A R T I C L E S
Henry and Townsend
pathway. DMST (3) and its tetrahydrobisfuran congener 22
undergo two separate rounds of O-methylation in vivo, which
must occur prior to the final oxidative transformation to
aflatoxin.16 Each round is catalyzed by rigorously substrate-
specific transferases with the capacity for neither cross-
methylation of the C-5 or C-7 phenols nor methylation out of
order (Scheme 5). The C-5 phenol is methylated first by
AflO17-19 and C-7 second by AflP.20-22 The logic of this order,
and rationale for O-methylation at C-5 particularly, is now
apparent. The angular xanthones 3 and 22 formed biochemically
are susceptible to acid-catalyzed isomerization to the thermo-
dynamically favored linear forms, which would derail the
aflatoxin pathway. The O-methylation to form sterigmatocystin
(30) prevents this from occurring. As an aside, we suspect that
the linear xanthone sterigmatin (32), isolated from the sterig-
matocystin producer Aspergillus Versicolor, may well be an
artifact of its isolation by silica gel chromatography.23-25
Realizing now that the angular xanthone specifically formed
in the biosynthetic reaction is thermodynamically disfavored
gave added dimension and further impetus to find a mechanistic
explanation for the anthraquinone w xanthone skeletal rear-
rangement in aflatoxin formation.
Incorporation Experiments. Whole-cell DIS1 fermentation
studies and the subsequent thin-layer and high-performance
liquid chromatographic (TLC and HPLC) analyses were carried
out as described.6 DIS1 is a deletion mutant of the aflatoxin
producer Aspergillus parasiticus defective in the first biosyn-
thetic step to norsolorinic acid (1, Scheme 1) and, thus, has
very low background amounts of aflatoxins.26 As noted above,
the intended fermentation experiments were complicated by our
inability to isolate either of the intended angular xanthones, 22
or 25, in pure form. This was so owing to the facile isomer-
ization of the desired angular xanthone to the linear, which could
be readily isolated as a homogeneous isomer. Therefore, linear
xanthones 21 and 24 were each incubated in pure form. The
results of these experiments were then compared to parallel
incubations containing mixtures of the linear and angular
xanthones, each administered immediately following hydro-
genolysis to maximize the proportion of the angular isomer.
Any differences in the results of these two pairs of incubations
were specifically attributed to the presence of either 9-hydroxy-
ldihydro-DMST (25) or dihydro-DMST (22). As the results of
the whole-cell and ground-cell experiments were nearly identi-
cal, only the whole-cell data will be presented. As expected,
neither of the linear isomers 21 or 24 was incorporated (data
Figure 4. Thin-layer chromatogram of DIS-1 whole-cell incubations
developed in 60% CHCl3, 30% EtOAc, and 10% HCO2H, photographed
under long-wave UV light. Lanes A-C are standards: (A) (top to bottom)
AFB1 (4), AFB2, AFG1, andAFG2; (B) linear and angular 9-deoxyxan-
thones 21 (upper) and 22 (lower); and (C) linear and angular 9-hydroxyx-
anthones 24 (upper) and 25 (lower). Lanes I-III: (I) chloroform extract of
DIS-1 control; (II) DIS-1 with 21/22 (3); and (III) DIS-1 with 24/25.
not shown), although the deoxyxanthone 21 gave a trace of
aflatoxin during the much longer whole-cell study. We regard
this observation as an artifact that could owe to thermodynami-
cally uphill equilibration of 21 with very small amounts of the
angular isomer 22 during the time of the experiment. Fermenta-
tions were next conducted with mixtures of the linear and
angular xanthones (21 with 22 and 24 with 25, characterized
by 1H NMR spectroscopy immediately prior to the incubations).
These experiments resulted in great increases of AFB2 and
AFG2 from the deoxy series, 21 with 22, attributable to the
incorporation of dihydro-DMST (22; Figure 4, lane II, and
Figure 5, chromatogram II). Significantly, no such increases
were measured from the 9-hydroxyl series, 24 with 25,
establishing no incorporation of angular 9-hydroxyldihydro-
DMST [25; Figure 4, lane III (identical to control lane I), and
Figure 5, chromatogram III]. The overwhelming incorporation
of dihydro-DMST, an established pathway intermediate, served
as a positive control and gave us confidence in the nonincor-
poration of angular 9-hydroxyldihydro-DMST (25). We judged
that 25 is not a precursor of DMST (22) or the aflatoxins.
(16) Udwary, D. W.; Casillas, L. K.; Townsend, C. A. J. Am. Chem. Soc. 2002,
124, 5294-5303.
(17) Yabe, K.; Matsushima, K.; Koyama, T.; Hamasaki, T. Appl. EnViron.
Microbiol. 1998, 64, 166-171.
(18) Yu, J. J.; Chang, P. K.; Payne, G. A.; Cary, J. W.; Bhatnagar, D.; Cleveland,
T. E. Gene 1995, 163, 121-125.
(19) Motomura, M.; Chihaya, N.; Shinozawa, T.; Hamasaki, T.; Yabe, K., Appl.
EnViron. Microbiol. 1999, 65, 4987-4994.
Discussion
(20) Keller, N. P.; Dischinger, H. C.; Bhatnagar, D.; Cleveland, T. E.; Ullah,
A. H. J. FASEB J. 1991, 5, A822-A822.
An apparent cytochrome P450, AflN, and a probable NADPH-
dependent reductase, AflM, are implicated by gene “knockout”
experiments to be responsible for the rearrangement of the
anthraquinone versicolorin A (2) to demethylsterigmatocystin
(3). To account for both oxidative cleavage and reductive loss
of the C-6 hydroxyl in this overall process, the simplest
mechanism must invoke two rounds of oxidation and a single
reductive step. Whole-cell5 and cell-free7 experiments with
6-deoxyversicolorin A (7, Scheme 6) gave no incorporation into
aflatoxin under conditions where VA (2) itself was readily
(21) Keller, N. P.; Dischinger, H. C.; Bhatnagar, D.; Cleveland, T. E.; Ullah,
A. H. J. Appl. EnViron. Microbiol. 1993, 59, 479-484.
(22) Yu, J. J.; Cary, J. W.; Bhatnagar, D.; Cleveland, T. E.; Keller, N. P.; Chu,
F. S. Appl. EnViron. Microbiol. 1993, 59, 3564-3571.
(23) Hatsuda, Y.; Ishida, M.; Matsui, K.; Hara, S.; Hamasaki, T. Agric. Biol.
Chem. 1972, 36, 521.
(24) Hamasaki, T.; Matsui, K.; Isono, K.; Hatsuda, Y. Agric. Biol. Chem. 1973,
37, 1769-1770.
(25) Fukuyama, K.; Tsukihara, T.; Katsube, Y.; Hamasaki, T.; Hatsuda, Y.;
Tanaka, N.; Ashida, T.; Kakudo, M. Bull. Chem. Soc. Jpn. 1975, 48, 1639-
1640.
(26) Mahanti, N.; Bhatnagar, D.; Cary, J. W.; Joubran, J.; Linz, J. E. Appl.
EnViron. Microbiol. 1996, 62, 191-195.
9
3728 J. AM. CHEM. SOC. VOL. 127, NO. 11, 2005