BIOCHEMICAL AND MOLECULAR CHARACTERIZATION OF THE PSEUDOMONAS-LEMOIGNEI POLYHYDROXYALKANOATE DEPOLYMERASE SYSTEM

Pseudomonas lemoignei has five different polyhydroxyalkanoate (PHA) de polymerase genes (phaZ1 to phaZ5), which encode the extracellularly lo calized poly(3-hydroxybutyrate) (PHB) depolymerases C, B, and D, poly( 3-hydroxyvalerate) (PHV) depolymerase, and PHB depolymerase A, respect ively. Four of the five genes (phaZ1 to phaZ4) have been cloned, and o ne of them (phaZ1) was studied in detail earlier (D. Jendrossek, B. Mu ller, and H. G. Schlegel, Fur. J. Biochem. 218:701-710, 1993). The fif th PHA depolymerase gene (phaZ5) was identified by colony hybridizatio n of recombinant Escherichia coli clones with a phaZ5-specific oligonu cleotide. The nucleotide sequence of a 3,704-bp EcoRI fragment was det ermined and found to contain two large open reading frames (ORFs) whic h coded for a polypeptide with significant similarities to glycerol-3- phosphate dehydrogenases of various sources (313 amino acids; M(r), 32 ,193) and for the precursor of PHB depolymerase A (PhaZ5; 433 amino ac ids; M(r), 44,906). The PHV depolymerase gene (phaZ4) was subcloned, a nd the nucleotide sequence of a 3,109-bp BamHI fragment was determined . Two large ORFs (ORF3 and ORF4) that represent putative coding region s were identified. The deduced amino acid sequence of ORF3 (134 amino acids; M(r), 14,686) revealed significant similarities to the branched -chain amino acid aminotransferase (IlfE) of enterobacteria. ORF4 (1,7 12 bp) was identified as the precursor of a PHV depolymerase (567 amin o acids; M,, 59,947). Analysis of primary structures of the five PHA d epolymerases of P. lemoignei and of the PHB depolymerases of Alcaligen es faecalis and Pseudomonas pickettii revealed homologies of 25 to 83% to each other and a domain structure: at their N termini, they have t ypical signal peptides of exoenzymes. The adjacent catalytic domains a re characterized by several conserved amino acids that constitute puta tive catalytic triads which consist of the consensus sequence of serin e-dependent hydrolases including the pentapeptide G-X-SX-G, a conserve d histidine and aspartate, and a conserved region resembling the oxyan ion hole of lipases. C terminal of the catalytic domain an approximate ly 40-amino-acid-long threonine-rich region (22 to 27 threonine residu es) is present in PhaZ1, PhaZ2, PhaZ3, and PhaZ5, Instead of the threo nine-rich region PhaZ4 and the PHB depolymerases of A. faecalis and P. pickettii contain an approximately 90-amino-acid-long sequence resemb ling the fibronectin type III module of eucaryotic extracellular matri x proteins. The function of the fibronectin type III module in PHA dep olymerases remains obscure. Two types of C-terminal sequences apparent ly represent substrate-binding sites; the PHB type is present in the P HB depolymerases of A. faecalis and P. pickettii and in PhaZ2, PhaZ3, and PhaZ5, and the PHV type is present in the PHV-hydrolyzing depolyme rases (PhaZ4 and PhaZ1). phaZ1 was transferred to A. eutrophus H16 and JMP222. All transconjugants of both strains were able to grow,vith ex tracellular PHB as a carbon source and produced translucent halos on P HB-containing solid media. PhaZ1, PhaZ2, PhaZ4, and PhaZ5 were purifie d from P. lemoignei and from recombinant E. coli; the processing sites of the precursors in E. coli were the same as in P. lemoignei, and si milar substrate specificities were determined for the wild-type and th e recombinant proteins. All PHA depolymerases hydrolyzed PHB at high s pecific activities. PhaZ1 and PhaZ4 additionally cleaved PHV, and PhaZ 4 hydrolyzed poly(4-hydroxybutyrate). None of the depolymerases was ab le to hydrolyze polylactide or PHA consisting of monomers with more th an five carbon atoms. While the wild-type depolymerase proteins were g lycosylated and found to contain glucose and N-acetylglucosamine, none of the recombinant proteins was glycosylated. PHB hydrolysis was depe ndent on divalent cations such as Ca2+ and was inhibited by the presen ce of EDTA.