Transition-state interactions revealed in purine nucleoside phosphorylase by binding isotope effects

Article abstract of DOI:10.1021/ja7104398

The binding of [5′-³H]inosine to human purine nucleoside phosphorylase results in an equilibrium binding isotope effect (BIE) of 1.5%, and transition state formation causes an intrinsic KIE of 4.7%. These values reflect atomic vibrational disto

Full text of DOI:10.1021/ja7104398

Published on Web 01/30/2008
Transition-State Interactions Revealed in Purine Nucleoside Phosphorylase
by Binding Isotope Effects
Andrew S. Murkin,† Peter C. Tyler,‡ and Vern L. Schramm*,†
Department of Biochemistry, Albert Einstein College of Medicine, Bronx, New York, and Carbohydrate Chemistry
Team, Industrial Research Ltd., Lower Hutt, New Zealand
Received November 19, 2007; E-mail: vern@aecom.yu.edu
Enzymatic transition-state (TS) analogues capture the lowered
activation energy for catalysis as binding energy and TS formation
has been explained in both dynamic and thermodynamic terms.1,2
Thermodynamic interpretation of catalysis and TS analogue binding
proposes tight binding of reactants at the TS. This binding is
converted directly into binding energy of TS analogues. The
dynamic interpretation of catalysis proposes formation of the TS
by a coordinated dynamic excursion of the protein that does not
require tight binding at the TS.2 Tight binding of TS analogues in
the dynamic model of TS formation involves conversion of the
dynamic TS excursion into a static, stable structure organized around
the chemically stable TS analogue. Here an experimental test of
these hypotheses is achieved for human purine nucleoside phos-
phorylase (PNP) through a comparison of chemically intrinsic
kinetic isotope effects (KIEs) from TS formation3 and binding
isotope effects (BIEs) for TS analogue interactions.4
Isotope effects, structural studies, and computational analysis of
PNP have established a ribooxacarbenium TS with reaction
coordinate motion from His257 causing electron repulsion from
the ribosyl oxygen to the hypoxanthine leaving group (Figure 1).3,5,6
Immobilization of the 5′-hydroxyl group in the Michaelis complex
generates a 1.5% BIE from [5′-3H]inosine with an additional 4.7%
intrinsic chemical KIE arising from TS distortion and electronics.
Mutation of His257 alters the kinetics, inhibition, structure, BIEs,
and KIEs found with human PNP.7
5′-3H BIEs and KIEs of inosine with PNP allow distinction
between bond vibrational distortions that occur upon binding of
the substrate from those distortions that arise from chemistry at
the TS. Comparing the [5′-3H]inosine isotope effects with those of
[5′-3H]ImmH and [5′-3H]DADMe-ImmH (pM TS analogues)
permits direct analysis of the bond distortions experienced at the
TS with those that occur on the binding of TS analogues.
While small bond distortional changes result from converting
unbound [5′-3H]inosine to the Michaelis complex, greater distor-
tional change results from formation of the TS (BIE 1.5% vs KIE
4.7%). If TS analogue binding mimics the distortional geometry
of the TS, we might expect TS analogues to show BIEs similar to
the combined KIE and BIE found in catalysis (V/K KIE of ∼6%).
If the binding of TS analogues converts a dynamic excursion that
forms the TS into a collapsed protein conformation, larger BIEs
would be expected for TS analogues than for inosine.
ImmH and DADMe-ImmH are potent inhibitory TS analogues
of PNP with Kd values of 58 and 11 pM, respectively.7-9 The Km/
Kd values of 690 000 and 3 700 000 relative to inosine indicate
substantial capture of the TS effect (Figure 1). These compounds
were synthesized with one of the 5′-hydrogens labeled with 3H and
separately with 14C as a remote label. Tritium was introduced by
reduction of the corresponding 5′-aldehydes with Na[3H]BH4, giving
Figure 1. Phosphorolysis of inosine catalyzed by PNP. The oxacarbenium-
ion TS, its chemically stable analogues, ImmH and DADMe-ImmH, and
positions of isotopic and remote labels are shown.
mixtures of the 5′-R and 5′-S monolabeled isomers. 14C was
incorporated into the 9-deazahypoxanthine moiety of ImmH and
the methylene bridge of DADMe-ImmH (see Supporting Informa-
tion).
Competitive binding of [5′-3H]ImmH and [5-14C]ImmH and of
[5′-3H]DADMe-ImmH and [methylene-14C]DADMe-ImmH was
performed using the ultrafiltration method.10,11 Solutions of 50 mM
KPO4 (pH 7.4), 20-60 µM inhibtor (∼4:1 ratio of 3H:14C by dpm)
and 10 µM PNP were applied to an ultrafiltration apparatus
(quadruplicate 100 µL samples), and 30 psi argon was applied until
approximately half of the solution had passed through the dialysis
membrane (10 kDa MW retention limit) into the lower well (90-
120 min). Samples from both upper and lower wells (30 µL) were
mixed with ten mL of scintillation fluid (Perkin-Elmer UltimaGold)
and counted for at least six cycles of 10 min each. Spectral
deconvolution of 3H and 14C was performed as described previously
using a 14C standard in a matrix identical to the BIE samples. The
BIEs were calculated according to:
14CT/14CB - 1
3HT/3HB - 1
BIE )
(1)
where 14CT and 14CB are the 14C counts in the top and bottom wells,
3
3
3
respectively, and HT and HB are the H counts in the top and
bottom wells, respectively.
ImmH and DADMe-ImmH yielded large BIEs of 13% and 29%,
respectively, which dwarf the 1.5% BIE and the 4.7% KIE measured
for the substrate inosine (Table 1).7 Thus, both TS analogue
inhibitors of PNP undergo larger bond vibrational distortions than
occur at the TS.
Factors contributing to a 5′-3H BIE include the degree of
hyperconjugation between the σ* orbitals of the 5′-C-H bonds and
neighboring electron density at C-4′ and O-5′. The contribution to
the BIE due to C-4′ interactions was estimated by model calculations
(Gaussian 98, B3LYP/6-31G** theory) comparing free ImmH, with
† Albert Einstein College of Medicine.
‡ Industrial Research, Ltd.
9
2166
J. AM. CHEM. SOC. 2008, 130, 2166-2167
10.1021/ja7104398 CCC: $40.75 © 2008 American Chemical Society

C O M M U N I C A T I O N S
Table 1. 5′-3H Binding Isotope Effect Data for Human PNP
His257 in the crystal structure (Nδ-O-5′ distance ) 2.84 Å)
elevates the calculated BIEs at all 5′-OH dihedral angles by 4%
(Figure 2b). As the base is brought closer, the calculated BIEs
increase sharply by as much as 6% at 2.6 Å, which is within error
limits of the X-ray structure.
ligands
K
d or Km
BIEa
[5′-3H]-, [5′-14C]Ino
40 µM
58 pM
11 pM
1.015 ( 0.003 (9)b
1.126 ( 0.005 (32)
1.292 ( 0.012 (27)
[5′-3H]-, [5-14C]ImmH
[5′-3H]-, [14C]DADMe-ImmH
If each of the above factors contributes multiplicatively to the
overall BIE, combinations may account for a 5′-3H BIE of up to
28% (1.06 × 1.14 × 1.06), a value close to the observed value of
29.2% for [5′-3H]DADMe-ImmH. Additional factors that cannot
be easily observed and modeled may also contribute to this BIE of
unprecedented magnitude. For instance, a 3° distortion of the sp3
geometry of C-5′ imposed by active site interactions has been shown
by Horenstein et al. to account for an observed isotope effect of
5%.12 Additionally, if the H-bond to the 5′-OH is stronger than in
our simplified model, the 5′-C-H bonds would become looser,13
leading to a larger BIE.
a BIE ( standard deviation. The number of replicates is given in
parentheses. b From reference 7.
Prior to these studies, the largest BIE reported for any protein-
ligand system was 8.5% for [4-3H]NAD+ with lactate dehydroge-
nase.14 The magnitude of this BIE on the binding of a substrate is
remarkable, owing to substantial bond loosening that occurs at the
site of hydride acceptance in the cofactor. Interestingly, the
association of [1-18O]oxamate to the NADH-bound form of this
enzyme resulted in a BIE of -1.6%, and this, to our knowledge,
was the only previous determination of an IE for the binding of an
inhibitor.15 The large BIEs demonstrated here illustrate substantial
bond distortion remote from the site of bond breaking upon binding
of TS analogues.
The BIEs for ImmH and DADMe-ImmH increase with the
binding affinity (Table 1). Thus, forces at the 5′ position increase
according to the energy difference between free and bound states.
Distortion of [5′-3H]inosine at the TS and/or combined with the
Michaelis complex is small compared to the TS analogues,
supporting large bond vibrational distortions of TS analogues but
not of inosine at its ribooxacarbenium ion TS.
Acknowledgment. This work was supported by NIH Grant
GM41916.
Figure 2. Calculated effects of 5′-hydroxyl orientation and polarization
on the 5′-3H BIE. (a) The O5′-C5′-C4′-N4′ dihedral angle of “unbound”
ImmH models was varied and compared to the “bound” model having
dihedral angles fixed to those in the crystal structure, and the corresponding
BIEs were calculated using ISOEFF98. Results are shown for the pro-R
and pro-S hydrogens, separately and as an equally weighted average. (b)
An imidazole molecule was added to the “bound” ImmH model in panel a
with fixed orientation as found in the crystal structure, but with Nδ-O-5′
distance varying from 2.84 to 2.20 Å. The BIE is the average for HR and
HS. The green curve in panel a can be seen as the slice at an N-O distance
of 10 Å.
Supporting Information Available: Synthetic methods for selected
compounds, molecular model calculations, and derivation of eq 1. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Wolfenden, R. Bioorg. Med. Chem. 1999, 7, 647-652.
(2) Schramm, V. L. Arch. Biochem. Biophys. 2005, 433, 13-26.
(3) Lewandowicz, A.; Schramm, V. L. Biochemistry 2004, 43, 1458-1468.
(4) Schramm, V. L. Curr. Opin. Chem. Biol. 2007, 11, 529-536.
(5) Fedorov, A.; Shi, W.; Kicska, G.; Fedorov, E.; Tyler, P. C.; Furneaux, R.
H.; Hanson, J. C.; Gainsford, G. J.; Larese, J. Z.; Schramm, V. L.; Almo,
S. C. Biochemistry 2001, 40, 853-860.
(6) Nunez, S.; Antoniou, D.; Schramm, V. L.; Schwartz, S. D. J. Am. Chem.
Soc. 2004, 126, 15720-15729.
little rotational restriction of the 5′-hydroxyl, to ImmH bound in
the active site, which X-ray crystallographic data (PDB ID 1RR6)7
indicates is fixed with an O5′-C5′-C4′-N4′ dihedral angle of
59.3° (Figure 2a). Unbound ImmH exists predominantly with a
dihedral angle between -60° and +60°, predicting a geometric BIE
contribution upon binding of 6% or less. In a similar manner, having
the H-O5′-C5′-C4′ torsion angle fixed by H-bonding to His257
could account for an additional 14% of the 5′-3H BIE, as has been
shown by Lewis and Schramm using an isopropanol-formate
model.11 H-bonding with an enzymatic base also polarizes the 5′-
OH, imparting negative charge character on O-5′ and enhancing
hyperconjugation, thereby loosening the 5′-C-H bonds. Pairing
ImmH with an imidazole molecule having the same geometry as
(7) Murkin, A. S.; Birck, M. R.; Rinaldo-Matthis, A.; Shi, W.; Taylor, E. A.;
Almo, S. C.; Schramm, V. L. Biochemistry 2007, 46, 5038-5049.
(8) Miles, R. W.; Tyler, P. C.; Furneaux, R. H.; Bagdassarian, C. K.; Schramm,
V. L. Biochemistry 1998, 37, 8615-8621.
(9) Lewandowicz, A.; Tyler, P. C.; Evans, G. B.; Furneaux, R. H.; Schramm,
V. L. J. Biol. Chem. 2003, 278, 31465-31468.
(10) Schramm, V. L. J. Biol. Chem. 1976, 251, 3417-3424.
(11) Lewis, B. E.; Schramm, V. L. J. Am. Chem. Soc. 2003, 125, 4785-4798.
(12) Horenstein, B. A.; Parkin, D. W.; Estupinan, B.; Schramm, V. L.
Biochemistry 1991, 30, 10788-10795.
(13) Maiti, N. C.; Zhu, Y.; Carmichael, I.; Serianni, A. S.; Anderson, V. E. J.
Org. Chem. 2006, 71, 2878-2880.
(14) LaReau, R. D.; Wan, W.; Anderson, V. E. Biochemistry 1989, 28, 3619-
3624.
(15) Gawlita, E.; Paneth, P.; Anderson, V. E. Biochemistry 1995, 34,
6050-6058.
JA7104398
9
J. AM. CHEM. SOC. VOL. 130, NO. 7, 2008 2167