Biosynthesis of P(3HB-co-3HHx) Copolymers by a Newly Engineered Strain of Cupriavidus necator PHB−4/pBBR_CnPro-phaCRp for Skin Tissue Engineering Application

Polyhydroxyalkanoates (PHAs) are biodegradable polymers synthesized by certain bacteria and archaea with functions comparable to conventional plastics. Previously, our research group reported a newly PHA-producing bacterial strain, Rhodococcus pyridinivorans BSRT1-1, from the soil in Thailand. However, this strain’s PHA synthase (phaCRp) gene has not yet been characterized. Thus, this study aims to synthesize PHA using a newly engineered bacterial strain, Cupriavidus necator PHB−4/pBBR_CnPro-phaCRp, which harbors the phaCRp from strain BSRT1-1, and characterize the properties of PHA for skin tissue engineering application. To the best of our knowledge, this is the first study on the characterization of the PhaC from R. pyridinivorans species. The results demonstrated that the expression of the phaCRp in C. necator PHB−4 had developed in PHA production up to 3.1 ± 0.3 g/L when using 10 g/L of crude palm kernel oil (CPKO) as a sole carbon source. Interestingly, the engineered strain produced a 3-hydroxybutyrate (3HB) with 2 mol% of 3-hydroxyhexanoate (3HHx) monomer without adding precursor substrates. In addition, the 70 L stirrer bioreactor improved P(3HB-co-2 mol% 3HHx) yield 1.4-fold over the flask scale without altering monomer composition. Furthermore, the characterization of copolymer properties showed that this copolymer is promising for skin tissue engineering applications.


Introduction
Skin tissue engineering is a promising alternative to replacing conventional dressings and autologous skin transplants for wound-healing purposes [1]. It aims to support skin regeneration when a clinical disorder influences the healing process, for instance, burn injuries, diabetic ulcers, vascular disease, immunocompromised conditions, severe damage, and aging [2]. Presently, a number of bioengineered skin structures have been generated, some of which are commercially available. However, the material used to fabricate scaffolds for engineered skin must possess crucial physical and mechanical properties to keep the wound from fluid loss and contamination [3]. Moreover, it should be easy to handle, biocompatible, biodegradable, and have identical structural and mechanical properties to natural skin.
To date, various bacterial strains have been reported with the ability to synthesize P(3HB-co-3HHx) copolymer, including the bacterial strain in the genus Rhodococcus [15][16][17][18][19][20]. However, with changing the bacterium and its PHA synthase (PhaC), which is the crucial gene that initiates PHA accumulation in prokaryotes, combined with the carbon sources, polymers with different monomeric compositions and suitable mechanical properties are achieved [21,22]. For instance, Budde et al. (2011) reported two Ralstonia eutropha strains that accumulate high levels of P(3HB-co-3HHx) from palm oil. These strains express a newly characterized PhaC gene from the strain Rhodococcus aetherivorans I24 [18]. Furthermore, Jeon et al. (2014) reported the production of P(3HB-co-3HHx) containing up to 40 wt% 3HHx monomers using engineered R. eutropha strains comprising deletions of the acetoacetyl-CoA reductase (phaB) genes and replacing the native PHA synthase with phaC2 from R. aetherivorans I24 using butyrate as a carbon source [19].
Our group recently isolated a newly bacterial strain of R. pyridinivorans BSRT1-1 from the soil in Thailand [23,24]. This strain can produce P(3HB) using fructose as a carbon substrate. However, the PhaC gene of this strain has not been characterized so far. Therefore, this study aims to biosynthesize the PHA using a newly engineered strain of Cupriavidus necator PHB − 4/pBBR_CnPro-phaC Rp , which replaces the native PhaC with PhaC from R. pyridinivorans BSRT1-1, and characterize the properties of the synthesized polymer to apply as a scaffold for skin tissue engineering application. To the best of our knowledge, this is the first study on the characterization of the PhaC gene from R. pyridinivorans species.

Bacterial Strains
All studied bacterial strains and plasmids are listed in Table 1. A bacterial strain of R. pyridinivorans BSRT1-1 was grown on Tryptic Soy Agar (TSA) (HiMedia, Maharashtra, India). C. necator was cultured on nutrient-rich agar (NR) (10 g/L meat extract, 10 g/L peptone, and 2 g/L yeast extract) with pH 6.8 at 30 • C. Escherichia coli was cultured in Luria-Bertani (LB) medium at 37 • C. Kanamycin at a final concentration of 50 µg/ mL was added to the media if necessary. All bacterial strains were stored at 4 • C for further study.

2.2.
Cloning of PHA Synthase Gene from R. pyridinivorans BSRT1-1 C. necator PHB − 4/pBBR_CnPro-phaC Rp , an engineered strain of C. necator PHB − 4, was constructed through transconjugating pBBR1MCS-2 plasmid with phaC1 promotor from C. necator (pBBR-CnPro) insert with the PHA synthase gene obtained from R. pyridinivorans BSRT1-1(phaC Rp ). The gDNA of R. pyridinivorans BSRT1-1 was extracted using the phenolchloroform technique [25] following the manufacturer's instructions. The obtained gDNA was used as a template for PCR amplification of phaC Rp . A forward (Fwd_phaC Rp ) and reverse primer (Rev_phaC Rp ) were designed to amplify phaC Rp . The pBBR1MCS2 plasmid was used as the shuttle and expression vector, while C. necator PHB -4, the PHA-negative mutant strain, was used as a host strain. The PCR product was digested with HindIII and ApaI restriction enzymes (Thermo Fisher Scientific, Waltham, MA, USA), and then, it was ligated with HindIII and ApaI-digested pBBR1MCS2 using DNA Ligation Kit Ver.2.1 (Takara Bio Inc., Kusatsu, Japan) according to the manufacturer's instructions. The resultant recombinant plasmids were transformed into E. coli 10G (Lucigen corporation, Middleton, WI, USA), E. coli S17-1 and further transconjugated to C. necator PHB − 4 as described  [26]. The successful transformants were verified by colony PCR and DNA sequencing. The DNA sequencing was conducted by Integrated DNA Technologies (IDT) Inc. (Singapore, Republic of Singapore). Finally, the culture of C. necator PHB − 4/pBBR_CnPro-phaC Rp was stored as glycerol stock in nutrient-rich (NR) broth medium with 30% v/v of glycerol supplemented with kanamycin (50 µg/mL) at −80 • C. Table 1. Bacterial strain and plasmids used in this study.
The inoculum of C. necator PHB − 4/pBBR_CnPro-phaCRp was prepared by transferring three full loops of a single colony in 50 mL of NR medium containing kanamycin (50 µ g/mL) and incubated at 30 °C with shaking at 200 rpm for 8 h. After that, the 3% v/v of prepared inoculum was transferred into the PHA production medium (MM) and -butyrolactone (GBL) (Sigma-Aldrich, St. Louis, MO, USA), 1,4 butanediol (BDO) (Merck, Darmstadt, Germany), sodium valerate, and sodium hexanoate (Table 3), were added to the culture medium at a final concentration of 2 g/L to generate the different PHA monomers.
The inoculum of C. necator PHB − 4/pBBR_CnPro-phaCRp was prepared by transferring three full loops of a single colony in 50 mL of NR medium containing kanamycin (50 µ g/mL) and incubated at 30 °C with shaking at 200 rpm for 8 h. After that, the 3% v/v of prepared inoculum was transferred into the PHA production medium (MM) and The mineral medium (MM) for PHA biosynthesis was comprised of 4.0 g/L of NaH 2 PO 4 , 4.6 g/L of Na 2 HPO 4 , 0. The inoculum of C. necator PHB − 4/pBBR_CnPro-phaC Rp was prepared by transferring three full loops of a single colony in 50 mL of NR medium containing kanamycin (50 µg/mL) and incubated at 30 • C with shaking at 200 rpm for 8 h. After that, the 3% v/v of prepared inoculum was transferred into the PHA production medium (MM) and incubated at 30 • C with 200 rpm for 48 h. The bacterial cells were collected by transferring the bacterial culture to a 50 mL centrifuge tube and centrifuged at 8000 rpm, 4 • C for 10 min. After that, the supernatant was discarded, and the cell pellets were washed with distilled water. Next, the cell pellets were transferred to a pre-weighed bijoux bottle, then frozen at −20 • C for overnight, and then lyophilized using a freeze dryer. Lastly, the lyophilized cells were weighed to calculate the DCW (g/L). The PHA content and monomer composition of PHA were analyzed by gas chromatography (GC) analysis [31].

Scale-Up in the 70 L Stirred-Tank Bioreactor
The fermentations were carried out in a 70 L stirred-tank bioreactor (B.E. MARUBISHI Co., Ltd., Pathumthani, Thailand) with a working volume of 50 L to improve biomass and P(3HB-co-3HHx) yield by C. necator PHB − 4/pBBR_CnPro-phaC Rp . The seed culture was prepared in NR media supplemented with kanamycin (50 µg/mL). The bioreactor containing MM media was sterilized at 121 • C for 30 min, cooled, and then inoculated with 3% (v/v) of seed culture. Batch cultivation was performed at 30 • C with a controlled pH of 6.8. A pH controller maintained the culture media by adding 3M of NaOH or H 3 PO 4 . The aeration rate and agitation speed were fixed at 0.25 vvm and 200 rpm, respectively. The PHA content and cell biomass were analyzed at 12 h time intervals for 72 h of fermentation. The fermentation experiments were performed in triplicates, and average values were calculated.

PHA Extraction and Purification
The extraction of PHA accumulated in the cells was performed using the chloroform extraction method [23]. First, 1 g of the freeze-dried cells was dissolved into a stirred chloroform (100 mL) at room temperature. After 3 days of incubation, the cell solution was filtrated through Whatman No. 1 filter paper to eliminate the cell residues. The filtrated solution was then concentrated to approximately 10 mL using a rotary evaporator at a temperature of 50 • C, followed by dropwise adding into a robustly stirred, ice-cold methanol (100 mL). The purified PHA was then harvested and air-dried for 3 days. The 1 H-NMR was performed to confirm PHA's chemical structure or monomer composition using a Bruker Advance 500 spectrometer (Bruker, Road Billerica, MA, USA) at 30 • C. About 3 mg of the purified PHA was dissolved in 1 mL of deuterated chloroform (CDCl 3 ) at a concentration of 25 mg/mL. Tetramethysilane was used as an internal chemical shift reference and NMR analysis was proceeded at 500 MHz with 64 scans. The spectra were calculated for each monomer by comparing with a previous publication.

Thermal Properties of Purified PHA
The thermal characteristic of the produced polymer was determined using a DSC-60 (Shimadzu, Kyoto, Japan) instrument. The flow rate of nitrogen was fixed at 30 mL/min. Approximately 5 mg of purified PHA was filled into an aluminum pan and heated from 25 to 200 • C with a 15 • C/min heating rate. The heated samples were then maintained at 200 • C for 2 min and subjected to rapid quenching to −40 • C. They were then heated again from −40 to 200 • C at a 15 • C/min heating rate. The glass transition temperature (T g ) and melting temperature (T m ) were observed from the DSC thermogram.

Molecular Weight of Purified PHA
The polymer's molecular weight was evaluated by gel permeation chromatography (GPC) (10A GPC system, Shimadzu, Kyoto, Japan). About 1 mg purified polymer was dissolved in 1 mL chloroform and analyzed at a column temperature of 40 • C. Polystyrene standards were used to generate a calibration curve. Molecular weight was reported as polystyrene equivalents.

Analytical Methods
The dry cell weight (DCW) of the bacterial cell was determined following the method described by Trakunjae et al. (2021) [23]. The analysis was performed by transferring 1 mL of the cell culture suspension to a pre-weighed Eppendorf tube, added in triplicate, then harvested by centrifugation at 8000 rpm for 10 min. After that, the biomass was washed with distilled water twice. Then, the cell pellet was frozen at −20 • C overnight, followed by lyophilization using a freeze dryer for 48 h. Finally, freeze-dried Eppendorf tubes were weighed again to verify the stability of DCW.
The PHA content and monomers composition of the harvested cell were analyzed by gas chromatography (GC) according to the method described by Braunegg et al. (1978) [31]. The 15-20 mg of freeze-dried cells were weighed in a screw-capped test tube, followed by adding 2 mL of chloroform and 2 mL of methanolysis solution (15% v/v sulfuric acid and 85% v/v methanol). The tubes were heated at 100 • C for 140 min and then cooled at room temperature. After that, 1 mL of distilled water was added, then vortexed for 1 min to achieve the layers of phase separation. The lower layer, containing hydroxyacyl methyl esters, was transferred into the tube containing sodium sulfate anhydrous to eliminate residue water. Then, the 500 mL of hydroxyacyl methyl esters solution and 500 mL of 0.2% v/v of caprylic acid methyl ester (CME) solution were mixed in a GC vial. In this case, the CME solution was used as an internal standard. The sample was analyzed by GC2014 (Shimadzu, Kyoto, Japan) equipped with an AOC-20i auto-injector (Shimadzu, Kyoto, Japan), Restek RTX-1 capillary column, and a flame ionization detector (FID). The injector temperature was set at 270 • C, and the temperature of column was set at 70 • C and slowly increased at a rate of 10 • C/min to 280 • C, while the detector temperature was established at 280 • C. Nitrogen was used as a carrier gas.

Statistical Analysis
All experiments were performed in triplicates. The obtained experimental data are indicated as the mean value of the triplicate with standard error.

Characterization of the PHA Synthase Gene
PhaC is the crucial enzyme for the synthesis of PHA, where the monomers are polymerized of PHA into PHA polymers. The substrate specificity of the PhaC determines the types of monomers incorporated into the PHA polymer [32,33]. In this study, the nucleotide and amino acid sequence of phaC Rp , a newly isolated PHA-producing strain, was obtained and used for the initial BLAST search. The nucleotide and amino acid length of phaC Rp were 1755 bp and 584 amino acids, respectively. According to the BLASTn analysis, the phaC Rp revealed significant similarity with the phaC of R. rodochrous ATCC BAA870 (CP032675.

PHA Biosynthesis and Heterologous Expression of phaCRp in C. necator PHB -4
PHA biosynthesis was performed in a 250 mL flask to evaluate PHA production by the newly engineered strain of C. necator PHB − 4/pBBR_CnPro-phaCRp and to select the best carbon and precursors for further studies. C. necator PHB − 4/pBBR_CnPro-phaCRp was cultured under nitrogen-limiting conditions at 30 °C for 48 h using different carbon sources and precursor or structurally related carbon sources. Of seven carbon sources, i.e., glucose, fructose, sucrose, glycerol, molasses, crude palm kernel oil (CPKO), and palm oil, CPKO was found to be the best carbon source for cell growth and PHA biosynthesis by C. PhaCs have been classified into four classes according to their primary sequences, substrate specificity, and subunit composition [33]. The class I, III, and IV PhaCs desire shortchain length monomers containing C3-C5 carbon chain lengths, such as 3-hydroxybutyrate (3HB). In contrast, class II PhaCs exhibit higher activities toward the monomers with medium-chain length comprising C6-C14 carbon chain lengths, such as the C6 monomer 3-hydroxyhexanoate (3HHx). However, PHA produced from short-chain-length and mediumchain-length monomer combinations has superior thermal and physical properties compared to the homopolymers. [33][34][35]. In 1999, Rehm and Steinbuchel reported that class I and class II PhaCs include enzymes containing only one type of subunit with molecular weights (M W ) between 61 and 68 kDa [36]. Furthermore, six conserved blocks of amino acid sequences were identified, revealing areas of higher similarity. In contrast, the N-terminal region, approximately 100 amino acids relative to type I PhaCs, is highly variable [37]. In this study, the NJ phylogenetic tree showed that the phaC Rp is a consistent clade with the PhaC of those bacteria in class I PHA synthases (Figure 1). Similarly, the phaC from R. ruber is encoding for a class I PHA synthase [37]. In addition, the gene coding for a putative PHA depolymerase (PhaZ), which is expected to be the enzyme involved in PHA mobilization, was reported downstream of the R. ruber phaC. In addition, R. aetherivorans revealed two putative class I PhaC genes identified as phaC1 Ra and phaC2 Ra [18]. Among the Rhodococcus species, R. aetherivorans I24 was extensively studied for PHA production [18]. When comparing the amino acid sequences of phaC Rp with phaC1 Ra and phaC2 Ra , it was shown that the phaC Rp and phaC1 Ra have very analogous sequences arrangement, which is 80% similar, while showing 39% similarity with the phaC2 Ra .

PHA Biosynthesis and Heterologous Expression of phaC Rp in C. necator PHB -4
PHA biosynthesis was performed in a 250 mL flask to evaluate PHA production by the newly engineered strain of C. necator PHB − 4/pBBR_CnPro-phaC Rp and to select the best carbon and precursors for further studies. C. necator PHB − 4/pBBR_CnPro-phaC Rp was cultured under nitrogen-limiting conditions at 30 • C for 48 h using different carbon sources and precursor or structurally related carbon sources. Of seven carbon sources, i.e., glucose, fructose, sucrose, glycerol, molasses, crude palm kernel oil (CPKO), and palm oil, CPKO was found to be the best carbon source for cell growth and PHA biosynthesis by C. necator PHB − 4/pBBR_CnPro-phaC Rp ( Table 2). Approximately 6.3 ± 0.4 g/L of DCW and 49 wt%DCW of PHA content were achieved when using CPKO as a carbon source, whereas 3.7 ± 0.3, 2.1 ± 0.2, 0.5 ± 0.0, 0.4 ± 0.0, and 0.3 ± 0.0 g/L of DCW were obtained when using PO, fructose, molasses, glycerol, glucose, and sucrose as a carbon source, respectively. In addition, the PHA contents of 41 ± 3, 36 ± 2, 30 ± 5, 1.1 ± 0, and 0.6 ± 0 wt%DCW were achieved when using PO, fructose, molasses, glycerol, and sucrose as a carbon substrate. However, this strain was not able to accumulate the PHA content when using glucose as a carbon source. Interestingly, the 2 mol% of 3HHx monomers were successfully incorporated by C. necator PHB − 4/pBBR_CnPro-phaC Rp without adding any precursor or structurally related carbon sources. Nevertheless, the low incorporation of 3HHx monomers fraction might be due to the high intracellular concentrations of 3HB-CoA, which limit 3HHx incorporation into the PHA produced by the C. necator PHB − 4/pBBR_CnPro-phaC Rp [18].
Plant oils have been reported as an excellent carbon source for PHA production by many bacterial strains. For example, the biosynthesis of P(3HB-co-3HHx) directly from palm oil has been reported by Budde et al. (2011) using recombinant Ralstonia eutropha strains expressing heterologous PhaCs from the bacterium R. aetherivorans I24 [18]. R. aetherivorans I24 was extensively studied for PHA production among the Rhodococcus species. Budde et al. [18] reported the production of P(3HB-co-3HHx) by R. eutropha (C. necator) Re2000 and Re2001, which were engineered bacterial strains containing phaC1 Ra and phaC2 Ra , respectively. The results revealed that strains Re2000 and Re2001 produced 5.98 and 1.09 g/L of P(3HB-co-3HHx) incorporated with 1.1 ± 0.3 and 1.5 ± 0.2 mol% when using palm oil as a sole carbon source, respectively. The P(3HB-co-3HHx) production from strain Re2000 was slightly higher than in this study, in which C. necator PHB -4/pBBR_CnPro-phaC Rp obtained 3.1 g/L of P(3HB-co-3HHx). However, this strain produced a molar fraction of 2 mol% 3HHX incorporated in P(3HB-co-3HHx), which was higher than those produced by strains Re2000 and Re2001.
Furthermore, the production of P(3HB-co-3HHx) copolymer including high 3HHx monomer fraction by C. necator strain Re2160/pCB113 was studied using CPKO as a carbon substrate [38]. In addition, the combinations of CPKO and oil palm tree trunk sap (OPTS) as a carbon substrate for the biosynthesis of P(3HB-co-3HHx) using the recombinant C. necator strain Re2058/pCB113 were reported by Muragan et al. (2016) [39]. Therefore, this study selected CPKO as the carbon source for further experiments.
Apart from carbon sources, several types of precursors or structurally carbon substrates were tested to enhance the P(3HB-co-3HHx) production and improve the ability of C. necator PHB − 4/pBBR_CnPro-phaC Rp for synthesizing different types of PHA monomers. The 2 g/L of five different precursors, i.e., sodium 4-hydroxybutyrate, E. coli S17 chromosome (Strain for conju pBBR1MCS-2 The Broad-ho pBBR_CnPro pBBR1MCS pBBR_CnPro-phaCRp pBBR_CnP pyridinivorans B
The inoculum of C. necator PHB − 4/pBBR_CnPro-phaCRp was prepared by transferring three full loops of a single colony in 50 mL of NR medium containing kanamycin (5 µ g/mL) and incubated at 30 °C with shaking at 200 rpm for 8 h. After that, the 3% v/v o prepared inoculum was transferred into the PHA production medium (MM) and -Butyrolactone, and 1,4-Butanediol. Moreover, the result shows that the addition of precursors decreased cell biomass.
To compare the PHA production ability of an engineered strain with the wild-type strain, the PHA biosynthesis of C. necator H16, C. necator PHB 4, R. pyridinivorans BSRT1-1, and C. necator PHB − 4/pBBR_CnPro-phaC Rp were performed using 10 g/L CPKO as sole carbon source without the addition of precursors. The results showed that the C. necator PHB − 4/pBBR_CnPro-phaC Rp produced 3.1 g/L of total PHA containing 2 mol% of 3HHx fraction. Meanwhile R. pyridinivorans BSRT1-1, which is the wild-type strain, and C. necator PHB -4, which is the negative control strain, cannot produce PHA when using CPKO as the sole carbon source (Table 4). However, compared to the well-studied PHA-producing strain C. necator H16, C. necator PHB − 4/pBBR_CnPro-phaC Rp produced less PHA. Nevertheless, this result revealed that with modifying the bacterium and its PhaC, combined with the carbon sources, polymers with different monomeric compositions are obtained [21,22]. Table 4. Comparison of PHA biosynthesis by C. necator H16, C. necator PHB -4, R. pyridinivorans BSRT1-1, and C. necator PHB − 4/pBBR_CnPro-phaC Rp using 10 g/L CPKO as sole carbon sources.

Scaling Up P(3HB-co-3HHx) Production in a 70 L Stirred-Tank Bioreactor
The enhancement of P(3HB-co-3HHx) production by C. necator PHB − 4/pBBR_CnPro-phaC Rp was carried out with a batch cultivation in a 70 L stirred-tank bioreactor containing 50 L of MM. A total of 10 g/L of CPKO was applied as a sole carbon source. The temperature, pH, aeration rate, and agitation speed were fixed at 30 • C, 6.8 and 0.25 vvm, and 200 rpm, respectively. During the first 48 h of fermentation, the growth of C. necator PHB − 4/pBBR_CnPro-phaC Rp gradually increased, then remained constant until 60 h; after that, it slightly decreased at 72 h ( Figure 2). Similarly, the PHA accumulation was increased in the bioreactor during the 48 h of fermentation but significantly decreased at 72 h ( Figure 2).
The enhancement of P(3HB-co-3HHx) production by C. necator PHB − 4/pBBR_CnPro-phaCRp was carried out with a batch cultivation in a 70 L stirred-tank bioreactor containing 50 L of MM. A total of 10 g/L of CPKO was applied as a sole carbon source. The temperature, pH, aeration rate, and agitation speed were fixed at 30 °C, 6.8 and 0.25 vvm, and 200 rpm, respectively. During the first 48 h of fermentation, the growth of C. necator PHB − 4/pBBR_CnPro-phaCRp gradually increased, then remained constant until 60 h; after that, it slightly decreased at 72 h ( Figure 2). Similarly, the PHA accumulation was increased in the bioreactor during the 48 h of fermentation but significantly decreased at 72 h ( Figure 2).
. The highest production of P(3HB-co-3HHx) was at 48 h when the biomass, PHA content, and 3HHx mol% were 7.7 ± 0.6 g/L, 56 ± 5 wt%DCW, and 2 mol%, respectively (Figure 2). According to the results, the biomass of C. necator PHB − 4/pBBR_CnPro-phaCRp in a 70 L stirred-tank bioreactor improved by 1.4-fold compared with the shake flask scale. However, this approach cannot increase PHA accumulation and the mol% fraction of 3HHx. This result might be due to the limitation of single-batch production [21,41]. Batch cultivations for manufacturing PHA are straightforward to operate but intrinsically limited productivity. The highest allowed concentration of nitrogen and carbon supply at the beginning of the fermentation batch is limited by the physiological criteria of the The highest production of P(3HB-co-3HHx) was at 48 h when the biomass, PHA content, and 3HHx mol% were 7.7 ± 0.6 g/L, 56 ± 5 wt%DCW, and 2 mol%, respectively ( Figure 2). According to the results, the biomass of C. necator PHB − 4/pBBR_CnPro-phaC Rp in a 70 L stirred-tank bioreactor improved by 1.4-fold compared with the shake flask scale. However, this approach cannot increase PHA accumulation and the mol% fraction of 3HHx. This result might be due to the limitation of single-batch production [21,41]. Batch cultivations for manufacturing PHA are straightforward to operate but intrinsically limited productivity. The highest allowed concentration of nitrogen and carbon supply at the beginning of the fermentation batch is limited by the physiological criteria of the production strain. Nevertheless, the development of PHA production using batch fermentation by various types of bacteria has been reported [42].
3.4. Characterization of P(3HB-co-3HHx) produced by C. necator PHB − 4/pBBR_CnPro-phaC Rp The 1 H NMR was performed to confirm 3HB and 3HHx units in the polymer fraction of PHA copolymer synthesized by the strain C. necator PHB − 4/pBBR_CnPro-phaC Rp . Figure 3 shows the 1 H NMR spectrum of H4 and H6, which represent the C4 methylene groups and C6 methyl-group in the 3HHx unit, respectively, implying the actual formation of the P(3HB-co-3HHx) copolymer [37,43]. However, the signal of H4 and H6 appears indistinct due to the low mol% fraction of 3HHx. tography (GPC). The weight-average molecular weight (Mw) of 6.27 × 10 5 Da was obtained with the polydispersity (Mw/Mn) of 1.70. Typically, the molecular weight of the PHA varies for several reasons, for instance, the availability of precursors for PHA synthesis, the type of PhaC, the expression level of PhaCs synthases, and the availability of enzymes that hydrolyze PHA [34].  In this study, the molecular weight of the P(3HB-co-2 mol% 3HHx) copolymers produced by C. necator PHB − 4/pBBR_CnPro-phaC Rp was analyzed by gel permeation chromatography (GPC). The weight-average molecular weight (M w ) of 6.27 × 10 5 Da was obtained with the polydispersity (M w /M n ) of 1.70. Typically, the molecular weight of the PHA varies for several reasons, for instance, the availability of precursors for PHA synthesis, the type of PhaC, the expression level of PhaCs synthases, and the availability of enzymes that hydrolyze PHA [34].
The molecular weight of P(3HB-co-3HHx), which has a high 3HHx monomer percentage, is typically significantly lower than that of P(3HB) homopolymer synthesized by wild-type C. necator H16. [46] Thermal properties of P(3HB-co-3HHx) copolymer synthesized by C. necator PHB − 4/ pBBR_CnPro-phaC Rp were analyzed using DSC. Defining the polymer thermal stability is essential for understanding the chemical recycling of polymer materials [38]. Values for DSC analysis were recorded from the second heating to eradicate the thermal history of the previous samples. The melting temperature (T m ) and glass transition temperature (T g ) of P(3HB-co-3HHx) were 160.18 and 6.04 • C, respectively ( Figure 4). The melting temperature of P(3HB-co-3HHx) copolymers are commonly lesser than those of P(3HB). However, there was no significant correlation between these parameters and 3HHx molar fractions [47].
PHAs have been reported for application for wound repair and skin tissue engineering applications [48][49][50][51]. The chemical properties that require consideration involve lowering the melting point and glass transition temperature. These characteristics rely upon the polymer's monomeric composition and molecular weight. PHA copolymers with high molecular weights can overcome these weaknesses [48]. However, after implantation, the main concern is the outcome of the products from PHA degradation and whether the PHA residuals are biocompatible with the cells at locations other than where they were proposed [49]. Among members of the PHAs family, P(3HB-co-3HHx) are promising biomaterials for the application in tissue engineering because they are biodegradable, biocompatible, and low cytotoxic. main concern is the outcome of the products from PHA degradation and whether the PHA residuals are biocompatible with the cells at locations other than where they were pro posed [49]. Among members of the PHAs family, P(3HB-co-3HHx) are promising bio materials for the application in tissue engineering because they are biodegradable, bio compatible, and low cytotoxic.
In this study, P(3HB-co-2 mol% 3HHx) copolymer was produced by a newly eng neered strain of C. necator PHB − 4/pBBR_CnPro-phaCRp, which possesses physical an chemical properties to be applied for wound repair or skin tissue engineering application Previously, Li et al. (2008) studied the blending of P(3HB)/P(3HB-co-3HHx) an P(3HB)/P(3HB-co-4HB) copolymers in several ratios for fabricating the nanofiber 3D sca folds and matrices cast solution for testing with human epidermal cells (keratinocyte [50]. Investigations of keratinocyte attachment and proliferation have shown that the sig nificant promotion of cell adhesion and proliferation could be obtained by fibrous matr ces. There was a striking resemblance between the porosity structure and the collage ECM organization. This study proved that the PHA-based matrices have the potential fo In this study, P(3HB-co-2 mol% 3HHx) copolymer was produced by a newly engineered strain of C. necator PHB − 4/pBBR_CnPro-phaC Rp , which possesses physical and chemical properties to be applied for wound repair or skin tissue engineering application.
Previously, Li et al. (2008) studied the blending of P(3HB)/P(3HB-co-3HHx) and P(3HB)/P(3HB-co-4HB) copolymers in several ratios for fabricating the nanofiber 3D scaffolds and matrices cast solution for testing with human epidermal cells (keratinocytes) [50]. Investigations of keratinocyte attachment and proliferation have shown that the significant promotion of cell adhesion and proliferation could be obtained by fibrous matrices. There was a striking resemblance between the porosity structure and the collagen ECM organization. This study proved that the PHA-based matrices have the potential for application in skin tissue engineering [50]. Furthermore, Tang et al. (2008) fabricated the nanostructured fibrous scaffolds of poly(3HB-co-5 mol% 3HHx), poly(3HB-co-7 mol% 4HB), and poly(3HB-co-97 mol% 4HB) copolymers [51]. They found that the tensile strength and Young's modulus of these scaffolds were comparable to those of human skin. The histological analysis confirmed the electrospun PHA scaffolds were well-tolerated in vivo after subcutaneous implantation [51].

Conclusions
This is the first study on the characterization of PhaC of R. pyridinivorans species. The strain C. necator PHB − 4/pBBR_CnPro-phaC Rp , a newly engineered strain of C. necator PHB − 4 with phaC Rp . CPKO and sodium hexanoate were found to be the best carbon source and precursor for P(3HB-co-3HHx) copolymer production by this strain, respectively. Under fermentation conditions, this strain can accumulate 56 ± 5 wt%DCW of PHA containing 2 mol% of 3HHx and produce 7.7 ± 0.6 g/L of biomass using a 70 L stirred-tank bioreactor. Furthermore, the properties of the P(3HB-co-3HHx) copolymer have demonstrated that this copolymer is promising for skin tissue engineering applications.
Author Contributions: Conceptualization, writing-original draft preparation, C.T.; review and editing, C.T., P.V., and K.S.; project administration, and funding acquisition, P.V. and C.T.; methodology and analyzed the data, C.T., S.Z.N., A.B., W.A., and P.J. All authors have read and agreed to the published version of the manuscript.