compounds, because 1 had 6,8-anti stereochemistry. In related
compounds with the same tricyclic skeleton 3,6-syn- and 3,6-
anti-isomers gave diagnostic carbon chemical shift values
corresponding to each skeleton and long-distance substituent
effects were hardly observed.10 Therefore, we assigned 3,6-
syn stereochemistry for 1.
In 6 H-10a and H-12a were coupled to H-11 by 8.5 Hz
indicating that H-11 was axial, suggesting that 6 could be
represented by either 6a or 6b (Figure 3). Structure 6a
Merobatzelladine B (2) had a molecular formula of
C19H36N3 as the free base, which was analyzed by HRESIMS.
The 1H and 13C NMR data of 2 coincided well with those of
1 except for the absence of the olefinic carbons. Interpretation
of the COSY and HSQC data and comparison of the NMR
data with those of 1 demonstrated that 2 had the tricyclic
portion identical with that of 1: the identity of the relative
stereochemistry was implied by the superimposable NMR
data and confirmed by the ROESY data. The molecular
formula and the NMR data suggested that 2 had two saturated
linear side chains composed of 10 carbons in total. The
lengths of the side chains were determined by interpretation
of 13C NMR data. 13C chemical shifts of three carbons from
the terminus of the side chains were assigned by the HMBC
data as δ 14.3 (2C; C-5′ and C-5′′), 23.6 (2C; C-4′ and C-4′′),
and 32.8/32.9 (C-3′ and C-3′′), whereas HSQC spectrum
allowed the assignment of C-1′/C-1′′ and C-2′/C-2′’ signals,
leaving no unassigned carbon signals. Therefore, both side
chains are n-pentyl group.
Then, we sought to determine the absolute stereochemistry
of merobatzelladine B (2). It was reported that related cyclic
guanidines could be converted to the triamines by reduction of
the guanidyl group followed by acidic hydrolysis.11 The
resulting triamine would be a good substrate for the modified
Mosher analysis. However, our attempts to reduce 1 or 2 with
NaBH4 were unsuccessful. Then we turned our attention to use
two of the nitrogen atoms as a scaffold to introduce a fused
six-membered ring in which a secondary hydroxyl group was
preinstalled (Scheme 1).11 Even though 1 decomposed under
Figure 3. Possible structures of 5 (5a and 5b) and 6 (6a and 6b)
with NOESY correlations shown by solid arrows.
accounted for all the observed NOESY cross peaks. How-
ever, NOESY data were not consistent with 6b: the distances
between H2-1′ and H-10b are shorter than those between H2-
1′ and H-10b, but cross peaks not between the former pairs
but between the latter pairs were observed; the same was
true of the NOESY cross peaks and the distances between
protons on C-1′′ and C-12. Therefore, 6 should be represented
by structure 6a. With the structure of 6 established, 5 should
be the diastereomer of 6 differing in the stereochemistry at
C-11. H-11 in 5 was in the equatorial position because all
the C-10 and C-12 methylene protons appeared as a broad
doublet. Therefore, compound 5 was represented by structure
5a, because structure 5b was a conformational isomer of 6a.
The absolute stereochemistry of 6 was determined by
application of the modified Mosher’s method.12 Compound
6 was converted to the (S)- and (R)-MTPA esters (7 and
8, respectively). In both 7 and 8, the OMTPA group
Scheme 1
.
Introduction of the Fourth Ring System in
Merobatzelladine B (2)
1
occupied the axial position as demonstrated by the H
NMR data. Even though the distribution of the signs of
chemical shift differences in axial esters are not completely
uniform, previous studies suggested that this methodology
could be applicable to the axial esters.13,14 Analysis of the
(7) Breitmaier, E.; Voelter, W. Carbon-13 NMR Spectroscopy; VCH:
Weinheim, 1990; pp 236-238.
(8) Heys, L.; Moore, C. G.; Murphy, P. J. Chem. Soc. ReV. 2000, 29,
57–67.
(9) (a) Nagasawa, K.; Koshino, H.; Nakata, T. Tetrahedron Lett. 2001,
42, 4155–4158. (b) Snider, B. B.; Busuyek, M. V. J. Nat. Prod. 1999, 62,
1707–1711. (c) Cohen, F.; Overman, L. E. J. Am. Chem. Soc. 2001, 123,
10782–10783.
the reported reaction condition to conjugate with 1,3-ditosyloxy-
2-benxyloxypropane, 2 was stable enough to be converted to a
mixture of diastereomeric tetracyclic derivatives 3 and 4. They
were separated by HPLC and the benzyl group was removed
by catalytic hydrogenation to afford 5 and 6 whose relative
stereochemistry was assigned by interpretation of the NMR data
in DMSO-d6 as described below.
(10) Snider, B.; Chen, J. Tetrahedron Lett. 1996, 39, 6977–6980.
(11) Wang, Q.; Lo¨nnberg, H. J. Am. Chem. Soc. 2006, 128, 10716–
10728.
(12) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem.
Soc. 1991, 113, 4092–4096.
(13) Oku, N.; Matsunaga, S.; Wada, S.; Watabe, S.; Fusetani, N. J. Nat.
Prod. 2000, 63, 205–209
.
Org. Lett., Vol. 11, No. 12, 2009
2657