The crystal structure of galactitol-1-phosphate 5-dehydrogenase from Escherichia coli K12 provides insights into its anomalous behavior on IMAC processes

Article abstract of DOI:10.1016/j.febslet.2012.07.073

Endogenous galactitol-1-phosphate 5-dehydrogenase (GPDH) (EC 1.1.1.251) from Escherichia coli spontaneously interacts with Ni²⁺-NTA matrices becoming a potential contaminant for recombinant, target His-tagged proteins. Purified recombinant, untagged GPDH (rGPDH) converted galactitol into tagatose, and d-tagatose-6-phosphate into galactitol-1-phosphate, in a Zn²⁺- and NAD(H)-dependent manner and readily crystallized what has permitted to solve its crystal structure. In contrast, N-terminally His-tagged GPDH was marginally stable and readily aggregated. The structure of rGPDH revealed metal-binding sites characteristic from the medium-chain dehydrogenase/reductase protein superfamily which may explain its ability to interact with immobilized metals. The structure also provides clues on the harmful effects of the N-terminal His-tag. Structured summary of protein interactions: GPDH and GPDH bind by molecular sieving (View interaction) GPDH and GPDH bind by x-ray crystallography (View interaction) GPDH and GPDH bind by cosedimentation in solution (View interaction).

Full text of DOI:10.1016/j.febslet.2012.07.073

FEBS Letters 586 (2012) 3127–3133
journal homepage: www.FEBSLetters.org
The crystal structure of galactitol-1-phosphate 5-dehydrogenase
from Escherichia coli K12 provides insights into its anomalous behavior on IMAC
processes
María Esteban-Torres a, Yanaisis Álvarez b,c, Iván Acebrón b, Blanca de las Rivas a, Rosario Muñoz a,
Gert-Wieland Kohring d, Ana María Roa e, Mónica Sobrino e, José M. Mancheño b,
⇑
a Laboratorio de Biotecnología Bacteriana, Instituto Ciencia y Tecnología de Alimentos y Nutrición (ICTAN), CSIC, Juan de la Cierva 3, 28006 Madrid, Spain
b Grupo de Cristalografía y Biología Estructural, Instituto de Química Física Rocasolano, CSIC, Serrano 119, 28006 Madrid, Spain
c Centro de Estudios Avanzados de Cuba, CITMA, Carretera de S. Antonio km 2, 17100 Habana, Cuba
d Microbiology, Saarland University, Campus, Geb. A1.5, D-66123 Saarbruecken, Germany
e Screening & Compound Profiling, CIB, GlaxoSmithKline, Santiago Grisolía 4, 28760 Tres Cantos, Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Endogenous galactitol-1-phosphate 5-dehydrogenase (GPDH) (EC 1.1.1.251) from Escherichia coli
spontaneously interacts with Ni2+-NTA matrices becoming a potential contaminant for recombinant,
target His-tagged proteins. Purified recombinant, untagged GPDH (rGPDH) converted galactitol into
Received 16 May 2012
Revised 16 July 2012
Accepted 27 July 2012
Available online 8 August 2012
tagatose, and D
-tagatose-6-phosphate into galactitol-1-phosphate, in a Zn2+- and NAD(H)-dependent
manner and readily crystallized what has permitted to solve its crystal structure. In contrast,
N-terminally His-tagged GPDH was marginally stable and readily aggregated. The structure of
rGPDH revealed metal-binding sites characteristic from the medium-chain dehydrogenase/reductase
protein superfamily which may explain its ability to interact with immobilized metals. The structure
also provides clues on the harmful effects of the N-terminal His-tag.
Edited by Miguel De la Rosa
Keywords:
Crystal structure
Dehydrogenase
Structured summary of protein interactions:
D-Tagatose-6-phosphate
Galactitol
Metal binding-site
Zn-metalloenzyme
GPDH and GPDH bind by molecular sieving (View interaction)
GPDH and GPDH bind by x-ray crystallography (View interaction)
GPDH and GPDH bind by cosedimentation in solution (View interaction)
Ó 2012 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction
groups: (i) proteins with metal-binding sites, (ii) proteins with ex-
posed histidine clusters, (iii) proteins that bind to the overexpres-
Currently immobilised metal affinity chromatography (IMAC) is
one of the most widely used techniques for the purification of re-
combinant proteins [1,2] due to the standardization of methods
for the efficient production of His-tagged proteins [3] and to the
fact that the presence of His-tags in recombinant proteins is not
normally harmful to their structures [4]. Despite in most cases
IMAC leads to high protein purity upon single-step chromato-
graphic purification [1], it is now recognized that other proteins
can be co-purified with the target, His-tagged protein. In Esche-
richia coli these ‘‘contaminants’’ have been classified in four
sed His-tagged protein, and (iv) proteins with affinity for agarose-
based matrixes [1,5]. Notably, the presence of these accompanying
proteins is not predictable nor universal since many of them are
stress-responsive proteins what suggest that experimental vari-
ables such as culture conditions, media composition and also the
particular bacterial strain, may have important effects on their
production.
We show that endogenous GPDH and rGPDH from Escherichia
coli spontaneously interact with NTA matrices what has permitted
its efficient purification and further characterization. The enzyme
catalyzed the oxidation of galactitol producing the rare sugar
tagatose, and the reduction of
D-tagatose-6-phosphate, which
Abbreviations: GPDH, endogenous galactitol-1-phosphate 5-dehydrogenase;
rGPDH, recombinant, untagged GPDH; IMAC, immobilized metal affinity chroma-
tography; NTA, nitrilotriacetic acid
rendered galactitol-1-phosphate. Pure rGPDH could be crystallized
and its structure solved in contrast to the recombinant,
N-terminally His6-tagged protein (His6-GPDH), which showed
marginal stability in solution and was therefore reluctant to
⇑
Corresponding author. Fax: +34 915642431.
E-mail address: xjosemi@iqfr.csic.es (J.M. Mancheño).
0014-5793/$36.00 Ó 2012 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.febslet.2012.07.073

3128
M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
crystallize, revealing a harmful effect of the tag on the GPDH
structure.
and scaled using MOSFLM [15] and SCALA from the CCP4 program
suite [16]. Intensities were converted to structure-factor ampli-
tudes using TRUNCATE also from the CCP4 suite [16].
GPDH is a member of the medium-chain dehydrogenase/reduc-
tase (MDR) protein superfamily [6] which is composed of homodi-
meric or homotetrameric proteins that contain two tetrahedrally
coordinated zinc ions per subunit, one catalytic at the active site,
and one structural [7]. The structure of rGPDH reveals a structural
zinc coordinated by four cysteine residues and a catalytic metal-
binding site, occupied by a Ni2+ion, which is coordinated by three
protein ligands (C38, H59, E60) and a water molecule. The presence
of this last ion within the catalytic metal-binding site indicates that
this center may be involved in the interaction with the Ni–NTA
matrix.
The crystal structure of rGPDH was determined by the molecu-
lar replacement method with PHASER from the PHENIX program
[17] using the atomic coordinates of threonine-3-dehydrogenase
from Thermus thermophilus (PDB code 2DQ4) as search model.
The program CHAINSAW [18] from CCP4 [16] was used to prepare
the search model. The estimated Matthews coefficient was
2.40 Å3 Daꢁ1, corresponding to 49% solvent content and two rGPDH
molecules within the asymmetric unit. Refinement was performed
with PHENIX [17]. Non-crystallographic symmetry restraints were
initially applied, but these were removed before the final
refinement cycles. Isotropic individual temperature factors were
refined, with the TLS parameters added in the final stages of refine-
ment. After several further rounds of restrained and TLS refinement
and manual correction using Coot [19] the structural model was
finally refined to an R-factor of 21.3% and an Rfree of 27.3%. Valida-
tion of the structure was performed with MOLPROBITY [20]. Data
processing and structure refinement statistics are shown in
Table 1.
2. Materials and methods
Identification of galactitol-1-phosphate 5-dehydrogenase was
done by mass spectrometry techniques from tryptic and chymo-
tryptic peptides obtained from in-gel digestions of protein bands
from the SDS–PAGE. MALDI-MS and MS/MS data were combined
through the BioTools 3.0 program (Bruker Daltonik) to search the
non-redundant protein databases NCBInr and SwissProt using the
Mascot software (Matrix Science, London, UK) [8].
The gene gatD from Escherichia coli (K12) coding galactitol-1-
phosphate 5-dehydrogenase was PCR-amplified using the primers
F-his_gatD (50-GGTGAAAACCTGTATTTCCAGG GCATGAAATCAGTG-
GTGAATGAT) and R-his_gatD (50-ATCGATAAGCTTAGTTAGCTA
TTATCAGGGAATGAGCAACACTTT, for producing His6-GPDH, and
F-r_gatD (50-TAACTTTAAGAAGGAGATATACATATGAAATCAGTGGT-
GAATGAT) and R-his_gatD for producing rGPDH. Both protein
variants were prepared essentially as described [9]. Purification in-
volved an IMAC step on HisTrap FF Ni-affinity column (GE Health-
care) and an ion exchange on HiTrap Q HP column (GE Healthcare).
Purity was checked by MALDI-TOF-MS [10].
Table 1
Data-collection and refinement statistics.
rGPDH
PDB ID: 4A2C
Data collection
Beamline
Wavelength (Å)
Space group
Unit-cell parameters (Å)
ID14–1 (ESRF)
0.9334
P21
a = 43.7, b = 77.1,
c = 107.5
Km determinations for galactitol oxidation were done from
1 mM to 100 mM galactitol in 100 mM Tris–HCl buffer pH 9.0, con-
taining 0.5 mM ZnCl2 and 1.8 mM NAD+ with rGPDH concentra-
a
c
= 90°, b = 95.36°,
= 90°
Resolution range (Å)
No. of measured reflections
No. of unique reflections
Mean I/r (I)
Completeness (%)
29.81–1.87 [1.97–1.87] a
211,151
tions of 200 and 100
D
l
g/ml, respectively. Km determinations for
-tagatose-6-phosphate reduction were done from 0.1 to 2 mM
-tagatose-6-phosphate (Sigma–Aldrich) in 100 mM Bis–Tris buf-
56,481 [7753]
9.2 [2.5]
95.9 [95.9]
3.7 [3.7]
8.9 [52.8]
20.2
D
fer pH 6.5, containing 0.5 mM ZnCl2 and 0.3 mM NADH with
rGPDH concentrations of 5 and 2.5 g/ml, respectively. The reac-
Multiplicity
Rsym(%)b
l
Wilson B factor (Å2)
Refinement
tion was started by addition of the substrate and the change of
extinction was followed at 340 nm using an Ultrospec 2100 pro
photometer from GE Healthcare.
The oligomeric state of His6-GPDH and rGPDH in solution was
studied by analytical gel-filtration and analytical ultracentrifuga-
tion techniques as described previously [11]. Thermal stability
studies were done by differential scanning fluorimetry assays
[12] with the LightCycler 480 Real Time PCR System (Roche). Anal-
ysis of the data was done with the LightCycler Protein Melting Soft-
ware (Roche).
Resolution range (Å)
Rwork(%)/Rfree(%)
Molecules and no hydrogen atoms
Protein molecules
Protein
29.81–1.87
21.2/27.3
2
5312
3
2
608
Ni2+
Zn2+
Water
Average B-factors (Å2)
Main chain
27.1
31.6
37.8
20.6
34.7
Side chain
Crystallization of GPDH (and rGPDH) was done using the sparse
matrix method [13] essentially as described [14]. Optimization of
Ni2+
Zn2+
the crystallization conditions to hanging drops containing 2
the protein solution (9.3 mg/ml in 20 mM Tris–HCl, pH 8.0 with
0.1 M NaCl, 0.04% (w/v) sodium azide) and 1 l of reservoir solu-
ll of
Water
Rms deviation from ideality
Bonds (Å)
Angles (°)
l
0.007
1.127
tion (24% (w/v) PEG 2000 MME, 0.1 M Bis–Tris, pH 6.5) rendered
high quality diffraction crystals. rGPDH crystals for X-ray analysis
were transferred to an optimized cryoprotectant solution (reser-
voir solution plus 15% (v/v) glycerol) for ꢀ5 s and then cryocooled
at ꢁ173 °C in the cold nitrogen-gas stream. Diffraction data were
recorded on a Q210ADSC CCD detector (Area Detector Systems
Corp.) at beamline ID14-1 at the European Synchrotron Radiation
Facility (ESRF) (Grenoble, France). The images were processed
Ramachandran plot
(Favored/allowed/generously allowed/disallowed) 97/99.9/0/0.1
(%)c
a
Values for the highest resolution shell are given in square brackets.
Rsym
b
=
RhklRi|Ii(hkl)ꢁ<I(hkl)>|/RhklRiIi(hkl), where Ii(hkl) is the intensity for the
ith measurement of an equivalent reflection with indices h, k, and l and <I(hkl)> is
the average intensity.
c
Calculated using MOLPROBITY [20].

M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
3129
3.2. Biochemical and biophysical characterization of recombinant
variants of GPDH
3. Results and discussion
3.1. Endogenous GPDH interacts NTA-matrixes
Two recombinant variants of the dehydrogenase were produced
in E. coli BL21 (DE3) cells: His6-GPDH, which contained an N-termi-
nal His6-tag followed by a TEV cleavage site, and recombinant
GPDH, which is equivalent to the endogenous enzyme.
Both enzymes exhibited dehydrogenase activity, converting
galactitol (galactitol-1-phosphate is not commercially available)
into tagatose. This reaction was dependent on both Zn2+ and
NAD+ (Table 2). The lower but significant activity detected in the
absence of Zn2+ may be due to Zn2+ contaminations in the buffer
substances. Since the results obtained with His6-GPDH were hardly
reproducible due to its instability (see below) a subsequent, thor-
ough kinetic analysis was carried out only for the untagged
protein.
Initial small-scale expression trials aimed at producing the
unrelated protein esterase Q88Y25_Lacpl from the lactic acid
bacteria Lactobacillus plantarum WCFS1 in E. coli BL21 (DE3) cells
revealed poor levels of expression, which were subsequently im-
proved by the co-overexpression of the enzyme with molecular
chaperones GroES/GroEL from E. coli [21]. In these conditions, a
minor but significant peak was observed in the eluate from the Hi-
sTrap FF column. The imidazole concentration required for elution
of the protein was ꢀ 25 mM, indicating a weakly bound protein [5].
The mass value of m/z obtained by MALDI-TOF mass spectrometry
(37362 Da) was lower than the expected one for the complete pro-
tein (38628 Da), therefore suggesting either that the expressed
protein was not the expected one or that it was a truncated form
of the putative esterase.
In-gel trypsin and chymotrypsin digestions and subsequent
mass spectrometry analyses unambiguously identified the protein
as the galactitol-1-phosphate 5-dehydrogenase from E. coli
(UniProt code: P0A9S3). The sequence coverage was 32% and 60%
for the cleavage with trypsin and chymotrypsin, respectively.
Some peptides containing Asn residues were identified as the
deamidated species (chymotryptic peptides comprising residues
47–57, 47–61 and 47–64), and a few methionine oxidations were
also observed (chymotryptic peptides 211–227, 278–290 and
327–343; tryptic peptides 203–225 and 280–304).
Galactitol-1-phosphate is expected to be the main substrate for
oxidation activity of rGPDH, but this is not commercially available.
In comparison, the unphosphorylated galactitol has to be consid-
ered as a substrate with low oxidation activity and consequently
a low affinity apparent Km value of 25.8 mM (
determined (Fig. 2). Importantly, the reverse reaction of
reduction was not detectable. Reduction of -tagatose-6-phos-
r
= 3.3 mM) was
D
-tagatose
D
phate expressed an increase in specific activity of more than
500-fold as compared to galactitol oxidation (Table 2) and an
apparent Km of 1 mM, which clearly demonstrates the expected
substrate specificity of this enzymebecause galactitol-1-phosphate
is the product of this reaction. Among other polyols as substrates
only L-sorbitol exhibited oxidation activity and very low activity
was found with 1,2-hexanediol (Table 2).
Therefore, endogenous (untagged) galactitol-1-phosphate 5-
dehydrogenase from E. coli spontaneously interacts with HisTrap
FF Ni-affinity columns, similarly to other proteins from E. coli
which exhibit high affinity for divalent cations [5]. In this sense,
GPDH shares similarity to oxydoreductases from the medium-
chain dehydrogenase/reductase (MDR) protein superfamily [6]
that contain two distinct metal-binding sites namely a structural
Zn2+-binding site formed by four Cys residues and a catalytic me-
tal-binding site (Fig 1). In particular, GPDH shows a sequence iden-
tity of ꢀ27% to threonine 3-dehydrogenase from Thermus
Analytical gel-filtration chromatography and ultracentrifuga-
tion assays in conjunction with MALDI-TOF-MS analysis of the in-
tact proteins were used to study the oligomeric state(s) of rGPDH
and His6-GPDH in solution. The first approach revealed that both
variants behave as dimeric species (rGPDH: 63 kDa; His6-GPDH:
72 kDa) and also that His6-GPDH is metastable in solution, forming
large aggregates (Fig 3a).
Conversely, sedimentation velocity analyses revealed that
rGPDH (19 and 30 lM) is very homogeneous with an s value of
thermophilus (PDB entry: 2DQ4) and also to L-threonine dehydro-
4.5 0.1 S (Fig. 3b). Additionally, the results obtained from the sed-
imentation equilibrium experiments indicate that the molecular
weight of this species (70.7 kDa for the analysis at 14,000 rpm
and 77.2 kDa at 18,000 rpm) is consistent with the mass of a dimer
(rGPDH monomeric theoretical mass is 37390 Da; Fig. 3c). This
conclusion is definitively supported by the crystal structure of
rGPDH which reveals it is a dimeric assembly (see below).
genase from Thermococcus kodakaraensis (PDB entry: 3GFB) [22].
Protein ligands involved in metal coordination in the latter two
dehydrogenases are also present in GPDH (Fig 1), suggesting that
they play a similar role in this enzyme, presumably explaining its
high affinity for Ni2+ ions immobilized in the NTA matrix.
Similar experiments carried out with His6-GPDH indicated that
it is heterogeneous, in agreement with the analytical gel-filtration
assays. The sedimentation velocity data indicated that despite the
Table 2
rGPDH activity at different conditions and with different substrates.
Substrate
20 mM
Cation
0.5 mM
NAD+
1.8 mM
NADH
0.3 mM
Spec. activity
(U/mg)
Activity
(%)
Galactitol
Galactitol
Galactitol
Galactitol
Galactitol
Zn2+
Zn2+
Mg2+
ꢁ
+
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
+
0.060
0.001
0.016
0.020
0.001
31.500
0.001
0.090
0.001
0.003
0.004
0.003
0.010
100
–
26.7
33.3
–
52,500.0
–
150
–
5.0
6.7
5.0
16.7
ꢁ
+
+
ꢁ
ꢁ
ꢁ
ꢁ
+
+
+
+
+
+
D-Tagatose-6-P
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
Zn2+
D-Tagatose
+
Fig. 1. Multiple amino acid sequence alignment of GPDH with threonine 3-
dehydrogenase from Thermus thermophilus (PDB entry: 2dq4) and with L-threonine
L-Sorbitol
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
D-Sorbitol
dehydrogenase from Thermococcus kodakaraensis (PDB entry: 3gfb). Conserved
positions are shown in red characters; positions conserved in GPDH and either 2dq4
or 3gfb are shown as white characters in a blue box. Residues that coordinate the
catalytic Zn2+ ion are in green boxes, and the four cysteine residues that coordinate
the structural Zn2+ ion are within cyan boxes.
Mannitol
Ribitol
Xylitol
1,2-Hexanediol

3130
M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
Fig. 2. Lineweaver–Burk plots for rGPDH from E. coli with galactitol (a) and
oxidation reaction at two different enzyme concentrations: 100 g/ml (closed symbols) and 200
two enzyme concentrations: 2.5 g/ml (closed symbols) and 5.0 g/ml (open symbols).
D
-tagatose-6-phosphate (b) as substrates. Four different experiments were done for the galactitol
l
l
l
g/ml (open symbols). Reduction of -tagatose-6-phosphate was studied at
D
l
major species corresponded to dimeric His6-GPDH (s value
4.8 0.2 S), this species only represented ꢀ25% out of total of spe-
cies, whose s values were higher and broadly distributed (not
shown).
As a whole, these results suggest that the presence of the His6-
tag at the amino-terminus of the enzyme affects negatively to its
stability. In agreement with this, protein thermal stability was
studied by differential scanning fluorimetry since this method
can be applied to samples prone to aggregate [12] due to the small
amount and low concentration of protein required. This approach
revealed an apparent Tm value of 63.07 0.06 °C for rGPDH and
59.96 0.12 °C for freshly prepared His6-GPDH. The resulting
of the structure folds (CATH)) of the members of the MDR protein
superfamily within the PDBe (http://www.ebi.ac.uk/pdbe). This
analysis reveals that there exist 158 entries of homooligomers for
the MDR group (3.90.180.10), which are distributed as: 110 di-
meric (69.6%), 46 tetrameric (29.1%) and 1 hexameric (1.3%). Since
just 7 (PDB entries: 1uuf, 1qor, 1o8c, 1o89, 1wly, 1xa0, and 1iyz)
out of 110 dimeric assemblies, and 24 out of 46 tetrameric assem-
blies come from bacteria, it can be inferred that bacterial MDR
members are strongly biased towards tetrameric assemblies.
Inspection of these bacterial, dimeric assemblies reveals a similar
subunit arrangement, which is the one observed for rGPDH, where
the 6-stranded b-sheet of the Rossmann fold of one subunit inter-
acts with the equivalent b-sheet of the accompanying subunit
forming an extended 12-stranded b-sheet across the dimer inter-
face to create a large interface between subunits (Fig. 4a). Analysis
with PISA (http://www.ebi.ac.uk/msd-srv/prot_int/cgi-bin/piserv-
er) indicates that formation of the rGPDH dimer buried 1726 Å2
of the 15,156 Å2 accessible surface of each monomer, giving a sol-
vent-accessible surface area of 26,860 Å2 for the dimer. A second
D
Tm (3.11 0.12 °C) is within the range observed in other systems
[23]. Hence, it is obvious that the His6-tag has an adverse, destabi-
lizing effect on the enzyme which we believe it can be explained
since the N-terminal segment of rGPDH provides some clues about
this (see below).
3.3. Structural basis of the anomalous IMAC behavior of GPDH and a
possible explanation for the adverse effect of the N-terminal His6-tag
point of contact is the lobe loop surrounding the structural Zn2+
-
binding site. Here, E94 forms a salt bridge with R228 from the
other subunit (3.3 Å), and the aromatic ring of F99 stacks against
the imidazole ring of H253 and undergoes a hydrophobic interac-
tion with F229 which in turn packs against the aliphatic chain of
The crystal structure of apo rGPDH determined at 1.87 Å resolu-
tion revealed that the enzyme is a dimer. This result is notable as
deduced from an analysis of the protein assemblies (exploration

M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
3131
Fig. 3. Analysis of the oligomeric state of rGPDH in solution. (a) Analytical gel-filtration analysis of rGPDH and His6-GPDH. The elution profiles of both proteins (red, rGPDH;
black, His6-GPDH) are shown together with the eluted positions of some standard proteins. Inset, semilog plot of the molecular mass of standard proteins (kDa) versus their
Kav values. (b) Sedimentation coefficient distribution c(s) corresponding to the sedimentation velocity of purified rGPDH (30
lM). (c) Sedimentation equilibrium data of
rGPDH (19 M) at 14,000 rpm (open circles) and 18,000 (open squares). The best fit to the data sets (solid line curves) corresponded to an ideal dimeric species. Residuals
l
from the fits are shown in the two lower panels.
K97. A hydrogen bond is also observed between N
and the carbonyl oxygen of R278.
e
2 atom of N102
each one of the above potential ligands may play a role in the cat-
alytic mechanism of rGPDH. Similarly to the structural Zn2+-bind-
ing site, conformation of protein ligands from the catalytic sites
from both subunits is essentially identical, although in this case
small differences in coordination distances are observed: for in-
stance, distances between E60 and Ni2+ are 3.1 Å and 2.6 Å in sub-
units A and B, respectively and 3.1 Å and 2.4 Å between Ni2+ and
the closest water molecule in subunits A and B, respectively.
The presence of a Ni2+ ion within the catalytic metal-binding
site of rGPDH supports the notion that this particular environment
may mediate the interaction with the IMAC matrix. In this sense, it
is important to remark that the presence of a Ni2+ ion in this metal
binding-site does not preclude the interaction since pure rGPDH
interacts with the matrix (not shown) and therefore initial
Ni2+-depletion conditions are not required. The existence of this
interaction may open up the possibility to configure straightfor-
ward purification protocols for other members of the MDR
superfamily by IMAC.
Obviously, the potential interaction between the IMAC matrix
and solvent exposed histidine residues cannot be discarded. In this
regard, the crystal structure of rGPDH reveals an additional Ni2+ion
coordinated to the imidazole rings of H267 from both subunits, a
water molecule and the carboxylate group of E300 from a symme-
try related molecule (Fig. S1). Despite this metal-binding site is
crystallographic in that it results from the protein crystal packing
it cannot be discarded as a latent Ni2+-binding site in solution
interacting with the matrix.
Based on the reported behavior of several metalloproteins [5],
the most plausible explanation for the intrinsic ability of rGPDH
to interact with Ni2+-NTA matrices is the presence of metal-bind-
ing sites within the structure. Spherical bunches of electron densi-
ties were observed in both the structural and catalytic sites, which
were assigned to Zn2+ and Ni2+ ions, respectively since this resulted
in B-factors for the metals in better agreement with those of the
atoms from the inner-sphere coordination. Average B-factors for
the atoms of the inner sphere coordination of the Zn2+ ions were
20.9 Å2 (chain A) and 23.0 Å2 (chain B), and those for zinc ions
are 19.4 Å2 and 21.7 Å2, respectively, whereas the corresponding
values for Ni2+ ions were 31.2 Å2 (chain A) and 28.7 Å2 (chain B),
and 41.6 Å2 (chain A) and 34.1 Å2 (chain B), respectively.
The structural zinc ion is bound within the lobe loop adjacent to
the catalytic domain, close to the subunits interface. The Zn2+ is
tetrahedrally coordinated by C89, C92, C95 and C103 (Fig. 4b).
Structural metal-binding site from both subunits are almost per-
fectly superimposable, with the coordination distances between
Zn2+ and the ligands varying between 2.2 and 2.3 Å. Despite this
loop makes intersubunit interactions (see above) there is no clear
correlation between oligomeric state of MDR superfamily mem-
bers and presence or absence of zinc: whereas the overall trend
is that mammalian MDRs are dimeric and contain zinc and bacte-
rial MDRs are tetrameric and lack zinc [24] it is notable that only
one (together with rGPDH) out of the 7 dimeric bacterial MDRs
(PDB code 1uuf) contains zinc.
The N-terminal end of rGPDH is not flexible but rigid as indi-
cated by: firstly, the N-terminal methionine of rGPDH is perfectly
defined in the electron density map with the side chain being bur-
ied within the structure (Fig. 4d); secondly, the b1 strand (residues
1–6)is tightly packed between the b2 strand (residues 13–17) and
the extended segment formed by residues 117–119; thirdly, a salt
bridge is formed between E16 carboxylic group and the N-terminal
Conversely, the active site of rGPDH harbors a nickel ion that is
bound by C38, H59, E60 and a water molecule, with E144 being in
close proximity also (Fig. 4c). Despite the subtle details of the coor-
dination of the catalytic zinc cannot be inferred from this structure
it is now recognized that there exist important differences in zinc
ligation along the catalytic mechanism of MDRs [25] and therefore

3132
M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
Fig. 4. Crystal structure of rGPDH. (a) Dimeric assembly of rGPDH within the monoclinic crystals. The 6-stranded b-sheet of the Rossmann fold of one subunit interacts with
the equivalent b-sheet of the accompanying subunit forming an extended 12-stranded b-sheet. (b) View of the composite omit map calculated using the program SFCHECK
[26] contoured at1.0r around the structural zinc ion and (c) the nickel ion at the active site. The structural zinc ion is tetrahedrally coordinated by C89 (2.3 Å), C92 (2.3 Å), C95
(2.3 Å) and C103 (2.3 Å). (c) The active site nickel ion is coordinated by C38 (2.5 Å), H59 (2.2 Å), E60 (3.1 Å) and a water molecule (3.1 Å). E144 is also in close proximity to the
metal (3.5 Å). (d) N-terminal region of rGPDH. The first four residues of the enzyme are defined in the composite omit map (except the flexible K2 side chain) indicating that it
is a rigid segment. It is bridged by b2 strand and the loop between residues 117–119. The figures have been prepared with PyMol (http://pymol.sourceforge.net/).
[3] Hochuli, E., Bannwarth, W., Döbeli, H., Gentz, R. and Stüber, D. (1988) Genetic
amino group. It is obvious that incorporation of additional residues
at the N-terminal end (as in His6-GPDH) eliminates this last inter-
approach to facilitate purification of recombinant proteins with a novel metal
chelate adsorbent. Biotechnology 6, 1321–1325.
action what may destabilize locally the protein. This fact together
with the expected, increased flexibility within this region due to
the presence of the complete tag (His6-tag plus TEV cleavage site)
may result in an improper folding of this region making His6-GPDH
a metastable molecular species.
[4] Carson, M., Johnson, D.H., McDonald, H., Brounillette, Ch. and DeLucas, L.J.
(2007) His-tag impact on structure. Acta Crystallogr. D Biol. Crystallogr. 63,
295–301.
[5] Bolanos-García, V.M. and Davies, O.R. (2006) Structural analysis and
classification of native proteins from E. coli commonly co-purified by
immobilised metal affinity chromatography. Biochim. Biophys. Acta 1760,
1304–1313.
[6] Persson, B., Hedlund, J. and Jörnvall, H. (2008) The MDR superfamily. Cell. Mol.
Life Sci. 65, 3879–3894.
Acknowledgements
[7] Auld, D.S. and Bergman, T. (2008) The role of zinc for alcohol dehydrogenase
structure and function. Cell. Mol. Life Sci. 65, 3961–3970.
[8] Perkins, D.N., Pappin, D.J., Creasy, D.M. and Cottrell, J.S. (1999) Probability-
based protein identification by searching sequence databases using mass
spectrometry data. Electrophoresis 20, 3551–3567.
[9] Curiel, J.A., DelasRivas, B., Mancheño, J.M. and Muñoz, R. (2011) The pURI
family of expression vectors: a versatile set of ligation independent cloning
plasmids for producing recombinant His-fusion proteins. Prot. Express. Purif.
76, 44–53.
JMM and RM thank the Ministerio de Ciencia e Innovación for
their research grantsBFU2010-17929, CSD2006-00015 and
AGL2011-22745, CSD2007-00063 FUN-C-FOOD and Comunidad
de Madrid S2009/AGR-1469, respectively. JMM thanks ESRF (Gre-
noble, France) for provision of synchrotron radiation facilities.
[10] Acebrón, I., Curiel, J.A., de las Rivas, B., Muñoz, R. and Mancheño, J.M. (2009)
Cloning, production, purification and preliminary crystallographic analysis of
a novel glycosidase from the food lactic acid bacterium Lactobacillus plantarum
CECT 748T. Prot. Express. Purif. 68, 177–182.
[11] de las Rivas, B., Fox, G.C., Angulo, I., Martínez-Ripoll, M., Rodríguez, H., Muñoz,
R. and Mancheño, J.M. (2009) Ornithine transcarbamylase from Lactobaillus
hilgardii: structural insights into the oligomeric assembly and metal binding. J.
Mol. Biol. 393, 425–434.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.febslet.
2012.07.073.
[12] Niesen, H., Berglund, H. and Vedadi, M. (2007) The use of differential scanning
fluorimetry to detect ligand interactions that promote protein stability. Nat.
Protoc. 2, 2212–2221.
[13] Jancarik, J. and Kim, S.H. (1991) Sparse matrix sampling: a screening method
for crystallization of proteins. J. Appl. Crystallogr. 24, 409–411.
[14] Angulo, I., Acebrón, I., de las Rivas, B., Muñoz, R., Rodríguez-Crespo, J.I.,
Menéndez, M., García, P., Tateno, H., Goldstein, I.J., Pérez-Agote, B. and
Mancheño, J.M. (2011) High-resolution structural insights on the sugar-
References
[1] Block, H., Maertens, B., Spriestersbach, A., Brinker, N., Kubiek, J., Fabis, R.,
Labahn, J. and Schäfer, F. (2009) Immobilized-metal affinity chromatography
(IMAC): A Review. Methods Enzymol. 463, 439–473.
[2] Malhotra, A. (2009) Tagging for protein expression. Methods Enzymol. 463,
239–258.

M. Esteban-Torres et al. / FEBS Letters 586 (2012) 3127–3133
3133
recognition and fusion tag properties of a versatile b-trefoil lectin domain
from the mushroom Laetiporus sulphureus. Glycobiology 21, 1349–1361.
[15] Leslie, A.G. (1999) Integration of macromolecular diffraction data. Acta
Crystallogr. D Biol. Crystallogr. 55, 1696–1702.
[22] Bowyer, A., Mikolajek, H., Stuart, J.W., Wood, S.P., Jamil, F., Rashid, N., Akhtar,
M. and Cooper, J.B. (2009) Structure and function of the -threonine
L
dehydrogenase (TkTDH) from the hyperthermophilic archaeon Thermococcus
kodakaraensis. J. Struct. Biol. 168, 294–304.
[16] The CCP4 suite: programs for protein crystallography (1994) Acta Crystallogr.
D Biol. Crystallogr. 50, 760–763.
[23] Vedadi, M., Niesen, F.H., Aliali-Hassani, A., Fedorov, O.Y., Finerty, P.J., Wasney,
G.A., Yeung, R., Arrowsmith, Ch., Ball, L.J., Berglund, H., Hui, R., Marsden, B.D.,
Nordlund, P., Sundstrom, M. and Edwards, A.M. (2006) Chemical screening
methods to identify ligands that promote protein stability, protein
crystallization, and structure determination. Proc. Natl. Acad. Sci. U S A 103
(43), 15835–15840.
[24] Banfield, M.J., Salvucci, M.E., Baker, E.N. and Smith, C.A. (2001) Crystal
structure of the NADP(H)-dependent ketose reductase from Bemisia argentifolii
at 2.3 Å resolution. J. Mol. Biol. 306, 239–250.
[17] Adams, P.D., Afonine, P.V., Bunkóczi, G., Chen, V.B., Davis, I.W., Echoo, N.,
Headd, J.J., Hung, L.W., Kapral, G.J., Grosse-Kunstleve, R.W., McCoy, A.J.,
Moriarty, N.W., Oeffner, R., Read, R.J., Richardson, D.C., Richardson, J.S.,
Terwilliger, T.C. and Zwart, P.H. (2010) PHENIX: a comprehensive Python-
based system for macromolecular structure solution. Acta Crystallogr. D. 66,
213–221.
[18] Stein, N. (2008) CHAINSAW:
a program for mutating pdb files used as
templates for molecular replacement. J. Appl. Crystallogr. 41, 641–643.
[19] Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular
graphics. Acta Crystallogr. D 60, 2126–2132.
[20] Davis, I.W., Leaver-Fay, A., Chen, V.B., Block, J.N., Kapral, G.J., Wang, X., et al.
(2007) MolProbity: allatomcontacts and structure validation for proteinsand
nucleic acids. Nucleic Acids Res. 35, W375–W383.
[21] Álvarez, Y., Esteban-Torres, M., Acebrón, I., de las Rivas, B., Muñoz, R.,
Martínez-Ripoll, M. and Mancheño, J.M. (2012) Preliminary X-ray analysis of
twinned crystals of the Q88Y25_Lacpl esterase from Lactobacillus plantarum
WCFS1. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 67, 1436–1439.
[25] Baker, P.J., Brittona, K.L., Fisher, M., Esclapez, J., Pire, C., Bonete, M.J., Ferrer, J.
and Rice, D.W. (2012) Active site dynamics in the zinc-dependent medium
chain alcohol dehydrogenase superfamily. Proc. Natl. Acad. Sci. U S A 106 (3),
779–784.
[26] Vaguine, A.A., Richelle, J. and Wodak, S.J. (1999) SFCHECK: a unified set of
procedures for evaluating the quality of macromolecular structure-factor data
and their agreement with the atomic model. Acta Crystallogr. Sect. D 55, 191–
205.