Discussion
Since P. graminis was described by Nitschke 1870 (Fuckel 1870), over 300 species have been recorded on graminaceous hosts, and many more on non-grass hosts. However, Parbery recognized that there are fewer species associated with grasses and established that there were 95 valid graminicolous Phyllachora species world-wide based on morphological characteristics (Parbery 1967). In the most complete study of Phyllachora species in North America, Orton (1944) identified 45 morphological species from more than 100 host species (Orton 1944). While this likely represents a significant overestimation of the true number of species in North and Central America, it does demonstrate the vast number of hosts on which Phyllachora species have been reported. Our results based on both herbarium and contemporary samples of infected hosts indicate that there are far fewer species ofPhyllachora in the Americas than indicated by Orton (1944) and Parbery (1967), and the species that are present have a greater host range than previously thought. The predominant species in this study,P. sp. 3 , has a broad geographic and host range with the capacity to infect maize throughout South, Central and North America as well as seven grass species and two dicot species. This phylogenetic species also includes isolates of 11 morphologically determined species ofPhyllachora (P. chaetochloae, P. diplocarpa, P. epicampsis, P. euphorbiaceae, P. graminis, P. heraclei, P. junci, P. maydis, P. rottboelliae, P. sylvatica, and P. vulgata ) from herbarium samples collected in the Dominican Republic, Germany, India, Mexico, the Philippines, Trinidad and Tobago and the U.S., indicating global distribution of this species. This expanded host range also now complicates the taxonomic status and the name to be retained by this genetic group. An isolate of P. graminis collected fromAgropyron repens from Germany was designated as the lectotype specimen for the genus (Clements and Shear 1931), and the isolate ofP. graminis examined in this study was also collected fromA. repens in Germany, indicating that P. graminis may have precedence for the species name of P. sp. 3 . This would have ramifications for P. maydis as well as several otherPhyllachora species in P. sp. 3 (Table 1; Figure 3) that appear to be synonyms of P. graminis . This is based solely on sequence data from the ITS and/or LSU region and further multi-gene phylogenetic studies of a larger representation of type material from herbaria and contemporary Phyllachora samples from additional hosts is needed for a thorough taxonomic assessment of this genus.
The three maize-infecting species, P. sp. 1, P. sp. 2 andP. sp. 3 , have overlapping geographic and host ranges, providing the opportunity for co-infection and genetic exchange. Co-infection on the same leaf tissue by P. sp. 3 and P. sp. 1 was observed on four occasions with herbarium samples (BPI893232_1 and BPI893232_2 , BPI893231_1 and BPI893231_2, BPI893226_1 and BPI893226_2, BPI893230_1 and BPI893230_2) from 3 counties in Indiana. A recent fungal community analysis of tar spot lesions on maize found a similar trend with two distinct OTUs occurring on 21 of 22 maize leaf samples from Michigan (McCoy et al. 2019). A similar phenomenon has also been observed in Albugo candida , another biotrophic pathogen with a broad host range (McMullan et al. 2015). Races of A. candida were not able to infect a host on their own but were able to co-infect with a race-specific isolate that suppressed host immunity in that host. The offspring of any genetic introgression or recombination resulted in a race with an expanded host range able to infect both plants infected by the parental strains of A. candida . A whole genome comparison of these A. candida races found a mosaic-like genome structure with large portions conserved between races, as well as regions with only 89% sequence similarity. This scenario may explain the wide host range and variation in morphology between hosts in Phyllachora species. Sexual reproduction in P. maydis followed by discharge of infective ascospores commonly occurs on corn leaves annually in maize producing regions of the U.S. (Kleczewski et al. 2019, Groves et al. 2020b). The presence of multiple maize infecting species in the midwestern US, and even on a single infected leaf, combined with frequent sexual recombination, ascospore release and infection, could result in novel populations and/or species of Phyllachora that are more virulent on maize or that have an expanded host range. This may also explain why P. sp. 3 has such a broad host range whereasP. sp. 1 and P. sp. 2 were only found on maize. Individual populations may gain the ability to infect a new host but are still able to sexually recombine with the rest of the population on the original host species. Given the geographic overlap of many grass species in Central, South and North America, small populations of P. sp. 3may have adapted to infect a novel grass species, while maintaining the ability to recombine with the larger P. sp. 3 complex, resulting in the expansion of the host range without specialization and speciation.
Speciation has likely occurred in instances where geographic isolation of a new host prevented further introgression with the original population. As maize is commonly grown from Argentina to Canada, it represents a common host for which distinct Phyllachorapopulations may infect and recombine resulting in potentially new and more virulent populations that are still part of the same species. It is unclear if geographic or genetic barriers lead to speciation between the closely related P. sp. 1, P. sp. 2 and P. sp. 3 , but the significant overlap in host and geography would indicate a genetic barrier. While P. sp. 1 and P. sp. 2 were only recovered from maize, our sampling scheme was strongly biased towards maize. It is possible that P. sp. 1 and P. sp. 2 are present on other grass and non-grass hosts in Central and North America and were not sampled in this study. These non-sampled hosts, if only infected by one of the Phyllachora species, may represent the isolation that led to adaptation and speciation.
For now the name Phyllachora maydis will be retained by P. sp. 3 as the P. maydis type material (BPI638553) clustered with this group. However, the presence of three maize-infecting species, the lack of type material of P. graminis , and the potential taxonomic synonymy with P. graminis and several other Phyllachoraspecies makes it difficult to determine which of the maize infecting species will retain the name P. maydis . Therefore, we recommend referring to P. sp. 1, P. sp. 2 and P. sp. 3 as thePhyllachora maydis species complex until further morphological and multi-gene phylogenetic studies can properly delineate these species.
In this work, we conducted the most comprehensive assessment ofPhyllachora maydis reported to date and provided evidence that our understanding of this species and genera is limited and requires significant attention. The reasons for the emergence of tar spot, caused by three different species of Phyllachora that have been present in Central America, Mexico and the Caribbean for over 75 years is still unclear. Several scenarios may explain the recent emergence and severity of tar spot caused by Phyllachora spp . in the upper Midwest of the U.S. While P. sp. 2 and P. sp. 3 have been present in both Mexico and Puerto Rico for the last century, it is possible that when the fungus was able to be dispersed via wind and rain to the U.S. it could not overwinter in colder climates and the disease could not get established. In fact, according to the herbarium specimens, P. sp. 3 has been present in the U.S. since the 1940s in California and Arizona on native grasses but not maize. However, recent studies have demonstrated that Phyllachora spp. can overwinter in Illinois (Kleczewski et al. 2019). Shorter and warmer winters due to climate change could be playing a role in the ability ofPhyllachora spp. to survive further north in the U.S. Changes in climate patterns during the growing season may also have an impact on this disease as increased temperature and precipitation may promote epidemics of this disease. Finally, a change in maize genetics may also play a role in the increased severity of tar spot. Since maize breeding programs were not selecting for resistance to tar spot, any partial resistance that may have been present in U.S. germplasm may have been lost through genetic drift. The loss of this resistance may not have been noticed until Phyllachora spp. arrived in the primary maize growing region of the U.S. The disease remains of minor importance in Mexico and Central American maize production, as resistance to this disease would be selected for in breeding programs. The most likely scenario for the emergence of tar spot in the U.S includes a combination of these factors: 1) introduction of multiple species ofPhyllachora from Mexico, Puerto Rico or other Central American countries through movement of infected plant tissue or possible long-distance movement via wind, rain, hurricane/tropical storm system, etc., 2) change in climate in the Midwestern maize growing region more hospitable to the growth, reproduction and survival of Phyllachora spp. ; and 3) lack of resistance in maize germplasm grown in the Midwestern U.S.