Leonid V. Chernomordik, PhD, Head, Section on Membrane Biology
Evgenia Leikina, DVM, Senior Research Assistant
Kamran Melikov, PhD, Staff Scientist
Elena Zaitseva, PhD, Research Fellow
Jean-Philippe Richard, PhD, Visiting Fellow
Andrew Chen, BS, Postbaccalaureate Fellow

Membrane fusion is a common stage of cell fusion in development, enveloped virus infections, and exocytosis. Recently, we extended our work on the fusion mechanisms from a relatively well-characterized fusion mediated by influenza virus hemagglutinin (HA) to poorly understood developmental fusion. In our work on HA-mediated fusion, we explored fusion between protein-free liposomes and HA-expressing cells and viral particles. We identified distinct fusion intermediates that precede lipid mixing and are dissimilar in their stability and propensity to complete fusion. The properties of these intermediates depend on liposome composition and, apparently, on the surface density of HAs. To explore the pathway of developmental fusion, we focused on the transmembrane protein EFF-1 of Caenorhabditis elegans, shown previously to be essential for fusion. We demonstrated that expression of EFF-1 drives hemifusion and fusion. Interestingly, syncytiogenesis requires EFF-1 in both fusing cells. Thus, the first example of a developmental fusogen activity, EFF-1-mediated fusion, shares hemifusion steps with pathways of viral and intracellular fusion but differs from these reactions in the homotypic organization of the fusion machinery. The analysis of the mechanisms by which fusion proteins as diverse as HA and EFF-1 drive fusion will, it is expected, yield new insights into mechanisms of ubiquitous fusion reaction.

C. elegans as a homotypic fusogen

Leikina, Chernomordik; in collaboration with Podbilewicz, Sapir, Shemer, Suissa, Valansi

At the outset of our investigation, the only protein fusogens that had been established were the proteins that mediate viral and intracellular fusion. Candidate fusogens for diverse and fundamental cell-cell fusion processes were characterized mainly at the genetic level and in vivo. In C. elegans hermaphrodites, 300 out of a total of 959 somatic nuclei reside in syncytial cells that originate through programmed and stereotyped cell-cell fusions in living embryos and larvae. Using genetic screens, we identified the transmembrane protein EFF-1 as a C. elegans candidate developmental fusogen that is required and sufficient to mediate fusion in the context of the living animal. However, we thought it possible that EFF1 mediates fusion by regulating an unidentified fusogen present on the surface of C. elegans cells. Potential membrane fusion proteins must meet several gold standards to be defined as fusogens: first, genetics and in vitro biochemical assays must demonstrate that the protein is necessary for membrane fusion events; second, cell-biological approaches must show that the protein is expressed and active at the fusion site; third, expression of the protein in heterologous cells must be sufficient to induce cell-cell fusion. While there are many candidate fusogens, only intracellular SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptors), membrane glycoproteins from various enveloped viruses, and type I proteins from nonenveloped reoviruses have passed the three tests.

To determine whether EFF1 is a bona fide fusogen or a regulator of fusion driven by another, yet unidentified EFF-1 fusogen, we expressed three EFF-1 isoforms in heterologous Sf9 insect cells. Our selection of these cells for investigation was informed by the fact that the cells normally do not form syncytia but have been used for cell fusion studies induced by viral fusogens. We have shown that the C. elegans transmembrane protein EFF-1, which is expressed at the surface of insect cells, initiates cell fusion and produces multinucleate syncytia. Thus, we established EFF-1 as the first bona fide developmental fusogen. We found that EFF-1-mediated fusion involves a hemifusion intermediate characterized by membrane mixing without cytoplasm mixing and identified it as a key intermediate in viral and intracellular fusion.

Importantly, syncytiogenesis requires EFF-1 to be expressed in both fusing cells. To test whether the same mechanism applies in vivo, we conducted genetic mosaic analysis of C. elegans and found that homotypic epidermal fusion needs EFF-1 in both cells. This is the first example of developmental fusogen activity. EFF-1-mediated fusion shares hemifusion steps with viral and intracellular fusion, but differs from these reactions in the homotypic organization of the fusion machinery. While viral fusogens are located at one of the fusing membranes, EFF-1 is required in both fusion partners. EFF-1 is also distinct from SNARE-dependent intracellular fusion where two fusing membranes carry different but complementary sets of protein fusogens. Homotypic machinery may provide better control of a developmental reaction than is required for heterotypic virus-host cell fusion during infection. For example, homotypic fusion may prevent fusion with cells at the edges of a multinucleated cell, allowing better control of syncytia size and shape. This mechanistic aspect of cell fusion during syncytia formation is critical for the normal development of many organs in nematodes and the formation of various tissues in mammalian organs as diverse as muscles, bones, placenta, and eye. We expect that, in other heterologous systems, EFF-1 expression may be used to fuse cells, with potential applications for gene therapy and manipulation of stem cell fates.

The dependence of viral fusion intermediates on the composition of the target membrane

Leikina, Chernomordik; in collaboration with Bailey, Markovic, Zhukovsky

Fusion mediated by hemagglutinin (HA), the envelope glycoprotein of influenza virus, is one of the best-characterized fusion reactions. To invade the cell, influenza virus binds to sialic acid receptors at the cell surface and delivers viral RNA into the cytosol by fusing the viral envelope with the membrane of acidified endosomes. Different experimental approaches have characterized the structure of the initial neutral-pH conformation of HA and the restructuring of HA at acidic pH while various experimental systems have been used to study the pathway of HA-mediated fusion. However, the mechanisms that couple low-pH-dependent restructuring of HA and membrane rearrangement remain elusive.

Conformational changes in acidified HA in the absence of a target membrane result in HA inactivation, which is detected as a decrease in the fusion rate after application of an additional low-pH pulse in the presence of the target membrane. Even when the membranes are already committed to fusion and would complete it at neutral pH, early fusion intermediates gradually lose fusion potential with time if fusion is blocked. We explored HA- and membrane-involving intermediates that precede lipid mixing and precede even the irreversible commitment to eventual lipid mixing. Characterization of the early intermediates might be facilitated by decreasing the area of contact and slowing the fusion reaction. Therefore, we replaced erythrocytes, often used as target membranes, with liposomes, which are protein-free lipid bilayer vesicles, and inhibited the fusion of HA-cells to liposomes by lowering the temperature.

To explore early intermediates in HA-mediated membrane fusion and their dependence on the composition of the target membrane, we studied lipid mixing between HA-expressing cells and liposomes containing phosphatidylcholine (PC) with different hydrocarbon chains. The composition of the target membrane affects the stability of fusion intermediates before lipid mixing. A more fusogenic target membrane effectively blocks nonproductive release of the conformational energy of HA, and, even for the same liposome composition, HA forms two types of fusion intermediates that are dissimilar in their stability and propensity to fuse. The two types of distinct intermediates observed in HA-cell/liposome fusion, and the effects of the target membrane on their properties, might be explained by the heterogeneity of the cell membrane and, in particular, by heterogeneous distribution of HA molecules at the cell surface. Given that the surface density of fusion proteins determines the rates, extents, and even phenotype of fusion, we hypothesize that, for liposomes of the same composition, fusion and inactivation patterns might differ in a manner dependent on the local HA concentration in the membrane patch under the liposome. We suggest that liposomes bound to regions of the cell with a high density of HA yield at 4°C lipid mixing in the case of dioleoyl phosphatidylcholine-cholesterol liposomes (DOPC-LS) and, in the case of distearoyl phosphatidylcholine-cholesterol liposomes (DSPC-LS), stable fusion intermediates; however, the latter do not inactivate with time but yield lipid mixing only at 37°C. In contrast, liposomes bound to regions with a low density of HA yield, in the case of DOPC-LS, fusion intermediates that do not support lipid mixing at 4°C but are stable and produce fusion upon a rise in temperature to 37°C. DSPC-LS at the low-density HA sites establish short-lived intermediates that yield lipid mixing only at 37°C and only if low-pH incubation at 4°C was relatively short.

We see that a more fusogenic target membrane and, possibly, a higher HA surface density can prevent HA inactivation at low temperatures. Assuming that both fusion and inactivation involve an irreversible conformational change of HA, the increase in fusogenicity of the target membrane or HA surface density might favor the fusion pathway versus the inactivation pathway. The suggested relationship between fusion and inactivation patterns and local HA density is supported by the lack of either partial lipid mixing or partial inactivation phenotypes in liposome fusion to viral particles with a homogeneously high density of HA. The diversity of the fusion intermediates emerging from our study emphasizes the importance of local membrane composition and local protein concentration in fusion of heterogeneous biological membranes.

COLLABORATORS

Austin L. Bailey, PhD, Cephalon, Inc., Hoboken, NJ
Tamar Gattegno, MSc, Technion-Israel Institute of Technology, Haifa, Israel
Irina Kolotuev, MSc, Technion-Israel Institute of Technology, Haifa, Israel
Ingrid Markovic, PhD, Center for Drug Evaluation and Research, FDA, Bethesda, MD
Benjamin Podbilewicz, PhD, Technion-Israel Institute of Technology, Haifa, Israel
Amir Sapir, PhD, Technion-Israel Institute of Technology, Haifa, Israel
Gidi Shemer, PhD, Technion-Israel Institute of Technology, Haifa, Israel
Meital Suissa, MSc, Technion-Israel Institute of Technology, Haifa, Israel
Clari Valansi, MSc, Technion-Israel Institute of Technology, Haifa, Israel
Mikhail A. Zhukovsky, PhD, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA

For further information, contact lchern@helix.nih.gov.