Stabilisation of Viral Membrane Fusion Proteins in Prefusion Conformation by Structure-Based Design for Structure Determination and Vaccine Development

Abstract

The membrane surface of enveloped viruses contains dedicated proteins enabling the fusion of the viral with the host cell membrane. Working with these proteins is almost always challenging because they are membrane-embedded and naturally metastable. Fortunately, based on a range of different examples, researchers now have several possibilities to tame membrane fusion proteins, making them amenable for structure determination and immunogen generation. This review describes the structural and functional similarities of the different membrane fusion proteins and ways to exploit these features to stabilise them by targeted mutational approaches. The recent determination of two herpesvirus membrane fusion proteins in prefusion conformation holds the potential to apply similar methods to this group of viral fusogens. In addition to a better understanding of the herpesviral fusion mechanism, the structural insights gained will help to find ways to further stabilise these proteins using the methods described to obtain stable immunogens that will form the basis for the development of the next generation of vaccines and antiviral drugs.

Keywords: fusion mechanism; fusogen; glycoprotein B; herpesvirus; membrane fusion; postfusion; prefusion; protein stabilisation; structure-based design; viral fusion protein.

Figure 2

Examples of stabilising modifications. Viral membrane fusion protein trimers stabilised in prefusion conformation by different mutations (red) depicted in ribbon representation [105] with insets showing specific examples of each stabilisation in detail. One protomer of each trimer is emphasised (dark grey) with structural elements essential for the fusion process coloured according to Figure 1: fusion peptides (light red), central helix (purple) and short helix (blue). PDB accession numbers are given in parenthesis for each structure. (A) Respiratory Syncytial Virus F glycoprotein [67] stabilised through interprotomer, intramolecular disulfide bond formation (yellow) and a T4 fibritin trimerisation domain (with protomers coloured grey, orange and light blue). (B) Human metapneumovirus F glycoprotein [57] stabilised by residue substitution to proline located at the hinge region between the short and central helix acting as helix breaker to prevent helix extension. (C) Human immunodeficiency virus envelope glycoprotein [77] stabilised through cavity filling mutations showing a phenylalanine inserted into the protein cavity of a neighbouring protomer (white surface). (D) Severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) spike protein [63] stabilised by charge balancing, whereby a repulsive force is replaced by an attractive force by replacing aspartate by asparagine to enable hydrogen bonding (dotted lines between red (oxygen) and blue (nitrogen) coloured atoms).

Figure 1

Three classes of viral membrane fusion proteins. Viral fusion class models illustrated by key domains in different states of the fusion process. The fusion proteins go through a number of conformations that drive the membrane fusion from approximation through hemifusion to fusion pore opening. Examples of structures are available for the pre- and postfusion states in all classes in contrast to intermediate conformations that are unstable and mostly extrapolated. Fusion loops or peptides are shown in pink, transmembrane domains as white cylinders while loop regions as well as regions not characteristic for the class are depicted as grey lines. (A) In class I fusion proteins, the long central coiled-coil domain (purple cylinders) is extended by refolding of the loop region connecting the short α-helical region (blue cylinder). After fusion loop anchoring the insertion state collapses, driven by the backfolding of the C-terminal portion of the helix (dark orange cylinder) to form a 6-helix bundle. (B) Class II fusogens rearrange after pH triggering, during insertion before the final trimer is formed. The fusion loops are located at the tip of a β-sheet in domain II (light orange box), while the motion is driven by a rotation of domain III (blue box) around domain I (green box) using a short hinge region (brown). (C) Class III is defined by an α-helical core similar to class I (purple cylinders), with a potential formation of a helical extension during insertion (blue cylinder), while the fusion loops are located at the tip of a β-sheet domain (light orange box) similar to class II.