Abstract
Non-self-recognition during asexual growth of Neurospora crassa involves restriction of heterokaryon formation via genetic differences at 11 het loci, including mating type. The het-6 locus maps to a 250-kbp region of LGIIL. We used restriction fragment length polymorphisms in progeny with crossovers in the het-6 region and a DNA transformation assay to identify two genes in a 25-kbp region that have vegetative incompatibility activity. The predicted product of one of these genes, which we designate het-6OR, has three regions of amino acid sequence similarity to the predicted product of the het-e vegetative incompatibility gene in Podospora anserina and to the predicted product of tol, which mediates mating-type vegetative incompatibility in N. crassa. The predicted product of the alternative het-6 allele, HET-6PA, shares only 68% amino acid identity with HET-6OR. The second incompatibility gene, un-24OR, encodes the large subunit of ribonucleotide reductase, which is essential for de novo synthesis of DNA. A region in the carboxylterminal portion of UN-24 is associated with incompatibility and is variable between un-24OR and the alternative allele un-24PA. Linkage analysis indicates that the 25-kbp un-24-het-6 region is inherited as a block, suggesting that a nonallelic interaction may occur between un-24 and het-6 and possibly other loci within this region to mediate vegetative incompatibility in the het-6 region of N. crassa.
IN filamentous fungi, non-self-recognition among individuals of the same species occurs during both the sexual and vegetative phases of the life cycle. During the sexual phase, non-self-recognition is mediated by the mating-type locus. Two strains must be of opposite mating type for sexual reproduction to occur (review in Coppinet al. 1997). Nonself recognition during vegetative growth in filamentous fungi is regulated by het (for heterokaryon incompatibility; Glass and Kuldau 1992) or vic (for vegetative incompatibility; Leslie 1993) loci. During vegetative growth, filamentous fungi grow as multinucleate hyphal filaments that undergo repeated fusion to form a mycelial network. Hyphal fusion can also occur between different individuals to form a heterokaryon in which genetically dissimilar nuclei coexist in a common cytoplasm. However, if two individuals differ at any het loci, productive heterokaryon formation is blocked and the two individuals are said to be “vegetatively incompatible.” Hyphal fusion between vegetatively incompatible strains usually results in compartmentalization and subsequent death of the fusion cells (Garnjobst and Wilson 1956; Bégueretet al. 1994; Jacobsonet al. 1998).
Two genetic systems that mediate vegetative incompatibility in filamentous fungi have been characterized (Glass and Kuldau 1992; Bégueretet al. 1994). In nonallelic systems, an interaction between specific alleles at different het loci mediates vegetative incompatibility. In allelic systems, a vegetative incompatibility reaction results when alternative alleles at the same het locus occur together in the same cell. Allelic interactions are thought to be the major mode of vegetative incompatibility in Neurospora crassa, where heterokaryon formation is governed by at least 11 loci: 10 het loci and 1 matingtype locus (Beadle and Coonradt 1944; Garnjobst 1955; Wilson and Garnjobst 1966; Perkins 1988). Heterokaryons or partial diploids formed between isolates that contain different alleles at any one of these loci results in inhibited growth rates and abnormal colony morphology (Mylyk 1975; Perkins 1975). Of these 11 loci, allelic differences at het-6 lead to one of the most severe vegetative incompatibility reactions. Partial diploids or heterokaryons that have alternative alleles at het-6 grow ~100 times slower than compatible heterokaryons (Smithet al. 1996), are aconidial, and exhibit hyphal compartmentation and subsequent death (Jacobsonet al. 1998).
The het-6 locus was originally identified in strains that displayed inhibited hyphal growth due to heterozygous duplications in the left arm of linkage group II (Mylyk 1975). After several weeks a subset of self-incompatible het-6 partial diploids escapes from growth inhibition to near wild-type growth rates. Escape is correlated with a deletion of at least a 35-kbp region within either of the duplicated LGIIL segments, suggesting that het-6 is located within this deleted region (Smithet al. 1996). Our objective in this study was to identify the molecular determinants that mediate vegetative incompatibility at the het-6 locus. We provide data that het-6 function is governed by at least two tightly linked genes in the het-6 region and that both loci act through a nonallelic mechanism to mediate vegetative incompatibility.
N. crassa strains used in this study
MATERIALS AND METHODS
Strains and culture conditions: N. crassa strains (Table 1) were grown on Vogel's medium with or without 1.5% agar, and growth supplements, where required (Davis and de Serres 1970). Crosses were performed on synthetic crossing medium by standard methods (Davis and de Serres 1970; Perkins 1986). Partial diploid analysis (Mylyk 1975; Perkins 1975) was used to identify collected wild strains that were functionally Oak Ridge (OR) or non-OR in the het-6 region. For example, when het-6PA haplotype strains are crossed to strains with the translocation T(IIL→IIIR)AR18 (which is het-6OR haplotype, Figure 1), one-third of the viable progeny contain duplications of het-6 region and are self-incompatible. By contrast, partial OR haplotype strains show near-wild-type growth → rates and diploid progeny from crosses between T(IIL IIIR)AR18 and het-6 normal asexual reproduction. Strains that carry the non-het-6OR haplotype are distinguishable from het-6OR haplotype strains by this assay. Cosmids representing the het-6 region of linkage group (LG) II were selected from the Orbach/Sachs library [Fungal Genetics Stock Center (FGSC), Department of Microbiology, University of Kansas Medical School, Kansas City, KS] as described previously (Smith and Glass 1996; Smithet al. 1996). Escherichia coli DH5α (GIBCO BRL, Burlington, Ontario) was used as a recipient for bacterial transformations.
DNA isolation and PCR amplification: N. crassa genomic DNA was isolated by the method of Oakley et al. (1987). Plasmid and cosmid DNAs were isolated by the alkaline lysis procedure of Birnboim and Doly as described in Sambrook et al. (1989) with or without further purification by CsCl-ethidium bromide gradients or by passage through plasmid purification columns (QIAGEN, Chatsworth, CA).
Oligonucleotides for PCR amplifications were synthesized with an Applied Biosystems (Foster City, CA) 390 PCR-Mate DNA synthesizer (Carleton University Biology Department). Sequences of primers used in this study are given in Figures 3 and 6. PCR amplifications (Saikiet al. 1988) were with an Amplitron II (Barnstead/Thermolyne, Dubuque, IA) with either the Expand PCR System (Boehringer Mannheim, Laval, Québec) or with Taq DNA Polymerase (GIBCO BRL). First-strand cDNAs from 74-OR23-1VA (gift of P. J. Vierula) were used with the 3′ rapid amplification of cDNA ends (RACE) adapter primer (Frohman 1990) and internal primers from het-6 (Figure 3) and un-24 (Figure 6) for PCR analyses of transcripts. Where necessary, PCR products were cloned in the hygR vector pCB1004 (Carrollet al. 1994) or into pCRII (Invitrogen, Carlsbad, CA).
DNA hybridizations: Genomic DNA was digested with restriction endonucleases (New England Biolabs, Mississauga, Ontario, and Boehringer Mannheim), subjected to electrophoresis in 0.8% agarose, 1 × TAE (40 mm Tris-acetate, 1 mm EDTA), and transferred to Hybond membranes according to the membrane manufacturer's recommendations (Amersham, Oakville, Ontario). Bacterial colonies containing the Sachs/Orbach N. crassa genomic library in pMOcosX (FGSC; Orbach 1994) were handled by standard protocols (Vollmer and Yanofsky 1986). Probe DNAs were labeled by the random primer method (T7 QuickPrime; Pharmacia, Baie d'Urfe, Québec) with [α-32P]dCTP (Amersham) or with fluorescein-dUTP (ECL; Amersham) and hybridizations were according to the membrane manufacturer's protocol.
DNA sequencing: Subclones were sequenced on both strands with [α-35S]dATP (Amersham) and Sequenase (Pharmacia) on a model S2 sequencing apparatus (Life Technologies, Burlington, Ontario) by the method of Sanger et al. (1977) or using the ABI automated sequencing procedure (Mississauga, Ontario) at the Biotechnology Laboratory (University of British Columbia). Single-strand sequencing of the 3′ region of un-24+ from wild collected strains was with LI-COR (Lincoln, NE) 4200L automated sequencing at Canadian Molecular Research Services (Ottawa, Ontario).
DNA transformation of N. crassa: DNA transformation of spheroplasts of N. crassa strains was performed by methods similar to those described in Royer and Yamashiro (1992). Approximately 1 μg of QIAGEN- or CsCl-purified DNA was mixed with 100 μl of a spheroplast suspension (~8.0 × 107 spheroplasts/ml). Spheroplasts transformed with the hygromycin resistance vectors pMP6 and pMOcosX (Orbach 1994) and pCB1004 (Carrollet al. 1994) were selected on plates with a final hygromycin B concentration of 235 units/ml.
RESULTS
Location of cosmids with respect to het-6: The het-6 region maps to the left arm of linkage group II (LGII). Genetic analysis using translocations specific for the left arm of LGII showed that wild-type strains could be separated into two categories, OR and non-OR, on the basis of vegetative incompatibility mediated by the het-6 region (Mylyk 1975; Perkins 1975). We constructed strains C2(2)-1 and C9-2 that are near isogenic to standard Oak Ridge background strains at all het loci (Table 1; Smithet al. 1996), but that contain alternative haplotypes in the het-6 region (het-6OR haplotype and het-6PA haplotype) and het-c (het-cOR and het-cPA; the het-c locus is ~12 map units centromere-proximal to het-6; Figure 1). The PA (Panama) alleles in the het-6 region and het-c in these strains were derived from a wild-type strain from Panama (FGSC 1131), introgressed into the OR background through five (RLM58-18) and six [C2(2)-5] backcrosses (Table 1; Smithet al. 1996).
We had previously isolated cosmids that spanned the het-6 region (Smithet al. 1996). We mapped het-6 with respect to these cosmids by analyzing progeny from a cross between C5-44 ro-7 un-24 het-6OR het-cOR; trp-1 A and RLM58-18 het-6PA het-cPA pyr-4; inl a (Table 1; Figure 1) for restriction fragment length polymorphisms (RFLPs). Of 220 random ascospores, 15 progeny were recombinant for ro-7 and un-24-het-6, 27 for un-24-het-6 and het-c, and 8 for het-c and pyr-4. No progeny that contained crossovers between un-24 and het-6 were detected. On the basis of recombination frequencies in the surrounding regions and the fact that ~19 kbp separates the two genes (below), we expected about 3 progeny recombinant for these two markers. Genomic DNA was isolated from 49 of the recombinant progeny, digested with HindIII, Southern blotted, and probed with all cosmids shown in Figure 1. Two progeny, C8c-106 and -353, had crossovers in the left portion of cosmid G8:G1, just distal to un-24. Nine strains had a crossover proximal to het-6 in cosmid X14:C1 (Figure 1). These crossovers position het-6 within an ~40-kbp region covered by cosmids G8:G1 and X14:C1 (Figure 1).
Transformation assays indicate that two genes mediate het-6 incompatibility: In previous studies with het loci, it was observed that the introduction of an alternative het allele via transformation resulted in either the failure of transformants to regenerate (presumably due to the inviability of transformants carrying alternative het alleles; Glasset al. 1990) or transformants that displayed growth characteristics that were identical in phenotype to incompatible partial diploids or heterokaryons (Saupeet al. 1996). The introduction of cosmids G8:G1 and X14:C1 (Figure 1) into spheroplasts of het-6PA haplotype (strains 2190 and RLM58-18; Table 1) gave 50 and 40 times fewer transformants, respectively, than when these same cosmids were introduced into het-6OR haplotype spheroplasts (strains 74-OR23-1V and RLM57-30). This number of transformants is no more than 5% of the number of transformants that regenerate when vector DNA alone is introduced into het-6OR haplotype or het-6 PA spheroplasts. Based on this assay, the gene(s) responsible for het-6 incompatibility activity must be located in the region covered by cosmids G8:G1 and X14:C1 (Figure 1). Two nonoverlapping subclones, TLP-1-31 (~22 kbp) and TLP-1-50 (~12 kbp), were prepared by digesting G8:G1 DNA with NotI and ligating each fragment separately into the pMOcosX vector. Each subclone was transformed separately into RLM58-18 (het-6PA haplotype) and RLM57-30 (het-6OR haplotype) spheroplasts. The introduction of either TLP-1-31 or TLP-1-50 into RLM58-18 and RLM57-30 spheroplasts resulted in regeneration of ⩾40 times fewer RLM58-18 (PA) vs. RLM57-30 (OR) transformants (Figure 2). Thus, each subclone carries at least one gene with het-6-associated incompatibility activity. Incompatibility activity was further identified in a 10.5-kbp EcoRI fragment of TLP-1-31 and a 4.8-kbp HindIII fragment of TLP-1-50, which were cloned separately as p31Eco-16 and p50-8, respectively (Figure 1). Both p31Eco-16 and p50-8 exhibited het-6-mediated incompatibility activity on the basis of transformation assays.
Map of some relevant genetic markers on N. crassa LGIIL (Perkinset al. 1982) in parental strains C5-44 and RLM58-18 and selected C8c progeny with crossovers that identify the location of het-6 function. The “un-24−OR” allele has OR-incompatibility activity and is temperature sensitive for growth. Shown below LGIIL maps are the extent of the translocated region in T(IIL → IIIR)AR18 and cosmids used in this study with an expansion of G8:G1 showing the positions of subclones TLP-1-31, TLP-1-50, p31Eco-16, and p50-8. Numerous RFLPs in the region covered by these cosmids distinguish parental strains C5-44 and RLM58-18.
Characterization of the incompatibility gene in TLP-1-50: We determined the DNA sequence of the insert in p50-8 and identified a single large open reading frame (ORF) that encodes a putative 680-amino-acid (aa) polypeptide with a predicted molecular weight of 77.8 kD. The smallest fragment with incompatibility activity via transformation assays is a ~2.4-kbp ApaI/HindIII fragment that contains the entire 680-aa ORF (Figure 3). Constructs in which a 201-bp internal fragment, the C terminus, or the translation start site were deleted lacked incompatibility activity in transformation assays.
We designated the 680-aa ORF in p50-8 as het-6OR (GenBank accession no. AF206700). The putative translation initiation sequence for het-6OR, CTTCATGGCT, is similar to the N. crassa consensus sequence CAM MATGGCT (Edelmann and Staben 1994). The promoter region of het-6OR contains a large polypyrimidine-rich tract. Similar tracts in the 5′ untranslated region of other genes, such as the eukaryotic cytoplasmic RNA-binding protein PABP, ribosomal proteins, and the elongation factors EF1α and EF2, have been correlated to growth-dependent translational control (reviewed in Hornstein et al. 1999). A second polypyrimidine-rich tract is present in the 3′ region of het-6OR, immediately following the predicted stop codon. There are no predicted introns in het-6OR based on a search for N. crassa intron consensus sequences (Bruchezet al. 1993; Edelman and Staben 1994).
Spheroplasts of RLM58-18 (het-6PA haplotype) and RLM57-30 (het-6OR haplotype) were transformed with either TLP-1-50 (contains het-6OR activity and hygR), TLP-1-31 (contains un-24OR activity and hygR), or pMP6 (contains hygR gene) DNA and plated onto medium containing hygromycin B. Significantly fewer regenerating PA colonies compared to OR colonies, and relative to the pMP6 control transformations, indicate the presence of two closely linked incompatibility genes within cosmid G8:G1.
A BLAST (Altschulet al. 1990) search identified three regions of predicted amino acid sequence similarity among HET-6OR and HET-E, which is involved in nonallelic vegetative incompatibility in Podospora anserina (Saupeet al. 1995), and TOL, which mediates vegetative incompatibility of opposite mating types in N. crassa (Shiu and Glass 1999; Figure 4). The three regions of similarity among HET-6, HET-E, and TOL are in the same order in all three proteins. The first 18-aa block showed ~50% similarity, the second 34-aa block showed ~36% similarity, and the third 10-aa block showed ~90% similarity between the three proteins. Additional BLAST searches did not identify any other proteins that shared regions common to HET-6OR, HET-E, and TOL. Other protein structure motifs were also not identified in HET-6. An mRNA from het-6OR was not detected by Northern analysis; however, 3′ rapid amplification of cDNA ends (Frohman 1990) was used to obtain partial cDNAs. DNA sequence analysis of these cDNAs, from primer site 6VP4 (Figure 3) to the 3′ end of the cDNA, verified that no introns occur in this region of het-6OR. Poly(A) tails of >20 bases in length at positions 305 and 329 bases downstream of the putative stop codon were detected in two different het-6OR cDNAs.
Map of region within p50-8 containing het-6OR gene with selected restriction enzyme recognition sites. Extent of a single ORF (het-6OR) and primer positions are given below the restriction map. GenBank accession numbers are AF206700 (het-6OR) and AF208542 (het-6PA). Primer sequences are as follows: 6VP1, 5′-TCCCTCCGATATCCTTCG-3′; 6VP2, 5′-AAATCCTCCCTCTCCGCC-3′; 6VP3, 5′-CGGTAACCTGTTCAGCT-3′; 6VP4, 5′-GAGAGGTCTACAAATCC-3′; 6VP5, 5′-CCCGCTAAGCCAAGGAGTCC-3′; 6VP9, 5′-CAGCAT C A T G A T G A G C A G T A-3′; 6VP10, 5′-CGCACGTAGTCCTCGTAG-3′; 6VP15, 5′-CCCCAAGCTTGCAACTTCC-3′. Selected cloned PCR or restriction fragments shown at bottom were tested for incompatibility activity via transformation assays into het-6OR and het-6PA haplotype strains as given at bottom right; + and − indicate that construct has or does not have, respectively, incompatibility activity.
The het-6PA allele differs from het-6OR: When a het-6OR probe was hybridized to genomic DNA from strains with the het-6OR haplotype (74-OR23-1V, FGSC 3223, P4450, and FGSC 3212), a single band of either ~4.0 or 5.2 kbp was present in each strain (Figure 5A). However, when the same het-6OR probe was hybridized to genomic DNA from non-het-6OR haplotype strains (P4468, FGSC 3199, P4471, and RLM58-18), no bands were detected (Figure 5A). By contrast, a probe made from the pan-2+ gene hybridized to a band in genomic DNA from all the het-6OR and non-het-6OR strains. These data suggested that the non-het-6OR haplotype strains either lacked a het-6 gene or that the non-het-6OR gene in these strains was highly divergent in DNA sequence from het-6OR.
Using the primer pair 6VP3 and 6VP5 (Figures 3 and 5B), PCR amplification products from the het-6OR and non-het-6OR strains could be distinguished by digestion with MboI. On the basis of DNA sequence analysis of MboI sites, we expected, and observed, that functional het-6OR strains had two bands of ~500 bp each (Figure 5B). By contrast, the functional non-OR strains all had a single band of ~950 bp, suggesting that there are two alternative het-6 allele types and that they share some DNA sequence similarity.
We cloned the entire het-6PA allele from a genomic library constructed in pUC118 of strain RLM58-18 (Table 1). The het-6PA allele (GenBank accession no. AF208542) differs significantly in both DNA and predicted amino acid sequence from het-6OR. There are 444 nucleotide sequence differences between het-6OR and het-6 PA over a 2100-bp region, including two 3-bp gaps and one 24-bp gap in the alignment. All the deletion/insertion events identified maintain the same ORF. Overall, the het-6OR and het-6PA alleles show only 78% DNA sequence identity; differences are scattered throughout the entire sequence with, on average (±S.D.), 11 ± 4.5 differences every 50 bp. The two predicted protein sequences are only 68% identical but both contain the three conserved blocks of aa sequence that are also found in HET-E and TOL (Figure 4). Both conservative and nonconservative amino acid differences occur between HET-6OR and HET-6PA. As with the DNA sequence comparison, the amino acid differences between the predicted het-6OR and het-6PA proteins are scattered throughout the ORFs.
Three regions of predicted amino acid sequence similarity in N. crassa: HET-6 (6-PA from het-6PA, 6-OR from het-6OR), P. anserina HET-E (Pa-e, accession no. L28125; Saupeet al. 1995), and N. crassa TOL (Nc-t, accession no. AF085183; Shiu and Glass 1999). Amino acid positions are given at ends of conserved blocks. At the bottom, identical sites in all sequences are marked *; positions at which two residues occur are marked +.
Two alternative alleles, het-6PA and het-6OR, can be distinguished from population samples by DNA hybridization or PCR-based markers. (A) Strains that are functionally OR [74-OR23-1VA (74OR), FGSC numbers 3212, 3223, and P4450)] at the het-6 region based on partial diploid tests have HindIII restriction fragments that hybridize to a het-6OR probe (arrowheads). DNA from functionally non-OR strains (FGSC number 3199, P4468, P4471, and 58PA = RLM58-18, the het-6 PA haplotype standard strain) do not hybridize to a het-6OR probe but do hybridize to a probe made from the pan-2 gene (bands with no arrowheads, used as an internal positive control; the pan-2 gene was a gift from J. Grotelueschen and R. Metzenberg). The pan-2 fragment is polymorphic in the population sample and segregates independently from het-6. Size standard (L) is λ-DNA digested with HindIII. (B) MboI digestion of PCR products from primer pair 6VP3/6VP5 (Figure 3) yields two fragments of ~500 bp in functional OR strains whereas functional non-OR strains exhibit a single fragment of ~950 bp.
The het-6PA allele was cloned into the pCB1004 vector and used for transformation assays with C9-2 (het-6OR haplotype) and C2(2)-1 (het-6PA haplotype) strains. The introduction of the het-6PA allele into either strain did not affect transformation efficiencies or the growth phenotype of transformants. These results were repeated with six independently derived het-6PA clones from three different strains [FGSC1131, RLM58-18, and C2(2)-1]. Four het-6OR clones obtained from two strains (74-OR23-1V and C9-2) by methods similar to those used to obtain het-6PA clones all had incompatibility activity when transformed into het-6PA haplotype strains. Thus, the het-6PA allele did not confer incompatibility activity, suggesting that het-6 incompatibility function requires additional factors in the het-6PA region to elicit an incompatibility reaction in het-6OR haplotype strains.
Identification of the incompatibility gene in p31Eco-16: We used PCR to amplify portions of the 10.5-kbp p31Eco-16 fragment (Figure 6) to delineate regions that conferred incompatibility activity. Three of these PCR products, 6JP3/6JP2, 6JP3/6JP6, and 6JP10/6JP6, mediated incompatibility when introduced into C2(2)-1 spheroplasts (het-6PA haplotype) but not with C9-2 spheroplasts (het-6OR haplotype; Figure 6). The smallest region identified as having incompatibility activity was defined by primers 6JP10 and 6JP6. We refer to the incompatibility activity of these clones as un-24OR activity.
Three putative ORFs occur on the 6JP10/6JP6 fragment (Figure 6). ORF1 is 929 aa and shows amino acid sequence identity to the large subunit of type I ribonucleotide reductases (GenBank accession no. AF171697; Smithet al. 2000). Constructs that carry the entire ORF1 complement a temperature-sensitive mutation, un-24, in the large subunit of ribonucleotide reductase of N. crassa (Figure 6). The ~300-aa ORF2 shares sequence identity to membrane glycoproteins and to mucin-like proteins due mainly to an abundance of serine, proline, and especially threonine residues. The ~400-aa ORF3 has no significant similarity to other genes in the databases and is in the opposite orientation to ORFs 1 and 2.
We used RACE to determine if transcripts with poly-adenylate tails occurred in the expected positions downstream from the three putative ORFs. First-strand cDNAs from 74-OR23-1VA (het-6OR haplotype) were used with the 3′ RACE adapter primer and internal primer 6JP11 for ORFs 1 and 2 or primer 6JP12 for ORF3 (Figure 6). Southern blots of the PCR products were probed with 6JP3/6JP2 fragment. Amplification products corresponding to an ORF3 cDNA were not identified. However, two hybridizing bands were detected in PCR products from cDNAs using 6JP11: a well-defined fragment of ~1.8 kbp and a more diffuse band of ~800 bp. DNA sequence analysis of the 1.8-kbp cDNA showed that predicted introns for ORF1 were spliced out and that it terminated with a poly-A tail at ~250 bases downstream from the translation stop site. We cloned 26 putative cDNAs by gel purifying PCR products in the 800-bp range. Only three of these clones hybridized to ORF1. DNA sequence analysis of these three clones showed that none contained a poly(A) insert downstream of ORF2. All three clones also had sequence of unknown origin that did not align to genomic sequence from this region. Based on these analyses, we conclude that only un-24OR is transcribed and henceforth refer to un-24OR as the gene with incompatibility activity in p31Eco-16.
Mutant strains that carry the un-24− allele cannot grow at temperatures >34° except under high osmotic pressure (Smithet al. 2000). The un-24− allele differs from the un-24OR allele by three transition mutations in the 5′ region of the gene. To determine whether the un-24− allele conferred incompatibility, we amplified and cloned the 6JP3/6JP6 fragment carrying the un-24− allele from strain C8c-164 (Table 1) and cotransformed it (with pCB1004) into C9-2 (het-6OR haplotype) and C2(2)-1 (het-6PA haplotype) spheroplasts. The introduction of un-24− DNA into het-6PA haplotype spheroplasts resulted in incompatibility activity at both 30 and 37°. Thus, the mutations in un-24 that result in temperature-sensitive growth do not concomitantly abolish un-24OR incompatibility activity at restrictive temperatures.
Map of region of p31Eco-16 with the un-24OR gene showing selected restriction enzyme recognition sites. Of three potential open reading frames, transcripts could be verified from only ORF 1, which corresponds to the large subunit of ribonucleotide reductase, un-24OR (GenBank accession no. AF171697). Introns in un-24OR are indicated by parentheses. The locations of selected oligonucleotide primers used in this study are given below ORFs. Primer sequences are as follows: 6JP2, 5′-CCTGAGGTGTATGAGGG-3′; 6JP3, 5 ′-G G T G A C A C G G C G C T GTG-3′; 6JP6, 5′-GTGCGGGCTTAACCGCTG-3′; 6JP9, 5′-T T G C C C A T G G T G G G T TCG-3′;6JP10, 5′-GGACGAGT T C G A C T C G G C-3′; 6JP11, C T C C G G A T G A G G T T G CCG-3′; 6JP12, 5′-CGGCAACCTCATCCGGAG-3′. Selected cloned products shown at bottom were tested for (a) ability to complement (+) or not complement (−) un-24, a temperature-sensitive mutation in the large subunit of ribonucleotide reductase and (b) the presence (+) or absence (−) of incompatibility activity based on transformation assays into het-6OR and het-6PA haplotype strains.
The carboxyl terminus of UN-24 is variable between un-24OR and un-24PA: To compare the function and sequence of un-24OR to the alternative PA allele, we cloned the 6JP3/6JP6 fragment containing un-24PA allele. We found RFLPs only between un-24OR and un-24PA with AluI, MboI, MspI, RsaI, and TaqI in the 1.5-kbp 3′ coding region. The un-24OR allele contains a large insertion within this 1.5-kbp 3′ coding region that is unique to N. crassa (Smithet al. 2000; Figure 7). This region is also required for un-24OR incompatibility function, based on transformation assays (Figure 6).
The un-24PA and un-24OR alleles differ at 33 of the 113-aa positions at the C terminus (Figure 7) and by an insertion/deletion of 18 bp in this region. We determined the DNA sequence from this region from six wild-collected strains, three (3223, P4450, and 3212) with the het-6OR haplotype and three (P4468, 3199, and P4471) with the non-het-6OR haplotype (Table 1). The number of amino acid substitutions observed within either the het-6OR compatible (4) or het-6OR incompatible (3) groups (represented by O in Figure 7) was significantly fewer than the number of amino acid substitutions (33) observed between each group (represented by X in Figure 7). All the het-6OR haplotype strains contained a six-amino-acid block not found in the non-het-6 OR haplotype strains. These data show that, as with het-6, polymorphisms occur in un-24 that are correlated with het-6 specificity.
To determine if the un-24PA allele confers incompatibility and/or complements temperature sensitivity in un-24 strains, we introduced the un-24PA allele (in pCB1004) into C2(2)-1 (het-6PA haplotype), C9-2 (het-6OR haplotype), and C8c-164 (un-24− and het-6OR haplotype) spheroplasts. Although five independently derived un-24PA clones were used (two from RLM58-18, three from FGSC 1131), transformation frequencies and the phenotype of transformants were identical between het-6OR haplotype and het-6PA haplotype transformants, indicating that un-24PA does not confer incompatibility activity in an Oak Ridge genetic background. In addition, C8c-164 strains transformed with the un-24PA allele grew well at permissive temperature, but slowed and then stopped growing when transferred to restrictive temperature. Similar to het-6PA, these results indicate that there are additional factors in the het-6 region that mediate incompatibility by un-24.
DISCUSSION
In this study, RFLP analysis of progeny with crossovers that flank het-6 incompatibility activity indicated that het-6 function is encoded by a segment covered by cosmids G8:G1 and X14:C1. This placement of het-6 function corroborates that of an earlier study (Smithet al. 1996) in which het-6OR/PA partial diploids escape from self-incompatibility through deletions of at least 35 kbp of OR DNA covered by G8:G1 or of ⩾70 kbp of PA DNA covered by G8:G1, X14:C1 and part of X5:F11. These data indicate that OR and PA DNA in this region carry reciprocal incompatibility activities. By transformation analyses, we determined that two genes, het-6OR and un-24OR, in G8:G1 confer incompatibility when introduced into strains containing the het-6PA haplotype. However, the alternative alleles het-6PA and un-24PA did not cause incompatibility reactions when introduced into strains that were genetically typed as het-6OR. The best explanation for these observations is that incompatibility function in the het-6 region is mediated through a nonallelic mechanism involving at least three different loci.
PCR-RFLP assay reveals two general forms of the unique N. crassa C-terminal region of UN-24. The un-24OR allele is always associated with the het-6OR allele in strains that are functionally OR for the het-6 region and the PA allele types of un-24 and het-6 co-occur in strains that are functionally non-OR for the het-6 region. Sequence analysis of the un-24+ C terminus region for four functionally OR (74OR = 74-OR23-1VA, FGSC nos. 3212, 3223, and P4450) and four non-OR strains (FGSC no. 3199, P4468, P4471, and 58PA = RLM58-18) reveals major differences between the two allelic forms (indicated by Xs between top four OR sequences and lower four PA sequences), but not within OR nor PA forms (Os at top and bottom of sequences, respectively).
The predicted het-6 gene product is unique in sequence data banks except that it shares three regions of similarity to proteins predicted for het-e in P. anserina and tol in N. crassa, both of which are involved in nonallelic heterokaryon incompatibility function. Matingtype incompatibility requires the products of the mat A-1, mat a-1, and tol genes. TOL contains coiled-coil (amino acid positions 177–211) and leucine-rich (aa positions 804–823) domains and is postulated to interact with mating-type proteins, or gene products that are regulated by mating type, to mediate vegetative incompatibility (Shiu and Glass 1999). In P. anserina, the het-e gene product, which resembles a GTP-binding protein (concensus sequences between aa 300 and 480) with β-transducin repeats near the C terminus (Saupeet al. 1995), is also postulated to interact with the het-c product, which has characteristics of a glycolipid transfer protein (Saupeet al. 1994). Both the het-e and het-c loci have multiple allelic specificities and only specific combinations of het-e and het-c alleles confer vegetative incompatibility (Bégueretet al. 1994; Saupe et al. 1994, 1995). The three conserved regions shared by HET-6 (aa 53–242), TOL (aa 337–500), and HET-E (aa 19–144) are distinct from the previously inferred motifs in TOL and HET-E. Detection of these conserved regions in what are otherwise very different vegetative incompatibility factors is interesting since the phenotype of incompatibility reactions is similar between different filamentous fungi and involves such common stages as hyphal compartmentation by septal plug formation, vacuolization, and death (Bégueretet al. 1994; Jacobsonet al. 1998). These conserved regions, therefore, may represent specific incompatibility domains involved in either the inactivation of het gene products during the sexual phase or in protein-protein interactions with common partners that, in turn, may initiate the vegetative incompatibility reaction. The function of these conserved regions can now be methodically explored.
Approximately 19 kbp distal to het-6, we identified a second incompatibility gene, un-24OR, that encodes the large subunit of a class I ribonucleotide reductase. By comparing regions of DNA having incompatibility function to those that complement un-24, a temperaturesensitive mutation in the OR form of the large subunit ribonucleotide reductase, it is evident that the catalytic activity is not coextensive with incompatibility function (Figure 6). The temperature-sensitive form also displays vegetative incompatibility function, even at restrictive temperatures. Incompatibility function by a large sub-unit ribonucleotide reductase has not been documented previously; however, its involvement in cell cycle control and redox chemistry make it an interesting candidate for this role. The enzyme is required for the conversion of ribonucleotide precursors to deoxyribo-nucleotides used in DNA synthesis and, therefore, plays a key role in cell division. The active form of the enzyme is a tetramer comprised of large and small subunit dimers, and substrate reduction is believed to occur by a long-range radical transfer pathway through the holoenzyme (review in Eklundet al. 1997). UN-24 in N. crassa contains a large, unique inserted region near the carboxyl terminus and the two forms of this unique region (Figure 7) are correlated with het-6OR and non-het-6OR incompatibility function. Based on structural studies of ribonucleotide reductase from E. coli (Uhlin and Eklund 1994), the C terminus appears to occupy a central position near the active site, where the large and small subunit dimers interact in forming a functional tetrameric holoenzyme. Incompatibility reactions by un-24OR into het-6PA haplotype cells may disrupt the quaternary structure of the enzyme and result in loss of function, nucleotide imbalance, and cell cycle arrest. Alternatively, toxicity of the incompatible complex may be due to nonspecific transfer of the free radical to other substrates (reviewed in Eklundet al. 1997). However, the introduction of un-24PA DNA does not inhibit the growth of het-6OR haplotype strains, suggesting that incompatibility is not simply due to the formation of “toxic” PA/OR large subunit dimers. Our finding that un-24PA also does not complement the un-24 temperature-sensitive phenotype suggests that an additional factor(s) is also involved in catalytic function of the UN-24 PA form. A possible explanation for these observations is that distinct PA and OR forms of the small subunit of ribonucleotide reductase also exist in N. crassa and that specific combinations of large and small subunits affect both catalytic and incompatibility activity of the enzyme.
We found no evidence of crossovers in the 19-kbp region between un-24 and het-6 among the 220 progeny analyzed in this study, although we expected to find about three crossovers in this intergenic region. The additional six wild-type strains that we surveyed in this study also fell into two functional het-6 groups, non-OR or OR, on the basis of partial diploid analyses. The strains that typed as het-6OR haplotype contained OR-like alleles at both un-24 and het-6, while strains that typed as non-het-6OR haplotype had PA-like alleles at both un-24 and het-6. Data from 40 wild strains from a Louisiana sugarcane field (N. Mir-Rashed, D. J. Jacobson and M. L. Smith, unpublished results) and from 126 wild isolates collected throughout the N. crassa species range (N. Mir-Rashed, D. J. Jacobson and M. L. Smith, unpublished results) also showed two allelic specificities based on both partial diploid and molecular analyses. These collected wild N. crassa strains represent an unknown, but large, number of meiotic cell divisions. Taken together, these observations indicate that there are only two allelic specificities in the het-6 region and that recombination is either blocked between un-24 and het-6, or that recombination events in this region result in nonviable progeny. The DNA and amino acid sequence divergence between allele types at un-24 and het-6 characterized in this study was significant, especially for het-6. The alternative het-6OR and het-6PA alleles have 78% DNA identity and their predicted products showed only 68% amino acid identity, suggesting that recombination has not occurred in the het-6 region for an evolutionarily significant period of time. Polymorphisms associated with allelic specificity that have been maintained for long time periods and that predate speciation have been reported for the het-c locus of N. crassa (Wuet al. 1998).
This study describes two genes, un-24OR and het-6OR, that are associated with vegetative incompatibility in the het-6 region of N. crassa. Both genes appear to function through a nonallelic mechanism whereby the OR forms, but not the PA forms, are active. Additional het gene(s) carrying PA specificity are hypothesized to be closely linked to the un-24-het-6 gene pair. A close physical association and lack of recombination between un-24 and het-6 were also observed, suggesting that the het-6 region operates as an incompatibility gene complex.
Acknowledgments
We acknowledge the technical assistance of R. Tropiano and L.Johns and the reviewers' comments. The work described in this article was supported by grants from the Natural Sciences and EngineeringResearch Council of Canada to M.L.S. and N.L.G.
Footnotes
-
Communicating editor: R. H. Davis
- Received June 14, 1999.
- Accepted April 5, 2000.
- Copyright © 2000 by the Genetics Society of America
