The Rust Fungi: Extreme Specialists with a Complex Life Cycle

These remarkable plant parasites (>7,000 species in the monophyletic order Pucciniales, Basidiomycota) illustrate nicely some general attributes of the CLC. As a group they are among the most highly evolved fungi and may have undergone as many as 300 million years of coevolution with their hosts since emerging from a common ancestor. From the inception of rust evolutionary studies in the early 1900s, this storied ancestor has broadly assumed to have been a parasite of tropical ferns (very early vascular plants, member of pteridophytes) in the Carboniferous Period of the Paleozoic era (e.g., Savile 1955; Leppik 1959).

From these hosts arguably the rusts diverged through geological time and changing climate to become parasites successively of gymnosperm trees (ancient and then more recent conifers in the Abietaceae, Cuppressaceae, and Taxaceae), then the arborescent angi- osperms, and ultimately the modern, herbaceous angiosperms (Leppik 1959). Many if not most of the major host groups likely served as ‘gene centers’ or stepping stones in the further diversification of the rust clade on successor host taxa (Leppik 1959).

In the case of the extant heteroecious grass rusts (discussed later), the herbaceous hosts originally were members of the Berberidaceae (barberry family), and ultimately that rust group extended to the sedges and grasses worldwide. Rust evolution as recounted in the many papers of Leppik and Savile was construed as being one of continual adaptation to phylogenetically younger hosts (parallel cladogenesis or co-speciation), with only occasional jumps geographically associated but taxonomically distinct hosts. Alternatively, however, it may have been primarily the latter (Hart 1988).

Significantly, the rusts appear to have been derived exclusively from parasitic ancestors (i.e., not secondarily from free-living forms as is believed to be the case for many parasitic animals), though this cannot be stated with assurance. Parasitism appears at the base of the Basidiomycota, including the subphylum Pucciniomycotina (which contains the rusts), although the mycorrhizal basidiomycetes in the sister subphylum Agaromycotina have s everal independent origins from saprophytic ancestors (James et al. 2006; Kohler et al. 2015). Unlike most fungi, the rusts do not grow saprophytically in nature and thus are intimately coevolved with the living host. This obligately biotrophic nature has set the stage for specialization and the most complex life cycle among fungi and probably in the entire living world.

Let us now turn to the details of why this is the case for the rusts. During the course of their life cycle, the rusts as a group are distinctive in exhibiting up to five morphological stages (and correspondingly diverse nuclear, structural, and ecological states), each typified by a characteristic spore form (OFig. 6.4b). Moreover, these stages occur in a definite sequence and typically on two taxonomically unrelated plant hosts. In the basic life cycle a complete sequence of spore types—which characterizes a so-called long-cycled or macrocy- clic rust—consists of teliospores, basidiospores, spermatia, aeciospores, and urediniospores (older terminology varies, see Chap. 22 in Webster and Weber 2007). So, unlike the complex life cycle of the frog, rust multiplication occurs at many points in its cycle. Apart from the basidiospores, which are borne naked on modified basidia extending from the germinating teliospores, all spore stages are produced within specialized structures (sori) called, respectively, spermogonia, aecia, uredinia, and telia.

This basic macrocyclic life cycle is modified (abbreviated) in various ways. For example, species lacking the uredinial stage are termed demicyclic; those lacking aecia and uredinia are called microcyclic or short-cycled. Two significant consequences of a microcyclic life cycle are that many such rusts do not undergo normal sexual recombination (spermogonia are absent) and all such rusts are autoecious (basidiospores reinfect the same host species). In contrast, there is a conventional sexual cycle in macrocyclic rusts and these parasites can be autoecious or heteroecious (requiring two dissimilar hosts to complete the life cycle) (Terminology varies here as well and some authors consider microcyclic rusts to produce only teliospores and basidiospores; e.g., Ono 2002; Agrios 2005).

Puccinia graminis, the black stem rust fungus, is distributed worldwide and attacks hundreds of species of cereals and grasses in at least 54 genera (Leonard and Szabo 2005). Like almost all the other rusts, this species consists of highly specialized forms each attacking certain host genera and frequently only specific cultivars (at which level they are specific enough to be termed races; Chap. 2). For example, a specialized form, P graminis f. sp. tritici, attacking wheat causes devastating epidemics attributable in part to the rapid evolution of races virulent only to specific lines of wheat (see gene-for-gene interactions in Chap. 3 and Singh et al. 2011). Not surprisingly, this pathogen is the focal point of rust investigations and is one of the most intensively studied of all organisms.

The fungus is a macrocyclic, heteroecious species that forms uredinia and telia on the Gramineae and alternate stages only on various native Berberis (barberry) or rarely on Mahonia hosts. While most of the contemporary attention is on the graminaceous hosts because of their economic importance, in an evolutionary context alluded to above, barberry is the primary host and grasses the secondary hosts. The stem rust life cycle contrasts with other cycles represented in the Puccinia genus. For instance, P. asparagi (asparagus rust) is a macrocyclic species but autoecious, while P malvacearum (hollyhock rust) is microcyclic. For P graminis, or any other heteroecious rust to complete its life cycle, geographic overlap of the alternate hosts (or effective dispersal of appropriate infective forms) is obviously a necessity. Details of the life cycles of various rusts can be found in Bushnell and Roelfs (1984), Roelfs and Bushnell (1985), and Webster and Weber (2007). Here we focus on the evolution of complexity or simplicity in the life cycle.

The early dogma is confused as to the respective origins of macrocyclic, typically heteroecious, rusts and the microcyclic, autoecious, category (Jackson 1931; Leppik 1959, 1961). Hart’s cladistics analysis (1988) shows that heteroecism is well nested within his cladogram, i.e., heteroecism is a relatively recent, derived condition that evolved several times. There is no evidence that it is the primordial state. Based on the 30 rust genera examined, Hart concluded that several short cycle, mainly tropical and warm climate rusts are the most basal (i.e., ‘primitive’) as indicated below in footnote 2. This contradicts much of the early work cited above that argued the heteroecious long cycle is primitive, based largely on the faulty premise that since such cycles occur on supposedly ancient hosts they must also be the incipient form. (Of course, until Hart’s cladistical treatment, the same ‘logic’ was used by others to argue the opposite case, that the microcycle was primitive.)

Heteroecism, perhaps having arisen fortuitously fairly early in rust evolution if not as the primordial state, was maintained arguably because susceptible tissue was scarce when basidi- ospores were being discharged. Inspection of the macrocycle suggests that heteroecism may have arisen from mating between rusts individually pathogenic to one of the two unrelated host species, with the resulting dikaryon able to infect both (Chap. 2 and Buss 1987, p. 158). Whatever its origin, the advent of host alternation must have carried considerable benefit because the increased distance of travel between two hosts involves higher costs in population mortality than travel between tissues of the same plant. (The practical implication of host alternation is that frequently an effective disease control strategy for heteroecious rusts is to eradicate the alternate host.) Heteroecism presumably fostered evolution of uredinia. The incipient uredinium would have provided for multicycles of urediniospores to offset population mortality in travel to a new host. Evolution of spermogonia and aecia would have facilitated outcrossing, nuclear transfer, and dikaryotization in a much more efficient manner than the former process of hyphal fusion from germinating basidiospores.

Just as environmental conditions perhaps selected for heteroecism and the associated lifecycle stages, so likely did the environment set the stage for evolution of secondary autoecism and the microcyclic rusts (Jackson 1931; Savile 1976; Anikster and Wahl 1979). How this occurred also has been actively debated, but there may have been an immediate conversion to the microcycle on the aecial host, where telia take the place of aecia on the alternate host of the ancestral rust. In fact, this pattern exists so consistently in nature that it has become known as Tranzschel’s Law after the Russian mycologist who formalized the generalization. He inferred that when the alternate host for the aecial stage of a suspected heteroecious rust is unknown it should be sought on plant species infected by microcyclic rusts with morphologically similar teliospores to the heteroecious rust (Shattock and Preece 2000; Webster and Weber 2007, Chap. 22). Of course, the search will be fruitless if the alternate host is extinct. And, this raises the important caveat of Hart (1988): while many rust researchers have doggedly assumed that the absence of stages is explicable simply by ignorance of the ‘missing’ host, in reality these taxa may be primitively short cycle.

Whatever the contested primacy of the microcycle versus macrocycle in geological time, the adaptive benefit of a microcycle is strikingly apparent among extant species in arctic, alpine, or desert areas. Here the growing season for the host and hence the parasite may be only a few weeks (Savile 1953, 1971a,b, 1976). Convergent similarities among rust species include not only autoecism and short-cycling, but also self-fertility and suppression of sper- mogonia. Savile (1953, 1976) contrasts the life cycles of heteroecious rusts along a latitudinal gradient from temperate zones to the arctic. In the former, most of the species have a macrocycle similar to that described above for the temperate Puccinia.

Numerous uredinial generations occur and the rusts can persist with the alternate hosts relatively far (many meters or more) apart. As one moves northward, time available for the repeating stage decreases progressively; near the tree line the identities of the plant species involved become obvious because the rust cannot survive unless the alternate hosts are within about 50 cm of each other. Just inside the tree line the plants must be contiguous. Beyond the tree line the only heteroecious rusts are those that can abbreviate the life cycle, for example, by self-fertility or by producing dispersible (diasporic) rather than sessile teliospores, or by ‘double-tracking’ (Savile 1953, 1976) the life cycle: Chrysomyxa, for example, overwinters as a dikaryotic perennating mycelium rather than as a telium in the leaves of its evergreen host, Ledum.

In the spring this mycelium produces uredinia and telia nearly simultaneously. These sori mature to produce their respective spore forms. The urediniospores infect Ledum; the teliospores germinate in situ to produce air-borne basidiospores that infect the alternate host Picea. The life cycle thus is abbreviated because gene recombination (teliospore/basidiospore phase) and dispersal/multiplication (urediniospore phase) proceed concurrently. (For an analogous discussion of environmental constraints as they influence amphibian life cycles, see Wilbur and Collins [1973]. It is notable that the life cycles in closely related species of sea urchins and relatives may run the full gamut of stages or involve loss of larval stages and accelerated development [Raff 1987]. Moran and Whitham [1988] discuss analogous life cycle reduction in the aphid Pemphigus.)

So, what do the welter of details and the diversity of life cycles among the rusts tell us about evolution of CLCs? First, by their adaptive adjustments in the cycle, these fungi exemplify the postulate that natural selection acts to adjust the length of time spent in a particular life stage to maximize lifetime reproductive success. ‘Telescoping’ of the rust life cycle by the dropping of one or more stages is analogous to progenesis, a term used in developmental biology to refer to abbreviated ontogeny by accelerated (precocious) sexual maturation. Gould (1977: Chap. 9) reviews some of the many examples among animals, including animal parasites (e.g., the ces- todes and certain copepods; see also Moran and Whitham 1988).

Thus, there are numerous instances where a parasitic species is progenotic with respect to free-living relatives or related parasites. This has been accomplished apparently by the adoption of a simplified life cycle from what was originally a CLC by deletion of the alternate host. Why heteroecism is retained seems a more challenging question. It has been traditionally interpreted adaptively, though retention may reflect constraint; this aspect is considered further in the next section.

Second, what we see in the complex cycles of the rust fungi may reflect an historical legacy—an extreme means to regulate nuclear access to the germ line. As we saw earlier, animals, plants, and fungi all accomplished this regulation differently. The rust fungi appear to have done so in particularly extreme fashion. An interesting and imaginative interpretation has been advanced by Buss (1987 p. 158) and is worth quoting at length below (emphasis added):

The forces which spawned the unusual life cycles of fungi are the forces which challenged metazoans to develop elaborate mechanisms of historecognition. Competition within organisms for access to the germ line arises not only through mutation, but also through fusion between conspecifics. The potential consequences of fusion for integrity of the individual varied as a function of the primitive conditions of the plant, animal, and fungal kingdoms with respect to cell membrane architecture and cellularization. Plants were not challenged. Those animal groups that were challenged responded by evolving sensitive mechanisms of self-recognition.

The fungi turned the challenge to their own advantage inventing from it novel life cycles. A life cycle trait, fusibility, challenged the different clades differently, producing a phylogenetic restriction in those groups which allow fusion. One group, the fungi, responded with a unique developmental innovation which subsequently fueled the evolution of complex, idiosyncratic life cycles!' [emphasis added]