Characterization and expression of XPH1, the hydrophobin gene of the lichen-forming ascomycete Xanthoria parietina

Authors


  • 1

    The EMBL Nucleotide Sequence Database accession numbers for the XPH1 (H1 of Xanthoria parietina) sequences reported in this paper are AJ250794 (cDNA) and AJ287227 (genomic DNA).

Author for correspondence: Rosmarie Honegger Fax: +41 1634 82 04 Email: rohonegg@botinst.unizh.ch

Summary

  •  Molecular genetics of wall surface hydrophobicity of the lichen-forming ascomycete Xanthoria parietina was explored. In a previous study XPH11, a class 1 hydrophobin, was isolated from the symbiotic phenotype collected in nature.
  •  A genomic library of aposymbiotically cultured Xanthoria parietina was constructed and screened for the hydrophobin gene XPH1. Gene expression was explored with in situ hybridization techniques.
  •  The encoded protein contained 110 amino acids, including a secretion signal; the open reading frame was interrupted by two introns. XPH1 was differentially expressed in the symbiotic phenotype, high expression being evident in medullary hyphae of the vegetative thallus and of the thalline margin of apothecia; these are in contact with the algal layer and prevent the thalline interior from becoming waterlogged, a prerequisite for gas exchange of the photobiont. No signal was detected in either the hydrophilic cortex, the conidiomata, the hymenial and subhymenial layers of apothecia, or in the aposymbiotically cultured mycobiont. The data for Northern blot analyses correlated with those of in situ hybridization studies.
  • XPH1 may fulfil important roles for the functioning of the symbiotic relationship.

Introduction

Lichens are the symbiotic phenotype of a taxonomically diverse group of nutritionally specialized fungi that acquire carbohydrates from a population of unicellular green algae or cyanobacteria. The species name of lichens refers to the fungal partner which, as an ecologically obligate biotrophic fungus is found in nature almost exclusively in symbiosis with its green algal or cyanobacterial photobiont. However, the majority of lichen-forming fungi are physiologically facultative biotrophs and thus can be axenically cultured in the aposymbiotic state. A high percentage of lichen-forming fungi, the so-called macrolichens, differentiate a morphologically and anatomically complex thallus, either leaf-shaped or shrubby (foliose or fruticose), in which the photoautotrophic partner is housed and actively shifted by specialized contacting fungal hyphae into an optimal position for illumination and gas exchange. Such thalli are the product of a complex hyphal polymorphism, comprising tissue-like (pseudoparenchymatic), hydrophilic zones as peripheral cortical layers, which enclose a system of aerial hyphae with water-repellent wall surfaces; these create an air-filled zone around the photobiont layer in the thalline interior, which is a prerequisite for the successful gas exchange of both partners of the symbiosis (for reviews see Honegger, 1991, 1993, 1997). Tissue-like peripheral cortical layers absorb water and dissolved mineral nutrients; these move within the fungal cell wall (apoplast) immediately underneath the superficial hydrophobic coat to reach the algal and medullary layers (Honegger, 1991, 1997). Mycobiont-derived secondary metabolites may crystallize either on or within the hydrophilic cortex (cortical) or on the wall surfaces of aerial hyphae in the thalline interior (medullary lichen products), thus enhancing their hydrophobicity.

The present study focuses on the molecular genetics of wall surface hydrophobicity in Xanthoria parietina, a common and widespread, foliose macrolichen with an internally stratified thallus. Bright yellow anthraquinones crystallize on and within the topmost part of the peripheral cortex. No medullary secondary metabolites are formed, yet the hyphal and green algal wall surfaces in the thalline interior are hydrophobic. As shown in low temperature scanning electron microscopy (LTSEM) preparations of specimens that had been cryoimmobilized and freeze-fractured in the fully hydrated state, no free water accumulates on these hydrophobic surfaces in the medullary and algal layers; this prevents the photobiont cell population from getting waterlogged at high levels of thalline hydration (Honegger & Peter, 1994; Honegger et al., 1996).

A thin rodlet layer, structurally resembling the rodlet layer on aerial structures of nonlichenized fungi (review: Wessels, 1997), was resolved in transmission electron microscopy (TEM) studies of freeze-etch replicas of aerial hyphae from the thalline interior of numerous taxa of lichen-forming ascomycetes (Honegger, 1982, 1984a, 1986, 1991, 1997), Xanthoria parietina included (Scherrer et al., 2000). It is this rodlet layer that seals the wall surfaces of the mycobiont and photobiont of macrolichens with a thin, water-repellent coat and channels the fluxes of solutes in the apoplastic continuum. Mycobiont-derived medullary secondary metabolites may crystallize on and within this layer and thus enhance its hydrophobicity (Honegger, 1991).

Hydrophobic wall surface layers of aerial structures in nonlichenized fungi have been explored in considerable detail (for reviews see Wessels, 1997, 1999, 2000). The rodlet layer was shown to be formed by hydrophobins, a class of small, secreted fungal proteins of approx. 100 amino acids, with eight cysteine residues in a conserved spacing and a similar hydropathy pattern (Wessels, 1997). The expression of hydrophobins is connected to various aspects of fungal morphogenesis. Different hydrophobins are involved in the development of vegetative aerial structures and sporulation, fruit body formation and parasitic or mutualistic interactions (Wessels, 1997, 1999, 2000; Wösten & Wessels, 1997).

A previous investigation focused on the rodlet layer on the hydrophobic cell wall surfaces of the algal and medullary layers of the symbiotic phenotype of Xanthoria parietina. The protein XPH1, isolated from washed cell wall fragments, showed all the characteristics of class I hydrophobins, including interfacial self-assembly into a rodlet layer, and the protein-encoding sequence of the XPH1 gene was identified (Scherrer et al., 2000). The present study investigated the structure of the entire hydrophobin gene XPH1, and its expression in the lichen thallus (i.e. the symbiotic phenotype) and in aposymbiotic cultures of Xanthoria parietina.

Materials and Methods

Organisms and culture conditions

The symbiotic phenotype of Xanthoria parietina was collected on an asbestos composite roof near our institute at the periphery of the city of Zurich, Switzerland.

An axenic multispore isolate of Xanthoria parietina, which was cultured in our laboratory (Honegger, 1990), was used for constructing the genomic library. The mycelium was grown for 3 months in a shaken liquid nutrient broth (mineral medium (Deason & Bold, 1960) plus 0.5% malt extract, 0.2% glucose, 0.025% proteose peptone, 0.025% casamino acids, 0.025% yeast extract) prior to the extraction of genomic DNA or total RNA. For expression experiments, multispore isolates of specimens collected in Zurich were grown for 5 months on either a glass fibre or a cellulose filter overlying agarized nutrient medium with reduced nitrogen content (without proteose peptone, casamino acids and yeast extract). All cultures were kept at 15°C in continuous illumination.

Nucleic acid isolation and electrophoresis

Lichen thalli or mycelium of the cultured mycobiont, respectively, were frozen and ground in liquid nitrogen after the addition of small amounts of aluminium oxide. Genomic DNA was extracted either after the modified protocol of Dellaporta et al. (1983; Scherrer et al., 2000) or according to the method of Borges et al. (2001). The temperature of the first extraction step proved to be important because of the high polysaccharide content of the lichen samples. By extracting at room temperature fewer carbohydrates were dissolved than at 40°C to 60°C, as recommended in standard procedures, thus the inhibitory effect on the subsequent enzymatic reactions was reduced.

For the isolation of λ phage DNA, the phages were grown according to the method of Burmeister and Lehrach (1996). The λ phage DNA was isolated following the instructions of the λ Phage Prep Protocol (Stratagene, La Jolla, CA, USA). For the preparation of plasmid DNA, a single colony of Escherichia coli was grown and plasmid DNA isolated by alkaline lysis (Sambrook et al., 1989). For sequencing reactions the plasmid DNA was prepared with the Plasmid Mini Kit (Qiagen, Crawley, UK).

For expression experiments the air dried lichen thalli were rehydrated for 8 h in the light prior to grinding in liquid nitrogen. Total RNA was extracted with the hot phenol method (Dudler & Hertig, 1992).

DNA and RNA electrophoresis and blotting were carried out using standard procedures (Sambrook et al., 1989). After blotting, the nucleic acids were bound to membranes (Hybond N+, Amersham Bioscience Europe, Dübendorf, Switzerland) in an ultraviolet crosslinker (Amersham).

Northern and Southern blot analysis

The cDNA fragment of XEH1 cloned in pCR 2.1 (Invitrogen, Groningen, The Netherlands; Scherrer et al., 2000) was used as a probe for Northern and Southern blots. For radioactive detection, probes were labelled with 32P by random priming (reaction mixture and enzyme: Boehringer Mannheim, Germany) and cleaned with Nick columns (Amersham Pharmacia Biotech Europe, Dubendorf, Switzerland). Hybridization and detection were carried out using standard procedures (Sambrook et al., 1989). Two different systems were used for nonradioactive detection: (1) the probe was biotinylated with the Random Octamer Labeling System (Tropix, Bedford, MA, USA); (2) with the AlkPhos direct Kit (Amersham Pharmacia, Cambridge, UK), the cDNA fragment was labelled directly with the enzyme. Hybridization and detection was performed either according to the protocol of Southern light (Tropix) or Alkphos direct (Amersham Pharmacia Biotech Europe).

Construction of the genomic library

DNA of axenically cultured, aposymbiotic Xanthoria parietina was partly digested with an optimized concentration of Sau3AI (Boehringer Mannheim) for 1 h at 37°C. After inactivation of the enzyme, the DNA was loaded on a preparative agarose gel. Fragments 9–20 kb long were cut out and electroeluted in a dialysis bag. Electroelution and the following concentration steps were carried out according to Sambrook et al. (1989). Subsequent to the fill-in reaction with dGTP and dATP, the fragments were ligated into the arms of λFixII vector (Stratagene). Recombinant λ phages were packed using the Gigapack III Gold Packaging extract (Stratagene). Ligation and packaging was performed according to the manufacturers instructions. The sizes of the two libraries constructed were determined by titration. The resulting values of 400 pfu µl−1 and 1000 pfu µl−1, respectively (total volume of each: 500 µl) were a promising basis for the successful screening for the XpH1 gene. Library 1 was screened directly; library 2 was amplified and stored at –80°C for further investigation.

Screening, subcloning and sequencing

The radioactively labelled cDNA fragment of XEH1 was used to screen the library. Plaques were lifted on nitrocellulose membranes (Schleicher & Schüll, Feldbach, Switzerland) and denatured according to Sambrook et al. (1989). The DNA of positive phage clones was isolated and digested with different restriction enzymes. Positive fragments were subcloned into the vector pBluescript SK+ (Stratagene, La Jolla, CA, USA) and sequenced three times on both strands. The following primers were used: M13 forward, M13 reverse, xpgf (5′-GATTTCAGCCAGCCAGCTCGC-3′), xpgr (5′-AGATGGCGAAACCCACG-3′), XpH1P7 (5′-TGGGACTTGGCTCATTTTCG-3′) and XpH1P8 (5′-CCTTGCCATAGACTATGGAC-3′).

The sequencing reaction was performed with an ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Forster City, CA, USA) following the instructions of the manufacturer and analysed in a 373 DNA Sequencer (Applied Biosystems). The sequences were analysed using the GCG version 8 software (Wisconsin Sequence Analysis Package Genetics Computer Group, Madison, Wisconsin, USA).

Reverse transcription and polymerase chain reaction

Reverse transcription was carried out according to Dumas et al. (1995) with 25 U AMV Reverse Transcriptase (Boehringer Mannheim) and 10 µg total RNA. The reaction was performed for 45 min at 42°C in 50 µl of the following reaction mixture: 10 mm Tris-HCl pH 8.3 (at 42°C), 0.6 mm anchored (A/C/G) oligo-dT primer (24-mer, MWG-Biotech, Münchenstein, Switzerland), 8 mm KCl, 1.6 mm MgCl2, 5 µg bovine serum albumin (BSA), 5 U RNase inhibitor (Pharmacia), 14 mmβ-mercaptoethanol, 1 mm of each dNTP and 4 mm sodium pyrophosphate.

Polymerase chain reaction (PCR) was performed with 5 µl of the reverse transcription (RT) reaction mixture and 1 U Taq DNA Polymerase (Boehringer Mannheim) in 50 µl reaction volume. The reaction mixture contained PCR buffer (supplied with the enzyme, with 1.5 mm MgCl2), 200 µm of each dNTP, 5 µm oligo-dT as reverse primer, and 5 µm specific forward primer (MWG-Biotech), which was designed on the basis of the sequence of the genomic clone of XPH1 (XpP6: 5′-ACATCATCGCCTTTGTCTCC-3′). Thirty-five cycles with the following profile were programmed: denaturation for 1 min at 94°C, annealing for 30 s at 49°C, extension for 1 min at 72°C, with an initial denaturation of 2 min and an additional terminal extension step of 6 min (Perkin Elmer Gene Amp PCR System 9600, Perkin Elmer Ltd, Beaconsfield, UK). The reverse-transcription polymerase chain reaction (RT-PCR) product was cloned into the vector pCR 2.1 using the TOPO TA Cloning Kit (Invitrogen).

In situ hybridization

Digoxigenin-based (DIG), nonradioactive in situ hybridization experiments were carried out essentially following the protocol published by Jackson (1991) and Long et al. (1996). Thallus fragments of lobes and apothecia were fixed with FAA (3.7% formaldehyde, 5% glacial acetic acid, 50% ethanol), dehydrated and embedded with Paraplast (Sigma). Sections, 8 µm thick, were fixed to slides covered with Vectabond (Reactolab S. A, Servian, Switzerland). After hybridization at 42°C, the sections were washed once with 2 × SSPE (0.3 m NaCl, 2 mm ethylenediaminetetraacetic acid (EDTA), 20 mm NaH2PO4, pH 7.4) at 50°C and three times with 0.2 × SSPE at 50°C, 37°C and room temperature, respectively. The RNase treatment was omitted. The sections were stained with Western blue (Promega, Southampton, UK) for 6 h. The staining reaction was stopped with TE (1 mm EDTA, 10 mm Tris-HCl, pH 8.0). After mounting in 50% glycerol and sealing of the cover slip with nail polish the sections were examined in either a Zeiss Photomikroskop 2 or a Zeiss Axioplan Microscope with DIK optics.

The cDNA fragment of XPH1 (c. 700 bp) was cloned in both directions into the pCR 2.1-TOPO vector (Invitrogen). To generate a positive control, a 460-bp fragment of the region coding for the nuclear small ribosomal subunit (SSU rDNA) of Xanthoria parietina was amplified by PCR with the primers nu-SSU-0072-5′-and nu-SSU-0497-3′ (Gargas & Taylor, 1992; Gargas & DePriest, 1996) and cloned into the pGEM-T easy vector (Promega). Antisense and sense RNA probes were prepared after linearization of both vectors with either BamHI, NcoI or Sall by in vitro transcription (including DIG-UTP, Roche Diagnostics, Lewes, UK) with T7 or SP6 RNA polymerase (NEB, Frankfurt am Main, Germany and Boehringer Mannheim, respectively) according to the manufacturers instructions.

Scanning electron microscopy

Low-temperature scanning electron microscopy (LTSEM) of frozen-hydrated specimens was performed according to Honegger et al. (1996). Pycnidia were prepared for conventional scanning electron microscopy (SEM) studies according to Honegger (1984b).

Results

A genomic λ phage library of aposymbiotically cultured Xanthoria parietina was constructed and screened for the hydrophobin gene XPH1. Positively hybridizing EcoRI fragment (5 kb) and the positive XhoI fragment (3 kb) were subcloned into pBluescript SK+. The sizes of these fragments corresponded exactly to the ones determined on genomic DNA blots of the symbiotic phenotype and of the aposymbiotically cultured mycobiont (Fig. 1). The XhoI fragment was restricted with HindIII into a 1.6 kb and a 1.4 kb fragment and subcloned. The 1.6 kb fragment was sequenced twice completely from both directions. Because the 3′-end of the gene was situated on the 1.4 kb fragment, the corresponding region on the 5 kb fragment was sequenced on both strands.

Figure 1.

Southern blot analysis of the symbiotic and aposymbiotic phenotypes of Xanthoria parietina from Zürich: lichen thallus (XpZH) vs axenically cultured isolate (XpMycZH). Genomic DNA was digested with four different enzymes; HindIII cut within the gene. The slight difference in size of the positive fragments between the two phenotypes might be due to the lower purity of the DNA extract of the lichen thalli.

Molecular characterization of the XPH1 gene

The complete nucleotide sequence of the coding and flanking regions of the XPH1 gene was determined (Fig. 2). In addition, the cDNA fragment of XPH1 obtained by RT-PCR was sequenced. No discrepancies were found between the cDNA fragment and the genomic clone. Comparison of the two sequences indicated that the XPH1 gene contained two introns in the coding region of 50 bp and 47 bp. The introns contained splice sites following the GT/AG rule and consensus sequences as typically found in filamentous fungi (Gurr et al., 1987; 5′ site, GTANGT; 3′ site, PyAG). Sequences of a putative internal splice signal, proposed as the site of lariat formation were present in both introns (intron 1, CGCTAAC; intron 2, TACTGAT; Gurr et al. 1987; Unkles, 1992). The length of this transcript of approximately 0.95 kb was determined using Northern blots (Fig. 3). This is in good correlation with the genomic sequence. Assuming a poly(A) tail of about 250 bp, the 5′ untranslated region must be relatively short. The 3′ untranslated region is 389 bases long (without polyA tail). The longest open reading frame (ORF) encodes a protein with 110 amino acids. Although two candidate methionines for the translational start were found in frame upstream of the known sequence, only the first one showed context sequences that corresponded to the translation initiation consensus sequence (Gurr et al., 1987; Unkles, 1992). The N-terminus of the encoded protein was formed by a series of apolar amino acids typical for a signal sequence. Computer analysis (Signal P WWW Server, Center for Biological Sequence analysis (http://www.cbs.dtu.dk/services/signalp/). The method was described by Nielsen et al. (1997).) recognized the N-terminus as a leader sequence and located a most likely cleavage site between amino acid 18 and 19: ASA-AP. This does not correlate with the N-terminal sequencing of the highly homologous protein XEH1 of Xanthoria ectaneoides (Scherrer et al., 2000). Mature XEH1 starts with TTPG-; thus, the secretion signal is 27 amino acids long.

Figure 2.

Sequence of the XPH1 gene. (a) Gene structure. Since the transcriptional start was not determined, the 5′-end of the XPH1 transcript is indicated by a dotted line and the positions marked in (b) refer to the translational start. (b) Genomic and deduced amino acid sequence. Putative promoter elements are underlined. The coding region (bold letters) is interrupted by two introns and includes a secretion signal sequence. The start of the mature protein is marked with an arrow (↑) and the eight cysteines, in the characteristic arrangement of hydrophobins, are labelled with asterisks. The positions of the HindIII restriction site, the putative polyadenylation signal and the mRNA cleavage site are indicated.

Figure 3.

Northern blot analysis of aposymbiotically cultured isolates and of the symbiotic phenotype (lichen thallus) of Xanthoria parietina. The size of the XPH1 transcripts is approximately 0.95 kb (a) Lanes 1 and 2, cultures grown in shaken liquid nutrient medium. (b) Lanes 1 and 2, cultures grown on either a glass-fibre or a cellulose filter overlying agarized nutrient medium with reduced nitrogen content (without proteose peptone, casamino acids and yeast extract); lanes 3 and 4, symbiotic phenotype from Roscoff (F) and Zurich (CH); lane 5, wheat RNA as a control. (c) Expression of XPH1 in either apothecia or vegetative lobes of Xanthoria parietina collected in Zurich. Lanes 1 and 2, marginal lobes; lanes 3 and 4, central thallus areas with a high proportion of apothecia.

Since the transcriptional starting point was not determined, the following positions of putative core promoter elements are given in base pairs from the A (+1) of the putative translational start (see also Fig. 2). A putative TATA (–47) and three CAAT (–54, –65 and –107) boxes were located upstream of the translation initiation site. The TATA box of filamentous fungi often does not exactly follow the classical consensus sequence as predominant in other eukaryotes (Gurr et al., 1987); AT-rich regions were found in several genes, as in XPH1 here. CT-rich motifs are known to precede the transcriptional initiation site in the promoter region of filamentous fungi. These elements were also found in XPH1 at the positions –28, –104 and –123. The polyadenylation site was located 15 bases after the putative polyadenylation signal, which was found at position +798, 367 bases after the stop codon.

Expression of XPH1

In the lichen thallus, i.e. the symbiotic phenotype, XPH1 was highly expressed, but no or only a very weak signal could be detected in the aposymbiotically cultured mycobiont (Fig. 3). Different culture conditions were tested: the mycobiont was grown either in liquid or on agarized media. Although aerial hyphae were formed at the periphery of the cartilaginous colonies on agar plates, no XPH1 signal was detected. Aposymbiotically cultured Xanthoria parietina produces neither vegetative nor sexual sporulating structures. Therefore, we tested whether XPH1 might be fruitbody specific. RNA was extracted separately from lobe margins and from apothecia-bearing central parts of the thallus from the symbiotic phenotype, but XPH1 was expressed in both samples (Fig. 3c).

In situ hybridization

The spatial expression pattern of XPH1 in the lichen thallus was investigated by in situ hybridization experiments (Fig. 4a–d,g). A high level of XPH1 transcript was localized in the medullary layer where the fungus forms aerial hyphae with hydrophobic cell wall surfaces. In the conglutinate cortical cells, which are embedded in a hydrophilic mucilage, XPH1 expression was reduced or absent. Only slight expression was detected in fungal cells of the algal layer (Fig. 4a) where the rodlet layer had been ultrastructurally resolved (Scherrer et al., 2000). No expression was observed in conidiomata (Fig. 4g) and in the hymenial and subhymenial layers of apothecia (Fig. 4d). However, a strong signal was evident in the aerial hyphae of the medullary layer of the thalline margin (margo thallinus) around the apothecial disc, and XPH1 transcripts were also found at the transition zone between subhymenial and algal layer, or at the periphery of pycnidia where the hyphae switch from dense conglutinate to aerial growth (Fig. 4d,g). Control experiments with the sense XPH1 RNA probe resulted in only a slight background staining (Fig. 4b). No signal was found in negative controls with identical treatment, but with no probe added to the hybridization buffer (Fig. 4c).

Figure 4.

Spatial distribution of the XPH1 transcript in the different layers of the stratified symbiotic phenotype of X. parietina (uc, upper cortex; ph, photobiont layer; m, medulla; lc, lower cortex). (a–d,g) Light micrographs of in situ hybridization analyses; the same magnification is used in a–c. (e,f) Low-temperature scanning electron micrographs of a cross-fractured, frozen-hydrated apothecium with thalline margin. (h–i) Scanning electron micrographs of specimens in which the hydrophilic mucilage of the pycnidial cavity had been removed prior to conventional preparation that included chemical fixation, dehydration and critical point drying. Cross-sections of the vegetative thallus (a–c) were hybridized with either the antisense probe (a), the sense probe (b) or without probe (c). Note the strong signal in the medullary layer in (a). Apothecia of Xanthoria parietina are shown in (d–f), with the hymenium (h) containing the asci, paraphyses embedded in the gelatinous matrix, the conglutinate subhymenium (sh) adjacent to the thalline margin with the photobiont (ph) and medullary layers (m) with hydrophobic cell wall surfaces and the lower cortex (lc) as in vegetative lobes (e,f) (PH, photobiont cell). In the in situ hybridization experiments, no transcripts were found in the hymenium, but XPH1 was highly expressed in the transition zone between subhymenium and photobiont layer (see f), in the medulla and the fungal cells of the photobiont layer (d). No staining reaction was detected in the corresponding negative controls as in (b) and (c) (not shown). Pycnidia are shown in (g–i). Note the strong signal in medullary hyphae (asterisk) adjacent to the pycnidial wall (pw). No signal was detected on either the pycnidial wall, the interconnected system of conidiophores (cp) in the pycnidial cavity (py), or on the minute, rod-shaped conidia (arrows in i).

In aposymbiotic culture on agarized media, Xanthoria parietina forms cartilaginous, thallus-like colonies with conglutinate, pseudoparenchymatous cells and few aerial hyphae at their periphery (Honegger, 1996). No XPH1 expression was detected in either cell type (not shown).

Uniform staining throughout all layers of the thallus, indicating uniform access of the probe to the different cell types, was found in control sections hybridized with a SSU rRNA fragment (not shown).

Discussion

Properties of the XPH1 gene

The XPH1 gene was the first hydrophobin-encoding gene from a lichen-forming fungus in databases. The encoded protein, Xanthoria parietina hydrophobin 1 (XPH1), belongs to the class I hydrophobins, as determined by molecular and biochemical characteristics (Scherrer et al., 2000). By screening a λ phage library, constructed with DNA which was isolated from aposymbiotically cultured Xanthoria parietina, the entire XPH1 hydrophobin gene, flanking regions included, was isolated. So far, nothing is known about the genome size of lichen-forming fungi. Genome sizes of nonlichenized filamentous fungi range over 107–108 bp per haploid genome (Gurr et al., 1987). The number of positive clones in the screening procedure provides indirect evidence that our genomic bank covers, with high probability, the entire genome. Because the library was constructed with the DNA of aposymbiotically cultured Xanthoria parietina, it can be used for a wide range of future investigations on the biology of this lichen-forming ascomycete; no DNA of the green algal photobiont was included.

XPH1 shares the characteristics of protein-encoding genes from nonlichenized filamentous fungi. Most contain introns that are shorter than mammalian and plant introns, the average being approximately 70 bp (Unkles, 1992). The encoded XPH1 protein carries a leader sequence for secretion at its N-terminus. In previous biochemical investigations XPH1 was shown to be extracellularly located, since it was isolated from washed cell wall fragments (Scherrer et al., 2000). Discrepancies between the cleavage site predicted by computer analysis and the N-terminal sequence of the mature protein became also evident in other studies: five amino acids difference was detected in HCF-1 of Cladosporium fulvum (Spanu, 1997). In other C. fulvum proteins, post-translational processing was shown to occur in several steps (van den Ackerveken et al., 1993).

The core promoter elements of XPH1 were identified by sequence homology to corresponding elements in nonlichenized filamentous fungi. The identification of the promoter region of this gene might be the basis for additional investigations. Expression studies with reporter genes under the control of the XPH1 promoter might be a challenge for future experiments. Many routine molecular techniques, such as transformation, have yet to become established for lichen-forming fungi.

Hydrophobin gene expression

The XPH1 gene proved to be highly expressed in certain cell types of the symbiotic phenotype, indicating a constant need for the hydrophobin protein in the lichen symbiosis. In contrast, some hydrophobins of nonlichenized fungi are known to be expressed only during a limited time-window in order to play a role during developmental processes. Two examples are the following: the hydrophobin genes of ectomycorrhizal Pisolithus tinctorius, which are highly expressed during the early stages of root colonization, with an almost complete decrease after 2–3 d (Tagu et al., 1996); these proteins are unlikely to be involved in mantle formation and thus no longer needed for the maintenance of the symbiotic interaction; and MPG1, the hydrophobin gene of the rice blast fungus, Magnaporthe grisea, is highly expressed during adhesion to the hydrophobic host cuticle and again during conidiophore and conidium differentiation (Beckerman & Ebbole, 1996). By contrast, the continuous expression of XPH1 in medullary hyphae of growing and nongrowing areas of the lichen thallus strongly indicates that this protein is required for the maintenance of the symbiotic relationship.

No, or very low expression of XPH1 was detectable in aposymbiotically cultured Xanthoria parietina, both in liquid media and on cellulose or glass-fibre filters. No hydrophobin monomers could be recovered from liquid media. By contrast, large amounts of the hydrophobin SC3 are released by the monokaryon of Schizophyllum commune when growing in liquid media. Wösten et al. (1999) proposed that SC3 lowers the surface tension of aqueous solutions, which allows submerged hyphae to grow into the air; and covers emerging hyphae with a hydrophobic coat to prevent them from re-entering the liquid medium. One might speculate that XPH1 is upregulated only during symbiosis. However, the impact of different culturing conditions will have to be carefully tested in long-term experiments. In the plant-pathogenic fungi M. grisea and Trichoderma reesei, hydrophobin gene expression was shown to be significantly increased by N and C starvation (Talbot et al., 1993; Nakari-Setäläet al., 1997), and all of the six HCf genes of C. fulvum were upregulated under conditions of nutrient deprivation (Seger et al., 1999; Nielsen et al., 2001). By contrast, changes in C or N concentrations in the growth media had no effect on the expression of the three hydrophobin genes of the ectomycorrhizal basidiomycete P. tinctorius (Duplessis et al., 2001). In nature Xanthoria parietina is very nitrophilous and grows best at highly eutrophicated sites, growth rates of the symbiotic phenotype, expressed as radial increase of thallus rosettes, being 4–7 mm yr−1 (Honegger et al., 1996). Lichen-forming ascomycetes are very oligotrophic (i.e. require very low amounts of carbohydrates, and grow very slowly in axenic culture. Under strict N- and C-starvation a very long culturing period would be required to grow sufficient material for expression experiments. Transfer of mycelial mats of Xanthoria parietina after prolonged culturing in nutrient medium to N-depleted media, as performed in the present study, had no effect on hydrophobin gene expression. The main problem in experimental work with lichen-forming fungi is the near impossibility of resynthesizing morphologically complex symbiotic phenotypes routinely under sterile conditions. Thus, a large number of experiments, including XPH1 expression studies during the re-lichenization process, cannot yet be carried out.

The data from Northern blot analyses correlated very well with in situ hybridization studies of the symbiotic and aposymbiotic phenotypes of Xanthoria parietina. The weak XPH1 expression signal in hyphae of the algal layer was surprising since this was the site where a rodlet layer could be resolved in freeze-etch preparations (Scherrer et al., 2000). However, in in vitro self-assembly tests the protein concentration of the solution proved to be crucial. The hydrophobin proteins DGH1, DGH2 and DGH3, upon extraction from the lichenized basidiocarps of the tropical basidiolichen Dictyonema glabratum, formed distinct rodlets only at relatively low concentrations (Trembley et al., 2002a) as was observed in the SC3 hydrophobin of the non-lichenized basidomycete schizophyllum commune (Wösten et al. 1994). Similarly, XPH1 formed rodlets at low hydrophobin concentrations, but a smooth layer at higher concentrations (S. Scherrer, unpublished). The wall surface layer of medullary hyphae, which creates a strong hydrophobic discontinuity along which the fracture plane was running over considerable distances, was smooth in freeze-etch preparations (R. Honegger, unpublished), medullary hyphae being the site of greatest XPH1 expression. It is very likely that hydrophobin monomers, upon secretion by medullary hyphae, are passively transported within the fungal apoplast to the adjacent algal layer, where they form aggregates with the characteristic rodlet pattern.

Potential functions of hydrophobin proteins in the lichen symbiosis

Although immunolocation studies have to ultimately prove that XPH1 forms the rodlet layer within thalli of Xanthoria parietina, we conclude that there is a strong possibility that the functionally important wall surface hydrophobicity in the medullary and algal layers of this and a range of closely related taxa of lichen-forming ascomycetes is hydrophobin-based. Lining air channels and preventing them from getting water-logged was also postulated to be the main function of the hydrophobins in fruit bodies of S. commune and Agaricus bisporus (Lugones et al., 1999). The internally stratified thalli of macrolichens, arguably the most complex vegetative structures in the fungal kingdom, are the product of long cohabitation of the partners involved. These thalli can be regarded as sophisticated culturing chambers for a population of green algal or cyanobacterial cells, with wall surface hydrophobicity being a key element for the functioning of the symbiotic interaction (Honegger, 1998). Similarly the lichenized fruiting bodies of the tropical basidiomycete D. glabratum (Sprengel) D. Hawksw. (syn. Cora pavonium (Web.) E. Fries) comprise a system of aerial hyphae with hydrophobic wall surfaces in the photobiont layer and at the basis of the hymenium, which prevent these areas from getting waterlogged. Three hydrophobins, DGH1, DGH2 and DGH3 (GenBank Accession numbers AJ320545, AJ320546 and AJ320547, submitted by M. Trembley), all differentially expressed, were characterized and a rodlet layer resolved in situ and in vitro. DGH1 was immunolocalized on the wall surfaces of the aerial hyphae in the photobiont layer (Trembley et al., 2002a,b). Further studies in taxonomically diverse groups of lichen-forming fungi are required to determine to what extent hydrophobins are generating wall surface repellency at different types of mycobiont–photobiont interfaces.

Acknowledgements

Our sincere thanks to our colleagues Robert Dudler, Ernst Freydl, Eric Van der Graaff, Thomas Wicker and David Chevalier for technical advice and for many stimulating discussions, to Jean-Jacques Pittet for his competent help with the artwork, and to the Swiss National Science Foundation (grant No. 31-52981.97 to R.H.) for generous financial support.

Ancillary