- Email: [email protected]
Isolation of two discrete human interferon-γ (immune) subtypes by high-performance liquid chromatography
ANAIYTK‘AI
137,
RfOCfll~MtSTRY
Isolation
of Two
1 is-1
19 (1984)
Discrete
Human
High-Performance JOSEF’H FRIEDLANDER.
Interferon-y Liquid
(Immune)
by
Chromatography’
DINA G. FISCHER, AND MENACHEM
Received
Subtypes
RUBINSTEIN
June 22. 1983
A rapid procedure for isolation of two biologically active human interferon-r subtypes was developed. Crude interferon-y produced in a serum-free culture of peripheral blood mononuclear cells by mitogen stimulation was concentrated and partially purified by chromatography on controlled-pore glass. Following desalting and concentration by ultrafiltration. a step of cationexchange high-performance liquid chromatography was performed. A linear NaCl gradient (O.Ol0.4 M) ;at pH 7 was employed and four peaks of biological activity eluting at 0.17, 0.20. 0.26 (major peak). and 0.3 M were obtained. The major peak of biological activity coincided with two prcmtein peaks. Analysis of one fraction from the major activity peak by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a protein hand having an apparent molecular weight of 26,000. while an adjacent fraction of the same activity peak contained a protein band corresponding to a molecular weight of 2 I .OOO The specific acticity of both subtypes was 7-10 j, IO’ units/mg.
Among the various types of human interferon, native interferon-y (EN-y) (1) is the least characterized. This substance is produced in lymphocytes upon stimulation with mitogens or specific antigens and it differs significantly in its structure (2) and properties (3) from the virus-induced LV- and b-interferons. Recently, a method of human IFN-7 production based on the stimulation of lymphocytes by a combination of the phorbol ester 120tetradecanoylphorbol13-acetate and phytohemagglutinin has been reported (4). Later. two subtypes of EN-y were isolated by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis (5). These subtypes had apparent molecular weights of 20,000 and 25,000, they corresponded to Coomassie blu,z-stainable protein bands, and were antigenically cross-reactive to each other. However. the irreversible denaturation of IFN-
y by SDS resulted in an X5-90%’ loss of biological activity during electrophoresis (5-e). So far, a convenient method of obtaining pure and biologically active cellular IFN-y has not yet been described. Such a procedure is essential for studying the biological properties of the various IFN-7 subtypes. High-performance liquid chromatography offers the speed and resolution needed for convenient purification of peptides and proteins. Reverse-phase HPLC has been used for the purification to homogeneity of IFN-cu and IFN-fi (7.8). However, IFN-y (and many other proteins) are irreversibly denatured under the conditions used in reverse-phase HPLC. Recently. ion-exchange HPLC columns, which enable separation of proteins under mild conditions, became available. We now report a procedure for purification to homogeneity of two cellular IFN-7 subtypes without loss of their biological activity. MATERIALS
’ Work supported by a grant from InterYeda. Israel. ’ Abbreviations Iused: IFN. interferon: SDS. sodium dodecyl sulfate.
AND
METHODS
Matwials. Mono-$ HPLC columns (5 X 50 mm. particle size IO wrn) were obtained from
116
FRIEDLANDER.
FISCHER.
Pharmacia. Phytohemagglutinin (PHA. reagent grade), was from Burroughs Wellcome. Controlled-pore glass beads (PG- 120-200) were from Sigma. Peripheral blood was obtained from normal donors. Crude 1FN-y was prepared in serum-free cultures of mononuclear cells as described (4). It was harvested after 24 h of incubation and stored at 4°C. Assay oJ‘IFN-7. IFN activity was assayed in 96-well flat bottom microculture plates by the inhibition of cytopathic effect of vesicular stomatitis virus in human WISH (ATCC CCL25) cells. In the absence of an international IFN-y standard, the assay was calibrated against IFN-a reference standard G-023-90 l527, kindly provided by the National Institutes of Health (Bethesda, Md.), and a laboratory standard of IFN-7. The presence or absence of IFN-a activity was tested on bovine MDBK cells (ATCC CCL-22) using the same type of assay.
Chromatography oJ IFN-y with corrtrolledpore glass and ultrafiltration. Controlled-pore glass beads were added to culture supernatants containing IFN-7 (3000- 16,000 units/ml, 8-20 X 1O4 units/mg) at a ratio of 1: 100 (by vol) in a polypropylene centrifuge bottle (4.9). The mixture (1 liter) was stirred for 3 h at 4°C the beads were allowed to settle, the supernatant was aspirated, and the beads were packed into a siliconized glass column (1 X 15 cm). The column was first washed with phosphate-buffered saline (50 ml) and 1FN-r was eluted by tetramethylammonium chloride (0.5 M) in phosphate-buffered saline (10). Fractions of 4 ml were collected and assayed for antiviral activity. Fractions containing IFN-7 activity were combined, desalted and concentrated by ultrafiltration on a YM-10 membrane (Amicon). The concentrate was diluted 3 times with 20 ml of 2 mM Na phosphate (pH 7.0) and finally concentrated to about 0.5% of the original culture volume. Insoluble proteins were removed by centrifugation (SOOOg, 10 min, 4°C) and the clear supernatant was stored at 4°C until used.
High-performance HPLC was performed
liquid chromatography on an Altex Model 330
AND
RURlNSrElN
liquid chromatograph (Beckman Instruments). A Mono-S HPLC cation-exchange column was equilibrated with 10 mM sodium phosphate, pH 7.0. in 20% (by vol) ethylene glycol (Buffer A). The IFN-y preparation from the previous step was loaded at a flow rate of 0.5 ml/min. The column was then washed with Buffer A at the same flow rate for 30 min. followed by a 60-min linear gradient of NaCl (O-400 mM) in Buffer A. The protein elution profile was obtained by an automated fluorescamine monitoring system (I 1). Each fraction (1 ml) was then asayed for IFN-y activity and for protein content by the fluorescamine column bypass system ( 12). Crystalline bovine serum albumin (Miles) was used as a protein standard.
SDS-polyacrylamide
gel electrophoresis.
Protein samples were treated with SDS and 2% fl-mercaptoethanol (5 min. 100°C) and electrophoresed on a slab gel of 15% polyacrylamide ( 13). After electrophoresis, protein bands were visualized by a silver stain (14) and by Coomassie blue. RESULTS
Fractionation
of Interferon-y
The binding of IIN-? to controlled-pore glass.which was performed according to the literature with minor modifications was quantitative (>99%, as judged by analysis of the unbound fraction) and enabled efficient concentration from crude cultures. Quantitative elution of IF-N--, was obtained by tetramethylammonium chloride. Removal of salt and a 5- to 20-fold concentration wasachieved by ultrafiltration. In addition to concentration and desalting, a 5-fold purification was achieved at this step. The purification is summarized in Table 1. High performance liquid chromatography wasperformed on a Mono-S cation exchanger. Addition of ethylene glycol to the elution buffers improved the resolution considerably, probably by reducing hydrophobic interactions. Several peaks of biological activity elut-
CHROMATOGRAPHIC
ISOLATION
OF
TWO
TABLE PURIFICATION
Step Crude CPG pool Y M- 10 concentrate Mono-S (all fractions) Fraction 38 (26 k.Da) Fraction 39 (21 kDa)
Volume (ml)
Total protein (mid
2300 29 3.7
730 170 21
1 1
0.18 0.11
INTERFERON-~
Total activity (units)
ing at NaCl calncentrations of 0.17,0.20,0.26, and 0.30 M were consistently obtained and the peak at 0.26 M was predominant (Fig. 1). The major peak of biological activity eluting at 0.26 M NaCl coincided with two partially resolved protein peaks. This pattern was consistent in over 30 repetitions. Aliquots ( 10%) from each peak of biological activity were pooled separately. diluted to a NaCl concen-
FIG. 1. Fractionation of IFN--, by cation-exchange HPLC. A Mono-S cation-exchange column (5 Y 50 mm) was equilibrated with 10 mM sodium phosphate buffer. pH 7.0. containing 20% (by vol) ethylene glycol (Buffer A). Partially purilied IFN-y (31 mg, 1.8 X 10’ units, in 3.7 ml Buffer A) was loaded on the column at a flow rate of 0.5 ml/min. The column was eluted with Buffer A for 30 min followed by a 60.min linear gradient of NaCl (O400 mM) in Buffer A. Fractions of 1 ml were collected in polypropylene test tubes.
117
SUBTYPES
1
OF HUMAN
3x 2.9 x 1.8 x 7 x 1.2 x 0.8 x
INTERFERON-y
Specific activity (units/mg) 10’ 10’ 10’ 18 lOh IO”
Purification factor
4 x IO4 1.7 x 105 8.6 x 10’
1 4 21
6.7 x 106 7.4 x 10”
167 185
Cumulative recovery P:) 100 97 60 23 4 3
tration of less than 50 mM, and rechromatographed on the same column. The elution positions of each protein and activity peak remained the same and no redistribution of its biological activity was obtained. The yield of IFN--/ was low (20%): however, no change in the specific activity was obtained.
Clmucterization
of IFN-y
Subtypes
Analysis of SDS-polyacrylamide gel electrophoresis of fraction 38 gave a major protein band with an apparent molecular mass of 26,000 Da (subtype 26 kDa), while fraction 39 gave a major protein band with an apparent molecular mass of 2 1,000 Da (subtype 2 1 kDa) (Fig. 2). Fractions corresponding to the other peaks of biological activity exhibited several protein bands including a 26,000 band. The specific activities of both subtypes 26 kDa and 2 1 kDa were 7- 10 X 1O6units/mg. The other peaks of biological activity had lower specific activities ( l-4 X 10hunits/mg). Acid sensitivity was tested to confirm the identity of the various peaks. Aliquots from all fractions were treated with 0.5 M acetic acid (1 h, 25°C) and then tested for residual activity. It was found that the titer was reduced by three logs. In addition, no antiviral cross-reactivity (~0. I %) was exhibited in bovine cells by any of these peaks, confirming their identity as IFN-7. DISCUSSION
The present study describes a simple procedure for the puritication of human immune interferon. This protein has been partially pu-
118
FRIEDLANDER.
FISCHER.
A
AND
RLlBINSTEIN
8
FIG. 2. SDS-polyacrylamide gel elcctrophoresis of IFN-y. (A) Silver staining. Lane (a) Molecular weight markers (from top: phosphorylase 94.000; albumin 67,000; ovalhumin 43.000; carbonic anhydrase 30.000: trypsin inhibitor 20.100; lactalbumin 14.400): (h) Crude IFN-y ( 140 ng): tc) IFN-7 after controlled-pore glass chromatography (300 ng): (d) IF& after ultrahltration (300 ng); (e) molecular weight markers: (f) HPLC fraction No. 38 (100 ng); (g) HPLC fraction No. 39 ( 100 ng). (B) Coomassie blue staining. Lane (a) IFN-7 after controlled-pore glass chromatography: (b) molecular weight markers (see above A, a): (c) HPLC fraction No. 37 (4 pg): (d) HPLC fraction No. 38 (1 gg); (e) HPLC fraction No. 39 (I pg).
rihed by chromatography on a large number of adsorbents including controlled-pore glass, concanavalin A-Sepharose, Cibacron BlueSepharose, and a variety of ion exchangers (5, 15.16). These methods gave relatively low resolution. therefore requiring a multistep procedure in order to obtain significant purification. Consequently, such multistep procedures give a very poor yield. In some cases proteins can be purified to homogeneity by preparative polyacrylamide gel electrophoresis in the presence of SDS: however, in most cases, including that of IIN(5). a signiticant loss of biological activity is caused due to an irreversible denaturation by the detergent. In the present study, the usefulness of cation-exchange HPLC for the purification of IFN-~ was demonstrated. High resolution was obtained under nondenaturing conditions. thereby allowing the separation of two IFN-y subtypes in one step without a significant loss of biological activity. Since many other lymphokines are present in crude IFN-
y the procedure can be adapted in principle for their isolation. The 26,000- and 2 1,OOO-Da polypeptides which were resolved in the present study were previously identified as two UN-y subtypes (5). All the peaks ofantiviral activity obtained by HPLC were sensitive to low pH and showed no cross-reactivity on bovine cells, indicating that they are all IFN-y. More recently, we were able to further confirm their identity as IFN-y subtypes by neutralization with specific anti-IFN-y monoclonal antibodies (I 7). The two protein peaks overlapping the major biological activity peak (fractions 37-40) seemed to be only partly resolved. In fact, they were completely resolved as revealed by SDS-gel electrophoresis. This situation is common in HPLC due to detector limitation (18). The purity of the isolated peaks was demonstrated by SDS-gel electrophoresis which is an independent proof since no size separation was performed during the purification. The availability of these biologically active. isolated
CHROMATOGRAPHIC
ISOLATION
EN-y subtypes will enable comparison of their structure and their biological activities. ACKNOWLEDGMENTS We thank thusiasm. Dr. Haim Vemer and assay of Saniei Career
Dr. Michel Revel for his support and enJ. Shoham for helpful discussions, and Mr. and Mrs. Rachel Eisenstadt for production IFN-). M.R. is an incumbent of the Ionel Development Chair.
REFERENCES I. Stewart 11. W. E., Blalock. J. E., Burke, D. C., Chany. C.. Dunnick. J. K.. Falcoff. E., Friedman. R. M., Galasso. G. J.. Jokhk. W. K.. and Vilcek, J. T. (1980) .J. Immunol125, 2353-2355. 2. Gray. P. W.. Leung, D. W.. Pennica. D.. Yelverton, E.. Najarian. R.. Simonsen, C. C.. Derynck, R., Sherwood. P. J.. Wallace. D. M., Berger, S. L., Levinson. A. D.. and Goeddel. D. V. (1982) Nuiurc, (Lmdon) 295, 503-508. 3. Stewart II. W. E. (1979) The Interferon System, Springer-‘v’erlag. New York/Berlin. 4. Yip. Y. K., Pang, R. H. L., Urban. C.. and Vilcek. J. (I 98 I) Proc. IV&/. :tc& Sci. C:S.,l 78, 16011605. 5. Yip. Y. K.. Barrowclough. B. S.. Urban. C.. and Vilcek, J. ( 1982) prrx ,vu/i. -Icad Srx C:T.4 79. 18201824.
OF
TWO
INTERFERON--,
6. Yip, Y. K., Barrowclough,
SUBTYPES
119
B.S., Urban, C., and Vilcek, J. (1982) Science 215, 41 l-413. Rubinstein, M.. Rubinstein, S.. Familletti. P. C.. Miller, R. S., Waldman. A. A.. and Pestka, S. ( 1979) Proc. Natl. Acud. Scl. USA 76, 640-644. 8. Rubinstein, M. (1979) Anal. Biochem. 97, 1-7. 9. Georgiades, J. A. (1982) TES. Rep. Bid. Med. 41, 179-183. IO. Chadha. K. C.. and Sulkowski, E. ( 1982) J. Jntcrfirott Rt7.x 2, 229-234. I I. Bohlen. P., Stein. S., Stone, J., and Udenfriend, S. (lY77) Anal. Biochrm. 67. 438-445. 12. Stein, S., and Moschera. J. (1981) in Methods in Enzymology (Pestka. S., ed.). Vol. 79. pp. 7-16, Academic Press, New York. 13. Laemmli, U. K. (1970) Nature (L~ndo~~) 227, 680685. 14. Sammons, D. W., Adams. L. D.. and Mishizawa. E. E. (1981) E/rc,lrophore.Pi.P 2, 135-141. 15. Mizrahi, A., O’Mally. J. A., Carter, W. A.. Takatsuki. A.. Tamura. G.. and Sulkowski. E. (I 978) J. Hiol. C’hcm. 253. 1612-7615. 16. Wirwanovnska-Stewart. M., Lin. L. S., Braude. 1. A.. and Stewart II. W. E. (1980) hf&c. Jmmw~ol 17. 625-633. 17. Novick, D., Eshhar. Z.. Fischer, D. G., Friedlander. J.. and Rubinstein, M. ( 1983) E.44BO .I. 2, 15271530. 18. Snyder, L. R.. and Kirkland, J. J. (1979) 1,) Introduction to Modern Liquid Chromatography. 2nd ed.. p. 44, Wiley, New York.
137,
RfOCfll~MtSTRY
Isolation
of Two
1 is-1
19 (1984)
Discrete
Human
High-Performance JOSEF’H FRIEDLANDER.
Interferon-y Liquid
(Immune)
by
Chromatography’
DINA G. FISCHER, AND MENACHEM
Received
Subtypes
RUBINSTEIN
June 22. 1983
A rapid procedure for isolation of two biologically active human interferon-r subtypes was developed. Crude interferon-y produced in a serum-free culture of peripheral blood mononuclear cells by mitogen stimulation was concentrated and partially purified by chromatography on controlled-pore glass. Following desalting and concentration by ultrafiltration. a step of cationexchange high-performance liquid chromatography was performed. A linear NaCl gradient (O.Ol0.4 M) ;at pH 7 was employed and four peaks of biological activity eluting at 0.17, 0.20. 0.26 (major peak). and 0.3 M were obtained. The major peak of biological activity coincided with two prcmtein peaks. Analysis of one fraction from the major activity peak by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed a protein hand having an apparent molecular weight of 26,000. while an adjacent fraction of the same activity peak contained a protein band corresponding to a molecular weight of 2 I .OOO The specific acticity of both subtypes was 7-10 j, IO’ units/mg.
Among the various types of human interferon, native interferon-y (EN-y) (1) is the least characterized. This substance is produced in lymphocytes upon stimulation with mitogens or specific antigens and it differs significantly in its structure (2) and properties (3) from the virus-induced LV- and b-interferons. Recently, a method of human IFN-7 production based on the stimulation of lymphocytes by a combination of the phorbol ester 120tetradecanoylphorbol13-acetate and phytohemagglutinin has been reported (4). Later. two subtypes of EN-y were isolated by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis (5). These subtypes had apparent molecular weights of 20,000 and 25,000, they corresponded to Coomassie blu,z-stainable protein bands, and were antigenically cross-reactive to each other. However. the irreversible denaturation of IFN-
y by SDS resulted in an X5-90%’ loss of biological activity during electrophoresis (5-e). So far, a convenient method of obtaining pure and biologically active cellular IFN-y has not yet been described. Such a procedure is essential for studying the biological properties of the various IFN-7 subtypes. High-performance liquid chromatography offers the speed and resolution needed for convenient purification of peptides and proteins. Reverse-phase HPLC has been used for the purification to homogeneity of IFN-cu and IFN-fi (7.8). However, IFN-y (and many other proteins) are irreversibly denatured under the conditions used in reverse-phase HPLC. Recently. ion-exchange HPLC columns, which enable separation of proteins under mild conditions, became available. We now report a procedure for purification to homogeneity of two cellular IFN-7 subtypes without loss of their biological activity. MATERIALS
’ Work supported by a grant from InterYeda. Israel. ’ Abbreviations Iused: IFN. interferon: SDS. sodium dodecyl sulfate.
AND
METHODS
Matwials. Mono-$ HPLC columns (5 X 50 mm. particle size IO wrn) were obtained from
116
FRIEDLANDER.
FISCHER.
Pharmacia. Phytohemagglutinin (PHA. reagent grade), was from Burroughs Wellcome. Controlled-pore glass beads (PG- 120-200) were from Sigma. Peripheral blood was obtained from normal donors. Crude 1FN-y was prepared in serum-free cultures of mononuclear cells as described (4). It was harvested after 24 h of incubation and stored at 4°C. Assay oJ‘IFN-7. IFN activity was assayed in 96-well flat bottom microculture plates by the inhibition of cytopathic effect of vesicular stomatitis virus in human WISH (ATCC CCL25) cells. In the absence of an international IFN-y standard, the assay was calibrated against IFN-a reference standard G-023-90 l527, kindly provided by the National Institutes of Health (Bethesda, Md.), and a laboratory standard of IFN-7. The presence or absence of IFN-a activity was tested on bovine MDBK cells (ATCC CCL-22) using the same type of assay.
Chromatography oJ IFN-y with corrtrolledpore glass and ultrafiltration. Controlled-pore glass beads were added to culture supernatants containing IFN-7 (3000- 16,000 units/ml, 8-20 X 1O4 units/mg) at a ratio of 1: 100 (by vol) in a polypropylene centrifuge bottle (4.9). The mixture (1 liter) was stirred for 3 h at 4°C the beads were allowed to settle, the supernatant was aspirated, and the beads were packed into a siliconized glass column (1 X 15 cm). The column was first washed with phosphate-buffered saline (50 ml) and 1FN-r was eluted by tetramethylammonium chloride (0.5 M) in phosphate-buffered saline (10). Fractions of 4 ml were collected and assayed for antiviral activity. Fractions containing IFN-7 activity were combined, desalted and concentrated by ultrafiltration on a YM-10 membrane (Amicon). The concentrate was diluted 3 times with 20 ml of 2 mM Na phosphate (pH 7.0) and finally concentrated to about 0.5% of the original culture volume. Insoluble proteins were removed by centrifugation (SOOOg, 10 min, 4°C) and the clear supernatant was stored at 4°C until used.
High-performance HPLC was performed
liquid chromatography on an Altex Model 330
AND
RURlNSrElN
liquid chromatograph (Beckman Instruments). A Mono-S HPLC cation-exchange column was equilibrated with 10 mM sodium phosphate, pH 7.0. in 20% (by vol) ethylene glycol (Buffer A). The IFN-y preparation from the previous step was loaded at a flow rate of 0.5 ml/min. The column was then washed with Buffer A at the same flow rate for 30 min. followed by a 60-min linear gradient of NaCl (O-400 mM) in Buffer A. The protein elution profile was obtained by an automated fluorescamine monitoring system (I 1). Each fraction (1 ml) was then asayed for IFN-y activity and for protein content by the fluorescamine column bypass system ( 12). Crystalline bovine serum albumin (Miles) was used as a protein standard.
SDS-polyacrylamide
gel electrophoresis.
Protein samples were treated with SDS and 2% fl-mercaptoethanol (5 min. 100°C) and electrophoresed on a slab gel of 15% polyacrylamide ( 13). After electrophoresis, protein bands were visualized by a silver stain (14) and by Coomassie blue. RESULTS
Fractionation
of Interferon-y
The binding of IIN-? to controlled-pore glass.which was performed according to the literature with minor modifications was quantitative (>99%, as judged by analysis of the unbound fraction) and enabled efficient concentration from crude cultures. Quantitative elution of IF-N--, was obtained by tetramethylammonium chloride. Removal of salt and a 5- to 20-fold concentration wasachieved by ultrafiltration. In addition to concentration and desalting, a 5-fold purification was achieved at this step. The purification is summarized in Table 1. High performance liquid chromatography wasperformed on a Mono-S cation exchanger. Addition of ethylene glycol to the elution buffers improved the resolution considerably, probably by reducing hydrophobic interactions. Several peaks of biological activity elut-
CHROMATOGRAPHIC
ISOLATION
OF
TWO
TABLE PURIFICATION
Step Crude CPG pool Y M- 10 concentrate Mono-S (all fractions) Fraction 38 (26 k.Da) Fraction 39 (21 kDa)
Volume (ml)
Total protein (mid
2300 29 3.7
730 170 21
1 1
0.18 0.11
INTERFERON-~
Total activity (units)
ing at NaCl calncentrations of 0.17,0.20,0.26, and 0.30 M were consistently obtained and the peak at 0.26 M was predominant (Fig. 1). The major peak of biological activity eluting at 0.26 M NaCl coincided with two partially resolved protein peaks. This pattern was consistent in over 30 repetitions. Aliquots ( 10%) from each peak of biological activity were pooled separately. diluted to a NaCl concen-
FIG. 1. Fractionation of IFN--, by cation-exchange HPLC. A Mono-S cation-exchange column (5 Y 50 mm) was equilibrated with 10 mM sodium phosphate buffer. pH 7.0. containing 20% (by vol) ethylene glycol (Buffer A). Partially purilied IFN-y (31 mg, 1.8 X 10’ units, in 3.7 ml Buffer A) was loaded on the column at a flow rate of 0.5 ml/min. The column was eluted with Buffer A for 30 min followed by a 60.min linear gradient of NaCl (O400 mM) in Buffer A. Fractions of 1 ml were collected in polypropylene test tubes.
117
SUBTYPES
1
OF HUMAN
3x 2.9 x 1.8 x 7 x 1.2 x 0.8 x
INTERFERON-y
Specific activity (units/mg) 10’ 10’ 10’ 18 lOh IO”
Purification factor
4 x IO4 1.7 x 105 8.6 x 10’
1 4 21
6.7 x 106 7.4 x 10”
167 185
Cumulative recovery P:) 100 97 60 23 4 3
tration of less than 50 mM, and rechromatographed on the same column. The elution positions of each protein and activity peak remained the same and no redistribution of its biological activity was obtained. The yield of IFN--/ was low (20%): however, no change in the specific activity was obtained.
Clmucterization
of IFN-y
Subtypes
Analysis of SDS-polyacrylamide gel electrophoresis of fraction 38 gave a major protein band with an apparent molecular mass of 26,000 Da (subtype 26 kDa), while fraction 39 gave a major protein band with an apparent molecular mass of 2 1,000 Da (subtype 2 1 kDa) (Fig. 2). Fractions corresponding to the other peaks of biological activity exhibited several protein bands including a 26,000 band. The specific activities of both subtypes 26 kDa and 2 1 kDa were 7- 10 X 1O6units/mg. The other peaks of biological activity had lower specific activities ( l-4 X 10hunits/mg). Acid sensitivity was tested to confirm the identity of the various peaks. Aliquots from all fractions were treated with 0.5 M acetic acid (1 h, 25°C) and then tested for residual activity. It was found that the titer was reduced by three logs. In addition, no antiviral cross-reactivity (~0. I %) was exhibited in bovine cells by any of these peaks, confirming their identity as IFN-7. DISCUSSION
The present study describes a simple procedure for the puritication of human immune interferon. This protein has been partially pu-
118
FRIEDLANDER.
FISCHER.
A
AND
RLlBINSTEIN
8
FIG. 2. SDS-polyacrylamide gel elcctrophoresis of IFN-y. (A) Silver staining. Lane (a) Molecular weight markers (from top: phosphorylase 94.000; albumin 67,000; ovalhumin 43.000; carbonic anhydrase 30.000: trypsin inhibitor 20.100; lactalbumin 14.400): (h) Crude IFN-y ( 140 ng): tc) IFN-7 after controlled-pore glass chromatography (300 ng): (d) IF& after ultrahltration (300 ng); (e) molecular weight markers: (f) HPLC fraction No. 38 (100 ng); (g) HPLC fraction No. 39 ( 100 ng). (B) Coomassie blue staining. Lane (a) IFN-7 after controlled-pore glass chromatography: (b) molecular weight markers (see above A, a): (c) HPLC fraction No. 37 (4 pg): (d) HPLC fraction No. 38 (1 gg); (e) HPLC fraction No. 39 (I pg).
rihed by chromatography on a large number of adsorbents including controlled-pore glass, concanavalin A-Sepharose, Cibacron BlueSepharose, and a variety of ion exchangers (5, 15.16). These methods gave relatively low resolution. therefore requiring a multistep procedure in order to obtain significant purification. Consequently, such multistep procedures give a very poor yield. In some cases proteins can be purified to homogeneity by preparative polyacrylamide gel electrophoresis in the presence of SDS: however, in most cases, including that of IIN(5). a signiticant loss of biological activity is caused due to an irreversible denaturation by the detergent. In the present study, the usefulness of cation-exchange HPLC for the purification of IFN-~ was demonstrated. High resolution was obtained under nondenaturing conditions. thereby allowing the separation of two IFN-y subtypes in one step without a significant loss of biological activity. Since many other lymphokines are present in crude IFN-
y the procedure can be adapted in principle for their isolation. The 26,000- and 2 1,OOO-Da polypeptides which were resolved in the present study were previously identified as two UN-y subtypes (5). All the peaks ofantiviral activity obtained by HPLC were sensitive to low pH and showed no cross-reactivity on bovine cells, indicating that they are all IFN-y. More recently, we were able to further confirm their identity as IFN-y subtypes by neutralization with specific anti-IFN-y monoclonal antibodies (I 7). The two protein peaks overlapping the major biological activity peak (fractions 37-40) seemed to be only partly resolved. In fact, they were completely resolved as revealed by SDS-gel electrophoresis. This situation is common in HPLC due to detector limitation (18). The purity of the isolated peaks was demonstrated by SDS-gel electrophoresis which is an independent proof since no size separation was performed during the purification. The availability of these biologically active. isolated
CHROMATOGRAPHIC
ISOLATION
EN-y subtypes will enable comparison of their structure and their biological activities. ACKNOWLEDGMENTS We thank thusiasm. Dr. Haim Vemer and assay of Saniei Career
Dr. Michel Revel for his support and enJ. Shoham for helpful discussions, and Mr. and Mrs. Rachel Eisenstadt for production IFN-). M.R. is an incumbent of the Ionel Development Chair.
REFERENCES I. Stewart 11. W. E., Blalock. J. E., Burke, D. C., Chany. C.. Dunnick. J. K.. Falcoff. E., Friedman. R. M., Galasso. G. J.. Jokhk. W. K.. and Vilcek, J. T. (1980) .J. Immunol125, 2353-2355. 2. Gray. P. W.. Leung, D. W.. Pennica. D.. Yelverton, E.. Najarian. R.. Simonsen, C. C.. Derynck, R., Sherwood. P. J.. Wallace. D. M., Berger, S. L., Levinson. A. D.. and Goeddel. D. V. (1982) Nuiurc, (Lmdon) 295, 503-508. 3. Stewart II. W. E. (1979) The Interferon System, Springer-‘v’erlag. New York/Berlin. 4. Yip. Y. K., Pang, R. H. L., Urban. C.. and Vilcek. J. (I 98 I) Proc. IV&/. :tc& Sci. C:S.,l 78, 16011605. 5. Yip. Y. K.. Barrowclough. B. S.. Urban. C.. and Vilcek, J. ( 1982) prrx ,vu/i. -Icad Srx C:T.4 79. 18201824.
OF
TWO
INTERFERON--,
6. Yip, Y. K., Barrowclough,
SUBTYPES
119
B.S., Urban, C., and Vilcek, J. (1982) Science 215, 41 l-413. Rubinstein, M.. Rubinstein, S.. Familletti. P. C.. Miller, R. S., Waldman. A. A.. and Pestka, S. ( 1979) Proc. Natl. Acud. Scl. USA 76, 640-644. 8. Rubinstein, M. (1979) Anal. Biochem. 97, 1-7. 9. Georgiades, J. A. (1982) TES. Rep. Bid. Med. 41, 179-183. IO. Chadha. K. C.. and Sulkowski, E. ( 1982) J. Jntcrfirott Rt7.x 2, 229-234. I I. Bohlen. P., Stein. S., Stone, J., and Udenfriend, S. (lY77) Anal. Biochrm. 67. 438-445. 12. Stein, S., and Moschera. J. (1981) in Methods in Enzymology (Pestka. S., ed.). Vol. 79. pp. 7-16, Academic Press, New York. 13. Laemmli, U. K. (1970) Nature (L~ndo~~) 227, 680685. 14. Sammons, D. W., Adams. L. D.. and Mishizawa. E. E. (1981) E/rc,lrophore.Pi.P 2, 135-141. 15. Mizrahi, A., O’Mally. J. A., Carter, W. A.. Takatsuki. A.. Tamura. G.. and Sulkowski. E. (I 978) J. Hiol. C’hcm. 253. 1612-7615. 16. Wirwanovnska-Stewart. M., Lin. L. S., Braude. 1. A.. and Stewart II. W. E. (1980) hf&c. Jmmw~ol 17. 625-633. 17. Novick, D., Eshhar. Z.. Fischer, D. G., Friedlander. J.. and Rubinstein, M. ( 1983) E.44BO .I. 2, 15271530. 18. Snyder, L. R.. and Kirkland, J. J. (1979) 1,) Introduction to Modern Liquid Chromatography. 2nd ed.. p. 44, Wiley, New York.