A molecular target for suppression of the evolution of antibiotic resistance: Inhibition of the Escherichia coli RecA protein by N 6-(1-naphthyl)-ADP

Article abstract of DOI:10.1021/jm050113z

We report that N⁶-(1-naphthyl)-ADP (1) inhibits the Escherichia coli RecA protein in vitro. A novel rapid screen identified 1 as a potent inhibitor of RecA nucleoprotein filament formation, and further characterization established 1 as an ATP-competitive inhibitor of RecA-catalyzed ATP hydrolysis. 1 and other inhibitors of RecA activities represent a new approach for understanding the molecular targets and pathways involved in the evolution of antibiotic resistance in bacteria.

Full text of DOI:10.1021/jm050113z

5408
J. Med. Chem. 2005, 48, 5408-5411
Scheme 1a
A Molecular Target for Suppression of
the Evolution of Antibiotic Resistance:
Inhibition of the Escherichia coli RecA
Protein by N6-(1-Naphthyl)-ADP
Andrew M. Lee, Christian T. Ross,
Bu-Bing Zeng, and Scott F. Singleton*
School of Pharmacy, Division of Medicinal Chemistry &
Natural Products, University of North Carolina at Chapel
Hill, CB 7360, Chapel Hill, North Carolina 27599-7360
Received February 5, 2005
a Reagents: (a) RNH2, EtOH, reflux, 59%; (b) (1) POCl3, Proton
Sponge, PO(OMe)3, 4 °C, (2) nBu3NH, H2PO4, pyridine, 0 °C, 13%.
Abstract: We report that N6-(1-naphthyl)-ADP (1) inhibits the
Escherichia coli RecA protein in vitro. A novel rapid screen
identified 1 as a potent inhibitor of RecA nucleoprotein
filament formation, and further characterization established
1 as an ATP-competitive inhibitor of RecA-catalyzed ATP
hydrolysis. 1 and other inhibitors of RecA activities represent
a new approach for understanding the molecular targets and
pathways involved in the evolution of antibiotic resistance in
bacteria.
relative orientation, the adenosine moiety in RecA is
located in a wide crevice near the surface of the protein
(Figure 1A). In contrast, the ATP-binding sites of other
NTPases and kinases envelop adenine in a hydrophobic
pocket and maintain close contact with the edge and
both faces of the purine moiety (Figure 1B), mitigating
their potential interactions with N6-substituted adenos-
ine nucleotides.12,13 Hence, we reasoned that the ATP-
binding site of RecA could uniquely accommodate a
sterically demanding substituent at N6 of adenine. Our
initial approach to inhibition of RecA thereby relies on
the principles of negative design: the achievement of
specificity involves the introduction of substituents that
prevent the formation of nonproductive alternative
complexes with nontargeted enzymes.
We chose to synthesize N6-substituted ADP analogues
including those depicted in Scheme 1. Of the compounds
in this class we have investigated, we selected 1-3
because they bear the largest N6-substituents and
because the 5′-O-triphosphates related to 2 and 3 have
been shown to be poor substrates of protein kinases,
myosin, and kinesin.13 The syntheses began with the
amination of 6-chloropurine ribonucleoside.14 The re-
sulting nucleosides were readily converted to the 5′-O-
diphosphates (see Supporting Information).
In conjunction with the synthesis of potential inhibi-
tors, we developed a novel high-throughput screening
compatible assay to take advantage of the fact that
binding of ADP can induce the release of RecA from
ssDNA. In the context of mechanistic studies, we
recently demonstrated that binding of ATP or ADP to
a preformed RecA-ssDNA complex results in the re-
lease of the bound protein molecules, followed by reas-
sociation of the ternary complex in the active confor-
mation.7 This discovery led us to consider whether RecA
could be prevented from binding to ssDNA immobilized
on streptavidin paramagnetic particles (SA-PMP).
Capturing biotinylated-(dT)36 on SA-PMP allowed us
to determine whether RecA is prevented from forming
an NPF in the presence of a putative inhibitor by
measuring the amount of RecA in the supernatant using
a Bradford assay. In practice, the assay is conducted
by incubating the nucleotide with RecA and (dT)36, and
then immobilizing the ssDNA and any bound RecA. This
allows us to take advantage of the fact that binding of
Antibiotic resistance in pathogenic bacteria has enor-
mous human and economic consequences worldwide.
The rapid rate at which bacteria develop drug resistance
is due in large part to mutations arising during stress-
induced DNA repair1 and during the lateral transfer of
genes between organisms.2 The bacterial RecA protein
is essential to both of these processes. In addition, RecA
function is required for aspects of pathogenicity, includ-
ing antibiotic-induced responses to ciprofloxacin3 and
â-lactams,4 antigenic variation in Neisseriae,5 and the
induction of shiga toxin production.6 All RecA functions
require formation of an active nucleoprotein filament
(NPF) comprising multiple RecA monomers, ATP, and
single-stranded DNA (ssDNA). Hence, the discovery of
small molecules that suppress the formation of such
NPFs would be an important step in the development
of inhibitors for the suppression of the evolution and
transmission of antibiotic resistance. On the basis of our
recent investigations of RecA mechanism,7 we designed
putative small-molecule inhibitors of RecA and a rapid
microplate screening assay for its inhibition. Herein, we
report the synthesis of N6-(1-naphthyl)-ADP (1, Scheme
1) and demonstrate that it inhibits formation of the
RecA-ssDNA complex that controls bacterial recombi-
nation and stress-induced mutagenesis.
Like most ATPases, RecA can be inhibited in vitro
by ADP and a variety of nonhydrolyzable derivatives
of ATP.9,10 Our rationale for selective inhibitor design
was inspired by the Steitz laboratory’s initial observa-
tion that the orientation of ADP bound to protein in
RecA crystals was different from those of nucleotides
bound to related NTP-binding proteins.11 To extend this
investigation, we comparatively analyzed the RecA
crystal structure with those of homologous motor pro-
teins, such as F1-ATPase, T7 helicase, and the rho
transcription terminator, as well as other P-loop
NTPases and kinases. As a result of ADP’s unusual
* To whom correspondence should be addressed. Telephone: 919-
966-7954. Fax: 919-966-0204. E-mail: sfs@unc.edu.
10.1021/jm050113z CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/29/2005

Letters
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 17 5409
known about the ability of various nucleotides to bind
and activate (or deactivate) the RecA nucleoprotein
filament. Indeed, the assay results (Figure 2b) demon-
strate the expected trend in the relative NPF stabili-
ties: RecA-ADP < RecA-only. Moreover, ATP has the
same apparent effect on NPF stability as does ADP,
consistent with the hydrolysis of ATP to ADP during
the experiment. The fact that the amount of RecA
protein released in the presence or absence of ADP/ATP
is significantly less than that released by the addition
of NaCl (final concentration of 1.5 M) verifies that
substantial protein is coating the immobilized (dT)36.
The results for the nucleotides that are known to be
RecA substrates were qualitatively similar to those for
ADP/ATP.15
Using this assay, we investigated the release of RecA
induced by three synthetic ADP analogues, N6-(1-
naphthyl)-, N6-(benzyl)-, and N6-(2-phenethyl)-ADP (1,
2, and 3, respectively; Figure 2b). Compound 1 (100 µM)
inhibited NPF formation to the greatest extent of any
agent examined to date. As a control, we observed that
the N6-(1-naphthyl)-adenosine monophosphate did not
inhibit NPF formation, the extent of formation being
the same within error as that observed in the absence
of added nucleotide. We conclude that N6-(1-naphthyl)-
AMP does not bind RecA under these conditions.
Importantly, the efficacy of 1 relative to its monophos-
phate for inhibition of NPF formation parallels that
known for ADP and AMP.16 The observation that the
inhibitory effect of 1 responds to the phosphorylation
state of the nucleoside suggests that its action does not
result exclusively from interactions with the naphthyl
moiety.
In contrast to those results for 1, 100 µM 2 or 3
inhibited NPF formation to the same or lesser extent
as 100 µM ADP. In this context, we tentatively conclude
that 2 and 3 bind RecA similarly to ADP. Nucleotides 2
and 3 have been successfully used by the Shokat13a,b and
Mitchison13c laboratories to control the functions of
bioengineered (mutant) myosins, protein kinases, and
kinesins.17 The superiority of 1 for the inhibition of wild-
type RecA likely results from its spacious surface crevice
in which the ADP analogues are accommodated.
Figure 1. Comparison of the ATP-binding sites of RecA (A)
and smooth muscle myosin (B). (Left panels) The structures
were aligned with SwissPDB Viewer8 using the CR atoms of
the residues in the phosphate-binding loop (P-loop) and the
R-helix and â-strand adjacent to each (cyan). The ADP
molecules are shown with atoms colored by elemental identity.
(Right panels) The Connolly surfaces of the respective proteins
and the bound ADP molecules (stick models) are shown
following identical rotations to emphasize the different relative
orientations of the nucleotides within the binding clefts.
The formation of an active NPF in the presence of
ATP results in ATP hydrolysis. By monitoring the
steady-state kinetics of ATP hydrolysis by RecA in a
poly(dT)-dependent ATPase assay,18 we found that 1
exhibited fully competitive inhibition with respect to
ATP and noncompetitive inhibition with respect to
ssDNA (see Supporting Information). These results
indicated that 1 binds in or near the RecA ATP-binding
site, confirming our expectation that the ATP-binding
cleft is large enough to accommodate the sterically
demanding naphthyl moiety. Although relatively high,
the apparent inhibition constant for 1 (Kic ) 46 ( 6 µM)
is essentially identical to the S0.5 value for ATP19,20 and
within the range of Kd and Ki values for ADP.9b The
observation that the affinities of 1, ADP, and ATP are
similar recapitulates the principles of negative design
used to guide our choice of potential inhibitors.
Figure 2. (a) Relative stabilities of NPFs formed in the
presence or absence of nucleotide as measured by Menetski
and Kowalczykowski using the salt titration midpoint (STMP,
ref 10). (b) Relative inhibition of NPF formation in the presence
of various nucleoside di- and triphosphates (100 µM each).
ADP to RecA reduces the apparent stability of the
RecA-ssDNA complex (Figure 2a).10
Using the inhibition of NPF formation assay, we
screened the 16 canonical nucleoside di- and triphos-
phates (both ribo- and 2′-deoxyribonucleotides) for RecA
inhibition. The results clearly recapitulated what is
We then conducted preliminary tests of the specificity
of 1 for inhibition of RecA. In particular, we evaluated
the inhibition of the ATPase activities of rabbit muscle
pyruvate kinase, E. coli rho transcription terminator,

5410 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 17
Letters
particularly rigorous test of its biological activity be-
cause the RecA protein is constantly overexpressed from
the trc promoter, whereas the normal basal expression
level for RecA is low and it is overexpressed only after
DNA damage induces the SOS response.23 Control
experiments performed using 1 in the absence of mito-
mycin C revealed no effect of 1 on the viability of
permeabilized GY7313 cells. Moreover, a preliminary
confirmation of a correlation between in vivo and in vitro
activities was provided by the observation that treat-
ment of permeabilized cells with 2 or 3 (100 µM)
resulted in no significant reduction in SOS (â-galactosi-
dase) induction (data not shown).
The observation that 1 mitigates the induction of SOS
in permeabilized E. coli cells provides powerful proof
that RecA activities, including its role in processes
leading to the development and transfer of antibiotic
resistance genes, may be selectively controlled by small
molecules in living bacteria. Diphosphates such as 1 are,
however, likely to be of little therapeutic utility because
of membrane impermeability caused by the negative
charges on the 5′-diphosphate moiety at physiological
pH. Nevertheless, substantial progress has been made
in the development of effective strategies for the intra-
cellular delivery of nucleotide prodrugs (pronucleotides),
particularly for use as anticancer and antiviral thera-
peutics.24 Experiments to test the biological activities
of pronucleotides related to 1 are under way.
Figure 3. SOS induction in the absence and presence of 1
(100 µM). â-Galactosidase activities in permeabilized GY7313
bacteria were measured 2.5 h after treatment with mitomycin
C (MMC, 0.5 µg/mL) as described in the Supporting Informa-
tion. The gray and red bars represent activities measured after
SOS induction by MMC, while the white bars represent basal
activities measured without added MMC. The “∆recA” data
were obtained using GY7313 cells transformed by pTrc99A
without a recA gene.
and chicken muscle myosin. Pyruvate kinase is an
important glycolytic enzyme, the ATP binding site of
rho is a close structural homologue of the RecA site, and
the ATPase active site of myosin is known to tolerate a
variety of base-substituted ATP analogues.21,13b Taken
together, these three enzymes represent several classes
of ATPases and provide a reasonable test of the pro-
miscuity of 1. At concentrations up to 300 µM, 1 led to
no reduction in the rates of the three enzyme-catalyzed
reactions. These results provide initial evidence that 1
is a specific inhibitor of RecA and validate the principles
used to guide the design of 1.
As an initial test of the biological activity of 1, we
tested its ability to inhibit the induction of the SOS
response in permeabilized E. coli cells. As described
above, the RecA protein stimulates DNA repair and
mutagenesis processes by initiating the SOS response.
In the presence of ATP and Mg2+, the RecA protein
bound to ssDNA activates the LexA repressor for
autoproteolytic destruction and derepresses the genes
controlled by LexA-repressible promoters. To evaluate
the ability of 1 to modulate the SOS response to the
DNA-damaging agent mitomycin C, we measured â-ga-
lactosidase activity in E. coli strain GY7313 harboring
a lacZ reporter gene fused to the LexA-regulated sfi
promoter. RecA-activated autoproteolysis of LexA dere-
presses lacZ, and the â-galactosidase produced can be
quantitatively measured by an established colorimetric
assay.18
In summary, we have reported the discovery that N6-
(1-naphthyl)-ADP prevents formation of the RecA-DNA
filament that is essential for all RecA-associated func-
tions. The fact that 1 is an inhibitor of RecA activities
(IRA) provides essential proof of the principle that RecA
functions can be selectively controlled by synthetic small
molecules. Importantly, we have developed a novel
assay for nucleotide analogue-dependent inhibition of
NPF formation that is suitable for high-throughput
screening of synthetic inhibitor libraries. IRAs that are
active inside living bacteria will allow for the design of
future chemical biology experiments to tease apart long-
standing questions about the prokaryotic response to
genome-wide DNA damage.25 Moreover, we anticipate
that these studies will provoke the development of lead
compounds for inhibiting RecA-dependent processes
leading to the development and transfer of antibiotic
resistance.
Acknowledgment. This work was supported by a
grant from the National Institutes of Health to S.F.S.
(Grant GM58114). The authors thank Dr. Harold Kohn
and Mr. Andrew Brogan for carrying out the rho ATPase
assays, and Mr. Timothy Wigle and Mr. William Horton
for technical assistance.
For the investigation of ADP derivatives, the E. coli
cells were made permeable to nucleotides with a hypo-
tonic Tris-EDTA buffer at 4 °C. This treatment rendered
the cells permeable to propidium iodide,22 but impor-
tantly, they remained fully viable as judged by the
ability to form colonies following recovery from perme-
abilization (see Supporting Information for details).
Following permeabilization, â-galactosidase induction
in ∆recA GY7313 harboring a plasmid with the WT recA
gene expressed from the trc promoter was 7-fold greater
than the basal level 2.5 h after treatment of the
transformants with mitomycin C (Figure 3). Control
experiments using GY7313 cells harboring the plasmid
without a recA gene verify that the response is RecA-
dependent. When 100 µM 1 was added just prior to
treatment with mitomycin C, the level of induction was
reduced more than 55%. The significant inhibition of
â-galactosidase induction by 1 in this system is a
Supporting Information Available: Experimental pro-
cedures (including syntheses) and analytical data. This mate-
rial is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Hersh, M. N.; Ponder, R. G.; Hastings, P. J.; Rosenberg, S. M.
Adaptive mutation and amplification in Escherichia coli: two
pathways of genome adaptation under stress. Res. Microbiol.
2004, 155, 352-359. Matic, I.; Taddei, F.; Radman, M. Survival
versus maintenance of genetic stability: a conflict of priorities
during stress. Res. Microbiol. 2004, 155, 337-341. Foster, P. L.
Stress responses and genetic variation in bacteria. Mutat. Res.
2005, 569, 3-11.

Letters
(2) Beaber, J. W.; Hochhut, B.; Waldor, M. K. SOS response
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 17 5411
(15) For purposes of clarity, the data for these nucleotides and a
variety of synthetic analogues are not shown; a full account of
the structural basis for nucleotide selectivity will be published
elsewhere.
(16) Watanabe, R.; Masui, R.; Mikawa, T.; Takamatsu, S.; Kato, R.;
et al. Interaction of Escherichia coli RecA protein with ATP and
its analogues. J. Biochem. (Tokyo) 1994, 116, 960-966.
(17) For a recent review, see: Specht, K. M.; Shokat, K. M. The
emerging power of chemical genetics. Curr. Opin. Cell Biol. 2002,
14, 155-159.
(18) Berger, M. D.; Lee, A. M.; Simonette, R. A.; Jackson, B. E.; Roca,
A. I.; et al. Design and evaluation of a tryptophanless RecA
protein with wild type activity. Biochem. Biophys. Res. Commun.
2001, 286, 1195-1203.
(19) Stole, E.; Bryant, F. R. Reengineering the nucleotide cofactor
specificity of the RecA protein by mutation of aspartic acid 100.
J. Biol. Chem. 1996, 271, 18326-18328.
(20) The S0.5 and Kd values reported for ATP binding range from 2.5
to 200 µM. See ref 9 and the following: Kowalczykowski, S. C.
Interaction of recA protein with a photoaffinity analogue of ATP,
8-azido-ATP: determination of nucleotide cofactor binding pa-
rameters and of the relationship between ATP binding and ATP
hydrolysis. Biochemistry 1986, 25, 5872-5881.
(21) Tonomura, Y.; Imamura, K.; Ikehara, M.; Uno, H.; Harada, F.
Interaction between synthetic ATP analogues and actomyosin
systems. IV. J. Biochem. (Tokyo) 1967, 61, 460-472. Cremo, C.
R. Fluorescent nucleotides: synthesis and characterization.
Methods Enzymol. 2003, 360, 128-177.
(22) Bunthof, C. J.; van Schalkwijk, S.; Meijer, W.; Abee, T.; Hugen-
holtz, J. Fluorescent method for monitoring cheese starter
permeabilization and lysis. Appl. Environ. Microbiol. 2001, 67,
4264-4271.
(23) Sedgwick, S. G.; Levine, A.; Bailone, A. Induction of recA+-
protein synthesis in Escherichia coli. Mol. Gen. Genet. 1978, 160,
267-276. Casaregola, S.; D’Ari, R.; Huisman, O. Quantitative
evaluation of recA gene expression in Escherichia coli. Mol. Gen.
Genet. 1982, 185, 430-439.
promotes horizontal dissemination of antibiotic resistance genes.
Nature 2004, 427, 72-74. Hastings, P. J.; Rosenberg, S. M.;
Slack, A. Antibiotic-induced lateral transfer of antibiotic resis-
tance. Trends Microbiol. 2004, 12, 401-404.
(3) Bisognano, C.; Kelley, W. L.; Estoppey, T.; Francois, P.; Schren-
zel, J.; et al. A RecA-LexA-dependent pathway mediates cipro-
floxacin-induced fibronectin binding in Staphylococcus aureus.
J. Biol. Chem. 2004, 279, 9064-9071.
(4) Miller, C.; Thomsen, L. E.; Gaggero, C.; Mosseri, R.; Ingmer,
H.; et al. SOS response induction by beta-lactams and bacterial
defense against antibiotic lethality. Science 2004, 305, 1629-
1631.
(5) Kline, K. A.; Sechman, E. V.; Skaar, E. P.; Seifert, H. S.
Recombination, repair and replication in the pathogenic Neis-
seriae: the 3 R’s of molecular genetics of two human-specific
bacterial pathogens. Mol. Microbiol. 2003, 50, 3-13.
(6) Casjens, S. Prophages and bacterial genomics: what have we
learned so far? Mol. Microbiol. 2003, 49, 277-300.
(7) Roca, A. I.; Singleton, S. F. Direct evaluation of a mechanism
for activation of the RecA nucleoprotein filament. J. Am. Chem.
Soc. 2003, 125, 15366-15375.
(8) Guex, N.; Peitsch, M. C. SWISS-MODEL and the Swiss-
PdbViewer: an environment for comparative protein modeling.
Electrophoresis 1997, 18, 2714-2723.
(9) For examples, see the following and references therein: (a)
Moreau, P. L.; Carlier, M. F. RecA protein-promoted cleavage
of LexA repressor in the presence of ADP and structural
analogues of inorganic phosphate, the fluoride complexes of
aluminum and beryllium. J. Biol. Chem. 1989, 264, 2302-2306.
(b) Lee, J. W.; Cox, M. M. Inhibition of recA protein promoted
ATP hydrolysis. 1. ATPγS and ADP are antagonistic inhibitors.
Biochemistry 1990, 29, 7666-7676.
(10) Menetski, J. P.; Kowalczykowski, S. C. Interaction of recA
protein with single-stranded DNA. Quantitative aspects of
binding affinity modulation by nucleotide cofactors. J. Mol. Biol.
1985, 181, 281-295.
(11) Story, R. M.; Steitz, T. A. Structure of the recA protein-ADP
complex. Nature 1992, 355, 374-376.
(12) Chene, P. ATPases as drug targets: learning from their struc-
ture. Nat. Rev. Drug Discovery 2002, 1, 665-673. Mao, L.; Wang,
Y.; Liu, Y.; Hu, X. Molecular determinants for ATP-binding in
proteins: a data mining and quantum chemical analysis. J. Mol.
Biol. 2004, 336, 787-807.
(24) Krise, J. P.; Stella, V. J. Prodrugs of phosphates, phosphonates,
and phosphinates. Adv. Drug Delivery Rev. 1996, 19, 287-310.
Schultz, C. Prodrugs of biologically active phosphate esters.
Bioorg. Med. Chem. 2003, 11, 885-898. Meier, C.; Ruppel, M.
F.; Vukadinovic, D.; Balzarini, J. “Lock-in”-cycloSal-pronucleo-
tidessa new generation of chemical Trojan horses? Mini-Rev.
Med. Chem. 2004, 4, 383-394. Peyrottes, S.; Egron, D.; Lefebvre,
I.; Gosselin, G.; Imbach, J. L.; et al. SATE pronucleotide
approaches: an overview. Mini-Rev. Med. Chem. 2004, 4, 395-
408. Drontle, D. P.; Wagner, C. R. Designing a pronucleotide
stratagem: lessons from amino acid phosphoramidates of anti-
cancer and antiviral pyrimidines. Mini-Rev Med Chem 2004, 4,
409-419.
(13) For examples concerning kinases, myosin, and kinesin, see: (a)
Shah, K.; Liu, Y.; Deirmengian, C.; Shokat, K. M. Engineering
unnatural nucleotide specificity for Rous sarcoma virus tyrosine
kinase to uniquely label its direct substrates. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 3565-3570. (b) Gillespie, P. G.; Gillespie,
S. K.; Mercer, J. A.; Shah, K.; Shokat, K. M. Engineering of the
myosin-ibeta nucleotide-binding pocket to create selective sen-
sitivity to N(6)-modified ADP analogs. J. Biol. Chem. 1999, 274,
31373-31381. (c) Kapoor, T. M.; Mitchison, T. J. Allele-specific
activators and inhibitors for kinesin. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 9106-9111.0
(25) Cox, M. M. Recombinational DNA repair of damaged replication
forks in Escherichia coli: questions. Annu. Rev. Genet. 2001,
35, 53-82.
(14) Kikugawa, K.; Iizuka, K.; Ichino, M. Platelet aggregation inhibi-
tors. 4. N6 -substituted adenosines. J. Med. Chem. 1973, 16,
358-364.
JM050113Z