were both difficult to synthesize, as the corresponding alcohols could not be converted into the chloride easily, and prone to decomposition and polymer formation.111 Therefore an alternative synthesis was devised. The reduction of easily synthesized 2-carbonoylchalcones by sodium borohydride results in the complete reduction of the ketone to the methylene (Scheme 45).101b,112 Importantly, this route avoids exposure of 119a and 119b to acidic reaction conditions, thereby
minimizing the opportunity for decomposition. Chalcones 118a and 118b were accessed in one step from 2-hydroxy acetophenone. Formation of the phenol carbonate and subsequent reduction with sodium borohydride afforded 119a and 119b in 66 and 48% yield respectively.
Scheme 45
In previous communications, the sulfenofunctionalization reaction has shown significant changes in both stereo- and site selectivity based on the geometry of the alkene in question as well as the distance between the alkene and the nucleophile (Chapter 2). Therefore, efforts were directed at accessing substrates which could help elucidate these effects with respect to the functionalization with phenolic nucleophiles. The synthesis of geometrically pure (E)-alkenes necessary for these reactions was surprisingly difficult. To solve this problem, a number of approaches were considered (Scheme 46). These are (a) Wittig and Schlosser-Wittig reactions, (b) the stereospecific elimination from diastereopure alkyldiphenylphosphine oxides, i.e. Horner- Wittig, (c) the SN2 and SN2' displacement of allyl and benzyl halides with organometallic reagents, (d) an intermolecular cross-coupling of alkenyl and aryl halides and (e) olefin metathesis of a terminal alkene and a styrene.
. 4.4.2.1. Wittig and Schlosser-Wittig Reactions.113 The first attempt to form the (E)-alkene junction employed benzyltriphenylphosphonium bromide and 2-hydroxy-(4H)-dihydrochroman as the reaction partners. With a Na-counterion the E:Z selectivity was about 6:1. Therefore, the Schlosser modification was tested as an alternative approach. In this case, PhLi was used as the base to equilibriate the intermediate betaine before the final elimination. Unfortunately, under the best conditions tested with the Schlosser modification, the selectivity did not exceed 4:1.
Scheme 47
4.4.2.2. Horner-Wittig Reaction.114 The disappointing failure of the direct Wittig methodology suggested that the intrinsic E:Z selectivity for this substrate may not be great. Therefore, an alternative approach that utilizes the Horner method was generated. The α-oxo diphenylphosphines 123 are isolable intermediates, and separation of the diastereomers would lead to diastereopure 123. The elimination from α-oxo diphenylphosphines is stereospecifically syn and therefore 120 should be produced as the geometrically pure (E)-isomer. The percursor synthesis began with the nucleophilic attack of benzyldiphenylphosphine anion onto dihydrocoumarin. The resulting ketone was reduced with NaBH4 in 9:1 selectivity favoring the desired trans-isomer. However, the final elimination proved problematic. Although only (E)-120 was visible by 1H NMR spectroscopy, the reaction proved to be very irreproducible, with mixtures of starting material, low amounts of alkene and other unidentified products. Although a few conditions were tested, this route, too did not ultimately prove successful.
Scheme 48
4.4.2.3. Organometallic Reagents.115 The problems encountered setting the (E)-alkene geometry led to the investigation of alternative approaches that would take advantage of preformed geometrically pure alkenes. The C-C bond disconnect along the alkene tether led to the consideration of SN2-type reactions on activated centers. Thus, either the combination of an allyl organometal with a benzyl halide or of a benzyl organometal with an allyl halide would both lead to the desired product. Thus, a wideranging investigation into the possible conditions for such a coupling was performed.
Scheme 49
The Wurtz-type coupling and homocoupling of activated organometallic reagents was an expected side reaction.116 Unfortunately, in most of the conditions tested, homocoupling to 125 was observed. Primarily, the initial formation of organometallic reagents proved problematic as the organometallic species generated immediately reacted with the remaining reagent leading to homocoupling. Thus, no clean generation of cinnamyl organometallic reagents was possible. The homocoupling was problematic even when the conjugation was interrupted by replacement of the
aryl moiety with a phenethyl moiety. The generation of benzyl organometallic species was similarly problematic with predominantly bibenzyl derivative 124 observed as the primary constituent of reaction mixtures. Conditions that employed organometallic reagents with lower reactivities such as In117 and Zn118 also did not lead to successful couplings.
4.4.2.4. Synthesis of Vinyl Halogens. The cross coupling of various alkenyl halides and with aryl boronates119 or aryl silicon reagents120 is well developed. Therefore, vinyl halides of type 128 were targeted. Treatment of 126 with the Ohira-Bestmann reagent121 led to the formation of alkyne 127. However, Al-H reduction of the alkyne followed by I2 quench primarily afforded the terminal alkene. The Takai reaction122 of protected aldehyde 129 was also unsuccessful. Further optimization of this route was not attempted based on the success obtained in the cross-metathesis reaction.
Scheme 50
4.4.2.5. Alkene Metathesis. Olefin metathesis reactions that employ Grubbs-I and Grubbs-II type catalysts generally proceed with high (E)-selectivity.123 Thus, metathesis of a styrene with a terminal olefin was seen as a possible answer to the problem of E:Z selectivity. The requisite terminal alkenes 131 were prepared from 2-methoxybenzyl bromide by displacing the bromide with a preformed Grignard reagent (Scheme 51).124 Deprotection with sodium ethanethiolate in refluxing DMF produced terminal alkenes 131a and 131b in 89 and 68% yield over two steps.19f
The use of Hoveyda-Grubbs II catalyst 135 for the metathesis of alkenes 131 was unsuccessful. No reaction was observed at low temperature. When the temperature was raised to reflux, homocoupling product 136 was observed. The Grubbs-I-indenylidene type catalyst 134 was much more successful at catalyzing the reaction. Metathesis of 130a and 131a with excess styrene was found to deliver protected styrenes with high geometrical selectivity.125 Usually, addition of a second portion of catalyst was necessary for high conversion. Protection of the phenol as the methyl ether led to slightly higher yields in the reaction. Subsequent demethylation under identical conditions resulted in the formation of 120 and 120b in high yield.
Table 16. Grubbs Metathesis of Butenyl- and Pentenylphenols.
entry R n cat. styrene temp, oC time, h yield, %a
1 H 1 135 1.2 rt 24 0 2 H 1 135 1.2 reflux 6 n.d. 3 H 1 134 5 rt 36 62% 4 Me 1 134 5 rt 36 69% 5 H 2 134 5 rt 36 42% a Isolated yield.
The (E)-selectivity of the olefin metathesis reaction does not extend to isolated alkenes under cross-metathesis conditions.123 Therefore a new route was devised to access substrates with a two-methylene tether but with alkyl substituents. The Pd-mediated opening of 2- vinylchroman by malonate nucleophiles mediated by palladium had been studied by Murahashi in 1986.126 Under similar conditions, diethylmalonate intermediate 137 could be accessed rapidly (Scheme 52). The decarboxylation of the malonate was effected using Krapcho conditions.127
NaCl was sufficient to decarboxylate the diester, but the high temperatures and long reaction times necessary led to uncharacterized polymerization products were observed by 1H NMR spectroscopy. Instead, use of NaCN resulted in rapid decarboxylation at only 160 oC, with a minimal amount of side products.128 Under the optimized conditions, the desired ester substrate could be obtained in 85% yield.
Scheme 52
Ester 138 then served as a linchpin for diversification to study the effects of Lewis basic groups on the reaction. Hydrolysis yielded 139 in 93% yield, and reduction of the ester afforded 140 in 91% yield (Scheme 53). Ether 141 was synthesized from the corresponding bromoarene.
Scheme 53
4.4.3. Optimization of the Phenoxysulfenylation Reaction. The use of a phenolic