High-fidelity DNA ligation enforces accurate Okazaki fragment maturation during DNA replication

DNA ligase 1 (LIG1, Cdc9 in yeast) finalizes eukaryotic nuclear DNA replication by sealing Okazaki fragments using DNA end-joining reactions that strongly discriminate against incorrectly paired DNA substrates. Whether intrinsic ligation fidelity contributes to the accuracy of replication of the nuclear genome is unknown. Here, we show that an engineered low-fidelity LIG1Cdc9 variant confers a novel mutator phenotype in yeast typified by the accumulation of single base insertion mutations in homonucleotide runs. The rate at which these additions are generated increases upon concomitant inactivation of DNA mismatch repair, or by inactivation of the Fen1Rad27 Okazaki fragment maturation (OFM) nuclease. Biochemical and structural data establish that LIG1Cdc9 normally avoids erroneous ligation of DNA polymerase slippage products, and this protection is compromised by mutation of a LIG1Cdc9 high-fidelity metal binding site. Collectively, our data indicate that high-fidelity DNA ligation is required to prevent insertion mutations, and that this may be particularly critical following strand displacement synthesis during the completion of OFM.


6)
legend should include a statement identifying the source organisms used in the alignment of ligase aa sequences. 7) Fig. 1a: the label "Sc_Cdc17" seems like a mistake. I suppose the authors are referring to S. pombe Cdc17 (DNA ligase, which should be labeled Sp_Cdc17) and not S. cerevisiae Cdc17 (catalytic subunit of Pol apha-primase).
Reviewer #2 (Remarks to the Author): DNA ligase 1 provides a vital function in cells as it ligates processed Okazaki fragments during lagging strand synthesis. Surprisingly, the fidelity of Ligase 1 (LIG1) has not been studied at a mechanistic level. This tour-de-force study combing genetic, structural biology and biochemical results, by a team of outstanding scientists, whohave investigated the effects of a LIG1 variant in which two key glutamate resides have been substituted to alanines. These Glu residues are believed to be important for high-fidelity metal binding. Using the URA3 reporter gene assay, the authors first show that this LIG1 variant shows a slight increase in mutation frequency, which when combined with altered mismatch repair or FEN1 (Rad27) nuclease shows a synergistic increase in mutation frequency. The spectra of mutations show increased +1 insertion in homonucleotide runs. Further analysis suggest that these are occurring on the lagging strand. The authors then show in the absence of MSH2 and RAD27 yeast carrying the EE/AA LIG1 variant do not grow. Based on these data the authors present a working model for LIG1 fidelity and then present and biochemical and elegant structural data supporting this model. Specifically they biochemically that bringing a bulged C close to the ligatable nick diminished ligation by the WT LIG1,which was arrest at the 5-AMP state, but show that removal of the two EE residues in the variant allows ligation of the bulge. Furthermore control experiments show that both DNA ligases have relative equal activity on a nicked substrate. They further show structural that EE to AA substitutions creates a pocket that can accommodate a flipped out residue and allow for replication slippage and subsequent +1 insertional frameshift. This is an excellent study that provides an important new and innovative data providing mechanistic insight into LIG1 fidelity that should have a long and lasting impact on the field. There are very few problems with this work, the following issue needs clarification. Figure 1, Panel C shows the accumulation of mutations on the coding strand which is the leading strand, but then the authors argue that the mutations were arising on the lagging strand in Figure  2. The logic on lines 148-150 makes sense, but since the authors used PCR fragments to sequence the mutations in the URA3, it is not clear how they can specifically assign specific mutations to one strand or the other. Please clearly describe in the methods. It may be confusing to reader to show mutational spectra on the leading strand in Figure 1 and in Supplemental figures, but then argue that these arose on the lagging strand. Would redrawing these on the oppositve strand make more sense. Is there a specific Pol delta mutant that could be used to prove this hypothesis? Despite the beautiful renderings in Figure 6, an added supplemental movie that shows the WT structure morphing into the variant structure and subsequent accommodation of the bulged C would be very helpful to the reader.
Reviewer #3 (Remarks to the Author): The manuscript by Williams et al, makes an excellent case for the need for precise, high-fidelity ligation during Okazaki fragment maturation in order to maintain the fidelity of the genome. Comparing ligation efficiency of the wild-type and a mutant variant which exhibits low fidelity, the authors show that a mutation in the high-fidelity metal binding site of LIG1 allows for promiscuous ligation and introduction of insertion mutations. This work is a natural extension of their previous work characterizing the metal binding site of human LIG1. Moving into the yeast model, the authors have relied on their robust reporter assays to show the importance of the high-affinity metal binding site in coordinating efficient ligation. When the metal binding site was altered, the mutant variant displayed an increased amount of single base addition mutations compared to the wild-type strain, with mutations clustered mostly in GC runs. Deletions of both flap processing enzyme, Rad27 and proteins from the mismatch repair pathway (MSH2) accentuated the insertions in the LIG1 mutant variant strain. Biochemical assays reveal that the mutant variant was more efficient at ligating bulged (+1) substrates compared to the wild type ligase. Finally, the authors provided a crystal structure for the mutant variant and showed how the bulge in the substrate is accommodated in a specific pocket that would have normally been occupied by the Mg2+. The genetic experiments clearly show an in vivo impact of altering the high-fidelity metal binding site. These findings are ably supported by both biochemical experiments and the crystal structure of the mutant ligase. The data presented are of high quality and the authors have provided good explanations for their observations. Overall, the manuscript provides an excellent model for the high-fidelity metal binding site and the role it plays in maintenance of genome stability.
Minor comments: 1. The authors considered +1 additions to be a result of either deficiency in MMR or caused during flap processing by Rad27 (and/or Dna2). Would the authors expect higher rates of +1 insertions in an exonuclease deficient strain of DNA polymerase delta? 2. Ligation assays: Supplementary Figure 6: The ligation efficiency of the WT-LIG1 seems to vary drastically on the substrates. For example, almost 90% of the substrate is ligated in substrates 1c and 2c, whereas ~ only 10% ligation of other substrates (3c, 6c, 7c) are observed. Since similar concentration of enzyme was present in each reaction, one would expect the ligation efficiency to be similar in each substrate. Could substrate breathing be a reason for this varying ligation efficiency? Similarly, for Figure 5, could the bulge destabilize the 3' annealing and thus not provide a stable nick for the ligase activity? 3. One can expect the bulge to be also present on the downstream primer (in case of equilibrating flaps). The bulge would probably not be accommodated in a similar manner as shown in the crystal structure. How would this impact ligation efficiency? Fig. 1a: the label "Sc_Cdc17" seems like a mistake. I suppose the authors are referring to S. pombe Cdc17 (DNA ligase, which should be labeled Sp_Cdc17) and not S. cerevisiae Cdc17 (catalytic subunit of Pol apha-primase).

7)
Response: Thank you, this is the S. pombe protein; Sp_Cdc17.

Reviewer #2:
DNA ligase 1 provides a vital function in cells as it ligates processed Okazaki fragments during lagging strand synthesis. Surprisingly, the fidelity of Ligase 1 (LIG1) has not been studied at a mechanistic level. This tour-de-force study combing genetic, structural biology and biochemical results, by a team of outstanding scientists, who have investigated the effects of a LIG1 variant in which two key glutamate resides have been substituted to alanines. These Glu residues are believed to be important for high-fidelity metal binding. Using the URA3 reporter gene assay, the authors first show that this LIG1 variant shows a slight increase in mutation frequency, which when combined with altered mismatch repair or FEN1 (Rad27) nuclease shows a synergistic increase in mutation frequency. The spectra of mutations show increased +1 insertion in homonucleotide runs. Further analysis suggest that these are occurring on the lagging strand. The authors then show in the absence of MSH2 and RAD27 yeast carrying the EE/AA LIG1 variant do not grow. Based on these data the authors present a working model for LIG1 fidelity and then present and biochemical and elegant structural data supporting this model. Specifically they biochemically that bringing a bulged C close to the ligatable nick diminished ligation by the WT LIG1,which was arrest at the 5-AMP state, but show that removal of the two EE residues in the variant allows ligation of the bulge. Furthermore control experiments show that both DNA ligases have relative equal activity on a nicked substrate. They further show structural that EE to AA substitutions creates a pocket that can accommodate a flipped out residue and allow for replication slippage and subsequent +1 insertional frameshift. This is an excellent study that provides an important new and innovative data providing mechanistic insight into LIG1 fidelity that should have a long and lasting impact on the field. There are very few problems with this work, the following issue needs clarification.
Response: We thank the reviewer for their positive comments on the manuscript and are happy to make the requested changes. Figure 1, Panel C shows the accumulation of mutations on the coding strand which is the leading strand, but then the authors argue that the mutations were arising on the lagging strand in Figure 2. The logic on lines 148-150 makes sense, but since the authors used PCR fragments to sequence the mutations in the URA3, it is not clear how they can specifically assign specific mutations to one strand or the other. Please clearly describe in the methods. It may be confusing to reader to show mutational spectra on the leading strand in Figure 1 and in Supplemental figures, but then argue that these arose on the lagging strand. Would redrawing these on the oppositve strand make more sense. Is there a specific Pol delta mutant that could be used to prove this hypothesis?
Response: We apologize for the confusion and appreciate the reviewer's feedback on this important issue. Therefore, we have taken this opportunity to clarify our URA3 mutation spectra figures and explanations. The assignment of these mutations to the lagging strand is based on previous work using both mutation signatures (Pursell et al., Science (2007) (2015)) to define the leading and lagging strands for the URA3 reporter gene in its position adjacent to the closest replication origin, ARS306. These studies were performed using mutator variants of DNA polymerases δ and ε that had been biochemically characterized as having specific mutation signatures and ribonucleotide incorporation propensities. We now include this more detailed description in the text on page 8 lines 150-152, and have also described our logic for assigning mutations to each strand in the Methods section on pages 14-15.
We display the sequence of the URA3 coding strand in the mutation spectra as a matter of convention and to keep our analysis consistent with previous publications where this same reporter system was utilized. Because the +1 insertion mutations that arise in the low-fidelity cdc9-EE/AA mutant strain do not appear to have strict sequence specificity (Fig. 2c), it may be difficult to use a specific DNA polymerase delta mutant or other genetic trick to provide further support of the hypothesis that they are arising during lagging strand synthesis.
Despite the beautiful renderings in Figure 6, an added supplemental movie that shows the WT structure morphing into the variant structure and subsequent accommodation of the bulged C would be very helpful to the reader.
Response: We thank the reviewer for this excellent feedback and have added a supplemental movie (Supplementary Movie 1) that shows a structural morph between the un-bulged and bulged DNA-bound states of mutant LIG1 EE/AA -DNA complexes. The bulged extrahelical nucleotide in the insC complex flips into a pocket created by alanine mutations in the AdD (turquoise) and DBD (grey) domains.