Compositional regulation of poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) by replacement of granule-associated protein in Ralstonia eutropha.

BACKGROUND: Phasin (PhaP), a kind of polyhydroxyalkanoate (PHA) granule-associated proteins, has a role in controlling the properties of PHA granules surface, and is thought to have influence on PHA biosynthesis in PHA-producing bacteria. This study focused on the phaP1(Re) locus in Ralstonia eutropha as a site of chromosomal modification for production of flexible poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) [P(3HB-co-3HHx)] from soybean oil.
RESULTS: Considering the high expression level of phaP1(Re), phaJ(Ac) [encoding (R)-specific enoyl-CoA hydratase from Aeromonas caviae] was inserted into the downstream of phaP1(Re) on chromosome 1 of R. eutropha strain NSDG harboring phaC(NSDG) (encoding PHA synthase with broad substrate specificity). The constructed strain efficiently accumulated P(3HB-co-3HHx) on soybean oil with higher 3HHx composition when compared to the previous strain having phaJ(Ac) within pha operon. Insertion of the second phaC(NSDG) along with phaJ(Ac) at the phaP1(Re) locus led to incorporation of much larger 3HHx fraction into PHA chains, although the molecular weight was markedly reduced. The R. eutropha strains were further engineered by replacing phaP1(Re) with phaP(Ac) (encoding phasin from A. caviae) on the chromosome. Interestingly, the phasin replacement increased 3HHx composition in the soybean oil-based PHA with keeping high cellular contents, nevertheless no modification was conducted in the metabolic pathways. Kinetic and Western blot analyses of PHA synthase using cellular insoluble fractions strongly suggested that the phasin replacement not only enhanced activity of PHA synthase from A. caviae but also increased affinity especially to longer (R)-3HHx-CoA. It was supposed that the increased affinity of PHA synthase to (R)-3HHx-CoA was responsible for the higher 3HHx composition in the copolyester.
CONCLUSIONS: The downstream of phaP1(Re) was a useful site for integration of genes to be overexpressed during PHA accumulation in R. eutropha. The results also clarified that polymerization properties of PHA synthase was affected by the kind of phasin co-existed on the surface of PHA granules, leading to altered composition of the resulting P(3HB-co-3HHx). The phasin replacement is a novel engineering strategy for regulation of composition of PHA copolyesters.

Background

Polyhydroxyalkanoates (PHAs) are biopolyesters produced by a number of microorganisms as intracellular carbon and energy storage materials. PHAs have attracted industrial attentions as one of possible solutions for recent environmental problems caused by general petroleum-based plastics, because PHAs are eco-friendly polymeric materials that can be produced from renewable biomass resources, and show biodegradable and biocompatible properties [1, 2].

The most abundant PHA in nature is poly(3-hydroxybutyrate) [P(3HB)], which is generally synthesized from acetyl-CoA through three consecutive reactions by β-ketothiolase (PhaA), NADPH-dependent acetoacetyl-CoA reductase (PhaB) and PHA synthase (PhaC). Ralstonia eutropha (Cupriavidus necator) H16 is a well-studied P(3HB) producer [3, 4]. It has been known that the three genes for P(3HB) biosynthesis are clustered as phaC-A-B1 operon, and PhaC shows polymerization activity toward short-chain-length (R)-3-hydroxyacyl (3HA)-CoAs of C3–C5. However, the application range of P(3HB) is very limited due to the brittle and struggle properties attributed to the high crystallinity and high melting temperature. Poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) [P(3HB-co-3HHx)] is a PHA copolyester naturally produced by some bacteria such as Aeromonas caviae from vegetable oils and fatty acids [5]. This copolyester has been demonstrated to show more soft and flexible properties that are suitable for practical applications when compared to P(3HB) homopolymer [1, 5]. In A. caviae, (R)-3HA-CoA monomers of C4 and C6 are provided from 2-enoyl-CoA intermediates in β-oxidation by the function of (R)-specific enoyl-CoA hydratase (PhaJ), and successively polymerized by PhaC having unique substrate specificity (C4-C7) [6, 7]. Based on these facts, efforts have been made to construct recombinant bacteria for efficient production of P(3HB-co-3HHx). Several previous studies engineered R. eutropha H16 by introduction of heterologous PHA synthase and (R)-specific enoyl-CoA hydratase, resulting in biosynthesis of P(3HB-co-3HHx) composed of adequately high 3HHx fractions from vegetable oils (Fig. 1) [8-12]. Tsuge et al. have reported that the Asn149Ser/Asp171Gly double mutant of A. caviae PHA synthase, named PhaCNSDG, could accept more 3HHx unit than wild type PhaC in vivo [8]. Insomphan et al. reported that deletion of one of fadB homologs in R. eutropha harboring phaCNSDG increased 3HHx fractions in the soybean oil-based P(3HB-co-3HHx) [13]. Recently, an artificial pathway for synthesis of this copolyester from structurally unrelated fructose was developed in R. eutropha [14].

Fig. 1

Proposed P(3HB-co-3HHx) biosynthesis pathway from soybean oil through β-oxidation by R. eutropha NSDG-based recombinant strains. PhaA and BktB β-ketothiolases, PhaB1 NADPH-acetoacetyl-CoA reductase, PhaJ and PhaJ4a (R)-specific enoyl-CoA hydratases, FadB bifunctional (S)-specific 2-enoyl-CoA hydratase/(S)-3-hydroxyacyl-CoA dehydrogenase, PhaC N143S/D171G mutant of PHA synthase from A. caviae, Re R. eutropha, Ac A. caviae

In PHA-accumulating bacterial cells, PHA chains are gathered and make granules in cytoplasm, and the PHA granules are covered with various proteins called PHA granule-associated proteins, PGAPs [15, 16]. Phasins (PhaPs) are amphiphilic small size proteins and known to be major PGAPs widely distributed in various PHA producers [17, 18]. The functions of PhaPs are thought to not only control the properties of PHA granules surface, but also have influence on PHA biosynthesis. PhaP1 is the most abundant phasin in R. eutropha H16 [17]. The expression level is very high owing to the strong phaP1 promoter [19, 20], by which transcription is regulated by PHA granule-binding transcriptional factor PhaR [15, 21]. PhaP1 occupied approximately 5 % of the total proteins in the crude extract of the cells cultivated on fructose [17], and was estimated to cover 27–54 % of surface of the PHA granules [22]. It was demonstrated that PhaP1 participated in control of the size and number of intracellular PHA granules and consequent amount of PHA, as the phaP1-deficient strain of R. eutropha stored approximately half amount of P(3HB) as many smaller granules in comparison with the wild strain [17, 23]. Interestingly, PhaP1 did not bind to PhaC1 directly in two-hybrid assay [24], but formed a high-molecular weight complex along with PhaC1 and soluble P(3HB) oligomer in R. eutropha [25]. The PhaC1-PhaP1-PHB complex showed no lag phase in PhaC activity assay, which suggested that PhaC had an active form in the complex.

Aeromonas caviae also has phasin (PhaP) of which gene is organized as phaP-C-J operon on the chromosome [6]. PhaP belongs to a different class of phasin from PhaP1, as they share no significant homology. When phaPCJ genes were highly expressed in A. caviae, a numerous number of small PHA granules were accumulated within the cells [26]. Moreover, a recombinant strain of A. caviae overexpressing phaPC synthesized P(3HB-co-3HHx) with much higher 3HHx composition (46–52 mol %) than the strain overexpressing phaC alone (16–21 mol %) [26]. Although the detail for this phenomenon has not been elucidated, the result suggested some effects of PhaP on catalytic properties of PhaC. Recent in vitro analyses demonstrated that DNA/PHA-binding protein PhaM activated polymerization activity of PhaC1, and formed a high-molecular-weight complex with PhaC1 [27]. Another enzyme assay indicated that PhaP1 and PhaP inhibited PhaC1 activity, whereas they enhanced the activity of PhaC up to 2.4 to 3-fold [28].

In this study, we focused on the phaP1 locus as a site for chromosomal modification in R. eutropha for production of P(3HB-co-3HHx). The effects of insertion of PHA biosynthesis genes at the downstream of phaP1, as well as those of replacement of phaP on biosynthesis of PHA copolyester from soybean oil were investigated.

Results

Effects of insertion of phaJ/phaCNSDG at downstream of phaP1 on PHA biosynthesis

Previous studies have demonstrated that introduction of PhaJ capable of accepting the C6-substrate was effective to increase 3HHx composition of P(3HB-co-3HHx) produced by R. eutropha from vegetable oils [10-12]. While, the expression of phaP1, regulated by PhaR [15, 21], is highly induced at the PHA accumulation phase and the induced transcription level is one of the highest in R. eutropha H16 [19, 20]. We here examined insertion of phaJ and/or the second copy of phaCNSDG into the downstream of phaP1 on chromosome 1 of R. eutropha strain NSDG possessing phaCNSDG instead of native phaC in pha operon. The gene organization in the constructed strains NSDG-P1ReC, NSDG-P1ReJ, and NSDG-P1ReCJ are illustrated in Fig. 2.

Fig. 2

Schematic diagram of genotypes of R. eutropha NSDG-based recombinant strains. phaA , β-ketothiolase gene; phaB1 , NADPH-acetoacetyl-CoA reductase gene; phaR , a gene encoding PHA-binding transcriptional repressor; phaJ , (R)-specific enoyl-CoA hydratase gene; phaC NSDG, a gene encoding N143S/D171G mutant of PHA synthase from A. caviae; P and P , promoter regions of phaP1 and phaCAB1 in R. eutropha, respectively. Re R. eutropha, Ac A. caviae

These strains were cultivated in a nitrogen-limited synthetic medium containing 1 % (v/v) soybean oil as a sole carbon source, and the results of PHA biosynthesis are shown in Table 1. The insertion of the additional copy of phaCNSDG at the downstream of phaP1 gave almost no influence on production and composition of PHA. NSDG-P1ReJ, constructed by insertion of phaJ, produced similar amount of P(3HB-co-3HHx) with fourfold higher 3HHx composition compared to the parent strain NSDG. This increase of 3HHx fraction was slightly larger than that observed by the strain MF02 harboring phaJ within pha operon. Enzyme assay indicated that enoyl-CoA hydratase activities toward crotonyl-CoA (C4) and 2-hexenoyl-CoA (C6) in NSDG-P1ReJ (C4: 27.2 U/mg, C6: 12.0 U/mg) were approximately 3.0- and 2.3-fold higher than those in MF02 (C4: 9.1 U/mg, C6: 5.2 U/mg).

Table 1

PHA biosynthesis by R. eutropha recombinant strains from soybean oil

Strain	Dry cell mass (g/L)	PHA content (wt%)	PHA (g/L)	3HHx composition (mol %)
(phaP1 Re strains)
NSDG	6.9 ± 0.02	86 ± 4.3	5.9	1.6 ± 0.03
NSDG-P1ReC	6.7 ± 0.21	91 ± 0.3	6.0	1.8 ± 0.10
NSDG-P1ReJ	6.7 ± 0.05	89 ± 1.5	6.0	6.5 ± 0.09
NSDG-P1ReCJ	5.7 ± 0.18	84 ± 1.3	4.8	10.5 ± 0.26
MF02	6.8 ± 0.02	88 ± 0.51	6.0	4.6 ± 0.07
(phaP Ac strains)
NSDG-PAc	7.1 ± 0.18	89 ± 1.1	6.3	2.0 ± 0.05
NSDG-PAcC	6.4 ± 0.11	88 ± 1.5	5.6	2.1 ± 0.51
NSDG-PAcJ	7.0 ± 0.40	87 ± 1.1	6.1	8.8 ± 0.24
NSDG-PAcCJ	6.3 ± 0.07	79 ± 3.3	5.0	17.2 ± 0.18
MF02-PAc	7.2 ± 0.30	90 ± 1.2	6.5	6.4 ± 0.10

The cells were cultivated in an MB medium containing 1 % (v/v) soybean oil at 30 °C for 72 h (n = 3)

When phaCNSDG-phaJ genes were inserted at the site, the resulting strain NSDG-P1ReCJ accumulated P(3HB-co-3HHx) with 3HHx fraction of 10.5 mol % (Table 1). This was 6.6-fold larger than that in the copolymer synthesized by the strain NSDG, although the PHA production was decreased from 5.9 to 4.8 g/l.

Effects of replacement of phaP1 by phaP on PHA biosynthesis

phaP1 on chromosome 1 was replaced by phaP from A. caviae in the recombinant strains of R. eutropha expressing phaCNSDG (Fig. 2), in order to investigate the change of PHA biosynthesis profiles by phasin derived from the same source as PHA synthase. Interestingly, this genetic modification tended to increase 3HHx composition without serious reduction of PHA production from soybean oil in all the strains examined, as shown in Table 1. The compositional change was noticeable in the strains expressing phaJ and the additional phaCNSDG at the downstream of phaP, as the 3HHx fraction in P(3HB-co-3HHx) significantly increased to 17.2 mol % in NSDG-PAcCJ compared to 10.5 mol % in NSDG-P1ReCJ. The 3HHx compositions of the copolymers synthesized by NSDG-PAcJ and MF02-PAc (6.4–8.8 mol %) were also higher than those in the corresponding strains having native PhaP1 (4.6–6.5 mol %). The effect of the phasin replacement was not remarkable in the strains not expressing phaJ, such as NSDG and NSDG-P1C.

Molecular weights of PHAs accumulated in R. eutropha recombinant strains

The number-average-molecular weights (Mn) of P(3HB-co-3HHx) synthesized by the phaP-replaced strains NSDG-PAc and NSDG-PAcJ were 1.8- and 1.4-fold higher than those by the corresponding parent strains NSDG and NSDG-P1ReJ, respectively (Table 2). The results suggested that PhaP allowed PhaCNSDG co-existed on the granule surface to synthesize longer PHA chain than PhaP1. The strains NSDG-P1ReCJ and NSDG-PAcCJ accumulated PHA with drastically decreased Mn to approximately one-ninth of those obtained by the strains not having the second copy of phaCNSDG. The values of polydispersity index of the resulting PHAs were not changed by the replacement of phasin.

Table 2

Molecular weights of PHA synthesized by R. eutropha recombinant strains from soybean oil

Strain	PhaP	M n × 105	M w × 105	PDI
NSDG	PhaP1Re	7.5 ± 3.1	24.4 ± 5.23	3.56 ± 0.80
NSDG-PAc	PhaPAc	13.5 ± 2.3	42.3 ± 5.93	3.16 ± 0.26
NSDG-P1ReJ	PhaP1Re	10.0 ± 2.1	33.8 ± 6.05	3.52 ± 0.90
NSDG-PAcJ	PhaPAc	13.8 ± 1.2	48.1 ± 2.16	3.52 ± 0.25
NSDGP1ReCJ	PhaP1Re	1.2 ± 0.17	2.8 ± 0.25	2.28 ± 0.21
NSDG-PAcCJ	PhaPAc	1.5 ± 0.36	3.6 ± 1.39	2.35 ± 0.32

PHA was extracted and purified from the cells cultivated in an MB medium containing 1 % (v/v) soybean oil at 30 °C for 72 h (n = 3). M number-average molecular weight, M weight-average molecular weight, PDI polydispersity index (M w/M n)

Kinetic parameters of PHA synthase on the surface of PHA granules

PHA synthase and phasin were bound on the surface of PHA granules contained in a cellular insoluble fraction. The kinetic analysis of PhaCNSDG on PHA granules was done using the insoluble fraction of R. eutropha to investigate the effects of the phasin replacement on the catalytic properties of PhaCNSDG. We initially attempted the analysis for the strains NSDG and NSDG-PAc grown on soybean oil, however, PHA synthase activity of the insoluble fractions was too low to determine reliable parameters. Therefore, the strains NSDG-P1ReCJ and NSDG-PAcCJ, both harboring the second copy of phaCNSDG, were used for the sources of the insoluble fraction. When the polymerization activities of PhaCNSDG were determined toward (R)-3HB-CoA (C4) and (R)-3HHx-CoA (C6), no lag-phase of the absorbance change was observed. The reactions followed Michaelis–Menten kinetics, and the determined kinetic parameters are shown in Table 3. PHA synthases in both the strains showed lower activity to (R)-3HHx-CoA than to (R)-3HB-CoA. Interestingly, the replacement of PhaP1 by PhaP led to 2.6 to 2.8-fold increase in the Vmax values, and significantly reduced the Km values to one-fourth to one-tenth. The C6/C4 ratios of Vmax were almost the same as each other for the two synthase samples, while the C6/C4 ratio of Km was markedly decreased from 1.9 with PhaP1 (NSDG-P1ReCJ) to 0.71 with PhaP (NSDG-PAcCJ). These results indicated that the increase of affinity of PhaCNSDG by the phasin replacement was more significant to the longer C6-substrate. Consequently, the C6/C4 ratio of Vmax/Km of the synthase in the phaP-expressing strain was 3.0-fold larger than that in the strain expressing native phaP1, which was consistent with the increase of the 3HHx fraction in the resulting PHA. The amounts of the PhaCNSDG protein in the insoluble fractions of NSDG-P1ReCJ and NSDG-PAcCJ, determined by Western blot analysis, were similar level as shown in Fig. 3a, indicating that the differences in the kinetic parameters were not due to changes in the amount of PHA synthase on the granule surface. We also observed that (R)-specific enoyl-CoA hydratase activity to the C4- and C6-substrates were not different between the soluble fractions of the two strains (C4: 14.3–15.3 U/mg and C6: 1.1–1.3 U/mg, respectively). The similar amounts of PhaCNSDG and activities of PhaJ in the two strains indicated little influence of the replacement of phaP on transcription driven by the upstream promoter region.

Table 3

Kinetic parameters of PhaCNSDG in insoluble fractions of R. eturopha recombinant strains toward (R)-3HB-CoA (C4) and (R)-3HHx-CoA (C6)

Strain (genotype)	Substrate chain length	Km		Vmax		Vmax/Km
		(µM)	C6/C4 ratio	(µmol/min/mg-insolublea)	C6/C4 ratio		C6/C4 ratio
NSDG-P1ReCJ	C4	320	1.9	0.21	0.20	6.4 × 10−4	0.10
(phaP1 Re-phaC NSDG-phaJ Ac)	C6	614		0.041		6.6 × 10−5
NSDG-PAcCJ	C4	86	0.71	0.52	0.21	6.1 × 10−3	0.30
(phaP Ac-phaC NSDG-phaJ Ac)	C6	61		0.11		1.8 × 10−3

The insoluble fractions were prepared from the cells cultivated in an MB medium containing 1 % (v/v) soybean oil at 30 °C for 48 h

aDry weight of the insoluble fraction

Fig. 3

a Western blot analysis of insoluble fractions (12.5 µg) of R. eutropha strains using anti-PhaC antiserum. b SDS-PAGE analysis of PHA granule fractions (700 µg of dry granules). Lanes: M protein marker; 1, R. eutropha strain NSDG-P1ReCJ; 2, R. eutropha strain NSDG-PAcCJ. The boxed bands (no. 1–8) in b were subjected to protein identification by LC–MS/MS

Confirmation of phasin replacement and identification of proteins in PHA granule fractions

The proteins in PHA granule fractions of the strains NSDG-P1ReCJ and NSDG-PAcCJ, prepared by density-gradient ultracentrifugation, were separated by SDS-PAGE (Fig. 3b), and the major eight protein bands were subjected to LC–MS/MS analysis for identification. The details of the results are shown in Additional file 1: Table S1. The 24-kDa protein (no. 5) in NSDG-P1ReCJ was determined to be PhaP1. In the strain NSDG-PAcCJ, this band was disappeared and a smaller protein (no. 8) in large abundance, not present in NSDG-P1ReCJ, was identified to be PhaC. This result clearly indicated actual replacement of major phasin on PHA granules to PhaP in NSDG-PAcCJ. The protein band no. 3 was outer membrane protein (porin), and the bands of no. 1, 2, 4, 6, and 7 contained multiple kinds of proteins including various membrane proteins, ribosomal proteins, and hypothetical proteins. Several proteins involved in important cellular functions were also detected, such as elongation factor Tu and cell division protein FtsA in the no. 1 band and RNA polymerase subunit α in the no. 2 band. These proteins were considered to be the result of contamination of other cell fractions, but possibility of the presence on the surface as PGAPS could not be excluded, as observed and discussed previously [16, 31]. Although R. eutropha possesses many phasins [3, 24, 29, 30], only PhaP5 could be detected in the band no. 7 besides PhaP1 in NSDG-P1ReCJ.

Electron microscopy

The accumulation of PHA granules within R. eutropha H16, NSDG, NSDG-PAc, NSDG-P1ReCJ, and NSDG-PAcCJ grown on soybean oil were observed by transmission electron microscopy (Fig. 4). The wild strain H16 accumulated 10–16 of small PHA granules per one cell, whereas NSDG and NSDG-PAc formed one to three of much larger PHA granule(s). In the cells of NSDG-P1ReCJ and NSDG-PAcCJ, granules with various sizes were existed together. The size and number of PHA granules were greatly affected by the kind of PHA synthase, but not by phasin on the granule surface. The reason for these observations has been unclear yet, but the formation of less larger granules in the strain NSDG was similar to phenotype of the phaP1-disruptant [17]. The hydrophilicity of cytosolic side of the PhaCNSDG protein on PHA granules may be lower than that of PhaC, leading to formation of large granules with reduced surface area.

Fig. 4

Transmission electron micrographs of ultrathin sections of R. eutropha strains. Magnification: ×10,000; scale bar 0.5 μm

Discussion

The gene of the major phasin in R. eutropha H16, phaP1, is one of the most highly expressed genes at the PHA accumulation phase [19, 20]. Barnard et al. have achieved T7 promoter-driven high level production of recombinant proteins by using the phaP1 locus as the integration site for T7 RNA polymerase [32]. We here applied the downstream of phaP1 as an insertion site for PHA biosynthesis-related gene(s) in R. eutropha, because the induction profile and expression level of phaP1 higher than phaCAB1 were expected to be advantage for PHA biosynthesis. When phaCNSDG was solely inserted at the site in the strain NSDG, no change in P(3HB-co-3HHx) composition was observed, probably due to poor provision of (R)-3HHx-CoA through the intact channeling pathway from β-oxidation to PHA biosynthesis. Indeed, enhancement of the channeling pathway by expression of phaJ significantly increased 3HHx composition without impairment of PHA production, as observed previously [10, 12]. The results of the cultivation suggested that the expression level of the gene located in the phaP1 locus was higher than within pha operon, given the higher 3HHx composition in the resulting P(3HB-co-3HHx) and higher enoyl-CoA hydratase activity in NSDG-P1ReJ (phaP1-phaJ) than those in the strain MF02 (phaCNSDG-phaJ-phaAB1). The insertion of the second copy of phaCNSDG at the downstream of phaP1 led to drastic decrease of the molecular weight of PHA by less than one-tenth. Previous studies [33-35] reported a negative correlation of intracellular PhaC activity or phaC expression level with P(3HB) molecular weight, and one of possible reasons for this phenomenon may be due to large number of catalytic molecules (PhaCNSDG) relative to substrate molecules [(R)-3HA-CoAs]. The present study was consistent with the previous observations, and the similar effects of plasmid-borne multi copies of phaC on P(3HB-co-3HHx) biosynthesis has been also seen in recombinant R. eutropha grown on plant oils [11].

This study also demonstrated a novel engineering strategy for compositional regulation of P(3HB-co-3HHx); the phasin replacement. We previously observed that overexpression of phaC together with phaP in A. caviae significantly increased the 3HHx fraction in PHA produced from fatty acids when compared to overexpression of phaC alone [26]. One of possible explanation for this phenomenon was some changes of catalytic properties of PhaC in the presence of PhaP derived from the same source. However, such compositional change had not been observed when phaPC were introduced in R. eutropha PHB−4 [6]. We assumed that, even if PhaP affected catalytic properties of PhaC, the effects might be disturbed by the presence of large amount of native phasins covering the surface of PHA granules in R. eutropha. The phaP1 encoding major phasin was thus replaced by phaP by homologous recombination, and the results demonstrated that the phasin replacement tended to increase 3HHx fraction in P(3HB-co-3HHx) with keeping high PHA contents on soybean oil. Further kinetic analysis of PhaCNSDG using native granules clarified interesting changes of the catalytic properties depending on the origin of phasin. The replacement of PhaP1 by PhaP significantly increased the catalytic efficiency of PhaCNSDG that had been already active form on the PHA granule surface. It has been reported that both PhaP and PhaP1 enhanced the polymerization activity of PhaC toward (R)-3HB-CoA in in vitro assay using the free proteins not associated with PHA granules, where the activation by PhaP was slightly stronger than that by PhaP1 [28]. It should be noted that the enhancement of polymerization activity by the replacement of PhaP1 with PhaP on the native granules (2.5 to 2.7-fold) was much larger than that observed in the case using the free proteins (1.3-fold). In addition, the phasin replacement more specifically increased the affinity of PhaCNSDG toward the longer (R)-3HHx-CoA, as seen in the increase of C6/C4 ratio of the Km values (Table 3). It was feasible that the higher relative affinity of the polymerizing enzyme to the C6-substrate was responsible for the larger 3HHx fraction in PHA synthesized by the strain NSDG-PAcCJ. Ushimaru et al. proposed that phasin helps to withdraw a PHA chain from PhaC exhibiting lower turnover rate than PhaC1, and thus prevent the chain from aggregation that potentially blocks the release site in the active dimer [28]. This idea can well explain the activation of PhaC by phasins, but not enough to understand the great increase of the substrate affinity of PhaCNSDG, particularly to the C6-substrate, by the phasin replacement. Although the detailed mechanism has been remained to be cleared, it was supposed that phasin could interact with active PHA synthase on the granule surface. In the case of specific combination of PHA synthase and phasin such as PhaCNSDG and PhaP, the interaction might induce some structural change potentially affecting the affinity and specificity toward the substrates into PHA synthase (Fig. 5). The molecular weights of the accumulated PHA also tended to be increased by the phasin replacement. The higher activity of PhaCNSDG in the presence of PhaP on the granule surface would lead formation of longer polymer chains than the enzyme with PhaP1.

Fig. 5

Estimated situations of the surface of PHA granule in R. eutropha strains expressing phaC NSDG together with phaP1 (left) or phaP (right)

The expression of phaJ was very effective for increasing 3HHx composition in P(3HB-co-3HHx) synthesized from fatty acids and plant oils by recombinant R. eutropha strains [10-12]. However, we previously observed saturation of the provision of 3HHx unit by this strategy [12]. Another strategy for increasing 3HHx composition was suppression of 3HB unit formation by disruption of phaA encoding thiolase, although this accompanied reduction of PHA production [10]. The advantage of the present phasin replacement strategy was further increase of 3HHx composition even in the phaJ-expressing strains without impairment of PHA biosynthesis. The R. eutropha strains applying this novel strategy can efficienty produce P(3HB-co-3HHx) composed of 9–17 mol % 3HHx fractions, which was suitable for practical applications with adequate flexibility [5].

Conclusion

This study demonstrated that the downstream of phaP1 on chromosome 1 of R. eutropha was a useful site for integration of PHA biosynthesis genes to achieve the overexpression during the PHA accumulation phase. In addition, we found that replacement of the major phasin PhaP1 in recombinant R. eutropha by PhaP from A. caviae resulted in increase of 3HHx composition of P(3HB-co-3HHx) without serious decrease of the productivity on soybean oil.

A general engineering strategy for alteration of PHA copolymer composition has been modification of metabolic flux of monomer-supplying pathways. The phasin replacement is a novel strategy not directly modifying metabolic pathways, thus there is no need to consider unexpected negative effects on global metabolisms. It was also shown that catalytic properties of PHA synthase were affected by phasin co-existed on the surface of PHA granules, and significance of the effects appeared to be depended on the kind of phasin. These results were interesting in understanding the role of phasin in PHA biosynthesis. In terms of the application, the use of adequate pair of PHA synthase and phasin would allow us to obtain PHA copolymers with desired composition with high productivity.

Methods

Bacterial strains and plasmids

The bacterial strains and plasmids used in this study are listed in Additional file 1: Table S2. R. eutropha strains were cultivated at 30 °C in a nutrient-rich (NR) medium containing 10 g of meat extract, 10 g of polypeptone, 2 g of yeast extract in 1 L of tap water. E. coli strains were cultivated at 37 °C in a Lysogeny broth (LB) medium. Kanamycin (100 mg/L) or ampicillin (50 mg/L) was added to the medium when necessary.

Construction of plasmids and strains

DNA manipulations were carried out according to standard procedures, and the sequences of oligonucleotide primers used for PCR are shown in Additional file 1: Table S3. KOD-Plus-ver.2 or KOD-Plus-Neo DNA polymerase and Ligation High ver.2 (Toyobo, Otsu, Japan) were used for PCR amplification and ligation, respectively.

The upstream and downstream regions flanked to phaP1 were individually amplified by PCR from R. eutropha H16 genomic DNA as a template with primer sets of phaPReout-Fw/phaPRe-Inv1 and phaPReout-Rv/phaPRe-Inv2, respectively. The latter fragment was 5′-phosphorylated, ligated with the former, and the upstream–downstream fragment was subjected to the second PCR using phaPReout-Fw/phaPReout-Rv primers. The amplified upstream–downstream fusion was inserted into pK18mobsacB at the SmaI site, leading to construction of pK18mobsacB-ΔP1.

The vectors for replacement of phaP1 by phaP, phaP-phaCNSDG, phaP-phaCNSDG-phaJ, or phaPJ were constructed as follows. The coding region of phaCNSDG was amplified from pTA2-NSDG [10] by PCR using a primer set of pEE32R-CAcup-inv/pEE32R-CAcdown-inv. A linear fragment not containing phaC in pEE32 harboring phaPCJ [6] was prepared by inverse PCR with phaPRe-Inv1/phaPRe-Inv2 primers, and then ligated with the phaCNSDG fragment to obtain pEE32-NSDG harboring phaP-CNSDG-J. The fragments of phaP, phaP-CNSDG, and phaP-CNSDG-J were amplified with pEE32-NSDG as a template and primer sets of phaPAc-Fw/phaPAc-Rv, phaPAc-Fw/phaCNSDG-Rv, and phaPAc-Fw/phaJAc-Rv, respectively. The amplified fragments were 5′-phosphorylated and then individually inserted between upstream and downstream regions of phaP1 by ligation with an inverse PCR product obtained with pK18mobsacB-ΔP1 as a template and phaPRe-Inv1/phaPRe-Inv2 primers. The resulting plasmids were designated pK18mobsacB-PAc, pK18mobsacB-PAcC, and pK18mobsacB-PAcCJ, respectively. pK18mobsacB-PAcJ was constructed by inverse PCR for elimination of the phaCNSDG region from pK18mobsacB-PAcCJ with a primer set phaPAcdown-Inv/phaJAc-Fw, followed by self-ligation after 5′-phosphorylation.

pK18mobsacB-P1ReC, pK18mobsacB-P1ReJ and pK18mobsacB-P1ReCJ, used for insertion of the genes from A. caviae at downstream of phaP1, were constructed from pK18mobsacB-PAcC, pK18mobsacB-PAcJ and pK18mobsacB-PAcCJ, respectively. Linear fragments not having the phaP region were prepared from these plasmids by inverse PCR using phaPRe-Inv1/phaPAc-down-Inv primers. The coding region of phaP1 was amplified with R. eutropha H16 genome DNA and a primer set of phaPRe-Fw/phaPRe-Rv, and the phaP1 fragment was ligated with the corresponding inverse PCR product with the same direction as the phaP1 promoter.

All the pK18mobsacB-based plasmids were transferred from E. coli S17-1 by conjugation to R. eutropha strains NSDG or MF02, and the recombinant strains formed by pop-in-pop-out recombination were selected, according to the procedures reported previously [10].

PHA production by recombinant R. eutropha strains

PHA production by R. eutropha strains was carried out on a reciprocal shaker (115 strokes/min) at 30 °C in a 500 ml flask with a 100 ml of a nitrogen-limited mineral salts (MB) medium and 0.1 ml of trace-element solution [36]. Soybean oil was directly added to the medium at 1.0 % (v/v) as a sole carbon source. After the cultivation for 72 h, the cells were harvested, washed with cold 70 % ethanol, washed again with deionized water, and then lyophilized. The cellular PHA content and monomer composition were determined by gas chromatography (GC) after methanolysis of the dried cells in the preference of 15 % sulfuric acid [35]. Extraction and purification of the accumulated PHA from the dried cells, and determination of molecular weight by gel permutation chromatography (GPC) were performed as described previously [37].

Enzyme assay

Crotonyl-CoA was synthesized by condensation of crotonic anhydride (Tokyo Chemical Industry, Tokyo, Japan) with lithium salt of CoA-SH [38]. Chemical synthesis of (R)-3HB-CoA and trans-2-hexenoyl-CoA by a mixed-anhydride method, and enzymatic conversion of trans-2-hexenoyl-CoA to (R)-3HHx-CoA were performed as described previously [39], except for the use of recombinant PhaJ4a [12] as (R)-specific enoyl-CoA hydratase.

Ralstonia eutropha strains were cultivated for 48 h in an MB medium at 30 °C with 1.0 % (v/v) soybean oil as a carbon source. The grown cells were harvested, washed by Tris–HCl buffer (pH 7.5), and then disrupted by high pressure homogenization as described previously [36]. The disrupted cells were centrifuged (18,800g, 10 min, 4 °C) to separate the soluble fraction and insoluble fraction containing native PHA granules. PHA synthase activity of the insoluble fraction was assayed with the reaction mixture composed 0.05–0.5 mM (R)-3HB-CoA or (R)-3HHx-CoA, 1 mM 5,5′-dithiobis (2-nitrobenzoic acid) [DTNB] in 50 mM phosphate buffer (pH 7.2), and 10 µl of the insoluble fraction. The increase in absorbance at 412 nm corresponding to release of free CoA-SH was measured spectrophotometrically at 30 °C (ε412 = 14.5 × 103). Enoyl-CoA hydratase activity, including both (S)- and (R)-specific activities, in the soluble fraction was determined in the reaction mixture composed of 250 µM crotonyl-CoA or trans-2-hexenoyl-CoA in 50 mM Tris–HCl (pH 8.0) and 5 µl of the soluble fraction. The hydration of the enoyl-CoA substrates were monitored as decrease in absorbance at 263 nm at 30 °C (ε263 = 6.7 × 103 M−1 cm−1). (R)-specific enoyl-CoA hydratase activity in the soluble fraction was determined by the hydration of the enoyl-CoA substrates coupled with polymerization of the resulting (R)-3-hydroxyalyl-CoAs of C4 and C6 by using the insoluble fraction of R. eutropha NSDG-PAcC as a source of PHA synthase. The reaction mixture was composed of 0.20 mM trans-2-enoyl-CoA, 1 mM DTNB in 50 mM phosphate buffer (pH 8.0), 5 µl of the insoluble fraction of the strain NSDG-PAcC, and the increase of absorbance at 412 nm was monitored at 30 °C.

The aliquot of the insoluble fraction was lyophilized to determine the dry weight, and used to calculate specific activities. The proteins contained in 12.5 µg of the insoluble fraction were separated by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) with 15 % acrylamide gel, according to the standard procedure. Western blot analysis for PhaCNSDG was performed using a specific antiserum against PhaC [26]. A goat anti-rabbit IgG (Fc fragment specific)-alkaline phosphatase conjugate (Calbiochem, CA, USA) was used as the secondary antibody, and 1-Step NBT/BCIP plus suppressor (Thermo Fisher Scientific, MA, USA) was used to detect the signals according to the manufacture’s instruction. Protein concentration of the soluble fraction was determined by the method of Bradford with bovine serum albumin as the standard.

Transmission electron microscope

The cells of the R. eutropha strains cultivated in an MB medium with 1 % (v/v) soybean oil at 30 °C for 48 h were harvested, washed, and resuspended in 0.1 M phosphate buffer (pH 7.2) containing 0.1 M sucrose. The cells were fixed with 2.5 % (w/v) glutaraldehyde in 0.1 M phosphate buffer (pH 7.2) containing 0.1 M sucrose, and subsequently with 1 % (w/v) osmium tetroxide in the buffer. The fixed cells were dehydrated in a graded ethanol and embedded with epoxy resin Quetol 651 (Nissin EM, Tokyo, Japan). The ultrathin sections (80 nm) of the embedded cells were prepared with a Leica UC7 ultramicrotome. The sections on a copper grid were stained with EM stainer (Nissin EM) followed by Reynolds’s lead citrate at room temperature. The specimens were observed on a JEOL 1400Plus transmission electron microscope with an accelerating voltage of 100 kV.

Isolation of native PHA granule fractions and identification of proteins

The insoluble fractions of R. eutropha strains were loaded on the top of a discontinuous glycerol gradient of 2 mL of 88 % glycerol and 8 mL of 44 % glycerol in 100 mM Tris–HCl (pH 7.5). After the ultracentrifugation (210,000g, 40 min, 4 °C), the PHA granules at the interphase of 44–88 % glycerol were collected, following by washing with 100 mM Tris–HCl (pH 7.5). The granules resuspended in 1 mL of the buffer were subsequently loaded on a discontinuous sucrose gradient prepared with 1 mL of 1.8 M sucrose, 2.9 mL of 1.6, 1.4, and 1.2 M sucrose, and 1 mL of 1.0 M sucrose in the buffer, and then subjected to ultracentrifugation (210,000g, 2 h, 4 °C). The band of the PHA granules were observed at the interphase of 1.2–1.4 M sucrose for P(3HB-co-3HHx) composed of 10.5 mol % 3HHx (produced by the strain NSDG-P1ReCJ), and that of 1.0–1.2 M sucrose for P(3HB-co-3HHx) composed of 17.2 mol % 3HHx (produced by the strain NSDG-PAcCJ). The bands of PHA granules were recovered, washed, and resuspended in 100 mM Tris–HCl (pH 7.5).

The PHA granule suspension containing approximately 700 μg of dried inclusion was separated by SDS-PAGE with 4–15 % gradient gel (Bio-Rad, Hercules, CA, USA). The tryptic peptides were prepared from the SDS-PAGE gel, and subjected to mass spectrometric analysis using an LC–MS system equipped with Acquity UPLC apparatus (Waters, Milford, MA, USA) and Synapt High Definition Mass Spectrometer (Waters). LC was performed with a reversed-phase C18 column (75 µm × 150 mm, 1.7 µm particle size) with the mobile phase of acetonitrile gradient in 1 % formic acid at the flow rate of 0.5 μl/min. The mass spectrometry was done in the positive ion mode. The MS/MS spectra were analyzed by Protein Lynx Global Server (Waters).