Anion binding to a protein–protein complex lacks
dependence on net charge
TRAVIS T. WALDRON, MODESTOS A. MODESTOU,
AND
KENNETH P. MURPHY
University of Iowa, Roy J. and Lucille A. Carver College of Medicine, Department of Biochemistry,
Iowa City, Iowa 52242, USA
(RECEIVED August 27, 2002; FINAL REVISION December 23, 2002; ACCEPTED January 7, 2003)
Abstract
The binding of anions to proteins occurs in numerous physiological and metabolic processes. In an effort
to understand the factors important in these interactions, we have studied the weak binding of phosphate and
sulfate to a protein–protein complex using isothermal titration calorimetry. To our knowledge, this is the
first system in which the thermodynamics of anion binding have been determined calorimetrically. By
studying both phosphate and sulfate binding and using a range of pH values, the charge on the anion was
varied from ∼ −1 to −2. Surprisingly, no dependence of the binding energetics on the charge of the anion
was observed. This result indicates that charge–charge interactions are not the dominant factor in binding
and suggests the importance of hydrogen bonding in specifically recognizing and coordinating anions.
Keywords: Anion; binding; electrostatics; charge–charge; hydrogen bonding
The binding of anions to proteins is critical in numerous
physiological and metabolic processes at all levels of life.
Structural studies show some of the characteristics of phosphate and sulfate binding sites in proteins (Chakrabarti
1993; Copley and Barton 1994), but the energetics of these
binding events remain poorly understood. Knowledge of the
energetics of such interactions is required in order to understand the determinants of binding affinity and specificity
such as charge–charge interactions and hydrogen bonding.
Unfortunately, such binding events are typically very weak
and therefore are below the detection limit of most binding
measurement techniques. These problems contribute to the
lack of information in the literature.
We have characterized the thermodynamics of the interaction between anions and a protein–protein complex.
The 56 amino-acid, serine-protease inhibitor, turkey ovomucoid third domain (OMTKY3), binds with high affinity
(4.1 × 1010; Lu 1994) to the serine protease, porcine pancreatic elastase (PPE). Upon formation of the complex, an
Reprint requests to: Kenneth P. Murphy, Department of Biochemistry,
University of Iowa, 51 Newton Road, Iowa City, IA 52242, USA; e-mail:
[email protected]; fax: (319) 335-9570.
Article and publication are at http://www.proteinscience.org/cgi/doi/
10.1110/ps.0230703.
anion-binding site is formed in the cleft between the two
proteins with residues that coordinate the anion being contributed by both proteins. Previously, we have shown that
phosphate binds to this complex at pH 6 with an unfavorable enthalpy and a favorable entropy (Edgcomb et al.
2000). Here, we investigate the effect of the charge of the
anion on the binding thermodynamics by looking at the
effects of pH, counter-ion, and salt on phosphate binding
and by studying the binding of sulfate. We find that varying
the net charge on the anion from ∼ −1 to −2 has no effect on
the binding energetics.
Results
Figure 1 shows the linkage scheme in which an anion binds
to the complex but not to the free proteins. We have determined that the anions do not bind to either the protease or
the inhibitor alone using differential scanning calorimetry
(DSC) to show that the stabilities of both proteins are independent of anion concentration (data not shown).
The binding of anions to the PPE/OMTKY3 complex was
determined for sodium and potassium phosphate at pH 7,
sodium phosphate at pH 6, and sodium sulfate at pH 6.
Table 1 lists the enthalpies determined for each titration.
The data are depicted graphically in Figure 2 showing bind-
Protein Science (2003), 12:871–874. Published by Cold Spring Harbor Laboratory Press. Copyright © 2003 The Protein Society
871
Waldron et al.
Table 1. Observed enthalpies and standard deviations for
each titration
⌬Hobs, kJ/mole
SD
Potassium Phosphate, pH 7
0
0.05
0.1
0.2
0.25
0.377
0.5
0.75
6.44†
10.3
13.7
20
20
18
21
20
0.6
0.7
1
2
1
1
1
Sodium Phosphate, pH 7
0
0.05
0.1
0.1
0.175
0.25
0.5
6.44†
9.7
15
14.8
17
20
19
0.5
1
0.9
1
1
2
Sodium Phosphate, pH 7 containing 1 M
sodium chloride
0.1
13.4
0.8
Sodium Phosphate, pH 6
0
0.025
0.05
0.1
0.15
0.2
0.25
0.5
0.796
16.24†
18
20.6
23
26.5
26
28
29
32
1
0.7
2
0.7
1
3
2
3
Sodium Sulfate, pH 6
0
0
0.025
0.05
0.05
0.1
0.15
0.25
0.25
0.3
0.4
0.5
0.575
10
8
12.1
20
19
19
21
23
23
23
26
26
26
1
1
0.7
1
2
2
2
1
1
2
2
2
2
[Anion], M
Figure 1. Upon complex formation between the serine protease, porcine
pancreatic elastase, and serine protease inhibitor, turkey ovomucoid third
domain, an anion binding site is created in the cleft between the two
proteins. Anions do not bind to the free proteins.
ing curves for the anions binding to the complex. The curves
are offset as a result of the proton linkage of the system, as
the enthalpy of binding OMTKY3 to PPE is dependent upon
pH and buffer (Baker and Murphy 1997). In the case of the
phosphate experiments, the phosphate itself is buffering,
while in the case of sulfate, PIPES is used because sulfate
has no buffering capacity at pH 6.
By fitting the data in Table 1 to Equation 1 (see Materials
and Methods), we obtain the values for the binding constants and enthalpies given in Table 2. Within error, there is
no difference in binding affinity or enthalpy for any of the
four cases.
When 1 M sodium chloride is added to the 100 mM
sodium phosphate titration at pH 7, there is no effect on the
observed enthalpy (shown as an “X” in Fig. 2). The intrinsic
binding of PPE-OMTKY3 is also NaCl independent (Baker
and Murphy 1997). This indicates that phosphate binding is
also salt independent.
Discussion
The results presented here indicate that the charge–charge
interaction between the anion and the protein is not driving
binding. The identity of the protein groups coordinating the
anions was discussed previously (Edgcomb et al. 2000) and
are from a modeled structure of OMTKY3/PPE based on
homology with a structure of OMTKY3/SGPB (Huang et
al. 1995; Fig. 3). The OMTKY3/SGPB structure contains an
anion coordinated by residues from both proteins, tyrosine
32 and arginine 41 from SGPB, and tyrosine 20 and lysine
872
Protein Science, vol. 12
The buffering conditions for each set of titrations are given. In the case of
sodium sulfate, the solution also contained 50 mM PIPES.
† Extrapolated from the data of Baker and Murphy (1997).
55 from OMTKY3. By homology, tyrosine 35 from PPE is
involved in addition to the tyrosine and lysine of OMTKY3.
The following lines of evidence suggest the charge–
charge interaction between the lysine of OMTKY3 and the
anion is relatively unimportant. We have altered the net
charge on phosphate from ∼ −1.1 to −1.7 by varying the pH.
The net charge was changed to −2 by changing the identity
Anion binding to a protein–protein complex
Figure 2. Binding data for complex formation in the presence of various
concentrations of anions. Lines indicate the least squares fit to Equation 1.
Error bars represent the standard deviation of the observed presaturation
injection heats for a given titration. Titrations under various conditions are
as follows: potassium phosphate pH 7 (red circles), sodium phosphate pH
7 (blue squares), sodium phosphate pH 6 (dark green diamonds), sodium
sulfate pH 6 (light green triangles), and sodium phosphate titration at pH
7 with 1 M NaCl (box with “X”).
of the anion from phosphate to sulfate. The difference in
charge affects the charge density, as phosphate and sulfate
have radii that are very close to one another. Further, we
have changed the identity of the counter ion from sodium to
potassium. It was noticed that the relevant pKa of phosphate
is slightly altered depending on the identity of the counter
ion. None of these changes had any effect on the energetics
of binding. If charge was a dominant aspect of the binding,
we would expect these changes to significantly alter the
binding affinity. This lack of dependence on charge–charge
interactions is further confirmed by the addition of 1 M
NaCl to the titrations. While we would expect NaCl to
screen attraction between oppositely charged groups, no difference is observed.
The lack of charge dependence for this interaction indicates that other factors are responsible for the binding. ConTable 2. Binding constants and enthalpies of binding for anions
binding to the PPE-OMTKY3 complex at 25°C
Potassium Phosphate, pH 7
Sodium Phosphate, pH 7
Sodium Phosphate, pH 6
Sodium Sulfate, pH 6
Kanion
⌬Hanion,
kJ/mole
T⌬S
kJ/K/mole
⌬G
kJ/mole
7±3
9±5
10 ± 5
17 ± 7
18 ± 2
17 ± 4
17 ± 3
18 ± 2
23 ± 2
22 ± 4
23 ± 3
25 ± 2
−5 ± 1
−5 ± 1
−6 ± 1
−7 ± 1
Numbers are fits to data in Table 1 using Equation 1.
comitant with the binding of a charged group is the release
of water (desolvation). The positive ⌬H we observe is consistent with a considerable enthalpic cost for desolvating a
charged ion. Although enthalpically unfavorable, desolvation produces a large entropy gain for solvent release. We
find that binding of phosphate and sulfate to the protein–
protein complex is indeed entropically driven.
Extensive work by Luecke and Quiocho (1990) has demonstrated the importance of hydrogen bonding in phosphate
binding to phosphate binding protein (PBP). The crystal
structure of PBP shows extensive hydrogen bonding networks coordinating the anion (Luecke and Quiocho 1990).
The same group has also shown that the negatively charged
phosphate binds tightly in a pocket of negative electrostatic
potential indicating that, over shorter distances, other factors begin to dominate affinities (Ledvina et al. 1998). Surveys that examine the coordination of phosphate and sulfate
ions in proteins for which crystal structures are available
(Chakrabarti 1993; Copley and Barton 1994) indicate that
anion binding sites display partial charge compensation.
This suggests that dipolar interactions are important for coordination of anions.
The energetics we have measured for anions binding to a
protein–protein complex lead to the conclusion that charge–
charge interactions are not major determinants modulating
binding. Consequently, optimal hydrogen bonding and local
dipolar interactions likely are important factors controlling
affinity and specificity.
Materials and methods
Materials
High-purity PPE was purchased from Elastin Products Company
and used without further purification. It was found that the small
impurities present would be cleaved by PPE and diffuse away in
the course of the dialysis, without noticeable loss or degradation of
PPE itself, as determined by SDS-PAGE. OMTKY3 was purified
from turkey egg white trypsin inhibitor (Sigma) as described elsewhere (Swint and Robertson 1993). The final steps of the purification are extensive dialysis against water, followed by lyophilization. All other salts, acids, and bases were of the highest purity
available from either Amresco, Sigma, or Fischer.
Isothermal titration calorimetry
Titrations were performed on a Calorimetry Sciences Corporation
(CSC) isothermal titration calorimeter (ITC), Model 4200.
Samples were prepared by overnight dialysis (MWCO 3500) at
4°C of both inhibitor (800 ␮L of OMTKY3) and enzyme (3 mL of
PPE) against 4 L of the same buffer in which they were dissolved.
The details of the buffering system used for each experiment are
given in Table 1. The concentration of PPE in the cell was ∼100
␮M and the concentration of OMTKY3 in the syringe was ∼1 mM.
Both proteins were in the same concentration of anion (phosphate
or sulfate), or anion and NaCl.
www.proteinscience.org
873
Waldron et al.
Figure 3. (A) The cleft between Streptomyces griseus protease B (SGPB) and OMKTY3. Residues coordinating the phosphate are tyrosine 32 and arginine
41 from SGPB, and tyrosine 20 and lysine 55 from turkey ovomucoid third domain (OMTKY3). (B) The view from the homology-modeled structure of
porcine pancreatic elastase (PPE) and OMTKY3. Tyrosine 35 of PPE is in a homologous position to tyrosine 32 of SGPB. Ribbons of the proteases are
on the left side of each panel, and ribbons from OMTKY3 are on the right side of each panel.
Because the binding of OMTKY3 to PPE is stoichiometric, the
binding enthalpy is equal to the heat for a presaturation injection
(i.e., the area under one peak) divided by the moles of titrant
injected. The average of the binding enthalpies from each presaturation injection, excluding the first, was taken as the observed ⌬H
of binding. Extinction coefficients of 4400 M−1cm−1 (Swint and
Robertson 1993) and 5.2 × 104 M−1cm−1 (Shotton 1970) were used
for OMTKY3 and PPE, respectively, for concentration determinations. The stoichiometry, as determined using Bindworks version
3.0 (distributed by CSC), was close to one for all experiments.
Determining ⌬Hanion and Kanion
For a system in which the anion binds to the complex, but not to
the free proteins, the observed binding enthalpy will include the
enthalpy of the protein–protein interaction (⌬H0), plus the molar
enthalpy of anion binding multiplied by the fraction of complex to
which anion is bound (Edgcomb et al. 2000). Accordingly, the
observed heats as a function of anion concentration were fit using
NONLIN (Michael Johnson, University of Virginia) to the following equation:
⌬Hobs = ⌬H0 + ⌬Hanion
Kanion关anion兴
共1 + Kanion关anion兴兲
( 1)
where ⌬H0 is the enthalpy of binding in the absence of anion, and
⌬Hanion and Kanion are the binding enthalpy and association constant for anion binding to the complex (Edgcomb et al. 2000).
Acknowledgments
This work was supported by grant MCB-9808073 from the National Science Foundation.
874
Protein Science, vol. 12
The publication costs of this article were defrayed in part by
payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 USC section 1734
solely to indicate this fact.
References
Baker, B.M. and Murphy, K.P. 1997. Dissecting the energetics of a protein–
protein interaction: The binding of ovomucoid third domain to elastase. J.
Mol. Biol. 268: 557–569.
Chakrabarti, P. 1993. Anion binding sites in protein structures. J. Mol. Biol.
234: 463–482.
Copley, R.R. and Barton, G.J. 1994. A structural analysis of phosphate and
sulphate binding sites in proteins. Estimation of propensities for binding and
conservation of phosphate binding sites. J. Mol. Biol. 242: 321–329.
Edgcomb, S.P., Baker, B.M., and Murphy, K.P. 2000. The energetics of phosphate binding to a protein complex. Protein Sci. 9: 927–933.
Huang, K., Lu, W., Anderson, S., Laskowski, Jr., M., and James, M.N. 1995.
Water molecules participate in proteinase-inhibitor interactions: Crystal
structures of Leu18, Ala18, and Gly18 variants of turkey ovomucoid inhibitor third domain complexed with Streptomyces griseus proteinase B.
Protein Sci. 4: 1985–1997.
Ledvina, P.S., Tsai, A.L., Wang, Z., Koehl, E., and Quiocho, F.A. 1998. Dominant role of local dipolar interactions in phosphate binding to a receptor cleft
with an electronegative charge surface: Equilibrium, kinetic, and crystallographic studies. Protein Sci. 7: 2550–2559.
Lu, W. 1994. “Energetics of the interactions of ovomucoid third domain variants
with different serine proteinases.” Ph.D. thesis, Purdue University, West
Lafayette, IN.
Luecke, H. and Quiocho, F.A. 1990. High specificity of a phosphate transport
protein determined by hydrogen bonds. Nature 347: 402–406.
Shotton, D.M. 1970. Elastase. Methods Enzymol. 19: 113–140.
Swint, L. and Robertson, A.D. 1993. Thermodynamics of unfolding for turkey
ovomucoid third domain: Thermal and chemical denaturation. Protein Sci.
2: 2037–2049.