Co-expression of polyhydroxyalkanoate synthase and (R)-enoyl-CoA hydratase genes of Aeromonas caviae establishes copolyester biosynthesis pathway in Escherichia coli


  • Toshiaki Fukui,

    1. Polymer Chemistry Laboratory and the RIKEN Group of the Research Institute of Innovative Technology for the Earth (RITE), The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama 351-0198, Japan
    Search for more papers by this author
  • Satoru Yokomizo,

    1. Polymer Chemistry Laboratory and the RIKEN Group of the Research Institute of Innovative Technology for the Earth (RITE), The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama 351-0198, Japan
    Search for more papers by this author
  • Genta Kobayashi,

    1. Polymer Chemistry Laboratory and the RIKEN Group of the Research Institute of Innovative Technology for the Earth (RITE), The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama 351-0198, Japan
    Search for more papers by this author
  • Yoshiharu Doi

    Corresponding author
    1. Polymer Chemistry Laboratory and the RIKEN Group of the Research Institute of Innovative Technology for the Earth (RITE), The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama 351-0198, Japan
    Search for more papers by this author

*Corresponding author. Tel: +81 (48) 467-9402; Fax: +81 (48) 462-4667; E-mail:


Polyhydroxyalkanoate biosynthesis genes of Aeromonas caviae were expressed in Escherichia coli LS5218 (fadR atoC(Con)), and the polyhydroxyalkanoate-producing ability of the recombinants was investigated. A LS5218 strain harboring only phaCAc (polyhydroxyalkanoate synthase gene) did not accumulate any polyhydroxyalkanoate from dodecanoate in spite of the existence of translated polyhydroxyalkanoate synthase protein, whereas co-expression phaCAc and phaJAc ((R)-specific enoyl-CoA hydratase gene) resulted in the accumulation of P(3-hydroxybutyrate-co-3-hydroxyhexanoate) copolymer up to 7–11 wt% of dry cell weight from octanoate and dodecanoate. These results indicated that both phaCAc and phaJAc are essential for E. coli LS5218 to establish the polyhydroxyalkanoate biosynthesis pathway from alkanoic acids. The copolyester content in the strain expressing both the genes under the lac promoter control reached to 38 wt% from dodecanoate. Enzyme assays suggest that efficient monomer formation via β-oxidation by a high level expression of phaJAc was important to achieve a high polyhydroxyalkanoate content in the recombinant E. coli.


A variety of microorganisms is known to accumulate polyhydroxyalkanoates (PHA) as an intracellular carbon and energy storage material, and much research has been focused on the production of PHA in bacteria due to its potential use as a biodegradable thermoplastic [1, 2].

Escherichia coli is unable to synthesize any PHA of a high molecular mass; however, the biosynthesis pathway of poly(3-hydroxybutyrate) (P(3HB)) has been established by several laboratories in recombinant E. coli cells which expressed the Ralstonia eutropha PHA biosynthesis operon consisting of the genes of PHA synthase, β-ketothiolase, and NADPH-acetoacetyl-CoA reductase [3–6]. The efficient production of P(3HB) from glucose could be achieved by employing the filamentation-suppressed recombinant E. coli in a defined medium [7]. The biosynthesis of poly(3-hydroxybutyrate-co-3-hydroxyvalerate) copolyester from glucose and propionic acid was also performed in the E. coli fadR atoC(Con) mutant harboring the R. eutropha PHA genes [8]. Recently, recombinant E. coli expressing the PHA synthase gene of Pseudomonas aeruginosa[9] or P. oleovorans[10] was demonstrated to produce PHA consisting of medium-chain-length 3-hydroxyalkanoates (C6–C12) from alkanoic acids as a carbon source. In both studies, the fadB mutant of E. coli should be applied as a host to achieve the accumulation of quantity amounts of PHA in the recombinant cells.

Aeromonas caviae was isolated as a producer of a random copolymer of 3-hydroxybutyrate and 3-hydroxyhexanoate (P(3HB-co-3HHx)) from alkanoic acids and oils [11]. We have cloned and analyzed the PHA biosynthesis genes of A. caviae, including structural genes of PHA synthase (phaCAc) and (R)-specific enoyl-CoA hydratase (phaJAc) [12], and have proved that phaJAc is an essential gene for PHA biosynthesis from alkanoic acids in A. caviae[13]. In this study, we have investigated whether expression of the PHA biosynthesis genes of A. caviae confers the ability for P(3HB-co-3HHx) biosynthesis to E. coli. via fatty acid β-oxidation cycles.

2Materials and methods

2.1Bacterial strains, plasmids, genetic techniques, and culture conditions

E. coli LS5218 (fadR601, atoC2(Con)) [14], and plasmids pUC18, pK223-3 (Pharmacia), and pJRD215 [15] were used for the heterologous expression of PHA biosynthesis genes of A. caviae. Basic DNA recombinant techniques were carried out according to the standard procedures [16]. Site-directed mutagenesis was done using the unique site elimination procedure with a U.S.E. mutagenesis kit (Pharmacia). The mutagenic primers M1, M2, M3, M4 [12] and M5 [13] were used for the construction of pKBgB18 (M2 and M3), pJRDG1 (M2 and M4), and pJRDG3 (M1 and M5). PCR was performed to amplify the coding region of phaCAc with primers 5′-AGCATATGAGCCAACCATCTTATGGCCC-3′ for N-terminus and 5′-CCAGGGATCCTGCGCTCATGCGGCGTCC-3′ for C-terminus (underlined sequences show NdeI and BamHI sites, respectively). Cells were cultivated in a 100-ml M9 medium containing the indicated carbon sources and appropriate antibiotics at 37°C, and polyesters accumulated in the cells were analyzed by gas chromatography as described previously [17].

2.2Enzyme assay

Whole cell extracts of the recombinant E. coli were prepared by sonication of the harvested cells grown on an M9 medium containing 10 mM sodium dodecanoate at 37°C for 48–72 h. PHA synthase activity in the whole extracts was determined by a spectroscopic assay according to the method described by Valentin and Steinbüchel [18] using chemically synthesized (R)-3HB-CoA and 5,5′-dithiobis(2-nitrobenzoic acid). Enoyl-CoA hydratase activity in soluble fractions of the whole extracts was assayed by the hydration of crotonyl-CoA [13]. Protein concentrations were determined by using Bio-Rad assay solution and bovine serum albumin as the standard.

2.3Western blot analysis

For preparation of specific antibodies against the PHA synthase of A. caviae, a synthetic oligopeptide (N-CHYVKVRLNPVFACPTEEDAA-C) designed from the deduced C-terminus of the synthase protein was coupled with bovine serum albumin, then the resultant conjugate was injected into a rabbit four times at 0, 2, 6 and 8 weeks, and the antiserum was obtained after 11 weeks. The proteins were separated by SDS-PAGE with 7.5–15% polyacrylamide gels and transferred onto 0.45-μm pore diameter PVDF membranes (Millipore) for 1.5 h at 100 mA. The membranes were subsequently treated with 1% Blocking reagent (Boehringer), the specific anti-serum, and alkaline phosphatase-conjugated mouse anti-rabbit IgG, followed by the addition of Nitro blue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt to visualize the specific proteins.


3.1Construction of plasmids

In order to investigate the PHA biosynthesis in recombinant E. coli, several plasmids were constructed to achieve the functional expression of specific PHA biosynthesis genes of A. caviae, as illustrated in Fig. 1. Plasmids pEE32 [12] and pJRDG13 [13] have been reported previously. For construction of pKBgB18 in which transcription of phaCAc should be controlled by tac promoter, BglII and BamHI sites across phaCAc were created by site-directed mutagenesis on pEE32, and then the BglII/BamHI restriction fragment was ligated to BamHI-digested pKK223-3. pJRDG1 and pJRDG3 contain ORF1Ac and phaJAc, respectively, between the native promoter and terminator regions for the A. caviae pha operon. BglII and BamHI sites across phaCJAc, or two BglII sites across ORF1Ac-phaCAc were created on pEE32, and the BglII/BamHI or BglII restriction fragments were inserted into a broad-host-range vector pJRD215 at the BamHI site, resulting in the construction of pJRDG1 or pJRDG3. pUL32 encodes ORF1Ac having the first eight amino acids of β-galactosidase at the N-terminus, PhaCAc, and PhaJAc downstream of the lac promoter. A 2.7-kbp fragment including the three genes was amplified by PCR with introduction of the NdeI site into the proposed translation start codon of ORF1Ac and a BamHI site downstream of the stop codon of phaJAc. The PCR product was digested with NdeI and BamHI, and then subcloned into modified pUC118 plasmid in which EcoRI site has been converted to NdeI site by using pNdeI linker, to give pUL32.

Figure 1.

Schematic drawing of plasmids used for the expression of A. caviae PHA biosynthesis genes in E. coli LS5218. E, EcoRI; B, BamHI; Bg, BglII; N, NdeI; Ppha, native promoter for A. caviae pha operon; Ptac, tac promoter, Plac, lac promoter.

3.2PHA accumulation in recombinant E. coli LS5218

E. coli LS5218 (fadR601 atoC2(Con)) was transformed by the constructed plasmids and the recombinants were cultivated in M9 medium containing sodium alkanoates (10 mM), and then PHA accumulated in the cells were analyzed. The results are represented in Table 1, and the expression of phaCAc in E. coli cells was confirmed by Western immunoblot analysis, as shown in Fig. 2. E. coli LS5218 did not accumulate detectable PHA from either glucose or dodecanoate. Whereas, LS5218/pEE32 expressing the PHA genes from A. caviae under the native promoter control could synthesize P(3HB-co-3HHx) copolymer up to 7–11 wt% of dry cell weight with 7–9 mol% 3HHx fraction from octanoate and dodecanoate. Clearly, PHA synthase derived from phaCAc, which is capable of incorporating 3HHx unit, was functionally expressed in E. coli cells. This recombinant, in addition, accumulated P(3HV) homopolymer when pentanoate was fed as a carbon source, indicating that the 3HB unit was not generated from acetyl-CoA molecules due to the deficient of phbA (β-ketothiolase) and phbB (NADPH-acetoacetyl-CoA reductase) genes. Hexanoate was not a substrate for PHA synthesis for this recombinant, probably due to the toxic effects of the medium-chain alkanoic acid on the cells.

Table 1.  PHA accumulation in E. coli LS5218 recombinants harboring A. caviae PHA biosynthesis genesa
PlasmidCarbon sourcePHA content (wt%)Composition (mol%)
  1. 3HB, 3-hydroxybutyrate; 3HV, 3-hydroxyvalerate; 3HHx, 3-hydroxyhexanoate.

  2. aCells were cultivated in a M9 medium containing glucose (1% w/v) or sodium alkanoate (10 mM) for 72–96 h at 37°C.

pKBgB18, pJRDG13Dodecanoate58812
pKBgB18, pJRDG1Dodecanoate0
pKBgB18, pJRDG3Dodecanoate6937
Figure 2.

Western blot analysis of cell extracts of recombinant E. coli LS5218 harboring A. caviae PHA biosynthesis gene by using anti-PhaCAc antibodies. The whole cell extracts were prepared from E. coli LS5218 harboring no plasmid (lane 1), pEE32 (lane 2), pKBgB18 (lane 3), pKBgB18-pJRDG13 (lane 4), pKBgB18-pJRDG1 (lane 5), pKBgB18-pJRDG3 (lane 6), pJRDG13 (lane 7), and pUL32 (lane 8).

Although LS5218/pKBgB18 (phaCAc) expressed much more PHA synthase protein (Fig. 2, lane 3) than the strain harboring pEE32 (lane 2), this recombinant accumulated no PHA when grown on dodecanoate. Introduction of phaCAc alone was not enough to confer the PHA biosynthesis ability from alkanoates to E. coli. PHA was not also synthesized in the cells harboring pJRDG13 (ORF1Ac phaJAc) without the synthase gene. Double transformants of LS5218 harboring both pKBgB18-pJRDG13, and pKBgB18-pJRDG3 (phaJAc) accumulated P(3HB-co-3HHx) copolymers from dodecanoate, whereas the strain harboring pKBgB18-pJRDG1 (ORF1Ac) did not, in spite of the strong expression of the synthase gene (lane 5). These results clearly reveal that both phaCAc and phaJAc are essential for E. coli. LS5218 to achieve the PHA biosynthesis from alkanoic acids. In LS5218/pUL32 containing the three PHA biosynthesis genes under the lac promoter control, the amount of expressed synthase protein was similar to those of the other strains harboring pKBgB18, but this recombinant accumulated a significant content (38 wt%) of P(3HB-co-3HHx) with 17 mol% of 3HHx fraction from dodecanoate.

3.3Enzyme assay

PHA synthase and enoyl-CoA hydratase (including both (S)- and (R)-specific enzymes) activities in the recombinant E. coli strains were assayed, as shown in Table 2. PHA synthase activities were very low (6–15 U g−1) even in the recombinants harboring phaCAc. Furthermore, no significant difference of the synthase activity was observed between these phaCAc-positive strains and the control strain, LS5218, although the Western immunoblot analysis clearly demonstrated the existence of translated synthase protein in these recombinant cells.

Table 2.  PHA synthase and enoyl-CoA hydratase activities in E. coli LS5218 recombinants on dodecanoate
PlasmidEnzyme activitya
 PHA synthase (U g−1)Enoyl-CoA hydratase (U mg−1)
  1. aActivities in the crude extract from cells grown on M9 medium containing sodium dodecanoate (10 mM) for 48–72 h at 37°C.

pKBgB18, pJRDG136.20.93
pKBgB18, pJRDG18.70.78
pKBgB18, pJRDG3120.71

In the LS5218 strains harboring phaJAc under the native promoter (pEE32, pJRDG3, or pJRDG13), enoyl-CoA hydratase activities (0.71–0.93 U mg−1) were only slightly higher than that (0.56 U mg−1) in the control. The transcriptional activity of the promoter for the A. caviae pha operon seems to be weak in E. coli cells. The amount of PHA synthase in LS5218/pEE32 expressing phaCAc under the pha promoter was actually small as confirmed by the immunoblot analysis. In contrast, the hydratase activity in LS5218/pUL32 was more than 10-fold higher (13 U mg−1) than those in the other strains, probably due to the efficient expression of phaJAc encoding (R)-specific enoyl-CoA hydratase under the lac promoter control.


In the previous reports, biosynthesis of medium-chain-length PHA from alkanoates has been achieved in fadB mutants of E. coli expressing PHA synthase gene from P. aeruginosa[9] or P. oleovorans[10]. These reports have suggested that E. coli has unknown channeling pathway(s) from fatty acid β-oxidation intermediates to (R)-3HA-CoA monomer units for PHA synthase (Fig. 3b). However, no PHA or only trace amount of PHA has been synthesized when E. coli having intact β-oxidation pathway were used as host strains. The channel from β-oxidation to PHA biosynthesis seems to be inefficient in E. coli strains without mutations in the gene(s) for β-oxidation (Fig. 3a).

Figure 3.

Proposed models of PHA biosynthesis pathways in recombinant E. coli from alkanoic acids via β-oxidation cycles.

We demonstrate here that E. coli LS5218/pKBgB18, harboring only phaCAc, did not synthesize any PHA from dodecanoate even though a significant amount of the PHA synthase protein was detected in the cell extract by immunoblot analysis. While, the introduction of phaJAc, encoding (R)-specific enoyl-CoA hydratase, together with phaCAc into LS5218 resulted in the accumulation of P(3HB-co-3HHx) copolyester from dodecanoate at a detectable level. These facts provide the evidence that the co-expression of PhaCJAc is able to establish the P(3HB-co-3HHx) biosynthesis pathway in E. coli via the intact β-oxidation cycles. In the recombinant harboring phaCJAc, trans-2-enoyl-CoA intermediates from 4 to 6 carbon numbers, generated from the β-oxidation of longer alkanoic acids, can be converted to the corresponding (R)-3HA-CoA by phaJAc-derived (R)-hydratase, and subsequently polymerized to P(3HB-co-3HHx) by the function of PHA synthase encoded by phaCAc (Fig. 3c), as previously proposed in A. caviae[13].

P(3HB-co-3HHx) was accumulated up to 38 wt% in the cells of LS5218/pUL32, whereas the content was much lower in the strain harboring pEE32, pKBgB18-pJRDG13, or pKBgB18-pJRDG3. PHA synthase activities in the extracts of these recombinants were as low as that of the control strain LS5218, even in the extract of LS5218/PUL32. Much PHA synthase protein did not retain the polymerization activity, leading to the little amount of functional synthase in E. coli, nevertheless, such low synthase activity was enough to achieve the accumulation of quantity amount of the copolyester. Enoyl-CoA hydratase activity in LS5218/pUL32, which expressed phaJAc under lac promoter control, showed more than 10-fold higher hydratase activity of the others. These results of PHA production and enzyme assay suggest that the level of (R)-specific enoyl-CoA hydratase activity is an important factor for PHA biosynthesis in this recombinant system. Higher (R)-hydratase activity may enhance the intracellular carbon flux toward PHA biosynthesis via β-oxidation pathway, resulting in a significant accumulation of the polyesters.

In conclusion, expression of phaCJ of A. caviae can establish P(3HB-co-3HHx) biosynthesis pathway from alkanoic acids in E. coli cells. The information obtained here may be applicable for molecular breeding of efficient P(3HB-co-3HHx)-producing bacteria, for example, by employing oil-assimilating bacteria as hosts of phaCJAc genes.


This study was performed as a part of the development of Biodegradable Plastics supported by the New Energy and Industrial Technology Development Organization (NEDO).