Mechanistic Studies on the 1,2-Spin-Center Shift in Carbohydrate Systems with a Fluorenylcyclopropyl Radical Clock

The mechanism of the 1,2-spin-center shift in carbohydrate systems was studied with a fluorenylcyclopropyl radical clock. The 1,2-rearrangement of the acyl fluorenylcyclopropane group without opening of the cyclopropane ring provides the strongest evidence that the 1,2-spin-center shift in carbohydrate systems occurs through a concerted transition state without the intermediacy of a 1,3-dioxolanyl radical.

−3 The 1,2spin-center shift (1,2-SCS), also termed the Surzur−Tanner rearrangement, 4−6 has emerged as a reliable method for the functionalization of carbohydrates. 7,8A general scheme for the 1,2-SCS is depicted in Figure 1a: carbohydrate A undergoes C−X homolysis or group transfer to generate anomeric radical B. The anomeric radical is then translocated to the C2 position by the migration of the adjacent acyloxy group to the C1 position, which generates the more stabilized C2-alkyl radical C. Classically, a C2 radical would be reduced with an H atom donor to form deoxy sugars D (R 1 = H). 9,10Recent work has shown that the C2 alkyl radical can combine with excited-state palladium 11−13 or nickel 14 complexes to engage in a variety of transition-metal-catalyzed transformations to form C2-functionalized products with the general structure of D, where R 1 = H, D, I, CH 2 COR, or CH=CHR.
The mechanism of the 1,2-SCS (i.e., the transformation B → C, Figure 1a) has been investigated extensively since its initial discovery. 6,15The four possible mechanisms that have received the most attention are depicted in Figure 1b. 16The intermediacy of dioxolanyl radical I has been deemed unlikely based on radical clock experiments with acyl cyclopropanes.For example, under excited-state palladium conditions, glycosyl bromide 1 underwent a 1,2-SCS without opening of the cyclopropyl ring to generate 2 (eq 1). 11ESR spectroscopy 17,18 and experimental studies 19 on dioxolanyl radical intermediates 4 and 7 showed that cyclopropyl dioxolanyl radicals prefer to open to the cyclopropane ring over the dioxolane (eqs 2 and 3).Taken together, these experiments suggest that dioxolanyl radical intermediates such as I are not competent intermediates for 1,2-SCS in carbohydrates.These experimental findings have also been supported by computational studies. 20o account for the unlikelihood of a pathway that proceeds through dioxolanyl intermediate I, three concerted transition states have been proposed.A three-centered transition state such as II was proposed by Kocǒvsky, 21 but this mechanism was subsequently disproven (for nonconformationally restricted ester substrates) 22 by 18 O labeling studies showing that labeled glycosyl bromide 9 underwent clean transposition to afford deoxy sugar 10a (eq 4). 17A three-membered transition state like II would occur with retention of the 18 O label at the carbonyl oxygen atom, which would have led to the exclusive formation of 10b.The results from this 18 O labeling study provided evidence for a five-electron, five-membered concerted transition state such as III. 17,23oubt has been cast on transition state III by separate isotopic labeling studies with 17 O in the tetrahydropyranyl systems.When 17 O-labeled tetrahydropyran 11 was treated with tri-n-butyltin hydride, only 33% transposition of the carbonyl oxygen atom of the acyl group to the C2 carbon atom was observed (eq 5). 24This labeling experiment contradicts the possibility of a five-centered concerted transition state such as III. 24This discrepancy was reasoned by suggesting that the mechanism of 1,2-SCS is substrate-dependent and that tetrahydropyranyl radicals undergo 1,2-SCS, at least in part, through the three-membered transition state II or a chargeseparated transition state such as IV where scrambling of the label would be possible.The current explanation for the observed scrambling of the isotopic label is that the lack of substituents on tetrahydropyran 11 causes this substrate to rearrange via a looser charge-separated transition state (such as IV) than the transition state for the analogous glucosyl substrate 9, which has its charge-separated pathway inductively destabilized by adjacent acyloxy substituents. 25e sought to address the conflicting mechanistic picture for the 1,2-SCS in carbohydrate systems by formally excluding the possibility of I as a reactive intermediate using a fluorenylcyclopropyl radical clock.While radical clock experi-ments with unsubstituted acyl cyclopropyl substrates have been previously conducted for 1,2-SCS reactions of carbohydrates (for example, eqs 2 and 3), 11,18,19 these experiments only provide accurate information if the radical intermediate in question exists for longer than the rate of ring opening of an unsubstituted cyclopropane, which is roughly 1 × 10 8 s −1 at 25 °C. 26,27Considering that dioxolanyl intermediate I was not observed by ESR spectroscopy, it is unsurprising that intermediate I could not be trapped (if it exists) by cyclopropane ring-opening of B to C because EPR is on the nanosecond time scale. 28With a rate of 6 × 10 12 s −1 at 25 °C, 29,30 the cyclopropyl ring opening of a fluorenylcyclopropyl radical clock occurs sufficiently fast to trap a fleeting dioxolanyl intermediate I should it last for a time shorter than the EPR time scale.Moreover, it is unknown if transition metals influence the mechanism of the 1,2-SCS in carbohydrates to include dioxolanyl radical intermediates.1,2-Spin-center shifts in acyclic systems with copper 31 and palladium 32 reagents have been reported to proceed through mechanisms that deviate from the five-centered concerted transition state.In addition, a reported example of a 1,2-SCS in a carbohydrate (albeit from a non-anomeric radical intermediate) that proceeded through a 1,3-dioxolanyl radical intermediate suggests that the formation of such an intermediate in this context may be feasible. 33arbohydrate substrates bearing acyl fluorenylcyclopropyl radical clocks at C2 were prepared.Glycosyl bromide 13 34,35 and fluorenylcyclopropyl acid 14 36 were synthesized using known synthetic sequences.The DCC-mediated coupling of glycosyl bromide 13 and acid 14 proceeded sluggishly to afford acyl radical clock compound 15 as a nearly equal mixture of diastereomers (eq 6).Likewise, tetrahydropyranol 16 was coupled with acid 14 to afford the unsubstituted carbohydrate derivative 17 as an inconsequential mixture of four diastereomers (eq 7).
Both fully substituted carbohydrate 15 and unsubstituted tetrahydropyran 17 underwent 1,2-SCS without opening of the fluorenylcyclopropyl radical clock.When carbohydrate 15 was subjected to Ngai's excited-state palladium-catalyzed C2 reduction conditions, 11 rearrangement product 18 was observed in the crude reaction mixture (eq 8).To verify that the 1,2-SCS occurred without opening of the cyclopropane ring, rearrangement product 18 was subjected to methanolysis to afford fluorenylcyclopropyl methyl ester 19 and various The Journal of Organic Chemistry pubs.acs.org/jocNote benzoylated C2-deoxy sugars (eq 9).Under modified conditions to generate the anomeric radical, 37,38 tetrahydropyran 17 was found to rearrange to product 20 without cyclopropane ring opening (eq 10).These results suggest that 1,2-SCS of both fully elaborated sugars and simple pyrans does not rearrange through dioxolanyl intermediates such as I.
In conclusion, the rearrangements of radical clocks 15 and 17 via 1,2-SCS mechanisms without opening of their fluorenylcyclopropyl radical provide strong evidence that the 1,2-SCS occurs without the intermediacy of a dioxolanyl radical intermediate.While these radical clock experiments all but exclude the possibility of dioxolanyl intermediate I's existence, it should be noted that concerted processes II, III, and IV, which have been analyzed previously, 25 cannot be discriminated by the results of this radical clock study.We hope that by disproving the existence of I that future synthetic methods that leverage the 1,2-SCS need not consider the possibility of dioxolanyl radical I as a reactive intermediate. 11EXPERIMENTAL SECTION General Information. 1 H NMR and 13 C{ 1 H} NMR were measured at ambient temperature using Bruker AV-400 (400 and 100 MHz, respectively) and Bruker AVIII-400 (400 and 100 MHz, respectively) spectrometers unless otherwise noted.All spectroscopic data were reported as follows: chemical shifts in parts per million on the δ scale referenced from residual solvent peaks ( 1 H NMR: CDCl 3 δ 7.26 ppm; 13 C{ 1 H} NMR: CDCl 3 δ 77.16 ppm), multiplicity (s = singlet, br = broad, d = doublet, t = triplet, q = quartet, m = multiplet, AB = AB system, and ABX = ABX system), coupling constants (Hz), and integration.Multiplicity of carbon peaks was determined using HSQC experiments.Ratios of products were determined by 13 C{ 1 H} NMR experiments 39 and confirmed by 1 H NMR experiments.Infrared (IR) spectra were acquired by a Nicolet 6700 FT-IR spectrometer through attenuated total reflectance (ATR).Highresolution mass spectra (HRMS) were recorded using an Agilent 6224 accurate-mass time-of-flight spectrometer with atmospheric pressure chemical ionization (APCI) or electrospray ionization (ESI) ionization sources.Analytical thin layer chromatography was performed on silica gel 60 Å F 254 plates.All reactions were performed under a nitrogen atmosphere in glassware that had been flame-dried under a vacuum unless otherwise noted.Nondeuterated solvents were dried and purified through alumina prior to use.Aqueous solutions were prepared from nanopore water with a resistivity over 18 MΩ-cm.Light-promoted reactions were performed in 1 dram borosilicate glass vials without the use of any additional filters unless otherwise noted.All reagents were commercially available, unless otherwise stated.
1,2-Spin-Center Shift of Radical Clock 17.A reported procedure 37 was followed for the 1,2-spin-center shift of radical clock 17.A solution of thiol 17 (0.0681 g, 0.159 mmol) in toluene (1 mL) was added dropwise over 15 min to a refluxing (110 °C oil bath) solution of AIBN (0.0026 g, 0.016 mmol) and HSnBu 3 (0.206 mL, 0.763 mmol) in toluene (2.5 mL).The reaction mixture was stirred at reflux for 16 h.The reaction mixture was cooled to 25 °C and concentrated in vacuo.Analysis of the crude reaction mixture revealed poor conversion of starting material to product, but the presence of rearranged products 20a and 20b suggested that the rearrangement occurred without rearrangement of the fluorenylcyclopropyl group, which implies there is no intermediacy of a dioxolanyl radical intermediate (Figure S3 in the Supporting Information).