Structural Basis for the Distinct Membrane Binding Activity of the Homologous C2A Domains of Myoferlin and Dysferlin

Identification of myoferlin as a mitochondria-associated membranes component required for calcium signaling in PDAC cell lines

BACKGROUND

Pancreatic ductal adenocarcinoma is an aggressive cancer type with one of the lowest survival rates due to late diagnosis and the absence of effective treatments. A better understanding of PDAC biology will help researchers to discover the Achilles' heel of cancer cells. In that regard, our research team investigated the function of an emerging oncoprotein known as myoferlin. Myoferlin is overexpressed in PDAC and its silencing/targeting has been shown to affect cancer cell proliferation, migration, mitochondrial dynamics and metabolism. Nevertheless, our comprehension of myoferlin functions in cells remains limited. In this study, we aimed to understand the molecular mechanism linking myoferlin silencing to mitochondrial dynamics.

METHODS

Experiments were performed on two pancreas cancer cell lines, Panc-1 and MiaPaCa-2. Myoferlin localization on mitochondria was evaluated by immunofluorescence, proximity ligation assay, and cell fractionation. The presence of myoferlin in mitochondria-associated membranes was assessed by cell fractionation and its function in mitochondrial calcium transfer was evaluated using calcium flow experiments, proximity ligation assays, co-immunoprecipitation, and timelapse fluorescence microscopy in living cells.

RESULTS

Myoferlin localization on mitochondria was investigated. Our results suggest that myoferlin is unlikely to be located on mitochondria. Instead, we identified myoferlin as a new component of mitochondria-associated membranes. Its silencing significantly reduces the mitochondrial calcium level upon stimulation, probably through myoferlin interaction with the inositol 1,4,5-triphosphate receptors 3.

CONCLUSIONS

For the first time, myoferlin was specifically demonstrated to be located in mitochondria-associated membranes where it participates to calcium flow. We hypothesized that this function explains our previous results on mitochondrial dynamics. This study improves our comprehension of myoferlin localization and function in cancer biology.

The Dysferlin C2A Domain Binds PI(4,5)P2 and Penetrates Membranes

Dysferlin is a large membrane protein found most prominently in striated muscle. Loss of dysferlin activity is associated with reduced exocytosis, abnormal intracellular Ca2+ and the muscle diseases limb-girdle muscular dystrophy and Miyoshi myopathy. The cytosolic region of dysferlin consists of seven C2 domains with mutations in the C2A domain at the N-terminus resulting in pathology. Despite the importance of Ca2+ and membrane binding activities of the C2A domain for dysferlin function, the mechanism of the domain remains poorly characterized. In this study we find that the C2A domain preferentially binds membranes containing PI(4,5)P2 through an interaction mediated by residues Y23, K32, K33, and R77 on the concave face of the domain. We also found that subsequent to membrane binding, the C2A domain inserts residues on the Ca2+ binding loops into the membrane. Analysis of solution NMR measurements indicate that the domain inhabits two distinct structural states, with Ca2+ shifting the population between states towards a more rigid structure with greater affinity for PI(4,5)P2. Based on our results, we propose a mechanism where Ca2+ converts C2A from a structurally dynamic, low PI(4,5)P2 affinity state to a high affinity state that targets dysferlin to PI(4,5)P2 enriched membranes through interaction with Tyr23, K32, K33, and R77. Binding also involves changes in lipid packing and insertion by the third Ca2+ binding loop of the C2 domain into the membrane, which would contribute to dysferlin function in exocytosis and Ca2+ regulation.

Analysis of Dysferlin Direct Interactions with Putative Repair Proteins Links Apoptotic Signaling to Ca2+ Elevation via PDCD6 and FKBP8

Quantitative surface plasmon resonance (SPR) was utilized to determine binding strength and calcium dependence of direct interactions between dysferlin and proteins likely to mediate skeletal muscle repair, interrupted in limb girdle muscular dystrophy type 2B/R2. Dysferlin canonical C2A (cC2A) and C2F/G domains directly interacted with annexin A1, calpain-3, caveolin-3, affixin, AHNAK1, syntaxin-4, and mitsugumin-53, with cC2A the primary target and C2F lesser involved, overall demonstrating positive calcium dependence. Dysferlin C2 pairings alone showed negative calcium dependence in almost all cases. Like otoferlin, dysferlin directly interacted via its carboxy terminus with FKBP8, an anti-apoptotic outer mitochondrial membrane protein, and via its C2DE domain with apoptosis-linked gene (ALG-2/PDCD6), linking anti-apoptosis with apoptosis. Confocal Z-stack immunofluorescence confirmed co-compartmentalization of PDCD6 and FKBP8 at the sarcolemmal membrane. Our evidence supports the hypothesis that prior to injury, dysferlin C2 domains self-interact and give rise to a folded, compact structure as indicated for otoferlin. With elevation of intracellular Ca2+ in injury, dysferlin would unfold and expose the cC2A domain for interaction with annexin A1, calpain-3, mitsugumin 53, affixin, and caveolin-3, and dysferlin would realign from its interactions with PDCD6 at basal calcium levels to interact strongly with FKBP8, an intramolecular rearrangement facilitating membrane repair.

Orientation of the Dysferlin C2A Domain is Responsive to the Composition of Lipid Membranes

Dysferlin is a 230 kD protein that plays a critical function in the active resealing of micron-sized injuries to the muscle sarcolemma by recruiting vesicles to patch the injured site via vesicle fusion. Muscular dystrophy is observed in humans when mutations disrupt this repair process or dysferlin is absent. While lipid binding by dysferlin's C2A domain (dysC2A) is considered fundamental to the membrane resealing process, the molecular mechanism of this interaction is not fully understood. By applying nonlinear surface-specific vibrational spectroscopy, we have successfully demonstrated that dysferlin's N-terminal C2A domain (dysC2A) alters its binding orientation in response to a membrane's lipid composition. These experiments reveal that dysC2A utilizes a generic electrostatic binding interaction to bind to most anionic lipid surfaces, inserting its calcium binding loops into the lipid surface while orienting its β-sheets 30-40° from surface normal. However, at lipid surfaces, where PI(4,5)P2 is present, dysC2A tilts its β-sheets more than 60° from surface normal to expose a polybasic face, while it binds to the PI(4,5)P2 surface. Both lipid binding mechanisms are shown to occur alongside dysC2A-induced lipid clustering. These different binding mechanisms suggest that dysC2A could provide a molecular cue to the larger dysferlin protein as to signal whether it is bound to the sarcolemma or another lipid surface.

Redefining the architecture of ferlin proteins: Insights into multi-domain protein structure and function

Ferlins are complex, multi-domain proteins, involved in membrane trafficking, membrane repair, and exocytosis. The large size of ferlin proteins and the lack of consensus regarding domain boundaries have slowed progress in understanding molecular-level details of ferlin protein structure and function. However, in silico protein folding techniques have significantly enhanced our understanding of the complex ferlin family domain structure. We used RoseTTAFold to assemble full-length models for the six human ferlin proteins (dysferlin, myoferlin, otoferlin, Fer1L4, Fer1L5, and Fer1L6). Our full-length ferlin models were used to obtain objective domain boundaries, and these boundaries were supported by AlphaFold2 predictions. Despite the differences in amino acid sequence between the ferlin proteins, the domain ranges and distinct subdomains in the ferlin domains are remarkably consistent. Further, the RoseTTAFold/AlphaFold2 in silico boundary predictions allowed us to describe and characterize a previously unknown C2 domain, ubiquitous in all human ferlins, which we refer to as C2-FerA. At present, the ferlin domain-domain interactions implied by the full-length in silico models are predicted to have a low accuracy; however, the use of RoseTTAFold and AlphaFold2 as a domain finder has proven to be a powerful research tool for understanding ferlin structure.

The Dystrophin Node as Integrator of Cytoskeletal Organization, Lateral Force Transmission, Fiber Stability and Cellular Signaling in Skeletal Muscle

The systematic bioanalytical characterization of the protein product of the DMD gene, which is defective in the pediatric disorder Duchenne muscular dystrophy, led to the discovery of the membrane cytoskeletal protein dystrophin. Its full-length muscle isoform Dp427-M is tightly linked to a sarcolemma-associated complex consisting of dystroglycans, sarcoglyans, sarcospan, dystrobrevins and syntrophins. Besides these core members of the dystrophin-glycoprotein complex, the wider dystrophin-associated network includes key proteins belonging to the intracellular cytoskeleton and microtubular assembly, the basal lamina and extracellular matrix, various plasma membrane proteins and cytosolic components. Here, we review the central role of the dystrophin complex as a master node in muscle fibers that integrates cytoskeletal organization and cellular signaling at the muscle periphery, as well as providing sarcolemmal stabilization and contractile force transmission to the extracellular region. The combination of optimized tissue extraction, subcellular fractionation, advanced protein co-purification strategies, immunoprecipitation, liquid chromatography and two-dimensional gel electrophoresis with modern mass spectrometry-based proteomics has confirmed the composition of the core dystrophin complex at the sarcolemma membrane. Importantly, these biochemical and mass spectrometric surveys have identified additional members of the wider dystrophin network including biglycan, cavin, synemin, desmoglein, tubulin, plakoglobin, cytokeratin and a variety of signaling proteins and ion channels.

Calcium binds and rigidifies the dysferlin C2A domain in a tightly coupled manner

The membrane protein dysferlin (DYSF) is important for calcium-activated plasma membrane repair, especially in muscle fibre cells. Nearly 600 mutations in the DYSF gene have been identified that are causative for rare genetic forms of muscular dystrophy. The dysferlin protein consists of seven C2 domains (C2A–C2G, 13%–33% identity) used to recruit calcium ions and traffic accessory proteins and vesicles to injured membrane sites needed to reseal a wound. Amongst these, the C2A is the most prominent facilitating the calcium-sensitive interaction with membrane surfaces. In this work, we determined the calcium-free and calcium-bound structures of the dysferlin C2A domain using NMR spectroscopy and X-ray crystallography. We show that binding two calcium ions to this domain reduces the flexibility of the Ca2+-binding loops in the structure. Furthermore, calcium titration and mutagenesis experiments reveal the tight coupling of these calcium-binding sites whereby the elimination of one site abolishes calcium binding to its partner site. We propose that the electrostatic potential distributed by the flexible, negatively charged calcium-binding loops in the dysferlin C2A domain control first contact with calcium that promotes subsequent binding. Based on these results, we hypothesize that dysferlin uses a ‘calcium-catching’ mechanism to respond to calcium influx during membrane repair.

The Many Faces of DFNB9: Relating OTOF Variants to Hearing Impairment

The OTOF gene encodes otoferlin, a critical protein at the synapse of auditory sensory cells, the inner hair cells (IHCs). In the absence of otoferlin, signal transmission of IHCs fails due to impaired release of synaptic vesicles at the IHC synapse. Biallelic pathogenic and likely pathogenic variants in OTOF predominantly cause autosomal recessive profound prelingual deafness, DFNB9. Due to the isolated defect of synaptic transmission and initially preserved otoacoustic emissions (OAEs), the clinical characteristics have been termed "auditory synaptopathy". We review the broad phenotypic spectrum reported in patients with variants in OTOF that includes milder hearing loss, as well as progressive and temperature-sensitive hearing loss. We highlight several challenges that must be addressed for rapid clinical and genetic diagnosis. Importantly, we call for changes in newborn hearing screening protocols, since OAE tests fail to diagnose deafness in this case. Continued research appears to be needed to complete otoferlin isoform expression characterization to enhance genetic diagnostics. This timely review is meant to sensitize the field to clinical characteristics of DFNB9 and current limitations in preparation for clinical trials for OTOF gene therapies that are projected to start in 2021.

Functions of Vertebrate Ferlins

Ferlins are multiple-C2-domain proteins involved in Ca²⁺-triggered membrane dynamics within the secretory, endocytic and lysosomal pathways. In bony vertebrates there are six ferlin genes encoding, in humans, dysferlin, otoferlin, myoferlin, Fer1L5 and 6 and the long noncoding RNA Fer1L4. Mutations in DYSF (dysferlin) can cause a range of muscle diseases with various clinical manifestations collectively known as dysferlinopathies, including limb-girdle muscular dystrophy type 2B (LGMD2B) and Miyoshi myopathy. A mutation in MYOF (myoferlin) was linked to a muscular dystrophy accompanied by cardiomyopathy. Mutations in OTOF (otoferlin) can be the cause of nonsyndromic deafness DFNB9. Dysregulated expression of any human ferlin may be associated with development of cancer. This review provides a detailed description of functions of the vertebrate ferlins with a focus on muscle ferlins and discusses the mechanisms leading to disease development.

Myoferlin, a Membrane Protein with Emerging Oncogenic Roles

Myoferlin (MYOF), initially identified in muscle cells, is a member of the Ferlin family involved in membrane fusion, membrane repair, and membrane trafficking. Dysfunction of this protein is associated with muscular dysfunction. Recently, a growing body of studies have identified MYOF as an oncogenic protein. It is overexpressed in a variety of human cancers and promotes tumorigenesis, tumor cell motility, proliferation, migration, epithelial to mesenchymal transition, angiogenesis as well as metastasis. Clinically, MYOF overexpression is associated with poor outcome in various cancers. It can serve as a prognostic marker of human malignant disease. MYOF drives the progression of cancer in various processes, including surface receptor transportation, endocytosis, exocytosis, intercellular communication, fit mitochondrial structure maintenance and cell metabolism. Depletion of MYOF demonstrates significant antitumor effects both in vitro and in vivo, suggesting that targeting MYOF may produce promising clinical benefits in the treatment of malignant disease. In the present article, we reviewed the physiological function of MYOF as well as its role in cancer, thus providing a general understanding for further exploration of this protein.

Myoferlin, a multifunctional protein in normal cells, has novel and key roles in various cancers

Myoferlin, a protein of the ferlin family, has seven C2 domains and exhibits activity in some cells, including myoblasts and endothelial cells. Recently, myoferlin was identified as a promising target and biomarker in non-small-cell lung cancer, breast cancer, pancreatic adenocarcinoma, hepatocellular carcinoma, colon cancer, melanoma, oropharyngeal squamous cell carcinoma, head and neck squamous cell carcinoma, clear cell renal cell carcinoma and endometrioid carcinoma. This evidence indicated that myoferlin was involved in the proliferation, invasion and migration of tumour cells, the mechanism of which mainly included promoting angiogenesis, vasculogenic mimicry, energy metabolism reprogramming, epithelial-mesenchymal transition and modulating exosomes. The roles of myoferlin in both normal cells and cancer cells are of great significance to provide novel and efficient methods of tumour treatment. In this review, we summarize recent studies and findings of myoferlin and suggest that myoferlin is a novel potential candidate for clinical diagnosis and targeted cancer therapy.

Ferlin Overview: From Membrane to Cancer Biology

In mammal myocytes, endothelial cells and inner ear cells, ferlins are proteins involved in membrane processes such as fusion, recycling, endo- and exocytosis. They harbour several C2 domains allowing their interaction with phospholipids. The expression of several Ferlin genes was described as altered in several tumoural tissues. Intriguingly, beyond a simple alteration, myoferlin, otoferlin and Fer1L4 expressions were negatively correlated with patient survival in some cancer types. Therefore, it can be assumed that membrane biology is of extreme importance for cell survival and signalling, making Ferlin proteins core machinery indispensable for cancer cell adaptation to hostile environments. The evidences suggest that myoferlin, when overexpressed, enhances cancer cell proliferation, migration and metabolism by affecting various aspects of membrane biology. Targeting myoferlin using pharmacological compounds, gene transfer technology, or interfering RNA is now considered as an emerging therapeutic strategy.