1742
S. Kumar et al. / Biochimica et Biophysica Acta 1844 (2014) 1741–1748
at the C-terminus, and the C-terminal residues of SAT bind to the active
site of OASS, forming a cysteine synthase complex, which in turn
reduces the activity of OASS [18]. Previous structural and biochemical
studies have revealed that formation of cysteine synthase complexes
requires SAT to have an isoleucine residue, at its C-terminal end [9].
The BaSAT C-terminal end sequence is Gly, Asp, Gly, and Ile, which
should thus favor formation of the complex. BaSAT has been shown to
completely inhibit Leishmania donovanii OASS [19], indicating that this
SAT can form a cysteine synthase complex. OASS, as described above,
is the second enzyme of pathway, and is structurally and biochemically
well characterized in many organisms such as E. coli (PDB id 2BHT) [20],
M. tuberculosis (PDB id 2Q3D) [21], Arabidopsis thaliana (PDB id 1Z7W)
[22], and E. histolytica (PDB id 2PQM) [23].
Here we report the crystal structure of SAT from B. abortus strain
S19, both in its apo (uncomplexed) state at a 1.97-Å resolution and in
complex with CoA at a 1.87-Å resolution. Kinetic studies of BaSAT
with serine as well as with acetyl CoA are also reported. The inhibition
kinetics with final product cysteine shows competitive inhibition with
substrate serine and mixed inhibition with acetyl coenzyme A. The
structural, biophysical, and biochemical experiments indicate that
BaSAT exists as a hexamer while gel filtration indicates a trimeric
state. Comparison of apo and CoA complex crystal structures show
there is no structural difference between them, indicating lock and
key binding.
2.4. Data collection and processing
Crystals of BaSAT were equilibrated in cryo-solutions containing the
crystallization condition but with concentrations of PEG 3350 increas-
ing from 15% to 30%. These cryo-protected crystals were mounted in
cryoloops, and initial data were collected at the AIRF, JNU, New Delhi,
using a Bruker Microstar generator and MAR 345 imaging plate
detector. High-resolution (up to 2 Å) diffraction data were collected at
the DBT-European Synchrotron Radiation Facility beamline BM14.
The data set was indexed, integrated, and scaled using the HKL-2000
data processing software [24]. BaSAT-coenzyme A complex crystal dif-
fraction data were collected at the home source (AIRF, JNU). Data
were indexed, integrated, and scaled using Automar (Mar-research,
Germany) (Table 1).
2.5. Structure determination and refinement
The BaSAT structure was determined by the molecular replacement
method using B. melitensis SAT as a search model (PDB id 3MC4,
Abendroth et al., unpublished, Seattle Structural Genomics Center for
Infectious Disease). This structure shares very high sequence similarity
(~99%) with BaSAT. Crystals belonged to the P21 space group, and
Matthews coefficient [25] calculations suggested that there could
be twelve protomers in an asymmetric unit. Aligned sequences of
B. abortus SAT and Brucella melintensis SAT were input into Chainsaw
[26], and its output was used in the Molrep [27] module of CCP4 [28]
for molecular replacement. A single protomer, trimer, and hexamer
were used for molecular replacement. The latter yielded the best solu-
tion with lowest initial R value of 46.3% where two hexamers (a total
of 12 protomers) were present in an asymmetric unit. A single step of
10 cycles of refinement in REFMAC5 [29] resulted in a decrease of the
Rwork to 21.9%. This model was further refined by rounds of iterative
model building using the COOT graphics package [30] in combination
with REFMAC5 and phenix.refine [31]. The remaining residues were
built manually with COOT guided by the σ-weighted 2Fo-Fc and Fo-Fc
2. Materials and methods
2.1. Cloning, over-expression, and purification of B. abortus SAT
The cloning and over-expression of B. abortus SAT (BaSAT) is ex-
plained in detail in supplementary data. The over-expressed protein
was purified using Ni-NTA column and gel filtration chromatography
(Supplementary Fig. 1). The pure protein fractions were pooled and
concentrated using Centricon concentrators (Amicon, Millipore) and
taken for crystallization trials. The proteins were concentrated to
14 mg/ml in gel filtration elution buffer containing 50 mM Tris–HCl
(pH 8.0), 150 mM NaCl, 5% glycerol, and 5 mM β-mercaptoethanol.
Table 1
Data collection and refinement statistics.
Data set
BaSAT-native
BaSAT-CoA
2.2. Crystallization of BaSAT
X-ray source
Wavelength (Å)
Space group
DBT-BM14, ESRF, France Bruker Microstar
0.9787
1.54179
P21
R3
Crystallization trials of the purified BaSAT were carried out using
various pre-formulated screens in 96-well plates (Molecular Dimen-
sions, UK), with 100 μl of reservoir solution in each well. Hanging
drops of 200 nl each of protein and precipitant were dispensed using
the Mosquito™ (Molecular Dimensions, UK) crystallization robot
(at Advanced Instrumentation Research Facility (AIRF), JNU, New
Delhi) for vapor diffusion, and plates were incubated at 289 K or 277
K. The initial crystals were obtained using the Morpheus™ screen (Mo-
lecular Dimensions, UK). After replicating the conditions in 24-well
plates (Hampton Research, USA) and optimizing various physicochem-
ical parameters, the best crystals of BaSAT were obtained using 7% (w/v)
PEG 1000, 7% (w/v) PEG 3350, 5% (v/v) MPD, 0.12 M ethylene glycol, 0.1
M Tris–HCl (pH 7.4), and 50 mM MgCl2 at 289 K.
Unit cell parameters a, b, c (Å)
α, β, γ (°)
Resolution range (Å)
Completeness (%)
Unique HKLs
73.7, 256.8, 82.3
90.00, 91.20, 90.00
50–1.96 (1.99–1.96)
86.0 (83.0)
183943 (76642)
2.4 (2.2)
30.4 (2.9)
0.6
3.8 (32.8)
104.3, 104.3, 105.7
90.00, 90.00, 120.00
68.68–1.87 (1.92–1.87)
99.2 (92.3)
331387 (35422)
9.18 (8.1)
13.6 (1.8)
0.45
6.8 (56.9)
Multiplicity
Average bI/σIN
Crystal mosaicity (°)
Rmerge (%)
Refinement
Rwork (%)
18.7
24.3
18.3
23.4
Rfree (%)
Mean B factor (Å2)
Protein
36.5 (31.4 from
Wilson plot)
–
29.4 (27.8 from
Wilson plot)
67.2
CoA
Number of atoms
Protein
Water
22448
1158
–
3788
149
48
2.3. Crystallization of BaSAT in complex with CoA
CoA
Other
16
9
BaSAT protein (14 mg/ml) was mixed with 0.1 mM coenzyme A
(Sigma Aldrich, USA) and 2 mM L-serine prior to crystallization trials.
Three microliters of protein and 3 μl of reservoir solution were mixed
and allowed to equilibrate against the reservoir of 500 μl in 24-well
plates. The best crystals appeared in the condition containing 5% PEG
1000, 5% PEG 1500, 5% MPD, 0.1 M ethylene glycol, 0.1 M Tris–HCl
(pH 7.4), and 20–50 mM MgCl2 at 289 K.
RMS deviations
Bond length (Å)
Bond angles (°)
Ramachandran Plot Statistics
0.013
1.7
0.013
1.6
Favored/allowed/disallowed (%) 96.4/3.4/0.2
97.8/2.2/0
The values in the parenthesis are for highest resolution shell. Rfree was calculated using a
randomly selected subset of 5% of the reflections.