Thesis advisor: Welkin Johnson / Endogenous Retroviruses (ERVs) are “fossilized” retroviruses of a once exogenous retrovirus located in the genome of extant vertebrates. Retroviral infection results in a provirus integration into the host genome. An infection of a germline cell could lead to the provirus potentially being inherited by the offspring of the infected individual. Once in the genome, the provirus becomes subject to evolutionary processes and can become either lost or fixed in a population, remaining as “fossils” long after the exogenous retrovirus has gone extinct23. Notably, 8% of the human genome consists of ERVs30. Human Endogenous Retrovirus Type K (HERV-K)(HML-2) family is of particular interest. HERV-K integrations are as old as 30-35 million years, endogenizing before the separation of humans and Old World Monkeys. However, there are human specific insertions, some as young as 150,000 – 250,000 years, making them the youngest insertion in the human genome. There are over 90 insertions in the human genome; the bulk is shared by all humans44,47. Transcripts of HERV-K genes are upregulated in multiple cancer and tumor cell lines 14,39,46, as well as in HIV-1 infected patients 7,11,29. Just as there are human specific insertions of ERV-K, there are also Old World Monkey specific insertions44. I have identified an intact endogenous retroviral envelope open reading frame on chromosome 12 of the rhesus macaque genome. This viral envelope-encoding sequence, which I refer to as rhERV-K env, retains all the canonical features of a retroviral Env protein. An alignment between rhERV-K env and a consensus sequence of HERV-K, HERV-Kcon env, shows a 70% amino acid sequence identity. For experimental purposes, reconstructed HERV-K envelopes have been incorporated into virions of Human Immunodeficiency virus (HIV-1)19,26,49, Murine Leukemia Virus (MLV)12, and Vesicular stomatitis Virus (VSV)26,41,49. While these approaches have illuminated some aspects of HERV-K Env-mediated entry, to date a cell-surface receptor has not been identified for any ERV-K Env. This could be due to its low infectivity levels12,26,49, its seemingly broad cell tropism limiting identification of null cell lines26,49, or possibly the HERV-K consensus reconstructions are not an accurate representation of the progenitor HERV-K virus. I am interested in understanding how the ERV-K retrovirus accessed the human germline (some 150,000 – 250,000 years ago). To do this, I focused specifically on the envelope proteins of HERV-K and rhERV-K, with the goal of analyzing the ERV-K entry process. The identification and inclusion of rhERV-K Env in this study is meant to circumvent the possibility that the previously described consensus reconstructions of human HERV-K Env are not representative, and may also provide a means to compare the endogenization process in the human/ape and old-world monkey lineages. I focused on developing two systems for single-cycle infection, one based on Mason-Pfizer Monkey Virus (MPMV) (which has not been done before), and a second based on MLV, which has previously been reported on. MPMV, like HERV-K, is a betaretrovirus, and I reasoned that possibly using a betaretrovirus would overcome some of the low-infectivity issues associated with prior attempts using HIV and MLV. To develop a system for examining function of the ERV-K Env proteins, I addressed 3 issues: 1. Are the HERV-K Env and rhERV-K Env proteins expressed and properly processed? 2. Can they be incorporated into virions of a heterologous virus? 3. Are ERV-K pseudotyped virions infectious? I have answered these questions in the following thesis. / Thesis (MS) — Boston College, 2017. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

The triggers for the development of multiple sclerosis (MS) have not been fully understood to date. One hypothesis proposes a viral etiology. Interestingly, viral proteins from human endogenous retroviruses (HERVs) may play a role in the pathogenesis of MS. Allelic variants of the HERV-K18 env gene represent a genetic risk factor for MS, and the envelope protein is considered to be an Epstein–Barr virus-trans-activated superantigen. To further specify a possible role for HERV-K18 in MS, the present study examined the immunogenicity of the purified surface unit (SU). HERV-K18(SU) induced envelope-specific plasma IgG in immunized mice and triggered proliferation of T cells isolated from these mice. It did not trigger phenotypic changes in a mouse model of experimental autoimmune encephalomyelitis. Further studies are needed to investigate the underlying mechanisms of HERV-K18 interaction with immune system regulators in more detail.

L’herpèsvirus humain (HHV) 6 est un betaherpèsvirus largement répandu, associé à plusieurs maladies neuroinflammatoires, telles que des encéphalites ou la sclérose en plaques (SEP). Cependant, les mécanismes impliqués dans la neuropathologie induite par les deux espèces d’HHV-6, HHV-6A et HHV-B, sont peu connus. De plus, l’absence de modèle d’infection chez le petit animal a ralenti l’étude de la pathogénèse virale. Dans ce contexte, nous avons développé un modèle d’infection par HHV-6 chez des souris transgéniques, qui expriment la protéine CD46 humaine, identifiée comme récepteur cellulaire pour HHV-6. Nous avons pu démontrer une persistance de l’ADN viral d’HHV-6A, mais pas d’HHV-6B, dans le cerveau de souris transgéniques pendant plusieurs mois. De plus nos résultats montrent qu’HHV-6A induit la sécrétion de chimiokines pro-inflammatoires par les cellules neurales murines et provoque l’infiltration de cellules immunitaires dans le cerveau de souris infectées. Enfin, HHV-6A, mais pas HHV-6B, pourrait induire des réponses cellulaires chez les cellules murines via le récepteur de l’immunité innée TLR9 (toll-like receptor 9). En collaboration avec une équipe de Grenoble, nous avons ensuite montré que l’infection par HHV-6A induit l’expression de rétrovirus endogènes humains (HERV) dans des cellules mononuclées et des lignées neurales humaines. Ces HERV, en particulier leurs protéines d’enveloppe qui présentent des propriétés pro-inflammatoires, sont associés à diverses maladies autoimmunes dont la SEP. HHV-6A pourrait donc participer au développement de pathologies inflammatoires via l’induction de ces HERV. L’ensemble de ces travaux supporte ainsi l’existence d’un lien entre l’infection par HHV-6A et la neuroinflammation, et apporte de nouvelles pistes quant aux mécanismes potentiellement impliqués. / Human herpesvirus (HHV) 6 is a widely spread betaherpesvirus, which has been associated to several neuroinflammatory diseases, such as encephalitis or multiple sclerosis (MS). However, the mechanisms explaining the neuropathology induced by the two species of HHV-6, HHV-6A and HHV-6B, remain to be elucidated. Moreover, the lack of small animal model for HHV-6 infection has considerably hampered the study of viral pathogenesis. In this context, we have generated several lines of mice expressing the human CD46 protein, identified as a cellular receptor for HHV-6, and characterized the infection. We demonstrated that DNA of HHV-6A, but not HHV-6B, can persist in the brain of CD46 transgenic mice for several months after intracranial injection. Moreover our results show that HHV-6A induces chemokine secretion by in vitro cultured murine brain cells and provokes leucocyte infiltration in the brain of infected mice. Finally, HHV-6A, but not HHV-6B, could activate cellular responses in murine cells through binding to toll-like receptor 9. In collaboration with the team of P. Marche in Grenoble, we then showed that HHV-6A and HHV-6B infection induce the expression of envelope genes from human endogenous retrovirus W (HERV-W) in human blood mononucleated cells and human neural cell lines. Envelope proteins of HERV-W are known to exhibit strong pro-inflammatory properties and were associated to various autoimmune diseases, including multiple sclerosis. HHV-6A and HHV-6B could therefore participate in the development of inflammatory disorders via the activation of these HERV genes. Altogether this work supports the hypothesis of a link between HHV-6 infection neuroinflammation and opens new perspectives in the study of the mechanisms potentially involved.

Le sepsis est défini par l’apparition de dysfonctions d’organes, multiples et mortelles, causées par une réponse de l’hôte dérégulée suite à une infection. L’hétérogénéité de la maladie représente un défi clinique majeur au regard de la prise en charge thérapeutique, et à ce jour les marqueurs proposés ne suffisent pas à stratifier les patients. Les rétrovirus endogènes humains (HERV) pourraient être des marqueurs pertinents,compte tenu des propriétés immunosuppressives de leurs enveloppes et de leur expression dans des maladies inflammatoires et auto-immunes. Cette thèse a pour objectif de savoir dans quelle mesure les HERV sont exprimés et modulés, dans des conditions d’inflammation et d’immunosuppression. Pour cela,nous avons utilisé une puce à ADN haute densité permettant (i) l’analyse de la transcription de 363 689HERV et 1500 gènes, et (ii) une lecture fonctionnelle de l’activité des LTR. L’expression des HERV a été objectivée (i) dans un modèle ex-vivo de tolérance à l’endotoxine sur des cellules mononuclées du sang périphérique (PBMC) d’individus sains et (ii) sur sang total provenant d’individus sains et de patients en choc septique, stratifiés ou non en fonction du statut immunitaire. (1) De 5,6% à 6,9% des HERV sont exprimés dans le compartiment sanguin et environ 20% des LTR possèdent une fonction promotrice ou polyA, les deux fonctions étant mutuellement exclusives. (2) Le contenu du transcriptome HERV est modulé ex vivo dans le contexte de tolérance à l’endotoxine laissant apparaitre deux grands phénotypes transcriptionnels. L’expression de certains loci HERV est corrélée au statut immunitaire de patient septique.L’évaluation d’une signature moléculaire complexe sur une cohorte de validation, permet la séparation en deux groupes présentant des critères de sévérité distincts, suggérant les HERV/MaLR comme biomarqueurs de stratification. (3) L’analyse de la co-expression des gènes et des HERV a permis d’intégrer ceux-ci au sein de réseaux associées à la réponse de l’hôte et de proposer des hypothèses fonctionnelles. / Sepsis is defined as a life-threatening organ dysfunction caused by a dysregulated host response to infection.The heterogeneity of the disease present a major clinical challenge with regard to the therapeutic coverage,and this day the proposed markers are not enough to stratify patients. The human endogenous retrovirus(HERV) could be relevant markers, considering the immunosuppressives properties of their envelopes andtheir expression in inflammatory and autoimmune disease. The aim of this thesis is to know to what extentthe HERVs are expressed and modulated, in inflammatory and immunocompromised contexts. For this, weused a high density DNA chip allowing (i) the transcription analysis of 363,689 HERV and 1500 genes,and (ii) a functional reading of LTRs activities. The HERVs expression was objectified (i) in endotoxintolerance ex vivo model in peripheral blood mononuclear cells (PBMCs) of healthy volunteers and (ii) inwhole blood of healthy volunteers and septic shock patients, stratified or not according to immunity state.(1) Of 5,6% at 6,9% of HERVs are expressed in the blood compartment and around 20% of LTRs have apromoter or polyA function, both functions being mutually exclusive. (2) The HERV transcriptome ismodulated in ex vivo endotoxin tolerance model letting appear two higher transcriptional phenotypes. Theexpression of some HERVs loci are correlated of the immunity state of the septic shock patients. Theevaluation of molecular signature in validation cohort, allowed to separate in two patients groupspresenting different severity criteria, suggesting HERV/MaLR as biomarkers of stratification. (3) The coexpressedanalysis of genes and HERVs allowed to integrate these within signaling pathways associated atthe host immune response and to provide functional hypothesis.

Human endogenous retroviruses (HERVs) constitute about 8% of the human genome. There are around 4000 pol-containing retroviral integrations in the human genome, which makes it impractical to measure each of them separately. Therefore we developed a set of degenerate real time PCRs to detect major groups bearing sequence similarities to gammaretroviruses, one of the largest groups of human endogenous retrovirus, and betaretroviruses, some of which have integrated into the human genome most recently and which remain the most intact. It was found that, although both gammaretroviral and betaretroviral RNAs were broadly expressed in various healthy tissues including reproductive tissues and brain, a differential expression pattern was observed. My work further revealed that HERVE and HERVW, two gammaretroviral sequences, were ubiquitously and highly expressed in pathologic and normal female reproductive tissues with tissue specific patterns. Expression of HERVE was higher in endometriotic tissue than in normal endometrium. HERVE and HERVW RNAs were higher in normal ovarian tissue than in ovarian cancer. Besides these tissue- and neoplasia-related differences, there were wide differences in HERV expression among individuals. Next, a selective pattern of HERVW upregulation was demonstrated in SK-N-DZ, a neuroblastoma cell line, upon re-oxygenation after a period of hypoxia or with 5-azacytidine, a demethylating agent. Furthermore, broad and high expressions of gammaretrovirus-like transcripts in different brain areas analyzed were identified. The expression levels were variable among different donors. In conclusion a ubiquitous HERV expression was observed in tissues and cell lines, with various patterns. At this stage the data are not sufficient to conclude whether HERV has any physiological or pathological roles in humans. However, their differential expression patterns are compatible with functional roles of HERV in humans.

Einleitung: Der seit über 30 Jahren ausgiebig erforschte Transkriptionsfaktor p53 besitzt offenbar Funktionen, die über seine bekannte und gut untersuchte Aktivität als Tumorsuppressor hinausgehen. So scheint er auch an der Regulation der menschlichen Lebenserwartung - über die Vermittlung einer allgemeinen physischen Robustheit - sowie der weiblichen Fertilität beteiligt zu sein. Insbesondere Primaten zeichnen sich durch eine vergleichsweise lange Lebenserwartung und eine lange reproduktive Phase aus. Ob p53 hier eine Rolle spielen könnte, ist unbekannt. Unsere Arbeitsgruppe entdeckte vor einigen Jahren das humane endogene Retrovirus-K (HERV-K) Protein Np9, dessen Gen sich in mehreren Kopien nur bei Menschen, Schimpansen und Gorillas findet. Weitere Untersuchungen wiesen außerdem darauf hin, dass Np9 an den Tumorsuppressor p53 zu binden vermag. Ziele der Untersuchungen: Es stellte sich also die Frage, ob die Funktion von p53 durch die Bindung an Np9 moduliert werden kann. Eine derartige Modulation des multifunktionellen Transkriptionsfaktors wäre natürlich auf Hominiden beschränkt. In der vorliegenden Arbeit sollten einige Teilaspekte der Interaktion von p53 und Np9 näher untersucht werden. Material und Methoden: Für die Bindungskartierung von p53 und Np9 wurden GST-Pulldown-Analysen durchgeführt. Die GST-Protein-Plasmide wurden in E.coli BL21 transformiert und nach Induktion mit IPTG exprimiert. Sie dienten als „Fängerproteine“ und waren dank ihres Glutathion-S-Transferase-tags in der Lage an GST-Sepharose-Kügelchen zu binden. Der putative Interaktionspartner als „Beuteprotein“ wurde in vitro translatiert und in diesem Zuge auch mit 35S radioaktiv markiert. Dann wurde er mit den an die Beads gebundenen GST-Proteinen inkubiert und anschließend die Proben auf ein SDS-Gel aufgetragen und aufgetrennt. Das Gel wurde anschließend auf eine Membran übertragen und der Blot auf einen Radioaktivfilm aufgelegt, woraufhin die Protein-Protein-Bindungen anhand des radioaktiven Beuteproteins als Banden erkennbar waren. Abschließend wurde der Blot mit GST-Antikörper inkubiert, dann am Folgetag mit Anti-Mouse-Antikörper. Mittels ECL Substrat konnte nun die Bindung der GST-getaggten Proteine an die Sepharosebeads nachgewiesen werden. Für den Electrophoretic Mobility Shift Assay wurden verschiedene Versuchsansätze pipettiert, welchen nach einer Inkubationszeit das zuvor mit 32P radioaktiv markierte Oligonukleotid zugegeben wurde. Nach erneuter Inkubation wurden die Proben auf das nicht-denaturiende EMSA-Gel aufgetragen und elektrophoretisch aufgetrennt. Dabei wurden die Protein-Oligonukleotid-Verbindungen gemäß ihrer Ladung, Größe und Konformation getrennt. Die Gele wurden im Geltrockner getrocknet und direkt mit einer Verstärkerfolie auf den Radioaktivfilm in einer Radioaktivkassette aufgelegt. Ergebnisse: Zunächst war es notwendig, die Bindung der beiden Partner biochemisch zu kartieren. Dies geschah mittels der GST-Pulldown-Analyse, in der Fragmente der Proteine exprimiert, miteinander inkubiert und schließlich kopräzipitiert wurden. Es stellte sich heraus, dass p53 mit seinem C-Terminus an Np9 bindet. Np9 hingegen band mit seinen Aminosäureresten (aa) 1-64 (ohne den C-terminus mit den aa 65-74) an p53. Das Np9-Fragment 36-74 zeigte nur eine schwache Bindung an p53. Interessanterweise band das Np9-Fragment 36-64 stärker an p53 als Volllängen-Np9 (1-74), was auf eine die Interaktion hemmende Domäne im C-Terminus von Np9 hinweisen könnte. Um zu untersuchen, ob die Bindung von Np9 an den C-Terminus von p53 die p53-DNA-Interaktion beeinflusst, wurden Electrophoretic Mobility Shift Assays (EMSAs) durchgeführt. Es konnte gezeigt werden, dass Np9-zumindest in vitro-durch Bindung an die regulatorische Domäne von p53 und in Anwesenheit des p53-aktivierenden Antikörpers PAb421 in der Lage war, die spezifische Bindungsfähigkeit von p53 an DNA zu erhöhen und somit seine Funktion als Transkriptionsfaktor zu unterstützen. Schlussfolgerungen: Die Resultate weisen also erstmals darauf hin, dass das nukleäre HERV-K Protein Np9 spezifisch in Hominiden eine p53-abhängige Tumorsuppressoreigenschaft aufweisen könnte. Weitere Untersuchungen-insbesondere in vivo-sind nun notwendig. Dies könnte auch als Forschungsgrundlage zu endogenen Retrovirusproteinen beim Pferd dienen.:1 Einleitung und Zielsetzung der Arbeit 1 2 Literaturübersicht 2 2.1 Der Tumorsuppressor p53 2 2.2 Das Kernprotein Np9 7 2.2.1 Retroviren 7 2.2.2 Endogene Retroviren 8 2.2.3 Humane endogene Retroviren 8 2.2.4 HERV-K 10 2.2.5 Das nukleäre Protein Np9 10 3 Material und Methoden 15 3.1 Material 15 3.1.1 Chemikalien 15 3.1.2 Puffer und Lösungen 17 3.1.3 Antikörper 21 3.1.4 Enzyme 22 3.1.5 Reaktionskits 22 3.1.6 Bakterienstämme 23 3.1.7 Kulturmedien 23 3.1.8 Oligonukleotide für EMSA 23 3.1.9 Größenstandards 24 3.1.10 Plasmide 26 3.2 Methoden 28 3.2.1 Nukleinsäuretechniken 28 3.2.2 Protein-Methoden 31 3.2.3 Prokaryonten 40 4 Ergebnisse 42 4.1 Interaktion zwischen Np9 und dem Tumorsuppressorprotein p53 42 4.2 GST-Pulldown 43 4.2.1 Klonierung für die GST-Pulldown-Analysen 43 4.2.2 Induktion der Proteinexpression 48 4.2.3 GST-Pulldown-Experimente 53 4.3 EMSA (Electrophoretic Mobility Shift Assay) 57 4.3.1 Radioaktive Markierung der Sonden 58 4.3.2 EMSA-Experimente 58 5 Diskussion 63 6 Zusammenfassung 68 7 Summary 70 8 Literaturverzeichnis 72 Danksagung 86 Abbildungsverzeichnis 87 Tabellenverzeichnis 88 / Introduction: The transcription factor p53, extensively investigated for over 30 years, apparently has functions which exceeds his known and well examined activity as a tumor suppressor. It seems to be involved in the regulation of the human life expectancy – by providing a general physical robustness - as well as of the female fecundity. Primates too are characterized by a comparatively long life expectancy and long reproductive phases, yet the possible influence of p53 is unknown. Our research group has discovered some years ago the Np9 protein of human endogenous retrovirus K (HERV K), which is found in several copies only with humans, chimpanzees and gorillas. Other investigations by our group suggested that Np9 might be able to interact with the tumor suppressor p53. Objective of the investigations: To study whether the function of p53 can be modulated by the interaction with Np9. Such a modulation of the multifunctional transcription factor would of course be limited to hominids. In the present work some aspects of the interaction between p53 and Np9 were analysed. Materials and methods: For the mapping of the interaction of p53 and Np9, GST pulldown assays were carried out. The GST protein plasmids were transformed in E. coli BL21 and expressed after IPTG induction. They served as bait proteins and bound to GST sepharose beads because of their Glutathione S-transferase-tags. The putative interaction partner as a prey protein was translated in vitro and radioactively marked with 35S. After being incubated with the GST-proteins bound to the beads, the samples were transferred on a SDS gel and separated. The gel was transferred to a membrane and the blot was exposed to an X-ray film. Thus, the radioactively labelled prey protein forms bands that identify the protein-protein interaction. Finally the blot was incubated with GST antibody, then on the following day with anti-mouse antibody. Using ECL-substrate it was now possible to demonstrate that the GST-tagged proteins bound to the sepharose beads. For the Electrophoretic Mobility Shift Assay different samples were prepared and, after an incubation time, the oligonucleotide radioactively marked with 32P was added. After additional incubation it was transferred on non-denaturating EMSA gel and separated by electrophoresis. Thus the protein oligonucleotide conjugates were separated according to charge, size and conformation. The gels were dried in the gel dryer, transferred to a membrane and placed against an X-ray film in a cassette. Results: Initially a biochemical mapping of the binding of the two partners had to be carried out. This was done by means of the GST pulldown assay, in which fragments of the proteins were extruded, incubated together and finally co-precipitated. It turned out that the C-terminus of p53 bound to Np9. However, Np9 bound to p53 with his amino acid residues (aa) 1-64 (lacking the C-terminal aa 65-74). The Np9 fragment 36-74 showed only a weak binding to p53. Interestingly the Np9 fragment 36-64 was binding stronger to p53 than a full length Np9 (1-74), which could point to a C-terminal domain in Np 9 inhibiting the interaction. In order to examine whether the binding of Np9 to the C-terminal of p53 affects the interaction of p53 with DNA, Electrophoretic Mobility Shift Assays (EMSAs) were carried out. It could be shown that Np9 was able to raise the specific binding ability of p53 with DNA and to support therefore its function as a transcription factor, by binding to the regulatory domain of p53 in presence of the activating p53 antibody PAB421. Conclusions: The results show for the first time that, specifically in hominids, the nuclear HERV-K protein Np9 could have a tumor suppressing quality that is dependent on p53. Further investigations, in particular in vivo, are necessary. This could be the starting point for research on equine endogenous retrovirusproteins in horses.:1 Einleitung und Zielsetzung der Arbeit 1 2 Literaturübersicht 2 2.1 Der Tumorsuppressor p53 2 2.2 Das Kernprotein Np9 7 2.2.1 Retroviren 7 2.2.2 Endogene Retroviren 8 2.2.3 Humane endogene Retroviren 8 2.2.4 HERV-K 10 2.2.5 Das nukleäre Protein Np9 10 3 Material und Methoden 15 3.1 Material 15 3.1.1 Chemikalien 15 3.1.2 Puffer und Lösungen 17 3.1.3 Antikörper 21 3.1.4 Enzyme 22 3.1.5 Reaktionskits 22 3.1.6 Bakterienstämme 23 3.1.7 Kulturmedien 23 3.1.8 Oligonukleotide für EMSA 23 3.1.9 Größenstandards 24 3.1.10 Plasmide 26 3.2 Methoden 28 3.2.1 Nukleinsäuretechniken 28 3.2.2 Protein-Methoden 31 3.2.3 Prokaryonten 40 4 Ergebnisse 42 4.1 Interaktion zwischen Np9 und dem Tumorsuppressorprotein p53 42 4.2 GST-Pulldown 43 4.2.1 Klonierung für die GST-Pulldown-Analysen 43 4.2.2 Induktion der Proteinexpression 48 4.2.3 GST-Pulldown-Experimente 53 4.3 EMSA (Electrophoretic Mobility Shift Assay) 57 4.3.1 Radioaktive Markierung der Sonden 58 4.3.2 EMSA-Experimente 58 5 Diskussion 63 6 Zusammenfassung 68 7 Summary 70 8 Literaturverzeichnis 72 Danksagung 86 Abbildungsverzeichnis 87 Tabellenverzeichnis 88