Docking Simulations of Serotonin to COMT Active Site
We employ the MedusaDock package to generate possible ligand conformations within the protein active site, and utilize the MedusaScore package to evaluate each conformation generated by MedusaDock [12, 13]. MedusaDock enables flexible docking of both the ligand and the side-chain amino acids of the protein. We perform all ligand docking simulations using the crystal structure of human COMT (PDB: 3BWM) .
Prior to all docking simulations, we first stripped all crystallized ligands bound to COMT. Crystallographic waters that were in the active site were retained for our docking simulations. We performed 200 docking simulations for serotonin, where each simulation began with a different seed number and each conformation generated from a simulation was subsequently minimized. All conformations were then ranked according to their free energy value. The lowest energy structure is determined to be the native pose.
Binding Measurements using Surface Plasmon Resonance
Purified COMT protein (derived from porcine liver) and serotonin ligand were obtained from Sigma-Aldrich. Protein concentration was verified by performing a BCA assay (Thermo Scientific). Purified COMT was biotinylated at its surface lysine residues using EZ-Link NHS-LC-LC Biotin (Thermo Scientific) at a 1:1 mole ratio of biotinylating reagent to protein and subsequently purified using a 1 mL Sephadex G-25 medium spin column. Biotinylated COMT was buffered with 10 mM Tris pH 7.4, 1 mM MgCl2, 1 μM DTT and stored at 4°C until use.
Surface plasmon resonance (SPR) measurements were conducted at 25°C using a Biacore 2000 instrument with a previously established protocol . Biotinylated COMT was first loaded onto a streptavidin-coated flow cell (sensor chip SA or Biotin CAPture Kit, GE Healthcare) at a flow rate of 5 μL/min, followed by a buffer flow, introduction of serotonin (buffered in 10 mM Tris pH 7.4, 1 mM MgCl2, 1 μM DTT), and then a final wash with buffer (10 mM Tris pH 7.4, 1 mM MgCl2, 1 μM DTT). Here the flow rate was established at 5 μL/min.
The dissociation rate of the serotonin-COMT complex was determined by fitting the dissociation data to a single-exponential decay.
Here, the response unit from the SPR machine is denoted as R
. Once the dissociation rate is determined, the association rate of serotonin to COMT can be found by fitting to the following equation:
Radiolabeled SAM Activity Assay
Activity of COMT was measured through a series of experiments where the enzyme was incubated with the substrate catechol (dissolved in H2O, Fisher Scientific) and S-[methyl-3 H]-adenosyl-L-methionine (PerkinElmer, 10.0 Ci/mmol) . Reactions were carried out in a buffer containing 10 mM Tris (pH 7.4), 1 mM MgCl2, and 1 μM DTT with a total reaction volume of 100 μL.
To determine the linear range of the activity assay, SAM concentration (1 μM) was held constant while COMT concentration was varied (10 – 80 ng/μL). Each reaction was placed in a clean PCR tube and incubated at 37°C for 30 minutes. After incubation, reactions were terminated by addition of 100 μL of 1 M HCl to each tube. Radiolabeled catechol products were extracted from the reaction by adding 10 mL of scintillation fluid (MonoFlow I, National Diagnostics) and quantified using a scintillation counter. Activities were normalized for each reaction by performing a duplicate reaction with the COMT inhibitor, OR-486 (0.5 mg of inhibitor dissolved in 30 μL DMSO and 20 μL of 0.9% saline; OR-486 buffer alone does not inhibit reaction), and subtracting its radioactivity. We found that the linear range was between 10 and 40 ng/μL, and thus for all subsequent experiments we used a COMT concentration of 20 ng/μL.
Kinetic experiments were performed at a variety of catechol concentrations (0, 0.5, 1, 3.3, 5, 10, and 33 μM) while maintaining constant SAM (1 μM) and COMT (20 ng/μL) concentrations. Reactions were set-up for a variety of time points: 0, 5, 10, 30, 60, and 120 minutes. Each time-point reaction is incubated at 37°C and then quenched with 1 M HCl at its specified time point. Products were quantified as discussed above. Serotonin-inhibition experiments were conducted for the same catechol concentrations, but with the addition of 100 μM of serotonin (Sigma) to each reaction.
Kinetic parameters were determined by first fitting the data sets to a single-site Michaelis-Menten model (for reference), followed by a two-site Michaelis-Menten model as given by [17
where catechol is abbreviated as “Cat.” To determine the mechanism of serotonin-inhibition, we calculated the apparent K
) of inhibition for both a competitive and non-competitive inhibitor, and compared which K
yielded the lowest value. The apparent K
for a competitive inhibitor is given by
is the apparent K
for the serotonin-inhibited curve. The apparent K
for a noncompetitive inhibitor is given by
where Vmax,app is the apparent Vmax for the serotonin-inhibited curve.
Mouse Behavioral Experiments
Subjects were naïve, young adult (6–14 weeks old) outbred CD-1® (ICR:CrI) mice of both sexes (Charles River, Boucherville, QC). Animals were housed in groups of 2 to 6 with same-sex littermates, with food (Harlan Teklad 8604) and tap water available ad lib. Approximately equal numbers of male and female mice were used; no sex differences were noted so data from both sexes were pooled. All procedures followed the guidelines and regulations of the Canadian Council on Animal Care, and were approved by the McGill Downtown Animal Care and Use Committee.
Mechanical sensitivity was measured using the automated von Frey test. Mice were placed in individual transparent Plexiglas cubicles set atop a perforated metal floor (with 5-mm diameter holes 7 mm apart), and separated from each other by opaque dividers. Prior to testing, the mice were habituated to their surroundings for 1 h. A von Frey fiber with automatically increasing force (Ugo Basile Dynamic Plantar Aesthesiometer) was applied to the mid-plantar hind paw. Three separate withdrawal threshold determinations (on each hind paw) were taken at baseline (pre-injection) and then averaged. One withdrawal threshold determination (on each hind paw) was taken at every post-injection time point.
For the SAM pre-treatment study, after measuring baseline thresholds, mice received i.p. injections of S- (5’-Adenosyl)-L-methionine chloride (80 mg/kg) (Sigma), followed, 15 min later, by 20 μl i.pl. injections of serotonin hydrochloride (10 uM) (Sigma). Mice were tested at 15, 30, 45, 60, 90, 120, 180 and 240 min following the serotonin injection.
For the β-adrenergic antagonist study, mice received 20 μL i.pl. injections of serotonin hydrochloride (5 uM), followed, 15 min later, by i.p. injection of a cocktail of ICI118,551 (5 mg/kg) and SR59230A (50 mg/kg) (Tocris). Mice were tested at 15, 30, 45, 60, 90, 120, 180 and 240 min following the antagonist injection.
Pain hypersensitivity was quantified as the area over the time-threshold curve (post-injection), using the trapezoidal method.
We utilize Generalized Estimating Equations (GEE) to determine whether SAM pretreatment or co-administration of SR and ICI treatment versus serotonin alone resulted in significantly different paw withdrawal thresholds  across the repeated measurements. The response was the repeated measurements of paw withdrawal threshold at fixed intervals post-administration of treatment. This model included an intercept and a binary covariate reflecting treatment status (e.g., ICI118,551 and SR59230A treatment versus serotonin alone) as the only covariate, where parameter estimates for this covariate can be interpreted as the overall mean change over measurements in PWT due to treatment. We assume an AR correlation structure between repeated measurements, gaussian distribution of the data, and the identity link. We tested whether the estimate for our binary covariate for treatment was significantly different from zero (indicating a significant change in paw withdrawal threshold due to treatment effect) using Proc Genmod in SAS version 9.2, and report p-values from this test. We also report the estimated mean change in paw withdrawal threshold due to treatment effect ± the standard error of the mean change.