A dehydration-inducible gene in the truffle Tuber borchii identifies a novel group of dehydrins

Sequence analyses

The 771-bp-long M6G10 fragment [GenBank:DN601500] was one of the fruiting body-regulated Expressed Sequence Tag (EST) retrieved from a gene expression profiling study conducted in the ascomycetous truffle T. borchii [21]. A blastx search against NCBI (National Center for Biotechnology Information, NIH, Bethesda) databases [23] using as a query the M6G10 EST revealed the existence of a conserved, previously uncharacterised group of DHN proteins. This result was then confirmed by using the complete TbDHN1 mRNA sequence. Significant similarities (E-value < e-10) to other proteins (Table 1), including a plant dehydrin from Hordeum vulgare ([GenBank: AAD02257]; blastx E = 4e-25) (Fig. 1), were found only by masking off repeated or low complexity regions. Dehydrins, like the majority of LEA proteins, are low complexity proteins; in fact, applications masking low complexity regions, like SEG [24], masked between 30% and 71% of the amino acids of a LEA protein [12]. Since the main effect of masking low complexity regions is to reduce the number of amino acids available for alignment, blast searches were all performed with the filter for low complexity regions switched off.

Accession number	Biological role	organism	Identities (%)	E-value	Sequence found by
SNU00161	predicted protein	Stagonospora nodorum	41	6e-50	blastp
EAA70367	hypothetical protein FG10051	Gibberella zeae	40	4e-47	blastx
CAD70810	putative protein	Neurospora crassa	39	9e-44	blastx
EAA55454	hypothetical protein MG09261.4	Magnaporthe grisea	35	7e-39	blastx
EAA62484	hypothetical protein AN5324.2	Aspergillus nidulans	37	1e-37	blastx
EAL89059	conserved hypothetical protein	Aspergillus fumigatus	31	3e-35	blastx
EAL84332	conserved hypothetical protein	Aspergillus fumigatus	38	2e-27	blastx
EAK94850	DHN6*	Candida albicans	32	5e-21	blastx
EAK94909	DHN*	Candida albicans	39	8e-21	blastx
EAA54716	hypothetical protein MG05507.4	Magnaporthe grisea	28	1e-17	blastx
EAL87020	hypothetical protein Afu7g04520	Aspergillus fumigatus	28	2e-16	blastx
Q7S6B0	Predicted protein	Neurospora crassa	26	0.009	blastp

*Putative identifications provided by automatical annotation of C. albicans genome.

A blastp search on Stagonospora nodorum protein database characterized the predicted protein SNU00161 (The Broad Institute Accession Number) as another possible fungal homolog of TbDHN1. A tblastx search was performed querying all those genomic projects in which a proteomic annotation was not yet available. Other possible fungal homologs were thus identified in Botrytis cinerea [The Broad Institute Accession Number:00133947_B.cinerea_19866915837883], Fusarium verticillioides [The Broad Institute Accession Number:0031597_F.verticillioides_19866917223803], Chetomium globosum [GenBank:AAFU01000048], Coccidioides immitis [GenBank:AAEC01000142] and C. posadisii [TIGR Accession Number:gnl|TIGR_222929|contig:3250:c_posadasii]. A search in the EST databases showed that most of these fungal proteins are present as expressed sequences. Additionally, other DHN-like members were found within cDNA sequences of the fungi Verticillium dahliae [GenBank:BQ110173] and Trichoderma reseii [GenBank:CF872473], and the plants Tortula ruralis [GenBank:CN201388], Picea engelmannii x Picea sitchensis [GenBank:DR464575], Malus x domestica [GenBank:CV091950], Oryza sativa [GenBank:CA765427] and Saccharum officinarum [GenBank:CA105075]. For each different species we reported only one entry, because sequences from unfinished/unannotated genome projects and from EST databases are always redundant. Although blast searches were performed also on sequences from Basidiomycota and Glomeromycota, all the retrieved fungal sequences belong exclusively to ascomycetous fungi.

Only DHN-like proteins from T. borchii, Gibberella zeae, Neurospora crassa, Magnaporthe grisea, Aspergillus nidulans, A. fumigatus, S. nodorum and Candida albicans were further analysed. All the other potential fungal and plant DHN proteins were excluded because of poor sequence quality (especially for ESTs), i.e. high percentage of Ns or too short fragments, and because all tblastx searches were performed on an unfinished genome project.

On the basis of these physiochemical characteristics, TbDHN1 and its fungal homologs can be assigned to the large family of LEA.

Classification according to Wise's rules

The next step in the characterization of TbDHN1 and its homologs was their classification within the big family of LEA proteins. TbDHN1 and the other twelve fungal proteins were classified according to a set of rules defined by Wise [12]. Each LEA Class is defined by a range of percentages for all the physicochemical features which usually characterize LEA proteins: hydrophilicity, predicted secondary structure and amino acid composition. According to this rule set, G. zeae, A. nidulans, A. fumigatus and C. albicans proteins were placed into Class II; while the aromatic amino acid percentage of N. crassa and M. grisea proteins and the values of minimum hydrophobicity for T. borchii and S. nodorum proteins were out of the established ranges for this Class. It is important to point out that, especially for N. crassa, M. grisea and T. borchii proteins, values of aromatic amino acid percentage and minimum hydrophobicity are very close to the limit of Class II. The six fungal proteins not assigned to Class II clustered with Class IV, but Wise concluded that members of Class IV should be more appropriately housed in Class II and Class III. According to Wise's analysis of LEA amino acid composition, glycine is highly represented in Class II, while in Class III glycine is found only marginally more than expected by chance. Moreover, Class III LEA proteins have high helix content. On the basis of a high percentage of glycines in the six proteins and/or a lower percentage of helix content than the expected one for Class III, also proteins from N. crassa, M. grisea, T. borchii and S. nodorum can be assigned to Class II, the group of plant dehydrins.

Furthermore, blastp results obtained with TbDHN1 as query against the nr protein database shared no common hits with the ones obtained using as queries the fungal proteins already classified in Class III (Pfam PF02987).

A new common signature pattern

Members of plant DHN family are characterized by the presence of two PROSITE signature patterns: the S-segment S(5)-[DE]-x-[DE]-G-x(1,2)-G-x(0,1)-[KR](4) (with the exception of pea dehydrins, Arabidopsis thaliana COR47 and XERO2 and wheat cold-shock proteins) and the K-segment [KR]-[LIM]-K-[DE]-K-[LIM]-P-G. These common signature patterns are not present in the fungal sequences. However, another new repeated conserved signature pattern (CSP) was identified in sequences from filamentous fungi: [GRK]-[PV]-H-x-[ST]-x-x-x-N-[nonpolar amino acid]-[nonpolar amino acid]-D-P-[RTP]-V-D-[SN]. This deduced signature pattern identified at the C-terminal represents the first block of amino acids aligned by T-Coffee [25] and it is repeated from one to nine times within each sequence with slight variations (Fig. 2). The different number of repetitions of the CSP poses a challenging problem to the alignment of these regions, but T-Coffee aligned all the first and the last repetitions together, because further conserved residues flank these two repetitions. Not only TbDHN1 and the analysed fungal proteins, but all the previously cited fungal and plant sequences retrieved from blast searches show this CSP. It has to be underlined that sequences from C. albicans do not show this CSP but, in any case, they do share common physicochemical features with the other fungal sequences (Table 2).

As in plant DHNs, in TbDHN1, GenBank:EAA55454 from M. grisea, GenBank:EAA70367 from G. zeae, SNU00161 from S. nodorum, GenBank:CAD70810 from N. crassa and both proteins from C. albicans, there are repetitions of amino acid strings which are identical within the same sequence but different from one sequence to the other (data not shown).

Phylogeny, structure and localization prediction

The sequence alignment among TbDHN1 and its fungal homologs, selected plant LEA Class I, Class II, Class III and fungal Class III proteins (Table 3) generated by T-coffee was used to infer the phylogenetic relationship. NJ (data not shown) and Bayesian posterior probability analyses (Fig. 3) showed a congruent tree topology. Bayesian analyses yielded a phylogenetic tree in which clades corresponding to LEA Class I, Class II, Class III proteins are recognizable, although the LEA Class III sequence group is not well defined. A first clade, supported by a posterior probability value of 0.74, comprises all LEA Class I proteins and a Class III protein from P. sativum. A second clade, supported by a posterior probability value of 0.63, includes all LEA 2 proteins from plant, TbDHN1 and its fungal homologs. Fungal LEA Class II proteins clustered together with a posterior probability value of 0.78 and 61% NJ bootstrap value. Despite the lack of CSP, also C. albicans proteins are included in the same cluster. Notably, Swiss-Prot:Q9ZTR5_HORVU, the best hit of known function provided by blastx searches, falls in this clade. Plant and fungal Class III proteins clustered together but did not form a well separated clade supported by posterior probability and MP bootstrap values higher than 50%.

Accession numbera	Organism	Classified as
P04568	Triticum aestivum	LEA 1
P09443	Gossypium hirsutum	LEA 1
P11573	Raphanus sativus	LEA 1
P17639	Daucus carota	LEA 1
P22701	Triticum aestivum	LEA 1
P46514	Helianthus annuus	LEA 1
P46517	Zea mays	LEA 1
P46520	Oryza sativa	LEA 1
Q02973	Arabidopsis thaliana	LEA 1
Q05190	Hordeum vulgare	LEA 1
Q05191	Hordeum vulgare	LEA 1
Q07187	Arabidopsis thaliana	LEA 1
O04232	Solanum tuberosum	LEA 2
O22623	Vaccinium corymbosum	LEA 2
O48622	Spinacia oleracea	LEA 2
O65216	Triticum aestivum	LEA 2
P12253	Oryza sativa	LEA 2
P12950	Zea mays	LEA 2
P22239	Craterostigma plantagineum	LEA 2
P22240	Lycopersicon esculentum	LEA 2
P28639	Pisum sativum	LEA 2
P42758	Arabidopsis thaliana	LEA 2
Q07322	Daucus carota	LEA 2
Q9ZTR5*	Hordeum vulgare	LEA 2
O49816	Cicer arietinum	LEA 3
P13934	Brassica napus	LEA 3
P13939	Gossypium hirsutum	LEA 3
P14928	Hordeum vulgare	LEA 3
P20075	Daucus carota	LEA 3
P23283	Craterostigma plantagineum	LEA 3
Q39058	Arabidopsis thaliana	LEA 3
Q39873	Glycine max	LEA 3
Q40696	Oryza sativa	LEA 3
Q40869	Picea glauca	LEA 3
Q40929	Pseudotsuga menziesii	LEA 3
Q41060	Pisum sativum	LEA 3
Q00002	Alernaria alternata b	LEA 3
Q5UUY7	Antonospora locustae b	LEA 3
Q6FVF6	Candida glabrata b	LEA 3
Q6BSI4	Debaryomyces hansenii b	LEA 3
Q8SVY9	Encephalitozoon cuniculi b	LEA 3
Q6CP34	Kluyveromyces lactis b	LEA 3
Q7S1U2	Neurospora crassa b	LEA 3
Q7SH94	Neurospora crassa b	LEA 3
Q9Y7X6	Schizosaccharomyces pombe b	LEA 3
Q6C0I4	Yarrowia lipolytica b	LEA 3
Q6C052	Yarrowia lipolytica b	LEA 3
Q6C800	Yarrowia lipolytica b	LEA 3

^a Swiss Protein Database Accession Number; ^b Fungal organisms; *This is DHN6 from Hordeum vulgare (corresponding to GenBank: AAD02257) which represents the first hit of known biological role found by blastx using TbDHN1 as query sequence.

A multiple alignment among TbDHN1 and its fungal homologs were generated with T-coffee and used to build a Hidden Markov Model profile. This was used to query Swiss-Prot and Pfam databases, but no hit with a significant score was retrieved: the first hit was an Ice nucleation protein (Swiss-Prot:P09815), E-value = 0.076, from Pseudomonas fluorescens.

WoLF PSORTII predicted for TbDHN1 and the other twelve fungal proteins a probable nuclear/cytoplasmic localization, but no nuclear localization signals were found by PredictNLS program.

The case study of TbDHN1: gene structure

The entire sequence of TbDHN1 gene was obtained from a T. borchii cDNA and genomic library screening. TbDHN1 mRNA was a 2300-bp-long open reading frame coding for a predicted 351 amino acid long polypeptide with a predicted molecular mass of 34,8 kDa. The genomic sequence revealed the presence of a 79-bp-long intron after 178 nucleotides from the translation initiator ATG and a putative TATA-box located at position -35 nt from the transcription start site (Fig. 4). Besides the presence of 4 repetitions of the CSP, another peculiar feature has to be pointed out: a block of 46 amino acids is repeated within the sequence, separated only by a glutamic acid residue. The corresponding nucleotide sequences are almost identical with slight variations only in the third nucleotide of some codons (Fig. 4).

The genomic region amplified by DHNf/DHNr primers contains one KpnI restriction site but no site for HindIII and SmaI. As revealed by DNA gel blot data reported in Figure 5A, a single-band pattern was produced by hybridization of the TbDHN1 probe to genomic DNA digested with HindIII and SmaI and a double-band was produced by KpnI, thus indicating that TbDHN1 is encoded by a single copy gene in T. borchii genome.

Stress-regulated expression of TbDHN1

Previous experiments showed that TbDHN1 transcript was strongly upregulated in T. borchii fruiting bodies compared to vegetative mycelium grown on control medium [21], but, since fruit bodies can be collected only in field, we chose to further investigate TbDHN1 expression using vegetative mycelium grown in axenic conditions.

RNA gel-blot experiments were conducted to determine whether TbDHN1 transcription levels increase in the same conditions affecting plant DHNs transcription, e.g. low temperature stress and salinity stress. The RNA gel-blot data reported in Figure 5B showed that the TbDHN1 messenger increased immediately following transfer on NaCl medium for 30 min and such up-regulation is still evident in a NaCl treatment prolonged to 24 h. The same strong up-regulation response is also evident after a 48 h-long cold treatment. TbDHN1 messenger rapidly returned to basal levels when mycelia cultivated for 24 h on NaCl medium were shifted for 5 h on control medium, while in mycelia continually cultivated for 29 h on NaCl medium the up-regulation was still evident.

Why consider TbDHN1 and its homologs as DHN-like proteins?

In silico and experimental analyses suggest that TbDHN1 and its homologs should be classified as dehydrins. There is much evidence to support this claim. Firstly, all the protein sequences shared the typical physiochemical properties of the large family of LEA proteins. More precisely, they fitted the set of rules defined by Wise to define Class II, the group of all plant dehydrins, although Wise trained his application with plant proteins exclusively. Class II of plant LEA proteins is also characterized by the presence of two conserved signature patterns. TbDHN1 and its homologs do not have any of the already identified conserved signature patterns, but those proteins from filamentous fungi and plants show a new, previously uncharacterized CSP. This pattern is variously repeated in the fungal sequences: from the single repetition in N. crassa Q7S6B0 to the nine repetitions in A. nidulans GenBank:EAA62484 and A. fumigatus GenBank:EAL89059. Except for the border-line case of N. crassa Q7S6B0, the last repetition of CSP at the C-terminal of all the sequences is the most conserved one. The absence of the conserved signature patterns typical of plant dehydrins is not sufficient to exclude a protein from Class II of LEA proteins; in fact Wise includes Swiss-Prot:O22623 from Vaccinium corymbosum in Class II, although the sequence lacks both conserved signature patterns.

All this evidence is further supported by the phylogenetic inference: the tree topology shows how TbDHN1 and its fungal homologs cluster together with LEA Class II proteins. C. albicans proteins too fall in the same clade, although they lack the fungal CSP.

We can thus determine that TbDHN1 and its homologs are dehydrin-like proteins. EAK94850 (NCBI Protein Database Accession Number) and EAK94909 (NCBI Protein Database Accession Number) from C. albicans were already recorded in the protein database at NCBI as "dehydrins", but exclusively on the basis of their homology with DHN6 from H. vulgare, the same protein which represents the best hit of the known biological role for TbDHN1. Here, our physiochemical and structural analyses provide further substantial elements to their assignment as dehydrins.

TbDHN1 gene structure and expression pattern

In silico analyses on fungal dehydrin-like proteins were followed by an extensive study on TbDHN1 gene structure and expression pattern in stress vs. control conditions. Southern blot analysis showed that TbDHN1 is a single copy gene but, in the absence of complete T. borchii genome sequence data, we cannot exclude the presence of other DHN-like coding genes in T. borchii. Since a number of DHN-like genes from the same genome (e.g. N. crassa, M. grisea, C. albicans and A. fumigatus) are very different in nucleotide sequence, the TbDHN1 probe could be expected to fail to match other DHN-like genes. The presence of multiple DHN-like proteins in a single proteome leads to the hypothesis that they represent a multigenic family; in fact, concerning A. fumigatus genome, the three identified DHN-like proteins are localized on three different chromosomes: EAL89059 on chromosome 6, EAL84332 (NCBI Protein Database Accession Number) on chromosome 4 and EAL87020 on chromosome 7.

RNA gel-blot experiments determined that TbDHN1 transcription levels increase in the same conditions affecting plant DHNs transcription, e.g. low temperature stress and salinity stress. This strong upregulatory response is fully reversible, since basal levels of the TbDHN1 transcript were recovered following additional five-hour incubation on control medium.

The upregulation of M6G10 EST in fruiting body compared to vegetative mycelium [21] is probably strictly linked to the exposure of fruiting bodies to low temperatures during the autumn/winter season. T. borchii ascomata develop from November to April and the best harvests generally result from a period of poor growing conditions, such as drought and low temperature, immediately followed by optimal conditions of warmth and moisture [34]. Fruiting bodies analyzed in our previous work were in fact sampled in a natural truffle ground near Alba in Piedmont (Italy) during the December 1999 – March 2000 and December 2000 – March 2001 production seasons, when the average registered temperatures were 5.7°C and 6.3°C, respectively. This may explain why M6G10 EST was the most expressed transcript in fruiting bodies compared to vegetative mycelium grown at 24°C. The high up-regulation of TbDHN1 transcript in water-limiting conditions could represent one of the molecular mechanisms to face the threat of desiccation. Many other genes encoding for stress proteins were up-regulated during the reproductive stage [21], including a putative hsp12 [GenBank:CN488002] which shares 31% identity with hsp12 from S. cerevisiae. The latter is also supposed to be a LEA-like protein because it is highly expressed when there is an increased osmolality within the cell, such as during spore maturation [14]. It thus seems that in T. borchii there is a molecular mechanism involving a set of genes specifically induced in response to cellular dehydration resulting from developmental events like fruiting body formation/maturation or environmental stimuli such as osmotic stress and low temperature. Although a transient transformation of T. borchii mycelium has already been obtained [35], stable transformants cannot be exploited to provide a direct functional evidence to support this hypothesis.

Sequence isolation

The 771-bp-long M6G10 EST was isolated from an expression library constructed from a mature fruiting body (CF70) of Tuber borchii [21]. cDNA array hybridizations showed that this clone was the highest upregulated gene (163.7-fold) between the reproductive and vegetative stage.

Two specific oligonucleotides (DHNf: 5'-GGTACTCGCGAAACACACTCTG-3'); DHNr: 5'-TGGCCTTCTAGAATTAACTCG-3') were used as PCR primers to amplify the corresponding region on M6G10 EST. PCR reaction was performed in a 30 μl volume reaction which consists of 80 μM of each deoxynucleoside triphosphate, 50 pmol of each primer, 1 U of REDTaq DNA polymerase (Sigma), and the buffer supplied by the manufacturer of the REDTaq DNA polymerase. The PCR parameters were as follows: 94°C for 3 min; 94°C for 45 sec, 60°C for 45 sec, and 72°C for 1 min for 35 cycles; and a final cycle at 72°C for 10 min. The resulting DNA fragment was purified with the QIAquick PCR purification kit (Qiagen) and utilized as a probe for plaque hybridization analysis of two T. borchii libraries: a cDNA library of vegetative mycelium grown for 30 days on MMN liquid medium, constructed in the phage vector Uni-Zap XR, and a DNA library constructed in the phage vector EMBL4 (kindly provided by B. Lazzari and A. Viotti, Istituto Biosintesi Vegetali, CNR, Milano). Phage DNA extracted from hybridization-positive plaques of both screenings was sequenced by Genome Express (Meylan – France). The full-length cDNA and the corresponding genomic sequence were thus obtained and deposited to GenBank database [GenBank:DQ308610]. Hereafter we refer to M6G10 complete genomic and mRNA sequence as TbDHN1.

Homology search

Initial homology searches were conducted with M6G10 sequence against the non-redundant (nr) protein sequence and the Swiss-Prot databases at the NCBI by using blastx program [38]. To identify homologs within Expressed Sequence Tags collections, blastn and tblastn searches were run against the EST database at NCBI and the fungal plant pathogen database [39].

Genome and protein sequences (2005-06-31 release) from fungi and plants available at NCBI, the Broad Institute [40] and TIGR [41] were downloaded to a local database and formatted to be used with a local BLAST installation. The complete nucleotide sequence of TbDHN1 was used to query public databases using blastx and blastp programs. All blast results were loaded into a MySQL [42] database for further analyses and data mining.

On the basis of sequence similarities, other twelve, previously uncharacterised, fungal TbDHN1 homologs were identified (Table 1). All twelve fungal proteins derived from a conceptual translation of their genome. The following in silico analyses were performed on TbDHN1 and these twelve fungal hypothetical proteins.

Secondary and tertiary structure predictions

Secondary structure percentage composition was obtained by a four-state prediction for each amino acid using PHDsec from ProteinPredict server [43]. The four-state predictions were converted to percentage composition values in order to minimize the effects due to differences in length across the sequences as suggested by Wise [12].

The program T-Coffee [25] was used to build a multiple alignment with TbDHN1 and the hits obtained by homology searches (Table 1). The alignment was slightly manually optimized with GeneDoc [44]. Based on the multiple fungal alignment, a profile hidden Markov model (HMM) was built using HMMER [45] in order to search Swiss-Prot and Pfam for a sequence family's consensus.

Phylogenetic analysis

T-Coffee was used to build a multiple alignment among the TbDHN1 fungal homologs, plant LEA proteins recently classified by Wise [12] in Class I, Class II and Class III and fungal sequences previously reported as LEA proteins by Pfam database (Pfam Database Accession Number PF02987) (Table 3). Among all plant LEAs, twelve proteins were chosen from each LEA class to represent the highest number of different plant species. Only LEA Class I, II and III were considered because they are the only three groups in which conserved consensus sequence motifs were identified. The alignment was slightly manually optimized with GeneDoc.

The Neighbour-joining (NJ) was performed using PAUP* 4.0b10 [46]. Robustness of the internal branches was assayed by bootstrap analysis (1000 runs). Bayesian inference based upon posterior probability distribution of trees was performed with MrBayes [47] with the following settings: seed and swapseed = 1, generations = 1,000,000, runs = 2, chains = 4, sampling every 100^th generation, starting trees = random, other settings = default. Phylogenetic trees were rendered with TreeViewX [48].

Amino acid composition

The EMBOSS application Pepinfo [49] was used to calculate amino acid class (Aliphatic, Aromatic, Non-polar, Polar, Charged, Basic and Acidic) percentage compositions and hydrophobicity values based on the method of Kyte and Doolittle. As regards the determination of hydrophobicity, a 21 aa window was used as suggested by Wise [12]. Three values were considered for each sequence: the minimum, the maximum and the average hydrophobicity; negative hydrophobicity values indicate hydrophilicity.

Subcellular localization prediction

Subcellular protein localization and cleavage sites predictions were made with the WoLF PSORTII web server [50] selecting "fungi" as organism type; nuclear localization signals were predicted with the PredictNLS program [51].

Expression and DNA analyses

Expression studies were performed to characterize only TbDHN1 gene. T. borchii Vittad. mycelia (isolate ATCC 95640) were grown in the dark at 24°C on a synthetic solid medium (control medium, CM) containing CaCl₂· 2H₂O (66 mg/L), NaCl (25 mg/L), KH₂PO₄ (500 mg/L), (NH₄)₂HPO₄ (250 mg/L), MgSO₄· 7H₂O (150 mg/L), FeNa-EDTA (20 mg/L), and thiamine hydrochloride (1 mg/L), plus agar (15 g/L) and glucose (5 g/L). To allow easy medium replacement, myceliar cultures were grown on semipermeable cellophane membranes. To determine whether TbDHN1 transcription levels increase in the same conditions affecting plant dehydrin transcription, mycelia were first cultured on control medium at 24°C and then transferred to either a single component-changed media at 24°C in which NaCl was 1000-fold more concentrated than in CM (hereafter designated as NaCl treatment) or to the same control medium at 4°C (hereafter designated as cold shock treatment). NaCl treatments were performed using three periods of exposure to high salt concentration: 30 min, 24 h and 29 h. On the other hand, the cold shock treatment consisted of a 48 h-cold exposure. For "switch-off" experiments, 24 h-NaCl treated mycelia were returned for 5 h to control medium (hereafter designated as NaCl→CM). All treated and untreated mycelia were harvested after 30-days' growth.

Total RNA for Northern Blot analysis was extracted with a LiCl method [52] from T. borchii mycelia grown for: a) 30 min, 24 h and 29 h or under of NaCl treatment; b) 48 h of cold treatment; c) 5 h on CM after 24 h of NaCl treatment and d) 30 days at 24°C on CM.

Heat-denatured total RNAs (10 μ g for each sample) were fractionated on 1.8% formaldehyde-agarose gels and transferred onto a Hybond membrane (Amersham) by capillary elution. A TbDHN1 cDNA probe was obtained by PCR amplification on M6G10 EST as previously described and hybridized to RNA blots containing size-fractionated RNA.

The random-priming labelling of 30 ng DNA probe was carried out in the presence of [³²P]dCTP. Stringent washes were performed at 65°C with 0.5X SSC and 0.1% (w/v) SDS. Each filter was then de-hybridized and probed with T. borchii 28S ribosomal DNA (rDNA) to estimate the level of total RNA loaded in each lane. Dried filters were exposed to the phosphor screen for 24 hours. Signals were quantified with a Personal Imager FX using Quantity One (Version 4.5.0) software (Biorad). Normalization for differences in loading, labelling and hybridisation were made based on the 28S ribosomal signal. Normalized signals of all treatments were compared with the control signals. Expression data are representative of three independent experiments.

Genomic DNA samples for gel blot analysis (10 μg each) were obtained from mycelia, according to Henrion et al. [53], digested with HindIII, KpnI and SmaI and size fractionated on a 0.8% agarose gel. A chemiluminescent labelling kit (ECL direct nucleic acids labelling and detection system, Amersham) was used for labelling TbDHN1 genomic probe; blotting onto Hybond-N (Amersham Pharmacia Biotech), pre-hybridization, hybridisation and stringent washing were conducted according to the manufacturer's instructions.