Regioselective Introduction of Ethoxycarbonylmethyl and Cyanomethyl Groups into Quinoline and Isoquinoline$ Micheline Grignon-Dubois* and Fa�â za Diaba Laboratoire de Chimie Organique et Organome�Ê tallique, CNRS UMR 5802, Universite�Ê Bordeaux I, 351, Cours de la Libe�Ê ration, F-33405 Talence-Cedex, France A new regioselective route to ethoxycarbonylmethyl- and cyanomethyl-1,2-dihydro-N-methyl-quinolines and -isoquinolines starting from methyl-quinolinium or -isoquinolinium iodides and commercially available trimethylsilyl reagents is presented. As part of a programme directed toward the synthesis of new drugs from functionalized heterocyclic derivatives, we needed to introduce ethoxycarbonylmethyl and cyanomethyl groups on the C-2 position of the quinoline ring.Ethoxy- carbonylmethyl-N-methoxycarbonyldihydroquinolines have previously been obtained by treating ethyl(tributylstannyl) acetate with quinoline and methyl chloroformate.1 However, organostannyl reagents are often dicult to eliminate and are toxic and polluting.Moreover, this process seems to be limited to quinolines activated by alkyl chloroformate, whereas we needed an alkyl group on the nitrogen atom. In the case of 1-methylquinoliniums, Fukuzumi et al.2 obtained the C-2 alkoxycarbonylmethylene adduct as the only pro- duct or in a mixture with the C-4 regioisomer when using a large excess of ketene silyl acetal. Reaction of ethyl bromo- acetate with Reissert compounds3 has also been said to lead to 1-ethoxycarbonylmethyl-1-cyano-2-carbophenyl-1,2- dihydroisoquinoline, but in low yields.Concerning the cyanoalkyl dihydroquinoline derivatives, the C-4 (a-cyano- benzyl) adduct was obtained by treating quinolinium methiodide with phenylacetonitrile and sodium ethoxide.4 In contrast, there is no reference in the literature related to the introduction of cyanoalkyl groups on the C-2 position of quinolines. We report here a new regioselective route to ethoxycarbonylmethyl- and cyanomethyl-1,2-dihydro-N- methyl-quinolines and -isoquinolines starting from methyl- quinolinium or -isoquinolinium iodides (1, 2) and commer- cially available trimethylsilyl reagents.Results and Discussion Nucleophilic addition to quinolinium salts is well known for the functionalization of the quinoline ring.5 However, the regiochemistry of the addition is reported to be depen- dent on substituent e€ects as well as on the nature of the nucleophilic reagent, leading to a competition between C-2 and C-4 additions.5,6 We have recently demonstrated that sonochemical activation allows regiospeciRc C-2 addition of anions to quinolinium iodides in good to quantitative yields.7 In particular, the C-2 addition of the acetonyl anion, prepared in situ from acetone and sodium hydroxide, was systematically obtained when treating with N-methyl- quinolinium iodides.7a From these results we Rrst tried to generate ethoxycarbonylmethyl or cyanomethyl moieties using sodium hydroxide and ethyl acetate or acetonitrile.All our attempts to functionalize 1 in this way failed, and the 1-methyl-2-qinolone 3 was systematically obtained as the only product (68�}75% yield). In contrast with the behaviour of methyl ketones,7a and despite the sonochemical activation, the pKa of ethyl acetate or acetonitrile (24�}25 compared to 20 in the case of methyl ketones8) does not allow the proton abstraction leading to the anion to compete with the addition of the hydroxide anion to the methiodide (Scheme 1).This led us to use trimethylsilyl reagents. As expected, trimethylsilyl acetate (ETSA) alone did not react with com- pound 1.9 The same reaction conducted with 1 equivalent of sodium hydroxide led to 4a, but in only 8% yield along with 3 (34%). In order to avoid the quinoline formation, we decided to replace sodium hydroxide by �Puoride. Indeed, the formation of a Si�}F bond (142 kcal mol1) is usually a highly exothermic process, which provides the driving force for a number of useful synthetic reactions.10 This is the case of active-methylene silylated compounds like ETSA or tri- methylsilylacetonitrile (TMSAN), which are well known to add to the carbonyl double bond in the presence of �Puoride leading to silyl-Reformatsky products.11 When quinolinium methiodide 1 was treated with ETSA or TMSAN in the presence of dried alkali metal �Pouride in acetonitrile solution the nucleophilic addition of the methylene anion was systematically observed (Scheme 1).As expected on the basis of their respective nucleophilic power, caesium �Puoride led to better yields than potassium �Puoride (86 and 40% respectively). Attempts to replace acetonitrile by methylene chloride or THF and alkali metal �Puoride by tetrabutylammonium �Puoride were unsuccessful. These observations are consistent with a charge-controlled process. The reactions were also conducted with 2, leading to the C- 1 adducts 5a, 5b.Examination of Table 1 shows that yields are better with ETSA than with TMSAN, and with quino- line compared to isoquinoline. From a mechanistic point of view, these results are consistent with formation of the methylene anion by nucleophilic attack at the silicon, and its addition to the ortho position versus nitrogen, which is the most electrophilic site. It can be suggested that the process J. Chem. Research (S), 1998, 660�}661$ Scheme 1 Reagents and conditions: i, CH3Y, NaOH, CH3CN, ultrasound; ii, Me3SiCH2Y, CsF, CH3CN, reflux or sonochemical activation $This is a Short Paper as deRned in the Instructions for Authors, Section 5.0 [see J.Chem. Research (S), 1998, Issue 1]; there is there- fore no corresponding material in J. Chem. Research (M). *To receive any correspondence (e-mail: m.grignon@lcoo.u-bor- deaux.fr). 660 J. CHEM. RESEARCH (S), 1998is concerted and takes place in close proximity to the quino-linium iodide as depicted in Scheme 2.Compounds 4b, 5a, 5b are original.Compound 4a hadbeen previously isolated by Fukuzumi et al.2 when treating1 with the ketene triethylsilyl acetal of ethyl acetate andBu4NF. Only the NMR yield was reported (100%), andthe experimental protocol was not described. Moreover,it appears that 10 equivalents of the silyl reagent werenecessary for the reaction to occur. Our process, which iseasy to handle and only needs 1 equivalent of commerciallyavailable silyl reagents, constitutes a real improvement forsynthesizing 4a.ExperimentalFlash column chromatography techniques (302 cm column)were employed to purify crude products using 70¡Ó230 mesh alumina(activity II-III, CH2Cl2) under positive air pressure.Ultrasound-promoted reactions were carried out in a common ultrasoniclaboratory cleaner lled with thermostatted water at 0¡Ó5 8C.Synthesis grade acetonitrile (Aldrich) was dried on molecularsieves.Commercial caesium and potassium uorides (Aldrich, re-agent ACS) were dried prior to use in a domestic microwave oven.Methiodides 1, 2 were prepared according to literature procedures12by alkylation of quinoline or isoquinoline with methyl iodidein acetone solution.To a stirred solution of quinolinium iodide (2.02 g, 7.4 mmol)in acetonitrile (20 mL) was added uoride (KF or CsF, 8.1 mmol)and the trimethylsilyl reagent (Me3SiCH2Y, 8.1 mmol). The mixturewas then stirred at reux or sonicated at 0¡Ó5 8C until the startingmaterial completely reacted as monitored by TLC (SiO2, MeOH¡ÓMe2CO, 10: 90).The reaction mixture was ltered through Celiteand the ltrate evaporated to dryness. The residue was taken upwith cyclohexane (R CH2COOEt, 100 mL) or dichloromethane(R CH2CN, 100 mL) and insoluble materials, if present, wereremoved by ltration and analysed separately. The solution wasevaporated to dryness leading to an oil, which was chromato-graphed eluting typically with CH2Cl2 (Al2O3, activity II-III,70¡Ó230 mesh).2-Ethoxycarbonylmethyl-1-methyl-1,2-dihydroquinoline 4a.H(250 MHz, CDCl3) 1.22 (3 H, t, J 7.1 CH2CH3), 2.44 (1 H, dd,J 14.4, 7.6, CH2CO), 2.56 (1 H, dd, J 14.4, 5.3, CH2CO), 2.91 (3 H,s, NCH3), 4.05 and 4.12 (2 H, dq, J 10.7 and 7.1, CH3CH2), 4.47(1 H, ddd, J 7.6, 5.5 and 5.3, H-2), 5.78 (1 H, dd, J 9.5 and 5.5,H-3), 8 (1 H, d, J 8.1, H-8), 6.66 (1 H,dd, J 7.3 and 7.2, H-6), 6.93 (1 H, dd, J 7.2, and 1.5, H-5), 7.12(1 H, ddd, J 8.1, 7.3 and 1.5 Hz, H-7); C (62.89 MHz, CDCl3)14.2 (CH2CH3), 36.5 (NCH3), 38.0 (COCH2), 57.6 (C-2), 60.6(CH3CH2), 111.0 (C-8), 117.1 (C-6), 123.8 (C-3), 126.3 (C-4), 126.9(C-5), 129.2 (C-7), 121.8, 144.2 (C), 171.4 (CO); m/z 231 (M, 6.4),144 (M£¾ CH2COOEt, 100%); IR (neat) ~max/cm£¾1 1630 (C.C),1725 (C.O) (Found: C, 72.45; H, 7.52; N, 6.0; O, 14.01. C14H17NO2requires C, 72.70; H, 7.41; N, 6.06; O, 13.83%).2-Cyanomethyl-1-methyl-1,2-dihydroquinoline 4b.H (250 MHz,CDCl3) 2.37 (1 H, dd, J 16.3, 6.8, CH2CN), 2.47 (1 H, dd, J 16.3and 5.5, CH2CN), 3.01 (3 H, s, NCH3), 4.40 (1 H, ddd, J 6.8, 5.6and 5.5, H-2), 5.82 (1 H, dd, J 9.6 and 5.6, H-3), 6.56 (1 H, d, J9.5, H-4), 6.56 (1 H, d, J 7.6, H-8), 6.74 (1 H, dd, J 7.5 and 7.4, H-6), 7.00 (1 H, dd, J 7.4, and 1.4, H-5), 7.18 (1 H, ddd, J 7.6, 7.5and 1.4 Hz, H-7); C (62.89 MHz, CDCl3) 21.1 (CNCH2), 37.1(NCH3), 57.5 (C-2), 111.4 (C-8), 117.9 (C-6), 121.3 (C-3), 127.4(C-5), 127.7 (C-4), 129.8 (C-7), 122.0, 143.1 (C) 117.7 (C/N); m/z184 (M, 5.8), 144 (M£¾ CH2CN, 100%); IR (neat) ~max/cm£¾1 1640(C.C), 2240 (C/N) (Found: C, 78.48; H, 6.51; N, 14.99.C12H12N2requires C, 78.22; H, 6.57; N, 15.21%).1-Ethoxycarbonylmethyl-2-methyl-1,2-dihydroisoquinoline 5a.H(250 MHz, CDCl3) 1.17 (3 H, t, J 7.1, CH2CH3), 2.50 (1 H, dd,J 14.1 and 7.1, CH2CO), 2.71 (1 H, J 14.1, 6.2, CH2CO), 2.94 (3 H,s, NCH3), 4.04 (2 H, q, J 7.1), 4.80 (1 H, ddd, J 7.1, 6.2 and1.1, H-1), 5.35 (1 H, d, J 7.3 H-3), 6.03 (1 H, dd, J 7.3 and 1.3,H-4), 6.80¡Ó6.96 (2 H, m), 6.98 (1 H, ddd, J 7.5, 7.2 and 1.3),7.10 (1 H, ddd, J 7.3, 7.2, and 1.8 Hz); C (62.89 MHz, CDCl3)14.1 (CH2CH3), 36.3 (COCH2), 40.5 (NCH3), 59.3 (C-1), 60.5(CH3CH2), 97.6, 122.7, 124.7, 125.6, 127.6, 136.1 (6 CH), 128.1,132.4 (C), 171.7 (CO); m/z 231 (M, 7.6), 144 (M£¾ CH2COOEt,100%); IR (neat) ~max/cm£¾1 1610 (C.C), 1725 (C.O) (Found:C, 72.61; H, 7.43; N, 6.12; O, 13.63.C14H17NO2 requires C, 72.70;H, 7.41; N, 6.06; O, 13.83%).1-Cyanomethyl-2-methyl-1,2-dihydroisoquinoline 5b.H (250MHz,CDCl3) 2.46 (1 H, dd, J 16.4, 6.7, CH2CN), 2.63 (1 H, dd, J 16.4,6.7, CH2CN), 3.02 (3 H, s, NCH3), 4.69 (1 H, dd, J 6.7 and 6.1,H-1), 5.40 (1 H, d, J 7.3, H-3), 6.04 (1 H, d, J 7.3 Hz, H-4),6.90¡Ó7.40 (4 H, m); C (62.89 MHz, CDCl3) 19.2 (CNCH2),40.9 (NCH3), 59.4 (C-1), 98.3, 125.3, 126.0, 126.2, 132.0, 135.2(6 CH), 126.9, 132.1 (C), 118.7 (C/N); m/z 184 (M, 8.9), 144(M£¾ CH2CN, 100); IR (neat) ~max/cm£¾1 1625 (C.C), 2320 (C/N)(Found: C, 78.29; H, 6.49; N, 14.89.C12H12N2 requires C, 78.22;H, 6.57; N, 15.21%).We thank the Conseil Re gional d'Aquitaine for nancialsupport.Received, 9th April 1998; Accepted, 30th June 1998Paper E/8/02694GReferences1 T. G. Murali Dhar and C. Gluchowski, Tetrahedron Lett., 1994,35, 989.2 S. Fukuzumi, M. Fujita, S. Noura and J. Otera, Chem. Lett.,1993, 1025.3 J. Ezquerra and J. Alvarez-Builla, J. Chem. Soc., Chem.Commun., 1984, 54.4 N.J. Leonard and R. L. Foster, J. Am. Chem. Soc., 1952, 74, 3671.5 For reviews, see N. Y. Sidgewick and F. R. S. Sidgewick, TheOrganic Chemistry of Nitrogen, Clarendon Press, Oxford, 1966,p. 718; S. F. Dyke, Adv. Heterocycl. Chem., 1972, 14, 279;J. Gurnos, Quinolines, Wiley, New York, 1982.6 See, for example, N. J. Leonard and R. L. Foster, J. Am. Chem.Soc., 1951, 73, 3325; 1952, 74, 2110; J. Metzgzer, H. Larive ,E.-J. Vincent, R. Dennilauler, R. Baralle and C.Gaurat, Bull.Soc. Chim. Fr., 1967, 30; J. Metzgzer, H. Larive , E.-J. Vincentand R. Dennilauler, Bull. Soc. Chim. Fr., 1967, 46; G. T.Pilygun and B. M. Gutsulyak, Russ. Chem. Rev., 1963, 32, 167;J. W. Bunting and W. G. Meathrel, Tetrahedron Lett., 1971,133; S. Fukuzumi and S. Noura, J. Chem. Soc., Chem. Commun.,1994, 287; M. Maeda, Chem. Pharm. Bull., 1990, 38, 2577.7 (a) F. Diaba, I. Lewis, M. Grignon-Dubois and S. Navarre,J. Org. Chem., 1996, 61, 4830; (b) M. Grignon-Dubois, F.Diabaand M.-C. Grelier-Marly, Synthesis, 1994, 800; (c) M. Grignon-Dubois and A. Meola, Synth. Commun., 1995, 25, 2999.8 J. March, Advanced Organic Chemistry, Wiley Interscience, NewYork, 4th edn., 1992, ch. 8, p. 249.9 Silicon reagents are known to be weaker nucleophiles than tinreagents: see for example, Hosomi, Chem. Lett., 1979, 977.10 For reviews, see W. P. Weber, Silicon Reagents for OrganicSynthesis, Springer, Berlin, 1983; E. W. Colvin, Silicon OrganicSynthesis, Butterworths, London, 1981; G. G. Yakobson andN. E. Akhmetova, Synthesis, 1983, 169; J. H. Clark, Chem. Rev.,1980, 80, 429.11 R. Latouche, F. Texier-Boullet and J. Hamelin, Bull. Soc. Chim.Fr., 1993, 130, 535; S. Jolivet, S. Abdallah-El-Ayoubi, D. Mathe,F. Texier-Boullet and J. Hamelin, J. Chem. Res., 1996, 300;E. Nakamura, M. Shimizu and I. Kuwajima, Tetrahedron Lett.,1976, 1699.12 O. Doebner and W. Miller, Ber. Bunsenges. Phys. Chem., 1883,16, 2464; C. F. Dun, Adv. Heterocycl. Chem., 1964, 3, 1.Scheme 2Table 1 Condensation of Me3SiCH2Y/CsF/CH3CN withcompounds 1 and 2Substrate Reagent Y Conditions Product yield (%)1 COOEt 2 h, )))) 4a : 861 COOEt 2 h, reflux 4a : 841 CN 2 h, )))) 4b : 431 CN 3.5 h, reflux 4b : 272 COOEt 2 h, )))) 5a : 562 COOEt 2 h, reflux 5a : 772 CN 3 h, )))) 5b : 482 CN 3 h, reflux 5b : 51J. CHEM. RESEARCH (S), 1998 661