A. V. Dix et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2522–2524
2523
THF, CH Cl , CHCl , acetone, CH CN, and Et O). To
3
there was little enantiopreference (ca. 4% ee by optical
8
rotation ), so 6 was nearly racemic.
2
2
3
2
our surprise, the diastereomeric salts separated cleanly
by flash chromatography (2:1 EtOAc/CH Cl to remove
2
2
the dibenzoyl-D-tartrate salt of (+)-1, followed by meth-
anol to isolate the dibenzoyl-D-tartrate salt of (ꢀ)-1).
Unfortunately, the less polar dibenzoyl-D-tartrate salt
of (+)-1 was missed during UV monitoring of collected
fractions due to extensive dilution by band broadening.
The dibenzoyl-D-tartrate salt in the more polar metha-
nolic fractions was concentrated in vacuo, basified with
Figure 2 illustrates the effectiveness of each compound,
averaging the data (see Table 1 in Supplementary Mate-
rial). (±)-6 has modest activity, reducing bacteria growth
over 30 percent at 40 lg/mL. Carbamate (±)-5 had no
antibacterial effect. (±)-1 had good antibacterial potency,
reducing growth by almost 85 percent at 40 lg/mL but
failed to have any significant effect at 10 lg/mL. (ꢀ)-1
produced the most promising results, with over 99%
reduction of bacteria at 40 lg/mL and 50% reduction
at 10 lg/mL. Note that the micromolar concentration
range for these different molecular weight compounds
is narrow (0.113–0.165 lM at 40 mg/mL) and unlikely
to contribute to the dramatic effects observed.
1
0% NaOH (aq), and extracted with ethyl acetate to
20
obtain (ꢀ)-1 ½aꢁ ꢀ32.3 (c 0.0296, CHCl ) spectroscop-
D
ically identical to its racemate. H NMR spectra of the
3
4
1
6
(
R)-MTPA amide of (ꢀ)-1 versus that of (±)-1 showed
that (ꢀ)-1 was homochiral (>95% ee). The absolute con-
figuration assigned to (ꢀ)-1 is consistent with that of all
2
,5,7–9
known (ꢀ)-hexahydropyrrolo[2,3-b]indoles.
In this study, (±)-1 produced moderate inhibition of the
growth of S. aureus, while (ꢀ)-1 gave an impressive 43-
fold greater inhibition (see Table 2 in Supplementary
Material). (ꢀ)-1 is the eutomer. For antibacterial poten-
From 3, two derivatives of (±)-1 were also prepared
Scheme 2). Alkylative cyclization of 3 to furnish (±)-5
(
was based on a quinine (35% yield) or quinidine (53%
yield) modification of Ganesan’s original approach with
1
cy, hydrophobic substitutions at N appear to be pre-
1
0
DIEA, which did not work in our hands. Reduction of
±)-5 with 10 equiv of lithium aluminum hydride
ferred over hydrophilic ones, both in this study and in
1
,2
(
previous work. In contrast, much greater antibacterial
potency was obtained when N was unsubstituted (a
1
0
8
produced racemic debromoflustramine B, (±)-6, in
8–74% yield. As expected at room temperature, even
though quinine and quinidine are homochiral amines,
6
hydrogen bond donor/acceptor). Future studies should
also explore the relevance of substitutions on the ben-
zene ring, the effect of different substituents at 3a, hu-
man cell toxicity, and the scope of antibacterial activity.
a
b
Acknowledgments
3
N
N
N
N
H
H
CO2CH3
CH3
We are indebted to the Department of Biological Scienc-
es at Salisbury University for their advice and material
contribution to the microbiological studies in this pro-
ject. We particularly thank Julie Meeks for her assistance
with microbiological studies. A.V.D. and C.M.M. also
received financial support from a Henson Student
Research Grant. We also thank Dr. Fred Kundell
(
+)-5
(+)-6
Scheme 2. Reagents and conditions: (a) prenyl bromide, Bu
Zn(OTf) , quinine (35%) or quinidine (53%), toluene; (b) LiAlH
10 equiv), THF, reflux (68–74%).
4
NI,
2
4
(
Figure 2. Percent inhibition of the growth of Staphylococcus aureus colonies at 10 and 40 lg/mL concentrations of inhibitors relative to DMSO
control (100% growth).