Combined experimental and theoretical study of long-range interactions modulating dimerization and activity of yeast geranylgeranyl diphosphate synthase.

Abstract

We present here how two amino acid residues in the first helix distal from the main dimer interface modulate the dimerization and activity of a geranylgeranyl diphosphate synthase (GGPPs). The enzyme catalyzes condensation of farnesyl diphosphate and isopentenyl diphosphate to generate a C(20) product as a precursor for chlorophylls, carotenoids, and geranylgeranylated proteins. The 3D structure of GGPPs from Saccharomyces cerevisiae reveals an unique positioning of the N-terminal helix A, which protrudes into the other subunit and stabilizes dimerization, although it is far from the main dimer interface. Through a series of mutants that were characterized by analytic ultracentrifugation (AUC), the replacement of L8 and I9 at this helix with Gly was found sufficient to disrupt the dimer into a monomer, leading to at least 10(3)-fold reduction in activity. Molecular dynamics simulations and free energy decomposition analyses revealed the possible effects of the mutations on the protein structures and several critical interactions for maintaining dimerization. Further site-directed mutagenesis and AUC studies elucidated the molecular mechanism for modulating dimerization and activity by long-range interactions.

Full text

Page 1

Combined Experimental and Theoretical Study of Long-Range
Interactions Modulating Dimerization and Activity of Yeast
Geranylgeranyl Diphosphate Synthase
Chia-Hsiang Lo,†Ying-Hsuan Chang,†Jon D. Wright,§Shih-Hsun Chen,|
Daphne Kan,‡Carmay Lim,*,§and Po-Huang Liang*,†,‡
Institute of Biochemical Sciences, National Taiwan UniVersity, Taipei; Institute of Biological
Chemistry and Biomedical Sciences, Academia Sinica, Taipei; and Department of Biological
Science and Technology, National Chiao Tung UniVersity, Hsin-Chu, Taiwan
Received November 6, 2008; E-mail: phliang@gate.sinica.edu.tw; carmay@gate.sinica.edu.tw
Abstract: We present here how two amino acid residues in the first helix distal from the main dimer interface
modulate the dimerization and activity of a geranylgeranyl diphosphate synthase (GGPPs). The enzyme
catalyzes condensation of farnesyl diphosphate and isopentenyl diphosphate to generate a C20product as
a precursor for chlorophylls, carotenoids, and geranylgeranylated proteins. The 3D structure of GGPPs
from Saccharomyces cerevisiae reveals an unique positioning of the N-terminal helix A, which protrudes
into the other subunit and stabilizes dimerization, although it is far from the main dimer interface. Through
a series of mutants that were characterized by analytic ultracentrifugation (AUC), the replacement of L8
and I9 at this helix with Gly was found sufficient to disrupt the dimer into a monomer, leading to at least
103-fold reduction in activity. Molecular dynamics simulations and free energy decomposition analyses
revealed the possible effects of the mutations on the protein structures and several critical interactions for
maintaining dimerization. Further site-directed mutagenesis and AUC studies elucidated the molecular
mechanism for modulating dimerization and activity by long-range interactions.
Introduction
There are growing numbers of examples in which dimeriza-
tion is required for enzyme activity. For example, the protease
from human immuno-deficiency virus (HIV) has an active site
formed by two monomers, each of which provides a catalytic
Asp residue. The dimer interface is a potential target for an
anti-HIV drug.1The protease from severe acute respiratory
syndrome coronavirus2and the protease from human cytome-
galovirus show no enzyme activity in the monomeric state.3,4
A dipeptide prolyl protease 4, which cleaves the peptide bond
after the penultimate residue, requires a conserved His at the
dimer interface for dimer stability; mutation of this His to Glu
was found to disrupt dimerization and to decrease enzyme
activity by 300-fold.5Whereas previous work showed that the
mutation of a residue in the dimer interface disrupted dimer-
ization, this work reveals mutation of two amino acid (aa)
residues at the N-terminus of an enzyme, geranylgeranyl
diphosphate synthase (GGPPs), distal from the main dimer
interface can also disrupt dimerization and decrease enzymatic
activity. Further insight as to how mutations distal from the
main dimer interface can modulate dimerization was obtained
from molecular dynamics (MD) simulations of the wild-type
and mutant GGPPs and subsequent free energy decomposition
analyses.
GGPPs catalyzes the head-to-tail condensation of C15farnesyl
diphosphate (FPP) with a C5isopentenyl diphosphate (IPP) to
form the C20geranylgeranyl diphosphate (GGPP) product.6-8
This compound can be used to make chlorophylls or R-toceph-
erol,9ent-kaurene, or taxadiene.10It can be further elongated
to produce long-chain isoprenoid used in quinine biosynthesis.9
Furthermore, two GGPP molecules can condense to form
phytoene,11the precursor for many carotenoids. GGPP or FPP
can be used as a ligand for protein prenylation, an essential
post-translational modification for signaling proteins such as Ras,
Rho, Rab, and Rac.12,13Thus, FPP synthase (FPPs), which
†National Taiwan University.
‡Institute of Biological Chemistry, Academia Sinica.
§Institute of Biomedical Sciences, Academia Sinica.
|National Chiao Tung University.
(1) De Clercq, E. Nat. ReV. Drug DiscoVery 2002, 1, 13–25.
(2) Fan, K.; Wei, P.; Feng, Q.; Chen, S.; Huang, C.; Ma, L.; Lai, B.; Pei,
J.; Liu, Y.; Chen, J.; Lai, L. J. Biol. Chem. 2004, 279, 1637–1642.
(3) Drake, P. L.; Cole, J. L.; Waxman, L.; Hall, D. L.; Sardana, M. K.;
Kuo, L. C. J. Biol. Chem. 1996, 271, 7445–7449.
(4) Margosiak, S. A.; Vanderpool, D. L.; Sisson, W.; Pinko, C.; Kan,
C. C. Biochemistry 1996, 35, 5300–5307.
(5) Chien, C. H.; Huang, L. H.; Chou, C. Y.; Chen, Y. S.; Han, Y. S.;
Chang, G. G.; Liang, P. H.; Chen, X. J. Biol. Chem. 2004, 279, 52338–
52345.
(6) Sacchettini, J. C.; Poulter, C. D. Science 1997, 277, 1788–1789.
(7) Ogura, K.; Koyama, T. Chem. ReV. 1998, 98, 1263–1276.
(8) Liang, P. H.; Ko, T. P.; Wang, A. H. Eur. J. Biochem. 2002, 269,
3339–3354.
(9) Ru ˝diger, W.; Benz, J.; Guthoff, C. Eur. J. Biochem. 1980, 109, 193–
200.
(10) Jiang, Y.; Proteau, P.; Poulter, D.; Ferro-Novick, S. J. Biol. Chem.
1995, 270, 21793–21799.
(11) Olszewski, N.; Sun, T.; Gubler, F. Plant Cell 2002, 14, S61–S80.
(12) Dogbo, O.; Laferriere, A.; D’Harlingue, A.; Camara, B. Proc. Natl.
Acad. Sci. U.S.A. 1988, 85, 7054–7058.
(13) Kuzuguchi, T.; Morita, Y.; Sagami, I.; Sagami, H.; Ogura, K. J. Biol.
Chem. 1999, 274, 5888–5894.
Published on Web 02/26/2009
10.1021/ja808699c CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 4051–4062 9 4051

Page 2

catalyzes FPP through condensation of C10geranyl diphosphate
with IPP, has been used as a drug target to develop bisphos-
phonate drugs for treatment of bone resorption diseases such
as osteoporosis and infection by parasitic protozoa.14One of
the drugs was thought to target GGPPs.15
GGPPs belongs to the trans-prenyltransferase family, since
a trans-double bond is formed during the IPP condensation with
FPP. The members in this family, which synthesize the final
products of C15by FPPs, C20by GGPPs, C30by hexaprenyl
diphosphate synthase (HexPPs), and C40by octaprenyl diphos-
phate synthase (OPPs), share sequence homology and possess
similar 3D structures composed of 15 R-helices connected by
loops.16-22However, the GGPPs from Saccharomyces cereVi-
siae shows an unique positioning of the first N-terminal helix
(see Figure 1), which protrudes into and binds the opposite
subunit.20Deletion of this helix led to re-establishment of an
(14) Guo, R. T.; Cao, R.; Liang, P. H.; Ko, T. P.; Chang, T. H.; Hudock,
M. P.; Jeng, W. Y.; Chen, C.; Zhang, Y.; Song, Y.; Kuo, C. J.; Yin,
F.; Oldfield, E.; Wang, A. H. Proc. Natl. Acad. Sci. U.S.A. 2007, 104,
10022–10027.
(15) Goffinet, M.; Thoulouzan, M.; Pradines, A.; Lajoie-Mazenc, I.;
Weinbaum, C.; Faye, J. C.; Seronie-Vivien, S. BMC Cancer 2006, 6,
60.
(16) Guo, R. T.; Kuo, C. J.; Chou, C. C.; Ko, T. P.; Shr, H. L.; Liang,
P. H.; Wang, A. H. J. Biol. Chem. 2004, 279, 4903–4912.
(17) Hosfield, D. J.; Zhang, Y.; Dougan, D. R.; Broun, A.; Tari, L. W.;
Swanson, R. V.; Finn, J. J. Biol. Chem. 2004, 279, 8526–8529.
(18) Sun, H. Y.; Ko, T. P.; Kuo, C. J.; Guo, R. T.; Chou, C. C.; Liang,
P. H.; Wang, A. H. J. Bacteriol. 2005, 187, 8137–8148.
(19) Gabelli, S. B.; McLellan, J. S.; Montalvetti, A.; Oldfield, E.; Docampo,
R.; Amzel, L. M. Proteins 2006, 62, 80–88.
(20) Chang, T. H.; Guo, R. T.; Ko, T. P.; Wang, A. H. J.; Liang, P. H.
J. Biol. Chem. 2006, 281, 14991–15000.
Figure 1. Ribbon representation of the dimer structures of trans-prenyltransferaes. From left to right are the structures of FPPs in the 1st column (PDB
entries 1RQI, 1YHL, 1FPS, and 1YV5, for Escherichia coli, Trypanosoma cruzi, avian, and human FPPs, respectively), GGPPs in the 2nd column (PDB
entries 1WY0, 1WMW, 2DH4, and 2FVI for Pyrococcus horikoshii Ot3, Thermus thermophilus, yeast, and human GGPPs, respectively), HexPPs and OPPs
in the 3rd column (PDB entries 2AZJ and 1V4E for Sulfolobus solfataricus HexPPs and Thermotoga maritima OPPs, respectively). In these structures, the
left subunit is displayed in orange with the first helix in green, and the right subunit is in blue. Only in the yeast GGPPs structure (boxed), the first helix is
protruded from its own subunit to bind with the opposite subunit. The helix A is distant from the main dimer interface formed by residues mainly from
helices F and G. Figures plotted using the PyMol program.38
4052J. AM. CHEM. SOC. 9 VOL. 131, NO. 11, 2009
ARTICLES
Lo et al.

Page 3

equilibrium between the dimeric and monomeric forms of the
enzyme, but greatly favored monomer, and significant loss of
enzymatic activity.20
As helix A is distant from the main dimer interface (∼25
Å), it was not understood (i) how it contributes to the GGPPs
subunit dimerization and (ii) whether the entire helix is required
for dimerization and enzyme activity. In this study, site-directed
mutagenesis, in conjunction with analytical ultracentrifugation
(AUC) measurements and molecular dynamics (MD) simula-
tions, was used to gain insight into the long-range interactions
modulating dimerization.
Materials and Methods
Materials. PfuTurbo, the plasmid miniprep kit, DNAgel extrac-
tion kit, and Ni-NTA resin were purchased from Qiagen. The
protein expression kit (including the pET32Xa/LIC vector and
competent JM109 and BL21 cells) was obtained from Novagen.
The QuikChange site-directed mutagenesis kit was obtained from
Stratagene. Radiolabeled [14C]IPP (55 mCi/mmol) was purchased
from Amersham Pharmacia Biotech. Nonlabeled FPP and IPP were
obtained from Sigma. All commercial buffers and reagents were
of the highest grade.
Plasmid Construction for Mutant GGPPs. The gene encoding
GGPPs was cloned from S. cereVisiae genomic DNA as previously
reported.20GGPPs mutants were prepared by using the QuikChange
site-directed mutagenesis kit. The mutagenic oligonucleotides are
5′-atggaggccaagatagatGGGctgatcaataatgatcctgtttgg-3′ and 5′-ccaaa-
caggatcattattgatcagCCCatctatcttggcctccat-3′ for E7G, 5′-atggaggc-
caagatagatgagGGGatcaataatgatcctgtttgg-3′ and 5′-ccaaacaggatcat-
tattgatCCCctcatctatcttggcctccat-3′
gatagatgagctgGGGaataatgatcctgtttgg-3′ and 5′-ccaaacaggatcattatt-
CCCcagctcatctatcttggcctccat-3′ for I9G, 5′-atggaggccaagataga-
GGGGGGatcaataatgatcctgtttgg-3′ and 5′-ccaaacaggatcattattgat-
CCCCCCatctatcttggcctccat-3′ for E7G/L8G, 5′-atggaggccaagata-
gatgagGGGGGGaataatgatcctgtttgg-3′ and 5′-ccaaacaggatcattatt-
CCCCCCctcatctatcttggcctccat-3′ for L8G/I9G, and 5′-ggtcc-
agccaaaatgaaagcGGGGGTtcaaaaccttat-3′
ACCCCCgctttcattttggctggacc-3′ for L22G/I23G; the mutagenic
oligonucleotides for performing truncated mutagenesis are 5′-
ggtattgagggtcgcgagctgatcaataatgatcctgt-3′ and 5′-agaggagagttagagc-
ctcacaattcggataagtggtc-3′ for ∆(1-6), 5′-ggtattgagggtcgcctgat-
caataatgatcctgtttg-3′ and 5′-agaggagagttagagcctcacaattcggataag-
tggtc-3′ for ∆(1-7), and 5′-ggtattgagggtcgcatcaataatgatcctgtttggtc-
3′ and 5′-agaggagagttagagcctcacaattcggataagtggtc-3′ for ∆(1-8).
∆(1-9) and ∆(1-17) were previously generated.21
The mutagenic oligonucleotides for performing site-directed
mutagenesis to probe the important interactions in stabilizing
dimerization are 5′-ggagatgtatGGGaataggttGGGaataaaacaggcgg-3′
and 5′-ccgctgttttattCCCaaccatattCCCatacatctcc-3′ for L163G/
M167G, 5′-ggttccttttataaatcttGGGggtattGGGtatcagattagagatg-3′ and
5′-catctctaatctgataCCCaataccCCCaagatttataaaaggaacc-3′ for L200G/
I203G, 5′-ggagatgtatGGGaatatggttatttgaataaaacaggcgg-3′ and 5′-
ccgcctgttttattcataaccatattCCCatacatctcc-3′ for L163G, 5′-ggagatg-
tatttgaatatggttGGGaataaaacaggcgg-3′ and 5′-ccgcctgttttattCCC-
aaccatattcaaatacatctcc for M167G, 5′-ggttccttttataaatcttGGGggtat-
tatttatcagattagagatg-3′ and catctctaatctgataaataataccCCCaagatttat-
aaaaggaacc-3′ for L200G, 5′-ggttccttttataaatcttttgggtattGGGtat-
cagattagagatg-3′ and 5′-catctctaatctgataCCCaatacccaaaagatttataaaa-
ggaacc-3′ for I203G, 5′-ggtgtaccctccactataGGCaccgcaaattatatg-3′
and 5′-catataatttgcggtGGCtatagtggagggtacacc-3′ for N101G, 5′-
ggacaaggcttgAAGatatactggagagactttctgcc-3′ and 5′-ggcagaaagtctctc-
cagtatatCTTcaagccttgtcc-3′ for D145K, 5′-gattacgattttcaacgaa-
GCAttgatcaatctacataggggacaag-3′ and 5′-cttgtcccctatgtagattgatcaa-
for L8G,5′-atggaggccaa-
and5′-ataaggttttga-
TGCttcgttgaaaatcgtaatc-3′ for E134A, 5′-aaaacaggcggccttttcGCAt-
taacgttgagactcatg-3′ and 5′-catgagtctcaacgttaaTGCgaaaaggccgcctgtttt-
3′ for R175A, and 5′-ggccattcgttggttccttttataGCTcttctgggtattatttatcag-
3′ and 5′-ctgataaataatacccagaagAGCtataaaaggaaccaacgaatggcc-3′ for
M167G/N199A (with M167G as template).
Expression and Purification of the Mutant GGPPs. The
mutant GGPPs plasmids were used to transform E. coli JM109
competent cells that were streaked on a Luria-Bertanim (LB) agar
plate containing 100 µg/mL ampicillin. Ampicillin-resistant colonies
were selected from the agar plate and grown in 5 mL of LB culture
containing 100 µg/mL ampicillin overnight at 37 °C. The correct
constructs confirmed by sequencing were transformed to E. coli
BL21(DE3) for protein expression. The 60 mL of overnight culture
of a single transformant was used to inoculate 6 L of fresh LB
medium containing 100 µg/mL ampicillin. The cells were grown
to A600) 0.6 and induced with 1 mM IPTG at 16 °C. After 16 h,
the cells were harvested by centrifugation at 7000g for 15 min to
collect the cell paste. The enzyme purification was conducted at 4
°C. Cell paste was suspended in 75 mL of lysis buffer containing
25 mM Tris-HCl, at pH 7.5, and 150 mM NaCl. Cell lysate was
prepared with a French pressure cell press (AIM-AMINCO Spec-
tronic Instruments) and centrifuged at 17 000g to remove cell debris.
The cell-free extract was loaded onto the Ni-NTA column, which
had been previously equilibrated with lysis buffer. The column was
washed with 10 mM imidazole followed by 20 mM imidazole-
containing buffer. His-tagged GGPPs eluted with 100 mM imidazole
was dialyzed twice against 3 L of buffer (25 mM Tris-HCl, pH
7.5, and 150 mM NaCl) and then subjected to the Factor Xa (FXa)
protease digestion to remove the tag. The mixture was then passed
through another Ni-NTA column, and subsequently, untagged
GGPPs was eluted with 10 mM imidazole-containing buffer and
then dialyzed twice against 3 L of buffer (25 mM Tris-HCl, pH
7.5, and 150 mM NaCl) for storage. SDS-PAGE analysis was used
to check the purity of GGPPs and its mutants.
AUC Experiments. AUC measurements of the purified mutant
GGPPs proteins were performed at the approximate concentrations
of 0.25-1.0 mg/mL. The sedimentation coefficients (S) of the
enzymes were estimated with a Beckman-Coulter XL-A analytical
ultracentrifuge with an An60Ti rotor. The sedimentation velocity
experiment was performed at 40 000 rpm at 20 °C with a standard
double-sector aluminum centerpiece. Both sets of data were
analyzed with the program Sedfit version 9.4c (http://www.ana-
lyticalultracentrifugation.com) to calculate the molecular weights
and sedimentation coefficients as previously described.23The
Sednterp version 1.09 program (http://www.jphilo.mailway.com/)
was used to calculate solvent densities and viscosities. The observed
sedimentation profiles of a continuous size distribution (c(s)) can
be calculated from the following equation: a(r, t) ) ∫c(s)L(s, D, r,
t)ds + ε, where a(r, t) denotes the experimentally observed signal.
L(s, D, r, t) denotes the solution of the Lamm equation for a single
species and ε is the noise component. The dissociation constant
(Kd) of the dimer-monomer equilibrium is calculated by global
fitting of the sedimentation data with the SEDPHAT program.
Kinetic Measurements of the Mutant GGPPs. For enzymatic
activity measurements, each mutant GGPPs (0.2 µM E7G, L8G,
I9G, E7G/L8G, L8G/I9G, ∆(1-6), ∆(1-7), ∆(1-8), ∆(1-9), or
∆(1-17)) was used. The reaction was initiated in 200 µL of solution
containing 100 mM Hepes (pH 7.5), various concentrations of FPP
and [14C]IPP as specified below, 50 mM KCl, 0.5 mM MgCl2, and
0.1% Triton X-100 at 25 °C. The enzyme concentration used in all
experiments was determined from its absorbance at 280 nm (ε )
20 340 M-1cm-1). Measurements of the kinetic parameters for the
wild-type and the mutants followed our published procedure.24For
the IPP Kmdeterminations, 25 µM FPP was utilized to saturate the
enzyme, and the IPP concentrations from 0.25-50 µM varied with
the Kmof IPP for each mutant were employed. For the FPP Km
and kcatmeasurements, 0.25-50 µM FPP was used along with 50
(21) Kavanagh, K. L.; Dunford, J. E.; Bunkoczi, G.; Russell, R. G. G.;
Oppermann, U. J. Biol. Chem. 2006, 281, 22004–22012.
(22) Kloer, D. P.; Welsch, P.; Beyer, P.; Schulz, G. E. Biochemistry 2006,
45, 15197–15204.
(23) Chang, H. C.; Chang, G. G. J. Biol. Chem. 2004, 278, 23996–24002.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 11, 2009
4053
Dimerization of Geranylgeranyl Diphosphate Synthase
ARTICLES

Page 4

µM [14C]IPP. To measure the initial rate, 40-µL portions of the
reaction mixture were periodically withdrawn within 10% substrate
depletion and then mixed with 10 mM EDTA for reaction
termination. The radiolabeled products were then extracted with
1-butanol, and the radioactivity associated with aqueous and butanol
phases were separately quantified by using a Beckmann LS6500
scintillation counter. Data of initial rates versus substrate concentra-
tions were analyzed by nonlinear regression of the Michaelis-Menten
equation using the KaleidaGraph (Synergy software) to obtain the
Kmand Vmaxvalues.25The kcatwas calculated from Vmax/[E].
3D Model Building of Wild-Type GGPPs. The coordinates of
the GGPPs dimer (aa -3 to 335), including the bound Mg2+for
each subunit, were taken from the 1.98-Å crystal structure, PDB
entry 2DH4 (the first five residues, numbered -4 to 0, is a linker
to expose the FXa cleavage site, but the coordinates of the
N-terminal residue, Met -4, were missing). For each chain, the
missing loop between residues 310 and 322 was modeled using
the MODLOOP program26,27and inserted into the 3D structure.
Hydrogen atoms were added and minimized using the CHARMM28
version 34 program with the CHARMM27 all-atom parameter set.29
At the crystallization pH of 7.5, all Asp and Glu residues were
deprotonated, Arg and Lys residues were protonated, while His
residues were protonated according to the local environment using
the Whatif program.30This resulted in a net dimer charge of -12e,
which was neutralized by adding 12 sodium counterions at the
highest electronegativity locations with the constraints that each
counterion was g5 Å from the protein surface and g9 Å from
each other. The resultant system was solvated in a truncated
octahedron containing TIP3P water molecules,31resulting in a total
of 145 617 atoms.
MD Simulations of Wild-Type and Mutant GGPPs. To relieve
any bad contacts in the solvated wild-type model structure, the water
molecules were subjected to rounds of minimization with constraints
on the protein heavy atoms. The resulting solvated system was
subjected to MD at a mean temperature of 300 K using a 1 fs time
step, periodic boundary conditions, van der Waals (vdW) interac-
tions shifted to zero at 12 Å, and electrostatic interactions treated
via the particle mesh Ewald summation method.32Initially, 50 ps
of dynamics was performed with the protein atoms and Mg2+ions
restrained by a harmonic potential, which was then reduced for all
protein atoms and removed for the Mg2+ions after 50 ps. This
was followed by another 200 ps dynamics for the modeled loop
region (residues 310-322). The final structure (without water
molecules) was used as the starting point for simulations of the
wild-type GGPPs dimer and the L8G/I9G double mutant dimer (by
mutating the L8 and I9 side-chains to Gly). The root-mean-square
deviation (rmsd) of the protein backbone in this starting structure
from that in the X-ray structure is ∼0.5 Å.
Each structure was solvated again in a truncated octahedron
containing TIP3P water molecules. Subsequently, 50 ps of dynamics
was performed allowing only water molecules to move, followed
by 50 ps with the protein backbone atoms lightly restrained, and
another 50 ps allowing all atoms in loop regions to move, and finally
900 ps of unconstrained dynamics. The backbone rmsd from the
initial structure is 2.1 ( 0.1 Å for the wild-type dimer and 2.9 (
0.1 Å for the double mutant dimer. Coordinates were saved every
0.1 ps from the final 100 ps of dynamics for a total of 1000
coordinate sets for each GGPPs structure. These were used to
compute a mean MD structure and used in the free-energy
decomposition analyses below.
Free Energy Decomposition. The dimerization free energy of
the wild-type or mutant protein (denoted by p) in solution was based
on the thermodynamic cycle: It is given by
As conformational changes upon dimerization were neglected (see
below), the gas-phase dimerization free energy, ∆Ggas, was ap-
proximated as a sum of the changes in the gas-phase vdW (∆EgasvdW)
and electrostatic (∆Egaselec) energies upon dimerization, which were
calculated using the CHARMM program and forcefield with a
dielectric constant of 1 and a nonbond cut-off of 999 Å. The
solvation free energy, Gsolv, was estimated as a sum of the
nonelectrostatic (Gsolvnonel) and electrostatic (Gsolvelec) contributions.
The Gsolvnonelwas approximated by a linear function of the solvent
accessible surface area (SASA); that is, Gsolvnonel) γ × SASA,
where γ ) 7.2 cal/mol/Å2,33,34and the absolute SASA of the protein
was computed using the CHARMM program and forcefield with a
solvent probe radius of 1.4 Å. The Gsolvelecwas estimated by the
finite-difference solution to the linearized Poisson-Boltzmann
equation implemented in the CHARMM program, using a set of
radii optimized for solvation calculations using the CHARMM
forcefield.35The contribution of an individual residue i to the
dimerization free energy can be determined from (a) ∆EgasvdW(i)
and ∆Egaselec(i), the pairwise vdW and electrostatic interactions of
residue i in chain A with all residues in chain B, respectively, (b)
Gsolvnonel(i) ) γ × SASA(i), where SASA(i) is the SASA of residue
i in the dimer/monomer, and (c) Gsolvelec(i), which can be decom-
posed because, in using the linearized Poisson-Boltzmann equation,
superposition allows the net electrostatic potential at any point to
be expressed as a sum of contributions from the atomic charges of
residue i. Thus, Gsolvelec(i) can be computed by summing over all
charges, the product of the charge and the potential at the position
of the charge due to the atomic charges from residue i in the dimer/
monomer. We refer the reader to previous works36,37for details of
the per-residue free energy calculations.
The 1000 coordinate sets saved during the final 100 ps of each
simulation were used to compute the free energies in eq 1. For
each coordinate set, the first residue (Thr -3), which was found to
be very flexible, was omitted as well as all water molecules and
counter-ions to eliminate issues with boundary sizes. Following
previous work,37the 1000 coordinate sets were sorted into 10 evenly
distributed clusters based on the ∆Egaselecbetween the two GGPPs
chains computed using a 18 Å cutoff for each coordinate set (see
(24) Pan, J. J.; Chiou, S. T.; Liang, P. H. Biochemistry 2000, 39, 10936–
10942.
(25) Eftink, M. R. Methods Enzymol. 1995, 259, 478–512.
(26) Fiser, A.; Do, R. K. G.; Sali, A. Protein Sci. 2000, 9, 1753–1773.
(27) Fiser, A.; Sali, A. Bioinformatics 2003, 18, 2500–2501.
(28) Brooks, B. R.; Bruccoleria, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983, 4, 187–217.
(29) MacKerell, J. A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. L., Jr.;
Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.;
Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos,
C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E.,
III; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.;
Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. J. Phys.
Chem. B 1998, 102, 3586–3616.
(30) Vriend, G. J. Mol. Graph. 1990, 8, 52–56.
(31) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, M. L. J. Chem. Phys. 1983, 79, 926–935.
(32) Feller, S. E.; Pastor, R. W.; Rojnuckarin, A.; Bogusz, S.; Brooks, B. R.
J. Phys. Chem. 1996, 100, 17011–17020.
(33) Sharp, K. A.; Nicholls, A.; Fine, R. F.; Honig, B. Science 1991, 252,
106–109.
(34) Jayaram, B.; McConnell, K. J.; Surjit, B. D.; Beveridge, D. L. J. Comp.
Phys. 1999, 151, 333–357.
(35) Nina, M.; Beglov, D.; Roux, B. J. Phys. Chem. B. 1997, 101, 5239–
5248.
∆Gsln(p) ) ∆Ggas(p) + Gsolv(p·p) - 2Gsolv(p)(1)
4054J. AM. CHEM. SOC. 9 VOL. 131, NO. 11, 2009
ARTICLES
Lo et al.

Page 5

Figure S1 in the Supporting Information). Within each cluster, the
number of coordinate sets was counted, and the coordinate set
whose ∆Egaselecis closest to the mean ∆Egaselecof the cluster was
chosen as the representative dimer coordinate set. As the monomer
structure for the wild-type or double mutant GGPPs is not available,
the chain A coordinates from the 10 representative dimer coordinate
sets were used for the monomer coordinate sets. Previous works
(e.g., La Font et al.37and references therein) have also used the
structure of the unbound protein from that of the respective complex.
Since the different clusters contain different numbers of coordinate
sets, the ∆Gslnvalues derived from the 10 representative coordinate
sets were weighted according to the number of coordinate sets in
each cluster to give a final weighted average ∆Gslnand respective
standard deviation. For each of the 10 clusters, the per-residue
contribution to the dimerization free energy, ∆Gsln(i), was also
computed based on the representative coordinate set and assumed
to be the same for all coordinate sets belonging to the same cluster.
The standard deviation of the 1000 ∆Gsln(i) values yields the
uncertainty/error in the calculated ∆Gsln(i).
The motivation for using the above procedure, rather than
calculating the free energies for all 1000 coordinate sets is (i) to
speed up the free energy decomposition calculations and (ii) to
ensure that the most frequently occurring configurations contribute
more to the final free energy than the rarer configurations, which
are also represented, albeit with less weight. All simulations and
free energy decomposition calculations were carried out on a parallel
cluster of Intel Core2 dual core CPUs with the GNU/Linux
operating system.
Results
Expression and Purification of Mutant GGPPs. Wild-type
and the mutant GGPPs including those lacking the first couple
of aa residues ∆(1-6), ∆(1-7), ∆(1-8), ∆(1-9), and ∆(1-17)
as well as the site-directed mutants E7G, L8G, I9G, E7G/L8G,
L8G/I9G, L163G, M167G, L200G, I203G, L163G/M167G,
L200G/I203G, N101G, D145K, E134A, R175A, and M167G/
N199A were expressed in E. coli and purified as previously
described for the wild-type.21All the mutant enzymes were
found in the soluble fraction of the cell lysate except L163G/
M167G and L200G/I203G, which formed inclusion bodies. The
Table 1. Hydrodynamic Properties of Wild-Type and Mutant
GGPPs
yeast GGPPs
major quaternary
structure
sedimentation
coefficient (S)
molecular
weight (kDa)
frictional
ratio (f/f0)
1.19
1.20
1.14
1.06
0.99
1.02
1.18
1.18
1.09
1.08
1.08
wild-type
∆(1-6)
∆(1-7)a
∆(1-8)
∆(1-9)
∆(1-17)
E7G
L8Ga
I9G
E7G/L8G
L8G/I9G
dimer
dimer
mixture
monomer
monomer
monomer
dimer
mixture
monomer
monomer
monomer
4.8
4.8
N.A.
3.5
3.5
3.5
4.8
N.A.
3.4
3.4
3.4
78
75
N.A.
38
38
37
78
N.A.
40
39
38
aFor mixture of dimer and monomer, sedimentation coefficient and
molecular weight cannot be determined (N.A.).
Figure 2. Quaternary structures of wild-type and the truncated mutants determined from AUC experiments. AUC data of (A) wild-type, (B) ∆(1-6), (C)
∆(1-7), (D) ∆(1-8), (E) ∆(1-9), and (F) ∆(1-17) GGPPs are shown. The molecular mass of each mutant protein was deduced from the sedimentation
velocity data. From the data, wild-type is a dimer, ∆(1-6) is a mixture of dimer and monomer, whereas ∆(1-8), ∆(1-9), and ∆(1-17) are monomer.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 11, 2009
4055
Dimerization of Geranylgeranyl Diphosphate Synthase
ARTICLES