Biomimetic Design Results in a Potent Allosteric Inhibitor of Dihydrodipicolinate Synthase from Campylobacter jejuni

Article abstract of DOI:10.1021/jacs.5b12695

Dihydrodipicolinate synthase (DHDPS), an enzyme required for bacterial peptidoglycan biosynthesis, catalyzes the condensation of pyruvate and β-aspartate semialdehyde (ASA) to form a cyclic product which dehydrates to form dihydrodipicolinate. DHDPS has, for several years, been considered a putative target for novel antibiotics. We have designed the first potent inhibitor of this enzyme by mimicking its natural allosteric regulation by lysine, and obtained a crystal structure of the protein-inhibitor complex at 2.2 ? resolution. This novel inhibitor, which we named 'bislysine', resembles two lysine molecules linked by an ethylene bridge between the α-carbon atoms. Bislysine is a mixed partial inhibitor with respect to the first substrate, pyruvate, and a noncompetitive partial inhibitor with respect to ASA, and binds to all forms of the enzyme with a K_i near 200 nM, more than 300 times more tightly than lysine. Hill plots show that the inhibition is cooperative, indicating that the allosteric sites are not independent despite being located on opposite sides of the protein tetramer, separated by approximately 50 ?. A mutant enzyme resistant to lysine inhibition, Y110F, is strongly inhibited by this novel inhibitor, suggesting this may be a promising strategy for antibiotic development.

Full text of DOI:10.1021/jacs.5b12695

Article
pubs.acs.org/JACS
Biomimetic Design Results in a Potent Allosteric Inhibitor of
Dihydrodipicolinate Synthase from Campylobacter jejuni
Yulia V. Skovpen, Cuylar J. T. Conly, David A. R. Sanders,* and David R. J. Palmer*
Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, Saskatchewan, Canada S7N 5C9
*
S Supporting Information
ABSTRACT: Dihydrodipicolinate synthase (DHDPS), an en-
zyme required for bacterial peptidoglycan biosynthesis, catalyzes
the condensation of pyruvate and β-aspartate semialdehyde (ASA)
to form a cyclic product which dehydrates to form dihydrodipi-
colinate. DHDPS has, for several years, been considered a putative
target for novel antibiotics. We have designed the first potent
inhibitor of this enzyme by mimicking its natural allosteric
regulation by lysine, and obtained a crystal structure of the protein-
inhibitor complex at 2.2 Å resolution. This novel inhibitor, which
we named “bislysine”, resembles two lysine molecules linked by an
ethylene bridge between the α-carbon atoms. Bislysine is a mixed
partial inhibitor with respect to the first substrate, pyruvate, and a
noncompetitive partial inhibitor with respect to ASA, and binds to
all forms of the enzyme with a K near 200 nM, more than 300 times more tightly than lysine. Hill plots show that the inhibition
i
is cooperative, indicating that the allosteric sites are not independent despite being located on opposite sides of the protein
tetramer, separated by approximately 50 Å. A mutant enzyme resistant to lysine inhibition, Y110F, is strongly inhibited by this
novel inhibitor, suggesting this may be a promising strategy for antibiotic development.
INTRODUCTION
■
The need for novel antibiotics is well-documented; in
particular, there has been a striking lack of new drugs for
Gram-negative bacteria in the last three decades. Dihydrodi-
1
picolinate synthase (DHDPS) is considered a target for new
antibiotics due to its essential role in the diaminopimelate
(dap) pathway responsible for biosynthesis of meso-diaminopi-
melate and L-lysine, required components of the cell wall
2
,3
peptidoglycan. The chemical and kinetic mechanisms of
DHDPS are well-described: DHDPS is a Schiff-base-dependent
aldolase catalyzing the conversion of pyruvate and (S)-
aspartate-β-semialdehyde (ASA) to (4S)-hydroxy-2,3,4,5-tetra-
hydro-(2S)-dipicolinate, which spontaneously dehydrates to
Figure 1. Reaction catalyzed by DHDPS. Pyruvate condenses with
Lys166 prior to binding by ASA and subsequent aldol reaction.
4
dihydrodipicolinate (Figure 1). The reaction follows classical
ping-pong kinetics, with pyruvate required to bind and form a
Schiff base with an active-site lysine before ASA binds and
4
,5
reacts. Despite this knowledge, inhibitors designed to target
a mixed partial inhibitor with respect to ASA. The partial
inhibition means that ∼10% of the activity remains at saturating
concentrations of lysine. The inhibition is also highly
cooperative.
Structural studies help explain the observed cooperativity.
DHDPS from Gram-negative bacteria (and many Gram-
positives) are homotetrameric, and can be described as a
loose dimer of tight dimers. At the tight dimer interface, an
allosteric cavity binds two lysine molecules with their α-carbon
6
−9
the active site of DHDPS have not yielded promising results,
and it has not been demonstrated that this enzyme can be
inhibited effectively. Here we report that potent inhibition is
possible, by mimicking the natural regulation mechanism.
DHDPSs from Gram-negative bacteria are regulated by the
end-product of the dap pathway, L-lysine, by allosteric binding
at low concentrations of inhibitor. The DHDPS from
Campylobacter jejuni (CjDHDPS), for example, is inhibited
5
with an apparent IC of 65 μM. Complete analysis shows the
5
0
inhibition behavior to be complex: lysine is an uncompetitive
Received: December 4, 2015
partial inhibitor with respect to the first substrate, pyruvate, and
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
atoms ∼4 Å apart. Binding of one lysine molecule under-
standably affects binding of the adjacent molecule, and this is
the same HPLC method as for protected precursor 9. Each
enantiomer was tested as an inhibitor of DHDPS. The active
enantiomer was presumed to be the (R,R)-isomer, later confirmed
by protein crystallography.
10,11
supported by isothermal titration calorimetry.
Notably, Hill
coefficients for inhibition of CjDHDPS by lysine exceed 2,
suggesting that binding to one allosteric site affects the
Enzymology. Expression and purification of DHDPS and
dihydrodipicolinate reductase (DHDPR). Expression and purification
5
5
antipodal allosteric site. Indeed, inhibition is thought to result
of DHDPS and DHDPR from C. jejuni have been described. The
in part from altered protein motions throughout the tetrameric
obtained proteins were concentrated to 0.37 mg/mL for DHDPS and
1.29 mg/mL for DHDPR.
1
2−14
structure.
10
We hypothesized that design of a molecule that would
mimic two lysine molecules would be the most effective route
to a potent inhibitor. By making a single inhibitor molecule that
could occupy the allosteric site, a significant entropic advantage
can be gained by eliminating the second binding event.
Examination of crystal structures suggested that connecting the
α-carbon atoms with a two-carbon linker would result in the
closest biomimetic structure. Here we describe the synthesis of
that molecule, which we have named “bislysine” (Figure 2), the
Enzyme assays and inhibition studies. The initial velocity of the
DHDPS enzymatic reaction was determined in a coupled assay by
monitoring the decrease in absorbance at 340 nm due to oxidation of
−1
−1
5
NADH (ε340 = 6220 M cm ) as described previously. All kinetic
measurements were made in 100 mM HEPES buffer at pH 8.0. A one-
mL assay contained 0.16 mM NADH, 0.37 μg of DHDPS, 7.15 μg
DHDPR and varying concentrations of ASA (0.066−2.6 mM) and
pyruvate (0.20−3.7 mM). A 0.0084 mM solution of (R,R)-bislysine
(tetrahydrobromide salt) was used in inhibition experiments. All
kinetic studies were carried out at 25 °C and normal atmospheric
pressure.
Slow-binding kinetic determination. To study the time depend-
ence of inhibition of DHDPS by (R,R)-bislysine, experiments were
performed by triggering the reaction by addition of DHDPS. Each
1
5,16
progress curve was fit to eq 1,
vz − v
s
obst
[
P] = vt +
(1 − e−k ) + d
s
kobs
(1)
Figure 2. Binding mode of two L-lysine molecules in the allosteric site
as shown by X-ray crystallography (PDB code 4M19) (left); the basis
for the design (R,R)-bislysine (right).
where v and v are velocities at steady-state and at time zero, k is the
s
z
obs
frequency constant of the exponential phase, t is time, d is
displacement or finite intercept, and [P] is related to the amount of
product produced.
Steady-state kinetic studies. DHDPS was preincubated with the
inhibitor (and other components of the assay) for 1 min and the
reaction was initiated by addition of ASA. The concentration of one
substrate was varied, while the concentration of the other was kept
constant at saturating level (2.6 mM for ASA and 3.7 mM for
pyruvate). All data points represent at least three experiments. The
models for mixed partial (eq 2) and noncompetitive partial inhibition
(eq 3) were used to fit experimental data with SigmaPlot 10.0 (Systat
Software, San Jose, CA),
kinetics of inhibition, and the high-resolution crystal structure
of CjDHDPS bound to bislysine and pyruvate. We have also
performed similar studies with an allosteric site mutant, Y110F,
which we have shown previously to be nearly insensitive to
1
4
lysine inhibition.
EXPERIMENTAL PROCEDURES
■
Preparation of bislysine. Synthesis and characterization of
bislysine and its precursors are described in the Supporting
Information. HPLC separations were performed on an Agilent 1100
series instrument consisting of a quaternary pump, autosampler, diode
array detector (DAD) and evaporative light scattering detector
[
K
S]
β[S][I]h
+
h
v
s
K (αK )
s
i
=
=
h′
[S][I]h
V
[
S]
Ks
[I]
max
1
+
+
+
K (αK )
h′
h
Ki
(2)
(
(
ELSD). A chiral semipreparative Astec CHIROBIOTIC T column
25 cm × 10.0 mm, Supelco Analytical) was used for separation of
s
i
h
[
Ks
S]
β[S][I]
enantiomers. Ammonium acetate buffer (250 mM, pH 4.5) with 70%
methanol was used as a mobile phase for separation of the protected
racemic mixture and deprotected individual stereoisomers. The mobile
phase was degassed by sparging with helium for 20 min. Instrument
parameters were set to the following: flow rate, 2.0 mL/min; injection
volume, 25 μL; DAD, 254 nm; ELSD temperature, 50 °C; ELSD gain,
+
h
v
KsK
i
h
h
Vmax
[
Ks
S]
[I]
[S][I]
1 +
+ Ki
+
KsK
h
h
i
(3)
where Vmax is the maximum velocity, K is the dissociation constant of
S
the ES complex, v is the initial velocity. K is the inhibition constant, I
i
5
. Acquisition and processing of data was done using ChemStation LC
is the inhibitor concentration, S is the substrate concentration, and α
and β are proportionality constants. The cooperativity coefficients h′
and h correspond to (R,R)-bislysine binding to E and E:pyr,
respectively (eq 2). For eq 3, h is a cooperativity coefficient for
Software (Rev.B.04.02; Agilent Technologies Inc.). All experiments
were performed isocratically. The column was initially washed with
methanol and equilibrated with the mobile phase for 30 min. For
quantitative separations, a 7.3 mg/mL solution of a phthalimide-
containing methyl ester precursor (Supporting Information, com-
pound 9) of bislysine was prepared in methanol (Supporting
Information Figure 2). Only the DAD detector was used for runs
involving fraction collection. The middle fraction between peaks of
enantiomers was discarded to avoid cross-contamination. Fractions of
each enantiomer were combined and evaporated under reduced
pressure and moderate heating (40 °C) to remove all ammonium
(
R,R)-bislysine binding to the F and F:ASA forms of the enzyme.
vi − v
i(sat)
log
= log K − h log I
v0 − vi
(4)
Equation 4 is the Hill equation for partial inhibition, where v is the
i
velocity in the presence of the inhibitor, v is the velocity in the
0
absence of the inhibitor, vi(sat) is the reaction velocity at saturating
concentrations of inhibitor, and K is an apparent overall dissociation
constant.
X-ray crystallography of CjDHDPS:bislysine. Diffraction quality
crystals were grown in a solution of 0.2 M sodium acetate, 16% P4000,
0.1 mM Tris pH 8.5 using the hanging drop vapor diffusion method.
Protein and precipitant solutions where mixed in a 1:1 3 μL drop over
1
acetate. The H NMR spectrum of each sample was collected to
confirm that ammonium acetate had been removed. Each enantiomer
obtained was hydrolyzed and purified on a cation exchange column as
described in the Supporting Information. The purity of each
enantiomer was judged by HPLC. The 8.6 mg/mL solutions of
(R,R)- and (S,S)-bislysine were prepared in water and analyzed using
B
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
5
(
00 μL well. DHDPS was preincubated with 20 mM (±)-bisysine
tetrahydrochloride salt) prior to setting up crystallization trials.
Crystals were transferred to a cryoprotectant solution consisting of
0% ethylene glycol, 30% PEG 400, 60% precipitant solution with 10
Scheme 1
1
mM (±)-bisysine (tetrahydrochloride salt) and then flash-cooled in
liquid nitrogen.
Diffraction experiments were conducted using synchrotron
radiation at the Canadian Light Source (CLS) on the 08ID-01
beamlines using a wavelength of 0.9798 Å and equipped with a MAR
CCD 300 area detector. Crystals were kept under a stream of liquid
nitrogen and diffraction was carried out at 100 K. The images were
1
7
integrated and scaled using auto-XDS/XSCALE in-house at the CLS.
Pertinent data collection statistics and refinement parameters are given
in Supporting Information Table 1.
a
Conditions: (a) 1-bromo-4-chlorobutane, LDA, THF/10% HMPA,
−
78 °C, 1 h, 16%. (b) potassium phthalimide, DMF, 90 °C, 2 h, 50%.
+
(
c) 1:1 48% HBr :AcOH, 115 °C, 3 d; AG 50W-X2(H form) eluted
aq
CjDHDPS:bislysine crystals were found to be space group
P2 2 2 . The structure was determined by molecular replacement
using MOLREP from the CCP4 suite using the solvent stripped wild
type structure as the search model. The solution found four monomers
organized as a tetramer in the asymmetric unit. Solvent content
determined from Matthews coefficients was 42.25%. Further
refinements were made using PHENIX, with manual model
correction in COOT. Structure quality was assessed using
MOLPROBITY and validation tools in COOT. The final structure
indicates that 98.5% of residues are in Ramachandran favored
conformations, 1.5% are in acceptable conformations, and 0% in
unacceptable conformations. Coordinates were deposited in the RSCB
Protein Databank (PDB), code 5F1V.
X-ray crystallography of Y110F:bislysine. Diffraction quality
crystals were grown in a solution of 0.28 M sodium acetate, 30%
PEG 4000, 0.1 M TRIS pH 8.5 using the hanging drop vapor diffusion
method. Protein and precipitant solutions where mixed in a 1:1 3 μL
drop over 500 μL well. DHDPS was preincubated with 25 mM
with HBr , 88%.
aq
1
1 1
18
stereoisomers: (R,R)-, (S,S)- and meso-bislysine. Before
deprotection, the racemate was separated from the meso-
compound by conventional flash chromatography. The pair of
enantiomers can be resolved either before or after deprotection
using chiral HPLC (Supporting Information), but the presence
of the phthalimide groups allowed uv detection, which made
separation for preparative purposes much more convenient.
Deprotected compounds could be observed using evaporative
light-scattering detection. Reinjection of purified (R,R)-
bislysine showed that it contained no more than ∼2% of the
1
9
2
0
2
1
(
S,S)-isomer.
Inhibition. The desired (R,R)-enantiomer has the same
relative configuration as a pair of L-lysine molecules. This
enantiomer was an extremely effective inhibitor of CjDHDPS,
with an apparent IC ∼ 300 times lower than that of lysine
(
±)-bisysine (tetrahydrochloride salt) prior to setting up crystal-
50
lization trials. Crystals were transferred to a cryoprotectant solution
consisting of 10% ethylene glycol, 30% PEG 400, 60% precipitant
solution with 20 mM (±)-bisysine (tetrahydrochloride salt) and then
flash-cooled in liquid nitrogen.
(
Figure 3). The other stereoisomers showed inhibition no
greater than that consistent with contamination by traces of the
active (R,R)-isomer.
Diffraction experiments were conducted using synchrotron
radiation at CLS on the 08ID-01 beamlines using a wavelength of
As shown in the Supporting Information, the observed
behavior was consistent with a slow-onset, one-step reversible
inhibition mechanism. The analysis of progress curves
0
.9795 Å and equipped with a MAR CCD 300 area detector. Crystals
were kept under a stream of liquid nitrogen and diffraction was carried
out at 100 K. The images were integrated and scaled using auto-XDS/
XSCALE in-house at the CLS. Pertinent data collection statistics and
refinement parameters are given in Supporting Infomation Table 1.
Y110F:bislysine crystals were found to be space group P2 2 2 .
(
Supporting Information Figure 5) shows inhibition of
DHDPS in a time-dependent manner, where equilibrium
between the free inhibitor and enzyme species establishes
within the first minute. The values of kobs determined using eq
1
1 1
1
showed a linear relationship with inhibitor concentration,
The structure was determined by molecular replacement using
MOLREP from the CCP4 suite using the solvent stripped wild type
structure as the search model. The solution found four monomers
organized as a tetramer in the asymmetric unit. Solvent content
determined from Matthews coefficients was 40.49%. Further refine-
ments were made using PHENIX, with manual model correction in
COOT. Structure quality was assessed using MOLPROBITY and
validation tools in COOT. The final structure indicates that 98.0% of
residues are in Ramachandran favored conformations, 1.66% are in
acceptable conformations, and 0.34% in unacceptable conformations.
The coordinates were deposited as PDB code 5F1U.
which suggested that a slow binding event, rather than fast
binding followed by slow conversion to a tightly bound
complex, occurs. Values of the slope and intercept of this
15
4
−1 −1
graph give values of kon and k of (4.4 ± 1.2) × 10 M
s
off
−3 −1
and (6.2 ± 4.2) × 10 s , respectively. The ratio of these
values gives an apparent K value of 140 nM, close to the
i
observed IC50 value.
Steady-state kinetic analysis showed that the stronger binding
of (R,R)-bislysine relative to lysine affected the kinetic
mechanism of inhibition. Figure 4a shows the model for
allosteric partial inhibition of a two-substrate enzymatic process.
(R,R)-Bislysine is a mixed partial inhibitor with respect to
pyruvate, and a noncompetitive partial inhibitor with respect to
ASA as shown graphically in Figure 4b and 4c, respectively.
These results are consistent with an inhibitor that binds to the
allosteric site regardless of the presence of pyruvate, whereas
our previous results showed that lysine, an uncompetitive
partial inhibitor, does not bind to apoenzyme (or not in a way
RESULTS
■
Preparation of bislysine. Synthesis of the target inhibitor
began by generation of the known Boc-protected methyl ester
of 2,5-diaminoadipate from adipic acid in seven steps
(Supporting Information). The key step in the synthesis was
the alkylation of this product with 1-bromo-4-chlorobutane, as
shown in Scheme 1. This was followed by introduction of the ε-
amino groups using phthalimide, and acidic hydrolysis to form
the final product. The alkylation proceeded in a moderate yield,
but allowed alkylation of both α-carbon positions in a single
step. This nonstereoselective approach yielded all three
5
that affects the reaction). Table 1 shows the inhibition
constants and Hill coefficients calculated using eqs 2, 3, and 4.
Our results show that bislysine binds to all forms of the enzyme
with a K near 200 nM, by far the lowest value ever observed for
i
C
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. L-Lysine and (R,R)-bislysine inhibition of DHDPS activity at 0.16 mM ASA and 3.5 mM pyruvate (left), and an expansion of the same plot
right).
(
Table 1. Hill Coefficients and Inhibition Constants (as
defined in Figure 4) DHDPS Inhibition by Bislysine
a
b
Pyruvate
ASA
hE
1.0 ± 0.2
1.6 ± 0.1
hE:pyr
hF
1.7 ± 0.1
1.7 ± 0.1
hF:ASA
Ki1, nM
Ki2, nM
Ki3, nM
Ki4, nM
240 ± 50
170 ± 70
200 ± 20
200 ± 20
a
b
Fit to the mixed partial inhibition kinetic model (eq 2). Fit to a
noncompetitive partial inhibition kinetic model (eq 3).
been selectively complexed from solution. The tetramer
complexed with two molecules of bislysine is shown in Figure
5
A, and the omit map of one allosteric site is shown in Figure
Figure 4. (R,R)-Bislysine inhibition of DHDPS. (a) General scheme of
R,R)-bislysine inhibition, α (and α′) must be >0, while β (and β′)
must be <1 for inhibition. In a pure noncompetitive partial inhibition
mechanism, α and α ′ = 1; therefore, K 1 = K2, K 3 = K 4, whereas in a
5B. Figure 5C shows the equivalent binding mode of (R,R)-
bislysine and L-lysine in the allosteric site.
(
We previously observed that binding of lysine to CjDHDPS
results in changes to the volume of both the allosteric and
active sites of the enzyme, revealing one of the mechanisms by
which lysine can affect catalysis without making large-scale
changes to the protein structure. Specifically, lysine binding to
CjDHDPS reduces the volume of the allosteric site by 43%
(ignoring the contribution by lysine) and increases the volume
i
i
i
i
mixed partial inhibition mechanism, α and α ′ ≠ 1 (K 1 ≠ K 2, K3 ≠
i
i
i
K4). (b) Lineweaver−Burk plot of data obtained at a constant ASA
i
concentration of 2.6 mM. Concentration of (R,R)-bislysine: (●) 0 μM,
(
○) 0.084 μM, (■) 0.17 μM, (□) 0.42 μM, (◆) 0.84 μM, (◇) 1.7
μM. Solid lines are fit lines, obtained by globally fitting the mixed
partial model (R2 = 0.98) to the data. Residuals are shown in
14
of the active site by 22%. When we repeated this analysis
using our bislysine-bound structure, we saw that these effects
were similar, even enhanced: the allosteric site volume
decreases by 49% upon binding of bislysine, while the active-
site volume increases by 34%.
Supporting Information Figure 9. (c) Lineweaver−Burk plot of data
obtained at a constant pyruvate concentration of 3.7 mM.
Concentration of (R,R)-bislysine: (●) 0 μM, (○) 0.084 μM, (■)
0
.17 μM, (□) 0.42 μM, (◆) 0.84 μM, (◇) 1.7 μM. Solid lines are fit
lines, obtained by globally fitting the noncompetitive partial model (R
2
=
0.99) to the data. Residuals are shown in Supporting Information
Allosteric site mutant Y110F. We recently showed that
Tyr110, a residue within the allosteric site which forms a
hydrogen bond with the carboxylate group of a bound lysine
Figure 10.
14
inhibition of DHDPS. Tellingly, bislysine inhibition is also
cooperative (h = 1.6−1.7), in concert with our prior
observation that the antipodal allosteric sites do not behave
independently. It is clear that occupancy of the allosteric site
has a global effect on the tetrameric protein.
molecule, was critical for effective inhibition by lysine. Tyr110
is adjacent to Tyr111, a residue which is part of an apparent
catalytic triad that is integral to catalysis. Mutation of Tyr110 to
phenylalanine removes the hydroxyl group responsible for this
hydrogen bond, just one of the many polar interactions
between the inhibitor and the protein. The mutant Y110F has a
turnover number reduced by half, but is nearly insensitive to
inhibition by lysine, with an estimated IC > 40 mM. A crystal
X-ray crystallography. We crystallized CjDHDPS in the
presence of a racemic mixture of (R,R)- and (S,S)-bislysine, and
determined the structure to 2.2 Å resolution. As with previous
50
14
structures of CjDHDPS, the electron density difference map
showed clear positive density within the active site consistent
with the Schiff-base formed by Lys166 and pyruvate. The
resulting structure showed that bislysine was unambiguously
present in the allosteric site, and that the (R,R)-enantiomer had
structure of Y110F bound to lysine showed that the loss of this
hydrogen bond resulted in dissipation of several effects of lysine
14
binding observed in the wild-type enzyme. In particular, the
binding of lysine to Y110F reduces the volume of the allosteric
site by only 26%, considerably less than the 43% observed with
D
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
wild-type enzyme. Similarly, while lysine binding to wild-type
increases the volume of the active site by 22%, in the Y110F
mutant the active site shows a slight decrease in volume. These
results suggest that a drug targeting the allosteric site of
DHDPS might be prone to resistance, since a single mutation
can result in a functioning catalyst that is insensitive to
inhibition. However, (R,R)-bislysine is a very potent inhibitor
of Y110F, with an IC50 of 400 nM (Supporting Information),
indicating that (R,R)-bislysine inhibits this mutant nearly as
5
well as it does wild-type, and 10 times more effectively than L-
lysine. Clearly the advantages of the bislysine design, the
occupancy of the entire allosteric site by a single molecule,
outweigh the effects of that hydrogen bond. We crystallized
Y110F in the presence of bislysine, and the resulting structure
closely resembles that of wild-type DHDPS bound to inhibitor.
Changes in cavity volumes in the Y110F-bislysine structure are
consistent with the trend shown for lysine and bislysine
complexation with wild-type CjDHDPS; that is, the volume of
the allosteric site of Y110F decreases by 38% upon bislysine
binding, while the volume of the active site increases by 16%.
DISCUSSION
■
Several reports show that DHDPS is a potential target for drug
22
development; Escherichia coli AT997 (a mutant strain lacking
the DHDPS-encoding gene dapA) can be maintained on
nutrient medium only if the medium is supplemented with
2
,23,24
diaminopimelate,
and dapA was classified as “essential” for
25
26
Bacillus subtilis and Haemophilus influenzae. (This is not
universally the case, however: Schnell et al. showed that
27
Pseudomonas aeruginosa mutants lacking dapA are viable. )
The therapeutic potential has inspired the design of DHDPS
inhibitors, but effective inhibition has proved elusive. Efforts
have concentrated on the active site, resulting in millimolar-
range inhibitors including analogs of substrates pyruvate and
28−30
6,7,29−33
ASA,
analogs of product HTPA or DHDP,
and
mimics of enzyme-bound condensation products of ASA and
7
−9,34
pyruvate.
Targeting of the active site is the common
approach to rational design, since one typically has specific
information about the molecules that fit that site. There are
many examples of successful inhibitors which appear to capture
the binding energy associated with recognition of the transition
state, such as Schramm’s femtomolar inhibitors of methyl-
35
thioadenosine nucleosidase. In the case of DHDPS, however,
the precise binding location and conformation of the natural
regulator lysine is available, providing a natural template for
developing new inhibitors. We sought to establish that the
allosteric site can be targeted effectively, which might open a
new avenue to drug design.
Bislysine proved to be a very effective inhibitor, and the
crystal structure of (R,R)-bislysine bound to the allosteric site
shows how closely it mimics the binding of two (S)-lysine
molecules. The slow-onset character of the inhibition may be
due to the fact that bislysine binds to a site designed for egress
of a pair of smaller molecules. The X-ray structure shows that
amino acid side chains, His56 (and His56′) in particular, would
need to move to allow bislysine to bind (Figure 5b and 5c).
This necessary movement is the most likely explanation, given
that there has been no observation of an “open and closed
conformation” or similar significant conformational change.
We believed that the design of a single molecule to occupy
the entire allosteric site would be an effective inhibition
strategy, a concept often seen in the design of “bisubstrate
analog” inhibitors. Such ligands can make many specific
Figure 5. Crystal structure of CjDHDPS with (R,R)-bislysine bound at
the allosteric site. (A) (R,R)-bislysine (orange) bound at the allosteric
site of wild-type CjDHDPS (PDB 5F1V). Residues of monomers A
and B are shown in green and yellow, respectively. (B) Electron
density of bislysine bound to the allosteric site of DHDPS. The green
mesh is the electron density omit map scaled at 3σ. Monomer A is
green, monomer B (primed residues) is magenta. (C) Overlay of
crystal structures of (R,R)-bislysine (orange inhibitor, green protein)
and L-lysine (cyan inhibitor, white protein, PDB code 4M19) bound at
the allosteric site of CjDHDPS.
E
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
interactions with the protein, resulting in higher affinity and
enhanced selectivity. However, we did not predict that bislysine
would be such an effective inhibitor of the Y110F mutant. The
very weak inhibition of this mutant observed in the presence of
lysine led us to believe that the presence of a tyrosine residue in
this position was crucial to inhibition. Instead, bislysine is nearly
as effective an inhibitor of Y110F as of the wild-type DHDPS.
The result underscores that our view was overly reductionist,
and that multiple mechanisms may be used simultaneously to
effect inhibition.
Synthetic methods, supplementary kinetic methods, X-
ray crystallography statistics, and NMR spectra. (PDF)
AUTHOR INFORMATION
Corresponding Authors
david.sanders@usask.ca
dave.palmer@usask.ca
■
*
*
Notes
The authors declare no competing financial interest.
Allosteric inhibition is often characterized by very noticeable
changes in the shape of the enzyme. In the case of DHDPS,
however, the changes in shape are not immediately obvious. As
we have discussed previously, binding of lysine to DHDPS
results in small domain movements, effectively rotating the
ACKNOWLEDGMENTS
■
This work was supported by NSERC Discovery Grants to
D.R.J.P. and D.A.R.S., and NSERC PGS awards to Y.V.S. and
C.J.T.C. The authors thank Keith Brown, Ken Thoms, and
other staff of the Saskatchewan Structural Sciences Centre. The
protein crystallographic studies described in this paper were
performed at the CLS, which is supported by NSERC, the
National Research Council of Canada, the Canadian Institutes
of Health Research, the Province of Saskatchewan, Western
Economic Diversification Canada, and the University of
Saskatchewan.
14
portion of the (β/α) barrel that includes β-strands 4, 5, and 6
8
relative to the rest of the barrel by just under 4 degrees. This
subtle movement results in an increased volume of the active
site, and a decreased volume of the allosteric site. Any
expectation that the inhibitor has disrupted the active site is not
borne out. The key residues of the active site include K166,
which forms the crucial Schiff base with pyruvate, and the
“
catalytic triad” of Y137, T47, and Y111′. The relative positions
REFERENCES
of these residues in the crystal do change when lysine is bound
to the allosteric site, but none by more than about 0.5 Å. Even
assuming that these changes reflect a difference in the
conformation of the enzyme in solution, it is difficult to
estimate the extent to which such changes affect catalysis.
Lysine’s effects on the TIM barrel also result in small but
significant changes at the weak dimer−dimer interface,
specifically between residues found on helices 6, 7, and 8
■
(
1) Lee, J. H.; Jeong, S. H.; Cha, S.-S.; Lee, S. H. Nat. Rev. Drug
Discov. 2007, 6, DOI: 10.1038/nrd2201-c1031.
2) Bukhari, A. I.; Taylor, A. L. J. Bacteriol. 1971, 105, 844.
3) Dogovski, C.; Gorman, M. A.; Ketaren, N. E.; Praszkier, J.;
Zammit, L. M.; Mertens, H. D.; Bryant, G.; Yang, J.; Griffin, M. D. W.;
Pearce, F. G.; Gerrard, J. A.; Jameson, G. B.; Parker, M. W.; Robins-
Browne, R. M.; Perugini, M. A. Plos One 2013, 8, DOI: 10.1371/
journal.pone.0083419.
(
(
(
see Supporting Information, Figure 14). The bislysine-bound
(4) Blickling, S.; Renner, C.; Laber, B.; Pohlenz, H. D.; Holak, T. A.;
Huber, R. Biochemistry 1997, 36, 24.
DHDPS structure shows similar changes, and the cooperativity
observed for inhibition by bislysine, despite binding sites being
separated by over 50 Å, may be a manifestation of these
alterations in dimer−dimer interactions. It is noteworthy that
the structure of Y110F bound to lysine does not show these
changes relative to the unliganded structure, but bislysine-
bound Y110F does. The additional contacts to both monomer
surfaces within the allosteric site apparently overcome the loss
of interaction with Y110. It is clear that the summation of many
small effects upon the protein result in inhibition.
(
(
5) Skovpen, Y. V.; Palmer, D. R. J. Biochemistry 2013, 52, 5454.
6) Turner, J. J.; Gerrard, J. A.; Hutton, C. A. Bioorg. Med. Chem.
2
(
005, 13, 2133.
7) Turner, J. J.; Healy, J. P.; Dobson, R. C. J.; Gerrard, J. A.; Hutton,
C. A. Bioorg. Med. Chem. Lett. 2005, 15, 995.
8) Boughton, B. A.; Griffin, M. D. W.; O’Donnell, P. A.; Dobson, R.
(
C. J.; Perugini, M. A.; Gerrard, J. A.; Hutton, C. A. Bioorg. Med. Chem.
2008, 16, 9975.
(9) Boughton, B. A.; Hor, L.; Gerrard, J. A.; Hutton, C. A. Bioorg.
Med. Chem. 2012, 20, 2419.
(
(
10) Phenix, C. P.; Palmer, D. R. J. Biochemistry 2008, 47, 7779.
11) Muscroft-Taylor, A. C.; Soares da Costa, T. P.; Gerrard, J. A.
CONCLUSION
■
Biochimie 2010, 92, 254.
12) Voss, J. E.; Scally, S. W.; Taylor, N. L.; Atkinson, S. C.; Griffin,
(
Biomimetic design based on the X-ray crystal structure of lysine
molecules bound to the allosteric site of CjDHDPS resulted in
a very potent inhibitor, one whose design appears to protect it
from simple mutational resistance. The bis-inhibitor approach
clearly demonstrates how a single molecule that binds to two
sites (in this case two binding sites within a single allosteric
cavity) can be more effective than two molecules. The observed
cooperativity of (R,R)-bislysine verifies that inhibition is
achieved in part by structural effects across the tetrameric
enzyme. By demonstrating that this enzyme is susceptible to
potent inhibitors, we believe the dap pathway remains a viable
target for novel antibiotic research.
M. D. W.; Hutton, C. A.; Parker, M. W.; Alderton, M. R.; Gerrard, J.
A.; Dobson, R. C. J.; Dogovski, C.; Perugini, M. A. J. Biol. Chem. 2010,
2
(
85, 5188.
13) Atkinson, S.; Dogovski, C.; Downton, M.; Czabotar, P.; Dobson,
R. J.; Gerrard, J.; Wagner, J.; Perugini, M. Plant Mol. Biol. 2013, 81,
4
31.
(14) Conly, C. J. T.; Skovpen, Y. V.; Li, S.; Palmer, D. R. J.; Sanders,
D. A. R. Biochemistry 2014, 53, 7396.
15) Morrison, J. F.; Walsh, C. T. Adv. Enzymol. Relat. Areas Mol.
Biol.; John Wiley & Sons, Inc.: 1988, p 201.
(
(
(
16) Frieden, C. J. Biol. Chem. 1970, 245, 5788.
17) Kabsch, W. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010, 66,
1
25.
18) Vagin, A.; Teplyakov, A. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2010, 66, 22.
(
ASSOCIATED CONTENT
■
(
(
19) Matthews, B. W. J. Mol. Biol. 1968, 33, 491.
20) Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.; Davis,
*
S
Supporting Information
The Supporting Information is available free of charge on the
I. W.; Echols, N.; Headd, J. J.; Hung, L.-W.; Kapral, G. J.; Grosse-
ACS Publications website at DOI: 10.1021/jacs.5b12695.
Kunstleve, R. W.; McCoy, A. J.; Moriarty, N. W.; Oeffner, R.; Read, R.
F
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T. C.; Zwart, P. H.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010, 66, 213.
(
21) Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D: Biol. Crystallogr.
004, 60, 2126.
22) Hutton, C. A.; Perugini, M. A.; Gerrard, J. A. Mol. BioSyst. 2007,
, 458.
23) Yeh, P.; Sicard, A. M.; Sinskey, A. J. Mol. Gen. Genet. 1988, 212,
05.
24) Neidhardt, F. C.; Curtiss, R. Escherichia coli and Salmonella:
cellular and molecular biology, 2nd ed.; ASM Press: Washington, D.C.,
996.
25) Kobayashi, K.; Ehrlich, S. D.; Albertini, A.; Amati, G.; Andersen,
2
(
3
(
1
(
1
(
K. K.; Arnaud, M.; Asai, K.; Ashikaga, S.; Aymerich, S.; Bessieres, P.;
Boland, F.; Brignell, S. C.; Bron, S.; Bunai, K.; Chapuis, J.;
Christiansen, L. C.; Danchin, A.; Debarbouille, M.; Dervyn, E.;
Deuerling, E.; Devine, K.; Devine, S. K.; Dreesen, O.; Errington, J.;
Fillinger, S.; Foster, S. J.; Fujita, Y.; Galizzi, A.; Gardan, R.; Eschevins,
C.; Fukushima, T.; Haga, K.; Harwood, C. R.; Hecker, M.; Hosoya, D.;
Hullo, M. F.; Kakeshita, H.; Karamata, D.; Kasahara, Y.; Kawamura, F.;
Koga, K.; Koski, P.; Kuwana, R.; Imamura, D.; Ishimaru, M.; Ishikawa,
S.; Ishio, I.; Le Coq, D.; Masson, A.; Mauel, C.; Meima, R.; Mellado, R.
P.; Moir, A.; Moriya, S.; Nagakawa, E.; Nanamiya, H.; Nakai, S.;
Nygaard, P.; Ogura, M.; Ohanan, T.; O’Reilly, M.; O’Rourke, M.;
Pragai, Z.; Pooley, H. M.; Rapoport, G.; Rawlins, J. P.; Rivas, L. A.;
Rivolta, C.; Sadaie, A.; Sadaie, Y.; Sarvas, M.; Sato, T.; Saxild, H. H.;
Scanlan, E.; Schumann, W.; Seegers, J. F.; Sekiguchi, J.; Sekowska, A.;
Seror, S. J.; Simon, M.; Stragier, P.; Studer, R.; Takamatsu, H.; Tanaka,
T.; Takeuchi, M.; Thomaides, H. B.; Vagner, V.; van Dijl, J. M.;
Watabe, K.; Wipat, A.; Yamamoto, H.; Yamamoto, M.; Yamamoto, Y.;
Yamane, K.; Yata, K.; Yoshida, K.; Yoshikawa, H.; Zuber, U.;
Ogasawara, N. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 4678.
(
26) Akerley, B. J.; Rubin, E. J.; Novick, V. L.; Amaya, K.; Judson, N.;
Mekalanos, J. J. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 966.
27) Schnell, R.; Oehlmann, W.; Sandalova, T.; Braun, Y.; Huck, C.;
Maringer, M.; Singh, M.; Schneider, G. Plos One 2012, 7, DOI:
0.1371/journal.pone.0031133.
28) Coulter, C. V.; Gerrard, J. A.; Kraunsoe, J. A. E.; Moore, D. J.;
Pratt, A. J. Tetrahedron 1996, 52, 7127.
29) Hutton, C. A.; Southwood, T. J.; Turner, J. J. Mini-Rev. Med.
Chem. 2003, 3, 115.
30) Coulter, C. V.; Gerrard, J. A.; Kraunsoe, J. A. E.; Pratt, A. J.
Pestic. Sci. 1999, 55, 887.
31) Couper, L.; Mckendrick, J. E.; Robins, D. J.; Chrystal, E. J. T.
Bioorg. Med. Chem. Lett. 1994, 4, 2267.
32) Walters, D. R.; McPherson, A.; Robins, D. J. Mycol. Res. 1997,
01, 329.
33) Mitsakos, V.; Dobson, R. C. J.; Pearce, F. G.; Devenish, S. R.;
(
1
(
(
(
(
(
1
(
Evans, G. L.; Burgess, B. R.; Perugini, M. A.; Gerrard, J. A.; Hutton, C.
A. Bioorg. Med. Chem. Lett. 2008, 18, 842.
(
34) Boughton, B. A.; Dobson, R. C.; Gerrard, J. A.; Hutton, C. A.
Bioorg. Med. Chem. Lett. 2008, 18, 460.
35) Singh, V.; Evans, G. B.; Lenz, D. H.; Mason, J. M.; Clinch, K.;
(
Mee, S.; Painter, G. F.; Tyler, P. C.; Furneaux, R. H.; Lee, J. E.;
Howell, P. L.; Schramm, V. L. J. Biol. Chem. 2005, 280, 18265.
G
DOI: 10.1021/jacs.5b12695
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX