Inhibitors of ADP-Ribosylating Bacterial Toxins Based on Oxacarbenium Ion Character at Their Transition States

Article abstract of DOI:10.1021/ja038159+

The bacterial exotoxins, cholera toxin (CT), pertussis toxin (PT), and diphtheria toxin (DT), interfere with specific host proteins to cause tissue damage for their respective infections. The common toxic mechanism for these agents is mono-ADP-ribosylatio

Full text of DOI:10.1021/ja038159+

Published on Web 04/17/2004
Inhibitors of ADP-Ribosylating Bacterial Toxins Based on
Oxacarbenium Ion Character at Their Transition States
Guo-Chun Zhou,† Sapan L. Parikh,† Peter C. Tyler,‡ Gary B. Evans,‡
Richard H. Furneaux,‡ Olga V. Zubkova,‡ Paul A. Benjes,‡ and Vern L. Schramm*,†
Contribution from the Department of Biochemistry, Albert Einstein College of Medicine,
Bronx, New York 10461, and Carbohydrate Chemistry Team, Industrial Research Ltd.,
Lower Hutt, New Zealand
Received August 27, 2003; E-mail: vern@aecom.yu.edu
Abstract: The bacterial exotoxins, cholera toxin (CT), pertussis toxin (PT), and diphtheria toxin (DT), interfere
with specific host proteins to cause tissue damage for their respective infections. The common toxic
mechanism for these agents is mono-ADP-ribosylation of specific amino acids in GsR, GiR, and eEF-2 proteins,
respectively, by the catalytic A chains of the toxins (CTA, PTA, and DTA). In the absence of acceptor
proteins, these toxins also act as NAD+-N-ribosyl hydrolases. The transition-state structures for NAD+
hydrolysis and ADP-ribosylation reactions have oxacarbenium ion character in the ribose. We designed
and synthesized analogues of NAD+ to resemble their oxacarbenium ion transition states. Inhibitors with
oxacarbenium mimics replacing the NMN-ribosyl group of NAD+ show 200-620-fold increased affinity in
the hydrolytic and N-ribosyl transferase reactions catalyzed by CTA. These analogues are also inhibitors
for the hydrolysis of NAD+ by PTA with Ki values of 24-40 µM, but bind with similar affinity to the NAD+
substrates. Inhibition of the NAD+ hydrolysis and ADP-ribosyl transferase reactions of DTA gave Ki values
from 19 to 48 µM. Catalytic rate enhancements by the bacterial exotoxins are small, and thus transition-
state analogues cannot capture large energies of activation. In the cases of DTA and PTA, analogues
known to resemble the transition states bind with approximately the same affinity as substrates. Transition-
state analogue interrogation of the bacterial toxins indicates that CTA gains catalytic efficiency from modest
transition-state stabilization, but DTA and PTA catalyze ADP-ribosyl transferase reactions more from ground-
state destabilization. pH dependence of inhibitor action indicated that both neutral and cationic forms of
transition-state analogues bind to DTA with similar affinity. The origin of this similarity is proposed to reside
in the cationic nature of NAD+ both as substrate and at the transition state.
and nonactive A2 polypeptide (5.4 kDa).2 PT is also a hexameric
Introduction
protein with a molecular mass of 105 kDa and has an A-B
Cholera toxin (CT), pertussis toxin (PT), and diphtheria toxin
(DT) are exotoxins produced by Vibro cholerae, Bordatella
pertussis, and Corynebacterium diphtheriae and are the caus-
ative agents of the severe diarrhea and dehydration in cholera,
the whooping cough of pertussis, and the mucous membrane
damage in diphtheria. Decreases in childhood immunization and
world health programs make it of interest to develop new
antitoxin technologies. Targeting exotoxins with specific inhibi-
tors is an advantage over cytotoxic antibiotics since bacterial
resistance is not expected to develop with antitoxins.
CT is a hetero-oligomeric toxin composed of a binding
pentamer of B subunits and a loosely associated catalytic A
subunit (27.2 kDa), which enters the cytosol to exert CT’s
toxicity.1 Reduction of the disulfide bonds in the A subunit
yields the catalytically active A1 polypeptide (CTA) (21.8 kDa)
subunit structure.3 The catalytically active subunit is the 28 kDa
A protomer (PTA), which is activated by reduction of its internal
disulfide bond.4 The B protomer consists of five subunits and
is responsible for receptor recognition and passage of the A
protomer through the cell membrane and into the cytosol. DT
is a single-chain, 535-residue (60 kDa) proenzyme. After
selective proteolysis and reduction, two fragments, A and B
(21 kDa and 39 kDa, respectively), are formed.5 Fragment A
(DTA) is the catalytic peptide.6
These bacterial exotoxins have a common mechanism in the
use of NAD+ as an ADP-ribose donor to ADP-ribosylate
specific host proteins (Figure 1).7 CTA ADP-ribosylates the
(2) Mekalanos, J. J.; Collier, R. J.; Romig, W. R. J. Biol. Chem. 1979, 254,
5855-5861.
(3) Tamura, M.; Nagimori, K.; Murai, S.; Yajima, M.; Ito, K.; Katada, T.; Ui,
M.; Ishii, S. Biochemistry 1982, 21, 5516-5522.
(4) (a) Moss, J.; Stanley, S. J.; Burns, D. L.; Hsia, J. A.; Yost, D. A.; Myers,
G. A.; Hewlett, E. L. J. Biol. Chem. 1983, 258, 11879-11882. (b) Locht,
C.; Antoine, R. Biochimie 1995, 77, 333-340.
† Albert Einstein College of Medicine.
‡ Industrial Research Ltd.
(1) (a) Zhang, R.-G.; Westbrook, M. L.; Westbrook, E. M.; Scott, D. L.;
Otwinowski, Z.; Maulik, P. R.; Reed, R. A.; Shipley, G. G. J. Mol. Biol.
1995, 251, 550-562. (b) Zhang, R.-G.; Scott, D. L.; Westbrook, M. L.;
Nance, S.; Spangler, B. D.; Shipley, G. G.; Westbrook, E. M. J. Mol. Biol.
1995, 251, 563-573.
(5) Lory, S.; Collier, R. J. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 267-271.
(6) Lory, S.; Carroll, S. F.; Bernard, P. D.; Collier, R. J. J. Biol. Chem. 1980,
255, 12011-12015.
(7) Van Ness, B. G.; Howard, J. B.; Bodley, J. W. J. Biol. Chem. 1980, 255,
10710-10716.
9
5690
J. AM. CHEM. SOC. 2004, 126, 5690-5698
10.1021/ja038159+ CCC: $27.50 © 2004 American Chemical Society

TS Analogues of ADP-Ribosylating Exotoxins
A R T I C L E S
Figure 1. Route “A” is NAD+ hydrolysis, and route “B” is ADP-ribosylation by bacterial exotoxins. The nucleophile in the CTA reaction is Arg 201 in
sR. In PTA, the nucleophile is the thiolate anion of a Cys four amino acids from the C terminal residue in GiR. In DTA the nucleophile is the imidazole
G
nitrogen of diphthamide, a modified histidine in eEF-2 (see details in the text).
R-subunit of stimulatory GTP binding protein GsR at Arg2018-10
with the inversion of stereochemical configuration in the
N-ribosidic bond. GsR is a dual-function protein that stimulates
adenylate cyclase after it binds GTP. The slow hydrolysis of
its bound GTP through its intrinsic GTPase activity11-13 returns
GsR to its resting state. ADP-ribosylation of GsR results in the
inhibition of GsR’s intrinsic GTPase activity,14 leading to
continuous stimulation of adenylate cyclase to result in high
concentration of intracellular cAMP. This perturbation of the
signaling pathway causes the net efflux of ions and water into
the lumen of the small intestine.15 PTA ADP-ribosylates the
R-subunit of inhibitory GTP binding protein GiR at the cysteine
residue, four amino acids from the carboxyl terminus,16 also
with inversion of configuration at the thio-ribosyl bond. GiR
functions to inhibit adenylate cyclase when activated by GTP
binding and also exbihits the intrinsic GTPase activity.14 ADP-
ribosylation of GiR reduces GTPase activity of GiR and causes
continuous inhibition of the adenylate cyclase, resulting in lower
blood glucose and hypersensitivity to histamine.4b DTA cata-
lyzes ADP-ribosylation of a post-translationally modified his-
tidine (diphthamide) on eukaryotic elongation factor 2 (eEF-2)
and this reaction also demonstrates inversion of ribosyl stere-
ochemistry from â to R for the N-ribosidic bond. ADP-
ribosylated EF-2 is inactive in mediating polypeptide chain
elongation in ribosomal function and results in cell death.17 In
the absence of their physiological target proteins, these toxins
act as NAD+ glycohydrolases to slowly hydrolyze NAD+ to
ADP-ribose and nicotinamide (Figure 1).18
The family of ADP-ribosylating exotoxins is the main cause
of human tissue damage from their respective bacterial infec-
tions. Therefore, they are targets for development of new site-
specific treatment of infectious diseases, including cholera,
pertussis, diphthera and other bacterial infections that act by
ADP-ribosylation exotoxins. Antitoxins are known to confer
resistance in this disease model since diphtheria immunizations
use inactivated exotoxin as the antigen. Efficient inhibitors of
ADP-ribosyl transferases could prevent tissue damage from these
infections. The inhibition of bacterial exotoxins neither inhibits
bacterial growth nor places selective pressure on the bacteria.
Resistant strains would not be expected to develop, and
inhibitors against specific exotoxins could make a permanent
addition to antibiotic therapy.
Transition-state analogues are proposed to be powerful
inhibitors for the enzymes by trapping energy of catalysis as
binding energy.19 One approach to transition-state analogue
design is to establish the nature of the enzymatic transition state
and to synthesize chemically stable analogues with similar
features.20 Transition-state structures can be solved by analysis
of multiple intrinsic kinetic isotope effects using specific isotopic
labels of the normal substrates. This analysis has been ac-
complished for the actions of CTA on NAD+ hydrolysis,21 PTA
on NAD+ hydrolysis,22 ADP-ribosylation of G-protein peptide
Ri3 C2023 and recombinant GiR1 subunits,24 and DTA on NAD+
hydrolysis25 and ADP-ribosylation of eEF-2.26 The studies have
(8) Gill, D. M. Proc. Natl. Acad. Sci. U.S.A. 1975, 72, 267-271.
(9) Gill, D. M. J. Infect. Dis. 1976, 133 (Suppl.), s55-s63.
(10) Moss, J.; Vaughan, M. J. Biol. Chem. 1977, 252, 2455-2457.
(11) Casey, P. J.; Gilman, A. G. J. Biol. Chem. 1988, 263, 2577-2580.
(12) Moss, J.; Vaughan, M. AdV. Enzymol. 1988, 61, 303-379.
(13) (a) Cassel, D.; Selinger, Z. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 3307-
3311. (b) Cassel, D.; Eckstein, F.; Lowe, M.; Selinger, Z. J. Biol. Chem.
1979, 254, 9835-9838. (c) Freissmuth, M.; Gilman, A. G. J. Biol. Chem.
1989, 264, 21907-21914.
(18) (a) Katada, T.; Tamura, M.; Ui, M. Arch. Biochem. Biophys. 1983, 24,
290-298. (b) Kandel, J.; Collier, R. J.; Chung, D. W. J. Biol. Chem. 1974,
249, 2088-2097. (c) Moss, J.; Manganiello, V. C.; Vaughan, M. Proc.
Natl. Acad. Sci. U.S.A. 1976, 73, 4424-4427. (d) Moss, J.; Stanley, S. J.;
Lin, M. C. J. Biol. Chem. 1979, 254, 11993-11996. (e) Galloway, T. S.;
Van Heyningen, S. V. Biochem. J. 1987, 244, 225-230.
(19) Wolfenden, R. Acc. Chem. Res. 1972, 5, 10-18.
(14) Hepler, J. R.; Gilman, A. G. Trends Biochem. Sci. 1992, 17, 383.
(15) (a) Keller, M. T. Pediatr. Infect. Dis. 1986, 5, 5101-5105. (b) Field, M.;
Rao, M. C.; Chang, E. B. N. Engl. J. Med. 1989, 321, 800-806. (c) Field,
M.; Rao, M. C.; Chang, E. B. N. Engl. J. Med. 1989, 321, 879-883.
(16) (a) Katada, T.; Ui, M. Proc. Natl. Acad. Sci. U.S.A. 1982, 74, 3129-3133.
(b) West, R. E.; Jr.; Moss, J.; Vaughan, M.; Liu, T.; Liu, T.-Y. J. Biol.
Chem. 1985, 260, 14428-14430.
(20) Schramm, V. L. Acc. Chem. Res. 2003, 36, 588-596.
(21) Berti, P. J.; Schramm, V. L. J. Am. Chem. Soc. 1997, 119, 12069-12078.
(22) Schreuring, J.; Schramm, V. L. Biochemistry 1997, 36, 4526-4534.
(23) Schreuring, J.; Schramm, V. L. Biochemistry 1997, 36, 8215-8223.
(24) Schreuring, J.; Berti, P. J.; Schramm, V. L. Biochemistry 1998, 37, 2748-
2758.
(25) Berti, P. J.; Blanke, S. R.; Schramm, V. L. J. Am. Chem. Soc. 1997, 119,
12079-12088.
(26) Parikh, S. L.; Schramm, V. L. Biochemistry 2004, 43, 1204-1212.
(17) Wilson, B. A.; Collier, R. J. Curr. Top. Microbiol. Immunol. 1992, 175,
27-41.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5691

A R T I C L E S
Zhou et al.
Figure 2. Transition-state structures for ADP-ribosylating toxins and features common to inhibitors 1 and 2.
vacuo to give a syrupy compound. Chromatography (chloroform/ethyl
acetate/methanol, 5:2:1) of the resulting residue afforded 8 (2.16 g,
77%) as a syrup that was used in the next step without further
purification.
established that transition states for the ADP-ribosylation
reaction all have oxacarbenium ion character in the ribose with
varying degrees of nucleophilic participation by the attacking
nucleophpile (Figure 2).21-26 On the basis of these data,
inhibitors 1 and 2 were synthesized to be ribooxacarbenium ion
mimics of NAD+ analogues (Figure 2). These proposed
analogues of the exotoxin transition states are characterized in
this report.
(3R,4R)-3-[(3-Hydroxy-4-hydroxymethylpyrrolidin-1-yl)methyl]-
benzamide (9). A solution of 8 (2.1 g, mmol) in 7 N NH3 in methanol
(30 mL) was stirred overnight at room temperature and then concen-
trated to dryness. Chromatography (chloroform/ethyl acetate/methanol,
5:2:1 f dichloromethane/methanol/ammonia, 7:2.5:0.5) gave 9 as a
clear syrup (0.9 g, 56%). 1H NMR (CD3OD): δ 7.86 (s, 1H), 7.78 (d,
J ) 1.3 Hz, 1H), 7.52 (d, J ) 2.6 Hz, 1H), 7.41 (t, 1H), 4.04-4.00
(m, 1H), 3.51-3.75 (m, 4H), 2.94 (t, 1H), 2.72 (m, 1H), 2.58 (dd,
1H), 2.36 (m, 1H), 2.22-2.19 (m, 1H). 13C NMR: δ 175.2 (C), 142.8
(C), 138.0 (C), 136.5 (CH), 132.5 (CH), 132.2 (CH), 130.5 (CH), 77.0
(CH), 67.0, 65.9, 63.9, 60.2 (CH2), 54.1 (CH). M/Z calculated for
C13H19N2O3 (MH+): 251.1396. Found: 251.1381.
Experimental Section
Synthetic Procedures. N-tert-Butoxycarbonyl-(3R,4R)-3-acetoxy-
4-acetoxymethylpyrrolidine (5). Acetic anhydride (4.5 mL, 47 mmol)
was added dropwise to a stirred solution of N-tert-butoxycarbonyl-
(3R,4R)-3-hydroxy-4-hydroxymethylpyrrolidine (4) (2.0 g, 9.2 mmol)
in dry pyridine (15 mL) at 0 °C (Scheme 1). After 1 h at 0 °C, the
resulting solution was allowed to warm to room temperature overnight,
diluted with ethyl acetate (250 mL), washed with water (100 mL), dried
(MgSO4), and concentrated in vacuo to afford a syrup. Chromatography
(ethyl acetate/petroleum ether, 1:1) gave 5 (2.5 g, 90%) as a clear syrup
(3R,4R)-3-[(3-Hydroxy-4-methanesulfonylmethylpyrrolidin-1-yl)-
methyl]benzamide (10). Methanesulfonyl chloride (165 mg, 1.44
mmol) was added dropwise to a solution of 9 (300 mg, 1.2 mmol) in
pyridine (2 mL) at -30 °C under argon atmosphere and stirred for 1
h. Acetic anhydride (0.3 mL, 3.1 mmol) was added to the reaction
mixture at -15 °C and stirring was continued for 1 h to afford a
chloroform-soluble acetate of 10. The reaction mixture was diluted with
ice-cold water (20 mL), extracted with chloroform (50 mL), washed
with saturated NaHCO3, dried (MgSO4), and concentrated in vacuo to
afford a syrup. The crude residue was deacetylated in 7 N NH3 in
methanol (10 mL) by being stirred overnight at room temperature and
then concentrated to dryness. Chromatography (chloroform/ethyl
acetate/methanol, 5:2:1 f 5:2:3 f chloroform/ethyl acetate/methanol/
aq ammonia, 5:2:3:0.1) of the resulting residue afforded 10 (257 mg,
1
(Scheme 1). H NMR (CDCl3): δ 4.10 (m, 1H), 4.04 (d, J ) 6.9 Hz,
2H), 3.72 (dd, J ) 5.3 Hz, 1H), 3.59 (m, 1H), 3.42-3.31 (m, 2H),
2.64-2.55 (m, 1H), 2.06 (s, 6H), 1.46 (s, 9H). 13C NMR: δ 171.1
(C), 170.8 (C), 170.4 (C),154.4 (C), 79.9 (CH), (74.7, 74.0) (CH), 63.3
(CH2), 60.4 (CH), (50.8, 50.5) (CH2), (46.9, 46.7) (CH2), (42.6, 41.1)
(CH), 28.6 (CH3), 21.1 (CH3), 20.9 (CH3).
(3R,4R)-3-Acetoxy-4-acetoxymethylpyrrolidine Hydrochloride
(6). A solution of hydrogen chloride in 1,4-dioxane (26 mL of 4.0 M
solution) was added to a solution of 5 (2.2 g, 7.3 mmol) in
dichloromethane (40 mL), and the resulting mixture was stirred for 2
h at room temperature. On completion, the reaction was concentrated
1
65%) as a clear syrup. H NMR (CD3OD): δ 7.89 (s, 1H), 7.81 (d, J
1
in vacuo to afford 6 as the hydrochloride salt (1.7 g, 98%). H NMR
) 1.2 Hz, 1H), 7.54 (d, J ) 5.6 Hz, 1H), 7.43 (t, 1H), 4.26 (m, 2H),
4.13-4.11 (m, 1H), 3.84-3.7 (m, 2H), 3.08 (s, 3H), 3.03 (m, 1H),
2.89 (m, 1H), 2.70 (dd, 1H), 2.52-2.45 (m, 2H). 13C NMR: δ 172.5
(C), 139.4 (C), 135.6 (C), 134.2 (CH), 130.1 (CH), 129.9 (CH), 128.4
(CH), 73.7 (CH), 71.8, 62.9, 60.9, 56.6 (CH2), 50.0 (CH), 37.6 (CH3).
M/Z calcd for C14H21N2O5S (MH+): 329.1171. Found: 329.1160.
Adenosine Bisphosphonate 11. Tetrabutylammonium hydroxide
(40% in water) was added to a solution of methylene diphosphonic
acid (1.14 g, 6.48 mmol) in water (15 mL) until the pH was 10. The
solution was lyophilized, and the residue was dissolved in acetonitrile
and concentrated to dryness. A solution of this material and 2′,3′-O-
isopropylidene-5′-O-toluenesulfonyladenosine (2.5 g, 5.42 mmol) in
dry acetonitrile (10 mL) was stirred at room temperature for 16 h and
then concentrated to dryness. The residue in water was eluted through
(CDCl3): δ 5.13 (m, 1H), 4.13 (m, 2H), 3.69 (m, 1H), 3.62 (m, 1H),
3.40-3.27 (m, 2H), 2.75 (m, 1H), 2.04 (s, 6H). 13C NMR: δ 169.9
(C),169.3 (C), 72.8 (CH), 61.2 (CH2), 48.5 (CH2), 45.1 (CH2), 42.1
(CH), 20.0 (CH3), 19.9 (CH3).
3-Carboxamidobenzaldehyde (7). 3-Cyanobenzaldehyde (3.1 g,
23.6 mmol) in concentrated sulfuric acid (30 mL) was heated for 2 h
at 100 °C. After being cooled to room temperature, the reaction mixture
was diluted with ice-cold water (200 mL), extracted with ethyl acetate
(3 × 100 mL), washed with saturated NaHCO3, dried (MgSO4), and
concentrated in vacuo to afford a white solid. Recrystallization from
1
ethyl acetae afforded 7 as a white solid (3.2 g, 91%). H NMR (CD3-
OD): δ 10.0 (s, 1H), 8.40 (s, 1H), 8.17 (d, J ) 1.4 Hz, 1H), 8.08 (d,
J ) 1.3 Hz, 1H), 7.69 (br s, 1H), 7.66 (t, J ) 1.3 Hz, 1H), 7.44 (br s,
1H).
+
a column of Amberlyst A15 resin (NH4 form), and the eluant was
(3R,4R)-3-[(3-Acetoxy-4-acetoxymethylpyrrolidin-1-yl)methyl]-
benzamide (8). Sodium triacetoxyborohydride (2.2 g, 10.3 mmol) was
added to a stirred solution of 7 (1.7 g, 11.4 mmol) and 6•HCl (1.7 g,
7.2 mmol) in 1,2-dichloroethane (40 mL) and stirred overnight at room
temperature under argon atmosphere. The reaction mixture was then
washed with saturated NaHCO3, dried (MgSO4), and concentrated in
concentrated to dryness. Chromatography (iPrOH/H2O/aq NH3 8:1:1,
then 6:2:2) of the residue on silica gel afforded the ammonium salt 11
(1.94 g, 69%) as a glass. 1H NMR (D2O): δ 8.33 (s, 1H), 8.08 (s, 1H),
6.12 (d, J ) 3.2 Hz, 1H), 5.32 (dd, J ) 3.3, 6.2 Hz, 1H), 5.14 (dd, J
) 2.2, 6.2 Hz, 1H), 4.57 (s, 1H), 4.03 (m, 2H), 2.04 (dt, J ) 2.5, 19.8
Hz, 2H), 1.61 (s, 3H), 1.39 (s, 3H).
9
5692 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

TS Analogues of ADP-Ribosylating Exotoxins
A R T I C L E S
Scheme 1
Compound 12. A solution of tetrabutylammonium hydroxide in
water (40% solution, 2.13 mL, 3 equiv) was added to a solution of 11
(565 mg, 1.1 mmol) in acetonitrile (2 mL) and concentrated in vacuo
to dryness. This tetrabutylammonium salt of 11 was dissolved in
acetonitrile (1 mL) and added to a suspension of 10 (180 mg, 0.548
mmol) in acetonitrile (1 mL). After concentration to dryness, followed
by drying in vacuo, the residue was dissolved in acetonitrile (200 µL),
and the solution was stirred overnight at room temperature and then
concentrated to dryness. The resulting residue was dissolved in water
was used in the next step without further purification. M/Z calcd for
C27H36N7O11P2 (M - H)-: 696.1948. Found: 696.1952.
Compound 2. Formic acid (3 mL, 80%) was added to 12 (90 mg,
0.27 mmol) and stirred at room temperature for 5 h. The reaction
mixture was diluted with water and concentrated in vacuo to dryness.
Chromatography (acetonitrile/water, 4:1 f acetonitrile/water/aq am-
1
monia, 8:1:1 f 3:1:1) afforded 2 as a white solid (56 mg, 66%). H
NMR (D2O): δ 8.36 (s, 1H), 8.08 (s, 1H), 7.73 (s, 1H), 7.62 (d, J )
7.4 Hz, 1H), 7.55 (d, J ) 7.5 Hz, 1H), 7.35 (t, 1H), 5.96 (d, J ) 5.3
Hz, 1H), 4.65 (m, 1H), 4.48 (m, 1H), 4.43-4.40 (m, 2H), 4.24 (m,
2H), 3.99-3.87 (m, 4H), 3.64-3.52 (m, 2H), 2.55 (m, 2H), 2.16-
2.09 (m, 2H). 13C NMR: δ 171.8 (C), 154.9 (C), 152.1 (CH), 149.0
+
(2 mL) and passed through a column of Amberlyst-15, (NH4 form).
The crude product was purified by chromatography (iPrOH/water/aq
ammonia, 4:1:1) to give 12 as a white solid (240 mg, 63%), which
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5693

A R T I C L E S
Zhou et al.
Scheme 2
(C), 140.4 (CH), 134.5 (CH), 133.5, 130.8 (C), 129.8, 129.6, 128.9
(CH), 118.7 (C), 87.4, 84.1, 74.2, 72.3, 70.5 (CH), 64.2, 62.8, 60.4,
54.9, 49.3 (CH2), 46.6 (CH), (28.4, 26.7, 25.0) (CH2). M/Z calcd for
C24H32N7O11P2 (M - H)-: 656.1635. Found: 656.1630.
and nicotinamide spots and quantified by liquid scintillation to
determine the fraction of NAD+ converted to nicotinamide. The
inhibition of inhibitors 1 and 2 were competitive inhibitors with respect
to NAD+ (data are not shown here), and all kinetic data were fit to
equations for competitive inhibition using the Kaleidagraph program
(version 3.52).
Values of Km and kcat were determined by fits to the equation V )
kcat ES/(S + Km), where V is the initial rate, E is the enzyme
concentration, S is the substrate (NAD+ for Km(NAD) and ArgOMe for
Phosphate 13. A solution of 9 (0.06 g, 0.24 mmol) in trimethyl
phosphate (2 mL) was cooled to 0 °C, and phosphoryl chloride (0.088
mL, 0.96 mmol) was added. The solution was stirred at 0 °C for 4 h
and then quenched with water. Excess aq NH3 was added, and the
solution was concentrated to dryness. Chromatography afforded 13 (6
1
mg) as a white solid. H NMR (D2O): δ 7.89 (m, 2H), 7.72 (d, J )
Km(ArgOMe)) concentration, kcat is the catalytic turnover number, and Km
7.8 Hz, 1H), 7.60 (t, J ) 7.6 Hz, 1H), 4.53-4.48 (m, 3H), 3.95-3.82
(m, 2H), 3.78-3.69 (m, 1H), 3.63-3.58 (m, 1H), 3.38-3.29 (m, 2H),
2.61 (s, 1H).
High-resolution mass spectra were recorded on MARINER 5158
electrospray TOF mass spectrometer.
Synthesis of Inhibitor 1. Coupling of 13 and 3 was modified from
literature procedures.27-29 Briefly, 0.6 mg (1.65 µmol) of 13 as the bis-
(ammonium) salt and 3.0 mg (4.2 µmol) of adenosine 5′-monophos-
phomorpholidate 4-morpholine-N,N′-dicyclohexylcarboxamidine salt
(Sigma) were dissolved in 0.5 mL of formamide together with MnCl2
(1.0 mg) and MgSO4 (0.5 mg) (Scheme 2). The reaction mixture was
stirred at room temperature for 2 days.
is the Michaelis constant. The inhibition constant (Ki) values were
determined by the equation V/V0 ) (S + Km)/(S + Km(1 + I/Ki)) or
V/V1 ) (S + Km(1 + I1/Ki))/(S + Km(1 + I/Ki)), where V is the initial
rate with the inhibitor, V0 is the initial rate without the inhibitor, V/V1
is the ratio of the initial rates with two different inhibitor concentrations,
S is the substrate (NAD+) concentration, I is the inhibitor concentration,
I1 is the lowest inhibitor concentration, and Km is the Michaelis constant.
Inhibitors 1 and 2 were tested in assays on cholera toxin A (CTA),
diphtheria toxin A (DTA), and pertussis toxin A (PTA). Inhibition
assays for DTA were for ADP-ribosylation of yeast eEF-2, and those
for CTA were for ADP-ribosylation of ArgOMe. All toxins were tested
for inhibition of NAD+ hydrolysis.
The mixture was lyophilized and reconstituted in 1 mL of water,
and undissolved materials were removed by centrifugation. The aqueous
extract was purified by HPLC (90% of 100 mM ammonium acetate,
pH 5.0 and 10% of 50% aqueous methanol; 2 mL/min; 260 nm; 35
min). The product-containing fractions were lyophilized, and the
ammonium acetate salt was removed by C18 Sep-Pak cartridge (Waters)
eluted with 10 mM triethylammonium bicarbonate (TEAB) according
to the procedure recommended in http://dharmacon.com. The lyophi-
lized product yielded 1 as the bis(triethylammonium) salt (0.38 mg,
Assays for CTA. Cholera toxin A protomer from V. cholerae was
purchased from Sigma as a lyophilized powder containing less than
0.5% of cholera toxin B subunit according to SDS-PAGE. The A
protomer was reconstituted to 1.0 mg/mL (∼36.8 µM) as recommended
by the supplier. The activity was assayed at 30 °C in 200 mM sodium
phosphate, pH 7.5, and 20 mM DTT (dithiothreitol). The A protomer
was preactivated in the buffer for 30 min at 30 °C to reduce the disulfide
bond between the catalytic A1 polypeptide (CTA) and the noncatalytic
A2 polypeptide.2,18c
Determination of Km and kcat of NAD+ Hydrolysis by CTA.
Reaction mixtures of 20 µL contained 3.68 µM of CTA (preactivated)
and variable concentrations of NAD+ (0.712, 2.07, 5.08, 10.1, 20.1,
and 39.1 mM including 4.6, 4.0, 4.1, 3.3, 6.7, and 6.0 µCi/mL of
radioactive-labeled NAD+). Aliquots (4 µL) were taken from the
reaction mixtures at 60, 120, and 180 min and applied to TLC plates.
Determination of Km(NAD) and kcat of ADP-Ribosylation of
Arginine Methyl Ester (ArgOMe) by CTA. Reaction mixtures of 20
µL contained 0.92 µM of CTA (preactivated), 75 mM ArgOMe, and
variable concentrations of NAD+ (0.994, 2.07, 2.81, 10.1, 20.1, and
37.6 mM including 4.5, 4.0, 4.5, 3.3, 6.7, and 6.3 µCi/mL of radioactive-
labeled NAD+). Aliquots (4 µL) were taken from the reaction mixtures
at 90 and 120 min and applied to TLC plates.
1
26.8%). H NMR (D2O): δ 1.31 (t, 18H, Et3N), 2.15 (m, 1H), 3.10
(m, 2H), 2.23 (q, 12H, Et3N), 3.90-4.52 (m, 10H), 6.07 (d, 1H, J )
5.85 Hz), 7.47 (m, 1H), 7.56 (m, 1H), 7.77 (m, 2H), 8.21 (s, 1H), 8.45
(s, 1H) ppm. 31P NMR (D2O): δ -9.87, -9.95 ppm. MS (ESI): 660
(M+ + H), 761 (M+ + H + Et3N), 1319 (2M+ + H).
General Assay Procedure. The assays are based on the enzymatic
reactions of NAD+ hydrolysis and ADP-ribosylation (Figure 1). All
assays were initiated by the addition of NAD+, including carbonyl-14C
radioactive labeled NAD+. Aliquots (4 µL) were taken from the reaction
mixtures at designated times and loaded onto DEAE cellulose TLC
plates (EM Science) and developed in 70/30 of 95% EtOH/1 M
ammonium acetate, pH 7.5.18c The Rf of nicotinamide is 0.9, and that
of NAD+ is 0.2. TLC plates were cut into pieces containing the NAD+
Determination of Km(ArgOMe) for ADP-Ribosylation of Arginine
Methyl Ester (ArgOMe) by CTA. Reaction mixtures of 20 µL
contained 0.92 µM of CTA (preactivated), 2.38 mM NAD+ containing
3.8 µCi/mL of radioactive-labeled NAD+, and variable concentrations
of ArgOMe (2.0, 10.0, 30.0, 50.0, 75.0, and 100.0 mM). Aliquots (4
(27) Schimazu, M.; Shinozuka, K.; Sawai, H. Tetrahedron Lett. 1990, 31, 235-
238.
(28) Lee, J.; Churchil, H.; Choi, W.-B.; Lynch, J. E.; Roberts, F. E.; Volante,
R. P.; Reider, P. J. Chem. Commun. 1999, 729-730.
(29) Kennedy, K. J.; Bressi, J. C.; Gelb, M. H. Bioorg. Med. Chem. Lett. 2001,
11, 95-98.
9
5694 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

TS Analogues of ADP-Ribosylating Exotoxins
A R T I C L E S
Table 1. Kinetic Constants for NAD+ and Inhibitors 1 and 2
inhibitor 1
Km/K
inhibitor 2
toxin
reaction
substrate
Km (µM)
kcat (min-1
)
K (µM)
i
KT(µM)
Km/K
i
i
i
CTA
NAD+ hydrolysis
ADP-ribosylation
NAD+ hydrolysis
ADP-ribosylation
ADP-ribosylation
NAD+ hydrolysis
ADP-ribosylation
NAD+, H2O
10800.0
4700.0
8.7
20.0
0.66
17.4
10.9
24.4
620
431
0.78
32.7
23.6
39.7
330
200
0.48
NAD+, ArgOMe
NAD+, H2O
PTA
DTA
19.2
NAD+, peptide Ri3 C20
NAD+, Cys-GiR
27.023
3.023
40.024
0.1126
182.026
23.024
NAD+, H2O
85.026
48.2
30.5
1.8
0.2
32.7
19.1
2.6
0.3
NAD+, diphthamide-eEF2
6.0 for NAD+, 0.66 for eEF226
µL) were taken from the reaction mixtures at 30 and 60 min and applied
on TLC plates.
µM of eEF-2, and 15.4 µM of NAD+ (4.8 µCi/mL of radioactive-labeled
NAD+). Aliquots (4 µL) were taken from the reaction mixtures at 6,
10, and 15 min for 1 or 4, 7, and 12 min for 2 and applied to TLC
plates.
Inhibition of CTA on NAD+ by 1 and 2 for NAD+ Hydrolysis.
To the activated CTA (2.76 µM) were added variable concentrations
of inhibitors (3.67, 14.68, 73.4, and 183.5 µM of 1 or 5, 20, 80, 150,
300, and 500 µM of 2) and 2.07 mM of NAD+ (3.8 µCi/mL of
radioactive-labeled NAD+) in a final volume of 20 µL. Aliquots (4
µL) were taken from the reaction mixtures at 120 and 180 min and
applied to TLC plates. Initial rates for these reaction conditions were
maintained for 240 min.
Assays for pH dependence of kinetic parameters consisted of 200
µL reaction mixtures of 100 mM Tris-HCl (pH 7.0 to 9.0), 1 mM
EDTA, 20 mM DTT, and 0.8 nM of diphtheria toxin. Assays contained
200 000 cpm [32P]NAD+ to follow ADP ribosylation. The kcat and Km
for eEF-2 were determined at a fixed, saturating concentration of NAD+
(825 µM) and by varying the concentration of eEF-2 (0.5 to 25 µM).
kcat and Km for NAD+ were determined at a fixed, saturating concentra-
tion of eEF-2 (10 µM) and by varying the concentration of NAD+ (1
to 100 µM). Aliquots (30 µL) were removed at 0, 1, 2, and 4 minutes
and pipetted into 50 µL of 5 mM unlabeled NAD+ and precipitated by
adding 80 µL 30% TCA. The precipitate was washed twice with 500
µL 5% TCA solution, followed by 500 µL of acetone and dissolved in
1 mL of 10% SDS in 200 mM Tris-HCl, pH 8.0. The sample was
mixed with 10 mL of scintillation fluid and analyzed by scintillation
counting to quantitate radioactivity in the ADP-ribosylated product.
The kinetic parameters for ADP-ribosyltransferase activity were
analyzed by fitting data to the Michaelis-Menten equation from assays
preformed in triplicate in at least three independent experiments.
Inhibition assays included 500 pM of DTA, 10 µM eEF-2, 15 µM NAD,
and variable concentration of inhibitor (5, 20, 40, 120, 240, and 480
µM of 2). Aliquots were taken from reaction mixtures at 4, 8, and 12
min for 2 and analyzed as described above.
Inhibition of CTA by 1 and 2 for ADP-Ribosylation of ArgOMe.
Initial rate assays (20 µL) included 0.92 µM of activated CTA, variable
concentrations of inhibitors (3.67, 14.68, 73.4, and 183.5 µM of 1 or
5, 20, 80, 150, 300, and 500 µM of 2), 75 mM of ArgOMe, and 2.36
mM of NAD+ (4.0 µCi/mL of radioactive-labeled NAD+). Aliquots (4
µL) were taken from the reaction mixtures at 90 and 120 min and
applied on TLC plates. Reaction rates were linear for over 120 min.
Assays for PTA. Pertussis toxin A protomer (PTA, produced by B.
pertussis) was purchased from List Biological Laboratories, Inc. It was
reconstituted to 0.71 µM as recommended by the supplier. The assays
were conducted at 37 °C in 100 mM Tris‚HCl, pH 8.0, 1 mM EDTA,
and 20 mM DTT. The A protomer (PTA) was preactivated in the buffer
at room temperature for 30 min.22-24
Determination of Km and kcat for NAD+ Hydrolysis by CTA.
Reaction mixtures of 20 µL contained 0.143 µM of PTA (preactivated)
and variable concentrations of NAD+ (10.0, 20.0, 50.0, 92.6, 183.0,
and 319.0 µM including 0.54, 1.08, 2.7, 5.0, 4.5, and 3.75 µCi/mL of
radioactive-labeled NAD+). Aliquots (4 µL) were taken from the
reaction mixtures at 120 and 180 min and applied on TLC plates.
Hydrolysis of NAD+ by PTA and Inhibition by 1 and 2. Assays
were conducted in reaction mixtures (30 µL) of 0.143 µM activated
PTA, variable concentrations of inhibitors (3.67, 14.68, 73.4, and 183.5
µM of 1 or 5, 20, 80, 150, 300, and 500 µM of 2), and 50 µM of
NAD+ (2.7 µCi/mL of radioactive-labeled NAD+). Aliquots (4 µL)
were taken from the reaction mixtures after incubation for 3 h and
applied to TLC plates.
Results
Inhibition of CTA by 1 and 2. Hydrolysis of NAD+ by CTA
gave a Km for NAD+ of 10.8 ( 0.8 mM, a kcat of 8.7 ( 0.3
min-1, and kcat/Km ) 806 M-1 min-1 (Table 1). Inhibition of
this reaction by 1 and 2 gave Ki values of 17.4 ( 2.5 µM and
32.7 ( 3.3 µM (Table 1).
ADP-ribosylation of ArgOMe by CTA gave a Km for NAD+
of 4.7 ( 0.2 mM, a kcat of 20 ( 0.3 min-1, a kcat/Km ) 4255
M-1 min-1, and a Km for ArgOMe of 114 ( 12 mM (Table 1).
Inhibitors 1 and 2 gave Ki values of 10.9 ( 1.4 µM and 23.6 (
2.0 µM (Table 1) for this reaction. Both inhibitors 1 and 2 bind
to CTA with high affinity relative to NAD+.
Assays for DTA. Recombinant DTA was overexpressed and purified
to homogeneity.25 eEF-2 is the acceptor of ADP-ribosylation by DTA
and was purified from Saccharomyces cereVisiae according to reported
procedures.30 The activity was assayed at 37 °C in 100 mM Tris‚HCl,
pH 8.0, 1 mM EDTA, and 20 mM DTT.
Hydrolysis of NAD+ and Inhibition by 1 and 2. Assays were
conducted in reaction mixtures (30 µL) of 6 µM of recombinant DTA,
variable concentrations of inhibitors (14.68, 73.4, 183.5, and 367.0 µM
of 1 or 20, 80, 150, 300, and 500 µM of 2), and 0.33 mM of NAD+
(4.8 µCi/mL of radioactive-labeled NAD+). Aliquots (4 µL) were taken
from the reaction mixture after being incubated for 240 min and were
applied to TLC plates.
ADP-Ribosylation of eEF-2 by DTA and Inhibition by 1 and 2.
Initial rate assays (30 µL) included 670 pM (for 1) or 500 pM (for 2)
of recombinant DTA, variable concentrations of inhibitors (3.67, 14.68,
73.4, and 183.5 µM of 1 or 5, 20, 80, 150, 300, and 500 µM of 2), 6.7
Inhibition of PTA by 1 and 2. The hydrolysis of NAD+ by
PTA gave a Km value of 19.2 µM and a kcat value of 0.66 min-1
(Table 1). Inhibition of PTA by 1 and 2 gave values for Ki of
24.4 ( 1.3 µM and 39.7 ( 2.1 µM for NAD+ hydrolysis (Table
1). Therefore, inhibitors 1 and 2 bind to PTA with affinities
similar to the substrate NAD+.
Inhibition of DTA by 1 and 2. DTA hydrolyzes NAD+ with
kinetic constants of Km ) 85.0 ( 13.0 µM and kcat ) 0.11 (
0.04 min-1 26 1 and 2 inhibit DTA on NAD+ hydrolysis with
.
Ki values of 48.2 ( 1.3 µM and 32.7 ( 3.5 µM (Table 1). 1
and 2 bind slightly better than NAD+ to DTA for the NAD+
hydrolysis reaction.
(30) Jørgensen, R.; Carr-Schmid, A.; Ortiz, P. A.; Kinzy, T. G.; Andersen, G.
R. Acta Crystallogr. 2002, D58, 712-715.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5695

A R T I C L E S
Zhou et al.
Figure 3. Ionization of inhibitors 2 and 13 to form the cationic transition-state analogues.
Figure 4. pKa plots for inhibitors 2 and 13. The chemical shift of 2′-proton resonance was followed and referenced to TSP (3-(trimethylsilyl)propionic acid)
during pH titration. Titrations were in 200 mM sodium phosphate including 5% D2O, and the data were determined from the best fit to the equation δ )
δHA - [( δHA - δA)/(1 + 10(pKa - pH))],36 where δ is the observed chemical shift value and δHA and δA are the chemical shift values for a given proton at
the low- and high-pH limits. (a) pKa is 8.40 for 2. (b) pKa is 8.50 for 13.
Table 2. pH Dependence of DTA on NAD+ Hydrolysisa
DTA catalyzes the ADP-ribosylation of eEF-2 with kinetic
pH
7.0
7.5
8.0
8.5
9.0
constants of Km(NAD) ) 6.0 ( 2.0 µM and kcat ) 182 ( 15
min-1 26
.
Inhibition of DTA by 1 and 2 gave Ki values of 30.5
Km (µM)b 8.3 ( 2.0 8.5 ( 2.6 6.0 ( 2.0 5.9 ( 2.0 8.9 ( 1.7
k
cat (s-1
)
3.5 ( 0.4 3.1 ( 0.3 3.0 ( 0.3 3.1 ( 0.3 3.5 ( 0.4
( 3.3 µM and 19.1 ( 3.3 µM (Table 1) for the ADP-
ribosyaltion of EF-2 by DTA. Inhibitors 1 and 2 bind to DTA
less well than NAD+ for the ADP-ribosylation of eEF-2 by
DTA.
Ki (µM)d
Ki (µM)a
Ki (µM)b
29 ( 2
27 ( 2
1.1
21 ( 3
19 ( 3
2.3
19 ( 3
14 ( 2
5.4
12 ( 2
6.0 ( 0.8 3.4 ( 0.4
6.7 13.6
17 ( 2
a
Ki for the cation of 2, corrected by [I]protonated ) p*[I]total, where p )
pKa for 2 and 13 and pH Dependence of Kinetic Constants
for DTA. The pKa values for inhibitors 2 and 13 were
determined by pH titration using the 2′-proton chemical shift
(Figures 3 and 4). The pKa values were readily established to
be 8.40 for 2 and 8.50 for 13. The pKa for 1 can also be assigned
as 8.40 since the chemical environment of N1′ is similar to that
of 2. The pH dependence for the ADP-ribosylation of eEF-2
by DTA demonstrated that the Km values do not vary signifi-
cantly from pH 7.0 to 9.0 (Table 2). Likewise, kcat is independent
of pH over this range. Since the primary constants are unchanged
over a pH range that encompasses the pKa of the inhibitor,
altered pH can be used to test the relative inhibitory potential
of the N1′-neutral and N1′-cationic inhibitors. At pH 7.0, the
observed Ki is 29 µM, and the inhibitor is >90% cationic to
give a calculated Ki for the cation of 27 µM, assuming the
neutral form does not bind (Table 2). At pH 9.0, the Ki is 17
µM, to give a Ki of 3.4 µM if only the cation binds. Making
the assumption that only N1′-neutral 2 binds gives calculated
Ki values of 1.1 and 13.6 µM at pH 7.0 and 9.0, respectively.
1/(1 + 10(ph-pK )). Ki assuming only the neutral form of 2 binds to DTA.
b
a
These constants suggest that both neutral and cationic forms
bind with similar dissociation constants. However, steady-state
kinetic analysis cannot prove the ionic state of bound forms.
Proof awaits the availability of labeled enzyme and/or inhibitor
for enhanced NMR sensitivity.
Discussion
Inhibitor Design. NMR evidence has demonstrated that the
reactions of ADP-ribosylation by CTA, PTA, and DTA give
products characteristic of SN2 reactions with the inversion of
configuration at the 1′-anomeric carbon.31-33 However, KIE
studies and transition-state analysis demonstrate that the transi-
tion states for NAD+ hydrolysis by CTA, PTA, and DTA are
(31) Oppenheimer, N. J. J. Biol. Chem. 1978, 253, 4907-4910.
(32) Scheuring, J.; Schramm, V. L. J. Am. Chem. Soc. 1995, 117, 12653-12654.
(33) Oppenheimer, N. J.; Bodley, J. W. J. Biol. Chem. 1981, 256, 8579-8581.
9
5696 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

TS Analogues of ADP-Ribosylating Exotoxins
A R T I C L E S
highly dissociative.21,22,25 The transition state for ADP-ribosy-
lation of G-protein peptide Ri3 C20 and recombinant GiR1
µM)23 and recombinant GiR1 subunits by PTA (23 µM).24
Therefore, the interaction between NAD+ and PTA is not
strongly influenced by the nature of the nucleophile. Bond order
to the water nucleophile is low at the transition state of NAD+
hydrolysis by PTA,22 and the major character in this transition
state resides in the NAD+ structure. The inhibitors 1 and 2 show
similar binding constants (Ki ) 24.4 µM for 1 and 39.7 µM for
2) to NAD+ (Table 1).
The catalytic rate of PTA for NAD+ hydrolysis is kcat ) 0.01
s-1, and this increases to 0.67 s-1 in the presence of saturating
GiR.24 Here, the kenz/kchem rate enhancement is modest for NAD+
hydrolysis. Similar to DTA, inhibitiors 1 and 2 bind to PTA
with affinity not significantly different from the Michaelis
constant for NAD+. This pattern is consistent with enzyme-
based substrate destabilization, where features of the transition
state will not necessarily increase binding. Catalytic evidence
for substrate destabilization by CTA, DTA, and PTA is supplied
by the common observation that the ADP-ribosylating toxins
all catalyze the slow hydrolysis of NAD+ in the absence of their
nucleophilic ADP-ribose acceptors. Catalysis of NAD+ hy-
drolysis requires that binding in the Michaelis complex desta-
bilizes bound NAD+ toward the transition state, facilitating the
attack of the solvent water nucleophile. NAD+ chemical
destabilization is readily accomplished by the hydrophobic
interactions with the nicotinamide ring discussed below.
Inhibitor Action on DTA. The transition states for the NAD+
hydrolytic reactions of CTA and DTA are closely related, with
Pauling bond orders of approximately 0.02 to the nicotinamide
leaving group and 0.005 to the attacking water nucleophile.21,22
However, their kinetic constants for Km and kcat are quite
different (Km for CTA/Km for DTA ) 125 and kcat for CTA/kcat
for DTA ) 75). Catalytic efficiency for NAD+ hydrolysis is
low for both catalysts, but the kcat value of 0.15 s-1 for CTA
greatly exceeds the rate of 0.002 s-1 for DTA. These slow rates
place modest limits on the theoretical affinity of transition-state
analogues. Wolfendens’ construct for transition-state affinity
assumes limits of (substrate Kd) × (kchem/kenz), where kchem and
kenz are nonenzymatic and enzymatic solvolysis rates, respec-
tively, and Kd is the dissociation constant for substrate.19 At
pH 7.5, the kchem for NAD+ under our experimental conditions
is approximately 10-6 s-1. Thus, rate enhancements are 1.5 ×
105 for CTA but only 2 × 103 for DTA. The transition-state
analogues capture a significant fraction of the 1.5 × 105 rate
enhancement energy from CTA by binding up to 620-fold more
tightly than NAD+. However, in the case of DTA, where rate
enhancement is only 2 × 103, there is insufficient recognition
of the transition-state features to enhance binding. This result
illustrates an inherent property of transition-state analogue
design. Poor catalysts can capture only relatively small amounts
of binding energy from molecules that resemble the poorly
stabilized on-enzyme transition states.
23
subunits24 have more nucleophilic character at the transition
state, but their transition states are also dominated by riboox-
acarbenium character. Likewise, the transition state for the ADP-
ribosylation of eEF-2 by DTA has increased nucleophilic
participation but is dominated by its cationic character.26 Despite
their differences,20 oxacarbenium ion character (Figure 2) is
associated with all of these transition states. Inhibitors 1 and 2
were designed on the basis of the common ribooxacarbenium
ion character for these transition-state structures. Recently,
similar chemistry has been used to produce transition-state
analogues of purine nucleoside phosphorylase and 5′-methylth-
ioadenosine phosphorylase, N-ribosyltransferases with oxacar-
benium character at their transition states.34,35 In these cases,
inhibitors bind 105 to 106 times tighter than substrates.
Comparison of Inhibitors. Compound 13 was tested as an
exotoxin inhibitor but did not show inhibitory activity to CTA,
PTA, and DTA when tested to 1000 µM (data are not shown).
The ADP portion of NAD+ is clearly important for enzyme
recognition. Kinetic analysis showed that the inhibitory activity
of 1 and 2 is similar for all three toxins (Table 1). Therefore,
the ADP group is essential but the bridge oxygen of the
pyrophosphate is not important for inhibitor recognition at any
of these catalytic sites. This result is useful for the development
of antitoxins, since the methylene bridge in the pyrophosphate
group produces biological stability against phosphodiesterases.
Transition-State Analogue Binding to CTA. Transition-
state analysis from KIE indicated that NAD+ at the transition
state of CTA is almost a fully developed ribooxacarbenium ion
with bond orders of 0.02 to the nicotinamide and 0.005 to the
attacking water nucleophile.21 This expanded oxacarbenium
transition state is expected to closely resemble the electrostatics
of 1 and 2. The inhibitory results demonstrate that 1 and 2 bind
to CTA more strongly than NAD+, consistent with the capture
of transition-state energy.
The inhibitors 1 and 2 also have good inhibitory activity with
CTA for ADP-ribosylation of ArgOMe, suggesting that the
transition state of CTA on ADP-ribosylation of ArgOMe is also
dissociative even though the attacking nucleophile is present at
the catalytic site. However, the weak binding and poor nucleo-
philicity of ArgOMe is not expected to induce the same degree
of nucleophilic participation as the more tightly bound GsR
nucleophiles. Nucleophilic attack lags behind transition-state
formation, and the enzyme directs the stereochemistry of the
reaction center (1′-C) during departure of the leaving group until
the nucleophilic attack is complete. If CTA resembles DTA and
PTA in its transition-state characteristics, it would be expected
to be more nucleophilic in ADP-ribosyl transfer reactions than
in NAD+ hydrolysis. The nearly equivalent inhibition of CTA
by 1 and 2 for hydrolytic and ArgOMe ADP-ribosyl transfer
reactions supports the proposal that substantial oxacarbenium
character is present at the transition state for this small-molecule
acceptor of ADP-ribose.
The similar binding of NAD+ and inhibitors 1 and 2 to both
DTA and PTA suggest that the catalytic rate enhancement is
caused by ground-state destabilization where bound NAD+ is
more reactive to solvent water in the bound than in the free
state. Ground-state destabilization is easily accomplished simply
by placing the nicotinamide leaving group into a hydrophobic
environment. The partial positive charge on the nicotinamide
group is shifted toward the ribosyl ring to make it more
Inhibitor Action on PTA. The Km values for NAD+ binding
do not change much from NAD+ hydrolysis by PTA (19.2 µM)
to ADP-ribosylation of G-protein peptide Ri3 C20 by PTA (27
(34) Lewandowicz, A.; Tyler, P. C.; Evans, G. B.; Furneaux, R. H.; Schramm,
V. L. J. Biol. Chem. 2003, 278, 314654-31468.
(35) Dyson, H. J.; Tennant, L. L.; Holmgren, A. Biochemistry 1991, 30, 4262-
4268.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5697

A R T I C L E S
Zhou et al.
oxacarbenium-like. A hydrophobic pocket that surrounds the
nicotinamide group is known to be a feature of NAD+ binding
to DTA and PTA.24,25 The addition of eEF-2 increases the
reaction rate of DTA 1600-fold, but inhibitors 1 and 2 show no
improvement in affinity, also consistent with ground-state
destabilization as the primary mechanism for enhancing the rate
of catalysis. Inhibitors 1 and 2 do not include the contribution
from the attacking nucleophile, and bindings of 1 and 2 to DTA
on NAD+ hydrolysis are only slightly better than NAD+ binding
to DTA on NAD+ hydrolysis (Table 1).
From NAD+ hydrolysis by DTA (Km (NAD) ) 85 µM) to ADP-
ribosylation of eEF-2 by DTA (Km(NAD) ) 6.0 ( 2.0 µM, kcat
) 182 ( 15 min-1),26 NAD+ binding to DTA becomes stronger.
However, the most strongly bound component of the ternary
complex is eEF-2 (Km(EF-2) ) 0.66 ( 0.06 µM)26 rather than
NAD+. The attacking nucleophile in the ADP-ribosylation of
eEF-2 is diphthamide, and here the nucleophilic bond order
increases from 0.005 for water (in the hydrolytic reaction) to
0.03 (in the eEF-2 reaction). The increased nucleophile partici-
pation is responsible for the 1600-fold rate acceleration relative
to NAD+ hydrolysis.
pH Profile for ADP-Ribosylation of eEF-2 by DTA. The
pyrrolidine N pKa values of 2 and 13 are 8.40 and 8.50,
respectively (Figure 4). To establish the ionic form binding to
DTA, the pH dependence of kinetic parameters was measured.
NAD+ is unlike other substrates for N-ribosyltransferases in
being cationic at N1 in the substrate and being cationic at C1′
at the transition state (Figure 2). Electron migration to the
leaving group nicotinamide causes the cationic center to shift
from N1 of the leaving group to C1′ of the NMN-ribosyl group.
Transition-state mimicry of 1 and 2 relies on the cationic nature
of the transition state, but the distance of cation migration is
only about 1.5 Å on conversion of substrate NAD+ to NAD+
at the transition state.24,25 The Ki values measured for 2 in the
ADP-ribosylation of eEF-2 by DTA show little change over
2.0 pH units, conditions where the N-cation varies from 96 to
20% of the inhibitor species. The Km for NAD+ also shows
little pH dependence. The simplest explanation is that both
neutral and cationic forms bind to the enzyme with similar Ki
values of approximately 20 µM. If it is assumed that only the
cationic or neutral forms bind, the Ki values would vary
considerably across this pH range (Table 2). However, this
kinetic argument cannot prove the ionic state of bound 2. Direct
measurement requires isotopically labeled inhibitor or enzyme
or both.
Conclusions
Analogues of NAD+ at the transition states for the ADP-
ribosylating bacterial exotoxins bind more tightly to CTA, but
not to DTA or PTA. CTA binds weakly to NAD+ and gains
catalytic rate enhancement from transition-state stabilization.
DTA and PTA bind more tightly to NAD+ as substrate and
catalyze their respective reactions by ground-state destabilization
of the NAD+, making it more reactive to nearby nucleophiles
provided by the physiological substrates, eEF-2 and GiR, as well
as water nucleophiles. Since much of the catalytic rate enhance-
ment comes from proximity of the ADP-ribosyl acceptor,
transition-state analogues with only NAD+ features bind poorly.
Improved binding might be expected for analogues containing
features of NAD+ at the transition state, along with features of
the nucleophilic group.
Acknowledgment. This work was supported by NIH Re-
search Grant AI34342 and the New Zealand Foundation for
Research, Science & Technology.
JA038159+
9
5698 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004