10.1002/chem.201605757
Chemistry - A European Journal
COMMUNICATION
unfavorable equilibrium of the reaction in aqueous medium
(Keq ≈ 5 M-1).[5] Efforts to increase the yield of 1a by using a large
excess of nitromethane were unsuccessful, as nitromethane
spontaneously reacts with the newly formed Michael acceptor to
afford the corresponding 1,3-dinitro compound.[6] In line with the
proposed mechanism, the enzyme does not accelerate the
cleavage of 1-(4-methoxyphenyl)-2-nitroethan-1-ol or its
dehydration to 1a.
Scheme 2. Henry additions vs. condensations. A: Diverse enzymes catalyze
Henry additions of aldehydes and nitroalkanes to afford β-hydroxynitro
compounds by activating the carbonyl group of the aldehyde (for example, by
binding in an oxyanion hole). B: In contrast, RA95.5-8 utilizes amine catalysis
to promote Henry condensation of the same substrates, affording exclusively
nitrostyrenes as products.
The enzymatic Henry condensation was kinetically
characterized using
a
spectrophotometric assay. The
concentration of anisaldehyde was varied at several fixed
concentrations of nitromethane (see Supporting Information). The
intersecting lines in the resulting double-reciprocal plot suggest
fast and reversible Schiff-base formation between anisaldehyde
and Lys83. The data were fitted to a simplified equation for a bi-
substrate reaction to determine the steady-state parameters. The
Henry condensation has a ten-fold lower kcat value (0.018 ± 0.001
s-1) than the RA95.5-8-catalyzed Knoevenagel reactions,
although the Michaelis constants (KM,anisaldehyde = 193 ± 12 µM;
KM,nitromethane = 70 ± 6 mM) are relatively similar. Kinetic
characterization of the enzymatic retro-Henry cleavage of
nitrostyrene 1a gave steady-state parameters kcat = 0.056 ± 0.002
binding and ionization of nitromethane via double hydrogen
bonding, a strategy for activating nitro groups that is widely used
in organocatalysis.[7] A more systematic evolutionary campaign
can be expected to enhance this novel activity considerably.
Promiscuous additions of aldehydes and nitroalkanes to afford
β-hydroxynitro compounds have been reported for diverse
classes of enzyme (Scheme 2A).[8] Hydroxycyanide lyases, which
naturally promote the reversible attack of cyanide on aromatic
aldehydes, are particularly effective and stereoselective catalysts
for these reactions.[9] In fact, the addition of nitromethane to
aldehydes is so favorable in aqueous medium that it may be
readily attained using suitable buffers at close-to-neutral pH. On
the other hand, the subsequent dehydration of the initial Henry
adducts to afford nitrostyrenes generally requires high
temperatures or strongly basic conditions – a far more challenging
prospect for biocatalytic systems.[6, 10] RA95.5-8 circumvents this
problem by providing an alternative route to nitroolefins involving
amine catalysis (Scheme 2B). As a consequence, this enzyme
could be a valuable source of these useful synthetic building
blocks. In contrast, catalysis of Henry reactions by bovine serum
albumin, a model protein for promiscuous lysine-mediated
catalysis,[11] only affords the corresponding β-hydroxynitro
s-1 and KM, = 213 ± 20 µM. Other nitroolefins such as trans-β-
1a
nitrostyrene (1b) and (E)-1-bromo-4-(2-nitrovinyl)benzene (1c)
are also cleaved by RA95.5-8 with 1.3- and 2.7-fold higher kcat/KM
values than 1a.
KN.4, an optimized RA95.5-8 variant with 30-fold enhanced
activity for the Knoevenagel reaction,[4b] was also examined as a
catalyst for the Henry condensation, in light of the mechanistic
similarities between the two reactions. To our surprise, the
steady-state parameters (Tables S2 and S3) showed that the
optimized enzyme is actually about 20-fold less active for the
reaction than the parent retro-aldolase RA95.5-8. Because some
of the mutations present in KN.4 could conceivably still benefit the
Henry condensation, we used gene shuffling to dilute the number
of substitutions within the parent sequence (see SI).
Approximately three hundred members of the corresponding
compounds via non-specific interactions.[12]
Although the
library were assayed in multi-well plates using
a
enzymatic synthesis of nitrostyrenes is thermodynamically
unfavorable in aqueous medium, effective biocatalytic cascades
can be envisioned in which the initially formed nitroolefins are
transformed downstream either by reduction of the nitro group[13]
or functionalization of the double bond[4d, 14] Preparative Henry
condensations might also be achievable using suitable enzyme
preparations in organic solvents.
spectrophotometric assay to monitor condensation of
anisaldehyde and nitromethane at 370 nm. A variant with
improved activity under the assay conditions was sequenced and
characterized as pure protein. Of the original fifteen mutations in
KN.4, four were retained in the improved catalyst (S21R, I107V,
I113V and N135G), which exhibited a 2 to 3-fold higher kcatapp than
RA95.5-8 but a similar kcatapp/KMapp (Tables S2 and S3). This result
demonstrates that some KN.4 mutations are generally beneficial
for condensation reactions, whereas other changes boosted
Knoevenagel activity at the expense of alternative reaction
pathways. It also suggests that, as seen for other RA95.5-8
activities,[4b, 4c] it should be possible to improve the efficiency of
the enzyme-catalyzed Henry condensation by further rounds of
directed evolution. In fact, by simply screening our previously
In summary, we reported the first biocatalytic synthesis of
nitrostyrenes in aqueous medium — a new activity for the artificial
enzyme RA95.5-8. We have shown that single active site
mutations can boost this activity, or, in turn, mitigate its
competition with other enzymatic reactions involving
nitrostyrenes.[4d] Optimized retro-aldolase variants for the Henry
condensation could offer exciting opportunities for the design of
innovative enzymatic synthetic routes for producing or diversifying
a wide range of nitro compounds. Extrapolating from the
functional versatility of these substances in combination with
RA95.5-8, other activated substrates might be profitably
engineered retro-aldolase variants, we found that the K210R
app
mutation alone increases kcat
for the condensation of
anisaldehyde and nitromethane 4 to 6 fold (Table S2). The
guanidinium cation of the arginine side chain may promote
This article is protected by copyright. All rights reserved.