Matura-tion of c-type cytochromes requires covalent attachment of the heme cofac-tor to the protein, and there are at least five distinct biogenesis systems that catalyze this post-transl
Trang 1Order within a mosaic distribution of mitochondrial c-type cytochrome biogenesis systems?
James W A Allen1, Andrew P Jackson2, Daniel J Rigden3, Antony C Willis4, Stuart J Ferguson1 and Michael L Ginger5,6
1 Department of Biochemistry, University of Oxford, UK
2 Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK
3 School of Biological Sciences, University of Liverpool, UK
4 MRC Immunochemistry Unit, Department of Biochemistry, University of Oxford, UK
5 Sir William Dunn School of Pathology, University of Oxford, UK
6 Department of Biological Sciences, Lancaster University, UK
Keywords
bioinformatics; Ccm system; cytochrome c;
Diplonema papillatum; evolution; heme
lyase; lateral gene transfer; mitochondria;
post-translational modification; Trypanosoma
Correspondence
M Ginger, Department of Biological
Sciences, Lancaster University, Lancaster
LA1 4YQ, UK
Fax: +44 1524 593192
Tel: +44 1524 593922
E-mail: m.ginger@lancaster.ac.uk
(Received 7 February 2008, revised 3 March
2008, accepted 5 March 2008)
doi:10.1111/j.1742-4658.2008.06380.x
Mitochondrial cytochromes c and c1 are present in all eukaryotes that use oxygen as the terminal electron acceptor in the respiratory chain Matura-tion of c-type cytochromes requires covalent attachment of the heme cofac-tor to the protein, and there are at least five distinct biogenesis systems that catalyze this post-translational modification in different organisms and organelles In this study, we use biochemical data, comparative genomic and structural bioinformatics investigations to provide a holistic view of mitochondrial c-type cytochrome biogenesis and its evolution There are three pathways for mitochondrial c-type cytochrome maturation, only one
of which is present in prokaryotes We analyze the evolutionary distribu-tion of these biogenesis systems, which include the Ccm system (System I) and the enzyme heme lyase (System III) We conclude that heme lyase evolved once and, in many lineages, replaced the multicomponent Ccm sys-tem (present in the proto-mitochondrial endosymbiont), probably as a con-sequence of lateral gene transfer We find no evidence of a System III precursor in prokaryotes, and argue that System III is incompatible with multi-heme cytochromes common to bacteria, but absent from eukaryotes The evolution of the eukaryotic-specific protein heme lyase is strikingly unusual, given that this protein provides a function (thioether bond forma-tion) that is also ubiquitous in prokaryotes The absence of any known c-type cytochrome biogenesis system from the sequenced genomes of various trypanosome species indicates the presence of a third distinct mito-chondrial pathway Interestingly, this system attaches heme to mitochon-drial cytochromes c that contain only one cysteine residue, rather than the usual two, within the heme-binding motif The isolation of single-cysteine-containing mitochondrial cytochromes c from free-living kinetoplastids, Euglena and the marine flagellate Diplonema papillatum suggests that this unique form of heme attachment is restricted to, but conserved throughout, the protist phylum Euglenozoa
Abbreviations
ccm, cytochrome c maturation; EF-1a, elongation factor-1a; EST, expressed sequence tag; IMS, intermembrane space; KH test, Kishino– Hasegawa test; LGT, lateral gene transfer; ML, maximum likelihood; SOD, superoxide dismutase.
Trang 2Multiple pathways for heme cysteine
attachment
The c-type cytochromes are characterized by the
cova-lent attachment of heme to the apocytochrome through
thioether (carbon–sulfur) bonds (Fig 1) Numerous
examples of distinct c-type cytochromes have been
described in Bacteria and, more recently, in some
Ar-chaea, where they typically function in electron transfer
or at the catalytic sites of certain enzymes [2–9] {In
agreement with the nomenclature proposed in [1], we
refer to the three domains of life as Bacteria (formerly
Eubacteria), Archaea (formerly Archaebacteria) and
Eucarya (the eukaryotes) When using the expression
‘Bacteria’, we therefore refer to the domain; when
using the term ‘bacteria’, we refer generically to
non-archaean prokaryotes} However, the best known
examples from the c-type cytochrome family are
mito-chondrial cytochromes c and c1, which function as
essential electron transfer components of the
respira-tory chain [7,8,10]
The covalent attachment of two vinyl groups from
the heme cofactor to the thiols in the CXXCH
heme-binding motif of apocytochromes c is chemically far
from facile (X is any amino acid, except cysteine), and
there are multiple systems which catalyze this
post-translational modification in biology [2,4,11–15]
Sys-tems I and II are modular and widely distributed
amongst bacteria [2,4,6,12,14]; they have been studied
using a combination of genetic and biochemical
approaches [2,4,14,16,17] System I is understood best
in Escherichia coli, where it consists of eight dedicated essential proteins, named CcmA–H (Fig 2A), and a number of accessory proteins CcmA–H are all mem-brane anchored or integral memmem-brane proteins, and collectively function in the periplasm The biogenesis
of c-type cytochromes is a spatial and temporal prob-lem; in bacteria, both heme and apoprotein are synthe-sized in the cytoplasm and must be transported to the periplasm, where heme attachment occurs The apocy-tochrome polypeptide is translocated by the general type II secretion (Sec) proteins [18] How heme is transported remains an intriguing mystery
CcmA and CcmB are reminiscent of an ATP-dependent (ABC-type) transporter, and CcmA has been shown to hydrolyze ATP [19] However, no transport substrate has yet been identified; heme has been proposed, but much evidence weighs against this possibility [19–22] A more recent hypothesis is that CcmA and CcmB are required to release heme from the heme chaperone CcmE by coupling the free energy gained from ATP hydrolysis [21] CcmE is a key player in the Ccm system; it binds heme cova-lently as an intermediate in the cytochrome c biogen-esis pathway [23] This remarkable heme attachment occurs between a histidine residue and a heme vinyl group Heme attachment to CcmE is dependent on CcmC [24], an integral membrane protein with a number of interesting phenotypes arising from muta-tion in ccmC, some of which may be unrelated to c-type cytochrome biogenesis [25] CcmD is a very small ( 60 amino acids) integral membrane protein
Fig 1 Structures of (A) heme (Fe-protoporphyrin IX) and (B) heme bound to a polypeptide chain as in a typical c-type cytochrome, in which the vinyl groups of the heme are saturated by the addition of cysteine thiols that occur in a Cys-Xxx-Xxx-Cys-His motif (only the sulfur atoms
of the cysteines are shown), forming covalent bonds between heme and protein (C) Cartoon representation of heme attachment to protein
in mitochondrial cytochrome c The porphyrin ring is shown in blue and the heme iron atom in brown The cysteines of the CXXCH motif form covalent bonds to the heme, and the histidine acts as a ligand to the heme iron atom via a nitrogen atom The sixth ligand to the iron atom is the sulfur of a methionine residue located distantly from the CXXCH motif in the primary structure of the protein In bacterial c-type cytochromes, histidine (rather than methionine) is often the sixth iron ligand, and there are examples with cysteine, an N-terminal amino group, asparagine, lysine or a vacant coordination site There are few restrictions on the nature of the Xxx-Xxx residues.
Trang 3that mediates complex formation between CcmC and
CcmE [26] CcmF and CcmH are implicated in the
transfer of heme from holo-CcmE to
apocyto-chrome c, including the covalent heme attachment
step to produce the product holocytochrome E coli
CcmH is a fusion protein which includes the proteins
known as CcmH and CcmI in many bacteria CcmG
is a thioredoxin-like protein [27] that forms part of
an electron transfer chain Electrons are transferred
from the cytoplasmic protein thioredoxin, via the
multidomain membrane protein DsbD, to CcmG,
and then to the apocytochrome to reduce a disulfide bond that forms between the cysteines of the apocyto-chrome CXXCH heme-binding motif; these thiols must
be reduced for heme attachment to occur (reviewed in [2]) Such a reductive pathway is thought to be neces-sary in E coli, partly because the periplasm contains the strong, indiscriminate, disulfide-oxidizing protein DsbA
System II (Fig 2B) is less well understood than System I at the molecular level, but it seems very likely
to consist of four proteins [28] {Note: The
nomencla-Fig 2 Cytochrome c biogenesis systems found in bacteria Each of these systems can mature a wide variety of c-type cytochromes, including those with multiple hemes (A) System I (the Ccm system) in Escherichia coli Some uncertainties are designated with ‘?’; for example, what, if anything, is transported by the ABC-type transporter CcmAB, and how is heme transported from its site of synthesis in the cytoplasm to the periplasm? DsbD has two thiols amongst its eight transmembrane helices which are believed to accept reducing equiv-alents from thioredoxin (TrxA) These thiols, in turn, pass on the reducing power to periplasmic C- and N-terminal domains From there, reductant passes to the c-type cytochrome biogenesis apparatus, tentatively by the route shown; CcmG has been shown in some schemes
to be the electron acceptor from CcmH but, although there is experimental evidence for this order, more evidence indicates the arrange-ment shown in the figure DsbA is a strong, non-specific disulfide bond-oxidizing protein found in the periplasm of E coli Ultimately, the cysteine thiols of the apocytochrome CXXCH heme-binding motif become reduced to allow heme attachment Heme becomes covalently attached to the chaperone CcmE as an intermediate in the pathway The specific covalent attachment of heme to apocytochrome c is believed to involve CcmF and H (B) Cytochrome c biogenesis System II in a Gram-negative bacterium In some species, CcdA is replaced
by the protein DsbD shown in Fig 2A CcdA and ResA provide a pathway by which reductant is transferred to the apocytochrome to reduce
a disulfide bond in the CXXCH heme-binding motif ResB and ResC provide the covalent heme attachment function to produce the product holocytochrome c Heme delivery to the periplasm from the cytoplasm may also occur through the ResBC complex, but this is presently not certain Other names are in common use for ResA ⁄ B ⁄ C (ResA = CcsX = HCF164; ResB = CcsB = Ccs1; ResC = CcsA; and CcdA = CcsC).
Trang 4ture for System II c-type cytochrome biogenesis
pro-teins is somewhat inconsistent in the literature Here,
we adopt the names used for the various biogenesis
proteins found in Bacillus subtilis, i.e ResA (also
called CcsX or HCF164), ResB (also called CcsB or
Ccs1), ResC (also called CcsA) and CcdA (also called
CcsC)} These include a thioredoxin-like protein called
ResA (similar in structure to CcmG) [29] and CcdA, a
functional analogue of DsbD, or DsbD itself
(depend-ing on the organism); together, these apparently form
a pathway analogous to that observed in System I for
reducing a disulfide bond in the apocytochrome
CXXCH motif The heme attachment (and possibly
heme delivery) function of System II is catalyzed by
ResB and ResC Indeed, a fusion protein cloned from
Helicobacter pylori containing elements of ResB and
ResC was sufficient to mature c-type cytochromes
when expressed in the periplasm of a ccm deletion
strain of E coli [30]
Several recent studies have provided insight into the
flexible organization of prokaryotic c-type cytochrome
biogenesis pathways For example, in the Archaea and
some bacteria, a divergent System I has recently been
described [3], and some bacteria contain components
of both System I and System II [3,4,6] Although the
presence of multiple cytochrome c biogenesis systems
in a single bacterium might hint at possible
redun-dancy, additional c-type cytochrome maturation
com-ponents are sometimes required for heme attachment
to specific substrates For example, in the
e-proteobac-terium Wolinella succinogenes, the ccsA1-encoded heme
lyase is required for thioether bond formation to
the remarkable CX15CH heme-binding motif of the
multi-heme c-type cytochrome MccA [31]
System III for cytochrome c maturation consists of
a single primary component, the enzyme heme lyase,
which is found only in the mitochondrial
intermem-brane space (IMS) of animals, fungi and some protists
[11,32] {The kingdom Protista refers to those
eukary-otes that cannot be classified as animals, plants or
fungi: it includes protozoa and algae The protozoa [or
‘first (proto-) animals (zoa)’] are unicellular eukaryotes,
which lack the chitinous cell wall found in fungi} At
least in fungi, heme lyase is supplemented by the
flavo-protein Cyc2, which is thought to provide reducing
equivalents for the heme attachment process [33] The
biochemical study of heme lyase has proved
challeng-ing, and the molecular details of its enzymology are
still largely unclear
Finally, a distinctive example of a biogenesis system
that is required for the dedicated maturation of a
partic-ular substrate is provided by the recent description of
System IV for cytochrome c maturation Heme is
attached through a single thioether linkage to cyto-chromes b6and b from the b6fand bc complexes of oxy-genic phototrophs (cyanobacteria, plants, algae) and certain Bacillus species, respectively [34,35] The mecha-nism by which covalent heme attachment to Bacillus cytochrome b occurs is not yet known, but the identifi-cation of gene products from the green alga Chlamydo-monas reinhardtii that restore cytochrome b6formation
in four ccb mutants constitutes the initial step in the characterization of System IV, which appears to be conserved in all oxygenic phototrophs [36]
In species from the phylum Euglenozoa, which includes Euglena gracilis and the medically relevant trypanosomatids (Trypanosoma brucei, T cruzi and pathogenic Leishmania species), heme is uniquely attached to the mitochondrial c-type cytochromes by a single thioether bond within a F⁄ AXXCH heme-bind-ing motif [37–41] In an earlier study, we determined that, in the trypanosomatids, the occurrence of single-cysteine-containing mitochondrial cytochromes c and
c1 correlates with the absence from both nuclear and mitochondrial genomes of genes encoding any compo-nent of the known c-type cytochrome maturation systems; we also provided experimental evidence that, for the single-cysteine-containing T brucei cyto-chrome c, spontaneous (i.e uncatalyzed) maturation is unlikely [41] These results indicate that at least one further pathway for cytochrome c maturation awaits discovery in the trypanosomatids
In this article, we draw on the resources that are provided through the availability of numerous com-plete genome sequences and several ab initio modeling programs We consider in detail the evolutionary dis-tribution of the machinery for mitochondrial cyto-chrome c assembly throughout the Eucarya, and the possible origins of heme lyase Although the origin of the exclusively eukaryotic heme lyase remains mysteri-ous, replacement of a proto-mitochondrial System I pathway for c-type cytochrome maturation occurred multiple times during protist evolution With rare exceptions, these replacements probably occurred as a result of eukaryote-to-eukaryote lateral gene transfer (LGT) or endosymbiotic gene transfer of heme lyase
We also approach defining the limits of the distribu-tion of the single-cysteine heme-binding motif found in some mitochondrial cytochromes c
Mapping character traits onto a consensus view of eukaryotic phylogeny
The origin of the first eukaryotic cell has been debated for many years; during the 1980s and early 1990s, the
Trang 5available experimental evidence was generally
consis-tent with an evolutionary model (called the Archezoa
theory), which posited two early phases to eukaryotic
evolution: an ancestral phase, in which the hallmark
features of the eukaryotic cytoskeleton, endomembrane
system and nucleus were evolved, followed by the
sec-ond critical phase, which saw the acquisition of the
a-proteobacterial endosymbiont and the evolution of
the proto-mitochondrion Although the results from
some phylogenetic analyses conflicted with the model
formulated by Cavalier-Smith (discussed in [42]), the
Archezoa theory generally received robust support in
phylogenetic trees derived from the analysis of small
subunit rRNA or translation elongation factor
pro-teins Grouped at the base of many of these trees were
several eukaryotic lineages, including diplomonads
(represented by Giardia), the parabasalids (represented
by Trichomonas) and the Microsporidia [43,44] (and
reviewed recently in [45,46]) The distinctive
ultrastruc-ture of these organisms suggested that they apparently
possessed neither mitochondria nor other hallmark
eukaryotic organelles, such as peroxisomes and golgi,
and their status as Archezoa denoted that they were
believed to be ancestrally without these organelles We
now know that this is not the case; more recent
phylo-genetic treatments have resulted in the repositioning of
at least some formerly basal or ‘primitive’ eukaryotes
elsewhere within the eukaryotic tree [46–48]
Further-more, although the secondary loss of peroxisomes has
occurred numerous times in evolution, the
aforemen-tioned organisms crucially retain mitochondria, golgi
and other classically eukaryotic subcellular
compart-ments that have merely been remodeled beyond
obvi-ous or easy recognition [49–53] Thus, there are no
known examples of contemporary eukaryotes that lack
double-membrane-bound organelles of mitochondrial
descent; indeed, although difficult to prove, a popular
current viewpoint is that the acquisition of the
proto-mitochondrial endosymbiont could have been
coinci-dent with eukaryotic origins (see, for example, [47,54]
for a further discussion)
Although the position of the root for eukaryotic
evolution remains a contentious issue – Cavalier-Smith
has argued that the last common ancestor of all extant
eukaryotes diverged with the unikont–bikont split
(Fig 3) [55–57]; other results have suggested that it is
still not possible to discount a previously long-standing
view that the diplomonads and parabasalids belong to
the earliest diverging eukaryotic lineage [46,47,58] –
comparative interrogations of various morphological
and molecular character traits, as well as phylogenies
based on the analysis of multiple gene sets, have
resulted in a seemingly robust resolution of eukaryotic
diversity into six major groupings ([59] and reviewed in [46,47,60,61]) The framework provided by this resolu-tion is increasingly being used to inform on the evolu-tion of various fundamental aspects of eukaryotic biology, both within and between these major group-ings [55–57,62–66] It is this consensus view of eukary-otic evolution on which the comparative analysis described below is based
A phylogeny for mitochondrial c-type cytochrome maturation
Using the complete or draft nuclear and mitochondrial genome sequences indicated in supplementary Doc S1,
we mapped the distribution of mitochondrial cyto-chrome c maturation pathways onto a consensus view
of eukaryotic phylogeny (Fig 3) Our aim was to assess whether there was any obvious order to the otherwise mosaic distribution of mitochondrial cyto-chrome c biogenesis machineries that has previously been hinted at [67,68]
The presence of the Ccm system in higher plants and some unicellular eukaryotes [e.g the deeply diver-gent jakobid Reclinomonas americana, ciliates and the rhodophyte (red alga) Cyanidioschyzon merolae] has been described previously [69–74], whereas other eukaryotes, such as the animals, the chlorophyte green alga C reinhardtii and the malarial parasite Plasmo-dium falciparum (an apicomplexan) have heme lyase for maturation of mitochondrial cytochromes c [2,15,32,75–77] The mitochondrial genome sequences
of various excavate, algal, plant and ciliate taxa very clearly point to the presence of System I within the a-proteobacterial endosymbiont from which mitochon-dria evolved [69,70,72,78,79] However, taking into account the generally robust support for relationships within and between the taxonomic groups shown in Fig 3, our comparative genomic analysis can be used
to provide new insight into the evolution of mitochon-drial cytochrome c maturation Observations that are key to the discussion that follows in subsequent sec-tions are: (a) there is no evidence for the occurrence of heme lyase within the bikont supergroup Excavata; (b)
in the unikonts, heme lyase is the only c-type cyto-chrome maturation system present; (c) there is a mosaic distribution of the Ccm system and heme lyase within the Chromoalveolata and Plantae; (d) wherever the multicomponent Ccm system is used for mitochon-drial cytochrome c maturation, it is always partially encoded on the mitochondrial genome; this is perhaps unsurprising given that CcmC and CcmF are mito-chondrial integral membrane proteins containing mul-tiple predicted transmembrane helices Where a
Trang 6mitochondrial genome sequence is complemented by
the availability of a complete or draft nuclear genome
sequence for the same organism, the following always
holds true: (i) if components of the Ccm system are
encoded in the mitochondrial genome, further
dedi-cated Ccm components are also encoded in the nuclear
genome; (ii) there are no examples of eukaryotes
possessing multiple systems for mitochondrial
cyto-chrome c maturation Thus, even without the availabil-ity of a sequenced nuclear genome, the absence from a protozoan or algal mitochondrial genome of genes encoding Ccm components almost certainly provides a reliable indication that System I will not be used for the maturation of mitochondrial cytochromes c and c1 There are several green, red (rhodophyte) and chromist algae (belonging to the Chromalveolata), plus other
Fig 3 The phylogenetic distribution of the different pathways used for mitochondrial c-type cytochrome maturation in eukaryotes (A) Rela-tionships within and between five of the six eukaryotic supergroups – no relevant data for c-type cytochrome maturation in the sixth super-group, Rhizaria, are currently available The unikonts comprise the Amoebozoa (to which Dictyostelium discoideum and the human pathogen Entamoeba histolytica belong) and the Opisthokonts (the animals, fungi and various protozoa) The unikonts differ from the bikonts (which include the algae, land plants and many different protozoa) in that they possess (probably ancestrally [55]) only a single centriole (the barrel-shaped structure from which flagellar basal bodies are derived and which, in many eukaryotes, is also involved in the organization of the mito-tic spindle) The phylogeny reveals that, within some groups (e.g Viridiplantae), some species contain System I, whereas others contain System III; there were no examples of eukaryotes that contained multiple systems for the maturation of mitochondrial c-type cytochromes A more detailed overview of the distribution of mitochondrial c-type cytochrome maturation pathways in the Plantae is provided in (B) Lineages belonging to the Streptophyta are highlighted by the grey background The evolutionary relationships shown represent a consensus view of published data A complete list of species used to produce the phylogeny, including the databases searched, is provided in supplementary Doc S1 Species for which the identification of the mitochondrial c-type cytochrome biogenesis apparatus is based on the interrogation of a complete genome sequence are as follows: the choanoflagellate Monsiga brevicolis (System III); the amoebozoan Dictyostelium discoideum (System III); the chlorophyte green algae Chlamydomonas reinhardtii, Volvox carteri, Ostreococcus lucimarinus and Ostreococcus tauri (all System III); the red alga Cyanidioschyzon merolae; the ciliates Tetrahymena thermophila and Paramecium tetraurelia (System I); the Apicom-plexans Plasmodium falciparum, Toxoplasma gondii and Theileria parva (System III); the dinoflagellate Perkinsus marinus (System III); the oomycetes Phytophthora ramorum and Phytophthora sojae and the diatoms Thalassiosira pseudonana and Phaeodactylum tricomutum (collectively belonging to a group known as the stramenopiles or chromists) (all System III); the Heterolobosean Naegleria gruberi (System I) Entamoeba histolytica (Amoebozoa), Encephalitozoon cuniculi (Microsporidia), Cryptosporidium parvum (Apicomplexa) and the diplomonad Giardia intestinalis all contain degenerate mitochondria known as mitosomes, and the parabasalid Trichomonas vaginalis possesses hydro-genosomes; such degenerate forms of mitochondria lack a respiratory chain and therefore do not contain c-type cytochromes.
Trang 7protozoan species (the amoebozoan Acanthamoeba
cas-tellanii), for which no nuclear genome sequence is
available, but there is an accessible or annotated [79]
mitochondrial genome sequence on which no
compo-nent of the Ccm system is encoded Similarly, there is
extensive sequence coverage for the uniquely organized
(many small linear chromosomes of less than 8.3 kb in
length) mitochondrial genome of the ichthyosporean
Amoebidium parasiticum (belonging to the
Opi-sthokonta); from the sequence released thus far, this
genome also lacks genes encoding Ccm components
[80] Assuming that none of these species contain
mitochondrial c-type cytochromes with atypical
heme-binding motifs, we suggest that it is likely that
nuclear-encoded heme lyase is used for the maturation of their
mitochondrial cytochromes c and c1
We found a eukaryotic cytochrome c biogenesis
Sys-tem II only in those eukaryotes that contain
chlorop-lasts (data not shown), and we assume that, in these
cases, System II is used for the maturation of the
chloroplast c-type cytochromes, given the ancestral
relationship between chloroplasts and System
II-con-taining cyanobacteria [81,82] Where System II was
observed, a second c-type cytochrome biogenesis
appa-ratus was always present and is presumed to be
responsible for maturing the mitochondrial
cyto-chromes c and c1(e.g System I in Arabidopsis thaliana,
System III in C reinhardtii and chromist algae)
Simi-larly, genes encoding the four chloroplast proteins
recently shown to be required for single-cysteine
attachment to cytochrome b6in Chlamydomonas –
Sys-tem IV for c-type cytochrome biogenesis – were also
only present in phototrophic eukaryotes [36] The
absence of heme lyase from the excavates, the possible
origins of heme lyase and the molecular basis for the
mosaic distribution of Systems I and III in
chromalve-olates and the Plantae are the critical issues upon
which we focus in the remainder of this article
Was heme lyase ever present in
the Excavata?
A number of important human pathogens, such as
try-panosomes, Giardia and Trichomonas, as well as a
diverse assortment of free-living protozoa, are included
in the supergroup Excavata The validity of this
classi-fication was initially based on a number of shared
morphological features, but has more recently received
modest support from a variety of molecular
phyloge-nies [59,83–85] Support for the monophyly of the
Excavata is, however, equivocal [86]; indeed, the
possi-bility that the earliest diverging eukaryote was an
ancestor of diplomonads (Giardia) and parabasalids
(Trichomonas) has not yet been entirely dismissed [46,58] Interestingly, if we accept the emerging evi-dence that groups the Excavata together, a deep-branching status for the supergroup can be inferred from a variety of character traits A prime example is the distinctive mitochondrial genome of the jakobid
R americanawhich, in terms of both gene content and genome organization, more closely resembles an a-pro-teobacterial genome than any other mitochondrial gen-ome that has presently been sequenced [70,87] Like Reclinomonas, some of the other excavates currently sampled (Naegleria and Malwimomonas) contain the Ccm system for cytochrome c maturation (Fig 3) Others (Trichomonas vaginalis and Giardia intestinalis) lack a capacity for respiration, and c-type cytochromes are accordingly absent from their degenerate mito-chondria, making it impossible to assess which system for cytochrome c maturation would have been present
in their last aerobic ancestors In trypanosomatids, cytochromes c and c1 are present, but there is no rec-ognizable c-type cytochrome maturation system Thus, there is no evidence that heme lyase was ever present within the excavate supergroup
The recently described absence [41] of any known cytochrome c biogenesis system from the various try-panosomatids represents a particularly intriguing sce-nario, as it correlates with the attachment of heme to single-cysteine XXXCH mitochondrial cytochromes in these organisms Such single cysteine cytochromes are also present in other kinetoplastids (the trypanosoma-tid family evolved from a kinetoplastrypanosoma-tid ancestor) and the euglenids Euglena gracilis and E viridis [37–41] All of these protists belong to the phylum Euglenozoa (Fig 4A), but, in addition to the euglenids and kine-toplastids, the Euglenozoa includes a third major taxo-nomic group, a family of mostly free-living marine flagellates known as the diplonemids Recent phyloge-nies suggest that the diplonemids are likely to be a sis-ter group to the Kinetoplastida [88] Although there is
no genome project for a diplonemid, we have used the relatively simple experiment of determining the type of mitochondrial cytochromes present (either CXXCH
or XXXCH heme attachment) to look further at the evolution of cytochrome c biogenesis in the Excavata From a combination of spectroscopic methods and N-terminal sequencing (Fig 4), Diplonema papillatum unambiguously contains a single-cysteine c-type cyto-chrome (AGQCH heme-binding motif) Thus, all three major taxonomic groups of the Euglenozoa (diplone-mids, kinetoplastids and euglenids) contain single-cysteine mitochondrial cytochromes c, and hence it is likely that they all contain the same, as yet unidenti-fied, apparatus for maturation of cytochromes c, which
Trang 8is distinct from that found in any other organisms.
Analysis of all the available genome sequences and all
publicly accessible expressed sequence tag (EST)
collec-tions (including ESTs for the excavates
Malawimon-as californiana, M jakobiformis and R americana)
using blast reveals that, strikingly,
single-cyste-ine attachment of heme to mitochondrial
cyto-chrome c remains a characteristic that is unique to species from the phylum Euglenozoa Crucially, these analyses included the use of the draft nuclear genome sequence for Naegleria gruberi, an amoeboflagellate with an aerobic metabolism from the phylum Hetero-lobosea, the eukaryotes with the closest evolution-ary relationship to the Euglenozoa [47,59,85] The
Fig 4 Diplonema cytochrome c has only a single cysteine in its heme-binding motif (A) Probable evolutionary relationships within the phy-lum Euglenozoa, as suggested by taxon-rich small subunit rRNA phylogeny (B) Absorption spectrum of semi-purified D papillatum cyto-chrome c, recorded at 25 C with the protein in 50 m M Tris ⁄ HCl (pH 8.0) containing a few grains of disodium dithionite to reduce the heme iron The protein was purified from a culture of D papillatum strain ATCC50162 by SP-Sepharose chromatography Absorption maxima were
at 419.5, 523.5 and 554.0 nm Inset: reduced pyridine hemochrome spectrum of the same protein Pyridine hemochrome analysis was con-ducted according to Bartsch [133]: final concentrations of hydroxide and pyridine were 0.2 M and 30% (v ⁄ v), respectively, and a few grains
of dithionite were added The a-band peak maximum at 553.0 nm (indicated by the vertical broken line) diagnostically indicates heme attach-ment to the polypeptide via one cysteine residue [37,39,41,133–135] Diplonema was cultured in artificial seawater as described previously [136], and subjected to detergent extraction [41] prior to isolation of cytochrome c (C) Sequence alignment of the N-terminal 40 amino acids
of Diplonema cytochrome c, as determined by Edman degradation, and the N-terminal regions of cytochromes c from other organisms: Cf, Crithidia fasciculata; Dp, Diplonema papillatum; Eg, Euglena gracilis; iso, isoform; Sc, Saccharomyces cerevisiae; Tb, Trypanosoma brucei The c-type cytochrome heme-binding motif is highlighted in bold for each cytochrome Underlined residues denote differences between the major and minor isoforms of mitochondrial cytochrome c in D papillatum: Dpiso1 is the major form (75% of the total protein) and Dpiso2 is the minor form (25%) Cytochrome c as analyzed in (B) was further purified using a CM-Sepharose column before N-terminal sequencing Cysteine gives a blank (X) in the sequencing reaction unless appropriately alkylated [137]; thus X is what is expected and observed for cyste-ine covalently bound to a heme in a c-type cytochrome It is, however, clear that the first residue of the heme-binding motif of D papillatum cytochrome c is alanine not cysteine, and thus the cytochrome has a single cysteine heme-binding motif of the type found in other Eugleno-zoaons, rather than CXXCH as observed in typical mitochondrial cytochromes c.
Trang 9N gruberimitochondrial cytochromes c and c1contain
CAQCH and CSACH motifs, respectively; these
cytochromes are matured by cytochrome c biogenesis
System I
We postulated previously [41] that the acquisition of a
novel mitochondrial cytochrome c biogenesis system in
the Euglenozoa provided not only a driving force for the
loss of a pre-existing maturation system, but also the
evolutionary pressure to move from CXXCH to
XXXCH cytochromes c If increased taxon sampling
fails to detect the existence of an excavate heme lyase,
this is likely to influence which of the models discussed
below most parsimoniously explains the distribution of
cytochrome c maturation systems shown in Fig 3
Probing the origins of heme lyase
Heme lyase has, since its discovery [15], remained a
rather enigmatic enzyme: the origin of this
eukaryotic-specific protein is obscure and little is known about
the biochemistry of System III-dependent
cyto-chrome c maturation [13] From the analysis shown in
Fig 3, it is clear that, although animals and
Dictyoste-liumeach encode a single form of heme lyase, two
iso-forms of heme lyase are found in other eukaryotes At
least in Saccharomyces cerevisiae, the presence of two
lyases reflects the distinct substrate preferences of each
enzyme: either cytochrome c or c1, respectively
[15,32,75] In order to obtain an insight into the origin
of heme lyase and to explore a molecular explanation
for its evolutionary distribution, we performed a
phy-logenetic analysis, and also applied a number of
bioin-formatics tools that can be used to detect remote
structural similarities between different proteins that
are undetectable even by sensitive iterative database
searches
Assuming that the presence of multiple heme lyases
always reflects, as it does in yeast, the deployment of
one enzyme to catalyze the maturation of each
mito-chondrial c-type cytochrome, one aim with the
phylog-eny was to determine whether the transition from a
single heme lyase with broad substrate specificity to
dual enzymes, each with their own specificity for either
cytochrome c or cytochrome c1 [15,32], was likely to
have occurred just once or on a number of occasions
With the exception of their N-termini, which were
lar-gely unique to each taxonomic group, heme lyase
pro-tein sequences were reliably aligned Following the
omission of sequences corresponding to putative heme
lyases from the choanoflagellate Monsiga brevicolis and
the dinoflagellate Perkinsus marinus, a bootstrapped
maximum likelihood (ML) phylogeny robustly resolved
distinct heme lyase clades for the metazoan, fungal,
algal and apicomplexan sequences These clades were supported by bootstrap values greater than 75 (Fig 5) However, the relationships between these clades were not robust, and therefore could not be resolved satis-factorily Clearly, the arrangement of the basal nodes towards the root of the phylogeny is crucial to an understanding of the evolution of the c–c1 heme lyase distinction, and the number of origins in particular However, all heme lyases from Apicomplexa clustered together with reasonable robustness (bootstrap value, 87), largely due to the distinct N-termini shared by these proteins
The monophyly of all apicomplexan heme lyases points towards at least two origins of the c–c1distinction amongst eukaryotes: one prior to the divergence of the fungi and one affecting the alveolates [the group that includes the ciliates, apicomplexans and dinoflagellates (Fig 3)] Further origins of the c–c1distinction affecting diatoms (Thalassiosira pseudonana and Phaeodacty-lum tricomutum) and chlorophyte algae are possible, but increased taxon sampling is necessary to allow the reso-lution of these possibilities In the example of Dictyoste-lium, we cannot know whether the presence of a single heme lyase represents an ancestral state or the reverse transition of going from two distinct lyases to a single lyase of broader substrate specificity However, with regard to the opisthokonts, the presence of a single heme lyase in animals, but multiple lyases in the choanoflagel-late Monsiga brevicolis (Fig 3) and the fungi, points either to multiple origins for the c–c1 dichotomy or a loss of a heme lyase isoform from animals with, presumably, relaxation of the substrate specificity
To determine whether the monophyly of the apicom-plexan sequences was an artifact introduced by the biased base composition common to apicomplexan genomes, a neighbor-joining phylogeny was estimated with logdet genetic distances [89], which correct for base composition imbalance Monophyly of apicom-plexan sequences was still recovered after correction for base composition The result of the Kishino–Ha-segawa (KH) test also corroborated the view that there have been multiple origins for the c–c1 distinction Here, to test whether the optimal topology obtained from the ML and Bayesian inference (BI) trees was significantly more likely than a ‘single-origin’ scenario, the likelihood score of an alternative tree, in which all c- and c1-type sequences were reciprocally monophy-letic (i.e one simulating a single origin for the c–c1 dis-tinction), was compared with the optimal ML estimate using a KH test [90] and phylip v3.65 [91] A signifi-cant reduction in likelihood score when this constraint was enforced demonstrated that a single origin of the c–c1 distinction could be rejected – the alternative ML
Trang 10tree topology had a likelihood score of )24248.9,
which was significantly worse than the unconstrained,
optimal tree topology (Dln L = )54.2, P < 0.001)
To seek insight into the possible origin of System III
for cytochrome c maturation, we used a variety of
bio-informatics tools (as described in supplementary
Doc S1) to search for protein families distantly related
to heme lyase The application of these approaches
served only to highlight further the enigmas that
sur-round this fundamentally important enzyme; however,
as cytochromes c and c1 are matured within the
mito-chondrial intermembrane space (IMS), two possible
candidate proteins identified are nonetheless worthy of
mention Thus, after the obvious match to the heme
lyase domain itself, the first HHPRED result initially
appeared interesting A small portion, 34 residues, of the heme lyase was matched to a region of a Pfam entry for the Erv1⁄ Alr family of IMS proteins involved
in protein import into the IMS and export of mito-chondrial Fe⁄ S clusters into the cytoplasm [92–95] However, the heme lyase secondary structure predic-tion was not in good agreement with the four helical bundle architecture of the Erv1⁄ Alr sulfhydryl oxidase, and no other fold recognition method (below) flagged
up this putative relationship
The best 3D-Jury consensus fold recognition scores were obtained for the conserved domain of the human heme lyase but, at up to 45, did not reach the bench-mark significance cut-off of 50 [96] Once again the matched protein, superoxide dismutase (SOD), was
Fig 5 An unrooted, maximum likelihood (ML) phylogeny of heme lyase protein sequences A WAG substitution matrix was applied with among-site rate heterogeneity described by a gamma distribution estimated from the data Branch lengths are measured in substitutions per site Non-parametric bootstrap values from the ML analysis over 50, and their corresponding posterior probabilities from the Bayesian analy-sis, are shown adjacent to the nodes An asterisk denotes bootstrap values > 95 and posterior probabilities of 1.00 Full details of the meth-ods used for phylogeny construction and in the predictive modeling of heme lyase are provided in supplementary Doc S1 Clades are color coded by taxon: Fungi (red; c-type heme lyases are shaded lighter); Metazoa (yellow); Apicomplexa (blue; lighter and darker shading highlight distinct subclades); algal ⁄ stramenophile (green; lighter and darker shading highlight distinct subclades).