Abstract
Vegetative incompatibility in fungi limits the formation of viable heterokaryons. It results from the coexpression of incompatible genes in the heterokaryotic cells and leads to a cell death reaction. In Podospora anserina, a modification of gene expression takes place during this reaction, including a strong decrease of total RNA synthesis and the appearance of a new set of proteins. Using in vitro translation of mRNA and separation of protein products by two-dimensional gel electrophoresis, we have shown that the mRNA content of cells is qualitatively modified during the progress of the incompatibility reaction. Thus, gene expression during vegetative incompatibility is regulated, at least in part, by variation of the mRNA content of specific genes. A subtractive cDNA library enriched in sequences preferentially expressed during incompatibility was constructed. This library was used to identify genomic loci corresponding to genes whose mRNA is induced during incompatibility. Three such genes were characterized and named idi genes for genes induced during incompatibility. Their expression profiles suggest that they may be involved in different steps of the incompatibility reaction. The putative IDI proteins encoded by these genes are small proteins with signal peptides. IDI-2 protein is a cysteine-rich protein. IDI-2 and IDI-3 proteins display some similarity in a tryptophan-rich region.
MOST filamentous fungi, such as the ascomycete Podospora anserina, possess a genetic system regulating heterokaryon formation (for review see Begueretet al. 1994). The fusion of hyphae from different wild-type isolates generally leads to the formation of inviable heterokaryotic cells. The growth of these cells stops rapidly, and they are destroyed through a degenerative and lytic process. This is referred to as vegetative or heterokaryon incompatibility. Genetic control of this phenomenon has been elucidated in several ascomycetes, including Aspergillus nidulans (Grindle 1963), Neurospora crassa (Garnjobst 1953; Perkins 1988), and P. anserina (Rizet 1952; Bernet 1967). In Podospora, incompatibility can be triggered by coexpression of two antagonistic alleles from a single het locus (allelic incompatibility) or by coexpression of antagonistic alleles from two distinct het loci (nonallelic incompatibility; Bernet 1967). Among the nine known het loci of P. anserina, five are involved in nonallelic incompatibility (het-c, het-d, het-e, het-r, and het-v) and define three nonallelic incompatibility systems (het-c/het-e, het-c/het-d, and het-r/het-v systems; Bernet 1967). In nonallelic systems, the incompatible alleles can be brought together in the same nucleus in the progeny of an appropriate cross and, thus, vegetative incompatibility occurs in homokaryotic cells. In the case of the het-r/het-v system, two different alleles have been found at each locus in wild-type isolates: het-r and het-R at the het-r locus and het-V and het-V1 at the het-v locus. het-R het-V homokaryotic spores can germinate normally, but the growth of the mycelium stops rapidly, and the lethal reaction is triggered in most cells. Such homokaryotic strains can be obtained for each nonallelic incompatibility system and have been named self-incompatible strains (SI strains). However, the lethality of the het-R het-V strain is conditional (Labarere 1973). The strain can grow at 32°, and at this permissive temperature, it displays a phenotype similar to that of wild type. If the strain is transferred to 26° (nonpermissive temperature), the incompatibility is triggered and the autolytic reaction occurs rapidly and simultaneously in all cells. This property has allowed the characterization of some biochemical aspects of the incompatibility reaction. Soon after the induction of the incompatibility reaction, a rapid decrease of total RNA synthesis (Labarereet al. 1974) and the appearance of ∼20 new proteins were observed (Boucherieet al. 1981). The presence of new proteolytic activities (Begueret 1972; Begueret and Bernet 1973) and the appearance or enhancement of other enzymatic activities, such as phenoloxidases and dehydrogenases (Boucherie and Bernet 1977, 1978), have also been described. It was concluded that the incompatibility reaction may be associated with regulation of gene expression (Boucherieet al. 1981).
Suppressor mutations of nonallelic incompatibility have been isolated. These include mod-A1 (Belcour and Bernet 1969), mod-B1 (Bernet 1971), and mod-C1 mutations (Labarere and Bernet 1977). The mod-A1 mutation restores growth of all SI strains (het-C het-E, het-R het-V, and het-C het-D), but the lytic reaction still occurs in the quiescent mycelium as opposed to growing cells (Bernetet al. 1973). The mod-B1 mutation does not modify the phenotype of SI strains, but the presence of mod-A1 and mod-B1 mutations in SI genetic backgrounds leads to a total suppression of vegetative incompatibility, restoration of growth, and suppression of the lytic reaction. The third suppressor, mod-C1, is specific for the het-R het-V strain. This mutation leads to a complete suppression of vegetative incompatibility in the SI het-R het-V strain.
Cloning and characterization of the het-s locus (Turcqet al. 1991), het-C, and het-E genes (Saupe et al. 1994, 1995), and of the mod-A gene (Barreauet al. 1998) have not provided sufficient information to elucidate the mechanism of the incompatibility reaction leading to cell death.
The goal of our study was to identify new genes, different from het genes, whose expression is induced during the course of this cell death reaction. We took advantage of the properties of the het-R het-V SI strain to demonstrate that some mRNAs were highly induced during this reaction and to identify the corresponding genes. Differential hybridization was used to isolate three such genes that constitute new tools with which to dissect the molecular mechanisms of this cell death pathway.
MATERIALS AND METHODS
Strains and growth media: The het-R het-V strain was obtained from the progeny of the cross between the het-R het-V1 strain and the het-r het-V strain. The het-r het-V strain corresponds to the s wild-type isolate from the G. Rizet and J. Bernet collections. The het-R het-V1 strain (Bernet 1967) is isogenic to the s strain, except for the het-r and het-v loci. The het-R and het-V1 genes were present in the M wild-type isolate from the same collection. The mod-A1 mutation was isolated by Belcour and Bernet (1969) and was previously named mod1-1. The mod-B1 mutation was isolated by Bernet (1971) and first named mod2-1. The mod-C1 mutation was isolated by Labarere and Bernet (1977).
Solid media used for crosses and spore germination were adapted from corn meal agar media (Esser 1974). Dihydrostreptomycin medium is a corn meal agar medium supplemented with 6 g/liter dihydrostreptomycin (Bernet and Labarere 1969). For total RNA preparation, mycelia were grown in synthetic liquid medium (Esser 1974) in Roux bottles.
Escherichia coli XL1 Blue strain (Stratagene, La Jolla, CA) was used for plasmid propagation. E. coli cells were electroporated using the E. coli pulser (Bio-Rad, Richmond, CA) following the manufacturer's recommendations.
RNA preparation and in vitro translation of poly(A)+ mRNA: Mycelia grown in liquid medium were quickly cooled to 4° by adding ice cubes and were then washed with distilled water and collected by filtration. The wet mycelium was ground in a mortar with sterile sea sand in 1 vol of a buffer containing 10 mm Tris-HCl, pH 7.4, 100 mm LiCl, 10 mm EDTA, 0.1% SDS, and 0.1 mg/ml freshly dissolved proteinase K and 1 vol of a 50% phenol, 48% chloroform, and 2% isoamylic alcohol mixture. The aqueous phase was collected after centrifugation (5 min, 10,000 × g) and reextracted twice by the same mixture; the RNA was precipitated by adding 2 vol of ethanol, collected by centrifugation, and dissolved in water. Ethanol precipitation was repeated three times. Finally, the RNA samples were dissolved in 10 mm Tris-HCl, pH 7.4, 0.5 m LiCl, 10 mm EDTA, and 0.1% SDS buffer, heated at 100° for 3 min, and subjected to chromatography on a 0.3-ml microcolumn of poly-U sepharose (IBF-France) equilibrated in the same buffer. The mRNA samples were eluted with 0.6 ml of 10 mm Tris-HCl, pH 7.4, and 0.01% SDS, concentrated by ethanol precipitation in 0.2 m LiCl, and dissolved in a small volume of water at a final concentration of 1 mg/ml. Purified mRNA samples were translated in a reticulocyte lysate (Translation Kit, reticulocyte lysate; NEN-France) using [35S]methionine in 25-μl assays, as directed by the manufacturer. Two microliters of the translation products was used for one-dimensional gel electrophoresis analysis in 12% acrylamide gel as described by Laemmli (1970), and 22 μl was used for two-dimensional gel electrophoresis analysis.
Two-dimensional gel electrophoresis: This method was a modification of the high-resolution, two-dimensional electrophoresis method described by O'Farrell (1975). The proteins translated in vitro were separated according to their isoelectric point by isoelectric focusing in 4% acrylamide gels containing 8 m urea in the first dimension (O'Farrell 1975) and subsequently by SDS-polyacrylamide gel electrophoresis (10% acrylamide), according to their molecular mass, in the second dimension (Chua and Bennoun 1975). After electrophoresis, the proteins were detected by fluorography. The gel was treated in 50 ml EN3HANCE (NEN France), as described by the manufacturer, dried, and autoradiographed at –70° for 5 days using Kodak XOMatS films (Eastman Kodak, Rochester, NY).
cDNA library construction: The cDNA library was constructed using the c-Clone II cDNA Synthesis Kit (Clontech, Palo Alto, CA) from poly(A)+ RNA extracted from the het-R het-V strain that had been transferred for 8 hr to 26°. Firststrand synthesis was primed with random hexamers. EcoRI linkers were ligated, and the cDNAs were cloned into pUC18.
PCR amplification: PCR amplification on DNA (Saikiet al. 1988) was achieved using 100 ng of reverse primer, 100 ng of M13-20 primer, and DNA corresponding to 1.7 × 106 bacteria from a cDNA library pool in a 100-μl mixture. After 12 min at 95°, DNA was amplified for 34 cycles as follows: 95° for 30 sec, 42° for 2 min, and 72° for 2 min in a thermocycler (Perkin Elmer, Norwalk, CT).
Colony hybridization: Bacterial colonies were transferred to Hybond N+ membranes (Amersham, Arlington Heights, IL) and lysed for 3 min on 3MM paper (Whatman, Clifton, NJ) soaked in 5% SDS (w/v) and 2× SSC. DNA was fixed by a 2-min microwave treatment. Membranes were then washed twice in a 3× SSC solution at 65°. Hybridization and washing were performed as for Southern and Northern blots.
Southern and Northern blot analysis: Southern blots were performed as described by Southern (1975). Northern blot analysis was performed as described (Turcq and Begueret 1987), except for the RNA from het-C het-E strain. For this strain, mycelium was grown on cellophane discs. Cultures on cellophane discs allow changes of growth media by transfer from one petri dish to another. After 3 days, the mycelia from 10 plates were scraped from the cellophane and ground with sand in a 1-ml phenol-chloroform:1 ml extraction buffer (10 mm Tris-HCl, pH 7.4, 100 mm LiCl, 10 mm EDTA, and 0.1% SDS) mixture for 3 min. One milliliter of phenol-chloroform and 1 ml of extraction buffer were added, and the mycelium was ground one more minute. This operation was repeated three times. Total RNA was purified by three extractions with phenol-chloroform and then precipitated twice with ethanol. RNA samples were finally resuspended in water at a final concentration of 5 mg/ml. When necessary, quantification of radioactivity on blots was performed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
cDNA sequencing and computer analysis: The cDNA clones were sequenced on both strands by the dideoxy chain termination method (Sangeret al. 1977) using the Sequenase kit (U.S. Biochemical Corp., Cleveland, OH), following the manufacturer's instructions. The deduced amino acid sequences were compared with known proteins from databases using the program BLAST (Altschulet al. 1997) at the National Center for Biotechnology Information (ncbi).
RESULTS
Qualitative analysis of mRNA expressed during vegetative incompatibility: Previous characterization of the incompatibility reaction has shown that the first symptoms, i.e., formation of large vacuoles and streaming of cytoplasmic granules (Beisson-Schecroun 1962), occur in the het-R het-V SI strain ∼6 hr after the temperature shift from 32° to 26° and ends after ∼15–20 hr by a complete disappearance of the cytoplasm. During this lethal reaction, a regulation of gene expression takes place (Labarereet al. 1974; Boucherieet al. 1981). This regulation could occur at the transcriptional and/or translational level. We performed a qualitative analysis of the mRNA content at different times during the incompatibility reaction to investigate the regulation of individual genes at the mRNA level. Total RNA was extracted from cultures maintained at 32° or shifted to 26° for 2–6 hr. Poly(A)+ RNA was purified, and 1 μg of each purified sample was translated in vitro using rabbit reticulocyte lysate and [35S]methionine. The resulting polypeptides were analyzed by one- and twodimensional polyacrylamide gel electrophoresis. The abundance of the majority of the synthesized polypeptides was apparently unaffected by the induction of incompatibility reaction. However, some significant variations in the peptide pattern were observed (Figure 1, A and B). Synthesis of about a dozen polypeptides was abolished or strongly decreased after the shift. The abundance of two or three polypeptides was increased and at least 11 new polypeptides could be detected 2, 4, or 6 hr after the induction of incompatibility. Some of these new polypeptides appeared transiently: they were present 2 hr after the shift, but they disappeared 2 or 4 hr later. These results demonstrate that the mRNA content of cells in which incompatibility has been triggered is qualitatively different from that of normal cells. The progress of the incompatibility reaction is, therefore, associated with a regulation of the mRNA level of specific genes whose expression is either repressed or induced. These variations of mRNA level may result from variations in transcriptional efficiency and/or mRNA stability.
—Gel electrophoresis analysis of the variation in polypeptide content of in vitro–translated mRNA from the het-R het-V strain shifted to 26°. (A) Fluorogram of a monodimensional gel electrophoresis of 2 μl of the peptides synthesized with the mRNA extracted from the culture grown at 32° (0) or shifted for 2, 4, or 6 hr to 26°. (B) Fluorograms of two-dimensional gels of 22 μl of the same samples. The first dimension was isoelectric focusing (IEF). Polypeptides ranging from 15,000 to 120,000 D were separated in SDS gels in the second dimension. Solid symbols represent polypeptides whose synthesis is abolished or strongly decreased. Open symbols represent new polypeptides.
Construction of a cDNA library enriched in sequencescorresponding to genes induced during incompatibility (idi genes): Identification of genes whose expression is increased during incompatibility can provide information on cell functions involved in cell death. To characterize such genes, the strategy described in Figure 2 was developed. A cDNA library was constructed from poly(A)+ RNA of the het-R het-V strain transferred for 8hr at 26°. Double-stranded cDNA was synthesized and cloned into the pUC18 vector. The population of recombinant plasmids was used to transform E. coli. This library was arbitrarily limited to 10,000 clones. Clones corresponding to genes that are expressed in normal cells were subtracted. This was achieved by identification of such clones by colony hybridization with radiolabeled cDNA corresponding to mRNA of the SI strain grown at 32°. This step was repeated twice, and the positive clones were discarded. The library was thus reduced to 3800 clones. This subtractive library should contain cDNAs corresponding to idi genes whose mRNA is more abundant in the het-R het-V strain 8 hr after the shift from 32° to 26°.
—Strategy used for cloning idi genes.
Identification of cosmids and characterization of genomic loci containing idi genes: The inserts of the recombinant plasmids of the subtractive cDNA library were used in hybridization experiments on a cosmid library to identify the corresponding genomic loci. To decrease the complexity of the probe, the subtractive cDNA library was fractionated into nine pools, each containing ∼450 clones. One of them was used to prepare the pool-1 probe. The inserts of the recombinant plasmids from this pool were amplified by PCR and radiolabeled. This probe was then used for hybridization on the 3500 clones of the genomic cosmid library of a wild-type strain (Deleuet al. 1993).
Positive signals were obtained for 128 clones. Among them, 92 contained ribosomal DNA. The presence of rDNA sequences in the pool-1 might result from the use of random primers rather than oligo(dT) for the construction of the library. Such clones, however, should have been subtracted from this library unless the rRNA contaminants were absent or in low abundance in the poly(A)+ RNA preparation used for the subtraction step. This could be explained by differences in the efficiency of the poly(A)+ RNA purification step or could reflect different messenger:ribosomal RNA ratios under permissive and nonpermissive conditions.
The remaining 36 positive clones were analyzed further. Cosmid DNA was extracted, restricted by three different enzymes, and analyzed by Southern blot. The membrane was first hybridized with radiolabeled total cDNA prepared from the mRNA of the het-R het-V strain grown under the permissive condition (32°). This probe was named cDNA-0h. One or more positive bands were observed for each cosmid. Their intensity is probably correlated with the abundance of the corresponding cDNA molecules in the probe. Therefore, each cosmid contains at least one gene expressed under permissive conditions. The membrane was then probed with radiolabeled cDNA produced from the mRNA of het-R het-V cells transferred at 26° for 8 hr (cDNA-8h probe). Among the 36 selected cosmids, only 10 gave differential hybridization with the two probes and a stronger signal with the cDNA-8h probe. A very strong signal was observed with this latter probe for 5 of these cosmids. They might correspond to genes that are highly expressed when incompatibility is triggered. The remaining cosmids that did not give differential hybridization with the two probes contained restriction fragment(s) that were revealed by the pool-1 probe and undetectable with the cDNA-8h probe. This discrepancy may result from a difference in the complexity between the two probes: the pool-1 probe is much less complex than the cDNA8h probe. cDNA corresponding to genes that are expressed to a low level during the incompatibility reaction was present in the cDNA-8h probe and enriched in the subtractive library. Such sequences could reveal cosmid fragments only when pool-1 was used as a probe. cDNA corresponding to genes that are expressed in low levels in both physiological conditions might also be present in the two probes. Their low abundance has not allowed their subtraction from the cDNA-8h library. Such sequences could also reveal cosmid fragments only when pool-1 was used as a probe. This indicates that the subtractive library contains sequences that are preferentially expressed during vegetative incompatibility and, probably, also sequences whose expression is low and not modified during the progress of lethal reaction.
The 10 cosmids that gave differential hybridization with the two probes with a stronger signal with the cDNA-8h probe were further analyzed. The five restriction fragments that gave the strongest signals with the cDNA-8h probe and little or no signal with the cDNA-0h probe were subcloned from the cosmids. Cross-hybridization experiments revealed that among these five restriction fragments, only three were distinct. They were the 6.5-kb HindIII fragment from cosmid 1, the 4.5-kb HindIII fragment from cosmid 13, and the 1.5-kb EcoRI fragment from cosmid 2 (Figure 3). These three restriction fragments do not contain common sequences. Among the five remaining restriction fragments, two cross-hybridized. In the simplest situation in which one fragment defines one gene, we have identified 7 different idi genes. From two-dimensional electrophoresis analysis of the translation products of mRNA from autolytic cells, at least 11 additional major polypeptides were found to be present. Thus, we can expect to characterize up to 11 idi genes.
—Southern blot analysis of three cosmids encoding idi genes. Cosmid DNA was extracted and restricted with different enzymes, and fragments were separated in a 0.8% agarose gel. (A) The resulting blot was probed with a 32P-labeled cDNA probe specific to mycelium grown at 32° (cDNA 0 probe). (B) The same blot was probed with a 32P-labeled cDNA probe specific to mycelium grown at 32° and transferred 8 hr at 26° (cDNA 8h probe).
With another pool of the subtractive cDNA library, we have again identified the same 10 cosmids. No additional restriction fragments corresponding to new idi genes were selected. Each pool seems, therefore, to be representative of the entire subtractive library for sequences highly induced upon incompatibility. This indicates that we cannot expect to characterize many more highly induced genes from this subtractive library. Any remaining relevant genes might be induced only to a low level 8 hr after the induction of incompatibility and might belong to cosmids such as those bearing fragments that are revealed by the pool-1 probe and undetectable with the cDNA-8h probe.
Expression of the idi genes is conditional and associated with vegetative incompatibility in the het-R het-V strain: Each of the three selected restriction fragments isolated from cosmids 1, 2, and 13 should contain at least one gene preferentially expressed during incompatibility. The pattern of expression of the corresponding genes was examined by Northern blot analysis on total RNA extracted from the het-R het-V strain grown at 32° or transferred to 26° for different times (3, 6, or 8 hr). Results are reported in Figure 4. When the RNA was probed with the 6.5-kb HindIII fragment from cosmid 1, an mRNA of ∼1.1 kb was detected (Figure 4A). The corresponding gene was named idi-1. With the 4.5-kb HindIII fragment from cosmid 13 as a probe, an mRNA of ∼1 kb was observed (Figure 4B). The corresponding gene was named idi-2. The idi-1 and idi-2 mRNAs were not detected in total RNA of the SI strain grown at the permissive temperature. For these two genes, the amount of mRNA increases and reaches a maximum after 3 hr for idi-2 and still increases after 8 hr at 26° for idi-1. Thus, transcription of idi-1 and idi-2 is observed only after the temperature shift. When total RNA was probed with the 1.5-kb EcoRI fragment from cosmid 2, an mRNA of ∼1.3 kb was detected (Figure 4C). The corresponding gene was named idi-3. The idi-3 mRNA was also revealed at a low level when total RNA of the strain grown at 32° was used. After the transfer to 26°, the amount strongly increases and reaches a maximum after 6 hr under the nonpermissive condition. This maximal level is about 8-fold higher than that in cells grown under permissive conditions (Figure 4C). The expression of idi genes was quantified by comparison with that of another gene, the het-C gene whose expression level is similar to that of a housekeeping gene (Saupeet al. 1994) and constant during incompatibility. The maximal expression level of the three idi genes was found to be ∼10-fold higher than that of het-C. Thus, the expression of these three genes is conditional and highly induced in the het-R het-V strain after the temperature shift that triggers the incompatibility reaction.
—Northern blot analysis of idi gene expression. Total RNA was extracted from cells grown at 32° and from cells grown at 32° and then transferred for 3, 6, or 8 hr at 26°, as described in materials and methods. RNA samples (30 μg/lane) were separated by electrophoresis through formaldehyde-agarose gels and transferred onto nylon membranes. The Northern blot was successively probed with 32P-labeled idi-1 (A), idi-2 (B), and idi-3 (C) fragments. The 28S rRNA was measured as a control to establish the relative loading of RNA in each lane (D).
To determine if the expression of the idi genes is specific to the het-R het-V strain rather than simply associated with the temperature shift, total RNA of the het-r het-V strain was analyzed. This strain is isogenic to the het-R het-V strain, except for the het-r locus. het-r and het-V alleles are compatible, and the strain displays a wild-type phenotype at any temperature. Total RNA was extracted from this strain grown at 32°, or grown at 32° and transferred to 26° for 8 hr, and analyzed by Northern blot with the idi-1 (6.5-kb HindIII fragment), the idi-2 (4.5-kb HindIII fragment), or the idi-3 (1.5-kb EcoRI fragment) fragments as probes. We observed that under both conditions, the expression of the three idi genes is similar to that observed in the het-R het-V strain grown at 32°: no hybridization was observed for idi-1 and idi-2, and a faint band was observed for idi-3. Therefore, the induction of idi gene expression is associated with the development of vegetative incompatibility triggered by the incompatible het-R and het-V genes.
Expression of idi genes in different genetic backgrounds
Expression of idi genes in other SI strains: The presence of suppressor mutations affecting all three nonallelic incompatibility systems described in P. anserina (het-r/het-v, het-c/het-e, and het-c/het-d) led us to propose that common steps may exist in the vegetative incompatibility reactions triggered by different incompatible het genes (Labarere 1973). We have, therefore, investigated the expression of idi genes in the het-C het-E strain. In this strain, self-incompatibility results from the coexpression of the incompatible het-C and het-E genes. Self-incompatibility is partially suppressed by the addition of dihydrostreptomycin to growth medium (Bernet and Labarere 1969; Bernetet al. 1973). The het-C het-E SI strain can grow on such a medium, but cell lysis is not completely suppressed, and a high proportion of dead cells is observed in the quiescent mycelium. When the strain grown on medium containing dihydrostreptomycin is transferred onto a medium lacking this antibiotic, growth stops rapidly and cell lysis soon extends to the entire mycelium. Total RNA was isolated from the strain grown under the permissive condition and after a transfer for 15 hr on medium without dihydrostreptomycin. The samples were analyzed by Northern blotting. The result reported in Table 1 indicates that idi-1 and idi-3 genes are expressed in the het-C het-E SI strain. The mRNA levels of idi-1 and idi-3 were similar in the strain grown on dihydrostreptomycin-containing medium or grown on this medium and then transferred for 15 hr on a medium without dihydrostreptomycin. No increase of the levels of these mRNAs was observed when dihydrostreptomycin was removed, even when the incubation time was extended to 36 hr. In contrast, the idi-2 mRNA was not detected in the het-C het-E SI strain. This result allows us to distinguish between the incompatible reaction triggered by the het-r/het-v or the het-c/het-e systems.
There are common features in vegetative incompatibility, whether it is triggered by allelic or nonallelic, incompatible het genes. The progress of the cell death reaction appears very similar at a microscopic level. At the macroscopic level, both genetic backgrounds result in the formation of a barrage in the region where incompatible strains fuse. However, no suppressor common to allelic and nonallelic systems has been identified so far, suggesting that the pathways leading to cell death may be different when incompatibility results from the expression of allelic or nonallelic, incompatible het genes. As the expression of idi-1 and idi-3 genes seems to be a good marker to follow the progress of the incompatibility reaction triggered by the coexpression of nonallelic het genes, we have examined their expression during vegetative incompatibility under the control of the allelic het-s/het-S system. The het-s– and het-S– incompatible alleles have been cloned (Turcqet al. 1990), and strains containing these two incompatible alleles can be obtained by ectopic integration of the het-s allele by transformation of a het-S strain. The growth of this recombinant strain is highly impaired, but it can be partially restored on complete medium at 18° (V. Coustou, personal communication). RNA was extracted from mycelium grown under these sublethal conditions and analyzed by Northern blot. No mRNA could be detected for any of the three idi genes (Table 1). This provides additional evidence for the existence of different pathways in allelic and nonallelic vegetative incompatibility.
Expression of idi genes correlates directly with the cell death reaction associated with nonallelic vegetative incompatibility: The expression of idi genes could be implicated in different ways in vegetative incompatibility. Their protein products could be directly involved in the cell death reaction. Alternatively, the expression of the idi genes could be simply induced as a consequence of the cell death reaction. In this case, idi genes may, for instance, behave as HSP genes induced under different stress conditions. We have, therefore, verified that idi genes are not induced under heat shock conditions. Total RNA was extracted from the compatible het-r het-V strain grown at optimal growth temperature (26°) and then shifted for 5–30 min to 37°, a condition that is known to induce the expression of HSP genes in P. anserina (Loubradouet al. 1997). Analysis by Northern blot did not reveal any expression of the three idi genes (data not shown). Therefore, it is unlikely that the idi genes encode general stress proteins. This is also supported by the fact that their expression is not induced when cell death is triggered by the het-s/het-S allelic system. The induction of these genes seems to be more directly associated with nonallelic vegetative incompatibility.
To test this hypothesis, we have analyzed the expression of idi genes in different strains that contain mutations in mod genes that suppress incompatibility. These mutations were isolated using a screen based mostly on suppression of the lethality of SI strains containing different combinations of incompatible, nonallelic genes. These mod mutations have allowed us to identify two different consequences of vegetative incompatibility in the SI strains that can be at least partly dissociated: an inhibition of the growth and a cell lysis reaction. The mod-A1 mutation restores the growth of the different SI strains, but cell lysis is still present in the quiescent mycelium. This autolytic reaction can be suppressed by the addition of the mod-B1 mutation, which alone has no effect on vegetative incompatibility. The effects of mod-A1 and mod-B1 are similar on the het-R het-V, het-C het-E, and het-C het-D SI strains. In contrast, the mod-C1 mutation both restores the growth and abolishes autolysis of the het-R het-V SI strain only. It has no suppressor effect when incompatibility is triggered by the other nonallelic, incompatible het genes (Labarere and Bernet 1977). We have investigated the expression of idi genes in the presence of these different suppressors (Table 1). Northern blot analysis of the het-R het-V mod-C1 strain showed that no induction of idi-2 and idi-3 could be detected after the transfer from 32° to 26°, even after 8 hr at 26°. In comparison, the amount of these mRNA was already highly enhanced after 3 hr at 26° for the het-R het-V strain. This suppressor effect is incomplete for idi-1. A very low amount of RNA is detectable after 8 hr at 26°. However, this amount is ∼25 times lower than in the absence of the mod-C1 mutation. Thus, the increase in idi-1, idi-2, and idi-3 mRNA is either strongly or completely suppressed by the mod-C1 mutation.
Expression of the three idi genes has also been determined in the het-R het-V mod-A1 mod-B1 strain, a genetic background that prevents incompatibility. As shown in Table 1, no induction of idi-1 and idi-3 genes was observed when the strain was transferred from 32° to 26°. Under the same conditions, the expression of idi-2 was not modified and the mRNA level was similar to that detected in the het-R het-V strain. The individual presence of mod-A1 or mod-B1 mutations had quite different consequences. No effect of mod-A1 was detected; the mRNA level of the three idi genes was highly enhanced after the transfer to 26°, reaching a level similar to that observed for the het-R het-V strain. In the het-R het-V mod-B1 strain, the idi-1 and idi-3 genes are expressed at the same level as in the het-R het-V strain. The amount of the idi-2 mRNA is also increased, but to a limited extent–only one-tenth of that found in the het-R het-V strain when self-incompatibility has been triggered. Under the same conditions, idi-1 and idi-3 mRNA levels are stable. This suppression of the expression of idi-2 mRNA is the first evidence of an effect of the mod-B1 mutation in the absence of the mod-A1 mutation. Surprisingly, idi-2 gene expression was not decreased in the double mod-A1 mod-B1 mutant strain; therefore, this mod-B1 phenotype is apparent only in the absence of the mod-A1 mutation. This indicates that idi-3 expression is also regulated both by mod-A and mod-B functions, as are idi-1 and idi-3 genes. However, this control seems to be exerted in a different way.
Thus, in both het-C het-E and het-R het-V strains, unlike idi-2 induction, idi-1 and idi-3 induction was always correlated with occurrence of cell lysis.
The three idi genes encode small proteins with signal peptides: The idi-1, idi-2, and idi-3 fragments were used to screen the subtracted cDNA library, and 85, 37, and 81 cDNAs were recovered, respectively. These cDNAs were used to construct contigs that were sequenced. These sequences were translated from the first ATG, which initiate 603-, 471-, and 588-bp-long ORFs, respectively.
Translation products of the idi-1, idi-2, and idi-3 ORFs correspond, respectively, to an acid 201-(amino-acid) polypeptide named IDI-1, an acid 157-aa polypeptide named IDI-2, and a 196-aa polypeptide named IDI-3. The primary sequence of the IDI-2 polypeptide is rich in cysteine (8.9%) and tryptophan (3.8%) residues. The IDI-3 polypeptide is also rich in tryptophan residues (3.1%). From a search in protein sequence databases, no significant similarity was found between IDI proteins and known proteins. One significant feature of these three proteins is the existence of a putative signal peptide at the NH2 terminus (Figure 5), as determined using the SignalP program (Nielsenet al. 1997). Using the BLAST program, which allows the comparison between two proteins, we observed a significant similarity (expect = 7 × 10–4) between IDI-2 and IDI-3 proteins over a 40-aa stretch that includes three of the six tryptophan residues (Figure 5). This suggests that these two proteins may be related.
DISCUSSION
Vegetative incompatibility in P. anserina induces a lethal reaction that ends in cell lysis. This cell death reaction is characterized by an intensive vacuolization of the cytoplasm that could be related to the general cell swelling observed during necrosis rather than to apoptosis (Bujaet al. 1993). No nuclear DNA fragmentation was observed during the course of this reaction (C. Clavé, unpublished results). Here, we demonstrate that the mRNA content of cells undergoing lysis is different from that of vegetative cells. Some genes are switched on or stimulated (idi genes), and others are unaffected or switched off during this cell death reaction. This difference in mRNA contents is likely to result from a transcriptional regulation of gene expression. However, a regulation of mRNA stability in some fungi was described in relation with the control of carbon catabolite repression (Cereghino and Scheffler 1996) and nitrogen metabolite repression (Plattet al. 1996). Further characterization of idi genes and their promoters will allow us to estimate the relative influence of transcriptional regulation and mRNA turnover in the observed variations of the levels of idi transcripts.
—Predicted amino acid sequences from idi cDNAs. Potential signal peptides are underlined. The conserved domain between IDI-2 and IDI-3 is underlined in a dashed line. GenBank accession numbers for idi ORFs are AF067181 (idi-1), AF067182 (idi-2), and AF067183 (idi-3).
The idi-1, idi-2, and idi-3 genes characterized herein display a very high expression level, which suggests that they could be direct actors in rather than regulators of this reaction. Moreover, the three putative protein products are small proteins with a signal peptide. Taking into account the intensive vacuolization of the cells during the course of the incompatibility reaction, it can be hypothesized that these proteins might be vacuolar. IDI-2 and IDI-3 proteins display some similarity over a 40-aa stretch. This might suggest a partial redundancy between these two idi-encoded functions.
The major difference between the expression of these three idi genes is that the expression of idi-1 and idi-3 mRNA is induced by different nonallelic systems and suppressed by common suppressors of nonallelic systems when idi-2 mRNA is not. This could indicate that idi-1 and idi-3 genes are involved in general cell death mechanisms that are common to all nonallelic systems, unlike idi-2, in which induction is specific of the pathway triggered by the het-r/het-v system. The expression of idi-1 and idi-3 was not induced when the cell death reaction was triggered by the allelic het-s/het-S system. Thus, the steps controlled by idi-1 and idi-3 are not common to all incompatibility systems. However, some downstream steps may be common to all these systems. This is suggested by the high similarity in the cytological aspects of the cell death reaction that occurs after the fusion of the hyphae from incompatible strains regardless of the het genes involved.
The function of idi genes may be directly involved in steps leading to cell death, but these genes could also be involved in the survival of some cells that escape to this program of cell death in SI strains (Bernet 1965). It cannot be excluded that the function of these genes might not be directly related to cell disintegration or cell protection. The idi genes might be induced as a consequence of the control of their expression by het gene products. Indeed, it has been proposed that het genes would control some differentiation steps in the life cycle of the fungus, such as the transition from the quiescent to the proliferating stage or the formation of female organs (Bernet 1992), and that vegetative incompatibility would correspond to a disorder of this regulation caused by the formation of abnormal complexes between the products of antagonistic het genes (Begueretet al. 1994). This alternative view is supported by the expression pattern of idi-2, which is not strictly related to the occurrence of the cell lysis reaction. idi-2 might be involved in such differentiation steps. Inactivation of the idi genes and determination of the associated phenotypes will be used to discriminate between these three possible functions: cell death, cell survival, and differentiation.
Acknowledgments
We greatly thank Dr. H. Boucherie for assistance in the two-dimensional gel electrophoresis analysis and Dr. S. Saupe for critical reading of the manuscript. This work was supported by a grant from Association pour la Recherche sur le Cancer. N.B. is funded by the French Ministère de la Recherche et de l'Education Nationale.
Footnotes
-
Communicating editor: R. H. Davis
- Received May 26, 1998.
- Accepted June 11, 1998.
- Copyright © 1998 by the Genetics Society of America
