AmericanChemicalSociety.com

Introduction

Our goal is to control the one-pot addition of two disparate nucleophiles to a single ynoate acceptor in order to yield a chiral carboxylic acid derivative (eq 1).  Thus, simple ynoate electrophiles may be converted in one reaction vessel to densely functionalized, stereochemically complex cyclic or polycyclic systems.  Successful development of this method represents a valuable advance for a number of reasons:  1) Ynoate esters are readily available, via either carboxylation of acetylene or oxidation of propargyl alcohols.  2) Three-component coupling reactions are highly efficient.  3) These reactions rapidly construct complex synthons for target-oriented synthesis.

            A proposed mechanism for a representative heteroconjugate addition-Diels–Alder process is outlined in Figure 1.  First, a trialkylamine base (R₃N) deprotonates a thiol nucleophile, generating a thiolate anion and a trialkylammonium cation.  Conjugate addition by the thiolate provides an allenolate intermediate, and the allenolate is protonated by the trialkylammonium salt to yield an enoate ester.  The enoate ester then acts as a conjugate acceptor; in the example below, it is activated by a Lewis acid (M) and undergoes Diels–Alder cycloaddition.  Our ultimate goal is to control the formation of the new stereocenters by the use of a chiral auxiliary or a chiral metal catalyst.

Figure 1. Mechanism for Three-Component Coupling Reaction of Ynoates

Current Results

            We plan a three-pronged approach toward the development of this three-component coupling reaction: 1) establishment of the amine-catalyzed conjugate addition of various nucleophiles to ynoates, 2) use of the conjugate addition products as dienophiles in Diels–Alder cycloadditions, and 3) development of an asymmetric variant.  To date, the first goal has been accomplished and great progress toward the second goal has been achieved.  Goal 3 awaits optimization of the current racemic process (vide infra).

1. Amine-Catalyzed Conjugate Addition to Ynoates

            Our investigation of amine-catalyzed conjugate addition to ethyl propiolate shows promising results, as described in our preliminary proposal (Table 1).  Our early focus has been on thiol nucleophiles, which require only catalytic amounts of amine and afford excellent yields with high Z:E ratios.  Very recently, we have expanded the scope of this reaction to include phenol as a nucleophile, producing a vinylogous ester product.

Table 1.  Preliminary Results for Amine-Mediated Addition to Ethyl Propiolate

^aDetermined by ¹H NMR spectroscopy of the unpurified reaction mixture

^bAfter workup with aqueous acid

2.  Thioether Oxidations

            Although production of these thioethers occurs reliably, the thioether adducts themselves are not active in Diels–Alder reactions under thermal or Lewis acidic conditions.  We directed our efforts toward a one-pot two-step thioconjugate addition-oxidation sequence, to produce a sulfoxide or sulfone product under conditions that would also be amenable to Diels–Alder cycloaddition.  We found meta-chloroperbenzoic acid (m-CPBA) to be a convenient oxidant for our system because of its compatibility with our solvent of choice, CH₂Cl₂.  After verification that the purified thioether could be oxidized with m-CPBA, conditions for the optimal production of either the sulfoxide or the sulfone were developed (Table 2).  We found that 2 equiv m-CPBA was sufficient to achieve oxidation to the sulfoxide, and 3 equiv produced the sulfone.  Despite high conversion (>90%), however, isolated yields for these processes were consistently lower than 20%.  Close inspection of the ¹H NMR spectrum of the unpurified reaction mixture revealed a large number of unidentified byproducts. 

Table 2.  One-Pot Thioconjugate Addition-Oxidation Reaction

^aDetermined by ¹H NMR spectroscopy of the unpurified reaction mixture ^bIsolated yield after chromatography ^cND = Not Determined  ^dyield of sulfoxide

            We speculated that byproduct formation was caused by residual amine from the thioconjugate addition step.  Indeed, oxidation of the purified thioether proceeded in high yield with no byproduct formation.  The negative effect of the amine, however, could be mitigated by the addition of acidic additives.  Trifluoroacetic acid (TFA), Cu(OAc)₂, and LiClO₄ were all effective additives, but we chose LiClO₄ for further studies because of its compatibility with the Diels–Alder reaction we envisioned as the third step of our one-pot process.

3. Diels–Alder Cycloadditions

            The Diels–Alder portion of our reaction was first optimized using the purified sulfone oxidation product.  A survey of Lewis acid catalysts at various temperatures showed LiClO₄ to be the optimal catalyst for cycloaddition by cyclopentadiene, providing the product in high conversion and high endo:exo ratio (Table 3).  Fortunately, LiClO₄ is also compatible with the reaction conditions for the thioconjugate addition and oxidation reactions.  Moreover, it catalyzes Diels–Alder reactions of sulfoxides, viable alternative dienophiles for our one-pot process.

Table 3.  Cycloaddition of Cyclopentadiene to Z Sulfone

^aDetermined by ¹H NMR spectroscopy of the unpurified reaction mixture

            Very recently, we expanded the diene scope to include cyclohexadiene.  Again, LiClO₄ was the optimal catalyst (Table 4).  Surprisingly, no other catalysts proved capable of mediating cycloadditions by cyclohexadiene.  Furthermore, only CH₂Cl₂ appears to be a viable solvent at this date, for reasons that remain unclear.  Nonetheless, the diastereoselectivity for these reactions is remarkable: only the endo isomer can be detected by ¹H NMR spectroscopy.

Table 4.  Cycloaddition of Cyclohexadiene to Z Sulfone

^aDetermined by ¹H NMR spectroscopy of the unpurified reaction mixture

            After discovering that LiClO₄ positively affects the oxidation reaction as well as the Diels–Alder cycloaddition, we modified our procedure for our one-pot three-step thioconjugate addition-oxidation-Diels–Alder sequence.  By adding the LiClO₄ immediately prior to the oxidation step, then continuing under our normal conditions, we were able to achieve a 71% yield of the major isomer for the cyclopentadiene reaction (eq 2).  This yield is a substantial increase over the 15% yield observed when no additive is present during the oxidation step, and represents a promising precedent for expansion to other substrates.

Conclusion   

            We have established an effective, high-yielding method for our model system, the three-component coupling of ethyl propiolate, toluenethiol, and cyclopentadiene.  We are now aggressively moving forward with the expansion of the reaction scope to include other thiols and dienes, and we have begun to investigate a one-pot vinylogous ester formation-epoxidation process that also derives from ynoate esters.  Our other long-term goals include the development of asymmetric versions of our reaction.