A novel carbodiimide coupling method for synthetic peptides

Journal oflmmunological Methods, 129 (1990) 119-125

119

Elsevier JIM05552

A novel carbodiimide coupling method for synthetic peptides Enhanced anti-peptide antibody responses Carla Deen, Eric Claassen, Koen Gerritse, N e t t y D. Zegers and W i m J.A. Boersma T.N.O. Medical Biological Laboratory, POB 45, 2280 AA Rijswijk, The Netherlands

(Received18 August 1989,accepted2 January 1990)

Coupling of peptides to immunogenic protein carriers is required for the generation of anti-peptide antibody responses. Carbodiimides are hetero bi-functional coupling reagents that are utilized for coupling reactions through carboxyl and amino groups. The procedures generally used for carbodiimide coupling of peptides and proteins result in conjugates which generate immunodominant antibody responses directed against the neodeterminants on the carrier protein. These determinants are induced by the reaction of carrier and/or peptide with the coupling agent. We have investigated the potential inhibiting effeet of an imidazole intermediary on the formation of unwanted neodeterminants during carbodiimide coupling. The serum antibody responses elicited with the peptide-protein conjugates produced were evaluated in ELISA. We have modified and improved the coupling with a watersoluble carbodiimide (EDC) in such a way that a high response to the coupled peptide is obtained in association with negligible levels of anti-neodeterminant antibodies. Key words: Synthetic peptide; Covalentcoupling; Neodeterminant; Immunization; Antibodyresponse

Introduction

Both short (10-12 amino acids) and longer synthetic peptides (SP) generally require coupling to an immunogenic protein carrier molecule in order to elicit a specific anti-peptide antibody response (De Weck, 1974). Such a response is determined by the space filling structure of the peptide which in turn is influenced by the physicochemical character of the peptide in relation to

Correspondence to: WJ.A. Boersma, T.N.O. Medical Biological Laboratory, POB 45, 2280 AA Rijswijk, The Netherlands. Abbreviations: EDC, 1-ethyl-3-(dimethylaminopropyl) earbodiimide; MBS, m-mal~'rfidobenzoyI-N-hydroxysuecfnimide ester; COG, chicken gamma globulin; BSA,bovineserum albumin; SP, syntheticpeptide.

its micro-environment, e.g., the protein to which it is coupled (Briand et al., 1985; Boersma et al., 1988). The localization and orientation of the peptide to the protein can be influenced by the choice of the coupling method. Amino acid-specific coupling methods have been described for cysteine (Carlsson et al., 1978; King et al., 1978), using mostly maleimides (Yoshitaki et al., 1979) or free - S H (Carlsson et al., 1978) groups in covalent interactions. Other coupling methods that are frequently employed make use of e.g., free amino groups with glutaraldehyde or N-hydroxy succ'mimide as reactive groups. When peptides are employed for immunization to produce antibodies reactive with the native protein this sometimes requires the use of specific coupling methods to avoid certain residues (cysteines) being affected in the coupling procedure.

0022-1759/90/$03.50 © 1990 ElsevierSciencePublishers B.V. (BiomedicalDivision)

120 It may then be advantageous to have an efficient alternative coupling method available with a reagent that does not have the disadvantages of glutaraldehyde (unwanted crosslinking) or the Nhydroxy succinimides (lack of stability). Carbodiimides are hetero bi-functional coupling reagents that are mainly used for coupling - C O O H and - N H 2 groups (Goodfriend et al., 1964; Kurzer et al., 1967; Yamada et al., 1981). Conjugates prepared by the methods and procedures generally used for carbodiimide coupling of peptide and protein immunogens lead to weak anti-peptide antibody responses. The responses are predominantly directed to the immunodominant coupling or neodeterminants on the carrier protein (Briand et al., 1985). This results in relatively poor anti-peptide antibody responses and may explain why carbodiimide coupling has until now found only limited application in the production of SP conjugates. The immunodominant responses are probably induced by the acylurea adducts formed when the protein or ligand reacts with the coupling agent (Briand et al., 1985) and lead to the high titers against carrier neodeterminants that we have previously observed (Boersma et al., 1988). This phenomenon may be a direct result of the reaction conditions that are used for carbodiimide coupling, e.g., the rate of hydrolysis in a water solution negatively influences the efficiency of the covalent linkage between protein and peptide. For water-soluble carbodiimides the coupling conditions described and the ratio of ligand and carrier as well as the ratio of carbodiimide per carrier and/or ligand show a wide variation in a number of publications (Goodfriend et al., 1964; Ichimori et al., 1985; Staros et al., 1986; Posnett et al., 1988). Nevertheless the procedures generally induce marked anti-neodeterminant responses. In this paper we discuss the application of a modified and improved method for peptide coupling with carbodiimide using the water-soluble 1-etliyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and allowing the reaction to take place in an imidazole buffer. Imidazole has been used for carbodiimide coupling of nucleotides to protein by Chu et al. (1983) and more recently Cliollet and Kawashima (1985) investigated the role of imidazoles in the coupling of biotin to nucleotides. The intermediate product

formed in the reaction between the activated ligand or protein and imidazole seems to largely prevent the formation of the immunodominant acylurea adducts of the carbodiimides in the direct reaction with the ligand or protein. We have adapted the imidazole mediated carbodiimide coupling for the formation of peptide carrier conjugates in such a way that the specific antibody response to the ligand is considerably improved and the response to the neodeterminants is considerably reduced.

Materials and methods

peptides Peptides were synthesized using the Merrified solid-phase technique (Merrifield, 1963) with an automated peptide synthesizer: SAM-II, Biosearch, San Rafael, CA, U.S.A. The synthetic peptides (SP) were purified by chromatography on Sephadex G-15 and reverse phase HPLC as described in detail elsewhere (Van Denderen et al., 1989). Four different model peptides were used to evaluate the conditions of the carbodiimide coupling with respect to their effect on immunological responsiveness to the peptides. SP39 is a peptide derived from the hinge region of human IgG1 (EPKSCDKTHICPPCPA). SPill (APPVAGGPSVC) and SP37 (VVTVPSSNFGTQTYTCN) are peptides derived from, respectively, the CH2 and C H1 regions of human IgG2. Finally SP42b is a common sequence of Staphylococcus aureus enterotoxins B and C (CKFDQSKYLMMYNNDK). In the case of SPIII and SP42b a cysteine was added to the C terminus and the N terminus respectively to permit amino acid-specific coupling with m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS). Conjugation Keyhole limpet hemocyanin (KLH) (Calbiochem, 374811, San Diego, CA, U.S.A.), chicken gamma globulin (CGG) (Sigma, G-6516, St. Louis, MO, U.S.A.) and bovine serum albumin (BSA) (Sigma, A-9647, St. Louis, MO, U.S.A.) were used as immunogenic protein carriers.

121 The carbodiimide chosen for the coupling experiments was the water-soluble 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) (Sigma, E-7750, St. Louis, MO, U.S.A.) because of the expected low rate of hydrolysis in aqueous solution (Goodfriend et al., 1964). Of the various imidazole compounds compared (imidazole, N-methyl-imidazole and 2-methyl-imidazole) N-methyl-imidazole was selected on the basis of experience with coupling nucleotides to protein (unpublished results). The effect of Nmethyl-imidazole as a coupling medium was compared to the standard coupling conditions for EDC. In the regular coupling procedure, subsequently referred to as RC, the carrier (10 mg. m1-1) and the peptide (20 mg. m1-1) were dissolved in a mannitol buffer (0.35 M) pH 5.0. At room temperature EDC (various doses, see results section) in the same mannitol buffer was added dropwise with stirring (20 rain). The mixture was stirred overnight at 4 ° C. In the modified coupling procedure, referred to as MC, 0.5 M N-methyl-imidazole pH 6.0 (Aldrich, M5, 083-4, Brussels, Belgium) was used to dissolve the protein and peptide. After the addition of EDC the mixture was stirred for 30 rain at room temperature followed by dialysis (MW cut off 10,000) against phosphate-buffered saline (PBS, 0.01 M, pH 7.2) at room temperature. For coupling peptide specifically via cysteine, m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) (Pierce, 22310) was used as described in detail elsewhere (Boersma et al., 1988). Briefly, MBS coupling was performed by mixing MBS (200 mol/mol carrier protein) and the carrier on ice in PBS. After removal of the excess MBS the SP (100 mol/mol carrier) was added and the mixture stirred for 30 rain at room temperature.

Immunization Female BALB/c mice, 12-20 weeks of age, bred and raised in our Institute, were injected twice intraperitoneally with a 4 week interval using various doses of the conjugates. Groups of three mice were used for each dose. Alumina gel (Haaijman et al., 1988) or a water in oil emulsion (Boldaout et al., 1981) was used as adjuvant.

Response analysis Anti-peptide antibody responses in the serum were evaluated using an ELISA procedure. The sera were taken 7 and 21 days after each immunization. The ELISA procedure was performed as described in detail elsewhere (Boersma et al., 1988; Haaijman et al., 1988). Briefly, wells of 96 well polyvinyl chloride plates (Falcon 3911, Becton and Dickinson, Oxnard, CA, U.S.A.) were coated with antigen (synthetic peptides and conjugates of different peptides with an irrelevant carrier) (5-10 #g.m1-1) in PBS. After washing, post-coating with gelatine (1 mg. ml-l) and washing again the plates were incubated with different dilutions of the mouse serum samples. Detection of bound antibodies was performed with peroxidase conjugated rabbit anti-mouse immunoglobulins (RAM/PO, Dakopatts P260, ITK Diagnostics, Glostrup, Denmark) and o-phenylenediamine (Eastman Kodak 107 8054, Rochester, NY, U.S.A.) as substrate. Absorbance was measured wlth a Titertek Multiskan reader (Flow Laboratories, Irvine, Scotland) at 492 nm.

Results

Using the regular (RC) coupling method with EDC, conjugates of chicken gamma globulin (CGG) and SPill were prepared. The molar ratio of SP and EDC to carrier was varied (25-100 mol SP/mol CGG, 50 mol EDC/mol SP) and mice were immunized with different doses of each conjugate. Similar experiments were performed using MBS as coupling agent. To evaluate the serum anti-SP responses, conjugates of SPIII with BSA as carder and MBS and EDC as coupling agents were used. Using these conjugates it was possible to discriminate between anti-peptide and anti-neodeterminant immune responses. After immunization with a CGG*SP (EDC) conjugate, anti-peptide responses were detected with the free peptides and with a BSA*SP (MBS) conjugate. The response to BSA*SP(EDC) comprised both antibodies to EDC-derived neodeterminants as well as anti-peptide antibodies and the anti-neodeterminant response was calculated by subtraction. Alternatively, using an EDC conjugate with an irrelevant peptide the response

122 IMMUNIZATION WITH SPill CONJJGATES: ANTI PEPTIDE RESPONSES

2~

IMMLI~tZAT ION WITH [

CGG- S,PlECK~)

CGG- SP(MBS)

(-q O~ I.J < ff O o3 <

0

5

10

50 5 DOSE OF ANTIGEN ~g)

BSA-SP(MBS)~

BSA-SP(EEX:::)I

10

50

,:1 S#

Fig. 1. Anti-peptide responses after immunizationwith regular peptide-protein conjugateswith EDC. Conjugates of CGG and SPill were made using regular protocols for EDC and MBS coupling. BALB/c mice were immunizedtwicewith an interval of 4 weeks with 5, 10 and 50 #g of each conjugate. Sera were collected 7 days after the second immunization and tested on ELISA plates coated with BSA conjugatesof the peptide (MBS and EDC) and free peptide (serum dilutions 1/200, mean of three individual sera is shown). Binding of antibodies was demonstrated with peroxidaseconjugated rabbit anti-mouse Ig, with o-phenylenediamineas a substrate. Absorbance was measured with a Titetek multiskan at 492 nm.

to EDC neodeterminants could be determined directly. However, a disadvantage of the latter approach is that differences in coupling behavior due to different residues in the irrelevant peptides may lead to confusing results. Fig. 1 shows typical serum-antibody responses (dilution 1/100) of mice immunized with one of the conjugates of SPill coupled to C G G with EDC (left) or MBS (right). The coupling conditions for EDC were 50 mol S P / m o l C G G and 50 mol E D C / m o l SP; for MBS coupling 200 mol M B S / m o l C G G were applied. Conjugates of BSA (MBS and EDC) with the same peptide (SPIII) were used for analysis. Immunization with various doses of the CGG*SPIII(EDC) conjugate resulted in serum antibodies that reacted strongly with BSA*SPIII(EDC) conjugates (Fig. 1) and B S A / EDC conjugates of irrelevant peptides (data not shown). A response to BSA*SPIII(MBS) or free peptide was detected only when high doses of the

C G G * S P I I I ( E D C ) conjugate were used for immunization. Conjugates of CGG*SPIII(EDC) produced with lower concentrations of EDC in the coupling medium also led to high responses to the neodeterminants introduced by EDC, but did not elicit a detectable response against the peptide (not shown). In contrast to the conjugates of C G G and SPIII produced with MBS a peptide-specific response was elicited even after immunization with low doses of CGG*SPIII(MBS) conjugates (Fig. 1). For the anti-peptide response a clear-cut doseresponse relationship was determined for the amounts of CGG*SPIII(MBS) conjugate immunized. Using the same coupling conditions (RC) the results could be reproduced with various other peptides coupled with EDC or MBS to different carrier proteins. Following immunization none of the EDC conjugates elicited significant antipeptide antibody responses. The results were independent of the carrier protein employed. For all peptides used, antibody responses were obtained using MBS coupling. Using the modified EDC coupling (MC) BSA conjugates of SPill were prepared at a constant molar ratio of p e p t i d e / c a r r i e r (100 mol SP/mol BSA) and using various molar ratios of EDC and peptides (0.05-50 mol E D C / m o l SP). The composition of the conjugates was determined with polyclonal mouse sera reactive with neodeterminant (anti-CGG*SPIII(EDC) and SPill (antiCGG*SPIII(MBS)) obtained in the experiment described above (Fig. 1). Serial dilutions of the MC coupled conjugates were coated onto ELISA plates. For the SPIII conjugates produced with 0.05-2.5 mol E D C / m o l SP the results are shown in Fig. 2. From this figure it is apparent that at the lower doses of E D C ( < 0.5 mol E D C / m o l SP) the response suggesting the presence of neodeterminants ('EDC') is negligible whereas, in the same conjugate, the anti-peptide response suggests a quite acceptable level of peptides 'coupled' to the carrier. However, in control experiments, we observed that after extensive dialysis of the conjugates prepared in this way, peptides could adhere to carrier proteins even in the absence of a coupling reagent. These 'pseudo-conjugates' were not able to evoke an anti-peptide response (Fig. 3).

123 FRELATIVE REPRESENTATION OF SP AND "NEOOETERMINANTS" IN SP CONJUGATES

IMMUNIZATION WITH KLH--SP CONJUGATES: ANTI PEPTIDE RESPONSES

ANTI "EDC" o 8 O B

2 ]/]

2.5 0.5 O.25 OD5

SP 42B

SP 39

SP

37

P] J

I

I

I

i

I

I

I

1

2

3

4

5

6

-

ANTIGEN"

DILUTIONS

0

1

[-2LOG]

2

3

(INIT.DIL.

4

5

I

5~/ml)

Fig. 2. Relative representation of SP and 'neo'determinants in SP conjugates. BSA-SPIII conjugates were made according to the modified protocol using different concentrations of EDC (0.5-2.5 tool EDC/mol SP). Serial dilutions of these conjugates were coated onto ELISA plates and the relative proportions of SP- and EDC-associated neodeterminant in each conjugate were determined using polyclonal mouse antibodies to SP and EDC (serum dilutions 1/200). Binding of antibodies was demonstrated as in Fig. 1.

MOUSE ANTI-SPIll RESPONSES UPON IMMUNIZATION WITH B S A * E D C * S P l I I 1.50 -

0", LU 0.75 < gO C0

.<

O.OO -

50

10

2.5 t # t 3 L EDC.,/

0.5

0.25

0.05

"

-

tvtOt. ~

Fig. 3. Specific serum anti-peptide antibody responses elicited with SP*EDC conjugates according to the modified coupling protocol with EDC. Mice were immunized twice at an interval of 4 weeks with 25 /Lg of BSA-SPIII conjugates made with decreasing concentrations of EDC (0.05-50 tool EDC/mol SP). Sera were taken 7 days after the second/mmunization and tested on ELISA plates coated with CCK3-SPIII(MBS) (serum dilutions: 1/100, mean of three individual sera is shown). Binding of antibodies was demonstrated as in Fig, 1.

,

.CP*EDC ~o.w.IvIBS

6

BSA-SPtMBS)~

~'*EDC

SP*t~I8

-

BSA-SI::~DC,)~

SP*EDC

~ae~i~

SP

Fig. 4. Mice were immunized twice with an interval of 4 weeks with 25 ttg of KLH-SP (MBS and EDC) conjugates of three different peptides (SP37, SP39 and SP42B). The EDC conjugates were produced according to the modified protocol. Sera were :..c~ollectedon day 21 after the second immunization and were tested on ELISA plates coated with BSA conjugates of the peptide with MBS and EDC and free peptide. (serum dilutions 1/300, mean of three individual sera is shown). Binding of antibodies was demonstrated as in Fig. 1.

Six conjugates of BSA*SPIII(EDC) produced at different EDC/SP ratios were injected into mice (25 /~g/mouse) as a water in oil emulsion. The anti-peptide resonse decreased rapidly with decreasing EDC/SP ratios (Fig. 3) even though peptides had been demonstrated in the conjugates (Fig. 2). From this it was concluded that a certain level of neodeterminants in the conjugates used for immunization was necessary to ensure an antipeptide response in vivo. We chose 50 tool EDC/mol SP at a 100 mol SP/mol carrier ratio for the production of antipeptide sera. The application of this protocol to a variety of peptides demonstrated that under the conditions chosen a significant anti-peptide response is indeed ensured despite the presence of neodeterminants in the conjugates (Fig. 4). The modified method for coupling with EDC is, in some cases, even superior to other methods and has already proven to be a satisfactory alternative for the coupling of SP39 with MBS. MBS-mediated conjugate formation led to conjugates that were totally unable to induce the production of anti-SP39 antibodies.

124 Discussion

We have shown that the covalent coupling of peptides using water-soluble carbodiimides can be very efficient for the linkage of (synthetic) peptides and carrier proteins in order to produce conjugates able to elicit peptide specific responses. The buffer conditions chosen for carbodiimide coupling (N-methyl-imidazole) lead to a decrease in the proportion of immunodominant neodet,-rminants that are introduced during the coupling. Using the present conditions the reaction intermediate formed (imidazolide) apparently inhibits the formation of acylureas. In addition the imidazole group provides conditions favorable to coupling (Goodfriend et al., 1964; Kurzer et al., 1967; Yamada et al., 1981). In spite of the presence of EDC dependent 'neodeterminants' in the MC protocol the coupling conditions in the presence of imidazole lead to efficient coupling and subsequent optimal antipeptide responses. This may be explained by the observation of Erlanger (1973) that substitution of only a proportion of the available coupling sites leads to optimal hapten specific responses. Using the RC method, the immunosuppressive effect of acyl-urea adducts to protein produced using the regular protocol may be caused by the high substitution rate of EDC. This would lead to repetitive structures on the protein surface which could decrease the level" and change the isotype of the anti-hapten response (De Weck, 1974). Immunological evaluation of responses to SPcartier conjugates was chosen since non-specific adherence of peptides to protein carriers, i.e., pseudo-conjugates, may interfere with the other (non-immunological, e.g., chemical) methods for determining the Iigand substitution rate in the coupling procedure. W e did not observe antipeptide responses with pseudo-conjugates. This confirms the necessity for a covalent linkage between carrier and ligand. We therefore did not undertake a quantitative determination of the substitution rate, e.g., by amino acid analysis because it is questionable whether it is possible to distinguish between covalent and pseudo conjugates. For the coupling of nucleotides, which can be readily monitored by UV spectroscopy, the reaction was about 5-20 times more efficient in N-

methyl-imidazole than in the buffers (mannitol) that are generally used for carbodiimide coupling (unpublished results). Carbodiimide coupling is dependent mainly on the presence of - N H 2 and - C O O H groups. This suggests that specific coupling and the degree of substitution is regulated by the composition of the amino acid residue represented in the peptide. The composition of SPill was such that only the amino- and carboxy-terminal residues were used for coupling. We showed that a specific and significant response could also be elicited using the MC protocol with the other peptides used in this model study. Only minor differences in immunogenicity were observed. On the question of amino acid composition governing coupling efficiency, similar observations have been made using N-hydroxy-succinimide (Staros et al., 1986). In conclusion, a method is described for the coupling of peptides of immunological interest. Selective use of this and other coupling methods may lead to more efficient use of peptides and a better understanding of the influence of peptide structure and orientation on the immunogenicity of conjugates.

References Boersma, W.J.A., Claassen, E., Deen, C., Gerritse, K., Haaaj-

man, J.J. and Zegers, N.D. (1988) Antibodies to short synthetic peptides for specific recognition of partly denatured protein. Anal. Chim. Acta 213, 187. Bokhout, B.A., Van Gaalen, C. and Van der Heyden, P.J. (1981) A selected water-in-oil emulsion: composition and usefulness as an immunological adjuvant. Vet. Immunol. lmmunopathol. 2, 491. Briand, J.P., Muller, S. and Van Regenmortel, M.H.V. (1985) Synthetic peptides as antigens: Pitfalls of Conjugation Methods. J. lmmunol. Methods 78, 59. Carlsson, J., Drevin, H. and Axen, R. (1978) Protein thiolation and reversible protein-protein conjugation N-succinimidyl 3-(2-pyridyldithio)propionate a new heterobifunctionalreagent. Biochem.J. 173, 723. Chollet, A. and Kawashima, E.H. (1985) Biotin-labeled synthetic oligodeoxyribonucleotides:chemical synthesis and uses as hybridization probes. Nucleic Acids Res. 13, 1529. Chu, B.C.F., Wahl, G.M. and Orgel, L.E. (1983) Derivationof unprotected polynueleotides.Nucleic Acids Res. 11, 6513. Davis, M.B. and Preston, J.F. (1981) A simple modified earbodiimide method for conjugation of small-molecular weight compounds to immunoglobulin O with minimal protein crosslinking. Anal. Biochem.116, 402.

125 De Weck, A.L. (1974) Low molecular weight antigens. In: M. Sela (Ed.), The Antigens, Vol. II. Academic Press, New York, chapter 3. Erlanger, B.F. (1973) Principles and methods for the preparation of drug protein conjugates for immunological studies. Pharmacol. Rev., 25, 271. Goodfriend, T., Levine, L. and Fasman, G. (1964) Antibodies to bradykinin and angiotensin: A use of carbodiimides in immunology. Science 1,14, 1344. Haaijman, J.J., Coolen, J., Deen, C., Krtise, C.J.M., Zijlstra, J.J. and Radl, J. (1988) Monoelonal antibodies directed against human immunoglobulins: Preparation and evaluation procedures. In: S.B. Pal (Ed.), Reviews in Immunoassay Technology. Vol. 1. MacMillan, London, p. 59. Ichimori, Y., Kurokawa, T., Honda, S., Suzuki, N., Wakimasu, M. and Tsukamoto, K. (1985) Monoclonal antibodies to human interferon-gamma: I Antibodies to a synthetic carboxyl-terminal peptide. J. Immunol. Methods 80, 55. King, T.P., Li, Y. and Kouchoumian, L. (1978) Preparation of protein conjugates via intermolecular disulfide bond formation. Biochemistry 17, 1499. Kurzer, F. and Douraghi-Zadeh, K. (1967) Advances in the chemistry of carbodiimides. Chem. Rev. 67, 107.

Merrifield, R.B. (1963) Solid phase peptide synthesis. I. The synthesis of a tetrapeptide, J. Am. Chem. Soc. 85, 2145. Posnett, D.N., McGrath, H. and Tam, J.P. (1988) A novel method for producing anti peptide antibodies. Production of site-specific antibodies to the T-cell antigen receptor ,8-chain. J. Biol. Chem. 4, 1719. Staros, J.V., Wright, R.W. and Swingle, D.M. (1986) Enhancement by N-hydroxysulfosuccinimide of water-soluble carbodiimide-mediated coupling reactions. Anal. Biochem. 156, 220. Van Denderen, J., Hermans, A., Meeuwsen, T., Troelstra, C., Zegers, N., Boersma, W., Grosveld, G. and Van Ewijk, W. (1989) Antibody recognition of the tumor-specific bcr-abl joining region in chronic myeloid leukemia. J. Exp. Med. 169, 87. Yamada, H., Imoto, T., Fujita, K., Okazaki, K. and Motomura, M. (1981) Selective modification of Aspartic acid-101 in lysozyme by carbodiimide reaction. Biochemistry 20, 4836. Yoshitaki, S., Yamada, Y., lshikawa, E. and Masseyeff, R. (1979) Conjugation of glucose oxidase from aspergillus niger and rabbit antibodies using n-hydroxysuccinimide ester of n-(4-carboxycyclohexyimethyl)-maleimide. Eur. J. Biochem. 101,395.