Tài liệu Báo cáo khoa học: X-ray crystallographic and NMR studies of pantothenate synthetase provide insights into the mechanism of homotropic inhibition by pantoate docx

X-ray crystallographic and NMR studies of pantothenate
synthetase provide insights into the mechanism of
homotropic inhibition by pantoate
Kalyan Sundar Chakrabarti*, Krishan Gopal Thakur, B. Gopal and Siddhartha P. Sarma
Molecular Biophysics Unit, Indian Institute of Science, Bangalore, India
Keywords
competitive inhibition; NMR; pantothenate
biosynthesis; substrate binding; X-ray
crystallography
Correspondence
S. P. Sarma, 207, Molecular Biophysics
Unit, Indian Institute of Science, Bangalore
560012, India
Fax: +91 80 23600535
Tel: +91 80 22932839
E-mail:
*Present address
Department of Biochemistry and Howard
Hughes Medical Institute, MS009 Brandeis
University, Waltham, MA, USA
Database
The
1
H
N
,
15
N,
13
C
a

,
13
C
b
and
13
C¢ chemical
shift values of the dimeric N-terminal
domain of Escherichia coli pantothenate
synthetase have been deposited in
BioMagResBank (b.
wisc.edu) under the accession number 6940
The structure factor file and the atomic
coordinates of the dimeric N-terminal
domain of Escherichia coli pantothenate
synthetase bound to three molecules of
pantoate have been deposited in the Protein
Data Bank under the accession number
3GUZ
(Received 26 July 2009, revised 6 October
2009, accepted 24 November 2009)
doi:10.1111/j.1742-4658.2009.07515.x
The structural basis for the homotropic inhibition of pantothenate synthe-
tase by the substrate pantoate was investigated by X-ray crystallography
and high-resolution NMR spectroscopic methods. The tertiary structure of
the dimeric N-terminal domain of Escherichia coli pantothenate synthetase,
determined by X-ray crystallography to a resolution of 1.7 A
˚
, showed a
second molecule of pantoate bound in the ATP-binding pocket. Pantoate

binding to the ATP-binding site induced large changes in structure, mainly
for backbone and side chain atoms of residues in the ATP binding
HXGH(34–37) motif. Sequence-specific NMR resonance assignments and
solution secondary structure of the dimeric N-terminal domain, obtained
using samples enriched in
2
H,
13
C, and
15
N, indicated that the secondary
structural elements were conserved in solution. Nitrogen-15 edited two-
dimensional solution NMR chemical shift mapping experiments revealed
that pantoate, at 10 mm, bound at these two independent sites. The solu-
tion NMR studies unambiguously demonstrated that ATP stoichiometri-
cally displaced pantoate from the ATP-binding site. All NMR and X-ray
studies were conducted at substrate concentrations used for enzymatic
characterization of pantothenate synthetase from different sources [Jonczyk
R & Genschel U (2006) J Biol Chem
281,
37435–37446]. As pantoate bind-
ing to its canonical site is structurally conserved, these results demonstrate
that the observed homotropic effects of pantoate on pantothenate biosyn-
thesis are caused by competitive binding of this substrate to the ATP-bind-
ing site. The results presented here have implications for the design and
development of potential antibacterial and herbicidal agents.
Structured digital abstract
l
MINT-7301221: PS (uniprotkb:P31663) and PS (uniprotkb:P31663) bind (MI:0407)byx-ray
crystallography (

MI:0114)
l
MINT-7301241: PS (uniprotkb:P31663) and PS (uniprotkb:P31663) bind (MI:0407)by
nuclear magnetic resonance (
MI:0077)
Abbreviations
cPS, C-terminal domain of pantothenate synthetase; HSQC, heteronuclear single quantum coherence; nPS, dimeric N-terminal domain of
Escherichia coli pantothenate synthetase; PS, pantothenate synthetase; TEV, tobacco etch virus; TROSY, transverse relaxation optimized
spectroscopy.
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 697
Introduction
Inhibition by substrates is an important feature of
metabolic enzymes that remains poorly characterized
in terms of structure. An intriguing aspect of sub-
strate-induced reversible inhibition is that it may vary
substantially even among closely related homologues.
Pantothenate synthetases (PSs) from all species cata-
lyze the condensation of a molecule of b-alanine with
a molecule of pantoate in an ATP-dependent manner
to form pantothenate [1,2]. Pantothenate itself is an
important cofactor that is essential for CoA biosynthe-
sis, and functions in fatty acid synthesis and energy
metabolism [2]. The structure and function of PSs
from several prokaryotic and eukaryotic organisms
have been reviewed recently [2]. Of special interest are
the PSs from the bacteria Escherichia coli [3–6] and
Mycobacterium tuberculosis [7], and from the plants
Oryza sativa [8], Arabidopsis thaliana [9,10] and
Lotus japonicus [8]. Analysis of the primary sequences
of this enzyme from E. coli, M. tuberculosis and

A. thaliana have shown that there is a high degree of
conservation ( 40% sequence identity) among the
enzymes from various sources, with the E. coli and
A. thaliana enzymes sharing 42% [9] sequence identity,
and the M. tuberculosis and A. thaliana enzymes shar-
ing 36% sequence identity [11]. Recently, it has been
shown that, unlike their bacterial counterparts, the
plant PSs are subject to allosteric control, and that this
allosteric control arises from the homotropic effects of
one of the substrates, i.e. pantoate [8,9]. One of the
interesting features of the plant PSs is the presence of
a conserved 24 residue insertion in the sequence [9].
This sequence of amino acids, which is missing in the
sequence of bacterial enzymes, is thought to play a sig-
nificant role in the proposed allostery in plant PSs [9],
through long-range interactions.
All PSs are known to be dimeric, with distinct N-ter-
minal and C-terminal domains in each protomer.
Structural studies have shown that the catalytic site
residues, as well as the dimerization interface, lie in the
N-terminal domain of the bacterial enzyme. Enzymatic
and other biochemical studies have shown that for PS
from M. tuberculosis, pantoate, ATP and b-alanine
have K
m
values of 0.13 mm, 2.6 mm and 0.8 mm,
respectively. Enzymatic studies of plant PSs have indi-
cated that the affinity of these substrates echoes that
observed for the bacterial enzyme. However, pantoate
at concentrations above 5 mm has been shown to inhi-

bit the plant enzyme [8,9]. This inhibition by pantoate
is thought to occur by binding of pantoate at a nonca-
nonical site, although with much lower affinity
( 100-fold weaker) [8,9]. Knowledge of the canonical
binding site comes from the well-documented struc-
tures of the M. tuberculosis PS in the substrate-bound
forms [11,12]. ATP at high concentration ( 5mm)
has been shown to offset the inhibitory effect of panto-
ate on this enzyme. Furthermore, the plant PS exhib-
ited negative cooperativity for b-alanine when
pantoate was present at a high concentration. To date,
no information is available on the structural basis for
this inhibition of this enzyme. It is also hypothesized
that this negative cooperativity has no regulatory func-
tion but serves to ensure robust synthesis of pantothe-
nate from low amounts of pantoate [9]. Here, we
report the results of the structural studies of the panto-
ate-bound form of the N-terminal domain of E. coli
PS (nPS). The structure and substrate-binding interac-
tions of E. coli nPS have been studied using solution
NMR as well as X-ray crystallography. It is important
to note that the structure of the E. coli protein in the
substrate-bound form has not been reported [13].
Details of the structural basis for the biochemically
observed inhibition of plant PS by pantoate, as well as
of the alleviation of this inhibition by another sub-
strate of the enzyme, ATP, are presented [8,9].
Results
Solution properties of the N-terminal domain of
PS, the C-terminal domain of PS, and full-length

PS
Full-length PS is a dimer of 63 kDa. The protein was
found to aggregate at the concentrations used for the
NMR studies, and did not provide spectra of adequate
quality under a number of different sample conditions
(phosphate buffer, pH 6.8; acetate buffer, pH 5.5; and
Tris buffer, pH 7.5), in the presence of substrates
(10 mm pantoate, ATP, and b-alanine), in the presence
of protein stabilizers such as arginine, glutamate and
proline [14–17], or even in the presence of detergents
(Chaps [18] or dodecylphosphocholine) [17].
Like PS (283 amino acids), nPS (residues 1–176)
exists as a dimer in solution. Evidence for this comes
from gel filtration data and from the experimentally
determined rotational correlation time (s
c
) of 17.25 ns
for nPS, calculated from the measured average
15
N T
1
and T
2
relaxation time constants [19].
The C-terminal domain of PS (cPS; residues 177–
283) showed spectra of poor quality, similar in nature
to that of the full-length PS under similar conditions.
Coconcentration of cPS with nPS did not prevent
aggregation of cPS. It is possible that the C-terminal
Structure and binding studies of nPS K.S. Chakrabarti et al.

698 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
domain is responsible for the aggregation of the full-
length PS proteins at the concentration required for
NMR experiments.
The studies leading to structural and dynamic char-
acterization of the PS molecule were limited to the
nPS protein. The nPS protein has a V52A mutation,
which was confirmed by MS and DNA sequencing.
Sequence-specific resonance assignments and
secondary structure of nPS
As mentioned above, nPS is a dimeric molecule
( 40 kDa). Under the circumstances, the backbone
and side chain resonance assignments could only be
obtained on samples that were enriched in
2
H,
13
C and
15
N. Figure 1 shows the assigned
1
H–
15
N correlation
spectrum of nPS. Sequence-specific resonance assign-
ment for the backbone
1
H
N
,

13
C
a
,
13
C
b
,
13
C¢ and
15
N
nuclei of the homodimeric nPS was obtained from
transverse relaxation optimized spectroscopy (TROSY)
[20] versions of HNCA, HN(CO)CA, HNCACB,
HN(CO)CACB and HNCO [21–23] experiments. The
resonance assignments leading to the establishment of
sequential connectivity for Glu5–Glu20 are shown in
Fig. S1. The backbone assignment for the amide
1
H–
15
N pairs is 93% complete, whereas the backbone
and side chain
13
C
a
and
13
C¢ and

13
C
b
carbons are
assigned to the extent of 97%, 93% and 93%, respec-
tively (BMRB accession number 6940) [24]. The
1
H–
15
N correlation in the TROSY experiment is miss-
ing for Asp35, Asn101, Thr103, Glu104, Tyr108,
Phe127, Thr132, Val134, Leu137, Leu140, and Phe153.
These resonances were not visible even after denatur-
ation and refolding to facilitate back-exchange of
amide protons from bulk solvent (H
2
O). These reso-
nances, especially the ones concentrated near the dimer
interface, are missing, presumably owing to the fact
that the residues undergo intermediate exchange on the
NMR timescale. The secondary structure of nPS was
found to be, with some differences, nearly identical to
the N-terminal domain in the crystal structure reported
here (see below) and of the full-length structure
reported earlier [13]. The secondary carbon chemical
shift [25,26] and short-range
1
H
N
–

1
H
N
NOE pattern
[27] that define the secondary structure are shown in
Fig. S2. The 3
10
helices formed by Pro59–Gln61 and
Glu119–Ser122 that are present in the crystal struc-
tures of E. coli and M. tuberculosis PS, are missing in
the solution structure. Helix IX, which is a continuous
helix (16 residues) in the PS crystal structure, is broken
at Thr132 and resumes only at Phe138 in the solution
structure. In the absence of substrates, the crystal
structure of the E. coli PS has shown that the protein
has an extensive dimer interface constructed from
strands formed by Tyr108–Val111 from each subunit
in an antiparallel b-sheet arrangement. Under the con-
ditions of these solution studies, the resonances for
Tyr108 could not be identified. However, we have been
able to identify resonances for Val109, Asp110, and
Val111. Prediction of backbone u and w dihedral
angles from calculated
13
C secondary chemical shifts,
using the program talos [28], indicates that this region
adopts a b-strand conformation under solution condi-
tions. The occurrence of a b-strand is corroborated by
the absence of backbone H
N

–H
N
NOEs.
Mapping chemical shift changes upon substrate
binding
Binding of substrates to nPS was studied by both
NMR and X-ray crystallography. NMR studies that
involved monitoring changes in chemical shift
Fig. 1. (A) Sequence specifically assigned
two-dimensional
1
H–
15
N TROSY spectrum
of triple-labeled nPS in 10 m
M phosphate
buffer, 10 m
M NaCl, 2 mM dithiothreitol, and
0.01% NaN
3
(pH 6.8). For clarity, assign-
ments of only resolved resonances have
been indicated.
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 699
positions of backbone atom pairs in
1
H–
15
N correla-

tion spectra were conducted by addition of substrates
in the concentration range used for biochemical char-
acterization of PS [7–9]. Changes in chemical shifts
were observed at pantoate concentrations > 5 mm,
indicating weak binding coupled to an exchange rate
that is intermediate on the NMR timescale. Addition
of d-pantoate (10 mm) to nPS caused deviations in
chemical shifts of more than 0.03 p.p.m. for 24 resi-
dues. Addition of ATP brought about further changes
in the heteronuclear single quantum coherence (HSQC)
spectrum of nPS. The changes in chemical shifts in the
HSQC spectra for select residues that lie in the panto-
ate-binding and ATP-binding pocket of nPS are shown
in Fig. 2. The residues that showed deviations in chem-
ical shifts of more than 0.03 p.p.m. upon binding pan-
toate and ATP fell into three categories. Residues in
class I were almost uniquely shifted in the presence of
ATP (Fig. 2, row 1). Residues in class II showed a net
perturbation in chemical shifts in the presence of both
pantoate and ATP (Fig. 2, row 2). Finally, residues in
class III showed deviations in the presence of pantoate
but reverted back to the free protein resonance upon
addition of ATP (Fig. 2, row 3). The backbone atoms
of Val111 and Gly113, which are involved in dimeriza-
tion (see above), were also affected by substrate bind-
ing. However, the backbone of His106, whose side
chain is involved in dimerization, and a residue
Fig. 2. The overlay of selected regions of
1
H–

15
N correlation spectra, highlighting
cumulative changes in chemical shifts of
specific residues of nPS as a function of
substrate(s). The resonances of free nPS
are shown in red, those of nPS bound to
D-pantoate are shown in blue, and those of
nPS bound to
D-pantoate and ATP are show-
n in green. The arrows indicate the sign of
change in chemical shift of the panto-
ate-bound (P; in blue) and pantoate +
ATP-bound (PA; in green) protein as
compared with the unbound protein.
Numbers in parentheses indicate the
magnitude of change in p.p.m.
Structure and binding studies of nPS K.S. Chakrabarti et al.
700 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
immediately before the dimer interface, Gly102, were
unaffected by substrate binding (Fig. 2, row 4). The
absolute values of the deviations in chemical shift
when compared to the unbound protein for both pan-
toate and ATP are shown in Fig. 3A,B. It can be
clearly seen that there were two regions of the protein
that were significantly perturbed upon the addition of
pantoate. These two regions correspond to the panto-
ate-binding and ATP-binding sites of PS [11]. ATP
caused additional changes in the spectrum, particularly
for those residues that lie in the ATP-binding pocket.
Interestingly, in the presence of ATP, the His34 back-

bone H
N
chemical shift reverted back to the resonance
position observed for the free form of the protein. The
observed changes in resonance positions in the HSQC
spectrum upon addition of pantoate followed by ATP
can be reconciled if one takes into consideration that a
molecule of pantoate is bound in the ATP-binding
pocket in addition to its canonical binding site, and
that ATP displaces this molecule of pantoate.
b-Alanine and pantothenate did not bind nPS under
the experimental conditions. Attempts to study the
binding of ATP alone to the protein were unsuccessful,
as the protein precipitated upon addition of ATP in
the absence of pantoate. An important feature of the
NMR study is the fact that a single set of resonances
were observed for nPS in the absence or presence of
substrates. This strongly suggests that binding of pan-
toate and ATP preserves the dimer symmetry in solu-
tion. In an effort to unambiguously identify and
understand the nature of binding of pantoate at the
ATP-binding site, we also structurally characterized
the pantoate-bound form of nPS using X-ray crystal-
lography, as described below.
Solution and refinement of the structure using
X-ray crystallography
The crystal structure of the pantoate–nPS complex was
solved by molecular replacement using phaser [29] in
ccp4 [30]. The first solution with a log likelihood gain of
1995 was obtained using the edited coordinates of

E. coli PS (1IHO.pdb). The molecular replacement solu-
tion could be readily refined, and an initial examination
of the (mf
0
– Df
c
) maps in coot clearly showed the posi-
tion of pantoate within the active site cleft of nPS. The
data-processing parameters are listed in Table 1.
Maximum likelihood refinement, using refmac 5
[31], was used to improve the quality of the electron
density maps and facilitate further rebuilding and
improvement of the molecular model, until no unex-
plained electron density remained, and the R
cryst
and
R
free
values converged at 18.9% and 24.4%, respec-
tively. The electron density was missing or weak for
residues 113–124 in all cases, and this stretch of resi-
dues was therefore not rebuilt.
X-ray crystallographic structure of nPS bound to
pantoate
Cocrystallization trials of nPS with pantoate were
performed over a range of pantoate concentrations.
Crystals of the nPS–pantoate complex obtained at
50 mm pantoate diffracted to a resolution of 1.7 A
˚
.

The crystal structure of the nPS–pantoate complex
indicated that it is a dimer in the asymmetric unit. The
backbone structure of the dimeric N-terminal domain
of the E. coli PS determined here is identical to that
determined earlier for the full-length protein [13]. The
A
B
Fig. 3. The absolute value of deviation in chemical shift (p.p.m.) of
residues of nPS upon binding to (A)
D-pantoate as a function of
sequence, and (B)
D-pantoate and ATP as a function of sequence.
Residues that showed a deviation of more than 0.03 p.p.m. were
considered to have a direct interaction with the ligand(s) (see text
for details).
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 701
dimer interface in PS is extensive. The dimer interface
is stabilized by 19 hydrogen bonds and two salt
bridges, with a total buried surface area of 1240 A
˚
2
(Table S3). The structure of the complex showed two
molecules of pantoate bound to one monomer
(chain A) and a single pantoate bound to the second
monomer (chain B). Thus, pantoate binding occurs in
the canonical pantoate-binding site (site I) in both
monomers with full occupancy, and at the ATP-bind-
ing site (site II) in one monomer with full occupancy.
Figure 4A shows the dimeric nPS molecule bound to

three molecules of pantoate. The electron density for
the two molecules of pantoate bound at the active site
of nPS can be unambiguously distinguished in Fig. 4B.
The crystal structures of the protein were of high ste-
reochemical quality, in that all backbone u and w
dihedral angles occupy allowed regions of the Rama-
chandran map [32,33]. Comparison of the monomer of
the substrate-free form of E. coli PS [13] with the
monomers of nPS in the pantoate-bound form at full
occupancy indicates that the backbone C
a
atoms show
an overall deviation of 0.51 A
˚
.
Pantoate bound at site I
The physicochemical nature of binding of pantoate to
its canonical binding site on nPS is identical to that
observed in the case of the M. tuberculosis protein [11].
In fact, the residues of nPS involved in binding panto-
ate are highly conserved in the two proteins. Figure 5A
shows that the important substrate–protein interac-
tions that are observed for the M. tuberculosis PS–pan-
toate complex are conserved in the case of the E. coli
protein too. Figure 5 shows that the side chain atoms
of Gln61, i.e. Oe1 and Ne2, are hydrogen-bonded to
the O3 and O4 atoms of pantoate, respectively,
whereas the similar atoms of Gln155 are hydrogen-
bonded to the O2 and O3 atoms of pantoate, respec-
tively. These two residues are conserved in the E. coli

and M. tuberculosis proteins, and are involved in iden-
tical conserved interactions with pantoate. Further-
more, the pantoate molecule is hydrogen-bonded to
other protein residues via networks of water-mediated
hydrogen bonds. The backbone nitrogen and carbonyl
oxygen of Thr29 and the side chain oxygen of Ser54
are hydrogen-bonded to the O4 atom of pantoate via a
network that consists of two water molecules. The
backbone carbonyl oxygen of Phe56 is also hydrogen-
bonded to the O4 atom of pantoate via one of the
water molecules in the same network. The phenolic
oxygen of Tyr71 is hydrogen-bonded to the O1 atom
of pantoate via a water molecule included in a network
of five water molecules. Also hydrogen-bonded to the
O1 atom of pantoate via this cluster of water mole-
cules is the Ne nitrogen of the imidazole ring of His37.
In addition to the hydrogen-bonding interactions, the
pantoate molecule is also involved in hydrophobic con-
tacts with the protein, as shown in Fig. S3. The Ca
atoms of residues in site I do not exhibit more than
1A
˚
deviations as compared with the substrate-free
form of the protein.
Pantoate bound at site II
The structure of nPS distinctly shows the presence of a
second molecule of pantoate bound at site II in one of
the monomers, albeit at full occupancy. The pantoate
in site II (Fig. 5) has the same orientation as the pan-
toate molecule in site I. In contrast to the pantoate

bound to site I, this molecule has fewer direct interac-
tions with protein residues. This pantoate molecule is
anchored to the protein by a hydrogen bond between
the O2 atom of pantoate and the Nf atom of the side
chain of Lys39. However, there are a large number of
water-mediated networked interactions that stabilize
the pantoate at site II. The backbone carbonyl oxygen
of Val175 is hydrogen-bonded to the O2 and O3 atoms
of pantoate via a water molecule. This water molecule
also hydrogen bonds backbone nitrogen atoms
of Glu150 and Lys151 to the O2 and O3 atoms of
Table 1. Atomic refinement of models for nPS bound to two mole-
cules of pantoate per monomer.
Data collection
Wavelength (A
˚
) 1.5418
Resolution (A
˚
) 34.78–1.67 (1.76–1.67)
a
Space group P2
1
2
1
2
Unit cell parameters, a, b, c (A
˚
) 61.89, 77.50, 78.85
Total reflections 109 170 (13 893)

Unique reflections 43 231 (5676)
Completeness (%) 96.2 (87.5)
I ⁄ r(I) 12.8 (2.0)
R
merge
(%) 5.1 (46.4)
Redundancy 2.5 (2.4)
Refinement
R
cryst
(%) 18.7
R
free
(%) 24.4
rmsd from ideal bond
length (A
˚
) ⁄ angles (°)
0.024 ⁄ 2.03
No. of protein residues 332
No. of water molecules 279
No. of ligand molecules 3
Wilson B-factor (A
˚
2
) 20.3
Ramachandran plot (%)
Most favored regions (%) 99.1
Allowed regions (%) 0.9
a

The high-resolution shell is shown in parentheses.
Structure and binding studies of nPS K.S. Chakrabarti et al.
702 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
pantoate. The Nd atom of Asp152 is hydrogen-bonded
via one water molecule to the O3 atom of pantoate,
and via two water molecules to the O4 atom of panto-
ate. The Ne atom of His37 is hydrogen-bonded to the
O4 atom of pantoate via a water molecule, which is in
turn hydrogen-bonded to a network of water molecules
that are hydrogen-bonded to the O1 atom of pantoate
in site I. The O4 atom of pantoate in site II and the
O1 atom of pantoate in site I are hydrogen-bonded via
a water molecule. Structural stabilization of ligands by
networks of water-mediated hydrogen bonds is often
observed in the active sites of enzymes [34–36]. The
pantoate molecule also has important hydrophobic
interactions with the aliphatic carbons of the Lys39
side chain (Fig. S4). Residues EKD(15–152) and
Val175 are conserved in the case of the E. coli and
M. tuberculosis enzymes, and the corresponding resi-
dues in the A. thaliana protein are KKD(150–152) and
Ser175, respectively. Given the fact that the main types
of interaction between pantoate and the residues in
site II are water-mediated, it is almost certain that
these interactions will be conserved in A. thaliana PS
as well. An interesting feature of pantoate binding at
site II is that it takes the position of the adenine ring
in the ATP-binding site, as shown in Fig. 6. The oxy-
gen atoms of the pantoate are spatially coincident with
the nitrogen atoms of the adenine ring. The O1, O2

and O4 atoms of pantoate occupy the positions of the
A
B
Site I
Site II
Site I
N
N
C
Gln61
Gln155
Asp152
Lys151
Glu150
Tyr71
Phe56
Ser54
Thr29
His37
Lys39
Val175
Phe56
Ser54
Thr29
His37
Lys39
Val175
Glu150
Lys151
Asp152

Gln61
Gln155
Tyr71
Fig. 4. (A) Crystal structure of nPS bound to three molecules of pantoate. The subunits are shown in ochre and blue, respectively. Major
secondary structural elements are numbered sequentially, where a is used to denote an a-helix and b is used to denote a b-strand. The
dimer is formed through association of b5 from each subunit (circled region). The pantoate molecules bound at site I and site II (ATP-binding
site) are shown in green (carbon atoms) and red (oxygen atoms), respectively. (B) Stereo view of the electron density map of the active site
of nPS. At this high resolution, the electron densities that define pantoate in site I and site II are clearly visible. The atoms of the pantoate
molecule are labeled. The electron density clearly defines the backbones and side chains of the residues that line the active site.
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 703
N6, N1 and N9 atoms, respectively, of the adenine
moiety of ATP. The pantoate is bound in the sequence
Thr29–Leu33, which forms the loop connecting
strand 2 (Ala25–Thr29) to helix II (Asp35–Arg47), and
the sequence Gly149–Phe153 in the loop that connects
strand 5 (Ile145–Gly149) to helix X (Phe153–Met166).
Attempts at cocrystallization or soaking of the nPS
crystals with ATP and ⁄ or b-alanine were unsuccessful.
These results were consistent with the solution behav-
ior of the protein.
Comparison of PS structures
Comparison of monomer chains of nPS show that
there is no significant difference in B-factors for the
residues that lie in the canonical pantoate-binding site
(site I). However, Arg64–Leu75 show higher B-factors
in the B-chain than in the A-chain. These residues
form the ‘gate loop’, which is thought to play a signifi-
cant role in channeling the substrate into the active site
in the M. tuberculosis protein. The deviations of back-

bone atoms in this region are small. Significant
changes are observed for residues that line the site II
pantoate-binding pocket. The residues that are shifted
by more than 1 A
˚
upon pantoate binding are listed in
Table 2.
Figure 7 shows a superposition of chain A of E. coli
PS (1IHO.pdb) on the A-chain and B-chain of nPS in
the region of the HXGH (34–37) motif (where
X = Asp for E. coli PS and Glu for M. tuberculosis
and A. thaliana PS). As compared with the substrate-
free form, the imidazole rings of His34 in the A-chain
and B-chain of nPS move by  4A
˚
and 5.5 A
˚
,
whereas the backbone Ca atom moves by only
 0.9 A
˚
and  0.5 A
˚
, respectively. The backbone Ca
atom of Asp35 moves by as much as 1.62 A
˚
and
Fig. 6. An expanded view of the super-
posed substrate-binding sites of nPS (green)
and M. tuberculosis PS (magenta). In nPS,

the pantoate molecule takes the position of
the adenine ring of ATP. The carbon and
oxygen atoms of pantoate are shown in
yellow and red, respectively. The carbon
and oxygen atoms of ATP are shown in red,
the nitrogen atoms in blue, and the phos-
phorus atoms in orange (see text for
details).
Gln61
Gln155
Asp152
Lys151
Glu150
Val175
Phe56
Ser54
Thr29
His37
Lys39
Val175
Glu150
Lys151
Asp152
Gln155
Gln61
Tyr71
Lys39
His37
Thr29
Ser54

Phe56
Tyr71
Fig. 5. Protein–ligand and water-mediated protein–ligand hydrogen bond interactions (broken lines) that stablize the pantoate molecules in
site I and site II of nPS. Oxygen atoms of pantoate and water are colored magenta and ochre, respectively. Direct protein–ligand hydrogen
bonds are colored magenta, and water-mediated hydrogen bonds are colored red. Several networks of water-mediated hydrogen bonds can
be observed.
Structure and binding studies of nPS K.S. Chakrabarti et al.
704 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
0.34 A
˚
in the A-chain and B-chain of nPS. Finally, for
His37, the A-chain Ca atom is displaced by 0.72 A
˚
and the B-chain Ca atom deviates by 0.23 A
˚
, concomi-
tant with a very minor deviation in the positions of
the imidazole rings.
Comparison of side chain torsion angles for these
residues between the free and the liganded forms shows
that the v
1
torsion angle for His34 changes from )64°
in the former to 164° and )169° in the A-chain and B-
chain of nPS, respectively. The v
1
torsion angle for
Asp35 changes from ) 75.20° to )171.65° and )89.25°
in the A-chain and B-chain of nPS, and in the case of
His37, the v

1
torsion angle changes from 63.34° to
)67.80° and )52.11° for the A-chain and B-chain of
nPS, respectively. The imidazole ring of His34 points
inwards in the substrate-free form of E. coli PS as well
as in the ATP-bound form of M. tuberculosis PS. Thus,
this movement of the imidazole ring in the case of
His34 is necessary to accommodate the pantoate
molecule in the ATP-binding site (Fig. 8).
Binding of pantoate and ATP to nPS
The crystal structure of nPS has shown clearly the
bound form of pantoate in the ATP-binding site. The
enzymology of substrate binding has been studied in
detail in the case of PS from M. tuberculosis, which
has enzyme kinetic parameters comparable to those of
E. coli PS [7]. E. coli nPS is saturated at an ATP con-
centration of 10 mm. This strongly suggests that the
K
D
for ATP in the case of nPS is similar that that
observed for M. tuberculosis PS, for which it has been
shown to be 1.8 mm. The binding of the pantoate to
the ATP-binding site must be weaker, which is not sur-
prising, given that ATP binding at this site will be
thermodynamically more favored [11–13], because of
more extensive hydrogen bond and hydrophobic inter-
action networks, which stabilize the ATP in the ATP-
binding pocket. Evidence that the ATP replaces the
pantoate from site II is given by the pattern of changes
in the resonance positions for the backbone atoms of

His34 in NMR spectra. In the presence of ATP, the
His34 resonance reverts back to the position of the
Table 2. The residues that show > 1 A
˚
displacement upon binding
pantoate at both site I and site II.
Residue Displacement (A
˚
)
Arg18 1.06
His34 1.10
Asp35 1.75
Gly36 1.74
Arg64 1.89
Pro65 1.32
Glu66 1.57
Asp67 1.45
Arg70 1.10
Tyr71 1.02
Pro72 1.06
Thr74 1.17
Leu75 1.13
Gln76 1.41
Glu77 1.19
Lys96 1.38
Glu97 1.27
Tyr99 1.07
Pro100 1.57
Fig. 7. Superimposition of the A-chain of nPS (green), the B-chain
of nPS (cyan), and the substrate-free protein (1IHO.pdb; magenta).

The carbon and oxygen atoms of pantoate are shown in green and
red, respectively. The imidazole ring of His34 in the pantoate-bound
A-chain nPS is forced to move away from the pantoate when
compared to the free form of E. coli PS.
Fig. 8. (A) Superimposition of the A-chain of nPS (green), the PS
from M. tuberculosis in ATP-bound form (1N2B.pdb; blue), and the
substrate-free PS from E. coli (1IHO.pdb; magenta ⁄ blue). As seen
for the ATP-bound form of M. tuberculosis PS, the His34 imidazole
ring moves in to interact with the oxygen of the b-phosphate group
of ATP. On the other hand, in nPS, the imidazole ring of His34
moves out to accommodate the incoming pantoate.
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 705
unbound protein. Addition of ATP to the pantoate-
bound form of the protein displaces the pantoate from
the ATP-binding site. The bound ATP is stabilized by
a hydrogen bond between the imidazole side chain of
His34 and the a-phosphate group of ATP [11,12]. This
is most probably accompanied by a reversion of the
structural changes induced by the pantoate binding at
the ATP-binding site (see above). In light of this, it is
not unreasonable to expect that the substrate-binding
properties shown for E. coli PS may be invoked to
explain the observed biochemical properties of the
plant PSs regarding the substrate inhibition by panto-
ate and the alleviation of this inhibition by high con-
centrations of ATP.
Discussion
PSs from all known sources are dimeric, although the
contribution of this dimerization to the properties of

the enzyme is not clear. Kinetic and structural studies
of E. coli PS [3,8,13] and M. tuberculosis PS [7,11,12]
suggest that active sites of the dimer are independent
of each other, and that reactions occur at both sites
simultaneously. In plant PSs, dimerization has been
shown to have a very different effect on the kinetic
properties of the enzyme [9]. A study of the sequence
(Fig. 9) and the modeled structure of A. thaliana PS
(Fig. S5 and Table S1) shows that it has a large loop
at the dimer interface and fewer energetically favorable
N-terminal and C-terminal interdomain interactions,
such as the hydrophobic, hydrogen bond and ionic
interactions, than E. coli PS. Thus A. thaliana PS is
expected to be structurally more open than E. coli PS.
Importantly, the residues at the active site are highly
conserved, and thus the nature of the interactions
between enzyme and substrate are expected to be iden-
tical. M. tuberculosis PS has a significantly higher
number of such energetically favorable interactions
that stabilize the ‘closed’ conformation. For PS from
L. japonicas, the reported dissociation constants for
the first and second molecules of pantoate are
0.042 mm and 5.33 mm, respectively; that is, the sec-
ond pantoate molecule binds with 100-fold lower affin-
ity than the first. Thus, the substrate inhibition by
pantoate is consistent with the finding that pantoate
binds to the ATPase site as well, and is displaced from
this site at equimolar concentrations of ATP.
On the basis of kinetic studies, an enzymatic scheme
has been proposed [9] in which the second molecule of

pantoate binds to the pantoyl adenylate-bound form
of the protein, at a site other than the active site,
which then has a regulatory role in the kinetics of the
reaction. Using high-resolution solution NMR meth-
ods, we have shown that the N-terminal domain of
E. coli PS is capable of binding both pantoate and
ATP, at equivalent sites in the dimer, which is the first
step of the bicatalytic–unicatalytic–unicatalytic–bicata-
lytic mechanism. X-ray crystallographic studies have
unambiguously shown that this domain binds two mol-
ecules of one of the substrates, i.e. pantoate. It is
important to note that, at the level of sensitivity and
resolution available for these studies, pantoate was not
found to bind to any other site on the protein, leading
us to conclude that there is no additional regulatory
Fig. 9. The sequence alignment showing
the high degree of homology between the
E. coli, A. thaliana (A. th) and M. tuberculo-
sis (M. tb) PS. The alignment was made
using
CLUSTALW2. Asterisks indicate strictly
conserved positions. Colons and periods
indicate full conservation of strong and
weak groups, respectively. The C-terminal
domain is highlighted with cyan, and the
insertion sequence corresponding to Lys108
to Arg139 in the A. thaliana sequence is
highlighted in green.
Structure and binding studies of nPS K.S. Chakrabarti et al.
706 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS

site on the N-terminal domain of PS [9]. A kinetic
scheme to explain this observation is shown in Fig. 10.
Such competitive inhibition by one of the substrates
of a reaction has also been observed in the case of ade-
nylosuccinate synthetase, one of whose substrates,
IMP, binds to the GTP-binding site and inhibits the
reaction [37–39]. The binding of pantoate to the
ATP-binding site also reveals the origin of the negative
cooperativity with b-alanine [9] at high pantoate con-
centrations. The model of A. thaliana PS (Fig. S5)
shows that there is a 24 residue insertion between
Thr105 and His106 of the E. coli sequence [9]. The
HETWIRVER(131–139) motif, which is part of the 24
residue insertion, and is present in all plant PSs and
absent in bacterial PSs, has been shown to be impor-
tant for the observed allosteric behavior [9] of plant
PSs. This insertion sequence is rich in Gly residues
N-terminal to the HETWIRVER(131–139) motif,
suggesting that this loop may adopt an extended
conformation in A. thaliana PS. Mutations in or dele-
tion of this stretch of residues at the dimer interface are
known to affect the catalytic properties of A. thaliana
PS [9], as a consequence of disruption of long-range
allosteric interactions. Gly102, which is N-terminal to
the insertion and completely conserved in E. coli,
M. tuberculosis and A. thaliana, is virtually unaffected
by substrate binding (Fig. 2). His106 in E. coli PS,
which is located at the end of the insertion sequence in
A. thaliana PS, is also unaffected by substrate binding.
His106 is involved in intersubunit interactions via its

side chain, and is in a similar position to Glu132 in
A. thaliana PS. Val111 and Gly113, on the other hand,
are both affected by both pantoate and ATP. These
residues have their positional equivalents in Val137 and
Arg139 in A. thaliana PS, as shown in Fig. S5. As the
structure of A. thaliana PS is not known, the structural
basis of the HETWIRVER(131–139) insertion sequence
in establishing long-range allosteric interactions
remains to be elucidated. Interestingly, residues in the
dimer interface and those in the vicinity exhibit high
flexibility. Preliminary analysis of
15
N relaxation data
shows that residues 118–122 have a higher degree of
flexibility than other ordered regions of the protein.
Unfortunately, missing electron density for resi-
dues 113–124 prevents any structural interpretations
from being made. An important outcome of this study
is the potential use of bipantoate molecules [40] as
inhibitors of this enzyme, with the aim of developing of
antibacterial and ⁄ or herbicidal agents [41,42]. The end-
to-end distance between pantoate molecules in site I
and site II is  5A
˚
. Therefore, bipantoate-based inhib-
itors, tethered by a flexible linker, could be attractive
candidates for inhibition of pantothenate synthetase.
Experimental procedures
The isotopically enriched chemicals, i.e.,
15

NH
4
Cl,
13
C
6
H
12
O
6
,
13
C
6
2
H
5
1
H
7
O
6
,
12
C
6
2
H
5
1

H
7
O
6
, 2-ketovalerate
(
13
C
5
,3-
2
H), 2-ketobutyrate (
13
C
5
, 3,3-
2
H), and deuterated
dodecylphosphocholine, were purchased from Cambridge
Isotopes Limited (USA) or Isotec.
2
H
2
O (100%), d-panto-
lactone, ATP (disodium), b-alanine and pantothenate were
purchased from Sigma.
The plasmid containing the tobacco etch virus (TEV)–
maltose-binding protein construct was a kind gift from
B. V. Geisbrecht (University of Missouri, KS, USA) [43].
The plasmid containing the full-length E. coli PS-hexahis-

tidine tag construct was a kind gift from C. Abell and
coworkers (University of Cambridge, UK).
Cloning of nPS
Specific primers were designed (forward primer, 5¢-GAC
CAGCTTCATGTGGCCATCGTGCAGGTT-3¢; reverse
primer, 5¢-CACGATGGCCACATGAAGCT-3¢) to PCR-
amplify the coding regions of these domains in the panC
gene from E. coli genomic DNA [44]. The amplified prod-
uct was ligated into an appropriately digested pet21a plas-
mid vector between Nde1 and EcoR1 restriction sites, to
obtain the clone of the N-terminal domain of PS. She gene
starts with an unusual codon, GTG. The C-terminal
domain of PS was cloned as a cytb5-based fusion protein.
The gene corresponding to the C-terminal domain of PS
was cloned downstream of the cytb5 gene, between BamH1
and EcoR1 restriction sites [45].
Protein expression and purification
All proteins were expressed using E. coli BL21(DE3) as
host strain, and purified using the protocol described
below.
Fig. 10. A model for substrate inhibition under conditions of high
pantoate concentration. In this model, a second pantoate molecule
binds to the enzyme–substrate complex to form a catalytically inert
E.Pt.Pt* species, where Pt is a pantoate bound to the canonical
pantoate-binding site, and Pt* is a second molecule of pantoate
bound to the ATP-binding site. ATP at high concentrations can
displace Pt*, and the reaction goes to completion.
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 707
Purification of the N-terminal domain of PS

The cell lysate was loaded onto a Q-Sepharose column
pre-equilibrated with buffer (20 mm Tris, pH 8.0, 0.01%
azide). The protein was collected in the flow-through, and
the collected fractions, after being checking by denaturing
SDS ⁄ PAGE, were pooled together. The protein was passed
through an S-100 16 ⁄ 60 sephacryl (Pharmacia) gel filtration
column for final purification. The protein was later dialyzed
into appropriate buffer (20 mm sodium dihydrogen phos-
phate ⁄ disodium hydrogen phosphate, 20 mm NaCl, 2 mm
dithiothreitol, 0.01% sodium azide, pH 6.8) and concen-
trated to  0.7 mm.
Purification of the C-terminal domain of PS
The cell lysate was passed through a DEAE anion exchange
column. The protein was eluted with a linear gradient of
NaCl (0–500 mm), and the fractions containing the protein
were pooled together after being checking by denaturing
SDS ⁄ PAGE. The protein was passed through an immobi-
lized metal ion affinity chromatography (Ni
2+
) column.
The protein, which binds weakly to the column, was eluted
with a linear gradient of imidazole (0–500 mm).
TEV protease digestion
Cleavage of target proteins from the fusion host was
achieved following dialysis of the fusion proteins into the
TEV protease cleavage buffer (50 mm Tris, pH 8.0, 100 mm
NaCl, 5 mm dithiothreitol). The protein and TEV protease
were added in a 5 : 1 molar ratio to the digestion reaction,
which was continued for 16 h at 22 °C. TEV protease was
purified in the laboratory, using the method described by

Geisbrecht et al. [43].
Separation of the C-terminal domain of PS from
the fusion host
The cleaved C-terminal domain of the PS was separated
from the fusion host as well as the TEV protease by immo-
bilized metal ion affinity chromatography (Ni
2+
) chroma-
tography. Apo-cytb5 and TEV protease bind to the
column, and the protein of interest is collected in the
flow-through. The protein was dialyzed against 20 mm
phosphate buffer (pH 6.8) containing 20 mm NaCl, 2 mm
dithiothreitol, and 0.01% sodium azide.
Purification of the full-length PS
The cell lysate containing the full-length PS was passed
over a Q-Sepharose column. The protein was eluted with a
linearly increasing gradient of NaCl. Fractions containing
PS were pooled, and passed over a Ni
2+
-nitrilotriacetic acid
column. Protein was eluted from the metal affinity column
with a linearly increasing gradient of imidazole. The PS
was further purified in the ATP-bound form by passage
through a Sepharose blue column that is crosslinked with
cibacron blue dye [46], and then eluting the protein with a
buffer containing ATP.
Crystallization and data collection
For protein crystallization, the purified protein was
exchanged in a dilute buffer of 20 mm Tris (pH 8.0) con-
taining 50 mmd-pantoate [47]. The concentration of the

protein was  15 mg ⁄ mL. Crystallization experiments were
performed using the hanging-drop vapor diffusion method.
In each drop, 2 lL of protein was mixed with 2 lL of the
reservoir solution. The best crystals were obtained using
8–15% poly(ethylene glycol) 4000, 40–150 mm NaCl, and
100 mm acetate buffer (pH 5.0), at 20 °C. The crystal with
50 mm pantoate diffracted to 1.7 A
˚
. Images were indexed,
and reflections were integrated using the program mosflm
[48], and scaled and merged with scala [49].
For data collection, the crystals with 50 mm pantoate were
flash-frozen in a cryostream of N
2
gas at 100 K. Diffraction
data were collected on a Mar FRD generator with a Bruker
detector. Data reduction and scaling were performed with
the programs mosflm and scala (ccp4 [30]). Data-process-
ing statistics are given in Table 1. The pantoate-bound
enzyme crystallized in the space group P2
1
2
1
2. Both datasets
collected showed two molecules per asymmetric unit.
Molecular replacement, model building, and
refinement
The crystal structure of the N-terminal domain of the
enzyme was determined by the molecular replacement
method, using the program phaser [29]. The N-terminal

domain (residues 1–176) of subunit A of E. coli PS
(1IHO.pdb) was used as a model for molecular replacement.
Preparation of isotopically enriched samples of
nPS
Uniformly
2
H,
13
C,
15
N enriched ILV - methyl protonated
samples of nPS were prepared following protocols described
previously [50,51]. All perdeuterated samples of nPS were
subjected to unfolding–refolding. The levels of isotopic
enrichment for the different samples described above were
ascertained using electrospray MS.
Samples for NMR spectroscopy
All of the NMR samples (0.6–0.8 mm) were prepared in
20 mm phosphate buffer, 20 mm NaCl, 2 mm dithiothreitol,
and 0.01% NaN
3
(pH 6.8).
Structure and binding studies of nPS K.S. Chakrabarti et al.
708 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
NMR spectroscopy
NMR spectra were acquired on a Bruker-Avance spectrom-
eter operating at a proton frequency of 700 MHz, equipped
with a 5 mm triple-resonance cryoprobe fitted with a single
(z-axis) pulsed field gradient accessory. The
15

N relaxation
datasets [52] were acquired on a Varian INOVA 600 MHz
spectrometer equipped with a cryogenically cooled triple-
resonance (z-axis) pulsed field gradient probe. All NMR
spectra were acquired at 303 K. NMR data were processed
using nmrpipe ⁄ nmrdraw [53], and assigned using ansig
[54,55].
NMR titration studies
The stock solutions of substrates (100 mm) were prepared
by dissolving weighed amounts of substrates in water. The
d-pantolactone solution was hydrolyzed to d-pantoate as
described previously [47]. The substrates were added to pro-
tein solutions (0.5 mm) to final concentrations of 100 lm,
500 lm,1mm,5mm and 10 mm in buffers containing 90%
H
2
O ⁄ 10% D
2
O.
Chemical shift mapping
The changes in chemical shifts as a function of substrate
concentration were monitored by performing
1
H–
15
N
TROSY experiments with a digital resolution of  5Hz
per point for the
1
H dimension and  16.63 Hz per point

in the
15
N dimension. Shifts were calculated using the
equation
D ¼½Dd
2
HN
þð0:1Dd
N
Þ
2

1=2
where D is the cumulative chemical shift deviation, Dd
HN
is
the change in chemical shift of the proton, and Dd
N
is the
change in the chemical shift of the nitrogen [56].
Modeling and comparison of the structures
of PSs
dali [57] was used to superimpose the crystal structures of
PSs from E. coli and M. tuberculosis (1IHO.pdb and
1MOP.pdb) for structural alignment. The sequence align-
ment of PSs from E. coli, M. tuberculosis and A. thaliana
was achieved using clustalw2 [58]. Finally, modeller [59]
was used to model the A. thaliana PS structure, using the
structures of PSs from E. coli and M. tuberculosis.
Structural analyses

Analysis of the sites of substrate binding from NMR data,
and structural comparison of the E. coli, M. tuberculosis
and A. thaliana PSs with nPS, was performed using vmd
[60] or pymol [61].
Acknowledgements
The NMR and MS facilities at the Indian Institute of
Science are funded from grants by DBT and DST. K.
S. Chakrabarti is a recipient of a CSIR Senior
Research Fellowship. The authors thank K. V. Srini-
vas for help with preparing the clone of nPS, A. Aro-
ra, CDRI, Lucknow, for help in collecting the
relaxation data, and N. Srinivasan and S. Yamunadevi
for their help with modeling and analyzing the
A. thaliana PS protein.
References
1 Umbarger E & Brown B (1958) Isoleucine and valine
metabolism in Escherichia coli. J Biol Chem 233,
1156–1160.
2 Webb ME, Smith AG & Abell C (2004) Biosynthesis of
pantothenate. Nat Prod Rep 21, 695–721.
3 Miyatake K, Nakano Y & Kitaoka S (1979) Pantothe-
nate synthetase from Escherichia coli. Methods Enzymol
62, 215–219.
4 Cronan JE Jr, Little KJ & Jackowski S (1982) Genetic
and biochemical analysis of pantothenate biosynthesis
in Escherichia coli and Salmonella typhimurium.
J Bacteriol 149, 916–922.
5 Merkel WK & Nicholas BP (1996) Characterization
and sequence of the Escherichia coli panBCD gene
cluster. FEMS Microbiol Lett 143, 247–252.

6 Umbarger HE & Neidhardt FC (1996) Escherichia coli
and Salmonella typhimurium: Cellular and Molecular
Biology. American Society for Microbiology,
Washington, DC.
7 Zheng R & Blanchard JS (2001) Steady-state and
pre-steady-state kinetic analysis of Mycobacterium
tuberculosis pantothenate synthetase. Biochemistry 40,
12904–12912.
8 Genschel U, Powell CA, Abell C & Smith AG
(1999) The final step of pantothenate biosynthesis
in higher plants: cloning and characterization
of pantothenate synthetase from Lotus japonicus
and Oryza sativum (rice). Biochem J 341, 669–678.
9 Jonczyk R & Genschel U (2006) Molecular adaptation
and allostery in plant pantothenate synthetase. J Biol
Chem 281, 37435–37446.
10 Jonczyk R, Roncolni S, Rychlik M & Genschel U
(2008) Pantothenate synthetase is essential but not limit-
ing for pantothenate biosynthesis in Arabidopsis. Plant
Mol Biol 66, 1–14.
11 Wang S & Eisenberg D (2003) Crystal structures of a
pantothenate synthetase from M. tuberculosis and its
complex with substrates and a reaction intermediate.
Protein Sci 12, 1097–1108.
12 Wang S & Eisenberg D (2006) Crystal structure of the
pantothenate synthetase from Mycobacterium
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 709
tuberculosis, snapshots of the enzyme in action. Bio-
chemistry 45, 1554–1561.

13 von Delft F, Lewenden A, Dhanaraj V, Blundell TL,
Abell C & Smith AG (2001) The crystal structure of
E. coli pantothenate synthetase confirms it as a member
of the cytidylyltransferase superfamily. Structure 9,
439–450.
14 Golovanov AP, Hautbergue MG, Wilson AS & Lian L
(2004) A simple method for improving protein solubility
and long-term stability. J Am Chem Soc 126,
8933–8939.
15 Schobert B & Tschesche H (1978) Unusal solution
properties of proline and its interaction with proteins.
Biochim Biophys Acta 541, 270–277.
16 Samuel D, Kumar TK, Ganesh G, Jayaraman G, Yang
PW, Chang MM, Trivedi VD, Wang SL, Hwang KC,
Chang DK et al. (2000) Proline inhibits aggregation
during protein refolding. Protein Sci 9, 344–352.
17 Mitra A & Sarma SP (2008) Escherichia coli ilvN
interacts with the FAD binding domain of ilvB and
activates the AHAS I enzyme. Biochemistry 47 ,
1518–1531.
18 Grzesiek S, Anglister J & Bax A (1993) Correlation of
backbone amide and aliphatic side-chain resonances in
13
C ⁄
15
N-enriched proteins by isotropic mixing of
13
C
magnetization. J Magn Reson 101B, 114–119.
19 Fushman D, Weismann R, Thuring H & Ruterjans H

(1994) Backbone dynamics of ribonuclease T1 and its
complex with 2¢GMP studied by two-dimensional
heteronuclear NMR spectroscopy. J Biomol NMR 4,
61–78.
20 Pervushin K, Riek R, Wider G & Wuthrich K (1997)
Attenuated T2 relaxation by mutual cancellation of
dipole–dipole coupling and chemical shift anisotropy
indicates an avenue to NMR structures of very large
biological macromolecules in solution. Proc Natl Acad
Sci USA 94, 12366–12371.
21 Salzmann M, Pervushin K, Wider G, Senn H & Wuth-
rich K (1998) TROSY in triple-resonance experiments:
new perspectives for sequential NMR asignment of
large proteins. Proc Natl Acad Sci USA 95 , 13585–
13590.
22 Salzmann M, Wider G, Pervushin K, Senn H &
Wuthrich K (1999) TROSY-type triple-resonance
experiments for sequential NMR assignments of large
proteins. J Am Chem Soc 121, 844–848.
23 Yamazaki T, Lee W, Arrowsmith CH, Muhandiram
DR & Kay LE (1994) A suite of triple resonance NMR
experiments for the backbone assignment of
15
N,
13
C,
2
H labelled proteins with high sensitivity. J Am Chem
Soc 116, 11655–11666.
24 Chakrabarti KS & Sarma Siddhartha P (2006) NMR

assignment of
2
H,
13
C and
15
N labeled amino-terminal
domain of apo-pantothenate synthetase from E. coli.
J Biomol NMR 36, 38.
25 Spera S & Bax A (1991) Empirical correlation between
protein backbone conformation and C
a
and C
b
13C
nuclear magnetic resonance chemical shifts. J Am Chem
Soc 113, 5490–5492.
26 Wishart DS & Sykes BD (1994) The
13
C chemical-shift
index: a simple method for the identification of protein
secondary structure using
13
C chemical-shift data.
J Biomol NMR 4, 171–180.
27 Wu
¨
thrich K (1986) NMR of Proteins and Nucleic Acids.
Wiley, New York.
28 Cornilescu G, Delaglio F & Bax A (1999) Protein

backbone angle restraints from searching a database for
chemical shift and sequence homology. J Biomol NMR
13, 289–302.
29 McCoy AJ, Grosse-Kunstleve RW, Storoni LC &
Read RJ (2005) Likelihood-enhanced fast translation
functions. Acta Crystallogr D Biol Crystallogr 61,
458–464.
30 Collaborative Computational Project, Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr D Biol Crystallogr 50, 760–763.
31 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by the
maximum-likelihood method. Acta Crystallogr D Biol
Crystallogr 53, 240–255.
32 Ramachandran GN, Ramakrishnan C & Sasisekharan
V (1963) Stereochemistry of polypeptide chain configu-
ration. J Mol Biol 7, 95–99.
33 Ramachandran GN & Sasisekharan V (1968) Confor-
mation of polypeptides and proteins. Adv Protein Chem
23, 283–438.
34 Panigrahi SK & Desiraju GR (2007) Strong and weak
hydrogen bonds in the protein–ligand interface. Proteins
Struct Funct Bioinf 67, 128–141.
35 Gopalakrishnan B, Aparna V, Jeevan J, Ravi M &
Desiraju GR (2005) A virtual screening approach for
thymidine monophosphate kinase inhibitors as antitu-
bercular agents based on docking and pharmacophore
models. J Chem Inf Model 45, 1101–1108.
36 Sarkhel S & Desiraju GR (2004) N–H O, O–H O,
and C–H O hydrogen bonds in protein–ligand

complexes: strong and weak interactions in molecular
recognition. Protein Struct Funct Bioinform 54, 247–259.
37 Mehrotra S & Balaram H (2007) Kinetic characteriza-
tion of adenylosuccinate synthetase from the thermo-
philic archaea Methanocaldococcus jannaschii.
Biochemistry 46, 12821–12832.
38 Iancu CV, Borza T, Fromm HJ & Honzatko RB (2002)
IMP, GTP, and 6-phosphoryl-IMP complexes of recom-
binant mouse muscle adenylosuccinate synthetase.
J Biol Chem 277, 26779–26787.
39 Borza T, Iancu CV, Pike E, Honzatko RB & Fromm
HJ (2003) Variations in the response of mouse isozymes
of adenylosuccinate synthetase to inhibitors of physio-
logical relevance. J Biol Chem 278, 6673–6679.
Structure and binding studies of nPS K.S. Chakrabarti et al.
710 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS
40 Oniciu DC, Bell RPL, McCosar BH, Bisgainer CL,
Dasseux J-LH, Verdijk D, Relou M, Smith D, Regdink
H, Leemhuis FMC et al. (2006) Syntheses of pantolac-
tone and pantothenic acid derivatives as potential lipid
regulating agents. Synth Commun 36, 365–391.
41 Tuck KL, Saldanha SA, Birch LM, Smith AG & Abell C
(2006) The design and synthesis of inhibitors of
pantothenate synthetase. Org Biomol Chem 4, 3598–3610.
42 White EL, Southworth K, Ross L, Cooley S, Gill RB,
Sosa MI, Manouvakhova A, Rasmussen L, Goulding
C, Eisenberg D et al. (2006) A novel inhibitor of
Mycobacterium tuberculosis pantothenate synthetase.
J Biomol Screen 12, 1–6.
43 Geisbrecht BV, Bouyain S & Pop M (2006) An

optimized system for expression and purification of
secreted bacterial proteins. Protein Expr Purif 46, 23–32.
44 Blattner FR, Plunkett GI, Bloch CA, Perna NT,
Burland V, Riley M, Collado-Vides J, Glasner JD,
Rode CK, Mayhew GF et al. (1997) The complete
genome sequence of Escherichia coli K-12. Science 277,
1453–1462.
45 Mitra A, Chakrabarti KS, Hameed MSS, Srinivas KV,
Kumar GS & Sarma SP (2005) High level expression of
peptides and proteins using cytochrome b5 as a fusion
host. Protein Expr Purif 41, 84–97.
46 Scopes RK (1994) Protein Purification, Principles and
Practice. Springer, New York.
47 Stansly PG & Schlosser ME (1945) The biological
activity of pantolactone and pantoic acid. J Biol Chem
161, 513–515.
48 Leslie AGW (1992) Recent changes to the MOSFLM
package for processing film and image plate data. Joint
CCP4 + ESF-EAMCB Newslett Prot Crystallogr, No.
26.
49 Evans P (2006) Scaling and assessment of data quality.
Acta Crystallogr D Biol Crystallogr 62 , 72–82.
50 Goto NK, Gardner K, Mueller GA, Willis RC & Kay
LE (1999) A robust and cost-effective method for the
production of Val, Leu, Ile (d1) methyl-protonated
15
N-,
13
C-,
2

H-labelled proteins. J Biomol NMR 13 , 369–
374.
51 Gardner KH & Kay LE (1997) Production and incor-
poration of
15
N,
13
C,
2
H(
1
H-d1 methyl)isoleucine into
proteins for multidimensional NMR studies. JAm
Chem Soc 119, 7599–7600.
52 Farrow NA, Muhandiram R, Singer AU, Pascal SM,
Kay CM, Gish G, Shoelson SE, Pawson T, Forman-
Kay JD & Kay LE (1994) Backbone dynamics of a free
and phosphopeptide-complexed Src homology 2 domain
studied by
15
N NMR relaxation. Biochemistry 33,
5984–6003.
53 Delaglio F, Grsesiek S, Vuister GW, Zhu G, Pfeifer J &
Bax A (1995) NMRPipe: a multidimensional spectral
processing system based on UNIX pipes. J Biomol
NMR 6, 277–293.
54 Kraulis PJ (1989) ANSIG: a program for the assign-
ment of protein 1H 2D NMR spectra by interactive
graphics. J Magn Reson 84, 627–633.
55 Kraulis PJ, Domaille PJ, Campbell-Burk SL, van Aken

T & Laue ED (1994) Solution structure and dynamics
of Ras p21.GDP determined by heteronuclear three-
and four-dimensional NMR spectroscopy. Biochemistry
33, 3515–3531.
56 Cavanagh J, Fairbrother WJ, Palmer AG III, Rance M
& Skelton NJ (2007) Protein NMR Spectroscopy, 2nd
edn. Elsevier Academic Press, New York, NY, USA.
57 Holm L & Sanders C (1996) Mapping the protein uni-
verse. Science 273, 595–693.
58 Larkin MA, Blackshields G, Brown NP, Chenna R,
McGettigan PA, McWilliam H, Valentin F,
Wallace IM, Wilm A, Lopez R et al. (2007)
ClustalW2 and ClustalX version 2. Bioinformatics 23,
2947–2948.
59 Eswar N, Marti-Renom MA, Webb B, Madhusudan
MS, Eramian D, Shen M, Pieper U & Sali A (2007)
Comparative Protein Structure Modeling with
MODELLER. Curr Protoc Protein Sci. Nov; Chapter 2:
Unit 2.9.
60 Humphrey W, Dalke A & Schulten K (1996) VMD –
Visual Molecular Dynamics. J Mol Graph 14, 33–38.
61 DeLano WL (2002) The PyMOL Molecular Graphics
System. .
Supporting information
The following supplementary material is available:
Fig. S1. Sequential connectivity walk along the protein
backbone for Glu5–Glu20 of nPS.
Fig. S2. Secondary structure distribution of nPS in
solution along with the sequence, short-range NOE
pattern, and secondary chemical shifts of

13
Ca and
13
Cb atoms.
Fig. S3. Two-dimensional projection showing the pan-
toate–protein interactions in the canonical pantoate-
binding site (site I).
Fig. S4. Two-dimensional projection showing the pan-
toate–protein interactions in the ATP-binding site
(site II).
Fig. S5. Model structure of pantothenate synthetase
from Arabidopsis (green), constructed using modeller,
superimposed on the crystal structures of PS from
E. coli (1IHO.pdb; blue) and M. tuberculosis
(1MOP.pdb; red).
Table S1. E. coli residues (black) and aligned residues
in A. thaliana (pink).
Table S2. M. tuberculosis residues (black) and aligned
residues in A. thaliana (pink).
Table S3. List of atom–atom interactions across the
dimer interface.
K.S. Chakrabarti et al. Structure and binding studies of nPS
FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS 711
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising

from supporting information (other than missing files)
should be addressed to the authors.
Structure and binding studies of nPS K.S. Chakrabarti et al.
712 FEBS Journal 277 (2010) 697–712 ª 2010 The Authors Journal compilation ª 2010 FEBS