If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
A novel fluorogenic reporter substrate for 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase gamma-2 (PLCγ2): Application to high-throughput screening for activators to treat Alzheimer's disease
Institute for Drug Discovery, Purdue University, West Lafayette, IN 47907, USAIUSM-Purdue TREAT-AD Center, West Lafayette IN 47907, USADepartment of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN 47907, USA
Institute for Drug Discovery, Purdue University, West Lafayette, IN 47907, USAIUSM-Purdue TREAT-AD Center, West Lafayette IN 47907, USADepartment of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN 47907, USA
A rare coding variant in PLCγ2 (P522R) expressed in microglia induces a mild activation of enzymatic activity when compared to wild-type. This mutation is reported to be protective against the cognitive decline associated with late-onset Alzheimer's disease (LOAD) and therefore, activation of wild-type PLCγ2 has been suggested as a potential therapeutic target for the prevention and treatment of LOAD. Additionally, PLCγ2 has been associated with other diseases such as cancer and some autoimmune disorders where mutations with much greater increases in PLCγ2 activity have been identified. Here, pharmacological inhibition may provide a therapeutic effect. In order to facilitate our investigation of the activity of PLCγ2, we developed an optimized fluorogenic substrate to monitor enzymatic activity in aqueous solution. This was accomplished by first exploring the spectral properties of various “turn-on” fluorophores. The most promising turn-on fluorophore was incorporated into a water-soluble PLCγ2 reporter substrate, which we named C8CF3-coumarin. The ability of PLCγ2 to enzymatically process C8CF3-coumarin was confirmed, and the kinetics of the reaction were determined. Reaction conditions were optimized to identify small molecule activators, and a pilot screen of the Library of Pharmacologically Active Compounds 1280 (LOPAC1280) was performed with the goal of identifying small molecule activators of PLCγ2. The optimized screening conditions allowed identification of potential PLCγ2 activators and inhibitors, thus demonstrating the feasibility of this approach for high-throughput screening.
The scientific community continues to search for preventive and curative Alzheimer's disease (AD) treatments. Genome-wide association studies (GWAS) involving tens of thousands of late-onset Alzheimer's disease (LOAD) patients have identified dozens of gene variants that confer heightened risk or protection against developing LOAD [
PLCγ2 along with the closely related PLCγ1 comprise the two PLCγ isozymes and belong to the larger 13-member phospholipase C (PLC) enzyme family. Both isozymes convert PIP2 (phosphatidylinositol 4,5-bisphosphate, PtdIns(4,5)P2, PI(4,5)P2) into IP3 (1D-Myo-inositol 1,4,5-trisphosphate) and diacylglycerol (DAG). PLCγ2 is mainly expressed in immune cells, microglia, and granule cells of the dentate gyrus with lower levels of expression found on some epithelial cells of the vasculature. PLCγ2 is only sparsely expressed in other brain cells such as astrocytes and oligodendrocytes [
A nonsynonymous mutation in PLCG2 reduces the risk of Alzheimer's disease, dementia with Lewy bodies and frontotemporal dementia, and increases the likelihood of longevity.
]. This mutation was found to confer a protective effect where patients who carried this P522R mutation exhibited a slower rate of cognitive decline than patients who did not [
]. While the exact mechanism of this protective effect of the P522R PLCγ2 mutant is not fully understood, this mutation does induce a slight increase in enzymatic activity [
]. Therefore, a pharmacological intervention that increases the wild-type PLCγ2’s activity may provide a therapeutic strategy to lessen the cognitive decline in Alzheimer's disease patients.
In addition to PLCγ2’s importance in AD, other mutations in the PLCγ2 gene have been implicated in various autoimmune disorders and cancers [
]. For example, hypermorphic PLCγ2 mutants Ser707Tyr, Ala708Pro, Leu845_Leu848del, Leu848Pro, and Met1141Lys have all been shown to cause auto-inflammation and PLCγ2-associated antibody deficiency and immune dysregulation (APLAID) syndrome in humans [
Expanding the clinical features of autoinflammation and phospholipase Cγ2-associated antibody deficiency and immune dysregulation by description of a novel patient.
A novel N-ethyl-N-nitrosourea-induced mutation in phospholipase Cγ2 causes inflammatory arthritis, metabolic defects, and male infertility in vitro in a murine model.
The phospholipase Cγ2 mutants R665W and L845F identified in ibrutinib-resistant chronic lymphocytic leukemia patients are hypersensitive to the Rho GTPase Rac2 protein.
]. As PLCγ2 activity is clearly important for an array of homeostatic and aberrant disease processes, there is a need to identify specific modulators of PLCγ2 activity and assays amenable to discovering these compounds.
Previously, phospholipase C (PLC) enzymatic activity has been monitored using radiolabeled PIP2 [
] have been developed to study PLC activity. For PLCγ enzymes, the water soluble WH-15 primarily measures the intrinsic ability of the enzyme's active site to process a substrate. This is due to the fact that PLCγ enzymes generally reside in a catalytically competent form in a cell, but are physically removed from their physiological substrate, PIP2, which is found in the plasma membrane [
]. Upon activation, PLCγ enzymes undergo a conformational change and relocate to the plasma membrane where they can access their physiological substrate [
For the potential prevention and/or treatment of LOAD, currently, it is unclear how best for a small molecule to modulate the activity of wild-type PLCγ2. For example, is it more advantageous for a molecule to simply increase the intrinsic/basal activity of PLCγ2 for solution substrates? Would this even result in an increase in PLCγ2 activity intracellularly? Alternatively, does a molecule need to directly stimulate a conformational change such that PLCγ2 translocates to the membrane, essentially resulting in a constitutively active enzyme? Or, would this result in too much activation leading to deleterious effects? Another option is a molecule that only binds and increases the activity of PLCγ2 once PLCγ2 has been activated (i.e. moves to the membrane) via normal cellular pathways. Ideally, a multitude of tool molecules that could modulate PLCγ2 through various mechanisms would be highly useful for the research community. Therefore, we endeavored to create a new PLCγ2 substrate based upon WH-15 and an optimized high-throughput screening assay protocol designed specifically to identify PLCγ2 activators that act primarily via increasing the intrinsic enzymatic activity of PLCγ2. To accomplish this, first, the spectral properties of multiple “turn-on” fluorophores were determined. The most promising fluorophore was incorporated into a WH-15-like substrate and an assay protocol designed to identify activators was optimized. Next, the utility of this assay was shown in a small screen of the LOPAC1280 library. However, to identify tool molecules that work through alternative mechanisms, additional in vitro and cellular assays will need to be developed as well.
Materials and methods
Materials
Chemicals for synthesis were purchased from well-known vendors, including ThermoFisher Scientific and MilliporeSigma. WH-15 was purchased from KXT Bio (Durham, NC). All consumables (microtiter plates, pipette tips) and reagents for the assay (Sodium chloride, glycerol, HEPES, potassium chloride, calcium chloride, dithiothreitol and adenosine triphosphate) were purchased from ThermoFisher Scientific (Waltham, MA, USA). EGTA, bovine serum albumin, sodium cholate and LOPAC1280 was procured from MilliporeSigma (Burlington, Massachusetts USA). Stock solutions of reagents were prepared with water and stored at 4 or −20 °C following manufacturer's instructions.
Methods
Synthesis of fluorophore candidates
The synthesis or commercial source of fluorophore candidates and their respective carbamate derivative comparators is described in the Supporting Information.
Characterizations of fluorophores
Stock solutions of fluorophores were prepared at 5 mM in ethanol. The maximal wavelength for excitation and emission was determined at concentrations between 2–200 µM (∼0.1–0.4% v/v ethanol) of fluorophore diluted in 50 mM HEPES buffer (pH 7.4) using a Synergy-neo2 plate reader (Agilent, Santa Clara, CA, USA). Various concentrations of fluorophores and their carbamate derivatives were prepared in 50 mM HEPES buffer (pH 7.4) and their intensity using the corresponding maximum excitation and emission wavelengths were determined.
Synthesis of C8CF3-coumarin
Multi-gram synthesis of known intermediate 3 was mostly accomplished by following the procedures of Huang et al. [
] and references therein. There was one major exception, which owed to a difficulty that arose during the synthesis of the trisubstituted benzene fragment. We therefore synthesized the previously reported intermediate 2 in a new way (Scheme 1), which was inspired by the work of Bjørsvik et al. on vanillin and related compounds [
High selectivity in the oxidation of mandelic acid derivatives and in O-methylation of protocatechualdehyde: new processes for synthesis of vanillin, iso-vanillin, and heliotropin.
]. Thus, 3,4-dihydroxybenzaldehyde (1) was treated with 2.5 eq of NaH to generate a dianion, whose 3-position phenoxide more readily reacted than the 4-position phenoxide with the electrophile, 1-iodooctane. This procedure was found to be more amenable to multi-gram scale synthesis of 2 than the published procedure.
Scheme 1Novel synthesis of intermediate 2. Reagents and conditions: (a) NaH, 1-iodooctane, DMF, THF, 0 °C to rt to 50 °C, 29%.
Once intermediate 3 was in hand, the fluorescent reporter 7-amino-4-trifluoromethyl-coumarin (4) was converted to its isocyanate 5, which was conjugated to 3 via a carbamate linkage to give 6. Next, 6 was globally deprotected using TMSBr, then treated with MeOH, and the obtained product was treated with triethylammonium bicarbonate buffer and purified by reversed phase chromatography to furnish the substrate C8CF3-coumarin in tris•NEt3 salt form (Scheme 2). Chemical characterization of C8CF3-coumarin can be found in SI Figs. 1-5.
Scheme 2Completion of novel substrate C8CF3-coumarin. Reagents and conditions: (a) triphosgene, NEt3, DCE, 90 °C; (b) NEt3, DMAP, 4 Å molecular sieves, DCM, rt, 86%; (c) i. TMSBr, DCM, 5 °C to rt; ii. MeOH, rt; iii. aq TEAB, rt, 71%.
Plasmids containing the DNA sequence for human Phospholipase Cγ1 (PLCγ1) and Phospholipase Cγ2 (PLCγ2) protein (wild-type (WT) and P522R) were synthesized and codon-optimized for insect cell expression by GenScript (Piscataway, NJ). The protein construct consists of an N-terminal hexahistidine tag followed by a short linker peptide (GVDLGT) and a tobacco etch virus (TEV) protease cleavage site (ENLYFQS). As has been reported previously for other PLCγ isoforms [
], a C-terminal truncation was introduced into the PLCγ2 WT and P522R proteins such that the expressed protein corresponds to residues 1-1195. Proteins were produced using baculovirus infection of Sf9 cells, followed by lysis and passage over an IMAC column equilibrated with Ni2+. Protein was eluted from the column and stored in 50 mM Tris pH 8.0, 500 mM NaCl, and 5 % glycerol. As shown in SI Fig. 6, Protein purity was determined to be greater than 90% via SDS-PAGE gel analysis and protein identity was confirmed by Western blot analysis using a monoclonal anti-His antibody (GenScript, Cat. No. A00186). The protein was then aliquoted, flash-frozen in liquid nitrogen, and stored at −80 °C until further use. The PLCγ1 D1165H mutant protein was a kind gift from the Sondek group at the University of North Carolina.
C8CF3-coumarin was dissolved in water at 1 mM and stored at −80 °C. Optimal concentrations of PLCγ2 and C8CF3-coumarin were diluted with HEPES assay buffer (50 mM HEPES (pH 7.4), 70 mM KCl, 3 mM CaCl2, 3 mM EGTA, 2 mM DTT and 0.04 mg/ml of BSA (fatty acid-free)).
PLCγ2 Assay with C8CF3-coumarin
For PLCγ2 assay with C8CF3-coumarin, 0–40 nM of PLCγ2 and 8 µM of C8CF3-coumarin were prepared with HEPES assay buffer. 10 µl of enzyme and 10 µl of substrate were plated in the wells of a black 384-well microtiter plate (20 µl total volume). The final concentrations of PLCγ2 and C8CF3-coumarin were 0–20 nM and 4 µM, respectively. The plate was placed on a Synergy-neo2 plate reader (Agilent, Santa Clara, CA, USA), and fluorescence intensity was recorded (ex: 362 nm; em: 496 nm) every 2 min for 3 h at 25 °C.
ATP Titrations with C8CF3-coumarin in PLCγ2 assay
A serial dilution of ATP (0–100 mM) was prepared in HEPES assay buffer. 1 µl of ATP solution (0–100 mM) and 10 µl PLCγ2 (5 nM) were added to a 384-well plate and incubated for 15 min at room temperature. After incubation, 10 µl of C8CF3-coumarin (8 µM) was added (21 µl total volume). For the PLCγ2 assay with C8CF3-coumarin, final concentrations of ATP, PLCγ2, and C8CF3-coumarin were 0–4.8 mM, 2.4 nM, and 3.8 µM, respectively. The fluorescence intensities from the enzymatic reactions were monitored with ex: 362 nm and em: 496 nm, as described above. The slopes of reaction curves in the initial linear range (30 min) were used for analysis.
Determinations of kinetics parameters
For kinetic studies, 2 nM of PLCγ2 and 10–200 µM of C8CF3-coumarin were prepared with HEPES assay buffer. 10 µl of enzyme and 10 µl of substrate were plated on the wells of a black 384-well microtiter plate and fluorescence intensity was recorded as described above. The final concentrations of PLCγ2 and C8CF3-coumarin were 1 nM and 5–100 µM, respectively. The slopes of reaction curves in the initial linear range (30 min) were used for analysis. For the PLCγ1 assay with C8CF3-coumarin, a similar procedure was followed as mentioned above, with a final concentration of PLCγ1 (2 nM) and C8CF3-coumarin (5–100 µM).
Optimization of assay conditions
To find optimal concentrations of calcium chloride, PLCγ2 (5 nM) and C8CF3-coumarin (8 µM) were diluted with assay buffer (50 mM HEPES (pH 7.4), 70 mM KCl, 3 mM EGTA, 2 mM DTT and 0.04 mg/ml of BSA (fatty acid-free)) containing 0–15 mM of CaCl2. 10 µl of enzyme and 10 µl of substrate were plated into a black 384-well microtiter plate, and fluorescence intensity was recorded as described above. The final concentrations of PLCγ2 and C8CF3-coumarin were 2.5 nM and 4 µM, respectively. The slopes of reaction curves in the initial linear range (30 min) were used for analysis.
To obtain optimal concentrations of enzyme, substrate and BSA, PLCγ2 (0–40 nM) and C8CF3-coumarin (4–16) were diluted with assay buffer (50 mM HEPES (pH 7.4), 70 mM KCl, 3 mM CaCl2, 3 mM EGTA, 2 mM DTT) containing 0.02–0.08 mg/ml of BSA (fatty acid-free). 10 µl of enzyme and 10 µl of substrate were plated into a black 384-well microtiter plate, and their fluorescence intensity was recorded as described above. The final concentrations of PLCγ2 and C8CF3-coumarin were 0–20 nM and 2–8 µM, respectively. The slopes of reaction curves in the initial linear range (30 min) were used for analysis.
High throughput screen of the LOPAC1280
To determine Z’-scores for the assay with C8CF3-coumarin, 0, 5, and 10 nM of PLCγ2 and 8 µM C8CF3-coumarin were prepared with HEPES assay buffer. The final concentration of 0 and 5 nM of PLCγ2 served as controls for inhibition and activation (2x activation) in the assay with C8CF3-coumarin. The prepared enzyme (10 µl per well) and substrate (10 µl per well) were dispensed into the 384-well plate with 1 ml of priming using a liquid dispenser (Mantis, Formulatrix, Bedford, MA, USA). Final concentrations of enzyme and C8CF3-coumarin were 0, 2.5, and 5 nM for PLCγ2 and 4 µM C8CF3-coumarin. The enzymatic reactions were monitored by detecting their fluorescence and the slopes of the reaction curves in the initial linear range (30 min) were used for analysis.
When screening the LOPAC1280, 100 nl of compounds (10 mM in DMSO) or only DMSO were dispensed using an Echo 650 liquid handler (Beckman Coulter, Indianapolis, IN). 10 µl of PLCγ2 and 10 µl of C8CF3-coumarin in assay buffer were dispensed in the wells described above. The final concentration of 0 and 5 nM of PLCγ2 served as controls for inhibition and activation, respectively, in the assay with C8CF3-coumarin. The final concentrations of compounds were 50 µM in assay solution (20 µl) containing 0.5% v/v DMSO. The enzymatic reactions were monitored by detecting their fluorescence and the slopes of reaction curves in the initial linear range (30 min) were used for analysis.
To test the optimal concentration of sodium cholate for C8CF3-coumarin, PLCγ2 and C8CF3-coumarin were prepared in the HEPES assay buffer with 0–1% sodium cholate. 10 µl of PLCγ2 (5 nM) and 10 µl of C8CF3-coumarin (8 µM) were dispensed into the 384 well plates. The final assay mixture contains PLCγ2 2.5 nM and C8CF3-coumarin 4 µM. The enzymatic reactions were monitored by detecting their fluorescence in the plate reader, and the slopes of reaction curves in the initial linear range (30 min) were used for analysis.
The hit compounds were prepared at a stock concentration of 20 mM in DMSO, which were diluted to 2, 0.5, 0.165, 0.055, 0.02 and 0.005 mM with DMSO. Hit compounds (100 nl) with the indicated concentration were incubated with 10 µl of PLCγ2 at room temperature for 15 min. DMSO was used as the control (0 µM). The assay reaction was initiated by adding 10 µl of C8CF3-coumarin. The final assay mixture contains compounds (10, 2.5, 0.825, 0.275, 0.1, 0.025, 0 µM), PLCγ2 (2.5 nM), and C8CF3-coumarin (4 µM). The enzymatic reactions were monitored by detecting their fluorescence and the slopes of reaction curves in the initial linear range (30 min) were used for analysis. The dose-response for the hit compounds was tested as mentioned above in the assay buffer with and without 0.4% w/v sodium cholate.
Optimization of assay conditions for WH-15 with PLCγ2
The WH-15 assay was prepared by modifying the previously report method [
]. For PLCγ2 assay with WH-15, 0–32 nM of PLCγ2 and 80 µM of WH-15 were prepared with HEPES assay buffer with 0.04–0.12 mg/ml of BSA (fatty acid-free). 10 µl of enzyme and 10 µl of substrate were plated in the wells of a black 384-well microtiter plate (20 µl total volume). The final concentrations of PLCγ2 and WH-15 were 0–16 nM and 40 µM, respectively. The plate was placed on a plate reader and fluorescence intensity was recorded (ex: 355 nm; em: 535 nm) every 2 min for 3 h at 25 °C.
XY-69 Assay
Liposomal XY-69 was prepared by modifying previously reported methodologies [
] for optimal monitoring of PLCγ2 activity. Briefly, 28.8 µl of XY-69 (1 mM) in methanol, 712.8 µl of phosphatidylinositol 4,5-bisphosphate (PIP2, 1 mg/ml) in chloroform/methanol/water (20:9:1) and 280 µl 1-stearoyl-2-arachidonoyl-sn-glycero-3-phosphatidylethanolamine (lipid PE, 10 mg/ml) in methanol were added to the glass tube followed by drying under a gentle stream of N2 for 120 min. The mixture was re-dissolved in 1 ml of a 1:1 CHCl3:MeOH solution followed by drying under a gentle stream of N2 for 180 min for film formation. The obtained film was then subjected to high vacuum for at least 120 min to remove all trace of solvents. 3 ml of 20 mM HEPES (pH 7.4) was added to the dried film. The material then was pulsed using a 3 mm sonicator probe tip at a depth of ∼5 mm into the fluid with three cycles of pulse, 5 sec ON (20% output) and 10 seconds OFF under the ice. The lipid mixture then was processed by six cycles of freeze-thaw by submerging the tube for 3 minutes in a -78 °C acetone and dry ice bath followed by thawing for 5 min in a 35 °C water bath. The resulting lipid vesicles (3 ml, lipid PE: 1224 µM, PIP2: 216 µM and XY-69: 9.6 µM) were incubated at room temperature for 30 minutes. A 1/5th equivalent volume (200 µl) of 6X Lipid Vehicle Assay Buffer (LVAB, 120 mM pH 7.4 HEPES, 420 mM KCl, 4.5 mM CaCl2, 18 mM EGTA), 24 µl of 1 M DTT and 24 µl 10 mg/ml BSA were added to the 1 ml lipid vesicle solution (total 1.2 ml) followed by vortexing for 20 sec. The resulting lipid vesicle solution (LVS, 1020 μM Lipid PE, 180 μM PIP2, 8.0 μM XY-69, 20 mM HEPES, 70 mM KCl, 0.75 mM CaCl2, 3 mM EGTA, 8 mM DTT, 0.08 mg/ml BSA) was used for the PLCγ2 enzymatic assay. For enzymatic assay with liposomal XY-69, 1 µM of PLCγ2 WT or P522R mutant (20 mM HEPES, 150 NaCl, 5 %v/v glycerol, 1 mM DTT) was diluted with 1 X LVAB (20 mM HEPES, 70 mM KCl, 0.75 mM CaCl2, 3 mM EGTA, pH 7.4 and 0.08 mg/ml of BSA) to yield a final protein concentration of 2.67 nM. 15 µl of PLCγ2 WT or the P522R mutant were plated in the wells of a black 384-well microtiter plate with 0.5 ml of priming using a liquid dispenser (Mantis, Formulatrix, Bedford, MA, USA). 5 µl of LVS (described above) were added to the wells containing PLCγ2 WT or the P522R mutant (final concentrations: HEPES: 20 mM (pH 7.4), KCl: 70 mM, CaCl2: 0.75 mM, EGTA: 3 mM, lipid PE: 255 μM, PIP2: 45 μM. XY-69: 2 μM, DTT: 2 mM, BSA: 0.08 mg/ml). The plate was placed on a plate reader and its fluorescence intensity (ex: 485 nm; em: 520 nm) was recorded every 2 min for 120 min.
Data analysis
Each experiment was repeated a minimum of three times. The data are represented as the mean ± standard deviation (SD). One-way ANOVA with pairwise post hoc Tukey HSD comparison was used for statistical analysis.
Results and discussion
Selection of optimal substrate fluorophore
At the time of publication, WH-15 appears to no longer be readily available commercially. While WH-15 is a robust PLC substrate, since further experiments with WH-15 would require its synthesis, this afforded the opportunity to potentially improve upon WH-15 by exploring alternative “turn-on” fluorophores. This is due to the fact that the installation of the fluorophore on WH-15 occurs on an advanced intermediate in the synthesis. Hence, various different PLC substrates based upon the WH-15 chemical scaffold could be readily synthesized and tested if a potentially superior “turn-on” fluorophore was identified.
To accomplish this, first, various “turn-on” fluorophore candidates were sought in the literature. These “turn-on” fluorophores are thus named as they exhibit little fluorescence when conjugated to another chemical structure. However, once they are released, a large increase in fluorescence signal is observed. An ideal “turn-on” fluorophore for our application would have 1) a large Stokes shift (i.e. difference between excitation and emission wavelength) to decrease the background noise of fluorescent compounds in screening libraries; 2) a large fluorescence intensity ratio between the “on” and “off” forms to decrease the background fluorescence caused by the substrate itself; 3) a high fluorescence intensity to minimize the amount of substrate needed and increase the signal to noise ratio; 4) a longer (i.e. more red) excitation wavelength to minimize background fluorescence; 5) good solubility in aqueous solutions; and 6) a wide range of fluorophore concentrations that exhibit a linear concentration-dependent change in fluorescence intensity.
Next, we identified “turn-on” fluorophores from the literature, and these were filtered so as to retain only viable candidates for final synthesis of a useful assay substrate. Namely, the physicochemical properties of potential candidates were controlled so as to avoid introducing a significant molecular weight, charge, or lipophilicity increase to the final substrate. Further, synthetic tractability and the presence of an aniline or aniline-like amino handle for chemical conjugation via a carbamate linkage were important criteria. Then, to compare the optical properties of the selected candidates in a standardized format, we prepared a library of six fluorophore candidates (“on” form) along with their carbamate derivative comparators (“off” form). These carbamate comparators were used in place of fully assembled substrates due to the synthetic ease of making the carbamate derivatives, except in the case of WH-15. The compounds were purchased or prepared according to standard literature procedures that are provided in the Supporting Information.
As shown in Table 1, the maximum excitation and emission wavelength was determined for each fluorophore. Next, the brightness of each fluorophore was compared against 6-aminoquinoline, the “turn-on” fluorophore found in WH-15. The fluorescence intensity ratio of the fluorophore was then compared to the corresponding carbamate derivative “off” version of the fluorophore. Using the criteria above, 7-amino-4-trifluoromethyl-coumarin was identified as the most suitable fluorogenic reporter on the basis of its balanced optical property profile, including a relatively long emission wavelength, large Stokes shift, fluorescence intensity of the reporter relative to 6-aminoquinoline, and the fluorescence intensity ratio between the carbamate derivative version and the reporter. Therefore, 7-amino-4-trifluoromethyl-coumarin was chosen as the “turn-on” fluorophore for incorporation in the PLCγ2 substrate.
Table 1Spectrophotometric properties of fluorophore candidates.
Compound #
Structurea
Excitation, nm
Emission, nm (Stokes shift)
Relative fluorescence intensity to 6-aminoquinoline
Unconjugated/conjugated ratiob
7 R=H WH-15
342
530 (188)
1x
50
4 R=H 8 R=Boc
362
496 (134)
236x
26.4
9 R=H 10 R=Boc
330
436 (106)
490x
3.2
11 R=H 12 R=Boc
422
546 (124)
25x
17.8
13 R=H 14 R=Boc
468
572 (104)
3.7x
17.4
15 R=H 16 R=Boc
392
452 (60)
986x
323.8
(a) Chemical names of reporters: 7, quinolin-6-amine; 4, 7-amino-4-(trifluoromethyl) coumarin; 9, 7-amino-4-methylcoumarin; 11, 4-amino-9-(n-propyl)-1,8-naphthalimide; 13, 7-amino-4-nitro-2,1,3-benzoxadiazole; 15, 8-amino-BODIPY; (b) Ratio of fluorescence emission signal of reporter to that of model substrate (or WH-15 in case of 7) at the indicated excitation wavelength.
Following the identification of 7-amino-4-trifluoromethyl-coumarin as a suitable fluorophore reporter, we synthesized the novel substrate named C8CF3-coumarin by linking the fluorophore through a 4-hydroxy benzyl alcohol linker to the inositol phosphate headgroup (Scheme 2). C8CF3-coumarin was prepared following the methods described above and in the Supporting Information. This substrate differs from the earlier reported WH-15 only in the fluorophore reporter (6-aminoquinoline in WH-15). Similar to WH-15, C8CF3-coumarin was expected to undergo a cascade reaction shown in Fig. 1A upon enzymatic action, ultimately releasing the free fluorophore.
Fig. 1Enzymatic reactions of PLCγ2 and C8CF3-coumarin. (A) Scheme of enzymatic reaction of C8CF3-coumarin. (B) Enzymatic reaction profiles of C8CF3-coumarin (4 µM) with various concentrations of PLCγ2. (C) The slope (30 min) from the initial linear range of reaction profiles. The values are shown as the mean ± SD from three different experiments. (D) ATP titrations with PLCγ2 (2.5 nM) and C8CF3-coumarin (4 µM). Results were represented with relative slope of initial linear range (30 min) in reaction profiles (n == 3). (E) Average enzymatic profile from all runs of PLCγ2-WT (2.5 nM), PLCγ2-P522R (2.5 nM) and C8CF3-coumarin (4 µM). The values are shown as the mean ± SD from three different experiments.
With the C8CF3-coumarin substrate synthesized, we first determined whether PLCγ2 could enzymatically process the substrate. Enzymatic processing should result in an increase in fluorescence intensity as cleavage of the C8CF3-coumarin substrate by PLCγ2 results in an intramolecular rearrangement leading to the release of the free, “turned-on” 7-amino-4-trifluoromethyl-coumarin (Fig. 1A). As shown in Fig. 1B and C, when C8CF3-coumarin (4 µM) was incubated with various concentrations of PLCγ2, a dose-dependent increase in fluorescence was observed over time. This result shows that C8CF3-coumarin is indeed reactive with PLCγ2 and as expected, the activity increases with enzyme concentration. To further confirm the enzymatic processing of the C8CF3-coumarin substrate, the known PLCγ inhibitor, adenosine triphosphate (ATP) [
] was added to the PLCγ2/C8CF3-coumarin reaction mixture. As shown in Fig. 1D, ATP was indeed able to inhibit the processing of C8CF3-coumarin with an IC50 of 27.2 µM, further confirming the enzymatic processing of the substrate.
With the enzymatic processing of the C8CF3-coumarin substrate validated, next, the activity of the LOAD-protective P522R PLCγ2 mutant was determined and compared with the activity of an equal concentration of wild-type PLCγ2. This experiment was repeated independently three times with n = 3 wells per for each point. As shown in Fig. 1E and SI 7, the average initial slope (30 min) of the PLCγ2-P522R reaction curve is 1.74 ± 0.08-fold greater than and significantly different than that of the wild-type PLCγ2 reaction curve (p = 0.0014). The activity differences then were tested using WH-15 as the substrate (SI Fig. 8) where a slightly smaller increase was also observed (1.5 ± 0.16). While this result would appear to agree with the intracellular PLCγ2 activity observed in EGF-stimulated cells transfected with the P522R PLCγ2 mutant [
], it cannot be ruled out at this time that the in vitro increase in activity could be due to an effect of differences acquired during the expression/purification of these two proteins, such as differences in the active to inactive ratio, differences in post-translational modifications, etc.
Determination of kinetic parameters for C8CF3-coumarin and PLCγ2
To further characterize C8CF3-coumarin, the kinetic properties of the substrate with PLCγ2 were assessed (Fig. 2A and Table 2). Various concentrations of C8CF3-coumarin (5–100 µM) were tested with 1 nM of PLCγ2 in the presence of BSA (0.04 mg/ml). The slope from the initial linear range (30 min) of the reaction curves were used to determine the rate of release of 7-amino-4-trifluoromethyl-coumarin from the substrate. The kinetic parameters (Km = 28.6 ± 5.6 µM and Vmax of 1.24 ± 0.44 pmol/min/ng) for PLCγ2 and C8CF3-coumarin were similar to those of PLCγ1 and C8CF3-coumarin (Km = 47.1 ± 13.5 µM with a Vmax of 1.50 ± 0.09 pmol/min/ng) as shown in Fig. 2B. Similarly, the efficiency of the enzymatic reactions (kcat/Km) are similar between PLCγ2 and PLCγ1 (6.0 ± 1.1 µM−1min−1 and 4.8 ± 1.4 µM−1min−1, respectively). In addition, these kinetics are similar to those previously found for PLCγ1 with WH-15 and the endogenous substrate PIP2 (Table 2) [
]. Taken together, these results show that C8CF3-coumarin is processed by the PLCγ enzymes similar to previously reported substrates.
Fig. 2Kinetics profiles in enzymatic reactions of PLCγ2 and PLCγ1 with C8CF3-coumarin. (A) PLCγ2 (1 nM) and C8CF3-coumarin (5–100 µM) (B) PLCγ1 (2 nM) and C8CF3-coumarin (5–100 µM). Results were represented with slope of initial linear range (30 min) in reaction profiles. The values are shown as the mean ± SD from four different experiments.
As the ultimate goal for the C8CF3-coumarin substrate is to identify small molecule activators of PLCγ2, we next sought to determine optimal assay conditions to preferentially identify potential activators. Identification of activators via screening is often more difficult than inhibitors. This is due to the fact that activation is often small, thus resulting in only minimal changes to an output signal. As the substrate was already optimized to provide a larger signal due to the use of a brighter fluorophore, we next optimized the buffer components for optimal enzymatic activity.
As PLCγ2 is a calcium-dependent enzyme, we tested the effect of calcium concentration on the enzymatic activity (SI Fig. 9). We also observed a relationship between the concentration of BSA in the buffer and enzymatic activity. Therefore, the BSA concentration was also optimized (SI Fig. 10). To increase the likelihood of identifying small levels of activation, we then chose concentrations of PLCγ2 and substrate that gave a robust signal with minimal deviation that also allowed for determination of an increase in enzymatic activity. The final assay conditions are as follows: 2.5 nM PLCγ2, 4 µM C8CF3-coumarin, 0.04 mg/ml BSA, and 3 mM calcium chloride.
As small molecule activators often only exhibit a small increase in enzymatic activity versus the control enzyme, we sought to improve the deviation of the assay in order to identify such activators. To accomplish this, we optimized the conditions above to be ran in a kinetic fashion. The increase in data collected per well results in smaller deviations between wells when the slope is utilized for analysis instead of a single endpoint value. Utilizing the reaction slope instead of a single endpoint also helps to minimize false hits by negating the impact of test compounds with low levels of background fluorescence. The major disadvantage of using a kinetic assay is that it adds a significant amount of time to complete a screen of a large compound collection. However, such kinetic assays have been used to screen large compound collections to successfully identify small molecule activators previously [
] and most modern HTS facilities are equipped with the automation necessary to carry out kinetic screens efficiently.
With an optimized assay in place, we next compared the activity of the new C8CF3-coumarin to that of WH-15. As shown in SI Fig. 11, under identical conditions, the activity of PLCγ2 was nearly undetectable using WH-15. This is due to the fact that the WH-15 fluorophore has a much lower fluorescence intensity than that of the 7-amino-4-(trifluoromethyl) coumarin. When an attempt was made to optimize the WH-15 assay in the same manner as that of the C8CF3-coumarin substrate (SI Fig. 12), the optimized assay conditions were essentially the same as those initially reported for WH-1524. This resulted in a final optimized substrate concentration of 40 µM for WH-15, compared to only 4 µM for that of C8CF3-coumarin. This 10-fold difference in concentration can greatly increase the number of assays that can be run with the same amount of material when C8CF3-coumarin is used as the substrate. Additionally, the difference in fluorescence intensity potentially allows for smaller activity changes to be detected with the C8CF3-coumarin substrate.
High-throughput screening for PLCγ2 activators using the C8CF3-coumarin substrate
To determine if the optimized C8CF3-coumarin assay was indeed amenable for high throughput screening, a Z’-factor was determined for both an inhibition assay and the more challenging activation assay. Due to a lack of reliable small molecule activators and inhibitors, wells with double the concentration of PLCγ2 (5 nM) or no PLCγ2 were used as the activation and inhibition controls. As shown in Fig. 3, Z’-factors of 0.077 were calculated for activation and 0.727 for inhibition. These values suggest that the optimized assay conditions are feasible for a high throughput screen to identify activators and that inhibitors can also be readily identified.
Fig. 3Profiles of Z’ score in enzymatic reactions of PLCγ2 with C8CF3-coumarin. PLCγ2 – 0 nM (green color), 2.5 nM (blue color) and 5.0 nM (red color) with C8CF3-coumarin (4.0 µM). Results represent the slope of the initial linear range (30 min) in the reaction profiles.
Next, a pilot screen of 1280 compounds (50 µM) in single wells from the Library of Pharmacologically Active Compounds (LOPAC1280) was performed to identify potential compounds that could significantly increase the enzymatic activity of PLCγ2. A fairly high compound concentration of 50 µM was chosen to test the robustness of the assay against compound interference such as background fluorescence. Compounds that exhibited a >1.6-fold increase in activity versus the control were deemed as positive hits. This 1.6-fold cut-off point was chosen as it generally corresponded with the activation observed for the protective P522R PLCγ2 mutant vs. wild-type PLCγ2 using these assay conditions and was also greater than 3 standard deviations from the control mean.
After screening, 36 compounds were identified as primary activator hits from the primary screen. These 36 compounds then were cherry-picked and tested again in triplicate and 20 compounds showed reproducible activation of PLCγ2. One of the most common mechanisms of small molecule “activation” in microtiter plates is where test compounds displace adhered protein from the sides of the well. Enzymes that are in solution are generally able to bind and process substrate more efficiently than those that are adhered to the side of a well. While compound-induced displacement results in apparent activation of the enzyme in vitro, compounds that exert their activity through this mechanism typically do not reproduce activation in a cellular context and are therefore considered false positives.
In an effort to eliminate these false positives, 19 of these reproducible compounds (1 compound was not available commercially and was not tested) were checked for dose response and sensitivity to a mild detergent (n = 4). Most compounds that block the adherence of the enzyme to the walls of the well do not show a dose-dependence. Additionally, these compounds typically lose activity when in the presence of a mild detergent. The addition of a detergent also can eliminate other non-desirable activity such as effects due to compound aggregation.
As shown in Fig. 4, four compounds (Clonixin, Daphnetin, SKF-82958, and 5-nitro-2-(3-phenylpropylamino) benzoic acid) exhibited a dose-response using the original assay conditions. However, when these same compounds were tested in the presence of sodium cholate (0.4% w/v), which is near the critical micelle concentration [
] but does not severely impact enzymatic activity of PLCγ2 (SI Fig. 13), only Clonixin and 5-Nitro-2-(3-phenylpropylamino)benzoic acid showed a slight increase in a dose-dependent manner. This suggests that Daphnetin and SKF-82958 may be exhibiting their activity through an undesirable mechanism.
Fig. 4LOPAC1280 library screening. Dose-response profiles of four compounds with and without sodium cholate (0.4% w/v) in enzymatic reactions of PLCγ2 and C8CF3-coumarin. Results represent the slope of the initial linear range (30 min) in the reaction profiles. The values are shown as the mean ± SD from three different experiments. Each experiment had four well replicates. **P < 0.01, *P < 0.05 represents significant differences versus the control group.
When clonixin and 5-nitro-2-(3-phenylpropylamino) benzoic acid were incubated with the P522R PLCγ2 mutant, a dose-dependent increase in activity was observed that was essentially identical to that of the WT mutant (SI Fig. 14). Importantly, when these compounds were incubated in the presence of the C8CF3-coumarin substrate without enzyme, no change in fluorescence was observed (SI Fig. 14). This result suggests that these compounds’ activity may be independent of the slight increase in activity observed due to the P522R mutation. To further explore the effect of these compounds, their impact on the activity of the closely related homolog PLCγ1 and one of its activating mutants, D1165H, then was explored. As shown in SI Fig. 15, both of these compounds exhibited a similar dose-dependent increase in activity on PLCγ1 as they did for PLCγ2. The increase in activity of the PLCγ1 D1165H activating mutant induced by these compounds was slightly higher than either the WT PLCγ1 or PLCγ2, possibly suggesting that these compounds work in some additive fashion with this activating mutation.
Lastly, these two compounds were tested in the XY-69 liposome assay with PLCγ2. As shown in SI Fig. 16, neither of these compounds had any effect on PLCγ2 activity in this assay. Currently, it is unclear why these two assay types have vastly different results. It could be that these compounds modulate the activity of PLCγ2 while the protein is in solution, but when PLCγ2 undergoes a conformational change to interact with the membrane, the interaction between the protein and compound is lost. Or, perhaps the very modest increase in activity in solution is not enough to significantly impact the protein's activity when it is processing a membrane-bound substrate. In any case, far more detailed studies will need to be conducted with any putative in vitro PLCγ2 activator, including its effects in cellular studies, before a determination of its usefulness as a potential chemical probe to study PLCγ2 activity can be made.
While this screen was specifically optimized to identify activators of PLCγ2, inhibitors could also be readily identified. In fact, 41 compounds showed inhibition between 40-60% and 24 compounds showed inhibition >60% in the primary screen. None of these potential compounds were investigated further. However, the known PLCγ inhibitors edelfosine [
] were identified as potential inhibitor hits from the primary screen.
Taken together, these results indicate that C8CF3-coumarin can be used as a reporter substrate and the optimized assay can be utilized in high throughput screening to monitor the activity of PLCγ enzymes to identify activators (or inhibitors) for drug discovery programs.
Conclusions
High throughput screening is a common approach to identify hit compounds that can be further developed into lead and clinical candidates. In this study, a novel reporter substrate named C8CF3-coumarin was synthesized. C8CF3-coumarin then was characterized, an assay protocol was developed to identify small molecular activators of PLCγ enzymes, and a proof-of-principle high throughput screen was conducted using PLCγ2 and the LOPAC1280 library. These studies showed that C8CF3-coumarin could indeed monitor PLCγ enzyme activity and identify potential small molecule activators from a screen. The identification of such small molecule PLCγ2 activators would be of great utility to the Alzheimer's disease research community by serving as both chemical probes to study the modulation of PLCγ2 and as potential therapeutic agents. Additionally, this same assay could be utilized to screen for PLCγ inhibitors that may have utility in studying and potentially treating oncological and autoimmune diseases.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded in whole with federal funds from the National Institute on Aging, National Institutes of Health under grant number 1U54AG065181.
Conflicts of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Acknowledgements
We would like to thank the Purdue University High-Throughput Chemical Genomics Screening Facility for their assistance. We would also like to thank John Sondek at the University of North Carolina for the kind gift of PLCγ1 D1165H mutant protein.
A nonsynonymous mutation in PLCG2 reduces the risk of Alzheimer's disease, dementia with Lewy bodies and frontotemporal dementia, and increases the likelihood of longevity.
Expanding the clinical features of autoinflammation and phospholipase Cγ2-associated antibody deficiency and immune dysregulation by description of a novel patient.
A novel N-ethyl-N-nitrosourea-induced mutation in phospholipase Cγ2 causes inflammatory arthritis, metabolic defects, and male infertility in vitro in a murine model.
The phospholipase Cγ2 mutants R665W and L845F identified in ibrutinib-resistant chronic lymphocytic leukemia patients are hypersensitive to the Rho GTPase Rac2 protein.
High selectivity in the oxidation of mandelic acid derivatives and in O-methylation of protocatechualdehyde: new processes for synthesis of vanillin, iso-vanillin, and heliotropin.
00024-2&id=sc1.jpg){kind=link}
00024-2&id=sc2.jpg){kind=link}
00024-2&id=gr1.jpg){kind=link}
00024-2&id=gr2.jpg){kind=link}
00024-2&id=gr3.jpg){kind=link}
00024-2&id=gr4.jpg){kind=link}