The proposal that Sco proteins could play a role in copper delivery to COX within the process of COX assembly was first formulated based on the observation that their overexpression could
Trang 1Seeking the determinants of the elusive functions of Sco proteins
Lucia Banci1,2, Ivano Bertini1,2, Gabriele Cavallaro1 and Simone Ciofi-Baffoni1,2
1 Magnetic Resonance Center (CERM), University of Florence, Italy
2 Department of Chemistry, University of Florence, Italy
Introduction
The first member of the family of Sco (synthesis of
cytochrome c oxidase) proteins was identified in yeast
as a gene product essential for accumulation of the
mitochondrially synthesized subunit II (Cox2) of
cytochrome c oxidase (COX) [1] COX is the terminal
component of the respiratory chain, located in the
inner mitochondrial membrane of eukaryotes and in
the plasma membrane of many prokaryotes The
catalytic core of the enzyme is composed of the three
largest subunits (Cox1, Cox2 and Cox3), which are
highly conserved between prokaryotes and eukaryotes [2] Both Cox1 and Cox2 contain metal cofactors which are required for COX to function, and include one copper ion in Cox1 (termed CuB) and two copper ions forming a dinuclear centre in Cox2 (termed CuA) [3] The CuA centre acts as the primary acceptor of electrons coming from cytochrome c, which are then transferred, via a low-spin heme a moiety, to the catalytic site formed by CuB and a high-spin heme a3 where oxygen binding and reduction take place [4]
Keywords
copper; cytochrome c oxidase; redox; Sco;
thiol-disulfide
Correspondence
I Bertini, Magnetic Resonance Center,
University of Florence, Via Luigi Sacconi 6,
50019 Sesto Fiorentino, Italy
Fax: +39 055 457 4271
Tel: +39 055 457 4272
E-mail: ivanobertini@cerm.unifi.it
(Received 15 February 2011, revised 12
April 2011, accepted 18 April 2011)
doi:10.1111/j.1742-4658.2011.08141.x
Sco proteins are present in all types of organisms, including the vast major-ity of eukaryotes and many prokaryotes It is well established that Sco pro-teins in eukaryotes are involved in the assembly of the CuA cofactor of mitochondrial cytochrome c oxidase; however their precise role in this pro-cess has not yet been elucidated at the molecular level In particular, some but not all eukaryotes including humans possess two Sco proteins whose individual functions remain unclear There is evidence that eukaryotic Sco proteins are also implicated in other cellular processes such as redox signal-ling and regulation of copper homeostasis The range of physiological functions of Sco proteins appears to be even wider in prokaryotes, where Sco-encoding genes have been duplicated many times during evolution While some prokaryotic Sco proteins are required for the biosynthesis of cytochrome c oxidase, others are most likely to take part in different processes such as copper delivery to other enzymes and protection against oxidative stress The detailed understanding of the multiplicity of roles ascribed to Sco proteins requires the identification of the subtle determi-nants that modulate the two properties central to their known and poten-tial functions, i.e copper binding and redox properties In this review, we provide a comprehensive summary of the current knowledge on Sco proteins gained by genetic, structural and functional studies on both eukaryotic and prokaryotic homologues, and propose some hints to unveil the elusive molecular mechanisms underlying their functions
Abbreviations
BsSco, apo-Sco from Bacillus subtilis; IMS, intermembrane space; Trx, thioredoxin.
Trang 2The two copper ions in CuA are coordinated by two
bridging Cys sulfur atoms, two His nitrogen atoms,
and 2 weak ligands provided by a Met sulfur and a
backbone carbonyl oxygen [5] The highly covalent
and rigid Cu2S2 core of the CuA centre is considered
an important feature in determining its efficiency in
long-range electron transfer by virtue of a low
reorga-nization energy [6] Another important factor in this
respect is the electronic structure of CuA, which cycles
between a reduced Cu(I)–Cu(I) state and an oxidized
species consisting of a fully delocalized mixed-valence
pair with two equivalent Cu1.5+ions [7,8]
The proposal that Sco proteins could play a role in
copper delivery to COX within the process of COX
assembly was first formulated based on the observation
that their overexpression could rescue respiratory
defi-ciency in yeast mutants lacking the copper chaperone
Cox17 [9], a low molecular weight protein that is
local-ized within the cytoplasm and the mitochondrial
inter-membrane space (IMS) [10,11] Many subsequent
studies in eukaryotic organisms were performed along
the lines of this hypothesis, and contributed to drawing
a picture where Sco proteins function in COX
assem-bly by mediating copper transfer from Cox17 to the
CuA site of Cox2 [12] The details of the mechanism
by which Sco proteins accomplish this function,
how-ever, remain a controversial issue, which is complicated
by the fact that different mechanisms appear to
oper-ate in different organisms Long recognized evidence in
this sense comes from the observation that two Sco
proteins (Sco1 and Sco2) playing distinct roles are
required for maturation of the CuA site in humans
[13,14], whereas yeast, despite having two Sco proteins
as well, needs only one of them [9,15] Furthermore, to
make the matter more puzzling, the human proteins
have been proposed to fulfil additional functions
besides COX assembly, including mitochondrial redox
signalling [16] and regulation of copper homeostasis
[17]
Sco proteins are also found in prokaryotic
organ-isms, leading to the widespread postulation that their
function in COX assembly is conserved between
eukaryotes and prokaryotes [18] Although this
assumption is supported by experimental data, the
pre-cise mode of action of Sco proteins in the insertion of
copper into Cox2 is as uncertain in prokaryotes as it is
in eukaryotes, and can also differ in different
organ-isms [19] In addition, prokaryotic Sco proteins have
also been implicated in functions that are unrelated to
COX assembly, such as in regulation of gene
expres-sion [20] and in protection against oxidative stress [21]
The functional divergence of Sco proteins in
prokary-otes is apparent from the analysis of their genomes,
some of which contain genes coding for Sco proteins without having any genes coding for Cox2 [22]
By bringing together the available data on eukary-otic and prokaryeukary-otic Sco proteins, a complex scenario therefore emerges in which major questions arise as to which is the ancestral function of Sco proteins, how (and how many) other functions have evolved from that, and to what extent the mechanisms operating in prokaryotes are related to, and thus can be used to understand, those active in the more complex eukary-otes The answers to these questions involve the description of the molecular determinants that underlie the specific functional mechanisms of these proteins In this work, we review the current knowledge on pro-karyotic and eupro-karyotic Sco proteins, with the aim of providing a framework to rationalize the various func-tions of these proteins and the elusive factors that determine these functions
Occurrence and sequence features
of Sco proteins in eukaryotes and prokaryotes
To date, no systematic analysis of eukaryotic genomes has been carried out to identify genes that encode Sco proteins In the most comprehensive survey available, Sco-encoding genes (Sco genes hereafter) were identi-fied in 39 eukaryotic species and their exon–intron structure was examined to reconstruct their evolution-ary history [23] This analysis showed that eukevolution-aryotic Sco genes all descend from an ancestral gene already present in the last common ancestor of lineages that diverged as early as metazoans and flowering plants, i.e more than 900 million years ago Also, it showed that the genomes of vertebrates and flowering plants contain two Sco genes, which derive from two inde-pendent duplication events To complement and extend these data, we have searched Sco genes in a total of 66 eukaryotic species (27 animals, 18 fungi, 9 plants and
12 protists) including, in addition to those examined in [23], all species whose complete genome sequences are available at the NCBI as of December 2010 (http:// www.ncbi.nlm.nih.gov/genomes/leuks.cgi) A summary
of our results is shown in Table 1
Sco genes have been found in 61 of the 66 eukaryotes analysed, with the exceptions of the microsporidia Encephalitozoon cuniculi and Encephalitozoon
intestinal-is, the amoebae Entamoeba dispar and Entamoeba histolytica, and the apicomplexan Cryptosporidium parvum The absence of Sco genes in these organisms is not unexpected, as all of them are obligate intracellular parasites that contain degenerated mitochondria called mitosomes, which lack many of the functions of
Trang 3Table 1 Occurrence of genes encoding Sco proteins in eukaryotic organisms, sorted by taxonomic group Organisms that were analysed in [23] are highlighted in grey Gene and protein IDs reported as not available (n ⁄ a) indicate genes that were identified in [23] but not by our search (presumably due to incomplete genome sequences) For the number of Sco genes in Pan troglodytes, see text.
100497895
XP_002937049.1 XP_002935088.1
n ⁄ a
NP_001038697.1
n ⁄ a
n ⁄ a
n ⁄ a
n ⁄ a
100125923
NP_001073712.1 NP_001098963.1
9997
NP_004580.1 NP_005129.2
722074
XP_001116350.2 XP_001118271.1
100126824
NP_001035115.1 NP_001104758.1
100517855
XP_003126813.1 XP_003132044.1
2889365
XP_445160.1 XP_447458.1
852325
NP_009580.1 NP_009593.1
Trang 4canonical mitochondria including oxidative
phosphory-lation [24–26] This observation is thus in full agreement
with the notion that the primary function of eukaryotic
Sco proteins is in mitochondrial COX assembly Our
results also confirm that plants and vertebrates have
two Sco genes with the conspicuous exception of
Pan troglodytes, for which only a Sco1 homologue has
been found However, a tblastn search in the P
troglo-dytesgenome using human Sco2 as the query sequence
reveals a close match in a region of chromosome 22
(NCBI locus NW_001231014, between nucleotides
49882300 and 49882500) where no genes are thought to
reside It is therefore most likely that P troglodytes also
has a Sco2 homologue, whose recognition has been
hindered by an error in the current genome assembly
(Build 2.1)
In addition to plants and vertebrates, multiple Sco genes also occur in the fungi Saccharomyces cerevisiae and Candida glabrata, which have two such genes, and
in kinetoplast protozoa, which have three (apart from Leishmania braziliensis, which has two) A neighbour-joining tree built from the multiple alignment of all the Sco proteins identified (Fig 1) indicates that indepen-dent duplications occurred (a) in a common ancestor
of vertebrates, (b) in a common ancestor of land plants, (c) in a common ancestor of S cerevisiae and
C glabrata, and possibly of other fungi, and (d) in a common ancestor of kinetoplasts, where two duplica-tions occurred This scenario implies that in eukaryotes containing two or three Sco proteins these proteins have distinct physiological functions, which are not nec-essarily the same in organisms belonging to different
Table 1 (Continued).
830129
NP_566339.1 NP_568068.1
4346889
NP_001045964.1 NP_001063017.1
7471514
XP_002323592.1 XP_002306313.1
8074665
XP_002462290.1 XP_002453341.1
100247202
XP_002266556.1 XP_002263427.1
100282683
NP_001130056.1 NP_001149062.1
5416541
XP_001561796.1 XP_001562419.1
5068632 5069910
XP_001462953.1 XP_001465217.1 XP_001470536.1
5651436 5653126
XP_888624.1 XP_001682836.1 XP_001684203.1
3660582 4357233
XP_803555.1 XP_827193.1 XP_001218860.1
3537405 3540368
XP_805842.1 XP_807216.1 XP_809712.1
Trang 5Homo sapiens|NP 004580.1 Pan troglodytes|XP 001164786.1 1000
Macaca mulatta|XP 001118271.1 1000
Bos taurus|NP 001073712.1 Sus scrofa|XP 003132044.1 1000
979
Mus musculus|NP 001035115.1 1000
Xenopus tropicalis|XP 002937049.1 855
Branchiostoma floridae|XP 002613836.1 678
Strongylocentrotus purpuratus|XP 001199433.1 458
Ixodes scapularis|XP 002402970.1 Nematostella vectensis|XP 001641939.1 164
319
Drosophila melanogaster|NP 608884.1 Drosophila simulans|XP 002078186.1 1000
Drosophila virilis|XP 002051936.1 1000
Anopheles gambiae|XP 314900.3 979
Tribolium castaneum|XP 969355.1 830
Apis mellifera|XP 001122061.1 Nasonia vitripennis|XP 001605752.1 997
420
Pediculus humanus corporis|XP 002424862.1 Acyrthosiphon pisum|NP 001156100.1
418
Hydra magnipapillata|XP 002156667.1 317
347
253
Homo sapiens|NP 005129.2 Macaca mulatta|XP 001116350.2 1000
Bos taurus|NP 001098963.1 Sus scrofa|XP 003126813.1 898
866
Mus musculus|NP 001104758.1 1000
Xenopus tropicalis|XP 002935088.1 1000
394
Caenorhabditis elegans|NP 494755.1 841
Sorghum bicolor|XP 002453341.1 Zea mays|NP 001130056.1 1000
Oryza sativa|NP 001045964.1 1000
Vitis vinifera|XP 002263427.1 644
Arabidopsis thaliana|NP 566339.1 771
Populus trichocarpa|XP 002323592.1 1000
Ostreococcus lucimarinus|XP 001422358.1 Ostreococcus tauri|XP 003084388.1
1000
Micromonas RCC299|XP 002508419.1 932
865
299
Plasmodium falciparum|XP 001349003.1 Plasmodium knowlesi|XP 002257566.1 1000
Theileria annulata|XP 952475.1 959
Sorghum bicolor|XP 002462290.1 Zea mays|NP 001149062.1 1000
Oryza sativa|NP 001063017.1 1000
Arabidopsis thaliana|NP 568068.1 Populus trichocarpa|XP 002306313.1 770
Vitis vinifera|XP 002266556.1 998
1000
Leishmania infantum|XP 001470536.1 Leishmania major|XP 001684203.1 1000
Leishmania braziliensis|XP 001562419.1 1000
Trypanosoma brucei|XP 803555.1 Trypanosoma cruzi|XP 807216.1 1000
1000 410
Leishmania infantum|XP 001462953.1 Leishmania major|XP 888624.1 1000
Leishmania braziliensis|XP 001561796.1 1000
Trypanosoma brucei|XP 827193.1 Trypanosoma cruzi|XP 805842.1 946
1000
Leishmania infantum|XP 001465217.1 Leishmania major|XP 001682836.1 1000
Trypanosoma cruzi|XP 809712.1 806
Trypanosoma brucei|XP 001218860.1 1000
413
535
240
Ashbya gossypii|NP 984670.2 Lachancea thermotolerans|XP 002551538.1 752
Kluyveromyces lactis|XP 453226.1 544
Zygosaccharomyces rouxii|XP 002494544.1 483
Candida glabrata|XP 447458.1 Saccharomyces cerevisiae|NP 009593.1 569
986
Pichia pastoris|XP 002493635.1 558
Debaryomyces hansenii|XP 002769958.1 Pichia stipitis|XP 001383869.2 990
Candida dubliniensis|XP 002422402.1 950
752
Candida glabrata|XP 445160.1 Saccharomyces cerevisiae|NP 009580.1 962
879
Yarrowia lipolytica|XP 504291.1 500
Aspergillus nidulans|XP 662446.1 Aspergillus oryzae|XP 001824537.2 1000
Magnaporthe oryzae|XP 366930.1 990
628
Schizosaccharomyces pombe|NP 595287.1 833
Cryptococcus neoformans|XP 572441.1 771
Monosiga brevicollis|XP 001742585.1
Sco2 vertebrates
Sco2 land plants
Kinetoplasts
Animals
Sco1 land plants
Green algae Apicomplexans Sco1 vertebrates
Fungi Sco1 fungi
Sco2 fungi
Trang 6kingdoms, however As mentioned in the Introduction,
it is indeed well known that the physiological roles of
Sco2 in humans and yeast must be diverse It would
then be useful to assess experimentally the roles of the
duplicated proteins in kinetoplasts and land plants as
well In particular, determining the function of the
duplicated plant proteins would be especially
interest-ing, as in all plants one of the two proteins (indicated
as ‘Sco2 land plants’ in Fig 1) lacks the characteristic
CXXXC motif present in all the other Sco proteins
and is thus presumably unable to bind copper (a
CXXXG motif is found in the proteins from Oryza
sa-tiva, Sorghum bicolor and Zea mays, and a SXXXG
motif in those from Arabidopsis thaliana,
Popu-lus trichocarpaand Vitis vinifera)
The distribution of Sco proteins across prokaryotic
species is far more variable than in eukaryotes A
bioin-formatics analysis of 311 prokaryotic genomes (285
from Bacteria and 26 from Archaea) revealed that Sco
proteins are present in a large variety of species from
both Bacteria and Archaea, which in most cases (65 out
of 128, i.e about 51%) have more than one Sco gene,
and can have up to seven [22] On the other hand, 183
of the 311 organisms analysed (i.e about 59%) were
found to contain no Sco homologues, which appear to
be lacking altogether in some prokaryotic groups such
as cyanobacteria [22] In particular, by searching for
the co-occurrence in genomes of Sco and Cox2 genes, it
was pointed out that about 12% of prokaryotes have
Cox2 but not Sco genes, and about 6% have Sco but
not Cox2 genes These observations imply that some
prokaryotes (including for example cyanobacteria)
evolved a process of COX maturation where Sco is not required and, on the other hand, that Sco proteins in prokaryotes can also function outside COX assembly The supposedly broader functional range of prokary-otic Sco proteins with respect to their eukaryprokary-otic coun-terparts, which is most likely to be found in those prokaryotes where multiple duplications of Sco genes have occurred, is reflected in the higher variability of their amino acid sequences (21 ± 9% average pairwise identity in prokaryotes versus 38 ± 9% in eukaryotes)
A comparison of the sequence profiles based on the multiple alignments of eukaryotic and prokaryotic Sco proteins, respectively, in fact shows that highly con-served residues are more numerous in eukaryotes than
in prokaryotes, and are especially concentrated in the regions forming the copper-binding site (Fig 2) How-ever, prokaryotic and eukaryotic sequences appear to share most of their major features: all the residues that are highly conserved in prokaryotes, including copper ligands and two aspartates in a DXXXD motif, are present and highly conserved in eukaryotes as well, and the additional highly conserved residues in eukaryotes are generally those found most frequently (though being more variable) in the corresponding positions in prokaryotic sequences In this respect, the most remark-able differences are the presence in eukaryotes of a DEXXK motif downstream of the CXXXC motif which has no counterpart in prokaryotes, and two other changes also involving the occurrence of charged resi-dues in eukaryotes in the place of non-polar resiresi-dues in prokaryotes (Glu and Arg for Ala and Gly, respec-tively; see Fig 2)
Fig 2 Profile–profile comparison of eukaryotic and prokaryotic Sco protein sequences obtained using the program HHSEARCH [94] The profile
of eukaryotic sequences was constructed from their multiple alignment using the program HMMER [95], while that of prokaryotic sequences was taken from [22] Highly conserved residues (i.e residues occurring at a given position with probability > 0.5) are shown in bold Copper-binding residues are highlighted in yellow Positions where the two profiles differ most are highlighted in red.
Fig 1 Neighbour-joining tree built (using the program CLUSTALW [91]) from the multiple alignment of eukaryotic Sco proteins (constructed using the program MUSCLE [92]) Relevant subgroups are shown Numbers on branches are bootstrap values based on 1000 replicates The tree was visualized with the program TREEVIEW [93].
Trang 7Structural studies on eukaryotic and
prokaryotic Sco proteins: hints for
function
To date the three-dimensional structures of human
Sco1 and Sco2 have been determined in their apo- and
metal-loaded states Specifically, apo-, Cu(I)- and
Ni(II)-Sco1 [16,27] and Cu(I)-Sco2 [28] structures are
available From all these data it emerges that the
over-all structure contains a typical thioredoxin (Trx) fold
[29] with the insertion of further secondary structure
elements The Trx fold is constituted by a central
four-stranded b sheet (b3, b4, b6, b7) and three flanking a
helices (a1, a3, a4) (Fig 3) On this scaffold, a
b-hair-pin structure (b1 and b2) followed by a 310-helix (h1)
is inserted at the N terminus and one helix (a2)
followed by a strand (b5), the latter forming a parallel
b sheet with strand b4, are inserted between strand b4
and helix a3 (Fig 3) This fold topology belongs to a
subset of the Trx superfamily, present in
peroxiredox-ins and glutathione peroxidases [30] A specific
property of the eukaryotic Sco fold, absent in Trx and
Trx-like family members, is the presence of a b-hairpin
in the extended, solvent-exposed loop connecting helix
a3 and strand b6 (Fig 3) All of these points of
inser-tion are those typically tolerated in a Trx fold without
disruption of the overall structure [29]
A comparison of the structural and dynamic
proper-ties of apo- versus metal-loaded states of human Sco1
and Sco2 provides a detailed molecular view of the
metal-binding process In the apo forms, a large
num-ber of residues in the metal-binding area sample, in
solution, multiple local conformational states
exchang-ing with each other on the intermediate or slow NMR timescale (ls to ms) [27] (Fig 4) This effect is particu-larly observed in human apo-Sco2 which indeed, at variance with human apo-Sco1, displays a conforma-tional heterogeneity involving, in addition to the metal-binding site region, also the b sheet and the sur-rounding a helices which constitute the protein core of Sco2 [28] Cu(I) binding, however, is in both Sco pro-teins able to ‘freeze’ the above regions in an ordered, more rigid conformation (Fig 4) This behaviour can
be rationalized taking into account the spatial location
of metal ligands Cu(I) ion is in fact coordinated by the two Cys residues of the CXXXC conserved motif, located in loop 3 and helix a1, and by a conserved His which is far in the sequence from the CXXXC motif, i.e in the b-hairpin present in the extended, solvent-exposed loop (Fig 4) Therefore, the involvement in the metal-binding site of residues from two different regions of the protein contributes to produce a com-pact structure of the metal-loaded protein state with respect to the apo form The large conformational var-iability of the His-containing loop observed in the apo-Sco1 solution structure [27] indicates that backbone structural changes are necessary to locate the metal ligand His260 in the vicinity of the other two ligands, Cys169 and Cys173 This behaviour is also confirmed
by the crystal structures [16,27,31] Even if the loop segments of apo-Sco1 have a continuous electron den-sity with similar backbone conformations in all three independent molecules of the crystal [16], the loop
N
β2
C
θ1
β3 β4
β5 α1
α4
α3
β6
β1
Fig 3 Schematic picture of the fold topology of Sco proteins The
secondary structure elements of a typical thioredoxin fold are
shown in grey Additional secondary structure elements present in
Sco proteins (inserted at the N-terminus and between strand b4
and helix a3) are shown in red A specific property of the eukaryotic
Sco fold is the presence of an extended, solvent-exposed loop
con-taining a b-hairpin (shown in green) connecting helix a3 and strand
b6.
+Cu(I)
Fig 4 Illustration of how metal binding ‘freezes’ the conforma-tional heterogeneity of the metal-binding region in Sco proteins From an apo state characterized by conformational disorder in the CXXXC motif and the loop containing the histidine ligand, one com-pact conformer with the appropriate metal-ligand distances is selected upon metal addition The cysteine ligands are shown in yellow, the histidine ligand in blue, and the Cu(I) ion in light blue.
Trang 8acquires a more ordered conformation as a
conse-quence of Ni(II) binding [27] This higher order is
rec-ognized by a definition of the electron density map in
that region for both molecules of the asymmetric unit
of Ni(II)-Sco1 higher than that in the apo-Sco1 crystal
structure A further confirmation comes from the
sig-nificantly lower temperature factors of the atoms
belonging to the His-containing loop in the structure
of Ni(II)-Sco1 with respect to those in the apo-Sco1
structure Crystallization therefore most likely selects,
in apo-Sco1, the lowest-energy conformers between the
multiple ones present in solution Backbone
conforma-tional changes to allow the formation of a
Cu(I)-bind-ing site appear to be necessary also in the crystallized
apo-Sco1 state, in agreement with the demand of a
conformational sampling of Sco1 to bind the metal
ion
The coordination sphere of Ni(II) in the crystal
structure of human Ni(II)-Sco1 [27] is quite peculiar In
this structure, the two metal-binding Cys residues are
oxidized and form a disulfide bond and therefore are
not capable of binding the Ni(II) ion as thiolates Still,
the metal ion remains in contact with the S—S bond
with an Ni—S distance of 2.0–2.2 A˚, suggesting the
for-mation of bonds with the lone pairs of the sulfur atoms
The coordination sphere of Ni(II) is completed by
His260 (Ne2—Ni, 2.03–2.45 A˚), in agreement with the
solution structure of Ni(II)-Sco1 and a water molecule,
or more likely an anion such as chloride, arranged in a
distorted square planar geometry The redox state of
the Cys ligands differs from that found in Ni(II)-Sco1,
Cu(I)-Sco1 and Cu(I)-Sco2 solution structures, which
have both cysteines in the reduced state [27] This
differ-ent behaviour is due to the differdiffer-ent experimdiffer-ental
con-ditions, i.e aerobic versus anaerobic, and the presence
of 1 mm dithiothreitol disulfide-reductant The only
other available crystal structure of a metal-loaded Sco1
form (yeast Sco1) [31] also shows a quite unexpected
metal coordination The crystal has been obtained by
soaking apo-Sco1 crystals with copper ions and its
structure reveals a copper-binding site involving Cys181
and Cys216, two cysteine residues present in yeast Sco1
but not conserved in human Sco1 and Sco2 and not
belonging to the conserved CXXXC motif A possible
explanation of this result is that the soaking solution
contained Cu(II) rather than Cu(I) ions, and the Cu(II)
ions could then have catalysed oxidation of the
conserved cysteines, which therefore cannot bind
cop-per The copper ion was then bound at an adventitious
site formed by the non-conserved Cys181 and Cys216
plus the conserved His239 in the flexible long loop
These structural data on eukaryotic Sco proteins
indicate that, despite the full conservation of the three
metal-binding ligands, the metal-binding site has an intrinsic structural flexibility, indicating the absence of
a binding site structurally well organized to receive the metal The latter property can thus explain the efficient formation of a disulfide bond between the Cys ligands and the movement of the His ligand towards a copper-binding site located in a different position with respect
to the typical metal-binding site of the Sco proteins The His-ligand-containing loop indeed displays the largest backbone fluctuations from the apo- to the Cu(I)-bound state, positioning the imidazole ring of His260 about 10 A˚ from the sulfur atoms of the metal-binding Cys residues, in apo-Sco1 However, from an open apo conformation with local disorder, the struc-ture converts, upon metal binding, into a well defined compact state In particular, the His ligand coordina-tion is the key event which modulates the ordered⁄ dis-ordered state of the whole metal-binding region (Fig 4) Taking into account that disordered regions
in protein structures are often engaged in protein–pro-tein interactions, one may speculate that this loop modulates association–dissociation of Sco1 with its partners For example, it is possible that, once the mitochondrial copper chaperone Cu(I)-Cox17 interacts transiently with apo-Sco1 and donates its copper cargo
to Sco1, the His-containing loop structurally rear-ranges, thus allowing His binding and concomitant formation of the compact Cu(I)-Sco1 structure The formation of the stable compact Cu(I)-Sco1 state could thus constitute the important driving force of the cop-per transfer from Cox17 to Sco1
Solution and crystal structures are also available for prokaryotic Sco homologues Specifically, the solution structure of apo-Sco from Bacillus subtilis (BsSco), solved in 2003, was the first for this class of proteins [32] and, later, its crystal structure [33] as well as the solution structure of the Sco1 homologue from Ther-mus thermophilus [19] were solved Despite numerous efforts, solution or crystal structures of the copper forms of these prokaryotic Sco proteins were not obtained However, a combination of spectroscopic techniques was used to find that BsSco employs a sin-gle metal site to bind both Cu(I) and Cu(II) [34–37], the former via two cysteines plus a weakly bound, unidentified ligand, and the latter via two cysteines with unequal bond strengths, two O⁄ N donor ligands including at least one histidine, and possibly a weakly bound water molecule
Both NMR and crystallographic data on BsSco show structural properties very similar to those found for eukaryotic apo-Sco proteins Backbone conformational exchange processes have been detected in solution for the CXXXC metal binding motif and the
Trang 9His-containing loop of BsSco Accordingly, the RMSD
between the His-containing loop of the two
crystallo-graphically independent molecules A and B is quite
high (4.66 A˚ compared with an overall value of
0.14 A˚) Also, the average temperature B-factor of this
loop is 53.56, compared with the average B-factor of
30.81 for molecule A and 31.23 for molecule B, and the
loop has a weak electron density map The
CXXXC-containing loop also exhibits differences between
molecule A and molecule B (RMSD of 1.56 A˚),
although much less than the His-containing loop, with
the average B-factor (49.40) also higher than the protein
average In some structures of apo-BsSco obtained in
the presence of copper, a disulfide bridge is observed
between the Cys of the CXXXC motif, similarly to
what occurs for yeast Sco1 There seems to be only a
small energy barrier separating the disulfide-bonded
and non-disulfide-bonded conformations Taken
together, these observations indicate that (as in
eukary-otic apo-Sco proteins) both metal-ligand-containing
loops implicated in copper binding exhibit
conforma-tional plasticity in the structure However, an important
structural difference of the metal-binding region of
apo-BsSco with respect to the corresponding region of
eukaryotic apo-Sco proteins is that, in the latter, one
cysteine is located at one end of an a helix and the other
is in the preceding short loop region, and the thiolate
groups of the cysteines are only partially exposed In
contrast, the two Cys residues in BsSco are located in a
protruding loop that is exposed to the solvent The
backbone conformation of the His-containing loop,
which does not have the b-hairpin present in eukaryotic
Sco proteins, is also such that it largely exposes the His
ligand to the solvent in BsSco only These differences
suggest that apo-BsSco has a greater structural
flexibil-ity than human Sco proteins, which can account for the
impossibility of getting a compact Cu(I) state even upon
Cu(I) addition This indicates that specific amino acid
substitutions in critical points of the fold can largely
affect the structural flexibility of the metal-binding
region of Sco proteins, and as a consequence their
cop-per affinity Cu(II) binding, however, is able to
deter-mine in BsSco the formation of a complex with extreme
kinetic stability and picomolar affinity [35,36,38] A
two-step model for Cu(II) binding has been proposed in
which a rapidly formed intermediate state of
Cu(II)-BsSco, with low-micromolar metal affinity, is then
slowly converted into the stable final Cu(II)-bound
form [38] However, high ionic strength can induce
destabilization of the Cu(II)-BsSco complex and metal
release, indicating that structural flexibility of the
metal-binding site can be easily promoted also in this
case [36] In a physiological context, it could be possible
that, for BsSco as well as for human Sco proteins, the interactions with a specific protein partner can induce conformational changes of the metal-binding site, thus promoting the metal release to the CuAsite
Eukaryotic Sco proteins in the
In eukaryotes, a large number of nuclear genes are required for the proper assembly and function of COX [39] The most thoroughly characterized aspect of COX assembly is that of mitochondrial copper delivery
to the nascent holoenzyme complex, and in particular delivery of copper to the CuA site Such process involves Sco proteins, specifically two highly homolo-gous members of the family, Sco1 and Sco2, and Cox17 Solution structure of the latter protein shows that a highly conserved twin Cx9C motif forms two disulfide bonds which are essential for the formation of
an a-hairpin fold [40–42] The oxidoreductase Mia40 is responsible in the IMS for promoting both the forma-tion of the two disulfides and the folding of Cox17 [43–45] A flexible and completely unstructured N-ter-minal tail of Cox17 contains a CC motif which coordi-nates one Cu(I) ion [40] It was shown that Cu(I) binding is essential to Cox17 function and that, in spite
of its dual localization, the proposed functional role of Cox17 in mitochondrial copper delivery to COX is restricted to the IMS [46,47] A high copy suppressor screen of a yeast Cox17 null strain led to the identifica-tion of the two highly homologous proteins Sco1 and Sco2 [9] Both proteins are imported in the IMS through the TOM translocase which recognizes a typical mitochondrial targeting sequence present at the N-terminus of Sco1 and Sco2 [48] Then, they are arrested in the mitochondrial inner membrane through
a stop-transfer mechanism, in which a transmembrane helix, subsequent to the mitochondrial targeting sequence in both proteins, functions as a critical sort-ing signal that causes the arrest of the precursor dursort-ing the import reaction at the level of the inner membrane
as well as in its lateral insertion into the lipid bilayer, both processes being mediated by the TIM23 translo-case [48]
Yeast Sco1 is absolutely required in the activation
of COX [1,49] and in vitro it can receive copper from Cox17 [50], indicating that Sco1 functions downstream
of Cox17 in copper delivery to COX Copper-binding properties [51], mutational analysis of the metal-bind-ing CXXXC motif [52] and physical interactions with Cox2 [53] suggested that Sco1 specifically delivers cop-per to the CuA site in the Cox2 subunit [52,53] A ser-ies of conserved residues on the leading edge of the
Trang 10His-containing loop have been suggested to be
impli-cated in Cox2 interaction, but not in the interaction
with Cox17, thus indicating different surfaces on Sco1
for the interaction with Cox17 and Cox2 [54] The
copper transfer from Sco1 to Cox2 has never been
directly observed in vitro as all the attempts to stabilize
eukaryotic Cox2 domains have been unsuccessful so
far [55] At variance with Sco1 mutants, yeast Sco2
mutants lack an obvious phenotype associated with
respiration, even if, similarly to Sco1, Sco2 interacts
with the C-terminal portion of Cox2 [56] Although
Sco2, like Sco1, can restore respiratory growth in the
Cox17 null mutant, rescue in this case requires
addi-tion of copper to the growth medium [9] Sco2 does
not suppress a Sco1 null mutant, although it is able to
partially rescue a Sco1 point mutant [9] The ability of
Sco2 to restore respiration in Cox17 but not Sco1
mutants is taken as an indication that Sco1 and Sco2
have overlapping but not identical functions Most
parts of yeast Sco1 (N-terminal portion amino acids
1–75 and C-terminal portion amino acids 106–295) can
indeed be replaced by the corresponding parts of yeast Sco2 without loss of function, but a short stretch of 13 amino acids, immediately adjacent to the transmem-brane region, is crucial for Sco1 function and cannot
be replaced by its Sco2 counterpart [52,56] In summary, the Sco2 function in yeast still remains elusive
In contrast to yeast, both Sco1 and Sco2 are required in humans for cellular respiration and CuA biogenesis [13,57] They have non-overlapping, cooper-ative functions in copper delivery to the CuAsite [13] Both human Sco1 and Sco2 are copper-binding pro-teins [58] and have an affinity for Cu(I) higher than that of human Cox17 [55]; accordingly, Cu(I) is quanti-tatively transferred from Cox17 to Sco1 [59] and Sco2 [60] (Fig 5) They also have a similar affinity for Cu(I) [55] and can rapidly exchange it with each other [60] (Fig 5) These data therefore strongly indicate that both human Sco proteins can receive Cu(I) from Cox17 to donate it to the apo-CuAsite, thus determin-ing the formation of the final active centre (Fig 5) In
Cu(I)Cox17
apoCox17 3S-S
2GSH GSSG
Cu, 2e –
Cu(I)Sco1
2e – , Cu(I) apoCox2
Cu(I)Sco2
Cu(I)Cox17
Cu(I)
apoCox17 2S-S Cu(I)Cox17
Cu(I)
Cu(I), 2e–
Cu, 2e–
apoCox17 2S-S
Cytoplasm
IMS
Matrix
Fig 5 Pathway of copper insertion into the Cu A site of COX in humans The structures of Cox2 and of Cox17, Sco1 and Sco2 in their differ-ent metal or redox states are shown Cysteine residues involved in copper binding or disulfide bond formation are shown as yellow sticks Histidine Cu(I) ligands in Sco1 and Sco2 are shown as blue sticks Copper ions are shown as magenta spheres Cox17 can simultaneously transfer Cu(I) ion and two electrons to metallate oxidized apo-Sco1 Sco1 and Sco2 can act as copper chaperones and⁄ or thioredoxins, being implicated in copper transfer to apo-Cox2 and in a disulfide exchange reaction from Sco2 to Sco1 and toward apo-Cox2 Cys-reduced states
of Sco1 and Sco2 are also able to exchange Cu(I) between each other.