Experimental evidence for enzyme-enhanced coupled motion/quantum mechanical hydrogen tunneling by ketosteroid isomerase

Article abstract of DOI:10.1021/ja0732330

Although there are considerable data demonstrating that quantum mechanical hydrogen tunneling (HT) occurs in both enzymatic and nonenzymatic systems, little data exist that address the question of whether enzymes enhance the amount of HT relative to the c

Full text of DOI:10.1021/ja0732330

Published on Web 04/22/2008
Experimental Evidence for Enzyme-Enhanced Coupled
Motion/Quantum Mechanical Hydrogen Tunneling by
Ketosteroid Isomerase
Thomas C. Wilde, Grzegorz Blotny, and Ralph M. Pollack*
Department of Chemistry and Biochemistry, UniVersity of Maryland Baltimore County,
1000 Hilltop Circle, Baltimore, Maryland 21250
Received May 7, 2007; E-mail: pollack@umbc.edu
Abstract: Although there are considerable data demonstrating that quantum mechanical hydrogen tunneling
(HT) occurs in both enzymatic and nonenzymatic systems, little data exist that address the question of
whether enzymes enhance the amount of HT relative to the corresponding nonenzymatic reactions. To
investigate whether 3-oxo-∆5-steroid isomerase (ketosteroid isomerase, KSI) enhances HT relative to the
nonenzymatic (acetate-catalyzed) isomerization of ∆5-androstene-3,17-dione (1) to ∆4-androstene-3,17-
dione (3), R-secondary deuterium kinetic isotope effects (KIE) at C-6 of the steroid were determined for
both the KSI- and acetate-catalyzed isomerizations. The normal intrinsic secondary KIE for both wild type
(WT) KSI (1.073 ( 0.023) and acetate (1.031 ( 0.010) suggest the possibility of coupled motion (CM)/HT
in both the enzymatic and nonenzymatic systems. To assess the contribution of CM/HT in these reactions,
the secondary KIE were also measured under conditions in which deuterium instead of hydrogen is
transferred. The decrease in secondary KIE for WT (1.035 ( 0.011) indicates the presence of CM/HT
in the enzymatic reaction, whereas the acetate reaction shows no change in secondary KIE for deuterium
transfer (1.030 ( 0.009) and therefore no evidence for CM/HT. On the basis of these experiments, we
propose that KSI enhances the CM/HT contribution to the rate acceleration over the solution reaction.
Active site mutants of KSI (Y14F and D99A) yield secondary KIEs similar to that of WT, indicating that
mutations at the hydrogen-bonding residues do not significantly decrease the contribution of CM/HT to the
KSI reaction.
Introduction
ations have led to the conclusion that tunneling effects are
generally similar in enzymatic and nonenzymatic reactions.11
Although transition state (TS) stabilization is generally
accepted as the main source of the extraordinary rate accelera-
tions (up to 1017-fold)1 of enzymatic reactions, the details of
this stabilization have been the subject of much debate. Several
mechanisms have been proposed, including low-barrier hydro-
gen bonds,2,3 preorganization effects,4,5 near attack conforma-
tions,6 and protein dynamics effects.7 Several authors have
proposed that enzymes might accelerate hydrogen-transfer
reactions by enhancing quantum mechanical tunneling relative
to solution, either by inducing tunneling where it did not exist
in the solution reaction or by increasing the amount of tunneling
already present.8–10 In contrast, however, theoretical consider-
Although hydrogen tunneling (HT) is well-documented, both
in solution and in enzymatic reactions,12,13 there is little
experimental evidence that demonstrates enhanced HT in an
enzymatic reaction relative to the corresponding model reaction
in solution. Alston et al.8,14 found a secondary tritium kinetic
isotope effect (KIE) larger than the equilibrium isotope effect
(EIE) for proton transfer by triosephosphate isomerase, in
contrast to the hydroxide-catalyzed enolization of acetophenone,
which exhibits a KIE smaller than the EIE. They interpreted
these results in terms of an enhanced tunneling contribution to
the enzymatic reaction. Similarly, the secondary KIEs for yeast
aldolase15 and enoyl-CoA-hydratase16 are also larger than the
EIEs, again leading to the possibility of enhanced tunneling in
these enzymatic systems. However, Doll et al. reported that the
extent of tunneling for the nonenzymatic abstraction of a
(1) Radzicka, A.; Wolfenden, R. Science 1995, 267, 90–93.
(2) Cleland, W. W. Biochemistry 1992, 31, 317–319.
(3) Gerlt, J. A.; Gassman, P. G. J. Am. Chem. Soc. 1993, 115, 11552–
11568.
(4) Warshel, A. J. Biol. Chem. 1998, 273, 27035–27038.
(5) Cannon, W. R.; Singleton, S. F.; Benkovic, S. J. Nat. Struct. Biol.
1996, 3, 821–833.
(11) Olsson, M. H.; Siegbahn, P. E.; Warshel, A. J. Am. Chem. Soc. 2004,
126, 2820–2828.
(12) Bell, R. P. The Tunnel Effect in Chemistry; Chapman and Hall: London,
1980.
(6) Bruice, T. C.; Lightstone, F. C. Acc. Chem. Res. 1999, 32, 127–136.
(7) Hammes-Schiffer, S.; Benkovic, S. J. Annu. ReV. Biochem. 2006, 75,
519–541.
(13) Liang, Z. X.; Klinman, J. P. Curr. Opin. Struct. Biol. 2004, 14, 648–
655.
(8) Alston, W. C.; Kanska, M.; Murray, C. J. Biochemistry 1996, 35,
12873–81.
(14) Alston, W. C.; Haley, K.; Kanski, R.; Murray, C. J.; Pranata, J. J. Am.
Chem. Soc. 1996, 118, 6562–6569.
(9) Huskey, W. P.; Schowen, R. L. J. Am. Chem. Soc. 1983, 105, 5704–
5706.
(15) Biellmann, J. F.; O’Connell, E. L.; Rose, I. A. J. Am. Chem. Soc.
1969, 91, 6484–8.
(10) Cha, Y.; Murray, C. J.; Klinman, J. P. Science 1989, 243, 1325–1330.
(16) Bahnson, B. J.; Anderson, V. E. Biochemistry 1989, 28, 4173–4181.
9
10.1021/ja0732330 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 6577–6585 6577

A R T I C L E S
Wilde et al.
Scheme 1
hydrogen atom from ethylene glycol by coenzyme B12 (5′-
deoxyadenosylcobalamin, AdoCbl)17 is identical within error
to that of a very similar enzymatic reaction (hydrogen atom
abstraction from methylmalonyl-CoA by AdoCbl-dependent
methylmalonyl-CoA mutase),18 leading to the conclusion that
there is no enhanced quantum mechanical tunneling for this
enzyme. Valley and Fitzpatrick19 found no evidence for tun-
neling in either the nonenzymatic (acetate or phosphate) or
enzymatic (nitroalkane oxidase) deprotonation of nitroethane,
again providing no basis for enhanced tunneling in an enzymatic
reaction. Thus, there are only a few examples of studies that
either support or refute the hypothesis that enzymes enhance
quantum mechanical HT. The scarcity of experimental evidence
is due to the difficulty of designing a suitable system for
comparison. Not only must one find a reaction that occurs by
identical mechanisms both in solution and at the active site of
the enzyme, but the details of the rate-limiting step(s) must be
known, and both systems must be amenable to the measurement
of precise KIEs.
The ketosteroid isomerase (KSI)-catalyzed isomerization of
5-androstene-3,17-dione (1) to 4-androstene-3,17-dione (3)
provides just such a system, because both the enzymatic and
solution reactions have been extensively characterized,20–24 and
both the enzymatic and acetate-catalyzed reactions occur through
the same mechanism (Scheme 1). KSI (3-oxo-∆5-steroid
isomerase, EC 5.3.3.1.) uses a single active site base (Asp-38,
pKa 4.57)25 to transfer a proton from C-4 to C-6 of the steroid,
and this base can be mimicked by acetate (pKa 4.75) in the
corresponding nonenzymatic reaction. In both reactions, proton
abstraction from C-4 (enolization) generates a dienolate inter-
mediate (2), which is then reprotonated at C-6 (ketonization)
to form the conjugated product (3). KSI stabilizes the intermedi-
ate and the flanking TSs (TS1 and TS2) through hydrogen
bonding from Tyr-14 and Asp-99 to O-3 of the steroid.26,27 In
addition to the mechanistic information regarding these reac-
tions, free energy diagrams and the associated rate constants
for all steps have been determined for both the KSI-28 and the
acetate-catalyzed23 reactions (Figure 1), enabling intrinsic KIEs
to be calculated from the observed values.
To probe for the existence of tunneling in the ketonization
step, we determined secondary deuterium KIEs at C-6 for the
second step of the reaction (protonation at C-6) for both
enzymatic and nonenzymatic catalysis. This method is based
upon the study of Huskey and Schowen,9 in which they showed
that quantum mechanical HT can result in inflated secondary
hydrogen isotope effects for systems in which the isotopic
hydrogen is coupled to the transferring hydrogen. In this work,
we describe results of our investigation that provide evidence
for enzymatic enhancement of coupled motion (CM)/HT by KSI.
Materials and Methods
Materials. WT and mutant isomerases (Y14F and D99A) were
available from previous work.29 Substrate (5-androstene-3,17-dione)
was prepared as described previously.30 All chemicals used were
reagent grade or better. UV data were obtained on either a Cary 1
Bio or Cary 100 Bio spectrophotometer. IR spectra were obtained
on a ThermoNicolet Avatar 320 FTIR instrument equipped with
an attenuated total reflectance crystal for solid-sample analysis.
NMR spectra were acquired on either a Bruker 300 MHz or Bruker
Avance 500 MHz instrument. Chemical shifts are given in ppm.
5-Androstene, 6-deutero-3,17-dione (1-6D). Compound 1-6D
was prepared in seven steps, as shown in Scheme 2. This synthetic
route is based on the work of Palmer et al.31 in the preparation of
C-6 tritiated cholesterol derivatives.
(17) Doll, K. M.; Bender, B. R.; Finke, R. G. J. Am. Chem. Soc. 2003,
125, 10877–10884.
(18) Chowdhury, S.; Banerjee, R. J. Am. Chem. Soc. 2000, 122, 5417–
5418.
(19) Valley, M. P.; Fitzpatrick, P. F. J. Am. Chem. Soc. 2004, 126, 6244–
6245.
(20) Pollack, R. M. Bioorg. Chem. 2004, 32, 341–53.
(21) Houck, W. J. Ph.D. Thesis, University of Maryland Baltimore County,
Baltimore, MD, 2004.
(22) Pollack, R. M.; Thornburg, L. D.; Wu, Z. R.; Summers, M. F. Arch.
Biochem. Biophys. 1999, 370, 9–15.
(23) Zeng, B.; Pollack, R. M. J. Am. Chem. Soc. 1991, 113, 3838–3842.
(24) Houck, W. J.; Pollack, R. M. J. Am. Chem. Soc. 2003, 125, 10206–
10212.
(25) Pollack, R. M.; Bantia, S.; Bounds, P. L.; Koffman, B. M. Biochemistry
1986, 25, 1905–1911.
(26) Choi, G.; Ha, N. C.; Kim, S. W.; Kim, D. H.; Park, S.; Oh, B. H.;
Choi, K. Y. Biochemistry 2000, 39, 903–909.
(27) Thornburg, L. D.; Henot, F.; Bash, D. P.; Hawkinson, D. C.; Bartel,
S. D.; Pollack, R. M. Biochemistry 1998, 37, 10499–10506.
(28) Hawkinson, D. C.; Eames, T. C.; Pollack, R. M. Biochemistry 1991,
30, 10849–10858.
(29) Thornburg, L. D.; Goldfeder, Y. R.; Wilde, T. C.; Pollack, R. M. J. Am.
Chem. Soc. 2001, 123, 9912–9913.
(30) Pollack, R. M.; Zeng, B.; Mack, J. P. G.; Eldin, S. J. Am. Chem. Soc.
1989, 111, 6419–6423.
(31) Palmer, M. A.; Goad, L. J.; Goodwin, T. W. NouV. J. Chim. 1978, 2,
401–404.
Figure 1. Free energy diagram for the isomerization of 1 to 3 catalyzed
by acetate and by wild type (WT) KSI at 25 °C.
9
6578 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Hydrogen Tunneling in 3-oxo-∆5-Steroid Isomerase
A R T I C L E S
Scheme 2
3ꢀ-Tosyloxy-5-androstene-17-one (5). 32 A total of 2.884 g (10
mmol) of 3ꢀ-hydroxy-5-androstene-17-one (4) and 2.856 g (15
mmol) of tosyl chloride were stirred in 40 mL of dry pyridine for
20 h at room temperature. Pyridine was removed by rotary
evaporation under reduced pressure, and the remaining substance
was dissolved in ethyl acetate. The organic phase was washed with
1 N HCl, water, sodium bicarbonate, and water. After drying
(MgSO4) and filtration, the solvent was removed, and the residue
was purified by column chromatography by using silica gel (Merck
60) and hexane:ethyl acetate (1:1). Recrystallization from ethyl
acetate-hexane gave 85% yield (3.757 g); mp 150-152 °C (lit.
mp 153-155 °C);33 Rf 0.81 (hexane/ethyl acetate 1:1). Anal. Calcd
for C26H34SO4 (442.61): (C, 70.55; H, 7.74; S, 7.24; O, 14.45).
Found: (C, 70.48; H, 7.69; S, 7.02). IR, λmax (solid): 2935, 2896,
was extracted with ethyl acetate, washed with water, and dried
(MgSO4). Purification, as described for compound 5, gave 85%
yield (2.02 g); mp 128-130 °C (lit. mp 132-135 °C);34 Rf 0.60
(hexane/ethyl acetate 1:1). Two spots were apparent by thin-layer
chromatography (TLC), which are likely the two (R and ꢀ) isomers
of the C-6 alcohol group. Anal. Calcd for C19H28O2 (288.43): (C,
79.12; H, 9.78; O, 11.09). Found: (C, 78.77; H, 9.81; O, 11.05).
IR, λmax (solid): 3596, 3060, 2997, 2861, 1720 cm-1
.
3,5 Cyclo-androstane-6,17-dione (7). A total of 5 mL of Jones
reagent was slowly added to a solution of compound 6 (1.9 g, 6.59
mmol) in 20 mL of acetone (previously cooled to ∼5 °C). The
mixture was stirred at this temperature for 20 min, at which point
excess Jones reagent was destroyed by addition of methanol. Water
was added, and the product was extracted with ethyl acetate, washed
with water, and dried (MgSO4). Ethyl acetate was removed, and
the product was purified by column chromatography by using silica
gel (Merck 60) and hexane:ethyl acetate (1:1). Recrystallization
from methanol gave 63% yield (1.18 g); mp 188-190 °C; Rf 0.65
(hexane/ethyl acetate 1:1). Anal. Calcd for C19H26O2 (286.41): (C,
79.67; H, 9.15; O, 11.17). Found: (C, 79.33; H, 9.18; O, 11.14).
1
1734, 1362, 1174 cm-1; H NMR (CDCl3, 300 MHz): δ 0.86 (s,
3H, C-18 methyl), 0.99 (s, 3H, C-19 methyl), 2.44 (s, 3H, tosyl
methyl), 4.35 (m, 1H, C-4 H), 5.38 (m, 1H, C-6H), 7.34 (d, 2H, J
) 8 Hz), 7.79 (d, 2H, J ) 8 Hz)ν.
3,5 Cyclo-6-hydroxy-androstane-17-one (6). Compound 5
(3.642 g, 8.23 mmole) and potassium acetate (4.03 g, 41 mmol) in
an acetone/water (80 mL:20 mL) mixture were refluxed for 18 h.
The acetone was removed by rotary evaporation, and the residue
IR, λmax (solid): 2956, 1731, 1679 cm-1
.
3,5 Cyclo-6,17-D-androstane-6,17-diol (8). Compound 7 (1.00
g, 3.49 mmol) of and 467 mg (11.15 mmol) of sodium borodeu-
teride (NaBD4, >99 atom % D) were dissolved in 30 mL of
diglyme:THF (1:1) and heated for 24 h. Solvents were removed
by rotary evaporation under reduced pressure. Water and ethyl
acetate were added, and the product was extracted into the ethyl
acetate fraction and dried (MgSO4). After filtration and removal of
the solvent, the residue was purified by column chromatography
by using silica gel (Merck 60) and hexane:ethyl acetate (4:1).
Recrystallization from ethyl acetate gave 88% yield (905 mg); mp
181-183 °C. Anal. Calcd for C19H28D2O2 (292.45): (C, 78.03; D
+ H, 10.34; O, 10.94). Found: (C, 77.40; H, 10.31; O, 11.22). IR,
(32) We later found this synthetic method to be irreproducible, and we
have since developed an easier, more reproducible method of tosylation
based on the work of Velusamy et al. (Velusamy, S. Kirankunar, J. S.
Punniyamurthy, T. Tetrahedron Lett. 2004, 45, 203) In a flask equipped
with a Dean-Stark trap, 2.884 g (10 mmol) of 3-ꢀ-hydroxy-5-
androstene-17-one, 1.90 g (10 mmol) of p-toluenesulfonic acid
monohydrate and 119 mg (5 mol%) of cobalt hexahydrate in 40 mL
of 1,2-dichloroethane were stirred and refluxed for 45 minutes. Solvent
was removed under reduced pressure, and the residue was dissolved
in ethyl acetate. The organic phase was washed in 1 N HCl, water,
sodium bicarbonate, and water. After drying (MgSO4) and filtration,
the solvent was removed, and the residue was purified by column
chromatography by using silica gel (Merck 60) and hexane:ethyl
acetate (1:1). Recrystallization from ethyl acetate-hexane gave 83%
yield (3.667 g).
(33) Crabb, T. A.; Dawson, P. J.; Williams, R. O. J. Chem. Soc., Perkin
Trans. 1980, 1, 2535–2541.
(34) Blickenstaff, R. T.; Atkinson, K.; Breaux, D.; Foster, E.; Kim, Y.;
Wolf, G. C. J. Org. Chem. 1971, 36, 1271–1277.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6579

A R T I C L E S
Wilde et al.
1
λmax (solid): 3396, 3324, 2949, 2868, 2160, 1040 cm-1. H NMR
(CDCl3, 500 MHz): δ 0.27 (t, 1H), 0.61 (m, 1H), 0.77 (s, 3H),
0.93 (s, 3H).
The procedure above was repeated for the acetate-catalyzed
formation of side products, except that a 2.0 M potassium acetate
solution was used instead of dilute hydroxide. To repeat these
reactions in the absence of air, a vacuum was applied to a three-
neck round-bottom flask containing the reaction mixture (prior to
addition of 1), followed by addition of argon. This process was
repeated twice more prior to the addition of 1. Positive argon
pressure was maintained throughout the reactions to avoid introduc-
tion of air.
LC/MS analysis was performed by using an Agilent 1100 HPLC
connected to a Bruker Daltronics Esquire 3000 Plus (ion trap) mass
spectrometer equipped with an atmospheric pressure chemical
ionization (APCI) accessory. A 2.0 × 250 mm C18 column
(Beckman Ultrasphere 5 µm) was used with a 0.4 mL/min flow
rate. Isocratic elution by using 55:45 methanol:water mobile phase
(15 min) was followed by an 80:20 methanol:water column wash
with 0.5% acetic acid present throughout. Postcolumn mobile phase
was injected directly into the mass spectrometer for analysis. MS
parameters: nebulizer, 50.0 psi; dry gas, 4.0 L/min; dry gas
temperature, 350 °C; APCI temperature, 450 °C; positive ion mode
scanning, 120-800 m/z.
Deuteration of 1 and 1-6D at C-4. Compound 1 (1-2 mg) was
incubated in 20-30 mL of 1.5 M potassium acetate buffer dissolved
in D2O (containing 20% CH3OD and 1 mM ethylenediamine
tetraacetic acid, EDTA) for about 1 h to exchange the C-4 protons
for deuterium. The reaction was halted by extraction of the steroids
with methylene chloride (3 × 10 mL). The extracts were com-
bined and evaporated to dryness at reduced pressure and tempera-
ture (5-15 °C) to avoid isomerization. The remaining solids were
purified by HPLC by using a Waters Symmetry C18 column (19 ×
150 mm) on a Waters 600 with isocratic elution (70:30 methanol:
water). Fractions containing 5-androstene-4,4-dideutero-3,17-dione
(1-4D2, λmax ) 290 nm, retention time ) 7 min) were collected
and evaporated to dryness by using a vacuum pump at reduced
temperature (5-15 °C).
5-Androstene-6-D-3,17-diol-acetate (9). Compound 8 (855 mg,
2.92 mmol) and zinc acetate (2.137 g, 11.68 mmol) were dissolved
in 30 mL of glacial acetic acid and refluxed for 20 h. Solvent was
removed, and the product was extracted and purified as described
for compound 8. Recrystallization from methanol gave 87% yield
(958 mg); mp 160-161 °C (lit. mp (undeuterated) 158-159.5
°C);35 Rf 0.68 (hexane/ethyl acetate 4:1). Anal. Calcd for
C23H32D2O2 (376.53): (C, 73.36; H + D, 9.10; O, 16.99). Found:
(C, 73.29; H, 9.07; O, 17.53). IR, λmax (solid): 2948, 2877, 2210,
1730, 1246, 1040 cm-1 (lit. IR λmax (KBr): 1730, 1620, 1250, 1040
cm-1).35
5-Androstene-6-D-3,17-diol (10). Compound 9 (871 mg) was
added to a solution of 500 mg of potassium hydroxide in 25 mL of
methanol/water (4:1), and the solution was refluxed for 3 h. After
cooling, the precipitated solid was filtered off, washed with water
until the filtrate pH was neutral, dried, and recrystallized from ethyl
acetate to give 66% yield (443 mg); mp 176-179 °C (lit. mp
178-179).35 Anal. Calcd for C19H28D2O2 ·H2O (310.47): (C, 73.50;
H + D, 10.39; O, 15.46). Found: (C, 73.35; H, 10.35; O, 15.09).
IR, λmax (solid): 3466, 3199, 2941, 2881, 2826, 2227, 1654, 1434,
1374, 1056, 1032 cm-1 (lit. IR λmax (KBr): 3400, 1620, 1050
cm-1).35
5-Androstene-6-D-3,17-dione (1-6D). A total of 1 mL of Jones
reagent was added dropwise to 98.6 mg (0.34 mM) of compound
10 dissolved in 20 mL of acetone previously cooled to ∼0 °C. After
3 min, excess Jones reagent was destroyed by addition of methanol.
A cold, saturated NaCl solution was added to the reaction mixture,
and the product was extracted with an excess of ethyl acetate. The
organic phase was washed with water, dried (MgSO4), and filtered.
Solvent was removed by rotary evaporation under reduced pressure
at room temperature. The product was purified by column chro-
matography by using silica gel (Merck 60) and hexane/ethyl acetate
(6:4). Recrystallization from ether gave 39.5% yield (40 mg). IR,
The same procedure was used with the 1-6D substrate to create
a triply deuterated substrate, 5-androstene-4,4,6-trideutero-3,17-
dione (1-4D26D). Purity of these C-4 deuterated substrates was
assessed by HPLC, which showed only two peaks corresponding
to the substrate and a small amount of product. UV analysis showed
less than 2% of product. The extent of deuteration was assessed
λ
max (solid): 2966, 2941, 2884, 2860, 2824, 2236, 1736, 1708, 1456,
1392, 1374, 1022, 1012 cm-1. Mass spectrometry (MS, APCI, low
1
resolution) 288.1 m/z. H NMR (CDCl3, 500 MHz, integration of
the C-6H (5.36 ppm) peak relative to the C-4ꢀH (3.28 ppm) and
C-4RH (2.84 ppm) peaks in the 500 MHz 1H NMR of 1-6D shows
that >99% deuteration at C-6 was achieved by this method): δ
3.28 (dt, 4ꢀ-H), 2.84 (dd, 4R-H), 1.21 (s, 19-Me), 0.92 (s, 18-Me).
lit. δ 3.28 (dt, 4ꢀ-H), 2.84 (dd, 4R-H), 5.36 (dt, 6-H), 1.20 (s, 19-
Me), 0.90 (s, 18-Me).36 No signal was observed at 5.36. TLC of
the product showed only one spot. Rf (hexane/ethyl acetate (3:2))
0.66.
1
by H NMR (500 MHz) in CDCl3 (99.8 atom %D, 0.05% TMS,
Cambridge Isotope Laboratories). Integration of the remaining 1H
NMR C-4H peaks (3.36 and 2.84 ppm) relative to the C-18 methyl
peak (0.9 ppm) showed that >99% deuteration at both the C-4R
and C-4ꢀ positions was achieved by this process.
KIEs in H2O. As described by Vitullo and Logue,37 small KIEs
can be accurately determined in a noncompetitive manner by the
simultaneous UV monitoring of the reactions of unsubstituted and
isotopically substituted molecules. In these experiments, a UV-vis
spectrophotometer with a cell changer capable of switching between
six cuvettes was used to monitor six separate reactions simulta-
neously. KIEs in acetate buffer were determined by adding identical
amounts of acetate buffer and methanol to each cuvette, followed
by addition of 1 (dissolved in methanol) to three cuvettes and 1-6D
to the other three cuvettes. The UV absorbances were monitored
at 248 nm for 6-10 half-lives (∼20-60 h, 98.5-99.9% comple-
tion). Infinity points were observed to be stable. Final conditions
in the cuvettes were 3.3% methanol, 1 mM EDTA, 1-3 M acetate
buffer, and 25.0 °C. The concentrations of 1 and 1-6D were identical
to each other for each set of runs but varied between 6 and 40 µM
in separate sets.
Autoxidation during Isomerization. Autoxidation products
were identified by using LC/MS, H NMR, IR, and UV spectros-
1
copy and high-resolution MS (Michigan State University Mass
Spectrometry Facility). To acquire sufficient sample for these
analyses, a large-scale hydroxide-catalyzed isomerization was
performed. Methanol (120 mL) containing 100 mg of 1 was added
to 480 mL of 0.8 mM NaOH in a 1 L separatory funnel. The
solution was mixed and left open to air. After 40 min, the UV
spectrum had stopped changing, and the reaction mixture was
extracted with chloroform (3 × 60 mL). The organic layers were
combined, dried (MgSO4), and filtered. After evaporation of the
filtrate to dryness under reduced pressure, the dried extracts (114
mg) were dissolved in 4 mL of chloroform and loaded on a
preparatory TLC plate (silica gel, Merck 60). The first separation
by using a 60:40 hexane:ethyl acetate mobile phase did not provide
adequate separation of bands; therefore, the plate was run a second
time using the same solvent mixture. Individual bands were
extracted by using a 60:40 CHCl3:MeOH mixture, filtered, and
evaporated to dryness.
The data were edited to be evenly weighted according to changes
in absorbance.38 KIEs were calculated by averaging the first-order
rate constants of the three cuvettes containing 1 and dividing this
number by the average of the rate constants from the three 1-6D
(35) Van Lier, J. E.; Smith, L. L. J. Org. Chem. 1970, 35, 2627–2632.
(36) Xue, L. A.; Talalay, P.; Mildvan, A. S. Biochemistry 1990, 29, 7491–
7500.
(37) Vitullo, V. P.; Logue, E. A. J. Am. Chem. Soc. 1976, 98, 5906–5909.
9
6580 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Hydrogen Tunneling in 3-oxo-∆5-Steroid Isomerase
A R T I C L E S
Scheme 3
in dilute phosphate buffer in the following manner. A solution of
10 mL of methanol containing a ∼2:1 mixture of 1:1-6D was added
to 290 mL of 6 mM KPi buffer (pH 7.0, 1 mM EDTA). A 40 mL
aliquot of this pre-enzyme solution was extracted with methylene
chloride (3 × 10 mL). Solvent was removed under reduced pressure,
and the solids were redissolved in methanol. This solution was
labeled 0% reaction and stored at -80 °C to await HPLC separation
of substrate and product. A solution of 200 µM D99A (8 µL) was
added to the remaining 260 mL of pre-enzyme solution. Final
conditions were 3.3% methanol, [1] ) 11.2 µM, [1-6D] ) 4.5 µM,
and [D99A] ) 6.2 nM. Aliquots were taken at 10, 50, and 85%
reaction (determined by UV absorbance at 248 nm) and worked
up identically to the pre-enzyme aliquot. Finally, 1 µL of WT (336
µM) was added to the residual solution to drive the reaction to
completion, and the remaining solution was worked up and labeled
100% reaction.
The five time point extracts (0, 10, 50, 85, and 100% reaction)
were separated by HPLC as described earlier. The substrate and
product fractions were collected and evaporated to dryness under
high-vacuum at 5-15 °C. These fractions were then dissolved in
methanol and analyzed by electrospray mass spectrometry (ESI-
MS). Individual peaks from resulting mass spectra were integrated
by summing the observed intensities in 0.025 m/z increments over
the entire peak (e.g., for the 287 m/z peak, the intensities between
286.6 and 287.4 were summed). Peak integrations obtained were
used to determine C-6H(D) KIEs.
cuvettes. At least three such values were determined and averaged
to obtain each reported KIE.
Enzymatic KIEs were determined in a similar manner. Six
cuvettes were prepared with phosphate buffer (4 mM, pH 7.0, 1.0
mM EDTA), methanol, and substrate or deuterated substrate. Then,
identical volumes of enzyme (10-20 µL, diluted in 2.5 mg/mL
bovine serum albumin (BSA)) were added to all six cuvettes. Final
substrate concentrations in the cuvettes were between 6 and 12 µM.
Catalyst concentrations were as follows: acetate (1-3 M); WT (5-7
pM for 3.3% methanol, 195-225 pM for 20% methanol); Y14F
(80-125 nM); and D99A (12-16 nM). Enzyme-catalyzed reactions
were monitored at 248 nm for 90-120 min (7-10 half-lives,
99-99.9% completion). Data analysis was performed in a manner
identical to that used for acetate.
KIEs in D2O. KIEs in D2O were determined in a manner similar
to that used to determine KIEs in H2O with the following
modifications. A solution of 4 mM deuterated phosphate buffer
(KPi), containing 1 mM EDTA, was used for the enzymatic runs.
Acetate buffers (3-4 M, pD 7.0), containing 1 mM EDTA, were
prepared by dissolving potassium acetate and Na2EDTA·2H2O in
D2O (99.9 atom % D) and adjusting the pD with small quantities
1
of glacial acetic acid. The percent of H for all deuterated buffers
The ratio of the M + 1 and M + 2 peaks (287 and 288 m/z,
respectively) was used to determine the ratio of deuterated/
nondeuterated substrate and product at each time point (see
Supporting Information). Equations 2 and 3 use these ratios to
determine the isotope effect based on substrate and product ratios,
respectively,42,43
is calculated to be <0.2%. Deuterated methanol (CH3OD) was used
in place of protonated methanol, and the 2.5 mg/mL BSA solutions
used for enzyme dilution were made with deuterated 4 mM KPi
buffer. Substrates used were 5-androstene-4,4-dideutero-3,17-dione
(1-4D2 for kH) and 5-androstene-4,4,6-trideutero-3,17-dione (1-
4D26D for kD). Enzyme concentrations were increased by 5-8-
fold over those used in protonated solvents to account for primary
hydrogen and solvent KIEs.
Intrinsic KIE Calculation for the WT Reaction. Reactions
catalyzed by acetate,23 Y14F,29,39 and D99A29,39 all have clean
rate-limiting ketonization steps; therefore, the observed KIEs are
the intrinsic KIEs for these catalysts. However, ketonization is not
cleanly rate-limiting for the WT-catalyzed reaction;28 therefore, the
observed KIEs are smaller than the intrinsic KIEs. Because the
relative heights of the TSs for the individual kinetic steps are known
for the solvent systems used, the intrinsic KIEs can be calculated
from the observed KIEs by using eq 140
(
)
KIE ) kH ⁄ kD ) log (1 - F) ⁄ log [ 1 - F RSF ⁄ R0]
(2)
(3)
KIE ) kH ⁄ kD ) log(1-F) ⁄ log [(1 - F)RPF ⁄ R0]
where F is the fraction of reaction completed, RSF is the ratio of
[1-6D]/[1] at fraction F, RPF is the ratio [3-6D]/[3] at fraction F,
and R0 is the [1-6D]/[1] ratio at the start (or [3-6D]/[3] at the end)
of the reaction.
Direct Injection ESI-MS. Mass spectra of the substrate and
product were acquired by dissolving the steroid in methanol (100
µM) and injecting the solution directly into the mass spectrometer
(Bruker Daltronics Esquire 3000 Plus). A total of 50 µL/min of
steroid solution was mixed with 500 µL/min of 60:40 methanol:
0.5% acetic acid solution by using a T-connector. Mass spectrometer
parameters were as follows: total injection flow rate, 550 µL/min;
APCI temperature, 450 °C; dry gas (N2) temperature, 350 °C; dry
gas flow rate, 5.0 L/min; and selective ion monitoring between 282.5
and 292.5 m/z. Spectra were acquired for 6-10 min and averaged.
V ⁄ K ) k + C + C × EIE ⁄ 1 + C + C
) ( )
r
(1)
(
D
D
f
r
f
in which VD/K is the observed KIE on kcat / Km, kD is the intrinsic
KIE, Cf and Cr are the forward and reverse commitment factors,
respectively, and EIE is the EIE between reactant and product.
Because the microscopic rate constants for isomerization by WT
are known,38 Cf and Cr (and therefore the intrinsic KIE) can easily
be calculated. For the mechanistic scheme shown (Scheme 3), where
k5 (ketonization) is the isotopically sensitive step, Cf ) (k5 / k4)[(k3
+ k2) / k2] and Cr ) k6 / k7. Because k6 , k7,28 Cr ) 0, and it can
be ignored. At high methanol concentrations (20%) k2 . k3, so Cf
) k5 / k4, whereas at low methanol concentrations, k2 and k3 are
similar, so we must use Cf ) (k5 / k4)[(k3 + k2) / k2] for the intrinsic
KIE calculations. It should be noted that only the protonation of
the intermediate at C-6 is stereospecific.41
Results
Autoxidation during Base-Catalyzed Isomerization. Initial
attempts to measure the KIE of the acetate-catalyzed isomer-
ization in the absence of EDTA indicated that the reaction does
not proceed cleanly. Final UV absorbances were 10-15% lower
than calculated, and the kinetic data did not fit well to a first-
order equation. For these reasons, product analysis was
undertaken.
KIE of D99A by MS. A known mixture of undeuterated (1)
and C-6 deuterated (1-6D) compounds was isomerized by D99A
(38) The data were manually edited such that they were evenly weighted
according to changes in absorbance. Typically, five absorbance
readings per 0.001 AU increase were kept, and the rest of the data
was discarded.
(39) Wilde, T. C. Ph.D. Thesis, University of Maryland Baltimore County,
Baltimore, MD, 2005.
(40) Northrop, D. B. Annu. ReV. Biochem. 1981, 50, 103–131.
(41) Zawrotny, M. E.; Hawkinson, D. C.; Blotny, G.; Pollack, R. M.
Biochemistry 1996, 35, 6438–6442.
(42) Strictly speaking, these equations are appropriate for analyses when
the isotopic molecule is present in trace amounts. Because of the small
KIEs measured here, the errors introduced by using these simpler
equations are less than 0.5%, which is considerably smaller than the
experimental error.
(43) Melander, L.; Saunders, W. H. Reaction Rates of Isotopic Molecules;
Wiley: New York, 1980.
(44) Solaja, B. A.; Milic, D. R.; Dosen-Micovic, L. I. Steroids 1994, 59,
330–334.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6581

A R T I C L E S
Wilde et al.
Scheme 4
Table 1. Rate Constants (M-1s-1) for the Acetate- and KSI-Catalyzed (kcat/Km) Isomerizations of 1 at 25 °C
H transfer
D transfer
catalysta
1
1-6D
1-4D2
1-4D26D
acetate
WT
(28.1 ( 1.1) × 10-6 c
(13.7 ( 0.8 )× 107 d
(45.7 ( 1.9) × 105
7320 ( 100f
(27.2 ( 1.4) × 10-6
(13.3 ( 0.8) × 107 e
(43.3 ( 1.4) × 105
6910 ( 60g
(4.6 ( 0.3) × 10-6
(4.07 ( 0.05) × 107
(8.2 ( 0.8) × 105
890 ( 10
(4.4 ( 0.3) × 10-6
(3.99 ( 0.04) × 107
(8.0 ( 0.8) × 105
860 ( 10
WTb
Y14F
D99A
(75 ( 8) × 103
(70 ( 8) × 103
(7.53 ( 0.04) × 103
(7.25 ( 0.06) × 103
a 3.3% methanol was used, unless otherwise noted. b 20% methanol. c Literature values (Lit.) 30.9 × 10-6 21 and 32 × 10-6 M-1 -1 23 d
Lit. 24 ×
s
.
107 M-1 -125
s
and 19.8 × 107 M-1 -1 36 e
s
.
Lit. 19.8 × 107 M-1 -1 36 f
s
.
Lit. 8690 M-1 -1 48 g Lit. 8520 M-1 -1 48
s . s .
The products of the acetate-, hydroxide-, and WT-catalyzed
isomerizations of 1 in the absence of EDTA were analyzed by
LC/MS. The WT reaction shows essentially complete isomer-
ization of 1 to 3, with negligible side-product formation. The
acetate and hydroxide reactions, on the other hand, occur with
significant side-product formation, as evidenced by additional
peaks in the chromatogram. The most prominent side product
was identified as 4-androstene-3,6,17-trione (11), based on its
λmax (250 nm) and M + 1 of 301 m/z. This identification was
C-6H(D) r-Secondary KIEs. The R-secondary deuterium
KIEs at position C-6 of the substrate were obtained from the
pseudofirst-order rate constants in the presence of EDTA for
the acetate- and KSI-catalyzed isomerizations under nonsat-
urating conditions ([S] , KM) by using UV spectroscopy.
Reactions were monitored to greater than g98% completion,
and both acetate- and enzyme-catalyzed reactions showed
excellent pseudofirst-order kinetics. These KIEs were determined
for both hydrogen and deuterium transfer in H2O and D2O,
according to the reactions shown in Scheme 4 panels a and b,
respectively. The second-order rate constants of isomerization
for the different catalysts and conditions are given in Table 1.
1
confirmed by a comparison of the HPLC retention time, H
NMR, IR, UV, and mass spectra of the isolated product with
those for authentic 11 (Steraloids), as well as with literature
values.44,45 Finally, the high resolution mass spectrum of the
isolated product (M + 1 m/z ) 301.1805) is consistent with
the molecular formula (C19H24O3) of 11.
In order to appropriately analyze these data, it is necessary
to calculate the intrinsic isotope effect for TS2 from the observed
rate constants. This intrinsic isotope effect for WT was
calculated from eq 1, as described in the Materials and Methods,
by using the method described by Northrop.46 This model
corrects the observed isotope effect to account for the extent to
which the isotope-dependent step is rate-limiting. For the
mutants and acetate, the intrinsic effect is equal to the observed
effect because the rate-determining step is cleanly ketonization.
Intrinsic R-secondary KIEs for ketonization of the intermediate
are given in Table 2. To determine whether the secondary KIEs
determined in protonated and deuterated buffers are different,
the Student t test47 was applied to the intrinsic KIEs for the
enzymatic data.
The 1H NMR spectrum of the acetate-catalyzed isomerization
product mixture exhibits two major peaks in the vinyl proton
region (4-7 ppm), corresponding to the isomerization product
3 (5.75 ppm) and the side product 11 (6.22 ppm). Integration
of these peaks gave a ratio of about 3:1 favoring 3, whereas an
identical reaction under argon gave a ratio of 7:1. For the
hydroxide-catalyzed reactions, a similar analysis yielded 3:1 and
40:1 ratios, respectively. Although attempts to eliminate oxygen
did reduce formation of 11, a simpler and more efficient method
is the addition of EDTA to the buffer solution. Addition of 1
mM EDTA was shown to eliminate side-product formation
(HPLC).21
(45) Nangia, A.; Anthony, A. Synth. Commun. 1996, 26, 225–230.
(46) Northrop, D. B. Biochemistry 1981, 20, 4056–61.
9
6582 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Hydrogen Tunneling in 3-oxo-∆5-Steroid Isomerase
A R T I C L E S
Table 2. Observed and Intrinsic C-6 R-Secondary Deuterium KIEs for the Isomerization of 1 by Various Catalysts (UV Spectroscopy)
at 25 °C
H transfera
D transferg
catalystb
KIEobsd
KIEintr
KIEobsd
KIEintr
Student t testd
acetate
WT
1.031 ( 0.013
1.033 ( 0.011
1.055 ( 0.017
1.059 ( 0.014
1.075 ( 0.004
1.031 ( 0.013
1.066 ( 0.023e
1.073 ( 0.023g
1.059 ( 0.014
1.075 ( 0.004
1.030 ( 0.009
1.021 ( 0.005
1.029 ( 0.009
1.038 ( 0.008
1.038 ( 0.005
1.030 ( 0.009
1.029 ( 0.007f
1.035 ( 0.011h
1.038 ( 0.008
1.038 ( 0.005
ND
>98%
>99.5%
>90%
>99.9%
WTc
Y14F
D99A
a Error values represent one standard deviation. b 3.3% methanol was used, unless otherwise noted. c 20% methanol. d Probablility that the intrinsic
KIEs for H and D transfers are different. ND ) not determined. e Cf ) 0.99. f Cf ) 0.29. g Cf ) 0.33. h Cf ) 0.21.
KIE by MS. The C-6H(D) KIE for the D99A-catalyzed
isomerization in protonated buffer was also determined by MS.
In these experiments, a ∼2:1 mixture of 1 and 1-6D was
isomerized by D99A, and portions of this solution were
quenched at different time points during the reaction. At each
time point, substrate and product were separated by HPLC,
collected and individually analyzed by direct injection MS. KIEs
were determined by using the 287/288 m/z ratios at three reaction
times, as described in Materials and Methods (KIE in paren-
theses): 10% reaction (1.04), 50% reaction (1.08), and 85%
reaction (1.05, 1.10), giving an average value of 1.07 ( 0.03.
Reliability of the Isotope Effects. Because the R-secondary
KIEs were determined noncompetitively by simultaneous mea-
surements of the pseudofirst-order isomerization rates of pro-
tonated and deuterated substrates, it is necessary to demonstrate
that this method produces reliable isotope effects. Although
this technique has been reported to be effective for determining
KIEs to 1% accuracy,37 the use of separate cuvettes introduces
the possibility that the observed KIEs are due to slight
differences in the solutions or the conditions rather than to actual
isotope effects. One concern is temperature control of the
cuvettes. To check for warm/cool spots, six cuvettes containing
identical amounts of buffer (4 mM KPi, pH 7.0), methanol
(20%), substrate (5.3 µM), and WT (167 pM) were prepared,
and the reaction was monitored for 6-10 half-lives for each
cuvette. This experiment was repeated six times, and differences
between rate constants determined in any two wells were <1%,
demonstrating that temperature variations among solutions are
insignificant. As an added precaution, the cuvette wells contain-
ing protonated and deuterated substrates were switched from
run to run, such that any small differences in temperature
between wells would average out. For noncompetitive deter-
minations of KIEs, the purity of the reactants is an important
consideration, which was addressed by a two-step purification
process. The substrates were first purified by column chroma-
tography and then further purified by reverse-phase HPLC. LC/
MS analysis showed no detectable impurities for both 1 and
1-6D.
Discussion
Autoxidation during Isomerization of 1. In order to accurately
determine KIE, it is imperative that the reactions lead cleanly
to only one product. Oxidation during the isomerization of 1 is
a potential complicating reaction that must be considered.
Autoxidation of unsaturated 3-keto steroids was first described
in 1961 by Camerino et al., who observed the base-catalyzed
oxidation of R,ꢀ-unsaturated systems at C-4 and C-6.49 Ringold
and Malhotra subsequently detected the uncatalyzed autoxidation
and isomerization of purified ꢀ,γ-unsaturated 3-keto steroids
upon standing at room temperature.50 Finally, de la Mare et al.
observed the acid-catalyzed autoxidation of ꢀ,γ-unsaturated
3-keto steroids to form a number of C-6 oxidized products.51
Dienol(ate)s are thought to be the intermediates in all of these
reactions.49,51,52
We observed large quantities (15-25%) of the oxidation
product 4-androstene-3,6,17-trione (11), as well as other side
products, during both the acetate- and hydroxide-catalyzed
isomerizations of 1. Support for oxygen as a reactant in this
process is provided by the reduction in side-product formation
when 1 is incubated with hydroxide under argon. The fact that
oxidation side products are not completely eliminated is likely
due to incomplete oxygen removal, as observed previously.50
A more effective method of avoiding oxidative side-product
formation, which was employed by de la Mare et al.,51 involves
addition of the metal-chelating agent EDTA. We find 1 mM
EDTA to be sufficient to reduce formation of 11 to <1% of
the reaction products, giving clean conversion of 1 to 3 and
allowing the observation of first-order isomerization kinetics
from which KIEs can be determined.
In addition, both the direction (inverse versus normal) and
magnitude of the isotope effects were determined by a competi-
tive method for D99A. Mass spectral measurement of the KIE
by ESI-MS, by using a mixture of 1 and 1-6D, confirmed that
the KIE for the D99A reaction is normal and small (1.07 (
0.03), in agreement with the kinetic results by UV (1.059 (
0.014), validating the noncompetitive UV data. Because of the
larger errors of the mass spectrometric method, only data from
the UV method were used to calculate the reported KIEs.
Because of the slow rate of the acetate-catalyzed isomerization
(k ≈ 3 × 10-5 s-1 at the lowest acetate concentration), con-
tributions from the uncatalyzed isomerization reaction must be
considered for this system. By using data recently obtained21
in acetate buffer (pH 5.5), the uncatalyzed rate constant is
calculated to be ∼1 × 10-6 s-1, or 3% of the slowest acetate
rate. The observed KIE could therefore contain a small53
component due to a KIE on the uncatalyzed reaction. However,
an increase in the acetate concentration by 3-fold (thereby
decreasing the uncatalyzed contribution to 1%) does not
change the observed KIE, indicating that the contribution of
the uncatalyzed reaction to the observed KIE is small. Because
(47) Shoemaker, D. P.; Garland, C. W.; Nibler, J. W. Experiments in
Physical Chemistry, 6th ed.; McGraw-Hill: New York, 1996.
(48) Xue, L. A.; Talalay, P.; Mildvan, A. S. Biochemistry 1991, 30, 10858–
10865.
(49) Camerino, B.; Patelli, B.; Sciaky, R. Tetrahedron Lett. 1961, 16, 554–
559.
(50) Ringold, H. J.; Malhotra, S. K. Tetrahedron Lett. 1962, 15, 669–672.
(51) de la Mare, P. B. D.; Wilson, R. D. J. Chem. Soc., Perkin Trans. II
1977, 157–162.
(53) By assuming that the KIE of the uncatalyzed reaction is within 15%
of the observed KIE, the contribution from the uncatalyzed rate is
less than 0.5%.
(52) Frimer, A. A.; Gilinsky-Sharon, P.; Hameiri, J.; Aljadeff, G. J. Org.
Chem. 1982, 47, 2818–2819.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6583

A R T I C L E S
Wilde et al.
Scheme 5
the enzymatic rates are ca. 1000-fold faster than the buffer rate,
buffer contribution to the observed enzymatic KIEs is negligible.
Finally, it should be noted that this work provides a much
more precise knowledge of the kH / kD ratios (Table 2) than of
the kH and kD values themselves (Table 1). Although reproduc-
ibility of the observed KIEs is generally 0.5-1.5%, the variation
in individual rate constants is typically 4-5%. A similar result
was observed previously by Vitullo et al.37 and is likely due to
slight differences in temperature and solution from run to run
or day to day. Based upon the above analysis, we are confident
that the KIEs in Table 2 are accurate, although they are
somewhat higher than the values of 1.00 ( 0.0136 and 1.02 (
0.0248 for WT and Y14F, respectively, previously reported by
Xue et al. by using initial rate kinetics, a method that is more
prone to error.
In this case, the C-H bending mode is partially lost at the TS,
leading to an inflated isotope effect. By using this model, CM
between the transferred proton (from Asp-38 or acetic acid) and
the R-hydrogen at C-6 could account for an isotope effect larger
than that expected from an analysis of hybridization changes.
Computational studies9,61,63 modeling hydrogen transfers to and
from carbon indicate that semiclassical CM (without HT) can
lead to secondary KIE inflations on the order of 3-4%.
Therefore, CM without HT could account for part, but not all,
of the KIEs observed in Table 2 (1.031-1.075, H transfer).
However, much larger increases of secondary KIEs (up to 25%)
can be observed when CM is combined with tunneling of the
transferred hydrogen,9,62,63 suggesting that these reactions may
include a tunneling component.
One probe for the existence of HT makes use of the fact that
a particle’s ability to tunnel is inversely related to its mass. Thus,
replacement of the transferring proton with a deuteron decreases
the contribution of tunneling to the reaction, causing a smaller
inflation in the secondary KIE62 if CM/HT is significant. This
method has been used to provide evidence for CM/HT in the
conversion of formate to carbon dioxide by yeast formate
dehydrogenase. This reaction has an extremely inflated second-
ary KIE (1.22) relative to the EIE (0.89).64 When deuterium is
transferred, however, the secondary KIE drops to 1.07, indicat-
ing the existence of CM/HT in this system. Similarly, the
reaction of glutamase mutase with [5′-3H]-adenosylcobalamin
and L-glutamate shows a reduced secondary KIE when deute-
rium is abstracted from glutamate, which was interpreted in
terms of CM/HT.65 In contrast, the reaction catalyzed by enoyl-
CoA hydratase (crotonase), which also exhibits an inflated
KIE,16 shows no change in the secondary KIE upon replacement
of the transferred hydrogen with deuterium,66 indicating the
absence of CM/HT.
Table 2 gives the results of the secondary KIE for transfer
of a hydrogen and transfer of a deuterium in the isomerization
of 1 for catalysis by acetate, WT KSI, and the mutants, along
with the corresponding standard deviations. Each KIE is the
average of 9-18 individual measurements. The appropriate
statistical tool to determine whether there are significant
differences between the isotope effects for the transfer of H
and of D for a given catalyst is the t test. The results of this
analysis are given in the last column of Table 2. For WT, the
likelihood that the isotope effects for hydrogen transfer and
deuterium transfer are different is >98% (3.3% methanol) and
>99.5% (20% methanol). For D99A, it is also >99.5%, whereas
for Y14F, it is >90%. For acetate catalysis, the t test was not
carried out, because the KIEs are 1.031 and 1.030, respectively.
(Ab)normal r-Secondary KIEs. In this work, we have isolated
the intrinsic secondary KIE for the protonation of the intermedi-
ate with a deuterium at C-6 (Scheme 1, TS2) by correcting the
observed effects to account for the extent to which the isotope-
dependent step is rate-limiting. For both the enzymatic (WT)
and nonenzymatic reactions, the protonation step (TS2) is
stereospecific, with protonation occurring at the ꢀ-face of the
steroid.41 Thus, the TSs for these reactions are comparable.
Secondary hydrogen isotope effects are often used to esti-
mate the structure of TSs of both solution43,54 and enzymatic
reactions.55–57 These secondary isotope effects normally lie
between unity and the EIE. Isotope effects near unity indicate
a substrate-like TS, and isotope effects approaching the EIE
suggest a product-like TS. Hybridization changes are generally
considered to dominate isotope effects for a hydrogen attached
to a carbon undergoing a change in hybridization.58 Both
calculations and experimental evidence give R-secondary EIEs
for deuterium bound to a carbon changing from sp2 to sp3
hybridization of about 0.89.56,58–60 On the basis of hybridization
changes, C-6 deuterium KIEs between 1.00 and 0.89 are
expected for the isomerization of 1-6D(TS2), where a KIE close
to 1.00 would indicate an intermediate (2)-like TS, and a KIE
closer to 0.89 would suggest a product (3)-like TS. In contrast
to this expectation, however, the observed values for all of the
systems investigated are greater than unity, inconsistent with a
model based on hybridization changes. These values, then, must
be interpreted in terms of a different model.
CM/HT. Inflated isotope effects in hydrogen-transfer reac-
tions, such as those observed here, are diagnostic of CM or
CM combined with HT.9,61,62 CM occurs when one of the
vibrations of the R-hydrogen is part of the reaction coordinate
(Scheme 5), leading to an increase in the KIE as part of the
R-hydrogen’s vibrational mode is converted to a translation.62
(54) Collins, C. J., Bowman, N. S., Eds. Isotope Effects in Chemical
Reactions; Van Nostrand Reinhold: New York, 1970.
(55) Cleland, W. W., O’Leary, M. H., Northrop, D. B., Eds. Isotope Effects
on Enzyme-Catalyzed Reactions; University Park: Baltimore, 1977.
(56) Cleland, W. W. In Isotopes in Organic Chemistry: Secondary and
SolVent Isotope Effects; Buncel, E., Lee, C. C., Eds.; Elsevier Science:
Amsterdam, 1987; Vol. 7, pp 61–110.
(57) Cook, P. F., Ed. Enzyme Mechanism from Isotope Effects; CRC: Boca
Raton, FL, 1991.
(58) Matsson, O.; Westaway, K. C. AdV. Phys. Org. Chem. 1998, 31, 143–
248.
(59) Cook, P.; Blanchard, J. S.; Cleland, W. W. Biochemistry 1980, 19,
4853–4858.
(60) Hartshorn, S. R.; Shiner, V. J., Jr. J. Am. Chem. Soc. 1972, 94, 9002–
9012.
(63) Rucker, J.; Klinman, J. P. J. Am. Chem. Soc. 1999, 121, 1997–2006.
(64) Cook, P. F.; Oppenheimer, N. J.; Cleland, W. W. Biochemistry 1981,
20, 1817–1825.
(61) Saunders, W. H. J. J. Am. Chem. Soc. 1985, 107, 164–169.
(62) Klinman, J. P. In Enzyme Mechanism from Isotope Effects; Cook, P.,
Ed.; CRC: Boca Raton, FL, 1991, pp 127–147.
(65) Cheng, M. C.; Marsh, E. N. Biochemistry 2007, 46, 883–9.
(66) Bahnson, B. J.; Anderson, V. E. Biochemistry 1991, 30, 5894–5906.
9
6584 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Hydrogen Tunneling in 3-oxo-∆5-Steroid Isomerase
A R T I C L E S
Thus, WT and both mutants exhibit decreases in the R-secondary
deuterium KIE from 1.06-1.08 for hydrogen transfer to
1.03-1.04 for deuterium transfer, which is consistent with CM/
HT.
In contrast to the enzymatic reactions, the acetate-catalyzed
isomerization shows no change in KIE with deuterium transfer,
which is consistent with no CM/HT as part of the solution
reaction. These results indicate that CM/HT is occurring in the
KSI reaction, in contrast to the model acetate reaction, for which
there is no evidence for CM/HT.
Previous Report of Tunneling with KSI. Xue et al.36 previ-
ously interpreted a decrease in C-4R secondary isotope effects
from 1.11 ( 0.02 (for substrate deuterated at C-4R) to 1.06 (
0.01 (for substrate deuterated at both the C-4R and ꢀ positions)
for the KSI-catalyzed isomerization of 1 in terms of a tunneling
contribution to the enolization step. However, this interpretation
relies on the assumption that KSI-catalyzed proton abstraction
is 100% stereospecific for C-4ꢀ. Unfortunately, this assumption
is not entirely valid; in fact, KSI abstracts the C-4R proton ∼5%
of the time when both C-4 positions are protonated.41 Therefore,
deuteration at C-4R decreases the rate of the reaction by about
4% because of a primary KIE (assuming a primary effect of 5),
similar to the amount of KIE inflation observed. Furthermore,
later work28 demonstrated that enolization is not cleanly rate-
limiting, as assumed by Xue et al.,36 which introduces further
complications in the interpretation of these data. For these
reasons, these results cannot be used as evidence for HT during
enolization by KSI.
Normal KIE for KSI in the Absence of Tunneling. Although
a contribution from tunneling can explain part of the inflated
KIEs for the enzymatic proton transfer, all of the R-secondary
KIEs are still normal (∼1.03-1.04) when a deuterium is being
transferred (Table 2), which is still inconsistent with predictions
from considerations of hybridization changes by using an
unconjugated alkene as a model for the intermediate 2 (∼0.89).
These results are reminiscent of the observation by Anderson67
of an inflated R-secondary KIE for crotonase that is not
decreased when deuterium is transferred instead of hydrogen.66,67
Anderson suggested that EIEs for enolate formation may be
larger (more normal) than for simple alkenes because of looser
vibrational modes of the C-H bond in enolates than in simple
alkenes. If this is the case, a normal KIE might be expected for
protonation of an enolate ion, such as 2. Alternatively, a
contribution to the KIE from CM without tunneling could also
lead to an increase in the KIE. As noted above, CM without
tunneling has been calculated to lead to secondary KIEs
inflations of up to 3-4%.9,61,63
Figure 2. Diagram illustrating the increase in coupling for the KSI reaction
relative to the acetate reaction. For simplicity, acetaldehyde is used as a
model carbonyl compound.
transfer lags well behind rehybridization, leading to a TS in
which the carbon is almost completely sp3 hybridized and in
which there is little resonance stabilization of the partial negative
charge on the enolate carbon by the carbonyl group. In the
reverse direction (ketonization), most of the resonance stabiliza-
tion must be lost before the TS is reached. Thus, the TS is of
higher energy because of the minimal resonance stabilization,
and these reactions are slower than one might expect given the
equilibrium acidity of the carbonyl compounds. The asynchro-
nicity of the two processes, gain (or loss) of resonance
stabilization and proton transfer, indicates minimal, if any, CM
of the secondary proton and the transferring proton in the case
of acetate catalysis. On the other hand, the observed coupling
in the KSI reaction suggests that loss of resonance stabilization
of the enolate ion and proton transfer are more concerted than
in the acetate reaction (Figure 2), leading to a TS with greater
resonance stabilization than that of the TS for acetate catalysis.
Acknowledgment. This research was supported by a grant from
the National Institutes of Health (5RO1GM038155). We thank
Professor Judith Klinman for helpful discussions.
Supporting Information Available: Derivation of equations
for KIE analysis by MS. This material is available free of charge
via the Internet at http://pubs.acs.org.
Mechanistic Implications of Enhanced CM/HT. The finding
of CM for KSI stands in contrast to the well-documented
asynchronicity in nonenzymatic enolization/ketonization reac-
tions.68–71 In simple base-catalyzed enolate formation, proton
JA0732330
(69) Bernasconi, C. Acc. Chem. Res. 1992, 25, 9–16.
(70) Bernasconi, C. F.; Wenzel, P. J. J. Am. Chem. Soc. 1994, 116, 5405–
5413.
(67) Anderson, V. E. In Enzyme Mechanisms from Isotope Effects; Cook,
P. F., Ed.; CRC: Boca Raton, FL, 1991, pp 389–417.
(68) Bernasconi, C. AdV. Phys. Org. Chem. 1992, 27, 119–238.
(71) Saunders, W. H., Jr J. Am. Chem. Soc. 1994, 116, 5400–5404.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6585