INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Poly(3-hydroxybutyrate) [P(3HB)] was identified as an intriguing bacterial inclusion body more than seven decades ago (21) and is now classified as one of the many different types of bacterial polyesters with the common name polyhydroxyalkanoate (PHA). Recently, due to increased awareness of global environmental issues, PHA has come under investigation because of its inherent property as a biodegradable thermoplastic (4, 37).

The biosynthesis of PHA has been studied in great detail in Alcaligenes eutrophus (28, 29, 36). In this bacterium, the biosynthesis process is initiated by the condensation of two acetyl coenzyme A (acetyl-CoA) molecules to acetoacetyl-CoA catalyzed by the enzyme -ketothiolase (EC 2.3.1.9). The structural gene for this enzyme is designated phaA. Acetoacetyl-CoA is then reduced to the R enantiomer of 3-hydroxybutyryl-CoA by an NADPH-dependent acetoacetyl-CoA reductase (EC 1.1.1.36), the structural gene for which has been given the designation phaB. Both these enzymes have been thoroughly studied for several PHA-accumulating bacteria (8, 11, 12, 23, 25, 26, 38). Finally, the key enzyme, PHA synthase, which is encoded by a structural gene designated phaC, catalyzes the polymerization of (R)-3-hydroxybutyryl-CoA to P(3HB). All the three genes are constitutively expressed in A. eutrophus and form a single operon, with the order phaC-A-B.

PHA synthase genes from more than 20 different bacteria have been cloned and analyzed (20). The results show that PHA synthases are a class of highly versatile enzymes and are not specific to only one type of hydroxyalkanoic acid (HA) (42). Nevertheless, they can be broadly classified into two different types based on their primary amino acid sequences and substrate specificities. One type is active towards short-chain HA, consisting of three to five carbon atoms, and is represented by the PHA synthase of A. eutrophus. Some of the PHA synthases of this type were also found to incorporate into PHA 4- and 5-HA, such as 4-hydroxybutyric acid (4HB) (19), 4-hydroxyvaleric acid (4HV) (45), and 5HV (5). The other type, which is represented by the PHA synthase of Pseudomonas oleovorans, is active towards medium-chain-length HA, containing 6 to 14 carbon atoms. In addition, a few bacteria, such as Aeromonas caviae (7) and Rhodococcus ruber (10), have also been reported to have synthases that exhibit specificity for both short-chain and medium-chain-length HA.

PHAs containing monomer compositions that result in useful properties have constantly been sought. The physical char- acteristics of poly(3-hydroxybutyrate-co-4-hydroxybutyrate) [P(3HB-co-4HB)] copolymers, having 4HB contents ranging from 0 to 100 mol%, show their great potential as biodegradable materials (32). Since 4HB is a linear monomer, it has the added advantage of being able to be degraded by a lipase as well as a depolymerase (31). In Comamonas acidovorans, it was shown that the 4HB monomer content could be easily controlled by supplying substrate mixtures of 4HB and other carbon sources. This results in the ability to biosynthesize PHAs with wide ranges of elasticity and tensile strength and, most importantly, controlled rates of biodegradation (31, 32).

In addition to C. acidovorans, A. eutrophus (19) and Alcaligenes latus (15) have been reported to produce PHA containing 4HB as a monomer. Among these wild-type bacteria, only C. acidovorans is capable of producing PHA copolymers with a very high (>90 mol%) 4HB monomer content, more than 20% of the dry cell weight (31). Invariably, however, 4HB is incorporated into PHA only when related carbon sources, such as 4HB, -butyrolactone, and 1,4-butanediol, are provided in the culture media. An attempt was made to produce P(3HB-co-4HB) in recombinant Escherichia coli using an unrelated carbon source, such as glucose. For this purpose, succinate degradation genes from Clostridium kluyveri were introduced into E. coli together with the PHA biosynthesis genes from A. eutrophus. It is known from a previous study (41) that 4-hydroxybutyryl-CoA, which is an immediate precursor for the PHA synthase of A. eutrophus, occurs as an intermediate in the cofermentation of succinic acid and ethanol in Clostridium kluyveri. Therefore, it was anticipated that 4-hydroxybutyryl-CoA could be generated from succinic acid in E. coli if the genes coding for necessary enzymes were introduced. However, despite the theoretically attractive approach, a maximum of 2.8 mol% 4HB was incorporated into PHA by the recombinant E. coli (44).

To this end, C. acidovorans seems to be a very promising wild-type candidate with a metabolic pathway suitable for the production of P(3HB-co-4HB) with a 0 to 100 mol% 4HB monomer content. In this paper we report the cloning and characterization of the PHA biosynthesis genes of C. acidovorans DS-17 (JCM10181). In addition, the substrate specificity of C. acidovorans PHA synthase was evaluated by heterologous expression of the cloned PHA synthase gene. The discussion emphasizes the capacity of the metabolic pathways in C. acidovorans to produce PHA with a high 4HB monomer content.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. All strains and plasmids used in this study are listed in Table 1. C. acidovorans and A. eutrophus were grown in a nutrient-rich (NR) medium containing meat extract (10 g/liter), Bacto Peptone (10 g/liter), and yeast extract (2 g/liter) at 30°C. E. coli strains were grown at 37°C in Luria-Bertani (LB) medium. For maintenance of plasmids, 25 mg of tetracycline per liter or 50 mg of ampicillin or kanamycin per liter was added.

Analysis of PHA accumulation. PHA accumulation was analyzed in both one-stage and two-stage batch cultivation. For one-stage cultivation, filter-sterilized carbon source was added to 100 ml of mineral salts (MS) medium consisting of 0.9 g of Na₂HPO₄ · 12H₂O, 0.15 g of KH₂PO₄, 0.05 g of NH₄Cl, and 0.02 g of MgSO₄ · 7H₂O and supplemented with 0.1 ml of trace element solution (18). For two-stage cultivation, cells were initially grown in 100 ml of NR medium for 12 h before being harvested and transferred to the same volume of nitrogen-free MS medium. Different carbon sources were then tested for their ability to promote PHA synthesis. To determine the polyester content and composition, 10 to 25 mg of lyophilized cell material was subjected to methanolysis in the presence of 15% (vol/vol) sulfuric acid. The resulting hydroxyacyl methyl esters were then analyzed by gas chromatography (GC) (2). For PHA copolymer containing high levels of 4HB, solvent extraction was employed and the content (weight percent) was determined gravimetrically. Lyophilized cells were stirred for 2 days in chloroform at room temperature (39). The insoluble cellular material was then collected by filtration and methanolyzed for GC analysis to check for incomplete extraction, while the chloroform extract was concentrated prior to precipitation of the polymer in methanol. The precipitate was then washed in methanol and ether followed by drying in vacuo, and it was then subjected to gravimetric measurement and compositional analysis by GC.

Isolation and manipulation of DNA. Isolation of total genomic DNA (gDNA) from C. acidovorans, plasmid DNA isolation, agarose gel electrophoresis, and transformation of E. coli were carried out according to standard procedures (33). Restriction endonucleases and all other DNA-manipulating enzymes and kits were used according to the manufacturers' protocols. Transconjugation of A. eutrophus with E. coli S17-1 harboring broad-host-range plasmids was performed as described by Friedrich et al. (6).

Construction of gDNA library. gDNA isolated from C. acidovorans was partially digested with HindIII, and the genomic fragments were ligated to the HindIII site of a broad-host-range cosmid vector, pLA2917 (1). The resulting concatemers were then packaged in vitro with a Gigapack II packaging kit (Stratagene). Mature phage particles containing the genomic fragments were then used to infect E. coli S17-1. Transformants harboring the hybrid cosmids were selected by plating on LB agar plates containing tetracycline (12.5 mg/liter).

Southern blot analysis and colony hybridization. A 1.8-kbp Csp45I-AatI fragment harboring the PHA synthase gene of A. eutrophus was used as a probe. Preparation of labeled probe and visualization of successful hybridization were carried out with the digoxigenin nucleic acid labeling and detection kit (Boehringer Mannheim Biochemicals).

Nucleotide sequence analysis. DNA fragments to be sequenced were subcloned into pBluescript II KS(+), and nested sets of deletion clones were generated by using exonuclease III (14). DNA sequencing was carried out by the dideoxy chain termination method with the Prism 310 DNA sequencer (Applied Biosystems, Inc.) employing the dye terminator labeling procedure (Perkin Elmer Corp.). Nucleic acid sequence data and the deduced amino acid sequences were analyzed with GENETYX-MAC software (Software Development Co., Tokyo, Japan). Similarity searches were performed with the BLAST (Basic Local Alignment Search Tool) program and the NCBI (National Center for Biotechnology Information) databases.

Nucleotide sequence accession number. The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL, and GenBank nucleotide sequence databases with the accession no. AB009273.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Identification, cloning, and nucleotide sequence of the C. acidovorans PHA biosynthesis genes. The digoxigenin-labeled PHA synthase gene of A. eutrophus (digoxigenin-labeled phaC_Ae) was used as a heterologous hybridization probe to detect the homologous gene of C. acidovorans. A relatively strong single hybridization signal appeared in the 20-kbp HindIII-digested fragments of C. acidovorans gDNA. Based on this observation, a genomic library was constructed with HindIII-digested gDNA. Screening of the gDNA library by colony hybridization with the digoxigenin-labeled phaC_Ae probe resulted in the isolation of one positive monocolony harboring a recombinant cosmid, designated pLACa. To confirm the functional activity of the cloned gDNA fragment, pLACa was mobilized from the E. coli S17-1 host cells into a PHA-negative mutant strain of A. eutrophus, PHB4. The ability to accumulate PHA was restored to the mutant upon transconjugation. Cells of E. coli S17-1 harboring pLACa, however, did not accumulate PHA when they were cultivated in LB medium supplemented with 0.5% (wt/vol) glucose and/or 0.5% (wt/vol) 4HB.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Organization of the PHA biosynthetic genes of C. acidovorans in EV80 subfragment. (a) Subfragments relevant for nucleotide sequence analysis; (b) restriction map; (c) strategy for sequencing of the SH40 subfragment.

Structure of the putative gene product of ORF1. ORF1 (1,893 bp) starts 748 bp downstream from the SmaI site (Fig. 2). Several potential translation initiation codons were present in the region between 700 and 900 bp (at bp 748, 841, 874, 883, and 892), but the ATG beginning at position 748 was considered the most probable translation initiation codon based on the presence of a reliable Shine-Dalgarno sequence 6 bp upstream. ORF1 encodes a protein of 630 amino acids with a calculated relative molecular weight (M_r) of 69,068, and this gene was referred to as the PHA synthase gene of C. acidovorans (phaC_Ca) because the translated product showed strikingly high identities to the synthases of Alcaligenes sp. strain SH-69 (78.7%) and A. eutrophus (51.1%) (Fig. 3). In order to maximize homology, several gaps had to be introduced into the amino acid sequences of these synthases, mainly in the regions that correspond to the phaC_Ca-encoded primary structure between amino acids 338 and 389 (Fig. 3). This extra region encoded by phaC_Ca apparently contributes to the main difference between the primary structures of these synthases. The primary structure of the product of phaC_Ca, however, exhibited relatively lower identities to other synthases capable of incorporating short-chain HA, such as those of Rhodobacter sphaeroides (35.4% identity) (17) and Methylobacterium extorquens (33.1% identity) (46). The phaC_Ca product showed about 40% identity to the medium-chain-length PHA synthases encoded by phaC1 and phaC2 of both P. oleovorans (16) and Pseudomonas aeruginosa (43). Moreover, a highly conserved cysteine residue that has been well documented for the catalytic cycle (9) was also well preserved in the deduced amino acid sequence for phaC_Ca (Fig. 2). Computer identity search and promoter analysis did not reveal any other PHA-related genes or promoter region in the 736-bp DNA sequence immediately upstream of the tentative ribosome binding site (RBS).

View larger version (61K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence of a 3,973-bp region containing the phaCCa and phaACa genes with the amino acid sequences. Stop codons are indicated by asterisks. SD, Shine-Dalgarno site (RBS).

View larger version (85K): [in this window] [in a new window]   FIG. 3.   Alignment and identities of the deduced sequence of PHA synthase from C. acidovorans with those from Alcaligenes (Alc.) sp. strain SH-69 (GenBank accession no. U78047), A. eutrophus (29, 36), Aeromonas caviae (7), Rhodococcus ruber (30), and Chromatium vinosum (22), which have the ability to incorporate short-chain HA into PHA. Amino acids identical in at least four sequences are boxed in black.

Structure of the putative gene product of ORF2. ORF2 (1,182 bp), which was located immediately downstream of phaC_Ca, has a spacer 123 bp from the transcriptional termination codon of ORF1 (Fig. 2). A potential RBS preceded the start codon at a distance of 9 bp. ORF2 was calculated to encode a 393-amino-acid protein with an M_r of 40,238. This deduced amino acid sequence, when subjected to an identity search, was found to have significant identity (74.2%) to the primary structure of -ketothiolase from A. eutrophus (28). Therefore, ORF2 was concluded to represent a structural gene for C. acidovorans -ketothiolase, phaA_Ca. It also showed significantly high identities, i.e., >60%, to other thiolases known to be involved in the biosynthesis of PHA (27, 34). This prompted us to determine the nucleotide sequence of the downstream region of phaA_Ca with the anticipation of finding a third gene involved in PHA biosynthesis, i.e., the gene coding for acetoacetyl-CoA reductase. However, despite the similarity of the arrangement of phaC_Ca and phaA_Ca to the arrangement of genes in the PHA biosynthetic operon of A. eutrophus, a gene corresponding to the third gene, phaB, was not found in the region downstream of phaA_Ca.

Heterologous expression studies. Expression of the cloned 20-kbp C. acidovorans gDNA in A. eutrophus PHB4 complemented the mutant strain, resulting in the accumulation of significant amounts of PHA, as shown in Table 2. P(3HB) homopolymer, as well as the copolymers P(3HB-co-3HV) and P(3HB-co-4HB), could be synthesized by the recombinant PHB4 strain from suitable carbon sources. Use of fructose as a carbon source resulted in the production of P(3HB) homopolymer in both one-stage (results not shown) and two-stage cultivations. Comparison with the monomer composition of PHA produced by wild-type A. eutrophus showed that the ability to incorporate 3HB monomers was significantly improved in the recombinant PHB4 strain. When pentanoate was supplied as the carbon source, the 3HB monomer content was increased to about 30 mol% more than that in the wild-type A. eutrophus. A similar increase was also observed when 4HB was the carbon source. The final PHA content in the recombinant PHB4 strain was not very different from that in the wild type when either fructose or pentanoate was the carbon source. However, the PHA contents accumulated in the wild-type C. acidovorans incubated with pentanoate or 4HB were 43 and 31% of the dry cell weight, respectively, and the latter was greater than that produced by the wild-type A. eutrophus or the recombinant PHB4 strain.

View this table: [in this window] [in a new window]   TABLE 2.   PHA accumulation by A. eutrophus PHB4 complemented with pLACa and by wild types of A. eutrophus and C. acidovoransa

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The genes for the key enzymes involved in PHA biosynthesis, PHA synthases, from more than 20 different bacteria have been cloned and analyzed with the objectives of elucidating the mechanism of PHA biosynthesis and reproducing the steps leading to the synthesis of desired bacterial polyesters. In this study, we report the cloning and characterization of the PHA synthase gene (phaC_Ca) and the -ketothiolase gene (phaA_Ca) of C. acidovorans DS-17 (JCM10181). Based on the types of PHAs produced, both the PHA synthases of C. acidovorans and A. eutrophus show specificity for short-chain 3- and 4-hydroxyacyl-CoA thioesters. The main difference is that C. acidovorans is able to produce a homopolymer of 4HB in substantial quantities (>20% of the cell dry mass) from 4HB and 1,4-butanediol (32). This prompted us to clone and characterize the PHA biosynthesis genes of C. acidovorans, to study the substrate specificity of the phaC_Ca-encoded product, and to determine if the product shows any preference for 4-hydroxybutyryl-CoA as a monomer.

The nucleotide sequence analysis of a 4-kbp SmaI-HindIII (Fig. 1) subfragment revealed that phaC_Ca and phaA_Ca were arranged as they are in the PHA biosynthesis operon of A. eutrophus (phaC-A-B), but there was a difference for the third gene, the gene encoding acetoacetyl-CoA reductase (phaB), which was not located in the region immediately downstream of phaA_Ca. This distinguishes the PHA biosynthesis operon of C. acidovorans from all those that have been reported to date. Comparison of the deduced amino acid sequence of the phaC_Ca product by alignment with other known synthases showed the highest identities to the synthases that are specific for short-chain HAs, especially to the synthases of Alcaligenes sp. strain SH-69 and A. eutrophus (Fig. 3). In contrast, the primary structure of phaC_Ca showed much lower identities to the synthases of Rhodobacter sphaeroides (17), M. extorquens (46), and Rhodococcus ruber (30), which were 35.4, 33.1, and 28.8%, respectively.

The heterologous expression in A. eutrophus PHB4 of a 20-kbp gDNA fragment from C. acidovorans containing phaCA_Ca suggests that the substrate specificity of C. acidovorans PHA synthase is indeed similar to that of A. eutrophus. Conditions that favored the production of P(4HB) homopolymer in C. acidovorans did not enhance the incorporation of 4HB monomers by the PHB4 transconjugant. This indicates that, more than the specificity of the PHA synthase, the monomer-supplying pathway plays an important role in determining the type of PHA that can be produced. It is known that A. eutrophus is unable to produce P(4HB) homopolymer in quantities of more than 1 or 2% of the dry cell weight (24). But the coexpression of phaC_Ae and orfZ of Clostridium kluyveri, which putatively encodes a 4HB-CoA transferase, had enabled the construction of an E. coli strain that produces P(4HB) in larger quantities (13). These results support the idea that a metabolic environment which can supply the necessary precursors for PHA biosynthesis is the most crucial aspect of improving 4HB incorporation into PHA. In A. eutrophus, the channeling of 4HB into the 3-hydroxybutyryl-CoA production pathway may be more efficient than that into the 4-hydroxybutyryl-CoA production pathway, whereas the opposite may be true for C. acidovorans. No preference for 4-hydroxybutyryl-CoA was shown by the phaC_Ca-encoded product when expressed in A. eutrophus PHB4; instead, the 3HB content was clearly increased (Table 2). The PHA content in wild-type C. acidovorans when supplied with 4HB was higher than in wild-type A. eutrophus and the recombinant PHB4 strain. This further suggests that the ability of C. acidovorans to produce PHA with a high 4HB monomer content is due to its metabolic capacity to efficiently supply 4HB monomers rather than to the specificity of its PHA synthase.