A molecular ruler

In the initiation step of RNA interference, the multidomain ribonuclease Dicer processes double-stranded RNA (dsRNA) into small fragments that are ∼21–25 nucleotides long. Several models have been put forward to explain how Dicer produces RNA fragments of such an exact size, but structural insights have been lacking. However, in Science, Doudna and colleagues now describe the crystal structure of full-length Giardia intestinalis Dicer at 3.3-Å resolution.

The G. intestinalis Dicer contains the PAZ domain, which binds to the end of dsRNA, and the tandem RNase III catalytic domains that are characteristic of Dicer, but it lacks other domains that are found in Dicer in higher eukaryotes. However, G. intestinalis Dicer substituted for Schizosaccharomyces pombe Dicer in vivo, which indicates a conserved mechanism of Dicer-catalysed dsRNA processing. Using Er3+, which has an inhibitory effect on G. intestinalis Dicer, the authors showed that its RNase III domains use a two-metal-ion catalytic mechanism. Furthermore, the structure of G. intestinalis Dicer, which seems to be conserved, indicates how Dicer generates specifically sized RNA fragments. A flat, positively charged platform domain separates the PAZ domain from the RNase III domains by a distance of 65 Å. This distance matches the length of 25 base pairs of RNA. So, Dicer works as “...a molecular ruler that measures and cleaves ∼25 nucleotides from the end of a dsRNA.” REFERENCE MacRae, I. J. et al. Structural basis for double-stranded RNA processing by Dicer. Science 311, 195–198 (2006)

A riveting toxin

The bacterial toxin aerolysin kills cells by forming heptameric channels in plasma membranes, but the channel structure and the mechanisms of membrane insertion and channel formation are poorly understood. van der Goot and colleagues focused on a loop in domain 3 (DIII loop) of aerolysin, because it contains alternating charged and neutral residues that could form a transmembrane β-hairpin (the channel might therefore be a β-barrel). In The EMBO Journal, they describe their study of the topology of the DIII loop and its role in channel formation.

The authors restrained the DIII loop using disulphide bonds, and showed that this did not affect aerolysin heptamerization. However, it affected membrane insertion and channel formation. By carrying out cysteine-scanning mutagenesis coupled with channel analysis using planar lipid bilayers and cysteine-specific labelling, they identified residues in the DIII loop that line the channel lumen and the sequence that forms the turn of the β-hairpin. By mutating this hydrophobic WPLVG sequence, they showed that this turn drives membrane insertion, and they propose that once it has crossed a bilayer, it folds back perpendicular to the barrel axis. They modelled this rivet-like conformation for the membrane anchoring of aerolysin, and sequence alignments indicate that many other pore-forming toxins might also use this type of channel riveting. REFERENCE Iacovache, I. et al. A rivet model for channel formation by aerolysin-like pore-forming toxins. EMBO J. 25, 457–466 (2006)