YfcX Enables Medium-Chain-Length Poly(3-Hydroxyalkanoate) Formation from Fatty Acids in Recombinant Escherichia coli fadB Strains

Abstract

Expression of Escherichia coli open reading frame yfcX is shown to be required for medium-chain-length polyhydroxyalkanoate (PHA_MCL) formation from fatty acids in an E. coli fadB mutant. The open reading frame encodes a protein, YfcX, with significant similarity to the large subunit of multifunctional β-oxidation enzymes. E. coli fadB strains modified to contain an inactivated copy of yfcX and to express a medium-chain-length synthase are unable to form PHA_MCLs when grown in the presence of fatty acids. Plasmid-based expression of yfcX in the FadB⁻ YfcX⁻ PhaC⁺ strain restores polymer formation. YfcX is shown to be a multifunctional enzyme that minimally encodes hydratase and dehydrogenase activities. The gene encoding YfcX is located downstream from yfcY, a gene encoding thiolase activity. Results of insertional inactivation studies and enzyme activity analyses suggest a role for yfcX in PHA monomer unit formation in recombinant E. coli fadB mutant strains. Further studies are required to determine the natural role of YfcX in the metabolism of E. coli.

Polyhydroxyalkanoates (PHAs) are a class of biopolyesters that are receiving considerable attention for use as renewable plastics in packaging and personal hygiene items (40). A broad spectrum of properties can be obtained from PHAs by varying the monomer unit composition (38, 40). PHAs containing short-chain-length monomer units are thermoplastics, whereas PHAs containing medium-chain-length monomer units are elastomeric. While considerable progress has been made toward understanding and manipulating pathways for the formation of short-chain-length PHAs (23, 34, 38), such as poly-3-hydroxybutyrate and poly-3-hydroxybutyrate-co-3-hydroxyvalerate, our understanding of the biosynthesis of medium-chain-length PHAs (PHA_MCL) is not as extensive. Escherichia coli engineered with a PHA synthase possessing substrate specificity for medium-chain-length monomer units will only produce significant levels of PHA_MCL when grown in the presence of fatty acids if the activity of the fatty acid β-oxidation complex has been disrupted by mutation or chemical inhibition (37). Since β-oxidation activities are required to process longer-chain fatty acids to PHA_MCL monomer units (Fig. 1), the requirement of a disrupted fatty acid β-oxidation complex for PHA_MCL formation is counterintuitive.

FIG. 1.

Proposed pathways for PHA_MCL formation from fatty acyl-CoAs in recombinant E. coli fadB strains engineered with a PHA_MCL synthase. Reactions 1 through 4 indicate enzymes that may participate in the formation of (R)-3-hydroxyacyl-CoA from β-oxidation pathway intermediates. Reactions: 1, (R)-β-ketoacyl-CoA reductase (31); 2, 3-ketoacyl-acyl carrier protein reductase (FabG; 39); 3, 3-hydroxyacyl-CoA epimerase (activity of the FadB multienzyme; 28); 4, (R)-enoyl-CoA hydratase (31). Question marks refer to activities that have not been demonstrated in E. coli.

In the study described here, we examined the ability of fadB mutants of E. coli to produce PHA_MCL from longer-chain fatty acids. FadB encodes the α subunit of a multienzyme complex that is involved in the degradation of fatty acids in E. coli (4). It has been previously suggested (37) that FadB activities are involved in PHA_MCL formation in E. coli strain LS1298, a fadB mutant of E. coli, despite reports (9) that the strain is devoid of β-oxidation activities. These unexplained results encouraged us to evaluate the ability of fadB strains to produce PHA_MCL from longer-chain fatty acids and to search for alternative activities that may play a role in PHA_MCL formation.

Searches of the E. coli nucleotide sequence database identified enzymes that could contribute to PHA_MCL monomer unit formation in a fadB mutant on the basis of their homology to β-oxidation enzymes. Candidate genes were identified and isolated, and their enzyme activities were characterized. Enzymatic analyses and insertional inactivation studies suggest a requirement for yfcX, a previously uncharacterized open reading frame in E. coli, for PHA_MCL formation from fatty acids in fadB mutants.

MATERIALS AND METHODS

General methods.

Protein concentrations were determined by the Bradford dye-binding procedure (5) with bovine serum albumin as the standard and protein assay solution from Bio-Rad (Hercules, Calif.). Acetoacetyl coenzyme A (CoA) and saturated acyl-CoA substrates were purchased from Sigma (St. Louis, Mo.). Standards of 3-hydroxyalkanoic acids for gas chromatography (GC) analysis were obtained from Matreya, Inc. (Pleasant Gap, Pa.).

Bacterial strains and medium.

The bacterial strains, plasmids, and primers used in this study are listed in Table 1. Dehydrated LB (Luria-Bertani) broth (Miller), Bacto 2× YT, and Bacto M9 minimal salts (5×) were purchased from Difco Laboratories (Detroit, Mich.). Palmitic and decanoic acid stock solutions contained 2.5% (wt/vol) fatty acid and 10% (wt/vol) Brij-58, pH 7.0, and were sterilized by autoclaving before use. Room temperature stock solutions of palmitic acid were briefly heated at 50°C before addition to the culture medium to create a homogeneous solution. Antibiotics were added to culture medium where indicated at the following concentrations: chloramphenicol, 25 mg/liter; ampicillin, 100 mg/liter; kanamycin, 50 mg/liter; tetracycline, 15 mg/liter.

TABLE 1.

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Relevant characteristic(s)	Reference
Strains
E. coli
MBX763	MG1655 fadB::kan	This study
MG1655		Roberto Kolter, Harvard Medical School
LS1298	fadB::kan (strain originally called SR1 in early literature)	9
DH1	recA	Life Technologies
MBX245	LJ14 rpoS::Tn10 Nalr	Metabolix laboratory strain
LJ14	atoC512 atoA514 fadR601	CGSC 6939
1106	rpoS::Tn10	A. Eisenstark
DH5α	recA Δ(lacZ)M15	Life Technologies
MBX2014	LS1298 yfcX::tet; tetracycline fragment is inactive in chromosome	This study
S. cerevisiae Fy-2	matα ura3-52	Fred Winston, Harvard Medical School
Plasmids
pKPS1.2	1.68-kb phaC1 from pPOB10 in pKK223-3	This study
pPOB10	P. oleovorans phaC1 and open reading frame 2 in pBluescript SK(+)	Peoples and Sinskeya
pKK223-3	4.58-kb vector, pMB1 replicon, tac promoter, Ampr	Amersham Pharmacia Biotech
pTRCN-KPS1.2	1.68-kb phaC1 from pKPS1.2 in pTRCN	This study
pTRCN	4.17-kb vector, pMB1 replicon, trc promoter, lacIq, Ampr	11
pTRCN-KPS1.2N	1.72-kb P. oleovorans phaC1 with E. coli RBS in pTRCN	This study
pSU18-KPS1.2N	1.72-kb phaC1 with E. coli RBS in pSU18	This study
pSU18	2.3-kb vector, p15A replicon, lac promoter, Cmr	1
pTRCN-f714	2.1-kb yfcX in pTRCN	This study
pTRCABV	3.44-kb fragment containing yfcX and yfcY in pTRCN	This study
pTRCAV	1.34-kb fragment containing yfcY in pTRCN	This study
pYfcYX	3.44-kb fragment containing yfcY and yfcX in pTRCN	This study
pN-A5	1.14-kb N-terminal fragment of yfcY in pTRCN	This study
pTRCNfadBA#6	5.2-kb fadBA fragment from pCEM in pTRCN	This study
pCEM	9.4-kb fadBA fragment in plasmid pACYC177	35
pTRCN-CEMA	9.4-kb fadBA fragment from pCEM in pTRCN	This study
pTRCNfadBKanA#1	6.5-kb fadB-kanA fragment from LS1298 genomic DNA in pTRCN	This study
pSU18 MF/NA #1	6.5-kb fadB-kanA fragment from LS1298 genomic DNA in pSU18	This study
pTRCNf714tet	3.6-kb yfcX disrupted with tet in pTRCN	This study
pACYC184	Source of tet cassette for pTRCNf714tet	6
pMAKf714tet-A	3.6-kb yfcX tet fragment from pTRCN-f714tet in pMAK705	This study
pMAK705	Temperature-sensitive replicon, lac promoter, Cmr	13
pTRCNPOX1A	2.3-kb pox1 in pTRCN	This study
Primers
Posyn1-N	5′ CCGAATTCAGGAGGTTTTTATTATGAGTAACAAGAACAACGATGAGCTG 3′
Posyn1-nrSacII	5′ TTGGTCGGAGCCATGGCTTCGGTCATCAGG 3′
n-f714	5′ CCGAATTCTGGAGGCGGAATAATGG 3′
c-f714	5′ CGGGATCCTTATTGCAGGTCAGTTGCAG 3′
A6V.up	5′ GGAATTCGGATCCATGGGTCAGGTTTTACCGCTGGTTACC 3′
A6V.dw	5′ GGAATTCAAGCTTTTATTCCGCCTCCAGAACCATTGCCG 3′
A5-N	5′ CCGAATCAGGAGGTTTTTATTATGGGTCAGGTTTTACCGCTGGTTACCC GCCAGGG 3′
YfcY-nr-ClaI	5′ GGGATGTGCACGCCCCAGTGCTTCACGAGCAAA 3′
YfcX4	5′ CGCGGCAGCGGTGATCCTGATGACTGAATCC 3′
c-f714	5′ CGGGATCCTTATTGCAGGTCAGTTGCAG 3′
N-pox1	5′ GGGGTACCAGGAGGTTTTTATTATGACGAGACGTACTACT 3′
C-pox1	5′ GCTCTAGATCACTTGTTTATTTTCGATAA 3′
f714seq7	5′ AAACGGGATTAGCGCAGGGCACCAGC 3′
f714seq8	5′ TCGGTAAAGATGACGTATTCGAAGC 3′

^aInternational patent application WO 91/00917.

Strain MBX763 was constructed via transduction of strain MG1655 with a P1 lysate propagated from E. coli LS1298. Cells resistant to kanamycin were selected for further study. Strain MBX763 was unable to grow after 72 h on E2 minimal plates (21) supplemented with 5 mM oleate, 10 mM octanoate, or 20 mM butyrate.

Genetic manipulations.

All PCR fragments were sequenced and found to contain wild-type sequences except where noted.

(i) pKPS1.2.

A 1.68-kb BamHI/PstI fragment containing phaC1 of P. oleovorans with its native ribosome binding site (RBS) was isolated from pPOB10 (O. P. Peoples and A. J. Sinskey, international patent application WO 91/00917, January 1991) and ligated into vector pKK223-3 (Amersham Pharmacia Biotech, Piscataway, N.J.).

(ii) pTRCN-KPS1.2.

The 1.68-kb phaC gene was excised from pKPS1.2 with BamHI and HindIII and cloned into pTRCN (11).

(iii) pTRCN-KPS1.2N.

A 0.43-kb fragment was amplified by PCR from plasmid pTRCN-KPS1.2 with primers Posyn1-N and Posyn1-nrSacII. The PCR fragment, containing an E. coli RBS upstream of the N-terminal fragment of phaC1, was digested with EcoRI and cloned into the EcoRI/SmaI sites of vector pTRCN as an EcoRI/blunt fragment. The 1.72-kb intact synthase gene was reconstructed by subcloning a 1.33-kb SacII/HindIII C-terminal synthase fragment from pTRCN-KPS1.2 behind the PCR fragment.

(iv) pSU18-KPS1.2N.

The 1.72-kb EcoRI/HindIII phaC1 fragment of pTRCN-KPS1.2N was excised and cloned into pSU18 (1).

(v) pTRCN-f714.

The yfcX gene was isolated by PCR from DH1 genomic DNA with primers n-f714 and c-f714. The 2.1-kb PCR fragment was digested with EcoRI and BamHI and cloned into pTRCN.

(vi) pTRCABV.

The yfcY gene was isolated by PCR from E. coli MBX245 genomic DNA with primers A6V.up and A6V.dw. The 1.34-kb PCR fragment was digested with BamHI and HindIII and cloned into pTRCN, forming pTrcAV. The yfcY gene was isolated from pTrcAV as a BamHI/HindIII fragment and inserted into the BamHI/HindIII sites of pTRCN-f714, forming pTRCABV. Plasmid pTrcABV contains the trc promoter, followed by yfcX and yfcY.

(vii) pYfcYX.

Plasmid pYfcYX contains the trc promoter, followed by yfcY and yfcX. The 3.44-kb yfcYyfcX insert of the plasmid was constructed in a two-step procedure. A 1.14-kb N-terminal portion of the yfcY fragment was isolated by PCR from E. coli DH1 genomic DNA with primers A5-N and YfcY-nr-ClaI. The fragment was digested with EcoRI and subcloned into the EcoRI and SmaI sites of vector pTRCN, forming 5.28-kb intermediary plasmid pN-A5. A 2.63-kb fragment containing the C terminus of yfcY, as well as yfcX, was isolated by PCR from E. coli DH1 genomic DNA with primers YfcX4 and c-f714. The fragment was digested with SmaI and BamHI and subcloned into the SmaI and BamHI sites of pN-A5.

(viii) pTRCNfadBA#6.

A 9.4-kb BamHI fragment was isolated from plasmid pCEM (35) and cloned into the BamHI site of pTRCN, forming intermediary plasmid pTRCN-CEMA. Plasmid pTRCN-CEMA was digested with NaeI and MfeI, and the protruding ends were filled in with Klenow. A 5.2-kb fragment was isolated from an agarose gel and cloned into the SmaI site of pTRCN, forming plasmid pTRCNfadBA#6. The orientation of the fadBA fragment in pTRCNfadBA#6 is suitable for transcription from the trc promoter.

(ix) pTRCNfadBKanA#1.

Genomic DNA isolated from strain LS1298 was digested with MfeI and NaeI, and protruding ends of the digested fragments were filled in with Klenow. The resulting blunt-ended fragment was ligated into the SmaI site of vector pSU18, forming plasmid pSU18MF/NA#1. Plasmid pSU18MF/NA#1 was digested with MfeI and NaeI, and the resulting fragment was cloned into pTRCN that had been digested with SmaI. Plasmid pTRCNfadBKanA#1 was isolated from the kanamycin-resistant transformants. DNA sequencing confirmed that the anticipated 6.5-kb NaeI/MfeI fragment, containing the gene encoding kanamycin resistance inserted into the BalI site of fadB, had been isolated from the chromosome.

(x) pTRCNf714tet.

Plasmid pACYC184 (6) was digested with AvaI and XbaI, and protruding ends of DNA were filled in with Klenow. A 1.50-kb blunt-ended fragment encoding tetracycline resistance was isolated and inserted into the NruI site of yfcX in plasmid pTRCN-f714.

(xi) pMAKf714tet-A.

Plasmid pTRCNf714tet was digested with EcoRI and XbaI, and protruding ends of DNA were filled in with Klenow. A 3.5-kb blunt-end yfcX-tet fragment was isolated. Plasmid pMAK705 (13) was digested with SphI, treated with T4 DNA polymerase, and ligated to the previously prepared yfcX-tet fragment.

(xii) pTRCNPOX1A.

The gene encoding POX1 was isolated by PCR from genomic DNA of Saccharomyces cerevisiae strain FY-2 with primers N-pox1 and C-pox1. The 2.27-kb PCR product was digested with KpnI and XbaI and cloned into pTRCN (11). Bases 1328 and 1329 of the PCR product were found to be inverted compared to the sequence of the S. cerevisiae oxidase coding region reported by Dmochowska et al. (10). This inversion, yielding a serine instead of the predicted threonine, was observed in multiple clones.

Chromosomal disruption of yfcX.

Chromosomal disruption of yfcX in E. coli LS1298 was performed by homologous recombination with plasmid pMAKf714tet-A as previously described (25). Integration candidates were tested for chloramphenicol sensitivity and tetracycline resistance by replicate plating. Two-thirds of the colonies examined were chloramphenicol sensitive, signifying excision of the vector DNA from the chromosome and subsequent loss of the DNA from the cell. None of the examined colonies were found to be tetracycline resistant. Genomic DNA samples of chloramphenicol-sensitive integrants were screened for the presence of a tetracycline-disrupted yfcX allele by PCR with primers f714seq7 and f714seq8. Integrant MBX2014, containing a PCR product of the expected size for the tetracycline-disrupted yfcX gene, was selected for further analysis by Southern blotting. Southern analysis was performed as described by Sambrook et al. (33) by using the 2.17-kb EcoRI/BamHI yfcX fragment from plasmid pTRCN-f714 as a probe. Probes were labeled prior to use with the AlkPhos direct labeling kit (Amersham Pharmacia Biotech) in accordance with the directions supplied by the manufacturer. Southern prehybridizations, hybridizations, and washes were performed at 60°C. Signal detection was achieved with the CDP-Star chemiluminescent signal detection kit (Amersham Pharmacia Biotech) in accordance with the directions supplied by the manufacturer. Restriction maps of chromosomal fragments containing the wild-type yfcX allele were prepared from the 11.40-kb chromosomal fragment of E. coli listed in the GenBank database under accession no. AE000322.

Preparation of substrates.

3-Ketoacyl-CoA compounds were synthesized as follows. 3-Ketoacyl thiophenyl esters were prepared from Meldrum's acid and the appropriate acid chloride by a procedure similar to that reported earlier (27). The 3-ketoacyl thiophenyl ester derivatives were purified via silica gel chromatography with mixtures of ethyl acetate and hexane. 3-Ketoacyl-CoA thioester compounds were prepared by trans-thiol esterification reaction of CoA with the appropriate 3-ketoacyl thiophenyl esters. Typically, 0.3 mmol of the thiophenyl ester was mixed with 0.12 mmol of CoA in 50 mM KH₂PO₄, pH 7.5, containing Triton X (0.1%) and acetonitrile (25 to 50%). The reaction was monitored by C₁₈ reversed-phase analytical high-performance liquid chromatography (HPLC). The reactions were monitored for loss of CoA and formation of the 3-ketoacyl-CoA derivative. Conversions ranged from 20 to 95%, depending upon the chain length of the 3-keto ester. After completion of the reaction, the mixture was acidified with phosphoric acid to pH 3, extracted with ether to remove thiophenol and its disulfide, and purified by solid-phase extraction or preparative C₁₈ HPLC. Products were analyzed by analytical HPLC.

Δ²-Enoyl-CoA substrates were enzymatically synthesized from the corresponding acyl-CoAs with acyl-CoA oxidase purified from E. coli 1106/pTRCNPOX1A. Cells were cultured in 2× YT at 30°C, and recombinant protein expression was induced with 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG) when cells reached an optical density at 600 nm of 0.6. Cells isolated from 1 liter of culture were resuspended in 10 mM KH₂PO₄, pH 7.0, and lysed by sonication. Soluble proteins were loaded onto a Toyo Jozo DEAE fast protein liquid chromatography (FPLC) column (3 by 14 cm) equilibrated with 10 mM KH₂PO₄, pH 7.0 (buffer A). The column was washed with buffer A (200 ml), and acyl-CoA oxidase activity was eluted with a linear gradient of NaCl (0.2 liter plus 0.2 liter, 0 to 500 mM) in buffer A. Fractions were analyzed for acyl-CoA oxidase activity by measuring the increase in A₂₆₃ due to formation of the double bond in an assay mixture (1 ml) containing 0.2 M KH₂PO₄, pH 7.0, 34 μM acyl-CoA, and enzyme. One unit of oxidase activity was defined as the formation of 1 μmol of Δ²-enoyl-CoA (ɛ₂₆₃ = 6,700 M⁻¹ cm⁻¹; 3) per min. Acyl-CoA oxidase-containing fractions were combined, dialyzed (50,000-molecular-weight cutoff) overnight (buffer A, 2 × 3 liters), and concentrated to 1 ml in a Centriprep YM-50 (Millipore, Bedford, Mass.), yielding a precipitate. The precipitate was resuspended, transferred to an Eppendorf tube, and centrifuged (Eppendorf microcentrifuge, 10 min, 4°C). The resulting pellet was washed three times with buffer A and dissolved in 10 mM KH₂PO₄, pH 7.0, containing 20% glycerol. The above procedure typically yielded partially purified protein preparations with specific activities ranging from 4 to 18 U/mg.

Δ²-Enoyl-CoAs were prepared in a reaction mixture (1 ml) containing 0.2 M KH₂PO₄, pH 7; 34 μM saturated acyl-CoA; and 10 μl of acyl-CoA oxidase (total, 0.02 to 0.2 U). The concentrations of saturated acyl-CoA substrates (in 50 mM KH₂PO₄, pH 4.7) were determined at 259 nm (ɛ = 16.4 mM⁻¹ cm⁻¹) (36) prior to the reaction. The progress of the enzymatic synthesis was monitored at 263 nm and the reaction mixture was incubated at room temperature until all of substrate was consumed. The reaction mixture was transferred to a Centricon YM-50 (Millipore), and acyl-CoA oxidase was removed by centrifugation. The liquid flowthrough, containing 0.2 M KH₂PO₄, pH 7, and 34 μM acyl-CoA, was used immediately as the substrate for hydratase assays.

Enzyme activity analysis.

The substrate specificities of the enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and β-keto thiolase enzymes in YfcX or YfcYX expression strains (Fig. 2) were determined in crude extracts prepared from cells cultured at 37°C in 2× YT medium containing ampicillin. Recombinant protein expression was induced with 1 mM IPTG when cells reached an optical density at 600 nm of 0.6. Cells were harvested after a total culture time of 24 h and lysed by sonication. Cell lysates were clarified by centrifugation prior to use in enzyme assays. For thiolase and dehydrogenase assays, cell pellets were resuspended in 1 ml of lysis buffer containing 50 mM Tris, pH 8.2; 1 mM EDTA; 10 mM β-mercaptoethanol; 20% glycerol; and 0.2 mM phenylmethylsulfonyl fluoride. For hydratase assays, isolated cell pellets were resuspended in 2 ml of lysis buffer containing 10 mM KH₂PO₄, pH 7.0.

FIG. 2.

Enzymatic analysis of YfcX and YfcY.

Residual enoyl-CoA hydratase and 3-hydroxyacyl-CoA dehydrogenase activities in strains expressing insertion mutant forms of yfcX or fadB were measured in crude extracts prepared from cells cultured at 30°C in 2× YT medium containing ampicillin. Cultures were induced with 0.4 mM IPTG at an optical density at 600 nm of 0.6 and harvested after a total culture time of 24 h. Cell pellets were resuspended in 2 ml of lysis buffer prepared by dissolving one tablet of complete mini protease inhibitors (Roche Diagnostics, Indianapolis, Ind.) in 10 ml of 10 mM KH₂PO₄, pH 7.0.

Dehydrogenase activity was assayed in reverse by monitoring the decrease in A₃₄₀ due to the conversion of NADH to NAD⁺ (3). A typical assay mixture (1 ml) contained 100 mM KH₂PO₄, pH 7.0; 0.2 mg/ml bovine serum albumin; 0.1 mM NADH; and 30 μM 3-ketoacyl-CoA and was initiated by the addition of enzyme to the assay mixture. One unit of dehydrogenase activity was defined as the loss of 1 μmol of NADH (ɛ₃₄₀= 6,220 M⁻¹ cm⁻¹)/min.

Hydratase activity was assayed by measuring the decrease in A₂₆₃ due to the hydration of the double bond as described by Binstock and Schulz (3). The final assay mixture (1 ml) contained 17 μM Δ²-enoyl-CoA substrate; 0.2 M KH₂PO₄, pH 7.0; and 0.2 mg of bovine serum albumin per ml. One unit of hydratase activity is defined as the loss of 1 μmol of Δ²-enoyl-CoA (ɛ₂₆₃ = 6,700 M⁻¹ cm⁻¹)/min.

Thiolase activity was assayed by monitoring the thiolysis of β-ketoacyl-CoA substrates at 304 nm (ɛ = 16.9 × 10³ cm⁻¹ M⁻¹, acetoacetyl-CoA; ɛ = 15.6 × 10³ cm⁻¹ M⁻¹, 3-ketohexanoyl-CoA; ɛ = 14.5 × 10³ cm⁻¹ M⁻¹, 3-ketooctanoyl-CoA; ɛ = 13.5 × 10³ cm⁻¹ M⁻¹, 3-ketodecanoyl-CoA and 3-ketododecanoyl-CoA [2]) as previously described, with some modifications (8). The assay mixture (1 ml) contained 65 mM Tris, pH 8.2; 50 mM MgCl₂; 62.5 μM CoA; 62.5 μM 3-ketoacyl-CoA; and soluble crude cell extract. One unit of thiolase activity was defined as the loss of 1 μmol of β-ketoacyl-CoA/min.

PHA_MCL accumulation.

Cell cultures for PHA_MCL production were performed with host strain LS1298 (9). A starter culture of freshly transformed cells was cultured at 37°C overnight in a 500-ml Erlenmeyer flask containing 50 ml of LB medium and the appropriate antibiotics. Cells were diluted 1:100 into a 500-ml Erlenmeyer flask containing 50 ml of fresh medium, and the samples were cultured at 37°C until the A₆₀₀ reached 0.6. The cells were induced with 2 mM IPTG, and either palmitic acid or decanoic acid was added to the culture medium at a final concentration of 0.25% (wt/vol). Samples were cultured for an additional 72 h prior to harvest. Isolated cell pellets were washed with 20 ml of 1× minimal M9 salts, isolated by centrifugation, and washed with an additional 10 ml of 1× minimal M9 salts. The cell pellet was dried overnight in a lyophilizer.

The total polymer content of cell samples, as well as the monomer unit composition of the polymer, was determined with a simultaneous extraction-and-butanolysis procedure. Cell samples and 3-hydroxyalkanoic acid standards were heated at 110°C for 2 h in 2 ml of a mixture containing (by volume) 90% 1-butanol and 10% concentrated hydrochloric acid with 2 mg of benzoic acid per ml as an internal standard. The water-soluble components of the resulting mixture were removed by extraction with water (2 ml). The organic phase (1 μl; split ratio, 1:50; overall flow rate, 2 ml/min) was analyzed on an SPB-1 fused silica capillary GC column (30 m; 0.32-mm inside diameter; 0.25-μm film; Supelco; Bellefonte, Pa.) connected to a Hewlett-Packard gas chromatograph with the following temperature profile: 80°C, 2 min; linear temperature gradient of 10°C/min from 80 to 250°C; 250°C, 2 min.

Purification of YfcYX.

E. coli MBX763/pYfcYX was cultured at 30°C in LB medium supplemented with ampicillin and kanamycin. Recombinant protein expression was induced for 24 h with 0.4 mM IPTG when the culture reached an optical density at 600 nm of 0.6. Cell pellets from 2 liters of culture were resuspended in 50 ml of 10 mM KP_i, pH 7, containing complete mini protease inhibitor cocktail (Roche Diagnostics). Cells were lysed by sonication, and the cell lysate was clarified by centrifugation, yielding a soluble protein extract typically containing 11.2 U of hydratase activity per mg with the substrate Δ²-decenoyl-CoA. Soluble proteins were loaded (3 ml/min) onto a Toyo Jozo DEAE FPLC column (3 by 14 cm) that had been previously equilibrated with buffer A (10 mM KH₂PO₄ [pH 7.0], 50 mM NaCl). The column was washed with 200 ml of buffer A, and proteins bound to the column were eluted with a linear gradient of NaCl (0.2 liter plus 0.2 liter, 0 to 500 mM) in buffer A. Fractions (10 ml) containing hydratase activity with the substrate Δ²-decenoyl-CoA were pooled, dialyzed (2 × 3 liters, 10 mM KP_i [pH 7], 4°C, 24 h), and concentrated with a Centricon 30 (Millipore). Glycerol (20%) was added, and the samples were stored at −80°C.

Gel filtration of the purified YfcYX preparation was performed on a Superdex 200 HR 10/30 FPLC column (Amersham Pharmacia Biotech) preequilibrated in buffer C (10 mM KP_i [pH 7], 150 mM NaCl). A standard calibration curve was generated with blue dextran 2000, thyroglobulin (669 kDa), catalase (232 kDa), albumin (67 kDa), chymotrypsinogen A (25 kDa), ferritin (440 kDa), aldolase (158 kDa), ovalbumin (43 kDa), and RNase (13.7 kDa). For molecular weight determinations and enzyme activity analysis, an aliquot of YfcYX was loaded onto the gel filtration column and proteins were eluted (0.25 ml/min) in buffer C. A portion of each fraction was analyzed for hydratase, dehydrogenase, and thiolase with Δ²-decenoyl-CoA, 3-hydroxydecanoyl-CoA, and 3-ketodecanoyl-CoA as substrates.

RESULTS

Polymer formation in LS1298/pSU18-KPS1.2N/pTRCN.

Previous researchers have observed the presence of PHA monomer units with a chain length shorter than that of the initial fatty acid carbon feed in polymer samples isolated from E. coli fadB mutant strains engineered for PHA production (22, 29, 30, 31). The presence of the shorter monomer units suggests that the host strains retain some fatty acid degradation activities despite the disruption of the FadBA fatty acid β-oxidation complex. In this study, E. coli strain LS1298 (9) was chosen as the host to examine both the fatty acid degradation and PHA monomer unit-producing abilities of fadB mutant strains. The polymer formed in strain LS1298/pSU18-KPS1.2N/pTRCN was used as a benchmark with which to compare additional strains, as well as to confirm the ability of LS1298 to degrade fatty acids during PHA_MCL production (22, 30). Plasmid pSU18-KPS1.2N contains the gene encoding P. oleovorans synthase 1 (17) behind a strong E. coli RBS, and plasmid pTRCN is a vector control. For PHA production, cultures of LS1298/pSU18-KPS1.2N/pTRCN cells were grown at 37°C in LB medium with either 0.25% palmitic or decanoic acid as the carbon source. Synthesis of plasmid-encoded proteins was induced with IPTG, and all strains were allowed to accumulate PHA for 72 h. When palmitic acid was utilized as the carbon source, LS1298/pSU18-KPS1.2N/pTRCN produced 9.2% ± 0.6% PHA per unit of dry cell weight (Table 2). The accumulated PHA contained 55.1 ± 1.0 mol% 3-hydroxyoctanoate and 44.9 ± 1.0 mol% 3-hydroxydecanoate (Table 2). When decanoic acid was utilized as the carbon source, LS1298/pSU18-KPS1.2N/pTRCN produced 8.5% ± 0.6% PHA per unit of dry cell weight (Table 2). The accumulated PHA contained 37.7 ±3.1 mol% 3-hydroxyoctanoate and 62.3 ± 3.1 mol% 3-hydroxydecanoate (Table 2). Strain LS1298/pSU18/pTRCN, containing both vector controls, did not produce detectable levels of PHA of any monomer unit composition, suggesting that PHA was formed only in the presence of active synthase. The ability to produce PHA_MCL containing monomer units with differing chain lengths in LS1298/pSU18-KPS1.2N/pTRCN with either palmitic or decanoic acid as the carbon source confirmed the presence of fatty acid degradation activities within LS1298.

TABLE 2.

PHA_MCL accumulation in LS1298 and MBX2014 strains^a

Strain (relative genotype)	Carbon source	PHA content (% DCW)	PHA composition (mol %)
			3HH	3HO	3HD
LS1298/pSU18-KPS1.2N/pTRCN (fadB mutant strain; plasmid PhaC+)	Palmitate	9.2 ± 0.6	0	55.5 ± 1.1	44.5 ± 1.1
LS1298/pSU18-KPS1.2N/pTRCN (fadB mutant strain; plasmid PhaC+)	Decanote	8.5 ± 0.6	0	37.7 ± 3.1	62.3 ± 3.1
MBX2014/pSU18-KPS1.2N/pTRCN (fadB yfcX double mutant strain; plasmid PhaC+)	Palmitate	0
MBX2014/pSU18-KPS1.2N/pTRCN (fadB yfcX double mutant strain; plasmid PhaC+)	Decanote	0
MBX2014/pSU18-KPS1.2N/pTRCN-f714 (fadB yfcX double mutant strain; plasmid PhaC+ YfcX+)	Palmitate	27.7 ± 3.6	19.6 ± 0.9	40.8 ± 0.8	39.6 ± 0.4
MBX2014/pSU18-KPS1.2N/pTRCN-f714 (fadB yfcX double mutant strain; plasmid PhaC+ YfcX+)	Decanote	10.1 ± 0.5	6.7 ± 0.6	32.1 ± 0.6	61.1 ± 0.8

^aPHA content and composition were determined by GC analysis as described in Materials and Methods. Abbreviations: % DCW, percent dry cell weight; 3HH, 3-hydroxyhexanoic acid; 3HO, 3-hydroxyoctanoic acid; 3HD, 3-hydroxydecanoic acid.

Analysis of enzymatic activities in the disrupted fadB gene from LS1298.

Residual activities remaining in a disrupted β-oxidation complex might contribute to the fatty acid degradation observed during polymer formation in strain LS1298/pSU18-KPS1.2N/pTRCN. While it has been proposed that the disrupted fadB allele in LS1298 contains a defective yet active dehydrogenase, as well as functional hydratase and epimerase activities (37), no experimental data has been provided to support these claims. In fact, LS1298 has been reported to be deficient in β-oxidation activities when assayed with four-carbon-unit substrates (9). To determine if residual fadB activities are encoded by the disrupted fadBkan gene of LS1298, a genomic DNA fragment encoding the fadBkan gene was isolated and cloned into E. coli expression vector pTRCN. The resulting plasmid was named pTRCNFadBKanA#1. A control plasmid, pTRCNFadBA#6, was also constructed by cloning an equivalent fragment encoding the intact FadBA β-oxidation complex into pTRCN.

Plasmids were transformed into host strain DH5α, and dehydrogenase and hydratase activities were measured in crude cell extracts with trans-2-hexenoyl-CoA and 3-ketodecanoyl-CoA, respectively, substrates identified as having the preferred chain length for each activity in purified preparations of FadBA (3). Equivalent dehydrogenase activities were observed in strain DH5α/pTRCNFadBKanA#1, expressing the disrupted β-oxidation complex, and DH5α/pTRCN, containing the vector alone, yielding 0.36 and 0.37 U/mg, respectively. The dehydrogenase activity of strain DH5α/pTRCNFadBA#6, expressing the native β-oxidation complex, was substantially higher, yielding 12.5 U/mg. Hydratase activity in strain DH5α/pTRCNFadBKanA#1 (0.8 U/mg) was slightly higher than the activity observed in strain DH5α/pTRCN (0.3 U/mg) but significantly lower than the hydratase activity measured in the undisrupted complex in strain DH5α/pTRCNFadBA#6 (16.5 U/mg).

Identification of genes with homology to β-oxidation enzymes.

The absence of dehydrogenase activity and the minimal levels of hydratase activity observed in strain DH5α/pTRCNFadBKanA#1 suggest that additional enzymes may participate in fatty acid degradation during polymer formation in LS1298/pSU18-KPS1.2N/pTRCN. To determine if E. coli contains other enzymes capable of catalyzing the degradation of fatty acids, the amino acid sequence of the rat long-chain enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase α subunit of the mitochondrial trifunctional protein (20; GenBank accession no. D16478) was used to screen the E. coli nucleotide sequence database with the TBLASTN program from The National Center for Biotechnology Information. The rat trifunctional protein was chosen because it encodes a dehydrogenase activity as well as a long-chain hydratase activity, an enzyme whose gene has yet to be identified in E. coli (4, 7). Open reading frame f714, corresponding to yfcX on the physical map of E. coli K-12 (32), was found to be 39.5% identical to the gene encoding the rat trifunctional protein. YfcX is an uncharacterized protein that has been recently classified as a member of the crotonase superfamily of enzymes on the basis of its amino acid sequence homology to the rat mitochondrial crotonase (12). Alignments of the amino acid sequence of YfcX with that of FadB, the previously identified α subunit of the E. coli FadBA β-oxidation complex (2, 3, 9), were prepared to determine if active-site amino acid residues in FadB are conserved in YfcX. YfcX (714 amino acids) is 34% identical to FadB (729 amino acids), and its alignment spans the entire FadB amino acid sequence. In addition, amino acid residues important for hydratase and dehydrogenase activities in FadB are conserved in YfcX. G116 and E139, active-site residues for enoyl-CoA hydratase and 3-hydroxyacyl-CoA epimerase activity in FadB (41, 42), are conserved as G115 and E140, respectively, in YfcX. E119, a residue that is important for enoyl-CoA hydratase activity in FadB (14), is conserved as E118 in YfcX. H450 and E462, active-site amino acids for 3-hydroxyacyl-CoA dehydrogenase activity in FadB (15, 16), are conserved as H446 and E458, respectively, in YfcX.

Analysis of enzyme activities of proteins with homology to β-oxidation enzymes.

YfcX was identified as a possible candidate for the fatty acid-degrading activities observed during polymer production in LS1298 strains on the basis of its homology to enzymes associated with β-oxidation. To test the activities of the candidate enzymes, the gene encoding YfcX was isolated by PCR from genomic DNA and cloned into E. coli expression vector pTRCN (11). The overexpression plasmid was transformed into host strain DH5α and cultured in 2× YT medium at 37°C as described in Materials and Methods. Selected intermediates from the β-oxidation of saturated fatty acids were employed as substrates for the enzyme assays. With this approach, plasmid-based expression of YfcX from strain DH5α/pTRCN-f714 yielded significant dehydrogenase activity with β-ketoacyl-CoAs of 6, 8, and 10 carbon units, compared to that of control strain DH5α/pTRCN (Fig. 2). The dehydrogenase activity of DH5α/pTRCN-f714 with acetoacetyl-CoA as a substrate was less than that of control strain DH5α/pTRCN. Significant hydratase activity was observed in DH5α/pTRCN-f714, compared to that of control strain DH5α/pTRCN. Activity with Δ²-enoyl-CoA substrates of 6, 8, 10, and 12 carbon units was observed. Minimal activity was observed with crotonyl-CoA.

Chromosomal inactivation of yfcX.

Hydratase and dehydrogenase enzyme assays of YfcX suggest that the multifunctional-enzyme gene may encode activities that account for the fatty acid degradation observed during polymer production in FadB mutant strains engineered with a medium-chain-length synthase. To determine if the activities of YfcX are essential for polymer production in E. coli fadB strains, the yfcX gene was inactivated by insertion of a fragment encoding tetracycline resistance into its coding sequence. The resulting plasmid, pTRCN-f714tet, was assayed for dehydrogenase and hydratase activities in host strain DH5α with 3-ketodecanoyl-CoA as the substrate. Dehydrogenase activities of 0.20 and 0.23 U/mg were observed for strains DH5α/pTRCN and DH5α/pTRCN-f714tet, respectively, suggesting that an inactivated dehydrogenase was created upon disruption of yfcX. Hydratase activities of 0.44 and 0.1 U/mg were observed for strains DH5α/pTRCN and DH5α/pTRCN-f714tet, respectively, suggesting an inactivated hydratase activity. The disrupted yfcX gene was transferred to the chromosome of LS1298 by homologous recombination of plasmid pMAKf714tet-A. The disruption of yfcX in the chromosome of LS1298 was confirmed by Southern blotting in integrant candidate MBX2014 (Fig. 3).

FIG. 3.

Maps comparing restriction enzyme sites in the 12.92-kb predicted chromosomal fragment of MBX2014 (a) and the 11.40-kb predicted chromosomal fragment of LS1298 (b). (c) Southern blot of digested genomic DNA preparations from strains MBX2014 and LS1298. Each lane contains 2 μg of genomic DNA digested with the indicated enzymes. Lanes: 1, MBX2014 genomic DNA digested with SapI, HindIII, and SacII; 2, MBX2014 genomic DNA digested with SapI and SacII; 3, LS1298 genomic DNA digested with SapI, HindIII, and SacII; 4, LS1298 DNA digested with SapI and SacII.

Analysis of PHA production in a fadB yfcX mutant strain.

An inactivated yfcX gene in strain LS1298 provides a useful tool with which to study the role of YfcX during polymer formation in fadB mutant strains. Strain MBX2014/pSU18-KPS1.2N/pTRCN, the fadB yfcX double-mutant host strain with plasmid-based expression of PHA_MCL synthase and a pTRCN vector control, failed to accumulate any polymer when either palmitic or decanoic acid was utilized as the carbon source. This observation suggests a requirement for activities associated with yfcX for polymer production in E. coli fadB mutant strains. Plasmid-based expression of yfcX from pTRCN-f714 in strain MBX2014/pSU18-KPS1.2N/pTRCN-f714 was found to restore polymer formation, yielding 27.7% ± 3.6% (dry cell weight) PHA when palmitic acid was the carbon source and 10.1% ± 0.5% (dry cell weight) PHA when decanoic acid was utilized as the carbon source (Table 2).

Identification of a gene encoding thiolase activity upstream of yfcX.

Enzymatic assays of strains overexpressing YfcX revealed elevated levels of hydratase and dehydrogenase activities compared to those of wild-type strains. These enzymes are often associated with fatty acid β-oxidation complexes. Interestingly, YfcY, encoded by a gene located upstream of yfcX, was previously assigned as a putative thiolase in homology searches of the E. coli nucleotide sequence database in our laboratory (L. Madison, unpublished data). To determine if YfcY indeed possesses thiolase activity, crude extracts prepared from strain DH5α/pTrcABV, encoding YfcX and YfcY, were assayed. DH5α/pTrcABV contained significant thiolase activity, compared to that of control strain DH5α/pTRCN (Fig. 2). Thiolase activity in DH5α/pTrcABV was observed with β-ketoacyl-CoAs of 6, 8, 10, and 12 carbon units. Little to no thiolase activity was observed when acetoacetyl-CoA was used as the substrate.

Comigration of YfcY and YfcX as a high-molecular-weight complex during gel filtration.

The hydratase, dehydrogenase, and thiolase enzyme activities associated with bacterial fatty acid degradation, such as FadB and FadA of E. coli (2) or FaoA and FaoB of Pseudomonas fragi (19), have been shown to copurify in multienzyme complexes composed of multimers of subunits. Since YfcY and YfcX contain activities related to fatty acid degradation, purification of the enzymes coexpressed in strain DH5α/pYfcYX was performed. Throughout the purification process, hydratase activity was monitored with the substrate Δ²-decenoyl-CoA. Greater-than-ninefold purification of the enzyme, with respect to hydratase activity, was achieved upon passage of the crude MBX763/pYfcYX lysate over a DEAE FPLC column (Table 3). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) indicated that the purified protein fraction contained proteins with sizes equivalent to those of both YfcX and YfcY. The native molecular weights of proteins in the purified preparation were estimated by passing the sample over a Superdex 200 HR 10/30 gel filtration column. Peaks possessing native molecular masses of 435,510 Da (peak P2), 11,950 Da (peak P3), 3,700 Da (peak P4), and 1,460 Da (peak P5) were observed in the gel filtration chromatogram (Fig. 4a). Peak P1 migrated with the void volume of the column. Fractions in peak P2 (435,510 Da) were the only fractions that contained any protein detectable by SDS-PAGE (Fig. 4b). These fractions were found to be enriched in bands close to the predicted sizes of YfcX (77.1 kDa) and YfcY (46.5 kDa). The multiple bands observed may be due to degradation of the protein during purification, since protease inhibitors were not used. Enzyme activity analysis of fractions 7 through 14 demonstrated that thiolase, hydratase, and dehydrogenase activities coeluted in fractions 9 through 13 (Fig. 4c).

TABLE 3.

Partial purification of YfcYX^a

Purification step	Vol (ml)	Enzyme activity, (U/ml)	Total activity (U)	Sp act (U/mg)	Protein concn (mg/ml)	Total protein (mg)	Enzyme recovery (%)	Protein purification (fold)
Crude lysate	45	234.79	10,566	11.2	20.96	943.2	100	1.0
Supernatant	42	206.36	8,667	13.7	15.07	632.9	82	1.2
DEAE pool	40	166.60	6,664	111.1	1.5	60	63	9.9

^aThe progress of the purification was monitored by assaying fractions for hydratase activity with Δ²-decenoyl-CoA.

FIG. 4.

Analysis of protein content and enzyme activity in fractions obtained upon gel filtration of partially purified YfcYX preparations. (a) FPLC gel filtration chromatogram obtained upon loading of 1.08 mg of partially purified YfcYX onto a Superdex 200 HR 10/30 gel filtration column. Calculated molecular masses of eluted peaks: P1, void volume; P2, 435,510 Da; P3, 11,950 Da; P4, 3,700 Da; P5, 1,460 Da. (b) SDS-PAGE analysis of fractions collected by gel filtration chromatography. Samples were subjected to SDS-10% PAGE. Lanes MW contained molecular size markers. (c) Thiolase, hydratase, and dehydrogenase activities in fractions collected by gel filtration chromatography.

DISCUSSION

While previous researchers have demonstrated that significant PHA_MCL formation in genetically engineered E. coli proceeds only when the fatty acid β-oxidation system of the host strain is impaired (37), the ability of E. coli to degrade fatty acyl-CoAs without the activities of the FadB component of its β-oxidation system has not been addressed. Conversion of fatty acyl-CoAs to 3-hydroxyacyl-CoAs with a shorter chain length requires the β-oxidation system to possess acyl-CoA dehydrogenase, enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and β-ketothiolase activities for fatty acid degradation, as well as enoyl-CoA hydratase activity for 3-hydroxyacyl-CoA formation (Fig. 1). Strain LS1298, one of the fatty acid degradation-impaired strains capable of producing PHAs from fatty acids when engineered with a PHA_MCL synthase (22, 30), contains an insertionally inactivated fadB gene that has been reported to encode inactive enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and β-ketothiolase when assayed with enzyme substrates with a four-carbon-unit chain length (9). PHA_MCL formation in recombinant LS1298 strains from fatty acid carbon feeds is therefore unexplained.

In this study, the ability of LS1298 to form PHA monomer units with chain lengths shorter than that of the initial fatty acid feed was examined. First, residual enzyme activities remaining in the disrupted FadB component of β-oxidation were measured. Plasmid-based overexpression of a genomic DNA fragment containing the insertionally inactivated fadB gene from LS1298 yielded no 3-hydroxyacyl-CoA dehydrogenase activity and basal levels of enoyl-CoA hydratase activity when the enzymes were assayed with substrates with the preferred chain length for FadB activities. The absence of 3-hydroxyacyl-CoA dehydrogenase activity suggests that an enzyme other than the disrupted FadB enzyme contributes to the degradation of fatty acyl-CoAs observed during polymer formation. Searches of the translated E. coli genome for candidate enzymes capable of degrading fatty acyl-CoAs identified YfcX and YfcY, amino acid sequences containing homology to known multifunctional β-oxidation enzymes. Enzyme assays of extracts prepared from recombinant E. coli overexpressing YfcX confirmed that the polypeptide possesses enoyl-CoA hydratase activity, as well as 3-hydroxyacyl-CoA dehydrogenase activity. While the substrate specificity of the dehydrogenase component of YfcX is similar to the dehydrogenase activity observed in FadB (2), the hydratase activities of the two enzymes vary significantly. YfcX contains a hydratase activity with a preference for medium-chain-length substrates, whereas FadB contains an activity with a preference for short-chain substrates. The substrate specificity of YfcY is similar to that of the thiolase activity observed in FadA of the E. coli FadBA β-oxidation complex (2). Inactivation of yfcX in LS1298 via insertional mutagenesis yielded a strain that, when engineered with a PHA_MCL synthase, was unable to form PHA_MCL from palmitic or decanoic acid feeds. The ability to form PHA_MCL was restored to the mutant strain upon plasmid-based expression of yfcX. These observations suggest that yfcX encodes a gene product that contributes to the formation of PHA_MCL monomer units from fatty acids with longer chain lengths in strain LS1298.

The formation of PHA_MCL has previously been used as a tool with which to monitor the flow of carbon through the fatty acid β-oxidation pathways of plants (24). Here, we have used the formation of PHA_MCL to detect the presence of enzyme activities that are still capable of degrading fatty acids in E. coli strains in which the FadB component of the FadBA β-oxidation complex has been disrupted. This tool has allowed us to identify the genes responsible for the degradation and to characterize the activities of the encoded enzymes. The enzyme activities of YfcX and YfcY and the comigration of the proteins as a high-molecular-weight complex in gel filtration experiments suggest that YfcX and YfcY possess some similarity to enzymes associated with bacterial fatty acid β-oxidation complexes. The estimated native molecular mass of the YfcYX complex (435 kDa) does not, however, conform to the formation of a heterotetramer model as predicted with the E. coli FadBA (3) and P. fragi FaoAB (18) complexes. While this work was in progress, Olivera et al. (26) reported the presence of two separate β-oxidation pathways in Pseudomonas putida U and suggested that other microbes, including E. coli, may also have more than one set of β-oxidation activities. The need for a second enoyl-CoA hydratase activity for PHA production in E. coli fadB mutants was mentioned by the authors as support for a second set of β-oxidation activities in E. coli. Additional experiments are required to fully characterize the YfcYX complex and to determine if E. coli does indeed possess more than one β-oxidation pathway.