Identification and analysis of residues contained on β → α loops of the dual-substrate (βα)8 phosphoribosyl isomerase a specific for its phosphoribosyl anthranilate isomerase activity

Article abstract of DOI:10.1002/pro.331

A good model to experimentally explore evolutionary hypothesis related to enzyme function is the ancient-like dual-substrate (βα)₈ phosphoribosyl isomerase A (PriA), which takes part in both histidine and tryptophan biosynthesis in Streptomyces coelicolor and related organisms. In this study, we determined the Michaelis-Menten enzyme kinetics for both isomerase activities in wild-type PriA from S. coelicolor and in selected single-residue monofunctional mutants, identified after Escherichia coli in vivo complementation experiments. Structural and functional analyses of a hitherto unnoticed residue contained on the functionally important β → α loop 5, namely, Arg¹³⁹, which was postulated on structural grounds to be important for the dual-substrate specificity of PriA, is presented for the first time. Indeed, enzyme kinetics analyses done on the mutant variants PriA-Ser⁸¹Thr and PriA-Arg¹³⁹Asn showed that these residues, which are contained on β → α loops and in close proximity to the N-terminal phosphate-binding site, are essential solely for the phosphoribosyl anthranilate isomerase activity of PriA. Moreover, analysis of the X-ray crystallographic structure of PriA-Arg¹³⁹Asn elucidated at 1.95 A herein strongly implicates the occurrence of conformational changes in this β → α loop as a major structural feature related to the evolution of the dual-substrate specificity of PriA. It is suggested that PriA has evolved by tuning a fine energetic balance that allows the sufficient degree of structural flexibility needed for accommodating two topologically dissimilar substrates-within a bifunctional and thus highly constrained active site-without compromising its structural stability. Published by Wiley-Blackwell.

Full text of DOI:10.1002/pro.331

Identification and analysis of residues
contained on b fi a loops of the
dual-substrate (ba)8 phosphoribosyl
isomerase A specific for its phosphoribosyl
anthranilate isomerase activity
1
1,2
Lianet Noda-Garcıa, Aldo R. Camacho-Zarco, Karina Verdel-Aranda,1,2
´
3
2
3
´
¨
Helena Wright, Xavier Soberon, Vilmos Fu¨ lop, * and
1
´
Francisco Barona-Gomez *
1
´
Evolution of Metabolic Diversity Laboratory, Laboratorio Nacional de Genomica para la Biodiversidad (Langebio),
CINVESTAV-IPN, Km 9.6 Libramiento Norte, Carretera Irapuato—Leon, Irapuato, C.P. 36822, Mexico
´
´
2
´
´
Departamento de Ingenierı´a Celular y Biocatalisis, Instituto de Biotecnologıa, UNAM, Av. Universidad 2001, Cuernavaca,
´
C.P. 62210, Mexico
3Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, United Kingdom
Received 5 November 2009; Revised 25 December 2009; Accepted 28 December 2009
DOI: 10.1002/pro.331
Published online 11 January 2010 proteinscience.org
Abstract: A good model to experimentally explore evolutionary hypothesis related to enzyme
function is the ancient-like dual-substrate (ba)8 phosphoribosyl isomerase A (PriA), which takes
part in both histidine and tryptophan biosynthesis in Streptomyces coelicolor and related
organisms. In this study, we determined the Michaelis–Menten enzyme kinetics for both isomerase
activities in wild-type PriA from S. coelicolor and in selected single-residue monofunctional
mutants, identified after Escherichia coli in vivo complementation experiments. Structural and
functional analyses of a hitherto unnoticed residue contained on the functionally important
b fi a loop 5, namely, Arg139, which was postulated on structural grounds to be important for the
dual-substrate specificity of PriA, is presented for the first time. Indeed, enzyme kinetics analyses
done on the mutant variants PriA_Ser81Thr and PriA_Arg139Asn showed that these residues, which
are contained on b fi a loops and in close proximity to the N-terminal phosphate-binding site, are
essential solely for the phosphoribosyl anthranilate isomerase activity of PriA. Moreover, analysis
of the X-ray crystallographic structure of PriA_Arg139Asn elucidated at 1.95 A herein strongly
˚
implicates the occurrence of conformational changes in this b fi a loop as a major structural
feature related to the evolution of the dual-substrate specificity of PriA. It is suggested that PriA
has evolved by tuning a fine energetic balance that allows the sufficient degree of structural
flexibility needed for accommodating two topologically dissimilar substrates—within a bifunctional
and thus highly constrained active site—without compromising its structural stability.
Keywords: (ba)8-barrels; HisA and TrpF; phosphoribosyl isomerase A (PriA); dual-substrate
specificity; loops motion; conformational diversity
Grant sponsor: Conacyt, Mexico; Grant numbers: 50952-Q, 83039; Grant sponsor: Joint international project of the Royal Society,
United Kingdom.
´
´
*Correspondence to: Francisco Barona-Gomez, Evolution of Metabolic Diversity Laboratory, Laboratorio Nacional de Genomica para
´
´
la Biodiversidad (Langebio), CINVESTAV-IPN, Km 9.6 Libramiento Norte, Carretera Irapuato—Leon, Irapuato, C.P. 36822, Mexico.
E-mail: fbarona@ira.cinvestav.mx or Vilmos Fu¨ lo¨ p, Department of Biological Sciences, University of Warwick, Gibbet Hill Road,
Coventry CV4 7AL, United Kingdom. E-mail: vilmos@globin.bio.warwick.ac.uk
C
V
Published by Wiley-Blackwell. 2010 The Protein Society
PROTEIN SCIENCE 2010 VOL 19:535—543 535

Introduction
In recent years, enzyme promiscuity has ‘‘evolved’’
from being a biochemical curiosity to one of the
main themes in the field of enzyme evolution,1,2
with a broad impact on the biotechnological applica-
tions of enzymes.3,4 The broad occurrence of enzyme
promiscuity throughout all forms of cellular metabo-
lism is well acknowledged,5 being more frequent in
enzymes performing peripheral metabolic functions,
such as those involved in the biosynthesis of natural
products.6,7 Moreover, the feature of enzyme sub-
strate promiscuity has been suggested to endow
enzymes to be more evolvable,8 that is to reach a
phenotypic trait with the smallest number of possi-
ble changes. However, despite the potential role of
enzyme dual-substrate specificity on the evolution
of enzyme functions, contemporary ‘‘generalist’’
enzymes that may serve as good experimental mod-
els to test evolutionary hypothesis are virtually
nonexistent.
Figure 1. Enzymatic reactions of PriA, HisA, and TrpF.
(i) PriA, HisA, and TrpF catalyze Amadori rearrangements
(isomerizations) on analogous phosphoribosyl substrates.
PriA (R ¼ R1 or R2), HisA (R ¼ R1), or TrpF (R ¼ R2). The
products of ProFAR and PRA are N0-[(50-phosphoribulosyl)
formimino]-5-aminoimidazole-4-carboxamide ribonucleotide
(PRFAR) and 1-[(2-carboxyphenyl)amino]-1-deoxyribulose
5-phosphate (CdRP), respectively.
Such an enzyme model, which would allow ex-
perimental validation of evolutionary hypothesis,
exists in the ancient-like dual-substrate (ba)8 phos-
phoribosyl isomerase A (PriA), which takes part in
both histidine and tryptophan biosynthesis in Strep-
tomyces coelicolor and Mycobacterium tuberculosis.9
This study showed that PriA has both N0-(50-phos-
phoribosyl)anthranilate (PRA) isomerase (trpF, EC
5.3.1.24) and N0-[(50-phosphoribosyl)formimino]-5-
aminoimidazole-4-carboxamide ribonucleotide (Pro-
FAR) isomerase (hisA, EC 5.3.1.16) activities, under
physiologically relevant conditions. Remarkably, this
generalist contemporary enzyme is capable of accom-
modating two quite structurally dissimilar sub-
strates within a common active site (Fig. 1). Because
PriA is closely related to HisA at the sequence and
structural levels,10,11 understanding the molecular
mechanisms that allowed acquisition of PRA isomer-
ase activity in this (ba)8-barrel is a fundamental
question with broad evolutionary implications for
this protein fold.
˚
only 0.32 A (199 common Ca atoms), the distance
between the last residues of the functionally impor-
tant b ! a loops 1 and 6 that appear well defined in
the open conformer of PriA, that is, Gly23 (loop 1)
and Ala169 (loop 6), is of 4.83
respectively.11
˚
˚
A
and 4.11 A,
In addition to the catalytic role of Asp11 and
Asp130 common to the PRA and ProFAR isomerase
activities of PriA, key molecular interactions with
functional consequences, as probed by site-directed
mutagenesis and in vivo complementation experi-
ments using Escherichia coli trpF and hisA mutants,
have been reported.11 Among these, it was found
that Ser81, which interacts with a sulphate ion at
the N-terminal phosphate-binding site (PBS)
through a molecule of water, and Arg19, which
directly takes part in the C-terminal PBS, have im-
portant functional roles for the activities encoded by
the trpF and hisA genes in E. coli, respectively.
Moreover, a plausible molecular switch that may be
mediating the different PriA conformers, involving
an unusual CAHꢀꢀꢀO contact between Thr166 and
Asp171 was also identified. This interaction network
is completed by a hydrogen bond between Asp171
and His22, which brings about loops 1 and 6, closing
the active site. Mutation of any of these residues
was found to affect in vivo the activities of PriA.11
In this study, we determined the steady-state
Michaelis–Menten enzyme kinetics for both PRA
and ProFAR isomerase activities of wild-type PriA
from S. coelicolor and the mutants PriA_Arg19Ala,
PriA_Ser81Thr, and PriA_Arg139Asn. Particular em-
phasis was put on the latter two mutants because
mutation of Ser81 and Arg139 was found to affect the
PRA isomerase activity of this HisA-like enzyme.
Moreover, the X-ray crystallographic structure of
From a structural perspective, there is evidence
suggesting that PriA does exist in more than one
‘‘natural’’ conformation. Two independent wild-type
structures of PriA of S. coelicolor, adopting different
conformations, have been elucidated using X-ray
crystallography, under virtually identical conditions,
˚
and at the same atomic resolution of 1.8 A (PDBs:
1VZW10 and 2VEP11). Furthermore, the two putative
PriA conformers cooccur irrespective of the presence
of substrates because these structures have bound
sulphate ions, which can be taken as phosphate ana-
logs present in the substrates. Thus, 1VZW has been
suggested to adopt an open conformation, whereas
2VEP adopt a closed conformation. Interestingly,
when these conformers are structurally superim-
posed, despite the fact that the root mean square
deviation (RMSD) of the superimposed structures is
536 PROTEINSCIENCE.ORG
Role of Residues from Loops 3 and 5 on the PRA Isomerase Activity of PriA

Figure 2. Specific ProFAR isomerase activity of PriA measured at different pHs. Specific ProFAR isomerase activity (at 25ꢁC),
defined as the moles of ProFAR that are catalyzed (at saturating concentrations, which equals 24 lM) per minute per
milligram of purified PriA (U/min ¼ lM mgꢂ1 minꢂ1), was determined by triplicate. The Michaelis-Menten enzyme kinetic
(vo vs. [S]) shown in the inset (top-right corner) corresponds to ProFAR isomerase activity of PriA at a pH of 7.5, using an
enzyme concentration of 25 nM.
PriA_Arg139Asn elucidated at 1.95 A was obtained,
and this structure was used to explain the basis of
the functional role of Arg139. Based on these results,
it was suggested that PriA has evolved its dual-sub-
strate specificity by tuning a fine energetic balance,
which allows the sufficient degree of structural flexi-
bility needed for accommodating two topologically
(Asp11 and Asp130), and thus, the mechanism of reac-
tion can be assumed to be the same for both activ-
ities11; the kinetic parameters using ProFAR and
PRA as substrates were measured at this pH. Nota-
bly, our results disagree with the previously reported
ProFAR isomerase kinetic parameters of PriA in at
least one order of magnitude (Table I). This discrep-
ancy was found to stem from the fact that the previ-
ously reported parameters were obtained at pH
8.5,10 which is one unit above the optimum pH
experimentally determined herein. The differences
found for the kinetic parameters using PRA as sub-
strate (4-fold increase) may originate from a higher
experimental standard error in the previous deter-
mination with regards to our data.
˚
dissimilar substrates—within
a bifunctional and
thus highly constrained active site—without compro-
mising its structural stability.
Results
Steady-state enzyme kinetics of PriA and
selected mutants
Previous results obtained by in vivo complemen-
tation of a hisA minus E. coli mutant showed that
the mutant PriA_Arg19Ala lacks ProFAR isomerase
activity.11 This mutant was purified, and the steady-
state enzyme kinetics analyses for both activities
were performed. As expected, the level of PRA isom-
erase activity is similar to wild-type PriA; however,
Michaelis–Menten parameters were determined for
the isomerization of the substrates PRA and ProFAR
in wild-type PriA from S. coelicolor. The specific Pro-
FAR isomerase activity of PriA was measured at dif-
ferent pHs, and the optimum pH was found to be 7.5
(Fig. 2). Because both isomerase activities of PriA
are performed by the same set of catalytic residues
Table I. Michaelis–Menten Kinetic Parameters of PriA and Selected Mutants
ProFAR isomerase activity
PRA isomerase activity
Enzymea
PriAb
KM (lM)
3.6 6 0.7
28
4.1 6 0.9
3.9 6 0.8
3.6 6 0.5
kcat (sꢂ1
1.3 6 0.2
0.9
0.24 6 0.5
0.33 6 0.08
0.24 6 0.01
)
kcat/KM (lMꢂ1 sꢂ1
)
KM (lM)
5 6 0.08
4
8.3 6 1.7
Ild
kcat (sꢂ1
3.4 6 0.09
12
1.4 6 0.2
)
kcat/KM (lMꢂ1 sꢂ1
)
0.4
0.7
3b
0.2
0.0037c
Il
0.03b
0.06
0.08
0.07
PriA_Arg19Ala
PriA_Ser81Thrc
PriA_Arg139Asn
Il
Il
Il
a
Each datapoint is composed by at least three independent determinations using freshly purified enzyme, which was used
to determine both activities simultaneously.
b
Data from the second row in PriA corresponds to the previously reported kinetic parameters for this enzyme.10
Despite the PRA isomerase activity of this mutant could be measured, its kinetic parameters could not be obtained (refer
c
to text and Fig. 3).
d
Il, immeasurably low.
ꢀ
Noda-Garcıa et al.
PROTEIN SCIENCE VOL 19:535—543 537

Figure 3. PRA isomerase activity of PriA and selected mutants. (A) Specific PRA isomerase activities of wild-type and
selected mutants, using 1.5 lM of PriA_Arg139Asn (green), 0.4 lM of PriA_Ser81Thr (blue), and 0.04 lM of PriA. (B) Michaelis-
Menten enzyme kinetics (vo vs. [S]) of PriA (red, lower panel) and selected mutants (upper panel), showing that PriA_Ser81Thr
(blue) could no be saturated, and that PriA_Arg139Asn (green) lacks all PRA isomerase activity. [Color figure can be viewed in
the online issue, which is available at www.interscience.wiley.com.]
despite the fact that expression in trans of the mu-
tant PriA_Arg19Ala fails to rescue the histidine
auxotrophy of the E. coli hisA minus mutant Hfr
G611, its catalytic ProFAR isomerase efficiency was
found to be only five times less than wild-type PriA
(Table I).
of these residues. Similarly to mutation of Ser81 into
a Thr residue, it was hypothesized that by changing
Arg139 into a closely related Asn, the functional
analysis of the original residue might be facilitated.
A complete lack of PRA isomerase activity in this
mutant was confirmed in vivo and in vitro, using
enzyme saturation conditions (Fig. 3 and Table I).
Interestingly, the ProFAR isomerase activity of this
mutant was not significantly affected: the KM of
PriA_Arg139Asn for the substrate ProFAR remained
the same, whereas its kcat is only five times smaller.
Because both isomerase activities of PriA are accom-
plished by the same active site,11 this result sug-
gests that Arg139 does not perform a catalytic role
and that it is not related to binding of the substrate
ProFAR.
Based on similar in vivo complementation
experiments, but using the trpF minus E. coli mu-
tant FBG-Wf, the residue Ser81, which interacts
through a molecule of water with the sulphate ion
bound at the N-terminal PBS in the wild-type struc-
ture of PriA, was specifically implicated on its PRA
isomerase activity.11 In agreement with this, the mu-
tant PriA_Ser81Thr was purified, and the determina-
tion of its steady-state enzyme kinetics showed that
its ProFAR isomerase activity is similar to wild-type
PriA, whilst its PRA isomerase catalytic efficiency is
decreased by the mutation at least 200-fold. Given
that the enzyme assay used for determining PRA
isomerase activity follows disappearance of the sub-
strate produced in situ, using more than 200 lM of
anthranilic acid interferes with the fluorometric
detection of PRA, rendering it impossible to saturate
the enzyme. However, although the KM and kcat ki-
netic parameters for conversion of PRA could not be
obtained, the catalytic efficiency was calculated from
the lineal phase of the Michaelis–Menten equation
[Fig. 3(B) and Table I].
Structural analysis of PRA isomerase minus
PriA mutants
To understand the functional contribution of Ser81
and Arg139 toward the PRA isomerase activity of
PriA, crystallographic X-ray structural analyses of
their corresponding mutants were pursued. Unfortu-
nately, the mutant PriA_Ser81Thr could not be crys-
tallized despite several attempts. In contrast, the
structure of PriA_Arg139Asn (PDB: 2x30), mutated
at the functionally important b ! a loop 5, absent
from 1VZW, could be resolved. This protein was
purified and crystallized using similar conditions as
for wild-type PriA.13 The structure of the PriA_Ar-
g139Asn mutant was obtained at a resolution of 1.95
In addition to residue Ser81, the sulphate ion
bound at the N-terminal PBS of PriA interacts with
residues Gly83, Arg85, Gly104, Thr105, and Arg139
.
Interestingly, all these residues are located on the
‘‘hinges" of b ! a loops, which have been shown to
be structurally and functionally important within
the (ba)8-barrel fold,12 warranting further analysis
˚
A (Table II) and aligned with the structure of wild-
˚
type PriA, with a final RMSD of 0.24 A (213 common
Ca atoms). This allowed
a comparison to be
538 PROTEINSCIENCE.ORG
Role of Residues from Loops 3 and 5 on the PRA Isomerase Activity of PriA

Table 2. X-Ray Data Collection and Refinement
Statistics
Although the structures of wild-type PriA and
the mutant PriA_Arg139Asn are overall quite similar,
their loops 1, 5, and 6 were found to adopt signifi-
cantly different conformations (Fig. 4). Loops 1 and
6 seem to be more flexible in PriA_Arg139Asn than
in wild-type structure, and a marked rigidity of loop
5 was only witnessed in the structure of the mutant.
This observation was quantitatively supported by
comparing the ratio of the B-factors of the residues
contained on this loop over the whole structure. On
average, the residues in loop 5 have 1.9 times higher
B-factors than the corresponding values in the
PriA_Arg139Asn structure. This observation suggests
that loop 5 of PriA is particularly flexible and that
minor mutations can affect its dynamics, which in
turn can have large functional consequences. Indeed,
because this PriA mutant lacks its PRA isomerase
activity, but not its ProFAR isomerase activity, it is
tempting to speculate that Arg139 is involved in
mediating conformational changes, subsequently
allowing the active site to adopt the right conforma-
tion to bind and/or turnover PRA productively.
PriA_Arg139Asn (PDB: 2x30)
Data collection
Synchrotron, detector and
ESRF BM16 ADSC
Q210 CCD, 1.0074
P3121
63.6, 102.9
28–1.95 (2.02–1.95)
182830
˚
wavelength (A)
Space group
˚
Unit cell (a¼ b, c (A))
˚
Resolution (A)
Observations
Unique reflections
I/r(I)
18061
23.0 (2.4)
0.093 (0.652)
99.8 (100.0)
a
Rsym
Completeness (%)
Refinement
All nonhydrogen atoms
Water molecules
1846
126
Other solvent molecules
Rcryst
2 (sulfate)
0.216 (0.359)
17365 (1269)
0.273 (0.460)
696 (39)
0.218
b
Reflections used
c
Rfree
Reflections used
Rcryst (all data)b
2
˚
Mean temperature factor (A )
RMSDS from ideal values
39.2
˚
Bonds (A)
0.016
1.4
0.18
Detailed analysis of the molecular environment
of Arg139 turned out to be informative in this
respect. Despite the physicochemical differences
between Asn and Arg, that is, the lack of a carbon
atom and the presence of an amine (containing one
N atom), instead of a guanidinium (containing three
N atoms), these two residues are able to interact
with the sulphate ion, fortuitously bound at the N-
terminal PBS in both structures (Fig. 5). Interest-
ingly, the residue taking position 139 in either
structure adopts the same conformation, with their
lateral chains pointing in the same direction toward
the sulphate ion. However, when an Arg is located
at position 139, two of the nitrogen atoms of the
guanidinium group are able to form salt bridges
with the sulphate ion, in contrast with only one con-
tact that is found when Asn is taking this position.
Angles (ꢁ)
DPI coordinate error (A)
˚
Numbers in parentheses refer to values in the highest
resolution shell.
P P
P P
a
Rsym
¼
h|Ih,j ꢂ <Ih>|/
h<Ih>, where Ih,j is the
j
j
jth observation of reflection h, and <Ih> is the mean inten-
sity of that reflection.
b
P
P
Rcryst
¼
||Fobs| ꢂ |Fcalc||/ |Fobs|, where Fobs and
Fcalc are the observed and calculated structure factor
amplitudes, respectively.
c
Rfree is equivalent to Rcryst for a randomly selected subset
of reflections not used in the refinement.
performed between wild-type PriA (PDB: 2VEP,
chain A) and the mutant PriA_Arg139Asn (Fig. 4) in
terms of their temperature factors (B-factor), which
are known to correlate with protein dynamics.
Figure 4. Structural differences between PriA and PriA_Arg139Asn. (A) Wild-type PriA (PDB: 2VEP); and (B) PriA_Arg139Asn
mutant (this work, PDB: 2x30). The structures were colored from blue to red as the B-factors increase. Loop 5 is highlighted
with a black arrow. (C) Structural superimposition of wild-type PriA (green) and PriA_Arg139Asn (red). The sulphate ions bound
to the two PBS, as well as residues Arg/Asn139, Asp11, and Asp130, delimiting the boundaries of the active site, are also
shown.
ꢀ
Noda-Garcıa et al.
PROTEIN SCIENCE VOL 19:535—543 539

Figure 5. Molecular interactions of the residue adopting position 139. (A) Interactions of Arg in wild-type PriA; the electron
density of this structure has been reported previously11; and (B) Interactions of Asn in PriA_Arg139Asn. Hydrogen bonds are
drawn as thin lines. The SIGMAA14 weighted 2mFo-DFc electron density using phases from the final model is contoured at
˚
1r level, where r represents the root mean square deviation electron density for the unit cell. Contours more than 1.4 A from
any of the displayed atoms, other than Ser81, have been removed for clarity. Figure was drawn using Molscript15,16 and
rendered with Raster 3D.17 [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
This is the only change observed in the binding of
the N-terminal sulphate ion when PriA and
PriA_Arg139Asn are compared.
In the cases where the enzyme assays confirmed
that the inability to complement a trpF mutation
was related to a lack of PRA isomerase activity, that
is, PriA_Ser81Thr and PriA_Arg139Asn, it was partic-
ularly encouraging to find that these mutations do
not affect the mechanism of reaction, as suggested
by the fact that ProFAR could still be converted
quite efficiently. Indeed, functional analysis of these
ProFAR isomerase monofunctional PriA mutants
does bring about novel insights into the evolution of
the PRA isomerase activity in this HisA-like (ba)8-
barrel. An interesting observation is that their KM
and kcat kinetic parameters for conversion of ProFAR
imply that these residues, which were presumed to
be part of the N-terminal PBS as suggested by their
closeness with the sulphate ion, must be unrelated
to the binding of the phosphate moieties of the sub-
strates. Moreover, because it is safe to assume that
these residues do not perform a catalytic role, and
that ProFAR and PRA most likely are bound by
their common phosphate moieties, the question of
how these mutations have abolished the ability to
convert PRA is a complex one, as discussed later.
The kinetic parameters obtained for the ProFAR
isomerase activity are consistent with two possibil-
ities: (i) that mutation of Ser81 and Arg139, which
are contained on b ! a flexible loops, may have
affected the ability of PriA to adopt the natural con-
formation of the active site, compromising com-
pletely turnover of PRA, as well as that of ProFAR
5-fold; and (ii) that PriA uses residues Ser81 and
Arg139 for binding of PRA, but not of ProFAR, as
suggested by the unaffected KM parameter obtained
for the latter substrate. Furthermore, although the
kcat and KM kinetic parameters for PriA_Ser81Thr,
with regards to PRA, could not be obtained, our
Discussion
Previous characterization of HisA enzymes from Ther-
motoga maritima and E. coli, including single-residue
mutants, implied
a
correlation between results
obtained in vivo and in vitro.18 Based on these obser-
vations, an in vivo genetic screening for identifying
key residues involved in the evolution of the dual-sub-
strate specificity of PriA was adopted. Indeed, previ-
ous in vivo results implicated residues Arg19 and
Ser81 in the specific conversions of ProFAR and PRA,
respectively.11 Although this approach has simplified
the analysis of a large collection of mutants, the fact
that the enzyme kinetic parameters obtained for
PriA_Arg19Ala does not agree with the inability of
this mutant to complement a hisA mutation may raise
some doubts about the validity of this approach. False
negatives due to technical problems can be ruled out
because many mutants were simultaneously analyzed,
rendering reproducible results. An alternative expla-
nation to this discrepancy may have to do with the
recent appreciation that functional performance of
enzymes in vivo is not limited to catalytic proficiency,
as measured in vitro.19 Such ‘‘in vivo characteristics,’’
necessarily encoded by the enzyme’s primary
sequence, may relate to protein stability, protein
expression, and protein-protein interactions, amongst
other unknown context-dependent factors. Irrespective
of the nature of this discrepancy, the difficulty in iden-
tifying residues specifically related to the ProFAR
isomerase activity may be taken as a reflection of the
ancestral origin of this activity in PriA, which is con-
served in all HisA bacterial homologs.
540 PROTEINSCIENCE.ORG
Role of Residues from Loops 3 and 5 on the PRA Isomerase Activity of PriA

ability to detect PRA isomerase activity in vitro is
consistent with the idea that this residue is nor
involved in a discreet catalytic step or binding of the
phosphate moiety. Given that PRA and ProFAR have
in common a phosphoribosyl moiety, by which these
substrates must be primarily bound by PriA, but
that they differ in their aromatic anomeric group, it
could be that these residues are specifically involved
in recognition of the anthranilate moiety of PRA,
with a concomitant negative effect on the turnover
of this substrate. This would also suggest that the
C-terminal PBS, as opposed to our previous proposal
based solely in in vivo complementation assays, does
bind PRA.11 Interestingly, multisequence alignments
of HisA and PriA enzymes suggest so far that Ser81
and Arg139 are specific to PriA (F. B. -G. and L. N. -G.,
unpublished observations).
The structural analysis of PriA_Arg139Asn is in
agreement with the conclusion that Ser81 and Arg139
are involved in the specific recognition, and subse-
quent turnover, of PRA. In accordance with Jensen’s
hypothesis on the evolution of metabolic pathways
through enzyme recruitment,20 the most likely solu-
tion for the evolution of the PRA isomerase activity
in the scaffold of PriA would imply the acquisition of
changes related to the binding capabilities. However,
because both PRA and ProFAR have analogous phos-
phates, subtle changes not related to the PBS would
have to occur. As demonstrated by the structural
analysis, the N-terminal PBS of wild-type PriA and
PriA_Arg139Asn are identical. In contrast, loss of
PRA isomerase activity in PriA_Arg139Asn is con-
comitant with an increased rigidity of the b ! a
loop 5. In agreement with this, Ser81 is located rela-
tively far from the sulphate ion, but within the func-
tionally important b ! a loop 3. These observations
suggest that modification of the dynamics of b ! a
loops, which are generally accepted to be important
for this protein fold,12,21 may have a major role on
the evolution of the dual-substrate specificity of
PriA. In agreement with this, loops have been impli-
cated in the gain-and-loss of enzymatic activities in
other protein folds.22,23
bility, and thus further investigation of the motion of
the active site loops of this (ba)8-isomerase, by means
of X-ray crystallography of complexes between enzyme
and transition-state analogues or noncrystallographic
methods compatible with analysis in solution,21 are
very promising. The hypothesis along these lines
would be that PriA has evolved its PRA isomerase ac-
tivity by tuning a fine energetic balance that allows
the sufficient degree of structural flexibility needed
for accommodating two topologically dissimilar sub-
strates—within a bifunctional and thus highly con-
strained active site – without compromising its struc-
tural stability.
Materials and Methods
Site-directed mutagenesis
The mutant PriA_Arg139Asn was constructed using
the megaprimer method26 using as template the
plasmid pGEX-PriA-Sco.9 The sequences of the mu-
tagenic oligonucleotides were as follows: R139For,
50ccgcggc aat ggctggacccgcgacggcggcg and R139Rev,
50cgccgccgtcgcgggtccagcc att gccgcgg. All constructs
were sequenced before functional analyses.
Biochemical characterization of PriA
and its mutants
In vivo complementation of hisA and trpF minus
E. coli mutants was performed as described previ-
ously.11 Expression and purification of wild-type
PriA and its mutants was performed from the
expression vector pET-15b (Stratagene) as described
previously.13 Michaelis-Menten enzyme kinetics of
ProFAR and PRA isomerase activities were deter-
mined using reported protocols14,27 with minor modi-
fications, as discussed later. The kinetic parameters
were obtained fitting initial rates to the Michaelis–
Menten model using nonlinear fit analysis with the
publicly available program MicroCal Origin 5.0.
Each datapoint included in Table I represents at
least three independent experiments using freshly
purified enzyme, which was simultaneously assayed
for both activities.
The proposed protein conformational diversity of
PriA, which implies cooccurrence of different conform-
ers of the same protein in solution—independent of
an induced fit mechanism—could be related to the
evolution of the dual substrate specificity of PriA. A
link between the evolution of novel enzyme functions,
starting from a promiscuous enzyme activity as an ev-
olutionary raw material, and protein conformational
diversity, has been previously postulated.24,25 This
proposal stems from studies on antibodies, where it
has been shown that these can indeed exist as a mix-
ture of conformers, capable of promiscuously binding
‘‘unnatural" antigens.24 The appearance of PriA from
ProFAR was synthesized, purified, and quanti-
fied according to the method described previously,28
using phisGIE-tac and pBS111R plasmids, kindly
provided by V. Jo Davisson and Rebecca S. Myers.
The HisF subunit of the imidazol glycerol phosphate
synthase from T. maritima (thisF) was subcloned
from plasmid pET11c-thisF, kindly provided by Prof.
Reinhard Sterner and coworkers,29 into the plasmid
pET-15b (Novagen) by using the restriction sites
NdeI and BamHI. Plasmid pET15b-thisF was used
to purify this enzyme by Nickel affinity columns
¨
HisTrap HP (GE Healthcare) attached to an Akta
FPLC from Amersham Biosciences. ProFAR isomer-
ase reactions were performed at 25ꢁC and followed
spectrophotometrically at 300 nm, in the presence of
a
HisA-like monofunctional promiscuous enzyme
might be seen as an interesting example of this possi-
ꢀ
Noda-Garcıa et al.
PROTEIN SCIENCE VOL 19:535—543 541

50 mM TES-HCl buffer (pH 7.5), 200 mM of
ammonium acetate, in a Varian Cary 100 BIO UV-
visible spectrophotometer. An excess of purified
tHisF (2 lM final concentration) was added to the
reaction mix to avoid product inhibition.18
beamline BM16 at the ESRF, France, and we acknowl-
edge the support of Ana Labrador.
References
For the PRA isomerase enzyme assay, the gene
encoding for indol 3-glycerol phosphate synthase from
T. maritima (tIGPS) was subcloned from plasmid pET
tmtrpC 21a30 (kindly provided by R. Sterner) into a
modified version of pQE30 (Qiagen), termed pQEI
(F.B.-G. and L.N.-G., unpublished results), using the
restriction sites NdeI and HindIII to generate plasmid
pQEItmIGPS. Anthranilate phosphoribosyl transfer-
ase from Saccharomyces cerevisiae (yAnPRT)
expressed from plasmid pETyAnPRT 28a (also pro-
vided by R. Sterner), and tIGPS expressed from
pQEItmIGPS, were purified by Nickel affinity using
the same Nickel affinity—FPLC system. Initial rates
for PRA isomerase activity were measured at 25ꢁC
using enzymatically produced PRA in 50 mM Tris-
HCl (pH 7.5), 4 mM MgCl2, 4 mM EDTA, 0.4 lM
DTT, 0-200 lM anthranilic acid (Sigma), a 5-fold
molar excess of PRPP (Sigma), and 1 lM of yAnPRT.
Both reactions (PRA production and its subsequent
conversion) were followed fluorometrically in 96-well
plates (Nuc 96 Well Optical Bottom Plates) in a
TECAN infinite M1000 plate reader (excitation at 310
nm and emission at 400 nm). tIGPS was added to
avoid product inhibition.
1. Mannervik B, Runarsdottir A, Kurtovic S (2009) Multi-
substrate-activity space and quasi-species in enzyme
evolution: Ohno’s dilemma, promiscuity and functional
orthogonality. Biochem Soc Trans 37:740–744.
2. Tracewell CA, Arnold FH (2009) Directed enzyme evo-
lution: climbing fitness peaks one amino acid at a time.
Curr Opin Chem Biol 13:3–9.
3. Hult K, Berglund P (2007) Enzyme promiscuity: mech-
anism and applications. Trends Biotechnol 25:231–238.
4. Nobeli I, Favia AD, Thornton JM (2009) Protein prom-
iscuity and its implications for biotechnology. Nat Bio-
technol 27:157–167.
5. Khersonsky O, Roodveldt C, Tawfik DS (2006) Enzyme
promiscuity: evolutionary and mechanistic aspects.
Curr Opin Chem Biol 10:498–508.
6. O’Maille PE, Malone A, Dellas N, Andes Hess B, Jr,
Smentek L, Sheehan I, Greenhagen BT, Chappell J,
Manning G, Noel JP (2008) Quantitative exploration of
the catalytic landscape separating divergent plant ses-
quiterpene synthases. Nat Chem Biol 4:617–623.
7. Des Marais DL, Rausher MD (2008) Escape from
adaptive conflict after duplication in an anthocyanin
pathway gene. Nature 454:762–765.
8. Aharoni A, Gaidukov L, Khersonsky O, Mc QGS, Rood-
veldt C, Tawfik DS (2005) The ‘evolvability‘ of promis-
cuous protein functions. Nat Genet 37:73–76.
9. Barona-Gomez F, Hodgson DA (2003) Occurrence of a
putative ancient-like isomerase involved in histidine
and tryptophan biosynthesis. EMBO Rep 4:296–300.
10. Kuper J, Doenges C, Wilmanns M (2005) Two-fold
repeated (betaalpha)4 half-barrels may provide a mo-
lecular tool for dual substrate specificity. EMBO Rep 6:
134–139.
Crystallization, X-ray data collection, structure
determination, and refinement
The PriA_Arg139Asn mutant was expressed and
purified as reported previously for wild-type PriA.
Crystals were obtained after 48 hr of incubation in
the same condition as reported previously.13 X-ray
data were processed using the HKL suite of pro-
grams.31 Refinement of the structure was carried
out using the coordinates of 2VEP by alternate
cycles of REFMAC32 and manual refitting using O.33
Water molecules were added to the atomic model
automatically using ARP34 at the positions of large
positive peaks in the difference electron density,
only at places where the resulting water molecule
fell into an appropriate hydrogen bonding environ-
ment. Restrained isotropic temperature factor refine-
ments were carried out for each individual atom.
Data collection and refinement statistics are given
in Table II.
11. Wright H, Noda-Garcia L, Ochoa-Leyva A, Hodgson
DA, Fulop V, Barona-Gomez F (2008) The structure/
function relationship of a dual-substrate (betaalpha)8-
isomerase. Biochem Biophys Res Commun 365:16–21.
12. Ochoa-Leyva A, Soberon X, Sanchez F, Arguello M,
Montero-Moran G, Saab-Rincon
G (2009) Protein
design through systematic catalytic loop exchange in
the (beta/alpha)8 fold. J Mol Biol 387:949–964.
13. Wright H, Barona-Gomez F, Hodgson DA, Fulop V
(2004) Expression, purification and preliminary crystal-
lographic analysis of phosphoribosyl isomerase (PriA)
from Streptomyces coelicolor. Acta Cryst D60:534–536.
14. Read RJ (1986) Improved Fourier coefficients for maps
using phases from partial structures with errors. Acta
Cryst A42:140–149.
15. Kraulis PJ (1991) Molscript—a program to produce
both detailed and schematic plots of protein structures.
J Appl Cryst 24:946–950.
16. Esnouf RM (1997) An extensively modified version of
MolScript that includes greatly enhanced coloring
capabilities. J Mol Graphics Model 15:132.
17. Merritt EA, Murphy MEP (1994) Raster3d Version-
2.0—a program for photorealistic molecular graphics.
Acta Cryst D50:869–873.
Acknowledgments
We are indebted to Nobuhiko Tokuriki for useful dis-
cussions and critical reading of the manuscript, and to
the reviewers for their useful suggestions to improve
this manuscript. We thank V. Jo Davisson, Rebecca S.
Myers, and Reinhard Sterner for providing plasmids
and strains.Crystallographic data were collected at
18. Henn-Sax M, Thoma R, Schmidt S, Hennig M, Kirsch-
ner K, Sterner
R (2002) Two (betaalpha)(8)-barrel
enzymes of histidine and tryptophan biosynthesis have
similar reaction mechanisms and common strategies
for protecting their labile substrates. Biochemistry 41:
12032–12042.
542 PROTEINSCIENCE.ORG
Role of Residues from Loops 3 and 5 on the PRA Isomerase Activity of PriA

19. Yoshikuni Y, Dietrich JA, Nowroozi FF, Babbitt PC,
Keasling JD (2008) Redesigning enzymes based on
adaptive evolution for optimal function in synthetic
metabolic pathways. Chem Biol 15:607–618.
28. Davisson VJ, Deras IL, Hamilton SE, Moore LL (1994)
A plasmid-based approach for the synthesis of a histi-
dine biosynthetic intermediate.
137–143.
J Org Chem 59:
20. Jensen RA (1976) Enzyme recruitment in evolution of
new function. Annu Rev Microbiol 30:409–425.
21. Wang Y, Berlow RB, Loria JP (2009) Role of loop-loop
interactions in coordinating motions and enzymatic
function in triosephosphate isomerase. Biochemistry
48:4548–4556.
22. Park HS, Nam SH, Lee JK, Yoon CN, Mannervik B,
Benkovic SJ, Kim HS (2006) Design and evolution of
new catalytic activity with an existing protein scaffold.
Science 311:535–538.
29. Thoma R, Obmolova G, Lang DA, Schwander M, Jeno
P, Sterner R, Wilmanns M (1999) Efficient expression,
purification and crystallisation of two hyperthermosta-
ble enzymes of histidine biosynthesis. FEBS Lett 454:
1–6.
30. Merz A, Knochel T, Jansonius JN, Kirschner K (1999)
The hyperthermostable indoleglycerol phosphate syn-
thase from Thermotoga maritima is destabilized by
mutational disruption of two solvent-exposed salt
bridges. J Mol Biol 288:753–763.
23. Doucet N, Watt ED, Loria JP (2009) The flexibility of a
distant loop modulates active site motion and product
release in ribonuclease A. Biochemistry 48:7160–7168.
24. James LC, Tawfik DS (2003) Conformational diversity
and protein evolution--a 60-year-old hypothesis revis-
ited. Trends Biochem Sci 28:361–368.
31. Otwinowski Z, Minor W (1997) Processing of X-ray dif-
fraction data collected in oscillation mode. Macromol
Cryst A 276:307–326.
32. Murshudov GN, Vagin AA, Dodson EJ (1997) Refine-
ment of macromolecular structures by the maximum-
likelihood method. Acta Cryst D 53:240– 255.
25. Tokuriki N, Tawfik DS (2009) Protein dynamism and
evolvability. Science 324:203–207.
26. Xu Z, Colosimo A, Gruenert DC (2003) Site-directed
mutagenesis using the megaprimer method. Methods
Mol Biol 235:203–207.
33. Jones TA, Zou JY, Cowan SW, Kjeldgaard M (1991)
Improved methods for building protein models in elec-
tron-density maps and the location of errors in these
models. Acta Cryst A 47:110–119.
34. Perrakis A, Sixma TK, Wilson KS, Lamzin VS (1997)
wARP: Improvement and extension of crystallographic
phases by weighted averaging of multiple-refined
dummy atomic models. Acta Cryst D 53:448–455.
27. Hommel U, Eberhard M, Kirschner K (1995) Phosphor-
ibosyl anthranilate isomerase catalyzes
a reversible
amadori reaction. Biochemistry 34:5429–5439.
ꢀ
Noda-Garcıa et al.
PROTEIN SCIENCE VOL 19:535—543 543