The Endosomal Sorting Complexes Required for Transport (ESCRT) proteins mediate fundamental membrane remodeling events that require stabilizing negative membrane curvature. These include endosomal intralumenal vesicle formation, HIV budding, nuclear envelope closure and cytokinetic abscission. ESCRT-III subunits perform key roles in these processes by changing conformation and polymerizing into membrane-remodeling filaments. Here, we report the 4 Å resolution cryo-EM reconstruction of a one-start, double-stranded helical copolymer composed of two different human ESCRT-III subunits, CHMP1B and IST1. The inner strand comprises “open” CHMP1B subunits that interlock in an elaborate domain-swapped architecture, and is encircled by an outer strand of “closed” IST1 subunits. Unlike other ESCRT-III proteins, CHMP1B and IST1 polymers form external coats on positively-curved membranes in vitro and in vivo. Our analysis suggests how common ESCRT-III filament architectures could stabilize different degrees and directions of membrane curvature.

Mutations in ATP13A2 (PARK9), encoding a lysosomal P-type ATPase, are associated with both Kufor-Rakeb syndrome (KRS) and neuronal ceroid lipofuscinosis (NCL). KRS has recently been classified as a rare genetic form of Parkinson's disease (PD), whereas NCL is a lysosomal storage disorder. Although the transport activity of ATP13A2 has not been defined, in vitro studies show that its loss compromises lysosomal function, which in turn is thought to cause neuronal degeneration. To understand the role of ATP13A2 dysfunction in disease, we disrupted its gene in mice. Atp13a2(-/-) and Atp13a2(+/+) mice were tested behaviorally to assess sensorimotor and cognitive function at multiple ages. In the brain, lipofuscin accumulation, α-synuclein aggregation and dopaminergic pathology were measured. Behaviorally, Atp13a2(-/-) mice displayed late-onset sensorimotor deficits. Accelerated deposition of autofluorescent storage material (lipofuscin) was observed in the cerebellum and in neurons of the hippocampus and the cortex of Atp13a2(-/-) mice. Immunoblot analysis showed increased insoluble α-synuclein in the hippocampus, but not in the cortex or cerebellum. There was no change in the number of dopaminergic neurons in the substantia nigra or in striatal dopamine levels in aged Atp13a2(-/-) mice. These results show that the loss of Atp13a2 causes sensorimotor impairments, α-synuclein accumulation as occurs in PD and related synucleinopathies, and accumulation of lipofuscin deposits characteristic of NCL, thus providing the first direct demonstration that null mutations in Atp13a2 can cause pathological features of both diseases in the same organism.

Recently mutations in ubiquilin-2 were identified in patients with amyotrophic lateral sclerosis (ALS) and ALS/dementia providing direct evidence for the importance of this protein in neurodegenerative diseases. Histological studies have suggested that ubiquilin-1/-2 are associated with various pathological inclusions including Lewy bodies in Parkinson’s disease, neurofibrillary tangles in Alzheimer’s disease, polyQ inclusions in expansion repeat diseases and various proteinopathies associated with ALS and frontotemporal dementia. Using specific ubiquilin-2 antibodies and a series of transgenic mouse models of proteinopathies associated with neurodegenerative disease, we show that ubiquilin-2 preferentially associates with huntingtin polyQ expansion aggregates compared to α-synuclein, tau and several other types of protein inclusions. These results were confirmed by similar findings for ubiquilin-1 and -2 in human brain tissue sections, where accumulation was observed in huntingtin inclusions, but only infrequently in other types of protein inclusions. In cultured cells, ubiquilin-2 associates with huntingtin/polyQ aggregates, but this is not compromised by disease-causing mutations. Although ubiquilin proteins can function as chaperones to shuttle proteins for degradation, there is persistent co-localization between ubiquilin-2 and polyQ aggregated proteins during disease progression in the N586-82Q-C63 Huntington’s disease mouse model. Thus, the co-localization of ubiquilin-2 with the huntingtin aggregates does not appear to facilitate aggregate removal.

Frontotemporal lobar degeneration (FTLD) has been subdivided based on the main pathology found in the brains of affected individuals. When the primary pathology is aggregated, hyperphosphorylated tau, the pathological diagnosis is FTLD-tau. When the primary pathology is cytoplasmic and/or nuclear aggregates of phosphorylated TAR-DNA-binding protein (TDP-43), the pathological diagnosis is FTLD-TDP. Notably, TDP-43 pathology can also occur in conjunction with a number of neurodegenerative disorders; however, unknown environmental and genetic factors may regulate this TDP-43 pathology. Using transgenic mouse models of several diseases of the central nervous system, we explored whether a primary proteinopathy might secondarily drive TDP-43 proteinopathy. We found abnormal, cytoplasmic accumulation of phosphorylated TDP-43 specifically in two tau transgenic models, but TDP-43 pathology was absent in mouse models of Aβ deposition, α-synucleinopathy or Huntington’s disease. Though tau pathology showed considerable overlap with cytoplasmic, phosphorylated TDP-43, tau pathology generally preceded TDP-43 pathology. Biochemical analysis confirmed the presence of TDP-43 abnormalities in the tau mice, which showed increased levels of high molecular weight, soluble TDP-43 and insoluble full-length and ~35 kD TDP-43. These data demonstrate that the neurodegenerative cascade associated with a primary tauopathy in tau transgenic mice can also promote TDP-43 abnormalities. These findings provide the first in vivo models to understand how TDP-43 pathology may arise as a secondary consequence of a primary proteinopathy.Electronic supplementary materialThe online version of this article (doi:10.1007/s00401-013-1123-8) contains supplementary material, which is available to authorized users.