1930 Journal of Natural Products, 2004, Vol. 67, No. 11
Notes
spray ionization, using a Symmetry C18 column with a gradient
mobile phase of acetonitrile and water. HPLC purification was
performed with a Waters 515 pump and 2487 UV detector,
using a Zorbax RX-SIL normal-phase column (10 × 250 mm).
Molecular modeling was achieved by ab initio calculations with
a Hartree-Fock database using a Spartan 2004 system.
Solvents and other chemicals were purchased from Fisher
Scientific and Sigma-Aldrich Corporation.
Bioassays. Feeding assays with the predatory fish Fundu-
lus heteroclitus were performed in aquaria at the Georgia
Institute of Technology’s marine facility on Skidaway Island,
Georgia, as previously described.1 Consumption of treated and
control foods were compared using a Fisher’s exact test with
an alpha value of 0.05.
synthetic 1 (δ 6.81), whereas the methine proton of
synthetic 2 resonated at δ 6.04 (all in CDCl3).
Additional evidence supporting the identity of the natu-
ral product as 2,3,4-tribromopyrrole (1) came from spectral
analysis of intermediate 5. Strong NOEs were observed
between H-5 and both the methyl and methine protons of
the protecting group of 2,3,4-tribromo-1-(triisopropylsilyl)-
pyrrole (5). Because ab initio calculations predicted a 2.1-
2.5 Å separation of methine/methyl protons and H-5 for 5
and 4.8-5.2 Å separation if the pyrrole methine was at
position 4, the observed NOEs favor the hypothesis that
the pyrrole methine was adjacent to the nitrogen atom
rather than at position 4 of the pyrrole system. Thus, it
appears that a silyl protecting group did provide the steric
bulk to overcome the electronic factors involved in deter-
mining regioselectivity of this electrophilic substitution
reaction (resulting in bromination at positions 2, 3, and
4), whereas unprotected pyrrole (3), subjected to the same
bromination reagent, was selectively transformed to the
electronically favored product 2. Taken together, the
spectral data support our hypothesis that the tribrominated
pyrrole natural product of S. kowalevskii is indeed 2,3,4-
tribromopyrrole (1) and not 2,3,5-tribromopyrrole (2).
When predatory fish (Fundulus heteroclitus) were offered
squid-based food pellets containing synthetic 1 at the
concentration at which the natural product is found in S.
kowalevskii, eight of 14 fish rejected pellets containing
synthetic 1, whereas all fish consumed control pellets
(squid pellets without 1), indicating a significantly deter-
rent effect of 1 (p ) 0.002), similar to the deterrence of
natural 1.1 The six fish that consumed the treated pellets
appeared to be larger than the other eight, raising the
possibility that larger individuals who generally eat more
may be less sensitive to deterrent compounds. Thus, these
six fish were fed control pellets to near-satiation and then
were offered treated pellets a second time. Although three
of these fish were too satiated to eat pellets of any kind in
the second experiment, the remaining three fish rejected
pellets treated with 1 and then consumed a control pellet,
suggesting that the effectiveness of chemical defenses is
affected by hunger status.
Compared to the deterrent properties of 1, pyrrole (3)
at the same concentration was palatable to F. heteroclitus,
as all 12 fish offered pellets containing 3 consumed both
treated and control pellets. This indicates that bromination
plays a crucial role in the deterrence of 1, although
Kicklighter et al.1 showed that many brominated aromatic
compounds from marine worms do not deter predators at
natural and greater than natural concentrations. Unfor-
tunately, we were unable to test 2 due to its lability.
In conclusion, for the first time 2,3,4-tribromopyrrole (1)
has been unambiguously and regioselectively synthesized.
The NMR spectral data for synthetic 1 agree with previ-
ously reported data for 1 isolated from the marine worm
Saccoglossus kowalevskii, whereas these data differ from
that of synthetic 2. Additionally, synthetic 1 was found to
deter feeding by predatory fish.
N-(Triisopropylsilyl)pyrrole (4). n-Butyllithium in hex-
ane (19.3 mL of a 1.6 M solution, 30.8 mmol) was added
dropwise to argon-dried pyrrole (3) (1.94 mL, 28.0 mmol) in
distilled THF at -78 °C. Triisopropylsilyl chloride (6.00 mL,
28.0 mmol) was added after 10 min and the reaction warmed
to room temperature. The solvent was then removed, water
was added, and the resulting residue was extracted with
diethyl ether. The organic phase was then dried over anhy-
drous magnesium sulfate and concentrated by rotary evapora-
tion. N-(Triisopropylsilyl)pyrrole (4) was isolated as a colorless
oil (6.26 g, 28.0 mmol), in 100% yield: 1H NMR (CDCl3, 400
MHz) δ 6.79 (2H, t, J ) 1.8), 6.31 (2H, t, J ) 2.0), 1.44 (3H,
septet, J ) 7.6), 1.09 (18H, d, J ) 7.2); 13C NMR (100 MHz) δ
124.0, 110.0, 17.8, 11.7; FABMS m/z 224.2 (15), 223.2 (20),
180.2 (11), 153.2 (25), 154.2 (100), 137 (29), 136 (57), 138 (65),
106.6 (19); HRFABMS [M]+ m/z 223.1765 (calc for C13H25NSi
223.1756).
2,3,4-Tribromo-1-(triisopropylsilyl)pyrrole (5). N-(Tri-
isopropylsilyl)pyrrole (4) (2.00 g, 9.00 mmol) was dissolved in
distilled THF (5.0 mL) and cooled to -78 °C. NBS (4.78 g, 26.9
mmol) in THF (35 mL) was then added over 10 min, and the
reaction was warmed to room temperature overnight under
argon. Cold hexane was then added to the reaction mixture
to precipitate unreacted NBS and succinimide byproduct, and
the slurry was filtered through neutral alumina. The partially
brominated intermediate was dissolved in distilled THF (5.0
mL) and cooled to -78 °C. NBS (2.40 g, 13.5 mmol) in THF
(20 mL) was added as before and the product recovered as
above. Solvent was removed by rotary evaporation, and the
product was filtered through silica gel and recrystallized in
pentane at -78 °C to produce 5 as a pale yellow solid (4.00 g,
8.69 mmol) in 97% yield: mp 50 °C; 1H NMR (CDCl3, 400 MHz)
δ 6.85 (1H, s), 1.66 (3H, septet, J ) 7.6), 1.12 (18H, d, J )
7.6); 13C NMR (100 MHz) δ 125.7, 105.6, 104.6, 100.8, 34.1,
18.0; EIMS m/z 464.9 (32), 462.9 (88), 460.9 (87), 458.9 (30),
419.9 (9), 417.9 (24), 415.9 (25), 413.9 (9), 382 (26), 380 (47),
378 (25), 204.9 (9), 202.9 (22), 200.9 (9), 157.1 (95), 138.0 (39),
136.0 (38), 115.1 (100), 87.1 (52), 73.0 (47), 59.0 (58); HREIMS
[M]+ m/z 458.9055 (calc for C13H22Br3NSi 458.9051).
2,3,4-Tribromopyrrole (1). 2,3,4-Tribromo-1-(isopropyl-
silyl)pyrrole (5) (100 mg, 0.217 mmol) was dissolved in THF
(3.0 mL) at room temperature. TBAF (0.543 mL, 0.543 mmol)
as a solution in THF was slowly added, and the mixture stirred
for approximately 1 h. The mixture was then washed with
water (3×), sodium bicarbonate (3×), and brine (3×). The
organic materials were concentrated in an ice bath with a
stream of nitrogen gas to yield 1 as a pale yellow oil (59.3 mg;
0.195 mmol) in 90% yield: 1H NMR (CDCl3, 300 MHz) δ 8.81
(1H, br s), 6.81 (1H, d, J ) 3.2); 1H NMR (acetone-d6, 400 MHz)
δ 11.23 (1H, br s), 7.11 (1H, d, J ) 3.2); 1H NMR (diethyl ether-
d10, 500 MHz) δ 11.05 (1H, br s), 6.87 (1H, s); 13C NMR (CDCl3,
75 MHz) δ 119.5 (CH), 101.9 (C), 100.2 (C), 99.8 (C); 13C NMR
(acetone-d6, 100 MHz) δ 121.9 (CH), 101.7 (C), 101.2 (C), 99.5
(C); EIMS m/z 306.8 (34), 304.8 (98), 302.8 (99), 300.8 (37),
225.9 (14), 223.9 (28), 221.9 (15), 198.9 (9), 196.9 (20), 194.9
(10); HREIMS [M]+ m/z 300.7749 (calc for C4H4Br3N 300.7737).
2,3,5-Tribromopyrrole (2). NBS (1.92 g, 10.8 mmol) in
THF (10 mL) was added over a period of 10 min to a solution
of pyrrole (3) (242 mg, 3.60 mmol) stirring in THF (10 mL)
under nitrogen, at -78 °C. After 10 more minutes, the reaction
Experimental Section
General Experimental Procedures. Melting points were
measured with a Mettler Toledo FP62 melting point ap-
paratus. NMR spectra were acquired using a Bruker Avance
DRX 500, Bruker AMX 400, or Varian Mercury Vx 300
spectrometer, in CDCl3, deteurioacetone, or deuteriodiethyl
ether, and referenced to the residual light solvent. Low- and
high-resolution mass spectra were acquired on a Micromass
70SE instrument with FAB and EI ionization. LC-MS data
were generated using a HP Series 1100 system with electro-