Skip to main content

Mechanism of Holliday junction dissolution by the Bloom’s Syndrome complex

Final Report Summary - BLMCOMPLEX (Mechanism of Holliday junction dissolution by the Bloom’s Syndrome complex)


Mutations in the BLM gene give rise to Bloom’s Syndrome (BS) (Ellis et al., 1995), a rare genetic disorder that includes the development of various types of cancers at a young age. The product of the BLM gene is the BLM DNA helicase, a member of the RecQ family of DNA helicases (Ellis et al., 1995). The family is named after the Escherichia coli orthologue RecQ, which is required for the suppression of illegitimate recombination (Hanada et al., 1997). The human RECQ proteins, of which there are five, exhibit similar biochemical properties and are implicated in common DNA metabolic processes. However, these proteins do not have overlapping or redundant cellular functions, as mutations in different RECQ proteins lead to distinct clinical syndromes (Hickson, 2003). BLM helicase has been shown to be involved in the regression of stalled replication forks (Ralf et al., 2006), and also participates in DNA double-strand break repair mediated by homologous recombination (HR) (Karow et al., 2000). The diagnostic feature of cells derived from BS patients, that lack functional BLM, is an elevated frequency of sister chromatid exchanges (SCEs) that leads to an increased loss of heterozygosity (LOH). LOH is one of the mechanisms capable of promoting complete loss of function of tumour-suppressor genes in human cancers after crossover formation.
BLM interacts with three other proteins to form the BTRR complex: Topoisomerase IIIα (TOPOIIIα) (Johnson et al., 2000; Wu et al., 2000)], RMI1 (Yin et al., 2005) and RMI2 (Liu and West, 2008; Singh et al., 2008; Xu et al., 2008). This complex promotes the ‘dissolution’ of DNA structures containing double Holliday junctions (dHJs) that arise during HR (Wu and Hickson, 2003). In the absence of BLM, HJs are resolved by the actions of the cellular HJ resolvases MUS81-EME1, SLX1-SLX4 and GEN1 (Wyatt and West 2014).
The BTRR complex removes dHJs by a process known as HJ ‘dissolution’, so named because the products of the reaction are always non-crossovers. The process involves BLM-mediated branch migration (Karow et al., 2000), in which the two HJs are brought together by convergent migration in order that TOPOIIIα can remove the resultant hemicatenane. However, the details of how convergent migration of the two HJs might be imposed are currently unknown. One aspect of this project was to determine the mechanism by which the BTRR complex actively induces convergent directionality, and we proposed three alternative models. In the first, BTRR complexes bind to each HJ and translocate freely (i.e. bidirectionally) along the DNA: thus, only in around 25% of the cases, branch migration would be convergent. In the second, BTRR complexes assemble on each HJ, but direct interactions between two BTRR complexes allows spooling of the DNA through the complex in order to provide the hemicatenane structure for topoisomerase-mediated strand passage. In the third model, two BLM monomers bind each HJ, interact together to mediate spooling, and the TRR complex subsequently assembles to form a BTRR-bound hemicatenane. In models 2 and 3, the BLM helicase provides twin motor pumps for branch migration, in a manner analogous to the bacterial RuvAB branch migration complex (Hiom et al., 1996; Parsons et al., 1995).
To determine which of these dissolution models is correct, and to understand at the mechanistic level the processing of dHJs in human cells, the objectives of this project were as follows:
Objective 1: Determination of the mechanism of dHJ dissolution by the BTRR complex.
To determine which of the above models is correct, we have utilised electron microscopic and biochemical techniques. The tools needed are purified BTRR complex and BLM helicase and an appropriate DNA substrate containing a dHJ. In previous studies, the four proteins that constitute the BTRR complex were purified individually from bacteria and/or yeast cells. However, to enable more efficient purification of a stable BTRR complex, I used a multi-BAC vector that allows over-expression of all four proteins in insect cells. I successfully completed the cloning and purification processes for both the BTRR complex and the BLM helicase, with tags that do not destabilise the complex. In collaboration with Dr Alessandro Costa’s laboratory, we continue to work on obtaining the structure of the BTRR complex in its apo and DNA-bound forms, as well as the BLM helicase bound to an oligonucleotide-based HJ. The techniques include transmission electron microscopy using both negative staining with uranyl acetate and cryo-EM. The process of determining the structures is currently underway, and 2D classes have already been obtained for the BLM-HJ complex.
Size exclusion chromatography and native PAGE analyses of the BLM helicase protein preparations, as well as the EM data, indicate that BLM dimers are the prevalent state of the protein in solution, even in the absence of DNA.
Visualisation of the full BTRR complex has been possible using negative and positive staining. I have been able to produce and visualize the plasmid-based dHJ DNA substrate, and have optimized the binding conditions to observe the BTRR-dHJ complex by electron microscopy. Our initial results indicate that two BTRR complexes bind the dHJ at the junction points. This has been confirmed using immuno-gold techniques where gold-labelled antibodies against each of the protein components were able to detect protein dimers in the presence of the dHJ plasmid. These results were recently presented at the FASEB Conference on Helicases and Nucleic-Acid Based Machines at Steamboat Springs, Colorado, USA.
Objective 2: Define the interaction surfaces between BLM or BTRR complexes and how direct interactions contribute to the catalytic activities of the complex.
Our initial results indicate that BLM forms dimers in solution. In collaboration with Dr Alessandro Costa (LRI), we have used peptide arrays to identify the minimal essential regions of interaction between two BLM monomers and also between the components of the BTRR complex. Extensive protein alignments of the homologs of BLM allowed me to locate conserved residues within the interaction domains. I used site-directed mutagenesis to produce point mutations in this conserved region and cloned the BLM mutants into both mammalian expression vectors and the MultiBac system, which allows expression in insect cells. These tools will allow me to study the phenotype of these mutants both in vitro and in vivo (as indicated below).
In future work, I will use the purified mutant and wild-type versions of the enzymes to conduct in vitro pull-down assays that should confirm the predicted surfaces of interaction in the complex and between two different BLM monomers. I will also perform the biochemical characterisation of these mutants in comparison with the wild-type BLM and BTRR complex. Indeed, fine mapping of the interaction surfaces will not only allow us to understand the mechanism of dHJ dissolution, but may also provide novel information relating to the distinct phenotypes of individuals with Bloom’s Syndrome.
Objective 3: Determine the importance of BLM/BTRR dimer formation for in vivo functionality and correlate with mutations found in BS patients.
In order to study the in vivo phenotypes of the mutants derived from the biochemical studies described in Objective 2, I cloned the mutant and wild-type versions of BLM into mammalian expression vectors, and will determine the phenotype of Bloom’s Syndrome patient cells lines complemented with these mutant versions of BLM. I have also carried out similar experiments using the commercially available HAP1 BLM-/- knock out cell line, and in the future will generate my own CRISPR KO model using FLP-IN 293 cell lines. The complemented cell lines will be used to determine whether the results obtained in vitro are also true within a more physiological context. The complemented cell lines will also be used for immunolocalisation studies, as well as the co-immunoprecitipitation of components of the BTRR complex. We therefore hope to be able to determine how different mutations in BLM affect the recruitment of the BTRR complex to sites of DNA damage. We will analyse the sensitivity of mutant-complemented cell lines to DNA damaging agents. These studies will allow us to determine the impact of the BTRR complex/BLM helicase dimerization on its DNA repair-related functions. We will also determine the frequency of SCEs in our mutant-complemented cell lines, and determine whether dimerization of BLM or inability to form the BTRR complex has a significant impact on SCE formation and crossover suppression.
Objective 4: Single and double HJ resolution by the human resolvases.
Human cells possess two pathways of endonuclease-mediated HJ resolution that act in a coordinated fashion at G2/M and anaphase, respectively. These involve the SLX1-SLX4, MUS81-EME1 and GEN1 proteins, which are thought to act upon double HJs that escape the attentions of the BTRR complex, and also upon single HJs that BTRR cannot process. However, the cleavage of double HJs by human resolvases had not been demonstrated.
We therefore carried out mechanistic analyses of the actions of these proteins on plasmid-based single and double HJ substrates. We found that purified SLX1-SLX4 and GEN1, and to a limited extent MUS81-EME1, cleave recombination intermediates containing either single or double HJs in vitro. Our results show that both pathways of HJ resolution present in human cells are able to efficiently deal with a variety of covalently closed recombination intermediates that arise during the repair of double strand breaks in DNA.
This latter section of the work program has now been completed, and was presented at the ABCAM conference Mechanisms of Recombination: 50th Anniversary Meeting of the Holliday Model (Alicante, Spain). A manuscript including this research is now being prepared for publication.