Organic Letters
Letter
irradiation conditions, doubling the concentration of 1 does
not improve the yield (entry 4).
Scheme 1. In Situ Derivatization of Acyl Chloride
Intermediate 2
a
The use of UV-C and UV-A lamps was then investigated.
With a UV-C (λ = 100−280 nm) lamp, the yield was
decreased to 32% (entry 5). Considering reactor fouling was
also observed, no further optimizations were carried out with
conditions, replacing UV-B with UV-A (λ = 340−400 nm) also
led to a lower yield (49%) for the same irradiation time. Under
40 W blue LED (λ = 450 nm) light, no product could be
observed after 8 h of irradiation (Table 1, entry 7)
demonstrating that the visible region (λ = 400−700 nm) of
the emitted spectrum by halogen or mercury-vapor lamps
initially reported16,17 is not optimal. With regard to the large
excess of oxalyl chloride used originally, doubling the
concentration did not significantly increase the yield of 1
whereas decreasing the concentration 2-fold still gave trimethyl
1,2,4-cubanetricarboxylate 3 in 55% yield (entries 8 and 9).
Using the optimized conditions, replacing oxalyl chloride with
methyl chlorooxoacetate led to the desired product 3 in 8%
yield (entry 10). Finally, as expected, a control experiment
performed in the absence of light gave no product (entry 11).
Using the optimized conditions, the formed acid chloride
intermediate 2 was then trapped with various nucleophiles to
demonstrate further in situ derivatization while keeping the
two original ester groups intact. Several 1,2,4-tricarbonyl
cubane derivatives have been prepared (Scheme 1). Hydrolysis
of 2 gave the crude carboxylic acid 4 in 56% yield, similar to
that obtained by Irngartinger17 (cf. Figure 1). Doubling the
amount of 1 and increasing the reaction time to 8 h led to
isolation of 4 in 44% yield (Scheme 1). While 3 was obtained
subsequently by the addition of methanol after concentration
of the acyl chloride intermediate 2, attempting to trap the acyl
chloride similarly with more complex and bulkier alcohols at
room temperature did not lead to isolation of the
corresponding 1,2,4-triester cubanes, illustrating the hindered
nature of this acid chloride group. However, refluxing 2 in
toluene in the presence of 1 equiv or a slight excess of
(−)-menthol and bile acid methyl ester led to the
corresponding desired product 5 and 6 in 30% and 37%
yield respectively (Scheme 1). The use of hydroxyphthalamide
at room temperature enabled us to obtain the redox active
ester 7 in low yield. In most cases, excess triethylamine (5.0
equiv) was added in order to avoid protonation and
precipitation of the nucleophiles due to the formation of
hydrochloric acid during the photochemical step. Changing
alcohol for cyclohexyl mercaptan afforded 8 in 43% yield.
Trapping the acyl chloride 2 with ammonium chloride in the
presence of triethylamine led to the isolation of the
corresponding amide 9 in 22% yield. Primary amines such as
p-fluorobenzylamine and the bulky tert-butyl amine gave the
amides in 43−47% yield. Amide bond formation using the
para-anisidine cubane isostere19d led to 12 in reasonable yield.
Because aminocubanes are unstable,4 it was added as the
hydrochloric acid salt. Electron-poor and -rich anilines and
ampyrone gave the corresponding amides 13−15 in moderate
and good yields, respectively. Finally, secondary amides 16 and
17 obtained with morpholine and piperidine, which are
abundant drug substructures, were synthesized. Unfortunately,
attempts to form a carbonyl−carbon bond through Friedel−
Craft acylation in the presence of N-methyl-indole did not lead
to the corresponding ketone 19, probably due to the unlikely
a
Isolated yield based on 0.6 mmol of 1; 7−10% of 1 was recovered
after the reaction.
formation or stabilization of the corresponding cubyl acylium
ion.
With these 1,2,4-trisubstituted cubane derivatives in hand,
further transformations were explored. Starting from the ortho-
carboxylic acid 4, the redox-active ester 7 was obtained in 92%
5166
Org. Lett. 2021, 23, 5164−5169