(R/S)-lactate/2-hydroxybutyrate dehydrogenases in and biosynthesis of block copolyesters by Ralstonia eutropha

Abstract Bacterial polyhydroxyalkanoates (PHAs) are promising bio-based biodegradable polyesters. It was recently reported that novel PHA block copolymers composed of (R)-3-hydroxybutyrate (3HB) and (R)-2-hydroxybutyrate (2HB) were synthesized by Escherichia coli expressing PhaCAR, a chimeric enzyme of PHA synthases derived from Aeromonas caviae and Ralstonia eutropha. In this study, the sequence-regulating PhaCAR was applied in the natural PHA-producing bacterium, R. eutropha. During the investigation, (R/S)-2HB was found to exhibit strong growth inhibitory effects on the cells of R. eutropha. This was probably due to formation of excess 2-ketobutyrate (2KB) from (R/S)-2HB and the consequent l-valine depletion caused by dominant l-isoleucine synthesis attributed to the excess 2KB. Deletion analyses for genes of lactate dehydrogenase homologs identified cytochrome-dependent d-lactate dehydrogenase (Dld) and [Fe-S] protein-dependent l-lactate dehydrogenase as the enzymes responsible for sensitivity to (R)-2HB and (S)-2HB, respectively. The engineered R. eutropha strain (phaCAR+, ldhACd-hadACd+ encoding clostridial (R)-2-hydroxyisocaproate dehydrogenase and (R)-2-hydoroxyisocaproate CoA transferase, ∆dld) synthesized PHA containing 10 mol% of 2HB when cultivated on glucose with addition of sodium (RS)-2HB, and the 2HB composition in PHA increased up to 35 mol% by overexpression phaCAR. The solvent fractionation and NMR analyses showed that the resulting PHAs were most likely to be block polymers consisting of P(3HB-co-3HV) and P(2HB) segments, suggesting that PhaCAR functions as the sequence-regulating PHA synthase independently from genetic and metabolic backgrounds of the host cell. Key points (R/S)-2-hydroxubutyrates (2HB) caused l-valine deletion in Ralstonia eutropha (R)- and (S)-lactate/2HB dehydrogenases functional in R. eutropha were identified The engineered R. eutropha synthesized block copolymers of 2HB-containing polyhydroxyalkanoates on glucose and 2HB Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00253-023-12797-6.


Introduction
Environmental pollution caused by plastic wastes is now becoming more serious on a global scale (Narancic et al. 2018;Waring et al. 2018).Polyhydroxyalkanoates (PHAs), which are biopolyesters accumulated within microbial cells as a carbon and energy storage, are eco-friendly alternatives to petroleum-based plastics because they can be produced from biomass feedstocks and show high biodegradability not only in soil but also fresh and sea water environments (Guzik et al. 2020;Miyahara et al. 2020).Poly((R)-3-hydroxybutyrate) [P(3HB)] is the most abundant PHA in nature.In general, P(3HB) is synthesized from acetyl-CoA through three consecutive reactions catalyzed by β-ketothiolase (PhaA), NADPH-dependent acetoacetyl-CoA reductase (PhaB), and PHA synthase (PhaC).It is also known that PHA-producing microbes potentially synthesize copolymers composed of two or more hydroxyalkanoate units.Biosynthesis of PHA copolymers by addition of precursor compounds into the media and metabolic and enzyme engineering approaches have been extensively studied in order to overcome the stiff and hard properties of P(3HB) homopolymer.
PHA synthase, the key enzyme in the PHA biosynthesis, catalyzes polymerization of (R)-3-hydroxyacyl (3HA)-CoAs via terminal extension of growing polymer chain (Neoh et al. 2022).PHA synthases belonged to α/β-hydrolase superfamily and classified into four classes based on the subunit composition and substrate specificity.The class I and II synthases are both single-subunit enzymes but show distinctly different substrate specificity from each other.The class I enzymes, including that from Ralstonia eutropha (Cupriavidus necator) H16 (PhaC Re ), are specific to short-chainlength (R)-3HA-CoAs from C 3 to C 5 , while the class II synthases derived from Pseudomonas spp.exhibit polymerization activity towards medium-chain-length (R)-3HA-CoAs of C 6 and the longer.Some PHA synthases exceptionally mediate polymerization of both short-and medium-chainlength (R)-3HA-CoAs, such as the class I PHA synthase from Aeromonas caviae FA440 (PhaC Ac ) polymerizing (R)-3HA-CoAs of C 4 -C 6 (Fukui and Doi 1997), and the class II enzyme from Pseudomonas sp.61-3 having broad substrate specificity to 3HA-CoAs from C 4 to C 12 (Matsusaki et al. 1998).PHA synthases often accept not only 3HA-CoAs but also 4-hydroxyacyl-CoAs and 5-hydroxyacyl-CoAs (Taguchi and Matsumoto 2021), whereas no natural synthase capable of accepting 2-hydroxyalkanoate (2HA) units has been known so far.In 2008, the S325T/Q481K double mutant of PHA synthase derived from Pseudomonas sp.61-3 (named as PhaC STQK ) was found to exhibit copolymerization activity to (R)-2HA-CoAs with (R)-3HB-CoA (Taguchi et al. 2008), expanding the structural range of microbial polyesters.Recombinant strains of Escherichia coli harboring phaC STQK synthesized PHA copolymers of d-lactate and (R)-3HB where the highest d-lactate fraction reached up to 97 mol% (Nduko et al. 2013;Shozui et al. 2010;Shozui et al. 2011).It was further demonstrated that (R)-2-hydroxybutyryl (2HB)-CoA and glycolyl-CoA were also accepted as substrates for PhaC STQK (Taguchi and Matsumoto 2021).
Unlike DNA and protein synthesis, PHA synthase-mediated polymerization occurs with template-independent manner that results in random distribution of the comonomer units in the polyester chain.This random polymerization property of PHA synthase is one of the limitations in the biological polyester synthesis, because synthesis of sequence-regulated copolymers, often showing altered properties from the corresponding random copolymers (Kumar et al. 2001), had been difficult by using PHA synthases.Recently, a chimeric PHA synthase consisting of N-and C-terminal regions of PhaC Ac and PhaC Re , respectively, was constructed aiming to create an engineered synthase compatibly possessing the broad substrate specificity of PhaC Ac and high polymerization activity of PhaC Re (Matsumoto et al. 2009).In vivo characterization of the chimeric PhaC (designated PhaC AR ) in E. coli exhibited the polymerization activity to not only (R)-3HA-CoAs but also (R)-2HA-CoAs.It should be further noted that the resulting PHAs synthesized by PhaC AR were block copolymers comprising of (R)-3HA-rich and (R)-2HA-rich segments, unlike the copolymers synthesized by PhaC STQK (Matsumoto et al. 2009;Matsumoto et al. 2018;Arai et al. 2020;Satoh et al. 2022).One of the block copolymers, poly((R)-2HB-block-(R)-3HB) [P(2HB-b-3HB)], was demonstrated to show elastomerlike properties that were not seen for the random copolymer, poly((R)-2HB-random-(R)-3HB) [P(2HB-ran-3HB)] ( Kageyama et al. 2021).PhaC AR is thus considered as a sequenceregulating PHA synthase.
R. eutropha H16 is a well-studied PHA-producing bacterium.There are vigorous studies for biochemical and molecular biological analyses regarding PHA biosynthesis by R. eutropha, as well as metabolic engineering of this bacterium focusing on efficient production of useful PHA copolymers such as flexible poly((R)-3HB-co-(R)-3-hydroxyhexanoate) [P(3HB-co-3HHx)] (Mifune et al. 2010;Insomphun et al. 2014;Insomphun et al. 2015;Zhang et al. 2019).The genome analysis of R. eutropha clarified that the 7.4 Mbp genome contains many genes related to metabolisms of fatty acids and carboxylic acids, such as 31 isologs of acyl-CoA synthetase/CoA ligase, 65 isologs of acyl-CoA dehydrogenase, 22 isologs of β-ketothiolase, and so on (Pohlmann et al. 2006), suggesting more complexed and robust metabolisms of R. eutropha towards acyl moieties when compared to E. coli.
This study investigated the polymerization property of PhaC AR in R. eutropha having quite different genetic and metabolic backgrounds from E. coli, since the sequenceregulating polymerization by PhaC AR has been shown only in E. coli as the host bacterium so far.During the research, we observed high growth toxicity of 2HB to the cells of R. eutropha, and obtained new knowledge with respect to 2HB metabolism in this bacterium.Structural analyses of the 2HB-containing PHAs synthesized by the PhaC AR -equipped strains with 2HB supplementation strongly suggested the occurrence of the block copolymerization phenomenon in R. eutropha.

Bacterial strains and plasmids
The strains and plasmids used in this study are listed in Table 1.E. coli strains DH5α and S17-1 were used as hosts for general genetic engineering and transconjugation, respectively.R. eutropha strains were routinely cultivated at 30 °C in a nutrient-rich (NR) medium containing 1% (w/v) bonito extract (Kyokuto, Tokyo, Japan), 1% (w/v) polypeptone, and 0.2% (w/v) yeast extract dissolved in tap water.E. coli strains were cultivated at 37 °C in a Lysogeny broth (LB) medium.Kanamycin (100 μg/mL for E. coli and 250 μg/mL for R. eutropha strains) or ampicillin (100 μg/mL for E. coli) was added into the medium when necessary.

Plasmid and strain constructions
DNA manipulations were carried out according to standard procedures.PCR reactions were performed with KOD DNA polymerase variants purchased from Toyobo (Osaka, Japan).The sequences of oligonucleotide primers used in this study are shown in supplementary Table S1.Modifications of the chromosomes of R. eutropha were performed by homologous recombination using pK18mobsacB (Schäfer et al. 1994)-based suicide vectors.Generally, a tandem of up-and down-stream homologous regions (~1000 bp-length) flanking to the target locus was cloned into pK18mobsacB, and the resulting plasmid was used for gene deletion.For gene insertion, a DNA fragment of the gene in interest was inserted between the up-and down-stream regions in the vector, followed by homologous recombination by using the construct.The coding region of phaC AR was amplified by using a primer set of phaCar_Fw/phaCar_Rv with pBSP Re phaC AR pct (Matsumoto et al. 2018) as a template.A fragment containing ldh Cd -hadA Cd was amplified by using hadA-inf_ Fw/hadA-inf_Rv and pTTQ-ldhAhadA Cd _opt (Mizuno et al. 2018) as a primer set and template, respectively.A broad-host-range vector pBPP, which has been constructed for gene expression in R. eutropha under the control of phaP1 promoter (Fukui et al. 2011), was used for overexpression of phaC AR .
Transformation of R. eutropha was carried out by transconjugation using E. coli S17-1 harboring the mobilizable plasmid as a donor.In the case of pop-in-pop-out recombination using a pK18mobsacB-based suicide vector, the transformants with the desired genotype were identified by appropriate PCR and isolated, as described previously (Mifune et al. 2010).

PHA production
R. eutropha strains were cultivated in a 100 mL phosphatelimited mineral salt medium composed of 0.2 g of NH 4 Cl, 0.087 g of K 2 HPO 4 , 0.02 g of MgSO 4 •7H 2 O, and 0.1 mL of trace-element solution (Kato et al. 1996) in 100 mL of 40 mM 3-(N-morpholino)propansulfonate (MOPS) buffer (pH 7.2).A filter-sterilized solution of glucose was added to the medium with the final concentration of 2% (w/v) as a carbon source.Other supplements were added into the medium also from the filter-sterilized stock solution with the final concentrations as indicated in the text.Kanamycin (final concentration; 100 μg/mL) was added when necessary.The cells grown at 30 °C for 72 to 120 h with reciprocal shaking (120 strokes/min) were harvested, washed once with cold deionized water, and then lyophilized.The cellular PHA content and composition were determined by gas chromatography (GC) after direct methanolysis of the dried cells in the presence of 15% sulfuric acid as described previously (Kato et al. 1996).Intracellular PHA was extracted from the lyophilized cells by chloroform with stirring at room temperature for 72 h, and then purified by reprecipitation using cold methanol.

H and 13 C NMR analyses and solvent fractionation
The extracted PHA was dissolved in chloroform-d containing 0.05% tetramethylsilane and was subjected to 1 H and 13 C NMR analyses with 400-MR NMR spectrometer (Agilent, California, USA).Solvent fractionation of the PHA copolymer synthesized by recombinant R. eutropha was performed as below.A total of 30-50 mg of the purified PHA was dissolved in 20 mL chloroform with stirring for 72 h at room temperature, then the solution was mixed with 200 mL tetrahydrofuran (THF).The mixture was stirred for 18 h at 4 °C.The THF-soluble fraction was obtained by passing through a polytetrafluoroethylene membrane filter, and the polymer was recovered by reprecipitation with methanol.The residue on the membrane was dried up and collected as the THFinsoluble polymer fraction (Matsumoto et al. 2018).

Introduction of PhaC AR and a pathway for 2HB-CoA formation into R. eutropha
phaC AR , the gene of the chimeric PHA synthase, was inserted into chromosome 1 of R. eutropha strain NSDG-GG∆B1 to obtain the strain IF001 (Fig. 1A).This was achieved by replacing phaC NSDG (encoding N149S/D171G mutant of PHA synthase derived from A. caviae) in NSDG-GG∆B1 with phaC AR by homologous recombination, where the parent strain has been engineered to assimilate glucose (Orita et al. 2012) and glycerol (Fukui et al. 2014) as well as weaken (R)-3HB-CoA formation by deletion of phaB1.It was expected that the deletion of phaB1 potentially led to relative increase in comonomer fraction other than (R)-3HB in PHA (Zhang et al. 2019).
Further modification was aimed to supply (R)-2HB-CoA from 2-ketobutyrate (2KB) generated by deamination of l-threonine in l-isoleucine biosynthesis pathway (Fig. 1B).It has been reported that (R)-2-hydroxyisocaproate dehydrogenase (LdhA) and (R)-2-hydoroxyisocaproate CoA transferase (HadA) derived from Clostridium difficile (currently Clostridioides difficile) (Kim et al. 2006) were functional in the formation of (R)-2HB-CoA from 2KB for biosynthesis of 2HB-containing PHAs by engineered E. coli (Sudo et al. 2020;Mierzati et al. 2020).Here, a tandem of ldhA Cd -hadA Cd , codon-optimized for E. coli, was inserted at downstream of phaP1 encoding a PHA granule-associated protein (phasin) on chromosome 1 of R. eutropha, as shown in Fig. 1A (the strain IF002).As the expression of phaP1 by phaP promoter (P phaP ) was derepressed by mobilization of PhaR regulator from DNA onto surface of PHA granules (Pötter et al. 2002), the heterologous ldhA Cd -hadA Cd genes were expected to be highly expressed during PHA accumulation phase.

Inhibitory effect of 2HB on growth of R. eutropha and restoration by l-valine
PHA biosynthesis properties of the resulting strains IF001 and IF002 were examined in a mineral salt medium containing 2% (w/v) glucose as a sole carbon source.Taking into consideration that nitrogen limitation usually represses amino acid biosynthesis and 2KB is an intermediate in l-isoleucine biosynthesis, phosphate limitation using MOPS medium was applied for induction of PHA synthesis.These strains accumulated PHA containing small fractions of 3-hydroxyvalerate (3HV) unit (~0.1 mol%) with cellular contents of 39-45 wt% of the dry cell weight (entries 1 and 2 in Fig. 2 and supplementary Table S2).Unexpectedly, only trace amount of 2HB unit (< 0.1 mol%) was detected in PHA synthesized by IF002.
We then examined cultivation of the strains in the glucose medium supplemented with 0.25% (w/v) sodium (RS)-2HB as the precursor in order to confirm polymerization activity of PhaC AR towards 2HB-CoA in R. eutropha.However, both the strains showed no growth in the presence of (RS)-2HB.When the strain IF001 was cultivated in the media containing lower concentrations of (R)-2HB, (S)-2HB, or (RS)-2HB (0.005-0.01%(w/v)), the cells also could not grow under all the conditions examined (Fig. 3A).These results demonstrated a strong inhibitory effect of (R/S)-2HB on the cell growth of R. eutropha regardless stereo-configuration of the 2-hydroxyl group.It was assumed that 2KB, formed by dehydrogenation of (R/S)-2HB, gave some effects on metabolisms of branched-chain amino acids and consequent toxicity, because 2KB is an intermediate in biosynthesis of l-isole- ucine from l-threonine (supplementary Fig. S1).We found that, among branched-chain amino acids (l-isoleucine, l-valine, and l-leucine), only l-valine could restore the cell growth in the presence of 0.01% (RS)-2HB (Fig. 3B).Pyruvate, a precursor of l-valine, also restored the growth in the presence of (RS)-2HB when added with high concentration of 0.8% (w/v) (Fig. 3C).Additionally, supplementation of 0.01% 2KB also significantly impaired the growth of IF001 with similar degree to (R/S)-2HB, and the growth restoration by l-valine or pyruvate was again observed (Fig. 3D).These results strongly supported that the growth inhibition by (R/S)-2HB was due to depletion of l-valine caused by inhibitory effects of 2KB generated from (R/S)-2HB on l-valine biosynthesis.The same tendency was observed for IF002 harboring ldhA Cd -hadA Cd (data not shown).

Identification of dehydrogenases responsible for 2HB oxidation in R. eutropha
In the genome of R. eutropha H16, there is no gene highly homologous to previously known 2-hydroxyacid dehydrogenases.We thus considered the possibility that the conversion of (R/S)-2HB to 2KB would be mediated by some homolog(s) of lactate dehydrogenases (LDHs).The probable LDHs identified in R. eutropha H16 are listed in Table 2 and supplementary Table S3, and their gene loci on the chromosomes are shown in supplementary Fig. S2.Individual deletion strains for these probable LDHs were constructed.It should be noted that the two adjacent genes encoding predicted d-LDHs, h16_A1681 (ldhA1) and h16_A1682 (ldhA2), have a completely identical nucleotide sequence to each other, due to duplication of 1676 bp regions of 1,834,348-1,836,023 and 1,836,024-1,837,699 on chromosome 1.The Fig. 2 PHA biosynthesis by engineered strains of R. eutropha on 2% (w/v) glucose (entries 1 to 5), 2% (w/v) glucose with 0.05% (w/v) l-valine (entries 6 to 10), and 2% (w/v) glucose with 0.25% (w/v) sodium (RS)-2HB and 0.05% (w/v) l-valine (entries 11 to 15).The residual cell weight and PHA production are shown in gray and dark gray bars, respectively, and 3HV and 2HB compositions are shown as black and white circles, respectively.The cells were cultivated in a 100 mL phosphate-limited mineral salt medium containing the additives described above for 72-120 h at 30 °C (triplicate) 1 3 actual presence of the duplicated regions (or three repeats in some clones) in IF001 was confirmed by PCR analysis.The knockout of LdhA1 and LdhA2 was conducted by removing both h16_A1681 and h16_A1682 along with the intergenic region.The genes h16_B0093, h16_B0092, and h16_B0091, sharing 44.2%, 26.0%, and 40.1% identities to lutA, lutC, and lutB derived from Bacillus subtilis, respectively, are a homolog set of [Fe-S] cluster protein-dependent l-lactate utilization (LUT) system.It has been reported that the LUT system plays a major role in growth of B. subtilis on l-lactate (Chai et al. 2009).The LUT system-deficient derivative of R. eutropha IF001 was constructed by deletion of the gene cluster of h16_B0093-B0092-B0091.
The resulting IF001-based strains IF005, IF012, IF013, IF014, IF015, IF016, and IF023 were cultivated in the phosphate-limited synthetic media containing d-or l-lactate as a sole-carbon source (Fig. 4A or B), as well as in the glucose media supplemented with (R)-, (S)-, or (RS)-2HB (Fig. 4C, D or E).Among the strains examined, only IF014 lacking H16_A3091 (Dld) showed severely poor growth on d-lactate Fig. 3 Growth inhibition of R. eutropha IF001 by (R/S)-2HB (A) and the restoration by l-valine (B) or pyruvate (C), and growth inhibition of R. eutropha by 2KB and the restoration by l-valine or pyruvate (D).The cells were cultivated in a 3 mL phosphate-limited mineral salt medium containing 2% (w/v) glucose for 72 h at 30 °C with reciprocal shaking (170 strokes/min) (triplicate).Supplementations shown in each panel are 0.005% (w/v) sodium (R)-2HB, 0.005% (w/v) sodium (S)-2HB, 0.01% (w/v) sodium (RS)-2HB, 0.01% (w/v) sodium 2KB, 0.05% (w/v) l-valine, and 0.8% (w/v) sodium pyruvate (Fig. 4A), indicating the critical role of Dld in conversion of d-lactate to pyruvate.All the strains could grow on l-lactate, where IF015 lacking H16_B0093-B0092-B0091 (LutACB) exhibited lag-time in the growth (Fig. 4B).This suggested that LutACB partially participated in utilization of l-lactate although other enzyme(s) would support the growth on l-lactate.As expected, the lower growth ability on d-and l-lactate was related to resistance to (R)-and (S)-2HB, respectively.IF014 and IF015 could grow in the medium containing (R)-2HB and (S)-2HB, respectively, and the strain IF016 doubly lacking Dld and LutACB showed significant growth in the medium containing (RS)-2HB (Fig. 4E).These strains IF014, IF015, and IF016 were still sensitive to 2KB (Fig. 4F), which was consistent with the assumption that 2KB generated from (R/S)-2HB had the actual toxic property to the cells of R. eutropha.

Effects of overexpression of PhaC AR on PHA biosynthesis
The vector for overexpression of phaC AR was constructed by inserting phaC AR into a broad host range-expression vector pBPP at the downstream of P phaP .It has been demonstrated that P phaP in pBPP acted as a strong constitutive promoter when introduced into R. eutropha (Fukui et al. 2011).In the phosphate-limited medium containing 2% (w/v) glucose co-supplemented with 0.25% (w/v) sodium (RS)-2HB and 0.05% (w/v) l-valine, the strain IF017/pBPP-phaC AR accumulated PHA with 21 wt%, and the 2HB composition reached up to 35.0 mol% (Fig. 5 and supplementary Table S6).The molecular weights of the resulting 2HB-containing PHAs were comparable with those of PHAs synthesized by the engineered E. coli harboring phaC AR (Table 3).

Structural analyses of the 2HB-containing PHAs
The PHA extracted from the dried cells of IF017/pBPP-phaC AR (entry 17) was determined to be actually a terpolymer composed by 3HB, 3HV, and 2HB units by 1 H NMR analysis (Fig. 6 and supplementary Fig. S3).It was reported that, when 2HB and 3HA (3HB and 3HV) units were randomly distributed in the polymer chain, the  -40.6 ± 0.3 3.9 1.9 methine proton of 2HB unit showed a complex resonance pattern at 4.8-5.1 ppm based on four triad sequences of 3HA-2HB*-3HA, 3HA-2HB*-2HB, 2HB-2HB*-3HA, and 2HB-2HB*-2HB, and that of 3HB was also resolved at 5.2-5.4ppm due to dyad sequences of 3HB-3HB* and 2HB-3HB* (Matsumoto et al. 2018).The methine proton resonances in the extracted PHA both showed single pattern of 2HB-2HB*-2HB and 3HB-3HB* (Fig. 6A and supplementary Fig. S3A), indicating that the hetero-linkages between 3HB and 2HB units were less than detectable level.The same resonance patterns for methine protons were also observed for PHA synthesized by IF017/ pBBR1MCS-2 not overexpressing phaC AR (supplementary Fig. S3B).The carbonyl carbon resonance in the 13 C NMR spectrum of PHA synthesized by IF017/pBPP-phaC AR was resolved into four peaks arisen from dyad sequences of 3HB and 3HV units.Although the signal corresponding to 3HV-3HV* homo-linkage was too small to calculate dyad-sequence distribution, the rather high signal intensities of the hetero-linkages of 3HB*-3HV and 3HV-3HB* strongly suggested high randomness of the distribution of 3HB and 3HV units (Fig. 6B and supplementary Fig. S4).These results raised two possibilities that the PHAs synthesized by the recombinant R. eutropha were a block copolymer consisting of P(3HB-co-3HV) segment and P(2HB) segment [P(3HB-co-3HV)-block-P(2HB)], or blend of P(3HB-co-3HV) copolymer and P(2HB) homopolymer.We then examined solvent fractionation, because P(2HB) homopolymer in P(3HB)/P(2HB) blend was soluble in THF (Matsumoto et al. 2018).When about 40 mg of the polymer extracted from the cells of IF017/pBPP-phaC AR was treated with THF, a large proportion was THF-insoluble while a very low amount of polymer (< 1 mg) was recovered in the THF-soluble fraction.The 1 H NMR analysis of the THFinsoluble fraction detected the same resonances patterns of methine protons of 2HB, 3HB, and 3HV (Fig. 6C and supplementary Fig. S5A) as those of the polymer before the fractionation (Fig. 6A, supplementary Fig. S3A).It was notable that the methine protons of 3HB and 3HV units were still detected in the THF-soluble fraction along with probable methine resonance of 2HB-2HB*-2HB triad, although signals derived from unknown impurities were observed due to low recovery yield (Fig. 6D and supplementary Fig. S5B).Given that the THF-soluble fraction was not P(2HB) homopolymer, the PHA synthesized by the engineered R. eutropha is most likely a block copolymer consisting of P(3HB-co-3HV) and P(2HB) segments linked with a covalent bond.

Discussion
This study focused on the application of the sequence-regulating chimeric PHA synthase PhaC AR in the well-known PHA producer, R. eutropha.During the investigation, we observed a strong growth inhibitory effect of 2HB to R. eutropha cells (Fig. 3A).The biosynthesis of 2HB-containg PHAs by R. eutropha has been reported by using a mutant of PHA synthase derived from Pseudomonas sp.6-19 (Park et al. 2013), in which PHA synthesis was done by growthunassociated two-step cultivation in a nitrogen-free synthetic medium containing 2HB, and thus there was no description regarding the growth inhibition by 2HB.It was thought that the effect of 2HB on the growth was related to 2HB metabolism specific in R. eutropha, because the growth of recombinant E. coli harboring phaC AR was not significantly impaired by addition of 0.5% (w/v) 2HB (Matsumoto et al. 2018).
The growth inhibition of R. eutropha by 2HB was attributed to l-valine deficiency, and the present results supported the assumption that 2KB was the actual inhibitor.l-Isoleucine and l-valine are biosynthesized via the shared pathway in which the first reaction is catalyzed by acetohydroxyacid synthase (AHAS) (supplementary Fig. S1).l-Isoleucine is the end product when 2KB and pyruvate are condensed by AHAS, while l-valine is synthesized via condensation of two molecules of pyruvate by AHAS.Generally, biosynthesis of branched-chain amino acids is tightly regulated by several factors.In E. coli, there are three AHAS isozyme, AHAS I, II, and III, encoded by ilvBN, ilvGM, and ilvIH, respectively.It was reported that AHAS II and III prefer to 2KB than pyruvate as the substrate, and E. coli K12-derived strains lack AHAS II due to flame shift mutations (Li et al. 2017).R. eutropha possesses a unique AHAS encoded by ilvB (h16_A1035) and ilvH (h16_A1036), which are clustered on chromosome 1 along with ilvC (h16_A1037) encoding acetohydroxyacid isomeroreductase (AHAIR).Lu et al. reported that IlvBH Re showed 140-times higher catalytic selectivity towards 2KB than pyruvate, and was regulated through feedback inhibition by the branched-chain amino acids (Lu et al. 2015).It was plausible that the 2KB-preferring IlvBH Re predominantly mediated the condensation of 2KB and pyruvate over the condensation of two molecules of pyruvate when intracellular concentration of 2KB was increased by oxidation of 2HB supplemented into the medium.This would lead to relatively lower flux of l-valine synthesis than that of l-isoleucine synthesis and consequent deficiency of l-valine.The feedback inhibition of IlvBH Re by l-isoleucine, over-produced in the presence of 2HB, may further promote the l-valine depletion.The growth restoration by high concentration of pyruvate (Fig. 3C and D) might be because the intracellular concentration ratio of 2KB/pyruvate returned to the normal level.
The growth inhibition by the (R/S)-2HB-derived 2KB was based on the presence of endogenous dehydrogenation activity towards 2HB.We here found that Dld (H16_A3091) and LutACB (H16_B0093-B0092-B0091) played a role in assimilation of d-and l-lactate, and moreover, they also related to the cellular sensitivity against (R)-and (S)-2HB, respectively.These results demonstrated the metabolic functions of Dld and LutACB in dehydrogenation of (R)-and (S)-2-hydroxyacids of C 3 -C 4 , although the contribution of LutACB to the conversion of l-lactate to pyruvate was seemed to be partial (Fig. 4B).These are new knowledge regarding the specific properties of R. eutropha for metabolisms of short-chain-length 2-hydroxyacids.
Under the (RS)-2HB-supplemented condition, IF002 (phaC AR + , ldhA Cd -hadA Cd + ) produced PHA composed of 8.5 mol% of 2HB fraction owing to the polymerization ability of PhaC AR to 2HB-CoA (entry 12).Considering that PHA synthesized by IF001 (phaC AR + ) contained no 2HB fraction even with the supplementation of (RS)-2HB (entry 11), R. eutropha does not have endogenous CoA transferase or CoA ligase activity to 2HB, thus the heterologous HadA Cd functioned as the CoA transferase generating 2HB-CoA.This agreed with the previously reported characteristics of an endogenous propionate CoA transferase (H16_A2718) that showed no activity to 2HB despite the broad substrate specificity towards various acids including 3HB, lactate, and glycolate (Volodina et al. 2014).Meanwhile, the lack of Dld resulted in an increase in 2HB composition up to 10.7 mol% (entries 13 and 15), probably due to higher availability of (R)-2HB attributed to the weakened conversion of (R)-2HB to 2KB.The 2HB composition was further increased up to 35 mol% by introduction of pBPP-phaC AR .This was consistent with the previous observation that higher expression of PHA synthase tended to promote incorporation of the minor 3HHx unit in PHA (Dennis et al. 1998;Kawashima et al. 2015).The expression level of PHA synthase was an important factor for compositional regulation of the 2HBcontaing PHA in R. eutropha.
There was no information regarding comonomer distribution and molecular weights of 2HB-containing PHAs synthesized by the previously reported R. eutropha strains expressing the mutant of pseudomonad PHA synthase.We here showed that PHAs synthesized by IF017/pBPP-phaC AR in the presence of (RS)-2HB were most likely a block copolymer consisting of P(3HB-co-3HV) and P(2HB) segments, P[(3HB-co-3HV)-block-2HB]. The weight-average molecular weights were 3.3-6.6× 10 5 , which were comparable with those synthesized by recombinant E. coli expressing PhaC AR .It was notable that PHA with altered 2HB composition synthesized by IF002 having intact dld also showed block sequence (data not shown).These results strongly suggested that the sequence-regulating polymerization was owing to the specific property of the chimeric PHA synthase, but independent from the metabolic background of the host cells.
At present, a very small fraction of 2HB unit was detected in PHA produced by the IF002-based ldhA Cd -hadA Cd + strains when glucose was fed as a sole carbon source (entries 2-5).Considering the fact that strains not harboring ldhA Cd -hadA Cd incorporated no 2HB unit into PHA on glucose in both and E. coli (Sudo et al. 2020) and R. eutropha, LdhA Cd was functional to 2KB to some extent but the participation in the formation of (R)-2HB from 2KB would be insufficient, also as suggested above.Insufficient metabolic flux from l-threonine to 2KB would be another possible cause for the low 2HB fraction in PHA produced from glucose.Further engineering focusing on these points would be useful to establish R. eutropha strains capable of producing the 2HB-containing PHAs efficiently from structurally unrelated carbon sources.

Fig. 6
Fig. 6 Verification of block-copolymerization property of PHAs synthesized by the engineered R. eutropha by solvent fractionation. 1 H NMR (A) and 13 C NMR (B) spectra of the PHAs extracted from IF017/pBPP-phaC AR , and 1 H NMR spectra of the THF-insoluble frac-

Table 3
Molecular weight of PHA synthesized by IF017-based strains of R. eutropha Mw, weight-average molecular weight; PDI, poly dispersity index.a) Estimated from the 2HB composition