24 25 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 Biological Cryo-Imaging Group eBIC Visualising a DNA repairmachine Related publication: Shakeel S., Rajendra E., Alcón P., O’Reilly F., Chorev D. S., Maslen S., Degliesposti G., Russo C. J., He S., Hill C. H., Skehel J. M., Scheres S. H.W., Patel K. J., Rappsilber J., Robinson C.V. & Passmore L. A. Structure of the Fanconi anemiamonoubiquitin ligase complex. Nature 575 , 234-237 (2019). DOI: 10.1038/s41586-019-1703-4 Publication keywords: DNA repair; DNA crosslink; E3 ubiquitin ligase; Cryo-EM; Mass spectrometry; Cancer T he human genetic illness Fanconi Anaemia (FA) causes abnormal development, bone marrow failure and a lifetime risk of developing cancer. FA is known to deactivate one of more than 20 genes that code for proteins that function collectively to repair damaged DNA and are known as the FA repair pathway. The FA pathway repairs crosslinks within and between strands of DNA. The multi-protein FA core complex acts at the centre of this pathway to identify and signal the site of DNA crosslinks. When the FA pathway is defective, it results in human disease. Although the FA core complex had been characterised in cells, it had not been fully reconstituted in the lab. Researchers isolated an FA core complex and determined its structure, using Cryogenic Electron Microscopy (cryo-EM) at the Electron Bio- Imaging Centre (eBIC) to image the complex at high resolution. They created 3D reconstructions from the images and used them to build models for how the FA core complex assembles and functions. Using the models, researchers can now make new hypotheses for how the FA core complex functions and begin to define this DNA repair pathway in molecular terms. This will allow them to explain why specific patient mutations inactivate the pathway – and in the future, we may be able to predict whether certain changes might cause disease. This was a long-term investigation, which took ten years to go fromproject conception to structure determination. The success of the project in the final stage was dependent on the excellent support from eBIC. DNA crosslinks occur during normal cellular metabolism, or after exposure to chemicals such as chemotherapeutic drugs or alcohol. Crosslinks block the replication of DNA and transcription, and therefore they must be repaired to maintain genome integrity. Defects in the ability to repair DNA crosslinks result in Fanconi Anaemia (FA), a human disorder characterised by developmental defects, bone marrow failure, and predisposition to cancer 1 . FA patients harbour mutations in any one of 22 FANC (FA complementation group) genes, and these genes encode proteins that function in DNA crosslink repair 2 . At the heart of the FA DNA repair pathway is a megadalton, multi-protein E3 ubiquitin ligase, the FA core complex. This complex monoubiquitinates its substrate proteins at the site of DNA crosslinks, creating a signal that recruits other repair factors to excise and repair the DNA lesions. X-ray crystal structures of the RING finger subunit (FANCL) had been previously determined 3 and these provided important insights into how FANCL binds to its cognate E2 enzyme. A native FA core complex had previously been purified from cells 4 , showing that the catalytic module of the complex is comprised of three subunits (FANCL, FANCB and FAAP100). This catalytic module is a more specific and more efficient E3 ligase than FANCL alone. Studies using negative stain electron microscopy indicated that the catalytic module is a dimer of heterotrimers 5 . Still, progress in understanding the mechanistic basis of ubiquitination by the FA core complex had been slowed by the challenges in obtaining FA core complex in sufficient quantity and purity, an incomplete description of subunit functions and interactions, and a lack of high-resolution structures. In the present study, methods were developed to isolate an active FA core complex by simultaneously over-expressing all eight subunits in insect cells, allowing purification of milligram quantities of a pure recombinant complex. This sample, as well as a subcomplex containing a subset of the subunits, was imaged using cryo-EM and 3D structures were obtained (Fig. 1). Computational methods were used to reduce blurring in peripheral parts of the complex (multi-body refinement and particle subtraction followed by focussed classification and refinement), and newmethods were developed for local symmetry averaging. Together, these methods increased the resolution of the 3D reconstructions, allowing visualisation of the FA core complex for the first time. Previously determined structures accounted for only 12% of the mass of the FA core complex, and the sequences of many subunits did not resemble proteins of known structure, making it difficult to interpret the cryo-EM data at 4.2 Å resolution. Thus, the sample was also studied using complementary methods including native mass spectrometry, crosslinking mass spectrometry and biochemical reconstitution. By integrating these data, models for seven of the eight subunits could be built into the cryo-EMmaps (Fig. 1c). The three subunits of the catalytic module were each present in two copies within the FA core complex, resulting in a pseudosymmetry in the centre of the complex (Fig. 2). Interestingly, the structure showed that FANCB, FANCL and FAAP100 not only act as the catalytic module, but they are also the structural core of the complex, providing a scaffold to assemble the remaining five subunits. Only one copy of each of the remaining subunits could be identified in the maps. Examinationofthemodelsshowedthattwosubunits(FANCBandFAAP100) have strikingly similar structures despite no apparent sequence similarity. The models also showed that there are two RING finger subunits (FANCL), located at opposite ends of the FA core complex (Fig. 2b). RING fingers are critical for the E3 ligase function. The two copies of FANCL are asymmetrically positioned – one is located in the base and is surrounded by the subunits of the substrate recognition module. All three domains of FANCL are visible in the base of the cryo-EM map. In contrast, the second FANCL is at the top of the complex and the RING domain is not visible. It is likely that this domain is flexible and therefore blurred out in the maps. The different conformations of the two FANCL subunits suggest that they play distinct roles within the complex, and extensive interactions with other subunits likely explain why the catalytic module is a better E3 ligase than FANCL alone 4 . Asymmetric dimerisation has been observed for other RING fingers and therefore may be a general feature of E3 ligases. Finally, the distribution of patient mutations weremapped onto the overall structure (Fig. 3). Intriguingly, almost all of the patient mutations are found in the structural periphery of the FA core complex. This suggests that mutations are not well tolerated in the central core of the complex because they would likely compromise its structural integrity, thereby abrogating ubiquitin ligase activity. In agreement with this, the few patients that have mutations within the structural core of the complex (in FANCB or FANCL) are severely afflicted. On the other hand, mutations in the periphery would not disrupt the structural integrity of the complex and are therefore more tolerated. The new structural model of the FA core complex provides insights into monoubiquitination by this large, multi-subunit E3 ligase in DNA repair. The structure and the reconstituted monoubiquitination system will be used in future studies to understand the mechanisms and regulation of FA-mediated DNA repair. References: 1. Kottemann M. C. et al. Fanconi anaemia and the repair ofWatson and Crick DNA crosslinks. Nature 493 , 356-363 (2013). DOI: 10.1038/nature11863 2. Walden H. et al. The Fanconi Anemia DNA repair pathway: structural and functional insights into a complex disorder. Annu. Rev. Biophys. 43 , 257- 278 (2014). DOI:10.1146/annurev-biophys-051013-022737 3. Hodson C. et al. Structure of the human FANCL RING-Ube2T complex reveals determinants of cognate E3-E2 selection. Structure 22(2) , 337-344 (2014). DOI: 10.1016/j.str.2013.12.004 4. Rajendra E. et al. The genetic and biochemical basis of FANCD2 monoubiquitination. Mol. Cell. 54(5) , 858-869 (2014). DOI: 10.1016/j.molcel.2014.05.001 5. Swuec P. et al. The FA core complex contains a homo-dimeric catalytic module for the symmetric mono-ubiquitination of FANCI-FANCD2. Cell. Rep. 18(3) , 611-623 (2017). DOI: 10.1016/j.celrep.2016.11.013 Funding acknowledgement: Medical Research Council MC_U105192715 Corresponding author: Dr Lori A Passmore, MRC Laboratory of Molecular Biology,
[email protected] Figure 1: (a) Selected 2D class averages of the FA core complex. One class appears to be symmetric (labelled). (b) Focused classification and refinement (top, base) or multibody refinement (middle) resulted in three independent cryo-EMmaps that are shown separately. (c) Model of FA core complex subunits (cartoon) fitted into the EM density (isosurface), coloured by assigned subunits. The green star marks a channel with a diameter of approximately 23 A. Figure panels reproduced from Shakeel et al. (2019). Figure 2: (a) Surface representation of the FA core complex model, highlighting FANCB and FAAP100, which act as a molecular scaffold. FANCB, orange; FAAP100, yellow; regions where we are unable to distinguish FANCB and FAAP100, yellow–orange. (b) Surface representation of the FA core complex model, highlighting the two copies of FANCL (left). On the right, the two models of FANCL are shown in cartoon representation, fitted in the cryo-EMmap. Density for the URD and RING domains is not well defined in the top copy. Figure panels reproduced from Shakeel et al. (2019). Figure 3: Distribution of patient mutations per subunit are indicated on the FA core complex by red dots. Most mutations are found in the structural periphery of the complex.