Slow-binding inhibition of peptide deformylase by cyclic peptidomimetics as revealed by a new spectrophotometric assay

Article abstract of DOI:10.1016/j.bioorg.2004.01.001

A new spectrophotometric/fluorimetric assay for peptide deformylase (PDF) has been developed by coupling the PDF reaction with that of dipeptidyl peptidase I (DPPI) and using N-formyl-Met-Lys-AMC as substrate. Removal of the N-terminal formyl group by PDF renders the dipeptide an efficient substrate of DPPI, which subsequently removes the dipeptidyl units to release 7-amino-4-methylcoumarin as the chromophore/fluorophore. The PDF reaction is conveniently monitored on a UV-Vis spectrophotometer or a fluorimeter in a continuous fashion. The utility of the assay was demonstrated by determining the catalytic activity of PDF and the inhibition constants of PDF inhibitors. These studies revealed the slow-binding behavior of a previously reported macrocyclic PDF inhibitor. This method offers several advantages over the existing PDF assays and should be particularly useful for screening PDF inhibitors in the continuous fashion.

Full text of DOI:10.1016/j.bioorg.2004.01.001

BIOORGANIC
CHEMISTRY
Bioorganic Chemistry 32 (2004) 178–191
www.elsevier.com/locate/bioorg
Slow-binding inhibition of peptide
deformylase by cyclic peptidomimetics
as revealed by a new spectrophotometric assay
Kiet T. Nguyen, Xubo Hu, and Dehua Pei*
Department of Chemistry and Ohio State Biochemistry Program, The Ohio State University,
100 West 18th Avenue, Columbus, OH 43210, USA
Received 12 November 2003
Abstract
A new spectrophotometric/fluorimetric assay for peptide deformylase (PDF) has been de-
veloped by coupling the PDF reaction with that of dipeptidyl peptidase I (DPPI) and using N-
formyl-Met-Lys-AMC as substrate. Removal of the N-terminal formyl group by PDF renders
the dipeptide an efficient substrate of DPPI, which subsequently removes the dipeptidyl units
to release 7-amino-4-methylcoumarin as the chromophore/fluorophore. The PDF reaction is
conveniently monitored on a UV–Vis spectrophotometer or a fluorimeter in a continuous
fashion. The utility of the assay was demonstrated by determining the catalytic activity of
PDF and the inhibition constants of PDF inhibitors. These studies revealed the slow-binding
behavior of a previously reported macrocyclic PDF inhibitor. This method offers several ad-
vantages over the existing PDF assays and should be particularly useful for screening PDF
inhibitors in the continuous fashion.
Ó 2004 Elsevier Inc. All rights reserved.
Keywords: Peptide deformylase; Kinetics; Assay; Dipeptidyl peptidase; PDF inhibition
1. Introduction
Protein synthesis in prokaryotes initiates with an N-formylmethionyl-tRNAi, re-
sulting in N-terminal formylation of all nascent polypeptides. Peptide deformylase
* Corresponding author. Fax: 1-614-292-1532.
E-mail address: pei.3@osu.edu (D. Pei).
0045-2068/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2004.01.001

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
179
(PDF,1 EC 3.5.1.88) catalyzes the subsequent removal of the formyl group from the
majority of bacterial proteins [1]. PDF belongs to a new subclass of amide hydrolase,
which utilizes a tetrahedrally coordinated ferrous ion to activate the metal–bound
water for nucleophilic attack on the substrate formyl group [2–4]. Genetic studies
have shown that PDF is essential for bacterial survival [5–7], but has apparently
no catalytic function in mammals [8]. PDF is present in all bacterial organisms
whose genomes have been sequenced. These properties make PDF an attractive
target for designing novel, broad-spectrum antibacterial agents [9–11]. Many potent
PDF inhibitors have been reported in recent years, some of which showed excellent
antibacterial activity in vitro and in animal models [12–16]. One of these inhibitors
is currently in phase I clinical trials for the treatment of upper respiratory tract
infections.
To facilitate the mechanistic study of PDF and inhibitor screening, we and others
have developed several assay methods for this enzyme. One method monitors the re-
lease of formate, using a formate dehydrogenase (FDH) as the coupling enzyme
[17,18]. Due to the poor specific activity of formate dehydrogenases, this assay does
not work well in the continuous fashion for assaying potent inhibitors. A second as-
say employs N-formylmethionylleucyl-p-nitroanilide (f-ML-pNA) as substrate and
the deformylation reaction was coupled to Aeromonas aminopeptidase (AAP), which
hydrolyzes the product, Met-Leu-p-nitroanilide, to release the chromogenic p-nitro-
aniline [19]. This convenient, sensitive, and inexpensive assay has been the primary
method used in this laboratory for mechanistic studies. However, the coupling en-
zyme AAP is also a metallopeptidase and is strongly inhibited by most of the
PDF inhibitors. Finally, a third assay employs N-formyl-(b-thiaphenylalanine)-con-
taining peptide substrates that eliminate thiophenol upon PDF catalyzed hydrolysis
of the formyl group [20]. The released thiophenol is monitored spectrophotometri-
cally using EllmanÕs reagent, 5,50-dithiobis(2-nitrobenzoic acid) (DTNB). This assay
proves to be highly sensitive and convenient for high-throughput screening of non-
thiol-containing PDF inhibitors. Its drawback is that the substrate is accessible only
by multiple-step synthesis and can undergo spontaneous (albeit slow) release of thi-
ophenol in aqueous solution. Further, we have recently synthesized a macrocyclic
PDF inhibitor 1 (Fig. 1), in which the P10 and P30 side chains are covalently cross-
linked [16]. The macrocycle acts as a potent, slow-binding inhibitor. We experienced
major difficulties in its kinetic characterization using the existing assays. These diffi-
culties prompted us to develop yet another spectrophotometric assay of PDF in
which the PDF reaction is coupled to dipeptidyl peptidase I (DPPI, also called ca-
thepsin C). DPPI is a cysteine protease of high specific activity toward peptide sub-
strates [21]. Because PDF inhibitors do not generally inhibit DPPI, the assay was
conveniently carried out in the continuous fashion and used to fully characterize
the slow-binding inhibitor.
1
Abbreviations used: PDF, peptide deformylase; DPPI, dipeptidyl peptidase I; f-Met-Lys-AMC, 7-(N-
formylmethionyl-L-lysyl)amino-4-methylcoumarin.

180
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
Fig. 1. Structures of PDF inhibitors.
2. Materials and methods
2.1. Materials
Escherichia coli PDF was overexpressed in E. coli and purified to apparent
homogeneity with either Fe2þ or Co2þ as the divalent metal [4,18]. DPPI from bovine
spleen and yeast formate dehydrogenase were purchased from Sigma Chemical
(St. Louis, MO). Other chemicals were obtained from either Sigma–Aldrich (St. Louis,
MO) or Bio-Rad Laboratories (Hercules, CA) and Advanced ChemTech (Louisville,
KY). PDF inhibitors 1–3 were synthesized as previously described [8,16].
2.2. Analytical methods
Protein concentration was determined by Bradford assay using bovine serum
albumin as the protein standard. For PDF, the total protein concentration as deter-
mined by Bradford assay was corrected by a factor of 0.71 [4]. Substrate concentra-
tions were determined by base hydrolysis followed by measuring the absorbance of
7-amino-4-methylcoumarin at 360 nM (e ¼ 1:7 ꢀ 104 Mꢁ1 cmꢁ1).
2.3. 7-(Na-Benzyloxycarbonyl-Ne-t-butoxycarbonyl-
[N-Cbz-Lys(Boc)-AMC]
L-lysyl)amino-4-methylcoumarin
To a solution of Cbz-Lys(Boc)-OH (0.42 g, 1.1 mmol), 7-amino-4-methylcouma-
rin (0.17 g, 1.0 mmol) and HOBt (0.17 g, 1.1 mmol) in dichloromethane (10 mL)
was added EDC (0.23 g, 1.2 mmol) at 0 °C. The reaction was stirred overnight at
room temperature. The mixture was diluted with 80 mL of ethyl acetate and washed
with dilute HCl (2 ꢀ 20 mL), water (20 mL), 5% NaHCO3 (2 ꢀ 20 mL), and brine
(20 mL). The organic layer was dried over Na2SO4, filtered and concentrated to dry-
ness. The residue was subjected to flash chromatography using dichloromethane/eth-
1
yl acetate as eluent to afford 190 mg of a white solid (35% yield). H NMR (CDCl3,
250 MHz) d 9.22 (s, 1H), 7.54 (s, 1H), 7.25 (m, 6H), 6.00 (s, 1H), 5.84 (d, J ¼ 7:5 Hz,
1H), 4.99 (s, 2H), 4.65 (br s, 1H), 4.27 (m, 1H), 2.97 (m, 2H), 2.23 (s, 3H), 1.82 (m,
1H), 1.35 (m, 4H), 1.28 (s, 9H).

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
181
2.4. 7-(N-formylmethionyl-Ne-t-butoxycaronyl-
Lys(Boc)-AMC]
L-lysyl)amino-4-methylcoumarin [f-Met-
N-Cbz-Lys(Boc)-AMC (90 mg, 0.17 mmol) was dissolved in a mixture solvent
containing 1:1 (v/v) methanol and ethyl acetate (10 mL) and hydrogenated in the
presence of Pd/C at room temperature for 2.5 h. After the reaction was completed
the catalyst was removed by filtration and the filtrate was concentrated to dryness.
The residue was dissolved in dichloromethane (6 mL), N-formylmethionine (37 mg,
0.21 mmol), HOBt (32 mg, 0.21 mmol), and EDC (40 mg, 0.21 mmol) were sequen-
tially added at 0 °C. The mixture was stirred for 2 h at room temperature. After that
the mixture was diluted with 50 mL of ethyl acetate, washed with dilute HCl
(2 ꢀ 15 mL), water (15 mL), 5% NaHCO3 (2 ꢀ 15 mL), and brine (15 mL). The or-
ganic layer was dried over Na2SO4, filtered and concentrated to dryness. The residue
was purified by silica column chromatography using dichloromethane/ethyl acetate
1
as eluent to give 60 mg of product (64% yield for two steps). H NMR (CDCl3,
250 MHz) d 9.76, (s, 1H), 8.28 (s, 1H), 7.84 (d, J ¼ 7:5 Hz, 1H), 7.68 (d, J ¼ 7:5 Hz,
1H), 7.65 (s, 1H), 7.42 (m, 2H), 6.12 (s, 1H), 4.91 (m, 2H), 4.67 (m, 1H), 3.07 (m,
2H), 2.63 (m, 2H), 2.36 (s, 3H), 2.20–2.03 (m, 2H), 2.05 (s, 3H), 1.95 (m, 1H),
1.83 (m, 1H), 1.45 (m, 4H), 1.39 (s, 9H).
2.5. 7-(N-Formylmethionyl-L-lysyl)amino-4-methylcoumarin (f-Met-Lys-AMC)
f-Met-Lys(Boc)-AMC (50 mg) was dissolved in dichloromethane (2 mL) and tri-
fluoroacetic acid (1 mL) was added. After the mixture was stirred for 2 h, the volatile
substances were removed under vacuum to afford the product in quantitative yield.
The product was neutralized by dissolving in 1:1 (v/v) dichloromethane/isopropyl al-
1
cohol (10 mL) containing 10% triethylamine. H NMR (D6-DMSO, 250 MHz): d
10.45 (s, 1H), 8.37 (m, 2H), 8.04 (br s, 1H), 7.76 (m, 2H), 7.65 (br s, 3H), 7.50
(dd, J ¼ 2:0, 8.7 Hz), 6.27 (s, 1H), 4.44 (m, 2H), 2.79 (m, 2H), 2.46 (m, 2H), 2.40
(s, 3H), 2.04 (s, 3H), 1.85–1.49 (m, 8H). ESI-HRMS: calcd for C22H31N4O5Sþ
(M + Hþ): 463.2009. Found 463.2004.
2.6. 7-(Na-9-fluorenylmethoxycarbonyl-Nc-trityl-
[N-Fmoc-Gln(Trt)-AMC]
L-glutamyl)amino-4-methylcoumarin
To an ice-chilled solution of Fmoc-Gln(Trt)-OH (244 mg, 0.4 mmol) and triethyl-
amine (40 mg, 0.4 mmol) in tetrahydrofuran (7 mL), ethyl chloroformate (48 mg,
0.44 mmol) was added dropwise. The mixture was stirred at 0 °C for 10 min. 7-Ami-
no-4-methylcoumarin (68 mg, 0.4 mmol) was added and the solution was stirred
overnight at room temperature. The solvent was evaporated and the residue was dis-
solved in 50 mL of dichloromethane and washed with 0.5 M HCl (2 ꢀ 20 mL), water
(20 mL), 5% NaHCO3 (2 ꢀ 20 mL), and brine (20 mL). The organic layer was dried
over Mg2SO4; filtered, and concentrated. The crude product was purified by silica
gel chromatography using dichloromethane/ethyl acetate as eluent to give 40 mg of
1
a white solid. H NMR (CDCl3, 400 MHz): d 7.77 (d, J ¼ 5 Hz, 1H), 7.49 (s, 1H),

182
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
7.44 (s, 1H), 7.40–7.27 (m, 26H), 6.97 (s, 1H), 6.18 (s, 1H), 4.41 (br s, 1H), 4.15 (m,
2H), 2.80 (m, 2H), 2.55 (m, 2H), 2.39 (s, 3H).
2.7. 7-(N-Formylmethionyl-L-glutamyl)amino-4-methylcoumarin (f-Met-Gln-AMC)
N-Fmoc-Glutamyl(Trt)-AMC (40 mg, 0.053 mmol) was dissolved in 20% piperi-
dine dichloromethane (5 mL) and stirred at room temperature for 45 min. The reac-
tion was purified by silica gel chromatography using ethyl acetate/dichloromethane
as eluent to give 20 mg white solid. The residue was redissolved in dichloromethane
(7 mL), and N-formylmethionine (8.8 mg, 0.05 mmol), 1-hydroxy-7-azabenzotriazole
(0.037 mmol), and 1-ethyl-3-(-(dimethylamino)propyl)-carbodiimide (EDC) (7.6 mg,
0.038 mmol) were sequentially added at 0 °C. The reaction was stirred overnight at
room temperature and followed by dilution into 40 mL dichloromethane. The mix-
ture was washed with 0.5 M HCl (2 ꢀ 20 mL), water (20 mL), 5% NaHCO3
(2 ꢀ 20 mL) and brine (20 mL). The organic layer was dried over Mg2SO4; filtered,
and concentrated. The residue was purified by silica gel column chromatography us-
ing dichloromethane/ethyl acetate as eluent to give 15 mg of 7-(Na-formylmethionyl-
N-trityl-L-glutamyl)amino-4-methylcoumarin (58% yield). This purified compound
(15 mg) was dissolved in 4.5 mL trifluoroacetic acid, 0.25 mL thioanisole, 0.1 mL ani-
sole, and 0.15 mL ethanedithiol. The mixture was stirred at room temperature for 1 h
and followed with removing the volatile substances with nitrogen gas. The remaining
residue was triturated with ether (3 ꢀ 15 mL) to give 10 mg of white solid (quantita-
tive yield). The product was neutralized by redissolving in 1:1 dichloromethane/iso-
propyl alcohol containing 10% triethylamine (10 mL). The solvent was evaporated
1
by nitrogen gas and the residue was triturated with ether (2 ꢀ 15 mL). H NMR
(DMSO, 400 MHz) d 10.42 (s, 1H), 8.41 (d, J ¼ 4 Hz, 1H), 8.34 (d, J ¼ 4 Hz, 1H),
8.10 (s, 1H), 7.77 (m, 2H), 7.52 (d, J ¼ 4 Hz, 1H), 7.3 (s, 1H), 6.80 (s, 1H), 6.28
(s, 1H), 4.46 (m, 1H), 4.37 (m, 1H), 2.5 (m, 2), 2.4–2.2 (m, 6H), 2.15 (s, 1H), 2.05
(s, 3H), 1.92 (s, 3H). ESI-HRMS: calcd for C21H26N4O6SNaþ (M + Naþ):
485.1465. Found: 485.1465.
2.8. Continuous PDF assay with DPPI
Prior to use, DPPI was activated by the treating with 5 mM dithiothreitol (DTT)
for 30 min in 50 mM Hepes (pH 7.0), 10 mM NaCl. Unless stated otherwise, all as-
says were performed at room temperature (25 °C). Assay reactions were performed
in a quartz microcuvette (total reaction volume ¼ 200 lL) containing 5–150 lM sub-
strate, 50 mM Hepes (pH 7.0), 10 mM NaCl, 5 mM DTT, and 0.1 U of DPPI. Assay
reactions were initiated by the addition of 1–10 lL of Co(II)-PDF enzyme (final con-
centration of 0.5–2.5 nM) and monitored continuously at 360 nm with a Perkin–El-
mer Lambda 20 UV–Vis spectrophotometer. The initial rates were obtained from the
early part of the reaction progression curves (<60 s). To insure that the deformyla-
tion reaction is the rate-limiting, reactions at the lowest and highest substrate con-
centrations were repeated with doubled amount of DPPI (0.2 U). The amount
of background hydrolysis of f-Met-Lys-AMC or f-Met-Gln-AMC by DPPI was

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
183
determined under similar conditions but in the absence of PDF. This background
hydrolysis rate was subtracted from the observed reaction rates.
To determine the catalytic activity of DPPI toward Met-Lys-AMC, a stock solution
of f-Met-Lys-AMC (10 mM) in 50 mM Hepes (pH 7.0), 150 mM NaCl was incubated
overnight with excess Co(II)-PDF to allow complete hydrolysis of the N-formyl
moiety. The PDF was inactivated by heating at 95 °C for 30 min and removed by
centrifugation. DPPI assays were performed by adding 0.005 U of DPPI to a mixture
(200 lL) containing 50 mM Hepes (pH 7.0), 10 mM NaCl, 5 mM DTT, and
0.5–100 lM of Met-Lys-AMC. The reaction progress was monitored at 360 nm. The
catalytic activity of DPPI against Met-Gln-AMC was determined in a similar manner.
2.9. PDF inhibition assay
PDF inhibition assays were performed in a reaction mixture (200 lL) containing
100 lM f-Met-Lys-AMC, 0–200 nM inhibitor, 50 mM Hepes (pH 7.0), 10 mM NaCl,
5 mM DTT, and 0.1 U of DPPI. The background hydrolysis rate was measured by
monitoring the reaction for 30 s at 360 nm on a spectrophotometer. The PDF reac-
tion was then initiated by the addition of Co(II)-PDF (final concentration 4.0 nM)
and monitored continuously for another 60–120 s. Initial rate was calculated from
the early region of the reaction progress curve (0–30 s after addition of PDF) and
corrected by subtracting the background rate. The inhibition constant (KI) was cal-
culated according to the equation:
V ¼ ðVmax ꢀ ½SꢂÞ=ðKMð1 þ ½Iꢂ=KIÞ þ ½SꢂÞ:
PDF assay in the fluorescence mode was similarly performed except that the re-
action was monitored on an Aminco-Bowman Series 2 Luminescence spectrometer.
The excitation and emission wavelengths were set at 380 and 460 nm, respectively.
Activity and inhibition assays with Fe(II)-PDF were carried out in the same man-
ner as for Co(II)-PDF, except that the reaction buffer contained 1 mM tris(carboxy-
ethyl)phosphine instead of 5 mM DTT as the reducing agent [3].
2.10. Slow-binding inhibition of PDF
To determine the KI value, assay reaction (200 lL) contained 60 lM f-Met-Lys-
AMC, 0.15 U DPPI, 50 mM Hepes (pH 7.0), 30 mM NaCl, 5 mM DTT, and 0–
500 nM inhibitor 1. The reaction was initiated by the addition of Co(II)-PDF
(2.0 nM final concentration) and was monitored continuously at 360 nm on a UV–
Vis spectrophotometer. The initial reaction rate was obtained from the early region
of the reaction progress curve (0–60 s), during which the substrate to product conver-
sion was <20%. The obtained initial rates (which were corrected for background hy-
drolysis) were fitted to the Michaelis–Menten equation to produce the KI value. To
determine the KIꢃ value, the assay reaction (1000 lL) containing 50 mM Hepes (pH
7.0), 100 mM NaCl, 4.0 nM PDF, 0–40 nM inhibitor 1, and 100 lg/mL bovine serum
albumin was incubated on ice for 4 h. The mixture was then brought to room tem-
perature and adjusted to 3 mM DTT, and the reaction was initiated by the addition

184
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
of 0.5 U of DPPI and 70 lL of f-Met-Lys-AMC (final 200 lM). The initial reaction
rates were similarly obtained and fitted to the Michaelis–Menten equation to pro-
duce the KIꢃ value. Reactivation of the E ꢄ I* complex was carried out by incubating
a reaction solution (100 lL) containing 50 mM Hepes (pH 7.0), 150 mM NaCl, PDF
(40–160 nM), inhibitor 1 (50–200 nM), and 100 lg/mL BSA for 2 h on ice. The solu-
tion was rapidly diluted into 900 lL of a solution containing 50 mM Hepes (pH 7.0),
150 mM NaCl, 5 mM DTT, 270 lM f-Met-Lys-AMC, and 0.1 U of DPPI and the re-
action progress was monitored continuously at 360 nm.
3. Results and discussion
3.1. Assay design
An ideal assay should involve a highly efficient substrate for both PDF and the
coupling enzyme, and the PDF inhibitors should not significantly inhibit the cou-
pling enzyme. DPPI meets both of these requirements. First, since DPPI is a cysteine
protease and has an entirely different catalytic mechanism from PDF [21], we rea-
soned that the metal-chelating PDF inhibitors should not inhibit DPPI. Second,
DPPI efficiently hydrolyzes a wide variety of dipeptidyl AMC substrates to release
the corresponding dipeptides and the chromophore/fluorophore AMC [22]. Sub-
strate specificity studies of PDF have revealed a strong preference for an L-methio-
nine at the N-terminal position [23–25]. It has broad specificity at the penultimate
position, although amino acids with positively charged side chains are slightly pre-
ferred [25]. On the basis of the above considerations, we designed a dipeptide sub-
strate, f-Met-Lys-AMC (Fig. 2), for the new DPPI coupled assay. The presence of
a polar side chain at the penultimate position also improves the aqueous solubility
of the substrate. Sequential action by PDF and DPPI should release formate, dipep-
tide Met-Lys, and the AMC, which can be monitored by either its absorbance at
360 nm or fluorescence emission at 460 nm. Consistent with a previous report [22],
the hydrolysis of Met-Lys-AMC by DPPI is extremely fast, with a kcat value of
5.7 sꢁ1, KM of 2.8 lM, and a kcat=KM of 2.1 ꢀ 106 Mꢁ1 sꢁ1 at pH 7.0. The low KM value
makes DPPI a particularly suitable coupling enzyme.
3.2. Continuous DPPI coupled assay
Fig. 3A shows the time courses for the hydrolysis of f-Met-Lys-AMC (50 lM) by
various amounts of Co(II)-PDF and DPPI, monitored at 360 nm on a UV–Vis spec-
trophotometer. In the presence of 0.03 U of DPPI and 5.0 nM of PDF, the substrate
is rapidly hydrolyzed and the absorbance increased linearly with time (tracing A).
The initial rate for f-Met-Lys-AMC hydrolysis approximately doubles when the
amount of PDF is doubled (compare tracings A and B). However, when the amount
of DPPI is doubled while keeping the amount of PDF constant, there is no signifi-
cant increase in hydrolysis rate (compare tracings A and C). Further, there is no vis-
ible lag phase at the early regions of the curves. These results indicate that under the

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
185
Fig. 2. Reactions involved in the assay.
Fig. 3. (A) Time courses for the hydrolysis of f-Met-Lys-AMC by PDF. In a final volume of 200 lL (pH
7.0), f-Met-Lys-AMC (80 lM) was incubated with varying amounts of Co(II)-PDF and DPPI. Tracing A,
2 nM PDF and 0.03 U DPPI; tracing B, 4 nM PDF and 0.03 U DPPI; tracing C, 2 nM PDF and 0.06 U
DPPI; and tracing D, 0.03 U DPPI (no PDF). (B) Plot of initial rates against f-Met-Lys-AMC concentra-
tion. The curve was fitted to the data according to the equation, V ¼ kcat ꢀ ½Eꢂ ꢀ ½Sꢂ=ðKM þ ½SꢂÞ, to give a
kcat of 26 ꢅ 4 sꢁ1 and KM of 40 ꢅ 7 lM. (Data presented are the mean ꢅ SD for three independent experi-
ments.)
assay conditions, the PDF reaction is rate limiting. In the absence of PDF, we also
observed a small yet clearly detectable background hydrolysis of the substrate
(tracing D). To determine the origin of this activity, we treated DPPI with p-hydroxy-
a-bromoacetophenone [26], which selectively inhibits cysteine proteases [27] by
alkylating their active-site cysteines, prior to assay with Met-Lys-AMC and f-Met-
Lys-AMC. Similar reduction of activity toward both substrates was observed. This
suggests that the background signal was due to direct hydrolysis of f-Met-Lys-AMC
by DPPI (with f-Met-Lys and AMC as reaction products). Further kinetic analysis
(in the absence of PDF) revealed a kcat value of 0.06 sꢁ1, KM value of 82 lM, and a

186
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
k
cat=KM value of 760 Mꢁ1 sꢁ1 for DPPI hydrolysis of f-Met-Lys-AMC. In an attempt
to minimize the amount of background hydrolysis, we chemically synthesized an-
other substrate, f-Met-Gln-AMC, but found that the commercial DPPI preparation
also cleaved this substrate with similar catalytic efficiency, providing further evidence
that DPPI is responsible for the background signal. Fortunately, this background
reaction can be easily measured and corrected, and therefore does not complicate
the assay in any substantial manner. To correct for this background rate, we typi-
cally monitored the reaction for 30 s prior to the addition of PDF (Fig. 3A); the slope
of the line gives the background hydrolysis rate. After the addition of PDF, the
reaction is monitored for another 60 s to obtain the total hydrolysis rate. Subtraction
of the background rate from the latter gives the PDF reaction rate. The total amount
of substrate consumed during the entire 90 s assay period is well below <20%. The
slightly increased slope of tracing C relative to A reflects the increased amount of
background hydrolysis of f-Met-Lys-AMC as a result of doubling the amount of
DPPI.
f-Met-Lys-AMC is an excellent substrate of PDF and exhibits Michaelis–Menten
kinetics (Fig. 3B). For Co(II)-PDF, the catalytic constants are kcat of 26 ꢅ 4 sꢁ1, KM
of 40 ꢅ 7 lM, and kcat=KM of (6.5 ꢅ 1.1) ꢀ 105 Mꢁ1 sꢁ1 at pH 7.0. A kcat=KM value of
7.5 ꢀ 105 Mꢁ1 sꢁ1 was obtained for the same substrate by the formate dehydrogenase
assay in the end-point format. The catalytic constants toward Fe(II)-PDF under the
same conditions were similarly determined as 52 ꢅ 1 sꢁ1
,
39 ꢅ 2 lM, and
(1.3 ꢅ 0.1) ꢀ 106 Mꢁ1 sꢁ1, respectively. These activities are only slightly lower than that
of f-ML-pNA, the best PDF substrate previously reported (kcat=KM ¼ 1:9 ꢀ 106
M
M
ꢁ1 sꢁ1) [19]. Coupled with the large extinction coefficient of AMC (e ¼ 17; 000
ꢁ1 cmꢁ1), the current method is highly sensitive. Actually, the sensitivity of this assay
is further improved by monitoring the reaction in the fluorescence mode. This is critical
for assaying very potent PDF inhibitors, because a proper assay condition requires that
the enzyme concentration be lower than that of the inhibitor, which should in turn be
similar to the KI value.
3.3. Screening of PDF inhibitors
The utility of the new assay in PDF inhibitor screening was first demonstrated
with BB-3497 (Fig. 1, compound 2), a competitive inhibitor of PDF (IC50 ꢆ7 nM
against Ni(II)-EcPDF) [15]. Varying concentrations of BB-3497 were directly added
into the assay reaction and the reaction progress was monitored continuously at
360 nm on a spectrophotometer or at 460 nm on a spectrofluorimeter (excitation at
380 nm). Fig. 4 shows an example of the reaction progress curves obtained with
Co(II)-PDF, as monitored on a spectrofluorimeter. The fluorescence yield increased
linearly with time and the initial rates were calculated from the slopes of the lines.
Data fitting to the Michaelis–Menten equation gave a KI value of 23 ꢅ 5 nM against
Co(II)-PDF. A KI value of 9.3 ꢅ 1.9 nM was similarly determined for Fe(II)-PDF.
These values are in reasonable agreement with the reported IC50 value (ꢆ7 nM)
[15]. Using the same assay method, compound 3 (Fig. 1) gave a KI value of 17 ꢅ 2 nM
against Co(II)-PDF. Assays in the absorption mode gave very similar KI values for

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
187
Fig. 4. Inhibition of PDF by BB-3497. Reactions were carried out at pH 7.0 using f-Met-Lys-AMC as
substrate (40 lM) and in the presence of indicated amounts of inhibitor. The reaction was initiated by
the addition of 4 nM Co(II)-PDF as the final component and the fluorescence yield at 460 nm (excitation
at 380 nm) was monitored with time (no correction of background rate). Inset, plot of remaining activity
(after correction for background hydrolysis and relative to that in the absence of inhibitor) against inhib-
itor concentration. Data fitting against the equation, V ¼ ðVmax ꢀ ½SꢂÞ=ðKMð1 þ ½Iꢂ=KIÞ þ ½SꢂÞ, gave a KI
value of 23 ꢅ 5 nM.
these two inhibitors. Previous AAP assays with Co(II)-PDF (in the end-point for-
mat) gave KI values of 11 ꢅ 1 and 18 ꢅ 1 nM for compounds 2 and 3, respectively
[8]. We have also used this new assay method to successfully determine the KI values
for a variety of inhibitors against E. coli, Plasmodium falciparum, and human PDF
(to be reported elsewhere). Under the assay conditions, none of the PDF inhibitors
showed any inhibition of DPPI (assayed with Met-Lys-AMC as substrate).
3.4. Slow-binding inhibition of macrocycle 1
Compound 1 exhibited time-dependent inhibition of EcPDF; drastically different
IC50 values were obtained depending on the duration of preincubation of the enzyme
with the inhibitor. In our earlier work [16], only an apparent KI value (0.23–0.67 nM)
was obtained after preincubating PDF with compound 1 for 2 h. Time-dependent en-
zyme inhibition is usually caused by either time-dependent inactivation (e.g., covalent
modification) of the enzyme or slow-binding inhibition (slow conversion of an initial
E ꢄ I complex to a tighter E ꢄ I* complex). Since inhibition by 1 is reversible (vide infra),
slow-binding inhibition must be operative and can be described by equation
KI
k5
E þ I ¢ E ꢄ I ¢ E ꢄ Iꢃ;
k6
where KI is the equilibrium constant for the formation of the initial E ꢄ I complex,
whereas k5 and k6 are the forward and reverse rate constants for the slow inter-
conversion of E ꢄ I and E ꢄ I*, respectively. The overall potency of the inhibitor is
described by the equilibrium constant KIꢃ ¼ KI ꢄ k6=ðk5 þ k6Þ [28]. However, proper
kinetic characterization of this inhibitor proved very challenging using the existing

188
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
assays. Fig. 5A shows an example of PDF reaction progress curves in the absence
(tracings A and B) and the presence of compound 1 (50 nM, tracing C), using FDH
as the coupling enzyme [17,18]. Instead of decreasing PDF reaction rate with time (as
expected from the time-dependent conversion of E ꢄ I into E ꢄ I*), an apparent time-
dependent ‘‘activation’’ was observed. This lag phase is due to slow FDH reaction
during the early stage; as formate accumulates, the FDH reaction rates increases,
eventually reaching the steady state when the rate of FDH reaction equals that of
PDF reaction. The observed lag phase became even more pronounced when com-
pound 1 was added to the assay reaction (tracing C). Addition of large amounts of
FDH did not significantly improve the situation [compare tracings A (reaction
contained 2.5 U/mL FDH) and B (reaction contained 12.5 U/mL FDH)].
Fig. 5B shows the reaction progress curves of the same experiments, but with
DPPI as the coupling enzyme. The addition of compound 1 to the assay reaction re-
sulted in time-dependent inhibition. Importantly, none of the reaction progress
curves had any visible lag phase. Further, when assayed against Met-Lys-AMC,
compound 1–3 did not show any detectable inhibition of DPPI at concentrations
up to 500 nM (data not shown). Data fitting against Michaelis–Menten equation us-
ing the initial reaction rates (0–60 s) and the final steady-state rates (after preincuba-
tion for 4 h) produced KI and KIꢃ values of 96 ꢅ 16 and 0.31 ꢅ 0.14 nM, respectively,
for inhibitor 1. The latter is in excellent agreement with the apparent KI value pre-
viously determined using the FDH and AAP assays (0.23–0.67 nM) [16]. To deter-
mine the reverse rate constant k6, PDF and compound 1 was preincubated for 2 h
and then rapidly diluted into 900 lL of a reaction buffer containing f-MK-AMC.
The time-dependent recovery of PDF activity was monitored with time on a spectro-
photometer (Fig. 6). Data fitting the data against the equation
Fig. 5. (A) Time courses for FDH-coupled reactions. Each reaction (200 lL total volume) contained
50 mM K2HPO4 (pH 7.2), 6.0 nM Co(II)-PDF, and 225 lM f-Met-Lys-AMC. In addition, reactions A
and B contained 0.5 and 2.5 U of FDH, respectively, whereas reaction C contained 0.5 U FDH and
50 nM compound 1. (B) Time courses for DPPI-coupled reactions. Each reaction (200 lL total volume)
contained 50 mM Hepes (pH 7.0), 30 mM NaCl, 5 mM DTT, 3.0 nM PDF, 0.022 U of DPPI, 225 lM
f-Met-Lys-AMC, and indicated amounts of compound 1.

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
189
Fig. 6. Reactivation of inhibited PDF. Co(II)-PDF (80–100 nM) and compound 1 (50–150 nM) were in-
cubated on ice for 2 h. The resulting enzyme (100 lL) was rapidly diluted into a reaction solution (900 lL)
containing 270 lM f-Met-Lys-AMC and 0.1 U of DPPI. (A) 80 nM PDF and 50 compound 1; (B) 80 nM
PDF and 100 nM compound 1; (C) 120 nM PDF and 150 nM compound 1. Curves were fitted to the
t
data according to the equation, Abs360 ¼ vs½t ꢁ ð1 ꢁ ek6 Þ=k6ꢂ, to give a rate constant (k6) of 0.0037 ꢅ
0.0010 minꢁ1
.
Abs360 nm ¼ vs½t ꢁ ð1 ꢁ ek6tÞ=k6ꢂ:
gave the reverse rate constant k6 (0.0037 ꢅ 0.0010 minꢁ1). The forward rate constant,
k5, was calculated from the KI, KIꢃ, and k6 values to be 1.6 ꢅ 0.4 minꢁ1
.
4. Conclusion
In summary, we have developed yet another highly sensitive, convenient, and ra-
pid assay for peptide deformylase. Compared to the previously reported assays, a
major advantage of the current method is that the assay reaction can be carried
out in a continuous fashion in the presence of PDF inhibitors, making it particularly
useful for screening PDF inhibitors in a high-throughput setting and the kinetic
characterization of PDF inhibitors. This feature has allowed us to characterize the
slow-binding properties of macrocyclic PDF inhibitor 1, which had been difficult
to accomplish by the other methods. It should be noted that, because PDF inhibitors
usually do not inhibit the formate dehydrogenase either, the FDH assay [17,18] has
been widely used for PDF inhibitor screening. However, due to the poor specific ac-
tivity of FDH, the FDH assays were typically performed in an end-point fashion. As
such, the FDH assay is not suitable (at least not convenient) for the kinetic evalua-
tion of slow-binding inhibitors. Its lower sensitivity and end-point nature also make
the FDH assay less reliable for determining the KI values of exceptionally potent
PDF inhibitors. A drawback of the current assay is the presence of slight back-
ground hydrolysis of N-formylated substrates by DPPI. This is not a problem for
routine kinetic assays or inhibitor screening involving the wild-type PDF. However,
this background reaction may become significant and thus complicate the assay
results involving catalytically impaired PDF mutants.

190
K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
Considering the strengths and weaknesses of each of the PDF assays, we make the
following recommendations. For routine kinetic assays of wild-type and mutant
PDF in the absence of PDF inhibitors, the AAP assay [19] is most convenient. Both
FDH and DPPI assays may be employed to screen PDF inhibitors. If kinetic char-
acterization of a PDF inhibitor is desired or if the PDF inhibitor is exceptionally po-
tent (low KI value), the DPPI assay is the method of choice. Further, for applications
that require high substrate concentrations, we prefer the DPPI assay in the absor-
bance mode, as high concentration can lead to internal fluorescence quenching
and other complications. However, for determining the KI values of potent PDF in-
hibitors, the fluorescence mode is advantageous, as it is more sensitive allowing the
use of very low enzyme (e.g., 2 nM) and substrate concentrations.
Acknowledgment
This work was supported by a grant from the National Institutes of Health
(AI40575).
References
[1] T. Meinnel, Y. Mechulam, S. Blanquet, Biochimie 75 (1993) 1061–1075.
[2] P.T.R. Rajagopalan, X.C. Yu, D. Pei, J. Am. Chem. Soc. 119 (1997) 12418–12419.
[3] D. Groche, A. Becker, I. Schlichting, W. Kabasch, S. Schultz, A.F.V. Wagner, Biochem. Biophys.
Res. Commun. 246 (1998) 342–346.
[4] P.T.R. Rajagopalan, S. Grimme, D. Pei, Biochemistry 39 (2000) 779–790.
[5] D. Mazel, S. Pochet, P. Marliere, EMBO J. 13 (1994) 914–923.
[6] T. Meinnel, S. Blanquet, J. Bacteriol. 176 (1994) 7387–7390.
[7] P.S. Margolis, C.J. Hackbarth, D.C. Young, W. Wang, D. Chen, Z. Yuan, R. White, J. Trias,
Antimicrob. Agents Chemother. 44 (2001) 1825–1831.
[8] K.T. Nguyen, X. Hu, C. Colton, R. Chakratarti, M.X. Zhu, D. Pei, Biochemistry 42 (2003) 9952–
9958.
[9] D. Pei, Emerg. Ther. Targets 5 (2001) 1–18.
[10] C. Giglione, M. Pierre, T. Meinnel, Peptide deformylase as a target for new generation, broad
spectrum antimicrobial agents, Mol. Microbiol. 36 (2000) 1197–1205.
[11] Z. Yuan, J. Trias, R.J. White, Deformylase as a novel antibacterial target, Drug Discov. Today 6
(2001) 954–961.
[12] K.M. Huntington, T. Yi, Y. Wei, D. Pei, Biochemistry 39 (2000) 4543–4551.
[13] D.Z. Chen, D.V. Patel, C.J. Hackbarth, W. Wang, G. Dreyer, D. Young, P.S. Margolis, C. Wu, Z.-J.
Ni, J. Trias, R. White, Z. Yuan, Biochemistry 39 (2000) 1256–1262.
[14] C. Apfel, D.W. Banner, D. Bur, M. Dietz, T. Hirata, C. Hubschwerlen, H. Locher, M.G.P. Page, W.
Pirson, G. Rosse, J.-L. Specklin, J. Med. Chem. 43 (2000) 2324–2331.
[15] J.M. Clements, R.P. Beckett, A. Brown, G. Catlin, M. Lobell, S. Palan, W. Thomas, M. Whittaker,
S. Wood, S. Salama, P.J. Baker, H.F. Rodgers, V. Barynin, D.W. Rice, M.G. Hunter, Antimicrob.
Agents Chemother. 45 (2001) 563–570.
[16] X. Hu, K.T. Nguyen, C.L.M.J. Verlinde, W.G.J. Hol, D. Pei, Structure-based design of a macrocyclic
inhibitor for peptide deformylase, J. Med. Chem. 46 (2003) 3771–3774.
[17] C. Lazennec, T. Meinnel, Anal. Biochem. 244 (1997) 180–182.
[18] P.T.R. Rajagopalan, A. Datta, D. Pei, Biochemistry 36 (1997) 13910–13918.
[19] Y. Wei, D. Pei, Anal. Biochem. 250 (1997) 29–35.

K.T. Nguyen et al. / Bioorganic Chemistry 32 (2004) 178–191
191
[20] X.-C. Guo, P.T.R. Rajagopalan, D. Pei, Anal. Biochem. 273 (1999) 298–304.
[21] K. Ishidoh, D. Muno, N. Sato, E. Kominami, J. Biol. Chem. 266 (1991) 16312–16317.
[22] T.V. Tran, K.A. Ellis, C.-M. Kam, D. Hudig, J.C. Powers, Arch. Biochem. Biophys. 403 (2002) 160–
170.
[23] J.M. Adams, J. Mol. Biol. 33 (1968) 571–589.
[24] Y.J. Hu, Y. Wei, Y. Zhou, P.T.R. Rajagopalan, D. Pei, Biochemistry 38 (1999) 643–650.
[25] S. Ragusa, P. Mouchet, C. Lazennec, V. Dive, M. Meinnel, J. Mol. Biol. 289 (1999) 1445–1457.
[26] G. Arabaci, X.-C. Guo, K.D. Beebe, K.M. Coggeshall, D. Pei, J. Am. Chem. Soc. 121 (1999) 5085–
5086.
[27] E.T. Kaiser, R.W. Furlanetto, J. Am. Chem. Soc. 92 (1970) 6980–6982.
[28] J.F. Morrison, C.T. Walsh, Adv. Enzymol. 61 (1988) 201–301.