|Home | About | Journals | Submit | Contact Us | Français|
The fabA and fabB genes are responsible for anaerobic unsaturated fatty acid formation in Pseudomonas aeruginosa. Expression of the fabAB operon was repressed by exogenous unsaturated fatty acids, and DNA sequences upstream of the translational start site were used to affinity purify DesT. The single protein interaction with the fabAB promoter detected in wild-type cell extracts was absent in the desT deletion strain, as was the repression of fabAB expression by unsaturated fatty acids. Thus, DesT senses the overall composition of the acyl-coenzyme A pool to coordinate the expression of the operons for the anaerobic (fabAB) and aerobic (desCB) pathways for unsaturated fatty acid synthesis.
Bacteria regulate their membrane fatty acid composition to maintain an optimal fluidity to support growth under different environmental conditions (for a review, see reference 19). This critical role for fatty acid structure is consistent with the multiple mechanisms that have evolved in bacteria to control the composition of fatty acids incorporated into membrane phospholipids. In Escherichia coli, unsaturated fatty acid (UFA) synthesis occurs exclusively by the anaerobic type II biosynthetic pathway and requires the activities of the fabA and fabB genes (4). FabA is a β-hydroxydecanoyl-acyl carrier protein (ACP) dehydratase isomerase that forms cis-3-decenoyl-ACP, which is then elongated by the FabB condensing enzyme (4). The proportion of UFA produced is governed by the levels of FabA and FabB expression. Both genes are transcriptionally regulated by the FadR activator and the FabR repressor (2, 7, 8, 18). In the absence of ligand, FadR binds to the fabA and fabB promoters to activate transcription, but the binding of long-chain acyl-coenzyme A (acyl-CoA) to FadR promotes a conformational change that releases the protein from DNA (1, 15-17). Both saturated fatty acid (SFA)- and UFA-CoA function as FadR ligands; thus, in the presence of any exogenous fatty acid, fabA and, to a lesser extent, fabB expression is diminished due to release of the activator. The ligand that controls FabR repressor binding to the fabA and fabB promoters is unknown, but the deletion of this regulator increases expression of fabB and, to a lesser extent, fabA, resulting in an increase in the proportion of UFA in the membrane phospholipids (18).
Like E. coli, Pseudomonas aeruginosa uses fabA and fabB for the anaerobic formation of UFA via the type II pathway that are coordinately expressed as a fabAB operon (9). P. aeruginosa also has two aerobic pathways for UFA synthesis. One pathway is catalyzed by DesA, a membrane-associated Δ9 desaturase that introduces a double bond into fatty acids at the 2-position of existing membrane phospholipids (20). The second aerobic pathway is catalyzed by the DesCB proteins that together act as a Δ9 desaturase that introduces a double bond into SFA-CoA (20). DesT is a repressor that regulates the desCB operon based on an increase in desCB mRNA in a P. aeruginosa strain with a deleted desT gene (19). DesT binding to a DNA palindrome in the desCB promoter is both positively and negatively regulated by acyl-CoA. Although DesT binds SFA- and UFA-CoA with the same affinity, DesT binding to the desCB promoter is enhanced by the formation of the DesT complex with UFA-CoAs, whereas DesT is released from DNA when it binds SFA-CoA. This elegant regulatory property monitors the overall fatty acid composition of the acyl-CoA pool, allowing DesT to match desCB expression to the fatty acid structures that are present in the environment. The addition of UFA to P. aeruginosa results in a 2-fold repression of fabAB expression (3). This similarity to FadR regulation in E. coli led to a screen of knockout strains having deletions in the FadR-like transcription factors, but fabAB repression by exogenous UFA was not affected (3). This work identifies DesT as a UFA-dependent repressor of fabAB expression in response to exogenous UFA that functions to coordinate the expression of the anaerobic and aerobic pathways for UFA synthesis.
Materials used in this study were [γ-32P]ATP (specific activity, 6,000 Ci/mmol) (American Radiochemical Company), poly(dI-dC) (Roche), CNBr-activated Sepharose 4B beads (GE Healthcare), herring sperm DNA (Invitrogen), SYPRO Ruby protein gel stain (Bio-Rad), 6% DNA retardation gel (Invitrogen), 5× Novex TBE hi-density sample buffer and 5× Novex running buffer (Invitrogen), T4 polynucleotide kinase (NEB) and acyl-CoAs (Avanti). Acyl-ACPs were prepared by the acyl-ACP synthetase method (13). The P. aeruginosa strains PAO1 and PAO482 (ΔdesT) were as described previously (20).
The transcriptional start site for the fabA gene was identified using the 5′-RACE kit (Invitrogen) as described by Zhang et al. (19). The cDNA was generated from strain PAO1 mRNA using a gene-specific reverse primer (fabA-1r; 5′-CTTCGCCCGAACCTAGGGCGCGGCCGC), amplified using the dC-tailed cDNA using a nested fabA-specific primer (fabA-2r; 5′-CAGCCGAGGTAGAAACCGACCAGCTGCC) and the anchor primer, and amplified a final time using a fabA-specific primer (fabA-3r;5′-CGCCTTCGAAGTGACAGGCG) and the AUAP primer. The PCR product was gel purified and sequenced using the fabA-3r primer.
The upstream transcriptional start site was identified using the First Choice RLM-RACE kit (Ambion). Full-length mRNAs from strain PAO1 were selected by tobacco acid pyrophosphatase treatment, and an adapter was ligated to the 5′ end. The cDNA was generated using a random decamer primer, and the cDNA was amplified using gene-specific primers: fabA outer, 5′-TCGCGCCGGGTTATTGTATCG, and the 5′ RACE outer primer and also fabA inner (5′-AAGTCACAGTGTCCGGCGAACAG), and the 5′ RACE inner primer. The PCR fragment was gel purified and sequenced using the fabA inner primer.
Northern blotting was performed with 10 μg of total RNA separated by electrophoresis on a 0.8% agarose formaldehyde gel. The probe was generated by NcoI-HindIII digestion of pCS22 (pET28b-PAfabA). Northern transfer, hybridization, and washing were performed using standard procedures. Detection and quantification were done using a Phosphorimager (Molecular Dynamics).
The sequence of the fabAB oligonucleotide probe (OP) was 5′-GCGGGAATAAAGTGAACATCTGTTCGCCGGACACTGTG, that of the desCB oligonucleotide was 5′-GATACATCAGTGAACGCTTGTTGACTCGATTGCG, and that of the scrambled oligonucleotide was 5′-GACGCATTGGAACGAAAACGAATGCGCAACGGCAACAG. The complementary sequences were annealed by incubating the two single-stranded DNAs at 94°C for 2 min, followed by gradually decreasing the temperature to 14°C over the course of 4 h. The 366-bp fabAB promoter long probe (LP) was synthesized from strain PAO1 genomic DNA using primers 5′-CCTGACCGCCAACGCCCTGC and 5′-GCGAGCTCCTCAAAAATCCCTG. The double-stranded OP and LP were labeled with [32P]dATP using T4 polynucleotide kinase. The gel mobility shift assays and the purification of DesT with a carboxy-terminal His tag were performed as described by Zhang et al. (19).
The double-stranded fabAB OP was coupled to Sepharose 4B beads as described by Kerrigan and Kadonaga (11). Phosphorylated, double-stranded OP (30 nmol) was coupled to 1.5 g of CNBr-activated Sepharose 4B beads (GE Healthcare) and stored at 4°C until further use. Crude extracts were prepared by growing strain PAO1 to mid-log phase in Luria broth. Cells were harvested and resuspended in binding buffer: 20 mM Tris, 0.1 M KCl, 1 mM dithiothreitol, 10% glycerol, 0.1% (vol/vol) NP-40. Membranes and cell debris were removed by centrifugation at 200,000 × g for 1 h. Proteins were precipitated using 60% saturated ammonium sulfate, resuspended in 5 ml of binding buffer, and fractionated using a Sepharose S-75 size exclusion column. Column fractions that were positive for fabAB promoter binding based on gel mobility shift assays were pooled, and herring sperm DNA (Invitrogen) was added to minimize nonspecific DNA binding. These extracts were either treated with 100 μM oleoyl-CoA or left untreated and were loaded onto a 1.0-ml oligonucleotide affinity column that was washed four times with 2 ml binding buffer. Protein fractions were eluted by increasing the KCl concentration in the binding buffer by 0.1 M up to 1.0 M. The eluted fractions were concentrated, and the salt concentration was reduced to 0.1 M using a cup concentrator (5,000-molecular-weight cutoff). The fractions were then subjected to gel mobility shift assay and sodium dodecyl sulfate (SDS) gel electrophoresis. The gels were stained with SYPRO Ruby protein gel stain (Bio-Rad), and the trypsin-digested protein bands from the 0.4 M KCl elution fraction were identified by mass spectrometry using an electrospray ionization LTQ linear ion trap mass spectrometer (Thermo Electron). Protein/peptide assignments were made on the basis of the tandem mass spectrometry spectra. Comparison of the tryptic peptide masses to peptide masses of proteins from the strain PAO1 database (http://www.pseudomonas.com) was performed using MASCOT (6).
P. aeruginosa strains PAO1 and PAO482 (ΔdesT) were grown in M9 minimal medium with 0.4% (vol/vol) glycerol plus 0.5% (vol/vol) Brij 58 to an optical density at 600 nm (OD600) of 1.2, the culture was split, 0.01% (wt/vol) oleate was added to one culture, and the cells were grown for an additional 30 min. Cells were harvested by centrifugation, and total RNA was isolated using the RNAqueous purification kit (Ambion). The expression levels of fabA were determined using the forward primer fabA-for (5′-CCGGGTAACGCGCAACT), the reverse primer fabA-rev (5′-CCGACATCGCTGATGTGAAC), and the TaqMan probe (5′-CCGCCCCCAACATGCTGAT). Results were compared using the ΔCT method of normalization using the P. aeruginosa housekeeping gene rpoD as the calibrator. Fold changes were calculated using the 2−ΔΔCT method (12). The primers and probes were as described previously (19, 20).
The transcriptional start site for the fabAB mRNA was determined to locate the promoter region of the operon (Fig. (Fig.1).1). A total of four independent 5′-RACE experiments were performed using mRNA isolated from strain PAO1. Multiple experiments yielded the result shown in Fig. Fig.1A.1A. These data placed the start site for this transcript at 92 bases before the translational initiation codon for the fabA gene. A second, longer fabAB transcript was detected in a RACE experiment using a different set of primers (Fig. (Fig.1B).1B). The start site for this transcript was located 255 bases before the translational initiation codon. A DNA palindrome was identified in the +102 to +120 position relative to the transcriptional start site of the longer fabAB transcript and was located prior to the transcriptional start site in the shorter transcript (Fig. (Fig.1C).1C). This DNA palindrome was highly conserved in both sequence and position in the alignment of the sequences upstream of the predicted translational start sites of the fabAB operons in five Pseudomonas species (Fig. (Fig.1C).1C). The conserved location of this palindrome in the fabAB promoter in Pseudomonas species suggested that it was a functional regulatory element. The locations of the −10 and −35 sequences for both promoters are shown in Fig. Fig.1D.1D. The existence of two fabAB transcripts was confirmed by Northern blot analysis (Fig. (Fig.1E).1E). The sizes of the two transcripts corresponded to the predicted sizes based on the RACE results (Fig. (Fig.1A1A and and1B).1B). The longer transcript appeared to be the most abundant (Fig. (Fig.1E1E).
P. aeruginosa extracts were fractioned by oligonucleotide affinity chromatography followed by identification of the bound proteins by gel electrophoresis coupled with mass spectrometry (Fig. (Fig.2).2). The affinity chromatography column was synthesized using OP, a 38-bp oligonucleotide containing the DNA palindrome identified in the fabAB promoter (Fig. (Fig.1C).1C). Extracts were fractionated in both the presence and the absence of oleoyl(18:1)-CoA to simulate the intracellular formation of acyl-CoA in the presence of exogenous fatty acids. Gel shift experiments using 32P-labeled OP located protein fractions that had DNA binding activity. The major protein bound to the column in the absence of 18:1-CoA (Fig. (Fig.2)2) was PA0141, a hypothetical DNA binding protein with 74% similarity to the PvdS transcriptional regulator of Mycobacterium tuberculosis. A comparison of strain PAO1 to a PA0141 deletion mutant showed no alteration of fabAB expression (not shown). Minor amounts of PA1520, PA2667, and PA4315 were also bound to the column. The addition of 18:1-CoA to the extracts prior to chromatography resulted in the appearance of a major protein in the bound fraction that was identified as DesT by mass spectrometry (Fig. (Fig.2).2). These data show that DesT bound to OP in an 18:1-CoA-dependent manner and was the major cellular protein that associated with this palindrome.
Two probes—OP, the conserved palindrome, and LP, consisting of the entire fabAB promoter region (Fig. (Fig.1C)—were1C)—were 32P-labeled and used in gel mobility shift experiments to investigate the number of fabAB promoter binding partners present in P. aeruginosa total cell extracts. There was only a single binding partner for 32P-labeled OP detected in cell extracts from strain PAO1 (Fig. (Fig.3A).3A). The binding of [32P]OP by the cell extract was enhanced by the addition of 18:1-CoA to the extract, and binding to OP was not detected in cell extracts from strain PAO482 (ΔdesT). A second set of experiments was performed using 32P-labeled LP to reveal the number of binding partners for the entire fabAB promoter region in extracts from strain PAO1 (Fig. (Fig.3A).3A). Only a single gel shift was detected using LP with the wild-type cell extracts. This binding was enhanced by the inclusion of 18:1-CoA in the assay, and there was no binding to LP detected in extracts from strain PAO482 (ΔdesT). Therefore, only a single DNA partner for the fabAB promoter was detected in our experiments with total cell extracts, and its absence in extracts from strain PAO482 identified the protein as DesT.
Purified DesT bound to the fabAB promoter based on gel mobility shift experiments using LP (Fig. (Fig.3B).3B). This binding was competitively eliminated by the addition of the nonradioactive OP containing the oligonucleotide palindrome (Fig. (Fig.1C).1C). A scrambled oligonucleotide did not compete for binding. These data show that the oligonucleotide identified in the fabAB operon (Fig. (Fig.1C)1C) was the only DesT binding site in the fabAB promoter.
Previous work extensively detailed the relationship between acyl-CoA binding to DesT and DesT affinity for DNA (19). Both UFA-CoA and SFA-CoA bound to DesT with equal affinity; however, DesT bound to UFA-CoA had a high affinity for the desCB promoter, whereas SFA-CoA binding to DesT prevented DNA binding. We determined whether the acyl-CoA effects on DesT binding to the palindrome in the fabAB promoter were the same using gel mobility shift assays with a DesT concentration that bound 50% of the OP probe (Fig. (Fig.4).4). The presence of UFA-CoA (16:1 and 18:1) increased the binding of DesT to OP, whereas the presence of SFA-CoA (16:0 and 18:0) released DesT from the oligonucleotide. Acyl-ACP did not influence the binding of DesT regardless of the structure of the attached fatty acid. These data indicated that the rules established for the regulated DesT binding to the desCB promoter (19) were the same for DesT binding to the fabAB promoter.
Although the same basic rules for regulated DesT DNA binding were found in the fabAB promoter as in the desCB promoter, the structures of the two palindromes were not the same. The fabAB palindrome differs from the desCB binding site by two base pairs (Fig. (Fig.5).5). This difference suggests that DesT likely binds more tightly to the more “perfect” desCB palindrome than to the site within the fabAB promoter. Consistent with this idea, DesT had a lower affinity for the fabAB oligonucleotide probe than for the desCB oligonucleotide probe (Fig. (Fig.5).5). These data suggested that the regulation of desCB by DesT would be more stringent than the regulation of fabAB based on the lower affinity of DesT for OP.
The regulation of fabAB by exogenous fatty acids was examined by real-time PCR (Fig. (Fig.6).6). Expression of the fabAB operon was repressed 2-fold in strain PAO1. This result is comparable to the 2-fold repression of fabAB expression observed previously using a lacZ reporter system (3). There was no statistical difference between the expression of fabAB in strain PAO1 in the absence of oleate and in strain PAO482 (ΔdesT) in either the presence or the absence of oleate. These in vivo data led to the conclusion that DesT is the only UFA-dependent regulator of fabAB expression. The repression of fabAB was much less than the repression of desCB noted previously (19, 20). This distinct difference between the stringency of regulation between these two operons was reflected in the differences in the affinities of DesT for its DNA binding site in the promoters. Fatty acid β-oxidation is a major route for acyl-CoA metabolism in P. aeruginosa (10, 14), and acyl-CoAs are envisioned as a supplement to rather than a replacement for the de novo fatty acid biosynthetic pathway.
This work expands the role of DesT in the P. aeruginosa physiology and leads to a model for the coordinated regulation of the anaerobic and aerobic pathways for UFA formation by DesT in P. aeruginosa (Fig. (Fig.7).7). Exogenous fatty acids are taken up by P. aeruginosa and converted to their acyl-CoA derivatives via one of the several acyl-CoA synthetases in the organism. DesT binds with essentially equal affinity to the entire acyl-CoA pool (19). The DesT·SFA-CoA complex locks the protein in a conformation that does not permit DNA binding. On the other hand, the DesT·UFA-CoA complex has high affinity for DNA binding sites within the fabAB and desCB promoters, and the binding of DesT to these sites represses transcription from these two operons. The ability of DesT to sense the entire acyl-CoA pool and adjust gene expression based on the UFA:SFA ratio provides an elegant system for balancing the proportion of UFA produced by the two pathways and providing the acyltransferase systems with a stable composition of SFA and UFA thioesters to select from. The finding of two transcripts for fabAB in P. aeruginosa is similar to fabA regulation in E. coli, where there are also two transcripts, one primarily controlled by the FadR activator while the other is regulated by the FabR repressor (5, 7). FadR DNA binding is also regulated by long-chain acyl-CoA, but unlike DesT, any acyl-CoA that binds to FadR induces a conformational change that releases FadR from its DNA binding site. Thus, the cellular levels of FabA and FabB in P aeruginosa are dynamically regulated, but the transcription of these vital genes is never completely shut off by DesT. The role of DesT in controlling fabAB expression likely accounts for the fact that DesT is conserved in both protein sequence and chromosomal location in five Pseudomonas species examined, yet only P. aeruginosa has a desCB operon (Table (Table1).1). The ability of DesT to sense fatty acid structure makes it a key regulator of the proportion of UFA available for membrane phospholipid synthesis by coordinating the expression of the anaerobic and aerobic pathways for the formation of UFA.
We thank the Hartwell Center for mass spectrometry and protein identification.
This work was supported in part by the National Institutes of Health grant GM 34496, National Institutes of Health Cancer Center (CORE) Support Grant CA21765, and the American Lebanese Syrian Associated Charities.
Published ahead of print on 30 October 2009.