Synthesis of Inositol Phosphate-Based Competitive Antagonists of Inositol 1,4,5-Trisphosphate Receptors

Inositol 1,4,5-trisphosphate receptors (IP3Rs) are intracellular Ca(2+) channels that are widely expressed in animal cells, where they mediate the release of Ca(2+) from intracellular stores evoked by extracellular stimuli. A diverse array of synthetic agonists of IP3Rs has defined structure-activity relationships, but existing antagonists have severe limitations. We combined analyses of Ca(2+) release with equilibrium competition binding to IP3R to show that (1,3,4,6)IP4 is a full agonist of IP3R1 with lower affinity than (1,4,5)IP3. Systematic manipulation of this meso-compound via a versatile synthetic scheme provided a family of dimeric analogs of 2-O-butyryl-(1,3,4,6)IP4 and (1,3,4,5,6)IP5 that compete with (1,4,5)IP3 for binding to IP3R without evoking Ca(2+) release. These novel analogs are the first inositol phosphate-based competitive antagonists of IP3Rs with affinities comparable to that of the only commonly used competitive antagonist, heparin, the utility of which is limited by off-target effects.


Introduction
IP 3 Rs are essential links between receptors in the plasma membrane that stimulate phospholipase C and release of Ca 2+ from the endoplasmic reticulum (ER).The resulting cytosolic Ca 2+ signals regulate many diverse cellular processes. 3The three subtypes of IP 3 Rs expressed in vertebrates (IP 3 R1-3) are closely related proteins and they are each regulated by both (1,4,5)IP 3 (1, Fig. 1) and Ca 2+ , but they differ in their sensitivity to other forms of regulation and in their subcellular and tissue distributions. 1 Extensive structure-activity studies, [4][5][6][7][8] reinforced by a highresolution structure of (1,4,5)IP 3 bound to the IP 3 -binding core of IP 3 R1 (Fig. 1A), 9 established that the vicinal 4,5-bisphosphate moiety is essential for (1,4,5)IP 3 binding and the equa-torial 6-hydroxyl and 1-phosphate confer high affinity (Fig. 1B).All high-affinity agonists of IP 3 R have structures equivalent to these substituents.The only endogenous inositol phosphate likely to bind to IP 3 Rs under physiological conditions is (1,4,5)IP 3 , the immediate water-soluble product of phospholipase C-catalyzed hydrolysis of the membrane lipid phosphatidylinositol 4,5-bisphosphate.However, synthetic ligands of IP 3 Rs, including many inositol phosphates 7 and derivatives of adenophostins, [10][11][12] have provided insight into the structural determinants of IP 3 R activation.These ligands include analogs of (1,4,5)IP 3 that are resistant to degradation, 13 fluorescent analogs, 14 partial agonists, 6 and synthetic derivatives of adenophostins. 10 There are, however, no ligands of IP 3 R that distinguish effectively between IP 3 R subtypes, 5,15,16 and the only available antagonists have severe limitations. 17The commonly used antagonists are heparin, 2-aminoethoxydiphenyl borate (2-APB), xestospongins and high concentrations of caffeine.The limitations of these antagonists include off-target effects, notably interactions with other Ca 2+ channels, Ca 2+ pumps, G proteins and other signalling pathways; membraneimpermeability (heparin) and, for xestospongins, an inconsistent history of effectiveness as discussed recently. 17This study was undertaken with the aim of developing more effective antagonists of IP 3 R.
18-21 Some, 19,22 though not all, 21 studies have suggested that (1,3,4,6)IP 4 may be a partial agonist, namely that it less effectively activates IP 3 R than full agonists like (1,4,5)IP 3 .It seems likely that the inverted position of the 2-OH in (1,3,4,6)IP 4 (equivalent to the 3-OH of (1,4,5)IP 3 when the structures are compared in orientations likely to reflect their interactions with IP 3 R, Fig. 1B) is a major determinant of the reduced affinity. 5,20Although (1,3,4,6)IP 4 is produced endogenously from (1,3,4)IP 3 , it is unlikely to attain concentrations that regulate IP 3 Rs. 24Nevertheless, we chose (1,3,4,6)IP 4 to attempt development of novel antagonists of IP 3 R because it and analogs in which its free hydroxyls are modified (3, 4) are meso compounds that make synthesis more straightforward, and we had initially supposed that (1,3,4,6)IP 4 might have reduced efficacy. 19,22We previously reported that dimers of inositol phosphates are high-affinity partial agonists of IP 3 R. 6 We have now developed a family of antagonists of IP 3 Rs (5-12 in Fig. 1C) from the (1,3,4,6)IP 4 backbone by modification of its free hydroxyls and dimerization of the modified structures.Diesteric or dietheric linkages of various sizes (n = 1-3) were chosen for these 5-O-homodimers, which were synthesized by means of a diverse and versatile approach.The most useful of these ligands (8, 10 and 12) bind to IP 3 R1 with an affinity comparable to that of the best available competitive antagonist of IP 3 R, heparin, the utility of which is limited by its off-target effects.

Chemistry
Synthesis of IP 4 s and IP 5 .Phosphates 2-4 were all prepared from myo-inositol (13) (Scheme 1).Thus, tetrasodium (1,3,4,6)IP 4 (2) was synthesized from butanedione-derived acetal 14 25 following a previously published route. 26Pentasodium (1,2,3,4,6)IP 5 (4)  was reached via pentol 15a and pentakis phosphate 15b.Modifications on the perphosphorylation and hydrogenolysis protocols, 26 of an inositol biscyclohexylidene acetal originated synthetic scheme, 27 were applied in order to solely obtain the pentasodium salt.The preparation of butanoate 3 involved a novel approach.Thus, acetal 14 was initially selectively protected at the C-2 position as the PMB ether to yield 16a. Masking of the remaining C-5 hydroxyl as the benzyl ether gave the fully protected derivative 16b, which was very carefully deprotected 28 upon treatment with aqueous DDQ to reach free alcohol 16c.Introduction of the required butyryl group was performed by esterification with butyric anhydride.The resulting ester (16d) was then exposed to aqueous TFA to cleave both acetals, and the corresponding tetraol (17a) was formed quantitatively.Perphosphorylation of crude 17a was accomplished using a 1H-tetrazole solution in acetonitrile and dibenzyl N,Ndiisopropylphosphoramidate at ambient temperature, followed by direct oxidation of the intermediate phosphite with m-chloroperbenzoic acid at low temperature.Finally, the obtained benzyl tetrakisphosphate 17b was subjected to hydrogenolysis in ethanol/water in the presence of Pd/C and sodium bicarbonate (exactly one equivalent per phosphate group) to yield quantitatively the desired tetrasodium salt 3.
Synthesis of dimeric analogs of IP 4 and IP 5 .For the synthesis of dimers 5-12, we envisioned the retrosynthetic analysis depicted in Scheme 2. Dimers 5-12 could be reached from the corresponding polyols 18 applying sequentially perphosphorylation and global deprotection protocols.The key to obtain all these compounds, differentially substituted on C-2, from a common intermediate (19) was to introduce orthogonal protective groups (PG and PG′) at an early stage of the synthesis.In this way, 19 could serve as the sole precursor for both series (2-O-butyrylated and 2-O-phosphorylated derivatives) by selective removal of PG′.Esters and ethers 19 could, in turn, be prepared by dimerization of the corresponding monomers 20 using the appropriate linkers.Since this process involved the relatively hindered secondary alcohols 20, we were keen to explore the feasibility of this approach.Finally, starting from myo-inositol (13) selective introduction of the required protective groups was expected to lead to monomers 20.
Monobenzyl ether 21 (Scheme 3) was recognized as a suitable derivative, appropriately functionalized to play the role of 20.Moreover 21 is easily accessible 25,29,30 from myo-inositol through butanedione bisacetal 14.Direct dimerization of this compound was initially investigated using the Steglich esterification approach 31 and employing malonic (n = 1) and succinic acid (n = 2) as linkers (Scheme 3 and Table S1 in ESI †).However, these apparently simple couplings were found to be complicated, under various reaction conditions tested, by the formation of acetate 22 (in the first case) and N-acylureas 24a-c (in both cases).Thus, for malonic acid reactions, the presence of DMAP seemed to solely favor the decarboxylation process, regardless of the carbodiimide (DCC or DIC) and the solvent used. 32We could not securely determine whether this decarboxylation occurred prior to or after the first esterification.However, in other runs we isolated the N-acylureas 24a and 24b, suggesting that acetate 22 is formed from malonic monoester.Although replacing DMAP with DIPEA eliminated this problem, the only product isolated was N-acylurea 24b, in very low yield, whereas starting material was quantitatively recovered when EDC was used.On the other hand, the reactions performed in the absence of base 33 were productive, yielding the desired dimer (23a) along with the corresponding N-acylurea (24a or 24b).The best results were obtained in the case of the DCC-promoted coupling. 34Surprisingly, applying the same conditions (DCC in Et 2 O) for the coupling of 21 with succinic acid was unsuccessful.In order to reach dimer 23b the presence of DMAP was a crucial factor using either DCC or EDC in CH 2 Cl 2 . 34Again, the reaction with DCC furnished an inseparable mixture of dimer 23b and N-acylurea 24c, which was subsequently resolved upon hydrogenolysis.In contrast to esters 23a,b, the synthesis of dimeric ethers 25a,b was accomplished in a more facile way.Williamson etherifications, through the in situ formed (NaH) sodium alkoxide of 21, were initially attempted in DMF using the required diiodo-or dibromo-alkanes, but with poor results.Replacing halo-electrophiles with the more reactive ditosylates 26 35 and 27 36 and applying a protocol 37 which involved KOH as base and a more polar solvent (DMSO) furnished the desired dimers (25a,b) in a clean way and in good yields. 38ith the key intermediate dimers in our hands, we proceeded to the next steps, which involved installation of the butyryl and phosphate groups.Pd-catalyzed hydrogenolysis of 23a,b and 25a,b led to the corresponding diols 28, which were esterified upon exposure to butyric anhydride to give 29 in very good yields (Scheme 4).
Because both agonists (1 and 2) released the same amount of Ca 2+ at maximally effective concentrations, a comparison of EC 50 and K d values allows the effectiveness with which each promotes opening of the IP 3 R Ca 2+ channel to be determined.A partial agonist needs to occupy more receptors to elicit the same response, which is then reflected in a higher EC 50 /K d ratio (and a lower value for pEC 50 -pK d , where p denotes the negative log). 6(1,3,4,6)IP 4 and (1,4,5)IP 3 did not differ significantly in their pEC 50 -pK d values (Table S2 in ESI †) suggesting that (1,4,5)IP 3 and (1,3,4,6)IP 4 have similar efficacies.We conclude that (1,3,4,6)IP 4 is a full agonist with lower affinity than (1,4,5)IP 3 , in agreement with a previous report, 21 but inconsistent with suggestions that it is a partial agonist. 19 S2. † (4) (Fig. 1B).The analogs retained both the essential pharmacophore (Fig. 1B, blue), and the 5-hydroxyl and 6-phosphate groups [equivalent to the 6-hydroxyl and 1-phosphate of (1,4,5)IP 3 ] that increase binding affinity (Fig. 1B, green).
(K d ∼40 µM) 44 than (1,2,3,4,6)IP 5 (4, K d ∼ 23 µM) and with substantially lower affinity than the dimers of 4. These comparisons are consistent with our observation that 10 µM (1,2,4,5,6)IP 5 had no detectable effect on (1,4,5)IP 3 -evoked Ca 2+ release, 6 whereas the same concentration of 12 caused a 2.8-fold decrease in (1,4,5)IP 3 -sensitivity (not shown).A dimeric benzene with six attached phosphate groups (biphenyl 2,2′,4,4′,5,5′-hexakisphosphate) was recently reported to be a rather high-affinity (K d ∼ 200 nM) antagonist of IP 3 R, but it inhibited IP 3 5-phosphatase with very similar potency. 45Compounds 8, 10 and 12 are the most potent inositol phosphatebased antagonists of IP 3 R so far reported.The affinity of these antagonists for IP 3 R1 (K d 7-8 µM) is comparable to that of heparin (K d ∼ 4 µM), 17 but the new dimeric antagonists are smaller than heparin (M r ∼1200 and ∼5000, respectively), and less likely to interact with as many additional intracellular targets.None of these antagonists is membrane-permeant, but based on the versatility of our synthetic approach, it may be feasible to esterify the phosphate groups of the dimeric antagonists to allow them to cross the plasma membrane and then be de-esterified by endogenous intracellular esterases. 46perimental Chemistry Materials and methods.All commercially available reagentgrade chemicals and solvents were used without further purification.Dry solvents were prepared by literature methods and stored over molecular sieves.Whenever possible, reactions were monitored using commercially available precoated TLC plates (layer thickness 0.25 mm) of Kieselgel 60F 254 .Compounds were visualized by use of a UV lamp and/or phosphomolybdic acid (PMA) or Seebach's stains upon warming.Column chromatography was performed in the usual way using Merck 60 (40-60 mm) silica gel using as eluents the solvents indicated in each case.Yields are reported for isolated compounds with >96% purity, as established by NMR spectroscopy.FTIR spectra were obtained in a Nicolet 6700 spectrometer.NMR spectra were recorded with a 300 MHz Bruker Avancelli spectrometer ( 1 H: 300 MHz, 13 C: 75 MHz, 31 P: 121 MHz) or an Agilent 500/54 spectrometer ( 1 H: 500 MHz, 13 C: 126 MHz, 31 P: 202 MHz) using the deuterated solvent indicated.Chemical shifts are given in parts per million and J values in Hertz using solvent or TMS as an internal reference.Assignments of protons were confirmed based on 2D NMR experiments ( 1 H, 1 H COSY, HSQC, and HMBC, recorded using a standard pulse-program library).High resolution mass spectra (HRMS) were recorded on micrOTOF GC-MS QP 5050 Shimadzu single-quadrupole mass spectrometer.For each known compound 1 H and/or 13 C NMR spectra along with HRMS spectra were used to establish identity.
Esterification of malonic acid with 21.Malonic acid (280 mg, 2.69 mmol), and DCC (4.43 g, 21.5 mmol) were successively added to a solution of alcohol 21 (2.69 g, 5.39 mmol) in dry Et 2 O (50 mL).The resulting slurry was vigorously stirred under an Ar atmosphere at room temperature for 24 h, while the reaction progress was monitored by TLC.Upon completion, the solvent was removed in vacuo.The residue was triturated with Et 2 O and filtered.The solid was further washed with Et 2 O (25 mL) and the filtrates were concentrated in vacuo and the residue was purified with flash column chromatography (hexanes/EtOAc 5 : 1 to 2 : 1) to give 1.78 g (62%) of diester 23a and 510 mg (12%) of ureido derivative 24b.
General procedure A: preparation of 5,5′-ethers 25a,b.Alcohol 21 (1 mmol) was dissolved in a 4 : 1 mixture of toluene and DMSO (2.5 mL), powdered KOH (140 mg, 2.5 mmol) was added and the mixture was warmed up to 55 °C.Then, 26 35 or 27 36 (0.5 mmol) was added in one portion and the resulting slurry was heated at the same temperature for 120 h, while the progress of the reaction was monitored by TLC.Upon completion, the mixture was neutralized with the addition of a saturated aqueous NH 4 Cl solution.Then, water was added to dissolve all solids and the clear solution was extracted with toluene (50 mL) and CH 2 Cl 2 (2 × 50 mL).The combined organic phases were dried over Na 2 SO 4 , and concentrated in vacuo.The residue was purified with flash column chromatography (hexanes/EtOAc 5 : 1 to 1 : 1) to give ethers 25a,b.
General procedure B: preparation of diols 28.10% Pd/C (200 mg) was added to a solution of dibenzyl ether 23 or 25 (1 mmol) in MeOH (60 mL).This mixture was vigorously stirred under H 2 (1 atm) at room temperature for 24 h.Then, it was filtered through a pad of Celite®, which was further washed with MeOH (20 mL), CH 2 Cl 2 (20 mL), and MeOH (20 mL).Diols 28 were found to be sufficiently pure and used in the next steps without any further purification.
General procedure C: preparation of butyrates 29.Dry Et 3 N (0.56 mL, 4 mmol) and DMAP (50 mg, 0.4 mmol) were added to a solution of diol 28 (1 mmol) in dry CH 2 Cl 2 (10 mL) under an Ar atmosphere at room temperature.Butyric anhydride (0.50 mL, 3 mmol) was added and the mixture was stirred at room temperature until the full consumption of starting material (TLC monitoring, about 24 h).The reaction mixture was diluted with CH 2 Cl 2 (20 mL) and successively washed with saturated aqueous sodium bicarbonate solution (3 × 10 mL) and saturated brine (10 mL).The aqueous phase was backextracted with CH 2 Cl 2 (10 mL) and the combined organic phases were dried over Na 2 SO 4 and concentrated in vacuo.The residue was purified with flash column chromatography (hexanes/EtOAc 7 : 1 to 1 : 1) to give butyrates 29.
General procedure D: removal of acetal protecting groups.A 90% aqueous solution of TFA (10 mL) was added dropwsise to a solution of starting acetal (16d or 28 or 29, 1 mmol) in CH 2 Cl 2 (10 mL) at room temperature.The resulting mixture was stirred at the same temperature for 2 h.Then, the volatiles were removed under reduced pressure (40 °C).The residue was successively treated with toluene (10 mL) and absolute EtOH (3 × 10 mL) and each time the solvent was removed under reduced pressure.The resulting polyol was found to be sufficiently pure by NMR and used in the next step without any further purification.
General procedure E: phosphorylation of polyols.A 0.45 M solution of 1H-tetrazole in CH 3 CN (3 equiv.per OH) was added to a flask containing neat starting polyol (15a or 17a or 30, 1 mmol) under an Ar atmosphere at room temperature.Then, dibenzyl N,N-diisopropylphosphoramidite (1.6 equiv.per OH) was added dropwise over a period of 30 min.The resulting mixtrure was stirred for 24 h at room temperature, and an additional amount of the phosphorylating agent was added (0.3 equiv.per OH).After 24 h the reaction mixture was diluted with CH 2 Cl 2 (10 mL) and cooled to −50 °C.A solution of 70% m-CPBA (2.4 equiv.per OH) in CH 2 Cl 2 (1.6 mL per mmol m-CPBA) was added dropwise and the mixture was left to vigorously stir for 5 h at 0 °C.The reaction mixture was further diluted with CH 2 Cl 2 (120 mL) and successively washed with a 10% aqueous solution of sodium sulfite (2 × 150 mL), a saturated aqueous solution of NaHCO 3 (2 × 120 mL), and H 2 O (120 mL).The combined aqueous phases were back-extracted with CH 2 Cl 2 (100 mL).The combined organic phases were washed with saturated brine (120 mL), and dried over Na 2 SO 4 .The solvents were removed under reduced pressure and the residue was purified with flash column chromatography (initially hexanes/EtOAc 2 : 1 to 1 : 2 and then 2-5% CH 3 OH in EtOAc).
General procedure F: final deprotection.The starting benzyl phosphate (15b or 17b or 31, 1 mmol) was dissolved in EtOH (50-70 mL).Deionized H 2 O (50-70 mL) and NaHCO 3 (1 equiv.per phosphate group) were added.Then, 10% Pd/C (1 g) was added to the resulting emulsion and the mixture was vigorously stirred under H 2 (1 atm) at room temperature for the indicating period of time.The reaction progress was monitored by 1 H NMR. Upon completion the catalyst was removed by filtration through an LCR/PTFE hydrophilic membrane (0.5 mm); the membrane was washed with a 1 : 1 mixture of EtOH and deionized H 2 O (3 × 30 mL).The combined filtrates were evaporated under reduced pressure (55 °C), and the resulting residue was dried under high vacuum for 24 h to yield the desired phosphate salt.
Equilibrium binding of 3 H-(1,4,5)IP 3 and competing ligands to IP 3 R1.These assays were performed at 4 °C in 500 µL of CLM containing 1.5 mM MgATP, membranes (∼20 µg protein) prepared from Sf9 cells expressing rat IP 3 R1 (Sf9-IP 3 R1 cells), 3 H-(1,4,5)IP 3 (1.5 nM, 19.3 Ci per mmol, Perkin Elmer, Waltham, MA, USA) and appropriate concentrations of competing ligand.Non-specific binding was determined by addition of 10 µM (1,4,5)IP 3 (Enzo Life Sciences, Exeter, UK).Reactions were terminated after 5 min by centrifugation (20 000g, 5 min, 4 °C).The pellet was washed with 700 µL of CLM, resuspended in 200 µL of CLM, and radioactivity was determined by liquid scintillation counting.Culture of Sf9 cells, infection with baculovirus encoding rat IP 3 R1, and preparation of membranes were as described previously. 49uantification of IP 3 R1 expression by Western blotting, using an anti-peptide antiserum to IP 3 R1 49 was performed as described. 50nalysis.For each individual experiment, concentrationeffect relationships were fitted to a Hill equation using nonlinear curve-fitting (GraphPad Prism, version 5).From each experiment, pEC 50 or pIC 50 [−log of the half-maximally effective (EC 50 ) or inhibitory (IC 50 ) concentration in M], Hill coefficient (n H ), and the maximal response were obtained and then used for statistical analyses.All reported comparisons of ligand potencies rely on comparisons within experiments because EC 50 values for (1,4,5)IP 3 -evoked Ca 2+ release can vary between passages of cells.For convenience, figures illustrating concentration-effect relations show average results from several experiments, but the values ( pEC 50 , etc.) determined from fitting curves to individual experiments were used for statistical analyses.Most statistical comparisons were paired, and used Student's t-test or one-way ANOVA with Tukey's post hoc test as appropriate.P < 0.05 is considered significant.
The dose ratio (DR = EC′ 50 /EC 50 , where EC′ 50 and EC 50 are the EC 50 values for (1,4,5)IP 3 -evoked Ca 2+ release determined in the presence and absence of antagonist, respectively) was used to calculate the apparent affinity (K d ) of IP 3 R1 for antagonists from functional assays: From equilibrium competition binding experiments, the K d of competing ligands was calculated from the concentration (IC 50 ) required to cause 50% displacement of the specifically bound 3 H-(1,4,5)IP 3 :

Fig. 1
Fig. 1 Structures of the ligands used.(A) Key contacts between (1,4,5) IP 3 and residues within the IP 3 -binding core (IBC) of IP 3 R1 (Protein Data Bank 1N4K). 9P4 and P5 form the most extensive interactions (not all are shown) with residues in the β-(in green) and α-domains (in blue) of the IBC, respectively, pulling the two domains towards each other. 23P1 and the 6-hydroxyl enhance affinity through hydrogen bonding with one residue each within the α-domain.(B) (1,4,5)IP 3 (1) showing the essential vicinal (4,5)-bisphosphate (blue) and 6-hydroxyl and 1-phosphate required for high-affinity binding (green).Analogs (2-4) are shown in configurations most likely to reflect their interaction with IP 3 R to allow comparison with (1,4,5)IP 3 .Numbers in brackets refer to equivalent groups of (1,4,5)IP 3 .Differences from (1,4,5)IP 3 are highlighted in red.(C) Structures of the dimeric analogs and the acyl or alkyl linkers used.