Total synthesis and absolute structure of manzacidin A and C [3]

Full text of DOI:10.1021/ja002556s

10708
J. Am. Chem. Soc. 2000, 122, 10708-10709
Scheme 1
Total Synthesis and Absolute Structure of
Manzacidin A and C
Kosuke Namba, Tetsuro Shinada, Toshiyuki Teramoto, and
Yasufumi Ohfune*
Department of Material Science
Graduate School of Science, Osaka City UniVersity
Sugimoto, Osaka 558-8585, Japan
ReceiVed July 13, 2000
A number of bromopyrrole alkaloids have been found from
marine sponges which are known to exhibit pharmacologically
useful activities such as R-adrenoceptor blockers, antagonists of
serotonergic receptor, actomyosin ATPase activators, etc.1 Re-
cently, Kobayashi et al. have isolated a novel class of the
alkaloids, manzacidin A-C (1a-1c),2 from an Okinawan sponge
Hymeniacidon sp.3 which possess a unique structure consisting
of an ester-linked bromopyrrolecarboxylic acid and a 3,4,5,6-
tetrahydropyrimidine ring in which one of the amino groups is
attached to the C4 quarternary carbon center.4 Although manza-
cidins exhibit similar biological activities to those of other
bromopyrrole alkaloids, only preliminary tests have been carried
out, owing to the extremely small amount of samples available
from marine sources.2 In the following, we describe highly
stereoselective synthesis of both natural manzacidin A (1a) and
C (1c), which unambiguously assigned their absolute structures
to be (4S,6R)-1a and (4S,6S)-1c.
a (a) (1) Et3N, Boc-L-Phe-OSu, THF-MeOH (7:1), 0 °C, 2 h; (2)
p-TsOH, 2,2-dimethoxypropane, toluene, reflux, 2 h; (3) 0.1 equiv PdCl2,
1 equiv CuCl, O2, DMF-H2O (7:1), rt, 12 h (66% from 5). (b) TMSOTf,
2,6-lutidine, CH2Cl2, rt, 30 min. (c) TMSCN, ZnCl2, 2-PrOH, rt, 18 h
(81% from 3a, 87% from 3b, and 48% from 3c).
selectivity, while in our previous work the nitrile addition to a
six-membered ketimine intermediate occurred from the sterically
less demanding R-face (eq 1).6
To examine the stereochemical outcome of the nitrile addition
to the imine A, we prepared three types of Strecker precursors
having Boc-L-phenylalanyl amide 4a, its D-Phe isomer 4b, and
Boc-glycyl amide 4c, respectively, from (2S)-allylglycinol 5
readily available from L-allylglycine or Boc-L-aspartate (Scheme
1).7 After chemoselective removal of the Boc group of 4a with
TMSOTf/2,6-lutidine,8 the resulting amine was treated with
trimethylsilylnitrile (TMSCN) to produce in 81% yield the amino
nitrile (4R)-3a as a single diastereomer.9 The D-Phe isomer 4b
afforded (4S)-3b (87%), exclusively. These results clearly indi-
cated that in each case the nitrile addition to the imine occurred
from the opposite side of the C6 benzyl group, for example, B,10
similar to the six-membered case (vide supra). On the other hand,
the glycyl amide 4c gave a 1:5 mixture of (4R)-3c and (4S)-3c,
indicating that the cyanide preferentially attacks from the sterically
less hindered â-face when the C6 substituent is absent. Thus, not
only was the stereochemical outcome of the amino nitrile
formation clearly understood, but also the desired amino nitriles
3a and 3b, which correspond to the stereochemistry of 1a and
1c, were obtained, respectively, with excellent stereocontrol.
We next examined the conversion of 3a into (2S,4R)-2a. This
process requires initial oxidation of 3a to the imino ketone 6,
The stereochemical relationship between 1a and 1c has been
proposed as either the C4 or the C6 diastereomer. We presumed
that their C4 configuration would be the same S by considering
a plausible biosynthetic pathway which involves (R)- or (S)-
isonitrile intermediate as often seen in the structure of marine
natural products.5 Thus, diastereoselective construction of (2S,4R)-
and (2S,4S)-diamines, 2a and 2b, would lead to 1a and 1c,
respectively (Scheme 1). This route relies on a stereoselective
construction of the amino nitrile 3 by the Strecker synthesis of
the amino ketone 4. However, an asymmetric version of the
Strecker synthesis using an amide is unknown, and the stereo-
chemistry of a cyanide addition to the imine A is unpredictable
in this case because its C2 chiral center also affects the facial
* To whom correspondence should be addressed. Telephone: +81-6-6605-
2570. Fax: +81-6-6605-3153. E-mail: ohfune@sci.osaka-cu.ac.jp.
(1) Faulkner, D. J. Nat. Prod. Rep. 1998, 15, 113-158 and references
therein.
(2) Kobayashi, J.; Kanda, F.; Ishibashi, M.; Shigemori, H. J. Org. Chem.
1991, 56, 4574-4576.
(3) Several bromopyrrole alkaloids have been isolated from Himeniacidon
Sponge: Kobayashi, J.; Inaba, K.; Tsuda, M. Tetrahedron 1997, 46, 16679-
16682 and references therein.
(4) Recently, manzacidin D structurally related to 1 has been isolated from
“Living Fossil” sponge: Jahn, T.; Ko¨nig, G. M.; Wright, A. D.; Wo¨rheide,
G.; Reitner, J. Tetrahedron Lett. 1997, 38, 3883-3884.
(5) Isonitrile-containing marine natural products, see: (a) Wright, A. D.;
Ko¨nig, G. M.; Angerhofer, C. K.; Greenidge, P.; Linden, A.; Desqueyroux-
Faundez, R. J. Nat. Prod. 1996, 59, 710-716. (b) Kobayashi, J.; Tsuda, M.;
Nemoto, A.; Tanaka, Y.; Yazawa, K.; Mikami, Y. J. Nat. Prod. 1997, 60,
719-720.
(6) (a) Moon, S.-H.; Ohfune, Y. J. Am. Chem. Soc. 1994, 116, 7405-
7406. (b) Ohfune, Y.; Moon, S.-H.; Horikawa, M. Pure Appl. Chem. 1996,
68, 645-648. (c) Horikawa, M.; Nakajima, T.; Ohfune, Y. Synlett 1997, 253-
254. (d) Ohfune, Y.; Nanba, K.; Takada, I.; Kan, T.; Horikawa, M.; Nakajima.
T. Chirality 1997, 9, 459-462. (e) Ohfune, Y.; Horikawa, M. J. Synth. Org.
Chem. Jpn. 1997, 55, 982-993.
(7) Ouerfelli, O.; Ishida, M.; Shinozaki, H.; Nakanishi, K.; Ohfune, Y.
Synlett 1993, 409-410.
(8) (a) Sakaitani, M.; Ohfune, Y. J. Org. Chem. 1990, 55, 870-876. (b)
Shimamoto, K.; Ohfune, Y. J. Med. Chem. 1996, 39, 407-423.
(9) The structure of 3a was determined by comparison of its 1H NMR data
with those of the iso-propyl derivative 3d whose structure has been confirmed
by X-ray crystallographic analysis, see: Supporting Information.
(10) Although it is difficult to elucidate the conformation of the seven-
membered ring ketimine intermediate B, this was proposed as depicted in
Scheme 1 by means of the 1H NMR data of 3a and the X-ray structure of 3d.
Boatlike conformation of a six-membered ring ketimine has been reported,
see: Schulz, G.; Steglich, W. Chem. Ber. 1977, 110, 3615-3623. The high
stereoselectivity in the addition of a cyanide ion would be due to steric as
well as stereoelectronic reasons. It is noted that the nitrile group of the addition
product 3 locates in axial which has an antiperiplanar relationship with the
neighboring amino lone pair providing stereoelectronic stabilization to the
axial orientation of the nitrile group.
10.1021/ja002556s CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/13/2000

Communications to the Editor
J. Am. Chem. Soc., Vol. 122, No. 43, 2000 10709
Scheme 2
Scheme 3
a (a) O3, MeOH, -78 °C, 20 min (94%). (b) concd HCl, 0 °C, 8 h, rt,
15 h, and 100 °C, 18 h. (c) (1) Boc2O, NaHCO3, dioxane-H2O (1:1), rt,
12 h; (2) Me4NOH-5H2O, Boc2O, CH3CN, rt, 15 h; (3) CH2N2, Et2O
(70% from 6). (d) LiAlH4, Et2O, 0 °C, 2 h (81%). (e) PDC, CH2Cl2, rt,
15 h (55%). (f) TFA, CH2Cl2, rt, 30 min, then CH(OMe)3, catalytic
amount of concd HCl, reflux, 4 h (91%). (g) 2 equiv NaH, DMF, rt, 4 h
(100%).
a (a) 0.1 equiv MeReO3, urea-H2O2, 40 °C, 15 min (87%). (b) concd
HCl, 85 °C, 15 h (76%). (c) (1) 10% H2SO4, 120 °C, 40 h; (2) Boc2O,
3 equiv NaHCO3, dioxane/H2O (1:1), rt, 14 h; (3) Me4NOH-5H2O,
Boc2O, CH3CN, rt, 14 h; (4) CH2N2, AcOEt (79% from 13: 14a/14b/15
) 6:1:1). (d) Boc2O, Et3N, DMAP, THF (80% from 14a and 63% from
14b). (e) (1) 2.5 equiv 0.3 N NaOH, MeOH, rt, 24 h; (2) CH2N2, Et2O
(78% from 14c). (f) LiAlH4, Et2O, 0 °C, 2 h (61%). (g) PDC, CH2Cl2, rt,
15 h (62%). (h) TFA, CH2Cl2, rt, 30 min, then CH(OMe)3, catalytic
amount of concd HCl, reflux, 3 h (79%). (i) 11, 2 equiv NaH, DMF, rt,
4 h (68%).
where t-BuOCl/Et3N is often employed,6 and subsequent hydro-
lytic removal of the phenylalanyl group. However, not only this
reagent system but also other combinations of oxidants and bases
were unsuccessful,11 resulting in almost all cases in recovery of
the starting material due probably to steric reasons as well as the
low acidity of the C6 hydrogen. Fortunately, we found that the
oxidation with ozone was quite effective for this conversion to
give in 94% yield the desired imino ketone 6. We presumed that
the reaction proceeded via C since ozone is a small molecule
compared to other oxidants and has a bidentate chemical
characteristic acting as an internal base after oxidation of the
nitrogen (Scheme 2). On treatment with concentrated HCl, 6
underwent removal of the phenylpyruvic acid and simultaneous
hydrolysis of the nitrile group to afford diaminocarboxylic acid
7a, which without purification, was converted into protected ester
7b. This was reduced to diol 8a,12 whose oxidation with PDC
occurred from the sterically less hindered hydroxymethyl group
to give 9. Construction of the tetrahydropyrimidine ring was
performed by successive treatments with (1) TFA and (2) methyl
orthoformate to give 10 (91%). Completion of the synthesis now
required esterification of the bromopyrrolecarboxylate with 10.
This was accomplished by means of trichloroacetylbromopyrrole
was identical in all respects (1H and 13C NMR, HRMS, and HPLC
profile) including the sign of specific rotation with natural 1a
(lit.2 [R]27 -28 (c 0.67, MeOH)). Thus, the absolute structure
D
of natural (-)-1a was unambiguously assigned to (4S,6R)-1a.
The successful synthesis of 1a led us to examine the synthesis
of 1c from the (4S)-amino nitrile 3b. We again faced difficulty
upon oxidative removal of the D-Phe moiety of 3b which did not
produce desired imine (4S)-6 even by ozone. With an expectation
that 12 would give (4S)-6, 3b was converted into hydroxylamine
12 by trioxorhenium oxidation.15 Teatment of 12 with a base did
not give (4S)-6, but with concentrated HCl it produced unexpect-
edly cyclic urea 13 (76%). D-Phe was a byproduct (13/D-Phe )
5:1).16 We propose that the carbon-carbon fragmentation between
C6 and C7 of the protonated hydroxylamine D occurred to give
E which re-cyclized to 13 (Scheme 3). Treatment of the urea 13
with 10% H2SO4 gave a mixture of hydrolyzed products which,
upon protection, afforded a mixture of 14a (59%), 14b (11%),
and 15 (7%). Both the pyrrolidines 14a and 14b were converted
into 15 in three steps, respectively. The ester 15 was reduced to
diol 16a whose conversion into 1c was successfully performed
in the same manner as that of 1a from 9a. Synthetic 1c showed
completely identical spectral data with those of reported.2 The
1113 to give 1a ([R]27 -22.4 (c 0.52, MeOH)).14 Synthetic 1a
D
(11) Treatment of 3a with t-BuOCl gave the corresponding N-Cl deriva-
tive. However, subsequent dehydrochlorination did not proceed at all using
the following bases: DBU, DMAP, (i-Pr)2NEt, DABCO, KHMDS, NaH,
t-BuOK, and AgBF4. The combinations of other oxidants and bases were not
successful: PhIO, CuBr2/t-BuOLi, PhSeBr/H2O2, and Swern oxidation.
(12) Optical purity of the diols 8a and 16a was ascertained to be >95% ee
by converting them into the corresponding bis-(+)- or (-)-MTPA ester 8b
and 16b, respectively, see: Supporting Information.
sign of specific rotation of synthetic 1c ([R]28 +89.0 (c 0.73,
D
MeOH))17 was also the same plus as that reported for (+)-1c (lit.2
[R]27D +37 (c 0.23, MeOH)). Thus, both natural (-)-1a and (+)-
1c were found to possess the same 4S configuration.
(13) Synthesis of 2-trichloroacetylpyrrole, see: (a) Bailey, D. M.; Johnson,
R. E.; Albertson, N. F. Org. Synth. 1971, 51, 100-102. (b) Bailey, D. M.;
Johnson, R. E. J. Med. Chem. 1973, 16, 1300-1302.
In summary, we have synthesized both natural (-)-1a and (+)-
1c via a Strecker synthesis of amino ketones 4a and 4b,
respectively. Not only was sufficient synthetic material (14 steps,
14% overall for 1a, and 15 steps, 3.5% for 1c) obtained to permit
further pharmacological studies on these alkaloids, but also the
present route enables the synthesis of the enantiomers or diaster-
eomers of 1 using the combination of (R)-5 with D- or L-Phe.
Furthermore, general applications of the methods employed for
the oxidation of the amino nitriles 3a and 3b are currently being
investigated.
(14) The synthetic 1a showed a smaller [R]D value than the reported one
although the optical purity (>95% ee) of synthetic 1a had been confirmed
using the diol 8a.12 The [R]D value of synthetic 1a using other solvent: [R]27
D
-19.7 (c 0.65, CH3CN/0.1 N HCl ) 1:1).
(15) Trioxorhenium oxidation of an amine, see: Goti, A.; Nannelli, L.
Tetrahedron Lett. 1996, 37, 6025-6028.
(16) We propose that D-Phe was eliminated from D via the imine i although
the amino ketone ii could not be isolated from the reaction mixture due
probably to its subsequent decomposition.
Acknowledgment. This study was supported by a grant from the
Research for the Future Program from the Japan Society for the Promotion
of Science (JSPS). K.N. is grateful to JSPS for a Research Fellowship
for Young Scientists.
(17) The [R]D value of synthetic 1c was much larger than the reported
value in this case.12,14 The [R]D value of synthetic 1c using another solvent:
[R]23D +75.8 (c 0.5, CH3CN/0.1 N HCl ) 1:1). Since only a small amount of
1c was isolated from the sponge, the discrepancy of the [R]D value between
synthetic and natural 1c would be due to contamination of an impure material
in natural 1c which was inseparable even by repeated HPLC purification. In
fact, unidentifiable impure material is observed in the 1H NMR spectra of
natural 1c, see: Supporting Information.2
Supporting Information Available: Full experimentals and charac-
terization of the compounds 1a, 1c, 2-18 (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
JA002556S