Fatty acid transport in bacteria

Fatty acid uptake systems

Import of fatty acids in E. coli:the FadL-FACS system

The first discovered microbial transport systems for fatty acids was the FadL-FACS system from E. coli for the import of exogenous long-chain fatty acids. It was described in 1969 that the ability ofE. coli to degrade exogenous fatty acids was associated to the presence of a fatty acid-CoA synthetase (FACS). Characterization of this enzyme suggested the presence of another transport protein which was discovered later in 1978 (Nunn & Simons, 1978). The gene encoding this protein, named fadL , belongs to the fatty acid degrading (fad) regulon containing eight catabolic genes for fatty acids (Feng & Cronan, 2009), including fadD which codes for the FACS. The regulon is controlled by the fadR gene, whose product activates the fad genes upon binding to acyl-CoA. Both fadL andfadD are expressed in a basal level to allow detection of exogenous fatty acids by fadR (Nils Joakim Færgeman & Knudsen, 1997).
FadL is a long-chain fatty acid-specific transporter with a β-barrel structure present in the outer membrane of E. coli . The presence of lipopolysaccharides in the outer membrane renders the cell impermeable to hydrophobic molecules such as fatty acids. FadL is necessary for the uptake of exogenous fatty acids so that they can be used both as a source of energy and as a constituent of phospholipids and triacylglycerols, relieving the bacteria from spending energy and resources into synthetizing new fatty acids (Van Den Berg, 2005). The structure and mechanism of FadL has been studied by Van der Berg et al (Van Den Berg, Black, Clemons, & Rapoport, 2004). FadL is a long 14-strand β-barrel that contains special features: in the extracellular region two loops containing α-helices form a hydrophobic groove; in the intracellular region a hatch domain with three α-helices blocks the channel; the N-terminus extends through the barrel towards the extracellular regions; finally, the strand S3 shows a peculiar bend or kink that disrupts the β-sheet formation and forms a lateral opening (Figure 1A). Concerning the mechanism of transport, the authors initially considered at a first stage that the fatty acids would bind first to the hydrophobic groove, then they would diffuse to an internal high-affinity binding pocket where they would lead to conformation changes both in the N terminus and the hatch domain opening a path for their liberation in the periplasmic space (Van Den Berg, 2005). Nonetheless, this model was refuted after showing the rigidity of the hatch domain (Hearn, Patel, Lepore, Indic, & Van Den Berg, 2009). The importance of the lateral opening in the β-barrel caused by the S3 kink led to support the hypothesis of lateral diffusion, in which the fatty acids would not be liberated to the periplasmic space but rather through the opening space in the β-barrel to the outer membrane (Lepore et al., 2011; Figure 1B). A summary of studied residues from FadL can be found in Table 1, all of them conserved in homologue fatty acids transporters from other species.
Once the fatty acids have been transported by FadL through the outer membrane, they move to the inner membrane. Even though there are reports that proteins may be involved in the process (A. Azizan & Black, 1994), it is likely that this step happens spontaneously. A piece of evidence that supports passive diffusion from the outer membrane to the inner membrane, as well as a flip-flop movement inside each membrane, is the necessity of a proton motive force that can generate acidic conditions in the periplasm leading to the protonation of fatty acid and the subsequent increase in liposolubility (Azliyati Azizan, Sherin, DiRusso, & Black, 1999). Once in the inner membrane, these fatty acids are catalysed by FACS to form acyl-CoA molecules, which are destined to further catabolic or anabolic processes. FACS is a soluble protein that is recruited to the membrane to assist in the transport of fatty acids (Overath, Pauli, & Schaire, 1969). The mechanism of this recruitment is not known and it possibly happens due to conformational changes after binding of ATP and it has been observed to be assisted by the presence of D-lactate, among other conditions (Mangroo & Gerber, 1993). The activity of FACS is necessary for the metabolism of exogenous fatty acids and it is speculated that it promotes the transport of fatty acids through vectorial acylation analogous to the vectorial phosphorylation of sugars (Black & Dirusso, 2003).

Engineering the fatty acid import system of E. coli for biotechnological applications

The import of fatty acid into cells is important for several industrial processes, such as fermentation processes using fatty acids and waste oils as carbon source. The cellular catabolism of fatty acids leads to the formation of acetyl-CoA, which delivers the acetyl group to the citric acid cycle providing energy and intermediates to the cell. Furthermore, acetyl-CoA is a precursor for important metabolites, such as branched-amino acids (Amorim Franco & Blanchard, 2017).
Engineering efforts to produce 3-hydroxypropionate (bioplastic precursor) from fatty acids showed the importance of controlling the expression of the FadL-FACS system (B. Liu et al., 2019). While using a strong promoter for FACS resulted in an increase in productivity, it led to an important decrease when used for FadL. However, the use of medium or weak constitutive promoters for FadL allowed to increase the productivity. This exemplifies the common problem associated to overexpression of membrane proteins, which can increase membrane stress and decrease cell growth and overall productivity (Kang & Tullman-Ercek, 2018). The constitutive expression of FACS and FadL, next to other metabolic modifications, led to a 3-hydroxypropionate yield of 1.56 g/g when grown in a 5L bioreactor using palmitic acid as substrate. In another study where E. coli was engineered for the production of lycopene (N. Liu et al., 2020), overexpression of FACS led to a small increase in the lycopene titre (from 29 to 33 mg/g DCW). Nevertheless, growth of the engineered strain in a mix of glucose, waste oil and yeast extract allowed for a total yield of 94 mg/g DCW.
Another important set of applications for the import of fatty acids are whole cell biotransformations. In these, bacteria grown previously to reach a certain biomass density act as catalysers to modify fatty acids or derived compounds through a limited number of enzymatic steps in order to produce a compound of higher value. In such process, cells act as capsules containing enzymes and are not grown during the biotransformation process. Deletion of FACS and overexpression of FadL showed an increase on fatty acid hydroxylation when expressing the cytochrome P450 CYP153A from Marinobacter aquaeolei, a heterologous cytosolic enzyme in E. coli, suggesting that other proteins can replace the role of FACS in recruiting free fatty acids from the inner membrane and modifying them (Bae, Park, Jung, Lee, & Kim, 2014). Another study involving both fatty acids and hydroxy-fatty acids determined the impact of FadL expression on the biotransformation of these compounds to hydroxy-fatty acids and keto-fatty acids, respectively (Jeon et al., 2018). Enhanced expression of FadL led to a five-fold increase in the single step transformation of oleic acid and 10-hydroxyoctadecanoic acid to 10-hydroxyoctadecanoic acid and 10-keto-octadecanoic, respectively, as well as a two-fold increase in the multistep transformation of ricinoleic acid to the ester molecule ((Z)-11-(heptanoyloxy)undec-9-enoic acid). This study also showed the negative effects that excessive overexpression of FadL can have on overall productivity. Finally, in another study FadL was overexpressed in a strain expressing human Cav1 proteins (Shin et al., 2019). Cav1 proteins stimulate the formation of endosomes that excise from the inner membrane. The formation of endosomes increased the uptake rates of ricinoleic acid two-fold and caused a decrease of fatty acid toxicity in the inner membrane. However, overexpression of FadL in this strain did not lead to a further increase in fatty acid uptake.

Import of fatty acids in other bacteria

Besides the fadL-FACS system from E. coli , other systems for the import and assimilation of exogenous fatty acids have been studied in other bacteria. All bacterial fatty acid import systems were found to be dependent on energy supply whether in the form of ATP or in the form of a proton gradient (Calmes & Deal, 1976). One of the first systems studied was that of other gram-negative bacteria such asPseudomonas oleovorans and Caulobacter crecentus , which showed similar characteristics to the fadL-FACS system (Toscano & Hartline, 1973; Zalatan & Black, 2011). Due to the different plasma membrane structure, it has been observed that import of fatty acids in gram positive bacteria occurs differently than in gram-negative bacteria (Figure 2A and 2B). Although fatty acid metabolism and transport has not extensively been studied in gram positive model bacteria, such as B. subtilis, some studies have been performed on other gram-positive bacteria. In the lipophilic gram-positive bacteria Nocardia asteroides (Calmes & Deal, 1976), the fatty acid uptake system was, in contrast to the previously discussed systems, found to be constitutively expressed and able to import fatty acids as free fatty acids, which would be activated by a soluble FACS at a later and independent stage. With a KM of 870 μM, the transport process from N. asteroides was found to have much less affinity than that of E. coli (KM between 15 and 34 μM). The import of fatty acids in another gram-positive bacterium,Streptomyces coelicolor (Banchio & Gramajo, 1997), also showed a constitutive expression, as well as a dependency on pH levels in the environment, decreasing the uptake of long-chain fatty acids with increasing pH. The import system of this bacterium was found to be specific for ionized fatty acids (at pH 7 or higher), while protonated ones were claimed to use a passive process rather than an active one (at a pH lower than 7).
In the fadL-FACS system, FACS is associated to the membrane, linking the formation of acyl-CoA to transport. Nevertheless, the production of acyl-CoA from exogenous fatty acids is not conserved in all bacteria, and other ways of activating and incorporating exogenous fatty acids have been studied, such as acyl-ACP synthase in Vibrio harveyi,or fatty acid kinases in gram positive bacteria (Cronan, 2014). The different ways to in which exogenous fatty acids are activated and metabolized has been reviewed elsewhere (Yao & Rock, 2017).
The study of fatty acid import in bacteria can serve different interests, such as improving the fatty acid import capacities of organisms of industrial importance or identifying new drug targets for pathogenic bacteria able to import fatty acids from the host. Among the different bacteria that are of interest for industry, the gram-positive genus Rhodococcus has attracted special attention for their ability to withstand and degrade a wide variety of pollutant, including hydrocarbons and lignin-derived compounds (Kim, Choi, Yoo, Zylstra, & Kim, 2018). Some species of this genus, such as Rhodococcus opacus and Rhodococcus jostii, can be considered oleaginous, being able to accumulate high amounts of triacylglycerol when growing in carbon-excess conditions (Alvarez et al., 2019). A fatty acid importer from R. jostii was found in a cluster with genes involved in lipid metabolism after a homology search using known ACB-transporters for hydrophobic compounds. The function of this gene (ro05645 or ltp1) was confirmed after its overexpression, which led to an increased growth in media containing fatty acids as substrate, as well as the uptake of fluorescently labelled fatty acids (Villalba & Alvarez, 2014). Furthermore, the overexpression of ltp1 was used to increase both the growth (2.2-fold) and the lipid accumulation (3.5-fold) of R. jostii when growing in olive mill waste (Herrero, Villalba, Lanfranconi, & Alvarez, 2018).
Mycobacterium tuberculosis is the causing agent of tuberculosis, one of the deadliest diseases in the world. This bacterium is adapted to survive for long periods of time inside the human host, partially thanks to its ability to metabolize host-derived fatty acids. This ability is especially important, as these cells tend to reside within lipid-rich sites, such as inside foamy macrophages (Lovewell, Sassetti, & VanderVen, 2016). M. tuberculosis presents a cell envelope containing mycolic acid and glycolipids which acts as a barrier for hydrophobic molecules. A multiprotein complex was observed to be involved in the uptake of fatty acids during macrophage infection as well as in monocultures (Figure 2C). The first member of this complex to be identified was LucA, a protein also involved in the uptake of cholesterol (Nazarova et al., 2017). It was observed that a mutant lacking LucA was unable to incorporate fatty acids, and a transcriptomic analysis of such mutant revealed the involvement of genes from the Mce1 locus in the uptake of fatty acids. From this locus, gene rv0167coding for the putative permease YrbE1A was found to be directly involved in fatty acid import. A further screening of mutants lacking the ability to assimilate fatty acids when infecting macrophages revealed the participation of other proteins (Nazarova et al., 2019), such as MceD1, MceG and OmamB. While the function of OmamB is not known, a similar protein OmamA was found to interact with LucA and stabilize the fatty acid import complex. MceG is a putative ATPase that may be involved in providing the energy for the transport of the fatty acids. Finally, MceD1 and other proteins from the same locus such as MceA1 and MceC1, may have structural roles in the formation of the fatty acid import complex. In addition to the complex Mce1, another gene (rv1272 ) from M. tuberculosis showing homology to theltp1 gene from R. jostii as well as to the lipid A export gene msbA from E. coli was confirmed to import fatty acids when expressed in E. coli (Martin & Daniel, 2018). A Blast search of rv1272 against the SwissProt database showed that this gene is homologous (39.39% identity) to the uncharacterized transporter YfiC from Bacillus subtilis, suggesting that the role of this transporter might be associated to the transport of fatty acids or other hydrophobic compounds.

Fatty acid export systems

The fatty acid export system in E. coli and its engineering

In its natural environment, E. coli must face high concentrations of fatty acids and other hydrophobic compounds and therefore it must contain mechanisms to prevent toxic effects, such as export proteins. At the same time, secretion of endogenous fatty acids is not detected for wild type strains under normal conditions. Nevertheless, fatty acid secretion in E. coli was observed for the first time when expressing a thioesterease from the plant Umbelleria californica(Voelker & Davies, 1994). This enzyme was expressed to generate free fatty acids inside the cell. In normal conditions, free fatty acids do not accumulate intracellularly in bacteria, as they are directly transferred from acyl-ACP to glycerol-3-phosphate (Magnuson et al., 1993). On the other hand, plants possess thioesterases to liberate fatty acids from acyl-ACP and these free fatty acids are used for anabolic processes (Gerhardt, 1992). The oilseed from U. californica accumulates medium-chain fatty acids, mainly lauric acid (C12:0) (Davies, Anderson, Fan, & Hawkins, 1991), and the expression of its thioesterase in E. coli leads to the same fatty acid profile (Voelker & Davies, 1994). However, the accumulation of fatty acids was only observed when the β-oxidation pathway was disabled. Secretion of lauric acid was shown by the observation of extracellular laurate crystals when growing on solid media. Later engineering of E. coli showed that saturated and monounsaturated fatty acids with 12 till 18 carbon atoms can be secreted in an efficient way reaching 40 mg/L of extracellular fatty acids, but no specific efflux protein for fatty acids was identified (H. Liu et al., 2012).
The accumulation of medium-chain fatty acids is toxic to bacteria due to their ability to destabilize the membrane and interfere with essential activities, such as creation of proton gradients (Lennen et al., 2011). This toxic activity was used to identify proteins involved in fatty acid export that would protect the cells from the accumulation of medium-chain fatty acids when expressing the thioesterase from U. californica (Lennen, Politz, Kruziki, & Pfleger, 2013). The genes whose knock out led to important effects on cell viability under these conditions were rob, acrAB, tolC and, at a lower extent,emrAB . The activity of the first three genes is tightly linked, as rob is a regulatory gene that induced acrAB expression among other genes, and TolC is an unspecific porin from the outer membrane that forms a complex with acrAB. The AcrAB-TolC complex is a transporter from the Resistance-Nodule-Division (RND) superfamily that spans from the intracellular space until the outer side of the bacterial membrane, forming a channel whose mechanism is regulated by allosteric changes and a proton gradient (Wang et al., 2017; Zgurskaya & Nikaido, 1999). The complex is formed by three TolC components, six AcrA components and three AcrB components (Figure 3). The mechanism behind the binding of ligands to AcrB is not precisely known, but due to the large number of different substrates that the AcrAB-TolC complex can transport it is thought that there might be several binding sites in the same binding pocket (Nakashima, Sakurai, Yamasaki, Nishino, & Yamaguchi, 2011). In any case, the binding of a substrate changes the conformation of the AcrB monomer from the relaxed or access state to a binding state, which induces conformational changes in the rest of the complex. Through several interactions with AcrB, AcrA continues the conformational changes to TolC, so that it can change from the closed state to the open state, opening the channel. Finally, the AcrB subunit bound to the ligand changes to another conformation (open state), liberating the substrate into the open channel for its liberation to the extracellular medium (Wang et al., 2017).
Due to the broad specificity of the AcrAB-TolC system, the AcrB component has been engineered through directed evolution to improve the export rate of different hydrophobic molecules, such as medium-chain alcohols, alkanes and alkenes (Chen, Ling, & Chang, 2013; Fisher et al., 2014; Mingardon et al., 2015). However, no protein engineering of AcrB to improve fatty acid export has been documented. Genetic engineering to improve medium-chain fatty acid export in E. coliwas performed by overexpressing several potential transporters (J. Wu et al., 2019). Overexpression of either of three transport proteins, namely AcrE, MdtC and MdtE, was found to increase medium-chain fatty acid extracellular concentrations from 600 mg/L to values between 800 and 1100 mg/L. One must realize that these proteins are part of larger protein complexes whose other components were not overexpressed in the study.
The AcrEF complex displays homology to the well-studied AcrAB-TolC complex. Yet, the AcrEF complex has been observed to be expressed at lower levels than the AcrAB complex and it has shown an important function in cell division and chromosome segregation (Lau & Zgurskaya, 2005). Nevertheless, the AcrAB and AcrEF complexes are from a structural point of view highly related: the amino acid sequence of acrF shares 87% similarity with acrB, and acrE is homologous to acrA (80% AA similarity). Therefore, it is likely that acrE fulfils the same role as acrA in the acrEF-TolC complex as the channel that connects acrF and TolC. Yet, when overexpressing acrA and acrB from the acrAB-TolC complex separately, no increased fatty acid export was observed. Hence, the mechanism behind the overexpression leading to increased export activity of a channel protein that connects the inner membrane pump (which contains the substrate binding) and the outer membrane porin remains unsolved. Furthermore, in the same study three-fold higher extracellular medium-chain fatty acid concentrations were achieved when overexpressing the three transport proteins (acrE, mdtC and mdtE) simultaneously, while observing only a 10% reduction in the OD (J. Wu et al., 2019).

Export of fatty acids in other bacteria

As observed in the previous section, fatty acid export in E. coliseems to be mainly associated to protection against membrane-related toxic effects caused by large concentrations of medium and long-chain fatty acids (Desbois & Smith, 2010). Medium-chain fatty acids have antimicrobial effects affecting a wide range of bacteria (Huang, Alimova, Myers, & Ebersole, 2011). Therefore, it is expected that many bacteria present export systems similar to AcrAB-TolC to export toxic fatty acids, although the specific systems have not been studied to date.
Besides medium chain fatty acids, polyunsaturated fatty acids display antimicrobial properties and in reaction to this, several pathogenic bacteria have developed systems to prevent their toxicity. Some of these systems are the degradation of exogenous fatty acids or their incorporation into phospholipids (Jiang et al., 2019). However, the environments where some pathogenic bacteria must thrive in, such as the skin or the mouth, contain unsaturated fatty acids and therefore they need more advanced protection mechanisms such as fatty acid export systems (Choi et al., 2013; Parsons, Yao, Frank, Jackson, & Rock, 2012). In the gram-positive bacterium Staphylococcus aureus , an opportunistic pathogen found on the skin, a fatty acid export system was found when screening for strains resistant to linoleic acid (Alnaseri et al., 2019). The strain S. aureus FAR7 showed a mutation in the transcription factor farR, which led to the upregulation of thefarE gene, encoding an efflux pump from the RND superfamily. Also the physiological response to PUFAs of the gram-negative pathogenic bacteria Acinetobacter baumannii, was studied and revealed the upregulation of the adeJ gene, a component of the multidrug efflux pump adeIJK (Jiang et al., 2019). It was observed that the deletion of this gene increased the susceptibility to PUFAs, leading to a six-fold increased growth delay in the presence of docosahexaenoic acid. However, the mutation did not lead to an accumulation of PUFAs in the cell. Nevertheless, growth experiments showed the ability of adeIJK to affect membrane lipid homeostasis and to export lipids to the extracellular medium. These results suggested that the adeIJK pump fulfils a similar role as the emhABC pump fromPseudomonas fluorescens , which controls lipid homeostasis through the efflux of endogenous long-chain fatty acids, both saturated and monounsaturated, in response to temperature changes (Adebusuyi & Foght, 2011). Another studied fatty acid export system from a pathogenic bacterium is the farAB system from Neisseria gonorrhoeae , which shows high similarity to the emrAB system from E. coli, and whose deletion leads to susceptibility to the long-chain fatty acids oleic acid, linoleic acid and palmitic acid. The minimal inhibitory concentration decreased from 1600 μg/ml to 50 μg/ml in the case of unsaturated fatty acids and from 100 μg/ml to 12.5 μg/ml in the case of palmitic acid (Lee & Shafer, 1999).
Photosynthetic microorganisms are of special interest due to their ability to fix carbon from the atmosphere to produce industrially relevant compounds in a more efficient way than plants.Synechocystis sp. PCC 6803 is the model organism for cyanobacteria and it has been observed to secrete long-chain fatty acids, up to 13% of the cellular biomass, without any genetic modifications (X. Liu, Sheng, & Curtiss, 2011). The deletion ofSynechocystis sp. PCC 6803 genes sll0180 andslr2131, homologous to the respective E. coli genesacrA and acrB, showed an effect in the fatty acid secretion of an engineered strain of Synechocystis sp. PCC 6803 specialized in the extracellular production of fatty acids (Bellefleur, Wanda, & Curtiss, 2019). While the complementation with acrA did not lead to a recovery of the fatty acid secretion rates, the complementation of slr2131 with acrB allowed for an increase in both extracellular and intracellular fatty acid concentrations.
Another fatty acid export mechanism has been identified in another cyanobacteria, Synechococcus elongatus a. This system was found through the genomic and transcriptomic analysis of a mutant able to produce free fatty acids but resistant to their toxicity. Inactivation of this export system, composed by the genes rndA1and rndB1 , led to susceptibility to exogenous saturated medium-chain fatty acids and unsaturated long-chain fatty acids. Orthologs of rndB1 are found in most genomes from cyanobacteria, but not in those of Synechocystis sp. PCC 6803. The RndA1B1 system from S. elongatus allows for efficient secretion of oleic acid, but palmitate is not transported and therefore it accumulates in the cells (Kato et al., 2015).