ACS Medicinal Chemistry Letters
Technology Note
substitution, which can be effective for installation of
heteroatom substituents but is not broadly useful for the
formation of C−C bonds.7,8 Advances in sp2−sp2 cross-
coupling have facilitated the installation of aryl/heteroaryl
groups,9,10 but the installation of alkyl functionality remains
challenging. Common strategies rely on the use of reactive
nucleophiles that may limit functional group compatibility and
reaction scope11,12 or require multiple-step sequences to
convert alkene functionality installed via sp2−sp2 cross-
coupling into an alkyl group13 (Figure 1B). More recently,
radical-based approaches have been applied to the function-
alization of purine nucleosides via the Minisci reaction.14 In
these examples, carboxylic acids serve as radical precursors for
C−H alkylation, with regioselectivity governed by the
electronics and substitution pattern of the heterocycle
substrate.
Table 1. Reaction Conditions for Cross-Electrophile
Coupling
a
b
reaction conditions
as above
yield of 3
53%
47%
38%
4%
4CzlPN as photocatalyst
(TMS)3SiH as reductant
LiOH as base
2,6-lutidine as base
DME as solvent
MeCN as solvent
no nickel catalyst
22%
13%
8%
As an alternative, we envisioned sp2−sp3 cross-coupling of
halogenated nucleosides, which would enable direct and
regiospecific installation of saturated functionality in a single
step. By incorporating alkyl substituents, this transformation
would provide access to novel products with increased sp3
character, which has been correlated with higher success rates
from drug discovery through approval.15 Within this context,
we felt that cross-electrophile coupling would be a particularly
attractive strategy, as it would offer the advantage of drawing
from a large pool of commercially building blocks such as alkyl
bromides, resulting in increased structural diversity of the
newly accessible products. Methods for cross-coupling of aryl
halides and alkyl halides have been published using nickel,16
palladium,17 and cobalt18 catalysts in conjunction with
reducing metals such as zinc or manganese. In addition,
photoredox catalysis and electrochemistry have recently been
combined with nickel catalysis to enable aryl−alkyl cross-
electrophile couplings in the absence of stoichiometric metal
reductants.19,20
With our specific goals in mind, we chose to apply nickel/
photoredox dual catalysis to the direct cross-coupling of 6-
chloropurine with alkyl bromides (Figure 1C). Because of the
mildness of the reaction conditions generally employed in
visible light-mediated dual catalytic cross-couplings, we
expected that this method would be suitable for coupling at
a late stage and should thereby circumvent the need for
challenging multistep syntheses to be carried out in parallel.
We began our investigations by exploring reaction
conditions for sp2−sp3 coupling of nucleoside 1 with alkyl
bromide 2 to provide the alkylated derivative 3. We took our
initial inspiration from the conditions described in the original
report of photoredox and nickel-catalyzed cross-electrophile
coupling from Zhang et al.19a As a light source, we chose an
integrated photoreactor developed to improve ease of use as
well as consistency of illumination.21,22 After modifying a
number of parameters, we were delighted to find conditions
under which the fully unprotected nucleoside 1 could serve as
a suitable substrate for cross-coupling (Table 1). This result
was particularly consequential in light of our goal of designing
a strategy amenable to late-stage diversification. Because
protection of the nucleoside is not required, we envision that
coupling could be carried out as the very last step of a synthesis
using readily available alkyl bromides. As such, a single batch of
the desired nucleoside core could be prepared and then
broadly diversified in a single step without the need for
deprotection of the resulting product library, which might
comprise dozens or even hundreds of compounds.
3%
a
Standard conditions: 0.07 mmol 1, 2 equiv of 2, 2 mol %
photocatalyst, 10 mol % Ni catalyst, 1 equiv of silanol, 3 equiv of
b
1
base, 0.10 M. Determined by HNMR using mesitylene as internal
standard.
Our studies revealed DMF as the optimal solvent and
sodium carbonate as the base of choice. For convenience and
consistency of the ligand-to-metal-ratio, we employed the
preligated complex [Ni(dtbbpy)(H2O)4]Cl2 reported by
Molander and co-workers as our source of nickel.23 In
addition, while we expected that a large excess of the alkyl
halide might provide improved yields, we chose to limit the
loading of this coupling partner to 2 equiv in anticipation of
cases where the reagent might be custom-synthesized and quite
precious. Using the ubiquitous photocatalyst Ir[dF(CF3)-
ppy]2(dtbbpy)PF6 and tris(trimethylsilyl)silane, a promising
38% NMR yield of the desired product was observed. During
the course of the reaction, deleterious debromination side
products resulting from the reduction of 2 were often
observed. Replacing tris(trimethylsilyl)silane with the increas-
ingly common reagent tris(trimethylsilyl)silanol24 suppressed
this unproductive pathway and increased the observed assay
yield to 53%. We were also pleased to find that the organic
photocatalyst 4CzIPN provided comparable yields to the
optimal iridium catalyst. However, subsequent time studies
indicated that the rate of reaction using 4CzIPN was
significantly lower, and we therefore chose to proceed with
Ir[dF(CF3)ppy]2(dtbbpy)PF6 for further investigation.
During the course of our studies, we recognized the
potential for this reaction to proceed through a homolytic
aromatic substitution mechanism instead of the dual-catalytic
cross-coupling pathway originally proposed by MacMillan and
co-workers,19a given that direct radical substitution reactions
are known for related heterocyclic substrates.25 The dual-
catalytic mechanism would require nickel to undergo oxidative
addition and reductive elimination steps in order to effect the
desired coupling. By contrast, homolytic aromatic substitution
would proceed through direct addition of an alkyl radical to
the aryl chloride, followed by single-electron reduction and
elimination steps to provide the desired product. No nickel
catalyst would be required for the latter reaction pathway to
proceed. In order to test whether a direct-addition mechanism
might be operative, we performed a reaction under the newly
defined standard reaction conditions, but omitted the nickel
663
ACS Med. Chem. Lett. 2021, 12, 662−666