Structural Effects of Terminal Groups on Nonenzymatic and Enzymatic

Feb 15, 2008 - Department of Innovative and Engineered Materials, Tokyo Institute of Technology. , ‡. Chemical Analysis Team, RIKEN Institute...
0 downloads 0 Views 1MB Size
Download PDF
Biomacromolecules 2008, 9, 1071–1078

1071

Structural Effects of Terminal Groups on Nonenzymatic and Enzymatic Degradations of End-Capped Poly(L-lactide) Kenji Kurokawa,†,‡ Koichi Yamashita,‡ Yoshiharu Doi, and Hideki Abe*,†,‡ Department of Innovative and Engineered Materials, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226-8502, Japan, and Chemical Analysis Team, RIKEN Institute, 2-1 Hirosawa, Wako-shi, Saitama 351-0198, Japan Received November 15, 2007; Revised Manuscript Received January 8, 2008

Poly(L-lactide) (PLLA) with various alkyl ester chain end groups were synthesized by ring-opening polymerization of L-lactide in the presence of zinc alkoxide as a catalyst. The structural effect of chain end groups on the rate of enzymatic and nonenzymatic degradations for amorphous films of PLLA were investigated at 37 °C in a TrisHCl buffer solution (pH 8.6) with proteinase K and at 60 °C in a phosphate buffer solution (pH 7.4), respectively. The rate of enzymatic degradation for PLLA films was dependent on the carbon numbers of alkyl ester chain end groups, and the rates of PLLA samples with dodecyl (C12), tridecyl (C13), and tetracocyl (C14) ester end groups were much lower than those of the other samples. The surface morphologies of PLLA films after enzymatic degradation were characterized by scanning electron microscopy. After the enzymatic degradation, non-endcapped PLLA, PLLA with methyl (C1) and hexyl (C6) ester chain ends, were degraded homogeneously by proteinase K and the film surface was very smooth. In contrast, the PLLA with alkyl ester chain ends of carbon numbers over 12 were degraded heterogeneously by the enzyme, and the sponge-like network structure was formed on the film surface. These results indicated that the long alkyl ester groups at the chain ends of PLLA molecules aggregated in the amorphous films and the erosion rate was depressed due to the coverage of the aggregated terminal groups on the film surface. For the nonenzymatic degradation, the molecular weight of non-end-capped PLLA was remarkably decreased with progress of degradation. In contrast, the molecular weight of the endcapped PLLA gradually reduced at the initial stage of degradation and then the rate of degradation was accelerated. The decreases of molecular weight of PLLA by autocatalyzed degradation were retarded by the capping of carboxyl chain ends.

Introduction Poly(L-lactide) (PLLA) is presently one of the most attractive materials because of the most promising candidates as a replacement of petrochemical products. This aliphatic polyester is synthesized by either the direct polymerization of L-lactic acid by a fermentation method or by a ring-opening polymerization of cyclic dimer (L-lactide) in the presence of a catalyst such as stannous octoate or zinc metal.1–6 The raw materials such as lactic acid or lactide of PLLA are produced from renewable carbon sources such as sugar or corn. Thus, PLLA is an environmentally friendly material. However, PLLA is a hydrolyzable polymeric material because of having ester bonds. Therefore, the environmental or commodity applications require fundamental information about degradability such as enzymatic or nonenzymatic degradation. The nonenzymatic and enzymatic degradation of PLLA have been numerously investigated during the past two decades. It was reported that the autocatalyzed hydrolysis of the PLLA film in the absence of enzymes proceeds by a bulk-degradation mechanism and endo-chain scission.7–12 For nonenzymatic degradation, cleavage of the ester linkages by the terminal carboxyl groups results in a successive reduction in molecular weight of PLLA. * To whom correspondence should be addressed. E-mail: [email protected]. Telephone: +81-48-467-8000. Fax: +81-48-462-4631. † Department of Innovative and Engineered Materials, Tokyo Institute of Technology. ‡ Chemical Analysis Team, RIKEN Institute.

Proteinase K, a well-known PLLA-degrading enzyme, is a fungal serine protease of Tritiratium album and can hydrolyze the high molecular weight PLLA.13–15 It is reported that proteinase K is an endo- and exo-type enzyme.11 It has been established that the enzymatic degradation rate of PLLA are dependent on molecular weight,16 crystallinity,17–19 and crystal size.19 Moreover, polymer blending16,20–24 and copolymerization25–28 techniques can regulate the physical properties such as thermal properties and hydrolyzability of PLLA. However, there are few reports about techniques satisfied with both preserving the thermal properties and changing the degradability. Enzymatic degradation reaction progresses on the surface of materials because of the heterogeneous reaction between the water-insoluble polymer chains and water-soluble enzyme molecules. Therefore, it can be predicted that changing the surface properties results in the change of enzymatic degradability. The surface properties could be changed by an endcapping technique with various functional groups29–32 because chain end groups of materials are segregated on the surface of materials owing to free energy.33,34 In a previous study,32 we synthesized the end-capped PLLA samples with dodecyl ester groups at an R-carboxyl chain end and dodecanoyl groups at an ω-hydroxyl chain end and evaluated the degradation rates of their melt-crystallized films in the presence of proteinase K at 37 °C. End-capping with dodecyl and dodecanoyl groups could regulate the surface properties and depress the enzymatic degradation rates of crystallized PLLA films. In this article, we report the synthesis of PLLA samples endcapped at the R-carboxyl and ω-hydroxyl chain ends with the

10.1021/bm701259r CCC: $40.75  2008 American Chemical Society Published on Web 02/15/2008

1072 Biomacromolecules, Vol. 9, No. 3, 2008 Scheme 1. (a) PLLA capped at R-carbonyl chain ends (Cx-PLLA); (b) PLLA capped at both R-carbonyl and ω-hydroxyl chain ends (Cx-PLLA-Cx′)

various terminal groups and the evaluation of the solid-state structure of end-capped PLLA. The enzymatic and nonenzymatic degradation tests of amorphous PLLA films having a variety of length on alkyl groups were carried out in a TrisHCl buffer solution containing proteinase K and a phosphate buffer solution, respectively. The structural effect of chain ends on enzymatic and nonenzymatic degradation was investigated.

Experimental Section Materials. L-Lactide (L-LA) was purchased from Purac Co. and recrystallized from toluene solution under isothermal condition at 65 °C. Methanol (special grade abbreviated as SP, >99.8%) (Kanto Chemical Co.), hexanol (SP, >98.0%), undecanol (SP, >98.0%), dodecanol (SP, >99.0%), tridecanol (first class grade abbreviated as FC, >95.0%), tetradecanol (SP, >98.0%), octadecanol (SP, >99.0%), tetracosanol (FC, >97.0%) (Tokyo Chemical Industry Co.), hexanoic anhydride (FC, >95.0%), and dodecanoic anhydride (FC, >95.0%) (Kanto Chemical Co.) were used as received. All other materials were used without further purification. For the enzymatic degradation test, proteinase K (lyophilized powder, g30 units/mg) was purchased from Sigma and used as received. Preparation of Zinc-Based Catalyst. Diethylzinc/water (ZnEt2/ H2O) (1/0.6) catalyst was prepared by a reported method.35 ZnEt2 was allowed to react with deoxygenated water at a molar ratio of 1/0.6 (ZnEt2/H2O) in dry 1,4-dioxane, followed by freeze-drying of the reaction mixture. The obtained ZnEt2/H2O was used as a catalyst for non-end-capped PLLA. Zinc dialkoxide catalysts were prepared by a reaction of ZnEt2 with each alcohol (methanol, hexanol, undecanol, dodecanol, tridecanol, tetradecanol, octadecanol, and tetracosanol) at a molar ratio of 1/2 (ZnEt2/alcohol) in organic solvent at room temperature, similar to the reported method.31,32 As a solvent, dichloromethane was used for the preparation of catalysts from alcohols with carbon numbers of 1–14, while tetrahydrofuran was applied for alcohols with carbon numbers of 18 and 24. Synthesis of Functionalized PLLA Samples. Non-end-capped PLLA samples were synthesized by the ring-opening polymerization of L-LA in the presence of ZnEt2/H2O catalyst.31,32 The monomer, ZnEt2/H2O catalyst, and dichloromethane reacted under nitrogen atmosphere at 60 °C. The produced polyester was dissolved in chloroform and precipitated by hexane. The purified polymer was dried in vacuo at room temperature and was obtained as white flocculent polymer. PLLA samples end-capped with methyl (C1), hexyl (C6), undecyl (C11), dodecyl (C12), tridecyl (C13), tetradecyl (C14), octadecyl (C18), and tetracosyl (C24) ester at the R-carboxylic acid terminus (Scheme 1) were synthesized by ring-opening polymerization of L-LA in the presence of zinc alkoxides as a catalyst. The monomer and catalyst reacted under nitrogen atmosphere at 40 °C. Polymerizations of L-LA were performed in dichloromethane or in tetrahydrofuran. To remove the residual catalysts in the non-end-capped and endcapped PLLA samples, chloroform solutions (1.0% w/v) of the polymers were washed with 5% (v/v) acetic acid for 2 days at room temperature. Then the polymers were precipitated in methanol. The obtained C6-PLLA samples were then acylated at the ω-hydroxyl chain end with hexanoic anhydride. The chloroform solution of the PLLA sample (5% w/v) was mixed with 10% (v/v) acid anhydride in

Kurokawa et al. 20% (v/v) pyridine of chloroform solution and stirred for 2 days at 40 °C. PLLA samples end-capped with dodecanol were acylated at the ω-hydroxyl chain end with dodecanoic anhydride by a reported method.32 The chloroform solution of C12-PLLA (5% w/v) was mixed to react with dodecanoic anhydride at a weight ratio of 1/2 (PLLA/ acidic anhydride) in the presence of 20% (v/v) pyridine of chloroform solution at 40 °C for 2 days. The obtained polyesters were precipitated by methanol from chloroform solution. Preparation of Melt-Quenched Amorphous Film of PLLA. PLLA films of 0.1 mm in thickness were prepared as follows: The PLLA samples and 0.1 mm Teflon spacer were first sandwiched between two Teflon sheets. Then, the polymer sample was pressed at 200 °C and then immersed immediately into liquid nitrogen to preserve the amorphous state. Preparation of PLLA/Alcohol Blend Films. Solvent-cast films of binary blends of PLLA with alcohol (hexanol, dodecanol, and tetracosanol) were prepared by dissolving the polymeric binary mixtures in chloroform and casting on glass plates at room temperature. Then the blend films were melt-quenched as mentioned above. Analytical Procedures. The molecular structures of end-capped PLLA samples were determined by 1H NMR analysis with JEOL R-300 and R-400 spectrometers at room temperature using CDCl3 at a concentration of 5 mg/mL. The molecular weight of the polymers was determined by gel permeation chromatography measurements at 40 °C using a Shimadzu 10A system and 10A refractive index detector with Shodex K-806 M and K-802 columns. Chloroform was used as the mobile phase at a flow rate of 0.8 mL/min, and a sample concentration of 1.0 mg/mL was applied. The system was calibrated by using polystyrene standards with a low polydispersity. Thermal analysis of PLLA samples and PLLA/alcohol blends were conducted by differential scanning calorimetry (DSC). Samples of ca. 3 mg were weighed into aluminum pans and were analyzed using a Perkin-Elmer Pyris 1 equipped with a cooling accessory under a nitrogen flow of 20 mL/min. The samples were heated from 0 to 200 °C at a rate of 20 °C/min. The melting temperature (Tm) was determined as the peak temperature of the DSC endotherms. The glass transition temperature (Tg) of blend films was determined as the midpoint of the change in heat capacity. For the measurement of the Tg of PLLA samples, the samples maintained at 200 °C were rapidly quenched to 0 °C and then heated to 200 °C at a heating rate of 20 °C/min. The Tg was taken as the midpoint of the change in heat capacity. The wide-angle X-ray diffraction patterns of PLLA samples were recorded at 23 °C on a Rigaku RINT2500 system using nickel-filtered Cu KR radiation operated at 40 kV and 200 mA. The scan was carried out in the 2θ range of 6–60 ° at a scan speed of 2.0°/min. For cross-section observation of PLLA films, the samples were immersed in liquid nitrogen for a few minutes before cutting to prevent deformation of the cross-section. After coating the samples with Au, imaging was performed with a JSM-6330F scanning electron microscope (JEOL) operated at an acceleration voltage of 5 kV at room temperature. Degradation Tests. Square specimens with dimensions of 10 mm × 10 mm × 0.1 mm that weighed about 12 mg were cut from the melt-quenched amorphous films. For the enzymatic degradation experiment, each specimen was placed in vials with 5 mL of Tris-HCl buffer (pH 8.6) containing 1.0 mg of proteinase K. The enzymatic hydrolysis was carried out at 37 °C. For nonenzymatic hydrolysis studies, phosphate buffer (pH 7.4) was used. The vials were placed in an oven at 60 °C. At a given degradation time, three replicate specimens were taken out and washed with distilled water. These specimens were dried in vacuo at room temperature and then weighed.

Results Synthesis of End-Capped PLLA Samples. End-capped PLLA samples with alkyl ester groups at the R-carboxyl chain end were synthesized by a ring-opening polymerization of L-LA

Degradations of End-Capped Poly(L-lactide)

Biomacromolecules, Vol. 9, No. 3, 2008 1073

Table 1. Polymerization Condition, Molecular Weight, and Thermal Properties of End-Capped PLLA samplea PLLA C1-PLLA C6-PLLA C6-PLLA-C6′ C11-PLLA C12-PLLA C12-PLLA-C12′ C13-PLLA C14-PLLA C18-PLLA C24-PLLA

[L-LA]0/2[Zn] reaction polymer (mol/mol) time (h) yield (%) 250 300 150 250 200 200 300 300

48 50 22 48 20 13 48 20 20 22 22

78 30 73 99 82 96 99 97 88

DS0b (%) 91 88 100(-OH) 78 77 100(-OH) 91 94 86 71

Mn(GPC)c Mw/Mn(GPC)d Mn(calc)e Mn(NMR)f Tgg (°C) Tmh (°C) ∆Hmi (J/g) 41500 57100 58400 55400 53400 52700 56000 68400 56600 73000 59200

1.65 1.26 1.15 1.16 1.65 1.26 1.25 1.61 1.33 1.28 1.29

10800 31800 21400 29500 29700 28500 41900 38000

16800 24500 32300 33800 25200 23300 24400 27600 31600 36300 31100

59 61 60 60 60 58 58 61 61 58 59

171 178 177 177 178 173 175 177 177 174 175

46 54 51 54 46 54 53 48 51 49 48

a PLLA was synthesized by polymerization of L-LA with diethylzinc/water as a catalyst. Polymerization conditions; reaction temperature at 60 °C, in CH2Cl2. Cx-PLLA were synthesized by polymerization of L-LA with zinc alkoxides as a catalyst. The x is carbon number of alcohol used to prepare the zinc alkoxides. Polymerization condition; reaction temperature at 40 °C, in CH2Cl2 (carbon numbers of 1–14) or tetrahydrofuran (carbon numbers of 18 and 24). Cx-PLLA-Cx′ were synthesized by acylation of Cx-PLLA with acid anhydride. The x′ is a half of carbon number of acid anhydride. Polymerization condition; reaction temperature at 40 °C, in CHCl3. b Degree of substitution with terminal groups at the R-carboxyl chain end was estimated from 1H NMR spectra. c The number-average molecular weight was determined by GPC analysis using a polystyrene standards. d The polydispersity was determined by GPC analysis. e The number-average molecular weight was calculated from [L-LA]0/[Zn] ratio and polymer yield. f The number-average molecular weight was determined from 1H NMR spectra. g Glass-transition temperature was measured by DSC (second run) from 0 to 200 °C at a rate of 20 °C/min. h Melting temperature was measured by DSC (first run) from 0 to 200 °C at a rate of 20 °C/min. i Enthalpy of fusion was measured by DSC (first run) from 0 to 200 °C at a rate of 20 °C/min.

in the presence of zinc alkoxide as a catalyst. Non-end-capped PLLA was synthesized through similar procedures by using ZnEt2/H2O as a catalyst. The obtained PLLA samples were reacted with acid anhydride to prepare the samples with alkanoyl groups at ω-chain ends. Table 1 lists the polymer yields, number-average molecular weights (Mn), and polydispersity (Mw/Mn) of the end-capped PLLA samples. All of the endcapped PLLA samples with alkyl ester groups were obtained in satisfactory yield (73–99%), except for PLLA with a methyl ester chain end (30%). The Mn(GPC) and Mw/Mn of end-capped PLLA samples determined from GPC ranged from 51000 to 73000 and from 1.15 to 1.65, respectively. According to previous studies, the chain end structure of the obtained PLLA samples was characterized by 1H NMR analysis. The methine proton of the ω-hydroxyl-terminus L-lactyl unit was detected at 4.35 ppm (Figure 1c) for non-end-capped PLLA and PLLA with the R-alkyl ester chain end, in addition to the methyl (1.59 ppm (Figure 1a)) and methine (5.17 ppm (Figure 1b)) protons of L-lactyl unit in main-chain (Figure 1). The peaks from the methyl proton of the R-methyl ester chain end and methylene proton of other R-alkyl esters connecting with the L-lactyl unit were respectively detected at 3.74 and 4.13 ppm. The degree of substitution (DS0) at the R-chain end was determined qualitatively by the peak intensities of the methine proton at the ω-chain end and the methyl or methylene proton of the alkyl ester groups. The DS0 values of PLLA samples were ranged from 0.71 to 0.94, and the majority of the R-chain end of PLLA molecules was substituted with the alkyl ester groups (see Table 1). In the 1H NMR spectra of PLLA samples with ω-alkanoyl chain ends, the methine proton signal at 4.35 ppm arising from the ω-terminus L-lactyl unit was completely disappeared, and the signal from the methylene proton of the alkanoyl group appeared at 2.35 ppm. Assuming that the polymerization reaction took place at all points of zinc-alkoxide linkage in the catalyst, the molecular weight (Mn(calc)) of the obtained samples was calculated from the monomer/catalyst ([L-LA]0/[Zn]) ratio and the polymer yield. Also, the number-average molecular weight (Mn(NMR)) of end-capped PLLA samples were determined qualitatively from the peak intensities of the methine proton at the hydroxyl terminus (peak c in Figure 1) and the main chain of methine proton (peak b in Figure 1) in the 1H NMR spectra. As shown in Table 1, the Mn(NMR) values were consistent with the Mn(calc)

Figure 1. 1H NMR spectrum of typical end-capped PLLA sample (A) and enlarged spectra of chain end groups for C6-PLLA and C6-PLLAC6′ (B).

values, except for PLLA with methyl ester chain end (C1PLLA). These results indicate that all of Zn-alkoxide groups act as active sites of polymerization of L-LA and that chain transfer reactions rarely occurred during the polymerization reaction. In the case of C1-PLLA, the Mn(calc) value was less than half of the Mn(NMR) value. As shown in Table 1, the polymer yield of C1-PLLA was considerably low (30%). These results suggest that the active site of zinc dimethoxide catalyst was less than half of the zinc-methoxide group. The Mn(NMR) values of end-capped PLLA samples were lower than Mn(GPC) values. It has been reported that the Mn(GPC) values determined from GPC by using the calibration of polystyrene standards represent larger values than the absolute ones due to the difference in exclusion volume of PLLA and polystylene with the same molar mass.36 In this study, the ratios of Mn(NMR) and Mn(GPC) ranged from 0.40 to 0.60, and these values relatively coincided with the previously reported values.31 The averaged ratio was estimated to be 0.48, and the value was applied as

1074 Biomacromolecules, Vol. 9, No. 3, 2008

Kurokawa et al.

Figure 2. Weight loss of amorphous film of PLLA capped with R-alkyl ester groups during the enzymatic degradation by proteinase K (200 µg/mL): O PLLA; [ C6-PLLA; 9 C12-PLLA; b C18-PLLA.

Figure 3. Weight loss of PLLA films with different R-alkyl ester groups after enzymatic degradation for 24 h.

Figure 4. Scanning electron micrographs of the surfaces (A,C,E,G,I K) and cross sections (B,D,F,H,J,L) of end-capped PLLA films after enzymatic degradation for 24 h at 37 °C by proteinase K: (A,B) PLLA; (C,D) C1-PLLA; (E,F) C6-PLLA; (G,H) C12-PLLA; (I,J) C18-PLLA; (K,L) C24-PLLA films.

the pertinent correction factor to convert the Mn(GPC) values to the actual values in this study. Thermal properties of end-capped PLLA samples were characterized by DSC. Table 1 also lists the melting temperature (Tm), enthalpy of fusion (∆Hm), and glass transition temperature (Tg) of the obtained samples. The Tm, ∆Hm, and Tg values of samples respectively ranged from 173 to 178 °C, 46 to 54 J/g, and 58 to 61 °C, and each value was almost constant among the end-capped PLLA samples regardless of the chain end structure. Enzymatic Degradation Test. Enzymatic degradation test of end-capped PLLA samples were carried out in an aqueous solution containing proteinase K at 37 °C. For the degradation test, melt-quenched amorphous films of PLLA samples were applied. Figure 2 shows the enzymatic erosion profiles of typical end-capped PLLA samples with hexyl (C6), dodecyl (C12), and octadecyl (C18) ester chain ends, together with non-end-capped PLLA. Except for C12-PLLA, the erosion profiles of end-capped PLLA films revealed similar trends to non-end-capped PLLA film. The weight loss of the film increased nearly proportionally with time over 30 h, and the film was almost eroded after degradation for 72 h by the action of enzyme. At the later degradation stage, the weight loss of C18-PLLA was slightly smaller than the non-end-capped PLLA. For the C12-PLLA sample, the similar enzymatic erosion was detected at the initial stage of reaction for 5 h, after which the weight loss gradually increased with time. Figure 3 summarizes the weight loss values of all end-capped PLLA films after degradation for 24 h. The weight losses of the PLLA samples with alkyl ester chain ends with carbon numbers of 12–14 were apparently lower than the other samples, and the values were around a quarter of non-end-capped PLLA. The PLLA samples with shorter alkyl ester groups (C1, C6, and C11) and longer ones (C18 and C24) revealed almost identical weight loss values. Before enzymatic degradation, all of the PLLA films showed transparency regardless of the species of alkyl ester chain ends.

It is of interest to note that the PLLA samples with alkyl ester groups with carbon numbers over 11 became opaque and whitish films after degradation. In contrast, the films of PLLA samples with shorter alkyl groups (C1 and C6) held their transparent appearance even though the degradation proceeded. On the basis of the X-ray diffraction analysis of PLLA films after enzymatic degradation, it was confirmed that all of the films maintained an amorphous state. The film surface and cross section of PLLA samples before and after enzymatic degradation were characterized by SEM. Before degradation, the surface of each PLLA sample was almost flat. Figure 4 shows the typical SEM images of the surface and the cross section of the PLLA films after enzymatic degradation for 24 h. The films of non-end-capped PLLA and PLLA with short alkyl groups (C1 and C6) (Figure 4C,E) maintained a smooth surface, while the film thickness was apparently decreased (Figure 4D,F). In contrast, the surface of the C12-PLLA film was blemished after degradation for 24 h (Figure 4G). For the PLLA samples with longer alkyl ester groups (C18 and C24) (Figure 4I,K), the sponge-like network structure was formed on the surface of films in the depth direction of 5–10 µm (Figure 4J,L). The film thickness was almost unchanged. These results suggest that the enzymatic degradation progressed by a different manner between the PLLA samples with shorter alkyl ester groups and longer ones even though the erosion profiles revealed almost same trends. As shown in parts A and B of Figure 5, after enzymatic degradation for 48 h, the C6-PLLA film became further thinner, while the film surface was still smooth. Because only slight erosion took place after the initial stage of degradation, the surface of the C12-PLLA film after degradation at 48 h was almost identical with the film surface at 24 h. On the other hand, the enzymatic degradation of C18-PLLA films continuously proceeded, and the sponge-like network structure was formed in the remained film throughout after enzymatic degradation at 48 h (Figure 5E,F).

Degradations of End-Capped Poly(L-lactide)

Biomacromolecules, Vol. 9, No. 3, 2008 1075

Figure 5. SEMs of the surfaces (A,C,E) and cross sections (B,D,F) of end-capped PLLA films after enzymatic degradation for 48 h at 37 °C by proteinase K: (A,B) C6-PLLA; (C,D) C12-PLLA; (E,F) C18-PLLA films.

Figure 6. Weight loss of PLLA capped with alkyl groups at both R,ωchain ends during the enzymatic degradation by proteinase K: [ C6PLLA; ] C6-PLLA-C6′; 9 C12-PLLA; 0 C12-PLLA-C12′.

Figure 8. Weight loss (A), molecular weight (B), and polydispersity (C) changes of end-capped PLLA film during hydrolytic degradation in the absence of enzyme: O PLLA; 9 C12-PLLA; 0 C12-PLLA-C12′. Solid and dotted lines in (B) represent the calculated and simulated profiles by using the obtained kinetic parameters, respectively.

Figure 7. SEMs of the surfaces (A,C) and cross sections (B,D) of end-capped PLLA films after enzymatic degradation for 24 h at 37 °C by proteinase K: (A,B) C6-PLLA-C6′; (C,D) C12-PLLA-C12′.

A degradation test of PLLA samples capped with terminal groups at R,ω-chain ends was performed. As shown in Figure 6, the erosion profile for the C6-PLLA-C6′ sample was almost identical with that of C6-PLLA. The weight loss of the C12PLLA-C12′ sample was larger than that of C12-PLLA. Although the substitution of ω-hydroxyl groups with dodecanoyl groups for the C12-PLLA sample induced an increase of enzymatic erosion rate, the weight loss of film was still lower than that of non-end-capped PLLA. Figure 7 shows the SEM images of the surface and cross section of PLLA films capped at R,ω-chain ends. Similar to the other end-capped samples, the surface of films were flat before degradation. After degradation, the film of the C6-PLLAC6′ sample held a smooth surface. On the other hand, the surface of the C12-PLLA-C12′ sample became more rough compared with the C12-PLLA sample, and the formation of a sponge-like network structure the same as the C18-PLLA sample was confirmed. Nonenzymatic Hydrolysis. Nonenzymatic degradation test of PLLA samples was carried out in a phosphate buffer (pH

7.4) at 60 °C. Figure 8 shows the weight loss of films, numberaverage molecular weight (Mn), and polydispersity (Mw/Mn) of PLLA molecules during hydrolytic degradation. Slight loss of film weight (around 2–4 wt % of the initial film) was detected for non-end-capped PLLA during the initial 17 days, then the values continuously increased to reach 0.77 mg/cm2 (13 wt % of the initial film) at 44 days. Although the weight loss of the films was relatively small, the Mn value of non-end-capped PLLA sample drastically reduced from the initial stage of degradation. The Mn values of non-end-capped PLLA decreased nonlinearly from 19900 to 2800 during the degradation for 44 days. The Mw/Mn values increased from 1.61 to 2.24 at the initial stage of degradation, and then the values decreased to reach 1.13 at 44 days. For the C12-PLLA sample, weight loss of films was hardly detected during the initial 17 days. The Mn values gradually reduced with time during the initial 23 days, and the rates of reduction in Mn values were apparently lower than that of the non-end-capped sample. After the initial 23 days, the pronounced decrease in Mn values was observed. Together with the decrease in Mn, the Mw/Mn of end-capped PLLA samples increased.

Discussion In this study, hydrolytic degradation behaviors of PLLA samples end-capped with various alkyl groups under the presence and absence of an enzyme.

1076 Biomacromolecules, Vol. 9, No. 3, 2008

From the enzymatic degradation experiments, it was found that the rate of erosion was strongly dependent on the length of alkyl ester chain ends and that the degradation manners could be roughly divided into three types. The enzymatic degradation of PLLA films with shorter alkyl ester groups proceeded homogeneously from the surface of film, and the weight loss of the films increased proportionally with time. The film thickness decreased as degradation proceeded with keeping the surface smooth. The introduction of an alkyl ester chain ends with carbon numbers of 12–14 induced suppression of the degradation of PLLA film by proteinase K. The remarkable erosion took place only at the initial stage of the enzymatic reaction, followed by very small weight loss. The erosion profiles of PLLA films with longer alkyl ester groups were almost identical with non-end-capped PLLA films. However, the surface of PLLA films with longer alkyl ester groups was blemished, and the sponge-like network structure was formed on the film surface, indicating that the enzymatic degradation reaction of PLLA samples with longer alkyl ester groups occurs heterogeneously from the surface to inside of the film. The degradation profiles similar to the C12-PLLA sample have been found for the enzymatic degradations of polymer blends containing enzymatically active and inactive components. Gajria et al. performed enzymatic degradation of PLLA and poly(vinyl acetate) (PVAc) blends by using proteinase K.22 The PLLA molecules on the film surface of PLLA/PVAc miscible blends were degraded by the enzyme at the initial stage. After the initial degradation period, enzymatically inactive PVAc molecules covered on the films surface and the enzymatic erosion was inhibited.22 Kumagai and Doi and Abe et al. also proposed the same mechanism for the enzymatic degradations of the blends of bacterial poly(3-hydroxybutyric acid)(P(3HB))/PVAc37 and of the chemosynthesized racemic P(3HB) with high isotacticity by the P(3HB)-degrading enzyme.38 Proteinase K, used in this study, is a family of serine proteases, and the enzyme is capable of hydrolyzing ester bonds in the PLLA molecules randomly. In the cases of PLLA samples with alkyl ester chain ends with carbon numbers of 12–14, the PLLA molecules on the film surface are degraded by the enzyme. On the other hand, the alkyl chain end groups are difficult to be hydrolyzed by proteinase K and stay on the film surface. As a result, during the enzymatic reaction at the initial stage, the alkyl end groups may be concentrated on the film surface and eventually covered. With an increase of alkyl chain end concentration, the enzymatic erosion on the film surface is retarded at the later stage. It is expected that alkyl groups with carbon numbers below 11 are also enzymatically inactive components. However, the suppression of enzymatic erosion was hardly detected for PLLA with such shorter alkyl ester groups. Sheth et al. reported the enzymatic degradation study of PLLA and poly(ethylene oxide) (PEO) blends with proteinase K.39 Because PEO is the watersoluble component, the enzymatic erosion rates of the blends were accelerated by the addition of PEO. Similar degradation profile was also observed for P(3HB) and PEO blends in the presence of P(3HB)-depolymerases.40 Alkyl groups with carbon numbers below 11 may be easily released from the film surface to reaction solution during the hydrolysis of PLLA molecules due to their well solubility or ability of dispersion as a liquid state in aqueous solution. As a result, the PLLA samples with shorter alkyl ester groups revealed almost the same degradation profile with non-end-capped PLLA. In the case of PLLA samples with alkyl ester chain ends with carbon numbers over 18, the alkyl chain end groups may

Kurokawa et al.

assemble each other owing to the hydrophobic interaction and the regions concentrating chain end groups formed heterogeneously in the film. Such regions concentrated the hydrophobic component are difficult to be hydrolyzed by proteinase K, while the enzyme easily hydrolyzes the amorphous regions of pure PLLA molecules. Therefore, the film surface of PLLA with longer alkyl ester groups was partially hydrolyzed, and the remained regions concentrated with chain end groups appeared as sponge-like network structure. As a result, at the later degradation stage, the weight loss of the C18-PLLA film was smaller than that of the non-end-capped PLLA, as shown in Figure 2. A similar heterogeneous degradation profile has been observed for the enzymatic degradation of immiscible PLLA/ poly(-caprolactone) (PCL) blends by proteinase K.20 The immiscible blends formed a phase separation structure and the enzyme hydrolyzed the PLLA phase but not the PCL phase. As a result of partial degradation of the PLLA phase, many holes were formed on the surface and inside of the blend films. As shown in Table 1, the thermal properties of end-capped PLLA samples showed the almost constant values independent of the structure of chain end groups. Taking into account the weight composition of chain end groups in PLLA components, it is expected that significant differences in thermal properties among the end-capped PLLA samples are hardly detected by calorimetric analysis. To confirm the possibility of phase separation between PLLA molecules and alkyl chain end groups, thermal properties of PLLA with an addition of adequate amounts of alkyl alcohols were characterized by DSC. The PLLA with n-hexanol showed a single glass-transition temperature (Tg), and the Tg value decreased from 55 to 33 °C with an increase in n-hexanol content from 0 to 10 wt %. This result indicates that the n-hexanol molecule is miscible with PLLA. The Tg value of PLLA also decreased with addition of n-dodecanol while not proportional with n-dodecanol content, suggesting the partial dispersion of the n-dodecanol component in PLLA phase. In contrast, the Tg value of PLLA with an addition of n-tetracosanol was almost identical with pure PLLA independent of the n-tetracosanol content. Therefore, the ntetracosanol is immiscible with PLLA, and it is suggested that the PLLA and n-tetracosanol components form a phase separation structure. Thus, the compatibility of PLLA with alkyl alcohols is dependent on the carbon numbers of alcohol. It is predicted that a similar relation holds also for the PLLA molecules and alkyl ester chain ends. Owing to such a difference in the dispersing state of alkyl ester chain ends in the PLLA phase, the degradation profiles differed between the PLLA samples with alkyl ester chain ends with carbon numbers of 12–14 and with carbon numbers over 18. As shown in Figure 7, on the surface of the C12-PLLA-C12′ film after enzymatic degradation, a sponge-like network structure was formed. From the results of DSC for PLLA with ndodecanol, it is suggested that the alkyl group with carbon number of 12 is partially miscible with PLLA. Therefore, it is predicted that a part of the terminated groups with a carbon number of 12 was segregated from PLLA matrix. By introduction of a dodecanoyl group at the ω-chain end, the alkyl groups at both R,ω-chain ends may assemble more easily. The heterogeneous phase structure composed of the pure PLLA region and chain end concentrated region is formed in the C12PLLA-C12′. As a result of phase separation, the enzymatic erosion rate of C12-PLLA-C12′ was increased compared with that of C12-PLLA.

Degradations of End-Capped Poly(L-lactide)

Biomacromolecules, Vol. 9, No. 3, 2008 1077

In the nonenzymatic hydrolysis test, it was confirmed that the decrease in molecular weight was apparently different between non-end-capped and end-capped PLLA samples. It has been reported that the hydrolytic degradation of PLLA molecules involves both a noncatalyzed reaction and an autocatalyzed reaction in which the R-carboxylic acid chain end participates, while that of the autocatalyzed reaction dominantly proceeds in the absence of enzyme.10 In the autocatalyzed hydrolysis reaction, the free carboxylic acid of the R-chain end unit acts as an acid catalyst to promote hydrolysis of the ester bond in the PLLA main chain. Protection at the R-carboxylic acid chain end by alkyl ester groups resulted in retardation of the autocatalyzed hydrolysis reaction. As a result, the decrease in molecular weight of end-capped PLLA samples was depressed at the initial stage of degradation. However, the noncatalyzed hydrolysis reaction gradually took place to generate R-carboxylic acid chain ends, resulting in acceleration of the hydrolysis reaction with degradation time. Here, the kinetic analysis was carried out to estimate the rate constants of nonenzymatic degradation reaction of PLLA molecules. Previously, Cha and Pitt have estimated the hydrolysis rate constant of PLLA molecules from the changes in molecular weight during nonenzymatic degradation in phosphate buffer (pH 7.4) at 37 °C.10 They regarded that the autocatalyzed hydrolysis was a dominant reaction, and the rate constant kd of hydrolysis was determined by the following equation:

1nMn(t) ) 1nMn(0) - kdt

Conclusion

(2)

where Pn(t) is the number-average degree of polymerization at time t and Pn(0) is the initial number-average degree of polymerization. The deviation of 1/Pn(t) at t by a noncatalyzed reaction is given by:

d[1 ⁄ Pn(t)] ⁄ dt ) kd

(2a)

On the other hand, from the eq 1, the deviation of 1/Pn(t) at t by autocatalyzed degradation is given as:

d[1 ⁄ Pn(t)] ⁄ dt ) kd[1 ⁄ Pn(t)]

(1a)

Provided that a combination of noncatalyzed and autocatalyzed reactions occurs and that the chain scission is completely random by each hydrolysis reaction, the deviation of 1/Pn(t) at time t is given as a sum of two deviations given by the eqs 1a and 2a:

d[1 ⁄ Pn(t)] ⁄ dt ) kd1 + kd2[1 ⁄ Pn(t)]

where Pn(0) is the initial number-average degree of polymerization and DS0 is the degree of substitution at R-chain end before hydrolytic degradation. By using these expressions, the rates of noncatalyzed and autocatalyzed reactions were determined from the molecular weight data of nonenzymatic degradation experiments. Because the equations above should be valid for the period without significant weight loss of samples, the weighted regression against the data sets at no significant weight loss was applied for the evaluation of rate constants. The calculated kd1 and kd2 values were 3.5 ( 0.5 × 10-6 day-1 and 4.1 ( 0.6 × 10-2 day-1, respectively. Although the obtained kd2 value was one order larger than the value (6.7 × 10-3 day-1) reported by Cha and Pitt, the value seemed to reasonably take into account the difference of reaction temperature. As the Pn(0) value of non-end-capped PLLA sample was around 270, the rate of autocatalyzed hydrolysis reaction was about 40 times that of the noncatalyzed reaction at the initial stage of degradation. For the C12-PLLA sample, the Pn(0) and DS0 values were, respectively, 350 and 0.77. Therefore, the autocatalyzed hydrolysis reaction still occurred at 10-fold frequency of noncatalyzed reaction at the initial stage. However, it could be concerned that the overall degradation rate at the initial stage was reduced to one-quarter by the partial capping of carboxyl chain ends.

(1)

where Mn(t) is the number-average molecular weight of PLLA at time t. The kd value of compression-molded PLLA film was calculated to be 6.7 × 10-3 day-1 at 37 °C.10 However, as mentioned above, the nonenzymatic degradation of PLLA molecules occurs by both the noncatalyzed and autocatalyzed reaction competitively. Therefore, both the rate constants of noncatalyzed and autocatalyzed hydrolysis reactions were evaluated from our experimental data. When the hydrolytic degradation of PLLA molecules progresses completely via noncatalyzed reaction, the rate constant kd of hydrolysis is given by the following equation:41,42

1 ⁄ Pn(t) ) 1 ⁄ Pn(0) + kdt

d[1 ⁄ Pn(t)] ⁄ dt ) kd1 + kd2[1 ⁄ Pn(t) - DS0 × 1 ⁄ Pn(0)] (4)

(3)

where kd1 and kd2 are the rate constants of noncatalyzed and autocatalyzed hydrolysis reactions, respectively. If the R-carboxylic acid chain end is protected by alkyl ester groups, the deviation of Pn(t) is converted as the following equation:

The relationships between chain end structure and degradation profiles of PLLA end-capped with various terminal groups were investigated. Thermal properties of PLLA samples were hardly changed by the capping of chain ends with alkyl groups. The enzymatic degradation test of amorphous films of end-capped PLLA with proteinase K was carried out in a Tris-HCl buffer solution. The degradation rates of PLLA samples with alkyl ester chain ends with carbon numbers of 12–14 were much slower than those of the other end-capped samples. From the SEM observations of end-capped PLLA films after enzymatic degradation, it was found that the films of PLLA with alkyl end groups with carbon numbers below 11 were homogeneously degraded from the film surface by the enzyme, while that the PLLA samples with alkyl ester chain ends with carbon numbers over 12 were heterogeneously eroded. Such difference in degradation manner among the PLLA samples with different alkyl ester chain ends may be related to the dispersing state of alkyl ester chain ends in PLLA phase. The nonenzymatic degradation test of end-capped PLLA films was performed in a phosphate buffer solution at 60 °C. The hydrolytic degradation rates of end-capped PLLA films decreased compared with nonend-capped PLLA, indicating that the autocatalyzed degradation at the initial stage of degradation was inhibited by the protection of the R-carboxylic acid end with the alkyl ester group. From the kinetic analysis of nonenzymatic degradation, the rate constants of autocatalyzed and noncatalyzed hydrolysis reactions were estimated and the kd1 and kd2 values were 3.5 ( 0.5 × 10-6 day-1 and 4.1 ( 0.6 × 10-2 day-1, respectively. Acknowledgment. This work was supported by a grant for Ecomolecular Science Research provided by the RIKEN Institute.

References and Notes (1) Carothers, W. H. J. Am. Chem. Soc. 1932, 54, 761–772.

1078 Biomacromolecules, Vol. 9, No. 3, 2008 (2) Leenslag, J. W.; Gogolewski, S.; Pennings, A. J. J. Appl. Polym. Sci. 1984, 29, 2829–2842. (3) Libiszowski, J.; Kowalski, A.; Duda, A.; Penczek, S. Macromol. Chem. Phys. 2002, 203, 1694–1701. (4) Kricheldorf, H. R.; Berl, M.; Scharnagl, N. Macromolecules 1988, 21, 286–293. (5) Dubois, P.; Jacobs, C.; Jérôme, R.; Teyssié, P. Macromolecules 1991, 24, 2266–2270. (6) Eguiburu, J. L.; Fernandez-Berridi, M. J.; Cossío, F. P.; San Román, J. Macromolecules 1999, 32, 8252–8258. (7) Li, S.; Garreau, H.; Vert, M. J. Mater. Sci.: Mater. Med. 1990, 1, 123–130. (8) Li, S.; Garreau, H.; Vert, M. J. Mater. Sci.: Mater. Med. 1990, 1, 131–139. (9) Li, S.; Garreau, H.; Vert, M. J. Mater. Sci.: Mater. Med. 1990, 1, 198–206. (10) Cha, P.; Pitt, C. G. Biomaterials 1990, 11, 108–112. (11) Tsuji, H.; Ishida, T. J. Appl. Polym. Sci. 2003, 87, 1628–1633. (12) Lunt, J. Polym. Degrad. Stab. 1998, 59, 145–152. (13) Williams, D. F. Eng. Med. 1981, 10, 5–7. (14) Fukuzaki, H.; Yoshida, M.; Asano, M.; Kumakura, M. Eur. Polym. J. 1989, 25, 1019–1026. (15) Yamashita, K.; Kikkawa, Y.; Kurokawa, K.; Doi, Y. Biomacromolecules 2005, 6, 850–857. (16) Tsuji, H.; Miyauchi, S. Biomacromolecules 2001, 2, 597–604. (17) Cai, H.; Davé, V.; Gross, R. A.; McCarthy, S. P. J. Polym. Sci., Part B: Polym. Phys. 1996, 34, 2701–2708. (18) Tsuji, H.; Miyauchi, S. Polymer 2001, 42, 4463–4467. (19) Tsuji, H.; Miyauchi, S. Polym. Degrad. Stab. 2001, 71, 415–421. (20) Tusji, H.; Ishizaka, T. Macromol. Biosci. 2001, 1, 59–65. (21) Liu, L.; Li, S.; Garreau, H.; Vert, M. Biomacromolecules 2000, 1, 350–359. (22) Gajria, A. M.; Dave, V.; Gross, R. A.; McCarthy, S. P. Polymer 1996, 37, 437–444. (23) Macdonald, R. T.; McCarthy, S. P.; Gross, R. A. Macromolecules 1996, 29, 7356–7361.

Kurokawa et al. (24) Renouf-Glauser, A. C.; Rose, J.; Farrar, D.; Cameron, R. E. Biomaterials 2005, 26, 2415–2422. (25) Li, S.; Tenon, M.; Garreau, H.; Braud, M.; Vert, M. Polym. Degrad. Stab. 2000, 67, 85–90. (26) Jeon, O.; Lee, S. H.; Kim, S. H.; Lee, Y. M.; Kim, Y. H. Macromolecules 2003, 36, 5585–5592. (27) Reeve, M. S.; McCarthy, S. P.; Downey, M. J.; Gross, R. A. Macromolecules 1994, 27, 825–831. (28) Li, S.; McCarthy, S. P. Macromolecules 1999, 32, 4454–4456. (29) Lee, W. K.; Losito, I.; Gardella, J. A.; Hicks, W. L. Macromolecules 2001, 34, 3000–3006. (30) Ouchi, T.; Uchida, T.; Arimura, H.; Ohya, Y. Biomacromolecules 2003, 4, 477–480. (31) Kobori, Y.; Iwata, T.; Doi, Y.; Abe, H. Biomacromolecules 2004, 5, 530–536. (32) Kurokawa, K.; Yamashita, K.; Doi, Y.; Abe, H. Polym. Degrad. Stab. 2006, 91, 2333–2341. (33) Mayes, A. M. Macromolecules 1994, 27, 3114–3115. (34) Tanaka, K.; Jiang, X.; Nakamura, K.; Takahara, A.; Kajiyama, T.; Ishizone, T.; Hirao, A.; Nakahama, S. Macromolecules 1998, 31, 5148– 5149. (35) Teranishi, K.; Iida, M.; Araki, T.; Yamashita, S.; Tani, H. Macromolecules 1974, 7, 421–427. (36) Baran, J.; Duda, A.; Kowalski, A.; Szymanski, R.; Penczek, S. Macromol. Rapid Commun. 1997, 18, 325–427. (37) Kumagai, Y.; Doi, Y. Polym. Degrad. Stab. 1992, 36, 241–248. (38) Abe, H.; Matsubara, I.; Doi, Y. Macromolecules 1995, 29, 844–853. (39) Sheth, M.; Kumar, R. A.; Davé, V.; Gross, R. A.; McCarthy, S. P. J. Appl. Polym. Sci. 1997, 66, 1495–1505. (40) Kumagai, Y.; Doi, Y. Polym. Degrad. Stab. 1992, 35, 87–93. (41) Jellinek, H. H. G. Degradation of Vinyl Polymers; Academic Press: New York, 1955. (42) Doi, Y.; Kanesawa, Y.; Kunioka, M.; Saito, T. Macromolecules 1990, 23, 26–31.

BM701259R