INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The ability to degrade extracellular poly(3-hydroxybutyrate) (PHB) and related polyesters is widely distributed among bacteria and depends on the secretion of specific polyester depolymerases which hydrolyze the water-insoluble polyester to water-soluble monomers or oligomers. One of the best-studied polyester-degrading bacteria is Pseudomonas lemoignei (5). It belongs to the beta subclass of Proteobacteria and is related to the Burkholderia-Ralstonia rRNA sublineage. The catabolic abilities of P. lemoignei are restricted to the utilization of a few organic acids (acetate, butyrate, valerate, pyruvate, succinate, and 3-hydroxybutyrate), and polyesters such as PHB, poly(3-hydroxyvalerate) (PHV), and related short-chain-length polyhydroxyalkanoates (PHA_SCL). Sugars, alcohols, and amino acids do not support growth of P. lemoignei (5, 17). Recently, P. lemoignei (strain A62) was reisolated by application of a specific enrichment procedure with PHV as the sole source of carbon and energy (16).

P. lemoignei is unique among PHA-degrading bacteria because it is able to synthesize at least five different extracellular PHA depolymerases. Three PHA depolymerases are specific for PHB and copolymers of 3-hydroxybutyrate (3-HB) and 3-hydroxyvalerate (3-HV) with low 3-HV content (PHB depolymerases A, B and D). The activity of these enzymes with the homopolyester PHV is below 5% of the activity obtained with PHB as substrate. None of the three PHB depolymerases is able to produce clearing zones on opaque PHV granule-containing agar. The two remaining PHA depolymerases (PHB depolymerases C and PHV depolymerase) also degrade PHB, but are additionally able to hydrolyze PHV, with about 15 and 30% activity compared to PHB hydrolysis. PHB depolymerase C and PHV depolymerase produce clearing zones on opaque PHV agar. Since the latter depolymerase is expressed only during growth on odd-numbered carbon sources (PHV and valerate), this PHA depolymerase was named PHV depolymerase (17).

Five PHA depolymerase structural genes (phaZ1 to PhaZ5) of P. lemoignei were cloned in earlier studies (3, 10, 11). Four genes, namely phaZ1, phaZ2, phaZ3, and phaZ5, were identified as encoding PHB depolymerases C, B, D, and A, respectively, by agreement of the DNA-deduced amino acid sequences with the sequences derived from Edman degradation of the purified proteins. When the N-terminal amino acid sequence of the PHV depolymerase was compared with the DNA-deduced amino acid sequence of the remaining gene, phaZ4, 23 of 24 amino acids were identical. A threonine codon had been determined at the position of the seventh codon of the deduced mature depolymerase, but an alanine had been identified in the N-terminal amino acid of the purified PHV depolymerase (10). From these results, it was hypothesized that phaZ4 might encode a different, yet unknown, PHA depolymerase and that the true PHV depolymerase structural gene had not been cloned so far. In that case, a sixth PHA depolymerase gene should be present and prompted us to undertake the present study.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids used in this study are shown in Table 1. P. lemoignei was grown in Nakayama mineral salts medium (MM) with 0.4% (wt/vol) (R,S)-3-hydroxybutyrate as described recently (26). Recombinant E. coli strains harboring PHA depolymerase genes were grown in nutrient broth (NB) medium supplemented with ampicillin (Ap; 100 µg/ml) and IPTG (isopropyl--D-thiogalactopyranoside [0.2 g/liter]) at 37°C. Identification of recombinant Escherichia coli strains expressing PHA depolymerase genes was performed by clearing zone formation on NB-PHB agar (NB-Ap-IPTG underlay [25 ml], NB-Ap-IPTG agar supplemented with 0.2% [wt/vol] denatured PHA granule overlay [7 ml]) and on M9-Ap-PHB mineral salts medium supplemented with ampicillin, IPTG, proline (50 mg/liter), thiamine (1 mg/liter), glucose (1 g/liter), and glycerol (5 ml/liter) underlay (25 ml) plus a 7-ml overlay agar of the same composition, but with addition of PHA granules as described above.

Preparation of polymer suspension and assay of PHA depolymerases. The polymers PHB, PHV, poly(3-hydroxyoctanoate) (PHO), and poly(3-hydroxyoctanoate-co-3-hydroxydecanoate) (PHO/HD) were isolated from sodium gluconate-grown cells of Ralstonia eutropha H16 (PHB), sodium valerate-grown cells of C. violaceum (PHV), sodium octanoate-grown cells of Pseudomonas oleovorans (PHO), and sodium gluconate-grown cells of Pseudomonas putida KT2440 (PHO/HD) as described previously (10, 11). PHA depolymerase activity was assayed by measuring the decrease in the optical density (650 nm) of the respective polymer suspension at 37°C. The reaction mixture contained 50 mM Tris-HCl (pH 8) with 1 mM CaCl₂ and 60 µl of a 0.3% (wt/vol) suspension of denatured PHB granules (total volume, 1 ml). The reaction was started by the addition of 5 to 50 µl of depolymerase preparation. Alternatively, clearing zone formation of opaque agar media containing PHA granules in 100 mM Tris-HCl (pH 8) and 1 mM CaCl₂ upon the addition of 1 to 2 µl of PHA depolymerase solution and incubation at 37°C semiquantitatively indicated the activity of the depolymerase.

Purification of PHA depolymerases. PHV depolymerase from P. lemoignei was purified from 5 liters of cell-free culture fluid of 0.2% (wt/vol) sodium valerate-grown (30°C, 24 to 36 h) cells by (i) ammonium sulfate precipitation (35 to 75%), (ii) chromatography on DEAE-Sephacel, and (iii) chromatography on CM-Sepharose-CL6B as described earlier (17), except that one additional purification step (iv) was required (size exclusion chromatography on Sephadex G-75S).

Proteolytic digestion by endoproteinase GluC. Purified wild-type PHV depolymerase (340 µg) in 50 mM K₂HPO₄ (pH 7.5), containing 0.05% (wt/vol) sodium dodecyl sulfate (SDS), was incubated at 100°C for 3 min. After cooling to room temperature, 1/30 (11 µg) endoproteinase GluC from Staphylococcus aureus V8 was added, and the complete reaction mixture was incubated at 37°C for 28 h. The cleavage products were separated by denaturing SDS-polyacrylamide gel electrophoresis (PAGE) and subsequently blotted onto a polyvinylidene difluoride membrane at 5 mA/cm² for 50 min and stained with Coomassie blue.

Immunological techniques. Rabbit polyclonal antisera were raised against (i) PHV depolymerase purified from P. lemoignei and (ii) PHB depolymerase purified from Comamonas sp. The immunoglobulin G (IgG) fractions were purified from the antisera (three injections of 100 to 500 µg of purified protein in 50% [vol/vol] complete Freund's adjuvant [first injection], 50% (vol/vol) incomplete Freund's adjuvant [second injection], and without adjuvant [third injection]) by standard chromatography on protein A Sepharose and stored at 20°C. Purified PHA depolymerase proteins were tested for immunological relatedness by Ouchterlony double-immune-diffusion testing in 1% (wt/vol) agarose in 100 mM Tris-HCl (pH 7.6) (19). Precipitation bands were visualized by staining with Coomassie blue.

Molecular biology techniques. Standard molecular biology techniques were used throughout this study, including double-stranded DNA sequencing by dideoxy chain termination and primer walking.

Cloning of the PHV depolymerase structural gene phaZ6. Agarose gel-separated and size-fractioned (2 to 10 kbp) partially MboI-digested genomic DNA of P. lemoignei was ligated to a mixture of BamHI-linearized pUC expression vectors (pUC9, pUC9-1, and pUC9-2) and transformed in E. coli XL1-Blue. Colonies harboring recombinant plasmids were screened for halo formation on PHB- and PHV-containing NB and M9 agar (described above).

Nucleotide sequence accession number. The DNA sequence of phaZ6 is available under accession no. AF222999.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Biochemical and immunological characterization of the PhaZ4 depolymerase from recombinant E. coli and the wild-type PHV depolymerase from P. lemoignei. In order to elucidate whether phaZ4 or another gene encoded the PHV depolymerase of P. lemoignei, the recombinant PhaZ4 depolymerase and the wild-type PHV depolymerase were purified, and the relevant characteristics of both depolymerases were compared. Recombinant E. coli cells, which harbored phaZ4 (pSN874), were grown on NB-PHB and M9-PHB as well as on NB-PHV and M9-PHV agar plates at 37°C, respectively. Large clearing zones appeared on NB-PHB and M9-PHB medium after 2 days and indicated the functional expression of PhaZ4. However, no clearing zone formation was observed on PHV agar even after 1 week of incubation. Apparently, the activity of PhaZ4 with PHV as a substrate was very low. The recombinant PhaZ4 depolymerase and the wild-type PHV depolymerase were purified as described in Materials and Methods. The purified proteins were homogeneous, as judged by denaturing SDS-PAGE and silver staining. The ratios between the specific PHB and PHV depolymerase activity remained almost constant throughout purification and amounted to 25:1 for PhaZ4 and 3:1 for the wild-type PHV depolymerase (Table 2). Therefore, it appears unlikely that PhaZ4 represents the PHV depolymerase.

View larger version (30K): [in this window] [in a new window]   FIG. 1.   Amino acid alignment of PHA depolymerases PhaZ6 (PHV depolymerase) and PhaZ4 (PHB depolymerase E). The first 300 amino acids of the mature proteins are aligned (catalytic domain). The N-terminal amino acid sequences of purified PHV depolymerase and of four endoproteinase GluC-generated internal peptides are indicated by italic letters and by underlining, respectively. Mismatches between amino acids determined by Edman degradation and the DNA-deduced sequence of phaZ4 are indicated by asterisks (*).

Cloning of the PHV depolymerase structural gene (phaZ6). Due to the apparent high amino acid sequence similarity between PhaZ4 and the PHV depolymerase, the identification of a recombinant PHV depolymerase-containing clone by colony hybridization with a PHV depolymerase-specific oligonucleotide or by a specific PCR-based amplification of the PHV depolymerase structural gene appeared not to be promising. A less selective approach was chosen for the detection of PHA depolymerase activity-expressing clones: about 6,500 recombinant strains of a genomic library of P. lemoignei in E. coli XL1-Blue were screened for haloformation on opaque M9-PHB and NB-PHB agar. Eight positive clones (no. 26, 39, 46, 49, 62, 66, 109, and 110) were identified that produced large clearing zones on PHB agar and indicated the functional expression of a PHA depolymerase. After purification, these clones were tested for halo formation on M9-PHV and NB-PHV agar. Except for no. 62 and 110, all other six clones produced clearing zones on both PHV agar types. The diameters of the clearing zones were significantly smaller on PHV agar and appeared later than those on PHB agar. The plasmid DNA of all eight clones was isolated, and the DNA sequences of the corresponding insert ends were determined. Seven clones harbored already known PHB depolymerase genes, namely phaZ1 (clones 39, 49, and 109), phaZ2 (clones 26 and 46), phaZ3 (clone 62), and phaZ5 (clone 110). Clone 66, which showed the highest clearing zone on PHV agar, was different from all other clones, and the complete DNA sequence (2,595 bp) of this clone was determined for both strands. One open reading frame (ORF1) of 1,254 bp was identified that encoded the PHV depolymerase structural gene phaZ6 (see below).

Characterization of phaZ6. A potential ribosome binding site (5'-GGAGA-3') and a sequence with homology to ⁷⁰-dependent promoters (5'-TTGACT-N17-AAATCT-3') were present 7 and 172 bp upstream of phaZ6, respectively. Another potential ATG start codon 18 bp upstream of phaZ6 was unlikely to represent the initiation site of translation due to the absence of a ribosome binding motif. A potential hairpin structure followed by a 7-bp poly(T) sequence was identified 88 bp downstream of phaZ6 and might represent a factor-independent termination site of transcription. The G+C contents of phaZ6 and of the total sequenced DNA amounted to 54.5 and 52%, respectively, which is a slightly lower value than that of total genomic DNA (58%) (5). The codon usage of phaZ6 was partially different from that of other genes sequenced from P. lemoignei. This difference was most evident for the His, Asp, Asn, and Lys codons: all eight His codons had the sequence CAT (100%), compared to only 63% for CAT and 37% for CAC as the average of other sequenced genes of P. lemoignei. The codons for Asp, Asn, and Lys with A or T in the wobble position were preferentially used in phaZ6 (63, 68, and 75%) compared to only 40, 42, and 25% in other genes of P. lemoignei, respectively. Codons with a G or C in the wobble position were preferentially used for most of the remaining amino acids and resulted in an average G+C content of the total gene above 50%. Inspection of the G+C content of the sequenced DNA revealed a local minimum of 35 to 40% in the region up to 400 bp upstream of phaZ6.

Characterization of PhaZ6. phaZ6 coded for a protein of 417 amino acids (43,610 Da). The N-terminal deduced amino acid sequence had the characteristics of signal peptides of secretory precursors: two positively charged amino acids (K₅ and R₉) were followed by several hydrophobic residues, and a putative signal peptidase cleavage site was predicted between A₂₅ and L₂₆, resulting in a molecular mass of 41,048 Da for the mature protein. The DNA-deduced amino acid sequence was in complete agreement with the 21 amino acids that had been determined by Edman degradation of the purified PHV depolymerase of P. lemoignei from amino acid 26 onward and sustained the predicted signal peptidase cleavage site. The sequence included an alanine at position 7 of the mature protein and contained the complete sequences of the four endoproteinase GluC-derived peptide fragments of the PHV depolymerase (Fig. 1). These results confirmed that phaZ6 encoded the true PHV depolymerase. The homology of the PhaZ6 amino acid sequence (mature protein) to other extracellular PHB depolymerases varied between 25% (Streptomyces exfoliatus PHB depolymerase) and 50% identical amino acids for the PhaZ4 depolymerase of P. lemoignei. The homology of PhaZ6 to PhaZ4 was particularly high within the first 200 amino acids (72.5%, Fig. 1).

View larger version (61K): [in this window] [in a new window]   FIG. 2.   Domain structure of PHASCL depolymerases. An interpretation of the DNA-deduced amino acid sequences is shown. The last three letters of each depolymerase refer to the bacterial origin: P. lemoignei (Ple), A. faecalis strains T1 and AE122 (Afa), P. stutzeri (Pst), R. pickettii (Rpi), Acidovorax sp. (Asp), Comamonas sp. (Csp), C. acidovorans (Cac), C. testosteroni (Cte), and S. exfoliatus (Sex). For references describing the PHA depolymerase genes, see Fig. 3.

View larger version (36K): [in this window] [in a new window]   FIG. 3.   Alignment of catalytic amino acids of PHASCL depolymerases. An alignment of the regions around the catalytic triad amino acids serine (S), aspartate (D), and histidine (H) and around the putative oxyanion histidine (H) (indicated by boldface letters) is shown. The numbering refers to the mature polypeptides. Positions (pos.) of amino acids with hydrophobic (*) or small (+) side chains are indicated in the consensus sequence.

View larger version (45K): [in this window] [in a new window]   FIG. 4.   Alignment of PHASCL depolymerase C-terminal substrate-binding domains. An alignment of the C-terminal 61 amino acids is shown (numbering refers to PhaZ6 depolymerase). The substrate-binding domains of types A and B are aligned separately under A and B, respectively. Residues that are conserved in almost all sequences are indicated by boldface letters. For abbreviations of depolymerases and for definitions of abbreviations, see the legends to Fig. 2 and 3.

Purification and characterization of PhaZ6 from recombinant E. coli. PhaZ6 was purified from the periplasm of IPTG-induced NB-grown cells. The elution behavior of PhaZ6 was very similar to that of the wild-type PHV depolymerase during all stages of purification. The purified PhaZ6 depolymerase hydrolyzed PHA_SCL, such as PHB and PHV, at high rates, but was unable to degrade medium-chain-length PHA (PHA_MCL) such as PHO or P(HO/HD). The ratio between the specific PHB and PHV depolymerase activity remained constant at about 3:1 throughout all stages of purification and is in agreement with the same ratio that had been determined for the PHV depolymerase purified from P. lemoignei (Table 2). The purified PhaZ6 depolymerase was subjected to an Ouchterlony double-immune-diffusion test against a polyclonal antiserum that had been raised against the purified wild-type PHV depolymerase. The PHV depolymerase and the recombinant PhaZ6 depolymerase both cross-reacted with the antibodies, and the precipitation bands perfectly mingled without the formation of a spur (Fig. 5). This indicated that both proteins are immunologically identical. On the contrary, no cross-reaction occurred with purified PHB depolymerases A and B (Fig. 5). PhaZ6 depolymerase did not cross-react with antibodies raised against PHB depolymerase of Comamonas sp., and antibodies raised against the wild-type PHV depolymerase did not cross-react with PHB depolymerase A (PhaZ5) of P. lemoignei (data not shown).

View larger version (119K): [in this window] [in a new window]   FIG. 5.   Ouchterlony double-immune-diffusion assay. The immunological reaction of a polyclonal PHV depolymerase antiserum against PHASCL depolymerases in 1% agarose is shown. The central well (well 1) contained 10 µg of a PHV depolymerase antiserum (IgG fraction). The other wells contained 8 to 10 µg of purified recombinant PhaZ6 depolymerase (wells 7, 2, and 3), purified wild-type PHV depolymerase (well 4), purified PHB depolymerase A (30 µg, well 5), or purified PHB depolymerase B (30 µg, well 6). The reaction was run to completion for 18 h at room temperature, before the agarose was stained with Coomassie blue.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

P. lemoignei is unique among PHA-degrading bacteria because of its high number of extracellular PHA depolymerases (at least six) and its ability to degrade the homopolyester PHV in addition to PHB. Most other PHB-degrading bacteria are not able to degrade PHV and usually have only one PHB depolymerase. Five PHB depolymerase genes of P. lemoignei (phaZ1 to phaZ5) and the corresponding proteins have been described in earlier studies (3, 5, 10, 11, 18; for a recent review, see reference 9). In this report, we studied a novel PHA depolymerase of P. lemoignei which has high activities with PHV (and PHB) as a substrate and which is specifically synthesized during growth on PHV or odd-numbered carbon sources, such as valerate (PHV depolymerase). The structural gene of the PHV depolymerase, phaZ6, was undoubtedly identified by (i) complete agreement of the N-terminal amino acid sequence of the purified wild-type PHV depolymerase and the N-terminal amino acid sequences of four internal endoproteinase GluC-generated peptides with the DNA-deduced amino acid sequence of phaZ6 (Fig. 2), (ii) by cross-reaction of purified recombinant PhaZ6 protein with PHV depolymerase-specific antibodies (Fig. 5), and (iii) by the same ratio of PHV and PHB depolymerase activity of the purified recombinant enzyme and the wild-type PHV depolymerase (Table 2). PhaZ6 showed significant amino acid sequence homologies to all 14 PHB depolymerases known from databases (25 to 50% identity). PhaZ1 of P. lemoignei is the only PHB depolymerase, except PhaZ6, that has any significant activity with PHV (Table 2). However, the degree of amino acid homology of PhaZ6 to PHB depolymerase C (PhaZ1) was lower than to any other depolymerase of P. lemoignei. Therefore, it is impossible to correlate particular amino acids or amino acid sequences of PhaZ1 and PhaZ6 with the ability to utilize PHV as a substrate.

Comparison of the primary amino acid sequences of all available PHA_SCL depolymerases revealed that mature PHA_SCL depolymerases generally consist of three domains in the following sequential order: (i) a catalytic domain including a catalytic triad of serine, aspartate, histidine, and the oxyanion histidine; (ii) a linking domain between the catalytic domain and the C-terminal domain; and (iii) a C-terminal substrate-binding domain (Fig. 2). Each domain can be present in different isoforms: two types of catalytic domains are differentiated by the primary sequence and by a different sequential order of the lipase box and the other catalytic triad amino acids. The sequential order of the catalytic amino acids of all six P. lemoignei PHA depolymerases and of the depolymerases of Alcaligenes faecalis AE122, A. faecalis T1, Ralstonia pickettii, and Pseudomonas stutzeri is (i) histidine (oxyanion), (ii) lipase box serine, (iii) aspartate, and (iv) histidine (Fig. 3). These depolymerases degrade PHB to a mixture of 3-HB and oligomers of 3-HB. The sequential order in the other PHB depolymerases (Acidovorax sp., Comamonas sp., C. acidovorans, C. testosteroni, and S. exfoliatus) is lipase box serine, aspartate, histidine, and oxyanion hole histidine (Fig. 3). The catalytic domains of these depolymerases share relatively high amino acid homologies to each other, but have rather low similarities to the catalytic domains of all other PHB depolymerases. As far as has been analyzed, these depolymerases have high internal 3-HB dimer hydrolase activity in addition to the depolymerase activity. As a consequence, monomeric 3-HB is the major end product of PHB hydrolysis in the presence of an excess of these PHB depolymerases, and the amount of oligomeric end products is relatively low. At present, it is unclear whether the composition of the PHB hydrolysis end products can be generally predicted from the sequential order of the catalytic triad amino acids.

Three types of linking domains have been identified in bacterial PHA depolymerases, namely (i) threonine-rich region, (ii) fibronectin type 3 domain (Fn3), and (iii) a cadherin-like domain (Cad). Threonine-rich regions, which consist of an amino acid sequence approximately 40 amino acids in length highly enriched in threonine and other amino acids with hydroxylated and/or small side chains, have been found only in P. lemoignei PHA depolymerases PhaZ1, PhaZ2, PhaZ3, PhaZ5, and PhaZ6. Similar domains with either a high degree of serine residues or with a high proportion of glycine, alanine, serine, threonine, and proline are known to occur in many polymer hydrolases (6). These domains are assumed to constitute a highly flexible linker between adjacent domains and might facilitate the hydrolysis of neighboring polymer chains, while the enzyme remains bound to the substrate via its substrate-binding domain. Fn3 domains have been found in the PHB depolymerases of Acidovorax sp., A. faecalis T1 and AE122, Comamonas sp., C. acidovorans, C. testosteroni, and S. exfoliatus and in PhaZ4 of P. lemoignei. They also occur in other polymer hydrolases, such as chitinases, cellulases, pullulanases, and amylases (15). Recently, a depolymerase with a novel type of linking domain with high amino acid homology to a cadherin domain has been described for the P. stutzeri PHB depolymerase (21). Since the Fn3 domain of the A. faecalis T1 PHB depolymerase was successfully replaced by a threonine rich region without any significant effect on activity, whether the linking region has any additional sequence-specific function is questionable (19).

The C-terminal part of all PHA_SCL depolymerases consists of a PHA-specific substrate-binding domain. This has been shown experimentally for the PHB depolymerases of P. lemoignei (PhaZ4), A. faecalis, C. acidovorans, C. testosteroni, and P. stutzeri (1, 2, 21, 25). Two types of substrate-binding domains can be differentiated by sequence alignment (Fig. 4). However, some amino acids, including the motif HxxAGR* (with x and * representing an amino acid with a variable or hydrophobic side chain, respectively) are conserved in both types of binding domains. Recently, two PHB depolymerases with two substrate-binding domains have been described for A. faecalis AE122 and P. stutzeri (13, 20). Interestingly, both subtypes of substrate-binding domains were found in the A. faecalis AE122 depolymerase (Fig. 4). The selective advantage of two substrate-binding domains is not known.

The highest degrees of amino acid homology between (mature) PHA_SCL depolymerases were found for the PHB depolymerase of Comamonas sp. and C. testosteroni (98% identical amino acid); for the depolymerases of A. faecalis T1 and R. pickettii (88%); and for the depolymerases PhaZ1, PhaZ2, and PhaZ3 of P. lemoignei (69 to 74% identity). The high degrees of similarity could be explained by the close relatedness of both Comamonas strains and of A. faecalis T1 and R. pickettii or by gene duplication in the case of P. lemoignei. For other depolymerases, high degrees of amino acid identity were found only for selected domains. For example, the catalytic domains and substrate-binding domains of the A. faecalis T1 and the PhaZ5 PHB depolymerase of P. lemoignei shared 72 and 68% identity, respectively, but both depolymerases had different types of linking domains (Fn3 domains versus threonine-rich region). A comparable result was obtained by alignment of the PHV depolymerase PhaZ6 with the PHB depolymerase PhaZ4. In this case, only the first 200 amino acids of the catalytic domains were highly identical (73%). In particular, 26 of the 28 N-terminal amino acids of the mature proteins were identical (93% identity; Fig. 1). A similar result was found even for depolymerases from not related bacteria, e.g., S. exfoliatus and PhaZ5 of P. lemoignei: relatively high homologies were found for the substrate-binding domains (60%), but the catalytic domain and the linking domain belonged to different types (Fig. 2). It appears as if the bacteria have exchanged and randomly combined selected domains of PHA_SCL depolymerases during evolution. This assumption is supported by experimental data in at least one case: the codon usage of phaZ6 was slightly different for the codons Asp, Asn, and His from those of other genes that have been sequenced from P. lemoignei. The AAT and AAC codons (Asn) occurred 21 and 10 times in phaZ6, corresponding to 68 and 32%, respectively. The average values for all other (eight) genes of P. lemoignei that have been sequenced amounted to 42% (AAT) and 58% (AAC). This apparent difference in the frequency of both codons was mainly caused by the high frequency of AAT within the DNA sequence of the catalytic domain (17 codons [corresponding to 74%], compared to only 6 AAC codons [26%]). The distributions of both codons within the linker and substrate-binding domain were different (four AAT and four AAC codons, respectively, each corresponding to 50%). Apparently, the levels of codon usage can be different in different domains of multidomain peptides and might indicate domain exchange.