A Mechanism-Based Inhibitor Targeting the DD-Transpeptidase Activity of Bacterial Penicillin-Binding Proteins

Article abstract of DOI:10.1021/ja038445l

Penicillin-binding proteins (PBPs) are responsible for the final stages of bacterial cell wall assembly. These enzymes are targets of β-lactam antibiotics. Two of the PBP activities include DD-transpeptidase and DD-carboxypeptidase activities, which carry

Full text of DOI:10.1021/ja038445l

Published on Web 12/03/2003
A Mechanism-Based Inhibitor Targeting the
DD-Transpeptidase Activity of Bacterial Penicillin-Binding
Proteins
Mijoon Lee, Dusan Hesek, Maxim Suvorov, Wenlin Lee, Sergei Vakulenko, and
Shahriar Mobashery*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556
Received September 10, 2003; E-mail: mobashery@nd.edu
Abstract: Penicillin-binding proteins (PBPs) are responsible for the final stages of bacterial cell wall
assembly. These enzymes are targets of â-lactam antibiotics. Two of the PBP activities include
DD-transpeptidase and DD-carboxypeptidase activities, which carry out the cross-linking of the cell wall and
trimming of the peptidoglycan, the major constituent of the cell wall, by an amino acid, respectively. The
activity of the latter enzyme moderates the degree of cross-linking of the cell wall, which is carried out by
the former. Both these enzymes go through an acyl-enzyme species in the course of their catalytic events.
Compound 6, a cephalosporin derivative incorporated with structural features of the peptidoglycan was
conceived as an inhibitor specific for DD-transpeptidases. On acylation of the active sites of DD-
transpeptidases, the molecule would organize itself in the two active site subsites such that it mimics the
two sequestered strands of the bacterial peptidoglycan en route to their cross-linking. Hence, compound
6 is the first inhibitor conceived and designed specifically for inhibition of DD-transpeptidases. The compound
was synthesized in 13 steps and was tested with recombinant PBP1b and PBP5 of Escherichia coli, a
DD-transpeptidase and a DD-carboxypeptidase, respectively. Compound 6 was a time-dependent and
irreversible inhibitor of PBP1b. On the other hand, compound 6 did not interact with PBP5, neither as an
inhibitor (reversible or irreversible) nor as a substrate.
Penicillin-binding proteins (PBPs) are important bacterial
enzymes that carry out the final steps of the bacterial cell wall
assembly.1-3 Three of these important activities are the trans-
glycosylase, DD-transpeptidase, and DD-carboxypeptidase reac-
tions, which are performed on the surface of the bacterial plasma
membrane. The transglycosylase activity catalyzes the polym-
erization of lipid II to give rise to the polymeric glycosidic
backbone of the cell wall, repeating units of N-acetylmuramic
acid (MurNAc) and N-acetylglucosamine (GlcNAc). A uniquely
bacterial peptide is appended to the muramic acid residue (1).
While some structural variations to the peptide portion of the
bacterial peptidoglycan are seen, that shown for 1 is typically
found in all Gram-negative and some Gram-positive bacteria.
DD-Transpeptidases carry out the indispensable cross-linking
step of the cell wall. The cross-linking reaction gives the
structural rigidity to the cell wall that is critical for bacterial
survival. The DD-transpeptidase activity swaps one amide bond
for another, in the process of which the cross-linked cell wall
is formed (1 f 2 f 4; Figure 1A). The DD-carboxypeptidase
activity moderates the degree of cross-linking by removing the
terminal D-Ala of the peptidoglycan (Figure 1B), hence pre-
empting the possibility of cross-linking. Both the DD-carboxy-
peptidase and DD-transpeptidase activities utilize an active-site
serine strategy in their mechanisms. The serine experiences
acylation by the peptidoglycan (2) in both groups of enzymes.
In DD-transpeptidases, a second strand of peptidoglycan (3) binds
the active site to undergo the cross-linking reaction with the
acyl-enzyme species (2), to give rise to the cross-linked cell
wall (4). The acyl-enzyme species undergoes hydrolysis in DD-
carboxypeptidases (2 f 5), to result in a smaller peptide.
Tipper and Strominger proposed a number of years ago that
the backbone of a â-lactam antibiotic mimics the acyl-D-Ala-
D-Ala portion of the peptidoglycan structure.4 By so-doing,
â-lactam antibiotics also acylate the active site serine of DD-
carboxypeptidases and DD-transpeptidases. However, in contrast
to the case of the peptidoglycan that acylates the active site
and allows the terminal D-Ala to serve as the leaving group,
acylation of the active site by â-lactams leaves the “leaving
group” covalently tethered to the “acyl-enzyme” species. Hence,
the enzyme is essentially irreversibly inactivated, which is the
basis for how â-lactam antibiotics kill bacteria. Edified by this
information, we conceived of cephalosporin 6. This compound
has been incorporated with structural features that mimic the
first strand of the peptidoglycan (the red portion of 6), hence
the molecule should acylate the enzyme active site serine.
However, on serine acylation, the portions of the structure of
species 7 drawn in blue would mimic the second strand of the
(1) Van Heijenoort, J. Glycobiology 2001, 11, 25R-36R.
(2) Goffin, C.; Ghuysen, J. M. Microbiol. Mol. Biol. ReV. 2002, 66, 702-738.
(3) Massova, I.; Mobashery, S. Antimicrob. Agents Chemother. 1998, 42, 1-17.
(4) Tipper, D. J.; Strominger, J. L. Proc. Natl. Acad. Sci. U.S.A. 1965, 54,
1133-1141.
9
16322
J. AM. CHEM. SOC. 2003, 125, 16322-16326
10.1021/ja038445l CCC: $25.00 © 2003 American Chemical Society

Inhibitor Targeting DD-Transpeptidase Activity
A R T I C L E S
Figure 1. (A) Mechanistic steps in cross-linking of the cell wall by DD-transpeptidases. (B) Mechanistic steps in the activity of DD-carboxypeptidases.
peptidoglycan sequestered within the active site of a DD-
transpeptidase, on a trajectory for the cross-linking event. In
light of the fact that compound 6 would have structural features
that would occupy both peptidoglycan-binding subsites, we were
anticipating that this molecule should be specific for DD-
transpeptidases. For the same structural reason, we had antici-
pated that a molecule such as 6 would not be recognized by
DD-carboxypeptidases, which interacts with one strand of
peptidoglycan.
We disclose herein the multistep synthesis of compound 6.
Furthermore, we describe the evaluation of compound 6 with
PBP1b (which exhibits the DD-transpeptidase activity) and with
PBP5 (which is a DD-carboxypeptidase), both from Escherichia
coli. As anticipated, the compound inhibited the DD-transpep-
tidase but was not even recognized by DD-carboxypeptidase,
either as a substrate or as an inhibitor.
Compound 6 was synthesized according to the approach
outlined in Scheme 1. The key intermediate 12 was obtained
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16323

A R T I C L E S
Lee et al.
Scheme 1
by the Wittig reaction of the phosphonium salt 10 with aldehyde
11 in the presence of potassium trimethylsilanolate. Compound
12 was briefly hydrogenolyzed over Pd/C in a mixture of acetic
acid and MeOH. The resultant amine 13 was coupled with the
mixed anhydride 14 to give compound 15. Subsequent hydro-
genation of 15 over Pd/C resulted in compound 16. Jones
oxidation of the alcohol moiety in 16 gave a mixture of the
desired acid 17 and the sulfoxides 18a and 18b. The protected
D-Ala-D-Ala-OtBu (19) was coupled to the mixture of 17 and
18a,b in the presence of EDCI and HOBt, giving compounds
20 and 21a,b. The mixture was stirred in a solution of stannous
chloride and acetylchloride in DMF to allow the reduction of
the undesired 21a,b to the desired 20. This procedure was
reported previously for reduction of sulfoxides to sulfides.5
Selective deprotection of the Boc group of 20 was achieved by
(5) Kaiser, G. V.; Cooper, R. D. G.; Koehler, R. E.; Murphy, C. F.; Webber,
J. A.; Wright, I. G.; Van Heyningen, E. M. J. Org. Chem. 1970, 35, 2430-
2433.
9
16324 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003

Inhibitor Targeting DD-Transpeptidase Activity
A R T I C L E S
treatment with iodotrimethylsilane (TMSI) and 2,6-di-tert-butyl-
4-methylpyridine in acetonitrile. The resultant amine 22 was
coupled with the tripeptide 23 (N-Ac-L-Ala-γ-D-Glu(R-OtBu)-
L-Lys(ꢀ-N-Boc)) to give 24. A global deprotection of all the
protective groups of 24 in trifluoroacetic acid produced the
desired final cephalosporin 6.
To demonstrate that compound 6 is indeed an inhibitor of
DD-transpeptidase and not of DD-carboxypeptidases, we were
in need of these enzymes. These proteins are difficult to study
because they are anchored on the cytoplasmic membrane of
bacteria by amino acid stretches at either the N- or C-terminal
portions of PBPs. To obtain the proteins in soluble forms, we
set out to remove the membrane anchors from PBP1b and PBP5
of E. coli, a DD-transpeptidase and a DD-carboxypeptidase,
respectively. Both PBP1b6 and PBP57 have been cloned
previously, but these clones are generally not available. Hence,
we undertook to clone these two proteins within our laboratory.
Figure 2. Double-reciprocal plot of the observed rate constants for enzyme
inactivation Versus the concentrations of inhibitor 6.
Incubation of PBP1b with compound 6 resulted in time-
dependent loss of activity. A series of the experiments were
carried out at different inhibitor concentrations (0.4-3.0 mM),
at a fixed concentration of the protein. At certain time intervals,
aliquots were removed, diluted, and incubated with bocillin to
determine the amount of the active enzyme remaining by
fluorescence measurements of the bocillin-bound enzyme by
SDS-PAGE. The double-reciprocal plot of the observed rate
constants for enzyme inactivation (kobs) as a function of inhibitor
concentration (Figure 2)13 was plotted to determine the first-
order inactivation rate constant (kinact ) 0.0072 ( 0.0007 min)
and the inhibitor constant (Ki ) 2.5 ( 0.8 mM).
At first glance, the millimolar value for the inhibitor constant
gives the impression that the inhibitor might not conform well
to the requirements of the active site for binding. This is not
so, for the following reason. Many important metabolites are
present in millimolar concentrations; hence the enzymes that
process them have evolved to experience saturation with these
substrates in the same concentration range. For example, of the
21 enzymes of the glycolytic and tricarboxylic acid pathways,
10 exhibit Km values in the millimolar range. The nascent
peptidoglycan is readily available to the membrane-anchored
PBPs in the periplasmic space. The high effective local
concentration of peptidoglycan would facilitate its processing
by the PBPs for favorable entropic reasons. Indeed, we have
evidence that the PBPs that have been studied in this report
process their peptidoglycan substrates with Km values in the
millimolar range (unpublished data); hence the argument that
has been presented above holds in their cases.
PBP5 is anchored to the plasma membrane via a carboxy-
terminal 18-amino-acid8 domain in its native state. To avoid
accumulation of the overexpressed protein in the periplasm,
where the peptidoglycan (the PBP substrate) is located, the PBP5
gene lacking the signal peptide and the anchor regions was
cloned directly under the T7 promoter of the pET-24a(+) vector.
The goal was achieved by PCR-amplifying the PBP5 gene from
E. coli. The protein was expressed in the cytoplasm of E. coli
BL21 and was purified to apparent homogeneity after one
chromatographic step. The final yield was approximately 120
mg of protein per liter of culture.
A similar strategy was chosen for cloning of the ponB gene
encoding PBP1b. The periplasmic portion of the protein lacking
membrane anchor9 was amplified by PCR and cloned under
T7 promoter into pET-33Sh vector, which was derived from
commercially available pET-33b(+) (Novagene). The resulting
construct had the membrane anchoring motif replaced with a
His-Tag label, and the protein was expressed in the cytoplasm
of E. coli. For purification of the His-tagged PBP1b, a two-
step purification was used. The final yield of the protein was
20-25 mg of protein per liter of culture.
Both PBP1b10 and PBP511,12 are known to function in their
nonmembrane-anchored forms without involvement of other
proteins. With the availability of these two PBPs in our hands,
we attempted to evaluate compound 6 with them in enzymo-
logical experiments. Specifically, we investigated whether the
compound inhibited these PBPs or served as a substrate to them.
As it turned out, compound 6 was an inactivator of PBP1b, but
it neither served as a substrate to PBP5 nor inhibited it
(concentrations of up to 5 mM). Hence, PBP5 did not recognize
compound 6 as a suitable molecule for binding in the active
site. However, as anticipated by design, compound 6 not only
was recognized by the active site of PBP1b but also served as
an inactivator of the enzyme.
We conclude that compound 6 is an effective mimic of the
peptidoglycan structure, whereby it acylates only the active site
of the DD-transpeptidase (PBP1b). On enzyme acylation, species
7 should occupy both substites of the active site, resulting in
effective inhibition of the enzyme. Availability of cephalosporin
6 should be a boon to structural biological studies of DD-
transpeptidase. The knowledge of the structure of the acyl-
enzyme species 7 would help elucidate the manner in which
DD-transpeptidases sequester the two strands of the polymeric
peptidoglycan in the active site, an event that is critical for
(6) Pla, J.; Rojo, F.; de Pedro, M. A.; Ayala, J. A. J. Bacteriol. 1990, 172,
4448-4455.
(7) Stoker, N. G.; Broome-Smith, J. K.; Edelman A.; Spratt B. G. J. Bacteriol.
1983, 155, 847-853.
(13) Silverman, R. Mechanism-Based Enzyme InactiVation: Chemistry and
Enzymology; CRC Press: Boca Raton, LA, 1988; p 22.
(8) Jackson, M. E.; Pratt. J. M. Mol. Microbiol. 1987, 1, 23-28.
(9) Edelman, A.; Bowler, L.; Broome-Smith, J. K.; Spratt, B. G. Mol. Microbiol.
1987, 1, 101-106.
(14) Compound 6 was designed based on the peptidoglycan structure for Gram-
negative bacteria, since this is the most common version in nature. Because
Gram-negative bacteria have an outer membrane that excludes molecules
larger than 700 Da, we did not expect that cephalosporin 6 (molecular
weight of 985 Da) would be able to reach to the surface of the inner
membrane of Gram-negatives, where PBPs are. Hence, cephalosporin 6
has no significant antibacterial activity.
(10) Schwartz, B.; Markwalder, J. A.; Wang, Y. J. Am. Chem. Soc. 2001, 123,
11638-11643.
(11) Gutheil, W. G. Anal. Biochem. 1998, 259, 62-67.
(12) Hesek, D.; Suvorov, D.; Morio. K.; Lee, M.; Brown, S.; Vakulenko, S. B.;
Mobashery, S. (submitted).
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16325

A R T I C L E S
Lee et al.
survival of all bacteria. Such elucidation of the mechanism soon
would pave the way in devising the next generations of
antibiotics (â-lactam and otherwise) that would manifest their
activity by inhibition of the events that lead to cell wall cross-
linking.14
Supporting Information Available: Experimental procedures,
including those for cloning of the two PBPs used in this study,
their purification, synthetic protocols, and characterization of
all new compounds (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
Acknowledgment. This research was supported by the
National Institutes of Health. We acknowledge the generous
gift of compound 8 from the Otsuka Chemical Co. of Osaka,
Japan.
JA038445L
9
16326 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003