Fc(F) and Fc(P), as shown in Fig. 9. solvent, 0.03M Tris-hydrochloride buffer, pH 7.5, Molecular weight was estimated by gel containing O.1 m NaCl ; flow rate, 4 ml/hr, regulated filtration on a Sephadex G-150 column with a with a Varioperpex pump (LKB) ; temperature, 22°C. 1.0 ml sample solution (2-5 mg/ml) was applied, each constant flow rate of 4 ml/hr at 22•Ž. Equi fraction volume, being 1.0 ml. volumes of standard and sample solutions were applied to the column. The elution volume giving maximum UV absorbancy was

Sedimentation Constant and Molecular plotted against molecular weight on logarith Weight-To determine the sedimentation con mic abscissa according to Andrew's method stant S20, w, samples of different concentrations (10) (Fig. 10). ranging from 2-6 mg/ml were analyzed. The The elution volumes of Fab (F) and Fab sample solution was run at 55,430 rpm in a (P) were almost identical, and Fc (F) and Fc standard cell with the temperature maintained (P) also coincided with each other. The molec at 20°C. The results showed a sedimentation ular weights obtained for Fab and Fc frag constant of 3.7 S for both Fab(F) and Fab(P). ments from both fruit bromelain and papain

J. Biochem. PROTEOLYSIS OF IgG BY FRUIT BROMELAIN 503 digestion were 51,000 and 48,000, respectively. whole IgG to papain and fruit bromelain. Amino Acid Analysis-Amino acid com This experiment further implies that the positions of four kinds of fragments : Fab (F), breakdown of human IgG occurs within the Fab (P), Fc (F), and Fc (P) were analyzed and hinge region. The sensitivity of this partic their values relative to aspartic acid (including ular site was already pointed out in previous ,aspargine) were calculated (Table 1). experiments with various proteolytic enzymes Fab fragments from papain and fruit (1-5) and even chemical cleavage (13). The bromelain digests gave very similar values breakdown of IgG with detergent in the .except for lysine, proline, and glutamic acid. absence of enzyme and with and without Fc fragments from papain and fruit bromelain reducing agent has also been reported (14 ). digests were almost identical except for very It is most likely that the breakdown of IgG small deviations for serine, valine, leucine, into Fab and Fc-like fragments depends more and lysine. on its structural features than on the speci ficity of enzymatic cleavage.

DISCUSSION We would like to thank Dr. S. Ota, The Daiichi College of Pharmaceutical Science for his kind supply Based upon the above results it may be said of fruit bromelain (A), and Dr. T. Tada, Chiba Uni that fruit bromolain causes limited proteolysis versity for his generous supply of anti-subclass sera. of human IgG, yielding Fab (F) fragments We also wish to thank Miss T. Oshima for her tech having a molecular weight of 51,000 and S20, w nical asistance in ultracentrifugal analysis. :rate of 3.7, and Fc(F) fragments having a

molecular weight of 48,000 and S20,w rate of REFERENCES 3.5. These fragments were found to be closely similar to papain enzymatic fragments 1. A.G. Plaut, N. Calvanico, and T.B. Tomasi, J. Imnzunol., 108, 289 (1972). in their physical, chemical, and immunological 2. G.E. Connell and R.R. Porter, Biochem. J., 214, properties. This strongly suggests that the 53 (1971). cleavage point of fruit bromelain and that of 3. W.J. Mandy, M.M. Rivers, and A. Nisonoff, J. papain on human IgG could be very close to Biol. Chem., 236, 3221 (1961). each other. However fruit bromelain differs 4. R.R. Porter, Biochenz. J., 73, 119 (1959). remarkably from papain in its weak cleaving 5. M. Silvia and G. Victor, Stud. Cercet. Biochem., action on human IgG. This difference in 15, 55 (1972). resistance seems to be due to subclass speci 6. S. Ota, S. Moore, and W.H. Stein, Biochemistry, ficity. According to the analytical results of 3, 180 (1964). Jefferis et al. (12), using myeloma proteins 7. E. Franklin, J. Clin. Invest., 39, 1933 (1960). belonging to four kinds of IgG subclass, only 8. R. Axen, J. Porath, and S. Ernback, Nature, 214, 1302 (1967). .IgG 2 showed relative resistance to papain 9. J. Porath, R. Axen, and S. Ernback, Nature, 215, enzymatic attack in the presence of cysteine. 1491 (1967). One the other hand, as shown in this experi 10. P. Andrews, Biochem. J., 91, 222 (1964). ment, not only IgG 2 but also IgG 1 remained 11. D.H. Spackinan, W.H. Stein, and S, Moore, resistant to enzymatic attack by fruit bro Anal. Chem., 30, 1190 (1953). melain in the presence of cysteine. The fact 12. R. Jefferis, P.D. Weston, and D.R. Stanworth, that the fruit bromelain-resistant portion in Nature, 219, 646 (1968). volves two subclass components IgG 1 and 13. H.J. Cahnmann, R. Arnon, and M. Sela, J. Biol. :IgG 2 and the papain-resistant portion involves Chem., 240, PC2762 (1965). only one subclass component IgG 2 could ac 14. H. Jaquet, B. Bloom, and J.J. Cebra, J. Immunol., 92, 991 (1964). count for the difference in the resistance of

Vol. 74, No. 3. 1973