Crystalline-Structure-Dependent Enzymatic Degradation of

Feb 29, 2008 - ... Kyushu Institute of Technology, 2-4 Hibikino, Wakamatu-ku, Kitakyushu-shi, Fukuoka 808-0196, Japan, and Department of Biological an...
1 downloads 0 Views 393KB Size
Biomacromolecules 2008, 9, 1221–1228

1221

Crystalline-Structure-Dependent Enzymatic Degradation of Polymorphic Poly(3-hydroxypropionate) Bo Zhu,† Yong He,† Haruo Nishida,‡ Koji Yazawa,† Nariaki Ishii,† Ken-ichi Kasuya,§ and Yoshio Inoue*,† Department of Biomolecular Engineering, Tokyo Institute of Technology, Nagatsuta 4259-B-55, Midori-ku, Yokohama 226-8501, Japan, Eco-Town Collaborative R&D Center for the Environment and Recycling, Kyushu Institute of Technology, 2-4 Hibikino, Wakamatu-ku, Kitakyushu-shi, Fukuoka 808-0196, Japan, and Department of Biological and Chemical Engineering, Gunma University, Kiryu, Gunma 376-8515, Japan Received November 6, 2007; Revised Manuscript Received January 4, 2008

The crystalline structure dependence of enzymatic degradation behavior was investigated for the polymorphic poly(3-hydroxypropionate) (P3HP), which has a basic backbone chemical structure of bacterial poly(3hydroxyalkanoate)s (P3HAs). The P3HP films consisting of the β-, γ-, and/or δ-form crystal were cast or meltcrystallized as reported previously (Macromolecules 2005, 38, 6455; Macromolecules 2006, 39, 194–203) by controlling the molecular weight, crystallization temperature, and/or temperature of the melt. Their thermal properties, crystalline structures, morphologies, and 13C solid spin–lattice relaxation dynamics were characterized by the differential scanning calorimetry, the wide-angle X-ray diffraction, the small-angle X-ray scattering (SAXS), and the 13C solid-state NMR spectra (SNMR), respectively. Both the crystallinities and the lamellar thicknesses of P3HP films were found to decrease roughly in the order of β-form > (or ≈) γ-form > δ-form. From previous work, which indicates that the P3HA enzymatic degradation depends only on the degree of crystallinity and the lamellar thickness, their enzymatic degradation rates are then expected to increase in the order of β-form < (or ≈) γ-form < δ-form. Unexpectedly, their experimental P3HP enzymatic degradation rates in the presence of P3HA depolymerase isolated from Ralstonia pickettii T1 increase in the reverse order, i.e., δ-form < γ-form < β-form. The weight loss rate of the δ-form film is almost 1 order of magnitude smaller than that of the fastest degraded β-form film. It is then strongly indicated that the crystalline structure plays a strikingly decisive role in the enzymatic degradation of P3HP. In particular, only when the conformation of crystalline chain accords with that of the bacterial poly(3-hydroxybutyrate) (P3HB) sample, i.e., the 21 helix conformation, is the P3HP sample degraded as slow as the P3HB sample. The inherent reason responsible for the unique P3HP enzymatic degradation behavior has been further clarified by comparing the molecular interaction and dynamics of polymorphic P3HP crystals.

Introduction Biodegradable poly(hydroxyalkanoate)s (PHAs), especially the bacterial ones, which have properties and processabilities comparable to those of traditional plastics, offer an attractive route to alleviate environmental concerns.1 A considerable body of knowledge on the biodegradation of PHAs by PHA depolymerases has been accumulated in the last two decades.2 The PHAs enzymatic degradability depends on the chemical structure3 and comonomer-unit composition.4 Besides the primary structure, some solid-state properties or structures, e.g., crystallinity5 and lamellar thickness6 are also important factors affecting the enzymatic degradation of PHAs. For a given PHA, the rate of enzymatic erosion is regulated by a delicate balance between the amorphous phase preferential degradation5 and the crystalline-phase specific adsorption of enzyme.7 Normally, the degradation rate would decrease with increasing the lamellar thickness.6 Moreover, the conformation and conformational flexibility of chains in the crystalline phase are expected to greatly affect * To whom all correspondence should be addressed. Tel.: +81-45-9245794. Fax: +81-45-924-5827. E-mail: [email protected]. † Tokyo Institute of Technology. ‡ Kyushu Institute of Technology. § Gunma University.

the enzymatic degradation, as the enzymatic degradation occurs through hydrolyzing ester linkages of PHA main chains. On the other hand, the chain packing also contributes to the chain mobility and is expected to affect the biodegradation. However, the effects of chain packing and conformation on the PHA biodegradation have not received the necessary attention so far. Most probably, this is due to our poor understanding on the multiformity of PHA in crystalline structure. Until now, only limited results are available on the crystal modifications of bacterial poly(3-hydroxybutyrate) (P3HB),8 poly(L-lactide),9 and poly(3-hydroxypropionate) (P3HP).10 Most of the above-mentioned polymorphic crystals of PHAs were usually prepared by the stretching and annealing procedure rather than by the spontaneous crystallization. The molecular orientation induced by the flow in molten state and the deformation in solid-state deeply influence the crystalline morphology, leading to the orientation of crystal, the production of so-called shish-kebab structure, the great enhancement of lamellar thickness and crystallinity, and sometimes distortion of the crystal unit.11 In addition, the flow in molten state would reduce the entanglement and induce the molecular orientation in the final amorphous phase. In this way, the stretching procedure introduces some additional and likely uncontrollable interfering factors to the biodegradation of PHAs. It would be

10.1021/bm701220x CCC: $40.75  2008 American Chemical Society Published on Web 02/29/2008

1222

Biomacromolecules, Vol. 9, No. 4, 2008

difficult or complex to give a quantitative description on the morphology in the stretched phase. It is therefore difficult to make a comparison on the biodegradation between the stretched phases or between the stretched and the unstretched phases. In contrast, the crystalline morphologies, which spontaneously form during casting and melt crystallization, are normally simple, similar, and controllable and thus comparable. In fact, the enzymatic degradations of the unoriented cast and meltcrystallized films have been the most intensively studied3–7 and well-clarified to be controlled by the crystallinity5 and lamellar thickness.6 With the unoriented PHA film as a model substrate, it is possible to clarify the effect of crystalline structure on the biodegradation with considering the effects of the lamellar thickness and crystallinity. Recently, three crystal modifications of biodegradable poly(3hydroxypropionate) (P3HP), i.e. the β-, γ-, and new δ-form, have been found by us to crystallize spontaneously in the meltcrystallized and/or cast film without stretching.12 The β- and the γ-forms both contain two all-trans chains but differ in the molecular arrangement from the projection to the ab-plane,10a,c,e,12b while the δ-form consists of at least two 21-helix conformed chains.12d The multiformity of P3HP in the chain conformation and the packing naturally make P3HP an appropriate model substrate to explore the relationship between the crystalline structure and biodegradability. Furthermore, due to the similarity of poly(3-hydroxyalkanoate)s (P3HAs) to P3HP in chemical structure (P3HAs all share the 3HP skeleton) and crystalline structure (all P3HAs adopt the similar 21 helix and/or trans conformation when crystallized),13,14 the work on P3HP will consequentially shed light on the crystallization and biodegradation of P3HAs and further their relationship. The previous work10d has investigated the enzymatic degradations of the stretched R-form (21 helix conformed10a,d) and β-form P3HP samples and the cast γ-form-rich P3HP film and found that the degradation of cast γ-form-rich sample is the slowest. In this work, the unoriented polymorphic P3HP samples, including the new δ-form sample, were used. The unoriented P3HP films with the β-, γ-, and δ-form crystals, were prepared by casting or melt-crystallizing with referring our previous works.12 Their crystalline structures, morphologies, crystallinity, thermal properties, and dynamics were investigated in detail by wide-angle X-ray diffraction, small-angle X-ray scattering, differential scanning calorimetry, and solid-state 13C nuclear magnetic resonance spectroscopy. Then their enzymatic degradations were carefully evaluated in the presence of P3HA depolymerase isolated from Ralstonia pickettii T1. In particular, more attention was paid to the δ-form P3HP, not only because it is completely new but also because it, as the same as the P3HB crystallized spontaneously, adopts the 21-helix12a,c conformation. For comparison, the enzymatic degradation of the cast P3HB film is also investigated simultaneously. The effect of crystalline structure on the biodegradation has been clarified with clearing out the interfering from the simultaneous alteration of crystallinity and lamellar thickness. Finally, the inherent factor responsible for the unique P3HP enzymatic degradation has been discussed by comparing the molecular interaction and dynamics of polymorphic crystals.

Experimental Section Materials. Two chemosynthesized P3HP samples (P3HP18k, Mn ) 1.8 × 104, Mw/Mn ) 2.48; P3HP70k, Mn ) 7.0 × 104, Mw/Mn ) 2.03), prepared by the ring-opening polymerization of propiolactone, were kindly supplied by Tokuyama Co. (Tokyo, Japan). The bacterial P3HB sample (Mn ) 5.6 × 105, Mw/Mn ) 2.2), was supplied by

Zhu et al. Mitsubishi Gas Co., Ltd. (Tokyo, Japan). All the samples were purified from chloroform solution by precipitation in heptane. Preparation of Films for Enzymatic Degradation. Two kinds of P3HP films were prepared for the enzymatic degradation. The meltcrystallized films were compression molded under 5 MPa at given temperatures of melt for 3 min (containing 1.5 min of postannealing without pressure), then quenched to desired temperatures, and annealed for appropriate times to generate polymorphic crystalline phases. The cast films were prepared by casting the chloroform solutions with an appropriate concentration, 10 mg/mL, on Teflon dishes. The solvent was allowed to evaporate slowly at 25 °C for at least 1 week. Analytical Procedures. The measurements of wide-angle X-ray diffraction (WAXD) were carried out on a Rigaku RU-200 (Rigaku Co., Tokyo, Japan), working at 40 kV and 200 mA, with Ni-filtered Cu KR radiation (λ ) 0.15418 nm). Scans were made between Bragg angles of 5–60° at a scanning rate of 1° min-1. The small-angle X-ray scattering (SAXS) profiles were registered by the same instrument between Bragg angles of 0.1–2.5°. The intensity was registered with every 0.004°, and X-ray was collected for 20 s at each step. The melting behaviors of polymorphic P3HPs were monitored by a Pyris Diamond DSC instrument (Perkin-Elmer Japan Co., Yokohama, Japan). The scales of temperature and heat flow at different heating rates were carefully calibrated using an indium standard with nitrogen purging. The samples between 5 and 6 mg were heated at 10 °C/min. Solid-state 13C NMR spectra were recorded at 30 °C on a JEOL GSX-270 spectrometer (JEOL Co., Tokyo, Japan) operating at 270.1 MHz for 1H ) 60.8 MHz for 13C. Typical NMR experimental conditions were as follows: 1H 90° pulse length, 5.5 µs; crosspolarization (CP) contact time, 1 ms; recycle delay, 10 s; 1H decoupling field strength, 54.1 kHz; magic angle spinning (MAS) rate, 5 kHz. Chemical shifts were externally referenced to methyl carbons of hexamethylbenzene at 17.36 ppm. The conventional CPT1 pulse sequence developed by Torchia15 was used to monitor the spin–lattice relaxation, and the 13C spin–lattice relaxation time (T1C) of each phase was further estimated by the experiential exponential model. Enzymatic degradation rates were evaluated by the weight-loss measurement, as described below. At 37 °C, the films with initial mass of about 10 mg and initial dimension of 10 × 10 × 0.1 mm were incubated into 1.0 mL of 0.1 M phosphate buffer (pH ) 7.5) with 1.0 µg of extracellular P3HA depolymerase isolated from Ralstonia pickettii T1 (PhaZRpi).16 The reaction solution was incubated with shaking. The films were removed after reacting for a period of time, washed with distilled water, and dried to constant weight in vacuum before weight measurement. The average weight loss rate for three films was reported as the result of one sample. Control tests in the buffer solution free from enzyme showed no detectable weight losses over the time scale of the experiment.

Results Preparation of Polymorphic Crystals. As the growth of the δ- and γ-forms is favored during casting in the low molecular weight (MW) and the high MW P3HP samples,12a,c respectively, two P3HP samples with MWs of 7.0 × 104 and 1.8 × 104, i.e., P3HP70k and P3HP18k, were cast at 25 °C to prepare the γand the δ-forms, respectively. The cast P3HP70k and P3HP18k samples are denoted as γ(cast) and δ(cast), respectively, and their X-ray diffraction patterns are shown in Figure 1. It is known that the melt crystallizations of the γ- and the β-forms are favored thermodynamically at higher temperature and kinetically at lower temperatures, respectively.12b Thus, after melting at 130 °C, P3HP70k samples were crystallized at 0, 30, and 70 °C to prepare, respectively, the β-form, β+γ mixed form, and γ-form P3HP, which are denoted as β(IC0), β+γ(IC30), and γ(IC70), and their X-ray diffraction patterns are also shown in Figure 1. In addition, as found previously,12c some specific δ-nuclei can survive if the cast δ-form is melted

Crystalline Structure Dependence of Enzymatic Degradation

Figure 1. WAXD patterns of β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast).

Biomacromolecules, Vol. 9, No. 4, 2008

1223

Figure 3. Plot of ln[S3(Is - Ifl)] as a function of S1.81 for β(IC0).

i.e., δ(IC70) and δ(cast), are noted to have much longer long periods than the β- or γ-form containing films. The average thickness of the interface (Li), which may exist between the crystal and amorphous layers, could also be evaluated from the scatterings. As suggested by Koberstein et al.,17a the slit-smeared intensity (Is) in the Porod region, when the local density fluctuations (Ifl) subtract from it, follows

Is - Ifl ≈ (Ap ⁄ s3)e-38(σs)

Figure 2. Lorentz-corrected SAXS profiles of β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast). Table 1. Comparison of Crystalline Structures of the δ-, γ-, and β-forms of P3HP Crystalsa crystal form

unit cell (parameters)

12a,c

δ-form γ-form10c

β-form10a,c

orthorhombic (a (0.773), b (0.448), c (0.477)) orthorhombic (a (0.700), b (0.490), c (0.493))

conformation 21-helix trans trans

a Note: Schematic representations for the γ- and β-forms are available in ref 10c.

at a temperature below 120 °C, and the growth of the δ-form can be overwhelmingly preponderant at 70 °C after melting at a lower temperature, e.g., 110 °C. In this way, another δ-form sample has been prepared and is denoted as δ(IC70), which, as shown in Figure 1, shows the same X-ray diffraction as δ(cast), i.e., three characteristic diffraction peaks with 2θ values at 20.7°, 23.3°, and 30.0°. The available main characteristics of all three kinds of crystals are summarized in Table 1. Lamellar Morphology. The morphological structures of the P3HP films with polymorphic structures were probed by SAXS. Figure 2 shows the profiles of Lorentz-corrected intensity (Iq2) for these polymorphic films, where the scattering vector, q, equals to 4π sin θ/λ. On the basis of Bragg’s law, the lower the value of the peak position, the longer the long period, and vice versa. As noted in Figure 2, the long periods of the meltcrystallized P3HP70k films, i.e., β(IC0), β+γ(IC30), and γ(IC70), almost have no dependence on the crystalline structure. The cast P3HP70k sample, i.e., γ(cast), has a shorter long period than the melt-crystallized γ-form. In addition, the two δ films,

1.81

(1)

where σ is the variance of the Gaussian phase boundary, s ) 2 sin θ/λ, and AP is Porod’s asymptote. As shown in Figure 3, the plot of ln(IS - Ifl) versus s1.8 for the β(IC0) sample is obviously linear in the Porod region. Then, the value of σ could be calculated from the slope of the linear plot. In this way, the average thickness of the interface (Li), which is defined as 3σ,17 was evaluated for all samples, and the results are tabulated in Table 2. As noted, all the samples were found to contain a sizable interface, being well accordant with the NMR results shown later. The interface thicknesses for the melt-crystallized samples are very similar with each other, and the interface thicknesses for the cast samples are smaller than those of the corresponding melt-crystallized ones. The average long period (L), amorphous thickness (La), and crystal thickness (Lc) were calculated by the one-dimensional correlation function γ(r) assuming a two-phase model.18 Figure 4 displays an inverse Fourier transform to the scattering relation for the experimental scattering profile of the β(IC0) sample, which is obtained directly from the intensities smeared by a slit of infinite height. Before the inverse Fourier transformation, the subtraction of the local density fluctuations, extrapolation to S ) 0 by using Guinier’s law,19 and extrapolation to ∞ through the application of Porod’s law20 were all carried out. The first maximum in γ(r) corresponds to the long period, and the first intercept A of γ(r) with the abscissa occurs at

r ) Φ(l - Φ)L

(2)

where Φ is the volume fraction of one phase in semicrystalline stacks. Because of the quadratic nature of eq 2, it yields two possible solutions for Φ. Then the thicknesses of two-phase layers were calculated to be LΦ. A commercial program, i.e., Crystallinity Analysis for Windows Version 1.0 (Rigaku Co., Tokyo, Japan), which is built on the Vonk method,22 was applied in this work to deconvolute the X-ray diffraction profile of each sample into the amorphous halo and the crystalline peaks and calculate the mass crystallinity (Wc). Assuming the same density for all crystalline phases, the mass crystallinity was transformed into the bulk crystallinity (Φc) with considering the density difference

1224

Biomacromolecules, Vol. 9, No. 4, 2008

Zhu et al.

Table 2. Properties and Solid Structure of Polymorphic P3HP Films and P3HB sample

∆Hfa/mJ · mg-1

Tma /°C

Wcb/%

Φcb/%

Φclinc/ %

Lic/nm

Lc/nm

L1c/nm

L2c/nm

Lcc/nm

Lc - Li/nm

δ(cast) δ(IC70) γ(cast) γ(IC70) β(IC30) β(IC00) P3HB

90.3 80.1 92.3 82.4 69.2 65.4 91.4

81.2 82.5 73.4 79.7 78.6 78.6 179.3

42 38 54 53 52 55 63

41 37 53 52 51 54 n.d.d

38 36 53 53 53 57 n.d.

1.1 1.2 1.1 1.3 1.2 1.3 n.d.

8.8 9.7 5.9 6.9 7.1 6.7 n.d.

3.3 3.5 2.8 3.2 3.4 2.9 n.d.

5.5 6.2 3.1 3.7 3.7 3.8 n.d

3.3 3.5 3.1 3.7 3.7 3.8 n.d.

>2.2 2.3 2.0 2.4 2.5 2.5 n.d.

a ∆Hf and Tm are melting enthalpy and temperature, respectively, determined by DSC. b Wc and Φc are mass and bulk crystallinity, respectively, determined by WAXD. c Li, L, L1, L2, Lc, and Φclin are interface thickness, long period, thickness of thin layer, thickness of thick layer, thickness of lamella, and linear crystallinity determined by SAXS, respectively. d Not determined.

Figure 4. Normalized one-dimensional correlation function γ(r) for β(IC0). 23

between the crystalline and amorphous layer. As shown in Table 2, the bulk crystallinities of the two δ-form samples are close to or below 40%, while those of the others are above 50%. Therefore, the crystalline layer (Lc) was attributed to the thin one (L1) for the δ(IC70) and δ(cast) samples and to the thick one (L2) for the β(IC0), β+γ(IC30), γ(IC70), and γ(cast) samples. The linear crystallinity (Φclin) was then calculated to be Lc/L for each P3HP film. The calculated lamellar parameters for all the polymorphic films are tabulated in Table 2. As noted, even though the contribution of interface is considered, the crystal thickness for all the melt-crystallized samples still decreases in the order of β(IC0) > (or ≈) β+γ(IC30) > (or ≈) γ(IC70) > (or ≈) δ(IC70), and the lamellae of two cast films are thinner than those of the corresponding melt-crystallized ones. In addition, the cast γ-form crystal is found to be the thinnest among all the polymorphic crystals. Thermal Properties. In Figure 5 are shown the DSC heating thermograms of the polymorphic P3HP samples. The melting enthalpy, ∆Hf, and the melting temperature, Tm, for each sample are listed in Table 2. As the γ-form would reorganize into the perfect or thick phase via the melting recrystallization,12b it is reasonable that the γ-form samples present multiple melting peaks. Normally, no melt-recrystallization phenomenon can be found for the melt-crystallized δ-form. However, the cast δ-form film does show two melting peaks, among which the higher one should arise from the recrystallized phase, as evidenced by the DSC heating scans with various rates (not shown). As the δ-form is the thermally most stable,12b,c the melt-crystallized δ-form sample has the highest melting point among all the meltcrystallized P3HPs despite its smaller crystal thickness. In addition, the cast P3HPs, i.e., the δ(cast)- and the γ(cast)-forms, were found to melt at a lower temperature than the corresponding melt-crystallized ones, as their crystal thicknesses are thinner. By division of the melting enthalpy of one polymorphic sample by the crystallinity measured by the WAXD analysis,

Figure 5. DSC heating thermograms of β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast).

Figure 6. Solid-state CP/MAS 13C NMR resonances of β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast) at 30 °C.

the melting enthalpies of the β-, γ-, and δ-form crystals can be estimated to be 118.9 (β(IC0)), 174.2 (γ(cast)), and 215.0 (δ(cast)) or 210.8 mJ · mg-1 (δ(IC70)), respectively. Namely, the melting enthalpy decreases in the order of δ > γ > β, which is well in accord with the order of their thermal stability.12b,c Chain Dynamics. Figure 6 shows the CP-MAS 13C NMR spectra of the polymorphic samples for the enzymatic degradation experiment and the assignments of all carbon resonances. The downfield shoulder of the RCH2 resonance and the upfield one of the βCH2 resonance are attributed to the amorphous phase due to their faster relaxation dynamics than that of their counterparts (not shown). For comparison, the chemical shift of each carbon resonance for the crystalline phases of all polymorphic samples is also incorporated in Figure 6. As shown

Crystalline Structure Dependence of Enzymatic Degradation

Figure 7. 13C spin–lattice relaxation curve for RCH2 of (a) δ(cast), (b) γ(cast), and (c) β+γ(IC30).

in Figure 6, the two δ-form samples show very similar resonance lines, and so does the two γ-form samples, which confirm the sensitivity of the CP-MAS 13C NMR spectra to the crystalline structure of P3HP. The β- and γ-forms also have similar RCH2 and βCH2 carbon resonances, which is well in accord with their conformational similarity.12b However, a large difference of carbonyl carbon chemical shift (∼0.7 ppm) exists between the two crystal forms, due to the stronger intermolecular interaction in the γ-form.12b Then, it is reasonable to note that the line shape of the carbonyl carbon resonance of the β+γ(IC30) sample depends on the relative content of the β- or γ-form, as shown in Figure 6. With respect to the trans-conformed β- and γ-forms, the 21-helix conformed δ-form presents a 0.4–07 ppm downfield shift at the RCH2 resonance and a 0.3–0.6 ppm upfield shift at the βCH2 resonance.12d The small upfield shift for the βCH2 resonance occurred with adopting the 21-helix conformation is possibly attributable to the combined action of the recovery of the βCH2 from the γ-eclipsed interaction with the carbonyl oxygen (deshielding) and the γ-gauche effect relative to the ester oxygen (shielding).12d,25 The classic CPT1 pulse sequence15 was used to measure the 13 C spin–lattice relaxation times (T1C). In addition to the β(IC0), γ(IC70) and δ(IC70) samples,12d the other three P3HP samples, i.e., β+γ(IC30), γ(cast), and δ(cast), were also included in the present work to compare their dynamics. As shown in Figures 7 and 8, the decay curves of the RCH2 and βCH2 resonances were fitted by the empirical multiexponential function to evaluate the values of the T1C. Similar to the time-progressive magnetizations of the β(IC0), γ(IC70), and δ(IC70) samples,12d those of the γ(cast) and δ(cast) samples could not be fitted well by the monoexponential and dual-exponential functions but could by the triexponential function, indicating that the three samples consist of three phases. For comparison, the β+γ(IC30) was also fitted similarly. In this case, the relaxation dynamics of crystalline phase was averaged by fitting with considering the contribution of each phase. In Table 3 are tabulated the fitted results for all the six samples. Reasonably, the longest T1C is ascribed to the restricted local motion in the well-ordered

Biomacromolecules, Vol. 9, No. 4, 2008

1225

Figure 8. 13C spin–lattice relaxation curve for βCH2 of (a) δ(cast), (b) γ(cast), and (c) β+γ(IC30). Table 3. 13C Spin–Lattice Relaxation Times of All Polymorphic P3HP Samples T1C/s (fraction) δ(cast) δ(IC70)

12d

γ(cast) γ(IC70)12d β+γ(IC30) β(IC0)12d

βCH2 RCH2 β CH2 RCH2 βCH2 RCH2 βCH2 RCH2 βCH2 RCH2 βCH2 RCH2

crystalline

interface

amorphous

98 (41%) 121 (41%) 126 (47%) 148 (46%) 58 (51%) 67 (49%) 87 (51%) 108 (48%) 52 (60%) 56 (60%) 44 (49%) 49 (49%)

4.3 (30%) 4.5 (32%) 5.2 (25%) 4.9 (26%) 2.0 (23%) 2.5 (26%) 2.2 (27%) 3.4 (26%) 1.1 (27%) 1.3 (27%) 2.1 (14%) 2.8 (13%)

0.35 (29%) 0.36 (27%) 0.35 (28%) 0.43 (28%) 0.34 (26%) 0.21 (25%) 0.27 (22%) 0.24 (26%) 0.30 (13%) 0.34 (13%) 0.30 (37%) 0.38 (38%)

crystalline phase, and the shortest one should be associated with the fast motion in the amorphous phase. On comparison of the resonances relaxed at different extents (not shown), the phase associated with the medium T1C has been further clarified to be the mobile crystalline phase, as its resonance is nearly the same as that of the well-ordered crystalline phase.24 Considering that an interface has been detected by the SAXS analysis, it is reasonable to assign the mobile crystalline phase to the interface between the crystals and amorphous layers. As tabulated in Table 3, the value of T1C of the well-ordered crystalline phase is noted to depend on the crystalline structure. With considering the crystalline layer of the δ-form sample is thinner than those of the β- and γ-form samples, and the latter almost have no difference (except for those of the γ(cast) sample), the increase of T1C in the order of β < γ < δ should mostly result from the alteration of the crystalline structure rather than that of the crystal thickness. In fact, we were impressed by the fact that the crystal in the γ(cast) sample relaxes much slower than that in the β(IC0) sample, although the latter is much thicker. Furthermore, the T1C value of the crystalline phase was found to depend on the sample preparation method. In the same crystalline structure, the cast sample is noted to relax faster

1226

Biomacromolecules, Vol. 9, No. 4, 2008

Zhu et al.

Figure 9. Weight loss of a variety of P3HP films (β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast)) at 37 °C in the presence of P3HA depolymerase from R. pickettii T1 as a function of incubation time.

Figure 10. Weight loss rate for β(IC0), β+γ(IC30), γ(IC70), γ(cast), δ(IC70), and δ(cast) estimated from the slope of plots of weight loss vs incubation time in Figure 6.

than the melt crystallized one, which is quite attributable to its thinner thickness. On the whole, the T1C value of the crystalline phase ranks as β(IC0) < β+γ(IC30) < (or ≈) γ(cast) < γ(IC70) < δ(cast) < δ(IC70). Although the T1C value quite possibly underrates the molecular motion of the stable phase due to the larger contribution of molecular interaction to the relaxation dynamics in the stable phase, the T1C value still can be used to describe the relative difference in the chain mobility of P3HP.12d Thus, the molecular mobility of the crystalline phase for the six polymorphic samples should rank as β(IC0) > β+γ(IC30) > (or ≈) γ(cast) > γ(IC70) > δ(cast) > δ(IC70). Enzymatic Degradation. The enzymatic degradation of P3HP films at 37 °C was investigated in the presence of P3HA depolymerase isolated from Ralstonia pickettii T1. In Figure 9 are shown the plots of the normalized weight loss as a function of the incubation time for all samples. The total incubation of each film lasted no more than 12 h so that each film kept its original shape at the end of the test. The data of the cast bacterial P3HB were also included in Figure 9 for the sake of comparison. All the films continuously lost weight upon exposure to the P3HA depolymerase. Similar to that of the P3HB film, the normalized weight loss of all P3HP films shows almost a linear regression with incubation progressing in the experimental time range. Namely, the enzymatic degradation of P3HP films proceeds at a constant rate independent of the exposure time. In Figure 10 are shown the weight loss rates of all films calculated from the linear regression of the normalized weight loss data shown in Figure 9. As noted, all the P3HP films degraded faster than the bacterial P3HB one. However, each P3HP sample has its own unique enzymatic degradation rates, regardless of holding the same primary structure. Among all P3HP samples, the weight loss rate of the δ(cast) film (0.22 mg · cm-2 · h-1) is the slowest, which is about 1 order of magnitude smaller than that of the fastest degraded β(IC0) film (1.73 mg · cm-2 · h-1). Another δ-form sample, i.e., the δ(IC70) film, degraded a little faster (at 0.27 mg · cm-2 · h-1) than the δ(cast) film, possibly due to its smaller crystallinity.5 The two γ-form samples, i.e., the γ(cast) and γ(IC70) films, degraded, respectively, at 1.19 and 0.69 mg · cm-2 · h-1. Both of them degraded much faster than the δ-form samples but slower than the β(IC0) film. In addition, the γ(cast) film was noted to degrade faster than the γ(IC70) one, which is attributable to its thinner crystal thickness.6 It is reasonable that the β+γ(IC30) film, which consists of the γ- and β-form crystals, degraded at a medium rate (1.37 mg · cm-2 · h-1) with respect to the γ- and

β-form samples. It was also noted that except for the δ-form samples, the P3HP films degraded much faster than the bacterial P3HB one. On the whole, the weight loss rates of the P3HP films decrease in the order of β(IC0) > β+γ(IC30) > (or ≈) γ(cast) > γ(IC70) > δ(IC70) >(or ≈) δ(cast). One may note that the two δ-form samples have a lower molecular weight than the other polymorphic P3HP samples. Considering the possible coexistence of the endo and exo type activities for the P3HA depolymerase isolated from Ralstonia pickettii T1,26a,b the use of the low molecular weight δ-form samples would not damage but strengthen the above conclusion.

Discussion It is well-known that for a given P3HA, the enzymatic degradation rate decreases as a consequence of an increase of the crystallinity5 and the lamellar thickness.6 A similar trend is also expected to be observed in enzymatic degradation of P3HP. As tabulated in Table 2, under the experimental condition, both the crystallinity and the lamellar thickness of P3HP samples (except γ(cast)) decrease roughly in the order of β-form > (or ≈) γ-form > δ-form. If the P3HP enzymatic degradation behavior only depends on the crystallinity and the lamellar thickness, the enzymatic degradation rates are expected to increase in the order of β-form < (or ≈) γ-form < δ-form. However, the experimental result is in the inverse order, i.e., β-form > γ-form > δ-form, unambiguously indicating that the crystalline structure is one of the most important factors (more important than the crystallinity and lamellar thickness in this case) in controlling the enzymatic degradation of P3HP. In fact, one could be more impressed by the fact that the γ(cast) sample, which, with respect to the β(IC0) sample, has a comparable crystallinity but consists of much thinner crystals, degraded much slower than the β(IC0) sample, which is far from our previous understanding on the P3HA enzymatic degradation. Until now, we have clarified what kind of role the crystalline structure plays in the enzymatic degradation of P3HP, but the fundamental issue of why the polymorphic P3HPs differ in the enzymatic degradability still remains open. On review of the above data, it is easy to find that not only the enzymatic degradability but also the chain mobility depends on the crystalline structure. More than that, by comparing the crystalline structure dependence of the chain mobility with that of the enzymatic degradability, it is found that the lower the chain mobility, the poorer the enzymatic degradability. Such a

Crystalline Structure Dependence of Enzymatic Degradation

relevancy between the chain mobility and the enzymatic degradability for one crystal form could be reasonably explained, if we recall the enzymatic degradation process in a microcosmic view. During the enzymatic degradation, the mobile chains can be adjusted more easily to adopt some specific conformation, which is necessary for the further catalytic action, by interacting with the catalytic domain of enzyme. This is not an assumption at all, as much support, in addition to us, is available from the intense previous work on P3HA, although the effect of crystalline structure was usually not involved. For example, the amorphous chains are well-known to be much easier degraded by enzymes than the crystalline ones as they are softer, the enzymatic degradation preferentially starts from the mobile crystal edges rather than the chain folds of the lamellar surfaces, and the less ordered and thermally less stable regions along the b-axis of the single crystal are more sensitive to the enzymatic attack.26 In addition, the previous finding6 that the thicker P3HA lamellae are more difficult to be enzymatically degraded also supports our opinion, as the P3HA chains, in this case, are obviously more rigid. On the other hand, when taking all in all, one could naturally realize that the chain mobility possibly predicts the enzymatic degradability for one P3HA whether it is crystalline or not. The difference in chain mobility of the three crystal forms, which plays a decisive role in the enzymatic degradation of P3HP, could be ascribed to their difference in conformation and packing as discussed below. Both the β- and the γ-forms consist of the similar all-trans conformed chains but differ in the packing efficiency along the a-axis.12a,c,d The difference in the packing efficiency obviously reflects the difference in the packing energy, i.e., that in the level of the intermolecular interaction.12b,d The IR evidence indicates that an intermolecular interaction, involving intermolecular carbonyl dipole–dipole interaction, in the γ-form is much stronger than that in the β-form. Thus, it is reasonable that the difference in the chain mobility between two crystals should be mainly attributed to their different packing efficiencies. As to the δ-form, the packing way is still unknown. However, for the stability of a crystal, the contribution from the conformational energy is generally more important than the packing energy, as the depth of intramolecular potential energy minima is much more pronounced than that pertaining to minima of packing energy. The IR evidence indicates that the intramolecular interaction, possibly involving one stronger intramolecular dipolar interaction, in the δ-form is much stronger than that in the β- and γ-forms.12d It is then suggested that the 21-helix conformation,12c rather than the packing way, should be primarily responsible for the more depressed chain mobility of the δ-form in comparison to the β- and the γ-forms, although a strong intermolecular interaction is also detected in the δ-form. In fact, it is the δ-form P3HP, which conforms to the 21-helix similar to the crystalline P3HB, that exhibits a degradation rate nearest to that of P3HB among all the three kinds of crystals, confirming again the existence of the distinctive influence of the conformation. However, keep in mind that we emphasize the effect of the crystalline structure here does not indicate that the effect of the lamellar thickness could be neglected in evaluating the enzymatic degradation of P3HA, although the latter is possibly limited with respect to the former.

Conclusion The cast or the melt-crystallized P3HP films consisting of the β-, γ-, and/or δ-form were prepared for the enzymatic

Biomacromolecules, Vol. 9, No. 4, 2008

1227

degradation test by controlling the crystallization temperature, temperature of melt, and molecular weight of P3HP. A considerable interface has been detected in all the P3HP samples by the SAXS analysis, which is in accord with the 13C spin–lattice relaxation dynamics of the P3HP samples. Furthermore, these interfaces are attributable to the mobile crystalline phases, as their resonance lines are very similar to those of the crystalline cores. The crystallinities of these P3HP films were found to decrease roughly in the order of β > (or ≈) γ > δ, and so do their lamellar thicknesses (even with considering the contribution of interface). The enzymatic degradation rate of P3HP at 37 °C in the presence of P3HA depolymerase from Ralstonia pickettii T1 was found to decrease in the order of β-form > γ-form > δ-form, which is contrary to the predication on the basis of the crystallinity and the lamellar thickness, and thus indicates that the crystalline structure plays a strikingly decisive role in the P3HP enzymatic degradation. In particular, only when the P3HP chain, which has the basic backbone structure of the P3HAs, conformed to the 21 helix, similar to the P3HB, does it has a similar biodegradability to the crystalline P3HB. It was further suggested that the chain mobility decided by the packing way and conformation should be responsible for the unique P3HP enzymatic degradation behavior. Further investigations are now being considered to extrapolate this consideration to other enzyme/PHA systems. Acknowledgment. Bo Zhu gratefully acknowledges Japan Society for the Promotion of Science (JSPS) for providing the fellowship and the grant-in-aid to do this research at the Tokyo Institute of Technology.

References and Notes (1) (a) Doi, Y. In Microbial Polyesters; VCH Publishers: New York, 1990. (b) Inoue, Y.; Yoshie, N. Prog. Polym. Sci. 1992, 17, 571. (c) Yoshie, N.; Inoue, Y. In Biopolymers; Steinbüchel, A, Ed.; Wiley-VCH: Weinheim, 2001; Vol. 3,Chapter 6. (d) Feng, L.; Yoshie, N; Asakawa, N.; Inoue, Y. Macromol. Biosci. 2004, 4, 186. (e) Holmes, P. A., In DeVelopments in Crystalline Polymers; Bassett, D. C., Eds.; Elsevier: London, 1988; p 1. (f) Anderson, A. J.; Dawes, E. A. Microbiol. ReV. 1990, 54, 450. (2) (a) Lusty, C. J.; Doudoroff, M. Proc. Natl. Acad. Sci. U.S.A. 1966, 56, 960. (b) Yamada, K.; Mukai, K.; Doi, Y. Int. J. Biol. Macromol. 1993, 15, 215. (c) Mukai, K.; Yamada, K.; Doi, Y. Polym. Degrad. Stab. 1994, 43, 319. (d) Tanio, T.; Fukui, T.; Shirakura, Y.; Saito, T.; Tomita, K.; Kaiho, T. S.; Masamune, S. Eur. J. Biochem. 1982, 124, 71. (e) Kasuya, K.; Inoue, Y.; Tanaka, T.; Akehata, T.; Iwata, T.; Fukui, T.; Doi, Y. Appl. EnViron. Microbiol. 1997, 63, 4844. (f) Jendrossek, D.; Schirmer, A.; Schlegel, H. G. Appl. Microbiol. Biotechnol. 1996, 46, 451. (3) Baba, H.; Tanahashi, N.; Kumagai, Y.; Doi, Y. J. Chem. Soc. Jpn. 1992, 5, 527. (4) (a) Shimamura, E.; Scandola, M.; Doi, Y. Macromolecules 1994, 27, 4429. (b) Saito, Y.; Nakamura, S.; Hiramitsu, M.; Doi, Y. Polym. Int. 1996, 39, 169. (c) Doi, Y.; Kitamura, S.; Abe, H. Macromolecules 1995, 28, 4822. (d) Abe, H.; Doi, Y.; Aoki, H.; Akehata, T.; Hori, Y.; Yamaguchi, A. Macromolecules 1995, 28, 7630. (e) Kanesawa, Y.; Tanahashi, N.; Doi, Y.; Saito, T. Polym. Degrad. Stab. 1994, 45, 179. (f) Feng, L.; Wang, Y.; Inagawa, Y.; Kasuya, K.; Saito, T.; Doi, Y.; Inoue, Y. Polym. Degrad. Stab. 2004, 84, 95. (g) Yoshie, N.; Fujiwara, M.; Kasuya, K.; Abe, H.; Doi, Y.; Inoue, Y. Macromol. Chem. Phys. 1999, 200, 977. (h) Wang, Y.; Inagawa, Y.; Saito, T.; Kasuya, K.; Doi, Y.; Inoue, Y. Biomacromolecules 2002, 3, 828. (i) Cao, A.; Arai, Y.; Yoshie, N.; Kasuya, K.; Doi, Y.; Inoue, Y. Polymer 1999, 40, 6821. (j) Abe, H.; Doi, Y. Int. J. Biol. Macromol. 1999, 25, 185. (5) (a) Koyama, N.; Doi, Y. Macromolecules 1997, 30, 826. (b) Spyros, A.; Kimmich, R; Briese, B. H.; Jendrossek, D. Macromolecules 1997, 30, 8218. (6) Abe, H.; Doi, Y.; Aoki, H.; Akehata, T. Macromolecules 1998, 31, 1791. (7) (a) He, Y.; Shuai, X.; Kasuya, K.; Doi, Y.; Inoue, Y. Biomacromolecules 2001, 2, 1045. (b) Scandola, M.; Focarete, M. L.; Gazzano,

1228

(8)

(9)

(10)

(11)

(12)

(13) (14)

Biomacromolecules, Vol. 9, No. 4, 2008

M.; Matuszowicz, A.; Sikorska, W.; Adamus, G.; Kurcok, P.; Kowalczuk, M.; Jedlinski, Z. Macromolecules 1997, 30, 7743. (c) Focarete, M. L.; Ceccorulli, G.; Scandola, M.; Kowalczuk, Ma. Macromolecules 1998, 31, 8485. (a) Iwata, T.; Aoyagi, Y.; Fujita, M.; Yamane, H; Doi, Y.; Suzuki, Y.; Takeuchi, A.; Uesugi, K. Macromol. Rapid Commun. 2004, 25, 1100. (b) Orts, W. J.; Marchessault, R. H.; Bluhm, T. L.; Hamer, G. K. Macromolecules 1990, 23, 5368. (c) Yokouchi, M.; Chatani, Y.; Tadokoro, H.; Teranishi, K.; Tani, H. Polymer 1973, 14, 267. (a) Eling, W. B.; Gogolewski, S.; Pennings, A. J. Polymer 1982, 23, 1587. (b) Hoogsteen, W.; Postema, A. R.; Pennings, A. J.; Ten Brinke, G.; Zugenmaier, P. Macromolecules 1990, 23, 634. (c) Pan, P.; Kai, W.; Zhu, B.; Dong, T.; Inoue, Y. Macromolecules 2007, 40, 6898. (a) Suehiro, K.; Chatani, Y.; Tadokoro, H. Polym. J. 1975, 7, 352. (b) Wasai, T.; Saegusa, T.; Furukawa, J. Kogyo Kagaku Zasshi (J. Chem. Soc. Jpn., Ind. Chem. Sect.) 1964, 67, 601. (c) Furuhashi, Y.; Iwata, T.; Sikorski, P.; Atkins, E.; Doi, Y. Macromolecules 2000, 33, 9423. (d) Furuhashi, Y.; Iwata, T.; Kimura, Y.; Doi, Y. Macromol. Biosci. 2003, 3, 462. (e) Okamura, K.; Marchessault, R. H. In Conformation of Biopolymers; Ramachandran, G. N., Ed.; Academic: New York, 1967; Vol. 2, pp 709–720. (a) Iwata, T.; Fujita, M.; Aoyagi, Y.; Doi, Y.; Fujisawa, T Biomacromolecules 2005, 6, 1803. (b) Somani, R. H.; Yang, L.; Hsiao, B. S.; Agarwal, P.; Fruitwala, H. A.; Tsou, A. H. Macromolecules 2002, 35, 9096. (a) Zhu, B.; He, Y.; Asakawa, N.; Yoshie, N.; Nishida, H.; Inoue, Y. Macromol. Rapid Commun. 2005, 26, 581. (b) Zhu, B.; He, Y.; Asakawa, N.; Yoshie, N.; Nishida, H.; Inoue, Y. Macromolecules 2005, 38, 6455. (c) Zhu, B.; He, Y.; Asakawa, N.; Nishida, H.; Inoue, Y. Macromolecules 2006, 39, 194. (d) Zhu, B.; Kai, W.; Pan, P.; Yazawa, K.; Nishida, H.; Sakurai, M.; Inoue, Y. Submitted for publication. (a) Cornibert, J.; Marchessault, R. H. Macromolecules 1975, 8, 296. (b) Cornibert, J. Ph. D. Thesis, Université de Montréal, 1972. (a) Perego, G.; Melis, A.; Cesari, M. Makromol. Chem. 1972, 157, 269. (b) Prud’homme, R. E.; Marchessault, R. H. Macromolecules 1974, 7, 541. (c) Meille, S. V.; Konishi, T.; Geil, P. H. Polymer 1984, 25, 773. (d) Cornibert, J.; Marchessault, R. H., Jr.; Lenz, R. W. Macromolecules 1973, 6, 676. (e) He, Z.; Prud’homme, R. E. Macromolecules 1999, 32, 7655. (f) Grenier, D.; Prud’homme, R. E. Macromolecules 1983, 16, 302. (g) Duchesne, D.; Prud’homme, R. E. Polymer 1979, 20, 1199.

Zhu et al. (15) Torchia, D. A. J. Magn. Reson. 1978, 30, 613. (16) (a) Kasuya, K.; Inoue, Y.; Doi, Y. Int. J. Biol. Macromol. 1996, 19, 35. (b) Shirakura, Y.; Fukui, T.; Saito, T.; Okamoto, Y.; Narikawa, T.; Koide, K.; Tomita, K.; Takemasa, T.; Masamune, S. Biochim. Biophys. Acta 1986, 880, 46. (17) (a) Koberstein, J. T.; Morra, B.; Stein, R. S. J. Appl. Crystallogr. 1980, 13, 34. (b) Cruz, C. S.; Stribeck, N.; Zachmann, H. G. Macromolecules 1991, 24, 5980. (c) Stribeck, N.; Alamo, R. G.; Mandelkern, L.; Zachmann, H. G. Macromolecules 1995, 28, 5029. (18) (a) Balta-Calleja, F. J.; Vonk, C. G. In X-ray Scattering of Synthetic Polymers; Elsevier: Tokyo, 1989. (b) He, Y.; Zhu, B.; Kai, W. H.; Inoue, Y. Macromolecules 2004, 37, 3337. (c) Zhu, B.; He, Y.; Yoshie, N.; Asakawa, N.; Inoue, Y. Macromolecules 2004, 37, 3257. (19) Guinier, A.; Fournet, G. In Small-Angle Scattering of X-rays; John Wiley & Sons: New York, 1955. (20) (a) Porod, G. Kolloid-Z. 1951, 124, 83. (b) Porod, G. Kolloid-Z 1952, 125, 51. (c) Porod, G. Kolloid-Z. 1952, 125, 108. (21) Goderis, B.; Reynaers, H.; Koch, M. H. J.; Mathot, V. B. F. J. Polym. Sci., Polym. Phys. Ed. 1999, 37, 1715. (22) Vonk, C; G, J. Appl. Crystallogr. 1973, 6, 148. (23) Crescenzi, V.; Manzini, G.; Calzolari, G.; Borri, C. Eur. Polym. J. 1972, 8, 449. (24) (a) Kaji, H.; Horii, F. Macromolecules 1997, 30, 5791. (b) Kuwabara, K.; Gan, Z.; Nakamura, T.; Abe, H.; Doi, Y. Biomacromolecules 2002, 3, 1095. (25) (a) Ritcey, A. M.; Prud’homme, R. E. Macromolecules 1992, 25, 972. (b) Grover, S. H.; Guthrie, J. P.; Stothers, J. B.; Tan, C. T. J. Magn. Reson. 1973, 10, 277. (26) (a) Scherer, T. M.; Fuller, R. C.; Goodwin, S.; Lenz, R. W. Biomacromolecules 2000, 1, 577. (b) Iwata, T.; Doi, Y.; Tanaka, T.; Akehata, T.; Masakatsu, S.; Teramachi, S. Macromolecules 1997, 30, 5290. (c) Hocking, P. J.; Marchessault, R. H.; Timmins, M. R.; Lenz, R. W.; Clinton Fuller, R. Macromolecules 1996, 29, 2472. (d) Nobes, G. A. R.; Marchessault, R. H.; Briese, B. H.; Jendrossek, D. Macromolecules 1996, 29, 8330. (e) Iwata, T.; Doi, Y.; Kasuya, K.; Inoue, Y. Macromolecules 1997, 30, 833. (f) Iwata, T.; Doi, Y. Macromol. Chem. Phys. 1999, 200, 2429. (g) Lee, W.-K.; Iwata, T.; Abe, H.; Doi, Y. Macromolecules 2000, 33, 9535. (h) Sudesh, K.; Abe, H.; Doi, Y. Prog. Polym. Sci. 2000, 25, 1503.

BM701220X