14810 J. Phys. Chem.

B 2009, 113, 14810–14815

Aromatic Interactions in the Binding of Ligands to HMGCoA Reductase

Emily A. Kee, Maura C. Livengood, Erin E. Carter, Megan McKenna, and Mauricio Cafiero*
Department of Chemistry, Rhodes College, 2000 N. Parkway, Memphis, Tennessee 38112
ReceiVed: May 14, 2009; ReVised Manuscript ReceiVed: September 28, 2009

3-Hydroxy-3-methyglutaryl-coenzyme A (HMGCoA) reductase is the enzyme that catalyzes the rate-

determining step in cholesterol synthesis; it is also the target for statin drugs, which are competitive inhibitors
of the enzyme. We examine potentially important enzyme-ligand interactions currently not incorporated
into statin drug design: weak, induction/dispersion interactions between ligands and residue tyrosine 479 in
the HMGCoA reductase active site. HMGCoA is a large molecule with a long coenzyme A “tail”, and in
order to study the interactions of interest, it was necessary to find the smallest possible portion of the HMGCoA
molecule that would serve as a reasonable model for the entire molecule. Using this minimal model, we
calculated BSSE-corrected electronic interaction energies between the residue and the ligand molecule using
several DFT methods (local, hybrid, and gradient-corrected DFT methods) as well as MP2. We also performed
several in silico mutations of the tyrosine 479 residue to determine the potential effects of these changes on
protein-ligand interaction energies. Our work shows that this previously unexploited protein-ligand interaction
between tyrosine residue 479 and HMGCoA can be important in the design of future statin drugs. Per our
previous work, our results show that local DFT methods more closely match MP2 energy values for aromatic
binding than do hybrid or gradient-corrected DFT methods.

Introduction
In the second step of the 13-step cholesterol synthesis
pathway, 3-hydroxy-3-methyglutaryl-coenzyme A (HMGCoA,
Figure 1) is activated by enzyme HMGCoA reductase and
converted into mevalonic acid. Cholesterol-lowering drugs
known as statins bind to the active site of HMGCoA reductase,
inhibiting its ability to activate HMGCoA and ultimately
lowering LDL cholesterol levels. Statins primarily use dipole/
dipole and hydrogen bonds to bind to the active site of the
reductase, focusing on the portion of the active site dominated
by lysine 691 (K691), lysine 692 (K692), aspartic acid 690
(D690), and serine 684(S684), with minimal explicit incorpora-
tion of hydrophobic side-chain interactions. However, our
previous work has shown that in protein-substrate binding,
induction and dispersion forces and aromatic ring-ring binding
can be important.1 Thus, in the current work we focus on the
ring-ring interaction between HMGCoA and HMGCoA re- Figure 1. HMGCoA molecule structure from the crystal structure
ductase at tyrosine residue 479 (Y479) to determine if this where it is bound to HMG-CoA reductase.
traditionally weaker interaction plays a significant role in
protein-substrate binding. We believe that if this aromatic We suggest that statin drugs should incorporate induction and
interaction is important in binding the natural substrate, HMG- dispersion interactions explicitly, so we are evaluating the
CoA, to the enzyme then it can play an important role in binding importance of these interactions in the statin target enzyme using
novel statin drugs to the same active site in the enzyme. the natural substrate HMGCoA. We have used several DFT
Because of the immense profitability of statin drugs, there methods (local, hybrid, and gradient-corrected or GGA) and
has been much recent interest in binding to the HMGCoA second-order Moller-Plesset perturbation theory (MP2) to
reductase active site. Istvan and Diesenhofer2 show the crystal calculate the interaction energies between the HMGCoA ligand
structures of six different statin drugs in complex with HMG- and a truncated portion of HMGCoA reductase. MP2 was
CoA reductase, identifying important amino acid interactions selected on the basis of convenience and speed and it was used
in the binding of the ligands to the active site (mentioned above). asthereferencetocompareinteractionenergiesofsubstrate-protein
Pfefferkorn et al.3 demonstrated structural modifications that interactions calculated by DFT methods. Because of time and
aided in binding drugs to the active sites of HMGCoA reductase, computing constraints, calculating the electronic structure for
increasing the potential efficacy of statin drugs to inhibit larger portions of the protein would be unfeasible with methods
cholesterol synthesis. da Silva et al.4 performed modifications other than DFT. Thus, looking toward larger-scale calculations
on a variety of statin drugs and evaluated the resulting binding in the future, we evaluate a suite of DFT methods for their ability
affinities through computer modeling and experimental evaluation. to mimic MP2 results. Contrary to what would typically be
10.1021/jp904508j CCC: $40.75  2009 American Chemical Society
Published on Web 10/13/2009
Binding of Ligands to HMGCoA Reductase
Figure 2. HMG portion of HMGCoA in complex with amino acid
residues D690, K691, K692, and S684.
expected, the more rigorous semilocal DFT methods were less
accurate in predicting interaction energies than local DFT
methods when compared to MP2 energy values. This is due to
a diffuse, slowly varying density found between the molecules
in the ring-ring interaction, which compensates for gradient
overcorrection in the GGA methods.
Truhlar et al. used two DFT methods5,6 that they developed,
PWB6K and M05-2X, to approximate interaction energies of
models that are similar in structure and orientation to the
aromatic interactions in the binding of ligands to HMGCoA
reductase that we have studied. These results mirror our findings
that local DFT methods and the meta DFT methods PWB6K
and M05-2X best approximate aromatic π-π stacking interac-
tions, such as the ring interactions that we studied in the present
work. Because of a lack of availability of the methods developed
by Truhlar et al., we have not been able to incorporate PWB6K
or M05-2X into our calculations; however, the incorporation
of these methods in future work is a possibility.
Computational Methods
The structures of all ligand-amino acid complexes studied
here were isolated from the crystal structure of HMGCoA bound
to HMGCoA reductase;2 these include the complex of the HMG
portion of HMGCoA with each of the four amino acids S684,
D690, K691, and K692 (Figure 2) and the complex of
HMGCoA with Y479 (Figure 3). Hydrogen atoms were added
to all free valences in these structures, and oxygen atoms and
hydroxyl groups were added where necessary to cap free amino
acids. We optimized the positions of all added atoms using the
B3LYP7/6-31 g8 level of theory. When truncating the HMGCoA
molecule (see Figure 3 for truncated complexes), we replaced
the excised atoms with a hydrogen atom and reoptimized all
hydrogen atoms. The full HMGCoA-Y479 complex was
broken down into two portions for further analysis: a complex
between Y479 and the ring portion of HMGCoA and a complex
between Y479 and the phosphate group of HMGCoA (these
two portions lead to what we call the ring interaction and
phosphate group interactions; see Figure 4).
Counterpoise-corrected interaction energies for all molecular
complexes were computed using MP2 and density functional
theory methods9,10 HF, B3LYP, SVWN,11,12 HCTH407,14-16
PBE,23,23 TPSS,24 PW91,17 and SPL,18-20 using the 6-31++G**
and 6-311+g* basis sets. Mutant amino acid residues were
created by replacing the Y479 side chain with the appropriate
new side chains (phenylalanine (F), leucine (L), and alanine
(A); see Figure 5) and optimizing the resulting mutants using
J. Phys. Chem. B, Vol. 113, No. 44, 2009 14811
the B3LYP/6-31 g level of theory. The interaction energies of
all mutant models were calculated as above. The in silico
mutation from Y479 to F479 was performed on the basis of
the strong likelihood of occurrence (this requires only a point
mutation to occur naturally) and demonstrates the importance
of the hydroxyl group on tyrosine in binding because pheny-
lalanine is just tyrosine without the hydroxyl group. The in silico
point mutation to A479 was chosen on the basis of a smaller
likelihood of occurrence and, more interestingly, to analyze the
contribution to ligand binding from the phenol group in tyrosine.
Alanine is essentially tyrosine without the phenol, so this
mutation allows us to determine how much binding comes from
the ring and how much comes from the backbone of the residue.
L479 represents a likely mutation and is interesting because its
side chain can have only dispersion interactions with the ligand.
All calculations were made using Gaussian0325 on Pentium
dual core processors.
Electrostatic Interactions between HMGCoA and
HMGCoA Reductase
Table 1 lists the interaction energies between the HMG
portion of HMGCoA and four HMGCoA reductase residues
commonly accounted for in statin design: D690, K691, K692,
and S684. Figure 2 shows these four residues in complex with
the portion of HMGCoA used in the calculation of interaction
energies. It should be noted that the interactions possible
between the residues and the ligands are dominated by dipole/
dipole interactions and hydrogen bonds between the terminal
amino, hydroxyl, and carboxylic acid groups on the amino acid
residues and the hydroxyl and carboxylic acid groups on HMG.
The pairwise interactions between each residue and HMG are
on the order of 1-5 kcal/mol, and the work below demonstrates
that the tail end of the molecule involved in induction and
dispersion has significant interactions that rival these electrostatic
interactions in magnitude and thus has excellent potential for
novel statin design.
Interaction Model System
To model the electronic interaction energies of Y479 in
complex with HMGCoA, we wanted to identify the smallest
possible portion of HMGCoA necessary to maintain a calculated
interaction energy with Y479 consistent with what would be
obtained by using the entire HMGCoA molecule. In Figure 3,
we see Y479 in complex with several portions of HMGCoA.
(See Figure 1 for the entire HMGCoA molecule.) In our first
effort to calculate the binding energy, we truncated the HMG-
CoA molecule at the first point illustrated in Figure 3. The
HMGCoA molecule was cut off at the first oxygen-carbon bond
coming down from the top of the molecule, and the bond to
carbon was replaced by a bond to a hydrogen atom; this
molecular complex is referred to as truncation scheme 1. For
this first truncation scheme, we computed the interaction
energies using MP2 and three DFT methods: SVWN, B3LYP,
and HCTH407 (Table 2). From these results, we observed that
the local DFT method, SVWN, produced qualitatively similar
results to MP2 (this was the only DFT method that showed
attraction), and it was quantitatively similar to MP2 as well
(-11.83 kcal/mol compared to -8.24 kcal/mol, respectively).
A second truncation (truncation scheme 2, Figure 3) was
performed in an attempt to remove more of the CoA tail without
drastically changing the interaction energies calculated in
truncation scheme 1. In truncation scheme 2, only DFT methods
SVWN, B3LYP, and HCTH407 were used to model interaction
energies. Because the energies observed in truncation scheme
14812 J. Phys. Chem. B, Vol. 113, No. 44, 2009
Figure 3. Tyrosine 479 in complex with truncated portions of the coenzyme A tail of HMGCoA. Truncation schemes are described in the text.
Figure 4. Two components of the total complex made up of the side
chain of tyrosine 479 and the ring portion of HMGCoA: the ring
interaction and the phosphate interaction.
1 for SVWN produced similar results to MP2 interaction
energies, we used SVWN as a standard for comparison. With
the second truncation, no significant change in interaction energy
calculated by SVWN was observed (-11.83 kcal/mol for
truncation scheme 1 and -11.99 kcal/mol for truncation scheme
2). This led to a third truncation scheme; however, the resulting
interaction energies for the Y479-HMGCoA complex interac-
tion were significantly different from those observed in previous
truncation schemes (SVWN calculated an interaction energy of
-9.88 kcal/mol). From Figure 3, we can see that in truncation
scheme 3 a phosphate group directly across from the hydroxyl
group on the tyrosine residue was removed (this was not all
that was removed in going from truncation scheme 2 to 3; see
Figure 3), and this is what we believe caused the significant
change in interaction energy. In truncation scheme 4, we
reattached the phosphate group that was removed in truncation
scheme 3 and recalculated the interaction energies using MP2
and the DFT methods. This truncation scheme produced similar
interaction energies (-7.73 kcal/mol for MP2 and -11.59 kcal/
mol for SVWN) to truncation scheme 1 (-8.24 kcal/mol for
MP2 and -11.83 kcal/mol for SVWN), with each method giving
a decrease in interaction energy of 5% or less. Thus, we
determined that truncation scheme 4 would be the optimal
balance of size and accuracy. All of the following work is based
on truncation scheme 4.
Kee et al.
Protein-Ligand Binding: Theoretical Results
Counterpoise-corrected interaction energies of the four com-
plexes (Y479 in complex with HMGCoA in truncation scheme
4 and the three mutant complexes based on this same truncation
scheme) were first calculated using the 6-31++g** basis set
(Table 3). The counterpoise-corrected interaction energies were
then computed again at the 6-311+g* level to test for basis set
convergence. The results (Table 4) were very similar to those
observed using the smaller basis set, 6-31++g**, and differed
by less than 0.1 kcal/mol in most calculations. We can thus
conclude that polarization and the diffuse function on hydrogen
atoms are not crucial at this level of analysis; furthermore, we
surmised that no further expansion of the basis set was
necessary. The following analysis thus references interaction
energy values from Table 4 only. MP2 is used as an accurate
point of reference for comparison, though historically MP2
slightly overestimates these types of interaction energies when
compared to coupled-cluster calculations.26-30 This work em-
phasizes differences in interaction energies rather than absolute
energies, and thus MP2 is considered to be of sufficient
accuracy.
Table 4 shows the interaction energies between the Y479
residue and the entire HMGCoA truncation scheme 4 (labeled
“complex”). In Figure 3, we see that the ring on Y479 is in
close proximity to both the ring from HMGCoA and a phosphate
group. We decomposed the interaction from the total complex
into the contributions from each of these two pieces by creating
two subcomplexes (Figure 4): Y479 and the ring portion of
HMGCoA only (labeled “ring”) and Y479 and the phosphate-
group-containing moiety (labeled “phosphate”). Tables 3 and
4 show interaction energies for the entire complex and for each
of the two pieces.
The MP2 interaction energy values for Y479-HMGCoA
show that Y479 and HMGCoA are bound (negative, attractive
energy value) fairly strongly for a nonbonded interaction. The
decomposition in the ring portion of the complex shows that
the ring-ring interaction is attractive and comprises the bulk
of the total complex interaction energy. The phosphate interac-
tion is actually weakly repulsive according to the MP2 calcula-
tions. The interaction energies for the mutant models show that
Binding of Ligands to HMGCoA Reductase J. Phys. Chem. B, Vol. 113, No. 44, 2009 14813

Figure 5. Mutated residues (F, L, and A) at position 479 in full complex with the ring portion of HMGCoA.

TABLE 1: Counterpoise-Corrected Interaction Energies (kcal/mol) between Various Residues in HMGCoA Reductase and the
HMG Portion of HMGCoAa
6-311+g* HF MP2 SVWN SPL B3LYP HCTH407 PW91 PBE TPSS
D690 -0.28 -2.10 -7.52 -7.25 -0.69 -0.77 -2.70 -2.18 -0.90
K691 0.53 -3.45 -3.89 -3.76 -0.64 -1.56 -1.88 -1.46 -0.75
K692 0.02 -1.53 -3.46 -3.35 -0.41 -1.08 -1.48 -1.22
S684 -3.51 -5.17 -13.28 -12.88 -5.38 -4.00 -7.04 -6.40 -5.62
a
All calculations were performed using 6-311+g*.

TABLE 2: Counterpoise-Corrected Interaction Energies

(kcal/mol) between Tyrosine 479 and Various Truncated
Portions of the Coenzyme A Tail of HMGCoAa
truncation scheme MP2 SVWN B3LYP HCTH407
1 -8.24 -11.83 6.01 8.47
2 -11.99 5.83 8.31
3 -9.88 2.93 3.25
4 -7.73 -11.59 6.17 8.7
a
See Figures 2-5 for truncation schemes. All calculations were
performed using 6-311+g*.
all mutations considered result in less attraction in the total
complex. Likewise, the analysis of the ring interactions alone
shows that all of the mutations result in less attraction (and
repulsion in the case of A479) for this portion of the complex.
This decrease in binding for the Y479F mutation results from
the loss of polarization across the aromatic ring upon removal
of the hydroxyl group. The decrease in binding for the Y479L
mutation results from the replacement of a delocalized electron
system in tyrosine to a saturated hydrocarbon chain in leucine.
Finally, the decrease in binding for the Y479A mutation results
from simply going to a much smaller side chain with many
fewer electrons.
MP2 values for the interaction energies of the phosphate
portions of the complexes are weakly repulsive for the wild
type and slightly attractive (on the order of 1 kcal/mol) for all
mutations made. The phosphate group is strongly polar and
polarizable, which would account for its strong attractive
interaction with the mutant models. The repulsion in the wild
type is likely the result of the phosphate being too close to the
hydroxyl group on tyrosine, causing steric repulsion. The
removal of the hydroxyl group in the Y479F mutation results
in the interaction energy going from slightly repulsive to slightly
attractive, supporting this reasoning.
HF calculations show the interaction energies to be repulsive
for all complex and ring models (wild type and mutants),
disagreeing with MP2 values that are attractive for all but one
of these energies. This nicely illustrates the necessity of
incorporating electron correlation when computing interaction
energies in these weak, nonbonded systems. We found that
nonlocal DFT methods, which mostly resulted in repulsive
interaction energies, did not typically agree with MP2 results.
The local DFT methods considered, which generally resulted
in attractive interaction energies, were often in close agreement
with MP2 values. These results regarding local versus nonlocal
DFT methods are atypical of what is traditionally expected. The
ring interactions (which are dominated by dispersion and
typically require a rigorously nonlocal method) are better
approximated by local DFTs because of the cancellation of
errors between the diffuse, slowly changing electron densities
and the exchange/correlation energy functional.
Protein-Ligand Binding: Physiological Implications
Y479 in HMGCoA reductase binds to HMGCoA strongly.
The computed interaction energy of -7.73 kcal/mol is compa-
rable to the dipole-dipole and hydrogen bonds that bind the
HMG portion of HMGCoA to the enzyme via residues D690,
K691, K692, and S684 (between -1.53 and -5.17 kcal/mol).
Most statin drug molecules, which are competitive inhibitors
of this binding site in the enzyme, are small and have been
found to bind primarily to D690, K691, K692, and S684 and a
few surrounding residues. We suggest that the Y479 interaction
in the binding site should be considered in statin drug design.
14814 J. Phys. Chem. B, Vol. 113, No. 44, 2009 Kee et al.

TABLE 3: Counterpoise-Corrected Interaction Energies (kcal/mol) between Amino Acid Residues at the 479 Position and Two
Different Portions of the Coenzyme A Tail of HMGCoA
mutation 6-31++g** HF MP2 SVWN SPL B3LYP HCTH407 PW91 PBE TPSS
Y479 ring interactions 3.96 -4.86 -4.31 -4.07 2.76 2.1 0.32 0.88 2.22
phosphate interactions 4.24 0.58 -4.99 -4.62 2.96 4.76 1.28 1.84 2.91
complex 8.55 -12.07 -11.36 5.85 8.64 1.63 2.58 5.43
Y479F ring interactions 4.38 -4.19 -3.8 -3.58 3.21 2.63 0.78 1.29 2.65
phosphate interactions 1.62 -0.9 -4.83 -4.57 0.89 1.62 -0.46 -0.058 0.85
complex 7.82 -10.62 -10.03 4.85 6.46 0.88 1.63 4.32
Y479A ring interactions 1.96 0.14 -1.62 -1.5 1.49 0.95 0.36 0.67 1.3
phosphate interactions -0.47 -0.34 -0.24 -0.24 -0.29 -0.35 -0.35 -0.31 -0.38
complex 0.87 -1.3 -2.72 -2.58 0.49 -0.11 -0.82 -0.35 0.22
Y479L ring interactions 4.15 -1.39 -3.63 -3.38 3.02 3.10 0.96 1.39 2.49
phosphate interactions -1.06 -1.05 -1.10 -1.08 -0.46 -1.41 -1.14 -0.86 -0.71
complex 2.99 -3.76 -5.47 -5.19 2.06 1.22 -0.59 0.08 1.33

TABLE 4: Counterpoise-Corrected Interaction Energies (kcal/mol) between Amino Acid Residues at the 479 Position and Two
Different Portions of the Coenzyme A Tail of HMGCoA
mutation 6-311+g* HF MP2 SVWN SPL B3LYP HCTH407 PW91 PBE TPSS
Y479 ring interactions 3.69 -5.18 -4.33 -4.09 2.75 2.1 0.28 0.85 2.18
phosphate interactions 3.98 0.71 -5.1 -4.74 2.68 4.68 1.13 1.72 2.83
complex 8.27 -7.73 -12.03 -11.31 5.63 8.54 1.55 2.53 5.44
Y479F ring interactions 4.32 -4.51 -3.8 -3.57 3.23 2.66 0.77 1.29 2.62
phosphate interactions 1.54 -0.79 -4.82 -4.56 0.84 1.7 -0.45 -0.027 0.89
complex 7.22 -7.66 -10.44 -9.86 4.88 6.59 0.93 1.74 4.42
Y479A ring interactions 1.99 0.2 -1.59 -1.47 1.5 1.05 0.34 0.71 1.31
phosphate interactions -0.4 -0.33 -0.53 -0.52 -0.29 -0.5 -0.54 -0.57 -0.4
complex 0.91 -1.24 -2.86 -2.72 0.54 -0.028 -0.88 -0.49 0.2
Y479L ring interactions 4.19 -1.37 -3.57 -3.33 3.08 3.2 0.91 1.02 2.2
phosphate interactions -0.75 -1.04 -1.05 -1.03 -0.52 -1.39 -1.2 -0.85 -0.72
complex 3.01 -3.73 -5.34 -5.06 2.11 1.34 -0.56 0.15 0.00

In the likely event of mutation from Y479 to F479, our data
predicts that cholesterol synthesis will remain fairly unaffected
on the basis of small changes in interaction energies. For the
small yet finite possibility that tyrosine would naturally mutate
to either alanine or leucine, we can conclude that there would
be a significant drop in interaction energies, which may
adversely affect cholesterol synthesis. Thus, we see that the π-π
interactions specifically are necessary for the binding of HMG-
CoA to HMGCoA reductase.
Conclusions and Future Work
Through an analysis of interactions between ligands and the
truncated HMGCoA molecule, we were able to conclude that
the interactions between Tyr 479 and HMGCoA are very
important (8 kcals/mol). We believe that this π-π-stacking
aromatic interaction between the protein and ligand deserves
to be given more attention in the design of drugs that inhibit
the action of this enzyme. We have begun some molecular-
level rational drug design on an HMGCoA inhibitor, and our
next paper describes statin drug design using this weak
interaction explicitly. Preliminary work on this had included
the design of drug candidates and docking calculations that show
that conventional statins modified to take advantage of this
interaction can fit into the active site and dock competitively.
References and Notes
(1) Hofto, M. E.; Cross, J. N.; Cafiero, M. J. Phys. Chem. 2007, 111,
9651. Hofto, L. R.; Van Sickle, K.; Cafiero, M. Int. J. Quantum Chem.
2007, 108, 112. Hofto, M. E.; Godfrey-Kittle, A.; Cafiero, M. THEOCHEM
2007, 809, 125. Van Sickle, K.; Culberson, L. M.; Holzmacher, J. L.;
Cafiero, M. Int. J. Quantum Chem. 2007, 107, 1523. Godfrey-Kittle, A.;
Cafiero, M. Int. J. Quantum Chem. 2006, 106, 2035. Van SickleK., Shroyer,
M.; CafieroM. J. Phys Chem B,submitted for publication. Hofto, L. R.; Lee,
C.; CafieroM. J. Comput. Chem. 2007, 30, 1111.
(2) Istvan, E. S.; Deisenhofer, J. Science 2001, 292, 1160.
(3) Pfefferkorn, J. A.; et al. Bioorg. Med. Chem. Lett. 2007, 17, 4531.
(4) da Silva, V. b.; Taft, C. A.; Silva, C. H. T. P. J. Chem. Phys. A
2008, 112, 2007.
(5) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2005, 109, 5656.
(6) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2007, 3, 289.
(7) Becke, D. J. Chem. Phys. 1993, 98, 5648.
(8) Rassolov, V. A.; Pople, J. A.; Ratner, M. A.; Windus, T. L. J. Chem.
Phys. 1980, 72, 650.
(9) Hohenberg, P.; Kohn, W. Phys. ReV. 1964, 136, B864.
(10) Kohn, W.; Sham, L. J. Phys. ReV. 1965, 140, A1133.
(11) Slater, J. C. Quantum Theory of Molecular and Solids: The Self-
Contained Field for Molecular and Solids; McGraw-Hill: New York, 1974;
Vol. 4.
(12) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(13) Perdew, J. P.; Zunger, A. Phys. ReV. B 1998, 37, 785.
(14) Hamprecht, A.; Cohen, A. J.; Tozer, D. J.; Handy, N. C. J. Chem.
Phys. 1998, 109, 6264.
(15) Boese, D.; Doltsinis, N. L.; Handy, N. C.; Sprik, M. J. Chem. Phys.
2000, 112, 1670.
(16) Boese, D.; Handy, N. C. J. Chem. Phys. 2001, 114, 5497. See also
supporting material EPAPS document E.
(17) Burke, K.; Perdew, J. P.; Wang, Y. In Electronic Density Functional
Theory: Recent Progress and New Directions; Ed. Dobson, J. F., Vignale,
G., Das M. P., Eds.; Plenum Press: New York, 1998.
(18) Perdew, P. In Electron Structure of Solids ′91; Ziesche P., Eschrig
H., Eds.; Akademic Verlag: Berlin, 1991; p 11.
(19) Perdew, P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson,
M. R.; Singh, D. J.; Fiolhais, C. Phys. ReV. B 1992, 46.
(20) Perdew, P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson,
M. R.; Singh, D. J.; Fiolhais, C. Phys. ReV. B 1993, 48.
(21) Perdew, J. P.; Burke, K.; Wang, Y. Phys. ReV. B 1996, 54, 16533.
(22) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77,
3865.
(23) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1997, 78,
1396.
(24) Staroverov, V. N.; Scuseria, G. E.; Tao, J.; Perdew, J. P. J. Chem.
Phys. 2003, 119, 12129.
(25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Binding of Ligands to HMGCoA Reductase
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,
revision C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(26) Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem. A 2006, 110,
10656–10668.
J. Phys. Chem. B, Vol. 113, No. 44, 2009 14815
(27) Sinnokrot, M. O.; Sherrill, C. D. J. Am. Chem. Soc. 2004, 126,
7690–7697.
(28) Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem. A 2003, 107, 8377–
8399.
(29) Ringer, A. L.; Sinnokrot, M. O.; Lively, R. P.; Sherrill, C. D.
Chem.sEur. J. 2006, 12, 3821–3828.
(30) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M.; Tanabe, K.
J. Am. Chem. Soc. 2001, 124, 104–112.
(31) McGrath, M. P.; Radom, L. J. Chem. Phys. 1991, 94, 511.
JP904508J