The intracellular protozoan parasite Toxoplasma gondii must scavenge cholesterol and other lipids from the host to facilitate intracellular growth and replication. Enzymes responsible for neutral lipid synthesis have been identified but there is no evidence for enzymes that catalyze lipolysis of cholesterol esters and esterified lipids. Here we characterize several T. gondii serine hydrolases with esterase and thioesterase activities that were previously thought to be depalmitoylating enzymes. We find they do not cleave palmitoyl thiol esters but rather hydrolyze short chain lipid esters. Deletion of one of the hydrolases results in alterations in levels of multiple lipids species. We also identify small molecule inhibitors of these hydrolases and show that treatment of parasites results in phenotypic defects reminiscent of parasites exposed to excess cholesterol or oleic acid. Together, these data characterize enzymes necessary for processing lipids critical for infection and highlight the potential for targeting parasite hydrolases for therapeutic applications.

Sample Preparation for Lipidomics. For lipid analysis, 5 T25 flask of confluent monolayer of HFF were infected with parasites and grown for 24 hrs. Parasites were subsequently collected and resuspended in 1.5 ml of 1xPBS buffer and transferred into glass vials pre-loaded with 2:1 chloroform/methanol/1XPBS to a final ratio of 2:1:1 chloroform/methanol/1xPBS.. Samples were vigorously shaken and lipid extraction was performed explained above. Five vials of lipids extracted from each strain (wild type and ΔTgASh4 parasites) were dried under nitrogen gas flow. Lipids were temporarily stored on dry ice before being transferred to a -80 freezer awaiting analysis on Q-TOF. 

Lipid analysis using high-performance liquid chromatography-mass spectrometry. Mass spectrometry analysis was performed with an electrospray ionization source on an Agilent 6545 Q-TOF LC/MS in positive and negative ionization modes. For Q-TOF acquisition parameters, the mass range was set from 100 to 1200 m/z with an acquisition rate of 10 spectra/second and time of 100 ms/spectrum. For Dual AJS ESI source parameters, the drying gas temperature was set to 250°C with a flow rate of 12 l/min, and the nebulizer pressure was 20 psi. The sheath gas temperature was set to 300°C with a flow rate of 12 l/min. The capillary voltage was set to 3500 V and the fragmentor voltage was set to 100 V. For separation of nonpolar metabolites, reversed-phase chromatography was performed with a Luna 5 mm C5 100 Å LC column (Phenomenex cat # 00B-4043-E0). Samples were injected at 20 ul each. Mobile phases for positive ionization mode acquisition were as follows: Buffer A, 95:5 water/methanol with 0.1% formic acid; Buffer B, 60:35:5 isopropanol/methanol/water with 0.1% formic acid. Mobile phases for negative ionization mode acquisition were as follows: Buffer A, 95:5 water/methanol with 0.1% ammonium hydroxide; Buffer B, 60:35:5 isopropanol/methanol/water with 0.1% ammonium hydroxide. All solvents were HPLC-grade. The flow rate for each run started with 0.5 minutes 95% A / 5% B at 0.6 ml/min, followed by a gradient starting at 95% A / 5% B changing linearly to 5% A / 95% B at 0.6 ml/min over the course of 19.5 minutes, followed by a hold at 5% A / 95% B at 0.6 ml/min for 8 minutes and a final 2 minute at 95% A / 5% B at 0.6 ml/min. Raw files were converted to mzXML format with MSConvert (ProteoWizard) using the Peak Picking Vendor algorithm. Files were analyzed using the web-based XCMS platform (Tautenhahn et al., 2012) with the following settings: signal to noise threshold, 6; maximum tolerated m/z deviation, 30 ppm; frame width for overlapping peaks, 0.01; and peak width, 10-60 s. Integrated peak intensities were normalized between conditions by median fold change. Ions were matched to the METLIN database with a 10 ppm tolerance for mass error. The tables containing potential metabolite annotations for each mass feature were downloaded from the XCMS platform. In order to analyze the metabolites according to their chemical classes, metabolite names were transformed into universally readable chemical identifiers: Using python 3 and the pubchempy wrapper for the PubChem PUG REST API  (Fahy et al., 2007) metabolite names were searched against the PubChem database (Kim et al., 2018) and the PubChem CID and SMILES data were matched to the respective mass features. If a given mass feature could correspond to numerous metabolites, all potential metabolites were considered.

For annotation and classification of lipids, the PubChem CID’s were next searched against the LIPID MAPS^®Structure Database (LMSD) using their LIPID MAPS REST service using python 3 (Fahy et al., 2007). Finally, the resulting lipid annotations were matched to the original XCMS metabolomics data. Lipid annotations and their fold changes were plotted after grouping by “core” or “main class.” Plots were produced using python 3 with pandas, matplotlib, and seaborn packages. Boxes indicate the first, second (median), and third quartile ranges for fold change, while whiskers extend 1.5 times past the first or third quartile ranges. Values lying outside these ranges are indicated by circles.

Naming Scheme:

[Experiment number]_[Name]_[Date]_[ion mode]_[run length]_[wild-type (WT) or mutant (ash4)]_[replicate number].

Our published analysis compared negative and positive mode runs for lipid samples obtained from WT or delta_ash4 knockout parasites (n=5 from each strain).