ª 2009 John Wiley & Sons A/S

Chem Biol Drug Des 2009; 74: 343–348 doi: 10.1111/j.1747-0285.2009.00879.x

Research Article

Antibiotic Resistance in Bacteria: Novel

Metalloenzyme Inhibitors

Sung-Kun Kim1, Cynthe L. Sims2, Susan E. 3-(N-morpholino) propanesulfonic acid; PCI, Phenol:chloroform:isoamyl
Wozniak2, Stephanie H. Drude2 Dustin alcohol; PCR, polymerase chain reaction; ssDNA, single stranded DNA;
TBE, tris–borate EDTA; TEMED, tetramethylene diamine.
Whitson2 and Robert W. Shaw2,*
1
Received 30 April 2009, revised 10 August 2009 and accepted for publi-
Department of Chemistry and Biochemistry, Baylor University, cation 11 August 2009
Waco, TX 76706, USA
2
Department of Chemistry and Biochemistry, Texas Tech University,
Lubbock, TX 79409-1061, USA
*Corresponding author: Robert W. Shaw, [email protected] b-Lactam antibiotics such as penicillins, cephalosporins and carba-
penems are widely used as antimicrobial drugs because they are
b-Lactam antibiotics are among the most impor- relatively inexpensive and have been very effective (1). Pathogenic
tant drugs used to fight bacterial infection. Over- bacteria are becoming increasingly resistant to b-lactam antibiotics.
use and misuse of b-lactam antibiotics has A common reason for this resistance is the inheritance of genes for
caused the evolution of resistance mechanisms, the expression of efficient enzymes called b-lactamases that cata-
allowing pathogenic bacteria to survive antibiotic lyze the hydrolytic inactivation of the b-lactam rings of the antibiot-
treatment. The major source of resistance to b- ics. The metal-dependent metallo-b-lactamases are the most
lactam antibiotics occurs through production of
problematic of these enzymes because of their wide range of sub-
enzymes called b-lactamases capable of catalyz-
ing hydrolysis of the b-lactam rings in these drug strate specificity and current lack of clinically useful b-lactamase
compounds. The metallo-b-lactamases have inhibitors (2). Study of such enzyme inhibitors is important clinically
become a major threat due to their broad sub- for the purposes of developing compounds that can be used in com-
strate specificities; there are no clinically useful bination with existing b-lactam antibiotics. We have utilized a com-
inhibitors for these metalloenzymes. We have binatorial chemistry and in vitro screening technique called SELEX
obtained single-stranded DNA’s that are potent (Systematic Evolution of Ligands by EXponential Enrichment) (3,4) to
inhibitors of the Bacillus cereus 5/B/6 metallo-b- obtain potent single straded DNA (ssDNA) inhibitors of the metallo-
lactamase. These are rapid, reversible, non-com- b-lactamase from the organism Bacillus cereus 5 ⁄ B ⁄ 6. We report
petitive inhibitors of the metalloenzyme, with Ki
steady state enzyme kinetic and electronic spectral data that sug-
and Ki¢ values in the nanomolar range. The inhi-
bition patterns and metal ion dependence of gest that the ssDNA can rapidly, reversibly inhibit this metallo-
their inhibition suggest that the oligonucleotides b-lactamase in a fashion that suggests that the inhibitors interfere
alter the coordination of the active site metal with the active site metal ion(s) (5). Furthermore, combinations of
ion(s); inhibition is efficient and highly specific. the b-lactam antibiotic cephalexin with the ssDNA can lead to a
Microbiological growth experiments, using com- suppression of the growth of both Gram-positive (6) and Gram-neg-
binations of ssDNA with the b-lactam antibiotic ative bacteria on agar plates or in liquid cultures.
cephalexin, reveal that the inhibitor is capable of
causing cell death in liquid cultures of both
Gram-positive and Gram-negative metallo-b-lac-
Methods and Materials
tamase producing bacteria in the micromolar
concentration range.
T4-DNA ligase was purchased from Promega (Madison, WI, USA).
Key words: antibiotic resistance, inhibitor, metallo-b-lactamase, Restriction endonucleases NdeI and SacI were purchased from New
metalloenzyme England Biolabs (Ipswich, MA, USA), Inc. and were used according
to the manufacturer's recommendations. DNA molecular weight
Abbreviations: EPR, electron paramagnetic resonance; EXAFS, markers (BstNI digested pBR322 and BstEII digested k DNA) were
Extended X-ray absorption fine structure; LB, Luria-Bertani; MOPS, purchased from New England Biolabs, Inc. DEAE-Sephacel, Sephadex

This is an invited publication from the International Symposium on Organic Synthesis and Drug Develop-
ment (2008). CB&DD honorary editors of this manuscript: Dr. Rao Rapaka, Prof. Guoqiang Lin, Prof. Yi Pan
and Prof. Guigen Li.

343
Kim et al.

G-25 (superfine), CM-Sepharose CL 6B and various columns were sodium chloride. In a 1 cm cuvette, 1.0 mL of substrate was added
purchased from Pharmacia (Uppsala, Sweden) or Bio-Rad Laborato- and incubated in the spectrophotometer at 25 C for 3–4 min to
ries (Hercules, CA, USA). The Gene Clean II Kit was purchased from reach temperature equilibration and establish a blank rate. About
BIO101. Automated DNA sequencing was performed on ABI PRIS- 50 lL of diluted enzyme was added to initiate the reaction. The
MTM 310 Genetic Analyzer from Applied Biosystems Inc. (Foster enzyme was preincubated with or without the inhibitor in the buffer
City, CA, USA). Synthetic oligonucleotides for SELEX were synthe- for 15 min at 25 C.
sized using a Beckman Oligo 1000 M oligonucleotide synthesizer.
PCR reactions were carried out using a Perkin Elmer Certus Thermal The protein concentrations were determined by the method of
Cycler. Pfu polymerase was purchased from Stratagene (La Jolla, Lowry (11) using bovine serum albumin as a standard. This method
CA, USA). The cell suspensions were sonicated using a Heat System was used throughout for all protein determinations.
Ultrasonics, Inc. model W-375 sonicator (Ultrasonics, Inc., Farming-
dale, NY, USA). Phenol:Chloroform:Isoamyl alcohol (PCI) (25:24:1) and
electrophoresis grade agarose were obtained from Amresco (Solon, SELEX
OH, USA). Porcine carboxypeptidase A and hippuryl-L-phenylalanine One 61 base ssDNA was synthesized containing 30 bases of ran-
were purchased from Sigma Chemical Company (St. Louis, MO, domized sequence (N)30 between two primer regions encompassing
USA). PCR 20 bp low ladder, ethidium bromide, dimethysulfoxide SacI and NdeI restriction endonuclease recognition sites respec-
(DMSO), acrylamide, bisacrylamide, benzylpenicillin, cephalosporin C tively (underlined):
(potassium salt), ampicillin, cesium chloride, EDTA, ethanol, glucose,
sodium hydroxide, potassium hydroxide, rubidium chloride, urea,
MOPS, Tris, ZnSO4, and various other inorganic salts and organic sol- 5¢-CGCGAGCTCCGCGCG(N)30CGCGCGCATATGGCGC-3¢
vents of reagent grade or better were obtained from Sigma Chemi- Sac I Nde I
cal Co. Difco brand (Difco, Franklin Lakes, NJ, USA) bacto-agar, This template DNA was amplified by PCR with the corresponding primers, a
casamino acids, and yeast extract used to make all media and plates 5¢ Primer (16 residues) possessing a NdeI site:
were obtained through Fisher Scientific (Pittsburg, PA, USA). 5¢-GCGCCATATGCGCGCG-3¢
and a 3¢ Primer (15 residues) possessing a SacI site:
Enzyme purification and enzyme assays
5¢-CGCGAGCTCCGCGCG-3¢
The metallo-b-lactamase from B. cereus 5 ⁄ B ⁄ 6 was produced from
Escherichia coli MZ1 harboring the pRE2 ⁄ bla plasmid and purified
according to the procedures described previously (7). The purity was The purified stock enzyme in 150 mM ammonium sulfate, 10 mM
ascertained by specific activity, native and SDS PAGE, and DE-MALDI- sodium citrate (pH = 7.0), 1 mM ZnSO4, and 30% (v ⁄ v) glycerol was
TOF. For metallo-b-lactamase activity assays, the activities using heated for 30 min at 60 C to denature any possible trace contami-
cephalosporin C as substrate were determined as previously reported nant proteins. This enzyme is stable at 60 C. The enzyme was centri-
(8). The activity assay procedure was essentially the same as the fuged and the supernatant was collected. The enzyme was diluted
spectrophotometric method of Davies et al. (9) in which the substrate with dilution buffer [20 mM TA (tris–acetate) and 1 mM ZnSO4,
absorbance at 260 nm is continuously monitored during hydrolysis. pH = 7.0]. The synthesized library of 61-mer ssDNA described above
One unit of activity was defined as the amount of enzyme required to was used for SELEX selection. The ssDNA was incubated with the
catalyze the hydrolysis of 1 lmol of substrate in 1 min at 30 C. All enzyme at 30 C for 15 min in TA buffer with an appropriate concen-
assays were carried out near Vmax using 4.3 mM cephalosporin C tration of NaCl; the total reaction volume was 20 lL. The amounts of
(approximately 10 Km) dissolved in 50 mM MOPS ⁄ 1 mM ZnSO4, NaCl in the incubated buffer were adjusted from)50 mM. After
pH = 7.0. The assays were performed in a 0.1 cm pathlength quartz 15 min, glycerol was added to each sample to give 10% (v ⁄ v) glycerol
cuvette and the total reaction volume was maintained at 250 mL. as a final concentration. Samples were run in the 6% (w ⁄ v) polyacryl-
amide gel at 200 V for 25–30 min. The SELEX electrophoretic mobility
For b-lactamase I activity assays, we used a modification of the shift assay used 6% (w ⁄ v) polyacrylamide gels (29:1 monoacryla-
method described by Davies et al. (9). The enzyme sample was mide:bis-acrylamide) in 20 mM TA buffer (pH = 7.0), polymerized with
incubated with 20 mM EDTA (pH = 7.0) for 15 min at room tem- 0.07% (w ⁄ v) ammonium persulfate and 0.028% (v ⁄ v) TEMED.
perature prior to the assay. The enzymatic hydrolysis of 1.1 mM
benzylpenicillin in 10 mM sodium citrate (pH = 7.0) and 1 mM The enzyme:ssDNA complexes were separated from unbound DNA
EDTA was continuously monitored at 231 nm at 30 C. One unit on the 6% (w ⁄ v) polyacrylamide gels. The resulting gels were
of b-lactamase activity was defined as the amount of enzyme soaked in the incubation buffer with ethidium bromide for 10 min
required to hydrolyze 1 lmole of substrate in 1 min at 30 C and and destained in distilled, deionized H2O. The enzyme:ssDNA com-
pH = 7.0. plexes were visualized by UV illumination using TM-36 Chromato-
UVE transilluminator from UVP Inc. The bands containing bound
The assay of porcine carboxypeptidase A is based on the method DNA were excised from the gel and crushed by a disposable pip-
of Folk and Schirmer (10). The rate of hydrolysis of hippuryl-L-phen- ette tip in the PCR tube. The ssDNA was amplified with 2.5 U of
ylalanine is determined by monitoring the increase in absorbance at the pfu polymerase. The reaction mixture, including 200 ng of 5¢
254 nm (25 C, pH = 7.5). The enzyme was dissolved in 10% lith- primer (16 residues) and 100 ng of 3¢primer (15 residues), was sub-
ium chloride to a concentration of 1–3 U ⁄ mL. Hippuryl-L-phenylala- jected to 30 cycles of 45 seconds at 94 C, 45 seconds at 55 C,
nine (1 mM) was dissolved in 0.05 M Tris–HCl, pH = 7.5 with 0.5 M and 6 seconds at 72 C. This was followed by 10 min at 72 C to

344 Chem Biol Drug Des 2009; 74: 343–348

 Metallo-b-lactamase Inhibitors

allow all annealed primers to finish extending. The optimal 10· and sequenced as described above. Twenty-one cycles of SELEX
buffer for PCR was 100 mM Tris–HCl (pH = 8.8), 35 mM MgCl2, and used to obtain the single sequence; from the seventeenth cycle to
250 mM KCl. The final concentration of dNTP was 2 mM. The total the twenty-first cycle, the time-period of the incubation of ssDNA
reaction volume was 100 lL. and enzyme was 2.5 h. This 30-residue ssDNA synthesized as
described above and subsequently purified by electrophoresis on a
The PCR products were purified from 12% (w ⁄ v) polyacrylamide gel 12% (w ⁄ v) polyacrylamide gel for all further experiments. The
(29:1 monoacrylamide:bis-acrylamide). After cutting out the segment HPLC-purified 10-residue ssDNA was purchased from the Oligo
of the gel using a sharp scalpel or razor blade, the slice was trans- Factory (Holliston, MA, USA).
ferred to a microcentrifuge tube. The slice was crushed using a dis-
posable pipette tip. The slice was weighed to determine its volume
and 1–2 volumes of elution buffer (0.5 M ammonium acetate, 1 mM Bacterial growth experiments
EDTA (pH = 8.0), and 0.1% (w ⁄ v) SDS) was added. The tube was For the microbiological experiments, B. cereus 5 ⁄ B ⁄ 6 was grown at
incubated at 45 C on a rotary platform for 2.5–3 h. After centrifug- 30 C in 'S' broth (8) while E. coli transformed with the expression
ing the tube at 12 000 g for 1 min, the supernatant was transferred vector plasmid PRE2 ⁄ bla was grown at the enzyme induction tem-
to a fresh microcentrifuge tube. To avoid any fragments of poly- perature (42 C) in LB medium.
acrylamide, a plastic column containing glass wool was used to
centrifuge the supernatant. A one-half volume aliquot of elution
buffer was added to the remaining pellet, mixed and recentrifuged. Results and Discussion
The supernatant and gel fragments were poured into the plastic
column and spun for 15 second; 2–2.5 volumes of 100% ethanol The SELEX method is used to screen for nucleic acid aptamers.
was added to the sample from this column and placed at )20 C Aptamers are oligonucleotide or peptide molecules that have the
for 1 h and at )80 C for 10–15 min. The tube was spun for ability to bind to specific target molecules in a molecular recogni-
10–15 min. This ethanol precipitation step helps in the removal of tion fashion that rivals antibodies (15). While it may at first seem
ethidium bromide. The supernatant was discarded. The pellet was surprising that a bacterial enzyme that presumably has no contact
washed with 70% ethanol and was dried. with nucleic acids in cells during its normal function can be so effi-
ciently and specifically inhibited by nucleic acid aptamers, we have
To confirm that the PCR product was the same size as the original previously observed a strong association of the crude enzyme with
ssDNA containing 30 random bases the two samples were com- nucleic acids during early stages of the metallo-b-lactamase purifi-
pared on 12% (w ⁄ v) polyacrylamide (29:1 monoacrylamide:bis-acryl- cation scheme (7).
amide) and 8 M urea gel in TBE buffer (12).
Figure 1 shows two photographs of the identical polyacrylamide elec-
The plasmid pRE2 ⁄ bla was digested with restriction endonucleases trophoresis gel containing a mixture of single-stranded DNA bound to
NdeI and SacI (13). All these double-digestion mixtures were elec- the purified metallo-b-lactamase during the first cycle of SELEX. On
trophoretically separated on 1.0% (w ⁄ v) agarose gel in TBE buffer the left, the gel was soaked in an ethidium bromide solution and illu-
at 60 V in the absence of ethidium bromide for 3 h. The linear minated to reveal the bound single-stranded DNA. On the right, the
pRE2 vector and the metallo-b-lactamase gene fragment were then same gel was stained with a protein specific stain, coomaisse blue.
located by staining the gels in 5 mg ⁄ mL ethidium bromide solution Clearly, both the metallo-b-lactamase and DNA are present in this
and visualized under UV. The restricted linear pRE2 was then band. Under the conditions of this experiment, the metallo-b-lactam-
excised from the gels, and extracted by the Gene Clean Kit. ase is cationic and the anode is placed at the bottom of the gel. Nor-
mally, the enzyme would not even enter the gel, but because it is
The ssDNA was amplified by PCR to make double stranded DNA tightly bound to the anionic DNA present, the complex migrates
(dsDNA). After ethanol precipitation, the fixed regions were toward the gel bottom. Unbound DNA is not visible here because it
digested with restriction endonuclease NdeI and SacI. The has run off the bottom of the gel.
digested fragment was loaded on 12% (w ⁄ v) polyacrylamide gels
(29:1 monoacrylamide:bisacrylamide) and was then purified by the After 21 cycles of SELEX, the final 30 residue DNA sequence that
crush and soak method. Ligation of the fragments with the linear was obtained was:
pRE2 vector was accomplished with T4 DNA ligase at 4 C over-
night or at room temperature for 3 h. For each ligation, 100 ng of 5¢-d(AACCAAACTTGGATCGGTGCACATGTCGAA)-3¢
linearized pRE2 vector, 1.11 ng of fragment and 3 U of T4 DNA
ligase were mixed together in ligation buffer in a total volume of Figure 2 shows the steady state enzyme kinetic data of the inhibi-
10 mL. After incubation, the mixture was used to transform E. coli tion of the B. cereus 5 ⁄ B ⁄ 6 metallo-b-lactamase by various concen-
strain TAP 56 competent cell prepared by the Hanahan method trations of the 30 residue ssDNA. The Lineweaver–Burk plot shown
(14). Transformed cells were incubated at 30 C for 2–5 h and in Figure 2 is typical for rapid, reversible, non-competitive inhibition.
were then put into LB medium that contained 1.0% (w ⁄ v) casami- Indeed, a similar kinetic pattern was obtained for inhibition of the
no acids, 0.5% (w ⁄ v) yeast extract, 0.5% (w ⁄ v) sodium chloride metallo-b-lactamase by the divalent metal chelating compound
(adjusted to pH = 7.0 with NaOH) and 50 mg ⁄ mL ampicillin. The EDTA (data not shown). Slope and intercept replots of the primary
culture was incubated at 30 C overnight. The subcloned plasmid data revealed the dissociation constants for the enzyme-inhibitor
DNA was prepared by the boiling minipreparation method (12), complex (Ki) and the Michaelis complex (Ki¢) respectively and their

Chem Biol Drug Des 2009; 74: 343–348 345

Kim et al.

Figure 1: The evidence for a complex of the Bacillus cereus 5 ⁄ B ⁄ 6 metallo-b-lactamase and ssDNA. On the left, the gel was stained by
ethidium bromide. On the right, the gel was stained by Coomassie Brilliant Blue R250. About 20 lM enzyme and 1.5 lM ssDNA were used to
make the complex. The buffer used for incubation was 20 mM TA (pH = 7.0) and 1 mM ZnSO4.

values are listed in Table 1. Clearly, with all the inhibition constants In fact, all of the inhibition was associated with this 10 residue
in the nanomolar range, inhibition of the metallo-b-lactamase by sequence. When we synthesized the nine residue sequence upstream
the 30 residue DNA is quite effective. By comparison, the Ki value and the 10 residue sequence downstream of the sequence in Figure 3,
for EDTA was 3 lM. neither of these ssDNA's had any effect at all on the metallo-b-lac-
tamase activity (data not shown). Further analysis of the kinetic inhibi-
An analysis of the 30 residue ssDNA sequence for possible segments tion pattern for the 10 residue DNA yield the constants listed in
of expected secondary structure by the M-fold program (16) revealed Table 1; inhibition by the 10 residue DNA is very similar to that by the
the predicted 10 residue segment depicted diagrammatically in Fig- 30 residue DNA. The rest of our experiments were performed using
ure 3. When this 10 residue DNA was synthesized, this 10 residue the 10 residue ssDNA shown in Figure 3.
stretch was found to also inhibit the metallo-b-lactamase in a non-
competitive fashion as was true for the 30 residue sequence (Figure 2). As inhibition of metalloenzymes by metal chelators is expected to
be non-competitive, we tested the 10 residue DNA as a reversible
140 inhibitor of a different, typical zinc-dependent enzyme, porcine car-
boxypeptidase A. Even at 25 times the concentration of the Ki for
120
Table 1: Steady state kinetic parameters for inhibition of Bacil-
lus cereus 5 ⁄ B ⁄ 6 metallo-b-lactamase by single stranded DNA
100
Aptamer Enzyme form Ki (nM ) Ki¢ (nM)
80 30 residue ssDNA Zinc 0.92 11
10 residue ssDNA Zinc 0.31 1.5
60 10 residue ssDNA Cobalt 2.5 · 103 2.9 · 105

40

20

0
–3 –2 –1 0 1 2 3 4 5 6
1/[cephalosporin C] (mM–1)

Figure 2: Lineweaver–Burk plot of inhibition of Bacillus cereus

5 ⁄ B ⁄ 6 metallo-b-lactamase by the 30 residue single stranded DNA.
For the filled circles: [DNA] = 0 nM; open circle: [DNA] = 1 nM; filled Figure 3: Diagrammatic representation of the predicted second-
square: [DNA] = 2 nM; open square: [DNA] = 3 nM. Steady ary structure for the 10 residue single stranded DNA sequence
state kinetic assays were performed as described in Methods and derived from the longer 30 residue sequence. The dots are intended
Materials. to represent hydrogen-bonded base pairs.

346 Chem Biol Drug Des 2009; 74: 343–348

 Metallo-b-lactamase Inhibitors

1.6
1.4
1.2

Absorbance
1
Figure 4: Electronic absorbance
0.8
spectra of 0.7 mM Bacillus cereus
5 ⁄ B ⁄ 6 metallo-b-lactamase in 0.6
50 mM MOPS ⁄ 1 mM Co (II) 0.4
sulfate, pH = 7.0 in the absence 0.2
(dotted line) and presence (solid 0
line) of 1.6 mM ssDNA. The path 300 350 400 450 500 550 600 650 700 750
length was 1 cm. Wavelength (nm)

the metallo-b-lactamase, the DNA had no effect on the enzymatic A 1.2

activity of porcine carboxypeptidase. Hence, the DNA does not
indiscriminately chelate the active site zinc but rather is highly spe- 1

Absorbance at 600 nm
cific for the B. cereus 5 ⁄ B ⁄ 6 metallo-b-lactamase to which it obvi-
0.8
ously binds tightly. Furthermore, a concentration of the 10 residue
DNA 25 times the Ki for the metallo-b-lactamase had no effect on 0.6
the enzymatic activity of another, metal-independent b-lactamase.
Clearly, even though benzylpenicillin is a very good substrate for 0.4
both enzymes (hence their substrate – binding domains have some
similarities) the DNA has no effect whatsoever on the activity of 0.2
the B. cereus 569 ⁄ H ⁄ 9 metal-independent b-lactamase I. Clearly, 0
the DNA is not competing for the active site. Still further support 0 5 10 15 20
Time (h)
for the idea that the inhibition involves the active site metal(s)
comes from the fact that when the inhibition is carried in excess
B 5
zinc (1 mM) the inhibitor is sixfold less effective than in the
absence of exogenous zinc (data not shown). 4.5
4
Absorbance at 600 nm

We conclude that not only does the DNA bind the metallo-b-lac- 3.5
tamase tightly, but also that it does so by interfering with the coor- 3
dination of one or more active site zinc ions by binding closely to 2.5
the active site metal(s) and ⁄ or perhaps by coordinating one or more
2
of the active site metals. Either or both of these ideas are consis-
tent with non-competitive inhibition patterns observed in the pres- 1.5
ence of the 10 or 30 residue DNA's; neither inhibitor simply 1
competes for the substrate binding site. To completely evaluate the 0.5
possibility of tight-binding inhibition of the enzyme by the DNA, a 0
more extensive analysis of the inhibition of the enzyme by the DNA 0 5 10 15
Time (h)
at various enzyme concentrations will be required.
Figure 5: Panel A. Bacillus cereus growth in liquid cultures.
It is possible to prepare the apoenzyme of the metallo-b-lactamase Three 1 mL cultures of 'S' broth were inoculated with B. cereus
and reconstitute the some of the original enzymatic activity by 5 ⁄ B ⁄ 6 cells and shaken for 20 h at 30 C. The blue circles represent
the addition of cobalt (II) sulfate (18). Figure 4 shows the visible the control with only the B. cereus. The green circles represent the
electronic spectra of the Co (II)-reconstituted metallo-b-lactamase in B. cereus with 1 lM cephalexin (antibiotic). The circles represent
the absence (dotted line) and presence (solid line) of an excess of growth of B. cereus with 1 lM cephalexin and 0.3 mM 10 residue
the 10 residue DNA. The changes in the spectrum indicate a DNA. Panel B. Escherichia coli growth in liquid cultures. Two 1 mL
change in the coordination sphere of the active site metal and are cultures of LB broth were inoculated with E. coli TAP 56 cells trans-
thought to be associated with a perturbation of the charge transfer formed with our metallo-b-lactamase expression vector plasmid and
band from the metal to the only cysteine thiol in the protein. Simi- shaken for 10 h at 42 C. The diamonds represent growth of E. coli
lar changes, of somewhat larger magnitude, are observed in the in LB broth with 5 lM cephalexin and the circles represent growth
presence of an excess of substrates such as cephalosporin C of the E. coli in 5 lM cephalexin and 0.3 mM 10 residue DNA.
(17,18). Thus, we have spectral evidence that the aptamer binds
tightly to the enzyme in a fashion consistent with our kinetic data tuted form of the enzyme and serves to further demonstrate the
at or near the metal binding sites. The Ki and Ki¢ data of Table 1 exquisite specificity of the DNA aptamer in that the DNA is able to
show that the 10 residue DNA can also inhibit the cobalt-reconsti- distinguish between the zinc and cobalt forms of the enzyme even

Chem Biol Drug Des 2009; 74: 343–348 347

Kim et al.

though both enzyme forms possess the same enzymatic activity at References
the same active site.
1. Perez F., Endimiani A., Hujer K.M., Bonomo R.A. (2007) The
Figure 5A, B demonstrates that the combination of the 10 residue continuing challenge of ESBLs. Curr Opin Pharmacol;7:459–
DNA aptamer with a typical b-lactam antibiotic (cephalexin) can 469.
cause cell death over a period of several hours of both Gram-positive 2. Walsh T.R., Toleman M.A., Poirel L., Nordmann P. (2005) Metal-
and Gram-negative antibiotic resistant bacteria. In Figure 5A, the lo-b-lactamases: the quiet before the storm? Clin Microbiol
combination of DNA and antibiotic suppresses Gram-positive Rev;18:306–325.
B. cereus 5 ⁄ B ⁄ 6 growth in liquid culture with only a single dose of 3. Turek C., Gold L. (1990) Systematic evolution of ligands by expo-
DNA over a 20-h period at 30 C. Note that the control cultures with nential enrichment: RNA ligands to bacteriophage T4 DNA poly-
either antibiotic alone or inhibitor alone do not cause growth suppres- merase. Science;249:505–510.
sion. In the case of Gram-negative E. coli TAP 56 in Figure 5B, the 4. Ellington A.D., Szostak J.W. (1990) In vitro selection of RNA
growth was done at 42 C to induce the B. cereus metallo-b-lactam- molecules that bind specific ligands. Nature;346:818–822.
ase in the E. coli cells transformed with pRE2 ⁄ bla. Clearly under the 5. Shaw R.W., Kim S.-K. (2003) Inhibition of bacterial metallo-b-
conditions of the latter experiment, the DNA was easily able to enter lactamase. FASEB J;17:A981.
the intact E. coli cells; the 10 residue DNA has a molar mass of 6. Wozniak S.E., Shaw R.W. (2006) In vivo tests of nucleic-acid-
3 kDA. From separate studies of the growth of B. cereus and E. coli at based metallo-b-lactamase inhibitors in cultures of Bacillus
various concentrations of the 10 residue DNA (data not shown), we cereus 5 ⁄ B ⁄ 6. FASEB J;20:A897.
were able to determine for each bacterium a concentration of the 7. Shaw R.W., Clark S.D., Hilliard N.P., Harman J.G. (1991) Hyper-
DNA that results in 50 per cent cell death (LC50) in their respective expression in Escherichia coli, purification and characterization
liquid cultures. In the presence of 5 lM cephalexin, the LC50 values of the metallo-b-lactamase of Bacillus cereus 5 ⁄ B ⁄ 6. Protein
obtained for B. cereus 5 ⁄ B ⁄ 6 and E. coli TAP56 were 75 and 32 lM Expr Purif;2:151–157.
respectively. Clearly, these LC50 values are higher than the Ki and Ki¢ 8. Myers J.M., Shaw R.W. (1989) Production, purification and spec-
values obtained for purified enzyme. This might be a result of possible tral properties of metal-dependent b-lactamases of Bacillus
exonuclease activity in the bacterial cultures or perhaps the endoge- cereus. Biochim Biophys Acta;995:264–272.
nous zinc ion concentration in the culture media. Putative exonuclease 9. Davies R.B., Abraham E.P., Melling J. (1974) Separation, purifica-
activity might also explain why the E. coli culture of Figure 5B eventu- tion and properties of b-lactamase I and b-lactamase II from
ally begins to grow again. Bacillus cereus 569 ⁄ H ⁄ 9. Biochem J;143:115–127.
10. Folk J.E., Schirmer E.W. (1963) The porcine pancreatic carboxy-
peptidase A system. I. Three forms of the active enzyme. J Biol
Conclusions and Future Directions Chem;238:3884–3894.
11. Lowry O.H., Rosebrough N.J., Farr A.L., Randall R.J. (1951) Pro-
We have shown that ssDNA aptamers can be highly specific and tein measurement with the Folin-phenol reagents. J Biol
effective inhibitors of the B. cereus 5 ⁄ B ⁄ 6 metallo-b-lactamase Chem;193:265–275.
both in Gram-positive and Gram-negative antibiotic resistant bacte- 12. Sambrook J., Fritsch E.F., Maniatis T. (1989) Molecular cloning:
ria. They appear to act by specifically interfering with the active A Laboratory Manual, 2nd edn. New York: Cold Spring Harbor
site metal ions of the enzyme. Laboratory Press; 7.70–7.76.
13. Reddy P., Peterkofsky A., McKenney K. (1989) Hyperexpression
Our continuing investigations of these inhibitors involve testing their and purification of Escherichia coli adenylate cyclase using a
efficacy with other bacterial metallo-b-lactamases. Likewise, we vector designed for expression of lethal gene products. Nucleic
are also investigating bacterial metallo-b-lactamase inhibition with Acids Res;17:10473–10488.
other nucleic acid aptamers including dsDNA and RNA (15); such 14. Hanahan D. (1983) Studies on transformation of Escherichia coli
studies may provide further insight into the details of the mecha- with plasmids. J Mol Biol;66:557–580.
nism of the inhibition. Whether or not such aptamers can be used 15. Pan W., Clawson G.A. (2009) The shorter the better: reducing
as they are as in vivo enzyme inhibitors is also being investigated. fixed primer regions of oligonucleotide libraries for aptamer
Even if the latter goal cannot be realized, nucleic acid aptamers selection. Molecules;14:1353–1369.
provide an intriguing new lead to the inhibition of metallo-b-lacta- 16. Zuker M. (1989) On finding all suboptimal foldings of an RNA
mases that may result in the development of new small molecule molecule. Science;244:48–52.
inhibitors. As a result, we are actively pursuing structural data for 17. Bicknell R., Waley S.G. (1985) Cryoenzymology of Bacillus cereus
the inhibitors and inhibited enzyme complexes using X-ray diffrac- b-lactamase II. Biochemistry;24:6876–6887.
tion, EXAFS, EPR, and NMR. 18. Bicknell R., Schaffer A., Waley S.G., Auld D.S. (1986) Changes
in the coordination geometry of the active-site metal during
catalysis of benzylpenicillin hydrolysis by Bacillus cereus b-lac-
Acknowledgments tamase II. Biochemistry;25:7208–7215.

The authors thank the Robert A. Welch Foundation (D-1306 to R. W.

S.) and the TTU Office of Technology Commercialization for support.

348 Chem Biol Drug Des 2009; 74: 343–348