Synthesis and biological evaluation of platensimycin analogs

Article abstract of DOI:10.1016/j.bmcl.2009.02.006

Platensimycin (1) displays antibacterial activity due to its inhibition of the elongation condensing enzyme (FabF), a novel mode of action that could potentially lead to a breakthrough in developing a new generation of antibiotics. The medicinal chemistry

Full text of DOI:10.1016/j.bmcl.2009.02.006

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1623–1627
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and biological evaluation of platensimycin analogs
a
b
a
a
Hong C. Shen a, , Fa-Xiang Ding , Sheo B. Singh , Gopalakrishnan Parthasarathy , Stephen M. Soisson ,
Sookhee N. Ha a, Xun Chen c, Srinivas Kodali d, Jun Wang c, Karen Dorso d, James R. Tata a,
Milton L. Hammond d, Malcolm MacCoss a, Steven L. Colletti a
*
a Departments of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
b Natural Products, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
c Cardiovascular Diseases, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
d Infectious Diseases, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Platensimycin (1) displays antibacterial activity due to its inhibition of the elongation condensing
enzyme (FabF), a novel mode of action that could potentially lead to a breakthrough in developing a
new generation of antibiotics. The medicinal chemistry efforts were focused on the modification of the
enone moiety of platensimycin and several analogs showed significant activity against FabF and possess
antibacterial activity.
Received 18 December 2008
Revised 29 January 2009
Accepted 2 February 2009
Available online 7 February 2009
Ó 2009 Published by Elsevier Ltd.
Keywords:
Platensimycin
Platensimycin analogs
FabF
Inhibition of bacterial type II fatty-acid
synthesis
In 2006, the Merck Research Laboratories discovered platensi-
mycin (1, Fig. 1), a structurally unique natural product, from Strep-
tomyces platensis. Platensimycin exhibits remarkable antibacterial
activity against Gram-positive bacteria including multiple drug-
resistant strains such as vancomycin-resistant Enterococcus faecalis
(VREF) and methicillin-resistant Staphylococcus aureus (MRSA).1
The extraordinary antibacterial property of platensimycin is due
to its selective inhibition of b-ketoacyl-(acyl-carrier-protein
(ACP)) synthase I/II, also known as FabF/B condensing enzymes,
which play an essential role in the bacterial type II fatty-acid syn-
thesis (FASII). The novel scaffold and intriguing biological property
of platensimycin immediately captured the interests of numerous
research groups. Consequently, several elegant total syntheses
have been reported.2 In addition, medicinal chemistry studies have
been conducted by the Nicolaou group.3 Herein we report our ef-
forts to interrogate structure–activity relationship (SAR) and phar-
macophore of platensimycin, which could guide future design.
The X-ray crystallographic studies of platensimycin bound to
modified enzyme ecFabF(C163Q) allows the identification of sev-
eral critical hydrogen bonding and ionic interactions.1 Regarding
the benzoic acid fragment, carboxylate moiety formed salt bridge
interactions with two amino acid residues: H340 and H303, the
phenol hydroxy group showed water mediated H-bond interaction
with D265, and the anilide N–H group exhibited hydrogen bond
with T270 (Fig. 2). In addition, the benzene ring of platensimycin
poses an edge-to-plane
p-interaction with F400. Indeed, as the
Nicolaou group3a and we discovered,1c,d the absence of any of these
structural elements of platensimycin led to an almost complete
loss of activity. Based on the SAR and X-ray finding, it has been
hence concluded that the benzoic acid moiety acts as an essential
polar warhead to block the malonate binding pocket of the en-
zyme. As shown in Figure 3, the ketolide portion of platensimycin
is engaged in a hydrogen bond between the enone carbonyl oxygen
and A309. Moreover, a hydrogen bond between the cage ether oxy-
gen and T270 also contributes to the binding affinity.4 Figure 4 re-
veals the surface of the binding pocket, in which the ketolide is
exposed to solvent suggesting the feasibility of adding extra func-
tional groups to the enone. This observation, in addition to a poten-
tial covalent protein binding issue associated with the enone
group, led us to modify the enone moiety of platensimycin.
OH
H
O
N
O
O
HO
CO2H
1
Figure 1. Structure of platensimycin (1).
* Corresponding author. Tel.: +1 732 594 1755; fax: +1 732 594 9473.
E-mail address: hong_shen@merck.com (H.C. Shen).
0960-894X/$ - see front matter Ó 2009 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2009.02.006

1624
H. C. Shen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1623–1627
Furthermore, we envision that the poor pharmacokinetic profile of
platensimycin may be due in part to the enone component, which
may act as an electrophile in vivo causing its high clearance. Thus
by modifying the enone group, we hope to minimize the undesir-
able properties of platensimycin whereas retaining its antibacterial
activity.
We used three assays to monitor SAR study focusing on the en-
one modification of platensimycin. The assays were (a) whole cell
antibacterial activity using the wild type S. aureus, (b) inhibition of
FabF enzyme using cell free FASII assay,1f and (c) antisense fabH/
fabF sensitized strain of S. aureus (saFabF2 assay is the most sensi-
tive S. aureus antisense two-plate differential sensitivity cellular
assay.1e The data reported as MDC which is defined as the mini-
mum concentration of the compound showing differential zone
of clearance between antisense plates compared to control plate).
The data are summarized in Table 1. Occasional discrepancy
among these three assays can be attributed to the different nature
of these three assays. The data from saFabF2 (MDC) assay is the
cumulative inhibition result of both FabH and FabF, which share
common operon and as a result antisense knock down of one gene
knocks down both genes. Therefore, better sensitivity is generally
observed for the saFabF2 assay. On the other hand, FASII is a cell
free enzyme assay and does not capture this effect. MIC is the
inhibitory data from wild type sa whole cell assay and is much
coarser but most important.
Figure 2. Co-crystal structure (2.5 Å) of platensimycin (1, yellow) and ecFabF
(C163Q) showing the binding site. Key residues and hydrogen bonds involving the
aminobenzoic acid motif are labeled. (PDB code: 2GFX).
First the saturation of the olefin of the platensimycin enone pro-
vided analogs 2a–d. Although the olefin does not appear to be di-
rectly involved in interacting with the enzyme based on the X-ray
structure, the saturation of the double bond, however, caused a
minor conformational change as revealed by co-crystal structures
of 2aÁecFabF(C163Q) and 2cÁecFabF(C163Q). Figures 5 and 6 show
overlay of the crystal structures of platensimycin and dihydropla-
tensimycin (2a) and platensimycin and phenyl dihydroplatensimy-
cin (2c), respectively. The comparison suggest that the boat
conformation of platensimycin changed to the chair conformation
in the former case, and the twist chair in the latter case both result-
ing in a small but appreciable change of the positioning of ketolide
carbonyl as well as the aminobenzoic acid moiety. Consequently,
these changes led to a 4–8 fold loss of antibacterial activity against
wild type S. aureus and cell free FASII assay. These compounds
showed much profound loss of activity against the FabF2 antisense
assay. Thus, the binding pocket of platensimycin appears to be
quite rigid and does not accommodate substantial conformational
changes. It was then postulated that to increase of the sp2 charac-
Figure 3. Co-crystal structure (2.5 Å) of platensimycin (1, yellow) and ecFabF
(C163Q) showing the binding site. Key residues and hydrogen bonds involving the
tetracyclic enone acid and linker are labeled. (PDB code: 2GFX).
ter of the
a and/or b carbon of the ketolide carbonyl would be ben-
eficial to activity. Indeed, cyclopropyl analog 2d was fourfold more
active in FabF2 antisense assay than compound 2a and two fold
better in killing wt S. aureus. However, this substitution has five-
fold opposite effect against FASII assay.
Next, fused heterocyclic analogs containing a pyrazole or isox-
azole were examined to replace the electrophilic enone. The ra-
tional was to use hetero atoms of heterocycles to mimic the
enone carbonyl oxygen, and maintain the sp2 character of the ke-
tone a-carbon in platensimycin. Unfortunately, this strategy gener-
ated analogs 2e and 2f with much weaker activity. It is worth
noting that analog 2e was more active than 2f suggesting the pyr-
azole methyl substituent may disrupt the hydrogen bonding be-
tween the analog 2f and A309.
Using an heteroatom to replace the
a-carbon of the enone in
platensimycin, we also prepared lactone 2h and lactam 2i that dis-
played only very weak activities in all the assays presumably due
to the change of the carbonyl’s orientation. Finally, the fine-tuning
of platensimycin with minimum conformational change by adding
a methyl group at the b-position of enone resulted in analog 2g
that was equipotent as platensimycin in the antisense saFabF2 as-
say. The advantage of compound 2g is the more steric hindrance of
Figure 4. The binding pocket of platensimycin (1) in ecFabF (C163Q) (surface
coloring: red, oxygen; gray, carbon; blue, nitrogen; yellow, sulfur) (PDB code:
2GFX).

H. C. Shen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1623–1627
1625
Table 1
In vitro activity of platensimycin analogs
Compds
saMIC
g/mL)a
IC50, sa FASII
g/mL)
saFabF2 (MDC),
antisense, (lg/mL)
(
l
(
l
OH
H
N
O
HO
CO2H
O
O
0.5
2
0.5
2
0.01–0.04
O
1
0.4
O
O
2a
Figure 5. Overlay of 2.5 Å co-crystal structures of platensimycin (1, purple) and 2a
(yellow) and ecFabF (C163Q) showing the binding site. (PDB code: 3G0Y).
O
H
4
13
<20
2b
O
H
4
4
<20
0.1
0.3
O
2c
O
H
1
9.5
10.5
O
H
2d
O
16
O
N
2e
Figure 6. Overlay of 2.5 Å co-crystal structures of platensimycin (1, purple) and 2c
(yellow) and ecFabF (C163Q) showing the binding site. (PDB code: 3G11).
O
O
>64
75.3
2.4
>2.7
0.03
>2.7
>2.7
conformation. Future design should maximize the conformational
similarity of analogs to the cyclohexenone of platensimycin.
To obtain dihydroplatensimycin, a Pd/C-catalyzed hydrogena-
tion provided analog 2a in good yield (Scheme 1). Direct 1,4-addi-
tion to the enone of platensimycin with a large excess of Grignard
reagents provided the desired adducts 2b and 2c, respectively with
an excellent control of diastereoselectivity (>10:1). The observed
NMe
N
2f
4
O
2g
O
OH
OH
H
N
H
N
>64
>64
73.6
>200
O
O
a
b
O
O
O
2h
O
O
O
HO
HO
CO2H
CO2H
1
2a
O
c
N
O
OH
OH
H
H
2i
O
H
O
H
N
N
a
sa = Staphylococcus aureus.
O
O
O
O
HO
HO
CO2H
CO2H
2c
2b
the enone b-position would render the potential in vivo 1,4-addi-
tion pathway to enone less likely. Our SAR study strongly suggests
that the intricate interaction between platensimycin and the en-
zyme does not allow a significant change of the cyclohexenone
Scheme 1. Reagents and conditions: (a) 10% Pd/C, MeOH, rt, 16 h, 99%; reaction
performed as a Na salt (see Ref. 1c); (b) MeMgBr (10 equiv), THF–DME (2:1), rt, 10%;
(c) PhMgBr (15 equiv), THF–DME (1:1), rt to 50 °C, 16 h, 10%.

1626
H. C. Shen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1623–1627
diastereoselectivity is due to the bulky cage fragment that steri-
cally biased the 1,4-addition to proceed exclusively from the bot-
tom face.
2 and 3). It should be noted that the use of pyridine as the base
for the amide formation instead of triethyl amine gave substan-
tially higher yields.
The synthesis of other analogs commenced with the common
intermediate 3 derived from platensimycin via a mild amide
hydrolysis5 and methyl ester formation (Scheme 2). Subse-
The incorporation of a cyclopropyl group was achieved using
the Corey-Chaykovsky cyclopropanation (Scheme 4).6 The result-
ing cyclopropane intermediate 8 was obtained as a single diaste-
reomer, serving as a key intermediate leading to analog 2d
(Scheme 4).
To install a b-methyl group to platensimycin, the 1,4-methyl
addition to intermediate 8 followed by a Nicolaou’s dehydrogena-
tion protocol7 generated compound 9, which was eventually con-
verted to analog 2g over four steps (Scheme 5).8
quently, the introduction of an
a-formyl group with a concomi-
tant trans-esterification set the stage for condensation reactions
of intermediate 4 with hydroxylamine and methyl hydrazine to
generate isoxazole 5 (Scheme 2) and pyrazole 7 (Scheme 3),
respectively. The following hydrolysis, amide formation and
deprotection sequence then led to analogs 2e and 2f (Schemes
OH
H
O
N
a, b
c
O
O
OH
OMe
O
O
O
O
O
O
HO
CO2H
1
3
O
OEt
e
O
f
d
O
OMe
OEt
O
O
N
5
O
O
O
O
H
O
4
OMOM
CO2Me
OH
H
N
H2N
O
g-j
O
O
N
MOMO
HO
CO2H
6
2e
Scheme 2. Reagents and conditions: (a) NaNO2, HOAc, Ac2O, 0 °C to rt, 5 h, 60%; (b) LiOH, THF, H2O, rt, 16 h, 99%; (c) TMSCHN2, MeOH, 0 °C, 5 min, 90%; (d) 10% Pd/C, MeOH,
rt, 2 h, 97%; (e) NaH, EtOCHO, Et2O, rt, 16 h, 74%; (f) NH2OH HCl salt, EtOH, reflux, 16 h, 47%; (g) LiOH, THF, MeOH, H2O, rt, 2 h, 90%; (h) 6, HATU, pyridine, DMF, rt, 16 h, 29%;
(i) LiOH, THF, H2O, rt, 16 h; (j) HCl, THF, 40 °C, 24 h, 86% over two steps.
OH
H
b-e
O
N
a
O
O
OEt
OEt
O
NMe
HO
O
NMe
O
O
N
CO2H
H
N
O
2f
4
7
Scheme 3. Reagents and conditions: (a) NH2NHMe, EtOH, reflux, 16 h, 54%; (b) LiOH, THF, MeOH, H2O, rt, 2 h, 90%; (c) 6, HATU, pyridine, DMF, rt, 24 h, 38%; (d) LiOH, THF,
H2O, rt, 16 h; (e) HCl, THF, 40 °C, 16 h, 65% over two steps.
OH
H
N
a
b-d
O
O
OH
O
OMe
O
O
H
H
O
O
O
O
H
HO
H
CO2H
2d
3
8
Scheme 4. Reagents and conditions: (a) Me3S+IÀ, NaH, DMSO, rt, 16 h, 34%; (b) 6, HATU, pyridine, DMF, rt, 16 h, 38%; (c) LiOH, THF, H2O, rt, 16 h; (d) HCl, THF, 40 °C, 24 h, 50%
over two steps.
a
b
O
O
OMe
OMe
O
OMe
H
O
O
O
O
O
O
3
9
OH
H
N
c-f
O
O
O
HO
CO2H
2g
Scheme 5. Reagents and conditions: (a) MeLi (3 equiv), CuI (6 equiv), THF, 0 °C to rt, 1.5 h, 78%; (b) HIO3, cyclohexadiene, DMSO, 60 °C, 2 h, 20%; (c) LiOH, THF, H2O, rt, 3 h,
95%; (d) 6, HATU, pyridine, DMF, rt, 27%; (e) LiOH, THF, H2O, rt, 24 h; (f) HCl, THF, 40 °C, 24 h, 34% over two steps.

H. C. Shen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1623–1627
1627
b
a
O
OMe
O
O
OMe
O
OMe
O
H
O
O
O
O
OH
e-g
O
HO
O
OH
3
10
OH
H
c, d
O
N
OH
O
O
O
O
O
O
HO
CO2H
2h
11
Scheme 6. Reagents and conditions: (a) cat. OsO4, NMO, CH2Cl2, rt, 3 days; (b) NaIO4, acetone, water, rt, 4 days, 99% over two steps; (c) NaBH4, MeOH, rt, 1 h, then HCl, 49%;
(d) LiOH, THF, MeOH, H2O, rt, 4 h, 90%; (e) 6, HATU, pyridine, DMF, rt, 16 h, 19%; (f) LiOH, THF, H2O, rt, 16 h; (g) HCl, THF, 40 °C, 16 h, 53% over two steps.
OH
a
b-e
H
N
O
O
O
OMe
O
OMe
H
N
N
O
O
O
O
HO
O
O
HO
CO2H
10
2i
12
Scheme 7. Reagents and conditions: (a) MeNH2ÁHCl, NaBH(OAc)3, Et3N, THF/acetic acid (4:1), 16 h, rt, 37%; (b) LiOH, THF, MeOH, H2O, rt, 1 h, 99%; (c) 6, HATU, pyridine, DMF,
rt, 16 h, 30%; (d) LiOH, THF, H2O, rt, 16 h; (e) HCl, THF, 40 °C, 16 h, 12% over two steps.
K.; Wang, J. J. Am. Chem. Soc. 2006, 128, 11916; (c) Singh, S. B.; Herath, K. B.;
Wang, J.; Tsou, N.; Ball, R. G. Tetrahedron Lett. 2007, 48, 5429; (d) Zhang, C.;
Ondeyka, J.; Zink, D. L.; Burgess, B.; Wang, J.; Singh, S. B. Chem. Commun. 2008,
The replacement of cyclohexenone of platensimycin with a lac-
tone or lactam involved the oxidative degradation of the double
bond that produced aldehyde acid 10 (Scheme 6). The reduction
of aldehyde to alcohol, followed by an acid-mediated lactone for-
mation allowed the formation of lactone intermediate 11, which
was subsequently converted to product 2h. In a similar fashion,
the sequential reductive amination of compound 10 and lactam
formation yielded intermediate 12 during the synthesis of analog
2i (Scheme 7).
In conclusion, several platensimycin analogs were prepared to
test the hypothesis whether the enone moiety was essential for
platensimycin’s biological activity. The SAR strongly suggests that
the conformation of the enone moiety is crucial in positioning
the ketone carbonyl group to interact with A309 of FabF. Of all
the analogs prepared, those bearing the ketone moiety and similar
conformation, such as compound 2d and 2g, were generally quite
active in the saFabF2 antisense assay but none were more active
than platensimycin.
5034; (e) Young, K. et al Antimicrob. Agents Chemother. 2006, 50, 519; (f) Kodali,
S.; Galgoci, A.; Young, K.; Painter, R.; Silver, L. L.; Herath, K. B.; Singh, S. B.; Cully,
D.; Barrett, J. F.; Schmatz, D.; Wang, J. J. Biol. Chem. 2005, 280, 1669.
2. (a) Nicolaou, K. C.; Li, A.; Edmonds, D. J. Angew. Chem., Int. Ed. 2006, 45, 7086;
(b) Nicolaou, K. C.; Edmonds, D. J.; Li, A.; Tria, G. S. Angew. Chem., Int. Ed.
2007, 46, 3942; (c) Zou, Y.; Chen, C.-H.; Taylor, C. D.; Foxman, B. M.; Snyder,
B. B. Org. Lett. 2007, 9, 1825; (d) Nicolaou, K. C.; Tang, Y.; Wang, J. Chem.
Commun. 2007, 1922; (e) Li, P.; Payette, J. N.; Yamamoto, H. J. Am. Chem. Soc.
2007, 129, 9534; (f) Lalic, G.; Corey, E. J. Org. Lett. 2007, 9, 4921; (g)
Tiefenbacher, K.; Mulzer, J. Angew. Chem., Int. Ed. 2007, 46, 8074; (h) Nicolaou,
K. C.; Pappo, D.; Tsang, K. Y.; Gibe, R.; Chen, D. Y.-K. Angew. Chem., Int. Ed.
2008, 47, 944; (i) Kim, C. H.; Jang, K. P.; Choi, S. Y.; Chung, Y. K.; Lee, E.
Angew. Chem., Int. Ed. 2008, 47, 4009.
3. (a) Nicolaou, K. C.; Stepan, A. F.; Lister, T.; Li, A.; Montero, A.; Tria, G. S.; Turner,
C. I.; Tang, Y.; Wang, J.; Denton, R. M.; Edmonds, D. J. J. Am. Chem. Soc. 2008, 130.
ASAP; (b) Nicolaou, K. C.; Tang, Y.; Wang, J.; Stepan, A. F.; Li, A.; Montero, A. J.
Am. Chem. Soc. 2007, 129, 14850; (c) Nicolaou, K. C.; Lister, T.; Denton, R. M.;
Montero, A.; Edmonds, D. J. Angew. Chem., Int. Ed. 2007, 46, 4712.
4. As shown in Ref. 3a, all-carbon cage analogs carbaplatensimycin and
adamantaplatensimycin both gave about threefold less of activity with respect
to platensimycin.
5. Mandai, T.; Yamakawa, T. Synlett 2000, 862.
6. Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353.
References and notes
7. Nicolaou, K. C.; Montagnon, T.; Baran, P. S. Angew. Chem., Int. Ed. 2002, 41, 1386.
8. For analog 2g: 1H NMR (acetone-d6, 500 MHz) d 10.6 (1H, br s), 9.11 (1H, s), 7.65
(1H, d, J = 6.0 Hz), 6.48 (1H, d, J = 6.0 Hz), 5.79 (1H, s), 4.44 (1H, s), 2.66 (2H, m),
2.38 (4H, m), 2.10 (7H, m), 1.89 (3H, s), 1.70 (1H, m), 1.41 (3H, s), 1.24 (3H, s);
LCMS m/z: 456.19 (M+1), 478.21 (M+Na). For analog 2h: 1H NMR (acetone-d6,
500 MHz) d 10.2 (1H, br s), 8.96 (1H, s), 7.68 (1H, br s), 6.31 (1H, bs), 4.50 (1H, d,
J = 11.0 Hz), 4.38 (1H, s), 3.88 (1H, d, J = 11.0 Hz), 2.63 (2H, m), 2.32 (5H, m), 2.05
(5H, m), 1.99 (3H, s), 1.68 (1H, m), 1.53 (1H, m), 1.38 (3H, s); LCMS m/z: 446.52
(M+1).
1. (a) Wang, J.; Soisson, S. M.; Young, K.; Shoop, W.; Kodali, S.; Galgoci, A.; Painter,
R.; Parthasarathy, G.; Tang, Y. S.; Cummings, R.; Ha, S.; Dorso, K.; Motyl, M.;
Jayasuriya, H.; Ondeyka, J.; Herath, K.; Zhang, C.; Hernandez, L.; Allocco, J.;
Basilio, A.; Tormo, J. R.; Genilloud, O.; Vicente, F.; Pelaez, F.; Colwell, L.; Lee, S. H.;
Michael, B.; Felcetto, T.; Gill, C.; Silver, L. L.; Hermes, J. D.; Bartizal, K.; Barrett, J.;
Schmatz, D.; Becker, J. W.; Cully, D.; Singh, S. B. Nature 2006, 441, 358; (b) Singh,
S. B.; jayasuriya, H.; Ondeyka, J. G.; Herath, K. B.; Zhang, C.; Zink, D. L.; Tsou, N.
N.; Ball, R. G.; Basilio, Á.; Genilloud, O.; Diez, M. T.; Vicente, F.; Palaez, F.; Young,