Báo cáo khoa học: Influence of divalent cations on the structural thermostability and thermal inactivation kinetics of class II xylose isomerases pdf

Influence of divalent cations on the structural
thermostability and thermal inactivation kinetics
of class II xylose isomerases
Kevin L. Epting
1
, Claire Vieille
2
, J. Gregory Zeikus
2
and Robert M. Kelly
1
1 Department of Chemical Engineering, North Carolina State University, Raleigh, NC, USA
2 Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI, USA
Enzymes from hyperthermophiles are intrinsically ther-
mostable and thermoactive, with optimal temperatures
of activity often in excess of 100 °C [1]. Studies focus-
ing on the molecular basis for the thermostability of
these enzymes have revealed an array of subtle contri-
buting factors, including larger hydrogen bonding and
ion pairing networks, additional inter-subunit inter-
actions in oligomeric proteins, decreased labile amino
acid content, lower surface-to-volume ratios, and smal-
ler loops [2–4]. These factors can often be identified by
comparing the structures of homologous enzymes with
varying degrees of thermostability [5]. However, the
extent to which these individual factors contribute to
thermostability is highly specific to particular enzymes
[6].
A potential contributing factor to enhanced thermo-
stability that has not been examined in much detail is
the role that metals play in stabilizing and activating

enzymes from hyperthermophiles in comparison with
their less thermophilic counterparts. Xylose isomerase
Keywords
inactivation kinetics; metal cofactors;
thermostability; xylose isomerases
Correspondence
R. M. Kelly, Department of Chemical
Engineering, North Carolina State University,
Raleigh, NC 27695–7905
Fax: +1 919 515 3465
Tel: +1 919 515 6396
E-mail:
(Received 27 November 2004, accepted 20
January 2005)
doi:10.1111/j.1742-4658.2005.04577.x
The effects of divalent metal cations on structural thermostability and the
inactivation kinetics of homologous class II d-xylose isomerases (XI;
EC 5.3.1.5) from mesophilic (Escherichia coli and Bacillus licheniformis),
thermophilic (Thermoanaerobacterium thermosulfurigenes), and hyperther-
mophilic (Thermotoga neapolitana) bacteria were examined. Unlike the
three less thermophilic XIs that were substantially structurally stabilized in
the presence of Co
2+
or Mn
2+
(and Mg
2+
to a lesser extent), the melting
temperature [(T
m

) %100 °C] of T. neapolitana XI (TNXI) varied little in
the presence or absence of a single type of metal. In the presence of any
two of these metals, TNXI exhibited a second melting transition between
110 °C and 114 °C. TNXI kinetic inactivation, which was non-first order,
could be modeled as a two-step sequential process. TNXI inactivation in
the presence of 5 mm metal at 99–100 °C was slowest in the presence of
Mn
2+
[half-life (t
1 ⁄ 2
) of 84 min], compared to Co
2+
(t
1 ⁄ 2
of 14 min) and
Mg
2+
(t
1 ⁄ 2
of 2 min). While adding Co
2+
to Mg
2+
increased TNXI’s t
1 ⁄ 2
at 99–100 °C from 2 to 7.5 min, TNXI showed no significant activity at
temperatures above the first melting transition. The results reported here
suggest that, unlike the other class II XIs examined, single metals are
required for TNXI activity, but are not essential for its structural thermo-
stability. The structural form corresponding to the second melting transi-

tion of TNXI in the presence of two metals is not known, but likely results
from cooperative interactions between dissimilar metals in the two metal
binding sites.
Abbreviations
TIM, triosephosphate isomerase; XI, xylose isomerase.
1454 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS
(XI) (d-xylose ketol isomerase, EC 5.3.1.5) is an excel-
lent model system to consider in this regard, as diva-
lent metal cations are important for both stability and
activity of all known XIs [7]. XI converts d-xylose to
d-xylulose in vivo, but it also converts d-glucose to
d-fructose in vitro [8], hence its use for the commercial
production of high fructose corn syrup [9,10]. An
intracellular enzyme, XI is found in a number of bac-
teria that can grow on xylose [11], as well as in fungi
[12,13], and plants [14]. XIs can be divided into two
groups based on sequence comparisons: class I and
class II [15]. Class I XIs contain % 390 amino acids,
while class II XIs typically contain around 440 amino
acids and are distinguished from the class I enzymes
by a 30–40 amino acid N-terminal insert [10]. The
functional role of this N-terminal insert in the class II
enzymes is unknown. Divalent metal cations (chosen
from Co
2+
,Mg
2+
, and Mn
2+
) are essential for XI’s

stability and activity [16–19], but their relative import-
ance differs somewhat for class I and class II XIs [7].
A number of three-dimensional structures have been
solved for class I XIs [20–23] and have been shown to
be essentially identical, which explains the similar bio-
chemical and thermostability properties of these
enzymes [7]. Comparisons between the structures of
the most thermostable class I XIs from Thermus caldo-
philus and Thermus thermophilus to those from less
thermophilic class I XIs from Arthrobacter B3728 and
Actinoplanes missouriensis revealed common thermosta-
bilizing features: increased ion pairing, lower surface-
to-volume exposure, fewer exposed labile amino acids,
and shortened loops [24]. Similar comparisons among
class II XIs have not been reported.
Class II XIs have not been studied to the same
extent as class I enzymes, probably because none are
currently used commercially. Genes encoding several
class II XIs have been cloned from mesophilic, thermo-
philic, and hyperthermophilic bacteria, expressed in
Escherichia coli, and characterized biochemically
[25–29]. Although class I and class II XIs differ in
their metal specificities [18], active-site structure and
metal-binding residues are conserved across the two XI
classes [7]. In contrast to class I XIs, however, the
thermostability of class II XIs is more variable. No
obvious differences in the enzyme structures can
explain these variations in stability, although the more
thermophilic class II XIs contain additional prolines
and fewer thermally labile asparagine and glutamine

residues [26,30].
The crystal structures of class I and class II XIs
show that these enzymes are typically homodimers
or homotetramers, consisting of a triosephosphate
isomerase (TIM) barrel connected to a C-terminal loop
[21,31]. Each monomer contains an active site that fea-
tures two distinct metal binding sites. Metal M1 is
coordinated by four carboxylate groups, while metal
M2 is coordinated by one imidazole and three carb-
oxylate groups [32]. Metals M1 and M2 were initially
called structural and catalytic metals, respectively, due
to the fact that the M1 site remained geometrically
unchanged during catalysis, while the M2 site changed
upon substrate binding [33]. Subsequent studies have
shown, however, that both metals are needed for cata-
lysis [34]. The binding affinity appears to differ for the
two sites [17,35,36], and it varies with pH and the
type of metal [37]. Binding affinity, though, varies
in the same way in both enzyme classes, with
Mn
2+
>Co
2+
>Mg
2+
[19]. Metal specificity also
depends on both the nature of the substrate (i.e. glu-
cose or xylose) and on the enzyme class [26]. With glu-
cose as the substrate, class I enzymes are best activated
by Mg

2+
[38], or, in some cases, by a combination of
Mg
2+
and Co
2+
[11,39]. While Co
2+
best activates
the class II enzymes for activity on glucose, Mn
2+
is
preferred for activity on xylose [38]. However, Mn
2+
or Co
2+
provide superior thermal stabilization to both
XI classes [16]. Metal specificity appears to be related
to the residues surrounding the metal binding sites.
Several mutant XIs with altered metal specificity were
created by site-directed mutagenesis [40–42]. In gen-
eral, the resulting mutants had decreased specificity for
all metals compared to the wild-type enzyme.
Given that metals are needed for both stabilizing
and activating XIs, a question arises concerning the
relative importance of these cofactors across functional
temperature ranges. This issue was examined here
for class II XIs from the mesophiles E. coli (ECXI)
and Bacillus licheniformis (BLXI), for the moderate
thermophile Thermoanaerobacterium thermosulfurigenes

(TTXI), and for the hyperthermophile Thermotoga
neapolitana (TNXI). Results indicate that the roles that
divalent metals play in TNXI stabilization and activa-
tion differ from those in the less thermophilic enzymes.
Results
Structural thermostability of class II XIs
Four class II XIs (i.e. ECXI, BLXI, TTXI, and TNXI)
were investigated to determine the influence of specific
divalent metal cations on thermally induced denatura-
tion using differential scanning calorimetry (DSC). In
the absence of metal (i.e. apoenzymes) or in presence
of a single metal at 5 lm, all four XIs exhibited a sin-
gle irreversible thermal transition (Fig. 1). While the
nature of the metal present significantly affected the
K. L. Epting et al. Thermostability of class II xylose isomerases
FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS 1455
stabilization of BLXI and TTXI relative to their apo
forms, the effect was smaller for ECXI and almost
negligible for TNXI (Fig. 2). The melting temperatures
(T
m
, the temperature at the maximum of the heat
capacity profile) of ECXI, BLXI, and TTXI followed
the trend reported previously for class II XIs, with
Mn
2+
or Co
2+
providing greater stabilization than
Mg

2+
. This was particularly noticeable for BLXI;
Mn
2+
or Co
2+
increased the T
m
by 20 °C more than
did Mg
2+
. In all instances, but to varying extents, the
apoenzyme melted at a lower temperature than the
same enzyme containing any of the three metals
(Table 1). It was interesting to note that TNXI melting
curves at low metal concentrations (5 lm) led to T
m
values between 96.9 °C and 97.6 °C, barely above the
T
m
of the apoenzyme (96.4 °C). Atomic emission
spectroscopy analyses of TNXI at low metal concen-
trations (i.e. between 5 and 500 lm metal) suggest that
both metal binding sites are not occupied in these con-
ditions (data not shown). In the presence of excess
metal (i.e. 5 mm, to ensure occupation of both metal
sites), TNXI’s T
m
slightly increased to 100.5 °C
(Mn

2+
), 100.4 °C (Mg
2+
), and 100.0 °C (Co
2+
). The
melting behavior of TNXI was also examined in the
presence of Ni
2+
and Ca
2+
, two divalent metals
that inactivate the enzyme [11]. With T
m
values of
100.9 °C and 100.5 °C for 5 mm Ni
2+
and 5 mm
0
20
40
60
80
100
70 80 90 100 110 120
Excess heat capacity
(kcal/mole·K)
Mg
2+
Mn

2+
Co
2+
Apo
TNXI
0
20
40
60
80
50 60 70 80 90 100
Excess heat capacity
(kcal/mole·K)
Co
2+
Apo
Mn
2+
Mg
2+
TTXI
0
50
100
150
200
40 50 60 70 80 90
Excess heat capacity
(kcal/mole·K)
Mg

2+
Co
2+
Mn
2+
Apo
BLXI
0
10
20
30
40
50
60
40 50 60 70 80 90
Temperature (ºC)
Excess heat capacity
(kcal/mole·K)
Mg
2+
Mn
2+
Co
2+
Apo
ECXI
Fig. 1. Thermal denaturation of class II xylose isomerases. DSC
scans of TNXI, TTXI, BLXI and ECXI were run in 50 m
M Mops
(pH 7.0) with no metal (apo) or in the presence (at 5 l

M) of a single
metal.
0
5
10
15
20
25
BLXI TTXI ECXI TNXI TNXI*
∆
T (ºC)
Mg
2+
Co
2+
Mn
2+
Fig. 2. Effect of metals on the T
m
values of class II XIs. DT is the
difference between T
m
(enzyme in the presence of 5 lM single
metal) and T
m
(apoenzyme). For TNXI*, the metal concentration
was 5 m
M.
Table 1. Effect of activating divalent cations on melting tempera-
tures of class II XIs.

Enzyme
Melting temperature T
m
(°C)
Apo Mg
2+
Co
2+
Mn
2+
ECXI 50.8 56.2 56.2 57.3
BLXI
a
50.3 53.3 73.4 73.6
TTXI 64.1 83.0 86.0 86.1
TNXI 96.4 97.6 97.5 96.9
TNXI
b
– 100.4 100.0 100.5
a
Data from [26].
b
Buffer containing 5 mM metal chloride.
Thermostability of class II xylose isomerases K. L. Epting et al.
1456 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS
Ca
2+
, respectively, both metals stabilized TNXI to the
same extent as Mn
2+

,Mg
2+
,orCo
2+
.
The biochemical and biophysical properties of some
XIs have been studied in the presence of two different
metals (e.g. Mg
2+
and Co
2+
) [43–45]. For three of the
enzymes studied here, the T
m
in the presence of 5 mm
Mg
2+
plus 0.5 mm Co
2+
was slightly above that affor-
ded by the most stabilizing single metal: ECXI
(58.0 °C vs. 57.3 °C with Mn
2+
), BLXI (74.4 °C vs.
73.6 °C with Mn
2+
), and TTXI (87.0 °C vs. 86.1 °C
with Mn
2+
). In these cases, the trajectories of the

melting curves were similar to those obtained in the
presence of a single metal (data not shown). In con-
trast, TNXI’s melting curve in the presence of 5 mm
Mg
2+
plus 0.5 mm Co
2+
showed two transitions,
around 99.5 °C and 110 °C (Fig. 3). To determine
whether the relative concentrations of the two metals
affected TNXI melting behavior, TNXI’s melting curve
was recorded in the presence of 5 mm Mg
2+
plus
either 1 or 5 mm Co
2+
. In both cases, two transitions
were observed (data not shown) at the same tempera-
tures as in the presence of 5 mm Mg
2+
and 0.5 mm
Co
2+
. In fact, all TNXI melting curves in the presence
of any two of the three metals showed two transitions
(Fig. 3); the Mn
2+
⁄ Mg
2+
and Mn

2+
⁄ Co
2+
combina-
tions showed transitions at 99.6 °C ⁄ 114.2 °C and at
100.6 °C ⁄ 112.1 °C, respectively.
To further investigate the basis for the two transi-
tions, individual TNXI scans in the presence of 5 mm
Mg
2+
and 0.5 mm Co
2+
were stopped at 75, 90, 95,
100, 105, 110, and 115 °C (Fig. 4) to view the residual
soluble enzyme conformation on SDS and native
PAGE. All samples showed a single band at 50 kDa
on SDS ⁄ PAGE (not shown), as expected for the TNXI
monomer. On the native gel, however, all samples
showed three distinct bands, presumably the tetramer,
dimer, and monomer (Fig. 5). In comparison, in the
presence of a single metal, no soluble protein was pre-
sent at temperatures beyond the end of the single ther-
mal transition (not shown). All three forms of the
enzyme were observed on native PAGE over the entire
temperature range. Despite the decrease in soluble pro-
tein concentration (Fig. 4), the relative amounts of
each form appear to remain the same, indicating that
as the dimer dissociates into monomers, the monomers
unfold and aggregate. Hence, the amount of monomer
in the soluble fraction remains low.

0
20
40
60
80
100
70 80 90 100 110 120
Temperature (ºC)
Excess heat capacity
(kcal/mole·K)
Fig. 3. TNXI melting transitions in the presence of two different
metals. DSC scans were run in 50 m
M Mops (pH 7.0) in the pres-
ence of 5 m
M Mg
2+
⁄ 0.5 mM Co
2+
(dashed line), 5 mM
Mn
2+
⁄ 0.5 mM Co
2+
(black line), or 5 mM Mg
2+
⁄ 5mM Mn
2+
(gray
line).
0

20
40
60
80
100
70 80 90 100 110 120
Temperature (ºC)
Excess heat capacity (kcal/mole·K)
0
20
40
60
80
100
Percent soluble enzyme
Fig. 4. Soluble TNXI concentration during thermal denaturation in
the presence of 5 m
M Mg
2+
plus 0.5 mM Co
2+
. The concentration
of soluble protein (% of starting concentration, h) and the thermal
transitions are shown as functions of the temperature.
Fig. 5. Native PAGE of soluble TNXI at different temperatures dur-
ing thermal denaturation in the presence of 5 m
M Mg
2+
plus
0.5 m

M Co
2+
. Soluble fractions were loaded at 100 lg protein per
lane. Control: unheated TNXI in 50 m
M Mops (pH 7.0) containing
5m
M Mg
2+
plus 0.5 mM Co
2+
. Presumed tetramer, dimer and
monomer bands identified.
K. L. Epting et al. Thermostability of class II xylose isomerases
FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS 1457
To determine if TNXI’s two melting transitions in
the presence of two different metals were a result of
differences in metal binding affinity in the two metal
binding sites, point mutations were introduced selec-
tively into metal sites M1 (i.e. mutation E232K) and
M2 (i.e. mutation D309K). The double mutant
(E232K ⁄ D309K) was also created. In previous stud-
ies of class I XIs, the residues corresponding to
E232 and D309 were mutated to lysine and the crys-
tal structures of the mutants were determined. The
positive charge of the lysine’s e-amino group presum-
ably replaced the metal, while leaving the other
metal site unaffected. Both mutations eliminated
activity [21,31]. Here, as expected, none of the three
TNXI mutants were active. The three mutants
showed a single melting transition in the presence of

5mm Mg
2+
and 0.5 mm Co
2+
(Fig. 6). With a T
m
of 100.7 °C, the M1 site mutant behaved like TNXI
in the presence of an excess single metal. Both the
M2 site mutant (T
m
of 95.5 °C) and the double
mutant (T
m
of 96.5 °C) behaved like the apo-TNXI
(T
m
of 96.4 °C).
Kinetic inactivation
Previously, BLXI was shown to follow first order kin-
etic inactivation [26]. To determine whether the two
thermal transitions observed by DSC for TNXI in the
presence of two metals had any relevance to this
enzyme’s inactivation kinetics, the inactivation course
of TNXI in the presence of various metal concen-
trations and metal combinations was determined.
Figure 7 shows the inactivation courses of apo-TNXI
and of TNXI in the presence of 5 mm concentrations
of each of the three divalent cations (Mg
2+
,Mn

2+
,or
Co
2+
). With the exception of apo-TNXI, whose inacti-
vation was first-order, TNXI exhibited non-first order
inactivation in the presence of any single metal or
combination of metals (data not shown). Previously,
TNXI inactivation was proposed to proceed by a two-
step, sequential mechanism [46]:
E À!
k
1
E
1
b
À!
k
2
E
d
where E is the native, fully active enzyme (relative
activity of 1.0); E
1
is an intermediate with lower
activity (b < 1.0) than E; E
d
is the inactive enzyme
and k
1

and k
2
are the inactivation rate constants. Such
a mechanism can be modeled by the sum of two expo-
nential terms, where y
(t)
represents the fractional resid-
ual activity [47]:
0
20
40
60
80
100
120
140
80 85 90 95 100 105 110
Temperature (ºC)
Excess heat capacity
(kcal/mole·K)
E232K / D309K
D309K E232K
Fig. 6. Thermal denaturation of TNXI metal site mutants. DSC
scans of TNXI and its E232K, D309K, and E232K ⁄ D309K mutants
in buffer A.
Mn
2+
0
0.2
0.4

0.6
0.8
1
0 20406080
0 20406080
100 120
Time (min) Time (min)
Time (min)
Time (min)
Relative Activity
Mg
2+
0
0.2
0.4
0.6
0.8
1
Relative Activity
Co
2+
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120 140
0 20406080100120140
160 180 200

Relative Activity
Apo
0
0.2
0.4
0.6
0.8
1
Relative Activity
Fig. 7. Model fit to TNXI kinetic inactivation
in the presence and absence of activating
divalent cations. Model parameters and r
2
values are listed in Table 2. 106 °C(r),
104 °C(h), 102 °C(n), 99 °C(n), 96 °C(m),
90 °C(d), 87 °C(s).
Thermostability of class II xylose isomerases K. L. Epting et al.
1458 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS
y
ðtÞ
¼ 1 þ
b
k
1
k
2
À k
1

expðÀk

1
tÞ
À
b
k
1
k
2
À k
1
expðÀk
2
tÞ
ð1Þ
If there is no active intermediate (E
1
), b ¼ 0 and the
expression reduces to a first order decay. Equa-
tion (1) was used to fit TNXI inactivation data at
different temperatures and at various concentrations
and combinations of the three metals (Table 2;
Fig. 7). Half-lives of the enzymes were calculated by
setting y
(t)
in Eqn (1) equal to 0.5 and solving for
t
1 ⁄ 2
. The inactivation behavior analysis revealed the
sensitivity of TNXI kinetic stability to temperature.
For example, TNXI’s half-life in the presence of

Co
2+
drops from 59.5 min at 96 °C to 14 min at
99 °C. At 99 °C (i.e. close to T
m
), TNXI’s estimated
t
1 ⁄ 2
show that Mn
2+
is the most stabilizing metal
(t
1 ⁄ 2
of 84 min), followed by Co
2+
(t
1 ⁄ 2
of 14 min)
and then Mg
2+
(t
1 ⁄ 2
of 2 min). Adding 0.5 mm
Co
2+
to an excess of Mg
2+
(5 mm) tripled the t
1 ⁄ 2
compared to only Mg

2+
. Based on inactivation rate
constants for the two-step mechanism, the first step
proceeded at a much higher rate than the second
step (the ratios of rate constants k
1
⁄ k
2
were at least
15 for all cases at 99–100 °C).
Discussion
Despite the structural similarities shared by the class II
XIs compared here (all show ‡ 48% sequence identity),
there was significant variation in the degree of struc-
tural thermostabilization that the different divalent
metals provided. It is interesting to note that for the
most thermostable enzyme, TNXI, metals were less
important for stability than for the other two thermo-
stable enzymes, BLXI and TTXI. Unlike BLXI and
TTXI, apo- and halo-TNXI melted at similar tempera-
tures (Fig. 1). This observation points to the heigh-
tened structural rigidity of hyperthermophilic enzymes,
even within their functional temperature ranges [48].
All XIs are known to be active only in the presence of
divalent cations. Although there have been reports of
XIs active with only one metal site occupied [18],
mutating either metal binding site inactivated TNXI.
In general, activity level varies with substrate, type of
metal present, and XI class. With glucose as the sub-
strate, class I enzymes are best activated by Mg

2+
, fol-
lowed by Co
2+
and Mn
2+
, while class II enzymes are
best activated by Co
2+
, then Mg
2+
and Mn
2+
.In
contrast, with xylose as the substrate, class II XIs are
best activated by Mn
2+
[16]. It is interesting that some
XIs show maximum activity in the presence of both
Mg
2+
and Co
2+
at ratios of 5 : 1, respectively, or
higher [39,43–45,49]. The reasons for the differences in
the metal preference for specific enzymes and sub-
strates are not clear, but this is most likely related to
subtle conformational changes in the active site that
vary depending on the specific metal present in each
site.

In addition to maximizing activity, the presence of
both Mg
2+
and Co
2+
also enhanced the thermostability
of all the XIs studied when compared with Mg
2+
or
Co
2+
alone. While the melting curves of ECXI, BLXI,
Table 2. Effect of divalent metals cations on TNXI inactivation kinetics.
y
ðtÞ
¼ A
expðÀk
1
ÁtÞ
þ B
expðÀk
2
ÁtÞ
or y
ðtÞ
¼ 1 þ
b
k
1
k

2
À k
1

expðÀk
1
tÞ
À
b
k
1
k
2
À k
1
expðÀk
1
tÞ
Metal Conc (mM)T(°C) A k
1
(min
)1
)B k
2
(min
)1
) b t
1 ⁄ 2
(min) v
2

(· 10
)2
) r
2
Mn
2+
2 99 0.428 0.22 0.58 0.002 0.57 83.2 0.024 0.996
102 0.518 0.40 0.48 0.005 0.48 6.9 0.083 0.983
104 0.639 0.41 0.40 0.008 0.39 4.2 0.224 0.966
106 0.668 0.61 0.34 0.014 0.34 2.3 0.061 0.991
Co
2+
2 96 0.41 0.08 0.59 0.003 0.57 59.5 0.057 0.992
99 0.43 0.20 0.57 0.013 0.53 14 0.058 0.993
102 0.51 0.36 0.49 0.020 0.46 5.8 0.024 0.997
104 0.61 0.49 0.39 0.038 0.36 2.9 0.029 0.997
Mg
2+
92 0.534 0.24 0.48 0.001 0.48 12 0.080 0.984
2 96 0.739 0.52 0.29 0.011 0.29 2.3 0.126 0.987
99 0.771 0.57 0.26 0.027 0.24 1.9 0.145 0.986
102 0.946 0.67 0.11 0.056 0.10 1.3 0.384 0.968
None – 87 0.910 0.02 – – – 30 0.398 0.965
90 1.02 0.06 – – – 11.3 0.185 0.996
Mg
2+
⁄ Co
2+
5 ⁄ 0.5 100 0.42 0.48 0.58 0.023 0.55 7.5 0.014 0.998
Co

2+ a
0.5 100 0.64 0.45 0.35 0.017 0.34 3.0 0.013 0.998
a
Data from [29].
K. L. Epting et al. Thermostability of class II xylose isomerases
FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS 1459
and TTXI in the presence of two different metals all
resembled the single-metal cases, the TNXI curve
showed two transitions (Fig. 3). TNXI’s unusual melt-
ing behavior in the presence of Mg
2+
and Co
2+
was
noted previously [50]. Further investigations showed
that the relative size of the transitions changed when the
pH was increased from 7.0 to 7.9 [51], suggesting that
the unusual melting behavior was related to metal bind-
ing affinity. In fact, TNXI exhibits two melting transi-
tions in the presence of any two of the metals studied
(i.e. Mn
2+
,Co
2+
, and Mg
2+
) (Fig. 3). To confirm that
this melting behavior was due to the presence of two dif-
ferent metals in TNXI’s metal binding sites, mutations
to site M1 (E232K) and site M2 (D309K) were intro-

duced in TNXI. In either case, the enzyme contained at
most one metal per active site. Both TNXI mutants
showed single melting transitions. Studies of Streptomy-
ces class I XIs have shown that the metal binding sites
(M1 and M2) have different binding affinities [17,52]
and that the high and low affinity sites are different for
different cations [36]: M1 is the high affinity site for
Mg
2+
, while M2 is the high affinity site for Co
2+
and
Mn
2+
[16]. High and low affinity sites were also repor-
ted in class II XIs [19], and they are likely to be the same
as in class I XIs, since the overall structure and the metal
binding residues are conserved across both enzyme clas-
ses. The TNXI E232K mutant would therefore have
higher affinity for Co
2+
as only the M2 site is intact,
while the D309K mutant would have higher affinity for
Mg
2+
for similar reasons. In the presence of Mg
2+
and
Co
2+

, D309K TNXI has a T
m
(95.5 °C) slightly lower
than that of the apoenzyme, and E323K TNXI has the
same T
m
as TNXI in the presence of a single metal.
These results suggest that M2 is the only metal import-
ant for TNXI stability, while both metals are needed for
activity.
Cleavage of the Arthrobacter XI C-terminal loop by
thermolysin affected neither stability nor activity [53],
suggesting that thermal inactivation begins in the TIM
barrel. Site directed mutagenesis was used in various
studies to examine metal binding in XIs (reviewed in
[7]). Point mutations that trigger conformational chan-
ges in the active site residues destabilize XIs, and
mutations that alter the metal binding residues greatly
reduce or eliminate activity [21,37,40,54,55]. These
results suggest that the irreversible thermal unfolding
begins with movement of active site residues [56]. The
metal cofactors are thought to hold the active site in a
stable conformation that is lost when the metals are
removed [26]. While the metals play an essential part
in the catalytic mechanism, they are not required for
the enzyme to fold properly. Indeed, the two metal site
mutants constructed in this study behaved exactly as
TNXI in the different purification steps (in particular,
they could be heat-treated in the same conditions as
TNXI and remain soluble). They also showed melting

transitions as high as the apoenzyme, and their mobility
on native PAGE was identical to that of TNXI (data
not shown).
TNXI showed non-first order kinetic inactivation in
the presence of Co
2+
,Mg
2+
or Mn
2+
. These kinetics
suggest a complex, higher order process that can cor-
respond to numerous possible molecular mechanisms
[47]. TNXI inactivation could be modeled as the sum
of two exponentials, as shown in Fig. 7 and Table 2.
This general mechanism can be interpreted as a
sequential inactivation, with one or more catalytically
active intermediates. Non-first order inactivation has
been reported previously for the Thermotoga sp. and
Streptomyces murinus XIs [29,51,57], although instead
of a single model, the inactivation data were divided
into two first-order phases – a faster initial phase and
a slower later phase. However, TNXI showed first
order inactivation when covalently immobilized to
glass beads [51] or when heated in the absence of
metals (apo). These results suggest that the soluble
TNXI inactivates through one or more partially active
intermediates that are not as active as the native form.
These intermediates are unable to form when the
enzyme is physically attached to a surface.

Hartley et al. [7] proposed that all XIs follow a com-
mon heat inactivation pathway involving irreversible
conversion to an altered apoenzyme that cannot bind
metals, followed by unfolding. The pathway can be
described as T fi T* fi M fi A, where T is the active
tetramer which is converted to the inactive apotetra-
mer (T*). T* then dissociates into monomers (M) that
unfold and form aggregates (A), This can be applied
to the recombinant TNXI, which exists primarily as a
dimer [50], as T « D fi D* fi M fi A. Here the
tetramer (T) is in an equilibrium with the dimer (D),
which is stable until it is irreversibly converted into an
apodimer (D*), which can no longer bind metal. Inac-
tive dimer formation during heat treatment has been
observed for the class I Streptomyces XI [58], for
which inactivation was due to a change in the active
site region. Figure 4 shows that some TNXI remains
in its native form throughout the melting transition.
Presumably, TNXI does not denature until the all the
metals are lost. This pathway, however, would only
describe first order inactivation, as the only active spe-
cies are the tetramer (T) and dimer (D), which have
similar biochemical and biophysical properties for
TNXI [50]. It is possible that the active intermediate is
a form D¢ between D and D*, where only one of the
active sites in the dimer is active. The pathway could
Thermostability of class II xylose isomerases K. L. Epting et al.
1460 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS
be represented T « D fi D¢fiD* fi M fi A, which
would explain the non-first order inactivation. If each

monomer in the dimer inactivated independent of its
partner, the relative residual activity of the intermedi-
ate (b) would be 0.5. As the subunit interaction of
Phe59 is thought to help shield the active site [59], the
inactivation of one monomer may shift this important
residue, thus decreasing the catalytic activity of the
remaining monomer. This is supported by the influence
of temperature on the TNXI inactivation parameter b
(Table 2); at the lower temperatures b is close to 0.5,
but as temperature increases b decreases significantly.
While metals are important in XIs for catalysis, it
appears that their influence on structural stability var-
ies. While certain divalent cations stabilize some XIs
by up to 20 °C, metals played a relatively minor role
in the stabilization of the most thermostable (TNXI)
XI yet identified. The results of this study suggest that
subtle modifications in structure at high temperatures
that result from dissimilar metals bound to binding
sites in TNXI created a structurally stable, but catalyt-
ically inept, form of the enzyme. Additional efforts
with homologous cofactor-requiring enzymes spanning
large functional temperature ranges are needed to see
if the results observed here are more generally applic-
able. It would also be interesting to see whether forms
of TNXI that are structurally stable in the presence of
multiple metal cations (second melting transition)
could be rendered catalytically viable.
Experimental procedures
Bacterial strains
Recombinant TNXI was expressed in E. coli BL21(DE3)

(Novagen, Madison, WI, USA) carrying plasmid
pET22b(+) containing the T. neapolitana 5068 xylA gene
as an Nde I–Hin dIII insert (i.e. plasmid pTNXI22) [50].
E. coli HB101 carrying plasmid pCMG11-3 [28] was used
to overexpress recombinant TTXI. ECXI was expressed in
E. coli JM105 carrying plasmid pKKX7 [25], while BLXI
was expressed in E. coli HB101 carrying plasmid pBL2 [26].
Enzyme purification
All XIs were purified from 1-L cultures grown in LB med-
ium. After centrifugation for 10 min at 4000 g, the cells
were re-suspended in 50 mm Mops (pH 7.0) containing
5mm MgSO
4
plus 0.5 mm CoCl
2
(i.e. buffer A). The cells
were disrupted by two consecutive passes through a French
Pressure cell (Thermo Spectronic, Walthum, MA, USA)
using a pressure drop of 14 000 p.s.i. After centrifugation
at 25 000 g, the supernatant was heat-treated for 15 min at
85 °C (TNXI), 70 °C (TTXI), 60 °C (BLXI), and 50 °C
(ECXI). The precipitated material was separated by centrif-
ugation at 25 000 g for 30 min. The soluble fraction was
loaded on a DEAE–Sepharose Fast-Flow column equili-
brated with buffer A. The protein was eluted with a linear
0–0.5 m NaCl gradient in buffer A, and the active fractions
were analyzed by SDS ⁄ PAGE. Partially purified enzymes
were loaded onto a Q-Sepharose column and eluded with a
linear 0–0.5 m NaCl gradient in buffer A. Active fractions
were combined and concentrated in a stirred ultrafiltration

cell (Amicon, Beverly, MA, USA), dialyzed against buffer
A, and stored at 4 °C. Protein concentrations were assayed
by the method of Bradford [60] using bovine serum albu-
min as the standard.
Site directed mutagenesis
Point mutations were introduced into the T. neapolitana xylA
gene using the QuickChangeÔ Site-Directed Mutagenesis kit
(Stratagene, La Jolla, CA, USA). Residues Glu232 and
Asp309 were substituted with Lys to block the metal binding
sites. The oligonucleotides used for mutagenesis were syn-
thesized by Integrated DNA Technologies (Coralville, IA).
Plasmid pTNXI22 was used as the template for mutagenesis.
Oligonucleotide 5¢-GGACAGTTCCTCATC
AAACCAAAA
CCGAAAGAACCC-3¢ (mutation site underlined) and its
complement were used to construct mutation E232K. Oligo-
nucleotide 5¢-CTTCTTCTTGGATGGGACACC
AAACAG
TTCCCAACAAA-3¢ (mutation site underlined) and its com-
plement were used to construct mutation D309K. The double
mutant was produced by using the plasmid encoding the
E232K mutation as the template and repeating the mutagen-
esis protocol using the D309K primers. Mutations were veri-
fied by DNA sequencing performed by the Integrated
Biotechnology Laboratories Sequencing and Synthesis Facil-
ity (University of Georgia, Athens, GA, USA). The mutant
enzymes were expressed in E. coli BL21(DE3), and purified
following the same procedure as the wild-type enzyme.
EDTA treatment
The purified enzyme was dialyzed overnight at 4 °C against

1 L of 50 mm Mops (pH 7.0) containing 10 mm EDTA. It
was then dialyzed twice against 50 mm Mops (pH 7.0) con-
taining 2 mm EDTA, and finally dialyzed twice against
50 mm Mops (pH 7.0). The apoenzyme was stored at 4 °C
until use.
Differential scanning calorimetry
DSC experiments were performed on a Nano-Cal differen-
tial scanning calorimeter (Calorimetry Sciences Corp.,
Provo, UT, USA). To determine the scan rate, the TNXI
enzyme in buffer A was examined using a scan rate of 0.5
K. L. Epting et al. Thermostability of class II xylose isomerases
FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS 1461
and 1 °C min
)1
. There were no noticeable differences
between the results of the scans, therefore a scan rate of
1 °C min
)1
was used for all the comparative studies. Sam-
ples were scanned from 25 °C to 100 °C for ECXI and
25 °C to 125 °C for TTXI and TNXI. The reversibility of
the thermal transition was checked by reheating the samples
after cooling from the first scan. The apoenzymes were
scanned against 50 mm Mops (pH 7.0). To prepare the sin-
gle metal-containing enzymes, the apoenzyme was dialyzed
at 4 °C overnight against 50 mm Mops (pH 7.0) containing
5mm metal-chloride. The enzyme solution was then dia-
lyzed once against 1 L of 50 mm Mops (pH 7.0) to remove
unbound metal and scanned against the corresponding
dialysis buffer. DSC experiments (with apo- and single

metal-containing enzymes) were conducted with 1.3 ±
0.3 mgÆmL
)1
(TNXI), 1.2 ± 0.6 mgÆmL
)1
(TTXI), and
2.0 ± 0.7 mgÆmL
)1
(ECXI). For mixed-metal DSC experi-
ments, the apoenzyme was dialyzed against 50 mm Mops
(pH 7.0) containing 5 mm MgSO
4
plus either 0.5 mm CoCl
2
or 5 mm MnCl
2
and scanned against the dialysis buffer.
Enzyme concentrations were 0.9 mgÆmL
)1
(TNXI),
1.7 mgÆmL
)1
(TTXI), and 3.0 mgÆmL
)1
(ECXI). Higher
concentrations were used for ECXI because ECXI gave a
weaker signal during its thermal transition.
Enzyme assays
Enzyme activity was assayed routinely with glucose as the
substrate. TNXI (10–20 lg) was incubated in 200 lLof

50 mm Mops (pH 7.0 at room temperature) containing
2mm CoCl
2
and 1.0 m glucose at 80 °C for 10 min. The
reaction was stopped by transferring the tube to an ice
bath. The amount of fructose produced was determined by
the resorcinol–ferric ammonium sulfate–hydrochloric acid
method [61].
Enzyme kinetic inactivation
To determine the effect of specific metals on TNXI kinetic
stability, the apoenzyme (100–200 lgÆmL
)1
final concentra-
tion) was preequilibrated with 0.5 mm CoCl
2
, 0.5 mm
MnCl
2
,or2mm MgCl
2
in 50 mm Mops (pH 7.0) for
30 min at 30 °C (preincubation conditions that are known
to be sufficient for the metal to reach equilibrium between
the buffer and enzyme metal-binding sites [19]). One hun-
dred lL aliquots of the enzyme solution were then incuba-
ted in 0.1 mL MultiplyÒ-Safecup screw-cap microtubes
(Sarstedt, Newton, NC, USA) at various temperatures in
an oil bath for different periods of time. Inactivation was
stopped by transferring tubes to a room-temperature water
bath. Residual activity was determined using the assay des-

cribed above. Non-linear curve fitting of the inactivation
data was performed using the v
2
minimization procedure
of the origin software (Microcal Software, Inc., North-
ampton, MA, USA).
Acknowledgements
This work was supported in part through grants from
the NSF to RMK (Bes-0115734 and Bes-0317886), and
to CV ⁄ JGZ (Bes-0115754).
References
1 Vieille C & Zeikus J (2001) Hyperthermophilic enzymes:
sources, uses, and molecular mechanisms for thermo-
stability. Microbiol Mol Biol Rev 65, 1–43.
2 Ladenstein R & Antranikian G (1998) Proteins from
hyperthermophiles: stability and enzymatic catalysis
close to the boiling point of water. Adv Biochem Eng ⁄
Biotechnol 61, 37–85.
3 Sterner R & Liebl W (2001) Thermophilic adaptation of
proteins. Crit Rev Biochem Mol Biol 36, 39–106.
4 Petsko GA (2001) Structural basis of thermostability
in hyperthermophilic proteins, or ‘there’s more than
one way to skin a cat’. Methods Enzymol 334,
469–478.
5 Danson MJ & Hough DW (1998) Structure, function
and stability of enzymes from the archaea. Trends
Microbiol 6, 307–314.
6 Lebbink JHG, Kengen SWM, van der Oost J & de Vos
WM (1999) Glutamate dehydrogenase from hyperther-
mophilic bacteria and archaea: determinants of thermo-

stability and catalysis at extremely high temperatures.
J Mol Catal B: Enzym 7, 133–145.
7 Hartley BS, Hanlon N, Jackson RJ & Rangarajan M
(2000) Glucose isomerase: insights into protein engineer-
ing for increased thermostability. Biochim Biophys Acta
1543, 294–335.
8 Takasaki Y, Kosugi Y & Kanbayashi A (1969) Studies
on sugar-isomerizing enzyme: purification, crystalliza-
tion, and some properties of glucose isomerase from
Streptomyces sp. Agr Biol Chem 33, 1527–1534.
9 Bentley IS & Williams EC (1996) Starch conversion. In
Industrial Enzymology (Godfrey T & West SI, eds), pp.
339–357. Stockton Press, New York.
10 Bhosale SH, Rao MB & Deshpande VV (1996) Mole-
cular and industrial aspects of glucose isomerase. Micro-
biol Rev 60, 280–300.
11 Chen WP (1980) Glucose Isomerase (a review). Process
Biochem 15, 30–41.
12 Banerjee S, Archana A & Satyanarayana T (1994)
Xylose metabolism in a thermophilic mold Malbran-
chea-Pulchella var. sulfurea Tmd-8. Curr Microbiol 29,
349–352.
13 Harhangi HR, Akhmanova AS, Emmens R, van der
Drift C, de Laat WTAM, van Dijken JP, Jetten MSM,
Pronk JT & den Camp HJMO (2003) Xylose metabo-
lism in the anaerobic fungus Piromyces sp. strain E2
follows the bacterial pathway. Arch Microbiol 180,
134–141.
Thermostability of class II xylose isomerases K. L. Epting et al.
1462 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS

14 Kristo P, Saarelainen R, Fagerstrom R, Aho S & Kor-
hola M (1996) Protein purification, and cloning and
characterization of the cDNA and gene for xylose iso-
merase of barley. Eur J Biochem 237, 240–246.
15 Vangrysperre W, Vandamme J, Vandekerckhove J,
Debruyne CK, Cornelis R & Kerstershilderson H (1990)
Localization of the essential histidine and carboxylate
group in d-xylose isomerases. Biochem J 265, 699–705.
16 Bogumil R, Kappl R & Huttermann J (2000) Role of
the binuclear manganese (II) site in xylose isomerase. In
Metal Ions in Biological Systems (Sigel H & Sigel A,
eds), pp. 365–405.
17 Callens M, Kerstershilderson H, Vangrysperre W &
Debruyne CK (1988) d-xylose isomerase from Strepto-
myces violaceoruber: structural and catalytic roles of
bivalent-metal ions. Enzyme Microb Technol 10, 695–
700.
18 Marg GA & Clark DS (1990) Activation of glucose iso-
merase by divalent cations: evidence for two distinct
metal binding sites. Enz Microb Technol 12, 367–373.
19 Van Bastelaere PBM, Callens M, Vangrysperre WAE &
Kersters-Hilderson HLM (1992) Binding characteristics
of Mn
2+
,Co
2+
, and Mg
2+
ions with several d-xylose
isomerases. Biochem J 286, 729–735.

20 Carrell HL, Glusker JP, Burger V, Manfre F, Tritsch D
& Biellmann JF (1989) X-ray analysis of d-xylose iso-
merase at 1.9 A
˚
native enzyme in complex with sub-
strate and with a mechanism-designed inactivator. Proc
Natl Acad Sci USA 86, 4440–4444.
21 Jenkins J, Janin J, Rey F, Chiadmi M, van Tilbeurgh
H, Lasters I, De Maeyer M, Van Belle D, Wodak SJ,
Lauwereys M, et al. (1992) Protein engineering of xylose
(glucose) isomerase from Actinoplanes missouriensis.1.
Crystallography and site-directed mutagenesis of metal
binding sites. Biochemistry 31, 5449–5458.
22 Henrick K, Collyer CA & Blow DM (1989) Structures of
d-xylose isomerase from Arthrobacter strain B3728 con-
taining the inhibitors xylitol and d-sorbitol at 2.5 and 2.3
A
˚
resolution, respectively. J Mol Biol 208, 129–157.
23 Lavie A, Allen KN, Petsko GA & Ringe D (1994)
X-ray crystallographic structures of d-xylose isomerase:
substrate complexes position the substrate and provide
evidence for metal movement during catalysis. Biochem-
istry 33, 5469–5480.
24 Chang CS, Park BC, Lee DS & Suh SW (1999) Crystal
structures of thermostable xylose isomerases from Ther-
mus caldophilus and Thermus thermophilus: possible
structural determinants of thermostability. J Mol Biol
288, 623–634.
25 Batt CA, Jamieson AC & Vandeyar MA (1990) Identifi-

cation of essential histidine residues in the active site of
Eschericia coli xylose (glucose) isomerase. Proc Natl
Acad Sci USA 87, 618–622.
26 Vieille C, Epting KL, Kelly RM & Zeikus JG (2001)
Bivalent cations and amino acid composition contribute
to the thermostability of Bacillus licheniformis xylose
isomerase. Eur J Biochem 268 , 6291–6301.
27 Brown SH, Sjoholm C & Kelly RM (1993) Purification
and characterization of a highly thermostable glucose-
isomerase produced by the extremely thermophilic
eubacterium, Thermotoga maritima. Biotechnol Bioeng
41, 878–886.
28 Lee C, Bhatnagar L, Saha BC, Lee YE, Takagi M,
Imanaka T, Bagdasarian M & Zeikus JG (1990) Clon-
ing and expression of the Clostridium (Thermoanerobac-
ter) thermosulfurogenes glucose isomerase gene in
Eshericia coli and Bacillus subtilis. App Environ Micro-
biol 56, 2638–2643.
29 Vieille C, Hess JM, Kelly RM & Zeikus JG (1995)
XylA cloning and sequencing and biochemical-charac-
terization of xylose isomerase from Thermotoga neapoli-
tana. Appl Environ Microbiol 61, 1867–1875.
30 Sriprapundh D, Vieille C & Zeikus JG (2000) Molecular
determinants of xylose isomerase thermal stability and
activity: analysis of thermozymes by site-directed muta-
genesis. Protein Eng 13, 259–265.
31 Allen KN, Lavie A, Glasfeld A, Tanada TN, Gerrity
DP, Car SC, Farber GK, Petsko GA & Ringe D (1994)
Role of the divalent metal ion in sugar binding, ring
opening, and isomerization by d-xylose isomerase:

replacement of a catalytic metal by and amino acid.
Biochemistry 33, 1488–1494.
32 Collyer CA, Henrick K & Blow DM (1990) Mechanism
for aldose-ketose interconversion by d-xylose isomerase
involving ring-opening followed by a 1,2-hydride shift.
J Mol Biol 212, 211–235.
33 Whitlow M, Howard AJ, Finzel BC, Poulos TL,
Winborne E & Gilliland GL (1991) A metal-mediated
hydride shift mechanism for xylose isomerase based on
the 1.6 A
˚
Streptomyces rubiginosus structures with
xylitol and d-xylose. Proteins 9, 153–173.
34 Lambeir AM, Lauwereys M, Stanssens P, Mrabet NT,
Snauwaert J, van Tilbeurgh H, Matthyssens G, Lasters
I, De Maeyer M, Wodak SJ, et al. (1992) Protein engin-
eering of xylose (glucose) isomerase from Actinoplanes
missouriensis. 2. Site-directed mutagenesis of the xylose
binding site. Biochemistry 31, 5459–5466.
35 Bogumil R, Kappl R, Huttermann J & Sudfeldt C
(1993) X- and Q-band EPR studies on the two Mn
2+
substituted metal-binding sites of d-xylose isomerase.
Eur J Biochem 213, 118–1192.
36 Sudfeld C, Schaffer A, Kagi JHR, Bogumil R, Schulz
H, Wulff S & Witzel H (1990) Spectroscopic studies on
the metal-ion-binding sites of Co
2+
substituted d-xylose
isomerase from Streptomyces rubiginosus. Eur J Biochem

193, 863–871.
37 Van Bastelaere PBM, Kersters-Hilderson HLM &
Lambeir AM (1995) Wild-type and mutant d-xylose
isomerase from Actinoplanes missouriensis: metal-ion
dissociation constants, kinetic parameters of deuterated
K. L. Epting et al. Thermostability of class II xylose isomerases
FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS 1463
and non-deuterated substrates and solvent-isotope
effects. Biochem J 307, 135–142.
38 Van Bastelaere P, Vangrysperre W & Kerstershilderson
H (1991) Kinetic studies of Mg
2+
,Co
2+
, and Mn
2+
d-xylose isomerases. Biochem J 278, 285–292.
39 Gaikwad SM, Rao MB & Deshpande VV (1992) d-glu-
cose (xylose) isomerase from Streptomyces: differential
roles of magnesium and cobalt ions. Enzyme Microb
Technol 14, 317–320.
40 Cha J, Cho Y, Whitaker RD, Carrell HL, Glusker JP,
Karplus PA & Batt CA (1994) Perturbing the metal site
in d-xylose isomerase: effect of mutations of His-220 on
enzyme stability. J Biol Chem 269, 2687–2694.
41 Cha J (1999) Contribution of second metal binding site
for metal specificity of d-xylose isomerase. J Microbiol
Biotechnol 9, 757–763.
42 van Tilbeurgh H, Jenkins J, Chiadmi M, Janin J,
Wodak SJ, Mrabet NT & Lambeir AM (1992) Protein

engineering of xylose (glucose) isomerase from Actino-
planes missouriensis. 3. Changing metal specificity and
the pH profile by site-directed mutagenesis. Biochemistry
31, 5467–5471.
43 Gong C-S, Chen LF & Tsao GT (1980) Purification and
properties of glucose isomerase of Actinoplanes missour-
iensis. Biotechnol Bioeng 22, 833–845.
44 Boguslawski G (1983) Purification and characterization
of glucose isomerase from Flavobacterium arborescens.
J Appl Biochem 5 , 186–196.
45 Deshmukh SS & Shankar V (1996) Glucose isomerase
from thermophilic Streptomyces thermonitrificans: purifi-
cation and characterization. Biotechnol Appl Biochem
24, 65–72.
46 Sriprapundh D, Vieille C & Zeikus JG (2003) Directed
evolution of Thermotoga neapolitana xylose isomerase:
high activity on glucose at low temperature and low
pH. Protein Eng 16, 683–690.
47 Sadana A (1991) Biocatalysis: Fundimentals of Enzyme
Deactivation Kinetics. Prentice Hall, Englewood Cliffs,
NJ.
48 Sehgal AC, Tompson R, Cavanagh J & Kelly RM
(2002) Structural and catalytic response to temperature
and cosolvents of carboxylesterase EST1 from the extre-
mely thermoacidophilic archaeon Sulfolobus solfataricus
P1. Biotechnol Bioeng 80, 784–793.
49 Tucker MY, Tucker MP, Himmel ME, Grohmann K &
Lastick SM (1988) Properties of genetically overproduced
E. coli xylose isomerase. Biotechnol Lett 10, 79–84.
50 Hess JM, Tchernajenko V, Vieille C, Zeikus JG & Kelly

RM (1998) Thermotoga neapolitana homotetrameric
xylose isomerase is expressed as a catalytically active
and thermostable dimer in Escherichia coli. Appl Environ
Microbiol 64, 2357–2360.
51 Hess JM & Kelly RM (1999) Influence of polymolecular
events on inactivation behavior of xylose isomerase
from Thermotoga neapolitana 5068. Biotechnol Bioeng
62, 509–517.
52 Bogumil R, Huttermann J, Kappl R, Stabler R, Sud-
feldt C & Witzel H (1991) Visible, EPR and electron
nuclear double-resonance spectroscopic studies on the
two metal-binding sites of oxovanadium (IV) -substi-
tuted d -xylose isomerase. Eur J Biochem 196, 305–312.
53 Siddiqui KS, Rangarajan M, Hartley BS, Kitmitto A,
Panico M, Blench IP & Morris HR (1993) Arthrobacter
d-xylose isomerase: partial proteolysis with thermolysin.
Biochem J 289, 201–208.
54 Whitaker RD, Cho YJ, Cha JH, Carrell HL, Glusker
JP, Karplus PA & Batt CA (1995) Probing the roles of
active-site residues in d-xylose isomerase. J Biol Chem
270, 22895–22906.
55 Fuxreiter M, Bocskei Z, Szeibert A, Szabo E, Dallmann
G, Naray-Szabo G & Asboth B (1997) Role of electro-
statics at the catalytic metal binding site in xylose iso-
merase action: Ca
2+
inhibition and metal competence in
the double mutant D254E ⁄ D256E. Proteins 28, 183–
193.
56 Varsani L, Cui T, Rangarajan M, Hartley BS, Goldberg

J, Collyer C & Blow DM (1993) Arthrobacter d-xylose
isomerase: protein–engineered subunit interfaces. Bio-
chem J 291, 575–583.
57 Bandlish RK, Hess JM, Epting KL, Vieille C & Kelly
RM (2002) Glucose to fructose conversion at high tem-
peratures with xylose (glucose) isomerases from Strepto-
myces murinus and two hyperthermophilic Thermotoga
species. Biotechnol Bioeng 80, 185–194.
58 Gaikwad S, Rao M & Deshpande V (1993) Structure-
function relationship of glucose (xylose) isomerase from
Streptomyces: evidence for the occurrence of inactive
dimer. Enzyme Microb Technol 15, 155–157.
59 Farber GK, Glasfeld A, Tiraby G, Ringe D & Petsko
GA (1989) Crystallographic studies of the mechanism of
xylose isomerase. Biochemistry 28, 7289–7297.
60 Bradford MM (1976) A rapid and sensitive method for
quantitation of microgram quantities of protein utilizing
the principle of protein-dye binding. Anal Biochem 72,
248–254.
61 Schenk M & Bisswanger H (1998) A microplate assay
for d-xylose ⁄ d-glucose isomerase. Enzyme Microb Tech-
nol 22, 721–723.
Thermostability of class II xylose isomerases K. L. Epting et al.
1464 FEBS Journal 272 (2005) 1454–1464 ª 2005 FEBS