Occurrence and possible role of endophytic fungi associated with seed pods of Colophospermum mopane (Fabaceae) in Botswana

Endophytic fungi were isolated from different aged surface sterilised pods of Colophospermum mopane and succession patterns were investigated. Lignocellulolytic enzyme assays as well as histological and fine structural studies were used to investigate if succession was related to different lignocellulolytic abilities of the fungi. Representatives of the common genera (Alternaria, Phoma and Phomopsis) were qualitatively tested for lignocellulolytic enzyme activity. Samples of endophyte infected and uninfected pod pericarps were fixed and sectioned for light microscopy and TEM. Samples were also viewed with SEM. A fungal succession was evident as pods aged, and detached from the tree. Strains of all three genera demonstrated lignocellulolytic abilities. Microscopy studies suggested that Phoma was only capable of utilising moderately lignified mesophyll cells whereas Phomopsis and Alternaria could degrade heavily lignified fibres. This could explain the pattern of succession observed: Phoma colonised younger pods where more unlignified resources were available, whereas Phomopsis and Alternaria colonised older pods, thus making use of the remaining lignified resources. Changes in endophyte abundance and diversity are related to the age and degree of decay of the pods, and determined by the lignocellulolytic abilities of the fungi. The potential role of these endophytes is discussed. © 2006 SAAB. Published by Elsevier B.V. All rights reserved.


Introduction
One of the most important trees in Botswana is the Mopane (Colophospermum mopane (J.Kirk ex Benth.) J.Kirk ex J. Léonard: Fabaceae). The mopane tree provides fodder for both wild animals and livestock and is an important source of fire wood and shade. It is also the host of the 'mopane worm' or 'phane', the caterpillar form of the moth Imbrasia belina Westwood, which can occur on this host in high densities and forms an important part of the diet of the local people (Coates Palgrave, 2002).
Mopane is a leguminous tree that often forms monospecific stands and dominates the vegetation structure in the north east of Botswana, commonly on alluvial soils (Macala, 1996). It occurs in two forms (often inter mixed), either as a multistemmed shrub of about 1-2 m high, or as a small to medium-sized tree of 5-12 m which possesses an upright, narrow crown (Timberlake, 1997).
Little research has been undertaken to investigate the fungi associated with Mopane other than a study on mainly basidiomycetous pathogens (Smith and Shah-Smith, 1999) and mycorrhizae (Sinclair et al., 1998) and there are no published records of any fungi except for a Mycosphaerella sp. in Zimbabwe (Whiteside, 1966). Therefore, to initiate mycological studies, preliminary research was carried out to determine the diversity and occurrence of endophytic fungi associated with the leaves and pods of Mopane (Taylor et al., 2002).
The role of endophytes in non-graminaceous hosts is a much speculated topic and endophytic associations have been implicated to have a variety of effects on the host plant (Kriel et al., 2000;Wilson, 2000). Possible roles include mediation in plant-herbivore interactions (Hammon and Faeth, 1992), which maybe have a beneficial effect for the host (Miller, 1986;Carroll, 1991;Calhoun et al., 1992;Findlay et al., 1995;Azevedo et al., 2000;Miller et al., 2002), a negative effect (promoting a positive effect on the herbivores) (Gange, 1996;Johnson et al., 2003) or are apparently neutral (Lappalainen et al., 1999;Ahlholm et al., 2002). It has recently been acknowledged that endophytes could be an important source of bioactive compounds (Miller, 1986;Petrini et al., 1992;Strobel and Long, 1998;Tan and Zou, 2001;Schulz et al., 2002). Also endophytes have been suggested as potential biocontrol agents of plants (Dorworth and Callan, 1996), or alternatively as agents in promoting host resistance to disease (Kriel et al., 2000;Tomita, 2003).
Endophytic fungi are variously defined, as discussed by Stone and Petrini (1997), and those of non-graminaceous hosts can be divided into mutualistic symbionts or latent pathogens (Sinclair and Cerkaukas, 1996), depending on whether they are parasitic and can cause disease. However, due to the difficulty of determining the pathogenicity of a fungus occurring endophytically, a more encompassing definition is often used to include all fungi which colonise living plant tissues and which remain symptomless for the whole or only part of the period of colonisation (Stone et al., 2000).
The role of endophytes in Mopane is unknown. The present study was undertaken to provide quantitative and qualitative information on the endophytic fungi isolated from mopane pods, especially in relation to the ages of pods. The association of these fungi with the pods was also investigated using various forms of microscopy to elucidate and explain the abundance patterns seen in the aforementioned isolation studies. Finally, the lignocellulolytic abilities of the endophytic fungi were tested.

Collection site and details
The collection site was a stand of mopane trees and shrubs (C. mopane) adjacent to Madiela Health Post, which is situated on the A1 road, 45 km north of Mahalapye in south east Botswana. The pods were transported back to the University of Botswana in Gaborone in a cool box and were processed within 24 h of collection while being kept at 4°C. Collections were made according to pod maturity and were taken directly from randomly chosen trees and not from the ground unless otherwise stated. Green soft immature pods (estimated to be 1 month) were collected on January 31, 2003. On April 10, 2003, yellow pods (estimated to be 2 months) and brown pods (estimated to be 2.5 months) were collected and on October 14, 2003 brown pods (estimated to be 9 months) were collected. On October 19, 2002, brown pods (estimated to be 21 months) were hand-picked from the tree. Pods that were 2.5 months old refer to those that were harvested at the beginning of the reproductive season, while 9-month-old pods were harvested at the end of the growth season. Older pods (21 months) displayed visible signs of decomposition. The 21-month-old pods were placed into wire cages in contact with the ground and left for 2 months to determine whether there was a shift in fungal abundance and diversity between pods on the tree and those on the ground. These pods were then collected (estimated to be 23 months old) (December 4, 2002) as were pods occurring on the ground which had been there for an indefinite period (estimated to be 23 months old). The last three categories of pods, i.e. brown (21 months), ground-for 2 months (23 months) and groundindefinite period (23 months), did not contain viable seeds due to severe decomposition of the pericarp.

Isolation studies
Thirty-five pods of each age category were surface sterilised in a laminar flow cabinet as follows: The pods were first washed in sterile water for 1 min, after which they were washed in sterile water with approximately 2 ml of dishwashing detergent. The pods were then transferred to a beaker of 70% alcohol for 1 min followed by 8 min immersed in undiluted commercial bleach (3.5% sodium hypochlorite). Finally, the pods were rinsed in 70% alcohol for 30 s and were allowed to dry in the laminar flow cabinet before being dissected. Six pieces (5 × 5 mm) were excised from the following areas of the pod pericarp (not including the internal seed) and numbered as follows: the stalk (1), outer margin (2 and 3), inner margin (4) and pericarp wall (5 and 6). Together these pieces represented the entire pericarp. The pieces were placed around the outside of a Petri dish (labelled on the reverse with the appropriate numbers) containing malt extract agar (MEA, Biolab, MERCK, Gauteng South Africa) with a fungal growth retardant, rose bengal, at 0.033 g/l (Sigma-Aldrich Co, St Louis, USA), and supplemented with 0.2 g streptomycin sulphate (Sigma-Aldrich Co, St Louis, USA) dissolved in 2 ml sterile water. The plates were labelled 1-35, and the pod age indicated. The Petri dishes were incubated at 25°C in the dark and were checked after 5 days for growth. All isolates that developed were subcultured onto MEA for morphological identification. The percentage colonisation of pod pieces [(No. of pod pieces which yielded ≥ 1 isolates / Total no. of pod pieces) × 100] was calculated, as were isolation frequencies (No. of isolates yielded in total / Total no. of pod pieces) to demonstrate the degree of multiple colonisation from the samples taken.

Microscopy
Material used in the microscopy studies was derived as follows. After surface sterilisation and incubation, samples where only a single endophyte had developed were taken from culture and processed for microscopy. The developing endophytes were subcultured to MEA for positive identification. Other samples were derived from pods that were unsterilised and unincubated, or pods which had been sterilised and incubated, but where no fungi developed. The latter served as controls.
Pericarps of intact pods and pods in various stages of decomposition were sputter coated with gold and examined with a Philips XL30 ESEM. To test the efficacy of the surface sterilisation method, pericarp sections before and after sterilisation were compared. Pieces of pericarp were prepared for LM and TEM as follows: Samples were pre-fixed in 2.5% Glutaraldehyde in 0.1 M phosphate buffer and post-fixed in 1% OsO 4 . Samples were dehydrated in a graded ethanol series, embedded in LR White, sectioned and examined in a Zeiss Axioskop II light microscope or Tecnai 12 TEM. To confirm the presence of endophytes, sections of pericarp were surface sterilised and plated on DRBC medium. These cultured sections were also examined anatomically and with TEM.

Lignocellulolytic tests
Qualitative methods were used to assess the lignocellulolytic enzyme production of different strains of the fungi isolated ( Table 2). The protocols are outlined by Pointing (1999) and involved a cellulolytic enzyme assay using cellulose clearance agar and a lignolytic assay using Poly-R 478 (Sigma-Aldrich Co, St Louis, USA) agar. Positive reactions for lignocellulose  degrading enzymes involve clearance of the respective media around the growing fungal colony.

Isolation studies
Three fungal genera namely Phoma, Phomopsis and Alternaria occurred prominently. For the purpose of this study, the genera have not been separated into species, but the following observations were made. The Phoma isolates were similar to Phoma multirostrata (P.N. Mathur, S.K. Menon and Thirum.) Dorenb. and Boerema, but differed in that the conidial mass was deep pink. The cultures reached a diameter of approximately 60 mm in 7 days and were very variable, ranging from almost translucent to dark brown with sparse aerial mycelium, with some isolates producing denser grey aerial mycelium. Some colonies were observed to sector. In all cases, pycnidia were produced abundantly over the surface of the colony. No attempt was made to identify the Phomopsis isolates, but some sporulated and all displayed the morphology typical of Phomopsis cultures with woolly, dense yellowishwhite mycelium (55-70 mm in 7 days), and black stromatic structures produced all over the colony with some producing cream-coloured conidial masses after 28 days. The Alternaria isolates were considered possibly to resemble two species although there did appear to be a continuum in colony morphology from 'Alternaria sp. 1' which produced dense woolly colonies that were brownish-grey above and blackishbrown in reverse to 'Alternaria sp. 2' which produced dense woolly colonies that were greenish-grey above and brown below with varying amounts of pale pink. Most of the colonies grew at a rate of 60-80 mm in 7 days. The remaining fungal taxa were rarely isolated or remained sterile.
A total of 721 isolates were made from 1470 samples, and a summary of the results is given in Table 1. No isolates were obtained from 1-month green pods, but the number of isolates steadily increased as the pod matured up until the age of 9 months. The number of isolates appears to remain constant while the pods are attached to the tree, however, when the same pods are placed on the ground endophytic fungi decrease.
The colonisation rates indicated that many of the samples plated were not colonised (up to 70% of the pod age categories that yielded endophytes), and of the samples that yielded endophytes, many were colonised by more than one isolate. This was demonstrated by the higher isolation rates when compared with the colonisation rates. For the purpose of the microscopy studies, only samples that either yielded no endophytes, or only one species, were utilised. This was to ensure that anatomical and ultrastructural evidence of pod breakdown for each species could be studied separately. Of samples, which yielded only one species, the isolate was purified by isolation onto MEA and in no instances did more than one isolate develop. Fig. 1a-f illustrates the relative proportions of different fungal genera isolated from pods of different ages. A pattern of colonisation was evident. Initially, in younger pods collected from the tree (2-9 months), Phoma was dominant, whereas Alternaria gradually increased as the pod matured and Phomopsis remained at a constant level. However, after the pod had remained on the tree for 21 months, Phoma reduced dramatically whereas Alternaria became dominant. When the same pods were placed on the ground (23 months), the dominance of Alternaria decreased and with time, the numbers of miscellaneous taxa began to increase, with Phoma and Phomopsis much reduced.

Lignocellulolytic studies
The results of the lignocellulolytic assays are given in Table  2. It should be recognised that both tests are qualitative and only give an indication of the lignocellulolytic abilities of the different isolates, but do not provide quantitative information. In addition, the cellulose assay was very difficult to interpret and the cellulolytic abilities of the different isolates were therefore based not on clearance, but rather on growth rates as has been suggested previously (Abdel-Raheem and Shearer, 2002). For Alternaria strains, the fungi from the older pods demonstrated greater lignocellulolytic abilities, whereas the strains of Phoma and Phomopsis from the younger pods were more consistent in their lignocellulolytic abilities.

Microscopy
SEM studies revealed that the surface sterilisation technique used was highly effective. Yellow 2 months (Fig. 2a) and brown 2.5 months unsterilised pods (Fig. 2c) showed particles over their surfaces, but more in the case of brown pods as the pods were older. Both yellow (Fig. 2b) and brown pods (Fig. 2d) were clear of these particles after surface sterilisation.
Mature 9-months brown pods often had stomata occupied by fungal growth (Fig. 2e,f). It was not always apparent whether ? +/− − the fungi grew out of the pericarp through the substomatal chambers or whether the stomata formed a route of penetration for fungi from the outside. In most cases, it appeared that the penetrating hyphae originated from germinating spores on the pod surface that then penetrated the stomata (Fig. 2f). There were also instances of subcuticular penetration of the pericarp by hyphae originating from germinating spores in pods from all age classes (Fig. 2g). Subcuticular hyphae were also frequently observed (Fig. 2e,h). Most symptomless surface sterilised pericarps (i.e. those not displaying any outward signs of fungal colonisation) displayed profuse fungal growth after incubation and there was considerable histological and ultrastructural evidence for the breakdown of lignified walls. A few surface sterilised pods had no endophytes growing out of them and these served as controls. When these 2.5-months pods were examined anatomically it was clear that the surface sterilisation process did not damage the pods, as the overall structure was intact (Fig. 3a). Therefore, it was safe to conclude that fungi caused the breakdown of not only lignified, but also unlignified cell walls.
In the 9-months mature brown pods, the mesophyll was largely intact although some cells showed signs of degradation (Fig. 3b). Most mesophyll cells had thick, lignified walls. The fibres were heavily lignified.
In sections of unsterilised pods that had been on the ground for 23 months, the mesophyll and epidermal cells were almost totally decomposed (Fig. 3c). The only remaining part of the epidermal layer was the thick, cutinized wall. Fungal growth was abundant in mesophyll and epidermal regions (Fig. 3c). The upper layer of the heavily lignified fibres was also degraded and lignin has completely disappeared from the walls, while the inner cell layers also showed loss of lignin from the walls (Fig. 3c).
In 9-months pericarps that yielded Alternaria the mesophyll was completely broken down, including most of the lignified parenchyma cells (Fig. 3d). Most of the epidermal cells were also destroyed. Light micrographs showed that the lumen of fibres was heavily infested by Alternaria hyphae and that cell walls were digested by the enzymatic action of the fungus as was evident by the extensive white areas in the walls (Fig. 3e).
Mesophyll cell walls that had been broken down by Alternaria often appeared swollen (Fig. 3f).
Fine structural studies supported the evidence that Alternaria sp. had the ability to degrade the lignified walls of fibres in 9-month pod pericarps. Cell wall areas that were in close contact with hyphae displayed evidence of enzymatic degradation by the fungus (Fig. 4a-d). Some areas appeared to be etched by the action of lignocellulolytic enzymes secreted by the fungus (Fig.  4a), while other areas were completely broken down or in various stages of decomposition. Some cell walls only had a granular component remaining due to the action of lignocellulolytic enzymes (Fig. 4b). Hyphae of Alternaria were able to penetrate and grow right through the lignified walls as was evidenced by the numerous holes in the cell walls (Fig. 4c,d). Lignified mesophyll cell walls were almost completely degraded by Alternaria. The fine structure of the lignified walls showed that they were degraded to a much larger extent than the lignified walls of fibres. Only granular and fibrillar substances remained where the lignified component used to be and the cell walls were freely traversed by hyphae (Fig. 6a).
The other two isolated species; Phoma and Phomopsis also appeared to have lignin-degrading abilities, but to a lesser extent. Breakdown of lignified mesophyll cell walls were conspicuous in pericarp sections from pericarps of all 2-9month pods that only had Phoma associated with the tissues (Fig. 5a). As with Phomopsis, the cell walls had a typical "ghost" appearance in light micrographs, indicating the breakdown of some components of the walls (Fig. 5b). Unlike Alternaria and Phomopsis however, the distribution of Phoma was confined to the mesophyll layer (Fig. 5c). The fibre layer was intact and no evidence was found that this fungus had the ability to use this layer as a carbon source or to colonise the tissue (Fig. 5c). The fine structure of the lignified cell walls associated with Phoma also showed some evidence of enzymatic breakdown by the fungus. Areas in close contact with fungal hyphae appeared more electron-translucent, indicating dissolution of cell components due to enzymatic breakdown (Fig. 6b).
In the case of Phomopsis, isolated from the 9-month brown pods, this fungus also attacked lignified mesophyll cell walls (Fig. 5d) but to a lesser extent than Alternaria. In light micrographs, only a faint outline of where the lignified mesophyll cell wall used to be was often visible (Fig. 5e). This also resulted in "ghost" cell walls. Unlike Phoma, Phomopsis was able to colonise fibres, but in light micrographs no clear evidence of the breakdown of fibres was noted. There was also no clear ultrastructural evidence that this fungus had the ability to degrade the highly lignified walls of fibres even though hyphae were closely associated with them (Fig. 4e,f).
However, there was a dramatic distinction at the boundary between the fibres and mesophyll cells in that the fungus clearly broke down the walls of the latter (Fig. 4e).
Fine structural studies showed that during the initial stages of the breakdown of mesophyll cell walls, the inner layers first became granular and fibrillar, confirming some enzymatic action by Phomopsis (Fig. 6c). Later, the entire fine structure of the "ghost" cell walls seen in light micrographs appeared swollen, granular and traversed by fine electron dense lamellae (Fig. 6d). The lamellae probably represent cellulose microfibrils that have become more visible after the breakdown of the lignin component. These cell walls were always associated with conspicuous hyphae that were closely appressed to the cell wall (Fig. 6d).

Discussion
In this study, the definition of the fungi being isolated has been a problem. This is because the pod pericarp is not strictly a living tissue once the green pods turn yellow. The tissue may therefore be in the initial stages of senescence, which is part of the normal maturation process of pod development. After pod fill, while the pericarp tissues are still green and beginning to turn yellow, lignification of various tissues such as the epidermis and mesophyll occurs. This is associated with loss of cytoplasm from some of these cells, which may indicate that some of the pericarp tissues may be in the process of loosing their viability, but overall the pod is still alive/functional (Jordaan and Krüger, 2005). Only pericarp tissues of pods that have turned brown have lost their viability.
The pericarp is first colonised while the pods are becoming yellow (i.e. still alive/functional) (Jordaan, unpublished observations), therefore according to the definition of endophytes stated previously, the fungi are thus endophytes. Consequently it seems correct to refer to these fungi as endophytes, even though most of the study is concerned with the colonisation of essentially dead pericarp tissue. Furthermore, it must be taken into account that the green pods used in this study were still very young as they were sampled before pod fill. They, therefore, may not have been colonised via horizontal infection. It is only during the latter stages of pod fill that the green pods become lignified. Then the pod is most susceptible to colonisation because lignification of the epidermis causes the guard cells to lose their ability to regulate opening and closing of stomata, resulting in stomata staying permanently open (Jordaan and Krüger, 2005). As was demonstrated in the SEM micrographs, the open stomata appear to be an easy route of penetration of fungi into the pericarp. A study undertaken of endophytes from different aged leaves (young, mature senescent and dead) and pods (Taylor et al., 2002) revealed that the same fungi were found to colonise living leaf tissues as were isolated from the various aged pods in the present study.
Endophytes have been shown to produce the enzymes necessary to infect and colonise host plants (Carroll and Petrini, 1983;Sieber-Canavesi and Sieber, 1993;Urairuj et al., 2003), and to become established to varying degrees. In the above ground tissues of some hosts, endophyte infections are restricted (Stone, 1987(Stone, , 1988Stone et al., 1994;Stone and Petrini, 1997), whereas other studies have demonstrated more extensive colonisation (Deckert et al., 2001). In plant pathology literature, there are similar accounts of pathogens invading plant tissues, and then remaining latent (Sinclair and Cerkaukas, 1996), but when these pathogens become active they develop symptoms of disease, thus distinguishing them as latent pathogens and not purely mutualistic endophytes (Sinclair and Cerkaukas, 1996).
There are a number of reports of endophytes being capable of litter degradation of their hosts (Carroll and Petrini, 1983), that is of endophytes becoming active after senescence and continuing even after leaf fall (Wildman and Parkinson, 1979;Cabral, 1985;Sieber-Canavesi and Sieber, 1993;Hata et al., 1998;Kumaresan and Suryanarayanan, 2002;Osono, 2002). Recently, several studies have directly investigated the decomposing abilities of the endophytes at different stages of the succession using enzyme assays and decomposition tests (Osono and Takeda, 1999;Muller et al., 2001;Urairuj et al., 2003).
The aforementioned studies bear some relevance to the research carried out in the present study. In terms of diversity and abundance, over time the major Mopane fungal taxa seem to show a succession demonstrated by a relative change in abundance. Also, the numbers of isolates increase until the 9month brown-pod stage and they decreases, probably due to changes in fungal abundance possibly resulting from diminishing resources. Similar studies have directly investigated succession trends of endophytes in relation to the relative abilities of the different fungi to decompose plant tissues, and their ultimate role in litter degradation. Litter degradation may be a very important, but overlooked, ecological role of endophytes.
The differing abilities of endophytes to utilise various host substrate components and the associated changes in fungal succession and colonisation of tissues have been clearly demonstrated in previous studies (Osono, 2002;Osono andTakeda, 2001, 2002;Kumaresan and Suryanarayanan, 2002;Sieber-Canavesi and Sieber, 1993). For instance Osono and Takeda (2002) demonstrated that changes in fungal succession were related to available resources in the leaves (lignin and carbohydrates) and fungal succession and colonisation were a result of the different abilities of the fungi to utilise the various substrates.
The above-mentioned studies (Osono, 2002;Osono andTakeda, 2001, 2002;Kumaresan and Suryanarayanan, 2002;Sieber-Canavesi and Sieber, 1993) suggested that certain endophytic fungi may have a greater role in litter degradation than previously acknowledged. It was speculated that the lignocellulolytic abilities of the endophytic fungi would be necessary to enable them to colonise the interior of the leaf and that the fungi would maintain a presence in senescent and dead leaves, and be in a position to exploit the various carbon compounds, including recalcitrant lignocellulose resources, and thus play a part in litter decomposition (Osono and Takeda, 1999). In other substrate utilisation studies it has been demonstrated that endophytes which colonise different tissue types have varying lignocellulolytic abilities and that the patterns of substrate utilisation may enable partitioning of the substrate resources (Carroll and Petrini, 1983).
In the present study, tissue specialisation was recorded and substrate utilisation patterns appear to be dependent on the availability of the constituents of the tissues over time. The ability of the endophytes to penetrate the Mopane pod pericarp and infect tissues was directly demonstrated by microscopy studies and indirectly by lignocellulolytic assays. The endophytes are considered to result from horizontal infection, as none was isolated from very young green pods. Thus, during the earliest stages of pod development they are entirely absent. However, these fungi have been isolated from mature green pods (Jordaan, unpublished observations) and in the present study many were isolated from yellow (2 months) and brown pods (2.5 months) and they continue to increase in abundance until the pods mature. However while they remain on the tree, the relative abundance of the different fungi changes over time.
The extent of breakdown and colonisation of lignified tissues were distinct to each of the three dominant fungal species isolated. The presence of Alternaria sp. in incubated pericarp sections was always associated with the breakdown of both lignified mesophyll cell walls as well as heavily lignified fibres, suggesting an ability to secrete lignocellulolytic enzymes and to degrade lignin. The data are supported by the lignocellulolytic tests. Lignified mesophyll cells appeared to be less resistant to enzymatic breakdown as they were in a far more advanced state of degradation than the walls of fibres.
The TEM results in this study are significant in terms of demonstrating apparent ultrastructural degradation of lignin cell walls associated with hyphae, of all three dominant isolates. Lignin is a strong cell wall polymer that is generally considered as being resistant to most fungi (Brett and Waldron, 1990). In addition, the lignified mesophyll cell walls appear to be less resistant to breakdown than the lignified walls of the fibres as only Alternaria had the ability to break down the latter. This may be due to differences in composition, structural linkages and whether the lignin is of the condensed or the uncondensed type (Lewis and Paice, 1989). Lignin is water impermeable and insoluble in water (Fry, 2004). Mesophyll cell walls that had been broken down by Alternaria often appeared swollen, indicating that the lignin component had been removed. Therefore, the remaining wall components that often appear are sponge-like and swollen due to the uptake of water.
The major route whereby endophytes gain entry into Mopane pods may be through the stomata or cuticle. In the latter case, subcuticular hyphae were sometimes seen. Rhabdocline parkeri Sherwood, J.K. Stone and G.C. Carroll, an endophyte of Douglas fir (Pseudotsuga menziesii (Mirbel) Franco) also enters the host directly through the cuticle and cell wall (Stone, 1987). Subcuticular penetration was also seen during the infection of Fagus sylvatica L. leaves by the endophyte Discula umbrinella (Berk. and Broome) M. Morelet . Unlike young green mopane pods, the epidermis, including the guard cells of mature brown pods are highly lignified (Jordaan and Krüger, 2005), resulting in the stomata being permanently open, thereby serving as a route for fungal infection.
The changes in the abundance and diversity of the endophytes are therefore apparently related to the age and degree of decay of the pods, and are determined by the lignocellulolytic abilities of the fungi. It would appear that all three endophytes show lignocellulolytic activities. Alternaria demonstrates the greatest ability as it breaks down more heavily lignified tissue than Phoma and Phomopsis, although the lignocellulolytic assays were inconclusive being qualitative, but at least positively demonstrated the lignocellulolytic abilities of the fungi. A possible scenario for the action of the fungi is as follows: Phoma and Phomopsis initially dominate but these fungi possibly decrease because the substrate that they are able to degrade is already fully utilised and they do not have the abilities to degrade the more heavily lignified tissues. Alternaria on the other hand can utilise heavily lignified tissues, thus it persists and increases over time. Miscellaneous soil fungi may eventually replace all the 'endophytes'.
One possible advantageous role of endophytes in mopane is that Phoma and Phomopsis may act as early colonisers that degrade the pericarp, thus allowing the pericarp to become water permeable. This allows effective germination of seeds in an arid environment when conditions are favourable. As colonisation continues (i.e. by Alternaria), the protective pericarp becomes much more deteriorated and the seed loses viability. This is reflected by the short-term reproductive strategy of Mopane, in that it produces large quantities of viable but not particularly robust seeds with large cotyledons. Rapid recycling of seeds that do not germinate would add to nutrients available for the following year's growth. In contrast to this observation, Wilson (1993) suggested that endophytes may have actually caused the evolution of certain plant defences such as hard-shelled seeds, which may have evolved as a response to, among other things, attack by fungi.
C. mopane produces many large, one-seeded pods but there is no permanent seed bank. Therefore, effective decomposition of pods and non-germinated seeds and a resultant recycling of nutrients may be of prime importance to this species and other species that benefit from the resources released. As the pericarps of pods consist of highly lignified tissue, endophytes that display lignocellulolytic abilities may significantly contribute to the decay of pods.