Open Access Article
Luke I.
Dixon
a,
Michael A.
Carroll
*a,
Thomas J.
Gregson‡
b,
George J.
Ellames‡
b,
Ross W.
Harrington
a and
William
Clegg
a
aSchool of Chemistry, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK. E-mail: m.a.carroll@ncl.ac.uk; Fax: +44 (0)191 222 6929
bDepartment of Isotope Chemistry and Metabolite Synthesis, sanofi-aventis, Alnwick, Northumberland NE66 2JH, UK
First published on 15th July 2013
Aryl(alkynyl)iodonium salts have been demonstrated to be valuable precursors to a diverse range of heteroaromatic ring systems including aryl[1,2-a]imidazopyridines. Successful application, using the recently described aryl(alkynyl)iodonium trifluoroacetate salts, is described, highlighting for the first time that the regioselectivity of this process is both counter-ion and concentration dependent. Studies with a carbon-13 labelled substrate established that the reactions of alkynyliodonium salts are highly complex and that multiple mechanistic processes appear to be underway simultaneously.
As highly electron-deficient acetylenic species, alkynyliodonium salts are reactive partners in cycloaddition reactions3 including 1,3-dipolar cycloadditions11–14 and Diels–Alder chemistry.15–17 The hypernucleofuge nature of the iodoarene in alkynyliodonium salts also makes them excellent sources of carbenes,10 providing access to a wealth of cyclic species such as cyclopentenes18,19 and pyrroles;20 recently highly synthetically challenging cyanocarbenes have also been generated.21,22
In the same fashion, indoles, furopyridines, indenes, imidazopyridines, imidazopyrimidines, furotropones, furonaphthoquinones, thiazoles, selenazoles and benzofurans (Scheme 1) may also be formed.2,4 Cumulatively, these ring systems account for a wide range of known pharmacophores, yet the potential for alkynyliodonium salts in the preparation of heterocycles remains to be exploited.
![]() | ||
| Scheme 1 Heteroaromatics from alkynyliodonium salts.2,4 | ||
Commonly used anions in alkynyliodonium salts include mesylates,23 tosylates,24,25 triflates26,27 and tetrafluoroborates28,29 due to their low nucleophilicity, though in many of these cases addition of the anion to the β-acetylenic position was still observed.30 As such the development of alkynyliodonium salts with an intramolecular anion has been an area of continuing interest as a means of restraining the addition reaction.30–36 In contrast, alkynyliodonium trifluoroacetates (TFA) have received little attention,3,37 though it was recently shown by us that they are not only readily prepared38 but also available on a large scale (>0.5 mol).39
Despite the wealth of alkynyliodonium salts reported to date, very little is known about the effects of the counter-ion used or their solution state behaviour. A novel comparison is presented herein between the TFA salts and some of the more common alkynyliodonium salts, demonstrating for the first time that the anion used imparts a profound effect on the regioselectivity of arylimidazo[1,2-a]pyridine formation; it was also found that the outcome of the reaction could be manipulated through substrate concentration. This study highlights a dynamic solution state for alkynyliodonium salt derivatives and that, through control of experimental conditions, access to a plethora of substituted heteroaromatics may be achieved, with alternative regioisomers possible from the same starting material.
In 2004 Liu and co-workers reported the synthesis of a range of 2-arylimidazo[1,2-a]pyridines (cf. 3a) from the reaction of alkynyliodonium tosylates and 2-aminopyridine.40 Surprisingly, the analogous reaction of 1a afforded both 2- and 3-substituted imidazo[1,2-a]pyridines in roughly equal amounts (Scheme 2), as confirmed by X-ray crystallography (ESI† and Fig. 1).
![]() | ||
| Scheme 2 Synthesis of imidazo[1,2-a]pyridines from 1a. | ||
![]() | ||
| Fig. 1 Structures of the cations of 3a·HCl·H2O (left) and 4a·HCl·2H2O (right), with 40% probability displacement ellipsoids; N atoms are shown with shading. | ||
This intriguing production of two regioisomers led us to repeat the procedure reported by Liu and co-workers40 using the tosylate, 5, to confirm that just one regioisomer had indeed been produced. In our hands, the experiment showed that the major product of the reaction was in fact 4a and that, although this was dominant, trace amounts of 3a were also present (2%, 3a
:
44%, 4a); comparison of the 1H-NMR data with other reports supports this regiochemical assignment.41–44 Having optimized the reaction shown in Scheme 2 for a range of solvents and bases (see ESI†), fluorobenzene (PhF) (to minimise intermolecular insertion) and K2CO3 were chosen respectively, and at room temperature to minimize decomposition of the alkynyliodonium salts.
Heterogeneous bases and aprotic,45 non-coordinating solvents gave the best results. Despite slightly lower yields, PhF was chosen as it resulted in a ‘cleaner’ reaction facilitating isolation of the products. Comparison with other alkynyliodonium salts showed that 3a was undetectable using the triflate (6), whereas the cyclic iodane, 7 (see ESI†), produced a ratio of products between that observed for the trifluoroacetate and tosylate salts (Table 1). Only the phenyliodonium derivatives were investigated since variation of the second aromatic ring was previously found to have no effect in the preparation of 2-arylfuro[3,2-c]pyridines.38
Such a dependence on the anion used has never been reported previously in the formation of heteroaromatics from alkynyliodonium salts, though the literature suggests that the conversion of alkynyliodonium salts to alkenyliodonium salts appears to demonstrate stereoselectivity for the E- or Z-isomer depending on the counter-ion used.45–55 The influence of solvent on ion pair separation, and hence choice of counter-ion, has also been shown to affect the reactions of diaryliodonium salts.56
In addition to the counter-ion dependence of the reaction it was found that the concentration of the reactants also exerted an unexpected degree of control over the product distribution (Table 2 and Fig. 2).
![]() | ||
| Fig. 2 Concentration dependence for the reaction of 1a and 2. | ||
It should be noted that 1a was soluble in DCM and CHCl3 at all the concentrations tested, though not in PhF; however, in all three solvents, rapid dissolution of 1a was observed on addition of 2. Under the reaction conditions outlined in Scheme 1 the K2CO3 remained as a solid throughout. This suggests that the key reactive species is generated in situ as the same concentration dependence was observed for all three solvents (see ESI†).
Although some increase in yield was observed at higher dilutions, far more noticeable was the concentration-dependent regioselectivity favouring the 2-substituted regioisomer, 3a. To confirm this unexpected dependence, reactions were conducted using five different iodonium salt concentrations (Table 2 and Fig. 2). This ‘tuning’ of the reaction conditions has potential value, for example in drug discovery, where both regioisomers are accessible from a single set of reagents.
To establish whether the observed concentration dependence was restricted to the production of phenylimidazo[1,2-a]pyridines (3a and 4a) several alternative alkynyliodonium salts were also studied (1b–1d) and it was pleasing to note that a similar trend was found (Table 3).
The mechanism of imidazopyridine formation presented by Liu40 follows that previously reported by Wipf57 (and later Togo30) for the formation of thiazoles. An alternative route to the observed product has also been proposed by Ochiai.58 All of these options invoke a monomeric form of the aryl(alkynyl)iodonium salt as the starting species even though kinetic and spectroscopic evidence for other hypervalent iodine compounds has been reported that indicates the presence, in solution, not only of associated counter-ions, but also of higher-order structures (dimers, oligomers etc.).59–61 Such species have also been shown to be highly concentration-dependent60,61 and therefore contributions from these structures cannot be ruled out.
In addition these proposals rapidly result in loss of the counter-ion. However, to retain the influence of this component, and taking into account these prior mechanistic studies, we propose that intermediates such as 9 and 14 (Fig. 3) and 8 and 13 (Fig. 4) should also be considered in the mechanistic rationale since both the amino- and pyridinyl-nitrogen atoms of 2-aminopyridine are viable nucleophiles62 (Scheme 3: there may also be influence of the counter-ion in the subsequent steps due to the charged nature of the proposed intermediates).
![]() | ||
| Fig. 3 Proposed [10-I-4] intermediates, 14 and 18. | ||
![]() | ||
| Fig. 4 Potential intermediates of Michael addition. | ||
![]() | ||
| Scheme 3 Distribution of products from the reaction of [7′-1313C]-1. | ||
Further complexity is introduced following alkylidene carbene formation (the carbene can either cyclize directly or undergo 1,2-migration prior to cyclization), resulting in the formation of 3a or 4avia several different pathways.
As the acetylenic products of 1,2-migration have been reported to be highly reactive,58,63 especially within a basic environment, they may prove difficult to observe and as such we prepared the isotopically labelled [7′-1313C]-1a to investigate the process.
This preliminary 13C-labelling study generated [2-1313C]-3a and [3-1313C]-4a as expected (Scheme 3); however, the isotopomer [3-1313C]-3a was also present, highlighting that a competitive 1,2-migration was occurring (Scheme 4) and suggesting that at least three reaction pathways are in operation.
![]() | ||
| Scheme 4 Mechanism of isotopomer formation; [3-1313C]-3a. | ||
This new-found understanding is being applied to the preparation of a diverse range of heterocyclic ring systems which are of interest to our drug discovery programmes. Further work to resolve and differentiate the many mechanistic options is on-going.
CAUTION: Some hypervalent iodanes are potentially explosive and should be handled taking appropriate precautions.64–67
:
0 hexane–ether for 5 min then increasing to 3
:
7 over 120 min and holding at this solvent mixture until elution was complete) to give the product as a white crystalline solid (0.24 g, 1.22 mmol, 49%). Mp 132–133 °C (from DCM–petrol) (lit.,70 136–137 °C from cyclohexane); Rf 0.55 (4
:
1 ether–petrol); Found: C, 80.4; H, 5.3; N, 14.4. Calc. for C13H10N2: C, 80.4; H, 5.2; N, 14.4%; IR νmax/cm−1 (neat) 3130, 1632, 1502, 1475, 1447, 1369, 1353, 1304, 1273, 1246, 1203, 1145, 1077, 1027; δH (300 MHz, CDCl3; Me4Si) 8.10 (1H, d, H5 J 6.9), 7.97 (2H, d, H2′/H6′ J 7.2), 7.85 (1H, s, H3), 7.65 (1H, d, H8 J 9.0), 7.45 (2H, tapp., H3′/H5′ J 7.5), 7.34 (1H, t, H4′ J 7.5), 7.17 (1H, tapp., H7 J 6.9), 6.77 (1H, tapp., H6 J 6.0); δC (75 MHz, CDCl3; Me4Si) 146.41 (C2), 146.09 (C9), 134.27 (C1′), 128.99 (C3′/C5′), 128.27 (C4′), 126.55 C2′/C6′), 125.86 (C5), 124.79 (C7), 118.00 (C8), 112.66 (C6), 108.38 (C3); m/z (CI) 195 ([M + H]+, 100%), 95 (3), 80 (2), 52 (4). Found: [M + H]+, 195.0917. C13H11N2 requires 195.0917.
:
1 ether–petrol); Found: C, 80.3; H, 5.1; N, 14.3. Calc. for C13H10N2: C, 80.4; H, 5.2; N, 14.4%; IR νmax/cm−1 (neat) 1634, 1603, 1540, 1499, 1480, 1450, 1442, 1352, 1296, 1272, 1262, 1175, 1148, 1134, 1074, 1009; δH (400 MHz, d6-DMSO; Me4Si) 8.45 (1H, d, H5 J 6.9), 7.72 (1H, s, H2), 7.61 (1H, d, H8 J 8.7), 7.57 (2H, d, H2′/H6′ J 7.3), 7.46 (2H, tapp., H3′/H5′ J 7.8), 7.34 (1H, t, H4′ J 7.4), 7.21 (1H, tapp., H7 J 7.8), 6.86 (1H, tapp., H6 J 6.6); δC (100 MHz, d6-DMSO; Me4Si) 146.06 (C9), 133.07 (C2), 129.74 (C3′/C5′), 129.42 (C1′), 128.37 (C4′), 127.97 (C2′/C6′), 125.58 (C3), 125.00 (C7), 124.45 (C5), 118.09 (C8), 113.30 (C6); m/z (ESI) 195 ([M + H]+, 100%). Found: [M + H]+, 195.0905. C13H11N2 requires 195.0922.
:
1 ether–petrol); Found: C, 80.9; H, 5.7; N, 13.4. Calc. for C14H12N2: C, 80.7; H, 5.8; N, 13.5%.; IR νmax/cm−1 (neat) 3132, 1633, 1506, 1483, 1372, 1349, 1268, 1245, 1202, 1139; δH (500 MHz, CDCl3; Me4Si) 8.05 (1H, dtapp., H5 J 6.8, J 1.2), 7.84 (2H, d, H3′/H5′ J 8.1), 7.77 (1H, s, H3), 7.60 (1H, dd, H8 J 9.1, J 0.8), 7.23 (2H, d, H2′/H6′ J 8.1), 7.12 (1H, ddd, H7 J 9.1, J 6.8, J 1.3), 6.71 (1H, dtapp., H6 J 6.8, J 1.1), 2.38 (3H, s, Me); δC (125 MHz, CDCl3; Me4Si) 145.86, 145.55, 137.72, 130.89, 129.36, 125.90, 125.45, 124.41, 117.36, 112.21, 107.69, 21.22; m/z (ESI) 209 ([M + H]+, 100%). Found: [M + H]+, 209.1071. C14H13N2 requires 209.1073.
:
1 ether–petrol); Found: C, 80.9; H, 5.7; N, 13.4. Calc. for C14H12N2: C, 80.7; H, 5.8; N, 13.5%; IR νmax/cm−1 (neat) 2981, 1634, 1545, 1490, 1353, 1295, 1255, 1166, 1148, 1013; δH (500 MHz, CDCl3; Me4Si) 8.30 (1H, dtapp., H5 J 6.9, J 1.2), 7.66 (1H, overlapped s, H2), 7.65 (1H, overlapped d, H8 J 8.0), 7.44 (2H, d, H3′/H5′ J 8.0), 7.32 (2H, dd, H2′/H6′ J 8.0, J 0.6), 7.17 (1H, ddd, H6 J 9.1, J 6.9, J 1.3), 6.78 (1H, tdapp., H7 J 6.9, J 1.2), 2.43 (3H, s, Me); δC (125 MHz, CDCl3; Me4Si) 145.99, 138.13, 132.28, 129.88, 128.00, 126.36, 125.73, 123.94, 123.34, 118.21, 112.34, 21.27. m/z (ESI) 209 ([M + H]+, 100%). Found: [M + H]+, 209.1072. C14H13N2 requires 209.1073.
:
1 ether–petrol); Found: C, 66.1; H, 3.9; N, 13.8. Calc. for C11H8N2S: C, 66.0; H, 4.0; N, 14.0%; IR νmax/cm−1 (neat) 3124, 1632, 1508, 1476, 1338, 1306, 1272, 1242, 1144, 1090; δH (500 MHz, d6-DMSO; Me4Si) 8.49 (1H, d, H5 J 6.7), 8.23 (1H, s, H3), 7.89 (1H, d, H2′ J 2.8), 7.61–7.55 (2H, m, H4′/H5′), 7.54 (1H, d, H8 J 9.0), 7.21 (1H, tapp., H7 J 6.6), 6.86 (1H, tapp., H6 J 6.7); δC (125 MHz, d6-DMSO; Me4Si) 144.94, 141.44, 136.38, 127.16, 127.07, 126.46, 125.15, 121.43, 116.76, 112.46, 109.31; m/z (ESI) 201 ([M + H]+, 100%). Found: [M + H]+, 201.0480. C11H9N2S requires 201.0481.
:
1 ether–petrol); IR νmax/cm−1 (neat) 3090, 1690, 1637, 1576, 1501, 1483, 1343, 1330, 1299, 1264, 1225, 1169, 1154, 1128, 1087, 1019; δH (500 MHz, d6-DMSO; Me4Si) 8.61 (1H, d, H5 J 7.0), 7.93 (1H, dd, H2′ J 1.7, J 1.3), 7.85 (1H, s, H2), 7.76 (1H, dd, H5′ J 5.0, J 2.1), 7.64 (1H, d, H8 J 8.5), 7.54 (1H, dd, H4′ J 5.0, J 1.3), 7.29 (1H, ddd, H7 J 8.5, J 6.7, J 1.7), 6.99 (1H, tdapp., H6 J 6.8, J 1.1); δC (125 MHz, d6-DMSO; Me4Si) 145.18, 132.58, 128.97, 127.20, 127.08, 124.57, 124.27, 121.18, 121.06, 117.40, 112.88; m/z (ESI) 201 ([M + H]+, 100%). Found: [M + H]+, 201.0479. C11H9N2S requires 201.0481.
:
1 ether–petrol); Found: C, 57.3; H, 3.2; N, 10.1. Calc. for C13H9BrN2: C, 57.2; H, 3.3; N, 10.3%; IR νmax/cm−1 (neat) 2955, 2924, 1679, 1635, 1428, 1401, 1371, 1321, 1203, 1065, 1006; δH (500 MHz, d6-DMSO; Me4Si) 8.51 (1H, dtapp., H5 J 6.8, J 1.2), 8.42 (1H, s, H3), 7.91 (2H, d, H3′/H5′ J 8.7), 7.61 (2H, d, H2′/H6′ J 8.7), 7.56 (1H, dd, H8 J 9.1, J 1.0), 7.24 (1H, ddd, H7, J 9.1, J 6.8, J 1.3), 6.89 (1H, tdapp., H6 J 6.8, J 1.0); δC (125 MHz, d6-DMSO; Me4Si) 144.82, 143.13, 133.16, 131.55, 127.47, 126.87, 125.11, 120.59, 116.61, 112.34, 109.42; m/z (ESI) 275 ([81Br][M + H]+, 97%), 273 ([79Br][M + H]+, 100%). Found: [M + H]+, 273.0026. C13H10BrN2 requires 273.0022.
:
1 ether–petrol); Found: C, 57.1; H, 3.2; N, 10.2. Calc. for C13H9BrN2: C, 57.2; H, 3.3; N, 10.3%; IR νmax/cm−1 (neat) 3023, 1537, 1499, 1478, 1398, 1351, 1303, 1291, 1264, 1174, 1151, 1101, 1074, 1007; δH (500 MHz, CDCl3; Me4Si) 8.25 (1H, dtapp., H5 J 7.0, J 1.2), 7.67 (1H, s, H2), 7.66 (1H, d, H8 J 9.1, J 1.1), 7.62 (2H, d, H2′/H6′ J 8.6), 7.40 (2H, d, H3′/H5′ J 8.6), 7.19 (1H, ddd, H7 J 9.1, J 6.7, J 1.3), 6.80 (1H, tdapp., H6 J 6.8, J 1.1); δC (125 MHz, CDCl3; Me4Si) 146.24, 132.68, 132.38, 129.35, 128.15, 124.49, 124.39, 123.06, 122.05, 118.29, 112.76; m/z (ESI) 275 ([81Br][M + H]+, 98%), 273 ([79Br][M + H]+, 100%). Found: [M + H]+, 273.0026. C13H10BrN2 requires 273.0022.
:
0 hexane–ether for 5 min then to 3
:
7 over 120 min and holding at this solvent mixture until elution was complete), loading the sample in DCM, to give the products as a white, crystalline solids; 2-Phenyl-2/3-[13C]-imidazo[1,2-a]pyridine ([131313C]-3a) (0.03 g, 0.14 mmol, 20%) and 3-Phenyl-3-[13C]-imidazo[1,2-a]pyridine ([1313C]-4a) (0.04 g, 0.21 mmol, 30%).
:
1 ether–petrol); δH (400 MHz, CDCl3; Me4Si) 8.09 (1H, dt, H5 J 6.8, J 1.2), 7.97–7.93 (2H, m, H2′/H6′), 7.84 (0.82H, dd, H3 J 8.4, J 0.8), 7.84 (0.18H, d, H3 J 190.7) 7.63 (1H, d, H8 J 9.2), 7.43 (2H, tapp., H3′/H5′ J 8.0), 7.32 (1H, t, H4′ J 7.6), 7.15 (1H, ddd., H7 J 9.2, J 6.6, J 1.2), 6.75 (1H, dtapp., H6 J 6.6, J 0.8); δC (100 MHz, CDCl3; Me4Si) 145.82 (C2-label), 145.47 (d, C9 J 4.5), 133.76 (d, C1′ J 67.9), 133.76 (s, C1′), 128.82 (d, C3′/C5′ J 4.4), 128.82 (s, C3′/C5′), 128.08 (s, C4′), 126.14 (d, C2′/C6′ J 2.5), 125.66 (d, C5 J 7.9), 125.66 (s, C5), 124.79 (C7), 117.62 (d, C8 J 6.1), 117.62 (s, C8) 112.55 (C6), 108.21 (C3-label); m/z (CI) 196 ([M + H]+, 100%), 184 (12). Found: [M + H]+, 196.0947. C1213C1H11N2 requires 196.0950.
:
1 ether–petrol); δH (400 MHz, CDCl3; Me4Si) 8.31 (1H, dd, H5 J 6.8, J 1.2), 7.68 (1H, od, H2 J 12.4), 7.65 (1H, od, H8 J 8.8), 7.55–7.52 (2H, m, H2′/H6′), 7.49 (2H, tapp., H3′/H5′ J 8.0), 7.39 (1H, tt, H4′ J 7.2, J 1.2), 7.17 (1H, ddd., H7 J 9.2, J 6.4, J 1.2), 6.86 (1H, tapp., H6 J 6.8); δC (100 MHz, CDCl3; Me4Si) 146.21 (d, C9 J 9.1), 132.57 (d, C2 J 69.0), 132.52 (s, C9), 129.40 (d, C1′ J 67.4), 129.33 (d, C3′/C5′ J 4.2), 129.33 (s, C3′/C5′), 128.27 (s, C4′), 128.11 (d, C2′/C6′ J 2.7), 125.84 (C3-label), 124.34 (s, C7), 123.43 (s, C5), 118.09 (C8), 113.30 (C6); m/z (CI) 196 ([M + H]+, 100%). Found: [M + H]+, 196.0945. C1213C1H11N2 requires 196.0950.
X-ray crystal structures are available for compounds 3a, 3a·HCl·H2O, 4a·HCl·2H2O and 7 (see ESI†). CCDC 907271–907274 contain the supplementary crystallographic data for this paper.
We thank sanofi-aventis, Newcastle University and the EPSRC (equipment grant EP/F03637X/1) for funding. We also thank Prof. W. McFarlane and Dr C. Wills for contributions in the field of NMR and the EPSRC National Mass Spectrometry Service Centre, Swansea, UK.
Footnotes |
| † Electronic supplementary information (ESI) available: Full experimental details and spectra, tables of X-ray crystallographic data and results. CCDC 907271–907274. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c3ob41112e |
| ‡ Current address: Department of Isotope Chemistry, Covance Laboratories, Alnwick, Northumberland, NE66 2JH, UK. |
| § Although several column packings were evaluated, e.g. reverse phase silica, alumina (neutral, basic and acidic), with a range of solvents and additives, the Grace cartridges were found to provide satisfactory purification. |
| ¶ Due to volatility, solvent could not be fully removed; yield calculated from 1H-NMR. |
| This journal is © The Royal Society of Chemistry 2013 |