Membranes in cells have defined distributions of lipids in each leaflet, controlled by lipid scramblases and flip/floppases. However, for some intracellular membranes such as the endoplasmic reticulum (ER) the scramblases have not been identified. Members of the TMEM16 family have either lipid scramblase or chloride channel activity. Although TMEM16K is widely distributed and associated with the neurological disorder autosomal recessive spinocerebellar ataxia type 10 (SCAR10), its location in cells, function and structure are largely uncharacterised. Here we show that TMEM16K is an ER-resident lipid scramblase with a requirement for short chain lipids and calcium for robust activity. Crystal structures of TMEM16K show a scramblase fold, with an open lipid transporting groove. Additional cryo-EM structures reveal extensive conformational changes from the cytoplasmic to the ER side of the membrane, giving a state with a closed lipid permeation pathway. Molecular dynamics simulations showed that the open-groove conformation is necessary for scramblase activity.

Ca2+ -activated Cl À channels (CaCCs) are gated open by a rise in intracellular Ca 2+ concentration ([Ca 2+ ] i ), typically provoked by activation of G q -protein coupled receptors (G q PCR). G q PCR activation initiates depletion of plasmalemmal phosphatidylinositol 4,5-bisphosphate (PIP 2 ). Here, we determined whether PIP 2 acts as a signalling lipid for CaCCs coded by the TMEM16A and TMEM16B genes. EXPERIMENTAL APPROACHPatch-clamp electrophysiology, in conjunction with genetically encoded systems to control cellular PIP 2 content, was used to define the mechanism of action of PIP 2 on TMEM16A and TMEM16B channels. KEY RESULTSA water-soluble PIP 2 analogue (diC8-PIP 2 ) activated TMEM16A channels by up to fivefold and inhibited TMEM16B by~0.2-fold. The effects of diC8-PIP 2 on TMEM16A currents were especially pronounced at low [Ca 2+ ] i . In contrast, diC8-PIP 2 modulation of TMEM16B channels did not vary over a broad [Ca 2+ ] i range but was only detectable at highly depolarized membrane potentials. Modulation of TMEM16A and TMEM16B currents was due to changes in channel gating, while single channel conductance was unaltered. Co-expression of TMEM16A or TMEM16B with a Danio rerio voltage-sensitive phosphatase (DrVSP), which degrades PIP 2 , led to reduction and enhancement of TMEM16A and TMEM16B currents respectively. These effects were abolished by an inactivating mutation in DrVSP and antagonized by simultaneous co-expression of a phosphatidylinositol-4-phosphate 5-kinase that catalyses PIP 2 formation. CONCLUSIONS AND IMPLICATIONSPIP 2 acts as a modifier of TMEM16A and TMEM16B channel gating. Drugs interacting with PIP 2 signalling may affect TMEM16A and TMEM16B channel gating and have potential uses in basic science and implications for therapy. AbbreviationsCaCC, calcium-activated chloride channel; DrVSP, Danio rerio voltage-sensitive phosphatase; E rev , reversal potential; G q PCR, G q -protein coupled receptors; IP 3 , inositol triphosphate; PIP 2 , phosphatidylinositol 4,5-bisphosphate; PLC, phospholipase C; PIPK, PIP 5-kinase type Iγ; V m , membrane potential

Membranes in cells have defined distributions of lipids in each leaflet, controlled by lipid scramblases and flip/floppases. However, for some intracellular membranes such as the endoplasmic reticulum the scramblases have not been identified. Members of the TMEM16 family have either lipid scramblase and ion channel activity, or specific chloride channel activity. Although TMEM16K is widely distributed and associated with the neurological disorder autosomal recessive spinocerebellar ataxia type 10 (SCAR10), its location in cells, function and structure are largely uncharacterised. Here we show that TMEM16K is an ER-resident calcium-regulated lipid scramblase. Our crystal structures of TMEM16K show a scramblase fold, with an open lipid transporting groove. Additional structures solved by cryo-EM reveal extensive conformational changes extending from the cytoplasmic to the ER side of the membrane, giving a state with a closed lipid permeation pathway. Molecular dynamics simulations showed that the open-groove conformation is necessary for scramblase activity. Our results suggest mechanisms by which missense variants of TMEM16K could cause SCAR10 ataxia, providing new hypotheses to explore for therapy. 4 Cells and their organelles are enclosed by lipid bilayers and the lipid composition of either side of these membranes is controlled by active transporters (flippases and floppases) and passive scramblases, which equilibrate lipids between the membrane leaflets 1 . Many lipids are synthesized on the cytoplasmic side of the endoplasmic reticulum (ER) membrane which, unlike the plasma membrane (PM), has a symmetrical lipid distribution, suggesting a role for scramblases in the ER. To date, specific ER scramblases have not been identified and characterised. The ten members of the TMEM16 scramblase/channel family of integral membrane proteins show a surprising diversity of function, being either Ca 2+ -activated chloride channels (TMEM16A and B) 2-4 , or Ca 2+ -activated lipid scramblases with nonselective ion channel activity (TMEM16C, D, F, G and J) 5-8 . While some members of the family (A,B,F)reside in the plasma membrane, others, including TMEM16K 9 may function in intracellular membranes. TMEM16K is a widely distributed 10 , but relatively unstudied member of the TMEM16K family. Truncations and missense variants of TMEM16K (ANO10) are associated with the autosomal recessive spinocerebellar ataxia SCAR10 11,12 (as known as ARCA3 13-15 or ATX-ANO10 16 ). SCAR10 causing cerebellar ataxia, epilepsy and cognitive impairment with cerebellar atrophy noted on MRI brain and coenzyme Q10 deficiency found in muscle biopsy, fibroblasts and cerebrospinal fluid 11,12,17,18 . Some patients also have epilepsy and cognitive impairment 13,14 . Knockout studies in Drosophila 19 and mice 20 have suggested that loss of TMEM16K homologue function affects spindle formation 19 , Ca 2+ signalling 20 and apoptosis 19,20 .Structural studies have gone some way towards explaining how TMEM16 family members function as channels or lipid scrambla...

Maintenance of skeletal muscle is beneficial in obesity and Type 2 diabetes. Mechanical stimulation can regulate skeletal muscle differentiation, growth and metabolism, however the molecular mechanosensor remains unknown. Here, we show that SWELL1 (Lrrc8a) functionally encodes a swell-activated anion channel that regulates PI3K-AKT, ERK1/2, mTOR signaling, muscle differentiation, myoblast fusion, cellular oxygen consumption, and glycolysis in skeletal muscle cells. LRRC8A over-expression in Lrrc8a KO myotubes boosts PI3K-AKT-mTOR signaling to supra-normal levels and fully rescues myotube formation. Skeletal muscle targeted Lrrc8a KO mice have smaller myofibers, generate less force ex vivo, and exhibit reduced exercise endurance, associated with increased adiposity under basal conditions, and glucose intolerance and insulin resistance when raised on a high-fat diet, compared to WT mice. These results reveal that the LRRC8 complex regulates insulin-PI3K-AKT-mTOR signalling in skeletal muscle to influence skeletal muscle differentiation in vitro and skeletal myofiber size, muscle function, adiposity and systemic metabolism in vivo.