(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Maintenance of divergent lineages of the Rice Blast Fungus Pyricularia oryzae through niche separation, loss of sex and post-mating genetic incompatibilities [1]

['Maud Thierry', 'Phim Plant Health Institute', 'Univ Montpellier', 'Inrae', 'Cirad', 'Institut Agro', 'Ird', 'Montpellier', 'Umr Phim', 'Anses Plant Health Laboratory']

Date: 2022-09 

Many species of fungal plant pathogens coexist as multiple lineages on the same host, but the factors underlying the origin and maintenance of population structure remain largely unknown. The rice blast fungus Pyricularia oryzae is a widespread model plant pathogen displaying population subdivision. However, most studies of natural variation in P. oryzae have been limited in genomic or geographic resolution, and host adaptation is the only factor that has been investigated extensively as a contributor to population subdivision. In an effort to complement previous studies, we analyzed genetic and phenotypic diversity in isolates of the rice blast fungus covering a broad geographical range. Using single-nucleotide polymorphism genotyping data for 886 isolates sampled from 152 sites in 51 countries, we showed that population subdivision of P. oryzae in one recombining and three clonal lineages with broad distributions persisted with deeper sampling. We also extended previous findings by showing further population subdivision of the recombining lineage into one international and three Asian clusters, and by providing evidence that the three clonal lineages of P. oryzae were found in areas with different prevailing environmental conditions, indicating niche separation. Pathogenicity tests and bioinformatic analyses using an extended set of isolates and rice varieties indicated that partial specialization to rice subgroups contributed to niche separation between lineages, and differences in repertoires of putative virulence effectors were consistent with differences in host range. Experimental crosses revealed that female sterility and early post-mating genetic incompatibilities acted as strong additional barriers to gene flow between clonal lineages. Our results demonstrate that the spread of a fungal pathogen across heterogeneous habitats and divergent populations of a crop species can lead to niche separation and reproductive isolation between distinct, widely distributed, lineages.

The spread of infectious plant diseases associated with introduced fungal pathogens can have devastating effects on agrosystems, ecosystems, and human livelihoods. In addition to the burden of emerging fungal infections, most major crops or domesticated trees are colonized by introduced pathogens that are widespread and have been present for extended periods of time. An understanding of the factors underlying the colonization success of fungal pathogens is essential to limit the risk of future pandemics. Here, we use genotyping, genome sequencing and phenotypic assays of a large collection of isolates of the rice blast fungus Pyricularia oryzae to investigate the factors underlying population differentiation and differences in life history strategies in this major plant pathogen. We found four lineages of P. oryzae that displayed significant differences in host range and repertoires of virulence genes, and thrived in areas with different types of environmental conditions. We also found that sexual compatibility between lineages is limited by the lack of fertile female structures and early, post-mating, reproductive barriers. Our study shows that niche separation, loss of sex and post-mating genetic incompatibilities can contribute to the maintenance of population structure in widely distributed fungal crop pathogens.

Funding: This work was funded by the Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement (INRAE) (to TK), the Agence Nationale de la Recherche (ANR) (Grant ANR-18-CE20-0016 to PG), the Centre de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD) (to DT), the Agence Nationale de Sécurité Sanitaire de l’Alimentation (Anses) (to RI), the CGIAR Research Program on Rice (to DT). This work was also partially funded by Bayer Crop Science Singapore (to DT), whose role in the study was limited to the collection and selection of 211 isolates, and genotyping of all 886 isolates. Bayer Crop Science Singapore had no additional role in the design, collection and analysis of data, decision to publish, or preparation of the manuscript. Other funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

To further explore the factors underlying the origin and maintenance of population structure in rice-infecting P. oryzae, we analyzed genomic and phenotypic variation in isolates sampled across all rice-growing areas of the world. Our aims were to re-evaluate population subdivision and the reproductive mode using a large collection of samples, to further explore host specialization using a larger number of isolates and rice varieties, to examine differences in the geographic and climatic distribution of lineages, and to measure their interfertility. For this, we analyzed genetic variation using genotyping data with a high resolution in terms of genomic markers and geographic coverage, and measured phenotypic variation for a representative subset of the sample set. We also characterized the genetic variability and content of repertoires of effector proteins using whole-genome resequencing data for a representative set of isolates.

The lineage of P. oryzae that infects rice is the most widely studied at the population level, due to its agricultural importance. Analyses of populations from commercial fields and trap nurseries based on molecular markers and mating assays have consistently shown that non-Asian populations of the rice-infecting lineage harbor a single mating-type and are clonal, with sexual reproduction never observed in the field [ 42 , 43 ]. Populations with female-fertile strains, a coexistence of both mating types and signatures of recombination have been observed exclusively in Asia [ 42 – 46 ]. More recent population genomic studies of the rice-infecting lineage have revealed genetic subdivision into four main lineages, estimated to have diverged approximately 1000 years ago and displaying contrasting modes of reproduction [ 47 – 49 ]. Only one of the four lineages that infects rice (lineage 1, prevailing in East and Southeast Asia) displays a genome-wide signal of recombination, a balanced ratio of mating type alleles and a high frequency of fertile females, consistent with sexual reproduction [ 44 ]. All the other lineages (lineages 2, 3 and 4, present on several continents) have clonal population structures, highly biased frequencies of mating types and rare fertile females, suggesting that reproduction in these lineages is strictly asexual [ 44 , 49 ]. The recombining lineage has a nucleotide diversity that is four times greater than that of the clonal lineages [ 49 ]. In theory, gene flow remains possible between the majority of the lineages, because there are rare fertile strains and some lineages have fixed opposite mating types. This raises questions about the nature of the factors underlying the emergence of different rice-infecting lineages in P. oryzae, and contributing to their maintenance in the face of potential gene flow. Previous studies have helped to cement the status of P. oryzae as a model for studies of the population biology of fungal pathogens. However, most efforts to understand the population structure of the pathogen have been unable to provide a large-scale overview of the distribution of rice-infecting lineages and of the underlying phenotypic differences, because the number of genetic markers was limited [ 43 , 44 ], the number of isolates was relatively small [ 47 , 49 ] and because host-pathogen interaction phenotypes were the only phenotypes that were scored [ 43 , 44 , 47 ]. The contribution of differences in the geographic and climatic distribution of rice blast lineages and genetic incompatibilities between them, in particular, remains unknown.

Pyricularia oryzae (Ascomycota; syn. Magnaporthe oryzae) is a widespread model plant pathogen displaying population subdivision. Pyricularia oryzae is best known as the pathogen causing rice blast, which remains a major disease of Asian rice (Oryza sativa), but it is also a major constraint on the production of wheat (Triticum), finger millet (Eleusine coracana) and Italian millet (Setaria italica), a significant disease of ryegrass (Lolium) and an emerging pathogen of maize (Zea maydis) [ 33 – 38 ]. An in-depth genomic analysis of a large worldwide collection of P. oryzae isolates from 12 host plants revealed the existence of 10 distinct P. oryzae lineages, each associated mostly with a single main cereal crop or grass host [ 39 ]. Sequence divergence between lineages was low, of the order of about 1% [ 39 ], and analyses of gene flow and admixture provided evidence that the different host-adapted lineages were connected by relatively recent genetic exchanges, and therefore corresponded to a single phylogenomic species [ 40 , 41 ]. The reproductive biology of P. oryzae is typical of plant pathogenic Ascomycetes. Pyricularia oryzae has a haplontic life cycle. Thus, the multicellular state is haploid, and the breeding system is heterothallic, with mating occurring only between haploid individuals of opposite mating types. Pyricularia oryzae is hermaphroditic [ 42 ], but successful mating also requires that at least one of the partners to be capable of producing female structures (i.e. “female-fertile”).

Introduced fungal plant pathogens display a range of population structures, reflecting their highly diverse life-history strategies and invasion histories [ 8 , 11 , 12 ]. Demographic events (e.g. population bottlenecks, secondary contact) and natural selection during spread across heterogeneous environments may be associated with intricate changes in population dynamics and life history traits, or may cause such changes. Shifting to a new host is a primary life history trait change in introduced fungal pathogens, and a major route for the emergence of disease. Changes in host range are facilitated by certain features, such as mating within or on their hosts and strong selective pressure on a limited number of genes [ 13 , 14 ]. Reproduction within, or on plant hosts favors assortative mating with respect to host use, leading to a strong association between adaptation to the host and reproductive isolation [ 15 , 16 ] and the differentiation of pathogen populations adapted to different host populations, varieties or species [ 17 – 20 ]. Such changes in pathogen host range can result from immune-escape mutations. Typically, they occur in a small number of genes encoding small secreted proteins called effectors, which enable microbes to influence the outcome of host-pathogen interactions for their own benefit, and which may be subject to surveillance by the plant immune system [ 21 – 23 ]. Another frequent change in the life history traits of introduced fungal pathogens is the loss of sexual reproduction [ 24 , 25 ]. Introduced fungal pathogens may develop a clonal population structure over their entire introduced range or over parts of that range ([ 8 ] and references therein). The immediate causes of the loss of sexual reproduction include a lack of compatible mating partners in the introduced range (i.e. a lack of compatible mating types) [ 11 ], hybridization and hybrid sterility [ 26 , 27 ], and a lack of compatible alternate hosts for sexual reproduction [ 28 ]. Mating type loss may be driven by the extreme population bottlenecks associated with the colonization of a new host or new area. Shifts to asexual reproduction may also result from selection against reproduction [ 8 ], particularly for pathogens of domesticated crops and trees, due to the availability of large, homogeneous host populations, releasing constraints that maintain sexual reproduction in the short term [ 29 ]. These conditions may favor the asexual spread of new lineages carrying adaptive allelic combinations requiring shelter from recombination [ 30 ]. Finally, changes in life history traits can also result from variations in environmental conditions, including temperature in particular, during the course of pathogen spread [ 31 , 32 ]. Deciphering the complex interplay between changes in population structure and life history strategies during the spread of plant pathogenic fungi requires deep sampling and the integration of genetic, phenotypic, and environmental information throughout the native and introduced ranges.

Introduced fungal plant pathogens constitute a major threat to ecosystem health and agricultural production [ 1 – 3 ]. Research in plant pathology has revealed a seemingly inexhaustible stream of disease outbreaks or significant changes in the geographic location of diseases and host ranges of pathogens [ 4 , 5 ]. In addition to the burden of emerging fungal infections, most major crops or domesticated trees are colonized by introduced pathogens that are so widespread and have been present for so long, that they do not necessarily come to mind when one thinks about invasive fungi [ 6 , 7 ]. If fungal pathogens are to spread successfully over their host species distribution, they must overcome a series of barriers to invasion related to dispersal, host availability, competition with other microbes, and abiotic conditions [ 8 ]. An understanding of the factors underlying the colonization success of widespread fungal pathogens is essential to protect world agriculture against future pandemics [ 9 , 10 ].

Results

Reproductive barriers caused by female sterility and postmating genetic incompatibilities To elucidate how the clonal lineages have emerged from the more ancient recombining populations in lineage 1, we determined the capacity of the different lineages to engage in sexual reproduction. Using in vitro crosses with Mat1.1 and Mat1.2 tester strains, we analyzed the distribution of mating types and determined the ability to produce female structures (i.e. female fertility). We found that lineages 2, 3 and 4 were composed almost exclusively of a single mating type: 97% of lineage 2 and 4 isolates tested carried the Mat1.1 allele, while 97% of lineage 3 isolates carried the Mat1.2 allele. Only a small proportion (0–7%) of isolates from these lineages showed female fertility (Table 1). The mating type ratio was more balanced in lineage 1 (52% of Mat1.1; Table 1), with Mat1.1/Mat1.2 ratios ranging from 40/60 in the Yule cluster to 69/31 in the Baoshan cluster. Female fertility rates were also high in most of the clusters in lineage 1 (Yule: 79%; Baoshan: 67%; Laos: 37%; Table 1). Only in the International cluster female fertility was as low as those in the clonal lineages, with 4% fertile females (Table 1). These observations suggest that despite the generally low rate of female fertility, sexual reproduction between most lineages is possible, except between lineages 2 and 4, that have the same highly biased mating type ratio. We further assessed the likelihood of sexual reproduction within and between lineages, by evaluating the formation of sexual structures (perithecia), the production of asci (i.e. meiotic octads within the perithecia) and the germination of ascospores (i.e. meiospores in meiotic octads) in crosses between randomly selected isolates of opposite mating types from all four lineages (Fig 3). This experiment revealed a marked heterogeneity in the rate of perithecium formation across lineages. Clusters in lineage 1 had the highest rates of perithecium formation, with isolates in the Yule cluster, in particular, forming perithecia in 93% of intra-lineage crosses, and in more than 46% of inter-lineage crosses (Fig 3A and S9 Data). Due to their highly biased mating type ratios, isolates from lineages 2, 3 and 4 could only be crossed with isolates from other lineages. The proportion of these inter-lineage crosses producing perithecia was highly variable and ranged, depending on the lineages involved, from 0% to 83% (Fig 3A and S9 Data). In inter-lineage crosses involving the International cluster of lineage 1, the rate of perithecium formation was similar as in inter-lineage crosses involving clonal lineages 2, 3 and 4. None of the intra-lineage crosses attempted between isolates from the International cluster led to perithecium formation (Fig 3A and S9 Data). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Success of crosses between lineages 2–4 and clusters within lineage 1 with (A) proportion of crosses producing at least one perithecium, (B) scoring of ascus formation and ascospore germination for a subset of crosses. https://doi.org/10.1371/journal.ppat.1010687.g003 Perithecium dissection for a subset of crosses involving some of the most fertile isolates revealed that most inter-lineage crosses produced perithecia that did not contain asci or contained asci with low ascospore germination rates (Fig 3B and S10 Data). While 100% of intra-lineage crosses between isolates of the Yule cluster produced numerous germinating ascospores, only 33%, 56% and 7% of Yule x lineage 2, Yule x lineage 3 and Yule x lineage 4 crosses produced germinating ascospores, respectively. Together, our crossing experiments indicate that the three clonal lineages and the International cluster of lineage 1 are isolated from each other by strong pre- and early-postmating barriers, including breeding system isolation (differences in mating type and female sterility), and a combination of gametic incompatibility, zygotic mortality or hybrid non-viability. However, three of the clusters in lineage 1 had biological features consistent with an ability to reproduce sexually, suggesting possible sexual compatibility with clonal lineages 2, 3 and 4, and the International cluster.

Specialization to temperature conditions and rice subgroups Given the very broad distribution of P. oryzae and the strong environmental heterogeneity it encounters in terms of the nature of its hosts and the climate in which it thrives, we evaluated the variation of fitness over different types of rice host genotypes and temperature conditions. We measured growth rate and sporulation of 41 representative isolates cultured at different temperatures to test the hypothesis of adaptation to temperature (lineage 1 [yule cluster]: 11; lineage 2: 10; lineage3: 10; lineage 4: 10). For all lineages, mycelial growth rate increased with incubation temperature. This trend was more visible from 10°C to 15°C (increased mean mycelial growth of +2.22 mm/day) than from 25°C to 30°C (+0.05 mm/day) (growth curves and a full statistical treatment of the data are presented in S6 Text; data are reported in S11 Data). Fitting a linear mixed-effects model with incubation time, experimental replicate, and lineage of origin as fixed effects and isolate as a random effect, revealed a significant lineage effect at each incubation temperature [10°C: F(1, 364) = 7988, p-value<0.001; 15°C: F(1, 419) = 33161, p-value<0.001; 20°C: F(1, 542) = 40335, p-value<0.001; 25°C: F(1, 413) = 30156, p-value<0.001; 30°C: F(1, 870) = 52681, p-value<0.001]. Comparing least-squares means resulting from linear mixed-effects models at each temperature revealed a significantly lower growth rate of lineage 4 at 10°C relative to other lineages and a significantly higher growth rate of lineage 1 at 15°C, 20°C, 25°C and 30°C relative to other lineages. Almost no sporulation was detected at 10°C after 20 days of culture, with the few spores observed being not completely formed and divided by only one septum instead of two septa in mature conidia (sporulation curves and a full statistical treatment of the data are presented in S7 Text; data are reported in S12 Data). For all lineages, sporulation (weighted by mycelium colony size) increased with temperature from 15 to 25°C and dropped at 30°C. Significant lineage effects were observed at 10°C, 15°C and 30°C (Kruskal-Wallis tests; 10°C: H(3) = 9.27, p = 0.026; 15°C: H(3) = 17.5, p-value<0.001; 30°C: H(3) = 9.60, p = 0.022). Pairwise non-parametric comparisons of lineages based on Dunn’s test revealed significant differences between lineages 1 and 2 at 10°C, between lineage 1 and lineages 3 and 4 at 15°C, and between lineage 1 and 4 at 30°C. Together, measurements of mycelial growth and sporulation at different temperatures revealed differences in performance between lineages, but no clear pattern of temperature specialization. We assessed the importance of adaptation to the host, by challenging 45 rice varieties representative of the five main genetic groups of Asian rice (i.e. the temperate japonica, tropical japonica, aus, aromatic and indica subgroups of Oryza sativa) with 70 isolates representative of the four lineages of P. oryzae and the four clusters within lineage 1 (S2 Table and S13 Data). Interaction phenotypes were assessed qualitatively by scoring resistance (from full to partial resistance: scores 1 to 3) or disease symptoms (from weak to full susceptibility: scores 4 to 6), which is standard in rice blast phenotyping and documented by images [54–57]. Resistance/susceptibility scores were analyzed by fitting a proportional-odds model and performing an analysis of variance that revealed significant differences between groups of isolates (χ2(6) = 100, p-value<0.001), between rice genetic groups (χ2(4) = 161, p-value<0.001), and a significant interaction between these two variables (χ2(24) = 97, p-value<0.001) (S8 Text). The finding of a significant interaction between groups of isolates (lineages or clusters) and rice types indicates that the effect of the group of isolates on the proportion of compatible interactions differed between rice types, suggesting adaptation to the host. This is also supported by the specific behaviour of certain lineages. Isolates from lineage 2 had much lower symptom scores than the other lineages on all rice types except temperate japonica, and the isolates of the Yule cluster were particularly virulent on indica varieties (Fig 4A and S2 Table). In comparisons of rice genetic groups, significantly higher symptom scores were observed on temperate japonica rice than on the other types of rice, whereas the varieties of the aromatic genetic group were significantly more resistant to rice blast (Fig 4B). Together, these experiments therefore revealed significant differences in host range between lineages. However, this specialization to the host was not strict, because there was an overlap between host ranges (Fig 4C–4G). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Compatibility between 70 P. oryzae isolates and 45 rice varieties, representing five types of rice. Compatibility was measured as symptom scores [54,55] estimated using pathogenicity tests in controlled conditions. A: Symptom scores as a function of the lineage of origin of isolates, or cluster of origin for isolates from lineage 1; B: Symptom scores as a function of the type of rice; C-G: Symptom scores as a function of the lineage of origin of isolates, for each type of rice. Abbreviations: Y, Yule; I, International; B, Baoshan; L, Laos; 2, lineage 2; 3, lineage 3; 4, lineage 4; Ind, indica; TeJ, temperate japonica; TrJ, tropical japonica; Aro, aromatic. Shared small capitals indicate non-significant differences (pairwise comparisons after computing least-square means, with Tukey adjustment). All interactions between rice varieties and P. oryzae isolates were assessed in three independent experiments, and the median of the three symptom scores was used in calculations. Boxen plots were drawn using function boxenplot() with Python package seaborn 0.11.1. Starting with the median as centerline, each successive level (i.e., each successive box) outward contains half of the remaining data. Number of boxes was controlled using option k_depth = ‘trustworthy’. Black horizontal lines represent the median. https://doi.org/10.1371/journal.ppat.1010687.g004

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010687

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/