Journal of the American Chemical Society
Article
relative strengths of n → π* interactions in ( )-1 and ( )-2
will be better understood.
The energy of the toward isomer of dt[3.3]pCpTA-SO is 2.3
kcal mol−1 lower than that of the away isomer despite the
proximity of the sulfoxide and amide groups. This energetic
preference is a direct result of the stronger n → π* interaction
from the bridging sulfoxide oxygen to the nearest amide
carbonyl in the toward isomer (d = 2.70−2.77 Å, θ = 88−96°)
compared to the sulfoxide sulfur in the away isomer (d =
3.15−3.22 Å, θ = 84−100°). Further n → π* evidence comes
from the comparison of amide puckering (deviation of the
carbonyl atom from the Car−N−O plane, Θ) for amides in the
same molecule. In the toward isomer, the proximal carbonyl
receives nonbonding electron donation from the sulfoxide
oxygen and puckers toward the sulfoxide by 4.3−5.7°. The
other amides which do not receive electron donation maintain
their planarity with Θ ∼ 0°. The amides in the away isomer
which only receive donation from the sulfoxide sulfur also
maintain relative planarity with Θ up to 0.8°. These results
indicate that the strength of the n → π* interaction in the
sulfone ( )-2 should be considerably stronger than that in the
sulfide ( )-1. Similar computational analyses of (Sp)-1 and
(Sp)-2 reveal the presence of stronger n → π* interactions in
(Sp)-2. NBO analysis was performed to confirm orbital overlap
between the nonbonding orbitals of the sulfide or sulfone and
the amide π* (Figure 5), showing that the sulfone in (Sp)-2
results in better orbital overlap when compared to (Sp)-1.
Figure 6. DFT optimized geometries of (Sp)-1 (a) and (Sp)-2 (b)
dimers with intermolecular hydrogen bond (N···O) distances
highlighted.
Retrosynthetic analysis of ( )-1 and ( )-2 led to the
advanced tetra-carboxylic acid intermediate ( )-3a which
could come from hydrolysis of the tetra-ester ( )-4a.
Although dt[3.3]pCps with ester substituents are known in
the literature,63,72−74 an example with decks comprising two
ester units each (for a total of four) was unknown. Synthesis of
( )-4a necessitated the development of new macrocyclization
chemistry and could be envisioned to come from the known
thiol 5a and its precursor bromide 6, published in previous
work by Staab and co-workers.63 The synthesis of target
dt[3.3]pCpTAs beginning from 6 is depicted in Scheme 1.
Synthesis of 5a beginning from building block 6,75 available
in 4 steps from commercial 2,5-dibromo-p-xylene, proceeded
in the literature through a diazomethane esterification of the
corresponding 1,4-bis(thiomethyl)terephthalic acid.63 As a
safer alternative, we utilized TMSCHN2 for the synthesis of
5a, but to our dismay, the reaction proceeded in poor yields.
Attempts at Fischer esterification in methanol and sulfuric acid
at reflux led to the presumably stable, but unknown to the
literature, bis(thiolactone) which could only be characterized
Figure 5. Orbital plots of the sulfur nonbonding orbitals and the
amide π* orbitals in (Sp)-1 (a) and the sulfone oxygen nonbonding
orbitals and the amide π* orbitals in (Sp)-2 (b) obtained from NBO
analysis.
1
by H NMR, FT-IR, and GC-MS due to its incredibly poor
solubility (see SI). Despite the poor conversion of 6 to 5a,
enough 5a was accessed to attempt synthesis of ( )-4a by 1:1
coupling of 5a and 6. This reaction proceeded well to give a
mixture of the desired 5,8,14,17-regioisomer ( )-4a and its
achiral gem- 5,8,15,18-regioisomer 4b in a ca. 5:1 ratio in favor
of ( )-4a. Chemical intuition dictates that ( )-4a should be
the most stable regioisomer due to the minimization of
transannular steric interactions that are present in 4b. The
regiochemistry of ( )-4a and 4b could be further confirmed
experimentally by comparison of chemical shifts and coupling
constants with previously known tetrabromo-dt[3.3]pCps
whose structures were confirmed by X-ray crystallography
Although compound ( )-4a was obtained in small
quantities, a route to ( )-4a on larger scale was desired for
synthesis of ( )-1 and ( )-2. Revision of the synthetic route
took advantage of the in situ deprotection of a thioacetate
group of 5b to unveil a masked thiolate followed by
substitution and macrocyclization with one equivalent of 6
to give a mixture of ( )-4a and 4b. The desired regioisomer
( )-4a could be separated by careful fractional crystallization
from DCM/MeOH but was generally carried forward as a
mixture and separated as the final target ( )-1 through
column chromatography and fractional crystallization. Hydrol-
Energies of dt[3.3]pCpTAs obtained from DFT calculations
reveal a thermodynamic preference for the chair conformers of
both 1 and 2, with the chair anti-conformer of 1 favored by 0.2
kcal mol−1 and the chair syn-conformer of 2 favored by 0.6 kcal
mol−1 in the gas phase. Optimization of dimers of 1 and 2
reveals a slightly larger interaction energy for the 2 anti-dimer
(−45.9 kcal mol−1) compared to the 1 anti-dimer (−44.4 kcal
mol−1) indicating potential for stronger supramolecular
assembly of 2. Energies of the syn-dimers for both 1 and 2
were >5 kcal mol−1 higher in energy than the anti-conformers,
indicating the thermodynamic preference for anti-assembly,
consistent with previously studied [n.n]pCpTAs (Figure
6).24−26
Synthesis. The synthetic chemistry of dt[3.3]pCps differs
from that of hydrocarbon variants, as dt[3.3]pCps are readily
prepared by macrocyclic coupling of extensively prefunction-
alized 1,4-bis(bromomethyl)benzenes and 1,4-bis-
(thiomethyl)benzenes.57−64 Oftentimes, the dt[3.3]pCps
then undergo sulfur extrusion to give the hydrocarbon bridged
[2.2]pCps with precise regiochemical control.65−71
D
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX