Armitage2 1 Health Science Center, University of Utah, Salt Lake City, UT, USA;2Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, Minneapolis, MN, US
Trang 1Solution structure of Cu6 metallothionein from the fungus
Neurospora crassa
Paul A Cobine1, Ryan T McKay2,*, Klaus Zangger2,†, Charles T Dameron3and Ian M Armitage2
1
Health Science Center, University of Utah, Salt Lake City, UT, USA;2Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, Minneapolis, MN, USA;3Department of Chemistry and Biochemistry, Duquesne University, Pittsburgh, PA, USA
The 3D-solution structure of Neurospora crassa Cu6
-metal-lothionein (NcMT) polypeptide backbone was determined
using homonuclear, multidimensional1H-NMR
spectros-copy It represents a new metallothionein (MT) fold with a
protein chain where the N-terminal half is left-handed and
the C-terminal half right-handedly folded around a
cop-per(I)-sulfur cluster As seen with other MTs, the protein
lacks definable secondary structural elements; however, the
polypeptide fold is unique The metal coordination and
the cysteine spacing defines this unique fold NcMT is only
the second MT in the copper-bound form to be structurally
characterized and the first containing the
-CxCxxxxxCxC-motif This motif is found in a variety of mammalian MTs
and metalloregulatory proteins The in vitro formation of the
Cu6NcMT identical to the native Cu6NcMT was dependent
upon the prior formation of the Zn3NcMT and its titration
with Cu(I) The enhanced sensitivity and resolution of the
800 MHz1H-NMR spectral data permitted the 3D structure determination of the polypeptide backbone without the substitution and utilization of the NMR active spin 1/2 metals such as113Cd and109Ag These restraints have been necessary to establish specific metal to cysteine restraints in 3D structural studies on this family of proteins when using lower field, less sensitive1H-NMR spectral data The accu-racy of the structure calculated without these constraints is, however, supported by the similarities of the 800 MHz structures of the a-domain of mouse MT1 compared to the one recalculated without metal–cysteine connectivities Keywords: copper; metallothionein; Neurospora crassa; NMR; solution structure
Metallothioneins (MTs) are a ubiquitous class of proteins
occurring in both prokaryotes and eukaryotes [1] MTs are
known for their small size (< 7 kDa), the ability to
coordinate a diverse range of metals, a lack of definable
secondary structure, high cysteine content ( 30%), and
degeneracy in the remaining residues (e.g predominance of
cysteine, serine, lysine and no aromatic residues) The high
cysteine content and their spacing give the MTs a high
affinity for metals (e.g Kaof Zn–MT 1 · 1012M) [2,3]
While an essential physiological role has yet to be ascribed
to MT, there is no question that MTs are involved in the protection of cells against metal intoxication through the sequestration of the excess essential metal ions like copper and zinc, as well as nonessential metal ions, like cadmium, mercury and silver [4] Although the essential metals have critical structural, catalytic and regulatory roles in proteins, their cellular concentration must be carefully maintained through the use of pumps and sequestering peptides and proteins to avoid toxic effects Despite the high affinity for metals, MTs are also postulated to participate in an undetermined mechanism of facile metal exchange (i.e kinetically labile yet thermodynamically stable) with other proteins possessing substantially lower metal affinities [5–8] Since the MTs have little or no repetitive secondary structure, their tertiary structure is dependent on the number and type of metal ions they coordinate
The MT from the fungus Neurospora crassa, is a single domain MT made up of 25 residues, seven of which are cysteine residues [9] This small peptide coordinates six Cu(I) atoms via the seven cysteinyl sulfurs [10] N crassa metallothionein (NcMT) binds copper in a solvent-shiel-ded environment, which produces a luminescent core [11]
In vivo, the NcMT is induced only by copper; other transition metals (zinc, cadmium, cobalt, and nickel) do not induce transcription of NcMT mRNA although the protein will bind these metal ions in vitro [12,13] The
in vitrometal ion binding characteristics of NcMT mimic that of the mammalian b-domain, i.e NcMT binds zinc(II), cadmium(II) and cobalt(II) with a 3 : 1 metal to
Correspondence to I M Armitage, Department of Biochemistry,
Molecular Biology and Biophysics, University of Minnesota, 6-155
Jackson Hall, 321 Church Street S.E., Minneapolis, MN 55455, USA.
Fax: +1 612 625 2163, Tel.: +1 612 624 5977,
E-mail: armitage@msi.umn.edu
Abbreviations: MT, metallothionein; NcMT, Neurospora crassa
met-allothionein; aMT-1, a-domain of mouse MT-1; s c , correlation time.
Note: The PDB file has been assigned the Brookhaven Protein Data
Bank Accession no 1T2Y and the chemical shifts have been deposited
in the BMRB data bank under accession number 6290.
*Present address: 101 NANUC, University of Alberta, Edmonton,
AB Canada T6G 2E1.
Present address: Institute of Chemistry/Organic and Bioorganic
Chemistry, University of Graz, Heinrichstrasse 28, A-8010 Graz,
Austria.
(Received 30 June 2004, revised 2 September 2004,
accepted 7 September 2004)
Trang 2protein stoichiometry and Cu(I) with a 6 : 1 stoichiometry
[12] The NcMT cysteine arrangement is identical to the
first seven amino-terminal cysteines of the b-domain of
human MT (Fig 1), though the human MT has two
additional cysteines in this domain Similar cysteine
spacing is seen in the metal binding motifs of the
metalloregulatory proteins from Enterococcus hirae
(CopY), Saccharomyces cerevisiae (Ace1 and Mac1) and
Candida albicans (AMT1) though these proteins do not
share any other significant homology, Fig 1 Along with
the metal coordination number, the positioning of the
cysteines in the primary structure has been determined to
be the critical factor in determining the global fold of the
MTs [14] The structures of several MT isoforms have
been determined including that from yeast [15,16], crab
[17], sea urchin [14], an antarctic fish [18] and various
mammals [19–23] With the exception of the yeast Cu7MT
[16], the previous NMR solution structures of MTs,
including yeast MT [15], all relied on the substitution of
NMR active spin 1/2 metals such as113Cd and109Ag for
the determination of specific metal–cysteine restraints,
which were included in the subsequent structure
calcula-tions [24] While the mammalian copper(I) MTs have
evaded 3D structural elucidation by NMR and X-ray
crystallography [1], recent CD spectroscopic studies on the
Cu(I) and Ag(I) substituted b-fragment of mouse MT1
have revealed structural differences that may reflect a
different conformational fold for the Ag(I) substituted
b-domain compared to the native Cu(I) containing
fragment [25] The only copper(I) MT structure
deter-mined has been that of the yeast MT which has a
distinctly different cysteine arrangement, Fig 1 [15,16]
However, disagreement exists in the comparison of this
NMR derived structure of the yeast MT without the 13
C-terminal residues and the small angle X-ray scattering
pattern of the full length protein which has been attributed
to the core of the NMR structure being too compact
[15,16,26] The similarity in the metal binding properties
and cysteine spacing in NcMT with mammalian MT
makes it a candidate for the modelling of the Cu(I)–sulfur
cluster of the b-domain of mammalian Cu6MT and
potentially useful for other copper(I)-regulated proteins
that contain the -CxCxxxxCxC- and -CxCxxxxxCxC-motifs, Fig 1
The1H NMR resonance assignments for the Cu6NcMT purified from N crassa have been reported [27] and these were used as a template to compare and confirm a similar overall fold for the in vitro reconstituted, synthesized protein used in this study We report here the 3D NMR solution structure of the Cu6NcMT, obtained without establishing any metal–cysteine restraints, and discuss the formation
of the Zn3NcMT precursor molecule that was needed to obtain the Cu6NcMT structure
Materials and methods
Proteins and peptides The NcMT was synthesized chemically to avoid its prob-lematic induction and purification As noted previously by another group [28], N crassa can use several or a mixture of copper resistance mechanisms to detoxify excess copper In initial attempts, we found that exposure to copper did not consistently result in the induction of MT Therefore, a synthetic peptide corresponding to the NcMT was synthes-ized using Boc-Pam resin and Boc-chemistry protocols [29] The peptide was deprotected and cleaved from the resin using anhydrous liquid hydrogen fluoride, para-cresol and para-thiocresol (269–271 K for 70 min) Purification inclu-ded a diethyl ether wash under a nitrogen atmosphere and preparative trifluoroacetic acid (TFA)/acetonitrile [buffer A: 0.1% (v/v) TFA/water; buffer B 90% acetonitrile/10% water and 0.1% TFA, v/v/v) linear gradient HPLC (0–50% buffer B over 40 min) on a reverse phase HPLC column (Waters Delta-Pak PrepPak C18; 40 mm· 100 mm) Elu-tion was monitored at 214 nm, and the eluted peptide was freeze-dried for storage Subsequent analysis showed that the peptide was oxidized by the cleavage/purification treatment, therefore the peptide was reduced by suspension
in 6M guanidine/HCl, 100 mM Tris pH 8.5, and 150 mM
dithiothreitol with incubation at 42C for 2 h The reduced protein sample was then loaded onto a G25 (Pharmacia) size exclusion column (200 mm· 15 mm) previously equili-brated with 0.2% (v/v) TFA The protein fraction was collected, pooled and sealed anaerobically Protein concen-tration was determined by amino acid analysis and the reduction state of the sulfhydryls confirmed by dithiodi-pyridine assay [30] All subsequent manipulations of the apo- and metallated NcMT were performed under anaer-obic conditions, 5% H2and 95% N2(v/v), in a glove box (Atomspure Protector Glove Box, Labconco, Kansas City,
MO, USA)
Metal titrations Copper(I) titrations were performed as described by Byrd
et al [31] Protein samples with reduction state > 95% were used for all titrations Sequential additions of copper(I), as [Cu(CH3CN)4]ClO4 in 200 mM ammonium acetate pH 7.9, were made to the apo-NcMT Identical titrations were performed with Zn(II)3NcMT instead of the apo-NcMT The Zn3NcMT was prepared from apo-NcMT
by adding 3.5 molar equivalents of Zn(II) (added as Zn(II)-acetate) in a 10-fold excess of free cysteine to stabilize the
Fig 1 Amino acid sequence alignment Sequences shown are those of
N crassa MT (residues 2–26, NCBI accession No CAA26793), the
b-domain of Homo sapiens MT-2 A (residues 1–30, NCBI accession
No P02795), the b-domain of Mus musculus MT1 (residues 1–30,
NCBI Accession No AAH36990), AMT1 from Candida glabrata
(residues 70–113, NCBI accession No P41772), ACE1 from
Sac-charomyces cerevisiae (residues 70–110, NCBI accession No.
NP_011349), CopY from Enterococcus hirae (residues 123–145, NCBI
accession No Q47839), MAC1 from S cerevisiae (residues 150–187,
NCBI accession No NP_013734) and MT from S cerevisiae (Cup1–1)
(residues 9–49, NCBI accession No AAS56843) The cysteine residues
are highlighted in bold and the consensus sequence repeated.
Trang 3excess zinc Prior to the copper(I) titration of the NcMT
sample, it was treated with 50 lL of a 1 : 1 slurry of
Chelex-100 (Bio-Rad) in degassed MilliQ water to remove
any excess unbound/displaced metals The Chelex-100 was
removed by centrifugation The Cu(I)-thiolate emission
was monitored at 580 nm (excitation at 295 nm) with a
Perkin-Elmer LS 50B luminescence spectrometer The
spectra were collected at 23C in sealed screw-topped
fluorescence cuvettes (Spectrocell) The spectrometer was
equipped with a 350 nm band-pass filter to avoid second
order effects and used settings of 5 and 20 nm, respectively,
for the excitation and emission slits Cu(I)–NcMT for
NMR analysis was desalted on PD-10 (Pharmacia, G-25)
columns equilibrated with MilliQ water, pooled and
lyophilized Copper stoichiometry was quantified by flame
atomic absorption spectroscopy and amino acid analysis
The final Cu6NcMT contained 1.8 lmol NcMT and
10.9 lmol Cu(I)
NMR sample preparation
Lyophilized Cu6NcMT (handled under an argon
atmo-sphere) was dissolved in 500 lL 90% H2O, 10% D2O,
pH 6.5, 0.1 mM 2,2-dimethyl-2-silapentane-5-sulfonate (as
an internal reference) [32], and 0.02% NaN3 that was
degassed prior to protein addition under high vacuum and
re-pressurized to 1 atm under argon to avoid oxidation/
disproportionation of the Cu(I)–protein The sample was
loaded into a 5 mm NMR tube, capped and sealed with
parafilm
NMR spectroscopy and assignments of NcMT
A 2D1H,1H-TOCSY [33,34] and 2D1H,1H-NOESY were
collected at 10C on a Varian Inova 800 MHz NMR
spectrometer equipped with a axis gradient,
triple-resonance probe The sweep widths for both experiments
were 9000 Hz with 1024 and 512 complex points collected
in the directly and indirectly detected dimensions,
respect-ively The TOCSY was collected with 64 transients per
increment (40 ms mixing time with 300 ms delay between
acquisitions), and the NOESY was collected with 32
transients per increment (300 ms mixing period with
200 ms delay between acquisitions) Spectra were
zero-filled to twice the number of collected points and apodized
using a p/3 shifted sine bell before Fourier transformation
New 1H-chemical shift assignments at 10C from
2D-NOESY and TOCSY spectra were performed as described
previously [35] All spectra were processed usingNMRPIPE
[36], and analysed using the program PIPP [37] Coupling
constants for backbone amide to a hydrogen atoms (i.e
3JHNHa) were obtained from 1D-proton and 2D-COSY
spectra using deconvolution Due to the different
temper-atures used during the NMR experiments (10C in the
present study and 25C in [27]) there are expected to be
small differences in chemical shifts, typically larger for NH
protons However, the overall very good agreement in
these shifts can be confidently attributed to the protein
forming the same structure Gradient NOESY spectra
were acquired with WaterGate solvent suppression taken
from the gnoesywg pulse sequence in Protein Pack as
supplied by Varian Inc
Structure calculations for NcMT Distance restraints (50 kcalÆmol)1ÆA˚)2) for NcMT structure calculations were classified as short (1.8–2.7 A˚), medium (1.8–3.3 A˚), and long (1.8–5.0 A˚) based on their NOE intensities The upper bound of an NOE restraint was extended by 0.5 A˚ for each pseudoatom methylene, or methyl group Residues exhibiting 3J HNHa coupling constants of < 5 Hz, 5–8 Hz, 8–9 Hz, and > 9 Hz were assigned dihedral angle restraints of)60 ± 30, )105 ±
55, ) 120 ± 40, and )120 ± 30, respectively, with a force constant of 500 kcalÆmol)1Ærad)1 No metal atoms were included in the structure calculation Structures were calculated using X-PLOR 3.851 [38] on an Octane Power Desktop (SGI R12000 with IRIX 6.5) using the hybrid distance geometry-dynamical simulated annealing protocol [39] as described for mouse MT1 [23] Out of the 30 calculated structures, 12 were selected based on the complete absence of NOE violations greater than 0.5 A˚ and r.m.s.d for bond and angle deviations from ideality of less than 0.01 A˚ and 5, respectively
Structure calculations for mouse aMT All structure calculations involving the a-domain of the mouse MT, aMT-1, were performed as previously reported [23] with the exception that the metal–cysteine restraints were not used and therefore all metals were absent in the calculations To determine the precision [how well the individual structures calculated with a limited set of long-range NOEs (dij, j > i + 4) compare to each other] and accuracy (the similarity of the calculated structures with a reduced set of long-range NOEs to the structure calculated with all NOEs) of the calculated structures, all 22 long-range NOEs were first removed and then, by a random selection process reintroduced one by one for the structure calcula-tions Thereby, for each number of long-range NOEs, 10 random selections of reintroduced long-range NOE sets were made and 10 structures calculated for each set, giving a total of 100 structures calculated for each point in Fig 4
Results
Metal binding stoichiometry of NcMT
N crassacan express MT in response to excessive concen-trations of copper in the growth media NcMT purified from the mycelium of the fungus contains 6 Cu(I) ions per mole of protein [9] The copper(I) ions are bound to the cysteinyl sulfurs in a solvent-shielded Cu–S core [40] Direct titration of copper(I) into a synthesized apo-NcMT resulted
in a 4-Cu(I)NcMT This complex was luminescent, excita-tion at 295 nm led to an emission maximum at 580 nm, which is evidence for solvent protected Cu(I)-thiolates but the luminescence increased up to a plateau at 4 mole equivalents of Cu(I) per protein (Fig 2A,B) The 4 : 1 stoichiometry conflicts with the known native species Titrations with a variety of copper(I) sources, different types and concentrations of reductant, and with or without thermal treatments as used by Stillman and coworkers [41]
to select stable forms of the rabbit CuMT all resulted in the formation of CuNcMT The 1D1H NMR spectrum of the
Trang 4Cu4NcMT, prepared under any of these conditions, was
very broad and in contrast to the well defined native
Cu6NcMT 1H NMR spectrum [27] The unresolved 1H
NMR spectrum would be consistent with the sample
containing a mixture of interconverting NcMT structures
and/or structural instabilities In an attempt to restrict the
family of conformers formed during the copper(I) titrations,
we first titrated zinc into the apo-protein to form a
Zn3NcMT precursor The excess zinc was removed by
treatment of the sample with Chelex-100 The zinc
stoi-chiometry of the resultant protein, 3-Zn(II):1 NcMT, was as
expected from previous studies of NcMT and by analogy to
the well studied b-domains of the mammalian MTs [12]
Sequential titration of the Zn3NcMT with Cu(I) salts as
before yielded a luminescent core but in this case, the
stoichiometry was 6 Cu(I):1 NcMT, Fig 2C,D Most
importantly, the 1D1H-NMR pattern for the Cu6NcMT
is equivalent to the native Cu(I) proton spectrum [27]
Attempts to prepare the Ag(I) derivative of NcMT did not
result in a stoichiometry of 6 metals per mole of protein
Attempts to displace all of the zinc from the Zn3NcMT by
silver were unsuccessful and always resulted in the
forma-tion of a mixed metal species inappropriate for structural
studies The structure determination of Zn3NcMT was not
pursued as to our knowledge there are no reports of this
metal form from natural sources
Metal-cysteine restraints
To explore the feasibility of determining an MT structure
without metal–cysteine restraints, the structure of the
a-domain of mouse MT-1 without the experimentally
determined113Cd–Cys restraints was re-calculated Fig 3
shows the backbone fold of the average energy minimized
structures for the a-domain of mouse MT-1 with (dark
strand) and without (light strand) the inclusion of
experi-mentally determined 113Cd–Cys restraints The
super-imposed backbone structures are remarkably similar with
an r.m.s.d of 1.94 A˚ over all backbone atoms This value is
very similar to the r.m.s.d comparison between the a-domain of mouse MT-1 and the a-domain of rat MT-2, 2.14 A˚ [23], suggesting that accurate and precise MT structure determinations are possible without using metal
to cysteine restraints providing one has a sufficient number
of NOE constraints The lack of regular secondary structure
in MTs places an increased importance upon the long-range contacts for precision in structure calculations This can be quantitatively evaluated by recalculating the mouse MT1 a-domain structure in the absence of all long-range NOE restraints and 113Cd–Cys connectivities and then syste-matically and randomly reintroducing long-range NMR
Fig 3 Ribbon backbone diagrams of the a-domain of mouse MT-1 calculated with (dark) and without (grey)113Cd–Cys restraints The Cd atoms (rendered to van der Waals radii), cysteine sulfur, and Cd-S bonds are indicated by the large spheres, small spheres and black lines, respectively The N and C termini are labelled for orientation Dia-grams were generated using the program INSIGHTII (Molecular Simu-lations, Inc.).
Fig 2 In vitro copper reconstitutions of
N crassa MT monitoring Cu(I)-S luminescence
at 580 nm after excitation at 295 nm (A) The reconstitutions at 10 nmolÆmL)1of apo-NcMT with copper(I) the increasing additions
of copper the luminescence makes uniform steps (B) (C) Titrations of 50 nmolÆmL)1
Zn 3 NcMT with copper(I) to a stoichiometry
of 6 Cu(I) per mole of protein The sequential increase plotted in (D) reveals a plateau at six equivalents and that the quantum yield of relative luminescence per mole of protein is equal for the 4 and 6 Cu(I)-NcMT.
Trang 5restraints to determine the effect on the precision and
accuracy of the generated structure Figure 4A shows the
r.m.s.d of 10 sets of structures, generated as described in
Materials and methods, to the average minimized structure,
for each set of reintroduced long-range NOEs Determining
the deviation of the generated structure with a reduced
number of NOEs from the published MT1 structure
provides a measure of accuracy shown in Fig 4B As can
be seen, the structure of MT1 without metal–cysteine
restraints is sufficiently well defined as long as 10–15
long-range NOEs are used for the structure calculation and the
backbone r.m.s.d is kept below 2 A˚
NcMT structure
Two-dimensional 1H,1H-TOCSY and NOESY NMR
spectroscopy was used to assign the 1H resonances and
provide interproton distance restraints, respectively, on
NcMT The chemical shifts of NcMT determined here at
10C are similar to those of a previous study performed at
25C [27] Almost all the1H chemical shifts were visible in the TOCSY with the exception of the side chain of Cys5 (assigned from the NOESY), and all resonances of Gly1 (not identified in either experiment) Assignment of Gly1 was not possible; this was most likely due to rapid exchange
of the amide proton of Gly1 with the solvent, coupled with chemical shift overlap reported previously [27] The chem-ical shifts have been deposited in the BMRB data bank under accession number 6290 Examination of the chemical shifts showed no indication of secondary structure [42,43] Line widths of 1D-1H NMR resonances have been used as
an efficient method for determining the aggregation state of
a protein [44,45] The average line width of NcMT amide resonances from the 1D-1H spectrum was determined to be 5.6 Hz When six Cu(I) atoms are considered bound to NcMT, the correlation time (sc) and line width for the monomer are expected to be 2.6–3.7 ns and 4.5–6 Hz, respectively [34] The observed line widths for the reconsti-tuted Cu6NcMT are consistent with the sample being monomeric A summary of the observed NOEs is presented
in Fig 5A Inspection of the NOE patterns shows no evidence of regular secondary structure elements, which is in agreement with the information from the chemical shifts
Fig 4 Effect of the inclusion of long-range NMR restraints for
aMT-1 structure determination in the absence of metal–Cys restraints.
Long-range NMR restraints were removed and then systematically
and randomly re-introduced for mouse aMT-1 structure
calcula-tions (A) Precision indicated via the r.m.s.d of the family of 10
structures generated (compared to the minimized average) for each
number of long-range NOEs included in calculations (B) Relative
accuracy of the calculations indicated by comparing the aMT-1
minimized average structure for each number of long-range NOEs
included (in the absence of metal–Cys restraints), to the average
minimized structure of aMT-1 calculated with all NOEs and
inclu-ding the metal–Cys restraints Error bars in both panels indicate the
standard deviation.
dαN(i,i+3)
d NN (i,i+3)
dNN(i,i+2)
d NN (i,i+1)
G
10
Residue Number
20
15
10
5
0
D C G C S G A S S C N C G S G C S C S N C G S K
dβN(i,i+1)
dαN(i,i+1)
dαN(i,i+2)
A
B
Fig 5 NOE map for NcMT (A) Summary of inter-residue NOEs determined for NcMT that are typically used to indicate secondary structure Strong, medium and weak intensity NOE cross-peaks are indicated by tall dark, medium grey, and small white boxes, respect-ively The primary sequence is shown at the bottom in the one letter code (B) The total number of NOEs assigned for NcMT displayed on
a per residue basis Intra-residue, sequential (d ij ,j ¼ i + 1), medium (d ij ,i + 1 < j < i + 4), long (d ij , j > i + 4) range NOEs are des-cribed by solid, dotted crosshatched, white, and thick crosshatched columns, respectively.
Trang 6and is typical for this family of proteins The total number
of intraresidue, sequential, medium, and long-range NOEs
for each residue in the NcMT sequence is shown in Fig 5B
The majority of critical long-range restraints involve the side
chains of Ala8, Asn12, and Cys21 Despite the almost
complete assignment of1H resonances, the number of
long-range contacts was low Utilizing the 152 NOEs and 13
dihedral angle restraints, a total of 30 structures was
generated using the programXPLOR3.851 [38] From these
generated structures, 12 resulted in no NOE or dihedral
angle violations and all showed the same backbone fold
Subsequently, the 10 accepted structures (no NOE violation
> 0.5 A˚, an r.m.s.d for bond deviations from ideality of
less than 0.01 A˚ and an r.m.s.d for angle deviations from
ideality of less than 5) with the lowest energy were selected
which yielded a backbone r.m.s.d to the average minimized
structure of 0.79 A˚ for the well-defined region (residues
5–20) and 1.59 A˚ for the entire length of the protein The
final family of 10 NcMT structures is shown in Fig 6 with
the N- and C-termini labelled and the cysteine sulfur atoms
in the closest to mean structure drawn as spheres The
structure of NcMT has been deposited in the Protein Data
Bank under accession number 1T2Y The structural
statis-tics for NcMT are presented in Table 1 The program
PROCHECK-NMR[46] showed that > 90% of the ø, w angles
for the 10 structures fell into the core or allowed regions
Despite the complete absence of elements of regular
secondary structure, which is a quite common situation
for MTs, the backbone global fold of the 25-residue peptide
is well defined if one excludes the N and C termini It shows
a new polypeptide structure with the backbone being
wrapped around an empty space containing the copper–
sulfur cluster, on going from the N to C terminus, in a left
handed form for the first half of the molecule then in a right
handed form for the second half Such a mixed handedness
of the peptide part around the metal–cysteine cluster has so
far been found only in the C-terminal domain of blue crab
MT [17], whose overall fold is quite different from that of
NcMT While NcMT shares considerable sequence
similar-ities and a conservative positioning of the seven cysteine
residues with other members of the MT family of proteins,
none of the other structurally characterized MTs show the
same fold This is reflected in the r.m.s.d differences in
cysteine residue positions (backbone heavy atoms and sulfur
atoms) between NcMT and lobster [47], yeast [15,16], mouse
[23], blue crab [17], fish [18] and sea urchin [14] MTs as shown in Table 2 Such r.m.s.d differences in cysteine positions are characteristic of unrelated MT structures A good example to illustrate this point is the r.m.s.d differences between the a- and b-domains of sea urchin
MT of 4.98 A˚ for the cysteine backbone heavy atoms and 2.92 A˚ for the sulfur atoms For comparison the cysteine
Fig 6 Stereo drawing of a line representation
of the backbone of 10 NcMT structures which possessed the lowest overall energy, out of 12 generated that contained no NOE or dihedral angle violations The ensemble is least-squares fitted to the first structure and the N and C termini are labelled for orientation The cys-teine sulfur atoms of the closest to mean structure are shown as yellow spheres.
Table 1 Structural statistics for NcMT, for the 10 lowest overall energy structures out of 12 without any NOE or dihedral violations.
NOE restraints
r.m.s.d to average structure (A˚)
Energies (kcalÆmol)1)
E overall 82.40 ± 7.95
E bonds 6.07 ± 0.74
E angles 29.67 ± 2.35
E dihedral 1.75 ± 0.58
E improper 7.86 ± 0.46 r.m.s.d from idealized covalent geometry used within X - PLOR
Procheck-NMRb[46]
a Residues 5–20 b Number of residues out of all 10 structures Total non-glycine and non-proline is 180 Number of glycines is 60, with 10 end-residues, for a total of 250 residues.
Trang 7residue position r.m.s.d between the structurally related
b-domain of mouse MT1 and human MT2 are 0.57 and
2.10 A˚ for the backbone heavy atoms and the sulfur atoms,
respectively
Discussion
The solution structure of the Cu6NcMT, which was solved
without the acquisition and inclusion of specific metal–
cysteine NMR restraints, shows a novel polypeptide fold
and represents only the second copper MT structure to be
elucidated [15,16] Although other MTs show a strong
sequence similarity to NcMT (e.g 32% with the b-domain
of human MT2) the 3D structure of the polypeptide
backbone is completely different The unique fold of this
CuMT is a clear demonstration that MT protein folds are
largely determined by the constraints of metal–sulfur
connections and not the amino acid sequence N crassa
MT has a -CxCxxxxxCxC- motif that is found in a variety
of Cu(I) binding proteins (Fig 1) The lack of repetitive
secondary structures in the NcMT peptide backbone along
with the unsuccessful efforts to prepare the isomorphic and
homogeneous NMR active spin 1/2 Ag(I) derivative of the
native Cu(I) NcMT were factors which had inhibited its
structure elucidation A hallmark of the MT structures is the
dependence on the sequential position of the cysteine in the
primary sequence, the identity of the coordinated metal and
its coordination number [14,48] The paper by Bertini et al
[16] was the first to show that the acquisition of NMR data
at 800 MHz allows for the determination of MT structures
without metal–cysteine constraints The comparative study
on mouse MT1 at 800 MHz in this paper helps to validate
these studies The successful in vitro reconstitution of
Zn3NcMT with six equivalents of Cu(I) indicated that the
metal binding stoichiometry was the same as the b-domain
of mammalian MT [12] even though NcMT lacks the two
C-terminal cysteines found in the mammalian MTs These
two missing cysteines and therefore less metal coordination
might be the reason why NcMT needs to be in the Zn3-form
before it successfully binds six atoms of copper(I) Thus,
Zn3–NcMT might constitute a scaffold where the zinc can
be replaced by Cu(I) without the need for large structural rearrangements In other words the less favourable enthalpy changes by binding 6 Cu(I) atoms to only seven cysteines in NcMT are compensated by the smaller differences in entropy when copper substitutes zinc in Zn3–NcMT, rather than binding to apo-NcMT
The variable coordination number of copper, which can adopt a linear two-coordinate (digonal) or a distorted planar trigonal three-coordinate geometry with sulfur ligands [49] adds an additional level of complexity in solving the copper MT structures These coordination possibilities
of the copper(I) ions are perhaps responsible for the titration
of the apo-NcMT to a non-native (4 : 1) stoichiometry In the case of the NcMT, it seems that the coordination of zinc
in Zn3NcMT was necessary to conform/stabilize a structure such that the Cu(I) titration produced the native Cu6Cys7
metal–thiolate cluster rather than the Cu4Cys7made from apo-NcMT which gave rise to a family of rapidly intercon-verting structures The resultant Cu6Cys7 core appears to constrain the peptide sufficiently to enable its 3D solution structure to be determined The requirement that the synthetic Cu(I)MT be made from a Zn-protein in vitro to obtain an NMR spectrum identical to that of the native Cu(I)MT raises questions about how the organism controls the formation of the Cu(I)NcMT in vivo If the Zn(II)NcMT were to form in vivo when NcMT was induced by copper exposure one might expect that it would be a very short
Table 2 Cysteine r.m.s.d comparison (in A˚) to other MTs.
a b-domain of lobster Homarus americanus MT b Cu(I)-yeast
Saccharomyces cerevisiae MT.cb-domain of mouse Mus musculus
MT1 d a-domain of blue crab Callinectes sapidus MT e b-domain
of blue crab Callinectes sapidus MT f a-domain of the fish
Noto-thenia coriiceps MT. gb-domain of the fish Notothenia coriiceps
MT h
a-domain of sea urchin Strongylocentrotus purpuratus MT.
i b-domain of sea urchin Strongylocentrotus purpuratus MT.
Fig 7 Ribbon diagram of the NMR solution structure of NcMT with six Cu(I) atoms modelled into possible positions within the structure and the cysteine side chains turned to point towards the protein’s centre Cu(I) atoms are rendered according to their van der Waals radii The Cys side-chains are rendered as grey sticks with the Cys sulfur atom shown
in light grey The modelling of Cu(I) binding as depicted was accom-plished by attempting to best satisfy known bond lengths and altering the Cys chi side-chain angle to point the sulfur atoms towards the centre The figure was prepared using INSIGHTII (Molecular Simula-tions, Inc.) The backbone fold of the NcMT protein is shown in black with the carbons and sulfur of the cysteine side chains shown in dark grey and light grey, respectively The Cu(I) atoms are rendered as medium grey spheres with a van der Waals radius of 0.95 A˚ Taking into consideration the Cu(I)–S luminescence which indicate a pro-tected copper core, we have manually rotated the Cys side-chain 1 cys
and 2 angles to direct the sulfur atoms into the core of the protein.
Trang 8lived and hard to isolate species as copper ions would be
expected to readily displace the Zn(II) However, there are
no in vivo data to indicate that the organism has to form a
zinc precursor The requirement of Zn(II) in the formation
of Cu(I)MT clusters, in particular the b-domain, has been
observed for mouse MT [50] Additionally, the same spacing
of cysteine residues is also found in the repressor protein
CopY from Enterococcus hirae and it is apparent that the
metal-binding properties of this protein may require
the binding of Zn(II) in this site [51,52] The stability of
the Cu(I)-S core and the DNA-binding activity are
dependent on zinc binding In addition, the copper
chap-erone CopZ is a specific source of copper for the Zn-form
of CopY [51] However, apo-CopY does not demonstrate
this specificity suggesting a potential loss of structure that
confers this property A combination of the data suggests
that Zn(II)-binding to these cysteine-rich, copper-binding
sites orders a structure that is required for activity
A model structure of Cu6NcMT demonstrating the
shielding of the copper core, using one model with copper
atoms satisfying the Cu–S bond distances and minimizing
any interactions within the NMR structure, is shown in
Fig 7 While there is probably a multitude of models that
would fit the NMR data, there is no way to confirm any
from NMR studies using the NMR inactive Cu(I) metal
form of the protein The NcMT backbone structure
presented here uses no specific metal–Cys restraints and a
low number of long-range NOE constraints However, the
mouse a-MT-1 structures calculated with and without
specific metal–Cys restraints support the assertion that
metal–Cys restraints are not required for precise and
accurate structure determination of the protein backbone
with NMR data collected at 800 MHz The a-domain of
MT-1 was also used to assess the requirements for
long-range NMR restraints in structure calculations Although
both the length of the a-domain of mouse MT1 (31 vs 25
amino acids in NcMT) and the number of NOEs (256 vs
152 for NcMT) is larger for mouse MT1, the
back-bone r.m.s.d between structures calculated without
metal–cysteine restraints is comparable for NcMT (1.94 A˚
for mouse MT1 and 1.59 A˚ for NcMT)
Evaluation of the relationship between the number of
long-range NOEs and the resulting r.m.s.d is relevant in
MTs because of their lack of defined secondary structure In
more typical proteins, the translational and rotational
positions of the residues contained in a secondary structural
element (a-helix, b-sheet) might be spatially defined by as
little as a single long-range contact to the rest of the molecule
However, in MT, which is composed mostly of coils and
loops, the spatial orientation of relatively large regions
cannot be positioned by such a single critical restraint
In conclusion we have described a new MT fold for the
N crassa Cu6MT which consists of a half left- and half
right-handedly polypeptide backbone wrapped around the
copper(I)-cysteine cluster No direct information about the
metal–sulfur connectivities could be obtained by using an
isomorphic, NMR active metal substitute for Cu+, such as
Ag+, because a stable homogenous Ag substituted NcMT
could not be prepared Nevertheless, the use of data at
800 MHz on the reconstituted Cu6NcMT was sufficient to
allow for an accurate backbone fold to be determined
Acknowledgements This work was supported by NIH grant DK18778 to I.M.A NMR instrumentation was provided with funds from the NSF (BIR-961477) and the University of Minnesota Medical School K.Z thanks the Austrian Science Foundation FWF for financial support under project number P15289 We would like to thank Dr David Live and Dr Beverly
G Ostrowski for maintenance of the spectrometer facility and computers, Matt Vetting for help with software in modeling of the Cu(I)-NcMT and expertise in use of the Argon Chamber for NMR sample preparation, and Gu¨lin O¨z for critical reading of the manuscript We also gratefully acknowledge the University of Minne-sota Supercomputing Institute for Digital Simulation and Advanced Computation for use of their facilities for processing/analysis of NMR data and subsequent structure calculations.
References
1 Vasˇa´k, M & Hasler, D.W (2000) Metallothioneins: new func-tional and structural insights Curr Opin Chem Biol 4, 177–183.
2 Romero-Isart, N., Cols, N., Termansen, M.K., Gelpi, J.L., Duarte, R., Atrian, S., Capdevila, M & Gonzalez-Duarte, P (1999) Replacement of terminal cysteine with histidine
in the metallothionein alpha and beta domains maintains its binding capacity Eur J Biochem 259, 519–527.
3 Li, H & Otvos, J.D (1998) Biphasic kinetics of Zn 2+ removal from Zn metallothionein by nitrilotriacetate are associated with differential reactivity of the two metal clusters J Inorg Biochem.
70, 187–194.
4 Hamer, D.H (1986) Metallothionein Annu Rev Biochem 55, 913–951.
5 Vallee, B.L (1995) The function of metallothionein Neurochem Int 27, 23–33.
6 Jacob, C., Maret, W & Vallee, B.L (1998) Control of zinc transfer between thionein, metallothionein, and zinc proteins Proc Natl Acad Sci USA 95, 3489–3494.
7 Jacob, C., Maret, W & Vallee, B.L (1998) Ebselen, a selenium-containing redox drug, releases zinc from metallothionein Bio-chem Biophys Res Commun 248, 569–573.
8 Faller, P., Hasler, D.W., Zerbe, O., Klauser, S., Winge, D.R & Vasˇa´k, M (1999) Evidence for a dynamic structure of human neuronal growth inhibitory factor and for major rearrangements
of its metal-thiolate clusters Biochemistry 38, 10158–10167.
9 Lerch, K (1980) Copper metallothionein, a copper-binding pro-tein from Neurospora crassa Nature 284, 368–370.
10 Smith, T.A., Lerch, K & Hodgson, K.O (1986) Structural study
of the Cu sites in metallothionein from Neurospora crassa Inorganic Chem 25, 4677–4680.
11 Beltramini, M & Lerch, K (1983) Spectroscopic studies on Neurospora copper metallothonein Biochemistry 22, 2043–2048.
12 Beltramini, M., Lerch, K & Vasak, M (1984) Metal substitution
of Neurospora copper metallothionein Biochemistry 23, 3422– 3427.
13 Munger, K., Germann, U.A & Lerch, K (1987) The Neurospora crassa metallothionein gene J Biol Chem 262, 7363–7367.
14 Riek, R., Precheur, B., Wang, Y., Mackay, E.A., Wider, G., Gu¨ntert, P., Liu, A., Ka¨gi, J.H.R & Wu¨thrich, K (1999) NMR Structure of the sea urchin (Strongylocentrotus purpuratus) metallothionein MTA J Mol Biol 291, 417–428.
15 Peterson, C.W., Narula, S.S & Armitage, I (1996) 3D solution structure of copper and silver-substituted yeast metallothioneins FEBS Lett 379, 85–93.
16 Bertini, I., Hartmann, H.-J., Klein, T., Liu, G., Luchinat, C & Weser, U (2000) High resolution solution structure of the protein part of Cu7 metallothionein Eur J Biochem 267, 1008–1018.
Trang 917 Narula, S.S., Brouwer, M., Hua, Y.X & Armitage, I.M (1995)
Three-dimensional solution structure of Callinectes sapidus
metallothionein-1 determined by homonuclear and heteronuclear
magnetic resonance spectroscopy Biochemistry 34, 620–631.
18 Capasso, C., Carginale, V., Crescenzi, O., Di Maro, D., Parisi, E.,
Spadaccini, R & Temussi, P.A (2003) Solution structure of
MT_nc, a novel metallothionein from the Antarctic fish
Notothenia coriiceps Structure (Camb) 11, 435–443.
19 Arseniev, A., Schultze, P., Worgotter, E., Braun, W., Wagner, G.,
Vasˇa´k, M., Ka¨gi, J.H.R & Wu¨thrich, K (1988)
Three-dimen-sional structure of rabbit liver [Cd7]Metallothionein-2a in aqueous
solution determined by nuclear magnetic resonance J Mol Biol.
201, 637–657.
20 Robbins, A.H., McRee, D.E., Williamson, M., Collett, S.A.,
Xuong, N.H., Furey, W.F., Wang, B.C & Stout, C.D (1991)
Refined crystal structure of Cd, Zn metallothionein at 2.0 A˚
resolution J Mol Biol 221, 1269–1293.
21 Messerle, B.A., Scha¨ffer, A., Vasˇa´k, M., Ka¨gi, J.H & Wu¨thrich,
K (1990) Three-dimensional structure of human
[113Cd7]metal-lothionein-2 in solution determined by nuclear magnetic resonance
spectroscopy J Mol Biol 214, 765–769.
22 Schultze, P., Wo¨rgo¨tter, E., Bruan, W., Wagner, G., Vasˇa´k, M.,
Ka¨gi, J.H & Wu¨thrich, K (1988) Conformation of
[Cd7]-metal-lothionein-2 from rat liver in aqueous solution determined by
nuclear magnetic resonance spectroscopy J Mol Biol 203, 251–
268.
23 Zangger, K., O¨z, G., Otvos, J.D & Armitage, I.M (1999)
Three-dimensional solution structure of mouse [Cd7]-metallothionein-1
by homonuclear and heteronuclear NMR spectroscopy Protein
Sci 8, 2630–2638.
24 O¨z, G., Pountney, D.L & Armitage, I.M (1998) NMR
spectro-scopic studies of I ¼ 1/2 metal ions in biological systems Biochem.
Cell Biol 76, 223–234.
25 Bofill, R., Palacios, O., Capdevila, M., Cols, N., Gonzalez-Duarte,
R., Atrian, S & Gonzalez-Duarte, P (1999) A new insight into the
Ag+ and Cu+ binding sites in the metallothionein beta domain.
J Inorg Biochem 73, 57–64.
26 Sayers, Z., Brouillon, P., Svergun, D.I., Zielenkiewicz, P & Koch,
M.H (1999) Biochemical and structural characterization of
recombinant copper-metallothionein from Saccharomyces
cervi-siae Eur J Biochem 262, 858–865.
27 Malikayil, J.A., Lerch, K & Armitage, I.M (1989) Proton NMR
studies of a metallothionein from Neurospora crassa:
Sequence-specific assignments by NOE measurements in the rotating frame.
Biochemistry 28, 2991–2995.
28 Kneer, R., Kutchan, T.M., Hochberger, A & Zenk, M.H (1992)
Saccharomyces cerevisiae and Neurospora crassa contain heavy
metal sequestering phytochelatin Arch Microbiol 157, 305–310.
29 Schnolzer, M., Alewood, P., Jones, A., Alewood, D & Kent,
S.B.H (1992) In situ neutralization in Boc-chemistry solid phase
synthesis Int J Peptide Protein Res 40, 180–193.
30 Grassetti, D.R & Murray, J.F Jr (1967) Determination of
sulf-hydryl groups with 2,2¢- or 4,4¢-dithiodipyridine Arch Biochem.
Biophys 119, 41–49.
31 Byrd, J., Berger, R.M., McMillin, D.R., Wright, C.F., Hamer, D.
& Winge, D.R (1988) Characterization of the copper-thiolate
cluster in yeast metallothionein and two truncated mutants.
J Biol Chem 263, 6688–6694.
32 Markley, J.L., Bax, A., Arata, Y., Hilbers, C.W., Kaptein, R.,
Sykes, B.D., Wright, P.E & Wu¨thrich, K (1998)
Recommenda-tions for the presentation of NMR structures of proteins and
nucleic acids J Mol Biol 280, 933–952.
33 Braunschweiger, L & Ernst, R.R (1983) Coherence transfer by
isotropic mixing: application to proton correlation spectroscopy.
J Magn Reson 53, 521–528.
34 Cavanagh, J., Fairbrother, W.J., Palmer, A.G III & Skelton, N.J (1996) Protein NMR Spectroscopy: Principles and Practice, 1st edn Academic Press, San Diego.
35 Wu¨thrich, K (1986) NMR of Proteins and Nucleic Acids John Wiley & Sons Inc, New York.
36 Delaglio, F., Grzesiek, S., Vuister, G.W., Zhu, G., Pfeifer, J & Bax, A (1995) NMRPipe: a multidimensional spectral processing system based on UNIX pipes J Biomol NMR 6, 277–293.
37 Garrett, D.S., Powers, R., Gronenborn, A.M & Clore, G.M (1991) A common sense approach to peak picking in two-, three-, and four-dimensional spectra using automatic computer analysis
of contour diagrams J Magn Reson 95, 214–220.
38 Brunger, A.T (1992) X-PLOR, Version 3.1 In A System for X-Ray Crystallography and NMR Yale University Press, New Haven.
39 Nilges, M., Clore, G.M & Gronenborn, A.M (1988) Determi-nation of three-dimensional structures of proteins from inter-proton distance data by dynamical simulated annealing from a random array of atoms Circumventing problems associated with folding FEBS Lett 239, 129–136.
40 Beltramini, M & Lerch, K (1981) Luminescence properties of Neurospora copper metallothionein FEBS Lett 127, 201–203.
41 Green, A.R., Presta, A., Gasyna, Z & Stillman, M.J (1994) Luminescent probe of copper-thiolate cluster formation within mammalian metallothionein Inorganic Chem 33, 4159–4168.
42 Wishart, D.S., Sykes, B.D & Richards, F.M (1991) Relationship between nuclear magnetic resonance chemical and protein sec-ondary structure J Mol Biol 222, 311–333.
43 Wishart, D.S., Sykes, B.D & Richards, F.M (1992) The chemical shift index: a fast and simple method for the assignment of pro-teins secondary structure through NMR spectroscopy Biochem-istry 31, 1647–1651.
44 Li, M.X., Gagne, S.M., Spyracopoulos, L., Kloks, C.P.A.M., Audette, G., Chandra, M.S., Solaro, R.J., Smillie, L.B & Sykes, B.D (1997) NMR studies of Ca2+ binding to the regulatory domains of cardiac and E41A skeletal muscle troponin C reveal the importance of site I to energetics of the induced structural changes Biochemistry 36, 12519–12525.
45 Anglister, J., Grzesiek, S., Ren, H., Klee, C.B & Bax, A (1993) Isotope-edited multidimensional NMR of calcineurin B in the presence of the non-deuterated detergent CHAPS J Biomol NMR 3, 121–126.
46 Laskowski, R.A., MacArthur, M.W., Moss, D.S & Thorton, J.M (1993) PROCHECK: a program to check the sterochemical quality of protein structures J Appl Crystallogr 26, 283–290.
47 Brouwer, M., Winge, D.R & Gray, W.R (1989) Structural and functional diversity of copper-metallothioneins from the Amer-ican lobster Homarus amerAmer-icanus J Inorg Biochem 35, 289–303.
48 Ka¨gi, J.H.R (1991) Overview of metallothionein Methods Enz-ymol 205, 613–626.
49 Pickering, I.J., George, G.N., Dameron, C.T., Kurtz, B., Winge, D.R & Dance, I.G (1993) X-ray absorption spectroscopy of cuprous-thiolate multinuclear clusters in proteins and model sys-tems J Am Chem Soc 115, 9498–9505.
50 Bofill, R., Capdevila, M., Cols, N., Atrian, S & Gonzalez-Duarte,
P (2001) Zinc (II) is required for the in vivo and in vitro folding of mouse copper metallothionein in two domains J Biol Inorg Chem 6, 405–417.
51 Cobine, P., Wickramasinghe, W.A., Harrison, M.D., Weber, T., Solioz, M & Dameron, C.T (1999) The Enterococcus hirae copper chaperone CopZ delivers copper (I) to the CopY repressor FEBS Lett 445, 27–30.
52 Cobine, P.A., Jones, C.E & Dameron, C.T (2002) Role for zinc (II) in the copper (I) regulated protein CopY J Inorg Biochem.
88, 192–196.