Navigation

0 Hits

  • Previous / Next

46.6 Synthesis by Metal-Mediated Coupling Reactions

DOI: 10.1055/sos-SD-046-00149

Negishi, E.; Wang, G.Science of Synthesis, (200946239.

General Introduction

1,3-Dienes represent a myriad of naturally occurring biologically and medicinally active compounds, ranging from simple insect pheromones to complex oligoene macrolides and carotenoids as well as non-natural compounds of basic chemical and material chemical interest. Logically speaking, 1,3-dienes can be constructed by the following processes: (i) [1C+3C] processes, (ii) [2C+2C] processes, (iii) modification of 4C compounds, and (iv) fragmentation of >4C compounds.

Traditionally, various carbonyl alkenation (olefination) reactions represented by the [1C+3C] process, such as aldol and Wittig reactions, collectively served as the main routes to 1,3-dienes. One of the widely observed difficulties associated with these reactions, however, was that they often displayed unsatisfactory levels of stereoselectivity, although much improvement has been made over the past few decades.

Until around the 1960s and early 1970s, metal-mediated CC bond formation was achieved mostly through the use of alkali and alkaline earth metals, represented by lithium and magnesium. Unfortunately, however, organometals containing these metals by themselves are not generally well suited for the formation of the CC bond by their reaction with unsaturated organic electrophiles containing C(sp2)X or C(sp)X bonds, where X is a halogen or related leaving group. This limitation has been fundamentally overcome through incorporation of d-block transition metals as components of stoichiometric and catalytic reagents. Thus, the great majority of the reactions discussed in this review are either transition-metal-promoted or -catalyzed CC bond-formation processes. Of the 24d-block transition metals, only a few, including titanium, manganese, iron, copper, and zirconium, may be considered to be sufficiently inexpensive to permit their stoichiometric use, although the stoichiometric use of chromium, cobalt, and nickel may also be justified in some cases. The others, including palladium and rhodium, must be used catalytically in essentially all cases.

A wide variety of the currently known transition-metal-mediated CC bond-forming processes can be classified into just a few to several fundamentally discrete reaction patterns (reductive elimination, carbometalation, migratory insertion, and nucleophilic or electrophilic attack on ligands), exemplified as 1,3-diene-generating processes in Scheme 1. It should be emphasized here that these processes can also be observed with non-transition metals but that transition metal complexes can readily and widely participate in these processes, notably under catalytic conditions. It is also important to note that those processes shown in Scheme 1 represent reaction patterns indicating starting compoundproduct relationships rather than reaction mechanisms.

Scheme 1 Representative Patterns of CC Bond Formation Using Transition Metals, Shown as 1,3-Diene-Forming Processes

All of the processes shown in Scheme 1 are applicable to the construction of the four-carbon frame of 1,3-dienes via [2C+2C] processes. In the alkyne carbometalation and migratory insertion reactions, the conjugated dienylmetals generated must be further transformed to remove the metal groups. In the reaction shown last, nucleophilic (or electrophilic) attack on ligands, the four-carbon framework is actually generated via a [2C+2C] coupling process as the 1,3-enyne, which must then be converted into the 1,3-diene via nucleophilic attack on the alkyne moiety. Thus, the actual 1,3-diene-generating process should be classified as a modification of 4C compounds. Inasmuch as carbometalation and nucleophilic or electrophilic attack on ligands are also subjects of the preceding Section 46.5, efforts have been made to minimize unnecessary overlaps. Mainly to keep this review within a reasonable length, the [1C+3C] processes and those via fragmentation of >4C compounds are categorically excluded. Only the [2C+2C] processes involving alkenylalkenyl coupling and modification of 4C compounds will be discussed.

From the viewpoint of carbon skeletons, the alkenyl groups may be classified into eight structural types 18 (Scheme 2).

Scheme 2 The Eight Skeletal Types of the Alkenyl Groups

Alkenyl groups 1 and 2 are devoid of stereochemistry. Both regio- and stereochemical identities are unequivocally specified in groups of type 36 whereas specification of ste­reochemistry of types 7 and 8 requires detailed knowledge of the R1 and R2 groups. These eight types of alkenyl groups can be combined to generate as many as 36 types of 1,3-dienes, even if the stereochemical details of groups 7 and 8 are ignored. In this review, however, these 36 types of 1,3-dienes are divided into five product categories merely for practical reasons (Schemes 37). These are (i) 1,3-dienes containing one or two parent vinyl and/or alk-1-en-2-yl groups (a total of 15 structural types are represented by the two generalized structures shown in Scheme 3); (ii) 1,4-disubstituted 1,3-dienes (Scheme 4); (iii) trisubstituted 1,3-dienes, excluding those containing a vinyl or vinylidene group (Scheme 5); (iv) tetrasubstituted 1,3-dienes, excluding those containing one or two fully substituted alkenyl groups (i.e., 1,1,3,4-, 1,1,4,4-, and 1,2,3,4-tetrasubstituted only) (Scheme 6); and (v) tetra-, penta-, and hexasubstituted 1,3-dienes containing one or two fully substituted alkenyl groups (Scheme 7).

Scheme 3 1,3-Dienes Containing One or Two Parent Vinyl and/or Alk-1-en-2-yl Groups

Scheme 4 1,4-Disubstituted 1,3-Dienes

Scheme 5 Trisubstituted 1,3-Dienes, Excluding Those Containing a Vinyl or Vinylidene Group

Scheme 6 Tetrasubstituted 1,3-Dienes, Excluding Those Containing One or Two Fully Substituted Alkenyl Groups

Scheme 7 Tetra-, Penta-, and Hexasubstituted 1,3-Dienes Containing One or Two Fully Substituted Alkenyl Groups

For the synthesis of natural products and other fine chemicals containing 1,3-dienes, it is essential in most cases that all of the carbon substituents, R1, R2, R3, etc., can be placed as indicated without being coupled with any others. With this in mind, mainly those processes that satisfy this critical requirement as well as high levels, typically 98%, of regio- and stereoselectivities will be discussed.

Selective synthesis of regio- and stereodefined 1,3-dienes and oligoenes had been achieved mostly via carbonyl alkenation until about 1970. Even today, such methodology offers a dependable and widely applicable route to 1,3-dienes, and it is indeed widely used. Nonetheless, its general lack of high stereoselectivity (98%) makes isomer separation mandatory in most cases. Another conceptually discrete route to 1,3-dienes involves alkenylation with preformed alkenyl derivatives. The first set of the stereo- and regiocontrolled syntheses of 1,4-disubstituted 1,3-dienes by alkenylation via organoboron migratory insertion (Sections 46.6.1.1 and 46.6.1.2) were reported in 1973. Although a number of transition-metal-promoted syntheses of 1,3-dienes as exemplified by those discussed in Section 46.6.3 have also been developed, some even before 1973, they are not yet well suited for preparing regio- and stereodefined 1,3-dienes. Their further development is clearly needed.

The alkyne hydrometalationpalladium-catalyzed cross-coupling tandem process discovered in 1976 (see Scheme 20, Section 46.6.3) marked the advent of the first highly (98%) selective and widely applicable method for the synthesis of 1,3-dienes via alkenylation. There are two critical components in this method. One is the generation of the two required alkenyl reagents, and the other is their cross-coupling reaction. For generation of a stereo- and regiodefined alkenyl reagent, a wide variety of alkyne addition reactions, including syn- and anti-hydro-, carbo-, as well as halo- and other heterometalations (see Schemes 25 and 26, Section 46.6.3) have been developed and successfully used. For alkenylalkenyl coupling, the palladium-catalyzed cross-coupling reaction with alkenylzinc reagents has proven to be generally very satisfactory and dependable. However, several other metals, including aluminum, boron, copper, zirconium, and magnesium, have also been shown to be useful. Specifically, aluminum, boron, and zirconium have been the three most convenient and satisfactory metals for alkyne syn-hydrometalation. Similarly, zirconium-catalyzed carboalumination and carbocupration have proven to be two most satisfactory and mutually complementary alkyne syn-carbometalation reactions, while haloboration appears to be highly promising for providing a potentially general route to trisubstituted alkenes. Several other metals, including indium, manganese, silicon, and tin, have also been used in the palladium-catalyzed cross-coupling reaction. For various reasons, including reactivity, selectivity, cost, and toxicity, however, their use in place of several others mentioned above must be fully justified. Although many further developments are clearly desirable, the results summarized in this section suggest that as long as the two required alkenyl reagents are satisfactorily obtainable, the corresponding 1,3-dienes of the great majority of the 36 possible structural types shown in Schemes 37 may be satisfactorily prepared, even though far fewer penta- and hexasubstituted 1,3-dienes have thus far been synthesized.

It should be emphasized that, in the majority of cases, the palladium-catalyzed alkenylalkenyl cross-coupling reaction proceeds as desired, with essentially full retention of all structural features of the two alkenyl groups. One notable unexpected but useful exception to the generalization stated above is the nearly complete (typically 9798%) stereoinversion observed in the palladium-catalyzed cross-coupling reaction of 2-bromo-1,3-dienes that can be generated via palladium-catalyzed alkenylation of 1,1-dibromoalk-1-enes (Section 46.6.4.1). It is also fortunate and equally important that the unexpected ste­reoinversion observed with various conventional palladium catalysts can be completely (98%) prevented by the use of bis(tri-tert-butylphosphine)palladium(0) and other very active catalysts. Thus, in cases where stereoinversion is observed, the reaction can be directed to either inversion or retention (Tables 7 and 8), and both reactions have been successfully applied to the syntheses of some natural products (see Scheme 41, Section 46.6.4.1).

In addition to the palladium-catalyzed direct alkenylalkenyl coupling, related palladium-catalyzed 1,3-enyne and 1,3-diyne syntheses can be applied to the synthesis of 1,3-dienes, although only those via enynes are discussed in Section 46.6.4.2. It is noteworthy that the palladium-catalyzed alkynylalkenyl coupling with alkynylzinc reagents is of considerably wider synthetic scope than the corresponding HeckSonogashira alkynylation. Moreover, the palladium-catalyzed organozinc cross-coupling reaction is also readily applicable to the alkenylalkynyl coupling with alkynyl halides, which is not a viable option in the HeckSonogashira alkynylation.

The Heck alkenylation, which is known to proceed via an alkene carbopalladationβ-dehydropalladation tandem process, is often viewed as a variant of palladium-catalyzed cross-coupling reactions, but this reaction suffers from the lack of strict regio- and stereoselectivity. In view of its operational simplicity, however, it should be seriously considered in cases where it leads to satisfactory results, as exemplified in Section 46.6.5.1.

In addition to alkene carbopalladation observable in the Heck reaction, the corresponding alkyne carbopalladation has been developed as a useful reaction for preparing alkenes since the late 1980s. Unlike the Heck reaction of alkenes, the alkyne carbopalladation is fundamentally living and hence only stoichiometric in palladium. However, it can be combined in one pot with various processes, including the Heck reaction, various carbonylative trapping processes, cross-coupling reactions, and other nucleophilic trapping reactions, to devise useful and overall catalytic routes to 1,3-dienes and oligoenes (Section 46.6.5.2). This attractive and promising synthetic methodology, however, has not yet been well exploited in the synthesis of natural products and other related compounds. In sharp contrast, catalytic cyclization of enynoic acids via oxypalladation and other related oxymetalation reactions using silver and zinc has already been widely used to synthesize 1,3-diene-containing lactones (Section 46.6.5.3).

For the synthesis of oligoenes containing more than one 1,3-diene unit, the use of various combinations of the available 1,3-diene-producing methods, including those discussed in this review, would be desirable for optimizing the overall synthetic processes. In the palladium-catalyzed alkynylation methodology discussed above, for example, procurement of alkynes as alkene precursors is often critically important. In many such cases, carbonyl-to-alkyne conversion reactions, such as the CoreyFuchs reaction (see Scheme 38, Section 46.6.4.1) and methyl ketone to terminal alkyne conversion, have proven to be useful.

In cases where 1,3-dienes and oligoenes contain one or more α-chiral centers, the adjacent alkenyl units are most efficiently and satisfactorily constructed via carbonyl al­kenation. On the other hand, those 1,3-dienes that contain β-chiral centers can often be reliably assembled via palladium-catalyzed isoalkylalkenyl cross-coupling reactions. Later developments suggest that many dienyl and oligoenyl carbonyl alkenation reagents may be most efficiently and satisfactorily prepared via alkyne additionpalladium-catalyzed alkenylalkenyl cross-coupling reactions. Thus, the carbonyl alkenationpalladium-catalyzed alken­ylation synergy has emerged as an attractive methodology for the synthesis of 1,3-dienes and oligoenes. Further development of the palladium-catalyzed alkenylation, discovery and development of many additional metal-mediated selective alkyne addition reactions, and further exploration of cyclic carbopalladation- and heteropalladation-based 1,3-diene syntheses are clearly desirable, as can be seen from the discussions presented in this review.