PMC

Diffraction from mPGES1* crystals grown in 8.8 MAG doped with 1 mol % DOPC in the presence of ligand BI1. (A) Diffraction pattern recorded at beamline PXII, SLS, with a 10-fold attenuated 10 × 15 μm² beam. The sample-to-detector distance was set to 350 mm. (B) Zoomed-in view of the boxed area in (A) showing diffraction spots to 2.08 Å (arrows).

Additive lipid optimization of mPGES1* crystallization in 8.8 MAG at 4 °C with ligand BI2. (A, B) 5 mol % DOPC. (C, D) 0.5 mol % DOPC. (E, F) 1 mol % DOPC. (G, H) 2 mol % DOPC. Precipitant solutions contained 2 mM GSH, 8% (v/v) MPD, 0.4 M KNO₃, 0.1 M lithium citrate, and 70 mM ADA pH 7.1. Images were recorded with (B, D, F, H) and without crossed polarizers (A, C, E, G).

Diffraction from crystals of mPGES1* grown in nitrate-containing conditions. Diffraction pattern recorded with a crystal grown in the absence of ligand (A, B) at beamline 23ID-B, APS, and in the presence of ligand BI1 (C, D) at beamline I24, DLS. Rings indicating resolution (Å) are shown in yellow. Panel B and D are expanded views of boxed areas of panel A and C, showing reflection (arrows) to 7.0 Å and 3.0 Å, respectively. The sharp powder rings at about 3.7 Å originate from mesophase lipid that has crystallized.

Structure comparison of mPGES1* and LTC4S highlighting the marked difference in conformation at the C-terminus. (A) mPGES1* (PDB ID 4BPM, 2.08 Å, this work). (B) LTC4S (PDB ID 2UUH, 2.15 Å). Each enzyme consists of a trimer. Monomers are colored cream, green and brown. The C-terminus in one of the monomers in (A) is color-coded according to conformation type corresponding to the THC designation in mPGES1* as follows: turn (Phe152-Pro154), red; helix (Ala155-Leu162), blue; coil (Gly163-Trp171), magenta. To facilitate comparison, a similar color coding has been applied to the C-terminus of LTC4S. The approximate locations of the endoplasmic reticulum membrane boundaries are marked by horizontal lines (blue, lumen side; red, cytoplasmic side).

Crystal contacts in mPGES1* that involve the C-terminal fusion peptide. (A) Three trimers of mPGES1* are shown for context. Trimers 1 and 2 reside within the same layer in the crystal which is of Type 1. Trimer 3 is in an adjacent layer. (B) An expanded view of the boxed area in (A). All interactions (dashed lines), corresponding to contacts of <4.5 Å, that involve a single C-terminal THC segment (yellow monomer) are identified. The total number of interactions is six. Four originate from Arg166; the other two involve Leu162 and Leu168. Interactions that involve THC exist within and between layers. Distances are shown in Å. TM1 and TM2 refer to transmembrane helices.

Effect of nitrate salts on the crystallization of mPGES1*. mPGES1* crystals grown at 4 °C in the presence of 0.1 M NaNO₃ (A, B), 0.4 M NaNO₃ (C, D), and 0.1 M NH₄NO₃ (E, F). Images were recorded with normal light (A, C, E) and between crossed polarizers (B, D, F) 5 days post setup. The hosting mesophase consisted of 5 mol % DOPC in 8.8 MAG. Ligands BI1 and BI2 were present in (C, D) and (E, F), respectively. The precipitant solutions contained 2 mM GSH, 8% (v/v) MPD, and 70 mM ADA pH 7.1.

Impact of host lipid, pH, and temperature on the crystallization of mPGES1*. (A) 7.8 MAG at 20 °C. (B) 9.7 MAG at 20 °C. (C) 8.8 MAG at 20 °C. (D) 7.9 MAG at 4 °C. (E) 9.7 MAG at 4 °C. (F, G, and H) 8.8 MAG at 4 °C. All host lipids were doped with 5 mol % DOPC. The precipitant solution contained 2 mM GSH, 3.5–8% (v/v) MPD and 70 mM ADA pH 6.5 (A–F), pH 6.7 (G) or pH 7.1 (H). In 8.8 MAG, crystals were smaller, but with more three-dimensional bulk, at pH 7.1 (H) than at pH 6.5 or pH 6.7. Images were recorded 5 days postsetup. Typical crystals in each image are indicated by arrows.

Structure-based rational for design of mPGES1 construct. (A) In two different crystal forms of LTC4S (PDB codes: 2PNO and 2UUH), similar dodecamers are observed. The four trimers, which make up the dodecamer, each sit on the vertex of a tetrahedron and exchange C-terminal extensions with the other three trimers in the dodecahedron. (B) Superposition of the 3.5 Å EM structure of mPGES1 (yellow - PDB code: 3DWW) on a trimer of LTC4S. Note that mPGES1 lacks the C-terminal extension present in LTC4S. (C) The mPGES1* construct used in this study included residues 10–151 of mPGES1 and residues 130–150 of LTC4S (with a Leu to Arg mutation at position 147). A careful comparison suggested this construct might form similar dodecamers to those observed with LTC4S. The construct also includes an N-terminal hexa-His tag and a C-terminal FLAG tag, for purification. (D) The amino acid sequence of the mPGES1* construct is shown.

Image of and first diffraction from mPGES1* crystals. (A) Crystals grown at 4 °C. Compared to crystals at 20 °C, considerably less precipitated protein, which appears as a brownish aggregate (see Figure D), was observed at 4 °C. The precipitant solution included 3% (v/v) hexanediol, 7.8% (v/v) MPD, and 70 mM ADA pH 6.5. 9.9 MAG doped with 5 mol % DOPC was used to form the mesophase. (B) Diffraction from crystals in (A) recorded with the unattenuated, 10 × 10 μm² beam at beamline I24, Diamond Light Source. The sample-to-detector distance was set to 700 mm. The sharp reflections in the ∼20–40 Å region of the diffraction pattern, while sparse, are consistent with scattering from a crystal of a macromolecule and suggest that mPGES1* had been successfully crystallized.

Results of in meso crystallization trials with mPGES1*, illustrating the critical role of the additive lipid, DOPC, in generating the first crystal hits. Neither 9.9 MAG (A) nor 7.8 MAG (B), as host lipids, produced crystals in the absence of DOPC. Combining DOPC with the protein prior to reconstitution into the 9.9 MAG-based mesophase had no effect on crystallization (C). Doping 9.9 MAG with 5 mol % DOPC prior to mesophase formation and reconstitution produced microcrystals (D). An expanded view of the boxed area in (D) is shown in (E). A microcrystal, measuring ∼5 um in maximum dimension, resides at the tip of the arrow. In the absence of GSH, the crystals in (D) and (E) disappeared after 3 to 7 days (F). All trials were conducted at 20 °C using a precipitant solution that included 7.8% (v/v) MPD and 70 mM ADA pH 6.5.

Make up of screens used for optimizing mPGES1* crystallization. All screens contained 2 mM GSH and 70 mM ADA pH 6.5, unless otherwise indicated. (A) The MPD-Fine pH Grid screen. In a 96-well plate, pH was increased in 0.2 pH unit increments from 5.7 in column 1 to 7.9 in column 12. In the same plate, MPD concentration (% (v/v)) was increased from 3.5 in row 1 to 10.5 in row 8 in increments of 1% (v/v). (B) The 48 salts (1–48#) in Hampton Research HR2-245 salt screen kit were added to the basic condition with either 5 or 8% (v/v) MPD to the extent of 0.1 M (wells A1-D12) and 0.4 M (wells E1-H12), resulting two 96-well screens. (C) The 96 additives in the Hampton Research HR2-138 additive screen kit were doped into the basic crystallization condition with 5 or 8% (v/v) MPD to the extent of 10% of their original concentrations, resulting two 96-well screens. (D) The 45 buffers/pH combinations in the Hampton Research HR2-241 buffer/pH screen kit, covering a pH range from 2.2 to 11.0, were added to the basic conditions (5 or 8% (v/v) MPD) to a final concentration of 0.1 M, resulting in a 90-condition screen.

Purification of mPGES1*. (A) Size exclusion chromatographic analysis. V_o and V_t mark the void and total column volumes, respectively. The void volume peak likely represents aggregated protein. The elution volume (V_e) for mPGES1* is 69.1 mL, which corresponds to an apparent molecular weight of 126 kDa. The near Gaussian shaped elution profile is consistent with a protein that is monodisperse. (B) SDS-PAGE analysis of mPGES1* from the mPGES1* peak in the chromatogram in (A) visualized using Coomassie Blue stain. Lanes 1 and 2 represent 4 and 40 μg of protein, respectively. The purpose of the low loading used in Lane 1 was to reveal contaminant bands that might overlap with the major band. Indeed, a very faint band was observed above the main mPGES1* band at ∼19 kDa. High loading in Lane 2 was used to reveal minor contaminants that otherwise might go undetected. The smear seen on either side of the main mPGES1* band in lane 2 is likely due to high levels of DM in the sample loaded on the gel. DM may compete with SDS for binding sites on mPGES1* producing a sample with a protein charge/mass range that shows up as a smear. Molecular weight markers are in lane 3. The sample is estimated to have a purity of >95%. The protein migrates as a monomer at the expected molecular weight of ∼18 kDa.