Filovirus Glycoprotein Sequence, Structure and Virulence

Leading Ebola subtypes exhibit a wide mortality range, here explained at the molecular level by using fractal hydropathic scaling of amino acid sequences based on protein self-organized criticality. Specific hydrophobic features in the hydrophilic mucin-like domain suffice to account for the wide mortality range. Significance statement Ebola virus is spreading rapidly in Africa. The connection between protein amino acid sequence and mortality is identified here. Introduction Ebola filoviruses exhibit ~ 70% glycoprotein (GP) subtype similarities, with wide variations in virulence (from nearly 90% to nearly 0% mortality) [1]. Although the Marburg species has only 30% GP similarity to Ebola, its length and viral morphology are similar, and it also has high mortality levels ~ 50% [2,3]. The static GP domain structure of the most virulent Ebola subtype, Zaire or ZEBOV, whose structure bound to an antibody is shown in Fig. 1 of [4], is the basis of many studies [5-9]. The molecular basis for explaining the widely differential pathogenicity of the filoviruses, which depends on many steps from membrane penetration to selective disruption of cell-cell binding, remains elusive [1,6]. Two GP mechanisms have been suggested as possible origins of virulence variations, different folds (Ebola Zaire is the only structure known) [2], and different flexibility [6]. Ebola subtypes share 70% sequence similarity, and subtypes with more than 40% similarity generally have similar folds [10]. A characteristic feature of filoviruses is a large, disordered, and highly glycosylated mucin-like domain (MLD), which is a natural candidate for disruption of cell-cell binding [6]. Here we present the results of fractal hydropathic analysis which quantifies the subtype dependence of Ebola MLD flexibility on length scales smaller than the MLD length (~ 150 aa). Ebola and Marburg sequence similarity is only 30%, but our analysis shows how the