Vincent Mercier, Jorge Larios, Guillaume Molinard, Antoine Goujon, Stefan Matile, Jean Gruenberg, Aurelien Roux

Preprint posted on February 21, 2019

Feeling tense? Hop aboard! Endosomal ESCRT-III recruitment and ILV formation is regulated by endosomal membrane tension

Selected by Nicola Stevenson


During endosomal protein sorting, five large multisubunit protein assemblies, collectively known as the ESCRT machinery (Endosomal Sorting Complexes Required for Transport), assemble in a sequential manner on the endosome surface to generate intraluminal vesicles – vesicles that bud into the lumen of the endosome to generate multivesicular bodies. The fourth of these assemblies, ESCRT-III, is largely cytoplasmic and only transiently assembles on endosomes during sorting. Once recruited here it deubiquitinates and sequesters cargo in the vesicle and reorganises the membrane for budding and scission (1). Previous reports have shown that ESCRT-III depends on other members of the ESCRT machinery, such as VPS25 of ESCRT-II (2), to localise to membranes. However, in this new study, Mercier et al report that ESCRT-III is rapidly recruited to the endosome following a decrease in membrane tension, such as occurs during stimulated recycling of the EGF receptor. This provides a fascinating new regulatory link between the physical properties of organelles and the recruitment of large protein machines.  Such a link may allow fluctuations in the extracellular environment to propagate changes in cell behaviour independent of (or augmenting) specific signalling molecules.

Key findings

In this study the authors use two methods to induce changes in endosomal membrane tension inside the cell; 1) hypertonic shock with sucrose or NaCl and 2) transient permeabilization of the endosome membrane using the small peptide LLOMe. Changes in membrane tension were confirmed with the fluorescent lipid tension reporter, Lyso Flipper. In both cases, a reduction in membrane tension induced the transient relocation of the ESCRT-III component CHMP4B from a diffuse cytosolic localisation to EEA1 positive endosomes. This endosomal recruitment is selective for ESCRT-III, since ESCRT-0 and ESCRT-1 were not recruited to membranes upon hypertonic shock and STAM and TSG101 only partially so on LLOMe treatment. Recruitment is, however, in part dependent on the ESCRT-III nucleator ALIX and recruited assemblies are functional.

These results were confirmed using a third method in which recombinant CHMP4B was incubated with giant unilamellar vesicles. In isotonic conditions CHMP4B recruited slowly to these membranes, however switching to hypertonic solution increased CHMP4B recruitment three-fold. Changing membrane tension by pulling on the membranes with optical tweezers also demonstrated a negative correlation between increasing membrane tension and CHMP4B recruitment. Electron microscopy of membranes coated with CHMP4B were tubular and deformed indicating CHMP4B recruitment induced membrane remodelling.

Finally, to show membrane tension regulated recruitment of ESCRT-III is physiologically relevant, the authors treated cells with epidermal growth factor (EGF) to stimulate recycling of its receptor through endosomes. Treatment increased ESCRT dependent ILV formation concurrent with a reduction in endosome membrane tension as detected with Lyso Flipper. Such a reduction in membrane tension is likely brought about by the increase in membrane surface area as endosomal vesicles carrying the EGF receptor fuse with mature endosomes.

In conclusion, CHMP4B polymerises on endosomal membranes when membrane tension is lowered.


With the advent of atomic force microscopy and optical tweezers, the study of membrane tension and cell behaviour has become a more popular and attainable pursuit in recent years. Understandably, research in this area has been predominantly, but not exclusively, focused on the plasma membrane due to its accessibility and amenability to such techniques. However, the recent development of fluorescent membrane tension sensors by some of the authors of this study (3) has provided the opportunity to expand these investigations to the study of internal membranes. These should help to elucidate the full extent to which mechanosensation is important to cellular function. I chose to highlight this paper as it is a good example of the potential of this avenue of study. Current models of membrane trafficking events generally focus on chemical interactions between proteins, lipids and small molecules but this study demonstrates that this is only part of the picture. Clearly we need to look beyond biochemistry to biophysics to really understand this process.

Future questions

  • Why is ESCRT-III but not upstream components of ESCRT recruited in this way? Changes in membrane tension are likely to occur very rapidly following stimulated endocytosis and so one may imagine this would be relevant to the initiation of ESCRT-mediated sorting rather than later steps. Is ESCRT-III playing another role on the endosomes i.e. reducing membrane surface in a compensatory manner?
  • What are the key features of the CHMP4B or its recruitment factors that enable membrane tension sensing?
  • Is membrane tension a global mechanism for sensing and responding to changes in membrane flux between organelles or is this specific to ILV formation. Some external stimuli can increase anterograde secretion, for example synaptic transmission can stimulate the ER export of de novo synaptic receptors to the cell surface. Does this similarly result in changes in membrane tension and recruitment of proteins for membrane remodelling?


  1. Raiborg, C., Stenmark, H. The ESCRT machinery in endosomal sorting of ubiquitylated membrane proteins. Nature 458, 445-452 (2009)
  2. Teo, H., Perisic, O., Gonzalez, B. & Williams, R. L. ESCRT-II, an endosome-associated complex required for protein sorting: crystal structure and interactions with ESCRT-III and membranes.  Cell7, 559–569 (2004)
  3. Goujon, A., Colom, A., Straková, K., Mercier, V., Mahecic, D., Manley, S., Sakai, N., Roux, A., & Matile, S. Mechanosensitive Fluorescent Probes to Image Membrane Tension in Mitochondria, Endoplasmic Reticulum, and Lysosomes. Am. Chem. Soc (2019)


Posted on: 26th February 2019 , updated on: 27th February 2019

Read preprint (1 votes)

  • Author's response

    Vincent Mercier and Aurelien Roux shared


    1. Upstream components of ESCRT have cargo sorting and ESCRT-III nucleating functions. We think that tension only modulates nucleation and polymerization of ESCRT-III, which is coupled to membrane deformation (see Chiaruttini et al. Cell 2015), without affecting cargo sorting functions, presumably because the latter process is independent of membrane tension.

    2. As described in Chiaruttini et al. Cell 2015, CHMP4B/Snf7 polymerizes into a spiral spring that can buckle the membrane. Thus, deformation of the membrane is coupled to polymer elongation. If tension is high, the membrane cannot be deformed easily, and probably hinders ESCRT-III polymerization. Interestingly, we have measured that the polymerization energy of Snf7 is 4.10-4 N.m-1 (Chiaruttini et al. Cell) and that the threshold tension above which the polymerization rate of CHM4B is significantly reduced is of the same order (this preprint). This strongly supports that tension directly affects ESCRT-III polymerization.

    3. Indeed, you are right, and this mechanism of membrane traffic regulation by fluxes of membrane tension has initially been proposed by Michael Sheetz. If an incoming flux of vesicles reaches an organelle, this will bring down its tension because the membrane area expands, facilitating formation of new membrane carriers from this organelle. Clathrin coats, but also COPI coats have been shown to be tension-sensitive.

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here