Brook’s response is pretty good. I can provide a little more detail.
The spike protein of the virus both mediates binding to the ACE2 protein (which allows it to attach to the cell in the first place) and the fusion of the membranes. ACE2 is not involved in the fusion event, that is completely mediated by the spike, all ACE2 does is allow binding that brings the two membranes close together for a long time. The spike has two functional domains, one that is highly variable across coronaviruses that mediates attachment and is the reason different viruses attack different species and cell types, and one that is more highly conserved that triggers the membrane fusion. In order for the fusion to occur, the spike protein has to be processed by a protease that actually cuts the fusion domain apart from the binding domain. This does not make them fall off each other, they remain bound, but they no longer have a continuous backbone. This then allows a re-folding of the protein to a lower-energy state, which drives the fusion of the closely opposed membranes.
It appears from the literature I have found that the re-folding requires an acidic pH, suggesting that fusion probably requires endocytosis of the virus into the lysosome as it goes along for the ride with recycled ACE2 protein. (This is one of several reasons that chloroquine and hydroxychloroquine are being studied for efficacy, they are known to reduce the acidification of this cellular compartment.) They are still arguing about if the current virus has the spike protein cleaved during synthesis, or cleaved by proteases that are present in the lysosome where the ACE2 is recycled. One of the distinguishing characteristics of this virus compared to other coronaviruses is extra cuttable sequences between the two domains allowing more proteases to more easily cut the two domains apart, causing faster and more reliable viral entry. This has been noted in virulent strains of multiple other viruses in the past.
Brook’s response is pretty good. I can provide a little more detail.
The spike protein of the virus both mediates binding to the ACE2 protein (which allows it to attach to the cell in the first place) and the fusion of the membranes. ACE2 is not involved in the fusion event, that is completely mediated by the spike, all ACE2 does is allow binding that brings the two membranes close together for a long time. The spike has two functional domains, one that is highly variable across coronaviruses that mediates attachment and is the reason different viruses attack different species and cell types, and one that is more highly conserved that triggers the membrane fusion. In order for the fusion to occur, the spike protein has to be processed by a protease that actually cuts the fusion domain apart from the binding domain. This does not make them fall off each other, they remain bound, but they no longer have a continuous backbone. This then allows a re-folding of the protein to a lower-energy state, which drives the fusion of the closely opposed membranes.
It appears from the literature I have found that the re-folding requires an acidic pH, suggesting that fusion probably requires endocytosis of the virus into the lysosome as it goes along for the ride with recycled ACE2 protein. (This is one of several reasons that chloroquine and hydroxychloroquine are being studied for efficacy, they are known to reduce the acidification of this cellular compartment.) They are still arguing about if the current virus has the spike protein cleaved during synthesis, or cleaved by proteases that are present in the lysosome where the ACE2 is recycled. One of the distinguishing characteristics of this virus compared to other coronaviruses is extra cuttable sequences between the two domains allowing more proteases to more easily cut the two domains apart, causing faster and more reliable viral entry. This has been noted in virulent strains of multiple other viruses in the past.