36 37 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 2 1 / 2 2 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 2 1 / 2 2 Understanding the mechanism of axon degeneration Related publication: Figley, M. D., Gu,W., Nanson, J. D., Shi, Y., Sasaki, Y., Cunnea, K., Malde, A. K., Jia, X., Luo, Z., Saikot, F. K., Mosaiab, T., Masic, V., Holt, S., Hartley-Tassell, L., McGuinness, H. Y., Manik, M. K., Bosanac, T., Landsberg, M. J., Kerry, P. S., Mobli, M., Hughes, R.O., Milbrandt, J., Kobe, B., DiAntonio, A., Ve, T. SARM1 is a metabolic sensor activated by an increased NMN/NAD+ ratio to trigger axon degeneration. Neuron 109 , 1118-1136.e11 (2021). DOI: 10.1016/j.neuron.2021.02.009 Publication keywords: ARM domain; NADase; TIR domain; X-ray crystallography; Allostery; Cryo-EM S ARM1 (sterile alpha and TIR motif 1) is a key player in nerve fibre (axon) degeneration and a promising new therapeutic target for prevalent neurological diseases, including peripheral neuropathies and traumatic brain injury. In healthy nerve cells, SARM1 is present but inactive. Disease and injury activate SARM1, which results in rapid breakdown of the essential 'helper molecule' nicotinamide adenine dinucleotide (NAD + ) and ultimately destruction of the axon. Defining the molecular mechanisms upstream and downstream of SARM1 enzyme activity can yield inhibitors as leads to anti-neurodegenerative disease therapeutics. To understand the auto-inhibition and activation mechanisms of SARM1 at the molecular level, researchers needed to determine the structure of SARM1 in inactive and active states. Using data collected at the electron Bio-Imaging Centre (eBIC for Industry) , the Midlands Regional Cryo-EM Facility at the University of Leicester, and the Centre for Microscopy and Microanalysis at the University of Queensland, they determined a ~3.1 Angstrom resolution Cryo-Electron Microscopy (Cryo-EM) structure of human SARM1 in the inactive state. Using X-ray crystallography, they also determined a structure of the regulatory domain of Drosophila SARM1 in complex with an allosteric activator. Together these structures explain how this protein maintains an inhibited (off) state and the critical steps required for its activation. These results allowed them to propose a molecular mechanism for SARM1 activation. They also provide a molecular framework for the design of novel drugs targeting a wide range of neurodegenerative diseases. When the normal functions of nerve fibres (axons) are compromised by insults such as trauma, transport blockade, or chemical toxicity, distinct morphological and molecular changes known as Wallerian degeneration result in cytoskeletal disassembly and granular degeneration of the axon distal to the injury site. Axon loss is a common theme shared by several neurodegenerative diseases, including neuropathies, traumatic brain injury, and glaucoma, but no treatments currently exist that effectively target axonal breakdown. The protein SARM1 is a central player in axon loss 1,2 . In healthy nerve cells, SARM1 is present but inactive. Disease and injury activate SARM1, which results in rapid breakdown of the essential 'helpermolecule' nicotinamide adenine dinucleotide (NAD + ) and ultimately destruction of the axon 3-4 . Genetic knockout of SARM1 has previously been shown to prevent axon loss in models of neurological disease, including peripheral neuropathies and traumatic brain injury, whereas hyperactive SARM1 variants have recently been shown to be enriched in patients with amyotrophic lateral sclerosis. Agents that block SARM1 function could therefore be exciting therapeutic candidates for a variety of neurological conditions 5 . SARM1 consists of an N-terminal regulatory armadillo repeat (ARM) domain, two tandem sterile alpha motif (SAM) domains, and a C-terminal Toll/ interleukin-1 receptor (TIR) domain with self-association-dependent NAD + glycohydrolase (NADase) activity. In healthy and intact axons, SARM1 activation is restrained by NMNAT2 (nicotinamide mononucleotide adenylyltransferase 2), which is an axonal enzyme that synthesises NAD + using NMN (nicotinamide mononucleotide) and ATP as substrates. NMNAT2 is a labile protein and is rapidly degraded after axon injury, which leads to a rise in the axonal levels of NMN. Loss of NMNAT2 causes axon fragmentation that is dependent on SARM1, suggesting that the increase in axonal NMN levels caused by the absence of NMNAT2 can activate SARM1. In the publication, the authors first show, using a combination of cellular, biochemical andbiophysical assays, thatboththeNADaseandaxondegeneration functions of SARM1 are activated by an increase in the ratio of NMN to NAD + and that these two metabolites directly compete for binding to SARM1’s ARM domain. Next, to provide structural insight into this activation mechanism, they determined the cryo-EM structure of human SARM1 (hSARM1) in ligand-free inactive state, and the crystal structure of the ARMdomain of Drosophila SARM1 (dSARM1) in complex with the allosteric activator NMN. Cryo-EM data was collected at eBIC, the Midlands Regional Cryo-EM Facility at the University of Leicester, and Centre for Microscopy and Microanalysis at the University of Queensland. From the initial EM data collected at eBIC, only features corresponding to SAM domain rings were visible in the 2D classes, suggesting that the SARM1 particles are not very stable, leading to disordered Biological Cryo-Imaging Group eBIC ARMandTIR domains after freezing. By using grids that had been preparedwith a 6 nm continuous carbon film, the quality of the sample significantly improved; intact SARM1 particles with features corresponding to the ARMandTIR domains were clearly visible in the 2D classes (Fig. 1(a), and a 3D density map with an overall resolution of 3.1 Å was obtained, which allowed modelling of the entire octameric complex (Fig. 1(b)). The structure reveals a ring-shaped octameric protein complex (Fig. 1B). The tandem SAM domains self-associate head-to-tail to form the central inner ring. The SAM ring is surrounded by the ARM domains, each containing 8 ARM repeat motifs (ARM 1-8). The ARM domains have an open conformation in the absence of the activator NMN. This conformation is responsible of the inactive nature of the ligand-free SARM1 and is altered upon NMN binding. The TIR domains are wedged between the ARMdomains on the outer periphery for the ring (Fig. 1B). The NADase activity of the TIR domains depends on their self-association, but the ARM-TIR interaction prevents such self-association. To further elucidate the mechanism of NMN-dependant activation of SARM1, the crystal structure of NMN-bound Drosophila SARM1 ARM domain was determined. The ligand-free and NMN-bound ARM domain structures of the human and Drosophila proteins show conformational differences, with an RSMD of 2.56 Å over 292 Cα atoms. The NMN-bound structure shows the N-terminal ARM repeats 1-2 (ARM1-2), and the C-terminal ARM6-8 collapsing inwards around the ligand-binding site, resulting in a more compact conformation (Fig. 2(a)). There is no major change in the region spanning ARM3-5, which also comprises the TIR domain interface, indicating that NMN- related conformational changes do not directly affect ARM:TIR binding. Given that ARM8 is tethered to the octameric SAM ring and likely exhibits limited movement, the conformational changes upon NMN binding would translate to rotation of the ARMdomain relative to the SAMoctameric ring. Superimposition of ARM8 from the NMN-bound structure on the ligand-free structure shows significant movement of ARM3 towards the adjacent ARM domain, potentially leading to steric clash (Fig. 2(b)). 2D class averages of the hSARM1 octamer showing partial ARM domain rings indicate this steric clash may be responsible for disrupting the ARM:SAM interface and abolishing the ARM:TIR interaction (Fig. 1(a)). Based on the structural analyses, the authors propose a model for SARM1 activation (Fig. 2(c)). In the ligand-free state, SARM1 is kept inactive simply by keeping the TIR domains spatially separated. Binding of the allosteric activator NMN results in collapse and rotation of the ARM domains, resulting in destabilisation of the ARM domain ring and release of the TIR domains. The TIR domains are then able to self-associate to formcatalytic sites capable of cleaving NAD + , which ultimately leads to destruction of the axon. In conclusion, the new understanding of the molecular basis of SARM1 regulation and the identification of the allosteric NMN-binding site provide a foundation for development of therapeutics for a wide range of human neurodegenerative diseases. References: 1. Gerdts, J. et al. Sarm1-mediated axon degeneration requires both SAM and TIR interactions. Journal of Neuroscience 33 , 13569–13580 (2013). DOI: 10.1523/JNEUROSCI.1197-13.2013 2. Osterloh, J. M. et al. dSarm/Sarm1 is required for activation of an injury- induced axon death pathway. Science 337 , 481–484 (2012). DOI: 10.1126/ science.1223899 3. Essuman, K. et al. The SARM1 Toll/Interleukin-1 receptor domain possesses intrinsic NAD + cleavage activity that promotes pathological axonal degeneration. Neuron 93 , 1334-1343.e5 (2017). DOI: 10.1016/j. neuron.2017.02.022 4. Horsefield, S. et al. NAD + cleavage activity by animal and plant TIR domains in cell death pathways. Science 365 , 793–799 (2019). DOI: 10.1126/science. aax1911 5. Krauss, R. et al. Axons matter: the promise of treating neurodegenerative disorders by targeting SARM1-mediated axonal degeneration. Trends in Pharmacological Sciences 41 , 281–293 (2020). DOI: 10.1016/j. tips.2020.01.006 Funding acknowledgement: The work was supported by the National Health and Medical Research Council (NHMRC grants 1107804 and 1160570 to B.K. and T.V., 1071659 to B.K., and 1108859 to T.V.), the Australian Research Council (ARC) Laureate Fellowship (FL180100109 to B.K.) and DisarmTherapeutics. T.V. received ARC DECRA (DE170100783) funding. Corresponding author: Prof Bostjan Kobe, University of Queensland,
[email protected] Dr Thomas Ve, Griffith University,
[email protected] Figure 1: Overall structure of ligand-free hSARM1. (a). Representative 2D class averages of “SAM domain only” particles from initial data and “intact” particles obtained using a continuous carbon film are highlighted by red and blue boxes, respectively; (b). An electrostatic potential density map of ligand-free hSARM1 (EMD-23278); Map regions corresponding to the ARM, SAM, and TIR domains are displayed in blue, green, and pink, respectively. Figure 2: Model of SARM1 activation upon NMN binding; (a). Structural superposition of NMN-bound dSARM1 ARM (grey; residues 370-676) and ligand-free hSARM1 ARM (blue; residues 62-400); (b). Structural superposition of ARM8 in NMN-bound dSARM1 ARM (grey) and ligand-free hSARM1 ARM (blue) suggests that the ARM domain would rotate and potentially clash with the ARM domain of adjacent subunits (blue) upon NMN binding; (c). Schematic model of SARM1 activation.