Enrique
Gómez-Bengoa
,
Aitor
Landa
,
Aitziber
Lizarraga
,
Antonia
Mielgo
,
Mikel
Oiarbide
and
Claudio
Palomo
*
Departamento de Química Orgánica I, Universidad del País Vasco, Apdo. 1072, 20080, San Sebastián, Spain. E-mail: claudio.palomo@ehu.es; Fax: (+) 943015270
First published on 25th October 2010
A novel method for the direct, amine-catalyzed, highly enantioselective α-alkylation of aldehydes is described that is founded upon the use of electrondeficient allylic halides as alkylating agents and DABCO or DMAP as coadyuvant. Both experimental observations and DFT calculations, in support of a SN2′-type addition-elimination pathway involving ammonium salt intermediates, are provided.
Recent progress in this area has been made using the potential of aminocatalysis and its adaptability to specific reaction mechanisms, from intramolecular SN2 to intermolecular radical, and SN1 pathways (Scheme 1).4 In 2004, Vignola and List reported α-methyl proline-catalyzed intramolecular reactions of ω-haloaldehydes to provide three and five member carbo- and heterocycles enantioselectively.5,6 More recent studies carried out independently by the groups of MacMillan,7 Melchiorre and Petrini,8 Cozzi,9 and Jacobsen,10 have not only provided novel methods for the aminocatalytic, asymmetric α-alkylation of aldehydes,11 but have also demonstrated the compatibility of enamine activation with radical and cationic (SN1) reaction pathways. These innovative methods represent a clear advance, but often require more sophisticated alkylating reagents/catalytic systems or are restricted to a certain class of alkylating electrophiles. In this Edge article, we describe an amine-catalyzed asymmetric α-alkylation of aldehydes that follows a distinct mechanism, based on the use of electrondeficient allylic bromides as alkylating reagents.
Scheme 1 Current, mechanistically complementary, solutions to the catalytic asymmetric α-alkylation of unmodified aldehydes. |
Scheme 2 Amine-catalyzed α-alkylation of aldehydes with allyl halides A, assuming a reversible N–alkylation side reaction (EWG: electron withdrawing group). |
Initial attempts for experimental validation of the above hypothesis using 2-(halomethyl)acrylates unravelled two additional difficulties for this challenging transformation to overcome; namely (a) the inefficacy of simple acrylate esters towards the addition reaction of transient enamine species14 and (b) the known tendency of α-branched aldehydes to undergo base-promoted racemization. Thus, Scheme 3, attempted the reaction of hydrocinnamaldehyde 1a (R = CH2Ph) with allyl bromide 2, catalyzed by the chiral pyrrolidine 5a15 (20 mol %) in the absence of any base, or the presence of 1 equivalent triethylamine (TEA) or diisopropylethylamine (DIPEA) as acid scavengers, led to an unaltered starting aldehyde and halide (Table 1, entries 1–3).
Scheme 3 Direct enantioselective α-alkylation of aldehydes with 2-(bromomethyl) acrylates. |
Entry | R33N | Cat | T/°C | t/h | Conv. (%)b | ee (%)c |
---|---|---|---|---|---|---|
a General conditions: 1a (1.5 mmol), 2 (0.5 mmol), R33N (0.51 mmol), catalyst 5 (20 mol %) were stirred in CH2Cl2 (2 mL) at the indicated temperature. b Determined by 1H NMR. c Determined by chiral HPLC of crude product before chromatography. NR: no reaction. | ||||||
1 | -- | 5a | 20 | 24 | NR | -- |
2 | TEA | 5a | 20 | 24 | NR | -- |
3 | DIPEA | 5a | 20 | 24 | NR | -- |
4 | DBU | 5a | 20 | 24 | >95 | 0 |
5 | DABCO | 5a | 20 | 24 | >95 | 81 |
6 | DMAP | 5a | 20 | 24 | >95 | 79 |
7 | DMAP | 5a | 0 | 24 | 95 | 85 |
8 | DMAP | 5b | 0 | 36 | 70 | 89 |
9 | DMAP | 5b | –10 | 48 | >95 | 93 |
Further screening of other stoichiometric amine bases eventually identified DBU as a successful promoter of the reaction, to afford the alkylated product 6a, albeit as a racemic product (entry 4). The null enantioselection observed was primarily attributed to the strong basicity of DBU in connection to problem (b). Most importantly, the dramatic effect in reactivity observed, suggested that the actual alkylating agent under these reaction conditions might likely be the corresponding ammonium salt derived from halide A and DBU,16 rather than the halide itself. Based on this assumption, the alkylation reaction was examined using DABCO and DMAP, which have been shown by Mayr to possess optimal nucleophilicity/nucleofugacity while being less basic than DBU.17 Gratifyingly, when the above reaction was carried out in the presence of 1 equivalent of either DABCO or DMAP, the alkylated aldehyde 6a was formed after 24 h at room temperature, with ee's of 81% and 79%, respectively (entries 5 and 6). Further improvement of the ee was attained by decreasing the reaction temperature (0 °C, entry 7) or by shifting to the catalyst 5b (entries 8 and 9).15c
Exploration of the scope of the reaction with respect to the aldehyde substrate, revealed satisfactory results for a range of aldehydes with varying steric demand and functionality, Table 2. Thus, aldehydes with linear alkyl chains of variable length afforded good yields and ee's uniformly above 90% (entries 1–6); slightly lower reactivity was shown by branched chain aldehydes, but still high selectivity was achieved (entry 7). Similarly, a range of functional groups, including alkene, ether, ester, acetal, and carbamate, are tolerated (entries 8–15), although the alkoxyaldehyde 1i gave diminished ee for reasons not yet well understood (entry 12). In general, DABCO affords similar or even slightly higher ee’s than DMAP at r.t. or 0 °C, but at lower temperatures DMAP showed superior ee’s for most cases.18 Interestingly, α-alkylation of the chiral aldehyde citronellal took place with remarkable diastereoselectivity for the matched pair (entry 11). On the other hand, control experiments using the reaction adducts were informative with respect to their configurational robustness against acidic or basic conditions. For example, the measured values of the ee before and after silica gel chromatography for adduct 6a were essentially identical. No appreciable variation of ee was either observed even after stirring a solution of 6a in CH2Cl2 admixed with silica gel at room temperature for 16 h. In turn, while exposure of 6a and 6d to DMAP or DABCO (5 mol%) at 0 °C overnight did not affect their configurational integrity, essentially complete racemization took place in the presence of DBU (5 mol%) after 2–3 h at the same temperature. A brief exploration of the influence of the size of the ester group on the reaction outcome, indicated that methyl ester 3 is equally suitable than the parent ethyl ester (compare entries 1 and 16); however, the tert-butyl ester 4 showed reduced reactivity at −10 °C to produce 8a, and at room temperature the reaction was feasible using either DMAP or DABCO, although at the expense of a selectivity decrease (entry 17).
Entry | R | Product | T/°C | t/h | Yield (%)a | ee (%)b |
---|---|---|---|---|---|---|
a Yield of isolated product after chromatography. In parenthesis, results using DABCO instead of DMAP. b Determined by chiral HPLC. c Ratio of diastereomers. d Conversion after 52 h of reaction. | ||||||
1 | CH2Ph | 6a | 0 | 48 | 56 (45) | 90 (94) |
2 | –10 | 48 | 59 | 93 | ||
3 | CH2CH3 | 6b | 0 | 48 | 61 (50) | 86 (91) |
4 | –10 | 48 | 63 | 92 | ||
5 | (CH2)3CH3 | 6c | 0 | 48 | 64 (49) | 85 (92) |
6 | –10 | 48 | 61 | 94 | ||
7 | CH(CH3) 2 | 6d | 0 | 60 | 54 | 90 |
8 | CH2CH = CH2 | 6e | –20 | 60 | 63 | 95 |
9 | (CH2)2CH = CH2 | 6f | –20 | 60 | 69 | 93 |
10 | (R)-Citronellal | 6g | 20 | 48 | 68 | 75:25c |
11 | (S)-Citronellal | 6h | 20 | 48 | 70 | 97:3c |
12 | CH2OBn | 6i | –10 | 60 | 59 | 68 |
13 | (CH2)2CO2Me | 6j | –10 | 60 | 62 | 90 |
14 | CH2 CH(OMe)2 | 6k | –10 | 60 | 58 | 87 |
15 | (CH2)4NHBoc | 6l | –10 | 60 | 62 | 94 |
16 | CH2Ph | 7a | –10 | 60 | 57 | 95 |
17 | CH2Ph | 8a | 20 | 52 | 80d (90)d | 85 (79) |
The orthogonal functionality of the resulting alkylated adducts provides versatility to this catalytic methodology. For instance, Scheme 4, reduction of aldehyde 6e and smooth lactonization of the resulting alcohol, afforded a known compound 10,19 which served to establish the stereochemistry of the catalytic reactions. Further potential of the method could be inferred from the ring closing metathesis of adducts 6e and 6f, using Grubbs' catalyst 9,20 to yield the corresponding 5- and 6-membered carbocycles 11e and 11f, in high enantioselectivity.21
Scheme 4 Elaboration of adducts into hetero- and carbocycles. |
Complementing these experiments, which serve as a proof of concept of the validity of halides of type A as suitable electrophiles for the difficult to control amine-catalyzed asymmetric α-alkylation of aldehydes,2,4 the reaction mechanism was briefly investigated computationally. In this respect, DFT calculations for the model reaction involving the methyl acrylate 3, DMAP, and the enamine derived from acetaldehyde and pyrrolidine, suggest an asynchronous SN2′ pathway for the initial step leading to the ammonium salt 12 (Fig. 1a), with the C–N bond formation preceding the C–Br bond breaking. For the subsequent reaction of 12 with the enamine, the energy profile (Fig. 1b) indicates a two step addition-elimination sequence, with 13 being the intermediate formed after the rate limiting addition step.23,24 The alternative direct displacement (SN2) pathway for both the C–N and C–C bond forming processes was also calculated, resulting in much higher energy barriers in both cases.25 Consistent with the above reaction scheme is the fact that treatment of butyraldehyde with the preformed ammonium salt from 2 and DMAP provided, in the presence of 20 mol% of catalyst 5b, alkylated adduct 6b in essentially the same yield and ee as in entry 3 of Table 2.
Fig. 1 Reaction Coordinate Diagram for the model reaction computed at the B3LYP/6–31* level (enthalpies given in kcal mol−1) corresponding to: (a) initial formation of intermediate ammonium salt 12 (asynchronous SN2′); (b) key C–C bond forming step (two-step addition-elimination SN2′-type mechanism). |
Finally, it was found that substitution at the β-position of the 2-(bromomethyl)acrylate exerts a notable effect upon reactivity. As the example in Scheme 5 shows, the reaction of butyraldehyde with 14 in the presence of DMAP was extremely slow, but with DABCO proceeded to give product 15 in moderate yield after prolonged reaction time. Significantly, 15 was the only regioisomer detected, which was obtained in remarkable diastereo- and enantioselectivity.22
Scheme 5 Regio-, diastereo-, and enantioselective α-alkylation of aldehydes with β-substituted α-bromomethyl acrylate 14. |
Footnote |
† Electronic Supplementary Information (ESI) available: Experimental procedures, spectral data for all new compounds, stereochemical proofs, and detailed computational data. See DOI: 10.1039/c0sc00402b |
This journal is © The Royal Society of Chemistry 2011 |