4
Tetrahedron
non-ionizing , UV-active peak (at rt=0.46 min) which was
the aqueous workup was omitted, the yield was decreased from
consistent with the tosylate counterion.
the previous example (42% vs. 61% yield).
Within 5 minutes, three quarters of the starting material was
consumed with four new components present. The expected
primary amine 25 was present (m/z= 637.3 for M+H). A second
major component was consistent with the intermediate
iminophosphorane 27 (m/z= 711.3 for M+H). We were surprised
that iminophosphorane 27 appeared to survive chromatography
in the presence of water/methanol +0.01% TFA, although it may
be the case that the primary amine (25) which is observed was an
artifact due to decomposition of the iminophosphorane during the
quenching & LC/MS analysis conditions. More surprising was
the observation that the main component was consistent with
bridged bicycle 28, in which the iminophosphorane partially
cyclized to form a bridged bicyclic piperidinyl phosphonium ion
Taken together, these results suggest the reaction proceeds by
intial iminophosphorane formation (27), followed by
intramolecular alkylation of the iminophosphorane nitrogen.
Iminophosphoranes have been shown to be strongly nuclephilic
8
9
at nitrogen, participating in both intra- and intermolecular
alkylations. The resultant monocycle 28 can slowly cyclize to the
tricyclic 26, either in the presence or absence of water, although
it appears in the aqueous conditions the yield is significantly
improved. In this mechanism, the amine 25 could be rationalized
as occuring through iminophosphorane decomposition either
through adventitious water or as an artifact of the LC/MS
analysis. Unanswered in this proposal however is the question of
why significant amounts of iminophosphorane 27 remained after
21 hours when the initial cyclization of iminophosphorane 27 to
bicycle 28 had occurred very rapidly within the first 5 minutes.
As our mechanistic insights are essentially based solely on
LC/MS analysis, we cannot at this time rule out other
(observed m/z= 539.3, consistent with M+ for 28). While we did
not find precedent for this type of transformation, this structure
was generally consistent with the observed reactivity. At the 5
minute time-point, small amounts of tricyclic product 26 were
already observed.
possibilities as to the course of this reaction.
At approximately one hour, almost all of the starting material
was consumed and nearly half of the total was monocyclized 28.
At this time, the amount of amine 25 observed was significantly
decreased compared to the 5 minute time point, and the fully
cyclized tricycle 26 was increased.
The structure of tricycle 26 was confirmed by single-crystal X-
ray analysis of the compound as its p-TsOH salt (Figure 5).
With tricycle 26 in hand, protected as the tosamide, the final step
was to deprotect the sulfonamide. In the event, reaction of
sulfonamide 26 with lithium naphthalenide accomplished
deprotection of this group. Isolation of the free 1,4-
Whereas in the earlier reaction, the mixture was quenched after
3
h, when complete consumption of the starting material was
diazaadamantane 10 from the reaction mixture was challenging,
so instead the resultant free amine was trapped with di-tert-
butyldicarbonate to provide the Boc-protected 1,4-
observed, this time the reaction was allowed to age overnight, as
it appeared that there was a significant amount of
iminophosphorane 27 remaining. Surprisingly, a significant
amount of solid material had crystallized from the reaction
mixture after aging overnight. These were collected and
identified as the p-TsOH salt of tricycle 26. In this case, where
diazaisoadamantane 29. This material was isolated in a
straightforward manner, and could be deprotected to afford the
free 1,4-diazaadamantane 10 as a TFA salt without purification.
Figure 4. LC/MS traces for aliquots removed from the reaction at 5 minutes (black trace) and 21 hours (red trace). Detection by UV monitoring at 220nm.