HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Peterson, K. M. Articles by Srivastava, D.K. Search for Related Content PubMed PubMed Citation Articles by Peterson, K. M. Articles by Srivastava, D.K. Social Bookmarking                   What's this?

Influence of Glu-376 Gln mutation on enthalpy and heat capacity changes for the binding of slightly altered ligands to medium chain acyl-CoA dehydrogenase

Department of Biochemistry and Molecular Biology, North Dakota State University, Fargo, North Dakota 58105, USA

Reprint requests to: Dr. D.K. Srivastava, Department of Biochemistry and Molecular Biology, North Dakota State University, Fargo, North Dakota 58105, USA; e-mail: DK_Srivastava{at}ndsu.nodak.edu; fax: (701) 231-9657.

(RECEIVED December 13, 2000; FINAL REVISION May 4, 2001; ACCEPTED June 7, 2001)

¹ Present address: University of Konstanz, Faculty of Biology, M644, D-78457 Konstanz, Germany.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1101/

	Abstract

TOP Abstract Introduction Results Discussion Materials and methods References

 
We showed that the -CH₂ NH substitution in octanoyl-CoA alters the ground and transition state energies for the binding of the CoA ligands to medium-chain acyl-CoA dehydrogenase (MCAD), and such an effect is caused by a small electrostatic difference between the ligands. To ascertain the extent that the electrostatic contribution of the ligand structure and/or the enzyme site environment modulates the thermodynamics of the enzyme–ligand interaction, we undertook comparative microcalorimetric studies for the binding of 2-azaoctanoyl-CoA (-CH₂ NH substituted octanoyl-CoA) and octenoyl-CoA to the wild-type and Glu-376 Gln mutant enzymes. The experimental data revealed that both enthalpy (H°) and heat capacity changes (C_p°) for the binding of 2-azaoctanoyl-CoA (H°₂₉₈ = -21.7 ± 0.8 kcal/mole, C_p° = -0.627 ± 0.04 kcal/mole/K) to the wild-type MCAD were more negative than those obtained for the binding of octenoyl-CoA (H°₂₉₈ = -17.2 ± 1.6 kcal/mole, C_p° = -0.526 ± 0.03 kcal/mole/K). Of these, the decrease in the magnitude of C_p° for the binding of 2-azaoctanoyl-CoA (vis-à-vis octenoyl-CoA) to the enzyme was unexpected, because the former ligand could be envisaged to be more polar than the latter. To our further surprise, the ligand-dependent discrimination in the above parameters was completely abolished on Glu-376 Gln mutation of the enzyme. Both H° and C_p° values for the binding of 2-azaoctanoyl-CoA (H°₂₉₈ = -13.3 ± 0.6 kcal/mole, C_p° = -0.511 ± 0.03 kcal/mole/K) to the E376Q mutant enzyme were found to be correspondingly identical to those obtained for the binding of octenoyl-CoA (H°₂₉₈ = -13.2 ± 0.6 kcal/mole, C_p° = -0.520 ± 0.02 kcal/mole/K). However, in neither case could the experimentally determined C_p° values be predicted on the basis of the changes in the water accessible surface areas of the enzyme and ligand species. Arguments are presented that the origin of the above thermodynamic differences lies in solvent reorganization and water-mediated electrostatic interaction between ligands and enzyme site groups, and such interactions are intrinsic to the molecular basis of the enzyme–ligand complementarity.

Keywords: Medium chain acyl-CoA dehydrogenase; isothermal microcalorimetry; thermodynamics of ligand binding; Glu-376 Gln mutation; 2-azaoctanoyl-CoA; octenoyl-CoA

Abbreviations: MCAD, human liver medium-chain acyl-CoA dehydrogenase • FAD, flavin adenine dinucleotide • FcPF₆, ferricenium hexafluoro phosphate • G°, standard free energy change • H°, standard enthalpy change • S°, standard entropy change • K_a, association constant • C_p°, standard heat capacity change • OcaCoA, octanoyl-CoA • OceCoA, 2-octenoyl-CoA • aza-CoA, 2-azaoctanoyl-CoA • THF, tetrahydrofuran • EDTA, ethylenediamine tetraacetic acid • 3',5'-ADP, 3',5'-adenosine diphosphate • C_p,p, heat capacity for changes in the polar surface areas • C_p,np, heat capacity for changes in the nonpolar surface areas • A_p, changes in the water accessible polar surface areas • A_np, changes in the water accessible nonpolar surface areas

	Introduction

TOP Abstract Introduction Results Discussion Materials and methods References

 
Medium-chain acyl-CoA dehydrogenase (MCAD) shows a wide range of substrate specificity, and it interacts with a variety of differently substituted CoA-derivatives (for reviews, see Beinert 1963; Engel 1990; Thorpe and Kim 1995). This feature of the enzyme provided an opportunity to investigate the influence of a small variation in the ligand structure on the kinetic and thermodynamic properties of the enzyme–ligand complexes, as well as during the enzyme catalysis (Engel 1990; Srivastava et al. 1997; Peterson and Srivastava 1997; Peterson et al. 1998b; Vock et al. 1998; Cummings and Thorpe 1994; Powell et al. 1987; McFarland et al. 1982). In the past, a few such studies involved the influence of deletion of distal fragments (e.g., 3'-phosphate and 3',5'-ADP) of the coenzyme-A structure (Srivastava et al. 1995; Peterson and Srivastava 1997; Peterson et al. 1998b), utilization of increasing chain lengths of acyl/enoyl-CoAs (Beinert 1963; Engel 1990; Peterson et al. 1995; Kumar 1997; Nandy et al. 1996), and the effect of hetero-atom substitution in the fatty acid regions of CoA-ligands on the spectroscopic, kinetic, and thermodynamic properties of the enzyme (Engel 1990; Thorpe and Kim 1995; Cummings and Thorpe 1994; Powell et al. 1987; Trievel et al. 1995). These studies have sometimes been corroborated, but at other times opposed by the known structural features of the enzyme–ligand complexes (Peterson and Srivastava 1997; Peterson et al. 1998; Kim et al. 1993; Lee et al. 1996).

The X-ray crystallographic structures of both pig liver and recombinant human liver enzymes in the absence and presence of bound product (octenoyl-CoA) have been solved to atomic resolutions (Kim et al. 1993; Lee et al. 1996; Kim and Wu 1988). The structural data reveal that the ligand-binding pocket of the enzyme traverses almost to the center of the protein structure, and is predominantly comprised of hydrophobic amino acid residues (Kim et al. 1993). The binding of octenoyl-CoA results in exclusion of at least four structured water molecules from the enzyme site (Kim et al. 1993). These features are corroborated by the detailed microcalorimetric studies for the binding of octenoyl-CoA to both pig kidney and recombinant human liver MCADs (Srivastava et al. 1997; Peterson et al. 1998b). Consistent with the principle of the classic hydrophobic effect (Tanford 1980), the binding of octenoyl-CoA (a predominantly hydrophobic ligand) to the hydrophobic enzyme site pockets of pig kidney and human liver enzymes yields large negative C_p° values, albeit of substantially different magnitudes (Srivastava et al. 1997; Peterson et al. 1998b).

We recently undertook a comparative transient kinetic investigation for the binding of octenoyl-CoA and 2-azaoctanoyl-CoA (a redox inactive ligand in which the -CH group is replaced by NH) via the stopped-flow method (Peterson et al. 2000). Such studies revealed that although the above substitution did not influence the overall binding affinity of the ligands to the enzyme, it influenced the microscopic steps during the overall binding process (Peterson et al. 2000). The binding of these as well as other ligands to the enzyme was found to proceed via two steps (Johnson et al. 1992; Kumar and Srivastava 1994,1995; Srivastava et al. 1995): The first (fast) step involved the formation of the enzyme–ligand collision/Michaelis complex, followed by the isomerization of the latter complex during the second (slow) step. The latter step was found to be coupled to the protein conformational change:

(1)

A comparative account of the binding studies for the above ligands led to the proposition that the electrostatic field involving the carboxyl group of Glu-376 plays an important role in stabilizing the ground and transition state structures during the isomerization steps of the cognate enzyme–ligand complexes (Peterson et al. 2000). With these features in mind, we proceeded to determine the thermodynamic consequences of -CH₂ NH substitution in octanoyl-CoA on binding of ligands to the wild-type and E376Q mutant enzymes. The results and conclusion derived from such studies are contained in the following sections.

	Results

TOP Abstract Introduction Results Discussion Materials and methods References

 
Figure 1 shows the isothermal microcalorimetric data for the titration of wild-type MCAD with increasing concentrations of 2-azaoctanoyl-CoA in 50 mM phosphate buffer at pH 7.6, containing 0.3 mM EDTA and 10% glycerol. The top panel shows the effect of injecting 4 µL aliquots (the first aliquot being 1 µL) of 205 µM stock solution of 2-azaoctanoyl-CoA into 1.8 mL of 10.1 µM solution of wild-type MCAD. The addition of each aliquot follows the emergence of a negative (exothermic) peak because of the enzyme–ligand interaction. The area under each peak serves as a measure of the amount of heat produced on ligand binding. Note that as the titration progresses, the area under the peak becomes smaller because of increased occupancy of the enzyme site by 2-azaoctanoyl-CoA. The bottom panel of Figure 1 shows the plot of the amount of heat generated per injection as a function of the molar ratio of 2-azaoctanoyl-CoA to the wild-type MCAD. The solid smooth line is the best fit of the experimental data according to Wiseman et al. (1989) for the values of stoichiometry (n), the association constant (K_a), and the standard enthalpic changes (H°) of the enzyme-2-azaoctanoyl-CoA complex of 0.88 ± 0.01 (moles of bound ligand per mole of MCAD subunit), (3.1 ± 0.2) x 10⁶ M^-1, and -21.1 ± 0.2 kcal/mole, respectively. These values are similar to the preliminary microcalorimetric data of Peterson et al. (2000). The magnitude of K_a, discerned from the data of Figure 1, is translated to the standard free energy changes (G° = -RT ln K_a) of -8.9 kcal/mole, on consideration of the standard state being equal to 1 M. Given the enthalpy (H°) and free energy (G°) changes for the binding of 2-azaoctanoyl-CoA to the enzyme, the entropic contribution (TS°) for the above process can be calculated (TS° = -G° + H°) to be -12.2 kcal/mole. It should be emphasized that the titration experiment of Figure 1 was performed in triplicate, and the average values of n, G°, H°, and TS° were discerned to be 0.85 ± 0.04, -8.7 ± 0.2 kcal/mole, -21.7 ± 0.2 kcal/mole, and -12.9 ± 0.9 kcal/mole, respectively. The microcalorimetric titration data for the binding of octenoyl-CoA to MCAD yielded the average values of n, G°, H°, and TS° to be 0.98 ± 0.06, -8.3 ± 0.2 kcal/mole, -17.2 ± 1.6 kcal/mole, and -8.9 ± 1.8 kcal/mole, respectively (Peterson et al. 1998). A comparative account of these data reveals that whereas the magnitudes of n and G° for the binding of 2-azaoctanoyl-CoA are similar to those obtained for the binding of octenoyl-CoA, the H° and TS° values are considerably different (Peterson et al. 2000). Of the latter parameters, the H° and TS° values for the binding of 2-azaoctanoyl-CoA to MCAD are 4.5 and 4.1 kcal/mole more negative, respectively, (i.e., enthalpically more favorable, but entropically less favorable) than those obtained for the binding of octenoyl-CoA to the enzyme.

View larger version (22K): [in this window] [in a new window]   Fig. 1. Isothermal microcalorimetric titration data for the binding of 2-azaoctanoyl-CoA with wild-type MCAD at 25°C. The top panel shows the raw calorimetric data generated by titration of 1.8 mL of 10.13 µM wild-type enzyme with 60 injections (first injection: 1 µL, subsequent injections: 4 µL each) of 205.1 µM 2-azaoctanoyl-CoA. The area under each peak was integrated and plotted against the molar ratio of 2-azaoctanoyl-CoA to the enzyme in the bottom panel. The solid smooth line is the best fit of the data for the stoichiometry (n) of the enzyme-2-azaoctanoyl-CoA complex (moles of bound 2-azaoctanoyl-CoA per mole of the enzyme subunit) of 0.88 ± 0.01, association constant (Ka) of (3.1 ± 0.21) x 106 M-1, and the standard enthalpic changes (H°) of -21.1 ± 0.2 kcal/mole.

Effect of Glu-376 Gln (E376Q) mutation

Temperature dependence of the thermodynamic parameters
We investigated the temperature dependence of the thermodynamic parameters for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to the wild-type and E376Q mutant enzymes. All these experiments were repeated two to five times, often using different batches of enzyme and ligand preparations, and the standard deviations have been calculated from the average of replicate experiments (Table 1). A casual look at the experimental data of Table 1 reveals that, except for a few instances, the stoichiometry of the enzyme–ligand complexes is close to unity. We believe the deviation in stoichiometry from unity is attributable to our inability to determine precisely the "intact" active site concentrations of the enzymes from different batches. We noted that if the experimental data were reanalyzed after adjusting the concentrations of the enzyme (based on the observed stoichiometries of the enzyme–ligand complexes from initial analyses), the derived thermodynamic parameters were not altered significantly. Such parameters were, however, altered when we deliberately fixed the stoichiometry of the enzyme–ligand complexes (without adjusting the enzyme concentrations) as being equal to 1. Hence, we opted to analyze all the microcalorimetric titration data presented herein as such without any preconceived bias.

View larger version (13K): [in this window] [in a new window]   Fig. 2. Temperature dependence of H° for the interactions of 2-azaoctanoyl-CoA and octenoyl-CoA with wild-type MCAD. The experimental data for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA are represented by the closed and open circles, respectively. The solid smooth lines are the linear regression analysis of the experimental data for slope (Cp°) of -0.627 ± 0.050 kcal mole-1 K-1 and -0.526 ± 0.03 kcal mole-1 K-1, and intercept (at 0.0 K) of 167 ± 14 kcal/mole and 140 ± 9 kcal/mole for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA, respectively.

View larger version (12K): [in this window] [in a new window]   Fig. 3. Temperature dependence of H° for the interaction of 2-azaoctanoyl-CoA and octenoyl-CoA with E376Q mutant MCAD. The experimental data for 2-azaoctanoyl-CoA and octenoyl-CoA are represented by the closed and open circles, respectively. The solid smooth lines are the linear regression analysis of the experimental data for slope (Cp°) of -0.511 ± 0.03 kcal mole-1 K-1 and -0.520 ± 0.02 kcal mole-1 K-1, and intercept (at 0.0 K) of 139 ± 10 kcal/mole and 142 ± 5.6 kcal/mole for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA, respectively.

From the thermodynamic data in Table 1, it should be noted further that irrespective of the type of ligand or enzyme, the temperature-dependent decrease in H° is compensated by the corresponding increase in the TS° value, such that the G° practically remains constant over the entire temperature range. Clearly, like other enzyme–ligand interacting systems (Herron et al. 1986; Jin et al. 1993; Mukkur 1978; Sigurskjold et al. 1994), the enthalpy-entropy compensation effect is prevalent for the binding of both 2-azaoctanoyl-CoA and octenoyl-CoA to the wild-type and E376Q mutant enzymes, and its origin may lie in the weak interaction (presumably mediated via hydrogen bonding) between the enzyme–ligand species (Dunitz 1995).

Model building studies and the changes in the water accessible surface areas on enzyme–ligand interactions
The molecular model building studies were performed by using the coordinates of the dimeric form of pig liver MCAD (obtained from the Brookhaven Protein Data Bank) in the absence (file name: pdb3mdd.pdb, to be referred herein as the apo enzyme) and the presence of C₈CoA (file name: 3pdbmde.pdb the holo enzyme). A casual look at these coordinate files revealed that the number of crystallographically identifiable water molecules in the dimeric forms of the apo and holo structures were 166 and 182, respectively. Evidently, the binding of C₈CoA to the apo enzyme site accompanies incorporation of 16 additional water molecules, and most of these water molecules were confined either to the subunit contact region or to the coenzyme-A binding region. If the above information is realistic (particularly because many "mobile" water molecules remain undetected by the X-ray crystallographic technique), it would strengthen our overall thermodynamic conclusion (see Discussion).

Figure 4 shows the location of water molecules within a 10 Å perimeter of the C₈CoA binding site in both apo and holo structures of the enzyme. This was accomplished by superimposition of the backbones of the above protein structures followed by interfacing the water molecules around the holo-enzyme bound C₈CoA via the aid of the Insight-II (98) software. Selected water molecules from the apo (light circle) and holo (dark circle) enzyme structures are labeled as A and H, respectively (Fig. 4), following their residue numbers (found in the respective coordinate files). Note that whereas most of the water molecules occupy the same or similar positions in both apo and holo structures, some water molecules are displaced from their original position in the apo structure to a new position in the holo structure. In Figure 4, the water molecule 803H appears to be a new water molecule in the holo structure of the enzyme. It is noteworthy that four water molecules, 802A to 805A, present in the apo structure (in place of the tail end region of C₈CoA) is removed on binding of the ligand (Kim et al. 1993). Because no high resolution X-ray crystallographic data are available for either wild-type enzyme-2-azaoctanoyl-CoA complex or of E376Q mutant enzyme in the absence or presence of any of the ligands used herein, it would be difficult to predict, a priori, the distribution of water molecules in these unknown enzyme–ligand structures. However, on energy minimization of different enzyme–ligand complexes (see below), it has been noticed that the water molecule, 803H, appear to move (1 Å) closer to the amide nitrogen of 2-azaoctanoyl-CoA vis-à-vis its position in the octenoyl-CoA bound structure of the wild-type enzyme (data not shown). Based on these cursory observations, it is tempting to speculate that in the case of the wild-type enzyme-2-azaoctanoyl-CoA complex, a water molecule (possibly 803H) is involved in the hydrogen bonding between the carboxyl group of Glu-376 and the amide nitrogen of the ligand (see Discussion).

View larger version (28K): [in this window] [in a new window]   Fig. 4. Location of water molecules within 10 Å perimeter of the enzyme-bound C8CoA. The water molecules present in the pdb3mde.pdb (holo structure) and pdb3mdd.pdb (apo structure) data files are shown by the light and dark circles, respectively. Some of the water molecules have been denoted by their residue numbers. The suffixes -A and -H represent the water molecules of the apo and holo structures, respectively.

View larger version (27K): [in this window] [in a new window]   Fig. 5. Spatial relationships among selected residues in the energy-minimized structures of the wild-type MCAD–ligand complexes. The relative positions of Glu-376 and the enzyme-bound FAD and CoA-ligands (2-azaoctanoyl-CoA and octenoyl-CoA) are shown. Note that the 2'-ribityl (hydroxyl) oxygen and the isoalloxazine ring of FAD move toward the CoA-ligand in the case of the enzyme-2-azaoctanoyl-CoA vis-à-vis the enzyme-octenoyl-CoA complex. Besides, the -NH group of 2-azaoctanoyl-CoA moves toward the carboxyl oxygens (OE1 and OE2) of Glu-376 as compared to the -CH group of octenoyl-CoA.

in which C_p,p and C_p,np are the heat capacity functions for the changes in the polar and nonpolar surface areas, respectively. The magnitudes of C_p,p and C_p,np have been proposed to be -0.14 and 0.32 cal/mole/K/Å, respectively, by Spolar et al (1992), and -0.26 and 0.45 cal/mole/K/Å, respectively, by Murphy et al (1993).

View this table: [in this window] [in a new window]   Table 2. Changes in the solvent-accessible surface areas upon binding of 2-azaoctanoyl-CoA and octenoyl-CoA to pig liver wild-type and E376Q mutant MCADs, and the predicted Cp° values

	Discussion

TOP Abstract Introduction Results Discussion Materials and methods References

 
The experimental data presented in the previous section lead to the following conclusions: (1) The binding affinities of both wild-type and Glu-376 Gln (E376Q) mutant medium-chain acyl-CoA dehydrogenases (MCAD) for 2-azaoctanoyl-CoA and octenoyl-CoA are essentially identical. Hence, neither the -CH NH substitution in C₈-CoA nor the removal of a net negative charge from the enzyme site residue Glu-376 (i.e., via E376Q mutation) influences the G° value of the enzyme–ligand complexes. (2) The H° value for the binding of 2-azaoctanoyl-CoA to the wild-type enzyme is 4.5 kcal/mole more negative (favorable) than that for the binding of octenoyl-CoA, suggesting that the favorable enthalpic contribution for the binding of 2-azaoctanoyl-CoA is offset by the unfavorable entropic contribution by an equal magnitude. (3) The discriminatory feature in H° for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to the enzyme is abolished on E376Q mutation; both these ligands yield H° of -13 kcal/mole on binding to the E376Q mutant enzyme site. (4) Whereas the heat capacity change (C_p°) for the binding of 2-azaoctanoyl-CoA to the wild-type enzyme is 101 cal/mole/K more negative than that of octenoyl-CoA, the C_p° values for the binding of the above ligands to the E376Q mutant enzyme are essentially identical (i.e., -0.52 kcal/mole/K). However, in neither case could the experimentally determined C_p° value be predicted on the basis of the changes in the water-accessible polar and nonpolar surface areas.

The comparative microcalorimetry-based ligand binding studies presented herein emphasize the importance of detailed thermodynamic evaluations to discern the molecular basis of enzyme–ligand interactions. The physical forces that dictate enzyme–ligand complementarity remains elusive if only binding constants of the enzyme–ligand complexes are determined. The fact that the G° values for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to the wild-type and E376Q mutant enzymes are equal provides no information about the influence of either -CH NH substitution or the elimination of the negative charge of Glu-376 in stabilizing the corresponding enzyme–ligand complexes. The role of the above variants starts emerging on comparison of the H° and C_p° values of the enzyme–ligand complexes.

Because the H° value for the binding of 2-azaoctanoyl-CoA to the wild-type enzyme is 4.5 kcal/mole more negative (favorable) than that of octenoyl-CoA, and the above parameter remains invariant for the binding of both these ligands to the E376Q mutant enzyme ( -13 kcal/mole), it implies that the coulombic interaction in part plays an important role in stabilizing the enzyme–ligand complexes (Gilson and Honig 1988; Xiao and Honig 1999). The structural, spectroscopic, and model building data all suggest that the carbonyl groups of octenoyl-CoA and 2-azaoctanoyl-CoA are polarized on binding to the enzyme sites, resulting in generation of an initial partial positive charge at the carbonyl carbon, which can be delocalized to the neighboring atoms (Engel 1990; Thorpe and Kim 1995; Kim et al. 1993; Srivastava and Peterson 1998; Johnson et al. 1995; Nishina et al. 1992). In the case of 2-azaoctanoyl-CoA, the partial positive charge can be stabilized at the -NH center, but in the case of octenoyl-CoA, it can at best be stabilized at the carbon center. Given that both carboxyl oxygens (i.e., OE1 and OE2) of Glu-376 are closer to the than the centers (Fig. 5), the columbic interaction between the negatively charged carboxyl group of Glu-376 and the partial positive charges of the above ligands is likely to be more pronounced in the case of 2-azaoctanoyl-CoA as compared to octenoyl-CoA. This may be responsible for the 4.5 kcal/mole enthalpic advantage for the binding of 2-azaoctanoyl-CoA (vis-à-vis octenoyl-CoA) to the wild-type enzyme. Because Glu-376 Gln mutation eliminates the net negative charge of the carboxyl group of Glu-376, the discriminatory contribution of H° for the binding of the above ligands to the enzyme is lost, yielding a nearly identical H° ( -13 kcal/mole).

To substantiate or refute the above simplistic explanation, we considered the possibility of whether or not the higher enthalpic changes during the course of titration of the wild-type enzyme with 2-azaoctanoyl-CoA was because of slow oxidase activity (Johnson et al. 1992; Johnson et al. 1994; McFarland et al. 1982) of the enzyme. Such a possibility seemed likely because we observed previously that the titration of human liver enzyme with octanoyl-CoA yielded about 10 kcal/mole more negative H° value than that obtained with octenoyl-CoA (Srivastava et al. 1997). Furthermore, the experimental factor (e.g., the substrate type), which enhanced the rate of the oxidase reaction of MCAD, was also found to yield a more negative H° value. This was further supported by an independent demonstration that the yeast acyl-CoA oxidase catalyzed conversion of hexanoyl-CoA to hexenoyl-CoA, coupled with the reduction of O₂ to H₂O₂, was an exothermic process (unpubl. results). If 2-azaoctanoyl-CoA underwent a slow oxidase reaction during the course of the enzyme–ligand titration, it would yield a more negative H° value, and the magnitude of the latter would increase with an increase in temperature, yielding a higher C_p° value (as observed experimentally). To probe this possibility, we measured the amount of H₂O₂ (Gopalan and Srivastava 1997) formed during the time regime of the microcalorimetric titration of MCAD by 2-azaoctanoyl-CoA. The experimental data provided no evidence of formation of H₂O₂ on incubation of 2-azaoctanoyl-CoA with either wild-type or E376Q mutant enzyme for up to 3 to 4 hr. Hence, the enthalpic contribution for the binding of 2-azaoctanoyl-CoA is not because of some unforeseen artifact of the enzyme-catalyzed oxidase reaction.

A quantitative evaluation of the G° and H° values for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to the wild-type and E376Q mutant enzymes reveals that the binding of these ligands is primarily enthalpically driven. Depending on the type of ligand, the favorable enthalpic contribution is offset by unfavorable entropic contributions of different magnitudes, yielding nearly identical G° values for all the enzyme–ligand complexes. As discussed by Dunitz (1995), such an enthalpy-entropy compensation is a common property for the binding of ligands to proteins (Herron et al. 1986; Jin et al. 1993; Mukkur 1978; Sigurskjold et al. 1994; Srivastava et al. 1997; Peterson et al. 1998b; Qin and Srivastava 1998), and its prevalence is believed to be dictated by weak interactions (often mediated via the hydrogen bonding) between the cognate species (Dunitz 1995).

It is evident that the entropic loss for the binding of 2-azaoctanoyl-CoA to the wild-type enzyme is maximum (i.e., -12.9 kcal/mole), followed by that for the binding of octenoyl-CoA to the wild-type enzyme (-8.9 kcal/mole), and it is minimum for the binding of these ligands (with equal magnitudes) to the E376Q mutant enzyme. At this point, without taking into account of the C_p° values, the compensatory entropic contributions (for the individual enzyme–ligand complexes) can be attributed to the loss in rotational and translational freedoms of the interacting species. However, such a simplistic explanation fails to account for the temperature dependence of enthalpic changes and the relative magnitudes of the C_p° values for the individual enzyme–ligand complexes (see below).

Following the proposal of Edsall (1935) that the solvation /desolvation of nonpolar molecules in aqueous phase is the major contributor of the heat capacity changes (C_p°), several investigators provided molecular interpretation of the hydrophobic effect for several processes such as partitioning of nonpolar molecules between organic and aqueous phases, protein unfolding, binding of ligands to proteins, and others (Tanford 1980; Privalov 1979; Spolar et al. 1992; Murphy et al. 1993; Makhatadze and Privalov 1990, 1993). A large negative C_p° value for the protein–ligand interaction has been attributed to the predominance of the hydrophobic (yielding negative C_p°) over the polar (yielding positive C_p°) interactions (Livingstone et al. 1991; Spolar et al. 1989; Privalov 1979; Makhatadze and Privalov 1990, 1993). According to the classic view of the hydrophobic effect involving the enzyme–ligand interaction, the water molecules surrounding the hydrophobic region of the ligand, as well as those resident within the enzyme site pocket, are stripped out/excluded to the exterior bulk phase. These water molecules have low entropy and high heat capacity, and thus the release of such water molecules on enzyme–ligand interaction results in an increase in entropy and a decrease in heat capacity (Tanford 1980; Dill 1990; Rowe et al. 1998).

The magnitude of C_p° for protein unfolding, as well as in selected cases of enzyme–ligand interactions, has been predicted successfully on the basis of the changes in the water-accessible polar and nonpolar surface areas of both enzyme and ligand species (Spolar et al. 1992; Murphy et al. 1993). If the structural data of the enzyme–ligand complexes were known to atomic resolutions, then their C_p° values could be predicted easily by determining the solvent-accessible surface areas of the individual enzymes, ligands, and the enzyme–ligand complexes (Peterson et al. 1998b). When such an approach was undertaken for predicting the C_p° values for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to wild-type and E376Q mutant enzymes (see Table 2), the calculated C_p° values were found to be considerably different from those obtained experimentally. In fact, the predicted C_p° values were more or less the same for binding of the above ligands to the wild-type and mutant enzymes (Table 2). This is in marked contrast to the observation that the binding of a more polar ligand, namely, 2-azaoctanoyl-CoA, to the wild-type enzyme yields 101 cal/mole/K more negative C_p° value than that obtained for the binding of a less polar ligand, octenoyl-CoA. However, the C_p° values for the binding of both these ligands to the E376Q mutant enzyme were essentially the same. Clearly, these results are not explainable in light of the classic paradigm of the hydrophobic effect, according to which the magnitude of C_p° (for the binding of ligands to their cognate enzyme sites) should be related directly to the extent of hydrophobic interactions (Tanford 1980; Dill 1990). Several investigators have arrived at a similar conclusion in explaining protein–ligand, membrane–ligand, and protein–protein interaction data (Holdgate et al. 1997; Rowe et al. 1998; Ladbury et al. 1994; Raman et al. 1995; Jin et al. 1993; Guinto and Di Cera 1996; Lundbak et al. 2000).

Freire and his collaborators (Gomez et al. 1995) have pointed out that the absolute heat capacity of a protein is comprised of the contributions from (1) the covalent primary structure (containing the contributions from vibrational frequencies, arising from the stretching and bending modes of each valence bond and internal rotations), (2) the noncovalent secondary and tertiary structures, and (3) the hydration/dehydration of individual atoms. Of these, although the primary (covalent) structural terms contribute maximally to the absolute heat capacity (C_p) of proteins, the changes in heat capacity (C_p°) on protein-folding and/or the protein–ligand interactions are dictated primarily by the hydration/dehydration terms. However, in the case of protein–ligand interactions, sometimes it may be difficult to sort the contributions of one heat capacity term from the other. For example, if the ligand binding alters the protein conformation, and the latter proceeds in concomitance with the changes in the hydration/dehydration properties of certain residues, it would be difficult to assign unambiguously the heat capacity contributions from the noncovalent versus hydration terms (Ayala et al. 1995; Ladbury et al. 1995; Lundback et al. 2000). Likewise, if the ligand binding accompanies sequestration of water molecules (which are highly ordered and possess reduced mobility) in the vicinity of the ligand binding site, the observed C_p° value would be different from that predicted on the basis of hydration/dehydration features of the ligand and enzyme sites (Guinto and Di Cera 1996; O'Brien et al. 1998)). A qualitatively similar conclusion can be drawn if the protein ligand interaction is mediated via a water molecule (see below). Evidently, the generalization derived from solvation/desolvation of small molecular weight compounds, as well as from the protein folding/unfolding data, may not be universally applicable for predicting the C_p° values of the protein–ligand interacting systems.

The question arises as to what other structural feature or features contribute to the negative magnitudes of H° and C_p° on protein–ligand interactions, such as described herein. Although there is no definitive answer to this question, circumstantial evidence corroborates the notion that the origin of large negative H° and C_p° values lies partly in the solvent-mediated interaction between the charged species (Ladbury 1996; Morton and Ladbury 1996; Tame et al. 1996; Sleigh et al. 1999). Ladbury (1996) argues that the favorable enthalpic gain attributable to the water mediated (protein–ligand) interaction often overweighs the entropic penalty because of restrictions on the degree of freedom of the water molecule, thus resulting in a favorable free energy of interaction. In their detailed studies, Holdgate et al. (1997) described that the Arg136 His mutation in DNA gyrase creates a cavity that sequesters a water molecule, which mediates the interaction between novobiocin and the enzyme via hydrogen bonding. Such an arrangement yields large negative H° and C_p° values. Likewise, Bhat et al. (1994) reported that the water molecules entrapped between the hen egg white lysozyme antigen-antibody interfaces promote their interactions and yield large negative C_p° values. In light of these observations, it is tempting to speculate that the discriminatory features in thermodynamic parameters for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA to the wild-type enzyme are because of solvent-mediated interaction (albeit of different magnitudes) between the negatively charged carboxyl group of Glu-376 and the partially positively charged ligand species. We propose that the water molecules, which remain bound to the enzyme site phase, are reorganized on binding of ligands, and depending on their orientation, they bridge between the charged ligand and enzyme site residues (see Results). This may lead to a decrease in the heat capacity of water molecules on transition from their (initially resident) hydrophobic phase (possessing higher heat capacity) to the hydrogen-bonded phase involving the charged residues (possessing lower heat capacity). Because the E376Q mutation abolishes the net negative charge from the enzyme site, the solvent-mediated charge–charge interaction is not as pronounced as observed with the wild-type enzyme, and thus the ligand-dependent discrimination in H° and C_p° values for the binding of 2-azaoctanoyl-CoA and octenoyl-CoA is abolished. We are currently testing this hypothesis by using Glu-376 Asp mutation of MCAD (Peterson et al. 1998a), which creates an 18 ų cavity (Peterson and Srivastava 2000) and has the potential to entrap a water molecule, on the magnitudes of H° and C_p° for the binding of above ligands, and we will report on these findings subsequently.

	Materials and methods

TOP Abstract Introduction Results Discussion Materials and methods References

 
Coenzyme A was purchased from Life Science Resources. Trans-2-octenoic acid was purchased from Pfaltz and Bauer; n-hexyl isocyanate was purchased from Acros. All other reagents were of analytical grade. Unless otherwise stated, all experiments were performed in 50 mM potassium phosphate at pH 7.6, containing 10% glycerol and 0.3 mM EDTA.

The expression system for the Glu-376 Gln mutation of MCAD was obtained initially as a gift from Dr. Sandro Ghisla's laboratory. Subsequently, this and a variety of other site-specific mutations of MCAD were created in our laboratory using the Quick Change site-directed mutagenesis kit from Stratagene. As will be elaborated elsewhere, the overall procedure involved the annealing of two (5' and 3') synthetic primers containing the mutation site on to the double-stranded expression vector of wild-type MCAD, followed by extension of the primers by thermal cycling using PFU DNA polymerase. The reaction products were treated with Dpn-I (specific for methylated DNA) to cleave the parental (nonmutated) DNA template. The nicked vector DNA, incorporating the desired mutation, was transformed into Epicurian Coli XL1-Blue super-competent cells, and the selection was performed via antibiotic (ampicillin) resistance. Finally, the plasmid containing the desired mutations was transformed into E. coli TG1 cells for expression.

The expression and purification of wild-type and Glu-376 Gln mutant human liver MCADs were performed as described by Peterson et al. (1995). The wild-type enzyme was assayed by monitoring the reduction of ferricenium hexafluorophosphate (FcPF₆) at 300 nm ( = 4.3 mM^-1cm^-1; Powell and Thorpe 1988) in a reaction mixture containing 100 µM octanoyl-CoA and 350 µM FcPF₆. The A₂₈₀/A₄₅₀ ratio of the purified enzyme was 5. Glu-376 Gln mutant enzyme was essentially inactive. The extinction coefficients of both wild-type and mutant enzymes were taken to be 15.4 mM^-1cm^-1 at 446 nm (Thorpe et al. 1979).

Octenoyl-CoA was prepared by the mixed anhydride method of Bernert and Sprecher (1977) and purified on a C₁₈ reverse-phase, high pressure liquid chromatography (HPLC) column, equilibrated with 20 mM potassium phosphate at pH 7.0, and eluted with an increasing gradient of methanol as described by Kumar and Srivastava (1994). The ligand, 2-azaoctanoyl-CoA, was prepared essentially as described by Trievel et al. (1995). Briefly, a solution of n-hexyl isocyanate dissolved in THF was added, under a nitrogen stream, to a solution of coenzyme A in 0.25 M NaHCO₃. The mixture was stirred under nitrogen for 30 min, after which the pH was adjusted to 5.5 with acetic acid. The products were purified on a C₁₈ reverse-phase HPLC column equilibrated with 20 mM ammonium acetate at pH 5.5, developed with an increasing gradient of methanol. All ligands were stored as lyophilized powders and eluted as a single peak at 260 nm. The extinction coefficients of octenoyl-CoA and 2-azaoctanoyl-CoA were taken to be 20.4 mM^-1cm^-1 and 16.0 mM^-1cm^-1 at 258 nm and 260 nm, respectively (Trievel et al. 1995; Kumar and Srivastava 1994).

Isothermal titration microcalorimetry
The isothermal titration microcalorimetry experiments were performed on an MCS isothermal titration microcalorimeter from Microcal as described by Srivastava et al (1997). The sample cell was filled with 1.8 mL (1.36 mL effective volume) of the buffer (control) or enzyme solution (experiment). The injector was filled with 250 µL of the appropriate ligand. The titration was initiated by a preliminary first injection of 1 µL followed by 59 injections of 4 µL each. During the experiment, the enzyme solution was stirred at a constant rate of 400 rpm.

All calorimetric data were presented after correction for the background as described previously (Srivastava et al. 1997). The experimental data were analyzed according to Wiseman et al. (1989).

The data analysis produced three parameters, namely, stoichiometry (n), association constant (K_a), and the standard enthalpy change (H°) for the binding of the ligand to MCAD. The standard free energy change (G°) for the binding was calculated according to the relationship G° = -RT ln K_a. Given the magnitudes of G° and H°, the standard entropy change (TS°) for the binding process was calculated according to the standard thermodynamic equation, G° = H° - TS°.

Molecular model-building studies
All the molecular model building studies were performed on a Silicon Graphics (SGI)-O2 molecular modeling workstation with the aid of Insight-II (98), Biopolymers-98, Homology-98, and Discover-98 software developed by Molecular Simulations. The coordinates for the X-ray crystallographic structure of pig liver medium chain acyl-CoA dehydrogenase in the absence (pdb3mdd.pdb) and presence of C₈-CoA (pdb3mde.pdb) were downloaded from the Brookhaven Protein Data Bank. The structural data of these proteins were manipulated and displayed with the aid of Insight-II. All modeling studies were performed in the presence of the crystallographically identified water molecules in the protein structures. The enzyme-bound C₈-CoA structure was modified to adopt the structures of octenoyl-CoA and 2-azaoctanoyl-CoA with the help of the Biopolymer module of Insight-II. The Glu-376 Gln mutation was constructed by replacing the side chain of Glu-376 with Gln with the aid of Homology-98 software.

The energy minimizations of the enzyme–ligand complexes were performed using the consistent valence force field (CVFF) with the aid of Discover-98 software (Dauber-Osguthorpe et al. 1988). The overall protocol involved a combination of the steepest descent and conjugate gradient approach via the following sequence of steps: (1) By fixing all atoms, except for hydrogen, the appropriate structure was subjected to 200 iterations of steepest descent followed by 100 iterations of conjugate gradient until a derivative of 1 kcal/mole/Å was achieved. (2) By fixing the protein backbone, FAD, and CoA-ligands, 2000 iterations were performed via the above minimization steps until a derivative of 0.01 kcal/mole/Å was achieved. (3) Finally, by restraining the backbone structure, the energy minimization (15,000 iterations) was performed via the conjugate gradient approach until an absolute derivative in the range of 10⁴ to 10^-5 kcal/mole/ Å was achieved.

Strategically, the wild-type-octenoyl-CoA complex was first minimized followed by changing the octenoyl-CoA structure to 2-azaoctanoyl-CoA and reminimization of the resultant complex. The Glu-376 Gln mutation was created in the minimized wild-type enzyme–ligand complexes before subjecting the corresponding structures to further energy minimization via the above sequence of steps. The maximum derivative of all the minimized structures decreased to a level of 10^-4 to 10^-5 kcal/mole/Å, which is considered to be ideal for the globally minimized structure of the protein–ligand complexes of comparable dimensions (Roberts et al. 1986).

Using the energy-minimized structures of the individual enzyme–ligand complexes, we determined the solvent-accessible nonpolar and polar surface areas of the enzymes in the absence and presence of bound ligands, and those of the isolated ligands were determined according to Lee and Richards algorithm (Lee and Richards 1971), using a probe with a radius 1.4 Å via the ProState module of Homology-98 (Molecular Simulations).

	Acknowledgments

 
This work was supported by National Science Foundation grant, MCB-9904416.

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

TOP Abstract Introduction Results Discussion Materials and methods References

 
Ayala, Y.M., Vindigni, A., Nayal, M., Spolar, R.S., Record, Jr., M.T., and Di Cera, E. 1995. J. Mol. Biol. 253: 787–798.[CrossRef][Medline]

Baldwin, R. 1986. Proc. Natl. Acad. Sci. 83: 8069–8072.[Abstract/Free Full Text]

Beinert, H. 1963. Enzymes, 2d ed. 7: 447–466.

Bernert, J.T. and Sprecher, H. 1977. J. Biol. Chem. 252: 6737–6744.

Bhat, T.N., Bentley, G.A., Greene, M.I., Tello, D., Dall'Acqua, W., Souchon, H., Schwarz, F.P., Mariuzza, R.Y., and Poljak, R. 1994. Proc. Natl. Acad. Sci. 91: 1089–1093.[Abstract/Free Full Text]

Chaires, J.B. 1997. Biophys. Chem. 64: 15–23.[CrossRef][Medline]

Cummings, J.G. and Thorpe, C. 1994. Biochemistry 33: 788–797.[CrossRef][Medline]

Dauber-Osguthorpe, P., Roberts, V.A., Osguthorpe, D.J., Wolff, J., Genest, M., and Hagler, A.T. 1988. Proteins 4: 31–47.[CrossRef][Medline]

Dill, K.A. 1990. Biochemistry 29: 7133–7155.[CrossRef][Medline]

Dunitz, J.D. 1995. Chem. Biol. 2: 709–712.[CrossRef][Medline]

Edsall, J.T. 1935. J. Am. Chem. Soc. 57: 1506–1507.

Engel, P.C. 1990. Chemistry and Biochemistry of Flavoenzymes (ed. F. Muller), Vol. III, pp. 597–655, CRC Press, London.

Engst, S., Vock, P., Wang, M., Kim, J.J.P., and Ghisla, S. 1999. Biochemistry 38: 257–267.[CrossRef][Medline]

Gilson, M.K. and Honig, B.H. 1988. Proteins 3: 32–52.[CrossRef][Medline]

Gomez, J., Hilser, V.J., Xie, D., and Freire, E. 1995. Proteins 22: 404–412.[CrossRef][Medline]

Gopalan, K.V., and Srivastava, D.K. 1997. Anal. Biochem. 250: 44–50.[CrossRef][Medline]

Guinto, E.R. and Di Cera, E. 1996. Biochemistry 35: 8800–8804.[CrossRef][Medline]

Herron, J.N., Kranz, D.M., Jameson, D.M., and Voss, Jr., E.W. 1986. Biochemistry 25: 4602–4609.[CrossRef][Medline]

Holdgate, G.A., Tunnicliffe, A., Ward, W.H.J., Weston, S.A., Rosenbrock, G., Barth, P.T., Taylor, I.W.F., Puptit, R.A., and Timms, D. 1997. Biochemistry 36: 9663–9673.[CrossRef][Medline]

Jin, L., Yang, J., and Carey, J. 1993. Biochemistry 32: 7302–7309.[CrossRef][Medline]

Johnson, B.D., Mancini-Samuelson, G.J., and Stankovich, M.T. 1995. Biochemistry 34: 7047–7055.[CrossRef][Medline]