On the origin of species in Termitomyces

In this thesis I focussed on a mutualistic symbiosis; the fungus-growing termites and their fungal symbiont Termitomyces. The origin of this symbiosis between fungus-growing termites (Macrotermitinae) and Basidiomycete fungi of the genus Termitomyces has been estimated at approximately 30 million years ago in the rainforests of central Africa. The symbiosis has a single origin with no known reversals of both partners to a non-symbiotic state. The single origin of fungiculture and the absence of any reversals of the fungus to a free-living state are puzzling as both parties have retained independent reproduction and dispersal, which could potentially de-stabilize the symbiosis by leaving room for the fungus to escape or for the termite to domesticate more than one symbiont. For this thesis I studied the biology of the closest relatives of Termitomyces in order to reconstruct the biology of its ancestor. Furthermore, I studied the phylogeny and biology of the genus Termitomyces in more detail to understand what made the genus so diverse and successful.

In order to make inferences on the events leading to domestication of Termitomyces a well-supported phylogeny is needed. In chapter 2 I performed a phylogenomic analysis of  25 species of Termitomyces and 21 related non-domesticated species. To reconstruct the biology of the ancestor of Termitomyces I also studied the biology of the non-domesticated species. The phylogenomic analysis recovered the insect-faecal associated genus Arthromyces as the sister group of Termitomyces. I found that Arthromyces shares a suite of traits with Termitomyces, indicating that their common ancestor also possessed these traits. I hypothesize that this set of traits predisposed the ancestor of Termitomyces towards domestication.

One of traits found in many species related to Termitomyces is the production of asexual spores (conidia). I hypothesized that these conidia play a role in local dispersal of the fungus and potentially in faster substrate colonisation. In chapter 3 I conducted a population study on the insect-faecal associated and conidia-producing fungus Blastosporella zonata. I sequenced two highly variable genetic markers of 21 collections of B. zonata mushrooms at three collection sites in cloud forests near Murillo, Colombia. I found signatures of clonality within a collection site but usually not between sites, which indicates that conidia are mainly used for local dispersal. There was genetic diversity as well, particularly between collection sites, indicating an important role for sexual spores (basidiospores) and resulting in unique genotypes of most collections. Unexpectedly. by reconstructing fungal haplotypes and subsequent phylogenetic analysis on these, I found evidence for the existence of two cryptic biological species in B. zonata.

In the studies I conducted in chapter 2 I used five unidentified species of Tephrocybe. As the morphological descriptions as well as the DNA sequences of these species did not match any species already known, I chose to describe those species in chapter 4. The taxonomy for these five new species was challenging as morphological traits and biogeography did not align with the molecular phylogeny. Therefore, I discussed several alternative options to keep genera monophyletic. This ultimately led to the erection of four new genera: Australocybe, Nigrocarnea, Phaeotephrocybe and Praearthromyces.

The genus Termitomyces contains about 40 described species and harbours a lot of morphological diversity. Probably many species of Termitomyces remain undescribed as there is a strong bias towards the description of species that regularly produce mushrooms. In chapter 5 I assembled a large dataset of over 1500 DNA sequences of Termitomyces and used species delimitation software to sort sequences into hypothetical species based on the barcoding gap within the dataset. Using this approach, I recovered 87 phylogenetic species and for these species I collected specimen metadata such as continent of origin and termite host genus. A phylogenetic reconstruction of the 87 species revealed five main symbiont groups which corresponded with the five largest fungus-growing termite genera. I identified several factors which could be involved in speciation: fruiting mode (underground or aboveground), geographic separation, symbiont transmission mode and suppression of mushroom formation by the host termite. Finally, I demonstrated that mushroom morphology does not correlate with phylogeny and therefore, without additional support of molecular data, is not useful for taxonomy in this genus.

Uniparental vertical transmission of symbionts is an important factor in aligning reproductive interests of host and symbiont and thus stabilizing the symbiosis. In fungus-growing termites, horizontal transmission is the predominant transmission mode, but uniparental vertical transmission has evolved via the female reproductives of the genus Microtermes and via the male reproductives of the species M. bellicosus. In chapter 6 I studied populations of Microtermes and their symbionts in South Africa. I found high genetic diversity in symbiont populations and low host-specificity, which indicates that frequent horizontal symbiont exchange occurs as well as sexual reproduction, despite the absence of mushrooms of these species. I therefore argue that Microtermes species regularly acquire symbionts associated with species of other genera that do not suppress mushroom formation, such as species of the genus Ancistrotermes.

New species of mushroom-forming fungi are often described based on morphological features of the mushroom. In the genus Termitomyces this has caused a heavy bias towards the description of species, which regularly produce mushrooms, leaving species that rarely or infrequently fruit undescribed. In chapter 7 I describe, for the first time in this genus, a species of Termitomyces with no reported naturally produced mushrooms, based on DNA sequence evidence, biological data and asexual features.

In chapter 8, I discuss how investigating and understanding the biology of species closely related to fungal symbionts can assist in reconstructing the biology of the ancestor and thereby determine conditions that facilitated fungal domestication by an insect. Although my research focused on a specific group of fungus-growing insects, the fungus-growing termites, I discuss how my approach can be applied to different unrelated groups of fungus-growing insects as well. Finally I discuss how my work not only offers significant contributions to the knowledge on the origin and subsequent evolution of fungiculture in termites, but also contributed to general fundamental knowledge of mutualistic symbioses in general.