Fast Transport of RNA Granules by Direct Interactions with KIF5A/KLC1 Motors Prevents Axon Degeneration

Complex neural circuitry requires stable connections formed by lengthy axons. To maintain these functional circuits, fast transport delivers RNAs to distal axons where they undergo local translation. However, the mechanism that enables long distance transport of non-membrane enclosed organelles such as RNA granules is not known. Here we demonstrate that a complex containing RNA and the RNA-binding protein (RBP) SFPQ interacts directly with a tetrameric kinesin containing the adaptor KLC1 and the motor KIF5A. We show that binding of SFPQ to KIF5A/KLC1 motor complex is required for axon survival and is impacted by KIF5A mutations that cause Charcot-Marie-Tooth (CMT) Disease. Moreover, therapeutic approaches that bypass the need for local translation of SFPQ-bound proteins prevent axon degeneration in CMT models. Collectively, these observations show that non-membrane enclosed organelles can move autonomously and that replacing axonally translated proteins provides a therapeutic approach to axonal degenerative disorders.


32
Sensory and motor neurons transmit signals through axons than can exceed a meter in length.

33
Therefore, many axonal functions, including axonal survival pathways, depend on proteins that are locally 34 translated and replenished in axon terminals. Localized protein synthesis is enabled by the initial assembly

77
The RBP SFPQ is found in both cell bodies and axons of sensory neurons. However, the 78 mechanisms by which SFPQ and its critical RNA cargos are transported between these two locations is 79 not known. We utilized live cell imaging of DRG sensory neurons expressing Halo-tagged SFPQ to directly

100
The anterograde kinesins involved in axon transport are formed by two dimers of kinesin heavy 101 chain (KHC or KIF5) and two dimers of kinesin light chain (KLC) (Figure 2A). The KIF5 family is encoded by 102 three distinct genes, KIF5A, KIF5B and KIF5C, and the genome also contains several light chains, KLC1-4.

103
To identify motors that associate with SFPQ and might enable transport of these RNA granules, we took 104 an unbiased approach in which we immunoprecipitated endogenous SFPQ from DRG neurons and used 105 mass spectrometry to analyze the co-precipitated components. We detected known interactors, including     Together these data demonstrate that KIF5A and KLC1 are 131 appropriately localized to mediate transport of SFPQ-RNA granules from cell bodies to distal axons, and 132 thus KIF5A/KLC1 may represent a specialized motor for these non-membrane enclosed organelles.

135
As RNA is a critical component of the large SFPQ-containing granules that move rapidly within the 136 axons, we asked whether SFPQ that binds to KIF5A/KLC1 is also associated with RNA cargos. We expressed

142
Based on the above findings, we hypothesize that KIF5A/KLC1 distinctively mediates fast transport 143 of critical SFPQ-RNA granules from the cell soma to the axons. To identify the structural basis for this 144 specificity, we first asked whether the highly divergent C-terminal tail regions of KLC1 and KIF5A are 145 required for SFPQ binding. When we overexpressed either Myc-tagged WT KLC1 or its C-terminal mutant 146 (ΔTail) in HEK 293T cells, and assessed binding to SFPQ by co-precipitation studies, we find that the binding is reduced by approximately 50% in the absence of the C-terminal region of KLC1 ( Figure 3D-F). Similarly, 148 truncation of the highly variable tail region of KIF5A nearly abolished its interaction with SFPQ ( Figure 3G-149 I) although this did not prevent binding of KIF5A to KLC1 (Figure 3-figure supplement 1  supplement 1). When we mutated the critical tyrosine residue within the motif to alanine (Y527A) the 161 binding between SFPQ and KIF5A/KLC1 was dramatically reduced, demonstrating that this Y-acidic motif 162 is required for binding to KIF5A/KLC1 motor complex ( Figure 4B and 4C). To assess whether SFPQ binds 163 directly to KLC1 without requiring a membranous organelle or an adaptor component, we purified human 164 KLC1 and used isothermal titration calorimetry (ITC) to test direct binding by a long SFPQ peptide that 165 spans the Y-acidic motif. The SFPQ peptide binds directly to KLC1 with a Kd of 3.8 ± 2.3 μM and a binding 166 stoichiometry of 1 ( Figure 4D and Table 1). Consistent with data from co-immunoprecipitation studies 167 above, Y527A mutation prevents the SFPQ peptide from binding directly to KLC1 in ITC assays ( Figure 4D).

168
Thermodynamic parameters of ITC measurements are summarized in Table 1. Together these data 169 demonstrate that SFPQ directly binds to KLC1 in a process that relies on the Y-acidic motif and is abrogated 170 by the Y527A mutation. These studies suggest that SFPQ RNA-transport granules might directly associate with microtubule-dependent motors rather than requiring a membrane platform for intracellular 172 transport.

174
Direct binding of SFPQ to KIF5A/KLC1 is required for its transport in axons.

175
The Y527A mutant of SFPQ provides a tool that can be used to ask whether the direct interaction 176 between SFPQ and kinesin motors is responsible for autonomous transport of SFPQ-RNA granules along 177 microtubules. Such a transport mechanism would contrast with previous models for transport of non-178 membrane enclosed granules, as RNA granules that contain b-actin mRNA "hitchhike" on membrane

192
It is striking that the Y-acidic motifs in JIP1 and other proteins that link kinesins to membranous 193 organelles are usually located within highly accessible regions such as the carboxy terminus, while the Y-194 acidic motif in SFPQ is instead located within the highly structured coiled-coil domain. These structural differences suggest that SFPQ-RNA granules may interact with kinesin in a different manner than do 196 membranous organelles. Previous studies identified sequences within KLC1 that are not present in KLC2 197 and that specify binding to JIP1. One key residue in KLC1 is N343 within the TPR4 region of KLC1; mutation 198 of this residue to a serine, as observed in KLC2, abrogates interaction between JIP1 and KLC1 (Zhu et al.,       Although it is widely accepted that RNA-granules are non-membrane enclosed organelles that 268 move by microtubule-dependent transport, how such RNA-granules associate with motors and move 269 through the axoplasm is not yet known. Our data demonstrate a direct interaction between SFPQ and the 270 kinesin-1 cargo adaptor complex KLC1, and we show that this interaction is required for autonomous 271 axonal transport. In contrast to this direct transport system, a recent study by Liao et al. demonstrated

302
Among the three KIF5 genes encoding the kinesin-1 family of motors, the KIF5A gene is the only 303 one associated with the human neurological diseases CMT2D, HSP and ALS. Our results demonstrate that 304 binding of SFPQ to KLC1 complexed with KIF5A rather than KIF5B or KIF5C enables transport of SFPQ-RNA 305 granules and promotes axon survival. Based on these findings, we postulate that defective transport of 306 SFPQ-RNA granules is a major contributor to KIF5A-associated neurodegenerative disorders, rather than 307 axon degeneration in these disorders being the result of a generalized impairment of transport.

508
Mass Spectrometry: Antibody-conjugated protein G beads from KLC1 and SFPQ immunoprecipitates were

523
The mass spectrometer was programmed to perform a combination of targeted (Parallel Reaction