Eight species of the Botryosphaeriaceae (canker and dieback pathogens) were identified on native Syzygium cordatum in South Africa, based on anamorph morphology, ITS rDNA sequence data and PCR-RFLP analysis. The species identified were Neofusicoccum parvum, N. ribis, N. luteum, N. australe, N. mangiferae, Botryosphaeria dothidea, Lasiodiplodia gonubiensis and L. theobromae. Their pathogenicity on S. cordatum seedlings and a Eucalyptus grandis × camaldulensis clone was determined in glasshouse inoculation trials. Isolates of all identified species, except one of N. mangiferae, were more pathogenic on the Eucalyptus clone than on S. cordatum. Some of the species that cross-infected these hosts, such as N. ribis, N. parvum and L. theobromae, were amongst the most pathogenic on the Eucalyptus clone, while B. dothidea and L. gonubiensis were the least pathogenic. Results of this study illustrate that species of the Botryosphaeriaceae from native hosts could pose a threat to introduced Eucalyptus spp., and vice versa.
The Botryosphaeriaceae (Dothideales) is comprised of fungal species that have a wide geographic distribution and extensive host range, including Eucalyptus spp. (Myrtaceae) (von Arx & Müller, 1954; Crous et al., 2006). These fungi are latent and opportunistic pathogens that occur as endophytes in symptomless plant tissues and they can cause rapid disease development when plants are exposed to unsuitable environmental conditions such as drought, freezing, hot or cold winds, hail wounds or damage caused by insects or other pathogens (Fisher et al., 1993; Smith et al., 1996). Species of the Botryosphaeriaceae cause a wide variety of symptoms on all parts of Eucalyptus trees and on trees of all ages, but are mostly associated with cankers and dieback followed by extensive production of kino, a dark-red tree sap, and in severe cases mortality of trees (Smith et al., 1994, 1996; Old & Davison, 2000).
The Myrtaceae is a predominantly southern hemisphere angiosperm family that accommodates more than 3000 species, largely distributed in the tropical and temperate regions of Australasia, as well as Central and South America (Johnson & Briggs, 1981). Species of the Myrtaceae also form an integral part of the Southern African indigenous flora (Palgrave, 1977). In this context, the most widespread myrtaceous tree in South Africa is Syzygium cordatum (Palgrave, 1977). Eucalyptus species, native Australasian Myrtaceae, are the most widely grown trees in commercial forestry plantations, particularly in the tropics and southern hemisphere, including South Africa.
Movement of pathogens between native and introduced hosts has been recognized as a significant threat to plant communities (Slippers et al., 2005). Because of the potential threat of native pathogens to non-native Eucalyptus plantations, various recent studies considered fungal pathogens on native hosts in areas where Eucalyptus spp. are intensively planted (Wingfield, 2003; Burgess et al., 2006). These studies showed that pathogens which can cause severe diseases on Eucalyptus spp. also occur on native plants and thus pose a threat to Eucalyptus spp. Where plantations of non-native Eucalyptus spp. are established amongst closely related native myrtaceous trees, pathogens could cross-infect either the native or introduced host group and cause serious diseases (Burgess & Wingfield, 2001). For example, the rust fungus Puccinia psidii, which occurs on a variety of native Myrtaceae in South America, has become one of the main pathogens on exotic Eucalyptus spp. in that area (Coutinho et al., 1998).
In South Africa, species of the Botryosphaeriaceae are amongst the most important canker pathogens in plantations of non-native Eucalyptus spp., causing twig dieback, branch and stem cankers and mortality of diseased trees (Smith et al., 1994). These fungi have also recently been reported as endophytes from native South African trees closely related to Eucalyptus, such as S. cordatum and Heteropyxis natalensis (Smith et al., 2001). The Eucalyptus plantations mostly occur in the eastern part of the country where S. cordatum is widely distributed (Palgrave, 1977; Anonymous, 2002; Fig. 1). Thus, Botryosphaeriaceae that occur on this native tree could pose a threat to exotic Eucalyptus spp. and vice versa. However, there have not been any detailed studies on Botryosphaeriaceae on native hosts closely related to Eucalyptus in South Africa. Because of the economic importance of Eucalyptus plantations, as well as the need to protect native flora, identification and characterization of Botryosphaeriaceae from S. cordatum is of great concern.
Recent studies combined morphological characteristics and DNA sequence data to distinguish and identify species within the Botryosphaeriaceae (Denman et al., 2000; Zhou & Stanosz, 2001; Crous et al., 2006). Molecular approaches most commonly used to study Botryosphaeriaceae are comparisons of sequence data from the internal transcribed spacer (ITS) gene region of the rDNA operon (Denman et al., 2000; Zhou & Stanosz, 2001). However, some closely related or cryptic species of the Botryosphaeriaceae have been difficult to distinguish based on single gene genealogies. Comparisons of sequence data for multiple genes or gene regions were thus used to discriminate between these species (Slippers et al., 2004a, c). Furthermore, identification of large numbers of species has been facilitated by PCR restriction fragment length polymorphism (RFLP) techniques (Slippers et al., 2004b).
The aims of this study were to identify Botryosphaeriaceae occurring on native S. cordatum in South Africa, based on ITS rDNA sequence data, PCR-RFLP analysis and anamorph morphology. Isolates belonging to the Botryosphaeriaceae on S. cordatum and Eucalyptus were also compared, with special attention given to overlaps and the potential for cross infection. The pathogenicity of the Botryosphaeriaceae isolates from S. cordatum was furthermore tested on both a Eucalyptus clone and S. cordatum in glasshouse trials.
Materials and methods
Isolates used in this study were collected in surveys of Botryosphaeriaceae on native S. cordatum in different geographical regions of South Africa, in 2001 and 2002 (Table 1, Fig. 1). The 148 isolates that were collected from 11 S. cordatum sites during these surveys form the basis of this study. Between 5 and 45 trees were sampled from each site. From each tree, isolations were made from dying twigs and symptomless, visually healthy twig and leaf tissues. Leaves and twig portions (5 cm in length) were washed in running tap water and surface sterilized by placing them sequentially for 1 min in 96% ethanol, undiluted bleach (3·5–5% available chlorine) and 70% ethanol, then rinsed in sterile water. Treated twig portions were halved and pieces from the pith tissue (2 mm2) and segments of the leaves (3 mm2) were placed on 2% malt extract agar (MEA; 2% malt extract, 1·5% agar; Biolab) in Petri dishes. Following incubation for 2 weeks at 20°C under continuous near-fluorescent light and colonies resembling Botryosphaeriaceae with grey-coloured, fluffy aerial mycelium, were selected. These colonies were transferred to 2% MEA at 25°C and stored at 5°C. All isolates have been maintained in the Culture Collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa, and representative isolates were deposited in the collection of the Centraalbureau voor Schimmelcultures (CBS), Utrecht, the Netherlands.
Table 1. Isolates considered in the phylogenetic study and pathogenicity trials
Culture collections: CMW = Tree Pathology Co-operative Programme, Forestry and Agricultural Biotechnology Institute, University of Pretoria; KJ = Jacobs & Rehner (1998); ATCC = American Type Culture Collection, Fairfax, VA, USA; BRIP = Plant Pathology Herbarium, Department of Primary Industries, Queensland, Australia; CAP = culture collection of A.J.L. Phillips, Lisbon, Portugal; CBS = Centraalbureau voor Schimmelcultures, Utrecht, Netherlands; ICMP = International Collection of Microorganisms from Plants, Auckland, New Zealand; ZS = Zhou & Stanosz (2001).
Isolates sequenced in this study are given in bold.
Single conidial cultures from 21 isolates were grown on MEA for 7 days at 25°C in the dark. Template DNA was obtained from the mycelium using the modified phenol:chloroform DNA extraction method described in Smith et al. (2001). DNA was separated by electrophoresis on 1·5% agarose gels, stained with ethidium bromide and visualized under ultraviolet light. DNA concentrations were estimated against λ standard size markers.
The internal transcribed spacer (ITS) regions ITS1 and ITS2, and the intermediate 5·8S gene of the ribosomal RNA (rRNA), were amplified using the primer pair ITS1 and ITS4 (White et al., 1990). The PCR reactions were performed using the PCR protocol of Slippers et al. (2004b). PCR products were separated in a 1·5% agarose gel, stained with ethidium bromide and visualized under UV light. Sizes of PCR products were estimated against a 100 bp molecular weight marker XIV (Roche Diagnostics). The PCR products were purified using the High Pure PCR Product Purification kit (Roche Diagnostics).
DNA sequencing and analysis
Based on conidial morphology, the isolates of Botryosphaeriaceae from S. cordatum in South Africa were tentatively separated into eight groups. ITS rDNA sequences were determined for representative samples from all morphological groups (Table 1). To determine the identity and phylogenetic relationships of these isolates, ITS sequences of known species of the Botryosphaeriaceae were obtained from GenBank and included in the analyses (Table 1). The purified PCR products were sequenced using the same primers that were used for the PCR reactions. The ABI PRISM™ Dye Terminator Cycle Sequencing Ready Reaction kit (Perkin-Elmer) was used for sequencing reactions, as specified by the manufacturers. Sequence reactions were run on an ABI PRISM 3100™ automated DNA sequencer (Perkin-Elmer).
Nucleotide sequences were analysed using sequence navigator version 1·0·1. (Perkin-Elmer Applied BioSystems, Inc.) software and alignments were made online using mafft version 5·667 (http://timpani.genome.ad.jp/~mafft/server/) (Katoh et al., 2002). Gaps were treated as fifth character and all characters were unordered and of equal weight. Phylogenetic analyses of aligned sequences were carried out using paup (Phylogenetic Analysis Using Parsimony) version 4·0b8 (Swofford, 1999). Most-parsimonious trees were found using the heuristic search function with 1000 random addition replicates and tree bisection and reconstruction (TBR) selected as the branch swapping algorithm. Branches of zero length were collapsed and all multiple, equally parsimonious trees were saved. Branch support was determined using 1000 bootstrap replicates (Felsenstein, 1985). The trees were rooted using the GenBank sequences of Guignardia philoprina and Mycosphaerella africana, which are closely related to Botryosphaeriaceae. The sequence alignments and phylogenetic tree were deposited in TreeBASE as S1412, M2541.
PCR-RFLP fingerprinting techniques were applied to confirm the identity of isolates that were not sequenced and to identify the isolates that could not be separated based on ITS rDNA sequences. Amplicons obtained using primer pairs ITS1 and ITS4, or BOT15 (5′-CTGACTTGTGACGCCGGCTC-3′) and BOT16 (5′-CAACCTGCTCAGCAAGCGAC-3′) (Slippers et al., 2004a) were digested with the restriction endonuclease CfoI. The RFLP reaction mixture consisted of 10 µL PCR products, 0·3 µL CfoI and 2·5 µL matching enzyme buffer (Roche Diagnostics). The reaction mixture was incubated at 37°C overnight. Restriction fragments were separated on 1·5% agarose gel as described for PCR products. The results were compared with those of Slippers (2003).
Morphology and cultural characteristics
Fungal isolates were grown on 2% water agar (WA; Biolab) with sterilized pine needles placed onto the medium, at 25°C under near-UV light, to induce sporulation. Conidia that were released from pycnidia on the pine needles were mounted in lactophenol on glass slides and examined microscopically. Ten measurements of conidia were taken for each isolate. Measurements and digital photographs were taken using a light microscope, a HRc Axiocam digital camera and accompanying software (Carl Zeiss Ltd). Colony morphology and colour were determined from cultures grown on 2% MEA at 25°C under near-UV light. Colony colours (upper surface and reverse) were compared with those in the colour charts of Rayner (1970).
Fifteen isolates, representing eight species of Botryosphaeriaceae isolated from native S. cordatum in South Africa, were used in this study (Table 1). One isolate of Botryosphaeria dothidea and two isolates for each of the other seven species were randomly selected for inoculations. The isolates were grown on 2% MEA at 25°C under continuous near-fluorescent light for 7 days prior to inoculation.
Two-year-old trees of an E. grandis × camaldulensis clone (GC-540) and 1-year-old saplings of S. cordatum were selected for the pathogenicity trials under glasshouse conditions. Saplings of S. cordatum were raised from seeds taken from a single tree grown in the Kwambonambi (Kwazulu-Natal province) area. Trees and saplings selected for inoculations were grown in pots outside, and maintained in the glasshouse for acclimatization for 3 weeks prior to inoculation. Trees were inoculated during the spring-summer season (September 2003–February 2004). The glasshouse was subjected to natural day/night conditions and a constant temperature of approximately 25°C. Each of the isolates representing the different species was inoculated into the stems of 10 trees of each host species. Ten trees were also inoculated with sterile MEA plugs to serve as controls. The 160 inoculated trees, 10 for each fungal species and 10 as a control, were arranged in a randomized block design. The entire trial was repeated once under the same conditions, giving a total of 320 trees inoculated for each host species.
For inoculations, wounds were made on the stems of trees using a 6-mm-diameter (Eucalyptus clone) or a 4-mm-diameter (S. cordatum) cork borer to remove the bark and expose the cambium. Wounds were made between two nodes on the stems of trees approximately 250 mm (Eucalyptus) or 150 mm (S. cordatum) above soil level. Plugs of mycelium were taken from 7-day-old cultures grown on MEA using the same size cork borer, and were placed into the wounds with the mycelial surface facing the cambium. Inoculated wounds were sealed with laboratory film (Parafilm M, Pechiney Plastic Packaging) to prevent desiccation and contamination. Lesion lengths (mm) were measured 6 weeks after inoculation. The fungi were re-isolated by cutting small pieces of wood from the edges of lesions and plating them on 2% MEA at 25°C. Re-isolations were made from two randomly selected trees per isolate and tree species and from all trees inoculated as controls.
Pathogenicity for all isolates inoculated on the Eucalyptus clone and S. cordatum was determined based on the length of lesions (mm) that developed after 6 weeks. There was no significant difference between the two repeats of the pathogenicity trials and the data were therefore combined to represent one dataset for the analyses. Statistical analyses of the data were performed using sas statistical software (version 8, SAS Institute). The 95% confidence limits were determined for all means based on full model analysis of variance (anova). Differences between means were therefore considered significant at the P ≤ 0·05 level.
DNA sequence analyses
DNA fragments of approximately 600 bp were amplified. The ITS dataset consisted of 53 ingroup sequences, with G. philoprina and M. africana as outgroup taxa (Table 1). After alignment, the ITS dataset consisted of 593 characters; 432 uninformative characters were excluded, and 161 parsimony-informative characters were used in the analysis. The parsimony analysis (using heuristic searches) produced 276 most parsimonious trees of 414 steps (consistency index (CI) = 0·702, retention index (RI) = 0·915), one of which was chosen for presentation (Fig. 2).
The isolates considered in the phylogenetic analyses formed 12 clades, designated as groups I to XII (Fig. 2). These groups were resolved in two major clades that corresponded to species of Botryosphaeriaceae with Fusicoccum-like or Diplodia-like anamorphs. The Fusicoccum clade comprised six groups that represented: Neofusicoccum parvum and N. ribis (group I), N. mangiferae (group II), N. eucalyptorum (group III), N. australe (group IV), N. luteum (group V) and B. dothidea (group VI). Groups VII and VIII represented species with Lasiodiplodia anamorphs: Lasiodiplodia theobromae (group VII) and L. gonubiensis (group VIII). These two groups (VII and VIII) formed a distinct subclade (supported by 100% bootstrap value) within the Diplodia clade. The other major subclade within the Diplodia clade contained four groups corresponding to: D. mutila (group IX), D. corticola (group X), Diplodia sp. (= ‘Botryosphaeria obtusa’) (group XI) and Diplodia pinea (= Sphaeropsis sapinea) (group XII) (Fig. 2).
All the isolates obtained from S. cordatum in this study resided in seven groups (Fig. 2) as follows: N. parvum and N. ribis (group I), N. mangiferae (group II), N. australe (group IV), N. luteum (group V), B. dothidea (group VI), L. theobromae (group VII) and L. gonubiensis (group VIII).
Isolates that were not identified using DNA sequence comparisons were subjected to ITS PCR-RFLP analyses. Digests of the PCR products, obtained using primers ITS1 and ITS4, with CfoI produced two distinctive banding patterns. These profiles matched those of N. parvum/N. ribis (99 isolates) and N. luteum/N. australe (5 isolates) as shown by Slippers et al. (2004b). To further distinguish isolates of N. parvum from those of N. ribis, amplicons obtained using primers BOT15 and BOT16 were digested using the same restriction endonuclease. The two banding patterns obtained matched those of N. parvum (42 isolates) and N. ribis (57 isolates) as described by Slippers (2003). However, N. luteum and N. australe could not be separated using this technique.
Morphology and cultural characteristics
All 148 isolates of the Botryosphaeriaceae from S. cordatum produced anamorph structures on pine needles on WA within 2–3 weeks. No teleomorph (sexual) structures were observed. Based on conidial morphology, isolates were separated into eight groups. Five of these groups corresponded to Botryosphaeriaceae with Neofusicoccum anamorphs (Fig. 3a–f), one with a Fusicoccum anamorph (Fig. 3g) and two with Lasiodiplodia (Diplodia-like) anamorphs (Fig. 4a,b).
Representative samples from the groups emerging from morphological comparisons were identified based on ITS rDNA sequence comparison. As described earlier, isolates of N. parvum and N. ribis were separated based on PCR-RFLP analyses. Further morphological examination of isolates, identified based on DNA data, provided support for their identity.
Cultures of N. parvum were initially white with fluffy, aerial mycelium, becoming pale olivaceous grey from the middle of colony after 3–4 days; columns of the mycelium formed in the middle of colony reaching the lid; margins were regular; reverse sides of the colonies were olivaceous grey. Conidia were hyaline, smooth, aseptate and fusiform to ellipsoid (average of 420 conidia: 18·2 × 5·5 µm, l/w 3·3) (Fig. 3a). The 42 isolates were identified as N. parvum.
Colonies of N. ribis were initially white, becoming pale olivaceous grey from the middle of colony, with thick aerial mycelium reaching the lids of Petri dishes; margins were regular; reverse sides of the colonies were olivaceous grey. Conidia were hyaline, unicellular, aseptate, fusiform, apices tapered (average of 570 conidia: 21 × 5·5 µm, l/w 3·8) (Fig. 3b). The 57 isolates were identified as N. ribis.
The culture of the single B. dothidea isolate identified in this study produced greenish olivaceous appressed mycelium, its margins regular and the reverse sides of the colonies olivaceous grey to iron-grey. Conidiomata were readily formed in the middle of colonies after 3–4 days of incubation. Conidia were hyaline, smooth with granular contents, aseptate, narrowly fusiform (average of 10 conidia: 27·8 × 5·4 µm, l/w 5·1) (Fig. 3g).
Isolates of N. mangiferae produced pale olivaceous grey appressed mycelium, slightly fluffy on the edges of colonies, with sinuate margins, and the reverse sides of colonies were olivaceous. Conidiomata were readily formed in the middle of colonies after 3–4 days and covered the entire surface of the colonies within 7–10 days. Conidia were hyaline, fusiform (average of 300 conidia: 14·2 × 6·3 µm, l/w 2·25) (Fig. 3f). The 30 isolates were identified as N. mangiferae.
Cultures of N. luteum were initially white, becoming pale olivaceous grey from the middle of colonies within 3–4 days, with suppressed mycelium, moderately fluffy in the middle and with regular margins. A yellow pigment was noticeable after 3–5 days of incubation and was seen as amber yellow on the reverse side of Petri dishes; after 5–7 days colonies become olivaceous buff to olivaceous grey. Conidiomata were readily formed from the middle of colonies within 3–4 days and covered the whole surface of colonies within 7–10 days. Conidia were hyaline, fusiform to ellipsoid, sometimes irregularly fusiform, smooth with granular contents, unicellular, forming one or two septa before germination (average of 40 conidia: 18·9 × 6·3 µm, l/w 3·0) (Fig. 3d,e). The four isolates were identified as N. luteum.
Cultures of N. australe were very similar in morphology to those of N. luteum, but the yellow pigment produced in young cultures was brighter and a honey-yellow colour when viewed from the bottom of the Petri dishes. Conidiomata readily formed at the middle of colonies within 3–4 days and covered the colony surfaces within 7–10 days. Conidia were hyaline, fusiform, apices rounded, aseptate, rarely uniseptate (average of 70 conidia: 20·5 × 5·7 µm, l/w 3·6) (Fig. 3c). These conidia were slightly longer and narrower on average than those of N. luteum, which was also reflected in a higher l/w ratio. The seven isolates were identified as N. australe.
Isolates of L. theobromae produced initially white to smoke-grey fluffy aerial mycelium, becoming pale olivaceous grey within 5–6 days with regular margins; the reverse sides of the cultures were olivaceous grey to iron, becoming dark slate-blue after 7–10 days. Conidia were hyaline, aseptate, ellipsoid to ovoid, thick-walled with granular contents (average of 50 conidia: 27 × 14·7 µm, l/w 1·85) (Fig. 4b). Dark, septate conidia typical for this species were not observed in this study. The five isolates were identified as L. theobromae.
Isolates of L. gonubiensis were similar in culture morphology to those of L. theobromae. Conidia of L. gonubiensis were initially hyaline, unicellular, ellipsoid to obovoid, thick-walled with granular contents, rounded at the apex and occasionally truncate at the base. Aging conidia became cinnamon to sepia with longitudinal striations, forming one to three septa (average of 20 conidia: 33·9 × 18·9 µm, l/w 1·8) (Fig. 4a). The two isolates were identified as L. gonubiensis.
All Botryosphaeriaceae isolates tested for pathogenicity on the E. grandis × camaldulensis clone (GC-540) produced lesions within 6 weeks. No lesions developed on trees inoculated with sterile MEA plugs as controls. The fungi re-isolated from the lesions that developed on trees were the same as those used for inoculations. The original Botryosphaeriaceae species were re-isolated from all trees chosen for re-isolations. No Botryosphaeriaceae were re-isolated from the controls.
Statistical analyses showed that the mean lesion length for the majority of isolates used in the trial differed significantly from that of the controls (Fig. 5a). The longest lesions were produced by isolates of L. theobromae, while the size of lesions produced by B. dothidea and L. gonubiensis were not significantly different to those of the controls (Fig. 5a). The mean lesion lengths for different strains of the same Botryosphaeriaceae species were not significantly different, except for the isolates of L. theobromae. Thus, L. theobromae isolate CMW 14116 was significantly more pathogenic than isolate CMW 14114 (Fig. 5a).
All Botryosphaeriaceae isolates inoculated on S. cordatum saplings produced lesions within 6 weeks. However, the mean lesion lengths produced by majority of the isolates were not significantly different from those of the controls (Fig. 5b). Some trees inoculated as controls also developed small lesions, but no Botryosphaeriaceae could be re-isolated from these lesions, which appeared to represent wound reactions. The fungi re-isolated from the lesions on trees inoculated with fungal mycelium were the same as those used for inoculations. The longest lesions were produced by one isolate of N. mangiferae (CMW 14034) and the mean lesion length obtained for this isolate was significantly greater than that of the other isolate (CMW 14102) of the same species (Fig. 5b). However, there were no statistically significant differences between the lesion lengths for the different isolates of the other species of Botryosphaeriaceae (Fig. 5b). The mean lengths of lesions produced by one isolate of N. ribis (CMW 13992) and one isolate of L. theobromae (CMW 14116) were also significantly different from that of the control (Fig. 5b). All the other isolates inoculated onto S. cordatum saplings produced lesions that were not significantly different from those of the controls (Fig. 5b).
Isolates of all the Botryosphaeriaceae used in this study, except those of N. mangiferuae, were more pathogenic on the Eucalyptus clone than on S. cordatum. Analyses of variance showed that the interactions between mean lesion length produced by the species of Botryosphaeriaceae on the Eucalyptus clone and those on S. cordatum were statistically significant (P ≤ 0·001).
Eight species of the Botryosphaeriaceae were identified on native S. cordatum in South Africa in this study. They were N. ribis, N. parvum, N. luteum, N. australe, N. mangiferae, B. dothidea, L. theobromae and L. gonubiensis. The isolates were identified based on ITS rDNA sequence data, PCR-RFLP analysis and anamorph morphology. With the exception of B. dothidea and L. gonubiensis, this is the first report of all of these species of Botryosphaeriaceae on native S. cordatum. All eight species had the ability to infect and cause lesions on the stems of a E. grandis × camaldulensis clone and S. cordatum in glasshouse trials. Although lesions produced by most of the isolates on S. cordatum saplings were not significantly different from those on the controls, the pathogens could be re-isolated from these lesions. In the case of some species, such as N. ribis, L. theobromae and F. mangiferae, one isolate did not produce lesions that differed from those of the control, while the other isolate did. From these data, and knowledge of the fungi on other hosts, it was concluded that this group of fungi could be regarded as potential pathogens of Syzygium. However, apart from the isolates of B. dothidea and L. gonubiensis, all the other Botryosphaeriaceae in this study produced lesions on the Eucalyptus clone that were significantly different from those of the controls. They should be considered as potential threats to plantation grown Eucalyptus spp. in South Africa.
Neofusicoccum ribis was the dominant species collected from native S. cordatum in South Africa in this study. This fungus represented 38% of all isolates obtained and was found in most of the areas surveyed. This abundant and wide distribution on a native host might indicate that this species is native to this region. Neofusicoccum ribis has been reported from Eucalyptus spp. (Myrtaceae) in its native range in Australia and on non-native Eucalyptus spp. in plantations (Old & Davison, 2000), but has not been identified on Eucalyptus spp. in South Africa (Slippers et al., 2004a). These identifications should, however, be interpreted with caution, as the distinction between N. parvum and N. ribis had not been recognized at the time of these studies (Slippers et al., 2004a). Furthermore, N. ribis as identified in this study (using RFLPs) was also interpreted as representing the N. ribis sensu lato group rather than strictly conspecific populations with the type isolates of this species, as identified by Slippers (2003). Further analyses using sequence data for additional gene regions and other variable markers will be required to more clearly characterize populations and potential cryptic species in this group. Neofusicoccum ribis was one of the most pathogenic species of the Botryosphaeriaceae on the Eucalyptus clone in this study. This fungus should thus be considered as a potentially important pathogen of Eucalyptus spp. in South Africa.
Isolates of N. parvum represented 28% of the total number of isolates obtained in this study. Recent studies showed that N. parvum is an important and widely distributed pathogen of non-native Eucalyptus plantations in South Africa (Slippers et al., 2004b). The wide distribution of N. parvum on non-native and native Myrtaceae in South Africa raises intriguing questions, such as whether these populations are native or introduced and how they might be interacting with each other. The movement of this pathogen between these important host groups represents a potential threat for both groups and should be further investigated. Isolates of N. parvum used in this study also developed only slightly smaller lesions than those of closely related N. ribis, illustrating its potential threat to Eucalyptus plantations in South Africa.
Only one isolate obtained from S. cordatum was identified as B. dothidea (anamorph Fusicoccum aesculi). This species is one of the most commonly reported members of the Botryosphaeriaceae from a wide variety of hosts, including Eucalyptus spp. (von Arx & Muller, 1954; Smith et al., 2001). While B. dothidea was considered to be an important canker pathogen of Eucalyptus spp. in South Africa (Smith et al., 1994), some of these isolates that were the most pathogenic (Smith et al., 2001) were re-identified as N. parvum (Slippers et al., 2004b). Botryosphaeria dothidea was seldom encountered on Eucalyptus spp. in other studies on this host (Slippers et al., 2004b) and results of the present study suggest that it is probably not an important pathogen of this tree.
High numbers of isolates from S. cordatum were identified as N. mangiferae. This species is best known as a pathogen of mango (Mangifera indica) worldwide, particularly in Australia (Johnson et al., 1992). Neofusicoccum mangiferae was earlier reported under different names from mango in South Africa (Darvas, 1991). Interestingly, however, a recent comprehensive study of Botryosphaeriaceae from mango plantations in South Africa, using a combination of DNA-based techniques and morphological data, did not report this species (Jacobs, 2002). The fact that this fungus is highly pathogenic on S. cordatum might imply that it has been introduced into South Africa on other woody plants. Studies focused on the origin of N. mangiferae are likely to yield intriguing results, relevant to commercial forestry and to the protection of natural biodiversity in South Africa.
Neofusicoccum luteum and phylogenetically closely related N. australe were identified on S. cordatum in this study, but have not been recorded on Eucalyptus spp. in South Africa. Neofusicoccum australe is a recently described species (Slippers et al., 2004c) and the present study is the first to consider the pathogenicity of this fungus on Eucalyptus. Neofusicoccum luteum was highly pathogenic to the Eucalyptus clone and its occurrence on the related S. cordatum in South Africa is of concern. Neofusicoccum luteum and N. australe were not the most commonly encountered species of the Botryosphaeriaceae on S. cordatum, but their presence alone provides sufficient evidence that they are well established in the country.
Two Lasiodiplodia species were identified in this study. Lasiodiplodia theobromae was isolated from S. cordatum in subtropical areas of South Africa. This fungus is an opportunistic pathogen with an extremely wide host range, including more than 500 host plants, mostly in tropical and subtropical regions (Punithalingam, 1976), and has previously been isolated from exotic Acacia, Eucalyptus and Pinus spp. in South Africa (Crous et al., 2000; Burgess et al., 2003). Lasiodiplodia theobromae was the most pathogenic species to the Eucalyptus clone in this study. Although the two isolates of L. theobromae displayed different levels of pathogenicity, both were highly pathogenic. Lasiodiplodia theobromae might be considered a potentially important pathogen of Eucalyptus in South Africa and studies to consider its pathogenicity to different species and hybrid clones would be warranted. Another Lasiodiplodia species isolated from S. cordatum was recently described as L. gonubiensis (Pavlic et al., 2004) and was isolated from a geographical region with a moderate climate where L. theobromae was absent. Lasiodiplodia gonubiensis was only very mildly pathogenic to the Eucalyptus clone, even though it is most closely related to the highly pathogenic L. theobromae.
The results of this study provide an interesting insight into the diversity of Botryosphaeriaceae occuring on native S. cordatum in South Africa. Some of these fungi appear to be potentially important pathogens of Eucalyptus spp. and future surveys should recognize this fact. Clearly, additional studies such as the one presented here, considering the pathogenicity of these fungi, will be needed to better understand their importance. This study emphasises the threat of cross-infecting species of the Botryosphaeriaceae to both native and introduced Myrtaceae. In a recent study, Burgess et al. (2006) showed that there is no restriction to the movement of N. australe between native and planted eucalypts in Western Australia. Population studies on other species of the Botryosphaeriaceae are, therefore, planned to provide further insight into their movement between native and cultivated hosts in South Africa.
We thank the National Research Foundation (NRF), members of the Tree Protection Co-operative Programme (TPCP), the THRIP initiative of the Department of Trade and Industry, and the Department of Science and Technology (DST)/NRF Centre of Excellence in Tree Health Biotechnology (CTHB), South Africa for financial support. We also thank Dr B. E. Eisenberg who provided the statistical analyses.