Why do some fungi give up their freedom and become obligate dependants on their host?



Fungi have repeatedly acquired an ability to infect plants (Berbee, 2001). Many of these infections result in disease, others are of benefit to the host, and some do not appear to have any measurable effect. These fungi are called pathogens, mutualists or ‘endophytes’, respectively. Some of the plant pathogenic fungi, and all of the mutualists and endophytes, are capable of infecting their hosts without causing death – they are known as biotrophs (Schulze-Lefert & Panstruga, 2003). The degree of compatibility achieved by these biotrophs is at times truly remarkable: they are capable of establishing very close contact with plant cells, at times even breaching the cell walls and producing extensive ‘intracellular’ structures which are thought to facilitate the uptake of nutrients from their host (Panstruga, 2003). This is a testimony to the exquisite ability of these microbes to avoid and, perhaps, suppress the normal defence mechanisms that plants are known to possess against microbial invaders (Voegele & Mendgen, 2003). How this is achieved remains almost a complete mystery – and is possibly one of the most exciting frontiers in our understanding of plant pathogen interactions today (Ellis, 2006). Another phenomenon that is not understood is why and how some of these highly compatible biotrophic fungi become ‘obligate’, that is they lose the ability to grow and complete their life cycle on any substrate other than their living plant host. This is true for both the ancient interactions with very broad specificity, such as the arbuscular mycorrhizas (Brundrett, 2002), as well as for the more recently evolved ones exhibiting higher host specificity, for example the cereal rusts (Staples, 2000), powdery mildews (Both & Spanu, 2004) and the Oomycete downy mildews (Grenville-Briggs & West, 2005).

What is the physiological basis for obligate biotrophy?

A prevailing and long-standing hypothesis states that obligate symbionts have lost their ability to synthesise the essential organic compounds they require for growth, and they have an absolute requirement for a host which provides these nutrients. We could call this the ‘auxotrophy principle’. This deficiency might be due to a loss of individual genes or of whole pathways for the synthesis and metabolism of such compounds, as has been documented in bacterial parasites (Wernegreen, 2005; Dagan, 2006). In these organisms it is not possible to supplement the growth medium by simply adding one substance because a whole gamut of compounds is required. Recently, an opportunistic human pathogenic fungus has been shown to have developed auxotrophy associated with a change in host range spectrum (Scully & Bidochka, 2006; Scully & Bidochka, 2005). In the case of obligate biotrophic fungi infecting plants there is as yet no clear evidence of dramatic loss of biosynthetic and metabolic ability: the existing EST sequencing projects have not identified obviously absent pathways in the libraries analysed (Thomas et al., 2001; Soanes et al., 2002; Giles et al., 2003). A definite statement to this effect will have to await the completion of the genome sequencing programmes of obligate biotrophic mutualists and pathogens (for example Glomus intraradices, Melampsora larici-populina, Puccinia graminis, Blumeria graminis or Hyaloperonospora parasitica), but at present a wholesale genome reduction or gene degradation seems unlikely in these organisms.

‘It could be argued that the development of feeding structures is not necessary per se for survival: after all, the majority of fungi survive very well without them; in fact it might be that the essential factor is an appropriate regulation of the metabolic toolkit of the biotrophs required to deal with nutrient uptake and processing’

An alternative hypothesis is that in obligate biotrophs the regulation of metabolic gene expression is dependent on signals and signal gradients that are only found in host plants. That is, the host provides cues which are used by the endosymbionts to control changes in gene expression supporting the appropriate levels of metabolic capacity.

It is widely accepted that the development of specific infection structures such as appressoria and haustoria (the specialist penetration and feeding structures, respectively) is triggered by the perception of plant-derived signals (Hamer & Talbot, 1998; Tucker & Talbot, 2001). In some cases these signals are known and at least in part understood. For example surface topography is recognised by rusts and induces the oriented growth of hyphae (Staples, 2000); surface hydrophobicity stimulates the formation of appressoria in Magnaporthe (Talbot et al., 1996; Hamer & Talbot, 1998); cutin degradation products increase the development of appressoria in Blumeria (Both & Spanu, 2004). Note that there is no direct evidence for an agent that induces the development of specialised feeding structures such as haustoria or arbuscules, but the formation of these structures might well be reliant on essential plant-derived stimuli.

It could be argued that the development of feeding structures is not necessary per se for survival: after all, the majority of fungi survive very well without them; in fact it might be that the essential factor is an appropriate regulation of the metabolic ‘toolkit’ of the biotrophs required to deal with nutrient uptake and processing. In other words, to survive, the developing fungus must express the transporters, pumps and channels at the right time, the correct location and at the right levels. It must also activate and regulate the enzymes that catalyse catabolic and anabolic reactions to deal with nutritional and growth requirements over a sustained period in order to complete its life cycle.

At present there is some direct evidence to support this alternative hypothesis. We have recently observed that the expression of metabolic genes in barley powdery mildew is often co-ordinately regulated during both development and pathogenic attack (Both et al., 2005), and important metabolic and uptake functions are expressed in rust haustoria (Sohn et al., 2000; Voegele et al., 2001; Jakupovic et al., 2006). Arbuscular mycorrhizal fungi express fatty acid synthase activity exclusively in the intra-radical mycelium (Trepanier et al., 2005), and genes involved in the metabolism of nitrogenous compounds are differentially expressed between intra-radical and extra-radical mycelia (Govindarajulu et al., 2005). This hypothesis can be tested further: if true, then at least some of the coordination of expression should break down when the pathogen is made to germinate in vitro, where partial growth of the fungi is observed. Another test will be to see whether there is evidence of deficiencies and mutations in the regulatory elements (e.g. promoters) that control metabolic genes; a systematic analysis will be possible once the whole genomes of some obligate biotrophs are sequenced and compared to the genomes of related non-obligate related fungi. In this context it is interesting to note that the rapid evolution of gene regulatory elements resulting in altered gene expression may be the basis for evolutionary radiation (Gilad et al., 2006).

Evolution of an obligate relationship

Why would a parasitic or even a mutualistic symbiont ‘want’ to lose its ability to grow as a free living organism when this means becoming totally dependant on a host for survival? The answer possibly lies in the fact that evolution is not forward-looking but the result of a balance of fitness at any given time. In this conceptual frame, it is therefore necessary to postulate that under certain circumstances, retaining the ability to grow and reproduce as a free-living facultative symbiont had a greater cost than developing obligate host specialisation. Thus when the interaction reached the point that free life was redundant, it was dispensed with. In a way this may be similar to the situation seen in the evolution of host specificity: maintenance of a wide host range has a cost that penalises its maintenance if a more specialised existence may be cheaper in the ‘here and now’ (Chin & Wolfe, 1984).

Are there any observations that might support this notion? I am not aware of any incidence in which a fungus, be it pathogenic or mutualistic symbiont, was actually seen to lose its ability to grow in axenic culture. Conversely, many pathogenic fungi that are facultative saprotrophs need to be routinely ‘passaged’ through a host to retain vigorous virulence. Failure to do so often results in decreased virulence.

Taken together, these observations and the existence of obligate biotrophy, point to the costs of maintaining both a biotrophic and a free-living lifestyle that might lead to diversifying radiation: maintenance of a facultative symbiotic status can be sustained only if there is selective pressure balancing the costs. It remains to be seen if experimental evidence can be obtained to refute or corroborate these ideas.


I would like to thank Kim Hammond-Kosack and Paola Bonfante for a critical discussion of the issues raised in this article.