Synthesis of functionalized pyridinium salts bearing a free amino group

Tetrasubstituted N -methylpyridinium salts bearing a free tertiary amino group have been synthesized by a straightforward procedure starting from inexpensive starting materials. The key feature of the synthesis is the use of a proton as a simple effective protecting group to achieve selective N-methylation of the pyridine ring without attacking the amino group.


Introduction
In metabolism, NADH is involved in redox reactions that transfer a hydride from one molecule to another.This coenzyme can be found in both its oxidized form (NAD + ) and its reduced form (NADH), which can be used as a reducing agent.The very selective nature of NADH-based reductions has inspired the development of a number of synthetic functional analogues. 1ntroduction of functional groups that can participate in hydride transfer reactions offers the potential to induce new reactivity and selectivity patterns.However, the development of methods for the preparation of such analogues is hampered by the relatively high electrophilicity of pyridines and pyridinium salts, [2][3][4] which significantly complicates attempts to functionalize these structures.This contribution describes our efforts in the synthesis of pyridinium salts 1 bearing free tertiary amine functionalities, which could serve as mediators in hydride transfer reactions.Scheme 1. Retrosynthesis of pyridinium salts 1.

Synthesis of aminopyridines
Ketone 5 was easily prepared in two steps from commercially available compounds following established procedures (Scheme 2): 5,6 first, treatment of ethyl acetoacetate 9 with one equivalent of ammonium carbamate 8 in MeOH gave (Z)-ethyl 3-aminobut-2-enoate 7 in quantitative yield. 5 Subsequent condensation of 7 with 1,3-cyclohexanedione and ethyl orthoformate in refluxing acetic acid led to the desired ketone 5 in moderate but acceptable yield (54%). 6 Scheme 2. Synthesis of ketone 5 Attempts to prepare amine 4 either by reductive amination of ketone 5 or by Mitsunobu reaction of the corresponding alcohol 10 resulted in very low yields due to substantial formation of side products and decomposition.Therefore, a three step procedure involving reduction of ketone 5 to alcohol 11 followed by formation of the desired amine via treatment of 10 with MsCl and subsequent nucleophilic displacement with a secondary amine was evaluated.Reduction of ketone 5 with NaBH4 in MeOH gave the alcohol 10 in quantitative yield (Scheme 3).Subsequent treatment with MsCl in the presence of triethylamine led to a 4:1 mixture of chloride 11 and alkene 12, which were easily separated by column chromatography.Scheme 3. Conversion of ketone 5 to chloride 11.
We were pleased to find that chloride 11 reacted readily with a variety of secondary amines in refluxing acetonitrile 7 to afford the desired amino pyridines in good to excellent yields (Table 1).The best results were obtained with cyclic secondary amines (entries 2-5).The reaction with L-prolinol (entry 5) generated a 1:1 mixture of two diastereoisomers, which could be separated by column chromatography.Amino pyridines 14-17 were readily accessible in gram quantities by this convenient synthetic procedure.After purification by column chromatography these products proved to be very stable and easy to handle solids.In contrast to cyclic amines, diethylamine (entry 1) proved to be unreactive, and even after longer reaction times the conversion was still very low.

Synthesis of pyridinium salts
Preliminary experiments showed that, as expected, direct methylation of amine 14 led to inseparable mixtures of 18, in which the tertiary amino group was methylated, and 19, in which both nitrogen atoms were methylated, along with variable amounts of unreacted starting material (Table 2).
Obviously the amine functionality had to be protected for the preparation of the desired pyridinium salts.As outlined in Scheme 1, we thought that selective protonation of the tertiary amine would provide an effective way to achieve selective N-methylation of the pyridine ring. 8o examine whether selective monoprotonation was indeed possible and whether the protonated amines were sufficiently stable under methylation conditions, we studied the protonation of amine 14 with various Brønsted acids (Table 3).Initial experiments using HCl .Et2O failed, resulting in rapid decomposition of the starting amine (entry 1).Protonation with trifluoromethanesulfonic acid (entries 2 and 3) yielded the desired compound 20, but also led to the formation of unidentified side products.Protonation with 0.9 or 2.0 equivalents of p-toluenesulfonic acid afforded the mono-and diprotonated compounds 22 and 23, respectively, in good yields (entries 4 and 5).Unfortunately, the resulting trifluoromethanesulfonate and p-toluenesulfonate salts proved to be highly sensitive to air and difficult to handle.Eventually, tetrafluoroboric acid diethyl etherate proved to be the optimal choice, affording 24 in 95% yield (entry 6).The resulting salt was found to be stable towards air and moisture, and the corresponding bisprotonated species was not detected even when two equivalents of acid were added (entry 7).In an analogous manner, the monoprotonated compounds 26 and 27 were cleanly obtained in 97% yield (Scheme 4).

Methylation
With reliable procedures for the preparation of monoprotonated amino pyridines in hand, we studied the reaction of standard methylating agents such as methyl trifluoromethanesulfonate, trimethyloxonium tetrafluoroborate (Me3OBF4), dimethyl sulfate or methyl tosylate with ammonium salt 24.While treatment with MeOTos and Me2SO4 led to the formation of complex mixtures, the use of Me3OBF4 resulted in formation of the desired biscationic salt, albeit with low conversion.MeOTf gave more promising results and, therefore, further studies focused on this reagent.After some experimentation, we identified dioxane as ideal solvent for this transformation (Table 4).
Variation of the temperature revealed that an increase from rt to 50 ºC had a negative impact on conversion (entries 1 and 2).Because of the relatively low solubility of 24 in dioxane, solvent mixtures of dioxane and 10% DMF were tested to enhance solubility, but the reaction was slower in this case (entry 3).Notably, shorter reaction times led to higher conversion (entries 4-6), which is consistent with the observation that the resulting biscationic salts are unstable and probably decompose over extended reaction times.Lowering the concentration from 0.2 M to 0.1 M increased the conversion from 22% to 47% (entry 6), a trend that became even more obvious when the concentration was lowered to 0.02 M, resulting in an improved conversion of 70% (entry 8).As observed before higher temperatures led to lower yields (entry 9).At 0 ºC, the 1 H NMR spectra showed no remaining starting material after 1 h, but the formation of almost 20% of an unidentified side product was observed (entry 10).The use of an excess of MeOTf had a negative effect on conversion (entries 11 and 12).In summary, stirring a 0.02 M solution of 24 in dioxane at room temperature for 4 hours proved to be optimal for this reaction (entry 8).Under these conditions, conversions to the biscationic salt in the range of 70% were consistently obtained.Using the same procedure compound 26 was formed with 60% conversion (Scheme 5).

Preparation of N-methyl pyridium salts with a free amino function
In order to complete the synthesis of amino pyridinium salts we needed a suitable base for the deprotonation of the amonium salts that allowed easy removal of the resulting protonated base.We reasoned that sodium hydride would be ideal because the protonated base in this case is hydrogen gas.After methylation of pyridine 24, addition of 1.5 equivalents of sodium hydride to the reaction mixture in dioxane yielded the pyridinium salt 30 and minor amounts of the unmethylated aminopyridine 14 (Scheme 6).Due to the higher solubility of 14, it was possible to separate the two compounds simply by washing with pentane.The generated NaBF4 could then be removed by dissolving 30 in dichloromethane and subsequent filtration through an HPLC filter.In this way the desired product 30 was obtained in 64% yield.This procedure was also successfully applied to the synthesis of pyridinium salt 31, which was isolated in 57% yield.Scheme 6. Deprotonation of 28 and 29.

Conclusions
An efficient straightforward synthetic route to N-methylpyridinium salts containing a tertiary amine function has been developed.Selective N-methylation of the pyridine ring, which is the key step of the synthesis, has been achieved by protonation of the amino group and subsequent reaction with methyl triflate, followed by deprotonation with sodium hydride.Through this procedure amino-functionalized pyridinium salts become conveniently accessible in high purity.

Table 1 .
Reaction of chloride 11 with secondary amines a Pure isolated product after chromatography.bEstimatedby1H NMR analysis of the crude reaction mixture.c Combined yield of both diastereoisomers.

Table 2 .
Direct methylation of amine 14 a Estimated by 1 H NMR analysis of the crude reaction mixture.

Table 3 .
Protonation of amine 14 a Estimated by1H NMR analysis of the crude reaction mixture.

Table 4 .
Methylation with methyl trifluoromethanesulfonate in dioxane a Estimated by 1 H NMR analysis of the crude reaction mixtures.Scheme 5. Methylation of the protonated amino pyridine 26.