Expanding Access to Remdesivir via an Improved Pyrrolotriazine Synthesis: Supply Centered Synthesis

Pyrrolotriazine 1 is an important precursor to remdesivir. Initial results toward an efficient synthesis are disclosed consisting of sequential cyanation, amination, and triazine formation beginning from pyrrole. This route makes use of highly abundant, commoditized raw material inputs. The yield of triazine was doubled from 31% to 59%, and the synthetic step count was reduced from 4 to 2. These efforts help to secure the remdesivir supply chain.


Representative Procedures:
2-cyanopyrrole (Table S2,  With stirring (300 RPM), phosphoryl trichloride (38.3 mL, 409.9 mmol, 1.1 eq.) was slowly charged while maintaining internal temperature below 15 °C. The reaction mixture was stirred at 20 °C for another 30 minutes. The reaction was cooled to 0 -5 °C and pyrrole (25.0 g, 372.6 mmol, 1.0 eq.) was slowly charged while maintaining internal temperature below 15 °C. The reaction mixture was stirred at 20 °C for another 1h. The light brown solution was cooled to 0 -5 °C and water (75 mL) was added while maintaining internal temperature below 15 °C. The reaction was stirred for 5 minutes at same temperature. Solid hydroxylamine hydrochloride (28.48 g, 409.9 mmol, 1.1 eq.), acetic anhydride (38.7 mL, 409.9 mmol, 1.1 eq.) and pyridine (150 mL, 1.863 mol, 5.0 eq.) were added sequentially, the internal temperature was not allowed to exceed 15 °C. After the addition was complete, the reaction mixture was heated to 90 °C (heating block) for overnight (16 hours). After complete consumption of the iminium chloride salt, as indicated by 1 H NMR, the reaction mixture was diluted with water (250 mL) and transferred to the separating funnel. The product was extracted from dark brown reaction mixture by using ethyl acetate (3 X 400 mL). The combined organic layer was washed with 1M HCl (250 mL) and brine (250 mL), dried over sodium sulfate and concentrated using rotary evaporation under reduced pressure. The desired 2-cyanopyrrole (34.7 g, 90% (adjusted with 89% purity by HPLC)) was distilled out from dark brown oil under vacuum at 100-140 °C. The obtained product was used for next (amination) step (52 g CSTR batch; in Batch Amination with NH2Cl from CSTR section). NH4Cl (24.6 g) in MTBE (330 mL) was cooled to -5 °C (internal temperature) in a 2L-round bottom flask, and concentrated NH4OH (38.4 mL) was added. Sodium hypochlorite (10-15% solution, Sigma-Aldrich, 432 mL) was then added via addition funnel over 30 min. The mixture was stirred for 30 min, the layers were separated, and the organic layer was washed with brine (1 × 180 mL). The organic layer was dried over powdered CaCl2 (15 g) in the freezer for at least 1 h and kept at the same temperature. Approximate concentration is 0.74 M.
A 0.5 L three necked round bottom flask was equipped with a J-KEM internal temperature probe and a stirring bar. A solution of 1H-pyrrole-2-carbonitrile (10.0 g, 105.3 mmol, 97% purity) in MTBE (100 mL) was added, cooled in an ice bath (internal temperature between 5-10 °C) and NaH (6.3 g, 158.0 mmol, 60% in mineral oil Sigma-Aldrich) was added in portions.  The synthesis began with C-2 formylation of pyrrole and further its oxidation to nitrile. Both reactions are well known in literature, however, 2-formyl pyrrole is a low melting solid not easily distilled or recrystallized in good yield. Moreover, waste will be generated in the process of purifying the aldehyde, aldoxime or other intermediates. We wondered whether isolation of the aldehyde or aldoxime intermediate was necessary. Both reactions can be performed in same solvent, the iminium chloride salt could be used directly to form the nitrile in one-pot. Considering this, we began our efforts with formylation of pyrrole using POCl3 (1.1 eq.) in DMF (15V) (Table S1, entry 1). The complete conversion of pyrrole to iminium chloride salt was observed in 1h, analyzed by 1 H NMR. According to the literature, 1 the oxidation of 2formylpyrrole to 2-cyanopyrrole is feasible at higher temperature (125 °C). Keeping this in mind, the hydroxylamine hydrochloride (1.1 equiv.) was added to the reaction mixture and heated at 125 °C for 16 h (overnight). The reaction was dark and sluggish, however, gave 58% assay yield by 1 H NMR (mesitylene was used as internal standard). Further, the assay yield dropped to 25% by lowering the reaction temperature to 90 °C (Table S1, entry 2). The lower yields could be because of chemical incompatibility of hydroxylamine hydrochloride with residual POCl3 and related species which would need to be quenched. Thus, water or ethanol was added to the reaction mixture prior to addition of hydroxylamine hydrochloride.

One-pot Synthesis of 2-Cyanopyrrole from Pyrrole
The HCl generated in the course of the quench could be used as catalyst for the dehydration of aldoxime. Surprisingly, the reactions gave >80% assay yield under these conditions (Table S1, entry 3 and 4).

Exploring effect of amount of DMF and water on reaction:
The effect of amount of solvents like, DMF and water on reaction was examined (Table S1, entry 5-10).
The reaction with 15V of DMF and 5V of water provided 94% assay yield. Importantly, the results were reproduced at 25 g batch under similar reaction conditions (Table S1, entry 9). The reaction gave 72 % isolated yield (adjusted with 72% purity). However, the extraction of product was found difficult due to emulsion formation and was not feasible at large scale. This led us to further optimize our reaction conditions to make product isolation easy. Next, the reactions were attempted under basic conditions by activating the in situ formed oxime with acetic anhydride. Initially, the reactions were performed with pyridine (5 eq.) and acetic anhydride (1.2 equiv.) in DMF (Table 2S, entry 1 and 2). The reactions were smooth and provided excellent assay yields (>90%).

Reaction screening for alternative solvents and amount of ethanol required for reaction:
Exploring solvent volumes and amount of pyridine: Further, the reaction was screened for an alternative solvents, concentration and reagents. In this regard, few experiments were performed on 0.2 g scale, and it was found that DMF and acetonitrile works well. Acetonitrile would be better compared to DMF, as it could be easily evaporated from reaction which will be helpful for isolation of the product. With this results in hand, the reaction was scaled up to 25 g under similar reaction conditions (Table S2, entry 5). The reaction gave 81% yield with 58% purity, however, triethylphosphate generated by reaction of EtOH with residual POCl3 and its associated species was problematic for purification process. The addition of water instead of ethanol resolved this problem and reaction provided excellent yield and purity at 5 g and 25 g scale (Table S2, entry 6 and 7). In addition, inorganic bases, NaHCO3 and NaOH, were tried instead of pyridine giving lower yields (Table S2, entry 8 and 9). We ended up with pyridine as best base for the reaction. Importantly, the reaction reproducibility was demonstrated at 100 g scale (Table S2, entry 11). The product was purified by vacuum distillation and used for next step. As discussed in manuscript, in order to avoid dilute reaction conditions and increase throughput for Namination of 2-cyanopyrrole, the solid aminating reagents like, O-(4-nitrobenzoyl)hydroxylamine (A) and O-(diphenylphosphinyl)hydroxylamine (B) were planned to use. At the outset, in 10 mL glass vials, the solution of 2-cyanopyrrole (100 mg) in DMF (2 mL) was treated with aminating reagents (A and B; 1.5 eq.; individually) in presence of NaH (1.5 eq.). However, both reactions were very thick (almost dry) and unable to stir with magnetic stir bar. Then, the reactions were attempted with excess amount of DMF (Table  S3, entry 1 and 2) and heated at 80 °C overnight (16 h). The reactions showed >90% conversion (by HPLC method). The assay yields were derived by using 1 H NMR (mesitylene was used as internal standard). The assay yields showed well agreement with conversion. However, the reactions were highly diluted at this stage. Then, reactions were examined in EasyMax instrument with overhead stirrer employing lower amount of solvents. The reactions were carried out on 500 mg scale with both aminating reagents (A and B) in DMF (25V and 40V) under similar reaction conditions (Table 3S, entry 3 and 4). The results were excellent, showing ~90% conversion in both reactions.

Exploring different solvents and bases with O-(diphenylphosphinyl)hydroxylamine (B) reagent:
Next, we screened the effect of solvents (DMF and NMP) and bases (NaH, KO t Bu and KOH) with both aminating reagents (A and B). NMP and KO t Bu combination gave the best result. Both reagents showed 94% conversion and assay yields were in good agreement with conversion (Table 3S, entry 5 and 6). Further, these reactions conditions were taken up forward for large scale reaction. The reactions were examined to investigate the lowest solvent volume that could be used on EasyMax (Table 3S, entry 7-10).
The reactions were carried out on 1.0 g scale with lowest 10V and 15V of solvent. The reaction with 15V of solvent gave good conversion and assay yield.

Procedure for synthesis of N-amination of 2-cyanopyrrole:
A 50 mL glass reactor was equipped in EasyMax instrument with an internal temperature probe, overhead stirrer and nitrogen line. The flask was charged with anhydrous NMP (15 mL) and cooled to an internal temperature of 0 -5 °C under an atmosphere of nitrogen. With stirring (300 RPM), 2-cyanopyrrole (1.0 g. 10.86 mmol, 1.0 eq.) was charged. Then, solid KO t Bu (1.34 g, 11.94 mmol, 1.1 eq.) was added to reaction mixture. The temperature was raised to 10 °C, reactions stirred for 10 minutes. To the reaction mixture solid O-(diphenylphosphinyl)hydroxylamine (2.78 g, 11.94 mmol, 1.1 eq.) was charged in four portions over 10 minutes. The temperature of reaction was slowly increased to 80 °C and then reaction continued for overnight (16 hours). The consumption of the starting 2-cyanopyrrole was monitored with HPLC (95%) and assay yield (95%) was derived by using 1 H NMR.

Investigation of NH2Cl Extraction
3.00 g of NH4Cl (57.2 mmol) was added to a 250 mL Erlenmeyer flask along with 4.70 mL of 14.5 M NH4OH (68.1 mmol). The mixture was cooled to -5 °C and stirred on a magnetic hotplate. 72 mL of bleach (2.2 M, 158 mmol) was added via addition funnel over the course of 15 minutes. The concentration of the bleach was stated as 7.5% NaOCl (7.1% active chlorine) which translates to 15.8 wt% NaOCl and 2.2 M NaOCl as found by iodometric titration. Some of this apparent difference in concentration exists as a result of ambiguity in tradeterms of the bleaching industry. The same procedure was used with more concentrated bleach, but the portions of reagent were changed as follows: 4.08 g NH4Cl, 6.39 mL NH4OH, and 72 mL of 10.6% NaOCl were mixed. Care should be taken working with monochloramine. It is a reactive oxidant.
The mixture reacted for 10 minutes at -5 °C prior to sampling. The aqueous feed was then held at this temperature through course of the study. 8 mL aliquots were taken to explore extraction of NH2Cl into organic solvent under various conditions. The aqueous chloramine was added to a 20 mL scintillation vial and vigorously shaken with organic solvent for 30 seconds. The biphasic mixture was allowed to separate and 1 mL of the organic phase was removed and titrated by the iodometric titration (See below). The results are listed in the table below.
Iodometric Titration: 6.20 g of Na2S2O3·5(H2O) was dissolved in 250 mL of water and set aside. 0.5 g of starch was dissolved in 50 mL of water by heating (hot plate) to 80 °C for 10 minutes and then set aside. 0.8 g of NaI was dissolved in 200 mL of water, and 10 mL of AcOH and 10 mL of the starch indicator were added. 20 mL of the iodide solution was transferred to a 50 mL Erlenmeyer flask with stir bar. 1.00 mL of NH2Cl in organic solvent was added to the iodide solution which turned a purplish brown color. The oxidant solution was stirred rapidly on a magnetic hotplate. The thiosulphate solution was added dropwise until the solution became clear. The volume required to quench the oxidant was recorded and molarity of NH2Cl was recorded.
Perhaps the difference between the reported and observed values can be explained as follows. A significant amount of volatile NH2Cl could be lost as the literature procedure evaporates the organic layer prior to filtration. 3 Also, the concentration of NaOCl was not stated. Strength of commercial bleach varies widely.

Continuous Production of Chloramine
A continuous stirred tank reactor (CSTR) was constructed from a 100 mL Schlenk flask with a liquid fill level set at 100 mL. Bleach, NH4Cl/NH4OH, and MTBE feeds were positioned below the liquid level surface, and the dip tubes removing reaction fluids were placed at the top of the fluid surface level, the 100 mL fill level volume. The CSTR was equipped with a large oval shaped stir bar, stirred at 800 rpm, and cooled to -5 °C. The total flow rate of fluids entering the CSTR was set at 10 mL/min to give a residence time of 10 min. The exit flow was set at 12 mL/min, faster than the entering flow to ensure a reactor volume of 100 mL/min. The exit stream flowed into the bottom of a gravity liquid-liquid settler made from a simple pressure-equalizing addition funnel (25 mL

Batch Amination with NH2Cl from CSTR
2-Cyanopyrrole (9.21 g, 10.0 mmol) was added to a 500 mL 3-neck round-bottom flask along with 100 mL of DMF. The solution was cooled to 0 °C with an ice-bath. 5.00 g of 60 wt% NaH in mineral oil was added slowly with stirring, keeping temperature below 35 °C. The round-bottom was connected to two condensers connected in series and cooled to -20 °C with a chiller. The condensers were configured to drain into a 250 mL receiving flask which was used to recycle the MTBE.
After the CSTR reached steady state, the stream of NH2Cl in MTBE was connected to the basified pot of 2-cyanopyrrole. The NH2Cl was added at a rate of 5.67 mL/min. It was added in 5 portions, where each fraction was flowed into the cyanopyrrole for 15 minutes. After every 15 minute addition, the chloramine feed was removed from the amination pot. The reaction vessel was set to 30 °C and placed under vacuum. The MTBE distilled and was collected in the receiving flask. After 14 minutes of evaporation, vacuum was turned off and the vessel was returned to atmospheric pressure. The MTBE was recycled and returned to the feedline used to extract NH2Cl in the CSTR (Stream C).
After the fifth addition of chloramine, conversion was measured as 93%, and 360 mL of MTBE was recovered (84% recovery).

Continuous Amination in PFR with NH2Cl from CSTR
A CSTR was made as described above in a 22 mL Scintillation vial with a fill volume of 20 mL. The total flow rate in was 2 mL/min, the exit was programmed at 3 mL/min, and the residence time was 10 min.
The input feeds for the CSTR are as follows. 15.0 g of NH4Cl (286 mol), 23.5 mL of concentrated NH4OH and 25 mL of H2O were combined and stirred until the mixture was homogeneous (Stream A). The total volume was 60 mL, and pumped at a rate of 0.124 mL/min by syringe pump (Stream A). Aqueous NaOCl was pumped at a rate of 0.742 mL/min (Stream B). MTBE was pumped at a rate of 1.13 mL/min (Stream C). The exit dip tube was programmed to remove solution at a rate of 3 mL/min, and it transported the biphasic mixture to the gravity separator as described previously. Peristaltic pumps from Vapourtec were used for fluid transport.
The NH2Cl in MTBE was pumped at a rate of 1.13 from the gravity separator, and mixed with a solution of 2-cyanopyrrole anion (0.609 M in DMF) via a T-Mixer (Idexx, 0.02 " ID). The 2-cyanopyrrole mixture was flowed at 0.129 mL/min by syringe pump. The solution was made by dissolving 1.40 g of 2cyanopyrrole (15.2 mmol, 1.00 equiv.), 1.22 g NaH (60 wt%, 30.4 mmol, 2.00 equiv.) in DMF to reach a total volume of 25 mL. The reaction mixture flowed into a PFR constructed from PFA tubing (0.06" ID, 5.04 mL, 4 min tR). The reaction ran for 12 minutes before collecting sample for analysis. 89% conversion was observed. Typical multi-charge amination procedure:

Batch Amination with Multicharge of Chloramine
To a solution of 1H-pyrrole-2-carbonitrile (0.46 g, 5 mmol) in MTBE (5 mL) was added NaH (0.40 g, 10.0 mmol, 60% in mineral oil) in portions, and the reaction was stirred for 20 min at room temperature. DMF (5 mL) was added and MTBE was distilled under vacuum at 20-30 °C. NH2Cl in MTBE (5 mL, 0.59M solution) was added via syringe. The mixture was let to stir between 20-30 °C for 5 min and HPLC analysis was carried out. MTBE was distilled under vacuum at 20-30 °C and NH2Cl in MTBE (5 mL, 0.59M solution) was added via syringe. MTBE was distilled under vacuum at 20-30 °C and this cycle was repeated 2 more times. Mesitylene was added as internal standard and the reaction was assayed by quantitative 1 H NMR. This procedure was consistently repeated in 1, 5 and 30 mmol scale (Table S5). NH4Cl (3.0 g) in MTBE (55 mL) was cooled to -5 °C (internal temperature), and concentrated NH4OH (4.7 mL) was added. Commercial bleach (72 mL, Clorox ~7.5% NaOCl) was then added via addition funnel over 15 min. The mixture was stirred for 15 min, the layers were separated, and the organic layer was washed with brine (1 × 30 mL). The organic layer was dried over powdered CaCl2 in the freezer for at least 1 h and kept at the same temperature. Approximate concentration is 0.59 M.

Amination and Triazine Formation in One-Pot
For scale-up to 10 g, a few changes were implemented: 1) In order to increase the throughput of the overall process, DMF was reduced from 10V to 5V in relation to 1H-pyrrole-2-carbonitrile with no impact to the reaction profile; 2) Based on our findings on the relation of NaH vs chloramine (see manuscript for details), NaH was reduced from 2 to 1.5 equiv.; 3) The number of chloramine charges were reduced from 4 to 3. This procedure was employed at in 2.8 g and 10 g scale furnishing the N-amino-2-cyanopyrrole product in 92% assay yield (Table S6, entries 1 and 2). The N-amino-2-cyanopyrrole product was cyclized by adding 3 equiv. of formamidine acetate and heating at 90-95 °C for 16 h. Assay 1 H NMR using 1,3,5trimethoxybenzene as internal standard showed 76% assay yield to triazine, which was isolated in 64% isolated yield after reducing the DMF volume and addition of water (78% purity). Preliminary purification studies showed that >97% pure triazine can be obtained by recrystallizing the crude solid from boiling water/EtOH (1:1), however, further studies showed that trituration with MTBE was more efficient for mass recovery in good purity.
Aiming at further increasing the throughput, more concentrated solutions of chloramine were investigated. Instead of using Clorox (~7.5% NaOCl), a sodium hypochlorite 10-15% solution from Sigma-Aldrich was used to prepare a ca. 0.74M chloramine solution in MTBE. By the addition of 170 mL (~1.2 equiv.) of this solution to deprotonated 1H-pyrrole-2-carbonitrile (10.0 g scale) in DMF gave 98% conversion to the Namino-2-cyanopyrrole product (

Evaluation of Alternative Bases and Solvents for Addressing Safety in the Amination step
We concluded our amination investigation by addressing the hazards surrounding use of NaH in conjunction with DMF. An interesting solution is the deprotonation of 1H-pyrrole-2-carbonitrile in MTBE, then solvent swap to DMF before adding the chloramine (Table S7, entry 2). This alternative was employed in the 10 g batches previously discussed. Alternative solvents and bases were also evaluated. Prior experience at M4ALL Institute 5 suggested that NaH deprotonation can be successful in "glyme type" solvents. Diglyme and diethylene glycol dibutyl ether (DEDGBE) were explored furnishing high conversion to the desired N-aminated product (Table S7, Entries 3 and 4). THF and MTBE were also investigated, however, lower conversion was obtained (Table S7, entries 5 and 6). These results confirm the importance of the coordinating nature of the "glyme type" ethers which can be used as an alternative to DMF. Different bases like, NaHMDS, NaO t Bu and KO t Bu were examined, however the product is formed in reasonable conversion, the reaction profiles did not excel the use of NaH as base ( To a solution of 1H-pyrrole-2-carbonitrile (1 mmol) in DMF or alternative solvent (1 mL) was added NaH or alternative base, and the reaction was stirred for 30 min at room temperature. NH2Cl (4 mL, ca. 0.56 M in MTBE) was added via syringe. The reaction conversion was monitored by HPLC after 1 hour.