Efforts towards the identification of simpler platensimycin analogues-the total synthesis of oxazinidinyl platensimycin

Article abstract of DOI:10.1002/chem.200802568

The total synthesis and biological testing of oxazinidinyl platensimycin, the methyl ester analogues, and the simplified oxazinidinyl analogue were investigated to identify simpler platensimycin analogues. The study also investigated the possibilities of replacing enone moiety of platensimycin with the oxazinidinyl ring to create a potential bioisostere compound. The study examined the possibilities to replace the tetracyclic core structure of platensimycin with other motifs that are synthetically easier to access. Proposed methods can applied to the construction of other polycyclic motifs found in the other natural or non-natural molecules. The results show that the enone moiety of platensimycin's core structure is sensitive to modifications.

Full text of DOI:10.1002/chem.200802568

COMMUNICATION
DOI: 10.1002/chem.200802568
Efforts towards the Identification of Simpler Platensimycin Analogues—The
Total Synthesis of Oxazinidinyl Platensimycin
Jingxin Wang,[a] Vincent Lee,[b] and Herman O. Sintim*[a]
The discovery of platensimycin (1; Figure 1),[1] a novel an-
tibiotic produced by Streptomyces platensis, has generated
Analysis of the X-ray crystal structure of platensimycin in
a complex with a mutant version of the ecFabF enzyme re-
veals several polar interactions between the highly polar
benzoic acid moiety of 1 and the FabF residues.[1] Also, the
tetracyclic core of compound 1 makes both polar and van
der Waals interactions within the enzymeꢀs active site. The
extensive interactions between platensimycin and the FabF
enzyme account for the high affinity of platensimycin to the
FabF enzyme.[1] Platensimycin has a low in vivo activity
when administered by conventional routes.[1] This has been
attributed to the drugꢀs poor pharmacokinetic properties.
Because of platensimycinꢀs poor pharmacokinetics, we and
others have initiated programs to study the structure–activi-
ty relationship of platensimycin with the aim of identifying
parts of the molecule that are amenable to modifications or
deletion without adversely affecting biological activity.[4]
These studies should ultimately lead to simpler analogues
that are easier to synthesize and also with enhanced phar-
macokinetics.
Figure 1. Structures of platensimycin and analogues and interactions of
the C5 carbonyl moiety of 1 and 2 with the amide hydrogen atom of
A309 of ecFabFACHTUNGTRENNUNG(C163Q).
Herein, we report the total synthesis and biological testing
of oxazinidinyl platensimycin 2, the methyl ester analogue
32, and the simplified oxazinidinyl analogue 4 (Figure 1).
This limited set of analogues was designed to answer two
questions: a) Could the enone moiety of platensimycin be
replaced with the oxazinidinyl ring to create a potential bio-
isostere compound 2?[5] b) Could the tetracyclic core struc-
ture of platensimycin be replaced with other motifs that are
synthetically easier to access?
excitement in the scientific community due to its potent an-
tibacterial activity against Gram-positive bacteria including
vancomycin-resistant Enterococcus faecalis (VREF) and me-
thicillin-resistant Staphylococcus aureus (MRSA). Platensi-
mycin, a specific inhibitor of the bacterial FabF enzyme, is
the first new class of antibiotic discovered in more than four
decades.[2] Due to platensimycinꢀs unprecedented complex
architecture and remarkable in vitro activity, several groups
have embarked on the total syntheses of this molecule.[3]
We envisioned that the core structure of oxazinidinyl pla-
tensimycin 2 could be obtained by the strategy outlined in
Scheme 1. To afford tricycle 10 by pathway a as opposed to
[a] J. Wang, Prof. H. O. Sintim
Department of Chemistry and Biochemistry
University of Maryland, College Park, MD 20742 (USA)
Fax : (+1)301-314-9121
E-mail: hsintim@umd.edu
[b] Prof. V. Lee
Department of Cell Biology and Molecular Genetics
University of Maryland, College Park, MD 20742 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802568.
Scheme 1. Retrosynthetic analysis of oxazinidinyl platensimycin.
Chem. Eur. J. 2009, 15, 2747 – 2750
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2747

forming 11 by pathway b, a regioselective epoxide ring-
opening of intermediate 9 to afford tricycle 10 was desired
(see Scheme 2). DFT calculations indicated that the activa-
tion energy of pathway a (Scheme 2) was 9.8 kcalmolꢀ1
pound 16 to give 8. Subjecting compound 16 to 30 mol%
DBU in toluene yielded an epimeric mixture of 16: 8 in a
10:8 ratio after 21 h (see the Supporting Information). It
therefore appears that the undesired bis-alkene 16 is both
the kinetic and thermodynamic product.
We wondered if we could perform the ring-closing meta-
thesis in the presence of DBU. Should the ruthenium cata-
lyst be compatible with the amine base, then a dynamic
ring-closing resolution should ensue and both diastereomers
8 and 16 would be suitable for the ring-closing reaction. To
the best of our knowledge, a dynamic ring-closing metathe-
sis involving epimerization has not been reported in the lit-
erature.
Scheme 2. 6-exo-tet (pathway a) versus 5-exo-tet cyclization (pathway b).
lower than that of pathway b (see the Supporting Informa-
tion). This suggests that pathway a is more favorable than
pathway b. Also, for the optimized ground state geometry of
compound 9, the distance between the oxygen nucleophile
and the C1 atom was 3.2 ꢂ, whereas the oxygen nucleophile
was 3.7 ꢂ away from C2. Together, these analyses gave us
the confidence that one of our key steps towards the synthe-
sis of oxazinidinyl platensimycin 2 was viable and worth pur-
suing.
Pleasingly, subjecting an epimeric mixture of compounds
8:16 (ratio of 1:3.7) to Hoveyda–Grubbs II catalyst in the
presence of DBU and benzoquinone afforded the ring-
closed product in 69% yield (83% based on the starting ma-
terial; Scheme 4). In contrast, in the absence of DBU, the
desired product could be obtained in a meager 20% yield.
The benzoquinone additive was important for minimizing
the formation of the enone by-product 19.[6]
Initially, we set out to make bis-alkene 8, the requisite
substrate for the ring-closing metathesis reaction. Treatment
of commercially available vinylogous ester 12 with anion 13,
generated from a stannane precursor by a lithium–tin ex-
change gave enone 14 in good yield (84%). The next step
involved the allylation of enone 14 with allyl halide. Allyl
bromide gave the desired product in only a meager 34%
yield accompanied by a substantial amount of diallylated
product 17 (44% yield). Interestingly, changing the allyla-
tion reagent to allyl iodide improved the yield to 62%, and
the formation of the diallylated product was somewhat sup-
pressed. Conjugate addition of vinyllithium to enone 15, in
the presence of BF3·Et2O resulted in a 1:3.7 mixture of race-
mic 8 and 16, respectively (Scheme 3). Unfortunately, the
major product 16 was not suitable for the subsequent ring-
closing metathesis because the alkene moieties were trans to
each other. The low overall yield of the desired cis-bis-
alkene 8 led us to investigate the epimerization of com-
Scheme 4. Dynamic ring-closing metathesis. Reaction conditions:
Grubbs–Hoveyda II catalyst (five portions; each portion=3.2 mol%),
DDQ (20 mol%), DBU (46 mol%), toluene, reflux, 69% (83% based on
starting material). DDQ=2,3-dichloro-5,6-dicyanobenzoquinone, DBU=
1,8-diazabicyclo-[5.4.0]undec-7-ene.
With bicyclic compound 18 in hand, in gram quantities,
we proceeded with the epoxidation reaction using mCPBA.
The epoxide 20, obtained in 64% yield, was then subjected
to a tandem nucleophilic MeLi addition followed by a sub-
sequent epoxide ring-opening to afford tricycle 21 in 70%
yield. In line with our expectation, the alternative tricycle 11
(see Scheme 2) was not obtained. Oxidation of compound
21 with Dess–Martin periodinane was followed by debenzy-
lation with Pd/H2 to afford 22 without any incident
(Scheme 5). The aromatic side chain of compound 4 was
synthesized following the strategy shown in Scheme 6.
The stage was now set to couple the aromatic side chain
28 with the tricyclic ring core 22. The reductive amination
intermediate 29 has two sterically similar faces (shown as
faces A and B in Figure 2) and it appeared at first glance
that a mixture of diastereomers would be obtained during
the reductive amination step. We however postulated that
because one face (face B) contained oxygen, subtle stereo-
electronic factors could swing the selectivity towards the de-
sired product. The Felkin–Anh model[7] for the addition of
nucleophiles to iminium 29 favors addition from face A
(Figure 2) due to favorable interactions between the nucleo-
Scheme 3. a) LiCH2OBn, THF, ꢀ788C!RT, 12 h, 84 %; b) LDA,
HMPA, allyl iodide, 62% (67% based on starting material); c) Sn-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
2748
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2747 – 2750

Simpler Platensimycin Analogues
COMMUNICATION
phile and the low-lying bond. However, literature exam-
ples[8] indicated that for tricyclic systems such as 29, a Cie-
plak model[9] would be a more appropriate predictor for
product distribution: A hydride approach from face B
would lead to two sCꢀC interacting favorably with the devel-
oping bond, whereas a hydride approach from face A will
only lead to one sCꢀC interacting favorably with the develop-
ing s*CꢀH bond. sCꢀO bonds are low lying and not expected
to make any significant hyperconjugative contribution. Ad-
ditionally, transition-state dipole minimization between the
ꢀ
ꢀ
C O bond and the forming C N bond dictates that ap-
proach from face B should predominate.[10] Together, these
arguments strengthened our conviction that the desired
product would be obtained by reductive amination. Model
studies using a simplified side chain confirmed our predic-
tion that hydride attack from face B predominates (see the
Supporting Information).
Scheme 5. a) mCPBA, dichloromethane, 08C!RT, 12 h, 64%; b) MeLi,
THF, ꢀ788C!RT, 5 h, 70%; c) Dess–Martin reagent, dichloromethane,
12 h, 95%; d) 10 mol%, Pd/C, H2, MeOH, RT, 12 h, 90%. mCPBA=3-
chloroperoxybenzoic acid.
The end game of our synthesis proceeded smoothly. Re-
acting compounds 22 and 28 in methanol in the presence of
sodium cyanohydride gave compound 30; which was not pu-
rified but subjected to carbonyldiimidazole cyclization to
afford 31 in 76% yield. Deprotection of the methyl ester
and the MOM groups led to oxazinidinyl platensimycin 2 in
80% yield over two steps. Oxazinidinyl cyclohexaplatensi-
mycin 4 was synthesized following the strategy outlined in
Scheme 7 (see the Supporting Information).
Scheme 6. a) NIS, TFA, RT, 12 h, 100%, then, MOMCl, Hꢃnig base, di-
chloromethane, RT, 12 h, 97% ; b) CNCO2Me, PhMgBr, ꢀ788C!RT,
THF, 12 h, 78%; c) MOMCl, Hꢃnig base, dichloromethane, RT, 12 h,
78%; d) H2, 10 mol% Pd/C, MeOH, 69%; e) pyridine, benzene, RT!
508C, 5 h, 95%; f) K2CO3, MeOH: H2O=3: 1, RT, 10 h, 99%. NIS=N-
iodoACHTUNGTRENNUNGsuccinimide, TFA=trifluoroacetic acid.
Scheme 7. a) NaBH3(CN), MeOH, pH 4, RT, 5 h, 75%; b) (im)2CO,
20 mol% DMAP, benzene, 508C, 12 h, 76%; c) 1n LiOH, MeOH:
H2O=2:1, RT, 12 h, then, 2n HCl, MeOH: H2O=2:1, RT, 24 h, 80%; d)
2n HCl, MeOH:H2O=3:1, RT, 24 h, 92%. DMAP=4-dimethylamino-
pyridine.
Oxazinidinyl platensimycin 2 has a minimum inhibitory
concentration (MIC) of 90 mgmLꢀ1[11] for all of the three
bacterial strains tested; Staphylococcus aureus (Newman),
Streptococcus agalactiae (2603 V/R), and Bacillus subtillis
(3160). MICs of oxazinidinyl platensimycin methyl ester 32
and the platensimycin analogue with simpler core structure
4 were greater than 512 mgmLꢀ1 for all of the three bacterial
strains.
Despite its structural mimicry of platensimycin, oxazini-
dinyl platensimycin 2 is two orders of magnitude less potent.
Figure 2. Facial selectivity of the reductive amination reaction.
Chem. Eur. J. 2009, 15, 2747 – 2750
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2749

H. O. Sintim et al.
C. D. Taylor, B. M. Foxman, B. B. Snider, Org. Lett. 2007, 9, 1825–
1828; f) P. Li, J. N. Payette, H. Yamamoto, J. Am. Chem. Soc. 2007,
129, 9534–9535; g) K. P. Kaliappan, V. Ravikumar, Org. Lett. 2007,
9, 2417–2419; h) A. K. Ghosh, K. Xi, Org. Lett. 2007, 9, 4013–4016;
i) K. Tiefenbacher, J. Mulzer, Angew. Chem. 2007, 119, 8220–8221;
Angew. Chem. Int. Ed. 2007, 46, 8074–8075; j) C. H. Kim, K. P.
Jang, S. Y. Choi, Y. K. Chung, E. Lee, Angew. Chem. 2008, 120,
4073–4075; Angew. Chem. Int. Ed. 2008, 47, 4009–4011.
[4] For the biological activity of recently synthesized platensimycin ana-
logues, see: a) K. C. Nicolaou, T. Lister, R. M. Denton, A. Montero,
D. J. Edmonds, Angew. Chem. 2007, 119, 4796–4798; Angew. Chem.
Int. Ed. 2007, 46, 4712–4714; b) K. C. Nicolaou, A. F. Stepan, T.
Lister, A. Li, A. Montero, G. S. Tria, C. I. Turner, Y. Tang, J. Wang,
R. M. Denton, D. J. Edmonds, J. Am. Chem. Soc. 2008, 130, 13110–
13119.
[5] The carbamate moiety of the oxazinidinyl ring can maintain the hy-
drogen-bonding interactions with A309 of ecFabF. The oxazinidinyl
ring is however not strictly an isosteric replacement of the enone
ring core of platensimycin due to the lack of the C4 methyl group
found in the parent compound.
[6] S. H. Hong, D. P. Sanders, C. W. Lee, R. H. Grubbs, J. Am. Chem.
Soc. 2005, 127, 17160–17161.
[7] a) M. Chꢄrest, H. Felkin, N. Prudent, Tetrahedron Lett. 1968, 9,
2199–2204; b) A. Mengel, O. Reiser, Chem. Rev. 1999, 99, 1191–
1223.
Carbaplatensimycin 3, an analogue that has the ethereal
oxygen of platensimycin replaced with carbon, has a MIC of
1–2 mgmLꢀ1 (MRSA), which is close to that of the parent
compound.[4b] Our work and that of the Nicolaou and the
Merck groups suggest that the enone moiety of platensimy-
cin is more important than the ethereal oxygen with regard
to antibiotic activity.[4b,12] The lack of antibiotic potency of
the methyl ester analogue 32, even at millimolar concentra-
tions, suggests that bacterial esterases are unable to hydro-
lyze the ester analogues of platensimycin into active
drugs.[13]
In conclusion, we have presented novel strategies such as
the dynamic ring-closing metathesis reaction and nucleophil-
ic addition to a ketone followed by tandem epoxide ring-
opening reaction to construct the tricyclic core structure
found in platensimycin. These strategies should be applica-
ble to the construction of other polycyclic motifs found in
other natural or non-natural molecules. Our synthesis of ox-
azinidinyl platensimycin is concise and involves only 11
steps, starting from commercially available vinylogous ester
12. We also show that the enone moiety of platensimycinꢀs
core structure is sensitive to modifications.
[8] a) J. M. Hahn, W. J. Le Noble, J. Am. Chem. Soc. 1992, 114, 1916–
1917; b) M. Xie, W. J. Le Noble, J. Org. Chem. 1989, 54, 3836–3839;
c) G. Mehta, F. A. Khan, J. Am. Chem. Soc. 1990, 112, 6140–6142;
d) G. Mehta, F. A. Khan, B. Ganguly, J. Chandrasekhar, Chem.
Commun. 1992, 1711–1712; e) R. L. Halterman, M. A. McEvoy, J.
Am. Chem. Soc. 1990, 112, 6690–6693.
Acknowledgements
[9] a) A. S. Cieplak, J. Am. Chem. Soc. 1981, 103, 4540–4552; b) A. S.
Cieplak, B. D. Tait, C. R. Johnson, J. Am. Chem. Soc. 1989, 111,
8447–8462.
[10] The reductive amination was conducted in the polar protic solvent,
MeOH; therefore dipole minimization might not play as great a role
in product distribution.
This work was supported by the University of Maryland, College Park.
We thank Dr. Y. Wang and Prof. A. Vedernikov for help with computa-
tional work. We also thank an anonymous reviewer for helpful com-
ments.
[11] Using the conventional and usual twofold dilution method, the MIC
Keywords: antibiotics · enzymes · ring-closing metathesis ·
medicinal chemistry · platensimycin · total synthesis
of oxazinidinyl platensimycin
128 mgmLꢀ1 because the determined MIC value of 90 mgmLꢀ1 lies
between 64 and 128 mgmLꢀ1
2 for all the strains would be
.
[12] Dihydroplatensimycin, an analogue that lacks the enone moiety, is
less active than platensimycin. Also, this keto analogue is unstable
and forms cyclized products. See: S. B. Singh, H. Jayasuriya, J. G.
Ondeyka, K. B. Herath, C. Zhang, D. L. Zink, N. N. Tsou, R. G.
Ball, A. Basilio, O. Genilloud, M. T. Diez, F. Vicente, F. Pelaez, K.
Young, J. Wang, J. Am. Chem. Soc. 2006, 128, 11916–11920; S. B.
Singh, H. Jayasuriya, J. G. Ondeyka, K. B. Herath, C. Zhang, D. L.
Zink, N. N. Tsou, R. G. Ball, A. Basilio, O. Genilloud, M. T. Diez, F.
Vicente, F. Pelaez, K. Young, J. Wang, J. Am. Chem. Soc. 2006, 128,
15547.
[1] J. Wang, S. M. Soisson, K. Young, W. Shoop, S. Kodali, A. Galgoci,
R. Painter, G. Parthasarathy, Y. S. Tang, R. Cummings, S. Ha, K.
Dorso, M. Motyl, H. Jayasuriya, J. Ondeyka, K. Herath, C. Zhang,
L. Hernandez, J. Allocco, A. Basilio, J. R. Tormo, O. Genilloud, F.
Vicente, F. Pelaez, L. Colwell, S. H. Lee, B. Michael, T. Felcetto, C.
Gill, L. L. Silver, J. D. Hermes, K. Bartizal, J. Barrett, D. Schmatz,
J. W. Becker, D. Cully, S. B. Singh, Nature 2006, 441, 358–361.
[2] C. Walsh, Nat. Rev. Microbiol. 2003, 1, 65–70.
[3] For a review, see: a) K. Tiefenbacher, J. Mulzer, Angew. Chem.
2008, 120, 2582–2590; Angew. Chem. Int. Ed. 2008, 47, 2548–2555;
for total syntheses and synthetic studies towards platensimycin, see:
b) K. C. Nicolaou, A. Li, D. J. Edmonds, Angew. Chem. 2006, 118,
7244–7248; Angew. Chem. Int. Ed. 2006, 45, 7086–7090; c) K. C.
Nicolaou, D. J. Edmonds, A. Li, G. S. Tria, Angew. Chem. 2007, 119,
4016–4019; Angew. Chem. Int. Ed. 2007, 46, 3942–3945; d) G. Lalic,
E. J. Corey, Org. Lett. 2007, 9, 4921–4923; e) Y. Zou, C.-H. Chen,
[13] Platensimycin methyl ester is also not a potent antibiotic, see: S. B.
Singh, K. B. Herath, J. Wang, N. Tsou, R. G. Ball, Tetrahedron Lett.
2007, 48, 5429–5433.
Received: December 8, 2008
Revised: January 28, 2009
Published online: February 11, 2009
2750
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2747 – 2750