Uncovering the “secret” lives of vacuolar fusion pores in living cells

Abstract

Non‐expanding fusion pores have been predicted to constitute rate‐limiting metastable intermediates in a broad range of fusion processes ranging from exocytosis to cell–cell fusion. However, their existence and nature, in particular in intracellular non‐exocytic compartments, remained largely intangible. In this issue, D'Agostino et al demonstrate that non‐expanding pores are likely long‐lived steady‐state intermediates in yeast vacuoles allowing the cells to adjust their volume to rapidly changing physiology. These pores are stabilized against closure by SNAREs and the attached vacuolar vesicle tethering complex HOPS. Their expansion is controlled by increasing membrane tension.

Membrane fusion is an ubiquitary event occurring along the secretory and endocytic pathway, shuttling cargo between compartments and governing a variety of physiological events, such as hormone secretion/neurotransmitter release, or the digestion of intra‐ (autophagy) and extracellular (phagocytosis) substances in lysosomes/vacuoles. Following several decades of intensive research, membrane‐tethering factors (Rab and Rab‐effectors), proteins priming and triggering fusion (Munc13 and synaptotagmin family members), and the core components of the fusion machinery (SNARE and SM proteins) have been identified which choreograph a cascade of reactions culminating in lipid bilayer merger (Wickner & Rizo, 2017). However, the final steps—fusion pore formation and expansion—still remain largely elusive. Most progress has been made studying regulated exocytosis at the plasma membrane demonstrating reversible pore flickering (kiss and run/stay) and more recently dynamic fusion pores (Alabi & Tsien, 2013; Shin et al, 2018). Much less is known about fusion pores in intracellular compartments. A general assumption is that such pores quickly dilate after the completely zippered SNAREs have overcome the energetic barrier to initiate membrane fusion. In this issue of The EMBO Journal, D'Agostino et al (2018) provide astonishing results about the size and the dynamics/lifetime of the vacuolar fusion pore in living yeast cells.

Yeast vacuoles have an almost spherical appearance, vary in their number depending on the physiological state of the cell, and are usually tethered to each other. To adapt to alterations in environmental conditions, e.g., fast‐changing osmotic pressure, vacuoles undergo fragmentation/fission (hypertonic medium) or fusion (hypotonic medium), thereby avoiding potentially catastrophic rupture and vacuolar content release into the cytosol (Desfougeres et al, 2016). Due to their relatively large diameter, vacuoles can be tracked in living cells by conventional fluorescence microscopy, allowing the detection of content mixing and lipid mixing and the presence of large expanding fusion pores. In an elegant approach, D'Agostino et al make use of this experimental system, employing fluorescence dyes of different sizes and nature, fluorescence resonance energy transfer (FRET), and fluorescence recovery after photobleaching (FRAP) to survey the nature of vacuolar fusion pores (Fig 1A).

Figure 1. Vacuolar fusion pore intermediates.

(A) Vacuolar fusion pore expansion is driven by an increase in membrane tension. (B) Molecular dynamics simulations illustrating potential fusion pore intermediates. Please note for clarity the HOPS complex was not depicted.

Large expanding fusion pores were detected at 1% of the tethered vacuoles as membrane contact areas devoid of any fluorescence when analyzing the distribution of a lipophilic dye targeted to the lumenal leaflet and of a membrane‐anchored fluorescent protein facing the cytoplasmic side. The detection of both markers in the remaining contact area demonstrates the presence of two intact lipid bilayers and excludes the existence of extensive hemifusion diaphragms. Small fusion pores and potential hemifusion intermediates, below the resolution limits of light microscopy, were identified by FRAP experiments of contacting vacuoles. Remarkably, vacuolar contact areas permit the exchange of markers labeling the outer and inner leaflet lipid, indicating the presence of potentially long‐lived nanoscopic fusion pores. These small pores provide a remarkable size cutoff for lumenal content: Neither soluble fluorescence markers in the molecular weight range of 300–500 Da nor transmembrane proteins containing lumenal extensions of only six hydrophilic amino acids can pass the pore. Not surprisingly, pore formation depends on the vacuolar tethering/SNARE machinery. These pores expand in the presence of osmotic pressure changes adopting the vacuolar volume/surface ratio to environmental conditions. After triggering vacuolar fission events by hypertonic media, new nanoscopic fusion pores are established, indicating that these pores represent the default state to prepare cells for future osmotic challenges.

Moreover, coarse‐grain molecular dynamics simulations by D'Agostino et al suggest that 1.5‐nm fusion pores can form spontaneously in model membranes. However, their expansion to 3 nm requires additional energy (8 mN/m; Fig 1B). Further expansion can then occur spontaneously but may be restricted or slowed down by components tethering the attached vacuoles. The predominant localization of the expanding fusion pore in the vertex area of the vacuolar contact zone may be energetically less costly as pointed out by the authors. Concerning the reverse pathway, “rigid” fully assembled cis‐SNARE complexes and the attached HOPS complex will likely oppose fusion pore closure/hemifusion stalk formation. Interestingly, the experimentally observed “impermeability” is at odds with the modeled 1.5‐nm pore. However, models introducing lipid asymmetry and increased amounts of phosphatidylethanolamine, which affect membrane curvature, can accommodate fusion pores in the range of 0.5 nm.

Overall, these remarkable findings have important implications:

1. Compared to pores involved in exocytosis at the plasma membrane, which have a diameter of ≈ 2 nm, the observed permeability barrier of vacuolar pores suggests a diameter in the subnanometer range (Spruce et al, 1990). These differences in pore size may result from unique lipid/protein compositions at different transport steps. This also includes attractive forces within the pore interior or adhesive forces in the contact area of opposing membranes. Advanced imaging techniques, such as cryoelectron microscopy (cryo‐EM) and focused ion‐beam milling combined with scanning electron microscopy (FIB‐SEM), will be methods of choice to provide further structural information about the actual pore diameter and the local organization of the fusion machinery.

2. The requirement for increased membrane tension to open the fusion pore is in line with several theoretical calculations and recent experimental evidence analyzing exocytosis events at the plasma membrane (Kozlov & Chernomordik, 2015; Wang & Galli, 2018). Thus, membrane tension becomes a prominent biophysical and physiological regulator. In other words, despite the absence of the classical machinery triggering regulated exocytosis, “constitutive” fusion at the vacuole shows an unanticipated level of regulation. Actually, the sophisticated Ca²⁺‐triggered transmitter release optimized for synchronization speed at the neuronal synapse may employ the same biophysical membrane tension principle for fusion pore regulation. In the presence of Ca²⁺, the calcium sensor synaptotagmin binds anionic lipids, deforms the membrane, and regulates fusion pore expansion (Bendahmane et al, 2018). Considering that synaptotagmins can also form rings, the Ca²⁺‐triggered membrane deformation could provide a temporally synchronized and spatially concealed area of altered membrane tension regulating fusion pore expansion (Rothman et al, 2017).

Thus, the manuscript by D'Agostino et al studying vacuolar fusion and recent publications analyzing exocytosis at the plasma membrane, in combination with advanced imaging techniques improving the temporal and spatially resolution, have now set the stage to reveal the “secret” lives and structural composition of fusion pores at different transport steps.

The EMBO Journal (2018) 37: e100656