• The expression of hydrophobin genes DGH1, DGH2 and DGH3 in the lichenized fruitbody of Dictyonema glabratum (syn. Cora pavonia) is reported here, as well as immunolocalization of the DGH1 protein.
•In situ hybridization was conducted to determine expression patterns of hydrophobin genes DGH1, DGH2 and DGH3. The DGH1 protein was expressed in Escherichia coli and used to raise antibodies for immunolocalization of the DGH1 protein.
•DGH1 and DGH2 are expressed by hyphae in the photobiont layer and the lower stratum. DGH3 is newly expressed by hyphae in the boundary layer. DGH1 is immunolocalized in the electron-dense outer layer of fungal walls lining gas-filled spaces in the photobiont layer, in aerial hyphae in the gas-filled lower stratum and in the transition zone between the two strata.
•Dictyonema glabratum hydrophobins surround wall surfaces in the photobiont layer of the fruitbody. Hydrophobin layers probably function to keep interhyphal spaces water-free, seal the apoplast to enable water translocation, and define strata within the basidiocarp.
Lichen-forming fungi obtain carbohydrates and in some cases fixed nitrogen through symbiotic association with photoautotrophic partners (i.e. cyanobacteria or green algae or both). Lichen-forming fungi have polyphyletic origins (Hawksworth & Hill, 1984; Gargas et al., 1995; Lutzoni et al., 2001) and occur primarily in the Ascomycotina of which 13 250 species (or 46%) are lichenized (Hawksworth, 1988). Only about 50 species (0.3%) of more than 16 000 basidiomycetes are lichenized. In contrast to most ascolichens, basidiolichens form fruiting bodies that are similar to those made by nonlichen-forming basidiomycetes.
The lichen-forming basidiomycete Dictyonema glabratum (Sprengel) D. Hawksw. (syn. Cora pavonia (Web.) E. Fries) in symbiosis with the cyanobacteria Scytonema sp. is the only known example of a basidiolichen in which the fruitbody itself is lichenized. It develops a large and internally stratified basidiocarp that grows from a pseudomeristematic marginal curl (Fig. 1a). The fruitbody produces a hymenium that eventually covers the lower surface of the older regions (Fig. 1b). Photosynthetic rates and annual carbon gain in Dictyonema glabratum are higher than in other lichens, indicating that this symbiotic association is very efficient (Lange et al., 1994). The fungal uptake of photosynthates takes place primarily across a highly developed fungal haustorial–cyanobacterial plasma membrane interface (Roskin, 1970; Slocum & Floyd, 1977; Oberwinkler, 2001). Typical of all lichens, there are no specialized structures in the fruitbody for water allocation and gas exchange, which are both essential for photosynthesis, nitrogen fixation and respiration. Fruitbodies of Dictyonema glabratum, which grows in the tropics, undergo daily desiccation and hydration cycles in which water content (% dry weight) may fluctuate from less than 10% to 1000% (Lange et al., 1994). This presents a major challenge in maintenance of the symbiotic phenotype.
Hydrophobins are secreted fungal proteins that function in several capacities, including fungal water relations. Hydrophobins were first characterized in the nonlichenized basidiomycete Schizophyllum commune (Schuren & Wessels, 1990; Wessels et al., 1991) and have since been isolated from numerous fungi. Class I hydrophobins have the inherent ability to self-assemble at interfaces into amphipathic films that have a rodlet pattern on the hydrophobic side (Wösten et al., 1993; Wösten et al., 1994a; Wessels, 1997). Once assembled, hydrophobin membranes are highly stable and insoluble in boiling sodium dodecyl sulphate (SDS) solutions but can be rendered into monomers with trifluoroacetic acid (TFA) (Wessels et al., 1991). Hydrophobin layers function in various aspects of fungal development such as differentiation of aerial hyphae, asexual sporulation and fruitbody formation (Wessels, 1994, 1997). Hydrophobin layers are also involved in mediating attachment to surfaces (St Leger et al. 1992; Wösten et al., 1994b; Kershaw & Talbot, 1998) and in early ectomycorrhizal development (Martin et al., 1999a). Hydrophobins are assumed to be involved in creating hydrophobic wall layers in lichenized fungi (Honegger, 1997). Hydrophobins have been characterized from Dictyonema glabratum (Trembley et al., 2002) as well as from two lichen-forming ascomycetes (Scherrer et al., 2000).
The three hydrophobins DGH1, DGH2 and DGH3 have been cloned from the fruitbodies of Dictyonema glabratum (Trembley et al., 2002). The proteins are small (93–108 amino acids) and are 54–66% identical. They each contain eight cysteines that are present in the conserved pattern typical of hydrophobins and the hydropathy profiles of the polypeptides are consistent with class I hydrophobins (Wessels, 1997). When aqueous solutions of the hydrophobin extract are air-dried, the proteins self-assemble into a rodlet-patterned layer. A similar rodlet mosaic is found on the surface of hyphae lining interhyphal gas-filled spaces in the photobiont layer of the fruitbody (Trembley et al., 2002) and is thought to be formed by hydrophobins.
The aims of the present study were to investigate the expression of hydrophobin genes DGH1, DGH2 and DGH3 in the fruitbody of Dictyonema glabratum and to immunolocalize DGH1, the most abundant of the hydrophobins, with antibodies raised against a recombinant DGH1 protein. The possible involvement of hydrophobin layers in the maintenance of gas-filled interhyphal spaces, in creating an apoplastic continuum into which free water is forced and in defining strata within the fruitbody is discussed.
Materials and Methods
Basidiocarp collection and storage
Dictyonema glabratum was collected at an altitude of 1500 m in Parque nacional Tapantí (Provincia Cartago), Costa Rica, in January and February 1998. Basidiocarps were dried in the field and stored dry during shipment. They were then cleaned of adhering debris and stored at –20°C.
Hydrophobin protein extraction
Hydrophobins were isolated from fruiting bodies of Dictyonema glabratum using previously described hydrophobin extraction procedures (Wessels et al., 1991; de Vries et al., 1993) with slight modifications. Dry fruitbodies (2–4 g) were put into liquid nitrogen and ground to a fine powder. Homogenated fragments were extracted with a buffer (0.1 m Tris-HCl, pH 8.0, 10 mm MgSO4, 1 mm phenylmethylsulphonyl fluoride to remove soluble proteins and the cell walls were pelleted by centrifugation. This was repeated three times, at 4°C. The pellet was then extracted twice 10 min at 100°C in a buffered SDS solution (0.05 m Tris, pH 6.8, 2% SDS), followed each time by centrifugation. The pelleted material, containing SDS-insoluble proteins, was further extracted by heating at 65°C in five changes of chloroform–methanol (2 : 1), each time followed by centrifugation. The resulting pellet was air dried and suspended on ice in cold 100% TFA for 1 h. This slurry was sonicated in an ice-cold water bath three times for 30 s. Undissolved fragments were removed by centrifugation, the TFA supernatant was evaporated under a stream of filtered air and the pellet was resuspended overnight in 10 mm Tris, pH 8.0. Insoluble material was removed by centrifugation and the supernatant, containing SDS-insoluble but TFA-soluble proteins, was dialysed overnight at 4°C against two changes of H2O (Spectra/Por, 3500 MWCO, Serva, Heidelberg, Germany).
Dialysed protein extracts were centrifuged at 4°C for 10 min at 14 000 g to remove particulate matter. The supernatant was subjected to vigorous aeration in order to provide an interface upon which proteins could aggregate. Aggregated proteins were collected by centrifugation. The pellet was resuspended in TFA in order to render aggregates into monomers. The TFA was then evaporated and the proteins were then dissolved in water for at least 1 h. Protein solutions were analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) using either 15% or 18% polyacrylamide gels. Proteins were stained with either silver (Blum et al., 1987) or 0.1% (w/v) Coomassie Blue R250 (Neuhoff et al., 1988).
Production and purification of recombinant proteins
The cDNA sequence (EMBL accession number: AJ320544) encoding the mature DGH1 protein (i.e. without the secretion signal peptide) was amplified by polymerase chain reaction (PCR) with Pwo polymerase (Boehringer Mannheim, Mannheim, Germany) using two sequence-specific primers that contained three extra bases plus restriction sites BamHI at the 5′-end (EDGH-1 sense) 5′-CGCGGATCCACGACTCCTAAGCCTCC (this, and all other oligos are from Microsynth, Balgach, Switzerland) and HindIII at the 3′-end (EDGH-2 antisense) 5′-GGGAAGCTTTAGTTTGCTATAGTACCGC. The PCR product was digested with BamHI and HindIII and cloned into the expression vector p6xHis-DHFRS(0) (Stüber et al., 1990) such that the DHFRS of the p6xHis-DHFRS(0) vector is cut out of the final product and the histidine tag is upstream of DHG1. The expression plasmid was transformed into Escherichia coli M15 containing pRep4 (Stüber et al., 1990). Cells were selected for the presence of both plasmids using 100 µg ml−1 ampicillin (Amp) and 25 µg mml−1 kanamycin (Kan). The presence and in-frame sequence of the recombinant insert was verified by PCR and sequencing, respectively.
Escherichia coli cells containing the expression plasmid were grown according to the methods of Stüber et al. (1990) with some modifications to account for the assumed toxicity of the hydrophobin to the cells. Overnight cultures of cells grown at 37°C in 2× tryptone-yeast extract medium (TY) medium containing Amp and Kan were subcultured and grown to an optical density at 600 nm (OD600 nm) of at least 0.8. Cultures were then induced by 2 mm isopropylthiogalactoside (IPTG) for 3 h. Bacteria were harvested by centrifugation and lysed in buffer A (6 m guanidine-HCl, 10 mmβ-mercaptoethanol (β-ME) and 0.1 m NaH2PO4, 0.01 m Tris, pH 8.0). The lysate was loaded onto a column containing 1 ml of nickel-charged nitrilotriacteic acid (Ni-NTA) slurry resin (Hochuli et al., 1987). The column was washed with 10 ml buffer A, followed by 10 ml buffer B (8 m urea, 10 mmβ-ME and 0.1 m NaH2PO4, 0.01 m Tris, pH 8.0) and 10 ml buffer C (same as buffer B but pH 6.3). The recombinant protein was eluted with 10 × 1 ml buffer D (same as buffer B but at pH 5.9) and 10 × 1 ml buffer E (same as buffer B but at pH 4.5). Aliquots of buffers D and E were analysed by SDS-PAGE for the presence of the recombinant protein using 15% acrylamide gels, which were stained with either silver (Blum et al., 1987) or 0.1% (w/v) Coomassie Brilliant Blue R250 (Neuhoff et al., 1988). To remove urea and provide nondenaturing conditions, fractions containing the recombinant protein were pooled and dialysed overnight at 4°C in 4 m urea in phosphate-buffered saline (PBS) pH 7.5, then changed to 1 m urea–PBS, pH 7.5, for 4 h, followed by PBS, pH 7.5, alone for 4 h.
N-terminal sequencing of fusion protein
Purified Ni-NTA extracts were separated by SDS-PAGE. The 6xHis-DGH1 protein migrated at 15 kDa and was electroeluted from the gel using standard procedures (Bio-Rad Laboratories, Hercules, CA, USA). The N-terminal sequencing was carried out at the Protein Analysis Unit, University of Zürich, Switzerland.
In vitro rodlet formation
For self-assembly of the recombinant DGH1 Ni-NTA purified 6xHis-DGH1 was electroeluted from a polyacrylamide gel using standard procedures (Bio-Rad Laboratories), dialysed against H2O and 10 µl of the solution was placed on a Parlodion-coated copper transmission electron microscopy (TEM) grid and left to dry.
Production and purification of antiserum against 6xHis-DGH1
Dialysed, Ni-NTA-purified recombinant 6xHis-DGH1 protein solution was lyophilized and a total of 800 µg was used to raise polyclonal antisera in two rabbits (Eurogentec Bel SA, Herstal, Belgium). An initial immunization with Freund’s complete adjuvant was followed by boosts at 14, 28 and 56 d.
Purification of the antiserum from antibodies binding E. coli proteins was carried out using methods described by Ásgeirsdóttir (1994). Total protein extracts (1 mg ml−1 PBS, pH 8.0) of IPTG-induced cultures of empty M15(pRep4) E. coli cells were immobilized onto four polyvinyldifluoride (PVDF) membranes (Millipore, Bedford, MA, USA) (60 cm2 each) by incubation for 1 h. Membranes were blocked for 4 h with 5% skimmed milk in PBS, pH 8.0, and were then sequentially incubated for 1 h each in the antiserum (diluted 100× in PBS, pH 8.0, with 0.02% NaN3). The purified antiserum was used to detect DGH1 in Western blotting and immunocytochemistry.
Recombinant and native proteins (isolated using hydrophobin extraction methods) (Wessels et al., 1991; de Vries et al., 1993; Trembley et al., 2002) were separated by SDS-PAGE and electroblotted onto a nitrocellulose membrane (Bio-Rad Laboratories) by semidry transfer blotting (Multiphor II, Pharmacia LKB, Uppsala, Sweden). The membrane was placed overnight in a blocking solution containing 5% (w : v) skimmed milk powder in Tris-buffered saline (TBS), 0.05% Tween 20, pH 8.0 (TBST). The blot was incubated for 1 h with anti-DGH1 antiserum diluted 1 : 1000 in TBST and washed 2 × 10 min in TBST. Blots were incubated for 45 min in alkaline phosphatase-conjugated goat antirabbit IgG antibodies (Boehringer Mannheim) diluted 1 : 2000 times in TBST and washed again 2 × 10 min in TBST. Alkaline phosphatase was detected with the substrate BCIP (5-bromo-4-chloro-3-indolyl-phosphate p-toludine salt) and NBT (nitrotetrazolium blue chloride) (Sigma, St Louis, MO, USA) according to Harlow & Lane (1988).
In situ mRNA hybridization
Sense and antisense RNA probes labelled with 11-digoxigenin-UTP were synthesized by in vitro transcription using T7 (New England Biolabs, Beverly, MA, USA) or SP6 (Boehringer Mannheim) polymerase, according to the manufacturer’s instructions (Roche Diagnostics, Basel, Switzerland). pGEM-T Easy plasmids (Promega, Madison, WI, USA) containing 117 bp, 105 bp and 152 bp fragments specific for DGH1, DGH2 and DGH3, respectively (EMBL accession numbers AJ320544, AJ320545 and AJ320546, respectively) (Trembley et al., 2002) were linearized with restriction enzymes cutting in the polylinker and 1 µg of each was used as a template for probe synthesis.
In situ hybridization experiments using sense and antisense versions of the probes were carried out using protocols established for plant tissues (Jackson, 1991; Vielle-Calzada et al., 1999) with minor modifications. Dictyonema glabratum sections cut from the marginal edge and from 3 to 4 mm behind the edge were fixed in FAA (3.7% formaldehyde, 5% glacial acetic acid, 50% ethanol), dehydrated and embedded with Paraplast (Sigma). Sections 8–10 µm thick were fixed to Vectabond-covered slides (Merck, Darmstadt, Germany). Hybridization was carried out overnight at 55–65°C. For immunological detection, slides were incubated overnight with Western blue (Promega,). After stopping the reaction with 1× Tris-EDTA buffer (TE), slides were dehydrated through an ethanol series and mounted in 50% glycerol. The procedure was repeated twice for each probe. In each treatment, one to three slides were used. Each slide contained 10–20 sections originating from one to three different basidiocarps. In order to elucidate basidiocarp structure, sections of fruitbodies that had been embedded in Unicryl (British Biocell International, Cardiff, UK) were stained with basic fuchsin and methylene blue. All sections were examined in a Zeiss Axioplan microscope under bright field or DIC optics.
Small pieces excised from the edge and 3–4 mm behind the edge of fruiting bodies were chemically fixed in 1.5% acrolein (Fluka, Buchs, Switzerland) and 1.25% glutaraldehyde in 0.03 m phosphate buffer, pH 7.1, for 4 h, followed by overnight fixation in 2% buffered osmium tetroxide. Specimens were dehydrated in an acetone series and then infiltrated with a 1 : 1 Epon-Spurr mix (Plano, Wetzlar, Germany), sandwiched between Teflon-coated microscopy slides and polymerized at 60°C. Thick sections, 90–100 nm, were mounted onto TEM copper grids. Immunolocalization was carried out according to methods provided by British BioCell International. Grids were incubated for 20 min in buffer I-A (1% bovine serum albumin (BSA), 500 mm NaCl, PBS, pH 8.2). Sections were then incubated for 1–2 h in anti-DGH1 antiserum diluted 100–200× in buffer I-A. Sections were washed five times for 5 min with buffer 1-A then incubated for 1 h with goat antirabbit IgG antibodies conjugated with 30 nm gold particles (British BioCell) diluted 10–20× in buffer 1-A. Control reactions were conducted under the same conditions using preimmune serum instead of the anti-DGH1 antiserum or using the gold-conjugated secondary antibody (goat antirabbit IgG) alone. Sections were stained with uranyl acetate and Sato’s lead solution prior to examination in a Hitachi H7000 TEM at 50–60 kV. The procedure was repeated three times, using two to four grids for each treatment. Each grid contained two to four sections originating from two or three different basidiocarps.
Transmission electron microscopy and shadowing of rodlets
Solutions of protein extracts (10 µl) were placed on Parlodion-coated copper TEM grids and left to dry. Surface shadowing with platinum and carbon was done at an angle of approximately 45° in a Balzers BA 360 freeze-etch apparatus (Bal-Tec, Balzers, Liechtenstein). Specimens were examined in a Hitachi H7000 transmission electron microscope at 60 kV.
Low-temperature scanning electron microscopy
Low-temperature scanning electron microscopy (LTSEM) studies were conducted using the nondedicated Bio-Rad SP 2000 A cryotrans system, interfaced with a Hitachi S4000 scanning electron microscope. Basidiocarps were rehydrated overnight and small pieces were excised and fixed onto specimen holders using white paper glue. Specimens were plunged into subcooled liquid nitrogen, fractured with a cold rod and sputter-coated with a gold–palladium alloy. Specimens were maintained at –170°C and examined in the SEM at 20 kV.
Scanning Electron Microscopy
Small pieces of desiccated fruitbodies were sputter-coated with gold and were examined in the scanning electron microscope at 20 kV.
In situ localization of DGH1, DGH2 and DGH3 mRNA in fruitbodies of Dictyonema glabratum
The basidiocarps formed by Dictyonema glabratum are dorsiventral and have a pseudomeristematic rim which unfolds such that the fruitbody grows concentrically. Strata are not yet differentiated in the marginal, curled pseudomeristem (Fig. 2a) but become stratified about 1 mm behind the curl (Fig. 2b). Fully stratified regions are situated from 1 to 3 mm behind the curl. Antisense mRNA probes transcribed from cDNA sequences shown by Southern blot analysis to be specific for each of the DGH1, DGH2 and DGH3 genes (Trembley et al., 2002) were used to detect mRNA transcripts in the curl, in the younger stratifying and in the older stratified regions of Dictyonema glabratum. All three genes showed high expression levels in hyphae that surround the photobiont clusters in the centre of the curled margin (Fig. 2c–e) and, with the exception of DGH2, showed relatively lower levels in the loose, aerial hyphae at the periphery of the curl. DGH1 and DGH2 had higher expression levels overall compared with DGH3 (Fig. 2c–e). In the regions just behind the curl, where the fruitbody is starting to stratify, expression of all three transcripts is high in the hyphae of the differentiating photobiont layer and lower stratum (Fig. 2f–h). In the stratified region of the basidiocarp the expression patterns of DGH1 and DGH2 were not different. Transcript levels were high both in mycobiont hyphae that surround the photobiont clusters and in aerial hyphae in the uppermost parts of the lower stratum (Fig. 2i,j). By contrast, the expression pattern of DGH3 in this region changes completely from that observed in the younger regions. Expression in the photobiont layer and lower stratum stopped but hyphae of the boundary layer and hyphae in a cluster attached to the hymenium show strong DGH3 expression (Fig. 2k). Boundary hyphae per se and the hymenial-attachment hyphae are not visually detectable in the younger regions because the hyphae become differentiated later. Throughout the fruitbody, hydrophobin gene expression was not detected in the cyanobacterial colonies or in the hyphae of the upper stratum. Limited staining in the hymenium was inconsistently detected but, in general, there was no expression of any of the genes in this region. Negative control reactions with sense probes of DGH1, DGH2 and DGH3 transcripts showed no staining in the older regions (Fig. 2l–n) or in the younger regions or the curl (data not shown). The different strata are indicated in Fig. 2m.
Expression of DGH1 in E. coli and antibody production
A DGH1 fusion protein lacking the secretion signal peptide and containing an N-terminal histidine tag was expressed in E. coli, purified and used to raise polyclonal antibodies against DGH1. Proteins were extracted from bacterial pellets and were purified over a Ni-NTA affinity column under denaturing conditions. Analysis by SDS-PAGE revealed that a main band at 15 kDa, representing the recombinant protein, was eluted at pH 4.5 and pH 5.9. The fractions were pooled and dialysed together. The Ni-NTA purified 15 kDa protein was N-terminally sequenced and had a 10-residue N-terminal sequence identical to the expected sequence. The Ni-NTA purified fractions containing 6xHis-DGH1 were thus used for antibody production in rabbits.
Class I hydrophobins characteristically form rodlets when introduced to a hydrophilic/hydrophobic interface. The Ni-NTA-purified 15 kDa fusion protein was electroeluted from an SDS-PAGE gel and was air dried but failed to form a typical hydrophobin rodlet-patterned mosaic (data not shown). Total (native) hydrophobin and DGH1-enriched extracts isolated from Dictyonema glabratum form rodlet-patterned mosaics when dried (Trembley et al., 2002), yet when the DGH1 native protein was isolated from the total hydrophobin extract by electrophoresis, as was used for 6xHis-DGH1, it no longer formed rodlets as well (data not shown).
Western blot analysis with anti-DGH1 antiserum and bacterial protein extracts from the expression study showed that there was no strong recognition of any proteins prior to induction with IPTG (Fig. 3, lane 1). However, after induction, a protein of the expected size (15 kDa) was recognized by the anti-DGH1 antiserum (Fig. 3, lane 2). This protein was purified by Ni-NTA affinity chromatography (Fig. 3, lane 3). A 30 kDa bacterial protein that was coeluted during the affinity purification was also recognized by the anti-DGH1 antiserum (Fig. 3, lane 3). Thus, the antiserum was not highly specific and purification was necessary. In order to remove non-DGH1-specific antibodies, the antiserum was incubated several times with bacterial protein extracts that had been immobilized on membranes.
The anti-DGH1 antiserum recognized native proteins purified from fruitbodies of Dictyonema glabratum. Native hydrophobin protein extracts and SDS-soluble cell wall protein extracts were separated by SDS-PAGE and silver stained. The SDS-insoluble, TFA-soluble hydrophobin extracts as well as SDS-soluble extracts contained several proteins (Fig. 3, lanes 4 and 5). A Western blot with anti-DGH1 antiserum showed that DGH1 (14 kDa) was recognized, as was the 31 kDa DGH1 multimer (Fig. 3, lane 6). The 31 kDa protein is assumed to be a multimer of DGH1 because, previously, sequencing revealed that the N-termini of the two proteins were identical (Trembley et al., 2002). In addition, an 18 kDa protein was recognized by the antiserum, but recognition was relatively low compared with DGH1. The antiserum did not react with any of the other proteins migrating closely together at 6, 8 or 9 kDa (Fig. 3, lane 4). A Western blot of the total SDS-soluble protein extract showed that the anti-DGH1 antiserum did not recognize other types of cell wall proteins (Fig. 3, lane 7).
Immunolocalization of DGH1 in fruitbodies of Dictyonema glabratum
Polyclonal antibodies raised against the recombinant protein 6xHis-DGH1 were used to localize DGH1 in Dictyonema glabratum fruitbodies. DGH1 was visualized by incubation with gold-conjugated, goat antirabbit IgGs. Photobiont cells in the nonstratified, curled margin are surrounded by long, thin mycobiont hyphae and haustoria penetrate their cell walls (Fig. 4a). There was insignificant gold labelling in cells and cell walls of the mycobiont in this region (Fig. 4b). In the older, stratified regions, photobiont cells and the hyphae engulfing them are larger compared with those of the marginal curl (Fig. 4c). In the photobiont layer, DGH1 was immunolocalized in the cell walls of hyphae that face onto interhyphal gas-filled spaces (Fig. 4d). DGH1 was also detected in the electron-dense wall layer where hyphae abut an airspace (indicated by arrow in Fig. 4d). At the boundary between the photobiont layer and the lower stratum, gold labelling was observed in the confluent, electron-dense outer wall layer at hyphal/hyphal contact (indicated by arrow in Fig. 4e). In the lower stratum, gold labelling was observed in cell walls of the aerial hyphae, in the peripheral, electron-dense layer (Fig. 4f). In control reactions where preimmune serum was used instead of anti-DGH1 antiserum (Fig. 4g) or where gold-conjugated antibodies were used alone (not shown) no labelling was detected. A parallel experiment in which basidiocarp sections were fixed instead in Unicryl (British Biocell International, Cardiff, UK) and visualized with a 12-nm gold conjugate showed the same pattern of DGH1 localization as above (not shown).
Distribution of water and hydrophobins in stratified regions of Dictyonema glabratum
In order to visualize the possible role of hydrophobins in the functional anatomy of Dictyonema glabratum, ultrastructural, molecular and immunocytochemical results from the present and an earlier study were collated. Ultrastructural changes occur in the basidiocarp of Dictyonema glabratum when it is dehydrated or water-saturated. Mycobiont hyphae and photobiont cells shrink and collapse in desiccated basidiocarps (Fig. 5a), whereas in water-saturated fruitbodies all cells and their relatively thick cell walls are fully hydrated and turgid (Fig. 5b). Free water, however, is restricted from interhyphal spaces in the photobiont layer and lower stratum which remain gas filled in hydrated specimens (Fig. 5b). By contrast, interhyphal spaces in the upper stratum, boundary layer and hymenial layer become completely water-filled (Trembley et al., 2002). Water and hydrophobin distribution patterns are superimposed onto a diagrammatic interpretation of a hydrated basidiocarp (Fig. 5c). Interhyphal spaces that remain gas-filled are shown in white and those that become waterlogged in light grey. Mycobiont hyphae are shaded in medium and cyanobacteria in dark grey. Thick, black lines outline hyphae in which either hydrophobin genes are expressed or localized. The surface of hyphal–gas interfaces in the photobiont layer is covered by rodlets, as detected in freeze-etch preparations (Fig. 5d). Similar rodlets are formed by a hydrophobin extract containing either all hydrophobin proteins purified from Dictyonema glabratum (Fig. 5e) or primarily DGH1 (Trembley et al., 2002).
The present study describes the expression patterns of the hydrophobin genes DGH1, DGH2 and DGH3 in the lichenized fruitbody of Dictyonema glabratum. DGH1 was also expressed in E. coli as a fusion protein for antibody production. The anti-6xHisDGH1 antiserum was used to immunolocalize the DGH1 protein in Dictyonema glabratum.
DGH1, DGH2 and DGH3 hydrophobin genes are highly expressed throughout all differently aged regions of the Dictyonema glabratum fruitbody. In a similar manner, the hydrophobin gene XPH1 is highly expressed in the medullary layer in both young and old parts of the lichen-forming ascomycete Xanthoria parietina (Scherrer, 2000; Scherrer et al., 2002). Hydrophobins in lichens thus seem to be in demand in the initial and long-term development, differentiation and/or functioning of the symbiotic phenotype. All of the hydrophobin genes in Dictyonema glabratum have similar patterns of expression in the curl and younger regions of the basidiocarp but show differential expression in the older, stratified parts. It is possible that in young, developing parts the proteins work together synergistically to create hydrophobic surface layers, as proposed elsewhere (Stringer & Timberlake, 1995; Ásgeirsdóttir et al., 1998) or are functionally redundant.
In E. coli, DGH1 was expressed as a recombinant protein. The fusion protein failed to self-assemble into a rodlet mosaic. The reasons for this might be that the presence of the tag or the inability of bacteria to carry out the appropriate post-translational protein modifications (Stüber et al., 1990) interfered with polymerization. Self-assembly of a rodlet layer has not been reported for the two other histidine-tagged hydrophobins that were expressed in E. coli (Peñas et al., 1998; Tagu et al., 2001). It cannot be excluded that components such as polysaccharides (Martin et al., 1999b) may have an impact on the formation of rodlet layers. When the native DGH1 was isolated from the hydrophobin extract it also failed to form rodlets: thus, other components in the extract are also important.
The anti-6xHisDGH1 antiserum was used to immunolocalize DGH1 in Dictyonema glabratum. In the curl region, the detection of DGH1 is insignificant and thus lower than the levels expected from the in situ hybridization data. Reasons for this might be that fungal cell wall properties in the curl hinder accessibility, or protein accumulation is too low to allow detection. In the younger and older regions of Dictyonema glabratum, localization patterns of DGH1 are in agreement with the in situ hybridization results. DGH1 accumulates in fungal cell walls, which is in agreement with the presence of a signal peptide for secretion in the DGH1 gene (Trembley et al., 2002). DGH1 is concentrated in the electron-dense, outer cell wall layer of aerial hyphae and hyphae at hyphal–gas-filled interfaces in the photobiont layer and the lower stratum. In a similar manner, the hydrophobins SC4 and ABH1, produced in fruitbodies of the nonlichenized basidiomycetes S. commune and Agaricus bisporus, respectively, are localized at the surface of interior, gas-filled channels (Lugones et al., 1999). In Dictyonema glabratum, DGH1 is additionally localized at hyphal–hyphal interfaces in between two strata. Similarly, localization of the HydPt-1 hydrophobin at hyphal–hyphal interfaces was observed in early stages of the ectomycorrhizal Pisolithus tinctorius–Eucalyptus globulus interaction (Tagu et al., 2001). However, in contrast to DGH1, labelling was confined primarily to an electron-transparent cell wall layer. The DGH1 protein is localized in the photobiont layer of Dictyonema glabratum, where rodlet layers line air spaces. Hydrophobin extracts purified from Dictyonema glabratum self-assemble in vitro into similar rodlet layers. Similar to DGH1, SC4 and ABH1 are localized in wall surfaces where rodlet layers, which resemble the ones formed in vitro by extracts of the corresponding hydrophobins, are observed (Lugones et al., 1999). The rodlet mosaic in Dictyonema glabratum may be formed primarily by DGH1, but DGH2 may also be present.
The spatial distribution patterns of water and hydrophobins in the stratified region of Dictyonema glabratum were diagrammatically superimposed in order to elucidate the possible function(s) of hydrophobins (Fig. 5). The boundary layer, which is sealed by DGH3, is the switching point in water translocation within the fruitbody; free interhyphal water is forced to move into the interior underneath hydrophobic wall surfaces. A similar situation may occur on the lower side of the basidiocarp where DGH3 seals off the hymenial attachment strands. The photobiont layer is lined by DGH1 and DGH2 on the lower side and by the DGH3-lined boundary layer on the upper side. Thus, the apoplastic continuum of the photobiont layer is three-dimensionally sealed by hydrophobins. Hydrophobins in this capacity probably facilitate the passive, apoplastic translocation of water and nutrients from the hydrophilic periphery into and within the photobiont layer. Hydrophobin rodlet layers, that enclose the photobiont layer, probably function at the same time to keep interhyphal spaces water free during saturation, such that gas exchange, important for photosynthesis, nitrogen-fixation and respiration, can take place. DGH1 may also be involved in hyphal–hyphal attachment and/or fruitbody stratification.
In the fruitbody of S. commune, SC4 functions to keep internal channels gas filled. This was proven with SC4-deficient mutants whose air channels lack rodlets and fill up with water when the fruitbody is soaked (van Wetter et al., 2000). We speculate that the hydrophobins in Dictyonema glabratum fulfil similar roles. Ultimate proof of the functional role of hydrophobins in Dictyonema glabratum would require knock-out mutants. However, as is the case for all lichen-forming fungi, it is not possible to resynthesize routinely the symbiotic phenotype under laboratory conditions from aposymbiotic (separated) sterile cultures of Dictyonema glabratum and its cyanobacterial photobiont, which makes the testing of mutagenized, symbiosis-related genes impossible at this time.
Hydrophobins seem to be widespread or even ubiquitous in fungi. Assuming that polymerised hydrophobin layers function in Dictyonema glabratum like they do in fruitbodies of non-lichenized basidomycetes suggests interesting evolutionary aspects regarding its transition to lichanization. Is it possible that potential photoautotrophic partners simply colonized fruitbody primordia and developed in the sheltered environment within a fruitbody that, utilizing hydrophobin layers, had already been constructed in such a way that the challenges of water allocation and gas exchange had been resolved? Similarities in ultrastructure and the ways in which hydrophobins function in lichenized and nonlichenized fungi indicate that this may be true. Conversely, development of fungal structures such as haustoria and differentiation of unique strata suggest that morphological innovations in Dictyonema glabratum have taken place, even if the initial cohabitation may have required no modification at all.
Our sincere thanks to Annette Haisch, Verena Kutasi and Urs Jauch for excellent technical assistance, Jean-Jacques Pittet for help with the artwork, Dr Urs Wäspi for guidance and discussions on protein expression and purification, Joachim Kilian for collecting Dictyonema glabratum in the field and the Swiss National Science foundation for generous financial support (grant no. 31–52981.97 to R. H.).