The puzzle of two tandem acyl-CoA ligases of Pseudomonas putida F1

ABSTRACT

The Pseudomonas putida F1 genome and those of many other pseudomonads contain two tandem genes encoding acyl-CoA ligases Pput_1340 (fadD1) and Pput_1339 (fadD2) with Pput_1339 (fadD2) being the upstream gene. The fadD designation was assigned when both genes were found to complement the growth of an Escherichia coli acyl-CoA synthetase fadD deletion strain with oleic acid as sole carbon source. Site-directed mutagenesis showed that residues of the ATP/AMP domain required for function of E. coli FadD were also essential for full function of FadD1 and FadD2. Growth of the constructed ∆fadD1, ∆fadD2, and ∆fadD1∆fadD2 strains was tested in minimal medium with different chain length fatty acids as sole carbon sources. Lack of FadD1 significantly retarded growth with different chain length fatty acids and lack of both FadD1 and FadD2 further retarded growth. Derivatives of the ∆fabA∆desA unsaturated fatty acid auxotrophic strain carrying a deletion of either ∆fadD1 or ∆fadD2 were constructed. Growth of the ∆fabA∆desA∆fadD1 strain was very weak, whereas the ∆fabA∆desA∆fadD2 strain grew as well as the ∆fabA∆desA parent strain. Overexpression of either fadD1 or fadD2 restored growth of the ∆fabA∆desA∆fadD1 strain with fadD2 overexpression having a greater effect than fadD1 overexpression. The ∆fadD1 or ∆fadD2 genes are cotranscribed although the expression level of fadD1 is much higher than that of fadD2. This is attributed to a fadD1 promoter located within the upstream FadD2 coding sequence.

IMPORTANCE

Pseudomonas bacteria demonstrate a great deal of metabolic diversity and consequently colonize a wide range of ecological niches. A characteristic of these bacteria is a pair of genes in tandem annotated as acyl-CoA ligases involved in fatty acid degradation. The Pseudomonas putida F1 genome is annotated as having at least nine genes encoding acyl-CoA ligases which are scattered around the chromosome excepting the tandem pair. Since similar tandem pairs are found in other pseudomonads, we have constructed and characterized deletion mutants of the tandem ligases. We report that the encoded proteins are authentic acyl-CoA ligases involved in fatty acid degradation.

INTRODUCTION

Bacteria utilize fatty acids as a carbon source mainly by β-oxidation (1). Given a sufficient oxygen supply, fatty acids are oxidized to acetyl-CoA which enters central metabolism mainly via the tricarboxylic acid cycle (2). Exogenous long-chain fatty acids enter the cell through the FadL fatty acid transporter (1) and become activated by conversion to acyl-CoA thioesters by the FadD acyl-CoA ligase. This conversion allows the fatty acid to enter the β-oxidation cycle which converts the acyl chain to acetyl-CoA in the complete β-oxidation cycle (1). The acyl-CoAs can also be incorporated into phospholipids by the PlsB-PlsC acyltransferase enzymes (3). In Pseudomonas putida, 3-hydroxy acyl-CoA thioesters produced during β-oxidation of fatty acid synthesis cycle can also be used as the raw material for the synthesis of the polyhydroxyalkanoate storage compounds (4, 5). Unsaturated fatty acids also undergo β-oxidation, but the cis double bond must first be isomerized to the trans configuration (1). Pseudomonas putida F1, unlike some other P. putida strains, can oxidize and grow on aromatic hydrocarbons (e.g., toluene, benzene, ethylbenzene, and p-cymene) (6–8).

The biochemistry and genetics of the β-oxidation pathway of Escherichia coli have been well studied (1) and that of Salmonella enterica is similar. E. coli also has a fatty acid β-oxidation pathway that functions under anaerobic conditions (9). The acyl-CoA ligase reaction (also called acyl-CoA synthase and acyl-CoA synthetase) is a classical acyl-adenylate enzyme (10) that proceeds through an acyl-AMP (acyl adenylate) intermediate which is attacked by the CoA thiol to give acyl-CoA (Fig. 1A). Structural and mutational studies show that acyl-CoA ligases have two conserved regions, the C-terminal ATP/AMP binding region and the N-terminal fatty acid activation region (11). The three-dimensional structure analysis of a thermophilic bacterium acyl-CoA ligase shows that when the C-terminal ATP binding region binds ATP, the N-terminal fatty acid activation region opens and binds the fatty acid (12). Upon ligation of the fatty acid to CoA, the N-terminal FA activation region closes, which promotes the dissociation of the acyl-CoA (12).

Fig 1.

FadD properties, FCL motifs (also called FACS motif) and P. putida FadD complementation. (A) The acyl-CoA ligase reaction. (B) Growth of an E. coli fadD strain on oleate as sole carbon source. Transformants of the strain E. coli ΔfadDΔaraBAD strain, JW1794, were grown at 37°C on a minimal medium with oleate as the sole carbon source. Growth was tested in either the presence or the absence of arabinose, the inducer of the pBAD promoter of vector pBAD24M that drives expression of the genes. The strains tested were BW25113 (wild type), JW1794 carrying E. coli fadD, P. putida fadD1 or P. putida fadD2, or the vector plasmid, pBAD24M. (C) Sequence alignments of the ATP/AMP and FCL (also called FACS) motifs of the E. coli FadD and P. putida FadD enzymes called the P loop and L sequence in the Thermus thermophilus crystal structure (12). The stars denote the active-site threonine and glutamate residues. (D) Strain JW1794 expressing the wild-type fadD1, T215A, or E362A fadD1 mutant proteins. Wild-type fadD2, T215A, or the E369A fadD2 mutants, or the vector plasmid, pBAD24M were also tested. Strain JW1794 and derived strains lack the ability to grow on arabinose.

Note that the ability of P. putida strains to grow on a great variety of organic compounds including hydrocarbons is reflected in the diversity of putative β-oxidation related enzymes encoded in the genome. For example, in P. putida F1 genes annotated as encoding the first two enzymes of β-oxidation, acyl-CoA ligase and acyl-CoA dehydrogenase, are present in at least 9 and 20 copies, respectively, whereas E. coli has only two acyl-CoA ligases and a single acyl-CoA dehydrogenase (9, 13). A recent detailed review of β-oxidation in E. coli has appeared (1).

RESULTS

Complementation of an E. coli ∆fadD strain by P. putida fadD1 and fadD2

Given the incredible metabolic and genetic diversity of P. putida F1, it was not a given that the Pput_1340 (fadD1) and Pput_1339 (fadD2) genes encoded fatty acyl-CoA ligases. Hence, we tested the proteins encoded by the two putative fadD genes, Pput_1340 (fadD1) and Pput_1339 (fadD2), for acyl-CoA ligase activity. Both FadD1 and FadD2 functionally complemented an E. coli ∆fadD strain for growth with oleic acid as sole carbon source (Fig. 1B). A further test was to mutagenize two highly conserved sequence elements that comprise an ATP/AMP-binding motif (often called the fatty acid-CoA ligase motif) plus a less conserved third sequence element that partially overlaps the ATP/AMP-binding motif (14) (Fig. 1C).

The putative ATP/AMP signature motifs of FadD1 and FadD2 were tested by substitution of alanine for residues T215 and E362 of FadD1 and E369 of FadD2. The E. coli ∆fadD strain carrying plasmids encoding the mutant proteins was tested for complementation activity. The strain expressing the T215A mutant protein grew in the absence of induction, whereas the strains expressing the E362A and E369A mutant proteins failed to grow either in the presence or in the absence of induction (Fig. 1D). Hence, the lack of activity of the E362 and E369 proteins agrees with the E. coli FadD report (14). Growth of the strain expressing the FadD1 or FadD2 T215A protein is consistent with the finding that the residue is not strictly conserved in acyl-adenylate enzymes (10).

Growth phenotypes of P. putida fadD deletion strains

Single and double deletion mutants of P. putida fadD1 and fadD2 were obtained through homologous recombination. The ∆fadD1 and ∆fadD1∆fadD2 strains both showed a lag in growth in both solid and liquid LB medium (Fig. S1). Complementation with plasmid-encoded fadD1 restored wild-type growth to the ∆fadD1 strain, whereas plasmids expressing either fadD1 or fadD2 restored almost wild-type growth to the ∆fadD1∆fadD2 strain with fadD1 giving somewhat better growth than fadD2 in solid LB medium (Fig. S1). The phospholipid fatty acid compositions of the P. putida ∆fadD1, ∆fadD2, and ∆fadD1∆fadD2 strains showed no significant differences from that of the wild-type strains as determined by GC-MS (Table S1). Since LB medium is essentially fatty acid free (it cannot support the growth of fatty acid auxotrophs), the slower growth of the fadD1 and ∆fadD1∆fadD2 strains was surprising and suggested that FadD1 may play a role in central metabolism.

A more specific test was growth of the P. putida ∆fadD1 and ∆fadD2 strains in a minimal medium with various fatty acids as sole carbon sources. The P. putida fadD deletion strains grew almost identically with 0.1% CAA (casein acid hydrolysate) as a carbon source. With fatty acids as sole carbon source, the ∆fadD1 strain grew more slowly on all medium and long chain acids, and the ∆fadD1∆fadD2 strain had a more severe effect on growth (Fig. 2). In contrast, the ∆fadD2 strain grew comparably to the wild-type parental strain. Complementation with plasmid-encoded fadD1 restored the wild-type growth phenotype of the P. putida ∆fadD1 strain, whereas plasmids encoding either fadD1 or fadD2 restored growth of the ∆fadD1∆fadD2 strain (Fig. S2). These results indicate that FadD1 plays the major role in fatty acid utilization and may be the acyl-CoA ligase mainly responsible for the activation of exogenous free FA, whereas FadD2 plays an auxiliary role. Prior studies report the loss of fadD affects bacterial motility (15, 16). The swimming ability of P. putida ∆fadD1 (but not ∆fadD2) strains was also impaired and complementation with wild-type fadD1 or fadD2 genes restored motility (Fig. S3).

Fig 2.

Growth analysis of fadD deletion strains with a variety of fatty acid species as sole carbon source. No growth defects were observed in P. putida fadD deletion strains with casamino acids (CAA) as a carbon source. The ∆fadD1 and ∆fadD1∆fadD2 strains showed a diversity of significant defects when grown with fatty acids of different chain lengths. The analyses were done in triplicate.

Utilization of oleic acid for phospholipid synthesis by the P. putida ligases

Exogenous fatty acids can either be degraded by β-oxidation or be incorporated into membrane lipids. To test the latter pathway, we utilized a P. putida F1 unsaturated fatty acid auxotroph isolated previously (17). P. putida F1 has two pathways to synthesize the unsaturated fatty acids required for functional membrane phospholipids. The fabA gene encodes the enzyme that introduces the double bond during fatty acid synthesis, whereas the desA gene encodes an oxygen-requiring desaturase that introduces double bonds into saturated acyl chains of the membrane phospholipids (17). Strains lacking both enzymes (∆fabA∆desA) are auxotrophic for oleic acid or another unsaturated fatty acid (17).

To study the utilization of oleic acid for phospholipid synthesis, the P. putida, triple deletion strains ∆fabA∆desA∆fadD1 and ∆fabA∆desA∆fadD2 were constructed. In plates containing oleic acid, the ∆desA∆fabA∆fadD1 strain grew very slowly, whereas growth of the ∆fabA∆desA∆fadD2 strain did not differ significantly from that of the ∆fabA∆desA strain (Fig. 3A). Plasmids encoding either fadD1 or fadD2 restored growth of the ∆fabA∆desA∆fadD1 strain, but the fadD1 strain required induction for robust growth (Fig. 3B). The FA compositions of the mutant and complemented strains were determined by GC-MS. The ∆fabA∆desA∆fadD1 and ∆fabA∆desA∆fadD2 strains both have decreased C16:1∆7 contents (Table S2). This fatty acid is produced by shortening of oleic acid (C18:1∆9) by one cycle of β-oxidation prior to incorporation into phospholipid. Both complemented strains restored the C16:1∆7 content of membrane phospholipids (Table S2). [1-¹⁴C]Oleate labeling showed that deletion of fadD1 or fadD2 both decreased incorporation of exogenous oleic acid by the ∆fabA∆desA strain and complementation with the wild-type alleles restored oleate incorporation (Fig. 4). These data indicate that FadD1 is mainly responsible for the activation of exogenous oleic acid for incorporation into phospholipids and for β-oxidation, whereas FadD2 plays only a minor role. Note that the nomenclature of the tandem acyl-CoA ligase genes in Pseudomonas strains differs probably because the gene encoding the more active acyl-CoA ligase was first discovered and named FadD1. The second gene was generally found by sequencing.

Fig 3.

Growth phenotypes of P. putida fadD deletion strains on oleic acid as sole carbon source in the ∆fabA∆desA strain and the complemented strains. (A) Deletion of fadD1 results in defective growth, whereas deletion of fadD2 had little or no effect on growth of the ∆fabA∆desA strain. (B) Plasmids encoding fadD1 or fadD2 restore growth of the ∆fabA∆desA∆fadD1 strain, and fadD2 is more efficient. Plasmid encoded fadD1 fully restores growth of the ∆fabA∆desA∆fadD1 strain but only with IPTG induction, whereas the plasmid encoded fadD2 gave good growth without IPTG induction. The medium was LB medium containing 5 mM oleic acid and 0.5% Brij58. The pSRK vectors are based on the broad-host-range plasmid pBBR1 and contain the E. coli lacI and lacZYA promoter (18).

Fig 4.

Deletion of either fadD1 or fadD2 affects incorporation of exogenous oleate. Deletion of either fadD1 or fadD2 decreased incorporation of exogenous oleic acid by the ∆fabA∆desA strain as assayed by [1-¹⁴C]oleate labeling. Complementation with plasmids encoding either fadD1 or fadD2 increased oleate incorporation. The numbers above the lanes are the incorporation values relative to the ∆fabA∆desA strain (value of 100).

To bypass the fatty acid auxotrophy of the ∆fabA∆desA strain, we expressed the des desaturase gene of Bacillus subtilis to probe possible other functions of FadD1. The plasmid encoding B. subtilis des was transformed into the ∆fabA∆desA∆fadD1 and ∆fabA∆desA∆fadD2 strains to produce strains that grew in the absence of oleic acid. However, growth was slower than in the presence of oleic acid (Fig. 5A). The slow growth of the B. subtilis des complemented strains relative to growth with oleate is probably due to the Des-catalyzed insertion of the double bond at the ∆5 position rather than toward the center of the acyl chains where the kink of the cis double bond gives maximal membrane fluidity. Intriguingly, the ∆fabA∆desA∆fadD1 complemented by the B. subtilis des desaturase grew extremely slowly (Fig. 5A). Plasmid encoded fadD1 or fadD2 genes restored the growth phenotype of the ∆fabA∆desA∆fadD1 strain complemented with B. subtilis des although restoration by fadD1 required IPTG induction (Fig. 5B). [1-¹⁴C]Acetate labeling of these strains showed that the ∆fadD1 deletion significantly decreased fatty acid synthesis and plasmid-encoded fadD1 or fadD2 restored fatty synthesis resulting from deletion fadD1 and the efficiency of fadD2 was much better than fadD1 (Fig. 6). However, although plasmid-encoded fadD1 largely restored fatty acid synthesis in the ∆fabA∆desA∆fadD1 strain, the incorporation was almost entirely into saturated fatty acids suggesting inhibition the Des desaturase (Fig. 6). GC-MS analysis of these strains confirmed the expected synthesis of ∆5 fatty acids (19, 20) as the sole unsaturated phospholipid fatty acid (Table S3).

Fig 5.

Growth phenotypes of strains complemented with B. subtilis des. (A) Expression of the B. subtilis des desaturase gene restores growth of the ∆fabA∆desA∆fadD1 strain in The absence of oleic acid condition. The ∆fabA∆desA∆fadD1 strain is significantly growth deficient relative to growth in the presence of oleic acid, whereas the ∆fabA∆desA∆fadD2 and the ∆fabA∆desA strains complemented with B. subtilis des showed essentially identical growth in the absence of oleic acid. LB medium containing 5 mM oleic acid and 0.5% Brij58. (B) Plasmids expressing fadD1 or fadD2 restored growth in the LB medium of the ∆fabA∆desA∆fadD1 strain complemented with B. subtilis des although fadD2 was more efficient than fadD1. The genes following the slashes(/s) are plasmid encoded.

Fig 6.

Deletion of fadD1 or fadD2 decreases fatty acid desaturation in ∆fabA∆desA strain complemented with B. subtilis des as assayed by [1-¹⁴C]acetate labeling. The genes following the slashes(/s) are plasmid encoded. The numbers above the lanes are the incorporation values relative to the ∆fabA∆desA strain (value of 100). SFA denotes saturated fatty acid chains.

Accumulation of free fatty acids in ∆fadD1 strains

We assayed the ∆fadD1 and ∆fadD2 strains for free fatty acid accumulation and found that ∆fadD1, but not ∆fadD2, showed accumulation. These assays were prompted by reports of free fatty acid accumulation in fadD strains of Sinorhizobium meliloti and E. coli (1, 21, 22) indicating that FadD plays a role in membrane phospholipid maintenance. In E. coli, free fatty acid accumulation is triggered by membrane stresses that activate PldA, an otherwise cryptic outer membrane phospholipase (21). The free fatty acids released by PldA hydrolysis are thought to be transported from the outer membrane to the cell cytosol where they are activated by FadD and recycled into phospholipids (21). These data argue that the accumulation of free fatty acids in the P. putida F1 ∆fadD1 and ∆fadD1∆fadD2 strains (Fig. 7) is due to an inability to recycle the fatty acids resulting from phospholipid degradation. Upon plasmid encoded expression of either FadD1 or FadD2, no free fatty acid accumulation was seen indicating that accumulation was the consequence of loss of FadD1 activity (Fig. 7). P. putida F1 lacks a recognizable PldA homolog. However, two P. aeruginosa genes, mlaZ and mlaY, that encode two proteins thought to act in concert, have been recently reported to function like E. coli PldA (23). Indeed, a P. aeruginosa strain lacking both mlaZ and mlaY has a compromised outer membrane permeability barrier which is restored by the expression of E. coli PldA (23). Convincing homologs (79% and 70% identity, respectively) of P. aeruginosa mlaZ (Pput_3981) and mlaY (Pput_3982) are found in P. putida F1 and, thus, may be involved in the production of free fatty acids we observed in P. putida F1 ∆fadD1 strains.

Fig 7.

Deletion of fadD1 results in extracellular free fatty acid accumulation. Both the ∆fadD1 and ∆fadD1∆fadD2 strains have significant free FA accumulation in the growth medium. The genes following the slashes(/s) are plasmid encoded. Complementation with plasmids that encode fadD1 or fadD2 blocks accumulation in the ∆fadD1 strain. SFA and UFA denote saturated and unsaturated acyl chains, respectively.

fadD1 and fadD2 are cotranscribed, and fadD1 has a dedicated promoter

The adjacent fadD2 and fadD1 genes could constitute a two gene operon although there is an unusually large intergenic region (319 bp) between the coding sequences of the two genes (Fig. 8A). To test the cotranscription possibility, we performed RT-PCR analyses using primers bracketing the intergenic region plus a portion of the coding sequences (Fig. 8B) and compared the RT-PCR and PCR products. A single band of the expected length was produced in both RT-PCR and PCR analyses demonstrating that the two genes are cotranscribed.

Fig 8.

Transcriptional analysis of the tandem fadD1 and fadD2 genes and detection of the transcription initiation site of fadD1. (A) Diagram of the chromosome segment in which fadD1 and fadD2 are located. (B) RT-PCR detection of the fadD1 mRNA shows that the two FadD genes are cotranscribed and hence constitute an operon. (C) PCR fragments used for obtaining the fadD1 transcription start. (D) Sequence of the transcription start analysis. (E) Location of the fadD1 transcription initiation site relative to the coding sequences of fadD2 and fadD1.

The finding that the FadD2 plays only a minor role in fatty acid metabolism could be due to poor enzymatic activity or to weak expression relative to FadD1. To test the relative expression levels of the two genes, the upstream 500 bp of the fadD1 coding sequence and the upstream 250 bp of the fadD2 coding sequence were fused with the lacZ gene of the expression vector. The translational fusions were transformed into the wild type or ∆fadD1 strains, and β-galactosidase activities were determined. The β-galactosidase activity of the fadD1 fusion in wild-type strain was 10- to 15-fold greater than that of the fadD2 fusion in minimal medium (0.1% CAA) (Fig. S4). This was also tested in the presence of fatty acids of various chain lengths. Fatty acid addition gave only a modest increase in β-galactosidase activity in the FadD1 fusion and had no detectable effect on the FadD2 fusion (Fig. S4). Indeed, the β-galactosidase activity of the fadD1 and fadD2 fusions was significantly decreased in medium containing oleic acid (Fig. 9). The β-galactosidase fusions indicate that fadD1 expression is much higher that fadD2 expression which argued that fadD1 must have its own promoter. This was the case as shown above (Fig. 8).

Fig 9.

Expression levels of the P_fadD1 and P_fadD2 promoters. (A) The β-galactosidase levels of lacZ fusions driven by P_fadD1 or P_fadD2 were detected on LB solid medium containing X-gal in the presence or absence of oleic acid. The significantly darker blue streaks of the P_fadD1 strain than those of the P_fadD2 strain indicate that the expression level of P_fadD1 is significantly higher than P_fadD2 in the wild-type strain, the control strain is P. putida F1 wild type. (B) β-Galactosidase activities of P_fadD1 and P_fadD2 in P. putida wild type without oleic acid; (C) β-Galactosidase activity of P_fadD1 and P_fadD2 in P. putida wild type with oleic acid. The error bars represent the standard deviation from the mean.

DISCUSSION

Although Pseudomonas species commonly have two adjacent acyl-CoA ligase encoding genes, they are differentially controlled. In the case of a given P. putida strain, the metabolic activities may reflect the substrate(s) on which the strain was isolated which provided selection for specific enzyme activities. The P. putida chromosomes show appreciable size diversity (5.73–6.3 Mb) and hence variable gene contents (24). Another complicating factor is that P. putida uses exogenous carbon sources to make a storage material, polyhydroxyalkanoate granules (a “bioplastic”), and this requires β-oxidation enzymes including acyl-CoA ligases (4, 5). Moreover, under carbon-limited conditions, degradation of the granules also requires β-oxidation (4, 5). Indeed, FadD1 of P. putida GPo1 is reported to bind to polyhydroxyalkanoate granules, whereas FadD2 remains unbound (25).

As shown above, although P. putida F1 fadD2 and fadD1 are cotranscribed, fadD1 is the more highly expressed gene due to a promoter located within fadD2 (Fig. 8). Expression of fadD1 is only modestly increased by growth in the presence of fatty acids, whereas fadD2 expression is not increased. Expression of both enzymes is inhibited by oleic acid, a seemingly perverse property (Fig. 9; Fig. S4). In P. putida U, the product of the fadD2 gene is reported to be unable to support growth on fatty acids and is proposed to be a cryptic gene. These workers were unable to construct strains lacking both fadD2 and fadD1. In P. putida, F1 ∆fadD2∆fadD1 strains were readily constructed and FabD2 was shown to be a minor contributor to growth on fatty acids. Double mutants were also constructed in P. putida strain KT2440 (26).

In P. aeruginosa PAO1, the upstream tandem acyl-CoA ligase gene is the more important. Each gene has a strong designated promoter although there is readthrough from the upstream gene that proceeds through the downstream gene (15). Expression of both genes is modestly increased (~ 2-fold) by growth on fatty acids although maximal transcription stimulation differs in that the upstream and downstream genes have different chain length specificities (oleate, upstream and octanoate, downstream). Mutant studies showed that loss of the upstream gene resulted in growth defects on all fatty acids tested, whereas loss of the downstream gene showed more modest defects except for growth on medium chain lengths (C8, C10) (15). As seen in P. putida F1 loss of both the upstream and downstream gene produced a stronger growth defect than the single mutant strains although slower growth on fatty acids remained presumably catalyzed by other acyl-CoA ligases (seven are annotated in strain PAO1). Thus, it seems that in P. aeruginosa, the upstream gene encodes the more important acyl-CoA ligase (15), whereas in P. putida F1, the downstream gene is the more important. The puzzle is why the pseudomonads have retained both genes when one produces a much more effective acyl-CoA ligase than the other. The two P. putida F1 genes and their products are only 73% identical at the nucleotide level and 63% identical at the protein level. Indeed, the proteins show only slightly less identity (55%) to E. coli FadD than to each another. A caveat to these studies is that fatty acid activation may be only a secondary occupation of an acyl-CoA ligase, and the primary target may be another molecule. If this is the case for FadD2 of P. putida F1, the nonfatty acid substrate must be efficiently activated given the low expression level of the fadD2 gene. On the other hand, the finding that the genes are cotranscribed suggests a common substrate. Substrate diversity is seen in the report that the best studied bacterial acyl-CoA ligase E. coli FadD known to initiate β-oxidation has recently been shown to accept acids containing a remarkable variety of aromatic rings as substrates (27).

A striking and unexpected result was that deletion of fadD1 prevented efficient complementation of the ∆fabA∆desA strain by the B. subtilis Des desaturase, whereas Des supported good growth in fadD2 strains. A possible explanation is that the free fatty acids accumulated by the fadD1 strain, but not by fadD2 strain, inhibit some steps of lipid metabolism that must be removed by β-oxidation. A possible target is the Des desaturase. Fatty acid inhibition of mammalian acyl-CoA desaturates has been reported (28, 29), but there are no data on membrane phospholipid desaturates such as the B. subtilis Des desaturase. Indeed, we can find no published enzymology on membrane bound phospholipid desaturates: all investigations are at the intact cell level. These include membrane topology models and mutagenesis studies but no enzyme purifications or attempts to assay desaturase activity in isolated cell membranes. Moreover, the B. subtilis Des electron donor is unclear (30). This situation can be attributed to the formidable difficulties in dealing with membrane proteins compounded by the membrane bilayer substrates of these desaturases. Thus, there is no straightforward system to test the possible effects of fatty acids on Des activity.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions

The strains and plasmids used are given in Table S4. E. coli and P. putida F1 strains were grown at 37°C and 30°C in Luria-Bertani (LB) medium containing (in g/L) (tryptone, 10; yeast extract, 5; NaCl, 10; pH 7.0). When required, antibiotics and inducers were added as follows (in μg/mL): sodium ampicillin, 100; kanamycin sulfate, 30; gentamicin, 30; tetracycline hydrochloride, 60; l-arabinose, 200; isopropyl-β-d-thiogalactoside (IPTG), 240 and 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-Gal), 100. Oleate was used at a final concentration of 5 mM. Bacterial growth was determined by growing on solid and liquid medium. The minimal medium was M9 medium which has following composition: 12.8 g/L Na₂HPO₄ •7H₂O, 3 g/L KH₂PO₄, 0.5 g/L NaCl, 1 g/L NH₄Cl, 2 mM MgSO₄, 0.1 mM CaCl₂. The strains were grown in baffled flasks with orbital shaking at 250 rpm or in an orbital shaking microplate reader. The New England Biolabs (NEB) Gibson Assembly Cloning Kit was used according to the NEB instructions.

Construction of P. putida F1 fadD1 and fadD2 deletion strains

Using the P. putida F1 genome DNA as template, PCR amplification was performed to obtain about 500 bp upstream and downstream fragments of the fadD1 or fadD2 genes, and the resulting DNA fragments were fused by overlap extension PCR. The primers used are given in Table S5. For fadD1 upstream and downstream, the primers fadD1 BamH1 up1 and fadD1 down1 and fadD1 up2 and fadD1 HindIII dn2 were used. For fadD2 upstream and downstream, the primers fadD2 BamH1 up1 and fadD2 down1 and fadD2 up2 and fadD2 HindIII dn2 were used, respectively. The DNA fragments were inserted into the Km-resistant suicide vector pK18mobsacB to produce suicide plasmids using BamHI and HindIII sites. After DNA sequencing, the recombinant plasmid was transformed into the E. coli S17-1 strain to give the donor strains for conjugation. The donor bacteria S17-1 (37°C) and the receptor P. putida F1 (30°C) were separately cultured in 5 mL liquid LB medium, respectively. The cells were collected by centrifugation. The bacterial pellet was washed in fresh LB medium three times and suspended each cell with 0.1 mL LB medium, and the bacterial suspension was spread on LB plates and incubated at 30°C for 24–48 h. The zygotes were scrapped from the plates washed with LB liquid medium suspended, and dilutions were plated on LB plates containing Km (30 µg/mL) and Amp (100 µg/mL) at 30°C for 24–48 h. A single recombinant was selected and cultured in LB liquid medium for 24–48 h. The culture was appropriately diluted and applied to LB plates containing 10% sucrose. Resultant single colonies were screened for the mutations by PCR amplifications and the PCR products sequenced.

Construction of the fadD1 triple mutant strain used the above method with the exception that the kanamycin (Km) resistance gene was ligated between the upstream and downstream the fadD1 fragments to improve screening efficiency. The primers are given in Table S5. The pKD4 plasmid (31) was used as a template to obtain the Km resistance gene fragment plus the FRT loci by primers pKD4 up(Gibson) with pKD4 dn(Gibson), the plasmid fragment was amplified by PCR with primer pK19 up(Gibson) and pK19 dn(Gibson) using the pK19mobsacB plasmid carrying Gm resistance as a template, and PCR was performed to amplify the upstream and downstream of the fadD1 gene about 500 bp by primers fadD1 up1(Gibson) with fadD1 dn1(Gibson) and fadD1 up2(Gibson) with fadD1 dn2(Gibson), respectively. Fusion of multiple fragments was performed using the Gibson Assembly Kit (NEB) to obtain the fadD1 knockout vector carrying Km resistance. The ∆fadD1 mutant strain was obtained by screening on LB plates containing sucrose and kanamycin and verified by PCR analysis and sequencing of the PCR product.

Complementation strains were constructed as followed. For fadD1 and fadD2, the targeted gene fragments were amplified by PCR via the primer fadD1 NdeI with fadD1 HindIII dn2 and fadD2 NdeI with fadD2 HindIII dn2, respectively, and then digested with NdeI and HindII, and the products ligated into the pBAD24M and pSRK vector digested with the same enzymes. For Bacillus subtilis des, the target gene fragment was amplified by overlap PCR for fusing Shewanella oneidensis des gene promoter through primer P_{S. oneidensis des} NdeI up with P_{S. oneidensis des} dn and B. subtilis des (overlap) up with B. sutibilis des BamHI dn and then digested with NdeI and BamHI, and the products ligated into the pSRK vector digested with the same enzymes. The complementation plasmids were transformed into the host strains by conjugation or by electroporation.

Growth analysis of fadD mutants

The test strain was inoculated in 5 mL of LB liquid medium and cultured overnight at 30°C. Transfer to a fresh medium to be tested at 1% inoculum and shakn in culture at 30°C. The OD₆₀₀ value of the bacterial liquid was measured periodically. The concentration of glucose in M9 medium was 0.2%, and the concentration of hydrolyzed casein was 0.1%. The final concentration of fatty acids of various chain lengths was 1 mM, and 0.5% Brij58 was added to solubilize the fatty acids. Three replicates were analyzed for each sample.

Thin-layer chromatography analysis of phospholipid acyl chains and free fatty acids in culture media

The fadD1 and fadD2 derivatives of the P. putida ∆fabA∆desA strain were grown to OD₆₀₀ 0.5 with 0.5 mM oleic acid and incubated for another 3 h at 30°C in the presence of [1-¹⁴C]oleic acid (final concentration of 0.1 µCi/mL). For B. subtilis des gene complemented fadD1 and fadD2 derivatives of the P. putida ∆fabA∆desA strain were grown to OD₆₀₀ 0.1 in LB medium and incubated for another 36 h at 30°C in the presence of [1-¹⁴C]acetate (final concentration of 1 µCi/mL). The complemented strains grow slowly and often take 3 days to show appreciable growth. The cells were extracted with methanol-chloroform (2:1), and the extracted phospholipids were extracted into chloroform and dried under nitrogen. The fatty acyl groups on phospholipids were then converted to their methyl esters by transesterification with 25% sodium methoxide, extracted into petroleum ether, taken to dryness under nitrogen. The methyl esters were resuspended in hexanes and loaded onto 20% silver nitrate TLC plates (Analtech) which were developed in toluene at −20°C. The inclusion of silver allows the separation of saturated and unsaturated esters, and low-temperature development allows the separation of double bond isomers. The plates containing the [¹⁴C]-labeled esters were analyzed by phosphorimaging using a GE Typhoon FLA700 Scanner and analyzed using the Image Quant TL program.

For the free fatty acid analyses, the P. putida ∆fadD1, ∆fadD2, ∆fadD1∆fadD2, wild type, and complemented strains were cultured in LB medium with a radioactive acid as follows. Five microliters of each strain was incubated at OD₆₀₀ 0.5 and then incubated for another 60 h for testing stationary phase cell membrane phospholipid recycling at 30°C in the presence of [1-¹⁴C]acetate (final concentration of 1 µCi/mL). After labeling, the cells were pelleted at 4,000 g for 10 min, and each 3 mL of supernatant was decanted into new glass tube. Glacial acetic acid (0.15 mL) was added to break down any fatty acid salts followed by 9 mL of methanol-chloroform (2:1) in supernatant. The free fatty acids were extracted into chloroform and dried under nitrogen. To each tube, 0.2 mL of ethyl ether/methanol (9/1) was added sufficient trimethylsilyldiazomethane (2 M in hexanes, Aldrich) to give a stable yellow solution. The tubes were capped and allowed to stand for at least 1 h. The excess trimethylsilyldiazomethane was then consumed by the addition of 5 µL aliquots of glacial acetic acid until the solution was colorless, and the resulting mixtures were carefully taken to dryness under nitrogen at 20°C, resuspended in hexanes, and loaded onto silver nitrate TLC plates (Analtech) as above. The plates containing the [¹⁴C] labeled esters were analyzed by phosphorimaging using a GE Typhoon FLA700 Scanner.

GC-MS analysis of phospholipid acyl chains

The strains were cultured in LB medium to OD₆₀₀ 0.5. If the medium contained oleic acid, the cells were washed three times with 0.5% Brij58 to remove oleate. The phospholipids were extracted as described above. The phospholipid fatty acyl group was converted to their methyl esters by transesterification with sodium methoxide, extracted into petroleum ether, taken to dryness under nitrogen, and analyzed by GC-MS using a highly polar chiral CP-Si88 column (Agilent Technologies). The CP-Si88 column allows baseline separation of the methyl esters of acids based on their double bond positions. The GC-MS analyses were done in the University of Illinois Carver Metabolomic Center using the parameters previously described (32).

Extraction of total RNA, cDNA synthesis, and amplification

Total RNA was purified using the RNAeasy Mini Kit (Qiagen). RNA was non-specifically converted to single-stranded cDNAs using the ProroScript First Stand cDNA Synthesis Kit (NEB Biolabs). The control samples were made during cDNA synthesis from the total RNA using the ProroScript First Stand cDNA Synthesis Kit without the addition of reverse transcriptase. All The resulting cDNA served as the template for PCR amplification of the fadD1 and fadD2 gene cluster, using specific primers fadD1 RT and fadD2 RT (Table S5) and an Eppendorf thermal cycler.

RLM-race

The 5′ends of fadD1 mRNA in P. putida F1 were mapped using RLM-RACE according to manufacturer instructions. To identify the 5′ends of the fadD1 mRNA, the PCR products by primer fadD1 inner (Table S5) were cloned into vector pCR 2.1 and sequenced using the M13 primer

β-Galactosidase assay

To construct the lacZ translational fusions in P. putida F1. 5′ fragments of fadD1 (PfadD1) and fadD2 (PfadD2) were fused in the correct translational reading frame to a large 3′ fragment of lacZ gene in vector pSRK Tc. The primers used are given in Table S5. The primer P_fadD1-lacZ up (Gibson) with P_fadD1-lacZ dn (Gibson), and P_fadD2-lacZ up (Gibson) with P_fadD2-lacZ dn to obtain fadD1 and fadD2, respectively, and the plasmid fragment was amplified by PCR with primer pSRK-lacZ up(Gibson) and pSRK-lacZ dn(Gibson) using the pSRK-lacZ plasmid (33) as a template. The fusion of two fragments was performed using the Gibson Assembly Kit to obtain the fadD1 and fadD2 fusion vectors. The fusion vectors were transferred to the wild-type strain of P. putida F1, and β-galactosidase activity of the strains were detected on X-gal containing plates. β-Galactosidase assays were also performed as described by Miller. Mid-log-phase cultures were collected by centrifugation, washed twice with Z buffer, and assayed for β-galactosidase activity after lysis with sodium dodecyl sulfate-chloroform. The data were obtained in triplicate in more than three independent experiments.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01267-24.

Supplemental material. aem.01267-24-s0001.docx.

Tables S1 to S5; Fig. S1 to S4.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.