Determination of the Extent of Protein Biotinylation by Fluorescence

A method was developed to determine the total amount of biotin present in biotinylated protein conjugates. Conjugates of bovine serum albumin, alkalin...
0 downloads 0 Views 133KB Size
Download PDF
Bioconjugate Chem. 1997, 8, 94−98

94

Determination of the Extent of Protein Biotinylation by Fluorescence Binding Assay Srivatsa V. Rao,† Kimberly W. Anderson,*,† and Leonidas G. Bachas*,‡ Department of Chemical and Materials Engineering and Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506. Received June 30, 1996X

A method was developed to determine the total amount of biotin present in biotinylated protein conjugates. Conjugates of bovine serum albumin, alkaline phosphatase, and horseradish peroxidase were used in this case study. The extent of biotinylation was determined by complete acid hydrolysis or by enzymatic digestion using proteinase K to release biotin from the biotinylated proteins, followed by sensitive detection of biotin using a coupled HPLC-binding assay system. This detection system is based on the enhancement of the fluorescence of streptavidin-FITC by biotin. The extent of biotinylation determined by this method was compared with the values obtained by a conventional colorimetric method that is based on the displacement of the dye 4-hydroxyazobenzene-2-carboxylic acid (HABA) from the binding sites of avidin. It was found that, because the described method determines the amount of liberated biotin after hydrolysis, it does not suffer from steric hindrance problems associated with the ability of biotin on a protein surface to displace HABA from avidin. Therefore, this method can provide a more accurate determination of the extent of biotinylation. It was also determined that the acid hydrolysis of the biotinylated protein was more effective in releasing the conjugated biotin compared to enzymatic digestion by proteinase K.

INTRODUCTION

Biotin has an extraordinarily high affinity for (strept)avidin with reported dissociation constants of ∼10-15 and ∼10-14 M for avidin and streptavidin, respectively (Green, 1990). Each (strept)avidin molecule has four binding sites that can bind biotin and biotinylated compounds. Further, a variety of biotinylating reagents are commercially available that can be used to attach biotin to proteins, nucleic acids, carbohydrates, and other biomolecules with relative ease (Wilchek and Bayer, 1990). This has led to an explosive growth in the use of biotin(strept)avidin systems in various biotechnological, analytical, and therapeutical applications (Diamandis and Christopoulos, 1991; Wilchek and Bayer, 1988). Biotinylation of proteins is generally achieved by an amidation reaction in which the free amino groups on lysine residues and the N terminus of proteins are reacted with an activated biotin derivative that is typically an N-hydroxysuccinimide (NHS) ester (Bayer and Wilchek, 1990). Sometimes, derivatives that introduce a spacer between the biotin and the protein surface are used to increase the accessibility of biotin toward (strept)avidin (Barbarakis et al., 1993a). Knowledge of the degree of biotinylation is of importance in the characterization of biotinylated proteins for use in optimized binding assays, bioreactors, and biosensors. Currently known methods for determining the extent of biotinylation of proteins can be broadly grouped into two categories: those that measure the accessible biotin on the protein, and those that measure the total biotinylation. 4-Hydroxyazobenzene-2-carboxylic acid (HABA) titration is one of the most commonly used methods to determine the extent of biotinylation (Bayer and Wilchek, 1990). In this case, HABA binds to avidin to give an absorption maximum at 500 nm. When biotin or biotin* Authors to whom correspondence should be addressed. † Department of Chemical and Materials Engineering. ‡ Department of Chemistry. X Abstract published in Advance ACS Abstracts, December 15, 1996.

S1043-1802(96)00080-8 CCC: $14.00

ylated proteins are added, biotin displaces HABA from avidin and the absorbance at 500 nm is reduced. The decrease in absorbance is then used to determine the extent of biotinylation. Because this method is based on an absorbance measurement, HABA titration suffers from low sensitivity. As a result, amounts typically in the order of several nanomoles of protein are required for the determination of the degree of biotinylation. More sensitive methods for biotin determination should reduce the amount of biotinylated protein sacrificed for this purpose. Der-Bailian et al. (1990) developed a method to determine the degree of biotinylation that is based on quenching of the natural fluorescence of avidin by biotin moieties on the biotinylated protein. This method, though simple, sensitive, and reproducible, suffers from steric hindrance effects (i.e., not all biotin on a protein can simultaneously bind to avidin) and thus underestimates the extent of biotinylation. Shah et al. (1994) have developed a different method that is based on the competition between fluorescein-labeled biotin and the biotinylated protein toward an anti-biotin monoclonal antibody. The fluorescence polarization of the fluorescein-labeled biotin was measured and correlated to the extent of biotinylation. This method also suffers from steric hindrance effects that lead to underestimation of the extent of biotinylation. However, the two techniques mentioned above may be useful in determining the amount of functional (i.e., available for binding) biotins on proteins, although there still may be some questions raised about the proper calibration of the methods; that is, biotin cannot be used as a standard in these assays because it may induce a different fluorescence signal than the biotinylated protein (Mei et al., 1994). Smith et al. (1991) have reported a different method for the determination of the degree of protein biotinylation. This requires biotinylation with an -aminohexanoic acid derivative of biotin. After acid hydrolysis, amino acid analysis yields the number of aminohexanoic acids per protein, which is equal to the degree of biotinylation of the protein. This method has picomolar sensitivity, but can be used only if the bio© 1997 American Chemical Society

Technical Notes

tinylated residue has incorporated an -aminohexanoic acid spacer arm. Consequently, despite the extensive use of biotinylated proteins in a variety of applications, the literature on their characterization is sparse (Kurosky et al., 1993; Miles and Garcia, 1995; Miller et al., 1994; Yem et al., 1989), especially in terms of a general method that determines the total number of biotin moieties attached per protein molecule. Therefore, it is desirable to develop a new method that allows the accurate determination of the degree of biotinylation. In this paper, we describe a method developed in our laboratory for the determination of the amount of biotin present in biotinylated proteins. This method is based on the complete acid hydrolysis of the biotinylated protein, followed by separation of the released biotin by HPLC, and determination of the amount of biotin present using a highly sensitive postcolumn reaction detection system that is based on a fluorescence binding assay. This method was also compared to an alternative hydrolysis approach that involved enzymatic digestion of the biotinylated protein using proteinase K. EXPERIMENTAL PROCEDURES

Reagents. Biotin, horseradish peroxidase (HRP) type VIA, biotinylated bovine serum albumin (b-BSA), biotinamidocaproyl-labeled bovine serum albumin (bcapBSA), biotinylated alkaline phosphatase (b-AP), avidin, N-hydroxysuccinimidobiotin (BNHS), sodium dodecyl sulfate (SDS), and proteinase K were all purchased from Sigma (St. Louis, MO). HABA and N,N-dimethylformamide (DMF) (ACS reagent grade), were obtained from Aldrich (Milwaukee, WI). Tris[hydroxymethyl]aminomethane (Tris) was from Research Organics (Cleveland, OH) and ethylenediaminetetraacetic acid (EDTA) from Mallinckrodt (St. Louis, MO). Streptavidin-FITC was purchased from Vector Laboratories (Burlingame, CA). Protein concentrations were determined according to the BCA protein assay (Pierce, Rockford, IL). Deionized water (Milli Q water purification system; Millipore, Bedford, MA) was used in the preparation of all solutions. HPLC with Postcolumn Reaction Detection. The experimental setup used in this work consisted of a Rainin (Woburn, MA) HPLC system, which was interfaced with a Macintosh computer (Apple Computer, Cupertino, CA). The system included a Rainin Rabbit solvent-delivery system and a Rheodyne (Berkeley, CA) Model 7125 injector with a 20-µL sample loop. The samples were separated using a reversed-phase 5 µm Microsorb C18 column (250 × 4.6 mm i.d.) (Rainin). An 80:20 (v/v) mixture of phosphate buffer, pH 7.0, and methanol was used as the mobile phase at a flow rate of 0.4 mL/min. The effluent stream from the HPLC column was mixed with the reagent stream containing 2 mg/L streptavidin labeled with fluorescein isothiocyanate (streptavidinFITC) prepared by diluting a stock solution (purchased as 1 mg/mL solution) with 100 mM phosphate buffer, pH 8.4. The reagent solution was pumped by an ISCO (Lincoln, NE) Model LC-2600 syringe pump at a flow rate of 1 mL/min and was merged with the column effluent through a tee-connector followed by a 10.0-m knitted open-tubular (KOT) reactor made from PTFE tubing (0.5 mm i. d., 14-mm helix diameter) (Przyjazny et al., 1993). The binding of biotin to streptavidin-FITC resulted in an enhancement of fluorescence intensity (Hentz and Bachas, 1995). Detection of the change in fluorescence was carried out by using a Fluorolog-2 spectrofluorometer (SPEX Industries, Edison, NJ) with a µ-fluorescence flow cell (20-µL cell volume; NSG Precision Cells, Farmingdale, NY). The excitation was set at 495 nm, and the

Bioconjugate Chem., Vol. 8, No. 1, 1997 95

Figure 1. Calibration plot for biotin using the streptavidinFITC-based postcolumn reaction detection system.

emission was monitored at 518 nm. The excitation and emission slit widths were each set at 2 mm. The SPEX spectrofluorometer was operated in the photon-counting mode. A calibration plot for biotin was prepared daily using biotin standards, and the corresponding peak areas were employed to estimate the amount of biotin present in each of the samples. A typical calibration plot is given in Figure 1. Biotinylation of HRP. For the preparation of biotinylated HRP (b-HRP), HRP was dissolved in 100 mM sodium hydrogen carbonate solution, adjusted to pH 8, to obtain a 10 mg/mL solution. Sufficient volume of a BNHS solution (100 mM BNHS in DMF) was added to 100 µL of an HRP solution to achieve an initial molar ratio of BNHS to HRP of 100. The biotinylation reaction was carried out in Reactivials (Pierce) for 3 h at room temperature. The reaction mixture was then dialyzed against phosphate buffer saline to remove unreacted BNHS. Acid Hydrolysis of Biotinylated Proteins. The biotinylated proteins were hydrolyzed with 6 M HCl in an evacuated vacuole at 110 °C for 18-24 h and dried in a vacuum centrifuge (Savant Instruments, Hickville, NY). The samples were then dissolved in a known amount of deionized water and analyzed for the amount of biotin present. Enzymatic Digestion of Biotinylated Proteins. The biotinylated proteins were digested with proteinase K and dissolved in 100 mM Tris-HCl buffer, pH 7.8, containing 5 mM EDTA and 0.5% (v/v) SDS. The molar ratio of proteinase K to biotinylated protein was 1:25 or 1:50. The digestion was carried out for 3 or 24 h at 37 °C. In all cases, the corresponding nonbiotinylated proteins were also processed under similar conditions and used as control samples. The digestion reaction was terminated by boiling the reaction mixture for 15 min. The samples were then used for biotin determination. RESULTS AND DISCUSSION

The HABA titration method provides an estimate of the accessible biotin but not the true extent of biotinylation of a protein, because some of the biotin moieties on the protein may not be accessible for binding to avidin as shown in Figure 2. In scenario A, where both the biotins are accessible to avidin, the HABA titration gives an accurate value of the extent of biotinylation. However, erroneous results may be obtained with highly biotinylated proteins because of possible cross-linking between

96 Bioconjugate Chem., Vol. 8, No. 1, 1997

Rao et al.

Figure 3. Separation and detection of free biotin released by a 24-h proteinase K digestion from b-HRP (moles of b-HRP/mole of proteinase K ) 25).

Figure 2. Depending on the nature of biotinylated protein, the HABA titration may give erroneous results: (A) degree of biotinylation is 2 as determined by HABA titration; (B) degree of biotinylation is determined to be 6; (C) degree of biotinylation is determined to be 4 for the same conjugate as in (B). Av, B, P, and H represent avidin, biotin, protein, and HABA, respectively.

avidin and biotinylated protein molecules (as a result of the availability of four binding sites on avidin) and/or steric hindrance effects, all of which may prohibit some of the biotin moieties on the surface of the protein from binding to avidin. The effect of steric hindrance is illustrated in Figure 2 (parts B and C), where for the same conjugate the accessible biotin in scenario B is 6 and in C is 4. This is also true for other assays that are based on the interaction between the biotinylated protein and avidin, streptavidin, or anti-biotin antibodies (DerBailian et al., 1990; Shah et al., 1994). Therefore, the HABA titration and the other methods that use intact biotinylated protein typically underestimate the extent of biotinylation and may not even be accurate at calculating the functional biotins on the protein molecule because of steric/cross-linking effects. Further, as mentioned earlier, additional questions may be raised about the ability of some of these methods to determine the degree of functional biotin on proteins, because free biotin and biotinylated protein have a different effect on the intrinsic fluorescence of avidin (Mei et al., 1994). Bayer and Wilchek (1990) have reported that one of the ways to circumvent the problems described above is to digest the biotinylated protein with proteinase K and determine the amount of free biotin released. This should give the true degree of biotinylation. As demonstrated in the present study (vide infra) and depending on the protein, the enzymatic digestion may lead to incomplete hydrolysis of the biotinylated protein and thus inaccurate results. To determine the total number of biotin moieties attached per protein molecule, a highly sensitive method was developed in our laboratory, which is based on complete hydrolysis of the biotinylated protein followed by a fluorescence binding assay that monitors the enhancement of fluorescence of streptavidin-FITC by biotin. Complete hydrolysis of the protein is necessary because biotin and biotinylated peptides could induce different

fluorescence enhancement to streptavidin-FITC (Barbarakis et al., 1993b), which may give erroneous results for the degree of biotinylation. By coupling a postcolumn reaction detection system based on this fluorescence binding assay to a HPLC separation, it is possible to verify the complete hydrolysis of the protein. It should be noted that the method has no interferences from amino acids that are not biotinylated, because these compounds do not induce a change in the fluorescence of streptavidin-FITC. The coupled HPLC-binding assay system also has good detection limits (4 × 10-13 mol of biotin), which reduces the amount of biotinylated protein needed for the assay. Proteinase K digestion and acid hydrolysis were investigated as methods for the release of biotin from the biotinylated proteins. It was reported earlier that, in about 3 h, proteinase K digests protein completely (Bayer and Wilchek, 1990; Ebeling et al., 1974). It was found, however, that 3 h was not enough to completely digest b-HRP, and therefore longer digestion times were used for this purpose. The chromatogram of the digest after a 24-h hydrolysis with proteinase K is shown in Figure 3, where biotin elutes out at 15.2 min (the hold-up time is 10.8 min). These data, in conjunction with a calibration plot for biotin (see Experimental Procedures), were used to calculate the extent of biotinylation, which was found to be 1.8 biotins per HRP molecule. Given that biotinylation was performed with a 100-fold excess of BNHS, this value of the degree of biotinylation is consistent with previous observations reporting that only 2 of the 7 amino groups of HRP (6 lysines and the N terminus) are accessible for modification by NHS esters (Paek et al., 1993; Zaitsu et al., 1992). The small peak at ∼19 min is most probably due to a biotinylated peptide and suggests incomplete hydrolysis of b-HRP even after a 24-h digestion with proteinase K. As anticipated, the chromatograms corresponding to the hydrolysis product of the nonbiotinylated proteins had no peaks, which indicates the selective nature of the postcolumn reaction detection system used toward the biotin moiety. As shown in Figure 3, prolonged digestion of b-HRP with proteinase K does not always give complete hydrolysis. This is more pronounced in Figure 4 , which shows the chromatogram obtained when b-BSA was digested for 3 h with proteinase K in amounts that were 25-fold less than b-BSA on a mole-to-mole basis. The additional peaks at retention times longer than that of biotin (biotin elutes at ∼15 min) indicate that the hydrolysis of the protein was incomplete. It should be mentioned that the

Technical Notes

Bioconjugate Chem., Vol. 8, No. 1, 1997 97 Table 1. Comparison of Different Methods Used To Estimate the Extent of Biotinylation of b-HRP and b-BSA method of determination HABA titration after enzymatic digestion with proteinase K, 3 h after enzymatic digestion with proteinase K, 24 h after acid hydrolysis

Figure 4. Separation and detection of free biotin released by a 3-h proteinase K digestion from b-BSA (moles of b-BSA/mole of proteinase K ) 25).

Figure 5. Separation and detection of acid-hydrolyzed b-BSA.

postcolumn reaction detection system used is highly selective and only detects biotin or biotinylated compounds. Thus, the additional peaks at 18.4 and 19.8 min are most probably due to biotinylated peptides. It was also found that the amount of proteinase K plays a role in the extent of hydrolysis and that an increase in the amount of proteinase K used causes an increase in the amount of biotin released. However, it was not possible to eliminate the additional chromatographic peaks by altering the amount of proteinase K used and/or the digestion time. The additional peaks were persistent even after 24 h of digestion with proteinase K. Therefore, proteinase K is not the ideal reagent to cause complete release of biotin, and this was verified as shown in Figure 4 by using the coupled HPLC-fluorescence system. Hydrolysis of proteins with concentrated HCl is commonly used to achieve complete digestion of proteins prior to amino acid analysis. Given the shortcomings of the proteinase K digestion, acid hydrolysis in evacuated vacuoles was performed on b-BSA, b-HRP, and b-AP. When b-BSA, b-HRP, and b-AP were analyzed for biotin, the chromatograms showed a single peak due to biotin. Figure 5 depicts a typical chromatogram of an acidhydrolyzed biotinylated protein. The values of the degree of biotinylation obtained by this method are higher than those determined by HABA titration (Table 1). The HABA titration gave degrees of biotinylation of 8.6 for b-BSA and 1.9 for b-HRP, while it was found by the method described above using acid hydrolysis combined with fluorescence binding assay that the ratios of (moles

b-HRP b-BSA 1.9 0.8 1.8 1.9

8.6 10.1 10.4 11.8

of biotin)/(mole of b-BSA) were 11.8 and 1.9, respectively. In the case of biotinylated alkaline phosphatase the chromatogram of the acid-hydrolyzed product also showed a single peak that corresponds to biotin. From these data the degree of biotinylation was estimated to be 6.2, while the enzymatically digested sample of the same conjugate gave a degree of biotinylation of 4.5. Because it is based on the determination of the amount of free biotin released from a biotinylated conjugate, the proposed method does not suffer from the bane of steric hindrance. Biotin and biotinylated analogs when reacted with streptavidin-FITC enhance the fluorescence intensity by different amounts (Barbarakis et al., 1993b). Therefore, complete hydrolysis, as reflected in a single peak for free biotin when using the HPLC-fluorescence detection method, is important for the accurate determination of biotin present and, thus, the degree of biotinylation. In the case of enzymatic digestion, the digest contains a few biotinylated peptides along with the released free biotin (e.g., see Figure 4). Therefore, enzymatic digestion does not provide accurate results. Biotinylating reagents that incorporate a spacer arm between biotin and the protein have also been used to reduce steric hindrance in the interaction with strept(avidin) (Barbarakis et al., 1992; Wilchek and Bayer, 1990). In that regard, a conjugate of BSA with a longchain biotin derivative (bcap-BSA) was studied to compare the proposed method with the HABA titration. The HABA titration gave a degree of biotinylation of 10.9, while the acid hydrolysis followed by the HPLCfluorescence detection gave a degree of biotinylation of 12.2. Thus, the HABA titration underestimates the degree of biotinylation of b-BSA (data discussed above) and bcap-BSA by 27% and 9.9%, respectively. The HABA titration has less relative error with bcap-BSA than with b-BSA, which is not surprising given the presence of the spacer arm between biotin and BSA. This allows for better accessibility of the attached biotin to avidin. However, even in bcap-BSA there are still on average 1.3 biotins per BSA molecule that are sterically hindered from binding to avidin. Complete acid hydrolysis of bcapBSA, on the other hand, provides the total amount of biotin conjugated to the protein. In conclusion, the method described in this paper is able to give an accurate estimation of the degree of biotinylation. The method has an excellent detection limit of 2 × 10-8 M for biotin. The acid hydrolysis to release biotin is superior to enzymatic digestion as it leads to complete hydrolysis of the protein. The enzymatic digestion is dependent on the amount of proteinase K used and the time of digestion. The currently described method should pose no problems when the degree of biotinylation of proteins prepared by using BNHS or longer chain esters of NHS is determined. ACKNOWLEDGMENT

This work was supported by grants from the National Science Foundation and the Department of Energy. We thank Michael Russ, from the Macromolecular Structure Analysis Facility, for acid hydrolysis of proteins.

98 Bioconjugate Chem., Vol. 8, No. 1, 1997 LITERATURE CITED Barbarakis, M. S., Daunert, S., and Bachas, L. G. (1992) Effect of Different Binding Proteins on the Detection Limits and Sensitivity of Assays Based on Biotinylated Adenosine Deaminase. Bioconjugate Chem. 3, 225-229. Barbarakis, M. S., Qaisi, W. G., Daunert, S., and Bachas, L. G. (1993a) Observation of “Hook Effects” in the Inhibition and Dose-Response Curves of Biotin Assays Based on the Interaction of Biotinylated Glucose Oxidase with Strept(avidin). Anal. Chem. 65, 457-460. Barbarakis, M. S., Smith-Palmer, T., Bachas, L. G., Chen, S., and Van Der Meer, W. (1993b) Enhancement of the Emission Intensity of Fluorescence-Labeled Avidin by Biotin and Biotin Derivatives. Evaluation of Different Fluorophores for Improved Sensitivity. Talanta 40, 1139-1145. Bayer, E. A., and Wilchek, M. (1990) Protein Biotinylation. Methods Enzymol. 184, 138-160. Der-Bailian, G. P., Gomez, B., Masino, R. S., and Parce, J. W. (1990) A Fluorometric Method for Determining the Degree of Biotinylation of Proteins. J. Immunol. Methods 126, 281285. Diamandis, E. P., and Christopoulos, T. K. (1991) The Biotin(Strept)avidin System, Principles and Applications in Biotechnology. Clin. Chem. 37, 625-636. Ebeling, W., Hennrich, N., Klockow, M., Metz, H., Orth, H. D., and Lang, H. (1974) Proteinase K from Tritirachium album Limber. Eur. J. Biochem. 47, 91-97. Green, N. M. (1990) Avidin and Streptavidin. Methods Enzymol. 184, 51-67. Hentz, N. G., and Bachas, L. G. (1995) Class-Selective Detection System for Liquid Chromatography Based on the Streptavidin-Biotin Interaction. Anal. Chem. 67, 1014-1018. Kurosky, A., Miller, B. T., and Knock, S. L. (1993) Kinetic Analysis of Biotinylation of Specific Residues of Peptides by High-Performance Chromatography. J. Chromatogr. 631, 281-287. Mei, G., Pugliese, L., Rostao, N., Toma, L., Bolognesi, M., and Finazzi-Agro´, A. (1994) Biotin and Biotin Analogues Specif-

Rao et al. ically Modify the Fluorescence Decay of Avidin. J. Mol. Biol. 242, 539-565. Miles, D., and Garcia, A. (1995) Separation of Biotin Labeled Proteins from Their Unlabeled Counterparts Using Immobilized Platinum Affinity Chromatography. J. Chromatogr. 702, 173-189. Miller, B. T., Rogers, E. M., Smith, J. S., and Kurosky, A. (1994) Identification and Characterization of O-Biotinylated Hydroxy Amino Acid Residues in Peptides. Anal. Biochem. 219, 240-248. Paek, S., Bachas, L. G., and Schramm, W. (1993) Defined Analyte-Enzyme Conjugates as Signal Generators in Immunoassays. Anal. Biochem. 210, 145-154. Przyjazny, A., Hentz, N. G., and Bachas, L. G. (1993) Sensitive and Selective Liquid Chromatographic Postcolumn Reaction Detection System for Biotin and Biocytin Using a Homogeneous Fluorescence-Linked Assay. J. Chromatogr. 654, 7986. Shah, D., Salbilla, V., Richerson, R., and Brown III, W. (1994) Determination of Biotin in Biotin-Conjugated Protein by Measuring Fluorescence Polarization. Clin. Chem. 40, 21122113. Smith, J. S., Miller, B. T., Knock, S. L., and Kurosky, A. (1991) Biotinylated Peptides/Proteins. Anal. Biochem. 197, 247-253. Wilchek, M., and Bayer, E. A. (1988) The Avidin-Biotin Complex in Bioanalytical Applications. Anal. Biochem. 171, 1-32. Wilchek, M., and Bayer, E. A. (1990) Biotin-Containing Reagents. Methods Enzymol. 184, 123-138. Yem, A. W., Zurcher-Neely, H. A., Richard, K. A., Staite, N. D., Heinrikson, R. L., and Deibel, M. R. (1989) Biotinylation of Reactive Amino Groups in Native Recombinant Human Interleukin-1β. J. Biol. Chem. 264, 17691-17697. Zaitsu K, Ohnishi, M., Nanami M., and Ohkura, Y. (1992) Evaluation of the Numbers of Functional Groups Introduced into Horseradish Peroxidase in Reactions with Four Heterobifunctional Reagents. Chem. Pharm. Bull. 40, 2205-2207.

BC960080P