|Home | About | Journals | Submit | Contact Us | Français|
Selectivity of six- vs. five- membered ring formation in C-H insertion on alkylsulfonyl diazoacetates is sensitive to the substrate structure and catalyst used.
Rhodium catalyzed insertion of carbenoids into C-H bonds has developed into a useful and powerful synthetic method.1 A number of total syntheses have been performed taking advantage of this methodology.2
A unique feature of this reaction, which is also its key advantage, is that no functionality needs to be present at the reactive center (C-H bond). In addition, formation of a carbon-carbon bond occurs – pivotal transformation in organic synthesis. However, with the advantage also comes a challenge - numerous potential reactive sites are commonly present in organic compounds. Consequently, selectivity becomes the central issue in this reaction.
Greater susceptibility of certain C-H bonds for insertion (such as ethereal or allylic) makes it possible to perform a selective intermolecular reaction on some classes of substrates.3 More generally, intramolecular reaction is used to control the selectivity, relying on conformational restrictions, or, particularly, on the usual overwhelming preference for formation of five-membered rings.4
We previously reported that carbene C-H insertion on diazosulfone and diazosulfonate substrates, in contrast, leads to preferential formation of six-membered rings when the sulfonyl group is incorporated into the forming ring.5 The subsequent report by the Du Bois group6 demonstrated that in case of diazosulfonates this preference persists on a variety of substrates, even when formation of a strained system is required. This unusual preference has been explained by different bond lengths and bond angle around SO2 fragment that favor the six-membered ring geometry, as in case of the nitrene insertion on similar substrates.7
In our follow-up studies of this reaction we have uncovered the tenuous nature of this preference on diazosulfone substrates. Here we report the result of these studies.
In our initial report6 we identified sulfone 1.1a as the major product of cyclization of diazosulfone 1. A more detailed analysis of the reaction mixture revealed the presence of small amounts of the cis-isomer, 1.1b, as well as five-membered isomers, 1.2a,b. The difficult separation was further complicated by concurrent equilibration of 1.2b to 1.2a during chromatography. Fortunately, it was found that equilibration of the mixture of 1.2a and 1.2b with DBU in CH2Cl2 at rt provided virtually exclusively 1.2a. Prior equlibration simplified the separation. Curiously, under these equilibration conditions the six-membered isomers 1.1a and 1.1b, form a mixture with a close to 1:1 ratio.
We continued by testing other catalysts for this transformation. Doyle’s Rh2(cap)4 (rhodium caprolactam), which was known to amplify C-H insertion selectivity,4b and Du Bois tethered Rh2(esp)28 catalysts proved effective for this transformation, showing results similar to rhodium (II) acetate. Rh2(cap)4 catalyst was unreactive at rt or at reflux in dichloromethane, necessitating the switch to the higher boiling solvent dichloroethane. Unexpectedly, Rh2(pfb)4 (rhodium(II) perfluorobutyrate) catalyst lead to reversal of the usual selectivity, giving primarily the five-membered products.
We then explored the effect of substitution adjacent to the reactive sites. Preparation of the necessary substrates is shown in Scheme 1. Substrates 1, 6 and 7 were prepared by our previously reported procedure,5 with a minor modification in case of 7 (switch of the solvent to DMF). Preparation of substrate 8 was achieved by acid-catalyzed SN1 substitution to make the sulfide, which was further processed as usual.
Decrease of substitution at the insertion site (substrate 6) predictably disfavored the insertion, leading to preferential formation of the five-membered products, with only small amounts of the six-membered sulfone. Notably, use of Rh2(pfb)4 made the preference for the five-membered ring complete, also greatly improving the yields of the product.
Less expectedly, substitution next to sulfone center would also increase the fraction of the five-membered products. Monomethyl substitution (subtsrate 7) lead to formation of close to equivalent amounts of five and six-membered products. Use of Rh2(pfb)4 amplified this bias to make five-membered sulfone prevalent. Notably, selectivity in formation of the center at β-position was observed for this substrate. The stereochemistry of the principal products is shown on Figures 1 and and2.2. Formation of small amounts (5 to 10% combined) of several products possibly containing the opposite configuration at β-position was observed. We were unable to sufficiently purify them for definitive identification. The dimethyl substituted diazosulfone, 8, was stable to rhodium catalysts at rt or reflux in methylene chloride, requiring reflux in dichloroethane to react. The reaction resulted in greatly prevalent formation of the five-membered product even with Rh2(OAc)4 catalyst. Use of Rh2(pfb)4 lead to complete selectivity towards the five-membered product at the expense of the decreased yield. The tendency can be explained by decreasing of the angle in C-C-SO2 fragment due to steric compression, favoring the five-membered ring geometry.
The structural and stereochemical assignments of the products have been performed on the basis of their interconversion, 1H, 13C, COSY, DEPT and 1D NOE difference spectra.
The relative configuration of stereocenters in 1.1b was assigned on the basis of the coupling constants between Hb and Hc (Figure 1). Smaller coupling constant and appearance of another coupling constant to Ha (confirmed by COSY) indicate the equatorial position of Hb. Similarly, diaxial arrangement of Hb and Hc is established in 7.1a, while indicating the cis-arrangement in 7.1b. Bidirectional NOE correlation between Ha and Hb established the configuration of the β-center. While formation of the cis-isomer of 8.1a appeared to occur in the reaction of 8 (by presence of characteristic peaks in the crude spectra), it appeared to recede after equilibration, and only 8.1a was isolated. It has been verified that 8.1a does not change upon treatment with DBU, unlike 1.1a and 7.1a, which produce mixtures of cis- and trans- isomers. This could be explained by unfavorable diaxial interaction with the methyl group that would be present in the cis- isomer of 8.1a, but not in 1.1b and 7.1b.
All five-membered products were isolated as single isomers after equilibration.
They all show similar patterns of chemical shifts and NOE correlations between protons. They have been assigned trans-configuration on the basis of NOE data and coupling constants. Observed NOEs and their relative intensities are consistent with the proposed structures, while several disagreements can be found with the cis-isomers. Additionally, the large coupling constant (10Hz) between Hc and He in 1.2a (Hc and He in 6.2a, Hd and Hh in 7.2a, Hc and Hf in 8.2a) indicate a dihedral angle that is close to either 0 or 180°. Weak NOE correlations between these hydrogens suggest against the small dihedral angle that would force these hydrogens to a close proximity. For 7.2a, relative configuration of the C5 center was unambiguously confirmed by a bidirectional NOE between Hb and Hd.
In summary, C-H insertion on the alkylsulfonyl diazoacetate substrates has demonstrated a sensitive nature of the earlier discovered preference for formation of six-membered rings. Substitution next to sulfone was found to tilt it towards the formation of five-membered sulfones. Unexpectedly, same influence is also exerted by Rh2(pfb)4 catalyst. This permits a degree of control over the reaction outcome to form either thiofuran or thiopyran 1,1-dioxides, both of which are useful intermediates in synthesis.
Further studies of this reaction will be reported in due course.
This work was supported by the American Chemical Society Petroleum Research Fund under grant No. PRF#46278-G1 and National Institutes of Health under grant No. GM085645. We thank Alena Kubatova for HRMS analyses. The work on TOF MS was supported by the National Foundation under grant No. CHE-0216038.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.