Transgenic plants producing peroxisomal polyhydroxy- alkanoate (PHA) from intermediates of fatty acid degradation were used to study carbon flow through the β-oxidation cycle. Growth of transgenic plants in media containing fatty acids conjugated to Tween detergents resulted in an increased accumulation of PHA and incorporation into the polyester of monomers derived from the β-oxidation of these fatty acids. Tween–laurate was a stronger inducer of β-oxidation, as measured by acyl-CoA oxidase activity, and a more potent modulator of PHA quantity and monomer composition than Tween–oleate. Plants co-expressing a peroxisomal PHA synthase with a capryl-acyl carrier protein thioesterase from Cuphea lanceolata produced eightfold more PHA compared to plants expressing only the PHA synthase. PHA produced in double transgenic plants contained mainly saturated monomers ranging from 6 to 10 carbons, indicating an enhanced flow of capric acid towards β-oxidation. Together, these results support the hypothesis that plant cells have mechanisms which sense levels of free or esterified unusual fatty acids, resulting in changes in the activity of the β-oxidation cycle as well as removal and degradation of these unusual fatty acids through β-oxidation. Such enhanced flow of fatty acids through β-oxidation can be utilized to modulate the amount and composition of PHA produced in transgenic plants. Furthermore, synthesis of PHAs in plants can be used as a new tool to study the quality and relative quantity of the carbon flow through β-oxidation as well as to analyse the degradation pathway of unusual fatty acids.
Plant peroxisomes are involved in a variety of metabolic functions including photorespiration, detoxification of reactive oxygen species and the catabolism of fatty acids and amino acids ( Olsen 1998). The β-oxidation and glyoxylate cycles participate in the breakdown of fatty acids to acetyl-coenzyme A (CoA) and its conversion to succinate, respectively. During germination of seeds storing triacylglycerides, the enzymes of both the β-oxidation and glyoxylate cycles are abundant in peroxisomes. Once photosynthesis is established, the activity of these two cycles greatly diminishes, while the enzymes involved in photorespiration become abundant. Reactivation of β-oxidation and glyoxylate cycles occurs during senescence. The intracellular concentration of hexose sugars or the flux of hexose sugars into glycolysis may provide important signals giving rise to changes in expression of genes involved in the glyoxylate cycle ( Graham et al. 1994 ; Ismail et al. 1997 ). It is not know whether the same signals could also be involved in regulation of the β-oxidation cycle. Although the enzymes of the β-oxidation and glyoxylate cycles are normally present at very low levels in photosynthetic leaves and developing seeds, analysis of transgenic Brassica napus expressing the California bay lauroyl-acyl carrier protein (ACP) thioesterase revealed an increase of isocitrate lyase (ICL) activity in leaves and of both ICL and acyl-CoA oxidase (ACOX) activities in developing seeds ( Eccleston & Ohlrogge 1998; Eccleston et al. 1996 ). Acyl-ACP thioesterases are plastidial enzymes catalysing the hydrolysis of acyl-ACPs, resulting in the termination of chain elongation by the fatty acid synthase complex. Transgenic plants expressing the lauroyl-ACP thioesterase were shown to accumulate lauric acid in seed triacylglycerides ( Voelker et al. 1992 ). In contrast, although the level of lauric acid synthesized in chloroplasts isolated from these transgenic plants is relatively high, no lauric acid is detected in leaves from the same plants ( Eccleston et al. 1996 ). These results suggested that the lauric acid synthesized in leaves is rapidly degraded through the peroxisomal β-oxidation cycle, a hypothesis supported by the increase in isocitrate lyase activity ( Eccleston et al. 1996 ). It is not known what regulates activation of the enzymes of the β-oxidation and glyoxylate cycles in this system, but the level of fatty acid intermediates could be such a factor.
Over the past decade, much attention has been focused on understanding the mechanisms implicated in the control of carbon fluxes to particular metabolic pathways. In many cases, this knowledge has practical benefits, such as increasing levels of starch, lipids or amino acids in storage tissues. Control over the carbon flux to the peroxisomal β-oxidation pathway has gained interest due to its potential impact on the accumulation of unusual lipids in transgenic oil crops ( Eccleston & Ohlrogge 1998), as well as on polyhydroxyalkanoate (PHA) production in plants ( Poirier 1999).
PHAs are polyesters of hydroxyacids naturally synthesized by a wide variety of bacteria ( Poirier et al. 1995 ; Steinbüchel 1991). PHAs have attracted considerable interest because of their plastic and elastomeric properties, as well as biodegradability, making them an interesting source of renewable and environmentally friendly polymers ( Poirier et al. 1995 ; Steinbüchel 1991). Synthesis of PHAs in agricultural crops is seen as an alternative to bacterial fermentation for the production of PHAs on a large scale and at low cost ( Poirier & Nawrath 1998; Poirier et al. 1992 ; Poirier et al. 1995 ). Expression of the Ralstonia eutropha (formerly Alcaligenes eutrophus) polyhydroxybutyrate (PHB) biosynthetic enzymes in the plastids of transgenic Arabidopsis thaliana led to the accumulation of PHB up to 14% of the leaf dry weight ( Nawrath et al. 1994 ). PHB is a plastic with poor physical properties, being relatively stiff and brittle. In contrast, medium-chain-length PHAs (MCL-PHAs) are co-polymers having physical properties ranging from soft plastics to elastomers, rubbers and glues ( De Koning 1995). MCL-PHAs are composed of 3-hydroxyacid monomers ranging from 6–16 carbons in length, and are synthesized in a number of pseudomonads from 3-hydroxyacyl-CoA intermediates generated by the β-oxidation of alkanoic acids ( Poirier et al. 1995 ; Steinbüchel 1991). It has recently been demonstrated that targeting of a PHA synthase from Pseudomonas aeruginosa into the peroxisomes of A. thaliana led to the synthesis of MCL-PHAs containing saturated and unsaturated monomers ( Mittendorf et al. 1998a ; Mittendorf et al. 1998b ). In these transgenic plants, PHA inclusions accumulated primarily in the peroxisomes and maximal synthesis was linked to the β-oxidation of reserve fatty acids in germinating seedlings.
In the present study, transgenic plants accumulating peroxisomal MCL-PHA were used as a tool to study factors influencing the activity of the β-oxidation cycle as well as the relative quantity and the nature of the fatty acids that can be channelled to β-oxidation and PHA synthesis.
Modulation of PHA synthesis by Tween fatty acid esters
To assess whether the flow of fatty acids to the peroxisomal β-oxidation and PHA biosynthetic pathways could be increased in plants, seedlings from transgenic A. thaliana line 3.3, which constitutively expressed the PhaC1 synthase from P. aeruginosa in peroxisomes ( Mittendorf et al. 1998a ; Mittendorf et al. 1998b ), were grown in liquid cultures supplied with various fatty acid conjugates. Tween-80 and Tween-20 are detergents of polyoxyethylenesorbitan esterified primarily to oleic acid (C18:1) and lauric acid (C12:0), respectively. It has been shown previously that plant cells can incorporate fatty acids present in Tween into their membrane lipids ( Terzaghi 1986a; Terzaghi 1986b; Terzaghi 1989). While the addition of 5% Tween-80 (equivalent to 28 m m oleic acid) resulted only in a marginal increase in PHA accumulation, addition of 5% Tween-20 (equivalent to 20 m m lauric acid) resulted in an approximate eightfold increase in PHA ( Table 1). Tween-80 feeding resulted in an increased proportion of all PHA monomers derived from the β-oxidation of oleic acid, namely 3-hydroxyhexadecenoic acid (H16:1), 3-hydroxytetradecenoic acid (H14:1), 3-hydroxydodecanoic acid (H12), 3-hydroxydecanoic acid (H10:0), 3-hydroxyoctanoic acid (H8) and 3-hydroxyhexanoic acid (H6) ( Table 1 and Fig. 1). PHA isolated from plants grown in media supplemented with Tween-20 showed significant increase in the proportion of all saturated 3-hydroxyacid monomers ranging from 6–12 carbons, which are primarily derived from the β-oxidation of lauric acid ( Table 1). Similar results were obtained for plant cultures fed with 2.5% Tween-80 and Tween-20, while addition of Tween backbone (Tween without fatty acids) had no significant effects on either PHA quantity or monomer composition compared to cultures grown in the absence of Tween (data not shown).
Table 1. Influence of Tween on the quantity and monomer composition of PHA produced in transgenic plants
a3-Hydroxyacid monomers are denoted by the prefix H. The monomer H10:1 and odd-chain monomers were not quantified.
Seedlings from phaC1-transformed line 3.3 were grown axenically in liquid medium containing half-strength Murashige and Skoog (MS) salts and 2% sucrose for the first 7 days, and in the same medium supplemented with 5% (w/v) Tween detergent for an additional 7 days.
PHAs produced from cultures supplied with Tween detergents were also analysed by 1H-NMR. One-dimensional 1H-NMR does not provide a level of resolution sufficient for the precise assignment of all peaks to particular carbon groups in PHA containing over 15 different saturated and unsaturated monomers. Nevertheless, comparative analysis of bacterial PHA containing saturated and unsaturated H6 and H8 monomers with linolenic acid and plant MCL-PHA clearly confirmed that PHA isolated from plant cultures supplied with Tween-80 contained unsaturated monomers and that their proportion greatly decreased in PHA isolated from cultures supplied with Tween-20 ( Fig. 2). These results indicate that addition of an external source of fatty acids in the form of a Tween conjugate can lead to an increased flow of fatty acids towards peroxisomal β-oxidation. It furthermore indicates that the amount of PHA synthase expressed in these transgenics was not limiting the synthesis of PHA in plants relying on the endogenous fatty acid pool, but rather that the supply of fatty acids to β-oxidation was a key factor.
Peroxisomal β-oxidation of unusual fatty acids
Plants can synthesize a broad range of unusual fatty acids which are mainly accumulated in their seed oils ( Van de Loo et al. 1993 ). Although plants producing unusual fatty acids, such as hydroxylated or epoxidized fatty acids, must be able to degrade them during germination, little knowledge is available on these catabolic pathways. Furthermore, there are examples of transgenic plants which accumulate in their seed triacylglycerides containing unusual fatty acids that they do not normally make, without apparent negative impact on seed germination, again indicating that appropriate degradation of unusual fatty acids occurs in these plants ( Broun & Somerville 1997; Lee et al. 1998 ).
To investigate the range of fatty acids which could be degraded by β-oxidation and analyse whether their degradation intermediates could be included into MCL-PHAs, unusual fatty acids were supplied exogenously to transgenic plants growing in liquid media containing 0.2% sucrose and 5% Tween-80. Growth of plants in media containing 0.2% sucrose generally lead to a higher level of PHA per g dry weight compared to plants growing in media containing 2% sucrose ( Table 1 versus Table 2). This is probably due to the higher shoot to root ratio of plants grown in low sucrose. While addition of 0.35 m m octanoic acid resulted only in minor changes in the composition of PHA, with a relatively small increase in the proportion of H8 monomer, addition of 1.8 m m tridecanoic acid resulted in a large increase in the proportion of H13 and H11 monomers, and a more modest increase in H9 and H7 monomers ( Table 2). Furthermore, addition of 0.47 m m of tridecenoic acid (C13:1 Δ12) and 0.87 m m 8-methyl nonanoic acid resulted in the synthesis of PHA containing additional novel monomers. For plants fed with C13:1 Δ12, the PHA contained 10, 7, 8 and 4 mol% of H13:1 Δ12, H11:1 Δ10, H9:1 Δ8 and H7:1 Δ6 monomers, respectively. Plants fed with 8-methyl nonanoic acid produced a PHA containing 33, 19 and 7 mol% 8-methyl H9, 6-methyl H7 and 4-methyl H5 monomers, respectively. In plant PHA isolated from cultures fed with C13:1 Δ12, the presence of monomers with terminal unsaturated bonds was further confirmed by 1H-NMR ( Fig. 2e). In all cases, changes in the plant PHA monomer composition following addition of unusual fatty acids can be explained by the incorporation of monomers derived from the 3-hydroxyacyl-CoA intermediates of the β-oxidation of these unusual fatty acids.
Table 2. . Influence of exogenous unusual fatty acids on PHA synthesis in plants
Seeds from phaC1-transformed line 3.3 were grown axenically in liquid medium containing half-strength MS salts and 0.2% sucrose for the first 7 days, and in the same medium supplemented with 5% (w/v) Tween-80 and free fatty acids for an additional 7 days. Monomers present in trace amounts (H10:1 and H16:1) were not quantified. n = 3–4; data are means ± SD.
Of the various enzymes involved in fatty acid β-oxidation, ACOX is thought to be a good indicator of the level of β-oxidation since it appears to be the main step controlling the flux through β-oxidation ( Aoyama et al. 1994 ; Holtman et al. 1994 ). The levels of both medium-chain and long-chain ACOX were measured in plants fed with 2.5% Tween-80, Tween-20 and Tween backbone ( Fig. 3). Addition of Tween-20 to the growth media increased medium-chain and long-chain ACOX activities in transgenic plants by approximately fourfold. In contrast, no significant changes in ACOX activity were observed with the addition of either Tween-80 or Tween backbone.
Modulating the monomer composition of MCL-PHAs using endogenous fatty acids
The physical properties of PHAs are dependent on their monomer compositions. MCL-PHA synthesized in transgenic A. thaliana expressing the peroxisomal PhaC1 synthase contains a high proportion of saturated and unsaturated long-chain monomers (40–50 mol% H12 and higher, 20–30 mol% polyunsaturated monomers), making the polymer soft and sticky. Better elastic properties and a higher melting point can be achieved by increasing the proportion of shorter and more saturated monomers. In a first strategy aimed at modulating the proportion of different fatty acids going to β-oxidation, and therefore influencing the proportion of various monomers in MCL-PHAs, we have expressed the peroxisomal PhaC1 synthase in various A. thaliana mutants deficient in fatty acid desaturation. The fad2-1 mutant is deficient in the extra-plastidial 18:1 desaturase, resulting in an increase in mono-unsaturated fatty acids and a decrease in di-unsaturated and tri-unsaturated fatty acids present in leaf, root and seed lipids ( Miquel & Browse 1992). The fad3/fad7-1/fad8 triple mutant has a large reduction in the 18:2 and 16:2 desaturases, resulting in a more than 10-fold decrease in tri-unsaturated fatty acids and a fivefold increase in di-unsaturated fatty acids ( McConn & Browse 1996). Transgenic plants expressing the peroxisomal PhaC1 synthase in the wild-type background were crossed to the fad2-1 and fad3/fad7-1/fad8 triple mutants, and F2 plants homozygous for the phaC1 gene in the different mutant backgrounds were selected.
The quantity and composition of the PHA synthesized in these various transgenic lines grown for 14 days in liquid media is shown in Table 3. Transgenic plants in the fad2-1 background produced a PHA with an increased proportion of monomers derived from mono-unsaturated fatty acids (H16:1, H14:1) or both mono-unsaturated and saturated fatty acids (H12, H10, H6), and a decrease in monomers derived from tri-unsaturated fatty acids (H14:3, H12:2, H8:1) and di-unsaturated fatty acids (H16:2, H14:2, H12:1). Transgenic plants in the fad3/fad7-1/fad8 background produced a PHA with a large decrease in monomers derived from tri-unsaturated fatty acids (H16:3, H14:3, H12:2 and H8:1) and an increase in monomers derived from di-unsaturated fatty acids (H16:2, H14:2, H12:1), as well as an increase in H8 derived from di-unsaturated, mono-unsaturated or saturated fatty acids. The changes in the PHA monomer composition produced in the various mutants reflected well the relative changes in abundance of fatty acids which are used as the source of substrates for PHA synthesis ( Fig. 1). The only exceptions are the relatively stable level of the H16:3 and H8 monomers in transgenic plants in the fad2-1 background. The amount of MCL PHA produced in the various lines was between 0.54 and 0.73 mg g–1 dry weight for 14-day-old plants, indicating that although the nature of the fatty acid going to β-oxidation was changed, the quantity remained relatively stable.
Table 3. . Composition of PHA produced in mutants affected in fatty acid desaturases
The monomer 10:1 and odd-chain monomers were not quantified.
Seedlings from phaC1-transformed line 3.3 (Columbia wild-type background) and lines expressing the same transgene in the fad2-1 and fad3/fad7-1/fad8 mutant backgrounds were grown axenically in liquid medium containing half-strength MS salts and 1% sucrose for 14 days.
c PHA content (mg g –1 dry weight) n = 3; data are means ± SD.
0.68 ± 0.06
1.3 ± 0.1
19 ± 1
22 ± 2
5.3 ± 0.2
4.7 ± 0.2
3.4 ± 0.2
1.4 ± 0.1
4.8 ± 0.3
3.0 ± 0.1
6.3 ± 0.3
15 ± 1
3.5 ± 0.2
0.3 ± 0.1
3.6 ± 0.1
7.4 ± 0.3
0.54 ± 0.06
2.3 ± 0.2
17 ± 1
15 ± 1
12 ± 1
11 ± 1
0.6 ± 0.1
0.79 ± 0.06
4.0 ± 0.1
15 ± 1
1.3 ± 0.1
10 ± 1
2.5 ± 0.1
0.7 ± 0.2
0.64 ± 0.06
8.4 ± 0.2
0.73 ± 0.03
1.4 ± 0.1
44 ± 1
5.5 ± 0.1
4.1 ± 0.1
11 ± 1
4.3 ± 0.1
2.6 ± 0.3
14 ± 1
0.11 ± 0.06
3.7 ± 0.2
0.3 ± 0.1
8.9 ± 0.1
Increased flow of carbon towards PHA in transgenic plants expressing an acyl-ACP thioesterase
Since experiments in liquid cultures indicated that the supply of fatty acids to the β-oxidation pathway and PHA synthesis could be increased by the exogenous addition of unusual fatty acids, such as lauric acid, we sought to determine whether a similar increase could be generated endogenously through genetic engineering.
The FatB3 gene encoding a medium-chain acyl-ACP thioesterase has recently been cloned from Cuphea lanceolata ( Martini et al. 1999 ; Töpfer et al. 1995 ). Expression of the full-length FatB3 cDNA in A. thaliana under the control of the B. napus napin promoter led to the accumulation of triacylglycerides in seeds containing 18 mol% capric acid, 3.5 mol% lauric acid and 3.4 mol% myristic acid (Y. Poirier, unpublished results), confirming that the FatB3 thioesterase from C. lanceolata has highest specificity for capryl-ACP ( Martini et al. 1999 ).
The FatB3 gene under the control of a truncated CaMV 35S promoter was transformed into transgenic A. thaliana line 3.3 expressing the PhaC1 synthase in the peroxisomes. Two independent transgenic lines, TP2.4 and TP2.11, homozygous for both the FatB3 thioesterase and PhaC1 genes, were analysed for fatty acid and PHA compositions. Seeds from the double transgenic line TP2.4 accumulated lipids containing 2.2 mol% capric acid, while seeds of line TP2.11 accumulated lipids containing 2.5 mol% capric acid. Leaves from both double transgenic lines did not show the presence of medium-chain length fatty acids (detection limit < 0.05 mol%), confirming that like B. napus and A. thaliana plants expressing a lauroyl-ACP thioesterase ( Eccleston et al. 1996 ; Hooks et al. 1999 ), A. thaliana transformed with a capryl-ACP thioesterase can accumulate medium-chain fatty acids in seed lipids but not in leaf lipids. Growth of the FatB3/phaC1 double homozygous transgenic plant lines was slower than either plants homozygous for only the phaC1 gene or heterozygous for both phaC1 and FatB3. The fresh weight of 26-day-old soil-grown FatB3/phaC1 double homozygous transgenic plants was approximately 30% of wild-type or phaC1-transformed plants. Reduction in growth has also been observed for some plants expressing only the CaMV35S–FatB3 gene construct, although the number of plants analysed were too small to clearly establish a correlation between level of thioesterase expression and reduction in plant growth.
PHA was extracted from 10-day-old double homozygous seedlings grown on agar plates and from leaves of 40-day-old soil-grown plants homozygous for the phaC1 gene and either homozygous or heterozygous (in a 1:2 ratio) for the FatB3 thioesterase. The data for the quantity and monomer composition of PHA are summarized in Table 4. Care should be taken in comparing the values of PHA in transgenic line PhaC1 3.3 grown for 10 days on plates containing 2% sucrose ( Table 4) with plants grown in a similar liquid media for 14 days ( Tables 1 and 3). Because most of the reserve lipids are in the cotyledons, this tissue is expected to have a larger amount of PHA compared to roots and leaves. Plants grown on agar plates for 10 days have large cotyledons, emerging primary leaves and a small root sytstem. In contrast, plants grown in liquid media for 14 days have a large root system and well developed primary leaves. Thus, the amount of PHA in plants grown on plates for 10 days (2.2 ± 0.3 mg g–1 dry weight, n = 4; mean ± SD) is consistently higher than for plants grown in liquid for 14 days (0.7 ± 0.1 mg g–1 dry weight, n = 5). The same transgenic plants grown in soil for 45 days have a PHA content of 0.10 ± 0.03 mg g–1 dry weight (n = 5), reflecting the fact that as the plant expands during post-germinative photosynthetic growth, the rate of increase in dry weight of the plant outpaces the synthesis of PHA in the peroxisomes ( Mittendorf et al. 1998a ).
Table 4. . PHA synthesis in transgenic plants expressing the phaC1 and FatB3 genes
Monomers 10:1 and 16:1 and odd-chain monomers were not quantified.
b Plants were grown under continuous illumination either for 10 days on solid medium (MS, 1% sucrose and 20 μg ml –1 hygromycin for TP lines or 50 μg ml–1 kanamycin for PHAC3.3 line) or for 40 days in soil.
c PHA content (mg g –1 dry weight).
Homozygous for both phaC1 and FatB3 genes.
Homozygous for phaC1 and either homozygous or heterozygous (in a 1:2 ratio) for FatB3.
Analysis demonstrated that 10- and 40-day-old double transgenic plants produced approximately three and eight times more PHA, respectively, in their vegetative tissues compared to plants expressing only the PhaC1 synthase ( Table 4). Furthermore, double transgenic plants contained a higher amount of saturated short-chain monomers, with H10 and H8 monomers reaching 74 mol% of the polymer in 40-day-old double transgenics, as compared to 18 mol% in plants expressing only the PhaC1 synthase.
The level of ACOX activity was measured in both TP2.4 and TP2.11 lines. A significant increase in both medium-chain and long-chain ACOX activity was measured in the double transgenic lines compared to the parental line expressing only the PhaC1 synthase ( Fig. 4).
It has been shown previously that exogenous Tween–fatty acid conjugates are imported into soybean cells where they are hydrolysed to give free fatty acids, which are in turn incorporated preferentially into extra-plastidial glycerolipids ( Terzaghi 1986a; Terzaghi 1986b). Shintani & Ohlrogge (1995) have also shown that addition of Tween–fatty acid esters to tobacco suspension cells caused a reduction in acetate incorporation into fatty acids which could be best explained by a biochemical or post-translational modification of the acetyl-CoA carboxylase. The effects of Tween on PHA accumulation in transgenic plants expressing a PHA synthase in the peroxisome clearly indicate that in addition of being included in membrane lipids, external fatty acids are also directed towards the peroxisomal β-oxidation cycle. Addition of Tween-80 or Tween-20 modified the PHA monomer composition by increasing specifically the proportion of 3-hydroxyacid monomers derived from the β-oxidation of the major fatty acid present in the conjugate.
It is striking that addition of Tween-20, containing primarily lauric acid, was more effective at increasing the amount of PHA than addition of Tween-80, which contains primarily oleic acid ( Table 1). Furthermore, larger shifts in PHA monomer composition were achieved with Tween-20 compared to Tween-80 ( Table 1). It is unlikely that differences in the affinity of the PHA synthase for the various 3-hydroxyacyl-CoA intermediates generated by the β-oxidation of oleic and lauric acids could account for this large difference since four out of six monomers generated by the degradation of these fatty acids are identical (i.e. H12, H10, H8 and H6) ( Fig. 1). These results thus indicate that the proportion of external fatty acids channelled towards peroxisomal β-oxidation is larger for Tween–laurate compared to Tween–oleate. This may be due to preferential penetration of Tween–laurate into the plant cells or to a higher rate of fatty acid hydrolysis from the Tween backbone. Alternatively, it is possible that this difference may also be due to an intrinsic metabolic response that specifically targets unusual fatty acids, such as lauric acid, towards β-oxidation, in comparison to usual fatty acids, such as oleic acid, which may be incorporated into membrane lipids (see below).
Addition of 2.5% Tween-20 to plant cells resulted in a large increase in ACOX activity, while addition of an equivalent amount of Tween-80 or Tween backbone had no significant effects on ACOX activity. This indicates that the increase in ACOX activity is not due to a non-specific toxic effect of detergents on cells, but rather is a response to the presence of fatty acids in the media, with plant cells being more responsive to unusual fatty acids compared to fatty acids normally present in membrane lipids. A significant increase in both medium- and long-chain ACOX activity has also been observed in shoots of transgenic A thaliana expressing constitutively the C. lanceolata capryl-ACP thioesterase ( Fig. 4). Analysis of transgenic B. napus expressing the lauroyl-ACP thioesterase revealed an increased activity of only the medium-chain ACOX, suggesting a correlation between the length of the synthesized fatty acid and the specific induction of a particular ACOX ( Eccleston & Ohlrogge 1998). Our results show no such direct correlation since FatB3 has high activity for C10 acyl-ACP while an increase in long-chain ACOX is evident in both transgenic lines tested. In plants fed with Tween-20, both medium- and long-chain ACOX were also increased. In contrast to rapeseed, constitutive expression of the California bay lauroyl-ACP thioesterase in A. thaliana was recently shown not to lead to increased activities of the various enzymes involved in β-oxidation, including ACOX ( Hook et al. 1999 ). These data, combined with our results on PHA synthesis in peroxisomes (see below), indicate that ACOX activity may not be a reliable indicator of the relative flow of fatty acids through β-oxidation, but that peroxisomal PHA may fulfil such a role.
Mechanisms are thought to exist in plants to exclude the accumulation of unusual fatty acids in membrane lipids ( Battey & Ohlrogge 1989; Stymne 1993). This concept is supported by the absence of medium-chain and hydroxylated fatty acids in membrane lipids from the vegetative tissues of transgenic plants expressing a medium-chain thioesterase and a hydroxylase, respectively ( Broun et al. 1998 ; Eccleston et al. 1996 ; Hooks et al. 1999 ; Van de Loo et al. 1995 ). Co-expression in transgenic A. thaliana of C. lanceolata FatB3 thioesterase in the plastid and PhaC1 synthase in the peroxisome resulted in an eightfold increase in PHA accumulating in 40-day-old shoots. Furthermore, the PHA monomer composition was strongly shifted towards H8 and H10 saturated monomers, reflecting the higher specificity of the FatB3 thioesterase towards capryl-ACP. Clearly, medium-chain fatty acids are being synthesized in the shoots of transgenic plants expressing FatB3 but are degraded by the β-oxidation pathway before their accumulation to detectable levels in membrane lipids. These results support the hypothesis that plant cells must have mechanisms which sense levels of free or esterified unusual fatty acids, resulting in changes in the activity of the β-oxidation cycle as well as in the removal and degradation of these unusual fatty acids.
Nearly 100 different hydroxyacid monomers have been shown to be included into bacterial PHAs, a large proportion of which are derived from the β-oxidation of unusual lipids or alkanoic acids added to the bacterial growth medium ( Steinbüchel & Valentin 1995). Feeding plant cells with a number of unusual fatty acids has shown that the plant β-oxidation pathway is also capable of generating a number of unusual substrates which can be included in MCL-PHAs, even from lipids which are not normally synthesized in significant quantities by the plant ( Table 2). For the different unusual fatty acids tested, all novel monomers (≥ 6 carbons) derived from degradation through peroxisomal β-oxidation were detected in PHA. These results indicate that analysis of the monomer composition of plant MCL-PHAs can be used as a tool to study the pathway of degradation of unusual fatty acids. For example, it is largely unknown how unusual fatty acids, such as hydroxylated or epoxydized fatty acids, are degraded in plants ( Gerhart 1993). Analysis of the monomer composition of plant PHA isolated from cells fed with unusual fatty acids could thus provide valuable information on the biochemistry of their degradation.
The physical properties of PHAs are strongly dependent on their monomer composition. Thus, while PHB homopolymer is a stiff and brittle plastic, poly(hydroxyvalerate–hydroxybutyrate) co-polymer is more flexible, while MCL-PHAs are typically soft elastomers. Gaining control over the monomer composition through modification of plant metabolism is therefore important for the success of producing PHAs in plants. In this study, we have used two strategies to modulate the quality of monomers channelled to PHA synthesis. In the first strategy, the PHA synthase was expressed in A. thaliana mutants deficient in the desaturation of lipids. PHA synthesized in these plants showed significant changes in monomer composition, such as reduction in tri-unsaturated and di-unsaturated 3-hydroxyacids ( Table 3). The changes in PHA monomer composition reflected well the relative proportion of various fatty acids present in the plant cell, with the exception of the stable level of H16:3 and H8 monomers in the fad2-1 mutant background. The reason for these discrepancies is unknown but could reflect changes in the availability of various 3-hydroxyacyl-CoAs to the PHA synthase due, for example, to changes in activities of different enzymes of the β-oxidation cycle ( Hooks et al. 1996 ). It is notable that the amount of PHA synthesized in the fad3/fad7-1/fad8 mutant background is not increased compared to wild-type ( Table 3). This indicates that large changes in the proportion of the usual fatty acids present in membrane lipids are not sufficient to increase the flow of fatty acids towards β-oxidation, reinforcing the concept that unusual fatty acids have a greater impact on this flow then usual fatty acids.
In the second strategy, both the quality and quantity of PHA synthesized were modulated. Shoot tissues of transgenic plants expressing the PhaC1 synthase in the peroxisomes accumulated a relatively low amount of PHA (0.1 mg g–1 dry weight in 40-day-old leaves) containing a high proportion of long-chain (59 mol% ≥ 12 carbons) or unsaturated (68 mol%) monomers, making the polymer a sticky soft material with low crystallinity and melting point. Co-expression of FatB3 thioesterase in the plastid and PhaC1 synthase in the peroxisome resulted in an eightfold increase in PHA accumulating in 40-day-old shoots, with the monomer composition being strongly shifted towards saturated short-chain monomers (81 mol% H6, H8 and H10), bringing the properties of the polymer towards an elastomer ( Table 4).
Unusual fatty acids containing functional groups, such as hydroxy, epoxy, keto, furanoid and cyclic groups, are attractive in the context of functionalizing PHAs as they represent a source of monomers with reactive side groups which can be used to change the properties of PHAs after extraction ( De Koning et al. 1994 ). It is expected that, as for the synthesis of medium-chain fatty acids in leaves, the synthesis in vegetative tissues of fatty acids with unusual side groups will lead to an increase in their degradation by the peroxisomal β-oxidation cycle. Thus, from this work, a working model emerges whereby the quantity and quality of the PHA synthesized in the vegetative tissues of plants could be modulated through the co-expression of enzymes responsible for the synthesis of unusual fatty acids in conjunction with a peroxisomal PHA synthase capable of using the β-oxidation intermediates from these unusual fatty acids to synthesize novel PHAs.
Plant material and growth conditions
Transgenic A thaliana line 3.3 expressing the P. aeruginosa PhaC1 synthase in the peroxisomes was transformed with the FatB3 gene by the vacuum infiltration method ( Bechtold et al. 1993 ). Transformants were selected on media containing Murashige and Skoog (MS) salts, 1% sucrose, 0.7% agar and 20 μg ml–1 hygromycin. Hygromycin-resistant plants were subsequently transferred to soil and grown under continuous fluorescent light at 19°C. For feeding experiments, axenic plants were grown under constant agitation (100 rev min–1) for 14 days in liquid media containing half-strength MS salts, 0.2–2% sucrose and Tween detergents or Tween backbone, either alone or in combination with free fatty acids. The fad2-1 mutant was obtained from the Nottingham Arabidopsis Biological Resource Center, while the fad3/fad7-1/fad8 triple mutant was provided by John Browse (Washington State University).
Production of Tween backbone
Tween from which fatty acids have been removed (Tween backbone) was prepared according to Terzaghi (1986a) with slight modifications. Ten g of Tween-80 was mixed with 30 ml of aqueous 1 m KOH and refluxed at 85°C for 20 h. The reaction mixture was neutralized with HCl and dried by lyophilization. The mixture was redissolved in diethyl ether and sufficient methanol to solubilize the Tween backbone. Following filtration through glass wool, the Tween solution was applied to a column packed with 100 g of 60–200 μm mesh silica gel 60 (Merck, Darmstadt, Germany) in diethyl ether. Free fatty acids were eluted with four column volumes of diethyl ether, then Tween backbone was eluted with four volumes of methanol. The dried Tween backbone was diluted with four volumes of water and extracted several times with hexane. Tween backbone contained less then 0.1% fatty acids.
The FatB3 gene from C. lanceolata was initially subcloned into the pART7 vector ( Gleave 1992) modified to contain only the first 200 bp of the CaMV 35S promoter, and finally inserted in the binary vector pBJ49 (B. Janssen, unpublished results). The binary Ti plasmid was introduced in the Agrobacterium tumefaciens strain GV3101 ( Koncz & Schell 1986) by electroporation.
ACOX activity measurements
Plant tissue (50 mg) was homogenized in 50 μl of cold extraction buffer containing 150 m m Tris–HCl (pH 7.5), 10 m m KCl, 10% (v/v) glycerol, 10 μm flavin adenine dinucleotide, 1 m mβ-mercaptoethanol, 1 m m EDTA, 0.1 m m PMSF, 0.01% Triton X-100 and 1% polyvinylpyrrolidone. The extracts were clarified by centrifugation at 14 000 g for 10 min at 4°C, and 20 μl of the supernatants were used per 1 ml of reaction mixture. ACOX activity was measured essentially as described by Hooks et al. (1996) using decanoyl-CoA and palmitoyl-CoA as substrates. Hydrogen peroxide production was followed at 500 nm in a UNICAM UV4 spectrometer.
PHA extraction and analysis
Extraction of PHA from plant material and analysis by gas chromatography and mass spectrometry (GC–MS) was performed as previously described ( Mittendorf et al. 1998a ; Mittendorf et al. 1998b ). Fresh or dried frozen plant material was ground in a mortar and lyophilized. The powder was extracted with methanol in a Soxhlet apparatus for 24 h followed by PHA extraction with chloroform for 24 h. The PHA-containing chloroform was concentrated using a Rotovapor and either filtered over glass wool or extracted once with water to remove residual solid particles. PHA was precipitated by the addition of 10 vol cold methanol and subsequently washed by two cycles of chloroform solubilization and methanol precipitation. PHA dissolved in chloroform was trans-esterified by acid methanolysis and analysed by GC–MS using a Hewlett-Packard 5890 gas chromatograph (HP-5MS column) coupled to a Hewlett-Packard 5972 mass spectrometer. In some experiments, a simplified procedure was used in which plant tissues were extensively extracted with methanol at 65°C, the residual material was trans-esterified by acid methanolysis and the PHA content was analysed by GC–MS utilizing the ion selective mode. Identification of monomers present in plant PHA was facilitated by the use of commercial standards and purified bacterial PHAs. 3-Hydroxyoctanoic acid, 3-hydroxydecanoic acid, 3-hydroxydodecanoic acid and 3-hydroxytetradecanoic acid were purchased from Sigma (St Louis) while 3-hydroxyhexanoic acid was obtained from D. Seebach (ETH, Zürich). MCL-PHA purified from P. putida grown on linoleic acid and in which the monomer composition was determined by GC–MS as well as NMR was obtained from G. Eggink (ATO-DLO, Wageningen). PHA containing 3-hydroxyoctenoic acid (H8:1 Δ7) and 3-hydroxyhexenoic acid (H6:1 Δ5) was obtained from B. Witholt (ETH, Zürich), and PHA containing 8-methyl-3-hydroxynonanoic acid and 6-methyl-3-hydroxyheptanoic acid was a gift from B. Hazer (Marmara Research Centre, Turkey). 1H-NMR spectroscopy was performed on a Bruker AC-P200 spectrometer at 200 MHz using deuteriochloroform as solvent and tetramethylsilane as internal standard with a sample concentration of ≤ 10 mg ml–1.
We thank Giovanni Ventre and Stephanie Stolz for their technical assistance. We are grateful to Baki Hazer, Robert Lenz, Gerrit Eggink, Michele Kellerhals and Bernard Witholt for providing MCL-PHA samples, and Bart Janssen for the pBJ49 vector. We are also grateful to John Browse for providing the fad3/fad7-1/fad8 mutant line. We thank Edward Farmer and Christiane Nawrath for critical reading of the manuscript. This work was supported, in part, by grants from the Ciba-Geigy Jubiläums Stiftung, the Herbette Foundation and Monsanto. V.B. and L.A. were recipients of fellowships from Colciencias of Colombia, and the Georgine Claraz Foundation, respectively.
EMBL accession number AJ131740(Cuphea lanceolata FatB3 gene).