Structure, synthesis, and biological properties of kalkitoxin, a novel neurotoxin from the marine cyanobacterium Lyngbya majuscula [24]

Full text of DOI:10.1021/ja005526y

J. Am. Chem. Soc. 2000, 122, 12041-12042
12041
Structure, Synthesis, and Biological Properties of
Kalkitoxin, a Novel Neurotoxin from the Marine
Cyanobacterium Lyngbya majuscula
Min Wu,† Tatsufumi Okino,† Lisa M. Nogle,†
Brian L. Marquez,† R. Thomas Williamson,†
Namthip Sitachitta,† Frederick W. Berman,‡
Thomas F. Murray,‡ Kevin McGough,§ Robert Jacobs,§
Kimberly Colsen, Toshinobu Asano,‡ Fumiaki Yokokawa,‡
Takayuki Shioiri,* and William H. Gerwick*,†
Figure 1. Representation of rotamers about C7, C8, C9, and C10 with
depiction of all heteronuclear and homonuclear couplings that were used
to define the relative stereochemistry at C7, C8, and C10 using the J-based
configuration approach.8
College of Pharmacy, Oregon State UniVersity
CorVallis, Oregon 97331
partial structures for 1 (Supporting Information). One partial
structure was composed of a sec-butyl group in which the methine
component was deshielded to a chemical shift (δ 2.28) consistent
with its being adjacent to a carbonyl. A second partial structure
was composed of a methylated tertiary amide group which existed
in two conformations (Supporting Information). A third partial
structure possessed a deshielded methylene (δ 3.35) that could
be sequentially connected by E.COSY to a second methylene
group, and by HSQMBC to a methine and high-field methyl
group. By E.COSY, an additional high-field methylene group (δ
1.10, 1.02) was adjacent to a methine which also bore a methyl
group. The fifth partial structure was composed of a similar
-CH2-CH-CH3 grouping; however, in this case, the methylene
group protons were deshielded to δ2 .31 and δ 2.55. The final
partial structure, based on E.COSY correlations, HSQMBC, and
chemical shift models,7 was composed of a thiazoline ring with
an ethylene appendage; this was further substantiated by EIMS
fragmentations (Supporting Information). HSQMBC data were
used to connect these partial structures and gave the full planar
structural assignment of 1.
Department of Physiology and Pharmacology
College of Veterinary Medicine
The UniVersity of Georgia, Room 2223
Athens, Georgia 30601
Department of Pharmacology
UniVersity of California at Santa Barbara
California 93106
Bruker Instruments, Inc., 19 Fortune DriVe
Manning Park, Billerica, Massachusetts 01821
Faculty of Pharmaceutical Sciences
Nagoya City UniVersity, Tanabe-dori, Mizuho-ku
Nagoya 467-8603, Japan.
ReceiVed August 15, 2000
Cyanobacteria from a diversity of marine and freshwater
habitats are known to produce neurotoxic secondary metabolites.1
Herein, we describe the complete stereostructure, synthesis, and
biological properties of kalkitoxin (1), a novel neurotoxic li-
popeptide from a Caribbean collection of Lyngbya majuscula.
The organic extract of this L. majuscula exhibited potent brine
shrimp and fish toxicity.2 Using these assays, the toxic metabolite
kalkitoxin (1), was isolated by sequential silica gel VLC, CC,
and normal-phase HPLC (12.8 mg, 0.3% of extract). Subse-
quently, bioassay-guided fractionation using a primary cell culture
of rat neurons in a microphysiometer3 or inhibition of IL-1â
stimulation of sPLA2 in hepatocarcinoma cells4 led to re-isolation
of 1 in small yield from various Caribbean collections of L.
majuscula.
The C3 stereochemistry of kalkitoxin was determined by
Marfey’s analysis. Kalkitoxin was ozonized and then hydrolyzed
in 6 N HCl to obtain cysteic acid. Marfey’s analysis of this
hydrolysate yielded L-cysteic acid, defining C3 as R. The limited
amount of kalkitoxin precluded determination of the C2′ stereo-
chemistry. The relative stereochemistry of the three chiral centers
within the aliphatic chain of kalkitoxin (C7, C8, C10) was
determined using the J-based configuration analysis method.8
3
The JCH values were measured by a modification of the
recently reported HSQMBC pulse sequence,6 and the 3JHH values
were determined utilizing the E.COSY pulse sequence.9 To
overcome the limited sample size of natural kalkitoxin (∼300
µg of 1 remained because of chemical instability) all data used
in this analysis were recorded on a Bruker 500 MHz DRX
spectrometer equipped with a Bruker 5 mm TXI CryoProbe
(benzene-d6). The relative stereochemistry at C7-C8 was sug-
3
Kalkitoxin (1) analyzed for C21H38N2OS indicated four degrees
of unsaturation; from 13C NMR analysis in DMSO-d6 two were
due to double bonds, one to a carbonyl group, and the remaining
one to a ring system.5 Data from E.COSY, HSQC, and a modified
HSQMBC6 experiments in benzene-d6 allowed deduction of six
gested by observation of a small (1.3 Hz gauche) JHH between
H7-H8, a large (6.1 Hz, anti) 3JCH between H8-C6, and a small
3
(<1 Hz, gauche) JCH between H7-C9 (see Figure 1). Stereo-
chemistry at positions C8 and C10 were related through the
intervening C9 diastereotopic methylene protons. The low-field
3
proton at C9 (H9a) showed a large (8.2 Hz, anti) JHH to H8,
† Oregon State University.
‡ The University of Georgia.
§ University of California at Santa Barbara.
(6) (a)Williamson, R. T.; Ma´rquez, B. L.; Gerwick, W. H.; Ko¨ve´r, K. E.
Magn. Reson. Chem. 2000, 38, 265-273. (b) By incorporation of a G-BIRDR
sequence during the INEPT transfer of the HSQMBC pulse sequence, an
Bruker Instruments, Inc., Nagoya City University.
(1) (a) Carmichael, W. W. Sci. Am. 1994, 270, 78-86. (b) Orjala, J.; Nagle,
D. G.; Hsu V.; Gerwick, W. H. J. Am. Chem. Soc. 1995, 117, 8281-8282.
(c) Berman, F. W.; Gerwick, W. H.; Murray, T. F. Toxicon 1999, 37, 1645-
1648.
1
effective JCH bond suppression and decoupling of remote protons is
achieved: Williamson, R. T.; Marquez, B. L.; Gerwick, W. H. (manuscript
in preparation).
(2) This brownish-colored L. majuscula was collected from 6 to 8 m water
depth (1 L compressed volume), 11 August, 1994, Playa Kalki, Curac¸ao
(Collection Code NAK-11Aug94-03); preserved at reduced temperature in
isopropyl alcohol until extraction; 4.32 g organic extract.
(3) Hirst, M. A.; Pichford, S. J. NIH Res. 1993, 5, 69.
(4) Tan, L. T.; Williamson, R. T.; Watts, K. S.; Gerwick, W. H.; McGough,
K.; Jacobs, R. J. Org. Chem. 2000, 65, 419-425.
(7) (a) Hawkins, C. J.; Levin, M. F.; Marshall, K. A.; van den Brenk, A.
L.; Watters, D. J. J. Med. Chem. 1990, 33, 1634-1638.; (b) Jalal, M. A. F.;
Hossain, M. B.; van der Helm, D.; Sanders-Loehr, J.; Actis, L. A.; Crosa, J.
H. J. Am. Chem. Soc. 1989, 111, 292-296.
(8) Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana,
K. J. Org. Chem. 1999, 64, 866-876.
(9) Griesinger, C.; Sørenson, O. W.; Ernst, R. R. J. Am. Chem. Soc. 1985,
107, 6394-6396.
(5) Wu, M. M.S. Thesis, Oregon State University, 1997.
10.1021/ja005526y CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/15/2000

12042 J. Am. Chem. Soc., Vol. 122, No. 48, 2000
Communications to the Editor
Scheme 1. Synthesis of (3R,7R,8S,10S,2′R)-Kalkitoxin (1)a
a (a) n-BuLi, THF; pTsCl. (b) LiAlH4, THF reflux. (c) 9-BBN,
ultrasound, THF; aq NaOH, H2O2. (d) MsCl, Et3N, DMAP, CH2Cl2. (e)
NaN3, DMF, 55 °C. (f) H2 (1 atm) 5% Pd-C, EtOAc. (g) (R)-2-
methylbutyric acid, DEPC, Et3N, DMF. (h) n-BuLi, THF; MeI. (i) TBAF,
THF. (j) Py‚SO3, Et3N, DMSO, CH2Cl2. (k) Phosphonate 5, NaHMDS,
THF. (l) MeMgBr, CuBr-DMS, THF/DMS, -30 °C. (m) aq LiOH, H2O2,
THF. (n) amino alcohol 7, EDCl‚HCl, iPr2NEt, DMAP, CH2Cl2. (o)
DAST, CH2Cl2, -20 °C. (p) H2S, Et3N, MeOH. (q) DAST, CH2Cl2, -20
°C.
Figure 2. Differences in 13C NMR shifts between natural kalkitoxin (1)
and four synthetic kalkitoxin stereoisomers.
cyclodehydration with DAST10k provided (3R,7R,8S,10S,2′R)-
kalkitoxin (1) in 48% overall from the amide 8 (Scheme 1).11
Comparison of 13C NMR chemical shifts between five syn-
thesized diastereoisomers and natural kalkitoxin showed very
small differences of less than 0.2 ppm (Figure 2). However, both
the 3S,7S,8R,10R,2′S- and 3R,7R,8S,10S,2′R-isomers showed
maximal 13C NMR differences of 0.026 ppm. The CD spectrum
of the 3S,7S,8R,10R,2′S-isomer was of equal intensity but opposite
sign to natural kalkitoxin. Correspondingly, the CD of the
3R,7R,8S,10S,2′R-isomer was essentially identical to natural
compound 1 (Supporting Information).
Natural (+)-kalkitoxin (1) was strongly ichthyotoxic to the
common goldfish (Carassius auratus, LC50 700 nM), potently
brine shrimp toxic (Artemia salina, LC50 170 nM), and potently
inhibited cell division in a fertilized sea urchin embryo assay (IC50
∼25 nM).12a Synthetic (+)-kalkitoxin (3R,7R,8S,10S,2′R) was
equally potent in the brine shrimp assay (LC50 170 nM).
Interestingly, synthetic (-)-kalkitoxin (3S,7S,8R,10R,2′S) was
relatively inactive as a brine shrimp toxin (LC50 9300 nM). In a
primary cell culture of rat neurons, natural kalkitoxin displayed
an exceptional level of neurotoxicity (LC50 3.86 nM), and its
effects were inhibitable with NMDA receptor antagonists.1c
Additionally, natural kalkitoxin is highly active in an inflammatory
disease model which measures IL-1â-induced sPLA2 secretion
from HepG2 cells (IC50 27 nM).4 Finally, preliminary evidence
suggests that kalkitoxin is an exquisitely potent blocker of the
voltage sensitive Na+ channel in mouse neuro-2a cells (EC50 of
1 ) 1 nM; EC50 of saxitoxin ) 8 nM).12b Consistent with many
known cyanobacterial metabolites, kalkitoxin appears to derive
from a mixed polyketide/nonribosomyl peptide synthetase path-
way.13
whereas the high-field proton (H9b) showed a small (4.4 Hz,
3
3
gauche) JHH to H8. Additionally, small (<1 Hz) JCH were
observed from H9a and H9b to C7, and a large (8.2 Hz, anti) 3JCH
from H9b to C14. The relative stereochemistry at C9 and C10
3
was determined by a large (9.4 Hz, anti) JHH between H9b and
3
H10 and a small (3.1 Hz, gauche) JHH for H9a-H10. Finally, a
large (7.4 Hz, anti) 3JCH was measured for H9a-C15. In summary,
these data strongly supported a 7R*, 8S*, 10S* relative stereo-
chemistry for 1. In combination with the above-determined 3R
absolute stereochemistry, the total number of stereochemical
possibilities was reduced to four.
To determine the absolute stereochemistry of natural kalkitoxin,
kalkitoxins having all possible configurations were synthesized;
(3R,7R,8S,10S,2′R)-kalkitoxin was found to be identical with the
natural substance 1. The chemical synthesis of natural 1 com-
menced with the known alcohol 2.10a Deoxygenation of the
corresponding tosylate, hydroboration, and conversion of the
resulting alcohol into the azide gave 3 in 39% overall yield from
2.10b-d Reduction of the azide, coupling with (R)-2-methylbutyric
acid using DEPC,10e and N-methylation gave N-methylamide 4
in 63% yield. O-Desilylation and oxidation was followed by a
Horner-Emmons reaction and simultaneously homologated and
introduced the (R)-phenylglycine-derived auxiliary required for
an asymmetric conjugate addition.10f Introduction of the C7 methyl
group by the method of Hruby10g proceeded smoothly to give a
methyl adduct 6 as a single isomer in 79% overall yield from 4.
Removal of the chiral auxiliary and coupling with (R)-amino
alcohol 710h yielded 8 in 94% yield. Following Wipf’s oxazoline-
thiazoline interconversion protocol,10i cyclodehydration of the
amide with DAST to the oxazoline10j and then treatment with
hydrogen sulfide afforded the thioamide. Finally, a second
Supporting Information Available: Experimental details and NMR
and CD spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA005526Y
(10) (a) Roush, W. R.; Palkowitz, A. D.; Ando, K. J. Am. Chem. Soc. 1990,
112, 6348-6359. (b) Brown, H. C.; Racherla, U. S. Tetrahedron Lett. 1985,
26, 2187-2190. (c) Crimmins, M. T.; O′Mahony, R. Tetrahedron Lett. 1989,
30, 5993-5996. (d) We could also obtain the azide 3 from the alcohol by our
one-pot procedure in 74% yield (see Mizuno, M.; Shioiri, T. Chem. Commun.
1997, 2165-2166). (e) Takuma, S.; Hamada, Y.; Shioiri, T. Chem. Pharm.
Bull. 1982, 30, 3147-3153 and references therein. (f) Romo, D.; Rzasa, R.
M.; Shea, H. A.; Park, K.; Langenham, J. M.; Sun, L.; Akhiezer, A.; Liu, J.
O. J. Am. Chem. Soc. 1998, 120, 12237-12254. (g) Li, G.; Patel, D.; Hruby,
V. J. Tetrahedron: Asymmetry 1993, 4, 2315-2318. (h) Ohfune, Y.;
Kurokawa, N. Tetrahedron Lett. 1984, 25, 1071-1074. (i) Wipf, P.; Miller,
C. P.; Venkatraman, S.; Fritch, P. C. Tetrahedron Lett. 1995, 36, 6395-6398.
(j) Lafargue, P.; Guenot, P.; Lellouche, J.-P. Heterocycles 1995, 41, 947-
958. (k) Lafargue, P.; Guenot, P.; Lellouche, J.-P. Synlett. 1995, 171-172.
(11) Information on the syntheses of kalkitoxin and its stereoisomers is
found in part in the Supporting Information and will be presented in detail
elsewhere.
(12) (a) Pruzanski, W.; Kennedy, B. P.; van den Bosch, H.; Stefanski, E.;
Vadas, P. Lab. InVest. 1997, 76, 171-178. (b) Manger, R. L.; Leja, L. S.;
Lee, S. Y.; Hungerford, J. M.; Hokama, Y.; Dickey, R. W.; Granade, H. R.;
Lewis, R.; Yasumoto, T.; Wekell, M. M. J. AOAC Intern. 1995, 78, 521-
527.
(13) Work in Oregon was supported by NIH CA 52955 and the MFBS
Center at OSU (ES03850), a JSPS fellowship to T.O.; work in Nagoya was
partially supported by Grants in Aid from Nagoya City University (to F.Y.)
and the Ministry of Education, Science, Sports, and Culture, Japan. We
gratefully acknowledge the assistance and permission of the CARMABI
Research Station, Curac¸ao in making collections of L. majuscula.