Membrane Curvature Promotes ER-PM Contact Formation via Junctophilin-EHD Interactions

Background:

Cardiomyocytes receive signals to contract in the form of action potentials at long invaginations of the plasma membrane (PM) called transverse tubules (t-tubules). T-tubules form extensive membrane contact sites with the endoplasmic reticulum (ER) called dyad junctions (see Figure 1A). Dyad junctions are important for passing the signal on to sarcomeres, the contractile units of cardiomyocytes: When the PM depolarises upon action potential arrival, voltage-gated L-type calcium channels (LTCCs) in the t-tubules open and calcium ions enter the cytosolic gap between the t-tubule and the ER.  This influx of calcium in turn leads to the opening of ryanodine receptor (RyR) channels in the ER, which release even higher quantities of calcium into the cytoplasm. Calcium then binds to the protein troponinin in the sarcomere and enables myosin to interact with actin filaments, leading to muscle contraction. The co-localisation of the PM-localised LTCC channels and the ER-localised RyR channels is clearly important for efficient calcium signalling and muscle contraction. The proteins responsible for co-localisation of the channels at dyad junctions are called junctophilins (JPHs). However, how junctophilins themselves are sorted to the dyad junction has until now been unknown.

In this preprint, Yang and colleagues decipher how junctophilins localise to curved plasma membrane regions. They show that Epsin15-homology domain containing proteins (EHDs) are required for junctophilin recruitment to curved membranes and identify the junctophilin domains required for EHD binding.

 

Method:

Different cell types, including cardiomyocytes that were either induced pluripotent stem cell differentiated (iPSC) or taken from rat embryos and the osteosarcoma cell line U2OS, were cultured on nanochips dotted with a regular array of nanopillar or nanobar structures. The cells wrap around these structures leading to curvature of the plasma membrane. Nanopillars with a diameter of 200-300 nm were used to resemble t-tubules. In the case of nanobars, a width of 200 nm and a length of 2 µm was used, leading to areas of curved and straight plasma membrane. Using fluorescent microscopy, the localisation of proteins could be investigated with regard to the nanopillar vs the flat surface in between, and with regard to the curved ends vs the flat sides of the nanobar (see Figure 1).

 

 

Key findings:

1. Components of dyad junctions localise to curved membranes in cells grown on nanopillar or nanobar grids

• • • The authors first checked the distribution of immuno-stained junctophilin JPH2 in iPSC-cardiomyocytes, which lack t-tubules, grown on the curvature-inducing surface. They observed that JPH2 accumulated at nanopillars or the ends of a nanobar, but the ER-resident protein translocon Sec61 was dispersed.
    • Other components of the dyad junction, such as LTCC (calcium channel in the PM) and RyR2 (calcium channel in the ER) also localised to the highly-curved membrane regions while the PM marker BFP-CAAX was present around the whole nanopillar.
    • E-Syt2 (another ER-PM tether), on the other hand, localised to the side walls (=flat membrane) of nanobars, showing that curvature preference is specific to junctophilin-mediated ER-PM contact sites.
    • Focused ion beam scanning electron microscopy revealed that ER-PM contacts preferentially form at nanopillars in cardiomyocytes. In WT U2OS cells, the ER-PM contact site marker MAPPER localised to nanobar ends. Both findings suggest that ER-PM contacts preferentially form at curved PM.

STIM1 and ORAI1 are recruited to ER-PM contact sites at curved regions of the PM upon ER calcium store depletion

• • • Before calcium store depletion, STIM1 was found in the intracellular ER network and ORAI1 was diffused throughout the PM.
    • After thapsigargin-induced ER calcium store depletion, STIM1 and ORAI1 co-localised at nanobar ends, suggesting that the proteins are preferentially recruited to ER-PM contact sites at curved PM.

The low complexity region (LCR) and the MORN repeats domain of JPH3 are both required for targeting to curved PM

• • • The MORN domain as well as LCR could bind to the PM independently of each other – deletion of just one did not hinder PM binding and when expressed in isolation, each localised to the PM. However, all of these contructs lost preference for curvature, indicating that both, the MORN domain and the LCR, are required for binding to curved PM regions.
    • The authors furthermore identified a polybasic motif within the LCR, which is required for PM binding, and showed that mutations of serine residues in the MORN/LCR region led to loss of curvature localisation. This is very interesting from a clinical point of view as these mutations are found in patients with hypertrophic cardiomyopathy.

JPHs bind to EHD proteins which are known to be important for t-tubule morphology

• • • Out of all JPH2-interacting proteins that are involved in curvature-dependent processes at the PM, only triple knock-down of EHD1, 2 and 4 disrupted JPH3’s localisation to nanobar ends.
    • Extraction of cholesterol changes the localisation of EHDs. Here the authors could show that this treatment disrupts not only EHD4’s localisation to nanobar ends but also that of JPH3.
    • Both, the MORN repeats and the LCR region were needed for binding to EHD as shown by co-localisation of MORN-LCR constructs with EHD4 in cells grown on a flat surface. Furthermore, only a construct containing both domains could be co-immunoprecipitated with EHD4.

 

What I like about the preprint:

This preprint initially caught my interest because it studies a membrane contact site that preferentially forms at a curved membrane. This architectural arrangement strikingly resembles the membrane contact site between the inner and outer mitochondrial membrane, which I investigate in my PhD research. I also really like the method of artificially inducing curvature in live cells with nanostructures. It is something I have not come across before and it has yielded really fascinating fluorescent images in this study.

 

Questions for the authors:

1. As far as I understood, JPH proteins are predominantly expressed in muscles and neurons. Why then are JPH3 and 4 expressed in osteosarcoma U2OS cells? Could they, together with EHDs, facilitate ER-PM membrane contact sites with other functions? Are there any known curved PM regions in U2OS cells that form ER contacts or may your findings lead to the discovery of new processes requiring ER-PM membrane contact sites at curved PM?
2. Can LCR with mutated polybasic region still bind to EHDs? Can pathogenic mutant variants still bind to EHDs?
3. Could you quantify the extent of ER-PM contact sites at nanopillars in WT cells vs cells with disrupted EHD-Junctophilin binding (or cells with cholesterol extracted EHD / EHD or junctophilin KD)? It would be interesting to know whether the junctophilin tether is required to localise PM-ER contact sites to the nanopillar or whether there are other tethers with curvature preference. In connection to this, does loss of EHD-junctophilin interaction lead to loss of LTCC and RyR2 colocalisation at nanopillars?

 

 

Bibliography:

Bhattacharyya, S. and T. J. Pucadyil (2020). “Cellular functions and intrinsic attributes of the ATP-binding Eps15 homology domain-containing proteins.” Protein Sci 29(6): 1321-1330. DOI: 10.1002/pro.3860

Lu, F. and W. T. Pu (2020). “The architecture and function of cardiac dyads.” Biophys Rev 12(4): 1007-1017. DOI: 10.1007/s12551-020-00729-x

Piggott, C. A. and Y. Jin (2021). “Junctophilins: Key Membrane Tethers in Muscles and Neurons.” Front Mol Neurosci 14: 709390. DOI: 10.3389/fnmol.2021.709390. DOI: 10.3389/fnmol.2021.709390

 

doi: https://doi.org/10.1242/prelights.38028

Read preprint (No Ratings Yet)