Abstract
BACKGROUND: Attine ants live in symbiosis with a basidiomycetous fungus that they rear on a substrate of plant material. This indirect herbivory implies that the symbiosis is likely to be nitrogen deprived, so that specific mechanisms may have evolved to enhance protein availability. We therefore hypothesized that fungal proteinase activity may have been under selection for efficiency and that different classes of proteinases might be involved. RESULTS: We determined proteinase activity profiles across a wide pH range for fungus gardens of 14 Panamanian species of fungus-growing ants, representing eight genera. We mapped these activity profiles on an independently obtained molecular phylogeny of the symbionts and show that total proteinase activity in lower attine symbionts peaks at ca. pH 6. The higher attine symbionts that have no known free-living relatives had much higher proteinase activities than the lower attine symbionts. Their total in vitro proteinase activity peaked at pH values around 5, which is close to the pH that the ants maintain in their fungus gardens, suggesting that the pH optimum of fungal proteinases may have changed after the irreversible domestication of evolutionary more derived fungal symbionts. This notion is also supported by buffering capacities of fungus gardens at pH 5.2 being remarkably high, and suggests that the fungal symbiont actively helps to maintain garden acidity at this specific level. Metalloproteinases dominated the activity profiles of lower attine gardens and may thus represent the ancestral type of proteinase production, whereas serine proteinase activity dominated the activity profiles of the higher attine gardens reared by Trachymyrmex and Sericomyrmex, suggesting that there may be trade-offs in the production of these enzyme classes. Remarkably, the single symbiont that is shared by species of the crown group of Atta and Acromyrmex leaf-cutting ants mostly showed metalloproteinase activity, suggesting that recurrent changes in enzyme production may have occurred throughout the domestication history of fungus-garden symbionts. CONCLUSIONS: Proteinase pH optima and buffering capacities of fungal symbionts appear to have evolved remarkable adaptations to living in obligate symbiosis with farming ants. Although the functional roles of serine and metalloproteinases in fungus gardens are unknown, the differential production of these classes of proteolytic enzymes suggest that substrate specificity may be important and that trade-offs may prevent the simultaneous upregulation of both classes of enzymes.