Probing the Phytopathogenic Blackleg Fungus with a Phytoalexin Homolog

Full text of DOI:10.1021/jo971702i

416
J . Org. Chem. 1998, 63, 416-417
Sch em e 1
P r obin g th e P h ytop a th ogen ic Bla ck leg
F u n gu s w ith a P h ytoa lexin Hom olog
M. Soledade C. Pedras* and Francis I. Okanga
Department of Chemistry, University of Saskatchewan, 110
Science Place, Saskatoon SK S7N 5C9, Canada
Received September 12, 1997
Plants produce a vast array of chemical defenses, some
of which are synthesized de novo in response to biotic and
abiotic stress, i.e., phytoalexins.1 The detoxification of
phytoalexins is a mechanism of virulence in diverse phyto-
pathogenic fungi.2 This mechanism allows fungi to invade
plants, which may lead to significant tissue damage and
concomitant yield losses. An environmentally attractive
strategy to control such plant pathogens could involve
application of selective inhibitors of those fungal enzymes
utilized in the detoxification of phytoalexins. In this con-
nection, we have established that the phytoalexin brassinin
(1) is metabolized3 and detoxified4 by the virulent blackleg
fungus [Phoma lingam (Tode ex Fr.) Desm., asexual stage
of Leptosphaeria maculans (Desm.) Ces. et de Not.], one of
the most important pathogens of the oilseeds rapeseed
(Brassica napus, B. rapa), canola (B. napus, B. rapa), and
mustard (B. juncea).5 In addition, the dithiocarbamate
group of brassinin (1) was thought to be essential for its
antifungal activity. Nevertheless, despite the relatively
higher antifungal activity of mono- and dichlorobenzyl
dithiocarbamates, their detoxification by virulent isolates of
the blackleg fungus was not averted.6 We have also dem-
onstrated that this pest can detoxify other brassica phytoal-
exins, including brassicanal A (2),7 brassilexin (3),8 and
cyclobrassinin (4)9 via different pathways, suggesting that
its efficacy in disarming host plants has a broad enzymatic
base.10 To further probe the selectivity of the enzymes
involved in detoxification of brassinin (1), the metabolism
and antifungal activity of phytoalexin analogues are being
investigated. Because methyl tryptaminedithiocarbamate
(5) was significantly more inhibitory to P. lingam than
brassinin, we examined its transformation by fungal cells.
Here we disclose unique chemical aspects of the metabolic
pathway leading to the fungal detoxification of 5, as well as
the unprecedented in vitro rearrangement of the dithiocar-
bamate S-oxide 10 to the 2-oxindole derivative 11.
Preliminary experiments were carried out to determine
the time required for complete metabolism of dithiocarbam-
ate 5 by virulent isolates of P. lingam. Cultures were
incubated with 5,11 samples were withdrawn at 0-24 h
intervals, the mycelia were filtered off, and the broth was
extracted with Et2O. Extracts were analyzed by TLC and
HPLC to determine the optimum incubation time for isola-
tion of possible metabolic intermediates. Subsequently,
fractionation of Et2O extracts obtained from larger scale
cultures afforded compounds 6-12 (Scheme 1).12 Com-
pounds 8 and 9 were the major metabolites obtained from
fungal transformation of 5, representing ca. 65% of the total
amount of isolated metabolites, followed by acetyltryptamine
7 (15%) and minor components 6 and 10-12 (less than 5%).
The structures of compounds 6-8 were readily determined
by comparison of their spectroscopic data with those of
authentic samples.13 The structures of all other metabolites
were deduced from analyses of their spectroscopic data as
discussed below and corroborated by synthesis.
* To whom correspondence should be addressed. Tel: (306) 966-4772.
Fax: (306) 966-4730. E-mail: pedras@sask.usask.ca.
(1) (a) Brooks, C. J .; Watson, D. G. Nat. Prod. Rep. 1985, 427-59. (b)
Brooks, C. J .; Watson, D. G. Nat. Prod. Rep. 1991, 367-389.
(2) (a) Daniel, M., Purkayastha, R. P., Eds. Handbook of Phytoalexin
Metabolism and Action; Marcel Dekker: Inc.: New York, 1995; p 650. (b)
Van Etten, H. D.; Mathews, D. E.; Mathews, P. S. Annu. Rev. Phytopathol.
1989, 27, 143-164.
(3) First isolation of brassinin (1): Takasugi, M.; Katsui, N.; Shirata,
A.; J . Chem. Soc., Chem. Commun. 1986, 1077-1078. Metabolism of
brassinin: Pedras, M. S. C.; Taylor, J . L. J . Org. Chem. 1991, 56, 2619-
2621.
The HRMS analyses of 9-11 indicated for each compound
a molecular formula of C12H14N2OS2, which formally repre-
sented addition of one oxygen unit to dithiocarbamate 5. The
1H NMR spectra and optical rotations of both 9 and 11
suggested that they were structural isomers (four magneti-
(4) (a) Pedras, M. S. C.; Borgmann, I.; Taylor, J . L. Phytochem. (Life Sci.
Adv.) 1992, 11, 1-7. (b) Pedras, M. S. C.; Taylor, J . L. J . Nat. Prod. 1993,
56, 731-738.
1
cally nonequivalent methylene protons) and chiral. The H
NMR spectrum of 10 displayed resonances comparable to
those of 5 (within 0.1 ppm), except for the SCH3 (δ 2.35 for
10 vs δ 2.56 for 5) and CH2N (δ 3.80 for 10 vs δ 4.06 for 5)
groups. Furthermore, the 13C NMR spectrum of 10 was
similar to that of 5 (within 2 ppm) except for resonances
(5) For a review on the blackleg fungus see, for example: Pedras, M. S.
C.; Se´guin-Swartz, G. Can. J . Plant Pathol. 1992, 14, 67-75.
(6) Pedras, M. S. C.; Khan, A. K.; Smith, K. C.; Stettner, S. L. Can. J .
Chem. 1997, 75, 825-828.
(7) First isolation of brassicanal A (2): Monde, K.; Katsui, N.; Shirata,
A. Chemistry Lett. 1990, 209-210. Metabolism of brassicanal A: Pedras,
M. S. C.; Khan, A. Q. J . Agric. Food Chem. 1996, 44, 3403-3407.
(8) First isolation of brassilexin (3): Devys, M.; Barbier, M.; Loiselet, I.;
Rouxel, T.; Sarniguet, A.; Kollmann, A.; Bousquet, J . Tetrahedron Lett. 1988,
29, 6447-6448. Metabolism of brassilexin: Pedras, M. S. C.; Khan, A. Q.;
Taylor, J . L. In Phytochemicals for Pest Control; Hedin, P. A., Hollingworth,
R. M., Masler, E. P., Miyamoto, J ., Thompson, D. G., Eds.; ACS Symposium
Series 658; American Chemical Society: Washington, DC, 1997; pp 155-
166.
(11) Dithiocarbamate 5 (higher antifungal activity than brassinin (1))
was administered to cultures of virulent isolates of P. lingam at lower
concentration (final concentrations 1.6 × 10-4 M and 2.0 × 10-4 M) than
brassinin (final concentration 5.0 × 10-4 M).
(12) Cultures were incubated with 5 up to 3 weeks; 6, 7, and 9-11 were
detected in cultures after 24 h of incubation, whereas 8 and 12 were detected
after 5 days of incubation.
(9) First isolation of cyclobrassinin (4) see ref 3. Metabolism of cyclo-
brassinin: Pedras, M. S. C.; Okanga, F. I. Chem. Commun., in press.
(10) For a recent review on cruciferous phytoalexins see ref 8.
(13) Authentic samples of 6 and 8 were purchased from Aldrich Chemical
Co.; 7 was obtained upon standard acetylation of 6 (Ac2O/Py).
S0022-3263(97)01702-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/14/1998

Communications
J . Org. Chem., Vol. 63, No. 3, 1998 417
Sch em e 2
S-oxide 13 is attributed to the additional CH2 group, which
reduces the propensity for elimination of dithiocarbamic acid
or equivalent (cf. 14). The photosensitized cyclization of
dithiocarbamate 5 to pyrroloindole 9 was achieved in 70%
yield, following a procedure previously reported20 for the
corresponding methyl carbamate. A similar oxygenation of
Nb-acetyltryptamine yielded the corresponding pyrroloindole
product in lower yield.21 Although these photooxygenation
reactions emphasized the likelihood of having similar pro-
cesses occurring nonenzymatically, incubation of 5 or 7
under conditions identical with those used for fungal cul-
tures but in the absence of the fungus, followed by HPLC
analysis of sample extracts indicated that both 5 and 7 were
stable for at least two weeks. Finally, in vitro oxidation of
acid 8 with DMSO/HCl yielded acid 12.18a Although a
similar reaction could explain the presence of 12 in fungal
cultures incubated with 5, control experiments indicated
that such a process did not occur under the experimental
conditions utilized for fungal metabolism of 5.
attributable to thiocarbonyl (δC 194.5 in 10 vs δC 198.8 in
5) and SCH3 (δC 13.3 in 10 vs δC 18.1 in 5) groups. These
results and the FTIR spectrum of 10 (905 cm-1) indicated
the presence of an S-oxide group.14 On the other hand,
comparison of the NMR spectral data of 9 and 11 with those
of 5 indicated that the structural differences resided in the
pyrrole moieties.15 Compared with 5, the 1H NMR spectrum
of 9 displayed signals for only 12 protons and the 13C NMR
spectrum indicated two sp3 carbons (δC 87.2, and 86.2)
instead of two sp2 carbons in 5. These results, together with
2D-NMR spectral data, suggested a pyrroloindole ring
system for compound 9. Similarly, the 2-oxindole substruc-
ture of dithiocarbamate 11 was deduced from NMR spectral
data. The structure of the minor acid 12 followed also from
spectroscopic data. Compounds 9-11 do not appear to have
been previously described; the absolute configurations of 9
and 11 remain to be determined. The chemical structures
of metabolites 9-12 were confirmed unambiguously by
chemical synthesis, of which the conversion of S-oxide 10 to
oxindole 11 is particularly noteworthy.
We have previously demonstrated that the unusual
brassinin S-oxide (13), although rather unstable at room
temperature, can be prepared in modest yield (ca. 30%) by
reaction of dithiocarbamate 1 with m-CPBA;16 similar reac-
tion of 5 with m-CPBA yielded S-oxide 10 in ca. 50% yield.
Surprisingly, 10 rearranged to oxindole 11 on standing in
CDCl3 solution at room temperature.17 This transformation
proceeded much slower in CH2Cl2 or CHCl3 but at a similar
rate in CH2Cl2 in the presence of TsOH. Furthermore,
oxidation of 5 with m-CPBA, followed by acidification of the
reaction mixture with TsOH, yielded oxindole 11 directly.
These results suggest that formation of 11 might be initiated
by protonation of the indole ring at C-3,18 followed by
intramolecular oxygen transfer from CdSdO to C-2 with
regeneration of the dithiocarbamate group (Scheme 2).
Compounds containing a dithiocarbamate S-oxide group are
rarely isolable and have been characterized mostly as
mixtures.19 Structures similar to 10 (i.e., S-oxide 13) appear
to have been described only once.3,4 However, the rear-
rangement observed for S-oxide 10 was not noted with 13,
which decomposed to brassinin (1) and undetermined prod-
ucts on standing. Although the decomposition of certain
dithiocarbamate S-oxides appears to lead mainly to the thiol
carbamates (sulfur extrusion), particularly in the presence
of acid,19 we have not isolated such products from 10 or 13.
The relatively higher stability of dithiocarbamate 5 and the
corresponding S-oxide 10 in relation to brassinin (1) and
A route for the metabolism of the brassinin homologue 5
by P. lingam is proposed in Scheme 1. There appear to be
two major pathways, one leads to acid 8 via tryptamine (6)
and acetyltryptamine (7) and the other yields dithiocarbam-
ate 9; neither 8 or 9 appeared to be metabolized further. In
addition, a minor pathway may lead to products 11 and 12
(e1%). Although 11 appeared as a likely precursor of acid
12, we have been unable to observe the fungal transforma-
tion of 11 into 12. Surprisingly, although 5 is significantly
more inhibitory to P. lingam than 10, S-oxide 10 adminis-
tered to fungal cultures was reduced to mostly 5, in less than
48 h. Control experiments indicated that 10 was stable
under those incubation conditions for several days, suggest-
ing that the transformation of 10 to 5 was an enzymatic
process. In additional experiments, we determined that the
fungal transformation of 5 occurred significantly faster in
the absence of light (3-5 days vs 2 weeks), with a similar
metabolite profile being detected. Finally, we determined
that none of the metabolic products 8-12 showed noticeable
inhibitory activity against P. lingam.
In conclusion, the metabolism of dithiocarbamate 5 by
virulent isolates of P. lingam is a detoxification process;
however, this process proceeded much slower than that of
the naturally occurring homologue 1. Both fungal degrada-
tion pathways of dithiocarbamates 1 and 5 yielded carboxylic
acids, none of which appeared to be inhibitory against P.
lingam. Clearly, further work must be carried out to probe
the specificity of the detoxifying enzymes and to understand
the intriguing effect of light on the rate of the fungal
transformation of 5. Nonetheless, our results strongly
suggest that phytoalexin analogues can be designed to deter
fungal phytopathogens more effectively than phytoalexins.
Moreover, such analogues might have the advantage of
acting synergistically with the natural disease resistance
factors of plants.
(14) Compound 10 could exhibit E/ Z isomerism, as reported previously
for brassinin S-oxide.3 Nevertheless, considering the sharpness of its NMR
resonances, it appeared that only one isomer was present and an
stereochemistry would allow intramolecular hydrogen bonding.
(15) Spectroscopic data are in complete agreement with the proposed
structures. See the Supporting Information.
E
(16) Prepared as reported in ref 4.
(17) Compound 11 was isolated in 40% yield together with 5 (20%) and
polar unidentified products.
(18) (a) Szabo-Pusztay, K.; Szabo, L, Synthesis 1979, 276-277. (b) J oule,
J . A.; Mills, K.; Smith, G. F. Heterocyclic Chemistry, 3rd ed.; Chapman &
Hall: London, 1995; p 317.
Ack n ow led gm en t. Support for this work was obtained
in part from the Natural Sciences and Engineering Re-
search Council of Canada and the University of
Saskatchewan.
(19) Segall, Y.; Casida, J . E. J . Agric. Food Chem. 1983, 31, 242-246.
(20) Nakagawa, M.; Okajima, H.; Hino, T. J . Am. Chem. Soc. 1977, 99,
4424-4429.
(21) Kametani, T.; Kanaya, N.; Ihara, M. J . Chem. Soc., Perkin Trans. 1
1981, 959-963.
Su p p or tin g In for m a tion Ava ila ble: Synthesis and spectro-
scopic data for 5, 9, 10, and 11 (3 pages).
J O971702I