Aicardi-Goutières Syndrome (AGS) is a single gene, autoimmune disorder, with variable onset in the first year of life. Its clinical features exhibit similarities to several autoimmune diseases and congenital viral infections. AGS can result from mutations in ADAR1, TREX1 and SAMHD1 as well as any of the three genes that encode the protein subunits of the RNase H2 enzyme. It is hypothesised that impairment of nucleic acid metabolism results in abnormal nucleic acid species within the cell. This in turn is thought to cause the aberrant immune response that leads to AGS. The RNase H2 complex contains the catalytic RNASEH2A subunit and the auxiliary RNASEH2B and RNASEH2C subunits, which are thought to provide structural support and facilitate interactions with additional cellular proteins. RNase H2 can cleave the RNA strand of an RNA:DNA hybrid as well as 5’ of a single ribonucleotide embedded in dsDNA. Therefore, RNase H2 may have roles in several cellular processes, including DNA replication and repair, transcription, and viral infection. The aim of this PhD project was to investigate the physiological functions of RNase H2. The localisation of the RNase H2 proteins was investigated using EGFP-tagging and fluorescent microscopy. The interaction between the PIP-box of RNASEH2B and PCNA was found to localise RNase H2 and not RNase H1 to nuclear replication foci during S-phase. This suggests that RNase H2 is the dominant RNase H activity during DNA replication. Stable cell lines expressing EGFP-RNASEH2B and an alternative isoform, EGFP-RNASEH2Balt, were generated and used to perform a protein-protein interaction screen by GFP-Trap and mass spectrometry. The results indicate putative physical interactions between RNASEH2B and other factors involved in DNA replication and repair. Further evidence for a role in DNA repair was revealed when mammalian RNase H2 null cells were treated with hydroxyurea. Low doses of hydroxyurea increased ribonucleotide incorporation into genomic DNA and impaired S-phase progression. In contrast to wild-type cells, RNase H2 null cell proliferation also failed to recover from this replicative stress after HU withdrawal. However, the ribonucleotide content of genomic DNA from these cells did return to pre-hydroxyurea treatment levels. This suggests that an alternative repair pathway exists in mammalian cells, which can remove ribonucleotides from DNA in the absence of RNase H2, but that this pathway is also harmful to the cells. There is evidence that TREX1 facilitates viral infection while SAMHD1 has been shown to restrict viral infection. Therefore, experiments were performed to investigate if RNase H2 could be a viral facilitator or restriction factor. Ribonucleotides can be incorporated into viral DNA, so RNase H2 could act as a restriction factor by nicking and damaging the pre-integration complex. However, RNase H2 could also function as a facilitator of infection by processing viral RNA:DNA hybrid by-products and thus prevent the host immune response. The data obtained during this PhD project provides further evidence that RNase H2 is involved in DNA replication and repair and has contributed to the understanding of the function of RNase H2 in the cell. However, it is still unknown how mutations in RNase H2 lead to the pathology of AGS.

Aicardi-Goutières syndrome (AGS) is an autosomal recessive encephalopathy with low incidence. The disease is caused by mutations in the genes encoding for TREX1, SAMHD1, ADAR, IFIH1 and the three genes encoding for the heterotrimeric RNase H2 enzyme. Biallelic mutations in any of the genes cause elevated type I interferon levels in the cerebrospinal fluid (CSF), the hallmark of AGS. In AGS patients, increased type I interferon levels cause massive inflammation in the brain that leads to mental and physical retardation that likely cause death in early childhood. AGS shows significant overlap with the prototypic autoimmune disease systemic lupus erythematosus (SLE). Like AGS patients, SLE patients are also characterized by increased type I interferon levels, anti-nuclear autoantibodies (ANAs) and arthritis. Moreover, heterozygous mutations in TREX1, SAMHD1 and RNase H2 are also found in a small fraction of SLE patients. Due to the genetic, molecular and clinical overlap, AGS is regarded as a monogenic variant of SLE. This overlap allows for the investigation of SLE pathomechanisms using genetically engineered mouse models with AGS-related mutations. In order to generate a mouse model that allows for the identification of pathomechanisms in AGS patients with mutations in the genes encoding for the RNase H2 enzyme, we generated mice with deficiency for the RNase H2 enzyme. Mice with complete deficiency for the RNase H2 enzyme (Rnaseh2c-/- or Rnaseh2bKOF/KOF) died perinatally or were stillborn. Mouse embryonic fibroblasts (MEFs) from E14.5 Rnaseh2bKOF/KOF embryos displayed impaired proliferation that was caused by the accumulation of MEF cells in G2/M of the cell cycle which increased with cultivation time or if MEF cells were isolated from E18.5 Rnaseh2bKOF/KOF embryos. Gene expression analysis of E14.5 fetal liver cells revealed a robust p53-mediated DNA damage response with the cell cycle inhibitor cyclin- dependent kinase inhibitor 1a (Cdkn1a, p21) being the most up-regulated gene. We found increased numbers of phosphorylated histone H2AX (γH2AX) in fetal liver and thymus cells from E18.5 Rnaseh2bKOF/KOF embryos, indicative of DNA double-strand breaks. Finally, we could show increased ribonucleotide loads in genomic DNA from embryos that were completely deficient for the RNase H2 enzyme. Collectively, we have demonstrated that complete RNase H2 deficiency causes persistent genomic ribonucleotide loads that render the DNA instable and prone to DNA strand breaks. DNA damage leads to the activation of p53 that in turn activates the cell cycle inhibitor p21 that inhibits cell cycle progression and likely causes accumulation of RNase H2-deficient cells in G2/M. To bypass early lethality we also generated bone marrow chimera and cell type-specific knockouts of the Rnaseh2b gene. While fetal liver cells of E14.5 Rnaseh2bKOF/KOF embryos could maintain hematopoiesis of irradiated recipient mice for almost one year, bone marrow cells from these primary recipients failed to reconstitute lethally irradiated mice in a secondary transfer. In line with this observation, VavCre-mediated deletion of the Rnaseh2b gene caused a more than hundred fold reduction of peripheral blood B cells, while B cell numbers remained unaltered upon CD19Cre-mediated deletion that occurs much later in B cell development. These data suggested that RNase H2 deficiency leads to the accumulation of genomic ribonucleotides that might cause the accumulation of a so far uncharacterized DNA damage species with increasing cell cycle passages. The data further supported our hypothesis that the impact of RNase H2 deficiency is determined by the number of cell proliferation. Finally, an epidermis-specific knockout of the Rnaseh2b gene displayed the most dramatic phenotype. Knockout mice were characterized by hyperpigmentation, hair loss and spontaneous ulcerations of the skin. Microscopically, these mice displayed moderate thickening of the epidermis and dermal fibrosis as indicated by increased collagen deposition. Macroscopic skin phenotypes were completely dependent on p53 expression, since concomitant deletion of the p53 gene rescued mice from hyperpigmentation, hair loss and ulcerations. This data demonstrated that Rnaseh2b deficiency in the epidermis may also lead to DNA damage and subsequent p53 activation as shown for fetal liver from E14.5 RNase H2-deficient embryos. Preliminary data also indicate an increased incidence of cancer formation in RNase H2/p53 double knockouts, identifying the RNase H2 enzyme as an important tumor suppressor.

Aicardi-Goutières syndrome (AGS) is an autosomal recessive encephalopathy with low incidence. The disease is caused by mutations in the genes encoding for TREX1, SAMHD1, ADAR, IFIH1 and the three genes encoding for the heterotrimeric RNase H2 enzyme. Biallelic mutations in any of the genes cause elevated type I interferon levels in the cerebrospinal fluid (CSF), the hallmark of AGS. In AGS patients, increased type I interferon levels cause massive inflammation in the brain that leads to mental and physical retardation that likely cause death in early childhood. AGS shows significant overlap with the prototypic autoimmune disease systemic lupus erythematosus (SLE). Like AGS patients, SLE patients are also characterized by increased type I interferon levels, anti-nuclear autoantibodies (ANAs) and arthritis. Moreover, heterozygous mutations in TREX1, SAMHD1 and RNase H2 are also found in a small fraction of SLE patients. Due to the genetic, molecular and clinical overlap, AGS is regarded as a monogenic variant of SLE. This overlap allows for the investigation of SLE pathomechanisms using genetically engineered mouse models with AGS-related mutations. In order to generate a mouse model that allows for the identification of pathomechanisms in AGS patients with mutations in the genes encoding for the RNase H2 enzyme, we generated mice with deficiency for the RNase H2 enzyme. Mice with complete deficiency for the RNase H2 enzyme (Rnaseh2c-/- or Rnaseh2bKOF/KOF) died perinatally or were stillborn. Mouse embryonic fibroblasts (MEFs) from E14.5 Rnaseh2bKOF/KOF embryos displayed impaired proliferation that was caused by the accumulation of MEF cells in G2/M of the cell cycle which increased with cultivation time or if MEF cells were isolated from E18.5 Rnaseh2bKOF/KOF embryos. Gene expression analysis of E14.5 fetal liver cells revealed a robust p53-mediated DNA damage response with the cell cycle inhibitor cyclin- dependent kinase inhibitor 1a (Cdkn1a, p21) being the most up-regulated gene. We found increased numbers of phosphorylated histone H2AX (γH2AX) in fetal liver and thymus cells from E18.5 Rnaseh2bKOF/KOF embryos, indicative of DNA double-strand breaks. Finally, we could show increased ribonucleotide loads in genomic DNA from embryos that were completely deficient for the RNase H2 enzyme. Collectively, we have demonstrated that complete RNase H2 deficiency causes persistent genomic ribonucleotide loads that render the DNA instable and prone to DNA strand breaks. DNA damage leads to the activation of p53 that in turn activates the cell cycle inhibitor p21 that inhibits cell cycle progression and likely causes accumulation of RNase H2-deficient cells in G2/M. To bypass early lethality we also generated bone marrow chimera and cell type-specific knockouts of the Rnaseh2b gene. While fetal liver cells of E14.5 Rnaseh2bKOF/KOF embryos could maintain hematopoiesis of irradiated recipient mice for almost one year, bone marrow cells from these primary recipients failed to reconstitute lethally irradiated mice in a secondary transfer. In line with this observation, VavCre-mediated deletion of the Rnaseh2b gene caused a more than hundred fold reduction of peripheral blood B cells, while B cell numbers remained unaltered upon CD19Cre-mediated deletion that occurs much later in B cell development. These data suggested that RNase H2 deficiency leads to the accumulation of genomic ribonucleotides that might cause the accumulation of a so far uncharacterized DNA damage species with increasing cell cycle passages. The data further supported our hypothesis that the impact of RNase H2 deficiency is determined by the number of cell proliferation. Finally, an epidermis-specific knockout of the Rnaseh2b gene displayed the most dramatic phenotype. Knockout mice were characterized by hyperpigmentation, hair loss and spontaneous ulcerations of the skin. Microscopically, these mice displayed moderate thickening of the epidermis and dermal fibrosis as indicated by increased collagen deposition. Macroscopic skin phenotypes were completely dependent on p53 expression, since concomitant deletion of the p53 gene rescued mice from hyperpigmentation, hair loss and ulcerations. This data demonstrated that Rnaseh2b deficiency in the epidermis may also lead to DNA damage and subsequent p53 activation as shown for fetal liver from E14.5 RNase H2-deficient embryos. Preliminary data also indicate an increased incidence of cancer formation in RNase H2/p53 double knockouts, identifying the RNase H2 enzyme as an important tumor suppressor.

Ribonucleotide excision repair (RER) is an RNase H2-dependent DNA-repair mechanism removing mis-incorporated ribonucleotides to maintain DNA stability. Decreased RNase H2 activity leads to an accumulation of ribonucleotides in the DNA, destabilizing and eventually damaging the DNA. This results in double-strand breaks, chromosome abberations, impaired segregation of defective chromosomes, and the formation of micronuclei with unstable nuclear membranes. Upon breakdown of the mironuclear envelope, the released chromatin triggers a cGAS-STING-dependent immune cascade that stimulates the production of type I interferons and cytokines. RNase H2 deficiency directly contributes to autoinflammation and autoimmunity and might further play a role in several types of cancer, aging and neurodegeneration. Therefore, RNase H2 activity is a promising diagnostic and prognostic marker. However, until today, no method for quantification of RNase H2 activity has been validated for a clinical use. Herein, a standard operating procedure for a high-throughput FRET-based whole cell lysate RNase H2 assay is implemented and validated delivering standard curves, statistical benchmarks and standardization to an externally validated control. Providing high sensitivity and strict linearity over a wide working range, the assay is applicable to various human cell or tissue samples with overall methodological assay variability from 8.6% to 16%. Human T cells were identified as a suitable cell type for the implementation of a clinical screening method, showing relatively small inter-individual variability when activity is normalized to cell number. Indeed, decreased RNase H2 activity was detected in T cells from one patient with systemic sclerosis and two patients with systemic lupus erythematosus who carried RNASEH2 mutations known to disrupt enzyme function in vitro compared with a control group of 24 healthy donors. With these findings, this dissertation provides fundamentals for the implementation of an RNase H2 assay screening method in the clinical setting. For the actual clinical application, however, the establishment of a significantly larger control group is necessary. This might allow identification of further inter-individual variables influencing RNase H2 activity and facilitate the determination of a threshold below which a reduction of RNase H2 activity is likely to become clinically relevant. Phenotypic effects of RNASEH2 mutations are often assessed in experiments with recombinant enzyme. However, this does not allow conclusions to be drawn about the extent to which the mutations affect enzyme activity through transcriptional, post-transcriptional, translational, or post-translational processes. In contrast, direct measurement of RNase H2 activity in cell lysates facilitates a more comprehensive assessment of the clinical relevance of genetic variants. To assess whether HeLa cell models are suitable for studying the intracellular effects of RNASEH2 mutations on enzyme activity, the well-studied RNASEH2B mutation A177T was inserted into HeLa cells using CRISPR/Cas9. However, due to high variability of RNase H2 activity in the HeLa cell clones after transfection and clonal selection, and low targeting efficiency, this approach has limited potential in HeLa cells. Direct use of immortalized cell lines derived from patient tissue, or a CRISPR/Cas approach in iPS cells might be more promising. As a secondary result, this study provides evidence that intracellular RNase H2 activity is increased in S, G2, and M phase of the cell cycle. In addition, increased RNase H2 activity was seen after stimulation with LPS and IL-2, and especially the mitogen PMA suggesting various pathways of RNase H2 activity regulation.:CONTENTS CONTENTS 5 ABBREVIATIONS 8 ABSTRACT 13 ZUSAMMENFASSUNG 15 1 INTRODUCTION 17 1.1 RIBONUCLEOTIDE EXCISION REPAIR IN HEALTH AND DISEASE 17 1.1.1 Role and function of RER in mammals 17 1.1.2 Clinical relevance of dysfunctional RER 18 1.1.2.1 Aicardi-Goutières syndrome – a paradigm for autoimmunity 19 1.1.2.2 RNase H2 and malignancy 21 1.1.2.3 DNA-damage, aging and neurodegeneration 21 1.2 RIBONUCLEASE H2 21 1.2.1 Genetic and biochemical characteristics 21 1.2.2 Rnase H2 activity assays 23 1.2.3 Known mutations and their effects on enzymatic function 24 1.3 CRISPR/CAS: A TOOL FOR TARGETED MUTAGENESIS 27 1.4 AIM OF THIS THESIS 28 2 MATERIAL AND METHODS 29 2.1 MATERIAL 29 2.1.1 Chemicals and Reagents 29 2.2.1.1 Nucleic acids 29 2.2.1.2 Enzymes 30 2.2.1.3 Antibodies 31 2.2.1.4 Buffers and solutions 31 2.2.1.5 Cell growth media 32 2.2.1.6 Basic chemicals and reagents 32 2.1.2 Consumables 34 2.1.3 Kits 34 2.1.4 Devices 35 2.1.5 Cell lines 36 2.1.6 Animal clones 36 2.1.7 Software, databases and websites 37 2.2 METHODS 37 2.2.1 Cell methods 37 2.2.1.1 Isolation of primary cells 37 2.2.1.2 Cell culture 39 2.2.1.3 Cell analysis 41 2.2.2 Nucleic acid methods 42 2.2.2.1 DNA isolation via isopropanol precipitation 42 2.2.2.2 DNA isolation via ethanol precipitation in 96-well plates 42 2.2.2.3 DNA quantification 43 2.2.2.4 Primer design for PCR 43 2.2.2.5 Polymerase chain reaction (PCR) and gel electrophoresis 43 2.2.2.6 Purification of PCR products 44 2.2.2.7 Cloning methods 44 2.2.2.8 DNA Sequencing 45 2.2.3 Protein methods 46 2.2.3.1 Protein extraction from cells 46 2.2.3.2 Protein quantification 47 2.2.4 Statistics and informatics 47 2.2.5 CRISPR/Cas methods 47 2.2.5.1 sgRNA design 47 2.2.5.2 Cloning of plasmids containing sgRNA cassettes 48 2.2.5.3 Design of repair templates 49 2.2.5.4 Generating genetically modified HeLa cells using CRISPR/Cas9 49 2.2.5.5 Genotyping of cell clones generated by CRISPR/Cas9 50 2.2.6 RNase H2 activity 50 2.2.7 Control group inclusion criteria 50 3 RESULTS 51 3.1 ESTABLISHING THE RNASE H2 ASSAY 51 3.1.1 Method establishment 51 3.1.1.1 Methodological approach 51 3.1.1.2 Assay workflow and normalization 51 3.1.1.3 Establishing basic assay settings 54 3.1.1.4 Time-resolved measurement 54 3.1.1.5 Establishing controls 56 3.1.1.6 Fluorescence standard curves 62 3.1.1.7 Interpretation of the fluorescence progress curve 62 3.1.1.8 Steady-state kinetics: Definition of assay end-points 65 3.1.1.9 Standardization to externally validated controls 66 3.1.1.10 Ruggedness 68 3.1.1.11 Influence of cell cycle and stimulation on RNase H2 activity 70 3.1.2 Assay precision 70 3.1.2.1 Coefficient of variation 70 3.1.2.2 Experimental design 71 3.1.2.3 Error levels I – III: from linear regression to pipetting error 71 3.1.2.4 Error level IVa and IV: quantification error 75 3.1.2.5 Error levels V and VI: cell preparation errors 77 3.1.2.6 Calculation of individual CVs 79 3.1.2.7 Replication of individual assay steps and the effective CV 81 3.1.2.8 Inter-assay variability the use of standards 82 3.1.3 RNase H2 activity of different cell types 82 3.2 ESTABLISHING A SCREENING STRATEGY FOR RNASE H2 ACTIVITY 85 3.2.1 Choice of cell type and cell isolation 85 3.2.2 Recruitment of the control group 86 3.2.3 Biological variability of RNase H2 activity in B cells and T cells 86 3.2.4 Sample size and effect size 89 3.2.5 Reduced RNase H2 Activity in T Cells of Patients with Systemic Autoimmunity 91 3.3 GENERATION OF AN RNASEH2BA177T CELL MODEL 93 3.3.1 Experimental design 93 3.3.2 Genotyping results 94 3.3.3 Impact of the RNASEH2B A177T mutation on RNase H2 activity 95 4 DISCUSSION 98 4.1 RNASE H2 ASSAY 98 4.1.1 Qualitative validity 98 4.1.1.1 Assay end-points 98 4.1.1.2 Determination of RNase H2 activity from enzyme progress curves 100 4.1.1.3 Normalization 102 4.1.1.4 Validation and control of systematic errors 104 4.1.2 Quantitative considerations 107 4.1.2.1 Sensitivity, precision and replication 107 4.1.2.2 Applicability for high-throughput analysis 108 4.1.3 Perspective 108 4.2 RNASE H2 ACTIVITY SCREENING IN HUMAN CD3+ CELLS 109 4.3 CELL MODELS FOR PATHOGENIC RNASE H2 VARIANTS 112 4.4 RNASE H2 FUNCTION AND REGULATION 113 4.4.1 RNase H2 and transcription 113 4.4.2 RNase H2 kinetic parameters 115 4.4.3 RNase H2 activity during the cell cycle and induction by PMA 115 4.4.4 RNase H2 activity in different cell types 117 REFERENCES 119 APPENDIX 134 APP. 1: ASSAY SUBSTRATES 134 APP. 2: ANALYSIS OF ERROR SOURCES 134 Biological errors 134 Procedural errors 137 APP. 3: QUBITTM PROTEIN ASSAY PERFORMANCE CHARACTERISTICS 139 APP. 4: ‘ACCURACY’ AND RELATED TERMS 140 APP. 5: CHARACTERISTICS OF THE SYSTEMIC SCLEROSIS PATIENT SSC1 141 APP. 6: RNASE H2 SUBUNIT PROTEIN EXPRESSION IN DIFFERENT TISSUES 142 APP. 7: PARADIGM CALCULATION OF THE EFFECTIVE METHODOLOGICAL CV 143 APP. 8: GENOTYPING RESULTS OF CRISPR/CAS9-GENERATED HELA CLONES 144 APP. 9: RNASE H2 ASSAY STANDARD OPERATING PROCEDURE 146 SOP 1 cell preparation and lysis 146 SOP 1.1 Material and reagents 146 SOP 1.2 Assay planning 146 SOP 1.3 Prepare cell pellets 147 SOP 1.4 Lysis 147 SOP 2 Qubit™ protein assay 148 SOP 2.1 Material 148 SOP 2.1 Working procedure 148 SOP 3 RNase H2 assay 150 SOP 3.1 Material and reagents 150 SOP 3.2 Prepare a plate layout and a pipetting scheme 151 SOP 3.3 Prepare the reaction buffer and substrates 151 SOP 3.4 Prepare your lysate premix (volume B, 65 µl) 151 SOP 3.5 Prepare the photometer 152 SOP 3.6 Start the reaction by adding volume A (55 µl) to the reaction plate 152 SOP 3.9 Insert the plate, perform gain adjustment and start the test run 152 SOP 3.10 Data analysis 152 SOP 4 Figures and Charts 155 RNase H2 assay work flow 155 Assay substrates 156 Chart A. Corrected CVs of all error levels 157 Estimation of the effective CV for a planned experiment 158 Chart B: RNaseH2 assay working range for different cell types 159 Chart C: Approximate cell yield of biological material 159 Chart D: Plate layout 160 Chart E: Pipetting scheme 160 Pipetting work flow 161 Chart F: Fluorescence raw data table 161 Calculation of standard catalytic activity using standard curves 162 Inter-assay comparability 163 ACKNOWLEDGMENTS 164 DECLARATIONS 165 / Ribonukleotid-Exzisionsreparatur ist ein RNase-H2-abhängiger DNA-Reparaturmechanismus, der durch die Entfernung fälschlich eingebauter Ribonukleotide die Integrität und Stabilität der DNA erhält. Eine verminderte RNase-H2-Aktivität führt zu einer Anhäufung von Ribonukleotiden in der DNA, wodurch die DNA destabilisiert wird und schließlich Schaden nimmt. Das Resultat sind Doppelstrangbrüche, Chromosomenabberationen, eine gestörte Segregation der defekten Chromosomen und die Bildung von „Mikrokernen“ mit instabilen Kernmembranen. Bei Zerfall dieser Mikrokern-Hüllen löst das freiwerdende Chromatin eine cGAS-STING-abhängige Immunkaskade aus, welche die Bildung von Typ-I-Interferonen und Zytokinen stimuliert. Verminderte RNase-H2-Aktivität trägt dadurch direkt zur Entstehung von Autoinflammation und Autoimmunität bei und spielt wahrscheinlich auch als Malignitätsfaktor einiger Karzinome, sowie bei Alterungsprozessen und Neurodegeneration eine Rolle. Daher kann RNase-H2-Aktivität als ein vielversprechender diagnostischer und prognostischer Marker angesehen werden. Bisher etablierte Methoden zur Messung der RNase-H2-Enzymaktivität verfügen jedoch nicht über die Standardisierung und Validierung, welche für den klinischen Einsatz notwendig sind. Diese Dissertation implementiert eine Standardvorgehensweise, Standardkurven und statistische Kenngrößen für einen FRET-basierten RNase-H2-Assay. Der Assay ist für die Anwendung mit Zelllysaten validiert, und liefert standardisierte Ergebnisse. Durch eine hohe Sensitivität und eine strikte Linearität über einen großen Arbeitsbereich kann der Assay in vielen verschiedenen Zell- oder Gewebetypen angewendet werden. Die Gesamt-Variabilität beträgt dabei zwischen 8,6 % bis 16 %. Aufgrund einer relativ niedrigen inter-individuellen Schwankung der zellulären RNase-H2-Aktivität sind menschliche T-Zellen ein geeigneter Zelltyp für klinische Vergleichsstudien. So konnte in T-Zellen einer Patientin mit Systemischer Sklerose und zweier Patientinnen mit Systemischem Lupus Erythematodes, welche bekannte heterozygote RNASEH2 Mutationen aufwiesen, eine verminderte RNase-H2-Aktivität im Vergleich zu einer Kontrollgruppe mit gesunden Probanden gefunden werden. Diese Dissertation liefert die Grundlagen für die Implementierung eines RNase-H2-Assays als klinisches Diagnostikum. Für eine tatsächliche klinische Anwendung ist jedoch die Etablierung einer deutlich größeren Kontrollgruppe notwendig. Dadurch könnten einerseits weitere interindividuelle Einflussgrößen auf die RNase-H2-Aktivität identifiziert werden. Andererseits könnte dies die Festlegung eines Schwellenwerts ermöglichen, unter welchem sich eine Reduktion der RNase-H2-Aktivität wahrscheinlich klinisch manifestiert. Zur Beurteilung der phänotypischen Auswirkungen von RNASEH2 Mutationen werden häufig Experimente mit rekombinanter RNase-H2-durchgeführt. Dies lässt allerdings keine Aussagen darüber zu, inwiefern die Mutationen die Enzymaktivität durch transkriptionelle, post-transkriptionelle, translationale oder post-translationale Prozesse beeinflussen. Die direkte Messung der RNase-H2-Aktivität in Zelllysaten ermöglicht eine umfassendere Bewertung der klinischen Relevanz genetischer Varianten. Zur Einschätzung ob HeLa-Zell-Modelle geeignet dafür sind, die intrazellulären Auswirkungen von RNASEH2-Mutationen auf die Enzymaktivität zu untersuchen, wurde mittels CRISPR/Cas9 die vielseits publizierte RNASEH2B-Mutation A177T in HeLa-Zellen eingefügt. Jedoch fand sich nach der Transfektion und Zellklonierung eine sehr hohe Variabilität der RNase-H2-Aktivität zwischen den HeLa-Zellklonen. Aufgrund der zudem relativ niedrigen Targeting-Effizienz scheinen HeLa-Zellen für diese Fragestellung ein wenig geeigneter Zelltyp zu sein. Die direkte Verwendung von immortalisierten Zelllinien aus Patientengewebe, oder die Anwendung von CRISPR/Cas in iPS-Zellen könnten vielversprechender sein. Als Nebenbefund fand sich in dieser Dissertation eine erhöhte RNase-H2-Aktivität in der S-, G2- und M-Phase des Zellzyklus, sowie nach der Stimulation mit LPS und IL-2, sowie insbesondere dem Mitogen PMA. Dies liefert Hinweise zu möglichen intrazellulären Regulationswegen der RNase-H2-Aktivität, über welche bisher wenig bekannt ist.:CONTENTS CONTENTS 5 ABBREVIATIONS 8 ABSTRACT 13 ZUSAMMENFASSUNG 15 1 INTRODUCTION 17 1.1 RIBONUCLEOTIDE EXCISION REPAIR IN HEALTH AND DISEASE 17 1.1.1 Role and function of RER in mammals 17 1.1.2 Clinical relevance of dysfunctional RER 18 1.1.2.1 Aicardi-Goutières syndrome – a paradigm for autoimmunity 19 1.1.2.2 RNase H2 and malignancy 21 1.1.2.3 DNA-damage, aging and neurodegeneration 21 1.2 RIBONUCLEASE H2 21 1.2.1 Genetic and biochemical characteristics 21 1.2.2 Rnase H2 activity assays 23 1.2.3 Known mutations and their effects on enzymatic function 24 1.3 CRISPR/CAS: A TOOL FOR TARGETED MUTAGENESIS 27 1.4 AIM OF THIS THESIS 28 2 MATERIAL AND METHODS 29 2.1 MATERIAL 29 2.1.1 Chemicals and Reagents 29 2.2.1.1 Nucleic acids 29 2.2.1.2 Enzymes 30 2.2.1.3 Antibodies 31 2.2.1.4 Buffers and solutions 31 2.2.1.5 Cell growth media 32 2.2.1.6 Basic chemicals and reagents 32 2.1.2 Consumables 34 2.1.3 Kits 34 2.1.4 Devices 35 2.1.5 Cell lines 36 2.1.6 Animal clones 36 2.1.7 Software, databases and websites 37 2.2 METHODS 37 2.2.1 Cell methods 37 2.2.1.1 Isolation of primary cells 37 2.2.1.2 Cell culture 39 2.2.1.3 Cell analysis 41 2.2.2 Nucleic acid methods 42 2.2.2.1 DNA isolation via isopropanol precipitation 42 2.2.2.2 DNA isolation via ethanol precipitation in 96-well plates 42 2.2.2.3 DNA quantification 43 2.2.2.4 Primer design for PCR 43 2.2.2.5 Polymerase chain reaction (PCR) and gel electrophoresis 43 2.2.2.6 Purification of PCR products 44 2.2.2.7 Cloning methods 44 2.2.2.8 DNA Sequencing 45 2.2.3 Protein methods 46 2.2.3.1 Protein extraction from cells 46 2.2.3.2 Protein quantification 47 2.2.4 Statistics and informatics 47 2.2.5 CRISPR/Cas methods 47 2.2.5.1 sgRNA design 47 2.2.5.2 Cloning of plasmids containing sgRNA cassettes 48 2.2.5.3 Design of repair templates 49 2.2.5.4 Generating genetically modified HeLa cells using CRISPR/Cas9 49 2.2.5.5 Genotyping of cell clones generated by CRISPR/Cas9 50 2.2.6 RNase H2 activity 50 2.2.7 Control group inclusion criteria 50 3 RESULTS 51 3.1 ESTABLISHING THE RNASE H2 ASSAY 51 3.1.1 Method establishment 51 3.1.1.1 Methodological approach 51 3.1.1.2 Assay workflow and normalization 51 3.1.1.3 Establishing basic assay settings 54 3.1.1.4 Time-resolved measurement 54 3.1.1.5 Establishing controls 56 3.1.1.6 Fluorescence standard curves 62 3.1.1.7 Interpretation of the fluorescence progress curve 62 3.1.1.8 Steady-state kinetics: Definition of assay end-points 65 3.1.1.9 Standardization to externally validated controls 66 3.1.1.10 Ruggedness 68 3.1.1.11 Influence of cell cycle and stimulation on RNase H2 activity 70 3.1.2 Assay precision 70 3.1.2.1 Coefficient of variation 70 3.1.2.2 Experimental design 71 3.1.2.3 Error levels I – III: from linear regression to pipetting error 71 3.1.2.4 Error level IVa and IV: quantification error 75 3.1.2.5 Error levels V and VI: cell preparation errors 77 3.1.2.6 Calculation of individual CVs 79 3.1.2.7 Replication of individual assay steps and the effective CV 81 3.1.2.8 Inter-assay variability the use of standards 82 3.1.3 RNase H2 activity of different cell types 82 3.2 ESTABLISHING A SCREENING STRATEGY FOR RNASE H2 ACTIVITY 85 3.2.1 Choice of cell type and cell isolation 85 3.2.2 Recruitment of the control group 86 3.2.3 Biological variability of RNase H2 activity in B cells and T cells 86 3.2.4 Sample size and effect size 89 3.2.5 Reduced RNase H2 Activity in T Cells of Patients with Systemic Autoimmunity 91 3.3 GENERATION OF AN RNASEH2BA177T CELL MODEL 93 3.3.1 Experimental design 93 3.3.2 Genotyping results 94 3.3.3 Impact of the RNASEH2B A177T mutation on RNase H2 activity 95 4 DISCUSSION 98 4.1 RNASE H2 ASSAY 98 4.1.1 Qualitative validity 98 4.1.1.1 Assay end-points 98 4.1.1.2 Determination of RNase H2 activity from enzyme progress curves 100 4.1.1.3 Normalization 102 4.1.1.4 Validation and control of systematic errors 104 4.1.2 Quantitative considerations 107 4.1.2.1 Sensitivity, precision and replication 107 4.1.2.2 Applicability for high-throughput analysis 108 4.1.3 Perspective 108 4.2 RNASE H2 ACTIVITY SCREENING IN HUMAN CD3+ CELLS 109 4.3 CELL MODELS FOR PATHOGENIC RNASE H2 VARIANTS 112 4.4 RNASE H2 FUNCTION AND REGULATION 113 4.4.1 RNase H2 and transcription 113 4.4.2 RNase H2 kinetic parameters 115 4.4.3 RNase H2 activity during the cell cycle and induction by PMA 115 4.4.4 RNase H2 activity in different cell types 117 REFERENCES 119 APPENDIX 134 APP. 1: ASSAY SUBSTRATES 134 APP. 2: ANALYSIS OF ERROR SOURCES 134 Biological errors 134 Procedural errors 137 APP. 3: QUBITTM PROTEIN ASSAY PERFORMANCE CHARACTERISTICS 139 APP. 4: ‘ACCURACY’ AND RELATED TERMS 140 APP. 5: CHARACTERISTICS OF THE SYSTEMIC SCLEROSIS PATIENT SSC1 141 APP. 6: RNASE H2 SUBUNIT PROTEIN EXPRESSION IN DIFFERENT TISSUES 142 APP. 7: PARADIGM CALCULATION OF THE EFFECTIVE METHODOLOGICAL CV 143 APP. 8: GENOTYPING RESULTS OF CRISPR/CAS9-GENERATED HELA CLONES 144 APP. 9: RNASE H2 ASSAY STANDARD OPERATING PROCEDURE 146 SOP 1 cell preparation and lysis 146 SOP 1.1 Material and reagents 146 SOP 1.2 Assay planning 146 SOP 1.3 Prepare cell pellets 147 SOP 1.4 Lysis 147 SOP 2 Qubit™ protein assay 148 SOP 2.1 Material 148 SOP 2.1 Working procedure 148 SOP 3 RNase H2 assay 150 SOP 3.1 Material and reagents 150 SOP 3.2 Prepare a plate layout and a pipetting scheme 151 SOP 3.3 Prepare the reaction buffer and substrates 151 SOP 3.4 Prepare your lysate premix (volume B, 65 µl) 151 SOP 3.5 Prepare the photometer 152 SOP 3.6 Start the reaction by adding volume A (55 µl) to the reaction plate 152 SOP 3.9 Insert the plate, perform gain adjustment and start the test run 152 SOP 3.10 Data analysis 152 SOP 4 Figures and Charts 155 RNase H2 assay work flow 155 Assay substrates 156 Chart A. Corrected CVs of all error levels 157 Estimation of the effective CV for a planned experiment 158 Chart B: RNaseH2 assay working range for different cell types 159 Chart C: Approximate cell yield of biological material 159 Chart D: Plate layout 160 Chart E: Pipetting scheme 160 Pipetting work flow 161 Chart F: Fluorescence raw data table 161 Calculation of standard catalytic activity using standard curves 162 Inter-assay comparability 163 ACKNOWLEDGMENTS 164 DECLARATIONS 165

The activation of the innate immune system is the first line of host defence against infection. Nucleic acids can potently stimulate this response and trigger a series of signalling cascades leading to cytokine production and the establishment of an inflammatory state. Mutations in genes encoding nucleases have been identified in patients with autoimmune diseases, including Aicardi-Goutières syndrome (AGS). This rare childhood inflammatory disorder is characterised by the presence of high levels of the antiviral cytokine interferon-α in the cerebrospinal fluid and blood, which is thought to be produced as a consequence of the activation of the innate immunity by unprocessed self-nucleic acids. This thesis therefore aimed to define the role of one of the AGS nucleases, the Ribonuclease H2 (RNase H2) complex, in innate immunity, and to establish if nucleic acid substrates of this enzyme were able to induce type I interferon production in vitro. The AGS nucleases may function as components of the innate immune response to nucleic acids. Consistent with this hypothesis, RNase H2 was constitutively expressed in immune cells, however, its expression was not upregulated in response to type I interferons. RNase H2-deficient cells responded normally to a range of nucleic acid PAMPs, which implied that a role for RNase H2 as a negative regulator of the immune response was unlikely, in contrast to the reported cellular functions of two other AGS proteins, TREX1 and SAMHD1. Therefore, no clear evidence was found for the direct involvement of RNase H2 in the innate immune response to nucleic acids. An alternative model for the pathogenesis of disease hypothesises that decreased RNase H2 activity within the cell results in an accumulation of RNA:DNA hybrids. To investigate the immunostimulatory potential of such substrates, RNA:DNA hybrids with different physiochemical properties were designed and synthesised. Methods to purify the hybrids from other contaminating nucleic acid species were established and their capacity as activators of the innate immune response tested using a range of in vitro cellular systems. A GU-rich 60 bp RNA:DNA hybrid was shown to be an effective activator of a pro-inflammatory cytokine response exclusively in Flt3-L bone marrow cultures. This response was completely dependent on signalling involving MyD88 and/or Trif, however the specific receptor involved remains to be determined. Reduced cellular RNase H2 activity did not affect the ability of Flt3-L cultures to mount a cytokine response against the RNA:DNA hybrid. These in vitro studies suggested that RNA:DNA hybrids may be a novel nucleic acid PAMP. Taken together, the data in this thesis suggest that the cellular function of RNase H2 is in the suppression of substrate formation rather than as a component of the immune response pathways. Future studies to identify endogenous immunostimulatory RNA:DNA hybrids and the signalling pathways activated by them should provide a detailed understanding of the molecular mechanisms involved in the pathogenesis of AGS and related autoimmune diseases.

京都大学 / 0048 / 新制・課程博士 / 博士(農学) / 甲第22493号 / 農博第2397号 / 新制||農||1076(附属図書館) / 学位論文||R2||N5273(農学部図書室) / 京都大学大学院農学研究科食品生物科学専攻 / (主査)教授 保川 清, 教授 佐々木 努, 教授 橋本 渉 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM