Portal de Periódicos da CAPES

Ectomycorrhizas – out of Africa?

2006; Wiley; Volume: 172; Issue: 4 Linguagem: Inglês

10.1111/j.1469-8137.2006.01930.x

1469-8137

Ian J. Alexander,

Ecology and Vegetation Dynamics Studies

Ectomycorrhizas are the most frequent and widespread mycorrhizal type in the forests and woodlands of cool-temperate and boreal latitudes. Tree species from all the major plant families that provide the dominant trees of these regions (e.g. Pinaceae, Fagaceae, Betulaceae, Salicaceae) habitually form ectomycorrhizas under natural conditions. The ectomycorrhizal (ECM) habit shows particular adaptations for nutrient capture in temperate and boreal forests (Read & Perez-Moreno, 2003). On soils of low pH, where litter decomposition, for reasons of climate or litter quality, is slow, ectomycorrhizas are produced abundantly in soil-surface organic layers, and ECM fungi mobilize nitrogen and phosphorus from organic residues. There are strong links between plant mycorrhizal status and ecosystem carbon dynamics, because the foliage of temperate and boreal ECM tree species has significantly lower mean litter decomposition rates than that of comparable arbuscular mycorrhizal species (Cornelissen et al., 2001). For these reasons, and also (one suspects) because the overwhelming majority of active researchers live and work in temperate regions, ectomycorrhizas have long been thought of as primarily a feature of temperate and boreal forests. However, it has been known for at least 50 yr (Peyronel & Fassi, 1956) that ECM trees also occur in tropical forest. It is true that most tropical tree taxa form arbuscular mycorrhizas (Alexander, 1989a), but an ecologically important minority of taxa form ectomycorrhizas. This includes, inter alia, the Dipterocarpaceae (Lee, 1998) and Fagaceae (Corner, 1972), many legumes in Caesalpinioideae (Alexander, 1989b), and Myrtaceae in the subfamily Leptospermoideae (Moyersoen et al., 2001). A glance at this list of plant families is, in itself, enough to emphasize the importance of ectomycorrhizas in tropical ecosystems. Dipterocarps extend from East Africa and Madagascar, through India, Bangladesh and Sri Lanka, to Southeast Asia, ranging from south China in the north to Papua New Guinea in the south. There are two genera in South America. Dipterocarps make up 80% of the canopy trees and up to 40% of the understorey in Southeast Asian lowland and montane rainforest, and dominate the dry monsoonal forests of north India, Burma and Thailand. ECM legumes form groves or extensive monodominant stands in the rainforests of the Guineo–Congolian basin (Newbery et al., 1988), while others cover vast areas in the Zambezian miombo woodlands of East and South–Central Africa, and the Sudanian savannah woodlands of the sub-Sahara. One of this group of ECM legumes (Intsia) extends into the dipterocarp forests of Southeast Asia, while another (Dicymbe) forms monodominant stands on poor soils throughout the Guyana shield of South America. The ecology and selective advantage of the ECM habit for host trees in the tropics is less well researched and well understood than in temperate and boreal forest. However, ECM trees in tropical forests occur in 'monodominant' stands, or stands where they make up a high percentage of basal area, or they attain 'family' dominance over very large areas. They appear to function in concert with one or more of a suite of other characters (large seeds, shade tolerance, mast fruiting, litter quality) to promote the success of their hosts (Alexander & Lee, 2005). 'Such a date places the evolution of angiosperm ectomycorrhizas close to the origin and early radiation of angiosperms, and well before the oldest fossils attributable to angiosperm lineages that nowadays form ectomycorrhizas.' Did the ECM habit arise in boreal and temperate forest, where ectomycorrhizas are most widespread today and where their selective advantage is clear, only later moving into tropical latitudes? Or might it have arisen independently, conceivably originally, in the tropics? Modern plant phylogenies show that the ECM habit has arisen independently over the course of evolution in Pinaceae and several disparate lineages of angiosperms (Fitter & Moyersoen, 1996; Bruns & Shefferson, 2004). Ectomycorrhizas are clearly 'younger' than the ancient arbuscular mycorrhizal symbiosis, which has been around since the earliest stages of land-plant evolution (Remy et al., 1994). The oldest known fossil ectomycorrhizas date from 50 million yr ago (mya) (LePage et al., 1997), but it seems certain that the symbiosis predates this by some time, because the Pinaceae and many of the angiosperm families whose members now form ectomycorrhizas were extant well before 50 mya (Wing & Boucher, 1998; Wang et al., 2000), along with the major fungal lineages with modern ECM representatives (Berbee & Taylor, 2001). The discovery (Ducousso et al., 2004) that the Sarcolaenaceae (an endemic family from Madagascar, sharing a common ancestor with Dipterocarpaceae) are ECM has been seen as evidence that ectomycorrhizas evolved in that lineage before dipterocarps drifted away from Madagascar/East Africa with the India/Seychelles land mass around 88 mya. In this issue, Moyersoen (pp. 753–762) argues that the evolution of ectomycorrhizas can be pushed even further back, to before 135 mya. This argument is based on his confirmation that the South American dipterocarp Pakaraimaea dipterocarpacea, endemic to Guayana and basal in the dipterocarp clade, is ECM. Because dipterocarps originated in Gondwana (Dayanandan et al., 1999; Ducousso et al., 2004), Moyersoen reasons that the ancestors of Dipterocarpaceae must have already had the capacity to form ectomycorrhizas prior to the separation of South America from Africa by the Atlantic around 135 mya. Such a date places the evolution of angiosperm ectomycorrhizas close to the origin and early radiation of angiosperms, and well before the oldest fossils attributable to angiosperm lineages that nowadays form ectomycorrhizas (Willis & McElwain, 2002). Nevertheless, Moyersoen's hypothesis is very attractive and, as he points out, it gains support from an apparently parallel situation in the tribe Amherstieae of Caesalpiniaceae. This is a primarily African clade whose members have long been known to be ECM (Alexander, 1989b), and where the isolated South American genus Dicymbe has also recently been shown to form ectomycorrhizas (Henkel et al., 2002). Assuming a common Gondwanaland ancestor for African and South American Amherstieae, a similar argument can be made for that ancestor to have been ECM prior to the continental break-up. One might also point out that the gymnosperm Gnetum, whose pantropical distribution could reflect a Gondwanan origin (Markgraf, 1929), is also ECM (Fassi, 1957). First, one could argue that the ancestors of the ECM dipterocarps and legumes were not themselves ECM prior to the break-up of Gondwanaland; they merely carried in their genome the information that would allow their descendents independently to become ECM at some later time on either side of the Atlantic. Second, an increasing number of dated molecular phylogenies suggest that long-distance transoceanic dispersal may be more important in explaining angiosperm disjunctions, such as those between Africa and South America, than previously supposed, and that this can obscure any effect of tectonic history, previously assumed to have been the major cause of their biogeographical patterns (Pennington et al., 2004). If this is the case, then the date of 135 mya for the evolution of ectomycorrhizas becomes suspect. With particular reference to the taxa mentioned here, Wikström et al. (2001) obtained a date for the origin of the Dipterocarpaceae, between 14 and 28 mya, precluding continental drift as an explanation for the observed distribution, although this date has been disputed (Givnish & Renner, 2004). Lavin et al. (2004) concluded that the age of most legume clades, including those with present-day distributions on either side of the Atlantic, precluded a large influence of continental vicariance on distribution patterns, and that long-distance dispersal was likely to be the primary explanation. Similarly, molecular data suggest that Gnetum probably reached its disjunct pantropical range either through transoceanic dispersal or through a Laurasian expansion followed by southwards spread well after the break-up of Gondwanaland (Won & Renner, 2006). So, while Moyersoen's reasoning is sound, his conclusion about the date of ECM evolution should be treated with caution until the debate over the role of vicariance vs long-distance dispersal in these key host taxa is resolved. What is missing from the above discussion is any consideration of fungi. This is often the case in mycorrhizal matters, where a decidedly phytocentric approach prevails, but it is good to remember that the evolution of ectomycorrhizas is likely to have been at least as important for the fungi involved as for the hosts. Just as in plants, the ECM symbiosis has evolved independently among multiple lineages of fungi (Bruns & Shefferson, 2004). The major radiations of the ECM basidiomycete lineages took place certainly by the Eocene–Oligocene (Bruns et al., 1998) and probably earlier in the Cretaceous (Berbee & Taylor, 2001). As Moyersoen points out, if ectomycorrhizas were present on the ancestors of dipterocarps and caesalpinioid legumes prior to the rifting of the South Atlantic, then one would expect that comigration of symbionts would result in at least some of their modern-day fungal associates showing similar disjunct distributions. It may be significant that ectomycorrhizas formed by Sebacinales were recorded both by Moyersoen on Pakaraimaea, and by Henkel et al. (2004) on Dicymbe. Sebacinales are the most basal group of basidiomycetes with known ECM members (Weiss et al., 2004), and a dated phylogeny of this group – including material from South America, Africa and Asia – could be very informative. The Russulaceae are another group that would repay attention. Although Moyersoen, in his necessarily restricted sampling, did not record Russulaceae in association with Pakaraimaea, Henkel et al. (2002) found Russulaceae in association with Dicymbe, and they show high abundance and diversity as ECM associates of dipterocarps and legumes in both Africa and Asia (Buyck et al., 1996; Lee et al., 2003). Many years ago, Pirozynski (1983) suggested that Russulaceae were warm-adapted ECM fungi that had their origins in Africa, and recent work seems to support this view (Buyck, 1997). Here again, if a dated phylogeny were possible, it might reveal whether South American Russulaceae have a Gondwanan origin, and shed further light on the question of where and when the ECM habit first arose. I thank Ursula Eberhardt for useful discussions.