SCIENCE science.orgmodel by use of density functional theory
calculations showed how the Pd ligand can
be designed to inhibit the unwanted second,
nonproductive C(sp^2 )–H bond activation (see
the figure, top). The energy of the transition
state for the withdrawal of C(sp^2 )–H hydro-
gen from the first product a,b-unsaturated
carboxylate intermediate reveals why one
class of ligands was more effective (see the
figure, bottom). For the complex with the L8
ligand, which forms a five-membered ring
transition state, this reaction is disfavored
by 6.8 kcal/mol, whereas the L33 ligand that
formed a six-membered ring transition state
had no barrier. Wang et al. took advantage of
this to develop another useful reaction with
this system, that of the synthesis of g-alkyli-
dene butenolides.
This process of reaction development by
using experience, logical thinking, and com-ORGANIC CHEMISTRYDirecting carboxylic acid dehydrogenation
A palladium ligand can activate carbon-hydrogen bonds yet avoid product olefin reactions
By Yoshiharu IwabuchiE
nzymes can activate functional groups
on molecules that are normally unre-
active. For example, methane mono-
oxygenase activates carbon-hydrogen
(C–H) bonds to form methanol ( 1 ).
Chemists have been able to mimic this
chemistry using transition metal compounds
to activate C–H bonds in organic compounds
( 2 ) and convert them into value-added com-
pounds. Targeting specific C–H bonds in
alkyl chains is more challenging and has
long been thought to require enzymes, but
strategies have emerged to activate and tar-
get particular C–H bonds in the presence of
other more reactive functional groups, often
relying on the ligands on the metal centers
( 3 ). On page 1281 of this issue, Wang et al.
( 4 ) describe a catalytic system that efficiently
activates the b-methylene C–H
bonds of fatty acids to create
unsaturated acids, a reaction
that had previously been con-
sidered impossible through
direct C–H activation methods
because of the higher reactivity
of the olefin products.
Chemical routes to activating
C–H bonds must not only over-
come their high bond strengths
(enthalpies of ~100 ± 15 kcal/
mol) but also target specific
bonds and not activate weaker
bonds in other functional
groups of the molecule. One
strategy has been to use pal-
ladium (Pd) to displace H and
form Pd–C bond intermediates.
Regioselectivity requires bring-
ing the active Pd catalytic center
into close proximity to the C–H
bond. The ligand coordinating
the Pd atom plays a substantial
role, with an appropriate topol-
ogy to simultaneously promote
the bond formation with the C
and to break the bond between
C and H. For example, the b-
methylene groups of alkyl am-
ides can be targeted for aryla-
tion with Pd catalysts ( 5 ).Wang et al. extended the above approach
and developed a catalytic system that effi-
ciently proceeds with the regioselective de-
hydrogenation of fatty acids. This reaction
had not been realized because the C(sp^2 )–H
bond of the product of this reaction, a,b-
unsaturated carboxylic acid, is more reactive
toward Pd than is the C(sp^3 )–H bond of the
substrate carboxylic acid. The newly formed
double bond would react with the regener-
ated catalytically active species immediately
after the reaction proceeds to form a chelate
and deactivate the catalyst.
The authors show that the undesired reac-
tion can be inhibited with the right ligand.
Tautomeric, bidentate pyridone ligands
formed six-membered ring transition states
but had limited dehydrogenation yields be-
cause of unwanted product reactivity. A de-
tailed analysis of the reaction mechanism
putational chemistry exploited
an unexplored chemical space,
which nature has been refin-
ing through repeated evolu-
tion from chance mutation.
Although enzymes acquired by
living organisms and molecular
catalysts created by chemists
have different compositions
and shapes, they can have many
things in common. The a,b-
dehydrogenation of carboxylic
acid takes place as a central pro-
cess in fatty acid metabolism by
withdrawal of a-hydrogen from
acyl–coenzyme A thioester and
hydride transfer of b-hydrogen
to flavin cofactor in the fatty
acid hydrogenase ( 6 ). The ap-
proach of Wang et al. should
strengthen the effective meth-
odology of chemistry to make
the impossible possible and
lead to further development. jREFERENCES AND NOTES- M. O. Ross, A. C. Rosenzweig, J. Biol.
Inorg. Chem. 22 , 307 (2017). - R. H. Crabtree, Chem. Rev. 85 , 245
(1985). - T. W. Lyons, M. S. Sanford, Chem. Rev.
110 , 1147 (2010). - Z. Wang et al., Science 374 , 1281
(2021). - G. Chen et al., Science 353 , 1023
(2016). - C. Thorpe, J.-J. P. Kim, FASEB J. 9 , 718
(1995).
10.1126/science.abm4457
Graduate School of Pharmaceutical
Sciences, Tohoku University,
Aobayama, Sendai 980-8578, Japan.
GRAPHIC: N. CARY/Email: [email protected]
SCIENCE
BASED ON Y. IWABUCHI
1.33
2.17 1.422.18 1.41
1.34L8 ligand L33 ligandFive- versus six-membered rings
In the second, undesired activation step, the L8 catalyst would form a five-membered ring
that is uphill in energy by almost 7 kcal/mol. The L33 catalyst forms a six-membered ring
isoenergetic with the reactants and deactivates. Selected bond distances are in ångstroms.Hydrogen
Carbon
Nitrogen
OxygenPalladiumChlorineActive and overactive catalysts
The Pd catalysts with either the L8 or the L33 ligand activate the C–H bond to dehydrate
the acid and form the double bond product. The ligand L33 attacked the remaining C–H
bond, leading to catalyst deactivation and product binding, unlike the L8 ligand.
Abbreviations: Ac, acetate; R, alkyl.ROOHPd(OAc) 2
Ligand
Oxidant ROOPd MROR OHOOPd MH HHH1 st C–H activation2 nd C–H activationCatalyst deactivation Desired productL8 or L33L333 DECEMBER 2021 • VOL 374 ISSUE 6572 1199Stopping after the first bond
Wang et al. developed a palladium (Pd) catalyst that can activate the C(sp^3 )–H bond
of a carboxylic acid to form a double bond next to the –COOH group. An important
feature in ligand design was avoiding attack on the remaining C(sp^2 )–H bond, which
would deactivate the catalyst and bind it to the product.