Organic & Biomolecular Chemistry

IP 3 receptors are channels that mediate the release of Ca 2+ from the intracellular stores of cells stimulated by hormones or neurotransmitters. Adenophostin A (AdA) is the most potent agonist of IP 3 receptors, with the β -anomeric adenine contributing to the increased potency. The potency of AdA and its stability towards the enzymes that degrade IP 3 have aroused interest in AdA analogs for biological studies. The complex structure of AdA poses problems that have necessitated optimization of synthetic conditions for each analog. Such lengthy one-at-a-time syntheses limit access to AdA analogs. We have addressed this problem by synthesizing a library of triazole-based AdA analogs, triazolophostins, by employing click chemistry. An advanced intermediate having all the necessary phosphates and a β -azide at the anomeric position was reacted with various alkynes under Cu( I ) catalysis to yield triazoles, which upon deprotection gave triazolophostins. All eleven triazolophostins synthesized are more potent than IP 3 and some are equipotent with AdA in functional analyses of IP 3 receptors. We show that a triazole ring can replace adenine without compromising the potency of AdA and provide facile routes to novel AdA analogs. ( 31 P coupled, ipso carbons of POCH 2 Ph); 31 P NMR (202.4 MHz, CDCl − 1.49, − 1.920, − 2.092; 19 F NMR (470.68 MHz, CDCl 113.585; Anal. for C 82 H 81 FN 3 O 18 P 3 C, 65.29; H, 5.41; N, 2.79. Found: C, 64.98; H, 5.11; N, 2.93.


I. Introduction
Many biological processes are regulated by changes in the intracellular concentration of Ca 2+ . 1 A major pathway for these Ca 2+ signals is the release of Ca 2+ from intracellular stores within the endoplasmic reticulum (ER) via IP 3 receptors (IP 3 R). 2 IP 3 R are a family of intracellular Ca 2+ channels that are expressed largely in ER membranes. 3 IP 3 (Fig. 1A, 1) is produced when cell-surface receptors stimulate phospholipase C activity. IP 3 then binds to IP 3 R causing its channel to open and release Ca 2+ into the cytosol. 4 The ability of IP 3 to interact with IP 3 R is terminated by its dephosphorylation or phosphorylation by specific 5-phosphatase and 3-kinase enzymes. 5 While terminating the Ca 2+ -mobilizing ability of IP 3 , these steps also initiate recycling of inositol to the lipid from which IP 3 is generated and the synthesis of more complex bioactive phosphoinositols including the pyrophosphates, 6 many of which have been recent synthetic targets. 7 Adenophostin A (AdA), a fungal metabolite isolated from the culture broth of Penicillium brevicompactum, is a more potent agonist than IP 3 of IP 3 R. 8 Moreover, AdA is resistant to the enzymes that metabolize IP 3 . Many analogs of AdA have been synthesized to address structure-activity relationships and the determinants of the increased potency of AdA. 9 These studies revealed that the three phosphates and 2″-OH (color coded in Fig. 1A, 2), which structurally mimic the essential pharmacophore of IP 3 , 10 are essential for AdA activity. The nucleobase/surrogate with β-stereochemistry is indispensible for the enhanced potency. 9b,11 Recent studies have suggested that a cation-π interaction between adenine and a cationic residue (Arg504) of the IP 3 R might be responsible for the increased potency of AdA. 9b,12 Additional base-modified analogs of AdA would be useful in both testing this proposal and as potent and metabolically stable agonists of IP 3 R to probe Ca 2+ signaling mechanisms.
The major hurdle in developing such AdA analogs lies in the synthesis of these structurally complex base-modified analogs (disaccharide nucleotides). 9e,13 The traditional synthesis involves N-glycosylation of a nucleobase with an orthogonally protected disaccharide derivative followed by chemoselective demasking and phosphorylation of specific hydroxyl groups in the nucleoside (Fig. 1B). This early introduction of the nucleobase necessitates an almost individually tailored synthesis (branching at an early stage of synthesis) for each base-modified analog. Further complications arise from the lack of a generally applicable method for glycosylation and phosphorylation of such complex molecules. For instance glycosylation of nucleobases is challenging with respect to yield 14 and regioselectivity, 13b,15 and each glycosylation condition has to be optimized individually. Chemoselective O-phosphorylation in the presence of nucleobases having nucleophilic groups/sites is also challenging. 16 Thus synthesis of analogs by a one-at-a-time method is cumbersome, time-consuming and impractical. We herein report a general strategy for making a combinatorial library of structurally diverse AdA analogs that are potent agonists of IP 3 R.

II. Results and discussion
Imidophostin ( Fig. 2A), though weaker than AdA, is more potent than IP 3 , suggesting that even a 5-membered aromatic heterocyclic ring can be a partial functional mimic of adenine. As triazole is isosteric with imidazole, we envisioned that triazole-based AdA analogs would also be potent agonists of IP 3 R. This and the earlier reports on the use of click reaction in biologically important molecules 17 prompted us to adopt azidealkyne click chemistry to make a library of triazolophostins, AdA analogs wherein the nucleobase is replaced by substituted triazoles. These molecules can easily be prepared by reacting appropriately protected tris-phosphate derivative II having an azide at the anomeric position with aryl or heteroaryl alkynes under Cu(I) catalysis (Fig. 2B). Such an approach would solve three major issues associated with conventional AdA analog synthesis: (i) ensure the necessary β-anomeric stereochemistry of the base-surrogate, (ii) as the triazole-based base-surrogate is introduced only at the penultimate step (after phosphorylation), the interference of nucleobases during phosphorylation can be avoided and (iii) diversification at the penultimate step provides access to many structurally diverse analogs from a common advanced intermediate.
We have synthesized orthogonally protected disaccharide 4 in thirteen steps from D-glucose and D-xylose in a convergent manner as reported previously. 18 Acetolysis 19 of the ketal 4 gave tetraacetate 5 as anomeric mixtures in very good yield. Lewis acid catalysed azidation of the acetate 5 gave anomerically pure β-azide 6. 20 Methanolysis of the acetate groups in 6 provided the triol 7 in excellent yield. Phosphitylation of the triol 7 using phosphoramidite followed by in situ oxidation using m-CPBA provided the fully protected trisphosphate 8. Compound 8 with three phosphates in their appropriate relative positions and β-oriented azide is the common advanced intermediate for coupling with different alkynes. Copper(I) catalysed click reaction of azide 8 with various alkynes provided fully protected triazolophostins 9a-k. As the reaction is catalyzed by copper(I), 1,4-substituted triazoles 21 were obtained exclusively in all the cases. Global removal of the benzyl protecting groups by transfer hydrogenolysis provided triazolo-  phostins 10a-k in very good yields (Scheme 1). Though in a recent elegant communication, Potter et al. reported a functionally active triazole based cADPR analog, 22 ours is the first report on the use of click chemistry to generate a library of IP 3 R agonists.
The abilities of the newly synthesized ligands to evoke Ca 2+ release via IP 3 R were assessed using permeabilized cells that express only type 1 IP 3 R. 23 A low-affinity Ca 2+ indicator trapped within the ER allowed the effects of IP 3 to be directly measured using a fluorescence plate-reader equipped to allow automated additions. All the triazolophostins fully release the IP 3 -sensitive Ca 2+ stores and all are more potent than IP 3 (Table 1, Fig. 3). It is interesting to note that the parent analog triazolophostin 10a having only the five-membered ring is 13-fold more potent than IP 3 . This remarkable AdA-like potency of 10a suggests that a single triazole ring can effectively replace the adenine base in AdA without compromising potency. In contrast, imidophostin was previously reported to be only slightly (1.3-fold) more potent than IP 3 , and it was therefore suggested that a fused bicyclic nucleobase/surrogate is essential for AdA-like potency. 18a Among all triazolophostins, the 3-fluoro derivative 10f is the most potent, and equipotent to AdA: it is the most potent AdA analog without a purine nucleobase. Of the several AdA analogs reported in the literature, very few match the potency of AdA. 9c,12b,13b,14,18a,24,25 In this context, our late-stage diversification method is important because it allows many potent analogs to be made in an easy and combinatorial way from the same advanced intermediate.
The docking study suggests that triazolophostin 10a and AdA exhibit similar interactions with the IP 3 -binding core of IP 3 R1 (IBC, residues 224-604). 24 The triazolyl ring of 10a forms a cation-π interaction with Arg504 similar to the proposed interaction between the adenine of AdA and Arg504. This supports previous reports that a cation-π interaction between the base and IP 3 R contributes to the increased potency of AdA and analogs. 9c,24 All other triazolophostin analogs showed a similar interaction. The small differences in activity among these analogs seem to arise from steric, rather than electronic factors. For instance, attaching aromatic rings with different electrostatic surface potentials to the triazole ring did not cause any systematic change in biological activity. The fact that all the analogs are both more potent than IP 3 , and show a conserved cation-π interaction between the triazole ring and Arg504 suggests that this critical interaction is unaffected by aromatic substitution, irrespective of the electronic nature of the aromatic substituent. The single bond connection between the triazole ring and the substituent (aromatic ring) allow torsional freedom (freely rotatable) and hence the sterically favorable non-coplanar conformation of the two aromatic rings. As a result of such non-coplanarity (lack of conjugation), the electron density on the triazole ring cannot be influenced by the aromatic unit. This could be the reason why there is no clear trend in activity when the aromatic ring is modified with electron-deficient or electron-rich substituents.
One reason for the increased potency of AdA is believed to be its conformational flexibility allowing it to adopt an optimal conformation for binding to IP 3 R. 26 The additional conformational freedom due to the single bond linkage of this hetero-biaryl unit (rather than fused rings in AdA) allows the second ring to orient to an electronically favorable position, such that it can have additional electrostatic interactions. Docking results suggest that the most potent analog, 10f, has additional interactions between F and IP 3 R side-chains (D566, Q565 and Q507) that may enhance its activity (Fig. 4).

III. Conclusions
In conclusion, addressing the problems associated with conventional one-at-a-time synthesis of AdA analogs, we have synthesized a library of new analogs using azide-alkyne click chemistry. The simple reaction conditions, high yield and atom-economy of the click reaction to install base-surrogates are advantageous over the conventional Vorbruggen method of installation of a nucleobase. By this easy and rapid method, we have generated a variety of ligands with diverse structures from a common advanced intermediate. All the analogs synthesized by this method are more potent than IP 3 and some (10a, 10f, 10g, 10k) are as potent as AdA. Though replacement of purines with pyrimidines or other aromatic groups is reported to give agonists that are more potent than IP 3 , none of the previously reported non-purinated AdA analogs was as potent as AdA. We have shown that AdA-like potency can be achieved without a nucleobase. Our study reveals that a five-membered aromatic ring, triazole, is sufficient as a base surrogate to exhibit AdAlike potency. Easy access to potent, non-metabolizable ligands of IP 3 R and the ease of conjugation with reporter groups such as fluorescent probes, photoaffinity probes, etc. offers avenues for advanced biological exploration. This report that triazoles can be functional mimics of nucleobases has broader implication and might trigger such attempts in other nucleobasederived signaling molecules.

General methods
Chromatograms were visualized under UV light and by dipping plates into either phosphomolybdic acid in MeOH or anisaldehyde in ethanol, followed by heating. 1 H NMR, COSY, NOESY and HMQC spectra were recorded on a 500 MHz NMR spectrometer. Proton chemical shifts are reported in ppm (δ) relative to the internal standard tetramethylsilane (TMS, , integration and peak identification). All NMR signals were assigned on the basis of 1 H NMR, 13 C NMR, COSY and HMQC experiments. 13 C spectra were recorded with complete proton decoupling. Carbon chemical shifts are reported in ppm (δ) relative to TMS with the respective solvent resonance as the internal standard. All NMR data were collected at 25°C. The concentration of the compounds for 1 H NMR was 5 mg per 0.5 mL and for 13 C NMR it was 20 mg per 0.5 mL for protected compounds and 5-7 mg per 0.5 mL for final com-pounds in the case of 1 H and 13 C NMR. Each triazolophostin (10a-k) was quantified using the Briggs's/Ames's phosphate assay. Melting point was determined using melting point apparatus and is uncorrected. Flash column chromatography was performed using a silica gel 230-400 mesh. Wherever needed, the reactions were carried out under an argon or nitrogen atmosphere employing oven dried glassware. X-ray intensity data measurements of freshly grown crystals of 4 were carried out at 293-296 K on a Bruker-KAPPA APEX II CCD diffractometer with graphite-monochromatized (MoKα = 0.71073 Å) radiation. The X-ray generator was operated at 50 kV and 30 mA. Data were collected with a scan width of 0.3°a t different settings of φ (0°, 90°and 180°) keeping the sample to detector distance fixed at 40 mm and the detector position (2θ) fixed at 24°. The X-ray data collection was monitored by SMART program. All the data were corrected for Lorentzian, polarization and absorption effects using SAINT and SADABS programs. SHELX-97 was used for structure solution and full matrix least-squares refinement on F 2 . Molecular and packing diagrams were generated using Mercury-3.1. Geometrical calculations were performed using SHELXTL and PLATON.

Synthesis of
To a solution of disaccharide 4 (1.0 g, 1.41 mmol) and Ac 2 O (0.4 mL, 4.24 mmol) in dry DCM (20 mL), 0.05 g of freshly prepared H 2 SO 4 -silica 19 was added. The reaction mixture was stirred at room temperature for 12 h under a nitrogen atmosphere and the reaction was monitored using TLC. When the starting material disappeared, the reaction mixture was quenched by the addition of solid NaHCO 3 . The solid material was filtered off and the organic layer was washed with saturated aqueous NaHCO 3 solution. The organic layer was dried over Na 2 SO 4 and evaporated under reduced pressure. The crude product thus obtained was purified by flash column chromatography using 40% ethyl acetate in petroleum ether as the eluent to obtain the known 18a tetraacetate 5 (0.96 g, 90%) as a sticky mass.

Biological assay
Ca 2+ release from the intracellular stores of saponin-permeabilized DT40 cells expressing only type 1 IP 3 R was measured using a low-affinity Ca 2+ indicator (Mag-fluo-4) as described previously. 23 Briefly, the ER was loaded with Ca 2+ to the steady state for ∼120 s by addition of 1.5 mM MgATP in medium containing p-trifluoromethoxyphenylhydrazone (FCCP) to inhibit mitochondria. IP 3 , AdA or triazolophostins were added with cyclopiazonic acid (10 µM) to inhibit further Ca 2+ uptake. Ca 2+ release was assessed 10-20 s after addition of the analog, and expressed as a fraction of the ATP-dependent Ca 2+ uptake.