– the reduction of carbonyl compounds is one of the most important synthetic reactions – the catalytic enantioselective reduction of C=O has been achieved using:
* chiral oxazaborolidines and other related boronates (H3B as a source of hydrogen)
* transition metal catalysts (H2 as a source of hydrogen)
– first described by Itsuno et al. who observed that valinol reacts with 1 mol eq. of borane by producing 1 ml eq. of hydrogen gas and giving rise to the alkoxyborane derivative shown below:
– the aminoalkoxyborane derivatives (A and B) shown below are a result of the reaction of valinol with 2 eq. of borane (producing 2 mol eq. of hydrogen gas)
– the resulting aminoalkoxyborane (A or B) was found to catalyze the enantioselective reduction of PhCOMe
– the optical yield of the reduction was found to depend on the relative amounts of valinol and borane
– maximum optical yeild is reached with a borane-valinol ratio of 2.0
– the optical yeild remains almost constant within the borane-valinol ratio range of 2.0-3.0
– Itsuno et al. observed significantly higher optical yeilds when the hydrogens attached to the carbon atom of the terminal hydroxyl group were replaced by bulky groups, such as phenyls
– in the case of ketones other than aromatic ones optical yeilds were lower
– the optical yeild increases with the increasing difference in the volume of the substituents of the ketone
– an unusual relation between the optical yeild and reaction temperature was observed (studied by Itsuno et al. using methyl-tert-butyl ketone as a test system)
– the catalyst was found to work with more efficiency near 0 °C than at -78 °C
– reductions of functionalized ketones were studied
Other related reductions: (by Itsuno et al.)
– best optical yeilds were observed in the case of halohydrin formation
– the halohydrins were converted to form optically active epoxides without rasemization
– the reduction works best with chlorinated acetophenones
– two years latter Corey et al. developed the ideas of Itsuno et al. further and described a new and better catalyst (an oxazaborolidine derived from diphenylprolinol)
– the oxazaborolidine derived from diphenylprolinol gave better enantioselectivities for arylalkyl ketones than diphenylvalinol based derivatives
– Corey et al. proposed a mechanism for the catalytic reduction
– proline based oxazaborolidines are also known as CBS (Corey,Bakshi,Shibata) catalysts
– the better performance of CBS catalysts, relative to the performance of valinol-based catalysts, was related (by Corey et al.) to the higher angle strain on the partial B=N double bond at the 5,5-ring fusion
– the angle strain disturbs PI-resonance (A) and exposes the lone pair of the nitrogen atom (B) for borane to coordinate
– in THF (needed to stabilize highly polar reactive intermediates) the borane atom is not totally coordinated to the catalyst:
– the bicyclic (CBS) catalyst is capable of binding the borane more tightly than the related monocyclic system
– the more strained the B-N bond, the higher the proportion of catalyst present as a borane complex (ready to operate as a chiral catalyst)
– computational studies on the CBS catalyst indicate that not all atoms adjacent to the borane and nitrogen atoms of the partial B=N bond lie in the same plane (for the related torsion angles at 0 °C/180 °C +/- 22 °C see THA 3,1563(1992))
– similar distortions were not observed with monocyclic oxazaborolidines
– the rigidity of the structure of CBS catalysts would also orient the borane to coordinate selectively on one of the faces of the oxazaborolidine ring
– coordination on the faces would involve:
* an attack on the less hindered side of the ring system (kinetic control)
* the formation of a 5,5,-cis-fused ring system is favoured over that of the highly strained 5,5-trans-fused system (thermodynamic control)
– in the formation of borane adducts of CBS catalysts only one adduct (lowering angle strain) is formed selectively
– other isomers of borane-oxazaborolidine adducts have also been considered, e.g.
– the system containing a hydride-bridged 6-ring was found to be more stable than the other diborane adducts
– the formation of hydride-bridged adducts indicates that the hydrogens of borons “scramble” in a mixture of borane and oxazaborolidine(s)
– this hydrogen – deuterium exchange “scrambling” has been observed experimentally [Tlahuext and Contreras, THA 5, 395 (1994)]
– an X-ray study on a borane adduct of a CBS catalyst (a B-methylated derivative) proves that the borane atom coordinates to the nitrogen atom
– the X-ray structure of the N-adduct proves that the formation of N-adducts is possible and probably even favoured over the other adducts; nevertheless, the involvement of borane O-adducts of oxazaborolidines (as reactive intermediates) cannot completely be ruled out
– the mechanism of catalysis in the case of monocyclic systems has been proposed to be controlled by factors partially different from those controlling CBS catalysis
– the selectivity of the formation of borane cis/trans-adducts of monocyclic oxazaborolidines (e.g. those derived from valinols) has been calculated to be too low to fit the experimentally observed enantioselectivities, e.g. in the case of the simple model shown below:
– computational studies on simple models imply that the next step in the mechanistic cycle of catalysis should show significant selectivity
…………
– one of the most significant consequences of the N-coordination of borane to an oxazaborolidine is the substantially enhanced acidity of the ring boron [intramolecular stabilization through the partial PI-bond between the boron and adjacent nitrogen atom is not possible in the N-adduct]
– computational studies on the formation of N-, O- and N,O-(di)adducts (related to LUMO energies) imply that the parent oxazaborolidine is the weakest Lewis acid (highest LUMO energy), the borane N,O-diadduct being the strongest (lowest LUMO energy)
– the more the borane coordinates to the N- and O- atoms of an oxazaborolidine ring the less the ring boron is stabilized by partial PI-bonding
– not only are there differences in the Lewis acidities of the borane N- and O-adducts, but there are also many possible orientations from which a Lewis base (in this case a ketone) can best approach the ring boron
– in the case of borane N-adducts the orientation of the dipole moment favours the coordination of ketones
– in the case of borane O-adducts the orientation of the dipole moment is not particularly favourable; the incoming Lewis base has to approach the ring boron in the plane of the ring (this inhibits binding)!
– the orientation of the dipole moment of the borane N-adduct of the parent oxazaborolidine implies that the ketone (or any Lewis base) could react to form a borane-ketone cis-adduct
– the 5,5-diphenyl substituents direct the ketone to favour the anti-conformation over the syn-conformation (see the figure below)
– the structures of both the syn- and anti-adducts a borane-formaldehyde complex coordinating to the parent oxazaborolidine have been generated and assessed using computational methods
– these simple models, extended with two phenyl groups on C-5 of the oxazaborolidine ring (the orientations of the phenyls were set on the basis of the orientations of the corresponding hydrogens), show how hindered the syn-conformation is in the case of oxazaborolidines bearing bulky substituents on C-5
– plausible conformations of both the anti- and syn-adducts and the related transition states of the hydride transfer have been studied computationally
– the hydride transfer taking place in the borane-ketone adducts of oxazaborolidines has been proposed to lead to the formation of an intramolecular adduct of an alkoxyborane, which in turn results in the formation of an aminoborane (A)
– the aminoborane can react further to form an oxazadiboretane (structure B)
– the oxazaboretane system (B) may undergo a number of reactions, one of which leads to the regeneration of the catalyst, whereas another leads to the formation of an alkoxyborane adduct analogous to the original borane adduct of oxazaborolidine
– NMR studies performed on the products of the related stoichiometric reduction carried out in the presence of Et3N gave the alkoxyborane
– a few other examples:
– further NMR studies on the products formed in the reduction of acetaldehyde with the same catalyst led to the structural interpretations shown below:
– on the basis of data obtained with 13C-NMR studies, it is not clear whether the species the signals originate from are oxazadiboretanes or their related openchain isomers (of which the latter ones are shown in the figure above)
– the results indicate that the first (of the two) hydride transfer(s) occurs with higher enantioselectivity than the second
– in addition to the mechanism of the regeneration of the catalyst discussed above, another plausible pathway has been proposed on the basis of computational studies carried out on hydride-bridged adducts of borane coordinated to oxazadiboretanes
– the regeneration of oxazaborolidine catalysts used in the enantioselective reduction of ketones was proposed to involve the hydride-bridged adduct shown below (two conformers; H2C=O as a model of the ketone and the parent oxazaborolidine as a model of the catalyst)
– the energies of the insertion of borane into oxazadiboretanes are rather low relative to those involved with most energy requiring/liberating steps in the reduction
…………
– the latest mechanistic proposal (on the basis of a computational AM1 study) is shown below:
– in contrast to the results of NMR studies on the model reaction of CBS reduction (H3C-CHO as a model of ketones), an alkoxyborane adduct (structurally analogous to that of the related borane adduct) is not included in the mechanism
– although the mechanism of the CBS reduction is not completely clear at this time, the stereochemical outcome of the reduction can easily be predicted
– boranes other than BH3 can also be used as a source of hydrogen in CBS reductions; e.g. catecholborane shown below
– in addition to the enantioselective synthesis of epoxides (Itsuno et al.), the products of these enantioselective reductions have been converted to many valuable compounds
a) The enantioselective synthesis of ALPHA-amino-acids (including unnatural ones)
b) The enantioselective synthesis of ALPHA-hydroxy-acids
c) The enantioselective synthesis of 1-deuterio primary alcohols
d) The enantioselective synthesis of benzylic thiols
e) The enantioselective synthesis of oxiranes
– in the case of CBS reductions, the coordination site of the ketone is usually determined by the difference in bulkiness of the substituents (RL and RS), but other selection mechanisms also exist
– any effect making one of the two lone pairs on the carbonyl oxygen atom of a ketone more basic than the other should work, e.g. CBS reduction of benzophenones
– the lone pair “a” (trans to the donating group) should be more basic
– two transition states:
– the formation of A (stabililzed by PI-electron donation from the p-OR group) should be favoured over B
– the stereochemical outcome of the reduction corresponds to the transition state A
– high selectivities are observed (although both substituents of the ketone being reduced are almost equally bulky)
– polymer-bound chiral oxazaborolidines have also been shown to work with enantioselectivities similar to those of their free monomeric analogs, e.g.
– the reduction of acetophenone using this polymeric catalyst gave 95% ee (the corresponding monomeric catalyst gave 97% ee)
– it has been shown that oxazaborolidines of which the basicity of the ring nitrogen has been reduced can also be utilized in the enantioselective reduction performed using oxazaborolidines
– in these catalysts:
* the basicity of the ring nitrogen has been reduced by an electron withdrawing substituent (e.g. Me-SO2)
* one of the bulky 5,5-substituents has been removed (one face of the ring has been made more accessible than the other)
* the bulky 5-substituent of the H3B O-adduct affords an axial conformation in which it:
a) is almost orientated against the plane of the sp2-hybridized oxygen of the ring, e.g.
b) substituents 4 and 5 are trans about the ring (otherwise repulsive interactions between the substituents will exist)
the substituents of the ring:
c) will block one face of the borane-oxazaborolidine O-adduct (the incoming ketone can approach the Lewis acidic boron with greater ease than from the face opposite the 5-substituent)
d) will direct the coordination of the incoming ketone towards an equatorial conforamation [the substituent of the boron (e.g. H) being cis to the bulky 5-substituent]
e) orient the coordinated H3B optimally (if the 5-substituent is in an equatorial conformation, the borane bound to the ring oxygen will reside far from the carbonyl carbon of the coordinated ketone)
(a complex in which the ketone is in an equatorial conformation would have even more problems)
