Conopressin, a nonapeptide disulfide CFIRNCPKG amide present in cone snail venom, undergoes a facile cleavage at the Cys6−Pro7 peptide bond to yield a disulfide bridged b 6 ion. Analysis of the mass spectral fragmentation pattern reveals the presence of a major fragment ion, which is unambiguously assigned as the tripeptide sequence IRN amide. The sequence dependence of this unusual fragmentation process has been investigated by comparing it with the fragmentation patterns of related peptides, oxytocin (CYIQNCPLG amide), Lys-vasopressin (CYFQNCPKG amide), and a series of synthetic analogues. The results establish the role of the Arg4 residue in facilitating the unusual N−C α bond cleavage at Cys6. Structures are proposed for a modified disulfide bridged fragment containing the Cys1 and Cys6 residues. Gas-phase molecular dynamics simulations provide evidence for the occurrence of conformational states that permit close approach of the Arg4 side chain to the Cys6 C β methylene protons.

No abstract

The spike (S) protein is a trimeric, membrane-anchored fusion protein that enables coronaviruses, such as the SARS-CoV-2, to recognize and fuse with their hosts' cells. While the prefusion and postfusion structures of the ectomembrane domain of the spike protein are available, the corresponding organization of its transmembrane domain is obscure. Since the transmembrane and ectomembrane domains of fusion proteins are conformationally linked, an understanding of trimerization and transmembrane conformations in the viral envelope is a prerequisite to completely understand viral fusion by the spike protein. To address this, we computationally explored the self-assembly of the SARS-CoV-2 spike transmembrane domain, starting first by determining the membrane boundaries of the spike transmembrane helix. Using atomistic molecular dynamics simulations, we found the spike protein transmembrane domain to be plastic, and the transmembrane helix to be very dynamic. The observed movements of the helix changed the membrane embedded sequence, and thereby affected the conformational ensemble of the transmembrane assembly in Martini coarse grained simulations, even flipping the super-helical handedness. Analysis of the transmembrane organization of the spike transmembrane helix provided rich insights into the interfaces utilized to self-associate. Moreover, we identified two distinct cholesterol binding regions on the transmembrane helix with different affinities for the sterol. The cholesterol binding pockets overlapped with regions involved in the initiation of transmembrane protein-protein interaction. Together, the results from our multiscale simulations not only provide insight into understudied trimeric helical interfaces in biomembranes, but also enhance our understanding of the elusive transmembrane conformational dynamics of SARS-CoV-2 spike and more generally of viral fusion proteins. These insights should enable the inclusion of the conformations of the spike protein transmembrane domain into the prevalent models of virus fusion.