pH-dependence in the hydrolytic action of the human fragile histidine triad

Article abstract of DOI:10.1002/ejoc.200500499

The human fragile histidine triad protein (Fhit) is a member of the HIT family of enzymes, which catalyze hydrolysis or nucleotidyltransfer reactions of dinucleoside polyphoshates. Fhit catalyzes the magnesium ion-dependent hydrolysis of P^1-5′-O-adenosine-P^3-5′-O- adenosine triphosphate (Ap₃A) to adenosine-5′-O-phosphate (AMP) and the magnesium complex of adenosine-5′-O-diphosphate (ADP) by a double displacement mechanism, with the formation of an adenylyl enzyme as an intermediate. Fhit also catalyzes the hydrolysis of adenosine-5′- phosphoimidazolide (AMP-Im) and adenosine-5′-phospho-N-methylimidazolide (AMP-N-MeIm). The pH-dependence of these reactions is reported herein, in which the principal conclusions are as follows: The action of wild-type Fhit on MgAp₃A is diffusion-limited and on AMP-Im and AMP-N-MeIm largely diffusion-limited and largely pH-independent. The actions of specifically mutated H96G-Fhit on AM-Im, and on AMP-N-MeIm, are not diffusion-limited and are pH-dependent. The actions of mutated forms of Fhit, H94G-Fhit, H98G-Fhit, and H94/98G-Fhit, are also not diffusion-limited and are pH-dependent. Log plots of kinetic parameters against pH show breaks that indicate a group on the enzyme must be protonated for maximal activity. Extensive analysis shows that the imidazole ring of His94 is not essential for the hydrolysis of MgAp₃A or AMP-imidizolides, and the imidazole ring of His98 engages in binding the substrates. In the hydrolysis of AMP-Im, Fhit and its His94, His96, and His98 variants bind the monoanionic form of AMP-Im, and the proton required for formation of imidazole in the hydrolytic process originates with an acid/base group of the enzyme. Fhit and several variants also catalyze the hydrolysis of p-nitrophenyl-AMP. Wiley-VCH Verlag GmbH & Co. KGaA, 2005.

Full text of DOI:10.1002/ejoc.200500499

FULL PAPER
DOI: 10.1002/ejoc.200500499
pH-Dependence in the Hydrolytic Action of the Human Fragile Histidine Triad
Kaisheng Huang,[a] Abolfazl Arabshahi,[a] and Perry A. Frey*[a]
Dedicated to Professor Wojciech Stec on the occasion of his 65th birthday
Keywords: Enzymes / Hydrolases / Enzyme catalysis / Hydrolysis / Kinetics
The human fragile histidine triad protein (Fhit) is a member
of the HIT family of enzymes, which catalyze hydrolysis or
nucleotidyltransfer reactions of dinucleoside polyphoshates.
Fhit catalyzes the magnesium ion-dependent hydrolysis of
P1–5Ј-O-adenosine-P3–5Ј-O-adenosine triphosphate (Ap3A)
to adenosine-5Ј-O-phosphate (AMP) and the magnesium
complex of adenosine-5Ј-O-diphosphate (ADP) by a double
displacement mechanism, with the formation of an adenylyl
enzyme as an intermediate. Fhit also catalyzes the hydrolysis
of adenosine-5Ј-phosphoimidazolide (AMP-Im) and adeno-
sine-5Ј-phospho-N-methylimidazolide (AMP-N-MeIm). The
pH-dependence of these reactions is reported herein, in
which the principal conclusions are as follows: The action of
wild-type Fhit on MgAp3A is diffusion-limited and on AMP-
Im and AMP-N-MeIm largely diffusion-limited and largely
pH-independent. The actions of specifically mutated H96G-
Fhit on AM-Im, and on AMP-N-MeIm, are not diffusion-lim-
ited and are pH-dependent. The actions of mutated forms of
Fhit, H94G-Fhit, H98G-Fhit, and H94/98G-Fhit, are also not
diffusion-limited and are pH-dependent. Log plots of kinetic
parameters against pH show breaks that indicate a group on
the enzyme must be protonated for maximal activity. Exten-
sive analysis shows that the imidazole ring of His94 is not
essential for the hydrolysis of MgAp3A or AMP-imidizolides,
and the imidazole ring of His98 engages in binding the sub-
strates. In the hydrolysis of AMP-Im, Fhit and its His94,
His96, and His98 variants bind the monoanionic form of
AMP-Im, and the proton required for formation of imidazole
in the hydrolytic process originates with an acid/base group
of the enzyme. Fhit and several variants also catalyze the
hydrolysis of p-nitrophenyl-AMP.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
UDPGlc + E-His166 h E-His166_UMP + Glc–1–P
(1)
Introduction
The fragile histidine triad (Fhit) protein is encoded by
the human gene FHIT and, presumably by way of a signal-
ing activity, functions as a tumor suppressor.[1] Tumor cells
and cells in tumor-derived lines display deletions or ab-
errant transcripts of FHIT.[2–7] Fhit is a dimer of identical
20 kDa subunits, each of which contains a triad of histidine
residues, His94-Val95-His96-Val97-His98 that is conserved as
HxHxHxx in other species, with the residues “x” being
hydrophobic. The structures of Fhit and other histidine
triad (HIT) proteins reveal an overall chain fold for the di-
mer that is similar to the fold of a single subunit in galac-
tose-1-P uridylyltransferase (GalT) from E. coli.[8–13] GalT
catalyzes the reaction of galactose-1-phosphate (Gal-1-P)
with uridine-5Ј-diphosphate glucose (UDPGlc) to form uri-
dine-5Ј-diphosphate galactose UDPGal and glucose-1-
phosphate (Glc-1-P) by a double displacement mechanism
[Equation (1) and Equation (2)].
E-His166-UMP + Gal-1-P h E-His166 + UDPGal
(2)
The nucleophilic catalyst His166 is the central residue in a
conserved triad of two histidines and one glutamine, His164
Pro165-His166-Gly167-Glu168 [14–24]
-
.
As an enzyme, Fhit catalyzes the Mg2+-dependent hy-
drolysis of P1–5Ј-O-adenosine-P3–5Ј-O-adenosine triphos-
phate (Ap3A) to adenosine-5Ј-O-phosphate (AMP) and the
Mg2+ complex of adenosine-5Ј-O-diphosphate (ADP) by
cleavage of the Pα–O bond and not the Pβ–O bond.[25–27]
The hydrolysis of a P-chiral analog of Ap3A in H218O pro-
ceeds with overall retention of configuration at P, consistent
with a double displacement mechanism of hydrolysis, as in
the reaction of GalT.[28] The apparent similarities of the
active site triads in Fhit and GalT, and the similar chain
folds of the two enzymes inspired the proposition of similar
mechanisms for the two reactions. According to this hy-
pothesis, the action of Fhit also follows a double displace-
ment mechanism, in which the ultimate adenylyl group ac-
ceptor is water [Equation (3) and Equation (4)].
[a] Department of Biochemistry, University of Wisconsin – Madi-
son,
1710 University Avenue, Madison, Wisconsin 53726, USA
E-mail: frey@biochem.wisc.edu
5198
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5198–5206

pH-Dependence in the Hydrolytic Action of the Human Fragile Histidine Triad
FULL PAPER
Ap3A + EFhit-His96 h EFhit-His96-AMP + ADP
EFhit-His96-AMP + H2O h EFhit-His96 + AMP
(3)
(4)
structure, the upstream histidine main-chain carbonyl
group is in position to accept a hydrogen bond from the
imidazole ring of the nucleophilic histidine in the covalent
nucleotidyl enzyme intermediate, as illustrated in Figure 1.
Acting in this capacity, the main-chain carbonyl group
could maintain the protonated, imidazolium moiety in the
covalent intermediate. This should potentiate the nucleoti-
dyl transfer activity of the intermediate because the neutral
species would be chemically far less reactive. A nucleoside-
5Ј-phosphoimidazolide is not basic enough to exist in the
imidazolium form at physiological pH; for example the
value of pKa for the imidazolium group of UMPIm is
5.68.[30]
Evidence supporting this mechanism includes the obser-
vation of a trace of protein-bound 32P when Fhit is mixed
with the poor substrate [α-32P]ATP (adenosine-5Ј-O-tri-
phosphate), the observation of 11% adenylylation of Fhit
by [8,8Ј-3H]Ap3A based on tritium incorporation, the ob-
servation of Fhit-catalyzed exchange of [3H]ADP into
Ap3A, and the observation of overall retention of configu-
ration at phosphorus consistent with double inversion in
two steps.[27–29]
The mechanism in Figure 1 cannot be tested by X-ray
crystallography until sufficiently high resolution is attained
to obtain images of protons in the GalT-His166-UMP inter-
mediate. We have undertaken to study this mechanism kin-
etically in chemically rescued reactions of the specifically
mutated Fhit, Fhit-H96G. We here report our studies of
pH-dependence in the actions of H96G-Fhit, wild-type
Fhit, and other variants of Fhit on Ap3A, adenosine-5Ј-
phosphoimidazolide (AMP-Im) and adenosine-5-phospho-
N-methylimidazolide (AMP-N-MeIm).
The crystal structures of GalT and Fhit strongly suggest
that nucleotidyl group transfer catalyzed by His166 in GalT
and His96 in Fhit proceeds with participation of the main-
chain carbonyl group of the conserved histidine two resi-
dues upstream, His164 in GalT and His94 in Fhit. In each
Results
Catalytic Properties of H96G-Fhit
The phosporamidase activity of H96G-Fhit toward
AMP-Im corresponds to the chemical rescue of the vari-
ant.[29] The complex of H96G-Fhit with AMP-Im should
be analogous to the covalent intermediate EFhit-His96–AMP
in Equation 3 and Equation 4 (2 in Figure 1), in that the
imidazole ring of AMP-Im should occupy the vacated subs-
ite for the imidazole ring of His96, and the AMP moiety
should occupy the same binding site as in the covalent in-
termediate. Then H96G-Fhit catalyzes the hydrolysis of
AMP-Im, as in the hydrolysis of EFhit-His96–AMP in Equa-
tion (4) (2 to 3 in Figure 1). The kinetic parameters kcat and
kcat/Km for hydrolysis of AMP-Im by the variant are about
1/50th those for the hydrolysis of Ap3A by wild-type
Fhit.[29] The parameters for the action of wild-type and
H96G-Fhit are listed in Table 1, together with values for
other variants of Fhit.
H96G-Fhit also catalyzes the hydrolysis of AMP-N-
MeIm at a maximum rate that is comparable to the value
for the hydrolysis of AMP-Im.[29] The value of Km in the
Figure 1. A mechanism for Fhit-catalyzed hydrolysis of MgAp3A.
Table 1. Kinetic parameters for hydrolysis of substrates by Fhit and selected variants.[a]
Substrate
Enzyme
Ap3A
kcat
Ap3A
kcat/Km
AMP-Im
AMP-Im
kcat/Km
AMP-N-MeIm
AMP-N-MeIm
kcat/Km
[a]
kcat
kcat
wt-Fhit[b]
H96G[b]
H94G
H98G
H94/98G
3.5 .1
ND
0.48 .01
0.07 .01
0.05 .01
1160 72
ND
1083 48
0.47 .02
0.27 .01
4.4 .1
642 16
35 1
737 33
11.3 .1
2.1 .1
4.2 .1
0.62 .06
–
0.24 .02
0.07 .03
717 25
1.9 .01
–
0.91 .03
0.57 .02
0.51 .01
0.06 .01
31
1
9.7 .6
[a] Kinetic parameters are kcat [s–1] and kcat/Km [m–1 s–1]. [b] Previously reported in ref.[29]
Eur. J. Org. Chem. 2005, 5198–5206
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5199

K. Huang, A. Arabshahi, P. A. Frey
FULL PAPER
hydrolysis of AMP-N-MeIm is significantly higher than in ionization of an enzyme-substrate complex, and that for
the hydrolysis of AMP-Im, presumably because of the kcat/Km refers to an ionization of either the free enzyme or
greater steric requirements of the N-methylimidazole group the free substrate. The ionizations governing the pH-rate
relative to imidazole. The equivalent maximum velocities profiles cannot be assigned to AMP-Im (Scheme 1) because
for the two substrates suggest that the rate-limiting step is the value of pKa for the phosphoimidazolide is 5.6 (see Ex-
the same for AMP-N-MeIm and AMP-Im within the Mi- perimental Section). Therefore, the pKa values in Figure 1A
chaelis complexes.
and B must correspond to ionizations of the complex
H96G-Fhit·AMP-Im and free H96G-Fhit, respectively.
Moreover, the values of Km for AMP-Im are essentially pH-
independent (data not shown), so that the enzyme must be
binding the anion of AMP-Im shown on the right side of
Scheme 1.
pH-Dependence of Hydrolysis by H96G-Fhit
To address the question of the protonation state of the
imidazole moiety in the nucleotidyl intermediate EFhit
-
His96–AMP in wild-type Fhit, we studied the pH-depen-
dence for the action of Fhit-H96G on AMP-imidazolides.
The pH-rate profiles for the hydrolysis of AMP-Im in Fig-
ure 2 show that both kcat and kcat/Km are pH-dependent,
with breaks in plots of log kcat and log kcat/Km against pH,
with high activity at pHs below 7 and much lower activity
at high pHs. The profiles are compatible with the require-
ment that a single ionizing group in its conjugate acid form
potentiates optimal enzymatic activity. The values of pKa
for the principal ionizations controlling the rate are 7.06 for
kcat and 6.87 for kcat/Km. The value for kcat refers to an
Scheme 1.
In the hydrolysis of AMP-Im, the imidazole ring must
acquire a proton to form free imidazole (Scheme 2). While
this proton ultimately arises from the water that is cleaved
in the overall hydrolytic process, proton transfer is presum-
ably mediated by an acid/base group of the enzyme. This
group may be the one that must be protonated for optimal
activity. Obvious candidates are histidine residues in the his-
tidine triad, which include His94 and His98 in H96G-Fhit.
The imidazole ring of His94 is not in the active site (Fig-
ure 1), but that of His98 is in contact with the substrate and
might serve as a proximal proton source in the hydrolysis
of AMP-Im.
Scheme 2.
If the phosphoimidazolide groups of AMP-Im and the
covalent intermediate EFhit-His96-AMP react in their pro-
tonated states as in Figure 1, then the pH-rate profile for
the H96G-Fhit-catalyzed hydrolysis of AMP-N-MeIm
should differ from that for the hydrolysis of AMP-Im. This
is because the imidazole ring in AMP-Im is not protonated
in neutral solution (pKa = 5.6) and the N-methyl group in
AMP-N-MeIm maintains the quaternary nitrogen in the
imidazole ring and therefore does not require a proton to
react. Scheme 3 illustrates the hydrolysis of AMP-N-MeIm.
In effect, the N-methyl group serves as a surrogate proton
in the hydrolysis.
Figure 2. pH-Rate profiles for hydrolysis of AMP-Im by H96G-
Fhit. Values of kcat and Km were determined in the pH range 5.5–
8.5 at 26.5 °C as described in the Experimental Section. Panel A
depicts a plot of log kcat against pH. The solid line is calculated
from the fit of data points to Equation (6), resulting in a value of
pKa = 7.06 0.03. Panel B depicts a plot of log kcat/Km against pH.
The fit of data points to Equation (6) yields a value of 6.87 0.03
for pKa, and the solid line is calculated for this value.
Scheme 3.
5200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5198–5206

pH-Dependence in the Hydrolytic Action of the Human Fragile Histidine Triad
FULL PAPER
The pH-rate profiles in Figure 3A and B for the hydroly- of log kcat/Km against pH is more complex, bell-shaped, and
sis of AMP-N-MeIm are unlike those in Figure 2 for the shows two breaks corresponding to pKa-values of 5.72 and
reaction of AMP-Im. The profiles for kcat and kcat/Km show 7.84. The MgAp3A does not undergo ionizations in this pH
two breaks, with identical values of 5.3 for the pKa on the range, so that the transitions in Figure 4B must represent
acid side and 7 on the alkaline side. The higher value of ionizations of the free enzyme (H96G-Fhit).
pKa is likely to correspond to the same ionization as in the
hydrolysis of AMP-Im. The lower pKa must represent the
ionization of a group on H96G-Fhit, in which the conjugate
acid state is less active than the unprotonated state. The
results show that the enzyme can be “overprotonated” for
the hydrolysis of AMP-N-MeIm, presumably because of the
charge type of AMP-N-MeIm as a zwitterion brought
about by the N-methyl group.
Figure 4. pH-Rate profiles for hydrolysis of Ap3A in 0.55  imid-
azole by H96G-Fhit. Panel A is a plot of log kcat against pH at
26.5 °C in the pH range of 6.0–8.9. The data were fitted to Equa-
tion (5) and gave 7.09 0.03 as the value of pKa. The solid line is
calculated for a pKa value of 7.09. Panel B is a plot of log kcat/Km
against pH, with the data points fitted to Equation (8). The two
values of pKa were 5.72 0.18 and 7.84 0.03, and the solid line is
calculated for these values.
Figure 3. pH-Rate profiles for hydrolysis of AMP-N-MeIm by
H96G-Fhit. Panel A is a plot of log kcat against pH for rates mea-
sured at 26.5 °C as described in the Experimental Section. The solid
line represents the fit of data points to Equation (8) with two values
pH-Dependence of Wild-Type Fhit
of pKa, =5.35 0.08 and 7.43 0.03. Panel B is a plot of log kcat
Km against pH with the data fitted to Equation (7). The solid line
is calculated based on 5.35 and 7.42 as the values of pKa.
/
Until recently, the best substrate for wild-type Fhit has
been MgAp3A. However, the wild-type enzyme also cata-
lyzes the hydrolysis of AMP-Im and AMP-N-MeIm equally
While H96G-Fhit alone does not catalyze the hydrolysis as well, the difference being that the magnesium require-
of MgAp3A, it is chemically rescued by free imidazole and ment for action on Ap3A is absent in the hydrolysis of
catalyzes
the
imidazole-dependent
hydrolysis
of AMP-imidazolates.[29] The similarities in the kinetic param-
MgAp3A.[29] The pH-dependence of this reaction at 0.55  eters for the three substrates extend to the pH-dependence,
imidazole is shown in Figure 4. In Panel A, the plot of log as shown in Figure 5. Two facts stand out in the pH-rate
kcat against pH shows increasing rate with increasing pH, a profiles. First, the bell-shaped profiles are shallow and do
slope of +1, and a break corresponding to pKa = 7.09: This not approach slopes of +1 and –1 at low and high pHs.
is the value of pKa for imidazole, and the rate-dependence Second, the rate constants are remarkably similar for the
on pH is assigned to the reaction of the enzyme–substrate three substrates. Both of these facts suggest that the rate-
complex with free imidazole, which must be in the form of limiting process may be similar for these substrates, and
its conjugate base to react in the chemical rescue. The plot that diffusion may be rate limiting.
Eur. J. Org. Chem. 2005, 5198–5206
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5201

K. Huang, A. Arabshahi, P. A. Frey
FULL PAPER
Km are determined in the presence and absence of a vis-
cogen at several relative viscosities ηrel The parameters kao
and ka are defined as the values of kcat/Km in the absence
and presence of viscogens, respectively, and the ratios kao/ka
are plotted against ηrel When diffusion is not rate limiting
the slope of this plot is zero, and when diffusion is fully rate
limiting the slope is +1. Figure 6 shows that the hydrolysis
of MgAp3A by the action of wild-type Fhit is fully diffusion
controlled, and the hydrolyses of AMP-Im or AMP-N-
MeIm by action of H96G-Fhit are not at all limited by
diffusion. The hydrolyses of AMP-Im or AMP-N-MeIm by
wild-type Fhit are significantly diffusion limited as well.
Therefore, diffusion limitation explains the shallow pH-de-
pendencies and similarities in kinetic parameters in Fig-
ure 5. In contrast, diffusion does not limit rates in the ac-
tions of H96G-Fhit.
Importance of His94 and His98
The conserved histidine residues in the HIT enzymes are
not equally important in catalysis, as shown by the kinetic
parameters in Table 1. While His96 is essential in the hy-
drolysis of Ap3A, His94 and His98 are less important. Mu-
tation of His94 to glycine has very little effect on the values
of kcat/Km in the hydrolysis of MgAp3A or AMP-Im, and
values of kcat are lower by seven- or eightfold in the action
of H94G-Fhit relative to the wild-type enzyme. This may
be attributed to the absence of the imidazole ring of His94
in the active site: the main chain carbonyl group of His94
is in hydrogen bonded contact with the imidazole ring of
His96 (Figure 1).
Figure 5. pH-rate profiles for hydrolysis of AMP-Im, AMP-N-
MeIm, and Ap3A by wild-type Fhit. Panel A shows plots of log
kcat against pH at 26.5 °C for the hydrolysis of AMP-Im (◊), AMP-
N-MeIm (◊) and Ap3A (᭡). Panel B shows plots of log kcat/Km
against pH for AMP-Im (◊) and AMP-N-MeIm (◊).
Mutation of His98 brings about much greater decreases
in kinetic parameters. Values of kcat and kcat/Km for H98G-
Fhit are 1/50th and 1/2500th the values for wild-type Fhit,
respectively. H98G-Fhit is also much less active against
AMP-Im and AMP-N-MeIm than the wild-type enzyme,
17 and 30-fold less active in terms of kcat. A difference from
the reaction of Ap3A is that the values of kcat/Km are only
20- to 50-fold lower for H98G-Fhit. Double mutation of
His94 and His98 further degrades the activity, especially
against the AMP-phosphoramidates.
To assess the impact of diffusion on the rate constants,
the effects of a viscogen on the rate constants were deter-
mined by the method of Brower and Kirsch,[34] with the
results shown in Figure 6. In this method, the values of kcat
/
The pH-rate profiles for the actions of H94G-, H98G-,
and the double variant H94/98G-Fhit with the three sub-
strates are shown in Figure 7, panels A to D. The data for
hydrolysis of AMP-Im by H94G- and H94.98G-Fhit in
panels A and B are similar in displaying one break and a
descending limb with increasing pH corresponding to pKa-
values of 6.8 to 7.9. In this respect, the profiles are analo-
gous to those for H96G-Fhit in Figure 2. The data are con-
sistent with a requirement for an acid/base group in its pro-
tonated form in the hydrolysis of AMP-Im. The results for
H98G-Fhit are different (part b of panel A), and the log
kcat data are not fitted to an equation. The log kcat/Km plot
is more analogous to the others, and this will be explained
in a later section.
Figure 6. Viscosity effects on hydrolysis by H96G-Fhit and wild-
type Fhit. Hydrolyses of AMP-Im, AMP-N-MeIm and Ap3A were
studied as a function of viscosity at 26.5 °C. The kinetic data are
o
plotted as ka = kcat/Km, ka = ka (ηrel = 1). (◊) Hydrolysis of AMP-
Im by H96G-Fhit, (◊) hydrolysis of AMP-N-MeIm by H96G-Fhit,
(᭡) hydrolysis of AMP-Im by wild-type Fhit, (∆) hydrolysis of
AMP-N-MeIm by wild-type Fhit, (᭢) hydrolysis of Ap3A by wild-
type Fhit.
The pH-rate profiles for the hydrolysis of AMP-N-MeIm
in panels C and D show little or no pH-dependence for kcat
5202
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5198–5206

pH-Dependence in the Hydrolytic Action of the Human Fragile Histidine Triad
FULL PAPER
in the actions of H98G- and H94/98G-Fhit. They further
confirm and extend the observation in Figure 3 that the
ЈЈoverprotonationЈЈ of the enzyme can inhibit the hydrolysis
of AMP-N-MeIm. The pKa-value for the enzymatic group
that must be unprotonated is 5.69 (curve b, panel D).
Role of His98
Curves b in Figure 7A and B suggest a role for His98 in
the hydrolysis of AMP-Im. This is clarified by considering
the pH-dependencies of Km in the hydrolysis of the two
AMP-phosphoramidates. Values of log Km are plotted
against pH in Figure 8 for the hydrolysis of AMP-Im and
AMP-N-MeIm catalyzed by H98G-Fhit. The data for
AMP-Im describe a straight line with a slope of +1,
whereas the values of log Km for AMP-N-MeIm are inde-
pendent of pH. Inasmuch as the mutation of His98 has lit-
tle effect on kcat (curve b in panel A), His98 must be re-
quired for binding AMP-Im, and it has no such effect on
the binding of AMP-N-MeIm. The hydrolysis of AMP-Im
requires proton transfer to the departing imidazole, and
His98 is in position to transfer a proton to a leaving group.
It is reasonable to assign this role to His98 in the hydrolysis
of AMP-Im. The hydrolysis of AMP-N-MeIm does not re-
quire proton transfer, and its binding by H98G-Fhit is pH-
independent.
Figure 8. Plots of log Km against pH for the hydrolysis of AMP-
Im and AMP-N-MeIm by H98G-Fhit. The closed symbols are for
the hydrolysis of AMP-Im and the open symbols for the hydrolysis
of AMP-N-MeIm by H98G-Fhit.
p-Nitrophenyl-AMP as a Substrate
Figure 7. pH-Rate profiles for the hydrolysis of AMP-Im and
AMP-N-MeIm by variants of Fhit. Panel A. Log kcat is plotted
against pH for the hydrolysis of AMP-Im by: a, H94G-Fhit, pKa =
7.18 0.06; b, H98G-Fhit; and c, H94/98G-Fhit, pKa = 7.86 0.05.
Panel B. Log kcat/Km is plotted against pH for the hydrolysis of
AMP-Im by: a, H94G-Fhit, pKa = 7.11 0.17; b, H98G-Fhit, pKa
= 6.75 0.13; and c, H94/98G-Fhit, pKa = 6.95 0.13. Panel C.
Log kcat is plotted against pH for the hydrolysis of AMP-N-MeIm
by: a, H98G-Fhit and b, H94/98G-Fhit, pKa1 = 5.61 0.13 and
pKa2 = 7.60 0.06. Panel D. Log kcat/Km is plotted against pH for
the hydrolysis of AMP-N-MeIm by: a, H98G-Fhit and b, H94/98G-
Fhit, pKa1 = 5.69 0.13 and pKa2 = 7.61 0.13.
To further define the specificity of Fhit for leaving
groups, we investigated p-nitrophenyl-AMP as a substrate.
It is indeed a poor substrate for Fhit. The value of kcat at
pH 7.0 is 1.6 s–1 and the value of Km is 2.5 m. H96G-Fhit
displays no detectable activity toward p-nitrophenyl-AMP.
H94G- and H98G-Fhit are less active than wild-type Fhit,
with respective values of 0.09 and 0.01 s–1 for kcat, and 4.1
and 4.6 m for Km. Still, Fhit displays remarkable promis-
cuity in catalyzing the hydrolysis of AMP phosphodiesters,
Eur. J. Org. Chem. 2005, 5198–5206
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5203

K. Huang, A. Arabshahi, P. A. Frey
FULL PAPER
phosphoramidates, and phosphoanhydrides. The AMP- Diffusion Limitations in the Action of Fhit
phosphodiesterase activity of Fhit can be added to the
The data in Figure 6 show that the action of wild-type
many catalytic activities of HIT proteins.[35]
Fhit is diffusion-limited, but Table 1 shows that the appar-
ent second-order rate constant for the hydrolysis of Ap3A
by Fhit (kcat/Km) is only 1.2×106 –1 s–1. Typical values of
rate constants for substrates binding to enzymes are
108 –1 s–1, about two orders of magnitude higher than for
the diffusion-limited enzymatic hydrolyses catalyzed by
Fhit. This paradox can be explained in several ways. Calcu-
lations of values of kcat/Km depend on knowledge of the
exact molar concentration of the enzyme under investiga-
tion, and they are burdened by the assumption that every
enzyme molecule displays exactly the same activity. But an
assumption of homogeneity in the structure and activity of
enzyme molecules in solution is invalid. In an enzyme with
catalytically functional acid (AH) and base (B) groups, the
groups may exist mainly as AH and B, but the groups will
be in a micro-equilibrium that includes the species A– and
BH+ resulting from proton transfer. When the latter is the
active form, the fraction of active enzyme molecules can be
quite small, 1% or less. This is known as reverse proton-
ation and is well-precedented.[36] Reverse protonation of
catalytic groups in Fhit is a likely explanation for the para-
doxically low value of kcat/Km.
Discussion
Phosphoramidase Activities of Wild-Type Fhit
Because wild-type Fhit catalyzes the hydrolysis of Ap3A,
presumably by the double-displacement mechanism of
Equation 3 and 4, and H96G-Fhit is completely inactive
against Ap3A, in agreement with the essential role of His
96 as the nucleophilic catalyst, it is of considerable interest
that AMP-Im and AMP-N-MeIm are also substrates for
wild-type Fhit. The kinetic parameters for wild-type Fhit
show that AMP-Im and AMP-N-MeIm react nearly as ef-
ficiently as MgAp3A. Clearly, the imidazole rings of AMP-
Im and AMP-N-MeIm cannot be interacting with Fhit in
the manner postulated above in hydrolysis by H96G-Fhit.
The imidazole and N-methylimidazole rings of AMP-Im
and AMP-N-MeIm are likely to react in the Fhit-catalyzed
reactions simply as leaving groups in place of MgADP. Un-
like the hydrolysis of MgAp3A, Fhit-catalyzed hydrolysis of
AMP-Im and AMP-N-MeIm is not dependent on Mg2+. In
the reaction of AMP-Im, His98 appears to mediate proton
donation to the leaving imidazole. In this mechanism,
AMP-Im (or AMP-N-MeIm) react with His96 to form the
same covalent adenylyl-Fhit intermediate as in the reaction
of MgAp3A in Equation (3) (E-His96-AMP). This interme-
diate undergoes hydrolysis in the second step according to
Equation (4).
Experimental Section
Materials: Imidazole and MgCl2 were purchased from Aldrich.
AMP, ADP, ATP, potassium phosphoenolpyruvate, Ap3A, reduced
β-nicotinamide adenine dinuceotide (NADH), myokinase, pyruvate
kinase, and lactic dehydrogenase were purchased from Sigma. Ace-
tone, methanol, KCl, NaCl, N-(2-hydroxyethyl)piperazine-NЈ-2-
ethanesulfonic acid (HEPES), 3-(morpholino)propanesulfonic acid
(MOPS), 2-(cyclohexylamino)ethanesulfonic acid (CHES), NaOH,
and HCl were obtained from Fisher Scientific. Ready gel and gel-
code blue stain reagent were from BioRad and Pierce, respectively.
AMP-Im and AMP-N-MeIm were synthesized as described.[29] Hu-
man Fhit and the mutated variants were isolated and purified from
E. coli strain SG100 sells transformed from pSGA02 carrying the
wild-type or mutated gene according to the published method.[29,31]
Somewhat analogous to the plot of log Km against pH in
Figure 8, a similar plot for the hydrolysis of MgAp3A by
H98G-Fhit displays a positive slope (data not shown). This
is compatible with a role for His98 in binding MgAp3A, as
implied by the mechanism in Figure 1.
Phosphoramidase Activities of H96G-Fhit
Because of the absence of His96, H96G-Fhit cannot cat-
alyze the hydrolysis of AMP-Im or AMP-N-MeIm by the
same mechanism as Fhit. It must react with AMP-Im and
AMP-N-MeIm in the manner of chemical rescue of the ade-
nylyl-Fhit intermediate, E-His96-AMP in Equation (3). The
pH-dependences in Figs. 2 and 3 constitute documentation
that H96G-Fhit binds the monoanionic form of AMP-Im
Assays of Fhit: The coupled assay monitoring AMP formation
spectrophotometrically at 340 nm was employed.[29] The coupling
enzymes were myokinase (adenylate kinase), pyruvate kinase, and
lactate dehydrogenase. Initial rates were measured in a Cary 50
spectrophotometer equipped with a thermostatted cuvette holder.
The rates observed as decreasing A340 due to consumption of
NADH were twice the rate of AMP-Im hydrolysis and, when used
(Scheme 1) as the substrate. At some point in the mecha- to assay the wild-type enzyme, three times the rate of Ap3A hydrol-
ysis. The detailed procedure was as follows: The assay was carried
out at 26.5 °C in 1.00 mL solutions containing 0.1  MOPS or
CHES or 0.55  imidazole-HCl buffer, as well as 300 µ NADH,
300 µ phosphoenolpyruvate, 100 µ ATP, 75 m KCl, 2 m
MgCl2, and10–40 units of each coupling enzyme at various pHs.
Care was taken to ensure that the coupling enzymes did not limit
the rate. The buffer, coupling enzymes, and substrates for the coup-
ling enzymes were separately incubated in the cell holder at 26.5 °C
for 20 min and then mixed just before adding either a substrate
(AMP-Im, AMP-N-MeIm, or Ap3A) or Fhit. The A340 was moni-
tored for 2 min to obtain the background rate (if any) and allow
nism, the imidazole ring of AMP-Im must acquire a proton
to depart as imidazole. This appears to be a function of
His98 in the action of wild-type Fhit on AMP-Im (Fig-
ure 8), and His98 may well serve as the proton donor in the
hydrolysis of AMP-Im by H96G-Fhit. However, this is not
established by the present data, and the mechanism of pro-
ton transfer to AMP-Im remains unknown. Comparison of
the results for AMP-Im and AMP-N-MeIm confirm that
the two react by mechanisms that differ in their require-
ments for proton transfer.
5204
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5198–5206

pH-Dependence in the Hydrolytic Action of the Human Fragile Histidine Triad
FULL PAPER
the temperature to stabilize at 26.5 °C, then the Fhit was added to
Y = A + B/(1 + [H+]/K)
(9)
initiate the enzymatic reaction. The initial rate of decreasing A340
(NADH) was measured. The rates were calculated by use of ε340
=
Acknowledgment
6220 –1·cm–1 for NADH. The subunit concentration of Fhit was
determined by use of ε280 = 8310 –1·cm–1 calculated from the
amino acid sequence of Fhit evaluated by the published method.[32]
This research was supported by Grant No. GM 30480 from the
National Institute of General Medical Sciences, U. S. Public Health
Service.
pH-Rate Profiles: The assay procedure was adapted for use from
pH 5.5 to pH 9.0, where the coupling enzymes displayed sufficient
activities, and initial rates were measured in triplicate. The good
buffers HEPES, MOPS and CHES were employed to cover the pH-
range. The data at each pH were fitted to the Michaelis–Menten
equation using KaleidaGraph to evaluate Vmax (kcat) and Km at that
pH. This was repeated at each pH. The non-enzymatic rates for
AMP-N-MeIm as the substrate were significant and had to be sub-
tracted from the measured initial rates after adding the enzyme.
Data were collected at 0.25 or 0.5 pH intervals, and the data were
replicated with both buffers at pHs corresponding to buffer
changes to ensure the absence of buffer effects. The profiles of log
kcat and log kcat/Km vs. pH were plotted and the data computer
fitted to Equation (5), Equation (6), Equation (7), or Equation (8)
using the programs HABELL, HBBELL, BELL or BEL2H,
respectively, of Cleland.[33]
[1] M. Ohta, H. Inoue, M. G. Cotticelli, K. Kastury, R. Baffa, J.
Palazzo, Z. Siprashvili, M. Mori, P. McCue, T. Druck, C. M.
Croce, K. Huebner, Cell 1996, 84, 587–597.
[2] K. Huebner, P. Hadaczek, Z. Siprashvili, T. Druck, C. M.
Croce, Biochim. Biophys. Acta 1997, 1332, M65–M70.
[3] K. Huebner, P. N. Garrison, L. D. Barnes, C. M. Croce, Ann.
Rev. Genet. 1998, 32, 7–31.
[4] L. Ji, B. Fang, N. Yen, K. Fong, J. D. Minna, J. A. Roth, Can-
cer Res. 1999, 59, 3333–3339.
[5] L. Y. Fong, V. Fidanza, N. Zanesi, L. F. Lock, L. D. Siracusa,
R. Mancini, Z. Siprashvili, M. Ottey, S. E. Martin, T. Druck,
P. A. McCue, C. M. Croce, K. Huebner, Proc. Natl. Acad. Sci.
USA 2000, 97, 4742–4747.
[6] K. R. Dumon, H. Ishii, L. Y. Fong, N. Zanesi, V. Fidanza, R.
Mancini, A. Vecchione, R. Baffa, F. Trapasso, M. J. During,
K. Huebner, C. N. Croce, Proc. Natl. Acad. Sci. USA 2001, 98,
3346–3351.
[7] H. Ishii, N. Zanesi, A. Vecchione, F. Trapasso, S. Yendamuri,
M. Sarti, R. Baffa, M. J. During, K. Huebner, L. Y. Fong,
C. M. Croce, FASEB J. 2003, 17, 1768–1770.
[8] C. Brenner, P. N. Garrison, J. Gilmour, D. Peisach, D. Ringe,
G. A. Petsko, J. M. Lowenstein, Nature Struct. Biol. 1997, 4,
231–238.
c
log y = log
log y = log
log y = log
log y = log
(5)
(6)
(7)
(8)
1 + [H+]/Ka
c
[9] J. E. Wedekind, P. A. Frey, I. Rayment, Biochemistry 1995, 34,
1 + Kb/[H+]
11049–11061.
[10] C. D. Lima, M. G. Klein, L. B. Weinstein, W. A. Hendrickson,
Proc. Natl. Acad. Sci. USA 1996, 93, 5357–5362.
[11] L. Holm, C. Sander, Trends Biochem. Sci. 1997, 22, 116–117.
[12] C. D. Lima, K. L. D’Amico, I. Naday, G. Rosenbaum, E. M.
Westbrook, W. A. Hendrickson, Structure 1997, 5, 763–774.
[13] M. Ohta, H. Inoue, M. G. Cotticelli, K. Kastury, R. Baffa, J.
Palazzo, Z. Siprashvili, H. C. Pace, P. N. Garrison, A. K. Rob-
inson, L. D. Barnes, A. Draganescu, A. Rösler, G. M. Black-
burn, Z. Siprashvili, C. M. Croce, K. Huebner, C. Brenner,
Proc. Natl. Acad. Sci. USA 1998, 95, 5484–5489.
[14] T. L. Field, W. S. Reznikoff, P. A. Frey, Biochemistry 1989, 28,
2094–2099.
c
1 + [H+]/Ka + Kb/[H+]
c
1 + [H+]/Ka + Kb/[H+] + Ko/[H+]2
[15] L. J. Wong, P. A. Frey, J. Biol. Chem. 1974, 249, 2322–2324.
[16] L. J. Wong, P. A. Frey, Biochemistry 1974, 13, 3889–3894.
[17] L. J. Wong, K. F. Sheu, S. L. Lee, P. A. Frey, Biochemistry 1977,
16, 1010–1016.
[18] S. L. Yang, P. A. Frey, Biochemistry 1979, 18, 2980–2984.
[19] K. F. Sheu, J. P. Richard, P. A. Frey, Biochemistry 1979, 18,
5548–5556.
In fitting the data, y values were the experimental pH-dependent
app
app
values of k or kcat/K , c was the pH-independent value arising
cat
mt
from the fitting procedure, and Ka and Kb were the fitted values of
acid dissociation constants.
The Effects of Viscosity on Kinetic Parameters: The kinetic param-
eters of both wild type and mutated Fhit with different substrates
were measured at varying relative viscosity (ηrel = η/η0) in 100 m
MOPS buffer containing sucrose at 26.5 °C. Approximate values of
relative viscosity of the sucrose used in this study are the follow-
ing:[34] 0% sucrose, ηrel = 0, 14% sucrose, ηrel = 1.47, 24% sucrose,
ηrel = 2.16, 32% sucrose, ηrel = 3.15. A plot of ka/ka0 (ka = kcat/Km,
ka0 = ka as ηrel = 1) vs. relative viscosity (ηrel) was obtained.
[20] A. Arabshahi, R. S. Brody, A. Smallwood, T. C. Tsai, P. A.
Frey, Biochemistry 1986, 25, 5583–5589.
[21] J. Kim, F. Ruzicka, P. A. Frey, Biochemistry 1990, 29, 10590–
10593.
[22] T. L. Field, W. S. Reznikoff, P. A. Frey, Biochemistry 1989, 28,
2094–2099.
[23] S. Geeganage, P. A. Frey, Methods Enzymol. 2002, 354, 134–
148.
[24] J. E. Wedekind, P. A. Frey, I. Rayment, Biochemistry 1996, 35,
11560–11569.
Ionization of AMP-Im: A 40 m AMP-Im solution in D2O was
adjusted to the desired pD by addition of DCl, and the chemical
shifts of C2(H) in the imidazole ring of AMP-Im were measured
at pHs from 4.5 to pH 7.5 at 22 °C in a Bruker 200 MHz NMR
spectrometer. The chemical shift-values were plotted against pH
and fitted to Equation (9) using the TITR program of Cleland. The
value of pKa obtained for the ionization of AMP-Im according to
Scheme 1 was 5.6 0.1.
[25] Y. Huang, P. N. Garrison, L. D. Barnes, Biochem. J. 1995, 312,
925–932.
[26] L. D. Barnes, P. N. Garrison, Z. Siprashvili, A. Guranowski,
A. K. Robinson, S. W. Ingram, C. M. Croce, M. Ohta, K.
Huebner, Biochemistry 1996, 35, 11529–11535.
[27] C. D. Lima, M. D. Klein, W. A. Hendrickson, Science 1997,
278, 286–290.
Eur. J. Org. Chem. 2005, 5198–5206
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5205

K. Huang, A. Arabshahi, P. A. Frey
FULL PAPER
[28] A. Abend, P. N. Garrison, W. D. Barnes, P. A. Frey, Biochemis-
try 1999, 38, 3668–3676.
[29] K. Huang, A. Arabshahi, Y. Wei, P. A. Frey, Biochemistry 2004,
43, 7637.
[30] F. J. Ruzika, P. A. Frey, Bioorg. Chem. 1993, 21, 238–248.
[31] C. Brenner, H. C. Pace, P. N. Garrison, A. K. Robinson, A.
Rosler, X. H. Liu, G. M. Blackburn, C. M. Croce, K. Huebner,
L. D. Barnes, Protein Eng. 1997, 10, 1461–1463.
[32] S. C. Gill, P. H. von Hippel, Anal. Biochem. 1989, 182, 319–
326.
[33] W. W. Cleland, Methods Enzymol. 1979, 63, 103–138.
[34] A. C. Brouwer, J. F. Kirsch, Biochemistry 1982, 21, 1302–1307.
[35] C. Brenner, Biochemistry 2002, 41, 9003–9014.
[36] P. A. Sims, T. M. Larsen, R. R. Poyner, W. W. Cleland, G. H.
Reed, Biochemistry 2003, 42, 8298–8306.
Received: July 6, 2005
Published Online: October 10, 2005
5206
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5198–5206