Structural and kinetic studies on adenylosuccinate lyase from Mycobacterium smegmatis and Mycobacterium tuberculosis provide new insights on the catalytic residues of the enzyme

Article abstract of DOI:10.1111/febs.12730

Adenylosuccinate lyase (ASL), an enzyme involved in purine biosynthesis, has been recognized as a drug target against microbial infections. In the present study, ASL from Mycobacterium smegmatis (MsASL) and Mycobacterium tuberculosis (MtbASL) were cloned, purified and crystallized. The X-ray crystal structure of MsASL was determined at a resolution of 2.16 A. It is the first report of an apo-ASL structure with a partially ordered active site C3 loop. Diffracting crystals of MtbASL could not be obtained and a model for its structure was derived using MsASL as a template. These structures suggest that His149 and either Lys285 or Ser279 of MsASL are the residues most likely to function as the catalytic acid and base, respectively. Most of the active site residues were found to be conserved, with the exception of Ser148 and Gly319 of MsASL. Ser148 is structurally equivalent to a threonine in most other ASLs. Gly319 is replaced by an arginine residue in most ASLs. The two enzymes were catalytically much less active compared to ASLs from other organisms. Arg319Gly substitution and reduced flexibility of the C3 loop might account for the low catalytic activity of mycobacterial ASLs. The low activity is consistent with the slow growth rate of Mycobacteria and their high GC containing genomes, as well as their dependence on other salvage pathways for the supply of purine nucleotides.

Full text of DOI:10.1111/febs.12730

Structural and kinetic studies on adenylosuccinate lyase
from Mycobacterium smegmatis and
Mycobacterium tuberculosis provide new insights on the
catalytic residues of the enzyme
Sanchari Banerjee1, Monika J. Agrawal1,*, Diptimayee Mishra1,*, Siddharth Sharan1,
Hemalatha Balaram2, Handanhal S. Savithri3 and Mathur R. N. Murthy1
1 Molecular Biophysics Unit, Indian Institute of Science, Bangalore, India
2 Molecular Biology and Genetics Unit, Jawaharlal Nehru Centre of Advanced Scientific Research, Bangalore, India
3 Department of Biochemistry, Indian Institute of Science, Bangalore, India
Keywords
Adenylosuccinate lyase (ASL), an enzyme involved in purine biosynthesis,
has been recognized as a drug target against microbial infections. In the
adenylosuccinate lyase; catalytic activity;
purine nucleotide supply; slow growth rate;
present study, ASL from Mycobacterium smegmatis (MsASL) and Myco-
bacterium tuberculosis (MtbASL) were cloned, purified and crystallized.
The X-ray crystal structure of MsASL was determined at a resolution of
X-ray crystallography
Correspondence
ꢀ
2.16 A. It is the first report of an apo-ASL structure with a partially
ordered active site C3 loop. Diffracting crystals of MtbASL could not be
obtained and a model for its structure was derived using MsASL as a tem-
plate. These structures suggest that His149 and either Lys285 or Ser279 of
MsASL are the residues most likely to function as the catalytic acid and
base, respectively. Most of the active site residues were found to be con-
served, with the exception of Ser148 and Gly319 of MsASL. Ser148 is
structurally equivalent to a threonine in most other ASLs. Gly319 is
replaced by an arginine residue in most ASLs. The two enzymes were cata-
lytically much less active compared to ASLs from other organisms.
Arg319Gly substitution and reduced flexibility of the C3 loop might
account for the low catalytic activity of mycobacterial ASLs. The low
activity is consistent with the slow growth rate of Mycobacteria and their
high GC containing genomes, as well as their dependence on other salvage
pathways for the supply of purine nucleotides.
M. R. N. Murthy, Molecular Biophysics
Unit, Indian Institute of Science,
Bangalore – 560012, India
Fax: +91 80 2360 0535
Tel: +91 80 2293 2458
E-mail: mrn@mbu.iisc.ernet.in
*These authors contributed equally to this
work.
(Received 20 November 2013, revised 15
January 2014, accepted 23 January 2014)
doi:10.1111/febs.12730
Structured digital abstract
• purB and purB bind by x-ray crystallography (View interaction)
Abbreviations
AICAR, 5-aminoimidazole-4-carboxamide ribotide; ASL, adenylosuccinate lyase; BaASL, Bacillus anthracis ASL; BsASL, Bacillus subtilis ASL;
CeASL, Caenorhabditis elegans ASL; CMLE, 3-carboxy-cis,cis-muconate lactonizing enzyme; EcASL, Escherichia coli ASL; ESRF, European
Synchrotron Radiation Facility; FtASL, Francisella tularensis ASL; HsASL, Homo sapiens ASL; LdASL, Leishmania donovani ASL; LpASL,
Legionella pneumophila ASL; MsASL, Mycobacterium smegmatis ASL; MS, Mycobacterium smegmatis; MtbASL, Mycobacterium
tuberculosis ASL; MTB, Mycobacterium tuberculosis; PaASL, Pyrobaculum aerophilum ASL; PDB, Protein Data Bank; PfASL, Plasmodium
falciparum ASL; PvASL, Plasmodium vivax ASL; SaASL, Staphylococcous aureus ASL; SAMP, succinyl-adenosine monophosphate; TbASL,
Trypanosoma brucei ASL; TmASL, Thermotoga maritima ASL.
1642
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
4-carboxamide ribotide or SAMP [6,7]. A general
scheme for catalysis of SAMP by ASL proposed on
the basis of biochemical studies [5,8,9] is shown in
Fig. 1. It proceeds by abstraction of a proton from Cb
atom of the succinyl group by a catalytic base leading
to the formation of a carbanion intermediate that is
stabilized by resonance [10]. Protonation at N1 or N6
by a catalytic acid leads to Ca-N bond cleavage and
release of AMP and fumaric acid as products.
Introduction
Purine nucleotides are synthesized in eukaryotes and
prokaryotes by two distinct pathways: the low energy
costing salvage pathway and the ATP dependent
de novo pathway. Both pathways are active in most
organisms. However, in obligate protozoan parasites
such as Toxoplasma gondii, Plasmodium falciparum
and Leishmania donovani, only the salvage pathway
has been found [1]. The pathway utilized for obtaining
purine may depend on the cellular environment, as
well as the requirements of the given organism. In
Mycobacteria, both pathways are known to exist,
although they have not been extensively investigated
[2–4]. Adenylosuccinate lyase (ASL; EC 4.3.2.2) is the
only enzyme common to both the pathways. In the de
novo biosynthetic pathway, it displays dual substrate
specificity and catalyzes two nonsequential reactions.
The same active site is involved in both the reactions.
These two reactions correspond to the cleavage of
5-aminoimidazole-(N-succinylocarboxamide) ribotide
to 5-aminoimidazole-4-carboxamide ribotide and
fumarate in one step and cleavage of succinyl-adeno-
sine monophosphate (SAMP) to AMP and fumarate
in another step [5]. It has been demonstrated that the
reactions involve general acid–base catalysis with
b-elimination of the succinyl group from the substrates
leading to the release of fumarate, which leaves the
active site before the release of 5-aminoimidazole-
ASL is a housekeeping enzyme that is found in
many organisms. It plays a crucial role in cellular rep-
lication, purine nucleotide cycle [11] and cellular
metabolism [12]. In humans, point mutations in the
purB gene encoding ASL that lead to functional defi-
ciency cause autosomal recessive disorders such as aut-
ism, mental retardation, epilepsy and degeneration of
muscles [13–15]. It has also been observed that the
nucleotide requirements of most normally dividing
human cells (except liver and activated T-lymphocytes)
are derived from the salvage pathway, whereas those
of rapidly dividing tumor cells are derived by de novo
biosynthesis [16], making ASL an attractive target for
anticancer drugs [17]. Additionally, ASL could also be
a suitable target for antibacterial drugs [18].
In the present study, we determined the X-ray crys-
tal structure of ASL from nonpathogenic Mycobacte-
ꢀ
rium smegmatis (MsASL) at a resolution of 2.16 A.
We also crystallized ASL from pathogenic Mycobacte-
rium tuberculosis (MtbASL). However, these crystals
Fig. 1. General scheme showing cleavage
of SAMP by ASL. The five steps involved
are: (1) Proton abstraction by the catalytic
base (B) from the CB atom of succinyl
group; (2) stabilization of the carbanion
intermediate formed; (3) protonation at N1
or N6 atom of AMP moiety by a catalytic
acid; (4) cleavage of the Ca-N bond
between AMP and succinyl groups; and
(5) release of the products.
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1643

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
Table 1. Data collection statistics.
did not diffract X-rays. In the absence of diffracting
crystals, a structural model for MtbASL was obtained
by homology modeling using MsASL structure as a
template. These unliganded structures reveal the most-
open conformation of the C3 loop observed so far in
the structures of aspartase–fumarase superfamily of
enzymes. The partially ordered C3 loop in MsASL
structure provides a basis for identification of the cata-
lytic residues. Structural comparison and phylogenetic
analysis show that the mycobacterial enzymes are
more similar to the human enzyme rather than other
bacterial ASLs. A comparison of the kinetic constants
of ASL from several organisms demonstrates that the
mycobacterial enzymes have the lowest catalytic activ-
ity. The activity of MtbASL is particularly low. This
could be a result of the slow growth rate of Mycobac-
teria and the high GC content of their genomes, as
well as their dependence on adenosine kinase (a sal-
vage pathway enzyme) for the supply of AMP. Our
structural results for MsASL could be used as a model
system to understand the structure–function relation-
ship in MtbASL. The structure of MsASL will also be
useful for exploring possibilities with respect to the
design of specific inhibitors.
MsASL
Wavelength (nm)
ꢀ
Resolution limits (A)
ꢀ
Highest resolution shell (A)
0.9537
50.0–2.16
2.28–2.16
5
Exposure time (s)
Crystal to detector distance (mm)
Space group
ꢀ
Cell parameters (A)
240
P21212
a = 74.6, b = 176.83,
c = 73.3
2
8.3 (27.4)a
253 669
52 292
Protomers in the asymmetric unit
Rmerge (%)
Total observations
Unique observations
Mean redundancy
Completeness (%)
Mean (I/rI)
4.85
98.9 (96.4)a
11.3 (4.2)a
aValues given in parentheses are for reflections of the highest reso-
ꢀ
lution shell extending from 2.28 to 2.16 A.
Table 2. Refinement statistics.
MsASL
ꢀ
Resolution (A)
50.0–2.16
49 552
20.4
Reflections used in refinementa
Rwork (%)a
Rfree (%)b
23.9
Results and Discussion
c
ꢀ
rms bond lengths (A)
0.0044
0.76
rms bond angles (º)c
Ramachandran statistics
1 Most favored (%)
MsASL structure determination
92.6
6.4
Crystals of MsASL diffracted X-rays to a resolution
ꢀ
of 2.16 A at the BM-14 beam line of the European
2 Additionally allowed (%)
3 Generously allowed (%)
4 Outliers (%)
1.0
Synchrotron Radiation Facility (ESRF). The data col-
lection statistics for these crystals are summarized in
Table 1. The structure of MsASL could be determined
using polyalanine coordinates of human ASL as the
phasing model. The model obtained was manually
adjusted and the side chains were built. After refine-
ment of positional parameters, 518 water molecules
were added based on 2Fo-Fc and Fo-Fc maps. Rwork
and Rfree factors of the final refined structure are
20.4% and 23.9%, respectively. The final refinement
statistics are summarized in Table 2.
0.0
2
ꢀ
Mean B-factor for protein atoms (A )
ꢀ
Mean B-factor for water atoms (A )
20.8
34.0
2
aRwork defined as Σ(Fo–Fc)/ΣFo, where Fo and Fc refer to observed
and calculated structure factor amplitudes. Rfree is defined similarly
for the reflections that were excluded from the refinement.
b5% of these reflections were not included in the refinement and
were used for calculation of R-free.
crmsd of the observed bond lengths and bond angles from the
standard values.
Electron density is not observed in both the protomers
for the first three residues at the N-terminus and the
last three at the C-terminus. Density is absent for resi-
due 76 in protomer A and residue 49 in protomer B,
as well as for a short stretch of three residues (90–92)
in both A and B. A long stretch of residues from 272–
287 that corresponds to the C3 loop (close to the
active site and harbouring the signature sequence
Model quality of MsASL
The final electron density map was of good quality
throughout most of the polypeptide main chain
(Fig. 2A). Two protomers, A and B are present in the
asymmetric unit of the crystal. A crystallographic two-
fold axis generates the other two protomers (C and D)
of the tetrameric MsASL. The observed electron den-
sity for the two protomers in the asymmetric unit is
largely similar, although with some minor differences.
‘
ASL structures determined so far [12,19,20]. Density
278GSSxxPxKxN287′) is disordered in all the native
1644
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
and seven of these residues are common to both the
protomers.
A
Validation of the structure using PROCHECK showed
that 92.6%, 6.4% and 1.0% of the residues are in the
most favored, additionally allowed and generously
allowed regions, respectively, of the Ramachandran
map [21]. There were no outliers in the final structure.
No additional density that might correspond to bound
ligands apart from water molecules was observed.
X-ray crystal structure of MsASL
ASL belongs to the aspartase/fumarase superfamily of
enzymes, all members of which share characteristic ter-
tiary and quaternary structures. This superfamily also
includes argininosuccinate lyase, d2-crystallin, class II
fumarase, L-aspartase and 3-carboxy-cis,cis-muconate
lactonizing enzyme (CMLE). The sequence identity
across the members of the superfamily is ~ 10–20%,
yet they are uniquely characterized by a common fold
and their ability to catalyze b-elimination reactions of
succinyl containing substrates leading to the formation
of fumarate, except in the CMLE catalyzed reaction,
which results in the formation of lactone [22].
B
MsASL structure adheres to the fold of the enzymes
of the aspartase/fumarase superfamily. Each protomer
of MsASL is composed of three distinct domains: N-
terminal D1 (residues 1–100), central D2 (residues
101–366) and C-terminal D3 (367–474) domains
arranged in the shape of a dumbbell (Fig. 3A). The
structure is predominantly helical, with each protomer
consisting of 22 helical segments and four b-strands.
The four strands that occur in domain D2 are orga-
nized as two anti-parallel b-sheets, each consisting of
two strands (b-sheet I: residues 142–148 and 151–157;
b-sheet II: residues 267–269 and 353–354).
ASL is biologically active only as a homotetramer
as the active site residues are from three highly con-
served regions, C1 (residues 95–110), C2 (residues 146–
157) and C3 (residues 272–288) contributed by three
independent protomers [23]. MsASL forms a tight tet-
ramer (Fig. 3B) with three independent stable inter-
faces (A/B, A/D, A/C) associated with buried surface
Fig. 2. Quality of the final electron density map. (A) Stereo view of
a representative section (residues 238–256) of the map contoured
at 1.5 r along with the fitted polypeptide chain (pink). (B) Stereo
view of the 2Fo-Fc electron density map for the partially ordered
residues 271–287 of the C3 loop (red), which is disordered in
most of the apo-ASL structures. The map has been contoured at
0.5 r.
2
2
2
ꢀ
ꢀ
ꢀ
areas of 5890 A , 3620 A and 3270 A respectively.
There are 58, 12 and 15 hydrogen bonds and 12, eight
and three salt bridges in the three interfaces, respec-
tively.
was observed for residues 272–279 in protomer A,
although at a lower contour level of 0.5 r (Fig. 2B).
The density for C3 loop residues 272–284 is absent in
protomer B. Residues Arg17, Glu115 and Gln360 have
been built in two alternate conformations, each with
half occupancy. The side chains of 11 residues in pro-
tomer A and 18 residues in protomer B have been
truncated according to the observed electron density
The C3 loop, which is disordered in most apo-ASL
structures, could be built in protomers A and D
related to the crystal by two-fold. The mean B-factor
for the main chain atoms of residues 272–287 is
2
ꢀ
71.1 A , which is substantially higher than those of all
2
main chain atoms of protomer A (32.9 A ) and
ꢀ
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1645

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
A
B
Fig. 3. MsASL X-ray crystal structure. Cartoon representation of the (A) protomer and (B) tetramer. The three domains in MsASL protomer
are labeled D1–D3. The N- and C-terminal residues are marked. The four protomers in MsASL tetramer in (B) are labeled P–A to P–D and
colored in blue, yellow, pink and green, respectively. The residues forming the active sites are shown in black. The view in (B) is related to
that of (A) by a rotation of 90º. The inset is an enlarged view of one of the active sites formed by the conserved regions C1, C2 and C3,
contributed by three separate protomers.
2
ꢀ
protomer B (31.7 A ), reflecting their flexibility and
partial disorder. The C3 loop was mostly disordered in
subunits B and C. MsASL is the first reported unli-
ganded ASL structure in which the C3 loop is ordered
at least partially. Upon ligand binding, this loop
becomes ordered in ASLs from Escherichia coli
(EcASL; Protein Data Bank (PDB) code: 2PTQ) [20],
Staphylococcous aureus (SaASL; PDB code: 2X75)
[18], Bacillus anthracis (BaASL; PDB code: 2PFM)
and Trypanosoma brucei (TbASL; PDB code: 4EFC),
whereas it remains disordered in ASLs from Francisel-
la tularensis (FtASL; PDB code: 4EEI) and Homo sapi-
tase/fumarase superfamily with striking resemblance to
MsASL and HsASL. The structural alignment of the
two mycobacterial ASLs and HsASL is shown in
Fig. 5. Validation of the tetrameric structure using
PROCHECK showed that 93.2%, 6.1% and 0.6% of the
residues are in the most favored, additionally allowed
and generously allowed regions, respectively, of the
Ramachandran map. Ser277 of chains A and C are
present in the disallowed region of the Ramachandran
map. This serine is in the partially disordered C3 loop.
Superposition of Ca atoms of MtbASL and MsASL
ꢀ
tetramers yielded an rmsd of 0.42 A.
ens (HsASL; PDB code: 2VD6 and 2J91).
A
The conformation of the C3 loop in MtbASL was
identical to that observed in the template when
MsASL, EcASL or SaASL with ordered C3 loops was
used for modeling. On the other hand, modeling
MtbASL based on structures from which the C3 loop
was omitted showed variable conformations for the
loop (Fig. 4C), suggesting that the loop is flexible and
might assume different conformations depending on
the environment.
superposition of the ordered C3 loop of MsASL with
those of EcASL and SaASL (C3 loop conformation in
BaASL and TbASL is similar to that of EcASL) shows
that the loop has the most-open conformation in
MsASL (Fig. 4A). In the native and L-aspartate bound
forms of aspartase B, an enzyme of the aspartase/
fumarase superfamily, the C3 loop is in open and
closed conformations, respectively [24]. Superposition
of these structures on MsASL confirms that the C3
loop of MsASL has the most-open conformation
(Fig. 4B).
Structural analysis
A
sequence-based phylogenetic analysis of ASLs,
whose structures have been determined, unexpectedly
showed that the mycobacterial ASLs are more closely
related to the higher eukaryotic ASLs such as HsASL
and Caenorhabditis elegans ASL (CeASL) than to
lower eukaryotic ASLs such as TbASL and Plasmo-
dium vivax ASL (PvASL) or other bacterial ASLs
(Fig. 6). Although amino acid sequence identity
Homology modeling of MtbASL structure
In the absence of diffracting crystals, MtbASL tetra-
meric structure was modeled using MsASL structure
because the amino-acid sequence identity between the
two enzymes is 84.5%. The modeled structure of
MtbASL shows a similar conserved fold of the aspar-
1646
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
A
B
Fig. 4. C3 loop conformation in MsASL
and related proteins. (A) Comparison of
the conformations of the C3 loop in apo-
MsASL (pink), EcASL (green, PDB code:
2PTQ) and SaASL (blue, PDB code: 2X75).
The loop is most open in MsASL. (B)
Comparison of the C3 loop conformation
of MsASL (pink) with unliganded (cyan,
PDB code: 3R6Q) and L-aspartate bound
(orange, PDB code: 3R6V) aspartase B. (C)
Comparison of the conformation C3 loop
modeled in MtbASL. The conformation of
the loop obtained by using the templates
MsASL with C3 loop and MsASL, EcASL
and SaASL without C3 loop are colored in
pink, cyan, green and blue, respectively.
The location of the active site is
C
represented by bound ligands, AMP and
fumaric acid in (A) and (C) and L-aspartate
in (B).
between MsASL and human ASL is 33%, it is only
20% between the E. coli and mycobacterial enzymes.
Structure based comparisons also reveal a similar phy-
logeny (results not shown). The organization of the
three structural domains of mycobacterial enzymes is
similar to that of human ASL. The rmsd values
observed upon structural superposition of MsASL
with HsASL, CeASL, PvASL and EcASL are 1.7, 1.5,
MsASL and EcASL results in rmsd values of 2.4, 1.5
ꢀ
and 3.2 A, respectively. D2 appears to be the most
conserved domain, whereas D3 is the most variable
(Fig. 7A). D3 has fewer interactions with the major
domain D2 compared with domain D1.
To compare the relative disposition of D3 with
respect to D2 in different ASL structures, ASL pro-
tomers were transformed using the matrix obtained
by superposing domain 2 of all other ASLs on
MsASL D2. D3 was found to occupy two extreme
positions in EcASL and MsASL (Fig. 7A). The dispo-
sition of D3 found in EcASL was also observed in
most of the other enzymes, including TbASL, PvASL,
ꢀ
2.3 and 2.4 A, respectively. The rmsd values upon
independent structural superposition of the three
domains of MsASL and human ASL protomers are
ꢀ
1.1, 1.1 and 1.4 A for D1, D2 and D3, respectively.
Similar structural superposition of the domains of
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1647

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
Fig. 5. Structural alignment of MsASL,
MtbASL and HsASL.
Fig. 6. A sequence based phylogenetic
tree of all ASLs for which structures are
available in the PDB. The mycobacterial
ASLs are evolutionarily closer to the higher
eukaryotes (human and C. elegans)
compared to other bacterial ASLs.
1648
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
A comparison of the disposition of D3 with respect
to D2 across the members of aspartase/fumarase
superfamily shows that the positional heterogeneity of
D3 observed in MsASL, EcASL and TmASL is also
found in this superfamily (data not shown). All the
structures of class II fumarase and L-aspartase depos-
ited in the PDB have a D3 conformation similar to
that of MsASL. All argininosuccinate lyases and d-2
crystallin structures in the PDB have a D3 conforma-
tion resembling that of EcASL. The two intact struc-
tures of CMLE found in the PDB have a D3
conformation similar to that of TmASL. It has been
demonstrated that ASL interacts with glutamyl-tRNA
synthetase [25], leading to linkage between protein bio-
synthesis and nucleotide metabolism. It has been pro-
posed that D3 might be involved in this interaction
[19]. Therefore, the disposition of D3 with respect to
D2 as observed in ASL structures might influence this
interaction and be of physiological importance.
A
B
Kinetic parameters
The kinetic parameters of the two enzymes for the sec-
ond reaction of de novo purine biosynthetic pathway
in both directions were determined at a pH of 7.6 and
temperature of 37 ºC. The variation of their activities
as a function of pH was studied over the range 5.2–
9.6. The optimum pH values for the catalytic activities
of MsASL and MtbASL were 8.0 and 7.6, respectively
(data not shown). The Michaelis–Menten curves of the
two mycobacterial enzymes for SAMP, AMP and
fumarate are shown in Fig. 8. The kinetic parameters,
kcat, Km and kcat/Km of the two mycobacterial ASLs
Fig. 7. Analysis of domain
3 disposition. (A) Ribbon diagram
representing superposition of protomers in ASLs from
M. smegmatis (pink), human (PDB code: 2VD6, orange) and E. coli
(PDB code: 2PTS, green). Domain organization in MsASL
resembles that of eukaryotic human ASL more than that of
bacterial EcASL. The structure of domain 2 is mostly invariant in
the three ASLs. Largest differences are observed for domain 3.
The three domains are labeled as D1, D2 and D3. (B) The two
extreme conformations of D3 in MsASL and EcASL. The two
helices appearing immediately after D2 are labeled as 1 and the
next two helices as 2. TmASL (grey) represents the intermediate
conformation.
are shown in Table 3. The catalytic efficiencies (kcat
/
Km) of MsASL for SAMP, AMP and fumarate are
~ 40-, 42- and ten-fold greater than the corresponding
values for MtbASL.
A comparison of the specific activities of all the
biochemically characterized ASLs is provided in
Table 4. Apart from the mycobacterial enzymes, these
include Bacillus subtilis ASL (BsASL) [23], EcASL
[20], Plasmodium falciparum ASL (PfASL) [26],
HsASL [7] and Leishmania donovani ASL (LdASL)
[27]. Most activity measurements on ASLs have been
performed at 25 ºC. However, the catalytic activity of
the mycobacterial enzymes was very low at this tem-
perature. The activity increased linearly with tempera-
ture until 50–55 ºC and started reducing when the
temperature was raised further (data not shown).
Therefore, physiological conditions (temperature of
37 ºC and pH 7.6), where moderate activity of myco-
bacterial ASLs was observed, were selected for activ-
ity measurements. Mycobacterial ASLs, particularly
Legionella pneumophila ASL (LpASL) and Pyrobacu-
lum aerophilum ASL (PaASL). By contrast, in
HsASL, the position of D3 with respect to D2 was
similar to that in MsASL. Disposition of D3 in BaA-
SL, Thermotoga maritima ASL (TmASL) and SaASL
is intermediate to that in EcASL and MsASL
(Fig. 7B). The length and conformation of the loop
following the helical residues 389–410 of MsASL is
variable in ASL structures and might contribute to
the relative disposition of domain D3 with respect to
D2. The length of this loop is shortest in ASLs with
a D3 conformation similar to EcASL and longest in
MsASL and HsASL.
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1649

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
A
B
Fig. 8. Plots of specific activities in
(lmolÁmgÀ1ÁminÀ1) of (A) MsASL (in red)
and (B) MtbASL (in blue) as a function of
substrate concentration fitted to the
Michaelis–Menten equation.
Table 3. Kinetic parameters determined at pH 7.6 and 37ꢀC for the
forward and reverse reactions with SAMP catalyzed by MsASL and
MtbASL.
Table 4. Comparison of kinetic parameters for SAMP catalysis
across ASL homologs.
kcat/Km
À1
kcat (sÀ1
)
Km (lM)
(lM ÁsÀ1
)
MsASL
MtbASL
Assay conditions
1 SAMP
a kcat (sÀ1
b Km (lM)a
MsASL
MtbASL
EcASL
HsASL
PfASL
37 °C, pH 7.6
37 °C, pH 7.6
25 °C, pH 7.0
25 °C, pH 7.5
25 °C, pH 7.4
25 °C, pH 7.0
25 °C, pH 7.0
0.7
0.1
43.7
204.2
–
0.02
0.0005
–
)
0.7 Æ 0.01
0.1 Æ 0.02
43.7 Æ 2.6
204.2 Æ 48.2
4.1
c kcat/Km (lM ÁsÀ1
)
)
)
0.02
0.0005
À1
287.0
7.5
10.6
32.0
24.0
3.5
26.3
0.2
2 AMP
a kcat (sÀ1
b Km (lM)a
)
0.8 Æ 0.07
277.8 Æ 45.4
0.03
0.1 Æ 0.01
134.0 Æ 27.0
0.0007
LdASL
BsASL
28.0
1.3
1.2
0.4
c kcat/Km (lM ÁsÀ1
À1
3 Fumarate
a kcat (sÀ1
b Km (mM)5
)
0.7 Æ 0.03
3.9 Æ 0.5
0.2
0.1 Æ 0.01
5.7 Æ 1.2
0.02
c kcat/Km (lM ÁsÀ1
À1
kcat/Km of the enzymes shows that HsASL is the
most active enzyme. Activities of other ASLs are
comparable, although lower than HsASL. Both
MsASL and MtbASL showed very low activity, with
MsASL being one order of magnitude less, and
MtbASL being three orders of magnitude less, than
those of other bacterial ASLs. The human enzyme is
~ 40 000 times more efficient compared to mycobacte-
rial ASLs.
a
As a result of high Km values, MsASL did not reach saturation
when titrating with AMP, whereas MtbASL did not reach satura-
tion with SAMP and fumarate. Thus, the apparent Km values for
the substrates are reported.
MtbASL, were observed to have low specific activities
compared to other ASLs despite using near optimal
conditions. A comparison of the catalytic efficiency
1650
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
Catalytic residues of mycobacterial ASLs
A
ASL catalysis (Fig. 1) follows a general acid–base
reaction involving two separate groups to function as
the catalytic acid and base [26]. The catalytic base is
involved in proton abstraction (Fig. 1, step 1), fol-
lowed by protonation by the catalytic acid (Fig. 1, step
3). In earlier studies, residues equivalent to His78 and
His149 of MsASL were proposed to function as the
acid and base, respectively [28]. In later studies,
His149 or an activated water molecule and Ser279
were proposed to function as the acid and base,
respectively [20]. In a more recent study on SaASL, it
was proposed that residue equivalent to His149 and
Lys285 of MsASL might function as the catalytic acid
and base, respectively [18]. It is therefore of interest to
examine the position and conformation of these pro-
posed catalytic residues with respect to the putative
ligand binding site in MsASL structure and their
potential to function in catalysis.
B
Figure 9A shows the distance of His78 and His149
from the N1 and N6 atoms of AMP in apo-MsASL
and holo-HsASL structures. His78 does not show a
large variation in its conformation and Ca position
between the apo- and holo-structures. Its Ne2 atom is
Fig. 9. Positions of the proposed catalytic residues in apo-MsASL
(pink) with respect to the bound ligands (yellow), AMP and fumaric
acid (FMR). (A) Structural superposition of apo-MsASL and holo-
HsASL (cyan, PDB code: 2VD6) showing the position of Ne2 of
His78 and His149 with respect to the N1 and N6 atoms of the
adenine ring of AMP. The position of Ne2 of His149 with respect to
Cb of FMR is also shown. The equivalent histidine residues in holo-
HsASL are His86 and His159. (B) Structural superposition of apo-
MsASL and holo-EcASL (green, PDB code: 2PTQ) showing the
position of MsASL Ser279 and equivalent EcASL Ser295 with
respect to Cb of FMR. Closure of the C3 loop might bring Ser279
close to the substrate. The color of the labels of the distances
shown indicates the color of the corresponding structure.
ꢀ
at a distance > 5.0 A from N1 of the adenine ring of
SAMP in all the structures, meaning that His78 is unli-
kely to function as a catalytic acid. The conformation
of His149 in apo-MsASL resembles that observed in
the holo-structures. The Ne2 atom of MsASL His149
is close to the N6 atom of the adenine ring of SAMP
ꢀ
with distances of 4.1 and 3.7 A in MsASL and
HsASL, respectively. By contrast, it is more distant
from the Cb atom of the fumarate moiety with dis-
ꢀ
tances of 4.7 and 4.5 A in MsASL and HsASL,
respectively, suggesting that His149 is unlikely to func-
tion as a base extracting proton from fumarate but
might function as the catalytic acid protonating N6 of
SAMP.
group of SAMP in all ASLs except SaASL. The con-
formation of Lys285 in MsASL is similar to that
observed in the holo-structure of EcASL. The equiva-
lent lysine in EcASL shows strong hydrogen-bonding
interactions with the carboxylate oxygens of SAMP.
Mutation of this lysine has been shown to increase Km
of the enzyme for SAMP by several folds in BsASL
[23], suggesting that Lys285 has a role in substrate
binding. Lys285 might contribute to substrate binding
by its interaction with the carboxylate group. It might
also act as a base as the mutant is devoid of catalytic
activity.
The identity of the catalytic base of ASLs has been
extensively debated [18,20]. Figure 10A shows the
position of the residues Lys285 and Ser279 with
respect to fumarate. Both of these residues, present on
the highly flexible C3 loop, have been proposed to
function as base in one or other of ASL structures
[18,20]. Based on the available ASL crystal structures
conformation of the C3 loop in MsASL, EcASL and
SaASL structures could be considered as most-open,
most-closed and intermediate, respectively. The confor-
mation of Lys285 is highly variable in the apo- and
holo-structures of ASLs. It is observed that substrate-
binding brings this lysine within interacting distance of
His149 and the carboxylate oxygens of the succinyl
In EcASL, the distance between the Oc atom of the
serine residue (Ser295) equivalent to MsASL Ser279
ꢀ
and Cb atom of the fumarate moiety is 3.1 A
(Fig. 9B), suggesting that it could function as the cata-
lytic base. In MsASL, the Ca of Ser279 is displaced
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1651

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
A
Fig. 10. Conformation of Ser279 in apo-
MsASL. (A) Stereo view of the
orientations of the proposed catalytic
bases, Ser279 and Lys285 in apo-MsASL
(pink) after superposition on holo-EcASL
(green) and holo-SaASL (blue). (B) Stereo
view of the orientations of Ser279 and
Lys285 in apo-MsASL (pink), apo-AspB
(cyan, PDB code: 36RQ) and holo-AspB
(orange, PDB code: 36RV) after structural
superposition. The equivalent Ser318 in
apo-AspB is initially far away from the
active site and is brought closer to the
substrate in the holo-structure. The bound
ligands at the active site are shown in line
representation.
B
ꢀ
away from the Ca of Ser295 of EcASL by 8.8 A. How-
ever, its side chain has a similar orientation as in
EcASL and points towards the active site with a dis-
residues are mostly conserved. In Mycobacteria, two
residues near the active site, proposed to be important
for substrate binding, were found to be different
[20,30]. The residue corresponding to MsASL-Ser148
and MtbASL-Ser146 is either retained as serine or
replaced by threonine in most other ASLs. This varia-
tion may not be the cause of drastic reduction in the
activity of mycobacterial ASLs because the substitution
does not lead to changes in the hydrogen bonding inter-
actions near the active site. The second residue is a con-
served arginine in all ASLs except in actinobacteria (a
phylum that includes the genus Mycobacteria), where it
is replaced by a glycine (Gly319 in MsASL). The dras-
tic mutation from arginine to glycine will lead to a
large cavity. In MsASL, this cavity is partially filled by
the guanidino group of Arg103 of MsASL (Arg101 of
MtbASL), although the orientation of the guanidino
group is opposite to that of other ASLs (Fig. 11).
Therefore, Arg319Gly substitution might be partially
responsible for the reduced efficiency of mycobacterial
enzymes. It would be of interest to examine the
catalytic properties of Gly319Arg mutant of MsASL.
The low value of kcat of mycobacterial ASLs sug-
gests that the active site might not be easily accessible
and the rate-limiting step of mycobacterial ASL cataly-
sis could be slow substrate binding or product release.
This might be a result of sluggish loop or subdomain
movements of the enzyme and slow re-organization of
the residues required for catalysis. The sequence at the
ꢀ
tance of 9.9 A between its Oc atom and Cb of the
fumarate moiety. It is possible that substrate binding
induces closure of the C3 loop reducing the distance
between Ser279 and Cb of the succinyl group, as seen
in EcASL, enabling proton abstraction by Ser279.
Similar closure of the C3 loop bringing the equivalent
serine residue closer to the substrate has also been
observed in L-aspartase apo- and holo-structures
(Fig. 10B). Hence, it is possible that Ser279 functions
as the catalytic base. In the light of these results, it is
likely that His149 functions as the catalytic acid and
Lys285 or Ser279 functions as the catalytic base in
mycobacterial ASLs, as well as in the aspartase/fuma-
rase superfamily. Further mutational experiments fol-
lowed by kinetic studies might establish the identity of
the catalytic base.
Low catalytic activity of mycobacterial ASLs
Variation in the catalytic activity of ASLs from differ-
ent species could be a result of sequence alterations or
change in the geometry of the active site or change in
the accessibility of the active site to substrates and
products. Alignment of sequences of ASLs from
various organisms (~ 150 ASL sequences representing
different phyla) shows that the proposed active site
1652
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
once in several hours. Between the two mycobacterial
species of the present study, Mycobacteria smegmatis
(MS) is the faster growing (6-h cycle) compared to
Mycobacteria tuberculosis (MTB), which divides once
in 24 h (Goodfellow and Magee, 1998). A comparison
of the kcat for the cleavage of SAMP in all the kineti-
cally characterized bacterial ASLs shows an inverse
relationship with the corresponding doubling time
(Table 5). Similar differences in the efficiency of the
enzymes from Mycobacteria and other organisms as a
result of differences in growth rates have been reported
previously [33,34]. Furthermore, the mycobacterial
genomes have a high GC content of ~ 65% and hence
their requirement of adenine nucleotides required for
DNA or RNA synthesis is correspondingly low.
Therefore, the low activity of ASL might suffice for
the slow growth rate and low AT content of these
bacteria [33,34].
Fig. 11. Arg319Gly substitution in MsASL (pink). This residue is a
strictly conserved Arg except in Actinobacteria. The guanidine
group of Arg103 of MsASL, although oriented differently, partially
fills this cavity. The residues equivalent to 319 and 103 of MsASL
are also shown for EcASL (green) and HsASL (yellow). The active
site is shown by bound ligands, AMP and fumaric acid (FMR),
colored in cyan.
The second possible reason for the lower activity of
mycobacterial ASLs could be the preference in these
bacteria for the salvage pathway over the de novo bio-
synthetic pathway for deriving purine nucleotides.
MTB genome codes for several purine salvage enzymes
[35]. Mycobacteria microti and Mycobacteria avium
have very low de novo purine biosynthetic activity and
the pathway appears to be absent in Mycobacteria
leprae [36]. Adenosine kinase, a salvage pathway
enzyme supplying adenine nucleotides, is present in all
eukaryotes as well as in Mycobacteria, whereas it is
absent in most of the intracellular bacteria [37,38].
Furthermore, adenosine kinase has been shown to play
a key role in supplementing adenine nucleotides in
M. leprae [36]. Therefore, low catalytic activity of
mycobacterial ASLs might be compensated by adeno-
sine kinase. However, the function of ASL cannot be
entirely substituted by adenosine kinase because it has
been demonstrated that, although adenosine kinase
deficient strains of MTB can grow in vitro [37], ASL
deficient strains fail to survive. This suggests that the
purB gene is essential for mycobacterial growth [39–41]
and could be a target for developing specific inhibitors.
Mycobacterial enzymes structurally resemble the
human enzyme more than bacterial enzymes although
they are almost 40 000 times slower than the human
N-terminal segment of the flexible C3 loop is (EL-
VTEGFAP) in MsASL, (FKQKTIAGE) in EcASL
and (KEIEEPFEK) in HsASL. The loop appears to
undergo open–close conformational transition in the
catalytic cycle. The human enzyme has six charged res-
idues in this segment compared to three in EcASL and
two in MsASL. The flexibility of the C3 loop may
depend on these charged residues. Thus, the human
enzyme with six charged residues may be most flexible.
The loop is found to be disordered in the ligand
bound structures (PDB code: 2VD6, 2J91) of HsASL.
With an intermediate charge, the loop is disordered in
the ligand free E. coli enzyme but gets ordered upon
ligand binding [20]. By contrast, with minimal charges,
the loop in MsASL is partially ordered even in the
unliganded enzyme. The dynamic nature of this loop
may contribute to the rate of catalysis leading to an
efficient human enzyme and sluggish mycobacterial
enzymes.
Many enzymes of purine salvage and biosynthetic
pathways have been shown to be up- or down-regu-
lated allosterically by the pathway intermediates or
purine nucleotides [31,32]. Whether mycobacterial
ASLs are also subject to allosteric regulation remains
to be investigated.
Table 5. Relationship between the doubling time of the bacteria
and kcat for SAMP catalysis.
Doubling time
kcatÁ(sÀ1
)
Importance of ASL in Mycobacteria
MtbASL
MsASL
BsASL
EcASL
24 h
0.1
0.7
1.3
4.1
6 h
The replication cycle of mycobacterial species is much
longer than those of other bacteria. Thus, although
E. coli divides once every 20 min, Mycobacteria divide
120 min
20 min
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1653

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
cific primers: MsASL-sense primer 5′-GCTAGCCATATG
TTC GTG ACG ATT CCG AAT-3′ and MsASL-anti-
sense primer 5′-GGATCCTTAAAGCTT CAA GAT CGC
GCC CGA-3′ (the underlined nucleotides represent restric-
tion sites). The purB gene of Mycobacterium tuberculosis
strain H37rv was amplified by PCR using gene-specific
primers: MtbASL-sense primer 5′-GCTAGCCATATG
AGC ATT CCC AAC GTG CTG-3′ and MtbASL-anti-
sense primer 5′-GGATCCTTAAAGCTT AAG AAT TGC
ACC CGG CGT-3′. The sense primers had NheI and
NdeI restriction enzyme sites, whereas the antisense prim-
ers had BamHI and HindIII restriction enzyme sites. Both
MsASL and MtbASL genes were cloned in pRSET C vec-
tor between NheI and BamHI sites. The strategy used for
cloning resulted in the addition of fourteen amino acids
from the vector sequence at the N-terminus of the protein
including a hexa-histidine tag useful for Ni2+-NTA (nitril-
otriacetic acid) affinity purification. MsASL and MtbASL
proteins were over-expressed in E. coli BL21 (DE3) pLysS
cells by the addition of 0.3 mM isopropyl thio-b-D-galacto-
side. The induced cultures of MsASL and MtbASL were
grown for further 6 h at 30 ºC and 12 h at 20 ºC, respec-
tively. The proteins were purified in a two-step process
starting with Ni2+-NTA affinity chromatography, fol-
lowed by gel filtration using a Superdex S-200 preparative
column (GE Healthcare, Little Chalfont, UK). The puri-
fied proteins were in the buffer containing 50 mM Tris
(pH 7.5) and 200 mM NaCl. The proteins were concen-
trated using an Amicon centrifugal filter device (molecular
weight cut-off of 30 000; Millipore, Billerica, MA, USA).
The molecular masses were confirmed by MALDI-TOF
and corresponded to 53 and 52 kDa for MsASL and
MtbASL, respectively. Protein concentrations were esti-
mated using a Jasco V-630 spectrophotometer (Jasco Inc.,
Easton, MD, USA) and an extinction coefficient of
31 400 MÀ1ÁcmÀ1 as calculated by EXPASY PROTPARAM [42].
The concentration obtained was verified by the Bradford
assay [43].
enzyme. It would be of great interest to design an
inhibitor for mycobacterial enzymes that does not
affect the human enzyme.
Conclusions
ASLs from MS and MTB were cloned, over-expressed,
purified and crystallized. Crystals of MsASL diffracted
X-rays to high resolution and its structure could be
ꢀ
determined at a resolution of 2.16 A. This is the first X-
ray crystal of an unliganded ASL to be reported with a
partially ordered C3 loop in its most-open conforma-
tion. Crystals of MtbASL diffracting X-rays to useful
resolution could not be obtained and hence its tetra-
meric structure was modeled using MsASL co-ordinates
as a template. The X-ray crystal structure of MsASL
and the modeled structure of MtbASL adhere to the
fold that is common to all members of the aspartase/
fumarase superfamily. The structures of mycobacterial
ASLs were found to be closer to HsASL compared to
other bacterial ASLs. A comparison of MsASL struc-
ture with other ASL structures suggests that His149
and Lys285 or Ser279 are most likely to function as the
catalytic acid and catalytic base, respectively.
MsASL was found to be catalytically more active
than MtbASL. However, both the enzymes were much
slower compared to ASLs of other organisms. Most of
the active site residues of mycobacterial enzymes were
found to be conserved, with only two exceptions:
Ser148 and Gly319 of MsASL, which correspond to
threonine and arginine residues, respectively, in most
of the other organisms. We speculate that the C3 loop
at the active site is most dynamic in the human
enzyme and least dynamic in mycobacterial enzymes.
The substitution of Gly in Mycobacteria for Arg of
human ASL and the nature of dynamics of the C3
loop might account for the low activity of mycobacte-
rial ASLs. It is likely that the low catalytic activity of
MsASL and MtbASL suffices for their slow growth
rates, dependence on other salvage pathways for ade-
nine nucleotides and their GC-rich genomes. However,
basal expression of ASL is important for the sustained
multiplication and infection of Mycobacteria. The
structures of these mycobacterial ASLs might therefore
be useful for the development of specific inhibitors as
drug candidates for mycobacterial infections.
Structural studies
Crystallization
Crystals for MsASL were obtained using a microbatch
method [44]. Crystallization was set up with 2 lL of
10 mgÁmLÀ1 protein and 2 lL of the crystallization condi-
tion containing 0.1 M sodium Hepes (pH 7.5), 1.4 M sodium
citrate and 0.1 M betaine hydrochloride. The crystals dif-
ꢀ
fracted X-rays to a resolution of 2.5 A at home-source [Mi-
crostar Ultra II Rotating Anode X-ray Generator (Bruker
Instruments, Inc., Bellerica, MA, USA) equipped with a
MAR imaging plate detector (Marresearch GmbH, Norder-
Materials and methods
ꢀ
stedt, Germany)] and a resolution of 2.16 A on the BM14
Cloning and protein purification
beam line at the ESRF synchrotron facility (Grenoble,
France).
The purB gene encoding ASL of Mycobacterium smegma-
tis strain MC2 155 was amplified by PCR using gene-spe-
1654
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
restraints were applied during the course of refinement. The
progress of refinement was monitored using 5% (2615) of
the 52 292 independent reflections, which were used for the
calculation of free R-factor. The structure of MsASL was
validated using PROCHECK [57] of the CCP4 software suite.
The coordinates and the structure factors of MsASL have
been deposited in the Protein Data Bank under code
4NLE.
Crystals for MtbASL were obtained using microbatch
and hanging drop methods [45]. Crystallization was set up
using 2 lL of 5 mgÁmLÀ1 protein and 2 lL of different
crystallization conditions. Crystallization was set up for
native protein, as well as for the protein incubated with
various concentrations of SAMP, AMP and fumarate.
Attempts were made to enhance the crystal quality by
adjusting crystallization conditions and including additives
to the crystallization droplet. The crystals were also soaked
for varying time periods in different cryoprotectants such
as polyethylene glycol and glycerol. However, crystals of
quality useful for structural studies could not be obtained.
Homology modeling of MtbASL
The tetrameric structure of MtbASL was modeled using
MODELLER, version 9v7 [58,59], based on the co-ordinates of
MsASL as the template. The output with the lowest molec-
ular probability density function was selected and its ste-
reochemical quality was verified using PROCHECK.
X-ray diffraction data collection and processing of
MsASL
Crystals of MsASL were transferred to the crystallization
condition containing 5% glycerol as the cryoprotectant for
a few seconds and then mounted in cryo-loops. They were
tested by exposing them to Cu-Ka radiation at 100 K
maintained using a stream of nitrogen gas. After confirm-
ing the quality, the crystals were frozen in liquid nitrogen
and sent to BM-14, ESRF (Grenoble, France) for data col-
lection. The frames were processed using MOSFLM [46] in the
Structural analysis
Sequence alignment was carried out using either CLUS-
TAL W [60] or MULTALIN [61]. Structure based sequence
alignment was carried out using the MUSTANG-MR struc-
tural sieving server [62]. Superposition of all structures
was achieved using SSM SUPERPOSE [63] of COOT and ALIGN
[64]. The interactions in the oligomer were identified using
PDBEPISA [65]. The secondary structure was evaluated using
PDBSUM [66]. Phylogenetic analyses were carried out using
MEGA, version 4.0 [67]. All figures were prepared using PY-
MOL [68].
ꢀ
orthorhombic space group P222 with a = 74.6 A,
ꢀ
ꢀ
b = 176.83 A and c = 73.3 A. The resulting intensities were
scaled using SCALA of the CCP4 software suite [47,48].
Structure determination of MsASL
Systematic absences in the diffraction data indicated the
presence of two 21 screw axes along the a- and b-directions.
Hence, the space group of the crystal was set to P21212. A
BLAST search [49] revealed that MsASL shares 33%
Biochemical studies
Steady-state kinetics of both MsASL and MtbASL for the
forward (cleavage of SAMP) and reverse (formation of
SAMP) reactions of the second reaction catalyzed by ASL
in de novo purine biosynthesis were followed by monitoring
the decrease or increase, respectively, in the absorbance of
SAMP at 290 nm as a function of time using a double
beam JASCO V-630 spectrophotometer at pH 7.6 and tem-
perature 37 ºC. The Km and Vmax for the forward reaction
were determined by titrating SAMP over a range of con-
centrations. The Km and Vmax for the reverse reaction with
the two substrates (AMP and fumarate) were determined
independently by keeping one substrate at saturating con-
centration and titrating the other substrate over a range of
concentrations. A mixed buffer (pH range 4.0–10.0) con-
taining succinic acid, sodium dihydrogen phosphate and
glycine in the ratio of 2 : 7 : 7 was used for the reactions
in order to minimize the effect of buffer components on
enzyme activities.
ꢀ
sequence identity with HsASL. The 2.0-A resolution struc-
ture of HsASL (PDB code: 2VD6) was retrieved from the
2
PDB. It has an average B-factor of 28.8 A . The structure
ꢀ
of MsASL was determined by molecular replacement using
the co-ordinates of the polyalanine chains of the AB dimer
of 2VD6 as the phasing model in PHASER [50] of the CCP4
software suite. The quality of the model obtained from
phaser was improved using the automatic model building
software BUCCANEER [51] of the CCP4 software suite and
AUTOBUILD [52] of PHENIX, version 1.7-650 [53] followed by
manual fitting using COOT [54]. The individual atomic posi-
tions and the isotropic temperature factors were refined
using REFMAC 5 [55] of the CCP4 software suite. During the
final stages of refinement, TLS refinement [56] was used,
where each of the protomer in the asymmetric unit con-
sisted of one TLS group. This resulted in a decrease of
Rwork and Rfree by 0.4% and 0.5%, respectively. The aver-
For determining the kinetic parameters for SAMP in the
forward reaction, 35 lg of MsASL and 125 lg of MtbASL
were used in a final assay volume of 500 lL and titrating
SAMP concentrations over the range of 5–300 lM.
2
ꢀ
age B-factor of the final structure is 32.3 A . The B-factors
specified refer to the temperature factors after adding the
TLS contribution. No noncrystallographic symmetry
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1655

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
5 Ratner S (1972) Argininosuccinases and
For determining the kinetic parameters for AMP in the
reverse reaction, 35 lg of MsASL and 125 lg of MtbASL
were used in a final assay volume of 500 lL with an almost
saturating fumarate concentration of 8 mM and titrating
AMP concentrations of 5–350 lM.
Adenylosuccinases. In The Enzymes (Boyer PD, ed),
pp. 167–197. Academic Press, New York.
6 Bridger W & Cohen LH (1968) The kinetics of
adenylosuccinate lyase. J Biol Chem 243, 644–650.
7 Stone R, Zalkin H & Dixon J (1993) Expression,
purification, and kinetic characterization of
recombinant human adenylosuccinate lyase. J Biol
Chem 268, 19710–19716.
For determining the kinetic parameters for fumarate in
the reverse reaction, 35 lg of MsASL and 125 lg of
MtbASL were used in a final assay volume of 500 lL with
almost saturating AMP concentration of 350 lM and titrat-
ing fumarate concentrations of 500 lM to 10 mM.
8 Hanson KR & Havir EA (1972) The enzymic
elimination of ammonia. In The Enzymes (Boyer PD,
ed), pp. 75–166. Academic Press, New York.
9 Casey PJ & Lowenstein JM (1987) Purification of
adenylosuccinate lyase from rat skeletal muscle by a
novel affinity column. Stabilization of the enzyme, and
effects of anions and fluoro analogues of the substrate.
Biochem J 246, 263–269.
The reactions were initiated by the addition of sub-
strate to the sample cuvette after setting the absorbance
reading with respect to the reference cuvette to zero. The
amount of substrate consumed was estimatÀed1 using the
difference extinction coefficient of 4.05 mM ÁcmÀ1 [69].
Specific activity was calculated as lmoles of substrate
converted to product in 1 min by a milligram of enzyme.
Kinetic parameters were obtained by fitting the data to
the Michaelis–Menten equation using GRAPHPAD PRISM,
version 5.02 (GraphPad Software, Inc., San Diego, CA,
USA).
10 Porter DJ, Rudie NG & Bright HJ (1983) Nitro
analogs of substrates for adenylosuccinate synthetase
and adenylosuccinate lyase. Arch Biochem Biophys 225,
157–163.
11 Van den Berghe G, Bontemps F, Vincent M & Van den
Bergh F (1992) The purine nucleotide cycle and its
molecular defects. Prog Neurobiol 39, 547–561.
12 Toth EA & Yeates TO (2000) The structure of
adenylosuccinate lyase, an enzyme with dual activity in
the de novo purine biosynthetic pathway. Structure 8,
163–174.
Acknowledgements
We thank the staff at MBU X-ray laboratory and
the beam line scientists at BM14, ESRF (Grenoble,
France) for help during the X-ray diffraction data
collection. Sourav Roy (JNCASR) kindly assisted in
the analysis of the kinetic experiments. MRNM, HSS
and HB thank the Department of Biotechnology and
the Department of Science and Technology, India,
for providing financial support. SB thanks the Coun-
cil for Scientific and Industrial Research (CSIR),
India for Junior & Senior Research Fellowships. We
thank Fathima Benazir for critically reading the
manuscript.
13 Jaeken J, Wadman S, Duran M, Van Sprang F, Beemer
F, Holl R, Theunissen P, De Cock P, Van den Bergh F
& Vincent M (1988) Adenylosuccinase deficiency: an
inborn error of purine nucleotide synthesis. Eur J
Pediatr 148, 126–131.
14 Maaswinkel-Mooij P, Laan L, Onkenhout W, Brouwer
O, Jaeken J & Poorthuis B (1997) Adenylosuccinase
deficiency presenting with epilepsy in early infancy.
J Inherit Metab Dis 20, 606–607.
ꢁ
15 Verginelli D, Luckow B, Crifo C, Salerno C & Gross
M (1998) Identification of new mutations in the
adenylosuccinate lyase gene associated with impaired
enzyme activity in lymphocytes and red blood cells.
Biochim Biophys Acta 1406, 81–84.
References
1 Berens RL, Krug EC & Marr JJ (1995) Purine and
pyrimidine metabolism. In Biochemistry and Molecular
Biology of Parasites (Marr JJ & Muller M, eds), pp.
89–117. Academic Press Ltd, London.
16 Jackson RC & Harkrader RJ (1981) The contributions
of de novo and salvage pathways of nucleotide
biosynthesis in normal and malignant cells. In
Nucleosides and Cancer Treatment (Tattersall MHN &
Fox RM, eds), pp. 18–31. Academic Press, Sydney.
17 McLeod HL, Cassidy J, Powrie RH, Priest DG, Zorbas
MA, Synold TW, Shibata S, Spicer D, Bissett D &
Pithavala YK (2000) Pharmacokinetic and
2 Ducati RG, Breda A, Basso LA & Santos DS (2011)
Purine salvage pathway in Mycobacterium tuberculosis.
Curr Med Chem 18, 1258–1275.
3 Malathi V & Ramakrishnan T (1966) Biosynthesis of
nucleic acid purines in Mycobacterium tuberculosis
H37Rv. Biochem J 98, 594.
4 Le Nours J, Bulloch EM, Zhang Z, Greenwood DR,
Middleditch MJ, Dickson JM & Baker EN (2011)
Structural analyses of a purine biosynthetic enzyme
from Mycobacterium tuberculosis reveal a novel bound
nucleotide. J Biol Chem 286, 40706–40716.
pharmacodynamic evaluation of the glycinamide
ribonucleotide formyltransferase inhibitor AG2034. Clin
Cancer Res 6, 2677–2684.
18 Fyfe PK, Dawson A, Hutchison M-T, Cameron S &
Hunter WN (2010) Structure of Staphylococcus aureus
1656
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS

S. Banerjee et al.
Mycobacterial adenylosuccinate lyase
adenylosuccinate lyase (PurB) and assessment of its
potential as a target for structure-based inhibitor
discovery. Acta Crystallogr Sect D: Biol Crystallogr 66,
881–888.
transition in Escherichia coli aspartate
transcarbamoylase. Nat Struct Mol Biol 8, 423–426.
31 Smith JL, Zaluzec EJ, Wery J-P, Niu L, Switzer RL,
Zalkin H & Satow Y (1994) Structure of the allosteric
regulatory enzyme of purine biosynthesis. Science 264,
1427–1433.
19 Toth EA, Worby C, Dixon JE, Goedken ER, Marqusee
S & Yeates TO (2000) The crystal structure of
adenylosuccinate lyase from Pyrobaculum aerophilum
reveals an intracellular protein with three disulfide
bonds. J Mol Biol 301, 433–450.
32 Goodfellow M & Magee JG (1998) Taxonomy of
mycobacteria. In Mycobacteria (Gangadharam PRJ &
Jenkins PA, eds), pp. 1–75. Chapman & Hall, London.
33 Sang PB & Varshney U (2013) Biochemical properties
of MutT2 proteins from Mycobacterium tuberculosis
and M. smegmatis and their contrasting antimutator
roles in Escherichia coli. J Bacteriol 195, 1552–1560.
34 Wheeler PR & Ratledge C (1994) Metabolism of
Mycobacterium tuberculosis. In Tuberculosis
pathogenesis, protection, and control (Bloom BR, ed),
pp. 353–385. ASM Press, Washington DC.
20 Tsai M, Koo J, Yip P, Colman RF, Segall ML &
Howell PL (2007) Substrate and product complexes of
Escherichia coli adenylosuccinate lyase provide new
insights into the enzymatic mechanism. J Mol Biol 370,
541–554.
21 Ramachandran G, Ramakrishnan CT & Sasisekharan
V (1963) Stereochemistry of polypeptide chain
configurations. J Mol Biol 7, 95–99.
22 Puthan Veetil V, Fibriansah G, Raj H, Thunnissen A-
MW & Poelarends GJ (2012) Aspartase/fumarase
superfamily: a common catalytic strategy involving
general base-catalyzed formation of a highly stabilized
aci-carboxylate intermediate. Biochemistry 51, 4237–
4243.
35 Cole S, Brosch R, Parkhill J, Garnier T, Churcher C,
Harris D, Gordon S, Eiglmeier K, Gas S & Barry CR
(1998) Deciphering the biology of Mycobacterium
tuberculosis from the complete genome sequence.
Nature 393, 537–544.
36 Wheeler PR (1987) Enzymes for purine synthesis and
scavenging in pathogenic mycobacteria and their
distribution in Mycobacterium leprae. J Gen Microbiol
133, 3013–3018.
23 Brosius JL & Colman RF (2002) Three subunits
contribute amino acids to the active site of tetrameric
adenylosuccinate lyase: Lys268 and Glu275 are
required. Biochemistry 41, 2217–2226.
37 Long MC, Escuyer V & Parker WB (2003)
Identification and characterization of a unique
adenosine kinase from Mycobacterium tuberculosis. J
Bacteriol 185, 6548–6555.
24 Fibriansah G, Veetil VP, Poelarends GJ & Thunnissen
A-MW (2011) Structural basis for the catalytic
mechanism of aspartate ammonia lyase. Biochemistry
50, 6053–6062.
38 McNally T, Helfrich RJ, Cowart M, Dorwin SA,
Meuth JL, Idler KB, Klute KA, Simmer RL,
Kowaluk EA & Halbert DN (1997) Cloning and
expression of the adenosine kinase gene from rat and
human tissues. Biochem Biophys Res Commun 231, 645–
650.
25 Gendron N, Breton R, Champagne N & Lapointe J
(1992) Adenylosuccinate lyase of Bacillus subtilis
regulates the activity of the glutamyl-tRNA synthetase.
Proc Natl Acad Sci 89, 5389–5392.
26 Bulusu V, Srinivasan B, Bopanna MP & Balaram H
(2009) Elucidation of the substrate specificity, kinetic
and catalytic mechanism of adenylosuccinate lyase from
Plasmodium falciparum. Biochim Biophys Acta 1794,
642–654.
39 Xu G, Liu B, Wang F, Wei C, Zhang Y, Sheng J,
Wang G & Li F (2013) High-throughput screen of
essential gene modules in Mycobacterium tuberculosis: a
bibliometric approach. BMC Infect Dis 13, 227.
40 Sassetti CM, Boyd DH & Rubin EJ (2003) Genes
required for mycobacterial growth defined by high
density mutagenesis. Mol Microbiol 48, 77–84.
41 Sassetti CM & Rubin EJ (2003) Genetic requirements
for mycobacterial survival during infection. Proc Natl
Acad Sci 100, 12989–12994.
27 Boitz JM, Strasser R, Yates PA, Jardim A & Ullman B
(2013) Adenylosuccinate synthetase and
adenylosuccinate lyase deficiencies trigger growth and
infectivity deficits in Leishmania donovani. J Biol Chem
288, 8977–8990.
28 Lee TT, Worby C, Bao Z-Q, Dixon JE & Colman RF
(1999) His68 and His141 are critical contributors to the
intersubunit catalytic site of adenylosuccinate lyase of
Bacillus subtilis. Biochemistry 38, 22–32.
42 Gasteiger EHC, Gattiker A, Duvaud S, Wilkins MR,
Appel RD & Bairoch A (2005) Protein Identification
and Analysis Tools on the ExPASy Server. In The
Proteomics Protocol Handbook (Walker JM, ed), pp.
571–607. Humana Press, New Jersey.
29 Segall ML & Colman RF (2004) Gln212, Asn270, and
Arg301 are critical for catalysis by adenylosuccinate
lyase from Bacillus subtilis. Biochemistry 43, 7391–
7402.
43 Bradford MM (1976) A rapid and sensitive method for
the quantitation of microgram quantities of protein
utilizing the principle of protein-dye binding. Anal
Biochem 72, 248–254.
30 Macol CP, Tsuruta H, Stec B & Kantrowitz ER (2001)
Direct structural evidence for a concerted allosteric
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS
1657

Mycobacterial adenylosuccinate lyase
S. Banerjee et al.
44 Chayen NE, Shaw Stewart P, Maeder D & Blow D
(1990) An automated system for micro-batch protein
crystallization and screening. J Appl Crystallogr 23,
297–302.
56 Winn M, Isupov M & Murshudov G (2001) Use of
TLS parameters to model anisotropic displacements in
macromolecular refinement. Acta Crystallogr Sect D:
Biol Crystallogr 57, 122–133.
45 McPherson A (1985) Crystallization of macromolecules:
general principles. Methods in Enzymology. 114, 112–
120.
57 Laskowski RA, MacArthur MW, Moss DS &
Thornton JM (1993) PROCHECK: a program to check
the stereochemical quality of protein structures. J Appl
Crystallogr 26, 283–291.
46 Leslie AG & Powell HR (2007) Processing diffraction
data with Mosflm. In Evolving methods for
macromolecular crystallography, pp. 41–51. Springer,
Netherlands.
ꢂ
58 Sali A & Blundell TL (1993) Comparative protein
modelling by satisfaction of spatial restraints. J Mol
Biol 234, 779–815.
ꢂ
59 Fiser A & Sali A (2003) Modeller: generation and
47 CCP4 (1994) The CCP4 suite: programs for protein
crystallography. Acta Crystallogr D Biol Crystallogr 50,
760–763.
refinement of homology-based protein structure models.
Methods Enzymol 374, 461–491.
48 Winn MD, Ballard CC, Cowtan KD, Dodson EJ,
Emsley P, Evans PR, Keegan RM, Krissinel EB, Leslie
AG & McCoy A (2011) Overview of the CCP4 suite
and current developments. Acta Crystallogr Sect D:
Biol Crystallogr 67, 235–242.
60 Higgins DG & Sharp PM (1988) CLUSTAL: a package
for performing multiple sequence alignment on a
microcomputer. Gene 73, 237–244.
61 Corpet F (1988) Multiple sequence alignment with
hierarchical clustering. Nucl Acids Res 16, 10881–10890.
62 Konagurthu ASRC, Schmidberger JW, Irving JA, Lesk
AM, Stuckey PJ, Whisstock JC & Buckle AM (April
2010) MUSTANG-MR structural sieving server:
applications in protein structural analysis and
crystallography. PLoS ONE 5, 1–7.
49 Altschul SF, Gish W, Miller W, Myers EW & Lipman
DJ (1990) Basic local alignment search tool. J Mol Biol
215, 403–410.
50 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn
MD, Storoni LC & Read RJ (2007) Phaser
crystallographic software. J Appl Crystallogr 40, 658–
674.
63 Krissinel E & Henrick K (2004) Secondary-structure
matching (SSM), a new tool for fast protein structure
alignment in three dimensions. Acta Crystallogr Sect D:
Biol Crystallogr 60, 2256–2268.
51 Cowtan K (2006) The Buccaneer software for
automated model building. Acta Crystallogr Sect D:
Biol Crystallogr 62, 1002–1011.
64 Cohen G (1997) ALIGN: a program to superimpose
protein coordinates, accounting for insertions and
deletions. J Appl Crystallogr 30, 1160–1161.
65 Krissinel E & Henrick K (2007) Inference of
macromolecular assemblies from crystalline state. J Mol
Biol 372, 774–797.
52 Terwilliger TC, Grosse-Kunstleve RW, Afonine PV,
Moriarty NW, Zwart PH, Hung L-W, Read RJ &
Adams PD (2007) Iterative model building, structure
refinement and density modification with the PHENIX
AutoBuild wizard. Acta Crystallogr Sect D: Biol
Crystallogr 64, 61–69.
66 Laskowski RA, Hutchinson EG, Michie AD, Wallace
AC, Jones ML & Thornton JM (1997) PDBsum: a
Web-based database of summaries and analyses of all
PDB structures. Trends Biochem Sci 22, 488–490.
67 Tamura K, Dudley J, Nei M & Kumar S (2007) MEGA4:
molecular evolutionary genetics analysis (MEGA)
software version 4.0. Mol Biol Evol 24, 1596–1599.
68 DeLano WL (2002) The PyMOL molecular graphics
system.
53 Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis
IW, Echols N, Headd JJ, Hung L-W, Kapral GJ &
Grosse-Kunstleve RW (2010) PHENIX: a
comprehensive Python-based system for
macromolecular structure solution. Acta Crystallogr
Sect D: Biol Crystallogr 66, 213–221.
54 Emsley P, Lohkamp B, Scott W & Cowtan K (2010)
Features and development of Coot. Acta Crystallogr
Sect D: Biol Crystallogr 66, 486–501.
69 Tornheim K & Lowenstein JM (1972) The purine
nucleotide cycle The production of ammonia from
aspartate by extracts of rat skeletal muscle. J Biol Chem
247, 162–169.
55 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by the
maximum-likelihood method. Acta Crystallogr Sect D:
Biol Crystallogr 53, 240–255.
1658
FEBS Journal 281 (2014) 1642–1658 ª 2014 FEBS