Similar to humans, bacteria store their genetic material in the form of DNA and arrange it into structures called chromosomes. In fact, most bacteria have a single circular chromosome. Bacteria multiply by simply dividing in two, and before that happens they must replicate their DNA so that each of the newly formed cells receives one copy of the chromosome.

Occasionally, mistakes during the DNA replication process can cause the two chromosomes to become tangled with each other; this prevents them from separating into the newly formed cells. For instance, the chromosomes can become physically connected like links in a chain, or merge into one long string. This kind of tangling can result in cell death, so bacteria encode enzymes called Xer recombinases that can untangle chromosomes. These enzymes separate the chromosomes by cutting and rejoining the DNA strands in a process known as Xer recombination.

Although Xer recombinases have been studied in quite some detail, many questions remain unanswered about how they work. How do Xer recombinases interact with DNA? How do they ensure they only work on tangled chromosomes? And how does a protein called FtsK ensure that Xer recombination takes place at the correct time and place?

Bebel et al. have now studied the Xer recombinase from a bacterium called Helicobacter pylori, which causes stomach ulcers, using a technique called X-ray crystallography. This enabled the three-dimensional structure of the Xer recombinase to be visualized as it interacted with DNA to form a Xer-DNA complex. Structures of the enzyme before and after it cut the DNA show that Xer-DNA complexes first assemble in an inactive state and are then activated by large conformational changes that make the DNA bend. Bebel et al. propose that the FtsK protein might trigger these changes and help to bend the DNA as it activates Xer recombination.

Further work showed that the structures could be used to model and understand Xer recombinases from other species of bacteria. The next step is to analyze how FtsK activates Xer recombinases and to see if this process is universal amongst bacteria. Understanding how this process can be interrupted could help to develop new drugs that can kill harmful bacteria.

In organisms with circular genomes, homologous recombination-mediated repair behind a stalled replication fork can join the two nascent daughter chromosomes, resulting in a chromosome dimer (Barre et al., 2001). Dimer formation prohibits proper segregation of the genetic information at cell division (Figure 1A), and must be repaired to produce viable progeny. In bacteria and archaea, chromosome dimers are monomerized by members of a large family of tyrosine recombinases, the Xer recombinases (Blakely et al., 1991, 1993). These enzymes act by promoting recombination between two DNA sites, called dif. The dif site is normally present in a single copy at the replication terminus (Carnoy and Roten, 2009; Kuempel et al., 1991), but it is duplicated in chromosome dimers, so that intramolecular recombination results in separation (‘resolution’) of the two chromosome copies (Figure 1A). Removal of the xer genes or the dif site results in increased DNA content, activation of the SOS response, cell filamentation, and cell death (Britton and Grossman, 1999; Debowski et al., 2012a; Hendricks et al., 2000; Pérals et al., 2000; Val et al., 2008). Besides chromosome dimer resolution, Xer recombinases can support plasmid resolution and mobilization of the cholera toxin phage CTXϕ and pathogenicity islands (Das et al., 2013; Fischer et al., 2010).

Figure 1

Download asset Open asset

In E. coli, Xer recombination is carried out by cooperation of two similar enzymes XerC and XerD (37% identity) that act together to bind and recombine dif sites (Blakely et al., 1993). Many other organisms (including Lactococcus, Helicobacter, Campylobacter spp. and archaea) employ a single Xer recombinase system (Carnoy and Roten, 2009; Cortez et al., 2010; Debowski et al., 2012a; Le Bourgeois et al., 2007; Leroux et al., 2013), with a prime example, XerH/dif_H, found in Helicobacter pylori, a gastric pathogen implicated in peptic ulcer disease and gastric cancer.

Xer recombinases are members of the tyrosine site-specific recombinase superfamily, a large group of enzymes that catalyze DNA breakage and rejoining using a conserved tyrosine nucleophile (Grindley et al., 2006; Guo et al., 1997; Midonet and Barre, 2014; Nunes-Düby et al., 1998). Tyrosine recombinases promote various programmed DNA rearrangements including the monomerization of phage, plasmid and chromosome multimers, resolution of hairpin telomeres, and the movement of virulence and antibiotic resistance carrying integrative mobile genetic elements (including phages and transposons)(Grindley et al., 2006; Jayaram et al., 2015; Midonet and Barre, 2014). In addition, tyrosine recombinases (as exemplified by Cre and Flp) provide powerful genetic engineering tools that are widely used to carry out mutagenesis and DNA insertion in eukaryotic chromosomes (Nagy, 2000; Turan et al., 2011).

Tyrosine recombinases share a common chemical mechanism that involves step-wise breakage and exchange of four DNA strands in pairs, proceeding through a characteristic four-way Holliday junction (HJ) DNA intermediate (Figure 1B)(Gopaul and Duyne, 1999; Grindley et al., 2006; Holliday, 2007). They cut each DNA strand with a polarity creating a covalent 3’ phosphotyrosyl protein-DNA linkage and a free 5’ hydroxyl group. The DNA ends then go on to join with the complementary ends of the partner DNA strand generating the recombined products. All DNA cleavage and rejoining reactions take place in an ordered protein-DNA synaptic complex, comprising four recombinase molecules holding the two recombination partner DNA molecules together.

An unusual feature of Xer recombination at dif is that it requires an accessory factor, FtsK (Aussel et al., 2002; Debowski et al., 2012a; Le Bourgeois et al., 2007; Leroux et al., 2013; Nolivos et al., 2010; Steiner et al., 1999). This DNA motor protein localizes to the bacterial cell division septum and contributes to segregating the sister chromosomes into the daughter cells by translocating towards their replication termini. On chromosome dimers, FtsK stops at the Xer-bound dif sites and activates recombination, triggering resolution of the dimers to monomers (Aussel et al., 2002; Grainge et al., 2011; May et al., 2015). Without FtsK, Xer-dif synaptic complexes are formed, but do not lead to final recombination products (Aussel et al., 2002; Diagne et al., 2014; Grainge et al., 2011; Zawadzki et al., 2013). Regulation by FtsK is critical to ensure that Xer recombination takes place only in the correct spatial and temporal context – at the division septum, when genome replication has been completed – thereby ensuring faithful genome segregation.

Here, we present two crystal structures of the H. pylori XerH recombinase in complex with its recombination site dif_H. Together with associated biochemical and in vivo data, these first Xer-DNA complex structures shed light on potential regulatory mechanisms of the recombination pathway. Remarkably, the overall shape and DNA conformation of initially formed XerH-dif_H synaptic complexes is considerably different from those of other tyrosine recombinases, such as Cre-loxP that has served as a model system for the family. The unanticipated conformation of the pre-cleavage synaptic complex suggests a possible model for why Xer proteins require external activation, and in comparison with the post-cleavage complex structure provides clues for how FtsK might stimulate recombination activity. Our structures provide a resource to construct models for other Xer synaptic complexes, including that of the heterotetrameric E. coli XerC/D system.

In this work, we investigated the structural and mechanistic bases of Xer site-specific recombination. Using the homomeric XerH-dif_H system from H. pylori as a model system, we report the first high-resolution crystal structures of Xer-dif synaptic complexes. These structures demonstrate that Xer proteins follow the established chemical pathway of tyrosine recombinases and reveal how the reaction steps can be choreographed by small differences in the arms of the dif_H recombination sites. Together with associated biochemical data, the structures also show that XerH-dif_H synaptic complexes initially assemble in an inactive state with straight DNA, which must undergo a major conformational change for catalytic activation, as previously observed in single molecule fluorescence studies of XerC/D-dif recombination (Diagne et al., 2014; Zawadzki et al., 2013). Our post-cleavage XerH-dif_H synaptic complex structure elucidates the structural nature of this conformational change, which involves major rearrangement of the protein-protein interfaces and DNA bending (Video 1). With molecular modeling, we extend our structural and mechanistic findings to the prototypical E. coli XerC/D-dif system, demonstrating that our structures provide a resource for understanding the mechanism of other Xer recombinases.

The thoroughly characterized Cre site-specific recombination reaction has long been considered a paradigm for tyrosine recombinase-based DNA rearrangements (Gopaul and Duyne, 1999; Grindley et al., 2006; Guo et al., 1997; Van Duyne, 2015). Unusually, Xer recombinases require an external factor (generally the cell division protein FtsK) for catalytic activation. This feature is essential for Xer’s chromosome maintenance function, but does not have any analogy in the Cre system. Therefore, the structural basis of Xer recombination and its activation have remained debated. From our structural, biochemical, and microbiological data on H. pylori XerH-dif_H recombination we can now propose a mechanistic model for XerH recombination and the regulatory function of FtsK, as follows (Figure 6):

Figure 6 with 1 supplement see all

Download asset Open asset

The recombination process starts with XerH binding to the dif_H site on the H. pylori chromosome (Figure 6A). Two XerH subunits bind cooperatively to one dif_H site (Figure 3—figure supplement 1; Figure 6A), each forming a clamp around one arm of dif_H (Figure 2C). The dif_H left arm has higher affinity for XerH than the right arm (Figure 3A,B), leading to the second XerH molecule binding to the right arm more often than vice versa. Cooperative binding of XerH involves extensive contacts between the two subunits bound to each dif_H, as revealed by our structures (1650.6 Ų surface area with ΔG = −12.6 kcal/mol and 1515.5 Ų with ΔG = −14.5 kcal/mol between molecules A-B and A’-B’ respectively in the pre-cleavage structure; and 1379.8 Ų surface area with ΔG of −13.8 kcal/mol in the post-cleavage structure). Despite the near-perfect dyad symmetry of the dif_H arms, the bound XerH subunits are structurally distinct. The left arm subunit assumes a conformation primed for DNA cleavage, whereas the right arm subunit is forced in an inactive state. This asymmetric configuration defines the order of all subsequent cleavage and strand transfer events.

Two XerH-bound dif_H sites then interact to create a tetrameric synaptic complex, as we see in our crystal structures (Figure 2B; Figure 6B). The sites align in antiparallel, and intersubunit interactions across the synaptic interface anchor the αO helices of the right arm-bound XerH subunits on their left arm-bound partner, creating a ‘circular’ domain-swapped arrangement, similar to the ones described for other tyrosine recombinases such as Cre and λ integrase (Biswas et al., 2005; Guo et al., 1997).

The pre-cleavage synaptic complex crystal structure contains nearly straight dif_H DNA (Figure 2B; Figure 6B), and XerH is in a catalytically inactive conformation (Figure 3G). Correspondingly, we were unable to observe dif_H recombination or XerH-mediated cleavage of intact dif_H DNA substrates in vitro (Figure 6—figure supplement 1, and data not shown). We cannot exclude the possibility that XerH does still catalyze transient strand cleavage of these substrates, followed by rapid efficient re-ligation making cleavage undetectable, but there is no evidence to support this idea. In contrast, nicked dif_H suicide substrates are cleaved efficiently, and our crystal structure in which XerH-mediated cleavage of nicked dif_H has occurred showed a sharply bent DNA conformation, consistent with a direct link between DNA bending and cleavage activity. The crystals giving our two structures were grown under different conditions (see Materials and methods), but XerH was active (on nicked dif_H substrates) in both conditions (data not shown), supporting our view that both structures are biologically relevant. We hypothesize that our in vitro systems lack a stimulatory factor that is required for normal recombination between intact dif_H sites. Septum-borne FtsK was shown to be required for XerH recombination in H. pylori (Debowski et al., 2012a), so we propose that this is the missing factor for XerH activation in vitro. FtsK-promoted rearrangement of the XerH-dif_H synaptic complex might bring the catalytic tyrosines of the left arm-bound subunits into their active positions close to the scissile phosphates and sharply bend the DNA as seen in our structures (Figures 4B and 6C; Video 1). Another possibility is that FtsK mediates formation of a different synaptic complex with the right arm-bound subunits activated for cleavage. However, this would be inconsistent with smFRET studies of XerC/D complexes showing that FtsK activates pre-formed synaptic complexes without major remodeling (Zawadzki et al., 2013). Analogy with the XerC/D system suggests that FtsK might directly interact with the C-terminal ~20 amino acids of XerH (Yates et al., 2006) or the back of its C-terminal domain (Keller et al., 2016). However, the role of FtsK in the H. pylori Xer system still requires further substantial investigation, and it remains possible that the bent-DNA configuration required for DNA cleavage can be reached from the pre-cleavage structure without the intervention of FtsK.

Following activation, XerH-catalysed DNA strand cleavage and rejoining presumably follow the conserved tyrosine recombination pathway (Figure 1B), with the XerH subunits bound to the left arms performing the first strand exchanges. The resulting HJ is then resolved by a reciprocal set of chemical reactions catalysed by the XerH subunits bound to the right arms.

DNA bending is a prerequisite for activity of many molecular machines, including transcription factors, topoisomerases and recombinases (Dong and Berger, 2007; Kim et al., 1993; Lee et al., 2013; Yang and Steitz, 1995). In most of these cases, DNA bending and the downstream function are performed by the same protein or complex, or DNA bending depends on ubiquitously available factors such as IHF. For tyrosine recombinases, the paradigm suggests that sharp DNA bending that is required for DNA cleavage and energetically inexpensive strand exchange, is introduced by the recombinase itself concomitant with DNA binding and synapsis. Here, we show that this is not the case for XerH, which binds and assembles its recombination substrates in an almost straight conformation. The required DNA bending then presumably occurs upon activation by an external factor that is present only in a particular spatial and temporal context, providing an elegant regulatory mechanism to ensure faithful chromosome segregation.

Our modeling of XerC/D-dif indicates that the architectures of the pre- and post- cleavage synaptic complexes are probably conserved between XerC/D and XerH (Figure 5B). The DNA conformational changes that we observe for XerH-dif_H are also consistent with smFRET studies of XerC/D-dif synaptic complexes (Diagne et al., 2014; Zawadzki et al., 2013), which imply a conformational change upon FtsK-mediated activation consistent with the differences between our models of the pre-cleavage and post-cleavage complexes. Thus, we expect that the mechanism of catalytic activation proposed here for XerH – preassembly of a synaptic complex with almost straight DNA, followed by FtsK-induced protein and DNA rearrangements activating the complex for catalysis – may be conserved in the XerC/D system.

In summary, our structures of XerH-dif_H complexes demonstrate that catalysis by the Xer family of site-specific recombination systems follows the established tyrosine recombinase pathway, with the reaction steps being choreographed by small differences in the arms of the dif recombination sites. Contrary to previous assumption, bending of the dif sites does not occur concomitantly with synaptic complex assembly but at a post-synaptic step, when the accessory factor FtsK might be needed to license the reaction by promoting a conformational change. We also provide structural insight into the conformational change required for activation (Video 1), and extend our findings to other Xer recombinases through modeling. Our structural insights help us to better understand how Xer proteins function and how they have adapted the tyrosine recombinase machinery for their unique genome maintenance function. In the long term, our data can also help to improve XerH-based genetic engineering tools that have been recently introduced for markerless gene deletions in H. pylori (Debowski et al., 2012b).

1. 1. Bebel A
   2. Barabas O

   (2016) Crystal structure of XerH site-specific recombinase bound to difH substrate: pre-cleavage complex

   Publicly available at the RCSB Protein Data Bank (accession no: 5JK0).

   http://www.rcsb.org/pdb/search/structidSearch.do?structureId=5JK0
2. 1. Bebel A
   2. Barabas O

   (2016) Crystal structure of XerH site-specific recombinase bound to palindromic difH substrate: post-cleavage complex

   Publicly available at the RCSB Protein Data Bank (accession no: 5JJV).

   http://www.rcsb.org/pdb/search/structidSearch.do?structureId=5JJV

1. 1. Adams PD
   2. Afonine PV
   3. Bunkóczi G
   4. Chen VB
   5. Davis IW
   6. Echols N
   7. Headd JJ
   8. Hung LW
   9. Kapral GJ
   10. Grosse-Kunstleve RW
   11. McCoy AJ
   12. Moriarty NW
   13. Oeffner R
   14. Read RJ
   15. Richardson DC
   16. Richardson JS
   17. Terwilliger TC
   18. Zwart PH

   (2010) PHENIX: a comprehensive Python-based system for macromolecular structure solution

   Acta Crystallographica Section D Biological Crystallography 66:213–221.

   https://doi.org/10.1107/S0907444909052925

   • PubMed
   • Google Scholar
2. 1. Afonine PV
   2. Grosse-Kunstleve RW
   3. Echols N
   4. Headd JJ
   5. Moriarty NW
   6. Mustyakimov M
   7. Terwilliger TC
   8. Urzhumtsev A
   9. Zwart PH
   10. Adams PD

   (2012) Towards automated crystallographic structure refinement with phenix.refine

   Acta Crystallographica Section D Biological Crystallography 68:352–367.

   https://doi.org/10.1107/S0907444912001308

   • PubMed
   • Google Scholar
3. 1. Aihara H
   2. Kwon HJ
   3. Nunes-Düby SE
   4. Landy A
   5. Ellenberger T

   (2003) A conformational switch controls the DNA cleavage activity of lambda integrase

   Molecular Cell 12:187–198.

   https://doi.org/10.1016/S1097-2765(03)00268-5

   • PubMed
   • Google Scholar
4. 1. Arnold PH
   2. Blake DG
   3. Grindley ND
   4. Boocock MR
   5. Stark WM

   (1999) Mutants of Tn3 resolvase which do not require accessory binding sites for recombination activity

   The EMBO Journal 18:1407–1414.

   https://doi.org/10.1093/emboj/18.5.1407

   • PubMed
   • Google Scholar
5. 1. Aussel L
   2. Barre FX
   3. Aroyo M
   4. Stasiak A
   5. Stasiak AZ
   6. Sherratt D

   (2002) FtsK Is a DNA motor protein that activates chromosome dimer resolution by switching the catalytic state of the XerC and XerD recombinases

   Cell 108:195–205.

   https://doi.org/10.1016/S0092-8674(02)00624-4

   • PubMed
   • Google Scholar
6. 1. Barre FX
   2. Aroyo M
   3. Colloms SD
   4. Helfrich A
   5. Cornet F
   6. Sherratt DJ

   (2000) FtsK functions in the processing of a Holliday junction intermediate during bacterial chromosome segregation

   Genes & Development 14:2976–2988.

   https://doi.org/10.1101/gad.188700

   • PubMed
   • Google Scholar
7. 1. Barre FX
   2. Søballe B
   3. Michel B
   4. Aroyo M
   5. Robertson M
   6. Sherratt D

   (2001) Circles: the replication-recombination-chromosome segregation connection

   PNAS 98:8189–8195.

   https://doi.org/10.1073/pnas.111008998

   • PubMed
   • Google Scholar
8. 1. Biswas T
   2. Aihara H
   3. Radman-Livaja M
   4. Filman D
   5. Landy A
   6. Ellenberger T

   (2005) A structural basis for allosteric control of DNA recombination by lambda integrase

   Nature 435:1059–1066.

   https://doi.org/10.1038/nature03657

   • PubMed
   • Google Scholar
9. 1. Blakely G
   2. Colloms S
   3. May G
   4. Burke M
   5. Sherratt D

   (1991)

   Escherichia coli XerC recombinase is required for chromosomal segregation at cell division

   The New Biologist 3:789–798.

   • PubMed
   • Google Scholar
10. 1. Blakely G
    2. May G
    3. McCulloch R
    4. Arciszewska LK
    5. Burke M
    6. Lovett ST
    7. Sherratt DJ

    (1993) Two related recombinases are required for site-specific recombination at dif and cer in E. coli K12

    Cell 75:351–361.

    https://doi.org/10.1016/0092-8674(93)80076-Q

    • PubMed
    • Google Scholar
11. 1. Blakely GW
    2. Davidson AO
    3. Sherratt DJ

    (1997) Binding and cleavage of nicked substrates by site-specific recombinases XerC and XerD

    Journal of Molecular Biology 265:30–39.

    https://doi.org/10.1006/jmbi.1996.0709

    • PubMed
    • Google Scholar
12. 1. Britton RA
    2. Grossman AD

    (1999)

    Synthetic lethal phenotypes caused by mutations affecting chromosome partitioning in bacillus subtilis

    Journal of Bacteriology 181:5860–5864.

    • PubMed
    • Google Scholar
13. 1. Carnoy C
    2. Roten CA

    (2009) The dif/Xer recombination systems in proteobacteria

    PLoS One 4:e6531.

    https://doi.org/10.1371/journal.pone.0006531

    • PubMed
    • Google Scholar
14. 1. Chen Y
    2. Narendra U
    3. Iype LE
    4. Cox MM
    5. Rice PA

    (2000) Crystal structure of a flp recombinase-Holliday junction complex: assembly of an active oligomer by helix swapping

    Molecular Cell 6:885–897.

    https://doi.org/10.1016/S1097-2765(00)00086-1

    • PubMed
    • Google Scholar
15. 1. Cortez D
    2. Quevillon-Cheruel S
    3. Gribaldo S
    4. Desnoues N
    5. Sezonov G
    6. Forterre P
    7. Serre MC

    (2010) Evidence for a xer/dif system for chromosome resolution in archaea

    PLoS Genetics 6:e1001166.

    https://doi.org/10.1371/journal.pgen.1001166

    • PubMed
    • Google Scholar
16. 1. Das B
    2. Martínez E
    3. Midonet C
    4. Barre FX

    (2013) Integrative mobile elements exploiting Xer recombination

    Trends in Microbiology 21:23–30.

    https://doi.org/10.1016/j.tim.2012.10.003

    • PubMed
    • Google Scholar
17. 1. Debowski AW
    2. Carnoy C
    3. Verbrugghe P
    4. Nilsson HO
    5. Gauntlett JC
    6. Fulurija A
    7. Camilleri T
    8. Berg DE
    9. Marshall BJ
    10. Benghezal M

    (2012a) Xer recombinase and genome integrity in Helicobacter pylori, a pathogen without topoisomerase IV

    PLoS One 7:e33310.

    https://doi.org/10.1371/journal.pone.0033310

    • PubMed
    • Google Scholar
18. 1. Debowski AW
    2. Gauntlett JC
    3. Li H
    4. Liao T
    5. Sehnal M
    6. Nilsson HO
    7. Marshall BJ
    8. Benghezal M

    (2012b) Xer-cise in Helicobacter Pylori: one-step transformation for the construction of markerless gene deletions

    Helicobacter 17:435–443.

    https://doi.org/10.1111/j.1523-5378.2012.00969.x

    • PubMed
    • Google Scholar
19. 1. Diagne CT
    2. Salhi M
    3. Crozat E
    4. Salomé L
    5. Cornet F
    6. Rousseau P
    7. Tardin C

    (2014) TPM analyses reveal that FtsK contributes both to the assembly and the activation of the XerCD-dif recombination synapse

    Nucleic Acids Research 42:1721–1732.

    https://doi.org/10.1093/nar/gkt1024

    • PubMed
    • Google Scholar
20. 1. Dong KC
    2. Berger JM

    (2007) Structural basis for gate-DNA recognition and bending by type IIA topoisomerases

    Nature 450:1201–1205.

    https://doi.org/10.1038/nature06396

    • PubMed
    • Google Scholar
21. 1. Emsley P
    2. Lohkamp B
    3. Scott WG
    4. Cowtan K

    (2010) Features and development of coot

    Acta Crystallographica Section D Biological Crystallography 66:486–501.

    https://doi.org/10.1107/S0907444910007493

    • PubMed
    • Google Scholar
22. 1. Ennifar E
    2. Meyer JE
    3. Buchholz F
    4. Stewart AF
    5. Suck D

    (2003) Crystal structure of a wild-type cre recombinase-loxP synapse reveals a novel spacer conformation suggesting an alternative mechanism for DNA cleavage activation

    Nucleic Acids Research 31:5449–5460.

    https://doi.org/10.1093/nar/gkg732

    • PubMed
    • Google Scholar
23. 1. Fischer W
    2. Windhager L
    3. Rohrer S
    4. Zeiller M
    5. Karnholz A
    6. Hoffmann R
    7. Zimmer R
    8. Haas R

    (2010) Strain-specific genes of Helicobacter Pylori: genome evolution driven by a novel type IV secretion system and genomic island transfer

    Nucleic Acids Research 38:6089–6101.

    https://doi.org/10.1093/nar/gkq378

    • PubMed
    • Google Scholar
24. 1. Gopaul DN
    2. van Duyne GD

    (1999) Structure and mechanism in site-specific recombination

    Current Opinion in Structural Biology 9:14–20.

    https://doi.org/10.1016/S0959-440X(99)80003-7

    • PubMed
    • Google Scholar
25. 1. Gopaul DN
    2. Guo F
    3. Van Duyne GD

    (1998) Structure of the Holliday junction intermediate in Cre-loxP site-specific recombination

    The EMBO Journal 17:4175–4187.

    https://doi.org/10.1093/emboj/17.14.4175

    • PubMed
    • Google Scholar
26. 1. Grainge I
    2. Lesterlin C
    3. Sherratt DJ

    (2011) Activation of XerCD-dif recombination by the FtsK DNA translocase

    Nucleic Acids Research 39:5140–5148.

    https://doi.org/10.1093/nar/gkr078

    • PubMed
    • Google Scholar
27. 1. Grindley ND
    2. Whiteson KL
    3. Rice PA

    (2006) Mechanisms of site-specific recombination

    Annual Review of Biochemistry 75:567–605.

    https://doi.org/10.1146/annurev.biochem.73.011303.073908

    • PubMed
    • Google Scholar
28. 1. Guo F
    2. Gopaul DN
    3. van Duyne GD

    (1997) Structure of cre recombinase complexed with DNA in a site-specific recombination synapse

    Nature 389:40–46.

    https://doi.org/10.1038/37925

    • PubMed
    • Google Scholar
29. 1. Guo F
    2. Gopaul DN
    3. Van Duyne GD

    (1999) Asymmetric DNA bending in the Cre-loxP site-specific recombination synapse

    PNAS 96:7143–7148.

    https://doi.org/10.1073/pnas.96.13.7143

    • PubMed
    • Google Scholar
30. 1. Hendricks EC
    2. Szerlong H
    3. Hill T
    4. Kuempel P

    (2000) Cell division, guillotining of dimer chromosomes and SOS induction in resolution mutants (dif, xerC and xerD) of Escherichia coli

    Molecular Microbiology 36:973–981.

    https://doi.org/10.1046/j.1365-2958.2000.01920.x

    • PubMed
    • Google Scholar
31. 1. Holliday R

    (2007) A mechanism for gene conversion in fungi

    Genetical Research 89:285–304.

    https://doi.org/10.1017/S0016672308009476

    • PubMed
    • Google Scholar
32. 1. Jayaram M
    2. Ma CH
    3. Kachroo AH
    4. Rowley PA
    5. Guga P
    6. Fan HF
    7. Voziyanov Y

    (2015) An overview of tyrosine Site-specific recombination: From an flp perspective

    Microbiology Spectrum 3:.

    https://doi.org/10.1128/microbiolspec.MDNA3-0021-2014

    • PubMed
    • Google Scholar
33. 1. Jo CH
    2. Kim J
    3. Han AR
    4. Park SY
    5. Hwang KY
    6. Nam KH

    (2016) Crystal structure of thermoplasma acidophilum XerA recombinase shows large C-shape clamp conformation and cis-cleavage mode for nucleophilic tyrosine

    FEBS Letters 590:848–856.

    https://doi.org/10.1002/1873-3468.12109

    • PubMed
    • Google Scholar
34. 1. Kabsch W

    (2010) XDS

    Acta Crystallographica. Section D, Biological Crystallography 66:125–132.

    https://doi.org/10.1107/S0907444909047337

    • PubMed
    • Google Scholar
35. 1. Karaca E
    2. Bonvin AM

    (2011) A multidomain flexible docking approach to deal with large conformational changes in the modeling of biomolecular complexes

    Structure 19:555–565.

    https://doi.org/10.1016/j.str.2011.01.014

    • PubMed
    • Google Scholar
36. 1. Keller AN
    2. Xin Y
    3. Boer S
    4. Reinhardt J
    5. Baker R
    6. Arciszewska LK
    7. Lewis PJ
    8. Sherratt DJ
    9. Löwe J
    10. Grainge I

    (2016) Activation of Xer-recombination at dif: structural basis of the FtsKγ-XerD interaction

    Scientific Reports 6:33357.

    https://doi.org/10.1038/srep33357

    • PubMed
    • Google Scholar
37. 1. Kelley LA
    2. Mezulis S
    3. Yates CM
    4. Wass MN
    5. Sternberg MJ

    (2015) The Phyre2 web portal for protein modeling, prediction and analysis

    Nature Protocols 10:845–858.

    https://doi.org/10.1038/nprot.2015.053

    • PubMed
    • Google Scholar
38. 1. Kim JL
    2. Nikolov DB
    3. Burley SK

    (1993) Co-crystal structure of TBP recognizing the minor groove of a TATA element

    Nature 365:520–527.

    https://doi.org/10.1038/365520a0

    • PubMed
    • Google Scholar
39. 1. Krissinel E
    2. Henrick K

    (2007) Inference of macromolecular assemblies from crystalline state

    Journal of Molecular Biology 372:774–797.

    https://doi.org/10.1016/j.jmb.2007.05.022

    • PubMed
    • Google Scholar
40. 1. Kuempel PL
    2. Henson JM
    3. Dircks L
    4. Tecklenburg M
    5. Lim DF

    (1991)

    Dif, a recA-independent recombination site in the terminus region of the chromosome of Escherichia coli

    The New Biologist 3:799–811.

    • PubMed
    • Google Scholar
41. 1. Le Bourgeois P
    2. Bugarel M
    3. Campo N
    4. Daveran-Mingot ML
    5. Labonté J
    6. Lanfranchi D
    7. Lautier T
    8. Pagès C
    9. Ritzenthaler P

    (2007) The unconventional Xer recombination machinery of Streptococci/Lactococci

    PLoS Genetics 3:e117.

    https://doi.org/10.1371/journal.pgen.0030117

    • PubMed
    • Google Scholar
42. 1. Lee I
    2. Dong KC
    3. Berger JM

    (2013) The role of DNA bending in type IIA topoisomerase function

    Nucleic Acids Research 41:5444–5456.

    https://doi.org/10.1093/nar/gkt238

    • PubMed
    • Google Scholar
43. 1. Leroux M
    2. Rezoug Z
    3. Szatmari G

    (2013) The xer/dif site-specific recombination system of Campylobacter jejuni

    Molecular Genetics and Genomics 288:495–502.

    https://doi.org/10.1007/s00438-013-0765-5

    • PubMed
    • Google Scholar
44. 1. Li Z
    2. Ye Y
    3. Godzik A

    (2006) Flexible structural neighborhood--a database of protein structural similarities and alignments

    Nucleic Acids Research 34:D277–D280.

    https://doi.org/10.1093/nar/gkj124

    • PubMed
    • Google Scholar
45. 1. Luscombe NM
    2. Laskowski RA
    3. Thornton JM

    (1997) NUCPLOT: a program to generate schematic diagrams of protein-nucleic acid interactions

    Nucleic Acids Research 25:4940–4945.

    https://doi.org/10.1093/nar/25.24.4940

    • PubMed
    • Google Scholar
46. 1. May PF
    2. Zawadzki P
    3. Sherratt DJ
    4. Kapanidis AN
    5. Arciszewska LK

    (2015) Assembly, translocation, and activation of XerCD-dif recombination by FtsK translocase analyzed in real-time by FRET and two-color tethered fluorophore motion

    PNAS 112:E5133–E5141.

    https://doi.org/10.1073/pnas.1510814112

    • PubMed
    • Google Scholar
47. 1. McCoy AJ
    2. Grosse-Kunstleve RW
    3. Adams PD
    4. Winn MD
    5. Storoni LC
    6. Read RJ

    (2007) Phaser crystallographic software

    Journal of Applied Crystallography 40:658–674.

    https://doi.org/10.1107/S0021889807021206

    • PubMed
    • Google Scholar
48. 1. Midonet C
    2. Barre FX

    (2014) Xer Site-Specific recombination: Promoting vertical and horizontal transmission of genetic information

    Microbiology Spectrum 2:.

    https://doi.org/10.1128/microbiolspec.MDNA3-0056-2014

    • PubMed
    • Google Scholar
49. 1. Nagy A

    (2000) Cre recombinase: the universal reagent for genome tailoring

    Genesis 26:99–109.

    https://doi.org/10.1002/(SICI)1526-968X(200002)26:2<99::AID-GENE1>3.0.CO;2-B

    • PubMed
    • Google Scholar
50. 1. Nolivos S
    2. Pages C
    3. Rousseau P
    4. Le Bourgeois P
    5. Cornet F

    (2010) Are two better than one? Analysis of an FtsK/Xer recombination system that uses a single recombinase

    Nucleic Acids Research 38:6477–6489.

    https://doi.org/10.1093/nar/gkq507

    • PubMed
    • Google Scholar
51. 1. Nunes-Düby SE
    2. Kwon HJ
    3. Tirumalai RS
    4. Ellenberger T
    5. Landy A

    (1998) Similarities and differences among 105 members of the int family of site-specific recombinases

    Nucleic Acids Research 26:391–406.

    https://doi.org/10.1093/nar/26.2.391

    • PubMed
    • Google Scholar
52. 1. Page R
    2. Grzechnik SK
    3. Canaves JM
    4. Spraggon G
    5. Kreusch A
    6. Kuhn P
    7. Stevens RC
    8. Lesley SA

    (2003) Shotgun crystallization strategy for structural genomics: an optimized two-tiered crystallization screen against the thermotoga maritima proteome

    Acta Crystallographica Section D Biological Crystallography 59:1028–1037.

    https://doi.org/10.1107/S0907444903007790

    • PubMed
    • Google Scholar
53. 1. Pargellis CA
    2. Nunes-Düby SE
    3. de Vargas LM
    4. Landy A

    (1988)

    Suicide recombination substrates yield covalent lambda integrase-DNA complexes and lead to identification of the active site tyrosine

    The Journal of Biological Chemistry 263:7678–7685.

    • PubMed
    • Google Scholar
54. 1. Pettersen EF
    2. Goddard TD
    3. Huang CC
    4. Couch GS
    5. Greenblatt DM
    6. Meng EC
    7. Ferrin TE

    (2004) UCSF chimera--a visualization system for exploratory research and analysis

    Journal of Computational Chemistry 25:1605–1612.

    https://doi.org/10.1002/jcc.20084

    • PubMed
    • Google Scholar
55. 1. Pérals K
    2. Cornet F
    3. Merlet Y
    4. Delon I
    5. Louarn JM

    (2000) Functional polarization of the Escherichia coli chromosome terminus: the dif site acts in chromosome dimer resolution only when located between long stretches of opposite polarity

    Molecular Microbiology 36:33–43.

    https://doi.org/10.1046/j.1365-2958.2000.01847.x

    • PubMed
    • Google Scholar
56. 1. Rother M
    2. Rother K
    3. Puton T
    4. Bujnicki JM

    (2011) ModeRNA: a tool for comparative modeling of RNA 3d structure

    Nucleic Acids Research 39:4007–4022.

    https://doi.org/10.1093/nar/gkq1320

    • PubMed
    • Google Scholar
57. 1. Serre MC
    2. El Arnaout T
    3. Brooks MA
    4. Durand D
    5. Lisboa J
    6. Lazar N
    7. Raynal B
    8. van Tilbeurgh H
    9. Quevillon-Cheruel S

    (2013) The carboxy-terminal αN helix of the archaeal XerA tyrosine recombinase is a molecular switch to control site-specific recombination

    PLoS One 8:e63010.

    https://doi.org/10.1371/journal.pone.0063010

    • PubMed
    • Google Scholar
58. 1. Steiner W
    2. Liu G
    3. Donachie WD
    4. Kuempel P

    (1999) The cytoplasmic domain of FtsK protein is required for resolution of chromosome dimers

    Molecular Microbiology 31:579–583.

    https://doi.org/10.1046/j.1365-2958.1999.01198.x

    • PubMed
    • Google Scholar
59. 1. Subramanya HS
    2. Arciszewska LK
    3. Baker RA
    4. Bird LE
    5. Sherratt DJ
    6. Wigley DB

    (1997) Crystal structure of the site-specific recombinase, XerD

    The EMBO Journal 16:5178–5187.

    https://doi.org/10.1093/emboj/16.17.5178

    • PubMed
    • Google Scholar
60. 1. Terwilliger TC
    2. Grosse-Kunstleve RW
    3. Afonine PV
    4. Moriarty NW
    5. Zwart PH
    6. Hung LW
    7. Read RJ
    8. Adams PD

    (2008) Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard

    Acta Crystallographica Section D Biological Crystallography 64:61–69.

    https://doi.org/10.1107/S090744490705024X

    • PubMed
    • Google Scholar
61. 1. Turan S
    2. Galla M
    3. Ernst E
    4. Qiao J
    5. Voelkel C
    6. Schiedlmeier B
    7. Zehe C
    8. Bode J

    (2011) Recombinase-mediated cassette exchange (RMCE): traditional concepts and current challenges

    Journal of Molecular Biology 407:193–221.

    https://doi.org/10.1016/j.jmb.2011.01.004

    • PubMed
    • Google Scholar
62. 1. Val ME
    2. Kennedy SP
    3. El Karoui M
    4. Bonné L
    5. Chevalier F
    6. Barre FX

    (2008) FtsK-dependent dimer resolution on multiple chromosomes in the pathogen vibrio cholerae

    PLoS Genetics 4:e1000201.

    https://doi.org/10.1371/journal.pgen.1000201

    • PubMed
    • Google Scholar
63. 1. Van Duyne GD

    (2001) A structural view of cre-loxp site-specific recombination

    Annual Review of Biophysics and Biomolecular Structure 30:87–104.

    https://doi.org/10.1146/annurev.biophys.30.1.87

    • PubMed
    • Google Scholar
64. 1. Van Duyne GD

    (2015) Cre recombinase

    Microbiology Spectrum 3:.

    https://doi.org/10.1128/microbiolspec.MDNA3-0014-2014

    • PubMed
    • Google Scholar
65. 1. van Zundert GC
    2. Rodrigues JP
    3. Trellet M
    4. Schmitz C
    5. Kastritis PL
    6. Karaca E
    7. Melquiond AS
    8. van Dijk M
    9. de Vries SJ
    10. Bonvin AM

    (2016) The HADDOCK2.2 web server: User-friendly integrative modeling of biomolecular complexes

    Journal of Molecular Biology 428:720–725.

    https://doi.org/10.1016/j.jmb.2015.09.014

    • PubMed
    • Google Scholar
66. 1. Vonrhein C
    2. Blanc E
    3. Roversi P
    4. Bricogne G

    (2007) Automated structure solution with autoSHARP

    Methods in Molecular Biology 364:215–230.

    https://doi.org/10.1385/1-59745-266-1:215

    • PubMed
    • Google Scholar
67. 1. Yang J
    2. Yan R
    3. Roy A
    4. Xu D
    5. Poisson J
    6. Zhang Y

    (2015) The I-TASSER Suite: protein structure and function prediction

    Nature Methods 12:7–8.

    https://doi.org/10.1038/nmeth.3213

    • PubMed
    • Google Scholar
68. 1. Yang W
    2. Steitz TA

    (1995) Crystal structure of the site-specific recombinase gamma delta resolvase complexed with a 34 bp cleavage site

    Cell 82:193–207.

    https://doi.org/10.1016/0092-8674(95)90307-0

    • PubMed
    • Google Scholar
69. 1. Yates J
    2. Aroyo M
    3. Sherratt DJ
    4. Barre FX

    (2003) Species specificity in the activation of Xer recombination at dif by FtsK

    Molecular Microbiology 49:241–249.

    https://doi.org/10.1046/j.1365-2958.2003.03574.x

    • PubMed
    • Google Scholar
70. 1. Yates J
    2. Zhekov I
    3. Baker R
    4. Eklund B
    5. Sherratt DJ
    6. Arciszewska LK

    (2006) Dissection of a functional interaction between the DNA translocase, FtsK, and the XerD recombinase

    Molecular Microbiology 59:1754–1766.

    https://doi.org/10.1111/j.1365-2958.2005.05033.x

    • PubMed
    • Google Scholar
71. 1. Zawadzki P
    2. May PF
    3. Baker RA
    4. Pinkney JN
    5. Kapanidis AN
    6. Sherratt DJ
    7. Arciszewska LK

    (2013) Conformational transitions during FtsK translocase activation of individual XerCD-dif recombination complexes

    PNAS 110:17302–17307.

    https://doi.org/10.1073/pnas.1311065110

    • PubMed
    • Google Scholar
72. 1. Zheng G
    2. Lu XJ
    3. Olson WK

    (2009) Web 3DNA--a web server for the analysis, reconstruction, and visualization of three-dimensional nucleic-acid structures

    Nucleic Acids Research 37:W240–W246.

    https://doi.org/10.1093/nar/gkp358

    • PubMed
    • Google Scholar

Thank you for submitting your article "Structural snapshots of Xer recombination reveal activation by synaptic complex remodeling and DNA bending." for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by a Reviewing Editor and John Kuriyan as the Senior Editor. The following individual involved in review of your submission has agreed to reveal his identity: Gregory D Van Duyne.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Understanding the mechanism of Xer recombination at chromosomal dif sites is acknowledged to be a long-standing, challenging problem. In this regard, the manuscript provides an extensive structural and biochemical analysis of the Helicobacter Xer-dif recombination machine in a well-written and well-documented manner, and that it reveals new details of the steps involved in this important reaction. Showing for the first time high resolution structures of a Xer recombinase complexed with DNA and these structures was unanimously considered to be useful for modelling and understanding other Xer recombinases.

This said, there was some debate as to the novelty and impact of the insights they propose. One reviewer felt that the model in which inactive, un-bent synaptic Xer/dif complexes form and are then induced to bend and cleave by FtsK is intriguing and would constitute a major advance in understanding Xer/dif recombination (if correct). Another reviewer mentioned that the structures of the pre-cleavage (and catalytically inactive) and post-cleavage (on a nicked 'suicide' substrate) synaptic complexes are important and provide some surprises, but that their mechanistic significance is more incremental, with the final model (at least as depicted) adding little to what has been published previously.

In going forward, attention should be paid to depicting and improving the discussion of what new insights can gained from the present work to resolve this debate – i.e., how specifically do the structural snapshots and associated biochemistry inform or change the field's understanding of Xer recombination mechanism in a meaningful way? In addition, there are several other issues that need to be addressed before a final decision can be reached. These include:

1) The paper overreaches in trying to reinforce the biological significance of the work. In the Abstract for example, the statement that the work will "open up new avenues towards the antibacterial treatment" is far-fetched. The same applies to the statement (in the Abstract and elsewhere) that the work provides mechanistic insight into the action of FtsK in stimulating the reaction. Similar overreach is present in the Discussion. In general, statements relating to how FtsK activates Helicobacter Xer recombination or how the present study will aid drug discovery should be downplayed, as there are no direct experiments or data that relate to these points. Relevant to the FtsK, would it be appropriate to bring up and cite the subsequent Zawadzki work (May et al. 2015) on FtsK activation as monitored in single-molecule FRET.

2) Figure 3E. Was the Xer recombination observed in E. coli FtsK-dependent (Figure 3E and supporting information)? If the authors still wish to focus on the importance of FtsK in activating Xer recombination, this is an important experiment. If it is FtsK-dependent, is this expected given the differences between Helicobacter FtsK and E. coli FtsK (and would one expect Helicobacter Xer to be acted on by E. coli FtsK, if indeed this is the case?). Some comment on this question is needed.

3) In Figure 6 and the associated text, there is a problem with the proposal that the binding of the two recombinases to dif is temporally sequential in a mechanistic sense. How can this be? The observed affinities for the two dif arms are presumably determined by the relative off-rates, meaning that the left site is more often occupied than the right site, and leading to the second molecule binding to the right site more often than vice versa. Why not simply restate this way? Along these lines, a substantial part of Figure 6 is a repeat of what is in Figure 1. It is suggested that Figure 6 be deleted and that relevant new panels be incorporated into Figure 5.

4) The evidence that XerH does not cleave linear dif sites shown in supplemental Figure 6 seems weak. E. coli XerCD can cleave dif sites in vitro and in vivo in the absence of FtsK, but this cleavage is non-productive (XerC generates and resolves HJs or cleaves and re-ligates; XerD is inactive). By analogy, XerH bound to the left half-site of dif resembles XerC in its properties (i.e., the more active half). The authors have made the opposite assignment in their model of XerC/XerD/dif shown in Figure 5, which is consistent with current views that XerD catalyzes the first strand exchange when activated by FtsK. However, this assignment implicitly assigns XerD half to be the more active one, based on the data presented. The synaptic complex structure determined in this work might instead correspond to the alternative 'XerC cleavage' synaptic complex: the one that undergoes futile cycling. The reaction schematic shown in Figure 6 implies that the more reactive dif half-sites initiate strand exchange (with the help of FtsK) and the inactive half-sites resolve the HJ intermediate, based on the structural models and biochemistry presented. This is the opposite of the currently favored mechanism for E. coli Xer/dif recombination. Please comment on this point.

5) The cleaved complex confirms that Xer enzymes proceed through a Cre-like pathway once activated, with sharply bend DNA arms, etc. Unfortunately, the structure has been symmetrized to be Cre-like, possibly masking asymmetric features that may be present. In particular, it may be difficult to conclude much from the inter-subunit interactions present, since the asymmetries introduced by the right half-site have been lost. This point should be noted in the text.

6) The pre-cleavage complex was crystallized without divalent ions, at pH 5.5. The post-cleavage complex was crystallized at pH 7 in the presence of a high concentration of divalent ions. An alternative interpretation of the current results is that the observed structures reflect the crystallization conditions and that the synaptic complex is actually bent (to some extent) at netural pH with Mg²⁺ present. The authors should address this alternative explanation in the paper.

7) The cooperativity of XerH binding to dif is actually very low, given the similar half-site vs. full site binding constants. The wording in the manuscript is somewhat misleading in this respect – it is surprising how low the cooperativity is, given the extent of inter-subunit interactions on a single dif site. The authors should report what this interaction surface is and discuss the lack of strong cooperative binding in light of the structure.

8) It would be useful to report whether a bent, but un-nicked duplex substrate can be reconstructed (modeled) from the cleaved complex. This would indicate whether the higher energy bent form of dif is likely to resemble the cleaved configuration, or alternatively, whether the cleaved complex undergoes additional bending and subunit rotations that occur after cleavage has occurred.

9) The authors suggest that the suicide substrates used are cleaved by XerH because the nicks facilitate the required bend. Usually, the energetic cost of DNA bending is mostly related to unstacking of the bases, which nicking does not solve. Do the current structures explain how the nicks might be specifically allowing higher cleavage for this system – e.g., through relieving a strained backbone configuration? Or are they simply allowing the cleavage product to accumulate? Is the structure of an XerH catalytic mutant bound to nicked dif sites expected to be bent?

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Structural snapshots of Xer recombination reveal activation by synaptic complex remodeling and DNA bending" for further consideration at eLife. Your revised article has been favorably evaluated by John Kuriyan (Senior editor), a Reviewing editor, and two reviewers.

The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below. Note that these issues can all be dealt with through textual changes, and do not require any additional experiments.

1) The models in Figures 1, 2 and 6 cause some confusion as to how the structures might be interpreted in light of the XerH mechanism. Please consistently draft these figures with the strand going from 5' [on the left] to 3' [on the right] placed on 'top' in each of the views [i.e. viewing from the minor groove side in each case]. The authors should also acknowledge that an alternative model is consistent with their data in which FtsK is required to make the alternative synaptic complex and/or explain why such a model isn't favored.

2) Subsection “Modeling of E. coli XerC/D synaptic complex based on the XerH-difH structures”. In the absence of supporting data indicating that XerH will preferentially resolve HJs by bottom strand exchange, this claim should be removed. See also Discussion paragraph 7 in reference to this point.

3) Discussion section, paragraph one. Aside from the crystal structure reported here, there is other evidence that XerH/dif sites are assembled in an inactive synaptic complex in solution (e.g., Zawadzki et al. should be noted). Is there additional evidence for XerH/difH?

4) Authors' response #1 and related manuscript sections. What is the evidence that XerH does not cleave dif sites in vitro? This has been a tricky issue in the field, since it is hard to distinguish cleavage followed by rapid ligation from lack of cleavage. Unless the authors can provide data (e.g., using 5'-bridging phosphorothiolates) that demonstrate lack of cleavage, it would be better to take the position that it is not yet clear whether cleavage occurs in vitro. It is also difficult to explain activity in E. coli with such a model. In contrast, the behavior is readily explained if XerH bound to the left end behaves like XerC (as in an alternative model). Related to this is the question of whether cleavage is required for stabilization of the synaptic complex, independent of FtsK. This is true of λ int and several catalytic Cre mutants form much weaker synaptic complexes relative to wild-type Cre, suggesting that this could also be true in XerH as well. The nicked XerH/DNA complex also supports this idea. Thus, the synaptic complex structure observed for XerH could equally well represent a stable intermediate that can reach the transient bent configuration required for cleavage on its own. Please resolve/comment on this issue.

5) Authors' response #6. The reviewers requested that the authors note the possibility that the synaptic complex structure may be influenced by crystallization conditions (low pH and/or lack of divalents). The authors acknowledge this in their response to reviewers, but do not appear to have modified the manuscript. Given that this structure is central to the authors' proposals for Xer recombination, the readers should be made aware that the conditions for the two structures are quite different. Please revise the manuscript accordingly.

6) The authors note that several unbound Xer crystal structures show different domain relationships, but continue to promote the idea that a specific conformational change takes place upon DNA binding (as shown in the video). To echo the original review, it is unlikely that the XerD alone structure represents a true "conformational intermediate," rather than just one of an ensemble of configurations present when Xer subunits are not bound to dif. The three Xer crystal structures in the PDB have their NTDs in completely different conformations. The ideas in the discussion and video illustrating that the NTD has to undergo a significant change when completing the clamp on the dif half-site was assumed many years ago by most people in the field and is true for all tyrosine recombinases. This point should clarified or else its discussion (and the movie) should be removed from the paper.

7) The distinction between understanding the mechanism of recombination vs. regulation of this process is important, but not made clear (Abstract and Introduction sections). The authors imply that the mechanism of XerC/XerD recombination is still unknown.

8) Introduction paragraph six. The structures do not delineate the pathway (delete this) but do shed light on potential regulatory mechanisms (reword).

9) In the same section the structure does not explain why activation is required. It suggests a possible model, as described by the authors, which differs from at least one other XerCD model.

10) Results paragraph one. The current mechanism is based on much more than the series of Cre-loxP structures cited. A review covering tyrosine recombinase mechanisms (there are many) might be more appropriate here.

11) Subsection “Structure of the XerH-difH synaptic complex” need to be redrafted to accurately to describe the historical record (in doing so, the authors should consider point 9 and earlier comments regarding alternative models). The current text is incomplete and in part misleading. In the absence of FtsK, HJs formed by XerC were detected in early work and this led to the hypothesis that the XerC active conformation formed preferentially and FtsK switched the conformation of an XerC active complex to one in which XerD was active. Single-molecule FRET experiments by Zawadzki and colleagues led to the different hypothesis that this is not the case; rather, the major synaptic complex that forms in the absence of FtsK is an 'XerD potentially active conformation,' which is catalytically dead. A small proportion of XerC active complexes nevertheless form and are responsible for the XerC HJs. Nevertheless, the dominant pre-active XerD synaptic complexes are the ones activated by XerD. This is consistent with the results here, but the historical development of these ideas should be explicitly chronicled.

12) Subsection “Structure of the XerH-difH synaptic complex” paragraph two. DNA duplex, not oligonucleotide.

[…] In going forward, attention should be paid to depicting and improving the discussion of what new insights can gained from the present work to resolve this debate – i.e., how specifically do the structural snapshots and associated biochemistry inform or change the field's understanding of Xer recombination mechanism in a meaningful way? In addition, there are several other issues that need to be addressed before a final decision can be reached.

We hope that our revisions have added clarity with respect to the novelty of our study. We have now included a paragraph at the beginning of Discussion to summarize the novel data and insights brought about by our work. We have also modified the text throughout, particularly the final paragraph of the discussion, to better emphasize the impact of our results for understanding of Xer recombination and tyrosine recombination more generally.

The main novel aspects of our study and their impact are summarized below for your information:

We report the first high-resolution structures of an Xer recombinase in synaptic complex with its DNA substrates. Such a structure has been long sought, as its absence prevented proper understanding of this important family of site-specific recombination systems. Structural models relying on available structures of the Cre/lox system and others left several questions unanswered. Especially with regard to the molecular mechanism of Xer activation at chromosomal dif sites, it remained unclear why Xer proteins require activation in the first place, and what the molecular assemblies and rearrangements preceding activation might look like (all previous tyrosine recombinase-DNA structures reflected macromolecular assemblies highly similar to the activated state). In this manuscript, we present two structures of the XerH-difH synaptic complex captured at distinct stages of the recombination reaction, which shed light on critical steps of the reaction, and the structural changes involved in XerH activation.

Our wild-type XerH-difH synaptic complex structure reveals a cleavage-incompetent conformation with almost straight DNA that has never been observed for tyrosine recombinases before. This structure shows that conformational changes are required for recombination activation, and suggests that an external factor may be required to facilitate the transition. Based on published evidence that XerH recombination requires FtsK in H. pylori (Debowski et al., 2012) and by analogy with the XerC/D system, it seems reasonable to predict that the need for this conformational change may be key to the regulation of Xer proteins, and that FtsK may be directly involved in facilitating this transition. While this model requires further validation, our structure will help to create testable experimental hypotheses.

We also present a post-cleavage XerH-difH synaptic complex structure (trapped using nicked suicide DNA substrates). This structure resembles previously reported post-cleavage structures of other tyrosine recombinases with strongly bent DNA, confirming that the chemistry of catalysis by Xer proteins likely resembles canonical tyrosine recombination. Comparison of our pre- and post-cleavage structures also illustrates the structural rearrangements required for activation of Xer recombination. Unexpectedly, these changes include a major rearrangement of the intermolecular protein interfaces and significant DNA bending. This is consistent with previously published smFRET data (Zawadzki et al., 2013, see below), and supports the idea that pre-formed Xer recombinase synaptic complexes can be remodeled without energetically inefficient disassembly and reassembly of the complex.

Our structures also show that strong DNA bending is not necessarily concomitant with synaptic complex assembly in tyrosine recombination, at least for the XerH-difH system.

Our structural and biochemical data reveal how a small asymmetry in the dif_H site helps to establish correct ordering of strand cleavages and ligations in recombination by the homomeric Xer recombinase XerH.

Modeling extends our data to the prototypical XerC/D system, indicating that the XerH-dif_H structures provide a useful resource for understanding other Xer recombinases.

1) The paper overreaches in trying to reinforce the biological significance of the work. In the Abstract for example, the statement that the work will "open up new avenues towards the antibacterial treatment" is far-fetched. The same applies to the statement (in the Abstract and elsewhere) that the work provides mechanistic insight into the action of FtsK in stimulating the reaction. Similar overreach is present in the Discussion. In general, statements relating to how FtsK activates Helicobacter Xer recombination or how the present study will aid drug discovery should be downplayed, as there are no direct experiments or data that relate to these points. Relevant to the FtsK, would it be appropriate to bring up and cite the subsequent Zawadzki work (May et al. 2015) on FtsK activation as monitored in single-molecule FRET.

In accord with the comments of the reviewers, we have removed all the comments regarding the potential use of our structures for antibiotic discovery. We have also rephrased the text referring to FtsK activation of Xer recombination, to clarify how far our data go and where hypothesis starts.

We do however maintain that our structural and biochemical data provide some important insights into the activation of XerH recombination by FtsK. First, our observations that wild-type XerH-dif_H synaptic complexes assemble in a cleavage-incompetent conformation and that XerH alone is unable to cleave wild-type DNA substrates in vitro indicates that the help of an external factor is needed for activation. In agreement with this idea, literature evidence from Debowski et al., 2012 showed that XerH recombination is critically dependent on FtsK in H. pylori, and strains lacking the C-terminal part of FtsK show a strong filamentous phenotype and no detectable XerH-dif_H recombination. Therefore, it seems reasonable to presume that FtsK activates XerH in vivo in H. pylori, as E. coli FtsK activates XerC/D-dif recombination in E. coli. Second, comparison of the inactive XerH-dif_H pre-cleavage complex with our activated post-cleavage complex structure indicates that XerH activation involves large conformational changes in pre-formed synaptic complexes. This is also in line with smFRET data reported by Zawadzki et al., 2013 and Diagne et al., 2014 for XerC/D-dif complexes.

Thus, we predict that FtsK is directly involved in activating XerH in H. pylori, probably via facilitating the conformational changes observed in our structures. Nevertheless, we acknowledge that our data do not provide clear-cut evidence for the exact function of FtsK, and do not address exactly how FtsK acts on the molecular or structural level. Insights into these questions will require further investigations, and we have now clarified this further in the manuscript (Discussion section).

Reference to May et al., 2015 has been added (Introduction section).

2) Figure 3E. Was the Xer recombination observed in E. coli FtsK-dependent (Figure 3E and supporting information)? If the authors still wish to focus on the importance of FtsK in activating Xer recombination, this is an important experiment. If it is FtsK-dependent, is this expected given the differences between Helicobacter FtsK and E. coli FtsK (and would one expect Helicobacter Xer to be acted on by E. coli FtsK, if indeed this is the case?). Some comment on this question is needed.

The experiment shown in Figure 3E used a standard E. coli lab strain with wild type FtsK. To check if XerH-dif_H recombination was dependent on E. coli FtsK in this assay, we tested recombination in the E. coli strain DS9041, which lacks the C-terminal part of FtsK (FtsK-C) (kindly provided by Prof. Dave Sherratt). Note that full deletion of FtsK is lethal and cannot be used in these assays, but FtsK-C deletion was previously shown to be sufficient to abolish XerC/D activation (Recchia et al., 1999; Steiner et al., 1999). Plasmid-based intramolecular XerH-dif_H recombination assays in this strain showed similar levels of dif_H-galK-dif_H cassette excision as was observed in the wild-type strain, suggesting that deletion of FtsK-C does not affect this reaction.

Chromosome resolution by XerH-dif_H recombination in its native H. pylori is known to require FtsK (Debowski et al., 2012). So, to check if chromosome dimer resolution by XerH-dif_H recombination can occur in E. coli, we replaced the dif site of the E. coli chromosome with dif_H-Km and supplemented this strain with an XerH expression plasmid. This showed a phenotype consistent with loss of chromosome dimer resolution (~10% of elongated E. coli cells that are unable to divide), irrespective of the presence or absence of E. coli FtsK-C. This showed that XerH-dif_H recombination on chromosomes is compromised in E. coli, and E. coli FtsK cannot fully complement for the function of H. pylori FtsK. These results are consistent with previous reports showing that the Xer-FtsK interactions can be species-specific (Yates et al. 2003).

The fact that we still observe detectable levels of intramolecular XerH-dif_H recombination on plasmid-borne dif_H-galK-dif_H cassettes in E. coli may be due to the high sensitivity of this assay (using a protein overexpression plasmid and multi-copy reporter plasmids), or it may indicate that factors other than E. coli FtsK-C (e.g. DNA bending by supercoiling, or other E. coli proteins) can activate XerH in this setting. However, elucidating the exact function of H. pylori FtsK and the effects of other interaction partners on XerH recombination requires further investigations, which are beyond the scope of the current study.

3) In Figure 6 and the associated text, there is a problem with the proposal that the binding of the two recombinases to dif is temporally sequential in a mechanistic sense. How can this be? The observed affinities for the two dif arms are presumably determined by the relative off-rates, meaning that the left site is more often occupied than the right site, and leading to the second molecule binding to the right site more often than vice versa. Why not simply restate this way?

We agree with the reviewer on this point. We have rephrased the corresponding text (Discussion section) and figure legend as requested, and avoid using the term ‘sequential binding’ throughout.

Along these lines, a substantial part of Figure 6 is a repeat of what is in Figure 1. It is suggested that Figure 6 be deleted and that relevant new panels be incorporated into Figure 5.

We agree that Figure 6 was unnecessarily repetitive. We have now removed panels E-G and keep only the parts that are needed to illustrate the model we propose for XerH recombination. For the benefit of the broad readership of eLife, however, we prefer to keep this model figure separate from the other data figures.

4) The evidence that XerH does not cleave linear dif sites shown in supplemental Figure 6 seems weak. E. coli XerCD can cleave dif sites in vitro and in vivo in the absence of FtsK, but this cleavage is non-productive (XerC generates and resolves HJs or cleaves and re-ligates; XerD is inactive). By analogy, XerH bound to the left half-site of dif resembles XerC in its properties (i.e., the more active half). The authors have made the opposite assignment in their model of XerC/XerD/dif shown in Figure 5, which is consistent with current views that XerD catalyzes the first strand exchange when activated by FtsK. However, this assignment implicitly assigns XerD half to be the more active one, based on the data presented. The synaptic complex structure determined in this work might instead correspond to the alternative 'XerC cleavage' synaptic complex: the one that undergoes futile cycling. The reaction schematic shown in Figure 6 implies that the more reactive dif half-sites initiate strand exchange (with the help of FtsK) and the inactive half-sites resolve the HJ intermediate, based on the structural models and biochemistry presented. This is the opposite of the currently favored mechanism for E. coli Xer/dif recombination. Please comment on this point.

We agree with the referee that assignment of XerC and XerD to the individual XerH subunits in our XerH-dif_H synaptic complexes is ambiguous. Our XerH cleavage experiments might indeed be reporting on ‘futile cycling’ by the subunit bound at the left arm and we cannot fully exclude the possibility that our structures reflect a ‘futile cycling’ state. However, we note that currently there is no evidence that XerH is subject to the same ‘futile cycling’ mechanism as the XerC/D-dif system. (In fact, XerC/D itself is also not subject to this phenomenon when acting on plasmid-borne cer and psi sites.) Instead, our results indicate that while first XerH cleavage preferentially occurs on the left arm, resolution of a Holliday Junction substrate (mentioned in the Discussion section as “data not shown”) occurs by cleavage of the right dif_H arm. Drawing analogy with the XerC/D-dif system, this observation would map XerC to the XerH subunit bound to the right arm (because it is XerC that resolves preformed HJ substrates with or without FtsK), and thus XerD to the left arm-bound subunit. This, together with the observed lack of cleavage on native (un-nicked) substrates (Figure 6—figure supplement 1), indicates that XerH might not be affected by futile cycling. Instead, XerH synaptic complexes appear to assemble directly in the correct configuration, but normally remain inactive in the absence of FtsK. Introduction of nicks in the DNA substrates might mimic the effect of FtsK, allowing activation of left arm cleavage as seen in our cleavage assays and the post-cleavage structure.

We now state this ambiguity and explain our choice for XerC/D assignment (subsection “Modeling of E. coli XerC/D synaptic complex based on the XerH-difH structures”).

5) The cleaved complex confirms that Xer enzymes proceed through a Cre-like pathway once activated, with sharply bend DNA arms, etc. Unfortunately, the structure has been symmetrized to be Cre-like, possibly masking asymmetric features that may be present. In particular, it may be difficult to conclude much from the inter-subunit interactions present, since the asymmetries introduced by the right half-site have been lost. This point should be noted in the text.

We appreciate that our synaptic complex with symmetrized dif_HLP might not fully reflect the subunit arrangement and interactions in the native XerH-dif_H complex. However, we note that the overall structure of each XerH subunit and their interactions with the respective dif_H DNA arms in the XerH-dif_HLP complex and the XerH – native dif_H (pre-cleavage) complex are very similar (see subsection “Conformational rearrangements in the post-cleavage synaptic complex”), suggesting that substrate symmetrization does not have a profound effect on the synaptic complex structure. We note that crystal structures of Cre/lox synaptic complexes also showed highly similar arrangements with symmetrized and native substrates (Guo et al. 1997, Guo et al. 1999, Ennifar et al. 2003). Moreover, we would like to emphasize that – despite the use of symmetrized substrates – the conformations of the two XerH subunits bound to the two arms of one difHLP site are quite different and the XerH-difHLP structure is actually ‘asymmetric’ in this sense (i.e. the complex has only dyad symmetry connecting the two difH sites across the synaptic interface; there is no symmetry connecting the protein molecules bound to two halves of one difHLP site). Since we observe very similar ‘asymmetry’ in the native pre-cleavage XerH-difH complex, this feature appears to be intrinsic to the natural system. We have now revised the text (last paragraph in subsection “Conformational rearrangements in the post-cleavage synaptic complex”) to make this clearer.

6) The pre-cleavage complex was crystallized without divalent ions, at pH 5.5. The post-cleavage complex was crystallized at pH 7 in the presence of a high concentration of divalent ions. An alternative interpretation of the current results is that the observed structures reflect the crystallization conditions and that the synaptic complex is actually bent (to some extent) at netural pH with Mg²⁺ present. The authors should address this alternative explanation in the paper.

We acknowledge that the crystallization conditions may have an influence. However, we have tested XerH activity under the exact crystallization condition used for the pre-cleavage complex (pH 5.5, no metal ions) and observe a good level of DNA cleavage with nicked DNA substrates, indicating that the crystallization conditions are not primarily responsible for the inactive conformational state we observe in our structure. In the Cre/lox system, both pre- and post-cleavage synaptic complexes invariably show a similar DNA bend and inter-subunit arrangement, irrespective of the crystallization conditions which included pH of 5 – 7.6 and various divalent ion (Mg²⁺ or Ca²⁺⁾ concentrations.

7) The cooperativity of XerH binding to dif is actually very low, given the similar half-site vs. full site binding constants. The wording in the manuscript is somewhat misleading in this respect – it is surprising how low the cooperativity is, given the extent of inter-subunit interactions on a single dif site. The authors should report what this interaction surface is and discuss the lack of strong cooperative binding in light of the structure.

All binding constants mentioned in the manuscript were measured with full-length DNA substrates; the substrates dif_HL and dif_HR contained random DNA sequence in place of the second dif_H arm. Since these substrates can still bind two XerH subunits (Figure 3 —figure supplement 1B-D) and all complex species were pooled for K_D calculations, the reported values do not reflect true ‘half-site’ binding constants. We apologize for this confusion, and have now revised the manuscript text to make this clear (subsection “difH asymmetry dictates the order of binding and cleavage events”).

To evaluate cooperativity, we have prepared Hill plots based on our EMSA data, by plotting the logarithm of the fractional binding of the DNA substrate, against the logarithm of the protein concentration. This gave a high value for cooperativity, in agreement with the very sharp transition of all unbound DNA to all bound (in dimeric complex) observed on our EMSA gels with native dif_H (Figure 3 —figure supplement 1A). Similarly, high cooperativity was observed for XerC/D – dif binding (e.g. Cao et al., 1997 and Spiers and Sherratt, 1997), and is consistent with the extensive inter-subunit interactions observed in our structures (now reported in Discussion section paragraph four), as also pointed out by the referees. However, we are aware that the exact level of cooperativity cannot be evaluated from our data and have rephrased the corresponding sections in the text.

8) It would be useful to report whether a bent, but un-nicked duplex substrate can be reconstructed (modeled) from the cleaved complex. This would indicate whether the higher energy bent form of dif is likely to resemble the cleaved configuration, or alternatively, whether the cleaved complex undergoes additional bending and subunit rotations that occur after cleavage has occurred.

We have performed the modeling as recommended, and find that the broken DNA strand can be joined in the cleaved complex structure with acceptable geometry, to reconstruct a bent but un-nicked substrate. However, the modeled DNA strand clashes with the protein in a loop region just before helix αN. This is now stated in the manuscript (paragraph two subsection “Conformational rearrangements in the post-cleavage synaptic complex”). This observation suggests that the protein might block proper bending of intact DNA, and changes in its conformation may be required to enable full DNA bending. Since the affected protein segment is closely connected to the catalytic tyrosine (on helix αN), its movement – perhaps facilitated by interactions with FtsK – may even explain the direct link between DNA bending and catalytic activation. Alternatively, cleavage might occur on partially bent DNA substrates, with additional bending occurring after cleavage. Interestingly, with the nicked substrates, the clash can be avoided thanks to the missing phosphate groups (see also our response to point 9).

While these hypotheses are intriguing, further experiments are required to evaluate the actual order of DNA bending and cleavage events.

9) The authors suggest that the suicide substrates used are cleaved by XerH because the nicks facilitate the required bend. Usually, the energetic cost of DNA bending is mostly related to unstacking of the bases, which nicking does not solve. Do the current structures explain how the nicks might be specifically allowing higher cleavage for this system – e.g., through relieving a strained backbone configuration? Or are they simply allowing the cleavage product to accumulate? Is the structure of an XerH catalytic mutant bound to nicked dif sites expected to be bent?

Modeling of a putative bent, but un-cleaved XerH-dif_H complex (based on the post-cleavage complex structure as described in point 8) indicated that the DNA might have somewhat strained backbone geometry, and the DNA appears to clash with the protein at the phosphate group exactly one nucleotide downstream of the scissile phosphate. Thus, we believe that the nicks favor the bent DNA conformation (paragraph two subsection “Conformational rearrangements in the post-cleavage synaptic complex”) by helping to increase the flexibility of the DNA strands at the bending points and relieving the strained backbone conformation, and they may also help avoid unfavourable interactions with the protein upon DNA bending due to the absence of the phosphate implicated in the clash. Accordingly, we would expect a catalytic mutant XerH to bind to nicked dif_H substrate roughly in the same conformation as the wild type protein, with a sharp DNA bend.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

[…] 1) The models in Figures 1, 2 and 6 cause some confusion as to how the structures might be interpreted in light of the XerH mechanism. Please consistently draft these figures with the strand going from 5' [on the left] to 3' [on the right] placed on 'top' in each of the views [i.e. viewing from the minor groove side in each case]. The authors should also acknowledge that an alternative model is consistent with their data in which FtsK is required to make the alternative synaptic complex and/or explain why such a model isn't favored.

We appreciate this comment and understand the reviewers’ concern regarding our drawings. Unfortunately, presenting the structures with viewing from the minor groove side would obscure visibility of important features of the structures (such as the DNA and the swapped C-terminal helix). Moreover, to conform with the previously published nomenclature of the difH arms we prefer to keep the ‘left arm’ on the left and the ‘right arm’ on the right side in our figures. This means that the schematic models in Figures 1 and 6 should be drawn with the strand going 3’ to 5’ on the top to be consistent with the structure figures. However, to address the reviewers’ concern and avoid confusion, we have now inverted the representation of the DNA sequence in Figure 2A (as well as in Figure 4A and in the figure supplements) so that it is now consistent with all other figures. In addition, we have clearly labeled both 5’ and 3’ ends of both strands on all DNA molecules in all drawings (see Figures 1, 2 and 6), and state the directionality of the DNA in the corresponding legends.

Concerning the possible alternative model for the role of FtsK, this possibility is now mentioned and discussed in the manuscript (Discussion section paragraph five).

2) Subsection “Modeling of E. coli XerC/D synaptic complex based on the XerH-difH structures”. In the absence of supporting data indicating that XerH will preferentially resolve HJs by bottom strand exchange, this claim should be removed. See also Discussion paragraph 7 in reference to this point.

All claims referring to HJ resolution by XerH have now been removed from the manuscript as requested.

3) Discussion section, paragraph one. Aside from the crystal structure reported here, there is other evidence that XerH/dif sites are assembled in an inactive synaptic complex in solution (e.g., Zawadzki et al. should be noted). Is there additional evidence for XerH/difH?

This point has now been noted in the text. For the XerH/difH system in particular, apart from our structural and biochemical data, we are not aware of any additional literature evidence showing that synaptic complexes initially assemble in an inactive conformation.

4) Authors' response #1 and related manuscript sections. What is the evidence that XerH does not cleave dif sites in vitro? This has been a tricky issue in the field, since it is hard to distinguish cleavage followed by rapid ligation from lack of cleavage. Unless the authors can provide data (e.g., using 5'-bridging phosphorothiolates) that demonstrate lack of cleavage, it would be better to take the position that it is not yet clear whether cleavage occurs in vitro. It is also difficult to explain activity in E. coli with such a model. In contrast, the behavior is readily explained if XerH bound to the left end behaves like XerC (as in an alternative model). Related to this is the question of whether cleavage is required for stabilization of the synaptic complex, independent of FtsK. This is true of λ int and several catalytic Cre mutants form much weaker synaptic complexes relative to wild-type Cre, suggesting that this could also be true in XerH as well. The nicked XerH/DNA complex also supports this idea. Thus, the synaptic complex structure observed for XerH could equally well represent a stable intermediate that can reach the transient bent configuration required for cleavage on its own. Please resolve/comment on this issue.

The reviewer is concerned about the possibility of “futile” DNA cleavage and ligation in the XerH-difH system. We have no evidence that this occurs in the XerH-difH system, but we cannot exclude the possibility based on our data. We now discuss the alternative possibilities raised by the reviewers in the manuscript.

5) Authors' response #6. The reviewers requested that the authors note the possibility that the synaptic complex structure may be influenced by crystallization conditions (low pH and/or lack of divalents). The authors acknowledge this in their response to reviewers, but do not appear to have modified the manuscript. Given that this structure is central to the authors' proposals for Xer recombination, the readers should be made aware that the conditions for the two structures are quite different. Please revise the manuscript accordingly.

We now point out the difference in crystallization conditions in the manuscript as requested.

6) The authors note that several unbound Xer crystal structures show different domain relationships, but continue to promote the idea that a specific conformational change takes place upon DNA binding (as shown in the video). To echo the original review, it is unlikely that the XerD alone structure represents a true "conformational intermediate," rather than just one of an ensemble of configurations present when Xer subunits are not bound to dif. The three Xer crystal structures in the PDB have their NTDs in completely different conformations. The ideas in the discussion and video illustrating that the NTD has to undergo a significant change when completing the clamp on the dif half-site was assumed many years ago by most people in the field and is true for all tyrosine recombinases. This point should clarified or else its discussion (and the movie) should be removed from the paper.

The respective discussion and movie have been removed in the current version of the manuscript.

7) The distinction between understanding the mechanism of recombination vs. regulation of this process is important, but not made clear (Abstract and Introduction section). The authors imply that the mechanism of XerC/XerD recombination is still unknown.

The last sentence of the Abstract has been rephrased to better distinguish between catalytic mechanism and its regulation; and the sentence that was in the Introduction has been deleted to avoid incorrect implications.

8) Introduction paragraph six. The structures do not delineate the pathway (delete this) but do shed light on potential regulatory mechanisms (reword).

Corrected.

9) In the same section the structure does not explain why activation is required. It suggests a possible model, as described by the authors, which differs from at least one other XerCD model.

Rephrased.

10) Results paragraph one. The current mechanism is based on much more than the series of Cre-loxP structures cited. A review covering tyrosine recombinase mechanisms (there are many) might be more appropriate here.

Corrected.

11) Subsection “Structure of the XerH-difH synaptic complex” need to be redrafted to accurately to describe the historical record (in doing so, the authors should consider point 9 and earlier comments regarding alternative models). The current text is incomplete and in part misleading. In the absence of FtsK, HJs formed by XerC were detected in early work and this led to the hypothesis that the XerC active conformation formed preferentially and FtsK switched the conformation of an XerC active complex to one in which XerD was active. Single-molecule FRET experiments by Zawadzki and colleagues led to the different hypothesis that this is not the case; rather, the major synaptic complex that forms in the absence of FtsK is an 'XerD potentially active conformation,' which is catalytically dead. A small proportion of XerC active complexes nevertheless form and are responsible for the XerC HJs. Nevertheless, the dominant pre-active XerD synaptic complexes are the ones activated by XerD. This is consistent with the results here, but the historical development of these ideas should be explicitly chronicled.

We now describe the historical development of the available models for XerC/D activation in the current version of the manuscript as requested. (Results section first paragraph)

12) Subsection “Structure of the XerH-difH synaptic complex” paragraph two. DNA duplex, not oligonucleotide.

Done.

Author details

1. Aleksandra Bebel

   Structural and Computational Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

   Contribution

   AB, Designed research and analyzed the results, Wrote the manuscript, Solved the crystal structures, Performed biochemical assays, Performed in vivo assays

   Competing interests

   The authors declare that no competing interests exist.

2. Ezgi Karaca

   Structural and Computational Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

   Contribution

   EK, Performed structural analysis and modeling, Contributed to writing of the manuscript

   Competing interests

   The authors declare that no competing interests exist.

3. Banushree Kumar

   Structural and Computational Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

   Contribution

   BK, Performed in vivo assays

   Competing interests

   The authors declare that no competing interests exist.

4. W Marshall Stark

   Institute of Molecular, Cell and Systems Biology, University of Glasgow, Glasgow, United Kingdom

   Contribution

   WMS, Wrote the manuscript, Provided E. coli plasmids and strains

   Competing interests

   The authors declare that no competing interests exist.

5. Orsolya Barabas

   Structural and Computational Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

   Contribution

   OB, Designed research and analyzed the results, Wrote the manuscript

   For correspondence

   barabas@embl.de

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0002-2873-5872

• 1,987

  Page views

• 312

  Downloads

• 10

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.