Poster Presentation The 3rd Prato Conference on Pore Forming Proteins 2015

The structure of anticipated folding core of the RTX-proteins (#48)

Lucia Motlová 1 2 , Ladislav Bumba 2 , Radovan Fišer 1 2
  1. Department of genetics and microbiology, Faculty of Science Charles University in Prague, Prague, Czech Republic
  2. Institute of Microbiology, Czech Academy of Sciences, Prague, Czech Republic

Repeats-in-ToXin (RTX) protein family of Gram-negative bacteria includes wide range of proteins with various biological functions comprising pore-forming toxins. They are characterized by export across the bacterial envelope via the type I secretion system and the C-terminal nonapeptide glycine- and aspartate-rich repeats, that fold into characteristic parallel β-roll structure with two facing sheets of β-strands linked by calcium-loaded turns. In different RTX proteins a C-terminal signal sequence with variable length and sequence is located. In our investigation of structure of adenylate cyclase toxin (CyaA) from bacterium Bordetella pertussis we found three anti-parallel beta-strands which do not facing each other but they are in one plane of  β-roll structure like ladder end continuing by turn-back to the other plane of β-roll in the very C-terminus of repeats which are closing the typical β-roll structure. This region in RTX structure is crutial for folding of whole toxin and bears a structural resemblance to corresponding regions in all solved RTX-proteins structures, which are extracellular lipase from Pseudomonas species, extracellular lipase from Serratia marcescens, zinc-dependent protease from Serratia species, alkaline protease from Pseudomonas aeruginosa, extracellular protease PrtC from Erwinia chrysanthemi and psychrophilic metalloprotease from an Antarctic Pseudomonas species. We present a simple algorithm useful for quantitative comparison of secondary structures of proteins with low sequence homology. We analyzed the C-terminal anti-parallel beta-strands structure between all solved RTX structures by similarity in sequences and secondary structures using this tool. The sequence identity in this region is in general very low (from 6 to 17 % between non-related structures, from 33 to 52 % between functionally related structures) but the secondary structure is very similar.