of 17

Edinburgh Research Explorer

0 views17 pages

Download

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Edinburgh Research Explorer Conflict RNA modification, host-parasite co-evolution, and the origins of DNA and DNA-binding proteins1 Citation for published version: McLaughlin, PJ & Keegan, LP 2014, 'Conflict
Edinburgh Research Explorer Conflict RNA modification, host-parasite co-evolution, and the origins of DNA and DNA-binding proteins1 Citation for published version: McLaughlin, PJ & Keegan, LP 2014, 'Conflict RNA modification, host-parasite co-evolution, and the origins of DNA and DNA-binding proteins1', Biochemical Society Transactions, vol. 42, no. 4, pp Digital Object Identifier (DOI): /BST Link: Link to publication record in Edinburgh Research Explorer Document Version: Peer reviewed version Published In: Biochemical Society Transactions General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact providing details, and we will remove access to the work immediately and investigate your claim. Download date: 09. Aug. 2019 Enzymatic RNA modification, host-virus conflicts and the origins of DNA and DNAbinding proteins Paul J. McLaughlin 1 and Liam P. Keegan 2,3,4 1 Institute of Structural and Molecular Biology, Michael Swann Building, School of Biological Sciences, Kings Buildings 2 Department of Molecular Biology and Functional Genomics Stockholm University S Stockholm, 3 Centre for Integrative Physiology, Hugh Robson Building, George Square, Edinburgh, EH8 9XDThe University of Edinburgh, Edinburgh, UK. 4 Corresponding author Telephone number Fax number addresses: Character count: 50,345 Running title: ADARs and immunity Abstract Nearly one hundred and fifty different enzymatically modified forms of the four canonical residues in RNA have been identified. For instance, enzymes of the ADAR (adenosine deaminase acting on RNA), family convert adenosine residues to inosines in cellular dsrnas. Recent findings show that DNA Endonuclease V enzymes have undergone an evolutionary transition from cleaving 3 to deoxyinosine in DNA and ssdna to cleaving 3 to inosine in dsrna and ssrna in humans. Recent work on dsrna-binding domains of ADARs and other proteins also shows that a degree of sequence-specificity is achieved by direct readout in the minor groove. However the level of sequence specificity observed is much less than that of DNA major groovebinding helix-turn-helix proteins. We suggest that the evolution of more sequence-specific binding proteins following the opening up of the major groove by the RNA to DNA genome transitions represents the major advantage that DNA genomes have over RNA genomes. We propose that a hypothetic RNA modification, a ribose reductase acting on genomic dsrna (RRAR) must have existed. We discuss why this is the most satisfactory explanation for the origin of DNA. The evolution of this RNA modification and later steps to DNA genomes are likely to have been driven by cellular conflicts with viruses. RNA modifications continue to be involved in host-virus conflicts; in vertebrates edited cellular dsrnas with inosine-uracil base pairs appear to be recognized as self RNA and to suppress activation of innate immune sensors that detect viral dsrna. Introduction ADAR (adenosine deaminase acting on RNA) enzymes deaminate adenosine bases to inosines in cellular dsrnas (Keegan et al., 2004). ADARs evolved from adenosine deaminases acting on trnas (ADATs), that convert adenosine bases to inosine in trna anticodons. Inosine at position 34 in trna anticodons permits a wider range of possible base-pairings and facilitates wobble decoding at the third position in eight trna types in Eukaryotes and one trna in E. coli. ADARs evolved from ADATs partly by addition of dsrna-binding domains. Inosine behaves like guanosine in Watson-Crick base pairing during cdna synthesis and translation so that ADAR editing of dsrna is detected as A to G base changes between genomic and cdna sequences. Pre-mRNA editing by ADARs often leads to changes codon meaning and production of edited isoforms of ion channel subunits and other CNS proteins in vertebrates and especially in Drosophila. Recent findings show that inosine-uracil (I-U) base pairs introduced by ADAR editing in vertebrate cellular dsrna (Fig. 1 A), also assist innate immune discrimination between self and non-self RNAs (Karikó et al., 2005; Vitali and Scadden, 2010). Animal cells use a range of Pattern Recognition Receptors (PRRs) to distinguish viral RNAs from cellular RNAs, activating interferon signalling in response to viral RNAs (Takeuchi and Akira, 2010). The mechanisms by which modified bases such as inosine in cellular RNAs modulate the activities of antiviral PRRs is an important area for future work and more understanding of modified base recognition by proteins is required. Inosine is only one of approximately one hundred and fifty known enzymatic modifications in RNA (Machnicka et al., 2013). Many of these were first discovered in trnas and rrnas but 5-meC and 6-meA have now been found to be widely distributed and present at specific locations in many mrnas. An evolutionary shift from DNA cleavage to RNA cleavage by Endonuclease V enzymes recognizing inosines or deoxyinosines Endonuclease V in bacteria is a DNA repair enzyme that cleaves phosphpdiester bond 3 to deoxyinosines arising from deamination of adenosines. Endo nuclease V flips out the deoxyinosine base into a recognition pocket(figure 1 B) (Dalhus et al., 2009). Endo V retains the cleaved DNA product after the reaction probably to allow ordered recruitment of repair the strand cleavage and remove doexyinosine. However the human homolog of Endo V does not cleave DNA 3 of deoxyinosine. In fact, surprisingly, human Endo V cleaves dsrna and ssrna 3 to inosines (Morita et al., 2013; Vik et al., 2013). Human Endo V does also bind at deoxyinosine in DNA but, requires a ribose 2 hydroxyl on the adjacent base which is involved in the cleavage reaction, the DNA (Vik et al., 2013). It will be interesting to establish when Endo V specificity switched between dsrna and DNA. The switch seems to be less complete in mouse Endo V than it is in human Endo V. Possibly the increased level of editing in primates compared to mice that is associated with dsrna formed by pairing of highly sequence similar Alu repeats has helped to recruit a new ribonuclease to help clear these edited RNA. Bacterial Endo V also recognizes a wide range of base mismatches and aberrant structures in DNA. Human Endo V might also cleave dsrnas having other modified bases or structural features. The switch between DNA and RNA substrates is rather simple for Endo V proteins because they use the minor groove to recognize base mismatches. In dsrna the movement of the phosphate strands away from one another is stopped by a steric clash in the minor groove between the ribose 2 hydroxyl and the ribose of the next base on the same strand (Fig. 2 A,B) (Dickerson et al., 1982). This steric clash brings the phosphate strands closer together over the major groove and makes the major groove narrower and much deeper in A- helix RNA than in the major groove in the more familiar B-helix of DNA. The B-DNA strand is narrower, the base edges in the major groove are brought closer to the surface and the major groove is wider (Fig 2 A,B). DNA is still able to adopt the A-helix conformation of dsrna during transcription and during replication by DNA polymerases. In the case of Endo V nuclease the transition between DNA and dsrna substrates may involve a recent change in preference towards dsrna, however, during the evolution of life many minor groove binding proteins may have undergone a switch form recognizing dsrna to recognizing DNA. It has been proposed that ribonucleotide reductases arose only at a late stage in molecular evolution, as genome-encoded proteins. Ribozymes are unlikely to have carried out the reduction of ribose earlier because the catalytic mechanism involves radicals that are difficult to control and damaging to the nucleic acid of a ribozyme (Poole et al., 2002). In now clear also that binding to the minor groove in DNA is still the mechanism used by a surprising number of key genomic processes. If these minor groove processes evolved in RNA genomes then the sizes and coding capacities achieved by ancient RNA genomes may have been much greater than we have anticipated. When the structure of a DNA target complex of the major DNA mismatch-recognition protein MutS was determined, it was found that the protein does not use the major groove to recognize the aberrant base, even though the capacity to distinguish between correct and incorrect bases should be higher in the major groove (Fig. 1 A) (Seeman et al., 1976). Instead MutS interacts with the minor groove of DNA (Lamers et al., 2000). This observation of minor groove interaction has since been extended to some other DNA mismatch recognition proteins. Minor groove dsrna sequence recognition by ADAR double strand RNA binding domains (dsrbds), versus major groove sequence recognition by helix-turn-helix proteins Sequence specificity of double-stranded RNA-binding domain sequences binding long A-helical dsrna was expected to be negligible. Specific adenosine recognition by ADAR dsrbds was thought to depend mainly on the unique molecular shapes of the range of different RNA hairpins in which edited adenosine residues occur in edited pre-mrnas. However, Fredrick Allain and colleagues in Zurich now call attention to the fact that there is some sequence specificity in binding of ADAR2 dsrbds to regular dsrna (Masliah et al., 2013; Stefl et al., 2010; Stefl et al., 2006). The ADAR2 dsrbd2 helix α1 lies along the dsrna minor groove and makes sequence-specific contacts (Fig. 2 A,B). Other dsrbds such as ADAR2 dsrbd1 similarly exhibit some sequence specificity. In the minor groove the guanine amino group is the most distinctive base-specific feature and is particularly important for sequence specific recognition as seen in the dsrna complex of the A. aeolicus RNaseIII dsrbd (Masliah et al., 2013). Sequence specific activities of RNaseIII on different substrates are strongly affected by positioning of guanine bases as positive and negative determinants. Multiple dsrdbs are present in many dsrna-binding proteins and the two dsrbds in ADAR2 combine to allow increased sequence specificity. ADARs also obtain significant further sequence specificity from the binding of the catalytic deaminase domain (Goodman et al., 2012), positioned to adjacent to dsrbd2 on dsrna (Rice et al., 2012), which contacts the target adenosine and probably also some preferred flanking the adenosine. The sequence specificity of minor groove binding by dsrdbs must be contrasted however with the much more impressive sequence specificity achieved by DNA-binding proteins. In the dsrdb-dsrna complex the protein domain steps over the deep and inaccessible major groove, making contacts only with the phosphate backbones. The lambda repressor DNA-binding domain contains a version of the ancient helix-turn-helix (HTH) fold that is one of the oldest structures used by proteins to recognize specific DNA sequences. The second, recognition alpha helix of the helix-turn-helix fold in lambda repressor penetrates the major groove of the DNA in the complex with a lambda phage operator DNA (Fig. 2A,B right side), (Jordan and Pabo, 1988). A much larger number of sequence-specific contacts are made by amino acid side chains of the lambda repressor recognition helix with the bases in the major groove than ADAR2 dsrbd makes in the minor groove of dsrna. This difference is due to intrinsic features of dsrna and DNA. The major groove is superior to the minor groove in allowing base sequences to be discriminated by proteins making hydrogen bonds and other contacts with exocyclic amino and methyl groups (Fig. 1) (Seeman et al., 1976). Different base pairs are much less distinguishable from the minor groove sides of the bases. HTH motifs are present in ancient promoter-proximal DNA binding gene-specific regulatory protein such as proteins MYB that are also conserved between eukaryotes and archaea. Although the HTH motifs and the details of how they interact with the major groove differs between eukaryotic and bacterial gene-specific regulators the derivation of the DNA binding domains from common ancestors is not in doubt. A modified RNA genome intermediate in the transition from RNA genomes to DNA genomes. The major groove in dsrna can be accessed by proteins when it is widened by unusual base pairing or by an extensive supporting folded RNA structure (Hermann and Westhof, 1999). Protein access to the major groove in a long genomic dsrna (Steitz, 1993), would require a significant structure alteration that would be difficult to maintain in a replicating molecule. The helix-turn-helix (HTH) major groove-binding proteins had evolved already when eukaryotes and bacteria separated (Aravind et al., 2005), suggesting that DNA must have evolved already also. However, evolutionary comparison studies on DNA polymerases show a surprising diversity in the origins of these proteins, suggesting that complete DNA genomes appeared after the separation of bacteria from eukaryotes or perhaps even several times independently in prokaryotes and eukaryotes (Forterre, 2006; Forterre and Grosjean, 2000, 2009). We suggest that DNA bases first arose as modified RNA bases at key points in genomic dsrna and that full DNA genomes arose only much later. A hypothetical ribose reductase acting on genomic dsrna (RRAR) could have converted RNA bases to DNA bases in a dsrna genome. This would have allowed the evolution of major groove-binding HTH-motif proteins to evolve before complete DNA genomes were formed. The local modification process envisaged for the RNA genome would have operated rather similarly to current DNA methylation Most of the difficulty in explaining how the RNA to DNA genome transition came about is due to the need to base the explanation on the current actions of ribonucleotide reductase producing free monodeoxynucleotides. If DNA arose at a late evolutionary time in cells that were already complex, then the evolutionary search process that discovered DNA will need to have been carried out directly on the RNA genome itself. It is likely that many enzymatic modifications of genomic dsrna were tested over the course of evolution till the one modification was discovered that offered a significant advantage. New enzymatic RNA modifications are likely to have appeared over the whole history of life and many may have disappeared again if they found no situation in which they offered a benefit. To understand why studying RNA modification leads to the conclusion that DNA must have first evolved as an enzymatic modification of RNA it is helpful to consider the example of the queueosine (Q) base modification in trnas. There are two general types of mechanisms involved in enzymatic modifications of RNA. Most commonly, particular bases in a specific RNA molecule are modified, sometimes in serial multistep processes in which the chemical structure of this single base becomes more and more elaborated from the original canonical genome-encoded base. In a much less common RNA modification process, a complicated modified base such as queueosine (Q) is partly presynthesised off the target RNA. In this case free guanosine is put through a series of modifications to form a prequeueosine base which. This is then substituted for a guanosine at position 34 in some specific trnas by a specific transglycosylase enzyme. Further modifications then occur on the pre-q base after insertion into the trna transcript. Based on our knowledge of the range of postsynthetic enzymatic RNA modifications occurring at position 34 in different trna it is easy to explain the evolution of the substitutional queueosine modification mechanism. The initial evolution of the queueosine base was by the usual type of postsynthetic modification of trna transcripts at position 34. Evolutionary selection for improved accuracy of decoding by the final enzymatically modified trnas defined the required chemical structure for the modified base. The subsequent evolution of a specific transglycosylase allowed the early part of the base modification process to become a separate presynthetic process carried out on a single nucleoside off the RNA. Let us now attempt to carry out a thought experiment in which we imagine that we know nothing about post-synthetic modifications in trnas. All we know is that a particular enzymatically modified form of guanosine is made off the trna and then inserted into a trna where it plays a critical role. How would we explain the evolution of queueosine in this case? This would be very difficult. We could not account for how either the structure of the queueosine base nor how the insertion into trnas specifically had ever managed to evolve. This thought experiment on the evolution of the queueosine base illustrates the same difficulty that we face in trying to explain the origin of DNA. We cannot satisfactorily explain the origin of DNA by starting from the current production of DNA mononucelotides by ribonucleotide reductases. That starting point leaves far too many questions unanswered. A lost or still undiscovered RNA modification that reduced ribose directly in dsrna genomes or caused the loss of the 2 hydroxyl in some way is a more likely explanation for production of the first DNA bases. Different RNA modifications that have been conserved in different RNA classes are very finely tailored to the functional requirements of these RNAs. Many of the modifications conserved in rrnas or trnas provide increased stability or increase the fidelity of translation. Ribosomal RNAs use an alternative RNA modification on the ribose 2 hydroxyl. If introduced in dsrna it might reduce any tendency to hydrolysis of the RNA genome that was caused by the presence of the 2 hydroxyl on ribose (Poole et al., 2000). Most importantly however ribose 2 methylation does not remove the 2 hydroxyl. The steric clash that forces dsrna into the A-helix conformation would still occur and formation of the B- helix as it does with deoxyribose. Ribose 2 methylation is useful to stabilise rrnas but a deoxyribose is required to give the wide major groove of the DNA helix. Cells will have discovered this by an evolutionary search process. Postsynthetic ribose reduction in stable, folded or catalytic RNAs is likely to have been harmful in ribozymes, rrnas or trnas. In genomic RNA however it immediately conferred some benefit. Viruses may have been first to evolve the ribose reduction to protect their genomes from host defences. Continuing conflicts with viruses are likely to have driven subsequent steps in DNA evolution also. As to the identity of the original ribose reductase acting on genomic dsrna (RRAR), it is worth considering the ribonucleotide reductases (RNRs) first, even if this only serves to outline requirements for this RNA modification activity. Current RNRs act on free ribonucleotide di- or tri-phosphates. Taking the closest possible analogy to the queueosine base discussed earlier one possible RRAR could have been an enzyme ancestral to the current Class I and II ribonucleotide reductases (RNRs) that acted on a ribose within an RNA strand. A larger substrate binding pocket would have been needed in the hypothetical RRAR ancestor than in the current RNRs. The C3 carbon of the target ribose is in the deepest part of the active site in Class I and II ribonucleotide reductases and more space will have been required for a continuous RNA strand to exit. The core of the active site would have been conserved in the subsequent conversion to RNRs. Curr
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks