obtain relatively narrow linewidths. In this case the NMR data
were also supported by HR-ESI-MS data. In this study we reported
that during the coupling of the peptide chain to the bisindole, an
inversion of the tryptophan moiety occurred, giving rise to two
epimeric Arg8 conjugates of the VLB derivative. The identification
of the epimers was based on a comparison of the characteristic
features of the tryptophan alpha proton in the 1H NMR spectrum of
the conjugate and that of smaller model systems. From a biological
point of view it was shown that the two epimers had markedly
different bioactivity, again showing that unambiguous structural
characterization is inevitable.
The NMR spectroscopic conclusion of these investigations is that
although the big-molecule character of bisindoles is even more pronounced
when they are bound to other fragments, with adequate
experimental parameters and model systems ambiguities can be
overcome.
3.4. Conformational aspects of VLB analogues
In 1969 the X-ray structure of VCR-methiodide [1,80] was determined,
leading to the general conclusion that the structure of both
VCR and VLB had thus been fully established. However, the motivation
to find more efficient but less toxic VLB analogues initiated the
exploration of the structure–bioactivity relationship of bisindoles,
leading to attempts to determine their conformational features in
solution. The earliest attempt to gain insight into the solution conformation
of VLB was performed by de Bruyn et al. in 1980 [56]. The
authors in this study gave complete 1H NMR assignments measured
at 360 MHz in different non-polar solvents and binary solvent mixtures
for VLB, the latter being used to avoid spectral congestion
in the 1H NMR spectrum. Based on the determined coupling constants,
the conformation of the velbanamine moiety in VLB using a
Karplus-type equation was deduced. This conformation included a
slightly flattened chair conformation for the piperidine ring with NC(
5) bond being in quasi-equatorial position, while the azanonene
ring was deemed to be in a boat-chair conformation, stabilized by
an intramolecular hydrogen bond between the C(16) COOMe and
the C(20) OH.
Later Hunter et al. [81] showed that this 3D structure in solution
was erroneous, drawing attention to the fact that the ‘classical’
strategy of assigning a 1H spectrum on the basis of chemical shifts
and multiplicities, followed by a conformational analysis using coupling
constants, fails in the case of these compounds. According to
the authors the main reason for this failure is that bisindoles contain
numerous small spin systems isolated from each other. In order
to obtain a more reliable stereochemical picture, this couplingconstant-
based conformational assignment was complemented by
the spatial connectivity information determined from NOE spectroscopy.
Based on these connectivities the authors concluded that
the velbanamine moiety of VLB took a conformation where the
piperidine ring was severely flattened and both the ethyl group and
the nitrogen lone pair were in equatorial positions. Besides these,