Further Studies on the Site-Specific Protein Modification by Microbial

Ajinomoto Company Inc., Pharmaceutical Research Laboratories, 1−1 ... Publication Date (Web): August 23, 2001 ... Chemical Reviews 0 (proofing), ...
0 downloads 0 Views 149KB Size
Bioconjugate Chem. 2001, 12, 701−710

701

Further Studies on the Site-Specific Protein Modification by Microbial Transglutaminase Haruya Sato,* Eiko Hayashi, Naoyuki Yamada, Masanobu Yatagai, and Yoshiyuki Takahara Ajinomoto Company Inc., Pharmaceutical Research Laboratories, 1-1 Suzuki-cho, Kawasaki-ku, Kawasaki-shi, 210-8681, Japan. Received October 28, 2000; Revised Manuscript Received May 20, 2001

A guinea pig liver transglutaminase (G-TGase)-mediated procedure for the site-specific modification of chimeric proteins was recently reported. Here, an alternative method with advantages over the recent approach is described. This protocol utilizes a microbial transglutaminase (M-TGase) instead of the G-TGase as the catalyst. M-TGase, which has rather broad structural requirements as compared to the G-TGase, tends to catalyze an acyl transfer reaction between the γ-carboxamide group of a intact protein-bound glutamine residue and various primary amines. To demonstrate the applicability of the M-TGase-catalyzed protein modification in a drug delivery system, we have utilized recombinant human interleukin 2 (rhIL-2) as the target protein and two synthetic alkylamine derivatives of poly(ethyleneglycol) (PEG12; MW 12 kDa) and galactose-terminated triantennary glycosides ((Gal)3)) as the modifiers. For the M-TGase-catalyzed reaction with PEG12 and (Gal)3, 1 mol of alkylamine was incorporated per mole of rhIL-2, respectively. Peptide mapping of (Gal)3-modified rhIL-2 ((Gal)3rhIL-2) by liquid chromatography-electrospray ionization mass spectrometry (LC-ESI/MS) suggested that the Gln74 residue in rhIL-2 was site specifically modified with (Gal)3. The PEG12-rhIL-2 and (Gal)3-rhIL-2 conjugates retained full bioactivity relative to the unmodified rhIL-2. In pharmacokinetic studies, PEG12-rhIL-2 was eliminated more slowly from the circulation than rhIL-2, whereas (Gal)3rhIL-2 accumulated in the liver via hepatic asialoglycoprotein receptor binding. The results of this study expand the applicability of the TGase-catalyzed methodology for the preparation of protein conjugates for clinical use.

INTRODUCTION

Our goal is to develop methods for site-specific protein modification in order to generate protein conjugates for clinical use. The advent of recombinant DNA technology has brought with it the rapid development of protein therapeutics. Cytokines and other biological response modifiers, thrombolytics, adhesion molecules, agonist and antagonist peptide fragments of growth factors, and receptors all have widespread applications. The full exploitation of this new and powerful therapeutic armory requires solutions to several problems. First, most parenterally administered proteins are rapidly cleared from the circulation by the reticuloendothelial system, kidney, spleen, or liver. Second, the targeting efficiency of the administered proteins to the site of action is low. Attaching poly(ethyleneglycol) (PEG, also called polyoxyethylene (POE))1 or an organ specific targeting device specifically to the proteins can improve the therapeutic utility by controlling the pharmacokinetic profiles in the body (1, 2, 3, 4). In this aspect, the product homogeneity can play an essential role in the licensing aspects of the protein conjugates as therapeutic agents. The most popular method is the thiol-selective modification of an introduced cysteine residue by using sitedirected mutagenesis and either maleimide-based or haloacetic-based reagents (5, 6, 7). A less popular method is the modification of vicinal OH and NH2 groups on N-terminally introduced serine or threonine residues by using site-directed mutagenesis and aminooxy-functionalized reagents (8, 9). In addition to the above methods, * To whom correspondence should be addressed. e-mail; [email protected].

we have recently devised a novel method for the sitespecific incorporation of alkylamine derivatives into specific glutamine residues by using genetic engineering techniques and guinea pig liver transglutaminase (GTGase) (10). This method is superior to the aforementioned procedures, because of its special limitation to glutamine residues for the incorporation and its mild reaction conditions. In fact, thiol-reactive reagents can modify the side chains of histidine and lysine, and the R-amino groups of the peptide to a lesser extent, whereas Nterminal oxidation with periodate can potentially damage the protein. Furthermore, our G-TGase-catalyzed method is applicable to the dual and site-specific incorporation by using a special substrate sequence (TG2; AQQIVM) for G-TGase as the appended tag (11, 12). In fact, a chimeric protein of human interleukin 2 (rTG2-IL-2), in which TG2 was fused to the N-terminus of IL-2, was 1 Abbreviations: TGase, transglutaminase; M-TGase, microbial TGase; G-TGase, guinea pig liver TGase; IL-2, interleukin 2; hIL-2, human IL-2; rhIL-2, recombinant hIL-2; PEG, poly(ethyleneglycol); PEG12, poly(ethyleneglycol) donor substrate (average MW, 12 kDa); (Gal)3, galactose-terminated triantennary glycosides; (Gal)3-rhIL-2, conjugate between (Gal)3 and rhIL-2; PEG12-rhIL-2, conjugates between PEG12 and rhIL2; ESI/MS, electrospray ionization mass spectrometry; LC-ESI/ MS, liquid chromatography-ESI/MS; ASGP-R, asialoglycoprotein receptor; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; Gln, glutamine: CTLL, cytotoxic T lymphocyte line; TFA, trifluoroacetic acid; Tris-HCl, tris(hydroxymethyl)aminomethane hydrochloride; RP-HPLC, reversedphase high-performance liquid chromatography; MW, molecular weight.

10.1021/bc000132h CCC: $20.00 © 2001 American Chemical Society Published on Web 08/23/2001

702 Bioconjugate Chem., Vol. 12, No. 5, 2001

Sato et al.

Figure 1. Synthesis of (Gal)3.

dually and site-specifically modified with alkylamine derivatives of PEG (MW 10 kDa). However, considering the therapeutic applications of our G-TGase-catalyzed method, one might fear some immunogenicity or antigenicity of the constructed chimeric proteins. To try to overcome the present limitations of our methodology, we have developed a more applicable method for the site-specific incorporation of alkylamines into intact proteins, by using a microbial transglutaminase (M-TGase) in place of G-TGase as the catalyst. M-TGase, which is produced by the microorganism Streptoverticillium sp. strain s-8112, catalyzes the acyl transfer reaction between the γ-carboxyamide groups of glutamine residues in proteins and various primary amines in the same way as mammalian TGases (13, 14) The amino acid sequence of the M-TGase is very different from those of mammalian TGases. Furthermore, MTGase has a rather broader substrate specificity for amine acceptor glutamine substrates in proteins than those of the mammalian TGases in our preliminary studies. To demonstrate the usefulness of the M-TGase for constructing protein conjugates, we have selected recombinant human interleukin 2 (rhIL-2) as a target protein. Although none of the six glutamine residues in rhIL-2 were available as substrates for G-TGase, as described before (10) one reactive Gln residue, which could serve as a substrate for M-TGase, was identified in this paper. For conjugation studies, we have used two model substrates as the circulating carrier and the targeting device. One is a synthetic poly(ethyleneglycol) derivative (PEG12, MW: 12 kDa) with a straight chain alkylamine at one end, which is a useful polymer for prolonging the circulating lifetimes of proteins in vivo, as described before (10, 12). The other is a galactose-terminated triantennary glycoside with a straight chain alkylamine

at one end ((Gal)3) (15). (Gal)3 is an artificial ligand for the hepatic asialoglycoprotein receptor (ASGP-R), which is uniquely localized on hepatocytes and recognizes exposed and branched galactose residues on serum glycoproteins (16, 17). Since the uptake of glycoproteins by this receptor is both a high-affinity and high-capacity process, (Gal)3 would be utilized as a homing device to target proteins specifically around hepatocytes. We report here the stoichiometrically precise, sitespecific incorporation of alkylamine derivatives (PEG12 and (Gal)3) into intact rhIL-2 by using M-TGase. To demonstrate the modification site in rhIL-2, peptide mapping of (Gal)3-modified rhIL-2 ((Gal)3-rhIL-2) by liquid chromatography-electrospray ionization mass spectrometry (LC-ESI/MS) was carried out. We further compared the biological activities and the pharmacokinetic profiles of PEG12-modified rhIL-2 (PEG12-rhIL2) and (Gal)3-rhIL-2 to those of unmodified rhIL-2. MATERIALS AND METHODS

Materials. M-TGase was purified from the culture supernatant of Streptoverticillium sp. s-8112 according to the modified method of Ando et al. (14). R-Carboxymethyl-ω-methoxypolyoxyethylene, with an average molecular weight of 12 kDa, was purchased from Nippon Oils & Fats Co., Ltd., Japan. rhIL-2 was purified from inclusion bodies, as described by Tsuji et al. (18). Other chemicals were of reagent grade. Synthesis of PEG12. The artificial substrate PEG12 was synthesized from R-carboxymethyl-ω-methoxypolyoxyethylene and N-Boc-1,5-diaminopentane (Fluka, Buchs, Switzerland) as described (10). Synthesis of (Gal)3. The synthesis of (Gal)3 was carried out according to the modified method of Murahashi et al. (15) as follows (Figure 1).

Site-Specific Protein Modification by Microbial Transglutaminase

(a) Synthesis of 2-[2-(2-Aminoethoxy)ethoxy]ethanol. 2-[2-(2-Chloroethoxy)ethoxy]ethanol (36.5 mL, 251 mmol) was added to a solution of potassium phthalimide (50.6 g, 273 mmol) in dimethylformamide (300 mL), and the solution was stirred and heated with an oil bath (100 °C) for 17 h. The suspension was cooled to room temperature and was filtered to remove the precipitate. The filtrate was concentrated under reduced pressure and was extracted with dichloromethane (300 mL). The extract was concentrated under reduced pressure to afford N-[2-(2-hydroxyethoxy)ethoxy]ethylphthalimide as a yellow oil: 1H NMR (300 MHz, CDCl3) δ 2.58 (br, 1H), 3.51-3.57 (m, 2H), 3.57-3.71 (m, 2H), 3.72-3.80 (m, 2H), 3.87-3.96 (m, 2H), 7.68-7.80 (m, 2H), 7.82-7.90 (m, 2H); 13C NMR (75 MHz, CDCl ) δ 37.19, 61.72, 67.91, 69.98, 3 70.32, 72.43, 123.25, 132.06, 133.95, 168.34. The above product was dissolved in ethanol (1.2 L). To this solution was added 80% hydrazine hydrate (16.8 mL, 277 mmol), and the mixture was mechanically stirred and heated under reflux for 2 h. The reaction mixture was cooled to room temperature and was filtered to remove the precipitate. The filtrate was concentrated under reduced pressure and was extracted with dichloromethane (500 mL). The extract was concentrated under reduced pressure to afford 2-[2-(2-aminoethoxy)ethoxy]ethanol as a yellow oil: yield 34.2 g; 1H NMR (300 MHz, CDCl3) δ 2.90 (t, J ) 5.2 Hz, 2H), 3.18 (br, 3H), 3.53-3.70 (m, 8H), 3.70-3.77 (m, 2H); 13C NMR (75 MHz, CDCl3) δ 41.33, 61.46, 70.11, 70.35, 72.63, 72.66. (b) Synthesis of 2-{2-[2-(Benzyloxycarbonylamino)ethoxy]ethoxy}ethanol. 2-[2-(2-Aminoethoxy)ethoxy]ethanol (34.2 g) was dissolved in distilled water (500 mL) and was cooled with an ice-water bath. To this solution were added alternately, in about five equal portions, sodium bicarbonate (19.5 g) and benzyloxycarbonyl chloride (125 mL, 33% solution in toluene), and the resulting mixture was stirred at room-temperature overnight. The reaction mixture was separated from the aqueous layer, which was extracted with ethyl acetate (100 mL) twice. The combined organic phases were washed with saturated aqueous sodium chloride, dried over anhydrous magnesium sulfate, and concentrated under reduced pressure. The residue was purified by column chromatography over silica gel, which was eluted with ethyl acetate to give 2-{2-[2-(benzyloxycarbonylamino)ethoxy]ethoxy}ethanol as a colorless oil; yield 32.1 g; 1H NMR (300 MHz, CDCl3) δ 2.58 (br, 1H), 3.35-3.44 (m, 2H), 3.54-3.67 (m, 8H), 3.71 (br, 2H); 5.10 (s, 2H), 5.47 (br, 1H), 7.27-7.42 (m, 5H); 13C NMR (75 MHz, CDCl3) δ 40.82, 61.65, 66.65, 70.08, 70.28, 70.34, 72.51, 128.05, 128.09, 128.46, 136.52, 156.43. (c) Synthesis of 1-O-{2-[2-(Benzyloxycarbonylamino)ethoxy]ethoxy}ethyl-β-D-galactopylanose Tetraacetate. Thirty-two grams of 2-{2-[2-(benzyloxycarbonylamino)ethoxy]ethoxy}ethanol and 22 g of β-Dgalactopylanose pentaacetate were dissolved in dichloromethane (400 mL), which was cooled in an ice-water bath. To this mixture was added a boron trifluoride diethyl ether complex (13.9 mL) dropwise, and the resulting mixture was stirred at room-temperature overnight. The reaction mixture was washed with saturated aqueous sodium bicarbonate, dried over anhydrous magnesium sulfate, and concentrated under reduced pressure. The crude product was purified by column chromatography over silica gel, which was eluted with ethyl acetate/n-hexane (1:1 f 2:1, v/v) to give 1-O-{2-[2(benzyloxycarbonylamino)ethoxy]ethoxy}ethyl-β-D-galactopylanose tetraacetate as a colorless oil; yield 13.7 g; 1H NMR (300 MHz, CDCl ) δ 1.98 (s, 3H), 2.04 (s, 6H), 3

Bioconjugate Chem., Vol. 12, No. 5, 2001 703

2.14 (s, 3H), 3.40 (dd, J ) 5.2, 10.4 Hz, 2H); 3.52-3.78 (m, 9H), 3.85 (br t, Japp ) 7.1 Hz, 1H), 3.95 (m, 1H), 4.064.24 (m, 2H), 4.52 (d, J ) 7.9 Hz, 1H), 5.00 (dd, J ) 3.5, 10.3 Hz, 1H), 5.11 (s, 2H), 5.20 (dd, J ) 7.9, 10.5 Hz, 1H), 5.37 (br d, Japp ) 3.5 Hz, 1H), 7.3-7.4 (m, 5H). (d) 1-O-[2-(2-(Aminoethoxy)ethoxy)ethoyl]-β-D-galactopylanose Tetraacetate. To a solution of 1-O-{2[2-(benzyloxycarbonylamino)ethoxy]ethoxy}ethyl-β-D-galactopylanose tetraacetate (13.7 g) in ethanol (300 mL) was added p-toluenesulfonic acid monohydrate (4.27 g). The mixture was added to 10% palladium on carbon (1.0 g) under argon and was stirred under atmospheric pressure of hydrogen at room-temperature overnight. After removing the hydrogen under reduced pressure and changing the atmosphere to argon, the reaction mixture was filtered to remove the palladium on carbon and was evaporated. The residue was dissolved in acetonitrile (300 mL), and the solution was evaporated under reduced pressure. This procedure was repeated twice to remove the residual ethanol. The residue was dissolved in acetonitrile (40 mL) again, and 4-methylmorpholine (2.5 mL) was added to this solution to afford a solution of 1-O[2-(2-(aminoethoxy)ethoxy)ethyl-β-D-galactopylanose tetraacetate (solution A). (e) Synthesis of tert-Butyloxycarbonyl-L-glutamylL-glutamic Acid r,γ-Benzyl Ester. Boc-L-Glu-OBzl (4.73 g, Peptide Institute, Inc., Osaka, Japan) and Nmethylmorpholine (1.54 mL) were dissolved in tetrahydrofuran (100 mL). The solution was cooled with a dry ice-ethanol bath. Ethyl chloroformate (1.35 mL) was added, and the mixture was stirred for 5 min. To the reaction mixure was slowly added a mixture of L-Glu(OBzl)-OBzl p-toulenesulfonate (7.0 g, Peptide Institute, Inc., Osaka, Japan) and N-methylmorpholine (1.54 mL) in dimethylformamide (50 mL). The dry ice-ethanol bath was removed and replaced with a sodium chloride-ice bath, and then the reaction mixture was stirred overnight and allowed to warm gradually to room temperature. The precipitate was removed by filtration, and the filtrate was concentrated under reduced pressure. The residue was dissolved in ethyl acetate and was washed successively with aqueous 10% citric acid, saturated aqueous sodium chloride, saturated aqueous sodium bicarbonate, and saturated aqueous sodium chloride. The organic phase was dried over anhydrous magnesium sulfate and was concentrated under reduced pressure to give 8.85 g of the tert-butyloxycarbonyl-L-glutamyl-L-glutamic acid R,γ-benzyl ester as pale yellow crystals; 1H NMR (300 MHz, CDCl3) δ 1.40 (s, 9H), 1.86-1.94 (m, 1H), 1.98-2.06 (m, 1H), 2.1-2.3 (m, 4H), 2.3-2.5 (m, 2H), 4.30-4.35 (m, 1H), 4.62-4.64 (m, 1H), 5.00-5.20 (m, 6H), 5.25 (d, J ) 7.0 Hz, 1H), 6.41 (d, J ) 7.0 Hz, 1H), 7.2-7.3 (m, 15H). (f) Synthesis of Compound 1. tert-ButyloxycarbonylL-glutamyl-L-glutamic acid R,γ-benzyl ester (8.85 g) was dissolved in trifluoroacetic acid (25 mL), stirred for 3 h under room temperature, and then concentrated under reduced pressure. The residue was dissolved in ethanol (200 mL), and the solvent was evaporated under reduced pressure to remove the residual trifluoroacetic acid. After repetition of this procedure, the obtained residue was dissolved in methanol (25 mL) and was neutralized with N-methylmorpholine. The solvent was evaporated, and the residue was dissolved in dichloromethane (220 mL) (solution B). N-Boc-6-aminohexanoic acid (3.16 g) was dissolved in dichloromethane (220 mL) and was cooled with an icewater bath. To this solution were added N-hydroxysuccinimide (1.73 g), dicyclohexylcarbodiimide (3.1 g), and solution B, and then the resulting mixture was stirred

704 Bioconjugate Chem., Vol. 12, No. 5, 2001

at room temperature overnight. The precipitate was removed from the reaction mixture by filtration, and the filtrate was washed successively with aqueous 10 % citric acid, aqueous saturated sodium chloride, and aqueous 10 % sodium bicarbonate, dried with anhydrous magnesium sulfate, and was finally concentrated under reduced pressure. The crude product was purified by column chromatography over silica gel, which was eluted with n-hexane/ethyl acetate (1:1 f 1:2, v/v) to give compound 1 as a colorless solid; yield 4.24 g; 1H NMR (300 MHz, CDCl3) δ 1.30-1.34 (m, 2H), 1.39-1.48 (m, 11H), 1.601.64 (m, 2H), 1.93-2.06 (m, 2H), 2.13-2.26 (m, 6H), 2.34-2.48 (m, 2H), 3.06-3.10 (m, 2H), 4.55-4.65 (m, 3H), 5.06-5.20 (m, 6H), 6.50-6.60 (br, 2H), 7.3-7.4 (m, 15H). (g) Synthesis of Compound 2. Ten percent of palladium on active carbon (900 mg) was added to the solution of compound 1 (4.24 g) in tetrahydrofuran (150 mL) and ethyl acetate (150 mL), and then the mixture was stirred under an atmospheric pressure of hydrogen at room-temperature overnight. The reaction mixture was filtered to remove the catalyst and was concentrated under reduced pressure to give compound 2 as a colorless solid; yield 3.39 g; 1H NMR (300 MHz, CDCl3) δ 1.341.52 (m, 13H), 1.61-1.67 (m, 2H), 1.91-2.00 (m, 2H), 2.15-2.24 (m, 2H), 2.26 (t, J ) 7.5 Hz, 2H), 2.35-2.43 (m, 4H), 3.03 (t, J ) 7.0 Hz, 2H), 4.39 (dd, J ) 5.0, 9.0 Hz, 1H), 4.44 (dd, J ) 5.0, 9.0 Hz, 1H). (h) Synthesis of Compound 3. Compound 2 (3.39 g) was dissolved in dimethylformamide (20 mL) and was cooled with an ice-water bath. To this solution were added N-hydroxysuccinimide (2.58 g) and dicyclohexylcarbodiimide (4.63 g), and the mixture was stirred for several minutes with cooling, after which solution A (1-O-{2-[2-(aminoethoxy)ethoxy]ethyl}-β-D-galactopylanose tetraacetate in acetonitrile) was added. The resulting mixture was stirred and cooled in a water bath overnight. The precipitate was removed by filtration, and the supernatant was then concentrated under reduced pressure. The residue was dissolved in ethyl acetate (300 mL) and was washed successively with an aqueous 10% citric acid solution, aqueous saturated sodium chloride, and aqueous saturated sodium bicarbonate. The organic phase was dried over anhydrous magnesium sulfate and was concentrated under reduced pressure. The crude product was purified by column chromatography over silica gel, which was eluted with dichloromethane/ methanol (20:1 f 15:1, v/v) to give compound 3 as a colorless solid; yield 5.87 g; 1H NMR (300 MHz, CDCl3) δ 1.30-1.36 (m, 2H), 1.42 (s, 9H), 1.46-1.52 (m, 2H), 1.60-1.66 (m, 2H), 2.01 (s, 9H), 2.02 (s, 9H), 2.04 (s, 9H), 2.09 (s, 9H), 2.2-2.4 (m, 2H), 3.08-3.12 (m, 2H), 3.53.6 (m, 34H), 3.70-3.75 (m, 6H), 3.93-3.98 (m, 6H), 4.10-4.20 (m, 6H), 4.38-4.43 (m, 2H), 4.55-4.57 (m, 3H), 4.72-4.76 (m, 1H), 5.02-5.06 (m, 3H), 5.17-5.22 (m, 3H), 5.39 (d, J ) 3.5 Hz, 3H), 6.58-6.62 (m, 1H), 7.04-7.08 (m, 1H), 7.12-7.18 (m, 1H), 7.24-7.26 (m, 1H), 7.807.85 (m, 1H). (i) Synthesis of (Gal)3. Compound 3 (5.87 g) was dissolved in 50% trifluoroacetic acid in dichloromethane (50 mL), stirred for 1 h under room temperature, and then concentrated under reduced pressure. The residue was dissolved in ethanol (200 mL) and was concentrated under reduced pressure to remove the residual trifluoroacetic acid. After repetition of this procedure, the resulting residue was dissolved in dichloromethane (30 mL), cooled with an ice-water bath, and then neutralized with 28% sodium methoxide in methanol. The solvent was evaporated under reduced pressure, and the remaining solution was subjected to column chromatography

Sato et al.

over silica gel (dichloromethane/methanol ) 30/1 f 5:1, v/v). The obtained product was dissolved in methanol (85 mL) and was cooled with an ice-water bath. To this solution was added 28% sodium methoxide in methanol (320 µL). This mixture was removed from the cooling bath and was stirred at room temperature for 5 h. The resulting mixture was cooled with an ice-water bath and was neutralized with Dowex 50W cation-exchange resin that had been washed with methanol. The resin was then removed by filtration. The filtrate was concentrated under reduced pressure to give 1.95 g of (Gal)3 as an amorphous solid. The purity of the (Gal)3 was more than 95% as analyzed by RP-HPLC with an Inertsil ODS-2 column (4.6 × 150 mm; GL Sciences Inc.), which was developed by a linear gradient of 4-20% acetonitorile containing 0.1% TFA for 30 min at a flow rate of 1 mL/ min; ESI-MS; (M + H)+ 1270.0 (MW 1269.4 Da). M-TGase-Catalyzed Modification of rhIL-2. The M-TGase-catalyzed modification of rhIL-2 was carried out with the following concentrations of reagents in 0.2 M Tris/HCl (pH 7.5) at 25 °C for 12 h: PEG12 or (Gal)3, 1.25 mM; protein, 5 µM; and M-TGase; 0.18 units/ml. At the end of the incubation, the reaction mixtures were analyzed by SDS-PAGE under the same conditions used for rTG1-IL-2 (10). To remove the residual free PEG12 and (Gal)3, the reaction mixtures were absorbed onto a Sep-Pak C8 cartridge (Millipore Co., Milford, MA) according to the same method used for rTG1-IL-2 (10). The eluted protein fractions were directly applied to a YMCC8-AP column (4.6 × 150 mm; YMC Co., Ltd.), which was developed by a linear gradient of 54%-60% acetonitrile containing 0.1% TFA for 30 min at a flow rate of 1 mL/ min. Each fraction was collected, and the different protein peaks were separately pooled and analyzed by reversedphase HPLC (RP-HPLC) and SDS-PAGE. After the analysis, most of the main peak fractions were directly applied to a Molcut L LGC column (Millipore Co., Milford, MA), and the buffer was exchanged with Dulbecco’s phosphate-buffered saline (D-PBS). The samples were used for bioactivity and pharmacokinetic studies. The remainder of the main fraction for (Gal)3-rhIL-2 in the RP-HPLC was used for peptide mapping. Protease Digestion of Modified and Unmodified rhIL-2. Protease digestions of (Gal)3-rhIL-2 and rhIL-2 were done according to the modified method for rhIL-2 (18). One hundred micrograms of reduced and S-carboxymethylated (Gal)3-rhIL-2 and rhIL-2 were each suspended in 180 µL of 0.05 M ammonium hydrogen carbonate buffer (pH 7.9) and were treated with 20 µL of V8 protease solution (Wako Pure Chemical Industries, Ltd., 100 µg/mL) for 22 h at 37 °C. After lyophilization of the reaction mixtures, the dry samples were respectively suspended in 45 µL of 0.05 M ammonium hydrogen carbonate buffer (pH 7.9) and were treated with 5 µL of L-1-(tosylamido)-2-phenylethyl chloromethyl ketone-treated trypsin (Sigma, 400 µg/mL) for 4 h at 37 °C. ESI-MS Analysis. An LCQ iontrap mass spectrometer (Thermoquest Co., San Jose, CA) was used for all experiments. The MS detector settings are as follows: temperature of the heated capillary, 200 °C; spray voltage, 3 kV; scan range, 300-1850 m/z. The proteasedigested peptides of modified and unmodified IL-2 were lyophilized to remove the buffer and were dissolved in 0.1% TFA. The lyophilized peptides were loaded onto a capillary chromatography column (Vydac C18, 0.3 mm i.d. × 300 mm), and the separated peptides were subjected to an electrospray ion source in the LCQ spectrometer. To achieve a stable electrospray over the course of the gradient elution, 50% MeOH containing 1% acetic acid

Site-Specific Protein Modification by Microbial Transglutaminase

was used as a sheath liquid at a flow rate of 3 µL/min. The elution conditions consisted of four successive linear increases of acetonitrile in 0.1% TFA, i.e., 4.5 to 4.5% in 10 min, 4.5 to 63% in 70 min, 63 to 81% in 5 min, and 81 to 81% in 5 min at a flow rate of 3 µL/min. Mass spectra were obtained by scanning the range of m/z 300-1850. The most abundant peptides were further analyzed by collisonal-induced dissociation (CID) mass spectrometry. In Vitro Biological Assays. The bioactivities of PEG12-rhIL-2, (Gal)3-rhIL-2, and unmodified rhIL-2 were determined by using the IL-2 dependent murine cell line CTLL-2 (19). The amount of IL-2 activity was determined in units/mg by using the rhIL-2 standard (Collaborative Biomedical Products Co., Inc.) and was then expressed as the percent residual bioactivity as compared to the purified rhIL-2. All samples were tested in quadruplicate. Pharmacokinetics of Modified rhIL-2. The modified and unmodified rhIL-2 were administered intravenously into male mice (C57BL/6, 6W, /, Charles River Breeding Labs, Japan). At specific time intervals, described in the Results, the mice were bled and sacrificed, and the liver and kidney were excised. Blood was collected in heparin-coated 1.5 mL Eppendorf tubes. Plasma was immediately prepared by centrifugation at 4 °C. Each organ was rinsed with ice-cold saline, weighed, and homogenized by adding enough isotonic phosphate buffer containing 0.2% BSA (pH 7.4) to yield a 10% homogenate (w/v). The IL-2 concentrations in plasma, liver, and kidney were determined by an ELISA procedure using the IL-2 Enzyme Immunoassay kit (Cayman Chemical Co., Michigan). Analytical Methods. The M-TGase enzyme activity was measured by a colorimetric hydroxamate procedure with N-carbobenzoxy-L-glutaminyl-glycine (20). SDSPAGE was carried out with the PhastSystem (Amersham Pharmacia Biotech AB, Uppsala, Sweden) using a 20% (w/v) acrylamide gel under reducing conditions, according to the same procedure used for rTG1-IL-2 (10). The concentrations of the unmodified and modified forms of rhIL-2 were quantified by using an extinction coefficient (E280) of 1.2 × 104 M-1. NMR spectra were determined at 300 MHz (1H) with a spectrometer operating in the Fourier Transform (FT) mode. Chemical shifts are given in ppm (δ) relative to tetramethylsilane as the internal standard. RESULTS

Synthesis of (Gal)3. The synthesis of (Gal)3 includes three stages (Figure 1). The first step involves the preparation of the building block of glutamic acids. The second stage involves the incorporation of the tri(ethylene glycol) spacer into β-D-galactopyranose tetraacetate. In the final stage, the galactopyranosyl-derivatized compounds are introduced into the building block of glutamic acids. The results of the ESI-MS and RP-HPLC analyses of the reaction product indicate that (Gal)3 is consistent with the assigned structure. M-TGase-Catalyzed PEG12-Modification of rhIL2. Figure 2A shows the SDS-PAGE analysis of the PEG12-incorporated product for rhIL-2. In the presence of PEG12 and M-TGase, the rhIL-2 band migrating at about 15 kDa is decreased, and one broad band, migrating at about 35 kDa, is visible (lane 3). The higher molecular weight band that is slightly visible at about 38 kDa could be due to M-TGase (MW 37.8 kDa). Since the time course analysis of the reaction products by SDS-PAGE revealed only one new band (data not

Bioconjugate Chem., Vol. 12, No. 5, 2001 705

Figure 2. SDS-PAGE analysis of the M-TGase-catalyzed PEG12- (A) and (Gal)3- (B) modified rhIL-2. (A) Lane 1; MW Marker Lane 2; unmodified rhIL-2 Lane 3; rhIL-2 + PEG12 + M-TGase Lane 4; purified PEG12-rhIL-2 (B) Lane 1; MW Marker Lane 2; unmodified rhIL-2 Lane 3; rhIL-2 + (Gal)3 + M-TGase Lane 4; purified (Gal)3-rhIL-2

shown), the band migrating at 35 kDa presumably represents the PEG12-incorporated rhIL-2, with a molar ratio of 1:1. The attachment of one PEG12 molecule to rhIL-2 causes a greater increase in size than that predicted from its molecular mass, presumably due to the bulkiness of the PEG12-rhIL-2 conjugate (PEG12-rhIL2). After the separation of the modified and unmodified rhIL-2 by RP-HPLC (data not shown), the isolated PEG12-rhIL-2 migrated as a single band at about 35 kDa in lane 4 (yield: 29%). M-TGase-Catalyzed (Gal)3-Modification of rhIL2. Figure 2B shows the SDS-PAGE analysis of the (Gal)3-incorporated product for rhIL-2. The electrophoretic pattern of the reaction mixture demonstrates that the rhIL-2 band migrating at 15 kDa is almost invisible, and one new band, migrating at 17 kDa, is present (lane 3), indicating that one mole of (Gal)3 per rhIL-2 is incorporated by M-TGase, similar to the PEG12-modification. The attachment of one (Gal)3 (MW; 1269 Da) caused a reasonable increase (about 1000 Da) in size, representing the single (Gal)3-incorporated product of rhIL-2 ((Gal)3rhIL-2). The M-TGase band at 38 kDa is stronger than that for the PEG12-modification in Figure 2A, due to differences in the enzyme activity/mg of the M-TGase used for the individual reactions. To characterize the (Gal)3-rhIL-2, we carried out the separation of the modified and unmodified rhIL-2 by RPHPLC (Figure 3). Two separate peaks (a and b) were observed. The large peak that eluted at about 12 min contained M-TGase and (Gal)3. Peak b corresponds to the unmodified rhIL-2 (data not shown). The protein band from peak a, which migrates at 17 kDa (lane 4 in Figure 2(B)), corresponds to the (Gal)3-rhIL-2. A fraction of peak a (yield: 20%) was collected and used for characterization. Although the incorporation efficiency of (Gal)3 was higher than that of PEG12, as determined by the SDSPAGE analysis of the reaction mixture, the yield of the (Gal)3-rhIL-2 was inversely lower than that of PEG12rhIL-2. The reason for the low recovery is assumed to be the insufficiency of the separation between (Gal)3-rhIL-2 and unmodified IL-2 in the RP-HPLC. ESI-MS Analysis of (Gal)3-rhIL-2. Figure 4 shows the LC-ESI/MS chromatography of the protease-digested peptides containing Gln residues of rhIL-2 and (Gal)3rhIL-2. The identities of the respective peaks containing Gln residues, together with the theoretical and measured molecular masses, are displayed in Table 1. A comparison of the chromatography derived from rhIL-2 and (Gal)3-rhIL-2 reveals that only VT-1 (residues 69-76, VLNLAQSK) was not observed in (Gal)3-

706 Bioconjugate Chem., Vol. 12, No. 5, 2001

Figure 3. Purification of (Gal)3-rhIL-2 on the YMC-C8AP column: a, peak containing (Gal)3-rhIL-2; b, peak containing unmodified rhIL-2.

rhIL-2, and a new peak (VT-1(+(Gal)3) was observed at 41 min. Since the ESI/MS spectra of the VT-1 (+(Gal)3) gave a strong peak at m/z 1063.1, as the double-charged protonated molecular ion ([M + 2H]2+, predicted m/z; 1063.5 Da) of the (Gal)3-incorporated VT-1 ((Gal)3 ((C52H96N6O29)); MW 1269.2 Da, VT-1; MW 871.5 Da) (Figure 5), (Gal)3 was site-specifically incorporated at Gln74 in rhIL-2. In Vitro Biological Assays. The bioactivities of PEG12-rhIL-2 and (Gal)3-rhIL-2 are shown in Table 2.

Sato et al.

The control, purified rhIL-2, has an activity of 1.43 × 107 units/mg as hIL-2. The bioactivities of PEG12-rhIL-2 and (Gal)3-rhIL-2 are 1.35 × 107 units/mg (94%) and 1.53 × 107 units/mg (107%), respectively, indicating that both modified forms of rhIL-2 retain full bioactivity relative to rhIL-2. Pharmacokinetic Studies. The blood plasma curves for PEG12-rhIL-2 and rhIL-2 given intravenously to mice are compared in Figure 6. PEG12-rhIL-2 showed a prolonged circulation half-life in comparison with unmodified rhIL-2. The R and β half-lives of rhIL-2 were 2.2 and 11 min, respectively. For PEG12-rhIL-2, these half-lives were increased to 8.8 and 75 min, respectively. The area under the plasma concentration vs time curve (AUC0-inf) of PEG12-rhIL-2 (442 ng‚h/mL) was about 10 times greater than that of unmodified rhIL-2 (46 ng‚h/ mL). Figure 7 shows the tissue distribution of (Gal)3-rhIL-2 and rhIL-2 after i.v. administration in mice. (Gal)3rhIL-2 predominantly accumulated in the liver, at about a 5-fold higher concentration than that of unmodified rhIL-2. When (Gal)3-rhIL-2 was coadministered with asialoorosomucoid, which is a well-known high-affinity ligand for the hepatic ASGP-R, the liver accumulation of the conjugate was reduced to the same level as that of rhIL-2. This result indicates that the hepatic targeting of (Gal)3-rhIL-2 is via the ASGP-R. DISCUSSION

Selective protein drug delivery and targeting seek to achieve the optimal arrival of a drug at its site of action in a manner that is appropriate for the disease and the drug, which would lead to a significant reduction in the possibility of drug side effects. Controlling the circulating lifetimes or tissue distributions of therapeutic proteins is one of the main areas for developing drug delivery

Figure 4. Comparison of LC-ESI/MS chromatography of trypsin- and V-8 protease-digested peptides containing Gln residues of rhIL-2 (a) and (Gal)3-rhIL-2 (b). A; Base peak chromatogram, B-H; [M + H]+ ion chromatograms for the fragments VT-1, VT-1(+(Gal)3), VT-2, VT-3, VT-4, VT-5, and VT-6, as described in Table 1.

Site-Specific Protein Modification by Microbial Transglutaminase

Bioconjugate Chem., Vol. 12, No. 5, 2001 707

Table 1. Sequence Information for the Observed Fragments Containing Gln Residues of RhIL-2 and (Gal)3-rhIL-2 Following LC-ESI/MS

a

fragment

residues

predicted mass (M + H)+ (m/z)

observed mass (M + H)+ (m/z)

sequenceb

VT-1 VT-1(+(Gal)3) VT-2 VT-3 VT-4 VT-5 VT-6

69-76 69-76 9-15 55-62 125-133 21-31 121-133

872.5 1063.5a 859.5 1058.4 1023.5 1292.7 1570.8

872.8 1063.1 859.7 1058.8 1023.8 1293.0 1571.1

VLNLAQSK VLNLAE(-(Gal)3)SK KTQLQLE HLQCLEEE CQSIISTLT LQMILNGINNY WITFCQSIISTLT

This value was described as (M + 2H)2+. b All cysteine (C) residues were carboxymethylated.

Figure 5. ESI/MS spectra from the LC-ESI/MS analysis shown in Figure 4a: mass spectrum of VT-1 of rhIL-2 at 42 min. b: Mass spectrum of (Gal)3-incorporated VT-1 (VT-1(+(Gal)3) of (Gal)3-rhIL-2 at 41 min. Table 2. IL-2 Bioactivity of Modified RhIL-2 bioactivitya protein

units/mg

% activityb

rhIL-2 PEG12-rhIL-2 (Gal)3-rhIL-2

1.43 × 1.35 × 107 1.53 × 107

100 94 107

107

a The amount of IL-2 activity was determined in units/mg by using the IL-2 standard on murine CTLL-2 cells. b The activity is expressed as the precent residual bioactivity as compared to the rhIL-2 bioactivity.

systems. By using M-TGase as the catalyst, we have developed a new methodology for the site-specific incorporation of alkylamine derivatives of a hepatic targeting device ((Gal)3) and PEG (PEG12) into intact proteins and controlling the pharmacokinetic profiles of the protein drugs. The progress of our application study for the TGasecatalyzed modification of therapeutic proteins depends on the adoption of M-TGase instead of G-TGase as the catalyst. Although a few attempts have been made to characterize the substrate specificity of M-TGase, using synthetic peptide derivatives, the characterizations of the substrate recognition of the enzyme have not considered the tertiary structure (14, 21). From our preliminary studies, M-TGase has lower substrate specificity for the amine acceptor site around the Gln residues in proteins than G-TGase. In the case of the G-TGase-catalyzed modification studies for rhIL-2, there was a need to

Figure 6. Plasma concentration time courses of PEG12-rhIL-2 and rhIL-2 in mice. The plasma concentrations were determined by ELISA after intravenous administration (50 µg as rhIL-2/ kg) into C57BL/6 mice (6W, /). Data are presented as means ( standard deviations.

construct a chimeric protein of hIL-2, in which the substrate peptide was fused to the N-terminus of hIL-2

708 Bioconjugate Chem., Vol. 12, No. 5, 2001

Sato et al. Table 3. Comparison of the Amino Acid Sequence Surrounding the Gln Residues in RhIL-2 Gln residues

sequencea,b

location

Gln11 Gln13 Gln22 Gln57 Gln 74 Gln126

APTSSSTKKT Q LQLEHLLLDL TSSSTKKTQL Q LEHLLLDLQM LQLEHLLLDL Q MILNGINNYK PKKATELKHL Q CLEEELKPLE KPLEEVLNLA Q SKNFHLRPRD EFLNRWITFC Q SIISTLT-COOH

in helix A in helix A in helix A in helix B near helix B in helix D

a The Gln residues in rhIL-2, as well as the charged amino acids, are printed in boldface. b The carboxy terminus of rhIL-2 is indicated by COOH

Figure 7. Tissue distributions of (Gal)3-rhIL-2 and rhIL-2 in mice. Each sample (50 µg as rhIL-2/kg) was injected intravenously into mice, without or with asialoorosomucoid (1320 µg/ kg). After 2 min of circulation, the uptake by the liver and kidney and the remaining fraction in the serum were determined (percentage of the injected dose). Data are presented as means ( standard deviations.

(10). On the other hand, we could site-specifically incorporate alkylamine derivatives of PEG or a hepatic targeting device into intact rhIL-2 by using M-TGase. In fact, we have found a lot of other substrate proteins for M-TGase, such as human interleukin 6, human interferon R, and so on (data not shown). The reason for the lower substrate specificity of M-TGase is assumed to be as follows. It is known that the catalysis of TGase takes place in an acylation-deacylation pathway; this reaction proceeds through the formation of Michaelis-type acyl-enzyme intermediates and the subsequent transfer of acylenzyme intermediates (22, 23). Specifically, there is evidence that these enzymes have extended active sites whose interactions with substrate peptides bear directly on their binding affinity for and/or their catalytic efficiency toward glutamine residues (24). These secondary interactions between the target proteins and the active site of TGase are thought to determine the reactivity for TGase. When compared with the structures of the two TGases, by Kanaji et al. (13), the secondary structure of the region around the active site Cys residue of M-TGase is similar to that of G-TGase, whereas the amino acid sequence of the M-TGase is very different from that of G-TGase. However, the molecular weight for M-TGase (37.9 kDa) is much smaller than those of the mammalian TGases, represented by G-TGase (76.6 kDa). Thus, the smaller M-TGase could easily interact with the substrate site of the target proteins, as compared to G-TGase, and the subsequent transfer of the acyl-enzyme intermediates results in the lower specificity toward glutamine residues. This characteristic of M-TGase is assumed to be a special feature among the TGases. To confirm the broad specificity of M-TGase for the Gln substrate in intact proteins, further systematic studies for its reactivity toward various proteins will be necessary. Our characterization studies of (Gal)3-rhIL-2 by LCESI/MS showed that the Gln74 residue in rhIL-2 was modified with (Gal)3 in a site-specific manner. In addition, the LC-ESI/MS analysis of PEG12-rhIL-2 also showed that only the VT-1 fragment (residues 69-76) of rhIL-2 was site-specifically modified with PEG12 (data not shown). The consistent results of the incorporation site between (Gal)3 and PEG12 indicate that only Gln74, out of the six Gln residues in rhIL-2 (Gln11, 13, 22, 57, 74, and 126), can be a substrate for M-TGase.

From the stereoview of the hIL-2 mutant (PDB Code: 3INK) represented by a Protein Data Bank program (25), all of the side-chains of the glutamine residues in hIL-2 are located in the exposed region. Since the location of reactive glutamine residues at the surface of the respective proteins is a common feature for TGase-catalyzed modifications, every glutamine residue in hIL-2 has the possibility of being an amine acceptor substrate in this category. Thus, the amino acid sequences surrounding the Gln residues in hIL-2 play a key role for the reactivity. In fact, the reactive Gln residues for TGase are located in a sequence containing a high proportion of charged and polar amino acids (26). Furthermore, in substrate proteins with known three-dimensional structures, i.e., melittin and glucagon, the reactive glutamine residues are located in an R-helix with positively charged amino acids clustered either NH2- or COOH-terminally, relative to this residue (27, 28) Table 3 shows the amino acid sequences and the secondary structures surrounding the Gln residues in hIL-2 (29, 30). Comparing the location of each Gln residue, the sequences around Gln57 and Gln74 would be most abundant in polar and (positively) charged amino acids. Thus, the site-specificity at Gln74 appears to be reasonable. Since it is impossible to derive a consensus sequence for the M-TGase-catalyzed modification at present, the reason for the reaction selectivity at Gln74 and not for Gln57 is immediately clear to us. The differences in the amino acid sequence and the peptide flexibility around Gln57 and Gln74 would dramatically affect the binding affinity for the acyl-enzyme intermediate between M-TGase and rhIL-2. Our pharmacokinetic studies showed that the sitespecific introduction of the PEG12 molecule prolonged the circulation lifetime of rhIL-2. To increase the time further, either enhancing the PEG length or adding multiple copies of PEG chains at one site would be necessary. In fact, this type of PEG-protein conjugate, in which one mole of high molecular weight (40 kDa) branched PEG is incorporated per mole of interferon R-2a, has shown sufficient prolongation in its pharmacokinetic profiles (31). On the other hand, (Gal)3-rhIL-2 could accumulate in the liver via the hepatic ASGP-R. To achieve hepatic delivery of proteins, many groups have synthesized galactosylated proteins, in which some galactose residues are covalently attached to the lysine side chains of the proteins (32, 33). However, since N-linked oligosaccharides containing multiple terminal Gal residues bind to the hepatic ASGP-R with high affinity, more Gal residues must be incorporated per target protein, which would reduce the bioactivity. Since such conjugates showed product heterogeneity and decreased bioactivity as compared to the unmodified proteins, it was difficult to develop them as therapeutic agents. Thus, the methodology for the hepatic targeting of proteins by using (Gal)3

Site-Specific Protein Modification by Microbial Transglutaminase

Figure 8. The strategy of the site-specific modification of proteins by the use of M-TGase (i) and G-TGase (ii).

would be generally applicable, as compared to the previous method, as detailed in the following points. (i) This methodology can achieve the hepatic delivery of proteins with one point of incorporation of the ligand at a substrate Gln residue. (ii) The incorporation of (Gal)3 would not affect the function of the proteins. In addition to the above points, (Gal)3-mediated targeting has another advantage over the previous hepatic delivery systems. Although there are some reports of the utilization of artificial tri- or tetraantennary ligands for ASGP-R for hepatic targeting devices, these were developed for carrying DNAs, antisense oligonucleotides, cholesterols, and liposomes, but not proteins, into hepatocytes (34, 35, 36, 37). Then, they selected the highest affinity ligands (Kd, nanomolar affinity) for the devices to favor the efficient uptake of the molecules via ASGPR-mediated endocytosis. In contrast, (Gal)3 is a lower affinity ligand for the ASGP-R (Kd, several hundreds nM to µM affinity) as compared to the previous ligands. This affinity would be sufficient for binding to the ASGP-R after administration, but insufficient for internalization into hepatocytes. In fact, we found that the (Gal)3 had a special feature that allowed it to escape from ASGP-R-mediated endocytosis after binding to the ASGP-R on hepatocytes in our preliminary studies, whereas the high-affinity ligand was rapidly internalized into the cells (data not shown). Thus, (Gal)3-protein conjugates would mainly accumulate around the extracellular region of hepatocytes, not within the cells. Since many endogenous proteins, such as cytokines, are produced and released to act locally, and since these types of molecules have multiple functions in the body, the delivery of systematically administered proteins to the site of action would be necessary to reduce the systemic toxicity. With this viewpoint, our hepatic targeting system is a potential methodology for developing therapeutic proteins toward liver diseases. To demonstrate the utility of the (Gal)3 ligand as a hepatic targeting device, further studies, such as kinetic studies of the uptake of (Gal)3-protein conjugates into hepatocytes, will be necessary. On the basis of the previous observations and the present studies of TGase-catalyzed modification, the strategy of our site-specific modification of proteins can be described as follows (Figure 8). In cases where one would like to introduce alkylamine derivatives into intact proteins, M-TGase would be suitable for their modification (1). Even if the target protein lacks reactive Gln residues, the site-specific introduction is still possible, because a short substrate sequence for G-TGase can be genetically introduced at the terminus of the target

Bioconjugate Chem., Vol. 12, No. 5, 2001 709

protein (2). This methodology can be generally applied to therapeutic proteins in cases where the incorporation of alkylamine derivatives does not affect the bioactivity of the proteins. Although alkylamine derivatives of PEG and an artificial ligand for the hepatic ASGP-R were employed as the modifier in this study, other functional alkylamines may be introduced to tailor the delivery and targeting of proteins. Furthermore, other kinds of TGases with different substrate specificities could be applicable for constructing protein conjugates. In conclusion, we have successfully developed a new methodology for the site-specific modification of intact proteins by the use of M-TGase. The high homogeneity of the constructed conjugates and the ability to control the pharmacokinetic profiles will improve the therapeutic indices of proteins. ACKNOWLEDGMENT

The authors are grateful to Dr. Yuji Iwashita for helpful discussions. We also thank Daisuke Ejima and Hiromi Sumikawa (Ajinomoto Co., Inc.) for their advice concerning purification study and structural information. LITERATURE CITED (1) Abuchowski, A., Kazo, G. M., Verhoest, C. R., Jr., Van Es, T., Kafkewitz, D., Nucci, M. L., Viau, A. T., Davis, F. F. (1984) Cancer therapy with chemically modified enzymes. I. Antitumor properties of poly(ethylene glycol)-asparaginase conjugates. Cancer Biochem. Biophys. 7, 175-86. (2) Katre, N. V., Knauf, M. J., Laird, W. J. (1987) Chemical modification of recombinant interleukin 2 by poly(ethylene glycol) increases its potency in the murine Meth A sarcoma model. Proc. Natl. Acad. Sci. U. S. A. 84, 1487-91. (3) Takakura, Y., Masuda, S., Tokuda, H., Nishikawa, M., Hashida, M. (1994) Targeted delivery of superoxide dismutase to macrophages via mannose receptor-mediated mechanism. Biochem. Pharmacol. 47, 853-8. (4) Kondo, S., Segawa, T., Tanaka, K., Izawa, K., Hashida, M., Kanematsu, T. (1996) Mannosylated superoxide dismutase inhibits hepatic reperfusion injury in rats. J. Surg. Res. 60, 36-40. (5) Goodson, R. J., Katre, N. V. (1990) Site-directed pegylation of recombinant interleukin-2 at its glycosylation site. Biotechnology (N. Y.) 8, 343-6. (6) Chilkoti, A., Chen, G., Stayton, P. S., Hoffman, A. S. (1994) Site-specific conjugation of a temperature-sensitive polymer to a genetically engineered protein. Bioconjugate Chem. 5, 504-7. (7) Kuan, C. T., Wang, Q. C., Pastan, I. (1994) Pseudomonas exotoxin A mutants. Replacement of surface exposed residues in domain II with cysteine residues that can be modified with poly(ethylene glycol) in a site-specific manner. J. Biol. Chem. 269, 7610-6. (8) Geoghegan, K. F., Stroh, J. G. (1992) Site-directed conjugation of nonpeptide groups to peptides and proteins via periodate oxidation of a 2-amino alcohol. Application to modification at N-terminal serine. Bioconjugate Chem. 3, 138-46. (9) Gaertner, H. F., Rose, K., Cotton, R., Timms, D., Camble, R., Offord, R. E. (1992) Construction of protein analogues by site-specific condensation of unprotected fragments. Bioconjugate Chem. 3, 262-8. (10) Sato, H., Ikeda, M., Suzuki, K., Hirayama, K. (1996) Sitespecific modification of interleukin-2 by the combined use of genetic engineering techniques and transglutaminase. Biochemistry 35, 13072-80. (11) Sato, H., Yamada, N., Shimba, N., Takahara, Y. (2000) Unique substrate specificities of two adjacent glutamine residues in EAQQIVM for transglutaminase: identification and characterization of the reaction products by electrospray ionization tandem mass spectrometry. Anal. Biochem. 281, 68-76.

710 Bioconjugate Chem., Vol. 12, No. 5, 2001 (12) Sato, H., Yamamoto, K., Hayashi, E., Takahara, Y. (2000) Transglutaminase-mediated dual and site-specific incorporation of poly(ethylene glycol) derivatives into a chimeric interleukin-2. Bioconjugate Chem. 11, 502-9. (13) Kanaji, T., Ozaki, H., Takao, T., Kawajiri, H., Ide, H., Motoki, M., Shimonishi, Y. (1993) Primary structure of microbial transglutaminase from Streptoverticillium sp. strain s-8112, J. Biol. Chem. 268, 11565-72. (14) Ando, H., Adachi, M., Umeda, K., Matsuura, A., Nonaka, M., Uchio, R., Tanaka, H., Motoki, M. (1989) Purification and characterization of a novel transglutaminase derived from microorganisms. Agric. Biol. Chem. 53, 2613-2617. (15) Murahashi, N., Ishihara, H., Sasaki, A., Sakagami, M., Hamana, H. (1997) Hepatic accumulation of glutamic acid branched neogalactosyllipid modified liposomes. Biol. Pharm. Bull. 20, 259-66. (16) Ashwell, G., Harford, J.(1982) Carbohydrate-specific receptors of the liver. Annu. Rev. Biochem. 51, 531-554. (17) Spiess, M. (1990) The asialoglycoprotein receptor: a model for endocytic transport receptors. Biochemistry 29 (43), 10009-10018. (18) Tsuji, T., Nakagawa, R., Sugimoto, N., Fukuhara, K. (1987) Characterization of disulfide bonds in recombinant proteins: reduced human interleukin 2 in inclusion bodies and its oxidative refolding. Biochemistry 26, 3129-34. (19) Gillis, S., Ferm, M. M., Ou, W., Smith, K. A. (1978) T cell growth factor: parameters of production and a quantitative microassay for activity. J. Immunol. 120, 2027-32. (20) Folk, J. E., Cole, P. W. (1966) Mechanism of action of guinea pig liver transglutaminase. I. Purification and properties of the enzyme: identification of a functional cysteine essential for activity. J. Biol. Chem. 241, 5518-25. (21) Nonaka, M., Matsuura, Y., Motoki, M. (1996) Incorporation of lysine- and lysine dideptides into Rs1-casein by Ca2+independent microbial transglutaminase. Biosci. Biotechnol. Biochem. 60, 131-133. (22) Folk, J. E. (1983) Mechanism and basis for specificity of transglutaminase-catalyzed epsilon-(gamma-glutamyl) lysine bond formation. Adv. Enzymol. Relat. Areas Mol. Biol. 54, 1-56. (23) Lorand, L., Parameswaran, K. N., Stenberg, P., Tong, Y. S., Velasco, P. T., Jonsson, N. A., Mikiver, L., Moses, P. (1979) Specificity of guinea pig liver transglutaminase for amine substrates. Biochemistry 18, 1756-65. (24) Gorman, J. J., Folk, J. E. (1984) Structural features of glutamine substrates for transglutaminases. Role of extended interactions in the specificity of human plasma factor XIIIa and of the guinea pig liver enzyme. J. Biol. Chem. 259, 900710.

Sato et al. (25) Brandhuber, B. J., Boone, T., Kenney, W. C., McKay, D. B. (1987) Three-dimensional structure of interleukin-2. Science 238, 1707-9. (26) Aeschlimann, D., Paulsson, M., Mann, K. (1992) Identification of Gln726 in nidogen as the amine acceptor in transglutaminase-catalyzed cross-linking of laminin-nidogen complexes. J. Biol. Chem. 267, 11316-21. (27) Terwilliger, T. C., Eisenberg, D. (1982) The structure of melittin. II. Interpretation of the structure. J. Biol. Chem. 257, 6016-22. (28) Sasaki, K., Dockerill, S., Adamiak, D. A., Tickle, I. J., Blundell, T. (1975) X-ray analysis of glucagon and its relationship to receptor binding. Nature 257, 751-7. (29) Mott, H. R., Baines, B. S., Hall, R. M., Cooke, R. M., Driscoll, P. C., Weir, M. P., Campbell, I. D. (1995) The solution structure of the F42A mutant of human interleukin 2. J. Mol. Biol. 247, 979-94. (30) Bazan, J. F. (1992) Unraveling the structure of IL-2 [letter]. Science 257, 410-3. (31) Drugs R D. 2 (1999) P 278-280, PEG interferon-alpha2a. PEG IFN-alpha-2a, Pegasys, RO 253036. (32) van der Sluijs, P., Bootsma, H. P., Postema, B., Moolenaar, F., Meijer, D. K. (1986) Drug targeting to the liver with lactosylated albumins: does the glycoprotein target the drug or is the drug targeting the glycoprotein? Hepatology 6, 7238. (33) Fujita, T., Nishikawa, M., Ohtsubo, Y., Ohno, J., Takakura, Y., Sezaki, H., Hashida, M. (1994) Control of in vivo fate of albumin derivatives utilizing combined chemical modification. J. Drug Target. 2, 157-65. (34) Haensler, J., Szoka, F. C., Jr. (1993) Synthesis and characterization of a trigalactosylated bisacridine compound to target DNA to hepatocytes. Bioconjugate Chem. 4, 85-93. (35) Plank, C., Zatloukal, K., Cotten, M., Mechtler, K., and Wagner, E. (1992) Gene transfer into hepatocytes using asialoglycoprotein receptor mediated endocytosis of DNA complexed with an artificial tetra-antennary galactose ligand. Bioconjugate Chem. 3. 533-9. (36) Biessen, E. A., Vietsch, H., Van Berkel, T. J. (1994) Cholesterol derivative of a new triantennary cluster galactoside directs low- and high-density lipoproteins to the parenchymal liver cell. Biochem. J. 302, 283-9. (37) Sliedregt, L. A., Rensen, P. C., Rump, E. T., Van Santbrink, P. J., Bijsterbosch, M. K., Valentijn, A. R., Van der Marel, G. A., Van Boom, J. H., Van Berkel, T. J., Biessen, E. A. (1999) Design and synthesis of novel amphiphilic dendric galactosides for selective targeting of liposomes to the hepatic asialoglycoprotein receptor. J. Med. Chem. 42, 609-18.

BC000132H