Convergent Synthesis of the Potent P2Y Receptor Antagonist MG 50-3-1 Based on a Regioselective Ullmann Coupling Reaction

MG 50-3-1 (3, trisodium 1-amino-4-{4-[4-chloro-6-(2-sulfophenylamino)-1,3,5-triazin-2-ylamino]-2-sulfophenylamino}-9,10-dioxo-9,10-dihydroanthracene 2-sulfonate) is the most potent and selective antagonist (IC50 4.6 nM) for “P2Y1-like” nucleotide-activated membrane receptors in guinea-pig taenia coli responsible for smooth muscle relaxation. Full characterization of the compound, however, e.g., at the human P2Y1 receptor, which is a novel potential target for antithrombotic drugs, as well as other P2 receptor subtypes, has been hampered due to difficulties in synthesizing the compound in sufficient quantity. MG 50-3-1 would be highly useful as a biological tool for detailed investigation of signal transduction in the gut. We have now developed a convenient, fast, mild, and efficient convergent synthesis of 3 based on retrosynthetic analysis. A new, regioselective Ullmann coupling reaction under microwave irradiation was successfully developed to obtain 1-amino-4-(4-amino-2-sulfophenylamino)-9,10-dioxo-9,10-dihydroanthracene 2-sulfonate (8). Four different copper catalysts (Cu, CuCl, CuCl2, and CuSO4) were investigated at different pH values of sodium phosphate buffer, and in water in the absence or presence of base. Results showed that CuSO4 in water in the presence of triethylamine provided the best conditions for the regioselective Ullmann coupling reaction yielding the key intermediate compound 8. A new synthon (sodium 2-(4,6-dichloro-1,3,5-triazin-2-ylamino)benzenesulfonate, 13) which can easily be obtained on a gram scale was prepared, and 13 was successfully coupled with 8 yielding the target compound 3.


Introduction
Extracellular nucleotides such as ATP, ADP, UTP or UDP are the physiological ligands for P2 purinergic receptors, which are localized in the cell membrane [1][2][3][4]. They are subdivided into G protein-coupled or metabotropic P2Y receptors (GPCRs), and ligand-gated ion channels (LGICs) or ionotropic receptors, termed P2X. Both families, GPCRs as well as LGICs constitute important novel drug targets [1][2][3][4]. In recent years, it was estimated that up to 50% of available drugs act via stimulation or blocking of GPCRs [5]. The P2Y 1 and the P2Y 12 receptor, for example, both of which are activated by ADP, are involved in platelet activation and aggregation [6][7][8][9]. The P2Y 12 receptor is already an established target of antithrombotic drugs like clopidogrel (Plavix ® ) and ticagrelor (Brilinta ® ) [10][11][12]. P2Y 1 antagonists may also have potential as novel antithrombotic drugs [6,7,13,14]. "P2Y 1 -like" receptors have been described in a number of early publications in the field to mediate smooth muscle relaxation in the intestine. Later on it was found that the intestinal preparations expressed both, P2Y 1 and P2Y 11 receptors [15].
The anthraquinone derivatives Reactive Blue 2 (RB-2, 1, Figure 1) and Cibacron Blue 3GA (2, Figure 1) have been identified as new lead compounds in drug discovery [16][17][18][19]. They have been found to interact with a variety of nucleotide-binding proteins in the human body [17,19], including a number of different P2Y and P2X receptor subtypes as well as ecto-nucleotidases [17,19,20]. MG 50-3-1 (3, Figure 1) is an isomer of compound 2, in which the sulfonate group in ring D is shifted from the meta-to the ortho-position with regard to the aminoanthraquinone substituent. Compound 3 has been described to be the most potent and selective antagonist for "P2Y 1 -like" receptors expressed in guinea-pig taenia coli, exhibiting an IC 50 value of 4.6 nM [21]. Full characterization of the compound, however, has been hampered due to difficulties in synthesizing the compound in sufficient amounts, therefore, 3 could not be investigated so far at human P2Y 1 receptors and at the other P2Y receptor subtypes. The compound could be a very useful pharmacological tool for basic research and target validation. The described linear synthesis, however, was reported to provide only 5.5% overall yield [21]. In our hands, the published approach was not successful at all. Therefore we developed a novel strategy for the preparation of the target compound 3. Herein we describe a convergent access towards compound 3 which will eventually allow full characterization and in vitro as well as in vivo evaluation of the drug, and thus may provide information valuable for the development of antagonists for P2Y 1 and P2Y 1 -like receptors. Furthermore, the compound will be a useful biological tool for investigating purinergic signalling, for example in the intestine.

Results and Discussion
Previous studies showed that the substitution pattern in the 4-position of the anthraquinone moiety plays a crucial role for the ability of the compounds to antagonize P2Y receptor subtypes, such as P2X1 and P2Y 1 -like [21], P2X2 [22], P2Y 2 [23], and P2Y 12 receptors [24,25] and to inhibit nucleoside triphosphate diphosphohydrolase (NTPDase) isoenzymes [26] and ecto-5'-nucleotidase [27]. Recently we developed a microwave-assisted Ullmann coupling reaction of bromaminic acid with a diverse range of aniline derivatives in the presence of elemental copper (Cu 0 ) in sodium phosphate buffer [28,29]. In the present study we examined the impact of the buffer pH, and the use of different copper catalysts at different pH values on the described microwave-assisted Ullmann coupling reaction. We were especially interested in the question of how regioselectivity could be achieved in the presence of two nonequivalent amino groups on the aromatic system. This is an important and challenging task, especially in case of the coupling reaction of bromaminic acid (4) with 2,5-diaminobenzenesulfonic acid (5) to yield 8, which represents a key step in the synthesis of MG 50-3-1 (3) [21] with typically low yield (10%) [21]. For direct comparison of the developed reaction we examined the coupling of bromaminic acid (4) with the isomeric 2,4-diaminobenzenesulfonic acid (6).

Optimization of the Ullmann Coupling Reaction of Bromaminic Acid with Aniline
In order to systematically optimize the microwave-catalyzed Ullmann coupling reaction [28] of bromaminic acid with anilines, we initially investigated the effects of the sodium buffer pH in the presence of four different copper catalysts having three different oxidation states (0, I and II) in a model reaction, namely the coupling reaction of bromaminic acid sodium salt (4) with aniline yielding Acid Blue 25 (AB-25, 7) as outlined in Table 1. It should be noted that the pH values were measured at the start of the reaction at 23 °C (see Tables 1, 2 and 4) as the reaction mixtures turned acidic during the course of the reactions due to the formation of hydrogen bromide.   Tables 2 and/or 4 for final pH values. c Conversion was estimated by RP-TLC using a mixture of acetone/water (1:4) as eluent; this was possible because all components (starting material and product) have different colors, bromaminic acid is red, while AB-25 is blue and the by-product is dark-red to violet.
Elemental copper (Cu) and copper(I) chloride (CuCl) gave almost the same results: they differed only in two cases, when water (pH 7, entry 1) was used as a solvent, or in acidic buffer (NaH 2 PO 4 , pH 4.8, entry 2), Cu being superior in both cases. The reaction occurred in the presence of Cu within 20-25 min with ca. 50% conversion. In the case of CuCl no conversion at all was observed in water or acidic media (pH 7 and 4.8, entry 1 and 2, respectively, Table 1), even when the mixture was harshly irradiated in the microwave oven for 150 min at 120 °C.
However, in the presence of different mixtures of phosphate buffer (entry 3-7, Table 1, neutral to basic pH values) the reaction went to completion within only 5 min (100% conversion). Next we examined the effect of the oxidation state II, represented by two different catalysts, copper(II) chloride (CuCl 2 ) and copper(II) sulfate (CuSO 4 ). Although both compounds are water-soluble salts and are therefore homogeneously distributed in the solution, they catalyzed the Ullmann coupling reaction less efficiently than the water-insoluble catalysts (Cu and CuCl), indicating that the Ullmann reaction can be catalyzed heterogeneously (Table 1). When water or acidic buffer were applied in the presence of a Cu (II) catalyst (entry 1 and 2, Table 1) product was not detectable on an RP-TLC plate, even when harsh conditions using the microwave oven at 120 °C for 2.5 h (150 min) were applied. At more alkaline pH values (Table 1, entry 3), ca. 50% conversion was achieved (entry 3, Table 1), but the reaction took more than 2 h (125 min) in the microwave oven. Further increases in the basicity of the sodium phosphate buffer (entries 4-7, Table 1) eventually resulted in 100% conversion also for the Cu(II) catalysts, but required longer reaction times in comparison to the reactions catalyzed by Cu or CuCl.
Mechanistic studies of Ullmann reactions have previously identified Cu(I) as the single active catalytic species under various conditions [30]. However, the formation of an organometallic intermediate from elemental copper is also conceivable [31]. In a recent study we found elemental copper (Cu) to be superior to CuCl and CuSO 4 in three different protocols in the presence or absence of microwave irradiation [28]. Our current-extended-experimental results using a simple model reaction indicate and confirm [28] that Cu (0) appears to be the catalytic species in the performed Ullman coupling reactions.

Regioselective Ullmann Coupling Reaction of Bromaminic Acid with 2,5-Diaminobenzene Sulfonate
It had been previously reported that 1-amino-4-(4-amino-2-sulfophenylamino)-9,10-dioxo-9,10dihydro-anthracene 2-sulfonate (8) could be obtained in 10% yield using CuCl as a catalyst in the presence of Na 2 SO 3 in water at 40-60 °C in a reaction time of 8 h [21]. We closely followed the described procedure and tried it more than three times on different days, but all of our attempts failed, and we never observed any traces of 8; the meta-isomer 9 was also not formed. The reaction did not proceed and we isolated only the starting material bromaminic acid (4). Then we gradually increased the temperature up to 90 °C and prolonged reaction times (up to 24 h), but without any success. Besides the starting compound 4 we now also isolated the hydrolysis product sodium 1-amino-4hydroxyanthraquinone-2-sulfonate [28].
Therefore we performed a systematic study of the coupling of 2,5-diaminobenzene sulfonic acid (5) with bromaminic acid (4) in the presence of different catalysts and in different solvent systems with the goal to obtain product 8 ( Table 2). This reaction may lead to two different isomers containing an ortho-(8) or a meta-sulfonate group (9) on ring D ( Figure 1 and Table 2). At first, we examined catalysis by elemental copper, which had been identified as the most powerful catalyst in the coupling of aniline with bromaminic acid (see Table 1), in the presence of different mixtures of the phosphate buffer. As outlined in Table 2 the meta-sulfonate 9 (meta with respect to the aminoanthraquinone) was the predominant product in most cases, especially when the yield was high (>50%; entries 3-6, Table 2). The best ratio to obtain the desired ortho-isomer 8 (ortho:meta 1:1) was observed at pH 4.8 and 7.0 (entry 1 and 2, Table 2), however conversion and yield were low. Next we examined the effects of CuCl and CuSO 4 in a sodium phosphate buffer mixture (pH 7.0, entry 8 and 9, Table 2). Within 5 min of microwave irradiation at 120 °C 100% conversion and a very good yield (>75%) was observed, but the (ortho:meta) ratio was dramatically reduced to 1:6. > 75 a Reaction conditions: A mixture of bromaminic acid (40.5 mg, 0.1 mmol), 2,5-diaminobenzene sulfonic acid (56.4 mg, 3 eq.), catalyst (5 mol %) and 5 mL of different buffer mixtures was irradiated in the microwave for 5-10 min at 120 °C. b Conversion and the ortho:meta sulfonate ratio was estimated by RP-TLC using a mixture of acetone/water (1:4) as eluent; this is possible because all components (starting material and product) have different colors: the starting material is red, while the product is blue and the by-product is dark-red or violet. c Yield was estimated based on RP-TLC results.
As a subsequent step we investigated the impact of triethylamine (Et 3 N) in water on the ortho:meta ratio of products 8 and 9 using three different catalysts ( Table 3). The use of NEt 3 in aqueous media is known to convert CuSO 4 to the complex [Cu(NEt 3 ) 4 ] 2+ , which could then be reduced to Cu(I) [32,33]. In the absence of Et 3 N the reaction was incomplete (30-60% conversion), but it was directed towards more ortho-isomer in the following sequence of catalysts: Cu > CuCl > CuSO 4 (entries 1-3, Table 3). The use of three equivalents of Et 3 N improved the yield. However, the ortho:meta ratio was decreased; in case of Cu a ratio of 1:1 could be observed, but conversion and yield were relatively low compared to the use of CuCl and CuSO 4 (entries 4-6, Table 3). In conclusion of this part, CuSO 4 was found to be the best catalyst in the Ullmann coupling reaction of bromaminic acid with aniline derivative 5 in water in the presence of Et 3 N, since it showed quantitative conversion (100%) and an excellent yield (>75%) with an acceptable ortho:meta ratio of 1:3 (entry 6, Table 3). > 75 a Reaction conditions: A mixture of bromaminic acid (40.5 mg, 0.1 mmol), 2,5-diaminobenzene sulfonic acid (56.4 mg, 3 eq.), catalyst (5 mol %) and 5 mL of water with or without Et 3 N (3 eq.) was irradiated by microwave for 5-20 min at 120 °C. b Conversion and the ortho:meta sulfonate ratio was estimated by RP-TLC using a mixture of acetone/water (1:4) as eluent; this is possible because all components (starting material and product) have different colors, the starting material is red, while the product is blue and the by-product is dark-red or violet. c Yield was estimated based on the RP-TLC results.

Regioselective Ullmann Coupling Reaction of Bromaminic Acid with 2,4-Diaminobenzene Sulfonate
For a direct comparison and to further investigate the scope of the developed Ullmann coupling reaction and its regioselectivity we reacted bromaminic acid (4) with 2,4-diaminobenzenesulfonic acid (6) using different copper catalysts. At first we examined the effect of different pH values of the sodium phosphate buffer in the presence of elemental copper (Cu) heating the reagents in a microwave oven at 120 °C for 5-10 min (see Table 4). We noticed that the reaction in the presence of Cu as a catalyst produced both possible isomers, ortho-and para-substituted.
Especially under basic conditions (pH 8.2 and 9.4, entry 5 and 6 respectively, Table 4) 100% conversion was observed and the coupling was in favour of the para-isomer (with respect to the aminoanthraquinone) with an ortho:para ratio of 1:3. The isolation of both products (10 and 11) was easy due to the big difference in their R f values. The purification and isolation of the two isomers proceeded by flash column chromatography using reversed phase silica gel. In neutral or acidic media the conversion was only about 25% or less (entries 1-3 Table 4) indicating the importance of basic conditions for the reaction. 1:3 >50 a Reaction conditions: A mixture of bromaminic acid (40.5 mg, 0.1 mmol), 2,5-diaminobenzene sulfonic acid (56.4 mg, 3 eq), Cu (5 mol %) and 5 mL of different buffer mixtures was irradiated by microwave for 5-10 min at 120 °C. b Conversion and the ortho:para sulfonate ratio was estimated by RP-TLC using a mixture of acetone/water (1:4) as eluent, this is possible because all components (starting material and product) have different colors: the starting material is red, while the product is blue and the by-product is dark-red or violet. c Yield was estimated based on the RP-TLC results.
Finally we investigated water as a solvent in the presence or absence of triethylamine using different copper catalysts: elemental copper (Cu), copper(I) chloride (CuCl), and copper(II) sulfate (CuSO 4 ). The mixture was irradiated by microwave at 120 °C for 5-20 min (Table 5). Interestingly only ortho-isomer was detected in the water system without triethylamine with an ortho:para ratio of >99:1. However, the reaction did not go to completion and only 50% conversion could be observed. The addition of triethylamine dramatically increased the conversion to 100% in case of CuCl or CuSO 4 as catalysts, whereas in the case of elemental copper (Cu) it was still only about 80%. The ortho:para ratio was dramatically reduced to 5:1 by applying CuCl (entry 5, Table 5). Nevertheless, the CuCl procedure was favorable due to a reaction conversion of 100%, an estimated isolated yield of >75%, and the easiness of isolating and purifying both isomers by reversed phase column chromatography.
It can thus be concluded that the best choice of catalyst in the microwave-catalyzed Ullmann coupling reaction depends on several factors, including (i) the pH value of the reaction mixture, since the protonation state of the aniline and the redox potential of the copper species are pH-dependent; (ii) the chemical nature of the reaction partners; (iii) the presence of complexing agents; and (iv) the desired regioselectivity of the reaction. Therefore the reaction conditions have to be optimized in each case, and depending on the reaction sequence a different copper species may be preferred. 1 >50 a Reaction conditions: A mixture of bromaminic acid (40.5 mg, 0.1 mmol), 2,5-diaminobenzenesulfonic acid (56.4 mg, 3 eq), catalyst (5 mol %) and 5 mL of water with or without Et 3 N (3 eq) was irradiated by microwave for 5-20 min at 120 °C. b Conversion and the ortho:meta sulfonate ratio was estimated by RP-TLC using a mixture of acetone/water (1:4) as eluent. This is possible because all components (starting material and product) have different colors, the starting material is red, while the product is blue and the by-product is dark-red or violet. c Yield was estimated based on the RP-TLC results.

Synthesis of MG 50-3-1 (3)
After having improved the preparation of compound 8-the key step in the synthesis of MG 50-3-1 (3)-we analyzed the possible retrosynthetic pathways of 3. Two different strategies can be envisaged: (i) a linear synthesis (pathway A, blue color, Scheme 1), which has been previously described in the literature for the synthesis of 3 (total yield of 5.5%) [21], and (ii) a new, convergent synthesis (pathway B, red color, Scheme 1).
Using the linear synthesis described in the literature (pathway A) [21] the final coupling reaction failed in our hands. The precursor of product 3 (the dichlorotriazinyl derivative 12, Scheme 1) was found to be unstable under the conditions used for the coupling reaction (aqueous basic media), and 12 was either partially or completely hydrolyzed to the corresponding hydroxychlorotriazinyl-or dihydroxytriazinyl-anthraquinone derivative. This means that the described synthesis of MG 50-3-1 (3) according to pathway A failed in our hands at two critical steps, (i) the regioselective Ullmann coupling reaction described above; and (ii) the final coupling reaction step (pathway A, Scheme 1). Therefore we applied a new, convergent approach designed by retrosynthetic analysis as outlined in Scheme 1, pathway B, and finally succeeded in obtaining compound 3. As mentioned above we systematically improved the key step in the synthesis of compound 3, the Ullman coupling reaction of bromaminic acid (4) with 2,5-diaminobenzensulfonic acid (5) by applying CuSO 4 (5 mol %) as a catalyst in water in the presence of 3 equivalents of triethylamine (entry 6, Table 3). The mixture was irradiated by microwave at 120 °C for 20 min. Product 8 (the Western part of the target molecule 3) was obtained in 13% isolated yield while its isomer 9 was obtained in 38% isolated yield (Scheme 2). The two isomers (ortho-isomer 8 and meta-isomer 9) where readily separated by flash column chromatography on reversed phase silica gel (RP-18) using a gradient of acetone/water as described in the experimental part. The Eastern part of the target structure 3 was synthesized by the coupling of cyanuric acid chloride (14) with 2-aminobenzenesulfonic acid (15) in the presence of sodium bicarbonate (NaHCO 3 ) in acetone/water (2:1) at 0-5 °C in an ice-bath for 30 min. Compound 13 was isolated in 95% yield. The last challenge was the final coupling reaction of the Western part 8 and the Eastern part 13, which required basic conditions (K 2 CO 3 ) that might lead to hydrolysis. Since we had developed a simple reaction yielding gram amounts of compound 13 with excellent yield (95%) the use of 13 in excess did not constitute any problem.

Scheme 1. Two proposed retrosynthetic analysis pathways for MG 50-3-1 (3): Linear synthesis (pathway A, blue color) and convergent synthesis (pathway B, red color
Finally 13 was successfully coupled with 8 in basic media (K 2 CO 3 ) using acetone/water (1:1) as a solvent at room temperature in a reaction time of 7 h. In order to quantitatively convert the precious reagent 8, the use of 3 equivalents of compound 13 was essential. The target compound MG 50-3-1 (3) was obtained in excellent isolated yield (97%). The total isolated yield over three steps was 12%.

General
All solvents and reagents were used as purchased. Thin-layer chromatography was performed using TLC aluminum sheets with silica gel 60 F 254 , or TLC aluminum sheets RP silica gel 18 F 254 (Merck, Darmstadt, Germany). Coloured compounds were visible at daylight; other compounds were visualized under UV light. Flash column chromatography was performed on a Büchi system using silica gel RP-18 (Merck, Darmstadt, Germany). 1 H-and 13 C-NMR data were collected on a Bruker Avance 500 MHz NMR spectrometer at 500 MHz ( 1 H), or 126 MHz ( 13 C), respectively. DMSO-d 6 was used as a solvent. Chemical shifts are reported in parts per million (ppm) relative to the deuterated solvent, i.e., DMSO,  1 H: 2.49 ppm, 13 C: 39.7 ppm, coupling constants J are given in Hertz and spin multiplicities are given as s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), br (broad). The purities of isolated products were determined by LC-ESI-MS (Applied Biosystems API 2000 LCMS/MS, HPLC Agilent 1100) using the following procedure: the compounds were dissolved at a concentration of 0.5 mg/mL in H 2 O-MeOH = 1:1, containing 2 mM NH 4 CH 3 COO. Then, 10 L of the sample was injected into an HPLC column (Phenomenix Luna 3m C18, 50 × 2.00 mm). Elution was performed with a gradient of water-methanol (containing 2 mM NH 4 CH 3 COO) from 90:10 to 0:100 for 30 min at a flow rate of 250 L/min, starting the gradient after 10 min. UV absorption was detected from 200 to 950 nm using a diode array detector. Purity of the compounds was determined over the whole range (200 to 950 nm) and proved to be ≥95%. For microwave reactions a CEM Focused TM Microwave Synthesis type Discover apparatus was used. A freeze dryer (CHRIST ALPHA 1-4 LSC) was used for lyophilization.

General Procedure A: Coupling Reaction of Bromaminic Acid with Aniline
To a 5 mL microwave reaction vial equipped with a magnetic stirring bar were added bromaminic acid sodium salt (4, 40.5 mg, 0.1 mmol), aniline (27.5 µL, 3 eq) and 5 mol % catalyst (Cu, CuCl, CuCl 2 , or CuSO 4 ) followed by 5 mL of different solvent systems consisting of different concentrations of sodium phosphate buffer {Na 2 HPO 4 (pH 9.4) [0.20 M] and NaH 2 PO 4 (pH 4.8) [0.12 M]} or water (pH 7). The mixture was capped and irradiated in the microwave oven (100 W) for 5-150 min at 120 °C. Then the reaction mixture was cooled to rt, and the conversion was determined by RP-TLC using acetone/water (1:4) as the mobile phase. To a 5 mL microwave reaction vial equipped with a magnetic stirring bar were added bromaminic acid sodium salt (4, 40.5 mg, 0.1 mmol), 2,5-diaminobenzene sulfonic acid (5) or 2,4-diaminobenzene sulfonic acid (6) (56.4 mg, 3 eq), and 5 mol % catalyst (Cu, CuCl, or CuSO 4 ) followed by 5 mL of different solvent systems consisting of different concentrations of sodium phosphate buffer {Na 2 HPO 4 (pH 9.4) [0.20 M] and NaH 2 PO 4 (pH 4.8) [0.12 M]}, water (pH 7) or water in the presence of three equivalents of triethylamine. The mixture was capped and irradiated in the microwave oven (100 W) for 5-20 min at 120 °C. Then the reaction mixture was cooled to rt, and the conversion was determined by RP-TLC using acetone/water (1:4) as eluent.