Thesis advisor: Welkin E. Johnson / Tetherin/BST2 is an interferon-inducible antiviral factor that restricts the egress of numerous enveloped viruses including HIV-1. Consequently, many viruses have evolved mechanisms to actively or passively evade restriction by tetherin. Most studies conducted to date focused on the tetherin-evasion mechanism of complex retroviruses like HIV and SIV, which encode accessory proteins like Vpu and Nef respectively to counteract tetherin-mediated restriction. However, there is a wide gap in knowledge in understanding how simple retroviruses (that includes alpharetroviruses, some betaretroviruses and gammaretroviruses) that lack obvious accessory proteins like HIV-1 Vpu and SIV-Nef, evade restriction by tetherin. In this dissertation, I have established that Simian retrovirus type-3, a prototypical type-D betaretrovirus, isolated from Asian macaques, is restricted by human tetherin but not by rhesus macaque tetherin. This differential sensitivity indicated that SRV-3 has a mechanism to evade tetherin-mediated restriction. I have identified the SRV-3 envelope (Env) glycoprotein as the viral determinant of tetherin antagonism, and have also found that SRV-3 envelope expression in-trans was sufficient to rescue a heterologous virus from tetherin. SRV-3 Env resulted in cell-surface down-modulation of rhesus tetherin, and this mechanism of tetherin-antagonism is independent of the SRV-3 Env trafficking pathway. The target specificity of SRV-3 Env overlapped a stretch of five residues (G14DIWK18) in the rhesus tetherin cytoplasmic tail that are absent from human tetherin. Additionally, I was able to show that SRV-3 Env physically interacts with rhesus tetherin by targeting the G14DIWK18 motif. SRV-3 belongs to a large supergroup of retroviruses, called the RDR Interference Supergroup. Due to this reason, I screened additional RDR envelope glycoproteins for their ability to antagonize a panel of tetherin homologs. All the RDR envelopes tested were sensitive to human tetherin but exhibited anti-tetherin activity when tested against a panel of tetherin homologs from squirrel monkey, baboon, dog and cat. I also found that several non-RDR gammaretroviral envelope glycoproteins also have anti-tetherin function. Thus, tetherin-antagonism is not just restricted to the envelope glycoproteins of retroviruses in the RDR interference supergroups but extends to other non-RDR gammaretroviruses as well. To my knowledge, this is the first characterization of gamma-type envelopes as tetherin antagonists. Thus, in the absence of a dedicated tetherin antagonist, many simple retroviruses in the beta- and gammaretrovirus genera may evade tetherin-mediated restriction through neo-functionalization of their envelope glycoproteins. We speculate that the evolutionary success of the gamma-type envelope may be due, at least in part, to this anti-tetherin function. / Thesis (PhD) — Boston College, 2018. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

Du fait de leur présence à la surface de la particule virale, les glycoprotéines d’enveloppe E1 et E2 du virus HCV jouent un rôle essentiel dans sa morphogenèse ainsi que lors de son entrée dans la cellule hôte. Jusqu’à récemment, les travaux de recherche sur les glycoprotéines d’enveloppe du virus HCV se sont essentiellement focalisés sur E2 car elle est la protéine d’attachement du virus. De plus, elle est la cible majeure des anticorps neutralisants et il a été longtemps postulé qu’elle était la protéine de fusion du virus. Cependant, les récentes publications de la structure de E2 ne mettent pas en évidence la présence d’un peptide de fusion et sa structure ne correspond pas aux critères attendus pour une protéine de fusion, suggérant que la glycoprotéine E1 seule ou en association avec E2 pourrait être responsable de l’étape de fusion. La structure de la région N-terminale de E1 (acides aminés 192 à 270) a récemment été résolue et a mis en évidence la présence d’une épingle à cheveux formée par 2 feuillets beta (β1 et β2) suivie par un segment de 16 acides aminés qui forme une hélice alpha (α1) flanquant 3 feuillets beta antiparallèles (β3, β4 et β5). En plus de la caractérisation de ces structures secondaires de E1, une région qui se situe au milieu de la protéine (approximativement entre les résidus 274 et 292) a été proposée avoir un rôle actif au cours du processus de fusion et elle pourrait correspondre à un peptide de fusion.Nous nous sommes basés sur ces travaux récents pour investiguer le rôle fonctionnel de la glycoprotéine E1 par une approche de mutagenèse dirigée des résidus conservés dans la région N-terminale et dans la région du potentiel peptide de fusion, dans le contexte d’un clone infectieux du HCV. Comme attendu, nos résultats indiquent que ces mutations introduites dans E1 n’ont aucun effet sur la réplication virale. Cependant, vingt-et-un parmi les vingt-huit mutants produits conduisent à une atténuation ou une perte de l’infectiosité virale. D’une manière très intéressante, deux mutants atténués, le T213A et le I262A, se sont montrés moins dépendants au co-récepteur claudine-1. D’autre part, nous avons montré que ces mutants utilisent un autre récepteur de la famille des claudines (claudine-6) pour l’entrée virale, indiquant ainsi un changement de dépendance à son co-récepteur claudine-1. A l’opposé, deux autres mutants, le L286A et le E303A, se sont révélés avoir une plus grande dépendance au co-récepteur claudin-1 pour l’entrée dans les cellules d’hépatome. Au cours de ce travail, nous avons également identifié une mutation intéressante à proximité du potentiel peptide de fusion. Cette mutation, G311A, conduit à la sécrétion de particules virales entières mais non infectieuses, suggérant un défaut d’entrée cellulaire pour ce virus. De façon très surprenante, nous avons également identifié une mutation (D263A) qui conduit à la sécrétion de particules virales dépourvues d’ARN génomique. Une caractérisation plus poussée de ce mutant a de plus révélé une modification dans la co-localisation subcellulaire entre l'ARN viral et la glycoprotéine E1, mettant en évidence pour la première fois un dialogue croisé entre E1 et l'ARN génomique du HCV lors de la morphogenèse du virus.En conclusion, nos observations permettent d’identifier précisément les régions spécifiques de la protéine E1 qui jouent un rôle dans l’assemblage et l’entrée du virus dans la cellule, mettant en évidence le rôle majeur de la glycoprotéine E1 au niveau des différentes étapes du cycle infectieux du HCV. / Being part of the viral particle, HCV envelope glycoproteins E1 and E2 play an essential role in virion morphogenesis as well as in HCV entry into liver cells. These glycoproteins form a non-covalent heterodimer, and until recently, research on HCV envelope glycoproteins has been mainly focused on E2. Indeed, this glycoprotein is the receptor-binding protein, it is also the major target of neutralizing antibodies and it was postulated to be the fusion protein. However, the recent publications of the structure of E2 do not show the presence of a fusion peptide and its structure does not fit with what one would expect for a fusion protein, suggesting that E1 alone or in association with E2 might be responsible for the fusion step. Concerning E1, only the crystal structure of the two-fifth N-terminal region, comprising amino acids 192 to 270, has been reported. This partial structure reveals a complex network of covalently linked, intertwined homodimers. The overall fold of the N-terminal E1 monomer consists of a beta-hairpin (β1 and β2) followed by a segment composed of a 16 amino-acid long alpha-helix (α1) flanking a three-strand antiparallel beta-sheet (β3, β4 and β5). In addition to the characterization of secondary structures within E1, a region located in the middle of the polypeptide (approximately between aa 274 and 292) has been suggested to play an active role during the fusion process and might potentially act as a fusion peptide. We took advantage of these recently published data to further investigate the functional role of HCV glycoprotein E1 by using a site-directed mutagenesis approach targeting conserved amino acids in the N-terminal region as well as in the region postulated to contain the fusion peptide in the context of an infectious clone. As expected, our results indicate that these mutations have no effect on virus replication. However, twenty-one out of twenty-eight mutations led to attenuation or inactivation of infectivity. Interestingly, two attenuated mutants, T213A and I262A, were less dependent on tight junction protein claudin-1, a co-receptor for HCV. Instead, these mutant viruses relied on another claudin (claudin-6) for cellular entry, indicating a shift in receptor dependence. In contrast, two other mutants, L286 and E303, were more dependent on claudin-1 for cellular entry into hepatoma cells cells. We also identified an interesting mutation downstream of the putative fusion peptide, G311A, which leads to the release of non-infectious particles having a defect in cellular entry. Finally, an unexpected phenotype was also observed for D263A mutant, which was no longer infectious but led to the secretion of viral particles devoid of genomic RNA. Further characterization of the D263A mutant revealed a change in subcellular co-localization between HCV RNA and E1, highlighting for the first time a crosstalk between HCV glycoprotein E1 and the genomic RNA during HCV morphogenesis.In conclusion, our observations allowed for the identification of specific regions in the E1 glycoprotein that play a role in virion assembly and entry, highlighting the major role played by this protein at different steps of the HCV infectious cycle.

La transmission mère-enfant (TME) est un modèle naturel permettant d’explorer le rôle des anticorps neutralisants dans la protection vis-à-vis de l’infection par le VIH-1 ainsi que les caractéristiques des virus transmis aux enfants malgré la présence d’anticorps maternels. Au cours de mes travaux, nous avons confirmé l’importance de la région V2 de la glycoprotéine de surface (gp120) du VIH-1 dans la résistance du virus à la neutralisation. L’analyse des propriétés de variants transmis à l’enfant et de variants maternels a mis en évidence une grande hétérogénéité de leurs propriétés biologiques sans pour autant identifier de propriétés spécifiques conférant un avantage sélectif aux virus transmis. Cependant, nous avons montré que les virus transmis étaient plus sensibles à la neutralisation par deux anticorps monoclonaux humains largement neutralisants récemment identifiés, PG9 et PG16, une observation qui pourrait avoir des applications intéressantes en termes de prévention de la TME ou de développement vaccinal. Enfin, nous avons identifié deux résidus très conservés de la gp120 impliqués dans la sensibilité à la neutralisation par PG9 et/ou PG16. / Mother-to-child transmission (MTCT) provides a natural model for studying the role of neutralizing antibodies in preventing HIV-1 infection, and the characteristics of the virus transmitted to the infants despite the presence of maternal antibodies. During my thesis works, we confirmed that the V2 domain of the surface envelope glycoprotein (gp120) of HIV-1 plays a major role in resistance to neutralization. The analysis of properties of variants transmitted to infant and maternal variants showed a wide spectrum of their biological properties, but we did not identify any specific property conferring a selective advantage for transmission of the virus to the infant. Nevertheless, we showed that the transmitted variants were more sensitive to neutralization by the two recently described broadly neutralizing monoclonal antibodies PG9 and PG16. This observation should have interesting applications in terms of prevention of MTCT or vaccine development. Finally, we identified two gp120 cross-clade conserved residues involved in neutralization sensitivity to PG9 and/or PG16.

Le virus de l’hépatite C (HCV) est l’agent responsable de l’hépatite C, maladie qui touche environ 3% de lapopulation mondiale. Une des caractéristiques de cette infection est son évolution dans 60 à 90% des casvers des formes chroniques avec des complications sévères telles que la cirrhose et le carcinomehépatocellulaire. Un des handicaps majeurs de la recherche sur le HCV est l’absence de systèmes decultures in vitro efficaces et de modèles animaux adaptés car le HCV n’infecte que l’homme et le chimpanzé.l’anticorps D32.10. Pour cela, nous avons développé un test de cellbindinget nous avons montré quel’interaction des particules virales sériques (HCVsp) radiomarquées à l’Iode 125 avec les celluleshépatocytaires (Huh‐7 et HepaRG) est spécifique et saturable impliquant des sites de haute et faible affinité.De plus, l’anticorps D32.10 est capable d’inhiber spécifiquement et efficacement les interactions de hauteaffinité entre les HCVsp et les cellules HepaRG avec une IC50 ≤ 0,5 μg/ml. Nous avons mis en évidence quel’inhibition est plus efficace lorsque nous utilisons sélectivement une population de particules HCVenveloppées exprimant fortement E1E2. Récemment, nous avons développé un système d’infection originaldes cellules HepaRG qui sont des cellules progénitrices du foie par les HCVsp et avons montré quel’infection, la réplication et la propagation dépendent de l’état de prolifération/différenciation de cescellules. Nous avons aussi démontré que les particules virales produites dans ce système contiennent del’ARN viral, expriment les protéines d’enveloppe E1E2 et sont infectieuses. Des études préliminairesmontrent que l’anticorps D32.10 inhibe fortement l’infection (95% à 80% aux jours 14 et 21 aprèsinfection) vraisemblablement au niveau des étapes précoces du cycle viral.Dans un second temps, nous avons recherché la prévalence des anticorps de même spécificité que le D32.10(anti‐E1E2A,B) dans différents groupes de patients HCV positifs afin de déterminer leur significationbiologique. Par un test ELISA utilisant les peptides biotinylés E1, E2A et E2B dans la phase de capture, nousavons démontré que la réponse anticorps anti‐E1E2A,B était présente dans 90% des cas chez les patientsqui guérissent spontanément avec des titres élevées (≥ 1/1000). Cette réponse humorale est absente ourare (< 10%) chez les patients porteurs chroniques non traités ou non répondeurs aux traitementsantiviraux. Une étude longitudinale a été réalisée chez des patients non répondeurs ou répondeursdéveloppant une réponse virologique soutenue à une bithérapie standard, interféron pégylé plus ribavirine.L’analyse statistique des résultats a montré que les anticorps anti‐E1E2A,B pouvaient être prédictifs de laréponse au traitement avec une spécificité et une valeur prédictive positive de 100%.La convergence des résultats in vitro et in vivo supporte un rôle neutralisant de l’anticorps monoclonalD32.10, permettant d’envisager son utilisation en immunothérapie. / Hepatitis C Virus (HCV) is the major etiological agent of liver disease in the world with approximately180 million people who are seropositive. The majority (60‐90%) of infected individuals progressesto chronic hepatitis that increases their risk for developing cirrhosis and hepatocellular carcinoma.One of the major limitations of HCV research is the lack of efficient in vitro culture systems andappropriateanimal models. vitro direct cell‐binding assay and an infection system of the human HepaRG cell line were developedby using HCVsp. The HepaRG cells possess potent ability to acquire a mature hepatocyte phenotype.The E1E2‐specific mAb D32.10 was shown to inhibit efficiently and specifically high affinityinteractionsthrough glycosaminoglycans and the CD81 tetraspanin between HCVsp and HepaRGcells with an IC50 = 0.5 μg/ml. This inhibition was more efficient when E1E2‐positive envelopedHCVsp were used selectively for binding studies (IC50 < 0.5 μg/ml). Establishment of infection,replication and propagation of HCVsp were shown to depend on the proliferation/differentiationstage of HepaRG cells. Persistent HCV infection in HepaRG cells could be obtained with production ofE1E2/RNA(+) infectious HCV particles. Preliminary data showed a complete early inhibitory effect ofthe D32.10 mAb on virion RNA production in HepaRG culture supernatants (95% at D14 and 80% atD21 post‐infection).Furthermore, the detection of the anti‐E1E2/D32.10‐binding peptide antibodies during natural HCVinfection demonstrated significant prevalence (90%) of these antibodies: (1) in patients whorecovered spontaneously from HCV infection with high titers compared to patients with chronichepatitis C, and (2) in patients who are complete responders compared to non responders toantivirals. Kinetic analyses revealed that the anti‐E1E2/D32.10‐like humoral response appeared veryearly with high titers (≥ 1/1000) and was associated with complete virus eradication. The positiveand negative predictive values (ROC curve analysis) for achieving or not a sustained viral response toantiviral therapy are 100% and 86%, respectively, reflecting diagnostic accuracy. The anti‐E1E2/D32.10‐binding peptide antibodies may thus predict the outcome of HCV infection andrepresent a new relevant pronostic marker in serum for the HCV diagnosis.Convergence of in vitro and in vivo data strongly support the neutralizing activity of the D32.10 mAb,and thus immunotherapeutic potential of this unique anti‐E1E2 D32.10 mAb.

Hepatitis C Virus (HCV) infects more than 170 million people worldwide but no vaccine is available yet. HCV entry may represent a promising target for therapies and is mediated by two envelope glycoproteins, E1 and E2, assembled as heterodimer onto the virus surface. However, how E1 and E2 dialog, structurally rearrange and act together during these steps remain poorly defined. In this work, we aimed to clarify the interrelation of E1E2 during virus entry, thus opening ways to potential new therapeutic strategies. We first investigated whether a strong genetic divergence between E1E2 heterodimers may highlight distinct functions. We observed that B-cell derived E1E2 were specialized for B-cell infection, suggesting that new functions can emerge from the E1E2 conformational plasticity. In a second approach, we identified a conserved dialog between E1 and the domain III of E2 that was critical for virus binding and fusion. Moreover, a computational model predicted a strong co-evolution between E1 and E2 as well as potential structural rearrangements, suggesting that HCV E2 is likely a fusion protein able to fold over via its domain III through the mediation of E1. Altogether, these different works highlight that E1 and E2 are involved in complex dialogs that regulate the heterodimer folding and functions, suggesting that E1E2 heterodimer is more likely a single functional protein entity than an association of two proteins with specific functions.

Les glycoprotéines d'enveloppe du virus de l'immunodéficience humaine de type 1 (VIH-1) sont impliquées dans une étape importante du cycle de réplication, l’entrée virale. La liaison de la glycoprotéine gp120 à CD4 contribue à la fixation du virus à la cellule cible et déclenche des changements conformationnels dans celle-ci afin de permettre à la gp120 se lier au récepteur de chimiokine CCR5 ou CXCR4. Contrairement à tous les enveloppes du groupe phylogénétique M, qui possède une sérine à la position 375, ceux du clade CRF01_AE possèdent une histidine. Ce résidu fait partie d'une cavité "Phe43" dans laquelle le résidu 43 de CD4 fait contact avec la gp120. Ici, nous avons évalué les conséquences fonctionnelles du résidu 375 en le remplaçant par une serine dans deux enveloppes CRF01_AE (CM244 et 92TH023), la sérine étant présente dans tous les autres clades du groupe M. Nous avons observé que la réversion de l'acide aminé 375 vers une sérine entraine une perte de fonctions des deux Envs CRF-AE tel que mesuré par une perte de l'infectivité et leur capacité à médier la fusion cellule à cellule. Le phénotype observé était la conséquence d’une très faible interaction avec CD4, qui n’était pas le résultat d’une enveloppe défectueuse ou instable. Par conséquent, la modification de l’acide aminé en position 375 a aussi altéré la sensibilité de neutralisation du virus par sCD4, qui en était diminuée. De plus, on a observé que certaines mutations des couches topologiques du domaine interne de la gp120 ont permis un rétablissement partiel de la fonctionnalité en restaurant l’interaction avec CD4. Les niveaux d’infectivité et de fusion des mutants des couches topologiques se rapprochant des enveloppes sauvages de CRF01_AE suggèrent une coévolution entre la cavité phe43 et les résidus du domaine interne de la gp120. Une compréhension des différences qui existent entre les enveloppes de CRF01_AE et les enveloppes de virus du groupe M, permettra d’avoir une idée sur la fonctionnalité des glycoprotéines d'enveloppe. Il serait possible de mettre en évidence des mécanismes impliqués dans les changements conformationnels des Envs qui permettent l’évasion des anticorps dirigés contre elle. De plus, ces possibles différences dans les enveloppes de CRF01_AE pourraient être exploitées pour le développement des différentes approches vaccinales. / The envelope glycoproteins (Env) from human immunodeficiency virus type 1 (HIV-1) mediate viral entry. The binding of the HIV-1 gp120 glycoprotein to CD4 contributes to the attachment of the virus to the target cell and triggers conformational changes in gp120 that allow high-affinity binding to the chemokine receptor CCR5 ou CXCR4. Contrary to all other Envs from the same phylogenetic group M, which possess a serine at position 375, the majority of CRF01_AE strains possess a histidine. This residue is part of the “Phe43” cavity, where residue 43 of CD4 engages with the gp120. Here we evaluated the functional consequences of replacing this residue in two CRF01_AE Envs (CM244 and 92TH023) by a serine, present in all the other clades from group M. We observed that reversion of the 375 amino acid to a serine resulted in a loss of functionality of both CRF_AE Envs, as measured by a loss in infectivity and their ability to mediate cell-to-cell fusion. The observed phenotype was the result of a weak interaction with CD4, which was not due to defective processing or trimer stability of these Envs. Therefore, modification of the amino acid at position 375 has also altered the sensitivity of virus neutralization by sCD4, which was reduced. Importantly, mutation of certain residues of the gp120 inner domain layers were able to partially restore wild-type levels of infectivity and cell-to-cell fusion to both CRF01-AE H375S Envs, suggesting a co-evolution of the Phe43 cavity and the gp120 inner domain. An understanding of the differences between the CRF01_AE envelopes and envelopes from group M presenting a serine at position 375 will provide a better knowledge about the functionality of the envelope glycoproteins. Among other things, it would be possible to identify the mechanisms involved in conformational changes of glycoproteins as well as those involved in the escape of envelope recognition by Env specific antibodies. Moreover, these possible differences in CRF01_AE envelopes could be exploited for the development of different vaccine approaches.

Le Virus de l’Hépatite C (VHC) infecte 170 millions de personnes dans le monde mais aucun vaccin n’est encore disponible. Le processus d’entrée du VHC dans les hépatocytes représente une cible prometteuse pour le développement de stratégie thérapeutique et est finement régulé par un nombre par les deux glycoprotéines d’envelope du VHC, E1 et E2, assemblé sous la forme d’un hétérodimère incorporé à la surface des particules virales. Cependant, comment E1 et E2 dialoguent, modifient leurs conformations et se coordonnent mutuellement au cours de l’entrée reste encore à être définit. Dans ce travail, nous avons souhaité clarifier l’interrelation entre E1 and E2 au cours de l’entrée afin d’ouvrir la voie à de potentiels stratégies thérapeutiques. Nous avons tout d’abord examiné si une importante divergence génétique entre des hétérodimères E1E2 pouvait être liée à l’existence de fonctions particulières. Nous avons observé une spécialisation des E1E2 isolé des Lymphocytes B pour l’infection de ces mêmes cellules mais pas des hépatocytes, suggérant que de nouvelles fonctions peuvent émerger de la plasticité conformationel de E1E2. Dans un second temps, nous sommes parvenus à identifier un dialogue conservé entre E1 et le domaine III de E2 (E2 DIII), critique pour les processus d’attachement et de fusion du VHC. Nous avons aussi montré grâce à une approche bio-informatique l’existence d’une co-évolution très importante entre E1 et E2. Cette approche a également prédit de potentiel changement de conformations au sein de l’hétérodimère, suggérant que E2 est sans doute une protéine de fusion capable de se replier sur elle-même via le repliement de son domaine III et l’aide de E1. Ainsi, ces différents travaux soulignent l’implication de E1 et E2 au sein de dialogues fins et complexes, qui régulent à la fois les conformations et les fonctions de l’hétérodimère. Ainsi, cela suggère que l’hétérodimère E1E2 représente plutôt une unité fonctionnelle et structurale unique, plutôt que l’association de deux protéines aux fonctions distinctes. / Hepatitis C Virus (HCV) infects more than 170 million people worldwide but no vaccine is available yet. HCV entry may represent a promising target for therapies and is mediated by two envelope glycoproteins, E1 and E2, assembled as heterodimer onto the virus surface. However, how E1 and E2 dialog, structurally rearrange and act together during these steps remain poorly defined. In this work, we aimed to clarify the interrelation of E1E2 during virus entry, thus opening ways to potential new therapeutic strategies. We first investigated whether a strong genetic divergence between E1E2 heterodimers may highlight distinct functions. We observed that B-cell derived E1E2 were specialized for B-cell infection, suggesting that new functions can emerge from the E1E2 conformational plasticity. In a second approach, we identified a conserved dialog between E1 and the domain III of E2 that was critical for virus binding and fusion. Moreover, a computational model predicted a strong co-evolution between E1 and E2 as well as potential structural rearrangements, suggesting that HCV E2 is likely a fusion protein able to fold over via its domain III through the mediation of E1. Altogether, these different works highlight that E1 and E2 are involved in complex dialogs that regulate the heterodimer folding and functions, suggesting that E1E2 heterodimer is more likely a single functional protein entity than an association of two proteins with specific functions.

L'hépatite B chronique reste un problème majeur de santé publique. Sous traitement par analogues nucléos(t)idiques (NUCs), l'objectif thérapeutique ultime est la clairance de l'antigène (Ag) HBs. Nous avons étudié l'influence de la variabilité des protéines d'enveloppe, impliquées dans l'entrée cellulaire du virus et cibles de la réponse immune, sur la clairance de l'Ag HBs. Des patients traités par NUCs ayant obtenu une clairance de l'Ag HBs (resolvers) ont été appariés à des non-resolver. Deux mutations combinées sT125M/sP127T, caractéristiques des non-resolver, étaient associées à une baisse de l'antigénicité prédite. L'analyse par séquençage haut débit montrait une plus grande variabilité du gène S chez les non resolver. Des tests fonctionnels portant sur des particules virales mutées en sT125M et sP127T sont en cours. Ces données moléculaires sont en faveur de l'existence de "motifs" spécifiques dans le gène S associés à la persistance de l'Ag HBs sous traitement par NUCs / Hepatitis B virus (HBV)-related chronic infection remains difficult to eradicate. On treatment by nucleos(t)ide analogues (NUCs), HBs Antigen (Ag) clearance is the ultimate but difficult therapeutic goal. Our aim was to investigate how variability of HBV envelope protein, crucial in viral cellular entry and targeted by host immune response, could play a role in HBsAg clearance. HBV chronically infected patients, treated by NUCs with HBsAg clearance (resolver) were matched with patients without HBsAg clearance (non resolver). Combined mutations sT125M/sP127T, associated with HBsAg persistence, displayed a lower predicted antigenicity. Ultra Deep Sequencing of S gene showed a higher variability in non resolver. Functional assays on viral particles including sT125M and sP127T mutations versus reference particles are in progress. As a conclusion, molecular features observed in non NR argue in favor of a different pattern in HBV S characteristics according to variable NUCs efficiency

Abstract Yellow head virus (YHV) has caused mass mortalities in Penaeus monodon shrimp farmed throughout Southeast Asia since it was first discovered in the early 1990’s. YHV possesses a positive-sense, single-stranded RNA genome and a rod-shaped enveloped virion. Together with the closely related gill-associated virus (GAV) identified in P. monodon shrimp in Australia, it is classified in the genus Okavirus, family Roniviridae within the order Nidovirales. YHV particles contain only three structural proteins, a nucleocapsid (N) protein (p20) protein and two envelope glycoproteins gp116 and gp64. In this study, the glycosylation status of gp116 and gp64 extracted from YHV virions was characterized in detail, including the identification of active N-linked glycosylation sites and the nature of the attached carbohydrates. This was achieved by optimizing and applying a combination of methods that included SDS-PAGE followed by carbohydrate-specific staining of gels or probing of membrane-bound proteins using lectins with different carbohydrate specificities, enzymatic removal of N-linked carbohydrates and a variety of mass spectrometry techniques. In these analyses, it was found that N-linked glycans are the major contributor to the higher estimated mass of gp116 and gp64 by SDS-PAGE compared to those estimated from their deduced amino acid sequences. Neither gp116 nor gp64 were found to posses O-linked glycans. Mannose residues were identified to be the major glycan component of carbohydrates linked to gp116 and gp64 and are possibly the sole component of carbohydrate linked to gp64. Unlike gp64, other glycans such as terminal N-acetyl--D-galactosamine and N-acetyl--D-glucosamine were identified to be attached to gp116. Assuming that glycosylation processes in shrimp mimic those of vertebrates that are known in more detail, the nature of the glycans attached to gp116 suggests that they might be added and modified during the transportation of the protein from the endoplasmic reticulum (ER) to the trans-Golgi network (TGN). Mass spectrometry analyses of tryptic peptides derived from the native glycoproteins and following their enzymatic deglycosylation, generated approximately 81% (gp116) and 66% (gp64) coverage of their predicted amino acid sequences. Detailed mass spectrometry analyses of peptides derived from the deglycosylated proteins identified that most of the potential N-linked glycosylated site in the virion envelope glycoproteins, 6 of 7 present in gp116 and 3 of 4 present in gp64 were identified to be modified by glycans. In gp116, one site was not identified and in gp64 one site was not utilized. As phosphorylation has been shown to affect nucleocapsid protein (N) functioning in vertebrate nidoviruses, SDS-PAGE using two phosphoprotein-specific staining methods, as well as mass spectrometry methods, were employed to examine whether the YHV N protein present in virions is phosphorylated. The protein staining methods provided contradicting results and no phosphate-containing peptides were identified by mass spectrometry. The apparent absence of phosphate in the N protein was also supported by its isoelectric point (pI ~10) determined by isoelectric focusing and two-dimensional electrophoresis (2-DE) analysis, which was very similar to that predicted (pI = 9.98) from its deduced amino acid sequence. Taken together, the data suggest that the YHV N protein encapsulated within virions is not phosphorylated. The RNA-binding capability of the GAV N protein was assessed using an electrophoretic mobility shift assay (EMSA) technique. Full-length and variously truncated forms of the GAV N protein expressed in bacteria were assessed in the assays. It was found that the full-length recombinant N protein bound to RNA in a sequence non-specific manner. Analysis of the five truncated N protein constructs localized the RNA-binding domain to a 50 amino acid sequence in the N-terminal region residing between Met11 and Arg60. A motif rich in proline and arginine residues, which are commonly found in other RNA-binding proteins, occurred in first 18 amino acids of this region. Although RNA-binding was not sequence-specific, the data suggest that this region of the GAV N protein is the most likely site at which it interacts with and nucleates viral genomic RNA during nucleocapsid formation. A synthetic peptide spanning the 18 amino acid of the putative RNA-binding domain was shown to possess RNA-binding properties similar to the recombinant protein fragment. These results indicated that the 18 amino acid, proline and arginine rich motif (MPVRRPLPPQPPRNARLI) in the N-terminal region of the GAV N protein confers its RNA-binding function. Using an immuno-co-precipitation assay, a host protein was found to interact abundantly with the GAV N protein in infected lymphoid organ cells. Mass spectrometry analysis identified the protein as -actin. Immuno-histochemistical double-labeling methods in conjunction with observations made using confocal and electron microscopy revealed that actin and the N protein were co-located in cytoplasm of infected cells. Electron microscopy suggested that interaction of the two proteins occurs before nucleocapsid envelopment within virions, suggesting that -actin might be involved in transporting the N protein or the nucleocapsid from their sites of synthesis to the rough endoplasmic reticulum where the virion acquires its envelopes. In summary, the research described in this thesis has advanced understanding of the YHV/GAV proteome through the identification of the glycosylation sites in the envelope glycoproteins gp116 and gp64, and demonstrating that nucleocapsid protein encapsulated within virion is unlikely to be phosphorylated. Functional studies have also shown that the nucleocapsid protein binds RNA non-specifically through an 18 amino acid domain near its N-terminus and that it binds and co-localizes with -actin in infected cells, suggesting that -actin may play role in trafficking N protein in infected cells.

The envelope glycoprotein (Env) of Human Immunodeficiency Virus 1 (HIV-1) has developed several immune-evasion mechanisms to avoid the induction of neutralising antibodies, including immunodominant non-neutralising epitopes, conformational flexibility of conserved epitopes, and spontaneous subunit dissociation, thus impeding vaccine development. Here, chemical modification of Env-based vaccine antigens is explored to overcome these obstacles. Firstly, covalent fixation of Env by chemical cross-linking was used to stabilise the conformationally flexible structure and prevent subunit dissociation. Cross-linked Env constructs showed reduced binding of many non-neutralising antibodies whilst largely maintaining antibody recognition by broadly neutralising antibodies. Compared to unmodified material, immunisation with some of these cross-linked proteins led to the induction of significantly increased antibody titres targeting the conserved CD4 binding site of Env despite similar overall antibody titres. These refocused antibody responses resulted in increased serum neutralising titres compared to animals receiving unmodified protein. Secondly, an epitope masking strategy was developed to reduce or eliminate the immunogenicity of neutralisation-irrelevant surfaces. This was achieved using site-selective addition of theoretically immunosilent glycoconjugates to lysine residues. Masking of model protein hen egg lysozyme (HEL) led to site-selective loss of antibody binding to the modification sites in vitro, which translated into refocusing of antibody responses from masked to unmasked epitopes in vivo. Mutant HIV-1 and influenza virus surface glycoproteins were designed that had lysine residues removed from close proximity to the respective broadly neutralising epitopes, but added throughout the remaining surface. Masking of these mutant proteins with second-generation glycoconjugates led to predictable perturbations of antibody binding in vitro. However, administration of these modified glycoproteins revealed unexpectedly that the masking glycans were highly immunogenic in vivo. Thus, this strategy may well prove useful if truly non-immunogenic glycoconjugates can be identified. Taken together, these chemical modifications of vaccine antigens may allow focused targeting of specific antigenic regions for increased B cell recognition, and may thus be a valuable tool for vaccine antigen design.