Comprehensive phylogenetic analysis of the ribonucleotide reductase family reveals an ancestral clade

  1. Audrey A Burnim
  2. Matthew A Spence
  3. Da Xu
  4. Colin J Jackson
  5. Nozomi Ando

  • Curated by eLife

    eLife logo

    Evaluation summary

    Ribonucleotide reductases (RNRs) have fascinated biologists and chemists, as these enzymes catalyze the conversion of ribonucleotides (NDPs or NTPs) to deoxynucleotides (dNDP or dNTPs), which are essential for DNA biosynthesis in all organisms. Given this role, they have been postulated to be the link in the transition from an RNA/protein to a DNA world. In addition, RNRs use an array of protein, metal-based, and nucleotide radicals for the reaction they catalyze. This paper creatively combines two methods of analysis to propose a new evolutionary model for the diversification observed for the RNR family into the three classes: I, II and III. The work is of interest to students of molecular evolution, RNRs and colleagues interested in the origin of life.

    (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 #1, Reviewer #2 and Reviewer #3 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

Ribonucleotide reductases (RNRs) are used by all free-living organisms and many viruses to catalyze an essential step in the de novo biosynthesis of DNA precursors. RNRs are remarkably diverse by primary sequence and cofactor requirement, while sharing a conserved fold and radical-based mechanism for nucleotide reduction. Here, we structurally aligned the diverse RNR family by the conserved catalytic barrel to reconstruct the first large-scale phylogeny consisting of 6779 sequences that unites all extant classes of the RNR family and performed evo-velocity analysis to independently validate our evolutionary model. With a robust phylogeny in-hand, we uncovered a novel, phylogenetically distinct clade that is placed as ancestral to the classes I and II RNRs, which we have termed clade Ø. We employed small-angle X-ray scattering (SAXS), cryogenic-electron microscopy (cryo-EM), and AlphaFold2 to investigate a member of this clade from Synechococcus phage S-CBP4 and report the most minimal RNR architecture to-date. Based on our analyses, we propose an evolutionary model of diversification in the RNR family and delineate how our phylogeny can be used as a roadmap for targeted future study.

Article activity feed

  1. Version published to 10.7554/elife.79790 on eLife
  2. eLife

    Evaluation summary

    Ribonucleotide reductases (RNRs) have fascinated biologists and chemists, as these enzymes catalyze the conversion of ribonucleotides (NDPs or NTPs) to deoxynucleotides (dNDP or dNTPs), which are essential for DNA biosynthesis in all organisms. Given this role, they have been postulated to be the link in the transition from an RNA/protein to a DNA world. In addition, RNRs use an array of protein, metal-based, and nucleotide radicals for the reaction they catalyze. This paper creatively combines two methods of analysis to propose a new evolutionary model for the diversification observed for the RNR family into the three classes: I, II and III. The work is of interest to students of molecular evolution, RNRs and colleagues interested in the origin of life.

    (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 #1, Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    Ribonucleotide reductases (RNR) share low sequence similarity, which makes it challenging to infer their phylogeny with traditional methods. To accurately decipher their evolutionary history and evolutionary relationships between different clades the authors combined a structure-based workflow developed by Spence et al. (Reference 1) and a state-of-the-art evo-velocity analysis. Thus, they present a convincing phylogenetic map of RNR, among which they found a clade Ø unknown before and determined its cryo-EM model. One strength of this study is that the analysis pattern utilized in this paper can give a good example of the analysis of protein families which are highly diverse in sequence but share an overall conserved structure core, and thus this analysis pipeline may be implemented in other protein families. The weakness of this study is that the catalytic function of RNRs from the novel clade Ø is not well characterized, such as the ferritin-like domain. It would be interesting to design comprehensive biochemical experiments on this novel clade RNR and maybe the authors will do that in the near future. On the basis of the large-scale phylogeny of RNR, the authors studied three extension/insertion regions of RNR, including N-terminal ATP-cone, C-termini of class II RNR and finger-loop-motif of Class III RNR. These discoveries systematically reveal the plasticity and evolvability of RNRs, which may lead to a model to depict the complete evolutionary history of RNRs. Another weakness is that the descriptions on these extensions/insertions are somewhat scattered and lack a summary model/illustration/table to unify all discoveries. Overall, this manuscript can promote the understanding of RNRs and protein evolution, and the methodology utilized in this study may offer a reference of other diverse protein families.

  4. eLife

    Reviewer #2 (Public Review):

    This paper provides a unique evolutionary model for the striking diversification of RNRs using compelling arguments based on a phylogeny using common elements of α subunit structures and a common chemical mechanism for NDP reduction, supported by the recently reported "evo-velocity" method.

    The strength of the paper is in pages 1-9.5 and Figures 1-4: Pages 1-9.5 present the authors new approach to understanding the evolution of the three classes of RNR (Figure 2). Class I uses a dimetallo-tyrosyl radical (diiron or dimanganese) or dimetallo-cofactor (iron-iron, iron-manganese with no protein radical), class II uses adenosylcobalamin (cobalt-dependent) and class III uses a glycyl radical cofactor generated by an Activating Enzyme that uses an 4Fe4S cluster and S-adenosylmethionine. This diversity and the requirement for small molecule or second protein subunit (β) (Figure 2) for the radical initiation process, confused the literature early on. The authors suggest that their analysis methods have allowed discovery of a new clade they call "oh" (/O). This new clade describes the relationship between the three classes, in my opinion, in a chemically and biologically very satisfying way. They first use an ensemble of hidden Markov models guided by substantial structural information on the large (common) subunit α of all RNRs (α, a 10 stranded α/β barrel with a finger loop in the center of the barrel, as shown in Figure 1) which houses a cysteine that becomes the thiyl radical that initiates catalysis in all three classes. Their generic model (describes insertions and deletions within α, and is shown in Figure 1A of their manuscript). This approach provided an accurate alignment of approx 7000 RNR sequences, which has previously been very challenging despite thousands of sequences in genomes, due to very limited sequence homology thought to be associated in part with horizontal gene transfer (viruses and hosts). This phylogenic method based on structure is then partnered with a new method called "evo-velocity", described by Hie et al and Hie and Kim et al (potential reviewers) which does not depend on multiple sequence alignments. The sequence data set used for the evo-velocity analysis is the same set of information used in the author's phylogenetic reconstruction using structure. The two methods together provide the basis for their new approach. Their studies suggest that the class III is the most ancestral RNR. They discovered a "new" (/0) clade and propose that the class I and II RNRs evolved from this new clade. They suggest that α in (/0) clade is the simplest of all αs (426 amino acids vs 759 amino acids in the most studied E coli class Ia α). Thus, the while the class III is the oldest of the RNRs (which has been and continues to be debated even recently), the author's model proposes that the last common ancestor from class III involves classes /0, I, and II with the subsequent divergence of class I and II from /0.

    The analysis of Burnim et al, in my opinion, allows the community to rationalize "more easily" the evolution of the three RNR classes. While we will never know the answer to this evolutionary question, their analysis is the first use of structures (Figure 1a) and mechanistic chemistry which we currently understand reasonably well, which allows a very satisfying picture to emerge. I need to clarify that I am a chemist/biochemist and do not really know very much about evolution. However, the inability to use sequence alignments alone to build phylogeny due to minimal sequence conservation has made understanding of evolution of these enzymes quite challenging to investigate. It turns out the first structures of RNR (E. coli Ia RNR and the L leichmannii class II RNR are almost superimposable, but one could not identify from efforts to align sequences. even the three cysteines in the active site, until we knew where to look.

    I would strongly suggest that this paper be reviewed by Sjöberg and her collaborators in Sweden who have been interested in this question for decades. Their papers were always a challenge for me, and the agreement, whether class III or class II was the most ancient RNR was continually being debated. I think they will be very excited by the analysis presented in this paper and perhaps can articulate how the Ando et al paper makes the picture of RNR class relationships clearer. In my opinion, the power of structure and chemistry is apparent from the authors' data presented.

    Weakness. Through their careful analysis, the author's learned much additional information about RNRs and helped the community to think about the evolutionary question of the classes in unique ways (pages 1-9.5). However, in pages 9.5 to 21, the paper enters into additional fascinating, but complex analyses. They focus on the the origin of the 100 amino acid ATP cone domain involved in regulation of RNR activity in distinct ways which are found in all three RNR classes (p 9.5-14). The cone domain is also found in NrdR transcription factors. Recently two groups have found that this domain can bind more than one nucleotide. At any rate, the domain story is very interesting, but its function and nucleotide regulation still has many unresolved issues. Their analysis as presented changes our thinking about this domain and its role in RNR which controls their activity by altering the proteins quaternary structures (organism specific and complex). Their results while interesting, to the uninitiated to the RNR world, becomes complex very rapidly and hard to digest. I suggest that their results would be better described in an independent publication with additional background.

    The second and third topics introduced (14-17 and 17-19) involve the C-terminal domains of the class II RNRs and the issue of the finger loop-motif in the class III RNRs, respectively. Both topics are very interesting from the discoveries the authors have made in their new analysis approach, but again, require much more background. In addition,some of the Figures need better figure legends and more contrasting colors to maximize points they want to make. The data is again very interesting and important, but will be hard for the uninitiated in RNRs to follow. More background is required.

  5. eLife

    Reviewer #3 (Public Review):

    In this manuscript it has been found that there is a deeply diverged ribonucleotide reductase class that can potentially be the ancestor of both class I and class II ribonucleotide reductases. Furthermore, the structure of a representative member of the new class was characterized with cryo-EM and SAXS. I found the manuscript very interesting and of high relevance. A weakness though was that I did not see anything written about enzyme activity and if the small subunit contains any free radical in the manuscript, which means that we cannot be sure that it really is a ribonucleotide reductase although the homologies and the ability of dTTP to induce dimerization is a strong indicator of that.

    Another conclusion in the manuscript was that the last common ancestor of the ribonucleotride reductase classes had the ATP cone-mediated allosteric regulation that we see in approximately half to the ribonucleotide reductase today. However, although the analysis presented is interesting, I think that it is still an open question whether the last common ancestor had an ATP cone or not. Many species contain more than one class of ribonucleotide reductase and because it is a mobile element, it can easily jump from one class to another.

  6. Version published to 10.1101/2022.04.23.489257v1 on bioRxiv