Engineering the Monomer Composition of Polyhydroxyalkanoates Synthesized in Saccharomyces cerevisiae

Abstract

Polyhydroxyalkanoates (PHAs) have received considerable interest as renewable-resource-based, biodegradable, and biocompatible plastics with a wide range of potential applications. We have engineered the synthesis of PHA polymers composed of monomers ranging from 4 to 14 carbon atoms in either the cytosol or the peroxisome of Saccharomyces cerevisiae by harnessing intermediates of fatty acid metabolism. Cytosolic PHA production was supported by establishing in the cytosol critical β-oxidation chemistries which are found natively in peroxisomes. This platform was utilized to supply medium-chain (C₆ to C₁₄) PHA precursors from both fatty acid degradation and synthesis to a cytosolically expressed medium-chain-length (mcl) polymerase from Pseudomonas oleovorans. Synthesis of short-chain-length PHAs (scl-PHAs) was established in the peroxisome of a wild-type yeast strain by targeting the Ralstonia eutropha scl polymerase to the peroxisome. This strain, harboring a peroxisomally targeted scl-PHA synthase, accumulated PHA up to approximately 7% of its cell dry weight. These results indicate (i) that S. cerevisiae expressing a cytosolic mcl-PHA polymerase or a peroxisomal scl-PHA synthase can use the 3-hydroxyacyl coenzyme A intermediates from fatty acid metabolism to synthesize PHAs and (ii) that fatty acid degradation is also possible in the cytosol as β-oxidation might not be confined only to the peroxisomes. Polymers of even-numbered, odd-numbered, or a combination of even- and odd-numbered monomers can be controlled by feeding the appropriate substrates. This ability should permit the rational design and synthesis of polymers with desired material properties.

Polyhydroxyalkanoates (PHAs) have received considerable interest as renewable-resource-based, biodegradable, and biocompatible plastics with a wide range of potential applications. More than 150 different hydroxyalkanoic acids have been identified as PHA constituents (13, 24). These PHA monomers can be synthesized in a variety of configurations to produce a wide range of material properties. Physiological data and enzymatic studies have shown that there are two distinct classes of PHAs based on the number of carbons in the monomers (9). Short-chain-length PHA (scl-PHA) polymers are composed of monomers containing 3 to 5 carbon atoms, whereas medium-chain-length PHA (mcl-PHA) polymers are composed of monomers containing 6 to 14 carbon atoms (17).

It has been demonstrated recently that mcl-PHA can be synthesized in the peroxisomes of Saccharomyces cerevisiae and Pichia pastoris when the mcl-PHA polymerase from Pseudomonas aeruginosa is expressed and targeted into this organelle (19, 20). However, the yield of peroxisomally produced PHA in yeast is limited by the number and volume of peroxisomes. In addition, induction of peroxisomal β-oxidation is highly sensitive to catabolite repression. However, it is unclear from these previous studies whether PHA synthesis is also possible in the cytosol of these yeasts. Cytosolic expression of PHA pathways bypasses the challenges associated with the peroxisomes and could potentially take advantage of the large carbon fluxes present during growth. It has been shown previously that poly-β-hydroxybutyrate is synthesized in the cytosol of S. cerevisiae if the scl-PHA polymerase from Ralstonia eutropha is expressed in this cell compartment (16). This finding indicates that native S. cerevisiae is capable of synthesizing monomers of the correct enantiomeric configuration for the polymerase enzyme. We have shown recently that mcl-PHA can be synthesized in the cytosol if the mcl-PHA polymerase from Pseudomonas oleovorans is expressed in S. cerevisiae (27) and hypothesized that mcl-PHA precursors are likely made based on peroxisomal enzymes that remain in the cytoplasm.

To synthesize mcl-PHA in the cytosol of S. cerevisiae based on β-oxidation intermediates, key peroxisomal proteins, including Faa2p, Fox1p, and Fox2p, must be active in the cytosol together with PHA polymerase (Fig. 1). Enzymes destined for the peroxisomal matrix are imported from the cytosol in a process involving specific targeting signals. Two different signals have been identified which are believed to be sufficient for transporting proteins into the peroxisome. One is C-terminal peroxisomal targeting signal 1 (PTS1), which is present in the majority of peroxisomal matrix proteins, and the other is peroxisomal targeting signal 2 (PTS2), which is located within the N-terminal 30 amino acids of some peroxisomal proteins such as Fox3p (5). PTS1 consists of the C-terminal tripeptide SKL or its conservative variants ([S/A/C][K/R/H][L/M]). Pex5p is the receptor for PTS1, whereas importing PTS2-carrying proteins is dependent on Pex7p (14).

FIG. 1.

β-Oxidation pathway in S. cerevisiae.

The S. cerevisiae pex5 mutant is viable but accumulates peroxisomal, leaflet-like membrane structures and is deficient in the import of peroxisomal matrix enzymes with an SKL-like import signal such as Fox2p (The Peroxisome Website [http://www.peroxisome.org/]). The acyl coenzyme A (acyl-CoA) oxidase Fox1p follows a novel, non-PTS1, import pathway that is also dependent on Pex5p (14). In pex5 mutants, both Fox2p and Fox1p are found in the cytosol but Fox3p is located in peroxisomes (25).

Activation of fatty acids entering S. cerevisiae can be mediated by at least four different acyl-CoA synthetase gene products, Faa1p to Faa4p (1, 7, 15). One of these enzymes, Faa2p, is a peroxisomal protein that carries a PTS1-like targeting sequence, while the other three enzymes do not show any obvious peroxisomal targeting sequences. A pex5 mutant is expected to retain Faa2p in the cytosol, enabling cytosolic fatty acid activation.

To test whether S. cerevisiae is able to synthesize increased levels of mcl-PHA in the cytosol, we have expressed the P. oleovorans mcl-PHA polymerase (11) in the cytosol of a pex5 receptor mutant. We expected that the peroxisomal proteins would remain localized in the cytosol due to the disrupted peroxisomal protein import mechanism, thus completing a PHA synthesis pathway. In addition, an scl-PHA polymerase was expressed both in the cytosol and in peroxisomes of S. cerevisiae to study the localization of intermediates from fatty acid metabolism and to explore using these metabolites for PHA synthesis.

MATERIALS AND METHODS

Strains and media.

Plasmids were maintained and propagated in Escherichia coli DH5α. All of the S. cerevisiae strains used in this study are listed in Table 1. S. cerevisiae BY4743, BY4741-YIL160C, and BY4743-YDR244W, which is a heterozygous pex5 mutant strain, were obtained from Invitrogen (Carlsbad, CA). Strains wt-16-4 and pex5-16-2 were sporulated from BY4743-YDR244W, and pex5-3c11 was made by mating two haploid pex5 strains by standard protocols (22). S. cerevisiae strains harboring a PHA polymerase gene were grown in SD medium (0.67% yeast nitrogen base without amino acids, 2% glucose, amino acids). For PHA production, a stationary-phase culture grown on glucose was harvested by centrifugation and the cells were washed once in water and resuspended at a 1:10 dilution in fresh SOG1 medium containing 0.67% yeast nitrogen base without amino acids, 1% glycerol, 0.4% Tween 80, and the appropriate fatty acids. When cultivating pex5 mutants, cultures were supplemented with Geneticin (100 μg/ml). The cultures were then grown on SOG1 medium for 5 to 6 days before being harvested for PHA analysis. The medium utilized either a phosphate (5 mM) or a citrate acid (5 mM) buffer to control the pH from 4.5 to 7.0.

TABLE 1.

S. cerevisiae strains used in this study

Strain	Genotype	Note	Origin
D603	Mata/α ura3-52/ura3-52 lys2-801/lys2-801 met/met his3/his3 ade2-101/ade2-101 reg1-501/reg1-501	Diploid, wild type	Carlson et al. (2) and Leaf et al. (16)
YPH499	Mataura3-52 lys2-80 ade2-101 trp-Δ63 his3-Δ200 leu2-Δ1	Haploid, wild type	This study
YPH500	Matα ura3-52 lys2-80 ade2-101 trp1-Δ63 his3-Δ200 leu2-Δ1	Haploid, wild type	This study
BY4743	Mata/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 ura3Δ0/ura3Δ0	Diploid, wild type	Catalog no. 95400-BY4743; Invitrogen
BY4743-YDR244W	Mata/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 ura3Δ0/ura3Δ0 PEX5/pex5::kanMX4	Diploid, heterozygous pex5 mutant	Catalog no. 95400-23603; Invitrogen
BY4741-YIL160C	Matahis3Δ1 leu2Δ0 ura3Δ0 met15Δ0 fox3::kanMX4	Haploid, fox3 mutant	Catalog no. 95400-2319; Invitrogen
pex5-3c11	Mata/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 ura3Δ0/ura3Δ0 pex5::kanMX4/pex5::kanMX4	Diploid, homozygous pex5 mutant	This study
pex5-16-2	Matahis3Δ1 leu2Δ0 ura3Δ0 lys2Δ0 met15Δ0 pex5::kanMX4	Haploid, pex5 mutant	This study
wt-16-4	Matα his3Δ1 leu2Δ0 ura3Δ0 lys2Δ0	Haploid, wild type	This study

Cloning procedure.

Plasmid p2TG1T-700(H) (Fig. 2) was constructed from plasmid p2TG1T(H), which contains the 2μm origin of replication, a HIS3 marker, the TEF1 promoter, and the URA3 termination sequence (3). The P. oleovorans mcl-PHA polymerase gene (phaC1) was isolated from plasmid pPT700 (12) by using a ClaI and EcoRI digest and was ligated into similarly digested p2TG1T(H). The P. oleovorans mcl-PHA polymerase gene (phaC1) containing the PTS1 peroxisomal targeting sequence was obtained by PCR. This peroxisomal targeting sequence was shown by Elgersma et al. to target expression of malate dehydrogenase (MDH3) to the peroxisomes in S. cerevisiae (5). The primers used were 5′-ATTATCGATGAGTAACAAGAACAACGATGAG-3′ and 5′-GGAATTCAACGCTCGTGAACGTAGG-3′, which create an upstream ClaI restriction site and a downstream EcoRI restriction site (underlined). The 3′ primer modified the phaC1 gene by adding a triple-amino-acid peptide (SKL) to the 3′ end. The PCR product was digested with ClaI and EcoRI and ligated into similarly digested p2TG1T(H) to create p2TG1T-755(H).

FIG. 2.

Vectors for mcl-PHA polymerase (phaC1) gene and scl-PHA synthase (phbC) gene expression. Plasmid p2TG1T-700(H) contains the TEF1 promoter, the 2μm origin, a HIS3 marker, and the URA3 terminal (Term.) sequence and expresses the P. oleovorans mcl-PHA polymerase. Plasmid p2TG1T-755(H) contains the mcl-PHA polymerase with a peroxisomal targeting sequence (PTS). Plasmid p2TG1T-500(H) contains the R. eutropha scl-PHA synthase. Plasmid p2TG1T-566(H) contains the R. eutropha scl-PHA synthase with a PTS.

The R. eutropha scl-PHA synthase gene was isolated from plasmid p2DPT-S(U) (2) by using ClaI and EcoRI and ligated into similarly digested p2TG1T(H). The resulting plasmid was named p2TG1T-500(H). R. eutropha scl-PHA synthase containing the peroxisomal targeting sequence was obtained by PCR cloning of p2TG1T-500(H). The primers used were 5′-ATTATCGATGGCGACCGGCAAAGGCGCGGC-3′ and 5′-GGAATTCACAATCTAGCCACAGCTCTTGCCTTGGCTTTGACGTAT-3′. The 3′ primer modified the phbC gene by adding a six-amino-acid peptide (ARVARL) to the 3′ end, which was shown by Hahn et al. to target scl-PHA synthase to the peroxisome of maize (8). The PCR product was digested with ClaI and EcoRI and ligated into similarly digested p2TG1T(H) to create p2TG1T-566(H). The plasmids were transferred into the S. cerevisiae strains by the lithium acetate procedure (23).

Analysis of PHA.

PHA content was determined by gas chromatography-mass spectrometry (GC-MS) analysis of dichloroethane extracts of dried cell material subjected to propanolysis (16). Briefly, the cells were harvested by centrifugation, dried overnight at 101°C, washed six times with warm methanol (65°C), and redried. Dry cell material (20 to 60 mg) was then incubated in a boiling water bath for 2 h in a mixture of 0.5 ml of dichloroethane plus 0.5 ml of acidified propanol. The organic phase was extracted once with water, and the organic phase was then used for GC-MS analysis. Samples were analyzed with a Kratos MS25 GC/mass spectrometer with a DB-WAX column.

RESULTS

Cytosolic expression of the mcl-PHA polymerase in wild-type and heterozygous pex5 yeast strains.

S. cerevisiae strains BY4743, wt-16-4, BY4743-YDR244W, D603, YPH499, and YPH500 were transformed with PHA polymerase plasmid p2TG1T-700(H). The recombinant yeast was grown in defined medium containing 0.5 g/liter lauric acid as the carbon source. The cytosolic expression of the mcl-PHA polymerase resulted in the production of detectable levels of PHA. Cytosolic polymer levels reached approximately 0.015% of the total cell dry weight (CDW) in S. cerevisiae BY4743, while polymer levels reached about 0.026% of the CDW in BY4743-YDR244W, which is 1.7 times higher than in the BY4743 PHA strain. Figure 3B shows the GC-MS trace obtained from a sample of S. cerevisiae BY4743-YDR244W harboring p2TG1T-700(H). The C₁₂ (3-hydroxydodecanoic acid), C₁₀ (3-hydroxydecanoic acid), C₈ (3-hydroxyoctanoic acid), and C₆ (3-hydroxyhexanoic acid) PHA peaks are all clearly visible, indicating that these monomers are present in the PHA polymers. The mass-to-charge ratios of all peaks were checked against the PHA produced by an E. coli strain expressing the P. oleovorans PHA polymerase. In haploid wild-type strain wt-16-4, PHA accumulated to about 0.025% of the CDW.

FIG. 3.

GC-MS analysis of PHA produced by S. cerevisiae BY4743-YDR244W when lauric acid (C₁₂) was used as the carbon source. Panels: A, S. cerevisiae BY4743-YDR244W; B, S. cerevisiae BY4743-YDR244W harboring p2TG1T-700; C, S. cerevisiae BY4743-YDR244W harboring p2TG1T-755. The positions of C₁₂, C₁₀, C₈, and C₆ PHAs are indicated. Only peaks with a mass-to-charge ratio of 131 are shown.

Yeast strains harboring plasmid p2TG1T-755(H), which expresses a peroxisomally targeted mcl-PHA polymerase, were used as positive controls. With the same cultivation method, PHA accumulated to 0.042%, 0.053%, and 0.054% of the CDW in the peroxisomes of BY4743-YDR244W, BY4743, and wt-16-4, respectively (Table 2 and Fig. 3C).

TABLE 2.

mcl-PHA content and monomer composition synthesized by different yeast strains when lauric acid (C₁₂) was used as the carbon source

Host	Plasmid	PHA content (% of CDW)	Mean composition of PHA (%, wt/wt) ± SD
			C12	C10	C8	C6
BY4743	p2TG1T-700	0.015 ± 0.001	59 ± 2	16 ± 1	22 ± 1	3 ± 1
BY4743	p2TG1T-755	0.054 ± 0.004	39 ± 4	24 ± 1	30 ± 5	8 ± 2
BY4743-YDR244W	p2TG1T-700	0.026 ± 0.003	55 ± 4	25 ± 5	16 ± 2	4 ± 3
BY4743-YDR244W	p2TG1T-755	0.042 ± 0.006	45 ± 3	23 ± 4	25 ± 2	6 ± 2
pex5-3c11	p2TG1T-700	0.053 ± 0.004	71 ± 3	15 ± 2	12 ± 2	2.0 ± 0.5
pex5-3c11	p2TG1T-755	0.046 ± 0.001	82 ± 2	6 ± 1	10 ± 3	2.1 ± 0.5
pex5-16-2	p2TG1T-700	0.031 ± 0.009	85 ± 4	12 ± 5	2.0 ± 0.5	1.0 ± 0.5
pex5-16-2	p2TG1T-755	0.028 ± 0.007	83 ± 3	8 ± 5	8 ± 2	1.4 ± 0.5
wt-16-4	p2TG1T-700	0.025 ± 0.013	67 ± 2	18 ± 4	7 ± 1	9 ± 4
wt-16-4	p2TG1T-755	0.054 ± 0.016	36 ± 5	28 ± 4	25 ± 4	12 ± 3
BY4741-YIL160C (fox3)	p2TG1T-700	0.047 ± 0.013	100	NDa	ND	ND
BY4741-YIL160C (fox3)	p2TG1T-755	0.13 ± 0.05	90 ± 2	6 ± 1	4 ± 2	ND

^aND, not detectable.

No cytosolic mcl-PHA was detected in wild-type yeast strains D603, YPH499, and YPH500.

Medium pH values of 4.5 to 7 were explored by using citrate and phosphate buffers. For all pH values, the cytosolic PHA content reached about 0.025% of the CDW in heterozygous pex5 yeast strain BY4743-YDR244W. However, the CDW was significantly lower for pH values higher than 6.0.

Composition of cytosolic mcl-PHA produced in heterozygous pex5 mutants.

To determine the influence of carbon source on PHA monomer composition, recombinant yeast cells were grown in SOG1 medium containing one of the following fatty acids: oleic acid (C₁₈, 1 g/liter), tetradecanoic acid (C₁₄, 0.5 g/liter), tridecanoic acid (C₁₃, 0.5 g/liter), lauric acid (C₁₂, 0.5 g/liter), undecanoic acid (C₁₁, 0.3 g/liter), or decanoic acid (C₁₀, 0.3 g/liter). The results of the analysis are summarized in Table 3. The data demonstrate that the PHA monomer composition is strongly dependent on the externally fed fatty acids. When C₁₀ fatty acids were used as the carbon source, C₁₀ PHA accounted for about 72% of the total biopolymer while no C₁₂ PHA was detected. Similarly, recombinant yeast grown on tridecanoic acid (C₁₃) and undecanoic (C₁₁) acid produced PHAs containing odd-chain monomers ranging from C₁₃ to C₇, with the major monomer components being C₁₃ and C₁₁ PHAs.

TABLE 3.

Cytosolic PHA content and monomer composition produced by S. cerevisiae BY4743-YDR244W harboring p2TG1T-700(H) when different even-numbered or odd-numbered fatty acids were used as the carbon source

Carbon source	PHA content (% of CDW)	Mean composition of PHA (%, wt/wt) ± SD
		C12	C10	C8	C6	C13	C11	C9	C7
C14, tetradecanoic acid	0.0022 ± 0.0009	NDa	32 ± 2	45 ± 3	22 ± 2
C10, decanoic acid	0.015 ± 0.006	ND	72 ± 2	23 ± 3	4 ± 1
C13, tridecanoic acid	0.017 ± 0.005					37 ± 3	26 ± 5	28 ± 3	8 ± 2
C11, undecanoic acid	0.009 ± 0.002					ND	39 ± 3	30 ± 2	13 ± 3

^aND, not detectable.

When the recombinant yeast cells were grown in medium containing tetradecanoic acid (C₁₄), only trace amounts of C₁₀, C₈, and C₆ PHAs were detected. No PHA was detected in cultures grown on oleic acid.

Cytosolic mcl-PHA synthesis in pex5 and fox3 mutant strains.

S. cerevisiae pex5-3c11 (homozygous diploid pex5 mutant strain), pex5-16-2 (haploid pex5 mutant strain), and BY4741-YIL160C (haploid fox3 mutant strain) were transformed with plasmids expressing either the PTS1-tagged or nontagged mcl-PHA polymerase [p2TG1T-755(H) and p2TG1T-700(H), respectively]. These mutant strains cannot grow on external fatty acids as the sole carbon source, so the culture media were supplemented with additional glycerol (1 to 3%, wt/vol) to enable growth without repression of peroxisomal enzymes. Lauric acid (0.4 g/liter) was used as the carbon source for PHA synthesis. The media were not buffered. After 5 to 6 days of culturing, the cells were harvested and analyzed for PHA.

S. cerevisiae strains pex5-3c11 and pex5-16-2 expressing the untargeted mcl-PHA polymerase from plasmid p2TG1T-700(H) accumulated PHA to approximately 0.053% and 0.031% of their CDW, respectively. Similar to the wild-type yeasts, the PHA in the pex5 mutants consisted of C₁₂, C₁₀, C₈, and C₆ monomers, with the C₁₂ monomer representing about 70 to 85% of the total biopolymer (Table 2). pex5 mutants harboring the peroxisomally targeted mcl-PHA polymerase expressed from plasmid p2TG1T-755(H) showed similar results, which were expected if the peroxisomal import mechanism was indeed interrupted (Table 2).

The haploid fox3 mutant yeast (strain BY4741-YIL160C) harboring the plasmid for untargeted polymerase [p2TG1T-700(H)] accumulated PHA based on lauric acid to about 0.047% of its CDW; however, the polymer contained only C₁₂ monomers. When the mcl-PHA polymerase was targeted to the peroxisomes as previously done by Marchesini and Poirier (18), the yeast accumulated PHA to approximately 0.13% of the CDW. The PHA was composed of C₁₂, C₁₀, and C₈ monomers, with the C₁₂ monomers representing the largest fraction, confirming data presented by Marchesini and Poirier (18). Since complete β-oxidation is not possible in the fox3 mutant, the C₈-C₁₀ precursors are likely derived from fatty acid biosynthesis.

Composition of cytosolic mcl-PHA synthesized in homozygous pex5 mutants.

To investigate the influence of the carbon source on PHA monomer composition in pex5 mutants, homozygous diploid S. cerevisiae strain pex5-3c11 was grown in SOG1 medium containing either oleic acid (C₁₈, 0.5 g/liter), tridecanoic acid (C₁₃, 0.4 g/liter), lauric acid (C₁₂, 0.4 g/liter), undecanoic acid (C₁₁, 0.2 g/liter), or decanoic acid (C₁₀, 0.2 g/liter). Table 4 shows that the PHA monomer composition is dependent on the nature of the external fatty acids. Recombinant yeast grown on C₁₃, C₁₂, C₁₁, and C₁₀ fatty acids produced PHA composed primarily of C₁₃, C₁₂, C₁₁, and C₁₀ monomers, respectively. These monomers represent about 45 to 77% of the total accumulated PHA (Table 4). Interestingly, when undecanoic acid was used as the carbon source, in addition to odd-chain-length PHAs, even-chain PHA monomers including C₁₂, C₁₀, C₈, and C₆ were detected. These even-chain precursors may originate from fatty acid biosynthesis, as proposed by Marchesini and Poirier (18). When the pex5 mutant was grown in medium containing glycerol and oleic acid or only glycerol, the culture accumulated PHA composed of C₈ and C₆ monomers. This also supports the conclusion that fatty acid biosynthesis provides precursors for PHA synthesis in yeast.

TABLE 4.

Cytosolic PHA content and monomer composition synthesized by S. cerevisiae pex5-3c11 harboring p2TG1T-700(H) when different fatty acids were used as the carbon source

Carbon source(s)	PHA content (% of CDW)	Mean composition of PHA (%, wt/wt) ± SDa
		C14	C13	C12	C11	C10	C9	C8	C7	C6
Oleic acid (C18)	0.0095 ± 0.0039					9 ± 4		57 ± 5		34 ± 2
Tridecanoic acid (C13)	0.051 ± 0.010		77 ± 4		16 ± 3		3.8 ± 0.5		2.1 ± 0.5
Lauric acid (C12)	0.054 ± 0.004			71 ± 3		15 ± 2		12 ± 2		2.0 ± 0.5
Undecanoic acid (C11)	0.040 ± 0.005			0.2 ± 0.1	46 ± 5	0.5 ± 0.3	25 ± 4	14 ± 5	7 ± 2	7 ± 3
Decanoic acid (C10)	0.052 ± 0.019					45 ± 5		40 ± 4		15 ± 4
Only glycerol	0.0013 ± 0.0006							83 ± 3		17 ± 2
Homoserine + C12	0.050 ± 0.020			58 ± 4		19 ± 3		19 ± 3		4 ± 2
Pyruvate + C12	0.063 ± 0.013	1.1 ± 0.4		73 ± 3		11 ± 2		11 ± 3		4 ± 2
Acetate + C12	0.045 ± 0.007	29 ± 2		71 ± 2
Formate + C12	0.069 ± 0.002	32 ± 2		68 ± 2

^aBlank, not detectable; all media contain 1 to 3% glycerol.

Factors influencing cytosolic mcl-PHA synthesis in pex5 mutant yeast.

Before GC-MS analysis, all dried cell samples were washed five or six times with warm methanol to remove free fatty acids, lipids, and 3-hydroxyacyl-CoA from the cells (19). The methanol washing removes the 3-hydroxy monomer peaks observed in plasmid-free negative control cultures. When a 5 mM citric acid buffer was used to control the pH, as was done for the wild-type and heterozygous pex5 mutant cultivation, the polymerase expressing pex5 mutants incorporated less than 15% of the available 3-hydroxy precursor monomers into PHA. This limited utilization of the 3-hydroxy precursor could be the result of cytosolic mcl-PHA polymerase inhibition. In the absence of the citric acid buffer, the pex5 mutant incorporates more than 50% of the available 3-hydroxy precursor into PHA. While S. cerevisiae is not capable of growing solely on citric acid, it does appear to affect PHA synthesis. Qi et al. reported that free CoA is a strong inhibitor of mcl-PHA polymerase (21). Citric acid is a tricarboxylic acid cycle intermediate. A major citrate synthesis route occurs via the condensation of acetyl-CoA and oxaloacetate. The availability of external citrate may affect the intracellular concentration of CoA, which could be a factor influencing the activity of the mcl-PHA polymerase.

Some additional factors that could influence PHA synthesis were explored. When yeast cells were cultivated in SOG1 medium containing C₁₂ fatty acid, externally added succinate (5 g/liter), malate (1 g/liter), oxaloacetate (1 g/liter), phosphate (0.5 g/liter), serine (1 g/liter), glycine (1 g/liter), bovine serum albumin (0.5 g/liter), and NaCl (5 g/liter) had no apparent influence on PHA synthesis. Pyruvate (1 g/liter), acetate (0.5 g/liter), and formate (0.5 g/liter) were tested as alternative carbon sources and in an attempt to reduce intracellular CoA concentrations. The use of pyruvate, acetate, or formate as a carbon source produced higher final biomass concentrations (data not shown). In addition, pex5 mutants grown on these substrates accumulated PHA with C₁₄ monomers (Table 4). These C₁₄ PHA precursors were likely synthesized through the fatty acid biosynthesis pathway and degraded in the cytosol by the β-oxidation enzymes that were not transported into the peroxisomes. Homoserine catabolism involves CoA. It was added to the media to a final concentration of 0.2% with the expectation of reducing free CoA levels, but it had no significant effect on PHA accumulation.

R. eutropha scl-PHA synthase expression in the cytosol and in peroxisomes.

S. cerevisiae strain BY4743 was transformed with either the nontargeted or the targeted PHA synthase plasmid [p2TG1T-500(H) or p2TG1T-566(H), respectively]. The recombinant yeast cells were grown in defined medium containing oleic acid (1 g/liter) as the carbon source. Cytosolic expression of the scl-PHA synthase resulted in the synthesis of PHA, which accumulated to 0.02% of the CDW. In the strain expressing the peroxisomally targeted enzyme, the PHA content was approximately 0.8% of the CDW.

The carbon source was varied to test the effect on monomer composition of peroxisomally produced PHA. The recombinant yeasts were grown on SOG1 medium containing one of the following fatty acids: lauric acid (C₁₂, 0.5 g/liter), tridecanoic acid (C₁₃, 0.5 g/liter), and a mixture of 0.25 g/liter lauric acid and 0.25 g/liter tridecanoic acid. The results are summarized in Table 5. When the peroxisomally targeted synthase strain was fed even-chain fatty acids, the accumulated PHA was composed of approximately 97 to 99% C₄ monomers, with the balance being C₈, C₆, and C₅ monomers. Similarly, feeding an odd-numbered (C₁₃) fatty acid resulted in a PHA copolymer composed of approximately 6% C₄ and 94% C₅ monomers. When the yeast cells were fed a mixture of C₁₂ and C₁₃ fatty acids, polymer levels reached approximately 7% of the CDW. The peroxisomally synthesized PHA was composed of 84% C₄ and 16% C₅ monomers, with the balance being C₆ and C₈ monomers.

TABLE 5.

Peroxisomal PHA content and monomer composition synthesized by S. cerevisiae BY4743 harboring p2TG1T-566(H) when different fatty acids were used as the carbon source

Carbon source	PHA content (% of CDW)	Mean composition of PHA (%, wt/wt) ± SD
		C4	C5	C6	C8
C18, oleic acid	0.8 ± 0.04	97.2 ± 0.5	1.4 ± 0.4	1.2 ± 0.2	0.2 ± 0.1
C12, lauric acid	3.8 ± 0.1	98.9 ± 0.3	0.18 ± 0.03	0.6 ± 0.2	0.3 ± 0.1
C13, tridecanoic acid	1.6 ± 0.1	6 ± 1	94 ± 1	NDa	ND
C12 and C13, lauric acid and tridecanoic acid	6.9 ± 0.1	84 ± 2	16 ± 1	0.3 ± 0.1	0.2 ± 0.1

^aND, not detectable.

DISCUSSION

It has been shown previously that expression of the mcl-PHA synthase targeted to the peroxisomes of S. cerevisiae leads to mcl-PHA synthesis in this organelle (19). In this study, we expressed the P. oleovorans mcl-PHA polymerase in the cytosol of wild-type and pex5 and fox3 mutant yeast cells. The pex5 mutation disrupts the transport of peroxisomal proteins with the PTS1 signal into the organelle, thus creating a functional cytosolic PHA pathway. The Fox3p enzyme, which possesses PTS2, is transported into the peroxisomes through the Pex7p transporter (Fig. 4). Expressing a nontargeted P. oleovorans mcl-PHA polymerase in pex5 mutants permitted the synthesis of mcl-PHA in the cytosol. As shown in Table 2, the pex5 heterozygous yeast strain produced 1.7 times more PHA than did wild-type yeast strain BY4743 harboring untargeted polymerase. This is likely due to a higher concentration of peroxisomal β-oxidation enzymes in the cytosol of the pex5 heterozygous mutant. The level of cytosolic PHA synthesized by the pex5 homozygous mutant is similar to the level synthesized by wild-type yeast expressing a peroxisomally targeted polymerase. No PHA was detected in wild-type yeast strains D603, YPH499, and YPH500; it is believed these strains have mutations in their fatty acids metabolisms.

FIG. 4.

Creation of a PHA synthesis pathway in the cytosol of an S. cerevisiae pex5 mutant. PhaC, PHA polymerase; Fox3, 3-ketoacyl-CoA thiolase; Pex7, peroxisomal signal receptor for the N-terminal signal (PTS2) of peroxisomal matrix proteins; FAS1, β subunit of fatty acid synthetase; FAS2, α subunit of fatty acid synthetase; CEM1, mitochondrial β-keto-acyl synthase; OAR1, mitochondrial 3-oxoacyl-acyl carrier protein (ACP) reductase; Pxa1/2, a heterodimeric peroxisomal ATP-binding cassette transporter complex. AcCoA, acetyl coenzyme A.

In contrast to our results, Marchesini and Poirier (18) reported that no PHA was detected in a pex5Δ0 mutant (BY4742-YDR244W) expressing the PHA synthase. They concluded that a functional peroxisome is required to synthesize mcl-PHA in yeast. We presumably worked with the same original strain and found that this pex5 strain cannot grow on a nonfermentable carbon source like glycerol or ethanol. This strain is suspected of having incomplete mitochondria. All of the pex5 haploid mutants used in this study were obtained through sporulation of the heterozygous diploid yeast BY4743-YDR244W. These pex5 haploid mutants are able to grow on glycerol. Cell growth on glycerol is important for the expression of peroxisomal proteins, as they appear to be completely repressed on glucose. Based on our data, a functional peroxisome is therefore not required to synthesize mcl-PHA in the cytosol of yeast.

S. cerevisiae strains harboring the scl-PHA synthase from R. eutropha produced PHA in the peroxisomes at up to 7% of the CDW. The scl-PHA was composed of C₈-C₄ monomers. These results confirm those obtained by De Oliveira et al. (4); however, the polymer levels in our study are about 100 times higher than in the previous study. The difference could be a result of using a different yeast strain, a different promoter system, or a different medium. In addition, we found PHA containing C₈ monomer. Thus, the R. eutropha polymerase is able to catalyze also the polymerization of C₈ precursors. The peroxisomal location of the targeted enzyme is based on previous research using the same targeting sequence (8), our previous work examining peroxisomal PHA synthesis (27), and additional biochemical data presented in the present study. For instance, PHA can be synthesized from oleic acid in the heterozygous pex5 strain (BY4743-YDR244W) expressing a peroxisomal polymerase, but no PHA was synthesized from oleic acid in the same strain expressing a cytosolic polymerase.

Wild-type yeast strains (BY4743 and wt-16-4) and the pex5 heterozygous strain have an intact peroxisomal β-oxidation pathway. It was an unexpected result to find these yeast cells capable of producing PHAs in the cytosol. One possible explanation is that β-oxidation enzymes are synthesized in the cytosol and then imported into the peroxisomes posttranslationally. This creates a window where they may be temporarily active in the cytosol. In some studies, 15 to 25% of the β-oxidation enzyme activities were found in the cytosol of yeast (6, 10, 14, 26). This explains why wild-type yeast can synthesize PHA in the cytosol and why PHA synthesized in wild-type yeast and the pex5 mutants contains PHA monomers with C backbones different in length than the fed fatty acids. Our data suggest that β-oxidation can occur, at least partially, in the cytosol of S. cerevisiae and is not confined solely to peroxisomes.

Another alternative carbon source for PHA precursors is fatty acids biosynthesis (18). When the pex5 mutant was grown in medium containing only glycerol, C₈ and C₆ monomers were detected in the synthesized PHA. In addition, if undecanoic acid (C₁₁) was used as the carbon source, the pex5 mutants produced PHA containing some even-chain PHA monomers. These even-chain-length monomers may have been synthesized by a fatty acid biosynthesis pathway and then degraded by the β-oxidation enzymes in the cytosol. So both external fatty acid and native fatty acid biosynthesis pathways may contribute to the observed PHA synthesis.

Wild-type and heterozygous pex5 yeast expressing a cytosolic PHA polymerase did not produce PHA from oleic acid (C₁₈). However, PHA synthesis from oleic acid was observed in strains expressing a peroxisomal polymerase (19, 27). These results suggest that β-oxidation intermediates cannot traverse the peroxisome membrane and that the nontargeted mcl-PHA polymerase is not transported into the peroxisomes. A possible explanation for why cytosolically expressed polymerase cannot produce mcl-PHA from oleic acid is that the degradation of oleic acid, which is an unsaturated fatty acid containing a double bond, occurs via a different pathway than that of saturated fatty acids. PHA synthesized by the homozygous pex5 mutants in oleic acid medium contains only C₁₀, C₈, and C₆ monomers, which are likely derived from fatty acid biosynthesis.

In this study, both scl-PHA and mcl-PHA were synthesized in either the cytosol or the peroxisome from intermediates of fatty acid metabolism. The composition of the PHA was strongly influenced by the genetic background of the yeast host, the monomer specificity of the polymerase, the cellular compartment in which the polymerase was active, and the substrate supplied in the medium. The data presented provide a basis for controlling the composition and thus the properties of the synthesized PHA. For instance, homopolymers can be synthesized by the fox3 mutant (BY4741-YIL160C) expressing the cytosolic mcl-PHA polymerase (Table 2). Polymers of even-numbered, odd-numbered, or a combination of even- and odd-numbered monomers can be controlled by feeding the appropriate substrates like fatty acids and glycerol (Tables 3, 4, and 5). In addition, the distribution of the monomers can also be influenced by feeding substrates like pyruvate and acetate along with a fatty acid (Table 4). The strategies presented all hold the potential of creating polymers with novel and desirable material properties. While the current S. cerevisiae strains make modest amounts of PHA, the findings are serving as a basis for future work which is aimed at creating a more viable PHA production system with improved rates and yields. Strategies for improving the strains include examining how flux to the PHA pathways can be increased by modifying feeding schemes or by expressing enzymes like acyl-CoA synthetase, acyl-CoA dehydrogenase, and/or trans-2-enoyl-CoA hydratase in the cytosol. In addition, S. cerevisiae is a popular eukaryotic model system and the knowledge and strategies developed with this system could be transferred to other eukaryotic systems. Microbes with interesting metabolic characteristics like the lipid-accumulating “oleaginous yeasts” could be a promising host, or the knowledge and strategies could be adapted to higher eukaryotic systems like plants.