Deciphering the mechanism of processive ssDNA digestion by the Dna2-RPA ensemble

Single-stranded DNA (ssDNA) commonly occurs as intermediates in DNA metabolic pathways. The ssDNA binding protein, RPA, not only protects the integrity of ssDNA, but also directs the downstream factor that signals or repairs the ssDNA intermediate. However, it remains unclear how these enzymes/factors outcompete RPA to access ssDNA. Using the budding yeast Saccharomyces cerevisiae as a model system, we find that Dna2 — a key nuclease in DNA replication and repair — employs a bimodal interface to act with RPA both in cis and in trans. The cis-activity makes RPA a processive unit for Dna2-catalyzed ssDNA digestion, where RPA delivers its bound ssDNA to Dna2. On the other hand, activity in trans is mediated by an acidic patch on Dna2, which enables it to function with a sub-optimal amount of RPA, or to overcome DNA secondary structures. The trans-activity mode is not required for cell viability, but is necessary for effective double strand break (DSB) repair.

This study employs a variety of biochemical, genetic, and single-molecule biophysical methods to shed light on Dna2-RPA synergy. Evidence is furnished that RPA regulates Dna2 processivity. Importantly, the authors document "trans" synergy between Dna2 and RPA that is dependent on a novel "AC motif" in Dna2. Interestingly, a mutation in the AC motif creates a separation-of-function dna2 mutant that is viable but defective in HR intermediate processing. Overall, this study furnishes insights into a hitherto obscure mechanism germane for understanding the processing of HR intermediates and mechanisms of genome maintenance. The analyses have been conducted exceptionally well and the conclusions are supported by the results.
I have just a few suggestions for the authors to consider, as follows.
1. The Introduction can be expanded to better discuss the mechanisms of Okazaki fragment maturation and DNA end resection. Figure 1, the processivity of the RPA-Dna2-ssDNA ensemble is likely independent of the helicase activity of Dna2 as no ATP was present. While the motor activity of Dna2 has been well-characterized in bulk biochemical assays, it would be interesting to see whether the addition of ATP has any impact on Dna2 processivity in the DNA curtain analysis.

In
3. In Figure 3A, the mapped interface between Dna2 and RPA largely resides at the N-terminus and the OB-fold region of Dna2. There is a well-known dna2-2 allele with the P504S mutation. The P504 residue is located right at the hinge area connecting the OB and the nuclease domains of Dna2. It would be interesting to ask whether the P504S mutation affects Dna2-RPA-ssDNA ternary complex formation.
4. The N-terminus of RPA is suggested by the authors to be crucial for complex formation. Besides dna2 mutants, it would be good to also test a Rfa1-N mutant, for instance, the rfa1-t11 mutant. Figure 3D, the hybrid NAB-Dna2∆248N protein provides a tool to secure cis while voiding the trans action. The authors could also test NAB-Dna2∆248N-AC or NAB-Dna2∆405N in the same assay. If the AC domain is needed only for trans-action, then NAB-Dna2∆248N-AC or NAB-Dna2∆405N should behave similarly as NAB-Dna2∆248N.

In
Reviewer #2: Remarks to the Author: This is a very nice manuscript detailing the interactions and effects of same between RPA and Dna2.
The work features a sophisticated application of multiple state of the art methods that all help to reinforce the main conclusions that RPA affects Dna2 processivity and that the interaction is focused on the N-terminus of RPA. There is more. My comments are generally minor.
The summary is too long by a factor of 2. The main points are not encountered until one has read almost 2/3 of it. This will discourage potential readers.
I assume these are all proteins from yeast, but that should be mentioned somewhere early for general readers.
Line 96. What salt is being referred to? It is not mentioned here and it is hard to know what is meant by "salt" at most points in the manuscript. KCl is mentioned late in the text and NaCl is mentioned in some of the figures. Neither is the physiological salt in cells; e.g., the concentration of Cl ion in yeast is generally maintained at < 1 mM regardless of the NaCl or KCl concentration outside the cells. This should be acknowledged and the word "salt" replaced with the actual salt being used in most places.
Reviewer #3: Remarks to the Author: In this is a very intersting manuscript Shen and colleagues combined single-molecule curtains, biochemical approaches and in vivo studies to investigate the interaction between yeast Dna2 DNA helicase and ssDNA binding protein RPA. Imortantly, the authors discovred a region at the N-terminus of RPA1 wich ineracts with Dna2 and allosterically activates Dna2 processive translocaion. This interaction with RPA is also mportant for the Dna2 recruitment to ssDNA under physiological conditions. The reciprocal site (or at least one of the sites), an acidic patch, on Dna2 has also been identified. In addition to Dna2 activation, RPA's more classic role of melting secondary structures of DNA seems important for longer range processivity of Dna2.
Overall, the experiments reportd here are carefully designed and expertly eecuted. The results are convincing and will be important to the fild as they will elevate our understanding of the Dna2-RPA.
A few points: 1. In the model the authors present ( Fig. S6 and corresponding discussion), the potential role of the RPA-Dna2 interaction in removing RPA from ssDNA during processive DNA degadation seems to be overlooked. Similarly, it would be informative if the authors discuss how the activation by RPA works in the context of Dna2 activation by other players in long range DSB resection. It has been shown recently that human CtIP activates DNA2 in long range end resection (Ceppi et al 2020 PNAS); would it be reasonable to expect that a similar interaction exists in the yeast system and how would this fit with RPA-Dna2 interaction?

Our point-by-point response to the Critiques:
We are very pleased that all three reviewers find our work of general interest. We would like to sincerely thank all three referees for their time reviewing our manuscript and their excellent comments to improve our study. As documented below, we strived to address all the comments from the referees and incorporate all their excellent suggestions in the revised manuscript, which, as a result of revision according to the critiques, is stronger than the original version.

REFEREE 1:
This referee noted that "Overall, this study furnishes insights into a hitherto obscure mechanism germane for understanding the processing of HR intermediates and mechanisms of genome maintenance. The analyses have been conducted exceptionally well and the conclusions are supported by the results." The referee made a number of suggestions to help us add new mechanistic insight to the story. We are grateful for these suggestions, and below, we document how we have incorporated them into the revised manuscript.

The referee asked us to expand the introduction section to better discuss the mechanism of Okazaki fragment and DNA end resection.
Our response: We would like to thank the referee for this suggestion, and we apologize for the lack of an introduction section due to the format for the original submission. We reorganized our manuscript to match the format for Nature Communication and to include an introduction section, where the lagging strand maturation and DNA end resection were discussed.

The referee asked us to examine the impact of ATP presence on Dna2 processivity in the DNA curtains analysis.
Our response: We have performed this analysis as instructed and compared Dna2-catalyzed digestion of RPA-GFP coated long ssDNA substrate in the absence or presence of 1 mM ATP. The result, included as Figure S2H, showed that the presence of ATP has no significant impact on Dna2 processivity in DNA Curtains analysis.

Knowing the functional importance of the OB-fold domain in Dna2, the referee asked us to test dna2-2 (dna2-P504S) for ternary complex formation with RPA and ssDNA.
Our response: During the revision process, we generated dna2-P504S construct and purified the mutant protein (Figure S7A(vi)). dna2-P504S had a low yield comparable to dna2-∆501N., Interestingly, purified dna2-P504S is severely defective in the digestion of 5' overhanging DNA regardless of the absence or presence of RPA (Figure S3E(i)). dna2-P504S also failed to support the formation of a stable ternary complex (Figure S3E(ii)), which is consistent with the role of Dna2 OB-fold in the ternary complex formation.
Our response: Again, we would like to thank this referee for the suggestion, and we have specified our system, budding yeast -Saccharomyces cerevisiae, in the revised abstract.
3. The referee asked us to specify the actual salt, viz. NaCl or KCl, used instead of using the general word "salt" since neither is the physiological salt in cells; e.g., the concentration of Cl ion in yeast is generally maintained at < 1 mM regardless of the NaCl or KCl concentration outside the cells. The referee also asked us to acknowledge this.
Our response: We completely agree with the referee and now have the type of salt used specified in the main text as suggested. We also acknowledged in the revised methods under the "Dna2 nuclease assays" section that neither KCl nor NaCl represents the physiological salt in cells.

REFEREE 3:
This referee noted that "Overall, the experiments reported here are carefully designed and expertly executed. The results are convincing and will be important to the field as they will elevate our understanding of the Dna2-RPA." This referee made a few comments to help us improve the manuscript, especially the discussion of our findings. We are truly grateful for these comments and now have them incorporated into the revised manuscript as detailed below.
1. The referee noted that the potential role of the RPA-Dna2 interaction in removing RPA from ssDNA during processive DNA degradation seems to be overlooked in the discussion, and also suggest we discuss how the other players in DNA end resection, such as CtIP/Sae2, may affect the function of Dna2-RPA pair.
Our response: We appreciate these suggestions and now have them incorporated into the revised discussion section. Indeed, our data from analyzing the NAB-dna2-∆248N fusion did suggest that the cis action within the Dna2-RPA ensemble can remove RPA from ssDNA likely through DNA degradation. However, knowing that the fusion protein fails to turnover from long ssDNA, dissembling of the ternary complex upon the completion of the ssDNA digestion may require the trans action between the AC motif and the free RPA molecule. We also added a section to discuss the potential impact of other resection factors including CtIP/Sae2 on RPA-Dna2 interaction. Although the reported stimulation of human DNA2 by CtIP is ATP-dependent and likely relies on the motor activity of DNA2, the Sgs1 helicase, which also interacts with Dna2, on the other hand, possesses an acid patch that has been shown to interact with Rfa1 (Hegnauer, EMBO J, 2012). Given the resection defects that we observed with dna2-AC, it will be interesting to further investigate the interplay of AC motifs from Sgs1 and Dna2 with Rfa1-N during DNA end resection.
2. The referee asked us to elaborate and clarify the RPA dynamics mentioned in the discussion.
Our response: We apologize for this confusion. Indeed, we were referring to the dynamic interactions between RPA and its partners as this referee pointed out. We have added a sentence to clarify this in the discussion section as "Our study thus provides a new paradigm on recognizing the dynamic nature of RPA, where, through different modes of interaction, RPA can actively deliver its bound ssDNA to downstream enzymes/factors and coordinate with them to further resolve DNA secondary structures encountered." 3. To confirm the dual role of RPA in the long-range nuclease activity of Dna2, the referee asked us to examine the digestion of a long ssDNA overhang without secondary structures or combine NAB-dna2-∆248N with a heterologous RPA.
Our response: We thank the referee for this suggestion. As shown in Figure 4D (first panel), removal of the internal hairpin located 80-90 nt from the labeled 3' end released the pausing at this position observed in Figure 4C, which affirmed the role of trans action between Dna2 and RPA in overcoming DNA structures. We also tested heterologous RPA with NAB-dna2-∆248N fusion as shown below. While free yeast RPA can facilitate the fusion protein to overcome the structural barrier, human RPA or E.coli SSB strongly inhibits the NAB-dna2-∆248N, which makes the interpretation difficult and we chose not to include this experiment in the revised manuscript.