Conformational stability of the N-terminal amino acid residues of mutated recombinant pigeon liver malic enzymes.

Abstract

Pigeon liver malic enzyme has an N-terminal amino acid sequence of Met-Lys-Lys-Gly-Tyr-Glu-Val-Leu-Arg-. Our previous results indicated that the N-terminus of the enzyme is located at or near the enzyme's active center involved in Mn(II)-L-malate binding and is also near to the subunits' interface. In the present study, the conformational stability of the various deletion (delta) and substitution mutants at Lys2/Lys3 of the enzyme was investigated with chemical and thermal sensitivities. The lysine residue at position 2 or 3 seems to be crucial for the correct active site conformation, probably through an ion-pairing with Glu6. Deletion at Lys2 or Lys3, delta(K2/K3), and the double mutant K(2,3)E were much less stable than the wild-type enzyme towards chemical denaturation. Kinetic analysis of the thermal inactivation at 58 degrees C of the recombinant enzymes indicated that mutation at position 3 to alanine (K3A) endows the protein with extra stability compared with the wild-type enzyme. K3A is also stable towards chemical denaturation. The concentration of urea that causes half unfolding, [urea]0.5, for K3A is 3.25 M compared with 2.54 M for the wild-type enzyme. The K3A mutant of malic enzyme might therefore have potential practical applications.

Full text

Page 1

Protein Engineering vol.11 no.5 pp.371–376, 1998
Conformational stability of the N-terminal amino acid residues of
mutated recombinant pigeon liver malic enzymes
Wei-Yuan Chou1, Shih-Ming Huang and
Gu-Gang Chang1
Department of Biochemistry, National Defense Medical Center, Taipei 100,
Taiwan
1To whom correspondence should be addressed
Pigeon liver malic enzyme has an N-terminal amino acid
sequence of Met-Lys-Lys-Gly-Tyr-Glu-Val-Leu-Arg-. Our
previousresultsindicatedthattheN-terminusoftheenzyme
is located at or near the enzyme’s active center involved
in Mn(II)–L-malate binding and is also near to the subunits’
interface. In the present study, the conformational stability
of the various deletion (∆) and substitution mutants at
Lys2/Lys3 of the enzyme was investigated with chemical
and thermal sensitivities. The lysine residue at position 2 or
3seemstobecrucialforthecorrectactivesiteconformation,
probably through an ion-pairing with Glu6. Deletion at
Lys2 or Lys3, ∆(K2/K3), and the double mutant K(2,3)E
were much less stable than the wild-type enzyme towards
chemical denaturation. Kinetic analysis of the thermal
inactivation at 58°C of the recombinant enzymes indicated
that mutation at position 3 to alanine (K3A) endows the
protein with extra stability compared with the wild-type
enzyme. K3A is also stable towards chemical denaturation.
The concentration of urea that causes half unfolding,
[urea]0.5, for K3A is 3.25 M compared with 2.54 M for the
wild-type enzyme. The K3A mutant of malic enzyme might
therefore have potential practical applications.
Keywords: enzyme engineering/malic enzyme/N-terminal amino
acid/proteinstability/proteinunfolding/site-specificmutagenesis/
unfolding intermediate
Introduction
The function of a protein relies entirely on its correct three-
dimensional structure, which is maintained primarily by non-
covalent bonding (Timasheff, 1993). Among the major con-
cerns in protein engineering is the seeking of a functional
mutant protein that is more stable than the wild-type enzyme
(WT). Unfortunately, the amino acid residues involved in
function are not optimized for stability and there seems to be
an inverse relationship between the effect of mutation on the
activity and stability (Shoichet et al., 1995). In this paper, we
identify a stable mutant of malic enzyme which has comparable
specific activity to the WT.
L-Malate is an important product in the food industry and
has been successfully industrialized with an immobilized
biocatalyst system (Tosa and Shibatani, 1995). Cytosolic malic
enzyme(EC1.1.1.40)catalyzesthereversibleoxidativedecarb-
oxylation of L-malateto yield pyruvate andCO2with concomit-
ant reduction of NADP?to NADPH. A stabilized malic
enzymemayhavepotentialapplicationinthebiotransformation
of NADP?to NADPH. We have immobilized malic enzyme
on Sepharose beads and found that the immobilized malic
© Oxford University Press
371
enzyme is more stable than the soluble form (Chang et al.,
1993). However, the immobilized enzyme lost ~50% of its
original activity. Alternative biotechniques are worthy of fur-
ther exploration.
We have isolated the full-length pigeon liver malic enzyme
cDNA and successfully expressed it in Escherichia coli cells
(Chou et al., 1994). The complete polypeptide chain of each
subunit is composed of 557 amino acids with Met-Lys-Lys-
Gly-Tyr-Glu-Val-Leu-Arg- as the N-terminus (Chou et al.,
1994). More recently, we have characterized the possible
involvement of Lys2/Lys3 in the Mn(II)–L-malate binding and
the subunits’ interactions (Chou et al., 1997). Arg9 was also
proposed to be critical for the Mn(II)–L-malate binding (Huang
et al., 1998). To clarify further the roles of N-terminal amino
acid residues in the active center of malic enzyme and to find
a stable mutant of this enzyme, the conformational stability of
various mutants around Lys2/Lys3 were characterized in this
work. We found that K3A is stable towards chemical and
thermal denaturation and is a potential candidate for future
practical applications.
Materials and methods
Site-specific mutagenesis, expression and purification of the
recombinant pigeon liver malic enzymes
Site-specific mutagenesis of pigeon liver malic enzyme was
carried out by polymerase chain reaction (PCR) as described
previously (Chou et al., 1997). The entire mutated cDNAs
were examined by dideoxy chain termination sequencing
(Sanger et al., 1977) to exclude any unexpected mutations
resulting from the PCR.
The mutated plasmids were transformed into BL21 bacteria.
The expressed recombinant proteins did not fuse with any tag
fragment or fusing protein introduced by the pET-21b vector;
after sonication, they were purified stepwise on a Q-Sepharose
and an adenosine 2?,5?-bisphosphate agarose column according
to a published procedure (Chang et al., 1991). All purified
enzymes were subjected to SDS–PAGE to examine the purity
(Wei et al., 1994).
Electrophoresis under non-denaturing conditions was per-
formed in duplicate with an 8–25% gradient gel in Pharmacia-
Biotech PhastSystem at 350 V h as described by Chang et al.
(1994). After electrophoresis, one gel was stained for protein
with Coomassie Brilliant Blue R-250 and the other gel plate
was stained with enzymatic activity (Chang et al., 1994).
Abbreviations
∆(K2/K3), single deletion mutant of the enzyme with deletion
at Lys2 or Lys3, which will yield the same mutant; ∆(K2K3),
double deletion mutant of the enzyme with deletions at Lys2
and Lys3; ∆G4, single deletion mutant of the enzyme with
deletion at Gly4; ∆(K2G4/K3G4), double deletion mutant of
the enzyme with deletions at Lys2 and Gly4 or Lys3 and
Gly4; ∆(K2K3G4), triple deletion mutant of the enzyme
with deletions at Lys2, Lys3 and Gly4; ∆(1–16), N-terminal
by guest on April 11, 2012
http://peds.oxfordjournals.org/
Downloaded from

Page 2

W.-Y.Chou et al.
truncated mutant with deletion of N-terminal amino acid
residues 1–16; K2A, K3A, K2E, K3E and Y5A, point mutants
with lysine at positions 2 or 3 and tyrosine at position 5
replaced by alanine (A) or glutamate (E), respectively; K(2,3)A
and K(2,3)E, double mutants with Lys2 and Lys3 replaced
by alanine or glutamate, respectively. The native (folded),
intermediate and unfolded forms of the enzyme are represented
by N, I and U, respectively.
Enzyme assay and protein determination
Malic enzyme activity was assayed according to a published
procedure (Chang et al., 1992). A 1 ml volume of reaction
mixture contained triethanolamine–HCl buffer (66.7 mM, pH
7.4), L-malate (0.419 mM, after correction for the Mn2?
chelation), NADP?(0.177 mM, after correction for the Mn2?
chelation), Mn2?(3.87 mM, after correction for the L-malate
and NADP?chelations) and an appropriate amount of enzyme.
The formation of NADPH at 30°C was monitored continuously
at 340 nm with a Perkin-Elmer Lambda 3B spectrophotometer.
One unit of enzyme activity was defined as the amount of
enzyme that catalyzed an initial rate of 1 µmol NADPH formed
per minute under the assay conditions. A molar absorption
coefficient of 6.22?103M–1cm–1for the NADPH was used
in calculations.
Protein concentration was determined by the protein–dye
binding method of Bradford (1976) using purified natural
pigeon liver malic enzyme (Chang and Chang, 1982) as
standard. An Mrof 260 000 for the tetrameric enzyme was
used in calculations.
Chemical stability of the recombinant malic enzymes in urea
solution
The urea solutions were freshly prepared daily. WT or mutant
malic enzymes were exposed to various concentrations of urea
at 30°C for 30 min and assayed for residual enzyme activity.
Our previous results established that the enzymatic activity is
a more sensitive probe than fluorescence in monitoring the
conformational changes of malic enzyme (Huang et al., 1998),
which confirmed that for a more complex protein, physical
probes are less suitable and functional probes are better
adapted (Garel, 1992). Furthermore, we have demonstrated
that denaturation of malic enzyme in urea is an instantaneous
process; 30 min of incubation is sufficient for the unfolding
process to reach an equilibrium state (Huang et al., 1998).
Control experiments were simultaneously performed with the
same amount of urea being added in the assay mixture.
Residual enzyme activity (Et/E0) was plotted versus urea
concentration, where E0and Etare the enzyme activity in the
control and experimental groups, respectively.
The whole data sets were fitted to Equation 1, which
describes a two-state unfolding mechanism with the unfolding
curve that has pre- and post-transitional baselines (Scheme 1)
(Santoro and Bolen, 1988; Pace, 1990).
Scheme 1
KN–U
≥
NU
(yN ?mN[D]) ? (yU? mU[D]) exp{–(∆G(H2O)–
m [D])/RT}
1 ? exp{ – (∆G(H2O)– m [D])/RT}
Yobs?
(1)
where yobsis the observed signal change, yN, yU, mNand mU
372
are the intercepts and slopes of the pre- and post-transitional
baselines, respectively; yNand yUrepresent the signals of the
folded and unfolded states, respectively. ∆G(H2O)denotes the
intrinsic free energy change in the absence of denaturant and
m represents the dependence of the ∆G on denaturant. [D]
denotes urea concentration. T is the absolute temperature and
R is the gas constant. The concentration of urea that causes
half unfolding, [urea]0.5, was estimated by dividing ∆G(H2O)by
m (Pace et al., 1989; Pace, 1990).
The denaturation curves were also analyzed by a three-state
model (Scheme 2) in which the whole data sets were fitted to
Equation 2 (Morjana et al., 1993).
Scheme 2
KN–I
≤
KI–U
≤
N
|
U
yN? yIexp{ – (∆G(H2O),N→I– mN→I[D])/RT} ?
yUexp{ – (∆G(H2O),N→I– mN→I[D])/RT}
exp{ – (∆G(H2O),I→U– mI→U[D])/RT}
1 ? exp{ – (∆G(H2O),N→I– mN→I[D])/RT} ?
exp{ – (∆G(H2O),N→I– mN→I[D])/RT}
exp{ – (∆G(H2O),I→U– mI→U[D])/RT}
yobs?
(2)
where yIis the intrinsic signal of the putative intermediate
state I. ∆G(H2O),N→I and ∆G(H2O),I→Uare the free energies
extrapolating to [D] ? 0 for the N → I and I → U
processes, respectively. mN→Iand mI→Uare the m values for
the corresponding processes.
The unfolding curves in the transition region were also
analyzed by the procedure of Pace et al. (1989). The apparent
equilibrium constant (Kapp) during the unfolding process was
estimated according to Equation 3 (Pace, 1986).
Kapp? fU/(1 – fU) ? fU/fN? (yN– y)/(y – yU)
where fNand fUare the fractions of enzyme that are present
in the folded and unfolded states, respectively. The linear free
energy relationships of the unfolding process were tested by
fitting the ∆G data to Equation 4 (Pace, 1986; Santoro and
Bolen, 1988).
∆G ? ∆G(H2O)– m[denaturant]
where ∆G ? –RTlnKapp. In cases where the ∆G versus
denaturant concentration plot was non-linear, Equation 5 was
applied (Matouschek et al., 1994).
∆G ? ∆G(H2O)– m1[denaturant] ? m2[denaturant]2
(3)
(4)
(5)
Thermal stability of the recombinant malic enzymes
The sensitivity of WT and various mutants of malic enzyme
to thermal denaturation was examined by heating the enzyme
solution at 58°C for various time intervals. The logarithm of
residue enzyme activity was plotted versus incubation time,
which was fitted to a double exponential equation (Equation
6) according to a three-state model (Scheme 3) (Violet and
Meunier, 1989).
Scheme 3
k1
k2
N
|
U
→→
RA ? [EAN– EAIk1/(k1– k2)] e–k1t? [EAIk1/(k1– k2)] e–k2t
(6)
by guest on April 11, 2012
http://peds.oxfordjournals.org/
Downloaded from

Page 3

Mutation of malic enzyme
where k1and k2are the inactivation rate constants for the
N → I and I → U processes, respectively; RA is the residual
enzyme activity observed at time t and EANand EAIare the
fractional enzyme activities of the native and the intermediate
forms of the enzyme with respect to the original enzyme
activity, respectively.
Results
Expression and purification of the recombinant pigeon liver
malic enzyme
The WT and the various deletion and substitution mutants of
malic enzyme were successfully expressed and purified to
apparent homogeneity by two successive Q-Sepharose and
adenosine-2?,5?-bisphosphate agarose columns. A major pro-
tein band corresponding to an Mrof 65 000 was observed in
SDS–PAGE for all recombinant enzymes. On examining the
enzyme by PAGE under native conditions, all recombinant
enzymes were shown to possess a tetrameric structure with
Mr260 000 or a mixture of tetramers and dimers. The only
exceptions were ∆(1–16) and ∆(K2K3G4), which possessed a
dimeric structure (Chou et al., 1997).
Chemical stability of the cloned malic enzymes
The mutants’ sensitivity towards chemical denaturation was
examined by monitoring the unfolding process in urea solution.
Figure 1 shows the unfolding curves for WT and various
mutants according to a two-state model (Scheme 1) or a
three-state model (Scheme 2). Although the data seem to fit
excellently to Equations 1 and 2, it should be noted that
because physical probe fluorescence changes did not reveal
the structural details in malic enzyme (Huang et al., 1998)
and residual enzyme activity was used as the structural probe,
which, unfortunately, also reflects the irreversible denaturation,
this therefore precludes the use of enzyme activity data for
the calculation of absolute thermodynamic values relating to
protein stability. In practice, the ∆G(H2O)values deviated greatly
from each other. This is because some mutants are multi-state,
which causes the m values of some mutants to deviate from
the WT. Nevertheless, examination of the unfolding curves
allows a qualitative comparison of the relative stability between
mutants and WT malic enzyme. We therefore used [urea]0.5
as an index of protein stability. As judged by eye, it is not
difficult to differentiate the unfolding pattern and chemical
stability between the WT and various mutants.
The mutants can be divided into two groups depending on
the unfolding mechanism. The group I mutants conform to a
two-state unfolding mechanism (Table I); within this group,
the K3A and ∆(1–16) mutants are more stable than WT (Figure
1a). The [urea]0.5values for K3A and ∆(1–16) were 3.25 and
3.15 M, respectively, which were significantly larger than that
for the WT ( [urea]0.5? 2.54 M). ∆(K2K3) was slightly more
stable than WT with [urea]0.5? 2.70 M. However, examination
of the unfolding curve of ∆(K2K3) (Figure 1a) indicates that
this mutant may be better grouped with other mutants in this
group, which contains ∆G4, ∆(K2G4/K3G4), ∆(K2K3G4) and
K(2,3)A. These mutants are slightly more unstable than WT
(Figure 1b). On the other hand, group II mutants, consisting
of Y5A, ∆(K2/K3) and K(2,3)E, conform to a three-state
mechanism, which suggests that at least an intermediate state
isinvolvedintheunfoldingprocess.Themulti-phasicunfolding
curve for K(2,3)E is more pronounced than that for ∆(K2/K3)
or Y5A. The unfolding data of the latter two mutants fit better
to a three-state model, especially at high urea concentrations.
373
Fig. 1. Chemical denaturation of the recombinant pigeon liver malic
enzymes by urea. WT and mutant malic enzymes in Tris–HCl buffer (30
mM, pH 7.4) were mixed with various concentrations of urea. The residual
enzyme activity was recorded after 30 min at 30°C. (a) Group I mutants,
which are stable and conform to a two-state unfolding mechanism. (s) WT
(17.2 µg/ml); (d) K3A (10 µg/ml); (u) ∆(1–16) (14 µg/ml); and (n)
∆(K2K3) (22.3 µg/ml). (b) Group I mutants, which are unstable and
conform to a two-state unfolding mechanism. (s) WT; (d) ∆G4 (10.7 µg/
ml); (u) ∆(K2G4/K3G4) (33.6 µg/ml); (j) ∆(K2K3G4) (14.7 µg/ml); and
(n) K(2,3)A (13.8 µg/ml). (c) Group II mutants, which are unstable and
conform to a three-state unfolding model. (s) WT; (d) Y5A (12.1 µg/ml);
(u) K(2,3)E (20 µg/ml); and (n) ∆(K2/K3) (47.7 µg/ml) mutants. The
symbols are data points. The solid lines through these data are computer-
fitted results according to Equation 1 in (a) and (b) and Equation 2 in (c).
The dashed lines in (c) are fitted results for WT, Y5A and ∆(K2/K3)
according to Equation 1. The fitted curve for K(2,3)E according to
Equation 1 was poor and is not shown in (c). For comparison, WT (s) is
shown in all panels.
The calculated parameters for ∆(K2/K3) and Y5A are more
satisfactory when fitted to a three- than to a two-state model.
The unfolding data for WT can be explained by both models.
by guest on April 11, 2012
http://peds.oxfordjournals.org/
Downloaded from

Page 4

W.-Y.Chou et al.
Table I. Chemical stability of the recombinant pigeon liver malic enzymes in urea denaturation
Group Mutant[Urea]0.5,N→U(M)[Urea]0.5,N→I(M) [Urea]0.5,I→U(M)
Group Ia
WTb
K3A
∆(1–16)
∆(K2K3)
∆(K2G4/K3G4)
∆(K2K3G4)
∆G4
K(2,3)A
WTb
Y5A
∆(K2/K3)
K(2,3)E
2.54 ? 0.15
3.25 ? 0.20
3.15 ? 0.21
2.70 ? 0.29
2.50 ? 0.16
2.45 ? 0.46
2.42 ? 0.22
2.32 ? 0.10
Group IIc
2.52 ? 0.7
2.05 ? 0.12
0.97 ? 0.06
0.24 ? 0.12
3.22 ? 3.35
4.34 ? 1.41
2.17 ? 0.59
2.79 ? 0.56
aMutants conform to a two-state unfolding model.
bThe results fitting to both two- and three-state models are presented for WT.
cMutants conform to a three-state unfolding model.
However, fitting the data to a three-state model produces a
large error in the [urea]0.5,I→Uvalue (Table I).
The results shown in Figure 1c clearly indicate the shift of
the K(2,3)E and ∆(K2/K3) denaturation curves to lower urea
concentration (medium points at 0.24–2.79 and 0.97–2.17 M
for K(2,3)E and ∆(K2/K3), respectively, compared with the
WT at 2.52–3.22 M (Table I).
Further evidence for the Y5A, ∆(K2/K3) and K(2,3)E
mutants conforming to a three-state model in chemical dena-
turation is shown in Figure 2b, which illustrates a non-linear
free energy relationship for these mutants, thereby implying
the existence of unstable, native-like folding intermediates
(Jonsson et al., 1996). WT and ∆G4 also demonstrate slight
curvature in the ∆G versus [urea] plot (Figure 2b). Other
mutants exhibited apparently linear ∆G versus [urea] plots
(Figure 2a).
Thermal stability of the cloned malic enzymes
The above results suggest a multi-step unfolding process for
some of the mutant malic enzymes. This conclusion was
further corroborated by the thermal denaturation experiments.
The K3A and ∆(1–16) mutants are also much more stable than
WT towards thermal inactivation. The kinetics of the thermal
denaturation of WT and mutant malic enzymes were examined
by monitoring the enzyme activity loss at 58°C versus incuba-
tion time. A semilogarithmic plot of the data was not linear,
indicating that the inactivation process does not follow first-
order kinetics and hence is not compatible with a two-state
model. The data were then fitted to an equation (Equation 6)
that describes a three-state model (Scheme 3).
All recombinant malic enzymes are fitted well to Equation
6 (Table II). The unstable mutants in urea, ∆(K2/K3) and
K(2,3)E, have very small k2values for the I → U process,
indicating an extremely stable intermediate state in these
two mutants. However, it should be noted that the thermal
inactivation is an irreversible process. The intermediate state
detected in the thermal inactivation does not necessarily
represent the same form that is observed in the chemical
denaturation.
Quaternary structure of the recombinant malic enzymes
during chemical denaturation
The quaternary structure of WT and various mutants during
urea denaturation was examined by SDS–PAGE under non-
denaturing conditions. WT starts to dissociate at a urea
374
Fig. 2. Free energy relationships of unfolding of the recombinant pigeon
liver malic enzymes. The Gibb’s free energy (∆G) values in the transitional
region shown in Figure 1 were plotted versus urea concentration. (a)
Mutants that conform to a linear free energy relationship. (s) K(2,3)A;
(d) ∆(K2K3); (u), ∆(K2G4/K3G4); (j) ∆(K2K3G4); (n) K3A; and
(m) ∆(1–16). (b) WT and mutants that conform to a non-linear free energy
relationship. (s) WT; (d) ∆(K2/K3); (u) Y5A; (j) ∆G4; and (n) K(2,3)E.
The symbols are data points. The lines through these data are computer-
fitted results according to Equation 4 in (a) and Equation 5 in (b).
by guest on April 11, 2012
http://peds.oxfordjournals.org/
Downloaded from

Page 5

Mutation of malic enzyme
Table II. Inactivation of recombinant wild-type and various mutants of pigeon liver malic enzyme at 58°C
MutantResidual activity (%)a
EAN
EAI
k1,N→I(min–1)
k2,I→U(min–1)
Wild-type
∆(K2/K3)
∆G4
∆(K2K3)
∆(K2G4/K3G4)
∆(K2K3G4)
∆(1–16)
K2A
K2E
K3A
K3E
Y5A
K(2,3)A
K(2,3)E
2
3
2
1.00 ? 0.03
1.09 ? 0.13
1.00 ? 0.03
1.00 ? 0.03
1.00 ? 0.01
1.00 ? 0.03
1.00 ? 0.02
1.00 ? 0.04
1.00 ? 0.02
0.99 ? 0.02
1.00 ? 0.01
1.00 ? 0.09
1.00 ? 0.02
1.00 ? 0.08
0.86 ? 0.33
0.04 ? 0.02
1.08 ? 0.06
0.30 ? 0.12
0.07 ? 0.06
0.93 ? 0.27
0.71 ? 0.05
1.16 ? 0.64
1.01 ? 0.06
0.45 ? 0.25
0.68 ? 0.12
1.08 ? 0.12
0.6 ? 0.10
0.05 ? 0.12
0.82 ? 0.13
1.31 ? 0.61
2.65 ? 1.31
0.50 ? 0.17
0.35 ? 0.04
2.06 ? 5.7
1.34 ? 0.46
0.78 ? 2.9
4.83 ? 1.3
0.47 ? 0.28
0.73 ? 0.39
4.35 ? 11.2
0.68 ? 0.28
0.78 ? 0.26
2.31 ? 1.70
–b
0.61 ? 0.06
0.02 ? 0.05
0.03 ? 0.10
0.23 ? 0.02
0.11 ? 0.09
0.76 ? 2.9
0.13 ? 0.01
0.12 ? 0.05
0.12 ? 0.01
0.29 ? 0.05
0.09 ? 0.02
–b
23
7
9
52
3
25
54
20
14
21
4
aResidual enzyme activity after 10 min of incubation at 58°C as compared with the original value.
bA very small fitted value was obtained.
concentration of 1.74 M, but we did not observe any enzym-
atically active dimers or monomers in the gel in the presence
of urea. No monomers were detected for ∆(K2K3G4) up to
3.64 M urea, where less than 3% enzyme activity remains.
For K(2,3)E, various aggregates form at urea concentrations
?2.73 M. No enzyme activity was detectable for these aggreg-
ates. The biphasic phenomenon is therefore not a dissociation–
denaturation process.
Discussion
We have gathered sufficient data to suggest that there are at
least two structural domains in the active center of malic
enzyme, i.e. the nucleotide binding domain for NADP?and
the Mn(II)–L-malate binding domain for the substrate (Chou
et al., 1994, 1996). We have proposed that the biphasic thermal
or chemical denaturation of malic enzyme is the result of
unequal unfolding rates of the N-terminal Mn(II)–L-malate
domain and the NADP?domain (Huang et al., 1998). In our
previous studies, adenosine 2?,5?-bisphosphate was found to
be an essential fragment for NADP?binding (Lee and Chang,
1987). During the purification stage of the mutant proteins,
all mutants showed similar affinities for adenosine 2?,5?-
bisphosphate agarose affinity columns and we observed no
significant difference in KmNADPvalues for all mutants com-
pared with that of the WT (Chou et al., 1997). These results
suggest that the conformation of the mutants in their NADP?
binding site is intact. The unfolding process may involve an
intermediate state with unfolded N-terminal Mn(II)–L-malate
domain and intact NADP?domain (Huang et al., 1998).
Our present results confirm the relevance of Lys2/Lys3 in
maintaining the integrity of the N-terminal domain. In ∆(K2/
K3) and K(2,3)E mutants the N-terminal domain is much more
unstable than the native state. The corresponding destabilizing
energies for the K(2,3)E and ∆(K2/K3) mutants are ~–8.44
and –5.74 kcal/mol, respectively. However, the putative inter-
mediate state of these two mutants seems to have a longer
half-life than that of the WT, as suggested by the extremely
small k2,I→Uvalues observed in the thermal inactivation experi-
ments (Table II). Our results therefore demonstrate a possible
method for freezing an otherwise unstable folding intermediate,
allowing its detection, trapping and characterization.
Warshel and colleagues (Warshel, 1978; Warshel et al.,
375
1988) have predicted, by theoretical derivation, that ionic
interactions are the major factor involved in enzyme catalysis.
Because Lys2/Lys3 is positively charged, it is reasonable to
postulate that ion-pairing or hydrogen bonding is involved in
the functional roles. Sequence analysis of malic enzyme by
the method of Garnier et al. (1978) or Eisenberg et al. (1984)
predicts a helical structure at the N-terminal 1–4 or 1–9
residues, respectively. One of the negatively charged residues
that might interact with Lys2/Lys3 is Glu6, which could form
an (i, i ? 4) ion pair with Lys2 or an (i, i ? 3) ion pair with
Lys3. Our results support the involvement of such ion pairing
in stabilizing the malic enzyme N-terminal domain conforma-
tion and the (i, i ? 4) interaction is stronger than the (i, i ? 3)
pair. This conjecture is compatible with the literature regarding
the stability of the small peptides or proteins (Scholtz et al.,
1993). According to the N-end rule proposed by Varshavsky
(1992), replacing a lysine residue by a glutamate residue will
increase the conformational stability of a protein in vivo. Our
results seem to suggest that this rule does not apply to the
in vitro conditions.
Our results indicate that the ∆(K2/K3) mutant has a pro-
nounced effect on the active-site conformation but the single
or double mutants at Lys2/Lys3 have similar kinetic properties
to those for the WT (Chou et al., 1997). On the other hand,
the ∆(1–16) mutant has an overall catalytic efficiency (kcat/
KmMnKmNADPKmMal) that is four orders of magnitude lower
than the WT value (Chou et al., 1997). This decrease in
enzyme activity must be due to removal of essential groups
other than Lys2/Lys3. The most likely candidate is Arg9
(Huang et al., 1998).
On the basis of the above discussion, Lys2/Lys3 is probably
involved in a structural role to maintain a proper conformation
for the catalysis to proceed. Any factors that alter the correct
conformation will have some effects on the catalytic activity
of the enzyme. Removing Lys2/Lys3 alters the active site
conformation, thereby affecting the correct lining of some
other essential groups, e.g. Arg9 with Mn(II)–L-malate, that
influence the substrate and metal binding.
Because Lys2/Lys3 may not be directly involved in binding
or catalysis, mutations at these residues are good candidates
for creating a stable mutant of malic enzyme without sacrificing
too much activity. Between the two stable mutants K3A and
by guest on April 11, 2012
http://peds.oxfordjournals.org/
Downloaded from