Structural basis for RNA-duplex unwinding by the DEAD-box helicase DbpA

  1. Jan Philip Wurm

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This manuscript is interesting to a broad audience in the general fields of RNA and structural biology. It provides detailed and important molecular insight into one of the mechanisms by which ATP-fueled RNA helicases can cause the local destabilisation of terminal base-pairs and eventually contribute to RNA structure remodelling and it is a prime example of how crystallographic high-resolution snapshots of conformational intermediates can be combined with sophisticated NMR techniques and assays into a comprehensive model. The manuscript would benefit from a broader and more explicit comparative discussion including the limitations of the proposed model, because DbpA is a rather specialised RNA helicase and because the double-stranded RNA substrates were specifically designed to exclusively investigate unwinding from the side of a short 5'-overhang.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

DEAD-box RNA helicases are implicated in most aspects of RNA biology, where these enzymes unwind short RNA duplexes in an ATP-dependent manner. During the central step of the unwinding cycle, the two domains of the helicase core form a distinct closed conformation that destabilizes the RNA duplex, which ultimately leads to duplex melting. Despite the importance of this step for the unwinding process no high resolution structures of this state are available. Here, we employ nuclear magnetic resonance spectroscopy and X-ray crystallography to determine structures of the DEAD- box helicase DbpA in the closed conformation, complexed with substrate duplexes and single- stranded unwinding product. These structures reveal that DbpA initiates duplex unwinding by interacting with up to three base-paired nucleotides and a 5’ single-stranded RNA duplex overhang. These high-resolution snapshots, together with biochemical assays, rationalize the destabilization of the RNA duplex and are integrated into a conclusive model of the unwinding process.

Article activity feed

  1. eLife

    Evaluation Summary:

    This manuscript is interesting to a broad audience in the general fields of RNA and structural biology. It provides detailed and important molecular insight into one of the mechanisms by which ATP-fueled RNA helicases can cause the local destabilisation of terminal base-pairs and eventually contribute to RNA structure remodelling and it is a prime example of how crystallographic high-resolution snapshots of conformational intermediates can be combined with sophisticated NMR techniques and assays into a comprehensive model. The manuscript would benefit from a broader and more explicit comparative discussion including the limitations of the proposed model, because DbpA is a rather specialised RNA helicase and because the double-stranded RNA substrates were specifically designed to exclusively investigate unwinding from the side of a short 5'-overhang.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.

  2. eLife

    Reviewer #1 (Public Review):

    Due to the considerable free energy of base-pair formation, the rapid reorganization of RNA structure frequently requires protein-assisted dissociation of RNA strands (chaperoning), as they otherwise remain in kinetically trapped conformations. RNA helicases of the DEAD-box and related families have long been described to act in this context as ATP-fuelled single-strand RNA binding proteins, but mechanistic insight into how DEAD-box helicases can cause local RNA strand separation has been limited.

    The present work demonstrates that the DEAD-box helicase DbpA can locally 'unzip' double-stranded RNA in the 5'-to-3' direction (starting from a short 5' single-stranded overhang fixed by RecA_C, with unwinding on/ by RecA_N, after closure of the domains), and it elegantly provides detailed molecular insight into this process (crystal structure snapshots along the pathway, validation by NMR techniques and evidence for conformational sampling, biochemical activity assays).

    These results are remarkable because DbpA had previously been described to act on RNA duplexes preferentially in the 3'-to-5' direction (Diges and Uhlenbeck (2005); Biochemistry 44, 7903; information currently not mentioned in the manuscript) and because only for this latter direction of DEAD-box helicase action a mechanistic model had previously been suggested (Mss116p, Mallam et al., 2012, unwinding on/ by RecA_C (during domain closure?), less detailed overall). Moreover, also the translocating and processive DEAH-box helicases act in the 3'-to-5' direction, with fairly well understood mechanistic details (reviewed in e.g. Bohnsack et al., 2021, Biol. Chem. 402, 561, currently not mentioned in the manuscript).

    However, the unwinding activity of DbpA observed in the manuscript requires very specifically constructed RNA substrates. Furthermore, the author demonstrates that DbpA is NOT preferentially recruited to junctions of double-stranded RNAs with 5'-single-stranded overhangs if the 5'-overhangs are sufficiently long to be accommodated on the composite RNA binding surface of RecA_N and RecA_C. The presented model starts nevertheless with the assumption that 'the RNA duplex is recruited to the transiently-formed, closed-conformation' (p.23, l.456 and Fig. 8). These limitations of the approach and of the resulting model could be pointed out to the reader as such much more explicitly, possibly also invoking auxiliary protein-targeting factors rather than just a stochastic process of binding 5'-junctions (p.24, l.491).

    Also, a pre-orientation of the RecA_N and RecA_C domains by auxiliary factors in the natural context (with examples from other helicases) may be considered in the discussion (rather than only a random tumbling of the RecA_N domain, p.23, l.454). In particular, it would be interesting here how the author now mechanistically interprets his previous observation (Wurm et al., 2021) that DbpA is strongly stimulated to act on substrates 'in trans', if the HP92 hairpin contains 6-12 nt of 5' single-stranded overhang 'in cis'. Could these nucleotides potentially pre-occupy the composite RNA binding surface? This data is currently not mentioned in the manuscript.

    Finally, the proposed model is presented as 'the' general mode of action for DEAD-box helicases, whereas even for the very specialized DbpA, with its requirement to bind the HP92 hairpin, the precise natural RNA targets are unknown and unzipping in the 5'-to-3' direction may be a minor mode of action. A more balanced and extended discussion is warranted here, including the difference between translocating DEAH helicases and locally unwinding DEAD (DExD) helicases, the question of directionality and role auxiliary factors, and the explicit reference to data (see the specific points) which is not currently explained by the presented model and potentially in conflict. It may be useful to present some of these questions and references already in the introduction, such that also less specialized readers can better appreciate where and to which degree the present work provides previously unknown molecular details and conceptual novelty.

    From a technical point-of-view, the work is very solid and represents an enormous tour-de-force, especially considering that this is presented as a single author work. The methods, although rather specialized, are sufficiently well described to be followed by a broader audience, and the illustrations and cartoons are very clear and consistent. The text is generally easy to follow, although occasionally some of the phrasing can be improved for clarity.

  3. eLife

    Reviewer #2 (Public Review):

    The manuscript by Jan Philip Wurm structurally characterizes a long sought after catalytic intermediate of RNA double strand melting by an ATP-dependent RNA helicase of the DEAD-box family. These enzymes are the most abundant class of RNA remodelling enzymes and have essential functions during RNA metabolism throughout all organisms. In the past several structures of unwound single stranded (ss) RNA products bound to DEAD-box helicases together with intermediates of the coupled ATP hydrolysis reaction have been solved. However, it remained elusive how these enzymes bind and destabilize the double stranded (ds) RNA starting material. This knowledge gap is now closed with the current paper. The author exploits the fact that the enzmye DbpA from E.coli contains a specific RNA binding domain fused to the central helicase (RecA) domains, which facilitates stable RNA substrate binding. This set-up allows co-crystallization of several RNA substrates with ds or ss sections together with the helicase bound to the ATP-analog ADP-BeF3, and several well resolved crystal structures are obtained.

    The substrate-bound structure shows that the helicase binds to the ss-dsRNA junction contacting the backbone of 3 ss and 3 ds nucleotides (nts), maintaining a continuously stacked conformation of the RNA bases that is compatible with a hybridized dsRNA. This structure provides a molecular explanation for the long-established preference and increased activity of many DEAD-box helicases on dsRNAs with short ssRNA overhangs. The author then shows that indeed also DbpA yields higher helicase activity on overhang-containing substrates, but that after an optimum, further increases in overhang length reduce unwinding rates again. Given that 3 ss nts were consistently found to be bound by DbpA, it would be interesting to see whether such substrates would actually give an even higher helicase activity than the tested 2 ss nt overhang substrate.

    Another structure with a ssRNA product is elucidated for comparison. It shows the same unwound conformation of ssRNA with a contorted backbone that is incompatible with ds hybridization, which has in the past been observed in several other RNA-bound DEAD-box helicase structures. The main difference to the dsRNA substrate lies in the binding of nts #5 and #6 which are flipped out of the base stacking conformation. Interestingly, in three of the six protomers crystallized per assymmetric unit the ssRNA takes up the same stacked conformation as in the ds substrate bound state. This observation could indicate an intrinsic binding flexibility for both RNA conformations in DbpA. However, it would be important to show that the RNA in this region is not involved in crystal packing in any of the three structures, which could influence/distort its spatial arrangement.

    Finally, EPR measurements are used to show that DbpA also samples the catalytically active, closed conformation in absence of bound ATP(-analoga) and substrate RNA, which could indicate that a preformed active state can even exists before ATP/RNA binding.

    Where the paper is currently falling a bit short is in explaining why the double stranded portion of the RNA should melt. Where does the energy come from? This is one outstanding question, that was not possible to answer in the past without the current structural information. The suggested passive "fraying" is vague and not fully convincing. Just because ATP binding/hydrolysis can happen in the presence of already ssRNA ("futile cycle") this does not mean that the distorted/unstacked binding mode of ssRNA is not induced upon ATP binding/hydrolysis. Actually, one could argue to the contrary, that the high ATPase rates in the presence of ssRNA are suggesting that ssRNA poses less "resistance" to the coordinated motions that have to occur in the RNA and ATP binding sites to yield triphosphate binding/hydrolysis together with the distorted RNA conformation, thus making this state energetically more favorable and more populated (=higher kcat).

    Here, a more extensive comparison of available DEAD-box helicase structures from different states of the catalytic cycle could potentially be instructive (with a focus on the coordination environment of the RNA backbone for nucleotide 5 and its connections/relay to the ATPase active site). Furthermore, a comparison of binding affinities for different RNA species (e.g. ssRNA, dsRNA with 0/1/2/3nt ssRNA overhangs) in the presence of different ATP hydrolysis intermediate analoga (ADP-BeF3, ADP+PO4, ADP+vanadate, ADP) would be useful.
    Importantly, throughout the paper is a joy to read, it shows carefully controlled experiments with high quality data that is concisely described and clearly illustrated. The subject matter is highly interesting for a wide range of biochemists, and structural biologists, and of fundamental importance for RNA biology.

  4. eLife

    Reviewer #3 (Public Review):

    Using NMR a non-hydrolysable ATP analogue is identified that can stall substrate complexes. Then the author reports crystal structures with a hairpin substrate complexes and a single-stranded product complex all attached to the HP92 RNA that he has previously shown to be recognized by the RRM and one of the RecA domains. By comparing the crystallographic structures with previously reported structures very similar conformations are observed. However, two types of closed single-stranded substrate conformations are seen which suggests a mechanism of RNA substrate binding and unwinding coupled to ATP hydrolysis. The role of 5' single-stranded overhang regions is explored by biochemical activity assays as a function of 5' overhang. Finally, using RNA substrate spin labelled at two positions NMR PREs observed are consistent with transient samples of the substrate-bound closed structure already in the absence of nucleotide.

    The work is technically sound and reports an interesting combination of crystallographic structures, supported by biochemical activity assays and NMR to identify transient interactions. The conclusions are well supported by the data shown. An interesting model is proposed for unwinding mechanisms by RNA helicases depending on 5' single-stranded overhangs, that may be of more general relevance.

  5. Version published to 10.1101/2022.02.23.481582v1 on bioRxiv