Engineering Escherichia coli to produce branched-chain fatty acids in high percentages

Branched-chain fatty acids (BCFAs) are key precursors of branched-chain fuels, which have cold-flow properties superior to straight chain fuels. BCFA production in Gram-negative bacterial hosts is inherently challenging because it competes directly with essential and efficient straight-chain fatty acid (SCFA) biosynthesis. Previously, Escherichia coli strains engineered for BCFA production also co-produced a large percentage of SCFA, complicating efficient isolation of BCFA. Here, we identified a key bottleneck in BCFA production: incomplete lipoylation of 2-oxoacid dehydrogenases. We engineered two protein lipoylation pathways that not only restored 2-oxoacid dehydrogenase lipoylation, but also increased BCFA production dramatically. E. coli expressing an optimized lipoylation pathway produced 276 mg/L BCFA, comprising 85% of the total free fatty acids (FFAs). Furthermore, we fine-tuned BCFA branch positions, yielding strains specifically producing ante-iso or odd-chain iso BCFA as 77% of total FFA, separately. When coupled with an engineered branched-chain amino acid pathway to enrich the branched-chain α-ketoacid pool, BCFA can be produced from glucose at 181 mg/L and 72% of total FFA. While E. coli can metabolize BCFAs, we demonstrated that they are not incorporated into the cell membrane, allowing our system to produce a high percentage of BCFA without affecting membrane fluidity. Overall, this work establishes a platform for high percentage BCFA production, providing the basis for efficient and specific production of a variety of branched-chain hydrocarbons in engineered bacterial hosts.     

Human very-long-chain acyl-CoA synthetase: cloning, topography, and relevance to branched-chain fatty acid metabolism

Very-long-chain acyl-CoA synthetases (VLCS) activate very-long-chain fatty acids (VLCFA) containing 22 or more carbons to their CoA derivatives. We cloned the human ortholog (hVLCS) of the gene encoding the rat liver enzyme (rVLCS). Both hVLCS and rVLCS contain 620 amino acids, are expressed primarily in liver and kidney, and have a potential peroxisome targeting signal 1 (-LKL) at their carboxy termini. When expressed in COS-1 cells, hVLCS activated the VLCFA lignoceric acid (C24:0), a long-chain fatty acid (C16:0), and two branched-chain fatty acids, phytanic acid and pristanic acid. Immunofluorescence and immunoblot studies localized hVLCS to both peroxisomes and endoplasmic reticulum. In peroxisomes of HepG2 cells, hVLCS was topographically oriented facing the matrix and not the cytoplasm. This orientation, coupled with the observation that hVLCS activates branched-chain fatty acids, suggests that hVLCS could play a role in the intraperoxisomal reactivation of pristanic acid produced via alpha-oxidation of phytanic acid.     

A novel high-throughput screening assay for putative antidiabetic agents through PPARalpha interactions

As natural peroxisome proliferator-activated receptor-alpha (PPARalpha) ligands, high levels of fatty acids and glucose could lead to hyperactivation of PPARalpha, like that seen in diabetes. Important diabetes research goals are to uncover new metabolic or signaling pathways involved in hyperglycemic cellular injury and to develop therapeutics for preventing or reversing this injury. Consequently, 1040 putative antidiabetic agents were screened for their ability to 1) affect PPARalpha lipid binding, 2) directly bind PPARalpha, and 3) alter PPARalpha transactivation in the presence of high glucose. A high-throughput fluorescent binding assay was developed to examine each compound's ability to restore fatty acyl-CoA binding to PPARalpha in the presence of high glucose concentrations. Approximately 1% of the compounds restored acyl-CoA binding by 60% or more. These compounds directly interacted with PPARalpha with high affinity (nM K(d)s), validating the primary screen. Furthermore, these compounds altered PPARalpha transactivation, and 1 strongly reversed the hyperactivation of PPARalpha found in the presence of clofibrate and high glucose levels.     

A phytol-enriched diet induces changes in fatty acid metabolism in mice both via PPARalpha-dependent and -independent pathways

Branched-chain fatty acids (such as phytanic and pristanic acid) are ligands for the nuclear hormone receptor peroxisome proliferator-activated receptor alpha (PPARalpha) in vitro. To investigate the effects of these physiological compounds in vivo, wild-type and PPARalpha-deficient (PPARalpha-/-) mice were fed a phytol-enriched diet. This resulted in increased plasma and liver levels of the phytol metabolites phytanic and pristanic acid. In wild-type mice, plasma fatty acid levels decreased after phytol feeding, whereas in PPARalpha-/- mice, the already elevated fatty acid levels increased. In addition, PPARalpha-/- mice were found to be carnitine deficient in both plasma and liver. Dietary phytol increased liver free carnitine in wild-type animals but not in PPARalpha-/- mice. Investigation of carnitine biosynthesis revealed that PPARalpha is likely involved in the regulation of carnitine homeostasis. Furthermore, phytol feeding resulted in a PPARalpha-dependent induction of various peroxisomal and mitochondrial beta-oxidation enzymes. In addition, a PPARalpha-independent induction of catalase, phytanoyl-CoA hydroxylase, carnitine octanoyltransferase, peroxisomal 3-ketoacyl-CoA thiolase, and straight-chain acyl-CoA oxidase was observed. In conclusion, branched-chain fatty acids are physiologically relevant ligands of PPARalpha in mice. These findings are especially relevant for disorders in which branched-chain fatty acids accumulate, such as Refsum disease and peroxisome biogenesis disorders.     

Metabolism of phytol to phytanic acid in the mouse, and the role of PPARalpha in its regulation

Phytol, a branched-chain fatty alcohol, is the naturally occurring precursor of phytanic and pristanic acid, branched-chain fatty acids that are both ligands for the nuclear hormone receptor peroxisome proliferator-activated receptor alpha (PPARalpha). To investigate the metabolism of phytol and the role of PPARalpha in its regulation, wild-type and PPARalpha knockout (PPARalpha-/-) mice were fed a phytol-enriched diet or, for comparison, a diet enriched with Wy-14,643, a synthetic PPARalpha agonist. After the phytol-enriched diet, phytol could only be detected in small intestine, the site of uptake, and liver. Upon longer duration of the diet, the level of the (E)-isomer of phytol increased significantly in the liver of PPARalpha-/- mice compared with wild-type mice. Activity measurements of the enzymes involved in phytol metabolism showed that treatment with a PPARalpha agonist resulted in a PPARalpha-dependent induction of at least two steps of the phytol degradation pathway in liver. Furthermore, the enzymes involved showed a higher activity toward the (E)-isomer than the (Z)-isomer of their respective substrates, indicating a stereospecificity toward the metabolism of (E)-phytol. In conclusion, the results described here show that the conversion of phytol to phytanic acid is regulated via PPARalpha and is specific for the breakdown of (E)-phytol.     

Pristanic acid and phytanic acid: naturally occurring ligands for the nuclear receptor peroxisome proliferator-activated receptor alpha

Phytanic acid and pristanic acid are branched-chain fatty acids, present at micromolar concentrations in the plasma of healthy individuals. Here we show that both phytanic acid and pristanic acid activate the peroxisome proliferator-activated receptor alpha (PPARalpha) in a concentration-dependent manner. Activation is observed via the ligand-binding domain of PPARalpha as well as via a PPAR response element (PPRE). Via the PPRE significant induction is found with both phytanic acid and pristanic acid at concentrations of 3 and 1 microM, respectively. The trans-activation of PPARdelta and PPARgamma by these two ligands is negligible. Besides PPARalpha, phytanic acid also trans-activates all three retinoic X receptor subtypes in a concentration-dependent manner. In primary human fibroblasts, deficient in phytanic acid alpha-oxidation, trans-activation through PPARalpha by phytanic acid is observed. This clearly demonstrates that phytanic acid itself, and not only its metabolite, pristanic acid, is a true physiological ligand for PPARalpha. Because induction of PPARalpha occurs at ligand concentrations comparable to the levels found for phytanic acid and pristanic acid in human plasma, these fatty acids should be seen as naturally occurring ligands for PPARalpha.These results demonstrate that both pristanic acid and phytanic acid are naturally occurring ligands for PPARalpha, which are present at physiological concentrations.     

Fatty-acid biosynthesis in a branched-chain alpha-keto acid dehydrogenase mutant of Streptomyces avermitilis

Fatty-acid biosynthesis by a branched-chain alpha-keto acid dehydrogenase (bkd) mutant of Streptomyces avermitilis was analyzed. This mutant is unable to produce the appropriate precursors of branched-chain fatty acid (BCFA) biosynthesis, but unlike the comparable Bacillus subtilis mutant, was shown not to have an obligate growth requirement for these precursors. The bkd mutant produced only straight-chain fatty acids (SCFAs) with membrane fluidity provided entirely by unsaturated fatty acids (UFAs), the levels of which increased dramatically compared to the wild-type strain. The levels of UFAs increased in both the wild-type and bkd mutant strains as the growth temperature was lowered from 37 degrees C to 24 degrees C, suggesting that a regulatory mechanism exists to alter the proportion of UFAs in response either to a loss of BCFA biosynthesis, or a decreased growth temperature. No evidence of a regulatory mechanism for BCFAs was observed, as the types of these fatty acids, which contribute significantly to membrane fluidity, did not alter when the wild-type S. avermitilis was grown at different temperatures. The principal UFA produced by S. avermitilis was shown to be delta 9-hexadecenoate, the same fatty acid produced by Escherichia coli. This observation, and the inability of S. avermitilis to convert exogenous labeled palmitate to the corresponding UFA, was shown to be consistent with an anaerobic pathway for UFA biosynthesis. Incorporation studies with the S. avermitilis bkd mutant demonstrated that the fatty acid synthase has a remarkably broad substrate specificity and is able to process a wide range of exogenous branched chain carboxylic acids into unusual BCFAs.     

Effect of clofibrate on the CoA thioester profile in rat liver.

Acute and chronic treatment with clofibrate increased the total CoA content of rat liver and altered the profile of the various CoA thioesters. There resulted a 2–3 fold increase in the contents of long chain acyl CoA, acetyl CoA and free CoA, contrasting with significant decreases found in succinyl CoA, malonyl CoA and acetoacetyl CoA contents. It is postulated that the known increase in fatty acid binding protein and/or the increased extramitochondrial β-oxidation of fat by an increased peroxisomal population may direct the compartmentation and metabolic fate of fatty acids and their CoA derivatives following clofibrate treatment.     

Very long chain fatty acid beta-oxidation by rat liver mitochondria and peroxisomes

Crude mitochondrial fractions were isolated by differential centrifugation of rat liver homogenates. Subfractionation of these fractions on self-generating continuous Percoll gradients resulted in clearcut separation of peroxisomes from mitochondria. Hexacosanoic acid beta-oxidation was present mainly in peroxisomal fractions whereas hexacosanoyl CoA oxidation was present in the mitochondrial as well as in the peroxisomal fractions. The presence of much greater hexacosanoyl CoA synthetase activity in the purified preparations of microsomes and peroxisomes compared to mitochondria, suggests that the synthesis of coenzyme A derivatives of very long chain fatty acids (VLCFA) is limited in mitochondria. We postulate that a specific VLCFA CoA synthetase may be required to effectively convert VLCFA to VLCFA CoA in the cell. This specific synthetase activity is absent from the mitochondrial membrane, but present in the peroxisomal and the microsomal membranes. We postulate that substrate specificity and the subcellular localization of the specific VLCFA CoA synthetase directs and regulates VLCFA oxidation in the cell.     

beta-ketoacyl-acyl carrier protein synthase III (FabH) is a determining factor in branched-chain fatty acid biosynthesis

A universal set of genes encodes the components of the dissociated, type II, fatty acid synthase system that is responsible for producing the multitude of fatty acid structures found in bacterial membranes. We examined the biochemical basis for the production of branched-chain fatty acids by gram-positive bacteria. Two genes that were predicted to encode homologs of the beta-ketoacyl-acyl carrier protein synthase III of Escherichia coli (eFabH) were identified in the Bacillus subtilis genome. Their protein products were expressed, purified, and biochemically characterized. Both B. subtilis FabH homologs, bFabH1 and bFabH2, carried out the initial condensation reaction of fatty acid biosynthesis with acetyl-coenzyme A (acetyl-CoA) as a primer, although they possessed lower specific activities than eFabH. bFabH1 and bFabH2 also utilized iso- and anteiso-branched-chain acyl-CoA primers as substrates. eFabH was not able to accept these CoA thioesters. Reconstitution of a complete round of fatty acid synthesis in vitro with purified E. coli proteins showed that eFabH was the only E. coli enzyme incapable of using branched-chain substrates. Expression of either bFabH1 or bFabH2 in E. coli resulted in the appearance of a branched-chain 17-carbon fatty acid. Thus, the substrate specificity of FabH is an important determinant of branched-chain fatty acid production.     

Changes in enzymatic activities involved in glucose metabolism by acyl-CoAs in Trypanosoma cruzi

This study describes the effect of some saturated and unsaturated free fatty acids and acyl-CoA thioesters on Trypanosoma cruzi glucose 6-phosphate dehydrogenase and hexokinase activities. Glucose 6-phosphate dehydrogenase was sensitive to the destabilizing effect provoked by free fatty acids, while hexokinase remained unaltered. Glucose 6-phosphate dehydrogenase inhibition by free fatty acids was dependent on acid concentration and chain length. Both enzymes were inhibited when they were incubated with acyl-CoA thioesters. The acyl-CoA thioesters inhibited glucose 6-phosphate dehydrogenase at a lower concentration than the free fatty acids; the ligands glucose 6-phosphate and NADP+ afforded protection. The inhibition of hexokinase by acyl-CoAs was not reverted when the enzyme was incubated with ATP. The type of inhibition found with acyl-CoAs in relation to glucose 6-phosphate dehydrogenase and hexokinase suggests that this type inhibition may produce an in vivo modulation of these enzymatic activities.     

The role of 4'-phosphopantetheine in t' biosynthesis of fatty acids, polyketides and peptides

The peptide part of CoA, 4'-phosphopantetheine, has been identified as an essential cofactor in in the biosynthesis of fatty acids, polyketides, depsipeptides, peptides, and compounds derived from both carboxylic and amino acid precursors, like rapamycin. The cofactor is attached to a unique protein moiety, referred to as acyl carrier protein, aminoacyl carrier protein, or peptidyl carrier protein. These carrier proteins are either associated to enzyme complexes (type II) or integrated within multifunctional polypeptide chains (type I). The cofactor is added in a post-translational modification reaction by highly specific transferases, acting on CoA. The functions of carrier proteins in directed condensation reactions are: (1) the selection of substrates to be attached as thioesters, (2) the stabilization of intermediates, (3) the presentation of intermediates for modification by associated enzyme activities, (4) facilitation of the directed condensation reactions of two adjacent intermediates, and (5) assistance in the termination reaction(s) leading to product release.     

Glucose directly links to lipid metabolism through high affinity interaction with peroxisome proliferator-activated receptor alpha

The pathophysiology of diabetes is characterized not only by elevated glucose but also elevated long chain fatty acid levels. We show for the first time that the peroxisome proliferator-activated receptor-alpha (PPARalpha) binds glucose and glucose metabolites with high affinity, resulting in significantly altered PPARalpha secondary structure. Glucose decreased PPARalpha interaction with fatty acid metabolites and steroid receptor coactivator-1 while increasing PPARalpha interaction with DNA. Concomitantly, glucose increased PPARalpha interaction with steroid receptor coactivator-1, DNA binding, and transactivation of beta-oxidation pathways in the presence of activating ligands. Heterodimerization of PPARalpha to the retinoid X receptor-alpha resulted in even larger increases in transactivation with the addition of glucose. These data suggest that PPARalpha is responsible for maintaining energy homeostasis through a concentration-dependent regulation of both lipids and sugars and that hyperglycemic injury mediated by PPARalpha occurs not only indirectly through elevated long chain fatty acid levels but also through direct action of glucose on PPARalpha.     

Novel form of lipolysis induced by leptin.

Hyperleptinemia causes disappearance of body fat without a rise in free fatty acids (FFA) or ketones, suggesting that leptin can deplete adipocytes of fat without releasing FFA. To test this, we measured FFA and glycerol released from adipocytes obtained from normal lean Zucker diabetic fatty rats (+/+) and incubated for 0, 3, 6, or 24 h in either 20 ng/ml recombinant leptin or 100 nM norepinephrine (NE). Whereas NE increased both FFA and glycerol release from adipocytes of +/+ rats, leptin increased glycerol release in +/+ adipocytes without a parallel increase in FFA release. In adipocytes of obese Zucker diabetic fatty rats (fa/fa) with defective leptin receptors, NE increased both FFA and glycerol release, but leptin had no effect on either. Leptin significantly lowered the mRNA of leptin and fatty acid synthase of adipocytes (FAS) (p < 0.05), and up-regulated the mRNA of peroxisome proliferator-activated receptor (PPAR)-alpha, carnitine palmitoyl transferase-1, (CPT-1), and acyl CoA oxidase (ACO) (p < 0.05). NE (100 nM) also lowered leptin mRNA (p < 0.05) but did not affect FAS, PPARalpha, ACO, or CPT-1 expression. We conclude that in normal adipocytes leptin directly decreases FAS expression, increases PPARalpha and the enzymes of FFA oxidation, and stimulates a novel form of lipolysis in which glycerol is released without a proportional release of FFA.     

Fatty acid-related phylogeny of myxobacteria as an approach to discover polyunsaturated omega-3/6 Fatty acids

In an analysis of 47 aerobic myxobacterial strains, representing 19 genera in suborders Cystobacterineae, Nannocystineae, Sorangiineae, and a novel isolate, "Aetherobacter" SBSr008, an enormously diverse array of fatty acids (FAs) was found. The distribution of straight-chain fatty acids (SCFAs) and branched-chain fatty acids (BCFAs) supports the reported clustering of strains in the phylogenetic tree based on 16S rRNA genes. This finding additionally allows the prediction and assignment of the novel isolate SBSr008 into its corresponding taxon. Sorangiineae predominantly contains larger amounts of SCFA (57 to 84%) than BCFA. On the other hand, Cystobacterineae exhibit significant BCFA content (53 to 90%), with the exception of the genus Stigmatella. In Nannocystineae, the ratio of BCFA and SCFA seems dependent on the taxonomic clade. Myxobacteria could also be identified and classified by using their specific and predominant FAs as biomarkers. Nannocystineae is remarkably unique among the suborders for its absence of hydroxy FAs. After the identification of arachidonic (AA) FA in Phaselicystidaceae, eight additional polyunsaturated fatty acids (PUFAs) belonging to the omega-6 and omega-3 families were discovered. Here we present a comprehensive report of FAs found in aerobic myxobacteria. Gliding bacteria belonging to Flexibacter and Herpetosiphon were chosen for comparative analysis to determine their FA profiles in relation to the myxobacteria.     

Inhibition of swarming motility of Pseudomonas aeruginosa by branched-chain fatty acids.

Pseudomonas aeruginosa is capable of moving by swimming, swarming, and twitching motilities. In this study, we investigated the effects of fatty acids on Pseudomonas aeruginosa PAO1 motilities. A branched-chain fatty acid (BCFA)--12-methyltetradecanoic acid (anteiso-C15:0)--has slightly repressed flagella-driven swimming motility and completely inhibited a more complex type of surface motility, i.e. swarming, at a concentration of 10 microg mL(-1). In contrast, anteiso-C15:0 exhibited no effect on pili-mediated twitching motility. Other BCFAs and unsaturated fatty acids tested in this study showed similar inhibitory effects on swarming motility, although the level of inhibition differed between these fatty acids. These fatty acids caused no significant growth inhibition in liquid cultures. Straight-chain saturated fatty acids such as palmitic acid were less effective in swarming inhibition. The wetness of the PAO1 colony was significantly reduced by the addition of anteiso-C15:0; however, the production of rhamnolipids as a surface-active agent was not affected by the fatty acid. In addition to motility repression, anteiso-C15:0 caused 31% repression of biofilm formation by PAO1, suggesting that BCFA could affect the multiple cellular activities of Pseudomonas aeruginosa.     

Branched-chain fatty acid biosynthesis in Escherichia coli

β-Ketoacyl-acyl carrier protein (ACP) synthase III (KASIII) catalyzes the first elongation step in straight-chain fatty acid (SCFA) biosynthesis in Escherichia coli. Overproduction of the corresponding KASIII gene, or the Brassica napus KASIII gene has previously been observed to lead to an increase in the amount of shorter-chain fatty acids produced by E. coli. In this study it is shown that overexpression of the KASIII gene, which initiates branched-chain fatty acid (BCFA) in Streptomyces glaucescens, does not lead to a change in the fatty acid profiles of E. coli. E. coli produces trace levels of BCFAs when grown in the presence of isobutyric acid, but the amounts of these are not significantly altered by expression of the S. glaucescens KASIII gene. In contrast, the amounts of BCFAs produced from isobutyryl CoA in vitro by E. coli cell-free extracts can be increased at least four-fold by the presence of the S. glaucescens KASIII. These observations suggest that in vivo production of isopalmitate by E. coli expressing the S. glaucescens KASIII is limited by availability of the appropriate BCFA biosynthetic primers. Journal of Industrial Microbiology & Biotechnology (2001) 27, 246–251. Received 10 January 2001/ Accepted in revised form 13 July 2001