Some transition-metal catalysts control organic reactions so that, given a choice of two mirror-image products, only one forms. The metal atom in these catalysts has been ignored as a source of control — until now.
In a classic episode of The Simpsons, Homer encounters problems when he enters a three-dimensional world beyond the flat environs of his cartoon existence. Chemists can appreciate his culture shock because, although flat organic molecules are relatively easy to make, three-dimensional molecules are much more challenging — especially those containing chiral carbon atoms, which have four different groups attached. Chiral molecules exist as one of two mirror-image forms (enantiomers), each of which interacts differently with other chiral entities. Because most biological molecules are chiral, it is essential to find ways of making enantiomerically pure compounds for drug discovery and other biochemical applications.
Chiral catalysts are invaluable in this respect, because they churn out chiral organic products as single enantiomers from flat starting materials. Such catalysts contain structures that act as templates for chirality in the products, and are said to be stereogenic. Usually these catalysts consist of one or two chiral ligand molecules bound to a central metal atom, so the stereogenicity comes from carbon atoms in the ligands. But, reporting on page 933 of this issue, Hoveyda and colleagues1 demonstrate that 'stereogenic-at-metal' catalysts can also be used effectively to form enantiomerically pure products in a carbon–carbon bond-forming reaction — that is, the source of chirality is a metal atom to which four different ligands are attached. This represents a new way of relaying chiral 'information' to organic molecules.
The ideal catalyst for a particular class of reaction would react with any substrate for that reaction. In practice, it is difficult to find catalysts with such wide scope, and so, for any given reaction, there is often one catalyst that performs better than another. Frustratingly, it isn't always clear in advance which catalyst will be best. It is therefore useful to have several different catalysts (each with their own combinations and/or arrangements of ligands) that can be tested in new reactions to find which one optimizes the yield of products — or, for chiral reactions, which one optimizes the transfer of chiral information from the catalyst to the product.
A case in point is the olefin metathesis reaction. In this process, two olefins (which contain carbon–carbon double bonds, C=C) react in the presence of a catalyst so that the C=C bonds in the reactants break, and new C=C bonds form between the fragments (Fig. 1a). Metathesis means 'exchange of parts', and the reactants can be thought of as pairs of dancers, who switch partners as if they were taking part in a square dance2. The metathesis reaction has become incredibly useful for organic synthesis. Indeed, in 2005 the Nobel Prize in Chemistry was awarded to Robert Grubbs, Richard Schrock and Yves Chauvin for their pioneering work in this area.
There are two main classes of catalyst for olefin metathesis: the molybdenum-containing catalysts devised by Schrock, and the ruthenium-based variety developed by Grubbs. The ruthenium catalysts are most popular because they can be used in the widest range of solvents and are not compromised by most of the chemical groups that are commonly found in organic molecules. The molybdenum catalysts are more sensitive, and are typically reserved for use with less reactive olefins and for substrates that have few chemical groups that might foul up the catalyst. But they have a useful role in metathesis reactions that yield enantiomeric products, partly because it is easy to incorporate chiral ligands into molybdenum catalysts, and also because these catalysts are fairly good at transferring chiral information from their ligands to reaction products.
The work reported by Hoveyda and colleagues1 takes molybdenum catalysts to their highest level of performance yet. The authors' new catalyst, with its stereogenic-at-metal design, performs metathesis reactions that provide greater yields than the existing compounds, and also provides higher levels of chiral induction, so that products are obtained as essentially only one enantiomer. This is directly attributable to the ability of the stereogenic molybdenum atom to direct the interactions of C=C bonds in reactants with the catalyst, controlling the alignment of the bonds so that only one enantiomeric product can form. But there is another unusual aspect of the catalyst's design that is key to its success: the ligands.
The simplest ligands in metal complexes are monodentate — they bind to the metal through a single atom (Fig. 1b). But monodentate ligands present a design challenge for chiral catalysts, and especially so for metathesis catalysts. In a reaction catalysed by a metal complex, the starting materials form a complex with the metal atom; this brings the starting materials close together so that they can react. The alignment of the reactants in the complex determines the chiral outcome of the reaction, and is governed by the other ligands around the metal, which jostle with the reactants for space. For maximum chiral induction, the ligands should be as close as possible to the newly forming bond between the reactants.
But in ruthenium metathesis catalysts, the monodentate chiral ligand sits on the opposite hemisphere of the intermediate complex to that of the reactants, minimizing the potential for chiral induction. The same problem occurs in catalysts for many other reactions: chiral monodentate ligands tend to move far away from reactant-binding sites. Nevertheless, Grubbs and co-workers have found ways to successfully 'drape' the stereogenic element of monodentate ligands around the central metal atom of a metathesis catalyst towards the reactants, so that products are formed predominantly as one enantiomer3,4.
Another problem with monodentate ligands in chiral catalysts is that they tend to rotate around the point of attachment to the metal. Such rotation changes the catalyst's conformation, thus altering the chiral environment around the metal. This is catastrophic, because each conformation has a different reactivity and provides a different degree of chiral induction. Such difficulties are usually overcome by using bidentate ligands, which bind to metals through two atoms (Fig. 1c). These ligands are less floppy than monodentate ones, and have fewer conformations to complicate (or limit) the degree of chiral induction. Bidentate ligands have allowed the successful development of chiral catalysts for alkene metathesis in both the ruthenium5 and molybdenum6,7 classes. But bidentate ligands don't always have the flexibility to allow large structural changes around the catalytic metal atom, as is sometimes required for metathesis reactions.
The ligands in Hoveyda and colleagues' catalyst1 are monodentate (Fig. 1d). This is necessary for efficient chiral induction from the stereogenic molybdenum atom. But, remarkably, the complexes are configurationally stable, preventing the problems usually caused by monodentate ligand rotation. Furthermore, because the flexible monodentate ligands accommodate large structural changes during reactions, the catalyst is more reactive than analogous catalysts that use bidentate ligands. The stereogenic-at-metal design thus yields an exceptionally reactive catalyst that also provides high levels of chiral induction.
The authors use their catalyst in an elegant, short synthesis of a structurally complex, naturally occurring compound, quebrachamine (see Fig. 4 on page 936). They find that the catalyst distinguishes between three different olefins in the achiral starting material, sequentially reacting only the required two, thus yielding the product essentially as a single enantiomer in high yield. This reaction could not have been achieved using previously available catalysts. Moreover, their catalyst performs the reaction in the presence of a tertiary amine (an organic base containing a nitrogen to which three hydrocarbon groups are attached) in the substrate, which is typically a troublesome motif in metathesis reactions.
Hoveyda and colleagues' stereogenic-at-molybdenum complex has the potential to become an all-purpose chiral catalyst for olefin metathesis reactions. More broadly, the authors have discovered a bold new design for chiral catalysts that will inspire the development of future generations of catalysts, not only for olefin metathesis, but also for many other catalytic reactions.