Revisiting the Mechanism of the Triosephosphate Isomerase Reaction: The Role of the Fully Conserved Glutamic Acid 97 Residue

Article abstract of DOI:10.1002/cbic.201100116

An analysis of 503 available triosephosphate isomerase sequences revealed nine fully conserved residues. Of these, four residues-K12, H95, E97 and E165-are capable of proton transfer and are all arrayed around the dihydroxyacetone phosphate substrate in the three-dimensional structure. Specific roles have been assigned to the residues K12, H95 and E165, but the nature of the involvement of E97 has not been established. Kinetic and structural characterization is reported for the E97Q and E97D mutants of Plasmodium falciparum triosephosphate isomerase (Pf TIM). A 4000-fold reduction in k_cat is observed for E97Q, whereas the E97D mutant shows a 100-fold reduction. The control mutant, E165A, which lacks the key catalytic base, shows an approximately 9000-fold drop in activity. The integrity of the overall fold and stability of the dimeric structure have been demonstrated by biophysical studies. Crystal structures of E97Q and E97D mutants have been determined at 2.0 A resolution. In the case of the isosteric replacement of glutamic acid by glutamine in the E97Q mutant a large conformational change for the critical K12 side chain is observed, corresponding to a trans-to-gauche transition about the Cγ-Cδ (χ³) bond. In the E97D mutant, the K12 side chain maintains the wild-type orientation, but the hydrogen bond between K12 and D97 is lost. The results are interpreted as a direct role for E97 in the catalytic proton transfer cycle. The proposed mechanism eliminates the need to invoke the formation of the energetically unfavourable imidazolate anion at H95, a key feature of the classical mechanism.

Full text of DOI:10.1002/cbic.201100116

DOI: 10.1002/cbic.201100116
Revisiting the Mechanism of the Triosephosphate
Isomerase Reaction: The Role of the Fully Conserved
Glutamic Acid 97 Residue
Moumita Samanta,[a] M. R. N. Murthy,[a] Hemalatha Balaram,[b] and Padmanabhan Balaram*[a]
An analysis of 503 available triosephosphate isomerase se-
quences revealed nine fully conserved residues. Of these, four
residues—K12, H95, E97 and E165—are capable of proton
transfer and are all arrayed around the dihydroxyacetone phos-
phate substrate in the three-dimensional structure. Specific
roles have been assigned to the residues K12, H95 and E165,
but the nature of the involvement of E97 has not been estab-
lished. Kinetic and structural characterization is reported for
the E97Q and E97D mutants of Plasmodium falciparum triose-
phosphate isomerase (Pf TIM). A 4000-fold reduction in kcat is
observed for E97Q, whereas the E97D mutant shows a 100-
fold reduction. The control mutant, E165A, which lacks the key
catalytic base, shows an approximately 9000-fold drop in activ-
ity. The integrity of the overall fold and stability of the dimeric
structure have been demonstrated by biophysical studies.
Crystal structures of E97Q and E97D mutants have been deter-
mined at 2.0 ꢀ resolution. In the case of the isosteric replace-
ment of glutamic acid by glutamine in the E97Q mutant a
large conformational change for the critical K12 side chain is
observed, corresponding to a trans-to-gauche transition about
the CgꢀCd (c3) bond. In the E97D mutant, the K12 side chain
maintains the wild-type orientation, but the hydrogen bond
between K12 and D97 is lost. The results are interpreted as a
direct role for E97 in the catalytic proton transfer cycle. The
proposed mechanism eliminates the need to invoke the forma-
tion of the energetically unfavourable imidazolate anion at
H95, a key feature of the classical mechanism.
Introduction
The selective pressure of molecular evolution is expected to
result in the conservation of key amino acid residues and
three-dimensional structural features that are critical for bio-
chemical function. Enzymes from diverse organisms, but which
catalyse the same chemical reaction, can have substantially dif-
ferent sequences, with relatively small numbers of completely
conserved residues, when large datasets are aligned. The rela-
tionship between three-dimensional structure and catalytic
function in enzymes is often obscured by the complexities of
the three-dimensional structure. Karplus and Kuriyan noted
that “The engineering principles underlying protein design are
truly baroque, with evolutionary tinkering resulting in molecu-
lar machines that may be effective and efficient but whose
three-dimensional structures are often so complicated as to
obscure the mechanism of action”.[1]
The availability of a growing dataset of TIM sequences and
three-dimensional structures from diverse organisms prompted
us to revisit some outstanding issues relating to the mecha-
nism of the isomerization reaction. Three residues—K12, H95
and E165—have been established as critical residues in bind-
ing and catalysis.[10–14] Curiously, the small polar substrates
DHAP and GAP are anchored to the protein surface through
hydrogen-bonding interactions involving the backbone NH
groups of S211, G232, G233 and G171, with the positively
charged e-amino group of K12 providing the only obvious
electrostatic interaction, being located at a distance of 3.13 ꢀ
(PDB ID: 1NEY)[15] from the oxygen atom bonded to C3 of the
substrate. This observation led Knowles (1991) to quote Paul-
ing (1960),[16] who suggested that “a main chain ꢀNHꢀ bond is
worth about half a charge”, suggesting that interaction with
the neutral backbone “effectively balances the dianionic
charge of the phospho group”.[2] The phosphate binding
energy has been linked to the transition state stabilization that
Triosephosphate isomerase (TIM) is a ubiquitous enzyme
controlling the central step in glycolysis, catalysing the isomeri-
zation of dihydroxyacetone phosphate (DHAP) and glyceralde-
hyde-3-phosphate (GAP).[2,3] The isomerization involves proton
transfer between adjacent carbon atoms of the substrate and
the reaction proceeds at a rate limited only by diffusion.[4] The
classic studies of Rose and Knowles on TIM are central to the
historical development of mechanistic enzymology.[5–7] In re-
viewing a long line of incisive studies,[6,7] Knowles character-
ized TIM as an enzyme fashioned by “evolution to perfec-
tion”.[8] Clearly, random mutations and phenotypic selection
have designed an enzyme active site that has reached the
limits of catalytic efficiency.[9]
[a] M. Samanta, M. R. N. Murthy, Prof. P. Balaram
Molecular Biophysics Unit, Indian Institute of Science
Bangalore-560012 (India)
E-mail: pb@mbu.iisc.ernet.in
[b] H. Balaram
Molecular Biology and Genetics Unit
Jawaharlal Nehru Centre for Advanced Scientific Research
Bangalore-560064 (India)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100116.
1886
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896

Mechanism of the Triosephosphate Isomerase Reaction
leads to the large observed catalytic rate enhancement.[17,18]
A
critical role for K12 was established by the early work of
Knowles, which demonstrated that mutation of this residue
resulted in a dramatic loss in activity.[12] Subsequent theoretical
work points to the key role for K12 in transition state stabiliza-
tion.[19] More recently, Richard has elegantly demonstrated that
the cationic side chain of K12 stabilizes the transition state by
as much as 7.8 kcalmolꢀ1.[13] By a two-part substrate approach,
it has been demonstrated that the contribution of K12 to the
transition state stabilization is 6.3 kcalmolꢀ1 greater than that
provided by the ammonium ion in the K12G mutant.[20] This
study establishes that precise orientation and tethering of the
alkyl ammonium side chain of K12 is central to the catalytic
rate enhancement. Go et al. (2010) have more recently pointed
out that “the phosphate gripper motif of TIM is unlikely to
contribute any significant enthalphic advantage, but will be
effective in an entropic sense by liberating waters of hydra-
tion”.[13] The carboxylate of E165 acts as a base initiating the
first step of proton abstraction from the substrate and com-
pletes the catalytic cycle by retransfer of the proton to form
the isomerized product. H95 mediates the proton transfer.
The careful experimental work of Knowles’ group led to the
conclusion that the neutral imidazole side chain of H95 func-
tions as an electrophile in the catalytic process.[21] This unusual
situation would require the formation of an imidazolate anion,
a process characterized by a very high pKa value (ꢁ14). Com-
bined QM/MM calculations carried out by Bash and Karplus
lent support to this unusual mechanism,[19] which, as Knowles
acknowledges, “runs counter to the prejudices of mechanistic
chemistry”.[21] The availability of a high-resolution crystal struc-
ture of an enzyme/substrate complex (yeast TIM/DHAP com-
Figure 1. View of the glutamic acid 97 (E97) residue along with the active
site residues K12, H95 and E165 with DHAP bound at the active site in the
yeast TIM/DHAP complex structure (PDB ID: 1NEY). All figures were generat-
ed with Pymol (http://www.pymol.org).
catalytic activity. Crystal structure determination of both mu-
tants has provided insights into the role of residue 97 and
established the dramatic reorientation of the critical K12 side
chain in the E97Q mutant.
plex) (PDB ID: 1NEY),[15] together with the growing volume of Results
sequences and related three-dimensional structures, prompted
Characterization of mutants
us to re-examine the distribution of the fully conserved resi-
dues in the vicinity of the enzyme active site. Inspection of a
dataset of 503 TIM sequences, from diverse organisms of arch-
aeal, bacterial and eukaryotic origins, reveals only nine fully
conserved residues: K12, T75, H95, E97, C126, E165, P166, G209
and G228 (for all the fully conserved residues, the numbering
schemes are identical for both Pf and yeast triosephosphate
isomerase). Of these, E97 is positioned directly at the active
site in close proximity to K12 and H95 as shown in Figure 1.
The presence of an ionizable side chain strategically positioned
between K12 and H95 suggests a potential role in the proton
transfer reaction.
The protein mutants were characterized by mass spectrometry
and by determination of their crystal structures. Although the
presence of each mutation was confirmed by sequencing of
the cloned gene, it was felt that an unambiguous characteriza-
tion of the mutation in the protein was both desirable and im-
portant, in view of the differences between our results on the
E97D mutant and the results of Knowles and co-workers on
the corresponding mutant of the chicken enzyme.[22] The de-
tails of the ESI MS characterization are presented in Figure S1
in the Supporting Information.
The availability of single crystals facilitated further characteri-
zation. Figure 2A and B provide a view of the electron density
in the vicinity of the site of the mutations. In the E97D mutant,
characterization of the mutation is readily accomplished by fit-
ting the observed electron density, as a result of the shorten-
ing of the side chain. In contrast, the electron density does not
permit any distinction between the glutamic acid and the glu-
tamine side chains. In the case of the E97Q mutant, a dramatic
reorientation of the proximal K12 side chain is observed. This
is presumably a consequence of the loss of the favourable
interaction between the positively charged e-amino group and
the negatively charged carboxylate as a result of the E97Q
mutation. This is further elaborated in the Discussion section.
Nearly twenty years ago, Knowles and co-workers reported
that a single-residue mutant, E97D, retained full catalytic activi-
ty,[22] an observation that suggests no direct role for E97 in cat-
alysis, despite its proximity to the active site and its complete
conservation. Intrigued by this observation, we have generated
and characterized, structurally and kinetically, the active site
mutants E97Q and E97D in Plasmodium falciparum triosephos-
phate isomerase (Pf TIM). The E97Q mutant, in which the ioniz-
able glutamic acid side chain is replaced by the isosteric, non-
ionizable glutamine side chain, is almost completely inactive.
The aspartic acid replacement results in side chain shortening
in the E97D mutant, which displays significantly attenuated
ChemBioChem 2011, 12, 1886 – 1896
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1887

P. Balaram et al.
are located on a single subunit. The side chain of residue 97
does indeed make hydrogen bond contacts to the fully con-
served T75 residue from the other subunit (see the Discussion).
We therefore examined the concentration dependence of
activity for the wild-type enzyme and for the E97Q and E97D
mutants over a wide protein concentration range of 0.2 mm to
4 mm. Figure S3 shows that over this concentration regime
there is no change in the specific activities of the wild-type
and mutant enzymes, suggesting that the replacements at resi-
due 97 have not significantly altered the stability of the dimer.
It might be noted that monomeric TIMs engineered by Wieren-
ga’s group retain the overall (a/b)8 barrel fold, but display an
approximately 1000-fold fall in activity, suggesting a complete
dependence of catalytic efficiency on dimer integrity.[23]
Figure 2. Electron-density maps (2 jFoꢀFc j contoured at 1.0s) at the active
site residues and residue 97 in: A) E97Q (3PSW), and B) E97D (3PSV) struc-
tures.
Biophysical studies
Enzyme activity
Figure S4 shows the far-UV CD and fluorescence emission
spectra of the wild-type enzyme and of the mutants E97Q and
E97D. The identical natures of the spectra suggest that the
overall fold and the tryptophan residue (W11 and W168) envi-
ronments are largely unperturbed by site-specific mutations at
position 97. Furthermore, the thermal melting profiles deter-
mined by monitoring of CD ellipticity at 222 nm are almost
identical for the three proteins. The positions of the fluores-
cence maxima (331 nm)[24] were identical for the three proteins
and were concentration-independent over the concentration
range from 0.5 to 20 mm (Figure S5). Analytical gel filtration
profiles (Figure S6), determined over a protein concentration
range from 0.5 to 20 mm, reveal identical elution volumes for
the wild-type enzyme and for the E97Q and E97D mutants,
confirming the dimeric nature of all three proteins, even at the
lowest concentration studied.
A comparison of the activities of the mutants and the wild-
type enzyme as a function of substrate concentration is shown
in Figure S2. Table 1 summarizes the kinetic parameters deter-
Table 1. Kinetic parameters for the wild-type enzyme and for the mu-
tants E97Q, E97D and E165A.
Enzymes
kcat [sꢀ1
]
Km [mm]
kcat/Km [mmꢀ1 sꢀ1
]
wild-type
E97Q
E97D
4.30ꢂ0.03ꢂ103
1.00ꢂ0.15
0.35ꢂ0.05
1.20ꢂ0.20
0.50ꢂ0.02
0.43ꢂ0.05
1.22ꢂ104
0.90
8.40ꢂ10
0.93
4.20ꢂ0.80ꢂ10
0.45ꢂ0.08
E165A
mined for the Pf TIM WT and the three mutants E97Q, E97D
and E165A. The E97Q and the E165A mutants exhibit approxi-
mately 4000- to 9000-fold drops in kcat, suggesting the almost
complete abolition of catalytic activity upon mutation. The
E97D mutant shows an appreciably higher kcat value, albeit still
100 times lower than that of the native enzyme. The back-
ground activities of the host AA200 cells and the host cells
transformed with the vector (pTrc99A) were also measured.
Analysis of the crystal structures
Crystals of the Pf TIM mutants E97Q and E97D each diffracted
to 2.0 ꢀ, with one protein dimer in the asymmetric unit. In
both cases, loop 6 was in the “open” conformation and the
active site was unliganded. In the liganded, loop-closed confor-
mation of TIM, the flexible loop 6 serves to sequester the reac-
tive enediol intermediate and to expel water molecules from
the immediate vicinity of the reactive functionalities.[9,31,32] De-
spite attempts to crystallize the mutants in the presence of
ligands, none of the crystals examined incorporated the ligand.
In each of the mutants, electron density corresponding to an
ethylene glycol molecule and a sulfate ion could be clearly
visualized. Figure 3 compares the active site structure of un-
liganded Pf TIM in the loop 6 open conformation (PDB ID:
1YDV)[25] with the corresponding residues in the two mutant
forms. In the case of Pf TIM, and indeed TIMs from other sour-
ces, ligand binding and loop 6 closure leaves the K12, E97 and
H95 residues unchanged, with a small movement of the E165
residue. Inspection of Figure 3B immediately reveals a large
conformational change involving the K12 side chain in the
case of the E97Q mutant. In all TIM structures determined so
far, the K12 side chain adopts a fully extended conformation.
The
0.001 mmolsꢀ1 mgꢀ1 may be compared with those obtained for
the Pf TIM WT (108 mmolsꢀ1 mgꢀ1), E97Q (0.046 mmolsꢀ1 mgꢀ1
observed
background
specific
activity
of
)
and E97D (1.6 mmolsꢀ1 mgꢀ1). These results suggests that the
shortening of the acidic side chain at position 97 in the E97D
mutant results in a significant reduction (100-fold) in activity.
Nevertheless an appreciable level of isomerase activity is still
observable. In contrast, the reduction in activity of the E97Q
mutant is ꢁ4000-fold whereas that observed for the E165A
mutant, which lacks the catalytic group involved in the initial
proton abstraction step, is ꢁ9000-fold. The replacement of the
glutamic acid at position 97 by glutamine can result only in
changes in electrostatic and hydrogen-bonding interactions,
because the two residues are almost identical in shape and
size, differing only in the ionizabilities of the side chains. The
TIM active site is located close to the dimer interface, although
all the residues important for the chemical steps of catalysis
1888
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896

Mechanism of the Triosephosphate Isomerase Reaction
Figure 3. Comparative views of the active site residues and the key interacting distances in: A) the Pf TIM wild type (PDB ID: 1YDV), B) E97Q (PDB ID: 3PSW),
and C) E97D (PDB ID: 3PSV).
In contrast, in the E97Q structure, the dihedral angle c3 (Cgꢀ the amino group at position 12. In this case a NꢀO distance of
Cd) has a gauche(ꢀ) conformation (c3=ꢀ75.18). In all other
structures, c3 has a trans conformation (c3ꢁ1808). Both in the
yeast TIM/DHAP complex (PDB ID: 1NEY)[15] and in the un-
liganded Pf TIM WT structure (PDB ID: 1YDV)[25] the e-amino
group of K12 forms a hydrogen bond with the carboxylic acid
side chain of E97. The observed NꢀO distances are 2.81 ꢀ in
the yeast/DHAP complex and 2.82 ꢀ in the unliganded Pf TIM
wild type. In the E97Q mutant, the conformational change in
the lysine side chain results in a loss of the hydrogen-bonding
interaction; the observed NꢀO distance increases to 4.8 ꢀ.
In the case of the E97D mutant, the K12 side chain adopts
the fully extended conformation observed in all other TIM
structures. The shortening of the side chain at position 97
from Glu to Asp results in a loss of the hydrogen-bonding in-
teraction between the carboxylic acid group at position 97 and
4.8 ꢀ is observed.
Mutation of the fully conserved Glu(E)97 residue of Pf TIM to
Gln (Q) results in a dramatic reduction in isomerase activity.
The carboxamide group of the glutamine side chain is isosteric
with the carboxylic acid moiety of the glutamic acid side chain.
They differ, however, in an important respect in their hydro-
gen-bonding abilities. Figure 4 compares the hydrogen bonds
formed by the residue 97 side chain in the Pf TIM wild type
and in the E97Q and E97D structures. In the native enzyme,
the carboxylic acid group of E97 is involved in multiple hydro-
gen-bonding interactions. One oxygen atom participates in
three potential hydrogen-bonding interactions with residues
from the other subunit. Three short inter-subunit interactions
can be identified: with the Thr75 side chain hydroxy group
(OꢀO: 2.6 ꢀ), the Thr75 backbone NH (NꢀO: 3.06 ꢀ) and the
Figure 4. Hydrogen-bonding interactions involving the residue 97 side chain with surrounding residues and water molecules: A) in the unliganded Pf TIM wild
type (PDB ID: 1YDV), and B) in E97Q (PDB ID: 3PSW), and C) in E97D (PDB ID: 3PSV). Dark grey and light grey represent residues from different subunits.
ChemBioChem 2011, 12, 1886 – 1896
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1889

P. Balaram et al.
Ser73 backbone CO (OꢀO: 3.26 ꢀ). Although the state of ioni-
zation of the carboxylic acid group cannot be directly ascer-
tained through the location of the H atom, the observation of
a rather short OꢀO distance between the Ser73 backbone CO
group and the Glu97 side chain oxygen is suggestive of—but
does not definitively establish—an un-ionized carboxylic acid
group at this position. This is consistent with the participation
of the oxygen in three hydrogen-bonding interactions, acting
as a donor in one case and an acceptor in the other two. The
second side chain oxygen atom of Glu97 makes a hydrogen
bond to the e-amino group of Lys12 and a water molecule
(W566), which in turn is hydrogen-bonded to the water mole-
cule W633, which also interacts with the S73 backbone carbon-
yl group. These two water molecules are conserved in all TIM
structures from diverse organisms, with the exceptions of the
tetrameric enzymes from the thermophilic organisms Methano-
caldococcus janaschii (PDB ID: 2H6R)[26] and Pyrococcus woesei
(PDB ID: 1HG3).[27] In the E97Q mutant, the hydrogen-bonding
interactions across the dimer interface are maintained. The
large conformational change in the K12 side chain presumably
arises as a result of the close proximity of the amide NH2 of
Q97 and the e-NH2 of K12. It should be noted that in fitting of
primary amide side chains to electron densities, distinctions
between the two possible orientations of the amide group are
generally made, by inspection of the potential hydrogen-bond-
ing interactions and steric clashes.[28] In the liganded TIM struc-
tures, a conserved water molecule forms a hydrogen-bonded
bridge between the lysine e-NH2 group and one of the phos-
phate oxygen atoms. This is demonstrated in the case of the
Pf TIM wild type/3PG complex (PDB ID: 2VFI) shown in Fig-
ure 5A, in which W2167 acts as the bridge. In the unliganded
TIM structures, several water molecules form a hydrogen-
bonded cluster at the unoccupied active site (Figure 6). Ligand
binding results in a displacement of these water molecules. In
the E97Q structure, displacement of the Lys e-amino group
from the position occupied in the wild-type enzyme places it
in the position of the bridging water molecule, resulting in the
formation of a hydrogen bond to one of the oxygen atoms of
the sulfate anion, located near the active site (Figure 5B). Inter-
estingly, in the E97D structure, in which K12 occupies a posi-
tion identical to that in the wild-type enzyme, the bridging
water molecule (W431) forms a hydrogen bond to the Lys e-
amino group and the sulfate ion (Figure 5C). The identical ori-
entation of the K12 side chain in the wild-type and the E97D
structures suggests that the hydrogen-bonding interaction be-
tween the e-ammonium group and the residue 97 carboxylate
group is not critical for positioning the catalytically crucial K12
side chain. The absence of any effect of the E97Q mutation on
the stability of the dimer is established by the absence of con-
centration dependence of enzyme activity, spectroscopic prop-
erties and the elution volumes in analytical gel filtration experi-
ments. This is consistent with the crystallographic observa-
tions, which reveal no change in interactions involving this res-
idue across the subunit interface. The dramatic loss in activity
is also consistent with structural evidence, which shows a large
change in the position of the critical e-amino group of K12.
Previous work has established that mutation of the K12 residue
results in almost complete loss of activity. Lodi et al. (1994) re-
ported kcat values of 0.018 sꢀ1 for K12M, 48 sꢀ1 for K12R, 6.3 sꢀ1
for K12H (pH 6.1) and 0.89 sꢀ1 for K12H (pH 7.5). The corre-
sponding yeast wild-type enzyme has
a kcat value of
8700 sꢀ1.[12] More recently, Go et al. have reported a kcat value
of 0.6 sꢀ1 for the K12G mutant of TIM.[13] These results suggest
that precise orientation of the K12 residue is essential for
optimal activity of the enzyme. Furthermore, the unchanged
K12 orientation, together with impaired activity in the E97D
mutant, points to a direct role for the E97 carboxylate in the
catalytic process. The hydrogen bond between E97 and K12
side chains, both of which are completely conserved, is also
suggestive of an important role for this ionizable side chain in
the proton-transfer reaction.
Figure 5. Important interactions of the critical residue K12 side chain: A) in the Pf TIM/3PG structure (PDB ID: 2VFI), and B) in E97Q (PDB ID: 3PSW), and C) in
E97D (PDB ID: 3PSV).
1890
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896

Mechanism of the Triosephosphate Isomerase Reaction
Figure 6. Water network at the active site of TIM: A) when unliganded (PDB ID: 1YDV), and B) liganded (3PG, PDB ID: 2VFI). Dark grey and light grey represent
residues from different subunits.
Discussion
formation of an imidazolate anion, which then reabstracts a
proton from the C1 position of the enediol, with subsequent
reprotonation of the substrate by the carboxylic acid at
E165.[11] Lodi and Knowles argued that the neutral state of
His95 is a consequence of its position at the positive end of
the macrodipole produced by the a helix formed over the resi-
dues 97–102.[33] The formation of an imidazolate anion is, how-
ever, a process that is not really anticipated on the basis of the
estimated acidity of neutral imidazole (pKa ꢁ14). In an influen-
tial overview, Knowles raised a question: “Why, when every-
thing else about this enzyme seems chemically so reasonable,
it should fail to use the more powerful electrophile [imidazo-
lium ion of protonated His95], remains a puzzle.”[9] Neverthe-
less, this proposal has been completely accepted, with support
being derived from combined QM/MM calculations carried out
by Karplus and co-workers.[19,34] Furthermore, FTIR investiga-
tions have established that in enzyme/substrate complexes,
the carbonyl stretching frequency of DHAP is sensitive to mu-
tations of H95 (H95Q, H95N), but insensitive to mutations at
K12. Binding of DHAP to wild-type yeast TIM results in a shift
of the carbonyl stretching frequency from 1732 cmꢀ1 in the
free ligand to 1713 cmꢀ1 in the bound state. In contrast, the
observed stretching frequencies in the H95Q and H95N mu-
tants (from 1732 to 1742 cmꢀ1) were unchanged, suggesting
that the e-ammonium group of Lys12 does not substantially
contribute to polarization of the carbonyl group of DHAP.[11]
Inspection of the yeast TIM/DHAP complex structure (PDB ID:
1NEY)[15] reveals that H95 is geometrically better positioned
than the Lys12 e-amino group to form a hydrogen bond to the
carbonyl oxygen of DHAP [N(H95)ꢀO dist: 2.72 ꢀ, aCON(H95):
121.28; N(K12)ꢀO: 3.07 ꢀ, aCON(K12): 105.38]. The geometry
Revisiting the proton transfer mechanism in TIM
The stereospecific abstraction of the pro-R hydrogen from the
hydroxymethylene group of DHAP by the carboxylate of E165
and the subsequent formation of the enediolate intermediate
constitutes the initial, critical step in the isomerization reac-
tion.[5] The K12 e-amino group is well positioned to provide
electrostatic stabilization of the enediolate intermediate, a
common feature of the aldose-ketose isomerases, as pointed
out by Rose nearly half a century ago.[5,29] His95 is the other
key residue directly involved in facilitating the proton-transfer
process. The extensive work of Knowles and his collaborators
has established an important attribute of H95: namely that the
imidazole side chain is neutral (uncharged).[21] Inspection of the
crystal structures of TIMs from different sources establishes
that the positions of the side chains of the Lys12, His95 and
Glu97 residues are identical, and remain unchanged both in
loop 6 open, unliganded structures and in loop 6 closed, li-
ganded structures. This suggests a largely preexisting orienta-
tion of these three active-site residues. Substrate binding and
loop closure results only in the movement of the catalytically
critical E165 side chain.[14,30] Loop 6 dynamics are of central im-
portance in the TIM reaction.[31] Loop closure over the bound
substrate serves to dehydrate the active site and to position
the hinge residue E165 for the proton-abstraction step. The
placement of this side chain in a “hydrophobic cage” has been
suggested to be responsible for an increase in the basicity of
the E165 carboxylate group.[32] The work of Knowles and co-
workers led to the suggestion that H95 functions as an electro-
phile, donating a proton from its neutral state, resulting in the
ChemBioChem 2011, 12, 1886 – 1896
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1891

P. Balaram et al.
of approach is more favourable in the former case. However,
the e-amino group can contribute very substantially to the sta-
bilization of the endiolate anion both electrostatically and
through hydrogen bonding, because of rehybridization of the
oxygen atom in the endiolate. Ionic hydrogen bonds are also
substantially stronger than the interactions between neutral
donors and acceptors.[35] Interestingly, the H95Q mutant
showed a 500-fold reduction in enzyme activity, leading to the
suggestion that an alternative mechanism of proton transfer is
operative in the mutant.[11]
At this juncture, the following question may be raised. Can
the protonated ammonium group of K12 serve to transfer a
proton to the oxygen atom at the C2 position of the enedio-
late anion, leading to the neutral enediol? Bash et al. noted in
1991 that although the pKa of the enediol has not been mea-
sured, the values for enols are in the neighbourhood of 11.[19]
The pKa of the Lys e-ammonium group might be slightly lower,
so this transfer could indeed be favourable. In contrast, the pKa
for the process His!Hisꢀ+H+ is of the order of 14. His95 in its
neutral form would then serve to abstract the proton from the
C1 OH group of the neutral enediol, with subsequent ketoniza-
tion and protonation of the C2 carbon by E165. Rapid proton
transfer through the hydrogen bond network involving H95,
E97 and K12 would restore the initial protonation states of
these residues, thereby completing the catalytic cycle. In par-
ticular, a 1808 flip of the His95 side chain about the CbꢀCg
bond can transiently change the positions of the protonated
and deprotonated N atoms, facilitating proton abstraction
from the enediol and also proton transfer to re-establish
charge states after the catalytic step. Indeed, a reaction-driven
imidazole ring flip has been proposed at the active site of a
serine protease.[36] Knowles did in fact anticipate “the seductive
possibility… that the proton transfers are mediated merely by
rotation of the appropriate side chain: the bidentate carboxyl
group for the manipulation of carbon-bound protons, and the
bidentate imidazole group for handling the oxygen-bound
protons”.[9] In the mechanism, proposed on the basis of the re-
sults presented here, summarized in Scheme 1, His95 is neutral
In this study, a 4000-fold loss of activity and a large confor-
mational change of the K12 side chain has been established
for the E97Q mutant of Pf TIM. It would therefore appear rea-
sonable to assume a critical role for the fully conserved E97
residue in the catalytic process. Are the K12 and the E97 resi-
dues functioning only in precisely orienting the anionic sub-
strate and the catalytically important H95 residue? The E97
side chain forms a strong hydrogen bond to the K12 e-amino
group (NꢀO, 2.8 ꢀ). The planar carboxylic acid of E97 and the
imidazole group of H95 are aligned approximately parallel and
are close-packed, with an interplanar distance of ꢁ3.4 ꢀ.
Knowles and co-workers have pointed out that the neutral
group of imidazole of H95 forms a hydrogen bond (=N···HꢀN
3.04 ꢀ) with the backbone NH of E97. This would orient the
imidazole such that the protonated nitrogen would be proxi-
mal to the substrate, and led to the proposal of a step involv-
ing proton abstraction from the neutral imidazole.[21]
Scheme 1. Schematic representation of the proposed proton-transfer mechanism in the triosephosphate isomerase catalytic cycle.
1892
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896

Mechanism of the Triosephosphate Isomerase Reaction
in the resting state of the enzyme, as established by Knowles
and co-workers. The candidate “electrophile” in this alternative
mechanistic proposal, however, is the e-ammonium group of
K12, the Brçnsted acidity of which might be more evenly
matched with the basicity of the initially formed enediolate
anion. His95 performs a critical—albeit more conventional—
catalytic role in acting as a base for proton abstraction from
the enediol. Bash et al. (1991) implicitly acknowledged the pos-
sibility of dynamic effects in noting: “Neither the structural
data, nor the calculations can exclude the possibility that the
transiently formed protonated His95 is the catalytic acid, nor
that a concerted proton transfer between substrate oxygen
atoms occurs via a neutral histidine as a relay”.[19] In a compre-
hensive discussion of the role of Lys12 in the isomerization
reaction catalysed by TIM, Richard and co-workers raised the
issue of whether the intermediate is a charged enediolate or a
neutral enediol. They suggest “that there is a strong catalytic
imperative to the avoidance of full transfer of a proton to the
enediolate oxyanion”. They argue that “full proton transfer to
the enediolate oxyanion could eliminate the large stabilizing
electrostatic interaction with the cationic side chain of
Lys12”,[13] which undoubtedly is a critical interaction for the
catalytic process. The alternative mechanistic proposal consid-
ered by us for the now classical isomerization reaction cata-
lysed by triosephosphate isomerase is based on the finding
that the E97Q mutant is 4000 times less active than the wild-
type enzyme. This may be compared with the 500-fold loss of
activity of the H95Q mutant.[11] The only striking structural dif-
ference between the wild-type enzyme and the E97Q mutant
is the dramatic reorientation of the K12 side chain. Indeed, out
of all the reported TIM structures, this is a sole example of a
large change in the position of the K12 side chain. Conserva-
tion of the E97 residue in all TIM sequences so far reported
suggests a critical role. The TIM active site is intricately con-
structed through the involvement of highly directional hydro-
gen-bonding interactions between the subunits, ensuring that
dimerization is a prerequisite for optimal catalytic competence.
The well-characterized monomeric TIMs engineered by Wieren-
ga and co-workers are at least 1000 times less active than their
dimeric counterparts.[23] Careful examination of the intersub-
unit interactions involving Glu(E) and Gln(Q) at position 97 led
to the conclusion that this mutation leaves the dimer interface
unperturbed. Indeed, experimental studies revealed no differ-
ences in the concentration dependence of biophysical parame-
ters of the wild-type and mutant enzymes. These observations
support the direct involvement of E97 in the catalytic cycle.
In the E97D mutant, the acidic group has been retained, but
the side chain is significantly shortened. This mutant shows a
significantly lower catalytic activity, with a reduction in kcat of
ꢁ100-fold. This observation is in contrast to that reported for
the E97D mutant of the chicken enzyme,[22] which shows no
loss of activity upon mutation. The reason for the differences
between these earlier studies and our present report are not
clear. Careful examination of the sequences and the crystal
structures of the chicken (PDB ID: 1TPH) and Pf enzymes (PDB
ID: 1O5X) establishes almost complete conservation of residues
implicated in substrate binding and catalysis. Indeed the facile
mutational interconversion of the codons for the E and D resi-
dues by changes at the third position (E: GAA, GAG; D: GAU,
GAC) suggests that conservation must be driven by a strong
phenotypic selection process. Inspection of the crystal struc-
ture of the E97D mutant reveals that the dimer interface inter-
actions are maintained and the orientations of the carboxylate
group of D97 and the imidazole group of H95 are also closely
similar to those in the wild-type enzyme (Figures 4 and 5). The
K12 side chain maintains the same position as in the wild-type
enzyme, despite the absence of a hydrogen-bonding interac-
tion with the carboxylate of D97. The corresponding NꢀO dis-
tances in the E97D structures are 4.83 ꢀ and 4.47 ꢀ. The fall in
activity might therefore arise from a disruption of the hydro-
gen bond network involving residues K12, E97 and H95.
In all of the TIM sequences reported so far, multiple se-
quence alignment reveals only nine completely conserved resi-
dues, of which four are capable of accepting and donating
protons. These four residues—two acidic (E165 and E97) and
two basic (H95 and K12) residues—are arrayed around the
small polar substrate. Whereas critical roles have been pro-
posed for three of these residues—E165, H95 and K12—in
binding and catalysis, no direct involvement of the conserved
E97 residue has been suggested. The results of this study es-
tablish that the isosteric E97Q replacement results in a dramat-
ic loss of catalytic activity and a large structural change in the
critical K12 side chain. Shortening of the side chain at posi-
tion 97, with retention of the acidic functionality in the E97D
mutant, leads to a 100-fold drop in catalytic efficiency, despite
the retention of the K12 side chain conformation identical to
that observed in the wild-type TIM. We believe that these re-
sults provide compelling evidence of direct involvement of the
E97 side chain in the proton-transfer reaction. The mechanistic
proposal advanced by us eliminates the need to invoke the
intermediacy of a chemically unlikely species, the imidazolate
anion at H95.
Conclusions
In summary (Scheme 1), the proton-transfer cycle in the TIM-
catalysed isomerization reaction begins with proton abstrac-
tion by the carboxylate anion of E165. A hydrogen bond net-
work involving the enediolate anion, the e-ammonium group
of K12, the carboxylic acid of E97 and the neutral imidazole of
H95 might provide a facile channel for proton hopping during
the catalytic cycle. K12 might indeed play a critical role by ab-
stracting a proton from the neutral E97 carboxylic acid, even
as it loses a proton to the enediolate oxygen to yield a neutral
enediol. It is conceivable that this concerted proton transfer
occurs via a five-coordinate transition state at the e-amino
group of Lys12, thereby avoiding the need for full proton
transfer to form a neutral amino group. Although the existence
of pentacoordinate nitrogen species has been the subject of
both experimental and theoretical debate, Olah and co-work-
ers have suggested that this possibility may be considered for
hydrogen because of its small covalent radius.[37] H95 can now
act in a more conventional role as a base, abstracting the
proton from the OH group on C1, setting in motion the re-
ChemBioChem 2011, 12, 1886 – 1896
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1893

P. Balaram et al.
E165A mutant, was also generated as an enzyme activity control in
which the key catalytic base was replaced by alanine. For construc-
tion of this mutant the wild-type gene was used as the template
and the single mutagenic 5’ primer 5’-GATAAT GTTATT TTGGCA
TATGCA CCTTTA TGGGCT ATTGGT AC-3’ was used with the restric-
tion site NdeI (detailed protocols provided in the Supporting Infor-
mation).
arrangement leading to reprotonation at C2 by transfer of the
proton from the neutral E165 residue. Although dissection of
the individual steps of proton transfer is useful in following
through the states of protonation of acidic and basic residues
during the catalytic cycles, it must be emphasised that all
steps—proton transfer and side-chain flips—could occur in a
near concerted manner. Although the resting state of the
enzyme has been assumed to have a protonated (uncharged)
E97 residue, it is conceivable that the anionic, ionized form
might also mediate proton transfer between a transient imida-
zolium ion at H95 and a neutral e-amino group at K12. Hydro-
gen-bond networks, when coupled to dynamic motions of cat-
alytic residues, can indeed contribute to the high efficiency of
enzymes involved in hydrogen-transfer reactions.[38] All the
fully conserved residues in triosephosphate isomerase occur di-
rectly at the active site or in close proximity. Three attributes
of triosephosphate isomerase, and indeed multimeric enzymes
in general, are considered important for function. These are
1) proper folding into a native three-dimensional structure,
2) specific subunit association to give functional multimers,
and 3) precise orientation of the active-site residues together
with suitable conformational flexibility to accommodate the
dynamic requirements of catalysis.[39] The availability of a large
sequence dataset for enzymes suggests that there are multiple
solutions to the problems of protein folding and specific pro-
tein–protein interactions leading to oligomerization. Catalytic
chemistry is, however, more severely constrained, resulting in
the overwhelming preponderance of fully conserved residues
at the active site. This view echoes a four-decade-old paper by
John Maynard Smith[40] that suggests, as paraphrased by
Knowles, that “enzymes represent the rare end products of an
extensive search through protein sequence space”.[2]
Protein expression and purification: E. coli AA200 cells (with de-
leted inherent TIM gene), transformed with the recombinant
vector containing the mutant TIM gene, was grown at 378C in Ter-
rific broth containing ampicillin (100 mgmLꢀ1). Cells were induced
by use of IPTG (isopropyl-b-d-thiogalactopyranoside, 300 mm) at
0.6–0.8 OD600 nm and were harvested by centrifugation. Cells were
resuspended in lysis buffer [TrisHCl (pH 8.0, 20 mm), ethylene-
diaminetetraacetate (EDTA, 1 mm), phenylmethanesulfonylfluoride
(PMSF, 0.01 mm), dithiothreitol (DTT, 2 mm) and glycerol (10%)]
and disrupted by sonication. After centrifugation and removal of
the cell debris, the supernatant was subjected to ammonium sul-
fate precipitation. The protein fraction containing TIM was selec-
tively precipitated above 70% ammonium sulfate saturation. This
precipitate was collected by centrifugation and resuspended in
buffer A [TrisHCl (pH 8.0, 20 mm), DTT (2 mm) and glycerol (10%)].
Removal of nucleic acid was effected by polyethylene imine precip-
itation. All purification steps were carried out at 48C. The protein
was dialysed extensively against buffer A at 48C overnight, purified
with the aid of an anion-exchange column (Q-sepharose, HR 60)
and eluted with a linear gradient of NaCl (0–1m). The fractions
containing the protein were pooled and precipitated by addition
of ammonium sulfate up to a concentration of 75%. The precipi-
tated protein was dissolved in buffer A and subjected to gel filtra-
tion chromatography (Sephacryl-200), with equilibration with the
same buffer in an AKTA BASIC FPLC system. Protein purity was
checked by 12% SDS-PAGE. The masses of the mutant proteins
were confirmed by LC ESI MS with an ion trap mass spectrometer
or a Q TOF system. Protein concentration was determined by Brad-
ford’s method[44] with use of BSA as a standard. Protein yields of
40–50 mgLꢀ1 of E. coli culture was obtained.
Experimental Section
Enzyme assay: Kinetic measurements were carried out by the
method of Plaut and Knowles[45] at room temperature. The TIM-
promoted conversion of GAP (glyceraldehyde 3-phosphate) into
DHAP (dihydroxyacetone phosphate) was coupled to the sub-
sequent reaction by use of the coupling enzyme a-glycerol phos-
phate dehydrogenase. In this coupled assay reaction, a total
reaction volume of 1 mL contained TEA (pH 7.6, 100 mm), EDTA
(5 mm), NADH (0.5 mm), a-glycerol phosphate dehydrogenase
(20 mgmLꢀ1) and GAP, to which TIM was added at the end. Where-
as in the case of the wild-type enzyme the assay could be started
with 10 ng of the protein, in the cases of the mutants, 1000 ng of
E97D, 10000 ng of E97Q and 10000 ng of E165A were needed to
initialize the reaction. Substrate concentrations were varied from
0.25 mm to 4.0 mm. The enzyme activity was determined by
monitoring the decrease in absorbance of NADH at 340 nm. The
extinction coefficient of NADH was taken to be 6220mꢀ1 cmꢀ1 at
340 nm.[46] The experiments were repeated at least three times
from independent purification batches. The values for the kinetic
parameters (Km, kcat) were determined by fitting to Michaelis–
Menten equation with the aid of Graphpad Prism software (Ver-
sion 5 for Windows, Graphpad Software, San Diego, California,
USA, http://www.graphpad.com).
Mutagenesis: The Plasmodium falciparum triosephosphate isomer-
ase gene was cloned into the expression vector pTrc99A, called
pARC1008.[41] The protein was overexpressed into AA200, an ampi-
cillin-resistant E. coli strain, from which the host TIM gene has been
deleted.[42] To probe the role of the residue at 97, two single mu-
tants—E97Q and E97D—were constructed by site-directed muta-
genesis by the single-primer method.[43]
Because of the lack of availability of a restriction site at the desired
position of mutation, a two-step process was followed:
Step 1) Generating the E97 int clone with the introduction of the re-
striction site of ECoRV at the desired position of mutation: In this
step, the wild-type gene was taken as the template and the single
5’ primer 5’-GTTATT ATTGGT CATTTT GATATC AGAAAA TATTTC
CATGAA ACCG-3’ was used with the ECoRV restriction site, to gen-
erate the E97 int clone.
Step 2: The intermediate clone was taken as the template and the
mutant clones E97Q and E97D were generated with the subse-
quent removal of the restriction site of ECoRV and introduction of
the required mutations. The 5’ primers used for generating the
E97Q and E97D clones from the template E97 int clone were 5’-
GTTATT ATTGGT CATTTT CAGAGA AGAAAA TATTTC CATGAA ACCG-
3’ and 5’-GTTATT ATTGGT CATTTT GATAGA AGAAAA TATTTC
CATGAA ACCG-3’, respectively. For this study, another mutant, the
Fluorescence spectroscopy, circular dichroism and size-exclusion
chromatography: Experimental details are provided in the Sup-
porting Information.
1894
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896

Mechanism of the Triosephosphate Isomerase Reaction
Table 2. X-ray diffraction data collection statistics.
Table 3. Crystallographic refinement statistics.
E97Q
E97D
E97Q
E97D
PDB ID
space group
unit cell a [ꢀ]
b [ꢀ]
c [ꢀ]
3PSW
P21221
50.91
53.75
174.48
90
3PSV
P21221
51.06
53.77
174.70
90
resolution range [ꢀ]
number of subunits/asymmetric unit
number of reflections used
% observed
Rwork [%]
Rfree [%]
45.8–1.99
2
30019
89.3
20.1
25
43.7–2.0
2
31746
94.9
18.5
23
a=b=g [8]
resolution range [ꢀ]
no. of reflections
no. of unique reflections
completion [%][b]
overall R merge[%][b]
multiplicity[b]
45.80–1.99
217608
33609
90.20 (93.80)
19.80 (53.50)
7.20 (5.00)
15.23 (2.88)
0.30
43.70–2.00
286503
33483
95.40 (95.30)
12.10 (43.00)
9.0 (8.0)
28.28 (5.84)
0.30
Model quality
number of protein atoms
number of water molecules
average B factor [ꢀ2]: protein
water
RMSD from ideal: bond length [ꢀ]
bond angle [8]
Ramachandran statistics
most-allowed region [%]
allowed region [%]
generously allowed region [%]
disallowed region [%]
4358
475
20.43
31.45
0.007
0.96
4341
456
18.44
29.56
0.007
0.98
[b]
hI/sIi
average mosaicity
93.8
6.2
0
93.6
6.4
0
[b] Represents highest-resolution shells.
0
0
Crystallization of E97Q and E97D: A protein concentration of
10 mgmLꢀ1 was used for crystallization experiments based on the
hanging drop method, in which each crystallization droplet con-
tained protein (ꢁ10 mgmLꢀ1, 3 mL) with cocktail buffer (3 mL) and
was equilibrated with reservoir solution (500 mL). The E97Q crystal
used for data collection was obtained under these conditions: PEG
(24%), HEPES buffer (pH 7.0, 100 mm), DTT (0.5 mm), sodium azide
(0.5 mm) and Li2SO4 (10 mm). The E97D crystal was obtained under
these conditions: PEG (16%), HEPES buffer (pH 7.0, 100 mm, DTT
(0.5 mm), sodium azide (0.5 mm) and Li2SO4 (10 mm).
Acknowledgements
P.B. is deeply indebted to N. V. Joshi for his analysis of TIM se-
quences and helpful discussions. We thank Kallol Gupta for the
mass spectral analysis of the mutant proteins. M.S. was support-
ed by a Senior Research Fellowship from the Council of Scientific
and Industrial Research (India). Research on TIM was supported
under a grant from the Department of Science and Technology
(India). X-ray diffraction and mass spectrometry facilities were
supported by program grants from the Department of Science
and Technology and Department of Biotechnology (India).
Data collection and processing: The crystals were transferred into
crystallization buffer containing ethylene glycol (20%) as cryopro-
tectant, before mounting on the cryoloop. X-ray diffraction data
were collected with a Rigaku rotating anode generator and a MAR
Research image plate detector system. The data sets were indexed,
integrated and scaled with the aid of DENZO and SCALEPACK from
the HKL2000 suite of programs.[47] Table 2 summarizes the crystallo-
graphic parameters and the data collection statistics.
Keywords: active sites · enzyme mechanisms · isomerases ·
mutagenesis · proton transfer
[1] M. Karplus, J. Kuriyan, Proc. Natl. Acad. Sci. USA 2005, 102, 6679–6685.
[2] J. R. Knowles, Phil. Trans. R. Soc. Lond. B 1991, 332, 115–121.
[3] R. K. Wierenga, E. G. Kapetaniou, R. Venkatesan, Cell. Mol. Life Sci. 2010,
67, 3961–3982.
Structure solution and refinement: The structures of E97Q and
E97D were solved by use of the molecular replacement program
PHASER from the CCP4 package.[48] The native Pf TIM crystal struc-
ture (PDB ID: 1O5X) was used as the starting model. The coordi-
nates of 1O5X were modified by removal of the loop 6 residues,
ligand, water molecules and alternate conformations. Refinement
of both structures was carried out by use of REFMAC,[49] with 20 in-
itial cycles of rigid body refinement followed by 50 cycles of re-
strained refinement. Loop 6 residues, ligand and water molecules
were subsequently added into the electron density maps, j2FoꢀFc j
and jFoꢀFc j contoured at 1s and 3s respectively. The model build-
ing was done with use of COOT.[50] In both structures, two subunits
were observed in the asymmetric unit. The existence of the E97Q
and E97D mutations was confirmed from the difference Fourier
maps. Water molecules were first located automatically by COOT
and validated if a peak was observed above 3s on a difference
map and 1.5s on a double difference map. Both structures were
refined to reasonable Rwork and Rfree values. The geometries of the
final structures were examined with the aid of the PROCHECK pro-
gram[51] in the CCP4 package. The refinement statistics for the
structures are summarized in Table 3.
[4] S. C. Blacklow, R. T. Raines, W. A. Lim, P. D. Zamore, J. R. Knowles, Bio-
chemistry 1988, 27, 1158–1167.
[5] S. V. Rieder, I. A. Rose, J. Biol. Chem. 1959, 234, 1007–1010.
[6] W. J. Albery, J. R. Knowles, Biochemistry 1976, 15, 5627–5631.
[7] W. J. Albery, J. R. Knowles, Biochemistry 1976, 15, 5621–5626.
[8] W. J. Albery, J. R. Knowles, Biochemistry 1976, 15, 5631–5640.
[9] J. R. Knowles, Nature 1991, 350, 121–124.
[10] R. T. Raines, E. L. Sutton, D. R. Straus, W. Gilbert, J. R. Knowles, Biochem-
istry 1986, 25, 7142–7154.
[11] E. A. Komives, L. C. Chang, E. Lolis, R. F. Tilton, G. A. Petsko, J. R. Knowles,
Biochemistry 1991, 30, 3011–3019.
[12] P. J. Lodi, L. C. Chang, J. R. Knowles, E. A. Komives, Biochemistry 1994,
33, 2809–2814.
[13] M. K. Go, A. Koudelka, T. L. Amyes, J. P. Richard, Biochemistry 2010, 49,
5377–5389.
[14] M. Alahuhta, R. K. Wierenga, Proteins Struct. Funct. Bioinf. 2010, 78,
1878–1888.
[15] G. Jogl, S. Rozovsky, A. E. McDermott, L. Tong, Proc. Natl. Acad. Sci. USA
2003, 100, 50–55.
[16] L. Pauling, The Nature of the Chemical Bond, 3rd ed., Cornell University
Press, New York, 1960.
[17] T. L. Amyes, A. C. O’Donoghue, J. P. Richard, J. Am. Chem. Soc. 2001, 123,
11325–11326.
ChemBioChem 2011, 12, 1886 – 1896
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1895

P. Balaram et al.
[18] J. R. Morrow, T. L. Amyes, J. P. Richard, Acc. Chem. Res. 2008, 41, 539–
548.
[32] M. M. Malabanan, T. L. Amyes, J. P. Richard, Curr. Opin. Struct. Biol. 2010,
20, 702–710.
[19] P. A. Bash, M. J. Field, R. C. Davenport, G. A. Petsko, D. Ringe, M. Karplus,
Biochemistry 1991, 30, 5826–5832.
[20] M. K. Go, A. Koudelka, J. P. Richard, J. Am. Chem. Soc. 2010, 132, 13525–
13532.
[21] P. J. Lodi, J. R. Knowles, Biochemistry 1991, 30, 6948–6956.
[22] S. C. Blacklow, K. D. Liu, J. R. Knowles, Biochemistry 1991, 30, 8470–
8476.
[23] a) T. V. Borchert, R. Abagyan, R. Jaenicke, R. K. Wierenga, Proc. Natl.
Acad. Sci. USA 1994, 91, 1515–1518; b) T. V. Borchert, J. P. Zeelen, W.
Schliebs, M. Callens, W. Minke, R. Jaenicke, R. K. Wierenga, FEBS Lett.
1995, 367, 315–318; c) M. Salin, E. G. Kapetaniou, M. Vaismaa, M. Laju-
nen, M. G. Casteleijn, P. Neubauer, L. Salmon, R. K. Wierenga, Acta Crys-
tallogr. Sect. D Biol. Crystallogr. 2010, 66, 934–944.
[24] P. Pattanaik, G. Ravindra, C. Sengupta, K. Maithal, P. Balaram, H. Balaram,
Eur. J. Biochem. 2003, 270, 745–756.
[25] S. S. Velanker, S. S. Ray, R. S. Gokhale, S. Suma, H. Balaram, P. Balaram,
M. R. N. Murthy, Structure 1997, 5, 751–761.
[26] P. Gayathri, M. Banerjee, A. Vijayalakshmi, S. Azeez, H. Balaram, P. Balar-
am, M. R. N. Murthy, Acta Crystallogr. Sect. D Biol. Crystallogr. 2007, 63,
206–220.
[27] H. Walden, G. S. Bell, R. J. Russell, B. Siebers, R. Hensel, G. L. Taylor, J.
Mol. Biol. 2001, 306, 745–757.
[33] P. J. Lodi, J. R. Knowles, Biochemistry 1993, 32, 4338–4343.
[34] Q. Cui, M. Karplus, J. Am. Chem. Soc. 2002, 124, 3093–3124.
[35] M. Meot-Ner, Chem. Rev. 2005, 105, 213–284.
[36] E. L. Ash, J. L. Sudmeier, R. M. Day, M. Vincent, E. V. Torchilin, K. C.
Haddad, E. M. Bradshaw, D. G. Sanford, W. W. Bachovchin, Proc. Natl.
Acad. Sci. USA 2000, 97, 10371–10376.
[37] K. O. Christe, W. W. Wilson, G. J. Schrobilgen, R. V. Chirakal, G. A. Olah,
Inorg. Chem. 1988, 27, 789–790.
[38] J. P. Klinman, Chem. Phys. Lett. 2009, 471, 179–193.
[39] a) Z. D. Nagel, J. P. Klinman, Nat. Chem. Biol. 2009, 5, 543–550; b) S. D.
Schwartz, V. L. Schramm, Nat. Chem. Biol. 2009, 5, 551–558.
[40] J. Maynard Smith, Nature 1970, 225, 563–564.
[41] J. Ranie, V. P. Kumar, H. Balaram, Mol. Biochem. Parasitol. 1993, 61, 159–
169.
[42] A. Anderson, R. A. Cooper, J. Gen. Microbiol. 1970, 62, 329–334.
[43] A. R. Shenoy, S. S. Visweswariah, Anal. Biochem. 2003, 319, 335–336.
[44] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
[45] B. Plaut, J. R. Knowles, Biochem. J. 1972, 129, 311–320.
[46] B. L. Horecker, A. Kornberg, J. Biol. Chem. 1948, 175, 385–390.
[47] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307–326.
[48] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storo-
ni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658–674.
[49] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect. D Biol.
Crystallogr. 1997, 53, 240–255.
[28] S. C. Lovell, J. M. Word, J. S. Richardson, D. C. Richardson, Proc. Natl.
Acad. Sci. USA 1999, 96, 400–405.
[29] I. Rose, Autobiography, The Nobel Prize in Chemistry, http://www.nobel-
prize.org 2004.
[50] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D Biol. Crystallogr. 2004, 60,
2126–2132.
[30] a) E. Lolis, G. A. Petsko, Biochemistry 1990, 29, 6619–6625; b) R. C. Da-
venport, P. A. Bash, B. A. Seaton, M. Karplus, G. A. Petsko, D. Ringe, Bio-
chemistry 1991, 30, 5821–5826; c) Z. Zhang, S. Sugio, E. A. Komives,
K. D. Liu, J. R. Knowles, G. A. Petsko, D. Ringe, Biochemistry 1994, 33,
2830–2837.
[51] R. A. Laskowski, M. W. MacArthur, D. S. Moss, J. M. Thornton, J. Appl.
Crystallogr. 1993, 26, 283–291.
[31] a) N. S. Sampson, J. R. Knowles, Biochemistry 1992, 31, 8482–8487;
b) J. C. Williams, A. E. McDermott, Biochemistry 1995, 34, 8309–8319;
c) Y. Wang, R. B. Berlow, J. P. Loria, Biochemistry 2009, 48, 4548–4556.
Received: February 16, 2011
Published online on June 10, 2011
1896
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1886 – 1896