Growth of Enterococcus faecalis ∆plsX strains is restored by increased saturated fatty acid synthesis

ABSTRACT The Enterococcus faecalis acyl-acyl carrier protein (ACP) phosphate acyltransferase PlsX plays an important role in phospholipid synthesis and exogenous fatty acid incorporation. Loss of plsX almost completely blocks growth by decreasing de novo phospholipid synthesis, which leads to abnormally long-chain acyl chains in the cell membrane phospholipids. The ∆plsX strain failed to grow without supplementation with an appropriate exogenous fatty acid. Introduction of a ∆fabT mutation into the ∆plsX strain to increase fatty acid synthesis allowed very weak growth. The ∆plsX strain accumulated suppressor mutants. One of these encoded a truncated β-ketoacyl-ACP synthase II (FabO) which restored normal growth and restored de novo phospholipid acyl chain synthesis by increasing saturated acyl-ACP synthesis. Saturated acyl-ACPs are cleaved by a thioesterase to provide free fatty acids for conversion to acyl-phosphates by the FakAB system. The acyl-phosphates are incorporated into position sn1 of the phospholipids by PlsY. We report the tesE gene encodes a thioesterase that can provide free fatty acids. However, we were unable to delete the chromosomal tesE gene to confirm that it is the responsible enzyme. TesE readily cleaves unsaturated acyl-ACPs, whereas saturated acyl-ACPs are cleaved much more slowly. Overexpression of an E. faecalis enoyl-ACP reductase either FabK or FabI which results in high levels of saturated fatty acid synthesis also restored the growth of the ∆plsX strain. The ∆plsX strain grew faster in the presence of palmitic acid than in the presence of oleic acid with improvement in phospholipid acyl chain synthesis. Positional analysis of the acyl chain distribution in the phospholipids showed that saturated acyl chains dominate the sn1-position indicating a preference for saturated fatty acids at this position. High-level production of saturated acyl-ACPs is required to offset the marked preference of the TesE thioesterase for unsaturated acyl-ACPs and allow the initiation of phospholipid synthesis.

are activated by the FakA/B system to form the acyl-PO 4 required for G3P acylation at the sn1-position to initiate phospholipid synthesis (Fig. 1A) (2).
E. faecalis is an opportunistic pathogen with the high level of antibiotic resistance and can cause hospital-acquired diseases (9). E. faecalis inhabits the gastrointestinal tracts of humans and other animals as a facultative anaerobe (9). E. faecalis synthesizes membrane phospholipids from either de novo synthesized fatty acids or from exogenous free fatty acids and PlsX plays a key role in both processes (10)(11)(12). Like S. pneumoniae, E. faecalis has clustered fab genes and its plsX gene is located adjacent to the acpB gene encoding an auxiliary ACP involved in the regulation of fatty acid synthesis (10)(11)(12). However, the E. faecalis ∆plsX strain differs from that of S. pneumoniae in that the E. faecalis ∆plsX strain shows extremely weak growth and is deficient in phospholipid acyl chain synthesis, whereas the S. pneumoniae ∆plsX strain grows normally (8).
We report that the E. faecalis ∆plsX strain is a fatty acid auxotroph and growth was restored by supplementation with appropriate exogenous fatty acids. Although E. faecalis encodes TesE, a functional thioesterase, this enzyme does not restore growth as was reported for the S. pneumoniae ∆plsX strain (8). However, on high-level overexpres sion of TesE, growth was restored. A ∆plsX ∆fabT strain lacking the FabT repressor had a growth phenotype similar to that of the ∆plsX strain, although it showed enhanced de novo phospholipid acyl chain synthesis compared to the ∆plsX strain. We identified a ∆plsX suppressor strain that allowed wild-type growth by giving increased synthesis of saturated fatty acyl-ACP species which following PlsX-catalyzed conversion to acyl-PO 4 initiate phospholipid synthesis by acylation of the sn1-position of G3P. Overexpression of an enoyl-ACP reductase had a similar effect. TesE has a marked preference for the cleavage of unsaturated acyl-ACPs, and overproduction of saturated acyl-ACPs overcomes this preference.

Materials
The 15-methyl palmitic acid was purchased from Cayman Chemicals and all the other fatty acids, antibiotics, phospholipase A2, ortho-nitrophenyl-β-galactoside, and agmatine sulfate were purchased from Sigma-Aldrich. The media were purchased from Fisher Scientific. The DNA polymerase, restriction endonuclease, and T4 ligase were purchased from New England Biolabs. Sodium [1-14 C]acetate (specific activity, 57.0 mCi/mmol) and [1-14 C]stearic acid (specific activity, 53.0 mCi/mmol) were provided by Moravek, Inc, and the [1-14 C]oleic acid (specific activity, 55 mCi/mmol) was purchased from American Radiolabeled Chemicals. Ni-NTA resin was from Invitrogen, and the DNA purification kits were from Qiagen. Silver nitrate silica gel thin layer plates were purchased from Analtech and M17 Broth was from Becton Dickenson. All the other reagents were of the highest available quality. Oligonucleotide primers were synthesized by Integrated DNA Technologies, and DNA sequencing was performed by ACGT, Inc.

Bacterial strains, plasmids, and incubation
The bacterial strains and plasmids used in this study are listed in Table S1, and the primers used for this study are listed in Table S2. E. coli cultures were incubated at 37℃ in Luria-Bertani medium (tryptone, 10 g/L; yeast extract, 5 g/L; NaCl, 10 g/L), whereas E. faecalis cultures were grown at 37℃ in M17 medium (BD DIFCO) or in AC medium (tryptone, 10 g/L; yeast extract, 10 g/L; K 2 HPO 4 , 5 g/L; glucose, 1 g/L). Antibiotics were added at the following concentrations (in milligram per liter): sodium ampicillin, 100 for E. coli; kanamycin sulfate 50 for E. coli; chloramphenicol 30 for E. coli, and 5 for E. faecalis; erythromycin at 250 for E. coli and 5 for E. faecalis. Fatty acids were added at 0.1 mM unless otherwise stipulated.

Construction of the E. faecalis ∆plsX ∆fabT strain
The construction of E. faecalis ∆plsX ∆fabT strain used plasmid pQZ149 containing the plsX gene knockout cassette described in the previous work (11,12). E. faecalis ∆fabT cells transformed with the plasmid were selected on AC medium agar plates with 5 mg/L erythromycin and 100 mg/L 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal) at 30℃. A blue colony was selected and streaked on AC agar plates with the same components described above at 42℃ to verify plasmid integration into the genome. The blue colony was incubated in AC liquid medium in the presence of 0.1 mM palmitate at 30℃ for 4 hours, and the culture was shifted to 42℃ overnight. This process was repeated several times, and the culture was diluted and spread on AC agar plates with 0.1 mM palmitate and 100 mg/L X-Gal at 42℃. The white colonies were picked for colony PCR to verify the deletion of the plsX gene.

E. faecalis thioesterase assay
The plasmid for expressing E. faecalis thioesterase was constructed by inserting its coding region amplified by primer set EftesE NdeI F and EftesE EcoRI R into the vector pET28b NdeI and EcoRI sites. To express E. faecalis thioesterase, E. coli Rosetta cells transformed with the expression plasmid above were incubated at 37℃ to optical density at 600 nm (OD 600 ) of 0.6 and then induced with 1 mM isopropyl β-D-1-thiogalac topyranoside for 4 hours. The cells were harvested, washed with phosphate-buffered saline (PBS), resuspended with lysis buffer containing 50 mM sodium phosphate (pH 8.0), 0.3 M NaCl, 10 mM imidazole, and 1 mM dithiothreitol (DTT) and then lysed with a French Pressure Cell. The supernatant after centrifugation was loaded onto the Ni-NTA column, and the protein was eluted with 0.25 M imidazole, dialyzed, and stored with 20% glycerol at −80℃. To test the function of purified E. faecalis thioesterase in vitro, various species of E. faecalis acyl-ACPs were prepared as described in the previous work (11)(12)(13) and then mixed with the purified E. faecalis thioesterase in buffer containing 50 mM Tris-HCl (pH 8.0). The reaction was incubated at 37℃ for 1 hour, and the products were analyzed by conformation-sensitive 2 M urea-18% PAGE.

Growth measurements of E. faecalis strains
The cultures were started at OD 600 of 0.01 in M17 medium at 37℃ with or without the presence of a single exogenous fatty acid. The OD 600 values of the culture were measured every 2 hours in triplicate for 8 hours.

Construction of E. faecalis plsX, fabK, fabI, and tesE overexpression plasmids
The construction of E. faecalis plsX overexpression plasmid was slightly modified from that described previously (12), whereas the plsX gene was ligated to the pQZ28 vector derived from pZL277 vector by replacing the original chloramphenicol resistance gene with the erythromycin resistance gene from the pBVGh vector (12,14,15).
To construct the E. faecalis fabK overexpressing plasmid, the fabK gene was amplified from genomic DNA using primer set EffabK NcoI F and EffabK EcoRI R and then inserted into the NcoI and EcoRI sites of vector pQZ28 (12,14,15). For construction of the fabI overexpression plasmid the fabI gene was amplified using primers EffabI NcoI and EffabI EcoRI and inserted into vector pQZ28 as above.
To construct the E. faecalis tesE overexpressing plasmid, the tesE coding sequence together with its promoter was amplified using primer set EftesE promoter BamHI F and EftesE SmaI R and the produced fragment was inserted into the BamHI and SmaI sites of low-copy-number shuttle vector pTRKL2. To construct the E. faecalis tesE controllable overexpression plasmid, an agmatine-inducible system was used (10,16). The fragment containing the promoter of putrescine transcarbamylase (aguB) and the regulator gene aguR was amplified using primer set EfaguR EcoRI R and EfaguB promoter R whereas the tesE coding region was amplified using primer set EftesE F and EftesE NcoI R. The two fragments synthesized above were combined through overlap PCR using primer set EfaguR EcoRI R and EntesE NcoIR and the resulting fragment was inserted into the NcoI and EcoRI sites of pQZ28 vector.

β-galactosidase assays
The construction of reporter plasmids expressing E. coli β-galactosidase from E. faecalis fabT promoter, fabI promoter, or fabO promoter irrespectively has been described in the previous work (12). To construct a reporter plasmid expressing β-galactosidase from the E. faecalis thioesterase promoter, the promoter and the first 35 bp of E. faecalis tesE coding regions (−274 to +35) relative to the tesE gene GTG initiation codon were amplified using primer set EftesE promoter plus 35 PstI F and EftesE promoter plus 35 SalI R and inserted into PstI and SalI sites at the 5′ end of a promoterless lacZ gene on plasmid pBHK322 constructed in the previous work (12).
To assay β-galactosidase activity, E. faecalis strains transformed with lacZ reporter plasmids above were incubated to mid-log phase at 37℃, harvested through centrifuga tion, washed by PBS, resuspended in Z buffer, lysed with chloroform and sodium dodecyl sulfate, and assayed for β-galactosidase activity. The data were collected in triplicate.

Thin layer chromatography (TLC) analysis of radioactive-labeled fatty acid methyl esters from bacterial phospholipids
To test de novo acyl chain biosynthesis, 5 mL of E. faecalis cultures was inoculated at OD 600 of 0.1 in AC medium and incubated at 37℃ for 6 hours in the presence of 1 mCi/L sodium [1-14 C]acetate with or without 0.1 mM of a single exogenous fatty acid. The cells were lysed with methanol-chloroform (2:1) solution, and the phospholipids were extracted by chloroform and then dried under nitrogen. The fatty acyl groups were methylated by 25% (wt/vol) sodium methoxide, extracted by hexanes, and processed for TLC analysis on Analtech silica gel containing 20% silver nitrate in toluene at −20°C. The TLC plates containing the [ 14 C]-labeled fatty acid methyl esters were exposed and quantitated by phosphorimager analysis on a GE Typhoon FLA700 Scanner and the data were analyzed by ImageQuant TL software.
To test the incorporation of unsaturated fatty acids, E. faecalis strains were started at OD 600 of 0.1 in 5 mL of AC medium containing 0.1 µCi/mL [1-14 C]oleic acid with 0.1 mM non-radioactive oleate and cultured at 37℃ for 6 hours. The cells were washed with phosphate-buffered saline and the phospholipids were extracted, methylated, and analyzed as described above.
To test the incorporation of saturated fatty acids, E. faecalis strains were started at OD 600 of 0.1 in AC medium containing 0.1 µCi/mL [1-14 C]stearic acid plus 0.1 mM non-radioactive palmitate, cultured at 37℃ for 6 hours, and processed for TLC analysis as described above.

Gas chromatography-mass spectrometry analysis of fatty acid components of cell membrane
E. faecalis strains were inoculated at an OD 600 of 0.1 in AC medium with or without the presence of 0.1 mM of a single fatty acid and cultured at 37℃ for 6 hours. The conversion of phospholipids to fatty acid methyl esters was the same method used for TLC analysis above, and the extracted products were sent for gas chromatography-mass spectrometry analysis.

Phospholipase A2 (PLA 2 ) treatment for the positional analysis of fatty acids of bacterial phospholipids
The process to determine the positional analysis of the acyl chains of phospholipids has been described previously (17). The extracted phospholipids were hydrated at 25℃ for 20 min with 0.5-mL reaction buffer containing 0.1 mM Tris-HCl (pH 7.5), 10 mM CaCl 2 , and 10 mM MgCl 2 followed by sonication at 4℃. Ten units of PLA2 (from bee venom, Sigma) dissolved in the same buffer was added and the reaction was incubated at 37℃ for 3 hours 0.5 mL of 0.5 M NaCl was added, and the reaction was terminated by adding 2 mL of methanol-chloroform (2:1). The organic phase was separated, collected, and 1-mL chloroform was added to complete the extraction. The digestion products were dried under nitrogen, and the lysophospholipid fraction was submitted for mass spectrometric analysis.

Loss of the E. faecalis plsX gene leads to growth deficiency and disruption of lipid metabolism
Unlike the S. pneumoniae ∆plsX strain (8) which grows normally, the E. faecalis ∆plsX strain showed only trace growth (very minute colonies) on M17 agarose plates which lack free fatty acids where the FA2-2 wild-type strain grew well (Fig. 1B). To test whether the growth deficiency of the ∆plsX strain was related to the inhibition of the synthesis of phospholipid acyl chains, we performed [1-14 C]acetate labeling followed by argentation TLC, which showed a 20-fold decrease in phospholipid acyl chain synthesis (Fig. 1C). The ∆plsX strain also showed a three-fold decrease in [ 14 C]oleic acid incorporation relative to the wild-type strain (Fig. 1D). Moreover, the ∆plsX strain synthesized dramatically increased levels of abnormally long-chain phospholipid acyl chains. Both saturated (C18:0) and unsaturated (C20:1 and C22:1) species were found although these are only trace components in the phospholipids of the wild-type strain (Fig. 1E). These abnormally long chains indicate reduced rates of transfer of acyl chains to G3P (or 1-acyl G3P) such that acyl-ACPs undergo additional cycles of elongation. This phenomenon was first seen in E. coli cultures starved for G3P (18). All ∆plsX phenotypes were abolished on the introduction of a plasmid-borne plsX gene into the ∆plsX strain ( Fig. 1B through E). These data indicated that the growth defect of the E. faecalis ∆plsX strain was likely caused by poor acylation of G3P, which resulted in deficient phospholipid synthesis.
Note, that we have measured fatty acid synthesis indirectly by analysis of the phospholipid acyl chains. Although the acyl chains are made by the fatty acid synthesis pathway, we analyze the chains after they become phospholipid acyl groups. Hence, decreased incorporation of [1-14 C]acetate can be due to the decreased synthesis per se or decreased transfer of acyl chains to the G3P phospholipid backbone or both. For example, when position sn1 of G3P is not acylated, position sn2 cannot be acylated.

Exogenous fatty acids restore the growth of the E. faecalis ∆plsX strain
To test the effects of various species of exogenous fatty acids on the growth of the E. faecalis ∆plsX strain, the strain was streaked on M17 agarose plates containing a single exogenous fatty acid at 0.1 mM. Supplementation with palmitic acid (C16:0) or oleic acid (C18:1, cis-9) restored the growth of the ∆plsX strain ( Fig. 2A). Stearic acid (C18:0) failed to support growth on M17 agarose plates but supported good growth in M17 liquid medium (Fig. S1). A possible explanation is the agarose trapping of micelles of the very poorly soluble stearic acid. Growth could also be restored by 15-methyl palmitic acid (15-methyl C16:0) which can be distinguished by mass from endogenously synthesized palmitic acid (Fig. 2B). However, the short-chain fatty acid, octanoic acid (C8:0), and the hydroxy acid 12-hydroxyoctadecanoic acid (12-OH C18:0) failed to allow growth of the ∆plsX strain (Fig. 2B). Growth of the E. faecalis ∆plsX strain was not supported by myristic acid (C14:0), a minor component of the acyl chains of the wild-type strain phospholipids ( Fig. 2C and Fig. 1D). Supplementation with palmitoleic acid (cis-9 C16:1) or linoleic acid (C18:2) failed to improve the growth of the ∆plsX strain (Fig. 2D), which could be caused by toxicity of these acids. In addition, supplementation with eicosenoic acid (cis-11 C20:1) or erucic acid (cis-13 C22:1) gave increased growth of the ∆plsX strain (Fig. 2E). To test the effects of exogenous fatty acids on the expression of fab genes in the ∆plsX strain, we measured lacZ gene expression driven by the fabT promoter (12). Relative to the ∆plsX strain incubated without exogenous fatty acids, lacZ expression was strongly increased in cells incubated with palmitic acid (threefold increase) or oleic acid (sixfold increase), whereas no change was observed in cultures grown with octanoic acid (Fig.  2F).

Growth of the ∆plsX strain is restored by high-level expression of the TesE thioesterase
One possible explanation for the growth deficiency of the E. faecalis ∆plsX strain might be the lack of a functional acyl-ACP thioesterase which would block the release of free fatty Research Article mSphere acids from de novo synthesized acyl-ACPs. The lack of free fatty acids would preclude the production of acyl-PO 4 by the fatty acid kinase system. Absence of acyl-PO 4 for PlsY acylation of G3P position sn1 would block the initiation of phospholipid synthesis (Fig.  1A). Note that exogenous fatty acids bypass the need for a thioesterase because these are directly converted to acyl-PO 4 species (4).
To test the thioesterase hypothesis, we identified an E. faecalis gene we call tesE (thioesterase Enterococcus) (locus tag EF_RS01825), which encodes a protein 39% identical to S. pneumoniae TesS. TesE was expressed and purified from E. coli and tested for the cleavage of various acyl species of both E. faecalis ACPs. Conformationsensitive gel electrophoresis showed almost complete cleavage of acyl derivatives of both AcpA and AcpB to holo-ACP ( Fig. 3A and B) indicating that E. faecalis TesE enc odes a functional thioesterase. However, TesE is a weak thioesterase, probably even weaker than S. pneumoniae TesS, a reportedly weak enzyme (8). TesE shows a distinct preference for the cleavage of the unsaturated acyl-ACP, oleoyl-ACP, over the saturated acyl-ACP, palmitoyl-ACP (Fig. 3C), whereas S. pneumoniae TesS is reported to cleave these substrates equally well (8). A strong preference for oleoyl-ACP cleavage was seen in cell extracts (Fig. S2) as well as with the purified TesE. Expression of E. faecalis tesE  Table S3. N/A denotes no additions.
Research Article mSphere gene assayed by a transcriptional fusion to β-galactosidase showed a 50% decrease in β-galactosidase activity in the ∆plsX strain compared to wild-type strain (Fig. 3D). However, overexpression of E. faecalis TesE strain from its native promoter on a low copy number plasmid failed to enhance the growth of the ∆plsX strain. We therefore used the regulated agmatine expression system (16) to allow the construction of a TesE expression plasmid. Exogenous agmatine controls the expression of agmatine degradation genes (16). A high copy plasmid (pQZ28) containing the aguR gene (which encodes a LuxR family transcriptional regulator of agmatine degradation) together with the promoter for the agmatine degradation genes was used to give inducible expression by agmatine addition. On induction with agmatine, the ∆plsX strain grew without fatty acid supple mentation, whereas no growth was seen in the absence of induction ( Fig. 3E and Fig. S3). Therefore, the growth deficiency of the ∆plsX strain is due to insufficient expression of the chromosomal tesE gene. Note that we made extensive attempts to delete tesE from the E. faecalis chromosome without success suggesting that it has an essential role in metabolism.

E. faecalis ∆plsX ∆fabT strain has essentially the same growth phenotype as the ∆plsX strain
The above data suggested that the growth deficiency of E. faecalis ∆plsX strain might result from a deficiency in de novo fatty acid biosynthesis caused by weak expression of fab genes. To test this premise, the ∆plsX ∆fabT strain was constructed to increase the expression of the fab genes (12). Lack of the FabT repressor increases the expression of the fatty acid synthesis genes by about threefold. However, the E. faecalis ∆plsX ∆fabT and ∆plsX strains showed similar growth-defective phenotypes ( Fig. 4A through C). As expected, exogenous fatty acid supplementation restored the growth of the ∆plsX ∆fabT strain (Fig. 4B). Examination of the fatty acyl chain compositions of the phospholipids showed a higher proportion of the long-chain unsaturated fatty acyl chains C20:1, C22:1, and C24:1 in the ∆plsX ∆fabT strain (Fig. 4D). Labeling with [1-14 C]acetate showed a two-fold increase in de novo phospholipid acyl chain fatty acid synthesis in the ∆plsX ∆fabT strain relative to the ∆plsX strain (Fig. 4E) as expected from increased fab gene expression. The increased levels of ≥C20-long unsaturated long-chain fatty acyl chains in the ∆plsX ∆fabT strain phospholipids result from increased elongation capacity. Indeed, the level of the C24 unsaturated acyl chain (presumably C24∆17) of the ∆plsX ∆fabT strain showed an increase of >10-fold relative to the ∆plsX strain (Fig. 4D). These data show that defective acylation of G3P rather than a lack of acyl chain synthesis causes the defective growth of the ∆plsX strain, and the wild-type level of TesE expression is too low to provide relief.

Increased synthesis of saturated fatty acyl chains in the E. faecalis ∆plsX suppressor strain
An E. faecalis ∆plsX suppressor colony isolated from a ∆plsX M17 agarose plate had a wild-type growth phenotype (Fig. 5 and B). Examination of the acyl chain components of the phospholipids showed a higher proportion of saturated acyl chains relative to the E. faecalis wild-type strain (Fig. 5C). When compared to the ∆plsX strain, the ∆plsX suppressor strain had a sixfold lower level of the C20:1 long-chain unsaturated acyl chain (Fig. 5C). However, the suppressor strain incorporated almost the same level of [1-14 C]oleic acid as the ∆plsX strain (Fig. 5D). Labeling with [1-14 C]acetate indicated a 12-fold increase in phospholipid acyl chain synthesis in the ∆plsX suppressor strain compared with the ∆plsX strain (Fig. 5E) compared to only a two-to threefold increase in the expression of the fab gene cluster as assayed by the fab-lacZ fusion (Fig. 5F-H). Moreover, expression of an E. faecalis tesE-lacZ fusion in the ∆plsX suppressor strain showed an about two-fold increase relative to the ∆plsX strain (Fig. 5I). High-throughput genomic sequencing (Oxford Nanopore confirmed by Illumina sequencing) of the ∆plsX suppressor strain identified a nonsense mutation (C1066 to T) in the fabO gene, which converted Gln356 to a termination codon resulting in a truncated FabO (β-ketoacyl-ACP synthase I) protein that lacked the last 56 residues. These residues are part of a β-sheet buried within the structure of the S. pneumoniae FabF (PDB 2ALM) and hence their loss is expected to destabilize and inactivate FabO. Note that unsaturated acyl chain synthesis is not blocked by the loss of FabO activity because the FabF 3-ketoacyl-ACP synthase II has weak FabO activity (10), which results in a strain having decreased synthesis of unsatura ted acyl chains. Strikingly, the growth of the ∆plsX strain could be also recovered by overexpression of an E. faecalis enoyl-ACP reductase, either FabI or FabK (Fig. 6A and B).
Overexpression of FabI or FabK results in greatly increased saturated acyl chain synthesis at the expense of unsaturated acyl chains (10,11) and gave a 30-fold (FabI) or 20-fold (FabK) increase in de novo acyl chain synthesis compared to the ∆plsX strain lacking overexpression (Fig. 6C). However, the FabK overexpression strain grew more slowly than Research Article mSphere the FabI overexpression strain perhaps due to the alteration of cell membrane stability by excess saturated fatty acyl chains (Fig. 6B) (10).

E. faecalis preferentially acylates the sn1-position of glycerophospholipids with saturated fatty acids
To test the acyl chain composition of the sn1-position of E. faecalis phospholipids, the ∆plsX strain was grown with or without exogenous fatty acids (Fig. 7A). Growth of the ∆plsX strain increased when palmitic acid was present, indicating that saturated fatty acids in the G3P sn1-position enhance growth (Fig. 7A). Moreover [1-14 C]acetate Inhibition of oleic acid incorporation in the ∆plsX and ∆plsX suppressor strains was also seen by GC-MS (Table S4). (E) De novo phospholipid acyl chain synthesis determined by [1-14 C]acetate labeling of wild-type, ∆plsX, and ∆plsX suppressor strains. The isotopic compound incorporated is given above the autoradiograms.
The numbers above the lanes are the radioactive label incorporations relative to the value for the wild-type strain (100). The red numbers are to focus the reader on the relevant data. low-copy vector at the right of ∆plsX suppressor was deleted since it was replaced with the strain overexpressing tesE gene from the agmatine-induced promoter on the pQZ28 high-copy vector in this study.
Research Article mSphere labeling showed that palmitic acid was more efficient than oleic acid in stimulating acyl chain synthesis in the ∆plsX strain (Fig. 7B). Unlike incorporation of [1-14 C]oleic acid ( Fig. 1D and Fig. 5D), labeling with [1-14 C]stearic acid showed very similar incorpo ration into the cell membrane phospholipids of wild-type and ∆plsX strains (Fig. 7C).
To determine the acyl chains esterified to the sn1-position of phospholipids in the E. faecalis wild-type phospholipids, phospholipase A2 (PLA 2 ) was used to cleave the acyl chains at sn2-position and convert phosphatidylglycerol, the major phospholipid species, to 1-acylphosphatidylglycerol (17). Mass spectral analysis of the derived 1-acylphosphati dylglycerol indicated that about 70% of acyl chains were saturated, mainly C16:0 and C18:0, indicating that PlsY favors the incorporation of saturated acyl chains in de novo synthesis of glycerophospholipids (Fig. 7D).

DISCUSSION
The growth of E. faecalis ∆plsX strain is extremely poorly. This is in marked contrast to the normal growth reported for S. pneumoniae ∆plsX strains (8). The poor growth of the E. faecalis ∆plsX strain results from an inability to efficiently acylate the G3P precursor of phospholipid synthesis, as shown by the dramatic increase in acyl chains of ≥C20. The growth defect of the E. faecalis ∆plsX strain can be overcome by the overproduction of saturated acyl-ACP species or by supplementation with exogenous fatty acids. The first clue to the role of increased saturated acyl-ACP species came from the ∆plsX suppressor strain that restored growth by expression of a truncated FabO protein. FabO encodes a 3-ketoacyl-ACP synthase I that elongates the cis-3-decenoyl-ACP required to initiate unsaturated fatty acid synthesis (10). However, the loss of FabO does not block unsatura ted fatty acid synthesis in E. faecalis because the FabF 3-ketoacyl-ACP synthase II can weakly elongate cis-3-decenoyl-ACP and provide decreased, albeit sufficient, unsatura ted fatty acid synthesis (10). The loss of FabO diverts nascent acyl-ACPs to saturated phospholipid acyl chain synthesis of wild-type, ∆plsX, ∆plsX/p-fabI, and ∆plsX/p-fabK strains assayed by [1-14 C]acetate labeling. In panel B, the growth curve for each strain was measured from independent triplicate cultures. The numbers above the lanes are the radioactive label incorporation relative to the value for the wild-type strain (100). The red numbers are to focus the reader on the relevant data. In panel C, increased phospholipid acyl chain synthesis in the ∆plsX/p-fabK strain (88% of the wild-type strain) was detected in an independent experiment. GC-MS analysis of the phospholipid acyl chains of the E. faecalis wild-type, ∆plsX, and ∆plsX/p-fabK and ∆plsX/p-fabI strains is given in Table S4.
Research Article mSphere acyl-ACP species (Fig. 5C). Consistent with this result, the growth deficiency of the E. faecalis ∆plsX strain was also overcome by overexpression of the FabK or FabI enoyl-ACP reductases, which results in increased saturated acyl chain synthesis ( Fig. 6A through C). On overexpression, these enoyl-ACP reductases intercept the trans-2 decenoyl-ACP intermediate of the FabN dehydrase/isomerase resulting in the blockage of unsaturated fatty acid synthesis (10). Both FabO truncation and enoyl-ACP reductase overexpression bypass the loss of PlsX by increasing the synthesis of saturated acyl chains at the expense of unsaturated acyl chains. The most straightforward mechanism for bypass of loss of PlsX would be cleavage of the saturated acyl-ACPs by TesE to give fatty acids for the conversion to the acyl-PO 4 species required for PlsY acylation of G3P position 1. TesE cleaves unsaturated acyl-ACPs much more readily than saturated acyl-ACPs ( Fig. 3 and Fig. S2). However, the level of TesE activity produced by the chromosomal tesE gene is too low to provide sufficient saturated acyl-ACPs to allow normal growth of the ∆plsX strain. However, TesE overexpression increased the cleavage of saturated acyl-ACPs by outcompeting the unsaturated acyl-ACPs for access to TesE resulting in the production of saturated acyl chains. The saturated fatty acids are converted to acyl-PO 4 that acylates the sn1-position of G3P and initiates phospholipid synthesis. This mechanism explains the rescue of growth of the ∆plsX strain by supplementation with palmitic acid, which is directly converted to palmitoyl-PO 4 .
Although this mechanism explains the data reported above, caveats remain. We have been unable to delete the tesE gene suggesting that it may have a physiological role other than acyl-ACP cleavage. Another complication is that the FakA fatty acid kinase activity requires a FakB fatty acid-binding protein. E. faecalis expresses four FakB proteins, whereas S. pneumoniae expresses three such proteins, each binding a defined fatty acid species (19). However, we have yet to detect fatty acid-binding specificity in any of the four E. faecalis FakB proteins (12). Moreover, the specificities of the E. faecalis G3P acyltransferases, PlsY and PlsC, have not been explored. Yet another complication is that the two E. faecalis ACPs have different biases in the PlsX reaction (11). AcpA favors the formation of acyl-phosphates, whereas AcpB favors the formation of acyl-ACPs. A further puzzle is the requirement for high levels of saturated acyl-ACPs to bypass the lack of PlsX. This is a puzzle because strains of E. faecalis totally blocked in fatty acid synthesis grow well with only oleic acid supplementation (15).
The saturated acyl-ACP requirement may be due to the loss of the substrate-channel ing function of PlsX (7). Recent reports indicate that PlsX shuttles between the cytosol and membrane and that in vivo function of the enzyme requires membrane binding which is proposed due to PlsX-binding acyl-PO 4 and presenting this substrate to the membrane-bound PlsY (6,7). In the absence of PlsX, PlsY may show a strong preference for saturated versus unsaturated acyl-PO 4 species. Further work will be required to understand these complex interactions.
It may seem surprising that the E. faecalis ∆plsX strain behaves differently from the S. pneumoniae ∆plsX strain since both species encode functional thioesterases that could provide fatty acids for acyl-PO 4 synthesis. However, the E. faecalis thioesterase appears weaker than that of S. pneumoniae, poorly expresses and strongly favors the cleavage of unsaturated acyl-ACPs, whereas the S. pneumoniae thioesterase has no such bias (8). Moreover, TesE is not coregulated with the fatty acid synthesis genes (Fig. S4). In agreement with Parsons and coworkers (8), we suspect that the TesE and TesS activities are the side reactions of an esterase or an acyltransferase lacking its acyl acceptor.