| |
[Molecules: 7] [Related articles/posters: 107 037 103 012 062 ] |
Scheme 2
The commercially available 1,2,3,4-tetrahydroisoquinoline 2a was first reacted with trifluoroacetic anhydride (TFAA) to afford the N-protected compound 2b. Under aluminium trichloride catalysis, compound 2b underwent clean Friedel-Crafts acylation as described in the literature and afforded compound 3b as a single isomer in a 86% yield. We confirmed the formation of a single isomer by 1H 200 MHz NMR spectroscopy and semi-empirical calculations are in good agreement with this regioselectivity. Under acidic conditions, the cleavage of the N-COCF3 bond proceeded smoothly to give 3a. We performed the same Friedel-Crafts acylation on the N-methyl derivative 2c ( the latter compound was obtained by quaternization of isoquinoline with methyl iodide followed by reduction of the intermediate isoquinolinium salt with sodium borohydride ). Under the same conditions used for the conversion of 2b into 3b, compound 3c was isolated with great difficulty in a 35% yield. After hydrolysis, the reaction mixture is very acidic and it is necessary to add a large amount of sodium hydroxide before extracting the amine 3c. It is well known that isoquinoline derivatives are not very stable in aqueous basic media and we think that 3c may decompose by ring opening. It was possible to improve the yield of the acylation reaction leading to 3c, on a 1g scale (55% yield). It must be emphasized that the regioselectivity of this reaction is similar to that obtained with 2b (confirmed by 1H 200 MHz NMR spectroscopy).
The transformation of the acetyl derivatives was performed under reinvestigated Willgerodt-Kindler conditions leading to 4b with a 51% yield and to 4c in a 55% yield.
Having at hand the required ester, the final conversion into the carboxamide moiety was more difficult than anticipated probably due to the presence in the same molecule of both a carboxylic acid derivative and a tertiary amine. Finally, the best method involving a complex between lithium aluminium hydride and the appropriate amine (4 equiv.) gave a 60% yield of a mixture of amide 1 and 2-aminopyridine. The decreasing of the amount of 2-aminopyridine (2 equiv. for example) resulted in poor yields in 1. The remaining 2-aminopyridine was removed after bulb-to-bulb distillation followed by HPLC on a C8 column. Compound 1 was unambigously identified by 1H and 13C NMR spectroscopy.
Association constants of receptor 1 with various amines
The binding constants (Ka) were calculated from data from the classical NMR titration method described in the literature assuming a 1:1
complexation mode. When a complexation occured, a shift of the signal
corresponding to the NH2 group of the amine guest was observed. A Job plot for
1 and 4-aminomethylpyridine gave a maximum at about x = 0.5 (mol
fraction of guest) indicating a 1:1 complexation mode.
Table 1 Association constants of receptor 1 with some amines (py = pyridine):
Amine Kass (M-1) o free amine (ppm) max (ppm)
tert-C4H9NH2 ----- 1.53 no shift
C4H9NH2 220 1.24 0.63
2-py-NH2 ? dilution effect -----
Ph-CH2NH2 42 1.54 0.5
4-py-CH2NH2 250 1.56 0.36
2-py-CH2NH2 6.5 1.65 0.25
As can be seen from Table 1, no association was observed in the case of the bulky tert-butylamine whereas good values of Ka were observed in the cases of butylamine and 4-aminomethylpyridine. On the other hand, 2-substituted pyridine derivatives gave very weak binding with 1. Despite the fact that the NMR method gave no result with 2-aminopyridine, we think that the binding with 1 might occur to a small extent if we consider the great difficulties encountered in the removal of this amine from the crude reaction mixture.