Báo cáo khoa học: The interaction of the Escherichia coli protein SlyD with nickel ions illuminates the mechanism of regulation of its peptidyl-prolyl isomerase activity ppt

The interaction of the Escherichia coli protein SlyD with
nickel ions illuminates the mechanism of regulation of its
peptidyl-prolyl isomerase activity
Luigi Martino
1,2
, Yangzi He
1,
*, Katherine L. D. Hands-Taylor
1
, Elizabeth R. Valentine
1
, Geoff Kelly
3
,
Concetta Giancola
2
and Maria R. Conte
1
1 Randall Division of Cell and Molecular Biophysics, King’s College London, London, UK
2 Department of Chemistry ‘P. Corradini’, University of Naples ‘Federico II’, Italy
3 MRC Biomedical NMR Centre, National Institute for Medical Research, London, UK
Introduction
The Escherichia coli sensitive to lysis D (SlyD) protein
was originally discovered as a host factor required for
E-protein-mediated cell lysis upon infection with bacte-
riophage /X174 [1]. It was then found to be identical
to wondrous histidine-rich protein, independently iden-
tified as a persistent contaminant of histidine-tagged
Keywords
FK506-binding protein (FKBP); nickel;
peptidyl-prolyl cis-trans isomerase (PPIase);

SlyD; structure
Correspondence
M. R. Conte, Randall Division of Cell and
Molecular Biophysics, King’s College
London, Guy’s Campus, London SE1 1UL,
UK
Fax: +44 0 2078486435
Tel: +44 0 2078486194
E-mail:
*Present address
Department of Molecular Biology, University
of Aarhus, Gustav Wieds Vej 10C, DK-8000,
Aarhus C, Denmark
Database
Structural data are available in the Protein
Data Bank under the accession number
2KFW
(Received 14 May 2009, revised 15 June
2009, accepted 17 June 2009)
doi:10.1111/j.1742-4658.2009.07159.x
The sensitive to lysis D (SlyD) protein from Escherichia coli is related to
the FK506-binding protein family, and it harbours both peptidyl-prolyl
cis–trans isomerase (PPIase) and chaperone-like activity, preventing aggre-
gation and promoting the correct folding of other proteins. Whereas a
functional role of SlyD as a protein-folding catalyst in vivo remains
unclear, SlyD has been shown to be an essential component for [Ni–Fe]-
hydrogenase metallocentre assembly in bacteria. Interestingly, the isomer-
ase activity of SlyD is uniquely modulated by nickel ions, which possibly
regulate its functions in response to external stimuli. In this work, we inves-
tigated the solution structure of SlyD and its interaction with nickel ions,

enabling us to gain insights into the molecular mechanism of this regula-
tion. We have revealed that the PPIase module of SlyD contains an addi-
tional C-terminal a-helix packed against the catalytic site of the domain;
unexpectedly, our results show that the interaction of SlyD with nickel ions
entails participation of the novel structural features of the PPIase domain,
eliciting structural alterations of the catalytic pocket. We suggest that such
conformational rearrangements upon metal binding underlie the ability of
nickel ions to regulate the isomerase activity of SlyD.
Abbreviations
FKBP, FK506-binding protein; HsFKBP12, Homo sapiens FK506-binding protein; IF, insert in flap; ITC, isothermal titration calorimetry;
MtFKBP17, Methanococcus thermolithotrophicus FK506-binding protein; PPIase, peptidyl-prolyl cis–trans isomerase; SlyD, sensitive to
lysis D.
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4529
recombinant proteins purified by immobilized metal
affinity chromatography [2]. SlyD has been suggested
to bind to divalent cations, including Ni
2+
,Zn
2+
and
Co
2+
, via its C-terminal domain, a stretch of approxi-
mately 50 amino acids containing several short clusters
of potential metal-binding residues such as histidines,
cysteines, aspartates, and glutamates [2–5]. Whereas
the N-terminal region of SlyD shares primary sequence
homology with the ubiquitous FK506-binding protein
(FKBP) family of peptidyl-prolyl cis–trans isomerases
(PPIases), this C-terminal tail appears to be a unique

feature of SlyD bacterial proteins [3,5] (Fig. 1).
The functional profile of SlyD is rather intriguing.
As a member of the FKBP family, SlyD harbours
prolyl isomerase activity, which is responsible for
accelerating the rate-limiting trans-to-cis isomerization
step in protein folding [6–9]. More recent work, how-
ever, has shown that SlyD associates a PPIase function
with a proficient chaperone-like activity, preventing
aggregation and promoting the correct folding of other
proteins [10–14]. SlyD displays high affinity for
unfolded proteins, irrespective of their proline content,
in a manner evocative of the E. coli trigger factor,
which also combines PPIase and chaperone abilities
[10,15,16]. The chaperone-like activity of SlyD appears
to reside in a characteristic insertion within the PPIase
domain when compared to eukaryotic FKBPs, called
the ‘insert-in-flap’ (IF) domain (Fig. 1) [17]. The IF
domain is also a trait of the archaeal FKBP from Met-
hanococcus thermolithotrophicus (MtFKBP17) [18,19],
conferring, as in this case, chaperone-like competence
to the protein [20].
To date, the physiological role of SlyD as a chaper-
one assisting with protein folding in vivo has remained
unclear [10,13,21]. Nonetheless, a function for SlyD in
the [Ni–Fe]-hydrogenase biosynthetic pathway has
recently emerged, with the identification of SlyD as an
essential component of the hydrogenase metallocentre
assembly, probably serving as a nickel supplier for the
formation of [Ni–Fe] clusters [22,23]. Not only does
this concur with the ability of SlyD to bind nickel ions,

but, notably, nickel ion binding to SlyD provides the
means to reversibly regulate its PPIase activity [9].
Consistent with this, the PPIase ability of SlyD has
been shown not to be critical for its role in hydroge-
nase biosynthesis [24].
Further investigations have uncovered a key interac-
tion between SlyD and the hydrogenase accessory
factor HypB [23]. E. coli HypB contains two metal-
binding sites – a high-affinity site in the N-terminal
region and a low-affinity site within its GTPase
domain – and both are required for hydrogenase matu-
ration [25–27]. However, in contrast to other bacterial
HypBs, the E. coli protein lacks additional storage
capacity for nickel in the form of a histidine-rich
stretch found in other organisms [28–30]. It has been
proposed that HypB interaction with SlyD may there-
fore circumvent this deficit, by recruiting extra metal-
binding capacity to the system. In support of this
hypothesis, whereas the PPIase domain of SlyD is
required to interact with an N-terminal proline-con-
taining stretch of HypB, the putative metal-binding
region, comprising residues 146–196, is strictly essential
for the role in hydrogenase biosynthesis [23].
The primary sequence homology of SlyD with other
FKBP proteins terminates around residue 139, incor-
porating the IF domain, which is also found in
archaea (Fig. 1). The C-terminal tail is present in SlyD
variants from other bacteria, with some degree of
sequence conservation (Fig. 1); this has been suggested
to be unstructured and an easy target for proteolytic

degradation [9,10].
Fig. 1. Alignment of SlyD bacterial proteins
and FKBP homologues. The alignment was
obtained using
T-COFFEE (.
uk/Tools/t-coffee/index.html). Invariant
residues are boxed in black, and conserved
residues in grey. The secondary structure
elements are superposed on the amino acid
sequence. The names of proteins of differ-
ent species are as follows: SlyD_ECOL,
E. coli; SlyD_HAEIN, Haemophilus influen-
zae; SlyD_AERHY, Aeromonas hydrophila;
SlyD_TREPA, Trepomena pallidum; SlyD_
HELPY, Helicobacter pylori; SlyD_HELPJ,
He. pylori J99; MtFKBP17, M. thermolitho-
trophicus; HsFKBP12, H. sapiens.
Structure and interactions of SlyD L. Martino et al.
4530 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
In this study, we have investigated the solution
structure of E. coli full-length SlyD, and found an
atypical PPIase domain containing an additional C-ter-
minal a-helix packed against the rest of the domain.
This is in full agreement with the very recently
published structure of the N-terminal part of SlyD,
although a functional role for this extra structural ele-
ment has not been established in this work [31].
Intriguingly, our investigations reveal that this C-ter-
minal helix, unprecedented for the FKBP family of
proteins, is involved in nickel ion binding, causing con-

formational rearrangements in the PPIase domain and
modulating its isomerase activity. The basis of this
molecular switch will be discussed.
Results
Structure determination
To characterize the structure of E. coli SlyD, prelimin-
ary NMR analysis was applied to full-length SlyD
(encompassing residues 1–196) and a truncated N-ter-
minal fragment, homologous to other FKBP proteins,
spanning residues 1–146 (SlyD1–146). Comparison of
1
H-
15
N HSQC spectra of the two molecules (data not
shown) indicated that, although many resonances were
directly superimposable, a number of well-resolved sig-
nals appeared to be shifted in the context of the dele-
tion mutant, suggesting potential intramolecular
interactions involving the PPIase core domain and a
number of residues beyond Glu146. As, in our hands,
the purified recombinant full-length SlyD appeared to
be stable in the conditions required for NMR analysis,
structural determination of the wild-type protein was
undertaken.
SlyD folds into two domains and a long,
unstructured C-terminal tail
The three-dimensional structure of E. coli SlyD was
determined using standard heteronuclear multidimen-
sional NMR techniques as described in Experimental
procedures. In solution, SlyD folds into two globular

domains, namely the PPIase domain and the IF
domain, bisected by a deep cleft. The PPIase domain
consists of two polypeptide segments, spanning resi-
dues 1–69 and 129–154, and the insert fragment, com-
prising residues 76–120, constitutes the IF domain
(Fig. 2). The partitioning of the polypeptide chain cre-
ates a pair of antiparallel strands at the base of the
cleft linking the two domains. These connecting seg-
ments span residues 70–75 and 121–128, respectively,
and act as a flexible hinge for a bending motion
between the domains (see below). These fragments are
not well defined, and few long-range NOE contacts to
the other domains could be unambiguously assigned,
although there is evidence of local structure in the
turns spanning regions 71–75 and 122–126.
The relative orientation of the PPIase and IF
domains is also undefined. Because both domains
establish contacts with residues located within the con-
necting segments, they do not tumble fully indepen-
dently of each other in solution. Nonetheless, no
unambiguous contacts between them could be
detected, and a fixed orientation could not be found in
our investigation (Fig. S1). This is in agreement with
previous structural studies of the archaeal homologue
MtFKBP17 [32] and with our backbone relaxation
analysis (Fig. 2c); in fact, estimates of the rotational
correlation times for the two domains, based on analy-
sis of T1 ⁄ T2 ratios, gave significantly different values
for the PPIase and IF domains, 13.6 and 11.2 ns
respectively. Furthermore,

1
D
NH
residual dipolar cou-
plings for SlyD were measured in liquid crystalline
media; however, attempts to find a single value for the
magnitude and rhombicity of the alignment tensor
using the maximum likelihood method [33] failed,
suggesting that the two domains could not align to a
single external axis.
Therefore, each domain was superimposed sepa-
rately to calculate the rmsd. A final family of 20 super-
imposed structures for the PPIase domain and the IF
domain is shown in Fig. 2; the overall values of rmsd
between the family and the mean coordinate position
are 0.749 and 0.828 A
˚
for backbone atoms in second-
ary structure regions, respectively. The structure calcu-
lation statistics are given in Table 1, and a
representative structure is reported in Fig. 3. The
structural quality, in terms of restraints violation and
deviation from the ideal geometry, was checked with
the program procheck-nmr (Table 1).
Our structure was also compared with the very
recent structure of the N-terminal fragment of SlyD,
encompassing residues 1–165 [31]. The rmsd values for
the PPIase and the IF domains are, respectively, 1.51
and 1.67 A
˚

over structured regions (defined in
Table 1), underscoring the fact that the structure of
the domains remains largely unaffected in the context
of the intact full-length protein. This is consistent with
the finding that the region encompassing residues 153–
196 of SlyD appears to be largely unstructured in our
study. Severe spectral overlap prevented us from
obtaining unequivocal sequence-specific assignment for
the majority of the residues in this stretch; however, a
comparative analysis of the HSQC spectra of full-
length SlyD and SlyD1–146 positively identified the
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4531
resonances arising from the C-terminal tail as a cluster
of signals around 8–8.5 p.p.m. characterized by
reduced {
1
H}-
15
N NOE values (data not shown).
Structure of the PPIase domain
The PPIase domain of SlyD possesses a b4–b5a–b5b–
a1–b2–b3–a4 topology, and folds to generate a twisted
four-stranded antiparallel b-sheet wrapped around the
a1-helix and flanked by the a 4-helix (Figs 2 and 3).
The numbering of the secondary structure elements
adopted here reflects the convention used for other
FKBP proteins (see below and Fig. 3). The a1-helix
displays a marked amphipathic character and sits on
A

B
C
Fig. 2. Analysis of structure and backbone dynamics of SlyD.
Superimposition of the backbone traces for the 20 lowest-energy
structures of SlyD (A) for the IF domain (traces showing resi-
dues 75–121) and (B) for the PPIase domain (residues 1–70 and
127–152). The N-termini and C-termini and secondary structure ele-
ments are indicated. The relative orientation of the two domains
with respect to each other is undefined; although they are not fully
mobile with respect to each other, long-range contacts could not
be unambiguously detected in this study (Fig. S1). (C) Backbone
relaxation analysis showing T1, T2 and {
1
H}-
15
N NOE values for
SlyD measured at 18.8 T and 298 K.
Table 1. Summary of structural statistics for SlyD.
NMR restraints SlyD
Total distance restraints (inter-residue)
Short–medium range
(residue i to I + j, j = 1–4)
728
Long range (residue i to I + j, j > 4) 474
Hydrogen bonds 23
Total dihedral angle restraints 230
/ 115
w 115
Restraint violations
Distance restraint violation > 0.2 A

˚
None
Dihedral restraint violation > 5° None
Average rmsd (A
˚
) among the 20 refined structures
Residues PPIase 1–69,
129–149
IF 76–121
Backbone of structured regions
a
0.749 0.828
Heavy atoms of structured regions 1.481 1.674
Backbone of all residues 0.778 1.228
Heavy atoms of all residues 1.501 1.899
Ramachandran statistics of 20 structures
Percentage residues in
Most favoured regions 87.8
Additional allowed regions 8.9
Generously allowed regions 2.2
Disallowed regions 1.1
a
Residues selected on the basis of
15
N backbone dynamics. PPI-
ase domain: 1–38, 41–68, and 129–148; IF domain: 76–83, 90–96,
and 105–120.
Structure and interactions of SlyD L. Martino et al.
4532 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
one portion of the b-sheet surface, with its polar flank

largely solvent-exposed and the apolar face making
hydrophobic contacts with several residues of the
b-sheet. The shorter a4-helix is packed against one
edge of the b-sheet and terminates in a sharp turn,
after which the unstructured C-terminal tail begins. It
was annotated as the a4-helix to avoid confusion with
the a2-helix found in archaeal proteins (see below and
Fig. 3). A reverse turn, comprising residues 64–69,
follows on from the b2-strand, and is found in most
FKBP structures, including Homo sapiens FKPB
(HsFKBP12) and MtFKBP17 (Figs 2 and 3). This is
stabilized by hydrophobic interactions with the
a1-helix, the b2-strand and the b3-strand, but it also
establishes a few contacts with the interconnecting
segments.
The structure of the PPIase module of SlyD closely
resembles the structure of the PPIase domain of
AB C
DE
F
Fig. 3. Structural comparison of HsFKBP12, MtFKBP17, and SlyD. Top panel: cartoon representations of the representative structures for
HsFKBP12 (A), MtFKBP17 (B), and SlyD (C). The flexible tail of SlyD has been truncated at residue 153. Lower panel: topological comparison
of HsFKBP12 (D), MtFKBP17 (E), and SlyD (F). The N-termini and C-termini are indicated. The secondary structure elements are labelled
according to the convention adopted for HsFKBP12.
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4533
FKBPs, as expected from primary sequence analysis;
however, there are interesting differences. One of the
closest structural homologues is the PPIase domain of
HsFKBP12 [34–36], although SlyD lacks the N-termi-

nal b1-strand that lies antiparallel to the b2-strand in
the human protein (Fig. 3). The b1-strand is also miss-
ing in MtFKBP17 [32], and in both proteins the short
structured N-terminal segment makes hydrophobic
contacts with the a1-helix and the b-sheet.
The b5-strand of SlyD is split into two halves,
namely b5a and b5b, separated by a five-residue bulge
bearing a close resemblance to the structure of
HsFKBP12. Conversely, the helix insertion observed in
MtFKBP17 (a2-helix; Fig. 3) appears to be confined
to the archaeal kingdom, and is not conserved in bac-
terial SlyD.
One of the most interesting features of the SlyD
structure is the presence of a novel helical extension to
the PPIase fold, termed the a4-helix. This elaboration
of the PPIase domain structure is thus far unique to
SlyD; it spans residues 144–149, and is almost entirely
missing in the truncated version (SlyD1–146). This
helix connects to the rest of the domain through a
well-defined segment extending from the b3-strand,
and is positioned at the convex side of the b-sheet near
to the ends of the b4-strand and b5b-strand. Residues
in the a4-helix establish a network of contacts with
Asp6, His38, Leu35, and Ala142, and undergo addi-
tional interactions with residues 151–153, which create
a tight turn following the a4-helix. Although long-
range NOE contacts could also be assigned between
Gly150, His151, Val152 and His153 in this turn and
Leu35 and Ala142 in the core domain, the dynamic
backbone analysis indicates that residues beyond

Ala149 experience intrinsic mobility on the nanosecond
to picosecond timescale. The position of this turn in
the structure as obtained from the structure calculation
therefore has to be considered as one of the possible
conformations.
This novel C-terminal extension of SlyD PPIase does
not obscure the putative peptidyl-prolyl binding side;
however, our results implicate it in the molecular
switch triggered by nickel ions (see below). Further-
more, it appears to be conserved in all of the SlyD
variants on the basis of primary sequence conservation
(Fig. 1).
Structure of the IF domain
The IF domain of SlyD displays a b6–a3–b9–b8–b7
topology, and folds to generate a four-stranded anti-
parallel b-sheet bordered by a short a-helix (a3-helix)
(Figs 2 and 3). This helix connects the b6-strand with
a partially flexible loop leading to the b9-strand. Phe84
and Val87 on the a3-helix engage in interactions with
Val112, Ile109, Val117 and Phe96 on the b-sheet, gen-
erating a hydrophobicity-stabilizing cluster that is the
core of the domain.
The IF domain of SlyD is aligned with the IF
domain of the archaeal homologous MtFKBP17 [32]
(Fig. 3). The main difference between these two
domains is the longer loop connecting the b9-strand
and b8-strand in SlyD. In our structure, several loops
are not as well defined as in the archaeal protein – this
is supported by relaxation analysis, although spectral
overlap prevented us from obtaining a complete set of

assignments and parameters for residues in this
domain. Inspection of the backbone relaxation, espe-
cially the {
1
H}-
15
N heteronuclear NOE values
(Fig. 2C), suggests a higher degree of intrinsic disorder
for the entire IF domain than for the PPIase module.
This picture agrees with the recently reported observa-
tion that the IF domain in isolation was unable to
adopt a stable fold in solution, and, when present in
the intact SlyD protein, it was found to destabilize the
PPIase domain [37]. The structural flexibility and plas-
ticity of the IF domain may constitute a necessary
feature for an efficient chaperone-like activity.
Consistent with its chaperone-like role and in line
with the archaeal counterpart, the SlyD IF domain
exhibits a large exposed hydrophobic surface with
potentially high affinity for unfolded or partially
folded proteins (Fig. S2). The IF domain might there-
fore perform a double activity: preventing aggregation
of unfolded substrates, and orientating them to facili-
tate their insertion within the PPIase domain. This is
in agreement with the degree of relative flexibility
observed in the structure. Conformational changes
involving bending motions of the hinge between the
two domains might in fact modulate access to and
from the PPIase active site. The presence of the IF
domain, which is unique in archaeal FKPBs and bacte-

rial SlyD, coupled with the hinge-bending motion
between the two domains, could enable the protein to
sample the surrounding space for potential ligands and
aid their interaction with the PPIase active site.
Comparison of PPIase domains and the
FK506–rapamycin interaction
The PPIase domain fold is highly conserved within the
large family of FKBP and FKBP-like proteins, and it
has also been found in parvulins, another group of
cis–trans prolyl isomerases [38,39]. The conserved moi-
ety of this fold appears to constitute the minimum
structural frame for PPIase activity, and includes the
Structure and interactions of SlyD L. Martino et al.
4534 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
b4-strand, b5-strand, b2-strand and b3-strand, and the
loop–helix–loop that forms the a1-helix [32,40,41]. The
active site of the extensively studied HsFKBP12 has
been mapped to a hydrophobic cavity delineated by
the concave surface of the b-sheet, the a1-helix, and
several loops [36]. The immunosuppressive agents
FK506 and rapamycin, which act as potent inhibitors
of the PPIase activity of the FKBPs, have been shown
to be lodged within the active site crevice of
HsFKBP12, cushioned by Tyr26, Phe36, Asp37,
Phe46, Phe48, Val55, Ile56, Trp59, Tyr82, Ile90, Ile91,
and Phe99 [36]. The corresponding binding pocket in
SlyD is structurally similar (Fig. 4), with the hydro-
phobic residues Tyr13, Val23, Asp24, Leu32, Tyr34,
Leu41, Ile42, Leu45, Tyr68 and Phe132 in analogous
positions, respectively, to those of the residues in the

conserved side chains in the human protein (Fig. 4).
The main difference lies in the position of Tyr82 ⁄ 68,
located in the reverse turn following the b2-strand,
also observed by Weininger et al. [31]. Furthermore,
Ile90 and Ile91, which reside in the loop between the
b2-strand and b3-strand of HsFKBP12, do not have
direct equivalents in SlyD. Nonetheless, because of the
relative mobility of the PPIase and IF domains, hydro-
phobic residues, such as Met124 and Leu125 in the
interconnecting segments, might be able to relocate in
the close vicinity of the crevice and undergo transient
interactions with the ligand. Also, importantly, the
additional a4-helix of SlyD does not affect the shape
of the cavity or obscure its entrance, as underscored
by comparable values of solvent-accessible surface
areas for the binding pockets of HsFKBP12 (Protein
Data Bank ID: 1FKF) and SlyD (460 ± 20 and
490 ± 80 A
˚
, respectively).
Notably, the exact mechanism of the PPIase cata-
lytic process uncertain, as is the role of the conserved
hydrophobic residues within the common domain fold
[8,39,41]. Unexpectedly, parvulins and a number of
FKBP-like proteins, such as the trigger factor, do not
bind FK506, despite the high structural homology with
genuine FKBPs, adding conviction to the view that the
shape of the cavity as well as its charge distribution
might determine substrate specificity [39,41]. The issue
of whether FK506 influences the PPIase activity of

SlyD has been somewhat unclear in the past [9]. Scholz
et al. [10] have recently demonstrated that FK506 is
capable of inhibiting the refolding activity of SlyD,
with an apparent binding affinity for the protein esti-
mated to be about three orders of magnitude weaker
than that for HsFKBP12. To characterize further the
interaction between FK506 and SlyD and, most impor-
tantly, to assess which regions of the protein make
contact with the ligand, we carried out a series of
1
H-
15
N HSQC NMR experiments, monitoring the
backbone amide chemical shift changes in SlyD upon
titration with FK506. This sensitive method allows the
detection of amide chemical shift perturbations caused
by direct binding or conformational changes induced
by ligand interaction, and can therefore demarcate the
regions directly affected by complex formation. Upon
addition of FK506, several SlyD resonances belonging
A
B
C
turn
Fig. 4. Structural alignment between the PPIase domains of SlyD
and HsFKBP12. (A) Ribbon representation of the superimposed
structures (HsFKBP12 in green and SlyD in yellow). The superimpo-
sition of the PPIase domains was performed using
DALI (http://
ekhidna.biocenter.helsinki.fi/dali_server/) (Z 9.7, rmsd $ 1.8 A

˚
over
107 residues; identity 22%). (B, C) Magnification of the active site
crevices of HsFKBP12 and SlyD, respectively (in grey). Selected
key catalytic residues showing a direct correspondence between
the two proteins are shown in stick representation and labelled.
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4535
to the PPIase domain, on and around the active cleft,
disappeared, whereas others on the IF domain experi-
enced a chemical shift variation on the fast equilibrium
timescale (Fig. S3). Similar results were obtained with
rapamycin (Fig. S3). In both cases, the titration was
terminated at a protein ⁄ drug ratio of 1 : 1, as further
additions of the largely water-insoluble agents caused
the formation of a white precipitate, and no further
change in the spectra was observed.
Although these data are not conclusive, they indicate
that both FK506 and rapamycin interact weakly with
SlyD. Although it cannot be excluded that the canoni-
cal PPIase site is implicated in the interaction, our
results show that the IF domain is clearly perturbed
by the presence of the ligand, on and around the
exposed hydrophobic patch.
Interaction of SlyD with nickel ions
SlyD is a unique FKBP protein, as its PPIase activity
is modulated by the presence of nickel ions [9]. The
nickel ions might therefore exert an important switch-
like regulatory control over the different functions of
SlyD, but the molecular basis of this attractive mecha-

nism remains uncertain. Because the PPIase activity of
the truncated SlyD1–146 was unaffected by nickel ions,
it was proposed that the C-terminal tail could be
responsible for binding metal ions and for the resultant
conformational change observed upon nickel ion bind-
ing [9]; however, this hypothesis leaves the question
open as to how this structural effect would be sensed
by the PPIase domain.
To achieve a deeper understanding of this regulatory
mechanism, the interaction between SlyD and nickel
ions was investigated using an array of biophysical
techniques. NMR titrations were employed to map the
binding site of the nickel ion on SlyD, and around a
1 : 1 nickel ion ⁄ protein ratio, several signals of the
protein disappeared in the
1
H-
15
N HSQC spectra
(Fig. S4). Intriguingly, the resonances perturbed by the
presence of nickel ions can all be mapped within the
PPIase domain, involving mainly, but not exclusively,
residues in the novel extension of the PPIase fold
(Figs 5 and S4). Moreover, a section of the PPIase
core domain, at or near the catalytic pocket, was also
affected. A more detailed mapping analysis of the
extent of the chemical shift variation upon nickel ion
binding was impeded by the loss of the perturbed
resonances, which could be attributable to either an
intermediate equilibrium of the complex and ⁄ or the

paramagnetic effect of the nickel ion (see below).
Nonetheless, our results clearly indicate that the ability
of SlyD to interact with nickel ions is not confined to
the unstructured C-terminal tail, as previously sus-
pected, but that the binding of at least one nickel ion
per molecule occurs on the PPIase domain, probably
affecting the conformation of the PPIase binding site
(see below). These results therefore provide the first
molecular explanation of the modulation of PPIase
activity of SlyD by nickel ions. These findings are not
in conflict with previously reported data, indicating
that the PPIase domain in isolation did not bind nickel
ions, because the putative PPIase fragment used in this
study terminated at residue 146, and so did not encom-
pass the full domain and lacked the key C-terminal
helical extension [9]. The NMR titration was con-
ducted up to a final nickel ion ⁄ SlyD ratio of 3 : 1;
nonetheless, further nickel ion additions beyond the
1 : 1 point caused only general line-broadening and
protein precipitation. The issue of the stoichiometry of
this interaction, however, deserved further attention, as
previous reports suggested 1 : 1, 3 : 1 or even higher
nickel ion ⁄ protein ratios [2,9,24]. To address this key
point and to further characterize such interaction
events, we employed isothermal titration calorimetry
(ITC) and CD techniques. ITC is largely used to inves-
tigate binding reactions by measuring the heat gener-
ated or absorbed in the binding event and thereby
providing the binding constant, the stoichiometry and
the enthalpy change (DH°) of the interaction [42,43].

For the nickel ion–SlyD interaction, carried out at
298 K and pH 7.25, the integrated heat data showed
that the process of nickel ion binding to the protein is
composed of one clear binding event (Fig. 6). The
binding isotherm corresponding to this reaction has
been obtained using an independent-site model [43],
revealing a stoichiometry of one nickel ion per protein,
an association constant of 4.16 · 10
5
m
)1
, and enthal-
pic (DH°) and entropic (TDS°) contributions of )166
and )134 kJÆmol
)1
respectively (Table 2). These nega-
tive values of enthalphy and entropy are typical of a
thermodynamic process describing metal coordination
by a protein molecule, with specific amino acid side
chains adopting a rigid conformation around the metal
ion [44,45]. The binding event is enthalpically driven,
suggesting that the formation of new interactions
between the nickel ion and the protein is the key
feature of the binding process.
Further evidence that SlyD interacts with nickel ions
with a 1 : 1 stoichiometry is provided by the analysis
of the changes in the far-UV CD spectra of the protein
upon titration with nickel ions (Fig. 6). Dramatic
changes in the molar ellipticity of SlyD were in fact
observed up to a 1 : 1 nickel ion ⁄ protein molar ratio,

and further additions of the ligand caused only minor
alterations in the CD spectra. This behaviour is in
Structure and interactions of SlyD L. Martino et al.
4536 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
A
B
Fig. 5. Structure mapping of the chemical
shift perturbations for the PPIase domain of
SlyD upon nickel ion binding. The positions
of residues that disappear in the
1
H-
15
N HSQC spectra upon complex forma-
tion are indicated in red on the protein sec-
ondary structure (A) and surface (B).
Selected perturbed residues are labelled.
The novel a4-helix appears to be signifi-
cantly involved in the interaction, and resi-
dues on the b5b-strand are suggested to
undergo conformational rearrangements in
the nickel ion-bound protein (see text).
KJ·mol
–1
[Ni
2+
]/[SlyD]
–20
–40
–60

–80
–100
–120
–140
–160
0.0 0.5 1.0 2.0 2.51.5
0
[θ]
215 nm
10
–3
(deg·dmol
–1
·cm
2
)
[Ni
2+
]/[SlyD]
–0.8
–0.9
–1.0
–1.1
–1.2
–1.3
0123456
4
3
2
1

0
–1
[θ] 10
–3
(deg·dmol
–1
·cm
2
)
λ (nm)
200 210 220
230
240
250
Power (µJ·s
–1
)
–2
0
2
4
6
8
10
12
0 2500 5000 7500
Time (s)
10 000
A
B

C
D
Fig. 6. Analysis of SlyD–nickel ion interaction. (A) Far-UV CD spectra of apo-SlyD (straight line) and SlyD in the presence of NiCl
2
(dotted
line) at a protein ⁄ nickel ion molar ratio of 1 : 1. The secondary structure content estimated by CD spectral analysis gave the following values:
apo-SlyD: 5% a, 47% b, 22% turn, and 25% irregular; SlyD–nickel ion: 5% a, 38% b, 20% turn, and 36% irregular). (B) Plot of the molar
ellipticity at 215 nm as function of the nickel ion ⁄ SlyD molar ratio. Interpolation of the experimental data (filled squares) with an equation
(dotted line) based on an independent binding sites model gives a stoichiometry of one nickel ion per protein molecule and a binding con-
stant of 2 · 10
5
M
)1
. (C) Raw titration data show the thermal effect of 10 lL injections of 400 lM NiCl
2
solution into a colorimetric cell filled
with 40 l
M SlyD solution at pH 7.25; the heat effect reveals an exothermic effect during the interaction. (D) Normalized heat of interaction:
data were obtained by integrating the raw data and subtracting the heat of ligand dilution into the buffer. The dashed line represents the
best fit obtained by a nonlinear least squares procedure based on an independent binding sites model.
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4537
agreement with the NMR and ITC results, indicating
that a major binding event occurs with a 1 : 1 stoichi-
ometry, accompanied by distinctive conformational
rearrangements in the protein. The far-UV CD spec-
trum of free SlyD, in the range 190–250 nm, shows a
well-defined minimum at 215 nm and a shoulder cen-
tred at 230 nm. Although the shapes of the curves are
very similar overall, the molar ellipticity is appreciably

less negative in the CD spectrum of the protein in a
1 : 1 complex with nickel ions than for the apo-pro-
tein, indicating a decrease in the secondary structure
content in the protein upon nickel ion binding. The
hyperbolic curve obtained by plotting the intensity of
the CD signal at 215 nm versus the nickel ion ⁄ protein
concentration molar ratio (Fig. 6) fits well with an
equation describing a simple model of a 1 : 1 interac-
tion (see Experimental procedures). Most importantly,
the association constant derived in this analysis is in
excellent agreement with the binding constant obtained
by ITC (Table 2), indicating that both techniques are
following the same process. The conformational
changes associated with the SlyD–nickel ion interac-
tion could therefore explain the higher values obtained
for the enthalpic and entropic contributions when
compared to other protein–nickel ion systems studied
[44,45], on the basis that the ITC phenomenon mea-
sured here is the result of both a molecular association
event and a concurrent conformational rearrangement.
To examine further such a conformational change,
deconvolution analysis of the far-UV CD spectra was
performed using dichroweb (see Experimental proce-
dures). For the apo-protein, the assessed a ⁄ b content is,
overall, consistent with its solution structure (Fig. 6),
but a marked decrease in the b-strand content was esti-
mated for the protein in the complex (without apprecia-
ble changes in the a-helical regions). This is particularly
interesting, as it might suggest that nickel ion binding
promotes the disruption of the b-sheet catalytic core of

the PPIase domain; this agrees well with the results of
the NMR titration experiments, where Leu32, Asp33,
Tyr34, Leu35 and His36 on the b-sheet appeared to be
perturbed by the metal ion interaction (Figs 5 and S4).
Discussion
In this work, we have investigated the solution
structure and molecular interactions of SlyD, a bacte-
rial protein related to the FKBP family of prolyl
isomerases. As for many members of the PPIase super-
family, an explicit function for SlyD in assisting with
protein folding in vivo remains, as yet, uncertain [46].
A number of the prolyl isomerases have been shown to
be involved in many other cellular processes [47,48]
and, likewise, SlyD has been identified as a key player
in the [Ni–Fe]-hydrogenase biosynthetic pathway [22].
SlyD consists of a long, unstructured C-terminal tail
preceded by two independently folded modules, the
PPIase domain and the IF domain, with isomerase and
chaperone-like properties respectively (see above). In
the final stages of this investigation, the solution struc-
ture of an N-terminal part of SlyD, SlyD(1–165), was
also determined [31]. A comparison of these reports
shows that the structure of the individual domains is
largely conserved within the context of the full-length
protein, and that the C-terminal region beyond resi-
due 157 is highly unstructured and independent of the
rest of the molecule.
Given that the SlyD structure bears unmistakable
similarities to that of MtFKBP17, it is conceivable that
these two domains work synergistically in SlyD, in line

with what has been suggested for the archaeal counter-
part [32]. This is corroborated by the finding that
insertion of the IF domain of SlyD into HsFKBP12
considerably boosts the chaperone-like activity of the
latter [17], and by the recent observation that the IF
domain of SlyD is directly involved in the binding of
unfolded proteins and peptides [31]. The relative flexi-
bility of the two domains revealed in the solution
structure implies a degree of domain swivelling that
might facilitate access to the PPIase catalytic pocket
and thereby enhance the ability of SlyD to act as a
folding catalyst. Collectively, the observations from the
NMR analyses and the existing literature point
towards a stepwise mechanism of catalysis whereby the
IF domain performs the initial docking of the peptide,
perhaps ideally positioning it for insertion within the
PPIase active site.
Notably, FK506 and rapamycin also appear to be
transiently anchored to the IF domain of SlyD in our
NMR chemical shift analysis, possibly mimicking the
recognition process of target peptides, consistent with
what has been suggested by Weininger et al. [31]. As
expected by comparison with other FKBP proteins,
key catalytic residues within the PPIase domain of
SlyD also experience some perturbation in the NMR
titrations upon ligand binding (see above). Nonethe-
Table 2. Results of the interpolation analysis for the binding of SlyD to nickel ions determined using ITC and CD.
n DH° (kJÆmol
)1
) TDS° (kJÆmol

)1
) DG°
298 K
(kJÆmol
)1
) K
b, ITC
(M
)1
) K
b, CD
(M
)1
)
1.0 ± 0.1 )166 ± 15 )134 ± 10 )32 ± 8 (4.1 ± 0.8) · 10
5
(2.0 ± 0.1) · 10
5
Structure and interactions of SlyD L. Martino et al.
4538 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
less, we believe that our data are not conclusive, and
whether SlyD is a genuine FK506-binding protein is
yet to be established.
The PPIase domain of SlyD emerges as being dis-
tinctive and rather intriguing, as its catalytic activity is
uniquely inhibited by the presence of nickel ions [9].
The overall architecture of this PPIase domain shows a
high degree of correspondence with other FKPB
domains, despite low sequence identity. Nonetheless, a
significant difference stems from the occurrence of a

novel helical extension to the PPIase module. Previous
structure predictions correctly anticipated the presence
of an FKBP-like domain in SlyD, but did not detect
the elaborations to the canonical fold revealed by our
study, which add valuable information to the expand-
ing databases of PPIase domains. The novel a4-helix
packs closely against one edge of the b-sheet in the
FKBP domain, with no apparent detrimental effect on
the shape and accessibility of the catalytic site. Unex-
pectedly, our investigations have shown that this addi-
tional structural feature does confer on the PPIase
domain the specific ability to bind to nickel ions, by a
mechanism never revealed before. The lack of struc-
tural information led to the suggestion that the puta-
tive unstructured C-terminal fragment, rich in
histidines, cysteines, aspartates, and glutamates, was
entirely responsible for the observed ability of SlyD to
bind metal ions, on the assumption that its PPIase
domain would terminate on or around residue 146 [9].
The structural results combined with the biophysical
characterization of the interaction with nickel ions pre-
sented here demonstrate unambiguously the existence
of one clear binding event of micromolar affinity and
1 : 1 stoichiometry that occurs on the PPIase domain,
in particular involving the novel structural features,
thereby assigning a new function (metal ion binding)
to this class of domains. However, most importantly,
our CD experiments have shown that nickel ion bind-
ing promotes loss and ⁄ or rearrangement of part of the
b-sheet within the PPIase catalytic pocket; the NMR

chemical shift analysis completes this picture by map-
ping such conformational changes to residues 32–36,
the only b-strand stretch exhibiting chemical shift per-
turbation upon interaction. Intriguingly, Leu32 and
Tyr34 have been tentatively identified as key catalytic
residues by comparison with HsFKBP12 (see above),
prompting suggestions that this structural perturbation
could be responsible for the reported loss of isomerase
activity of SlyD in the presence of nickel ions. As
chemical shift mapping is unable to discriminate
between residues directly in contact with the ligand
and residues undergoing conformational changes upon
binding, it is not straightforward at this stage to delin-
eate the side chains in SlyD responsible for nickel ion
coordination. Potential candidates may include histi-
dines (i.e. His36, His38, His149, His151, and His153),
glutamates (i.e. Glu146) and⁄ or aspartates (i.e. Asp33)
(Figs 5 and S4). The detailed structural basis of the
SlyD–nickel ion interaction therefore remains to be
elucidated, and further studies are currently underway;
nevertheless, this article provides the first molecular
explanation of the observed modulation of the isomer-
ase activity of SlyD by nickel ions, supporting the
hypothesis of a switch mechanism for SlyD, possibly
in response to environmental cues.
Our experiments could not detect further binding of
SlyD to nickel ions beyond the 1 : 1 stoichiometry.
This, however, does not exclude the possibility that the
C-terminal tail of SlyD beyond residue 156, rich in
metal-binding residues, would also be capable of inter-

acting with nickel ions. In fact, our NMR titration
experiments were hampered by line-broadening and
precipitation past the 1 : 1 molar ratio and by the
severe spectral overlap for amide protons beyond resi-
due 156; CD experiments would be unable to detect
binding events that do not cause conformational
changes; and ITC methodologies could be inadequate
for systems with an association constant equal to or
lower than 10
3
, as the heat associated with such inter-
actions would be negligible. As a different stoichiome-
try was detected using different techniques that do not
rely on the same biophysical parameters [24], it is con-
ceivable that the C-terminal stretch is also involved in
nickel ion binding. This would agree well with the
observation that, in E. coli, the polyhistidine stretch-
deficient HypB is counterbalanced by the C-terminal
extended SlyD, possibly to allow for extra nickel stor-
age. However, our experiments unambiguously show
that the first nickel ion occupies the higher-affinity site
on the PPIase domain, so any binding in the unstruc-
tured tail would be substantially (at least 100-fold)
weaker than the first nickel binding and, contrary to
previous suggestions, would not trigger conformational
rearrangements in SlyD or the resultant modulation of
its isomerase activity.
The physiological relevance of the nickel ion-binding
site on the PPIase domain remains to be established.
SlyD has been shown to interact with a 77-mer N-ter-

minal fragment of HypB, with a proline-rich stretch
(encompassing residues 28–36) partaking in such recog-
nition [23]. Notably, this fragment is in the vicinity of
the high-affinity nickel ion-binding motif in HypB
[CXXCGC(2–7)]. Whereas the PPIase activity of SlyD
has been shown not to be required for hydrogenase
assembly [24], it is unknown whether such activity may
actually be detrimental to the process. It may be
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4539
possible to envisage a mechanism in which the binding
of nickel ions to SlyD would ensure that its isomerase
activity is switched off, to prevent unwanted cis–trans
isomerization of the prolines adjacent to the high-affin-
ity metal-binding motif in HypB. Another model may
imply that the interaction between SlyD and HypB
somehow weakens the nickel ion affinity of the latter;
nickel ions thereby released by HypB may be readily
sequestered by the PPIase site of SlyD. Although coher-
ent with the fact that SlyD appears to interact in the
proximity of the nickel ion high-affinity motif of HypB,
the exact mechanism of the SlyD–HypB interaction
remains to be established, and this hypothesis awaits
testing. Little is known about the determinants of SlyD
responsible for HypB interaction. A role for the IF
domain in establishing contacts with HypB is supported
by the observation that the deletion mutant lacking
stretch 107–111 does not retain its chaperone ability or
its ability to crosslink HypB in vivo [23]. The structure
of SlyD maps fragment 107–111 to the b8-strand, argu-

ing that such a deletion mutant would probably not
retain its native conformation, and therefore substanti-
ating the hypothesis that a fully functional IF domain
is indeed essential for such recognition; moreover, the
IF domain has been shown to be responsible for the
binding of several unfolded protein and peptides sub-
strates [31]. Further experiments are needed to under-
stand the molecular basis of the interaction between
SlyD and HypB, and this could shed light on the role
of the atypical nickel ion-binding PPIase domain of
SlyD in the hydrogenase assembly.
In conclusion, the molecular view of the SlyD–nickel
ion binding mode presented here provides a compre-
hensive structural context for the interpretation of
in vitro and cellular binding studies. Together, the
structure of SlyD and its interactions suggest a plausi-
ble mechanism for a molecular switch between its dif-
ferent functions, and lay the groundwork for further
experimentation.
Experimental procedures
Plasmid construction and protein expression
Full-length SlyD and the truncated SlyD1–146 mutant were
subcloned from E. coli strain BL21 chromosomal DNA into
a pQE60 vector (Qiagen, Hilden, Germany). Each recombi-
nant protein was cloned with and without a C-terminal
hexahistidine tag, and all of the proteins were expressed in an
M15 (pREP4) E. coli strain (Qiagen). The untagged full-
length SlyD (SlyD) was used for the structural and biophysi-
cal studies reported here. For NMR, cells were grown on
minimal media enriched with 0.8 gÆL

)1
[
15
N]ammonium
chloride and 2 gÆL
)1
[
13
C]glucose, at 37 °C, until a D
600 nm
of
0.6 was reached, and then induced with 1 mm isopropyl thio-
b-d-galactoside. Cells were harvested 4 h after induction,
resuspended in 20 mm Tris ⁄ HCl, 100 mm KCl and 10 mm
imidazole at pH 8, and lysed by sonication. After centrifuga-
tion at 39 700 g for 40 min, the soluble fraction was purified
by affinity chromatography on an Ni
2+
–nitrilotriacetic acid
resin (Qiagen), following the manufacturer’s protocol. The
eluted protein was dialysed in 20 mm Tris ⁄ HCl, 100 mm KCl
and 1 mm dithiothreitol at pH 7.25, and loaded on a 5 mL
Hi-trap DEAE column (GE Healthcare, Uppsala, Sweden).
The protein was eluted with a linear 0–2.0 m KCl gradient in
buffer A [50 mm Tris ⁄ HCl, 0.1 mm EDTA, and 10% (v ⁄ v)
glycerol, pH 7.25], and dialysed in 20 mm Tris ⁄ HCl, 100 mm
KCl, and 1 mm dithiothreitol (pH 7.25) (NMR buffer).
NMR spectroscopy
For NMR studies, pure SlyD was concentrated to 0.8 mm
in a volume of 700 lL. NMR spectra were recorded at

298 K on Varian Inova spectrometers operating at 14.1 and
18.8 T, and on Bruker Avance spectrometers at 14.1 and
16.4 T, equipped with triple resonance cryoprobes. The
1
H,
15
N and
13
C resonance assignments for SlyD will be
reported elsewhere. All NMR data were processed using
nmrpipe ⁄ nmrdraw [49] and analysed using xeasy [50].
Distance restraints were obtained from
15
N-edited and
13
C-
edited NOESY experiments; backbone dihedral angles were
determined using talos software [51]. Hydrogen-bonded
amide protons were detected by performing a series of
1
H-
15
N HSQC experiments up to 10 h after the protein was
buffer-exchanged in D
2
O. T1, T2 and {
1
H}-
15
N NOE

experiments were performed using the pulse sequences
adapted from standard schemes, and analysed using
nmrpipe.
1
D
NH
residual dipolar couplings were measured
at 298 K in a liquid crystalline phase composed of $ 5%
alkyl-poly(ethylene glycol) C8E5 ⁄ n-octanol in NMR buffer
[52]. Precise measurements of
1
J
NH
splittings were obtained
from
1
J
NH
-modulated 2D spectra [53].
For NMR titration experiments FK506 (Sigma) and
rapamycin (Caltag-MedSystems) were dissolved in 96% eth-
anol at a concentration of 22 mm and added in increasing
amounts to 600 lL of 0.5 mm
15
N-labelled protein in NMR
buffer.
1
H-
15
N HSQC spectra were recorded at ligand ⁄ pro-

tein molar ratios of 0.33 : 1, 0.66 : 1, 0.99 : 1, 1.22 : 1 and
1.55 : 1 to follow the resonances perturbed by ligand bind-
ing to the protein. The backbone amide assignments of
SlyD were transferred in the complex to the resonances that
would have the smallest Dd
AV
. The weighted average of
15
N and
1
H
N
chemical shift variation was calculated as fol-
lows: Dd
AV
¼f0:5½ðDd
1
H
N
Þ
2
þð0:2Dd
15
NÞ
2
g
1=2
. Titrations
were repeated with 96% ethanol alone to minimize the risk
of buffer interference.

For the titration with nickel ions, a solution of 45 mm
NiCl
2
in NMR buffer was added to the
15
N-labelled
Structure and interactions of SlyD L. Martino et al.
4540 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
protein, dissolved in the same buffer, at 0.6 : 1, 1.2 : 1,
1.8 : 1, 2.4 : 1 and 3.0 : 1 nickel ion ⁄ protein molar ratios.
Structure calculation
The solution structure of SlyD was calculated using a
combined torsion angle and Cartesian coordinates dynam-
ics protocol executed in xplor [54]. The structures were
calculated from random starting coordinates on the basis
of 1202 NOE distance restraints, including 728 short-
range connectivities (residue i to residue I + j, where
1<j £ 4), 474 long-range connectivities (residue i to resi-
due I + j, where j > 4), 230 dihedral angle restraints,
and 23 hydrogen bond distance restraints. NOEs observed
at 100 ms were classified as strong, medium or weak
(< 2.8, 3.8, and 5.5 A
˚
, respectively) on the basis of peak
intensities calibrated internally using known distances. The
structures were analysed using molmol [55], pymol [56],
and procheck-nmr [57]. Structures were displayed using
molmol and pymol. The final family comprises the 20
structures of the lowest total energy from a total of 100
calculated structures; structure statistics are shown in

Table 1.
ITC experiments
The nickel ion solutions were prepared by diluting concen-
trated stocks of metal ions (100 mm) in a 20 mm Tris ⁄ HCl
buffer (pH 7.25). Metal ion titrations were performed at
298 K using a high-sensitivity CSC-4200 ITC microcalorim-
eter from Calorimetry Science (Lindon, UT, USA). Before
each ITC experiment, the pH of each solution was checked,
the reference cell was filled with deionized water and the
protein solution was degassed for 2–5 min to eliminate air
bubbles. Care was taken to start the first addition after
baseline stability had been achieved. Measurement from the
first injection was discarded from the analysis of the inte-
grated data, in order to avoid artefacts due to the diffusion
through the injection port occurring during the long equili-
bration period, locally affecting the protein concentration
near the syringe needle tip. In each titration, 10 lLofa
solution containing 200–600 lm NiCl
2
was injected into
a solution of SlyD (30–40 lm) in the same buffer, using a
computer-controlled 250 lL microsyringe. In order to allow
the system to reach equilibrium, a spacing of 400 s was
applied between each ligand injection. Heat produced by
nickel ion dilution was verified to be negligible by perform-
ing a control titration of NiCl
2
into the buffer alone, under
the same conditions. Integrated heat data obtained for the
titrations were fitted using a nonlinear least-squares

minimization algorithm to a theoretical titration curve,
using the software bindwork. DH° (reaction enthalpy
change in kJÆ mol
)1
), K
b
(binding constant in m
)1
) and n
(number of binding sites) were the fitting parameters. The
reaction entropy was calculated using the relationships
DG° = ) RT lnK
b
(R = 8.314 JÆmol
)1
ÆK
)1
, T = 298 K),
and DG° = DH°)TDS°.
The measurements were performed at two different final
Ni(II) ⁄ protein molar ratios: in a first set of experiments, a
final molar ratio of 6 : 1 was reached, whereas in the sec-
ond set, the measurements were repeated up to a 2 : 1
molar ratio to obtain a better characterization of the bind-
ing event centred on a stoichiometry of 1 : 1.
CD
CD spectra were recorded with a Jasco J-715 spectropola-
rimeter equipped with a Peltier-type temperature control
system (model PTC-348WI), and calibrated with an aque-
ous solution of 0.06% d-10-(1)-camphorsulfonic acid at

290 nm. Experiments were conducted in the same buffer
(20 mm Tris ⁄ HCl, 100 mm KCl, pH 7.25) and at same tem-
perature (298 K) used for the ITC measurements. The
molar ellipticity per mean residue, [h] (degÆcm
2
Ædmol
)1
), is
calculated from the equation h½¼h½
obs
ðmÞ=10 lC, where
[h]
obs
is the ellipticity (deg), m is the mean residue molecu-
lar weight (124 Da), C is the protein concentration
(gÆmL
)1
), and l is the optical path length of the cell (cm).
Cells with 0.1 cm path length and protein concentrations of
about 0.2 mgÆmL
)1
were used to record CD spectra
between 205 and 250 nm, and cells with 0.01 cm path
length and a protein concentration of 2 mgÆmL
)1
were used
to measure CD spectra between 190 and 250 nm. A time
constant of 16 s, a 2 nm bandwidth and a scan rate of
5nmÆmin
)1

were used to acquire the data. The spectra were
signal-averaged over at least five scans, and baseline
corrected by subtracting a buffer spectrum. To estimate the
secondary structure content, curve fitting was performed
using dichroweb [58].
The association between the protein and the nickel ions
was analysed by following the changes in molar ellipticity
at 215 nm ½h
215
upon nickel ion addition to a solution
containing the protein. The binding constant (K
b
) was
determined using the simplest binding model, assuming that
one molecule of ligand binds to one molecule of protein.
The molar ellipticity at 215 nm could therefore be defined
as the sum of the contributions of free and bound protein:
½h
215
¼½PÁ½h
P
þ½PLÁ½h
PL
ð1Þ
where [P] and [PL] represent the concentration of the free
and bound protein, respectively, and [h]
P
and [h]
PL
are

the corresponding molar ellipticity signals at 215 nm. The
concentration of the bound protein could be written
as the difference between the total concentration of
ligand in solution [L]
tot
and the fraction of unbound
ligand [L]:
½PL¼½L
tot
À½Lð2Þ
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4541
The constant of binding of the reaction K
b
is defined as:
K
b
¼
½PL
½PÁ½L
ð3Þ
The total concentrations of the protein [P]
tot
and the
nickel ions [L]
tot
per each CD measurement is given by:
½P
tot
¼½PÁð1 þ K

b
Á½LÞ ð4Þ
½L
tot
¼½LÁð1 þ K
b
Á½PÞ ð5Þ
If Eqn (4) is combined with Eqn (5), a quadratic expres-
sion is obtained that can be solved to express the concen-
tration of the free ligand, [L], in terms of the total
concentration of ligand and protein. This can be substituted
into Eqn (2) to give an analytical expression for [PL]:
If Eqns (4), (6) and (1) are combined, an expression of
[h]
215
as function of the molar ratio between the total
concentration of nickel ions and the total concentration of
the protein is obtained.
Acknowledgements
M. R. Conte is indebted to the Wellcome Trust for
financial support. The Centro Interdipartimentale di
Metodologie Chimico-Fisiche (CIMCF, University of
Naples ‘Federico II’) is gratefully acknowledged for
technical support for CD measurements. K. L. D.
Hands-Taylor is a recipient of an EPSRC-case PhD
studentship. E. R. Valentine acknowledges support
from a National Science Foundation International
Postdoctoral Research Fellowship, grant num-
ber OISE-0601986. We thank G. Mastroianni, F. Frat-
ernali, P. Brown, R. Tata and R. Hagan for help at

the initial stage of this project.
References
1 Maratea D, Young K & Young R (1985) Deletion and
fusion analysis of the phage phi X174 lysis gene E. Gene
40, 39–46.
2 Wulfing C, Lombardero J & Pluckthun A (1994) An
Escherichia coli protein consisting of a domain homolo-
gous to FK506-binding proteins (FKBP) and a new
metal binding motif. J Biol Chem 269, 2895–2901.
3 Roof WD & Young R (1995) Phi X174 lysis requires
slyD, a host gene which is related to the FKBP family
of peptidyl-prolyl cis-trans isomerases. FEMS Microbiol
Rev 17, 213–218.
4 Bolanos-Garcia VM & Davies OR (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.
5 Roof WD, Horne SM, Young KD & Young R (1994)
slyD, a host gene required for phi X174 lysis, is related
to the FK506-binding protein family of peptidyl-prolyl
cis-trans-isomerases. J Biol Chem 269, 2902–2910.
6 Brandts JF, Halvorson HR & Brennan M (1975) Con-
sideration of the possibility that the slow step in protein
denaturation reactions is due to cis-trans isomerism of
proline residues. Biochemistry 14, 4953–4963.
7 Balbach J & Schmid FX (2000) Prolyl Isomerization and
Its Catalysis in Protein Folding. Oxford University
Press, Oxford.
8 Schmid FX (2002) Prolyl isomerases. Adv Protein Chem

59, 243–282.
9 Hottenrott S, Schumann T, Pluckthun A, Fischer G &
Rahfeld JU (1997) The Escherichia coli SlyD is a metal
ion-regulated peptidyl-prolyl cis ⁄ trans-isomerase. J Biol
Chem 272, 15697–15701.
10 Scholz C, Eckert B, Hagn F, Schaarschmidt P, Balbach
J & Schmid FX (2006) SlyD proteins from different spe-
cies exhibit high prolyl isomerase and chaperone activi-
ties. Biochemistry 45, 20–33.
11 Scholz C, Thirault L, Schaarschmidt P, Zarnt T, Faatz
E, Engel AM, Upmeier B, Bollhagen R, Eckert B &
Schmid FX (2008) Chaperone-aided in vitro renatur-
ation of an engineered E1 envelope protein for detection
of anti-Rubella virus IgG antibodies. Biochemistry 47,
4276–4287.
12 Scholz C, Schaarschmidt P, Engel AM, Andres H,
Schmitt U, Faatz E, Balbach J & Schmid FX (2005)
Functional solubilization of aggregation-prone HIV
envelope proteins by covalent fusion with chaperone
modules. J Mol Biol 345, 1229–1241.
13 Han KY, Song JA, Ahn KY, Park JS, Seo HS & Lee J
(2007) Solubilization of aggregation-prone heterologous
proteins by covalent fusion of stress-responsive Escheri-
chia coli protein, SlyD. Protein Eng Des Sel 20, 543–549.
14 Graubner W, Schierhorn A & Bruser T (2007) DnaK
plays a pivotal role in Tat targeting of CueO and func-
tions beside SlyD as a general Tat signal binding chap-
erone. J Biol Chem 282, 7116–7124.
½PL¼½L
tot

À
À1 À K
b
ð½P
tot
À½L
tot
Þþ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð1 þ K
b
ð½P
tot
À½L
tot
ÞÞ
2
þ 4K
b
½L
tot
q
2K
b
ð6Þ
Structure and interactions of SlyD L. Martino et al.
4542 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS
15 Hesterkamp T, Hauser S, Lutcke H & Bukau B (1996)
Escherichia coli trigger factor is a prolyl isomerase that
associates with nascent polypeptide chains. Proc Natl

Acad Sci USA 93, 4437–4441.
16 Patzelt H, Ru
¨
diger S, Brehmer D, Kramer G, Vor-
derwu
¨
lbecke S, Schaffitzel E, Waitz A, Hesterkamp T,
Dong L, Schneider-Mergener J et al. (2001) Binding
specificity of Escherichia coli trigger factor. Proc Natl
Acad Sci USA 98, 14244–14249.
17 Knappe TA, Eckert B, Schaarschmidt P, Scholz C &
Schmid FX (2007) Insertion of a chaperone domain
converts FKBP12 into a powerful catalyst of protein
folding. J Mol Biol 368, 1458–1468.
18 Maruyama T & Furutani M (2000) Archaeal peptidyl
prolyl cis-trans isomerases (PPIases). Front Biosci 5,
D821–D836.
19 Maruyama T, Suzuki R & Furutani M (2004) Archaeal
peptidyl prolyl cis-trans isomerases (PPIases) update
2004. Front Biosci 9, 1680–1720.
20 Furutani M, Ideno A, Iida T & Maruyama T (2000)
FK506 binding protein from a thermophilic archaeon,
Methanococcus thermolithotrophicus, has chaperone-like
activity in vitro. Biochemistry 39, 453–462.
21 Barik S (2006) Immunophilins: for the love of proteins.
Cell Mol Life Sci 63, 2889–2900.
22 Zhang JW, Butland G, Greenblatt JF, Emili A & Zam-
ble DB (2005) A role for SlyD in the Escherichia coli
hydrogenase biosynthetic pathway. J Biol Chem 280,
4360–4366.

23 Leach MR, Zhang JW & Zamble DB (2007) The role
of complex formation between the Escherichia coli
hydrogenase accessory factors HypB and SlyD. J Biol
Chem 282, 16177–16186.
24 Zhang JW, Leach MR & Zamble DB (2007) The pept-
idyl-prolyl isomerase activity of SlyD is not required for
maturation of Escherichia coli hydrogenase. J Bacteriol
189, 7942–7944.
25 Leach MR, Sandal S, Sun H & Zamble DB (2005) Metal
binding activity of the Escherichia coli hydrogenase
maturation factor HypB. Biochemistry 44, 12229–12238.
26 Dias AV, Mulvihill CM, Leach MR, Pickering IJ,
George GN & Zamble DB (2008) Structural and
biological analysis of the metal sites of Escherichia coli
hydrogenase accessory protein HypB. Biochemistry 47,
11981–11991.
27 Gasper R, Scrima A & Wittinghofer A (2006)
Structural insights into HypB, a GTP-binding protein
that regulates metal binding. J Biol Chem 281,
27492–27502.
28 Rey L, Imperial J, Palacios JM & Ruiz-Argueso T
(1994) Purification of Rhizobium leguminosarum HypB,
a nickel-binding protein required for hydrogenase
synthesis. J Bacteriol 176
, 6066–6073.
29 Fu C, Olson JW & Maier RJ (1995) HypB protein of
Bradyrhizobium japonicum is a metal-binding GTPase
capable of binding 18 divalent nickel ions per dimer.
Proc Natl Acad Sci USA 92, 2333–2337.
30 Olson JW, Fu C & Maier RJ (1997) The HypB protein

from Bradyrhizobium japonicum can store nickel and is
required for the nickel-dependent transcriptional regula-
tion of hydrogenase. Mol Microbiol 24, 119–128.
31 Weininger U, Haupt C, Schweimer K, Graubner W,
Kovermann M, Bruser T, Scholz C, Schaarschmidt P,
Zoldak G, Schmid FX et al. (2009) NMR solution
structure of SlyD from Escherichia coli: spatial separa-
tion of prolyl isomerase and chaperone function. J Mol
Biol 387, 295–305.
32 Suzuki R, Nagata K, Yumoto F, Kawakami M, Nem-
oto N, Furutani M, Adachi K, Maruyama T & Tanok-
ura M (2003) Three-dimensional solution structure of
an archaeal FKBP with a dual function of peptidyl
prolyl cis–trans isomerase and chaperone-like activities.
J Mol Biol 328, 1149–1160.
33 Warren JJ & Moore PB (2001) A maximum likelihood
method for determining D(a)(PQ) and R for sets of
dipolar coupling data. J Magn Reson 149, 271–275.
34 Michnick SW, Rosen MK, Wandless TJ, Karplus M &
Schreiber SL (1991) Solution structure of FKBP, a ro-
tamase enzyme and receptor for FK506 and rapamycin.
Science 252, 836–839.
35 Moore JM, Peattie DA, Fitzgibbon MJ & Thomson JA
(1991) Solution structure of the major binding protein
for the immunosuppressant FK506. Nature 351, 248–
250.
36 Van Duyne GD, Standaert RF, Karplus PA, Schreiber
SL & Clardy J (1993) Atomic structures of the human
immunophilin FKBP-12 complexes with FK506 and
rapamycin. J Mol Biol 229, 105–124.

37 Zoldak G, Carstensen L, Scholz C & Schmid FX (2009)
Consequences of domain insertion on the stability and
folding mechanism of a protein. J Mol Biol 386, 1138–
1152.
38 Sekerina E, Rahfeld JU, Muller J, Fanghanel J, Ra-
scher C, Fischer G & Bayer P (2000) NMR solution
structure of hPar14 reveals similarity to the peptidyl
prolyl cis ⁄ trans isomerase domain of the mitotic regula-
tor hPin1 but indicates a different functionality of the
protein. J Mol Biol 301, 1003–1017.
39 Vogtherr M, Jacobs DM, Parac TN, Maurer M, Pahl
A, Saxena K, Ruterjans H, Griesinger C & Fiebig KM
(2002) NMR solution structure and dynamics of the
peptidyl-prolyl cis–trans isomerase domain of the trig-
ger factor from Mycoplasma genitalium compared to
FK506-binding protein. J Mol Biol 318, 1097–1115.
40 Kay JE (1996) Structure–function relationships in the
FK506-binding protein (FKBP) family of peptidylprolyl
cis-trans isomerases. Biochem J 314 (Pt 2), 361–385.
41 Ikura T, Kinoshita K & Ito N (2008) A cavity with an
appropriate size is the basis of the PPIase activity.
Protein Eng Des Sel 21, 83–89.
L. Martino et al. Structure and interactions of SlyD
FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS 4543
42 Wiseman T, Williston S, Brandts JF & Lin LN (1989)
Rapid measurement of binding constants and heats of
binding using a new titration calorimeter. Anal Biochem
179, 131–137.
43 Freire E, van Osdol WW, Mayorga OL & Sanchez-Ruiz
JM (1990) Calorimetrically determined dynamics of

complex unfolding transitions in proteins. Annu Rev
Biophys Biophys Chem 19, 159–188.
44 Grossoehme NE, Mulrooney SB, Hausinger RP &
Wilcox DE (2007) Thermodynamics of Ni2+, Cu2+,
and Zn2+ binding to the urease metallochaperone
UreE. Biochemistry 46, 10506–10516.
45 Zambelli B, Danielli A, Romagnoli S, Neyroz P, Ciurli S
& Scarlato V (2008) High-affinity Ni2+ binding selec-
tively promotes binding of Helicobacter pylori NikR to
its target urease promoter. J Mol Biol 383, 1129–1143.
46 Go
¨
thel SF & Marahiel MA (1999) Peptidyl-prolyl cis–
trans isomerases, a superfamily of ubiquitous folding
catalysts. Cell Mol Life Sci 55, 423–436.
47 Fischer G & Aumuller T (2003) Regulation of peptide
bond cis ⁄ trans isomerization by enzyme catalysis and
its implication in physiological processes. Rev Physiol
Biochem Pharmacol 148, 105–150.
48 Lu KP, Finn G, Lee TH & Nicholson LK (2007) Prolyl
cis–trans isomerization as a molecular timer. Nat Chem
Biol 3, 619–629.
49 Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J
& Bax A (1995) NMRPipe: a multidimensional spectral
processing system based on UNIX pipes. J Biomol
NMR 6, 277–293.
50 Bartels C, Xia T, Billeter M, Gu
¨
ntert P & Wu
¨

thrich K
(1995) The program XEASY for computer supported
NMR spectral-analysis of biological macromolecules.
J Biomol NMR 6, 1–10.
51 Cornilescu G, Delaglio F & Bax A (1999) Protein back-
bone angle restraints from searching a database for
chemical shift and sequence homology. J Biomol NMR
13, 289–302.
52 Ru
¨
ckert M & Otting G (2000) Alignment of biological
macromolecules in novel non-ionic liquid crystalline
media for NMR experiments. J Am Chem Soc 122,
7793–7797.
53 Tjandra N, Grzesiek S & Bax A (1996) Magnetic field
dependence of nitrogen-proton J splitting in
15
N-
enriched human ubiquitin resulting from relaxation
interference and residual dipolar coupling. J Am Chem
Soc 118, 6264–6272.
54 Bru
¨
nger AT (1993) X-PLOR Manual, Version 3.1. Yale
University, New Haven, CT.
55 Koradi R, Billeter M & Wuthrich K (1996) MOLMOL:
a program for display and analysis of macromolecular
structures. J Mol Graph 14, 51–55, 29–32.
56 Merritt EA & Bacon DJ (1997) Raster3D: photorealis-
tic molecular graphics. Methods Enzymol 277 , 505–524.

57 Laskowski RA, Rullmannn JA, MacArthur MW,
Kaptein R & Thornton JM (1996) AQUA and PRO-
CHECK-NMR: programs for checking the quality of
protein structures solved by NMR. J Biomol NMR 8,
477–486.
58 Whitmore L & Wallace BA (2004) DICHROWEB, an
online server for protein secondary structure analyses
from circular dichroism spectroscopic data. Nucleic
Acids Res 32, W668–W673.
Supporting information
The following supplementary material is available:
Fig. S1. Structural traces of full-length SlyD.
Fig. S2. Electrostatic surface potential for SlyD.
Fig. S3. Interactions of SlyD with FK506 and rapamy-
cin.
Fig. S4. NMR titration experiments of SlyD with
Ni(II) and table of chemical shift perturbations of
SlyD upon interaction with Ni(II).
This supplementary material can be found in the
online article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Structure and interactions of SlyD L. Martino et al.
4544 FEBS Journal 276 (2009) 4529–4544 ª 2009 The Authors Journal compilation ª 2009 FEBS