300
Bioconjugate Chem. 1998, 9, 300−303
TECHNICAL NOTES Synthesis of Heterobifunctional Poly(ethylene glycol) with a Reducing Monosaccharide Residue at One End Teruo Nakamura, Yukio Nagasaki, and Kazunori Kataoka* Department of Materials Science and Technology, Science University of Tokyo, Noda 278, Japan. Received September 15, 1997; Revised Manuscript Received December 15, 1997
A new synthetic method for a heterobifunctional poly(ethylene glycol) (PEG) having a monosaccharide moiety at one end was created. PEG with a reducing monosaccharide residue at the R-end, which is linked to a defined position of the sugar molecule, could be prepared via the anionic polymerization of ethylene oxide (EO) initiated with a potassium alkolate of a protected monosaccharide such as 1,2;5,6-di-O-isopropylidene-D-glucofuranose (DIGL), 1,2;3,4-di-O-isopropylidene-D-galactopyranose (DIGA), and 1,2-O-isopropylidene-3,5-O-benzylidene-D-glucofuranose (IBGL). The resulting PEGs possess the corresponding sugar molecule at the R-chain end and a hydroxyl group at the ω-chain end. The ω-chain end could be converted to several functional groups such as allyl, amino, and hydroxycarbonyl groups in high yield. Such heterobifunctional PEGs possessing a reducing monosaccharide residue at the R-end are one of the promising tools for bioconjugate chemistries.
INTRODUCTION
Poly(ethylene glycol) (PEG) is a polymer possessing peculiar physicochemical properties such as high water solubility and flexibility. In addition, its bioinert characteristic increases an opportunity for biomedical applications. Chemical modification of biologically active compounds with PEG, designated PEGylation, led to alteration of their pharmacokinetics (1). As we reported previously, polymeric micelles, composed of PEG-poly(peptide) block copolymers, behave as long-circulating vehicles of antitumor agents in vivo (2). The hydrophilic character of the PEG outer shell along with the small size of the micelle (approximately tens of nanometers) confers a reduced nonspecific uptake to reticuloendothelial systems. By installing homing devices onto the surface of the micelles, we enhanced cellular uptake at a target tissue. Carbohydrates are known as one of the promising candidates for a homing moiety. Recent progress in glycobiology revealed that carbohydrateprotein interaction requires a ligand with well-defined structural compatibility (3). For instance, the permeability of intestinal epithelium to macromolecular substances was improved by installing a reducing glucose residue rather than a nonreducing moiety on them (4). This enhanced permeability may be attributed to receptor-mediated endocytosis via glucose transporters, mediating the facilitative diffusion of glucose, through specific binding with the reducing glucose residue. This result suggests that the designed sugar moieties can be utilized as a targeting device. However, to design a sugar derivative as a homing unit, regiospecific conjugation as mentioned above and spatial arrangement of plural sugar residues are key factors in demonstrating the target function. If an accommodating spacer, such as PEG, can * To whom all correspondence should be addressed.
be introduced at a sugar residue site-specifically, it will provide a superior device for incorporating a potent targeting moiety into the compounds desired to be delivered. Thus, we report herein a new synthetic route for a heterobifunctional PEG with a reducing monosaccharide residue at one end with a defined position linkage of PEG and the sugar moiety. A functional group was introduced at the other end for further derivatization. Our strategy is to initiate anionic polymerization of ethylene oxide (EO) from the hydroxyl of site-specifically protected sugar molecules. EXPERIMENTAL PROCEDURES
Polymerization of EO from Protected Sugars. To a solution of 1,2;5,6-di-O-isopropylidene-D-glucofuranose (DIGL) (5) (260 mg, 1 mmol) in tetrahydrofuran (THF) was added potassium naphthalene (1 mmol) as a THF solution to convert a hydroxyl group at the C-3 position to a potassium salt. EO (115 mmol) was added to the solution via a cooled syringe, and the mixture was stirred in a water bath for 2 days. The reaction mixture was then poured into ice-cold ether (approximately 1 L) to obtain a white precipitate. Polymerizations of EO with the potassium alkolate form of 1,2;3,4-di-O-isopropylidene-D-galactopyranose (DIGA) (5) and 1,2-O-isopropylidene-3,5-O-benzylidene-D-glucofuranose (IBGL) (6) were carried out in a procedure similar to the DIGL-initiated EO polymerization. Removal of Protective Groups from the Sugar End Group. PEG with a sugar derivative end group was dissolved in 8:2 (v/v) trifluoroacetic acid-water and allowed to stand for 1 h at room temperature. The reaction solution was diluted with water, and the solvent was removed under reduced pressure followed by lyophilization from water.
S1043-1802(97)00179-1 CCC: $15.00 © 1998 American Chemical Society Published on Web 02/05/1998
Technical Notes
Bioconjugate Chem., Vol. 9, No. 2, 1998 301 Table 1. Results of Polymerization of Ethylene Oxide (EO) Initiated with Potassium Alkoxide of DIGLa Mn
run
feed [EO]0/[DIGL]0
time (d)
calcd
obsd
Mw/Mn
yield (%)
1 2 3
60 115 230
2 2 4
2600 5000 10000
2400 4300 7600
1.05 1.09 1.17
88 83 80
a
Mn and Mw/Mn were determined by GPC.
Figure 1. GPC chromatogram of PEG with a DIGL end group (the same sample as run 2 in Table 1).
ω-End Functionalization of Sugar-PEG. After the polymerization of EO (115 mmol) initiated with DIGL (130 mg, 0.5 mmol) in THF, a THF solution of potassium naphthalene (1 mmol) was added to the polymerization mixture and the mixture stirred at room temperature for 30 min. To the solution was added allyl bromide (0.43 mL, 5 mmol) via a syringe, and the mixture was stirred at room temperature for 4 h. Then, the reaction solution was mixed with chloroform and washed with a saturated NaCl aqueous solution several times to eliminate impurities from the PEG derivative. The organic phase was dried with sodium sulfate followed by concentration and then poured into ice-cold ether. The resultant white precipitate was dried in vacuo. Under these reaction conditions, the conversion of the ω-end from potassium alkoxide to the allyl group was almost complete, forming DIGL-PEG-allyl heteroPEG quantitatively. For the introduction of an amino group at the ω-end, 100 mg of DIGL-PEG-allyl and a 100-fold molar excess of 2-aminoethanethiol hydrochloride were dissolved in degassed water containing 10 wt % sodium chloride and the mixture was stirred in a water bath at 40 °C for 3 h. Purification was carried out in the same way as the sugar-PEG-allyl preparation. RESULTS AND DISCUSSION
As we reported previously, potassium alkoxide possessing an acetal moiety initiates the polymerization of EO to yield PEG with an acetal group at the R-end without any side reaction (7). If the potassium salt of DIGL acts as an initiator for EO polymerization, then a heterobifunctional PEG with sugar derivative at the R-end should be obtained. To confirm the hypothesis, synthesis of R-DIGL-ω-hydroxyl-PEG was attempted. The precipitate obtained after polymerization was dried in vacuo and was analyzed by gel permeation chromatography (GPC) as shown in Figure 1. The resultant material had a number average molecular weight (Mn) of 4300 with a unimodal and narrow molecular weight distribution (MWD) of 1.09 (Mw/Mn). Results of polymerization using DIGL as an initiator under several conditions are summarized in Table 1. The value of the
Figure 2. 400 MHz 1H NMR spectra of PEGs with a sugar residue at the R-end in D2O. DIGL-PEG (upper) and deprotected product Glc(3)-PEG (lower).
Mn could be controlled by the initial monomer/initiator ratio while retaining a narrow MWD. Removal of protective groups from the sugar derivative at the PEG chain end was achieved under acidic condition (8). Figure 2 shows the 1H NMR spectrum of the PEG before and after the TFA treatment. In the lower spectrum, anomeric protons of the reducing glucose residue at 4.66 ppm [H-1(β)] and 5.23 ppm [H-1(R)] were clearly observed at the same time as the disappearance of signals at 1.24, 1.40, 1.45, and 1.52 ppm originating from isopropylidene groups, observed in the upper spectrum, indicating the complete deprotection of the sugar derivative at the end of the polymer. Typical protons of the free sugar residue in Figure 2 (lower) further supported the achievement of cleavage. Worth noticing, revealed by the 1H NMR spectrum of the deprotected polymer, was an increased ratio of R-anomer to β-anomer of the glucose residue on coupling with PEG. The R/β ratio from the 1H NMR spectrum of the polymer is estimated to be 45/55; on the other hand, the ratio is 37/ 63 for intact glucose. The reason for this increased stability of the R-anomer may be the intramolecular interaction of the PEG chain with the sugar moiety. EO polymerizations with IBGL and DIGA as the initiators were also carried out to obtain sugar-ended PEGs with a different linkage position and a different sugar residue, respectively (see Scheme 1). In the case of IBGL, PEG with a hydroxyl group at both chain ends was produced in approximately 35% as a byproduct due to a trace amount of water in IBGL. The larger Mn of the polymer with hydroxyl ends permitted the isolation of PEG with a sugar end group by gel filtration. In the DIGA case, the polymerization succeeded in producing the targeted compound adequately. Removal of protective groups on the sugar derivative on these polymers was also achieved by TFA treatment to convert them to
302 Bioconjugate Chem., Vol. 9, No. 2, 1998
Nakamura et al.
Scheme 1
reducing types. On the whole, the polymerization of EO proceeded smoothly with the sugar derivative initiators. The chances of exploiting the sugar-PEG for drug targeting or surface modification can be improved by ω-end derivatization to an appropriate function. The living nature of the propagating end of the PEG chain allows the introduction of an additional functionality to the ω-end via a successive chemical reaction in the same reaction vessel. The product obtained after ω-end allylation was analyzed by GPC to reveal the formation of the polymer having an Mn of 6000 with a unimodal MWD (Mw/Mn ) 1.06) (Figure 3). In the 1H NMR spectrum of the polymer (Figure 4), signals that can be attributed to DIGL and the allyl group were observed. The signal intensities of methyl protons in the isopropylidene unit, oxymethylene protons in PEG, and methine protons in the allyl end group along with C-1 protons in the sugar moiety were 12, 590, and 2.2, respectively. The Mn of the polymer was then calculated to be 6500, assuming one sugar residue per each polymer chain, which is in good agreement with that from the GPC measurements. This result is consistent with the formation of heterobifunctional PEG possessing sugar and allyl groups at the R- and ω-termini, respectively. The allyl end can be converted to an amino group or a carboxyl group by radical addition of thiol compounds possessing the corresponding functional group. Although the procedure using a radical initiator at 70 °C in anhydrous DMF was previously reported for preparing amino-ended PEG (9), addition of thiols to the allyl group under mild conditions without any catalyst (10) should be attractive from the standpoint of polymer end-group derivatization. With the procedure described herein, the conversion to the amine end was accomplished in 91% yield, estimated by the intensity ratio of isopropylidene to the allyl unit from the 1H NMR spectrum (Figure 4), which was proved by the appearance of two methylene protons of the 2-aminoethanethiol unit at 2.85 and 3.20
Figure 3. GPC chromatogram of PEG with R-DIGL and R-allyl end groups.
ppm with suitable intensities. On the contrary, amination in pure water was achieved in 26% conversion. Formation of a PEG-solute complex may change the chain conformation in the vicinity of the ω-end, enhancing the possibility of thiol access to the allyl group. Derivatization using mercaptoacetic acid proceeded under the same conditions. From the 1H NMR spectrum of the resultant polymer, there was no signal to assign to the allyl group, while the methylene proton of the mercaptoacetic acid unit was clearly observed at 3.25 ppm as a singlet. A great excess of mercaptoacetic acid in the reaction solution, however, partially cleaved isopropylidene on the sugar derivative. The ω-end amina-
Technical Notes
Figure 4. 400 MHz 1H NMR spectrum of PEG with R-DIGL and ω-allyl end groups in CDCl3.
Figure 5. 400 MHz 1H NMR spectrum of ω-aminated PEG with a R-DIGL end group in CDCl3.
tion or hydroxycarbonylation of sugar-PEGs increases their usefulness in the preparation of a variety of bioconjugates. In conclusion, the successful synthesis of a heterobifunctional PEG with a sugar end via EO polymerization using a sugar derivative as an initiator was demonstrated. The sugar residue was introduced onto PEG without any linkage reagent, and the PEG linking
Bioconjugate Chem., Vol. 9, No. 2, 1998 303
position on the sugar was defined by protective groups introduced to the hydroxyl on the sugar molecule. The procedure reported here can be applied to an array of carbohydrates to provide a series of sugar-PEGs. Furthermore, ω-end functionalization of the sugar-PEG was attained. These sugar-PEGs may be applicable to preparation of a wide variety of bioconjugates as well as polymeric micelles with many sugar residues on their surface, which should be quite useful in active targeting of drugs and biologically active compounds. A part of this study was supported by a Grant-in-Aid for Scientific Research on Priority Areas of “SuperBiosystem Constructed by Cognitive Multidimensional Glyco-Molecules” (09240233), The Ministry of Education, Science, Sports and Culture, Japan. LITERATURE CITED (1) Harris, J. M., Ed. (1992) Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications, Plenum Press, New York. (2) Kataoka, K., Kwon, G. S., Yokoyama, M., Okano, T., and Sakurai, Y. (1993) J. Controlled Release 24, 119-132. (3) Lee, Y. C., and Lee, R. T. (1995) Acc. Chem. Res. 28, 321327. (4) Koyama, Y., Ishikawa, M., Iwamoto, M., and Kojima, S. (1992) J. Controlled Release 22, 253-262. (5) Whister, R. L., and Wolfrom, M. L., Eds. (1963) Methods in carbohydrate chemistry, Vol. 2, pp 318-325, Academic Press, New York. (6) Whister, R. L., and Wolfrom, M. L., Eds. (1963) Methods in carbohydrate chemistry, Vol. 1, pp 198-201, Academic Press, New York. (7) Nagasaki, Y., Kutsuna, T., Iijima, M., Kato, M., Kataoka, K., Kitano, S., and Kadoma, Y. (1995) Bioconjugate Chem. 6, 231-233. (8) Christensen, J. E., and Goodman, L. (1968) Carbohydr. Res. 7, 510-512. (9) Cammas, S., Nagasaki, Y., and Kataoka, K. (1995) Bioconjugate Chem. 6, 226-230. (10) (a) Lee, R. T., and Lee. Y. C. (1974) Carbohydr. Res. 37, 193-201. (b) Koyama, Y., Umehara, M., Mizuno, A., Itaba, M., Yasukouchi, T., Natsume, K., Suginaka, A., and Watanabe, K. (1996) Bioconjugate Chem. 7, 298-301.
BC970179B
