Teratosphaeria stem canker of Eucalyptus: two pathogens, one devastating disease

Summary 
Background
 Teratosphaeria gauchensis and T. zuluensis are closely related fungi that cause Teratosphaeria (previously Coniothyrium) stem canker disease on Eucalyptus species propagated in plantations for commercial purposes. This disease is present in many countries in which Eucalyptus trees are planted, and continues to spread with the international trade of infected plant germplasm. Taxonomy Fungi, Ascomycota, Pezizomycotina, Dothideomycetes, Dothideomycetidae, Capnodiales, Teratosphaeriaceae, Teratosphaeria. Identification The causal agents form dark masses of pycnidia that are visible on the surface of distinct stem cankers that typically form on young green stem tissues. Accurate diagnosis of the causal agents requires DNA sequence data. Host range Nine species of Eucalyptus are known to be affected. Of these, E. grandis and its hybrids, which include some of the most important planting stock globally, appear to be particularly vulnerable. Disease symptoms Small necrotic lesions develop on young green stem tissue. These lesions coalesce to form large cankers that exude gum. Epicormic shoots develop below the girdling canker and, in severe cases, trees die. Useful websites Mycobank, https://www.mycobank.org; Publications of the Forestry and Agricultural Biotechnology Institute (FABI), https://www.fabinet.up.ac.za/index.php/journals.


I N T RO D U C T I O N
Teratosphaeria stem canker is an important disease of Eucalyptus species planted outside their native range for the production of wood and wood products in countries with sub tropical and tropical climates (Gezahgne et al., 2005;Wingfield et al., 1996). The disease is caused by two closely related species of Dothideomycete fungi in the genus Teratosphaeria (Order: Capnodiales; Family: Teratosphaeriaceae) (Crous et al., 2009a). Intriguingly, T. gauchensis and T. zuluensis cause Teratosphaeria stem canker with identical symptoms, independently in different parts of the world (Cortinas et al., 2006c).
Only the asexual states of T. gauchensis and T. zuluensis are known. Consequently, few morphological characters of taxo nomic relevance are available to discriminate between the patho gens. Accurate identification and disease diagnosis relies entirely on DNA sequence comparisons (Roux et al., 2002). Microsatellite genotyping has also strongly influenced our understanding of the global prevalence of these two species. This review seeks to illustrate that these molecular genetic techniques have been indispensable in the study of these fungi and, further, that ge nomic tools will define the future understanding of their biology and management.
This pathogen profile reviews the available knowledge re garding Teratosphaeria canker and its causal agents, gathered subsequent to the first description of this disease more than two decades ago (Wingfield et al., 1996). The apparent and worrying ease with which Teratosphaeria canker has spread to plantations across the globe is discussed, and the growing threat of this dis ease to plantation forestry is also considered.

A G LO BA L LY I M P O R TA N T H OST
Eucalyptus is a large genus (>740 species) in the Myrtaceae, a family comprised predominantly of flowering trees and some shrubs (Rejmánek and Richardson, 2011). The genus is native to Australia and some surrounding islands in the Pacific, but many of its members thrive outside of their natural environment as a result of rapid growth and drought tolerance. Consequently, Eucalyptus species have been transported across the globe for use as shade, ornamental trees and wind barriers (Santos, 1998). Together with Pinus and some other tree genera, Eucalyptus trees have also become one of the main components in both Northern and Southern Hemisphere forest plantation operations (Carle and Holmgren, 2008;Del Lungo et al., 2006).
Eucalyptus trees are valuable not only for their timber, but also for nontimber commodities, such as essential oils and honey (Rejmánek and Richardson, 2011). Globally, Eucalyptus planta tions established for the production of wood products have in creased in magnitude exponentially over the past century, and comprised an estimated 20 million hectares in 2008 (Rejmánek and Richardson, 2011). This equates to almost onequarter of the 92 million hectares of Eucalyptus in natural Australian forests (Australia's State of the Forests Report, 2013). The incredible suc cess with which these trees have been grown outside their native range has largely been attributed to an initial escape from nat ural enemies (Wingfield, 2003). This diseasefree 'honeymoon' period has, however, been relatively shortlived, and disease and pest problems have become an important constraint to the sustainability of Eucalyptus plantations globally (Burgess and Wingfield, 2017).

D I S E A S E I M PAC T
An unknown stem canker of Eucalyptus grandis was first re ported in 1988 in a plantation in the vicinity of KwaMbonambi in KwaZulu Natal, South Africa (Wingfield et al., 1996). The pathogen was described as Coniothyrium zuluense (now known as Teratosphaeria zuluensis), and the disease was referred to as Coniothyrium (now Teratosphaeria) canker (Wingfield et al., 1996). This stem canker disease damaged large areas of Eucalyptus in the subtropical forestry regions of South Africa, necessitating the removal of numerous E. grandis stands (Cortinas et al., 2010).
Teratosphaeria canker was known only from South Africa until the early 2000s (Van Zyl et al., 2002b). Surveys of Australian plantations and natural forests have also failed to detect similar symptoms or pathogens, although closely related species causing leaf spots have been identified (Andjic et al., 2010;Crous et al., 2006). For this reason, it was initially hy pothesized that the disease originated in South Africa, possibly as a result of a host shift (Slippers et al., 2005) from a closely related plant, probably one in the Myrtaceae (Wingfield et al., 1996). This would have been similar to the situation that has occurred with various other Eucalyptus canker pathogens in the Cryphonectriaceae (Nakabonge et al., 2006;Van der Merwe et al., 2010). However, Teratosphaeria canker has subsequently emerged in many other subtropical and tropical areas of the world in which Eucalyptus is planted for commercial forestry (Fig. 1). It now affects plantations in several Asian (Cortinas et al., 2006a;Gezahgne et al., 2003a;Van Zyl et al., 2002b), African (Gezahgne et al., 2005;MuimbaKankolongo et al., 2009;Roux et al., 2005) and South American (Cortinas et al., 2006c;Gezahgne et al., 2003a) countries, as well as in Mexico, Hawaii and Portugal (Cortinas et al., 2004;Roux et al., 2002;Silva et al., 2015). T. gauchensis

T. zuluensis
Although trees die only in severe cases of the disease, the impact of Teratosphaeria canker on the forestry industry can be devastating. Stem cankers decrease the quality of timber to the extent that it cannot be used for construction (Gezahgne et al., 2003a). In addition, the wood is stained with red gum that ex udes from the cankers (Gezahgne et al., 2003b;Old et al., 2003), decreasing the value of other timber products. The cost and qual ity of pulping are also negatively affected because the cankers hinder the debarking process.

T H E PAT H O G E N E M E RG E N C E O F A C RY P T I C S PEC I E S
As the incidence of Teratosphaeria stem canker increased in plan tations outside South Africa, a large collection of isolates became available to study the genetics of T. zuluensis (Cortinas et al., 2006b). Rather than enabling a population genetics study, geno types of these isolates revealed two distinct taxa isolated from can kers on Eucalyptus that were indistinguishable based on symptoms (Cortinas et al., 2006c). The isolates accommodating individuals from southern Africa, Asia and Mexico represented T. zuluensis. In contrast, a second group emerged comprising isolates from South America, and these were provided with the name T. gauchensis (Fig. 1). Thus, two species, resulting in identical symptoms, oc curred on the same hosts in different parts of the world.
The lack of clear morphological features for the Teratosphaeria canker pathogens has contributed to a series of changes in their taxonomy (Fig. S1, see Supporting Information). The few taxo nomically useful morphological characteristics, such as conidial size and development, are inconspicuous and difficult to deter mine (Cortinas et al., 2006c;Crous et al., 2009a;Wingfield et al., 1996). The canker pathogen was identified as a Dothideomycete fungus and originally described in the genus Coniothyrium (Wingfield et al., 1996). DNA evidence later revealed a close re lationship to Mycosphaerella in the order Capnodiales, causing it to be transferred sequentially to three traditionally asexual genera associated with the sexual genus Mycosphaerella, namely Colletogleopsis, Kirramyces and Readeriella (Andjic et al., 2007;Cortinas et al., 2006a;Crous et al., 2007). Finally, to escape the debates surrounding the appropriate asexual genus and in antici pation of the 'one fungus one name' principle (Hawksworth et al., 2011;Taylor, 2011), these Eucalyptusinfecting fungi were trans ferred to the family Teratosphaeriaceae (Crous et al., 2007) and placed in the sexual genus Teratosphaeria (Crous et al., 2009a).

Disease symptoms
Teratosphaeria zuluensis and T. gauchensis have very different geographical distributions (Fig. 1). It is intriguing that two al lopatric pathogens cause indistinguishable disease symptoms independently. New infections typically occur during spring, after which Teratosphaeria canker is first visible as small (2-5 mm) ne crotic lesions on Eucalyptus stems (Wingfield et al., 1996) (Fig. 2), with young stem tissues particularly vulnerable (Old et al., 2003). These lesions become elliptical as they grow in size (Cortinas et al., 2006c), penetrate the vascular cambium (Old et al., 2003) and eventually merge with neighbouring lesions to form cankers filled with gum, also known as kino pockets (Wingfield et al., 1996;Fig. 2). Stem malformation typically ensues and the bark covering these cankers often cracks vertically, creating a 'cat eye' appearance and causing the gum to exude (Cortinas et al., 2006c). In the case of severe infections on susceptible clones, cankers girdle the stems, epicormic shoots develop and the tops of the trees die (Old et al., 2003;Wingfield et al., 1996).
A study by Van Zyl et al. (2002a) showed that the pathoge nicity of different isolates of T. zuluensis collected from the same area of disease occurrence varied considerably. The majority of isolates (>70%) did not give rise to noticeable lesions on 6month old E. grandis seedlings and appeared to be nonpathogenic. The remainder of the isolates produced lesions ranging between 15 and 61 mm, and the most virulent isolates (causing the longest lesions) also induced swelling and necrosis of stem tissue around the inoculation site.

Pathogen identification
Teratosphaeria zuluensis and T. gauchensis grow within host tissue, but are visible on the surface of cankers as dark pyc nidia, producing large numbers of dark, singlecelled conidia (Wingfield et al., 1996). During moist conditions, black conidial tendrils extend from the canker surfaces (Cortinas et al., 2006c), but sexual structures of T. zuluensis and T. gauchensis have never been observed in the field or in culture (Wingfield et al., 1996).
Morphological distinction between T. zuluensis and T. gauchensis is not possible with any reasonable level of con fidence. The two species have overlapping conidial lengths, although T. gauchensis conidia tend to be slightly longer and narrower (Silva et al., 2015). The conidia of T. gauchensis develop from polyphialidic conidiogenous cells, whereas T. zuluensis has percurrently proliferating monophialidic conidiogenous cells and the culture morphology is similarly variable for both species (Cortinas et al., 2006a(Cortinas et al., , 2006cVan Zyl et al., 2002b) (Fig. S2, see Supporting Information). Culture margins are grey or white and may be smooth or irregular. The centres are typically varying shades of olive or grey that darken with age. From below, cul tures range from dark green to rust/brown and may have lighter white or olivaceous bands at the margins.
Molecular data provide the only reliable means to iden tify the species of Teratosphaeria causing Eucalyptus cankers. Phylogenetic studies group the two species into closely related clades with 26 fixed polymorphisms between them in the internal transcribed spacer (ITS), βtubulin, elongation factor 1α (EF1α) and ATP6 genes (Cortinas et al., 2004;Gezahgne et al., 2005;Van Zyl et al., 2002b). A 20basepair fragment in the EF1α in tron is absent in T. gauchensis, but present in T. zuluensis, pro viding a useful marker for distinction between the species.
Intraspecific phylogenetic structure exists in both pathogens, but is especially common in T. zuluensis (Cortinas et al., 2006c;Jimu et al., 2014;Silva et al., 2015). Teratosphaeria zuluensis iso lates from some localities form wellsupported subclades, but isolates from localities with moderate to high genetic diversi ties, such as South Africa and China, intersperse with those from other locations (Chen et al., 2011;Cortinas et al., 2010). These subclades are not substantiated by biological information and only one study has reported a difference in the temperature range and pathogenicity of a wellsupported group of T. zuluen sis isolates from Thailand (Van Zyl et al., 2002b).
The lack of clear geographical boundaries and biological information between intraspecific clades in T. zuluensis and T. gauchensis precludes the description of additional spe cies. However, considering the intraspecific variability, it may be relevant to refer to these species as species complexes: a group of monophyletic isolates with significant genetic dif ferences that are not detected by standard barcoding or di agnostic procedures (Chen et al., 2016). Further distinction only becomes necessary if clinical, diagnostic or ecological relevance can be attached to different isolates in the complex (Chen et al., 2016), which is presently not the case for T. zulu ensis or T. gauchensis.

Ecological factors
The Teratosphaeria stem canker pathogens grow very slowly in culture, taking 6 weeks at 25 °C to reach a diameter of 40-50 mm (Cortinas et al., 2006c). This poor growth probably indicates a reliance on living host tissue, and therefore a bio trophic lifestyle (Wingfield et al., 1996). The two species dif fer in the extremes of their temperature range: T. gauchensis appears to tolerate lower temperatures (c. 10 °C) better than T. zuluensis, which, in turn, grows at 35 °C where T. gauche nsis growth shows a marked decrease (Silva et al., 2015). At their optimum range of 20-25 °C, T. zuluensis achieves a larger maximum diameter than T. gauchensis (Cortinas et al., 2006c).
The optimum temperature for growth of T. zuluensis and T. gauchensis might explain why canker outbreaks occur in sub tropical and tropical plantations. Countries such as South Africa and Ethiopia have Eucalyptus plantations in both subtropical and temperate regions, yet Teratosphaeria canker is damaging only in subtropical areas (Gezahgne et al., 2005;Wingfield et al., 1996). Recently, however, T. gauchensis has been identified from Teratosphaeria canker symptoms on E. globulus in Portugal, a country that has a moderate Mediterranean climate (Silva et al., 2015). These authors also noted that T. gauchensis isolates from the cooler northern part of the country were less tolerant to high temperatures (35 °C), implying that the pathogen may be adapt ing to the cooler environment.
Teratosphaeria gauchensis has been isolated from seven different Eucalyptus plantation species (excluding hybrids) and T. zuluensis from four (Table 1). Currently, their only shared hosts are E. camaldulensis and E. grandis (Jimu et al., 2015;Van Zyl et al., 2002b), two of the most widely planted Eucalyptus species in tropical and subtropical regions. Ironically, E. grandis and its hybrids appear to be highly sus ceptible to Teratosphaeria canker. The remainder of the known host species, especially E. cloeziana, E. paniculata and E. pro pinqua, are less common (FAO, 1981). Similarly, although E. globulus is widely planted, it is typically planted in tem perate climates in which the stem canker pathogens do not flourish (FAO, 1981). Therefore, it is likely that the different hosts of T. gauchensis and T. zuluensis reflect the distribution of Eucalyptus species in global plantations rather than true host preferences of these fungi.

R E L AT I O N S H I P TO E UC A LY P T U S L E A F PAT H O G E N S
Teratosphaeria zuluensis and T. gauchensis are the only known stem canker pathogens in a clade of species predominantly as sociated with Eucalyptus leaves (Crous et al., 2009a;Quaedvlieg et al., 2014) (Fig. 3). The majority of these leaf associates cause leaf spots of minor importance. However, some, such as T. de structans, T. cryptica, T. nubilosa and T. pseudoeucalypti, are responsible for major economic losses to the forestry industry (Burgess and Wingfield, 2017;Cândido et al., 2014;Hunter et al., 2011). Surprisingly, DNA evidence has shown that the two stem canker pathogens are each more closely related to Eucalyptus leafassociated species than they are to each other (Quaedvlieg et al., 2014). Based on the ITS and EF1α genes, T. gauchensis groups closest to T. majorizuluensis, T. foliensis and T. stellenbos chiana (Quaedvlieg et al., 2014;Silva et al., 2015) (Fig. 3). With the exception of T. stellenboschiana, discovered in Stellenbosch, South Africa, these leaf pathogens were all described from east ern Australia (Andjic et al., 2010;Crous et al., 2009bCrous et al., , 2006Summerell et al., 2006). It seems unlikely that two species causing the same disease symptoms are not each other's closest relatives (Gezahgne et al., 2005;Old et al., 2003). An association with Eucalyptus leaves also appears to be ancestral, as other species in the phyloge netic clade are all leaf associates (Crous et al., 2009a) (Fig. 3). Pathogenicity to Eucalyptus stems would, consequently, have had to evolve separately in T. zuluensis and T. gauchensis. In fact, T. gauchensis has been isolated from leaf specks on E. maidenii and E. tereticornis (Pérez et al., 2009), and T. zuluensis has been isolated as an E. grandis leaf endophyte (Marsberg et al., 2014). These results suggest that they retain some affinity for leaves and support the notion that they could have emerged from leafassociated ancestors.

P OSS I B L E O R I G I N S
Some diseases of Eucalyptus, such as Cryphonectria canker and Myrtle rust, have emerged from pathogens moving to exotic Eucalyptus plantations from native plant hosts (Glen et al., 2007;Nakabonge et al., 2006). Others are caused by patho gens that are known to have been introduced from the native range of Eucalyptus (Wingfield, 2003;Wingfield et al., 2008). The latter pathogens are typically of little consequence in nat ural Eucalyptus forests, but can become critically important in commercial monoculture stands (Burgess and Wingfield, 2002;Park et al., 2000).

Clues from population genetics studies
Population genetics studies do not support a South African ori gin for T. zuluensis, as originally hypothesized by Wingfield et al. (1996). Isolates from China and Malawi have greater gene and genotypic diversities than South African T. zuluensis popula tions (Chen et al., 2011;Cortinas et al., 2010), incongruent with the higher genetic diversity expected in the native range of a species (Hunter et al., 2008;McDonald, 1997). Teratosphaeria zuluensis populations elsewhere in Africa have low genetic diver sity, consistent with recent introductions (Dlugosch and Parker, 2008;Jimu et al., 2016a). All T. zuluensis populations analysed thus far have been predominantly clonal with strong population differentiation.
The population genetic structure of T. gauchensis in South America does not necessarily rule out a host jump from a native plant. South American populations of T. gauchensis, distributed between Argentina and Uruguay are surprisingly cohesive and diverse (Cortinas et al., 2011). Surveys of native Myrtaceae in Uruguay, however, identified only one species of Teratosphaeria that also occurs on Eucalyptus, and concluded that this species jumped from Eucalyptus onto the native plant and not the other way around (Pérez et al., 2013). Similarly, the five other members of the Mycosphaerellaceae and Teratosphaeriaceae that were identified on native Myrtaceae in the study of Pérez et al. (2013) are known Eucalyptus associates and do not represent a shift from native plants to Eucalyptus. In Africa, T. gauchensis populations in different countries are not structured and share genotypes, but genetic diversity is much lower than in South America (Jimu et al., 2016b). The genetic diversity and genotype identities in Zimbabwe implicate this country as the source of T. gauchensis introductions into Uganda and Ethiopia (Jimu et al., 2016b). The African condition reflects the anthropogenic movement of T. gauchensis between forestry plantations. Therefore, it is possible that the cohesive South American populations are a result of multiple introduc tions of the pathogen and high levels of trade in plant material between the different regions.

Australia: the source?
The majority of Eucalyptus leaf diseases, caused by species of Mycosphaerella and Teratosphaeria, are believed to have origi nated in Australia. Population genetics studies have confirmed high genetic diversities in eastern Australia for T. nubilosa (Hunter et al., 2008;Pérez et al., 2012) and T. epicoccoides (Taole et al., 2015). Teratosphaerianubilosa is devastating in planta tions, both in Australia and globally (Hunter et al., 2009), and T. epicoccoides has been associated with severe outbreaks in an Australian plantation (Carnegie, 2007), but these species do not cause significant damage in natural forests, probably because of the long evolutionary association with their Eucalyptus hosts. The species richness of Mycosphaerella and Teratosphaeria spe cies in natural Eucalyptus forests appears to be considerable, as surveys typically identify several new species (e.g. Andjic et al., 2010, Crous et al., 2009b, Summerell et al., 2006. Because of the numerous Teratosphaeria leaf pathogens known from Australia, the many that are probably still unknown and the minor symp toms that often occur on Eucalyptus trees in their natural envi ronment, it is plausible that T. zuluensis and T. gauchensis also occur in these forests, but in the absence of disease symptoms. Highthroughput sequencing (Kemler et al., 2013) and cul turedependent studies (Marsberg et al., 2014) have confirmed the presence of T. zuluensis as an E. grandis endophyte. The E. grandis trees analysed were from a South African plantation af fected by Teratosphaeria canker. Although they do not support an Australian origin, these studies indicate that T. zuluensis can exist as an endophyte (Fig. 4). Similarly, Jimu et al. (2016c) showed that T. zuluensis occurs on the seed and seed capsules of E. grandis trees displaying symptoms of Teratosphaeria canker. This is partic ularly significant, as the initial outbreak in South Africa was pre dominantly from seedderived E. grandis (Wingfield et al., 1996), and the study thus provided the first experimental evidence of Teratosphaeria species moving globally with seed consignments. Jimu et al. (2016c), however, also noted that infected seed does not yield infected seedlings, suggesting that horizontal rather than vertical pathogen transmission causes disease (Fig. 4).
The spread of Eucalyptus across plantations globally, and the diseases that tend to ensue, provides strong evidence that pests and pathogens are unwittingly moved around with Eucalyptus plant material (Jimu et al., 2016a(Jimu et al., , 2016cTaole et al., 2015). Teratosphaeria species have no known association with animal vectors, and the population genetics of pathogenic Teratosphaeria species studied thus far implies limited disper sal and gene flow. The different geographical distributions of T. zuluensis and T. gauchensis complicate the hypothesis that these species are natural Eucalyptus associates, as a common endophytic origin should have resulted in concurrent introduc tions. However, the numerous fungi naturally associated with Eucalyptus and the intraspecific variability of T. zuluensis and T. gauchensis imply that a diverse suite of potential pathogens resides within natural Eucalyptus forests. Combined with the fre quent movement of Eucalyptus germplasm from Australia to and between other countries, multiple introductions of Eucalyptus endophytes may have facilitated the emergence of different pathogens in the different environments. Current evidence there fore suggests that the Teratosphaeria canker pathogens occur naturally with Eucalyptus and have been cointroduced into plan tations with the germplasm (plants or seeds) of their hosts.

P U TAT I V E I N F EC T I O N C YC L E
The life cycles of the stem pathogens T. gauchensis and T. zu luensis have never been elucidated. However, evidence gath ered during the course of the past two decades provides a likely scenario that emerges in Eucalyptus plantations (Fig. 4). Although they are known only as pathogens of stem tissues, both T. gauchensis and T. zuluensis have been isolated from healthy Eucalyptus leaves (Marsberg et al., 2014;Pérez et al., 2009). In addition, a metagenomics study has revealed the pres ence of T. zuluensis in the seeds and seed capsules of trees suf fering from Teratosphaeria canker (Jimu et al., 2016c). Seedlings grown from clean seed did not contain the pathogen, suggesting that it is not vertically transferred from seed to plant, but rather that seedlings are infected horizontally after they begin to grow. Infection of the green tissues of Eucalyptus trees most probably occurs from conidia produced on cankers and other plant tissues carrying asymptomatic infections. This would be similar to many tree pathogens in the Botryosphaeriaceae, such as Diplodia sap inea (Bihon et al., 2011a(Bihon et al., , 2011b and Botryosphaeria dothidea (Marsberg et al., 2017), which infect healthy plant tissues and remain dormant until environmental conditions or host suscepti bility allow disease symptoms to develop. Studies showing that T. gauchensis and T. zuluensis can exist endophytically as latent pathogens in healthy plant tissues have yet to be conducted. However, the fact that T. zuluensis is found on seed capsules in the absence of symptoms (Jimu et al., 2016c), and that Kemler et al. (2013) found Teratosphaeriaceae to be one of the most common groups of fungi existing in healthy Eucalyptus tissues, provides strong evidence that these canker pathogens have a la tent asymptomatic phase in their life cycle.

C U R R E N T CO N C E R N S
The spread of T. gauchensis and T. zuluensis to plantations across the world presents an immense risk to Eucalyptus forestry. Despite the known threats, pathogens continue to be introduced into new areas, illustrated by the recent emergence of T. gauche nsis in Portugal and Zimbabwe (Jimu et al., 2015;Silva et al., 2015). The infection of E. globulus by T. gauchensis in Portugal is specifically alarming, as it represents a mild climate and a new host that usually only occurs in this mild climate. The expanded range of Teratosphaeria canker therefore increases the risk of ad aptation to different environmental conditions.
The introduction of new genotypes to plantations already affected by the disease also remains concerning. The popula tion size and genetic diversity of pathogens are regarded as important measures of their adaptability (McDonald and Linde, 2002;McDonald and McDermott, 1993); multiple introductions may increase this adaptability and, consequently, the ability to cause disease. The genetic diversity of T. gauchensis and T. zuluensis in certain parts of the world, especially China and Zimbabwe, implies that multiple genotypes have been intro duced independently (Chen et al., 2011;Jimu et al., 2016b). In areas with a high genetic diversity, more than one genotype occurs on a single host (Chen et al., 2011;Jimu et al., 2016b). In addition, Asian and African populations of T. zuluensis and South American populations of T. gauchensis show evidence of recombination, despite the lack of a known sexual state in either stem canker pathogen (Cortinas et al., 2011;Jimu et al., 2016a).
For several years, T. gauchensis and T. zuluensis were known as species causing the same disease, but in distinct geographi cal locations. Both species, however, continue to spread and, in 2014, T. gauchensis and T. zuluensis were detected in the same E. grandis plantation in Uganda (Jimu et al., 2014). One year later, T. gauchensis was found to be widespread in the southern African country of Zimbabwe (Jimu et al., 2015), a region in which only T. zuluensis was previously known from South Africa, Mozambique and Malawi . Undoubtedly, anthropogenic movement of plant material to and amongst African countries has brought these two pathogens into close proximity (Fig. 1).
The consequences of virulent T. gauchensis and T. zuluensis strains infecting the same host simultaneously are unknown. One species may simply outcompete and replace the other, as has been observed between the Dutch elm disease pathogens Ophiostoma ulmi and O. novoulmi (Brasier, 2001). Only the asexual forms of T. gauchensis and T. zuluensis are known and, in the absence of sexual reproduction, the opportunity for the acquisition of novel alleles is limited. History has shown that reproductive isolation in closely related allopatric fungi is often incomplete, and that contact between such species often leads to hybridization (Olson and Stenlid, 2002). This could be an ideal means for T. zuluensis and T. gauchensis to circumvent their ap parent low rates of recombination.
The outcome of hybridization between two pathogens may be variable, ranging from unfit hybrids to a novel lineage that outcompetes its parents or establishes in a new niche (Olson and Stenlid, 2002). Examples of the latter situation include hy brids within the rust genus Melampsora (Newcombe et al., 2000;Spiers and Hopcroft, 1994) and the oomycete genus Phytophthora (Érsek and Nagy, 2008), which have led either to expanded host ranges or the collapse of host resistance (Olson and Stenlid, 2002). Between the extremes of unsuccessful to successful hy bridization lies the possibility of introgression. Even shortlived hybrids may backcross, acting as 'genetic bridges' that facilitate gene transfer between parent taxa (Brasier, 2001;Harrison and Larson, 2014). For example, despite the clear fitness advantage of O. novoulmi, it has acquired vegetative compatibility genes from O. ulmi via introgression (Brasier, 2001). Advantageous characteristics are bound to introgress more easily as a result of favourable selection (Harrison and Larson, 2014) and, given ideal conditions, genes that influence important characteristics, such as pathogenicity, temperature sensitivity or host range, may be shared between T. zuluensis and T. gauchensis.

F U T U R E PROS PEC T S
After the initial outbreak of Teratosphaeria canker in South Africa (Wingfield et al., 1996), the majority of susceptible E. grandis trees were replaced with resistant clones (Cortinas et al., 2010). Teratosphaeria zuluensis perpetuated in the susceptible clones that were not replaced (Cortinas et al., 2010), but Teratosphaeria canker became only a minor problem in the remainder of the plantation (Fig. 2e). Similarly, in Ethiopia, stands of healthy E. ca maldulensis border infected E. camaldulensis stands, implying that they are resistant to T. gauchensis (Gezahgne et al., 2005). Both examples highlight the value of resistance breeding and se lection in combating this disease. Molecular genetic studies have played an important role in elucidating the taxonomy and geographical distribution of Teratosphaeria stem canker pathogens. Phylogenies based on DNA sequences have substantially improved our understanding of the taxonomy of these fungi (e.g. Cortinas et al., 2006c), and studies with microsatellite markers have unravelled their genetic diversity (or lack thereof) globally (e.g. Cortinas et al., 2010). However, little is known of their mode of infection, interaction with Eucalyptus hosts and the underlying basis of their specificity to stem rather than leaf tissues.
The genomics era has enabled many advances in the field of plant pathology (Aylward et al., 2017), as exemplified by the dis covery of T. zuluensis as an endophyte and seed contaminant (Jimu et al., 2016c;Kemler et al., 2013). The application of 'omics' tech niques to T. zuluensis and T. gauchensis will provide the next step towards addressing the current knowledge gaps, and will clearly accelerate the acquisition of new knowledge regarding these important tree pathogens. Both species are currently targeted for wholegenome sequencing (GOLD projects Gp0311872 and Gp0311873; https://gold.jgi.doe.gov). These sequences will enable comparative genomics studies between the two stem pathogens, as well as their leafassociated relatives, for which several genomes are in the pipeline. The initial aims of these studies will be to de termine the mating type systems of T. zuluensis and T. gauchensis, evaluate their risk of hybridization and discover which adaptations characterize their habitat switch from leaves to stems.

SU PP O R T I N G I N F O R M AT I O N
Additional supporting information may be found in the online version of this article at the publisher's web site: Fig. S1 Taxonomic history of Teratosphaeria zuluensis and T. gauchensis. Fig. S2 Culture morphology of Teratosphaeria gauchensis (row a, b) and T. zuluensis (row c, d) from above (row a, c) and below (row b, d). Cultures were grown on malt extract agar for 3 weeks at 26 °C, using an initial 5mm mycelial plug and 65mm Petri dishes.
Text S1 Phylogenetic methods and GenBank® accession num bers used to construct Fig. 3.