Concise total syntheses of variolin B and deoxyvariolin B

Article abstract of DOI:10.1021/jo050523v

The total synthesis of the marine alkaloid variolin B has been achieved in 8 steps and 17% overall yield, starting from commercially available 4-chloro-2-methylthiopyrimidine. The key reaction involves the tandem deoxygenation and cyclization of a triarylmethanol using a combination of triethylsilane and trifluoroacetic acid. In addition, the deoxygenated analogue was prepared in 6 steps and 23% overall yield, starting from the same starting material.

Full text of DOI:10.1021/jo050523v

Concise Total Syntheses of Variolin B and Deoxyvariolin B
Regan J. Anderson,† Jonathan B. Hill,† and Jonathan C. Morris*,‡
Department of Chemistry, University of Canterbury, Christchurch, New Zealand, and School of Chemistry
and Physics, University of Adelaide, Adelaide, SA, Australia
jonathan.morris@adelaide.edu.au
Received March 15, 2005
The total synthesis of the marine alkaloid variolin B has been achieved in 8 steps and 17% overall
yield, starting from commercially available 4-chloro-2-methylthiopyrimidine. The key reaction
involves the tandem deoxygenation and cyclization of a triarylmethanol using a combination of
triethylsilane and trifluoroacetic acid. In addition, the deoxygenated analogue was prepared in 6
steps and 23% overall yield, starting from the same starting material.
Introduction
kaemia) cell line (IC50 ) 210 ng/mL), as well as moderate
antiviral activity. Further investigations into variolin B’s
biological activity have been hampered by a lack of
material. This has led to considerable interest in the
synthesis of variolin B, with several groups reporting
syntheses of the variolin core and/or fragments of the
core.4,5 In 2001, we reported6 the first total synthesis of
variolin B. Since then, two other total syntheses have
been completed.7 The majority of the synthetic ap-
proaches build up the tricyclic core of variolin B in a
linear fashion from an heteroaryl component using
conventional heterocyclic chemistry.4,5a,5b,5d In contrast,
we have employed a highly convergent approach based
on the recognition of a hidden symmetry element in
Marine organisms have provided a vast array of
structurally diverse and biologically interesting struc-
tures.1 However, to date, few marine natural products
have become useful pharmaceutical drugs. One of the
major reasons for this is the difficulty in obtaining
adequate quantities of the marine organism, as secondary
metabolites from marine organisms are usually produced
in very low quantities and, therefore, large amounts of
the organism are required.2 Harvesting large quantities
of marine organisms is difficult and expensive, and can
also have a major ecological impact. As a consequence,
chemical synthesis is playing an increasingly important
role in providing supplies of marine natural products.
A recent example of this is variolin B (1), a novel
marine alkaloid which was isolated from the rare, and
difficult-to-access, Antarctic sponge Kirkpatrickia varia-
losa by the group of Blunt and Munro.3 Preliminary
investigations established that variolin B (1) displayed
strong in vitro activity against the P388 (murine leu-
(4) Syntheses of fragments of the variolin core structure: (a) AÄ lvarez,
M.; Ferna´ndez, D.; Joule, J. A. J. Chem. Soc., Perkin Trans. 1 1999,
249. (b) AÄ lvarez, M.; Ferna´ndez, D.; Joule, J. A. Synthesis 1999, 615.
(c) Minguez, J. M.; Vaquero, J. J.; Alvarez-Builla, J.; Castan˜o, O.;
Andre´, J. L. J. Org. Chem. 1999, 64, 7788. (d) Fresneda, P. M.; Molina,
P.; Bleda, J. A. Tetrahedron 2001, 57, 2355. (e) AÄ lvarez, M.; Ferna´ndez,
D.; Joule, J. A. J. Chem. Soc., Perkin Trans. 1 2002, 471.
(5) Syntheses of the variolin core structure (a) Fresneda, P. M.;
Molina, P.; Delgado, S.; Bleda, J. A. Tetrahedron Lett. 2000, 41, 4777.
(b) Mendiola, J.; Minguez, J. M.; Alvarez-Builla, J.; Vaquero, J. J. Org.
Lett. 2000, 2, 3253. (c) Anderson, R. J.; Morris, J. C. Tetrahedron Lett.
2001, 42, 311. (d) AÄ lvarez, M.; Ferna´ndez, D.; Joule, J. A. Tetrahedron
Lett. 2001, 42, 315. (e) Mendiola, J.; Baeza, A.; Alvarez-Builla, J.;
Vaquero, J. J. J. Org. Chem. 2004, 69, 4974. (f) Fernandez, D.; Ahaidar,
A.; Danelon, G.; Cironi, P.; Marfil, M.; Perez, O.; Cuevas, C.; Albericio,
F.; Joule, J. A.; Alvarez, M. Monatsh. Chem. 2004, 135, 615. (g)
Ahaidar, A.; Fernaindez, D.; Perez, O.; Cuevas, C.; Albericio, F.; Joule,
J. A.; Alvarez, M. ARKIVOC (Gainesville, FL, United States), 2004,
iv, 74.
† University of Canterbury.
‡ University of Adelaide.
(1) Blunt, J. W.; Copp, B. R.; Munro, M. H. G.; Northcote, P. T.;
Prinsep, M. R. Nat. Prod. Rep. 2004, 21, 1.
(2) Urban, S.; Hickford, S. J. H.; Blunt, J. W.; Munro, M. H. G. Curr.
Org. Chem. 2000, 4, 765.
(3) (a) Perry, N. B.; Ettouati, L.; Litaudon, M.; Blunt, J. W.; Munro,
M. H. G. Tetrahedron 1994, 50, 3987. (b) Trimurtulu, G.; Faulkner,
D. J.; Perry, N. B.; Ettouati, L.; Litaudon, M.; Blunt, J. W.; Munro, M.
H. G.; Jameson, G. B. Tetrahedron 1994, 50, 3993.
(6) Anderson, R. J.; Morris, J. C. Tetrahedron Lett. 2001, 42, 8697.
10.1021/jo050523v CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/15/2005
6204
J. Org. Chem. 2005, 70, 6204-6212

Syntheses of Variolin B and Deoxyvariolin B
SCHEME 1
SCHEME 2
variolin B. In this paper, we present full details of our
synthetic strategy and its application to the total syn-
thesis of variolin B and the nonnatural analogue, deoxy-
variolin B.
Results and Discussion
1310 with n-butyllithium at -95 °C as detailed by
Undheim and co-workers.9 The lithio species is extremely
reactive, and we observed that warming above -95 °C
led to decomposition of the reagent. However, when the
bath temperature was maintained at -95 °C, 5 reacted
with acid chloride 11 to form triarylmethanol 10 in an
optimized yield of 56% (Scheme 2, left-hand side). A
major contaminant is the biaryl alcohol 14, which is not
easily separated by chromatography. Therefore, the
mixture was reacted with manganese dioxide, which
allowed the contaminant to be readily removed. Three
equivalents was required for complete consumption of the
acid chloride, and slow addition of pre-cooled (-95 °C)
reagents11 was vital to minimize decomposition of the
lithio species.12
Our retrosynthetic strategy for variolin B (1) is outlined
in Scheme 1, with the key compound being the core
structure 2, which contains the appropriate functionality
to introduce the highly polar amino groups.8 Examination
of 2 reveals a hidden symmetry element, which leads to
the disconnection of the N10-C10a bond indicated in
Scheme 1. Thus, it was anticipated that cyclization of the
symmetrical triarylmethane 3 would generate the desired
core structure with appropriate functionality to allow
elaboration to variolin B. Triarylmethane 3 should be
obtained by deoxygenation of the alcohol 4, which, in
turn, could be rapidly constructed by reaction of the
known9 lithio species 5 (2 equiv) with acid chloride 6.
When this work was initiated, there were no reports on
the preparation of the tricyclic core of the variolins, so it
was decided to start with the commercially available acid
chloride 11, which would lead to the deoxy-analogue 7.
This would establish the feasibility of our approach and,
if successful, would provide insights into the structure-
activity relationships of variolin B.
Deoxygenation of triarylmethanol 10 was attempted
using a variety of methods. Deoxygenation via the Barton
ester13 was complicated by difficulties in obtaining the
required xanthate cleanly. Numerous methods were
screened for the direct conversion of the tertiary alcohol
10 to the triarylmethane 9. These included NaCNBH3/
Synthesis of the Deoxyvariolin Core 8. Lithio
species 5 was generated by treatment of iodopyrimidine
(10) Iodide 13 is readily prepared by stirring chloride 12 in 55%
aqueous HI solution. See the Supporting Information and the refer-
ences in footnote 9 for further details.
(7) (a) Molina, P.; Fresneda, P. M.; Delgado, S.; Bleda, J. A.
Tetrahedron Lett. 2002, 43, 1005. (b) Molina, P.; Fresneda, P. M.;
Delgado, S. J. Org. Chem. 2003, 68, 489. (c) Ahaidar, A.; Ferna´ndez,
D.; Perez, O.; Danelon, G.; Cuevas, C.; Manzanares, I.; Albericio, F.;
Joule, J. A.; AÄ lvarez, M. Tetrahedron Lett. 2003, 44, 6191 (d) Ahaidar,
A.; Ferna´ndez, D.; Danelon, G.; Cuevas, C.; Manzanares, I.; Albericio,
F.; Joule, J. A.; AÄ lvarez, M. J. Org. Chem. 2003, 68, 10020.
(8) Converting a thiomethyl group into a amino group is a common
procedure. For a recent example, see: Adlington, R. M.; Baldwin, J.
E.; Catterick, D.; Pritchard, G. J. J. Chem. Soc., Perkin Trans. 1 1999,
855.
(11) It was found that when a specialized piece of glassware first
described by Suzuki and Noyori was used, efficient control over the
internal temperature could be achieved. Addition of the n-BuLi through
the spiral port ensures that the n-BuLi solution is efficiently pre-cooled
to -95 °C before it reacts with the iodide. See: Suzuki, M.; Noyori, R.
In Organocopper Reagents: A Practical Approach; Taylor, R. J. K., Ed.;
Oxford University Press: New York, 1994; pp 201-206.
(12) This reaction can also be carried out more simply using Barbier-
type conditionssn-butyllithium was added dropwise to a mixture of
iodide 13 and acid chloride 11 in THF at -95 °C. See the Experimental
Section for full details.
(9) Lithio species 5 is generated by iodine-lithium exchange of
4-iodo-2-methylthiopyrimidine (13) using n-BuLi in THF at -95 °C.
See ref 5c and Majeed, A. J.; Antonsen, Ø.; Benneche, T.; Undheim,
K. Tetrahedron 1989, 45, 993.
(13) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans.
1 1975, 1574.
J. Org. Chem, Vol. 70, No. 16, 2005 6205

Anderson et al.
TABLE 1. Optimization of the Reaction Conditions for
the Synthesis of 8 from 10
access to the tricyclic core of variolin B in just three steps
from commercially available materials.
entry
equiv of TES
equiv of TFA
yield of 8 (%)
Synthesis of the Variolin Core 2. Our attention
turned to the application of this methodology to the
synthesis of the oxygen-containing core structure 2.
Using the conditions established for the synthesis of
triarylmethanol 10, the double-addition of lithio species
5 to acid chloride 622 was examined (Scheme 4). Unfor-
tunately, triarylmethanol 4 was isolated in only 14%
yield. Prolonged reaction at -95 °C offered no improve-
ment to the yield; it appeared that decomposition of the
lithio species predominated over addition to the hindered
carbonyl group of 6. As the equivalent Grignard reagent
is reported to be stable at 0 °C,23 we hoped that it might
react more readily with the acid chloride 6. However,
despite attempting this reaction under a variety of
conditions, 4 was never observed in the product mixture.
As the instability of 5 was hampering our efforts at
producing 4, it was decided to prepare 4 by the reaction
of the 3-lithiopyridine derived from 21 with the sym-
metrical ketone 20 as detailed in Scheme 4. The sym-
metrical ketone 20 was readily available by reaction of
lithio species 5 with 0.5 equiv of diethyl carbonate at -95
°C. The 3-lithiopyridine had previously been prepared by
reaction of 2-chloro-4-methoxypyridine (21)24 with either
LDA or LTMP at -78 °C, but significant amounts of
starting material were recovered in both cases.23b,c Comins
and LaMunyon have shown that LDA is too weak a base
for the complete deprotonation of 2-, 3-, and 4-methoxy-
pyridines.25 Consequently, we decided to examine the
lithiation of 21 with butyllithium reagents, even though
it has been reported that attempts to use the stronger
alkyllithium bases (n-, s- or t-BuLi) failed due to compet-
ing substitution or halogen-metal exchange.23c While the
use of t-BuLi at -78 °C resulted in nucleophilic attack
at the pyridine ring, the use of n-BuLi resulted in an
efficient lithiation at -95 °C.26 Reaction of the generated
lithio species with the ketone 20 at -78 °C resulted in
the desired triarylmethanol 4 being obtained in 76%
yield, constituting a significant improvement on the
former approach.
1
2
3
4
5
2.0
4.0
8.0
8.1
7.8
8.3
8.2
8.2
4.3
2.0
15
18
22
34
33
ZnI2,14 Pd-C/AlCl3/cyclohexene,15 P/I2/HI,16 H2/Pd-C,
LiAlH4/ AlCl3,17 NaBH4/TFA,18 and H2/Adams’ catalyst/
TFA19sall of which resulted in either no reaction or the
formation of complex mixtures. It appeared that the
π-deficient rings of the pyrimidines were undergoing
reduction in preference to hydrogenolysis of the C-O
bond. This impasse was finally resolved by using a
combination of triethylsilane (TES) with trifluoroacetic
acid (TFA) (Scheme 2 and Table 1).20
An initial attempt using 1.2 equiv of TES in refluxing
TFA produced a mixture of several products. While the
desired triarylmethane 9 was not among these (as H
1
NMR spectroscopy clearly showed the presence of two
nonidentical pyrimidine rings in each product), we were
pleased to find that one of the products was, in fact, the
variolin core structure 8, albeit in ∼5% yield. The other
products isolated from the TES/TFA reaction were iden-
tified as 15, 16 and 17 (Scheme 3). These products had
not been deoxygenated, suggesting that they were the
result of acid-promoted reactions. Indeed, when triaryl-
methanol 10 was refluxed in neat TFA, the compounds
15, 16, and 17 were formed in the ratio of 14:1:4.
While one might conclude that upon deoxygenation of
10, the triarylmethane 9 cyclizes under the reaction
conditions to afford the tricyclic core 8, the presence of
compounds 15 and 16 led us to consider an alternate
mechanism whereby cyclization of 10 might precede
deoxygenation (Scheme 3). In this process, the initially
formed species is 18, which would be a common inter-
mediate for all three cyclized products 8, 15 and 16.
Reduction of 18 with TES, followed by elimination of
water would afford the desired product 8, while acid-
catalyzed rearrangements would lead to the formation
of 15 and 16 as detailed in Scheme 3. Compound 17,
which still contains a chlorine atom, cannot arise from
18 but presumably forms by rearrangement of 10 in a
process similar to that shown for 16.21
Reaction of 4 with TES/TFA, under the optimized
conditions used on 10, resulted in the formation of
tricyclic core 2 in a disappointing yield of 9% yield
(Scheme 5 and Table 2, entry 1). The major product of
the reaction was compound 22, isolated in 54% yield.
Careful optimization of the reaction conditions allowed
We reasoned, therefore, that decreasing the amount
of TFA and increasing the amount of TES in the reaction
should minimize the formation of side-products 15-17
and lead to greater yields of the desired variolin core 8.
Indeed, this was the case, with 8 being isolated in 34%
yield under optimized conditions (see Table 1). While this
yield was modest, it should be noted that we have gained
(22) Acid chloride 6 was generated in 97% yield from 2-chloro-4-
methoxynicotinic acid (19) by reaction with oxalyl chloride. Acid 19
was obtained in 4 steps from commercially available materials by
modifying a synthesis of the bromo-analogue described by Kasum, B.;
Prager, R. H. Aust. J. Chem. 1983, 36, 1455. See the Supporting
Information for more details.
(23) Lepreˆtre, A.; Turck, A.; Ple´, N.; Knochel, P.; Que´guiner, G.
Tetrahedron 2000, 56, 265.
(14) Lau, C. K.; Dufresne, C.; Be´langer, P. C.; Pie´tre´, S.; Scheigetz,
J. J. Org. Chem. 1986, 51, 3038.
(24) 2-Chloro-4-methoxypyridine (21) was produced in three steps
and 48% overall yield from 4-nitropyridine-N-oxide. Details are
provided in the Supporting Information. See also (a) Talik, Z. Roczniki
Chem. 1962, 36, 1313. Chem. Abstr. 1963, 59, 6358b (b) Walters, M.
A.; Shay, J. Tetrahedron Lett. 1995, 36, 7575. and (c) Connon, S. J.;
Hegarty, A. F. Tetrahedron Lett. 2001, 42, 735.
(15) Olah, G. A.; Prakash, G. K. S. Synthesis 1978, 397.
(16) Lee, W. Y.; Park, C. H.; Kim, Y. D. J. Org. Chem. 1992, 57,
4074.
(17) Brewster, J. H.; Osman, S. F.; Bayer, H. O.; Hopps, H. B. J.
Org. Chem. 1964, 29, 121.
(25) Comins, D. L.; LaMunyon, D. H. Tetrahedron Lett. 1988, 29,
773.
(26) Quenching with MeOD after 1 h of reaction at -78 °C gave the
C3-deuterated pyridine in ∼84% yield (judging by 1H NMR spectros-
copy), together with some starting material (∼7%) and a small amount
of addition product (∼9%). When the lithiation was repeated at -95
°C, the addition product was virtually eliminated.
(18) Gribble, G. W.; Leese, R. M. Synthesis 1977, 172.
(19) Peterson, P. E.; Casey, C. J. Org. Chem. 1964, 29, 2325.
(20) For a review of ionic hydrogenation see Kursanov, D. N.; Parnes,
Z. N.; Loim, N. M. Synthesis 1974, 633.
(21) It is interesting to note that 17 could not be converted into 16.
Exposure of 17 to TFA at reflux did not lead to cyclization.
6206 J. Org. Chem., Vol. 70, No. 16, 2005

Syntheses of Variolin B and Deoxyvariolin B
SCHEME 3
SCHEME 4
SCHEME 5
the yield of 2 to be improved to 47% (entry 2). Interest-
ingly, no products analogous to 15 and 16 were observed
in this reaction. Ether 22 is presumably formed in a
manner similar to the deoxy-equivalent 17 (Scheme 3)
whereby the hydroxyl group attacks one of the pyrimidine
rings and the C-C bond breaks to relieve the strain of
the resulting epoxide. It was reasoned that if this side
reaction could be blocked, higher yields of 2 could be
obtained. To test this hypothesis, it was decided to protect
the hydroxyl group as the acetate 23. Acetylation of 4
proved to be a slow reaction, with poor conversion.
However, if the reaction to form 4 was quenched with
acetyl chloride, acetate 23 was generated in 78% yield.
When 23 was treated with TES and TFA under es-
sentially the same conditions that were used for alcohol
4, the variolin core 2 was formed as the sole product in
75% yield, without any trace of 22 (Entry 3). This
modification was also applicable in the deoxy-series, with
the cyclization of the acetate 2427 affording 8 in 69% yield,
with no trace of the side-products 15-17 (Entry 4).
Synthesis of Variolin B (1) and Deoxyvariolin B
(7). Having developed a concise route to the variolin core,
deprotection of the methoxy group of 2 and substitution
of the thiomethyl groups with ammonia would complete
the synthesis of variolin B (Scheme 6). The use of boron
tribromide failed to remove the methoxy protecting
group, producing complexes that precipitated out of
solution. Heating 2 with sodium ethanethiolate in DMF
(27) The acetate 24 was synthesized using the Barbier-type reaction,
followed by an acetyl chloride quench. Despite the crude material
looking clean by NMR spectroscopy, the purified acetate 24 was
obtained in only 50% yield. However, if the crude reaction mixture
was submitted directly to the cyclization conditions, 8 could be obtained
in 50% overall yield (based on acid chloride 11).
J. Org. Chem, Vol. 70, No. 16, 2005 6207

Anderson et al.
TABLE 2. Synthesis of Core Structures 2 and 8 Using TES/TFA
% yield
entry
substrate
conditions
2/8
22
1
2
3
4
4
4
23
24
8 equiv TES/4 equiv TFA/70°C
9
47
75
69
54
28
0
8 equiv TES/2 equiv TFA/ClCH2CH2Cl/100°C
8 equiv TES/2 equiv TFA/ClCH2CH2Cl/100°C
8 equiv TES/2 equiv TFA/1ClCH2CH2Cl/100°C
n/a
SCHEME 6
acetate salt. The free base was obtained by neutralization
of the salt, with concentrated aqueous ammonia solution,
in 80% yield from 27. Synthetic variolin B was compared
to a natural sample provided by Blunt and Munro and
exhibited identical spectroscopic properties to those
reported for the natural material.3 It was equally potent
in the P388 assay as the natural product.
The deoxygenated analogue 7 was readily generated
from core structure 8 using a similar sequence to that
used for variolin B (Scheme 6).29 Of note is that this
analogue was found to have essentially the same activity
in the P388 assay (IC50 ) 157 ng/mL) as variolin B,
which indicates that the hydroxyl group is not essential
to the biological activity of variolin B.
In summary, variolin B has been produced in 8 steps
and 17% overall yield, starting from the commercially
available pyrimidine 12. Highlights of our synthesis
include the tandem deoxygenation/cyclization to form the
core variolin skeleton and the straightforward functional
group manipulation required to introduce the necessary
functionality of variolin B. The analogue deoxyvariolin
B (7) was prepared in 6 steps, in 23% overall yield, and
was found to be as biologically active as the natural
product. This synthetic strategy is readily amenable to
the production of analogues and will provide material to
allow further investigation of the biological properties.
at 95 °C did result in deprotection, but partial displace-
ment of the thiomethyl groups by sodium ethanethiolate
had also occurred. It was decided, therefore, to carry out
the methoxy deprotection after the amino groups had
been introduced.
Experimental Section30
(2-Chloropyridin-3-yl)bis-[2-(methylsulfanyl)pyrimidin-
4-yl]methanol (10). The reaction vessel11 was charged with
iodide 13 (5.00 g, 19.8 mmol) and freshly distilled THF (270
mL), and cooled to -95 °C (MeOH/liquid N2). A solution of
n-BuLi in hexanes (1.57 M, 12.6 mL, 19.8 mmol) was carefully
added to the stirred solution over 20 min, while maintaining
the temperature at -95 °C. After 20 min, a solution of acid
chloride 11 (1.17 g, 6.67 mmol) in THF (10 mL) was added to
the dark mixture over 5 min. After 3.5 h at -95 °C, the
reaction was quenched with methanol (1 mL) and allowed to
warm overnight. Excess THF was removed in vacuo, and the
residue was taken up in water/EtOAc. After separation of the
layers, the aqueous layer was re-extracted with EtOAc (x2).
The organic layers were combined and washed with saturated
brine solution. After drying with MgSO4 and concentration,
the crude material was purified by flash chromatography on
silica, eluting with 25-50% EtOAc/hexanes, to afford triaryl-
methanol 10 (1.24 g, ∼93% pure by 1H NMR spectroscopy) in
addition to a number of other side products, one of which was
the secondary alcohol 14. Fractions which contained mixtures
of 10 and 14 were subjected to oxidation by MnO2 (2 g) in
refluxing benzene (12 mL). Filtration through Celite (rinsing
with boiling EtOAc) and concentration in vacuo gave a
mixture, which was readily separated by chromatography on
silica, eluting with 25-50% EtOAc/hexanes, to yield a further
343 mg of 10 (∼95% pure by 1H NMR spectroscopy). Thus, 10
Oxidation of 2 with 2 equiv of MCPBA at -40 °C gave
predominantly the bis-sulfoxide 25 as a 1:1 mixture of
diastereomers. While substitution of the sulfoxides with
ammonia would take us one step closer to the target, we
felt the use of a protected amine would give a more easily
handled product. Thus, heating the crude oxidized mix-
ture in neat p-methoxybenzylamine resulted in clean
substitution of both sulfoxide groups to afford bis-amine
26, in 78% yield for the two steps. Removal of the
methoxy group was now revisited and treatment with an
excess of sodium ethanethiolate in DMF at 55 °C cleanly
produced the pyridinol 27 in 95% yield, without affecting
the p-methoxybenzyl groups.
DDQ oxidation failed to effect the deprotection of the
p-methoxybenzyl groups, while the use of TFA28 led to
removal of only the p-methoxybenzyl group on the
pendant pyrimidine ring. However, standing 27 in neat
triflic acid at room-temperature resulted in the removal
of both protecting groups. After neutralization, the crude
material was purified by flash chromatography on re-
versed-phase silica to give variolin B (1) as its trifluoro-
(29) It has also been prepared by AÄ lvarez and Joule’s group, using
7-azaindole as starting material. See refs 5d and 7d for details.
(30) General procedures are detailed in the Supporting Information.
(28) Smith, A. B.; Rano, T. A.; Chida, N.; Sulikowski, G. A.; Wood,
J. L. J. Am. Chem. Soc. 1992, 114, 8008.
6208 J. Org. Chem., Vol. 70, No. 16, 2005

Syntheses of Variolin B and Deoxyvariolin B
was obtained as an orange foam (1.58 g, ∼93% pure, 56%
yield). An analytical sample was obtained by flash chroma-
tography on reversed phase silica, eluting with 75% MeOH/
water, to give an off-white solid. Mp: 54-58 °C. 1H NMR (300
MHz, CDCl3): δ 2.49 (s, 6H, SCH3), 6.33 (s, 1H, OH), 7.17 (dd,
J ) 4.9, 7.8 Hz, 1H, H-5), 7.23 (dd, J ) 2.0, 7.8 Hz, 1H, H-4),
7.40 (d, J ) 5.1 Hz, 2H, H-5′), 8.38 (dd, J ) 2.0, 4.9 Hz, 1H,
H-6), 8.56 (d, J ) 5.1 Hz, 2H, H-6′). 13C NMR (75 MHz,
CDCl3): δ 14.2 (SCH3), 79.4 (Ar3COH), 114.8 (C-5′), 121.7 (C-
5), 136.8 (C-3), 138.9 (C-4), 149.1 (C-6), 151.5 (C-2), 158.0 (C-
6′), 168.8 (C-4′), 171.8 (C-2′). IR (CDCl3): υmax 3381, 1549. MS
(EI): m/z (rel intensity) 391/393 (59/25), 356 (100), 266/268
(71/29),125(68),45(53).HRMS(EI): CalcdforC16H1435ClN5O32S2
(M+) 391.0328, found 391.0332.
addition to several unidentified products. Repeated flash
chromatography on silica, eluting with 37-70% EtOAc/ hex-
anes, gave the variolin core 8 as a yellow solid (143 mg, 34%).
1
Mp: 210-211 °C. H NMR (500 MHz, CDCl3): δ 2.68 (s, 3H,
2′-SCH3), 2.74 (s, 3H, 9-SCH3), 7.36 (d, J ) 5.1 Hz, 1H, H-5′),
7.52 (dd, J ) 4.5, 8.2 Hz, 1H, H-3), 7.83 (d, J ) 6.8 Hz, 1H,
H-7), 8.07 (d, J ) 6.8 Hz, 1H, H-6), 8.52 (d, J ) 5.1 Hz, 1H,
H-6′), 8.61 (dd, J ) 1.5, 4.5 Hz, 1H, H-2), 8.65 (dd, J ) 1.5, 8.2
Hz, 1H, H-4). 13C NMR (75 MHz, CDCl3): δ 14.3 (2′-SCH3),
14.9 (9-SCH3), 101.4 (C-5), 108.3 (C-6), 112.9 (C-5′), 120.7 (C-
4a), 120.8 (C-3), 128.0 (C-4), 137.5 (C-5a), 139.9 (C-7), 141.8
(C-2), 143.0 (C-10a), 155.1 (C-9), 156.6 (C-6′), 160.9 (C-4′), 172.4
(C-2′). MS (EI): m/z (rel intensity) 339 (100), 306 (10), 220
(14), 133 (12). HRMS (EI): Calcd for C16H13N532S2 (M+)
339.0612, found 339.0617.
Synthesis of 10 via Barbier-Type Conditions. Iodide 13
(1.576 g, 6.3 mmol) was dissolved in THF (17 mL) and
transferred via cannula to a pre-cooled solution of the acid
chloride 11 (0.500 g, 2.8 mmol) in THF (8 mL) at -96 °C. A
solution of n-BuLi in hexanes (1.7 M, 3.68 mL, 6.3 mmol) was
then added dropwise over 10 min with vigorous stirring. The
dark red/brown solution was stirred at -96 °C for a further
2h; then methanol (400 µL) was added over 1 min. The deep
red solution was stirred at room temperature for 2 h, then
diluted with CH2Cl2 (75 mL), washed with sat. aq. NaHCO3
solution (100 mL), brine (100 mL), dried with Na2SO4, filtered,
and concentrated in vacuo. This crude material was dissolved
in benzene (15 mL), MnO2 (1 g) added, and the black suspen-
sion was stirred at reflux for 2 h. This was then filtered
through a Celite plug and concentrated in vacuo. The red/
brown residues were purified by flash chromatography on
silica, eluting with 50% EtOAc/hexanes, to give the triaryl-
methanol 10 as an orange solid (0.521 g, ∼95% pure, 40%
yield).
Other compounds isolated from this reaction are as follows:
9-(Methylsulfanyl)-5-[2-(methylsulfanyl)pyrimidin-4-
yloxy]pyrido[3′,2′:4,5] pyrrolo[1,2-c]pyrimidine (16). This
could be further purified by reversed phase preparative HPLC,
eluting with 85% CH3CN/water + 0.05% TFA, to give a yellow
solid. Mp: 205-207 °C, sweats at 203 °C. 1H NMR (300 MHz,
CDCl3): δ 2.21 (s, 3H, 2′-SCH3), 2.72 (s, 3H, 9-SCH3), 6.62 (d,
J ) 5.4 Hz, 1H, H-5′), 6.96 (d, J ) 6.6 Hz, 1H, H-6), 7.39 (dd,
J ) 4.8, 8.2 Hz, 1H, H-3), 7.49 (d, J ) 6.6 Hz, 1H, H-7), 7.85
(dd, J ) 1.5, 8.2 Hz, 1H, H-4), 8.39 (d, J ) 5.4 Hz, 1H, H-6′),
8.56 (dd, J ) 1.5, 4.8 Hz, 1H, H-2). 13C NMR (75 MHz,
CDCl3): δ 13.9 (2′ SCH3), 14.6 (9-SCH3), 102.2 (C-5′), 104.8
(C-6), 115.6 (C-4a), 119.0 (C-5), 119.7 (C-3), 125.4 (C-4), 126.2
(C-5a), 135.4 (C-7), 139.0 (C-10a), 141.5 (C-2), 153.5 (C-9),
158.8 (C-6′), 168.5 (C-4′), 173.2 (C-2′). MS (EI): m/z (rel
intensity) 355 (94), 308 (17), 230 (100), 143 (19), 113 (18), 73
(15). HRMS (EI): Calcd for C16H13N5O32S2 (M+) 355.0562,
found 355.0559.
(2-Chloropyridin-3-yl)bis-[2-(methylsulfanyl)pyrimidin-
4-yl]methyl acetate (24). The reaction vessel11 was charged
with a solution of iodide 13 (3.296 g, 13.1 mmol) in freshly
distilled THF (40 mL), and cooled to -95 °C (MeOH/liquid N2).
A solution of acid chloride 11 (1.043 g, 5.9 mmol) in THF (10
mL) was transferred to the reaction vessel via cannula. A
solution of n-BuLi in hexanes (1.6 M, 8.17 mL, 13.1 mmol)
was then added dropwise over 20 min with vigorous stirring.
The dark red/brown solution was stirred at -96 °C for a
further 2h, then acetyl chloride (2.11 mL, 29.6 mmol) was
added over 5 min. The deep red solution was stirred at room
temperature for 16 h, then diluted with CH2Cl2 (100 mL),
washed with sat. aq. NaHCO3 solution (200 mL), brine (150
mL), dried with Na2SO4, filtered, and concentrated in vacuo.
This crude material was purified by repeated flash chroma-
tography on silica, eluting with 1:1 EtOAc/hexanes, to give the
acetate 24 as an orange solid (1.300 g, 98% pure, 50% yield).
Mp: 58-61 °C. 1H NMR (500 MHz, CDCl3): δ 2.30 (s, 3H,
COCH3), 2.39 (s, 6H, SCH3), 7.26 (dd, J ) 4.9, 7.8 Hz, 1H,
H-5), 7.30 (d, J ) 4.9 Hz, 2H, H-5′), 7.89 (dd, J ) 1.7, 8.1 Hz,
1H, H-4), 8.39 (dd, J ) 2.0, 4.9 Hz, 1H, H-6), 8.51 (d, J ) 4.8
Hz, 2H, H-6′). 13C NMR (75 MHz, CDCl3): δ 14.0 (SCH3), 21.3
(COCH3), 85.9 (C(Ar)3OAc), 114.7 (C-5′), 121.1 (C-5), 134.0 (C-
3), 141.3 (C-4), 149.2 (C-6), 150.3 (C-2), 157.4 (C-6′), 166.1 (C-
4′), 168.6 (COCH3), 172.6 (C-2′). IR (KBr): υmax 1757, 1549,
1400, 1348, 1207, 1086, 1061, 903. HRMS (ESI): Calcd for
C18H1735ClN5O232S2 (MH+) 434.0512, found 434.0512.
(2-Chloropyridin-3-yl)-[2-(methylsulfanyl)pyrimidin-
4-yl]-[2-(methylsulfan yl)pyrimidin-4-yloxy]methane (17).
This could be further purified by flash chromatography on
reversed phase silica, eluting with 80% MeOH/water, followed
by flash chromatography on silica, eluting with 50% EtOAc/
1
hexanes, to give a viscous gum. H NMR (300 MHz, CDCl3):
δ 2.43 (s, 3H, 2′′-SCH3), 2.44 (s, 3H, 2′-SCH3), 6.58 (d, J ) 5.9
Hz, 1H, H-5′′), 7.19 (d, J ) 4.9 Hz, 1H, H-5′), 7.27 (dd, J )
4.8, 7.6 Hz, 1H, H-5), 7.43 (s, 1H, Ar2CHOAr), 7.84 (dd, J )
1.8, 7.6 Hz, 1H, H-4), 8.31 (d, J ) 5.9 Hz, 1H, H-6′′), 8.38 (dd,
J ) 1.8, 4.8 Hz, 1H, H-6), 8.54 (d, J ) 4.9 Hz, 1H, H-6′). 13C
NMR (75 MHz, CDCl3): δ 13.91 (SCH3), 13.93 (SCH3), 74.3
(Ar2CHOAr), 103.2 (C-5′′), 113.4 (C-5′), 122.6 (C-5), 132.5 (C-
3), 137.6 (C-4), 149.3 (C-6), 150.1 (C-2), 157.8 (C-6′), 158.0 (C-
6′′), 165.3 (C-4′), 166.8 (C-4′′), 172.3 (C-2′′), 172.9 (C-2′). MS
(ES): m/z (rel intensity) 392/394 (100/40), 356 (19), 250/252
(39/18). HRMS (ES): Calcd for C16H1535ClN5O32S2 (MH+)
392.0407, found 392.0409.
9-(Methylsulfanyl)-5a-[2-(methylsulfanyl)pyrimidin-4-
yl]-5aH-pyrido[3′,2′: 4,5]pyrrolo[1,2-c]pyrimidin-5-one
(15). Triarylmethanol 10 (500 mg, 1.28 mmol) was heated with
TFA (0.82 mL, 11 mmol) at 80 °C for 10 h. After removal of
excess TFA in vacuo, the residue was dissolved in CH2Cl2 (65
mL) and neutralized with 5% aqueous NaHCO3 solution (40
mL). After extraction with CH2Cl2, the extracts were worked
up as usual. 1H NMR spectroscopic analysis of the crude
material showed it to contain ketone 15 as the major product,
together with lesser amounts of 16 and 17. Flash chromatog-
raphy on silica, eluting with 30-50% EtOAc/hexanes, followed
by flash chromatography on reversed phase silica, eluting with
75% MeOH/water, gave 15 as a yellow solid (∼95% pure by
1H NMR, 70 mg, 15%). An analytical sample was obtained by
reversed phase preparative HPLC, eluting with 75% CH3CN/
water + 0.05% TFA, to give a yellow solid. This compound
slowly decomposes over a matter of weeks at room tempera-
9-(Methylsulfanyl)-5-[2-(methylsulfanyl)pyrimidin-4-
yl]pyrido[3′,2′:4,5]pyr rolo[1,2-c]pyrimidine (8).
1. Preparation from Triarylmethanol 10
Triarylmethanol 10 (97% pure, 497 mg, 1.23 mmol) was
heated with TFA (0.41 mL, 5.3 mmol) and TES (1.65 mL, 10.3
mmol) in a sealed Young’s tube at 75 °C for 66 h. After
concentration of excess reagents in vacuo, the residue was
dissolved in CH2Cl2 (65 mL) and neutralized with 5% aqueous
NaHCO3 solution (40 mL). After extraction with CH2Cl2 the
extracts were worked up in the usual manner. Analysis of the
1H NMR spectrum of the crude mixture revealed the presence
of the title compound, and lesser amounts of 16 and 17, in
1
ture. Mp: 52-55 °C. H NMR (500 MHz, CDCl3): δ 2.38 (s,
3H, 2′-SCH3), 2.52 (s, 3H, 9-SCH3), 5.96 (d, J ) 6.6 Hz, 1H,
H-6), 6.76 (d, J ) 6.6 Hz, 1H, H-7), 7.12 (dd, J ) 4.9, 7.4 Hz,
1H, H-3), 7.14 (d, J ) 4.9 Hz, 1H, H-5′), 7.96 (dd, J ) 1.6, 7.4
J. Org. Chem, Vol. 70, No. 16, 2005 6209

Anderson et al.
Hz, 1H, H-4), 8.47 (d, J ) 4.9 Hz, 1H, H-6′), 8.66 (dd, J ) 1.6,
4.9 Hz, 1H, H-2). 13C NMR (75 MHz, CDCl3): δ 14.08 (SCH3),
14.14 (SCH3), 70.8 (C-5a), 107.1 (C-6), 112.66 (C-5′), 112.71
(C-4a), 118.8 (C-3), 134.6 (C-4), 136.9 (C-7), 154.6 (C-9), 156.2
(C-2), 158.1 (C-6′), 164.3 (C-10a), 165.4 (C-4′), 172.8 (C-2′),
193.2 (C-5). IR (CDCl3): υmax 1732, 1414. HRMS (ES): Calcd
for C16H14N5O32S2 (MH+) 356.0640, found 356.0642.
H-3′′′), 6.98 (d, J ) 5.4 Hz, 1H, H-5′), 7.33-7.42 (m, 6H, H-3,
H-6, PMB H-2′′, PMB H-2′′′), 7.63 (d, J ) 6.8 Hz, 1H, H-7),
8.27 (dd, J ) 1.0, 5.2 Hz, 1H, H-2), 8.31 (d, J ) 5.4 Hz, 1H,
H-6′), 8.56 (br m, 1H, H-4), 10.36 (br t, J ) 5.4 Hz, 1H, 9-NH).
13C NMR (75 MHz, CDCl3): δ 44.3 (9-NHCH2Ar), 45.1 (2′-
NHCH2Ar), 55.3 (both OCH3), 100.6 (C-5), 101.5 (C-6), 107.9
(C-5′), 113.9 (PMB C-3′′ or PMB C-3′′′), 114.0 (PMB C-3′′′ or
PMB C-3′′), 120.0 (C-3), 121.9 (C-4a), 128.4 (C-4), 128.6 (PMB
C-2′′′), 128.8 (PMB C-2′′), 130.3 (PMB C-1′′), 131.4 (PMB C-1′′′),
138.4 (C-5a), 139.5 (C-2), 143.0 (C-7), 143.4 (C-10a), 149.0 (C-
9), 157.4 (C-6′), 158.7 (PMB C-4′′ or PMB C-4′′′), 158.9 (PMB
C-4′′′ or PMB C-4′′), 162.1 (C-4′ or C-2′), 162.2 (C-2′ or C-4′).
IR (KBr): υmax 3240, 2959-2835 (series of weak bands), 1560,
1512. MS (EI): m/z (rel intensity) 517 (56), 397 (12), 382 (12),
121 (100). HRMS (EI): Calcd for C30H27N7O2 (M+) 517.2226,
found 517.2242.
Deoxyvariolin B (7). Bis-amine 29 (180 mg, 0.35 mmol)
was dissolved in triflic acid (2 mL) and the deep orange/red
solution stirred at room temp for 2.5 h. The solution was cooled
to 0 °C, and aq NH3 was added dropwise until a yellow
precipitate had formed. Sat. aq. NaHCO3 was added carefully
until bubbling ceased. The suspension was filtered and the
yellow/green solid washed with H2O. The crude material was
dissolved in THF (40 mL), Merck DIOL (40 µm APD) (1.5 g)
added and the suspension slowly diluted with hexanes (60 mL)
with vigorous swirling, which led to coating of the DIOL with
the product. The suspension was loaded onto a DIOL column
(8.5 g) and flash chromatography, eluting with 2% MeOH/CH2-
Cl2, gave deoxyvariolin B (7) as a yellow microcrystalline solid
(84 mg, 88%).Mp: dec at ∼220-240 °C (lit.5d 160-162 °C). 1H
NMR (500 MHz, (CD3)2SO): δ 6.51 (s, 2H, 2′ NH2), 7.05 (d, J
) 5.4 Hz, 1H, H-5′), 7.57 (dd, J ) 4.8, 8.2 Hz, 1H, H-3), 7.62
(d, J ) 6.3 Hz, 1H, H-7), 7.67 (d, J ) 6.3 Hz, 1H, H-6), 8.22 (d,
J ) 5.4 Hz, 1H, H-6′), 8.44 (dd, J ) 1.5, 4.8 Hz, 1H, H-2), 8.5
(br, 1H, 9-NH), 8.90 (dd, J ) 1.5, 8.2 Hz, 1H, H-4), 9.4 (br,
1H, 9-NH). 13C NMR (75 MHz, (CD3)2SO): δ 99.6 (C-5), 101.9
(C-6), 107.1 (C-5′), 121.0 (C-3), 121.7 (C-4a), 129.4 (C-4), 138.3
(C-5a), 140.3 (C-2), 143.0 (C-10a), 144.0 (C-7), 149.9 (C-9),
158.2 (C-6′), 161.6 (C-4′), 163.6 (C-2′). IR (KBr): υmax 3304,
3142-2849 (series of weak bands), 1647, 1574, 1514, 1472,
1456, 1269. MS (EI): m/z (rel intensity) 277 (100), 236 (54).
HRMS (EI): Calcd for C14H11N7 (M+) 277.1076, found 277.1072.
2. Preparation from Acetate 24.
Triarylmethyl acetate 24 (745 mg, 1.72 mmol) was heated
with TFA (0.264 mL, 3.43 mmol), TES (2.19 mL, 13.7 mmol)
and 1,2-dichloroethane (4 mL) in a sealed Young’s tube at 100
°C for 3 days. The orange solution was diluted with CH2Cl2
(350 mL), washed with sat. aq. NaHCO3 solution, brine, dried
with MgSO4, filtered, and concentrated in vacuo. This crude
material was purified by repeated flash chromatography on
silica, eluting with 37% EtOAc/hexanes, to give the variolin
core 8 as a yellow microcrystalline solid (424 mg, 95% pure,
69%).
3. Preparation of 8 from Iodide 13.
The reaction vessel11 was charged with a solution of the acid
chloride 11 (1.004 g, 5.7 mmol) in THF (40 mL) and cooled to
-95 °C (MeOH/liquid N2). A solution of iodide 13 (3.164 g, 12.6
mmol) in freshly distilled THF (10 mL) was added via cannula.
A solution of n-BuLi in hexanes (1.58 M, 7.94 mL, 12.6 mmol)
was then added dropwise over 20 min with vigorous stirring.
The dark red/brown solution was stirred at -96 °C for a
further 2h then acetyl chloride (2.04 mL, 28.5 mmol) was
added over 5 min. The deep red solution was stirred at room
temp for 1 h, then diluted with CH2Cl2 (100 mL) and washed
with sat. aq. NaHCO3 then brine, dried with Na2SO4, filtered,
and concentrated in vacuo. This crude material was then
transferred to a Young’s tube and left at 0.03 mmHg for 16 h
to remove any 1-iodobutane. To this residue, was added 1,2-
dichloroethane (10 mL), TFA (0.879 mL, 11.4 mmol) and TES
(7.27 mL, 45.6 mmol), and the resulting dark red/brown
solution incubated at 100 °C for 3 d. The dark brown solution
was diluted with CH2Cl2 (400 mL), washed with sat. aq.
NaHCO3 solution, brine, dried with MgSO4, filtered, and
concentrated in vacuo. This crude material was purified by
flash chromatography on silica, eluting with 10% EtOAc/CH2-
Cl2, to give the variolin core as a yellow microcrystalline solid
(985 mg, 98% pure, 50% yield over 2 steps).
2-Chloro-4-methoxynicotinoyl chloride (6). Carboxylic
acid 1922 (1.27 g, 6.77 mmol), oxalyl chloride (2.4 mL, 28 mmol)
and DMF (3 drops) were stirred in CH2Cl2 at room tempera-
ture for 19 h, with a CaCl2 drying tube used to exclude
atmospheric moisture. After concentration in vacuo, the crude
product was purified by Kugelrohr distillation (150 °C, 0.07
mmHg) to give acid chloride 6 as a white crystalline solid (1.36
g, 97%). Mp: 66-69 °C, sweats at 62 °C. 1H NMR (500 MHz,
CDCl3): δ 3.98 (s, 3H, OCH3), 6.89 (d, J ) 5.9 Hz, 1H, H-5),
8.36 (d, J ) 5.9 Hz, 1H, H-6). 13C NMR (126 MHz, CDCl3): δ
56.8 (OCH3), 106.4 (C-5), 123.9 (C-3), 146.1 (C-2), 152.1 (C-6),
162.9 (C-4), 164.3 (C)O). IR (CDCl3): 1792, 1578, 1302, 1047.
MS (EI): m/z (rel intensity) 170/172 (100/34), 112/114 (27/9).
HRMS (EI): Calcd for C7H535ClNO2 (M+ - Cl) 170.0009,
found 170.0007.
Bis-[2-(methylsulfanyl)pyrimidin-4-yl]methanone (20).
The reactor was charged with iodide 13 (3.90 g, 15.5 mmol)
and distilled THF (47 mL), and cooled to -97 °C (MeOH/N2).
A solution of n-BuLi in hexanes (1.55 M, 10.0 mL, 15.5 mmol)
was slowly added to the stirred solution over 21 min, while
maintaining the temperature at -97 °C. After 30 min, a
solution of diethyl carbonate (0.94 mL, 7.8 mmol) in THF (4
mL) was added over about 3 min. After 15 min at -97 °C, the
reaction was allowed to warm to -35 °C over 2 h, and then to
room temperature. The reaction mixture was shaken with sat.
aq. NH4Cl solution, extracted with EtOAc, and subjected to
standard workup. The crude material was partially purified
by Kugelrohr distillation (160 °C, 0.03 mmHg after removing
more volatile components), then subsequently by flash chro-
matography on silica, eluting with 25, 30 and then 50% EtOAc/
9-(4-Methoxybenzylamino)-5-[2-(4-methoxybenzylami-
no)pyrimidin-4-yl]py rido[3′,2′:4,5]-pyrrolo[1,2-c]pyrimi-
dine (29). Under atmospheric conditions, a solution of m-CP-
BA (102 mg, 591 µmol) in chloroform (10 mL) was cooled to
-55 °C (MeOH/N2) and added dropwise, over 15 min, to a
similarly cooled stirred solution of core structure 8 (100 mg,
295 µmol) in chloroform (10 mL). The solution was left to warm
to room temperature over 1 h, then washed with sat. aq.
NaHCO3 solution (25 mL), brine (25 mL), dried with Na2SO4,
filtered, and concentrated in vacuo to leave a yellow solid. This
crude material, which was used without purification, was
predominantly a mixture of diastereomeric bis-sulfoxides 28
with the following spectroscopic characteristics. 1H NMR (500
MHz, CDCl3): (as most signals for the diastereoisomers
coincide, they are all quoted as multiplets) δ 3.03 (m, 3H), 3.23
(m, 3H), 7.62-7.65 (m, 1H), 7.77 (m, 1H), 8.23 (m, 1H), 8.63
(m, 1H), 8.66-8.69 (m, 1H), 8.78-8.80 (m, 1H), 8.85 (m, 1H).
The crude oxidized material (109 mg, 0.29 mmol) was heated
with toluene (2 mL) and p-methoxybenzylamine (192 µL, 1.46
mmol) at 100 °C for 3 d. The solution was allowed to cool to
room temperature, diluted with CH2Cl2 (20 mL), and washed
with sat. aq. NaHCO3 solution, brine, dried with Na2SO4,
filtered and concentrated in vacuo. The crude material thus
obtained was purified by repeated flash chromatography on
neutralized silica, eluting with 0.5% MeOH/CH2Cl2 (0.1%
TEA), to give the bis-amine 29 as an orange foam (109 mg,
72% over 2 steps). Mp: 54-60 °C. 1H NMR (300 MHz,
CDCl3): δ 3.808 (s, 3H, OCH3), 3.811 (s, 3H, OCH3), 4.69 (d, J
) 5.9 Hz, 2H, 2′-NHCH2Ar), 4.89 (d, J ) 5.4 Hz, 2H, 9-NHCH2-
Ar), 5.52 (br, 1H, 2′ NH), 6.89-6.93 (m, 4H, PMB H-3′′, PMB
6210 J. Org. Chem., Vol. 70, No. 16, 2005

Syntheses of Variolin B and Deoxyvariolin B
hexanes, to give ketone 20 as a pale yellow solid (1.14 g, 53%).
raphy on silica, eluting with 80% EtOAc/hexanes, gave acetate
23 (308 mg, 68%). Mp: 151-154 °C. 1H NMR (500 MHz,
CDCl3): δ 2.30 (s, 3H, OCOCH3), 2.39 (s, 6H, SCH3), 3.49 (s,
3H, OCH3), 6.76 (d, J ) 5.6 Hz, 1H, H-5), 7.18 (d, J ) 5.4 Hz,
2H, H-5′), 8.23 (d, J ) 5.6 Hz, 1H, H-6), 8.39 (d, J ) 5.4 Hz,
2H, H-6′). 13C NMR (126 MHz, CDCl3): δ 14.0 (SCH3), 21.4
(OCOCH3), 55.9 (OCH3), 84.5 (Ar3COAc), 107.0 (C-5), 114.8
(C-5′), 122.5 (C-3), 149.7 (C-6), 152.1 (C-2), 156.6 (C-6′), 166.3
(C-4), 168.16 (C-4′ or C)O), 168.22 (C)O or C-4′), 171.7 (C-
2′). IR (KBr): 1755, 1572, 1549, 1346, 1207, 1038. MS (EI):
m/z (rel intensity) 463(2), 428 (51), 423/421 (61/27), 420 (55),
404 (24), 370 (37), 369 (100), 268 (32), 125 (48), 83 (31). HRMS
(EI): Calcd for C19H18N5O332S2 (M+ - Cl) 428.0851, found
428.0860.
1
Mp: 106-107 °C. H NMR (500 MHz, CDCl3): δ 2.51 (s, 6H,
SCH3), 7.54 (d, J ) 4.9 Hz, 2H, H-5), 8.79 (d, J ) 4.9 Hz, 2H,
H-6). 13C NMR (75 MHz, CDCl3): δ 14.2 (SCH3), 114.9 (C-5),
158.8 (C-6), 159.2 (C-4), 173.2 (C-2), 190.7 (C)O). IR (KBr):
1695, 1564, 1553, 1420, 1356, 1319. MS (EI): m/z (rel intensity)
278 (100), 214 (23), 125 (76). HRMS (EI): Calcd for C10H10-
N4O32S2 (M+) 278.0296, found 278.0289.
(2-Chloro-4-methoxypyridin-3-yl)bis-[2-(methylsulfa-
nyl)pyrimidin-4-methanol (4).
1. Preparation by Reaction of 5 with 6.
The reaction vessel11 was charged with iodide 13 (4.44 g,
17.6 mmol) and freshly distilled THF (180 mL), and cooled to
-95 °C (MeOH/N2). A solution of n-BuLi in hexanes (1.55 M,
11.4 mL, 17.7 mmol) was added dropwise to the stirred solution
over 19 min, while maintaining the temperature at -95 °C.
After 30 min, a solution of acid chloride 6 (1.21 g, 5.86 mmol)
in THF (9 mL) was added to the dark mixture over 7 min.
The reaction was stirred for 4.5 h at -95 °C, quenched with
MeOD (1.2 mL) and allowed to warm to room temperature.
Excess THF was removed in vacuo, and the residue was taken
up in water/EtOAc and extracted with EtOAc (x4). The
extracts were washed with brine and dried over MgSO4.
Partial concentration of the EtOAc solution gave a precipitate,
which was removed by filtration and washed with EtOAc.
Concentration of the filtrate and washings gave the crude
material. Flash chromatography on silica, eluting with 55-
80% EtOAc/ hexanes, gave triarylmethanol 4 as a tan solid
(383 mg, 14% yield, ∼93% pure by 1H NMR spectroscopy).
Formation of Variolin Core 2.
1. Preparation from Triarylmethanol 4.
A mixture of triarylmethanol 4 (100 mg, 0.237 mmol) and
TFA (37µL, 0.48 mmol) was dissolved in 1,2-dichloroethane
(0.5 mL). The resulting orange solution was transferred to a
Young’s tube, fitted with a rubber septum, containing TES
(0.30 mL, 1.9 mmol). With a strong stream of argon flowing,
the septum was replaced with a Teflon screw-cap, and the
sealed reaction vessel was heated at 100 °C for 43 h. After
cooling, the vessel was opened and the contents diluted with
CH2Cl2 (12 mL). The solution was neutralized with 5% aq
NaHCO3 solution (8 mL) and worked up with CH2Cl2 accord-
ing to the standard procedure. Flash chromatography on
silica, eluting with 48-75% EtOAc/hexanes, gave in order of
elution: a. 4-Methoxy-9-(methylsulfanyl)-5-[2-(methyl-
sulfanyl)pyrimidin-4-yl]pyrido[3′,2 ′:4,5] pyrrolo[1,2-c]-
pyrimidine (2) as a yellow solid (41 mg, 47%). Mp: 192-194
1
Mp: 135-137 °C. H NMR (500 MHz, CDCl3): δ 2.49 (s, 6H,
SCH3), 3.43 (s, 3H, OCH3), 6.55 (s, 1H, OH), 6.76 (d, J ) 5.4
Hz, 1H, H-5), 7.39 (d, J ) 5.4 Hz, 2H, H-5′), 8.25 (d, J ) 5.4
Hz, 1H, H-6), 8.46 (d, J ) 5.4 Hz, 2H, H-6′). 13C NMR (75 MHz,
CDCl3): δ 14.1 (SCH3), 55.8 (OCH3), 78.0 (Ar3COH), 107.1 (C-
5), 113.7 (C-5′), 124.8 (C-3), 149.8 (C-6), 152.3 (C-2), 157.2 (C-
6′), 165.9 (C-4), 171.0 (C-4′), 171.1 (C-2′). IR (CDCl3): 3344
(br), 1576, 1551, 1348, 1209, 1043. MS (EI): m/z (rel intensity)
421/423 (65/29), 387 (19), 370 (21), 369 (21), 342 (38), 338 (57),
296 (44), 268 (60), 260 (100), 125 (98), 112 (22). HRMS (EI):
Calcd for C17H1635ClN5O232S2 (M+) 421.0434, found 421.0448.
1
°C. H NMR (500 MHz, CDCl3): δ 2.65 (s, 3H, 2′-SCH3), 2.70
(s, 3H, 9-SCH3), 4.03 (s, 3H, OCH3), 6.92 (d, J ) 5.4 Hz, 1H,
H-3), 7.40 (d, J ) 5.4 Hz, 1H, H-5′), 7.71 (d, J ) 6.8 Hz, 1H,
H-7), 7.98 (d, J ) 6.8 Hz, 1H, H-6), 8.48 (m, 2H, H-2, H-6′).
13C NMR (75 MHz, CDCl3): δ 14.1 (SCH3), 14.9 (SCH3), 55.6
(OCH3), 101.9 (C-3), 102.5 (C-5), 108.5 (C-6), 110.8 (C-4a), 117.7
(C-5′), 135.9 (C-5a), 138.6 (C-7), 143.5 (C-2), 144.3 (C-10a),
154.2 (C-9), 155.6 (C-6′), 159.6 (C-4), 161.1 (C-4′), 171.2 (C-
2′). IR (CDCl3): 1570, 1468, 1285, 1013. MS (EI): m/z (rel
intensity) 369 (100), 354 (13), 112 (10), 70 (10). HRMS (EI):
Calcd for C17H15N5O32S2 (M+) 369.0718, found 369.0720.
2. Preparation by Reaction of 20 with 21.
2-Chloro-4-methoxypyridine (21) (0.633 g, 4.41 mmol) was
dissolved in freshly distilled THF (18 mL). The flask was cooled
to below -90 °C (acetone/N2), and a solution of n-BuLi in
hexanes (1.55 M, 2.9 mL, 4.5 mmol) was slowly added over 17
min to the stirred solution, keeping the temperature below -90
°C. The orange solution was then stirred at -78 °C for 1 h, by
which time it had become a wine-red color. The reaction
mixture was again cooled to below -90 °C, and a solution of
ketone 20 (1.14 g, 4.09 mmol) in THF (10 mL) was added over
11 min while maintaining the temperature below -90 °C. The
dark mixture was then stirred at -78 °C for 3.5 h, quenched
with methanol and allowed to warm to room temperature. The
reaction mixture was shaken with sat. aq. NH4Cl solution,
extracted with EtOAc and subjected to standard workup. Flash
chromatography on silica, eluting with 80% EtOAc/hexanes,
gave triarylmethanol 4 as a pale yellow solid (1.32 g, 76%).
(2-Chloro-4-methoxypyridin-3-yl)bis-[2-(methylsulfa-
nyl)pyrimidin-4-yl]me thyl acetate (23). 2-Chloro-4-meth-
oxypyridine (21)23 (0.141 g, 0.980 mmol) was dissolved in
freshly distilled THF (4 mL). The flask was cooled to below
-90 °C and a solution of n-BuLi in hexanes (1.6M, 0.61 mL,
0.98 mmol) was added over 2 min to the stirred solution. The
orange solution was then stirred at -78 °C for 1 h, by which
time it had become a wine-red color. The reaction mixture was
again cooled to below -90 °C and a solution of ketone 20 (271
mg, 0.974 mmol) in THF (2.2 mL) was added over 2 min. The
dark mixture was stirred at -78 °C for 3 h, whereupon freshly
distilled acetyl chloride (0.30 mL, 4.2 mmol) was added. After
warming to room temperature overnight, the reaction mixture
was shaken with sat. aq. NaHCO3 solution, extracted with
EtOAc, and subjected to standard workup. Flash chromatog-
b. (2-Chloro-4-methoxypyridin-3-yl)-[2-(methylsulfa-
nyl)pyrimidin-4-yl]-[2-(met hylsul-fanyl)pyrimidin-4-
yloxy]methane (22) as a viscous gum (28 mg, 28%). Tritu-
ration with n-hexane gave an off-white solid. Mp: 113-121
1
°C. H NMR (500 MHz, CDCl3): δ 2.36 (s, 3H, 2′-SCH3), 2.44
(s, 3H, 2′′-SCH3), 3.84 (s, 3H, OCH3) 6.57 (d, J ) 5.9 Hz, 1H,
H-5′′), 6.80 (d, J ) 5.9 Hz, 1H, H-5), 7.17 (d, J ) 4.9 Hz, 1H,
H-5′), 7.79 (s, 1H, Ar2CHOAr), 8.27 (d, J ) 5.9 Hz, 1H, H-6),
8.30 (d, J ) 5.9 Hz, 1H, H-6′′), 8.49 (d, J ) 4.9 Hz, 1H, H-6′).
13C NMR (75 MHz, CDCl3): δ 13.88 (SCH3), 13.90 (SCH3), 56.2
(OCH3), 71.8 (Ar2CHOAr), 103.5 (C-5′′), 106.4 (C-5), 112.9 (C-
5′), 120.8 (C-3), 150.7 (C-6), 152.5 (C-2), 157.1 (C-6′), 158.0 (C-
6′′), 166.0 (C-4), 167.1 (C-4′), 167.3 (C-4′′), 172.1 (C-2′), 172.4
(C-2′′). MS (EI): m/z (rel intensity) 421/423 (94/43), 387 (21),
374/376 (33/13), 296 (79), 280 (39), 246 (79), 212 (100), 184
(28), 147 (43), 125 (58), 112 (45), 104 (48), 77 (53), 76 (50), 70
(44). HRMS (EI): Calcd for C17H16N5O235Cl32S2 (M+) 421.0434,
found 421.0444.
2. Preparation from Acetate 23.
A mixture of acetate 23 (112 mg, 0.241 mmol) and TFA (37
µL, 0.48 mmol) was dissolved in 1,2-dichloroethane (0.5 mL).
The resulting orange solution was transferred to a Young’s
tube, fitted with a rubber septum, containing TES (0.31 mL,
1.9 mmol). With a strong stream of argon flowing, the septum
was replaced with a Teflon screw-cap, and the sealed reaction
vessel was heated at 80 °C for 26 h. After cooling, the vessel
was opened and the contents diluted with CH2Cl2 (15 mL).
The solution was neutralized with 5% aq NaHCO3 solution (8
mL) and the phases separated. The aqueous layer was further
extracted with CH2Cl2 and the organic extracts were worked
J. Org. Chem, Vol. 70, No. 16, 2005 6211

Anderson et al.
up according to the standard procedure. Flash chromatography
on silica, eluting with 48% EtOAc/hexanes, gave core structure
2 (67 mg, 75%).
5.5 Hz, 1H, H-3), 6.88-6.92 (m, 4H, PMB H-3′′, PMB H-3′′′),
7.03 (d, J ) 6.8 Hz, 1H, H-6), 7.06 (d, J ) 5.7 Hz, 1H, H-5′),
7.32 (d, J ) 8.4 Hz, 2H, PMB H-2′′′), 7.40 (d, J ) 8.4 Hz, 2H,
PMB H-2′′), 7.68 (d, J ) 6.8 Hz, 1H, H-7), 8.05 (d, J ) 5.5 Hz,
1H, H-2), 8.29 (d, J ) 5.7 Hz, 1H, H-6′), 10.94 (br t, J ) 5.5
Hz, 1H, 9-NHPMB), 15.7 (br, 1H, OH). 13C NMR (75 MHz,
CDCl3): δ 44.3 (9-NHCH2Ar), 45.2 (2′-NHCH2Ar), 55.3 (both
OCH3), 100.3 (C-6), 100.5 (C-5), 106.9 (C-5′), 107.5 (C-3), 111.4
(C-4a), 114.1 (PMB C-3′′, PMB C-3′′′), 128.8 (PMB C-2′′), 129.2
(PMB C-2′′′), 130.2 (PMB C-1′′, PMB C-1′′′), 137.5 (C-5a), 142.8
(C-2), 143.8 (C-7), 145.4 (C-10a), 149.7 (C-9), 158.6 (C-6′), 158.9
(PMB C-4′′), 159.0 (PMB C-4′′′), 159.5 (C-4′), 159.8 (C-4), 160.0
(C-2′). IR (KBr): νmax 3223, 2820-2835 (series of weak bands),
1574. MS (EI): m/z (rel intensity) 533 (38), 413 (19), 412 (18),
411 (15), 121 (100). HRMS (EI): Calcd for C30H27N7O3 (M+)
533.2175, found 533.2185.
Variolin B (1). Pyridinol 27 (31 mg, 0.058 mmol) was
dissolved in neat TfOH (0.5 mL) under atmospheric conditions.
The flask was stoppered, and the deep red solution was left
at room temperature for 17 h. The reaction was cooled in ice
and carefully quenched with ice-water (1 mL), followed by
conc aq NH3 solution (1 mL), which produced a bright yellow
precipitate. After mixing thoroughly to ensure complete neu-
tralization, the precipitate was filtered through a fine frit,
washing with water and then MeOH/water (1:1). The solid was
then suspended in MeOH/CH2Cl2 (a small amount of TFA was
added to convert the free base to the more soluble salt form)
and adsorbed onto a small amount of reversed phase silica for
flash chromatography, eluting with 50% MeOH/water then
70% MeOH/water + 0.5% TFA. The bright yellow fractions
were combined and concentrated in vacuo to give variolin B
as its trifluoroacetate salt. The salt was neutralized by
sonication with conc aq NH3/MeOH to give the free base 1.
Concentration in vacuo (35 °C, 0.03 mmHg to remove am-
monium trifluoroacetate) gave variolin B (1) as a brown-orange
solid (13.6 mg, 80%). Mp: dec at ∼320 °C (lit3a 45 °C, dec). 1H
NMR (500 MHz, (CD3)2SO): δ 6.81 (d, J ) 5.6 Hz, 1H, H-3),
6.97 (s, 2H, 2′ NH2), 7.14 (d, J ) 5.6 Hz, 1H, H-5′), 7.23 (d,
J ) 6.7 Hz, 1H, H-6), 7.63 (d, J ) 6.7 Hz, H-7), 8.17 (d, J )
5.6 Hz, 1H, H-2), 8.27 (d, J ) 5.6 Hz, 1H, H-6′), 8.45 (br, 1H,
9-NH), 9.79 (br, 1H, 9-NH), 16.05 (s, 1H, OH). 13C NMR (75
MHz, (CD3)2SO): δ 99.6 (C-5), 100.3 (C-6), 106.0 (C-5′), 107.5
(C-3), 111.2 (C-4a), 137.1 (C-5a), 143.1 (C-2), 144.6 (C-7), 144.9
(C-10a), 150.3 (C-9), 158.3 (C-4′), 159.8 (C-4), 160.0 (C-6′), 161.4
(C-2′). IR (KBr): 3085, 1670, 1572, 1477, 1458, 1298. HRMS
(ES): Calcd for C14H12N7O (MH+) 294.1103, found 294.1096.
4-Methoxy-9-(4-methoxybenzylamino)-5-[2-(4-methox-
ybenzylamino)pyrim idin-4-yl]pyrido [3′,2′:4,5]pyrrolo-
[1,2-c]pyrimidine (26). Core structure 2 (37 mg, 0.10 mmol)
was dissolved in CHCl3 (5 mL) under atmospheric conditions
and cooled to -40 °C (CH3CN/CO2). A solution of m-CPBA in
CHCl3 (10 mg/mL) was similarly cooled and added dropwise
to the solution of 2 until TLC analysis indicated the consump-
tion of all starting material (after about 2 equiv of m-CPBA).
The solution was warmed to room temperature and neutralized
with sat. aq. NaHCO3 solution. This was extracted with CH2-
Cl2 and a standard workup gave a yellow solid which was
predominantly a mixture of diastereomeric bis-sulfoxides 25.
The crude mixture was used without purification; however,
the bis-sulfoxides had the following spectroscopic character-
istics: 1H NMR (500 MHz, CDCl3): (As most signals for the
diastereoisomers coincide, they are all quoted as multiplets.)
δ 3.02 (m, 3H), 3.19 (m, 3H), 4.12 (m, 3H), 7.00-7.01 (m, 1H),
7.98-7.99 (m, 1H), 8.12-8.14 (m, 1H), 8.48-8.49 (m, 1H),
8.64-8.67 (m, 1H), 8.79-8.81 (m, 1H).
The crude oxidized material 25 was heated with an excess
of p-methoxybenzylamine (0.15 mL, 1.1 mmol) at 85 °C for 15
h. The red paste thus obtained was directly purified by flash
chromatography on silica, eluting with 2.5-4% MeOH/CH2-
Cl2. This was chromatographed again on silica, eluting with
50% EtOAc/CH2Cl2 then 100% EtOAc, to give bis-amine 26
as a yellow-orange solid (43 mg, 78% over two steps). Mp: 74-
1
77 °C. H NMR (500 MHz, CDCl3): δ 3.81 (s, 6H, both PMB
OCH3), 3.99 (s, 3H, 4-OCH3), 4.66 (d, J ) 5.6 Hz, 2H,
2′-NHCH2Ar), 4.85 (d, J ) 5.5 Hz, 2H, 9-NHCH2Ar), 5.51 (br,
1H, 2′ NHPMB), 6.82 (d, J ) 5.6 Hz, 1H, H-3), 6.89-6.91 (m,
4H, PMB H-3′′, PMB H-3′′′), 7.00 (d, J ) 5.2 Hz, 1H, H-5′),
7.29 (br, 1H, H-6), 7.34 (d, J ) 8.5 Hz, 2H, PMB H-2′′′), 7.39
(d, J ) 8.5 Hz, 2H, PMB H-2′′), 7.43 (br d, J ) 6.3 Hz, 1H,
H-7), 8.16 (d, J ) 5.6 Hz, 1H, H-2), 8.26 (d, J ) 5.2 Hz, 1H,
H-6′), 10.39 (br t, J ) 5.5 Hz, 1H, 9-NHPMB). 13C NMR (75
MHz, CDCl3): δ 44.3 (9-NHCH2Ar), 44.9 (2′-NHCH2Ar), 55.3
(both PMB OCH3), 55.5 (4-OCH3), 101.3 (C-5), 101.6 (C-3),
101.8 (C-6), 111.4 (C-4a), 112.2 (C-5′), 113.96 (PMB C-3′′′ or
PMB C-3′′), 114.03 (PMB C-3′′ or PMB C-3′′′), 128.5 (PMB
C-2′′′), 128.8 (PMB C-2′′), 130.5 (PMB C-1′′), 131.4 (PMB C-1′′′),
137.5 (C-5a), 141.5 (C-2), 141.9 (C-7), 144.7 (C-10a), 148.7 (C-
9), 154.6 (br, C-6′), 158.7 (PMB C-4′′′), 158.9 (PMB C-4′′), 159.4
(C-4), 160.9 (C-2′), 162.6 (C-4′). IR (KBr): 3242, 2997-2833
(series of weak bands), 1570. MS (EI): m/z (rel intensity) 547
(65), 427 (14), 426 (9), 425 (12), 136 (38), 135 (53), 121 (100),
112 (21), 77 (24). HRMS (EI): Calcd for C31H29N7O3 (M+)
547.2332, found 547.2334.
Acknowledgment. We thank PharmaMar SA, Tres
Cantos, and the University of Canterbury for financial
assistance, Professors M. H. G. Munro and J. W. Blunt
for their support of this work and the provision of an
authentic sample of variolin B, Professors P. J. Steel
and A. J. Phillips for helpful discussions, and Mr Bruce
Clark for mass spectrometry experiments. R.J.A. ac-
knowledges the New Zealand Vice-Chancellors’ Com-
mittee for scholarship support.
4-Hydroxy-9-(4-methoxybenzylamino)-5-[2-(4-methox-
ybenzylamino)pyrim idin-4-yl]pyrido [3′,2′:4,5]pyrrolo-
[1,2-c]pyrimidine (27). A solution of sodium ethanethiolate
(80%, 202 mg, 1.92 mmol) in dry DMF (2 mL) was added to a
solution of bis-amine 26 (106 mg, 0.193 mmol) in DMF (2 mL),
and the mixture was stirred at 55 °C for 7.5 h. After cooling,
aq NH4Cl solution was added and the mixture was extracted
with EtOAc. The organic extracts were washed three times
with water to remove DMF and worked up as usual. Flash
chromatography on silica, eluting with 4% MeOH/CH2Cl2, gave
pyridinol 27 as a yellow solid (98 mg, 95%). Mp: 195-196 °C.
1H NMR (500 MHz, CDCl3): δ 3.80 (s, 3H, OCH3), 3.81 (s, 3H,
OCH3), 4.61 (d, J ) 5.5 Hz, 2H, 2′-NHCH2Ar), 4.85 (d, J ) 5.5
Hz, 2H, 9-NHCH2Ar), 5.35 (br, 1H, 2′ NHPMB), 6.76 (d, J )
Supporting Information Available: General experimen-
tal details, details of the preparation of 13, 19, and 21, and
copies of 1H and 13C NMR spectra of all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JO050523V
6212 J. Org. Chem., Vol. 70, No. 16, 2005