Structure-activity relationships of the peptide deformylase inhibitor BB-3497: Modification of the methylene spacer and the P1′ side chain

Article abstract of DOI:10.1016/S0960-894X(03)00532-8

Structural modifications to the peptide deformylase inhibitor BB-3497 are described. In this paper, we describe the initial SAR around this lead for modifications to the methylene spacer and the P1′ side chain. Enzyme inhibition and antibacterial activity data revealed that the optimum distance between the N-formyl hydroxylamine metal binding group and the P1′ side chain is one unsubstituted methylene unit. Additionally, lipophilic P1′ side chains that closely mimic the methionine residue in the substrate provided compounds with the best microbiological profile.

Full text of DOI:10.1016/S0960-894X(03)00532-8

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2709–2713
Structure–Activity Relationships of the Peptide Deformylase
Inhibitor BB-3497: Modification of the Methylene Spacer
and the P10 Side Chain
Stephen J. Davies,* Andrew P. Ayscough, R. Paul Beckett, Ryan A. Bragg,
John M. Clements, Sheila Doel, Christine Grew, Steven B. Launchbury,
Gemma M. Perkins, Lisa M. Pratt, Helen K. Smith, Zoe M. Spavold, S. Wayne Thomas,
Richard S. Todd and Mark Whittaker
British Biotech Pharmaceuticals Limited, Watlington Road, Oxford OX4 6LY, UK
Received 26 March 2003; revised 19 May 2003; accepted 22 May 2003
Abstract—Structural modifications to the peptide deformylase inhibitor BB-3497 are described. In this paper, we describe the initial
SAR around this lead for modifications to the methylene spacer and the P10 side chain. Enzyme inhibition and antibacterial activity
data revealed that the optimum distance between the N-formyl hydroxylamine metal binding group and the P10 side chain is one
unsubstituted methylene unit. Additionally, lipophilic P10 side chains that closely mimic the methionine residue in the substrate
provided compounds with the best microbiological profile.
# 2003 Elsevier Ltd. All rights reserved.
Ribosomal protein synthesis in bacteria is initiated by
N-formylation of methionyl-tRNA by formyl-
transferase. Most mature proteins, however, do not
retain the N-formyl group or the terminal methionine
residue. Deformylation is therefore a crucial step in
bacterial protein biosynthesis and the metalloenzyme
responsible, peptide deformylase (PDF), is essential for
bacterial growth.1 Bacterial PDF is now widely
recognised as an attractive target for antibacterial
chemotherapy.2
cation the optimum distance between the metal binding
group and the P10 side chain is evaluated and the pre-
ferred P10 substituents for enzyme binding affinity and
antibacterial activity are discussed.
During the course of this investigation, several X-ray
crystal structures of PDF enzymes were elucidated.5ꢀ8
The structural information provided by these studies
We reported previously3 that the N-formyl hydroxyl-
amine BB-3497 (Fig. 1) is an effective inhibitor (IC50=7
.
nM) of the Escherichia coli PDF Ni enzyme, exhibiting
potent antibacterial activity both in vitro and in vivo. In
a subsequent communication,4 we identified the N-for-
myl hydroxylamine and hydroxamic acid structural
motifs to be the optimum metal binding groups on the
pseudopeptidic backbone of BB-3497. Further to this
study, we have conducted a systematic modification of
the other sites indicated in Figure 1. In this communi-
*Corresponding author. Tel.: +44-1235-441732; fax: +44-1235-
441509; e-mail: stephen.davies@evotecoai.com
Figure 1. BB-3497 structure–activity relationships.
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00532-8

2710
S. J. Davies et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2709–2713
was particularly valuable in terms of designing and
modelling P10 side chain modifications of BB-3497 that
could occupy the well-defined S10 pocket.
The chemical modifications described here are an
attempt to understand the key structural requirements
that may lead to molecules with an improved anti-
bacterial profile and pharmacological properties.
Chemistry
To examine the optimal distance between metal binding
group (MBG) and the P10 position (Table 1), three dif-
ferent approaches were used. The first explored route
involved the synthesis of compound 4 containing an
ethyl spacer and is shown in Scheme 1. Conversion of
the carboxylic acid analogue 19 to the mixed anhydride
using ethyl chloroformate, followed by reduction with
sodiumborohydride gave the primary alcohol 2, which
was subjected to Mitsunobu conditions with Cbz-
NHOBn to give N,O-bisprotected hydroxylamine 3.
Hydrogenolysis of 3 and formylation with N-formyl
benzotriazole10 (BtCHO) gave 4 after purification by
preparative HPLC.
Scheme 2. (i) BnONH2, NaCNBH3, 5 N HCl (pH 4.5), MeOH, rt
.
H-tert-leucine NMe2, DMF, 0 ꢁC to rt, then separate diastereoisomers,
45%; (iv) H2, 10% Pd/C, MeOH, rt, quant.
82%; (ii) CDI, HCO2H, Et2O, rt, 94%; (iii) EDAC HCl, HOAt, then
To further probe the optimum distance between the
MBG and the P10 position, a second approach was
based on preparing those compounds which did not
contain the methylene spacer group (Scheme 2).
Reductive amination of 2-oxo-hexanoic acid–sodium
salt 5 with O-benzylhydroxylamine and acidic workup
with sulfuric acid afforded 6 in excellent yield. Sub-
sequent N-formylation using N-formyl imidazole (pre-
pared in situ fromcarbonyl diimidazole and formic
acid) and standard peptide coupling conditions using
binding affinity. The non-diastereoselective approach
used for the preparation of compound 16 is shown in
Scheme 3.
Standard Evans methodology12 using the (S)-oxazolidi-
none 10 and hexanoyl chloride generated 11 in excellent
yield. Acetylation of 11 provided 12 with high dia-
stereoselectivity (>85% d.e.). Reductive amination with
11
tert-leucine-NMe2 gave 8 as a mixture of diastereo-
isomers. Separation of the isomers by preparative
HPLC and removal of the O-benzyl protecting group by
Pd-catalysed hydrogenation provided the target
compounds 9a and 9b.
We also prepared an analogue of BB-3497 containing a
methyl group on the methylene spacer in order to eval-
uate the effect of substitution in this position on enzyme
Scheme 3. (i) nBuLi, hexanoyl chloride, THF, 86%; (ii) NaHMDS
.
(1 M in THF), acetyl chloride, THF, 80%; (iii) (a) BnONH2 HCl,
NaOAc; (b) NaCNBH3, acetic acid, 70% (two steps); (iv) (a) BtCHO,
.
HOAt, then H-tert-leucine NMe2, DMF, 0 ꢁC to rt, 82%; (vi) H2, 10%
Pd/C, MeOH, rt 80%.
Scheme 1. (i) (a) Ethyl chloroformate, Et3N; (b) NaBH4, THF, 36%;
(two steps); (ii) PPh3, DIAD, CbzNHOBn, THF, 30%; (iii) (a) H2,
Pd/C; (b) BtCHO, THF, 32% (two steps).
THF, 56%; (b) LiOH, H2O2, THF, H2O, 64%; (v) EDAC HCl,

S. J. Davies et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2709–2713
2711
hyde 18, followed by oxime formation using O-benzyl-
hydroxylamine in the presence of sodium acetate
provided 19. Saponification of the ethyl ester was fol-
11
lowed by amide coupling with tert-leucine-NMe2 to
give 20 in good yield. Reduction of the oxime using
sodiumcyanoborohydride provided the protected
hydroxylamine as a mixture of diastereoisomers which
were separated by chromatography to give the desired
(R)-diastereoisomer 21.13 Hydroxylamine 21 was N-for-
mylated10 using BtCHO and subsequent removal of the
O-benzyl protecting group by catalytic hydrogenation
afforded the target compound 23. A more detailed pro-
cedure for the preparation of the majority of com-
pounds in this paper is available in the patent
literature.14
Results and Discussion
All compounds were screened for activity against the
PDF.Ni enzyme and selected Gram-positive and Gram-
negative bacterial strains.15
Scheme 4. (i) HCOOEt, NaOEt, 23–45%; (ii) NH2OBn, NaOAc, aq
EtOH, 40–70%; (iii) NaOH, MeOH; quant; (iv) H-tert-leucine-NMe2,
EDC, HOAt, DMF, 45–75%; (v) NaCNBH3, AcOH, then separate
diastereoisomers, 30–45%; (vi) BtCHO, THF; 86%; (vii) H2, Pd/C,
MeOH, quant.
Any attempt at increasing the length of the spacer group
(4) or incorporating even a small substituent (16) resul-
ted in loss of activity, relative to BB-3497 (Table 1).
Further modification, by complete removal of the
spacer group (9a+9b) showed significant loss of PDF
enzyme activity. The rigid geometry around the metal
binding group and the distance to the P10 pocket allows
for little modification in this region of the inhibitor.
O-benzyl hydroxylamine and sodium cyanoborohydride
gave 13 as a mixture of diastereoisomers. Subsequent N-
formylation followed by oxidative hydrolysis of the
chiral auxilliary gave the carboxylic acid precursor 14.
11
Standard coupling conditions using tert-leucine-NMe2
provided the compound 15. Hydrogenolysis under
standard conditions afforded the N-formyl hydroxyl-
amine 16 as a mixture of diastereoisomers.
The activity of P10 analogues of BB-3497 against the
PDF enzyme is consistent with a well defined hydro-
phobic S10 pocket. In correlation with the structural
information obtained from X-ray data, this pocket is
capable of accommodating small alkyl, cycloalkyl and
benzylic groups. Variation in the length of the P10 alkyl
chain (Table 2) (23a–23c, 23e–23h) revealed that n-butyl
was optimal (BB-3497). However, the enzyme also
accommodated a variety of other substituents. Branch-
ing of the alkyl chain (23i–23k), addition of a cycloalkyl
substituent (23l–23n) and the presence of unsaturation
or a sulfur atom( 23–23s) in the P10 side chain all gave
To examine the effect of P10 side-chain modification on
PDF enzyme activity and antibacterial activity, a non-
stereospecific route was developed to access analogues
of BB-3497 as shown in Scheme 4.
The compounds detailed in Table 2 (23a–w) were pre-
pared fromthe corresponding ethyl esters 17, with the
P10 side chain already in place. Formylation of 17 with
ethyl chloroformate/sodium ethoxide to give the alde-
Table 1. PDF enzyme inhibition and antibacterial activity of BB-3497 and analogues with a modified spacer group
.
E. coli PDF Ni IC50 (nM)
Compd
X
E. coli MIC (mM)
S. capitis MIC (mM)
BB-3497CH
4
16
9a
9b
712.5
1000
600
25
2
CH2CH2
CHMe
n=0 (R-isomer)
n=0 (S-isomer)
>200
>200
>200
>200
>200
>200
>200
>200
20%@1mMa
30%@1mMa
Ecoli , Escherichia colia; S. capitis, Staphylococcus capitis.
aIC50 expressed as percentage inhibition at a concentration of 1 mM.

2712
S. J. Davies et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2709–2713
Table 2. PDF enzyme inhibition and antibacterial activity of BB-3497 and its P10 analogues
P10 (R)
E. coli MIC (mM)
S. capitis MIC (mM)
.
E. coli PDF Ni IC50 (nM)
Compd
23a
23b
23c
BB-3497
23d
23e
23f
Me
Et
n-Pr
100
70
50
7
70
10
30
>200
200
50
12.5
>200
25
>200
>200
>200
25
>200
25
200
n-Bu (R-isomer)
n-Bu (S-isomer)
n-Pentyl
n-Hexyl
100
23g
23h
23i
23j
23k
23l
23m
23n
23p
23q
23r
n-Heptyl
n-Octyl
i-Pr
i-Bu
i-Pentyl
40
50
50
10
20
8
20
6
30
20
30
20
20
>200
>200
>200
100
>200
>200
>200
200
50
200
c-Pentylmethyl
c-Pentyl
c-Hexylmethyl
Allyl
But-3-enyl
But-2-ynyl
EtSCH2
12.5
>200
50
12.5
>200
25
>200
200
100
50
25
50
25
23s
23t
>200
50
Bn
25
23u
23v
23w
Ph(4-Cl)
(4-MeO)PhCH2
1-Piperidylmethyl
500
200
>1000
>200
>200
>200
>200
>200
>200
.
A 2-fold difference in PDF Ni enzyme potency can be regarded as significant when comparing IC50 values between compounds.
rise to potent PDF enzyme inhibitors. Incorporation of
a benzylic group was acceptable (23t) but a substituent
on the ring (23v) resulted in a 10-fold reduction in PDF
inhibition. Introduction of a basic nitrogen (23w) and a
directly linked aromatic group (23u) were detrimental to
PDF inhibition.
optimise the microbiological profile. Preliminary results
from this study are reported in the following communi-
cation.11
In conclusion, a series of BB-3497 analogues containing
modifications to the spacer group and the P10 side chain
have been synthesised and were tested in the PDF
enzyme and whole cell assays. The optimum distance
between the metal binding group and the P10 position
on the pseudopeptide backbone of BB-3497 was found
to be an unsubstituted methylene unit. The optimum
groups in the P10 position for PDF enzyme inhibition
and antibacterial activity were found to be the n-butyl
and the cyclopentylmethyl substituent.
A limited number of compounds having the (S)-con-
figuration at the P10 centre were assayed and shown to
be less potent against the PDF enzyme when compared
to the (R)-diastereoisomer. This is highlighted in the
case of BB-3497 and 23d. The data also suggest a mod-
est advantage in activity against Gram-positive organ-
isms is gained by replacing the n-butyl in BB-3497 with
a cyclopentylmethyl substituent (23l).
Compounds with similar PDF activity often have a
markedly different antibacterial profile (e.g., 23m vs
23t). These differences may result in part from relative
lipophilicities but other factors enabling access to the
bacteria and/or susceptibility to efflux mechanisms may
be important. Modifications adjacent to the metal
binding group and to the n-butyl substituent are limited
due to the steric requirements for binding the active site
metal and size of the hydrophobic S10 pocket.5ꢀ8 Sub-
stituents that closely mimic the methionine of the sub-
strate provide optimal PDF enzyme inhibition.
Following this study we have focused on the modifi-
cation of substituents in the P20 and P30 positions to
Acknowledgements
The authors wish to thank Drs. Stephen East and Gilles
Pain for support and contributive discussions.
References and Notes
1. Yuan, Z.; Trias, J.; White, R. J. Drug Discov. Today 2001,
6, 955.
2. (a) Pei, D. Emerg. Ther. Targets 2001, 5, 23. (b) Giglione,
C.; Meinnel, T. Emerg, Ther. Targets 2001, 5, 41. (c) Giglione,

S. J. Davies et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2709–2713
2713
C.; Meinnel, T. Exp. Opin. Ther. Targets 2001, 5, 415. (d)
Clements, J. M.; Ayscough, A. P.; Keavey, K.; East, S. P.
Curr. Med. Chem.-Anti-Infect. Agents 2002, 1, 239.
3. Clements, J. M.; Beckett, R. P.; Brown, A.; Catlin, G.;
Lobell, M.; Palan, S.; Thomas, W.; Whittaker, M.; Wood, S.;
Salama, S.; Baker, P. J.; Rodgers, H. F.; Barynin, V.; Rice,
D. W.; Hunter, M. G. Antimicrob. Agents Chemother. 2001,
45, 563.
10. Katritzky, A. R.; Chang, H.-K.; Yang, B. Synthesis 1995,
503.
11. Davies, S. J.; Ayscough, A. P.; Beckett, R. P.; Clements, J.
M.; Doel, S.; Pratt, L. M.; Spavold, Z. M.; Thomas, W.;
Whittaker, M. Bioorg. Med. Chem. Lett. See following article.
doi 10.1016-S0960-894X(03)00533-X
12. Evans, D. A.; Morrissey, M. M.; Dorow, R. L. J. Am.
Chem. Soc. 1985, 107, 4346.
4. Smith, H. K.; Beckett, R. P.; Clements, J. M.; Doel, S.;
East, S. P.; Launchbury, S. B.; Pratt, L. M.; Spavold, Z. M.;
Thomas, W.; Todd, R. S.; Whittaker, M. Bioorg. Med. Chem.
Lett. 2002, 12, 3595.
5. Chan, M. K.; Gong, W.; Rajagopalan, P. T. R.; Hao, B.;
Tsai, C. M.; Pei, D. Biochemistry 1997, 36, 13904.
6. Becker, A.; Schlichting, I.; Kabsch, W.; Groche, D.;
Schultz, S.; Wagner, A. F. Nat. Struct. Biol. 1998, 5, 1053.
7. Becker, A.; Schlichting, I.; Kabsch, W.; Schultz, S.;
Wagner, A. F. J. Biol. Chem. 1998, 273, 11413.
8. Guilloteau, J. P.; Mathieu, M.; Giglione, C.; Blanc, V.;
Dupuy, A.; Chevrier, M.; Gil, P.; Famechon, A.; Meinnel, T.;
Mikol, V. J. Mol. Biol. 2002, 320, 951.
9. (a) The succinate portion of compound 1 was prepared by
employing an asymmetric alkylation on an Evans N-hexanoyl
oxazolidinone with t-butyl bromoacetate. See: Beckett, R. P.;
Crimmin, M. J.; Davis, M. H.; Spavold, Z. Synlett 1993, 137.
(b) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem.
Soc. 1982, 104, 1737.
13. Separation of the diastereoisomers was achieved by flash
chromatography using ethyl acetate/hexane as solvent eluants.
Based on our knowledge of BB-3497 and the asymmetric route
employed to determine the absolute stereochemistry,16 we were
able to compare 1H NMR, HPLC retention times and thin layer
chromatography of the different P10 analogues relative toBB-3497
anditsprecursors.OnaC18 Phenomenex Luna 50ꢂ4.6 mm HPLC
column, the (R)-diastereomer always eluted before the corre-
sponding (S)-diastereomer. For a select number of compounds,
both diastereoisomers of the N-formyl hydroxylamine derivative
.
23 were assayed against the PDF Ni enzyme and the more active
diastereoisomer assigned the designated stereochemistry.
14. British Biotech Pharmaceuticals. Patent Application WO
99/39704, Aug. 12, 1999.
15. PDF inhibition assays and MICs were performed as
described previously.3
16. Pratt, L. M.; Beckett, R. P.; Davies, S. J.; Launchbury,
S. B.; Miller, A.; Spavold, Z. M.; Todd, R. S.; Whittaker, M.
Bioorg. Med. Chem. Lett. 2001, 11, 2585.