Biomacromolecules 2001, 2, 1110-1115
1110
Polylactones. 55. A-B-A Triblock Copolymers of Various Polypeptides. Syntheses Involving 4-Aminobenzoyl-Terminated Poly(E-caprolactone) as B Block Hans R. Kricheldorf* and Karsten Hauser Institut fu¨r Technische und Makromolekulare Chemie, Bundesstrasse 45, D-20146 Hamburg, Germany Received March 19, 2001; Revised Manuscript Received July 10, 2001
A telechelic poly(-caprolactone) having a degree of polymerization (DP) around 25 and two 4-aminobenzoyl chain ends was used as a macroinitiator for the ring-opening polymerization of various R-amino acid N-carboxyanhydrides (NCAs). Glycine-NCA, L-alanine-NCA, L-phenylalanine-NCA, and γ-benzyl-Lglutamate-NCA served as monomers and the NCA/macroinitiator ratio was varied between 20:1, 40:1, and 100:1. In the case of L-Phe-NCA, a ratio of 200:1 was also used. It was demonstrated by means of model studies, that the NCAs may react almost quantitatively with the 4-aminobenzoyl end groups despite their relatively low nucleophilicity. The isolated triblock copolymers were characterized by viscosity measurements and by 1H NMR spectroscopy with regard to their composition (which in most cases paralleled the feed ratios). However, in the case of γ-Bzl-Glu-NCA mixtures of di- and triblock copolymer were obtained. The secondary structures of the solid copolymers were examined by IR spectroscopy and 13C NMR CP/ MAS spectroscopy. It was found that the R-helix/β-sheet ratio of the poly(L-Ala) and poly(L-Phe) blocks increases with their average length, according to the NCA/macroinitiator ratio. Introduction Over the past 3 decades biodegradable polymers have found several important pharmaceutical and medical applications,1-6 and far more potential applications are currently under investigation. For most applications the properties of the homopolymers such as poly(-CL) or polylactide need to be modified. Syntheses of random or block copolymers is the most obvious and promising strategy to meet the requirements of individual applications. Block copolymers consisting of polylactone and polyether blocks or consisting of two different polylactone blocks have been synthesized and characterized by numerous research groups and chemical companies. Very little research activity was focused on block copolymers consisting of polylactone and polypeptide blocks.7 The combination with peptide blocks introduces a new quality of hydrolytic stability into the copolymers, because enzymatic degradation is required, whereas water alone (even at pH 5-8) does not hydrolyze the peptide bonds. Furthermore, the H bonds may undergo favorable interaction with other peptides or proteins, allowing for stabilization of peptide hormones or for specific binding to tissues and proteinous surfaces. In a previous publication,8 we have reported on the synthesis of telechelic poly(-CL) having two 4-aminobenzoyl end groups starting from a ring-expansion polymerization of -CL (Scheme 1). In addition to an easy synthesis, the 4-aminobenzoyl end groups have the advantage that their low nucleophilicity prevents an aminolytic cleavage of the poly(-CL) chains during synthesis and storage. The relatively low nucleophilicity of the 4-aminobenzoyl end groups (when compared to R-amino acid chain ends) raises the
question, if they are reactive enough as initiators of NCAs, so that A-B-A triblock copolymers can be obtained (Scheme 2). If the rate of initiation is too slow relative to the rate of propagation a mixture of triblock copolymers, diblock copolymers and unreacted poly(-CL) may be the result. Therefore, it was the main purpose of this work to find out, if the telechelics of structure 1 are useful as macroinitiators for the preparation of A-B-A type peptide block copolymers. The problems and features of primary amine-initiated polymerizations of NCAs have recently been reviewed9,10 and, thus, do not need here a detailed discussion. Various kinds of diblock and A-B-A triblock copolymers containing polypeptide blocks have been reported in the literature. Yet, to the best of our knowledge, only one publication has appeared so far7 dealing with block copolymers of polypeptides and a biodegradable polyester. In that publication, amine-encapped poly(L-lactide) was used as macroinitiator of several NCAs yielding diblock copolymers. Experimental Section Materials. -Caprolactone (-CL) was purchased from Aldrich Co. (Milwaukee, WI) and distilled over freshly powdered calciumhydride in vacuo. The previously described 4-aminobenzoyl-terminated poly(-caprolactone), poly(-CL), having a DP of 24 ( 1, was used for all polymerizations.8 Glycine, L-alanine, L-phenylalanine, and L-glutamic acid were gifts of DEGUSSA AG (Hanau, Germany). Gly-NCA, L-Ala-NCA, and L-Phe-NCA were prepared from the silylated N-methoxycarbonyl derivatives by means of PBr3 as described previously.10 They were twice recrystallized from tetrahydrofuran/ligroin, whereby dry charcoal was used
10.1021/bm0100561 CCC: $20.00 © 2001 American Chemical Society Published on Web 09/15/2001
A-B-A Triblock Copolymers
Biomacromolecules, Vol. 2, No. 4, 2001 1111
Scheme 1
Scheme 2
during the first recrystallization. γ-Benzyl L-glutamate-NCA (γ-Bzl-Glu-NCA) was prepared according to the literature,11,12 except that commercial diphosgene (Aldrich Co.) was used instead of phosgene. The melting points of all NCAs agreed with the literature data,9 and the 1H NMR spectra (recorded in dry acetone-d6) agreed with the expected structure and purity. N-Butyl-4-aminobenzoate was purchased from Aldrich Co. and used as received. n-Butyl (N-Chloroacetyl)-4-aminobenzoate. n-Butyl 4-aminobenzoate (0.05 mol) and pyridine (0.05 mol) were dissolved in dry ethyl acetate (120 mL), and chloroacetyl chloride (0.05 mol) diluted with dry ethyl acetate (30 mL) was added dropwise with stirring. The reaction mixture was stirred for 20 h at 20 °C and was then filtered from the precipitated pyridine hydrochloride. The filtrate was concentrated in a vacuum until the product began to crystallize. It was isolated by filtration and dried at 40 °C in a vacuum. Polymerizations of the NCAs. A 100 mL glass reactor equipped with one gas inlet/outlet tube was silanized using dichlorodimethylsilane, followed by repeated washings with dry toluene and dry diethyl ether. The reactor was dried for 2 h at 60 °C in a vacuum (p ) 0.1 mbar) and, then, refilled with dry argon. A solution of NCA (10 mmol) in 50 mL of dry dioxane or dichloromethane and the required amount (0.5, 0.25, or 0.1 mmol) of the telechelic poly(-caprolactone) having two 4-aminobenzoyl end groups were introduced into the reactor under dry argon. The reaction mixture was stirred until the poly(-caprolactone) was dissolved, and then, the reaction vessel was allowed to stand for 5 days at room temperature. Afterward, the reaction mixture was poured into 500 mL diethyl ether under strong stirring. The precipitated polymer was isolated by filtration and dried at 40 °C in a vacuum.
Measurements. The inherent viscosities were measured with a automated Ubbelohde viscometer thermostated at 25 °C. The 400 MHz 1H NMR spectra were recorded with a Bruker AM-400 FT NMR spectrometer in 5 mm o.d. sample tubes. CDCl3, CDCl3/TFA mixtures, or DMSO-d6 containing TMS served as solvent and shift reference. All 75.5 MHz 13C NMR CP/MAS spectra were obtained on a Bruker MSL 300 FT-NMR spectrometer. Samples of 200-250 mg were measured in 7 mm rotors made of ZrO2 at spinning rates between 4.2 and 4.9 MHz. A contact time of 3 ms and a repetition time of 4 s were used as “standard conditions”. The chemical shifts were referenced to Me4Si as follows. The carbonyl signal of solid glycine was used as tertiary standard (170.09 ppm) and referenced to solid adamantane. The shift of solid adamantane was set equal to that of an adamantane solution in chloroform containing Me4Si. The IR spectra were recorded from KBr pellets on a Nicolet “Impact 410” FT-IR spectrometer. Results and Discussion Model Studies and Polymerizations. To find out if the 4-aminobenzoyl end groups react almost quantitatively with the NCAs, 1H NMR spectroscopy seemed to be the most promising analytical method. Acylation of the amino group should cause a downfield shift of the neighboring aromatic protons (ortho to the nitrogen). To check the extent of this shift effect, commercial n-butyl 4-aminobenzoate was acylated with chloroacetyl chloride. The 1H NMR spectra presented in Figure 1 illustrate the significant downfield shift of the “f-protrons” after acylation (from 6.65 to 7.60 ppm). The telechelic poly(-CL) 1 was then used as initiator of the ring-opening polymerization of Gly-NCA in dioxane at 20 °C, and the monomer macroinitiator ratio was varied from 20:1 over 40:1 to 100:1 (see nos. 1-3 and footnote a in Table 1). For the 20:1 copolymer, a weak signal of the aromatic protons ortho to the nitrogen of unreacted 4-aminobenzoyl units was still detectable around 6.65 ppm (compare f-protons in Figure 1A). However, for the 40:1 and 100:1 copolymers, no such signal was observed (Figure 2), and thus, the spectra of these copolymers clearly prove that at least for a monomer/macroinitiator ratio g40:1 a quantitative acylation has taken place. This result means, in turn, that
1112
Biomacromolecules, Vol. 2, No. 4, 2001
Kricheldorf and Hauser
Figure 1. 400 MHz 1H NMR spectra (recorded in CDCl3/TMS) of (A) 4-aminobenzoic acid butyl ester; (B) N-(chloroacetyl)-4-aminobenzoic acid butylester. Table 1. Amino Acid/-Caprolactone Triblock Copolymers, Initiated with a Poly(-CL) Having Two p-Aminobenzoyl End Groups polymer no. amino acid NCA:EGa DPexpb yieldc (%) ηinh (dL/g) 1 2 3 4 5 6 7 8 9 10 11 12 13
Gly Gly Gly L-Phe L-Phe L-Phe L-Phe L-Ala L-Ala L-Ala γ-Bzl-L-Glu γ-Bzl-L-Glu γ-Bzl-L-Glu
10:1 20:1 50:1 10:1 20:1 50:1 100:1 10:1 20:1 50:1 10:1 20:1 50:1
13 24 46 13 25 48 102 15 27 55 17 38 82
68 71 69 75 72 67 80 72 76 69 66 60 70
0.32d 0.33d 0.28d 0.23e 0.20e 0.29e 0.36e 0.23d 0.21d 0.24d 0.14f 0.17f 0.21f
a Units: mol of NCA/mol of NH end groups ()1/2 monomer/macro2 initiator ratio). b Determined by 1H NMR spectroscopy. c Yield after the d precipitation into diethyl ether. Measured at 25 °C with c ) 2 g/L in CH2Cl2:TFA ) 2:1. e Measured at 25 °C with c ) 2 g/L in CH2Cl2:TFA ) 4:1. f Measured at 25 °C with c ) 2 g/L in DMF.
true A-B-A triblock copolymers were indeed obtained. The signal assignments of the NH protons “a”, “b”, and “c” in Figure 2 are based on two aspects: first, their intensity ratios and second their reduced intensity when deuterated trifluoroacetic acid (CF3CO2D) was used as cosolvent (resulting in a partial H/D exchange).
With L-Ala-NCA, an analogous series of polymerizations was performed (nos. 8-10, Table 1) and analogous results were obtained. The 1H NMR spectra proved again that for monomer/initiator ratios g40:1 a quantitative reaction of the 4-aminobenzoyl end groups had taken place. However these NMR measurements had to be performed in neat CF3CO2D, because the NH-signal of the poly(L-Ala) chains obscured the aromatic signals of the 4-aminobenzoyl unit, when a mixture of CDCl3 and CF3CO2D was used as in the case of Gly-NCA. When L-Phe-NCA was used as monomer, a M/I ratio of 200:1 was included in the standard series of polymerizations. Regardless if L-Phe-NCA or γ-Bzl-L-GluNCA were polymerized, the extent of the reaction with 4-aminobenzoyl end groups could not be determined, because the strong signals of the aromatic side chains and the NH signal obscured the signals of the 4-aminobenzoyl groups. The chemical composition of the isolated block copolymers was determined by 1H NMR spectroscopy. From the signal intensities of the amino acids and the signal intensities of -CL or 4-aminobenzoyl units an apparent DP was calculated. In the case of Gly, L-Ala, and L-Phe, the experimental DPs are somewhat higher than the monomer/ macroinitiator ratios for low feed ratios of 20:1 and 40:1. This difference may be attributed to a fractionation effect. Since both the telechelic poly(-CL)s and the polypeptides
A-B-A Triblock Copolymers
Biomacromolecules, Vol. 2, No. 4, 2001 1113
Figure 2. 400 MHz 1H NMR spectrum (recorded in CDCl3:TFA ) 2:1) of a glycine/-caprolactone triblock copolymer (no. 2, Table 1).
should possess a certain noncorrelated molecular weight distribution, those block copolymers containing relatively long poly(-CL) segments and shorter peptide blocks possess a higher solubility in organic solvents and will not precipitate from the reaction mixture. Furthermore, it may be taken into account that diblock copolymers, if present at all, will be more soluble in the “nonsolvent” than the triblock copolymers. Hence, an “extraction” of these diblock copolymers might have contributed to the somewhat higher than calculated DPs. However, 1H NMR spectra of the soluble fraction did not give any indication of unreacted 4-aminobenzoyl groups. Therefore, the assumption of “extracted” diblock copolymers has no experimental basis. The rather high DPs of the γ-Bzl-Glu-copolymers may partially be the result of an incomplete conversion of the initiator. In contrast to the other peptides studied in this work the polyglutamate has a relatively good solubility in dioxane and adopts an R-helical conformation after a DP of 8.9 Therefore, the 4-aminobenzoyl groups have to compete with the more reactive amino groups of the growing glutamate chains at any time and lose this competition. In the case of the other three NCAs, the growing oligopeptides form β-sheet structures, which precipitate from the reaction mixture (as observed). The reactivity of the amino end groups on the surface of the β-sheet structures is sterically hindered as demonstrated and discussed in the literature many times.9,10 Therefore, the aminobenzoyl end groups of the telechelic poly(-CL) 1 can here successfully compete with the growing peptide chains. The poor solubility of polypeptides in organic solvents was a serious hindrance for the determination of absolute molecular weights. On the other hand, it allowed us to check if unreacted poly(-CL) 1 can be extracted from the isolated
Figure 3. IR spectra of two L-alanine/-caprolactone triblock copolymers: (A) NCA/end group ratio ) 10:1 (no. 8, Table 1); (B) NCA/end group ratio ) 50:1 (no. 10, Table 1).
block copolymers by means of chloroform. However, this test proved to be negative. The inherent viscosities of the Gly, L-Ala, and L-Phe block copolymers were recorded in mixtures of chloroform and trifluoroacetic acid (TFA). All values were higher than that of the macroinitiator 1 (having a ηinh value of 0.15 dL/g in CHCl3/TFA 4:1). Therefore, these inherent viscosities and extraction experiments also agree with a covalent binding between polypeptide and poly(CL) blocks. However, within a series of increasing M/I ratios, no simple trend was observed. These inconsistencies may be explained by conformational changes, such as the formation of hairpin structures with intramolecular H bonds in the case of longer poly(Gly) chains. Poly(L-Ala) and poly(L-Phe) form R-helices above a DP of 10 even in neat TFA.9,10 Poly(L-Phe) is insoluble in neat TFA or CHCl3/TFA mixtures regardless of the secondary structure. Therefore, the complete solubility of the L-Phe block copolymers in CHCl3/TFA may be taken as experimental evidence, that neat poly(L-Phe) chains not attached to the macroinitiator are absent. The lower inherent viscosities of the L-Ala or L-Phe block copolymers prepared with M/I ) 40 relative to those prepared with M/I ) 20 may be explained by formation of R-helices which represents shrinkage of the hydrodynamic volume relative to fully solvated loose chains.
1114
Biomacromolecules, Vol. 2, No. 4, 2001
Table 2. Chemical Shifts (δ) of Poly(R-amino Acid)s in NMR-CP/MAS Spectra
Kricheldorf and Hauser 13C
poly(R-amino δ of the R- and 31-helixa δ of the β-sheet structurea acid) CO R-C β-C CO R-C β-C (Gly)n (L-Ala)n (L-Phe)n (γ-Bzl-L-Glu)n a
172.1 168.9 176.8 175.2 175.4
42.0 42.9 52.8 61.3 56.8
15.5 35.0 25.9
169.2
44.3
172.2 169.0 172.2
49.3 53.2 51.1
20.3 39.3 29.7
Chemical shifts δ in ppm.
Figure 4. 75.5 MHz 13C NMR CP/MAS spectrum of a poly-caprolactone having two 4-aminobenzoyl end groups 1.
Secondary Structures of the Solid State. The secondary structure of the solid block copolymers was characterized by two methods. It has been known for several decades13-18 that the IR bands resulting from the vibration of the peptide groups are sensitive to conformational changes such as the R-helix vs the β-sheet structure. The IR spectra of the block copolymers confirmed that the poly(Gly) blocks exclusively existed in the β-sheet structure and the poly(γ-Bzl-L-Glu) blocks mainly (but not exclusively) in the R-helical structure. As illustrated by Figure 3, the R-helix/β-sheet ratio of the poly(L-Ala) blocks increased with higher M/I ratios of the NCA polymerization. This finding is in perfect agreement with the expected triblock structure and an increasing length of the poly(L-Ala) blocks attached to the macroinitiator. An analogous but less pronounced trend was observed in the IR spectra of the L-Phe block copolymers. In addition to IR spectroscopy, 13C NMR CP/MAS spectroscopy was used to characterize the secondary structure of the solid copolymers. As published previously,19-24 the 13C NMR signals of the CO, C-R, and C-β carbons are sensitive to conformational changes of the solid polypeptides. The chemical shift values of the relevant conformations are summarized in Table 2. Figure 4 displays the 13C NMR spectrum of the macroinitiator for comparison. Figure 5, parts A and B, demonstrates that the poly(Gly) blocks exist in the β-sheet structure and the poly(γ-Bzl-L-Glu) blocks predominantly (around 90%) in the R-helical conformation. The CO signals in Figure 6, parts A and B, indicate that the R-helix/β-sheet ratio of the poly(L-Phe) blocks increases with higher M/I ratios. An analogous result was found for the L-Ala block copolymers. Therefore, both the IR and the 13C
Figure 5. 75.5 MHz 13C NMR CP/MAS spectra (A) a glycine/caprolactone triblock copolymer (no. 3, Table 1); (B) a γ-benzyl-Lglutamate/-caprolactone triblock copolymer (no. 11, Table 1).
Figure 6. 75.5 MHz 13C NMR CP/MAS spectra of two L-phenylalanine/-caprolactone triblock copolymers, A: NCA/end group ratio ) 10:1 (no. 4, Table 1) B: NCA/end group ratio ) 100:1 (no. 7, Table 1).
NMR spectra confirm that the lengths of the poly(L-Ala) and poly(L-Phe) blocks increase with higher M/I ratios. This observation is a good argument against the hypothesis that the growing peptide chains cleaved the poly(-CL) blocks
Biomacromolecules, Vol. 2, No. 4, 2001 1115
A-B-A Triblock Copolymers Scheme 3
preparation of A-B-A triblock copolymers (2) when GlyNCA, L-Ala-NCA, or L-Phe-NCA are used as reaction partners. In these cases the initially formed oligopeptides form β-sheet structures with the consequence that the reactivity of the amino end groups is reduced by steric hindrance. In the case of the soluble oligo(γ-Bzl-L-Glu) blocks, the formation of β-sheet structures is less pronounced, and the propagation is faster than the initiation over the whole course of the NCA polymerization. Hence, part of the macroinitiator remains unchanged. As will be demonstrated in another publication25 this problem will be overcome with a telechelic poly(-CL) having more reactive aliphatic amino groups. References and Notes
Table 3. Fractions of R-Helix and β-Sheet Structure of All Triblock Copolymers polymer no. 1 2 3 4 5 6 7 8 9 10 11 12 13
fractions of R-helix (%)
fraction of β-sheet (%)
0-5 0-5 0-5 20 ( 10
95-100 95-100 95-100 80 ( 10
35 ( 10 60 ( 10 40 ( 10 55 ( 10 75 ( 10 90 ( 5 95 ( 5 95 ( 5
65 ( 10 40 ( 10 60 ( 10 45 ( 10 25 ( 10 10 ( 5 5(5 5(5
by aminolysis of the ester groups followed by a reinitiation of the NCAs via the liberated HO-CH2 end groups (Scheme 3). Such a series of side reaction would result in the generation of numerous short peptide blocks which in the case of L-Ala and L-Phe cannot adopt an R-helix structure. In summary, the conformations observed for the solid block copolymers are in satisfactory agreement with the expected triblock structure. A rough calculation of the fractions of R-helix and β-sheet structure of all block copolymers is compiled in Table 3. Conclusion The results obtained in this work allow the conclusions that the telechelic poly(-CL) 1 is an useful initiator for the
(1) Targeting of Drugs; Gregoriadis, G., Senior, J., Trouet, A., Eds.; Plenum Press: New York and London, 1982. (2) Degradation Phenomena on Polymeric Biomaterials; Planck, H., Dauner, M., Renardy, M., Eds.; Springer-Verlag: Berlin, Heidelberg, Germany, and New York, 1992. (3) Plastics from Microbs; Mobley, D. P.; Ed.; Hanser Publ.: Mu¨nchen, Wien, and New York, 1994. (4) Kricheldorf, H. R.; Kreiser-Saunders, I. Macromol. Symp. 1996, 103, 85. (5) Degradable Polymers; Albertsson, A.-C.; Karlson, S.; Eds.; Hu¨thig & Wepl Verlag: Heidelberg, Oxford, England, 1998. (6) Schroeter, J. Kunstoffe 2000, 90, 1. (7) Gotsche, M.; Keul, H.; Ho¨cker, H. Macromol. Chem. Phys. 1995, 196, 3891. (8) Kricheldorf, H. R.; Hauser, K. Macromolecules 1998, 31, 614. (9) Kricheldorf, H. R. R-Amino Acid N-Carboxyanhydrides and Related Heterocycles; Springer-Verlag: Berlin, Heidelberg, Germany, and, New York, 1987; Chapter 2. (10) Kricheldorf, H. R. In Models of Biopolymers by Ring-Opening Polymerization; Denczek, S., Ed.; CRC Press: Boca Raton FL, 1990; Chapter 1. (11) Blout, E. R.; Karlson, R. H. J. Am. Chem. Soc. 1956, 78, 941. (12) Overell, B. G.; Petrow, V. J. Chem. Soc. 1959, 232. (13) Krimm, S.; Kuroina, K. Biopolymers 1968, 6, 401. (14) Komoto, T.; Kiu, K. Y.; Minoshima, Y.; Oya, M.; Kawai, T.; Makromol. Chem. 1973, 168, 261. (15) Moore, W. H.; Krimm, S. Biopolymers 1976, 15, 2465. (16) Tonioli, C.; Palunbo, M. Biopolymers 1977, 16, 219. (17) Hoh, K.; Nakahara, T.; Shimanouchi, T.; Oya, M.; Uno, K.; Iwakura, Y. Biopolymers 1968, 6, 1759. (18) Heitz, F.; Marchal, E.; Spach, G. Macromolecules 1975, 8, 145. (19) Kricheldorf, H. R.; Mu¨ller, D. Polym. Bull. 1981, 6, 101. (20) Kricheldorf, H. R.; Mu¨ller, D. Polym. Bull. 1982, 8, 495. (21) Kricheldorf, H. R.; Mutter, M.; Mayer, F.; Mu¨ller, D.; Fo¨rster, H. Biopolymers 1983, 22, 1357. (22) Kricheldorf, H. R.; Mu¨ller, D.; Stulz, J. Makromol. Chem. 1983, 184, 1407. (23) Kricheldorf, H. R.; Mu¨ller, D. Macromolecules 1983, 16, 615. (24) Ando, I.; Saito, H.; Tabete, R.; Shoji, A.; Ozaki, T. Macromolecules 1984, 17, 457. (25) Kricheldorf, H. R.; Hauser, K. Manuscript in preparation.
BM0100561
