(Received for publication, September 9, 1996, and in revised form, December 6, 1996)
From the Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565, Japan
The interaction of the subunit with the
2 subunit of tryptophan synthase is known to be
necessary for the activation of each subunit and for the catalytic
efficiency of the
2
2 complex. To
elucidate the roles of hydrogen bonds in the interaction site between
the
and
subunits for subunit association, eight mutant
subunits at five hydrogen bonding residues (N104D, N104A, N108D, N108A,
E134A, E135A, N157D, and N157A) were constructed, and the thermodynamic
parameters of association with the
subunit were obtained using a
titration calorimeter. The N104D and N104A mutations remarkably
decreased the stimulation activities, the association constants, and
the association enthalpies. Although the association constant and the
stimulation activities of E134A were reduced in the absence of salt,
the change in the association enthalpy was relatively small, and the
addition of salt could repair its defects. The substitutions at
positions 135 and 157 did not affect the stimulation activity and
decreased the Gibbs energy of association corresponding to the defect
in 1 mol of hydrogen bond. The present results suggest that the
subunit which has a mutation at position 104 cannot fold into an intact
conformation upon complex formation, resulting in reduced stimulation
activities. The hydrogen bond with Asn-104, which is a conserved
residue among 16 microorganisms, was especially important for
/
interaction and mutual activation.
Protein-protein interactions play a central role in many
physiologically important reactions. The tryptophan synthase complex is
an excellent model enzyme for exploring the molecular recognition mechanism in protein-protein interaction (1, 2). Bacterial tryptophan
synthase is an 2
2 complex. The separated
and
subunits catalyze inherent reactions termed the
and
reactions, respectively (3-5). When the
and
subunits combine
to form the
2
2 complex, the enzymatic
activity of each subunit is stimulated by 1 to 2 orders of magnitude.
It has been reported that this activation is due to conformational
changes in both subunits upon
/
subunit interaction (reviewed
in Ref. 1).
Some intermolecular electrostatic interactions (hydrogen bonds and salt
bridges) and hydrophobic (packing) interactions were observed in many
strongly bound complexes (6-8). In particular, electrostatic
interaction is thought to play a critical role in protein-protein
interaction (8-11). We focused on hydrogen bonds in the interaction
site between the and
subunits of tryptophan synthase, termed
herein "
/
subunit hydrogen bonds." A thermodynamic study of
the association of proteins having mutations at interface residues is
important to elucidate the mechanisms of the subunit interaction in the
tryptophan synthase complex. Also, it should provide useful information
for understanding the energetic contribution of the hydrogen bonds that
stabilize the protein complex. The amino acid sequences of the
and
subunits from Escherichia coli are highly similar to
those from Salmonella typhimurium (12); therefore, the
three-dimensional structures have been considered to be almost
identical (13). Residues that form hydrogen bonds in the interaction
site between the
and
subunits were identified from the x-ray
structure of the
2
2 complex from S. typhimurium (14). Table I identifies the hydrogen
bonding partner in the
subunit for the
subunit residues and the
percent conservation of each residue (12, 15-26). In this paper, eight
mutant
subunits at five positions were constructed except for
position 133, which forms a hydrogen bond at an atom in the main chain
(Table I). The roles of hydrogen bonding residues in the subunit
interaction were investigated by activity measurements and titration
calorimetry of the association. We will discuss the contribution of
electrostatic interaction and the roles of hydrogen bonding residues in
the subunit association in the complex.
|
Materials
MutantFive /
hydrogen
bonding residues of the
subunit from E. coli were
replaced by means of site-directed mutagenesis using synthetic
oligonucleotides. The sequences of the template DNAs and the
oligonucleotides are shown in Table II. Glu-134 and
Glu-135 could be directly substituted with Ala by a single code
mutagenesis (E134A and E135A). Asn residues at positions 104, 108, and
157 were substituted with Asp (N104D, N108D, and N157D) or Ala (N104A, N108A, and N157A). The eight mutant
subunits from E. coli were purified as described (27).
|
The subunits of tryptophan
synthase from E. coli can be obtained as a dimer form
(
2). The E. coli strain SP974 transformed with plasmid pUC9BG-2 (a gift from Dr. Somerville, Purdue University, West Lafayette, IN) coding the wild-type
subunit of tryptophan synthase from E. coli was grown (28), and the protein was
purified as described (29).
Methods
Protein ConcentrationsProtein concentrations of the
wild-type and mutant subunits were estimated from the absorbance at
278.5 nm, assuming
E1%1 cm = 4.4 (30, 31).
The
subunit was estimated from the absorbance at 280 nm, assuming
E1%1 cm = 6.5 (32).
The apparent association constant
(Ka) and association enthalpy
(Ha) of the
subunit with the
subunit
were calorimetrically determined at 40 °C using an OMEGA titration
calorimeter from MicroCal, Inc., as already described (29, 33). Prior
to the experiments, the sample proteins were dialyzed against a 50 mM potassium phosphate buffer, pH 7.0, containing 0.1 mM dithiothreitol, 5 mM EDTA, and 0.2 mM pyridoxal 5
-phosphate. For the experiments in the
presence of additional salt, 300 mM KCl was added to the buffer. The solution of the
subunits in a cell (0.015-0.04
mM, 1.3115 ml) was titrated with the solution of the
subunits (0.1-0.3 mM) using a 250-µl long needle
injection syringe. The
Ha and Ka values were calculated using the ORIGIN computer
program (MicroCal, Inc.). The experimental data were analyzed using a single set of identical sites model, where two
subunit binding sites per
2 subunit were identical and were independent
of each other. The association Gibbs energy
(
Ga) and entropy
(
Sa) can be calculated from experimental
Ka and
Ha values using the
following equations,
![]() |
![]() |
The forward reaction, which is the
conversion of indole-3-glycerol phosphate to indole, was measured by
changes in the absorption at 340 nm of NADH produced in the reaction
coupled with glyceraldehyde-3-phosphate dehydrogenase (34). The
reaction, which is the formation of L-tryptophan from
indole and L-serine, was assayed utilizing the difference
in absorption between indole and L-tryptophan at 290 nm
(35). 1 unit of activity in each reaction is the conversion of 0.1 µM substrate to product in 20 min at 37 °C.
The inherent activities of the subunits in the forward
reaction
were determined in the absence of the
2 subunit in 0.1 M Tris-HCl buffer, pH 8.0, at 37 °C. The stimulation
activities in the forward
and
reactions were determined in the
presence of a constant amount of the wild-type or mutant
subunits
and various amounts of the
subunit. The assay conditions for the stimulation activities were 50 mM potassium phosphate
buffer, pH 7.0, at 37 °C, the same conditions as used for the
calorimetry, in the presence and absence of 300 mM KCl. For
the stimulation activities in the
reaction, the inherent activities
of the
subunit at various concentrations were measured and
subtracted.
To
determine the effect of the deletion of the hydrogen bonds on the
stimulation activities, the five mutant subunits having Ala at the
intersubunit hydrogen bonding site (N104A, N108A, E134A, E135A, and
N157A) were investigated. We also investigated three Asp mutant
subunits (N104D, N108D, and N157A) to determine the effect of the
introduction of a charged residue at these positions. All the mutant
enzymes should lack one hydrogen bond between the
and
subunits.
These mutant
subunits retained the inherent activities in the
forward
reaction (about 0.1 units/mg), indicating that the
substitutions did not affect the monomer structure of the protein.
Fig. 1 shows the stimulation activities in the reaction of several
subunits (wild type, N104D, N104A, N108D,
N108A, and E134A) as a function of concentration of the
subunit.
The other substitutions (E135A, N157D, and N157A) had little effect on
the stimulation activities (data not shown). In the absence and
presence of 300 mM KCl, the titration curves of N104D,
N104A, and N108D
subunits were more gradual than that of the
wild-type
subunit, indicating weaker association with the
subunit. The curve of N108A
subunit was sharp and similar to that
of the wild type, suggesting that the weakened association of N108D
subunit is not due to the lack of a hydrogen bond. In the absence of
KCl, the titration curve of E134A
subunit was more gradual than
that of the wild type, but the weakened association could be repaired by the addition of 300 mM KCl. In the forward
reaction,
similar titration curves were found (data not shown). These results
indicate that the hydrogen bonds, except for position 104, are not
required for the mutual activation.
Titration Calorimetry of the Interaction of Mutant
The mutual activation between
the and
subunits is induced by the formation of the
2
2 complex. To determine the effect of
deletion of a hydrogen bond on the
/
subunit interaction, the
thermodynamic parameters of association of mutant
subunits with the
subunit were examined using ITC. Fig. 2 shows the
calorimetric titration curves of the
subunit with the wild-type and
mutant
subunits at 40 °C, pH 7.0. The titration curves of four
mutant
subunits (N104D, N104A, N108D, and E134A) had a more gradual transition than that of the wild-type
subunit, indicating the decreases in the association constants (Ka) of these mutant
subunits. The thermodynamic parameters of association were
calculated by the computer program ORIGIN (29, 33), assuming that the
two
subunit binding sites of the
2 subunit were
identical and independent of each other (i.e. single set of
identical sites model), and are shown in Table III. The
independent or interactive two sites model was also tried, but these
models failed to fit the titration curves. All titration curves of the
mutant enzymes demonstrate a similar stoichiometry of about 0.7
chain/
chain. The departure from the expected 1:1 stoichiometry
probably results from the presence of a fraction of inactive
2 subunit, which cannot bind the
subunit (29). The
thermodynamic parameters of association are independent of the binding
number because they are estimated by the molar concentration of the
subunit.
|
The negative Ga values of almost all mutant
subunits were decreased compared with that of the wild-type
subunit. The changes in the
Ga varied with
the substituted positions. The change in the
Ga of the N108A
subunit was the smallest
among the mutant enzymes. However, there was a subtle change in the
Ha, and it was compensated by a change in the
T
Sa. The
Ga values for E135A, N157D, and N157A
subunits were affected by 2.0-3.9 kJ mol
1, and these
defects resulted from an unfavorable increase in the
T
Sa values. Remarkable changes
were observed for N104D, N104A, N108D, and E134A
subunits. Their
Ka values were decreased by 1 order of magnitude,
and the decreases in the negative
Ga values
were 7.1-9.1 kJ mol
1 compared with that of the wild-type
subunit. The negative
Ha values of N104D,
N104A, and N108D
subunits were unfavorably decreased compared with
that of the wild-type
subunit. The decrease in
Ha of the E134A
subunit was relatively
smaller than that of the N104D, N104A, or N108D
subunits although
the Ka value of the E134A
subunit was similar to
those of the others.
Electrostatic
interactions have been reported to contribute to many intermolecular
associations based on x-ray crystallographic analysis (8, 9),
thermodynamic analysis (10), and computer simulation (11). To
investigate the electrostatic contribution to the interaction between
the and
subunits, the effect of salt was examined by ITC. Fig.
3a shows the titration curve for the
wild-type
subunit in the absence and presence of 300 mM KCl. The thermodynamic parameters of association are shown in Table
III. The Ka values showed that the complex formation of the wild-type tryptophan synthase was increased by the addition of
KCl. The negative
Ha of the wild-type
subunit in the presence of KCl was lower than that in the absence of
additional salt (Table III). However, the favorable increase in the
Sa overcame the unfavorable
Ha and resulted in a favorable decrease in
Ga.
The effects of the substitutions at the interface on the electrostatic
properties of association were then examined by ITC in the presence of
300 mM KCl for the mutant subunits at positions 104 (N104D and N104A) and 134 (E134A), which have decreased
Ka values in the absence of salt (Fig. 3 and Table
III). The negative
Ha values of N104D and
N104A
subunits were decreased by the addition of KCl in a way
similar to that of the wild type, whereas the Ka
values were slightly decreased. In the case of E134A
subunit, the
Ka value was greatly increased by the addition of
KCl. The negative
Ha value of E134A
subunit in the presence of KCl was larger than that of the wild type, in contrast with that in the absence of KCl.
Recently, it has been reported that monovalent cations
change the conformation of the tryptophan synthase complex and activate it (36-41). The preferential binding of a cation (such as
Na+ or K+) to a specific site in the subunit has high affinity, and the site is saturated at 20 mM cation concentration (38). In this study, because we
treated the sample in 50 mM potassium phosphate buffer, the
concentration of K+ might be enough to saturate the cation
binding sites of the tryptophan synthase complex even in the absence of
additional KCl. Therefore, the increase in the Ka
value of the wild type on the addition of salt observed here (Table
III) should not result from a conformational change due to the effect
of the cation. Alternatively, the present results suggest the presence
of electrostatic repulsion between the
and
subunits. The
addition of salt might diminish this repulsion and cause an increase in
the affinity. Similar stabilizing effects of salts on the oligomer
structure have been reported for some proteins (42-47).
We examined whether there is an electrostatic repulsion between the and
subunits due to the complex structure (14). The charges on the
subunit interface were calculated using the Insight II-Delphi program
(48), and it was found that both subunits were relatively negatively
charged at the subunit interface. A qualitative estimation using the
Insight II-Docking program showed that the electrostatic effect acts
unfavorably on the association between the
and
subunits,
whereas the van der Waals interaction is favorable.
In the cases of the mutant subunits at position 104, the
presence of salt to shield the electrostatic repulsion did not improve
the affinity for the
subunit (Table III). In contrast, the weakened
affinity of E134A
subunit was repaired by the presence of salt and
was close to that of the wild type (Table III).
Shirley et al. (49) have reported that
the contribution of a hydrogen bond to the Gibbs energy of unfolding is
5.4 ± 2.5 kJ mol1 using mutant proteins. Fersht
(50) has also evaluated it to be 2.1-7.5 kJ mol
1. In the
present study, the average
Ga of eight
mutant proteins was 5.2 kJ mol
1 (Table III), and it fell
within the ranges reported previously. However, the other thermodynamic
parameters,
Ha and
Sa, varied with the substituted positions
(Table III). This indicates that hydrogen bonds play different roles in
the association depending on their positions (see below).
The effects on the
thermodynamic parameters of association of the mutant subunits at
positions 135 and 157 (E135A, N157D, and N157A) could be considered to
be due to the deletion of only 1 mol of hydrogen bond (Table III). In
each case, the reasonable decrease in the negative
Ga value corresponds to the deletion of one
hydrogen bond as reported previously (49, 50). Although the changes in
the
Ha and
T
Sa of these mutant
subunits
were mostly within experimental error, there is a tendency for the unfavorable
Ga to be accompanied by an
unfavorable change in the positive
T
Sa and partially compensated by
a favorable change in the negative
Ha. These
results agree with the conclusion of Privalov and Makhataze (51) that
the hydrogen bonds between the polar groups are stabilized
entropically. Connelly et al. (52) have studied the effect
of the substitution of a hydrogen bonding residue on the protein-ligand
binding using ITC and have found similar results. Because the changes
observed for the substitutions at positions 135 and 157 correspond to
the deletion of one hydrogen bond as described above, the effects due
to the mutations must be localized in the substituted region.
The substitutions at positions 135 and 157 (E135A, N157D, and N157A)
had no effect on the stimulation activity, and the changes in the
thermodynamic parameters are in the ranges of the contribution from one
hydrogen bond as described above (Table III), suggesting that these
substitutions do not result in a defect in the intact conformation of
the complex. These residues are located on the edge of the interface
with the subunit, far from the intersubunit tunnel, and the
partner residues of the
subunit (Met-15 and Tyr-181 of the
subunit) are not highly conserved. Therefore, it can be concluded that
these hydrogen bonds are not very important for the mutual
activation.
Due to the substitutions at
position 104 (N104D and N104A), the stimulation activities were
decreased (Fig. 1), and the decreased Ka values of
these mutant subunits (Table III) might be related to the reduced
stimulation activities. We have reported that the complex formation
couples with the folding and the rearrangement events in the
or/and
subunits, and the folding might occur not only at the subunit
contact surface but also at other parts in the molecules (53). The
unfavorable decreased
Ha values of the mutant
enzymes at position 104, compared with that of the wild type (Table
III), might suggest that the mutant proteins cannot fold into an intact
conformation upon complex formation. Therefore, the stimulation
activities of their resulting complexes were lower than that of the
wild-type complex. Furthermore, the affinities of N104D and N104A with
the
subunit could not be improved by the addition of salt (to seal
electorostatic repulsion) (Table III), suggesting that these
mutant
subunits could not associate with the
subunit to fold
the similar conformation to the
2
2 complex of the wild-type protein.
Position 104 is apparently located far from the intersubunit tunnel.
However, in the 2
2 complex form, the Asn
residue at position 104 of the
subunit contacts the residue at
position 292 of the
subunit in a sharp turn in the
trypsin-sensitive "hinge" region (Table I). This hinge region is
included in the long stretch from 260-310 (with a loop conformation),
which forms one side of the intersubunit tunnel of the
subunit
(14). The substitution (28, 54) or proteolysis (55-57) in the
"hinge" region of the
subunit reduces the stimulation activity,
the substrate affinity, and the association between the
and
subunits. Thus, this region is believed to play a critical role in the
conformational change upon the subunit association (1). The present
results revealed the special role of 104; the hydrogen bonding at
position 104 is necessary to maintain the hinge of the
subunit in
an intact form, and a substitution at this position causes a
significant defect in the stimulation activity. Regarding other
mutations at position 104 of the
subunit (58), it has also
been reported that the N104S mutant complex decreases the activity in
the
and
reactions. Asn-104 and its hydrogen bonding partner
Gly-292 of the
subunit are completely conserved residues among the
tryptophan synthase from 16 different organisms (Table I). It can be
concluded that the hydrogen bond at Asn-104 of the
subunit with
Gly-292 of the
subunit is critical to induce a perfect complement
association.
Substitutions at positions 108 and 134 might also involve conformational changes because the changes in the thermodynamic parameters could not be interpreted based on the deletion of only one hydrogen bond as described above.
At position 108, Asp mutation caused decreases in the
Ka value by 1 order of magnitude and in a negative
Ha (Table III). Asn-108 of the
subunit
contacts the residue at position 290, which is Ala in S. typhimurium and is replaced by Glu in E. coli in the
hinge region of the
subunit. The extreme substitution with a
bulkier and more polar residue at position 108 (N108D) might prevent
intact conformation of the complex, but hydrogen bonding at position
108 is not required for complex formation as shown in results of N108A.
Another mutation at this position (N108S) also does not affect the
activity of the complex (58).
The Ga value of E134A
subunit in the
absence of KCl was remarkably affected by the substitution, although
the change in the
Ha was relatively small
(Table III). In the case of E134A
subunit, the conformational
change due to the complex formation might correspond to that of the
wild type, despite its low Ka value. This is also
supported by the results in which the
Ga was
improved by the addition of the KCl, and the negative
HaM-W and positive
T
SaM-W values could
be explained by the deletion of one hydrogen bond as described for the
substitutions at positions 135 and 157. Glu-134 is 40% exposed to the
solvent even in the complex form (Table I). The percentage of
conservation of Glu-134 is 75%, but its hydrogen bonding partner
Gln-19 of the
subunit is not highly conserved (Table I). Present
results indicate that this residue is not required to induce complement
association although the substitution of Glu-134 affected the affinity
for the
subunit.
The hydrogen bonding residues located in the
/
subunit interface play various roles in the association and
mutual activation, depending on their positions. The deletion of the
hydrogen bond at Asn-104 affected the thermodynamic parameters of
association with the
subunit and remarkably reduced the stimulation
activity in the
reaction. It can be concluded that the hydrogen
bond at Asn-104 of the
subunit is especially important for intact association with the
subunit and mutual activation of the
complex.
DNA sequence analysis was carried out by an automated DNA sequencer (Applied Biosystems, Inc.) at the Research Center for Protein Engineering, Institute for Protein Research, Osaka University.