Cyclooxygenase-2 catalysis and inhibition in lipid bilayer nanodiscs

Abstract

Cyclooxygenases (COX-1 and COX-2) oxygenate arachidonic acid (AA) to generate prostaglandins. The enzymes associate with one leaflet of the membrane bilayer. We utilized nanodisc technology to investigate the function of human (hu) COX-2 and murine (mu) COX-2 in a lipid bilayer environment. huCOX-2 and muCOX-2 were incorporated into nanodiscs composed of POPC, POPS, DOPC, or DOPS phospholipids. Size-exclusion chromatography and negative stain electron microscopy confirm that a single COX-2 homodimer is incorporated into the nanodisc scaffold. Nanodisc-reconstituted COX-2 exhibited similar kinetic profiles for the oxygenation of AA, eicosapentaenoic acid, and 1-arachidonoyl glycerol compared to those derived using detergent solubilized enzyme. Moreover, changing the phospholipid composition of the nanodisc did not alter the ability of COX-2 to oxygenate AA or to be inhibited by various nonselective NSAIDs or celecoxib. The cyclooxygenase activity of nanodisc-reconstituted COX-2 was reduced by aspirin acetylation and potentiated by the nonsubstrate fatty acid palmitic acid to the same extent as detergent solubilized enzyme, independent of phospholipid composition. The stabilization and maintenance of activity afforded by the incorporation of the enzyme into nanodiscs generates a native-like lipid bilayer environment to pursue studies of COX utilizing solution-based techniques that are otherwise not tractable in the presence of detergents.

Keywords: Arachidonic acid; Aspirin; Cyclooxygenase; Nanodisc; Nonsteroidal anti-inflammatory drugs; Phospholipid.

Figure 1. Biophysical Characterization of POPC:muCOX-2

(A) SDS-PAGE gel depicting the results from the dual affinity purification of nanodisc-reconstituted COX-2. Lane 1, molecular weight marker; lane 2, detergent solubilized COX-2 control; lane 3 MSP1E3D1(+) control; lane 4, crude nanodisc incorporation mixture following detergent removal with BioBeads; lane 5, FLAG column flow through; lanes 6–9, FLAG affinity column wash fractions; lane 10, purified nanodisc-reconstituted COX-2 eluted from IMAC affinity column. (B) SEC Analysis of incorporation: muCOX-2 reconstituted into POPC nanodiscs (solid line); empty POPC nanodiscs (dotted line).

Figure 2. Negative Stain EM Analysis of POPC:muCOX-2

(A) Representative negative-stain EM micrograph of POPC:muCOX-2. (B) 2D k-means class averages of POPC:muCOX-2 particles. 6,069 particles were windowed into 150 × 150 pixel individual images, aligned to each other, and subjected to ten cycles of multireference alignment and k-means classification. The numbers at the bottom right designate the number of particles used in each average and the box size of each image is 44nm × 44nm. The images suggest that the COX-2 homodimer lies at the edge of the nanodisc scaffold. However, this is due to the nature of the sample preparation for negative-stain EM analysis, which flattens the sample prior to analysis. (C) Model depicting muCOX-2 (PDB id 3HS5) bound to one leaflet of the POPC nanodisc. Nanodisc scaffold proteins and POPC lipids are colored green and gray, respectively. The muCOX-2 homodimer is colored blue, with the MBD of each monomer colored red.

Figure 3. Product Formation by Nanodisc-reconstituted and Detergent Solubilized huCOX-2

RP-HPLC product profiles derived from the incubation of [1-¹⁴C]-AA with (A) nanodisc-reconstituted and (B) detergent solubilized huCOX-2. A Waters Symmetry C18 column (4.6×250mm) was eluted at 1 mL/min flow rate with a stepwise gradient of acetonitrile:water:acetic acid 37.5:62.5:0.01 (by vol.) for 15 minutes followed by 70:30:0.01 (by vol.) for 15 minutes, and then with methanol. The peak marked with an asterisk is an artifact from switching from the first solvent to the second solvent.

Figure 4. Comparison of the Cyclooxygenase Activity of Nanodisc-reconstituted huCOX-2 Comprised of Different Phospholipids

Bar graphs depicting the relative activity of huCOX-2 reconstituted in nanodiscs containing different phospholipids (gray bar) compared to detergent solubilized FLAG huCOX-2 (black bar). Cyclooxygenase activity was measured using an oxygen electrode and 100μM AA as the substrate. Normalized V_MAX values were derived from the mean values of three measurements ± SEM and a value of 100% was assigned for the cyclooxygenase activity of detergent solubilized FLAG huCOX-2. Relative activities were calculated as: POPC:huCOX-2, 90%; DOPC:huCOX-2, 102%; DOPS:huCOX-2, 104%; POPS:huCOX-2, 100%.

Figure 5. NSAID Mediated Inhibition of COX-2 in Detergent and Nanodiscs

Bar graphs depicting the ability of NSAIDs to inhibit COX-2 that has been reconstituted into nanodiscs with different phospholipids compared to inhibition of detergent solubilized COX-2 is shown for (A) muCOX-2 and (B) huCOX-2. A value of 100% is assigned for the cyclooxygenase activity of uninhibited FLAG COX-2, with the remaining cyclooxygenase activity after inhibition normalized to the uninhibited enzyme. Error bars represent the standard deviation between triplicate measurements. Abbreviations are as follows: IBP, ibuprofen; Napx, naproxen; FBP, flurbiprofen; Indo, indomethacin; Meclo, meclofenamic acid; CBX, celecoxib.

Figure 6. Aspirin Inhibition of muCOX-2 in Detergent and Nanodiscs

(A) Detergent solubilized muCOX-2 (solid line) and POPC:muCOX-2 (dashed line) were incubated with 500μM aspirin for the indicated times and then assayed for oxygenase activity using an oxygen electrode. Values represent the average of triplicate measurements ± SEM and were corrected for the production of monohydroxy product. Chiral analysis of 15-HETE formed by aspirin-acetylated (B) detergent solubilized and (C) nanodisc-reconstituted huCOX-2. A Chiralpak AD column (4.6×250mm) was eluted at 1 mL/min flow rate with a solvent of hexane/ethanol 100:2 (by vol). 15R-HETE represented greater than 98% of the total products formed for each analysis.

Figure 7. PA Potentiates AA Oxygenation in Nanodiscs

Bar graphs depicting the potentiating affect of AA oxygenation in the presence of PA for detergent solubilized and nanodisc-reconstituted (A) huCOX-2 and (B) muCOX-2. The oxygenation of 5μM AA in the absence of PA is set to 100% (black bar), while the percent increase in the oxygenation of 5μM AA in the presence of 25μM PA is depicted by the gray bars. Values represent the average of triplicate measurements ± SEM.