Design at the atomic level: Design of biaryloxazolidinones as potent orally active antibiotics

Article abstract of DOI:10.1016/j.bmcl.2008.10.011

We have developed a first generation of hybrid sparsomycin-linezolid compounds into a new family of orally bioavailable biaryloxazolidinones that have activity against both linezolid-susceptible and -resistant Gram-positive bacteria as well as the fastidi

Full text of DOI:10.1016/j.bmcl.2008.10.011

Bioorganic & Medicinal Chemistry Letters 18 (2008) 6175–6178
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design at the atomic level: Design of biaryloxazolidinones as potent orally
active antibiotics
Jiacheng Zhou b,––, Ashoke Bhattacharjee b, Shili Chen b, Yi Chen b,àà, Erin Duffy a, Jay Farmer b,
,
Joel Goldberg b,à, Roger Hanselmann b, Joseph A. Ippolito a, Rongliang Lou b, Alia Orbin b,§, Ayomi Oyelere b,–
,
Joe Salvino b,||, Dane Springer b, , Jennifer Tran b,àà, Deping Wang a,§§, Yusheng Wu b, Graham Johnson a,b,
*
a Department of Structure-Based Drug Design, Rib-X Pharmaceuticals Inc., 300 George Street, Suite 301, New Haven, CT 06511, USA
b Department of Medicinal Chemistry, Rib-X Pharmaceuticals Inc., 300 George Street, Suite 301, New Haven, CT 06511, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We have developed a first generation of hybrid sparsomycin–linezolid compounds into a new family of
orally bioavailable biaryloxazolidinones that have activity against both linezolid-susceptible and -resis-
tant Gram-positive bacteria as well as the fastidious Gram-negative bacteria Haemophilus influenzae
and Moraxella catarrahalis. The convergent synthesis of these new compounds is detailed.
Ó 2008 Elsevier Ltd. All rights reserved.
Received 12 May 2008
Revised 1 October 2008
Accepted 2 October 2008
Available online 7 October 2008
Keywords:
Oxazolidinone
Biaryloxazolidinone
Structure-based drug design
Linezolid
Sparsomycin
Gram-negative bacteria
X-ray crystal structure
Ribosome
Oral antibiotics
Hybrid antibiotics
In the preceding letter1 we outlined how the X-ray struc-
tures of linezolid and sparsomycin bound to the 50S ribosome
were used to design a new family of oxazolidinones. Specifi-
cally, we described how the previously unknown proximity
and juxtaposition of the linezolid and sparsomycin-binding sites
led us to the design of new compounds to bridge and occupy
the binding sites of both of these antibiotics. In this letter, we
describe how these extended and somewhat complex hybrid
molecules were transformed into potent and selective next gen-
eration biaryloxazolidinone antibiotics with an expanded anti-
bacterial spectrum.
Compound 1 described in our companion letter1, was pre-
pared using an alkenyl bridge element to connect the desired
thymine ring of sparsomycin and the phenyloxazoldinone ele-
ment of linezolid. Close inspection of the structure of compound
1 in complex with the 50S ribosomal subunit1 suggested that the
placement of the terminal thymine ring could be further opti-
mized. As shown in Figure 1, the condensed heteroaryl face of
the key interacting partner of the 50S ribosomal subunit, Ade-
nine 2602 (colored white; Escherichia coli numbering of 23S
* Corresponding author. Tel.: +1 203 848 6934; fax: +1 203 624 5627.
E-mail address: gjohnson@rib-x.com (G. Johnson).
Present address: 1517 Caywood Road, Lodi, NY 14860, USA.
Present address: Wyeth Research, CN 8000, Princeton, NJ 08543, USA.
Present address: 4890 Briarwood Drive, Macungie, PA 18062, USA.
Present address: Georgia Tech, 3305 IBB Building, GA 30332, USA.
rRNA), participates in an offset
p stack interaction with the ter-
à
§
minal thymine element in Compound 1. Computational studies
indicated that this is not the optimal relative orientation for a
–
||
Present address: Cephalon Inc., 383 Phoenixville Pike, Malvern, PA 19355, USA.
p–p
interaction. As such, we prepared three analogs (2, 3, and
4) on the biaryl scaffold (Fig. 2) in which we had sought to opti-
mize the overlap between the terminal heterocycle and the
Present address: Goodwin and Procter, 53 State Street, Boston, MA 02109, USA.
Present address: 91 Golden Hill Drive, Guilford, CT 06437, USA.
Present address: Biogen Idec, 14 Cambridge Center, Cambridge, MA 02142, USA.
Present address: Incyte Corporation, Experimental Station, E336, Route 141 and
àà
p
§§
––
ribosomal base. Specifically we focused our attention on both
shortening the chemical functionality bridging the terminal
Henry Clay Road, Wilmington, DE 19880, USA.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.10.011

6176
J. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6175–6178
O
O
NH
O
O
HN
N
NH
O
F
NHAc
1
Figure 1. Key interaction of compound 1 (yellow) with Adenine 2602 (A2602) (white) of the 50S ribosomal binding site.
O
O
O
O
O
O
N
N
N
NH
NH
NH
F
F
F
O
NHAc
NHAc
NHAc
N
N
2
3
4
Figure 2.
aromatic group to the biaryloxazolidinone and on modifying the
nature of the terminal aromatic group.
Table 1
Optimization of terminal aromatic interactions
Table 1 shows the results from our efforts to optimize antibi-
otic–ribosome interactions. In all cases, a simple benzylic amine
functionality delivered the aromatic rings for productive interac-
tions with A2602, and simple heterocycles such as 3 were as po-
tent as condensed heterocycles. Exquisite potency across the
streptococci was maintained by all three compounds (2, 3, and
4). In addition, all were potent against the linezolid-resistant
enterococcal strain. Importantly, all analogs showed potency
against Haemophilus influenzae RD1, a moderately susceptible
clinical strain, although their MIC90s (not shown) were not opti-
Linezolid
2
3
4
Intrinsic affinity
E. coli translation IC50
Selectivity
(
l
M)
4.6
Y
<0.02
Y
0.05
Y
0.04
Y
MIC (lg/ml)
S. pneumoniae 02J1175 mef (A)
2
1
32
16
8
0.5
0.5
2
8
1
0.2 5
0.5
2
8
0.5
0.25
0.2 5
4
4
0.5
S. pyogenes Msr 610 erm (B)
E. faecalis P5 (linezolidÒ
)
H. influenzae parent strain (RD1)
H. influenzae 895 (AcrB-KO)
All minimal inhibitory concentration (MIC) determinations were carried out under
NCCLS conditions.3 Ribosomal translation was carried out as described in our
companion letter.1 All compounds were greater than 100-fold selective for E. coli
over rabbit reticulocyte enzymes.
mal (8–16 lg/ml). Their molecular properties, in particular the
predictions for Caco-2 cell permeability, a surrogate for oral bio-
availability, were significantly worse than linezolid and therefore
were expected to have limited oral bioavailability. This liability
was confirmed in the comparatively poor in vivo efficacy of
these compounds in
shown).
a murine peritonitis model (data not
cacy. In some cases this design strategy led us away from the incor-
poration of a terminal heterocycle (compounds 9 and 11). In order to
be effective, we devised synthetic routes and key intermediates that
would allow us to probe several lines of analoging in a parallel and
efficient manner. Using these synthetic approaches we designed
and generated analogs 5–11 (Fig. 3).ààà As noted earlier1 the highest
energy interaction with the ribosome are associated with the biaryl
and oxazolidinone residues. From our computational analysis, the
diversity of remaining residues incorporated in compounds 5–11
are readily accommodated by the ribosome but do not provide for
specific interactions. Their major influence is therefore on
modifying the various molecular properties associated with cell
penetration.
Balancing the biaryl scaffold for candidate compounds: In con-
junction with a series of amino acid biaryl oxazolidinones (details
to be published separately), these simplified analogs formed the
base for a predictive computational model that described activity
against H. influenzae. The model is a simple six-descriptor model
that requires a balance between factors that contribute to both affin-
ity and those that speak to drug-like properties (data not shown).
Combining this model with accurate Qikprop2 predictions for
Caco-2 cell permeability and aqueous solubility as well as being
guided by our structural knowledge within the ribosome, we initi-
ated an optimization program. Our goals were to maintain or im-
prove activity across the Gram-positive bacteria, boost activity
against H. influenzae, and balance properties required for in vivo effi-
As shown in Table 2, these compounds showed superior
activity across the MIC panel, which included clinically relevant
and challenging Gram-positive strains and were >100-fold selec-
tive for bacterial ribosomes over those of rabbit reticulocyte
(data not shown). Moreover, they showed consistent and signif-
Full details of the structure-based approach, described above will be published
elsewhere. Briefly, from iterative solutions of new oxazolidinone-containing com-
pounds bound to their ribosomal target, we ranked new analogs on how they
improved the interaction energies and shape complementarity (or ‘‘fit”) to the
binding pocket. This was accomplished using our proprietary suite of computational
tools, AnalogTM, which combines a grow-search-and-score algorithm with a highly
accurate molecular properties calculator, QikProp. These ideas were filtered based on
mathematical models that point in the direction of improved biological activity. MIC
data from a broad range of substituted bis-aryloxazolidinones were incorporated as
thetraining sets for several computational models of H. influenzae inhibition. These
models were based on either a linear or neural network approach and were used to
icant activity against H. influenzae strains, with most MIC90
s
against H. influenzae of 4 g/ml or better. Additionally, all com-
l
pounds showed oral efficacy in
a
Streptococcus pneumoniae
ààà
Although derived here from first principles by use of the ribosomal structure,
several examplesof substituted biaryl oxazolidinones have been previously
reported.5–8 However, the importanceof the therapeuticaly valuable heterocyclic-
generate
oxazolidinones.
a
target list of substituted alkyl and heteroalkylamino bis-aryl
substituted aminoalkyl variants reported hererepresent
important compound class.
a new advance to this

J. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6175–6178
6177
O
O
O
O
O
N
O
N
N
N
N
N
NH
N
F
F
F
N
N
N
NHAc
NHAc
NHAc
N
N
HN
5
6
7
N
O
O
O
O
O
O
N
N
N
N
NH
NH
N
F
F
F
NC
N
NHAc
NHAc
NHAc
8
9
10
O
N
O
O
N
NH
F
F
NHAc
11
Figure 3.
Table 2
Potent, orally efficacious designer oxazolidinones
Bacterial strain (MIC in
lg/ml)
5
6
7
8
9
10
11
Linezolid
E. faecalis ATCC29212-P5 (linezolidR)
E. faecalis 1069 vanB
1
0.5
1
1
0.25
0.25
0.5
2
2
8
1
1
1
0.5
1
0.5
0.5
0.25
0.5
4
2
8
4
1.30
1
2
0.5
1
2
0.5
1
1
0.125
0.125
0.125
2
2
4
32
2
16
2
1
2
60.25
0.5
0.5
0.125
0.25
0.125
4
2
8
8
<0.8
60.25
0.5
1
0.25
0.25
0.125
0.5
0.25
1
60.25
1
E. faecium A6349 vanA + (linezolidR)
S. aureus 67-0 MRSA
0.5
0.25
0.25
0.25
4
4
8
8
2.00
S. pneumoniae 02J1175 mef (A)
S. pneumoniae 02J1258 erm (B) + L4
S. pyogenes msr610 erm (B)
H. influenzae RD1 parent
H. influenzae 1100
H. influenzae A1950
0.125
0.125
0.125
2
1
8
1
16
16
32
32
5.8
MIC90 H. influenzae
PD50 (mg/kg/day) S. pneumoniae 02J1175 mef (A)
4
2.50
1
4
4
1.70
10.70
28.40
All minimal inhibitory concentration (MIC) determinations were carried out under NCCLS conditions.3
N3
Cl
H2N
PPh3, H2O, THF
I
O
1. 4-Hydroxymethylphenylboronic acid
O
K2CO3 , Pd(PPh3)4, toluene/EtOH/H2O
O
O
F
N
NaN3, DMF
F
N
O
F
N
O
H
F
N
O
O
H
2. MsCl, DIEA, DMF
H
H
AcHN
AcHN
AcHN
AcHN
12
13
14
15
RCH2Br, DIEA
DMF
or
RCHO, DMF
NaBH(OAc)3
1. Propargylamin, DMF
2. Boc2O, K2CO3,THF/H2O
3. NaN3, NH4Cl, CuI. DIEA
DMF
Tetrazole
Hunig's base
CH3CN
1. Trimethylsilylacetylene
2. TBAF, AcOH
4. HCl, 1,4-dioxane
N
R
N
H
N
N
N
N
N
H
O
N
N
O
N
N
N
H
F
N
F
N
O
5
O
O
O
F
N
H
H
O
F
N
7
O
AcHN
H
AcHN
H
8
AcHN
N
N
AcHN
N
2
3
4
9
R=4-Quinolyl
R=4-Pyridyl
R=2-Furanyl
R=CN
N
O
F
N
O
6
H
10 R=4-Isoxazolyl
11 R=CH2CH2F
AcHN
Scheme 1.
mouse peritonitis model (Table 2). These values were generally
better than the highly orally-bioavailable linezolid, whose oral
PD50 in this model is 5.8 mg/kg/day.
Synthesis: Compounds 2–4 (Fig. 2) and 5–11 (Fig. 3) were pre-
pared as shown in Scheme 1. Palladium-mediated coupling of
hydroxymethylphenyl boronic acid with intermediate 124 followed

6178
J. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6175–6178
by mesylation and chloride displacement afforded compound 13 in
excellent overall yield. Displacement of chloride from 13 by tetra-
zole afforded both possible tetrazole regioisomeric products 5 and
6. Displacement of chloride from 13 with propargylamine followed
by copper-catalyzed 2 + 3 cycloaddition of azide afforded the ami-
nomethyltriazole 7. Reaction of 13 with sodium azide in DMF gave
intermediate 14). Reaction of 14 in a copper-mediated 2 + 3 cycload-
dition of trimethylsilylacetylene followed by deprotection afford the
N-linked triazole 8. Finally, phosphine-mediated reduction of 14
gave the amine 15. Using either alkylation or reductive amination
afforded target compounds 2–4, 9, 10, and 11.
References and notes
1. Zhou, J. et al. Bioorg. Med. Chem. Lett. 2008, 18, 6179
2. Molecular properties calculation software available from Schrödinger
LLC.
3. NCCLS. M7-A5—Methods for Dilution Antimicrobial Susceptibility Tests for
Bacteria That Grow Aerobically; Approved Standard—Fifth Edition, 2001. NCCLS
Document M100-S12/M7 (ISBN 1-56238-394-9). NCCLS, 940 West Valley Road,
Suite 1400, Wayne, Pennsylvania 19087-1898, USA.
4. Lee, C. S.; Allwine, D. A.; Barbachyn, M. R.; Grega, K. C.; Dolak, L. A.; Ford, C. W.;
Jensen, R. M.; Seest, E. P.; Hamel, J. C.; Schaadt, R. D.; Stapert, D.; Yagi, B. H.;
Zrenko, G. E.; Genin, M. J. Bioorg. Med. Chem. 2001, 9, 3243.
5. Hutchinson, D. K. Expert Opin. Ther. Patents 2004, 14(9), 1309.
6. US 5254577.
7. US 6689779.
8. US 7192974.
Acknowledgments
We thank Laura Lawrence, Timothy McConnell, Bradford King,
and Xiang Luo, for the biological data incorporated in this letter.