Comparison of the thiol-dependent antioxidant systems in the ectomycorrhizal Laccaria bicolor and the saprotrophic Phanerochaete chrysosporium


  • Mélanie Morel,

    1. Unité Mixte de Recherches 1136 Interactions Arbres/Microorganismes INRA/Nancy Université, IFR 110 Génomique Ecologie et Ecophysiologie Fonctionnelles. Faculté des Sciences BP 239, 54506 Vandoeuvre-lès-Nancy Cedex, France
    Search for more papers by this author
  • Annegret Kohler,

    1. Unité Mixte de Recherches 1136 Interactions Arbres/Microorganismes INRA/Nancy Université, IFR 110 Génomique Ecologie et Ecophysiologie Fonctionnelles. Faculté des Sciences BP 239, 54506 Vandoeuvre-lès-Nancy Cedex, France
    Search for more papers by this author
  • Francis Martin,

    1. Unité Mixte de Recherches 1136 Interactions Arbres/Microorganismes INRA/Nancy Université, IFR 110 Génomique Ecologie et Ecophysiologie Fonctionnelles. Faculté des Sciences BP 239, 54506 Vandoeuvre-lès-Nancy Cedex, France
    Search for more papers by this author
  • Eric Gelhaye,

    1. Unité Mixte de Recherches 1136 Interactions Arbres/Microorganismes INRA/Nancy Université, IFR 110 Génomique Ecologie et Ecophysiologie Fonctionnelles. Faculté des Sciences BP 239, 54506 Vandoeuvre-lès-Nancy Cedex, France
    Search for more papers by this author
  • Nicolas Rouhier

    1. Unité Mixte de Recherches 1136 Interactions Arbres/Microorganismes INRA/Nancy Université, IFR 110 Génomique Ecologie et Ecophysiologie Fonctionnelles. Faculté des Sciences BP 239, 54506 Vandoeuvre-lès-Nancy Cedex, France
    Search for more papers by this author

Author for correspondence:
Nicolas Rouhier
Tel: +33 3 83 68 42 25
Fax: +33 3 83 68 42 92


  • • Sequencing of the Laccaria bicolor and Phanerochaete chrysosporium genomes, together with the availability of many fungal genomes, allow careful comparison to be made of these two basidiomycetes, which possess a different way of life (either symbiotic or saprophytic), with other fungi. Central to the antioxidant systems are superoxide dismutases, catalases and thiol-dependent peroxidases (Tpx). The two reducing systems (thioredoxin (Trx) and glutathione/glutaredoxin (Grx)) are of particular importance against oxidative insults, both for detoxification, through the regeneration of thiol-peroxidases, and for developmental, physiological and signalling processes. Among those thiol-dependent antioxidant systems, special emphasis is given to the redoxin and methionine sulfoxide reductase (Msr) multigenic families.
  • • The genes coding for these enzymes were identified in the L. bicolor and P. chrysosporium genomes, were correctly annotated, and the gene content, organization and distribution were compared with other fungi. Expression of the Laccaria genes was also compiled from microarray data.
  • • A complete classification, based essentially on gene structure, on phylogenetic and sequence analysis, and on existing experimental data, was proposed.
  • • Comparison of the gene content of fungi from all phyla did not show huge differences for multigenic families in the reactive oxygen species (ROS) detoxification network, although some protein subgroups were absent in some fungi.


The availability of genomes from saprotrophic, mutualistic and pathogenic fungi now provides the opportunity to understand the processes used by soil fungi either to colonize decaying wood and soil litter or to interact with living plants within their ecosystem. For example, the genome of the ectomycorrhizal fungus Laccaria bicolor reveals only a single gene encoding an endoglucanase with a cellulose-binding domain and no gene for exocellobiohydrolase, indicating a poor ability to degrade plant cell-wall polysaccharides. There is also little evidence of the oxidative systems necessary for lignin degradation (Martin et al., 2008). Compelling evidence has arisen over the last decade demonstrating that reactive oxygen species (ROS) and reactive nitrogen species (RNS) play a central role in pathogen defense in plants (Jennings et al., 1998; Govrin & Levine, 2000; Egan et al., 2007; Floryszak-Wieczorek et al., 2007; Molina & Kahmann, 2007; Takemoto et al., 2007) and also in mutualistic symbioses (Gafur et al., 2004; Tanaka et al., 2006). A successful pathogen and symbiont must be able to overcome or suppress a complex array of ROS-mediated host plant defenses. In fact, microbial suppression of ROS-mediated defenses as a result of the secretion of ROS-scavenging enzymes, such as superoxide dismutase (SOD) and catalase, which convert ROS into less reactive species, has been documented in plant pathogens, endophytes and symbionts (Jennings et al., 1998; Gafur et al., 2004; Tanaka et al., 2006; Molina & Kahmann, 2007). Besides their toxicity towards lipids, proteins and DNA, and finally towards whole organisms, ROS have also, depending on their concentration, important physiological roles whereby they act as signalling molecules (D’Autréaux & Toledano, 2007). In fungi, ROS are involved in cell differentiation, in mycorrhiza establishment and development, and in virulence and pathogenicity (Aguirre et al., 2005; Baptista et al., 2007; Egan et al., 2007; Molina & Kahmann, 2007). At the protein level, ROS and RNS can directly regulate protein function or enzyme activity through post-translational modifications (e.g. formation of disulfide bonds, glutathionylation or nitrosylation) or through amino acid oxidations, the two most sensitive amino acids being cysteine and methionine. In fungi, these ROS are mainly generated in mitochondria upon over-reduction of the electron transport chain and through various oxidases, especially in the peroxisome (Aguirre et al., 2005; Takemoto et al., 2007; Herrero et al., in press). Various systems, involving proteins that often belong to multigenic families, have been developed to regulate tightly the concentration of some ROS (especially superoxide ions, hydrogen peroxide and peroxynitrite). Major ROS scavenging enzymes include SOD, catalases, catalase/peroxidases and thiol-dependent peroxidases (Tpx) which regroup glutathione peroxidases (Gpx) and peroxiredoxins (Prx) also called thioredoxin peroxidases (Fig. 1). In the absence of selenoproteins in fungi, all the Tpxs are ubiquitous nonheme peroxidases, which use reactive thiol groups and the sulfenic acid chemistry for their reactivity and use thioredoxin (Trx) or glutaredoxin (Grx) systems for their regeneration (Chae et al., 1994). Contrary to catalases, which only reduce hydrogen peroxide, they are generally able to reduce a broader range of substrates, including peroxynitrites and phospholipid hydroperoxides (Fig. 1). On the other hand, some Tpxs are rather specifically involved in ROS sensing and signalling mechanisms. In Saccharomyces cerevisiae, glutathione peroxidase (ScGpx3) oxidized by hydrogen peroxide can in turn oxidize a transcription factor, yeast activator protein 1 (so-called Yap1), to activate the antioxidant defense genes (Delaunay et al., 2002). In Schizosaccharomyces pombe, a 2-Cys Prx, called Tpx1, can transfer a redox signal to Pap1 (the AP1 transcription factor of S. pombe) to activate it by oxidation at low H2O2 concentration (Vivancos et al., 2005). At higher H2O2 concentration, the overoxidation of this 2-Cys Prx inhibits the Pap1-dependent pathway, but it participates in the regulation of another pathway that is dependent on the mitogen-activated protein kinase (MAPK) Sty1 (Veal et al., 2004; Vivancos et al., 2005). The sulfinic acid formed on these 2-Cys Prx is regenerated by ATP-dependent enzymes called sulfiredoxins (Srxs) (Biteau et al., 2003). Another class of antioxidant proteins that uses a thiol-dependent regeneration mechanism and participates in ROS scavenging is called methionine sulfoxide reductase (Msr). Indeed, as the oxidation of methionine into methionine sulfoxide (MetSO) is a reversible process (the reduction is catalysed by Msr), and as it would not systematically result in protein inactivation, the cyclic oxidation/reduction steps could serve as an important antioxidant mechanism (Levine et al., 1996).

Figure 1.

The reactive oxygen species (ROS) detoxifying and signalling network in fungi is linked to protein thiol oxidation. Superoxide dismutase (SOD) is the only enzyme able to prevent toxic association of the superoxide anion with nitric oxide. Once hydrogen peroxide (H2O2) is formed, it can be degraded by catalases, thiol-dependent peroxidases (Tpx) or in some cases by glutaredoxins (Grx) and glutathione S-transferases (GST). Thiol-dependent peroxidases are certainly crucial enzymes because they are able to reduce a broad range of substrates (peroxynitrite (ONOO-) hydrogen peroxide but also peroxidized lipids), and some Tpx are involved in H2O2 sensing and signalling, especially by regulating the AP1 transcription factor. If ROS are not degraded, they can damage all macromolecules, but thiol groups of cysteines in particular. The formation of disulfide bonds, protein glutathione adducts or sulfenic acids are reversible reactions catalyzed by either Grx or Trx, or both. The formation of sulfinic (RSO2H) or sulfonic (RSO3H) acids is considered to be an irreversible reaction, except in the case of some eukaryotic 2-Cys peroxiredoxins (Prx), whose oxidation of the peroxidatic cysteine into sulfinic acid is reversed by sulfiredoxins (Srx). Cat, catalase; Grx, glutaredoxin; GSH, reduced glutathione; SOD, superoxide dismutase; Srx, sulfiredoxin; Tpx, thiol peroxidase; Trx, thioredoxin.

Both Tpx and Msr require a thiol-assisted recycling system, dependent either on glutathione, Grx or Trx. In addition, Grx and Trx have been shown to participate in a large number of processes, from defence against oxidative stress to cellular processes, such as development, cell differentiation, host–pathogen interactions, virulence, reproduction and assimilation pathways, in particular that of sulfur (Muller, 1991; Collinson et al., 2002; Garrido & Grant, 2002; Trotter & Grant, 2002; Missall & Lodge, 2005a,b; Malagnac et al., 2007; Thön et al., 2007).

Several recent reviews have described the functions of antioxidant systems in fungi, mainly in Ascomycetes (Carmel-Harel & Storz, 2000; Grant, 2001; Herrero & Ros, 2002; Herrero & de la Torre-Ruiz, 2007; Toledano et al., 2007; Herrero et al., in press). However, except for a recent article on Grxs and a partial analysis of Paracoccidioides brasiliensis antioxidant systems through an expressed sequence tag (EST) analysis, no comprehensive survey of sequenced fungal genomes has been conducted on the redoxin (Trx, Grx, Srx and Prx) and Msr gene families (Campos et al., 2005; Herrero et al., 2006). So far, thiol-dependent antioxidant systems of an ectomycorrhizal fungus has never been characterized at a genome-scale and it is not known if symbiotic fungi have gained or lost specific pathways, as a result of biotrophy, in comparison to saprotrophic fungi. In the present study, we characterized the complete set of genes coding for thiol antioxidant systems in the recently sequenced genomes of the saprotrophic white rot Phanerochaete chrysosporium (Martinez et al., 2004) and of the ectomycorrhizal symbiont L. bicolor (Martin et al., 2008). The goals of this comparative genomic analysis were as follows: to organize growing fungal genomic data and thus obtain a clear classification of the predicted redoxins and Msr proteins; to find specific genes in some fungi and understand whether the way of life and biological traits of these fungi correlate with variations of the antioxidant gene content; to survey the level of transcripts in various tissues and environmental conditions; to conduct phylogenetic analyses of key proteins within the Mycota; and to provide a basis for the functional analysis of these families.

Materials and Methods

Bioinformatic genome analysis

The DNA sequences and the deduced peptide sequences for the reducing and thiol-dependent antioxidant systems were retrieved from the L. bicolor S238N-H82 and P. chrysosporium whole-genome databases at the US Department of Energy Joint Genome Institute (JGI): and, respectively. To confirm the gene model intron/exon structures, and to determine if the predicted sequences were expressed, 38 913 high-quality L. bicolor ESTs, clustered into 13 782 tentative consensi (, and 13 811 P. chrysosporium ESTs available at the National Center for Biotechnology Information (NCBI, were used for blast searches and aligned to the gene models. The curated peptide sequences were used to search against the NCBI, Broad-MIT ( and Duke University ( databases using blastp or tblastn. To study, in further detail, the relationship of antioxidant proteins in filamentous fungi, paralogs from L. bicolor and P. chrysosporium and other fungi were used for phylogenetic analysis using clustalw and the neighbour-joining tree algorithm in mega, version 4 (Tamura et al., 2007). The cellular location of putative proteins was predicted using predotar (, psort (, targetp (, mitoprot ( and wolfpsort (

Whole-genome microarray analysis

For transcript profiling, the RNA extraction from free-living mycelium, ectomycorrhizas and fruiting body and calculation of the expression levels was as described in Section 9 of the Supplemental Online Information in Martin et al. (2008). The complete expression data set is available as series # GSE9784 at the Gene Expression Omnibus at NCBI (

Results and Discussion

The release of the L. bicolor and P. chrysosporium genomes (Martinez et al., 2004; Martin et al., 2008) allowed a genome-wide inventory to be made of the genes coding for the reducing and thiol-dependent antioxidant systems in these basidiomycetes. We compiled all the predicted sequences of interest in the current genome assemblies, curated the gene models generated by automated annotations and edited them when needed. Additional searches using fungal sequences available in databases were performed on both the L. bicolor and P. chrysosporium whole-genome annotations and EST data sets to identify missing genes. In order to obtain a comprehensive census and an improved classification of those multigenic families in fungi, the L. bicolor and P. chrysosporium gene contents were compared with nine other fungal genomes and with a few bacterial, archaeal and eukaryotic model organisms (Table 1). In the course of this in silico analysis, we correlated specific genomic features in the antioxidant system to the lifestyle of these fungi (i.e. symbiotic, saprophytic or pathogenic) or to their phylogeny (i.e. Ascomycota, Basidiomycota or Zygomycota).

Table 1.  Comparative analysis of the gene content of the reducing and thiol-dependent antioxidant systems in various sequenced basidiomycetes (Bas), ascomycetes (Asc) and zygomycetes (Zyg) and in model organisms from other kingdoms (Mam, mammals; Bact, bacteria; Arch, archaea)
 OrganismsReducing systemsThiol-dependent antioxidant systems
TRTrxGRGrxGrx CxxCGrx CGFSGrx CxxSTpx (Prx/Gpx)2-Cys Prx1-Cys PrxPrxQPrx IIGpxSrxMsrA/B
  1. Gpx, glutathione peroxidase; GR, glutathione reductase; Grx, glutaredoxin; Msr, methionine sulfoxide reductase of type A or B; Prx, peroxiredoxin; Srx, sulfiredoxin; Tpx, thiol peroxidase; TR, thioredoxin reductase; Trx, thioredoxin. For thioredoxins, only small proteins with WCGPC/WCPPC/WCQPC active sites were considered. The total number of genes for each family is shown in bold. For the Grx and Tpx families, subgroups have been distinguished according to the existing classification.

MamHomo sapiens2214 22 −6/6411611/3
PlantsPopulus trichocarpa319336205119/62124615/5
Arabidopsis thaliana219231144139/82115815/9
BactEscherichia coli1214 31 −3/111111/1
Thermotoga maritima1 −  − −3/−12−/−
ArchAeropyrum pernix1 −  − −3/−13−/−
BasLaccaria bicolor1415 22 16/1112211/1
Phanerochaete chrysosporium1434 12 17/1122211/1
Coprinopsis cinereus1214 12 16/1112211/1
Ustilago maydis1214 12 15/112211/1
Cryptococcus neoformans1215 13 14/21111211/1
AscSaccharomyces cerevisiae2317 23 25/32111311/1
Candida albicans1116 32 15/41112411/1
Magnaporthe grisea1114 12 14/111211/1
Neurospora crassa1214 12 14/111211/1
Aspergillus nidulans1214 12 15/121211/1
ZygRhizopus oryzae1135 13 16/3312322/1

Genome screening of the reducing systems

Most fungi do not possess ascorbate or analogous molecules, and have no homolog of ascorbate peroxidases or enzymes involved in ascorbate regeneration (dehydroascorbate reductase and monodehydroascorbate reductase) (this study). The antioxidant systems used for ROS and RNS detoxification are therefore essentially dependent on the glutathione/Grx or Trx systems. For the Trx family, only proteins of < 200 amino acids, showing classical WCGPC, WCPPC or even WCQPC active sites, and a high homology outside the active sites, were considered in this study. Larger proteins possessing a Trx domain with a WCGPC active site, but associated with other modules or Trx-like proteins with divergent CxxC active sites, have not been studied because these proteins are generally not characterized and nothing is known about their function and reactivity. For the Grx family, only those with recognized active sites (CPYC, CPFC, CGFS, CSYC and CPYS) and/or sufficient homology for the glutathione-binding site were analysed. This explains why there are no true Trx or Grx homologs in Thermotoga maritima or Aeropyrum pernix, although some related proteins are possible candidates for catalyzing dithiol–disulfide exchanges (Table 1). Except for plants with large expanding multigene families, the total gene content varied only slightly among fungi, Escherichia coli and humans. The major difference was a higher number of Grx genes in some fungi, such as S. cerevisiae or Candida albicans (seven and six genes), compared with other organisms (usually four to five genes), and a higher number of Trx genes in L. bicolor and P. chrysosporium (four genes) compared with all other organisms (one to three genes). The analysis of the Trx and Grx families within fungal genomes is detailed in the following sections.

Genome analysis of the thiol-dependent antioxidant systems

In the absence of ascorbate peroxidases, the major peroxide-degrading enzymes are catalases or catalase peroxidases and Tpxs. The Tpx family comprises the Prxs, or Trx peroxidases, and the so-called Gpxs (Rouhier & Jacquot, 2005). Nevertheless, only mammalian Gpxs use glutathione for their regeneration, whereas all plant, bacterial and fungi Gpxs probably use Trx for their reduction (Tanaka et al., 2005; Navrot et al., 2006). The Tpx family distribution between the organisms analysed is heterogeneous. Again, plants developed large multigene families (15 to 17 members), whereas the thermophilic bacteria, T. maritima, or the archaea, A. pernix, displayed a restricted number of genes (three members), with no Gpx homolog (Table 1). Mammals exhibited an intermediate situation with 12 members, but six Gpxs, which are essentially glutathione-dependent enzymes, in contrast to most other Gpxs. In addition, some are selenoproteins. Even between the different fungi, the number of Tpxs varied from five to nine. At this stage, we do not know whether these differences in gene distribution provided an evolutionary advantage to the organisms concerned. The analysis of the Tpx family within fungal genomes will be further discussed under ‘Thiol-peroxidases and sulfiredoxin in fungi’.

Sulfiredoxins are proteins, found specifically in eukaryotes, which regulate the oxidation state of some members of the Tpx family (2-Cys Prx subgroup) by reducing sulfinic acids formed on the peroxidatic cysteines. Thus, indirectly they act in concert with thiol-peroxidases for ROS sensing and signalling. Only one gene is normally present in eukaryotic genomes, and among the fungal genomes analysed here, a Srx was only identified in the two Hemiascomycetes (S. cerevisiae and C. albicans) and in one basidiomycete (Cryptococcus neoformans), whereas two genes were identified in the zygomycete Rhizopus oryzae. Strikingly, these two genes showed a higher homology to mammalian Srx than to fungal Srx. This is the only organism analysed, to date, that has been found to possess two Srxs. In fact, an Srx was also found by blast searches in all hemiascomycetes and in the archeascomycete, S. pombe (data not shown). From an evolutionary point of view, the following hypotheses can be put forward to explain this: the gene has been acquired in all hemiascomycetes after the split with euascomycetes, and independently in an archeascomycete, a zygomycete and a basidiomycete; or the gene was originally present in the fungal ancestor but lost in most clades except in hemiascomycetes. Whether a different mechanism has evolved in other fungal genera, or whether they do not possess this ROS signalling pathway, is still unknown.

Through the reduction of MetSO back to methionine, Msr is involved in antioxidant systems and uses reducing equivalents, generally Trx, for its regeneration. Upon methionine oxidation, two types of MetSO are formed, S-enantiomers and R-enantiomers. This coincides with the existence of two classes of Msr, called MsrA and MsrB, reducing respectively the S-epimer and the R-epimer. Except for T. maritima and A. pernix, which exhibit no clearly identified MsrA and MsrB ortholog, and plants presenting expanded families, most organisms analysed possess a single gene of each class. Only R. oryzae possesses two MsrA genes, whereas Homo sapiens exhibit three MsrB genes. In mammals, both MsrA and MsrB were found in mitochondria, but the cellular location is not known in other organisms. As oxidation is ubiquitous in the cells, we speculate whether mechanisms such as alternative splicing or translational regulation allow a larger distribution of Msr within eukaryotic cells.

Glutathione/glutaredoxin reducing system in fungi

Analysis of the draft genomes of L. bicolor and P. chrysosporium has revealed the existence of both monothiol and dithiol Grx. L. bicolor possesses five Grx, one containing a CPYC active site (LbGrx2.1), one with a CPHS motif (LbGrx2.2), one with a CPYS motif (LbGrx6) and two with a CGFS motif (LbGrx4 and LbGrx5). The predicted genes and proteins have been named according to their orthologs in S. cerevisiae. These five proteins clustered into four different clades on a phylogenetic neighbor-joining tree; the group containing Grx6 orthologs was split into two subgroups containing ascomycetous or basidiomycetous and zygomycetous sequences (Fig. 2).

Figure 2.

Phylogenetic tree of the glutaredoxin family. Gene models for Laccaria bicolor and Phanerochaete chrysosporium glutaredoxins (Grx) are listed in Tables 2 and 3. Other accession numbers are as follows: CcGrx1, EAU92516; CcGrx4, EAU85130; CcGrx5, EAU90056; CcGrx6, EAU91098; CnGrx1, XP_775768; CnGrx4.1, XP_569769; CnGrx4.2, XP_569770; CnGrx5, XP_568962; CnGrx6: XP_569285, UsGrx1: XP_761095, UsGrx4: XP_760370, UsGrx5: XP_759014, UsGrx6, XP_761582; ScGrx1, NP_009895; ScGrx2, NP_010801; ScGrx3, NP_010383; ScGrx4, NP_011101; ScGrx5, NP_015266; ScGrx6, P38068; ScGrx7, NP_010274; CaGrx1, XP_721348; CaGrx2.1, XP_719021; CaGRx2.2, XP_721347; CaGrx4, XP_720477; CaGrx5, XP_711746; CaGrx6, XP_711401; NcGrx1, XP_961585; NcGrx4, XP_964063; NcGrx5, XP_960860; NcGrx6, XP_965167; MgGrx1, XP_360072; MgGrx4, XP_369841; MgGrx5, XP_364697; MgGrx6, XP_366151; AnGrx1, XP_661819; AnGrx4, XP_680836; AnGrx5, XP_661908; AnGrx6, XP_868846; RoGrx1, RO3G_01888; RoGrx4, RO3G_11698; RoGrx5.1, RO3G_16507; RoGrx5.2, RO3G_14195; RoGrx6, RO3G_08630. Accession numbers for Rhizopus oryzae sequences come from the Broad Institute databases ( The tree was constructed using Mega4. Basidiomycetes are shaded in grey.

Although the active site of LbGrx2.2 has an unusual monothiol motif (CPHS), the protein clustered in the dithiol CPYC group. In fact, only one gene was initially annotated because the CPHS and the CPYC isoforms from L. bicolor were located consecutively on the same scaffold. The fact that they display only 51.7% identity suggests that the two genes originated from an ancient duplication event. The presence in this CPYC group of the R. oryzae Grx5.1 (RoGrx5.1), exhibiting a CGFS active site, was striking. Nevertheless, except for the active site, most other features associated with CPYC proteins were conserved and those associated with CGFS Grx were not. In particular, RoGrx5.1 does not possess the conserved cysteine found in position 117 in ScGrx5, which might be involved in the deglutathionylation reaction in yeast (Supporting Information Fig. S1) (Tamarit et al., 2003). These two examples highlight the complexity of the Grx family. Indeed, the active-site sequences are not always the only criteria used to determine the Grx subgroup and, for example, some Grx with a monothiol active site could use a dithiol mechanism when an additional cysteine outside the active site is used in some reactions. Most other fungi only possess a single isoform with a CPYC active site, except for S. cerevisiae, which exhibits two CPYC proteins (ScGrx1 and ScGrx2), and C. albicans, which has two CPYC isoforms (CaGrx2.1, CaGrx2.2) and one CPFC isoform (CaGrx1) (Table 1, Supporting Information Fig. S1). According to prediction software, the L. bicolor isoforms LbGrx2.1 and LbGrx2.2 are mitochondrial and P. chrysosporium Grx1 (PcGrx1) is located in the nucleus (Tables 2 and 3). Nevertheless, PcGrx1, LbGrx2.1, LbGrx2.2 and ScGrx2 exhibit in-frame methionines that could correspond to putative transcription or translation initiation sites, allowing their cytosolic localization. This would lead to a differential subcellular distribution of these proteins within the cell. For example, yeast ScGrx2 is located in both the cytosol and mitochondria, whereas ScGrx1 is located exclusively in the cytosol (Porras et al., 2006).

Table 2.  Major features of the Laccaria bicolor reducing and thiol-antioxidant systems
Gene productaGene modelsbProtein IDcLengthdPutative localizationeRedox centersfExonsgESTsh
LbPrx Q1eu2.Lbscf0003g06110308972169CGCTRQACQF5Yes
LbPrx Q2fgenesh3_pg.C_scaffold_8000221325727226N/CGCTNQACGF5Yes
Lb2-Cys PrxestExt_Genewise1_human.C_200095174328219CFTFVCPTEIL7Yes
Lb1-Cys PrxestExt_Genewise1_human.C_40887172363224CFTPVCTTEL7Yes
Table 3.  Major features of Phanerochaete chrysosporium reducing and thiol-antioxidant systems
Gene productaGene modelsbProtein IDcLengthdPutative localizationeRedox centersfExonsgESTsh
PcGrx4fgenesh1_pm.C_scaffold_700002310446239C, NWAEPC/RCGFS3Yes
PcPrx Q1fgenesh1_pg.C_scaffold_120000516867237CGCTKEACEF5Yes
PcPrx Q2e_gwh2.12.297.1124208218NGCTTQACGF5Yes
Pc2-Cys Prxfgenesh1_pg.C_scaffold_190000368807211C/NFTFVCPTEIL5Yes
Pc1-Cys Prx.1fgenesh1_pm.C_scaffold_200005210009222CFTPVCTTEL7No
Pc1-Cys Prx.2fgenesh1_pm.C_scaffold_2000052126313229CFTPVCTTEL7No

A second subgroup, clustering independently from the dithiol group described above, includes LbGrx4 and PcGrx4, as well as ScGrx3 and ScGrx4 (Fig. 2). All these proteins possess a CGFS active-site motif and a Trx domain in the N-terminal part. Although it contains a WAxPC sequence that is reminiscent of the authentic Trx active-site motif, WCGPC, the N-terminal end is very divergent between fungi (Supporting Information Fig. S2). The Grx domain, including the CGFS active site, is by far more conserved among the fungi analysed. It is notable that R. oryzae is the only fungus possessing a protein with two Grx domains repeated in tandem, a situation previously described for mammalian and human PICOT proteins, whereas proteins with three Grx domains also exist in plants (Rouhier et al., 2006a; Herrero & de la Torre-Ruiz, 2007). The two S. cerevisiae monothiol Grx, with a hybrid Trx–Grx structure (ScGrx3 and ScGrx4) are both localized in the nucleus, and the Trx domain is important for nuclear import (Molina et al., 2004). LbGrx4 was predicted to be cytosolic, and algorithms predicted a cytosolic or nuclear localization for PcGrx4. However, this has not been determined experimentally, and the significance of in silico prediction of proteins targeted to the nucleus is lower than for mitochondrial/plastidic proteins.

Another monothiol Grx subgroup clustered separately in the phylogenetic tree (Fig. 2). It also contains proteins with a CGFS active site, but these proteins do not have the N-terminal Trx motif. The size of these proteins is thus shorter than those described previously. L. bicolor and P. chrysosporium each possess one isoform of this group, named LbGrx5 and PcGrx5. Homologs have been found in all fungal genomes analysed (Table 1, Supporting Information Fig. S3). The prediction of a mitochondrial location for these proteins is reliable because it is predicted for all the fungal Grx5 proteins, and this has been confirmed experimentally in S. cerevisiae and in S. pombe (Rodríguez-Manzaneque et al., 2002; Chung et al., 2005). In S. cerevisiae, this protein is involved in the formation of Fe–S clusters (Rodríguez-Manzaneque et al., 2002). A null mutant accumulates free iron in the cells, is hypersensitive to external oxidants and shows protein hypercarbonylation compared with wild-type cells, an indication of intracellular oxidative conditions above normal levels (Rodriguez-Manzaneque et al., 1999).

Additionally to these classical dithiol and monothiol Grx proteins, it appears that all fungi possess an additional Grx subgroup with CPYS active-site sequences (Herrero et al., 2006). The proteins clustered close to dithiol Grxs, but independently and in two classes dependent on the phylum (Fig. 2). S. cerevisiae possesses two different isoforms, one with a CPYS active site and one with a CSYS active site. P. chrysosporium and L. bicolor, as well as all other fungi, have only one CPYS isoform. The Aspergillus nidulans isoform exhibits a slightly divergent CPFS motif. The predicted localization is variable depending on the organism considered. The protein is predicted to be nuclear for P. chrysosporium, A. nidulans and R. oryzae, mitochondrial for S. cerevisiae, secreted for L. bicolor, Coprinopsis cinerea, Magnaporthe grisea, Neurospora crassa and C. albicans, or cytoplasmic for C. neoformans and Ustilago maydis. The N-terminal parts are very divergent, but all proteins, except A. nidulans Grx6, are predicted to have a transmembrane domain at this end (Supporting Information Fig. S4). So far, the role of this new group of Grx is not known.

As far as Grx are involved in thioltransferase reactions, they need to be reduced in the cell. The reducing system is composed of glutathione and of a NADPH-dependent pyridine nucleotide disulfide oxidoreductase called glutathione reductase (GR). L. bicolor has only one gene coding for GR, whereas three are found in the P. chrysosporium genome. The sequences are highly conserved among species, especially the CVNVGCVP motif containing the catalytic disulfide (Tables 2 and 3). The three isoforms of P. chrysosporium are very close to each other and the genes are located sequentially on the same scaffold, strongly suggesting that a duplication event occurred quite recently. The L. bicolor GR displays c. 68% identity with the three P. chrysosporium GR genes. R. oryzae also exhibits three different genes, whereas other fungi analysed possess only one GR gene. In the neighbour-joining phylogenetic tree, the GR proteins clustered according to fungal evolution as we can distinguish a basidiomycete and an ascomycete group, with the only zygomycete sequence available being closer to ascomycetes (data not shown). As glutathione and Grx are present in most subcellular compartments, a GR should be present in each compartment to keep glutathione reduced. From the software programs used for cellular localization prediction, the L. bicolor GR would be located in the cytosol. Whether this protein can also be targeted to mitochondria has not been established. P. chrysosporium GR1 and GR2 (PcGR1 and PcGR2) could be directed to mitochondria, whereas P. chrysosporium GR3 (PcGR3) is predicted to be cytosolic. It has been shown in S. cerevisiae that a single GLR1 gene encodes both the mitochondrial and cytosolic forms of GR, with the presence of two translation initiation sites in the same transcript (Outten & Culotta, 2004). The authors have shown that the presence of an A in position –3 is favourable to translation initiation. When looking carefully at P. chrysosporium GRs, putative alternative translation initiation sites, possessing the preferred consensus sequence AXXAUG, are found in PcGR1 and PcGR2, suggesting that the two proteins can be dually targeted. Overall, three different cytosolic isoforms and two mitochondrial isoforms could be produced in this organism. Whether such a high number of GR isoforms has an impact of the fungal physiology will await additional experimental studies.

Thioredoxin reductase/thioredoxin reducing system in fungi

The so-called Trx system is usually constituted by a Trx and a pyridine nucleotide-disulfide oxidoreductase named NADPH-dependent thioredoxin reductase (NTR). It is an ubiquitous system as these two components have been identified in organisms of the different kingdoms (i.e. Archaebacteria, Eubacteria, Plantae, Mycota) and other higher eukaryotes. Thioredoxins are small proteins possessing a well-conserved WCGPC catalytic site and, as a result of these two redox-active cysteines, they are able to reduce oxidized target proteins by dithiol–disulfide exchanges. The further reduction of Trxs occurs through NTR. Among kingdoms, two forms of thioredoxin reductases exist: low-molecular-weight isoforms (approx. 70 kDa as a dimer) are found in bacteria, fungi and plants, whereas high-molecular-weight isoforms (approx. 110 kDa as a dimer) are found in mammals and in some organisms such as Plasmodium falciparum or Drosophila melanogaster (Arnér & Holmgren, 2000). In the well-known S. cerevisiae system, two distinct cytosolic and mitochondrial NTRs have been characterized (Pedrajas et al., 1999). Only one gene encoding NTR has been found in the other sequenced fungal genome analysed (Table 1). Nevertheless, using green fluorescent protein (GFP) fusion, it has been shown, in C. neoformans, that the thioredoxin reductase is found both in cytoplasm and mitochondria (Missall & Lodge, 2005a). Alternative splicing has also been reported for NTR in other organisms (Agorio et al., 2003; Reichheld et al., 2005). For instance, in P. chrysosporium, an N-terminal extension recognized as a transit peptide has been predicted, by several programs (psort, mitoprot and targetP), to target the protein to mitochondria. Because of the generally low conservation of this region, the presence of N-terminal transit sequences can be easily misannotated. Whether NTRs are also dually targeted to the cytosol and to mitochondria in other fungi, through such extensions, or even without, has yet to be proven, but for most fungi analysed, the presence of Trx in mitochondria would indicate that a NTR is also required in this compartment.

In the fungal genomes analysed, between one and four different Trx isoforms have been found. The difference in gene content for Trx in L. bicolor and P. chrysosporium (that have four Trx isoforms) compared with other fungi (that have one to three Trx isoforms) could be explained by a duplication event (Table 1). Indeed, out of a total of four Trx genes in L. bicolor, three (LbTrx1, LbTrx2 and LbTrx3) share similar structures with exon size strictly conserved and a high sequence identity between LbTrx1 and LbTrx2 (88%) (Supporting Information Fig. S5, and data not shown). In addition, two of these genes are repeated in tandem. At present, we do not know whether this difference would provide a significant physiological advantage and whether this has led to specific functions for the additional Trx isoforms.

A phylogenetic tree only constructed with fungal Trxs is very difficult to analyse, in particular because of the small number of analysed sequences. We have therefore added several Trxs from plants, mammals, bacteria and other fungi, to improve the reliability of our analysis (Fig. 3). It appears that fungal Trxs are mainly distributed into four subgroups. For instance, PcTrx1 and PcTrx3, and LbTrx1, LbTrx2 and LbTrx3, cluster mostly with other basidiomycete sequences in a group, specific to fungi, which also contains ScTrx3 and RoTrx1. A second group contains most Trxs from ascomycetes. A third group containing basidiomycete Trxs, and particularly PcTrx4, is close to mammalian cytosolic Trx. Orthologs of PcTrx4 are only found in a limited number of fungal genomes and we have added sequences from Hebeloma cylindrosporum, Trametes versicolor and Gloeophyllum trabeum to delineate this subgroup in more detail. PcTrx2 is a member of the fourth group of fungal Trxs, specific to basidiomycetes and closely related to the mammalian mitochondrial Trxs. Analysis of PcTrx2, which exhibits an N-terminal extension, as well as other members of this group, suggests that these Trxs might be also located in mitochondria. Nevertheless, ScTrx3 and CnTrx1, which have been shown experimentally to be mitochondrial, cluster independently (Pedrajas et al., 1999). It is interesting to note that the PcTrx2-containing group belongs to a larger cluster, including also bacterial Trxs, chloroplastic Trxs m, x and y, and the mitochondrial mammalian Trxs. As a result of the presence of bacterial Trxs and the prokaryotic origin of Trxs m, x and y (Meyer et al., 2002), it is tempting from this phylogenetic tree to conclude that the members of this cluster derive from a prokaryotic ancestor, thus confirming the symbiotic origin of chloroplasts and mitochondria. On the contrary, the plant mitochondrial Trxs o, chloroplastic Trxs f and the Trxs h, which cluster independently, have been proposed to have an eukaryotic origin, confirming the higher diversity of Trxs in plants (Meyer et al., 2002, Gelhaye et al., 2005). With this logic, fungal Trx members clustering outside the fourth group would have an eukaryotic origin.

Figure 3.

Phylogenetic tree of the thioredoxin (Trx) family. Gene models for Laccaria bicolor and Phanerochaete chrysosporium Trxs are listed in Tables 2 and 3. Accession numbers for the sequences of the nine fungi analysed are as follows: AnTrx1, XP_657774; AnTrx2, XP_681840; CaTrx, XP_719372; CcTrx1, EAU92470; CcTrx2, EAU92826; CnTrx1, XP_569667; CnTrx2, XP_569429; MgTrx, XP_361762, NcTrx1, XP_962887; NcTrx2, XP_962263; RoTrx1, RO3G_13608; ScTrx1, NP_013144; ScTrx2, NP_011725; ScTrx3, NP_010006; UmTrx1, XP_762659; UmTrx2, XP_757517. Accession numbers for proteins from other organisms are listed below: At, Arabidopsis thaliana. AtTrxCXXS1, At2g40790; AtTrxCXXS2, At1g11530; AtTrxf1, At3g02730; AtTrxf2, At5g16400; AtTrxh1, At3g51030; AtTrxh2, At5g39950; AtTrxh3, At5g42980; AtTrxh4, At1g19730; AtTrxh5, At1g45145; AtTrxh7, At1g59730; AtTrxh8, At1g69880; AtTrxh9, At3g08710; AtTrxm1, At1g03680; AtTrxm2, At4g03520; AtTrxm3, At2g15570; AtTrxm4, At3g15360; AtTrxo1, At2g35010; AtTrxo2, At1g31020; AtTrxx, At1g50320; AtTrxy1, At1g76760; AtTrxy2, At1g43560; EcTrx, Escherichia coli (NP_756559); GtTrx, Gloeophyllum trabeum (EB003034); HcTrx, Hebeloma cylindrosporum (CK995145); HsTrx1, Homo sapiens (NP_759908); HsTrx2, Homo sapiens (NM_012473); MmTrx1, Mus musculus (BAA04881); MmTrx2, Mus musculus (EDL29654); PfTrx, Pseudomonas fluorescens (YP_263040); RnTrx2, Rattus norvegicus (NP_445783); SeTrx, Salmonella enterica (NP_457831); TvTrx, Trametes versicolor (EB077553); VvTrx, Vibrio vulnificus (NP_759908). Accessions number for Rhizopus oryzae sequences were obtained from the Broad Institute databases ( The tree was constructed using Mega4. Basidiomycetes are shaded in grey.

Thiol-peroxidases and sulfiredoxin in fungi

The genome of L. bicolor and P. chrysosporium has revealed the existence of respectively seven and eight Tpxs. Among the thiol-peroxidase family, five subgroups (2-Cys Prxs, 1-Cys Prxs, Prx Q or bacterioferritin comigratory protein (BCP), Prx II and Gpx) have been distinguished in plants and the same situation seems to have evolved in fungi (Table 1, Fig. 4) (Rouhier & Jacquot, 2005). The difference between the two fungi comes from the presence of an additional isoform of the 1-Cys Prx type in P. chrysosporium (Tables 1–3). The distribution among fungi is variable, with no 2-Cys Prx in U. maydis, M. grisea, N. crassa and A. nidulans, and no Prx Q in R. oryzae (Table 1). Nevertheless, as there are always some gaps in those sequenced genomes, we cannot completely exclude that some orthologs exist in these species. However, in the case of the 2-Cys Prx, the absence of these proteins in the above-mentioned species is correlated with the absence of Srx, an enzyme that seems to interact only with Tpx of the 2-Cys Prx subgroup, supporting the absence of this regulatory pathway in some fungi. Another difference is the number of Gpx, with variations from one in most fungi to four in C. albicans. It is not clear, at present, whether all the sequences are expressed. Independently of the Tpx subgroup considered, it appears from the phylogenetic analysis that the sequences from basidiomycetes mostly cluster together in separate clades, although a 1-Cys Prx isoform from P. chrysosporium (Pc1-Cys Prx.1) and a Gpx isoform from C. neoformans (CnGpx2) are more divergent and group separately (Fig. 4).

Figure 4.

Phylogenetic tree of the thiol-dependent peroxidase (Tpx) family. Gene models for Laccaria bicolor and Phanerochaete chrysosporium Tpxs are listed in Tables 2 and 3. Other accession numbers are as follows: CcPrxII.1, EAU81656; CcPrxII.2, EAU86553; UmPrxII.1, XP_759324; UmPrxII.2, XP_759094; NcPrxII.2, CAE76545; NcPrxII.1, XP_327166; MgPrxII.1, XP_368384; MgPrxII.2, XP_366634; CaPrxII.2, XP_720512; CaPrxII.1, XP_715909; AnPrxII.2, XP_681961; AnPrxII.1, XP_661291; ScPrxII (AhpC/cTpx3), P38013; CnPrxII, XP_572206; RoPrxII.1, RO3G_16017; RoPrxII.2, RO3G_08417; Ca1-CysPrx, XP_716935; Mg1-CysPrx, XP_362792; Cc1-CysPrx, EAU84705; Um1-CysPrx, XP_762551; Nc1-CysPrx, XP_325886; An1-CysPrx.1, XP_661577; An1-CysPrx.2, XP_659296; Sc1-CysPrx, P34227; Cn1-CysPrx (CnTsa3), XP_567781; Ro1-CysPrx, RO3G_02627; Ca2-CysPrx (CaTsa1), XP_716082; Cc2-CysPrx, EAU83474; Sc2-CysPrx.1 (ScTsa1, cTpx1), P34760; Sc2-CysPrx.2 (ScTsa2, cTpx2), Q04120; Cn2-CysPrx (CnTsa1), XP_571871; Ro2-CysPrx.1, RO3G_02462; Ro2-CysPrx.2, RO3G_12289; Ro2-CysPrx.3, RO3G_12682; CaPrxQ, XP_717695; MgPrxQ, XP_367592; CcPrxQ.1, EAU85749; CcPrxQ.2, EAU86548; UmPrxQ.1, XP_759358; UmPrxQ.2, XP_760998; NcPrxQ, XP_331323; AnPrxQ, XP_661905; ScPrxQ (Dot5/nTPx), P40553; CnPrxQ (CnTsa4), XP_569732; CcGpx, EAU85361; MgGpx, XP_367549; UmGpx, XP_757931; CaGpx1, XP_714295; CaGpx2, XP_714294; CaGpx3, XP_714296; CaGpx4, XP_714081; AnGpx, XP_660450; NcGpx, XP_329893; ScGpx1, P36014; ScGpx2, P38143; ScGpx3, P40581; CnGpx1, XP_570772; CnGpx2, XP_771839; RoGpx1, RO3G_14907; RoGpx2, RO3G_04463; RoGpx3, RO3G_14537. Accession numbers for Rhizopus oryzae sequences were obtained from the Broad Institute databases ( The tree was constructed using Mega4. Basidiomycetes are shaded in grey.

Within the Prx II group, there are two clearly separated subclasses containing proteins from all phyla (i.e. ascomycetes, basidiomycetes and zygomycetes). Nevertheless, S. cerevisiae and C. neoformans only possess one of the two isoforms. Overall, the identity between the sequences is very low (Supporting Information Fig. S6). These proteins, homologous to the peroxisomal membrane protein of 20 kDa (so-called PMP20 proteins), are 160–180 amino acids long. They are predicted to be mainly cytosolic proteins (Tables 2 and 3), although ScPrx II (also named cTPxIII) and most other members, including LbPrx II.1, LbPrx II.2 and PcPrxII.2, possess a putative peroxisomal signal sequence, such as AKL or AHL, in the C terminus. Experimental data based on GFP fusion indicated that ScPrx II (cTPxIII) is present in the cytosol (Park et al., 2000), whereas the C. boidinii PMP20 would effectively be associated with the inner side of the peroxisomal membrane (Horiguchi et al., 2001). When looking carefully at the sequence, the second cysteine, found 25 amino acids after the peroxidatic sequence in most plant orthologs, is only present in the two R. oryzae Prx II proteins. As there is no apparently conserved cysteine that could be used for recycling of the peroxidatic cysteine, the regeneration of the latter cysteine would proceed by direct reduction of the sulfenic acid formed during catalysis, as demonstrated with the C. boidinii PMP20 (Horiguchi et al., 2001).

In the 1-Cys Prx subgroup, all the fungi analysed possess a member of this group, with A. nidulans and P. chrysosporium containing two Prx isoforms. However, whereas the sequences from the A. nidulans group are very similar, those of P. chrysosporium are more distantly related, suggesting that the A. nidulans isoforms have been duplicated recently, whereas the two P. chrysosporium isoforms may result from an older duplication event (Table 1, Fig. 4). Sc1-Cys Prx and An1-Cys Prx.1 exhibit an N-terminal extension, which compartmentalizes the protein in mitochondria (Supporting Information Fig. S7) (Pedrajas et al., 2000). For the orthologs in other fungi, it is not clear whether this extension is absent or not detected by the annotation process. Hence, Lb1-Cys Prx, Pc1-Cys Prx.1 and Pc1-Cys Prx.2 are predicted to be cytosolic proteins (Tables 2, 3).

Concerning the 2-Cys Prx subgroup, the distribution is heterogeneous, ranging from zero (U. maydis, M. grisea, N. crassa and A. nidulans) to three (R. oryzae) isoforms, depending on the organism analysed (Table 1). All these proteins possess the conserved peroxidatic cysteine and a recycling C-terminal cysteine. The position of and the motif containing this recycling cysteine differ slightly in S. cerevisiae, with a VLPCN motif instead of a VCPA motif (Supporting Information Fig. S8). As demonstrated in S. cerevisiae, the proteins of this class are probably cytosolic, although some software programs also predicted a nuclear localization for Lb2-Cys Prx (Tables 2 and 3).

In the Prx Q subgroup, there is a subclass specific to basidiomycetes (Prx Q1), although there is no identified homolog in C. neoformans, and a subclass (Prx Q2) that is common to all fungi (Table 1, Fig. 4). The Prx Q1 proteins are predicted to be located in the cytosol, although no experimental evidence for this exists (Tables 2 and 3). The Prx Q2 subclass includes generally larger proteins with an N-terminal extension (Supporting Information Fig. S9). According to the nuclear localization of the S. cerevisiae protein (ScPrx Q2) (also called Dot5 or nuclear thioredoxin peroxidase (nTpx)), most of the proteins of this group, and especially LbPrx Q2 and PcPrx Q2, are predicted to be nuclear (Tables 2 and 3, and data not shown) (Park et al., 2000). The two cysteines putatively implicated in the catalytic and recycling mechanisms are separated by four amino acids. It is noticeable that two of them (UmPrxQ2 and CaPrxQ) do not possess the recycling cysteine, suggesting they are using an alternative regeneration mechanism (Supporting Information Fig. S9).

Concerning the Gpx clade, all Gpx from basidiomycetes group together except for one C. neoformans isoform (Fig. 4). Some ascomycetes, such as S. cerevisiae and C. albicans, and the zygomycete, R. oryzae, have extended families: three to four genes compared with only one gene in other ascomycetes analysed (Table 1). This could be related to specific functions. Indeed, it has been demonstrated in S. cerevisiae, for example, that Gpx3 is specialized in the oxidation and activation of the AP1 transcription factor (Delaunay et al., 2002). As there are no particular N-terminus or C-terminus targeting sequences, all Gpx are predicted and assumed to be cytosolic (Tables 2 and 3). This has been confirmed for the two Gpxs of C. neoformans (Missall et al., 2005). From results obtained with the S. cerevisiae Gpxs, all the fungal Gpxs are probably Trx-dependent enzymes. Indeed, the sulfenic acid formed on the peroxidatic cysteine is attacked by a recycling cysteine to form an intramolecular disulfide bridge, which is subsequently reduced by Trxs. Analysis of the amino acid sequence alignment indicates that these two cysteines are conserved in all Gpxs analysed, except for RoGpx3 (Supporting Information Fig. S10) (Tanaka et al., 2005). Whether this enzyme is functional and can use another regeneration system is not known, but this situation has also been found for a few other Gpxs and especially one in Chlamydomonas reinhardtii and Synechocystis (accession numbers: EDP03176 and slr1992).

Methionine sulfoxide reductases in fungi

Methionine sulfoxide reductases are important repairing and antioxidant enzymes, able to reduce free or peptide-bound MetSO into methionine. They use a thiol-assisted catalytic mechanism similar to the one used by most Tpx (i.e. sulfenic acid chemistry and a thiol-dependent regeneration mechanism). All fungi, except for R. oryzae, only possess one member of each class – MsrA or MsrB (Table 1) – raising the question of their compartmentalization as methionine oxidation could occur everywhere in the cell. When analysing the sequences, some MsrA and MsrB seem to possess an N-terminal extension, which would direct the protein into mitochondria (Supporting Information Figs S11, S12). In particular, LbMsrA, PcMsrA and PcMsrB, but not LbMsrB, are predicted to be mitochondrial enzymes (Tables 2 and 3). Again, for the proteins that do not possess such an extension, it is not always clear whether it is absent or not detected during the automatic or manual annotation processes. In addition, we should bear in mind that some proteins can be targeted to a subcellular compartment without clear targeting sequences and, as seems to be the case in fungi, a protein possessing an N-terminal extension can be dually targeted (Outten & Culotta, 2004; Porras et al., 2006).

The two subgroups are different in terms of primary and tertiary sequences, as attested by the phylogenetic tree and the amino acid sequence alignments (Fig. 5 ,Supporting Information Figs S11, S12). In addition, there is no apparent clustering of the sequences based on the fungi phylum (Fig. 5). The classification of MsrA and MsrB is usually based on the number and the position of the cysteines involved in the catalytic and regeneration mechanisms, and for MsrB on the presence of two CxxC motifs able to bind a zinc atom (Rouhier et al., 2006b). The identity between the sequences ranges from 37 to 75% for MsrA, from 37 to 78% for MsrB, but only from 1 to 14% between the two subgroups. All fungal MsrA analysed belong to the same functional group as they display two conserved cysteines: the catalytic cysteine is included in a GCFWG motif in the N-terminus part; and the single recycling cysteine is present in a GYxC motif in the C-terminus part, in a position similar to that of the Mycobacterium tuberculosis enzyme (Supporting Information Fig. S11) (Taylor et al., 2003). Similarly, all MsrB exhibit identical properties, comprising the two CxxC motifs for zinc binding, the catalytic cysteine in the C-terminus end, in a RHCVN motif, and the recycling cysteine in the middle of the sequence, in a CGWPA motif (Supporting Information Fig. S12).

Figure 5.

Phylogenetic tree of the methionine sulfoxide reductase (Msr) family. Gene models for Laccaria bicolor and Phanerochaete chrysosporium thiol-dependent peroxidases (Tpxs) are listed in Tables 2 and 3. Other accession numbers are as follows: CcMsrA, EAU87877; UmMsrA, XP_757993; NcMsrA, XP_323148; MgMsrA, XP_361769; CaMsrA, XP_719602; AnMsrA, XP_662118; ScMsrA, NP_010960; CnMsrA, XP_572086; RoMsrA1, RO3G_08254; RoMsrA2, RO3G_14738; CcMsrB, EAU90603; UmMsrB, XP_757875; NcMsrB, XP_957530; MgMsrB, XP_365794; CaMsrB, XP_718384; AnMsrB, XP_659536; ScMsrB, NP_009897; CnMsrB, XP_568991; RoMsrB, RO3G_14699. Accession numbers for Rhizopus oryzae sequences were obtained from the Broad Institute databases ( The tree was constructed using Mega4. Basidiomycetes are shaded in grey.

Expression analysis of L. bicolor genes coding for the reducing and antioxidant systems in various fungal tissues

From the EST analysis in L. bicolor summarized in Table 2, only four genes (LbGrx2.1, LbGrx2.2, LbTrx2 and LbPrxII.1) were not firmly established as ‘expressed genes’; however, except for LbGrx2.2, which was not identified in the first annotated version of the L. bicolor genome and thus not present on the whole-genome DNA chips, the expression of the three other genes was confirmed by the microarray analysis. Contrary to L. bicolor, approximately half of the P. chrysosporium genes listed in Table 3 are found in EST databases. Using semiquantitative reverse transcription–polymerase chain reaction (RT-PCR), the expression of all of those genes, except for MsrA and MsrB that were not tested, was confirmed in P. chrysosporium growing in liquid culture (data not shown).

Taking advantage of the design of this whole-genome DNA chips containing the 20 614 gene models first identified in Laccaria, we investigated the expression of all those genes in three different tissues (differentiating fruit body, free-living mycelium and ectomycorrhizal tissues derived from the interaction between Poplar or Douglas fir and L. bicolor S238N) (Martin et al., 2008). These three tissues are the major developmental stages in the life cycle of ectomycorrhiza (ECM) fungi. In mutualistic symbioses, including ECM symbiosis, the host plant has a strong impact on the physiology and fitness of the symbiosis. It was thus crucial to evaluate the impact of an alternate host, such as poplar and Douglas fir, on the Laccaria transcriptome (Martin et al., 2008). Figure 6 presents the modifications of the transcript level in Poplar ECM, Douglas fir ECM and fruiting bodies compared with free-living mycelium. All of these genes are expressed in the different tissues with intensities of between 2000 and 38 000 (data not shown), but the expression of most of them is not significantly regulated in ECM or fruiting bodies compared with free-living mycelium. Among the significantly regulated transcripts, Grx1 is up-regulated in the two mycorrhizal systems analysed compared with mycelium. The expression of Prx Q1 and Gpx is increased both in ECM and fruiting bodies compared with the expression of mycelium. The expression of Prx Q2 and Prx II.1 was only increased in Douglas fir ECM. By contrast, some genes of these multigenic families are down-regulated. For example, Trx2 and 2-Cys Prx are down-regulated in ECM compared with free-living mycelium, and the expression of Trx2 was also repressed in fruiting bodies. By contrast, the GR and 1-Cys Prx transcript levels were only significantly changed in fruiting bodies. Using semiquantitative RT-PCR, the expression of two of these regulated genes (Lb2Cys Prx and LbTrx) was analysed in all conditions mentioned before. The decreased expression observed with the microarray analysis was confirmed for both genes, although the mean fold change is lower (data not shown).

Figure 6.

Differential expression of genes from the Laccaria antioxidant and reducing systems, as revealed by the NimbleGen Laccaria whole-genome expression array. The log 2 values of the mean fold change of these genes in Poplar ectomycorrhizas (ECM), Douglas fir ECM and fruiting bodies compared with free-living mycelium (FLM) are shown. Fold changes of > 2, with a P-value of < 0.05 and a posterior probability of differential expression (PPDE) value of > 0.95 ( were considered as significantly regulated and are indicated by asterisks.

The significance of all these expression changes is not yet clear, but the requirement of antioxidant may be essential for the following reasons. Free-living mycelium mimics the hyphae growing in soil and litter as facultative saprotrophs. This mycelium is probably exposed to abiotic stresses, such as pH changes, nutrient starvation and drought stress, and also to deadly interactions with other microorganisms. Mycelium in ECM tissues should compensate for the ROS generated by the colonized plants. Finally, fruit body formation involves striking development stresses, including apoptosis, which probably require a finely tuned redox metabolism. Thus, as ROS synthesis occurs during many processes, from developmental processes to plant–fungi interactions, the regulation of some fungal antioxidant proteins is finally not very surprising. There could be a fine tuning, as only some members of the Tpx family are, for example, up-regulated during the establishment of ECM, whereas others are simply down-regulated or do not see their expression vary. In addition, independently to ROS signalling, Trxs, and Grxs in particular, via their capacity to regulate a great number of enzymes by redox regulation, are potentially involved in a large variety of metabolic processes.


Comparison of the gene content of fungi from all phyla did not show huge differences for multigenic families in the ROS detoxification network, although some protein subgroups were totally absent in some fungi. Most of the time this corresponded to individual gene losses, as at least some members of a given phylum conserved the genes. The case of Srx may be different because this gene has been identified in hemiascomycetes but in only one isolated basidiomycete and one zygomycete. P. chrysosporium and L. bicolor, two fungi belonging to basidiomycetes, but with different ways of life, do not exhibit many differences in their gene content. While the predicted gene content is much higher in L. bicolor than in P. chrysosporium (20 614 genes compared with 10 048 genes), the latter exhibited three more genes (two GR and one 1-Cys Prx) in the families analysed compared with L. bicolor. More generally between the basidiomycetes analysed, although some are plant pathogens (U. maydis), human pathogens (C. neoformans), saprobic fungi (P. chrysosporium and C. cinereus) or ECM fungi (L. bicolor), the difference in gene content is weak. The major difference is a more developed Trx family in L. bicolor and P. chrysosporium (four genes; probably resulting from duplication events) compared with the three other basidiomycetes.

The gene content is not the only important point to consider because, as a result of differential transcription or translation initiation, a given gene can generate several proteins in the cells. This is especially crucial for the subcellular distribution of the proteins. It has already been demonstrated several times (ScGR, ScGrx2) that a protein can be dually targeted and the analysis of the gene sequences provides additional candidates for such regulation.


Financing from the ANR program (06-2_138585) to M. M., E. G. and N. R. is acknowledged. The genome sequencing of Laccaria bicolor H82 was funded by the US Department of Energy's Office of Science, Biological, and Environmental Research Program and by the University of California, Lawrence Berkeley National Laboratory, under contract no. DE-AC02-05CH11231, by the Lawrence Livermore National Laboratory under contract no. DE-AC52-07NA27344, and by the Los Alamos National Laboratory under contract no. DE-AC0206NA25396. The EST sequencing and transcriptome analysis were funded by the US Department of Energy, INRA ‘AIP Séquençage’, Région Lorraine and the European Network of Excellence EVOLTREE grants.