ACS Catalysis
Research Article
Scheme 3. Reaction Scale-Up
the case of 34−37, is challenging under photochemical
conditions due to a fast (photo)catalyzed protodeborylation
process in the presence of a base. In the case of 36, a
continuous-flow approach was particularly beneficial to
increase the yield as a result of the shorter reaction time and
hence increased efficiency.
because of its ability to act as a carbonyl mimicking group,29
with higher stability toward the in vivo metabolism. We
therefore evaluated the scope with respect to a broad variety of
trifluoromethylalkenes using primary and secondary BAs.
Alkyne-bearing substrates (Scheme 2, 64, 65) underwent the
reaction smoothly, without observing the concomitant addition
of the radical on the alkyne moiety.
To further broaden our scope, a wide range of Michael
acceptors were tested. The Giese products were obtained in
good to moderate yields (41−56). The product formation was
conditioned by the electrophilicity of the alkene, thus
providing an excellent yield in the case of 2-benzylidinemalo-
nonitrile (45) and a lower yield in the case of acrylonitrile
(44). It is worth mentioning that in the case of benzyl acrylate
(42), flow conditions were found to be particularly beneficial,
with a lower formation of byproducts probably derived from
the labile benzylic scaffold.26 We next turned our attention to
the functionalization of dehydroalanine (Dha), an amino acidic
residue whose unsaturated backbone can serve as a radical
acceptor. Indeed, in recent years, the seminal work by Davis’
group on “post-translational mutagenesis” using the dehy-
droalanine residue as free-radical trapping has inspired many
researchers to install side chains on peptides and proteins,
employing several radical sources to engage in this trans-
formation.27a In this context, we decided to utilize our
methodology for the functionalization of protected Dha,
affording a variety of structurally modified unnatural protected
amino acids and the protected amino acid leucine itself in good
yields (50−56). As already highlighted by Molander, the N-
(Boc, Boc) protection of Dha results in higher reactivity
because of the lower electron density on the alkene moiety in
comparison with the singly N-Boc-protected Dha.27b Interest-
ingly, the Dha-containing dipeptide 56 could be functionalized
as well in flow (no desired product could be observed in
batch), unveiling the possibility for the selective modification
of peptide residues in complex organic and biologically
relevant structures.
Elimination Scope. We then further proved the versatility
of this protocol by applying it to intramolecular radical−polar
crossover elimination reactions, namely, defluorinative alkyla-
tion and allylation reactions. Under the conditions reported by
Ley’s group,6a the fast protonation of the intermediate
carbanion limits the applicability for elimination reactions.
Xu and co-workers circumvented the problem employing a
non-protic solvent (DCM) but still used a Lewis base in a
defluorinative alkylation method.28 Pleasingly, under our
additive-free conditions, the same E1cb-type fluoride elimi-
nation predominated over protonation and delivered the
desired gem-difluroalkene. This moiety is indeed interesting
Interestingly, we could functionalize heteroaromatic styryl
systems bearing the indole moiety (72) and a benzo[d][1,3]-
dioxole group (66, 67) under mild reaction conditions. In
addition, a −BPin-substituted styryl system could also be
successfully alkylated in moderate yield without the loss of the
pinacolborate scaffold (74). Notably, a derivative of mestranol
was synthesized in moderate yield (75, 50%), proving the
functional group tolerance of this mild approach in natural
products. With a similar elimination mechanism, allylation
reactions were performed as well (76−82). The desired
allylated products were obtained in moderate yield in the case
of phenyl allyl sulfone (76−79), while increased yields could
be achieved installing a more electron-withdrawing group such
as an ethyl ester in the starting material (80 and 81). The
competitive protonation step especially lowered the yield in
the case of 77, where the Giese product was predominantly
formed (77′).
Limitations. During the evaluation of the substrate scope,
we found that methyl BA was unreactive under our optimized
conditions in both batch and continuous flow. Similarly, tert-
butyl BA only gave a modest yield (15%), while the use of aryl
BAs was restricted to electron-rich moieties and only afforded
the product deriving from a fast hydrogen atom-transfer step
involving the highly reactive aryl radical.
Reaction Scale-Up. Having demonstrated the wide
applicability of this reaction manifold, we sought to prove its
robustness via performing large-scale continuous-flow reaction.
Increasing the reaction scale to 2 mmol in flow in a 10 mL
reactor, the alkylated product from 2-vinyl pyridine was
afforded in 68% yield (throughput: 142 mg/h), opening the
possibility to achieve a successful scale-up. The same reaction
in batch could only afford 6.09 mg/h. Similarly, the 2 mmol
reaction employing Dha as a radical acceptor delivered the
unnatural amino acid presented in Scheme 3 in 75% yield
(throughput: 334 mg/h). These results clearly show the
superiority of flow chemistry for an efficient scale-up of
photochemical reaction.
CONCLUSIONS
■
In conclusion, we have demonstrated that the oxidation
potential of alkyl BAs can be tuned by means of a hydrogen-
10868
ACS Catal. 2021, 11, 10862−10870