Identification of a Hydrophobin Gene That is Developmentally Regulated in the Ectomycorrhizal Fungus Tricholoma terreum

Abstract

The symbiosis between ectomycorrhizal fungi and trees is an essential part of forest ecology and depends entirely on the communication between the two partners for establishing and maintaining the relationship. The identification and characterization of differentially expressed genes is a step to identifying such signals and to understanding the regulation of this process. We determined the role of hydrophobins produced by Tricholoma terreum in mycorrhiza formation and hyphal development. A hydrophobin was purified from culture supernatant, and the corresponding gene was identified. The gene is expressed in aerial mycelium and in mycorrhiza. By using a heterologous antiserum directed against a hydrophobin found in the aerial mycelium of Schizophyllum commune, we detected a hydrophobin in the symbiosis between T. terreum and its native pine host Pinus sylvestris. The hydrophobin was found in aerial mycelium of the hyphal mantle and also in the Hartig net hyphae, which form the interface between both partners. Interestingly, this was not the case in the interaction of T. terreum with a host of low compatibility, the spruce Picea abies. The differential expression with respect to host was verified at the transcriptional level by competitive PCR. The differential protein accumulation pattern with respect to host compatibility seen by immunofluorescence staining can thus be attributed at least in part to transcriptional control of the hyd1 gene.

Symbiotic interactions are highly dependent on communication between the partners for initiation and maintenance. One ecologically important symbiosis is the ectomycorrhiza formed between basidiomycete fungi and their host trees (for reviews, see references 4, 17, and 35). Some trees form associations with many different fungi, but some fungi interact with only one host plant. Tricholoma terreum usually is associated with pine (Pinus sylvestris), but not with spruce (Picea abies) (14, 34). Ectomycorrhiza formation is essential if Tricholoma is to produce mushrooms, which therefore are detected normally only under pine. This host specificity can be used to identify genes involved in interactions with the native hosts that are not expressed in interactions with nonnative hosts. Expressing these genes may be important for successful in vitro synthesis of mycorrhiza.

T. terreum forms mycorrhiza in dual culture with pines within 4 to 6 weeks and with spruce in 4 to 6 months (25). Both mycorrhizae form fully developed mantles and Hartig's nets (11), but the spruce mycorrhizae lack the fine palmetti structures and symbiotic short roots may also lack parts of Hartig's net.

Genes specific for mycorrhization probably encode proteins involved in nutrient transfer or structural components of fungal/host interaction. Hydrophobins are a class of fungal cell wall proteins involved in making cell-cell or cell-surface contact (13, 47, 49, 50, 53). They are essential for the growth of hyphae from a hydrophilic medium into the air. The hydrophobin layer is a sheet of rodlets up to 10 nm in diameter and 100 nm in length (49, 55) that surrounds the hyphae and can assemble in vitro.

Hydrophobins in the wood-rotting fungus Schizophyllum commune were first described when four hydrophobin genes were identified (22, 30, 41, 42). One of these, sc3, encodes the hydrophobin protein specific for aerial mycelium (32, 42, 43, 48, 56). The protein is excreted from the hyphal tip into the growth medium (33, 54). When it encounters an interphase between hydrophilic and hydrophobic surfaces, it self-assembles into highly ordered structures (7, 8, 53, 55). Wild-type aerial hyphae accumulate high levels of hydrophobin on their surfaces. This protein layer lowers the surface tension of water, and only then can aerial hyphae develop (48, 56). Deletion of the sc3 gene, which encodes a hydrophobin in S. commune, results in a colony with little or no aerial mycelium (40). Wild-type aerial hyphae accumulate high levels of hydrophobin on their surfaces. The Sc3 hydrophobin and the three fruitbody-specific hydrophobins of S. commune Sc1, Sc4, and Sc6 can be distinguished biochemically from the pathogenicity-linked class II hydrophobins, e.g., cerato-ulmin (46).

Hydrophobins have been detected in zygomycetes, ascomycetes, and multiple basidiomycetes (for reviews, see references 49 and 51). Different roles in the fungal life cycle have been attributed to hydrophobins, e.g., in cell wall integrity (Sc3 disruptants are osmolabile at the hyphal apices), lining air channels for gas exchange in lichens and fruitbodies, covering conidiospores and making them water repellent, as protectants against the mammalian immune system or plant defenses, and as elicitors of plant defense or promoting the contact of hyphae with plant surfaces in pathogenic interactions (for review, see reference 49). In ectomycorrhiza, hydrophobins also have been detected among the proteins corresponding to expressed sequence tags of Pisolithus tinctorius (26, 36, 37, 38). In a preliminary study, only hydrophobins in the hyphal mantle were detected (37). A more thorough investigation using immunogold labeling of the protein in ultrathin sections of mycorrhiza detected hydrophobins in the Hartig net where every fungal cell was covered by hydrophobin (39). The influence of a hydrophobin layer on nutrient exchange is unclear, as the function of the symbiotic tissue is not impaired by rendering the surface of the hyphal wall impenetrable to water (27, 39). However, hydrophobins do not cover every portion of every hypha, and even in assembled Sc3 protein layers, pores are visible that permit diffusion of oligosugar or even oligopeptide moieties (51, 53).

In a previous paper (25), we showed that compatible and incompatible T. terreum ectomycorrhizal interactions differed in the levels of protein that reacted with an antibody directed against the S. commune Sc3 protein. Our objectives in this study were to purify T. terreum hydrophobin from culture supernatant, to identify and analyze the corresponding gene(s), and to perform expression analyses of the gene(s). These studies allowed us to determine the specificity of the gene product(s) for mycorrhiza and to evaluate the regulation of the gene(s) in response to signals from the plant.

MATERIALS AND METHODS

Cultivation and purification of hydrophobin.

T. terreum strain KR 7216, from Karl-Heinz Rexer, Philipps-University Marburg, was used throughout the study. Cultures were maintained on Melin Nokrans medium (MMN; 19). Mycelia grown for hydrophobin production were inoculated into flasks of MMN and maintained as surface cultures without shaking for 1 month at 20°C with 12 h of light per day.

We extracted hydrophobin from 1-month-old cultures grown saprophytically. The mycelium was pelleted by centrifugation (30 min, 3,000 × g), as suggested by the purification procedure described for the hydrophobin of S. commune (52). Air was bubbled through the culture supernatant to induce the formation of stable hydrophobin aggregates that were pelleted by centrifugation (30 min, 7,000 × g). The pellet was redissolved in distilled water and again pelleted by centrifugation. The dried pellet was redissolved in ice-cold trifluoroacetic acid (2 h at 4°C) and centrifuged (10 min, 3,000 × g), and the pellet was dried under a constant stream of air for at least 2 h. The pellet was suspended in buffer (100 mM Tris-HCl [pH 8.0], 60% ethanol) and loaded onto a preparative denaturing protein gel (12.5%; Mini-Protean II; Bio-Rad, Munich, Germany) after denaturation for 10 min at 95°C with 1 volume of sample buffer (50 mM Tris-HCl [pH 6.8], 2% sodium dodecyl sulfate [SDS], 5% β-mercaptoethanol, 10% glycerol, 0.01% bromophenol blue). Separation on the gel was for 15 min at 80 V and 60 min at 140 V (25 mM Tris-HCl [pH 8.8], 192 mM glycine, 0.1% SDS). The silver-stained gel was documented with a video documentation unit (Bio-Rad). The major band at 23 kDa was blotted from the gel onto a polyvinylidene difluoride membrane and subjected to automated N-terminal amino acid sequencing using phenylthiohydantoin labeling (PTH analyzer; Applied Biosystems, Weiterstadt, Germany).

Cloning and sequencing of hyd1.

From the N-terminal amino acid sequence, we designed a set of degenerate primers (5′-CCICTICCXGGXGGITCIAAG[T/C]TA-3′; I = inositol, X = G, A, T, or C). We conducted reverse transcription-PCR (RT-PCR) on mRNA from the mycelium (DNA and RNA extraction) (45) with this primer and an adapter-oligo(T) primer (5′-GACTCGAGTCGACATCGA-T₁₇-3′) using Omniscript reverse transcriptase (Qiagen, Hildesheim, Germany). Hydrophobin messenger was amplified (35 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s) using the primer deduced from the N terminus and the adapter primer (5′-GACTCGAGTCGACATCGA-3′). The amplified cDNA fragment was cloned (into vector pCR; Invitrogen, Leeks, The Netherlands) using Escherichia coli DH5α (29) and was sequenced (LiCor; MWG, Eberswalde, Germany). Using the cloned fragment as a probe, a 3.6-kb EcoRI genomic DNA fragment was cloned, identified by colony hybridization, and sequenced (21). To amplify the entire cDNA, two primers were derived from the genomic sequence (hydstart, 5′-ATGTTCTCTAAAGTCGCTCTC-3′; hydstop, 5′-CAAGTTGATGGGAGAGCAGCC-3′). Sequence analyses and alignments were performed with CLUSTAL V, and standard procedures were performed (20, 29).

Expression analyses.

Northern blot analyses were performed under high-stringency conditions (29) with an 859-bp SalI restriction fragment covering the open reading frame as a probe.

Western blottings were performed with the purified fractions of protein, cell wall extracts (in 50 mM Tris-HCl [pH 6.8]-2% SDS after 10 min at 100°C), and crude cell extracts. Anti-Sc3 and anti-Sc4 antisera were kindly provided by J. G. H. Wessels (Groningen, The Netherlands).

Mycorrhiza synthesis.

Mycorrhizae were synthesized as previously described (5, 9, 25). Tree seedlings of pine (Pinus sylvestri) and spruce (Picea abies) were germinated from surface-sterilized seeds (Staatsklenge Nagold, Nagold, Germany) under sterile conditions on germination medium (5) and incubated like dual cultures under a day-night regime of 20 and 17°C with 70 and 80% humidity, respectively, and 14 h of light. After a 5-cm-long main root developed, seedlings were placed on top of a cellophane membrane in a petri dish containing medium (modified MMN; 19) covered with a sterile cellophane membrane. Leaves were placed outside the plate through holes made in the rim of the petri dishes by use of a hot steel probe under sterile conditions. The root was sandwiched with a second cellophane membrane. Plates were sealed with parafilm and incubated as described before. After secondary roots were visible, the fungus (macerated from a saprophytic culture or pieces cut from surface cultures) was inoculated onto the lower cellophane and the second membrane was replaced. The resealed petri dishes were replaced in the incubators, and the development of mycorrhiza was checked visually at regular intervals.

Microscopy and immunodetection.

The mycorrhiza were frozen in Tissue-Tec (Sakuro Finetek, Marburg, Germany) at −24°C for 1 h and then cut (2800 Frigocut-E; Leica, Wetzlar, Germany) in 10- to 40-μm-thick sections. Thin sections were mounted on poly-l-lysine coated slides and examined with a Zeiss Axiophot fluorescence microscope (Carl Zeiss, Jena, Germany) equipped with a CCD camera (Junior Spot; Visitron, Munich, Germany). Fungal and plant nuclei were stained for inspection with DAPI (4′,6′-diamidino-2-phenylindole) after mounting for microscopy in 0.1 M Tris-HCl (pH 8.0)-50% glycerol.

We used rhodamin-labeled anti-rabbit immunoglobulin G (IgG) and anti-tubulin antibody (Sigma, Munich, Germany) for immunofluorescence microscopy on cryomicrotome cuttings (see above) as described by Fischer and Timberlake (10).

Competitive PCR.

For competitive PCR, the hydstart and hydstop primers were used to amplify cDNA from mycorrhizal mRNA preparations and to amplify a larger, intron-containing genomic competitor fragment (0.01 to 100 pg) that was added to the reaction mixture. To ensure equal quantities of fungal transcripts for competitive PCR, a control amplification of rRNA/ITS was used (primers: NS1, GTAGTCATATGCTTGTCTC, and NS8, TCCGCAGGTTCACCTACGGA; denaturation, 3 min at 95°C, 30 cycles of 2 min at 95°C, 35 s at 50°C, and 30 s at 72°C, and final polymerization, 10 min at 72°C) (3, 24). This step amplified both fungal and plant sequences that were separated by restriction with MboI, thus allowing the quantification of the fungal RNA in the complete purified RNA preparation.

Nucleotide sequence accession number.

A 3.6-kb EcoRI genomic DNA fragment was sequenced, and the sequence was submitted to GenBank (T. terreum hyd1, accession no. AY048578).

RESULTS

Purification of hydrophobin from T. terreum and identification of hyd1 gene.

The hydrophobin from culture supernatant contained a highly enriched fraction of peptides that, when resolved by SDS-polyacrylamide gel electrophoresis (PAGE), had a major band at 23 kDa (Fig. 1), a size similar to that of the S. commune hydrophobin Sc3. We cut this band from the gel and sequenced the N-terminal amino acid end. The first amino acid was ambiguous, but the next seven amino acids were accurately identified ([T]PLPGGSS[Y]) and could be used to design a primer for PCR amplification of a corresponding gene fragment.

FIG. 1.

Purification of hydrophobin from culture supernatant of T. terreum. The trifluoroacetic acid extract of high-molecular-mass complexes from aerated culture supernatant showed a prominent band after SDS-PAGE at 23 kDa that was used for further characterization and N-terminal amino acid sequencing.

The RT-PCR fragment was approximately 400 bp in length. The full genomic sequence was obtained from a cloned 3.6-kb EcoRI fragment that was identified from Southern blot analysis with the PCR fragment as a probe. The genomic sequence has two introns of standard sizes (53 and 74 nucleotides) with consensus splice sites. Two putative transcription elements (CAAT and TATAA box) were identified in the genomic sequence, and a single methionine start codon was detected 18 amino acids upstream from the identified amino acid terminus of the mature and purified protein. The propeptide sequence was highly hydrophobic, as expected for a secretion signal of a cell wall structural protein. The sequence of the N-terminal peptide was different from that predicted by the genomic sequence. This difference (position 9 in the amino acid sequence) did not prevent amplification and probably resulted from ambiguity in the peptide sequence.

Expression of hyd1 and hydrophobin protein localization.

Transcript levels were investigated under different growth conditions. The cDNA of hyd1 was detected in vegetative mycelium, but not in fruitbodies (Fig. 2). High levels of transcript also were found in aerial hyphae and intermediate levels in mycorrhiza (data not shown).

FIG. 2.

Northern blot analysis of hyd1 expression in different mycelia of T. terreum. A SalI restriction fragment covering the hyd1 coding region was used as a probe to detect hyd1 mRNA (A) in hyphae after submerged culture of 1 week (lane 3), 2 weeks (lane 2), and 3 weeks (lane 4), in aerial hyphae (lane 1), and in RNA prepared from fruitbody tissue (lane 5). Staining of total RNA was used as a loading control (B).

Anti-Sc3 antibody cross-reacted with purified Hyd1 (Fig. 3), and in unpurified crude extracts only one major band reacted. In contrast, anti-Sc4 antibody did not react (data not shown). We used anti-Sc3 antiserum to localize Hyd1 of Tricholoma to aerial hyphae (Fig. 4A, B). The Hyd1 hydrophobin was found at the perimeter of the cell in an immunodetectable form, which is expected if it is found on the outside of the cell wall. This localization was found only in aerial hyphae.

FIG. 3.

Western blot analysis of cross-reactivity of anti-Sc3 antibody with T. terreum hydrophobin. Two independently purified hydrophobin preparations of T. terreum were used to examine cross-reactivity of anti-Sc3 antiserum. For a control, a similar preparation of Sc3 from culture supernatant of S. commune was used.

FIG. 4.

Immunofluorescence microscopy of hydrophobin Hyd1. Localization of the Hyd1 protein was achieved by immunofluorescence staining using anti-Sc3 antiserum on submerged and aerial hyphae of T. terreum. (A) The hydrophobin was detected at the perimeter of the cell in aerial mycelium. (B) No fluorescence was observed in substrate hyphae. The accumulation of hydrophobin in symbiotic tissue of T. terreum with pine was shown by staining with anti-Sc3 antiserum (C and E) in the hyphal mantle and the Hartig net of compatible mycorrhiza between T. terreum and pine (C to F). (G and H) Host specificity was shown in comparison to an interaction with the non-native host spruce, where hardly any staining was visible for the Hartig net. The mantle showed comparable levels of the protein regardless of compatibility. Bars, 50 μm.

We also used anti-Sc3 antiserum for in situ analysis of Hyd1 accumulation in symbiotic tissue. Hyd1 was found in both mantle and Hartig net of mycorrhizae formed with pine (Fig. 4C to F). Staining was prominent in the hyphal mantle but could also be detected around every cell within the Hartig's net. Mycorrhiza of different ages had similar staining, indicating that the expression did not change grossly during the symbiosis.

Host specificity.

T. terreum and spruce form an ectomycorrhiza with both mantle and Hartig's net. Hyd1 accumulated primarily in the mantle, with little or no staining detected in the Hartig's net (Fig. 4G and H).

The mRNA levels of hyd1 were measured for both the pine and the spruce mycorrhiza, and the mRNA levels were equalized based on the amounts of fungal internal transcribed spacer (ITS) sequences present. From 100 ng of RNA (containing both fungal and plant sequences), the ITS fragments were amplified and cut with MboI for differentiation between fungal and plant sequences. In the pine mycorrhiza (0.8 μg) and in the spruce mycorrhiza (0.42 μg) were fungal ITS sequences. Thus, we used a 1.9-fold excess of spruce mycorrhizal RNA for competitive PCR. The hyd1 transcript level was 8- to 10-fold higher in the pine than in the spruce mycorrhiza (Fig. 5). These results are consistent with the observed difference in protein accumulation and suggest transcriptional regulation of hyd1 with respect to the host tree as well as differences between aerial and submerged hyphae.

FIG. 5.

Competitive PCR for quantification of hyd1 transcripts in compatible versus low-compatibility ectomycorrhiza. Since the amounts of fungus were different in both interactions, the quantities of mycorrhizal RNA containing equal amounts of fungal RNA were first determined. Using 100 ng of compatible (top row) and 190 ng of low-compatibility (bottom row) RNA, a competitive PCR was performed, amplifying the transcripts for hyd1 present in the mRNA (lower bands) as well as a larger, genomic fragment (upper bands) added in defined amounts to the sample. Equal quantities for competitor and intrinsic mRNA signals were used to determine amounts of specific mRNA in total fungal RNA.

DISCUSSION

We purified a class I hydrophobin from culture supernatant of T. terreum. Based on the deduced amino acid sequence, the eight cysteine moieties were appropriately spaced and the hydrophobicity pattern was typical. We also observed the biochemical features expected for a class I hydrophobin: self-aggregation at air/water interphases, multiple high-molecular-weight bands in SDS gels, and a single band seen only after trifluoroacetic acid incubation.

The DNA sequence predicts a 108-amino-acid residue protein. The mature protein sequence begins with amino acid 19. The 18-amino-acid leader sequence is sufficiently hydrophobic to be involved in protein excretion, as seen with other hydrophobins (e.g., references 13 and 47, and also reference 12). The 23-kDa band recovered after trifluoroacetic acid treatment does not correspond to the deduced mass of 8,846 Da of the 90-amino-acid mature peptide encoded by the gene. Posttranslation modifications are possible, including O-glycosylation sites at Thr19 and Ser42. The Sc3 hydrophobin of S. commune has an N-terminal modification of 16 to 22 O-linked mannosyl residues (7). Similar glycosylation of Hyd1 seems likely since the anti-Sc3 antibody reacts primarily with the modified parts of the glycosylated Sc3 hydrophobin.

Hyd1 is expressed in vegetative hyphae and can be purified from culture supernatant, as can other class I hydrophobins (e.g., references 2, 23, 30, and 50). Hyd1 also is expressed in mycorrhiza with indications of host specificity. The hyd1 gene is differentially expressed in vegetative submerged, aerial, and fruitbody hyphae. In S. commune, hydrophobins with differential expression patterns are known (28) and the regulation occurs at the transcriptional level (31). The Sc3 hydrophobin in S. commune is regulated by the mating type genes (15, 16, 18, 44) and by nuclear spacing in a dikaryon (1, 32). Our T. terreum strain is dikaryotic, and the expression is similar to that of sc3.

Hyd1 is the major vegetative hydrophobin in T. terreum, as only one major band was detected in the culture supernatant of vegetatively grown dikaryotic mycelium. However, at least one more hydrophobin is expected in T. terreum, based on additional signals in Southern blot analyses. For S. commune and Agaricus bisporus, three fruitbody-specific hydrophobins and two fruitbody-specific hydrophobins, respectively, are known in addition to a vegetative hydrophobin (6, 23, 30). Sequence similarity among hydrophobins usually are limited, which might mean that the cross-hybridizing hydrophobin is allelic to hyd1 rather than being a representative of other hydrophobins with probable different function (compare A. bisporus; 23). To test this hypothesis, the two monokaryotic strains present in the dikaryon should be separated and investigated for their hydrophobin Southern banding pattern.

Hyd1 has the general structure necessary for the biochemical properties of a class I hydrophobin, with a consensus sequence of X_2-28(9/27)-C-X_5-9(6)-C-C-X_4-39(32)-C-X_8-23(13)-C-X_5-9(5)-C-C-X_6-18(12)-C-X_2-13(5) (C = cysteine, X = any amino acid; the subscripts are the ranges of the number of amino acids, and the numbers in parentheses are the numbers of amino acids for hyd1 from T. terreum; 46, 47). In sequence alignments using the deduced amino acid sequence, the sequence similarity is limited (up to 41% identity). To a typical hydrophobin from aerial mycelium, Sc3, similar levels of identity are detectable (38% amino acid identity), and the same is true for the hydrophobin HydPt1 from P. tinctorius, a hydrophobin for which a function in mycorrhiza has been described (39). Thus, hydrophobins are important not only in the aerial mycelium but also in the symbiotic tissues of the Hartig's net.

By using anti-Sc3 antiserum, Hyd1 was localized in aerial hyphae and found to be coating the symbiotic tissue in a manner resembling the coating of the aerial hyphae. In contrast, aggregated, fruitbody-specific hydrophobins have been found only on surfaces of fungal tissue, such as the lining of air channels in fruit bodies or lichens or on peel tissue. The coating of every hypha in a tissue that is expected to be involved in nutrient exchange between the symbionts was not expected. However, the same pattern was found in another ectomycorrhiza between P. tinctorius and eucalypt roots. Immunogold labeling clearly showed the same, single-cell coating occurrence of a specifically expressed hydrophobin in this mycorrhiza (39).

The T. terreum hyd1 gene expression pattern suggests that hydrophobins might be involved in host recognition and the specificity of the fungus for a specific host tree. Such differences could have a large impact on the ecosystem as the different mycorrhizal fungi share some host trees and may connect the trees in the woodland. The identification of a differentially expressed gene with host specificity will allow us to search for substances used in communication between the partners. This search has been problematic due to the slow and error-prone screening process, which involves in vitro synthesis of mycorrhiza. The detection of the hyd1 gene product may provide both a simpler and more sensitive screen. In addition, characterization of the promoter of the hydrophobin gene and the role of the protein in the development of compatible mycorrhiza will help understanding the regulatory mechanisms involved. Such experiments would require gene knock-out and overexpression studies that must await the development of a transformation system for Tricholoma.