Chemical modification of erythropoietin: an increase in in vitro activity by guanidination

Biochimica et Biophysica Acta, 1038(1990) 125-129

125

Elsevier BBAPRO33604

Chemical modification of erythropoietin: an increase in in vitro activity by guanidination Rika Satake, Hiroyuki Kozutsumi, Makoto Takeuchi and Katsuhiko Asano Pharmaceutical Laboratory, Kirin Brewery Co., Maebashi, Gunma (Japan)

(Received16 August1989)

Keywords: Erythropoietin;Chemicalmodification;Guanidination Human recombinant erythropoietin (rHuEPO) was chemically modified with several group-specific reagents in order to study the role of each kind of amino-acid residue in its biological activity. Guanidination of the amino groups of the lysine residues yielded derivatives that showed higher activities in vitro than native rHuEPO, whereas amidination had no effect on the activity. By contrast, modification of the positive charges of the lysine residues to neutral or negative charges, such as in carbamylation, trinitrophenylation, acetylation or succinylation, caused a significant loss of rHuEPO activity. Chemical modification of other amino-acid residues, such as arginine and tyrosine residues or carboxyl groups, also led to loss of activity.

Introduction Erythropoietin is a glycoprotein hormone which regulates the number of peripheral erythrocytes [1,2,]. Recently, recombinant human erythropoietin (rHuEPO) was successfully produced in Chinese hamster ovary (CHO) cells [3], and is now available for clinical use. Over the last few years, the biological [4-6] and pharmacological [7,8] properties of rHuEPO have been extensively studied, and its clinical value has been clearly demonstrated [9,10]. However, little is known about the molecular mechanisms of the biological activity of rHuEPO. To identify the functional domains of rHuEPO, Sytkowski and Donahue [11] used anti-peptide antiserum against erythropoietin (EPO) and pointed out the importance of the region corresponding to the aminoacid residues 99-129, while Wang et al. [12] suggested the crucial role of the disulfide bridge between Cys-29 and Cys-31.

Abbreviations: rHuEPO, recombinant human erythropoietin;CHO, Chinese hamster ovary; TNBS, 2,4,6,.trinitrobenzenesulfonicacid; EDC, 1-ethyl-3-(3'-dimethylaminopropyl)carbodiimide;GDMP, 1guanyl-3,5-dimethylpyrazole; SDS-PAGE, SDS-polyacrylamidegel electrophoresis. Correspondence: KatsuhikoAsano,PharmaceuticalLaboratory,Kirin BreweryCo., Ltd.,1-2-2, Souja-machi,Maebashi,Gunma,371, Japan.

We have previously tried to determine the active domain of rHuEPO by partial proteolysis of the molecule, but no active peptide fragment has so far been found (Satake et al., unpublished data). In this report, we show that the activity of rHuEPO is sensitive to chemical modification of lysine, arginine and tyrosine residues, as well as carboxyl groups, and discuss the influence of these modifications, particularly changes in the net charge of the molecule, on the biological activity. Materials and Methods Materials

rHuEPO was highly purified to homogenity from the conditioned media of the CHO cell line expressing the human erythropoietin gene by ion-exchange, reversephase and gel chromatography. The purity was more than 99% by SDS-PAGE and in vivo specific activity was measured at 200 000 I U / m g protein, o-Methylisourea hydrogen sulfate, aminoguanidine nitrate, 2,3butanedione, phenylglyoxal and 1,2-cyclohexanone were purchased from Aldrich (Belgium). Potassium cyanate and tetranitromethane were obtained from Sigma (U.S.A.). 2,4-pentadione, ethylacetiimide hydrochloride, 2,4,6,-trinitrobenzenesulfonic acid (TNBS) and 1-ethyl3-(3'-dimethylaminopropyl) carbodiimide hydroehloride (EDC) were obtained from Wako Pure Chemical (Japan). 1-Guanyl-3,5-dimethylpyrazole (GDMP) was synthesized by the method of Bannard et al. [13] from aminoguanidine and 2,4-pentadione.

0167-4838/90/$03.50 © 1990 ElsevierSciencePublishersB.V.(BiomedicalDivision)

126

Chemical modification Guanidination. rHuEPO

was guanidinated with o-methylisourea [141 or G D M P [15]. The solution of rHuEPO (10 m g / m l ) was mixed with an equal volume of 1.0 M o-methylsourea (pH adjusted to 10.5 with NaOH). The reaction was allowed to proceed for 6 days at 4 o C, and at appropriate time intervals aliquots were removed and dialyzed against 1 mM HC1 to stop the reaction, rHuEPO was added at the concentration of 3 m g / m l in 0.7 M G D M P (pH adjusted to 9.9 with NaOH). The mixture was left for 5 days at 4 ° C and the reaction was subsequently stopped by adding HCI to a pH of 3.0, and the excess reagent was removed by dialysis against 1 mM HCI. Amidination. rHuEPO was amidinated by a modified method of Zollock et al. [16]. rHuEPO (2 m g / m l ) in 0.25 M borate buffer (pH 10.0) was mixed with an equal volume of 0.4 M ethylacetiimide in the same buffer. The reaction was allowed to proceed at room temperature for 4 h and was stopped by dialysis against water. Carbamylation. rHuEPO (1 mg/ml) was carbamylated with potassium cyanate by the method of Plapp et al. [14]. Trinitrophenylation. rHuEPO (1 m g / m l ) was trinitrophenylated with TNBS by the method of Plapp et al [14]. Acetylation. rHuEPO (2 m g / m l ) in 0.3 M phosphate buffer (pH 7.2) was incubated with an equal amount of acetic anhydride at 0 ° C for 1 h and the reaction was stopped by dialysis against water [17]. Succinylation. rHuEPO (2 m g / m l ) in 0.5 M NaHCO 3, (pH 8.0) containing 0.2 M NaC1, was incubated with a 15-fold molar excess of succinic anydride at 15 ° C for 1 h and the reaction was stopped by dialysis against water [18]. Modification of arginine residues. Arginine residues of rHuEPO were modified with 2,3-butanedione [19], 1,2,cyclohexanone [20] or phenylglyoxal [21], as described previously. Nitration. Tyrosine residues of rHuEPO were modified with tetranitromethane [22]. Modification of carboxyl groups, rHuEPO (1 mg/ml, pH adjusted to 4.5 with HCI) was modified at room temperature by 0.02 M EDC employing 1.0 M glycinamide [23]. Over a period of 60 min, aliquots were removed at appropriate time intervals and the reaction was stopped by the addition of 2.5 vol of 1.0 M sodium acetate (pH 4.75).

Analysis The number of modified amino-acid residues was determined principally by amino-acid analysis, using an Hitachi 835 amino-acid analyzer, after hydrolysis of the samples in 6 M HC1 in sealed evacuated tubes at l l 0 ° C for 24 or 48 h [14]. The number of free amino groups was determined with TNBS [24].

In vitro biological activity was measured by determining the incorporation of 59Fe into cultured rat bone marrow cells after incubation with samples [25] and in vivo biological activity was determined by the exhypoxic polycythemic mouse bioassay. [26]. The amount of sialylated oligosaccharides was determined by paper electrophoresis of oligosaccharides released from derivatives by hydrazinolysis [27]. The digestion of the derivatives (1 m g / m l ) with pepsin (2.5 # g / m l ) in 0.1 M acetate buffer (pH 4.5) was performed at 37 ° C for 2 h. At appropriate time intervals, aliquots were analyzed by SDS-PAGE, as described by Laemmli [28], using a 12.5% polyacrylamide gel. The derivatives (0.4 m g / m l ) were also digested with plasmin (0.1 U / m l ) in 0.01 M phosphate buffer (pH 7.0) containing 0.155 M NaC1 at 3 7 ° C for 5 h. Results

Derivatives of rHuEPO and their activities The derivatives of rHuEPO obtained in this study, the numbers of amino acids modified in each derivative, and their activities, are summarized in Table I. Guanidination of the lysine residues significantly increased the activity of rHuEPO in vitro. About seven

TABLE I

Derivatives of rHuEPO and their activities Amino

Dervative

Total No.

No. of

Activity % a

acid

or reagent

of residues

modified residues

in vitro

Lys

guanidino

Arg

7.3 b 8.0 c

260 240

50 56

amidino

6.9

100

n.d.

carbamayl

6.4

<1

n.d.

trinitrophenyl

4.8

<1

n.d.

acetyl

6.8

< 10

n.d

succinyl

4.8

<1

n.d.

2.0

< 20

n.d.

cyclohexanone

5.0

< 10

n.d.

phenylglyoxal

1.0

< 10

n.d.

butanedione

8

in vivo

12

Tyr

nitro

4

1.0 2.0

20 < 10

n.d.

Giu Asp

arnido

18

4.0 7.0

60 < 10

n.d. n.d.

a Relative to native rHuEPO taken as 100, calculated by the ratio of specific activities (2.2-105 I U / m g protein in vitro, 2.0.105 I U / m g protein in vivo). b Modified with o-methylisourea. ¢ Modified with GDMP. n.d., not done.

127 out of eight lysine residues were guanidinated with o-methylisourea, while all eight lysine residues were modified with G D M P . However, the extent of increased activity was almost the same for both derivatives. Amino-acid analysis showed that no other modification occurred in either of the guanidinated r H u E P O derivatives. Amidination of amino groups in r H u E P O led to derivatives containing about seven amidinated e-amino groups of lysine, as determined by amino-acid analysis. The derivatives retained the positive charges of t h e amino groups as did guanidinated derivatives, but the activity remained unchanged. To understand the role of the positive charges of lysine residues in the activity of rHuEPO, the positive charges were neutralized b y carbamylation, trinitrophenylation and acetylation, or were changed to negative charges by succinylation. About six e-amino groups of the lysine residues were carbamylated according to the detection of homocitrulline by amino-acid analysis. Trinitrophenylation and acetylation also led to derivatives containing five to seven modified e-amino groups, respectively. As shown in Table I, the in vitro activities of all these derivatives were either lost or were less than 10% of that of native rHuEPO. Succinylation of free e-amino groups also caused complete loss of in vitro activity. These results suggest that the positive charges of the lysine residues, though not fully identified, are important for the in vitro activity of rHuEPO. Neutralization of the positive charges of arginine residues of r H u E P O with 2,3-butanedione, 1,2-cyclohexanone or phenylglyoxal also caused a significant loss of in vitro activity, but no particular arginine residue essential for the activity of r H u E P O was identified in the molecule, r H u E P O was also inactivated by both the nitration of two tyrosine residues and the modification of seven carboxyl groups, whereas 60% of the activity remained after the modification of four carboxyl groups.

20G ~J

l ,

No, of modified Iltslne residues

Fig. 1. Effect of guanidination on the activity of rHuEPO, rHuEPO was incubated with o-methylisourea(O) or GDMP (A). Aliquots were removed at indicated times to measure the number of lysine residues guanidinated, and their in vitro activities were measured as described in Materials and Methods. In vitro activities (%) are calculated by the ratio of specific activities and given as a percent of the control. Each point is the mean 5=S.E. for three determinations.

G D M P , and both these derivatives showed the same increase in activity. Fig. 2 shows the dose-response curves of guanidinated and native r H u E P O for in vitro activity. 59Fe incorporation into bone marrow cells was enhanced by exposure of the cells (6.25.10 5 cells) to r H u E P O at concentrations from 1.0 to 1 0 0 . 1 0 -15 m o l / m l . Although enhancement was also elicited by guanidinated rHuEPO, its potency was 2- to 3-fold that of native rHuEPO. In addition, there was a significant difference in the slope of the dose-response curves between guanidinated and native rHuEPO. Bone marrow cells showed higher sensitivity to guanidinated r H u E P O than to native rHuEPO. On the other hand, the guanidinated deriva-

Characterization of guanidinated rHuEPO The synthesis of a guanidinated r H u E P O with a higher in vitro activity than native r H u E P O led us to characterize this derivative further. Fig. 1 shows that the activity of the derivative in vitro increased with an increase in the number of guanidinated lysine residues. N o significant difference was detected in the effect of guanidination reagents such as o-methylisourea and G D M P on the increased acitivity of rHuEPO. These results suggest that there is no single guanidinated lysine residue specifically involved in the activation of rHuEPO. On the other hand, a peptide-mapping study of both guanidinated derivatives suggests that the guanidination of Lys-152 is not important for the activation of r H u E P O (data not shown), because only Lys-152 was not guanidinated by o-methylisourea, whereas all of the lysine residues were guanidinated by

t

o

~000

i.

|

3~

I

I

1

10

rlm|~

I

100

I

I000

( xilFnmi/ml )

Fig. 2. Dose-response curve of guanidinated and native rHuEPO. 6.25-10 5 bone marrow cells were incubated with rHuEPO guanidinated with GDMP (O), containing all eight lysine residues modified, or native rHuEPO (o). After 24 h culture, 59Fe-incorporation into the cells was measured. Each point is the mean:t: S.E. for three determinations.

128 TABLE II

Amount of sialylated oligosaccharides obtained from guanidinated and native rHuEPO The amount of sialylated oligosaccharides was analyzed by paper electrophoresis as described in Materials and Methods. N, neutral; A1, mono-sialylated; A 2, di-sialylated; A3, tri-sialylated; A 4, tetrasialylated; and Atotal, A 1 + A 2 + A 3 + A 4. Fraction N

A1

A2

A3

A4

Atota I

Native rHuEPO

1.96

13.3

10.3

31.1

43.4

98.0

Guanidinated rHuEPO a

6.97

17.3

12.9

32.8

30.1

93.0

Guanidinated rHuEPO b

1.86

14.4

31.3

43.7

98.1

8.72

a Guanidinated with o-methyllisourea. b Guanidinated with GDMP.

tive showed only half the biological activity in vivo compared with native rHuEPO, as shown in Table I. This discrepancy in the in vitro and in vivo activities might have occurred because the conditions used in this assay were most suitable for the measurement of in vivo biological activity of both rHuEPO and human urinary EPO [29], but were not optimized for these derivatives. These derivatives may have different in vivo properties such as dose-response curves and speed of effect from native rHuEPO. It is well-known that the oligosaccharide chains of HuEPO are important for biological activity in vivo [30-33]. When terminal oligosaccharide chains of HuEPO are modified, the activity of HuEPO is abolished completely in vivo, but enhanced by 2- or 3-fold in vitro [31]. To check whether guanidination affects the terminal sialic acids of rHuEPO, N-linked oligosaccharide chains of guanidinated rHuEPO were removed by hydrazinolysis at 100 ° C for 16 h, and their charge distil-

O

2

4

InwlmtlN time (hrs) Fig. 3. Digestion of guanidinated and native rHuEPO by proteinase. rHuEPO guanidinated with G D M P (@) and native rHuEPO ( o ) were incubated with plasmin at 3 7 ° C and aliquots were removed at indicated times and analyzed for intact molecule content. The content of intact molecules is measured by densitometry on the gel as compared to the undigested control and given as a % of the control. Each point is the mean + S.E. for three determinations.

bution was analyzed by paper electrophoresis as described previously [27]. As shown in Table II, N-linked oligosaccharide chains of r H u E P O were separated into five fractions (N, A1-A4). The percentage of neutral sugar chains (N), as well as the ratio among acidic chains (A1-A4), was essentially the same between guanidinated and native rHuEPO. This result clearly indicated that guanidination did not alter the charge distribution of the sugar chains of rHuEPO. Therefore, the discrepancy in the in vivo and in vitro activities is not due to the removal of the terminal sialic acids of rHuEPO during guanidination. To investigate the possibility that guanidinated rHuEPO was more susceptible than native rHuEPO, in vivo, to plasma proteinases such as plasmin, guanidinated rHuEPO was digested with plasmin under the conditions described in Materials and Methods. Fig. 3, however, shows that guanidinated r H u E P O was more resistant to plasmin than native rHuEPO. Guanidinated rHuEPO was also resistant to other proteinases, such as pepsin (data no shown). Discussion In the present study we showed that the activity of rHuEPO was sensitive to chemical modifications of the lysine, arginine and tyrosine residues, as well as carboxyl groups. Among these, modification of the lysine residues with various lysine specific reagents resulted in a wide range of changes in the activity of modified rHuEPO. Modifications that changed the positive charges of the lysine residues to neutral or negative charges, such as carbamylation, trinitrophenylation, acetylation or succinylation, caused a substantial loss in the vitro activity, whereas amidination which left the total number of positive charges of lysine unchanged did not effect the activity. It is clear that the positive charges of the lysine residues are essential for the activity of rHuEPO in vitro. A most noteworthy finding in this study is the marked increase in the activity of guanidinated r H u E P O in vitro. The dose-response curve of guanidinated r H u E P O (Fig. 2) shows the possibility that there is a difference in the affinity to the receptor between guanidinated and native rHuEPO. A major difference between guanidinated and native r H u E P O is the size of the side chain at the lysine residue, suggesting that the larger side chain of the lysine residue is favorable for approach to its receptor. However, this is unlikely because amidinated rHuEPO, which has the same number of positive charges and almost the same size of side chains at the lysine residue as guanidinated rHuEPO, showed essentially the same activity as native rHuEPO. Therefore, it seems probable that the guanidino groups, together with their positive charges, play an important role in the interaction between the receptors and rHuEPO.

129

Acknowledgements T h e a u t h o r s w i s h to t h a n k M . K u b o t a , M. W a d a a n d D r . N . N a g a n o f o r in v i v o b i o a s s a y . Y. I n a g a k i is a l s o thanked for the synthesis of GDMP.

References 1 Graber, S.E. and Krantz, S.B. (1978) Annu. Rev. Med. 29, 51-66. 2 Spivak, J.L. and Graber, S.E. (1980) Johns Hopkins Med. J. 146, 311-320. 3 Lin, F.-K., Suggs, S., Lin, C.-H., Browne, J.K., Smalling, R., Egrie, J.C., Chen, K.K., Fox, G.M., Martin, F., Stabinsky, Z., Badrawi, S.M., Lai, P.-H. and Goldwasser, E. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7580-7584. 4 Egrie, J.C., Strickland, T.W., Lane, J., Aoki, K., Cohen, A.M., Smalling, R., Trail, G., Lin, F.-K., Browne, J.K. and Hines, D.K. (1986) Immunobioiogy 172, 213-224. 5 Browne, J.K., Cohen, A.M., Egrie, J.C., Lai, P.H., Lin, F.-K., Strickland, T., Watson, E. and Stebbing, N. (1986) in Cold Spring Harbor Symposium on Quantitative Biology 51 (Harris, R.G., eds), pp. 693-702, Cold Spring Harbor, New York. 6 Dessypris, E.N., Gleaton, J.H. and Armstrong, D.L. (1987) Br. J. Haematol. 65, 265-269. 7 Masunaga, H., Goto, K. and Ueda, M. (1987) Acta Haematol. Jpn. 50, 1119-1125. 8 Misago, M., Chiba, S., Tsukada, J., Kikuchi, M. and Eto, S. (1988) Acta Haematol. Jpn. 51,967-974. 9 Esehback, J.W., Egrie, J.C., Downing, M.R., Browne, K.J. and Adamson, J.W. (1987) N. Engl. J. Med. 316, 73-78. 10 Winears, C.G., Oliver, D.O., Pippard, M.J., Reid, C., Downing, M. and Cotes, P.M. (1986) Lancet ii, 1175-1178. 11 Sytkowski, A.J. and Donahue, A. (1987) J. Biol. Chem. 262, 1161-1165. 12 Wang, F.F., Kung, C.K.-H. and Goldwasser, E. (1985) Endocrinology 116, 2286-2292. 13 Bannard, R.A.B., Cassleman, A.A., Cockburn, W.F. and Brown, G.M. (1958) Can. J. Chem. 36, 1541-1549.

14 Plapp, B.V., Moore, S. and Stein, W.H. (1971) J. Biol. Chem. 246, 939-945. 15 Robinson, N.C., Neurath, H. and Walsh, K.A. (1973) Biochemistry 12, 414-420. 16 Zollock, D.T. and Niehaus, W.G., Jr. (1975) J. Biol. Chem. 250, 3080-3088. 17 Riordan, J.F. and Vallee, B.L. (1972) Methods Enzymol. 25, 494-499. 18 Habeeb, A.F.S.A., Cassidy, H.G. and Singer, S.J. (1958) Biochim. Biophys. Acta 29, 587-593. 19 Riordan, J.F. (1973) Biochemistry 12, 3915-3923. 20 Patthy, L. and Smith, E.L. (1975) J. Biol. Chem. 250, 565-569. 21 Werber, M.M., Moldovan, M. and Sokolovsky, M. (1975) Eur. J. Biochem. 53, 207-216. 22 Nestler, J.E., Chanko, G.K. and Strauss, J.F. (1985) J. Biol. Chem. 260, 7316-7321. 23 Carraway, K.L. and Koshland, D.E., Jr. (1972) Methods Enzymol. 25, 616-623. 24 Fields, R. (1971) Biochem. J. 124, 581-590. 25 Goldwasser, E., Eliason, J.E. and Sikkema, D. (1975) Endocrinology 97, 315-323. 26 Cotes, P.M. and Bangham, D.R. (1961) Nature 191, 1065-1067. 27 Takeuchi, M., Takasaki, S., Miyazaki, H., Kato, T., Hoshi, S., Kochibe, N. and Kobata, A. (1987) J. Biol. Chem. 263, 3657-3663. 28 Laemmli, U.K. (1970) Nature 227, 680-685. 29 Hammond, D., Shore, N. and Movassaghi, N. (1968) Ann. NY Acad. Sci. 68, 516-527. 30 Mcmullen, A.I. (1960) Nature 185, 102-103. 31 Dordal, M.S., Wang, F.F. and Goldwasser, E. (1985) Endocrinology 116, 2293-2299. 32 Goto, M., Akai, K., Murakami, A., Hashimoto, C., Tsuda, E., Ueda, M., Kawanishi, G., Takahashi, N., Ishimoto, A., Chiba, H. and Sasaki, R. (1988) Biotechnology 6, 67-71. 33 Takeuchi, M., Inoue, N., Strickland, T.W., Kubota, M., Wada, M., Shimizu, R., Kozutsumi, H., Takasaki, S. and Kobata, A. (1989) Proc. Natl. Acad. Sci. USA 86, 7819-7822.