Introduction

Laccases (benzendiol:oxygen oxidoreductase, EC 1.10.3.2) are (blue) multi-copper enzymes that catalyze the oxidation of a range of aromatic compounds, such as mono-, di- and polyphenols, aromatic amines and methoxy-phenols, under reduction of oxygen to water (Messerschmidt 1997; Burke and Cairney 2002). They are widespread in wood-rotting fungi and have been repeatedly linked to lignin degradation. Alone or in concert with other enzymes, laccases can directly depolymerize the lignin macromolecule. Laccases catalyze the removal of an electron from phenolic hydroxyl groups of native lignin, thereby forming free phenoxy radicals. These lignin-derived radicals are further oxidized to quinones. Furthermore, laccases decarboxylate phenolic and methoxyphenolic acid structures of lignin and cause their demethylation or demethoxylation (Leonowicz et al. 2001).

Besides their prominent role in delignification, fungal laccases participate in various physiological processes. Laccases have been repeatedly linked to fruiting body formation in ascomycetes (Labarère and Bernet 1978; Hermann et al. 1983; Broxholme et al. 1991) and basidiomycetes (Wessels 1993; Kües and Liu 2000). In Schizophyllum commune, Agaricus bisporus and Lentinula edodes for example, extracellular enzymatic activities alter during the life cycle, in particular during fruiting body formation and maturation (Phillips and Leonard 1976; Leatham and Stahmann 1981; Ohga et al. 1999). One proposed laccase function in fruiting body development is mediating hyphal–hyphal aggregation by chemical cross-linking via oxidative polymerization of phenolic cell wall constituents (Broxholme et al. 1991; Zhao and Kwan 1999). Some fungal laccases take part in the synthesis of catechol- and napthalenediol-melanins (Langfelder et al. 2003; Nosanchuk and Casadevall 2003) and possibly in pigment production in mushroom tissues and basidiospores (Leatham and Stahmann 1981; Vnenchak and Schwalb 1989). In addition, respiratory enzymes may be inhibited by laccase-dependent oxidation of phenolic compounds during sporulation (Weaver et al. 1970).

In plant and animal pathogens, the production of melanins helps during infection by stabilizing cell walls and cell turgor and mediating protection against adverse environmental conditions and enzymatic degradation (Langfelder et al. 2003; Nosanchuk and Casadevall 2003). Laccases also contribute to pigmentation and environmental protection of non-virulent fungi, as shown for asexual conidia in Aspergillus species (Clutterbuck 1972; Tsai et al. 1999). In Podospora anserina, female-sterile long-living mutants are laccase-negative. Under normal senescent conditions, laccase activity may act in defeating oxidative stress and detoxification of phenolic compounds (Marbach et al. 1994; Fernández-Larrea and Stahl 1996). Laccase defense reactions (protection against host oxidative responses, detoxification of adverse phenolic compounds, pigmentation) are also known in wood-decay fungi during interactions with living trees (Burke and Cairney 2002) and between interacting mycelia of different organisms (Iakovlev and Stenlid 2000). In ecto- and ericoid mycorrhiza, laccases are postulated to act protectively against soil pollutants and host defense compounds, in lignin depolymerization and polyphenol degradation, in the release of nitrogen from insoluble protein–tannin complexes, in pigmentation and in humus formation (Burke and Cairney 2002).

More than 60 different laccases from fungi are known in sequence (Kumar et al. 2003). These enzymes have a typical structure, featuring four copper atoms located in two copper centers as part of their overall catalytic apparatus, being responsible for the electron transfer during redox reactions. The mononuclear center (T1) with one copper atom (type-1 Cu, blue, with a maximum absorbance at around 600 nm) serves as the primary electron acceptor from the substrate. Electrons are then transferred to the trinuclear cluster (T2/T3) composed of a binding site (T2) for one copper atom (type-2 Cu) with weak absorbance in the visible spectrum and a binding site (T3) for two coupled copper atoms (type-3 Cu) characterized by an absorbance at about 330 nm. The trinuclear cluster is the dioxygen-binding site and reduces molecular oxygen upon receipt of four electrons. Two histidines and one cysteine serve as ligands for type-1 Cu at the T1 center and eight histidines for binding of type-2 and type-3 Cu at the T2/T3 cluster (Messerschmidt 1997; Ducros et al. 1998; Hakulinen et al. 2002; Piontek et al. 2002). The histidines and the cysteine are spread over four highly conserved amino acid regions. These regions are considered fungal laccase signature sequences (L1–L4) and include not only residues involved in copper binding, but also non-copper-ligating residues responsible for conformational functions (Kumar et al. 2003).

The high sequence conservation of the laccase signature sequences has been exploited to design DNA primers or in hybridization to clone one or more laccase genes from a variety of organisms. A gene family of five non-allelic laccase genes has been identified in the white rotting fungi Trametes villosa (Yaver and Golightly 1996; Yaver et al. 1996) and Pleurotus sajor-caju (Soden and Dobson 2001; GenBank CAD45381). Four laccase genes are known in P. ostreatus (Giardina et al. 1995, 1996, 1999; GenBank AJ344434), three in the plant pathogen Rhizoctonia solani (Wahleithner et al. 1996), the white-rots Pycnoporus cinnabarinus (Eggert et al. 1998; Temp et al. 1999; Otterbein et al. 2000), Trametes I-62 (Mansur et al. 1997) and L. edodes (GenBank BAB84354, BAB84355, and BAB84356) and one in the white-rot S. commune (Hatamoto et al. 1999) and in the saprophyte Coprinellus (Coprinus) congregatus (Kim et al. 2001). In Coprinopsis cinerea (Coprinus cinereus), three complete genes (lcc1, lcc2, lcc3) were described by Yaver et al. (1999). In addition, Bottoli et al. (1999) published partial sequences from three genes that appeared to be distinct from genes lcc1, lcc2 and lcc3 (Kües 2000). In this study, we show that the saprophyte Coprinopsis cinerea has at least eight different laccase genes within the haploid genome. This presents the largest laccase gene family reported so far from a single fungus.

Materials and methods

Strains and culture conditions

Homokaryon AmutBmut (A43mut, B43mut, pab-1) and monokaryons AT8 (A43, B43, trp-3, ade-8), FA2222 (A5 B6, acu-1, trp-1.1;1.6), LT2 (A6, B6, trp-1.1;1.6), PG78 (A6, B42, pab-1, trp-1.1;1.6), LN118 (A42, B42, ade-2, trp-1.1;1.6) and 218 (A3, B1, trp-1.1;1.6; Kertesz-Chaloupková et al. 1998) were cultivated at 37 °C in liquid YMG/T medium (Granado et al. 1997) to isolate genomic DNA by the method of Zolan and Pukkila (1986). Pools of an indexed genomic library of homokaryon AmutBmut in Escherichia coli NM554 and single cosmids from this library were grown and harvested as described by Bottoli et al. (1999). E. coli XL1-Blue and DH5α (Stratagene) were used as hosts in routine cloning.

DNA techniques

Pools of cosmid colonies or single clones of the library were screened for laccase genes by whole-cell PCR, using degenerate or specific primers (Table 1) in a total volume of 25 µl containing 10 mM Tris, pH 8.8, 50 mM KCl, 0.1% Triton X-100, 1.5 mM MgCl2, 0.2 mM dNTPs, 0.4 µM each primer (4 µM for degenerate primers), 1 unit of Taq polymerase and approximately 5 ng of DNA or 1 µl of whole bacterial cells from plate-washes. PCR conditions were 2 min (PCR with isolated DNA) or 5 min (whole-cell PCR) initial denaturation at 94 °C, followed by 35 cycles of 30 s at 94 °C, 30 s at an annealing temperature specific to a primer set (Table 1), a primer set-specific extension time (Table 1) at 72 °C and a final extension step at 72 °C for 10 min.

Table 1 Primer sets and PCR conditions
Full size table

Isolation of cosmid DNA, DNA restriction analysis and Southern blotting were performed using standard methods (Bottoli et al. 1999). Hybridization probes were digoxigenin-labeled and detected by chemiluminescence, using the respective kits from Roche (Mannheim, Germany). Hybridization probes for lcc4, lcc5 and lcc6 were labeled by PCR, using primer sets lcc4.1, lcc5.1 and lcc6.1 (Table 1), respectively, on genomic DNA from strain AmutBmut. For the cosmids that were positive with degenerate primers (Table 1) but negative with the specific primers for lcc4, lcc5 and lcc6, we used the degenerate primers on selected cosmids (2-3E, 2-5H, 7-2G, 35-4B) to generate specific hybridization probes. Fragments of cosmid DNA containing full-length laccase genes were subcloned into standard vectors (Stratagene, Amsterdam, The Netherlands). pBluescript KS+ constructs were: (1) pLcc5AmutBmut carrying an 8-kb PstI fragment from cosmid 30-10A with lcc5, (2) pLcc7AmutBmut carrying a 12-kb SacI fragment from cosmid 7-2G with lcc7 and (3) pLcc8AmutBmut carrying a 4-kb EcoRI fragment from cosmid 35-4B with lcc8. pBC SK+ constructs were: (1) pLcc4AmutBmut carrying a 13-kb EcoRI fragment from cosmid 22-12F with lcc4 and (2) pLcc6AmutBmut carrying a 9-kb XbaI fragment from cosmid 36-2B with lcc6. The 10-kb plasmid pE8.1 originated from a deletion event in cosmid 22-12F in E. coli that brought together in close vicinity the laccase gene lcc4 and the pab1 + marker gene present in the cosmid backbone (Bottoli et al. 1999).

For TA cloning into pBluescript KS+ (Marchuk et al. 1991), PCR products were amplified from genomic AmuBmut DNA with the specific primer sets lcc1.2, lcc1.3, lcc1.4, lcc2.1 and lcc3.1 (Table 1), to give pLcc1.2AmutBmut, pLcc1.3AmutBmut, pLcc1.4AmutBmut (all with a partial lcc1 gene), pLcc2AmutBmut (full-length lcc2, carrying a PCR artefact G-to-T at position 2,015) and pLcc3AmutBmut (full-length lcc3). A full-length copy of lcc1 in pLcc1AT8 from strain AT8 was amplified with primer set lcc1.1 (Table 1).

Sequencing of plasmid DNA was performed by the Göttingen Genomics Laboratory. Sequences were edited and assembled using the Staden package 2001 (http://www.mrc-lmb.cam.ac.uk/pubseq/). Alignments of DNA and amino acid sequences were generated with ClustalX (http://www-igbmc.u-strasbg.fr/BioInfo/ClustalX/Top.html) and manual adjustments were made in GeneDoc ver. 2.6.002 (http://www.psc.edu/biomed/genedoc/). Phylogenetic analysis of amino acid sequences were performed with Mega ver. 2.1 (http://www.megasoftware.net/). Signal peptides were predicted with SignalIP ver. 2.0 (http://www.cbs.dtu.dk/services/).

Results

Identification of laccase genes in C. cinerea homokaryon AmutBmut

Previously, a set of degenerate primers (Table 1) deduced from the laccase signature sequences of fungal laccases produced DNA fragments of an expected size of about 140 bp in half of the 60 cosmid pools from the indexed genomic library of homokaryon AmutBmut. Specific primers obtained from sequences of subcloned PCR fragments assigned three distinct laccase gene sequences (originally named lac1, lac2, lac3) to 16 of the positive pools (Bottoli et al. 1999; Table 2). In this study, we used the same sets of specific primers (lcc4.1, lcc5.1, lcc6.1; Table 1) to identify single cosmids in specific pools that bear a copy of one of the three laccase genes. Following subcloning of specific cosmid fragments, we established the laccase gene sequences. Two different EcoRI fragments from E. coli cosmid clone 22-12F contained the same complete laccase gene sequence—one of the fragments originated from a deletion event in E. coli (see Materials and methods). The nucleotide sequences of the three different AmutBmut laccase genes differed markedly from the lcc1, lcc2 and lcc3 of C. cinerea strain IFO 8371 (Yaver et al. 1999). They were deposited in GenBank under the names lcc4, lcc5 and lcc6 (accession numbers AY338760, AY338761, AY338762, respectively).

Table 2 Coprinopsis cinerea laccase genes identified in specific pools of the genomic library of strain AmutBmut
Full size table

No fragments were amplified from AmutBmut genomic DNA with the lcc1 primer set lcc1.1 (Table 1), designed from the outermost 5′ and 3′ ends of the published IFO 8371 lcc1 sequence. Primer sets lcc2.1 and lcc3.1 (Table 1) from the outermost 5′ and 3′ ends of the published lcc2 and lcc3 sequences from strain IFO 8371 produced PCR products from genomic DNA of homokaryon AmutBmut of expected sizes. We cloned and sequenced the 3-kb lcc2 and the 2.8-kb lcc3 PCR products of the AmutBmut genes (GenBank accession numbers AY338758, AY338759). Compared with strain IFO 8371, five nucleotide exchanges and a deletion of three adjacent nucleotides were detected in the lcc2 gene of strain AmutBmut and five nucleotide exchanges in the lcc3 gene.

In the pools of the AmutBmut genomic library that previously yielded signals in PCR with the degenerate primers but not with the primers specific for either lcc4, lcc5 or lcc6 (primer sets lcc4.1, lcc5.1, lcc6.1; Table 1; Bottoli et al. 1999), we identified 21 single cosmids by PCR with the degenerate primer set (Table 1). Comparative restriction digests and Southern blot analysis using PCR-amplified probes from selected cosmids (see Materials and methods) divided them into four groups [13 (from 12 different pools), four, three and one molecules, respectively]. Primer sets lcc2.1 and lcc3.1 (Table 1) identified gene lcc2 in the group of 12 cosmid pools and lcc3 in the group of four (Table 2). Restriction fragments from cosmids of the two groups containing so far unknown laccase genes (cosmid 7-2G from the group of one pool, cosmid 35-4B from the group of three pools, respectively; Table 2) were subcloned (see Materials and methods) and sequenced. These new laccase genes were named lcc7 and lcc8 (GenBank accession numbers AY338763, AY338764).

When we analyzed the 60 pools of the library with primers specific for the seven different laccase genes (primer sets lcc2.2, lcc3.2, lcc4.2, lcc5.2, lcc6.2, lcc7.1, lcc8.1; Table 1), we obtained signals for an additional three pools with the lcc2 primers, for eight extra pools with the lcc3 primers, for nine additional pools with the lcc8 primers and none with the others (Table 2). Amongst those 20 extra pools, there were seven (pools 3, 34, 42, 58, 60 with lcc3, pools 16, 48, 60 with lcc8) which have not been identified by Bottoli et al. (1999) with the degenerate primer set (Table 1). As expected from PCR with isolated genomic DNA, primer set lcc1.1 from gene lcc1 of strain IFO 8371 (Table 1) did not yield any DNA amplification with any of the 60 pools of the AmutBmut genomic library.

Identification of gene lcc1 in a collection of Coprinopsis strains

A screen of different C. cinerea strains with lcc1 primer set 1 (Table 1) resulted in successful amplification of the lcc1 alleles from genomic DNA from monokaryons FA2222, AT8, LT2 and PG78. Like AmutBmut DNA, genomic DNA from monokaryons LN118 and 218 did not yield any PCR products with these primers. Due to the lack of an lcc1 sequence from homokaryon AmutBmut, we chose to analyze the lcc1 gene from strain AT8 by cloning and sequencing the 2.4-kb lcc1 PCR product (GenBank accession number AY338756). The AT8 sequence differed from strain IFO 8371 at 43 positions (42 nucleotide exchanges, one extra nucleotide in an intron of the AT8 gene).

We then tested different combinations of primers obtained from sequencing the AT8 lcc1 gene (Table 2) in PCR against AmutBmut DNA. A PCR product of 0.9 kb was obtained with primer set lcc1.2, a 1.3-kb product with primer set lcc1.3 and a 1.7-kb product with primer set lcc1.4 from isolated genomic AmutBmut DNA, but unfortunately not from any of the 60 pools of the cosmid library. Products of the same sizes were also amplified from genomic DNA of strains LN118 and 218, strongly suggesting gene lcc1 to be present in homokaryon AmutBmut as well as in monokaryons LN118 and 218.

The three PCR products from strain AmutBmut were subcloned and sequenced. The 1,670-bp partial lcc1 gene sequence obtained was deposited in GenBank under accession number AY338757. Comparison with the lcc1 sequences from strains AT8 and IFO 8371 revealed high nucleotide identity (90 mismatches with AT8, 75 mismatches with IFO 8371) and high amino acid identity (three changes over the 446-amino acid sequence available from strain AmutBmut, see below). We conclude therefore that there are at least eight different laccase genes within a haploid genome of C. cinerea.

Intron structure

The positions of the ATG start codons and the introns in the nucleotide sequences of all eight genes cloned in this work were deduced by alignment with the cDNA sequences of lcc1, lcc2 and lcc3 of strain IFO 8371 (Yaver et al. 1999) and by comparison with the conserved sequences of C. cinerea intron splice junctions (Seitz et al. 1996). We identified a total of 84 introns in the eight cloned laccase genes. However, the actual number of introns in the individual genes varies between seven and 13 (Fig. 1). The lengths of the introns are typical for C. cinerea (Seitz et al. 1996) and range between 47 nucleotides and 87 nucleotides, with an average base number of 59±4 nucleotides. The consensus sequence for the 5′ splice junction of the 84 introns is G/C81%g100%t99%a/g94%a/c80%g93%t/c88% and for the 3′ junction t/c97%a100%g100%G/C73% (exon bases shown in upper case, intron bases in lower case), resembling the consensus sequences defined by Seitz et al. (1996). At least one internal CTNA consensus sequence for lariat formation (Seitz et al. 1996) is also found in all but nine of the introns. As with other C. cinerea genes (Seitz et al. 1996), there is no preference in the eight laccase genes for intron insertion at codon phase 0, 1 or 2.

Fig. 1
figure 1

Intron positions within the eight laccase genes (lcc1–lcc8) of Coprinopsis cinerea. Horizontal lines indicate the laccase genes, vertical bars indicate introns. Dotted lines mark introns that interrupt the coding sequence of the different genes at exactly the same codon position

Full size image

When comparing the eight genes over their whole length, introns interrupt the coding sequences at 16 different positions (Fig. 1). The 13 introns found in lcc2 and lcc3 are all at the same positions. All other genes differ from these and each other by the gain or loss of one or more introns at variable positions. Only four of the intron positions (positions 2, 4, 5, 13) are conserved in all eight laccase genes. At 15 of the total of 16 positions, two or more genes have an intron (Fig. 1). Only the first intron predicted in lcc8 is unusual, residing at a unique position eight nucleotides behind the ATG start codon (Fig. 2). If not a true intron, gene lcc8 would translate from the only possible 5′ in-frame localized ATG triplet into a protein product with an unusual 161 extra amino acids at the N-terminus, mainly consisting of serine, threonine, aspartic acid, glycine and alanine residues.

Fig. 2
figure 2

Alignment of the deduced amino acid sequences of the C. cinerea laccases. Lcc1 is from monokaryon AT8; all other aligned sequences are from homokaryon AmutBmut. Shown below the alignment in circles are the amino acids in Lcc3 of IFO 8371 (Y158, N164) that differ from the sequence of AmutBmut and amino acids in the available Lcc1 sequence from AmutBmut (R243, V319, V383) that differ from those in the sequences from IFO 8371 and AT8. Black triangles above the alignments mark amino acids translated from codons interrupted by introns in phase 0 (arrow between two amino acids) or in phases 1 or 2 (arrow above amino acid). Numbers refer to introns shown in Fig. 1. Signal peptides are underlined. The cysteine and the histidine residues involved in copper binding are indicated by white boxes and numbered according to the copper type (1, 2, 3 for type-1, type-2, type-3, respectively) they bind. The laccase signature sequences L1–L4 (Kumar et al. 2003) are indicated by gray boxes

Full size image

Sequence comparison of all 84 introns by ClustalX alignment showed higher identity among some of the introns within a gene (e.g. lcc2 introns 4, 8, 16 have intron identities of 32–51%, lcc5 introns 4, 7, 14 have identities of 35–49%), compared with introns present at the same position in the different genes (e.g. the four introns present in all genes). Most striking are the introns at position 4, which vary considerably in sequence (14–33% identity) and in size (53–87 bp). Thus, evolutionary force conserved the intron positions but not their nucleotide sequence. In an overview of all introns, sequence identity between introns is in the range 6–63%.

The laccase proteins

The deduced amino acid sequences of the eight complete laccase genes isolated in this study vary in length over 516–567 amino acids (Fig. 2) in a range typical for fungal laccases (Yaver et al. 1999). Over their whole length, these proteins show sections of high amino acid conservation interrupted by short stretches of more variable sequences (Fig. 2, Table 3). The amino acid sequences of Lcc1 from strain AT8 and Lcc2 from strain AmutBmut (Fig. 2) are identical to the corresponding sequences from strain IFO 8371 (Yaver et al. 1999). The three amino acid changes detected in the available AmutBmut Lcc1 sequence (Q243R, A319V, A383V) are all at positions that are less conserved between the different C. cinerea laccases (Fig. 2). In contrast, the two amino acid differences found in Lcc3 from AmutBmut compared with the IFO 8371 sequence are at conserved positions but of comparable nature (Y158F, N164D; Fig. 2).

Table 3 Identities and similarities (%) of the amino acid sequences of the eight laccases from C. cinerea
Full size table

All eight C. cinerea laccases have the conserved cysteine and the ten conserved histidines which serve as copper ligands (Ducros et al. 1998; Piontek et al. 2002) spread over the laccase signature sequences (Fig. 2). We also found perfect matches to the laccase signature sequences L1–L4 recently proposed by Kumar et al. (2003). Consistent with an extracellular function established for other well characterized fungal laccases, including C. cinerea Lcc1 (Heinzkill et al. 1998, Yaver et al. 1999), signal peptides were predicted for seven of the eight proteins, with the neural networks and the hidden Markov models in SignalIP ver. 2.0 giving high probabilities (P>0.79). The lengths of the signal peptides range over 16–22 amino acids (Fig. 2). For Lcc8, only the hidden Markov model predicted a signal peptide with a lower probability (P=0.41; Fig. 2).

Identities between the eight proteins range from 46% to 77% and similarities from 57% to 83%, with Lcc2 and Lcc3 being the most similar (Table 3). When compared with laccases from C. congregatus, T. villosa, P. ostreatus, P. cinnabarinus and S. commune, the C. cinerea laccases Lcc2 and Lcc3 show higher similarities to laccases from other species than to the other six proteins from the same species. In the neighbor-joining tree, Lcc2 and Lcc3 cluster with CLAC2 from C. congregatus (Fig. 3). Lcc6 and Lcc7 represent another pair of C. cinerea laccases with higher identity and similarity to each other (71% identity, 79% similarity) than to any other of the laccases (Table 3, Fig. 3). Lcc1 shows highest similarity towards the Lcc6/Lcc7 cluster and to Lcc4 (Table 3, Fig. 3). Lcc5 and Lcc8 of C. cinerea seem to be about as close to the other C. cinerea laccases as to enzymes from the other species (Fig. 3).

Fig. 3
figure 3

Neighbor-joining tree of the deduced amino acid sequences of the eight C. cinerea (Cci) laccases and selected laccases of other basidiomycetes. The tree is calculated with p-distances using Mega ver. 2.1, based on a ClustalX alignment. Bootstrap values (500 replications) higher than 70% are indicated at branchings. The scale bar indicates a distance equivalent to 0.05 amino acid substitutions per site. Species and strains (with GenBank accession numbers) are: Cco Coprinellus congregatus CLAC2 (CAB69046), Pci Pycnoporus cinnabarinus Lac1 (AAG13724), Lcc3-1 (AAC39469), Lcc3-2 (AAD49218), Pos Pleurotus ostreatus POXA1b (CAA06292), POXC (Q12739), POX1 (Q12729), POXA3 (CAC69853), Sco Schizophyllum commune Lac (BAA31217), Rso Rhizoctonia solani Lcc1 (P56193), Lcc2 (S68118), Lcc3 (Q02079), Tvi Trametes villosa LCC1 (Q99044), LCC2 (Q99046), LCC3 (JC5355), LCC4 (Q99055), LCC5 (Q99056)

Full size image

Promoter structure

Promoter sequences of the eight AmutBmut genes are relatively little conserved. However, alignment of the 5′ upstream gene sequences revealed two subgroups of more closely related promoters. lcc2, lcc3 and lcc4 form one group (sequence identities: lcc2/lcc3 37%, lcc2/lcc4 26%, lcc3/lcc4 29%) and lcc5, lcc6 and lcc7 form another (sequence identities: lcc5/lcc6 36%, lcc5/lcc7 15%, lcc6/lcc7 25%). From lcc1, we did not have enough 5′ sequence for an extensive promoter analysis. However, the 106 bp available upstream to the ATG start codon match to the first group with 29–39% identity to the three other promoter sequences. In contrast, the upstream sequence from gene lcc8 did not reveal an obvious relationship to any of the other promoter sequences (data not shown).

DNA motifs conserved in all promoter sequences are not evident. Putative TATA elements, for example, are only found in lcc2, lcc3 and lcc4. One to three potential metal-response elements shown in P. ostreatus to function in copper regulation of laccase gene transcription (consensus TGCRNNC; Faraco et al. 2003) are found scattered over the different promoter sequences at non-conserved positions (for lcc2, lcc3, lcc4).

Discussion

In this study, we identified eight different laccase genes within the haploid genome of the C. cinerea homokaryon AmutBmut. This represents the largest laccase gene family known so far in any fungus. Up to five different laccase genes have been identified in other basidiomycetes (Yaver and Golightly 1996; Yaver et al. 1996; Soden and Dobson 2001; GenBank CAD45381), but in some instances when using dikaryotic strains, isolated genes have been shown to be alleles (Wahleithner et al. 1996). In most cases, degenerate primers were designed to isolate the genes, making use of the laccase signature sequences.

Yaver et al. (1999) and Bottoli et al. (1999) used different degenerate primers to detect laccase genes in C. cinerea. The forward primers, albeit somewhat different in sequence, were deduced from the same protein region, whereas the reverse primers originated from different regions. Because of two mismatches within the last six nucleotides at the 3′ end, the forward primer 5′-ATICAYTGGCAYGGIYTIYTI-3′ of Yaver et al. (1999) were unlikely to recognize our gene lcc8. In all other cases, the primer combination of Yaver et al. (1999) had either no mismatch or one at nucleotide 6, counting from the 3′ end (lcc2, lcc5), which was tolerated at least in the isolation of gene lcc2. Looking at the primer combination of Bottoli et al. (1999), genes lcc3 and lcc8 might not always efficiently be recognized by their forward primer (Table 1). In lcc3, there is one mismatch 3 nucleotides from the 3′ end of the primer and lcc8 has two mismatches at 4 and 8 nucleotides from the 3′ end, respectively. These mismatches may account for the fact that we found seven extra pools containing a copy of lcc3 or lcc8 that formerly were overlooked in the analysis by Bottoli et al. (1999). Gene lcc1 was not detected in the AmutBmut genomic library. This was not due to a failure in sequencing the degenerate primers but to a complete lack of lcc1 sequences in the library. Previous analysis of nine different genes suggested the 6-fold genomic AmutBmut library to be complete (Bottoli et al. 1999). The variable appearance of laccase genes in the library (ranging from 0 to 15 positive pools; Table 2), however, may indicate a preferential tolerance in E. coli of some C. cinerea fragments compared with others. Analysis of cosmid 22-12F with gene lcc4 showed that a genomic fragment was eliminated by deletion in E. coli, supporting this suggestion.

The presence of at least eight different laccase genes in a single organism raises the question why there are so many. From cDNA cloning, it is clear that three of the eight different laccase genes from C. cinerea (lcc1, lcc2, lcc3) are expressed within the vegetative mycelium when grown in rich FG4 medium (Yaver et al. 1999). It is noteworthy that the three genes known to be expressed in the vegetative mycelium form a group with more closely related promoter sequences, together with lcc4. Enzymatic activity was demonstrated in FG4 medium but only one enzyme (Lcc1) was detected and purified from the culture broth (Schneider et al. 1999). Under inducing conditions (copper, phenolic lignin composites, other aromatic compounds, such as 2,5-xylidine, coniferyl alcohol, ferulic acid), other basidiomycetes usually produce several laccase isoenzymes (Leonowicz and Trojanowski 1975; Palmieri et al. 2000; Soden and Dobson 2001). Addition of copper to the growth medium and confrontation of C. cinerea homokaryon AmutBmut with straw and wood induces laccase activity, but the enzyme(s) identity remains to be shown (M. Navarro-Gonzalez, unpublished data). Potential metal response elements (Faraco et al. 2003) were found in the promoter sequences of all eight genes.

In P. ostreatus, genes pox1 and poxc are expressed in potato/glucose/yeast extract broth (Giardina et al. 1995, 1996). The addition of copper markedly increases the transcription of poxc and induces POXA1b production (Giardina et al. 1999, Palmieri et al. 2000). In P. sajor-caju, lac3 appears to be constitutively expressed, whilst lac1, lac2 and lac4 are differentially regulated by nutrients and the presence of metallic and aromatic inducers (Soden and Dobson 2001). The three laccase genes from Trametes I-62 were shown to be differentially regulated by nutrients, growth phases and the inducer veratryl alcohol (Mansur et al. 1998). Differential expression patterns suggest different functions for the different laccases. Different enzyme properties support this assumption. POXC of P. ostreatus uses ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulphonic acid)], guaiacol and syringaldazine as substrates, but with individual pH optima. The enzyme is almost fully active in the temperature range 40–60 °C (Palmieri et al. 1993, Giardina et al. 1996). POXA1b is able to oxidize ABTS and syringaldazine at pH optima comparable with POXC, but acts at a lower temperature (20–50 °C) and does not react with guaiacol (Giardina et al. 1999). In the phylogenetic tree of laccases shown in Fig. 3, POXA1b and POXC do not cluster very closely together. Whether close clustering reflects similar enzyme properties is presently not known, since the products for most cloned genes are not or not well described. The deduced products of the expressed C. cinerea genes lcc2 and lcc3 cluster together with the acidic laccase CLAC2 from C, congregatus suggested to act in neutralization of acidic culture medium (Kim et al. 2001). The characterized enzyme Lcc1 of C. cinerea is active over a wide pH range (pH 4.0–10.0), with an optimum of pH 4.0 for ABTS and pH 6.5 for syringaldazine. It has a temperature optimum in the range 50–70 °C (Schneider et al. 1999).

In several species, laccase activity has been implicated in fruiting body formation, in mediating hyphal–hyphal aggregation and/or in tissue and spore pigmentation (see Introduction). In C. cinerea, Vnenchak and Schwalb (1989) reported laccase activity in maturing fruiting bodies at the stage when spores are stained by melanin. Other places of melanin production are the brown patches in the mycelial matting of aging cultures and the outer rind of the multicellular sclerotia acting as resting bodies under unfavorable environmental conditions (Kües et al. 2002). Our future studies will have to identify the functions of the eight laccase genes under vegetative growth conditions and in the different developmental processes. With so many highly similar genes, detection of expression will be easiest on the transcript level, using PCR with discriminative primers. Homokaryon AmutBmut with only one haploid genome is an ideal strain for such analysis. By mutation in both mating type loci, it is self-compatible and undergoes fruiting body and basidiospore formation (Walser et al. 2003). Its vegetative mycelium mimics a dikaryon with clamp cell production and light-regulated asexual sporulation (Kertesz-Chaloupková et al. 1998). A co-isogenic strain with wild-type mating type genes is available (P. Srivalai, unpublished data) to help the study of laccase gene expression in a normal monokaryon.