A Histidine-rich Metal Binding Domain at the N Terminus of
Cu,Zn-Superoxide Dismutases from Pathogenic Bacteria
A NOVEL STRATEGY FOR METAL CHAPERONING*
Andrea
Battistoni
§,
Francesca
Pacello,
Anna Paola
Mazzetti,
Concetta
Capo,
J. Simon
Kroll¶,
Paul R.
Langford¶,
Assunta
Sansone¶,
Giovanna
Donnarumma
,
Piera
Valenti, and
Giuseppe
Rotilio
From the Dipartimento di Biologia, Università
di Roma "Tor Vergata," Via della Ricerca Scientifica,
00133 Rome, Italy, ¶ Molecular Infectious Diseases Group,
Department of Paediatrics, Imperial College School of Medicine, St.
Mary's Hospital, Norfolk Place, London W2 1PG, United Kingdom, and
Istituto di Microbiologia-II, Università di Napoli,
Larghetto S. Aniello a Caponapoli, 2, 80138 Napoli, Italy
Received for publication, November 21, 2001, and in revised form, May 18, 2001
 |
ABSTRACT |
A group of Cu,Zn-superoxide dismutases from
pathogenic bacteria is characterized by histidine-rich N-terminal
extensions that are in a highly exposed and mobile conformation. This
feature allows these proteins to be readily purified in a single step by immobilized metal affinity chromatography. The Cu,Zn-superoxide dismutases from both Haemophilus ducreyi and
Haemophilus parainfluenzae display anomalous absorption
spectra in the visible region due to copper binding at the N-terminal
region. Reconstitution experiments of copper-free enzymes demonstrate
that, under conditions of limited copper availability, this metal ion
is initially bound at the N-terminal region and subsequently
transferred to an active site. Evidence is provided for intermolecular
pathways of copper transfer from the N-terminal domain of an enzyme
subunit to an active site located on a distinct dimeric molecule.
Incubation with EDTA rapidly removes copper bound at the N terminus but
is much less effective on the copper ion bound at the active site. This
indicates that metal binding by the N-terminal histidines is
kinetically favored, but the catalytic site binds copper with higher
affinity. We suggest that the histidine-rich N-terminal region
constitutes a metal binding domain involved in metal uptake under
conditions of metal starvation in vivo. Particular
biological importance for this domain is inferred by the observation
that its presence enhances the protection offered by periplasmic
Cu,Zn-superoxide dismutase toward phagocytic killing.
| |
INTRODUCTION |
Cu,Zn-superoxide dismutases
(Cu,Zn-SODs)1 (encoded in
bacteria by sodC gene(s)) are metalloenzymes that catalyze
the dismutation of the superoxide anion into oxygen and hydrogen
peroxide by the alternate reduction and oxidation of a copper ion that
constitutes the catalytically active redox center (1). Although for a
long time Cu,Zn-SODs have been considered almost exclusively eukaryotic cytosolic enzymes, more recent studies have established that this enzyme is present in a large number of bacterial species, where it is
exported to extracytoplasmic compartments (2). Since the first step in
the phagocytic oxidative burst is the single electron reduction of
molecular oxygen to superoxide by a transmembrane NADPH-oxidase complex
(3), and superoxide is unable to cross the cytoplasmic membrane (4), it
has been proposed that Cu,Zn-SOD protects bacteria from oxygen free
radicals generated by inflammatory cells and facilitates bacterial
survival within the host. This hypothesis has gained support from
different studies that have shown that sodC mutants
of Salmonella typhimurium (5-8), Neisseria meningitidis (9), and Haemophilus ducreyi (10) are
attenuated in animal models and by the demonstration that Cu,Zn-SOD
protects bacteria from macrophage killing (6, 11) or extracellular superoxide generated in vitro (5, 6, 9, 12).
The discovery that, at least in some bacterial species, Cu,Zn-SOD is
able to modulate virulence has suggested that this enzyme could
represent a target for novel antimicrobial strategies and has
stimulated studies on its structural and functional properties. In this
context the x-ray structure of different bacterial Cu,Zn-SODs have been
solved (13-17), and the spectroscopic and catalytic properties of some
of these enzymes have been investigated (17-21). These studies have
shown that prokaryotic and eukaryotic Cu,Zn-SODs derive from a common
ancestor gene and share a similar three-dimensional fold, based on a
flattened Greek-key eight-stranded 
-barrel and a similar
organization of the redox center. In the oxidized form the copper ion
is coordinated by the nitrogen atoms of four histidine residues; the
zinc ion is coordinated by three histidines and an aspartic acid
residue, and the two metal ions are simultaneously coordinated by a
single histidine residue (termed the "histidine bridge" or the
"bridging imidazolate"), in a structural motif so far found only in
Cu,Zn-SODs (22, 23). Relevant differences are, however, observed in the
organization of the active site channel and in the way subunits are
assembled. In particular, these studies have highlighted that bacterial
Cu,Zn-SOD may be monomeric or dimeric (15) and that small differences
at the dimer interface may finely modulate the enzyme activity and
stability (17).
Interestingly, whereas all eukaryotic Cu,Zn-SODs conform to a single
structural model that appears to have been strictly preserved throughout evolution (24), analysis of amino acid sequences from
Cu,Zn-SODs of different bacterial species suggests much greater variation, so individual enzyme variants may exhibit unique properties (15, 16). The most obvious differences between bacterial Cu,Zn-SODs include insertions and deletions in some of the major loops protruding from the -barrel, which could plausibly result in differences in the
active site channel architecture and subunit assembly, and the
substitutions of some of the conserved metal ligands that are expected
to affect significantly the enzyme activity. The functional
implications of these variations are still to be explored, but it is
likely that such differences may lead to modulation in the enzyme
activity in different bacteria.
In general, the bacterial Cu,Zn-SODs are least alike at their
N-terminal ends. In this study, we have investigated the metal binding
ability of N-terminal extensions present in a subset of Cu,Zn-SODs from
Gram-negative pathogenic bacteria. We have found that such extensions,
which are in a highly mobile conformation and contain several histidine
residues, bind divalent metal ions with high efficiency. We propose
that they represent high affinity functional metal binding domains,
perhaps involved in the uptake of the prosthetic metals of the enzyme
in environments where their concentration is very low.
| |
EXPERIMENTAL PROCEDURES |
Construction of Expression Plasmids--
Plasmids pIT4, pJSK130,
and pJSK40, bearing the sodC genes from H. ducreyi, Haemophilus parainfluenzae, and
Haemophilus influenzae, respectively, have been described
previously (25, 26). In order to ensure similar expression levels of
the different enzyme variants, all sequences encoding mature Cu,Zn-SODs
were cloned in plasmid pHEN-1 (27), under control of the
lacZ promoter. Plasmid pPHduSOD1, expressing wild type
H. ducreyi Cu,Zn-SOD, was obtained by PCR amplification with
the oligonucleotides HduFor1 (5'-CTAAGCTTAAGGAGATAAAATGAAATTA) and
HduRev (5'-CCTGAATTCTTATTTAATTACACCGCATGCC), using pIT4 as a DNA
template and the high fidelity polymerase ExpandTM (Roche
Molecular Biochemicals). The amplified DNA was digested with
HindIII and EcoRI and cloned in the corresponding
sites of pHEN-1. Plasmid pJSK326, expressing a truncated form of
H. ducreyi Cu,Zn-SOD (N-terminal deleted mutant)
lacking the first 22 amino acids (HGDHMHNHDTKMDTMSKDMMSM), was
obtained by a similar strategy, amplifying the coding sequence with
primers HduFor2 (5'-TTCCATGGCCGAAAAAATTGTAGTGCCT) and HduRev and
subsequently cloning the amplified DNA in the NcoI and
EcoRI sites of pHEN1. To construct plasmid pPHduSODFIRS,
expressing a mutant H. ducreyi enzyme lacking residues
11-22, the sequence encoding residues 1-10 was amplified with primers
HduFor1 and HduRevS (5'-GACCATGGTGTCATGATTATGCATA). The PCR product was
restricted with HindIII and NcoI and inserted in
the corresponding sites of plasmid pJSK326. Plasmids pPHpSOD and
pPHiSOD expressing wild type Cu,Zn-SODs from H. parainfluenzae and H. influenzae, respectively, were
constructed as follows. The sequences coding for the two mature enzymes
were amplified with primers HiFor
(5'-AACCATGGCCCATGACCATATGGCAAAAC) and HiRev
(5'-CTGAATTCTTATTTAATCACGCCACATGC), using plasmids pJSK130 and
pJSK40 as templates. The amplified DNA fragments were digested with
NcoI and EcoRI. Since there is an internal
NcoI site in these two sodC genes, the resulting
NcoI-NcoI (279 base pairs) and
NcoI-EcoRI (216 base pairs) DNA fragments were
separated by agarose gel electrophoresis, and the
NcoI-EcoRI fragment from H. parainfluenzae (identical to the corresponding DNA region of the
H. influenzae sodC gene) was cloned in plasmid pHEN-1, thus
obtaining plasmid pPHaeShort. Subsequently, the
NcoI-NcoI fragments from the two distinct
sodC genes were inserted in the NcoI site of
pPHaeShort. To construct plasmids pPHpSOD-3 and pPHiSOD-3, expressing
mutant proteins lacking the first three amino acid residues (HDH), DNA
was amplified with primers hemeFor2 (5'-AACCATGGCTATGGCAAAACCAGCAGGT)
and HiRev, using pPHpSOD and pPHiSOD as DNA templates,
respectively. After digestion with NcoI, the 270-base pair
NcoI-NcoI DNA fragments were purified and
inserted in the NcoI site of plasmid pPHaeShort. Plasmids
pPEcSODHis and pPXSODBHis, expressing Escherichia coli and
Xenopus laevis Cu,Zn-SODs fused to the 22 N-terminal
residues from H. ducreyi Cu,Zn-SODs, were obtained by
inserting a DNA fragment obtained by amplification with HduFor1 and
HduRevL into the HindIII and NcoI sites of
pPEcSOD81A (28) and pPXSODB (29). The complete nucleotide sequences of
all the PCR amplified DNA fragments were checked by the dideoxy chain
termination method.
Cu,Zn-SODs Expression and Purification--
Overexpression of
the different Cu,Zn-SODs was carried out in 71/18 E. coli
cells (30) grown at 37 °C in Luria Bertani broth (LB) containing 100 µg/ml ampicillin. When cells reached an A600 of 0.5, cultures were supplemented with 0.2 mM
isopropyl--D-thiogalactopyranoside, 0.25 mM
CuSO4, and 10 µM ZnSO4. The
periplasmic fraction was prepared as described previously (19). Wild
type and mutant enzymes were purified by immobilized metal affinity
chromatography on a nickel-nitrilotriacetic acid (NTA) resin (Qiagen).
The column was equilibrated with a buffer containing 300 mM
NaCl and 50 mM sodium phosphate, pH 7.8 (buffer A).
Proteins bound to the resin were eluted by washing the column
sequentially with 2-4 volumes of buffer A supplemented with 10, 20, 40, 80, and 250 mM imidazole. Fractions were analyzed by
denaturing polyacrylamide gel electrophoresis (SDS-PAGE). The same
procedure was used to test the affinity of Cu,Zn-SOD for different
metal ions on an iminodiacetic acid (IDA) resin, charged with nickel,
copper, zinc, cadmium, or cobalt according to the manufacturer's
instructions (Amersham Pharmacia Biotech). Cu,Zn-SOD samples purified
by immobilized metal affinity chromatography were analyzed for their
metal content using the PerkinElmer Life Sciences spectrometer
AAnalyst 300 equipped with the graphite furnace HGA-800. No
significant nickel contamination was found in Cu,Zn-SODs purified from
Ni-NTA, whereas the enzymes purified from IDA exhibited severe metal contamination.
After affinity purification on Ni-NTA columns, samples containing wild
type Cu,Zn-SODs (usually >90% pure) were concentrated and subjected
to gel filtration chromatography on a HiLoadTM 16/60
SuperdexTM FPLC column (Amersham Pharmacia Biotech). Wild
type H. ducreyi Cu,Zn-SOD purified from 71/18 was not
suitable for a spectral analysis of copper binding (31). Therefore, for
this purpose, the enzyme was expressed in H500 E. coli cells
(32) and purified by the procedure described above.
To purify truncated proteins lacking N-terminal histidines (obtained
from E. coli 71/18 cells harboring pJSK326 or pPHpSOD-3), periplasmic extracts were concentrated and dialyzed against 10 mM potassium phosphate, pH 7.0, and fractionated onto a
Whatman DE52 column equilibrated with the same buffer. Fractions
containing Cu,Zn-SOD were pooled, concentrated, and dialyzed against 20 mM Tris-HCl, pH 7.0, 0.15 M NaCl and injected
onto a High Load 16/60 Superdex 75 gel filtration FPLC column and
eluted with the same buffer. As a final step, the enzymes were further
dialyzed against 20 mM potassium phosphate, pH 6.5, and
subjected to ion exchange chromatography on a Mono-S HR 5/5 FPLC column
(Amersham Pharmacia Biotech) equilibrated with the same buffer, using a
0-0.1 M NaCl gradient. Whereas the H. parainfluenzae mutant eluted in a single peak containing the pure
enzyme, the H. ducreyi mutant eluted in two major peaks, one
of which corresponded to the heme-containing enzyme (31). This
Cu,Zn-SOD form was discarded, and the heme-lacking protein was
subjected to a second round of ion exchange chromatography under
identical conditions.
All purified proteins were concentrated and dialyzed against 10 mM potassium phosphate buffer, pH 7.0. Protein
concentration was determined by the Lowry method (33). The zinc and
copper content of purified Cu,Zn-SOD samples was verified by atomic
absorption. The metal content of enzymes purified from metal enriched
media was variable from one preparation to another. On average, the copper content of wild type H. parainfluenzae and H. ducreyi enzymes was between 1 and 2.5 copper ions/subunit, whereas
zinc content was between 1.3 and 1.8 zinc ions/subunit. The metal
content of these samples was diminished by treatment with EDTA
(followed by dialysis), indicating that a significant fraction of these metal ions is loosely bound on the enzyme surface. Mutant enzymes contained lower amounts of both metals.
Limited Proteolysis and N-terminal Sequence
Analysis--
H. ducreyi Cu,Zn-SOD (0.25 mg/ml) was
dissolved in 50 mM potassium phosphate buffer, pH 8.0, and
incubated at 20 °C in the presence of increasing amounts of
proteinase K (proteinase K/Cu,Zn-SOD ratios were between 0.01 and 1%).
At different times aliquots were withdrawn, and the proteolytic
digestion was stopped by boiling the sample in 2% SDS, 5%
mercaptoethanol. A similar procedure was used for trypsin digestion,
but in this case the reaction was carried out in ammonium bicarbonate
buffer at 37 °C. Samples from limited proteolysis experiments were
subjected to SDS-PAGE on 15% gels and electrotransferred onto
polyvinylidene difluoride membranes (Immobilon P, Millipore). Protein
bands were visualized by Coomassie Blue staining, excised from the
membrane, and subjected to several cycles of automated Edman
degradation performed with a model 473A pulsed liquid sequencer
(Applied Biosystems) with an on-line analyzer of phenylthiohydantoin
amino acids.
Preparation of the Copper-free Enzyme and Copper Reconstitution
Experiments--
Copper-free Cu,Zn-SODs were prepared following a
procedure described previously (34). Briefly, the enzymes were treated with excess potassium ferrocyanide to reduce copper and then dialyzed for 12-24 h at 4 °C against 0.1 M potassium phosphate
buffer, 50 µM KCN, pH 6.0. The samples were further
dialyzed twice for 24 h at 4 °C against phosphate buffer to
remove KCN. The metal content of copper-free proteins was evaluated by
atomic absorption spectroscopy. The final copper content of samples
used for reconstitution experiments contained between 0.3 and 0.6% of
the initial value, whereas no significant loss of zinc was observed.
Reconstitution experiments were carried out by adding stoichiometric or
sub-stoichiometric amounts of CuCl2 to the copper-free SOD
dissolved in water at 25 or 37 °C. The reconstitution process was
monitored spectrophometrically by registering protein absorption in the
visible region as a function of time with a PerkinElmer Life Sciences
Lambda 2 spectrophotometer thermostated at 25 or 37 °C, using 1-cm
path length cuvettes. In these experiments Cu,Zn-SODs were dissolved at
a concentration of 2.5 × 10
4
M.
The rate of copper binding at the active site was also evaluated by
measuring the increase in catalytic activity following the addition of
copper to copper-free Cu,Zn-SODs dissolved in water. Reconstitution
experiments at copper/subunit ratios below or equal to 1 were carried
out by adding variable amounts of proteins to a 5.8 × 107 M copper solution.
Reconstitution experiments at higher copper/subunit ratios were
obtained by adding the proteins (at a final concentration of 2.9 × 107 M) to copper solutions of
higher concentrations. The activity assays were carried out by the
pyrogallol method (35). To rule out a possible contribution to
enzyme activity of copper bound outside the active site, 0.1 mM EDTA was added to the assay buffer. The percent activity
values were obtained by the ratio between the activity of reconstituted
enzymes and that of an equal amount of fully reconstituted Cu,Zn-SODs.
All the activity values obtained during reconstitution experiments were
corrected by subtracting the contribution to activity of residual
copper present in copper-free proteins. Fully metallated H. parainfluenzae and mutant H. parainfluenzae and
H. ducreyi enzymes showed nearly identical activity values. An approximate activity value for the fully reconstituted wild type
H. ducreyi enzyme was obtained using as a standard the
reconstituted N-terminal deleted mutant.
Macrophage Killing Experiments--
The mouse macrophage-like
cell line J774 and the human macrophage-like line THP-1 were grown in
Dulbecco's modified Eagle's medium supplemented with 10% fetal calf
serum in an atmosphere of 5% CO2. 3 h before the
experiments the THP-1 cell suspension was supplemented with 1 µg/ml
phorbol 12-myristate 13-acetate. Macrophage cell concentration was
adjusted to 2 × 107 cells/ml for killing experiments.
E. coli GC4468 cells (36) expressing either wild type or
mutant (devoid of N-terminal histidines) H. parainfluenzae
and H. ducreyi Cu,Zn-SODs were grown overnight at 37 °C
in LB medium or in LB supplemented with 0.25 mM
CuSO4 and 10 µM ZnSO4. After growth cells were washed three times in phosphate-buffered solution without Ca2+ and Mg2+. The bacterial cell
concentration was then adjusted to 1 × 108 cells/ml
in phosphate-buffered solution without Ca2+ and
Mg2+.
Before presentation to macrophages, microorganisms were pre-opsonized
by exposing 2 × 107 bacteria (200 µl) to 50 µl of
fresh pooled serum plus 750 µl of phosphate-buffered saline (pH 7.4),
for 25 min at 37 °C with moderate agitation. 200 µl (~2 × 106) of opsonized bacteria were added to a mixture of 5%
fresh serum (40 µl), phosphate-buffered solution (560 µl), and 200 µl of culture medium with or without macrophages (4 × 106). 100-µl samples were withdrawn at 0, 30, 60, and 90 min in order to determine bacterial survival in serum and in
macrophages by counting colony-forming units in trypticase soy agar
(Oxoid). The number of surviving bacteria was obtained from at least
three independent experiments. In each experiment, the macrophage
killing assay was done in triplicate.
| |
RESULTS |
Histidine-rich N-terminal Sequences in Bacterial
Cu,Zn-SODs--
Fig. 1 shows an
alignment of the amino acid sequences of 10 Cu,Zn-SODs from
Gram-negative bacteria, which includes the four enzymes whose
three-dimensional structures have been solved and a few other enzyme
variants that are expected to possess a similar structure (15). Some of
these enzymes, namely those from the mucosal colonists and pathogens
H. ducreyi, H. parainfluenzae, H. influenzae, N. meningitidis, and Actinobacillus
pleuropneumoniae, show two features that were ignored during
previous analyses; they are characterized by an unusually large number
of histidine residues (which in the case of the H. ducreyi
and A. pleuropneumoniae enzymes approximates the 10% of
total residues) and by amino acid extensions at their N-terminal
sequence. These N-terminal extensions are of variable length and
display low sequence identity, but all of them initiate with a cluster
of histidines (2-4 residues) interspersed with negatively charged
residues. Similar His-rich regions have been found in several other
proteins able to bind nickel, zinc, or other transition metals. These
proteins include UreE from Klebsiella aerogenes (37), HypB
from Rhizobium leguminosarum (38), CooJ from
Rhodospirillum rubrum (39), SlyD from E. coli (40), ZnuA from different bacteria (Ref. 41 and references therein),
CzrB from Staphylococcus aureus (42), and Zrc1 from Saccharomyces cerevisiae (43). For some of these proteins,
it has been shown that the His-rich region is directly involved in metal binding, suggesting that the N-terminal sequence found in some
bacterial Cu,Zn-SOD may represent a protein domain able to bind
divalent metals.
View larger version (54K):
[in this window]
[in a new window]
|
Fig. 1.
Multiple alignment of the amino acid
sequences of 10 Cu,Zn-SODs from Gram-negative bacteria. The global
alignment has been deduced from optimal superimposition of available
bacterial Cu,Zn-SOD structures, as described previously (15). The
secondary structure elements are shown above the amino acid
sequences, using the following symbols: -strands,  ;
310 helix, G. The histidine residues present in
these bacterial Cu,Zn-SODs are shown in bold and are
underlined, and residues coordinating Cu2+ and
Zn2+ are indicated by C and Z,
respectively. Other known Cu,Zn-SOD sequences from Gram-negative
bacteria (Caulobacter crescentus, Legionella
pneumophila, and Francisella tularensis) have not been
included in the alignment because they show significant variations in
the primary structure which suggest a slightly modified enzyme
architecture (15) and do not possess N-terminal extensions rich in
histidine residues.
|
|
Protease Sensitivity of the N-terminal Extension of Bacterial
Cu,Zn-SODs--
Cu,Zn-SODs are in general very stable enzymes, highly
resistant to protease digestion. However, we found that preparations of
the H. ducreyi enzyme underwent rapid fragmentation,
accumulating a major form with a molecular mass (~16 kDa) comparable
to that of the enzyme produced by cells bearing pJSK326, which lacks
the initial 22 amino acids. It has been found previously that, due to
presumed proteolytic cleavage, the A. pleuropneumoniae
Cu,Zn-SOD used in a recent crystallographic study lacked the seven
initial amino acids and that the following residues were in a
disordered conformation (16). Since proteolytic enzymes usually cleave the polypeptide chain within highly flexible regions and not in rigid
structural elements (44), these observations suggest that the
N-terminal extension of bacterial Cu,Zn-SODs could be in an exposed and
mobile conformation. To probe the conformational state of its
N-terminal region, the H. ducreyi enzyme was subjected to
limited proteolysis experiments with proteinase K and trypsin. Fig.
2A shows the SDS-PAGE analysis
of H. ducreyi Cu,Zn-SOD digested with proteinase K as a
function of proteinase K/Cu,Zn-SOD ratio. The N-terminal sequence of
each digestion product was determined in order to map the sites of
higher accessibility. The N-terminal sequence of band a is Ser-Lys-Asp,
indicating that initial cleavage occurs within the N-terminal domain
and removes 15 amino acids from the enzyme. Band b, which appears after
digestion with large amounts of proteinase K, is a mixture of three
distinct protease digestion products whose amino acid sequences begin
with Val-Gly-Thr, Glu-Ser-Ala, and Tyr-Gly-Leu, respectively, and
derive from cleavages within the first -strand elements of the
enzyme. Incubation with trypsin (Fig. 2B) led to the
appearance of two digestion products (bands c and d), which were both
mapped within the N-terminal domain. The N-terminal sequence of band d
is Asp-Met-Met. Such a sequence indicates that the proteolytic attack
occurred at a position very close to the site initially recognized by
proteinase K, leading to the removal of 17 residues. Band c sequence
starts with Met-Asp-Thr and is due to a trypsin cleavage after the
first lysine residue in the polypeptide.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 2.
Limited proteolysis of H. ducreyi
Cu,Zn-SOD. A, SDS-PAGE analysis of proteolytic
fragments obtained after digestion with proteinase K. Proteolysis was
carried out for 10 min at the following proteinase K/Cu,Zn-SOD ratios
(w/w): lane 1, 0; lane 2, 0.01; lane
3, 0.05; lane 4, 0.1; lane 5, 0.2;
lane 6, 0.5; and lane 7, 1. Lane
8, molecular weight markers. B, SDS-PAGE analysis of
proteolytic fragments obtained after digestion with trypsin.
Proteolysis was carried out for 15 min at the following
trypsin/Cu,Zn-SOD ratios (w/w): lane 1, 0; lane
2, 0.01; lane 3, 0.05; lane 4, 0.1; lane 5, 0.2; lane 6, 0.5; and lane
7, 1. Lane 8, molecular weight markers.
|
|
It is noteworthy that the sites and rates of Cu,Zn-SOD cleavage by
trypsin and proteinase K were neither affected by preincubation of the
enzyme with zinc nor by preincubation with chelating agents, suggesting
that metal binding does not confer significant rigidity to the
N-terminal domain.
Such experimental results are consistent with a theoretical analysis of
H. ducreyi Cu,Zn-SOD stability carried out by the method
described by Guruprasad and co-workers (45) which is based on
the occurrence of specific dipeptides in stable and unstable proteins.
According to this analysis, the enzyme was found to have a high
instability index due to the presence in the N-terminal region of
several dipeptides typical of unstable proteins (Met-His, Met-Ser, and
Lys-Met).
The His-rich Region Confers High Affinity for Divalent Metal
Ions--
The metal binding ability of the His-rich regions of the
Cu,Zn-SODs from H. ducreyi, H. parainfluenzae,
and H. influenzae was tested by immobilized metal affinity
chromatography. The three wild type enzymes (Fig.
3, A, F, and G) are
all characterized by high affinity for Ni-NTA and may be purified
nearly to homogeneity in a single chromatographic step, starting from
periplasmic extracts. The removal of the whole N-terminal extension (22 amino acids) from the H. ducreyi enzyme drastically reduces
its affinity for the resin (Fig. 3B). In contrast, a mutant
enzyme that retains the first 10 residues (including the four
N-terminal histidines), but lacks residues 11-22, possesses a metal
affinity comparable to that of the wild type protein (Fig.
3C), thus indicating that residues 11-22 do not
significantly contribute to the enzyme affinity for the metal chelate
resin. Similarly, mutant Cu,Zn-SODs from H. parainfluenzae
(Fig. 3H) and H. influenzae (not shown) lacking the first three residues (His-Asp-His) display low affinity for nickel.
Fusion of the H. ducreyi N-terminal region to the Cu,Zn-SODs from E. coli (Fig. 3E) and X. laevis
(not shown) confers high affinity for Ni-NTA also to these two enzymes.
It should be noted, however, that the N-terminal deleted H. ducreyi Cu,Zn-SOD (Fig. 3B) shows a greater affinity
for Ni-NTA than the E. coli (Fig. 3D) or mutant
H. influenzae and H. parainfluenzae enzymes,
indicating that other surface-exposed histidines also contribute to its
affinity for metal chelate resins.
View larger version (107K):
[in this window]
[in a new window]
|
Fig. 3.
Purification of Cu,Zn-SODs by immobilized
metal affinity chromatography. Periplasmic extracts (1 ml) from
E. coli cells producing different Cu,Zn-SODs were loaded
onto a Ni-NTA column (formed by 1 ml of Qiagen resin). Proteins were
eluted by serial washing of the column with buffers (4 ml) containing
increasing amounts of imidazole and analyzed by SDS-PAGE. The
eight panels show purification of the following enzymes:
A, H. ducreyi Cu,Zn-SOD; B, mutant
H. ducreyi Cu,Zn-SOD lacking the initial 22 residues;
C, mutant H. ducreyi Cu,Zn-SOD lacking residues
11-22; D, E. coli Cu,Zn-SOD; E,
mutant E. coli Cu,Zn-SOD containing the H. ducreyi His-rich domain at the N terminus; F, H. parainfluenzae Cu,Zn-SOD; G, H. influenzae
Cu,Zn-SOD; H, mutant H. parainfluenzae Cu,Zn-SOD
lacking the His-rich region. Lane 1, molecular weight
markers; lane 2, periplasmic extract; lane 3, flow-through; lane 4, 10 mM wash; lane
5, 20 mM wash; lane 6, 40 mM
wash; lane 7, 80 mM wash; and lane 8, 250 mM wash.
|
|
Similar experiments were also carried out with an IDA resin charged
with different metal ions (nickel, zinc, copper, cadmium, and cobalt).
We have found that the H. ducreyi enzyme binds to this resin
with an affinity that is independent of the divalent metal ion used to
charge the resin.
Copper Binding by the N-terminal Domain--
The electronic
spectra in the visible region of wild type Cu,Zn-SODs from H. ducreyi and H. parainfluenzae (Fig.
4) exhibited unusual absorption bands
with respect to other bacterial or eukaryotic enzyme variants. The
spectrum of H. parainfluenzae Cu,Zn-SOD (Fig. 4A)
displays two maxima at 680 and 525 nm, whereas the H. ducreyi enzyme (Fig. 4B) is characterized by a broad
absorption peak centered at 630-650 nm. The precise location of the
peak varied from preparation to preparation, and the peak was
blue-shifted in samples containing more copper. The addition of
sub-stoichiometric amounts of copper to the H. ducreyi
enzyme induced a strong increase in absorbance and a further blue shift
of the absorption maximum up to 615 nm, whereas the addition of small
amounts of copper to the Cu,Zn-SOD from H. parainfluenzae
slightly increased the sample absorbance without affecting the maximum
wavelengths (not shown). In contrast, the H. ducreyi and
H. parainfluenzae mutants lacking the N-terminal histidines
exhibited a copper absorption band centered at 680 nm, typical of
Cu,Zn-SODs naturally lacking the N-terminal histidines. In line with
these observations, which suggest that the anomalous electronic spectra
of the two enzymes are due to copper bound at the N-terminal domain, we
observed that, upon pretreatment with EDTA, both the enzymes exhibited
a single absorption maximum close to 680 nm (Fig. 4, A and
B).
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 4.
Electronic absorption spectra of H. parainfluenzae and H. ducreyi
Cu,Zn-SODs. A, trace a, H. parainfluenzae
Cu,Zn-SOD as isolated (~12 mg/ml); trace b, trace
a 1 min after EDTA addition. B, trace a,
H. ducreyi Cu,Zn-SOD as isolated (~40 mg/ml); trace
b, trace a after the addition of 0.5 mM
CuCl2; trace c, trace b 1 min after the addition
of 1 mM EDTA. C, copper binding by copper-free
H. parainfluenzae Cu,Zn-SOD. The process of copper binding
by copper-free H. parainfluenzae Cu,Zn-SOD at 25 °C was
monitored by electronic absorption spectroscopy. Copper reconstitution
was started by the addition of 1 copper eq (5 × 104 M) to the enzyme dissolved in
water at a concentration of 8.6 mg/ml. Cu,Zn-SOD absorption spectra in
the 450-750 nm region were registered before the addition of copper
(trace a) and at the following times after copper
supplementation: trace b, 5 min; trace c, 15 min;
trace d, 30 min; trace e, 60 min; trace
f, 80 min; trace g, 120 min.
|
|
To investigate the possible relationship between the His-rich metal
binding domain and the active site of the enzyme, we prepared copper-free derivatives of the H. parainfluenzae and
H. ducreyi Cu,Zn-SODs and monitored the process of copper
binding in vitro. One eq of copper (i.e. two
copper ions/enzyme dimer) was added to the H. parainfluenzae
copper-free enzyme, and the electronic spectrum in the 450-800 nm
region was followed over time (Fig. 4C). Immediately after
the addition of copper, we observed the appearance of an absorption
peak centered close to 525 nm with only a small shoulder at longer
wavelengths. During incubation, a slow but progressive increase in
absorbance of the copper band at 680 nm, indicative of copper binding
in the active site, was observed, with the concomitant disappearance of
the 525-nm peak. The rate of this process was influenced by
temperature, being more than 5-fold faster at 37 than at 25 °C,
suggesting that higher temperatures may increase motions of the
N-terminal domain and favor metal translocation from the His-rich
region to an active site. Immediately after the addition of a second
equivalent of copper, the protein spectrum showed two well resolved
absorption bands at 525 and 680 nm, similar to those found with the
Cu,Zn-SOD purified from E. coli extracts (see Fig.
4A). The copper absorption peak at 525 nm disappeared
following the addition of EDTA (Fig. 4A). In contrast,
reconstitution of copper-free H. parainfluenzae mutant
enzyme with an equivalent of copper led to the rapid development of the
680 nm peak and showed no evidence of additional copper-binding sites.
Although the spectra reported in Fig. 4C do not identify a
clean isosbestic point, they suggest that copper is rapidly bound at
the N-terminal domain and then transferred to an active site. We
suggest that the lack of an isosbestic point might be due to an
increase in absorbance at low wavelengths following copper binding in
an active site due to a ligand-to-metal charge transfer transition
between the imidazole of His-61 (the bridging histidine) and copper
(46). Alternative explanations may be that this is due to the presence
of copper bound to additional sites on the enzyme surface or to the
simultaneous presence of copper in both sites in a subset of molecules.
Similar reconstitution experiments were performed with the copper-free
H. ducreyi enzyme (data not shown). In this case, however,
the large overlap between the absorption peaks due to copper binding at
the N terminus and at the active site (and possibly at other sites on
the enzyme surface) prevented a clear spectral demonstration of metal
exchange between the two copper-binding sites. In fact, addition of 1 eq of copper to the wild type copper-free enzyme produced an increase in absorption over a wide region with no well defined peaks, whereas upon addition of a further equivalent a peak close to 615 nm appeared, similar to that observed in the enzyme as isolated after addition of
copper (not shown). Also in this enzyme the removal of loosely bound
copper by EDTA rapidly produced a modification of the spectrum, with
appearance of the typical active site copper absorption band centered
close to 680 nm. The same band was observed after reconstitution with 1 copper eq of the H. ducreyi mutant lacking the N-terminal domain.
Kinetics of Activity Recovery by Reconstituted Copper-free
Enzymes--
The process of metal transfer from the N-terminal domain
to the active site was also monitored by following the kinetics of activity recovery by the copper-free enzyme. The buffer used to measure
SOD activity by the pyrogallol method contains chelating agents that
allow discrimination between activity due to copper bound at the active
site and that due to spurious or loosely bound metals. In fact, as
chelating agents rapidly remove the copper bound to His-rich domains,
the catalytic activity observed during the reconstitution process must
be exclusively attributable to copper bound in the active sites.
Reconstitution experiments in which a stoichiometric amount of copper
(1 copper eq/SOD subunit) was added to wild type and mutant copper-free
enzymes from H. parainfluenzae and H. ducreyi
were initially performed. The kinetics of activity recovery of
the four proteins are shown in Fig. 5. Mutant H. parainfluenzae enzyme regained more than 90%
activity in less than 5 min, indicating that this protein is able to
bind copper very rapidly in the catalytic site. In contrast, presumably due to copper trapping by the His-rich domain, the increase of catalytic activity of wild type enzyme was much slower (it reached 90%
activity only after more than 30 min of incubation). Under the same
experimental conditions, the N-terminal deleted H. ducreyi enzyme regained activity more slowly than mutant H. parainfluenzae enzyme, but differences between the two enzymes
were abolished when the reconstitution experiments were carried out
using 2 copper eq/SOD subunit (see also Fig.
6). This finding suggests, in line with
affinity chromatography experiments (see Fig. 3), that other site(s)
able to compete with the active sites for copper binding are present on
the surface of mutant H. ducreyi enzyme devoid of the
His-rich domain. Compared with the other enzymes, wild type H. ducreyi Cu,Zn-SOD showed much slower kinetics of activity recovery.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 5.
Kinetics of regaining activity by the
copper-free enzymes. Copper-free proteins at a concentration of
2.9 × 107 M were incubated
with an equivalent of copper (5.8 × 107
M), and the increase in catalytic activity over time was
measured by the pyrogallol method.  , H. ducreyi
Cu,Zn-SOD;  , H. parainfluenzae Cu,Zn-SOD;  , mutant
H. ducreyi Cu,Zn-SOD;  , mutant H. parainfluenzae Cu,Zn-SOD.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 6.
Kinetics of regaining activity by the
copper-free enzymes at variable copper/subunit molar ratios.
A, reconstitution of copper-free wild type H. parainfluenzae SOD carried out at the following copper/SOD subunit
molar ratios:  , 2:1;  , 1.4:1; , 1:1; *, 1:5; , 1:20;  ,
1:40. Reconstitution of mutant H. parainfluenzae enzyme at
the same copper/protein molar ratios was also performed obtaining in
all cases a nearly identical curve ( ). B, reconstitution
of copper-free wild type (filled symbols) and
N-terminal deleted mutant (open symbols) H. ducreyi enzyme, carried out at the following copper/SOD subunit
molar ratios: , 4:1;  , 2:1; , , 1:1; , 1:5;  , 1:10;
, 1:20; , , 1:40.
|
|
The process of copper incorporation into the active site was further
investigated by measuring the kinetics of recovery of activity as a
function of the copper/subunit ratio (Fig. 6). These experiments
confirmed that the recovery of activity that follows copper addition to
mutant H. parainfluenzae copper-free protein is extremely
fast, as 50% of the enzyme maximal activity was reached in less than
30 s under all conditions tested (the rate of reactivation of this
enzyme is too fast to show differences related to the copper/protein
ratio by the method we have used). Nearly identical kinetics of
activity recovery were observed for the wild type H. parainfluenzae enzyme when reconstitution was carried out with two
or more copper ions/SOD subunit. At lower copper/subunit molar ratios,
the rates of wild type H. parainfluenzae enzyme reactivation were slower (Fig. 6A); the process of copper incorporation
in an active site was negatively affected by decreasing the
copper/subunit ratio. The most likely explanation for such results is
that copper-deficient N-terminal sites compete with each other
and with the active sites for the copper ion bound at the N-terminal
domain of another Cu,Zn-SOD molecule. Also, it is possible that more
stable copper complexes involving histidine ligands from different
molecules may form when protein is in excess with respect to copper. In
any case, despite the long time required to activate the enzyme (at a
copper/subunit molar ratio equal to 1:40 about 28 min were necessary to
reach 50% maximal activity), we have observed that wild type H. parainfluenzae Cu,Zn-SOD is eventually able to reach an activity
value close to 100% of its maximal activity even at low copper/protein
ratios (see Fig. 6A and data not shown).
Reactivation kinetics of the H. ducreyi enzyme are much
slower than those of H. parainfluenzae Cu,Zn-SOD (Fig.
6B). The N-terminal deleted enzyme rapidly reached
50% of its maximal activity when it was reconstituted with two copper
ions/SOD subunit, but at lower copper/subunit molar ratios, the times
required to reach 50% maximal activity slightly increased, and the
enzyme failed to reach full activation. This result is consistent with
the hypothesis that, unlike the case of the mutant H. parainfluenzae enzyme, there must be another site with significant
copper affinity (or more sites with low affinity) on the enzyme surface
that compete(s) with the active sites for copper binding. It is
possible that this additional site might involve ligands from different
SOD dimers, as already observed in the crystal of A. pleuropneumoniae Cu,Zn-SOD (16) or the two facing histidine
residues (His-64) located at the dimer interface (31).
As in the case of the Cu,Zn-SOD from H. parainfluenzae, the
wild type H. ducreyi copper-free enzyme recovered activity
very slowly, and a strong decrease in the rate of catalytic
reactivation was obtained by lowering the copper/protein ratio in the
reconstitution reaction. For example, at a copper/subunit ratio equal
to 1:40 about 75 min are necessary to achieve 50% maximal activity
(compared with 28 min for the H. parainfluenzae enzyme and 5 min for the N-terminal deleted H. ducreyi mutant). On
the whole, the kinetics of H. ducreyi Cu,Zn-SOD
reconstitution suggest that the metal-binding sites present on the
surface of this enzyme compete with very high efficiency with the
active sites for copper binding. Moreover, although the fully
reconstituted N-terminal deleted H. ducreyi protein
(obtained at a Cu/subunit ratio of 2) showed an activity identical to
that of the reconstituted mutant and wild type H. parainfluenzae enzymes, we never succeeded in obtaining the same activity value for reconstituted H. ducreyi Cu,Zn-SOD. When
reconstitution was carried out at a copper/subunit ratio equal to 4, the H. ducreyi enzyme reached 97% activity of the
copper-saturated N-terminal deleted mutant. Lower copper/subunit
ratios resulted in lower levels of activity recovery, whereas higher
amounts of copper induced significant protein aggregation. This result
suggests that each H. ducreyi N-terminal domain might be
involved in the binding of two or more copper ions.
Spectroscopic analyses and reconstitution experiments indicate that the
copper ions bound by the N-terminal domains are subsequently slowly
transferred to an active site. Different mechanisms, including intramolecular and intermolecular copper transfer from the N-terminal domain to an active site, could be invoked to explain these results. To
investigate further the mechanism of copper translocation from the
N-terminal domain to the active site, a different kind of reconstitution experiment was carried out (Fig.
7). Copper was added to wild type
H. parainfluenzae enzyme at a 1:10 copper/subunit ratio, and
after 2.5 min the still largely inactive copper-enzyme complex
(possessing about 6% maximal activity) was added to the mutant protein
devoid of the N-terminal histidines at a final copper/subunit ratio
equal to 1:20 (50% subunits of wild type and 50% subunits of mutant
enzyme). The recovery of activity was monitored in parallel with that
of the mutant (copper/subunit ratio of 1:20) and wild type
(copper/subunit ratios of 1:10 and 1:20) enzymes. A net increase in the
rate of enzyme reactivation with respect to that of the wild type
enzyme was observed. This result strongly suggests that intermolecular
copper transfer may occur from the N-terminal domain of an SOD molecule
to an active site of an enzyme subunit located in a distinct Cu,Zn-SOD
dimer.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 7.
Intermolecular copper transfer.
Reconstitution was started by the addition of copper (5.8 × 107 M) to copper-free wild type
H. parainfluenzae enzyme (2.9 × 106 M), thus obtaining a mixture
with a copper/subunit ratio of 1:10. After 2.5 min 2.9 × 106 M mutant H. parainfluenzae enzyme lacking the N-terminal histidines was added
to the copper-enzyme complex (where copper is still largely bound at
the N terminus of the protein), thus obtaining a sample with a
copper/subunit ratio of 1:20. The recovery of activity of this sample
(*) was monitored in parallel with that of the native enzyme
reconstituted at 1:10 () and 1:20 () copper/subunit ratios and of
the mutant enzyme reconstituted at the copper/subunit ratio of 1:20
().
|
|
Copper Release from the Active Site of Bacterial
Cu,Zn-SODs--
After extensive dialysis, the EDTA-treated H. ducreyi Cu,Zn-SOD (either the as-isolated enzyme or the
copper-free reconstituted form) showed an absorption maximum at 660 nm,
whereas the dialyzed EDTA-treated H. parainfluenzae enzyme
showed a shoulder close to 525 nm. Moreover, a consistent decrease in
copper content after dialysis was observed in all Cu,Zn-SOD samples
(wild type and mutants lacking N-terminal His-rich domains). These
observations suggest that the active site copper can be lost by these
bacterial Cu,Zn-SODs, as described previously for the E. coli enzyme (19, 47). To test this hypothesis, Cu,Zn-SODs from
H. parainfluenzae, H. ducreyi, and E. coli were incubated at neutral pH in the presence of 0.1 mM EDTA at 37 °C, and the enzyme activity was assayed by the pyrogallol method at a series of time points. As shown in Fig.
8, all Cu,Zn-SODs underwent a progressive
decrease in activity, indicative of copper loss from the active site.
Under the same conditions bovine Cu,Zn-SOD activity was totally
unaffected by incubation in the presence of EDTA. The rate of loss of
enzyme activity was similar in wild type and mutant
Haemophilus enzymes, indicating that the N-terminal domain
is not directly involved in the process of metal loss from the active
site. It should be noted that the rate of inactivation of the E. coli enzyme was much faster than that of the
Haemophilus enzymes, possibly due to higher mobility of
loops forming the active site channel in the monomeric enzyme (19).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 8.
Copper loss by bacterial Cu,Zn-SODs.
Enzymes were incubated in 100 mM phosphate buffer, 0.1 mM EDTA, pH 7.0, at 37 °C. Aliquots were withdrawn at
the indicated times and immediately assayed by the pyrogallol method to
measure residual activity. Each data point is the mean of three
independent measurements. , E. coli Cu,Zn-SOD; ,
H. ducreyi Cu,Zn-SOD; , H. parainfluenzae
Cu,Zn-SOD; , mutant H. ducreyi Cu,Zn-SOD; , mutant
H. parainfluenzae Cu,Zn-SOD; , bovine Cu,Zn-SOD.
|
|
Macrophage Killing Assays--
We have tested the ability of wild
type and mutant Cu,Zn-SODs from H. parainfluenzae and
H. ducreyi to protect E. coli K12 cells from
phagocytic killing. The assays were carried out using murine or human
macrophages. Control experiments were performed to determine the serum
resistance of each bacterial strain in the presence of 5% fresh pooled
serum but without macrophages.
Bacteria used in phagocytic killing experiments were obtained by
culturing cells until they reached the stationary phase either in
standard LB medium or in LB supplemented with copper and zinc. In fact,
as observed previously for other Cu,Zn-SODs (28), the enzymes from
H. ducreyi and H. parainfluenzae produced in
E. coli cells grown in LB medium display a modest catalytic
activity, whereas both wild type and mutant enzymes show maximal
activation (about a 20-fold higher activity) when produced in media
supplemented with 200 µM CuSO4 and 10 µM ZnSO4. Under these conditions the Cu,Zn-SOD activity expressed by bacteria producing wild type or mutant
Cu,Zn-SOD was comparable (data not shown). Fig.
9 shows that the survival of recombinant
E. coli cells expressing wild type or mutant H. parainfluenzae Cu,Zn-SOD within the mouse macrophage-like line
J774 was similar to that of control cells when bacteria were pre-cultured in standard LB medium. In contrast, wild type H. parainfluenzae Cu,Zn-SOD was able to confer significant protection from phagocytic killing to E. coli cells grown in a medium
supplemented with metals. Interestingly, the mutant enzyme lacking the
N-terminal histidines provided a significantly lower protection.
Experiments carried out with the human macrophage-like line THP-1 gave
essentially identical results.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 9.
The His-rich domain influences E. coli cells resistance to phagocytic killing. Macrophage
killing of E. coli GC4468 cells expressing wild type and
mutant Haemophilus Cu,Zn-SODs. Bacteria were grown in
standard LB (open symbols) or in LB supplemented with copper
and zinc (filled symbols). The phagocytosis experiment was
carried out as described under "Experimental Procedures" with J774
murine macrophage-like cells. , wild type H. ducreyi
Cu,Zn-SOD; , mutant H. ducreyi Cu,Zn-SOD; , ,
H. parainfluenzae Cu,Zn-SOD; , , mutant H. parainfluenzae Cu,Zn-SOD , ; control cells bearing plasmid
pEMBL18.
|
|
Similar experiments were carried out with bacteria expressing the
H. ducreyi Cu,Zn-SOD. Overexpression of this Cu,Zn-SOD
conferred very efficient protection toward macrophage killing on
E. coli grown in copper- and zinc-supplemented media. In
this case, however, the difference in survival between E. coli cells expressing the wild type enzyme and the mutant protein
devoid of the N-terminal histidines was less pronounced (Fig. 9).
| |
DISCUSSION |
This investigation was prompted by inspection of the amino acid
sequences of a subgroup of Cu,Zn-SODs, all expected to have a nearly
identical monomer fold (15). The sequences of some of these enzymes
(from A. pleuropneumoniae, H. ducreyi, H. parainfluenzae, H. influenzae, and N. meningitidis) are characterized by the following two features that
attracted our attention for their potential ability to modulate metal
binding and/or uptake: the unusually large number of histidines and the
presence of His-rich N-terminal extensions. The presence of an
unusually flexible N-terminal domain could be predicted from the
recently solved x-ray structure of A. pleuropneumoniae
Cu,Zn-SOD (16). This enzyme was found to contain two copper
ions/subunit, one of which, as expected, was bound in the active site,
whereas the other was ligated in a tetrahedral geometry on the protein
surface by two neighboring histidine residues from each of two adjacent
SOD molecules. Such an x-ray analysis did not provide information on
the metal binding ability of the N-terminal extension because the
crystallized enzyme was devoid of the N-terminal histidines, due to
protease digestion during protein purification. The aim of this work,
therefore, was to evaluate whether such N-terminal extensions
effectively represent metal binding domains and to assess their ability
to modulate metal uptake in the active site.
Immobilized metal affinity chromatography experiments clearly
demonstrated that the N-terminal regions of the Cu,Zn-SODs from H. ducreyi, H. influenzae, and H. parainfluenzae are characterized by high affinity for divalent
metal ions and that such an affinity is largely mediated by the
His-rich region. The Cu,Zn-SOD from H. ducreyi showed some
affinity for Ni-NTA also upon removal of the N-terminal extension,
indicating that other residues (probably some other histidines located
on the enzyme surface) are able to interact with divalent metal ions,
as in the case of the A. pleuropneumoniae enzyme. The
affinity for divalent ions provided by Cu,Zn-SODs N-terminal regions
was comparable to that reported for several proteins engineered by the
introduction of a 6-histidine tag (48) and was sufficient to allow
purification of the enzyme in a single chromatographic step. Such a
property can be transferred to other proteins that have no affinity for
metal chelate resins.
The ability of the N-terminal region to bind transition metals with
high affinity was confirmed by the electronic spectra of the wild type
enzyme of H. ducreyi and H. parainfluenzae
compared with mutant enzymes lacking their N-terminal histidines. The
two mutant enzymes showed a copper absorption band centered at 680 nm,
typical of all Cu,Zn-SODs. In contrast, the wild type enzymes from both
organisms exhibited unusual spectra due to the overlapping of the
copper absorption band centered at 680 nm with another due to copper
bound to the N-terminal domain. The enzyme from H. parainfluenzae displayed two clearly distinct absorption maxima (680 and 525 nm), whereas the H. ducreyi enzyme showed a
broad absorption band with a maximum affected by copper content of the sample. The difference between the two enzymes probably reflects a
difference in the number or the nature of residues coordinating the
metal ion as well as the possible presence of additional metal-binding site(s) on the surface of the H. ducreyi enzyme.
We have also investigated the possibility that metals initially bound
by the N-terminal domains could be delivered to an active site. This
hypothesis has been tested spectroscopically, monitoring the changes in
the electronic spectra following copper addition, and by catalytic
assays, measuring the regain of activity of copper-free proteins.
Reconstitution experiments with copper-deficient H. parainfluenzae enzyme demonstrated that in vitro copper
is rapidly bound at the N terminus of the enzyme and then is slowly
transferred to an active site. The ability of the N-terminal domain to
compete efficiently with the active sites for copper binding
demonstrates that this protein region may act as a high affinity
binding domain able to capture efficiently metals from the environment.
The observation of a unidirectional transfer of copper from the
N-terminal histidines to an active site suggests that copper binding by
the N-terminal domain is kinetically favored, possibly due to the high
mobility of the N-terminal extension but that the active site binds
copper more tightly than the external domain. Similar reconstitution experiments were carried out with the H. ducreyi enzyme,
although in this case the overlap between the absorption bands due to
copper bound at the active site and at the N-terminal domain (and
possibly at other site(s)) prevented a spectral demonstration of copper transfer from the N-terminal region to an active site.
The kinetics of activity recovery reported in Figs. 5 and 6
confirmed that the N-terminal domain rapidly traps the available copper
ions as the two wild type enzymes showed a much slower recovery of
activity than the mutant enzymes lacking N-terminal histidines. These
experiments also suggested that the N-terminal domains from different
Cu,Zn-SOD molecules compete for copper binding. As a consequence, when
protein reconstitution is carried out with a large molar excess of
protein with respect to copper, the copper ion is transferred very
slowly to an active site.
Either an intra- or an intermolecular mechanism of copper transfer from
the N-terminal domain to an active site could be envisioned to explain
the reconstitution data obtained with the two wild type Cu,Zn-SODs.
However, the observation that the rate of copper transfer to the
active site is enhanced by the addition of an equimolar amount of
mutant protein to wild type H. parainfluenzae enzyme (with
copper already bound at its N-terminal domain (Fig. 7)) demonstrates
that copper transfer does not necessarily proceed via an intramolecular
pathway. If copper bound by the N-terminal histidines were necessarily
transferred intramolecularly, the rate of copper transfer should be
insensitive to the addition of molecules of mutant enzyme. Although
such a result does not totally exclude the possibility that
intramolecular transfer might occur in a subset of molecules, it is
more simply explained by an intermolecular mechanism of copper
transfer. This interpretation is consistent with the observation that
the N-terminal extension of the H. parainfluenzae enzyme is
rather short (9 amino acids) so that, also assuming the total absence
of constraints in its possibility to move, it is difficult to imagine,
based on the available structures of bacterial Cu,Zn-SODs, that it
could reach the active site cavity of the same subunit and mediate the
direct transfer of the copper ion. Similar considerations may apply to the enzymes from H. influenzae, N. meningitidis,
and A. pleuropneumoniae, in all of which the N-terminal
extensions are probably too short to allow direct intramolecular
transfer of the metal ion. Further work is required to understand
whether the N-terminal histidines exert their action by mediating
direct insertion of the copper ion into an active site of a neighboring
SOD dimer or by increasing the local concentration of the metal ions
around the catalytic site. Also, it is worth noting that the H. ducreyi enzyme may form multimeric complexes (31) and that
production of this enzyme in media not supplemented with metal ions
strongly induces formation of the oligomeric forms of the enzyme (not
shown). The formation of these high molecular weight Cu,Zn-SOD
complexes in vivo might favor the process of intermolecular
metal transfer.
Little is known about metal trafficking in the periplasmic space and
the mechanisms of metal uptake by periplasmic Cu,Zn-SOD, but the
discovery of a high affinity metal-binding site located on a very
mobile protein region in bacterial Cu,Zn-SOD suggests that this protein
domain could play some role in the process of copper and zinc binding
in vivo. In fact, recent work has established that several
metalloproteins, although easily able to acquire their metal cofactor
in vitro, require auxiliary factors to obtain the metal
in vivo (49-51). In the cytoplasm of eukaryotic and
prokaryotic cells, the delivery of copper to specific proteins is
mediated by a family of copper-transporting proteins, collectively
known as copper chaperones. Specific interactions between a chaperone and its target protein allows metal exchange between the two proteins, protecting the cell from the reactivity of free metal ions. This is the
case, for example, of eukaryotic Cu,Zn-SOD that, while possessing an
extraordinary high affinity for copper in vitro, is unable
to obtain its metal cofactor in vivo under physiologically low metal conditions in the absence of a specific copper transporter (49, 52). Although the mechanism of copper exchange between Cu,Zn-SOD
and its copper chaperone is still under investigation, it has been
established that the chaperone for Cu,Zn-SOD is characterized by three
distinct protein domains (53, 54). The central domain has a SOD-like
structure that physically interacts with Cu,Zn-SOD. The N- and
C-terminal domains are both involved in metal binding, with the
C-terminal domain, which is disordered in crystal structures and highly
mobile in solution, playing a crucial role for Cu,Zn-SOD activation
in vivo.
Bacteria do not possess genes encoding proteins homologous to the
eukaryotic copper chaperone for Cu,Zn-SOD (55), and copper uptake by
prokaryotic Cu,Zn-SODs is probably largely dependent on metal
availability in the periplasmic space. This hypothesis is consistent
with the observations that the activity of periplasmic Cu,Zn-SOD
overexpressed in bacteria is roughly proportional to the amount of
copper with which the medium is supplemented (28 and this work) and
that the E. coli sodC promoter is down-regulated by extracellular copper chelators (55). Whereas the outer bacterial membrane is highly permeable to small molecules (including metal ions)
present in the environment (56), copper availability in the periplasmic
compartment is regulated by pumps that mediate copper transport into
cells when its intracellular concentration is limiting or its excretion
when it is in excess (51, 57). In the present study, we have
investigated the metal-binding properties of the His-rich domains,
following their ability to bind copper and mediate its transfer to an
active site. This choice was prompted by the assumption that copper
availability is limited in cells and by our ability to monitor copper
binding either spectroscopically or by its effect on the enzyme
activity. In contrast, zinc, which is usually considered an abundant
element, is spectroscopically silent, and its contribution to activity
cannot be detected by indirect solution assays. However, the
possibility that the N-terminal domain could play a role also in the
binding of zinc should not be discarded. In fact, complex mechanisms
also regulate zinc concentration in the bacterial cell (58) and could
potentially limit zinc availability in the periplasm. As a matter of
fact, screening Vibrio cholerae and Pseudomonas
aeruginosa for genes induced during infection have identified
zinc-responsive regulators as putative virulence factors (59, 60).
The problem of obtaining adequate amounts of these trace metals must be
particularly acute for those pathogens that colonize host compartments
where free metal ions (defined as metal ions not sequestrated by other
metal-binding proteins) are virtually or entirely absent. We suggest
that a likely function for the His-rich domain present in some
Cu,Zn-SODs could be to facilitate metal uptake under conditions of
metal starvation. By virtue of their high affinity for divalent ions,
such N-terminal domains could efficiently trap the few available metal
ions and subsequently deliver them to an active site. Through
development of such a metal recruiting function, Cu,Zn-SODs may be
enabled to acquire these crucial cations in the presence of other
competing proteins in the crowded periplasmic environment. Although no
data are available concerning free copper concentration in mammalian
tissues, it has recently been shown that only a few mammalian cell
types are able to support a low level of copper incorporation into
cytosolic Cu,Zn-SOD in the absence of its copper chaperone (61),
implying that in most tissues free copper availability is low.
Interestingly, so far we have identified the His-rich metal binding
domain exclusively in Cu,Zn-SODs from mucosal colonists and pathogens.
These organisms inhabit an environment where free metal ions might be
particularly scarce, in contrast to the situation in which enteric or
free living bacteria (whose Cu,Zn-SODs lack this domain) are to be found.
The slow kinetics of copper transfer from the N-terminal domain to an
active site in vitro, with a consequent retarding effect on
the rate of catalytic activation of the protein compared with mutant
Cu,Zn-SODs devoid of N-terminal histidines, suggests that such protein
domains may provide a contingency rather than an obligatory function,
being needed perhaps only under conditions of severe metal deprivation.
Our results (Fig. 9) demonstrate that the ability of overexpressed
Cu,Zn-SODs to protect E. coli from phagocytic killing
requires pre-charging of the enzyme with catalytic copper before
bacteria enter the phagocyte. In this context, overexpression of
Cu,Zn-SODs containing the additional N-terminal domain protect E. coli more efficiently than overexpression of mutant enzymes devoid
of this high affinity metal-binding site. Such enhanced protection
might suggest a higher enzymatic activity of the wild type protein in the course of microbial infection. However, the metal binding function
of the N-terminal domain may be invoked as a plausible alternative
explanation. As reported previously for the E. coli enzyme
(19), Cu,Zn-SODs from Haemophilus species readily lose copper from their active site. Reduction of divalent copper at the
active site, which follows reaction with hydrogen peroxide generated at
high concentrations during the oxidative burst, may further decrease
copper affinity for the active site and favor its release. If such
copper loss occurs in the course of infection, we speculate that, where
present, the N-terminal domain could confer an advantage, either by
favoring reacquisition of copper by the Cu,Zn-SOD active site or by
shielding bacterial cells from toxic free radicals generated by the
reaction of free copper ions with hydrogen peroxide.
| |
FOOTNOTES |
*
This work was supported in part by the MURST-CNR program
L.95/95 "Biomolecole per la Salute Umana" (to A. B), by the CNR
target project on "Biotechnology" (to A. B), by a MURST PRIN grant
(to P. V. and G. R.), and by The Wellcome Trust (to J. S. K. and
P. R. L.).The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
§
To whom correspondence should be addressed: Dip. di Biologia,
Università di Roma "Tor Vergata," Via della Ricerca
Scientifica, 00133 Rome, Italy. Tel.: 39-0672594372; Fax:
39-0672594311; E-mail: andrea.battistoni@uniroma2.it.
Published, JBC Papers in Press, May 21, 2001, DOI 10.1074/jbc.M010527200
| |
ABBREVIATIONS |
The abbreviations used are:
Cu, Zn-SOD,
Cu,Zn-superoxide dismutase;
LB, Luria-Bertani broth;
PCR, polymerase
chain reaction;
PAGE, polyacrylamide gel electrophoresis;
NTA, nitrilotriacetic acid, IDA, iminodiacetic acid.
| |
REFERENCES |
TOP
ABSTRACT
INTRODUCTION
