Symbioses have driven a number of evolutionary innovations in the history of life. Genomic investigations into mutualistic symbioses between insects and bacteria, corals and dinoflagellates, have revealed that long and intimate associations can result in genomic modification of both symbionts, reflecting their symbiotic relationships (46-50). Their heterotrophic nature led fungi to evolve mutualistic symbioses with other organisms that provide them with various nutrients, namely carbon. The most well-studied examples are mycorrhizal fungi that form mutualistic associations with more than 90 % of extant plant species (51). Numerous molecular studies have characterized fungal and plant genes involved in the steps leading to functional symbiosis (51). On the other hand, genetic mechanisms of the lichen symbiosis, distributed among ca. 20 % of all fungal species and considered as one of the most successful mutualistic symbioses, remain largely unknown. Unlike other mutualistic symbioses between fungi and photoautotrophs, lichens are unique in the development of a symbiotic structure that often includes complete envelopment of algal/cyanobacterial cells by fungal hyphae (8). Within this symbiotic structure the fungus (mycobiont) optimizes illumination, gas exchange, water and nutrient supply to create a favourable and protective environment for the photosynthetic partner (photobiont) which produces photosynthetic products that nourish not only itself but the entire lichen.
Here, from the results of our comparative transcriptomic analysis between the non-symbiotic and symbiotic states of the Usnea hakonenesis system, we identify genes specifically expressed in the later lichenization stage, thallus differentiation, and hypothesize the genetic mechanisms involved in the development of a functional symbiotic structure. The advantage of the laboratory experiment is that it can control factors affecting the symbiotic interaction. However, resynthesized thalli of U. hakonensis cannot develop a symbiotic phenotype identical to that of the natural (20). To counter these shortcomings we focused on the significant differentially expressed genes (DEGs) that are consistent between the two symbiotic states. The imposition of consistency decreased the number of available DEGs but successfully filtered genes more relevant to the symbiosis.
The significant over-representation of the GO term ‘transmembrane transport’ among the algal consistently down-regulated genes in the symbiotic states may indicate active nutrient transports in the non-symbiotic culture where algal cells have direct contact with the medium. On the other hand, the GO term ‘oxidation-reduction process’ was over-represented both among the up- and down-regulated fungal genes in the symbiotic states. Although the assigned genes predicted to be involved in the oxidation-reduction process of various metabolic activities are similar, genes involved in defence against oxidative stress, such as alternative oxidase, thioredoxin, and thioredoxin reductase, were only identified among the down-regulated genes. In lichens, both enzymatic and non-enzymatic antioxidants are suggested to be involved in protection from oxidative stress (52). The down-regulation of these genes in the symbiotic states possibly indicates a shift in fungal preference for antioxidants upon symbiosis. The U. hakonensis system might also be well suited for investigation into stress physiology of lichens.
The non-symbiotic fungus and alga were grown on a medium under a laboratory condition that was totally different from the condition in nature. The over-representation of GO terms related to metabolic processes and transmembrane transports implies that the down-regulated genes include considerable number of genes induced under such specific culture conditions. To accomplish our aim to identify genes involved in the symbiotic interaction we further focused on the consistent up-regulated genes as ‘symbiosis-related genes’.
Establishment of the symbiotic interface
Inside the thalli fungal hyphae attach to algal cells with structures called haustoria. In the case of fruticose ascomycetous lichens with trebouxioid photobionts, including Usnea species, fungal hyphae are observed to grow into but do not penetrate the algal cell walls (2, 53-55). At a haustorium, a thick hydrophilic layer overlies the fungal cell wall the outer part of which is covered by a hydrophobic layer that spreads to the surface of algal cell walls (2, 56). The hydrophilic layer is known to absorb water at hydration and the hydrophobic layer to seal the apoplastic continuum of the fungus and the alga, together producing passive fluxes of water and nutritional solutes from the thallus exterior to the interior algal layer (Figure 2a) (2, 56, 57). The hydrophobic layer secures gas-filled zones inside the thallus even at full hydration which is a prerequisite for efficient CO2 diffusion to the algal cells. The symbiosis-related genes identified in this study include many genes that are likely to be involved in producing a functional symbiotic interface.
Cell wall modification
In U. hakonensis, the three fungal symbiosis-related glycoside hydrolase genes, Uhk_019559, Uhk_007999, and Uhk_005214, may function in the degradation of algal and fungal cell walls during the morphogenesis of haustoria (Figure 2b). Uhk_019559 and Uhk_007999 possibly degrade cellulose and hemicellulose in algal cell walls (58-60) when fungal hyphae grow to reach algal cells and develop haustorial structures at the contact sites (2, 53). Uhk_005214 may break down beta-1,3-glucan in fungal cell walls at haustoria where algal and fungal cell walls were reported to be thinner (61, 62), possibly causing local loosening of the fungal wall that could facilitate transport between the symbionts. Uhk_002614 may possibly be involved in the production of the hydrophilic layer comprises of polysaccharides (Figure 2c) (63).
Hydrophobic layer
In lichens, hydrophobins have been identified in the region where hyphae are in contact with algal cells (26, 56, 57, 64-66) called the hydrophobic layer. A class I hydrophobin gene identified in Xanthoria parietina (XPH1) was indicated to be indispensable for the maintenance of a symbiotic relationship (65). In the present study, Uhk_01658 showed significant up-regulated expression in both symbiotic states despite differences in the developmental stages and growth conditions of the two. Our results suggest that this hydrophobin could be the main component of the hydrophobic layer and may play an important role throughout the symbiotic interaction (Figure 2d).
The majority of lichens are known to deposit secondary metabolites most of which are unique to lichens. Crystals of secondary metabolites are often observed within and on the hydrophobic layer and are thought to enhance the hydrophobicity of the layer (2). A symbiosis-related gene encoding a polyketide synthase (Uhk_016999) showed high similarity only to a polyketide synthase identified in U. longissimi (GenBank accession: AEJ54468.1). The up-regulated expression of this gene in both symbiotic states suggests that this gene may be involved in the synthesis of symbiosis-specific metabolites specific to these two species or possibly to genus Usnea.
Lipids and fatty acids are also reported as components of the hydrophobic layer (28). Expression of genes involved in sphingolipid and phospholipid biosynthesis was up-regulated in the symbiotic states (Uhk_003277, Uhk_017380, and Uhk_017885). Sphingolipids and phospholipids are constitutive lipids of membrane microdomains and extracellular vesicles respectively (29, 30, 67, 68), both reported to be involved in the communication between symbiotic partners (69-73). The up-regulated expression of the genes involved in these fungal lipid metabolisms might suggest that lipids are not merely the components of the hydrophobic layer but also used as fungal tools to communicate with algal partners (Figure 2e).
At least two fungal symbiosis-related genes are predicted to encode lipases (Uhk_003579 and Uhk_013145). Joneson et al. (12) reported up-regulated expression of Lip 3 in the second stage (contact) of lichenization in the lichen-forming fungus Cladonia grayi. Inferred from extracellular activity of Lip3, they suggested that lipids in the hydrophobic layer could be the targets of the secreted lipase. We likewise assume that the two lipases might hydrolyse fungal (or possibly algal) lipids (possibly triacylglycerols) in the layer, releasing fatty acids instead. Glycerol released with fatty acids by lipid digestion may be recycled in fungal cells as indicated by the symbiosis-related expression of a glycerol transporter (Uhk_011461) and a glycerol dehydrogenase (Uhk_011759) (Figure 2f).
Nutrient flow between the fungal and algal partners
The mutualistic symbiosis of lichens is ensured by benefit which both fungal and algal partners have from the relationship. The mycobiont receives carbohydrates produced by the photosynthesis of the photobiont, whereas the photobiont receives a suitable environment and nutrients for the photosynthesis. To initiate and maintain such a relationship, nutrients must be sufficiently produced and efficiently transferred between the symbiotic partners. In this study, genes involved in photosynthesis and transport of carbohydrate, phosphate, and nitrogen were identified as symbiosis-related genes. The resynthesized thalli we used to extract RNA were protrusions formed on top of fungal colonies where neither fungal nor algal cells have direct access to nutrients in the medium (Figure 1). Therefore, the identified symbiosis-related transporter genes are most likely involved in nutrient transport from the symbiotic interface rather than that from the medium.
Carbon flow
Production
In the U. hakonensis system, the five carbon polyol ribitol, is produced by algal photosynthesis and transported to the mycobiont. (37-40, 74). The enhanced expression of photosynthesis genes in the symbiotic states implies modifications under the constraints of the symbiosis, that possibly guarantee adequate carbohydrate supply for the entire thallus (Figure 3a). The up-regulation of algal symbiosis-related gene Ttzw_014612, corresponding to the D1 reaction centre protein in photosystem II (PSII), could be a part of an adaptation to the symbiotic lifestyle that imposes higher light intensities on the photobiont than the non-symbiotic lifestyle (75). The D1 reaction centre protein is known as the primary target of light-induced oxidative damage. The immediate replacement of damaged D1 proteins is vital for photosynthetic organisms to avoid inactivation of PSII (76). In the green alga Chlamydomonas reinhardtii, expression of the D1 protein gene is mainly regulated at the level of translation while transcription has been reported to be constitutive (77). If analogous regulation can be supposed, the up-regulated expression of the D1 protein gene in the symbiotic states indicates an increase in the number of stock transcripts in cells, thereby enabling rapid synthesis of D1 proteins, i.e., rapid recovery from photodamage.
The symbiosis-specific expression of carbonic anhydrase (CA) (Ttzw_000076) that may function in CO2 transport and fixation is likely to indicate algal adaptation to CO2 conditions inside a symbiotic structure. In symbiosis, the layered structure of a thallus could change CO2 concentration around algal cells. The change in CO2 conditions may induce expression of CA and increased enzyme concentration could ensure efficient transport and fixation of CO2. The expansion of a specific subclass of cytoplasmic β CAs in the genome of Trebouxiaceae lichen-forming alga, Asterochloris glomerata, has been reported and a potential relationship to the lichen symbiosis was indicated (13). Besides the important roles CAs play in photosynthesis, they may be involved in metabolic activities such as biosynthesis of amino acids, lipids, and development of nitrogen-fixing root nodules (33, 34). The identification of symbiosis-related CAs in lichen-forming algae is intriguing because it raises the possibility of symbiotic-specific regulatory systems that may control algal metabolism via the CO2 condition inside a thallus.
Light-dependent regulation may also be involved in algal symbiosis-specific metabolism as indicated by the symbiosis-related ferredoxin:thioredoxin reductase (FTR) (Ttzw_014778). FTR is a key protein in the light-dependent regulation of carbon flow and various metabolic activities in photosynthetic organisms (35, 36). Richardson and Smith (40) reported that Trebouxia photobionts directly isolated from lichen thalli released ribitol from cells while cultured Trebouxia photobionts produced relatively more sucrose than ribitol and almost completely lost the ability to release fixed carbon. Such symbiosis-specific carbon metabolism could be regulated by the ferredoxin/thioredoxin system that links light to enzymatic activities. As light has been reported as an essential factor for the induction of a symbiotic phenotype (7), light condition in a thallus could probably be one of the triggers that initiate the regulatory system which induces algal metabolic activities required in the lichenization process (Figure 3b).
Taken together, the functions of these algal symbiosis-related genes may indicate not only enhanced photosynthetic activity of the photobiont but also adjustment of broader metabolic activities probably required to develop and stabilize the symbiotic lifestyle.
Transportation
Ribitol excreted from photobiont cells to the symbiotic interface should be readily taken up by the mycobiont. Recently, Yoshino et al. found that transporters similar to the fungal ribitol transporter AmLAT2 (41) tend to be duplicated in most lichen-forming species in Lecanoromycetes (78). They hypothesized that the duplication of ribitol transporter genes may be associated with lichenization events in Lecanoromycetes and is likely to enable the efficient utilization of ribitol. In line with their finding we identified three AmLAT2-like genes in the fungal transcriptome, two of which (Uhk_000003 and Uhk_000004) were up-regulated in the symbiotic states. Armaleo et al. (13) reported that one of the five putative ribitol transporters found in the C. grayi genome was induced in the early stage of lichenization. The up-regulated gene expression of the ribitol transporters in the early and late stages of lichenization may indicate their fundamental roles in lichenization (Figure 3c).
Phosphorus and Nitrogen flow
While carbohydrates are exclusively provided by the photobiont, other nutrients essential for various metabolic activities in lichens are provided by the mycobiont which has direct access to the surrounding environment. Nitrogen and phosphorus are known to affect the growth of lichens, and are suggested to be involved in the regulation of the symbiotic balance between the mycobiont and the photobiont (79-81).
The up-regulated expression of genes corresponding to fungal MFS phosphate transporter (Uhk_019687) and acid phosphatase (Uhk_006220), and algal transmembrane ATPase (Ttzw_021111) indicates that phosphorus is likely to be provided as inorganic phosphate (Pi) from the mycobiont to the photobiont (Figure 3d). In arbuscular mycorrhizal symbiosis, plant plasma membrane H+-ATPase is proposed to create a proton electrochemical gradient across the membrane that energizes the uptake of Pi by Pi transporters (82, 83). Algal Pi transporter genes were not identified as symbiosis-related but one algal H+/Pi symporter gene was significantly up-regulated in the natural thalli (Table 6).
Unlike carbon and phosphorous, no symbiosis-related transporters were identified for nitrogen transport between the symbionts. (Figure 3e and Table 7). Previous genetic studies on early stages of lichenization reported increased expression of fungal ammonium transporter genes (11, 13). However in the present study fungal ammonium transporter genes showed no up-regulated expression in either the resynthesized or the natural symbiotic state relative to the non-symbiotic state. The inconsistency between the two symbiotic states and the previous studies may simply reflect the difference in accessible nitrogen sources, or may possibly reflect stage-specific nitrogen metabolism. Further investigation is required to elucidate the mechanism of nitrogen transport in lichens. Nevertheless, symbiosis-specific nitrogen flow could be inferred from the symbiosis-related genes predicted to be involved in algal nitrogen metabolism ([Protein-PII] uridylyltransferase (Ttzw_019382) and glutamine amidotransferase (Ttzw_009103)). An interesting feature of PII signaling proteins is that they can integrate a nitrogen signal with a carbon signal and thereby coordinate the cellular balance of nitrogen and carbon (44, 45). The indicated involvement of PII proteins in algal symbiosis-specific nitrogen metabolism might enable the mycobiont to control the carbon metabolism of the photobiont via nitrogen supply, which would possibly help balance the growth of the symbionts.