Divergent unsaturated fatty acid synthesis in two highly related model pseudomonads

The genomes of the best‐studied pseudomonads, Pseudomonas aeruginosa and Pseudomonas putida, which share 85% of the predicted coding regions, contain a fabA fabB operon (demonstrated in P. aeruginosa, putative in P. putida). The enzymes encoded by the fabA and fabB genes catalyze the introduction of a double bond into a 10‐carbon precursor which is elongated to the 16:1Δ9 and 18:1Δ11 unsaturated fatty acyl chains required for functional membrane phospholipids. A detailed analysis of transcription of the P. putida fabA fabB gene cluster showed that fabA and fabB constitute an operon and disclosed an unexpected and essential fabB promoter located within the fabA coding sequence. Inactivation of the fabA fabB operon fails to halt the growth of P. aeruginosa PAO1 but blocks growth of P. putida F1 unless an exogenous unsaturated fatty acid is provided. We report that the asymmetry between these two species is due to the P. aeruginosa PAO1 desA gene which encodes a fatty acid desaturase that introduces double bonds into the 16‐carbon acyl chains of membrane phospholipids. Although P. putida F1 encodes a putative DesA homolog that is 84% identical to the P. aeruginosa PAO1, the protein fails to provide sufficient unsaturated fatty acid synthesis for growth when the FabA FabB pathway is inactivated. We report that the P. putida F1 DesA homolog can functionally replace the P. aeruginosa DesA. Hence, the defect in P. putida F1 desaturation is not due to a defective P. putida F1 DesA protein but probably to a weakly active component of the electron transfer process.

Fatty acids are required for the construction of the inner and outer membrane phospholipids of gram-negative bacteria (Cronan & Gelmann, 1975;Parsons & Rock, 2013). Bacteria, plants, and some parasites use the type II fatty acid synthesis system (FASII).
Compared with the FASI of eukaryotes, the FASII pathway utilizes separate enzymes that catalyze the individual steps rather than large multifunctional proteins (Yao & Rock, 2017). In many bacteria, FabA and FabB are responsible for the production of unsaturated fatty acids (UFAs) under both aerobic and anaerobic conditions.
We have now explored this asymmetry between these two highly related pseudomonads.

| Properties of the P. putida F1 DesA and FabA activities in vivo
In common with P. aeruginosa PAO1, P. putida F1 encodes homologs of DesA and FabA (Figure 1). The DesA and FabA proteins of the two species have 84% and 89% identity, respectively. However, P. putida F1 lacks the P. aeruginosa PAO1 desB and the misnamed desC gene (which encodes a putative electron transfer protein). The ∆fabA ∆desA double mutant of P. aeruginosa PAO1 is auxotrophic for UFA (Zhu et al., 2006). We previously studied P. putida ∆fabB strains and found these strains to be UFA auxotrophs under aerobic conditions (Dong et al., 2021). This was puzzling because P. putida encodes a protein having 84% identity to the P. aeruginosa DesA protein that in P. aeruginosa bypasses loss of the fabA gene (Zhu et al., 2006). Indeed, deletion of the P. putida desA gene had little or no effect on UFA synthesis raising the possibility that P. putida desA may be a pseudogene (Dong et al., 2021). To test this possibility, we introduced the P. putida desA into a P. aeruginosa PAO1 ∆fabA ∆desA strain and found that it fully restored growth and UFA synthesis to the P. aeruginosa PAO1 strain ( Figure 2a). Hence, the P. putida DesA is a functional desaturase that acts on phospholipid C16:0 saturated acyl chains (Zhu et al., 2006). We also tested the expression of the Des desaturase of B. subtilis and observed complementation although with very slow growth perhaps due to the atypical UFA chain synthesized (Figure 2a).
We tested the abilities of plasmids encoding E. coli fabA and P.
putida F1 fabA to complement the P. aeruginosa PAO1 ∆fabA ∆desA strain. Growth of the E. coli fabA complemented strain was significantly better than that of P. putida F1 fabA complemented strain both in the presence and the absence of IPTG induction (Figure 2a), suggesting that the E. coli fabA was more active than that of P. putida  (Table 1). These data demonstrate that like the P. aeruginosa PAO1 DesA, the P. putida F1 DesA is a Δ9 desaturase that desaturates phospholipid C16:0 saturated acyl chains to C16∆9 chains. Note that previous workers expressed a P.

| FabA is essential for UFA synthesis in P. putida F1
To study the roles of the fabA and desA genes in P. putida F1 UFA synthesis, deletion strains were constructed individually and in combination. The entire coding sequences of the ∆desA and ∆desA ∆fabA strains were deleted, whereas in the ∆fabA:Kan strain, an internal segment of the gene was replaced with a gene encoding kanamycin resistance ( Figure 3a). The P. putida F1 ∆desA ∆fabA and both ∆fabA strains were UFA auxotrophs and grew well in oleate-supplemented media ( Figure 3b). Deletion of the P. putida F1 desA had no effect on fatty acid synthesis or growth ( Figure 3b). Therefore, fabA is essential for UFA synthesis and desA is (at best) a weakly supplemental pathway.

| Analysis of the P. putida F1 fabA and fabB genes
The P. putida F1 fabA and fabB are adjacent genes with only 11 bp between the coding sequences. As mentioned above, we constructed two different ∆fabA strains. One construct removed the entire fabA coding sequence, whereas the other, called ∆fabA:Kan contained an internal deletion marked with a kanamycin cassette (Figure 3a). To our surprise, complementation of the P. putida F1 ∆desA ∆fabA strain from which the entire fabA coding sequence had been removed failed with plasmids encoding either E. coli fabA or P. putida F1 fabA (Figure 4). In contrast, when fabA and fabB were coexpressed, the strain grew well (see below). The ∆fabA:Kan internal deletion strain was readily complemented by the expression of either E. coli fabA or P. putida F1 fabA (data not shown). These results argued that the total deletion of the entire fabA coding sequence had removed a sequence required for fabB expression that remained in the internal deletion strain. These data led us to study the transcription of these genes.
Reverse transcription-PCR showed that the fabA and fabB are cotranscribed and thus constitute an operon as previously reported in P. aeruginosa PAO1 (Hoang & Schweizer, 1997) (Figure 5a). The most straightforward explanation of the differences between the two ∆fabA strains constructed is that a fabB promoter located within the fabA coding sequence was retained in the construction of the internal deletion in the ∆fabA:Kan strain (the 3′-end of which is 375 bp from the FabB ATG initiation codon). This hypothesis was tested by 5′-RLM-RACE analysis and showed a transcription initiation site 70 bp upstream of the fabB coding sequence and well within the fabA coding sequence (Figure 5b,c). Note that the fabAB promoter was mapped previously and is 120 bp upstream of the fabA coding sequence (Dong et al., 2021).
We measured the relative strengths of the fabAB and fabB promoters by using the promoters to drive the expression of a promoterless E. coli lacZ gene ( Figure 6). β-Galactosidase assays showed that the fabB promoter strength was about 14% that of the fabAB promoter.
To preclude any influence of fabF2 that could complicate the study of the role of the fabB promoter, a ∆fabA ∆fabF2 strain was constructed ( Figure 3a). FabF2 is expressed from a cryptic P. putida F1 gene (fabF2) which encodes a protein that has FabB activity and weak FabF activity (Dong et al., 2021). Expression of fabF2 is cryptic due to transcriptional repression by a protein encoded in the same The black boxes are the three histidine-rich regions, and the black stars indicate the conserved histidine residues that are catalytically essential for desaturase activity (Shanklin et al., 1994). gene cluster. When the P. putida ∆fabB strain was plated on fatty acid-free media suppressor colonies appeared are the result of mutational inactivation of the repressor protein (∆fabB ∆fabF2) strains did not accumulate suppressors (Dong et al., 2021).
There have been proposals that E. coli FabA and FabB might form a complex perhaps aided by a third protein but no definitive data have appeared, although there is a candidate protein (Sugai et al., 2001). To test this possibility, we constructed plasmids that mixed E. coli and P. putida genes, a FabA plus either FabF2 or a FabB. We also constructed plasmids that expressed fabF2 together with E. coli fabA or P. putida F1 fabA and used these to test the complementation of the P. putida F1 ∆desA ∆fabA (which is also FabB deficient due to loss of the promoter within fabA). As reported below, all combinations were functional arguing strongly against the complex formation.
Expression of E. coli fabA rescued the growth of the ∆fabA ∆fabF2 strain, whereas P. putida F1 fabA alone could not (Figure 4). This was confirmed by labeling of cultures with [1-14 C]acetate which showed that the E. coli fabA restored UFA synthesis to the ∆fabA ∆fabF2 strain (data not shown). As discussed above, the activity of E. coli FabA was significantly higher than that of P. putida F1 FabA in complementation of the P. aeruginosa PAO1 ∆fabA ∆desA strain.
Therefore, when P. putida F1 fabB is expressed from its own promoter, the high activity of E. coli FabA can fully restore UFA synthesis. This result suggests that in the P. putida F1 wild-type strain, the fabB promoter is needed to increase FabB levels to compensate for the low activity of its cognate FabA.
The lac promoter of the expression vector pSRK (Khan et al., 2008) was used to test the ability of the expression of P. putida fabA plus fabB to complement the P. putida F1 ∆fabA ∆desA F I G U R E 2 Complementation of the P. aeruginosa ∆fabA ∆desA strain by various desA genes. (a) Comparison the growth of the P. aeruginosa (Pa), ∆fabA ∆desA strain complemented with P. aeruginosa desA, P. putida F1 (Pp) desA, or B. subtilis (Bs) des. All three desaturases restored P. aeruginosa ∆fabA ∆desA growth in the absence of a fatty acid supplement. The growth phenotypes of the strains complemented with the two Pseudomonas desA genes were essentially the same, whereas the growth of strain complemented with B. subtilis des was much weaker than the desA complemented strains. (b) The [1-14 C]acetate labeled phospholipid fatty acids of the complemented derivatives of the P. aeruginosa PAO1 ∆fabA ∆desA strain. The P. aeruginosa ∆fabA ∆desA strain lacked UFA synthesis, whereas P. aeruginosa desA, P. putida desA, and B. subtilis des restored UFA synthesis to the P. aeruginosa ∆fabA ∆desA strain. The desA genes synthesized a C16∆9 UFA, whereas the B. subtilis des synthesized a C16∆5 UFA. Vec denotes the empty vector. Note: Pa △ AA denotes the P. aeruginosa ∆fabA ∆desA strain. The Ole (oleic acid) concentration was 1 mM. WT denotes wild type. The entry on the right side of the shill denotes a plasmid carried gene. strain. The P. putida fabA fabB plasmid restored robust growth with or without IPTG induction (Figure 7). When the P. putida  (Table 2).

| DISCUSS ION
Phospholipid

F I G U R E 6
Determination of the relative expression levels of P fabA and P fabB . (a) The expression levels of -galactosidase of lacZ fusions driven by P fabA or P fabB were detected on the LB solid medium containing X-gal, the blue streaks of the P fabA strain were significantly darker than those of the P fabB strain, indicating that the expression level of P fabA is significantly higher than P fabB in the wild-type strain, the control strain is P. putida F1 wild-type; (b) -Galactosidase activities of P fabA and P fabB in P. putida wild type.
synthesize UFAs in bacteria: an oxygen-dependent fatty acid desaturation pathway and the FabAB pathway. P. putida F1 differs from P. aeruginosa PAO1 in that its aerobic desaturation pathway fails to compensate for the loss of the FabAB pathway even when DesA is overexpressed (Dong et al., 2021). This argues that the lack of compensation is not the fault of the P. putida F1 DesA protein but of a component of the electron transfer process that is only weakly active. This electron transfer process is functional in P. aeruginosa F I G U R E 7 Complementation of the P. putida F1 ∆fabA ∆desA strain. Growth of derivatives of the P. putida F1 ∆fabA ∆desA strain with plasmids expressing either P. putida F1 fabA fabB (PpfabAfabB) or E. coli fabA with P. putida F1 fabB (EcfabAfabB).  Table 2.

F I G U R E 8
PAO1 and thus P. putida F1 DesA functionally replaced P. aeruginosa PAO1 DesA.
The fabA fabB gene cluster is ubiquitous in Pseudomonas, and it has been shown that the cluster is an operon in P. aeruginosa PAO1 and now in P. putida F1 (Figure 5a). However, P. putida F1 fabB has its own weak promoter located within fabA (Figure 5b). This is not the case in P. aeruginosa PAO1, where only a fabA fabB transcript was reported (Hoang & Schweizer, 1997). Based on Figure  FabA. This low activity seems an intrinsic property of the P. putida F1 FabA because expression of the gene from a lac promoter on a multicopy plasmid also gave only weak complementation. Hence, although

| Bacterial strains, plasmids, and growth conditions
The strains and plasmids used are given in Table S1. E. coli and P.

| Construction of P. putida F1 ∆desA, fabA, and ∆ fabA ∆desA strains
PCR amplification (primers are given in Table S2) was performed to obtain DNA fragments of about 500 bp upstream and downstream of either desA or fabA. The DNA fragments were then fused by overlap extension PCR. These DNA fragments were cloned into the suicide vector pK18mobsacB carrying resistance to Km to form a suicide plasmid. After DNA sequencing, the recombinant plasmid was transformed into the E. coli S17-1 strain to obtain a donor strain for conjugation. The donor bacteria S17-1 (37°C) and the receptor P. putida F1 (30°C) were separately cultured in a 5 ml liquid LB medium, respectively, and the cells were collected by centrifugation. The bacteria were washed three times with the fresh LB medium, then each cell pellet was suspended in 0.1 ml LB medium, the donor and recipient bacteria were mixed and plated on an LB plate, and cultured at 30°C for 24-48 h. The bacteria were then washed from the plate and the liquid was collected, diluted appropriately, spread on LB plates with kanamycin (30 μg/ ml) and ampicillin (100 μg/ml), and incubated at 30°C for 24-48 h.
A single recombinant was selected and cultured in the LB liquid medium lacking antibiotics for 24-48 h. The culture was appropriately diluted and applied to LB plates containing 10% sucrose. The single colonies that grew on the LB sucrose solid medium were screened by PCR (with the wild-type strain was as a control) for the expected gene mutations and the PCR products were sequenced to give the ∆desA single and ∆fabA ∆desA double-mutant strains.
Construction of the fabA internal deletion strain used the above method except that the kanamycin resistance gene was ligated be-  (Dong et al., 2021) as the template.

| Construction of complementation strains
The target gene fragment was amplified by PCR or overlapping PCR, and then digested with restriction endonuclease, and ligated into vector pSRK digested with the same enzyme and screened and verified.
The NEB Gibson Assembly Kit was also used to fuse multiple fragments. The complementation plasmids were transferred to the host cell by conjugation or electroporation. The method of conjugation is as above. Electroporation was performed as follows: 5 ml of host strain was cultured overnight, the cells of 1 ml of host strain were recovered and washed twice with 300 mM sucrose, concentrated to 100 μl, the complementation plasmid was added and then transfer to an electrophoresis cuvette and put on ice 10 min. Following electrophoresis at 2.5 kV, 1 ml of fresh medium was added. After growth for 1 h at 30°C, the cells were spread on a plate containing the appropriate antibiotics, and the resulting colonies screened after incubation.

| Thin layer chromatography (TLC) analysis of phospholipid acyl chains
The complemented or mutant strains were cultured in the LB medium with or without oleate and labeled with a radioactive acetate as fol-

| GC-MS analysis of phospholipid acyl chains
The strains were cultured in the LB medium at OD 600 0.5. Cultures were standardized and fatty acid methyl esters were generated as above and then analyzed by GC-MS using a highly polar chiral CP-

| RLM-RACE
The 5′ends of fabB mRNA in P. putida F1 were mapped using RLM-RACE according to the manufacturer's instructions. To identify the 5′ends of the fabB mRNA, the PCR products were cloned into vector PCR 2.1 and sequenced.

| ß-Galactosidase assays
To construct the lacZ translational fusions in P. putida F1. The 5′ fragments promoters of fabA(P fabA ) and fabB(P fabB ) were fused in the correct translational reading frame to a large 3′ fragment of lacZ gene in vector pSRK Tc. The P fabA and P fabB fusion vectors were transferred to the wild-type strain of P. putida F1, and the blue strain was detected on the X-gal-containing plate.
lacZ encodes β-galactosidase, and X-gal, when cleaved by βgalactosidase, produces galactose and 5-bromo-4-chloro-3-hy droxyindole. The latter then spontaneously dimerizes and oxidizes to 5,5′-dibromo-4,4′-dichloroindole blue, an insoluble dark blue product. ß-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 ßgalactosidase activity after lysis with sodium dodecyl sulfate chloroform. The data were obtained in triplicate in more than three independent experiments.

AUTH O R CO NTR I B UTI O N S
Huijuan Dong: Acquisition, analysis, or interpretation of the data; conception or design of the study; Writing of the manuscript.
Haihong Wang: Conception or design of the study. John E. Cronan: Acquisition, analysis, or interpretation of the data; conception or design of the study; Writing of the manuscript.

ACK N OWLED G M ENTS
We thank Dr. Charles O. Rock for strain PA068, Dr. Alexander Ulanov for gas chromatography-mass spectrometry analyses and Dr. Tatiana Kondakova for early contributions to this project.

FU N D I N G I N FO R M ATI O N
This work was supported by the National Institutes of Health (grant AI15650) from the National Institute of Allergy and Infectious Diseases.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

E TH I C A L S TATEM ENT
This article does not contain any studies with human participants or animals performed by any of the authors.