Discovery of the Elusive UDP-Diacylglucosamine Hydrolase in the Lipid A Biosynthetic Pathway in Chlamydia trachomatis

ABSTRACT Constitutive biosynthesis of lipid A via the Raetz pathway is essential for the viability and fitness of Gram-negative bacteria, including Chlamydia trachomatis. Although nearly all of the enzymes in the lipid A biosynthetic pathway are highly conserved across Gram-negative bacteria, the cleavage of the pyrophosphate group of UDP-2,3-diacyl-GlcN (UDP-DAGn) to form lipid X is carried out by two unrelated enzymes: LpxH in beta- and gammaproteobacteria and LpxI in alphaproteobacteria. The intracellular pathogen C. trachomatis lacks an ortholog for either of these two enzymes, and yet, it synthesizes lipid A and exhibits conservation of genes encoding other lipid A enzymes. Employing a complementation screen against a C. trachomatis genomic library using a conditional-lethal lpxH mutant Escherichia coli strain, we have identified an open reading frame (Ct461, renamed lpxG) encoding a previously uncharacterized enzyme that complements the UDP-DAGn hydrolase function in E. coli and catalyzes the conversion of UDP-DAGn to lipid X in vitro. LpxG shows little sequence similarity to either LpxH or LpxI, highlighting LpxG as the founding member of a third class of UDP-DAGn hydrolases. Overexpression of LpxG results in toxic accumulation of lipid X and profoundly reduces the infectivity of C. trachomatis, validating LpxG as the long-sought-after UDP-DAGn pyrophosphatase in this prominent human pathogen. The complementation approach presented here overcomes the lack of suitable genetic tools for C. trachomatis and should be broadly applicable for the functional characterization of other essential C. trachomatis genes.

the hydrolysis of UDP-2,3-diacylglucosamine (UDP-DAGn) to lipid X at the fourth step of the pathway. The conversion of UDP-DAGn to lipid X is carried out by either LpxH in beta-and gammaproteobacteria or LpxI in alphaproteobacteria; LpxH and LpxI are unrelated to each other in sequence or catalytic mechanism (5,6). Neither LpxH nor LpxI is found in C. trachomatis, leaving a critical gap in the knowledge of lipid A biosynthetic genes in this prominent human pathogen (Fig. 1C).
Due to a lack of robust tools for genetic manipulation in Chlamydiae, the identification and functional characterization of Chlamydia gene products has presented unique challenges. To overcome this obstacle and identify the elusive UDP-DAGn hydrolase in C. trachomatis, we developed a genetic complementation screen in an E. coli strain engineered to have temperaturecontrolled expression of lpxH. Using a C. trachomatis expression library, we identified a previously uncharacterized open reading frame (ORF)-C. trachomatis ORF 461 (ct461), renamed lpxGthat is conserved in all Chlamydia species. LpxG exhibits robust UDP-DAGn hydrolase activity in vitro and in vivo. Overexpression of LpxG resulted in the accumulation of lipid X in C. trachomatis and profoundly reduced bacterial infectivity, validating LpxG as a novel member of the UDP-DAGn hydrolases and highlighting the importance of regulated lipid A biosynthesis in Chlamydia pathogenicity. We suggest that the approach presented in this study may be widely applicable for functional screening of uncharacterized essential genes in C. trachomatis. A. (C) Conservation of lipid A biosynthetic enzymes in E. coli and C. trachomatis. Essential enzymes are marked by red blocks, nonessential enzymes by blue blocks, and the absence of an enzyme in the genome of C. trachomatis is shown by a white block. The existence of an LpxH functional ortholog, LpxI, in alphaproteobacteria is denoted by the orange block next to LpxH. Young et al. encoding the UDP-DAGn hydrolase activity in the lipid A biosynthetic pathway of C. trachomatis, we devised a temperaturesensitive genetic complementation screen in E. coli ( Fig. 2A). The temperature-sensitive plasmid pKJB5 containing E. coli lpxH was transformed into E. coli BL21(DE3), and the chromosomal copy of lpxH in these transformed cells was replaced with a ⌬lpxH::kan cassette from E. coli strain W3110A⌬HEc (6) by P1vir-mediated transduction to generate E. coli strain HY1. We reasoned that since lpxH is an essential gene in E. coli, depletion of the temperaturesensitive plasmid at 44°C will be lethal unless the LpxH UDP-DAGn hydrolase activity is provided by a functionally equivalent Chlamydia gene product.

Identification of the
We constructed a C. trachomatis expression library from a collection of E. coli clones harboring all C. trachomatis ORFs (ORFome) in pDONR221 (7), which were transferred en masse to pDEST17 to place the expression of the inserted C. trachomatis DNA under the control of a T7 promoter and a 5=-end ribosomal binding site. HY1 cells were transformed with this C. trachomatis expression library to generate a collection of expression strains, which we will refer to as HY1_Lib. The equivalent of~25,000 clones of HY1_Lib were plated and incubated at 44°C. In parallel, HY1 cells were transformed with an empty vector (HY1_VC) to account for the background emergence of spontaneous temperature-resistant HY1 variants. Temperature-resistant clones arose in HY1_Lib at an~5-fold-greater rate than in HY1 cells transformed with the empty vector (Fig. 2B), suggesting that a factor within the C. trachomatis expression library complemented for the loss of the UDP-DAGn hydrolase activity.
To identify the Chlamydia gene that complemented an lpxH deletion, we first used PCR to determine the size of the Chlamydia ORF(s) harbored by the plasmids in temperature-resistant E. coli HY1 clones. Nearly 40% of the HY1_Lib colonies surviving at 44°C contained an insert of~1,100 bp in pDEST17 (Fig. 2C). DNA sequencing of the 1,100-bp PCR product revealed that these plasmids contained ct461, which was renamed lpxG.
The Chlamydia UDP-DAGn pyrophosphatase LpxG is a Mn 2؉ -dependent, membrane-associated calcineurin-like phosphoesterase enzyme. The Chlamydia lpxG gene encodes a 37-kDa uncharacterized putative metallophosphoesterase. The fulllength protein is predicted to have an N-terminal transmembrane helix (Fig. 3), which was truncated from the ORF during construction of the original C. trachomatis library (7). We verified that the UDP-DAGn hydrolase activity of LpxH in E. coli was also complemented by full-length LpxG, as cells expressing full-length LpxG remained viable upon P1vir-mediated transduction of a ⌬lpxH:: kan cassette (data not shown).
A protein-protein BLAST search uncovered LpxG orthologs in other Chlamydia species (Fig. 3). Alignments of these sequences revealed a DXH(X)~2 5 GDXXDR(X)~2 5 GNHD motif (with X indicating any amino acid) that is characteristic of enzymes in the calcineurin-like phosphoesterase (CLP) family (8)(9)(10). Proteins with this classification utilize a cluster of two to three metal ions to facilitate hydrolysis, with residues from this conserved motif coordinating the metal cofactors (8). Additionally, LpxG possesses a GHXH motif (with representing hydrophobic residues) found in a subclass of CLP enzymes (Fig. 3) (11,12). of identity, with darkness of shade corresponding to degree of conservation. CLP motifs are marked in black, with the key amino acid patterns highlighted in red. "X" represents any amino acid, and "" denotes any hydrophobic residue. The DXH(X)~2 5 GDXXDR(X)~2 5 GNHD motif is indicated by a red bar, and the location of the predicted transmembrane helix in the Chlamydia LpxG is indicated in pink. Sequences were obtained from the NCBI server, and alignment was carried out using ClustalW (32).

Young et al.
To characterize the in vitro activity of LpxG, we expressed C-terminally His 10 -tagged LpxG and an LpxG D59A variant (indicated by an asterisk in Fig. 3) that contained an alanine substitution for the first aspartate residue in the DXH(X)~2 5 GDXXDR(X)~2 5 GNHD motif, predicted to be required for metal chelating and efficient catalysis of CLP enzymes (8)(9)(10). After isopropyl-␤-D-thiogalactopyranoside (IPTG) induction, cell-free extracts (CFEs) were prepared from E. coli C41(DE3) cells expressing either LpxG or LpxG D59A or cells harboring an empty vector. The extracts were then tested for lipid X production from UDP-DAGn using a thin-layer chromatography (TLC)-based radiographic assay in the presence of a mixture of unlabeled and ␤-32 P-labeled substrate. A significantly higher rate of lipid X accumulation was observed in cell extracts from cells expressing LpxG than in an equal amount of extracts prepared from the vector control strain (Fig. 4A). The lipid product generated in the LpxG reactions has a migration property similar to that of the lipid product generated with purified Haemophilus influenzae LpxH (13), supporting its identity as lipid X. In contrast, the CFE from cells expressing LpxG D59A displayed little activity, with a level of UDP-DAGn hydrolysis similar to that seen for vector control samples despite a significant level of protein expression, suggesting that the D59A substitution in the predicted metal-chelating motif significantly compromised the catalytic activity of LpxG.
As our in vitro assay with CFE of cells expressing LpxG verified an important catalytic residue in the predicted metal-chelating motif of the CLP family of enzymes, we next examined the metal dependency of LpxG using Ni 2ϩ -NTA-purified enzyme in the presence of detergent (Fig. 4B). The purified enzyme was preincubated with EDTA to remove any copurifying metals. LpxG was then diluted into reaction buffer containing one of the following chloride salts at 1 mM: Ca 2ϩ , Co 2ϩ , Cu 2ϩ , Fe 3ϩ , Mg 2ϩ , Mn 2ϩ , Ni 2ϩ , or Zn 2ϩ . Assay conditions where the di-or trivalent metal salt was replaced with 2 mM NaCl or no metal (i.e., in the presence of 1 mM EDTA) were also included as controls. The Mn 2ϩ assay condition showed activity that was 35-fold greater than that of the identical reaction with no metal (Fig. 4B). Co 2ϩ also enhanced activity, albeit at lower levels (3-fold). The other metal cofactors had little effect on activity. The activity of the enzyme was unaffected by the presence of NaCl, indicating that the ionic strength of the reaction was not responsible for the observed metal stimulation. Overall, our results indicated that Mn 2ϩ was the most efficient and thus the most likely cofactor for LpxG activity.
LpxG hydrolyzes the ␣-phosphorus atom of UDP-DAGn. While the other two known UDP-DAGn hydrolases, LpxH and LpxI, both catalyze the formation of lipid X, they do so through different mechanisms: LpxH attacks the ␣-phosphate of the substrate (5), whereas LpxI attacks at the ␤ position (6). To examine the mechanism employed by LpxG, purified enzyme was used to convert UDP-DAGn to lipid X in the presence of H 2 16 O and a mixture of H 2 18 O-H 2 16 O (70:30, vol/vol). The lipid X and UMP products were extracted using an acidic single-phase Bligh-Dyer system (14), as was reported previously (6). A sample of the LpxG reaction extraction product was analyzed by reverse-phase liquid chromatography coupled with electrospray ionization/mass spectrometry (LC-ESI/MS) operated in the negative ion mode. The UMP product, with a predicted [MϪH] Ϫ ion at m/z 323.029, was detected at m/z 323.028 (Fig. 5A). For the H 2 18 O-labeled reaction mixture,~70% of the UMP product contained 18 O, as shown by the presence of a much more intense peak at m/z 325.031 versus that at m/z 323.027 (Fig. 5A, bottom). The lipid X product, predicted to have a [MϪH] Ϫ ion at m/z 710.425, was also detectable at the predicted m/z. In this case, no mass shift in the lipid X peak was observed in the product of the reaction carried out in the presence of H 2 18 O (Fig. 5B). Taken together, our observations indicate that LpxG catalyzes the attack of water exclusively on the ␣-phosphorus atom of UDP-DAGn and incorporates the solventderived oxygen atom into UMP instead of lipid X (Fig. 5C).
Overexpression of LpxG in C. trachomatis leads to accumulation of chlamydial lipid X and a loss of bacterial infectivity. To determine whether LpxG is a bona fide UDP-DAGn hydrolase, we examined the consequence of LpxG overexpression in C. trachomatis. Since the overexpression of many lipid A enzymes is toxic in E. coli due to the accumulation of lipid A intermediates (15), we predicted that overexpression of LpxG but not of the inactive LpxG D59A mutant should similarly cause buildup of lipid X and bacterial toxicity, resulting in reduced Chlamydia infectivity.
C. trachomatis lymphogranuloma venereum (LGV) strain L2 (LGV-L2) was transformed with an LpxG overexpression vector as described previously (16). The plasmid contained the FLAGtagged lpxG gene under the control of an anhydrous tetracycline (ATc)-inducible promoter, and the expression of LpxG was confirmed by immunoblot analysis with the anti-FLAG antibody (Fig. 6A). ATc was added at 16 h postinfection (hpi) to induce the overexpression of LpxG. For the analysis of lipids, Vero cells infected with strain LGV-L2 containing the LpxG expression vector were lysed and their lipids extracted by the Bligh-Dyer method (14,17). The relative abundance of the lipids was examined by normal phase LC-ESI/MS operated in the negative ion mode. Significant accumulation of lipid X with a [MϪH] Ϫ ion at m/z 778.584 (predicted m/z 778.524) was observed in the LpxG overexpression cells compared to the vector control cells or cells overexpressing the inactive LpxG D59A mutant (Fig. 6B), confirming that LpxG participates in lipid A biosynthesis in Chlamydia. Importantly, the localization of LpxG and the LpxG D59A mutant within infected cells was indistinguishable (Fig. 6C), indicating that the LpxG D59A mutation did not alter the stability of LpxG or its localization to bacterial membranes. ATc-induced overexpression of LpxG but not of the LpxG D59A mutant also resulted in a significant reduction in the generation of infectious Chlamydia EBs (Fig. 6D), a phenotype similar to what had been reported with LOS inhibitors (18). Overall, these findings highlight the importance of maintaining controlled lipid A biosynthesis and membrane integrity for the ability of Chlamydia to infect cells.

DISCUSSION
C. trachomatis is a leading cause of infectious blindness and the most prevalent sexually transmitted bacterial infection (1,2,19). Despite the clinical importance of Chlamydia, a functional characterization of its gene products has been hampered by a lack of robust genetic tools, leaving many biologically important activities unresolved, including the elusive gene encoding the UDP-DAGn pyrophosphatase activity that is essential for lipid A biosynthesis in this bacterium. Our studies reinforce the notion that, despite the phylogenetic differences between C. trachomatis and the model organism E. coli, these bacteria share similar molecular signatures of metabolic and biochemical needs, making it possible to use E. coli as a heterologous host to characterize Chlamydia gene products (20)(21)(22). The screening strategy of using a Chlamydia genomic library for heterologous expression in E. coli provides an efficient method to identify other C. trachomatis genes with unknown functions, especially those that complement essential cellular processes in Gram-negative bacteria.
Such an approach is exemplified by our identification of the Chlamydia UDP-DAGn pyrophosphatase, LpxG. LpxG shares extremely low sequence identity with either LpxH (11%) or LpxI (9%), making it impossible to identify LpxG solely based on sequence conservation and highlighting LpxG as the founding member of a third family of UDP-DAGn pyrophosphatases. Importantly, overexpression of LpxG results in the accumulation of lipid X in Chlamydia and leads to a decrease of bacterial infectivity. This observation, together with our previous report of a stalled Chlamydia infectious cycle when lipid A biosynthesis is blocked by pharmacological inhibition of LpxC (18), emphasizes the functional importance of maintaining balanced lipid A biosynthesis to generate infectious EBs. Given the crucial role of an intact outer membrane environment for the proper functionality of Chlamydia outer membrane proteins and secretion systems, it is likely that there exists not only a regulatory mechanism influencing the rate of lipid A biosynthesis but also a lipid A sensing and signaling pathway orchestrating the life cycle of Chlamydia infection.
The discovery of LpxG as a UDP-DAGn pyrophosphatase that is distinct from LpxH in beta-and gammaproteobacteria and LpxI in alphaproteobacteria represents a major step forward in our understanding of the lipid A biosynthetic pathway in Gramnegative bacteria. Recently, a potent small-molecule inhibitor of E. coli LpxH has been discovered through high-throughput screening (23). The exceedingly low sequence identity of the unique UDP-DAGn pyrophosphatase LpxG compared with LpxH and LpxI forecasts distinct structural features within LpxG that could be exploited for developing highly specific antibiotics for treating Chlamydia infections without causing major alterations in resident microbial communities or leading to unintended antibiotic resistance among other coinfecting pathogens. Plasmids, bacterial strains, and growth conditions. The plasmids and bacterial strains used in this study are listed in Table S1 and S2, respectively, in the supplemental material (24,25). P1vir phage lysate preparation and infections were carried out following standard procedures (26). Luria-Bertani (LB) broth (Difco, Detroit, MI) was used as the growth medium for liquid culture, and LB broth supplemented with 15 g/ liter Bacto agar was used for solid-phase growth. Antibiotics were used at the following concentrations: 100 g/ml ampicillin (Amp), 50 g/ml kanamycin (Kan), 25 g/ml chloramphenicol (Cam).

Chemicals and reagents.
Construction of pDEST C. trachomatis genomic library. Each clone from the C. trachomatis ORF library harbored in E. coli (7) was inoculated into a single well of a microtiter plate containing 200 l of LB supplemented with Kan and incubated for 18 h at 37°C. The overnight cultures were pooled, and plasmids were extracted to yield the plasmid library pDONR_Ctlib. Gateway technology was employed to transfer the ORF inserts from pDONR_CtLib into pDEST17 to generate pDEST_CtLib. Gateway LR Clonase (Invitrogen, Carlsbad, CA) reactions were carried out according to specifications from the manufacturer, except that reactions were performed at 25°C for 24 h. Aliquots from these reactions were transformed into E. coli C41(DE3) cells and plated on LB agar supplemented with Amp. To ensure at least a 20-fold coverage of the entire library, more than 20,000 colonies were collected and pooled, followed by outgrowth for 2 h at 37°C in LB. The final cell sample containing pDEST_CtLib was termed C41(DE3)_CtLib.
Complementation screen. A P1vir lysate was generated from an ⌬lpxH::kan E. coli strain harboring lpxH on the pBAD33 vector (strain W3110A⌬HEc) (see Table S2 in the supplemental material) and used to transduce E. coli C41(DE3) harboring lpxH on the pET21a vector [C41(DE3)EcH] (see Table S2). Transductants were selected on LB-agar plates containing Kan and Amp. The chromosomal deletion of the lpxH gene was confirmed by sequencing, and this strain was designated C41(DE3)⌬HEc (see Table S2).
To generate a controllable lpxH expression strain, E. coli lpxH on a temperature-sensitive plasmid (pKJB5) was transformed into E. coli BL21(DE3) cells, and the ⌬lpxH::kan cassette from strain C41(DE3)⌬HEc was transduced by P1vir transduction at 30°C, the permissive temperature for the plasmid pKJB5. The resulting strain, HY1 (see Table S2 in the supplemental material), harbored a deletion of lpxH, as confirmed by sequencing.
To screen the C. trachomatis ORFome for gene(s) that complemented a temperature-mediated lpxH disruption, pDEST_CtLib was transformed into strain HY1 and incubated at 44°C. A portion of the transformed cells was incubated at 30°C to determine the transformation efficiency. The pKJB2 plasmid encoding E. coli lpxH and an empty pET16b plasmid were also transformed into HY1 as positive and negative controls, respectively. The number of colonies was scored and compared to the numbers in the controls.
Molecular biology methods. To generate an expression vector of LpxG with a C-terminal His 10 tag that was cleavable by tobacco etch virus (TEV) protease, QuikChange (Stratagene, La Jolla, CA) mutagenesis was used to insert additional nucleotides into pET21b (Novagen) to encode the TEV protease recognition site (ENLYFQG) (27), as well as four additional histidine residues needed to elongate the affinity tag. The resulting plasmid, pHSC, was confirmed by sequencing.
Ct461/lpxG from the original pDONR library lacks 21 residues of the N-terminal transmembrane helix. The gene encoding full-length LpxG was generated by megaprimer PCR to extend the N-terminal missing residues and then cloned into pHSC to yield LpxG followed by the TEV site and His 10 tag (pLpxGt). The LpxG D59A mutant was generated by point mutagenesis (pLpxGt_D59A). The presence of the correct sequence was confirmed by DNA sequencing.
In order to verify the UDP-DAGn hydrolase activity of LpxG, W3110A cells harboring either full-length Ct461 in a pBAD33 plasmid (Invitrogen) or an empty vector were transduced with a P1vir lysate prepared from W3110A⌬HEc, following standard protocols, to replace the chromosomal lpxH with a kan cassette. Colonies containing the ⌬lpxH::kan insertion can only be isolated from the cells harboring lpxG on the plasmid and not from cells containing an empty vector, confirming the functional complementation of LpxH in E. coli by full-length LpxG from C. trachomatis.
Protein expression and purification. Plasmids encoding full-length LpxG (pLpxGt), the LpxG D59A mutant (pLpxGt_D59A), or an empty vector (pET21t10) were transformed into E. coli C41(DE3) cells for protein expression. The cells (LpxG_t10, LpxG D59A _t10, and VC_t10) were grown at 30°C in LB supplemented with Amp until the optical density at 600 nm (OD 600 ) reached 0.7 to 0.8 and then induced with 1 mM IPTG for protein expression for 4 to 5 h. Cells were collected by centrifugation, resuspended in an ice-cold buffer containing 20 mM HEPES, pH 8.0, and passed twice through a French pressure cell (SIM-Aminco; Spectronic Instruments) at 18,000 lb/in 2 . The cell debris was removed by centrifugation at 10,000 ϫ g, and the supernatant was collected as cell-free extracts (CFEs) and stored at Ϫ80°C.
For protein purification, LpxG was extracted from diluted CFE (5 mg/ml in 300 mM NaCl, 10% glycerol, 20 mM HEPES, pH 8.0) with dodecyl-␤-D-maltoside (DDM; Avanti Polar Lipids, Alabaster, AL) at a final detergent concentration of 1% (wt/vol). The detergent-solubilized LpxG was subjected to ultracentrifugation at 100,000 ϫ g for 45 min, and the supernatant was purified by Ni-NTA affinity chromatography in the presence of 0.01% DDM. The final protein concentration ranged from 0.15 to 0.5 mg/ml. UDP-DAGn hydrolase activity assay. Autoradiographic assays for the hydrolase activity were similar to that previously described (5,13), but with slight modifications. The reaction mixtures were a final volume of 12.5 l in 0.6-ml polypropylene tubes and contained 20 mM HEPES, pH 8.0, 0.5% (wt/vol) BSA, 0.035% (wt/vol) DDM, 1 mM MnCl 2 , 100 M UDP-DAGn (prepared as previously described [28]), 1,000 cpm/l [␤-32 P]UDP-DAGn (prepared as previously described [13]), and protein or lysate sample. All reaction mixture components besides the protein sample were mixed to a volume of 10 l and equilibrated at 30°C for 10 min, after which 2.5 l of protein was added to start the reaction. The final protein concentrations in the assays ranged from 0.05 mg/ml to 1.0 mg/ml to maintain linear activity within the time frame being tested. Aliquots were taken from the reaction mixtures at various time intervals and spotted onto glass-backed silica gel thin-layer chromatography (TLC) plates (EMD Chemicals, Darmstadt, Germany). These plates were developed in a chloroform-methanol-water-acetic acid (25:15:4:2) tank system, and the data analyzed using PhosphorImager (GE Healthcare) as previously described (5,13). Young et al. Metal dependence of LpxG. To analyze the metal dependence of LpxG, a modified version of the autoradiographic assay described above was employed. First, a concentrated stock of EDTA was added to a sample of LpxG to obtain a final chelator concentration of 50 M EDTA; the protein was then incubated on ice for 30 min. Next, the sample was diluted 10-fold into various reaction mixtures similar to those described above, except that 1 mM MnCl 2 was replaced with one of the following chloride salts at 1 mM: Ca 2ϩ , Co 2ϩ , Cu 2ϩ , Fe 3ϩ , Mg 2ϩ , Mn 2ϩ , Ni 2ϩ , and Zn 2ϩ . Conditions in which the replacing component was 2 mM NaCl or 1 mM EDTA were also included as controls.
Analysis of LpxG reactions by mass spectrometry. The lipid products of the LpxG reaction were analyzed by mass spectrometry, employing a method similar to that used for analysis of the LpxI reaction products (6 (14). A 10-l aliquot of this material was analyzed by reverse-phase LC-MS using a Shimadzu LC system coupled to a TripleTOF 5600 quadrupole time-of-flight tandem mass spectrometer (AB Sciex, Framingham, MA). The MS instrumental settings for negative ion ESI and tandem MS (MS/MS) analysis of lipid species were as follows: ion spray voltage (IS), Ϫ4,500 V; current gas (CUR), 20 lb/in 2 (pressure); gas 1 (GS1), 20 lb/in 2 ; declustering potential (DP), Ϫ55 V; and focusing potential (FP), Ϫ150 V. Data analysis was performed using Analyst TF1.5 software. LC was operated at a flow rate of 200 l/min with a linear gradient as follows: 100% mobile phase A was held isocratically for 2 min and then linearly increased to 100% mobile phase B over 14 min and held at 100% B for 4 min. Mobile phase A consisted of methanol-acetonitrileaqueous 1 mM ammonium acetate (60:20:20, vol/vol/vol). Mobile phase B consisted of 100% ethanol containing 1 mM ammonium acetate. A Zorbax SB-C 8 reverse-phase column (5-m particle size and dimensions of 2.1 by 50 mm) was obtained from Agilent.
Overexpression of LpxG in C. trachomatis. C. trachomatis lymphogranuloma venereum (LGV) strain L2 was transformed with either pASK-GFP-L2 (29) or derivatives containing LpxG-FLAG or LpxG D59A -FLAG in place of green fluorescent protein (GFP) using previously described transformation methods (30). Vero cells were infected at a multiplicity of infection (MOI) of 3 with the various recombinant Chlamydia strains. Plates were spun at 3,000 rpm for 30 min at 10°C to synchronize infections. In all assays, cells were treated with 2 ng/ml ATc at 20 hpi. Samples were collected for Western blot analysis, lipid extraction, immunofluorescence, and determination of infectious progeny at specified time points.
For immunoblot analysis, cell extracts were normalized for total protein content by the Bradford assay, and equal amounts of protein were resolved by SDS-PAGE and transferred to 0.45-m nitrocellulose membranes. Proteins were detected by incubation of membranes with the antibodies indicated above, followed by fluorescently labeled secondary antibodies. The fluorescence signal was measured using the LI-COR Odyssey imaging system (LI-COR Biosciences).
For the immunofluorescence assay, Vero cells were seeded onto glass coverslips placed in a 24-well plate. At 30 hpi, the coverslips were fixed with 3% formaldehyde-0.025% glutaraldehyde at room temperature for 20 min. The cells were permeabilized with 0.2% Triton X-100 in phosphate-buffered saline (PBS), blocked with 3% bovine serum albumin (BSA) in PBS for 30 min, and stained with antibodies against the inclusion membrane protein IncG (31) and the FLAG epitope (F3165; Sigma). DAPI (4=,6-diamidino-2-phenylindole) was used for staining the nucleus. Confocal images were acquired using a Zeiss LSM 510 inverted confocal microscope.
Mass spectrometry analysis of lipid X of C. trachomatis. For lipid extraction, Vero cells were seeded into 6-well plates, with 4 plates per sample. At 40 hpi, samples were washed briefly with H 2 O and lysed with 200 l H 2 O per well. Each sample was pooled, and bacteria stabilized in a sucrose-phosphate-glutamate (SPG) buffer (8 mM Na 2 HPO 4 , 2 mM NaH 2 PO 4 , 220 mM sucrose, 0.50 mM L-glutamic acid). Samples were stored at Ϫ80°C until analyzed by mass spectrometry. Lipid X was extracted from the C. trachomatis sample by the acidic Bligh-Dyer method as described previously (14,17). The lipid samples were analyzed by normalphase LC-MS. Normal-phase LC was performed on an Agilent 1200 quaternary LC system equipped with an Ascentis silica high-performance liquid chromatography (HPLC) column (5-m particle size and dimensions of 25 cm by 2.1 mm; Sigma-Aldrich, St. Louis, MO). Mobile phase A consisted of chloroform-methanol-aqueous ammonium hydroxide (800: 195:5, vol/vol); mobile phase B consisted of chloroform-methanol-wateraqueous ammonium hydroxide (600:340:50:5, vol/vol); and mobile phase C consisted of chloroform-methanol-water-aqueous ammonium hydroxide (450:450:95:5, vol/vol). The elution program consisted of the following: 100% mobile phase A was held isocratically for 2 min and then linearly increased to 100% mobile phase B over 14 min and held at 100% B for 11 min. The LC gradient was then changed to 100% mobile phase C over 3 min, held at 100% C for 3 min, and finally returned to 100% A over 0.5 min and held at 100% A for 5 min. The LC eluent (with a total flow rate of 300 l/min) was introduced into the ESI source of a high-resolution TripleTOF 5600 mass spectrometer with instrumental settings as described above.
Infectivity of C. trachomatis overexpressing LpxG. Vero cells were seeded in 96-well plates and infected with various C. trachomatis strains. At 40 hpi, cells were subjected to hypotonic lysis by adding 160 l of H 2 O for 10 min at room temperature, followed by the addition of 40 l of 5ϫ SPG buffer and storage at Ϫ80°C. To determine the numbers of infectious particles, serial dilutions of the lysates were used to infect new monolayers of Vero cells. At 32 hpi, cells were washed with PBS and fixed with methanol for 20 min at room temperature. Samples were blocked with 1% BSA in PBS for 1 h and probed with rabbit anti-Slc1 antibodies (30), followed by secondary Alexa Fluor 555-labeled goat anti-rabbit antibodies (Thermo Fisher). Inclusions were counted on a Cellomics ArrayScan high-content imaging system (Thermo Fisher).