Note pubs.acs.org/joc
Catalytic Alkynylation of Cyclic Acetals and Ketals Enabled by Synergistic Gold(I)/Trimethylsilyl Catalysis Mathéo Berthet, Olivier Songis, Catherine Taillier, and Vincent Dalla* UNIHAVRE, FR 3032, URCOM EA 3221, 76600 le Havre, 25 rue Philippe Lebon, BP 540, 76058-F Le Havre, France S Supporting Information *
ABSTRACT: A completely regioselective and challenging gold(I)-catalyzed ring-opening of cyclic 1,3-dioxolanes and dioxanes by trimethylsilyl alkynes to set diol-derived propargyl trimethylsilyl bis-ethers is reported. This unprecedented and not trivial transformation does not operate with the catalytic methodologies recently reported for catalytic alkynylation of acyclic acetals/ ketals, and is uniquely enabled by the application of a recently introduced synergistic gold(I)−silicon catalysis concept capable of producing simultaneously catalytic amounts of two key players, a silicon-based Lewis superacid and a nucleophilic gold acetylide.
T
Scheme 1. Current State-of-the-Art for Catalytic Alkynylation of Acyclic Acetals, for TiCl4-Mediated Alkynylation of Cyclic Acetals and Examples of Bioactive Compounds Bearing a Functionalized Propargylic Diol Skeleton
aken separately, the incorporation of an alkynyl group into organic molecules1 and the transformations of acetals under an acidic activation mode2 are foundational to modern organic chemistry. In addition to the stimulating academic challenge of confronting these two topics, the great synthetic utility of the expected propargyl ether products has prompted several research groups to develop new strategies to enable highly effective SN1 alkynyl transfers from acetals.3 To tackle the strong current demand in developing safer, atomeconomical, and sustainable methodologies, catalytic approaches for these transformations constitute an emerging research theme. Accordingly, a few number of methodologies enabling the direct and catalytic alkynylation of acyclic acetals and ketals by C−H activation of terminal alkynes,4,5 including impressive enantioselective variants by Watson and co-workers,6 were reported quite recently (Scheme 1a). Similar transformations of cyclic acetals and ketals have been comparatively much less studied. As far as we are aware, only a handful of papers have been reported, and all mention use of trimethylsilyl alkynes and more than one equivalent of the strong Lewis acid TiCl4 as a mediator (Scheme 1b).2a,7 One of the most plausible reasons accounting for the absence of catalytic variants is the energetic cost associated with the ringopening of cyclic acetals, which combined with the weak nucleophilicity of terminal and TMS alkynes,8 renders a catalytic transformation particularly defiant.9 In addition to the inherent challenge of accommodating cyclic acetals/ketals to catalytic alkynylation, the potential synthetic value of the diol monopropargylic ether products, which comprise a number of important molecular targets (Scheme 1c) on one hand,10,11 while constituting a promising and yet so far under-appreciated platform for further elaborations on the other hand, is also © 2017 American Chemical Society
noteworthy. Whereas the targeted compounds could in principle also be accessible by the addition of a metallic acetylide on protected hydroxy aldehydes/ketones, the latter are available by relatively lengthy routes from the corresponding diols, and thus our approach planning the use of easily Received: July 21, 2017 Published: August 16, 2017 9916
DOI: 10.1021/acs.joc.7b01828 J. Org. Chem. 2017, 82, 9916−9922
Note
The Journal of Organic Chemistry available acetals/ketals (most of them being commercially available) establishes an additional improvement besides the catalytic aspect. These considerations prompted us to tackle the challenge of developing the first catalytic alkynylation of cyclic acetals and ketals. To reach this goal we sought to exploit a synergistic gold(I)−silicon catalysis concept recently introduced.12−14 The latter is characterized by two interlaced catalytic cycles capable of producing in situ and simultaneously small amounts of a silicon-based Lewis superacid dedicated to the activation of a pro-electrophile, and a gold alkynylide intended for trapping the resulting intermediate promptly due to an improved nucleophilicity compared with the parent alkynylsilane. We have launched this concept by reporting the first catalytic alkynylation of N,O-acetals.15,16 Assuming that our approach is transposable to a wider range of transformations, and thus bears the potential to become a truly general catalytic alkynylation methodology, we next identified the non trivial and challenging alkynylative ring-opening of cyclic acetals/ketals as new platform (Scheme 2).17 We now report the successful realization of these ideas to provide diol monopropargylic ethers with perfect regiocontrol and moderate diastereocontrol.
Table 1. Optimization of the Reaction Conditions for the Alkynylation of Model Acetals 2a and 2b
Scheme 2. Synergistic Gold(I)−Silicon Catalysis Concept (Left) and Synthetic Applications (Right)
a
entry
solvent
T (°C)
t (h)
1 2 3 4 5 6 7 8 9 10 11 12 13 14
C6D6 C6D6 C6D6 CDCl3 CHCl3 CDCl3 C6D6 CDCl3 CDCl3 CDCl3 CDCl3 CDCl3 CDCl3 CDCl3
85 85 50 85 85 85 85 85 50 r.t 50 85 50 50
3 5 24 18 24 48 1 1 h 30 6 4d 21 5 h 30 18 18
NMR yield of 3 (%)a
yield of 4 (%)b
100 100 97 100
92 81c 75 92 75 49d 58 37 51 38
70 70 48 70 55 30e 62 100 100
57d 90d 92d,f
NMR yield related to the internal standard 1-bromo,4-nitrobenzene. Yield of isolated product. cUsing 1.5 equiv of trimethylsilylphenylacetylene. dUsing 1 mol% of IPrAuCl and 1 mol% of AgOTf. eUsing 1 mol% of JohnPhos-AuNTf2. fReaction executed from 3.8 mmol of 2b instead of 0.224 mmol for all the other experiments in Table 1. b
total conversion into the TMS-ether 3a. A brief survey of less toxic solvents was undertaken from which chloroform appeared to be a safer alternative (Table 1, entries 4−5). Moreover, we verified that the reaction did not occur without the gold precatalyst or by directly employing phenylacetylene (Scheme 1).4,5 These observations are particularly revealing of the decisive and unique ability of the association TMS-alkynesgold(I) triflate/triflimidate catalysts to promote difficult alkynylation reactions, thus supplying an additional credit to our concept of synergistic catalysis. By using the accordant IPr gold(I) triflate catalyst, the conversion remained incomplete even after 48 h (Table 1, entry 6). This result indicates that the stronger Lewis acidity of the TMSNTf2 catalyst is essential for an efficient alkynylation of 2a. The evaluation of the second model acetal substrate, the monosubstituted component 2b, revealed a markedly different behavior. The dioxolane 2b, being the progenitor of a more stabilized oxonium intermediary, displayed superior reactivity as well as sensitivity, as demonstrated by shorter reaction time, poor NMR and isolated yields of 4b under varying reaction conditions using gold(I)triflimidate catalysts (Table 1, entries 7−11). The latter family seems unsuitable to promote this reaction in an efficient fashion, due to the extreme reactivity of the ancillary TMSNTf2 Lewis acid catalyst which most likely causes decomposition of 2b or of its oxocarbenium ion. Accordingly, shifting to the IPrAuCl/AgOTf catalyst system which generates in situ the softer TMSOTf Lewis acid12 solved this problem, and delivered the product 4b in an excellent yield from a reaction run at 50 °C over 18 h (Table 1, entries 12−14). Finally, the alkynylation of 2b, which slightly differs from its counterpart 2a by one methyl substituent, succeeded in much milder reaction conditions. With such flexible and fully adaptable reaction conditions in hand and a more global comprehension of the reaction
To validate our working hypothesis, we began this study by testing the different methodologies recently reported for the catalytic alkynylation of the more reactive acyclic acetals,4−6 and as we had anticipated none of them was capable of promoting any conversion of the cyclic substrates (Scheme 3). Scheme 3. Absence of Reactivity of Common Cyclic Acetals Using Catalytic Alkynylation Procedures Described for Acyclic Acetals4
To initiate the optimization study of the projected synergistic gold(I)−silicon catalytic alkynylation reaction, we chose the nonsubstituted acetal 2a as a model reaction partner. 2a was first reacted with 1.9 equiv of trimethyl(phenylethynyl)silane 1a in the presence of 1 mol% of IPrAuNTf2 in deuterated benzene at 85 °C (Table 1, entry 1). To our delight, a rapid disappearance of 2a was observed, resulting in a clean and 9917
DOI: 10.1021/acs.joc.7b01828 J. Org. Chem. 2017, 82, 9916−9922
Note
The Journal of Organic Chemistry
our method is complementary of a related alkynylation of aldehydes published by Gonzales et al.13 In the latter case, the harsh reaction conditions required for the alkynyl transfer to the aldehyde function makes inevitable a second SN1-type alkynylation of the primary propargylic OTMS ether intermediate, finally delivering a bis-alkynylated compound where the ether function was destroyed. As for the alkynylation of (rac)-(4R,5R)-4,5-dimethyl-2phenyl acetal 2g and its stereochemically undefined congener (rac)-4-methyl-2-phenyl (d.r 45:55) 2h, the alcohol products 4g and 4h were isolated in a good yield of 78%, provided that an alternative in situ desilylation mediated by TBAF is applied due to the acid-sensitivity of the benzylic/propargylic ether adducts. In addition to being obtained in good yield and d.r., compound 4h also resulted from a completely regioselective reaction for the primary alcohol (vide inf ra). As briefly mentioned above, in some specific cases our alkynylation resulted in unsatisfactory outputs, since we were occasionally not even able to isolate products. Driven by the idea of transforming these sensitive hydroxylic reaction products 4 into more robust derivatives, thus improving the efficacy of the process, we sought to apply one-pot derivatizations of the alkynylation products as a way to solve the problem. Some years ago, our group was involved in the development of an efficient O-silyl into O-benzoyl interconversion mediated by benzoyl fluoride and catalytic DMAP.18 This highly selective transprotection that occurred in mild conditions and was compatible with a large range of functional groups, was particularly useful to accomplish a deprotection/ reprotection through a one-pot two-step sequence. The possibility to combine both alkynylation and transprotection in a sequential one-pot procedure was thus investigated. We were delighted to observe that after full conversion toward the TMS-alcohol ascertained by analytical monitoring, the catalytic trans-protection was successfully achieved across diverse TMSalkynes/acetals(ketals) combinations, and furnished the desired esters in decent to good yields (Scheme 5).19 As hoped, this
parameters, we then examined the substrate scope and functional group tolerance of the reaction (Scheme 4). Scheme 4. Scope of the Catalytic Alkynylation of Dioxolanes and Dioxanes
a
IPrAuCl and AgOTf (1 mol % each) were used as the catalyst system. The in situ desilylation was effected with TBAF instead of HCl. c1.2 equiv (instead of 1.9 equiv) of trimethylsilylphenylacetylene were used.
b
Scheme 5. Scope of the Sequential One-Pot Alkynylation of Acetals and trans-Protection into Benzoate Derivatives
This gold(I)-catalyzed alkynylation mostly focuses on unbiased and thus challenging acetals/ketals devoid of any aromatic or mesomeric stabilizing substituent at the anomeric carbon. In these cases, the method is effective for preparing a wide range of monopropargyloxy-1,2-diols (1,3-dioxolane series, see 4c−e and 4j−m) and 1,3-diols (case of 1,3-dioxane, see 4f) in moderate to excellent yields, which were optimized and varied depending on the structural pattern, and thus the inherent stability and sensitivity, of the alkyne/acetal combinations used. Of particular note, the TBDPS-protected alkynol derivative 1c (TBDPS = tert-butyl dimethyl silyl) exhibited an exquisite reactivity with 2a to deliver the target product 4k in 85% yield. In contrast, masking the alcohol by a benzoate ester of greater Lewis base properties (alkyne 1d) altered the reactivity somehow, leading to only 36% yield of the alkynylated compounds 4l. Acetals/ketals bearing a π-stabilizing phenyl substituent, including the more classical acyclic benzaldehyde dimethyl acetal and dioxolanes with branching substituents on the 4/5 positions, expectedly appeared to be highly reactive substrates to furnish the propargylic methyl ether product 4i and the diol derivatives 4g and 4h within short reaction times (
