Crystal Structure of Methanobacterium thermoautotrophicum Phosphoribosyl-AMP Cyclohydrolase HisI † , ‡

Abstract

The metabolic pathway for histidine biosynthesis is interesting from an evolutionary perspective because of the diversity of gene organizations and protein structures involved. Hydrolysis of phosphoribosyl-AMP, the third step in the histidine biosynthetic pathway, is carried out by PR-AMP cyclohydrolase, the product of the hisI gene. The three-dimensional structure of PR-AMP cyclohydrolase from Methanobacterium thermoautotrophicum was solved and refined to 1.7 A resolution. The enzyme is a homodimer. The position of the Zn(2+)-binding site that is essential for catalysis was inferred from the positions of bound Cd(2+) ions, which were part of the crystallization medium. These metal binding sites include three cysteine ligands, two from one monomer and the third from the second monomer. The enzyme remains active when Cd(2+) is substituted for Zn(2+). The likely binding site for Mg(2+), also necessary for activity in a homologous cyclohydrolase, was also inferred from Cd(2+) positions and is comprised of aspartic acid side chains. The putative substrate-binding cleft is formed at the interface between the two monomers of the dimer. This fact, combined with the localization of the Zn(2+)-binding site, indicates that the enzyme is an obligate dimer.

Full text

Crystal Structure of Methanobacterium thermoautotrophicum Phosphoribosyl-AMP
Cyclohydrolase HisI†,‡
J. Sivaraman,|,§,¶,+Rebecca S. Myers,⊥Lorena Boju,|,§ Traian Sulea,|Miroslaw Cygler,|,§,¶ V. Jo Davisson,⊥and
Joseph D. Schrag*,|,§
Biotechnology Research Institute, National Research Council, Montreal, Quebec H4P 2R2, Canada,
Montreal Joint Center for Structural Biology, Montreal, Quebec H4P 2R2, Canada, Department of Biochemistry,
McGill UniVersity, Montreal, Quebec, Canada, and Department of Medicinal Chemistry and Molecular Pharmacology,
Purdue UniVersity, West Lafayette, Indiana 47907
ReceiVed March 14, 2005; ReVised Manuscript ReceiVed June 1, 2005
ABSTRACT: The metabolic pathway for histidine biosynthesis is interesting from an evolutionary perspective
because of the diversity of gene organizations and protein structures involved. Hydrolysis of phosphoribosyl-
AMP, the third step in the histidine biosynthetic pathway, is carried out by PR-AMP cyclohydrolase, the
product of the hisI gene. The three-dimensional structure of PR-AMP cyclohydrolase from Methanobac-
terium thermoautotrophicum was solved and refined to 1.7 Å resolution. The enzyme is a homodimer.
The position of the Zn2+-binding site that is essential for catalysis was inferred from the positions of
bound Cd2+ions, which were part of the crystallization medium. These metal binding sites include three
cysteine ligands, two from one monomer and the third from the second monomer. The enzyme remains
active when Cd2+is substituted for Zn2+. The likely binding site for Mg2+, also necessary for activity in
a homologous cyclohydrolase, was also inferred from Cd2+positions and is comprised of aspartic acid
side chains. The putative substrate-binding cleft is formed at the interface between the two monomers of
the dimer. This fact, combined with the localization of the Zn2+-binding site, indicates that the enzyme
is an obligate dimer.
Histidine biosynthesis is linked to nitrogen metabolism at
the level of both precursors and intermediates in the pathway,
and histidine is the only amino acid that derives part of its
structural carbon from a nucleotide (1). In the biosynthetic
pathway, ATP is utilized as a substrate and, following
ribosylation of the N1-position and dephosphorylation, the
resultant N1-phosphoribosyl-AMP (PR-AMP)1undergoes a
ring opening reaction catalyzed by PR-AMP cyclohydrolase
coded by the hisI gene (Figure 1). This overall transformation
renders a carbon and nitrogen from AMP available for the
subsequent incorporation into the imidazole ring of histidine.
While its role in histidine metabolism has been known
for many years, the PR-AMP cyclohydrolase chemistry draws
mechanistic analogies to the general class of purine/pyrimi-
dine deaminases. The PR-AMP cyclohydrolase hydrolyzes
the adenine ring of phosphoribosyl-AMP between N1and
C6(Figure 1), while for enzymes such as adenosine deami-
nase, the hydrolysis releases ammonia. Enzymes such as
adenosine deaminase are characterized by using zinc-
activated water to effect a nucleophilic attack. Previous
studies of the PR-AMP cyclohydrolase have confirmed that
its native function requires a zinc ion and classified the
protein as a metalloenzyme (2). Despite the similarity in the
overall reactions, the ligands and coordination of the catalytic
zinc vary among the deaminases, making the overall
mechanistic comparisons between PR-AMP cyclohydrolases
ambiguous. For instance, in adenine deaminase, the catalytic
zinc is stabilized by three histidines and an aspartic acid in
an octahedral coordination (3), metal ligands in cytidine
deaminase consist of three cysteines in a tetrahedral coor-
dination (4), and in yeast cytosine deaminase, the catalytic
zinc is bound by two cysteines and a histidine (5).
Histidine biosynthesis is one of the most interesting
metabolic pathways from an evolutionary perspective be-
cause of the diversity of gene organizations and protein
structures. An example of this diversity can be seen in the
gene structure of hisI. In many species, the gene exists as a
fusion with hisE. Yet, there are cases among protobacteria
and archaea where each gene is expressed as a monofunc-
tional protein, prompting the overall proposal that hisI and
hisE began as separate genes that underwent fusion in later
branches of evolution (6). Comparison of DNA sequences
of the hisI genes from the eukaryote Saccharomyces cer-
†Funding from CIHR Grant 200103GSP-90094-GMX-CFAA-19924
(to M.C.), the National Institutes of Health Grant GM067915 (to V.J.D.),
and funding of the Ontario Center for Structural Proteomics from the
Ontario Research and Development Challenge Fund is gratefully
acknowledged.
‡Coordinates and structure factors have been deposited in the Protein
Data Bank with code 1ZPS.
* Corresponding author: Dr. Joseph D. Schrag, Biotechnology
Research Institute, National Research Council, 6100 Royalmount
Avenue, Montreal, Quebec H4P 2R2, Canada. E-mail, Joe.Schrag@
nrc-cnrc.gc.ca; tel, (514)496-2557; fax, (514)496-5143.
|Biotechnology Research Institute, National Research Council.
§Montreal Joint Center for Structural Biology.
¶McGill University.
⊥Purdue University.
+Current address: Department of Biological Sciences, National
University of Singapore, Singapore.
1Abbreviations: PR-AMP, N1-(5′-phosphoribosyl)adenosine 5′-
monophosphate; TCEP, tris(2-carboxyethyl)phosphine; EDTA, (ethyl-
ene-diamine) tetraacetic acid.
10071Biochemistry 2005, 44, 10071-10080
10.1021/bi050472w CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/07/2005

eVisiae, the archaebacterium Methanococcus Vannielii, and
the eubacterium Escherichia coli indicates that these genes
are derived from a common ancestor; 27% of all nucleotides
are identical across the three species with conservation of
amino acids at 39% (7). These observations suggest that the
HisI function in the histidine biosynthetic pathway may have
developed prior to the divergence of these biological
kingdoms. However, neither a mechanistic consequence
regarding the HisI function nor its relationship with other
deaminases is apparent from these comparisons.
Biochemical characterization of the PR-AMP cyclohy-
drolase from M. Vannielii (2) demonstrated the importance
of both Zn2+and Mg2+metals in the catalytic mechanism.
The presence of Zn2+in the lysis buffer enhances the specific
activity and stability of the enzyme, and 1 equiv of bound
Zn2+could be removed only by extensive dialysis with
1,10-phenanthroline. We have proposed that a unique
sequence motif, C93(X)15C109H110(X)5C116, represents the Zn2+
binding domain (2). Additionally, the M. Vannielii enzyme
was reversibly inhibited by inclusion of EDTA in the assay
mixture, demonstrating that Mg2+is also required for
catalytic activity. The consensus sequence motif S67R[S/T]-
RXX[L/I]WXKG[E/A]TSG81 was suggested to constitute a
“P-loop type” region associated with Mg2+and phosphate
binding (2).
We report here the crystal structure of the phosphoribosyl
cyclohydrolase, HisI, from the archaea Methanobacterium
thermoautotrophicum. Analysis of the three-dimensional
structure in light of the biochemical characterization of the
homologous M. Vannielii enzyme identifies the active site
and the likely Zn2+- and Mg2+-binding sites. The location
of these sites implies that dimerization is required for HisI
function and that a functional relationship exists between
the active sites of the PR-AMP cyclohydrolase and cytosine
deaminase.
MATERIAL AND METHODS
Materials. With the exception of the substrate PR-AMP,
all chemicals and buffers were purchased from commercial
sources. Synthesis of PR-AMP was performed as previously
described (2).
Cloning. A pET15b plasmid, MT245, harboring the M.
thermoautotrophicum hisI gene was kindly provided by the
Ontario Center for Structural Proteomics, University of
Toronto. The His6-tagged protein produced from this vector
consistently aggregated during purification. The gene was
excised from the plasmid using BamHI and NdeI and was
inserted into a pET11a vector (Novagen, Darmstedt, Ger-
many) lacking a His tag. E. coli MC1061 cells were
transformed with the resulting plasmid, pLB1. PCR screening
identified colonies containing the insert. The plasmid was
purified from one colony designated pLB1-2 and was used
to transform E. coli BL21(DE3) Gold Magik cells. Cells
were plated on Luria-Bertani (LB) plates containing kana-
mycin and ampicillin. A glycerol stock was prepared from
a single colony after a 15 h culture at 37 °Cin5mLofLB
medium.
Expression and Purification. A50mLE. coli culture
prepared from a glycerol stock was incubated overnight at
37 °C and then transferred to1LofCircle Grow medium
(Bio 101, Inc., Carlsbad, CA) containing ampicillin (0.05
mg/mL) and kanamycin (0.1 mg/mL). After2hofgrowth
to an OD600 nm of 0.9, the production of the enzyme was
induced by the addition of IPTG to a final concentration of
1 mM. The culture was incubated for 20 h at room
temperature with shaking (250 rpm). Cells were harvested
by centrifugation at 4000gfor 20 min. The cells were lysed
by sonication in lysis buffer containing 50 mM Tris-HCl (pH
7.5), 0.5 M NaCl, 0.5 mM ZnCl2, 0.1% (w/v) Triton X-100,
5% (w/v) glycerol, 20 mM β-mercaptoethanol (BME), and
one tablet of CompleteC-EDTA free protease inhibitors
(Pharmacia, Piscataway, NJ). The cell lysate was cleared by
centrifugation (150 000g, 30 min, 4 °C), heated to 85 °C
for 15 min with gentle agitation, and centrifuged at 3000g
for 20 min to remove precipitated protein. The supernatant
was loaded onto a Mono Q HR 5/5 column (Pharmacia,
Piscataway, NJ) equilibrated with 50 mM Tris-HCl (pH 7.5)
and5mMβ-mercaptoethanol and washed with 10 column
volumes of equilibration buffer. Protein was eluted with a
20 column volumes linear gradient from 0 to 0.6 M NaCl in
50 mM Tris-HCl, pH 7.5. HisI-containing fractions were
pooled, and the protein was dialyzed against 50 mM Tris-
HCl (pH 7.5), 200 mM NaCl, 0.5 mM ZnCl2,and5mM
β-mercaptoethanol. The dialyzed protein was concentrated
to 10 mg/mL using a Centricon YM-30 ultrafiltration unit
(Millipore, Bellerica, MA) and was stored at 4 °C.
Dynamic Light Scattering. Dynamic light scattering
measurements were performed at room temperature on a
DynaPro-801 (Protein Solutions, Charlottesville, VA) instru-
ment. The sample was filtered through a 100 µm filter for
analysis.
Crystallization. Crystallization experiments were carried
out at room temperature using the hanging-drop, vapor-
diffusion method. Initial crystallization screening was per-
formed using the Hampton Research (Laguna Niguel, CA)
sparse-matrix screens I and II. After optimization of crystal-
lization conditions, the best crystals were produced by mixing
the protein solution at a concentration of 2 mg/mL in a 1:1
ratio with a reservoir solution containing 100 mM HEPES-
HCl (pH 7.5), 50 mM CdSO4, 1.6 M sodium acetate, and
10% (w/v) glycerol.
Single crystals were picked up directly from the drop using
nylon loops and were flash-cooled in a N2gas cold stream
at 100 K (Oxford Cryosystems, Oxford, U.K.). Diffraction
data were collected at beamline X8C at the National
Synchrotron Light Source, Brookhaven National Laboratory.
Data reduction was done using HKL2000 (8). The data used
for refinement were collected at a wavelength of 1.009 1 Å.
A total of 180 frames were collected with an oscillation angle
of 1°per frame. Data used for phasing were collected at a
wavelength of 1.282 Å, the Zn absorption maximum. Data
statistics are reported in Table 1.
FIGURE 1: Reaction catalyzed by phosphoribosyl-AMP cyclohy-
drolase HisI.
10072 Biochemistry, Vol. 44, No. 30, 2005 Sivaraman et al.

Structure Solution and Refinement. X-ray fluorescence
measurements indicated the presence of Zn in the crystals
flash-cooled directly from the drop. Initial phases were
determined by SAD methods from the data collected at the
Zn absorption edge using SOLVE/RESOLVE (9,10). The
partial model obtained accounted for 45% of the expected
residues. The model was improved by iterative cycles of
manual rebuilding using O (11), followed by refinement of
the model against a higher resolution “native” dataset using
CNS (12). The model was refined against data with Bijvoet
pairs unmerged because a significant anomalous signal was
apparent in the “native” data. Refinement included a bulk
solvent correction and an anisotropic overall B-factor refine-
ment. Noncrystallographic symmetry restraints were em-
ployed during the early refinements, but were not used in
later stages and for refinement of the final model. No σcutoff
was used during refinement. Validation of the model was
done using PROCHECK (13).
The identities of the heavy atoms contributing to the
anomalous signal were assigned by comparison of peak
heights in anomalous differences maps calculated using
phases from the refined model. All of the heavy atoms were
assigned as Cd2+ions because the relative peak heights in
the anomalous maps were identical in maps calculated from
data collected at the Zn absorption maximum (1.282 Å, ∆f′′)
5.8) and at the remote wavelength of 1.494 Å, where the Zn
anomalous contribution is very low (∆f′′ )0.64). The Cd2+
anomalous contribution, in contrast, is significant at both of
these wavelengths (∆f′′ )3.4 and 4.4, respectively). As-
signment as Cd2+is consistent with the high concentration
(50 mM) of CdSO4being essential for crystallization. This
interpretation explains why three-wavelength MAD phasing
around the Zn absorption edge was unsuccessful. Since all
the bound metal ions are Cd2+ions, the data for all three
wavelengths contains approximately equal anomalous con-
tributions and there are no differences that contribute to phase
information.
Docking of a Tetrahedral Intermediate. A model of 6-OH
PR-AMP, a possible transition state in the reaction catalyzed
by HisI, was docked into one of the two putative active site
clefts of the HisI homodimer. The ligand was first manually
positioned in two independent orientations that differed by
an approximate 2-fold rotation that relates the two ribose
phosphate groups on either end of the substrate’s adenine
ring. In both orientations, the 6-OH oxygen atom was placed
within coordination distance of the Cd2+ion representing
the putative catalytic Zn2+ion. Another Cd2+ion that likely
represents a bound Mg2+ion, as well as two water molecules
buried at the bottom of the binding site cleft, is also retained
from the crystal structure.
For each orientation of 6-OH PR-AMP, bound conforma-
tions were modeled by a search using a Monte Carlo
minimization (MCM) procedure (14-16). In each MCM
cycle, a starting conformation was generated by randomly
perturbing one or more dihedral angles of the ligand, with
the 6-OH oxygen as an anchor atom. This starting conforma-
tion then was subjected to an energy minimization in which
the ligand, ions, and water, as well as a predefined set of
protein residues extending 8 Å around the ligand, were
allowed to relax up to an rms gradient of 0.05 kcal/(mol‚Å).
Three thousand MCM cycles were carried out to sample six
rotatable bonds of the bound ligand (two rotors between the
purine moiety and the ribose rings, plus four rotors between
the ribose rings and their phosphate groups). The lowest
energy ligand conformation was re-docked into the initial
(crystallographic) protein geometry followed by an energy
minimization in the same binding site region up to an rms
gradient of 0.01 kcal/(mol‚Å), thus leading to a final model
for each binding orientation.
Conjugate gradient energy minimizations were carried
using the AMBER all-atom molecular mechanics force field
with the PARM94 parameter set for proteins (17) and the
GAFF parameter set for organic compounds (18). A distance
dependent (4rij) dielectric and an 8 Å nonbonded cutoff were
employed. The protonation state at physiological pH was
adopted with the exception of the Zn-chelating Cys resi-
dues, which were ionized. Partial atomic charges of 6-OH
PR-AMP were obtained by a two-stage-restrained fitting
procedure to the single-point HF 6-31G* electrostatic
potential (17,19). AMBER-compatible, charge-delocalized
models with tetrahedral and octahedral geometries were used
for the Zn2+and Mg2+ions, respectively (20). During all
energy minimizations, the distance between the 6-OH oxygen
atom of the transition state and the Zn2+ion was constrained
within the 2.0-2.5 Å range with a harmonic potential of
200 kcal/(mol‚Å2).
PR-AMP Cyclohydrolase Assay and Kinetic Constants.
Protein concentrations were determined by total protein
amino acid analysis (∆A280/mg 1.14). The cyclohydrolase
assay followed the established procedure (2). Briefly, routine
assays were conducted at 30 or 65 °CinaUV-vis
spectrophotometer monitoring changes in absorbance at 300
Table 1: Data Collection and Refinement Statisticsa
phasing
λ)1.282 Å refinementa
λ)1.0091 Å
Data Collection
unit cell
a(Å) 39.7 39.8
b(Å) 54.3 54.4
c(Å) 117.3 117.4
resolution range (Å) 50.0-1.86 50.0-1.70
wavelength (Å) 1.2823 1.0091
observed hkl 138718 157479
unique hkl 37937 47949(3809)
completeness (%) 91.8 91.3(70.3)
overall I/σI12.9 40.5(5.5)
Rsym (%) 4.4 3.3(12.6)
Refinement
resolution range (Å) 50.0-1.70
Rfree (%) no. reflections 23.5 (2338)
Rwork (%) no. reflections 20.7 (45193)
rmsd bond lengths (Å) 0.01
rmsd bond angles (deg) 1.9
mean coordinate error (Å) 0.20
Ramachandran Plot
favored region (%) 93.0
allowed region (%) 7.0
generously allowed (%) 0.0
disallowed region (%) 0.0
Average B-factors (Å2)
main chain atoms (no. atoms) 24.5 (1010)
side chain atoms (no. atoms) 27.6 (975)
overall protein atoms (no. atoms) 26.1 (1985)
waters (no. atoms) 29.5 (135)
metal ions (no. atoms) 32.1 (19)
acetate ions (no. atoms) 24.9 (20)
aHighest resolution shell in () )1.76-1.70 Å.
Crystal Structure of HisI Biochemistry, Vol. 44, No. 30, 2005 10073

nm (∆300 nm 6700 M-1cm-1) or 260 nm (∆260 nm 8020
M-1cm-1). Each assay contained 50 mM Tris-HCl (pH 7.5),
1 mM EDTA, 5 mM MgCl2, and 50 µM PR-AMP in a total
volume of 1 mL. The reaction was initiated by the addition
of 0.1-1µg of PR-AMP cyclohydrolase. Steady-state kinetic
constants were determined by varying the concentration of
PR-AMP from 5 to 150 µM. The pH rate profiles were
carried out in 30 mM Bis-Tris Propane-HCl (pH 6.5, 7.0,
7.5, 8.0, 8.5, 9.0) at 65 °C. Substrate saturation curves at
each pH were determined from six substrate concentrations
over the range of 2.5-100 µM PR-AMP with observation
at 260 nm.
Metal Exchange and Protein-Metal Ion Contents. All
buffers used for dilution and desalting of protein samples
were prepared in metal-free glassware following treatment
with Chelex (1,21). A 0.5 mL aliquot of M. thermoau-
totrophicum PR-AMP cyclohydrolase prepared as above was
dialyzed at 1 mg mL-1against 3×1 L of 0.1 mM
CdSO4to directly exchange the active site zinc with
cadmium. Verification of metal exchange was accomplished
by ICP-MS on the sample, measuring both zinc and
cadmium. All samples of PR-AMP cyclohydrolases were
desalted using Sephacryl S-200 HR MicroSpin columns.
Each sample was passed through two MicroSpin columns
before dilution into 2% HNO3. A calibration curve was
generated using standard addition of the metal of interest to
control against matrix effects. Briefly, an increasing known
amount of metal was added to each sample. Each sample
was then analyzed by ICP-MS typically using protein at 10
ppb. The resulting values were then plotted in Cartesian
fashion, with xand ycoordinates corresponding to the known
concentration of added metal versus the ICP-MS measure-
ment for that sample, respectively. The amount of metal in
the original sample was determined by the x-intercept of this
plot. ICP-MS samples were analyzed at the Campus-wide
Mass Spectrometry Center at Purdue University.
RESULTS
The molecular weight of the M. thermoautotrophicum HisI
monomer is 15 458 Da. SDS-PAGE analysis of the purified
protein under reducing conditions revealed two bands, one
corresponding to a monomer and the other to a dimer. Native-
PAGE showed a single band corresponding to a dimer. Con-
sistent with the presence of a dimer, light-scattering studies
indicated the presence of a single species with a molecular
weight of ∼37 kDa. Although the cysteine residues in the
protein are located at the dimer interface, they form metal-
binding sites and no intermolecular disulfide bonds are ob-
served. The observation of dimers in the SDS gels is, there-
fore, not likely to result from intermolecular disulfide bonds.
Crystallization and Phasing. Crystallization of HisI re-
quired high concentrations (50 mM) of Cd2+ions. The
crystals belong to the orthorhombic space group P212121,
with unit-cell parameters a)39.8, b)54.4, c)117.4 Å.
The crystals routinely diffract to 2 Å resolution and the best
crystals diffract to 1.7 Å resolution. A dimer constitutes the
asymmetric unit, and the solvent content is approximately
40%.
Zinc is required for enzymatic activity of M. Vannielii
PR-AMP cyclohydrolase (2) and presumably for all orthologs.
We, therefore, included 0.5 mM ZnCl2throughout the
purification and crystallization of M. thermoautotrophicum
HisI to stabilize the protein. We anticipated that bound Zn
atoms would provide a convenient source of phase informa-
tion. X-ray fluorescence measurements on flash-cooled
crystals showed the presence of Zn, and phases were obtained
by SAD methods from data collected at the Zn absorption
maximum. However, the high concentration of Cd2+in the
crystallization solution (50 mM) provided another potential
source of anomalous scatterers that could contribute some
of the observed bound metal ions. Anomalous difference
maps were used to distinguish between potential Cd2+and
Zn2+atoms. The peaks observed in anomalous difference
maps, calculated from datasets collected at the Zn absorption
maximum where the Zn contribution to the anomalous signal
is strong and at a remote wavelength, where the anomalous
contribution from Zn is very small, are nearly identical (see
Materials and Methods for details). This observation suggests
that none of the heavy atoms contributing to the anomalous
signal are Zn2+ions. All of the metal ions observed in the
electron density are, therefore, interpreted as Cd2+ions.
OVerall Structure. The asymmetric unit of the crystal
comprises a dimer. Chain A of the refined model includes
residues 8-131, and chain B includes residues 4-131. The
model also includes 19 Cd2+ions, 5 acetate ions and 129
water molecules and has been refined to an R-factor of 0.209
and Rfree of 0.236. The mean temperature factors reported in
Table 1 demonstrate that chain B is generally better ordered
than is chain A. The N-terminal 20 residues of chain A, in
particular, are poorly ordered, and only fragmented electron
density is observed for residues 16-21. Electron density for
the long side chains (Lys, Arg, Glu) are poorly defined for
many residues in chain A. The loop encompassing residues
115-120, however, is better ordered in chain A than in chain
B. Data statistics and the quality of the model are indicated
in Table 1.
Each monomer is composed of four β-hairpins (Figure
2A). On one side of the monomer, hairpins 1 (residues
23-42) and 3 (residues 76-98) pack in an antiparallel man-
ner, forming a four-stranded mixed β-sheet. Packed against
one face of this β-sheet are hairpin 2 (residues 54-69) and
the helix that is inserted between hairpins 1 and 2.
The overall dimensions of the dimer in the asymmetric
unit of the crystal are 40 ×50 ×53 Å3. The root-mean-
square (rms) deviation in the positions of CRatoms in the
two monomers is 0.39 Å. The dimer interface is formed by
interactions of hairpins 1, 3, and 4 of each monomer (Figure
2B). The β-sheets formed by hairpins 1 and 3 pack face to
face, primarily by interactions of hydrophobic residues, with
a large percentage being the branched residues valine,
isoleucine, and leucine. Hairpin 4 (residues 111-122) packs
against hairpin 3 of the other monomer within the dimer via
hydrogen bonding characteristic of parallel β-structure. This
domain swapping, in effect, extends the mixed β-sheet from
four strands to six strands. The total surface area buried in
the dimer interface is ∼3100 Å2, more than 40% of the total
molecular surface area of each monomer.
The most conspicuous features of the HisI model are the
two surface clefts that are formed at the interface between
the two molecules and are related by the dimer 2-fold
symmetry (Figure 3A). Four Cd2+ions are bound in each of
these surface clefts (Figure 3B). One of these metal ions,
Cd2+(1), is coordinated by the side chains of three cysteines.
10074 Biochemistry, Vol. 44, No. 30, 2005 Sivaraman et al.

Cys86 comes from one monomer, whereas Cys102 and
Cys109 come from the other monomer. The tetrahedral coor-
dination is completed by the oxygen atom from an acetate
molecule or by a water molecule (Figure 3C). Adjacent to
these cysteines is a cluster of three aspartic acid residues.
The carboxylate oxygen atoms (residues 85, 87, and 89 in
the loop between the β-strands of hairpin 3) and several water
molecules octahedrally coordinate another cadmium ion,
Cd2+(2), bound in the cleft (Figure 3C). A third Cd2+ion is
bound between the acidic and cysteine clusters and is
coordinated by Asp85 and Cys86 from one monomer,
Cys109 from the other monomer, and solvent molecules. The
fourth Cd2+ion is bound to residues His16 and Asp89.
Comparisons of 98 amino acid sequences of HisI orthologs
from various prokaryotic, archeal, and eukaryotic organisms
indicate that HisI was well-conserved during evolution.
Invariant residues in all these sequences are Ser60, Lys69,
Gly70, Asp87, Cys102, His103, and Cys109. Trp67 is
substituted by arginine in the sequence from Pyrococcus
furiosus, but the alkyls chain of arginine can also make
FIGURE 2: Ribbon diagram of HisI. (A) Stereoview of HisI monomer. β-hairpins 1-4 are colored yellow, red, green, and blue, respectively.
(B) Sequence alignment of M. thermoautotrophicum and M. Vannielii HisI. Secondary structure observed in the M. thermoautotrophicum
enzyme is indicated. The β-strands are colored according to the hairpins to which they contribute, with hairpin colors as in panel A. The
putative P-loop motif is marked in gray, and the Zn2+-binding cysteine residues are marked in black. This figure was smade using ALSCRIPT
(22). (C) HisI dimer viewed approximately down the noncrystallographic symmetry axis. Chain A is colored gold, and Chain B is colored
as in panel A. Note domain swapping of hairpin 4. This figure was made using PyMol (www.pymol.org).
Crystal Structure of HisI Biochemistry, Vol. 44, No. 30, 2005 10075

hydrophobic contacts. Asp85, Cys86, and Asp89 are not
conserved in the putative HisI sequence from Leptospira
interrogans (Swiss-Prot entry Q8F8K4), but the catalytic
competence of this protein is uncertain. These invariant and
highly conserved residues of HisI map to the cleft described
above (Figure 3A). Residues 60-70 (one β-strand of hairpin
2) and 102-109 (connection between hairpin 3 and 4) from
one monomer are at one end of the cleft, and residues 85-
89 (the loop between the two β-strands of hairpin 3) from
the other monomer are on the side of the cleft. The conserved
residues include the metal ion binding aspartates and
cysteines.
We were unable to crystallize HisI in the absence of Cd2+,
and even reduction of the Cd2+concentration destabilized
the crystals. This requirement reflects the Cd2+ions partici-
pating in intermolecular contacts crucial for crystal formation.
Two of the most prominent features in electron density maps
correspond to the Cd2+ions bound between symmetry related
molecules. The first metal atom is coordinated by Glu52B
from one dimer and by His103A and Cys86B from the
symmetry related dimer. The second is bound by Cys86A
and His103B from one dimer and by Glu123B from the
symmetry related dimer.
A DALI search using the HisI monomer as a search model
identified several similar proteins with Zscores ranging from
3.6 to 4.3. The highest score was for the N-terminal
β-sandwich domain of the subunit of the E. coli proton-
translocating ATP synthase (PDB code: 1E79 (23,24)).
These two domains superimpose with an rms deviation of
2.7 Å for 60 CRatoms. The -subunit inhibits the ATPase
activity in isolated F1-ATPase and is essential for the
coupling of proton translocation to ATP synthesis. The ATP-
binding site of the ATPase is, however, far from the
-subunit, and the significance of the structural similarity
between HisI and the ATPase subunit is not clear.
HisI is also structurally similar to the N-terminal domain
of the Thermus thermophilus ribosomal protein TL5 (PDB
code 1FEU (25)) and to the large subunit ribosomal protein
FIGURE 3: CONTINUED
10076 Biochemistry, Vol. 44, No. 30, 2005 Sivaraman et al.

L25 from E. coli (PDB code 1DFU (26)). The HisI model
superimposes on TL5 with an rms fit of 3.0 Å for 63 CR
atoms and on L25 with an rms fit of 3.3 Å for 58 CRatoms.
TL5 and L25 are functional analogues (27), both binding
the 5S rRNA. The rRNA-binding site of these proteins is
on the face of the molecule that forms the dimer interface
in HisI. Therefore, little functional similarity between HisI
and TL5/L25 is apparent.
PR-AMP Cyclohydrolase Kinetic Constants and pH Pro-
file. The specific activity of the purified protein was first
FIGURE 3: (A) Stereoview of the surface of the putative substrate-binding cleft showing the positions of conserved residues and Cd2+(1)
and Cd2+(2) ions in the cleft. Residues in cyan derive from chain A, and those in white derive from chain B. Asp and Cys residues in the
cleft are red and yellow, respectively. The blue (residues 60-74) and green (H103) residues are residues which are highly conserved in
HisI proteins from a wide variety of organisms. (B) Stereoview of the residues forming the cleft at the dimer interface. The oxygens are
colored in red, nitrogens in blue, Cd2+in orange, carbons of monomer A in cyan, and those of monomer B in white. (C) Coordination of
the putative metal binding sites. The site of Cd2+(1) has a tetrahedral coordination and is most likely the Zn2+-binding site (marked Zn in
the figure). The site of Cd2+(2) has six ligands and is most likely the Mg2+-binding site (marked Mg in the figure). Water molecules are
shown as red spheres. Coloring of atoms in monomers A and B are like in panel B. (D) The lowest energy conformer of the putative
transition state of PR-AMP in the cleft. Close-up of the active site of HisI with the model of the transition state bound to the enzyme.
Carbon atoms from residues in chain A are in cyan, and those from chain B are white. Hydrogen bonds and metal coordination bonds are
represented by black and green dashed lines, respectively. A schematic representation of the putative transition state is also shown.
Crystal Structure of HisI Biochemistry, Vol. 44, No. 30, 2005 10077

analyzed at 30 °C and pH 7.5 following the protocol
established for PR-AMP cyclohydrolase from M. Vannielii.
The measured specific activity (5.3 U/mg) was less than half
that of M. Vannielii PR-AMP cyclohydrolase (12 U/mg).
Therefore, the assay was repeated at the optimal growth
temperature of M. thermoautotrophicum (65 °C). The specific
activity at 65 °C was 24 U/mg, compared to 35 U/mg for
M. Vannielii PR-AMP cyclohydrolase (Table 2). The steady-
state kinetic constants were determined for M. thermo-
autotrophicum HisI at 65 °C and pH 7.5 using the established
protocols (2). The resulting constants, Km)14 (3µM,
kcat )8s
-1(kcat/Km)6×105M-1s-1) are in good
agreement with the constants determined for M. Vannielii
PR-AMP cyclohydrolase (2), Km)9.9 (1.7 µM, kcat )
4.1 (0.3 s-1(kcat/Km)4.1 ×105M-1s-1). The kcat/KmpH
profile of M. thermoautotrophicum HisI indicates a single
pKaat 7.5 (0.2 (Figure 4), in good agreement with M.
Vannielii PR-AMP cyclohydrolase (7.3 (0.1) (2). The
greater drop in M. thermoautotrphicum activity in the basic
range could be due to the difference in the pI between these
orthologs (M. thermoautotrophicum HisI pI )5.2, M.
Vannielii HisI pI )6.3).
Cadmium-Containing PR-AMP Cyclohydrolase Steady-
State Kinetic Analysis. The crystal structure of M. thermo-
autotrophicum HisI was solved in the presence of 50 mM
cadmium, and consequently, this ion replaced the Zn2+in
the active site of this zinc metalloenzyme. To determine if
the crystal structure represents the active form of the enzyme,
direct metal exchange through extensive dialysis was per-
formed to substitute cadmium for zinc. The metal content
of various preparations of the M. thermoautotrophicum HisI
were analyzed by ICP-MS. Protein samples analyzed after
the initial purification from E. coli in the presence of
2-mercaptoethanol and ZnCl2were found to have a 10:1
molar ratio of Zn2+to protein (Table 3). Cadmium, in the
original preparation, was measured at 1/50 000 of the zinc
content. These protein samples were then dialyzed against
CdSO4-containing buffer in the absence of 2-mercaptoethanol
and desalted through gel filtration before analysis by ICP-
MS. As indicated in Table 3, Cd2+substitution by direct
dialysis resulted in 0.95 Cd2+per subunit, with a residual of
0.3 Zn2+equiv coordinated with the protein. All other metals
were below the detection limits of the analysis. The resulting
kinetic constants for the Cd2+-containing enzyme, Km)2.9
(0.8 µM, kcat )0.21 s-1(kcat/Km)7×104M-1s-1),
indicate that the activity of the enzyme is comparable to that
of zinc-containing PR-AMP cyclohydrolase, with kcat/Km
affected by a 10-fold decrease.
DISCUSSION
The histidine biosynthetic pathway is highly conserved
across a wide variety of organisms. While in some organisms
hisI is a unique gene and in others a fusion with hisE creates
a bifunctional enzyme, its amino acid sequence is highly
conserved. The most highly conserved residues surround the
surface cleft formed at the dimer interface that defines the
putative catalytic site of the enzyme.
The steady-state kinetic and pH analyses of M. thermo-
autotrophicum HisI are consistent with the results from M.
Vannielii PR-AMP cyclohydrolase and indicate that the
crystallized protein is enzymatically competent. The pH
profile of kcat/Kmis nearly bell-shaped, although the data are
not adequate on the higher pH side to assess a second pKa.
On the low-pH side of the curve, the commitment factor
occurs at the pKathat is consistent with a cysteinyl side chain.
Consistent with the suggestion that a protein side chain is
being reflected in the pH dependence of the kcat/Kmis the
fact that the Cd2+-substituted enzyme behaved in a similar
manner. The substitution of cadmium for the active site zinc
caused a greater decrease in the commitment factor at higher
pH. However, the pH dependence of the protein reaction
kinetics is essentially unchanged, with the pKaof 7.55 for
the zinc-containing protein versus 7.54 with cadmium.
Direct metal exchange experiments to replace the active
site zinc with cadmium and subsequent kinetic analyses are
consistent with the interpretation of the HisI structure
depicting an active form of the enzyme. After desalting each
ortholog using Sephacryl S-200 spin columns, the Zn2+
content of the M. thermoautotrophicum HisI was 10 times
the amount determined for the ortholog from M. Vannielii
(Table 3). The amount of zinc associated with M. thermo-
autotrophicum HisI is consistent with the large number of
Cd2+ions observed in the crystal structure. Most of the metal-
binding residues are conserved in M. Vannielii HisI, yet
analysis of this isoform consistently results in a 1:1 molar
ratio of metal-to-protein. The larger number of Zn2+associ-
ated with the enzyme preparation from E. coli may reflect a
difference in the net charge of these two enzymes. Whereas
M. Vannielii has a total charge of -1atpH7,theM.
thermoautotrophicum enzyme has a charge of -7, thus
Table 2: Specific Activity as a Function of Temperature
30 °C
(µmol/min)/mg 65 °C
(µmol/min)/mg
M. Vannielii His I 12 35
M. thermoautotrophicum His I 5.3 24
FIGURE 4: pH profile of M. thermoautotrophicum PR-AMP
cyclohydrolase; 9, zinc-containing protein, pKa)7.55; b,
cadmium-containing protein, pKa)7.54.
Table 3: Summary of Metal Ion Analysesa
mol/subunit Zn2+Cd2+
M. thermoautotrophicum
original preparation 10/10 -
M. thermoautotrophicum
dialyzed against Cd2+
0.3 0.95/0.97
M. Vannielii His I 1/1 -
aDesalted using S-200 sepharose spin column. When two measure-
ments were made, each result is reported.
10078 Biochemistry, Vol. 44, No. 30, 2005 Sivaraman et al.

requiring a larger number of counterions. Extensive dialysis
of M. thermoautotrophicum HisI to replace zinc with
cadmium resulted in a single molar equivalent of the
replacement metal. Cadmium substitution does render the
enzyme active, although not all of the zinc was displaced.
The structure of the crystallized protein does provide a basis
for evaluating the structural features of the enzyme active
site. The catalytically unique features of the Cd2+-sub-
stituted enzyme are reflected in the absolute changes in
Vmax and Kmvalues at the various pH values (Table 4). The
relative impact of the metal substitution on these kinetic
constants is compensatory, rendering the overall enzyme
equal toward their substrate specificity at all pH values (Table
4, Figure 4).
The previous studies on the M. Vannielii PR-AMP cyclo-
hydrolase (2) demonstrated that catalysis requires both Zn2+
and Mg2+ions. The Zn2+is a high-affinity binding ligand,
while the Mg2+was shown to rapidly exchange. The
mechanistic correlate for the cyclohydrolase reaction and the
Zn2+-dependent enzyme are the family of nucleoside/
nucleotide hydrolases represented by adenosine deaminase
and cytidine deaminase (3,4). The required stoichiometry
of M. Vannielii PR-AMP cyclohydrolase for Zn2+is 1:1,
and the three cysteine residues in the sequence motif
C93(X)15C109H110(X)5C116 were suggested to be the Zn2+
ligands. Consistent with these observations is the three-
dimensional structure of the M. thermoautotrophicum en-
zyme, which shows that the equivalent cysteine residues,
Cys86, Cys102, and Cys109, form a metal binding site. Most
interesting is the obligate dimeric feature of this structure
which combines Cys86 from one monomer with two cys-
teines from the other monomer to complete the metal binding
site. A solvent molecule completes the tetrahedral coordina-
tion of this bound metal ion. Mutations of the corresponding
cysteine residues in the M. Vannielii enzyme confirm that
these residues form the Zn2+-binding site and each of them
is essential for enzymatic activity (D’Ordine et al., manu-
script in preparation). Therefore, a fully competent enzyme
requires dimerization to form the active site.
A requirement for free Mg2+was demonstrated for the
M. Vannielii enzyme (2), and the model indicates co-binding
of the substrates and metal. The triad of conserved aspartic
acid residues in the cleft is a likely place for Mg2+to bind.
As expected for the Mg2+ion, the Cd2+(2) ion bound in this
position in our structure is octahedrally coordinated. This
metal ion neutralizes the adjacent negative charges of the
carboxylate groups. The close proximity of this site to the
putative Zn2+-binding site is consistent with the requirement
of Mg2+for catalytic activity and is also consistent with the
decreases in activity observed upon mutation of the M.
Vannielii equivalent of Asp87 (D’Ordine et al., manuscript
in preparation).
The conserved sequence motif SR[S/T]RXX[L/I]WXKG-
[E/A]TSG previously suggested to form a P-loop in-
volved in Mg2+/phosphate binding (2) corresponds to the
sequence S60TSRGKLWLKGESSG74 in M. thermoautotrophi-
cum. Many of the highly conserved residues in HisI
sequences are located in this motif. The three-dimensional
structure reported here reveals that this motif forms one wall
of the surface cleft. No metal-binding sites are localized to
this loop and would indicate that the loop does not play the
role of a classical Mg2+and phosphate binding P-loop.
Nonetheless, the conservation of these residues and their
location imply that they play a crucial role in substrate
binding.
To investigate possible substrate-binding roles of con-
served residues in this motif, a model of 6-OH PR-AMP
was docked into one of the two putative active site clefts of
the HisI homodimer. 6-OH PR-AMP represents a tetrahedral
intermediate formed by the attack of the Zn-activated water
nucleophile to the purine C6atom of the PR-AMP substrate.
This would be the mechanistic corollary to the tetrahedral
intermediates formed in the purine/pyrimidine deaminases.
This substrate has an approximate 2-fold symmetry, and two
orientations of the elongated molecule are possible within
the cleft of PR-AMP cyclohydrolase. The two binding modes
were evaluated in detail. Flexible docking was achieved by
Monte Carlo sampling of rotations around all six of the single
bonds in the ligand, followed by energy minimization while
constraining the 6-OH oxygen atom within the tetrahedral
coordination geometry relative to the deduced Zn2+location.
The lowest energy model obtained for the HisI-transition
state complex is shown in Figure 3D.
This model suggests plausible roles for the conserved
residues in the HisI sequence. Specifically, the SR[S/T]RXX-
[L/I]WXKG[E/A]TSG consensus motif contributes to the
substrate binding. The side chains of Ser60 and Ser62 and
the backbone nitrogen atom of Thr61 hydrogen-bond the
adenosyl phosphate group, and side chain and backbone
nitrogen atoms of Lys69, Gly70, Ser72, and Ser73 hydro-
gen-bond to the phosphoribosyl phosphate. The modeling
suggests that Arg15, although not highly conserved among
PR-AMP cyclohydrolases, provides additional stabilization
of the adenosyl phosphate. The aromatic side chain of Trp67
interacts with both the adenine ring and the phosphoribosyl
ribose ring. The deduced position of the Mg2+ion suggests
that it interacts with the adenosyl ribose rather than with the
phosphate. The Asp87 carboxylate may also stabilize the
transition state by interacting with the purine amino group
at the tetrahedral C6atom.
D’Ordine et al. (2) noted the conservation of a histidine
(His103 in M. thermoautotrophicum) and suggested that this
residue may play a role in catalysis. In the crystal structure,
this residue is involved in coordination of a Cd2+ion bound
between two symmetry-related molecules contributing to
crystal contacts. This restraint likely influences the orientation
and position of the histidine side chain relative to its
conformation in solution. The model of the bound putative
tetrahedral intermediate 6-OH PR-AMP provides the ratio-
nale for the role of His103. As seen in Figure 3D, rotation
of the imidazole ring could allow a hydrogen bond to the
Table 4: Kinetic Parameters of Native M. Thermoautotrophicum
HisI (Zinc) Enzyme and M. Thermoautotrophicum HisI Dialyzed
against Cd2+(Cadmium)
Zinc Cadmium
pH kcat Km
kcat/Km
(mM-1s-1)kcat Km
kcat/Km
(mM-1s-1)
6.5 1.7 112 0.0152 0.13 32.6 0.0040
7 8.7 73.4 0.1185 0.31 5 0.0620
7.5 7.5 33.6 0.2232 0.35 1.9 0.1842
8 6.4 13.6 0.4706 0.29 0.75 0.3867
8.5 9.9 20.4 0.4853 0.28 0.9 0.3111
9 7.4 19.8 0.3737 0.2 0.9 0.2222
Crystal Structure of HisI Biochemistry, Vol. 44, No. 30, 2005 10079

6-OH group and directly contact the scissile C6-N1bond
of the transition state (∼3.4 Å). We could thus infer that
His103 would be capable not only of positioning the Zn-
activated water molecule in the ground state prior to
transition state formation, but also of assisting in the
conversion of the transition state toward product formation.
Mutation of this histidine decreased the activity of the M.
Vannielii enzyme by approximately 3 orders of magnitude
(D’Ordine et al., manuscript in preparation). A more detailed
description of the proposed reaction mechanism will be
presented in a subsequent paper (D’Ordine et al., manuscript
in preparation).
In conclusion, the three-dimensional structure of HisI is
consistent with the previous biochemical characterization of
the enzyme. The proposed Zn2+-binding site was identified
to include three conserved cysteine residues as proposed
previously, a likely binding site for the required Mg2+ion
has been localized, and the motif of conserved residues,
although not a classical Mg2+/phosphate-binding P-loop, is
suggested by modeling of a reactive intermediate to be the
substrate binding motif. The structure of the HisI catalytic
site is consistent with a reaction mechanism similar to those
of nucleoside deaminases.
REFERENCES
1. Singh, B. K. (1999) Plant Amino Acids, Marcel Dekker, Inc., New
York.
2. D’Ordine, R. L., Klem, T. J., and Davisson, V. J. (1999) N1-(5′-
phosphoribosyl)adenosine-5′-monophosphate cyclohydrolase: pu-
rification and characterization of a unique metalloenzyme, Bio-
chemistry 38, 1537-1546.
3. Kinoshita, T., Nishio, N., Nakanishi, I., Sato, A., and Fujii, T.
(2003) Structure of bovine adenosine deaminase complexed with
6-hydroxy-1,6-dihydropurine riboside, Acta Crystallogr., Sect. D
59, 299-303.
4. Johansson, E., Neuhard, J., Willemoes, M., and Larsen, S. (2004)
Structural, kinetic, and mutational studies of the zinc ion environ-
ment in tetrameric cytidine deaminase, Biochemistry 43, 6020-
6029.
5. Ireton, G. C., Black, M. E., and Stoddard, B. L. (2003) The 1.14
Å crystal structure of yeast cytosine deaminase: evolution of
nucleotide salvage enzymes and implications for genetic chemo-
therapy, Structure 11, 961-972.
6. Fani, R., Mori, E., Tamburini, E., and Lazcano, A. (1998)
Evolution of the structure and chromosomal distribution of
histidine biosynthetic genes, Origins Life EVol. Biosphere 28,
555-570.
7. Beckler, G. S., and Reeve, J. N. (1986) Conservation of primary
structure in the hisI gene of the archaebacterium, Methanococcus
Vannielii, the eubacterium Escherichia coli, and the eucaryote
Saccharomyces cereVisiae,Mol. Gen. Genet. 204, 133-140.
8. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray
diffraction data collected in oscillation mode, in Macromolecular
Crystallography, part A (Carter, C. W., Jr., and Sweet, R. M.,
Eds.) 276th ed., pp 307-326, Academic Press, New York.
9. Terwilliger, T. C. (2002) Automated structure solution, density
modification and model building, Acta Crystallogr., Sect. D 58,
1937-1940.
10. Terwilliger, T. C., and Berendzen, J. (1999) Automated MAD and
MIR structure solution, Acta Crystallogr., Sect. D 55 (Pt 4), 849-
861.
11. Jones, T. A., Zou, J.-Y., Cowan, S. W., and Kjeldgaard, M. (1991)
Improved methods for building models in electron density maps
and the location of errors in these models, Acta Crystallogr., Sect.
A47, 110-119.
12. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros,
P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges,
M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and
Warren, G. L. (1998) Crystallography & NMR system: A new
software suite for macromolecular structure determination, Acta
Crystallogr., Sect. D. 54, 905-921.
13. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton,
J. M. (1993) PROCHECK: a program to check the stereochemical
quality of protein structures, J. Appl. Crystallogr. 26, 283-291.
14. Chowdhury, S. F., Sivaraman, J., Wang, J., Devanathan, G.,
Lachance, P., Qi, H., Menard, R., Lefebvre, J., Konishi, Y., Cygler,
M., Sulea, T., and Purisima, E. O. (2002) Design of noncovalent
inhibitors of human cathepsin L. From the 96-residue proregion
to optimized tripeptides, J. Med. Chem. 45, 5321-5329.
15. Lin, L. Y., Sulea, T., Szittner, R., Vassilyev, V., Purisima, E. O.,
and Meighen, E. A. (2001) Modeling of the bacterial luciferase-
flavin mononucleotide complex combining flexible docking with
structure-activity data, Protein Sci. 10, 1563-1571.
16. Li, Z., and Scheraga, H. A. (1987) Monte Carlo-minimization
approach to the multiple-minima problem in protein folding, Proc.
Natl. Acad. Sci. U.S.A. 84, 6611-6615.
17. Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, I. R., Merz, K.
M., Ferguson, D. M., Spellmeyer, D. C., Fox, T., Caldwell, J.
W., and Kollman, P. A. (1995) A second generation force field
for the simulation of proteins, nucleic acids, and organic mol-
ecules., J. Am. Chem. Soc. 117, 5179-5197.
18. Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A., and Case,
D. A. (2004) Development and testing of a general amber force
field, J. Comput. Chem. 25, 1157-1174.
19. Cornell, W. D., Cieplak, P., Bayly, C. I., and Kollman, P. A. (1993)
Application of RESP chrages to calculate conformational energies,
hydrogen bond energies, and free energies of solvation, J. Am.
Chem. Soc. 115, 9620-9631.
20. Pang, Y. P., Xu, K., Yazal, J. E., and Prendergas, F. G. (2000)
Successful molecular dynamics simulation of the zinc-bound
farnesyltransferase using the cationic dummy atom approach,
Protein Sci. 9, 1857-1865.
21. Auld, D. S. (1988) Methods for metal substitution, Methods
Enzymol. 158,71-79.
22. Barton, G. J. (1993) ALSCRIPT a tool to format multiple sequence
alignments, Protein Eng. 6,37-40.
23. Gibbons, C., Montgomery, M. G., Leslie, A. G., and Walker, J.
E. (2000) The structure of the central stalk in bovine F(1)-ATPase
at 2.4 Å resolution, Nat. Struct. Biol. 7, 1055-1061.
24. Fedorov, R., Meshcheryakov, V., Gongadze, G., Fomenkova, N.,
Nevskaya, N., Selmer, M., Laurberg, M., Kristensen, O., Al-
Karadaghi, S., Liljas, A., Garber, M., and Nikonov, S. (2001)
Structure of ribosomal protein TL5 complexed with RNA provides
new insights into the CTC family of stress proteins, Acta
Crystallogr., Sect. D 57, 968-976.
25. Lu, M., and Steitz, T. A. (2000) Structure of Escherichia coli
ribosomal protein L25 complexed with a 5S rRNA fragment at
1.8-Å resolution, Proc. Natl. Acad. Sci. U.S.A 97, 2023-
2028.
26. Zvereva, M. E., Shpanchenko, O. V., Nierhaus, K., and Dontsova,
O. A. (2000) Ribosomal protein TL5 of T. thermophilus is
incorporated in the E. coli 50S ribosomal subunit, Dokl. Biochem.
374, 199-201.
BI050472W
10080 Biochemistry, Vol. 44, No. 30, 2005 Sivaraman et al.