Introduction

Commercial forestry in Uruguay using exotic species of Pinus and Eucalyptus has expanded rapidly during the course of the last two decades. Eucalyptus plantations now cover approximately 457,000 ha with >50% of this area being planted to Eucalyptus globulus (Petraglia and Dell’Acqua 2006). Other species commonly planted include E. grandis and E. dunnii. As in other areas of the world, the principal end-use of these plantations is in the pulp and paper industry, which is growing rapidly in Uruguay.

More than 100 species of Mycosphaerella have been identified from Eucalyptus, where they primarily cause leaf spots (Burgess et al. 2007; Crous 1998; Crous et al. 2007). Collectively, these leaf diseases are referred to as Mycosphaerella leaf disease (MLD) (Crous 1998). However, Crous et al. (2007) have shown recently that the genus Mycosphaerella is polyphyletic and therefore several species occurring on Eucalyptus have been transferred to the genus Teratosphaeria. Despite the changes in the species names, the common name for the associated disease has not changed. Thus, MLD on Eucalyptus is currently caused by both Mycosphaerella spp. and Teratosphaeria spp. (Crous et al. 2007; Hunter et al. 2009). Mycosphaerella leaf disease reduces the photosynthetic capacity of leaves causing premature defoliation, shoot die-back and decreased growth resulting in significant growth losses (Lundquist and Purnell 1987; Milgate et al. 2005; Mohammed et al. 2003; Park and Keane 1982a).

Despite the fact that many species of Teratosphaeria and Mycosphaerella infect Eucalyptus, few studies have considered the pathogenicity of these fungi. Certain species such as Teratosphaeria nubilosa (= Mycosphaerella nubilosa) and Teratosphaeria cryptica (= Mycosphaerella cryptica) act as primary pathogens infecting living leaf tissue and acquiring nutrients from them in a hemibiotrophic fashion (Hunter et al. 2009; Park and Keane 1982b). These two species are considered as the most important causing MLD worldwide (Carnegie and Ades 2002; Hunter et al. 2009; Park and Keane 1982a). In contrast, many Teratosphaeria and Mycosphaerella species on Eucalyptus are apparently mildly pathogenic or secondary colonisers of necrotic tissue, following infection by more pathogenic species, insect damage or from natural leaf senescence (Crous 1998; Crous et al. 2006).

Alfenas et al. (2004) reported four species of Mycosphaerella and Teratosphaeria in Brazil, namely M. marksii, M. parkii, T. suberosa and T. suttonii. However, there are no reports of MLD resulting in severe losses either in Brazil (Alfenas et al. 2004) or in the rest of South America. Severe defoliation of E. globulus was observed in Chile during 2000 and while M. grandis, M. lateralis, M. parkii, M. walkeri, T. africana and T. molleriana were associated with lesions, the primary cause of the defoliation was abiotic stress (Ahumada 2003). In Uruguay, several species of Mycosphaerella and Teratosphaeria have been reported, including M. marksii, M. walkeri, M. vespa and T. suberosa (Balmelli et al. 2004), but none are known as aggressive pathogens.

A recent population study of T. nubilosa suggests that the natural range of this species is eastern Australia (Hunter et al. 2008; 2009), coinciding with the natural distribution of E. globulus and E. nitens (Tibbits et al. 1997), two of the most susceptible species to T. nubilosa (Carnegie et al. 1998). The introduction of Eucalyptus spp. into many Southern Hemisphere countries during the course of the last century has allowed T. nubilosa to expand its geographical distribution considerably (Hunter et al. 2009). The first report of T. nubilosa (as Mycosphaerella molleriana) outside Australia was in 1950 in South Africa, although it is believed that it was present in that country as early as the 1930s (Lundquist and Purnell 1987). By 1982, T. nubilosa was known to be broadly distributed in New Zealand (Mohammed et al. 2003). The introduction of T. nubilosa into Western Australia most likely occurred between 1994 and 1999 (Maxwell et al. 2001). In 1995, it was reported in Kenya, Tanzania and Zambia (as Mycosphaerella juvenis) (Crous 1998) and in 2000 in Ethiopia (Hunter et al. 2008). Teratosphaeria nubilosa apparently arrived in Europe around 2001, where it was associated with E. globulus in four regions of Spain (Crous et al. 2004) and more recently in Portugal (Hunter et al. 2008). Given its widespread distribution as an alien invasive pathogen, it is interesting that T. nubilosa has not previously been reported from any South American country.

During the autumn of 2007, severe defoliation resulting from symptoms resembling those of MLD was observed on young E. globulus plantations in Uruguay. The change in the status of MLD in the country and the severity of the defoliation led to the suspicion that a new pathogen had been introduced into Uruguay. The aims of this study are to isolate and identify the causal agent of the disease. Once the pathogen had been identified, an investigation to determine the most probable origin of the pathogen was undertaken using 10 microsatellite loci. In this regard, the three possible areas of origin that were considered included Spain, because this country has acted as a source of seed and clonal material for Uruguayan forestry (Balmelli 2006); South Africa, the country where E. globulus was first introduced as a non-native for forestry in 1853 (Potts et al. 2004) and Australia, which is known to represent the native range of T. nubilosa and most Eucalyptus grown in Uruguay (Balmelli 2006; Hunter et al. 2008).

Materials and methods

Sample collection and lesion description

Diseased juvenile, intermediate and adult Eucalyptus leaves showing symptoms of MLD were collected during December 2007. Juvenile leaves were obtained from young trees, coppice growth and epicormic shoots on older trees. Plantations surveyed included E. globulus subsp. globulus, E. globulus subsp. maidenii and E. dunnii of different ages. Sampling was conducted in six provinces of Uruguay, namely Colonia, Lavalleja, Maldonado, Paysandú, Río Negro and Rocha. Eucalyptus leaves were air-dried and transported to a laboratory for detailed description of the lesions and isolations.

Fungal isolations

Lesions on the affected leaves were typical of those caused by a Mycosphaerella sp. or Teratosphaeria sp. with pseudothecia abundant on the abaxial surfaces. Fungal strains were isolated using the method described by Crous (1998). Lesions were excised from the leaves, hydrated for 2 h and attached to the undersides of Petri dish lids, with pseudothecia facing downwards over 2% Malt Extract Agar (MEA). After 24 h, germinating ascospores were observed on the surface of the agar, photographed using phase contrast microscopy and classified according to the germination patterns described by Crous (1998). Single germinating ascospores were transferred to fresh 2% MEA plates to obtain pure cultures, initially incubated at 15°C for 2 weeks and then at 20°C for a further 30 days. All cultures obtained in this study were deposited in the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa.

DNA extraction, PCR amplification and sequencing

Mycelium was scraped from the surface of actively growing cultures, freeze-dried and ground to a fine powder using sterile metal beads on a Mixer Mill type MM 301 Retsch® tissue lyser (Retsch, Germany) for 3 min at a frequency of 30 cycles s-1 DNA was isolated following the protocol described by Cortinas et al. (2006).

The internal transcribed spacers (ITS1 and ITS2) and the 5.8S gene of the rDNA operon were amplified using the primer pairs ITS-1 and ITS-4 (White et al. 1990). A portion of the beta tubulin-2 gene region was amplified using the primers Bt-2a and Bt-2b (Glass and Donaldson 1995). PCR reactions were conducted with the same reaction mixture as those utilised by Cortinas et al. (2006) with an initial denaturation at 95°C for 2 min, followed by 35 cycles of 30s at 94°C, 30 s at 57°C, 1 min at 72°C and a final elongation step of 7 min at 72°C.

For DNA sequencing reactions, PCR products were purified using Sephadex G-50 (Sigma Aldrich) in Centri-Sep spin columns (Princeton Separations) following the manufacturer’s instructions. Purified amplicons were sequenced using the ABI Prism Big Dye Terminator Cycle sequencing reaction kit v 3.1 (Applied Biosystems) following the manufacturer’s protocols. Both forward and reverse primers were utilised in all cases to sequence both strands.

Phylogenetic analyses

DNA sequences obtained in this study were analysed and edited in MEGA v 3.1 (Kumar et al. 2004). Analysed sequences were compared to representative Mycosphaerella and Teratosphaeria sequences deposited in GenBank, including T. nubilosa sequences reported from different countries (Table 1). All sequences obtained in this study were deposited in GenBank (Table 1).

Table 1 Mycosphaerella and Teratosphaeria isolates used in this study for phylogenetic analyses
Full size table

DNA sequence alignments were conducted in MAFFT v 6. employing the high-accuracy option of local alignment mode L-INS-I (Katoh et al. 2005) and manually aligned by inserting gaps when necessary in MEGA v 3.1 (Kumar et al. 2004). MODELTEST v 3.7 (Posada and Crandall 1998) was utilised to select the best nucleotide substitution model to fit the data. Using the Akaike Information Criterion (AIC), the Transitional Model (TIM + G) with a gamma distribution of the rate variation was then selected for a Bayesian analysis. Gaps were treated as missing data. Four Monte-Carlo Markov chains were run twice in MrBayes 3.1.1 (Ronquist and Huelsenbeck 2003) for two million generations starting from random trees; trees were sampled every 100th generation. The first 3000 trees were discarded (burn-in). The remaining 17,000 trees were utilised to obtain a consensus Bayesian tree (50% majority-rule), rooted using Neofusicoccum parvum (= Botryosphaeria parva) (AF343467) as the outgroup taxon.

For parsimony analysis, heuristic searches were conducted in PAUP v 4.0b10 (Swofford 2002) using Tree Bisection Reconnection (TBR) as the branch-swapping algorithm and the following parameters: MulTrees were on, starting trees were obtained via stepwise addition of 100 random replicates and remaining trees were added in single sequence fashion, MaxTrees was set at 100. Parsimony uninformative characters were excluded from the analysis. Finally, 1000 bootstrap replicates were conducted to determine confidence levels of branching points in the phylogenetic tree.

Genescan and population diversity

A total of 46 isolates collected from leaves on different trees from four provinces of Uruguay was selected to consider the diversity of the population (Table 2). Ten microsatellite loci were amplified using the fluorescently-labelled primers developed by Hunter et al. (2006). PCR reactions were conducted as described above, starting with an initial denaturation at 94°C for 2 min, followed by 35 cycles of 30 s at 94°C, 30 s at 60°C, 1 min at 72°C and a final elongation step for 45 min at 72°C. An exception was made in those reactions employing the MN-8 primer pair, where the annealing temperature was set at 54°C and not at 60°C. GeneScan-500 LIZ Size Standard (Applied Biosystems) was utilised to determine the allele sizes by electrophoresis on an ABI PRISM 3100 Genetic Analyser (Applied Biosystems). Genescan and Genemapper software packages (Applied Biosystems) were used for further data analysis.

Table 2 Origin of the Uruguayan Teratosphaeria nubilosa isolates used in this study
Full size table

Results

Sample collection and lesion description

Leaf spot and blotch symptoms were observed in all surveyed areas on E. globulus subsp. globulus and E. globulus subsp. maidenii, but they were less evident on E. dunnii. Similar leaf spots were observed on both juvenile and intermediate age foliage (Fig. 1). Leaf spots were present on leaves from the previous season’s growth as well as on juvenile leaves from the current season. Usually the second or third fully expanded leaf on new branches was already showing symptoms, suggesting the presence of a primary pathogen.

Fig 1
figure 1

Leaf symptoms, lesions and germinating ascospore of Teratosphaeria nubilosa on different hosts. a Adaxial surface of juvenile foliage of E. globulus subsp. maidenii b Abaxial surface of the same leaf in a c Intermediate foliage of E. globulus subsp. maidenii d E. globulus subsp. globulus (adaxial surface) e E. dunnii (adaxial surface) f Lesion on E. globulus subsp. maidenii (abaxial surface) g Lesion on E. dunnii (adaxial surface) h Germinating ascospore after 24 h

Full size image

The most common lesions observed were yellow to brown in colour, round to angular in shape, frequently coalescing to form larger blotches across the leaf surface (Fig. 1). As they aged, lesions became darker and pseudothecia became more visible, especially, but not exclusively, on the abaxial leaf surfaces (Fig. 1b, f). Lesion borders were surrounded by raised corky reddish to red-purple margins (Fig. 1f, g). These lesions were typical of MLD.

Evaluation of the severity of the disease was beyond the scope of the current study; however, remarkable differences in the impact of the disease were observed between Eucalyptus species. Leaves of both sub-species of E. globulus showed several spots per leaf and in severely affected plantations the diseased area covered up to 50% of the leaf lamina (Fig. 1a–d). On E. dunnii, however, typically only one to three small lesions were observed per leaf (Fig. 1e). Adult foliage also exhibited symptoms of MLD, but the characteristics of the lesions were different from those found on juvenile and intermediate foliage. Lesions were usually smaller than typical for the pathogen and they did not coalesce to form blotches. These lesions were more frequent on senescing leaves, characteristic of saprophytic or slightly pathogenic Mycosphaerella spp. and Teratosphaeria spp.

Fungal isolations

Germinating ascospores from leaf spots found on both juvenile and intermediate age foliage showed the type F germination pattern described by Crous (1998). The same germination pattern was observed in ascospores obtained from E. globulus subsp. globulus, E. globulus subsp. maidenii and E. dunnii. Both cells of the ascospores produced germ tubes that grew parallel to the long axis of the spore. Spores became distorted and slightly constricted at the medium septum and did not darken during germination (Fig. 1h). Conversely, ascospores isolated from adult leaves showed a different germination pattern to those from young and intermediate age foliage.

Phylogenetic analyses

PCR amplifications of the ITS region yielded amplicons of approximately 600 bp for all the isolates. Following alignment, the DNA sequence data contained a total of 580 characters, 397 of which were constant, 119 were variable and parsimony-uninformative and 64 were parsimony-informative. The consensus Bayesian phylogram showed that all isolates from Uruguay grouped together in a well supported clade (Bayesian support = 0.99) (Fig. 2). The ex-epitype isolate of T. nubilosa was also included in this clade, along with T. nubilosa isolates from different countries and hosts. Similarly, parsimony analysis resulted in a phylogenetic tree where all isolates from Uruguay and T. nubilosa grouped in a clade showing 100% bootstrap support (data not shown).

Fig 2
figure 2

Bayesian consensus tree obtained from the ITS sequence data. Bayesian support values are shown on the branching points. Uruguayan isolates are shown in bold and the number of isolates per province is provided in brackets. The outgroup branch has been shortened

Full size image

PCR amplifications of the beta tubulin-2 locus yielded amplicons of approximately 400 bp for all the isolates. The beta tubulin-2 sequences for all 20 Uruguayan isolates were identical (data not shown but see Table 1 for GenBank accession numbers). In addition, beta tubulin-2 sequences of Uruguayan isolates showed 100% similarity to the sequence for the ex-epitype culture of this species and to the sequence of a previously identified T. nubilosa isolate from Spain (data not shown but see Table 1 for GenBank accession numbers).

Genescan and population diversity

All ten microsatellite loci used to describe the genetic diversity of the isolates obtained in this study were monomorphic, showing only one allele per locus. In this way, allele sizes 267, 183, 309, 162, 380, 319, 216, 137, 224 and 100 bp were obtained from the loci MN-1, MN-2, MN-3, MN-4, MN-7, MN-8, MN-9, MN-10, MN-11 and MN-14 s respectively. This reflects the presence of only one multilocus haplotype in Uruguay. Interestingly, this multilocus haplotype was identical to the unique haplotype found by Hunter et al. (2009) in isolates of T. nubilosa from Portugal and Spain and referred to as the Iberian populations. Furthermore, all 10 microsatellite primer pairs developed by Hunter et al. (2006) were species-specific for T. nubilosa and amplified the expected DNA fragments in all 46 Uruguayan isolates. These results add substantial confidence to the DNA sequence-based identification of the pathogen in Uruguay as T. nubilosa.

Discussion

The nature and severity of the outbreak of MLD, as well as the leaf symptoms observed in the Uruguayan Eucalyptus plantations, are consistent with previous descriptions of T. nubilosa elsewhere (Crous 1998; Hunter et al. 2009; Park and Keane 1982a). The presence of leaf spots on leaves from the current season’s growth is characteristic of Teratosphaeria species that are primary pathogens. Moreover, the germination pattern of ascospores from the isolated fungus coincides with the descriptions of the germination pattern of T. nubilosa (Crous 1998; Hunter et al. 2009). The Uruguayan isolates were grouped in a clade with 0.99 of Bayesian support and 100% of bootstrap support in the phylogenetic analyses, which included the ex-epitype culture of this species and T. nubilosa isolates collected from seven countries on three continents. In addition, the beta tubulin-2 sequences of the Uruguayan isolates showed 100% similarity to the sequence for the ex-epitype culture of this species and to the sequence of a previously identified T. nubilosa isolate from Spain (Crous et al. 2004). All 10 microsatellite primer pairs that are species-specific for T. nubilosa (Hunter et al. 2006) amplified the expected amplicon successfully and the allele sizes for the Uruguayan and Iberian isolates were identical. These data allowed us to conclusively identify the causal agent of the new and severe outbreak of MLD in Uruguay as T. nubilosa.

The discovery of T. nubilosa in this study represents the first record of this important pathogen in South America. In a recent population genetic study on T. nubilosa, Hunter et al. (2008) suggested that the centre of origin of the pathogen is eastern Australia. They further showed that the pathogen had been introduced into Ethiopia, Kenya, New Zealand, Portugal, South Africa, Spain, Tanzania, Western Australia and Zambia (Hunter et al. 2008; 2009). It is thus intriguing that T. nubilosa has taken as long as it has to reach South America.

Teratosphaeria nubilosa was associated with severe leaf disease on E. globulus, which is the most highly susceptible species (Carnegie and Ades 2002; Carnegie et al. 1998; Milgate et al. 2005). Leaf disease caused by T. nubilosa can be extremely serious as was found in South Africa where commercial forestry with E. globulus was abandoned during the 1930s due to the impact of this disease (Hunter et al. 2009; Lundquist and Purnell 1987). Likewise, two attempts in 1971 and 1997 at afforestation with E. globulus have failed due to MLD in north-western Tasmania (Mohammed et al. 2003), and T. nubilosa caused almost complete defoliation on E. globulus subsp. maidenii in northern New South Wales, Australia (Carnegie 2007). Teratosphaeria nubilosa has been reported as a primary pathogen of twelve Eucalyptus spp. and hybrids, including E. grandis, which is widely planted in Uruguay (Crous 1998; Hunter et al. 2009; Park and Keane 1982a). The appearance of the pathogen in Uruguay is, therefore, serious and has the potential to limit forestry in this country.

The appearance of T. nubilosa in Uruguay raises substantial concerns regarding further spread of this pathogen to other countries in South America. As in Uruguay, E. globulus is extensively planted in several South American countries such as Chile, Colombia, Ecuador and Peru (Potts et al. 2004). The appearance of T. nubilosa in Uruguay increases the likelihood that it will now move to neighbouring countries. This could occur by natural spread, given that T. nubilosa produces windborne ascospores, a quality that confers high dispersion ability (Hunter et al. 2009; Park and Keane 1982b). Long-distance spread to countries further afield in South America is also likely to be enhanced by the apparent first occurrence of T. nubilosa in the region. This would most likely occur through the movement of planting material, which is commonly exchanged by companies owning land in more than one country of the region (Wingfield et al. 2008).

All ten microsatellite loci developed by Hunter et al. (2006) and used to analyse the Uruguayan population of T. nubilosa were monomorphic. This is despite the fact that these microsatellite markers have previously been shown as highly variable on a global suite of isolates of the pathogen, with a total of 68 multilocus haplotypes (Hunter et al. 2006; 2008; 2009). There seems little doubt that the pathogen in Uruguay is represented by a single clone. Teratosphaeria nubilosa is homothallic, where single individuals can undergo sexual reproduction alone (Park and Keane 1982b). An intriguing result of this study was that the single multilocus haplotype found in Uruguay, is the same haplotype found in Spain and Portugal, where the population of T. nubilosa is also represented by a single clone (Hunter et al. 2008; 2009). This result suggests strongly that the pathogen was introduced into Uruguay from one of these countries where its population has a very limited diversity, rather than from Australia, the area of origin of the fungus or an area such as South Africa where T. nubilosa is represented by a relatively diverse population (Hunter et al. 2009, 2008). Forestry companies in Uruguay have close associations with those particularly in the Iberian Peninsula and opportunities for an introduction into Uruguay from that area would have been great.

Commercial Eucalyptus plantations in Uruguay originated from seed and clonal material imported predominantly from Australia, Chile and Spain (Balmelli 2006). The movement of seed and asymptomatic vegetative material among countries is one of the most important vehicles for plant pathogen movement (Elmer 2001). As far as we know, there is no study showing that Mycosphaerella or Teratosphaeria species associated with MLD are seedborne; however, the widespread distribution of apparently Australian Mycosphaerella spp. on Eucalyptus strongly suggests that this is the case (Ciesla et al. 1996; Wingfield et al. 2008). Although there is no evidence to suggest that this has occurred, the introduction of clonal plant material could have also provided a means of introduction of the pathogen, particularly considering that symptoms of T. nubilosa are visible only after four weeks of incubation (Park and Keane 1982b).

Annual Eucalyptus disease surveys have been conducted in Uruguay during the course of the last ten years (M. J. Wingfield, unpublished data). These surveys have included special efforts to identify Mycosphaerella spp. associated with leaf diseases in the country and there has been no evidence to suggest that T. nubilosa was present. This suggests strongly that T. nubilosa in Uruguay represents a recent introduction of a pathogen that has spread rapidly in the country. The fact that T. nubilosa was first reported in the Iberian Peninsula in 2001 (Crous et al. 2004) leads us to believe that this pathogen entered Uruguay during the course of the last seven years.

The only economically viable management strategy for MLD is the development of resistant or tolerant Eucalyptus genotypes as has been true for various other diseases of these trees (Tibbits et al. 1997; Wingfield 2003). Eucalyptus globulus provenances are known to differ in their susceptibility to diseases, including MLD in Uruguay (Balmelli et al. 2004). In addition, there are excellent opportunities to plant non-susceptible Eucalyptus spp. and hybrids of E. globulus with species that have a low level of susceptibility to infection by T. nubilosa (Carnegie and Ades 2002; Carnegie et al. 1998; Milgate et al. 2005). Given the serious damage that this pathogen is able to cause, evaluation and development of resistant planting stock in Uruguay should be considered a priority.