Selective Approaches to a - and b -Arylated Vinyl Ethers

: We developed simple and efficient protocols for palladium-catalyzed regioselective a - and b -arylations of structurally diverse vinyl ethers. Both catalytic methods proceed under relatively mild reactions conditions applying to a broad substrate range including more complex compounds providing arylated glucal or isochromene. Lacking the common requirement of a large reagent excess, the transformations are highly economic and limiting the waste production. Results from computational studies (DFT) provided insight into the key factors determining the pronounced regioselectivities of the investigated reactions.


Introduction
Due to the presence of electron-enriched double bonds, vinyl ethers are molecules of great importance for organic chemistry.T hey exhibit unique properties and undergo av ariety of transformations leading to structurally diverse, complex molecules. [1] Consequently,t hey are considered as highly valuable synthetic building blocks. [2] However, stereochemically defined vinyl ether formations through transvinylation reactions,d ouble bond isomerizations,e liminations,o lefinations,o rc oupling reactions are commonly limited to the simplest (n-butyl or ethyl) or non-substituted derivatives. [3] Syntheses of substituted and more complex enol ethers such as a-a nd b-(benzyloxy)styrene have been described but their preparations require toxic, expensive or hard-to-obtain reagents making those methods less attractive. [4] Among the more recent methodologies for modifications of vinyl ethers are palladium-catalyzed Heck reactions. [5] However,t hese very attractive direct CÀHa ctivation methods have only been applied to simple model compounds such as n-butyl vinyl ethers (Scheme 1). [6] Furthermore,t hey are characterized by low regioselectivities and ap oor reagent economy with the requirement of large excesses of starting materials. [6a-c,m-t] Studies beyond model reactions [6f-l, 7] focussing on conversions of substituted vinyl ethers and their regioselective functionalizations are lacking.I nl ight of the above facts and inspired by the seminal work of Hallberg from 1993, [8] we initiated as tudy directed towards the development of regioselective Heck-type arylation reactions of benzyl-substituted vinyl ethers 1 (Scheme 1).

Results and Discussion
At the beginning of our investigation, as imple and efficient method for the preparation of benzyl-type vinyl ethers 1 was needed. In order to ensure that the subsequent reaction development was not affected by the presence of trace amounts of metallic contaminations, [9] the protocol by Matsubara [10] involving benzyl alcohols,calcium carbide,a nd as uperbase consisting of KOHi nD MSO was applied for synthesizing key starting materials 1.F or the initial screening and optimization of the reaction conditions,v inyl ether 1a was selected as representative starting material. a-Phenylations of 1a to give 6a were set in the focus first. Scheme 2 gives an overview on the findings,and all results are detailed in the Supporting Information (Table S1).
Under the optimized conditions with ac ombination of palladium acetate and dppp [1,3-bis(diphenylphosphino)propane] in the presence of Hünigsb ase (DIPEA) in DMF Scheme 1. Heck-type arylations of vinyl ethers as reported by Hallberg [8] and developed here. instead of phenyl triflate as aryl source it was apparent that water had ad eleterious effect on the yield due to partial hydrolysis of starting material 1a. [11] Similarly,the presence of even small amounts of atmospheric oxygen proved unfavorable leading to mixtures of a-and b-regioisomers.T hus,f or achieving ahigh a-arylation the catalysis had to be performed in degassed DMF under argon. Theu se of other bases than DIPEA led to lower conversions of 1a (Scheme 2). Interestingly,the regioselectivity remained unaffected. Among dioxane,P EG400, toluene,a nd DMF the latter solvent gave the best conversion of 1a.S creening of various palladium complexes revealed that Pd(OAc) 2 was the most effective catalyst metal species (Scheme 2). Thel igand type affected the regioselectivity of the arylation (Scheme 2and Table S1 in the Supporting Information) with dppp providing the best results for the a-phenylation. [12] Performing the catalysis at 50 8 8Cand 120 8 8Cinstead of 80 8 8Cled to longer reaction times and the formation of undesired by-products,r espectively. Applying iodobenzene (4a)instead of phenyl triflate afforded mixtures of a-and b-regioisomers.Phenyl nonaflate (5a)was an equally good aryl source as phenyl triflate (Scheme 2). Next, the substrate scope of the a-arylation process was investigated. Theo ptimal reaction conditions as depicted in Scheme 2were used as initial starting point. Thestructures of both coupling partners,vinyl ether 1 as well as aryl triflate 2, were varied. Noteworthy,b oth compounds were applied in a1.0:1.1 ratio thereby avoiding the previously required large excess of the arylating agent. Scheme 3shows the results.
In the first series of experiments (leading to products 6aj), the aryl group of the benzyl moiety of vinyl ether 1 was modified. Phenyl triflate (2a)s erved as aryl source.I nm ost cases the yields of the corresponding products were high, reaching 96 %for compounds 6c, 6e,and 6f.Diphenylmethyl vinyl ether (1b)reacted well too,leading to 6b in 93 %yield. Phenyl, methyl, and methoxy substituents were tolerated and the reactions proceeded without significant impact. Probably due to steric factors,a ne xtension of the reaction time from 3hto overnight was required for achieving ayield of 93 %of product 6h with a2 -methoxy-substituted aryl group.V inyl ethers with hetarenes led to lower yields of the corresponding products as found for 2-furyl-and 2-pyridinyl-substituted 6i and 6j,w hich were obtained in 79 %a nd 38 %y ield, respectively.F rom all of these results we conclude that chelating effects were not required for the high a-selectivity of the arylation. If chelation was possible (as for 1h-j), it had no effect or appeared to hamper the process.
In the next series of experiments,b enzyl vinyl ether (1a) was the starting material and the aryl triflate structure was varied (Scheme 3). While the reactions with 1-naphthyl, 2naphthyl-, and 4-substituted aryl triflates proceeded well, providing products 6k-n in yields ranging from 92 %to96%, the use of 4-nitro-containing aryl triflate 2o gave 6o in only 33 %y ield even after an extended reaction time (overnight). Thee lectron-withdrawing nitro substituent exhibited an egative effect on the product formation.
Thef inal four entries in Scheme 3r elated to product formations of 6p-s reveal that also combinations of starting materials with substituents on both vinyl ethers and aryl triflates could be applied. Although extended reaction times (overnight) proved beneficial in these cases,t he yields were generally high (up to 96 %). Theo nly exception was the catalysis leading to 6s,w hich was formed in only 24 %y ield. Scheme 2. Optimized procedure for the a-phenylation of 1a to give 6a and variations thereof;for details see Table S1 in the Supporting Information.
Scheme 3. Substrate scope of the a-arylation. Thus,the aforementioned negative impact of the nitro group on the aryl triflate became apparent here too. In his early work, Hallberg and co-workers had also reported b-arylations of enol ethers (Scheme 1). [8] Forachieving ah igh b-selectivity,c helation control was the key to success.W ithout this effect, the a/b-selectivity remained modest. Seeing ac hallenge there,t he development of ap rotocol for b-selective arylations of simple enol ethers was targeted. Considering Hallbergsc onditions as the starting point, phenyl iodide (4a)w as used as aryl source. Scheme 4r eveals the optimized conditions and presents results for b-selective arylations of various substrates.A detailed presentation of all optimized factors developed by studying reactions between 1a and 4a can be found in the Supporting Information (Table S2). Abrief summary is given here:first, the presence of dppp lowered the b-selectivity,and thus,p erforming the catalysis ligand-free was beneficial. Second, the palladium source affected the conversion of 1a but played am inor role in determining the a/b-selectivity of the vinyl ether arylation. Third, additives,i np articular chloride salts (LiCl, Bu 4 NCl, and Aliquat 336), had apositive effect on both conversion of 1a and b-regioselectivity.Fourth, the character of the base influenced the E/Z ratio of the barylation product. While with DIPEA the E isomer dominated, inorganic bases (NaOAc and K 2 CO 3 )g ave the Z isomer in preference.F ifth, DMF and THF proved to be the best solvents.B ecause the latter required al onger reaction time (24 hv ersus 3h for DMF) for full conversion of 1a, DMF was chosen for subsequent studies.Interestingly,among all tested seven solvents (DMF,P hMe,o ctafluorotoluene, DCE, THF,and MeCN) acetonitrile was the only one that led to a2:1 a-selectivity in the phenylation of 1a (at aconversion of 6% after 2h). Finally,t he best result was achieved with 6mol %o fp alladium dichloride as metal catalyst with 2.0 equiv of tetrabutylammonium chloride as additive and 1.2 equiv of DIPEA as base in DMF under argon for 2hat 80 8 8C. [13] Under these conditions,a18/82 mixture of 6a and 7a was obtained in 83 %y ield, where b-substituted 7a had an E/Z ratio of 44/56. Reactions between structurally related benzyl vinyl ethers with substituted arenes and 4a gave similar results (Scheme 4, top). Thus,the 6b/7b, 6c/7c, 6e/7e, 6g/7g,a nd 6h/7h mixtures were isolated in yields ranging from 81 %t o8 9% with the b-arylated products 7 being formed in significant excess over their a-arylated counterparts 6.P robably due to chelation, 2-furyl-substituted 1i showed an exclusive regioselectivity leading to 70 %o fbphenylated product 7i only.I ne ach case,aslight preference (ca. 2:3) of the Z diastereomers was observed. Foru nknown reasons,product 7e was an exception with an E/Z ratio of 1:9. 2-Pyridinyl-substituted 1j did not react at all under these conditions.C ouplings of benzyl vinyl ether (1a)w ith aryl iodides other than 4aproceeded well too (Scheme 4, bottom). While the transformations providing 6m/7m and 6n/7n gave similar results as described above,t he reactions with 4-nitro iodobenzene (4d)a nd 4-chloro iodobenzene (4e)w ere remarkable as b-arylated 7o and 7t,r espectively,w ere obtained as single regioisomers.M ost likely,t hese results were due to the electron-withdrawing nature of the additional substituents (nitro and chloro) on the aryl source,w hich supported the b-selectivity of the process.Ofinterest was also the coupling of iodoindole 4fwith 1a.Although the product yield was only 45 %, a1 :1 mixture of a-arylated 6u and bsubstituted 7u was obtained. This result is noteworthy because no protection of the indole NH was required to allow the catalysis to occur.
As the data in Scheme 4show,the development of ahighly b-selective arylation of simple vinyl ethers proved challenging. Even after an extensive experimentation, mostly mixtures of a-and b-substituted products were obtained. Although the latter products dominated, only af ew couplings proceeded with complete regiocontrol. Although we could not fully reach our goal, an interesting phenomenon came to our support which finally allowed us to obtain pure b-arylated products.W hile attempting to achieve higher regioselectivities,wefound different reaction rates of hydrolysis of the two regioisomers.Ashort optimization revealed that the use of a0.01 Msolution of formic acid in THF was most effective for this process.T hus,s tirring am ixture of a-arylated 6 and bsubstituted 7 in that reaction media overnight led to an exclusive hydrolysis of 6 allowing an easy isolation of 7 as pure regioisomer (Scheme 5). With minor variations the double bond geometry of 7 was retained. This was also true for product 7e,w here the E/Z ratio was particularly high. Further experiments showed that this approach was general and could be utilized in the synthesis of various b-arylated vinyl ethers.
With the intention to demonstrate the utility of the developed methodology and to extend its application, adomino diarylation involving two consecutive palladium-catalyzed CÀC-bond formations was attempted. To our delight, this approach proved successful. Hence,using the b-arylation conditions as summarized in Scheme 4, am ixture of 1- Although the precise details of the reaction sequence are unknown, we assume,b ased on the a-position of the newly introduced phenyl group stemming from 4a,t hat the intramolecular ring-closure (with b-arylation selectivity under the chosen reaction conditions) occurs first, and that the resulting palladated heterocylic intermediate then cross-couples with 4a in as econd step. [14] As second example,t he arylation of tribenzylated glucal 10 was chosen. This transformation was of interest for two reasons:f irst, it involved am ore complex starting material with various functional groups allowing directing effects and, second, the vinyl fragment was disubstituted and part of a6membered heterocycle.F urthermore,t ot he best of our knowledge,r egioselective glucal arylations have only been achieved with organometallic reagents until now, [15] whereas direct metal-catalyzed reactions of this type have remained unstudied. To our disappointment, the previously devised aarylation protocol [with phenyl triflate (2a)asaryl source as summarized in Scheme 3] proved ineffective,p roviding phenylated 11 in only 10 %y ield (Scheme 6). However,a n improvement was achieved by following the alternative approach [with phenyl iodide (4a)asaryl source as summarized in Scheme 4].N ow, 11 was obtained in 73 %y ield (with 74 %c onversion of 10). This result revealing ah igh aselectivity was unexpected because previous applications of this protocol had predominantly provided b-arylated products.A ssuming that this switch from b-to a-selectivity was linked to the disubstitution of the vinylic portion of 10,t he arylation behavior of compounds 12 and 13 was studied. While 12 proved inactive allowing no arylation with neither the a-n or the b-selective protocol, 13 reacted well and, confirming our aforementioned hypothesis, both approaches provided a-arylated product 14 in preference.Indetail, under the conditions shown in Scheme 3with phenyl triflate (2a)as aryl source, 14 was obtained as the sole product in 69 %yield (after 24 h). Finally,a pplying the second palladium catalysis with phenyl iodide (4a)a sa ryl source and using conditions summarized in Scheme 4until complete conversion, amixture of a-arylated 14 and b-arylated 15 resulted with the former product in a68:32 preference.Hence,itseems that enol ethers with a1 ,2-disubstituted vinyl unit have ap ronounced aarylation selectivity,which we have not been able to override by conditional changes until now.
In order to shed light on the key factors determining the regioselectivity of the vinyl ether arylation, aD FT computational study,specifically focused on the vinyl insertion into the Pd À Ph bond and its regioselectivity,w as carried out (see Supplementary Information for computational details). As noted before,t he regioselectivity is highly sensitive towards several factors in the reaction mixture,i ncluding the palladium source,the ligand (or its absence), the base,and the aryl (pseudo)halide.For favoring the a-selectivity,the presence of dppp as ligand and the use of an aryl triflate were key.A s previously observed by Jutand and co-workers in stoichiometric reactions, [16] dppp in combination with aryl triflates and Pd(OAc) 2 favors the generation of cationic ArPd-(dppp)] + species.I nc ontrast, aryl iodides provide mixtures of neutral and cationic species.I na ddition, the absence of ligand and the presence of chloride ions favor the vinyl insertion at the b-position.
To further analyse this effect, the free energy activation barrier for vinyl insertion into the PdÀPh bond using three different models was calculated:f irst, an electron-deficient cationic (dppp)-Pd complex, second, an eutral Pd complex that might be formed from PdCl 2 as precursor and DIPEA as additive,a nd finally third, an anionic electron-rich Pd complex that would be favored under excess of Cl À ions (Scheme 7). Although many other species could co-exist in solution, we aimed to identify at rend that correlated the electron density of the palladium center with the competition . Finally,the neutral palladium complex was located in between of the other two species,slightly favoring the a-position by 1.1 kcal mol À1 .These results can also explain the difficulties in improving the b selectivity,asthe absence of al igand prevents the formation of well-defined species,a nd consequently both neutral and anionic species can co-exist in solution.

Conclusion
In summary,w ed eveloped two protocols for a-a nd bselective arylations of benzyl vinyl ethers.B oth approaches are catalytic Heck-type cross-couplings,w here activity and the regioselectivity are determined by an umber of factors including the palladium species,t he presence or absence of al igand and its specific type,t he aryl source,a dditives,a nd fine details of the reaction conditions.Incontrast to previous studies,analmost equimolar substrate ratio was sufficient for high product formation, and the regioselectivity of the arylation did not rely on chelating effects.Aselective hydrolysis of the a-arylated product over its b-substituted analogue allows isolating the latter in case an unsatisfying a/b-selectivity characterizes the coupling reactions.M ore structurally complex enol vinyl ethers can be applied but both protocols lead to an a-arylation preference.H owever,d irect a-arylation of glycals developed here can potentially become avaluable tool in the synthesis of gliflozins medicines used to treat type II diabetes mellitus.ADFT computational study revealed the key factors leading to the pronounced regioselectivities,a nd the obtained results were in good agreement with the experimental observations.