Dual substrate specificity of bacillus subtilis PBP4a

Article abstract of DOI:10.1021/bi400211q

Bacterial dd-peptidases are the targets of the β-lactam antibiotics. The sharp increase in bacterial resistance toward these antibiotics in recent years has stimulated the search for non-β-lactam alternatives. The substrates of dd-peptidases are elements of peptidoglycan from bacterial cell walls. Attempts to base dd-peptidase inhibitor design on peptidoglycan structure, however, have not been particularly successful to date because the specific substrates for most of these enzymes are unknown. It is known, however, that the preferred substrates of low-molecular mass (LMM) class B and C dd-peptidases contain the free N-terminus of the relevant peptidoglycan. Two very similar LMMC enzymes, for example, the Actinomadura R39 dd-peptidase and Bacillus subtilis PBP4a, recognize a d-α-aminopimelyl terminus. The peptidoglycan of B. subtilis in the vegetative stage, however, has the N-terminal d-α-aminopimelyl carboxylic acid amidated. The question is, therefore, whether the dd-peptidases of B. subtilis are separately specific to carboxylate or carboxamide or have dual specificity. This paper describes an investigation of this issue with B. subtilis PBP4a. This enzyme was indeed found to have a dual specificity for peptide substrates, both in the acyl donor and in the acyl acceptor sites. In contrast, the R39 dd-peptidase, from an organism in which the peptidoglycan is not amidated, has a strong preference for a terminal carboxylate. It was also found that acyl acceptors, reacting with acyl-enzyme intermediates, were preferentially d-amino acid amides for PBP4a and the corresponding amino acids for the R39 dd-peptidase. Examination of the relevant crystal structures, aided by molecular modeling, suggested that the expansion of specificity in PBP4a accompanies a change of Arg351 in the R39 enzyme and most LMMC dd-peptidases to histidine in PBP4a and its orthologs in other Bacillus sp. This histidine, in neutral form at pH 7, appeared to be able to favorably interact with both carboxylate and carboxamide termini of substrates, in agreement with the kinetic data. It may still be possible, in specific cases, to combat bacteria with new antibiotics based on particular elements of their peptidoglycan structure.

Full text of DOI:10.1021/bi400211q

Article
pubs.acs.org/biochemistry
Dual Substrate Specificity of Bacillus subtilis PBP4a
Venkatesh V. Nemmara,† S. A. Adediran,† Kinjal Dave,† Colette Duez,‡ and R. F. Pratt*,†
†Department of Chemistry, Wesleyan University, Lawn Avenue, Middletown, Connecticut 06459, United States
‡Centre d’Ingen
́
́ ́ ̀ ̀
iere des Proteines, Universite de Liege, B-4000 Sart Tilman, Liege, Belgium
S
* Supporting Information
ABSTRACT: Bacterial DD-peptidases are the targets of the β-lactam
antibiotics. The sharp increase in bacterial resistance toward these
antibiotics in recent years has stimulated the search for non-β-lactam
alternatives. The substrates of DD-peptidases are elements of peptidogly-
can from bacterial cell walls. Attempts to base DD-peptidase inhibitor
design on peptidoglycan structure, however, have not been particularly
successful to date because the specific substrates for most of these
enzymes are unknown. It is known, however, that the preferred substrates
of low-molecular mass (LMM) class B and C DD-peptidases contain the
free N-terminus of the relevant peptidoglycan. Two very similar LMMC enzymes, for example, the Actinomadura R39 DD-
peptidase and Bacillus subtilis PBP4a, recognize a D-α-aminopimelyl terminus. The peptidoglycan of B. subtilis in the vegetative
stage, however, has the N-terminal D-α-aminopimelyl carboxylic acid amidated. The question is, therefore, whether the DD-
peptidases of B. subtilis are separately specific to carboxylate or carboxamide or have dual specificity. This paper describes an
investigation of this issue with B. subtilis PBP4a. This enzyme was indeed found to have a dual specificity for peptide substrates,
both in the acyl donor and in the acyl acceptor sites. In contrast, the R39 DD-peptidase, from an organism in which the
peptidoglycan is not amidated, has a strong preference for a terminal carboxylate. It was also found that acyl acceptors, reacting
with acyl−enzyme intermediates, were preferentially D-amino acid amides for PBP4a and the corresponding amino acids for the
R39 DD-peptidase. Examination of the relevant crystal structures, aided by molecular modeling, suggested that the expansion of
specificity in PBP4a accompanies a change of Arg351 in the R39 enzyme and most LMMC DD-peptidases to histidine in PBP4a
and its orthologs in other Bacillus sp. This histidine, in neutral form at pH 7, appeared to be able to favorably interact with both
carboxylate and carboxamide termini of substrates, in agreement with the kinetic data. It may still be possible, in specific cases, to
combat bacteria with new antibiotics based on particular elements of their peptidoglycan structure.
DD-Peptidases catalyze the final step(s) in bacterial cell wall
biosynthesis. These enzymes, therefore, are important antibiotic
targets and do, in fact, represent the site of action of the β-
lactams.1,2 They are known to catalyze the peptidoglycan cross-
linking reaction of Scheme 1, an amide aminolysis (trans-
and represent the killing sites of the β-lactam antibiotics. The
LMM group is believed to contain carboxypeptidases and
endopeptidases, which are engaged in peptidoglycan matura-
tion and degradation and are nonessential to bacterial survival,
in the short term at least, although they are also inhibited by β-
lactams.
A representative example of general peptidoglycan structure
(stem peptide dimer) is shown in Figure 1. In this figure, 1
represents a peptidoglycan dimer and (a) and (b) represent the
two structural variants, carboxylate and carboxamide, discussed
in detail below. One might imagine that the DD-peptidases
would show some recognition of elements of this structure. The
results of experiments conducted in vitro with individual
purified DD-peptidases are decidedly ambivalent overall.3 In
general, purified HMM DD-peptidases show no significant
affinity for any part of 1 or catalyze any reaction of it at
biologically significant rates. Certain LMMA DD-peptidases
catalyze the carboxypeptidase reaction of D-alanine-terminating
peptides but, in general, show no significant affinity for any part
of 1. LMMB and LMMC enzymes, however, have been shown
Scheme 1
peptidase) reaction involving a covalent acyl−enzyme inter-
mediate. Although it is clear that the substrates of these
enzymes must be elements of peptidoglycan, the source of their
individual specificity is, in general, not understood.3,4
The issue is complicated by the presence in bacteria of a
variety of DD-peptidases with apparently different roles in cell
wall construction, and acting at different times in the cell
cycle.5,6 They are usually subdivided into a high-molecular mass
(HMM) group, with subclasses A and B, and a low-molecular
mass (LMM) group, with subclasses A−C.7 The former act as
transpeptidases (Scheme 1), are essential to bacterial survival,
Received: February 18, 2013
Revised: March 25, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
MATERIALS AND METHODS
■
D-Amino acids and D-amino acid amides were commercial
products (Sigma, ChemImpex) and used as received, except D-
lysinamide, which was prepared as described in the Supporting
Information. The synthesis of 2 in this laboratory has been
previously described.20 Penicillin V (phenoxymethylpenicillin)
was purchased from Sigma and its amide synthesized in this
laboratory according to the method of Johnson21 by E.
Baukanas. N-(Phenylacetyl)glycyl-D-thiolactate was synthesized
as previously described.22
Synthesis. The synthesis of 4 is outlined in Scheme 2, and
the experimental details are given below.
D-6-(N,N-Dibenzylamino)-6-(benzyloxycarbonyl)-trans-
hex-2-enoic Acid Ethyl Ester (7). To a stirred solution of 6
(2.84 g, 7 mmol, 1 equiv)12 in dry THF (20 mL) (under an
inert atmosphere) was added (carbethoxymethylene)-
triphenylphosphorane (3.2 g, 9.2 mmol, 1.3 equiv) at room
temperature. The reaction mixture was stirred for 4 h, after
which solvent was evaporated. The crude product was purified
by chromatography on silica gel (9.5:0.5 hexane:ethyl acetate),
yielding the product 7 as a colorless oil (1.62 g, 60% yield): 1H
NMR (CDCl3, 300 MHz) δ 1.27 (t, J = 6.9 Hz, 3H), 1.86 (m,
2H), 2.05 (m, 1H), 2.32 (m, 1H), 3.35 (t, J = 7.5 Hz, 2H), 3.50,
3.88 (AB q, J = 15 Hz, 4H), 4.15 (q, J = 7.2 Hz, 2H), 5.16, 5.26
(AB q, J = 10.8 Hz, 2H), 5.62 (d, J = 16 Hz, 1H), 6.75 (quint, J
= 7.2 Hz, 1H), 7.2−7.45 (m, 15H).
Figure 1. Stem peptide dimer of B. subtilis showing the sites of
amidation and DD-peptidase cleavage. The structural variants addressed
in this paper are (a) carboxylate and (b) carboxamide.
to exhibit strong affinity for the stem peptide N-terminus
appropriate to the particular organism from which they were
derived.4,8,9 This result is seen in the kinetics of turnover of
specific peptide substrates and the affinity of specific inhibitors
such as β-lactams and boronic acids. The structural bases of
these specificities have been revealed by crystal structures of
various complexes of the enzymes with these compounds.10−12
One LMMC DD-peptidase in which these issues have been
studied in some detail is PBP4a of Bacillus subtilis. This enzyme
has been shown to have considerable specificity toward peptide
2 (kcat/Km = 2.2 × 104 s−1 M−1) and is strongly inhibited by
boronic acid 3 (Ki = 9 nM).4 The structural basis of this
specificity has been revealed by crystallography.13
D-α-Aminopimelic Acid Ethyl Ester (8). To a solution of 7
(1.02 g, 2.3 mmol, 1.0 equiv) in a mixture of methanol (30 mL)
and ethyl acetate (5 mL) was added 10% Pd on activated
carbon (Pd/C) (260 mg). The hydrogenation reaction was
conducted at 50 psi H2 at room temperature for 16 h. Pd/C
was removed by filtration through a Celite pad. The filtrate was
evaporated leaving behind an off-white solid, which was dried
1
under reduced pressure to give 8 (335 mg, 78%): H NMR
(CD3OD, 300 MHz) δ 1.21 (t, J = 6.9 Hz, 3H), 1.42 (m, 2H),
1.62 (m, 2H), 1.80 (m, 2H), 2.34 (t, J = 7.5 Hz, 2H), 3.5 (m,
1H), 4.2 (q, J = 7.2 Hz, 2H).
It has also been shown, however, that during the vegetative
stage of growth, B. subtilis cell walls contain amidated meso-
diaminopimelic acid,14,15 and thus, the peptidoglycan structure
would be represented by 1b. On the other hand, in B. subtilis
spores, the predominant peptidoglycan is not amidated, as in
1a.16 Peptidoglycan containing amidated diaminopimelic acid is
not uncommon and, in some cases, has established function.17
The actual in vivo role of B. subtilis PBP4a is not known,
although it is apparently expressed in the late vegetative stage of
growth18 and thus, as an endopeptidase, may be faced with 1b
as a substrate. On the other hand, if it is also active during
construction of the spore wall, it might also be expected to
hydrolyze 1a. The same issue must face other B. subtilis DD-
peptidases or penicillin-binding proteins (PBPs). B. subtilis
carries an extensive array of PBPs, some devoted to the
vegetative stage and others expressed during spore formation.19
The specificity of these enzymes for amidated versus non-
amidated peptidoglycan is, in general, not known. In this paper,
we explore the specificity of B. subtilis PBP4a for the specific
amidated acyl-donor and acyl-acceptor substrates 4 and 5, as
expressed in steady state kinetics. The analogy of the latter
substrate to an amidated penicillin is also explored.
(Z)-D-α-Aminopimelic Acid Ethyl Ester (9). To a stirred
solution of 8 and Na2CO3 in a water/dioxane (1:1) mixture (20
mL) at ice temperature was added benzyl chloroformate (0.35
mL, 3.3 mmol, 2.0 equiv) in dioxane (3 mL) dropwise over 10
min. The reaction mixture was then brought to room
temperature and stirred overnight. The mixture was extracted
with ether (3 × 20 mL), and the pH of the aqueous layer was
adjusted to 1.0 by addition of 1.0 M HCl. The acidified layer
was extracted with ethyl acetate (5 × 50 mL) and separated.
The combined organic layers were dried over Na2SO4 and
evaporated to dryness. The residue was dried further under
reduced pressure, and the resulting sticky gum was purified by
column chromatography on silica gel (4:3 hexane:ethyl acetate)
1
to yield 9 as a colorless oil (0.5 g, 95% yield): H NMR (D6−
DMSO, 300 MHz) δ 1.16 (t, J = 7.2 Hz, 3H), 1.31 (m, 2H),
1.49 (m, 2H), 1.65 (m, 2H), 2.26 (t, J = 7.5 Hz, 2H), 3.9 (m,
1H), 4.02 (q, J = 7.5 Hz, 2H), 5.05 (s, 2H), 7.35 (s, 5H), 7.6
(d, J = 8.1 Hz, 2H).
D - 6 - B e n z y l o x y c a r b o n y l a m i n o - 6 - [ N - ( 2 ′ , 4 ′ -
dimethoxybenzyl)carbamoyl]hexanoic Acid Ethyl Ester (10).
Compound 9 (0.5 g, 1.48 mmol, 1.0 equiv) in DMF (12 mL)
was added by syringe into a round-bottom flask containing
HATU (0.76 g, 2.0 mmol, 1.35 equiv) maintained under a
nitrogen atmosphere with constant stirring. The flask was
cooled with ice, and 2,4-dimethoxybenzylamine (0.3 mL, 2.0
B
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Scheme 2
mmol, 1.35 equiv) was added. DIPEA (0.64 mL, 3.7 mmol, 2.5
equiv) was then added dropwise over 20 min. The reaction
mixture was subsequently brought to room temperature and
stirred overnight. The solvent, DMF, was removed by rotary
evaporation, and the residue was extracted with ethyl acetate.
This solution was washed with 0.1 M HCl (20 mL), water, and
finally brine. The organic layer was separated, dried over
Na2SO4, and evaporated to dryness. The residual yellow oil was
purified by column chromatography on silica gel (3:7
hexane:ethyl acetate) to yield 10 as a colorless solid (440 mg,
70% yield): 1H NMR (D6− DMSO, 300 MHz) δ 1.16 (t, J = 7.2
Hz, 3H), 1.23 (m, 2H), 1.49 (m, 2H), 1.65 (m, 2H), 2.26 (t, J =
7.5 Hz, 2H), 3.73 (s, 3H), 3.76 (s, 3H), 3.9 (m, 1H), 4.02 (q, J
= 7.2 Hz, 2H), 4.15 (m, 2H), 5.02 (s, 2H), 6.45 (d, J = 7.8
Hz,1H), 6.53 (s, 1H), 7.04 (d, J = 8.4 Hz, 2H), 7.38 (m, 5H),
8.1 (m, 1H); ESI/MS(+) m/z 487.92 (M + 1), 510.27 (M +
23).
D - 6 - B e n z y l o x y c a r b o n y l a m i n o - 6 - [ N - ( 2 ′ , 4 ′ -
dimethoxybenzyl)carbamoyl]hexanoic Acid (11). LiOH (17
mg, 0.7 mmol, 1.0 equiv) was added to compound 10 (340 mg,
0.7 mmol, 1.0 equiv) dissolved in a water/THF (1:1) mixture
(20 mL) at 5 °C. The reaction mixture was stirred at 5−10 °C
until TLC showed complete conversion of the starting material.
After completion of the reaction, THF was removed by
evaporation and the pH of the aqueous layer adjusted to 1.0 by
addition of 1.0 M HCl. The aqueous layer was then extracted
with ethyl acetate (4 × 50 mL) and separated. The combined
organic layers were dried over Na2SO4 and evaporated to
dryness. The crude product was purified by column
chromatography on silica gel (9.3:0.7 DCM:MeOH) to yield
11 as a colorless solid (300 mg, 85% yield): 1H NMR (CD3OD,
300 MHz) δ 1.34 (m, 2H), 1.59 (m, 2H), 1.60 (m, 2H), 2.24
(t, J = 7.5 Hz, 2H), 3.76 (s, 3H), 3.78 (s, 3H), 4.05 (m, 1H),
4.3 (m, 2H), 5.07 (s, 2H), 6.43 (d, J = 7.8 Hz, 1H), 6.49 (s,
1H), 7.12 (d, J = 8.4 Hz, 2H), 7.20 (d, J = 7.8 Hz, 2H), 7.36
(m, 5H), 8.1 (m, 1H); ESI/MS(−) m/z 457.00 (M − 1).
D - 6 - B e n z y l o x y c a r b o n y l a m i n o - 6 - [ N - ( 2 ′ , 4 ′ -
dimethoxybenzyl)carbamoyl]hexanoyl-D-alanyl-D-alanine
Benzyl Ester (13). To a solution of 11 (200 mg, 0.44 mmol, 1.0
equiv) in DMF (3 mL) were added N-hydroxysuccinimide (63
mg, 0.55 mmol, 1.25 equiv) and 1-ethyl-3-[3-(dimethylamino)-
propyl]carbodiimide (105 mg, 0.55 mmol, 1.25 equiv) at ice
temperature. The reaction mixture was brought to room
temperature and stirred for 20 h. To this mixture was added a
solution of D-alanyl-D-alanine benzyl ester 124 (0.24 g, 0.66
mmol, 1.5 equiv) in DMF (3 mL) followed by the dropwise
addition of DIPEA (0.23 mL, 1.3 mmol, 3.0 equiv). The
mixture was stirred overnight, after which DMF was removed
by evaporation under vacuum. Water was added to the residue,
and the solid formed was collected by filtration and dried under
vacuum. The residual solid was recrystallized from 2-propanol
1
to yield 13 as a colorless solid (170 mg, 58% yield): H NMR
(D6− DMSO, 300 MHz) δ 1.07 (d, J = 6.9 Hz, 3H), 1.12 (d, J =
6.9 Hz, 3H), 1.28 (m, 2H), 1.43 (m, 2H), 1.60 (m, 2H), 2.06
(t, J = 6.8 Hz, 2H), 3.37 (s, 3H), 3.39 (s, 3H), 3.98 (m, 1H),
4.10 (m, 1H), 4.3 (m, 2H), 5.02 (s, 2H), 5.10 (s, 2H), 6.46 (d, J
= 8.4 Hz, 1H), 6.52 (s, 1H), 7.04 (d, J = 8.4 Hz, 2H), 7.36 (m,
5H), 7.93 (d, J = 8.1 Hz, 1H), 8.1 (m, 1H), 8.34 (d, J = 6.0 Hz,
1H); ESI/MS(+) m/z 691.0 (M + 1), 713.43 (M + 23).
D-6-Benzyloxycarbonylamino-6-carbamoylhexanoyl-D-
alanyl-D-alanine Benzyl Ester (14). TFA (1.5 mL) was added
to 13 (40 mg, 0.06 mmol) in a round-bottom flask at room
C
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
temperature, and the solution was stirred for 3.5 h. TLC
monitoring confirmed the disappearance of the starting
material. TFA was removed by evaporation, and the pink
solid thus obtained was washed with ether until the pink color
was completely removed. The crude product was purified by
preparative TLC on silica gel with 7.5% MeOH in DCM as the
0−0.30 μM, and 0−0.60 μM, respectively, for the R39 DD-
peptidase and 0.17 μM, 0.5 mM, 0−0.10 μM, and 0−6.0 μM,
respectively, for PBP4a. Compounds 20 and 21 were generated
from the hydrolysis of 2 and 4, respectively, in the presence of
PBP4a, and studied as inhibitors of this enzyme in situ as a
function of pH. Buffers for the latter experiments were 20 mM
MES (pH 6.0), 20 mM MOPS (pH 7.5), and 20 mM TAPS
(pH 8.8).
Computational Methods. Crystal structures of Actino-
madura R39 DD-peptidase [Protein Data Bank (PDB) entry
2XDM] and B. subtilis PBP4a (PDB entry 2J9P) with specific
ligands bound in the active site were used as the starting point
for model building and computations with Discovery Studio
version 2.5 (Accelrys). Tetrahedral intermediate structures of
the enzyme−substrate reaction were modeled with D-alanine or
D-alaninamide as the leaving group. A pH of 7.0 was set, with
positive charge on Lys52 (R39) and Lys55 (PBP4a) and
negative charge on Asp142 (Asp145), whereas His352 (PBP4a)
was either charged or uncharged depending on the environ-
ment desired. The CHARMm force field was employed for all
the computations.
In each case, the enzyme−substrate complex was hydrated
with a 20 Å sphere of water molecules centered at the active site
serine. The hydrated structure was energy-minimized using a
steepest descent algorithm (250 steps), followed by the
conjugate gradient method for a total of 1000 steps. The
energy-minimized structures were heated to 300 K in 10000
steps (10 ps) with coordinates saved every 100 steps. The
output structure of the equilibration run was then subjected to
a 1.0 ns production run at 300 K. Several structures were
selected from the output of each production run, and each of
these was energy-minimized using the protocol mentioned
above.
1
eluent, yielding 14 as a colorless solid (28 mg, 86% yield): H
NMR (CD3OD, 300 MHz) δ 1.27 (d, J = 7.2 Hz, 3H), 1.37 (d,
J = 7.5 Hz, 3H), 1.43 (m, 2H), 1.61 (m, 2H), 1.79 (m, 2H),
2.20 (t, J = 6.9 Hz, 2H), 4.05 (m, 1H), 4.31 (m, 2H), 4.45 (m,
1H), 5.07 (s, 2H), 5.14 (d, J = 5.4 Hz, 2H), 7.39 (s, 10H); ESI/
MS(+) m/z 541.0 (M + 1), 563.40 (M + 23).
D-6-Amino-6-carbamoylhexanoyl-D-alanyl-D-alanine (4).
To a solution of 14 (28 mg, 0.05 mmol) in methanol (20
mL) was added 10% Pd/C (10 mg). The hydrogenation
reaction was conducted at 40 psi H2 at room temperature for 5
h. Pd/C was removed by filtration through a Celite pad. The
filtrate was evaporated, leaving behind an off-white solid, which
was dried under reduced pressure to give 4 (7 mg, 50% yield):
1H NMR (D2O) δ 1.2 (m, 6H), 1.30 (m, 2H), 1.50 (m, 2H),
1.75 (m, 2H), 2.16 (t, J = 7 Hz, 2H), 3.85 (t, J = 6.6 Hz, 1H),
4.01 (q, J = 7.6 Hz, 1H), 4.14 (q, J = 7.6 Hz, 1H); high-
resolution ESI/MS(+) m/z 317.1816 (M + 1), calcd for
C13H25N4O5 317.1825.
Enzyme Kinetics. Steady state kinetics experiments were
conducted under conditions previously established to generate
high stable activity of Actinomadura R39 DD-peptidase and B.
subtilis PBP4a:4 20 mM sodium phosphate (pH 7.5) at 25 °C.
Measurements of enzyme activity with peptides 4 and 5 were
determined spectrophotometrically where the loss of the
peptide bond was monitored at wavelengths between 215
and 235 nm. Initial rates were fit to the Henri−Michelis−
Menten equation by a nonlinear squares procedure to yield
values of kcat and Km. The range of peptide substrate
concentrations employed was typically 0−1.0 or 0−2.0 mM,
and enzyme concentrations were typically 0.10−0.25 μM. In
cases where substrate concentrations approaching Km were not
achieved, values of kcat/Km were obtained by nonlinear least-
squares exponential fits to the progress curve. In the case of the
Actinomadura R39 DD-peptidase, the Km of the peptide
substrate 4 was determined as the Ki value from a competition
experiment in which N-(phenylacetyl)glycyl-D-thiolactate, 15,
was the spectrophotometric substrate (245 nm; Δε = 2500
cm−1 M−1; Km = 38.4 μM). Kinetics of aminolysis of 15 by
amino acids and amino acid amides were conducted as
previously described.23 Rate constants of inactivation of the
enzymes by the penicillins 16 and 17 were obtained from
similar experiments where total progress curves were fit to
Scheme 3 with Dynafit.24 Concentrations of enzyme, substrate
(15), 16, and 17 in these experiments were 0.14 μM, 0.5 mM,
RESULTS AND DISCUSSION
■
The substrate specificity of bacterial DD-peptidases in terms of
the structure of their presumed substrate, represented in Figure
1, is a matter of considerable interest because of its relevance to
antibiotic design. Considering the reactions catalyzed, trans-
peptidase, endopeptidase, and carboxypeptidase (Figure 1), and
the general mechanism of the reaction (Scheme 1), it is clear
that there are two elements of substrate specificity, that for the
acyl donor and that for the acyl acceptor.
PBP4a of B. subtilis is a LMM class C DD-peptidase, thought
likely to be an endopeptidase in vivo.4,25 The substrate
specificity of this enzyme is particularly interesting because of
the variable amidation of diaminopimelic acid exhibited in B.
subtilis peptidoglycan, as described in the introductory section.
To examine this issue, we first assessed the reactivity of PBP4a
with substrates 2, 4, and 5. As a control, the Actinomadura R39
DD-peptidase, another LMMC enzyme, was employed; as far as
we are aware, Actinomadura peptidoglycan does not contain
amidated diaminopimelic acid.9
Scheme 3
The synthesis of 4 was achieved as outlined in Scheme 2.
The orthogonally protected D-α-aminopimelic acid derivative
10 was obtained from the previously described versatile
intermediate 6. The ethyl ester 10 was then hydrolyzed to
produce the acid 11, which was coupled to D-alanyl-D-alanine
benzyl ester 12, yielding the protected tripeptide 13. The
carboxamide group of 13, protected as an N-(2,4-dimethox-
ybenzyl) derivative,26 was deprotected with trifluoroacetic acid
and the remaining benzyl groups by hydrogenation to afford
the required product 4. The more straightforward synthesis of 5
D
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Scheme 4
Table 1. Steady State Parameters for Peptide Hydrolysis by PBP4a and the R39 DD-Peptidase
enzyme
PBP4a
parameter
substrate 2
substrate 4
substrate 5
a
b
kcat (s−1
)
≥22
≥30
≥8.6
≥1.0
Km (mM)
cat/Km (s−1 M−1
≥1.0
≥1.0
k
)
2.2 × 104
1.0
(3.0 0.3) × 104
(8.6 0.2) × 103
(kcat/Km)/[kcat/Km(2)]
kcat (s−1
Km (mM)
cat/Km (s−1 M−1
(kcat/Km)/[kcat/Km(2)]
1.36
0.39
c
R39
)
7.4 0.9
16
1
≥1.29
≥1.0
(1.3 0.5) × 10−3
5.76 × 106
1.0
(30 6) × 10−3
5.2 × 105
0.09
k
)
(1.29 0.03) × 103
0.0022
a
b
c
From ref 13. From ref 4. From ref 20.
is outlined in Scheme 4 and described in detail in the
Supporting Information.
donor. The aminolysis of an acyl−enzyme intermediate should
pass through the same transition states and tetrahedral
intermediate as those from hydrolysis of the original amide
(Scheme 5).
Both the R39 DD-peptidase and PBP4a catalyzed hydrolysis
of the C-terminal amino acid residue of 4 and 5, D-alanine and
D-alaninamide, respectively, as demonstrated by electrospray
mass spectra of completed reaction mixtures (not shown).
Steady state rate parameters obtained from kinetics studies of
this reaction are presented in Table 1. These data show that the
R39 enzyme is strongly specific for the diacid 2 with
significantly less reactivity with the monoamides 4 and 5 and,
in particular, with 5 where the amide would be in the acceptor
site. The differences seem to be largely in the Km parameter,
suggesting differences in direct affinity for the active site,
although it should be noted that whether acyl−enzyme
intermediate formation or breakdown is rate-determining in
the presence of saturating substrate is currently unknown. In
the case of PBP4a, there is little difference in specificity for 2, 4,
and 5, suggesting that carboxylate and amide react equally well
in both sites. The common relatively high Km is notably
different from the situation with the R39 enzyme. It has been
suggested that this reflects a need for a protein conformational
change accompanying the reactions of PBP4a, perhaps deriving
from the presence of unreactive conformers in solution.4 The
striking difference in specificity for carboxylate versus amide
substrates between the two enzymes is most readily seen in
kcat/Km ratios (Table 1). The structural basis for these kinetics
results is discussed below.
Scheme 5
The acyl−enzyme intermediate, however, may be obtained
from another source, hydrolysis of a thiolester, for example. In
the case presented here, the steady state kinetics of aminolysis
of thiolester 15 was studied. Hydrolysis of the acyl−enzyme
intermediate from this substrate at saturation is probably rate-
determining with respect to thiolester hydrolysis because added
amine nucleophiles accelerate the observed rate at substrate
concentrations above Km. An example of this phenomenon is
shown in Figure 2A. The results of such experiments, the
determined k4 values (Scheme 5), are listed in Table 2. Some
amine acceptors at high concentrations also inhibited the
reaction (an example is shown in Figure 2B), most likely by
competition for the donor site (Scheme 5); this phenomenon
has been previously observed.23,27 In such cases, values of K5,
usually quite large (millimolar), are also given in Table 2.
Another assessment of the relative affinity of carboxylate and
amide for the acceptor site can be achieved by studying the
transfer of an acyl group to amine acceptors from a standard
E
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Table 2. Kinetic Parameters for the Aminolysis of 15,
Catalyzed by PBP4a and the R39 DD-Peptidase
a
enzyme
amine
k4 (s−1 M−1
63 1.5
)
K5 (mM)
PBP4a D-alanine (300)
D-alaninamide (400)
D-alanyl-D-alanine (280)
D-phenylalanine (80)
D-phenylalaninamide (84)
glycine (500)
130 16
b
1100 220
no
b
no
50 12
b
no
35
no
3
b
720 180
b
no
300 20
b
glycinamide (700)
640 300
490 95
(3.9 1.8) × 103
(3.2 1.4) × 104
no
ε-N-acetyl-D-lysine (94)
D-lysine (25)
23.4 1.5
b
no
b
D-lysinamide (16)
no
b
D-α-aminopimelic acid
150
2
no
(12.5)
meso-2,6-diaminopimelic acid 890 200
15.8 0.3
75 27
(40)
R39
D-alanine (300)
D-alaninamide (450)
D-phenylalanine (80)
D-phenylalaninamide
glycine (500)
150 74
b
49 16
(7.1 2.3) × 103
no
b
no
b
17
3
no
b
360 100
no
b
glycinamide
23 25
no
a
b
Highest concentration in parentheses. Not observed under the
conditions employed.
Another window into the specificity of the acceptor site can
be obtained by consideration of the relative reactivities of
penicillin V (16) and its amide (17) with these enzymes (Table
3). It is now accepted that β-lactams such as penicillins are
Table 3. Second-Order Acylation Rate Constants for
Inhibition of PBP4a and the R39 DD-Peptidase by β-Lactams
β-lactam
PBP4a
R39
16
(1.83 0.33) × 105
(4.23 0.02) × 103
0.023
(2.17 0.30) × 105
(9.1 3.0) × 10
0.00042
17
Figure 2. Effect of D-alaninamide (A) and N-ε-acetyl-D-lysine (B) on
initial rates of turnover of the thiolester 15 (0.5 mM), catalyzed by
PBP4a (0.17 μM). The points are experimental, and the lines
represent the fits of the data to Scheme 5 (see the text).
17/16
inhibitors of DD-peptidases to a considerable degree because of
their resemblance to D-Ala-D-Ala-terminated peptides.28,29
Thus, one might expect the carboxylate of a penicillin to bind
at the active side of a DD-peptidase in the same position as that
of the terminal carboxylate of a substrate. Crystal structures
generally support this proposition, at least as far as one can
judge from inert acyl−enzyme structures.10,30
The data of Table 3 show that penicillin V is an excellent
inhibitor of both PBP4a and the R39 DD-peptidase. In the
acylation transition state, the carboxylate group of 16 must fit
well into the active site of both enzymes. As generally expected
from precedent,31,32 the amide 17 is a considerably poorer
inhibitor of both enzymes, but as anticipated from the results
described above, it is a relatively much better inhibitor of
PBP4a than of the R39 DD-peptidase.
Evident from Table 2 is the much greater efficiency of D-
amino acid amides than D-amino acids in reaction with the
acyl−enzyme intermediate derived from reaction of 15 with
PBP4a. D-Lysinamide, the closest analogue to a “real” acceptor
(Figure 1), is particularly effective. The greater effectiveness of
the amide acceptor is more evident in Table 2 than in Table 1.
It may be that the diamide 18 would be an even better PBP4a
substrate than 2, 4, or 5. Further extension of the carboxamide
by N-substitution seemed unlikely to produce acyl acceptors; D-
alanyl-D-alanine, for example (Table 2), was only a weak
inhibitor. We are unaware of a bacterial peptidoglycan in which
amidated diaminopimelic acid is extended through the amide.
The structural basis of the different specificities of PBP4a and
the R39 DD-peptidase for amidated substrates was explored by
molecular modeling. Tetrahedral intermediate models, 19,
derived from the substrates of 4, 5, and 18 were constructed
from the crystal structures of a boronate transition state
analogue complex of the R39 DD-peptidase12 and an acyl−
enzyme complex of PBP4a.13 The initially constructed
structures, where more than one orientation of the amide
group was assessed, were subjected to molecular dynamics
In contrast to the results for PBP4a, the R39 enzyme
employed D-amino acid amide acceptors much less effectively
than the amino acids themselves (Table 2). These results for
both enzymes are thus in accord with those from the peptide
hydrolysis experiments described above (Table 1).
F
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
simulations as described in Materials and Methods. From these,
structures exhibiting the most favorable interactions with the
respective enzymes were selected for energy minimization,
using interaction energy calculations33 to subsequently choose
between them, where appropriate.
The active sites of PBP4a and the R39 DD-peptidase are very
similar and would be expected to interact with substrates and
their analogues in similar ways.12,13,34 In each case, there is a
strongly interacting oxyanion hole comprising two backbone
NH groups [Ser52 (49) and Thr413 (414); the residue number
of PBP4a is followed in parentheses by that of the
corresponding residue in the R39 DD-peptidase], which is
believed to stabilize anionic tetrahedral intermediates.
All models were initially constructed such that the tetrahedral
oxyanion occupied the oxyanion hole, and those interactions
were, in general, retained. In certain cases, contact between the
ligand and the oxyanion hole was lost during the molecular
dynamics run; such structures were considered nonproductive.
Other active site functional groups that either interact with
specific substrates or are believed to be directly involved in
catalysis are listed in Table 4, where they are subdivided into
Figure 3. Superimposition of the backbone atoms of the active site
residues of PBP4a (atomic colors) onto those of the R39 DD-peptidase
(turquoise). Only heavy atoms are shown.
interaction with 4 because the terminal carboxylate groups of β-
lactams34 and, most likely, substrates12 interact with both
Thr411 and Thr413 of the R39 DD-peptidase, and with their
analogues in other PBP structures.3 It seems unlikely, however,
that the mutation of Thr413 is specific for carboxamide binding
because the equivalent of Ser413 is found in most LMMC
enzymes, e.g., in Escherichia coli and other Gram-negative
species where diaminopimelic acid amidation does not usually
occur. More striking is the replacement of Arg351 with
histidine (His352) in the acyl-donor site. The side chain of
histidine, of course, may be cationic, like that of arginine, or
neutral, depending on its pKa and the pH of the medium.
To obtain evidence for the pKa of His352 of PBP4a, the
variation of the Ki value of the product inhibitors 20 and 21 was
determined as a function of pH (Table 5). The product 20
Table 4. Active Site Residues of PBP4a and the R39 DD-
Peptidase in Contact with Specific Substrates
R39
PBP4a
acyl-donor interaction
Trp139
Asp142
Tyr147
Arg351
Met414
Ser415
Ser49
Pro142
Asp145
Tyr150
His352
Leu415
Ser416
Ser52
Table 5. Inhibition of PBP4a by 20 and 21 as a Function of
pH
Ki (mM)
inhibitor
pH 6.0
pH 7.5
pH 8.8
acyl-acceptor interaction
20
21
0.9 0.1
0.28 0.08
0.24 0.06
0.48 0.08
0.38 0.04
Lys52
Lys55
>5
Ser298
Lys410
Thr411
Thr413
Ser299
Lys411
Thr412
Ser414
from 2 is known to occupy the donor site of the R39 DD-
peptidase11 and is likely to do so for PBP4a also.13 The Ki value
of neither 20 nor 21 shows significant variation between pH 7.5
and 8.8. This suggests the absence of a pKa of His352 in this
region, and thus the likelihood that its pKa is below 7.5. The
increase in Ki for 20 at pH 6.0 probably reflects the influence of
the increase of Km of the substrate below pH 7.5.23 The
apparently greater increase in Ki of 21 than 20 suggests that the
former inhibitor binds even more weakly to the protonated
form of the enzyme than does the latter. These data are
interpreted to mean that His352 contains a neutral imidazole
moiety at neutral pH. The structural models below, taken
together with the kinetic data provided above, support this
proposition.
those interacting with the acyl-donor segment of the substrate
and those interacting with the acyl-acceptor/leaving group. The
details of these interactions and their putative participation in
catalysis have been discussed previously.12,13
The superimposition diagram of Figure 3 emphasizes the
point made above concerning the similarity of the PBP4a and
R39 DD-peptidase active sites. Two differences in active site
residues between them, whose interactions with substrates
might change on amidation of the latter (1a vs 1b), are
apparent in Table 4 and Figure 3. The replacement of Thr413
in the R39 DD-peptidase with serine (414) in PBP4a may affect
G
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
moieties. It does not seem to interact significantly with His352.
On the other hand, the carboxamide of 4 interacts (NH-π) with
the neutral imidazole of His352 (Figure 4A). If His352 were
cationic, however, the carboxylate of 5 would interact tightly
with it (electrostatic) but the amide of 4 less so (Figure 5). The
structure in Figure 5 also shows the distortion of the β-3 strand
(which supports one element of the oxyanion hole) that occurs
throughout the molecular dynamics simulation. This suggests
protein-mediated interaction between the acyl-donor and
-acceptor sites, an issue that has previously been raised from
Figure 4 shows structures of peptides 4 (Figure 4A) and 5
(Figure 4B) in the PBP4a active site where His352 is presumed
−
to be neutral. These models show that both the CO2 and
CONH2 groups interact well with the acyl-donor site. The
carboxylate of 5 interacts directly [potential hydrogen bonds
(see Figure 4B)] with Ser416, with both the NH and the OH
Figure 4. Stereoviews of energy-minimized tetrahedral intermediate structures (19) formed upon reaction of specific tripeptide substrates with
PBP4a. The side chain of His352 is neutral. Only heavy atoms are shown except for hydrogens on the imidazole nitrogens of His352. (A) Peptide 4
and (B) peptide 5.
H
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 5. Stereoview of an energy-minimized tetrahedral intermediate structure (19) formed upon reaction of the specific tripeptide substrate 4 with
PBP4a. The side chain of His352 is protonated and thus cationic. Only heavy atoms are shown except for hydrogens on the imidazole nitrogens of
His352.
observations of crystal structures containing specific sub-
strates.35 The analogous R39 DD-peptidase models (Figure S1
of the Supporting Information) also seem to show a less
effective interaction of the amide of 4 than the carboxylate of 5
with the cationic (Arg351) donor site. Hence, the replacement
of Arg351 of the R39 DD-peptidase, which in vivo would only
come into contact with carboxylate substrates, with the neutral
(at pH 7) imidazole of His352 of PBP4a appears to be a
substitution favored by evolution to produce an enzyme that
might have to process both carboxylate and amidated
substrates. Models containing diamide 18 also show a snug
fit into the enzyme active site when the histidine is neutral
(Figure S2 of the Supporting Information). Figure 4B also
shows that the amide group of 5 interacts well in the acceptor
site, with apparent hydrogen bonds to the side chains of
Ser299, Lys411, and Ser414.
Arg351 of the R39 DD-peptidase appears to be essentially
completely conserved through LMMC DD-peptidases, most of
which interact with peptidoglycan carboxylates. FASTA
searches suggest that the histidine mutation is largely confined
to Bacillus sp. where diaminopimelate amidation might be
expected.14,36 A variety of other enzymes are required to
distinguish carboxylate and carboxamide groups, and many of
them use a strategy similar to that described above. For
example, certain amino acid aminotransferases employ an
arginine to accommodate a substrate γ-carboxylate and neutral
ligands to accommodate an analogous γ-amide.37 The class I E.
coli glutaminyl-tRNA synthetase employs neutral ligands for the
substrate γ-amide but an arginine to divert a γ-carboxylate into
a nonproductive complex.38 As an example, closer to home,
orthologues of LMMA E. coli PBP5 and PBP6 appear to use a
conserved arginine to bind the α-glutamyl carboxylate of
peptidoglycan substrates.39 On the other hand, orthologues of
PBP3 of Streptococcus pneumoniae, also LMMA DD-peptidases,
seem to have neutral ligands in the analogous position to
interact with the α-glutaminyl carboxamides of their
peptidoglycan.
The results described above strongly suggest that the PBP4a
DD-peptidase of B. subtilis is adapted to efficiently catalyze
reactions of peptidoglycan fragments containing both D-
aminocarboxylate and D-aminocarboxamide termini. In vivo,
therefore, this enzyme may well have a dual role, evolutionarily
selected by the presence of both termini in the peptidoglycan of
B. subtilis at different phases of growth.14,18 PBP4a is thought to
generally act as an endopeptidase in vivo.4,25 In contrast, the
R39 DD-peptidase, another LMMC enzyme, seems strongly
adapted to a single peptidoglycan structure. In principle,
inhibitor design could take advantage of the kind of dual
specificity found in PBP4a to target particular enzymes or
particular stages of the cell cycle. Design of DD-peptidase
inhibitors based on specific elements of peptidoglycan structure
remains an attractive goal, but not one easily achieved.3,40 The
affinity of other B. subtilis PBPs, including the essential HMM
enzymes, for carboxamide termini is an interesting issue, and
one that we are currently investigating.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic details for the preparation of peptide 5 and D-
lysinamide and structures of 2 and 4 bound to the
Actinomadura R39 DD-peptidase and of 18 bound to B. subtilis
PBP4a. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Telephone: (860) 685-2629. E-mail: rpratt@wesleyan.edu.
Fax: (860) 685-2211.
Funding
This research was supported by National Institutes of Health
Grant AI-17986 (R.F.P.) and in part by the Belgian Program on
Interuniversity Poles of Attraction initiated by the Belgian State,
I
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
protein 4a and its complex with a peptidoglycan-mimetic peptide. J.
Mol. Biol. 371, 528−539.
(14) Atrih, A., Bacher, G., Allmaier, G., Williamson, M. P., and Foster,
J. J. (1999) Analysis of peptidoglycan structure from vegetative cells of
B. subtilis 168 and role of PBPs in peptidoglycan maturation. J.
Bacteriol. 181, 3956−3966.
(15) Warth, A. D., and Strominger, J. L. (1971) Structure of the
peptidoglycan from vegetative cell walls of Bacillus subtilis. J. Bacteriol.
180, 4967−4973.
Prime Minister’s Office, Science Policy programming (IAP n°
P6/19). C.D. is a research associate of the FRS-FNRS, Belgium.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS
■
DCM, dichloromethane; DIPEA, diisopropylethylamine; DMF,
dimethylformamide; DMSO, dimethyl sulfoxide; EDC, N-
[(dimethylamino)propyl]-N′-ethylcarbodiimide; ESI/MS, elec-
trospray ionization mass spectrometry; HATU, O-(7-azabenzo-
triazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophos-
phate; HMM, high-molecular mass; LMM, low-molecular
mass; MES, 2-(N-morpholino)ethanesulfonic acid; MOPS, 3-
(morpholino)propanesulfonic acid; NMR, nuclear magnetic
resonance; PBP, penicillin-binding protein; TAPS, N-tris-
(hydroxymethyl)methyl-3-aminopropanesulfonic acid; TFA,
trifluoroacetic acid; THF, tetrahydofuran; TLC, thin-layer
chromatography.
(16) Warth, A. D., and Strominger, J. L. (1969) Structure of the
peptidoglycan in bacterial spores: Occurrences of the lactam of
muramic acid. Proc. Natl. Acad. Sci. U.S.A. 64, 528−535.
(17) Bernard, E., Rolain, T., Courtin, P., Hols, P., and Chapot-
Cartier, M.-P. (2011) Identification of the amidotransferase AsnB1 as
being reasonable for meso-diaminopimelic acid amidation in
Lactobacillus plantarum peptidoglycan. J. Bacteriol. 193, 6323−6330.
(18) Pederson, L. B., Murray, T., Popham, D. L., and Setlow, P.
(1998) Characterization of dacC, which encodes a new low-molecular
weight penicillin-binding protein in Bacillus subtilis. J. Bacteriol. 180,
4967−4973.
(19) Scheffers, D.-J. (2005) Dynamic localization of penicillin-
binding proteins during spore development in Bacillus subtilis.
Microbiology 151, 999−1012.
REFERENCES
■
(20) Anderson, J. W., Adediran, S. A., Charlier, P., Nguyen-Distec
̀
he,
(1) Waxman, D. J., and Strominger, J. L. (1983) Penicillin-binding
proteins and the mechanism of action of β-lactam antibiotics. Annu.
Rev. Biochem. 52, 825−869.
M., Frere, J.-M., Nicholas, R. A., and Pratt, R. F. (2003) On the
̀
substrate specificity of bacterial DD-peptidases: Evidence from two
series of peptidoglycan-mimetic peptides. Biochem. J. 373, 949−955.
(21) Johnson, D. A. (1953) Carboxy derivatives of benzylpenicillin. J.
Am. Chem. Soc. 75, 3636−3637.
(2) Macheboeuf, P., Contreras-Martel, C., Job, V., Dideberg, O., and
Dessen, A. (2006) Penicillin-binding proteins: Key players in bacterial
cell cycle and drug resistance processes. FEMS Microbiol. Rev. 30, 673−
691.
(22) Xu, Y., Soto, G., Adachi, H., Van der Linden, M. P. G., Keck, W.,
and Pratt, R. F. (1994) Relative specificities of a series of β-lactam-
reognizing enzymes towards the side chains of penicillins and acyclic
thioldepsipeptides. Biochem. J. 302, 851−856.
(3) Pratt, R. F. (2008) Substrate specificity of bacterial DD-peptidases
(penicillin-binding proteins). Cell. Mol. Life Sci. 65, 2138−2155.
(4) Nemmara, V. V., Dzhekieva, L., Sarkar, K. S., Adediran, S. A.,
Duez, C., Nicholas, R. A., and Pratt, R. F. (2011) Substrate specificity
of low-molecular mass bacterial DD-peptidases. Biochemistry 50,
10091−10101.
(23) Adediran, S. A., Kumar, I., Nagarajan, R., Sauvage, E., and Pratt,
R. F. (2011) Kinetics of reactions of the Actinomadura R39 DD-
peptidase with specific substrates. Biochemistry 50, 367−387.
(24) Kuzmic, P. (1996) Program DYNAFIT for the analysis of
enzyme kinetic data: Application to HIV proteinase. Anal. Biochem.
237, 260−273.
(5) Popham, D. L., and Young, K. D. (2003) Role of penicillin-
binding proteins in bacterial cell morphogenesis. Curr. Opin. Microbiol.
6, 594−599.
(6) Matthei, P.-J., Neves, D., and Dessen, A. (2010) Bridging cell wall
biosynthesis and bacterial morphogenesis. Curr. Opin. Struct. Biol. 20,
749−755.
(25) Duez, C., Zervosen, A., Teller, N., Melkonian, R., Banzubaze,
́
E.,
Bouillenne, F., Luxen, A., and Frere, J.-M. (2009) Characterization of
̀
the proteins encoded by the Bacillus subtilis yoxA-dacC operon. FEMS
Microbiol. Lett. 300, 42−47.
(7) Ghuysen, J.-M. (1991) Serine β-lactamases and penicillin-binding
proteins. Annu. Rev. Microbiol. 45, 37−67.
(26) Shimshock, S. S., Waltermire, R. E., and DeShong, P. (1991) A
total synthesis of ( )-tirandamycin B. J. Am. Chem. Soc. 113, 8791−
8796.
̀
(8) Leyh-Bouille, M., Nakel, M., Frere, J.-M., Johnson, K., Duez, C.,
Ghuysen, J.-M., Nieto, M., and Perkins, H. R. (1972) Penicillin-
sensitive DD-carboxypeptidases from Streptomyces strains R39 and K11.
Biochemistry 11, 1290−1298.
(27) Kumar, I., and Pratt, R. F. (2005) Transpeptidation of a specific
substrate catalyzed by the Streptomyces R61 DD-peptidase: Character-
ization of a chromogenic substrate and acyl acceptor design.
Biochemistry 44, 9971−9979.
(9) Ghuysen, J.-M., Leyh-Bouille, M., Campbell, J. N., Moreno, R.,
̀
Frere, J.-M., Duez, C., Nieto, M., and Perkins, H. R. (1973) Structure
of the wall peptidoglycan of Streptomyces R39 and the specificity and
profile of its exocellular DD-carboxypeptidase-transpeptidase for
peptide acceptors. Biochemistry 12, 1243−1251.
(28) Tipper, D. J., and Strominger, J. L. (1965) Mechanism of action
of penicillins: A proposal based on their structural similarity to acyl-D-
alanyl-D-alanine. Proc. Natl. Acad. Sci. U.S.A. 54, 1133−1141.
(29) Pratt, R. F. (2002) Functional evolution of the serine β-
lactamase active site. J. Chem. Soc., Perkin Trans. 2, 851−861.
(30) Silvaggi, N. R., Josephine, H. R., Kuzin, A. P., Nagarajan, R.,
Pratt, R. F., and Kelly, J. A. (2005) Crystal structures of complexes
between the R61 DD-peptidase and peptidoglycan-mimetic β-lactams:
A non-covalent complex with a “perfect penicillin”. J. Mol. Biol. 345,
521−533.
(10) McDonough, M. A., Anderson, J. W., Silvaggi, N. R., Pratt, R. F.,
and Kelly, J. A. (2002) Structures of two kinetic intermediates reveal
species specificity of penicillin-binding proteins. J. Mol. Biol. 322, 111−
122.
(11) Sauvage, E., Powell, A. J., Heilemann, J., Josephine, H. R.,
Charlier, P., Davies, C., and Pratt, R. F. (2008) Crystal structures of
complexes of bacterial DD-peptidases with peptidoglycan-mimetic
ligands; the substrate specificity puzzle. J. Mol. Biol. 381, 383−393.
(12) Dzhekieva, L., Rocaboy, M., Kerff, F., Charlier, P., Sauvage, E.,
and Pratt, R. F. (2010) Crystal structure of a complex between the
Actinomadura R39 DD-peptidase and a peptidoglycan-mimetic
boronate inhibitor: Interpretation of a transition state analogue in
terms of catalytic mechanism. Biochemistry 49, 6411−6419.
(13) Sauvage, E., Duez, C., Herman, R., Kerff, F., Petrella, S.,
(31) Holysz, R. P., and Stavely, H. E. (1950) Carboxy derivatives of
benzylpenicillin. J. Am. Chem. Soc. 72, 4760−4763.
(32) Varetto, L., DeMeester, F., Monnaie, D., Marchand-Brynaert, J.,
̀
Dive, G., Jacob, F., and Frere, J.-M. (1991) The importance of the
negative charge of β-lactam compounds in the interactions with active-
site serine DD-peptidases and β-lactamases. Biochem. J. 278, 801−807.
(33) Curley, K. C., and Pratt, R. F. (1997) Effectiveness of tetrahedral
adducts as transition-state analogs and inhibitors of the class C β-
Anderson, J. W., Adediran, S. A., Pratt, R. F., Frer
P. (2007) Crystal structure of the Bacillus subtilis penicillin-binding
̀
e, J.-M., and Charlier,
J
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
lactamase of Enterobacter cloacae P99. J. Am. Chem. Soc. 119, 1529−
1538.
(34) Sauvage, E., Herman, R., Petrella, S., Duez, C., Bouillenne, F.,
Frere, J.-M., and Charlier, P. (2005) Crystal structure of the
̀
Actinomadura R39 DD-peptidase reveals new domains in penicillin
binding proteins. J. Biol. Chem. 280, 31249−31256.
(35) Silvaggi, N. R., Anderson, J. W., Brinsmade, S. A., Pratt, R. F.,
and Kelly, J. A. (2003) The crystal structure of phosphonate-inhibited
D-Ala-D-Ala peptidase reveals an analogue of a tetrahedral transition
state. Biochemistry 42, 1199−1208.
(36) Hughes, R. C. (1970) Autolysis of isolated cell walls of Bacillus
licheniformis N.C.T.C.6346 and Bacillus subtilis Marburg strain 168.
Separation of the products and characterization of the mucopeptide
fragments. Biochem. J. 119, 849−860.
(37) Hirotsu, K., Goto, M., Okamoto, A., and Miyahara, I. (2005)
Dual substrate recognition of aminotransferases. Chem. Rec. 5, 160−
172.
(38) Bullock, T. L., Uter, N., Nissan, T. A., and Perona, J. J. (2003)
Amino acid discrimination by a class I aminoacyl-tRNA synthetase
specified by negative determinants. J. Mol. Biol. 328, 395−408.
(39) Chen, Y., Zhang, W., Shi, Q., Hesek, D., Lee, M., Mobashery, S.,
and Shoichet, B. K. (2009) Crystal structures of penicillin-binding
protein 6 from Escherichia coli. J. Am. Chem. Soc. 131, 14345−15354.
(40) Josephine, H. R., Charlier, P., Davies, C., Nicholas, R. A., and
Pratt, R. F. (2006) Reactivity of penicillin-binding proteins with
peptidoglycan-mimetic β-lactams: What’s wrong with these enzymes?
Biochemistry 45, 15873−15883.
K
dx.doi.org/10.1021/bi400211q | Biochemistry XXXX, XXX, XXX−XXX