Citrinin revisited: From monomers to dimers and beyond

Article abstract of DOI:10.1039/b600960c

Detailed chemical analysis of the solid phase fermentation of an Australian Penicillium citrinum isolate has returned the known compounds citrinin (1), phenol A acid (6), dihydrocitrinone (7) and dihydrocitrinin (8), together with a novel cytotoxic dimer, dicitrinin A (5). Dicitrinin A (5) was determined to be a dimerised artefact of the major co-metabolite citrinin, and its structure solved by spectroscopic analysis and chemical modification. Analysis of the products encountered during the controlled decomposition of citrinin led to the discovery of additional citrinin dimers and delineated a plausible mechanistic pathway linking all monomeric and dimeric citrinin degradation products. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b600960c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Citrinin revisited: from monomers to dimers and beyond
Benjamin R. Clark,a Robert J. Capon,*a Ernest Lacey,b Shaun Tennantb and Jennifer H. Gillb
Received 20th January 2006, Accepted 27th February 2006
First published as an Advance Article on the web 16th March 2006
DOI: 10.1039/b600960c
Detailed chemical analysis of the solid phase fermentation of an Australian Penicillium citrinum
isolate has returned the known compounds citrinin (1), phenol A acid (6), dihydrocitrinone (7) and
dihydrocitrinin (8), together with a novel cytotoxic dimer, dicitrinin A (5). Dicitrinin A (5) was
determined to be a dimerised artefact of the major co-metabolite citrinin, and its structure solved by
spectroscopic analysis and chemical modification. Analysis of the products encountered during the
controlled decomposition of citrinin led to the discovery of additional citrinin dimers and delineated a
plausible mechanistic pathway linking all monomeric and dimeric citrinin degradation products.
that led to the sequential transformation of citrinin into both
monomeric and dimeric products, and also includes a detailed
mechanistic analysis of the decomposition.
Introduction
Well known for the production of biologically active metabolites,
Penicillium citrinum Thom is a common filamentous fungus found
the world over. One of the best known P. citrinum metabolites is
citrinin (1), a polyketide mycotoxin first isolated from a P. citrinum
strain in 1931.1 The structure for citrinin (1) was first proposed
by Brown et al. in 1948,2 although subsequent studies were
needed to clarify the absolute stereochemistry3 and tautomeric
configuration.4 Extensive biosynthetic studies have been carried
out on 1 using a variety of isotopic labelling strategies.5,6 A large
number of citrinin derivatives have been isolated, most notably
by Curtis et al., who in 1968 described the isolation of seven
metabolites related to citrinin.7
Citrinin (1) is a well-known contaminant of a number of agricul-
tural products, and has been demonstrated to possess nephrotoxic
activity8 in addition to a number of other chronic toxic effects.9
As such, several studies have been carried out on its detoxification
via degradation. These studies concluded that decomposition of
citrinin occurs at◦>175 ◦C under dry conditions, and at lower tem-
peratures (>100 C) in the presence of water.10,11 Known citrinin
decomposition products include phenol A (2) and the formylated
derivative citrinin H2 (3).12 In most cases, decomposition of
citrinin coincides with a decrease in cytotoxicity, however, at least
one study13 noted that the decomposition of citrinin under aque-
ous conditions led to an increase in cytotoxicity. This increased
cytotoxicity was attributed to formation of citrinin H1 (4).13
The current report provides an account of a chemical investiga-
tion into an Australian P. citrinum isolate. In addition to describing
the isolation and structure elucidation of citrinin and a selection
of known analogues, this report presents the first account of a new
cytotoxic citrinin dimer, dicitrinin A (5). Recognizing that 5 could
be an artefact brought about during handling and storage, this
study includes a detailed chemical analysis of the decomposition of
citrinin under various conditions. This study defines the conditions
Results and discussion
The MeOH extract derived from a solid fermentation of P.
citrinum (MST-F10130) was concentrated in vacuo and the residue
fractionated by repeated reverse phase SPE and HPLC to yield
the known P. citrinum metabolites citrinin (1), phenol A acid
(6), dihydrocitrinone (7), and dihydrocitrinin (8), in addition to
the unprecedented citrinin dimer, dicitrinin A (5). The known
compounds 1, 6, 7 and 8 were identified by spectroscopic analysis
(HRESIMS, 1H and 13C NMR, UV-vis, and [a]D) and comparison
to literature data.8,14–19
Dicitrinin A (5) was isolated as a deep red solid that returned
a HRESI(+)MS pseudomolecular ion (M + Na, m/z 403.1520)
corresponding to a molecular formula (C23H24O5) requiring 12
double bond equivalents (DBE). Examination of the H NMR
spectrum revealed two isolated spin systems, designated fragments
A and B.
aCentre for Molecular Biodiversity, Institute for Molecular Bioscience, Uni-
versity of Queensland, Brisbane, QLD, 4072, Australia. E-mail: r.capon@
imb.uq.edu.au; Fax: +61 7 3346 2090; Tel: +61 7 3346 2979
1
bMicrobial Screening Technologies Pty. Ltd., Smithfield, NSW, 2164,
Australia. E-mail: elacey@microbialscreening.com; Fax: +61 2 9757 4515;
Tel: +61 2 9757 2586
1520 | Org. Biomol. Chem., 2006, 4, 1520–1528
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
compound of known trans relative stereochemistry, the 1H NMR
resonances reported for H-3 and H-4 in 10 were described as
multiplets, with no experimental measure of 3J3 ,4
20 In the absence
ꢀ
.
ꢀ
1
of experimental data we simulated H NMR spectra for 10 with
both natural trans and unnatural cis stereochemistries, returning
calculated 3J3,4 values of 4.0 and 8.2 Hz, respectively. The former
3
ꢀ
ꢀ
value was very close to the experimentally observed J3 ,4 value
for fragment B (3.6 Hz), suggestive of a trans-stereochemistry
about these methyls. In this respect fragment B can be viewed as a
decarboxylated ring B contracted “citrinin” analogue.
Fragments A and B account for all 1H and 13C NMR resonances
observed in the NMR spectra of 5, however, consideration of
the molecular formula confirmed that one DBE and one proton
remained unaccounted for, and the two fragments shared a
“common” oxygen atom. The missing DBE could be accounted
for by requiring that a single ring be formed in the fusion of
fragments A and B. Likewise, the “common” oxygen atom could
explained if the C-8 oxygen of fragment A was common with one of
either the C-6ꢀ or the C-8ꢀ oxygens evident in fragment B. Finally,
the missing proton could be attributed to a phenolic proton on
fragment B.
Given the partial structures attributed to fragments A and B,
and the observations listed above, there existed two alternative
ways to assemble the two fragments. The chemical shift for C-7ꢀ
(dC 104.7) dictated that it could not be oxygenated, thus C-7ꢀ in
fragment B must be linked to C-1 in fragment A. Given that the
C-8 oxygen in fragment A must be common with either the C-
6ꢀ or the C-8ꢀ oxygen in fragment B, the two alternate structures
for dicitrinin A are shown in Fig. 2. At this point the available
evidence was incapable of unambiguously determining which of
these two structures was correct. Efforts at observing an NOE
correlation between the phenolic OH and the C-5ꢀ methyl could
not be pursued as, despite the use of several alternative NMR
solvents, the phenolic OH resonance was not observed in any 1H
NMR data set. As an alternative, O-methylation of the phenolic
OH in 5 was attempted, in an effort to obtain a methyl ether that
would reveal the crucial NOE correlations.
The 1H NMR resonances associated with fragment A consisted
of two 2◦ methyls (dH 1.26 and 1.37), one deshielded methyl (dH
2.22), two methines (dH 3.55 and 5.46), and a deshielded singlet
(dH 7.03). A suite of gCOSY correlations confirmed that the 2◦
methyls and methines were mutually coupled, and this, together
with gHMBC correlations, established the partial structure as
indicated (see Fig. 1). All 13C NMR chemical shifts for fragment A
were consistent with the assigned structure, with C-1, C-6 and C-8
all deemed sp2 and oxygenated based on their deshielded chemical
shifts (dC 169.8, 170.4 and 157.6, respectively). The chemical shifts
for C-1 and C-6 in particular, were consistent with the quinone
methide indicated. The relative stereochemistry of the 2◦ methyls
in fragment A were determined to be trans based on 3J3,4 (<0.5 Hz)
which was comparable to 3J3,4 values observed for citrinin (1). In
this regard, fragment A resembles a C-7 decarboxylated “citrinin”
akin to decarboxycitrinin (9).
Fig. 1 Structure fragments for dicitrinin A (5). Arrows indicate key
gHMBC correlations, while bold bonds highlight gCOSY correlations.
The 1H NMR resonances associated with fragment B were
similar to those of fragment A, consisting of two 2◦ methyls (dH
1.30 and 1.36), a singlet methyl (dH 2.27), and two methines (dH
3.42 and 4.75). Again, gCOSY and gHMBC analysis (see Fig. 1)
established a partial structure which was strongly reminiscent of
the known citrinin analogue 10.20 Carbon resonances for C-7ꢀ and
C-8ꢀ (dC 104.7 and 137.3, respectively) in fragment B, which could
not be assigned by gHMBC analysis, were assigned by comparison
with literature data for 10 (dC 102.8 and 138.9, respectively). The
relative configuration about the two 2◦ methyls in fragment B
Fig. 2 Plausible alternative structures for dicitrinin A (5).
Small scale (ca. 100 lg) trial methylations of dicitrinin A (5) with
MeI proved promising, with LC-DAD-MS analysis confirming
the formation of a monomethylated product (m/z 395, M +
H). However, when the methylation was repeated on a larger
scale (1.6 mg) the major product (1.1 mg, 64%) proved to be
an unexpected dimethylated species 11 (m/z 409, M). The most
noticeable change in the 1H NMR spectrum of dimethyl dicitrinin
was assigned by consideration of 3J3 ,4 . Although a useful model
ꢀ
ꢀ
This journal is The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1520–1528 | 1521
©

View Article Online
Table 1 NMR data (600 MHz, d6-DMSO) for dicitrinin A (5) and dimethyl dicitrinin A (11)
Position
H #
Dicitrinin A (5)
Dimethyl dicitrinin A (11)
a
b
dC
dH (m, J/Hz)
gCOSY
gHMBC
dC
dH (m, J/Hz)
gHMBC
1
3
4
4a
5
6
7
8
8a
9
169.8
84.9
33.3
137.5
124.1
170.4
99.4
157.6
102.4
18.2
18.8
10.0
87.9
44.2
144.9
119.1
149.3
104.7
137.3
137.2
20.6
170.8
85.5
33.0
136.7
124.5
169.4
97.3
158.0
104.2
18.3
17.8
10.0
88.9
44.2
143.3
126.6
149.9
—
5.46 (q, 6.8)
3.55 (q, 7.2)
H-9
H-10
C-1, C-4, C-4a, C-9, C-10
C-4a, C-5, C-8a, C-10
5.59 (q, 6.6)
3.65 (q, 7.2)
C-1, C-4a
C-4a, C-5, C-8a, C-9
7.03 (s)
H-11
C-5, C-6, C-8, C-8a
7.60 (s)
C-5, C-6, C-8, C-8a
1.37 (d, 6.8)
1.26 (d, 7.2)
2.22 (s)
4.75 (dq, 3.6, 6.4)
3.42 (dq, 3.6, 7.0)
H-3
H-4
C-3, C-4
C-3, C-4, C-4a
1.39 (d, 6.6)
1.28 (d, 7.2)
2.27 (s)
4.85 (dq, 4.0, 6.5)
3.49 (obs)
C-3, C-4
C-3, C-4, C-4a
C-4a, C-5, C-6
10
11
3ꢀ
H-7
C-4a, C-5, C-6
H-4ꢀ, H-9ꢀ
H-3ꢀ, H-10ꢀ
C-4aꢀ, C-8ꢀ/8aꢀ, C-10ꢀ
C-4aꢀ, C-8ꢀ/8aꢀ, C-9ꢀ, C-10ꢀ
4ꢀ
4aꢀ
5ꢀ
6ꢀ
7ꢀ
8ꢀ
—
—
20.3
18.3
11.3
58.8
62.2
8aꢀ
9ꢀ
1.36 (d, 6.4)
1.30 (d, 7.0)
2.27 (s)
H-3ꢀ
C-3ꢀ, C-4ꢀ
1.40 (d, 6.5)
1.34 (d, 7.0)
2.38 (s)
4.14 (s)
3.84 (s)
C-3ꢀ, C-4ꢀ
C-3ꢀ, C-4ꢀ, C-4aꢀ
C-4aꢀ, C-5ꢀ, C-6ꢀ
C-6
10ꢀ
11ꢀ
6-Me
6ꢀ-Me
18.7
12.0
H-10ꢀ
C-3ꢀ, C-4ꢀ, C-4aꢀ
C-4aꢀ, C-5ꢀ, C-6ꢀ
C-6ꢀ
a Assignments were assisted by HSQC, HMBC, and DEPT experiments, and by comparison with related compounds. b Most 13C values have been
extracted from HMBC and HSQC spectra and assigned by comparison to 5. C-7ꢀ, C-8ꢀ and C-8aꢀ were not observed.
A (11), relative to dicitrinin A (5), was the appearance of two new
methyl singlets (dH 4.14 and 3.84, see Table 1) consistent with the
presence of two OMe groups. gHMBC correlations from these
new OMe groups to carbons at 169.4 and 149.9 ppm, respectively,
implied that both the C-6 and C-6ꢀ oxygens had been methylated.
While dicitrinin A (5) exhibited strong pseudomolecular ions in
ESI(+)MS and ESI(−)MS, dimethyl dicitrinin A (11) only ionised
in the positive mode, returning an exceptionally clean ESI(+)MS
spectrum. Even when HRESI(+)MS data was acquired in the
presence of a NaI reference, no (M + Na) ion was observed,
suggesting that 11 possessed an inherent positive charge. That 11
Fig. 3 Dimethyl dicitrinin A (11). Arrows indicate key ROESY correla-
tions. Extended resonance delocalization stabilizes the oxonium.
was cationic was further supported by its high water solubility
(compared to 5, which is EtOAc soluble and water insoluble), and
its immobility on silica TLC (cf. rf 11 = 0.0 vs. 5 = 0.25, with an
UV-vis studies
eluant of 99 : 1 EtOAc–HCO2H). This cationic character was also
apparent from UV-vis studies, described below. This unexpected
dimethylation not withstanding, a ROESY NMR experiment
carried out on 11 revealed informative correlations between the C-
6ꢀ OMe and C-5ꢀ methyl, and between the C-6 OMe and H-7 (see
Fig. 1), confirming the relative stereostructures of both dimethyl
dicitrinin A (11) and dicitrinin A (5). The formation and stability
of the dimethyl oxonium 11 can be attributed to extensive charge
delocalisation as proposed in Fig. 3.
Given the assigned structure for dicitrinin A (5) it seemed highly
likely that 5 was an artefact brought about by dimerization of
the abundant co-metabolite citrinin (1). As such, given that the
absolute stereochemistry of citrinin is known3 and the same as
citrinin encountered in this study (i.e. same [a]D), by inference
the absolute stereochemistry of dicitrinin A (5) and its dimethyl
analogue 11 can be assigned as shown.
The UV-vis spectrum of 5 was observed to undergo reversible
pH sensitive shifts (see Fig. 4). Under basic conditions a dramatic
shift was observed in the kmax of the electron-transfer band at 422 to
540 nm. Such bathychromic shifts are well known for deprotonated
phenolic compounds,21 although the magnitude of the shift was
unusually high in this case. Under acidic conditions another
change in the UV-vis spectrum occurred, with the appearance
of a new peak at 490 nm, attributed to the formation of a
protonated species. Curiously, all dicitrinins occur as red solids,
but are bright yellow in solution. This colour change may be
indicative of an amphoteric character, with the solid form being
composed partially or fully of a “salt” comprised of oxonium
and oxyanion species such as those given in Fig. 4. Consistent
with this hypothesis, dimethyl dicitrinin A (11), which is no longer
amphoteric, forms a yellow solid.
1522 | Org. Biomol. Chem., 2006, 4, 1520–1528
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
the oxonium moiety and disrupted the extended conjugation
(see Fig. 5). Further evidence for a nucleophilic addition adduct
was obtained by LC-DAD-MS analysis of a sample of 11 to
which NaOH had been added, which revealed a single peak (m/z
449, M + Na) consistent with the proposed product shown in
Fig. 5. This change in the UV-vis spectrum of 11 proved to be
reversible, although attempts to effect comparable changes with
nucleophiles other than hydroxide proved unsuccessful.
Decomposition studies
After elucidation of the structure of dicitrinin A (5), consideration
was given to the hypothesis that 5 was an artefact of the handling
and/or storage procedures. Initial decomposition studies were de-
signed to mimic the conditions encountered during fractionation:
i.e., mild heating, and exposure to mild acid and base.
Two samples of citrinin (10 mg each) were dissolved in aqueous
MeOH (4 mL), with one sample being acidified with a drop of
trifluoroacetic acid (TFA), and the other basified with a drop of
triethylamine (TEA). The two samples were heated at 50 ◦C in
sealed flasks for 7 d and their decomposition progress monitored
by LC-DAD-MS. The presence of TEA appeared to stabilise
the citrinin (1) and no significant decomposition was observed.
However, under TFA conditions decomposition did occur, yielding
products that were tentatively identified as phenol A (2), phenol A
acid (6) and decarboxycitrinin (9) based on LC-DAD-MS analysis.
No dimeric products were observed. This result prompted a second
set of decompositions at higher concentration, in the hope of
promoting dimerization. For solubility reasons these studies were
carried out in MeOH. Three samples of citrinin (1) (5 mg each)
were dissolved in a small amount of MeOH (100 lL), with a
drop of TFA added to one, TEA to another; and the third left
unmodified. All three samples were heated at 50 ◦C in sealed vials
over a period of 15 d, with the reaction progress monitored by LC-
DAD-MS. Under neutral conditions dicitrinin A (5) was formed
as a major product. Dicitrinin A (5) was also formed in both
the TFA and TEA modified samples, however yields were not as
high, with TEA slowing the decomposition and TFA favouring
the formation of monomeric products.
Fig. 4 UV-vis spectra of dicitrinin A (5) under different pH conditions,
and speculated products. Vial images are EtOAc–H2O partitions.
The dimethyl derivative 11 also displayed interesting changes
in its UV-vis spectrum (see Fig. 5). Not surprisingly, addition
of acid resulted in no change, however, under basic conditions
the prominent peaks at 449 and 379 nm disappeared entirely. It
was speculated that the dramatic change in the UV-vis spectrum
was due to nucleophilic addition of OH− at C-1, which quenched
To further investigate the identities of the decomposition prod-
ucts, the experiment was repeated on a larger scale (50 mg), under
neutral conditions in MeOH (0.5 mL). The reaction was halted
after 14 d and LC-DAD-MS analysis of the crude decomposition
mixture performed (see Fig. 6).
The resulting mixture was then fractionated by reverse phase
HPLC: details of the major products are presented in Table 2. The
major product was dicitrinin A (5), with an estimated yield of ca.
8 mg (16%). Other known compounds isolated from the mixture
included all of the monomeric species detected in the earlier
aqueous methanol decompositions. Three new dimeric products
(12–14) were isolated and their identities are addressed below.
A noteworthy feature of the decomposition products was their
different behaviour during ESIMS analysis. While all products
formed strong M + H ions under ESI(+)MS conditions, two
different ionisation processes were observed during ESI(−)MS
acquisition-leading to either M − H or M + 17 ions as the
dominant species. It is known that citrinin exists as a hydrate (15) in
aqueous solution (see Fig. 10),14 so it is reasonable to assume that
decomposition products bearing an analagous quinone methide
Fig. 5 UV-vis spectra of dimethyl dicitrinin A (11) under different pH
conditions, and speculated products.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1520–1528 | 1523
©

View Article Online
Table 2 Data for major compounds isolated from large-scale citrinin (1) decomposition
Compound
Yield/mg
Major ions (+ve)
Major ions (−ve)
UV-vis maximaa
Phenol A (2)
3.3 (6.6%)
4.2 (8.4%)
1.5 (3.0%)
0.9 (1.8%)
1.0 (2.0%)
<0.7 (1.4%)
8.0 (16%)
197 (M + H)
241 (M + H)
207 (M + H)
437 (M + H)
251 (M + H)
393 (M + H)
381 (M + H)
383 (M + H)
195 (M − H), 179 (M − 17)
239 (M − H)
282, 223 (sh), 202
315, 252, 214
336, 288, 228 (sh), 203
413 (sh), 394, 267, 232
409 (sh), 332, 236
450 (sh), 425 (sh), 378, 293 (sh), 269, 230
516 (sh), 451 (sh), 422 (sh), 385, 308, 279, 223b
408, 281, 267, 241
Phenol A acid (6)
Decarboxycitrinin (9)
Dicitrinin D (14)
Citrinin (1)
223 (M + H2O − H), 205 (M − H)
453 (M + H2O − H)
267 (M + H2O − H)
409 (M + H2O − H)
379 (M − H)
Dicitrinin C (13)
Dicitrinin A (5)
Dicitrinin B (12)
<0.8 (1.6%)
381 (M − H)
a Extracted from DAD. b This UV-vis spectrum does not correspond to any of those shown in Fig. 2, due to the presence of protonated species.
established by HRESI(+)MS analysis (m/z 383.1495, M + H),
was consistent with the oxidative cleavage of an aromatic methyl
in dicitrinin A (5). 1H NMR chemical shifts and 2D NMR
correlations for 12 were consistent with this assessment, and
gHMBC correlations confirmed placement of the aromatic methyl
at C-5ꢀ, implying that the C-5 methyl had been cleaved. gHMBC
correlations from H-3, H-4 and H3-10 to a common carbon (dC
110.6), assigned as C-4a, supported this hypothesis. The 13C NMR
resonance for C-4a (dC 110.6) in 12 was significantly shielded rela-
tive to C-4a in 5 (dC 137.5), as would be predicted by the placement
of a hydroxy group at C-5. The arguments presented above permit
a tentative structure assignment to dicitrinin B (12) as indicated.
The ESI(−)MS spectra of 12 was dominated by an M − H ion, as
predicted. A plausible mechanism for the oxidative transformation
Fig. 6 HPLC DAD trace (254 nm) from the decomposition of citrinin
of dicitrinin A (5) into dicitrinin B (12) is presented in Fig. 7.
(1). Details for major peaks are given in Table 2. Yellow peaks represent
monomeric species while red peaks correspond to dimeric species.
moiety could generate hydrated ions (i.e. M + H2O–H) under
ESI(−)MS conditions. Those products that bear a phenolic OH
might be expected to suppress this ion forming pathway, in favor
of the more direct formation of an M − H ion. Consistent with this
hypothesis, decarboxycitrinin (9) exhibits both M − H and M +
H2O − H ions in the ESI(−)MS, whereas phenols 2 and 6, which
lack the capacity to form hydrates show only M − H ions. In the
ESI(−)MS of dicitrinin A (5) the M − H ion dominates as hy-
dration at C-1 would disrupt the extended conjugation. This “hy-
drated ion” hypothesis proved a valuable tool in exploring possible
structures for the new citrinin dimers, dicitrinins B–D (12–14).
Fig. 7 Proposed mechanism for the oxidative conversion of dicitrinin A
(5) to yield dicitrinin B (12).
In the absence of NMR data the structures tentatively proposed
for dicitrinins C (13) and D (14) rely heavily on mass spectral
Of the new dicitrinins only dicitrinin B (12) returned useful 1H
NMR data. The 1H NMR spectrum of 12 was very similar to that
of 5, consisting of four 2◦ methyls (dH 1.21, 1.27, 1.28 and 1.32), an
aromatic methyl (dH 2.34), four methines (dH 3.24, 3.29, 4.57 and
5.01), a deshielded singlet (dH 6.29), and a broad exchangeable
resonance (dH 9.81). The molecular formula of 12 (C22H22O6)
and UV-vis analysis, and consideration of likely decomposition
outcomes. HRESI(+)MS analysis revealed molecular formulae
for 13 (C24H24O5) and 14 (C25H24O7) suggestive that the former
was a dicitrinin A analogue that had not undergone ring con-
traction, whereas the latter was a dicitrinin C precursor that had
not undergone decarboxylation. While by necessity speculative,
1524 | Org. Biomol. Chem., 2006, 4, 1520–1528
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
the proposed structures are consistent with the available data,
and represent plausible intermediates along the decomposition
pathway from citrinin to dicitrinins A and B. The ESI(−)MS for
dicitrinin C (13) and D (14) are dominated, as predicted, by M +
H2O − H ions. Plausible mechanisms for a ring contraction of
dicitrinin C (13) to yield dicitrinin A (5), and a decarboxylation of
dicitrinin D (14) to yield dicitrinin C (13), are shown in Fig. 8 and
9, respectively.
The mechanisms proposed in Fig. 7–9, that address decom-
position processes of oxidation, ring contraction and decarboxy-
lation, respectively, delineate key stages in the interconversion of
dicitrinins. These processes, together with the formation of citrinin
hydrate, keto/enol tautomerization, redox and hydroysis can also
be used to explain the formation of monomeric citrinin decom-
position products. A plausible citrinin decomposition pathway
linking all monomeric species mentioned in this report is shown
in Fig. 10. A key driver for the proposed pathway is the ability
of citrinin (1) to form the hydrate 15. A redox reaction between
1 and 15 can lead to the oxidized product dihydrocitrinone (7)
and the reduced product dihydrocitrinin (8). Likewise, 15 can ring
open to lead to 2, 3 and 6, or ring open and then ring contract to
give 10.
Having proposed the majority of the decomposition pathway,
the key “dimerization” event remains to be explained. A plausible
mechanism for dimerization of citrinin (1) is presented in Fig. 11.
This mechanism involves a heterocyclic Diels–Alder connection
between citrinin (1) and an alternate citrinin tautomer, followed
by the irreversible loss of CO2 and H2O to form a key interme-
diate. This hypothetical intermediate, which was not detected in
our studies, may undergo oxidation to re-establish the quinone
methide and extended conjugation, leading to dicitrinins A–
D. Alternatively this intermediate can undergo oxidative ring
opening to give citrinin H1 (4).
While the structures for dicitrinins B–D (12–14) are only
tentatively assigned, the discussion presented on dicitrinin A
(5), including spectroscopic and mechanistic observations, reveals
citrinin as an entry point into hitherto underexplored chemical
space. It is interesting to speculate on the range of Diels–Alder
products that could be obtained by challenging citrinin by an
array of natural and synthetic dienophiles. Curiously, nature has
provided a glimpse into such space with the timely report by
Kobayashi et al. in 200522 of perinadine A (16) from a marine-
derived strain of P. citrinum Perinadine A (16) represents a natural
mixed citrinin and pyrrolidine alkaloid Diels–Alder adduct.
Fig. 8 Proposed mechanism for the ring contraction conversion of
dicitrinin C (13) to yield dicitrinin A (5).
In the current study, all purified compounds were screened
against a range of test microorganisms, and against the mouse
NS-1 cell line. None possessed significant antimicrobial activity,
but citrinin (1) and dicitrinin A (5) were moderately active in the
NS-1 cytotoxicity assay, with LD99 values of 25 and 6.3 lg mL−1,
respectively.
Our experimental protocols and mechanistic hypotheses, de-
tailed above, for the conversion of citrinin (1) into both monomeric
and dimeric decomposition products provides a comprehensive
molecular basis for future analytical chemistry investigations into
citrinin-contaminated foods and their detoxification. To support
these studies we are also seeking to investigate the formation
Fig. 9 Proposed mechanism for the decarboxylation of dicitrinin D (14)
yield dicitrinin C (13).
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1520–1528 | 1525
©

View Article Online
Fig. 10 Proposed mechanism for the decomposition of citrinin (1) into monomeric species.
of Diels–Alder adducts between citrinin and a range of natural
and synthetic dienophiles. The results from these studies will be
reported at a future date.
Experimental
General experimental procedures have been described previously.23
NMR spectrum simulations were carried out using ACD HNMR
Predictor software, version 7.
Analytical HPLC gradients
Standard LC-DAD-MS analyses were carried out using the
following gradient: 1 mL min−1 gradient elution from 90% H2O–
MeCN (0.05% HCO2H) to MeCN (0.05% HCO2H) using a 5 lm
Zorbax StableBond C8 150 × 4.6 mm column.
Assays
Details for antimicrobial and cytotoxicity assays have been
described.24
Biological material
The fungal strain (MST-F10130) was isolated from a roadside soil
sample collected near Ardlethan in New South Wales, Australia.
The isolate was identified as Penicillium citrinum Thom on
morphological grounds.
Isolation
A solid fermentation (1 kg wheat, 21 d, 28 ◦C) was extracted
Fig. 11 Proposed mechanism for the dimerization of citrinin (1), and
subsequent formation of dicitrinins A–D and citrinin H1.
twice with MeOH (ca. 6 L). These extracts were combined and
1526 | Org. Biomol. Chem., 2006, 4, 1520–1528
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
concentrated in vacuo to an aqueous concentrate (2 L) and
triethylamine added to adjust to pH ca. 8.5. This was passed
through four parallel C18 SPE cartridges (4 × 10 g, Varian HF
C18) followed by sequential elution with 50% H2O–MeOH (4 ×
40 mL each) and MeOH (4 × 40 mL each). The aqueous eluant
was adjusted to pH ca. 3.5 with the addition of trifluoroacetic acid
(TFA) and once more passed through the same C18 SPE cartridges
followed by similar sequential elution to afford 50% H2O–MeOH
and MeOH fractions. All 50% H2O–MeOH and MeOH eluants
were evaporated to give a combined residue (ca. 7.9 g). This was
subjected to preparative HPLC (60 mL min−1 with gradient elution
of 70 to 40% H2O–MeCN (0.01% TFA) over 20 min followed by
MeCN (0.01% TFA) for 10 min, through a 5 mm Phenomenex
Luna C18(2) 50 × 100 mm column), giving 10 fractions.
Dimethyl dicitrinin A (11). Dicitrinin A (5) (1.6 mg) was
dissolved in dry EtOAc (0.5 mL) after which Na2CO3 (10 mg)
and an excess of MeI (50 lL) were added, and the mixture stirred
overnight at 50 ◦C. The mixture was then filtered and evaporated
to dryness. Purification of the product was carried out using C8
HPLC (3.2 mL min−1 gradient elution from 60 to 30% H2O–
MeCN (0.1% TFA) over 20 min, through a 5 lm Zorbax C8 10 ×
250 mm column), to give dimethyl dicitrinin A (11) as a yellow
solid (1.1 mg, 64%). UV-vis kmax(EtOH)/nm 449 (e/dm3mol−1cm−1
830), 379 (6800), 306 (sh) (4100), 282 (7400), 208 (10400); UV-
vis kmax(EtOH–NaOH)/nm 299 (e/dm3mol−1cm−1 2600), 251 (sh)
(4600), 219 (19200); 1H NMR (d6-DMSO, 600 MHz) see Table 1;
13C NMR (d6-DMSO, 150 MHz) see Table 1; ESI(+)MS m/z
409 (M); HRESI(+)MS m/z 409.2012 (M, C25H29O5 requires
409.2015).
One of the combined fractions was partitioned between bu-
tanol and water, and the butanol-soluble material subjected to
preparative HPLC (22 mL min−1 gradient elution from 70 to 30%
H2O–MeCN over 16 min, through a 5 lm Zorbax RX–C8 21.2 ×
250 mm column) to yield citrinin (1) (130 mg) and dihydrocitrinin
(8) (9 mg). Another of the combined fractions was also partitioned
between butanol and water, and the water-soluble material further
purified by HPLC (2.5 mL min−1 gradient elution from 85 to
33% H2O–MeCN, through a 5 lm Zorbax StableBond C18 9.4 ×
250 mm column) to give phenol A acid 6 (6 mg).
Decomposition Studies
For the small scale “dilute” solution decompositions, the following
procedure was used. Two duplicate samples of citrinin (1) (10 mg)
were dissolved in aqueous MeOH (1 : 1, 4 mL) in vials, and a drop
of TFA added to one and TEA to the other. The vials were then
sealed and the mixture stirred at 50 ◦C for 4 d. Reaction progress
was monitored by LC-DAD-MS.
One of the more polar fractions was re-subjected to preparative
HPLC (60 mL min−1 with isocratic elution of 79% H2O–MeCN
(0.01% TFA) over 30 min, through a 5 lm Phenomenex Luna C18(2)
50 × 100 mm column), followed by C18 SPE (Alltech C18 Extract-
clean 2 g cartridge, 10% stepwise gradient elution from 80 to 100%
MeOH) to give dihydrocitrinone (7) (18 mg) and dicitrinin A (5)
(8 mg).
The small scale “concentrated” solution decompositions were
carried out in sealed vials. Three duplicate samples of citrinin (1)
(5 mg) were each dissolved in MeOH (100 lL) in vials. A drop of
TFA was added to one vial, a drop of TEA to another, and no
additive to the third. The sealed mixtures were heated at 50 ◦C for
15 d, and reaction progress monitored by LC-DAD-MS.
The large-scale decomposition was carried out in the same
manner, using 50 mg (0.2 mmol) of 1 in 0.5 mL MeOH, for
15 d. Fractionation of the products formed in the large-scale
decomposition was carried out using C18 HPLC (22 mL min−1
gradient elution from 20% H2O–MeCN (0.1% TFA) to MeCN
(0.1% TFA) over 30 min, through a 5 lm Zorbax StableBond-C18
21.2 × 250 mm column) to yield phenol A (2) (3.3 mg, 6.6%),
phenol A acid (6) (4.2 mg, 8.4%), and dicitrinin A (5) (8.0 mg,
16%) as the major products. A small quantity of unreacted citrinin
(1) (1.0 mg, 2%) was also recovered, while decarboxycitrinin (9)
(estimated 1.5 mg) was isolated but decomposed before analysis.
Three minor dimers, dicitrinin B (12) (0.8 mg, 1.6%), dicitrinin C
(13) (0.7 mg, 1.4%) and dicitrinin D (14) (0.9 mg, 1.8%) were also
isolated. UV-vis spectra for dicitrinins B–D were extracted from
DAD data-reported extinction coefficients are relative only.
Citrinin (1). Identified by spectroscopic analysis. [a]D,
ESI( )MS, 1H and 13C NMR were in good agreement with
literature values.8,14
Dicitrinin A (5). Red solid; [a]D +73.9◦ (c 0.016, CHCl3); IR
mmax(CHCl3)/cm−1 3495, 2985, 2932, 1618, 1533, 1455, 1411, 1381;
UV-vis kmax(EtOH)/nm 446 (sh) (e/dm3mol−1cm−1 7300), 422
(10500), 401 (sh) (8900), 304 (3400), 293 (sh) (4500), 276 (11100),
267 (11000), 223 (12100); UV-vis kmax(EtOH–HCl)/nm 490
(e/dm3mol−1cm−11300), 384 (9500), 309 (7600), 286 (8200), 245
(7200), 223 (9100), 208 (10400); UV-vis kmax(EtOH–NaOH)/nm
540 (e/dm3mol−1cm−1 4600), 403 (8400), 329 (4000), 283 (sh)
(5400), 262 (10100), 218 (br) (18600); 1H NMR (d6-DMSO,
600 MHz) see Table 1; 13C NMR (d6-DMSO, 150 MHz) see
Table 1; ESI(+)MS m/z 381 (M + H); ESI (−) MS m/z 379 (M–
H); HRESI(+)MS m/z 403.1520 (M + Na, C23H24O5Na requires
403.1521).
Phenol A (2). Identified by spectroscopic analysis. ESI( )MS,
1H NMR and 13C NMR were in good agreement with literature
values.17,18
Phenol A acid (6). Identified by spectroscopic analysis. [a]D,
ESI( )MS, 1H NMR and 13C NMR were in good agreement with
literature values.16–18
Dicitrinin B (12). A red solid; UV-vis kmax(MeCN–H2O)/nm
408 (e/dm3mol−1cm−1 9200), 281 (7300), 267 (6500), 241 (10000);
dH(600 MHz, d6-DMSO) 9.81 (brs, OH), 6.29 (s, H-7), 5.01
(q, J 5.9, H-3), 4.57 (dq, J 6.4, 4.2, H-3ꢀ), 3.29 (obs, H-4ꢀ), 3.24 (q,
J 6.8, H-4), 2.34 (s, H3-11), 1.32 (d, J 6.4, H3-9ꢀ), 1.28 (d, J 6.9,
H3-10ꢀ), 1.27 (d, J 5.9, H3-9), 1.21 (d, J 6.8, H3-10); ESI(+)MS
m/z 787 (2M + Na), 383 (M + H); ESI(−)MS m/z 381 (M–
H); HRESI(+)MS m/z 383.1495 (M + H, C22H23O6 requires
383.1495).
Dihydrocitrinone (7). Identified by spectroscopic analysis. [a]D,
ESI( )MS, 1H and 13C NMR were in good agreement with
literature values.15,25
Dihydrocitrinin (8). Identified by spectroscopic analysis. [a]D,
ESI( )MS, 1H and 13C NMR were in good agreement with
literature values.19
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 1520–1528 | 1527
©

View Article Online
Dicitrinin C (13). A red solid; UV-vis kmax(MeCN–H2O)/nm
450 (sh) (e/dm3mol−1cm−1 2300), 425 (sh) (3200), 378 (7800), 293
(sh) (2900), 269 (10000), 230 (6600); ESI(+)MS m/z 393 (M + H);
ESI(−)MS m/z 841 (2M + Na + 2H2O − 2H), 409 (M + H2O −
H); HRESI(+)MS m/z 393.1703 (M + H, C24H25O5 requires
393.1702).
6 J. Barber, A. C. Chapman and T. D. Howard, J. Antibiot., 1987, 40,
245–248.
7 R. F. Curtis, C. H. Hassall and M. Nazar, J. Chem. Soc. C, 1968, 85–93.
8 P. Krogh, E. Hasselager and P. Friis, Acta Pathol. Microbiol. Scand.,
Sect. B: Microbiol. Immunol., 1970, 78, 401–413.
9 V. Betina, Mycotoxins: Chemical, Biological, and Environmental
Aspects, Elsevier, Amsterdam, 1989.
10 N. Kitabatake, A. B. Trivedi and E. Doi, J. Agric. Food Chem., 1991,
39, 2240–2244.
Dicitrinin D (14). A red solid; UV-vis kmax(MeCN–H2O)/nm
413 (sh) (e/dm3mol−1cm−1 8000), 394 (8800), 267 (10000), 232
(7100); ESI(+)MS m/z 895 (2M + Na), 437 (M + H); ESI(−)MS
m/z 453 (M + H2O − H); HRESI(+)MS m/z 437.1598 (M + H,
C25H25O7 requires 437.1600).
11 A. B. Trivedi, E. Doi and N. Kitabatake, J. Food Sci., 1993, 58, 229–
232.
12 M. Hirota, A. B. Menta, K. Yoneyama and N. Kitabatake, Biosci.,
Biotechnol., Biochem., 2002, 66, 206–210.
13 A. B. Trivedi, M. Hirota, D. Etsushiro and N. Kitabatake, J. Chem.
Soc., Perkin Trans. 1, 1993, 2167–2171.
14 J. Barber, J. L. Cornford, T. D. Howard and D. Sharples, J. Chem. Soc.,
Perkin Trans. 1, 1987, 2743–2744.
Acknowledgements
15 M. M. Chien, P. L. Schiff, Jr., D. J. Slatkin and J. E. Knapp, Lloydia,
1977, 40, 301–302.
The authors would like to acknowledge A. Hocking for taxonomic
studies and G. MacFarlane for acquisition of HRMS data.
B. Clark was the recipient of a SGSA scholarship from the
University of Queensland. This research was partially funded by
the Australian Research Council.
16 N. J. Cartwright, A. Robertson and W. B. Whalley, J. Chem. Soc., 1949,
1563–1567.
17 T. Roedel and H. Gerlach, Liebigs Ann., 1995, 885–888.
18 J. P. Brown, A. Robertson, W. B. Whalley and N. J. Cartwright, J. Chem.
Soc., 1949, 867–879.
19 J. Deruiter, J. M. Jacyno, R. A. Davis and H. G. Cutler, J. Enzyme
Inhib., 1992, 6, 201–210.
20 C.-H. Chen, C.-Y. Shaw, C.-C. Chen and Y.-C. Tsai, J. Nat. Prod., 2002,
65, 740–741.
References
1 A. C. Hetherington and H. Raistrick, Philos. Trans. R. Soc. London,
Ser. B, 1931, 220, 269–296.
21 A. I. Scott, Interpretation of the Ultraviolet Spectra of Natural Products,
Pergamon Press, Oxford, 1964.
2 J. P. Brown, N. J. Cartwright, A. Robertson and W. B. Whalley, Nature,
1948, 162, 72–73.
3 R. K. Hill and L. A. Gardella, J. Org. Chem., 1964, 29, 766–767.
4 S. Kovac, P. Nemec, V. Betina and J. Balan, Nature, 1961, 190, 1104–
1105.
5 L. Colombo, C. Gennari, D. Potenza, C. Scolastico, F. Aragozzini and
C. Merendi, J. Chem. Soc., Perkin Trans. 1, 1981, 2594–2597.
22 M. Sasaki, M. Tsuda, M. Sekiguchi, Y. Mikami and J. Kobayashi, Org.
Lett., 2005, 7, 4261–4264.
23 B. R. Clark, R. J. Capon, E. Lacey, S. Tennant and J. H. Gill, Org.
Biomol. Chem., 2006, DOI: 10.1039/b600959j.
24 B. Clark, R. J. Capon, M. Stewart, E. Lacey, S. Tennant and J. H. Gill,
J. Nat. Prod., 2004, 67, 1729–1731.
25 C. H. Hassall and D. W. Jones, J. Chem. Soc., 1962, 4189–4191.
1528 | Org. Biomol. Chem., 2006, 4, 1520–1528
This journal is
The Royal Society of Chemistry 2006
©