Microbial synthesis of a novel terpolyester P(LA‐co‐3HB‐co‐3HP) from low‐cost substrates

Summary Polylactide (PLA) is a bio‐based plastic commonly synthesized by chemical catalytic reaction using lactic acid (LA) as a substrate. Here, novel LA‐containing terpolyesters, namely, P[LA‐co‐3‐hydroxybutyrate (3HB)‐co‐3‐hydroxypropionate (3HP)], short as PLBP, were successfully synthesized for the first time by a recombinant Escherichia coli harbouring polyhydroxyalkanoate (PHA) synthase from Pseudomonas stutzeri (PhaC1Ps) with 4‐point mutations at E130D, S325T, S477G and Q481K, and 3‐hydroxypropionyl‐CoA (3HP‐CoA) synthesis pathway from glycerol, 3‐hydroxybutyryl‐CoA (3HB‐CoA) as well as lactyl‐CoA (LA‐CoA) pathways from glucose. Combining these pathways with the PHA synthase mutant phaC1Ps (E130D S325T S477G Q481K), the random terpolyester P(LA‐co‐3HB‐co‐3HP), or PLBP, was structurally confirmed by nuclear magnetic resonance to consist of 2 mol% LA, 90 mol% 3HB, and 8 mol% 3HP respectively. Remarkably, the PLBP terpolyester was produced from low‐cost sustainable glycerol and glucose. Monomer ratios of PLBP could be regulated by ratios of glycerol to glucose. Other terpolyester thermal and mechanical properties can be manipulated by adjusting the monomer ratios. More PLBP applications are to be expected.


Introduction
As increasing global warming and plastic pollution threaten human sustainability, materials from renewable biomass are attracting attention due to their biodegradability and environmentally friendliness. Polylactide (PLA) is a representative of bio-based biodegradable polyester synthesized via combination of microbial lactic acid (LA) fermentation and chemical polymerization of lactide (Sudesh and Iwata, 2008;Nampoothiri et al., 2010;Park et al., 2012a,b). PLA has been used in areas of biomedical implants, food packaging and drug delivery. However, the complicated synthetic process including fermentation for LA production, LA purification, lactide esterification and lactide ring-opening polymerization increases the cost, meanwhile the heavy metal residues in the final polymer products could limit its food or medical applications. In addition, PLA poor thermal and mechanical properties are also adverse to its large-scale applications.
Polyhydroxyalkanoates (PHA) are a family of diverse polyesters synthesized by a variety of bacteria as intracellular carbon and energy storage compounds (Li et al., 2007;Chen and Hajnal, 2015;Koller and Rodr ıguez-Contreras, 2015). As biorenewable and biodegradable materials, the diversity of PHA provides different physical properties to suit various applications (Chen, 2009;Brigham and Sinskey, 2012;Koller, 2014). Copolymerization of LA with other hydroxyalkanoate (HA) monomers via microbial synthesis is one of the effective methods to improve the physical properties of PLA or PHA (Li et al., 2010).
Recently, several engineered PHA synthases able to utilize lactyl-CoA (LA-CoA) and 3-hydroxybutyric-CoA (3HB-CoA) as substrates were reported (Taguchi et al., 2008;Yang et al., 2011;Ochi et al., 2013). To deliver LA-CoA in recombinant Escherichia coli, propionyl-CoA transferases (Pct) in alanine fermentation pathway of several organisms including Clostridium propionicum and Megasphaera elsdenii were expressed in host strains respectively. A 6 mol% LA fraction in the copolyester was achieved. 15 Subsequently, efforts were made on increasing the LA ratio in the copolyester by regulating the metabolic flux or evolving the PHA synthase (Jung et al., 2010;Yamada et al., 2010;Shozui et al., 2011), and also by the uses of different organisms in addition to commonly used E. coli, including Corynebacterium glutamicum, Ralstonia eutropha and Sinorhizobium meliloti (Song et al., 2012;Park et al., 2013;Tran and Charles, 2015). In addition, low-cost carbon substrates, such as glucose and xylose, were applied to synthesize LA-based copolyesters (Park et al., 2012a,b;Nduko et al., 2014;Salamanca-Cardona et al., 2014), which could reduce production cost and facilitate its industrialization.
As mentioned, PLA has the major deficiencies of poor flexibility, ductility and thermal resistance, and copolymerization is possibly effective to improve the PLA properties. Therefore, LA copolyesters were investigated with P(LA-co-3HB) as a representative (Taguchi et al., 2008;Jung et al., 2010;Yamada et al., 2010;Shozui et al., 2011;Yang et al., 2011;Ochi et al., 2013). However, mechanical properties of PHB are similar to PLA, therefore, other monomers such as 3-hydroxyvalerates (3HV), 3-hydroxyhexanoate (3HHx) and glycolate (GA) were introduced into the LA copolyesters for property improvements (Shozui et al., 2010a,b;Choi et al., 2016;Li et al., 2016). Remarkably, poly(3-hydroxypropionate) (P3HP), a relatively new PHA family member, has become very interesting due to its strong mechanical properties including an elongation at break of more than 600% and Young's modulus of 3 GPa (Andreeßen and Steinb€ uchel, 2010;Zhou et al., 2011a,b;Meng et al., 2012). Thus, 3HP monomers in LA copolyesters could very likely compensate for the shortcomings of PLA. This study attempted to biosynthesize a LA-containing terpolyester consisting of LA, 3HP and 3HB with improved properties over PLA homopolyester.

Results and discussion
Engineering a pathway for biosynthesis of P(LA-co-3HBco-3HP) from unrelated carbon source For the above-mentioned purpose, the substrate specificity of PHA synthase is the most important factor determining the monomer constituents incorporated into PHA. A mutant PHA synthase from Pseudomonas stutzeri strain 1317 (PhaC1 Ps ) with 4-point mutations at E130D, S325T, S477G and Q481K was prepared as stated below, which could incorporate LA and other 3-HAs into PLBP terpolyester. In addition, a 3HP synthetic pathway combined with LA and 3HB pathways were constructed to supply the monomers for the terpolyester synthesis (Fig. 1).
The engineering pathway allows adjustments of 3HB, LA and 3HP monomer ratios in the terpolyesters by feeding various ratios of glucose to glycerol. When a higher LA ratio is needed, LA can be added to the culture to supply additional LA precursor.
Engineering a PHA synthase able to polymerize LA The substrate specificity into PHA synthase is the most important factor determining the monomer constituents incorporated into PHA. Common PHA synthases are capable of polymerizing the 3-hydroxyalkanoates (3HAs) CoA with variable carbon chain lengths (Bernd, 2003). LA, which is the monomer of PLA, is a 2HA, which is not accepted by natural PHA synthases. To synthesize PLBP terpolyesters, a mutant of PHA synthase, able to incorporate LA and other 3-HAs at the same time was obtained via site-directed mutagenesis on type II PHA synthase PhaC1 Ps of P. stutzeri 1317 which has already a versatile substrate specificity (Park et al., 2015). Five mutation sites on phaC1 ps (phaC2 ps ), including E130D, Fig. 1. Metabolically engineered pathways for production of terpolyester P(LA-co-3HB-co-3HP) or PLBP from unrelated carbon source. Enzymes encoded by each gene are described below: phaA, b-ketothiolase; phaB, NADPH-dependent acetoacetyl-CoA reductase; ldhA, lactate dehydrogenase; pct, propionyl-CoA transferase; dhaB, glycerol dehydratase; dhaT, 1,3-propanediol dehydrogenase; aldD, aldehyde dehydrogenase; pcs', propanoyl-CoA synthetase; phaC, PHA synthase from Pseudomonas stutzeri strain 1317 (PhaC1 Ps ) with 4-point mutations at E130D, S325T, S477G and Q481K.  (Taguchi et al., 2008;Jung et al., 2010;Yamada et al., 2010;Yang et al., 2011;Chuah et al., 2013;Ochi et al., 2013). With various combinations of these site mutations, up to 20 phaC1 Ps and phaC2 Ps variants were constructed (Table S1). To compare their activities towards LA polymerization, the 3HB-CoA synthetic pathway including genes of b-ketothiolase (phaA), NADPH-dependent acetoacetyl-CoA reductase (phaB) of R. eutropha, LA-CoA synthetic pathway including lactate dehydrogenase (ldhA) of E. coli and propionyl-CoA transferase (pct) of Clostridium propionicum were constructed into a series of plasmids termed pBLPCAB-X (X represents the specific phaC variant in the plasmid). pBLPCAB-Xs were derived from pBHR68 (kindly donated by Prof Steinb} uchel of M} unster Univ/Germany), with the three genes of ldhA, pct and phaC variant inserted downstream the P re promoter. The LA polymerization activities of typical phaC variants were determined by expressing the corresponding plasmids in E. coli S17-1 (Table 1).

Construction of an effective PLBP synthetic system
Aimed to produce a novel terpolyester PLBP, three synthetic pathways for each constituent were constructed ( Fig. 1), including PHB and PLA synthetic routes that were reported. The P3HP synthetic pathway was focused on as it would improve the mechanical properties of the terpolyesters. P3HP could be synthesized from 1, 3-propanediol, glycerol and glucose (Andreeßen and Steinb€ uchel, 2010;Zhou et al., 2011a,b;Meng et al., 2015). Glycerol was chosen in this study to regulate 3HP ratio in the copolyesters, whereas glucose was the substrate for PLA and 3HB synthesis. The 3HP synthetic pathway from glycerol consisted of genes encoding glycerol dehydratase (dhaB) of Klebsiella pneumoniae, 1,3-propanediol dehydrogenase (dhaT) and aldehyde dehydrogenase (aldD) of Pseudomonas putida KT2442 and ACS domain of trifunctional propionyl-CoA synthetase (pcs') functioning as a CoA ligase of Chloroflexus aurantiacus (Andreeßen and Steinb€ uchel, 2010;Zhou et al., 2011a,b;Meng et al., 2012). The p3HP1p plasmid was constructed based on the pSEVA351 for P3HP production and contained genes of dhaT, aldD, dhaB and pcs' (Silva- Rocha et al., 2013). Gene fragments dhaT-aldD, pcs and dhaB were amplified from plasmid pZQ03, pDC02 and pZQ01 respectively. Subsequently, the pSEVA351 backbone was ligated with these three fragments by Gibson assembly. When strengthen the expression of dominant gene dhaB controlling glycerol utilization, the optimized plasmid p3HP2p exhibited enhanced efficiency in P3HP synthesis when it was co-expressed with pBHR68 in recombinant E. coli S17-1 (Table S3).
Finally, recombinant E. coli S-LA-3HP harbouring plasmids pLA and p3HP2p was obtained. When cultivated in LB medium supplemented with 20 g L À1 glucose and 10 g L À1 glycerol, a new terpolyester P(90.41 mol% 3HB-co-7.78 mol% 3HP-co-1.81 mol% LA) was successfully synthesized for the first time from unrelated carbon sources glucose and glycerol.

Nuclear magnetic resonance analyses of the PLBP terpolyester
The composition and monomer sequence distribution of P(3HB-co-3HP-co-LA) was confirmed by nuclear magnetic resonance (NMR) (Fig. 3). From the 1 H NMR spectra (Fig. 3A), there were not only well-characterized proton resonances of B(2), B(3) and B(4) in 3HB units (3HB abbreviated as B, 3HP monomer abbreviated as P and LA abbreviated as A) but also additional four proton resonances assigned to the 3HP units such as P(2), P (3) and LA units including A(2) and A(3) with identical intensities based on the previous studies (Park et al., 2013(Park et al., , 2015Meng et al., 2015). The molar ratio of 3HB, 3HP and LA in the PHA copolyester was 90.41%, 7.78% and 1.81%, respectively, as calculated using the intensity of B(3), P(3) and A(2). The individual carbon species in 3HB, 3HP and LA monomer were also identified by specific 13 C NMR (Fig. 3B). The expanded spectra of individual splitting resonance of carboxyl carbon B(1), P (1) and A(1) in the copolyester were split into multiple peaks (Fig. 3C), which were assigned to the 3HBcentred (B(1)*B), 3HP-centred (P(1)*P), LA-centred (A (1)*A) and the three units comonomer sequences [B(1) *P(1)*A(1)] (N*M represents the interaction of monomer N and M) (Meng et al., 2015;Park et al., 2015). This phenomenon was due to different sequence arrangements of 3HB, 3HP and LA monomers in the polymer chains, which is common in random PHA copolyesters Tripathi et al., 2012).
The tacticity distribution of tri-block copolyester was studied via 13 C NMR based on various stereosequences. The good resolution of methylene regions B (2) was chosen as an example for analysis. Two sharp peaks corresponding to B(2)*B and B(2)*P*A were observed (Fig. 3D). No split in each peak was visible, indicating that the stereo-sequence was isotactic. If the polymer is syndiotactic or atactic, some diad or quadruple peaks should be observed in 13 C NMR (Kemnitzer et al., 1993;Hocking and Marchessault, 1995). The above detailed analysis of 1 H NMR and 13 C NMR spectra clearly indicated that PHA was a random copolyester P(3HB-co-3HP-co-LA) with an isotactic microstructure.

Regulation of the monomer ratios in PLBP terpolyesters
Recombinant E. coli S17-1 harbouring genes described in Fig. 1 could produce terpolyesters PLBP consisting of LA, 3HB and 3HP when grown in a mixture of glucose and glycerol as substrates ( Table 2). The terpolyesters PLBP were fully synthesized from unrelated carbon sources, which was more economic for PHA synthesis (Hermann-Krauss et al., 2013;Povolo et al., 2013). Compositions of monomers LA, 3HB and 3HP in PLBP could be adjusted by changing the glucose-to-glycerol ratios.
Lactic acid ratio was as low as 1.81% in the terpolyester when PLBP was produced by the wild-type E. coli S17-1 harbouring genes described in Fig. 1. To increase the LA ratio in PLBP, the synthetic pathway was optimized and the competitive pathways of LA were deleted. Gene ldhA was inserted in pLA plasmid and was overexpressed to produce more LA as ldhA functions in converting pyruvate into LA (Jung et al., 2010). However, it was found that overexpression of ldhA alone could not enhance the LA content in the terpolyester (Table S4) as a gene dld could convert LA back into pyruvate (Dym et al., 2000;Choi et al., 2016). Unexpectedly, the recombinant strain-containing plasmid pLA' without ldhA, replacing pLA plasmid of the PLBP synthetic system, synthesized a PLBP terpolyester with an increased LA ratio under the same culture conditions (Fig. S1.2). The deletion of ldhA in pLA plasmid resulted in a transcriptional change in those genes downstream the same promoter for the ones that were closer to the promoter (Fig. S1.1). It was proven that the transcriptional level of pct was correlated with the LA ratio in PLBP terpolyesters from the RT-PCR results (Fig. S1.2). It suggested that PCT-mediated CoA transfer reaction that transformed LA into LA-CoA was the time-limiting step of LA polymerization. Meanwhile, to improve LA production, a competitive pathway, pyruvate-formate pathway, was weakened. Gene pflA, the activator of pyruvate-formate lyase, was knocked out in E. coli S17-1 (Zhu and Shimizu, 2004;Shozui et al., 2010b;Zhou et al., 2011a,b). The gene pflA knockout lead to an obvious decrease in formate production of the strain, whereas the lactate concentration in the culture medium increased oppositely (Fig. S2).
Escherichia coli S17-1 DpflA harbouring two plasmids of pLA' and p3HP2p was constructed and studied in shake flasks using different substrates combinations (Table 2 and Fig. 4). Obviously, the LA ratio in PLBP terpolyester produced by this recombinant was increased compared with the wild-type E. coli S17-1. The substrates added to the culture were utilized simultaneously indicating the synthesis of random copolyesters (Fig. 4). In general, several tendencies could be summarized from the data: there was a positive correlation between LA ratio and glucose concentration. Especially when glucose concentration was decreased to 10 g L À1 , the LA ratio was down to 0.79% because most of the glucose Fig. 3. NMR analysis of PLBP terpolyester. 1 H NMR spectra (A) and 13 C NMR spectra (B) of random copolyester P(3HB-co-3HP-co-LA) containing 90.41 mol% 3HB, 7.78 mol% 3HP and 1.81 mol% LA, respectively, and its expanded 13 C NMR spectra of carboxyl carbon [B(1), P(1), A(1)] area (C) and methylene regions (D) in the terpolyester. B, P and A refer to 3HB, 3HP and LA; numbering schemes were the same as molecular formulations of polyester indicated in the inset in (A). N*M represents the interaction of monomer N and M. "i" indicates "isotactic." Chemical shifts are in ppm and tetramethylsilane (TMS) was employed as an internal chemical shift standard. was utilized for cell growth; when extra D, L-LA was added into the medium, LA ratio increased sharply to more than 27%, whereas the cell dry mass (CDM) was deceased oppositely due to the toxicity of LA; as the substrate of P3HP, glycerol affected 3HP ratio in PLBP terpolyester in a similar relationship as that between glucose and LA ratio. Interestingly, 3HP ratio reached its peak value of 12.03% when glucose was reduced to 10 g L À1 . Synthesis of PLBP terpolyester with variable compositions was achieved. At a high glucose concentration of 30 g L À1 , cells grew to over 5.2 g L À1 containing close to 60% PHA, this was both the highest cell growth and the highest PHA accumulation compared with other low glucose concentration ( Table 2). The terpolyester consisted of 9% 3HP, 12% LA and 79% 3HB. The result indicated that high concentration of glucose favoured formation of cell mass and PHA, especially PHB. At the lowest glycerol concentration of 2 g L À1 , 3HP had the lowest content of 5% in the terpolyester. LA addition improved LA ratios in the terpolyester, but it had negative effect on cell growth (Table 2). Hence, several strategies exist to change the composition of the monomer level by adjusting the ratio of fed carbon sources.

Physical characterization of PLBP with different monomer compositions
Two types of PLBPs were extracted and the compositions of them were determined by GC. The physical properties of PLBPs including molecular mass, thermal properties and mechanical properties were studied ( Table 3). The weight average molecular mass of PLBP was less than 2 9 10 5 , which was lower compared with its homopolyester and P(LA-co-3HB) copolyester, yet approximated the same as weight average molecular mass of PLBV terpolyesters (Shozui et al., 2010b). In addition, PLBPs inherited the ductility of P3HP as it had an improved elongation at break of over 100% when the 3HP ratio reached 15%. The thermal parameters including T m , T g and DH m of PLBP were in between of those of individual homopolyesters (PLA, P3HB and P3HP) as blocks. Interestingly, PLBPs had two T m and the lower one was close to the T m of PHB. Possibly, the presence of 3HB-rich segments in the terpolyester led to this phenomenon. The incorporation of 3HP into P(LA-co-3HB) significantly improved the tensile strength and elongation at break compared with its copolyester P(LA-co-3HB).

Conclusions
Recombinant E. coli S17-1 DpflA consisting of three synthetic pathways of lactyl-CoA, 3-hydroxypropionyl-CoA and 3-hydroxybutyryl-CoA was able to polymerize the three monomers to form random terpolyester P(LAco-3HB-co-3HP) or PLBP under the catalysis of phaC1 Ps (Q481K S325T E130D S477G) cloned from P. stutzeri strain 1317 with four specific point mutations. The terpolyester compositions can be controlled by changing the ratios of three substrates including glucose, glycerol and LA. The terpolyesters have changing properties depending on monomer ratios, Table 2. PLBP production using various substrate concentrations by E. coli S17-1 DpflA harbouring two plasmids of pLA' and p3HP2p.
Glu (g L À1 ) Gly (g L À1 ) L A( gL À1 ) CDM (g L À1 ) PHA (wt%) 3HP ( The recombinant was cultivated in LB medium supplemented with different concentrations of glucose (glu), glycerol (gly) and LA for 48 h in shake flasks. The data are the averages of three parallel experiments. CDM, cell dry mass; PHA, the terpolyester; 3HP, 3-hydroxypropionate; LA, lactate. Fig. 4. The concentration of glucose, glycerol and lactate in shakeflask studies. E. coli S17-1 DpflA harbouring two plasmids of pLA' and p3HP2p was cultured in LB medium supplemented with 20 g L À1 glucose, 10 g L À1 glycerol and 2 g L À1 lactate. Blue line, glucose; red line, glycerol; green line, lactate. Error bars represent the standard deviation of experiments conducted in triplicates. especially an enhanced elongation at break compared with the homopolyesters of 3HB and LA as well as their copolyesters.

Experimental procedures
Microorganism, plasmid and shake-flask culture conditions All the microorganisms and plasmids used in this study are listed in Table 4. E. coli Trans 1T1 and JM109 were used as the host strains for genetic manipulation. E. coli S17-1 was used for polymer production (Simon et al., 1983). All the E. coli strains were cultured in LB medium. LB medium contains (g L À1 ): 10 tryptone, 5 yeast extract and 10 NaCl. Glycerol and glucose were added into LB medium as carbon sources. The seed cultures stored at À80°C were inoculated into LB medium and cultivated at 37°C for 12 h at 200 rpm/min on a rotary shaker for reactivation INFORS HT,Hong Kong,China). Subsequently, the seed cultures were inoculated into 500 mL conical flasks containing 50 mL LB medium with an inoculation volume ratio of 5%.  (Zaman et al., 2011). f P3HB was synthesized by bacteria (Li et al., 2011). g P3HP was synthesized by bacteria (Zhou et al., 2011a,b). h P(LA-co-3HB) was produced by recombinant E. coli (Yamada et al., 2011). i Sample weight of P(90.4 mol% 3HB-co-7.8 mol% 3HP-co-1.8 mol% LA) was 18.2 mg. j Sample weight of P(79.8 mol% 3HB-co-13 mol% 3HP-co-7.2 mol% LA) was 17.5 mg. When an antibiotic selection pressure was required, the medium was supplemented with ampicillin (100 lg mL À1 ), kanamycin (50 lg mL À1 ) or chloramphenicol (34 lg mL À1 ). To increase the LA ratio in the terpolyester, various amounts of D, L-LA (1, 2 or 5 g L À1 ) were added into the LB medium along with NaOH-modulating pH to neutral. When P3HP was the constituent of the polymer, 5 lM vitaminB 12 (VB 12 ) was added into the medium to maintain the activity of glycerol dehydratase (dhaB).

PHA analysis, extraction and purification
Cells were harvested by centrifugation (CR21 GIII; Hitachi, Tokyo, Japan) at 10 000 rpm/min for 8 min, then washed with distilled water and centrifuged again. CDM were measured after the concentrated cells were lyophilized at À65°C with five times air pressure (LGJ-10C; SiHuanKeXue, Beijing, China). Thirty to fourty milligrams of lyophilized cells was used for the transesterification reaction in which 2 mL of transesterification mixture and 2 mL of chloroform were added in each transesterification test tube (Kato et al., 1996). After a 4 h transesterification at 100°C, the PHA content and monomer compositions of the cells were assayed by gas chromatograph (GC-2014; Shimadzu, Suzhou, China) (Ouyang et al., 2007). The intracellular polymers were extracted using a Soxhlet extractor (Soxtec 2050; Foss, Hilleroed, Denmark). The extracted PHA was purified via precipitation when mixed with the 10-folds volume of icecold ethanol and dissolved in chloroform for film casting.

Metabolic flux analysis
Concentrations of LA, glycerol and glucose were determined using a high-performance liquid chromatography (HPLC) (LC-20A; Shimadzu) equipped with an ion exchange column (Aminexs HPX-87H; Bio-Rad, 7.8 9 300 mm 2 , Hercules, California, USA) and a refractive index detector (RID-10A; Shimadzu). Gene pct transcriptional level was assayed using RT-PCR. Total RNA was extracted from recombinant E. coli strains using RNA prep pure Cell/Bacteria Kit (Tiangen, Beijing, China). cDNA was synthesized using Fastquant RT Kit (Tiangen) and then real-time PCR (RT-PCR) was carried out for mRNA analysis with SuperReal PreMix (SYBR Green; Tiangen). 16S rRNA was used as the inner standard. The experimental procedures were described (Lv et al., 2015).

NMR analysis on PHA
The 1 H and 13 C spectra were obtained using a JEOL JNM-ECA 600 NMR spectrometer to measure the polymer compositions, the chemical microstructures and the monomer sequences. Tetramethylsilane was used as the internal standard.

Molecular mass and other properties of PHA
Molecular mass was determined via gel permeation chromatography (GPC Spectra System P2000; Shimadzu) equipped with a Shimadzu HSG60 column at 40°C. The melting temperature (T m ), enthalpy of fusion (DH m ) and glass-transition temperature (T g ) were measured via differential scanning calorimetry (DSC-60; Shimadzu) in a temperature ranging from À80°C to 200°C under a nitrogen atmosphere of 50 mL min À1 . Thermal stabilities of the materials were studied by a thermogravimetric analyser (TA-Q50; TA Instrument, New Castle, Delaware, USA). Three to five milligrams of each sample was loaded at temperature ranging from 10 to 400°C in a nitrogen atmosphere of 50 mL min À1 (Shen et al., 2009). Mechanical properties were studied using a materials testing machine (INSTRON 3365; Instron, Grove City, Ohio, USA) at room temperature at a speed of 5 mm min À1 . Table S1. Site specific mutations of two PHA synthases phaC Ps variants. Table S2. Comparison of various LA polymerizing enzymes. Table S3. P3HP synthetic ability of E. coli harboring p3HP1p and p3HP2p plasmids, respectively. Table S4. Effects of expressing gene ldhA on LA synthesis. Fig. S1.1. Comparison of plasmids structure between pLA and pLA'.