Signal Transduction by Tga3, a Novel G Protein {alpha} Subunit of Trichoderma atroviride

Trichoderma species are used commercially as biocontrol agentsagainst a number of phytopathogenic fungi due to their mycoparasiticcharacterisitics. The mycoparasitic response is induced whenTrichoderma specifically recognizes the presence of the hostfungus and transduces the host-derived signals to their respectiveregulatory targets. We made deletion mutants of the tga3 geneof Trichoderma atroviride, which encodes a novel G protein ${alpha}$ subunit that belongs to subgroup III of fungal G ${alpha}$ proteins. ${Delta}$ tga3mutants had changes in vegetative growth, conidiation, and conidialgermination and reduced intracellular cyclic AMP levels. Thesemutants were avirulent in direct confrontation assays with Rhizoctoniasolani or Botrytis cinerea, and mycoparasitism-related infectionstructures were not formed. When induced with colloidal chitinor N-acetylglucosamine in liquid culture, the mutants had reducedextracellular chitinase activity even though the chitinase-encodinggenes ech42 and nag1 were transcribed at a significantly higherrate than they were in the wild type. Addition of exogenouscyclic AMP did not suppress the altered phenotype or restoremycoparasitic overgrowth, although it did restore the abilityto produce the infection structures. Thus, T. atroviride Tga3has a general role in vegetative growth and can alter mycoparasitism-relatedcharacteristics, such as infection structure formation and chitinasegene expression.

INTRODUCTION

Top

Introduction

Several fungi belonging to the genus Trichoderma can act asmycoparasites and are used commercially as biological controlagents against plant-pathogenic fungi, such as Rhizoctonia solani,Botrytis cinerea, Sclerotium rolfsii, Sclerotinia sclerotiorum,and Pythium spp. (5, 6, 18). Mycoparasitic strains can penetrateand kill the host fungi. Mycoparasitism is accompanied by thesecretion of antifungal metabolites, such as peptaibol antibiotics(32, 47), and morphological changes, such as coiling aroundthe host and development of hooks and/or appressorium-like structures(11). Hydrolytic enzymes, such as chitinases, glucanases, andproteases, also are important for biocontrol activity, sincethey enable Trichoderma to degrade the host's cell wall andto utilize its cellular contents (20). Chitinase gene expressionis induced by colloidal chitin, fungal cell walls, or the chitinmonomer N-acetylglucosamine (29). In Trichoderma atroviridethe N-acetylglucosaminidase (NAGase)-encoding gene, nag1, hasa major impact on the induction by chitin of other chitinases(4). In mycoparasitic interactions between T. atroviride andR. solani, expression of ech42, which encodes endochitinase42, is triggered by a low-molecular-weight diffusible factorreleased from the host prior to physical contact between thetwo fungi (10, 50). Lectins in the host's cell wall can inducecoiling of the mycoparasite around the host (1, 24, 41).

Elucidation of the signaling pathways underlying mycoparasitismhas only recently begun. Heterotrimeric G proteins are composedof ${alpha}$ , ß, and ${gamma}$ subunits. G ${alpha}$ subunits play pivotal rolesin the recognition process, virulence, and virulence-dependentdevelopment in a number of plant-pathogenic fungi (3, 15, 26,27, 31). Fungal G ${alpha}$ subunits are highly conserved and can be dividedinto three major subgroups (3). In T. atroviride, the subgroupI G ${alpha}$ protein Tga1 affects light-induced conidiation and mycoparasitism-relatedcoiling (41). The corresponding protein in Trichoderma virensis involved in antagonism against S. rolfsii but not in antagonismagainst R. solani. ${Delta}$ tgaA loss-of-function mutants sporulate andhave coiling behavior similar to that of the wild type (37),suggesting that the subgroup I G ${alpha}$ subunits have different functionsin these mycoparasites. The T. virens subgroup II G protein ${alpha}$ subunit, TgaB, has no specific role in development or mycoparasitism(37). Fungal subgroup III G ${alpha}$ proteins also regulate morphologicaland developmental processes, such as germination, conidiation,and secondary metabolite production (28, 53), and they may increaseintracellular cyclic AMP (cAMP) levels by stimulating adenylylcyclase (28, 33, 39).

The objectives of this study were to isolate a subgroup IIIG ${alpha}$ subunit deletion mutant of T. atroviride and to determineits role in development and mycoparasitism. This report is thefirst report of the role of the Tga3-mediated G protein pathwayin mycoparasitism-related properties, such as host recognition,chitinase expression, and secretion.

MATERIALS AND METHODS

Top

Materials and Methods

Strains and culture conditions.
T. atroviride strain P1 (formerly Trichoderma harzianum ATCC74058) was used for this study and was grown on potato dextroseagar (PDA) (Difco, Detroit, Mich.) at 28°C until sporulation.B. cinerea and R. solani were used as pathogens and were obtainedfrom the collection of the Institute of Plant Pathology, Universitàdegli Studi di Napoli Federico II, Naples, Italy. Escherichiacoli JM 109 was the host for plasmid amplification.

T. atroviride was grown in liquid synthetic medium (SM) as describedpreviously (4). For induction experiments, the fungus was precultivatedfor 36 h in SM containing 1% (wt/vol) glycerol as the carbonsource, harvested by filtration, washed with sterile tap water,and transferred to fresh SM containing 1% (wt/vol) colloidalchitin or 1% (wt/vol) N-acetyl-ß-D-glucosamine. Totest the influence of exogenous cAMP (Sigma, St. Louis, Mo.),cAMP was added to PDA cooled to 48°C after autoclaving.

Cloning of tga3.
Two products ( $~$ 1,000 and $~$ 1,100 bp) were obtained by PCR amplificationof genomic DNA from T. atroviride strain P1 with degenerateoligonucleotide primers based on conserved regions of G ${alpha}$ sequences(primers G ${alpha}$ F [5'-GGNGCNGGNGARWSNGGNAA-3'] and GaR [5'-GTRTCNGTNGCNYDNGT-3']).The products were subcloned into pGEM-T (Promega, Madison, Wis.)and sequenced. For screening of a genomic ${lambda}$ BlueStar libraryof T. atroviride P1, the tga3-containing PCR fragment was radioactivelylabeled with ³²P by random priming and used as a probe. Sequencingof isolated clones was performed by using the following internalprimers: Tga3intF1 (5'-CTGCTTCGAGAACGTTACCTC-3'; bp 881 to 901),Tga3intF2 (5'-CGATGGCATTTACGCGAAG-3'; bp 1632 to 1650), Tga3intF3(5'-GTATTAACTTGCTGCGCTCATTG-3'; bp 2476 to 2498), Tga3intR1(5'-GTATTCGCTCAGCGCAACAC-3'; bp 913 to 932), Tga3intR2 (5'-GACAATGGTCGACTTGCCAC-3';bp 220 to 239), Tga3intR3 (5'-CAGGGCTGGCAATGATGAAAG-3'; bp –486to –466), and Tga3intR4 (5'-CTGTGCCAGCCTAAAAGTGTGC-3';bp –1287 to –1266).

Fungal transformation.
Agrobacterium-mediated transformation of T. atroviride was carriedout as previously described (51). Briefly, after cocultivationof a conidial suspension (10⁷ spores ml^–1) for 24 h withan Agrobacterium tumefaciens strain containing the disruptionconstruct pTSZ- ${Delta}$ tga3, in which the entire tga3 coding regionwas replaced by an hph (hygromycin B phosphotransferase-encoding)expression cassette (51), fungal transformants were selectedon PDA containing 200 µg of hygromycin B per ml. Transformantswere recovered and purified to mitotic stability by repeatedtransfer to selective medium and by two rounds of single-sporeisolation.

DNA and RNA manipulations.
Genomic DNA was isolated as previously described (16); Southernhybridizations were carried out as described by Sambrook etal. (43). RNA was isolated as described by Chomczynski and Sacchi(7). Northern blotting of total RNA was performed by using BiodyneBnylon membranes (Pall Europe, Portsmouth, United Kingdom) andapplying 20 µg of RNA. For hybridization, the radioactivelylabeled PCR product containing most of the tga3 structural genedescribed above was used. For analysis of chitinase gene transcription,a 1.9-kb SalI/XbaI fragment of nag1, a 1.0-kb PstI fragmentof ech42, and a 1.9-kb KpnI fragment of act1 were used as probes.

Real-time reverse transcriptase PCR.
RNAs were incubated with DNase I (1 U/µg of RNA; Fermentas,St. Leon-Rot, Germany) according to the manufacturer's instructionsto remove any remaining chromosomal DNA. First-strand cDNA synthesiswas carried out with an oligo(dT)₁₈ primer (0.5 µg/µl),a random hexamer primer (0.2 µg/µl), 0.45 µgof total RNA, and RevertAid H Minus Moloney murine leukemiavirus reverse transcriptase (200 U/µl; Fermentas) accordingto the manufacturer's instructions. For amplification of an $~$ 60-bp fragment of the actin-encoding gene act1, real-time PCRwas carried out with primers act1F (5'-TATGTGCAAGGCCGGTTTCG-3')and act1R (5'-GGTCTTCCGACAATGGAGG-3'). A 370-bp nag1 fragmentwas amplified with primers nag1intF (5'-GCAAATGCGCTGTGGCCCATC-3')and nag1intR (5'-GCTTGAAGGTGGTCGCGGTATC-3').

Real-time PCR amplification was carried out with the iCyclerreal-time detection system (Bio-Rad, Hercules, Calif.) in a50-µl mixture containing iQ SYBR Green Supermix (Bio-Rad),each primer at a concentration of 100 nM, 5 µl of sample,10 nM fluorescein, and 1x SYBR Green. The act1 fragment wasamplified with the following PCR program: initial denaturationfor 90 s at 95°C, followed by 50 cycles consisting of 95°Cfor 15 s, 57°C for 20 s, and 72°C for 20 s. nag1 wasamplified by 50 cycles of 95°C for 15 s, 61°C for 20s, and 72°C for 20 s. Successful amplification was verifiedby determination of the melting temperature and by agarose gelelectrophoresis. A nonamplification control and a control withRNA without reverse transcription to exclude DNA contaminantswere performed in parallel for every assay. All determinationswere performed three times with two different sample dilutions.The results were displayed in the PCR base line subtracted mode,where the software (Bio-Rad) determined the cycle number atwhich the reaction reached the threshold.

The efficiency (E) of the real-time PCR was calculated by amplificationof two different dilutions of the samples by using the equationE = 10^1/tc (12), and the threshold cycles (tc) were amendedfor an optimum efficiency of 2. To compare different samples,the threshold cycle of the analysis with the primers for nag1was corrected with a factor for the act1 amplification. Thesample of wild-type T. atroviride P1 harvested 10 h after replacementwith glucose (Gluc/10h sample) was assigned the factor 1 becauseit had the highest threshold cycle. As 3.32 cycles theoreticallygives a 10-fold increase in the amount of PCR product, the differencebetween the threshold cycle of the wild-type Gluc/10h sampleand the corrected threshold cycle of the other samples was dividedby 3.32, yielding X. Inducibility (I) was calculated by usingthe equation I = 10^X.

Protoplasting and determination of growth inhibition by SDS.
Protoplasts were prepared from mycelia grown for 15 h on PDAcontaining cellophane (Maildor, Gournay en Bray, France) aspreviously described (16) with lysing enzymes from T. harzianum(Glucanex; Sigma). The inhibitory effect of sodium dodecyl sulfate(SDS) was determined by measuring the radial growth rates onplates containing PDA supplemented with 0.01 or 0.02% SDS andcomparing these growth rates to the growth rates on plates containingPDA without SDS. Hyphal elongation was measured twice a day,and the growth rate was determined.

Biocontrol assays.
Tests of inhibition of B. cinerea spore germination were performedas described previously (4). Inhibition by culture filtrateswas adjusted for the biomass present in the culture. Plate confrontationassays with R. solani or B. cinerea as the host were performedas previously described (32, 50).

Enzyme assays.
Chitinase activity was determined in culture filtrates, in culturebroth including mycelia, and in intracellular samples. Mycelium-boundactivity was determined with medium containing mycelium, whichwas removed by centrifugation at 13,000 x g at 22°C for5 min before spectrophotometric analysis. For determinationof intracellular enzyme activities, harvested mycelium was groundin liquid N₂ and suspended in 1.5 volumes of protease inhibitorcocktail containing 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonylfluoride, 1 µM pepstatin A, and 0.6 µM leupeptinhemisulfate (all obtained from Sigma). Cells were broken bysonication for 1 min at 4°C (cycle 30 with 40% power; SonopulsSonicator [Bandelin, Berlin, Germany]), and the extract wascentrifuged for 20 min at 20,000 x g at 4°C. The supernatantwas collected and stored at –20°C after addition of10% glycerol.

Enzyme assays were performed as previously described (17) withp-nitrophenyl N-acetyl-ß-D-glucosaminide as the substratefor determination of N-acetylglucosaminidase activity and withp-nitrophenyl ß-D-N',N"-triacetylchitotriose as thesubstrate for determination of endochitinase activity (all chemicalswere obtained from Sigma). Enzyme activity was related to biomassby determination of the mycelial dry weight.

Measurement of intracellular cAMP levels.
Mycelia from strains grown for 72 h on PDA plates covered witha cellophane membrane were ground in liquid N₂ to a fine powder,weighed, and homogenized in 10 volumes of 0.1 M HCl. The sampleswere centrifuged at 600 x g at 22°C, and the cAMP contentwas measured with a Direct cAMP enzyme immunoassay kit (Sigma)used according to the manufacturer's instructions. cAMP levelswere related to the protein concentrations (as measured by theBio-Rad protein assay [Bio-Rad]) of the samples and were expressedas means ± standard deviations for three independentexperiments.

Determination of cell wall chitin and glucan contents.
Triplicate cultures of each strain were grown for 36 h in potatodextrose broth (PDB) (Difco). Mycelia were harvested, presseddry between sheets of filter paper, weighed, ground in liquidN₂, and suspended in distilled water. Cells were broken by sonicationfor 1 min at 4°C (cycle 50 and 70% power; Sonopuls Sonicator),and the extract was centrifuged for 10 min at 13,000 x g at4°C, washed four times with distilled water, and lyophilized.

Chitin content was determined as previously described (48) afterthe samples were hydrolyzed by boiling for 4 h with 4 M HCl.For glucan determination, the samples were hydrolyzed as describedby Nemcovic and Farkas (38), and the amount of glucan was determinedas described previously (35).

Nucleotide sequence accession number.
The nucleotide sequence of tga3 has been deposited in the GenBankdatabase under accession number AF452097.

RESULTS

Top

Results

Cloning and characterization of tga3.
PCR amplification resulted in two products that were approximately1,000 and 1,100 bp long, both of which correspond to G ${alpha}$ subunitsbased on amino acid sequences deduced from the sequenced PCRfragments. A 958-bp fragment encoded Tga1 (GenBank accessionnumber AY036905) (41). A 1,189-bp fragment, designated tga3,encoded a protein that is 88% identical (9) to Neurospora crassaGNA-3, which belongs to subgroup III of fungal G ${alpha}$ subunits asclassified by Bölker (3). The 1,189-bp PCR product wasused to screen a genomic ${lambda}$ BlueStar library and to identify multipleclones carrying 10- to 15-kb inserts. Based on sequencing withinternal primers Tga3intF1-F3 and Tga3intR1-R4, tga3 (GenBankaccession number AF452097) consists of a predicted 1,470-bpopen reading frame interrupted by five putative introns at positionsconserved in other fungal G ${alpha}$ homologues. The amino acid sequencewas 90% identical to the MAGA sequence (GenBank accession numberAF011340) from Magnaporthe grisea and 88% identical to the GNA-3sequence (GenBank accession number AF281862) from N. crassa.Most of the differences occurred in the N-terminal one-halfof the protein. Tga3 was only 49% identical to T. atrovirideTga1. Based on Southern analyses performed with genomic DNAof T. atroviride digested with several restriction enzymes andwith the 1,189-bp PCR product as the probe, tga3 is presentin a single copy in the genome of T. atroviride (data not shown).

tga3 expression during confrontation between T. atroviride and R. solani.
tga3 gene transcription was examined during confrontation ofT. atroviride with R. solani by isolating total RNA from differentstages during a plate confrontation assay (Fig. 1). Samplesfrom stage 1 (the distance between Trichoderma and the hostwas >10 mm; neither ech42 nor nag1 was expressed), stage2 (precontact; ech42 was expressed, but nag1 was not expressed),and stage 3 (direct contact; both nag1 and ech42 were expressed)(50), including controls in which T. atroviride was confrontedwith itself, were subjected to Northern analysis. tga3 was transcribedconstitutively whether the host was present or not (Fig. 1),although detection of the tga3 transcript required a long exposureof the Northern blot, suggesting that tga3 was not highly expressed.

View larger version (30K):
[in this window]
[in a new window]
FIG. 1. Transcription of tga3 during different stages of a direct confrontation of T. atroviride P1 with R. solani. A schematic representation of the different confrontation stages (stages 1, 2, and 3) of the mycoparasitic interaction is shown, and expression of the chitinase-encoding genes ech42 and nag1 is indicated by a plus sign or a minus sign. T, T. atroviride; R, R. solani. For Northern analysis of tga3 gene transcription, 20 µg of total RNA was hybridized with a $~$ 1.1-kb PCR fragment of the tga3 structural gene or a 1.9-kb KpnI fragment of the actin-encoding gene act1 as a probe. Lanes M, mycoparasitic interaction between Trichoderma and Rhizoctonia; lanes C, control, in which Trichoderma was confronted with itself.

Phenotype of tga3 disruptants.
Sixty-two percent of the transformants had single-copy integrationof the disruption cassette at the homologous tga3 gene locus.Three arbitrarily chosen disruptants, ${Delta}$ tga3-1, ${Delta}$ tga3-3, and ${Delta}$ tga3-5,were analyzed. The disruptants grew at $~$ 45% the rate of the wildtype on PDA and formed little aerial mycelium. The mutants formedflat, glossy colonies that continuously produced dark greenconidia (Fig. 2A) that were the normal size and shape. The ${Delta}$ tga3strains also sporulated in the dark, suggesting that the Tga3G ${alpha}$ subunit has a role in light-induced conidiation. The intracellularcAMP levels in the mutants (6.5 ± 1.3 pmol of cAMP/mgof protein) were less than one-half those found in the wild-typestrain (15.9 ± 4.2 pmol/mg) or a transformant with anectopic integration of the disruption cassette (13.5 ±2.9 pmol/mg). Addition of exogenous cAMP (2, 5, and 10 mM) didnot alter the slowly growing, hypersporulating phenotype ofthe mutants (Fig. 2A), although inclusion of 5 mM cAMP in thegrowth medium more than doubled the intracellular cAMP levels(35.5 ± 5.4 pmol of cAMP/mg of protein) in all threestrains tested compared to the wild-type strain grown on PDAwithout cAMP.

View larger version (84K):
[in this window]
[in a new window]
FIG. 2. Morphology of strain ${Delta}$ tga3-3 compared to the morphology of wild-type T. atroviride P1 (wt). (A) Colony morphology after growth on solid medium (PDA) for 7 days and effect of addition of 5 mM exogenous cAMP on the phenotype. (B) Microscopic examination of the T. atroviride wild type and ${Delta}$ tga3-3 mutant after growth in liquid culture (PDB) for 36 h at 28°C.

In liquid cultures in PDB or SM with 1% glycerol for 36 h, noconidiation occurred. The germination of conidia from the ${Delta}$ tga3mutants was both reduced and delayed compared to the germinationof conidia from wild-type strain P1. After 9 h of incubationat 28°C, $~$ 30% of the wild-type conidia, but only 10% of the ${Delta}$ tga3 conidia, had formed germ tubes, and after 11 h, 55% ofthe wild-type conidia and 33% of the ${Delta}$ tga3 conidia had germinated.After incubation for 36 h in liquid cultures, the ${Delta}$ tga3 mutantsproduced $~$ 30% less biomass than the comparable wild-type culturesproduced. The mycelia of ${Delta}$ tga3 strains were more highly branchedthan those of the wild-type strain (Fig. 2B).

Mycoparasitic ability of ${Delta}$ tga3 mutants.
After 3 to 4 days of cocultivation of wild-type strain P1 withR. solani, P1 began to overgrow and lyse the host fungus (Fig.3A), while the ${Delta}$ tga3 mutants were avirulent in similar cultures(Fig. 3B). Even when the mutants were grown until their myceliacovered about one-half of the plate before R. solani was inoculatedon the opposite side of the petri dish, the ${Delta}$ tga3 strains wereunable to parasitize the R. solani colonies (Fig. 3B). Thus,the Tga3 G ${alpha}$ subunit could have an important role in signal transductionduring host recognition and/or the mycoparasitic response ofT. atroviride. Similar results were obtained for the ${Delta}$ tga3 mutantswhen B. cinerea (Fig. 3C) or S. sclerotiorum (data not shown)was used in the confrontation assays. Typical morphologicalchanges associated with mycoparasitism (34) (e.g., growth ofthe mycoparasite alongside the host's hyphae and attachmentto these hyphae) were not observed in confrontations with the ${Delta}$ tga3 mutants (Fig. 3D). Addition of exogenous cAMP to the plateconfrontation assays (PDA supplemented with 2, 5, or 10 mM cAMP)did not restore the mycoparasitic behavior (i.e., overgrowthand host lysis) of the ${Delta}$ tga3 strains, although cultures supplementedwith 5 mM cAMP regained the ability to form mycoparasitism-relatedinfection structures (data not shown). Culture filtrates ofthe ${Delta}$ tga3 mutants from cultures grown on N-acetylglucosaminefor 10 h and on colloidal chitin for 48 h as sole carbon sourceshad 25 to 75% of the antifungal activity (inhibition of Botrytisspore germination) of the wild-type culture filtrates.

View larger version (79K):
[in this window]
[in a new window]
FIG. 3. Biocontrol assays for analyzing mycoparasitic ability. (A to C) Plate confrontation assays of T. atroviride wild-type strain P1 (wt) (A) and the ${Delta}$ tga3-3 mutant with R. solani (B) and B. cinerea (C). In the plates on the right in panels B and C, the tga3-negative mutant was grown until it covered about one-half of the plate, and then the host fungi were inoculated. Pictures were taken 4, 7, and 14 days (A) or 4 and 14 days (B and C) after inoculation. (D) Microscopic examination of the mycoparasitic interaction between the T. atroviride wild type or the ${Delta}$ tga3-3 mutant and R. solani. Attachment of the wild type to the host's hyphae is indicated by the arrow.

Extracellular chitinase activities in ${Delta}$ tga3 strains.
In the wild-type strain T. atroviride P1, no chitinase productionwas observed during incubation on glucose or glycerol for 5and 10 h (data not shown). When induced with N-acetylglucosamine,the wild type produced up to 300 U of NAGase per g (dry weight)for up to 10 h; after 12 h the NAGase activity began to declineagain (Table 1). Upon incubation on colloidal chitin for 48h, both NAGase activity (up to 1,900 U/g [dry weight]) and endochitinaseactivity (up to 205 U/g [dry weight]) also were induced in T.atroviride P1.

View this table:
[in this window]
[in a new window]
TABLE 1. Extracellular, mycelium-bound, and intracellular NAGase activities of N-acetylglucosamine-induced cultures of T. atroviride wild-type strain P1 and the ${Delta}$ tga3 mutant

As with the wild type, chitinase production was not observedin the ${Delta}$ tga3 mutants when they were cultured on glucose or glycerol.The extracellular NAGase activities of the N-acetylglucosamine-inducedcultures of the ${Delta}$ tga3 strains were 7 to 21% that of the wild-typestrain (Table 1). In cultures induced with colloidal chitin,both the NAGase and endochitinase activities of the ${Delta}$ tga3 mutantswere no more than 22 to 60% those of the wild type. These resultssuggest that Tga3 influences the production of extracellularendo- and exochitinases.

nag1 and ech42 transcription in ${Delta}$ tga3 mutants.
In wild-type strain P1, nag1 transcript formation was greatest5 h after transfer to a medium containing N-acetylglucosamine,lower after 10 h, and undetectable by 12 h. Following transferto colloidal chitin, nag1 and ech42 transcripts could be detectedat both 36 and 48 h. At all times tested, the ${Delta}$ tga3 mutant producedsignificantly more nag1 transcript than the wild type producedwhen the organisms were induced with N-acetylglucosamine, andit produced both nag1 and ech42 transcripts when the organismswere induced with colloidal chitin (Fig. 4).

View larger version (34K):
[in this window]
[in a new window]
FIG. 4. Analysis of transcription of the chitinase-encoding genes nag1 and ech42 in the T. atroviride wild type (wt) and the ${Delta}$ tga3-3 mutant strain following induction with N-acetylglucosamine or colloidal chitin: Northern analysis of nag1 gene transcription 5, 10, and 12 h after transfer to 1% N-acetylglucosamine (lanes N) and 36 and 48 h after transfer to 1% colloidal chitin (lanes C). A 1.9-kb SalI/XbaI fragment of nag1 was used as the probe. ech42 gene transcription was analyzed 36 and 48 h after transfer to colloidal chitin by using a 1.0-kb PstI fragment of ech42 as the probe. Hybridization with a 1.9-kb KpnI fragment of the actin-encoding gene act1 was used as a loading control.

The nag1 mRNA levels of selected samples were quantified byreal-time reverse transcriptase PCR. nag1 mRNA was barely detectablein both the wild type and the ${Delta}$ tga3-3 mutant 10 h after transferto glucose or glycerol. After induction with N-acetylglucosaminefor 10 h, the nag1 transcript levels were at their highest valuesin both strains; the wild type showed $~$ 2,700-fold induction,and ${Delta}$ tga3 showed $~$ 3,500-fold induction. After 12 h, the nag1 transcriptlevel for the wild-type strain had decreased to $~$ 600-fold inducibility,reflecting the results of the Northern blot analysis, whilethe nag1 transcript level for the ${Delta}$ tga3 strain had dropped onlyslightly to $~$ 2,650-fold inducibility. When organisms were inducedwith colloidal chitin, a similar pattern was observed; the ${Delta}$ tga3strain again had elevated nag1 transcript levels at both 36h (1.9-fold higher than the wild type) and 48 h (1.5-fold higherthan the wild type) after induction.

Localization of N-acetylglucosaminidase activity in ${Delta}$ tga3 mutants.
Although the ${Delta}$ tga3 mutant had elevated nag1 and ech42 transcriptlevels, its extracellular chitinase activities in induced cultureswere reduced. Similar levels of mycelium-bound NAGase activitywere found in the wild type and the mutant, but intracellularNAGase activity was significantly higher in the ${Delta}$ tga3 mutantthan in the wild-type control (Table 1).

Cell wall of the ${Delta}$ tga3 mutants.
We could not produce protoplasts from the ${Delta}$ tga3 strains evenif we incubated them for up to 20 h (normal incubation time,3 to 4 h) and added twice as much of the lysing enzymes. Resistanceto protoplasting by cell wall lytic enzymes may be due to achange in cell wall polymer content, an increase in cross-linking,or a change in the type of cross-linking occurring in the cellwall. Any of these changes could affect cell wall porosity (21).Nevertheless, neither the cell wall chitin nor the glucan contentof the ${Delta}$ tga3 mutants was significantly different from the contentof the wild type or the transformant with ectopic integrationof the disruption construct (data not shown). Growth inhibitionby SDS often is used as an indicator of cell wall integrity(8, 21), but the reduction in hyphal growth by the ${Delta}$ tga3 mutantsin the presence of 0.01 or 0.02% SDS was comparable to thatof the control strains (data not shown).

DISCUSSION

Top

Discussion

In filamentous fungi, ${alpha}$ subunits of heterotrimeric G proteins,which are activated upon ligand binding to a corresponding receptor,play crucial roles in regulating functions such as asexual development,sexual reproduction, and pathogenicity, in which fungi sensethe presence of mating partners by pheromones and pathogenicfungi respond to signals from their hosts (22, 25, 26, 28, 31,39, 44, 52). We identified and characterized a novel G protein ${alpha}$ subunit-encoding gene, tga3, of the mycoparasitic fungus T.atroviride, which groups with subgroup III of fungal G ${alpha}$ proteins.Thus, Trichoderma spp. have members of each G ${alpha}$ protein subgroup,as previously shown for, e.g., N. crassa and M. grisea (28,31, 45). The tga3 gene is transcribed constitutively at a lowlevel during different growth stages and irrespective of thepresence of a host fungus. Deleting the T. atroviride subgroupI G ${alpha}$ protein-encoding gene tga1 did not alter tga3 transcriptlevels (Zeilinger, unpublished data), which is consistent withresults obtained for N. crassa, where deleting any of the threeG ${alpha}$ subunit-encoding genes did not affect the expression of theother G proteins (28).

For generating ${Delta}$ tga3 knockout mutants, the hph (hygromycin Bphosphotransferase-encoding) gene was employed as a selectionmarker. This marker gene is frequently used for transformationof T. atroviride (34, 41, 50, 51), and its gene product so farhas not been reported to interfere with growth or the mycoparasiticability of the fungus. This is consistent with our results foran ectopic transformant used as a control in selected experiments,in which it behaved like the wild type.

${Delta}$ tga3 mutants grow slowly with nearly no aerial hyphae, and theycontinuously produce conidia when they are cultured on solidmedia, a phenotype similar to that reported for an N. crassa ${Delta}$ gna-3 mutant (28). However the Neurospora mutant can form conidiain submerged culture. Conidia from ${Delta}$ tga3 mutants also have reducedand delayed germination, a characteristic similar to a characteristicobserved for the gasC mutants of Penicillium marneffei (53).

Mutants of some subgroup III fungal G ${alpha}$ proteins, including Cryphonectriaparasitica CPG-2 (14), N. crassa GNA-3 (28), Ustilago maydisGpa3 (39), and Podospora anserina MOD-D (33), can be phenotypicallysuppressed by exogenous cAMP. The finding that the ${Delta}$ tga3 mutantphenotype was not altered by the addition of exogenous cAMPsuggests that these properties are mediated through a cAMP-independentpathway. Nevertheless, intracellular cAMP levels were significantlyreduced in ${Delta}$ tga3 mutants, suggesting that Tga3 can increase theinternal cAMP level.

In Trichoderma, as in other filamentous fungi, asexual sporulationis activated by light (2, 13, 42, 49), and no conidia are producedwhen the fungus is grown in the dark. When grown in the dark, ${Delta}$ tga3 mutants sporulated in the same manner as when they werecultured in the light, like tga1 antisense mutants (41). Thus,both G ${alpha}$ proteins (Tga1 and Tga3) act as negative regulators ofconidiation, perhaps by sharing downstream signaling components.Overlapping roles for G ${alpha}$ subunit proteins also are known in N.crassa, in which both GNA-1 and GNA-3 regulate conidiation (25,28), and in P. marneffei, in which GasA- and GasC-mediated signalingblocks conidiation and regulates production of a red pigment(52, 53).

Tga3 also has a major impact on mycoparasitism, as ${Delta}$ tga3 mutantswere avirulent in plate confrontation assays with R. solani,B. cinerea (Fig. 3), and S. sclerotiorum (data not shown). Theantifungal activity of chitinolytic enzymes and their role duringmycoparasitism are well established (29, 40, 46). While ${Delta}$ tga3strains had reduced extracellular endochitinase and N-acetylglucosaminidaselevels, nag1 and ech42 gene transcription and intracellularchitinase levels were increased in the mutants. Compared toAspergillus nidulans, in which the subgroup I G ${alpha}$ subunit FadAand the Gß subunit SfaD regulate cell wall porosityand chitin content (8), cell wall composition was not alteredin ${Delta}$ tga3 mutants. Thus, the observed accumulation of chitinolyticenzymes inside the cell and the retention of these enzymes inthe cell wall may be due to Tga3-mediated posttranscriptionalregulation of chitinase gene expression and its influence onchitinase secretion.

When chitinase-induced culture supernatants were used in invitro Botrytis spore germination assays, the antifungal activityof the ${Delta}$ tga3 mutants was reduced compared to the activity ofthe wild type and roughly reflected the extracellular chitinaseactivity. These experiments showed that the ${Delta}$ tga3 mutants canstill produce extracellular chitinases and suggested that thecomplete loss of mycoparasitism in direct confrontation assaysis caused by other defects. One hypothesis is that the ${Delta}$ tga3strains no longer recognize the host. This hypothesis is stronglysupported by the lack of mycoparasitism-related infection structures(e.g., growth alongside, attaching to, or coiling around thehost's hyphae) (11, 34) in dual cultures of the ${Delta}$ tga3 mutantswith a host fungus. In U. maydis, deleting the subgroup IIIG ${alpha}$ protein Gpa-3 rendered the fungus avirulent. The resultingmutants also were mating deficient, as they could not respondto the pheromone signal (39). Another hypothesis for the lossof mycoparasitism by the ${Delta}$ tga3 mutants is that hyphal polaritycould be changed due to the altered, hyperbranching phenotype,resulting in an inability to form the typical mycoparasitism-accompanyingmorphological changes.

Recognition of the host fungus by Trichoderma seems to playan important role in triggering the mycoparasitic response (1,36, 37). The host-derived signals identified so far are diffusible,low-molecular-weight compounds released from the host's cellwall by the action of Trichoderma-secreted chitinases, whichfunction as precontact inducers of antagonistic genes in themycoparasite and as elicitors of plant defense mechanisms (19,30, 50). Since most of the chitinase activity is retained insidethe cells of ${Delta}$ tga3 strains, one hypothesis for the loss of hostrecognition could be a reduction in or a complete lack of thisinducer, resulting in the failure of precontact induction ofmycoparasitism-relevant genes. In addition, lectins presenton the host's cell wall may be involved in host recognition,as they induce mycoparasitic coiling upon direct contact (23,24). When different lectins were used to coat nylon fibers forbiomimetic experiments, the lectins promoted attachment to andcoiling around the fibers, regardless of their sugar moiety,suggesting that lectins were not the specificity-determiningfactor in host recognition (41). Tga1 antisense transformantscould not overgrow R. solani and exhibited decreased coilingaround nylon fibers when they were tested in biomimetic assays(41), similar to ${Delta}$ tga3 mutants which showed a loss of mycoparasitism-relatedinfection structure formation upon direct contact with a livinghost fungus.

In summary, the data strongly support the hypothesis that theloss of mycoparasitism of ${Delta}$ tga3 mutants is a combination of indirecteffects (e.g., reduced growth, a defect in chitinase secretion)and direct effects (e.g., loss of infection structure formation).Furthermore, Tga3 directly influences the regulation of mycoparasitism-relatedchitinase gene transcription, as the mutants had enhanced nag1and ech42 mRNA levels.

The two G protein ${alpha}$ subunits of T. atroviride have overlappingroles in conidiation and mycoparasitism, so it is necessaryto identify the specific receptors and the signaling pathwaysused to transmit the host-derived signals to downstream effectorsin order to understand the complex biochemical processes thatresult in fungal biocontrol.

ACKNOWLEDGMENTS

This work was supported by grants from the Austrian Programmefor Advanced Research and Technology (APART 10764) of the AustrianAcademy of Sciences and from Fonds zur Förderung WissenschaftlicherForschung (grant FWF P15483). Part of the work was performedwithin the scope of the Marie Curie training site GEMCAT (contractHPMT-CT-2001-00243) supported by the European Community.

We thank C. P. Kubicek for helpful discussions and for criticallyreviewing the manuscript.

FOOTNOTES

* Corresponding author. Mailing address: Research Area of Gene Technology and Applied Biochemistry, Institute for Chemical Engineering, Vienna University of Technology, Getreidemarkt 9, A-1060 Vienna, Austria. Phone: 43-1-58801-17256. Fax: 43-1-5816266. E-mail: szeiling{at}mail.zserv.tuwien.ac.at.

REFERENCES

Top