Importance of disulfide linkage for constructing the biologically active human interleukin-2

Importance of disulfide linkage for constructing the biologically active human interleukin-2

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 257, No. 1, August 15, pp. 194-199,1987 Importance of Disulfide Linkage Biologically Active Human TAKAO ...

516KB Sizes 0 Downloads 17 Views

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 257, No. 1, August 15, pp. 194-199,1987

Importance of Disulfide Linkage Biologically Active Human TAKAO

Biotechnology

YAMADA,’ KOICHI

AKIRA KATO,

Laboratories, Ltd

Received

February

for Constructing Interleukin-2

the

FUJISHIMA, KENJI KAWAHARA, OSAMU NISHIMURA

AND

Central Research , Yodogawa-ku,

LXvision, Takeda Osaka 532, .Japan

17, 1987, and in revised

form

April

Chemical

Industries,

27,1987

Recombinant human interleukin-2 (rIL-2) produced in Escherichia coli possessesa free thiol group at Cys-125 and a disulfide linkage between Cys-58 and Cys-105, as in the case for natural human interleukin-2. Treatment of rIL-2 with 200 mM dithiothreitol resulted in the cleavage of the Cys-58-Cys-105 disulfide bond. The reduced form of rIL2 thus obtained retained only 10% of the in vitro biological activity of the native form, as measured by the ability to stimulate the growth of an IL-Z-dependent mouse natural killer cell line, NKCS. Far-uv circular dichroism studies indicated that the cleavage of the disulfide bond results in a decrease of a-helix content. Near-uv circular dichroism studies suggested that the native molecule is folded into a rigid tertiary structure, while the reduced form showed a spectrum similar to that of rIL-2 denatured in the presence of 6 M guanidine . HCl. The once-reduced molecule was readily reoxidized in the presence of IO PM cU2+ to form the native molecule with full biological activity. These results strongly demonstrate that the Cys-58-Cys-105 disulfide linkage in the IL-2 molecule is essential for constructing a rigid and biologically active form of IL-Z. o 1987 Academic Press, Inc.

Recent advances in recombinant DNA technology have led to the mass production of eukaryotic proteins by Eschemkhia coli cells. However, some of recombinant proteins produced in this way have been proved to be different from natural proteins. Such recombinant proteins as human interferon-a (1, Z), human interferon-y (3), human growth hormone (4), and human interleukin-2 (IL-2)2 (5-8) possess an additional methionine residue at their amino termini corresponding to the initiator methionine codon. In addition, recombinant calf prochymosin is present as an oligo-

merit form interlinked partly by disulfide bonds and as a reduced form in inclusion bodies of E. coli cells (9). Harris thinks that several normally soluble proteins are insoluble in E. coli cells because they lack proper disulfide bridges or have incorrect disulfide bridges (10). It is well known that the formation of proper disulfide bonds is often essential to the structure and function of proteins. Therefore, it is important to obtain information on disulfide bridges of E. co&derived recombinant proteins. IL-2 is a lymphokine produced by activated T lymphocytes and plays important roles in the proliferation and differentiation of T lymphocytes, as well as in the regulation of the immune systems (11-13). We have purified and characterized recombinant human IL-2 (rIL-2) produced in E. coli cells harboring the human IL-2 gene (5). More recently, we have succeeded in

1 To whom correspondence should be addressed. ’ Abbreviations used: IL-2, interleukin-2; rIL-2, recombinant interleukin-2; DTT, dithiothreitol; TFA, trifluoroacetic acid, DTNB, 5,5’-dithiobis(2-nitrobenzoic acid). 0003-9861/87 Copyright All rights

$3.00

0 1987 by Academic Press, Inc. of reproduction in any form resewed.

194

IMPORTANCE

OF

DISULFIDE

LINKAGE

separating recombinant human IL-2 and its methionylated form by utilizing the difference of their isoelectric points (14). We report here the determination of the location of the disulfide linkage in E. coliderived rIL-2. The importance of the disulfide linkage for constructing a rigid form of the IL-2 molecule with full biological activity is discussed. MATERIALS

AND

METHODS

Puri,fication of rIL-2. rIL-2 produced in E. coli N4830/pTB 285 cells harboring the human IL-2 gene was purified (5) and separated from the contaminating methionyl rIL-2, as described previously (14). Isolation of thiol-containing peptide fragments. The thiol-containing peptides were isolated, essentially in the same manner as that reported by Egorov et al. (15). rIL-2 (6.3 mg, 420 nmol) dissolved in 3 ml of 6 M guanidine. HCl-0.1 M NaCl-25 mM ammonium acetate (pH 5.0) was applied at a flow rate of 5 ml/h to a Thiopropyl-Sepharose 6B (Pharmacia) column (1.0 X 5.1 cm) equilibrated with the same buffer. The column was successively washed with the equilibration buffer, 0.1 M NaCl-25 mM ammonium acetate (pH 5.0) and 0.2 M acetic acid (pH 3.0). The gel was transferred to a tube with 6 ml of 0.2 M acetic acid (pH 3.0) containing 0.1 mg of pepsin (EC 3.4.23.1; Sigma), and then the tube was rotated end over end for 15 h at 3’7°C. After peptic digestion, the gel was again packed in a column. The column was washed with 0.2 M acetic acid (pH 3.0) and 25 mM ammonium acetate (pH 4.5), and then the residual 2-thiopyridyl groups of the gel were removed as 2-thiopyridone with 20 mM 2mercaptoethanol-25 mM ammonium acetate (pH 4.5). The thiol-containing peptides were eluted at a flow rate of 5 ml/h with 20 mM 2-mercaptoethanol-25 mM ammonium acetate (pH 8.0) after the column was washed with 25 mM ammonium acetate (pH 4.5) and 25 mM ammonium acetate (pH 8.0). The eluate was taken to dryness with a Speed Vat concentrator (Savant Instruments, U.S.A.). The powder thus obtained was dissolved in 0.1% TFA and subjected to reverse-phase HPLC with a Nucleosil

FOR

ACTIVE

HUMAN

INTERLEUKIN-2

195

5C18 column (0.8 X 30 cm, Macherey-Nagel Company, FRG). As a result, 137 nmol of four thiol-containing peptides (13 nmol of P-l, 48 nmol of P-2,14 nmol of P-3, and 62 nmol of P-4) was isolated starting from 420 nmol of rIL-2, as estimated on the basis of amino acid analysis. Reduction and reoxidation of r&2. rIL2 was reduced by incubating a rIL-2 solution in the presence of 200 mM DTT for 12 h at pH 5.0 and at room temperature. The reaction mixture was applied to a Sephadex G-25 (Pharmacia) column (2.4 X 20 cm) equilibrated with 2 mM DTT-5 mM ammonium acetate (pH 5.0), and reduced rIL2 was collected. Reoxidation of reduced rIL2 was performed as follows. Reduced rIL2 obtained above was applied to a Sephadex G-25 column (2.4 X 20 cm) equilibrated with 10 pM CuS04-5 mM ammonium acetate (pH 5.0). The eluate was kept at 4°C for 20 h and then subjected to reverse-phase HPLC with an Ultrapore RPSC column (1.0 X 25 cm, Beckman Instruments) (5). The reoxidized form thus obtained was lyophilized and dissolved in 5 mM ammonium acetate (pH 5.0). The protein recovery was about 70%. Quantitative analysis of the free thiol group. The free thiol group was determined by the method of Ellman (16) with cysteine as a standard. Reduced rIL-2 was subjected to analysis after DTT in its preparation was removed by reverse-phase HPLC. Assay of IL-2 activity. IL-2-activity was determined by the ability to stimulate the growth of an IL-2-dependent murine cell line, NKCS, as reported previously (1’7). Assay of protein. Protein was determined spectrophotometrically in 5 mM ammonium acetate (pH 5.0) based on the molar absorption coefficient of rIL-2 at 280 nm, 9.58 X lo3 M-l cm-’ (5). Protein in the preparation of reduced rIL-2 was determined by the method of Bradford (18) using BioRad protein assay reagent with rIL-2 as a standard. The protein of the standard was determined by absorbance at 280 nm as described above. Analysis of amino acid composition. The amino acid composition was determined on a 24-h hydrolysate with 6 N HCI at 110°C in the presence of 4% thioglycolic acid.

196

YAMADA

Amino acid analysis using ninhydrin was performed on a Hitachi Model 835 amino acid analyzer. Cys was determined as cysteic acid on a 24-h hydrolysate after performic acid oxidation. Circular dichroism (CD) studies. CD spectra were determined at room temperature on a Jasco Model J-20A spectropolarimeter. Spectral bandwidth was set at 1 nm. The light path for 200 to 250 nm was 0.022 cm and that for 250 to 310 nm was 1.0 cm. The solvent spectrum was obtained and subtracted from the protein spectrum for each sample. The data were expressed as the mean residue ellipticity (fl), which was calculated using the mean residue weight of IL-2 of 116. RESULTS

Location of Disuljide Linkage

ET

AL.

0.05OC

z i ,,.,,25_ G g i yI 9 o ;njt-

60 nradon

80 Time

100

(min)

FIG. 1. Reverse-phase HPLC of thiol-containing peptides derived from rIL-2. Elution was performed with a linear gradient of acetonitrile concentration in the presence of 0.1% TFA. Flow rate was 3.0 ml/ min. Peaks, P-l, P-2, P-3, and P-4, were subjected to amino acid analysis and indentified as follows. P-l (Cys-125~Ser-130); Ser (1.5), Glx (l.O), Cys (0.7), Ile (1.5). P-2 (Cys-125-Thr-131); Thr (0.9), Ser (1.6), Glx (lo), Cys (0.8), Be (1.5). P-3 (Cys-125-Ile-129); Ser (0.9), Glx (l.O), Cys (0.8), Be (1.5). P-4 (Cys-125-Thr-133); Thr (1.7), Ser (1.6), Glx (l.O), Cys (0.8), Ile (1.5), Leu (0.9).

The rIL-2 molecule is composed of 133 amino acids and contains the three cysteine residues, Cys-58, Cys-105, and Cys-125 (14). The quantitative analysis of the free thiol group by the DTNB method (16) showed column to give four peaks, P-l, P-2, P-3, that one cysteine residue possessesa free and P-4 (Fig. 1). Based on amino acid analthiol group (see Table I), while the other ysis, P-l, P-2, P-3, and P-4 were identified two cysteine residues are linked together as Cys-125-Ser-130, Cys-125-Thr-131, Cysvia a disulfide bridge. To determine the lo- 125-Ile-129, and Cys-125-Thr-133, respeccation of the disulfide linkage, rIL-2 was tively, all of which contained Cys-125 (Fig. applied on Thiopropyl-Sepharose and gel- 2). The results demonstrate that the rIL-2 bound rIL-2 was digested by pepsin. The molecule possessesa free thiol group at thiol-containing peptide fragments were Cys-125 and a disulfide linkage between eluted from the gel and subjected to re- Cys-58 and Cys-105, as in the case for natverse-phase HPLC with a Nucleosil 5C18 ural human IL-2 (19). Treatment of rIL-2 with 200 mM DTT resulted in the cleavage of the Cys-58-CysTABLE I 105 disulfide linkage in the rIL-2 molecule. NUMBEROFFREETHIOLGROUPSAND BIOLOGICAL The reduced form of rIL-2 thus obtained ACTIVITYFORTHREE FORMSOF rIL-2 retained only about 10% of the biological rIL-2 form

Number of free thiol groups per molecule

Biological activity (U/w)

125 130 -Cys-Cln-SW-lie-lie-Ser-Thr-Leu-Thr-OH

133

--P-1-

Native Reduced Reoxidized

0.91 2.62 0.90

34,000 3100 35,500

Note. Number of free thiol groups and biological activity were determined as described under Materials and Methods.

a

p-2-

-p-3+ PP-4

4

FIG. 2. Amino acid sequence of carboxyl-terminal region in rIL-2. Each peptide refers to the peaks shown in Fig. 1.

IMPORTANCE

OF

DISULFIDE

LINKAGE

activity of the native form (Table I) and was eluted later than the native molecule on reverse-phase HPLC (Fig. 3). The cellfree extract of E coli cells, when subjected to reverse-phase HPLC immediately after extraction, gave a major peak not with the retention time for the Cys-58-Cys-105 linked form but with that for the reduced form of rIL-2 (Fig. 3D). The lyophilized preparation from this peak was found to consist of apparently homogeneous recombinant human IL-2 as judged by SDSpolyacrylamide gel electrophoresis. The quantitative analysis of the free thiol group in this preparation indicated that the peak arose from reduced rIL-2. In addition, treatment of the cell-free extract with DTT resulted in little change in the peak area. These results, therefore, suggest that recombinant human IL-2 is present as a reduced form in E. coli cells..

A 0.01 __--

___---

0.01

1

r

\

A

-0 I - 0.01

C

I

E P NO z 0 0.01 : 2 k 9” (0 (

F-y-?-L 1 10

20

Retention

ACTIVE

HUMAN

3 0 z .I u N? 8 g-1 s zr ‘0 x --2 ”

197

INTERLEUKIN-2

. . . . .. . . .. . . .. . . .. . ..-.

-

200

210

220 230 Wavelength

240 (nm)

FIG. 4. Far-uv CD spectra of various forms of rIL2. -, Native form in 5 mM ammonium acetate (pH 5.0) (1.03 mg/ml); ---, reduced form in 2 mM DTT/5 mM ammonium acetate (pH 5.0) (0.945 mg/ml); -. -, reoxidized form in 5 mM ammonium acetate (pH 5.0) (0.692 mg/ml); * * ., denatured form in 6 M guanidinemHC1/5 mM ammonium acetate (pH 5.0) (0.796 mg/ml).

CD Spectrum Analysis

y----’

__iLI3

0

FOR

30

Time

(min)

FIG. 3. Reverse-phase HPLC of native form (A), reduced form (B), reoxidized form (C) of rIL-2, and cellfree extract of E. coli cells (D). Samples were applied to an Ultrapore RPSC column (1.0 X 25 cm, Beckman Instruments). The amounts of protein used were .Ol mg (A, B, and C) and 1.0 mg (D). Elution was performed with a linear gradient of acetonitrile concentration in the presence of 0.1% TFA. Flow rate was 3.0 ml/min.

To compare the structure of the reduced form with that of native rIL-2, CD studies were carried out. Far-uv CD spectra (Fig. 4) indicated that the ordered structures were present in both forms, which showed a minimum at 208 nm and a shoulder around 222 nm, as described for native rIL2 previously by Cohen et al. (20). The rIL2 molecule contains 60% a-helix, 36% /3sheet, and 4% remainder as estimated by the method of Provencher et al. (21), while the reduced form is poorer in the a-helix content than is the native form (Fig. 4). In near-uv CD spectra, native rIL-2 exhibited positive bands at 266 and 291 nm and a broad positive band between 270 and 285 nm (Fig. 5). The results suggest that the native form is folded into a rigid tertiary structure. On the other hand, the spectrum of the reduced form was quite similar to that of rIL-2 denatured in the presence of 6 M guanidine . HCl (Fig. 5), indicating that reduced rIL-2 fails to have a rigid tertiary structure. These results reveal that the Cys-58-Cys-105 disulfide linkage of IL-2 plays an indispensable role for creating a

198

YAMADA

I

ET

I

260

270

280 Wavelength

290 (nm)

300

FIG. 5. Near-w CD spectra of various forms of rIL2. -, Native form in 5 mM ammonium acetate (pH 5.0) (1.82 mg/ml); ---, reduced form in 2 mM DTT/5 mM ammonium acetate (pH 5.0) (1.89 mg/ml); -a -, reoxidized form in 5 mM ammonium acetate (pH 5.0) (0.692 mg/ml); . . *, denatured form in 6 M guanidine. HC1/5 mM ammonium acetate (pH 5.0) (0.796 mg/ml).

rigid and biologically active form of the IL2 molecule. The reduced form was readily reoxidized in the presence of 10 PM Cu’+. The reoxidized form thus obtained possessed the same specific activity as that of the native form (Table I). Its CD spectra (Figs. 4 and 5) and its elution profile on reverse-phase HPLC (Fig. 3) were exactly the same as those for the native form. Analysis of the free thiol group indicated that the reoxidized form contains a disulfide linkage between Cys-58 and Cys-105, as in the case for the native form. DISCUSSION

The location of the disulfide linkage in rIL-2 was determined by pepsin digestion of Thiopropyl-Sepharose-bound rIL-2. This method was quite effective in that only the thiol-containing peptide fragments were isolated. It should be emphasized that acidic conditions throughout the chromatography including pepsin digestion prevent the artifactual disulfide exchange reaction. In fact, we found the disulfide

AL.

exchange reaction occurred in the chromatography under neutral or basic conditions. The experimental results indicated that the rIL-2 molecule contains a free thiol group at Cys-125 and a disulfide bridge between Cys-58 and Cys-105, as is the case for natural human IL-2 produced by the JURKAT cell line (19). Peptide mapping of rIL-2 with pepsin confirmed this conclusion (data not shown). Robb et al. reported that treating natural human IL-2 with DTT resulted in a 70% loss of its biological activity (19). It is known that the mutant rIL-2 proteins with Cys-58 or Cys-105 substituted by other amino acid residues have little biological activity (22,23). We found that the reduced form of rIL-2 retained only 10% of the biological activity of the native form. The actual biological activity of the reduced form might be lower if reoxidation into the native form during the bioassay is taken into consideration. The reduced form was eluted as a more hydrophobic derivative than the native form on reverse-phase HPLC. Reverse-phase HPLC analysis indicated that recombinant human IL-2 is present as a reduced form in E. coli cells. It might be kept under inadequately oxidative conditions in inclusion bodies (24) because of its high level of expression. Some of this reduced form is gradually activated by the air oxidation during the purification process to form the Cys-5%Cys-105 linkage.3 In this study, however, we showed that the reduced form can be rapidly oxidized in the presence of 10 PM Cu2+ to become a native form with full biological activity. This method would be very helpful in obtaining a good yield of Cys-5%Cys-105-linked rIL2 from the cell-free extract of E. coli cells. Far-uv CD studies suggested that the ordered structures are present in both native and reduced forms and that the reduced form has less a-helix content than the native form. In near-uv CD spectra, the native form of rIL-2 showed some CD signals derived from asymmetric environments for the aromatic residues (25),

’ T. Yamada

and K. Kato,

unpublished

results.

IMPORTANCE

OF

DISULFIDE

LINKAGE

clearly indicating that the native form has a tertiary structure into which the aromatic residues are incorporated. On the other hand, the reduced form showed a spectrum similar to that of rIL-2 denatured in the presence of 6 M guanidine . HCl. This evidence suggests that the aromatic residues are not incorporated into a rigid structure and are exposed to the solvent in the reduced form. These results, therefore, strongly demonstrate that the Cys-5%Cys105 disulfide linkage in the IL-2 molecule plays an essential role in constructing a rigid and biologically active form of IL-2.

FOR

We thank Dr. Yukio Sugino and Dr. Atsushi Kakinuma of the Central Research Division for their encouragement and discussion throughout this work. We are also grateful to Dr. Kyozo Tsukamoto and Dr. Osamu Shiho for determining IL-2 activity and Dr. James R. Miller for reading the manuscript.

ENBECK,

9. SCHOEMAKER, 10.

Chem.

N. K., AND

ture

293,408-411.

5. KATO,

(Lonclwn) K., YAMADA,

ASANO,

T.,

(1985)

B&hem.

STEBBING,

H.,

Biophys.

Vol.

AND

MAR-

Science

4, pp. 127-185,

F. W.,

AND

Ac-

GALLO,

R.

193,1007-1008.

GILLIS, S., FERM, M. M., Ou, W., AND (1978) .I Immunol 120,2027-2032.

13.

SMITH,

Annu.

K. A. (1984)

SMITH,

K. A.

Rev. Immunol.

2,319-

333. T., KATO,

K., KAWAHARA,

0. (1986) Biochem. 135,837~843.

mun 15. EGOROV,

T. CARLSSON,

TADA,

K., AND

Biophys.

H., SHIHO,

Arch.

Biochem.

O., KUROSHIMA,

NISHI-

Res. Corn-

A., SVENSON, A., RYDEN, J. (1975) Proc. Natl. Acad

72,3029-3033. 16. ELLMAN, G. L. (1959) 70.

BRADFORD,

Chem.

256,

19.

ROBB, H.

D. A., LIN,

Hsu,

Y.-R.

N.,

Ross,

Res. (1985)

N. (1981)

L.,

AND

Sci. USA

Biophys.

82,

K., KOYAMA,

M.,

K. (1986) J. Immud

AND

K., ONDA, KAKINUMA,

Res. Cwmmun.

H., A. 130,

Anal.

M. (1976)

Biochem.

R. J., KUTNY, R. M., PANICO, R., AND CHOWDHRY, V. (1984)

Acad

Methods 72,248-254. M.,

MORRIS,

Proc.

NatZ.

Sci. USA 81.6486-6490.

20. COHEN,

F. E., KOSEN, P. A., KUNTZ, I. D., EPSTEIN, L. B., CIARDELLI, T. L., AND SMITH, K. A. (1986)

Science

234,349-352.

21. PROVENCHER, chemistry 22.

WANG,

S. W.,

AND

GLOCKNER,

Bie

J. (1981)

20,33-37.

A., Lu,

S., AND

MARK,

D. F. (1984)

Science

224,1431-1433.

Nu-

LIANG,

S.-M.,

ALETT,

24. KITANO,

THATCHER,

B. (1986) K., FUJIMOTO,

D. R., LIANG,

J. Biol

Chem.

S., NAKAO,

AND

M., WATANABE,

J. Biotechnol.

5, 77-

25. TIMASHEFF, S. N. (1970) in The Enzymes P. D., Ed.), Vol. II, pp. 371-443, Academic New York.

(Boyer, Press,

T., AND

NAKAO,

Y. (1987)

C.-M.,

261,334-337.

86.

692-699. 6. LIANG, S.-M., ALLET, B., ROSE, K., HIRSCHI, M., LIANG, C.-M., AND THATCHER, D. R. (1985)

Biochem.

A. H.,

12.

23.

T., KAWAHARA,

SUGINO,

BRASNEW,

D. A., RUSCETTI,

(1976)

18.

J., LIN, N., HARKINS, R. N., W. H., Ross, M. J., FODGE,

G., AND

MORGAN,

LAI,

260,14435-14439.

D., PRENDER,

K. (1986)

93,157-165.

L. J., ESTELL,

4. ALSON, K. C., FENNO, SNIDER, C., KOHR,

J. M.,

(Williamson, R., Ed.), ademic Press, London. 11.

KOTHS,

STON, F. A. 0. (1985) EMBOJ. 4, ‘775-780. HARRIS, T. J. R. (1983) in Genetic Engineering

H.-F.,

LEVINE, H. L., SLINKER, B., FIELDS, F., M. J., AND SHIVELY, J. (1981) J. Interferon 1,381-390.

J. Biol

A., AND

359,391-402.

AND TSUKAMOTO,

1. STAEHELIN, T., HOBBS, D. S., KNUG, C.-Y., AND PESTKA, S. (1981) J. Biol. 9750-9754.

T., ALTON,

R., BOOSMAN,

J. Chromatogr.

II.

3. ARAKAWA,

199

INTERLEUKIN-2

7. LAHM, H.-W., AND STEIN, S. (1985) J. Chromatogr. 326,357-361. 8. KUNITANI, M., HIRTZER, P., JOHONSON, D., HAL-

REFERENCES

R., PERRY,

HUMAN

14. YAMADA, MURA,

ACKNOWLEDGMENTS

2. WETZEL,

ACTIVE

J. 229,429-439.