INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Fluorescent pseudomonads belonging to the rRNA homology group I are able to synthesize and accumulate large amounts of polyhydroxyalkanoic acids (PHA)¹ consisting of various saturated 3-hydroxy fatty acids with carbon chain length ranging from 6 to 14 carbon atoms as carbon and energy storage compound (1). PHA isolated from these bacteria contained also constituents with double bonds or with functional groups such as branched, halogenated, aromatic, or nitrile side chains (2). The composition of PHA depends on the PHA synthases, the carbon source, and the involved metabolic routes (2-6). In Pseudomonas putida at least three different metabolic routes occur for the synthesis of 3-hydroxyacyl coenzyme A thioesters, which are the substrates of the PHA synthase (7). (i) -Oxidation is the main pathway when fatty acids are used as carbon source. (ii) Fatty acid de novo biosynthesis is the main route during growth on carbon sources that are metabolized to acetyl-CoA, like gluconate, acetate, or ethanol. (iii) Chain elongation reactions in which acetyl-CoA moieties are condensed to 3-hydroxyacyl-CoA is involved in the PHA synthesis during growth on hexanoate. Recently, recombinant PHA_MCL (MCL = medium chain length) synthesis was also obtained in a -oxidation mutant of Escherichia coli LS1298 (fadB) expressing PHA synthase genes from Pseudomonas aeruginosa (8, 9), indicating that the -oxidation pathway in E. coli provides precursors for PHA synthesis (8). From extended homologies of the primary structures of PHA_MCL synthases to PHA_SCL (SCL = short chain length) synthases (1), which occur in bacteria accumulating poly(3-hydroxybutyric acid) such as e.g. Alcaligenes eutrophus, it seems also likely that the substrate of PHA_MCL synthases is (R)-3-hydroxyacyl-CoA in pseudomonads. The main constituent of PHA of P. putida KT2442 from unrelated substrates such as gluconate is (R)-3-hydroxydecanoate (7, 10, 11). Thus, to serve as substrate for the PHA synthase, (R)-3-hydroxyacyl-ACP must be converted to the corresponding CoA derivative. This can be mediated in a one step reaction by an (R)-3-hydroxyacyl (ACP to CoA) transferase. Another possibility is the release of (R)-3-hydroxydecanoic acid by a thioesterase, and subsequent activation to the CoA derivative. Only few enzymes have been described catalyzing a similar reaction. Examples are the malonyl-CoA-ACP transferase, which catalyzes the transfer of the malonyl moeity from CoA to ACP (12), and (R)-3-hydroxydecanoyl-ACP-dependent UDP-GlcNAc acyltransferase, which catalyzes the transfer of hydroxydecanoyl moeity from ACP to UDP-GlcNAc (13, 14). In this study, we describe the isolation and characterization of P. putida KT2440 mutants, which are defective in the PHA synthesis via fatty acid de novo biosynthesis, and we identified and characterized the gene locus, which phenotypically complements these mutants. The gene product of phaG was purified, and the catalyzed reaction was identified.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Bacterial Strains, Plasmids, and Growth of Bacteria-- Pseudomonads and Escherichia coli strains as well as the plasmids used in this study are listed in Table I. E. coli was grown at 37 °C in Luria-Bertani (LB) medium. Pseudomonads were grown at 30 °C either in nutrient broth complex medium (0.8%, w/v) or in a mineral salts medium with 0.05% (w/v) ammonia (15).

Nitrosoguanidine Mutagenesis-- Mutagenesis was performed according to Miller (16). Cells were incubated for 15 min in the presence of 200 µg of N-methyl-N'-nitro-N- nitrosoguanidine/ml.

Polyester Analysis-- 3-5 mg of lyophilized cell material was subjected to methanolysis in the presence of 15% (v/v) sulfuric acid. The resulting methyl esters of the constituent 3-hydroxyalkanoic acids were assayed by gas chromatography according to Brandl et al. (17) and as described in detail recently (10).

Isolation, Analysis, and Manipulation of DNA-- Plasmid DNA was prepared from crude lysates by the alkaline extraction procedure (18). Total genomic DNA was isolated according to Ausubel et al. (19). All genetic procedures and manipulations of DNA were conducted as described by Sambrook et al. (20). DNA sequencing was carried out by the dideoxy chain termination method (21) with single-stranded or with double-stranded alkali-denatured plasmid DNA but with 7-deazaguanosine 5'-triphosphate instead of dGTP (22) and with -³⁵S-dATP using a T7 polymerase sequencing kit according to the manufacturer's protocol (Amersham Pharmacia Biotech). Synthetic oligonucleotides were used as primers, and the "primer-hopping strategy" (23) was employed. Analysis was done in 8% (w/v) acrylamide gels in buffer, pH 8.3, containing 100 mM hydrochloride, 83 mM boric acid, 1 mM EDTA, and 42% (w/v) urea in a S2-sequencing apparatus (Life Technologies, Inc.). Nucleic acid sequence data and deduced amino acid sequences were analyzed with the sequence analysis software package (version 6.2, June 1990) according to Devereux et al. (24). The nucleotide and amino acid sequence data reported here have been submitted to GenBank^TM under accession number AF052507.

Determination of the Transcriptional Start Site-- Total RNA was isolated as described by Oelmüller et al. (25). The determination of the transcriptional start site was done by a S₁ nuclease protection assay. The hybridization conditions for the S₁ nuclease protection assays were done as described by Berk and Sharp (26) and Sambrook et al. (20), and the S₁ nuclease reactions were conducted as described by Aldea et al. (27). DNA probes and dideoxynucleotide sequencing reactions for sizing the signals were performed with pBluescript SKBH13 DNA as a template. In the annealing reaction, the oligonucleotide (5'-GGGTATTCGCGTCACCT-3') complementary to positions 887 to 871 and the oligonucleotide 5'-CCGCATCCGCGCGATAG-3' complementary to positions 986 to 970, respectively, were used for ³⁵S labeling. For all mapping experiments, 25 µg of RNA was mixed with the labeled DNA fragments (10⁷ cpm/µg of DNA).

Polymerase Chain Reaction-- Polymerase chain reaction amplifications were performed in 100-µl volumes according to Sambrook et al. (20) in an Omnigene thermocycler (Hybaid Ltd., Teddington, U. K.) with Vent polymerase (New England Biolabs GmbH, Schwalbach, Germany). The following oligonucleotides were used as primers to amplify the coding region of phaG to construct plasmids pBHR-QG (derivative of pQE60 (Qiagen), insertion into NcoI/BamHI sites) and pBHR81 (derivative of pBBR1MCS-2 (28), insertion into EcoRI/BamHI sites), respectively: 5'-CATGCCATGGGAAGGCCAGAAATCGCTGTA-3', 5'-CGCGGATCCGATGGCAAATGCATGCTGCCC-3' (pBHR-QG); 5'-CGGAATTCAAGGAGTCGATGACATG-3', 5'-CGCGGATCCCGGCGCCCCGTGGCC-3' (pBHR81). Both plasmids possess artificial ribosome binding sites conserved for E. coli, and transcription is regulated by the lac promoter.

Preparation of Cell Extracts and Electrophoretic Methods-- Approximately 1 g (wet weight) of E. coli cells were suspended in 1 ml of buffer A (50 mM Tris hydrochloride, pH 7.4, 0.8% (v/v) Triton X-100, 10 mM MgCl₂, 10 mM EDTA, which was supplemented with 200 µg of phenylmethylsulfonyl fluoride per ml) and disrupted by sonification for 1 min at an amplitude of 14 µm in a W 250 sonifier (Branson Schallkraft GmbH, Germany). Soluble cell fractions were obtained as supernatants from 30 min of centrifugation at 50,000 × g and 4 °C. SDS- and mercaptoethanol-denatured proteins were separated in 11.5% (w/v) polyacrylamide gels in Tris-glycine buffer (25 mM Tris, 190 mM glycine, 0.1% (w/v) SDS (29) and stained with Coomassie Brilliant Blue (30).

Purification of Recombinant PhaG-His Tag and PhaG-- Recombinant PhaG-(His)₆ tag (C-terminal fusion) was purified from E. coli JM109-harboring plasmid pBHR-QG. Crude extract was subjected to Ni²⁺-nitrilotriacetic acid-agarose and washed twice with 20 mM imidazole, and the PhaG-(His)₆ tag was eluted with 250 mM imidazole. Purified PhaG-(His)₆ tag was used to raise anti-PhaG antibodies. Native PhaG was purified from Pseudomonas oleovorans ATCC 29347-harboring plasmid pBHR81 by native preparative PAGE (14% (w/v) polyacrylamide) applying the PrepCell 491 (Bio-Rad).

Analysis of (R,S)-3-Hydroxyacyl-CoA or ACP Thioester by High Performance Liquid Chromatography (HPLC)-- As a reference substance, (R,S)-3-hydroxydecanoyl-CoA was synthesized using 10 milliunits of acyl-CoA synthetase (Sigma) in 100 µl of 50 mM Tris-HCl, pH 7.5, containing 2 mM ATP, 5 mM MgCl₂, 2 mM coenzyme A, and 2 mM (R,S)-3-hydroxydecanoate. The reaction was stopped by the addition of 5 volumes of Dole's reagent (80% (v/v), 20% (v/v) n-heptane, 0.02 N H₂SO₄), and remaining free fatty acid was extracted with n-heptane. (R,S)-3-Hydroxydecanoyl-ACP was synthesized as described by Rock and Cronan were used (31). HPLC analysis was conducted with a RP18 column (nucleosil C18, 7 µm, Knauer) and 25 mM potassium phosphate buffer pH 5.3 as mobile phase. Thioesters were eluted with increasing acetonitrile gradient and detected with a diode array detector (DAD 540, Kontron) at a spectral range of 200 to 500 nm with a 0.8-nm spectral resolution.

Assay of Transfer of 3-Hydroxydecanoate from CoA to ACP-- The transferase assay was conducted in 100 µl of 50 mM Tris-HCl, pH 7.5, containing 5 mM MgCl₂, 2 mM dithioerythrol, 500 µM acyl carrier protein (Sigma), and 2 mM (R,S)-3-hydroxydecanoyl-CoA with a 100-µg protein of crude extract or 50 µg of purified PhaG. After incubation for 4 h at 37 °C, the reaction was stopped by the addition of Dole's reagent, and the reaction mixture was analyzed by HPLC.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Complementation of Mutants Effected in the PHA Synthesis via de Novo Fatty Acid Biosynthesis-- Mutants of P. putida KT2440, which are only deficient in the metabolic route-linking fatty acid de novo synthesis, were generated with nitrosoguanidine according to Miller et al. (16). Five mutants (PhAG_N) were identified, which accumulated PHA only up to 3% of the cellular dry weight (CDW) from gluconate but up to 85% PHA of CDW when cultivated on octanoate as the sole carbon source. The composition of the polymer was not affected. We constructed a library of EcoRI-digested P. putida KT2440 genomic DNA with the cosmid vector pVK100 (32) and the Gigapack II Gold Packaging Extract (Stratagene Cloning Systems, La Jolla, CA) in E. coli S17-1. Approximately 5,000 transductants were applied to minicomplementation experiments, with mutant PHAG_N-21 as recipient. One of the hybrid cosmids (pVK100::K18) harbored three EcoRI-fragments (3, 6, and 9 kbp) and enabled PHAG_N-21 to accumulate PHA from gluconate. Subcloning revealed that the 3-kbp EcoRI fragment (E3, pMPE3) complemented PHAG_N-21 and any other PHAG_N mutant exhibiting this phenotype. Complementation was not achieved by the hybrid cosmid pHP1016::PP2000 comprising the entire 7.3-kbp PHA synthase locus of P. aeruginosa PAO1 plus approximately 13 kbp of the upstream region or by the hybrid cosmid pHP1016::PP180 comprising the phaC2 gene of P. aeruginosa PAO1 plus approximately 16 kbp of the adjacent downstream region (10).

Determination of the Gene Locus and Nucleotide Sequence of phaG-- Fragment E3 was cloned into pBluescript SK, and the entire nucleotide sequence was determined (Fig. 1). It comprised 3,061 nucleotides with three ORFs (Fig. 1). The only ORF that was completely localized on this fragment was ORF2 with 885 nucleotides starting at position 911 and terminating at position 1795 (Fig. 1). ORF2 will be referred to as phaG. A putative S/D sequence was identified eight nucleotides upstream of the start codon. About 230 bp downstream of the translational stop codon a potential factor-independent transcription terminator was located (Fig. 1). ORF1 and ORF3 are localized only incompletely on E3 with ORF1 lacking the 5'-region and with ORF3 lacking the 3'-region. The amino acid sequence deduced from ORF1 revealed significant homologies to a hypothetical, not further characterized protein of Hemophilus influenzae (33). In contrast, the amino acid sequence deduced from ORF3 did not reveal any significant homology to proteins available from EMBL data base. Several other smaller ORFs were detected. However, none of them did obey the rules of Bibb et al. (34) for a coding region or was preceded by a reliable ribosomal binding site.

View larger version (56K): [in this window] [in a new window]   Fig. 1.   Nucleotide sequence of fragment E3. Amino acids deduced from the nucleotide sequence are specified by standard one-letter abbreviations. The promoter sequence ("10" and "35") is boxed. Putative ribosome binding sites are indicated by black bars and the letters S/D. The position of a tentative factor-independent transcriptional terminator downstream of phaG is indicated by arrows. An arrow starting with a dot indicates the transcription start site and direction of transcription.

Characterization of the phaG Translational Product-- The codon usages in phaG, ORF1 and ORF3 agreed well with typical P. putida codon preferences. The G + C content of 59.2 mol % for phaG was similar to the value of 60.7 to 62.5 mol % determined for total genomic DNA of P. putida (35). The phaG gene encodes a protein of 295 amino acids with a molecular mass of 33,876 Da. Sequence alignments of the amino acid sequence deduced from phaG revealed a 44% overall identity to the rhlA gene product of P. aeruginosa PG201 (Fig. 2). RhlA also consists of 295 amino acids and has a molecular mass of 32.5 kDa. This gene represents the 5'-terminal gene of a gene cluster consisting of the genes rhlA, rhlB, and rhlR. The first two genes encode proteins involved in rhamnolipid biosynthesis. The rhlB gene product exhibited rhamnosyltransferase activity, whereas the function of RhlA is not yet characterized but is necessary for effective rhamnolipid biosynthesis. RhlR represents a transcriptional activator acting upon ⁵⁴-dependent promoters (36). The C-terminal regions of RhlA and PhaG revealed high homology to a gene region (qin) of P. aeruginosa encoding the so-called "quinolone-sensitivity protein" (GenEMBL data library, accession number L02105) amounting to 50.6 and 40.1% to PhaG or to RhlA, respectively, in 249 overlapping residues (Fig. 2). This region comprises 1503 nucleotides. The N terminus of the qin gene was not exactly determined, and the homology as depicted in the data base extents only from nucleotide 207 to 566 of this sequence (Fig. 2). However, translation of this sequence in all six reading frames and a subsequent tBLASTn search resulted in the identification of homologies also in the upstream region of the suggested qin translational start codons but in different reading frames with the N-terminal region of PhaG and RhlA.

View larger version (46K): [in this window] [in a new window]   Fig. 2.   Homology of the phaG gene product to RhlA (40) and the putative qin gene product (GenEMBL data library, accession number L02105) of P. aeruginosa. That part of the amino acid sequence that was deduced from the improved open reading frame analysis of the qin nucleotide sequence is given in lowercase letters. Matching amino acids are boxed. Dashes indicate gaps, which were introduced to improve the alignment. Numbers indicate the positions of the amino acids in the respective proteins.

Identification and Regulation of the Promoter-- 244 bp upstream of phaG, a putative ⁷⁰-dependent promoter structure TTGCGCN₁₇TTGAAT (where N is a nucleoside) was identified. The promoter was verified by complementation studies of mutant PHAG_N-21 with subfragments of E3. The 2.2-kbp SalI-EcoRI subfragment (SE 22, pMPSE22) (Fig. 1, Table I), which lacked the above-mentioned promoter sequence, did not complement this mutant, whereas the 1.3-kbp BamHI-HindIII subfragment (BH13, pBHR75) (Fig. 1, Table I) of E3 conferred the ability to again synthesize PHA from simple carbon sources. In addition, the significance of this putative promoter structure was proved by S₁ nuclease protection with total RNA isolated from gluconate-grown and octanoate-grown cells of P. putida KT2440 harvested in the stationary growth phase. The transcriptional start site was identified 5 nucleotides downstream of the putative promoter consensus sequence at position 673 (Fig. 1, 3). For octanoate-grown cells only an extremely weak RNA signal was detected, whereas a strong signal occurred with RNA isolated from gluconate-grown cells (Fig. 3). This indicated a strong transcriptional induction of phaG under conditions of PHA synthesis via fatty acid de novo biosynthesis.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   S1 nuclease protection assays of the phaG transcripts. Lanes A, C, G, and T, standard sequencing reactions to size the mapping signals. RNA was isolated from gluconate-grown (lanes 1 and 3) or octanoate-grown (lane 4) cells of P. putida KT2440 (lanes 1 and 3) and A. eutrophus H16 (lane 2).

Heterologous Overexpression of phaG in E. coli-- A plasmid expressing a C-terminal His(6) tag fusion protein of PhaG was constructed. The resulting plasmid pBHR-QG enabled overexpression of phaG under lac promoter control in E. coli JM109 (Fig. 4). The fusion protein could only be purified under denaturing conditions by immobilized metal ion affinity (Fig. 5) and was used as antigen to raise antibodies.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Restriction maps of plasmids pBHR-QG (a) and pBHR81 (b). kb, kilobases.

View larger version (101K): [in this window] [in a new window]   Fig. 5.   a, heterologous expression of phaG-His tag in E. coli and purification. Cytoplasmic fractions obtained from cells of recombinant strains of E. coli grown in LB medium and fractions from batch purification with Ni2+-nitrilotriacetic acid-agarose were separated in 11.5% (w/v) polyacrylamide gels and stained to visualize protein with Serva blue R. M, molecular weight standards. Lane 1, crude extract of E. coli JM109 (pQE60); lane 2, crude extract of E. coli JM109 (pBHR-QG); lane 3, eluate after washing with 20 mM imidazole; lane 4, purified PhaG-His tag after elution with 250 mM imidazole. b, heterologous expression of phaG in P. oleovorans und purification of native PhaG. P. oleovorans-harboring pBHR81 was cultivated 16 h at 30 °C on mineral salts medium containing 1% (w/v) gluconate. Crude extracts were applied to native PAGE (PrepCell 491, Bio-Rad), and the first fraction with high absorption at 280 nm yielding purified PhaG was analyzed. M, molecular weight standards. Lane 1, crude extract of P. oleovorans (pBBR1MCS-2); lane 2, crude extract of P. oleovorans (pBHR81); lane 3, first protein eluate from native PAGE containing pure PhaG.

Functional Homologous and Heterologous Expression of phaG-- Functional expression, as revealed by complementation of mutant PHAG_N21, was obtained from plasmid pBHR81, a derivative of vector pBBR1MCS-2 (28) containing the coding region of phaG in sites EcoRI/BamHI (Fig. 4, Table II). Additionally, transfer of pBHR81 into P. oleovorans ATCC 29347, which is not capable of PHA synthesis from simple carbon sources, resulted in PHA accumulation from gluconate contributing to about 55% of CDW (Table II). Thus only functional expression of phaG in P. oleovorans established a metabolic link between fatty acid de novo biosynthesis and PHA synthesis. Expression of phaG in P. aeruginosa PAO1 based on plasmid pBHR81 revealed an ~40% increase in PHA accumulation (Table II). We also investigated functional expression of phaG in E. coli JM109-harboring plasmids pBHR81 and pBHR71 allowing functional expression of PHA synthase gene phaC1 (8), but no PHA accumulation was observed when cells were grown on glucose. Furthermore, transfer of pBHR81 into P. aeruginosa UO299 (rhlA) did not result in complemention of this mutant with respect to rhamnolipid synthesis (data not shown). Thus PhaG does not functionally replace RhlA. To evaluate whether PhaG exhibits PHA synthase activity, we cultivated the P. putida PHAG_N mutants harboring pBHR81 under nonlimited nitrogen conditions, which resulted in decreased PHA synthase levels and decreased PHA accumulation (37). No increase in PHA accumulation was observed when cells were grown on gluconate in the presence of PhaG (data not shown).

Enzymatic Assay of PhaG-- Native PhaG was purified from crude extracts of P. oleovorans (pBHR81) by native PAGE as described under "Experimental Procedures." Recombinant PhaG showed high mobility in native PAGE, which could be utilized for one-step purification (Fig. 5). PhaG was also identified by N-terminal amino acid sequencing.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   HPLC analysis of reaction products from enzymatic assay with PhaG. a, crude extracts from various bacteria harboring either (a) vector pBBR1MCS-2 (negative control) or (b) plasmid pBHR81 were employed for the enzymatic PhaG assay. d, purified PhaG was directly used for the assay (c) with heat-inactivated PhaG as negative control. 3-Hydroxydecanoyl-CoA (3HD-CoA) was provided as substrate, and the transfer of the acyl moeity to ACP was demonstrated (3-hydroxydecanoyl-ACP (3HD-ACP)). Peaks were identified based on their Rf values, by co-chromatography, and by their spectra. The identity of relevant peaks was indicated.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Phenotypical complementation of P. putida KT2440 PHAG_N mutants, which are affected in PHA biosynthesis based on fatty acid de novo biosynthesis, led to the identification and characterization of phaG as a new gene locus relevant for PHA biosynthesis in P. putida. The PHA synthesis pathway via -oxidation was not impaired in the PHAG_N mutants. PHAG_N mutants were not complemented with the PHA synthase locus of P. aeruginosa PAO1 and adjacent genomic region. Therefore, PHAG_N mutants are not defective in the PHA synthase locus, and most probably phaG is not closely linked to the PHA synthase locus. Furthermore, phaG is not in general essential for the synthesis of PHA in P. putida KT2440 but is only required for PHA synthesis and accumulation from gluconate or other simple carbon sources, which are catabolized to acetyl-CoA in this organism before PHA synthesis starts.

From results of labeling studies, nuclear magnetic resonance spectroscopy and gas chromatography-mass spectroscopy Eggink et al. (4) and Huijberts et al. (7, 38) concluded that the precursors of PHA_MCL biosynthesis from simple carbon sources are predominantly derived from (R)-3-hydroxyacyl-ACP intermediates occurring during the fatty acid de novo biosynthetic route. Since the constituents of PHB and PHA represent the R configuration, and since PHA_SCL and PHA_MCL synthases are highly homologous, the intermediates in fatty acid metabolism are presumably converted to (R)-3-hydroxyacyl-CoA before polymerization. Nevertheless, some other routes of PHA synthesis are also possible. Other conceivable alternatives are the release of free fatty acids by the activity of a thioesterase with a thiokinase, subsequently activating these fatty acids to the corresponding hydroxyacyl-CoA thioesters or chain elongation with -ketothiolase, or -oxidation of synthesized fatty acids. Evidence for the latter pathways in P. putida (7) was obtained and explains why phaG mutants are not completely defective in PHA_MCL biosynthesis from gluconate. Functional expression of either PHA synthase and accumulation of PHA_MCL from fatty acids indicate that PHA synthases are not utilizing (R)-3-hydroxyacyl-ACP derivatives as substrate (8, 9).

All mutants analyzed and complemented by phaG synthesized PHA to some extent (0.5-3% CDW) with a typical monomer composition of polyester derived from simple carbon sources, as far as detectable. However, analysis of mutant complementation studies and the genomic organization of phaG revealed no indication for the existence of another protein essential for the PHA synthesis from simple carbon sources in P. putida KT 2440. Therefore, most probably only one additional specific enzymatic step is required for PHA synthesis from gluconate that is not required for PHA synthesis from octanoate. This hypothesis was supported by the observation that only PhaG conferred the ability to synthesize PHA from gluconate to P. oleovorans, which lacks this capability (Table II). Furthermore, the analysis of enzymatic activity of PhaG strongly suggests that one enzyme is sufficient to link fatty acid de novo synthesis with PHA synthesis (Fig. 6). Evidence that PhaG is not directly involved in synthesis of PHA_MCL was provided by cultivations of the P. putida PHAG_N mutants (pBHR81) under nitrogen limited and nonlimited conditions. Under nonlimited conditions the level of PHA synthases and PHA_MCL accumulation is significantly decreased (37), and even in the presence of PhaG, no increase in PHA_MCL synthesis was observed.

Although no complementation of rhamnolipid synthesis in P. aeruginosa rhlA mutant UO299 was obtained with phaG expressed from plasmid pBHR81, the high degree of homology of phaG to rhlA and the qin region of P. aeruginosa, respectively, indicates a related function of these proteins. The exact function of the "quinolone sensitivity protein" has not yet been described. Quinolones such as nalidixic acid are synthetic antibiotics exhibiting strong antimicrobial effects on Gram-negative bacteria including P. aeruginosa. The rhlA gene product is involved in the rhamnolipid biosynthesis of P. aeruginosa PG201, which are synthesized as biosurfactants during the late exponential and stationary growth phases. Rhamnolipid biosynthesis proceeds by sequential glycosyl transfer reactions, each catalyzed by specific rhamnosyltransferases with TDP-rhamnose acting as a rhamnosyl donor, and 3-hydroxydecanoyl-3-hydroxydecanoate or L-rhamnosyl-3-hydroxydecanoyl-3-hydroxydecanoate acting as acceptors as proposed by Burger et al. (39, 40). 3-Hydroxydecanoate can be formed via -oxidation or via fatty acid de novo biosynthesis (41). A dimer consisting of two 3-hydroxydecanoic acid molecules is formed by a hitherto unknown mechanism. RhlA significantly enhanced the level of rhamnolipids in rhamnolipid-negative mutants of P. aeruginosa PG201 when it was coexpressed with the rhamnosyltransferase (RhlB) as compared with the expression of the isolated rhlB gene.

3-Hydroxyacyl-ACP intermediates provided by fatty acid biosynthesis are presumably the common intermediates of PHA and rhamnolipid biosynthesis from gluconate. If the ACP derivatives themselves do not serve as substrates for PHA synthases or enzymes involved in rhamnolipid synthesis for the condensation of two 3-hydroxydecanoyl moieties, they must be either directly transesterified to the corresponding CoA derivatives or transferred to CoA thioesters by the combined action of a thioesterase and a thiokinase. Various transacylases and acyltransferases have been described and well characterized catalyzing the direct transfer of an acyl moiety, e.g. (i) the malonyl-CoA-ACP transferase, which catalyzes the transfer of the malonyl moeity from CoA to ACP (12) and (ii) the hydroxydecanoyl-ACP-dependent UDP-GlcNAc acyltransferase, which catalyzes the transfer of hydroxydecanoyl moeity from ACP to UDP-GlcNAc (13, 14). The bacterial acyltransferase LpxA is one representative of a large family that possesses conserved repeating hexapeptides (42). Sequence analysis of membrane-bound glycerolipid acyltransferases revealed that these proteins share a highly conserved domain containing invariant histidine and aspartic acid residues separated by four less conserved residues in an HX₄D configuration (43). Site-directed mutagenesis of the invariant histidine resulted in lack of activity, indicating an essential role of this residue (43). Although no significant homology of PhaG to transacylases and acyltransferases was found, this highly conserved HX₄D mini-motif is also present in PhaG at positions 176-181 of the amino acid sequence (Fig. 1), suggesting a similar function of PhaG. The studies on heterologous expression of phaG and the enzymatic characterization of PhaG strongly suggests that PhaG catalyzes the conversion of (R)-3-hydroxyacyl-ACP to (R)-3-hydroxyacyl-CoA derivatives (Table II, Fig. 6), which serve as ultimate precursors for the PHA polymerization from unrelated substrates in pseudomonads proposed recently (4, 44).