How do nonspecific enzymes that help to correct RNA folding identify misfolded structures among similar, properly folded RNAs? It seems that careful discrimination has little to do with it.
Most RNAs adopt defined three-dimensional structures that allow them to function properly. But RNAs also often misfold into inactive structures, which can persist for a long time1. Fortunately, RNA chaperone proteins give misfolded RNAs the opportunity to refold correctly1,2. However, many RNA chaperones are nonspecific, and correctly folded and misfolded RNA structures tend to be very similar. On page 1014 of this issue, Bhaskaran and Russell3 describe the way one RNA chaperone, CYT-19, confronts this challenge.
CYT-19 is a member of the large group of DEAD-box proteins — enzymes that are involved in almost all aspects of RNA metabolism4. The name DEAD-box comes from the amino-acid sequence (D-E-A-D in single-letter code) of one of the evolutionarily conserved motifs in these proteins. DEAD-box proteins are often called RNA helicases, because they can unwind short RNA duplexes using ATP as an energy source. But in contrast to canonical helicase enzymes, which unwind duplexes by travelling along the nucleic-acid strands, DEAD-box proteins do not move on the RNA duplexes but load directly to double-stranded regions and then pry the strands apart5.
Among the targets of CYT-19 are misfolded group-I intron RNAs, which are catalytic RNAs that excise themselves from, and subsequently rejoin the ends of, a larger precursor RNA. It is thought that CYT-19 unwinds one or more of the short helices that are the building-blocks of larger RNA structures6. Like most DEAD-box helicases, CYT-19 unwinds duplexes indiscriminately and, given that the correctly folded and misfolded group-I introns form almost identical helices, it was not clear how CYT-19 distinguishes between these structures. The general assumption was that, somehow, CYT-19 unravels only the misfolded structures to promote correct folding.
Bhaskaran and Russell3 make the remarkable observation that CYT-19 unfolds both correctly folded and misfolded structures. However, misfolded RNAs are disassembled at a faster rate. Consequently, at any given time, the misfolded group-I intron RNAs have a better chance of unfolding and refolding correctly (Fig. 1, overleaf).
But why are misfolded RNAs disassembled faster than the correct structures? The authors show that mutated group-I introns that are functional but cannot pack all their helices together in the ultimate higher-order structure are readily unfolded by CYT-19. So it is possible that the reason for the slower unfolding rate of correctly folded RNAs is that crucial helices are sequestered in the higher-order structure and are thus protected from unwinding by CYT-19. This possibility is supported by an earlier study7 showing that CYT-19 cannot easily unwind duplexes hidden within an RNA structure. If this is true for other nonspecific DEAD-box chaperones, this would be an elegant and unexpectedly simple way by which these proteins might preferentially unravel misfolded RNA structures.
Bhaskaran and Russell also make another intriguing observation. Under certain conditions — for example, at reduced magnesium concentrations, a consequence of which is the weakening of higher-order RNA structures — CYT-19 increases the ratio of misfolded RNAs to correctly folded structures. This is remarkable, because misfolded group-I RNAs are thermodynamically less stable than the correct structures and, at equilibrium, the most stable species dominates in a mixture of possible structures.
The authors find that CYT-19 achieves the redistribution of RNA species against their thermodynamic equilibrium by establishing a 'kinetically controlled steady state' (Fig. 1b). In this state, CYT-19 constantly unwinds RNA structures in an ATP-dependent manner, so that the distribution of these structures is no longer dictated by their thermodynamic stabilities but, instead, by the rate at which they are actively unravelled by CYT-19, as well as by the rate of their refolding. Under these conditions, CYT-19 readily unravels both misfolded and correct structures. But when the RNA subsequently misfolds it does so faster than correctly refolding RNA, and therefore accumulates.
Similar ATP-driven redistribution of model RNA complexes by another DEAD-box protein has been reported8. So it seems that DEAD-box proteins can establish kinetically controlled steady states for various RNAs. The ability of these proteins to favour the accumulation of thermodynamically less stable RNA species seems particularly beneficial for processes in which RNAs must undergo extensive structural changes, such as during ribosome assembly and pre-messenger-RNA splicing. Both of these processes involve numerous DEAD-box proteins4.
The work of Bhaskaran and Russell3 implies that sequence nonspecificity of DEAD-box proteins is more of an asset than a disadvantage, at least for their function as RNA chaperones. Nonspecific chaperone proteins can work on many misfolded RNAs and at different positions within them, which obviates the need for a large number of highly specific proteins that would exclusively aid the folding of each of the many cellular RNAs. The potency of unfolding by the omnipresent and nonspecific DEAD-box proteins also offers one possible explanation for why some RNAs tolerate misfolding into fairly stable structures. This is because they can rely on DEAD-box proteins to rescue them. But an increased awareness of the versatility of nonspecific DEAD-box proteins highlights the next question: why does the cell use so many of these enzymes?
Herschlag, D. J. Biol. Chem. 270, 20871–20874 (1995).
Schroeder, R., Barta, A. & Semrad, K. Nature Rev. Mol. Cell Biol. 5, 908–919 (2004).
Bhaskaran, H. & Russell, R. Nature 449, 1014–1018 (2007).
Linder, P. Nucleic Acids Res. 34, 4168–4180 (2006).
Yang, Q. & Jankowsky, E. Nature Struct. Mol. Biol. 13, 981–986 (2006).
Mohr, S., Stryker, J. M. & Lambowitz, A. M. Cell 109, 769–779 (2002).
Tijerina, P., Bhaskaran, H. & Russell, R. Proc. Natl Acad. Sci. USA 103, 16698–16703 (2006).
Yang, Q., Fairman, M. E. & Jankowsky, E. J. Mol. Biol. 368, 1087–1100 (2007).