Large-scale expression, purification, and characterization of an engineered prostacyclin-synthesizing enzyme with therapeutic potential

Abstract

Recently, we reported that a novel hybrid enzyme (TriCat enzyme), engineered by linking human cyclooxygenase-2 (COX-2) with prostacyclin (PGI(2)) synthase (PGIS) together through a transmembrane domain, was able to directly integrate the triple catalytic (TripCat) functions of COX-2 and PGIS and effectively convert arachidonic acid (AA) into the vascular protector, PGI(2) [K.H. Ruan, H. Deng, S.P. So, Biochemistry 45 (2006) 14003-14011]. In order to confirm the important biological activity and evaluate its therapeutic potential, it is critical to characterize the properties of the enzyme using the purified protein. The TriCat enzyme cDNA was subcloned into a baculovirus vector and its protein was expressed in Sf-9 cells in large-scale with a high-yield ( approximately 4% of the total membrane protein), as confirmed by Western blot and protein staining. The Sf-9 cells' membrane fraction, rich in TriCat enzyme, exhibited strong TriCat functions (K(m)=3 microM and K(cat)=100 molecules/min) for the TriCat enzyme and was 3-folds faster in converting AA to PGI(2) than the combination of the individual COX-2 and PGIS. Another superiority of the TriCat enzyme is its dual effect on platelet aggregation: it completely inhibited platelet aggregation at the low concentration of 2 microg/ml and then displayed the ability to reverse the initially aggregated platelets to their non-aggregated state. Furthermore, multiple substrate-binding sites were confirmed in the single protein by high-resolution NMR spectroscopy, using partially purified TriCat enzyme. These studies have clearly demonstrated that the isolated TriCat enzyme protein functions in the selective biosynthesis of the vascular protector, PGI(2), and revealed its potential for anti-thrombosis therapeutics.

Figures

Fig. 1

Engineering the TriCat enzyme by linking the C-terminal position of COX-2 to the N-terminal position of PGIS through a transmembrane (TM) linker of 10 amino acid residues. The TriCat enzyme’s direct conversion of AA to PGI2, through its triple catalytic activities, is indicated by the pathway shown.

Fig. 2

(A) TriCat enzyme expression in Sf-9 cells using a BV expression system. The membrane proteins (20 µg) of ten Sf-9 cell plaques (A–J) transfected with the BV vector containing the TriCat enzyme’s cDNA were separated by 7% SDS–PAGE and then Western blot analysis was performed using either (I) anti-PGIS peptide antibody or (II) anti-COX-2 peptide antibody as described [13]. (B) Activity assays for the individual plaques shown in (A) were performed. 20 µg of the membrane protein were incubated with [14C]-AA (10 µM) and the supernatant was analyzed by HPLC-scintillation analysis as described in the Experimental Procedures. The amount of [14C]-6-keto-PGF1α (degraded [14C]-PGI2) produced by each individual plaque of Sf-9 cells was plotted. (C) Western blot analysis for the ‘H’ plaque of the Sf-9 cells expressing the 6His-TriCat enzyme at different days after the transfection. (D and E) Analysis of 6His-TriCat enzyme expression by electrophoresis (Coomassie Blue staining, (A) and Western blot analysis with anti-PGIS peptide antibody (B) using the Sf-9 cells from plaque ‘H’.

Fig. 3

Triple catalytic activity assay for the plaque ‘H’ (upper panel) and the control (transfected with vector alone, lower panel) Sf-9 cells. The crude membrane protein (100 µg) of the Sf-9 cells were incubated with [14C]-AA (10 µM) and then the supernatant was analyzed by HPLC-scintillation analysis as described in Fig. 2. The amount of [14C]-6-keto-PGF1α (degraded [14C]-PGI2) as well as the non-specific side-product, [14C]-HHT, and the unmetabolized [14C]-AA are all indicated in the panels.

Fig. 4

Comparison of the 6His-TriCat enzyme expressed in the Sf-9 cells (1), the wild type TriCat enzyme expressed in the mammalian cell lines, HEK293 (2) and COS-7 (3), and the pcDNA vector (4) expressed in HEK293 cells using activity assays (A) and Western blot analysis (B) as described in Fig. 2.

Fig. 5

HPLC analyses of proteins expressed with different expression systems. The assays were performed under the same conditions as in Fig. 3. [14C]-AA was added to: (A) HEK293 cells expressing the COX-2 enzyme, (B) Sf-9 cells containing BV-expressed COX-2-10aa-PGIS, or (C) HEK293 cells co-expressing the COX-2 and PGIS enzymes. The major products and by products have also been indicated. Also, the percentage of degraded [14C]-AA has been determined and is shown as the circled by-products in B and C.

Fig. 6

Comparison of the kinetic studies for the COX-2-10aa-PGIS prepared from Sf-9 cells (A and C) and the mixture of the individual COX-2 and PGIS (B) co-expressed in HEK293 cells [13]. Different concentrations of [14C]-AA (0.3–60 µM) were added to either the Sf-9 cells or the HEK293 cells in a total reaction volume of 30–100 µL. After a 0.5–5 min incubation, the reaction was terminated and the supernatant was analyzed by HPLC-scintillation analysis as described in the Experimental Procedures. The results were plotted as the average of multiple experiments, where n = 3.

Fig. 7

(I) Effects of the TriCat enzyme on anti-platelet aggregation. The platelet-rich plasma was incubated with 100 µM of AA at 37 °C in the presence of the TriCat enzyme [10 µg, (A)], COX-2 mixed with PGIS [10 µg, (B)] Sf-9 cells only (C), or PBS (D). The addition of the AA to the platelet cells is indicated with an arrow. (II) Dose-dependent inhibition of platelet aggregation by the TriCat enzyme. The platelet aggregation experiments were performed with the addition of increasing amounts of the 6His-COX-2-10aa-PGIS (■) or Sf-9 cells only (●) as described in the Experimental Procedures.

Fig. 8

(I) Comparison of the effects of anti-platelet aggregation for the BV-expressed COX-2-10aa-PGIS protein (A), the mixture of COX-2 and PGIS expressed in HEK293 cells (B), aspirin (1.3 mg/mL, (C) and untransfected Sf-9 cell protein (D). (II) Comparison of the metabolized AA in the platelet-rich plasma in the absence (A), and presence of the membrane-bound 6His-COX-2-10aa-PGIS (B) or the mixture of the individual COX-2 and PGIS (C). The platelet-rich plasma was incubated with PBS, (A), 10 µg of membrane-bound 6His-COX-2-10aa-PGIS (B) or the mixture of COX-2 and PGIS [10 µg total, (C] and then incubated with AA (2 mM) mixed with [14C]-AA (10 µM) for 10 min. The supernatant was collected by centrifugation and then profiled by the HPLC-scintillation analyzer (Fig. 3).

Fig. 9

(A) PAGE analysis of the partially purified 6-His-TriCat enzyme. The following samples: Sf-9 cells expressing the 6-His-TriCat enzyme [crude Sf-9 cell membrane (50 µg, Lane 1)] and the protein extracted by 1% octyl glucoside (50 µg, Lane 2), and then separated by ultracentrifugation (10 µg, Lane 3) and further purified by gel filtration (2 µg, Lane 4), were separated by 7% SDS-PAGE and stained by Coomassie blue. The molecular mass of the 6-His-TriCat enzyme is indicated on the right. (B) Determination of the trip-cat activities for the partially purified 6-His-TriCat enzyme in the conversion of AA to PGI2. 5 µg of the purified 6-His TriCat enzyme was incubated with [14C]-AA and the products were analyzed by the HPLC-scintillation analyzer as described in Fig. 2 and Fig. 3. The 6-keto-PGF1α (degraded [14C]-PGI2) has been labeled.

Fig. 10

Phase changes and intramolecular NOEs of ibuprofen (A and B) and U46619 (C and D) induced by COX-2-10aa-PGIS. Green and grey color peaks represent the negative phase (for free ligand) and positive phase (for the ligands bound to protein), respectively. The new intramolecular NOEs of ibuprofen (B) and U46619 (D) induced by the 6-His-TriCat enzyme in comparison with its free state (A and C) are boxed and labeled (B and D).