The ternary complex factor (TCF) subfamily of ETS‐domain transcription factors form ternary complexes with the serum response factor (SRF) and the c‐fos SRE. Extracellular signals are relayed via MAP kinase signal transduction pathways through the TCF component of the ternary complex. Protein–protein interactions between TCFs and SRF play an essential role in formation of this ternary complex. A 30 amino acid sequence encompassing the TCF B‐box is sufficient to mediate interactions with SRF. In this study we have identified amino acids which are critical for this interaction and derived a molecular model of the SRF binding interface. Alanine scanning of the Elk‐1 B‐box reveals five predominantly hydrophobic residues which are essential for binding to SRF and for ternary complex formation in vitro and in vivo. These amino acids are predicted to lie on one face of an α‐helix. Peptides encompassing the B‐box retain biological activity and have helix‐forming propensity. α‐Helix and ternary complex formation is disrupted by the introduction of helix‐breaking proline residues. Our results are consistent with a model in which the Elk‐1 B‐box forms an inducible α‐helix which presents a hydrophobic face for interaction with SRF. We discuss the wider applicability of our results to similar short protein–protein interaction motifs found in other transcription factors.

Pathogens can alter species composition and ecosystem function by causing direct and indirect effects on communities. Zoospores of the chytrid fungus Batrachochytrium dendrobatidis (hereafter, Bd), a pathogen implicated in worldwide amphibian declines, can be consumed by filter-feeding zooplankton and can damage mouthparts of infected amphibian larvae. Consequently, we hypothesized that this pathogen would affect the abundance of zooplankton and survival and feeding abilities of larval amphibian hosts. In turn, this could affect the algal food resources of zooplankton and tadpoles, which can include phytoplankton and periphyton. We tested these hypotheses by manipulating the presence of western toad (Anaxyrus boreas) larvae and Bd in outdoor mesocosms and quantifying the densities of amphibians, zooplankton, phytoplankton, and periphyton. Bd infections reduced amphibian larval densities, but this only weakly benefitted periphyton (a density-mediated indirect effect). After controlling for larval densities, mesocosms with Bd-exposed larvae had significantly more periphyton biomass than mesocosms with larvae but no Bd, consistent with infection-induced mouthpart damage reducing larval feeding rates on attached algae (a trait-mediated indirect effect). In fact, the estimated trait-mediated effect of Bd on periphyton biomass was larger than the densitymediated effect. However, we did not find evidence that Bd exposure triggered a switch to filter-feeding phytoplankton. The presence of Bd was associated with increased copepod abundance, consistent with zooplankton consuming chytrid zoospores. These results suggest that Bd has the potential to cause both trait-and density-mediated indirect effects that can alter community composition and perhaps the primary productivity of freshwater ecosystems.