Asymmetric total synthesis of (-)-VM55599: Establishment of the absolute stereochemistry and biogenetic implications

Article abstract of DOI:10.1021/ja017425l

The first asymmetric biomimetic total synthesis of VM55599 (13) has been achieved utilizing an intramolecular Diels - Alder cycloaddition as a key step. The synthetic material was utilized to elucidate the absolute stereochemistry of the natural product. The results are discussed in terms of a unified biogenesis of the paraherquamides and VM55599.

Full text of DOI:10.1021/ja017425l

Published on Web 02/16/2002
Asymmetric Total Synthesis of (-)-VM55599: Establishment
of the Absolute Stereochemistry and Biogenetic Implications
Juan F. Sanz-Cervera and Robert M. Williams*
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523
Received October 30, 2001
Abstract: The first asymmetric biomimetic total synthesis of VM55599 (13) has been achieved utilizing an
intramolecular Diels-Alder cycloaddition as a key step. The synthetic material was utilized to elucidate the
absolute stereochemistry of the natural product. The results are discussed in terms of a unified biogenesis
of the paraherquamides and VM55599.
Introduction
The paraherquamides1 (Figure 1), together with the brevian-
amides,2 marcfortines,3 and sclerotamides4 are secondary me-
tabolites of fungal origin that feature a common bicyclo[2.2.2]-
diazaoctane core. It has been postulated that this ring system is
generated through an intramolecular Diels-Alder cycloaddition
of the C5 moiety across the R-carbons of the amino acid
subunits, as depicted in Scheme 1.5
Everett and co-workers described in 1993 the isolation of
VM55599 (13, Figure 1), a minor metabolite from culture
extracts of a Penicillium sp. (IMI332995) that also produces
paraherquamide A (1), among other paraherquamides.6 Taking
Figure 1. 1, Paraherquamide A, R1 ) OH, R2 ) Me, R3 ) H2, X ) N; 2,
paraherquamide B, R1 ) H, R2 ) H, R3 ) H2, X ) N; 3, paraherquamide
C, R1 ) R2 ) CH2, R3 ) H2, X ) N; 4, paraherquamide D, R1 ) O, R2
) CH2, R3 ) H2, X ) N; 5, VM55596, R1 ) OH, R2 ) Me, R3 ) H2, X
) N+-O-; 6, VM55597, R1 ) OH, R2 ) Me, R3 ) O, X ) N; 7,
* To whom correspondence should be addressed. E-mail: rmw@
chem.colostate.edu
(1) (a) Yamazaki, M.; Okuyama, E.; Kobayashi, M.; Inoue, H. Tetrahedron
Lett. 1981, 22, 135-136. (b) Ondeyka, J. G.; Goegelman, R. T.; Schaeffer,
J. M.; Kelemen, L.; Zitano, L. J. Antibiot. 1990, 43, 1375-1379. (c) Liesch,
J. M.; Wichmann, C. F. J. Antibiot. 1990, 43, 1380-1386. (d) Blanchflower,
S. E.; Banks, R. M.; Everett, J. R.; Manger, B. R.; Reading, C. J. Antibiot.
1991, 44, 492-497.
(2) (a) Birch, A. J.; Wright, J. J. J. Chem. Soc., Chem. Commun. 1969, 644-
645. (b) Birch, A. J.; Wright, J. J., Tetrahedron 1970, 26, 2329-2344. (c)
Birch, A. J.; Russell, R. A. Tetrahedron 1972, 28, 2999-3008. (d) Bird,
B. A.; Remaley, A. T.; Campbell, I. M. Appl. EnViron. Microbiol. 1981,
42, 521-525. ((e) Bird, B. A.; Campbell, I. M. Appl. EnViron. Microbiol.
1982, 43, 345. (f) Robbers, J. E.; Straus, J. W. Lloydia 1975, 38, 355. (g)
Paterson, R. R. M.; Hawksworth, D. L. Trans. Br. Mycol. Soc. 1985, 85,
95-100. (h) Wilson, B. J.; Yang, D. T. C.; Harris, T. M. Appl. Microbiol.
1973, 26, 633-635. (i) Coetzer, J., Acta Crystallogr. 1974, B30, 2254-
2256.
paraherquamide E (VM54159), R1 ) Me, R2 ) H; 8, SB203105, R1
)
Me, R2 ) OH; 9, SB200437, R1 ) H, R2 ) H; 10, paraherquamide F
(VM55594), R1 ) H, R2 ) Me, R3 ) Me; 11, paraherquamide G
(VM54158), R1 ) OH, R2 ) Me, R3 ) Me; 12, VM55595, R1 ) H, R2 )
Me, R3 ) H.
Scheme 1. Proposed Formation of the Bicyclo[2.2.2]diazaoctane
Ring System in the Biosynthesis of the Paraherquamides,
Brevianamides, and Marcfortines
(3) (a) Polonsky, J.; Merrien, M.-A.; Prange, T.; Pascard, C. J. Chem. Soc.,
Chem. Commun. 1980, 601-602. (b) Prange, T.; Buillion, M.-A.;
Vuilhorgne, M.; Pascard, C.; Polonsky, J. Tetrahedron Lett. 1980, 22,
1977-1980.
(4) Whyte, A. C.; Gloer, J. B. J. Nat. Prod. 1996, 59, 1093-1095.
(5) (a) Porter, A. E. A.; Sammes, P. G. J. Chem. Soc., Chem. Commun. 1970,
1103. (b) Williams, R. M.; Kwast, E.; Coffman, H.; Glinka, T. J. Am. Chem.
Soc. 1989, 111, 3064-3065. (c) Williams, R. M.; Glinka, T.; Kwast, E.;
Coffman, H.; Stille, J. K. J. Am. Chem. Soc. 1990, 112, 808-821. (d) Sanz-
Cervera, J. F.; Glinka, T.; Williams, R. M. J. Am. Chem. Soc. 1993, 115,
347-348. (e) Sanz-Cervera, J. F.; Glinka, T.; Williams, R. M. Tetrahedron
1993, 49, 8471-8472. (f) Domingo, L. R.; Sanz-Cervera, J. F.; Williams,
R. M.; Picher, M. T.; Marco, J. A. J. Org. Chem. 1997, 62, 1662-1667.
(g) Stocking, E. M.; Sanz-Cervera, J. F.; Williams, R. M. Angew. Chem.
1999, 111, 880-883; Angew. Chem., Int. Ed. 1999, 38, 786-789. (h)
Stocking, E. M.; Williams, R. M.; Sanz-Cervera, J. F. J. Am. Chem. Soc.
2000, 122, 9089-9098. (i) Sanz-Cervera, J. F.; Williams, R. M.; Marco,
J. A.; Lo´pez-Sa´nchez, J. M.; Mart´ınez, M. E.; Gonza´lez, F.; Sanceno´n, F.
Tetrahedron 2000, 56, 6345-6358.
into account the structural similarities between these co-
occurring metabolites, these authors proposed that VM55599
(13) might indeed be a biosynthetic precursor of paraherquamide
A (1). The relative stereochemistry of VM55599 (13) was
assigned by Everett and co-workers through extensive 1H NMR
nOe experiments, but the small quantity of this compound that
was isolated precluded the determination of its absolute con-
figuration.
An important detail about the absolute configuration that these
authors hypothesized for VM55599 (13) involves the relative
(6) Blanchflower, S. E.; Banks, R. M.; Everett, J. R.; Reading, C. J. Antibiot.
1993, 46, 1355-1363.
9
2556 VOL. 124, NO. 11, 2002 J. AM. CHEM. SOC.
10.1021/ja017425l CCC: $22.00 © 2002 American Chemical Society

Asymmetric Total Synthesis of (−)-VM55599
A R T I C L E S
Scheme 2. Unified Biogenesis of the Paraherquamides and
configuration at C-14 that disposes the C-17 methyl residue syn-
to the bridging C5 moiety. In stark contrast, in all the other
paraherquamides with a methyl group in an equivalent position
the methyl group is disposed anti- to the bridging C5 moiety.
Therefore, if VM55599 (13) were indeed a biosynthetic precur-
sor to the paraherquamides as proposed, the configuration at
C-14 would have to be inverted at some point along the
biosynthetic pathway.
VM55599
Previous studies from our laboratories have shown that (S)-
isoleucine (L-Ile) serves as the precursor of the â-methyl-â-
hydroxyproline ring in paraherquamide A (1). This mandates
that the L-Ile side chain stereochemistry is retained at C-14 in
paraherquamides 7-10, 12 and that hydroxylation in paraher-
quamides 1, 5, 6, and 11 occurs with net retention at C-14.7
Accordingly, these results brought into question the capacity
of VM55599 (13) to serve as a biosynthetic precursor to the
paraherquamides. If, as one could reasonably speculate, L-Ile
is not only a biosynthetic precursor to the paraherquamides, but
also to VM55599 (13), the absolute stereochemistry of this
compound must be that depicted in Figure 1. Thus, the absolute
configuration of the bicyclo[2.2.2]diazaoctane ring system of
VM55599 would be anticipated to be enantiomorphic to that
of all other members of the paraherquamide family.
These experimental observations led us to propose a unified
biosynthesis of the paraherquamides and VM55599 (13), as
shown in Scheme 2.8 In this proposal, the biosynthetic precursors
of the paraherquamides and that of VM55599 would arise as
diastereomeric products of the Diels-Alder cycloaddition of a
common azadiene through two of four possible diastereomeric
transition states.
The minor product of this cycloaddition culminating in
VM55599 would constitute approach of the dienophile syn- to
the methyl group (as in structure B, Scheme 2). The major
product of the cycloaddition would be compound 15 or 16,
wherein the dienophile attacks the opposite face of the azadiene,
anti- to the methyl group, which would give rise to the various
paraherquamides.
This hypothesis was recently experimentally tested through
feeding experiments of racemic, doubly 13C-labeled compounds
13-16.9 These experiments revealed that only intermediate 15
was incorporated into paraherquamide A, while racemic, doubly
13C-labeled VM55599 (13), 14, and 16 were not incorporated.
While these results lend support to the proposed biogenesis and
provide a prediction as to the correct stereostructure of
VM55599, the absolute stereochemistry of VM55599 remained
uncertain, abrogating the connection of this metabolite to the
paraherquamide pathway. Considering that VM55599 (13) is
produced in very small quantities by the Penicillium sp. (>600:1
ratio of paraherquamide A:VM55599), a feeding experiment
with labeled L-Ile proved impractical.
HPLC.10 The two enantiomers readily separated, which were
then individually compared to an authentic sample of natural
VM55599. The first eluting enantiomer coincided in retention
time and optical rotation with the natural VM55599 (13), and
this procedure provided a way to prepare a few milligrams of
the natural antipode of VM55599 (13). Both enantiomers were
then transformed into their N-p-bromobenzoate derivatives
through reaction with p-bromobenzoyl chloride and 2,6-lutidine
in refluxing acetonitrile. However, these compounds failed to
produce a monocrystal of sufficient quality for X-ray diffraction
analysis in our hands. A plausible alternative involving an
enantiospecific total synthesis starting with a chiral, nonracemic
compound of known absolute configuration was subsequently
investigated.
Results and Discussion
In contemplating a stereochemically unambiguous asymmetric
total synthesis in this general family, we envisioned installing
the â-methylproline ring with the anticipated absolute stereo-
chemistry in a manner in which the critical stereogenic center
at C-14 would not be perturbed. Our recently reported racemic
synthesis of VM55599 could not be adapted to suit this
To determine the absolute configuration of VM55599 (13),
we first resolved synthetic, racemic VM555998c,9 by chiral
(7) (a) Stocking, E. M.; Sanz-Cervera, J. F.; Williams, R. M.; Unkefer, C. J.
J. Am. Chem. Soc. 1996, 118, 7008-7009. (b) Stocking, E. M.; Martinez,
R. A.; Silks, L. A.; Sanz-Cervera, J. F.; Williams, R. M. J. Am. Chem.
Soc. 2001, 123, 3391-3392.
(8) (a) Williams, R. M.; Sanz-Cervera, J. F.; Sanceno´n, F.; Marco, J. A.;
Halligan, K. J. Am. Chem. Soc. 1997, 119, 1090-1091. (b) Williams, R.
M.; Sanz-Cervera, J. F.; Sanceno´n, F.; Marco, J. A.; Halligan, K. Bioorg.
Med. Chem. 1998, 6, 1233-1241. (c) Stocking, E. M.; Sanz-Cervera, J.
F.; Williams, R. M. J. Am. Chem. Soc. 2000, 122, 1675-1683.
(9) Stocking, E. M.; Sanz-Cervera, J. F.; Williams, R. M. Angew. Chem., Int.
Ed. 2001, 40, 1296-1298.
(10) The HPLC separation was carried out with an (R,R) Whelk-O2 10/100 1
cm diameter × 25 cm length semipreparative column. A mixture of
heptane-i-PrOH, 82:18 was used as mobile phase. The detector was set at
220 nm. Racemic VM55599 (1 mg) was injected each time. Natural
VM55599 had a retention time of 11.5 min, while ent-VM55599 was eluted
at 14.9 min. The resolution was total (100% valley between peaks); this
was confirmed later through re-injection of the isolated peaks under the
same conditions, which showed no detectable amounts of the other
enantiomer.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2557

A R T I C L E S
Sanz-Cervera and Williams
Scheme 3. Conversion of L-Ile into the Key Azadiene Precursor
23
Scheme 4. Intramolecular Diels-Alder Cycloaddition of 23 and
the Asymmetric Total Synthesis of VM55599 (13)
limitation since a dehydro-â-methylproline residue was em-
ployed as the azadiene precursor.8c It was readily recognized
that, if a dehydrotryptophan residue could be employed in a
similar manner, generation of the azadiene species should not
perturb the â-position of the proline ring leading to optically
pure cycloadducts. This strategy has been successfully realized
as described below.
easily separated by chromatography on silica gel. Although both
compounds would have been potentially useful as substrates
for the synthesis of 13, only the major compound 20 was carried
forward. After the enolate of 20 was generated through treatment
with sodium bis(trimethylsilyl)amide in THF at -78 °C,
condensation with aldehyde 2113 furnished a mixture of three
diastereomers 22 in 87-90% combined yield. The relative
stereochemistry of these compounds was not determined, but
one of these isomers was separated for analytical identification
purposes by means of column chromatography on silica gel.
The MTM group in the mixture of diastereomers 22 was
removed through treatment with an excess of iodomethane in
acetone in the presence of an aqueous solution of NaHCO312 to
give a mixture of three diastereomers in 92% yield. Treatment
of the diastereomeric mixture with 50% aqueous formic acid13
at 80 °C for 45 min resulted in removal of the MTM group and
dehydration of the alcohol, giving exclusively compound 23 in
70% yield. No trace of the corresponding E-isomer was detected.
The configuration of the double bond was assigned through
comparison of 1H NMR spectral data with similar compounds.14
When compound 23 was treated with acetyl chloride at room
temperature for 14 days according to conditions recently
described by Liebscher and co-workers,15 a mixture of three
diastereomeric cycloadducts 14 (35% yield), 24 (15% yield),
and 25 (10% yield) resulted (Scheme 4). It is assumed that
acetylation of 23 yields the O-acyl lactim (A) that tautomerizes
Inspired by Nature, we selected L-Ile as the starting material
for the construction of the â-methylproline ring. The configu-
ration of the carbon that supports the methyl group in L-Ile is
S, and it can be transformed conveniently in five steps and 47%
overall yield into the optically pure â-methylproline derivative
17 in multigram amounts using a Hoffman-Loeffler-Freytag
sequence.11 Compound 17 was easily hydrolyzed to acid 18 in
96% yield with LiOH in a mixture of THF-EtOH-H2O 1:1:1.
The acid 18 was then coupled to glycine methyl ester hydro-
chloride with EDCI (1-(3-dimethylaminopropyl)-3-ethylcarbo-
diimide hydrochloride) in the presence of 3-hydroxybenzotri-
azole and triethylamine in dichloromethane to give the
corresponding dipeptide in 96% yield. The resulting crude
compound was deprotected with TFA at room temperature. The
crude deprotected dipeptide was treated with triethylamine to
neutralize the excess TFA, and the mixture of salts thus prepared
was heated in refluxing toluene to give the diketopiperazine 19
in 88% yield for three steps (Scheme 3).
The diketopiperazine 19 was protected as its N-methylthio-
methyl (MTM) derivative through formation of the sodium salt
with NaH in DMF at 0 °C, followed by reaction with
chloromethyl methyl thioether at room temperature.12 This
reaction furnished a mixture of the MTM-protected diketopip-
erazine 20 (61.5% yield) and an epimer (10% yield), which were
(13) Nakatsuka, S.; Miyazaki, H.; Goto, T. Tetrahedron Lett. 1980, 21, 2817-
2820.
(11) (a) Titouani, S. L.; Lavergne, J.-P.; Vaillefont, P.; Jaquier, R. Tetrahedron
1980, 36, 2961-2965. (b) Stocking, E. M.; Sanz-Cervera, J. F.; Unkefer,
C.; Williams, R. M. Tetrahedron 2001, 57, 5303-5320.
(14) Plate, R.; Rutger, R. J. F.; Ottenheijm, H. C. J. J. Chem. Soc., Perkin Trans.
1 1987, 2473-2480. See also ref 13.
(15) Shangde, J.; Wessig, P.; Liebscher, J. J. Org. Chem. 2001, 66, 3984-
3997.
(12) Corey, E. J.; Bock, M. G. Tetrahedron Lett. 1975, 16, 3269-3270.
9
2558 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Asymmetric Total Synthesis of (−)-VM55599
A R T I C L E S
to azadiene B, which suffers intramolecular Diels-Alder
cycloaddition from three of the four possible diastereomeric
transition states, followed by loss of acetate. We did not detect
the presence of any acetylated products or intermediates. The
cycloadducts obtained had identical spectroscopic properties to
the racemic compounds previously prepared in our laboratories.8c,9
except for being optically pure.
Finally, reduction of the cycloadduct 14 with an excess of
DIBAH in toluene at room temperature for 24 h gave synthetic
(-)-VM55599 (13) that was identical in all respects to natural
(-)-VM55599 (13), including optical rotation, CD and retention
time on chiral HPLC (see Supporting Information). Thus, natural
VM55599 (13) has the absolute configuration depicted (and
predicted) in Scheme 1.
This result rigorously confirms the predicted absolute ster-
eochemistry of VM55599, and provides additional experimental
support for the unified biogenesis proposed for this compound
and the paraherquamides (Scheme 2). It was quite surprising
to observe that cycloadduct 16, which contains the relative and
absolute stereochemistry of the paraherquamides, was not
detected from the cycloaddition reaction. The cycloadducts
obtained in our racemic synthesis8c gave (as O-methyl lactim
ethers), compounds stereochemically corresponding to 16:24:
25:14 in a ratio 3.7:1.6:1:2.6. In the present case, the ratio is
3.5:1.5:1:0. Thus, both laboratory cycloaddition reactions8c
display a proclivity for the formation of the VM55599 relative
stereochemistry with differing proportions of the other three
diastereomeric relationships. It is interesting to compare the syn:
anti- ratios with respect to the relative stereochemistry at C-20.16
In the present case, the syn:anti- ratio was 1.4:1. In our
previously reported racemic system, the syn:anti- ratio was
2.4:1. The ratio of cycloaddition of the isoprene-derived
dienophile across the azadiene with respect to the methyl group
in the proline ring was 5:1, in the present case, favoring approach
of the dienophile from the same face of the azadiene (see B,
Scheme 2) as the methyl group and 1.47:1 in the racemic
O-methyl lactim ether system.8c While structural differences in
the laboratory azadienes (see B, Scheme 4) and that for the
biological system (see A, Scheme 2) indeed exist, the intrinsic
facial selectivity of this cycloaddition appears not to be greatly
biased toward the paraherquamide stereochemistry. In the
biological system, the diastereochemical distribution is expressed
as >600:117 (corresponding to 15 or 16:13 or 14) as evidenced
by the complete lack of natural metabolites that would arise
from substances containing the anti-stereochemistry16 imbedded
in either 24 or 25.
Conclusions
This study has served to confirm the absolute stereochemistry
of the putative shunt metabolite, VM55599, that is consistent
with a unified biogenesis of the paraherquamides proposed
earlier8 and corrects a misassignment reported in the initial
isolation and structure determination of this natural product.6 It
is now also clear that the absolute stereochemistry of the entire
paraherquamide family arises from the single stereogenic center
at the â-position of the â-methylproline moiety that is fashioned
from the L-Ile side chain.7,9 The laboratory intramolecular
Diels-Alder cycloaddition recorded here again demonstrates
an unexpected proclivity for the formation of the VM55599
relative stereochemistry which runs in sharp contradistinction
to the stereochemical preference expressed in Nature. The steric
effect of the methyl group in the â-methylproline ring is
apparently not as important as originally thought in biasing the
facial approach of the dienophile. Although the oxidation state
of the putative azadiene species (A/B, X ) O or H2) in the
biological system remains uncertain at this time, the contrast
between the two biomimetic laboratory cycloaddition reactions8c
and that postulated to occur in the biosynthetic constructions
strongly implicates protein organization of the precyclization
substrate conformation to greatly favor production of the
paraherquamide relative stereochemistry. Whether protein ca-
talysis of this construction occurs, remains an open question.
Efforts to address these remaining subtle, yet significant, issues
are the subject of ongoing studies in these laboratories.
Acknowledgment. This work was supported by the National
Institutes of Health (Grant CA70375 to R.M.W.). Mass spectra
were obtained on instruments supported by the National
Institutes of Health Shared Instrumentation Grant GM49631.
J.F.S.-C. thanks the DGICYT of Spain for a research grant
(Proyecto PB98-1438). We thank Professor Robert W. Woody
and Dr. A.-Young M. Woody of the Biochemistry and Molec-
ular Biology Department at Colorado State University for
invaluable help with CD spectra.
(16) The syn-/anti-relationship refers to the relative stereochemistry between
the C-20 stereogenic center (VM55599 numbering) and the cyclic amino
acid residue (proline, â-methylproline, or pipecolic acid):
Supporting Information Available: Complete experimental
procedures, spectroscopic and analytical data for all new
compounds (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA017425L
The syn-selective cycloaddition referred to in Scheme 2 would result in
diastereomers with the syn relative configuration as shown by the general
structure above.
(17) The relative amount of paraherquamide to VM55599 harvested from
Penicillium sp. IMI332995 is >600:1. This ratio does not include other
paraherquamides (1-12) that are typically co-isolated with 1.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2559