Mechanism-based inhibition of an essential bacterial adenine DNA methyltransferase: Rationally designed antibiotics [1]

Full text of DOI:10.1021/ja003285o

976
J. Am. Chem. Soc. 2001, 123, 976-977
Communications to the Editor
Dam methylase, or similar regulatory adenine DNA MTases.
Adenine DNA MTase inhibitors would also act on RM systems
and provide further beneficial bacteriostatic effect.
Mechanism-Based Inhibition of an Essential Bacterial
Adenine DNA Methyltransferase: Rationally
Designed Antibiotics
Daphne C. Wahnon, Vincent K. Shier, and
Stephen J. Benkovic*
The PennsylVania State UniVersity
Department of Chemistry, 414 Wartik Laboratory
UniVersity Park, PennsylVania, 16802
ReceiVed September 6, 2000
The emergence of bacteria resistant to available antibiotics has
caused great concern in the medical community and has created
a need for the discovery of novel antibiotic agents. In the past
new antibiotics have been discovered by the random screening
of natural products and the subsequent identification of the target
protein. In addition to this drug discovery method, the recent
advances in genome sequencing have made it possible to envision
a complementary drug discovery method in which a bacterial
enzyme target is first identified and new antibiotic compounds
are discovered from the targeted inhibition of this enzyme.
Successful antibiotics would be inhibitors of an essential bacterial
enzyme that is unique to bacteria and has no mammalian
homologue. Here we report on the progress toward the develop-
ment of small-molecule selective inhibitors of an essential
bacterial N-6 adenine DNA methyltransferase (MTase), using a
mechanism-based multisubstrate adduct approach and demonstrate
the inhibition using CcrM (cell cycle regulated DNA MTase) from
the pathogenic Brucella abortus.
DNA methylation is an important and ubiquitous biological
event that encodes an additional level of information to the “four
base” genetic code. The majority of DNA MTases in bacteria
belong to restriction modification systems (RM) that protect the
bacterial host DNA from invading viruses. CcrM, however, lacks
the cognate restriction enzyme of RM systems, serves a critical
regulatory role in the correct progression of cell cycle events,1
and is essential to cellular viability.2 In addition, CcrM is
conserved among several bacterial species including pathogens
such as B. abortus, Helicobacter pylori, and Haemophilus
influenza.3 In mammals, DNA methylation is also critical in
regulating higher functions of the genome including replication,
and gene expression and is intimately linked to the biochemistry
of some cancers.4 N-6 adenine DNA MTase activity is not found
in mammalian cells, however, and instead occurs at C-5 cytosine.5
For these reasons CcrM is implicated as an ideal target for
antibacterial therapy. A second regulatory bacterial N-6 DNA
MTase Dam is not essential to cellular viability but has been
recently demonstrated to be essential to bacterial virulence in
Salmonella typhimurium.6 Inhibitors of N-6 adenine would result
in potent compounds against bacterial strains that contain CcrM,
Figure 1. (A) Reaction catalyzed by CcrM: a direct methyl transfer
from SAM to deoxyadenosine of GANTC. (B) Sinefungin, a natural
product methyltransferase inhibitor.
Enzymatic N-6 adenine DNA methylation (Figure 1A) proceeds
in a direct fashion in contrast to the mechanism observed in
solution.7 To access the exocyclic amine, which is normally
involved in Watson-Crick base pairing, the target adenine is
flipped away from the double helix and consequently positioned
proximally to the cofactor S-adenosylmethionine (SAM) to allow
direct methyl transfer to occur.8 In CcrM, the target base is
contained within the recognition sequence GANTC. Known
inhibitors of DNA methylation are analogues of SAM, including
the natural product sinefungin (Figure 1B). These compounds are
reasonable inhibitors but target all methylation events in the cell.
Clinical use of these compounds is limited by their toxicity.9
We reasoned that by extending from N-6 of the substrate
adenine into the SAM binding pocket and tethering the substrate
to the cofactor we might retain the inhibition observed with
cofactor analogues and induce selectivity from the additional
binding in the target adenine binding site. Similar multisubstrate
adduct approaches (the covalent attachment of two enzyme
substrates to form a single molecule) have been shown to increase
binding affinity and specificity of the target enzyme.10 Compounds
5-8 were designed as partial multisubstrate adducts for N-6
adenine DNA MTase.
Adapting the methodology developed for the synthesis of N-6
arylthiomethyl ribonucleosides11 it was possible to obtain com-
pounds 5-8 from adenosine, adenosine 5′-phosphate, 2′-deoxy-
adenosine, and 2′-deoxyadenosine 5′-phosphate. Hydrolysis of
N-acetyl homocysteine thiolactone under deoxygenated conditions
afforded the N-acetyl-DL-homocysteine that was used in the
subsequent N-6 alkylthiomethylation step (Scheme 1). Reaction
solutions were kept near pH 5 during the synthesis of the 2′-
(7) (a) Pogolotti, A. L., Jr.; Subramaniam, R.; Santi, D. V. J. Biol. Chem.
1988, 263, 7461. (b) Ho, D.; Wu, J.; Santi, D.; Floss, H. Arch. Biochem.
Biophys. 1991, 284, 264.
(8) (a)Allan, B.; Beechem, J.; Lindstom, W.; Reich, N. J. Biol. Chem. 1998,
273, 2368. (b) Holz, B.; Klimasauskas, S.; Seva, S.; Weinhold, E. Nucl. Acids
Res. 1998, 26, 1076.
(9) Ghosh, A. K.; Liu, W. J. Org. Chem. 1996, 61, 6175. Zwerygerth, E.;
Schillinger, D.; Kaufmann, W.; Roettcher, D. Trop. Med. Parasitol. 1986,
37, 255.
(10) (a) Broom, A. D. J. Med. Chem. 1989, 32, 2. (b) Inglese, J.; Blatchly,
R. A.; Benkovic, S. J. J. Med. Chem. 1989, 32, 937. (c) Inglese, J.; Benkovic,
S. J. Tetrahedron 1991, 47, 2351.
* To whom correspondence should be addressed. Telephone: 814-865-
2882. Fax: 814-865-2973.
(1) (a) Zweiger, G.; Marczynski, G.; Shapiro, L. J. Mol. Biol. 1994, 235,
472. (b) Stephen, C.; Zweiger, G.; Shapiro, L. J. Bacteriol. 1995, 177, 1662.
(2) Stephens, C.; Reisenauer, A.; Wright R.; Saphiro L. Proc. Natl. Acad.
Sci. U.S.A. 1996, 93, 1210.
(3) Wright, R.; Stephens, C.; Shapiro, L. J. Bacteriol. 1997, 179, 5869.
(4) Szyf, M. Pharmacol. Ther. 1996, 70, 1.
(5) (a) Malone, T.; Blumenthal, R.; Cheng, X. J. Mol. Biol. 1995, 253,
618. (b) Razin, A.; Szyf, M. Biochim. Biophys. Acta 1984, 782, 331. (c)
Adams, R. BioEssays 1995, 17, 139.
(11) (a) Kemal, O.; Reese, C. Synthesis, 1980, 1025-1028. (b) Yamazaki,
Y.; Suzuki, H.; Kamibayashi, A.; Watanabe, N.; Takahashi, N. Agric. Biol.
Chem. 1979, 43, 1945.
(6) Heithoff, M. D.; Sinsheimer, R. L.; Low, D. A.; Mahan, M. J. Science
1999, 284, 967.
10.1021/ja003285o CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/13/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 977
Table 1. Ki’s of Compounds 5-8
with compounds 7 and 8 even at concentrations 10-fold greater
than required to inhibit CcrM activity (Figure 2). Sinefungin, in
contrast, was found to inhibit both CcrM and HhaI. To test the
hypothesis that inhibition of adenine DNA MTase would lead to
inhibition of bacterial cell growth we tested compounds 7 and 8
in an in vivo cell growth assay. We carried out the cell growth
assays using C. crescentus as a model for the pathogenic B.
abortus since CcrM from B. abortus (B.CcrM) has a 65% identity
and 76% homology to CcrM from C. crescentus (C.CcrM).
Furthermore, these compounds inhibit CcrM from both bacteria
with nearly identical KI values (Table 1). Inhibition of cell growth
was observed in the presence of compounds 7 and 8 with an IC50
of about 500 µM. It is clear from these results that CcrM is
inhibited at much lower concentrations than is required to produce
an antibiotic effect in vivo. It is likely that these compounds do
not easily pass through the bacterial cell wall to contact the target
enzyme. We are now extending this work to include the synthesis
of non-nucleoside analogue inhibitors and have observed that
small molecule inhibitors, which are able to pass through the cell
wall, are good inhibitors of bacterial cell growth. For the most
effective inhibitors cell growth inhibition data can be correlated
to enzyme inhibition data supporting a mechanism-based inhibi-
tion of cell growth (data not shown).
deoxyadenosine analogues to avoid depurination side products.
Compounds 1-4 were purified using reverse phase HPLC,
cellulose chromatography or DEAE sephadex ion exchange
chromatography. N-acetyl deprotection of 1-4 was achieved using
acylase I and yielded enantiomerically pure 5-8 in a kinetic
resolution step. 5-8 were purified by cellulose chromatography
to remove salts and enzyme and were stable for a several days at
30 °C in 50 mM, pH 7.4 HEPES buffer (assay conditions)
although significant decomposition was observed in both basic
(>pH 8) and acidic aqueous solutions.
Scheme 1. Reaction and Conditions: (a) HOAc, 7:3 Ethanol:
Water, Reflux, 2 Days; (b) Acylase I, pH 7 phosphate buffer
We have previously reported on the purification and mechanism
of CcrM from Caulobacter crescentus.12 We now report on the
purification of CcrM from the pathogenic B. abortus (B.CcrM).
B.CcrM was overexpressed in Escherichia coli as an N-terminal
six histidine tagged protein (B.CcrM) and purified using standard
Ni NTA affinity chromatography procedures. This represents the
first purification of this CcrM homologue. In a typical purification,
10 mg of >95% pure protein was obtained per two liter of
bacterial culture. The purified B.CcrM is functional and methylates
the synthetic substrate hemi-methylated N-6 45/50 mer as
determined by the tritium incorporation filter-binding assay
previously described.
Compounds 5-8 were shown to inhibit B.CcrM activity with
Ki’s ranging from 5 to 30 µM (Table 1). Adenosine and
homocysteine individually had no effect on MTase activity. The
tethering of these two compounds results in a cooperative effect
on binding to B.CcrM and is consistent with the active site binding
of the partial multi substrate adduct inhibitor. Since 5-8 were
assayed in the presence of their respective N-acetylated precursors,
we assayed 1-4 separately to determine the effect of these
compounds. We observed no inhibition for 1-4 at concentrations
where 5-8 completely inhibited the enzymatic reaction. As can
be seen from Table 1, inhibition was not greatly influenced by
the presence of a 5′-phosphate group or the absence of a
2′-hydroxyl.
Figure 2. Selective Inhibition: Percent reaction for sinefungin, 7, and
8 for CcrM and HhaI. CcrM Assays: [Sinefungin] ) 75 µM, [I] )
100 µM. Hha I Assays: [Sinefungin] ) 200 µM, [I] ) 1 mM.
In conclusion, we have described the purification of a novel
cell cycle regulated adenine DNA MTase from the pathogenic
B. abortus and presented a mechanism-based strategy for the
selective inhibition of this enzyme. We have observed inhibition
comparable to the natural product sinefungin, but in contrast to
sinefungin, the observed inhibition is selective for adenine DNA
methylation. We have demonstrated that inhibitors of adenine
DNA MTase have antibiotic effects in a cell strain containing a
regulatory CcrM and we are now focusing on the synthesis and
screening of small-molecule combinatorial libraries to develop
lead antibiotic candidates. These studies demonstrate that the
essential regulatory bacterial adenine DNA MTases are an
attractive target for the design of novel antibiotics. As sequences
of bacterial genomes are completed and as additional bacterial
cell cycle regulators are discovered the available targets for
rationally designed antibiotics will expand.
Acknowledgment. We thank Rachel Wright (Department of Devel-
opmental Biology, Stanford University) for the Brucella abortus and
Caulobacter crescentus CcrM gene. This work was sponsored by DARPA
Grant MDA972-97-1-0008.
We used HhaI, a bacterial C5 cytosine DNA MTase to
determine whether our inhibitors showed selectivity toward
adenine DNA MTase. No inhibition of HhaI activity was observed
Supporting Information Available: Experimental details and char-
acterization data (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
(12) Berdis, A. J.; Lee, I.; Coward, J. K.; Stephens, C.; Wright, R.; Shapiro,
L.; Benkovic S. J. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 2874.
JA003285O