Studies on three E. coli DEAD-box helicases point to an unwinding mechanism different from that of model DNA helicases.

DEAD-box proteins participate in various aspects of RNA metabolism in all organisms. These RNA-dependent ATPases are usually regarded as double-stranded RNA unwinding enzymes, though in vitro this activity has only been demonstrated for a subset of them. Given their high biological specificity, their equivocal unwinding activity may reflect the noncognate character of the substrates used in vitro. Here, we pinpoint other reasons for this elusiveness. We have compared the ATPase and helicase activities of three E. coli DEAD-box proteins, CsdA, RhlE and SrmB. Whereas the ATPase activity of all proteins is stimulated (albeit to various degree) by long RNAs, only RhlE is stimulated by short oligoribonucleotides. Consistently, all three proteins can unwind RNA duplexes with long single-stranded extensions, but only RhlE is effective when extensions are short or absent. Another critical constraint concerns the length of the duplex region: in the case of RhlE, the ratio (duplex unwound)/(ATP hydrolyzed) drops 1000-fold upon going from 11 to 14 base pairs, indicating a low processivity. Remarkably, allowing for these constraints, all three proteins can unwind substrates with either 5' or 3' extensions (or no extension in the case of RhlE). This behavior, which contrasts with that of well studied SF1 DNA helicases, is discussed in the light of available structural and biochemical data.