Biophysical characterization of membrane-active regions of structural and non-structural proteins from dengue virus

  1. de Castro Nemésio, Henrique
Dirigida por:
  1. José Villalaín Boullón Director/a

Universidad de defensa: Universidad Miguel Hernández de Elche

Fecha de defensa: 25 de julio de 2014

Tribunal:
  1. Manuel José Estevez Prieto Presidente/a
  2. María Amparo Estepa Pérez Secretario/a
  3. Pablo Gastaminza Landart Vocal
  4. Senena Corbalán García Vocal
  5. Jesús Salgado Benito Vocal

Tipo: Tesis

Resumen

Dengue virus, part of the Flaviviridae family is the most prevalent virus in the human population and is responsible for the highest morbidity and mortality rates worldwide, with about 400 million infections estimated yearly in tropics and subtropics [109]. People infected by DENV present symptomatology with varying severity degrees: asymptomatic, Dengue fever (DF), Dengue haemorrhagic fever (DHF) and Dengue Shock syndrome (DSS). Despite its prevalence, there are currently neither effective antivirals nor antiviral drugs against DENV, being the sole eradication strategies based on the control of mosquito populations [55, 107, 117]. This virus is classified into three distinct serotypes with 69-78 % primary sequence identity [356] and it is composed of a nucleocapsid, formed by several molecules of protein C bound to the viral RNA genome, enclosed by a host derived lipid bilayer where 180 copies of proteins M and E are embedded as heterodimers. Its positive single-stranded RNA genome of some 11 kb possesses a single ORF that encodes a single polyprotein of about 3000 amino acids. After appropriate processing by viral and host proteases, this polyprotein gives rise to three structural proteins C, prM and E and seven non-structural proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5 [79, 119]. DENV enters cells by receptor mediated endocytosis and the decrease in pH inside endosomes triggers a conformational change that results in the insertion of the fusion peptide (located in protein E, a class II fusion protein) into the endosomal membrane, resulting in the release of the nucleocapsid into the cytoplasm [365]. Viral replication and assembly processes occur in ER-derived membranes, sequestered and reorganized by the virus. Following its assembly, the immature virions start their transit through the trans-Golgi network, where a decrease in pH triggers the furin cleavage of the peptide bond between the pr peptide and the M protein. When the virion is exocyted, the pr peptide is released and the viral particle is rendered fully mature. Both in entry and exit steps, all viral proteins require and use membranes to exert their functions, e.g. by modulating their structure to form replication complexes, to execute their fusion process or even to carry out viral assembly. The effects DENV proteins have on membranes are known, yet the mechanisms by which those occur are elusive, the proteins involved are even less known and the exact regions in those proteins that interact with membranes are virtually unknown. Therefore, we sought to highlight the membrane-active regions of DENV¿s structural proteins and the lesser known and most hydrophobic non-structural proteins. Protein C forms mainly ¿-helical dimers in solutions, where two regions possess interesting biophysical characteristics: one is highly hydrophobic, suggesting possible membrane interactions [215] while the other has a high positive charge concentration, what would point to a possible interaction with the negatively charged viral RNA genome [219]. It is known that this protein is responsible for the specific encapsidation of the genome [79, 219] and that it accumulates around ER-derived lipid droplets [218]. We have identified two membrane-active regions in this protein, corresponding to the highly hydrophobic region from residues 39 to 56 and its C-terminal signal sequence. Protein E is a class II fusion protein with its N-terminal region (residues 1 to 395) divided into three domains formed by beta structures [54, 119], an N-terminal domain I, flanked on one side by domain II (where the fusion peptide is located between residues 98 to 112) and on the other by domain III, the putative location of the receptor binding sites. Its C-terminal region would be composed of a stem region followed by two helical transmembrane domains. These regions contribute to the conformational changes required for membrane fusion and the transmembrane regions would be involved in protein-lipid and lipid-lipid interactions [75, 164, 285, 360]. We have found at least five membrane-active regions in this protein, coincident with the fusion peptide (residues 88 to 122), a proline-rich region involved in protein-protein interaction (residues 198 to 221), another previously described hydrophobic region (residues 270 to 309), a region coincident with the stem region (residues 406 to 422) and finally a region from residues 428 to 479 that contains part of the stem region and one of the previously described transmembrane domains. These results define the membrane interacting regions of these proteins and corroborate the importance of these in the interactions of native proteins. A peptide derived from protein C from residues 39 to 56, DENV2C6, located in a highly hydrophobic region in the C protein dimer, required for appropriate viral maturation, vesicle binding and viral assembly [215, 218] induces significant effects to membrane model systems. It binds with high affinity to several membrane models and ruptures them with no clear distinction between membranes with different composition, It can also modulate the thermotropic behaviour of membranes, inducing the formation of lipid phases with different peptide concentrations and altering the steady-state fluorescence anisotropy of probes inserted into the lipid palisade. This peptide significantly affects BMP (a lipid found predominantly in endosomes [161]) containing membranes, what suggests a possible function in viral fusion during infection. Moreover, FTIR experiments have shown that this peptide has different secondary structure in solution and in DMPC membranes, meaning that it is a two way protein-lipid interaction. Proteins NS4A and NS4B are highly hydrophobic proteins involved in several functions in the viral cycle. It is known that NS4A modulates the ATPase function of NS3 [265], co-localizes with other viral proteins in the replication complex and it can be found in ER membranes, modulating its curvature [264, 267]. Furthermore, this protein seems to play a role in the interferon response [263], inhibition of apoptosis through autophagy induction what increases the cellular lifespan [271]. A recently described topology model suggests two transmembrane segments separated by a membrane interacting domain and its C-terminal region is a 2k fragment, responsible for the translocation of NS4B into the ER lumen [264]. We could delineate two regions that affect the structure and integrity of the lipid membrane from residues 52 to 90 and 90 to 125, matching the transmembrane segments of the topological model. The 2k fragment does not seem to affect ER membranes, what would suggest conformational changes dependent on lipid composition, a fairly common process of membrane modulation in viral cycles. As for protein NS4B, it seems to be the most important protein in interferon response inhibition by inhibiting the STAT1 phosphorylation [273]. Its co-localization with dsRNA and proteins NS3 and NS5 suggests it might play a role in the viral replication [275]. Adding to that, this protein forms dimers through a cytosolic loop. These findings provide a clear indication of its hybrid lipid-lipid and protein-lipid interacting capabilities [276]. In its most recent topology model, this protein is composed of at least three transmembrane domains preceded by two membrane-interacting domains [272]. We highlighted four membrane-active segments in this protein from residues 50 to 80, 94 to 127, 163 to 190 and 210 to 240 and at least one segment coincident with the dimerization region that could interact with membranes. We conclude that this protein actively modulates and reorganizes membranes and also possibly interacts with other proteins. The prM protein is composed of two different regions, peptide pr from residues 1 to 91 and the M protein from residues 91 to 166, that is in turn composed of a stem region from residues 112 to 132 followed by two transmembrane segments [125, 208, 209, 223]. Apart from tis fusion preventive function, by protecting the fusion peptide, this protein plays a role in intracellular transport and apoptosis regulation [224-228]. It has been recently found that this protein forms oligomers in ER-derived membranes while reorganizing them [102]. We could determine at least on protein interacting region in the pr peptide that would match the region of contact with protein E and a membrane interacting region in the same moiety. As for protein M, the transmembrane segments were detected and the stem region modulates the thermotropic behaviour of membranes. These results provide the confirmation that this protein interacts with membranes and define the regions involved in this phenomenon Protein NS2A is the less characterized of all DENV proteins yet the literature points to its roles in important processes of the viral cycle. It is part of the replication complex [261], it is required in the viral assembly, shown by the fact that mutations in this protein significantly reduce the viral load [260], it is needed for proper NS1 processing [260] and , in concert with NS4A and NS4B, inhibits the interferon response [263]. It is a highly hydrophobic protein composed of at least five transmembrane and two membrane interacting regions [260]. Its location in membranes and hydrophobic affinity suggest that this protein might also reorganize membranes thus aiding the formation of the replication complexes. Using the above stated methodology, we defined at least two membrane interacting domains from residues 25 to 41 and 103 to 183. The first region matches one of the membrane interacting regions defined in the topology model and its effect is dependent on the lipid composition. This region was characterized in detail and its interaction with membranes depends not only on the lipid composition but also on the ionic strength of its environment. Its interaction with membranes is mainly electrostatic and its selective effect hints at a possible conformational change switching between membrane bound and membrane-free states, thus modulating membrane structure. These results yielded three published scientific papers, two already sent for their publication, several congress abstracts and this PhD. Thesis.