Correspondence: Sung-Hwan Yun, Department of Medical Biotechnology, Soonchunhyang University, Asan 336-745, South Korea. Tel.: +82 41 5301288; fax: +82 41 5303085; e-mail: firstname.lastname@example.org
Members of the Fusarium graminearum species (Fg) complex, which are homothallic ascomycetous species, carry two opposite mating-type (MAT) loci in a single nucleus for controlling sexual development. We investigated the roles of three (MAT1-1-1, MAT1-1-2, and MAT1-1-3) and two (MAT1-2-1 and MAT1-2-3) transcripts located at both loci in representative Fg complex species (F. graminearum and Fusarium asiaticum). In self-fertile F. graminearum strains, the transcript levels of MAT1-1-1, MAT1-2-1, and MAT1-2-3 peaked 2 days after sexual induction (dai) and then remained high until 12 dai, whereas MAT1-1-2 and MAT1-1-3 transcripts reached peak levels between 4 and 8 dai. In contrast, all of the MAT transcripts in self-sterile F. asiaticum strains accumulated at much lower levels than those in F. graminearum during the entire time. Targeted gene deletions confirmed that MAT1-1-1, MAT1-1-2, MAT1-1-3, and MAT1-2-1 were essential for self-fertility in F. graminearum, but MAT1-2-3 was not. All MAT-deleted strains (except ΔMAT1-2-3) produced recombinant perithecia when outcrossed to a self-fertile strain. These results indicate that developmental up-regulation of the individual MAT genes in both a proper fashion and quantity is critical for sexual development, and that alterations in the gene expression could be attributed to the variation in self-sterility among the Fg complex.
Fusarium graminearum (telomorph: Gibberella zeae), an ascomycetous fungus causing Fusarium head blight of cereal crops (McMullen et al., 1997), is considered a member of the F. graminearum species (Fg) complex, which consists of over 15 phylogenetically distinct species (also known as lineages) distributed worldwide (O'Donnell et al., 2000, 2004, 2008; Starkey et al., 2007; Yli-Mattila et al., 2009; Sarver et al., 2011). All F. graminearum sensu stricto strains (lineage 7) can produce sexual progeny (ascospores) without contact with a sexual partner, which is known to be important for initiating the disease cycle (Trail et al., 2002). However, this self-fertility varies among the other members of the Fg complex. For example, Fusarium asiaticum (lineage 6), which is widely distributed in Asia, exhibited a lower self-fertility than the highly fertile F. graminearum strains (Lee et al., 2012).
The sexual ability of the Fg complex is controlled by master regulators called mating-type (MAT) loci (Debuchy & Turgeon, 2006). Unlike their heterothallic relatives, the Fg complex strains carry two MAT loci (MAT1-1 and MAT1-2) in a single nucleus for controlling sexual development, but the structural organization of individual MAT genes is similar to those in Sordariomycetes fungi (e.g. Neurospora crassa, Podospora anserina, and Sordaria macrospora; Yun et al., 2000; Debuchy & Turgeon, 2006). Three (MAT1-1-1, MAT1-1-2, and MAT1-1-3) and one (MAT1-2-1) transcripts are located at both loci, among which the deduced product of MAT1-1-1 carries a DNA-binding motif called the alpha box, those of MAT1-1-3 and MAT1-2-1 contain an HMG box domain, and that of MAT1-1-2 includes a newly proposed DNA-binding PHP domain (Yun et al., 2000; Debuchy & Turgeon, 2006). An additional transcript, MAT1-2-3, has been proposed as a new MAT gene at the MAT1-2 locus in the heterothallic Fusarium verticillioides and F. graminearum (Martin et al., 2011). However, it contains no known DNA-binding motifs and its role(s) in sexual development are unknown. To date, gene deletion analyses have confirmed that both MAT loci are essential for sexual development in F. graminearum (Lee et al., 2003; Desjardins et al., 2004) but the functional requirement for the individual MAT genes, except MAT1-2-1, has not been intensively demonstrated. Recently, the transgenic strains deleted for MAT1-1-1 and MAT1-1-3, respectively, have become available (Son et al., 2011). Despite the importance of MAT loci in sexual development, transcriptional expression or regulation of MAT genes has remained largely unknown in filamentous fungi. Only a few reports are available (Leubner-Metzger et al., 1997; Czaja et al., 2011), and only the expression pattern of MAT1-1-2 is available from microarray analysis in F. graminearum (Hallen et al., 2007).
The functions of each MAT gene in a self-fertile S. macrospora have been determined; Smt A-1 and Smt A-3, which are comparable to MAT1-1-1 and MAT1-1-3, respectively, are dispensable for fruiting body formation (Klix et al., 2010). Transcript profiling analyses revealed that at least several hundred genes were differentially regulated in fungal strains lacking MAT genes, but which genes are directly controlled by the MAT loci is unknown. In addition, the sexual defect in asexual fungal species such as Alternaria alternata and Bipolaris sacchari is not attributable to the non-functionality of their MAT genes (Sharon et al., 1996). Rather, genes downstream in the regulatory pathways probably controlled by MAT proteins (i.e. the target genes of the MAT proteins) may be nonfunctional in these asexual species (Hornok et al., 2007). However, the variation in the capacity for sexual mating in the Fg complex at the level of MAT loci has not been investigated.
Therefore, the objectives of this study were (1) to compare the expression patterns of individual MAT transcripts in representative strains of fully self-fertile F. graminearum and self-sterile F. asiaticum to investigate the variation in sexual capacity within the Fg complex; and (2) to determine the functional roles of each MAT gene, including MAT1-2-3, in F. graminearum sexual reproduction.
Materials and methods
Fungal strains and culture conditions
Fusarium graminearum PH-1, Z3639, and Z3643 were used (Bowden et al., 2008), which belong to lineage 7 of the Fg complex (O'Donnell et al., 2000). T43ΔM2-2 was a MAT1-2-1-deleted strain derived from Z3643 (Lee et al., 2003). FgGFP-1, constructed from Z3643 in this study, was a self-fertile strain carrying a green fluorescence protein (GFP) gene. Three F. asiaticum strains (lineage 6) were isolated from Korean cereals: SCKO4 (Kim et al., 2005) from barley and the remaining two (ESR3R6 and ASR1R2) from husked seeds of rice harvested in 2010. The rice strains (ESR3R6 and ASR1R2) are available from the Korean Agricultural Culture Collection (KACC; http://www.genebank.go.kr) under KACC no. 46428 and 46429, respectively. Fusarium graminearum strains are highly self-fertile, whereas all F. asiaticum strains are self-sterile. These wild-type and MAT-deleted strains, derived from Z3643 or Z3639, were stored in 20% glycerol at −70 °C. For genomic DNA extraction, each strain was grown in complete medium (Leslie & Summerell, 2006) at 25 °C for 72 h. Sexual reproduction was induced on carrot agar (Leslie & Summerell, 2006), as described previously (Lee et al., 2003). For outcrosses, the mycelial plug of a MAT-deleted (ΔMAT) strain was placed on carrot agar and incubated at 25 °C for 7 days. A conidial suspension (105 conidia mL−1) of the FgGFP-1 strain was dropped onto mycelia of the ΔMAT strain and incubated for an additional 5–10 days (Lee et al., 2003).
Nucleic acid extraction, PCR conditions, and DNA gel blot hybridization
Fungal genomic DNA and total RNA were extracted as described previously (Leslie & Summerell, 2006; Kim & Yun, 2011). PCR primers (Supporting Information, Table S1) were synthesized by the Bioneer Corporation (Chungwon, Korea). Quantitative real-time PCR (qPCR) was performed with the SYBR Green Super Mix (Bio-Rad) using the first-strand cDNA synthesized from total RNA (Lee et al., 2010; Kim & Yun, 2011). The amplification efficiencies of all genes were determined as described previously (Kim & Yun, 2011). Gene expression was measured in three biological replicates from each time point. Statistical analysis was performed by anova by Duncan's multiple range test. GzRPS16 (FGSG_09438.3) and EF1A (FGSG_08811.3) were used as endogenous controls for data normalization (Kim & Yun, 2011). The amount of MAT1-1-2 transcript from a 2-day-old vegetative sample of ASR1R2 was used as a reference for comparison. DNA gel blot was prepared (Sambrook & Russell, 2001) and hybridized with biotinylated DNA probes to be prepared by BioPrime DNA labeling system (Invitrogen), followed by developing using a BrightStar® BioDetect™ Kit (Ambion). All procedures in chemiluminescent detection followed the protocol provided.
Vector construction and fungal transformation
DNA constructs for deletion of individual MAT genes from the genomes of F. graminearum Z3643 or Z3639 were created for a split marker recombination procedure (Catlett et al., 2003). The 5′ and 3′ flanking regions of the target MAT gene were amplified by PCR using the primers in Table S1. The geneticin resistance gene cassette was amplified from pBCATPH with the primers Hyg/for and Hyg/rev (Kim et al., 2008). The three amplicons were mixed in a 1 : 1 : 3 molar ratio, fused in a second round of PCR, and used as a template to generate split markers with the new nested primer sets (Table S1). The amplified products were added into the protoplasts of wild-type F. graminearum strains for transformation (Kim et al., 2011; Lee et al., 2011).
MAT transcript expression in F. graminearum and F. asiaticum
Using qPCR, we compared the accumulation of individual MAT transcripts at nine time points on carrot agar in six F. graminearum and F. asiaticum strains to determine the time course of gene expression during both the vegetative growth and the sexual cycle, as well as variation in the expression patterns between these species. The average PCR efficiency of the primer sets for individual MAT genes ranged from 1.93 to 1.99. In all self-fertile F. graminearum strains examined, all MAT gene transcripts accumulated more highly (with the levels ranging from c. 10- to 140-fold) during fruiting body (perithecia) formation than during growth of aerial mycelia: no significant differences in MAT transcript levels were found during vegetative growth (Fig. 1, Table S2). However, the pattern of transcript accumulation differed between MAT genes during perithecia formation. Accumulation of MAT1-1-1, MAT1-2-1, and MAT1-2-3 transcripts peaked at an early stage of sexual development (2 days after perithecial induction; c. 25- to 100-fold higher than during the vegetative growth), decreased abruptly at 4 dai, then subsequently increased, and remained at high levels until 12 dai, when mature perithecia formed. In contrast, MAT1-1-2 and MAT1-1-3 transcripts reached peak levels during the late stages of sexual development (between 4 and 8 dai; c. 10- to 20-fold higher than during the vegetative growth). Moreover, the average expression level of MAT1-1-2 and MAT1-1-3 transcripts at their peaks was c. 10-fold lower than that of MAT1-1-1, MAT1-2-1, and MAT1-2-3 transcripts at the peaks (Fig. 1, Table S2).
In all self-sterile F. asiaticum strains examined, the MAT1-1-1, MAT1-2-1, and MAT1-2-3 expression was also highly induced at the early stage, similar to those in F. graminearum described above, but the transcript levels during the entire sexual cycles were c. 10- to 20-fold lower than those in F. graminearum (Fig. 1, Table S2). The later sexual stage-specific patterns of MAT1-1-2 and MAT1-1-3 shown in F. graminearum were significantly altered in F. asiaticum. Neither MAT1-1-2 nor MAT1-1-3 was significantly induced at any time point during the sexual development compared with those during the vegetative growth (Fig. 1, Table S2).
Functional requirement for MAT genes in F. graminearum strains
Integration of a transforming DNA construct for gene deletion into the fungal genome via a double cross-over resulted in a F. graminearum Z3643 or Z3639 strain lacking individual MAT genes (designated ΔMAT1-1-1, ΔMAT1-1-2, ΔMAT1-1-3, ΔMAT1-2-1, and ΔMAT1-2-3; Fig. 2a). Targeted gene deletion was verified by DNA blot hybridization (Fig. 2b). In carrot agar cultures of the wild-type Z3643 or Z3639 strains, protoperithecia began to form at 3 dai and developed into fully fertile perithecia after 6–7 dai, which carried asci containing eight ascospores. However, those formed in the ΔMAT1-1-1, ΔMAT1-1-2, and ΔMAT1-1-3 strains were smaller than normal perithecia from wild-type cultures, and carried neither asci nor ascospores even 4 weeks after perithecial induction (Fig. 3). Barren perithecia in the ΔMAT1-1-1 strains were smaller than those in the ΔMAT1-1-2 and ΔMAT1-1-3 strains, but the numbers of barren perithecia from all of these ΔMAT strains were similar to those of fertile wild-type strains (Fig. 3). In addition, the ΔMAT1-2-1 strain (T43ΔM2-2) produced no perithecia on carrot agar, as reported previously (Lee et al., 2003). Unlike these MAT deletion strains, the ΔMAT1-2-3 strains produced a similar number of normal fertile perithecia to Z3643, demonstrating that MAT1-2-3 are dispensable for perithecia formation in F. graminearum (Fig. 3). The phenotypes of all of the MAT-deleted strains examined, other than fertility (e.g. mycelial growth, pigmentation, and conidiation), were not different from those of their wild-type progenitor (data not shown).
To determine whether self-sterile ΔMAT strains retain the ability to outcross, we set up sexual crosses of a transgenic F. graminearum (FgGFP-1) carrying a GFP gene to each of the ΔMAT1 strains, wherein the ΔMAT strains were forced to act as the female parent. All outcrosses except that of the ΔMAT1-2-3 strain produced morphologically normal mature perithecia with asci containing eight ascospores; the numbers of perithecia formed in the outcrosses were reduced to c. 30% of the level of the self or wild-type strains based on examination of more than 100 perithecia. All perithecia from each outcross examined yielded eight tetrads, of which four fluoresced (Fig. 4), indicating the occurrence of normal meiosis for production of recombinant progeny. However, the outcross of the self-fertile ΔMAT1-2-3 strain produced no recombinant perithecia, as in the wild-type strain [i.e. FgGFP1 (male) × Z3643 (female)].
Availability of individual MAT transcript expression profiles in various fungal strains provides clues to the variation in self-fertility among the Fg complex at the level of MAT loci. The differing expression pattern of individual MAT genes in all F. asiaticum strains compared with F. graminearum strains can be attributable to the defect in self-fertility in these strains. Failure to up-regulate MAT1-1-2 and MAT1-1-3, and reduced up-regulation of MAT1-1-1, MAT1-2-1, and MAT1-2-3 during the entire sexual cycle may cause a putative set of genes under the control of these MAT genes to be abnormally or not properly expressed, leading to self-sterility in F. asiaticum. Nevertheless, similarity in expression patterns of MAT1-1-1, MAT1-2-1, and MAT1-2-3 in all F. graminearum and F. asiaticum strains examined cannot exclude the possibility that the early induction pathway of sexual development controlled by these genes is not responsible for the self-fertility differences in these Fg complex strains. To test these postulates, a comparison of genome-wide expression profiles using combinations of wild-type F. graminearum and F. asiaticum strains and their MAT-deleted strains would be necessary. To date, several approaches have been used to identify the target genes of MAT loci in several filamentous fungi (Qi et al., 2006; Hallen et al., 2007; Keszthelyi et al., 2007; Klix et al., 2010; Bidard et al., 2011). For example, comparing transcription profiles during sexual development, or between a fertile fungal strain and its transgenic strain lacking a MAT gene (e.g. in P. anserina, F. verticillioides, and S. macrospora), provided several sets of genes differentially regulated in the mutant strains. However, the genes directly regulated by individual MAT genes remain undetermined.
The developmental up-regulation pattern and transcript abundance in two sets of MAT genes (a set of MAT1-1-1, MAT1-2-1, and MAT1-2-3, and the other of MAT1-1-2 and MAT1-1-3) provide new insight into functional role(s) of individual MAT genes for sexual development in F. graminearum, which are also supported by the phenotypic changes in the gene deletion strains. The former set of MAT genes can be considered key regulators of sexual development, particularly required for the early sexual stage for the following reasons. First, the gene expressions peaked at 2 dai, and the transcripts were more (at least 65-fold higher) abundant than those of the latter set of MAT genes at 2 dai. Secondly, the absence of perithecium-like structures in ΔMAT1-2-1 strain or the presence of barren perithecia in the ΔMAT1-1-1 strain, which were even smaller than those in the ΔMAT1-1-2 and ΔMAT1-2-3 strains, on carrot agar could be attributable to blockage of early events such as internuclear recognition, formation of ascogenous hyphae, and nuclear fusion. In addition, the gene functions are quite well conserved among many ascomycetous fungi, with the exception of the homothallic S. macrospora, in which SmtA-1 (comparable to MAT1-1-1) was dispensable for perithecia formation (Klix et al., 2010). In contrast, the latter set of MAT genes may be involved in the late stages of sexual development. Even though they were also confirmed as essential regulators of sexual development, the ΔMAT1-1-2 and ΔMAT1-1-3 strains retained the capacity to produce barren perithecia, indicating that their sexual development was blocked at the stages required for perithecia maturation. However, the function of these genes in sexual reproduction was not conserved among the fungal species examined. MAT1-1-2 was essential for the formation of sexual fruiting bodies in heterothallic P. anserina and homothallic S. macrospora (Klix et al., 2010), as well as in F. graminearum, but it seemed to have a redundant function along with MAT1-1-3 in the heterothallic N. crassa. MAT1-1-3, which was essential for sexual development in F. graminearum, was confirmed as a non-essential regulator in S. macrospora (Klix et al., 2010).
The function of a newly proposed MAT gene (MAT1-2-3) at the MAT1-2 locus was confirmed as non-essential for sexual development in F. graminearum. However, the expression pattern of MAT1-2-3 was similar to those of MAT1-1-1 and MAT1-2-1 in both F. graminearum and F. asiaticum, suggesting that it is also responsible for the defects in self-fertility in F. asiaticum, although it may have redundant functions. Sexual stage-specific MAT1-2-3 expression indicates that it is an additional MAT transcript at the MAT1-2 locus, although its regulatory capacity is unclear, since it contains no known DNA-binding motif (Martin et al., 2011).
Outcrosses of a ΔMAT strain to a self-fertile strain demonstrated that a nucleus carrying both MAT1-1 and MAT1-2 loci prefers a nucleus lacking at least one MAT gene, as well as a nucleus lacking all the genes at the MAT1-1 locus (Lee et al., 2003) for nuclear fusion when the two types of nuclei are present in ascogenous hyphae formed in the outcrosses. Thus, individual MAT genes except MAT1-2-3 at both MAT loci play a role in the nuclear choice mechanism during sexual development. However, whether this is mediated by pheromone pathways as in heterothallic species is uncertain, since the pheromone system is dispensable in the homothallic F. graminearum (Kim et al., 2008; Lee et al., 2008).
In conclusion, variations in the expression pattern and level of the two sets of MAT transcripts, which play a role in the early and late stages of sexual development, respectively, represent a possible cause of the variation in self-fertility in the Fg complex strains. However, the upstream regulatory mechanisms or signaling pathways that determine the differences in the expression of these MAT genes in F. graminearum and F. asiaticum remain unknown. Therefore, confirmation of this hypothesis requires further research, for example expression of these genes in F. asiaticum to mimic that in F. graminearum by swapping the promoters. Additionally, the different functional requirement of MAT1-1-1 and MAT1-1-3 for sexual development between homothallic F. graminearum and S. macrospora implies that some of the regulatory networks controlled by MAT proteins may not be conserved among filamentous ascomycetes.
This research was supported by a grant from the Next-Generation BioGreen 21 Program (No. PJ008210), Rural Development Administration, Republic of Korea, and by the Agricultural Research Center program of the Ministry for Food, Agriculture, Forestry and Fisheries, Republic of Korea.