G-Protein β Subunit of Cochliobolus heterostrophus Involved in Virulence, Asexual and Sexual Reproductive Ability, and Morphogenesis

Abstract

Previous work established that mutations in mitogen-activated protein (MAP) kinase (CHK1) and heterotrimeric G-protein α (Gα) subunit (CGA1) genes affect the development of several stages of the life cycle of the maize pathogen Cochliobolus heterostrophus. The effects of mutating a third signal transduction pathway gene, CGB1, encoding the Gβ subunit, are reported here. CGB1 is the sole Gβ subunit-encoding gene in the genome of this organism. cgb1 mutants are nearly wild type in vegetative growth rate; however, Cgb1 is required for appressorium formation, female fertility, conidiation, regulation of hyphal pigmentation, and wild-type virulence on maize. Young hyphae of cgb1 mutants grow in a straight path, in contrast to those of the wild type, which grow in a wavy pattern. Some of the phenotypes conferred by mutations in CGA1 are found in cgb1 mutants, suggesting that Cgb1 functions in a heterotrimeric G protein; however, there are also differences. In contrast to the deletion of CGA1, the loss of CGB1 is not lethal for ascospores, evidence that there is a Gβ subunit-independent signaling role for Cga1 in mating. Furthermore, not all of the phenotypes conferred by mutations in the MAP kinase CHK1 gene are found in cgb1 mutants, implying that the Gβ heterodimer is not the only conduit for signals to the MAP kinase CHK1 module. The additional phenotypes of cgb1 mutants, including severe loss of virulence on maize and of the ability to produce conidia, are consistent with CGB1 being unique in the genome. Fluorescent DNA staining showed that there is often nuclear degradation in mature hyphae of cgb1 mutants, while comparable wild-type cells have intact nuclei. These data may be genetic evidence for a novel cell death-related function of the Gβ subunit in filamentous fungi.

Filamentous fungi recognize and respond to signals from the environment and from host organisms by altering their growth and development. The creation of mutants deficient in signaling is key to understanding the regulation of the disease process of plant pathogens. The loss of a signaling element controlling a coordinated set of downstream events can result in an avirulent phenotype. Following the isolation of a G-protein α (Gα) subunit gene from Neurospora crassa (35), a large number of such genes were identified in filamentous fungi (4) and classified into groups according to conserved motif sequences (13). Most ascomycetes for which there are data (e.g., N. crassa) have three Gα subunit genes (10). One fungal Gα subunit group, defined by N. crassa GNA1, is most similar to the mammalian Gαi subunit class; this class includes consensus myristoylation sequences and the target of pertussis toxin ADP ribosylation (35). Mutations in members of this class often lead to defects in development and virulence. Less is known about the functions of members of the other two main classes of fungal Gα subunits (4, 11, 12, 13, 20, 44). The different classes are similar enough that functional redundancy can occur. In N. crassa, for example, loss of Gna-2 results in only subtle phenotypes, while deletion of GNA-2 from a strain that is already null for GNA-1 accentuates the phenotypes (2). A recurrent observation is that a particular Gα subunit may have different functions in different fungal species. For example, the N. crassa Gα subunit Gna-3 apparently belongs to a pathway that represses conidiation (16), while a different Gα subunit, the Gαi subunit homolog FadA, has this function in Aspergillus nidulans (1).

Fungal Gα subunits, like those in animal cells, likely form heterotrimers that are composed of α, β, and γ subunits and that, upon interaction with an activated heptahelical transmembrane receptor, dissociate from the βγ heterodimer. Either part of the dissociated trimer can activate downstream effectors. The presence of a Gβγ dimer has not yet been demonstrated in filamentous fungi, and relatively little is known about the signaling functions of the Gβ subunit. Among the plant pathogens, Gβ subunit genes are known to be required for the virulence and development of the chestnut blight fungus Cryphonectria parasitica (15), the rice blast agent Magnaporthe grisea (28), and Fusarium oxysporum f.sp. cucumerinum, which causes cucumber wilt (14). The homologs in A. nidulans (29), N. crassa (42), and the opportunistic basidomycete human pathogen Cryptococcus neoformans (37) are also involved in developmental pathways. The C. neoformans ortholog is required for mating but is dispensable for virulence (37), again emphasizing how different signaling pathways are put to different uses, depending on the fungal species.

In budding yeast cells, the mating pheromone signal is transduced by the Gβγ dimer STE4/STE18, while the corresponding Gα subunit serves to sequester the dimer. A mitogen-activated protein kinase (MAPK) cascade transmits the signal downstream to drive the expression of target genes; specificity is maintained despite the fact that the same module can signal for mating or filamentous growth (23, 31). Filamentous growth is mediated by MAPK cascades operating in parallel to or downstream from the G proteins Ras and Gpa2 (17, 26, 30). The pathways of yeast cells can provide insight into how filamentous fungi, including pathogens, transduce the particular signals that control their development (3). In C. parasitica, both α subunit and β subunit genes have been studied by targeted gene disruption. Both genes are required for pathogenesis, but the phenotypes conferred by mutations in the genes encoding the two subunits are not identical (15). The β subunit is required for the regulation of development in A. nidulans. A loss of function of the Gβ subunit can suppress the phenotypes resulting from the overactivation of the Gα subunit that occurs in mutants lacking the RGS (regulator of G-protein signaling) protein FlbA (29). The C. neoformans Gβ subunit gene GPB1 is required for mating and signals through an MAPK cascade analogous to that in budding yeast cells, while the Gα subunit Gpa1 signals through a distinct, cyclic AMP-dependent pathway (37).

The maize pathogen Cochliobolus heterostrophus attacks the host leaf, forming small appressoria that do not appear essential for penetration (13), which may occur through stomata or directly through the epidermis. Mutants in which the Gα subunit gene CGA1 is disrupted or deleted still cause disease symptoms on corn leaves (13), despite their altered growth pattern and nearly total loss of the capacity to form appressoria. Furthermore, once inside the leaf, CGA1 mutants are able to grow and overcome host defenses. One of several possible explanations for the ability of a signaling mutant to retain at least partial virulence is redundancy in the pathways that transduce signals from the host. At least two pathways are thought to transduce host and surface signals leading to the pathogenic development of M. grisea (6). In all filamentous fungal species for which sufficient information is available, several Gα subunit genes have been found (C. heterostrophus has three, CGA1, CGA2, and CGA3) (B. A. Horwitz and B. G. Turgeon, unpublished data). To help distinguish between redundancy in G-protein signaling pathways and alternate signaling routes through additional pathways, we have isolated and disrupted the sole C. heterostrophus Gβ subunit gene, CGB1. The phenotypes conferred by the loss of CGB1 overlap only partially with those of cga1 mutants, providing insight into heterotrimeric G-protein signaling pathways.

MATERIALS AND METHODS

Fungal strains, growth media, and nucleic acid isolation.

Wild-type (WT) C. heterostrophus strains C4 (MAT1-2 tox1⁺; ATCC 48331) and C5 (MAT1-1 tox1⁻; ATCC 48332) were described previously (33). Three original cgb1 mutant strains generated from strain C4 (C4ΔGβ1, C4ΔGβ4, and C4ΔGβ9) were purified by isolation of single conidia. One purified cgb1 mutant strain, C4ΔGβ4, was backcrossed to WT strain C5, and WT (N52-R-1, N52-R-2, and N53-R-6) and mutant (N52-R-3, N52-R-4, and N52-R-5) progeny were collected and screened for the cgb1 mutation. Progeny strain N52-R-5 (cgb1 MAT1-2 tox1⁻) was backcrossed again to strain C5 (cross 1400), and a tetrad was collected. One progeny mutant from the tetrad from cross 1400 (1400-1-3; cgb1 hygB^R tox1⁻ MAT1-1 ALB1) was crossed to strain C4 to confirm cosegregation of the hygB marker and the cgb1 mutation. To test reproductive ability, strain 14-1-3 was also crossed to strain CB12 (MAT1-2 alb1) (cross 1421). Four progeny strains (two black and two albino) from a tetrad from cross 1421 were backcrossed to WT test strains of the opposite color. Pigmented cgb1 progeny were also crossed to albino cgb1 progeny of the opposite mating type to test reproductive ability in a cgb1 homozygous cross.

Media, standard growth conditions, and mating assays were described previously (18, 36, 39). To examine the ability to conidiate, total conidia were collected from plates (100 by 15 mm) containing complete medium with xylose (CMX; three replicates were used) and counted by using a hemocytometer.

Genomic DNA was isolated from mycelium ground in liquid nitrogen as described for plant tissue (8) and used for PCR analysis.

Isolation of CGB1.

An internal fragment of the CGB1 gene was isolated by PCR amplification from genomic DNA of strain C4. In the first reaction, the following primers, designed to match peptide sequences IYAMHW and GHDNRV, respectively, were used: forward, 5′-AT(A/T/C)TA(T/C)GCNATGCA(T/C)TGG-3′, and reverse, 5′-ACNC(G/T)(G/A)TT(G/A)TC(G/A)TGNCCG-3′. A second, nested reaction was carried out with a 1-μl aliquot from the first reaction as a template and with the following primers, designed to match the peptide sequences VMTCAY and ATGSDD, respectively: forward, 5′-GTNATGACNTG(C/T)GCN-3′, and reverse, 5′-(A/G)TC(A/G)TCN(C/G)(T/A)NCCNGTNGC-3′. Cycling parameters were as follows: 2 min at 94°C; 35 cycles of 1 min at 94°C, 2 min at 50°C, and 2 min at 72°C; and 5 min at 72°C.

After electrophoresis of the PCR products, the gel was blotted and probed with a Cpgb-1 cDNA clone from C. parasitica, kindly provided to us by D. Nuss (15). After identification of the C. heterostrophus CGB1 internal fragments among the products of both single (899 bp) and nested (447 bp) reactions, upstream and downstream regions were isolated by additional PCRs with a GenomeWalker kit (Clontech) and with C. heterostrophus strain C4 genomic DNA as a template. For this, the following nested gene-specific primers were used: for the 3′ flank, outer, GSP3′-1, 5′- TCGGATATCAACGCCATTCAGTTCTTC-3′, and inner, GSP3′-2, 5′-GCCTCGTGTCGCCTGTTTGATATCCG-3′; and for the 5′ flank, outer, GSP5′-1, 5′-AGACCACCGCAGGCAACGTAGT -3′, and inner, GSP5′-2, 5′-GGACCTTGTTTGTTGTGTAGGC-3′. The PCR products were cloned in pCRScript (Stratagene) or pUC57 (Fermentas); the resulting clones are referred to as GW5′ and GW3′ (see Fig. 2A).

FIG. 2.

CGB1 gene replacement construct. (A) Construct strategy. The top line shows the chromosomal region carrying CGB1 (light gray arrow) and the 473-bp region replaced by the hygB resistance marker in the event of a double-crossover integration. Also shown are the 5′ (GW5′, 693 bp) and 3′ (GW3′, 460 bp) extensions obtained by PCR-based genome walking with primers GS5′-2 and GS3′-2. The bottom line shows the gene replacement construct consisting of GW5′-hygB-GW3′. The hygB resistance cassette from pUCATPH (22) was cloned between the two flanks as a SalI/XbaI fragment. Primers used for gene amplification and for checking integration events are indicated by arrows. (B) Confirmation of a double-crossover integration event at CGB1. Genomic DNAs from the strains indicated above the panels were used as templates for amplification with the primer pairs indicated below the panels. Strains 1400-1-3 and 1400-1-4 are cgb1 progeny from the backcross described in Materials and Methods (see also Fig. 3). In parentheses below each panel are the expected sizes of PCR products obtained with each primer pair.

Phylogenetic relationships.

Protein sequences were aligned by using CLUSTAL W, version 1.82. A distance matrix was computed from this alignment by using PROTDIST from the PHYLIP package (9) (http://bioweb.pasteur.fr). Default options were chosen, without bootstrapping. The resulting distance was scaled in units of the expected fraction of amino acids changed. The FITCH program (PHYLIP) was used to generate a phylogenetic tree; default options were chosen, with the exception that Arabidopsis was used as the outgroup species. The resulting tree was plotted with Phylodendron, version 0.8d (D. G. Gilbert; http://iubio.bio.indiana.edu/treeapp/treeprint-form.html).

Construction of a gene replacement vector.

The CGB1 disruption construct (see Fig. 2A) was prepared in two steps. First, a 693-bp fragment excised from GW5′ with HindIII and XbaI and the gene for resistance to hygromycin B (hygB) under the regulation of the A. nidulans TrpC promoter (excised from vector pUCATPH [22] with XbaI and SalI) were cloned into pBluescript (Stratagene) digested with HindIII and SalI. The resulting plasmid was digested with SalI and ligated to a 460-bp SalI fragment excised from GW3′. The desired orientation of this 3′-flanking region was confirmed by restriction enzyme digestion. The final construct was designated GbetaDXO. A 3.5-kb linear fragment that included the flanking regions and a selectable marker was obtained by PCR amplification with T3 and T7 primers and with GbetaDXO as a template. A total of 20 μg of this fragment was used to transform protoplasts as described previously (39), except that the enzyme Novozym 234 was replaced with β-d-glucanase (catalog no. 0439-1; InterSpex, San Mateo, Calif.) at a concentration of 2.9 mg/ml. After 24 h of recovery on regeneration medium (1 M sucrose, 0.1% yeast extract, 0.1% casein hydrolysate) (34), the plate was overlaid with water agar containing hygromycin B (Calbiochem) so that the final concentration of hygromycin B in the plate was 50 μg/ml.

Confirmation of a gene replacement event at CGB1.

To confirm the insertion of the transforming DNA and the concomitant deletion of 473 bp of CGB1 by a double-crossover event, the following primers were used in PCRs with genomic DNAs from WT and transformant strains as templates: Beta3as, 5′-GAATAGGCGCACACGTCC-3′; Gβ3215for, 5′-CACCACGTACGTGCTACAC-3′; TTrpC, 5′-GGTGTTCAGGATCTCGATAAG-3′; and PTrpC, 5′-GGTCGTTCACTTACCTTGCTTG-3′. In addition, the primers GSP7 (sense), 5′-ACTACGTTGCCTGCGGTGGTC-3′, and GSP12 (antisense), 5′-TCAGAGCCAGTGCCGAAAGC-3′, were used to amplify a 411-bp fragment from the portion of the coding region deleted upon double-crossover integration of the disruption construct.

Plant inoculation.

Mycelial mats of 7-day-old fungal cultures were scraped from plates (100 by 15 mm) of CMX agar, inoculated into 100 ml of liquid complete medium (CM) in 250-ml flasks, and shaken for 18 h at 30°C and 200 rpm. A 2-g sample of wet mycelium was diluted in 15 ml of 0.05% Tween 20 solution in water and ground in a 40-ml tissue grinder (Konte Glass Co., Vineland, N.J.) until cells were fragmented, as determined microscopically. Once the mycelia were well ground, 1 ml of mycelial suspension was diluted 16- and 256-fold. A total of 16 ml of mycelium dilution was sprayed onto 14-day-old corn plants (W64-A) by using a pressurized Preval Spray Gun Power Unit thin-layer chromatography sprayer (Alltech Associates, Deerfield, Ill.). The plants were incubated for 20 h at 22°C in a mist room and then transferred to a 23°C growth chamber with 16 h of light per day. Lesions were assessed daily, and photographs were taken at 5 days postinfection.

Microscopy.

For microscopic observation of vegetative hyphae and appressorium formation, WT (C5) and cgb1 mutant (14-1-3) strains were grown on CM agar plates for 7 days. Mycelial mats then were scraped from the plates and fragmented in 50 ml of sterile distilled water in a sterile stainless steel blender cup (ground for 30 s). About 50 μl of mycelial fragment suspension was placed on a microscope glass slide (Fisherbrand Plain; Fisher Scientific, Pittsburgh, Pa.) and incubated at 32°C in a mini-mist chamber (100- by 15-mm petri dish padded with water-saturated filter paper) for 6, 8, 14, and 24 h. Three replicates were used for each treatment. After each period of incubation, fungal material was stained with cotton blue.

For pigmentation evaluation on solid agar substrates, strains were grown on CMX agar plates, and fungal aerial masses were scraped from the surfaces of the plates 2 weeks after inoculation. A cross section of the agar substrate was cut from a pigmented region by using a surgical blade, and the section was viewed by using a Photomicroscope II microscope (Carl Zeiss, Oberkochen, Germany) with bright-field optics. For pigmentation evaluation in liquid, strains were incubated in 100 ml of liquid CM at room temperature with shaking for 48 h, and culture extracts were collected by filtration first through four layers of cheesecloth and then through a 0.2-μm-pore-size filter (Corning Glass Works, Corning, N.Y.).

To examine nuclear conditions, germinated mycelial fragments (obtained by using the methods described above) were stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Co., St. Louis, Mo.). For this, specimens were fixed with 100% ethanol followed by washing with 75% ethanol and then with water. DAPI-stained specimens were viewed by using the Photomicroscope II microscope with UV excitation filter BG3, beam splitter 450, and barrier filter 41. Images were captured on Elite Chrome 400 film (Kodak, Rochester, N.Y.) at exposure times of 5 to 10 s.

TUNEL staining.

Terminal deoxynucleotidyltransferase-mediated dUTP-fluorescein nick end labeling (TUNEL) was performed by using an in situ cell death detection kit (fluorescein; Roche Applied Science, Indianapolis, Ind.). Mycelia from test strains C5 (WT) and 1400-1-3 (cgb1 mutant) were scraped from CMX agar plates (7 days old), fragmented in a stainless steel blender, and inoculated into 200 ml of liquid CM. Mycelia were harvested after 20 h of incubation at room temperature with shaking. Fungal protoplasts were prepared by following the same procedure as that used for C. heterostrophus transformation (34), except for the replacement of Novozyme 234 with Glucanex (in a liquid preparation; kindly provided by C. M. Hjort, Novo Nordisk, Bagsvaerd, Denmark) in the enzyme-osmoticum solution (2 mg/ml). After adjustment of the final concentration to 10⁶/ml in STC (1.2 M sorbitol, 10 mM Tris, pH 7.5, 50 mM CaCl₂), protoplasts were stored at −80°C and thawed on ice just before use.

For the TUNEL assay, thawed protoplasts (40 μl each) were placed as drops on glass slides (Fisherbrand Colorfrost Plus; Fisher Scientific), air dried in a laminar hood for 1 h, and then fixed with 2.5% paraformaldehyde in phosphate-buffered saline (PBS [pH 7.4]) at 4°C overnight. Fixed protoplasts were washed three times with PBS, dehydrated in an ethanol series (50, 70, 90, and 100%), and air dried for 2 h. Just before the TUNEL assay, the dried protoplasts were rehydrated in an ethanol series (100, 90, 70, and 50%) and then resuspended in 10 mM sodium citrate (pH 6.0), permeabilized with 0.1% Triton X-100 in 10 mM sodium citrate (pH 6.0) at room temperature for 3 min, and washed three times with PBS. TUNEL labeling was performed according to the manufacturer's instructions, except that the incubation at 37°C was increased to 2 h. A negative control (labeling buffer only, without terminal deoxynucleotidyltransferase) and a positive control (pretreated with DNase I) were included for both strains.

After the TUNEL assay, the slides were washed three times with PBS, air dried briefly, mounted in mounting medium from a Prolong Antifade kit (Molecular Probes, Eugene, Oreg.), covered with a coverslip, and sealed with nail polish. Specimens were examined by using an Olympus BH-2 microscope. Fluorescence images were recorded on Kodak Elite Chrome 400 film at exposure times of 30 to 60 s with ×40 or ×100 fluorescence objectives.

Nucleotide sequence accession number.

The sequence of CGB1 has been deposited in GenBank with accession number AY211190.

RESULTS

Isolation of CGB1.

Nested PCR, hybridization with a full-length C. parasitica Cpgb-1 cDNA clone (15), and use of the GenomeWalker kit led to the cloning of the C. heterostrophus CGB1 gene (ChCGB1). The complete coding sequence was obtained from the clone. Four introns were located by comparison with the cDNA sequence obtained by RT-PCR. The four ChCGB1 introns are conserved in the A. nidulans SfaD gene, while introns 1, 2, and 4 are conserved in the N. crassa GNB-1 gene; the third intron is missing. An additional sequence overlapping and extending the genomic sequence by 1.5 kb upstream from the start codon and downstream from the stop codon was obtained by using CGB1 to query the C. heterostrophus genome database (provided by Celera Genomics for Torrey Mesa Research Institute [TMRI]/Syngenta). The deduced amino acid sequence of 351 amino acids shows high homology to those of other filamentous ascomycete Gβ subunits (e.g., 81% identity to A. nidulans SfaD [AAC33436], 80% identity to C. parasitica Cpgb-1 [AAC49838], and 76% identity to N. crassa GNB-1 [AAM53552]). The phylogenetic relationships among fungal Gβ subunits and several plant and animal Gβ subunits are shown in Fig. 1A. The ChCgb1 protein grouped with the other fungal Gβ subunits. STE4 of Saccharomyces cerevisiae grouped separately from Gβ subunits of the other ascomycetes and even other fungi.

FIG. 1.

C. heterostrophus CGB1, a Gβ gene. (A) Phylogenetic analysis of fungal Gβ subunit amino acid sequences. The predicted ChCgb1 protein sequence was aligned with fungal and other Gβ sequences. Protein accession numbers are indicated for each species. The scale bar indicates the calculated distance (0.1 is equivalent to an expected 10% amino acid change). (B) Organization of CGB1. The top line shows the CGB1 ORF (bp 1 to 1304) with intron positions as inverted triangles. Numbers above the triangles correspond to intron sizes in base pairs. Numbers below correspond to start codon, stop codon, and intron start positions. The bottom line shows the organization of the predicted Cgb1 protein (amino acids 1 to 351); the amino acid start position of each predicted WD repeat is indicated in parentheses. Below these diagrams is an alignment of WD repeats (WD1 to WD7) predicted by comparison with proteins in a WD repeat protein database. Below the alignment is the consensus sequence for WD repeats (for details, see the text and the BMERC website).

The ChCgb1 protein was used to query a database of WD repeat proteins (BioMolecular Engineering Research Center [BMERC], Boston University [http://bmerc-www.bu.edu]) (Fig. 1B). This analysis confirmed that the predicted ChCgb1 protein has seven repeats (probability of 1.0). WD repeats typically contain a GH dipeptide 11 to 24 residues from the start of the domain and a terminal WD dipeptide, which together represent the WD signature (Fig. 1B). There is flexibility in even the most conserved positions in the profile. For example, in repeat 2, YN appears instead of WD, but these are among the most common replacements found in the WD repeat database (Fig. 1B, bottom panel).

A search of the C. heterostrophus genome database (provided by Celera Genomics for TMRI/Syngenta) did not reveal any other Gβ subunit genes, although other genes predicted to encode WD repeat-containing proteins were present (B.-N. Lee and B. G. Turgeon, unpublished data). Known Gβ subunits share homology in the N-terminal region preceding the WD repeat domain (32). When the protein database was searched (BLASTP) by using the 62 N-terminal amino acids of ChCgb1, the fungal homologs were the top hits, followed by those of Dictyostelium discoideum and metazoans. This region displayed significant homology to other Gβ subunits (45% identity [significance score, E = 0.001] to mouse and human Gβ subunit 5).

Identification of cgb1 mutant strains.

To construct cgb1 mutants, protoplasts of C. heterstrophus strain C4 were transformed with linear DNA obtained from construct GbetaDXO. The 3.5-kb fragment consisted of the two ChCGB1 homologous regions flanking the selectable marker (Fig. 2A). Following plating of protoplasts on regeneration medium and selection for hygromycin B resistance, 12 candidate transformants were isolated from two transformation experiments. These were purified by isolation of single conidia and screened by PCR for those carrying a deletion at ChCGB1. Such events were confirmed for three transformants (C4ΔGβ1, C4ΔGβ4, and C4ΔGβ9). A double-crossover integration event at CGB1 would be expected to replace 473 bp of the CGB1 coding region (amino acids 101 to 258, including all of WD repeats 2, 3, and 4 and a large part of repeat 5) (Fig. 1B) with 2,107 bp containing the hygB selectable marker (Fig. 2A). PCR amplification with primer combinations Gβ3215for-TTrpC and PTrpC-Beta3as and with genomic DNAs from two candidate cgb1 mutant strains (1400-1-3 and 1400-1-4; see Materials and Methods for the origins of these strains) and the WT strain generated 1,488- and 841-bp products from the mutant strains, respectively (Fig. 2B). DNA from the WT strain did not yield products with these primer pairs. Primer pair GSP7-GSP12 generated a 411-bp product from the WT strain which was missing from the mutant strains (Fig. 2B). PCR products corresponding to the 5′ and 3′ sides of the integration (Fig. 2B) were sequenced for one transformant, confirming the double-crossover event.

Backcrosses (1402, 1404, and 1400) between WT strain C5 and three hygromycin B-resistant cgb1 progeny strains (N52-R-3, N52-R-4, and N52-R-5, respectively) recovered from the cross between original transformant C4ΔGβ4 and C5 showed that all three mutant progeny strains carried a second mutation (Fig. 3; shown here is the cross between N52-R-5 and C5). This second mutation was not linked to the cgb1 mutation, as demonstrated by the tetrad depicted in Fig. 3B. Four types of progeny were observed: dark green (hygB^S), light gray with white edges (hygB^S), gray (hygB^R), and white (hygB^R) (Fig. 3B). A backcross of one of the gray (hygB^R) progeny strains (Fig. 3B, 1400-1-3) to WT strains C4 (cross 1408) and CB12 (cross 1421) and analysis of both random and tetrad ascospore progeny revealed that the second mutation had been eliminated, leaving a strain with a single mutation at CGB1 (Fig. 3C). Two of the purified cgb1 strains from backcross 1400 (1400-1-3 and 1400-1-4) were used for further detailed characterization of strains with deletions at CGB1.

FIG. 3.

Purification of strains carrying the cgb1 mutation. (A) Morphology of WT strain C5 (left) and C4ΔGβ4 (right), a hygB^R transformant recovered from transformation of protoplasts with the CGB1 gene disruption vector. Strains were grown on CMX for 5 days. The WT was dark green (left); the double mutant was dark green with a white border (right). Both formed conidia. (B) Segregation of progeny from a tetrad from a cross between the WT and mutant strains shown in panel A. The mutant strain carried a second, untagged mutation, as revealed by segregation analysis (four types of progeny are shown). Progeny were grown on CMX (left) or CM (minus salts) plus hygromycin B (right) for 4 days. Progeny segregated 1:1:1:1 (dark green:dark gray:white:light gray with a white border). Dark gray progeny (strains 1400-1-3 and 1400-1-4; arrows) and white progeny were hygB^R, whereas dark green progeny and progeny that were light gray with a white border were hygB^S. Note that when first isolated from pseudothecia and transferred to CMX, the double-mutant parental progeny were white, but they darkened (to look like those in panel A, right) when transferred from the CMX plate on which they were originally isolated to a new CMX plate (not shown). If a cross is done again between this type of progeny and the WT, tetratype progeny result again. Thus, in the tetratype, parental-type progeny are dark green (hygB^S) and white (hygB^R), while non-parental-type progeny are dark gray (hygB^R) and light gray with a white border (hygB^S). PCR analysis confirmed that the dark gray (hygB^R) progeny carried the CGB1 mutation. (C) Progeny from a cross between the WT and dark gray (hygB^R) non-parental-type progeny (strain 1400-1-3; panel B, arrows) segregated 1:1 for dark green (hygB^S) (WT):dark gray (hygB^R), indicating a single mutation. Progeny were grown on CMX (left) or CM (minus salts) plus hygromycin B (right) for 4 days.

Characterization of cgb1 mutants. (i) Growth and pigmentation.

The radial rate of growth of cgb1 mutants was not significantly different from that of the WT (six replicates measured over 7 days); however, fewer aerial hyphae were produced. One striking difference was that the mutants accumulated considerably more dark pigment in older cultures whether grown on solid medium (Fig. 4A to D) or in liquid medium (Fig. 4G and H). The increased pigmentation of the mutants was contained in the hyphae, not secreted into the medium (Fig. 4G and H), and the older hyphae of the mutants were often fragmented (Fig. 4F).

FIG. 4.

The cgb1 mutant has increased pigmentation in vegetative mycelia. (A to F) Pigmentation of mycelia on the agar substrates of WT strain C5 (A to C) and cgb1 mutant strain 1400-1-3 (D to F). Fungal aerial masses were scraped off the surfaces of plates 2 weeks after inoculation. The agar substrate of the WT was only lightly pigmented (A), whereas that of the mutant was completely dark (D). Microscopic images of cross sections of agar substrates of the WT (B and C) and the mutant (E and F) showed that the increased pigmentation of the cgb1 mutant occurred within vegetative hyphae. Note that vegetative hyphae of the cgb1 mutant were often fragmented (F, arrows); the inset shows a higher magnification of the area with the arrows. (G and H) Liquid cultures of the same two strains as those shown in panels A and D. Note that the liquid culture of the cgb1 mutant (right) was much darker than that of the WT (left), but the filtrates showed no difference.

Hyphae of the WT growing on a glass slide followed a meandering or wavy pattern (Fig. 5A and C) and often produced appressoria at their tips (Fig. 5C and E). Hyphae of the cgb1 mutants grew along a straight path (Fig. 5B and D) and did not make appressoria (Fig. 5F).

FIG. 5.

The cgb1 mutant has abnormal hyphal growth and fails to produce appressoria. Mycelial fragments of WT strain C5 (A, C, and E) and cgb1 mutant strain 1400-1-3 (B, D, and F) were suspended in water, drops were placed on glass slides and incubated for 14 h at 32°C, and slides were stained with cotton blue-lactophenol. Photographs show mycelia from the older parts of the hyphae (A and B), the young hyphae at the growing edge (C and D), and the hyphal tips (E and F). Note that young hyphae of the WT were often curved and that appressoria formed frequently at the tips (E, arrow), whereas young hyphae of the mutant tended to be straight and no appressorium-like structures formed at the tips (F, arrow). Incubation for 6 or 24 h gave the same results (not shown).

(ii) Asexual reproduction.

cgb1 mutants made fewer than 100 conidia per plate, in contrast to the WT, with 10⁵ to 10⁶. The few conidia produced were normal in shape; however, in contrast to the WT conidia, which germinated and produced appressoria by 6 h at 32°C, the mutant conidia did not germinate. The second mutation carried by the original isolates suppressed the cgb1 conidiation phenotypes; original isolates (C4ΔGβ1, C4ΔGβ4, and C4ΔGβ9) and progeny of a cross between C4ΔGβ4 and C5 (N52-R-3, N52-R-4, and N52-R-5) showed a reduced but still significant ability to produce conidia. Partial conidiation was not restored upon subsequent transfers of these lines. The second mutation likely originated during transformation.

(iii) Sexual reproduction.

To check reproductive ability, four progeny (two black and two albino) from a tetrad recovered from cross 1421 were backcrossed again to WT test strains of opposite color. Pigmented cgb1 progeny were also crossed to albino cgb1 progeny of the opposite mating type to test reproductive ability in a cgb1 homozygous cross.

In crosses between albino and pigmented WT test strains, both white (Fig. 6A) and black pseudothecia formed. In crosses between an albino cgb1 mutant and a pigmented WT test strain, no albino pseudothecia (the color of the cgb1 mutant) were observed (Fig. 6B), indicating that the mutant was female sterile. The progeny of crosses between two cgb1 mutants were completely sterile (Fig. 6C); no pseudothecia formed. The fertility of pseudothecia that did form was normal in terms of numbers of asci and ascospores produced, and complete tetrads could be found (compare Fig. 6E with Fig. 6D). On average, asci produced by mutant strains (Fig. 6G) were smaller than those produced by WT strains (Fig. 6F).

FIG. 6.

cgb1 mutants are female sterile. (A) A cross between two WT test strains (strain C5, which is pigmented, and strain CB12, which is albino) yielded both black (white arrow) and white (red arrow) pseudothecia, indicating that both strains are female fertile. (B) A cross between the pigmented WT test strain shown in panel A and an albino cgb1 mutant (strain 1421-2-4) yielded only black pseudothecia (white arrow), indicating that the mutant is female sterile. (C) The result of a homozygous cross between the albino cgb1 mutant shown in panel B and a pigmented cgb1 mutant (strain 1400-1-3) is sterile, since both mutants are female sterile. (D and E) Crushed pseudothecia. Numbers of asci in pseudothecia from crosses A (shown in D) and B (shown in E) appear to be the same; complete tetrads (shown in insets) can be found in both types of crosses. (F and G) Mature asci from cross B (shown in G) are smaller, on average, than those from cross A (shown in F).

(iv) Virulence.

Since very few conidia were produced by cgb1 mutants and those that were found did not germinate, mycelial fragments were used to inoculate maize plants. No disease symptoms were found on plants inoculated with cgb1 mutants, while WT controls produced large lesions under these conditions (Fig. 7). T-toxin production was normal in cgb1 Tox1⁺ progeny (data not shown), indicating that T-toxin production does not depend on CGB1.

FIG. 7.

The cgb1 mutant is unable to cause disease on the host plant. Corn plants (3 weeks old) were inoculated with mycelial fragments (suspended in water) of WT strain C5 or CGB1 deletion strain 1400-1-3, and symptoms were recorded 7 days after inoculation. The WT caused lesions (white arrow), whereas the cgb1⁻ mutant failed to cause visible lesions on leaves of any age.

(v) Distribution of nuclei.

Mycelia of WT and mutant strains were stained with DAPI to determine whether the alteration from a wavy to a straight growth pattern in cgb1 mutants was related to an abnormal distribution of nuclei. In young hyphae, the number and distribution of nuclei appeared normal in the mutants. Mature hyphae of the mutants often showed an abnormal DAPI staining pattern (Fig. 8B to D). In the WT, nuclei could be clearly observed as punctate dots in the cytoplasm (Fig. 8A), while in the mutant, DAPI-stained material was diffuse and patchy in most of the cell, indicating the possible breakup of nuclei (Fig. 8C and D).

FIG. 8.

The cgb1 mutant appears to undergo nuclear degradation during vegetative growth. Mycelial fragments of the WT strain (A) and cgb1 mutant strain 1400-1-3 (B, C, and D) were suspended in water, drops were placed on glass slides and incubated for 5 h at 32°C, and slides were stained with DAPI. WT cells were typically multinucleate, and punctate nuclei (arrows) were generally evenly distributed in the hyphae. The hyphae of mutant cells had fewer obvious nuclei (B and C, white arrows); in some hyphal compartments, no punctate nuclei were seen (D). Apparent nuclear degradation, characteristic of apoptosis, was frequently observed in mutant cells (C and D, gray arrows). The inset shows a higher magnification of regions indicated by the arrows in panel C.

(vi) TUNEL.

The apparent breakup of nuclei indicated by the DAPI staining pattern might reflect a greater incidence or an earlier appearance of programmed cell death in cgb1 mutants. To test this notion, the free 3′OH termini of low- and high-molecular-weight DNAs with strand breaks can be labeled by terminal transferase. TUNEL staining of nuclei, which is based on this labeling reaction, is thus a hallmark of apoptosis. In the reaction used here, terminal transferase added fluorescein-labeled dUTP to these ends.

Attempts to use mycelium for the in situ TUNEL reaction were unsuccessful. We tried several procedures previously described for other filamentous fungi, including permeabilization by freezing in liquid nitrogen (24) and treatment with cell wall-degrading enzymes plus proteinase (21), but failed to obtain the expected results from positive controls, suggesting that the TUNEL methods developed for one system may not be suitable for another. We found that protoplasts prepared for transformation (34) could be used for the TUNEL assay relatively easily. Protoplasts of both the WT (Fig. 9E) and the cgb1 mutants (data not shown) pretreated with DNase I (positive control) showed typical TUNEL-positive cells at the same rates and intensities, while those from the negative control showed background staining only (Fig. 9A). TUNEL-positive cells were frequently observed in cgb1 mutant protoplasts even without DNase I treatment (Fig. 9C). In contrast, TUNEL-positive cells were only occasionally found in WT protoplasts without DNase I treatment (Fig. 9A). The fraction of TUNEL-positive WT protoplasts was 0.036 (329 protoplasts counted from three photographs; 99% confidence interval, 0.026). The fraction of TUNEL-positive cgb1 mutant protoplasts was 0.204 (240 protoplasts counted from four photographs; 99% confidence interval, 0.067). The TUNEL-positive staining in the mutants showed a fragmented pattern (Fig. 9C), while the positive control-stained nuclei appeared intact (Fig. 9E). These data suggest that cells of cgb1 mutants may undergo DNA fragmentation during vegetative growth.

FIG. 9.

Detection of apoptotic nuclei by the TUNEL assay. Protoplasts were obtained from WT and mutant hyphae and stained (see Materials and Methods). (A and C) Fluorescence images. (B and D) Phase-contrast images of the same fields (×100). (E) Positive control for TUNEL staining: WT protoplasts that were fixed, treated with DNase I, and stained. Note the intact but TUNEL-positive nuclei. The cgb1 mutant has similar bright fluorescent dots (C), indicative of nuclear DNA degradation, while WT protoplasts show only dull background fluoresence (A). Bar, 10 μm.

DISCUSSION

One can infer from the known functions of signaling proteins in other eukaryotes that a G-protein heterotrimer detects extracellular ligands through as-yet-unidentified receptors. The next step in the signal cascade is for the activated GTP-bound form of the Gα subunit, as well as the free Gβγ heterodimer, to transmit the signal to downstream effectors. Mutants carrying mutations in these subunits can provide insight into the functions of G-protein-coupled pathways. One should keep in mind, however, that the roles of the products of signaling gene orthologs often differ even between closely related species, suggesting that signaling proteins and their immediate targets may have evolved separately in different species (19, 41). It is thus essential to study the signaling networks of different pathogens in order to obtain a coherent view of the general properties and which properties are unique to a particular species or host-pathogen pair.

The high level of identity of CGB1 with other fungal Gβ genes and the structure of the predicted polypeptide (Fig. 1) are good evidence that ChCGB1 is a Gβ subunit-encoding gene. An exhaustive search of the C. heterostrophus genome sequence (TMRI/Syngenta) indicated that there is only one in the genome; therefore, disruption of this gene creates a strain that cannot form G-protein heterotrimers. Thus, we can conclude that G-protein heterotrimers are not essential for survival. There are a total of three Gα subunit-encoding genes in the genome (Lee and Turgeon, unpublished; M. Giloh and B. A. Horwitz, unpublished data; TMRI/Syngenta C. heterostrophus genome BLAST search). It was shown previously that one of these (CGA1) is essential for several developmental pathways of the corn pathogen C. heterostrophus. A loss of ChCGA1, which belongs to the fungal Gαi subunit class, results in a loss of the ability to form appressoria and female sterility, while virulence on corn is retained (13). A loss of CGA1 does not block asexual sporulation, and conidia of cga1 mutants are viable; however, cgb1 mutants produce very few conidia, and these do not germinate. A loss of CGA1 is lethal for ascospores, and in cga1 homozygous crosses, no pseudothecia are formed. In contrast, cgb1 mutants, although female sterile, can be crossed to the WT, and mutant ascospores are viable. We predict that functions lost in cga1 but not cgb1 mutants will be attributed to specific interactions of the Gα subunit Cga1 with a downstream effector.

The isolation of cgb1 mutants was accompanied by the discovery of an additional unlinked mutation (Fig. 3) which permitted increased conidation. This finding was also reported for the M. grisea ortholog MGB1, although in that case, spontaneous mutations occurred at low rates (27). This observation may be further studied by the identification of the relevant gene combined with the construction of additional mutant alleles of the Gβ subunit gene in both species. The gene corresponding to the second mutation is of interest because it may encode an interaction partner of the Gβ subunit.

A loss of the MAPK ChCHK1 results in a broad set of phenotypes, including a drastic loss of virulence and autolysis of mature mycelial colonies (19). A loss of the Gβ subunit does not result in age-related colony autolysis. These data show that at least some of the functions of ChChk1 are independent of the Gβγ dimer. chk1 mutants grow in the normal wavy pattern on a glass or hard plastic surface, in contrast to cgb1 and cga1 mutants, which grow in a straight path. All three types of mutants produce few or no appressoria. ChCHK1 mutants, like CGB1 mutants, are female sterile but can be crossed to a WT test strain. The phenotypes of the three signal transduction mutants are summarized in Table 1. The basis for the morphological defects shown by conidia germinating on a hard surface is not yet clear. These defects include straight growth of germ tubes, decreased branching, and lack of appressoria. A loss of G-protein signaling may make it impossible for the growing hyphae to detect the surface. Alternatively, the G protein may be needed for the output, i.e., the remodeling of the growing hyphal tip. An example of such remodeling is the momentary loss of polarity that occurs when an appressorium is formed. Appressorium formation and branch formation may share some features. The small appressoria of C. heterostrophus cells contain multiple nuclei, and in this respect they are like normal hyphal cells, except that they are wider and rounded, presumably because of a transient loss of the normal hyphal tip polarity. In budding yeast cells, the activity of the small rho-type G-protein Cdc42p determines the budding frequency (5), but there is no evidence to implicate heterotrimeric G proteins in the control of polarity. Although they are defective in certain morphogenetic transitions, G-protein mutants still essentially follow the normal fungal growth pattern. Mutants with defects in the establishment and maintenance of polarity have more drastic phenotypes and often must be studied with the help of temperature-sensitive alleles (25). Thus, heterotrimeric G-protein pathways, while being dispensable for the polar growth of hyphae, may modulate or interact with the machinery that sets up the main polar axes, changing them in response to extracellular signals or internally produced developmental cues. This scenario could explain the phenotypes of altered conidiation and appressorium formation and the absence of a wavy growth pattern for G-protein mutants. All may involve a transient or a permanent loss of polar growth in response to signals. Novel downstream effectors may be required to transmit such signals.

TABLE 1.

Summary of phenotypes associated with signal transduction gene mutations in C. heterostrophus

Characteristic	Phenotypea for:
	Chk1 (MAPK)	Cga1 (Gα subunit)	Cgb1 (Gβ subunit)
Conidiation	None	Nearly normal	None or extremely rare
Mating	Female sterile	Female sterile, ascospore lethal	Female sterile
Virulence	Greatly reduced	Nearly normal	Greatly reduced
Colony morphology	Age-dependent autolysis	Altered	Altered
Hyphal growth pattern on surface	Wavy (normal)	Linear	Linear
Appressoria	None	None or few	None or few
Pigmentation	Decreased	Nearly normal	Increased

^aNearly normal means that the function was not lost, but quantitative differences remained to be characterized in more detail. Female sterile means that pseudothecia were formed in a cross between the mutant and a WT tester, but the hyphae composing the pseudothecia were derived from the WT parent. Data are from this work and earlier studies (12, 18). Altered colony morphology means that colonies appeared different from WT colonies but that the overall growth pattern was retained.

We observed a staining pattern suggesting nuclear disintegration in vegetative hyphae (Fig. 8) and also found a marked increase in TUNEL staining of protoplasts obtained from mycelia of the cgb1 mutant relative to the WT (Fig. 9). Nuclear disintegration is characteristic of apoptosis and indicates that the mutant may undergo premature or abnormal cell death. This is a novel phenotype for fungal signal transduction mutants. TUNEL is nevertheless an indirect method, and the patterns shown in Fig. 8 and 9 need to be studied in more detail in C. heterostrophus and other fungi for which signal transduction mutants are available. Similar patterns of DNA staining have been observed in N. crassa strains undergoing vegetative incompatibility, suggesting that the interaction of incompatible het alleles leads to programmed cell death (24). There is precedent in the animal cell literature for either the promotion or the inhibition of apoptosis by G-protein signaling pathways. The Gβγ complex was implicated as a cell death mediator in cells transfected with a mutant amyloid precursor protein (11). In mouse cardiac myocytes, β₂-adrenergic signaling via a Gi-Gβγ-phosphatidylinositol 3-kinase pathway promotes survival, and inhibition of this pathway converts β₂-adrenergic signaling from survival to apoptosis (43). A loss of the Gβ subunit SfaD protects A. nidulans against conidial lysis by the plant antifungal protein osmotin (7) which, in yeast cells, induces apoptosis (27). The mechanism by which G-protein signaling protects against cell death appears to be a higher chitin content in the cell wall of osmotin-resistant strains (7). From our data, it appears that a signal normally carried by the Gβγ dimer prevents the death of hyphal cells. The phenomenon of vegetative incompatibility (23) may provide an ideal system for testing the prediction that the Gβ subunit is involved in signaling to prevent cell death.

Our interpretations of the developmental and cell death-related phenotypes corresponding to CGB1 are not the only ones possible. The fungal Gβ subunit may have roles independent of those of G-protein heterotrimers in addition to its well-characterized functions. In animal cells, there is evidence for cytoskeleton-dependent functions of heterotrimeric G proteins at sites other than the plasma membrane, where the receptors and the heterotrimers reside (38, 40). Filamentous fungi may provide an opportunity to identify such novel functions.

We can conclude that appressorium formation, meandering growth on the host or other surfaces, conidiation, and invasive growth on the leaf are regulated by pathways that are not a linear cascade from G protein to development. The identification of genes that are specifically activated downstream of each transducer will help to provide a more complete description of the regulatory hierarchy that transduces signals from the host to promote pathogenic development. A loss of the Gβ subunit provides a strain that has nearly normal growth but is severely deficient in conidiation and virulence. Loss-of-function mutants for the MAPK Chk1 are likewise viable but show greatly reduced virulence. These signal transduction mutants will be a valuable asset in efforts to define groups of genes whose expression is specific to the conidiation and disease development pathways.