Ester and hydroxamate analogues of methionyl and isoleucyl adenylates as inhibitors of methionyl-tRNA and isoleucyl-tRNA synthetases

Article abstract of DOI:10.1016/S0960-894X(01)00095-6

The structure-activity relationship on a series of ester and hydroxamate analogues of methionyl and isoleucyl adenylate has been investigated through introducing linkers between the 1′-position of ribose and adenine surrogates as methionyl-tRNA, and isoleucyl-tRNA synthetase inhibitors, respectively. The results indicate that ester analogue 23 was found to be a potent inhibitor of Escherichia coli methionyl-tRNA synthetase, and its interaction with the active site was proposed by a molecular modeling study.

Full text of DOI:10.1016/S0960-894X(01)00095-6

Bioorganic & Medicinal Chemistry Letters 11 (2001) 961–964
Ester and Hydroxamate Analogues of Methionyl and Isoleucyl
Adenylates as Inhibitors of Methionyl-tRNA and Isoleucyl-tRNA
Synthetases
Jeewoo Lee,a,* Sang Uk Kang,a Su Yeon Kim,a Sung Eun Kim,a Mee Kyoung Kang,a
Yeong Joon Job and Sunghoon Kimc
aLaboratory of Medicinal Chemistry, College of Pharmacy, Seoul National University, Shinlim-Dong, Kwanak-Ku, Seoul 151-742,
South Korea
bImaGene Co., Sung Kyun Kwan University, 300 Chunchundong, Jangangu, Suwon, Kyunggido 440-746, South Korea
cNational Creative Research Initiative Center for ARS Network, Sung Kyun Kwan University, 300 Chunchundong, Jangangu, Suwon,
Kyunggido 440-746, South Korea
Received 24 November 2000; accepted 31 January 2001
Abstract—The structure–activity relationship on a series of ester and hydroxamate analogues of methionyl and isoleucyl adenylate
has been investigated through introducing linkers between the 10-position of ribose and adenine surrogates as methionyl-tRNA, and
isoleucyl-tRNA synthetase inhibitors, respectively. The results indicate that ester analogue 23 was found to be a potent inhibitor of
Escherichia coli methionyl-tRNA synthetase, and its interaction with the active site was proposed by a molecular modeling study.
# 2001 Elsevier Science Ltd. All rights reserved.
Introduction
Aminoacyl adenylates, a mixed anhydride intermediate
generated during the reaction, have long been recog-
nized as the lead compounds in finding potent and
selective inhibitors, because they are known to bind
more tightly to the enzyme than the substrates, amino
acid and ATP, generally by two or three orders of
magnitude. Most aminoacyl adenylate analogues,
reported as enzyme inhibitors, have been designed with
a similar strategy, in which they ensure both tight bind-
ing and stability by replacing the labile anhydride bond
of aminoacyl-AMP with stable non-hydrolyzable bio-
isosteres. Among them, sulfamate bioisostere has been
Aminoacyl-tRNA synthetases (aaRSs) catalyze the
transfer of specific amino acids to their corresponding
tRNAs to form aminoacyl-tRNAs, which are used for
protein synthesis.1,2 Since the aminoacylation reaction is
essential in all living organisms, these enzymes have
attracted much attention as promising antibacterial tar-
gets to overcome the resistance problem caused by main-
line antibiotics.3,4 Obviously, the selective inhibition of
pathogen synthetases to the human cell counterpart is an
important issue when considered as a drug candidate.
*Corresponding author. Tel.: +82-2-880-7846; fax: +82-2-888-0649; e-mail: jeewoo@snu.ac.kr
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00095-6

962
J. Lee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 961–964
the most successful in terms of binding affinity to a
variety of aaRSs, such as AlaRS,5 ArgRS,6 HisRS,6
IleRS,7 ProRS,8 SerRS,9 ThrRS,6 and TyrRS.10
4-phenyltetrazole, followed by deprotection of the TBS
group provided key intermediates 7. For the syntheses
of ester analogues, alcohols 7 were esterified with N-Boc
methionine, or N-Boc isoleucine, and then deprotected
under acidic conditions, to give 8–13, respectively. For
the syntheses of hydroxamate analogues, alcohols 7
were condensed with N-Boc methionine hydroxamate or
N-Boc isoleucine hydroxamate by Mitsunobu reaction,
and then deprotected to afford final compounds 14–19,
respectively. Syntheses of one-carbon-elongated analo-
gues are described in Scheme 2. Alcohol 6 was dehy-
drated into alkene 20 by mesylation and subsequent
base-mediated elimination. Ozonolysis, followed by
NaBH4 reduction of double bond 20, provided the cor-
responding alcohol 21, whose hydroxyl was substituted
by heterocycles under Mitsunobu conditions, and then
deprotected to give 22 as key intermediates. By follow-
ing the same sequence described in Scheme 1, alcohols
22 were converted into ester analogues 23–26 and
hydroxamate analogues 27–30, respectively.
An extensive SAR study on sulfamate analogues of iso-
leucyl adenylate demonstrated that the binding region
of adenine moiety in Escherichia coli isoleucyl-tRNA
synthetase (IleRS) contained a wide hydrophobic
pocket large enough to afford three linear aromatic
rings.3
We recently reported that ester 1 and hydroxamate
analogue 2 of methionyl adenylate, as another stable
analogue of aminoacyl adenylate, were potent inhibitors
of methionyl tRNA synthetases (MetRS) isolated from
E. coli, M. tuberculosis, S. cerevisiae and human.11 As
part of our continuing effort to find potent and patho-
gen-selective inhibitors of MetRS and IleRS as potential
antibacterial agents, we herein describe extensive SARs
of ester and hydroxamate analogues of methionyl and
isoleucyl adenylates, in which heterocycles as adenine
surrogates are separated from ribose by one or two
carbon units.
Biological Results and Discussion
Synthesized methionyl and isoleucyl adenylate analo-
gues were evaluated as inhibitors of E. coli MetRS and
IleRS, respectively, which are mechanistically regarded
as class I type with similar structural folds and sequence
motifs.13 Inhibitory activities were determined by mea-
suring the decrease of the aminoacylation product,
the [35S]methionyl E. coli-tRNAMet or [3H]isoleucyl
E. coli-tRNAIle, in the presence of different chemical
concentrations.
Synthesis
Ester and hydroxamate analogues of isoleucyl adeny-
late, 3 and 4, were prepared from 20,30-isopropylidene
adenosine by following the previous procedure.11
Syntheses of two-carbon-elongated analogues are out-
lined in Scheme 1. Methyl 3,6-anhydro-2-deoxy-4,5-O-
isopropylidene-d-allo-heptonate (5) as a starting mat-
erial was prepared from d-ribose by a literature pro-
cedure.12 Primary alcohol of 5 was protected with the
TBS group, and its ester was reduced into corresponding
alcohol 6 by LiAlH4. The Mitsunobu reaction of 6 with
heterocycles, including adenine, N-benzyladenine and
Ester and hydroxamate analogues of methionyl adeny-
late, 1 and 2, as standards were examined with
IC50=5.0and 27 mM, respectively.11 The introduction
of one carbon (ester: 23 with IC50=3.6 mM; hydro-
Scheme 1. Synthesis of two-carbon linker analogues. Reagents and conditions: (a) H2SO4, acetone, 87%; (b) Ph3PCHCO2Me, MeCN, 85%;
(c) TBSCl, imidazole, THF, 97%; (d) LiAlH4, THF, 77%; (e) heterocycle, DEAD, PPh3, THF, 60–98%; (f) Bu4NF, THF, 70–98%; (g)
RCH(NHBoc)CO2H, DCC, DMAP, CH2Cl2, 70–95%; (h) TFA, anisole, 70–95%; (i) RCH(NHBoc)CONHOPMB, PPh3, THF, 58–90%.

J. Lee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 961–964
963
Scheme 2. Synthesis of one-carbon linker analogues. Reagents and conditions: (a) MsCl, NEt3, CH2Cl2, 95%; (b) t-BuOK, THF, 50%; (c) O3,
CH2Cl2 (d) NaBH4, 70% in two steps; (e) heterocycle, DEAD, PPh3, THF, 60–94% (f) Bu4NF, 87–96%; (g) RCH(NHBoc)CO2H, DCC, DMAP,
CH2Cl2, 63–93%; (h) TFA, anisole, 75–98%; (i) RCH(NHBoc)CONHOPMB, PPh3, THF, 60–75%.
xamate: 27 with IC50=26 mM) or two carbon (ester: 8
with IC50=8.3 mM, hydroxamate: 14 with IC50=56 mM)
as a linker between 10-position of ribose and adenine did
not significantly affect the inhibitory activity of the parent
compounds, 1 and 2. However, the order of inhibitory
activities was consistently expressed as one-carbon lin-
kerꢀparent>two-carbon linker in both the ester
(23ꢀ1>8) and hydroxamte (27ꢀ2>14) template. Par-
ticularly ester 23 is found to be the most potent one in
this series. The SAR analysis revealed that the catalytic
pocket of E. coli MetRS to bind the adenine moiety of
methionyl adenylate would be large enough to afford
structural variations. It appears that one-carbon linker
is the optimal length for inhibition. Replacement of the
adenine group with N-benzyladenine and 4-phenyltet-
razole, as more lipophilic and bulky, did not improve
their inhibitory activities in both ester (one-carbon lin-
ker: 24 and 25; two-carbon linker: 9 and 10) and
hydroxamate analogues (one-carbon linker: 28 and 29;
two-carbon linker: 15 and 16). However, these analogues
also represented a similar relationship, as shown in the
adenine series: ester analogues with one-carbon linker,
24 and 25 were found to be the most active in each series
(24>9, 15, 28 and 25>10, 16, 29), respectively (Table 1).
In isoleucyl adenylate analogues, ester 3 and hydrox-
amate 4 analogues unexpectedly showed poor inhibitory
activity to E. coli IleRS with IC50=183 and 181 mM,
respectively. The result was in strong contrast to sulfa-
mate analogues of isoleucyl adenylate previously repor-
ted, which represented the nanomolar range of enzyme
inhibitory activity.3 Structural modification on the ade-
nine part, as MRS inhibitors did above, also had no
influence on their inhibitory activities. According to the
SAR of sulfamate analogues of isoleucyl adenylate,
their inhibitory activities to E. coli IleRS improved as
the adenine moiety was removed from ribose, and
replaced by more lipophilic aromatic rings. The result
indicated that E. coli IleRS had a wide hydrophobic
pocket on the adenine binding site, which was enough
to interact with a long and lipophilic side chain. This
finding was also confirmed from two inhibitors, 12 and
13, with two-carbon linkers and a more lipophilic ade-
nine surrogate, which displayed better activity than the
parent compounds, 3 and 4, as compared to the sulfa-
mate analogues.
A recent report on the X-ray crystal structure of E. coli
MetRS14 prompted us to direct a receptor-guided
approach for investigating MetRS inhibitors. Based on
the data, we proposed a model of the active site docked
Table 1. Enzyme inhibitory activities of target compounds
MetRS
inhibitors
E. coli MetRS
(IC50, mM)
IleRS
inhibitors
E. coli IleRS
(IC50, mM)
1
2
5.0
27
3
4
183
181
8
9
8.3
11
12
13
17
18
19
26
30
>128
84
33.2
>256
56
10
14
15
16
23
24
25
27
28
29
63.6
>128
>128
>128
>128
>128
23
34
3.6
16.5
12.2
26
63
17
Figure 1. Proposed model of E. coli MetRS active site docked with
compound 23.

964
J. Lee et al. / Bioorg. Med. Chem. Lett. 11 (2001) 961–964
with a potent MetRS inhibitor 23 by molecular model-
ing (Fig. 1). The docking conformation of 23 into the
active site of MetRS was obtained from energy mini-
mization, followed by an alignment procedure on the
reported X-ray structure of aminoacyl adenylate bound
into aminoacyl-tRNA synthetases using GSAP.15 We
referred to a previous model of the active site proposed
by the site-directed mutagenesis of the methionine
binding site.16 The docking study was performed using
the DOCK procedure of the Sybyl 6.6. 23 was rigidly
docked into the binding site using graphical manipula-
tion with continuous energy monitoring. The manually
docked local energy minimized receptor–ligand complex
was subjected to an additional conjugate gradient mini-
mization using the minimization criteria. The result is
represented in Figure 1. In this model, two essential
interactions in methionine were examined as reported
previously.16 While the sulfur atom interacted with
phenolic hydroxyl of Tyr15 by hydrogen bonding, the
ammonium group coordinated with both carbonyl and
carboxylate of Asp52 by dipole-ionic and ionic interac-
tion. We also found two hydrogen bonds in adenosine,
whose 30-OH and 6-NH2 interacted with the NH of
Tyr15 and carbonyl of His24, respectively. Further
examination indicated that the enzyme contained a deep
and narrow hydrophobic pocket around the binding site
of the adenine base in order to afford binding of syn-
thesized MetRS inhibitors with methylene and ethylene
linkers. Although this model did not consider the parti-
cipation of water during the ligand–receptor interaction,
due to the lack of its X-ray structure, it will be helpful
for our continuing receptor-guided investigation into
MetRS inhibitors.
Acknowledgements
This work was supported in part by a grant (HMP-00-
CH-15-0014) from the Ministry of Health & Welfare,
R.O.K. the Brain Korea 21 Project to J. Lee, as well as
a grant from the National Creative Research Initiatives
to S. Kim.
References and Notes
1. van der Haar, F.; Gabius, H. J.; Cramer, F. Angew. Chem.,
Int. Ed. Engl. 1981, 20, 217.
2. De Pouplana, L. R.; Schimmel, P. Cell. Mol. Life Sci. 2000,
57, 865.
3. Schimmel, P.; Tao, J.; Hill, J. FASEB J. 1998, 12, 1599.
4. Renau, T. E.; Hecker, S. J.; Lee, V. J. Annu. Rep. Med.
Chem. 1998, 33, 121.
5. Ueda, H.; Shoku, Y.; Hayashi, N.; Mitsunaga, J.; In, Y.; Doi,
M.; Inoue, M.; Ishida, T. Biochim. Biophys. Acta 1990, 1080, 126.
6. Forrest, A. K.; Jarvest, R. L.; Mensah, L. M.; O’Hanlon,
P. J.; Pope, A. J.; Sheppard, R. J. Bioorg. Med. Chem. Lett.
2000, 10, 1871.
7. Hill, J. M.; Yu, G.; Shue, Y.-K.; Zydowsky, T. M.; Rebek,
J. US Patent 5,726,195.
8. Heacock, D.; Forsyth, C. J.; Kiyotaka, S.; Musier-Forsyth,
K. Bioorg. Chem. 1996, 24, 273.
9. Belrhali, H.; Yaremchuk, A.; Tukalo, M.; Larsen, K.; Ber-
thet-Colominas, C.; Leberman, R.; Beijer, B.; Sproat, B.; Als-
Nielsen, J.; Grbel, G.; Legrand, J.-F.; Lehmann, M.; Cusack,
S. Science 1994, 263, 1432.
10. Brown, P.; Richardson, C. M.; Mensah, L. M.; O’Hanlon,
P. J.; Osborne, N. F.; Pope, A. J.; Walker, G. Bioorg. Med.
Chem. 1999, 7, 2473.
11. Lee, J.; Kang, S. U.; Kang, M. K.; Chun, M. W.; Jo, Y. J.;
Kwak, J. H.; Kim, S. Bioorg. Med. Chem. Lett. 1999, 9, 1365.
12. Ohrui, H.; Jones, G. H.; Moffatt, J. G.; Maddox, M. L.;
Christensen, A. T.; Byram, S. K. J. Am. Chem. Soc. 1975, 97,
4602.
13. Arnez, J. G.; Moras, D. TIBS 1997, 22, 211.
14. Mechulam, Y.; Schmitt, E.; Maveyraud, L.; Zelwer, C.;
Nureki, O.; Yokoyama, S.; Konno, M.; Blanquet, S. J. Mol.
Biol. 1999, 294, 1287.
15. The structure of 23 was constructed using the molecular
modeling program Sybyl 6.6. Energy minimization was per-
formed using the Tripos force field and conjugate gradient
method until the rms energy gradient was below 0.005 kcal/
mol A on a Silicon Graphics INDIGO 2 workstation. The
charge was calculated by using the Gasteiger-Huckel method.
ˆ
In summary, the SAR on a series of ester and hydrox-
amate analogues of methionyl and isoleucyl adenylate
has been investigated through adenine surrogates bear-
ing carbon linkers between them and the 10-position of
ribose. In methionyl adenylate analogues, ester analo-
gues with a one-carbon linker appeared to be optimal
for the E. coli MetRS inhibitory activity, and 23
emerged as a potent candidate. SARs revealed that the
adenine binding site of E. coli MetRS would be a flex-
ible hydrophobic pocket large enough to afford bulky
adenine surrogates with one- or two-carbon linkers,
whose existence was examined by a modeling study. A
model of the E. coli MetRS active site derived from the
X-ray structure of the enzyme and inhibitor 23, was
proposed in order to explain the ligand–enzyme inter-
action. It will be helpful in the further study of the
receptor-guided inhibitor design.
˚
The energy-minimized structure of 23 was aligned to histidyl
adenylate (Arnez, J. G. et al. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 7144) using GASP (Genetic Algorithm Similarity
Program, Tripos). During the calculation, the conformation of
histidyl adenylate was fixed as a reference structure.
16. Kim, H. Y.; Ghosh, G.; Schulman, L. H.; Brunie, S.;
Jakubowski, H. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 11553.