V.R. Porter et al. / Biochemical and Biophysical Research Communications 403 (2010) 85–90
87
80, and 200
amikacin, paromomycin, neomycin B, ribostamycin, and tobramy-
cin; 0, 10, 25, 75, 125, 200, and 300 M for gentamicin; 0, 25, 75,
125, 200, 300, and 400 M for sisomicin). The time points utilized
lM for kanamycin A; 0, 1, 5, 10, 25, and 100 lM for
l
l
for kanamycin A were 0, 2, 4, 6, 8, and 10 min, whereas for all other
aminoglycosides, time points of 0, 1, 2, 3, 4, and 5 min were uti-
lized. For tobramycin, five additional time points at intervals of
20 or 30 s were used at lower concentrations of the aminoglyco-
side. The experiments were carried out in triplicate for each sub-
strate concentration with a negative control (no ANT(40)).
For the determination of Km and kcat for the NTP and dNTP
cosubstrates (Supplementary Figs. S5–6), reactions were per-
formed as described above with at least a 2-fold excess of kanamy-
cin A (5 mM for 2 and 1 mM NTP and dNTP trials and 1 mM for
every other NTP and dNTP concentrations) and varying concentra-
Fig. 2. TLC showing the formation of 40-AMP-kanamycin
A
(lane 2), 40-GMP-
kanamycin A (lane 3), 40-IMP-kanamycin A (lane 4), 40-CMP-kanamycin A (lane 5),
40-UMP-kanamycin A (lane 6), 40-dTMP-kanamycin A (lane 7), 40-AMP-amikacin
(lane 9), 40-GMP-amikacin (lane 10), 40-IMP-amikacin (lane 11), 40-CMP-amikacin
(lane 12), 40-UMP-amikacin (lane 13), and 40-dTMP-amikacin (lane 14) from the
parent drugs kanamycin A (lane 1) and amikacin (lane 8) by action of ANT(40).
material was partially consumed, but no new spots were observed
on the TLC when reacted with UTP or TTP.
The masses of novel aminoglycoside-NMPs were determined by
HRMS. To prepare the samples for MS analyses, methanol (25 lL)
was added to the reaction mixtures described above. The precipi-
tated protein was removed by centrifugation (14,000 rpm, rt,
10 min), and the supernatant was freeze-dried overnight. The ami-
noglycoside-NMP derivatives were dissolved in a minimum volume
of H2O prior to HRMS analysis. Fig. 3 displays representative exam-
ples of HRMS for some of the kanamycin A derivatives generated.
tions of NTPs and dNTPs (0, 25, 100, 200, 500
UTP and TTP; 0, 100, 200, 500 M, as well as 1 and 2 mM for
ATP; 0, 25, 50, 75, 100, 200, 500 M, and 1 mM for GTP; 0, 25,
50, 100, 150, 200, 500 M, as well as 1 and 2 mM for ITP; 0, 25,
50, 100, 150, 200, and 500 M for CTP; 0, 50, 100, 350, 500,
750 M, and 1.5 mM for dATP; 0, 50, 100, 400, 500, and 750
for dCTP; 0, 100, 200, 350, 500, and 750 M for dGTP; 0, 100,
200, 300, 350, 500, and 750 M for dUTP). The time points utilized
lM, and 1 mM for
l
l
l
l
l
lM
l
l
for all NTPs and dNTPs were 0, 1, 2, 3, 4, and 5 min. The experi-
ments were carried out from 2 to 4 times for each cosubstrate con-
centration with a negative control (no ANT(40)). The Km and kcat
values for all NTPs and dNTPs (except CTP) were determined by a
Michaelis–Menten curve fit of the data using Kaleidagraph 3.6
curve fitting software. The Km and kcat values for CTP were deter-
mined by Lineweaver–Burk analysis, as high enough concentra-
tions to obtain a Michaelis–Menten curve were not attainable
due to high background signal.
2.7. Antibacterial activity screen by disk diffusion and bioTLC assays
Thereaction conditionsdescribedfor TLCvisualization were used
for antibacterial assays. For the disk diffusion assays (Supplemen-
tary Fig. S7), aliquots of the total reaction mixtures (10 lL) were
loaded onto sterile disks placed on LB agar plates. Overlays of soft
LB agar (0.75 g of agar per 100 mL of LB broth) (6 mL) containing
Bacillus subtilis (60 lL of an overnight culture) were poured onto
the plates. After >10 h of incubation at 30 °C, plates were examined
for clear zones of inhibition. All novel NMP-aminoglycosides were
found to be inactive against B. subtilis. However, for neomycin B re-
acted with GTP and ATP, very small zones of inhibition could be ob-
served. To test if the activity came from unreacted parent neomycin
B or from the novel GMP-neomycin B and AMP-neomycin B, bioTLC
assays were performed as previously described [17]. Briefly, por-
tions of the reaction mixtures (10 lL) were loaded onto a TLC plate
and allowed to migrate using 3:2/MeOH:NH4OH. The plate was
air-dried for 2–3 h prior to being overlayed with soft LB agar
(12 mL) infused with B. subtilis (120 lL of an overnight culture).
The B. subtilis overlay was grown at 30 °C until clear zones of inhib-
ited growth were observed (overnight). The Rf values of the starting
materialandproductsonthestainedTLCscorrespondedtothe Rf val-
ues of the zones of inhibition on the overlay. The bioTLCs confirmed
thatthehalos of inhibitionseen inthe diskdiffusion assaysfor nucle-
otidylated neomycin B were solely caused by unreacted parent drug.
2.5. ATP-[32P]PPi exchange assays
To verify if there is a fast reversible exchange in the active site of
the enzyme that would result in the formation of [32P]ATP, ATP-
[
32P]PPi reactions (100
l
L) containing Tris–HCl (pH 7.5) (75 mM),
MgCl2 (10 mM), KCl (50 mM), TCEP (pH 7.0) (5 mM), ATP (5 mM),
aminoglycoside (100
M), and [32P]PPi (1 mM, 84.12 Ci/mM) were
run at 25 °C. The reactions were started by addition of ANT(40)
(0.1 M) and incubated for up to 30 min prior to quenching with
charcoal suspension (500 L) [1.6% (w/v) activated charcoal, 4.5%
l
l
l
(w/v) Na4P2O7, and 3.5% (v/v) perchloric acid in H2O]. The charcoal
was pelleted by centrifugation before being washed twice with the
wash solution (500
chloric acid in H2O], resuspended in H2O (500
l
L) [4.5% (w/v) Na4P2O7 and 3.5% (v/v) per-
L), and counted
l
by liquid scintillation. The experiments were done in duplicate
for each substrate concentration with a negative control (no
ANT(40)). No exchange was observed by this assay for all aminogly-
cosides tested, which is consistent with the previously reported
dramatically slower kinetics in the reverse direction for this en-
zyme when tested with AMP-kanamycin A as starting material [15].
3. Results and discussion
3.1. Heterologous expression and purification of ANT(40)
The 33.3-kDa ANT(40) from S. aureus was overexpressed in E. coli
as a NHis6-tagged protein in order to establish its aminoglycoside
substrate and nucleotide triphosphate cosubstrate profile in vitro.
Purification to homogeneity of the soluble protein was achieved
by Ni(II)–NTA affinity chromatography to yield about 7 mg of
ANT(40) per liter of culture.
2.6. TLC visualization and HRMS analysis of ANT(40) reactions
Reaction mixtures (25
(50 mM), MgCl2 (10 mM), DTT (1 mM), KCl (50 mM), aminoglyco-
side (3.2 mM), NTP (6 mM), and ANT(40) (14.4
M) were incubated
lL) containing Tris–HCl (pH 7.5)
l
at rt overnight and loaded onto a TLC plate (Silica gel F254 250 mm
thickness). Negative control reaction mixtures contained all re-
agents, but no ANT(40) enzyme. Visualization was performed using
a cerium-molybdate stain. The eluent systems utilized and Rf val-
ues observed are reported in Table S1. Representative examples
of TLCs are depicted in Fig. 2. In the case of gentamicin, the starting
3.2. Substrate and cosubstrate specificity of ANT(40)
To determine the substrate specificity of ANT(40), a variety of
commercially available aminoglycosides possessing or lacking a
40-hydroxyl functionality were utilized. To monitor product