- Email: [email protected]
Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers
Accepted Manuscript Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers Pablo Villalobos, Francisco Soto, Mauricio Baez, Jorge Babul PII:
S0300-9084(16)30166-3
DOI:
10.1016/j.biochi.2016.08.013
Reference:
BIOCHI 5047
To appear in:
Biochimie
Received Date: 26 December 2015 Accepted Date: 29 August 2016
Please cite this article as: P. Villalobos, F. Soto, M. Baez, J. Babul, Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers, Biochimie (2016), doi: 10.1016/ j.biochi.2016.08.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
1 2
Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers
3
Pablo Villalobos1, Francisco Soto1, Mauricio Baez2*, Jorge Babul1*.
4 1
6 7
2
Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Santiago, Chile.
RI PT
5
Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias Químicas y Farmacéuticas, Universidad de Chile, Santiago, Chile.
8
10
* Corresponding authors
SC
9
E-mail: [email protected] (MB); [email protected] (JB)
M AN U
11
Abstract
13
We have proposed an allosteric ATP inhibition mechanism of Pfk-2 determining the structure
14
of different forms of the enzyme together with a kinetic enzyme analysis. Here we
15
complement the mechanism by using hybrid oligomers of the homodimeric enzyme to get
16
insights about the allosteric communication pathways between the same sites or different ones
17
located in different subunits. Kinetic analysis of the hybrid enzymes indicate that homotropic
18
interactions between allosteric sites for ATP or between substrate sites for fructose-6-P have a
19
minor effect on the enzymatic inhibition induced by ATP. In fact, the sigmoid response for
20
fructose-6-P observed at elevated ATP concentrations can be eliminated even though the
21
enzymatic inhibition is still operative. Nevertheless, leverage coupling analysis supports
22
heterotropic interactions between the allosteric ATP and fructose-6-P binding occurring
23
between and within each subunit.
AC C
EP
TE D
12
1
ACCEPTED MANUSCRIPT
24
1. Introduction The allosteric regulation induced by MgATP is a common property of E. coli
26
phosphofructokinases Pfk-1 and Pfk-2 [1, 2]. Nevertheless, information about their regulation
27
cannot be transferred between them since both structures belong to different folds.
RI PT
25
In order to understand the structural basis of the allosteric mechanism of Pfk-2, we have
29
determined the X-ray structure of different forms of the enzyme, including complexes with
30
fructose-6-P (PDB ID: 3N1C [2]), or MgATP (PDB ID: 3CQD, [3]) which, with a thorough
31
kinetic analysis, led us to propose a mechanism for the allosteric inhibition [2]. In this work we
32
complement the mechanism by using hybrid oligomers to get insights about the allosteric
33
communication pathways between the same sites or different ones located in different subunits.
M AN U
SC
28
The rationale of our experimental approach is to eliminate or perturb the binding capacity
35
in one subunit while the partner subunit remains intact [4]. For example, homotropic interactions
36
for fructose-6-P binding sites in B. stearothermophilus phosphofructokinase and E. coli Pfk-1
37
have been studied by hybrid homotetramers, where the binding site for fructose-6-P was
38
eliminated in one or several subunits [5, 6]. A similar approximation was used to detect long
39
range interactions that accounts for substrate inhibition induced by ATP in Pfk-1 [7]. In that case,
40
a hybrid homotetramer, containing a single intact ATP binding site, failed to show substrate
41
inhibition by ATP, supporting the fact that communication arises between the subunits and not
42
within the subunits.
43
Structural organization of Pfk-2. One site for fructose-6-P and two binding sites for MgATP in
44
adjacent locations are found in each subunit (Fig. 1A). Like its non-allosteric homologues, the
45
functional structure required for catalysis in Pfk-2 is the homodimeric state. Key residues to
46
provide electrostatic and aromatic stacking interactions with the allosteric ATP are established by
AC C
EP
TE D
34
2
ACCEPTED MANUSCRIPT
the side chains of K27 and Y23 respectively. Both side chains are contributed from the partner
48
subunit due to the intertwined nature of the homodimeric interphase (Fig. 1B).
49
The saturation curves for fructose-6-P at different MgATP concentrations indicate that inhibition
50
occurs due to a decrease in the apparent affinity for fructose-6-P, together with a transition from
51
hyperbolic to sigmoidal behavior and partial reduction of the catalytic constant [2]. These
52
observations could be explained in part by the perturbation of a network of interactions between
53
basic residues and the phosphate moiety of fructose-6-P. In this way, the binding of MgATP
54
sequesters the side chain of K27 and provokes a reorientation of the side chain of R90 out the
55
active site (Figure 1B). Elimination of the formal charge of R90 by site-directed mutagenesis
56
increases the KMapp value for fructose-6-P in three orders of magnitude [8] supporting the idea
57
that the side chain reorientation of R90 must play a role in the MgATP inhibition. Importantly,
58
according to the crystal structures, these conformational changes occur in both subunits of the
59
Pfk-2 dimer [2]. However, the apparent diminution of fructose-6-P affinity is also accompanied
60
by a change of the sugar saturation curve from hyperbolic to a sigmoid shape, suggesting a
61
communication route between subunits [2].
TE D
M AN U
SC
RI PT
47
Hybrids enzymes show that homotropic interactions between fructose-6-P sites and
63
allosteric ATP sites are not fundamental for allosteric inhibition neither for the sigmoid response
64
observed for fructose-6-P at elevated concentration of the allosteric inhibitor. Conversely, the
65
sigmoid response for fructose-6-P seems more related with the catalytic turnover rather than
66
substrate binding. Further analysis performed in silico, shows that several heterotropic
67
interactions between fructose-6-P and allosteric sites for ATP are operating within and between
68
the Pfk-2 subunits. These interactions are created between key residues for ATP and fructose-6-P
69
binding.
AC C
EP
62
70
3
ACCEPTED MANUSCRIPT
71 72 73
RI PT
74 75 76
SC
77
81 82 83
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
EP
87
Figure 1. Dimeric structure of Pfk-2 showing the location of fructose-6-P (F6P), allosteric ATP (ATPa) and substrate ATP (ATPc). A. The image shows the Pfk-2-fructose-6-P complex (PDB ID 3N1C) and the MgATP molecules binding at the active and allosteric sites; the MgATP coordinates were obtained from the Pfk-2-ATP crystallographic structure (PDB ID 3CQD) after aligning both structures. The Pfk-2 monomers are colored blue and white, ATP and F6P are represented with licorice sticks and magnesium ions are shown as green spheres. B. Structural changes induced by F6P or MgATP. Location of the Y23 and K27 side chains of the Pfk-2MgATP and Pfk-2-fructose-6-P complexes are shown as yellow and green sticks respectively. Note that in each allosteric site the key residues Y23 and K27 are provided by the β3 and β2 strands of the partner subunit. However, to be consistent with the structural organization of Pfk-2, each allosteric site is considered to belong to only one subunit, irrespective of whether the two sidechains are supplied by the partner subunit. Key interactions: Y23 with the adenine ring of the allosteric ATP and K27 either with ATP or fructose-6-P. The R90 sidechain in the presence of fructose-6-P is seen pointing towards the active site and outwards in presence of MgATP. The catalytic base of Pfk-2, D256, does not show structural differences on the superposition.
AC C
86
TE D
84 85
79
M AN U
80
78
4
ACCEPTED MANUSCRIPT
103 104
2. Materials and Methods 2.1. Enzyme purification and site direct mutagenesis. Site directed mutagenesis was carried out
106
with the Gene TailorTMsystem using pET21 plasmid containing the Pfk-2 gene as template.
107
Protein purification was carried out as is described by [9].
108
2.2. Isolation and characterization of hybrids. To promote subunit exchange, mixtures of Pfk-2
109
mutants were incubated at 20 °C in 25 mM Tris buffer pH 8.2, 5 mM MgCl2, 5 mM DTT for 3 h
110
(except for hybrid D, 20 h) and hybrid dimers separation was achieved by ion exchange
111
chromatography (Mono Q column 5/10 (Pharmacia) using a Water’s system equipped with a UV
112
detector. A charged tag (R206E/K207D) was introduced in the surface of Pfk-2 to isolate the
113
hybrids by anion exchange. The presence of the tag did not modify the kinetic properties with
114
respect to the wild type enzyme (Fig. S1 and Table S1). Each hybrid was constructed on the basis
115
of a dimer composed by an active and inactive subunit (Table 1). Hybrid A was obtained from a
116
mixture of the inactive mutant D256N and the tagged mutant R206E/K207D. Hybrid B was
117
created by mixing the tagged mutant harboring the Y23A mutation with the inactive D256N
118
mutant. Hybrid C was obtained by mixing the double mutant D256N/Y23A with the tagged
119
mutant. Hybrid D was obtained mixing the D256N/R90Q mutant and the active tagged mutant.
120
2.3. Kinetic and binding properties of the hybrid enzymes. The MgATP-induced inhibition was
121
evaluated as a function of the fructose-6-P or MgATP concentration using two fixed co-substrate
122
concentrations [2]. Phosphofructokinase activity was measured by coupling the fructose-1,6bisP
123
production to the oxidation of NADH followed at 340 nm [1]. All kinetic assays were performed
124
at room temperature (25 °C) by employing 96-well plates and read in a microplate reader (Biotek,
125
Synergy). The binding constants for fructose-6-P or MgATP were evaluated using intrinsic
AC C
EP
TE D
M AN U
SC
RI PT
105
5
ACCEPTED MANUSCRIPT
fluorescence measurements in a Shimadzu PC-5031 spectrofluorimeter as is described by Guixé
127
et al. [10]
128
2.4. Treatment of kinetic and affinity data. Fluorescence binding assays were performed to
129
determine the fructose-6-P and MgATP affinity for the wild type enzyme and the R90Q and
130
Y23A mutants [8] [10]. Coherent with the presence of two MgATP molecules per Pfk-2
131
monomer [3], activity as a function of the MgATP concentration was fitted to the following
132
substrate inhibition model:
[E ]t × k cat × [S ] [S ] K + [S ]× 1 + M
K i
M AN U
v=
133
SC
RI PT
126
where kcat corresponds to the catalytic constant obtained from the initial steady-state velocity at
135
the described conditions, [S] equals the concentration of MgATP, KMapp is the apparent affinity of
136
MgATP and Ki is the apparent binding constant for MgATP acting as an inhibitory substrate. On
137
the other hand, data from activity assays as a function of the fructose-6-P concentration were
138
fitted to a Hill equation. All data were plotted and analyzed by nonlinear regressions using
139
SigmaPlot 11.0.
AC C
EP
TE D
134
6
ACCEPTED MANUSCRIPT
3. Results and Discussion
141
3.1. Mutants and hybrids design. In order to determine the independent behavior of sites present
142
in different subunits of an oligomer, or between sites present in the same subunit, hybrid
143
oligomers were created to break the expected symmetry by decreasing or abolishing the affinity
144
of allosteric or substrate sites in individual subunits [4]. Thus, there are two types of Pfk-2 hybrid
145
dimers, one to study heterotropic interactions between the allosteric MgATP and sugar sites
146
(hybrids B and C in table 1), and another to study homotropic interaction between the fructose-6-
147
P binding sites present in the dimer (hybrid D in table 1). For this purpose it is necessary to
148
eliminate or diminish the affinity of one allosteric site and/or the fructose-6-P binding site.
M AN U
SC
RI PT
140
Table 1 summarizes the kinetic properties of the homodimers required to create the
150
hybrids studied in this work. To create hybrids B and C it is necessary to avoid the stacking
151
interaction between the adenine moiety of the allosteric ATP and Y23 (Fig. 1B), so the residue
152
was replaced by alanine. This mutation diminishes the affinity of the allosteric ATP in one order
153
of magnitude with respect to the wild type enzyme, as indicated by the binding curves obtained
154
from intrinsic fluorescence measurements (Fig. S2A and Table S2). The diminished affinity
155
towards MgATP was attributed specifically to the alteration of the allosteric site, since the Y23A
156
mutant shows an attenuated enzymatic inhibition (a large Ki value together with a minor
157
increment of the KMapp for fructose-6-P), without changes in the catalytic parameters (Fig. S3 and
158
Table S3A and B). On the other hand, to create hybrid D it is necessary to disrupt the sugar
159
phosphate binding site, so R90 was replaced by glutamine (R90Q). This mutation increases both
160
the apparent KMapp [8] and the Kd for fructose-6-P by three orders of magnitude (Fig. S2B and
161
Table S2). The structural effects seem to be confined to the sugar site, since the allosteric ATP
162
binding was not modified by the mutation [8]. In other words, we utilized the Y23A mutant that
AC C
EP
TE D
149
7
ACCEPTED MANUSCRIPT
has a reduced affinity for the allosteric MgATP and the R90Q mutant with a decreased affinity
164
for fructose-6-P. Moreover, an additional mutation was added to each hybrid dimer (D256N) to
165
inactivate one active site and hence facilitate the kinetic analysis. The D256N mutant decreases
166
the catalytic activity in two orders of magnitude without a substantive alteration in the inhibitory
167
parameters [4].
168
SC
Table 1. Mutants used and properties of the heterodimers
RI PT
163
Mutant 2 Tag mutant: Active.
Heterodimer Hybrid A: Inactive subunit with intact allosteric ATP and fructose-6-P sites and active partner subunit.
a
D256N: Inactive.
Y23A/Tag Mutant: Active with decreased allosteric ATP binding.
Hybrid B: Intact allosteric and active sites in the same subunit.
Hybrid B
D256N/Y23A: Inactive, with decreased allosteric ATP binding.
Tag Mutant: Active.
TE D
Hybrid C:
Diagram Hybrid A
Hybrid C
EP
Intact allosteric and active sites in different subunits.
Tag Mutant: Active.
Hybrid D: Hybrid D To test inter subunit communication between fructose-6-P sites. An inactive subunit with decreased fructose-6-P binding and active partner subunit. The columns referred as mutant 1 and 2 represent the homodimers used to create the hybrids enzymes described in columns 3 and 4. The subunits of the dimer are highlighted in different colors following the color of the homodimeric mutants used to create them. Each active site is composed by a large domain (circle), a fructose-6-P binding site (triangle) and the ATPa binding site (rhombus). The interactions of Y23 with the allosteric ATP are shown in blue since formally Y23A is donated by the partner subunit. “X” symbolizes a mutation introduced in the respective site. The blue subunit contains the negative tag mutations (R206E/K207D). Hybrid A, one inactive subunit (mutant D256N) with intact allosteric and fructose-6-P binding sites. Hybrid B, one inactive subunit (mutant D256N) with the allosteric site mutated (mutant Y23A). Hybrid C, one inactive subunit (D256N) with its adjacent allosteric site mutated (mutant Y23A). Hybrid D, one inactive subunit (mutant D256N) with its fructose-6-P binding site mutated (mutant R90Q).
AC C
D256N/R90Q: Inactive, with decreased fructose6-P binding.
M AN U
Mutant 1 D256N: Inactive.
8
ACCEPTED MANUSCRIPT
169
3.2. Formation and isolation of Pfk-2 hybrid dimers. The formation of hybrid dimers or
171
heterodimers occurs due to the dynamic nature of the association and dissociation events of the
172
oligomers present in solution. Fig. 2A compares the anion exchange chromatograms obtained for
173
the inactive mutant (mutant D256N; upper panel), Pfk-2 containing a “tag of charge” (middle
174
panel), and the result of incubation of an equimolar mixture of both enzymes (bottom panel).
175
Coherently, the dimer containing two negatively tagged subunits elutes at higher NaCl
176
concentrations with respect to the D256N mutant (without the “tag of charge”), and the mixture
177
of both Pfk-2 forms results in a new peak between the tagged and untagged homodimers. The
178
new peak was collected and kinetically characterized, showing a kcat value of 25 s-1,
179
approximately half of the value observed for the wild type enzyme (56 s-1), confirming the
180
presence of a hybrid dimer created by a fully active and an inactive subunit (Fig. 2A, peak III). In
181
addition, we reasoned that if the hybrid dimer equilibrium exchange is allowed to reestablish, the
182
appearance of tagged and inactive parental homodimers should be observed. Therefore, to
183
reestablish the equilibrium exchange, the purified new peak was equilibrated in the exchange
184
buffer (using a desalting column) and reinjected into the anion exchange column. The sample
185
shows the presence of two additional peaks corresponding to the wild type and the inactive
186
mutants (data not shown). We observed that desalting of the new peak was a condition to recover
187
the parental homodimers, supporting the fact that elevated salt concentrations slows the dimer
188
equilibrium exchange. Addition of ligands like fructose-6-P and MgATP also slows the dimer
189
equilibrium exchange rate.
AC C
EP
TE D
M AN U
SC
RI PT
170
190
9
EP
Figure 2. Separation of hybrid enzymes using anion exchange chromatography. A. The figure shows the elution profile of the inactive mutant (D256N, upper panel), Pfk-2 modified with a negative charged tag (R206E/K207D, middle panel) and a mixture of both enzymes previously incubated for 3 h at 20 °C (bottom panel). The specific activity of the new peak (25 s1 ) was about half of that of the tagged enzyme (56 s-1). Anion exchange chromatography was carried out using a MonoQ HR 5/5 column and the proteins were eluted with a linear KCl gradient from 0.2 to 0.75 M in presence of 25 mM Tris pH 8.2, 1 mM DTT and 5 mM MgCl2. The presence of KCl or ligands (either MgATP or fructose-6-P) stops the exchange reaction avoiding hybrid re-equilibration. B. Kinetics of subunit exchange between the inactive mutant and the tagged Pfk-2. Samples of a mixture of tagged wild type and D256N mutant were withdrawn and injected into MonoQ HR column for each time. Symbols represent the temporal variation of absorbance peak area measured for peaks I (○), II (Δ) and III (●). The continuous lines represent an exponential function used to describe the temporal appearance of peak III or the disappearance of peaks I and II. The kinetic constant for the appearance of the hybrid was 0.03 min-1.
AC C
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
208
10
ACCEPTED MANUSCRIPT
209
Table 2. Kinetic parameters for the MgATP inhibition of the hybrid enzymes The constants were obtained from the kinetics in Fig. S4.
1 mM fructose-6-P
Enzyme
kcat s-1
KMapp (ATP), µM
Ki, mM
kcat (s-1)
Hybrid A
18±2
12±6
0.57±0.09
20.2±0.5
21.8±4.5
0.87±0.07
11.8±1.7
Hybrid B 10.8±1.0 15±2
11±3
1.1±0.1
20±1
Hybrid D
6.6±0.2
6±1
1.1±0.2
7.6±0.3
8.4±0.4
14±2
13±2
18±3
13±2
14 ±2
4.1±0.7
ND
M AN U
210
KMapp(ATP), µM Ki, mM
SC
Hybrid C
RI PT
0.05 mM fructose-6-P
In order to observe the kinetics of subunit exchange, equimolar mixtures of tagged wild
212
type and D256N mutant were injected into the column after several incubation times (between 10
213
and 1,080 min for hybrid A and between 3 and 27 h for hybrid D), and the concentrations of each
214
species was quantified using their chromatographic areas. As indicated in Fig. 2B, the relative
215
abundance of the hybrid dimer was 45 %, suggesting an unbiased subunit exchange. Fitting of the
216
exponential function to the time course of hybrid formation gives a rate constant between 2 and 6
217
x 10-4 s-1. This value is similar to the kinetic constant obtained for the slower process of the
218
unfolding of Pfk-2 determined from the kinetics of chemical denaturation (1.5 x 10-4 s-1) [11].
219
The agreement between both constants suggests that the dimer unfolding and dissociation is the
220
limiting step for dimer exchange [12].
221
3.3. Kinetic analyses of Pfk-2 hybrid enzymes. First we ask if each active site works
222
independently with respect to catalysis in the Pfk-2 dimer. For this purpose, a hybrid dimer with
223
an inactive subunit harboring the D256N mutation was created, denominated as hybrid A (Table
224
1, first row). Additionality, like each hybrid studied in this work, it contains a “charged tag”
225
(R206E/K207D) on the protein surface of one subunit to isolate the heterodimer hybrids by anion
AC C
EP
TE D
211
11
ACCEPTED MANUSCRIPT
exchange chromatography [13]. Fig. 3 shows the typical kinetic analysis performed to
227
characterize the allosteric inhibition of wild type Pfk-2, but performed with the hybrid A. The
228
allosteric effect is evidenced by a pronounced inhibition when the saturation curves are measured
229
as a function of the MgATP concentration (Fig. 3A). Curve fitting to models of substrate
230
inhibition retrieve the apparent substrate inhibition constant (Ki), which represents the apparent
231
affinity of the allosteric ATP for the E-ATP complex (see Material and Methods). Thus, this
232
apparent affinity is high (Ki = 0.57 mM) at low fructose-6-P concentrations and low (Ki = 14
233
mM) at elevated concentrations of the sugar-phosphate (Table 2).
SC
RI PT
226
M AN U
234
The allosteric inhibition can also be evaluated as an increase of the apparent KMapp for
236
fructose-6-P at elevated MgATP concentrations. In this case, the diminished apparent affinity for
237
fructose-6-P is accompanied by a change towards a sigmoid response (Fig. 3B). In the case of
238
hybrid A, the kinetic parameters for both substrates, and the regulatory kinetic properties were
239
very similar to the wild type values, except for the kcat value which is half of that of the wild type
240
enzyme (56 s-1, [2]). These observations support the idea that both subunits work independently,
241
at least with respect to the catalytic turnover of each active site.
244 245 246
EP
243
AC C
242
TE D
235
247 248
12
ACCEPTED MANUSCRIPT
Table 3. Kinetic parameters for fructose-6-P saturations of hybrid enzymes.
SC
RI PT
The constants were obtained from the kinetics shown in Fig. S5
Figure 3. Kinetic characterization of hybrid A. A. Initial velocities for hybrid A as function of the MgATP concentration measured at low (0.05 mM ○) and high (1 mM ●) fructose-6-P concentrations. The lines represent the best fit to an uncompetitive inhibition model and the resulting kinetic constants are shown in Table 2. B. Kinetic assays for hybrid A were performed with fructose-6-P as variable substrate at two fixed MgATP concentrations: 0.1 mM (∆) and 5 mM (▲). The continuous line represents a fit of each curve to the Hill equation; resulting parameters are shown in Table 3.
258
In order to disturb one allosteric site for ATP in the Pfk-2 dimer we designed hybrids B
259
and C. As is indicated in the second and third row of the table 1, both hybrids B and C carry out
260
the Y23A mutation in one subunit. The difference between them is the localization of the inactive
261
subunit with respect to the localization of the Y23A mutation. Nevertheless, taken into account
262
that the D256N mutation does not perturb the affinities for fructose-6-P or allosteric ATP binding
263
[2], both hybrids have the same symmetry with respect to the binding sites. Therefore, we
264
expected similar regulatory kinetics for both hybrids B and C.
TE D
EP
AC C
265
M AN U
249 250 251 252 253 254 255 256 257
266
13
ACCEPTED MANUSCRIPT
0.1 mM MgATP nH
kcat, s-1
KMapp(F6P),µM
nH
Hybrid A 32.8±0.3
35±2
1.36±0.05
24.4±0.6
258±27
1.9±0.2
Hybrid B
16.2
60.9
1.1
12.09±0.03
176.9±3.9
1.7±0.2
Hybrid C
27±1
42±7
1.0±0.3
23.0±0.4
218±15
1.1±0.2
Hybrid D
9.0±0.6
16±4
0.71±0.08
11.6±0.1
176±4
1.41±0.09
.
267
RI PT
KMapp(F6P), µM
Enzyme
kcat s-1
5 mM MgATP
The saturation curves for MgATP determined for hybrids B and C displayed inhibition by
269
substrate (Figs. S4A and S4B). In both cases, the co-substrate (fructose-6-P) increases the
270
apparent inhibition constant (Ki) for MgATP and both enzymes have essentially identical Ki
271
values either at low or high fructose-6-P concentrations (Table 2). Application of the substrate
272
inhibition model gives an Ki for MgATP of 0.87 ± 0.07 mM for the hybrid B and 1.1 ± 0.1 mM
273
for the hybrid C (Table 2), which is close but slightly higher than the Ki measured for hybrid A
274
(0.57 ± 0.09 mM, Table 2). Taken the value of Ki as an indication of the apparent allosteric
275
binding of ATP, affinities of hybrids B and C are similar between them but slightly diminished
276
with respect to hybrid A. In opposition, if both allosteric sites are required for full inhibition, then
277
we expected to observe a Ki = 7 mM, similar to that obtained for the homodimeric Y23A mutant
278
determined at equivalent fructose-6-P concentrations (Table S3A).
M AN U
TE D
EP
Elevated concentrations of MgATP (5 mM) increase the KMapp for fructose-6-P to 164 µM
AC C
279
SC
268
280
for hybrid B and up to 218 ± 15 µM in the case of the hybrid C (Table 3 and Fig.S5A and S5B).
281
These results indicate that the enzymatic inhibition is still operating in both hybrids, since the
282
increments of ATP concentration increase the apparent KM for fructose-6-P reaching a similar
283
value as the one calculated for hybrid A (KMapp = 258 µM, Table 3 and Figure 3B). Otherwise,
284
the value of KMapp for fructose-6-P should be similar to the Y23A mutant (66 µM, Table S3B).
14
ACCEPTED MANUSCRIPT
However, there is an important difference between hybrids B and C. In the case of hybrid C, the
286
saturation curves for fructose-6-P keep their hyperbolic shape (Fig. S5B), i.e., the hill number
287
remains close to 1 (Table 3) when the KMapp for fructose-6-P is increased by the presence of ATP.
288
Differently, the saturation curves for fructose-6-P for hybrid B display a sigmoid response like
289
hybrid A (Table 3 and Fig. S5A). Therefore, the sigmoid saturation for fructose-6-P observed at
290
elevated concentrations of ATP is uncoupled from the allosteric inhibition in hybrid C. This
291
effect, seems more related with the catalytic turnover than to substrate binding since the D256N
292
mutant, used to inactivate one subunit, has similar affinities for fructose-6-P and allosteric
293
binding [2].
M AN U
SC
RI PT
285
294
Based on these results we postulate that direct homotropic interactions between the
295
allosteric sites for MgATP seem not to be fundamental for allosteric inhibition in Pfk-2, since
296
hybrids B and C, with only one functional allosteric site, still present MgATP inhibition. To test the homotropic interactions between the fructose-6-P binding sites of both
298
subunits, hybrid D was characterized kinetically (Table 1, fourth row). This hybrid contains an
299
inactive subunit harboring the R90Q mutation, which decreases the affinity for fructose-6-P by
300
three orders of magnitude, while the other subunit remains unaltered with respect to the catalytic
301
power and sugar affinity. Like in the cases of hybrids B and C, hybrid D shows the characteristic
302
substrate inhibition pattern as a function of the MgATP concentration, with a Ki for MgATP of 1
303
mM, at low concentrations of fructose-6-P (Table 2 and Fig. S4C). Also, the KMapp for fructose-6-
304
P increases from 16 (at 0.1 mM MgATP) to 176 µM under inhibitory conditions (5 mM MgATP,
305
Table 3). However, the shape of the saturation curves for fructose-6-P turn from negative
306
cooperativity at low MgATP concentrations to slightly sigmoidal shape at higher concentrations
307
(Hill number from 0.7 to 1.4, Fig S5C.). This could imply that the sigmoid response for fructose-
308
6-P is not fully explained by direct homotropic communication between fructose-6-P sites
AC C
EP
TE D
297
15
ACCEPTED MANUSCRIPT
observed at elevated MgATP concentrations, but rather by heterotropic interactions between one
310
site for fructose-6-P and allosteric sites for ATP either localized in the same or in the next
311
subunit. Alternatively, the sigmoid response for fructose-6-P could be related to a catalytic effect
312
as is suggested by the hyperbolic curves obtained for the hybrid C. However, these results should
313
be taken with caution. Differently from hybrids B and C, the rate of exchange determined for
314
hybrid D was extremely low; it took about 12 h to reach equilibrium, while the other hybrid
315
dimers required less than 160 min. Moreover, the final population of hybrid was lower than that
316
expected for an unbiased exchange equilibrium (Fig. S6). These observations indicate that hybrid
317
D presents conformational changes beyond the perturbation of the fructose-6-P biding itself.
318
Therefore, the decrease in the cooperativity observed in hybrid D would be difficult to interpret
319
in terms of a mechanism of allosteric communication between both fructose-6-P binding sites.
M AN U
SC
RI PT
309
In order to obtain a whole picture of the allosteric MgATP inhibition mechanism,
321
structural information should be superimposed with the protein dynamics [14]. Based on SAXS
322
measurements, Pfk-2 displays a dynamic equilibrium between an open and a close state [15]. The
323
open state corresponds to the conformation of the free enzyme and the closed one is induced
324
either by fructose-6-P or allosteric ATP binding. The conformational changes were described as
325
domain motion between the interface and the major domain of each subunit that open or close the
326
cleft of each active site. Low frequency normal modes analyses of proteins dynamics have the
327
potential to describe functional conformational changes [16, 17]. Based in this hypothesis,
328
Mitternacht et al. have introduced a quantity called binding leverage, that measures the ability of
329
a binding site to couple to the intrinsic motions of a protein [18] and leverage coupling (LC) [19],
330
to determine the degree of connection between two sites. The LC analysis between the allosteric
331
site (represented by Tyr23) and the rest of structure was determined on the SPACER server [20].
332
Figure 4 shows how the strength of allosteric communication is mapped on the structure of Pfk-2
AC C
EP
TE D
320
16
ACCEPTED MANUSCRIPT
between Y23 and the rest of residues in the oligomer according to the LC values. Y23 presents
334
large LC values between residues of the sugar site localized either in the same or in the partner
335
subunit (Fig. 4). This web of interactions would be responsible for heterotopic interactions
336
between the allosteric ATP and the fructose-6-P binding sites. Conversely, no coupling was
337
observed (low LC values, blue color in Fig. 4) between the allosteric site and surface residues or
338
with the catalytic ATP site. Although, several residues that directly contact the sugar are coupled
339
(LC values over 20, table S4), Y23, Lys27 and Arg90 deserve a closer inspection.
SC
RI PT
333
The intrasubunit coupling predicted by the LC analysis would be rationalized structurally
341
by a simple and direct mechanism. In each monomer, fructose-6-P and the allosteric ATP are
342
positioned close to each other within the large cleft conformed by the major and minor domains
343
and, both are intimately related through the side chain of Lys27 (Fig. 1). X-ray structures of Pfk-
344
2 show the ε-amino group of Lys27 coordinated either with the phosphate of fructose-6-P (PDB
345
ID 3N1C) or the ɣ-phosphate of the allosteric MgATP (PDB ID 3CQD). Therefore, if the ε-
346
amino group of Lys27 prevents the simultaneous binding of fructose-6-P and allosteric MgATP,
347
the KMapp for fructose-6-P should increase due to a mutually exclusive binding mechanism.
348
Additionally, a partial mechanism would be plausible if Lys-27 decreases the simultaneous
349
energy binding of fructose-6-P and the allosteric MgATP with respect to their independent
350
binding event.
TE D
EP
AC C
351
M AN U
340
Although the kinetic analysis of the hybrids do not dissect between intra or inter negative
352
interactions between the allosteric ATP and fructose-6-P sites, intersubunit negative interactions
353
can be mapped at the level of Arg90 by the LC analysis. Like in the case of Lys27, the positive
354
charge of Arg90 is able to contact the phosphate group of the sugar. However, when MgATP
355
binds to the allosteric site, the side chain of Arg90 points outwards the active site [2], which in
356
turn should increase the KMapp for fructose-6-P. This conformational change requires a rotation of 17
ACCEPTED MANUSCRIPT
the χ1 dihedral angle from −64.38 to 70.2° [2]. In the “out orientation”, Arg-90 makes a π-cation
358
interaction with Trp-88 and an ion pair with Asp-87 whose side chain is in close contact with side
359
chains of the partner subunit [3]. This web of interactions is part of the hinge of each active site
360
and hence its formation could be sensitive with respect to the major conformational changes
361
described by SAXS or the LC analyses.
RI PT
357
In conclusion, kinetic characterization of hybrid dimers suggests that homotropic
363
interaction between fructose-6-P sites is not related with the sigmoid response for fructose-6-P
364
induced by ATP (hybrid D) and indicate that direct homotropic interactions between allosteric
365
ATPs are not required for full allosteric inhibition. Notably, the sigmoid response for fructose-6-
366
P can be decoupled from the increment of its apparent KM changing the relative localization of
367
the catalytic subunit with respect the allosteric site. This effect suggests that the sigmoid response
368
is related with the catalytic turnover rather than to substrate binding. Also, structural analysis and
369
in silico dynamics support negative heterotopic interactions (between allosteric ATP and
370
fructose-6-P sites) operating between and within each subunit where Lys27 and R90 would play a
371
key role decreasing the apparent affinity for fructose-6-P.
374 375 376 377
M AN U
TE D
EP
373
AC C
372
SC
362
378 379 380
18
384 385 386 387 388 389 390 391 392
EP
383
Figure 4. Leverage coupling analysis between the allosteric ATP site and the rest of the structure. The analysis was performed on Spacer Server using the Pfk-2-fructose-6-P complex (PDB ID 3N1C). In the case of the structure obtained in presence of MgATP (PDB ID 3CQD) the analysis shows similar results (data not shown). The leverage coupling between a key residue of the allosteric site (Tyr23, green) and the rest of structure is colored according to the scale at the top: low and high values of leverage coupling in blue and red respectively. Some key residues are represented: allosteric site (Tyr23 and Lys27), the superfamily-conserved catalytic motif NXXE residues (Asn187 and Gln190) and the sugar site (Arg90 and Lys27).
AC C
382
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
393 394 395
20
ACCEPTED MANUSCRIPT
Acknowledgments
397 398 399 400
This work was supported by Fondecyt 1090336 and 1130510, Conicyt, Chile (JB). Pablo Villalobos was supported by a Conicyt doctoral fellowship (21151101). We thank Dr. Victoria Guixé for the helpful comments on the manuscript. We also thank the reviewers for helpful suggestions and modifications.
401
RI PT
396
Author Contribution Statement
403 404 405
Conceived and designed the experiments: MB, PV, JB. Performed the experiments: MB, PV, FS. Analyzed the data: MB, PV, JB. Contributed reagents/materials/analysis tools: JB. Wrote the paper: MB, PV, JB
M AN U
SC
402
AC C
EP
TE D
406
21
ACCEPTED MANUSCRIPT
Supplementary Material
408 409
Figure S1. Effect of fructose-6-P on the tagged mutant R206E/K207D activity at two fixed MgATP concentrations
410 411 412 413 414 415 416 417
Table S1. Kinetic parameters for fructose-6-P saturations of the R206E/K207D mutant
418 419
Table S3A. Kinetic parameters for allosteric inhibition of the Y23A mutant measured as function of the MgATP concentration.
420 421
Table S3B. Kinetic parameters for the allosteric inhibition of the Y23A mutant measured as function of the fructose-6-P concentration
SC
Table S2. Parameters obtained from binding assays measured by intrinsic fluorescence
M AN U
Figure S3. Kinetic characterization of the Y23A mutant.
Figure S4. Inhibition of the hybrid enzymes by MgATP.
TE D
Figure S5. Fructose-6-P saturation curves for the hybrid enzymes. Figure S6. Kinetics of subunit exchange for hybrid D formation. Table S4. Leverage coupling analysis between the allosteric site (Y23) and the rest of structure.
EP
431
Figure S2. Effect of fructose-6-P on the intrinsic fluorescence of the wild type, R90Q/D256N and Y23A enzymes.
AC C
422 423 424 425 426 427 428 429 430
RI PT
407
22
ACCEPTED MANUSCRIPT
REFERENCES
EP
TE D
M AN U
SC
RI PT
1. Babul J (1978) Phosphofructokinases from Escherichia coli. Purification and characterization of the nonallosteric isozyme, J Biol Chem. 253, 4350-4355. 2. Cabrera R, Baez M, Pereira HM, Caniuguir A, Garratt RC & Babul J (2011) The Crystal Complex of Phosphofructokinase-2 of Escherichia coli with Fructose-6-phosphate: Kinetic and Structural Analysis of the Allosteric ATP inhibition, J Biol Chem. 286, 5774-5783. 3. Cabrera R, Ambrosio ALB, Garratt RC, Guixé V & Babul J (2008) Crystallographic Structure of Phosphofructokinase-2 from Escherichia coli in Complex with Two ATP Molecules. Implications for Substrate Inhibition, J Mol Biol. 383, 588-602. 4. Robey EA & Schachman HK (1985) Regeneration of active enzyme by formation of hybrids from inactive derivatives: implications for active sites shared between polypeptide chains of aspartate transcarbamoylase, Proc Natl Acad Sci U S A. 82, 361-365. 5. Ortigosa AD, Kimmel JL & Reinhart GD (2003) Disentangling the Web of Allosteric Communication in a Homotetramer: Heterotropic Inhibition of Phosphofructokinase from Bacillus stearothermophilus†, Biochemistry. 43, 577-586. 6. Fenton AW, Paricharttanakul NM & Reinhart GD (2004) Disentangling the Web of Allosteric Communication in a Homotetramer: Heterotropic Activation in Phosphofructokinase from Escherichia coli†, Biochemistry. 43, 14104-14110. 7. Fenton AW & Reinhart GD (2003) Mechanism of Substrate Inhibition in Escherichia coli Phosphofructokinase†, Biochemistry. 42, 12676-12681. 8. Cabrera R, Babul J & Guixé V (2010) Ribokinase family evolution and the role of conserved residues at the active site of the PfkB subfamily representative, Pfk-2 from Escherichia coli, Arch Biochem Biophys. 502, 23-30. 9. Baez M, Wilson Christian AM, Ramírez-Sarmiento César A, Guixé V & Babul J (2012) Expanded Monomeric Intermediate upon Cold and Heat Unfolding of Phosphofructokinase-2 from Escherichia coli, Biophys J. 103, 2187-2194. 10. Guixé V, Rodríguez PH & Babul J (1998) Ligand-Induced Conformational Transitions in Escherichia coli Phosphofructokinase 2: Evidence for an Allosteric Site for MgATP2- †, Biochemistry. 37, 13269-13275. 11. Baez M, Wilson CAM & Babul J (2011) Folding kinetic pathway of phosphofructokinase-2 from Escherichia coli: A homodimeric enzyme with a complex domain organization, FEBS Lett. 585, 2158-2164. 12. Park S-Y, Quezada CM, Bilwes AM & Crane BR (2004) Subunit Exchange by CheA Histidine Kinases from the Mesophile Escherichia coli and the Thermophile Thermotoga maritima†, Biochemistry. 43, 2228-2240. 13. Fenton AW & Reinhart GD (2002) Isolation of a Single Activating Allosteric Interaction in Phosphofructokinase from Escherichia coli†, Biochemistry. 41, 13410-13416. 14. Kern D & Zuiderweg ERP (2003) The role of dynamics in allosteric regulation, Current Opinion in Structural Biology. 13, 748-757. 15. Cabrera R, Fischer H, Trapani S, Craievich AF, Garratt RC, Guixé V & Babul J (2003) Domain Motions and Quaternary Packing of Phosphofructokinase-2 from Escherichia coli Studied by Small Angle X-ray Scattering and Homology Modeling, Journal of Biological Chemistry. 278, 12913-12919.
AC C
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
23
ACCEPTED MANUSCRIPT
16. Alexandrov V, Lehnert U, Echols N, Milburn D, Engelman D & Gerstein M (2005) Normal modes for predicting protein motions: A comprehensive database assessment and associated Web tool, Protein Science : A Publication of the Protein Society. 14, 633-643. 17. Ma J (2005) Usefulness and Limitations of Normal Mode Analysis in Modeling Dynamics of Biomolecular Complexes, Structure. 13, 373-380. 18. Mitternacht S & Berezovsky IN (2011) Binding Leverage as a Molecular Basis for Allosteric Regulation, PLoS Comput Biol. 7, e1002148. 19. Mitternacht S & Berezovsky IN (2011) Coherent Conformational Degrees of Freedom as a Structural Basis for Allosteric Communication, PLoS Comput Biol. 7, e1002301. 20. Goncearenco A, Mitternacht S, Yong T, Eisenhaber B, Eisenhaber F & Berezovsky IN (2013) SPACER: server for predicting allosteric communication and effects of regulation, Nucleic Acids Research. 41, W266-W272.
RI PT
476 477 478 479 480 481 482 483 484 485 486 487
AC C
EP
TE D
M AN U
SC
488
24
ACCEPTED MANUSCRIPT
The inhibitory mechanism of ATP on Pfk-2 is characterized creating hybrid dimers Homotropic interaction do not play a role in the enzymatic inhibition induced by ATP
AC C
EP
TE D
M AN U
SC
RI PT
Low frequency normal modes support inhibitory heterotropic interactions
S0300-9084(16)30166-3
DOI:
10.1016/j.biochi.2016.08.013
Reference:
BIOCHI 5047
To appear in:
Biochimie
Received Date: 26 December 2015 Accepted Date: 29 August 2016
Please cite this article as: P. Villalobos, F. Soto, M. Baez, J. Babul, Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers, Biochimie (2016), doi: 10.1016/ j.biochi.2016.08.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
1 2
Regulatory network of the allosteric ATP inhibition of E. coli phosphofructokinase-2 studied by hybrid dimers
3
Pablo Villalobos1, Francisco Soto1, Mauricio Baez2*, Jorge Babul1*.
4 1
6 7
2
Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Santiago, Chile.
RI PT
5
Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias Químicas y Farmacéuticas, Universidad de Chile, Santiago, Chile.
8
10
* Corresponding authors
SC
9
E-mail: [email protected] (MB); [email protected] (JB)
M AN U
11
Abstract
13
We have proposed an allosteric ATP inhibition mechanism of Pfk-2 determining the structure
14
of different forms of the enzyme together with a kinetic enzyme analysis. Here we
15
complement the mechanism by using hybrid oligomers of the homodimeric enzyme to get
16
insights about the allosteric communication pathways between the same sites or different ones
17
located in different subunits. Kinetic analysis of the hybrid enzymes indicate that homotropic
18
interactions between allosteric sites for ATP or between substrate sites for fructose-6-P have a
19
minor effect on the enzymatic inhibition induced by ATP. In fact, the sigmoid response for
20
fructose-6-P observed at elevated ATP concentrations can be eliminated even though the
21
enzymatic inhibition is still operative. Nevertheless, leverage coupling analysis supports
22
heterotropic interactions between the allosteric ATP and fructose-6-P binding occurring
23
between and within each subunit.
AC C
EP
TE D
12
1
ACCEPTED MANUSCRIPT
24
1. Introduction The allosteric regulation induced by MgATP is a common property of E. coli
26
phosphofructokinases Pfk-1 and Pfk-2 [1, 2]. Nevertheless, information about their regulation
27
cannot be transferred between them since both structures belong to different folds.
RI PT
25
In order to understand the structural basis of the allosteric mechanism of Pfk-2, we have
29
determined the X-ray structure of different forms of the enzyme, including complexes with
30
fructose-6-P (PDB ID: 3N1C [2]), or MgATP (PDB ID: 3CQD, [3]) which, with a thorough
31
kinetic analysis, led us to propose a mechanism for the allosteric inhibition [2]. In this work we
32
complement the mechanism by using hybrid oligomers to get insights about the allosteric
33
communication pathways between the same sites or different ones located in different subunits.
M AN U
SC
28
The rationale of our experimental approach is to eliminate or perturb the binding capacity
35
in one subunit while the partner subunit remains intact [4]. For example, homotropic interactions
36
for fructose-6-P binding sites in B. stearothermophilus phosphofructokinase and E. coli Pfk-1
37
have been studied by hybrid homotetramers, where the binding site for fructose-6-P was
38
eliminated in one or several subunits [5, 6]. A similar approximation was used to detect long
39
range interactions that accounts for substrate inhibition induced by ATP in Pfk-1 [7]. In that case,
40
a hybrid homotetramer, containing a single intact ATP binding site, failed to show substrate
41
inhibition by ATP, supporting the fact that communication arises between the subunits and not
42
within the subunits.
43
Structural organization of Pfk-2. One site for fructose-6-P and two binding sites for MgATP in
44
adjacent locations are found in each subunit (Fig. 1A). Like its non-allosteric homologues, the
45
functional structure required for catalysis in Pfk-2 is the homodimeric state. Key residues to
46
provide electrostatic and aromatic stacking interactions with the allosteric ATP are established by
AC C
EP
TE D
34
2
ACCEPTED MANUSCRIPT
the side chains of K27 and Y23 respectively. Both side chains are contributed from the partner
48
subunit due to the intertwined nature of the homodimeric interphase (Fig. 1B).
49
The saturation curves for fructose-6-P at different MgATP concentrations indicate that inhibition
50
occurs due to a decrease in the apparent affinity for fructose-6-P, together with a transition from
51
hyperbolic to sigmoidal behavior and partial reduction of the catalytic constant [2]. These
52
observations could be explained in part by the perturbation of a network of interactions between
53
basic residues and the phosphate moiety of fructose-6-P. In this way, the binding of MgATP
54
sequesters the side chain of K27 and provokes a reorientation of the side chain of R90 out the
55
active site (Figure 1B). Elimination of the formal charge of R90 by site-directed mutagenesis
56
increases the KMapp value for fructose-6-P in three orders of magnitude [8] supporting the idea
57
that the side chain reorientation of R90 must play a role in the MgATP inhibition. Importantly,
58
according to the crystal structures, these conformational changes occur in both subunits of the
59
Pfk-2 dimer [2]. However, the apparent diminution of fructose-6-P affinity is also accompanied
60
by a change of the sugar saturation curve from hyperbolic to a sigmoid shape, suggesting a
61
communication route between subunits [2].
TE D
M AN U
SC
RI PT
47
Hybrids enzymes show that homotropic interactions between fructose-6-P sites and
63
allosteric ATP sites are not fundamental for allosteric inhibition neither for the sigmoid response
64
observed for fructose-6-P at elevated concentration of the allosteric inhibitor. Conversely, the
65
sigmoid response for fructose-6-P seems more related with the catalytic turnover rather than
66
substrate binding. Further analysis performed in silico, shows that several heterotropic
67
interactions between fructose-6-P and allosteric sites for ATP are operating within and between
68
the Pfk-2 subunits. These interactions are created between key residues for ATP and fructose-6-P
69
binding.
AC C
EP
62
70
3
ACCEPTED MANUSCRIPT
71 72 73
RI PT
74 75 76
SC
77
81 82 83
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
EP
87
Figure 1. Dimeric structure of Pfk-2 showing the location of fructose-6-P (F6P), allosteric ATP (ATPa) and substrate ATP (ATPc). A. The image shows the Pfk-2-fructose-6-P complex (PDB ID 3N1C) and the MgATP molecules binding at the active and allosteric sites; the MgATP coordinates were obtained from the Pfk-2-ATP crystallographic structure (PDB ID 3CQD) after aligning both structures. The Pfk-2 monomers are colored blue and white, ATP and F6P are represented with licorice sticks and magnesium ions are shown as green spheres. B. Structural changes induced by F6P or MgATP. Location of the Y23 and K27 side chains of the Pfk-2MgATP and Pfk-2-fructose-6-P complexes are shown as yellow and green sticks respectively. Note that in each allosteric site the key residues Y23 and K27 are provided by the β3 and β2 strands of the partner subunit. However, to be consistent with the structural organization of Pfk-2, each allosteric site is considered to belong to only one subunit, irrespective of whether the two sidechains are supplied by the partner subunit. Key interactions: Y23 with the adenine ring of the allosteric ATP and K27 either with ATP or fructose-6-P. The R90 sidechain in the presence of fructose-6-P is seen pointing towards the active site and outwards in presence of MgATP. The catalytic base of Pfk-2, D256, does not show structural differences on the superposition.
AC C
86
TE D
84 85
79
M AN U
80
78
4
ACCEPTED MANUSCRIPT
103 104
2. Materials and Methods 2.1. Enzyme purification and site direct mutagenesis. Site directed mutagenesis was carried out
106
with the Gene TailorTMsystem using pET21 plasmid containing the Pfk-2 gene as template.
107
Protein purification was carried out as is described by [9].
108
2.2. Isolation and characterization of hybrids. To promote subunit exchange, mixtures of Pfk-2
109
mutants were incubated at 20 °C in 25 mM Tris buffer pH 8.2, 5 mM MgCl2, 5 mM DTT for 3 h
110
(except for hybrid D, 20 h) and hybrid dimers separation was achieved by ion exchange
111
chromatography (Mono Q column 5/10 (Pharmacia) using a Water’s system equipped with a UV
112
detector. A charged tag (R206E/K207D) was introduced in the surface of Pfk-2 to isolate the
113
hybrids by anion exchange. The presence of the tag did not modify the kinetic properties with
114
respect to the wild type enzyme (Fig. S1 and Table S1). Each hybrid was constructed on the basis
115
of a dimer composed by an active and inactive subunit (Table 1). Hybrid A was obtained from a
116
mixture of the inactive mutant D256N and the tagged mutant R206E/K207D. Hybrid B was
117
created by mixing the tagged mutant harboring the Y23A mutation with the inactive D256N
118
mutant. Hybrid C was obtained by mixing the double mutant D256N/Y23A with the tagged
119
mutant. Hybrid D was obtained mixing the D256N/R90Q mutant and the active tagged mutant.
120
2.3. Kinetic and binding properties of the hybrid enzymes. The MgATP-induced inhibition was
121
evaluated as a function of the fructose-6-P or MgATP concentration using two fixed co-substrate
122
concentrations [2]. Phosphofructokinase activity was measured by coupling the fructose-1,6bisP
123
production to the oxidation of NADH followed at 340 nm [1]. All kinetic assays were performed
124
at room temperature (25 °C) by employing 96-well plates and read in a microplate reader (Biotek,
125
Synergy). The binding constants for fructose-6-P or MgATP were evaluated using intrinsic
AC C
EP
TE D
M AN U
SC
RI PT
105
5
ACCEPTED MANUSCRIPT
fluorescence measurements in a Shimadzu PC-5031 spectrofluorimeter as is described by Guixé
127
et al. [10]
128
2.4. Treatment of kinetic and affinity data. Fluorescence binding assays were performed to
129
determine the fructose-6-P and MgATP affinity for the wild type enzyme and the R90Q and
130
Y23A mutants [8] [10]. Coherent with the presence of two MgATP molecules per Pfk-2
131
monomer [3], activity as a function of the MgATP concentration was fitted to the following
132
substrate inhibition model:
[E ]t × k cat × [S ] [S ] K + [S ]× 1 + M
K i
M AN U
v=
133
SC
RI PT
126
where kcat corresponds to the catalytic constant obtained from the initial steady-state velocity at
135
the described conditions, [S] equals the concentration of MgATP, KMapp is the apparent affinity of
136
MgATP and Ki is the apparent binding constant for MgATP acting as an inhibitory substrate. On
137
the other hand, data from activity assays as a function of the fructose-6-P concentration were
138
fitted to a Hill equation. All data were plotted and analyzed by nonlinear regressions using
139
SigmaPlot 11.0.
AC C
EP
TE D
134
6
ACCEPTED MANUSCRIPT
3. Results and Discussion
141
3.1. Mutants and hybrids design. In order to determine the independent behavior of sites present
142
in different subunits of an oligomer, or between sites present in the same subunit, hybrid
143
oligomers were created to break the expected symmetry by decreasing or abolishing the affinity
144
of allosteric or substrate sites in individual subunits [4]. Thus, there are two types of Pfk-2 hybrid
145
dimers, one to study heterotropic interactions between the allosteric MgATP and sugar sites
146
(hybrids B and C in table 1), and another to study homotropic interaction between the fructose-6-
147
P binding sites present in the dimer (hybrid D in table 1). For this purpose it is necessary to
148
eliminate or diminish the affinity of one allosteric site and/or the fructose-6-P binding site.
M AN U
SC
RI PT
140
Table 1 summarizes the kinetic properties of the homodimers required to create the
150
hybrids studied in this work. To create hybrids B and C it is necessary to avoid the stacking
151
interaction between the adenine moiety of the allosteric ATP and Y23 (Fig. 1B), so the residue
152
was replaced by alanine. This mutation diminishes the affinity of the allosteric ATP in one order
153
of magnitude with respect to the wild type enzyme, as indicated by the binding curves obtained
154
from intrinsic fluorescence measurements (Fig. S2A and Table S2). The diminished affinity
155
towards MgATP was attributed specifically to the alteration of the allosteric site, since the Y23A
156
mutant shows an attenuated enzymatic inhibition (a large Ki value together with a minor
157
increment of the KMapp for fructose-6-P), without changes in the catalytic parameters (Fig. S3 and
158
Table S3A and B). On the other hand, to create hybrid D it is necessary to disrupt the sugar
159
phosphate binding site, so R90 was replaced by glutamine (R90Q). This mutation increases both
160
the apparent KMapp [8] and the Kd for fructose-6-P by three orders of magnitude (Fig. S2B and
161
Table S2). The structural effects seem to be confined to the sugar site, since the allosteric ATP
162
binding was not modified by the mutation [8]. In other words, we utilized the Y23A mutant that
AC C
EP
TE D
149
7
ACCEPTED MANUSCRIPT
has a reduced affinity for the allosteric MgATP and the R90Q mutant with a decreased affinity
164
for fructose-6-P. Moreover, an additional mutation was added to each hybrid dimer (D256N) to
165
inactivate one active site and hence facilitate the kinetic analysis. The D256N mutant decreases
166
the catalytic activity in two orders of magnitude without a substantive alteration in the inhibitory
167
parameters [4].
168
SC
Table 1. Mutants used and properties of the heterodimers
RI PT
163
Mutant 2 Tag mutant: Active.
Heterodimer Hybrid A: Inactive subunit with intact allosteric ATP and fructose-6-P sites and active partner subunit.
a
D256N: Inactive.
Y23A/Tag Mutant: Active with decreased allosteric ATP binding.
Hybrid B: Intact allosteric and active sites in the same subunit.
Hybrid B
D256N/Y23A: Inactive, with decreased allosteric ATP binding.
Tag Mutant: Active.
TE D
Hybrid C:
Diagram Hybrid A
Hybrid C
EP
Intact allosteric and active sites in different subunits.
Tag Mutant: Active.
Hybrid D: Hybrid D To test inter subunit communication between fructose-6-P sites. An inactive subunit with decreased fructose-6-P binding and active partner subunit. The columns referred as mutant 1 and 2 represent the homodimers used to create the hybrids enzymes described in columns 3 and 4. The subunits of the dimer are highlighted in different colors following the color of the homodimeric mutants used to create them. Each active site is composed by a large domain (circle), a fructose-6-P binding site (triangle) and the ATPa binding site (rhombus). The interactions of Y23 with the allosteric ATP are shown in blue since formally Y23A is donated by the partner subunit. “X” symbolizes a mutation introduced in the respective site. The blue subunit contains the negative tag mutations (R206E/K207D). Hybrid A, one inactive subunit (mutant D256N) with intact allosteric and fructose-6-P binding sites. Hybrid B, one inactive subunit (mutant D256N) with the allosteric site mutated (mutant Y23A). Hybrid C, one inactive subunit (D256N) with its adjacent allosteric site mutated (mutant Y23A). Hybrid D, one inactive subunit (mutant D256N) with its fructose-6-P binding site mutated (mutant R90Q).
AC C
D256N/R90Q: Inactive, with decreased fructose6-P binding.
M AN U
Mutant 1 D256N: Inactive.
8
ACCEPTED MANUSCRIPT
169
3.2. Formation and isolation of Pfk-2 hybrid dimers. The formation of hybrid dimers or
171
heterodimers occurs due to the dynamic nature of the association and dissociation events of the
172
oligomers present in solution. Fig. 2A compares the anion exchange chromatograms obtained for
173
the inactive mutant (mutant D256N; upper panel), Pfk-2 containing a “tag of charge” (middle
174
panel), and the result of incubation of an equimolar mixture of both enzymes (bottom panel).
175
Coherently, the dimer containing two negatively tagged subunits elutes at higher NaCl
176
concentrations with respect to the D256N mutant (without the “tag of charge”), and the mixture
177
of both Pfk-2 forms results in a new peak between the tagged and untagged homodimers. The
178
new peak was collected and kinetically characterized, showing a kcat value of 25 s-1,
179
approximately half of the value observed for the wild type enzyme (56 s-1), confirming the
180
presence of a hybrid dimer created by a fully active and an inactive subunit (Fig. 2A, peak III). In
181
addition, we reasoned that if the hybrid dimer equilibrium exchange is allowed to reestablish, the
182
appearance of tagged and inactive parental homodimers should be observed. Therefore, to
183
reestablish the equilibrium exchange, the purified new peak was equilibrated in the exchange
184
buffer (using a desalting column) and reinjected into the anion exchange column. The sample
185
shows the presence of two additional peaks corresponding to the wild type and the inactive
186
mutants (data not shown). We observed that desalting of the new peak was a condition to recover
187
the parental homodimers, supporting the fact that elevated salt concentrations slows the dimer
188
equilibrium exchange. Addition of ligands like fructose-6-P and MgATP also slows the dimer
189
equilibrium exchange rate.
AC C
EP
TE D
M AN U
SC
RI PT
170
190
9
EP
Figure 2. Separation of hybrid enzymes using anion exchange chromatography. A. The figure shows the elution profile of the inactive mutant (D256N, upper panel), Pfk-2 modified with a negative charged tag (R206E/K207D, middle panel) and a mixture of both enzymes previously incubated for 3 h at 20 °C (bottom panel). The specific activity of the new peak (25 s1 ) was about half of that of the tagged enzyme (56 s-1). Anion exchange chromatography was carried out using a MonoQ HR 5/5 column and the proteins were eluted with a linear KCl gradient from 0.2 to 0.75 M in presence of 25 mM Tris pH 8.2, 1 mM DTT and 5 mM MgCl2. The presence of KCl or ligands (either MgATP or fructose-6-P) stops the exchange reaction avoiding hybrid re-equilibration. B. Kinetics of subunit exchange between the inactive mutant and the tagged Pfk-2. Samples of a mixture of tagged wild type and D256N mutant were withdrawn and injected into MonoQ HR column for each time. Symbols represent the temporal variation of absorbance peak area measured for peaks I (○), II (Δ) and III (●). The continuous lines represent an exponential function used to describe the temporal appearance of peak III or the disappearance of peaks I and II. The kinetic constant for the appearance of the hybrid was 0.03 min-1.
AC C
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
208
10
ACCEPTED MANUSCRIPT
209
Table 2. Kinetic parameters for the MgATP inhibition of the hybrid enzymes The constants were obtained from the kinetics in Fig. S4.
1 mM fructose-6-P
Enzyme
kcat s-1
KMapp (ATP), µM
Ki, mM
kcat (s-1)
Hybrid A
18±2
12±6
0.57±0.09
20.2±0.5
21.8±4.5
0.87±0.07
11.8±1.7
Hybrid B 10.8±1.0 15±2
11±3
1.1±0.1
20±1
Hybrid D
6.6±0.2
6±1
1.1±0.2
7.6±0.3
8.4±0.4
14±2
13±2
18±3
13±2
14 ±2
4.1±0.7
ND
M AN U
210
KMapp(ATP), µM Ki, mM
SC
Hybrid C
RI PT
0.05 mM fructose-6-P
In order to observe the kinetics of subunit exchange, equimolar mixtures of tagged wild
212
type and D256N mutant were injected into the column after several incubation times (between 10
213
and 1,080 min for hybrid A and between 3 and 27 h for hybrid D), and the concentrations of each
214
species was quantified using their chromatographic areas. As indicated in Fig. 2B, the relative
215
abundance of the hybrid dimer was 45 %, suggesting an unbiased subunit exchange. Fitting of the
216
exponential function to the time course of hybrid formation gives a rate constant between 2 and 6
217
x 10-4 s-1. This value is similar to the kinetic constant obtained for the slower process of the
218
unfolding of Pfk-2 determined from the kinetics of chemical denaturation (1.5 x 10-4 s-1) [11].
219
The agreement between both constants suggests that the dimer unfolding and dissociation is the
220
limiting step for dimer exchange [12].
221
3.3. Kinetic analyses of Pfk-2 hybrid enzymes. First we ask if each active site works
222
independently with respect to catalysis in the Pfk-2 dimer. For this purpose, a hybrid dimer with
223
an inactive subunit harboring the D256N mutation was created, denominated as hybrid A (Table
224
1, first row). Additionality, like each hybrid studied in this work, it contains a “charged tag”
225
(R206E/K207D) on the protein surface of one subunit to isolate the heterodimer hybrids by anion
AC C
EP
TE D
211
11
ACCEPTED MANUSCRIPT
exchange chromatography [13]. Fig. 3 shows the typical kinetic analysis performed to
227
characterize the allosteric inhibition of wild type Pfk-2, but performed with the hybrid A. The
228
allosteric effect is evidenced by a pronounced inhibition when the saturation curves are measured
229
as a function of the MgATP concentration (Fig. 3A). Curve fitting to models of substrate
230
inhibition retrieve the apparent substrate inhibition constant (Ki), which represents the apparent
231
affinity of the allosteric ATP for the E-ATP complex (see Material and Methods). Thus, this
232
apparent affinity is high (Ki = 0.57 mM) at low fructose-6-P concentrations and low (Ki = 14
233
mM) at elevated concentrations of the sugar-phosphate (Table 2).
SC
RI PT
226
M AN U
234
The allosteric inhibition can also be evaluated as an increase of the apparent KMapp for
236
fructose-6-P at elevated MgATP concentrations. In this case, the diminished apparent affinity for
237
fructose-6-P is accompanied by a change towards a sigmoid response (Fig. 3B). In the case of
238
hybrid A, the kinetic parameters for both substrates, and the regulatory kinetic properties were
239
very similar to the wild type values, except for the kcat value which is half of that of the wild type
240
enzyme (56 s-1, [2]). These observations support the idea that both subunits work independently,
241
at least with respect to the catalytic turnover of each active site.
244 245 246
EP
243
AC C
242
TE D
235
247 248
12
ACCEPTED MANUSCRIPT
Table 3. Kinetic parameters for fructose-6-P saturations of hybrid enzymes.
SC
RI PT
The constants were obtained from the kinetics shown in Fig. S5
Figure 3. Kinetic characterization of hybrid A. A. Initial velocities for hybrid A as function of the MgATP concentration measured at low (0.05 mM ○) and high (1 mM ●) fructose-6-P concentrations. The lines represent the best fit to an uncompetitive inhibition model and the resulting kinetic constants are shown in Table 2. B. Kinetic assays for hybrid A were performed with fructose-6-P as variable substrate at two fixed MgATP concentrations: 0.1 mM (∆) and 5 mM (▲). The continuous line represents a fit of each curve to the Hill equation; resulting parameters are shown in Table 3.
258
In order to disturb one allosteric site for ATP in the Pfk-2 dimer we designed hybrids B
259
and C. As is indicated in the second and third row of the table 1, both hybrids B and C carry out
260
the Y23A mutation in one subunit. The difference between them is the localization of the inactive
261
subunit with respect to the localization of the Y23A mutation. Nevertheless, taken into account
262
that the D256N mutation does not perturb the affinities for fructose-6-P or allosteric ATP binding
263
[2], both hybrids have the same symmetry with respect to the binding sites. Therefore, we
264
expected similar regulatory kinetics for both hybrids B and C.
TE D
EP
AC C
265
M AN U
249 250 251 252 253 254 255 256 257
266
13
ACCEPTED MANUSCRIPT
0.1 mM MgATP nH
kcat, s-1
KMapp(F6P),µM
nH
Hybrid A 32.8±0.3
35±2
1.36±0.05
24.4±0.6
258±27
1.9±0.2
Hybrid B
16.2
60.9
1.1
12.09±0.03
176.9±3.9
1.7±0.2
Hybrid C
27±1
42±7
1.0±0.3
23.0±0.4
218±15
1.1±0.2
Hybrid D
9.0±0.6
16±4
0.71±0.08
11.6±0.1
176±4
1.41±0.09
.
267
RI PT
KMapp(F6P), µM
Enzyme
kcat s-1
5 mM MgATP
The saturation curves for MgATP determined for hybrids B and C displayed inhibition by
269
substrate (Figs. S4A and S4B). In both cases, the co-substrate (fructose-6-P) increases the
270
apparent inhibition constant (Ki) for MgATP and both enzymes have essentially identical Ki
271
values either at low or high fructose-6-P concentrations (Table 2). Application of the substrate
272
inhibition model gives an Ki for MgATP of 0.87 ± 0.07 mM for the hybrid B and 1.1 ± 0.1 mM
273
for the hybrid C (Table 2), which is close but slightly higher than the Ki measured for hybrid A
274
(0.57 ± 0.09 mM, Table 2). Taken the value of Ki as an indication of the apparent allosteric
275
binding of ATP, affinities of hybrids B and C are similar between them but slightly diminished
276
with respect to hybrid A. In opposition, if both allosteric sites are required for full inhibition, then
277
we expected to observe a Ki = 7 mM, similar to that obtained for the homodimeric Y23A mutant
278
determined at equivalent fructose-6-P concentrations (Table S3A).
M AN U
TE D
EP
Elevated concentrations of MgATP (5 mM) increase the KMapp for fructose-6-P to 164 µM
AC C
279
SC
268
280
for hybrid B and up to 218 ± 15 µM in the case of the hybrid C (Table 3 and Fig.S5A and S5B).
281
These results indicate that the enzymatic inhibition is still operating in both hybrids, since the
282
increments of ATP concentration increase the apparent KM for fructose-6-P reaching a similar
283
value as the one calculated for hybrid A (KMapp = 258 µM, Table 3 and Figure 3B). Otherwise,
284
the value of KMapp for fructose-6-P should be similar to the Y23A mutant (66 µM, Table S3B).
14
ACCEPTED MANUSCRIPT
However, there is an important difference between hybrids B and C. In the case of hybrid C, the
286
saturation curves for fructose-6-P keep their hyperbolic shape (Fig. S5B), i.e., the hill number
287
remains close to 1 (Table 3) when the KMapp for fructose-6-P is increased by the presence of ATP.
288
Differently, the saturation curves for fructose-6-P for hybrid B display a sigmoid response like
289
hybrid A (Table 3 and Fig. S5A). Therefore, the sigmoid saturation for fructose-6-P observed at
290
elevated concentrations of ATP is uncoupled from the allosteric inhibition in hybrid C. This
291
effect, seems more related with the catalytic turnover than to substrate binding since the D256N
292
mutant, used to inactivate one subunit, has similar affinities for fructose-6-P and allosteric
293
binding [2].
M AN U
SC
RI PT
285
294
Based on these results we postulate that direct homotropic interactions between the
295
allosteric sites for MgATP seem not to be fundamental for allosteric inhibition in Pfk-2, since
296
hybrids B and C, with only one functional allosteric site, still present MgATP inhibition. To test the homotropic interactions between the fructose-6-P binding sites of both
298
subunits, hybrid D was characterized kinetically (Table 1, fourth row). This hybrid contains an
299
inactive subunit harboring the R90Q mutation, which decreases the affinity for fructose-6-P by
300
three orders of magnitude, while the other subunit remains unaltered with respect to the catalytic
301
power and sugar affinity. Like in the cases of hybrids B and C, hybrid D shows the characteristic
302
substrate inhibition pattern as a function of the MgATP concentration, with a Ki for MgATP of 1
303
mM, at low concentrations of fructose-6-P (Table 2 and Fig. S4C). Also, the KMapp for fructose-6-
304
P increases from 16 (at 0.1 mM MgATP) to 176 µM under inhibitory conditions (5 mM MgATP,
305
Table 3). However, the shape of the saturation curves for fructose-6-P turn from negative
306
cooperativity at low MgATP concentrations to slightly sigmoidal shape at higher concentrations
307
(Hill number from 0.7 to 1.4, Fig S5C.). This could imply that the sigmoid response for fructose-
308
6-P is not fully explained by direct homotropic communication between fructose-6-P sites
AC C
EP
TE D
297
15
ACCEPTED MANUSCRIPT
observed at elevated MgATP concentrations, but rather by heterotropic interactions between one
310
site for fructose-6-P and allosteric sites for ATP either localized in the same or in the next
311
subunit. Alternatively, the sigmoid response for fructose-6-P could be related to a catalytic effect
312
as is suggested by the hyperbolic curves obtained for the hybrid C. However, these results should
313
be taken with caution. Differently from hybrids B and C, the rate of exchange determined for
314
hybrid D was extremely low; it took about 12 h to reach equilibrium, while the other hybrid
315
dimers required less than 160 min. Moreover, the final population of hybrid was lower than that
316
expected for an unbiased exchange equilibrium (Fig. S6). These observations indicate that hybrid
317
D presents conformational changes beyond the perturbation of the fructose-6-P biding itself.
318
Therefore, the decrease in the cooperativity observed in hybrid D would be difficult to interpret
319
in terms of a mechanism of allosteric communication between both fructose-6-P binding sites.
M AN U
SC
RI PT
309
In order to obtain a whole picture of the allosteric MgATP inhibition mechanism,
321
structural information should be superimposed with the protein dynamics [14]. Based on SAXS
322
measurements, Pfk-2 displays a dynamic equilibrium between an open and a close state [15]. The
323
open state corresponds to the conformation of the free enzyme and the closed one is induced
324
either by fructose-6-P or allosteric ATP binding. The conformational changes were described as
325
domain motion between the interface and the major domain of each subunit that open or close the
326
cleft of each active site. Low frequency normal modes analyses of proteins dynamics have the
327
potential to describe functional conformational changes [16, 17]. Based in this hypothesis,
328
Mitternacht et al. have introduced a quantity called binding leverage, that measures the ability of
329
a binding site to couple to the intrinsic motions of a protein [18] and leverage coupling (LC) [19],
330
to determine the degree of connection between two sites. The LC analysis between the allosteric
331
site (represented by Tyr23) and the rest of structure was determined on the SPACER server [20].
332
Figure 4 shows how the strength of allosteric communication is mapped on the structure of Pfk-2
AC C
EP
TE D
320
16
ACCEPTED MANUSCRIPT
between Y23 and the rest of residues in the oligomer according to the LC values. Y23 presents
334
large LC values between residues of the sugar site localized either in the same or in the partner
335
subunit (Fig. 4). This web of interactions would be responsible for heterotopic interactions
336
between the allosteric ATP and the fructose-6-P binding sites. Conversely, no coupling was
337
observed (low LC values, blue color in Fig. 4) between the allosteric site and surface residues or
338
with the catalytic ATP site. Although, several residues that directly contact the sugar are coupled
339
(LC values over 20, table S4), Y23, Lys27 and Arg90 deserve a closer inspection.
SC
RI PT
333
The intrasubunit coupling predicted by the LC analysis would be rationalized structurally
341
by a simple and direct mechanism. In each monomer, fructose-6-P and the allosteric ATP are
342
positioned close to each other within the large cleft conformed by the major and minor domains
343
and, both are intimately related through the side chain of Lys27 (Fig. 1). X-ray structures of Pfk-
344
2 show the ε-amino group of Lys27 coordinated either with the phosphate of fructose-6-P (PDB
345
ID 3N1C) or the ɣ-phosphate of the allosteric MgATP (PDB ID 3CQD). Therefore, if the ε-
346
amino group of Lys27 prevents the simultaneous binding of fructose-6-P and allosteric MgATP,
347
the KMapp for fructose-6-P should increase due to a mutually exclusive binding mechanism.
348
Additionally, a partial mechanism would be plausible if Lys-27 decreases the simultaneous
349
energy binding of fructose-6-P and the allosteric MgATP with respect to their independent
350
binding event.
TE D
EP
AC C
351
M AN U
340
Although the kinetic analysis of the hybrids do not dissect between intra or inter negative
352
interactions between the allosteric ATP and fructose-6-P sites, intersubunit negative interactions
353
can be mapped at the level of Arg90 by the LC analysis. Like in the case of Lys27, the positive
354
charge of Arg90 is able to contact the phosphate group of the sugar. However, when MgATP
355
binds to the allosteric site, the side chain of Arg90 points outwards the active site [2], which in
356
turn should increase the KMapp for fructose-6-P. This conformational change requires a rotation of 17
ACCEPTED MANUSCRIPT
the χ1 dihedral angle from −64.38 to 70.2° [2]. In the “out orientation”, Arg-90 makes a π-cation
358
interaction with Trp-88 and an ion pair with Asp-87 whose side chain is in close contact with side
359
chains of the partner subunit [3]. This web of interactions is part of the hinge of each active site
360
and hence its formation could be sensitive with respect to the major conformational changes
361
described by SAXS or the LC analyses.
RI PT
357
In conclusion, kinetic characterization of hybrid dimers suggests that homotropic
363
interaction between fructose-6-P sites is not related with the sigmoid response for fructose-6-P
364
induced by ATP (hybrid D) and indicate that direct homotropic interactions between allosteric
365
ATPs are not required for full allosteric inhibition. Notably, the sigmoid response for fructose-6-
366
P can be decoupled from the increment of its apparent KM changing the relative localization of
367
the catalytic subunit with respect the allosteric site. This effect suggests that the sigmoid response
368
is related with the catalytic turnover rather than to substrate binding. Also, structural analysis and
369
in silico dynamics support negative heterotopic interactions (between allosteric ATP and
370
fructose-6-P sites) operating between and within each subunit where Lys27 and R90 would play a
371
key role decreasing the apparent affinity for fructose-6-P.
374 375 376 377
M AN U
TE D
EP
373
AC C
372
SC
362
378 379 380
18
384 385 386 387 388 389 390 391 392
EP
383
Figure 4. Leverage coupling analysis between the allosteric ATP site and the rest of the structure. The analysis was performed on Spacer Server using the Pfk-2-fructose-6-P complex (PDB ID 3N1C). In the case of the structure obtained in presence of MgATP (PDB ID 3CQD) the analysis shows similar results (data not shown). The leverage coupling between a key residue of the allosteric site (Tyr23, green) and the rest of structure is colored according to the scale at the top: low and high values of leverage coupling in blue and red respectively. Some key residues are represented: allosteric site (Tyr23 and Lys27), the superfamily-conserved catalytic motif NXXE residues (Asn187 and Gln190) and the sugar site (Arg90 and Lys27).
AC C
382
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
393 394 395
20
ACCEPTED MANUSCRIPT
Acknowledgments
397 398 399 400
This work was supported by Fondecyt 1090336 and 1130510, Conicyt, Chile (JB). Pablo Villalobos was supported by a Conicyt doctoral fellowship (21151101). We thank Dr. Victoria Guixé for the helpful comments on the manuscript. We also thank the reviewers for helpful suggestions and modifications.
401
RI PT
396
Author Contribution Statement
403 404 405
Conceived and designed the experiments: MB, PV, JB. Performed the experiments: MB, PV, FS. Analyzed the data: MB, PV, JB. Contributed reagents/materials/analysis tools: JB. Wrote the paper: MB, PV, JB
M AN U
SC
402
AC C
EP
TE D
406
21
ACCEPTED MANUSCRIPT
Supplementary Material
408 409
Figure S1. Effect of fructose-6-P on the tagged mutant R206E/K207D activity at two fixed MgATP concentrations
410 411 412 413 414 415 416 417
Table S1. Kinetic parameters for fructose-6-P saturations of the R206E/K207D mutant
418 419
Table S3A. Kinetic parameters for allosteric inhibition of the Y23A mutant measured as function of the MgATP concentration.
420 421
Table S3B. Kinetic parameters for the allosteric inhibition of the Y23A mutant measured as function of the fructose-6-P concentration
SC
Table S2. Parameters obtained from binding assays measured by intrinsic fluorescence
M AN U
Figure S3. Kinetic characterization of the Y23A mutant.
Figure S4. Inhibition of the hybrid enzymes by MgATP.
TE D
Figure S5. Fructose-6-P saturation curves for the hybrid enzymes. Figure S6. Kinetics of subunit exchange for hybrid D formation. Table S4. Leverage coupling analysis between the allosteric site (Y23) and the rest of structure.
EP
431
Figure S2. Effect of fructose-6-P on the intrinsic fluorescence of the wild type, R90Q/D256N and Y23A enzymes.
AC C
422 423 424 425 426 427 428 429 430
RI PT
407
22
ACCEPTED MANUSCRIPT
REFERENCES
EP
TE D
M AN U
SC
RI PT
1. Babul J (1978) Phosphofructokinases from Escherichia coli. Purification and characterization of the nonallosteric isozyme, J Biol Chem. 253, 4350-4355. 2. Cabrera R, Baez M, Pereira HM, Caniuguir A, Garratt RC & Babul J (2011) The Crystal Complex of Phosphofructokinase-2 of Escherichia coli with Fructose-6-phosphate: Kinetic and Structural Analysis of the Allosteric ATP inhibition, J Biol Chem. 286, 5774-5783. 3. Cabrera R, Ambrosio ALB, Garratt RC, Guixé V & Babul J (2008) Crystallographic Structure of Phosphofructokinase-2 from Escherichia coli in Complex with Two ATP Molecules. Implications for Substrate Inhibition, J Mol Biol. 383, 588-602. 4. Robey EA & Schachman HK (1985) Regeneration of active enzyme by formation of hybrids from inactive derivatives: implications for active sites shared between polypeptide chains of aspartate transcarbamoylase, Proc Natl Acad Sci U S A. 82, 361-365. 5. Ortigosa AD, Kimmel JL & Reinhart GD (2003) Disentangling the Web of Allosteric Communication in a Homotetramer: Heterotropic Inhibition of Phosphofructokinase from Bacillus stearothermophilus†, Biochemistry. 43, 577-586. 6. Fenton AW, Paricharttanakul NM & Reinhart GD (2004) Disentangling the Web of Allosteric Communication in a Homotetramer: Heterotropic Activation in Phosphofructokinase from Escherichia coli†, Biochemistry. 43, 14104-14110. 7. Fenton AW & Reinhart GD (2003) Mechanism of Substrate Inhibition in Escherichia coli Phosphofructokinase†, Biochemistry. 42, 12676-12681. 8. Cabrera R, Babul J & Guixé V (2010) Ribokinase family evolution and the role of conserved residues at the active site of the PfkB subfamily representative, Pfk-2 from Escherichia coli, Arch Biochem Biophys. 502, 23-30. 9. Baez M, Wilson Christian AM, Ramírez-Sarmiento César A, Guixé V & Babul J (2012) Expanded Monomeric Intermediate upon Cold and Heat Unfolding of Phosphofructokinase-2 from Escherichia coli, Biophys J. 103, 2187-2194. 10. Guixé V, Rodríguez PH & Babul J (1998) Ligand-Induced Conformational Transitions in Escherichia coli Phosphofructokinase 2: Evidence for an Allosteric Site for MgATP2- †, Biochemistry. 37, 13269-13275. 11. Baez M, Wilson CAM & Babul J (2011) Folding kinetic pathway of phosphofructokinase-2 from Escherichia coli: A homodimeric enzyme with a complex domain organization, FEBS Lett. 585, 2158-2164. 12. Park S-Y, Quezada CM, Bilwes AM & Crane BR (2004) Subunit Exchange by CheA Histidine Kinases from the Mesophile Escherichia coli and the Thermophile Thermotoga maritima†, Biochemistry. 43, 2228-2240. 13. Fenton AW & Reinhart GD (2002) Isolation of a Single Activating Allosteric Interaction in Phosphofructokinase from Escherichia coli†, Biochemistry. 41, 13410-13416. 14. Kern D & Zuiderweg ERP (2003) The role of dynamics in allosteric regulation, Current Opinion in Structural Biology. 13, 748-757. 15. Cabrera R, Fischer H, Trapani S, Craievich AF, Garratt RC, Guixé V & Babul J (2003) Domain Motions and Quaternary Packing of Phosphofructokinase-2 from Escherichia coli Studied by Small Angle X-ray Scattering and Homology Modeling, Journal of Biological Chemistry. 278, 12913-12919.
AC C
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
23
ACCEPTED MANUSCRIPT
16. Alexandrov V, Lehnert U, Echols N, Milburn D, Engelman D & Gerstein M (2005) Normal modes for predicting protein motions: A comprehensive database assessment and associated Web tool, Protein Science : A Publication of the Protein Society. 14, 633-643. 17. Ma J (2005) Usefulness and Limitations of Normal Mode Analysis in Modeling Dynamics of Biomolecular Complexes, Structure. 13, 373-380. 18. Mitternacht S & Berezovsky IN (2011) Binding Leverage as a Molecular Basis for Allosteric Regulation, PLoS Comput Biol. 7, e1002148. 19. Mitternacht S & Berezovsky IN (2011) Coherent Conformational Degrees of Freedom as a Structural Basis for Allosteric Communication, PLoS Comput Biol. 7, e1002301. 20. Goncearenco A, Mitternacht S, Yong T, Eisenhaber B, Eisenhaber F & Berezovsky IN (2013) SPACER: server for predicting allosteric communication and effects of regulation, Nucleic Acids Research. 41, W266-W272.
RI PT
476 477 478 479 480 481 482 483 484 485 486 487
AC C
EP
TE D
M AN U
SC
488
24
ACCEPTED MANUSCRIPT
The inhibitory mechanism of ATP on Pfk-2 is characterized creating hybrid dimers Homotropic interaction do not play a role in the enzymatic inhibition induced by ATP
AC C
EP
TE D
M AN U
SC
RI PT
Low frequency normal modes support inhibitory heterotropic interactions