Structure-activity relationships of novel antibacterial translation inhibitors: 3,5-Diamino-piperidinyl triazines

Article abstract of DOI:10.1016/j.bmcl.2006.07.052

Structure-activity relationships of the 3,5-diamino-piperidinyl triazine series, a novel class of bacterial translation inhibitors, are described. Optimization was focused on the triazine C-4 position in which aromatic substituents that contained electron-withdrawing groups led to potent inhibitors. The initial lack of antibacterial activity was correlated with poor cellular penetration. Whole cell antibacterial activity was achieved by linking additional aromatic moieties at the triazine C-4 position.

Full text of DOI:10.1016/j.bmcl.2006.07.052

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5451–5456
Structure–activity relationships of novel antibacterial
translation inhibitors: 3,5-Diamino-piperidinyl triazines
Yuefen Zhou, Zhongxiang Sun, Jamie M. Froelich, Thomas Hermann and Daniel Wall*
Anadys Pharmaceuticals, Inc., San Diego, USA
Received 15 June 2006; accepted 17 July 2006
Available online 4 August 2006
Abstract—Structure–activity relationships of the 3,5-diamino-piperidinyl triazine series, a novel class of bacterial translation
inhibitors, are described. Optimization was focused on the triazine C-4 position in which aromatic substituents that contained
electron-withdrawing groups led to potent inhibitors. The initial lack of antibacterial activity was correlated with poor cellular
penetration. Whole cell antibacterial activity was achieved by linking additional aromatic moieties at the triazine C-4 position.
Ó 2006 Elsevier Ltd. All rights reserved.
Bacterial infections are a leading cause of morbidity and
are major contributors toward top causes of mortality in
the United States, including lower respiratory diseases,
pneumonia, and sepsis.1 Coupled with mounting bacte-
rial resistance, an aging population, and a dearth of nov-
el therapeutics, there is an important unmet clinical need
to develop new antibiotics.2 One challenge in antibiotic
discovery is developing compounds with broad-spec-
trum activity. A proven drug target against diverse bac-
terial groups is the ribosome, where a number of
clinically important antibiotic classes bind including
macrolides, tetracyclines, aminoglycosides, and oxazo-
lidinones. Importantly, all of these antibiotics, which
were identified in whole cell screens, specifically target
rRNA, which represents a macromolecule class that
has largely been ignored in target driven drug discovery
programs.3 The recently solved co-structures of antibiot-
ics bound to the ribosome have opened structure-guided
strategies to discover small molecules that bind rRNA.4
Emerging from such an approach we have reported on a
novel series of synthetic aminoglycoside mimetics that
inhibit bacterial translation, presumably by interfering
with the ribosomal decoding site (A-site).5 Here we
expand on the series structure–activity relationships
(SAR).
The aminoglycosides are currently the best character-
ized class of small molecules that interact specifically
with RNA. In prior studies the two-amine groups of
the 2-deoxy-streptamine (2-DOS) scaffold, which is
conserved among aminoglycoside antibiotics, have
been identified as a key pharmacophore for binding
rRNA.6 Medicinal chemistry efforts to improve the
toxicological, microbiological, and pharmacological
properties of aminoglycosides have been hindered by
their complex chemistry. To circumvent this problem,
we have previously developed a synthetic mimetic of
2-DOS, the cis-3,5-diamino-piperidine (DAP).5 The
DAP ring retains the characteristic cis-1,3-diamine
configuration of 2-DOS which is an important feature
for RNA recognition via hydrogen bonding to base
edges.6 In contrast to aminoglycosides, the reduced
chemical complexity of the DAP heterocycle renders
this series amenable to rapid parallel synthesis. In
one optimal configuration two symmetrically posi-
tioned DAP moieties are directly linked to a triazine
core to form the di-DAP substituted triazine (DAPT)
(1) (Scheme 1). Similar to aminoglycosides, DAPT
molecules have been shown to bind the A-site within
16S rRNA, inhibit translation in vitro, promote trans-
lational miscoding in vivo, act as cidal agents and
protect mice from bacterial infections.5 Molecular
modeling studies based on crystal structures of amino-
glycosides bound to the A-site suggest that the DAP
groups serve as a key pharmacophore for rRNA bind-
ing. Here we describe the optimization of the DAPT
series around R1 (1), which plays a critical role for
obtaining antibacterial activity.
Keywords: Aminoglycosides; Antibiotics; Translation inhibitors;
Ribosome.
*
Corresponding author. Tel.: +1 858 530 3711; fax: +1 858 530
3644; e-mail: dwall@anadyspharma.com
Present address: Department of Chemistry and Biochemistry,
University of California, San Diego, La Jolla, CA 92093-0358, USA.
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.07.052

5452
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5451–5456
BocHN
NHBoc
NHBoc
NHBoc
Cl
N
Cl
Cl
N
Cl
R2 NH2
N
C
H
N
N
N
N
N
N
BocHN
NHBoc
N
N
NaHCO3 / THF
0 oC-r.t., 40 min.
NaHCO3 / THF
r.t.1h, 80 oC, 3-48h
N
HN
R2
Cl
HN
D
R2
A
B
2 M HCl in
dioxane/MeOH
r.t., o.n.
DAP
NH2
NH2
1
N
2
6
N
N
H2N
NH2
5N N3
4
R1
1
Scheme 1. Parallel synthesis of DAPT (1) compounds.
Table 1. Bacterial protein synthesis inhibitory activity for compounds 1
In these studies, lead optimization was focused on sub-
stitutions at the triazine C-4 position (1), which for con-
venience of discussion, this region of the molecule was
termed tailpiece. Scheme 1 shows a general route for
parallel synthesis of DAPT library compounds. In this
method, cyanuric chloride (A) was treated with an
amine (e.g., aromatic amine) at low temperature to form
intermediate B, followed by treatment with Boc-protect-
ed DAP (C) under heating conditions to give the
tri-substituted 1,3,5-triazine derivative D. Removal of
the Boc-protecting group with HCl in dioxane yielded
compound 1.7 Early library compounds showed that
halide, hydroxyl, amine or alkyl tailpieces (compounds
2–11) resulted in poor to modest activity in a bacterial
in vitro translation assay (Table 1). In contrast, aromatic
tailpieces, as exemplified by compounds 14 and 15,
resulted in potent translation inhibitors. Compounds
with aromatic heterocyclic tailpieces had mixed activi-
ties; an isoxazole tailpiece (13) exhibited potent inhibito-
ry activity, while a pyridine tailpiece (12) had moderate
activity (Table 1). These combined data suggest that pla-
nar aromatic amine tailpieces (R1) improved rRNA
binding. We hypothesize that stacking interactions with
rRNA base residues could account for improved activity.
Compounda
R1
IC50 (lM)b
2
3
4
5
6
7
Cl
42
92
NH2
45
62
NHOH
NHCH3
NHNH2
N(CH3)2
170
290
8
180
N
9
10
11
63
23
36
HN
N
NH
NH
N
H
N
12
32
HN
HN
13
14
15
9
7
8
O
N
HN
Although compound 14 was a reasonable inhibitor of
bacterial translation, it did not inhibit bacterial growth
(Table 2). To improve potency and establish antibacteri-
al activity, optimization was focused around the phenyl
tailpiece of compound 14. The results in Table 2 show
that electron-withdrawing substitution on the phenyl
ring of R2 improved inhibitory potency against bacterial
translation as exemplified in compounds 16–21. In
contrast, electron-donating substitutions, including
compounds 22–26, resulted in reduced potency com-
pared to the parent molecule (14). We hypothesize that
substituent-modulated changes in electronic properties
of the phenyl ring likely influence the stacking interac-
tions with rRNA bases.
HN
NH2
Kan
Gent
Tet
0.4
0.03
2.8
1.5
Cm
a Kan, kanamycin; Gent, gentamicin; Tet, tetracycline; and Cm,
chloramphenicol.
b Coupled in vitro transcription–translation assay using Escherichia
coli extracts.5
Escherichia coli and Staphylococcus aureus (Table 2). To
determine if the weak antibacterial activity was caused
by poor cellular penetration, this series of compounds
was tested against an impꢀ permeability mutant.8
Indeed, the MIC values of a number of the potent trans-
lation inhibitors (16, 17, 18, and 20) were significantly
Importantly, certain compounds (16, 18, and 20) that
were potent protein synthesis inhibitors in vitro showed
modest antibacterial activity on standard test strains of

Y. Zhou et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5451–5456
5453
Table 2. Antibacterial activitya and protein synthesis inhibitory
activity
accomplished by first coupling the aniline linker with
R5 followed by the same triazine chemistry shown in
Scheme 1.7 SAR of these library compounds revealed
that ortho substitutions universally correlated with weak
antibacterial activity, while substitutions at the meta and
para positions often resulted in compounds with sub-
stantially improved MIC values. Table 3 illustrates that
a para substitution (27) is more potent than an identical
ortho substitution (28). Similar antibacterial differences
can be seen between compounds 18 and 19 (Table 2).
In contrast, substituents in the meta or para position
often offer little discernable differences in potencies as
seen between compounds 29 and 30 (Table 3).
DAPT
NH
R2
Compound R2
IC50 Sa
(lM)
Ec
Ec imp+ Ec
impꢀ
14
16
17
7
4
>64 >64 >64
>64
O
N+
32
32
32
4
4
The SAR in Table 3 suggested that longer tailpieces,
which contained two aromatic rings, could improve
antibacterial activity on a standard test strain (e.g., 27,
29, and 30). To further test this SAR, the role of the
phenyl ‘linker’ was examined. Using a similar synthetic
route,7 Table 4 shows a comparison of the antibacterial
activity between compounds with and without an aro-
matic linker (X). In all cases the inclusion of an aromatic
linker (compounds 31, 33, 35, and 37) results in
improved antibacterial activity. These results are consis-
tent with the prior SAR suggesting the phenyl linker in
the tailpiece plays an important role in bacterial cell
penetration.
O-
4
>64 >64 >64
Br
18
3
16
64
32
4
F
NH
HN
19
5
>64 >64 >64
>64
4
F
F
F
F
20
21
22
23
24
25
26
5
16
32
32
O
S
Based on the established SAR small focused libraries
were synthesized around the DAPT series containing
a phenyl linker. In most cases substituents were
attached in the para position. Table 5 shows represen-
tative compounds from focused libraries where the
phenyl linker substituents (R5) consisted of aromatic
heterocycles (38), sulfonamides (39), a-ketoamides
(40), b-ketoamides (41–45), and amides (46, 47).
Again, these compounds were prepared using a similar
synthetic approach described above.7 As shown in
Table 5 the expansion of the tailpiece resulted in a
set of compounds that exhibited favorable antimicro-
bial activity. The MICs against E. coli were typically
in the range of 1–4 lg/mL, while activity against
Pseudomonas aeruginosa was generally more potent
(MICs frequently <1 lg/mL). In general, activity
against S. aureus and other Gram-positive bacteria
was less potent, with exceptions of compounds 38
and 46, suggesting that there is room for further opti-
mization. The IC50 values for in vitro translation inhi-
bition remained in a similar single digit lM range,
and these compounds generally exhibited low toxicity
towards human cells.
NH2
4
>64 >64 >64
>64
>64
8
O
13
15
11
14
11
32
64 >64
32 >64
64
OH
O
>64 >64 >64
>64 >64 >64
>64 >64 >64
>64
>64
>64
N
Sa, Staphylococcus aureus ATCC 25923; Ec, Escherichia coli ATCC
25922; Ec imp+, E. coli LMG194 (wild type); Ec impꢀ, E. coli LMG194
(isogenic permeable mutant).
a MIC (minimum inhibitory concentration) lg/mL.5
lowered when tested on the impꢀ mutant as compared to
the isogenic imp+ strain (Table 2). These results suggest
that the poor antibacterial activity of this series was
caused by poor cell penetration or efflux.
In summary, the DAPT series described here originat-
ed from a structure-guided approach to discover novel
aminoglycoside mimetics that bind rRNA and act as
bacterial protein synthesis inhibitors. This series
evolved through an iterative process of high through-
put library synthesis and testing, in part described
here, which discovered promising leads. To date over
1000 compounds have been synthesized and utility
has been shown in a mouse septicemia model.5 An
important feature of this series is the potent antimi-
crobial activity against a number of Gram-negative
To improve the antibacterial activity of this series, sub-
sequent libraries were made focusing on the modifica-
tion of the R2 phenyl ring. Initially, sample libraries of
about 30 compounds were synthesized with substitu-
tions at the ortho (R3), meta (R4) or para (R5) positions
of the phenyl ring that is directly linked to the triazine
core (Table 3). The synthesis of these compounds was

5454
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5451–5456
Table 3. Antibacterial activity and protein synthesis inhibitory activity
DAPT
NH
R3
R4
Linker
R5
Compound
R3ðorthoÞ
H
R4ðmetaÞ
OCH3
R5ðparaÞ
IC50 (lM)
Escherichia colia
Cl
HN
O
27
8
4
OH
Cl
HN
O
28
29
30
OCH3
H
30
5
>64
8
OH
O
H
H
NH
HO
OH
O
H
H
4
8
NH
HO
OH
a MIC (lg/mL) of strain ATCC 25922.
Table 4. Antibacterial activity and protein synthesis inhibitory activity
DAPT
X
Y
Compound
Xa
Y
IC50 (lM)
Escherichia colib
14
7
>64
NH
NH
NH
31
32
33
34
35
36
37
9
16
32
2
N
H
Cl
7
Cl
Cl
O
H N
10
14
4
Cl
N
OH
H
H
N
64
8
N
O
H
N
H N
N
N
H
OH
Br
—
5
16
2
N
OH
Br
O
9
HN
N
a Left sides of the molecules are bonded to DAPT via the designated N atom and the right sides are bonded to Y.
b MIC (lg/mL) of strain ATCC 25922.
pathogens with especially striking activity against
P. aeruginosa, a pathogen for which there is a clear
unmet clinical need to develop new antibiotics.2
Further attractions of this novel class include activity
against aminoglycoside-resistant bacteria and the rela-
tive ease in which the series can be synthesized.

Y. Zhou et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5451–5456
5455
Table 5. Antibacterial activity, protein synthesis inhibitory activity, and eukaryotic cytotoxicitya
DAPT
NH
R4
R5
Compound R4
R5
IC50 (lM)
MIC (lg/mL)
Escherichia coli Staphylococcus
CC50 (lM)a
Pseudomonas
aureus ATCC 25923 aeruginosa ATCC 27853
ATCC 25922
S
N
38
39
40
41
42
43
44
45
46
H
10
4
2
2
16
16
16
16
16
8
ND
33
68
O
S
OH
OH
H
4
1
4
4
4
4
2
1
16
N
O
H
N
O
O
4
60.5
61
61
1
92
NH
O
O
O
6
24
N
H
O
H
4
71
N
H
Cl
F
F
O
O
O
O
O
H
9
179
146
124
140
N
H
O
F
F
H
6
1
O
N
H
O
H
5
16
2
61
1
N
Br
H
HN
H
10
O
OH
HN
O
Cl
47
Gent
Cm
OH
7
2
<0.5
4
16
<0.5
8
60.5
0.5
180
>300
ND
0.03
1.5
>64
ND, not determined.
a Eukaryotic cytotoxiccity was determined in a proliferation assay with CEM T-cells after 72 h.5
Shlaes, D. M.; Projan, S. J.; Edwards, J. E. ASM News
2004, 70, 275.
Acknowledgments
3. (a) Hermann, T.; Tor, Y. Expert Opin. Ther. Patents 2005,
15, 49; (b) Vicens, Q.; Westhof, E. ChemBioChem 2003, 4,
1018.
4. (a) Franceschi, F.; Duffy, E. M. Biochem. Pharm. 2006, 71,
1016; (b) Poehlsgaard, J.; Douthwaite, S. Nat. Rev.
Microbiol. 2005, 3, 871; (c) Tenson, T.; Mankin, A. Mol.
Microbiol. 2006, 59, 1664.
We thank S. Fish for assistance with the in vitro trans-
lation assay and V. Banh for testing eukaryotic cytotox-
icity. This work was supported in part by a grant from
the National Institute of Allergy and Infectious Disease
(AI51105).
5. Zhou, Y.; Gregor, V. E.; Sun, Z.; Ayida, B. K.;
Winters, G. C.; Murphy, D.; Simonsen, K. B.; Vour-
loumis, D.; Fish, S.; Froelich, J. M.; Wall, D.;
Hermann, T. Antimicrob. Agents Chemother. 2005, 49,
4942.
References and notes
1. Anderson, R. N.; Smith, B. L. Natl. Vital Stat. Rep.; Center
Dis. Control 2005, 53, 1.
6. (a) Simonsen, K. B.; Ayida, B. K.; Vourloumis, D.;
Winters, G. C.; Takahashi, M.; Shandrick, S.; Zhao, Q.;
2. (a) Overbye, K. M.; Barrett, J. F. Drug Discov. Today 2005,
10, 45; (b) Rice, L. B. Biochem. Pharm. 2006, 71, 991; (c)

5456
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5451–5456
Hermann, T. ChemBioChem 2003, 4, 886; (b) Vourloumis,
7. Zhou, Y.; Vourloumis, D.; Gregor, V. E.; Winters, G. C.;
Hermann, T.; Ayida, B. K.; Sun, Z.; Murphy, D.; Simon-
sen, K. B. 2005. PCT WO 2005/028467 A1.
8. Sampson, B. A.; Misra, R.; Benson, S. A. Genetics 1989,
122, 491.
D.; Winters, G. C.; Takahashi, M.; Simonsen, K. B.; Ayida,
B. K.; Shandrick, S.; Zhao, Q.; Hermann, T. ChemBioChem
2003, 4, 879; (c) Vicens, Q.; Westhof, E. Biopolymers 2003,
70, 42.