• In forest soils, ectomycorrhizal and saprotrophic Agaricales differ in their strategies for carbon acquisition, but share common gene families encoding multi-copper oxidases (MCOs). These enzymes are involved in the oxidation of a variety of soil organic compounds.
• The MCO gene family of the ectomycorrhizal fungus Laccaria bicolor is composed of 11 genes divided into two distinct subfamilies corresponding to laccases (lcc) sensu stricto (lcc1 to lcc9), sharing a high sequence homology with the coprophilic Coprinopsis cinerea laccase genes, and to ferroxidases (lcc10 and lcc11) that are not present in C. cinerea. The fet3-like ferroxidase genes lcc10 and lcc11 in L. bicolor are each arranged in a mirrored tandem orientation with an ftr gene coding for an iron permease. Unlike C. cinerea, L. bicolor has no sid1/sidA gene for siderophore biosynthesis.
• Transcript profiling using whole-genome expression arrays and quantitative reverse transcriptase–polymerase chain reaction (qRT-PCR) revealed that some transcripts were very abundant in ectomycorrhizas (lcc3 and lcc8), in fruiting bodies (lcc7) or in the free-living mycelium grown on agar medium (lcc9 and lcc10), suggesting a specific function of these MCOs.
• The amino acid composition of the MCO substrate binding sites suggests that L. bicolor MCOs interact with substrates different from those of saprotrophic fungi.
Ectomycorrhizal (ECM) fungi establish a series of hyphal networks with different physiological activities, namely the mycorrhizal mantle, the intraradicular Hartig net, the rhizomorphs colonizing the soil and the extraradical mycelium growing on decaying organic matter (Cairney & Burke, 1996). In colonizing roots, ECM fungi have a direct access to reduced carbon from their host plant (Read & Perez-Moreno, 2003). The observation that some ECM fungi possess carbon isotopic signatures that are intermediate between those of purely biotrophic and purely saprotrophic fungi, however, suggests that several species possess significant saprotrophic capacity (Koide et al., 2008). The ECM extraradical mycelia, which have a central position at the soil–tree interface, can produce extracellular enzymes such as proteases and carbohydrate-active enzymes (Cairney & Burke, 1994; Bending & Read, 1997; Chalot & Brun, 1998), thus facilitating access to products of biopolymer degradation. In most boreal and temperate forests, ECM fungi dominate the decomposed litter and humus, where they apparently mobilize nitrogen and make it available to their tree hosts (Lindahl et al., 2007; Hobbie & Horton, 2007). Nonspecific enzymes, such as polyphenol oxidases, are involved in degradation of various organic compounds found in decomposing litter (Baldrian, 2006) and they might be released by ECM fungi to facilitate nutrient acquisition by the multitrophic associations (bacteria–ECM fungi–tree roots) (Cairney & Burke, 1994; Bending & Read, 1997).
Polyphenol oxidases, usually called multi-copper oxidases (MCOs), include laccases, ascorbate oxidases, and Fet3 ferroxidases. These enzymes typically bind four copper atoms in two highly conserved copper-binding centers (Messerschmidt & Huber, 1990). Laccases (p-diphenol: O2 oxidoreductases; EC 220.127.116.11), belonging to a group of blue copper oxidases, catalyze the one-electron oxidation of phenols, aromatic amines, and other electron-rich substrates by reducing molecular oxygen to water through an oxidoreductive multi-copper system (Thurston, 1994). They are widely distributed in numerous basidiomycetous and ascomycetous fungi (Thurston, 1994; Luis et al., 2004; Baldrian, 2006; Hoegger et al., 2006), but also in higher plants, insects, and bacteria (Mayer & Staples, 2002). Laccases provide a diverse array of biological functions. Their multiple functional roles in development (e.g. formation of fruiting bodies), lignin and humus degradation (Baldrian, 2006; Snajdr et al., 2008), metabolism (e.g. pigment formation and depolymerization of organic compounds), and pathogen–host interactions have been comprehensively reviewed by Thurston (1994) and Burke & Cairney (2002).
Fet3 ferroxidase (Fe(II): oxygen oxidoreductase, EC 18.104.22.168) is a plasma membrane protein in which the ferroxidase-containing domain responsible for ferroxidase activity is located on the external cell surface. The Fet3-Ftr1-based transport system, required for high-affinity iron uptake, consists of an oxidase (Fet3) and a permease (Ftr1) that work together to facilitate transmembrane iron transport (Askwith & Kaplan, 1998). The Fet3-Ftr1 system has been described in the white-rot fungus Phanerochaete chrysosporium (Larrondo et al., 2007), and in the yeasts Saccharomyces cerevisiae, Schizosaccharomyces pombe and Arxula adeninivorens (Wartmann et al., 2002; Canessa et al., 2005) and also appears to be present in many pathogenic fungi such as Ustilago maydis, Fusarium graminearum, Aspergillus fumigatus and Cryptococcus neoformans (Schrettl et al., 2004; Eichhorn et al., 2006; Park et al., 2007; Jung et al., 2008). Fet3 ferroxidases are able to catalyze the oxidation of a variety of organic compounds in addition to ferrous iron (Baldrian, 2006).
Laccaria bicolor is a ubiquitous ECM symbiont of hardwood and conifer roots found in different ecological niches. The draft-genomic sequence of L. bicolor has recently been published (Martin et al., 2008; Martin & Selosse, 2008). Analysis of the L. bicolor gene repertoire revealed that its genome is lacking carbohydrate-active enzymes involved in plant cell wall degradation, and, as a consequence, it is unable to use cellulose, pectins, and pectates as a carbon source. However, L. bicolor possesses expanded multigene families associated with hydrolysis of bacterial, fungal and microfaunal polysaccharides and proteins. It has been suggested (Martin & Selosse, 2008) that these secreted proteases, chitinases and glucanases are involved in the mobilization of nitrogen from organic substrates. As nitrogen-containing polymers in soils often are complexed with polyphenolic compounds, oxidation of polyphenols may be required before hydrolytic enzymes may access their nitrogenous substrates (Bending & Read, 1997).
Here, we characterize the complete set of genes coding for MCOs in the L. bicolor genome. This includes cataloguing the MCO genes and their predicted proteins, and comparing this multigene family with those of other fungi, such as the white-rot fungus P. chrysosporium and the coprophilic C. cinerea, which differ in their strategies of carbon acquisition.
Materials and Methods
Strain and culture conditions
The dikaryotic parental strain S238N and the sib-monokaryotic S238N-H82 of Laccaria bicolor (Maire Orton) (Di Battista et al., 1996) were grown at 25°C and maintained on Pachlewski medium agar plates. For transcript extraction, 5-mm agar plugs of the free-living vegetative mycelium from a stock culture were transferred onto cellophane-covered agar plates containing high sugar (20 g l−1 glucose and 5 g l−1 maltose) Pachlewski medium and were grown for 2 wk. The mycelium was snap-frozen in liquid nitrogen immediately after sampling and stored at −80°C until further analysis.
The inoculum of L. bicolor S238N was prepared by aseptically growing the mycelium in a peat–vermiculite nutrient mix (Duponnois & Garbaye, 1991). The formation of ectomycorrhizas is only induced with the dikaryotic strain S238N. Seeds of Douglas-fir (Pseudotsuga menziesii D. Don) from provenance zone 422 (Washington State, USA) were pretreated in moist peat at 4°C for 1 month to break dormancy. Before the inoculation, the top 15 cm at the nursery soil of the INRA (Institut National de la Recherche Agronomique) Center of Champenoux (a brown acidic soil) was sterilized with methyl bromide. One liter of fungal inoculum was mixed per 1 m3 of the nursery soil and seedlings were then sown. Nine months later, nursery-grown mycorrhized Douglas-fir seedlings were harvested, their root systems were washed in tap water and short roots forming mycorrhizas with L. bicolor were sampled under a stereomicroscope. A laser-scanning confocal microscopy image of a transverse section of these ectomycorrhizas is shown in fig. 1(b) in Martin et al. (2008). Fruiting bodies of L. bicolor S238N growing near the seedlings were sampled at the same time (see fig. 1a in Martin et al., 2008).
Ectomycorrhizas of L. bicolor–Populus trichocarpa (Torr. & Gray) (poplar) were synthesized by growing cuttings of P. trichocarpa for 3 months in pots containing Terragreen (Brenntag Lorraine, Toul, France) mixed with fungal inoculum in a peat–vermiculite mix (4 : 1, volume:volume). Ectomycorrhizas and fruiting bodies were snap-frozen in liquid nitrogen immediately after sampling and stored at −80°C until further analysis.
Ab initio genome annotation and manual curation
The sequencing, assembly, and annotation of the L. bicolor genome were described by Martin et al. (2008). All L. bicolor sequences from the monokaryon H82 are available at the Joint Genome Institute (JGI) website (http://www.jgi.doe.gov/laccaria) and have been deposited at GenBank/European Molecular Biology Laboratory/DNA Data Bank of Japan under project accession number ABFE01000000. Complete Coprinopsis cinerea (=Coprinus cinereus (Schaeffer 1774: Fries)) DNA and protein sequences were obtained from the C. cinerea strain Okayama genome database (Coprinus cinereus Sequencing Project, Broad Institute; http://www.broad.mit.edu), while sequence data for Phanerochaete chrysosporium Burdsall (Martinez et al., 2004) were obtained from the Joint Genome Institute website (http://genome.jgi-psf.org/Phchr1/Phchr1.home.html). A search for the presence of sid1/sidA homologs was carried out in the Ustilago maydis (DC.) Corda genome (http://www.broad.mit.edu/annotation/genome/ustilago_maydis/Home.html; Hoegger et al., 2006). Using BLAST search and the INTERPRO domains (IPR001117 and IPR002355) at the JGI website, we identified gene models coding for MCOs in the draft genome of L. bicolor S238N-H82. Gene prediction at JGI was performed using four gene predicters: EUGENE, FGENESH, GENEWISE and TWINSCAN, and gene models were selected by the JGI annotation pipeline (Martin et al., 2008). Selection of the MCO models was based on expressed sequence tag (EST) support, completeness, and homology to a curated set of proteins. Additionally, searches were performed with the use of a range of MCO sequences available from fungi at National Center for Biotechnology Information (NCBI) GenBank (http://www.ncbi.nlm.nih.gov/) and UNIPROT (http://expasy.org/) to probe the L. bicolor genome database using BLASTN, TBLASTN, and BLASTP algorithms at the JGI L. bicolor BLAST server and the INRA Laccaria DB (http://mycor.nancy.inra.fr/IMGC/LaccariaGenome/). The MCO sequences were also used in a TBLASTN query against the JAZZ (Martin et al., 2008) and ARACHNE (Labbé et al., 2008) L. bicolor sequence assemblies. The putative homologs that were detected were characterized based on conserved domains, identities, and E-values in comparison with know proteins. Laccaria bicolor MCO gene models were edited when needed. Sequences of MCO exons were confirmed by sequencing of the corresponding cDNAs.
For phylogenetic analysis, the MCO amino acid sequences were aligned with ClustalX version 1.81 (http://bips.u-strasbg.fr/fr/Documentation/ClustalX/) using the following multiple alignment parameters: gap opening penalty 15, gap extension penalty 0.3, and delay divergent sequences set to 25%; and the Gonnet series was selected as the protein weight matrix. The obtained alignments were adjusted manually with GeneDoc version 2.6.002 (http://www.psc.edu/biomed/genedoc/). Heterogenous regions in the alignments were excluded, keeping only regions where the assignment of positional homology was possible. Maximum parsimony analysis was performed using paup* version 4.0b10 (Sinauer Associates Inc., Sunderland, MA, USA) and trees were created in TreeViewX version 0.5.0 (http://darwin.zoology.gla.ac.uk/~rpage/treeviewx/index.html). Neighbor joining trees were constructed using the Jones–Taylor–Thornton (JTT) substitution rate matrix in mega version 3.1 (http://www.megasoftware.net/). Bootstrap analysis was carried out with 500 replicates.
RNA isolation, PCR amplification and cDNA sequencing
The vegetative free-living mycelium grown on Pachlewski medium, and the L. bicolor–Douglas-fir and L. bicolor–P. trichocarpa ectomycorrhizas and fruiting bodies were ground in liquid nitrogen and total RNA was isolated using the RNeasy Plant Mini kit (Qiagen, Courtaboeuf, France). The RNase-free DNase set (Qiagen) was used to digest DNA during RNA purification. Full-length doubled-stranded cDNAs corresponding to mRNAs expressed in mycelia were obtained using the SMART–PCR cDNA Synthesis Kit (Clontech, Palo Alto, CA, USA). PCR amplification of the full-length cDNA, from the start to the stop codon, with primers designed using the nucleotide sequences of manually annotated gene models, was performed on a GeneAmp 9600 thermocycler (Perkin-Elmer Instruments, Shelton, CT, USA) using the Advantage 2 Polymerase Mix (Clontech). Successful PCR reactions resulted in a single band on a 0.8% agarose gel (Bioprobe; QBiogene, Illkirch, France) in Tris borate-ethylenediaminetetraacetic acid (EDTA) (Sambroock et al., 1989) and stained with ethidium bromide (2 µg ml−1; Roche, Rosny-sous-Bois, France). The size of the band was estimated using a 1-kb ladder (Invitrogen, Cergy Pontoise, France). Amplified products were purified with the Multiscreen PCR plate system (Millipore Corporation, Boston, MA, USA) according to the manufacturer's instructions. The concentration of cDNA was estimated with the low DNA mass ladder (Invitrogen). Direct cDNA sequencing was performed on the CEQ 8000XL automated sequencer (Beckman Coulter, Fullerton, CA, USA). Two nanograms of purified template cDNA was labelled during a cycle sequencing reaction with 5 ng of CEQ DTCS-Quick Start Kit (Beckman Coulter) in a GeneAmp 9600 thermocycler (Perkin Elmer Instruments). All full-length cDNAs were sequenced by cDNA walking. Sequences of the cDNAs described here are available at the NCBI database under accession numbers FJ432084 to FJ432094.
Whole-genome expression oligoarray analyses
Accumulation of predicted MCO transcripts was detected in free-living mycelium of L. bicolor S238N, ectomycorrhizal root tips of Douglas-fir and poplar, and fruiting bodies of L. bicolor S238N using the L. bicolor whole-genome expression oligoarray data described in Martin et al. (2008). Two versions of oligoarrays were used for the present analysis. The oligoarrays (v.1.0) contained eight oligonucleotide probes for each gene model, including MCOs, and each oligonucleotide was synthesized in duplicate on the oligoarrays (Martin et al., 2008). The complete expression data set obtained with oligoarrays (v.1.0) is available as a series under accession number GSE9784 at the Gene Expression Omnibus at NCBI (http://www.ncbi.nlm.nih.gov/geo/). The oligoarrays (v.2.0) contained three oligonucleotide probes for each gene model, including MCOs, and each oligonucleotide was synthesized in duplicate on the oligoarrays (A. Kohler & F. Martin, unpublished results). To estimate the signal background and the resulting threshold value for significant expression, the mean intensity of 30 000 random probes present on the microarray was calculated. Gene models with expression exceeding the threshold by threefold or more were considered to be transcribed. Cross-hybridizations between arrayed oligonucleotide probes and conserved regions of MCO transcript sequences were observed for several MCO sequences. For example, the eight oligonucleotide probes of the oligoarray (v.1.0) designed for lcc1 cross-hybridized to conserved sequences in lcc5. By contrast, the eight oligonucleotide probes designed for lcc4 and lcc9 were specific for these transcripts only. The three oligonucleotide probes of the oligoarray (v.2.0) designed for lcc7 cross-hybridized to conserved sequences of lcc6. As a consequence, for all lcc sequences, the number of specific oligonucleotide probe sequences varied (Table 4). Two biological replicates were used in the transcript profiling carried out using the v.1.0 oligoarrays, whereas three replicates were used for the v.2.0 oligoarrays. The reported gene expression values therefore corresponded to the mean intensity of hybridization signals obtained for the specific oligonucleotide probes. A Cyber-T test was performed on the mean for each transcript (P < 0.05).
Table 4. Quantification by quantitative reverse transcriptase–polymerase chain reaction (qRT-PCR) and by exon expression array of the transcript levels of the nine laccase (lcc1 to lcc9) and the two ferroxidase (lcc10 and lcc11) genes in different conditions
The NimbleGen array analysis was carried out using a pure culture of Laccaria bicolor growing on Pachlewsky medium (mycelium; 20 g l−1 glucose), fruiting bodies (Fb) of L. bicolor and mycorrhizas of L. bicolor–Douglas-fir (Lb/Dg) and L. bicolor–Populus trichocarpa (Lb/Pt). Transcript levels in agar-grown mycelium were used as the control values. Two biological replicates were used for each treatment with NimbleGen oligoarrays (v.1.0; NG1) and three biological replicates were used for each treatment with NimbleGen oligoarrays (v.2.0; NG2). The numbers in the column ‘oligos’ correspond to the numbers of specific oligonucleotides used to calculate the mean intensity for the value. For qRT-PCR analysis, three biological replicates were performed for each treatment. The elongation factor 3 gene was used as the reference transcript. Gray-shaded cells correspond to developmental stages where some multi-copper oxidase (MCO) genes showed a differential expression. For NimbleGen oligoarrays, a Cyber-T test was performed on the mean for each transcript (*, P < 0.05; **, P < 0.01). For qRT-PCR analysis, a t-test was performed for each transcript. An asterisk indicates a P-value < 0.05.
Measurement of MCO transcripts in mycelium, L. bicolor–P. menziesii ectomycorrhizas and fruiting bodies was performed using a two-step quantitative RT-PCR (qRT-PCR) procedure. Total RNA was quantified with a spectrophotometer and then reverse-transcribed (80 ng per reaction) using the iScript cDNA Synthesis kit (#170-8891; Bio-Rad, Marnes La Coquette, France). cDNAs were used as templates in real-time quantitative PCR reactions with gene-specific primers designed using Primer 3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) and amplify 3.1 (http://engels.genetics.wisc.edu/amplify) (Table S4). The following criteria were used: product size between 100 and 400 bp, melting temperature 60°C and a GC percentage > 50%. Target gene expression was normalized to the gene encoding the L. bicolor elongation factor 3 (Protein ID 293350). Reactions of qPCR were run using the MJ-opticon2 DNA real-time PCR system (Bio-Rad, Hercules, CA, USA). The following cycling parameters were applied: 95°C for 3 min and then 40 cycles of 95°C for 30 s, 60°C for 1 min and 72°C for 30 s. A control with no cDNA was run for each primer pair. For data analysis, the geometric mean of the biological replicates (n = 3) was calculated. The primer efficiency ranged between 90% and 110%.
Identification of MCO genes in the genome of L. bicolor
Based on the high conservation of amino acid ligands, a consensus sequence for MCOs has been defined corresponding to Prosite PDOC00076, InterPro IPR001117, IPR002355, and Pfam 00394. The ab initio annotation and subsequent automated BLAST and INTERPRO searches of the L. bicolor draft genome sequence (Martin et al., 2008) identified 11 gene models having this conserved MCO domain (Table 1). According to Hoegger et al. (2006), the products of nine gene models belong to laccases sensu stricto (lcc1 to lcc9) and the other two sequences to ferroxidases (lcc10 and lcc11). lcc6, lcc7 and lcc8 belonged to tandem repeated sequences, that is, multiple copies of a similar nucleotide sequence within a < 10-kbp region of the genome. lcc6 and lcc7 were only separated by an intergenic region of 536 bp, and lcc7 and lcc8 by a sequence of 788 bp. The lcc10 and lcc11 genes were both located close (< 400 bp) to iron permease genes (homologous to S. cerevisiae ftr1), with start to start codons in a mirrored tandem orientation (Fig. 1). This tandem repeat structure was confirmed on the ARACHNE assembly (Labbéet al., 2008) and on the finished sequence (J. Grimwood & F. Martin, unpublished results). The sequences of the other lcc genes were randomly distributed in the genome. Three nontranscribed truncated gene models (JGI Protein ID 238280, ID 310410 and ID 315309, with a homology to laccase sequence < 1 kb) were identified and were considered as pseudogenes.
Table 1. Number and classification of multi-copper oxidases (MCOs) in basidiomycete genomes and presence of homologs of representative genes of the high-affinity iron uptake pathways
Classification of MCOs is according to Hoegger et al. (2006). Lbi, Laccaria bicolour; Cci, Coprinopsis cinerea; Pch, Phanerochaete chrysosporium; Cne, Cryptococcus neoformans; Uma, Ustilago maydis; ftr, iron permease; lcc, laccase.
Laccases sensu stricto
Fungal pigment MCOs
ftr1 homologs clustered in genome with MCOs
The length of the nucleotide sequence for the 11 lcc genes ranged between 2103 and 2409 bp (Supporting Information Table S1). The coding exon sequences for these lcc genes were confirmed by EST alignment and cDNA sequencing; they are interrupted by short introns of length and structure typical of L. bicolor (Martin et al., 2008). The actual number of introns in the individual MCO genes varies between nine and 14 (Table S1). Based on the intron positions, the lcc gene family clustered into the two distinct subfamilies corresponding to the laccases sensu stricto (lcc1 to lcc9) and to the ferroxidases (lcc10 and lcc11) (Fig. 2a–c). In the first subfamily (lcc1 to lcc9), introns were predicted at 18 different positions, with 10–14 introns per individual gene, of which eight are shared by all (Fig. 3). The 10 introns in lcc6, lcc7 and lcc8 are all at identical positions (Fig. 3). In the second subfamily (lcc10 and lcc11), seven intron positions were found to be conserved in both genes, but each gene also has two unique intron positions. However, lcc10 and lcc11, encoding putative ferroxidases, have no intron positions in common with the laccase sensu stricto genes (Fig. 3).
Codon alignments and analysis of synonymous substitution rates (ps) indicated more recent duplication events for the group of lcc1, lcc2, lcc3, lcc5 and lcc9. The ps values of pairs of genes in this group were significantly lower (from 0.13 to 0.61) compared with those of the other laccase genes, which were all around the saturation level (Fig. S1). The ps was also very low for the ferroxidases lcc10 and lcc11, at 0.20, and for the corresponding permease pair ftr1a/ftr1b, at 0.19. The nonsynonymous substitution rate was 4% for both combinations.
Identification of the regulatory cis-elements
A region consisting of 2000 bp immediately upstream of the ATG start codon (5′-promoting region) or the region between the stop codon and the start codon of the next coding region were analyzed for the presence of conserved cis-elements. The lcc1 and lcc3 upstream sequences were similar (60%), with some conserved portions showing up to 80% nucleotide sequence identity. The lcc10 and lcc11 upstream regions presented an identity of 80% as a result of the presence of an ftr1 gene in a mirrored tandem orientation (Fig. 1). The other upstream sequences of lcc genes showed 40–60% sequence identity. The distribution of several putative cis-acting elements in the promoter region of the 11 MCOs is shown in Table S2. The TATA box (TATAA/TA) was located between positions −35 and −61 from the ATG start codon, except for lcc5 and lcc8, where it occurred at positions −175 and −363, respectively.
One to five putative carbon catabolite repressor (CreA) consensus binding sequences (SYGGRG), identified as a repressor involved in glucose repression (Kulmburg et al., 1993), were identified in the promoter region of L. bicolor lcc genes, except in lcc7 where no consensus sequence was found. One or two consensus TGCRCNC sequences (Thiele, 1992) corresponding to metal-responsive elements (MREs) were found, except in lcc4, lcc5, lcc7, lcc9 and lcc11. One to six potential stress-responsive elements (STREs) were also detected within the consensus core CCCT sequence (Treger et al., 1998). Two to 10 inferred nitrogen factor binding sites (GATA) (Marzluf, 1997) were located in the promoter region, except for lcc6, where this site was lacking. Long pyrimidine-rich regions, typical for strong fungal promoters and usually located upstream of the TATA box, were not found in the upstream regions of L. bicolor lcc genes.
A putative polyadenylation signal, AATAA, which is a slight variation of the consensus polyadenylation signal AATAAA sequence (Proudfoot, 1991), was found 121 and 368 bp downstream of the stop codon in lcc6 and lcc8, respectively.
The predicted Lcc proteins
The deduced protein sequences of the nine genes encoding laccases sensu stricto showed a length typical for fungal laccases sensu stricto (from 504 for Lcc1 to 540 aa for Lcc8) (Table S1). By contrast, the ferroxidases Lcc10 and Lcc11 were longer (632 and 607 aa, respectively) (Table S1). All Lcc proteins have a signal peptide of 16–22 amino acids (Table S3), the fungal Lcc signature sequences L1 to L4 (Kumar et al., 2003), and the postulated substrate-binding loops (Larrondo et al., 2003; Kilaru et al., 2006). The TargetP and SignalP programs predicted a signal peptide for all mature Lcc proteins except for Lcc5, suggesting an extracellular location.
Sequence identities among the 11 deduced proteins ranged from 26 to 85%. Lcc4 had less similarity to other Lcc proteins (26–62% identity). Lcc10 and Lcc11 had the most divergent sequences of all the proteins examined. The pairs Lcc1/Lcc2, Lcc3/Lcc5, Lcc7/Lcc8, and Lcc10/Lcc11 were most similar to each other (> 74% identity). In the phylogenetic analysis of basidiomycete laccases sensu stricto (Fig. 2a), L. bicolor proteins clustered into three groups: (1) Lcc1, Lcc2, Lcc3, Lcc5 and Lcc9, which clustered with members of the large C. cinerea laccase family; (2) Lcc4, which had an unresolved position; and (3) Lcc6, Lcc7 and Lcc8, which grouped with the Lentinula edodes sequences. In the phylogenetic analysis of fungal ferroxidases, Lcc10 and Lcc11 of L. bicolor clustered with other basidiomycete ferroxidases (Fig. 2b), closest to Mco5 of P. chrysosporium and Lac1 of Auricularia polytricha.
The amino acid residues that act as Cu2+ ligands are highly conserved in MCOs (Messerschmidt & Huber, 1990). All expected ligands were numbered (Table 2) on the basis of whether they coordinated with the type 1, type 2, or type 3 Cu2+ centers. The 11 predicted L. bicolor enzymes contained the 10 conserved histidines and the one conserved cysteine of the copper-binding centers of MCOs, except in the T3 center in the L3 region for Lcc11 where a histidine (H) was exchanged with a proline (P). The predicted sequences of Lcc1 to Lcc9 perfectly matched the laccase signatures L1–L4.
Table 2. Comparison of the laccase signature sequences of Laccaria bicolor lcc1 to lcc11
The comparison of the C. cinerea Lcc1 sequence with those of L. bicolor laccases allowed us to evaluate putative interactions of the enzymes with their different organic substrates (Table 3). The aspartate (D) residue in the first substrate loop, the N-terminal region cysteine (C) residue, and amino acid residues in loops III and IV and in the β-hairpin loop C7–C8 are likely to be involved in ligand binding. Laccaria bicolor laccases only contained some of these residues, suggesting weaker ligand interactions (Table 3; Bertrand et al., 2002; Hakulinen et al., 2002; Larrondo et al., 2004).
Table 3. Sequence alignment of the potential substrate-binding loops of the Laccaria bicolor laccases Lcc1 to Lcc11, the laccase multi-copper oxidase 1 (Mco1) of Phanerochaete chrysosporium (Pc Mco1), the laccase Lcc1 of Coprinopsis cinerea (Cc Lcc1), the laccase LccIIIb of Trametes versicolor (Tv LccIIIb) and the laccase of Melanocarpus albomyces (Ma Lcc)
Putative mature amino acid sequences of the laccases sensu stricto had between three predicted N-glycosylation sites (N-X-S/T) for Lcc4 and Lcc9 and 11 for Lcc3 (Table S3). The putative ferroxidases Lcc10 and Lcc11 had 11 and 10 predicted N-glycosylation sites, respectively.
Gene expression of MCOs
By analyzing the position of L. bicolor ESTs (from fruiting bodies and free-living mycelium, available at INRA LaccariaDB) on scaffolds at the JGI website, we were able to identify ESTs corresponding to the annotated lcc3, lcc6, lcc9 and lcc10 genes among the ∼220 000 ESTs from L. bicolor S238N, confirming the expression of several MCO genes. RT-PCR amplification and further sequencing of full-length MCO cDNAs from RNA extracted from L. bicolor free-living mycelium confirmed the exon/intron structure and the expression of all curated MCO gene models. Expression of lcc genes of L. bicolor was also detected in free-living mycelium, ectomycorrhizas and fruiting bodies using custom whole-genome expression oligoarrays and qPCR (Table 4). Transcripts for lcc3 and lcc8 were very highly expressed in L. bicolor/Douglas-fir and L. bicolor/Populus trichocarpa ectomycorrhizas, whereas lcc7 was highly expressed in fruiting bodies. Lcc2 was highly expressed in ectomycorrhizal root tips of Douglas-fir. However, lcc9 and lcc10 were mostly expressed in the free-living mycelium grown on glucose-rich agar medium. The expression level of lcc11 was similar in mycelium, ectomycorrhizas, and fruiting bodies. The iron permease gene (homologous to S. cerevisiae ftr1; data not shown), located close to lcc10, was mostly expressed in the free-living mycelium grown on glucose-rich agar medium. The expression of the iron permease gene (data not shown), located close to lcc11, was similar in mycelium, ectomycorrhizas, and fruiting bodies.
Eleven full-length and three truncated (pseudogene) sequences coding for MCOs were identified in the draft genome of L. bicolor. According to the classification suggested by Hoegger et al. (2006), these enzymes belong to the laccases sensu stricto (lcc1 to lcc9) and the ferroxidases (lcc10 and lcc11). This number of MCO genes is higher than those of P. chrysosporium (five MCOs; Larrondo et al., 2004), Cryptococcus neoformans (six MCOs), and U. maydis (six MCOs; Hoegger et al., 2006), but lower than that of C. cinerea (17 MCOs; Kilaru et al., 2006) (Table 1). The genes encoding laccases sensu stricto are randomly distributed throughout the L. bicolor genome, except for the clustered lcc6, lcc7 and lcc8 genes. This organization contrasts with that of C. cinerea, where the 17 laccases are clustered at seven different loci in the genome (Kilaru et al., 2006), and with that of P. chrysosporium, where four MCO genes belong to the same small genomic region (Larrondo et al., 2004).
The structures of 11 MCO genes were inferred from the ab initio annotation, manual curation, and comparison of ESTs and full-length cDNA sequences with curated gene models. The number of introns in lcc sequences ranged between nine and 14 and is thus similar to that for other basidiomycete genes, such as the laccase genes from C. cinerea (seven to 18; Kilaru et al., 2006) and A. bisporus (14; Perry et al., 1993) and the MCO genes from P. chrysosporium (14 to 19; Larrondo et al., 2004).
The two L. bicolor ferroxidase genes are located within less than 400 bp of an iron permease gene (homologous to S. cerevisiae ftr1), with start-to-start codons in a mirrored tandem arrangement (Fig. 1). The fet3/ftr1-based transport system, required for high-affinity iron uptake has been described in the white-rot fungus P. chrysosporium (Larrondo et al., 2007). Such an arrangement suggests a common regulation and function of the genes in iron metabolism as proposed for the fet3/ftr1 homologs of S. pombe, P. chrysosporium and C. neoformans (Askwith & Kaplan, 1997; Larrondo et al., 2007; Jung et al., 2008) and shown to be present in many other filamentous fungi such as A. fumigatus and U. maydis (Fig. 2b, Schrettl et al., 2004; Hoegger et al., 2006).
Predicted proteins and putative enzymatic activity
The predicted sequence of Lcc11 is lacking the essential histidine (H) residue of the tri-nuclear copper center T2/T3, as defined by Kumar et al. (2003) for fungal laccases. Proteins with amino acids other than histidine, acting as Cu2+ ligands, are not likely to have optimal capabilities in electron transfer and in oxidation of various substrates (Jeuken et al., 2000). In Lcc10 and Lcc11, the lack of a cysteine (C) in the signature sequence L2 is shared with fungal ferroxidases, including Mco1 of P. chrysosporium, which has a strong ferroxidase and only a low laccase activity (Kumar et al., 2003). The amino acid residue located 10 amino acids downstream of the conserved cysteine in the L4 domain has an important effect on the redox potential of type 1 copper at the active site (Canters & Gilardi, 1993). Based on the substitution of this residue, laccases are classified into three types: class 1 (M, Met), 2 (L, Leu), and 3 (F, Phe) (Eggert et al., 1998). The L (Leu)-E (Glu)-A (Ala) triad at positions +6 to +8 downstream of the conserved cysteine is also considered to be important for a high redox potential (Xu et al., 1998). Laccases from L. bicolor may therefore be categorized in class 2, as each of these genes has an L (Leu) residue, except for Lcc5 (G). Moreover, the L (Leu)-E (Glu)-A (Ala) triad does not exist, and consequently laccases from L. bicolor seem to have a medium redox potential.
All 11 MCO proteins of L. bicolor have substrate-binding loops (Table 3) described in the three-dimensional structure analysis of crystallized laccases (Hakulinen et al., 2002; Larrondo et al., 2003). By comparison with the structure of MCO loops involved in substrate binding of Trametes versicolor, C. cinerea and Melanocarpus albomyces, some amino acids that are involved in contact with organic substrates are lacking in L. bicolor laccases, suggesting a low substrate specificity or that the L. bicolor laccases interact with different substrates than those of saprotrophic fungi. This is also suggested by the differential clustering of the basidiomycete laccases, which may at least partially reflect functional differences (Hoegger et al., 2006), and therefore suggests variable functions for the different groups. None of the L. bicolor sequences clustered with laccases of white-rot fungi which have been hypothesized to be specifically involved in wood degradation (Hoegger et al., 2006).
In the 5′-noncoding upstream region of fungal MCOs, several putative cis-acting elements known to modulate gene expression were reported. A DNA-binding transcriptional factor CreA of the Cys2-His2 zinc finger class involved in glucose repression was characterized in Aspergillus nidulans (Kulmburg et al., 1993he different L. bicolor lcc genes, many putative CreA-binding sites (ZYGGRG) were located in the 5′-upstream region and may explain their low expression in mycelium growing on high-glucose medium. However, the in vivo contribution of these putative regulatory elements to glucose repression remains to be demonstrated. Several putative MREs were also found in the 5′-noncoding upstream regions of the L. bicolor lcc genes. These putative MREs are identical to the consensus sequence (TGCRCNC) found in the promoters of metallothionein genes in higher eukaryotes (Varshney et al., 1986). It has been shown that protein factors can bind MREs of the laccase promoters from P. ostreatus only when copper is absent (Faraco et al., 2003). The location of the putative TATA element in this promoter region is similar to the location found in several other fungal genes, in which the TATA box is generally located 30 to 60 nucleotides upstream from the transcriptional start site (Hong et al., 2007). Paired TATA and CAAT elements have been identified in other fungal laccase promoters such as those of T. versicolor (Jönsson et al., 1995) and P. ostreatus (Giardina et al., 1995). TATA boxes are also present far from the initial translation site in Trametes sp. AH28-2 (Xiao et al., 2006). Although the distribution of these motifs is conserved, their absolute positions vary.
Speciation and evolutionary process
The phylogenetic analysis clearly supports the presence of multiple MCO genes in the ancestral basidiomycete species, predating the C. cinerea/L. bicolor split, that have been maintained during the evolutionary process. As in other fungal MCO multigene families, several duplication events before and/or after the C. cinerea/L. bicolor divergence seem to be responsible for the organization of this family (Kilaru et al., 2006). The clustering of L. bicolor genes lcc6 to lcc8, separated by less than 1 kb, strongly supports such events. Transposable elements were not found in this region, precluding the involvement of retrotransposition in the duplication event. Codon alignments and analysis of synonymous substitution rates (ps) indicated more recent duplication events for the group of lcc1, lcc2, lcc3, lcc5 and lcc9. The high numbers of shared intron positions (11 to 13) support their common origin. The latest version of ARACHNE (v.2.0) allowed us to determine the position of MCO genes on the pseudochromosomes (Labbéet al., 2008). It is interesting to note that these genes are found to be located at different positions in the L. bicolor genome, whereas genes from more recent duplications in C. cinerea are found in close proximity to each other (Kilaru et al., 2006).
Furthermore, stretches of 115–205 bp of noncoding sequences between, as well as upstream and downstream of, the gene pairs encoding ferroxidases and their corresponding ftr1 homologs are highly conserved (Fig. 1), supporting a recent duplication event similar to that for the group of lcc1, lcc2, lcc3, lcc5 and lcc9. Although lcc7 and lcc8 are clustered and display a high degree of homology, they present a very dissimilar expression pattern. The representation of putative promoter elements is also very different among them, which could explain their opposite regulation. Moreover, three nontranscribed truncated gene models (with a homology to the laccase sequence of < 1 kb) were identified (data not shown). Approximately 800 tandem duplications were identified in the L. bicolor genome (Martin et al., 2008). Gene duplication leading to expansion of gene families is a major process shaping the L. bicolor genome (Martin et al., 2008) and is extensive in this species compared with other sequenced basidiomycetes. The presence of three extra truncated genes with homology to laccase in the genome of L. bicolor (not shown) and one extra truncated laccase gene in C. cinerea (Kilaru et al., 2006), suggests, however, that some of the duplicated gene sequences might, however, be eliminated from the genome over time.
Laccaria bicolor MCOs were expressed at different levels at the developmental stages investigated (Table 4). Accumulation of lcc7 transcripts was high in fruiting bodies, suggesting that this laccase is involved in the formation of these aerial aggregated tissues. This is consistent with the findings of Vnenchak & Schwalb (1989), who reported laccase activity in maturing fruiting bodies in C. cinerea and in A. bisporus, respectively, and transcript analysis in C. cinerea performed by M. Navarro-Gonzaléz (unpublished). lcc3 and lcc8 were highly expressed in differentiated Douglas-fir and P. trichocarpa ectomycorrhizas, suggesting a role for these MCOs in the functioning of fungal symbiotic tissues. The expression of lcc11 was similar in all developmental stages investigated, suggesting a nonspecific tissue expression. As lcc9 and lcc10 were highly expressed in the free-living mycelium grown on glucose-rich medium, their expression is not regulated by glucose. They may have a role in nutrient mobilization from soil-derived components by ECM extraradical mycelium. ECM fungi, such as L. bicolor, dominate in decomposed humus and litter, where they mainly mobilize nitrogen during litter decomposition. However, in light of the low genetic potential of L. bicolor for the production of enzymes degrading plant cell wall polysaccharides (Martin et al., 2008), it seems highly unlikely that L. bicolor is able to degrade plant organic matter, or able to support a major part of its carbon metabolism through saprotrophic activities. Therefore, lcc9 and lcc10 of L. bicolor may be involved in releasing aromatic amines locked up within humus complexes. Laccaria bicolor is unable to mobilize carbon from plant cell wall residues accumulating in the organic matter. Microcosm experiments in which transcripts for MCOs could be measured in extraradical mycelium colonizing litter will confirm this allegation. Saprotrophic fungi with the capability to modify litter through the secretion of lignocellulose-degrading enzymes (Lindahl et al., 2007; Baldrian & Valaskova, 2008; Snajdr et al., 2008) could release substrates available to L. bicolor.
To date, ferroxidase genes have been found in all sequenced basidiomycetes except C. cinerea (Table 1). Instead, this saprotrophic fungus has a sid1/sidA gene involved in the biosynthesis of an iron-chelating siderophore (Hoegger et al., 2006) involved in iron acquisition under conditions of low soil iron availability (Haas, 2003). By contrast, L. bicolor, like P. chrysosporium and C. neoformans, has no sid1/sidA gene involved in siderophore biosynthesis, whereas U. maydis possesses both systems for high-affinity iron uptake (Hoegger et al., 2006; Eichhorn et al., 2006; this study). Therefore, L. bicolor depends on the enzymatic Fet3-type pathway (fet3-like ferroxidase and ftr1 iron permease) to regulate its iron requirements (Kosman, 2003). The transcripts coding for these iron acquisition systems were down-regulated in ECM root tips compared with mycelium growing on glucose-rich medium (Table 4). These results suggested that high iron levels could lead to a decrease in expression levels (Eichhorn et al., 2006; Jung et al., 2008).
Our observations highlight the need for the identification of the functions, substrate specificity and regulation of L. bicolor laccases and ferroxidases; the study of temporal variations in laccase and ferroxidase gene expression in forest ecosystems on L. bicolor ectomycorrhiza associated with coniferous (i.e. Douglas-fir) or deciduous (i.e. P. trichocarpa) species (Martin, 2001); the measurement of ECM root-tip phenoloxidases (Courty et al., 2005, 2006, 2007), and the quantification of the production of enzymes by laccase antibodies in soil and plant tissues.
We would like to thank the Joint Genome Institute (US Department of Energy) and the Laccaria Genome Consortium for access to the L. bicolor genome sequence before publication. EST sequencing and transcriptome analysis were funded by the US Department of Energy, INRA ‘AIP Séquençage,’ the European network of excellence EVOLTREE and Région Lorraine grants. The first author was funded by a grant from the French Ministry of Ecology and Sustainable Development (Biological Invasions program). Research in Göttingen was supported by the Deutsche Bundesstiftung Umwelt (DBU). We also thank the IFR 110 for access to the DNA sequencing and functional genomics facilities, and Christine Delaruelle (INRA Nancy) for her technical assistance with DNA sequencing.