Total synthesis and isolation of citrinalin and cyclopiamine congeners

Article abstract of DOI:10.1038/nature13273

Many natural products that contain basic nitrogen atoms-for example alkaloids like morphine and quinine-have the potential to treat a broad range of human diseases. However, the presence of a nitrogen atom in a target molecule can complicate its chemical synthesis because of the basicity of nitrogen atoms and their susceptibility to oxidation. Obtaining such compounds by chemical synthesis can be further complicated by the presence of multiple nitrogen atoms, but it can be done by the selective introduction and removal of functional groups that mitigate basicity. Here we use such a strategy to complete the chemical syntheses of citrinalinB and cyclopiamineB. The chemical connections that have been realized as a result of these syntheses, in addition to the isolation of both 17-hydroxycitrinalinB and citrinalinC (which contains a bicyclo[2.2.2]diazaoctane structural unit) through carbon-13 feeding studies, support the existence of a common bicyclo[2.2.2]diazaoctane-containing biogenetic precursor to these compounds, as has been proposed previously.

Full text of DOI:10.1038/nature13273

ARTICLE
doi:10.1038/nature13273
Total synthesis and isolation of citrinalin
and cyclopiamine congeners
Eduardo V. Mercado-Marin1*, Pablo Garcia-Reynaga1*, Stelamar Romminger2, Eli. F. Pimenta2, David K. Romney3,
Michael W. Lodewyk4, David E. Williams5, Raymond J. Andersen5, Scott J. Miller3, Dean J. Tantillo4, Roberto G. S. Berlinck2
& Richmond Sarpong1
Many natural products that contain basic nitrogen atoms—for example alkaloids like morphine and quinine—have the
potential to treat a broad range of human diseases. However, the presence of a nitrogen atom in a target molecule can
complicate its chemical synthesis because of the basicity of nitrogen atoms and their susceptibility to oxidation. Obtain-
ing such compounds by chemical synthesis can be further complicated by the presence of multiple nitrogen atoms, but it
can be done by the selective introduction and removal of functional groups that mitigate basicity. Here we use such a
strategy to complete the chemical syntheses of citrinalin B and cyclopiamine B. The chemical connections that have been
realized as a result of these syntheses, in addition to the isolation of both 17-hydroxycitrinalin B and citrinalin C (which
contains a bicyclo[2.2.2]diazaoctanestructuralunit) throughcarbon-13feeding studies, support the existenceofa common
bicyclo[2.2.2]diazaoctane-containing biogenetic precursor to these compounds, as has been proposed previously.
The prenylated indole alkaloids are an emerging class of natural pro- the isolation of two new citrinalins, provide support for a proposed
ducts typified by the presence of an indole ring, or derivatives thereof biogenesis of the subset of prenylated indole alkaloids that lack the
(that is, spirooxindole or pseudoindoxyl), decorated by one or more bicyclo[2.2.2]diazaoctane core.
prenyl groups or the vestige of a prenyl group. Isolates from this family
Biosynthetic connections
of natural products include citrinalins A and B (Fig. 1, 1and 2) and cyclo-
piamines A and B (4 and 6), which are the focus of this Article. The
modificationsoftheindolecoreintheprenylatedindolealkaloidfamily,
which occur by a reaction with dimethylallyl pyrophosphate1, results
in the introduction of a chromene unit as is found in (1)stephacidin A
(10; see blue highlighted portion) or a bicyclo[2.2.2]diazaoctane core
that is typical of many congeners, including 11 and 12 (ref. 2) (see red
highlighted portion).
Althoughstructurallysimilar, theprenylatedindolealkaloidsdisplay
a diverse range of bioactivities including anti-tumour, insecticidal, anthel-
mintic, calmodulin-inhibition and antibacterial properties3. The recent
discoveryofcitrinadins A(ref. 4)andB(ref. 5)(7and8)andPF1270A–
PF1270C6 (9a–9c) has added anunprecedented dimensiontothe struc-
tural motifs afforded by the Penicillium strains, and has raised several
questions as to the biogenesis of these structurally related alkaloids.
Recently, syntheses of citrinadins A and B have been achieved7,8. Par-
ticularly intriguing to us is a subset of this emerging subclass including
citrinalins A and B (1and 2) and cyclopiamines A and B (4and 6), which,
like the citrinadins, lack the bicyclo[2.2.2]diazaoctane framework and,
remarkably, possess an alkyl nitro group. Cyclopiamines A and B (4
and 6) were discovered in 1979 in a toxinogenic strain of Penicillium
cyclopium9, whereas citrinalins A and B (1 and 2) were discovered in
2010 in a strain of Penicillium citrinum10. Although natural products
that possess aryl nitro groups are known, those that contain aliphatic
nitro groups are extremely rare11. As a result, the citrinalins and cyclo-
A stimulating connection may be drawn between cyclopiamine A and
B via the intermediacy of nitronate iminium ion 5 (ref. 9) (Fig. 1). The
interconversion of 4 and 6 was demonstrated by heating either com-
pound in a mixture of dioxane and water or in dimethylformamide9
(DMF). This led to a proposal that 6, which is the more stable of the
two isomers (we have computed 6 to be 9.6 kcal mol21 lower in energy
than 4 in a DMF solvent model; see Supplementary Information), may
in fact be an isolation artefact. Given the likelihood that the citrinadins,
citrinalins and cyclopiamines are all oxidative degradation products of
a precursor containing a bicyclo[2.2.2]diazaoctane ring, such as marc-
fortine A (11; in the case of the citrinadins) or stephacidin A (10; in the
case of the citrinalins and cyclopiamines), we wondered whether the
citrinalins could be transformed to the cyclopiamines. On the basis of
this assumption, it is particularly baffling that, unlike cyclopiamines A
and B, which are related by an aza-Henry (or nitro-Mannich) reaction
as shown in Fig. 1 (4 u 6, via 5), citrinalin A and the originally pro-
posed structure of citrinalin B (3) would be related not by the formal
epimerization of the C22 stereocentre but rather by the nature of the
relative configuration of the C14 carbon (highlighted in 2 and 3). On
the basis ofthe connection between cyclopiamines A and B demonstrated
previously9, we intuited that the structure of citrinalin B may be better
represented by 2. To support this proposal, we undertook a computa-
tionalsimulationofthe1Hand13CNMRspectrathatwouldbeexpected
for the neutral andsaltformsof citrinalins A andB (SupplementaryInfor-
piamines, which also possess three nitrogen atoms in chemically dis- mation). As has been convincingly demonstrated in numerous cases,
tinct environments, are unusual and are therefore attractive targets for this method provides an accurate prediction of the structures of com-
synthesis. The synthetic studies described here have culminated in the plex natural products12. We found that the computed and empirical data
total syntheses of ent-citrinalin B (ent-2; ent, enantiomer) and cyclo- forthetrifluoroaceticacid salt formof citrinalin A isingoodagreement
piamine B (6), and, along with 13C feeding studies that have resulted in with those reported in ref. 10. The corrected mean absolute deviations
1Department of Chemistry, University of California, Berkeley, California 94720, USA. 2Instituto de Quimica de Sao Carlos, Universidade de Sao Paulo, CP 780, CEP 13560-970, Sao Carlos, SP, Brazil.
3Department of Chemistry, Yale University, PO Box 208107, New Haven, Connecticut 06520, USA. 4Department of Chemistry, University of California, Davis, California 95616, USA. 5Department of
Chemistry and Earth and Ocean Sciences, University of British Columbia, Vancouver, British Columbia V6T IZI, Canada.
*These authors contributed equally to this work.
3 1 8
| N A T U R E | V O L 5 0 9 | 1 5 M A Y 2 0 1 4
Macmillan Publishers Limited. All rights reserved
©2014

ARTICLE RESEARCH
O
O
O
NO2
NO2
22
NO2
22
N
N
H
N
14
O
O
O
14
HN
HN
HN
H
H
H
H
H
O
O
O
Citrinalin A (1)
Citrinalin B (2) (revised)
Citrinalin B (3) (original)
MeO
MeO
MeO
O
O
NO2
N
H
NO2
O
O
O
N
N
H
N
H
N
N
N
H
H
O
H
O
O
Cyclopiamine A (4)
5
Cyclopiamine B (6)
O
Me
O
O
Me
Me
NHMe
NHMe
OH
NHMe
H
N
H
O
N
H
N
H
O
O
HN
O
HN
HN
O
OR
OH
O
O
NMe2
O
O
Citrinadin A (7)
Citrinadin B (8)
PF1270A (R = (CH2)2CH3, 9a)
PF1270B (R = CH2CH3, 9b)
PF1270C (R = CH3, 9c)
O
O
O
O
N
NH
O
O
N
NMe
N
NMe
N
H
H
HN
HN
O
OH
H
O
H
O
Me
O
O
O
Marcfortine A (11)
Paraherquamide A (12)
N
O
O
N
N
H
H
H
(+)-stephacidin A (10)
Figure 1
|
Selected prenylated indole alkaloids. The prenylated indole
this family (for example 1 and 4) have emerged that do not possess this
alkaloid family encompasses over 80 natural products, some of which contain a structural motif. Me, methyl.
bicyclo[2.2.2]diazaoctanecoreas in 10, 11 and12. Recently, severalmembers of
(CMAD) in the 1H and 13C NMR resonances are 0.21 and 2.0 p.p.m., to proceed on the assumption that 2 most probably represents the cor-
respectively (the largest outliers are 1.0 and 5.2 p.p.m., respectively). How- rect structure of citrinalin B. Ultimately, a reanalysis of the NMR data
of citrinalin B, collected in MeOH-d4 (Supplementary Information),
corroborates the assignment of 2 as the true structure of citrinalin B.
ever, the computed data for the trifluoroacetic acid salt form of 3 (the
originally proposed structure of citrinalin B) differs significantly from
thatrecordedusingthenaturallyoccurringmaterial(CMADs, 0.45and
2.0 p.p.m.; largest outliers, 2.3 and 9.6 p.p.m. for 1H and 13C, respec-
tively). The best match to the reported spectral data was found to cor-
respond to 2 in its neutral form (CMADs, 0.12 and 1.6 p.p.m.; largest
outliers, 0.38 and 4.4 p.p.m. for 1H and 13C, respectively), which cor-
Synthesis
AsoutlinedinFig. 2,cyclopiamineB(6)canbeobtainedfromtheenan-
tiomer of citrinalin B (ent-2) by using a chromanone rearrangement
to forge the tetrahydroquinolone structural moiety found in the cyc-
roborates thepotentially similarbiosyntheticconnection that hasbeen lopiamines. In turn, ent-2 could be taken back using an ‘indole-to-
established forthe cyclopiamines (outlined in Fig. 1). Asa result, wechose spirooxindole’transformto fusedhexacycle 13. Fusedindole 13 would
MeO
O
O
H2N
NO2
NO2
O
O
N
H
N
H
N
H
O
N
H
N
HN
H
H
H
O
O
Cyclopiamine B (6)
Ent-citrinalin B (ent-2)
13
OMe
O
O
NC
HN
NC
HO
N
H
N
H
TIPSO
O
O
H
H
D-proline (17)
14
15
16
Figure 2
|
Retrosynthetic analysis plan for cyclopiamine B and citrinalinB. The syntheses of natural products 2 and 6 are expected to arise from common
intermediate 13. TIPS, triisopropylsilyl.
1 5 M A Y 2 0 1 4
| V O L 5 0 9 | N A T U R E | 3 1 9
Macmillan Publishers Limited. All rights reserved
©2014

RESEARCH ARTICLE
arise from tricycle 14, which may be prepared from diene 15, the t- face selectivity for the oxygenation of 25 is poor, attention was turned
butyldimethylsilyl variant of which was first prepared in ref. 13, and to the use of reagent control to achieve the desired diastereoselective
tetrahydroindolizinone 16, which would ultimately arise from D-proline oxygenation. Inthisregard, weweredrawntopeptide-derivedcatalysts
(17).
We initiated our synthetic studies with the protection of D-proline
developed for oxygenations27. Following an investigation of a focused
library ofpeptide catalysts, 32(Fig. 4b)emergedasthesuperiorcatalyst
(20 mol% loading) and provided hydroxyindolenine 28 in 83% yield
from25. Hydroxyindolenine28rearrangeswithheatingusingSc(OTf)3
over 2 h to afford pseudoindoxyl 33 (Fig. 4c) instead of the desired spi-
rooxindole. The equilibrium between pseudoindoxyls and spirooxin-
doles is well recognized and has been studied for the migration of C2
alkyl substituents28 and C2 aryl substituents29. However, despite pro-
longed heating, furtherrearrangement of pseudoindoxyl33 to the desired
spirooxindole was not observed. It is possible that an intramolecular
hydrogen bond stabilizes pseudoindoxyl 33 against further rearrange-
by t-butoxycarbonyl (Fig. 3), which was followed by the reduction of
thecarboxylicacidgroupandSwernoxidation oftheresultinghydroxyl
toaffordaldehyde18 (ref. 14). Alkynylative homologationof the aldehyde
group of18using the Ohira–Bestmann method15, followedbyremoval
of the t-butoxycarbonyl group and acylation with 2-cyanoacetyl chlo-
ride gives alkyne 19. This serves as a substrate for an unprecedented
formal cycloisomerization that probably proceeds via a metal vinyli-
dene intermediate16, anti-Markovnikov hydration and Knoevenagel
condensation to give tetrahydroindolizinone 16. At this stage, a SnCl4-
catalysed Diels–Alder [412] reaction17 between 16 and diene 15, and a
subsequent basic work-up, affords an enone (notshown), which is iodin-
ated to yieldiodoenone 20 (ref. 18). A mild hydrolysis of thenitrile group
of 20is achieved using Pt-complex21 (ref. 19) to afford thecorresponding
carboxamide,whichservesas a substrate for a Hofmann rearrangement
thatiseffectedwithphenyliodosylbistrifluoroacetatetoyieldcarbamate
22 (ref. 20). Suzuki cross-coupling of 22 with known boronic ester 23
(ref. 21) gives adduct 24, which is efficiently converted to fused indole
25usingtwosequentialreductions—allinaccordwiththeeffectivepro-
tocols established in ref. 22.
The face-selective oxygenation of C2/C3-fused indoles is a well-
established route to hydroxyindolenines, which serve as precursors to
the corresponding spirooxindoles23. We therefore reasoned that the
oxygenation of indole 25 (Fig. 4a) could be a path to the spirooxindole
structural moiety found in 1, 2, 4 and 6. On the basis of related prece-
dents for heteroatom-directed oxygenation7,24,25, we expected the car-
bamate group of 25 to direct oxygenation to the a-face and provide
28. Surprisingly, the use of Davis’ oxaziridine26 (29; 3.0 equiv.) leads to
26 and trace amounts of both hydroxyindolenine 28 and spirooxin-
dole 27 (spirooxindole 27 arises via the intermediacy of hydroxyindo-
lenine 26). A survey of other oxaziridines, including 30 and 31, leads,
at best (using 31), to a 1:1 ratio of thedesired hydroxyindolenine, 28, and
both hydroxyindolenine 26 and spirooxindole 27. Because the inherent
˚
ment (a bond distance of 2.24 A is computed for the pseudoindoxyl
carbonylgroupand N–Hproton ofthecarbamategroupin33;seeSup-
plementary Information). Evidence for a stabilizing intramolecular hydro-
genbond in 33 comes from the observation that hydroxyindolenine 26
(preparedbyoxidationof25withDavis’oxaziridine)rearrangesreadily
at room temperature in the presence of mild acid to spirooxindole 27;
a pseudoindoxyl generated from 26 would lack the analogous stabiliz-
ing hydrogen bond. However, the possibility exists that 26 proceeds to
anepoxideintermediate(seeAin insetinFig. 4c)that rearrangesto27.
The difficulty of further rearranging pseudoindoxyl 33 caused us to con-
sider alternative approaches that would produce the desired spiroox-
indole structural moiety of the citrinalins and cyclopiamines.
Amino compound 35 (Fig. 5) was prepared on the assumption that
an amino group, or some oxidized derivative thereof (for example the
corresponding hydroxylamine), could serve as a hydrogen-bond donor
toeffectstereoselectiveoxygenationoftheindoleC2–C3bondandthen,
by further oxidation to a nitroso or nitro group, remove the presumed
intramolecular hydrogen bond that may stabilize the pseudoindoxyl
form (as in 33). It seemed reasonable that this sequence would facili-
tate the eventual conversion of 35 to nitro spirooxindole compound
36. Initial experiments established that epoxidation of the chromene
ring was a competing reaction that occurred under various oxygena-
tion conditions. As such, we opted to effect a Wacker oxidation30 of 25
O
O
(7) Cat. [Ru(Cp)(MeCN)L2]PF6
98%
HN
H
BocN
H
NC
NC
(1)–(3)
(4)–(6)
HO
N
H
N
H
PPh2
65%
(3 steps)
87%
(3 steps)
O
O
N
L =
D-proline (17)
18
16
19
t-Bu
Me
Me
Me
Me
P
H
P
O
O
Me
P
O
Pt
O
OMe
Me
OH
HN
H
O
O
H
B
NC
H
21
(10), (11)
O
I
I
(8), (9)
N
H
N
H
+
O
NO2
70%
(2 steps)
68%
(2 steps)
O
O
23
20
22
O
O
OMe
N
O
NH
H
O
OMe
HN
H
O
O
(13), (14)
(12)
N
H
95%
82%
(2 steps)
N
H
O2N
H
O
25
24
Figure 3
|
Preparation of fused hexacycle 25. The use of a Diels–Alder
(6) 2-cyanoacetylchloride, Et3N, CH2Cl2, 0 uC to RT; (7) acetonitrile
bis[2-diphenylphosphino-6-t-butylpyridine] cyclopentadienylruthenium(II)
hexafluorophosphate (8 mol%), acetone and H2O, 70 uC; (8) 15, SnCl4, 278 to
reaction involving a proline-derived indolizidinone dienophile affords a key
tricycle that is advanced to hexacycle 25 by Suzuki coupling to boronic ester
23. Reagents and conditions are as follows: (1) di-t-butyl dicarbonate (Boc2O), 242 uC; (9) I2, 4-dimethylaminopyridine, pyridine and CCl4, 60 uC; (10)
NaHCO3, H2O and tetrahydrofuran (THF), room temperature (RT 5 23 uC); 21 (20 mol%), EtOH and H2O, RT; (11) phenyliodosyl bistrifluoroacetate,
(2) BH3NTHF, THF, 0 uC to RT; (3) (COCl)2, dimethylsulphoxide, CH2Cl2,
diisopropylethylamine, 278 uC; (4) dimethyl (diazomethyl)phosphonate,
K2CO3, MeOH, 0 uC to RT; (5) 4N HCl and dioxane, 0 uC to RT;
MeOH, RT; (12) dppfPdCl2 (10 mol%), K3PO4, DMF, 40 uC; (13) Zn dust,
NH4Cl, HCO2NH4, p-TsOH, MeOH, RT; (14) NaCNBH3, 1N HCl(aq.), 0 uC to
RT. dppf, (diphenylphosphino)ferrocene; Et, ethyl; t-Bu, t-Butyl; Ts, tosyl.
3 2 0
| N A T U R E | V O L 5 0 9 | 1 5 M A Y 2 0 1 4
Macmillan Publishers Limited. All rights reserved
©2014

ARTICLE RESEARCH
a
O
O
OMe
N
O
HN
O
O
OMe
N
OMe
HN
H
O
O
HN
O
HN
OH
O
(1)
H
H
N
H
O
27
N
H
N
H
O
H
OMe
25
26
O
HN
OH
O
N
H
Oxaziridines
Bu
Ph
O
Ph
O
N
O
H
H
H
28
N
N
N
S
S
S
O
O
O
O
O
O
29 (3.0 equiv.)
61% yield 26 (trace 28 and 27)
30 (3.0 equiv.)
1:4 28:(27+26)
31 (3.0 equiv.)
1:1 28:(27+26)
b
BnO
O
O
O
Me
OMe
N
N
OMe
HN
H
O
Me
HN
O
HN
CbzHN
HO2C
O
OH
(2)
O
O
N
H
O
HN
Me
83% yield
N
H
N
H
H
28
25
Indole oxidation catalyst (32)
c
O
O
H bond?
(2.24 Å apart)
OMe
OMe
N
HN
O
HN
O
O
OH
O
(3)
N
H
N
73% yield
N
O
H
H
O
H
H
28
33
O
O
O
OMe
OMe
N
O
HN
HN
OH
O
(4)
N
H
HN
O
N
84% yield
H
H
H
26
27
O
OMe
O
HN
O
N
H
O
N
H
H
A
Figure 4
|
Face-selective oxygenation of fused hexacycle 25. a, Oxidative
hydroxyindolenine (26) affords the corresponding spirooxindole (27).
Reagents and conditions are as follows: (1) oxaziridine (29, 30, or 31), CH2Cl2,
RT; (2) 32 (20 mol%), 4-dimethylaminopyridine, diisopropylcarbodiimide,
H2O2, CHCl3, 4 uC; (3) Sc(OTf)3, toluene, 110uC; (4) 23 mM HCl, CH2Cl2, RT.
Bn, benzyl; Cbz, carboxybenzyl; Ph, phenyl.
rearrangement studies of fused indole 25 with a range of oxaziridines leads
predominantly to the undesired, epimeric, hydroxyindolenine (26) and
spirooxindole (27). b, Use of indole oxidation peptide catalyst 32 to effect
oxidation yields the desired hydroxyindolenine (28). c, Hydroxyindolenine
28 rearranges to an undesired pseudoindoxyl (33), whereas the epimeric
toaffordchromanone34(Fig. 5), whichwouldbeadvantageousbecause extensive investigation, this task was effectively accomplished using a
the chromanoneunitis found in thecitrinalinsand cyclopiamines. Remark- modification of a known procedure32 by treating 36 with a variant of
ably,treatmentof35(followingremoval of the methoxycarbonylgroup Meerwein’ssalt(Me3OBF4), whichprobably leadstoamethylatedami-
in 34) with an excess of dimethyldioxirane (formed in situ from acet- dinium intermediate that is cleanly reduced with sodium cyanoborohy-
one and Oxone) affords spirooxindole 36 as the major product (dia- dride to give ent-citrinalin B (ent-2) in 66% yield (79% based on recovered
stereomeric ratio, 4:1) where the spiro centre is as desired and the nitro starting material). The spectroscopic data for the neutral form of ent-2
group has been installed. Studies of dimethyldioxirane oxidations of are fully consistent with previous data reported for the compound believed
indoles tospirooxindoles31 suggestthatspirooxindole36might arise from to be citrinalin B (ref. 10; corroborating the computational predictions
epoxide B (Fig. 5, inset). Therefore, it is possible that the introduction of and reanalysisinMeOH-d4), except for the sign of opticalrotation, which
the chromanone diminishes the participatory role of the indole nitro- is opposite. The structure of ent-2 was unambiguously confirmed by
gen lone pair leading, afterrearrangement (seedirectionof arrowin B), X-ray crystallographic analysis of its HCl salt. Ent-citrinalin B is easily
to 36. With spirooxindole 36 in hand, what remained was a selective convertedtocyclopiamine B(6)ontreatmentofent-2withsodiumhydride
removal of the tertiary amide carbonyl group by reduction, which had and heating (to effect the conversion of chromanone to tetrahydroqui-
to be accomplished in the presence of the chromanone and secondary nolone) and subsequent methylation of the resulting phenol. The struc-
amide carbonyl groups as wellas the newlyintroduced nitro group. After ture of cyclopiamine B (6) was also unambiguously confirmed by X-ray
1 5 M A Y 2 0 1 4
| V O L 5 0 9 | N A T U R E | 3 2 1
Macmillan Publishers Limited. All rights reserved
©2014

RESEARCH ARTICLE
O
O
Figure 5
|
Completion of the
syntheses of ent-citrinalinB and
cyclopiamine B. The total syntheses
of 2 and 6 required the identification
of conditions that accomplished the
oxidation of the amino group and
spirooxindole formation in one pot
as well as unique conditions for the
selective reduction of the tertiary
amide carbonyl group. The
OMe
N
OMe
HN
H
HN
O
O
O
O
(2)
(1)
N
H
N
H
96% yield
N
H
71% yield
H
H
O
34
25
O
O
H2N
O
O
(3)
O2N
N
H
rearrangement of ent-citrinalin B
(2) to cyclopiamine B (6) was also
demonstrated. Reagents and
N
56% yield
(4:1 d.r.)
N
H
O
HN
H
O
H
H
35
O
conditions are as follows: (1)
36
Pd(OAc)2 (40 mol%), benzoquinone,
H2SO4, MeCN and H2O, RT; (2)
Me2S, methanesulphonic acid, 40 uC;
(3) Oxone (10 equiv.), NaHCO3,
acetone and H2O, 0 uC to RT;
O
NO2
(4), (5)
66% yield
(79% b.r.s.m.)
(2 steps)
N
˚
O
(4) Me3OBF4, CH2Cl2, 4 A molecular
HN
O
sieves, 45 uC; (5) NaCNBH3, MeOH,
0 uC; (6) NaH, DMF, 60 uC; (7) MeI,
K2CO3, acetone, 60 uC. b.r.s.m.,
based on recovered starting material;
d.r., diastereomeric ratio; Oxone,
potassium peroxymonosulphate.
H
H
Ent-Citrinalin B (ent-2)
Ent-citrinalin B•HCl (ent-2•HCl)
MeO
NO2
H
O
(6), (7)
N
H
N
78% yield
(2 steps)
O
Cyclopiamine B (6)
O2N
O
O
N
H
O
N
H
H
O
B
crystallographic analysis. Thus, the synthesis of ent-2 and its conver-
sion to 6 show that ent-2 is the true structure of citrinalin B, albeit the
enantiomer of the naturally occurring material.
a
24
4
24
O
O
8
Biosynthetic considerations
8
4
NO2
6
6
The total syntheses of ent-citrinalin B (ent-2; 19 steps from D-proline,
5.5%overallyield)andcyclopiamineB(6;21 stepsfromD-proline, 4.3%
overall yield) not only unambiguously establish the structures of these
metabolites, butalsoprovidepossibleinsightintothebiogenesisofthese
natural products (especially as to the possible formation of the cyclo-
piamines from the citrinalins).
23
23
21
21
N
H
N
12
12
1
1
NH
15
O
O
17
13
17
28
13
HN
15
HN
H
H
26
OH
O
26
O
27
27
O
17-hydroxycitrinalin B (37)
Citrinalin C (38)
b
OH
HO
The citrinalins, and in turn the cyclopiamines, probably arise from
a bicyclo[2.2.2]diazaoctane precursor. However, such a precursor was
unknown before the findings that are reported herein (see below). Con-
sistentwithnumerousbiosyntheticstudiesoftheprenylatedindolealka-
loids, the structural features of 1, 2, 4 and 6 suggest that tryptophan,
prolineandtwoisopreneunitsarebiosyntheticprecursorstothesecom-
pounds. Although no biosynthetic studies on 1 and 2 or 4 and 6 or the
related citrinadins and PF1270 alkaloids has appeared, it has been sug-
gested that they are derived from bicyclo[2.2.2]diazaoctane precursors
that suffer the ‘loss’ of one diketopiperazine carbonyl group5. Through
the isolation of 17-hydroxycitrinalin B (37; Fig. 6a) and, more impor-
tantly, citrinalin C (38) following a series of stable isotope labelling exper-
iments (summarized in Fig. 6b; seeSupplementary Information), wehave
now obtained support for the possible biogenesis of the citrinalins and
cyclopiamines from a precursor bearing the bicyclo[2.2.2]diazaoctane
moiety.
[U-13C]anthranilic acid
OH
OH
O
HO
HO2C
H2N
H2N
NH2
[1-13C]glucose
CO2H
[U-13C]ornithine
O
NO2
N
H
O
HN
H
R
O
Citrinalin A (1) (R = H, α-NO2)
Citrinalin B (2) (R = H, β-NO2)
17-hydroxycitrinalin B (37) (R = OH, β-NO2)
Figure 6
|
Isolation of two new citrinalins and 13C labelling studies.
a, Structures of 17-hydroxycitrinalin B and citrinalin C. Two additional
citrinalins, 37 and 38, were isolated on refractionation and reanalysis of
secondary metabolites from P. citrinum F53. b, Summary of the 13C labelling
studies. 13C incorporation studies of P. citrinum F53 reveal that glucose (pink),
anthranilic acid (blue) and ornithine (red) are biosynthetic precursors to
TheNMRandmassspectroscopycharacterizationdatafor37isfully
consistent with the assigned structure. Moreover, the assigned relative
configurationfullycorroboratestherevisedstructureofcitrinalin B(2). the citrinalins.
3 2 2
| N A T U R E | V O L 5 0 9 | 1 5 M A Y 2 0 1 4
Macmillan Publishers Limited. All rights reserved
©2014

ARTICLE RESEARCH
O
O
O
NO2
NO2
+H2O; –CO2
then [O]
N
?
N
H
N
H
NH
O
O
O
HN
HN
HN
(this work)
H
O
H
H
O
O
O
Citrinalin B (2)
Citrinalin A (1)
Citrinalin C (38)
Figure 7
|
Biosynthetic proposal for citrinalins. Consistent with previous
(38), which is followed by a decarboxylation event and amine-group oxidation
reports on the bicyclo[2.2.2]diazaoctane congeners, the citrinalins probably
arise through an intramolecular Diels–Alder reaction to form citrinalinC
to the nitro group.
Byanalogyto citrinalin B (2), the absolute configuration of37was assigned synthetic methods reported here to the syntheses of other prenylated
as 1S,14R,16R,17R,22R. 17-hydroxycitrinalin B (37) was initially iso- indole alkaloids is ongoing and will be reported in due course.
lated from P. citrinum F53 grown in a nitrogen-depleted culture med-
ium. Stable isotope feeding studies with [U-13C]anthranilic acid and
[1-13C]glucose gave significant 13C labelling (Supplementary Informa-
tion). High levels of [U-13C]ornithine were also incorporated into 37,
and additional feeding studies with [U-13C]proline gave almost unde-
tectable labelling. Ornithine is a well-known biosynthetic precursor to
proline, but to our knowledge it has never been reported as an efficient
substrate for isotopic labelling of the putative proline-derived atoms in
the biosynthesis of prenylated indole alkaloids of fungal origin bearing
the bicyclo[2.2.2]diazaoctane moiety. The labelling investigations sug-
gest that 17-hydroxycitrinalin B (37) might arise from either 3-hydroxyl
ornithine, 3-hydroxy proline or by the late-stage oxygenation of the
citrinalin A, B or C skeleton.
Citrinalin C (38), isolated as a minor component from the culture
medium of P. citrinum F53, gives NMR and mass spectroscopic data
(Supplementary Table 4) that is fully consistent with the relative and
absoluteconfigurationillustratedforthisnaturalproduct. Theisolation
of 38, along with the congeners lacking the bicyclo[2.2.2]diazaoctane
structural moiety from P. citrinum F53, lends support to a bicyclo[2.2.2]
diazaoctane-containingprecursor, whicharisesfromacommittedintra-
molecular Diels–Alder cycloaddition step such as that studied in detail
for other congeners33. Hydrolysis of the amide bridge of citrinalin C
(38; Fig. 7), followed by decarboxylation, and amino-group oxidation
to the nitro group, as proposed in the biosynthesis of the structurally
related citrinadin B5, would then yield citrinalin A. These latter steps
are the subject of current biosynthesis studies.
METHODS SUMMARY
All reactions were performed under a nitrogen atmosphere using dry solvents
under anhydrous conditions, unless otherwise noted. Dry tetrahydrofuran, tolu-
ene, methanol, triethylamine, benzene and diethyl ether were obtained by passing
thecommercially available, oxygen-freesolventsthroughactivated aluminacolumns
from GlassContour. Dichloromethane was distilled over calcium hydride under a
nitrogenatmosphere. Yieldsrefertomaterialspurifiedusingsilicagelcolumnchro-
matography. Full experimental details and characterization data for all new com-
pounds (1HNMR,13CNMR,massspectrometry,infrared,Rf value), including14–36,
2 and 6, appear in Supplementary Information. Crystallographic data were collected
on a MicroSTAR-H APEX II (ChexStar: RUA #1091) instrument, and the Bruker
SAINT and SADABS software programs were used for integrating and scaling the
data, respectively. The CYLVIEW program (developed by C. Y. Legault) was used
for X-ray depictions. Computational analyses were conducted following confor-
mational searches using the MMFF94 force field (SPARTAN’10). Density func-
tional theory calculations were performed with GAUSSIAN09 (B3LYP/6-311G(d,p)
theory level). Full details are included in Supplementary Information.
Received 29 January; accepted 21 March 2014.
1. Stocking, E. M., Sanz-Cervera, J. F. & Williams, R. M. Reverse versus normal prenyl
transferases in paraherquamide biosynthesis exhibit distinct facial selectivities.
Angew. Chem. Int. Ed. 38, 786–789 (1999).
2. Finefield, J. M., Frisvad, J. C., Sherman, D. H. & Williams, R. M. Fungal origins of the
bicyclo[2.2.2]diazaoctane ring system of prenylated indole alkaloids. J. Nat. Prod.
75, 812–833 (2012).
3. Miller, K. A. & Williams, R. M. Synthetic approaches to the
bicyclo[2.2.2]diazaoctane ring system common to the paraherquamides,
stephacidins and related prenylated indole alkaloids. Chem. Soc. Rev. 38,
3160–3174 (2009).
A question that remained at this stage concerned the biogenesis of
citrinalin B. On the basis of observations of the cyclopiamine series9 4. Tsuda, M. et al. Citrinadin A, a novel pentacyclic alkaloid from marine-derived
fungus Penicillium citrinum. Org. Lett. 6, 3087–3089 (2004).
(see 4 R 6 in Fig. 1), we anticipated that citrinalin A (1) might be con-
verted to citrinalin B (2) via a nitronate iminium intermediate analogous
5. Mugishima, T. et al. Absolute stereochemistry of citrinadins A and B from marine-
derived fungus. J. Org. Chem. 70, 9430–9435 (2005).
to 5. In the event, heating a solution of a naturally occurring sample of
citrinalin A (1) in DMF-d7 at 100 uC for 20 h leads to a 1:1 ratio of 1
and 2 (with complete conversion to citrinalin B (2) after 60 h; see Sup-
plementary Fig. 22), confirming the connection of these metabolites
presumably by the same aza-Henry or nitro-Mannich epimerization
sequenceestablishedforthecyclopiamines9. However, wehaveobserved
some key differences. First, the epimerization in the citrinalin series
occursataqualitativelylowerrate(probablyowingtoanon-productive
proton transfer from the vinylogous imide N–H to the tertiary amine)
and higher temperature. In addition, we have not been able to achieve
anyobservableconversion ofcitrinalin Btocitrinalin A evenatelevated
temperatures (165 uC) over prolonged periods (24 h). Our current efforts
are focused on gaining a deeper understanding of these differences and
exploring the biosynthetic conversion of citrinalin C to citrinalin A.
6. Kushida, N. et al. PF1270A, B and C, novel histamine H3 receptor ligands
produced by Penicillium waksmanii PF1270. J. Antibiot. (Tokyo) 60, 667–673
(2007).
7. Bian, Z., Marvin, C. C. & Martin, S. F. Enantioselective total synthesis of
(2)-citrinadinA and revisionof its stereochemicalstructure. J. Am. Chem. Soc. 135,
10886–10889 (2013).
8. Kong, K. et al. An enantioselective total synthesis and stereochemical revision of
(1)-citrinadin B. J. Am. Chem. Soc. 135, 10890–10893 (2013).
9. Bond, R. F., Boeyens, J. C. A., Holzapfel, C. W. & Steyn, P. S. Cyclopiamines A and B,
novel oxindole metabolites of Penicillium cyclopium westling. J. Chem. Soc. Perkin
Trans. I 1751–1761 (1979).
10. Pimenta, E. F. et al. Use of experimental design for the optimization of the
production of new secondary metabolites by two penicillium species. J. Nat. Prod.
73, 1821–1832 (2010).
11. Parry, R., Nishino, S. & Spain, J. Naturally-occurring nitro compounds. Nat. Prod.
Rep. 28, 152–167 (2011).
12. Lodewyk, M. W., Siebert, M. R. & Tantillo, D. J. Computational prediction of 1H and
13C chemical shifts: a useful tool for natural product, mechanistic and synthetic
organic chemistry. Chem. Rev. 112, 1839–1862 (2012).
Conclusion
13. Jewett, J. C. & Rawal, V. H. Total synthesis of pederin. Angew. Chem. Int. Ed. 46,
6502–6504 (2007).
We have reported the total syntheses of the prenylated indole alkaloids
ent-citrinalinBandcyclopiamine B. Ourresultsunambiguouslyidentify
citrinalin B throughsynthesis, a reanalysis of thenaturally isolated mate-
rial and an X-ray crystallographic study. Our studies on the isolation of
metabolites from P. citrinum suggest that a bicyclo[2.2.2]diazaoctane-
containing metabolitesuch as citrinalin C (38) isanintermediate inthe
biogenesis of citrinalins A (1) and B (2) (Fig. 7). The extension of the
14. Omura, K. & Swern, D. Oxidation of alcohols by ‘‘activated’’ dimethyl sulfoxide.
A preparative, steric and mechanistic study. Tetrahedron 34, 1651–1660 (1978).
15. Ohira, S. Methanolysis of dimethyl (1-diazo-2-oxopropyl) phosphonate:
generation of dimethyl (diazomethyl) phosphonate and reaction with carbonyl
compounds. Synth. Commun. 19, 561–564 (1989).
16. Grotjahn, D. B. & Lev, D. A. A general bifunctional catalyst for the anti-Markovnikov
hydration of terminal alkynes to aldehydes gives enzyme-like rate and selectivity
enhancements. J. Am. Chem. Soc. 126, 12232–12233 (2004).
1 5 M A Y 2 0 1 4
| V O L 5 0 9 | N A T U R E | 3 2 3
Macmillan Publishers Limited. All rights reserved
©2014

RESEARCH ARTICLE
17. Kishi, Y. et al. Synthetic approach towards tetrodotoxin. I. Diels-Alder reaction of
alpha-oximinoethylbenzoquinones with butadiene. Tetrahedr. Lett. 11,
5127–5128 (1970).
33. Ding, Y. et al. Genome-based characterization of two prenylation steps in the
assemblyofthe stephacidinand notoamideanticanceragentsina marine-derived
Aspergillus sp. J. Am. Chem. Soc. 132, 12733–12740 (2010).
18. Johnson, C. R. et al. Direct alpha-iodination of cycloalkenones. Tetrahedr. Lett. 33,
Supplementary Information is available in the online version of the paper.
917–918 (1992).
19. Ghaffar, T. & Parkins, A. W. A new homogeneous platinum containing catalyst for
the hydrolysis of nitriles. Tetrahedr. Lett. 36, 8657–8660 (1995).
20. Moriarty, R. M., Chany, C. J. II, Vaid, R. K., Prakash, O. & Tuladhar, S. M. Preparation
of methyl carbamates from primary alkyl- and arylcarboxamides using
hypervalent iodine. J. Org. Chem. 58, 2478–2482 (1993).
Acknowledgements R.S. and P.G.-R. thank the US National Institutes of Health (NIH;
NIGMS RO1 086374) for financial support. R.S. is a Camille Dreyfus Teacher Scholar.
E.V.M.-M. acknowledges the US National Science Foundation (NSF) for a graduate
fellowship (GRFP). S.R., E.F.P. and R.G.S.B. are grateful to the Brazilian National Council
of Technological and Scientific Development (CNPq; grant 470643/2010-2) and the
21. Herzon, S. B. & Myers, A. G. Enantioselective synthesis of stephacidin B. J. Am.
˜
Sao Paulo Research Foundation (FAPESP; grant 2012/50026-3) for funding. D.E.W.
Chem. Soc. 127, 5342–5344 (2005).
and R.J.A. thank NSEPC for funding. M.W.L. and D.J.T. acknowledge support from the
NSF (CHE-0957416 and supercomputing resources through a grant from the XSEDE
programme: CHE-030089). S.J.M. is grateful to the NIH for support (GM096403). We
thank A. DiPasquale for solving the crystal structures of ent-2
NHCl, 6, 27 and 36
(supported by NIH Shared Instrumentation Grant S10-RR027172). We would like to
thank T. Lebold, R. M. Williams and D. Sherman for discussions.
22. Myers, A. G. & Herzon, S. B. Identification of a novel Michael acceptor group for the
reversible addition of oxygen- and sulfur-based nucleophiles. Synthesis and
reactivity of the 3-alkylidene-3H-indole 1-oxide function of avrainvillamide. J. Am.
Chem. Soc. 125, 12080–12081 (2003).
23. Marti, C. & Carreira, E. Construction of spiro[pyrrolidine-3,39-oxindoles] 2 recent
applications to the synthesis of oxindole alkaloids. Eur. J. Org. Chem. 2209–2219
(2003).
Author Contributions R.S. conceived and directed the synthetic aspects of the
research, and wrote the majority of the manuscript (with input from all authors) except
the section on biosynthesis, which was contributed by S.R., E.F.P., D.E.W., R.J.A. and
R.G.S.B. The synthetic plan was designed by R.S. with input from E.V.M.-M. and P.G.-R.
who carried out the plan under the supervision of R.S. Oxidation catalyst 32 was
provided by D.K.R. and S.J.M., who, along with P.G.-R., E.V.M.-M. and R.S., designed the
oxidation studies of 25, which were done by P.G.-R. The computational NMR studies of
1, 2 and 3 were designed and performed by M.W.L. and D.J.T. with input from P.G.-R.,
E.V.M.-M. and R.S. Biosyntheticstudies weredesignedandconductedbyS.R., E.F.P. and
R.G.S.B., who also isolated and characterized 3, 37 and 38. D.E.W. and R.J.A. provided
facilities and contributed to the purification, data analysis and structural analysis of 3,
37 and 38.
24. Guerrero, C. A. & Sorensen, E. J. Concise, stereocontrolled synthesis of citrinadin B
core architecture. Org. Lett. 13, 5164–5167 (2011).
25. Grubbs, A. W., Artman, G. D. III, Tsukamoto, S. & Williams, R. M. A concise total
synthesis of the notoamides C and D. Angew. Chem. Int. Ed. 46, 2257–2261
(2007).
26. Davis, F. A. & Stringer, O. D. Chemistry of oxaziridines. 2. Improved synthesis of
2-sulfonyloxaziridines. J. Org. Chem. 47, 1774–1775 (1982).
27. Kolundzic, F. et al. Chemoselective and enantioselective oxidation of indoles
employing aspartyl peptide catalysts. J. Am. Chem. Soc. 133, 9104–9111 (2011).
28. Gu¨ller, R. & Borschberg, H.-J. A stereoselective transformation of pseudoindoxyls
into oxindoles in a single operation. Tetrahedr. Lett. 35, 865–868 (1994).
29. Movassaghi, M., Schmidt, M. A. & Ashenhurst, J. A. Stereoselective oxidative
rearrangements of 2-aryl tryptamine derivatives. Org. Lett. 10, 4009–4012
(2008).
Author Information A sample of the P. citrinum strain F53 is deposited at the Brazilian
Collection of Environmental and Industrial Microorganisms under the accession code
CBMAI 1186. Crystallographic data for crystal structures ent-2NHCl, 6, 27 and 36
have been deposited at the Cambridge Crystallographic Data Centre (http://
www.ccdc.cam.ac.uk) under accession codes CCDC984477, CCDC984478,
CCDC984480 and CCDC 984479, respectively. Reprints and permissions information
is available at www.nature.com/reprints. The authors declare no competing financial
interests. Readers are welcome to comment on the online version of the paper.
Correspondence and requests for materials should be addressed to
30. Miller, D. G. & Wayner, D. D. M. Improved method for the Wacker oxidation of cyclic
and internal olefins. J. Org. Chem. 55, 2924–2927 (1990).
31. Zhang, X. & Foote, C. S. Dimethyldioxirane oxidation of indole derivatives.
Formation of novel indole-2,3-epoxides and a versatile synthetic route to
indolinones and indolines. J. Am. Chem. Soc. 115, 8867–8868 (1993).
32. Borch, R. F. A new method for the reduction of secondary and tertiary amides.
Tetrahedr. Lett. 9, 61–65 (1968).
R.G.S.B. (rgsberlinck@iqsc.usp.br) or R.S. (rsarpong@berkeley.edu).
3 2 4
| N A T U R E | V O L 5 0 9 | 1 5 M A Y 2 0 1 4
Macmillan Publishers Limited. All rights reserved
©2014