Full Paper
doi.org/10.1002/chem.202005261
Chemistry—A European Journal
cules. We have also earlier demonstrated that once carbonyla-
tion conditions with 13COgen as the CO source have been opti-
mized, the same reaction conditions can easily be transferred
to carbon-14 labeling.[26]
Table 1. Optimization of the one-pot synthesis of carboxamides starting
from the in situ generated NiII-acyl complex.[a]
With the optimized conditions for the one-pot protocol in
our hands, we proceeded to investigate the substrate scope of
this methodology (Scheme 7). Initially, the efficiency of the pro-
tocol with both primary and secondary amines was investigat-
ed, and yields of 68–100% of the desired carboxamides 1–4
and the 13C-variants (13C-1– C-5) were obtained as shown in
13
Entry
Deviations from initial conditions
Solvent
Conv.
[%][b]
section A. Some structural variance in the alkyl zinc reagents
on the reaction outcome was demonstrated (section B), which
include both carboxyl esters and a heterocyclic fragment de-
rived from the pharmaceutical pentoxifylline (compounds 6, 8,
1
2
3
4
5
6
7
8
none
THF
THF
THF
78
75
66
78
80
87
87
3
47
71
[97][c]
100
NEt3 (1.5 equiv) added to amine
NiL2 instead of NiL1
MeCN instead of THF
Bn2NH (2.0 equiv)
Py2S2 (2.0 equiv)
reactions run at 708C
No Py2S2
(3-phenylpropyl)-ZnBr (1.5 equiv)
reversed stoichiometry[d]
reversed stoichiometry[d]
reversed stoichiometry,[d] COgen[e]
13
13C-6– C-8). The three compounds were obtained in good to
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
THF
excellent yields (59–91%). Furthermore, the synthesis of four
known carboxamide-bearing pharmaceuticals, including fluta-
mide (9), bucinnazine (10), ramelteon (12), olaparib (13), and
their 13C-labeled counterparts, was achieved in good to high
yields (56–81%, section C). Only in the case of praziquantel (11
and 13C-11) was the yield down to 31%, a result which can be
ascribed to the less efficient transmetallation step with second-
ary alkyl zinc reagents as previously observed.[18a] On the other
hand, reverting to use of the less bulky cyclopropyl zinc bro-
mide restored the yields as demonstrated in the case of olapar-
ib (13 and 13C-13).
Unfortunately, adaptation of this amidation protocol to acet-
amide synthesis proved unrewarding. Nevertheless, we have
earlier reported the successful application of methyl palladium
complexes as a starting point for 11C- and 13C-isotope-labeling
of a variety of N-acetylated peptides.[27] On the other hand, our
method proved ideal for accessing pharmaceutical derivatives
with longer alkyl chains, a situation which would be problem-
atic with Pd-mediated aminocarbonylations due to b-hydride
elimination.[9] This included the synthesis of six variants of
commercial pharmaceuticals, linezolide (14), agomelatine (15),
prozac (16), lacosamide (17), melatonine (18) and gefitinib (19)
along with their corresponding 13C-isotope analogs (section D).
Interestingly, in the case of 19, a piperazine derivative of gefiti-
nib, this compound was shown to possess significant activity
towards human cancer cell lines.[28]
9
10
11
12
THF
[a] All reactions were set up with a two-chamber system on a 0.1 mmol
scale (see Supporting Information for further details). [b] Conversion was
determined by 1H NMR analysis using CH2Br2 as an internal standard.
[c] Yield of the isolated product. [d] NiL1 (1.5 equiv), CO (2.0 equiv), 3-
phenylpropyl zinc bromide (1.5 equiv), Py2S2 (1.5 equiv), Bn2NH (1.0 equiv,
limiting). Reaction time of step 3 was increased to 2 h. [e] SilaCOgen was
exchanged with COgen (2.0 equiv), PdCl2(COD) (5 mol%), HBF4·P(tBu)3
(5 mol%), Cy2NMe (3.0 equiv) in THF (1.0 mL) and the reaction time for
step 1 was increased to 4 h.
portance for this sequential transformation. Upon addition of
excess alkyl zinc reagent, a deleterious side reaction was ob-
served, giving rise to the carbonylative homodimer of the alkyl
zinc reagent, namely 1,7-diphenylheptan-4-one. This also low-
ered the yield of the carboxamide product down to 47%
(entry 9). To avoid this side reaction, titration of the alkyl zinc
reagent prior to use is recommended, such that all of the re-
agent is consumed during the transmetallation step. Lastly, in
an attempt to obtain a quantitative yield of the carboxamide,
the reaction was performed with the amine as the limiting re-
agent. Using 1.5 equivalents of NiL1 in step 1, the same
number of equivalents for the disulfide in step 2, and a reac-
tion time of 2 h for step 3 did not increase the yield of the re-
action (entry 10). But when changing the solvent back to THF,
a quantitative conversion was observed, both when using Si-
laCOgen and COgen as the CO-releasing molecules (entries 11
and 12).[22,25] We propose this solvent effect to be the result of
a cleaner generation of the NiII-acyl complex as a build-up of a
13C-side product is seen when using acetonitrile (a signal at
192.8 on the 13C NMR spectrum is observed suggesting the for-
mation of a stable Ni(CO) complex). Particularly attractive is
the observation that a quantitative yield of the carboxamide
can be attained with COgen as the CO source. Unlike SilaCO-
gen, both 13COgen and 14COgen are commercially available,
opening the door for carbon-13 and carbon-14 labeling of car-
boxamides without the pre-synthesis of the CO releasing mole-
For the substrate scope, certain structures proved reluctant
to efficient N-acylation with our procedure (section E). O-Acyla-
tion was shown to be competitive in attempts to prepare the
propyl derivative of paracetamol. In the unsuccessful case of
fentanyl, we believe steric hindrance of the amine nucleophile
was most likely detrimental for the N-acylation step. On the
other hand, significant decomposition was observed in the at-
tempted adaptation of our acylation chemistry with a cephalo-
1
sporin derivative as monitored by H NMR spectroscopy of the
crude reaction mixture.
The successful demonstration of this chemistry to carbon-13
labeling of aliphatic carboxamides holds promise for its adap-
tions to labeling with carbon-14 as earlier demonstrated with
Pd-catalyzed aminocarbonylations to 14C-labeled benzamides
and benzoic acids.[26] However, for 11C-labeling, significantly
faster reaction times would be compulsory because of the
Chem. Eur. J. 2021, 27, 7114 –7123
7118
ꢀ 2021 Wiley-VCH GmbH