Journal of the American Chemical Society
Article
despite the formal charge on the nitrogen in 6-ABC, is in
agreement with previous computational studies in which the syn-
periplanar geometry was found to offset the energy cost of
symmetric bending nearly as efficiently as the anti-periplanar
geometry in 1-fluoro-2-butyne.21 These direct interactions
between the alkyne π-bond and the propargylic C−N bonds
in 2-ABC and 6-ABC account for the asymmetry and alkyne
polarization.
Although the propargylic C−N bonds in 2-ABC and 6-ABC
lower the barriers to cycloaddition with azide 1 and diazo
compound 2, the two dipoles are affected differentially. For
instance, the barrier for the reaction of azide 1 with 2-ABC is
∼0.6 kcal/mol higher in energy than that for its reaction with 6-
ABC (21.9 versus 21.3 kcal/mol), whereas the same comparison
with diazo compound 2 shows a decrease of 0.2 kcal/mol (19.5
for 2-ABC versus 19.7 kcal/mol for 6-ABC).47 Such differences
can be exploited to develop chemoselective reactions between
similar dipoles that are mutually orthogonal.29,48,49
To understand the reactivity of 2-ABC with the intent of
exploiting differential reactivity toward different dipoles, we
employed distortion/interaction (strain−activation) analy-
sis.14,16,50−52 In particular, we sought to compare the reactivity
of 2-ABC with those of DIBO and DIBAC (Figure 3, Tables S2
and S3). We found that the transition states for the reaction of
each dipole with 2-ABC display the strongest interactions
(−10.9 and −14.3 kcal/mol for 1−2-ABC-TS and 2−2-ABC-
TS, respectively). Diazoacetamide 2 also provides a decrease in
distortion energies for both the dipole (16.8 kcal/mol) and the
cyclooctyne (3.7 kcal/mol) in 2−2-ABC-TS relative to both 2−
DIBO-TS and 2−DIBAC-TS.53 Meanwhile, azidoacetamide 1
displays both a dipole distortion energy (16.6 kcal/mol) and an
alkyne distortion energy (2.6 kcal/mol) in 1−2-ABC-TS that are
similar to those in 1−DIBAC-TS.
The origins of the favorable distortion and interaction
energies for both 2-ABC transition states became apparent
upon inspection of optimized geometries (Figures 3 and 4). The
propargylic C−N bond within 2-ABC facilitates bond
formation,22 resulting in shortened incipient bonds at the
internal N/C within the 2-ABC-TS relative to the corresponding
DIBO-TS and DIBAC-TS, for each 1,3-dipole.
Interactions between the aryl nitrogen in 2-ABC and the
acetamide in both dipoles 1 and 2 are evident from large
interaction energies in optimized transition state geometries. In
addition, 2-ABC is the sole constitutional isomer in which the
syn approach of substituents on the incoming dipole relative to
the azabenzo ring is favored for both azide 1 and diazo
compound 2 (Scheme 1). Having found that interactions in
both 1−2-ABC-TS and 2−2-ABC-TS enable each dipole to
overcome alkyne polarization in 2-ABC (Figure 2), we next
examined the nature of the interactions that enhance cyclo-
addition reactivity.
We found that azide 1 adopts a conformation containing an
intramolecular N···H−N hydrogen bond within a 5-membered
ring. That hydrogen bond is retained in 1−2-ABC-TS (Figure
4). There, significant stabilization via an nN→π*C=O interaction
from the aryl nitrogen to the acetamide carbonyl of azide 1 is
apparent from an N···C=O distance of 2.89 Å, N···C=O angle of
θ = 105.2°, and energy of 2.0 kcal/mol (Figure 4). Such an n→
π* interaction54 is unprecedented in a SPAAC. In contrast to
azide 1, the diazo compound 2 does not form an intramolecular
hydrogen bond. Instead, we found an intermolecular hydrogen
bond with an N···H−N distance of 2.17 Å and energy of 3.9
kcal/mol in 2−2-ABC-TS. That strong hydrogen bond leads to a
Figure 4. Comparison of interactions in 2-ABC cycloadditions with N-
methylazidoacetamide (1) and N-methyldiazoacetamide (2). (A)
Second-order perturbations obtained from an NBO analysis. (B) Key
stabilizing orbital interactions: N···C=O n→π* interaction with azide 1
and N···H−N hydrogen bond with diazo compound 2.
lower activation barrier for diazoacetamide 2 than that for azide
1.
Synthesis of ABC. We sought to experimentally test the
computational results. When considering the available synthetic
methods to access strained alkynes, we were challenged by the
inherent limitations posed by an azabenzo group. Common
routes to cyclooctynes rely on the synthesis of parent alkenes,
dibromination, and subsequent elimination of HBr (2 × ), often
requiring extended synthetic routes.30,37,38 A possible circum-
vention is the Friedel−Crafts reaction of electron-rich benzyl
phenyl ethers with tetrachlorocyclopropene, followed by
hydrolysis to generate a biaryl cyclopropenone and UV
irradiation to form a cycloalkyne.26 Unfortunately, an azabenzo
group is not compatible with a Friedel−Crafts reaction.
To overcome these challenges, we reasoned that we could
harness the high energy that is inherent in an alkylidene carbene
to accomplish a [1,2]-rearrangement that yields strained alkynes
(Scheme 2). In analogy to the Fritsch−Buttenberg−Wiechell
rearrangement55−57 (which is the second step of the Corey−
Fuchs reaction58), we sought to enlist the dehydrative
fragmentation of a 5-hydroxyalkyl-1H-tetrazole, accessed via
N-morpholinomethyl-5-lithiotetrazole that is generated in
situ.59,60 Specifically, we found that the nucleophilic addition
of N-morpholinomethyl-5-lithiotetrazole to commercially avail-
able ketone 3, which is a precursor to the antihistamine
loratadine (Claritin),61 proceeded smoothly in THF and
afforded tetrazole 4 after acid hydrolysis. Dehydration with
EDC in THF gave a tetraazafulvene intermediate, which
expelled dinitrogen to generate an unstable alkylidene
carbene.62 Its [1,2]-rearrangement63 afforded the desired
cyclooctyne. Adventitiously, this opportunistic route to ABC
affords a chloro group that is an ideal handle for functionaliza-
tion through well-established aryl chloride coupling chem-
9492
J. Am. Chem. Soc. 2021, 143, 9489−9497