Investigation of the Binding Profile and Specificity of Monoclonal Iga to Microbiota Communities Under Steady State and Inflammatory Conditions

Investigation of the binding profile and specificity of monoclonal IgA to microbiota communities under steady state and inflammatory conditions

Von der Fakultät für Mathematik, Informatik und Naturwissenschaften der RWTH Aachen University zur Erlangung des akademischen Grades einer Doktorin der Naturwissenschaften genehmigte Dissertation

vorgelegt von

Johanna Kabbert, M.Sc. aus Berlin, Deutschland

Berichter: Univ.-Prof. Dr. rer. nat. Oliver Pabst

Univ.-Prof. Dr.-Ing. Lars Blank

Tag der mündlichen Prüfung: 01.06.2021

Diese Dissertation ist auf den Internetseiten der Universitätsbibliothek verfügbar.

Kabbert J, Benckert J, Rollenske T, Hitch C.A, Clavel T, Cerovic V, Wardemann H and Pabst O, High microbiota reactivity of adult human intestinal IgA requires somatic mutations. J Exp Med (2020) 217 (11): e20200275. https://doi.org/10.1084/jem.20200275

D 82 (Diss. RWTH Aachen University, 2021)

Table of Contents Abstract...... 1 Zusammenfassung ...... 3 1 Introduction ...... 5 1.1 “Guests” in the gut ...... 5 1.2 Mucosal surfaces ...... 5 1.3 Microbiota composition and distribution in the intestine ...... 6 1.4 The microbiota shapes host immunity and physiology ...... 7 1.5 Organization of the gut mucosal immune system ...... 8 1.6 Origin of B cells and antigen-independent BCR diversification ...... 11 1.7 Structure of antibodies in mice and men ...... 12 1.8 Luminal transport of intestinal antibodies ...... 14 1.9 SIgA coats members of the intestinal microbiota ...... 17 1.10 Effector functions of intestinal SIgA ...... 18 1.11 T cell dependent and T cell independent generation of intestinal IgA ...... 20 1.12 Generation of affinity-matured IgA responses in Peyer’s patches ...... 22 1.13 How does luminal antigen reach follicles in Peyer’s patches? ...... 23 1.14 B cell encounter with antigen in Peyer’s patches ...... 23 1.15 SHM and CSR in germinal centers require AID activity ...... 24 1.16 Mechanisms leading to IgA CSR in Peyer’s patches ...... 26 1.17 Mechanism promoting increased affinity of antibody responses ...... 27 1.18 Plasma blast homing to the intestinal lamina propria ...... 30 2 Aims ...... 31 3 Materials and Methods Part I ...... 32 3.1 Animals ...... 32 3.2 Human fecal material ...... 32 3.3 Preparation of human tissue samples ...... 32 3.3.1 Human tissue samples for monoclonal antibody generation ...... 32 3.3.2 Flow cytometry analysis and single B cell sorting ...... 32 3.3.3 PCR amplification and expression vector cloning ...... 33 3.4 Recombinant antibody production and purification ...... 33 3.4.1 Re-transformation of plasmids ...... 33 3.4.2 Plasmid purification and Ig gene sequence analysis...... 33 3.4.3 Expression and Protein A based purification of recombinant antibodies ...... 34 3.5 Enzyme linked Immunosorbent Assay (ELISA) ...... 34 3.5.1 Anti-human IgG1 ELISA ...... 34

3.5.2 Polyreactivity ELISA of recombinant antibodies ...... 34 3.6 Deglycosylation of immunoglobulins ...... 35 3.6.1 Deglycosylation of recombinant Fc-IgG1 antibodies ...... 35 3.7 Generation of germ-line variants ...... 35 3.7.1 Ig sequence analysis of recombinant antibodies ...... 35 3.7.2 Cloning of germ-line variants ...... 35 3.8 Western blot ...... 36 3.8.1 Western blot analysis of germ-line variants ...... 36 3.8.2 Western blot analysis of deglycosylated recombinant antibodies ...... 36 3.9 Mass spectrometry (MS) ...... 37 3.9.1 Mass spectrometry of deglycosylated recombinant antibodies ...... 37 3.10 Fecal material collection and preparation ...... 37 3.10.1 Preparation of fecal material ...... 37 3.11 Flow cytometry of antibody-stained bacteria ...... 38 3.11.1 Bacterial flow cytometry and bacterial sort purification (FACS) ...... 38 3.12 16S rRNA gene amplicon sequencing and analysis ...... 39 3.12.1 Metagenomic bacterial DNA isolation ...... 39 3.12.2 PCR amplification of bacterial DNA ...... 39 3.12.3 16S rRNA gene amplicon Illumina sequencing ...... 40 3.12.4 16S rRNA gene amplicon analysis ...... 40 3.13 Microbiology ...... 40 3.13.1 In vitro cultivation of oligoMM12 bacteria ...... 40 3.14 Data analysis ...... 41 3.14.1 Formulas ...... 41 3.14.2 Statistical analysis ...... 41 4 Materials and Methods Part II ...... 42 4.1 Animals ...... 42 4.2 Intraperitoneal application of Tamoxifen ...... 42 4.3 FTY720 administration ...... 42 4.4 Surgical procedures ...... 43 4.4.1 4-OHT micro-injections of single Peyer’s patches ...... 43 4.5 Lymphocyte isolation ...... 44 4.5.1 Tissue preparation and cell isolation ...... 44 4.5.2 Cell isolation from blood ...... 44 4.5.3 Cell isolation from lymphoid organs ...... 45 4.5.4 Lymphocyte isolation from small intestinal lamina propria ...... 45

4.6 Epithelial cell isolation from the small intestine ...... 45 4.7 Cell stainings for flow cytometry analysis ...... 46 4.7.1 Staining of B cells and plasma cells ...... 46 4.7.2 Staining of small intestinal epithelial cells ...... 46 4.8 Histology ...... 47 4.8.1 Tissue preparation for histological analysis ...... 47 4.8.2 Cryotome tissue sections ...... 47 4.9 Confocal microscopy ...... 48 4.10 Data Analysis ...... 48 4.10.1 Statistical analysis ...... 48 5 Results Part I ...... 49 5.1 Microbiota-reactive IgA and IgG plasma cell derived antibodies are prevalent in the adult human small intestine ...... 49 5.2 Microbiota-reactive IgA mAbs consistently bind a major fraction of the microbiota .55 5.3 Donor-dependent variability of IgA mAb binding capacities to human gut bacteria .57 5.4 IgA-microbiota interactions are Fab-mediated and independent of glycosylation ....59 5.5 Intestinal microbiota-reactive IgA antibodies show broad VH gene usage ...... 60 5.6 Intestinal IgA shows broad but distinct binding to members of the gut microbiota ...62 5.7 High microbiota-reactive intestinal IgA mAbs are cross-species reactive ...... 67 5.8 IgA cross-species reactivity is not conferred by polyreactivity ...... 70 5.9 Microbiota-reactive intestinal antibodies carry high numbers of somatic mutations 73 5.10 High microbiota reactivity of human intestinal IgA requires somatic mutations ...... 76 5.11 Germ-line reversion of intestinal IgA mAbs does not cause polyreactivity ...... 79 5.12 Germ-line variants with sustained microbiota binding do not show relevant changes in their specificities to gut bacteria ...... 80 5.13 Intestinal IgA shows reduced binding to oligoMM12 fecal bacteria ...... 83 5.14 Incomplete binding to in vitro cultivated oligoMM12 bacteria ...... 85 5.15 High microbiota-reactive IgA targets distinct oligoMM12 bacteria ...... 87 6 Results Part II ...... 91 6.1 Systemic tamoxifen administration leads to stable eYFP expression in B cells ...... 92 6.2 Histological validation of eYFP expression in Peyer’s patches ...... 95 6.3 Frequencies of eYFP+ B cells increase over time in inductive gut compartments ...96 6.4 Single 4-OHT Peyer’s patch injections lead to local eYFP labeling of B cells ...... 98 6.5 Examination of 4-OHT dispersion after single Peyer’s patch injections...... 102 6.6 Spatial distribution of Cre-activity throughout the small intestine ...... 108 6.7 4-OHT after single Peyer’s patch injections is confined to the injection site ...... 109 6.8 eYFP+ B cells activated in a single PP home to the SI LP in steady state ...... 114 iii

7 Discussion ...... 125 7.1 General discussion and main findings ...... 125 7.2 Cross-species reactive antibodies – defining the term ...... 125 7.3 Different binding modes lead to cross-species reactivity of intestinal antibodies ... 126 7.4 Cross-species reactivity in the adult human gut requires somatic mutations and does not correlate with polyreactivity ...... 127 7.5 Cross-species reactive antibody responses in mice and humans ...... 128 7.6 Somatic mutations shape cross-species reactivity of intestinal antibodies ...... 131 7.7 Role of T cells in affinity-matured cross-species reactive IgA responses ...... 134 7.8 What determines the extent of SIgA-bacteria binding? ...... 134 7.9 Functional consequences of cross-species reactive IgA in the healthy gut ...... 136 7.10 Cross-species reactive antibodies are maintained in the inflamed gut ...... 137 8 Conclusion ...... 140 9 Appendix ...... 141 9.1 List of abbreviations ...... 141 9.2 OTU tables ...... 143 9.3 mAb and germ-line sequences ...... 149 9.4 List of Figures ...... 155 9.5 List of Tables ...... 157 10 Acknowledgements ...... 158 11 References ...... 160 Eidesstattliche Erklärung/Declaration of academic honesty ...... 177

Abstract Immunoglobulin A (IgA) is the dominant antibody isotype secreted into the intestinal lumen and plays a key role in gut homeostasis. IgA binds to luminal pathogens, toxins and members of the gut microbiota, thereby contributing to intestinal host immune-defense as well as maintaining a stable microbiota composition. In addition to IgA, in the human gut low levels of IgG are also present, which markedly increase during inflammation. Considering the vast changes of microbial consortia in the gut, it remains elusive how the host can generate and regulate relevant antibody-responses in this highly dynamic setting. Similarly, the specificity of human IgA and IgG to the microbiota and the mechanisms underlying antibody-microbiota interactions are still largely unknown. Here, we examined the binding capacity and specificity to the intestinal microbiota of a large set of monoclonal antibodies (mAbs) generated from intestinal IgA+ and IgG+ plasma cells (PCs) derived from human healthy donors (HD) or Crohn’s disease (CD) patients by bacterial flow cytometry. Our study revealed that a high frequency of IgA mAbs from both healthy and inflamed gut exhibited high microbiota-binding. Interestingly, we found similarly high microbiota-binding capacities among human IgG mAbs. To further determine the bacterial-binding profiles of microbiota-reactive human IgA mAbs, we sorted IgA-bound and unbound fecal bacteria and characterized their composition by 16S rRNA sequencing. Some bacterial taxa were commonly targeted by several mAbs, whereas others were selectively bound by single mAbs only. Notably, all high microbiota-reactive HD and CD derived IgA mAbs bound phylogenetically distant bacteria but showed in addition to common bacterial specificities unique binding profiles. We refer to this phenomenon as “cross- species reactivity”. Importantly, microbiota cross-species reactivity did not correlate with polyreactivity (i.e. binding to structurally non-related antigens) but was crucially dependent on accumulated somatic mutations. Accordingly, cross-species reactive IgA mAbs carried frequent somatic mutations and the majority of mAbs showed substantial loss of microbiota binding in their germ-line configuration. In addition, germ-line reversion did not cause these mAbs to gain polyreactivity. This strongly suggests that microbiota-reactive mAbs do not derive from originally polyreactive antibodies. We therefore propose that “early” polyreactive antibodies may become supplanted by highly mutated, affinity-matured antibodies during aging. The high numbers of somatic mutations suggest that cross-species reactive antibodies may be the result of ongoing selection during multiple rounds of affinity-maturation and most likely depend on T cell interactions. Notably, cross-species reactive antibodies did not enrich all members of a targeted species, implying that cross-species reactive IgA may target distinct genetic sub-strains within bacterial species or bacteria in a particular growth-state. Collectively, our data suggest that a system of affinity-matured, cross-species reactive antibodies is one dominant mechanism of IgA-microbiota interactions in the human gut. We propose that cross- species reactivity of IgA facilitates broad but specific binding to phylogenetically distant gut bacteria to efficiently interact with the microbiota in both the healthy and inflamed gut. The continuous exposure to varied but structurally-similar antigens might enable the selection for cross-species reactive IgA responses. Prospectively, identifying the microbiota-targeting profiles of IgA derived from healthy or inflamed gut may allow to develop anti-microbial mAbs as diagnostic or therapeutic tools.

Another aspect of intestinal immunity relates to the generation of IgA-producing PCs. However, how IgA responses to the microbiota are generated under steady state conditions is largely unknown. Also, it is uncertain which inductive compartments of the gut associated lymphoid tissue predominantly contribute to the generation and differentiation of intestinal IgA-secreting PCs. Additionally, there is still limited knowledge about the contribution of newly generated IgA+ PCs to the prevailing intestinal PC pool in homeostasis. To investigate these questions, we have established an in vivo mouse model to monitor the differentiation and migration kinetics of activated B cells. In this model, the local injection of 4-Hydroxy-Tamoxifen into a single intestinal Peyer’s patch (PP) leads to the irreversible eYFP-labeling of activated B cells, which allows to track activated B cells and all their progeny. Notably, in steady state and in the absence of neo-antigen challenge, we found a surprisingly high frequency of activated eYFP+ B cells in PPs. Our data further demonstrate that a marked frequency of eYFP+ B cells originating from a single PP re-circulate through different PPs, with only transient homing to the mesenteric lymph nodes. This suggests that B cells activated in a single PP may undergo affinity-maturation in multiple PPs leading to the generation of highly specific IgA+ PCs. Moreover, eYFP+ B cells generated in a single PP had the capacity to transition into IgA+ PCs and contributed to the PC pool in the small intestinal lamina propria up to 90 days after induction. In the future, our established model will allow to study the clonal overlap of PC- repertoires in different gut compartments and to identify functional differences in IgA+ PCs generated under healthy or inflammatory conditions. Therefore, our method of local B cell labeling in single PPs, may provide further understanding of the underlying mechanisms leading to cross-species reactive IgA responses to the microbiota.

Zusammenfassung Immunoglobulin A (IgA) ist der vorherrschende Antikörper-Isotyp im Darmlumen und spielt eine entscheidende Rolle für die Darm-Homöostase. IgA bindet Pathogene, Toxine und die intestinale Mikrobiota und trägt als wichtiger Bestandteil der intestinalen Immunabwehr maßgeblich zu einer stabilen Zusammensetzung der Mikrobiota bei. Zusätzlich zu IgA sind im menschlichen Darm auch geringe Mengen an IgG-Antikörpern vorhanden, die sich im Zuge von Entzündungen stark erhöhen. Eine Vielzahl von Faktoren nehmen Einfluss auf die Zusammensetzung der Darm-Mikrobiota. Wie im Gegenzug Antikörper-Antworten generiert und reguliert werden, um ein stabiles Gleichgewicht der Darm-Mikrobiota zu erhalten, ist noch immer weitestgehend unverstanden. Des Weiteren sind die Antigene der Darm-Mikrobiota, die von humanem IgA und IgG spezifisch erkannt werden, sowie die zugrunde liegenden Mechanismen solcher Antikörper-Mikrobiota-Interaktionen noch größtenteils unbekannt. In dieser Arbeit haben wir die Bindungskapazitäten und Spezifitäten von humanen monoklonalen Antikörpern (mAbs) für intestinale Mikrobiota mittels Durchflusszytometrie von Darmbakterien untersucht. Die humanen mAbs wurden von intestinalen IgA+ und IgG+ Plasmazellen (PCs) aus an Morbus Crohn (Crohn’s Disease, CD) erkrankten Patienten und gesunden Menschen (HD) generiert. Unsere Ergebnisse zeigen, dass eine Vielzahl humaner IgA-Antikörper aus dem entzündeten und dem gesunden Darm gleichermaßen die Fähigkeit haben, einen großen Anteil der Mikrobiota zu binden. Wir konnten ebenfalls zeigen, dass auch humane IgG- Antikörper Bakterien der Darm-Mikrobiota binden. Um die Bindungsprofile von intestinalen Mikrobiota-reaktiven IgA-mAbs genauer zu bestimmen, wurden von IgA-gebundene und nicht gebundene fäkale Bakterien durchflusszytometrisch sortiert und die Zusammensetzung beider Fraktionen mittels 16SrRNA-Sequenzierung charakterisiert. Es zeigte sich, dass einige Bakterien-Taxa gleichermaßen von einem Großteil der mAbs gebunden wurden, während andere Taxa lediglich von einzelnen mAbs gebunden wurden. Außerdem zeigten alle hoch Mikrobiota-reaktiven HD und CD IgA mAbs, zusätzlich zu ihrer Bindungskapazität für phylogenetisch diverse Bakterien, individuelle Bindungsprofile. Dieses Phänomen bezeichnen wir als „cross-species“ Reaktivität. Von entscheidender Bedeutung ist, dass „cross-species“ Reaktivität nicht per se auf Polyreaktivität beruht (d.h. Bindungen zu strukturell- unterschiedlichen Antigenen), sondern wesentlich von der Akkumulierung somatischer Mutationen abhängt. Entsprechend ist eine hohe Anzahl an somatischen Mutationen kennzeichnend für „cross-species“-reaktive IgA-Antikörper. Dahingegen war die Fähigkeit Mikrobiota zu binden, bei einem Großteil aller mutierten mAbs in deren Keimbahnkonfiguration deutlich reduziert. Zusätzlich führte die Reversion dieser mAbs in die Keimbahnkonfiguration nicht zu deren Polyreaktivtät. Dies legt nahe, dass Mikrobiota-reaktive IgA mAbs nicht von ursprünglich-polyreaktivien Antikörpern abstammen. Wir vermuten daher, dass polyreaktive Antikörper, die früh während der Entwicklung entstehen, im Laufe des Lebens durch mutierte und affinitätsgereifte Antikörper ersetzt werden. Die hohe Anzahl akkumulierter, somatischer Mutationen lässt auf eine kontinuierliche Selektion dieser „cross-species“-reaktiven Antikörper während mehrerer Affinitäts-Reifungsschritte schließen und erfordert wahrscheinlich die Interaktion mit T-Zellen. Interessanterweise wurden nicht alle Bakterien einer Spezies einheitlich von „cross-species“-reaktiven Antikörpern gebunden. Dies könnte auf eine gezielte Bindung „cross-species“-reaktiver IgA-Antikörper an genetische Untergruppen innerhalb einer 3

Bakterienspezies, oder aber an Bakterien in bestimmten Wachstumszuständen hindeuten. Zusammenfassend veranschaulichen unsere Daten, dass affinitätsgereifte, „cross-species“- reaktive Antikörper einen dominanten Mechanismus für die Interaktion zwischen IgA und der Mikrobiota im humanen Darm darstellen. Wir vermuten, dass „cross-species“ Reaktivität eine breite, aber dennoch spezifische Bindung von phylogenetisch diversen Darmbakterien ermöglicht, was sowohl im gesunden als auch im entzündeten Darm eine effiziente IgA- Mikrobiota Interaktion erlaubt. Vermutlich gewährleistet der kontinuierliche Kontakt mit vielfältigen, aber strukturell-ähnlichen Antigenen die Selektion von „cross-species“-reaktiven IgA-Antikörpern. Eine weitergehende Identifizierung der Darmbakterien, die von IgA unter homöostatischen oder entzündlichen Bedingungen gebunden werden, könnte perspektivisch die Entwicklung von Mikrobiota-spezifischen mAbs als diagnostisches oder therapeutisches Mittel ermöglichen. Ein weiterer Aspekt der intestinalen Immunität bezieht sich auf die Generierung von IgA- produzierenden PCs. Bisher ist nicht vollständig geklärt, welche induktiven Kompartimente des darmassoziierten-lymphatischen Gewebes maßgeblich an der Generierung von intestinalen IgA-Antworten unter homöostatischen Bedingungen beteiligt sind. Darüber hinaus ist noch ungewiss, inwiefern unter homöostatischen Bedingungen neu generierte IgA+ PCs einen Anteil zu bereits vorhandenen intestinalen PCs leisten. Um diese Fragen zu untersuchen, haben wir ein Mausmodell etabliert, das es erlaubt, die Differenzierung und Migrationskinetik aktivierter B-Zellen in vivo zu erfassen. In diesem Modell wird 4-Hydroxy-Tamoxifen lokal in einzelne Peyersche Platten (PPs) des Dünndarms injiziert. In der Folge erhalten B-Zellen, die in einer solchen PP aktiviert werden, eine irreversible Fluoreszenzmarkierung (eYFP), die es ermöglicht, aktivierte B-Zellen und ihre Nachkommen zu verfolgen. Wir konnten zeigen, dass im homöostatischen Zustand und in Abwesenheit von neo-Antigenen ein überraschend hoher Anteil aktivierter B-Zellen in PPs zu finden ist. Unsere Daten zeigen ferner, dass B-Zellen, die in einer bestimmten PP aktiviert wurden, durch verschiedene PPs re-zirkulieren und sich lediglich transient in den mesenterischen Lymphknoten aufhalten. Dies lässt vermuten, dass aktivierte B-Zellen Affinitätsreifungen in unterschiedlichen PPs durchlaufen, was zur Generierung von spezifischen IgA+ PCs führt. Zudem konnten wir zeigen, dass B-Zellen, die aus einem einzelnen PP stammen, zu IgA+ PCs differenzieren können und bis zu 90 Tage nach ihrer Induktion in der Lamina propria des Dünndarms nachweisbar sind. Die im Rahmen dieser Arbeit etablierte Methode, könnte in Zukunft dazu genutzt werden, klonale Überlappungen des PC-Repertoires in verschiedenen Darmkompartimenten zu untersuchen. Des Weiteren wäre es möglich, funktionelle Unterschiede von IgA+ PCs, die im gesunden oder im entzündeten Darm generiert wurden, zu untersuchen. Demzufolge ermöglicht die lokale B- Zell Markierung in einzelnen PPs auch, die zugrundeliegenden Mechanismen der „cross- species“ Reaktiviät von IgA-Antworten gegen die intestinale Mikrobiota in weitergehenden Experimenten gezielt zu untersuchen.

1 Introduction 1.1 “Guests” in the gut Evolution has provided the mammalian gut with an incredible array of defense mechanisms against invading pathogens, toxins, food derived antigens, but also enabled the mutualistic relationship between the host and in the gut residing commensals, termed the microbiota. The microbiota has been accompanying us ever since, and have become indispensable to the mammalian intestine. However, one might wonder how the host copes with the challenge of keeping the millions of prokaryotic “guests” in check, without mounting a constant inflammatory response and even established a mutualistic “tit for tat” scenario. Considering the vast diversity of microbial species colonizing the intestine, it is reasonable to believe that this reciprocal “dialogue” between host-immunity and the microbiota serves a purpose and has thus been evolutionary conserved ever since, resulting in sustained homeostatic host-microbiota symbiosis.

1.2 Mucosal surfaces Mucosal surfaces can be found throughout the entire mammalian body, at the nasal, respiratory, urogenital and gastrointestinal tract. These mucosal surface areas are typically composed of an epithelial layer that is overlaid by mucosal fluids, such as saliva, tears or mucus1-3. In mammals, mucosal surfaces constitute the largest areas within the body bridging the outside world and the interior environment of the host. Each of these mucosal compartments fulfill distinct physiological functions, such as gas exchange in the lung tissue or nutrient absorption in the intestine. Specialized mucosal epithelium enables the cross-talk of the host with the external environment, but thereby also posing a potential route of entry for a wide variety of pathogens and antigens. Nonetheless, mucosal surfaces, especially epithelial cells and associated mucosal fluids, represent a critical physiological barrier acting as first line defense against an environment rich in pathogens but also harmless antigens and commensal bacteria. Different subtypes of epithelial cells contribute to mucosal defense both by presenting a physical barrier but also by the secretion of humoral factors, e.g. mucins and defensins4. In addition to these innate mechanical and biochemical defense mechanisms, the mucosal tissue is also populated by large numbers of immune cells belonging to both innate and adaptive immunity. For example, intraepithelial lymphocytes (IELs) consisting of various T cell subsets constitute a resident part of the mucosal epithelium1,5,6. Furthermore, myeloid cells such as dendritic cells (DCs) or macrophages and lymphocytes such as B cells, T cells and innate lymphoid cells (ILCs) reside below the epithelium ready to act upon signals from epithelial cells1,3,7,8. Mammalian mucosal sites are thus highly interactive surfaces and interfaces with the outside world. A particular feature of the gastrointestinal mucosa is the dense colonization by commensal bacteria that are tolerated by the host and may even benefit from a thick mucus layer especially in the gut9,10. Thus, host immune responses enable both robust clearance of pathogens from mucosal surfaces but also facilitate ignorance or actively suppress inflammatory responses to non-pathogenic but beneficial commensals. One of the most critical humoral modulators of adaptive mucosal immunity specialized to keep microbial members and pathogens in check are antibodies, in particular immunoglobulin A (IgA)11-17. The mammalian immune system has evolved to invest considerable energy in the production of IgA that seems 5 hard-wired even in germ-free (GF) mice18-21. IgA constitutes the largest quantity of produced antibodies in the mammalian body, mostly found as secretory IgA (SIgA) at gut mucosal surfaces17,22-24. Intestinal SIgA has the capacity to shape microbial compositions, impact bacterial growth and metabolic functions and limit systemic dissemination of pathogens and bacteria-derived metabolites, thereby providing homeostatic host-microbiota interactions at intestinal interfaces15,23,25-29.

1.3 Microbiota composition and distribution in the intestine At birth, mammalian intestines are largely devoid of microbes, but soon after birth become colonized by a dense and complex population of bacteria (as well as eukaryotic microorganisms, archaea and viruses), termed the microbiota9,19,30,31. However, the bacterial colonization of the intestine differs in microbial diversity and density from birth to adulthood9,32- 34. At the outset of mammalian life, the type of delivery and mode of feeding significantly determine the community structure of commensals colonizing the neonatal gut. Here, components from maternal milk and passively transmitted maternal antibodies shape the initial microbial configuration in newborns, which is usually of lower diversity as compared to the adult gut. Maternally transmitted antibodies influence and shape the microbiota composition as early as after birth35-39 and facilitate the colonization of a select set of beneficial microbial members that are likewise beneficial in the maternal gut35,38-40. Thus, the transmission of maternal SIgA and also IgG enables the engraftment of favorable microbial members that have undergone “maternal pre-screening”31,36,38,40,41. The infant gut is dominated by microbial species such as Lactobacillus sp., Bifidobacterium sp., Escherichia coli, Staphylococcus, Bacteroides fragilis, and Streptococcus42,43. During early life, mammals experience highly dynamic changes such as nutritional changes and extensive exposure to environmental antigens that in turn influence the microbial composition in the gut9,31,33,34,44. Solid food intake coincides with a profound shift of prevalent microbial members33,45-47. Consequently, microbial diversity increases with age and reaches a relatively stable microbial configuration by approximately three years of age, then resembling the adult microbiota9,34,44,48. In the adult gut the density and diversity of the microbiota varies significantly throughout the intestinal tract, which may reflect an adaptation of gut bacteria to distinct immune niches31,34,49,50. The small intestinal mucosa is colonized by ~103/g bacterial cells as compared to a colonization densitiy of ~1012/g in the large intestine30,31,33,51 (Figure 1). The “healthy” adult microbiota composition is numerically dominated by species belonging to the major gut phyla Bacteroidetes (Bacteroidaceae, Prevotellaceae, and Rikenellaceae), Firmicutes (Lachnospiraceae, Ruminococcaceae), Actinobacteria (Bifidobacteria, Coriobacteriaceae) and Proteobacteria (Enterobacteriaceae) and less numerous by Fusobacteria, Verrucomicrobia (Akkermansia muciniphila) and Euryarchaeota (Methanobrevibacter smithii)32,49,52,53. The small intestine is typically dominated by members of Enterobacteriaceae and Lactobacillaceae, whereas the colon contains members of Bacteroidaceae, Lachnospiraceae, Prevotellaceae, Rikenellaceae, and Ruminococcaceae, respectively34,35,44. The microbiota is nevertheless a vulnerable ecosystem and the structure of microbial communities may be subject to alterations throughout life9,33,48,54,55. Therefore, adult microbiota configurations can show substantial inter-individual variation56, with highly unique sets of bacterial strains found in between individuals that may 6 result from factors including host genetics, environmental changes, diet, pharmaceutical interventions, lifestyle and gut physiology33,34,57-59. Despite marked inter-individual differences of microbial communities, a core set of microbial members (enterotypes) and microbial encoded genes can be found across individuals32,52,60. This strongly indicates that individual microbial communities nonetheless maintain conserved functionality54,57,60,61. Altogether, the intestinal microbiota and microbial-derived metabolites are essential to maintain host homeostasis shaping many aspects of host physiology and immunity9,10,31,56,62,63.

1.4 The microbiota shapes host immunity and physiology At the outset of mammalian life, newborns are not only largely devoid of bacteria but in addition only have a primitive and immature immune system and lack fully developed mucosal lymphoid tissue, despite already established Anlagen for lymphoid structures such as Peyer’s patches (PPs)6,64-66. Notably, commensals in addition to nutritional components such as retinoic acid (RA) seem to be indispensable for the development and maturation of host immunity and physiology10,13,22,47,55,66,67. Furthermore, microbial colonization influences the post-natal development of the gut epithelial barrier as well as the maturation and programming of host immune cells within intestinal lymphoid compartments31,47,55,66,68-70. The first encounter with commensal bacteria to the neonate intestine initiates the differentiation of microfold (M) cells in the PP overlying epithelium (FAE) and results in the activation of PP germinal centers (GCs) marked by an influx of B and T cells leading to the formation of secondary GCs resulting in increased numbers of IgA+ plasma cells (PCs) in the gut1,16,55,71,72. Thus, the maturation of mucosal lymphoid tissues and the generation of intestinal IgA antibodies typically requires the colonization with commensals10,55,56,62,67,72. Once stably acquired, the microbiota will become indispensable for the host. Then, the microbiota as well as microbial-derived products elicit a variety of beneficial functions that have major implications for host immunity and physiology9,62,73. Commensals are required for nutritional up-take, bile acid conversion, vitamin and hormone biosynthesis and provide colonization resistance to invading pathogens10,31,62,74. Moreover, the microbiota and microbial-derived metabolites provide immune modulatory stimuli9,10,73,75,76. Such immune-stimulatory and modulatory capacities have been described for segmented filamentous bacterium (SFB) and Akkermansia muciniphila that enhance intestinal IgA production and Th17 cell induction, associated with increased protection against pathogens9,31,77-79. Microbial-derived metabolites from the breakdown of proteins, fat and fibers literally affect the entire body and immune compartments. Commensals belonging to the Bacteroidetes and Firmicutes phylum, in particular Bacteroides sp. and Faecalibacterium prausnitzii play a central role in host physiology by providing short chain fatty acids (SCFAs), which are microbial derived metabolites from dietary fiber fermentation69,80. SCFAs have been shown to affect host metabolism and immune cells via binding to G-protein coupled receptors70,81 and thereby have been implicated to promote B cell differentiation into antibody secreting plasma cells (PCs) and enhance B cell metabolism and the generation of IgA inducing Tregs69,82,83. SCFA are also an important energy source for intestinal epithelial cells, thereby contributing to a robust gut barrier function84,85. The breach of the mucosal barrier results in the systemic spread of opportunistic-commensals leading to inflammatory responses that are associated with the development of pathologies such as inflammatory bowel disease 7

(IBD), obesity, type 2 diabetes, malnutrition and cardiovascular diseases. Similarly, major shifts in microbial composition, called dysbiosis, can have metabolic consequences associated with intestinal inflammation and chronic diseases9,34,63,86,87. Collectively, maintaining a balanced and beneficial microbiota composition is crucial for intestinal integrity, but this homeostatic state seems to require a bi-directional communication between the host and the microbiota34,74,88. A challenge for host-immunity remains how to appropriately discriminate between beneficial members of the microbiota and transiently invading pathogens23,88,89. Thus, major aspects of gut mucosal immunity comprise the fine-tuned balance between active immunity against pathogens and developing immune tolerance to beneficial commensals to both counteract pathogen challenges and to maintain beneficial host-microbiota symbiosis10,37,75,89-91.

1.5 Organization of the gut mucosal immune system The gut mucosal immune system is organized in inductive and effector sites2,6,92. The effector sites foremost comprise scattered immune cells in the mucosal tissue, whereas the inductive sites comprise the organized gut associated lymphoid tissue (GALT). The GALT thus acts as inductive sites for adaptive immune cell priming and differentiation. Activated and differentiated effector cells then home to gut mucosal tissues such as the lamina propria (LP) and the intestinal epithelium to constitute the effector site of the gut immune system2,6,92. The migration of immune cells via the lymphatic system and blood from inductive compartments to effector sites provides the basis for cellular and humoral immune responses. GALT comprises anatomically distinct and functionally specialized compartments that are adapted to local conditions, such as different pH-values, bacterial load and diversity as well as concentrations of nutrients and microbial metabolites. Hence, the GALT harbors the greatest number as well as the highest diversity of immune cells in the body comprising both lymphoid and myeloid cells2,6,17. In mice and humans, the most prominent inductive sites of the GALT are Peyer’s patches (PPs) that already develop in utero6,66,93. However, final maturation and further organization of PPs requires the colonization with microbiota to form organized secondary lymphoid follicles and chronic GCs structures, which are essential for the generation of intestinal IgA responses9,21,72,94. Although, PPs are probably the best known component of the inductive GALT1,23,64,65, additional structures such as the ceacal patch1 and the much smaller solitary isolated lymphoid tissues/follicles1,93 (SILTs or ILFs) that only form after birth and are scattered throughout the small and large intestinal mucosa and submucosa are part of the inductive GALT1,23,94-96. In addition to these GALT structures, the gut-draining mesenteric lymph nodes (MLNs) also constitutes an immune inductive compartment in the gut1,6,23,94 (Figure 1). Despite the ascribed immune-inductive capacities of these tissues in addition to PPs, their function and contribution to gut immunity are less well defined with major differences between species. Common to mature PPs are several large and distinct B cell enriched follicles with intervening T cell-rich areas. B cell follicles predominantly comprise IgM+(IgD+) mature naïve B cells64,97. Each B cell follicle is surrounded by a parafollicular T cell area and comprises large numbers of high endothelial venules (HEV) allowing for cell migration and lymphoid recirculation64,98,99. Although, B cell follicles in mature PPs vary in size and number throughout life, they in contrast to other encapsulated peripheral lymph nodes (pLNs), constitutively 8 contain active germinal centers (GCs) that develop from primary follicles upon encounter with microbial antigens soon after birth and are maintained thereafter in an active state64,65,72,97,100. Conversely to pLNs, PPs lack a capsule and lymphoid areas are instead separated from the antigen rich intestinal lumen by a single epithelial layer, the follicle associated epithelium (FAE) and the more diffuse areas immediately below the epithelium called the sub-epithelial dome (SED)6,64,99. Moreover, PPs lack afferent lymphatics in contrast to other pLNs6. Consequently, mature naïve B cells, in fact lymphocytes in general, enter PPs from the bloodstream via HEVs that are located both within follicles and T cell rich inter-follicular areas, whereas lymphatic vessels are primarily used for lymphocyte exit64,99,101. While pLNs receive antigen via afferent lymphatics and potentially blood, PPs sample antigen via the overlaying FAE, in particular via a unique epithelial cell type, known as microfold (M) cells. M cells are specialized enterocytes with a poorly developed brush boarder and a thin glycocalix layer making them proficient for luminal antigen up-take71 with subsequent trans-epithelial vesicular transport to underlying antigen presenting cells (APCs) such as DCs and macrophages1,64,102,103. In PPs antigen is then presented by APCs to both immature B and T cells1,64,103. Activated T cells preferentially differentiate into CD4+ T helper cells that in addition to DCs and cytokines induce the differentiation of antigen-specific B cells into IgA-committed plasmablasts13,15,24,104,105. PP derived IgA+ plasmablasts further proliferate and differentiate while migrating through the MLN and thoracic duct into the blood circulation until homing to mucosal effector sites (LP), where they complete their terminal differentiation into IgA-secreting plasma cells (PCs)15,17,22-24,99,101. Contrary to inductive GALT compartments and MLNs, at gut effector sites, antigen- experienced lymphocytes accumulate and persist as committed regulatory T cells, memory B and T cells or terminally differentiated IgA-producing PCs6,16,17,93 (Figure 1). The gut LP harbors the largest population of mature IgA+ PCs, with IgA producing PCs making up ~30-40% of all mononuclear cells. It is estimated that in mice and humans approximately 80% of the total PCs secrete IgA, making it the most dominant antibody isotype in the gut22,24,94,101. However, minor populations of IgM and IgG secreting PCs are also present in the human LP and are also found at low numbers in mice83,106-109. In contrast, increased numbers of IgG+ PCs in the gut are typically associated with intestinal inflammation such as inflammatory bowel disease (IBD) and dysbiosis46,78,83,108,110. Based on the high IgA levels predominantly produced in the SI LP, suggests that PPs, in addition to other GALT sites and the MLNs, may represent the major inductive sites of intestinal IgA responses15,22,24,65,93,94,111.

Figure 1: Distribution of bacterial load and organization of the gut immune system throughout the intestine. Schematic distribution of gut mucosal inductive compartments (GALT) including PPs, ILFs, the ceacal patch and the gut-draining mesenteric lymph nodes (MLNs). Naïve B and T cells are primed in inductive GALT compartments or the mesenteric lymph nodes (MLNs) and further differentiate while migrating through lymph, the thoracic duct and blood until homing to mucosal effector sites such as the lamina propria (LP) and intestinal epithelial cells (IECs), where they act as effector cells such as IgA-producing plasma cells or intra epithelial lymphocytes (IELs).

1.6 Origin of B cells and antigen-independent BCR diversification B cells originate from a population of pluripotent hematopoietic stem cells that differentiate into common lymphoid progenitor (CLP) cells112,113. The earliest B cell development and commitment to the B cell lineage occurs prenatally in the fetal liver, followed by continuing development in the bone marrow113. The classical ontogeny of B cell development in the bone marrow relies on complex differentiation processes, followed by transcriptional changes and selection, which control the release of mature naïve B cells into circulation that express a functional membrane bound B cell receptor (BCR)113-115. B cells in the bone marrow generate their recognition diversity to the vast amount of different antigen structures by undergoing rearrangement of the gene segments encoding for the BCR immunoglobulin (Ig) heavy (H) and light (L) chain variable regions. In this first antigen-independent diversification phase, B cell precursors in the bone marrow randomly assemble the exons encoding for the IgH and IgL chain variable regions from individual variable (V), diversity (D) and joining (J) gene segments into in frame transcripts, a process referred to as V(D)J somatic gene recombination116,117. The initiation of somatic recombination requires the activity of the recombination-activating gene (RAG) 1 and RAG2 endonuclease complex. This process is finalized by the non-homologous end-joining machinery, that joins broken DNA segments independent of extended homology116,118,119. Productive rearrangement and assembly of the V(D)J IgH and IgL chain exons in developing B cells enables the generation of a membrane bound BCR, which is expressed on the cell surface (Figure 2 A). Immature B cells typically express a first monomeric IgM-BCR on their cell surface that in association with the transmembrane proteins Igα and Igβ (CD79 a/b) forms a functional and mature BCR114,115,120-122 (Figure 2 A). The recombined IgH and IgL chains determine the unique antigen specificity of a newly formed BCR and the association with Igα/β allows downstream BCR signal transduction that further directs B cell fate114,115,121. Functional BCR signaling is required for further B cell maturation, survival and selection and the BCR must provide tonic signaling either spontaneously or via ligand interactions122,123. Immature IgM-expressing B cells emerging from the bone marrow, enter circulation and undergo final differentiation into mature naïve B cells in secondary lymphoid organs (SLOs), particularly in the spleen (SPL). There, alternative splicing of RNA-transcripts of H chain genes leads to the co-expression of a surface IgM/IgD-BCR. After leaving the SPL, peripheral mature naïve B cells characteristically express membrane bound IgM/IgD-BCR, CD45, MHCII as well as the early B cell markers CD20, CD19, CD40 and B220 (CD45R) that in mice indicates the commitment of CLPs to the B cell lineage113,122,124. Mature naïve B cells continue to circulate through blood and SLOs. In SLOs mature naïve B cells are ready to respond to antigen. B cells recognize cognate antigen through their membrane bound BCR99,114. Upon antigen binding the BCR cross-links and induces phosphorylation of the tyrosines within the immunoreceptor tyrosine-based activation motifs (ITAMs) in the cytoplasmatic domains of the BCR-associated transmembrane proteins Igα and Igβ (CD79 a/b) (Figure 2 A), which mediates intracellular signaling required for B cells proliferation and differentiation and initiates the internalization and processing of BCR-associated antigens99,125,126. BCR-associated antigen display on MHCII127 to CD4+ T cells typically leads to the differentiation of B cells into long-lived effector PCs that produce specific antibodies98,125. Additionally, B cells express toll-like receptors (TLRs) that belong to a group of pattern 11 recognition receptors (PRRs) and are able to recognize conserved microbial ligands such as lipopeptides (TLR 2), LPS (TLR 4, which is not expressed on human B cells), flagellin (TLR 5) or bacterial CpG-DNA (TLR 9). TLRs are a family of conserved innate immune receptors and signal (except TLR3 that uses the TRIF adapter protein) through the MYD88 adapter protein120,128,129. TLR signaling does not result in antigen internalization but leads to the secretion of cytokines and chemokines130,131. However, sufficient cross-linking of the BCR in addition to TLR co-stimulation might drive the proliferation of activated B cells into short-lived PCs without further co-stimulatory signals and cytokines needed120,129. Depending on the mode of antigen encounter, availability of co-stimulatory signals and T cell “help” determines whether activated B cells differentiate into short-lived PCs through extra-follicular differentiation132-134 or become GC-precursor cells that further undergo T cell dependent affinity-maturation resulting in long-lived antibody producing PCs97,98,135,136. PCs thus represent the terminal differentiation step of conventional bone marrow derived B cells, constituting the only cell type able to produce antibodies22,101,114,134,137 (Figure 2 B). Based on the induction mode and the mucosal effector compartment, local PCs then secrete antibodies belonging to either of the five isotypes (IgA, IgD, IgE, IgG and IgM)138-140.

1.7 Structure of antibodies in mice and men Antibodies play an integral role of the adaptive humoral immune system14,16,94,141,142. Antibodies are the secreted copies of the BCR. Structural differences between the membrane bound BCR and its secreted antibody form are merely located at the carboxy-terminal protein tail115, respectively (Figure 2 A, B).

Figure 2: Membrane bound BCR and secreted antibodies. (A) B cells express multiple copies of monomeric isotype-specific surface immunoglobulins (Igs) (black box) that in association with the Igα and Igβ (CD79 a/b) heterodimer (blue and orange) constitutes the B cell receptor (BCR) (dashed box). The first expressed cell-surface BCR-isotype by mature naïve B cells is the monomeric IgM isotype113,122. The membrane bound B cell receptor (BCR) is expressed as a transmembrane receptor protein that contains an outer antigen binding moiety and an integral membrane domain (green) at the carboxy-terminal part. The Igα and Igβ (CD79 a/b) heterodimer subunits are signaling molecules that

12 span the plasma membrane and the cytoplasmatic tails contain immunoreceptor tyrosine-based activation motifs (ITAMs)115. The Igα and Igβ (CD79 a/b) heterodimer constitutes the signal transduction moiety of the BCR complex that mediates downstream intracellular signaling via tyrosine phosphorylation within the the Igα and Igβ (CD79 a/b) cytoplasmatic ITAM domains upon cognate antigen binding to the outer membrane antigen-binding Ig domain of the BCR115,126. Cognate antigen BCR binding results in B cell activation, internalization and processing of BCR-bound antigen125. (B) Antibodies (black box) constitute the secreted form of the BCR (dashed box) and are produced by plasma cells (PCs) that are terminally differentiated B cells and act as adaptive humoral modulators of the mucosal immune system by binding to antigens (Figure adapted from115).

Similar to the BCR, secreted antibodies consist of two regions, the Ig Fc (constant fragment) and the Ig Fv (variable fragment) part. The Fc-domain defines the antibody isotype and governs biological effector functions of the antibody molecule at distinct compartments in the body143-145. The Fv is unique to each B cell and consists of two identical antigen recognition sites that can bind cognate antigen. Structurally, antibodies are made up of four polypeptide chains, two identical heavy (H) and two identical light (L) chains (either κ or λ) that are covalently linked by disulfide bonds. The stems of the two H chains form the Fc-fragment but also the C-tail region of the two L chains contribute to the Fc-fragment of the antibody. The CL and VL domains together with the VH domains form the antigen binding part (Fab) of the antibody molecule (Figure 3). In humans and mice all antibodies belong to one of the five antibody isotypes (IgA, IgD, IgE, IgG and IgM)139,140 with differing prevalence for each isotype (Figure 3). Of these isotypes, IgA and IgM are prevalent in mucosal secretions13,15,83. These, so-called secretory antibodies are expressed as polymers (mostly dimeric IgA, dIgA and pentameric IgM, pIgM)13,94,109. In both humans and mice, IgA production is without compare, exceeding the prevalence of all other isotypes combined15,22,29,146. Most of the produced dIgA is present at mucosal surfaces in particular in the gastrointestinal tract, in contrast to monomeric IgA found in serum15,83,109,143,147. A typical feature of intestinal dIgA is its high resistance to proteolytic degradation91,143,148. The mouse intestine is largely devoid of pIgM and monomeric IgG, whereas in the human intestines pIgM and IgG are also prevalent46,78,107,108,149- 151. Moreover, while the mouse gut only harbors one IgA isotype, humans express two IgA isotypes, IgA1 and IgA2 that differ in their biological features143,145,148. These distinguishing features of IgA1 and IgA2 determine their abundance throughout the intestine15,104,143,145. IgA1 is prominently found in the proximal small intestine and characteristically contains a hinge region that is easily cleaved by bacterial proteases, whereas IgA2 lacks this hinge region and is therefore more resistant to proteolytic cleavage15,143,148. This may explain the higher prevalence of secreted IgA2 in distal small intestine and colon15,93,145.

Figure 3: Molecular structure of antibody molecules and antibody isotypes. Antibodies (black box) are produced by PCs and constitute the secreted copies of the monomeric membrane bound BCR. The secreted monomeric antibody molecule (dashed box) consists of four polypeptide chains, two identical heavy (H) and two identical light (L) chains that are covalently connected by disulfide (SS) bridges. Both the H chains and the L chains consist of constant (CH, CL) and variable (VH, VL) domains. The VH and VL domains of both H and L chains constitute the variable (V) region and the CH and CL domains form the constant (C) region of the antibody. The VL and CL domains of the L chains together with the VH domain of the H chains contribute to the antigen-binding fragment (Fab). A monomeric antibody molecule has two identical antigen recognition sites that can directly bind cognate antigen (red). The Fab is unique to each mature naïve B cell, facilitated by the somatically recombined gene segments of the V(D)J elements of the V region. The CH domains of the H chains constitute the cristallizable fragment (Fc) that determines the antibody isotype and governs in addition to the Fc-decorating glycans the biological effector functions of the antibody e.g. Fc-receptor binding capacity and the prevalence of respective isotypes in functional locations. Plasma cells produce isotype-specific antibodies that belong to either of the five isotypes found in mammals (IgM, IgG, IgD, IgE and IgA). Of the five isotypes IgM is secreted as a pentamer (pIgM) and IgA mostly as a dimer (dIgA) and the single molecules connected tail-to-tail by the polypeptide joining (J, blue) chain. IgG, IgE and IgD are secreted as monomers (Figure adapted from139).

1.8 Luminal transport of intestinal antibodies In order to control infections but also to maintain homeostasis at gut mucosal sites, relevant quantities of antibodies need to be actively transported to intestinal mucosal surfaces. Polymeric IgA is transported via the polymeric Ig receptor (pIgR) across the mucosal epithelial barrier94,104 (Figure 4). PIgR is constitutively expressed as a transmembrane glycoprotein on 14 the basolateral side of epithelial cells in the small intestine (duodenum, jejunum) and colon94. Characteristically, polymeric antibodies contain joining (J) chains, which are highly conserved proteins connecting the Fc-parts of monomeric antibodies tail-to-tail, essential for the generation polymeric antibodies and enabling the trans-epithelial transport into the gut lumen152. For pIgR-mediated transport the J chain of polymeric antibodies (pIgs) becomes covalently bound to the secretory component (SC), a fragment of the pIgR23,94,153 (Figure 4). The complex of pIgs and pIgR is internalized via receptor-mediated endocytosis and transported to the apical side of polarized epithelial cells for exocytosis101,154. Following exocytosis at the apical membrane, the extracellular portion of the pIgR-pIg complex is subjected to proteolytic cleavage and leads to the release of secretory antibodies (SIgs) into the gut lumen. Free SC is also found in the gut lumen, resulting from transcytosis and cleavage of dIgA-unbound pIgR94,155. Similarely to dIgA is pIgM also ransported via the pIgR into the gut lumen94,104. In addition, mammary gland epithelial cells have been reported to express pIgR156. In fact, during lactation large quantities of dIgA are delivered into breast milk via pIgR-mediated transport and antibodies passively transferred to neonates38-41. Passively received maternal antibodies protect the infant gut against infections, help to establish the composition of beneficial commensals in the neonatal gut and bridge the period until a fully mature immune system is established36,38,39,157. Notably, large quantities of IgA have also been detected in bile of both mice and humans158-161. Thus, bile-derived IgA may contribute to luminal SIgA in the proximal small intestine162. However, in contrast to murine hepatocytes is the expression of pIgR on human hepatocytes still controversial, but the presence of IgA in human bile suggests pIgR-expression at least in bile duct associated tissue158,159. Conversely to polymeric antibodies, gut-luminal transport of IgG is not well defined, however, potential passive leakage through the epithelial layer, passage from plasma IgG into local secretions, or Fc-neonatal receptor (FcNR) mediated transport have been suggested149,163-165.

Figure 4: Polymeric immunoglobulin receptor (pIgR) mediated secretion of dimeric IgA (dIgA) into the gut lumen as SIgA. (1) Dimeric IgA (dIgA) is produced by lamina propria (LP) resident PCs, dIgA consists of two covalently bound monomers through disulfide bonds to the J chain (blue). (2) The polymeric Ig-receptor (pIgR) is expressed on the basolateral site of gut epithelial cells. (3) The extracellular portion of the pIgR known as secretory component (SC, green) recognizes and covalently binds to the J chain of dIgA. (4) Association of the SC part and dIgA-J chain results in SC conformational changes and final binding. (5) Final binding of the SC to dIgA results in internalization of pIgR-dIgA-J chain complexes and the complex is transcytosed. (6) Endocytosed vesicles containing the pIgR-SC- dIgA complexes are transcytosed in endosomes to the apical site of the cell. (7) During exocytosis at the apical site of the cell membrane, the SC-fragment is proteolytically cleaved from the pIgR-SC-dIgA complex. (8) The cleaved SC remains associated with dIgA and released as secretory IgA (SIgA) into the gut lumen. Also free SC of dIgA-unbound pIgR-SC is released into the gut lumen. Uncleaved pIgR- SC complexes are recycled. Secretory antibodies are formed by the combined function of PC-produced dIgA and epithelial cells expressing pIgR-SC (Figure adapted from155).

1.9 SIgA coats members of the intestinal microbiota SIgA in the gut lumen typically targets and clears pathogens, but also coats members of the gut microbiota under homeostatic conditions11,15,109,166. In fact, in mice and humans, a relevant proportion of the microbiota is constantly coated by endogenous SIgA, suggesting a continuous IgA response to the microbiota11,20,167-169. Flow cytometry analysis confirmed that ~10 - 60% of the overall gut bacteria are bound by endogenous SIgA, ranging from ~60% coating of total bacteria in the small intestine to ~10 - 20% in feces11,20,111,168,169. Thus, SIgA- coating of bacteria is substantially higher in the proximal small intestine compared to the colon, despite the higher bacterial load in the latter. Interestingly, although commensals in both the small and large intestine are coated by SIgA, only colonic bacteria that are also present in the small intestine are SIgA-coated in the colon11,20. This suggests that SIgA-coating influences the microbiota composition within different gut compartments and that IgA responses may predominately be elicited in the small intestine11,93,109. Moreover, high throughput 16S sequencing of SIgA-coated and non-coated bulk fecal bacteria revealed that distinct groups of gut bacteria are enriched in the SIgA-coated fraction compared to non-coated bacteria11,12,20,35,168,169. Here, SIgA-coated fractions showed prominent enrichment for bacteria of the genera Bifidobacterium, Proteobacteria, Bacteroides, Prevotella, Flavobacterium and Clostridium, in accordance with previously described common microbial targets of human intestinal IgA including the genera Ruminococcus, Clostridium, Roseburia and Blautia35,107,169,170. However, IgA-targeted commensals are still poorly characterized and the mechanisms by which they induce specific and context-dependent IgA responses are not fully clear. Considering the opportunistic potential of some commensals, there is a pressing need for understanding how the SIgA-microbiota system controls host-microbial homeostasis. Thus, the question still remains what determines which microbiota are targeted by SIgA and what defines a healthy microbiota on a population level. While some studies highlight the therapeutic potential of SIgA-coated microbiota derived from healthy donors to ameliorate disease states after complete fecal microbiota transfer, others recently outlined the causality for SIgA-coated microbiota with disease development and exacerbated pathologies9,109,166,171,172. Considering the latter, transfer of SIgA-coated bacteria from IBD donors or donors with nutritional deficiencies into germ-free recipient mice led to severe pathology and enhanced diet- dependent enteropathy as compared to animals colonized with non-coated bacteria170,172. Conversely, the relevance of endogenous SIgA-microbiota interactions was demonstrated by decreased microbial diversity and strong microbial species variability in mice with impaired IgA-responses111,173-175. Likewise, the complete lack of IgA in B cell deficient mice correlated with the systemic invasion of pathobionts and overgrowth of mucus associated bacterial species such as Desulfovibrionaceae, Mucispirillum sp. and Ruminococcus sp., emphasizing the crucial role of SIgA in controlling and shaping microbial communities176. Similarly, studies performed in pIgR- or J chain deficient mice, which lack luminal transport of dIgA, resulted in an overgrowth and systemic dissemination of disease-promoting Enterobacteriaceae and decreased neutralization of bacterial toxins associated with strong systemic inflammatory immune responses and increased disease susceptibility14,109,177-179. Collectively, SIgA is considered a key molecule for maintaining intestinal homeostatic interactions between the gut microbiota and host-immunity at mucosal surfaces15,23,91,180. 17

However, the precise mechanisms and the identity of relevant microbial antigens that induce microbiota specific IgA responses in the healthy gut are not entirely known.

1.10 Effector functions of intestinal SIgA In contrast to the classical functions that come to mind when thinking of antibody responses such as opsonization, complement activation and antibody mediated cellular toxicity, the functions of SIgA antibodies present in the intestinal lumen per se must be different considering the absence of other (phagocytic) immune cells and almost complete lack of complement in the gut lumen. Hence, one might need to reconsider these classical functions described for antibodies when looking at functional consequences of SIgA in the gut. Considering the phenomenon of constitutively SIgA-coated microbiota under steady state conditions, it is however not fully clear how SIgA-coating impacts and shapes the intestinal microbiota and what mechanisms enable such maintained interactions. Conversely to the lacking mechanistic insight of IgA affecting the microbiota, direct effects of SIgA regulating the intestinal ecosystem to protect the gut from infections and inflammation have been described in more detail for IgA- responses to pathogens and toxins. Moreover, intestinal IgA responses to pathogens are typically specific and require Fab-mediated binding to cognate antigen, whereas SIgA- microbiota interactions may involve both canonical and non-canonical binding. While canonical interactions rely on Fab-mediated binding to antigen, non-canonical SIgA-microbiota interactions are mediated by any other moiety of the antibody, but the Fab-domain25,29,91,180,181. In the context of infection immunity, the primary function of SIgA at gut mucosal surfaces is to facilitate the entrapment of microbes and foreign antigens in the mucus layer or lumen, thereby preventing direct contact to the epithelial surface – a process known as immune exclusion25,29,155,180. Immune exclusion by SIgA prevents epithelial damage by reducing the adherence/attachment to and penetration of the gut epithelium that could otherwise lead to local penetration of pathogens with subsequent systemic dissemination17,23,29,180. Mechanistically, immune exclusion is thought to act through agglutination, that is cross-linking of antigens such as viruses, bacteria or foreign proteins leading to the luminal clearance of SIgA-entrapped immune aggregates23,29,91. Additionally, a recently suggested alternative mode of SIgA-mediated immune exclusion is enchained growth, which involves cross-linking of dividing planktonic bacteria demonstrated for enteric Salmonella sp.182,183 (Figure 5). Other functions of SIgA, in addition to SIgA-mediated immune exclusion, describe impaired bacterial fitness via SIgA-binding to bacterial surface molecules such as flagella that abrogates motility and thus virulence of motile bacteria by impairing penetration of the gut epithelium141,180,184,185. Moreover, SIgA-binding to bacteria can modulate the expression of virulence genes and modulate metabolic functions associated with pro-inflammatory outcomes175,180,186,187. The formation of pIgR-dIgA-antigen complexes leads to removal of pathogens breaching into the LP through transcytosis into the gut lumen. SIgA can furthermore inhibit endotoxin activity and neutralize pathogens or viruses within epithelial cells (e.g. viral replication)90,141,179,180,188. Finally, SIgA-opsonization of pathogens enhances the luminal uptake through M cells, which promotes increased IgA-responses to sampled pathogens29,99,189 (Figure 5). Thus, intestinal IgA performs antigen exclusion and (intracellular) neutralization of toxins and viruses without causing tissue damage14,184. 18

While the mechanisms of SIgA-pathogen interactions are relatively well described, there is comparable little known about the mechanisms of how SIgA regulates the establishment and stability of the gut microbiota under homeostatic conditions. Understanding these mechanisms and functional consequences of SIgA-microbiota interactions in steady state have thus become a great research interest. It has therefore been proposed that in addition to pathogen clearance, IgA responses play a key role in actively regulating host-microbiota symbiosis25,27,91,161,180,190. Some mechanistic concepts of how IgA can do so, may indeed relate to SIgA-pathogen interactions, although with different outcomes for commensals. For example, SIgA-mediated cross-linking of commensals may have the opposite effect to immune exclusion in the context of SIgA-microbiota interactions. Here, SIgA-mediated enchainment of beneficial microbial members may promote immune inclusion/retention and thereby facilitate colonization of commensals in distinct niches (Figure 5). SIgA-dependent bacterial clumps in this scenario may be less likely to be expelled from mucus. This mode of SIgA-microbiota interactions may benefit low abundant or less competitive beneficial species. Indeed, SIgA-coating can lead to increased commensal adherence to epithelial cells, which has been described for Bifidobacterium lactis, and Lactobacillus rhamnosus191,192. Similarly, commensal retention rather than exclusion has also been described for species that are closely associated with the mucosal tissue such as segmented filamentous bacteria (SFB), flagellated members of Enterobacteriaceae and mucus degrading bacteria such as Akkermansia muciniphila9,11,109,170,193. Recently, it was described that the commensal species Bacteroides fragiles exploits SIgA-binding to competitively promote its establishment in the gut194, suggesting that some commensals can co-opt SIgA-binding to establish mucosal communities in a defined niche, thereby providing colonization resistance to exogenous competitors and to engender stable host-microbiota symbiosis. Notably, SIgA is also heavily glycosylated and thus provides additional non-canonical binding sites e.g. for bacterial lectin-like adhesins25,147,195,196. Such glycan-mediated SIgA-microbiota interactions have been shown for bacteria as well as viruses25,181,195-197. Accordingly, specific SIgA-microbiota interactions might not always be required for the modulation of the microbiota. Making use of high SIgA concentrations in mucus have been described for Bacteroides fragilis that effectively degrades SIgA-associated glycans and may use these glycans as a nutritional carbon source50,198-200. Moreover, SIgA glycan-mediated binding has the capacitiy to alter the gene expression186,187,193 e.g. of polysaccharide utilization loci in Bacteroides thetaiotaomicron, which in turn facilitated symbiotic colonization between Bacteroides thetaiotaomicron and beneficial mucus- embedded microbiota of the Firmicutes phylum201 (Figure 5). Whether the higher concentration of SIgA in mucus lead to increased binding of tissue/mucus associated bacteria or reflects a biased response towards these microbial species is not fully clear. Collectively, the process of immune exclusion may provide an efficient way to regulate the prevalence of highly abundant species, thereby also generating niches that facilitate the colonization of beneficial but low abundant or less competitive microbial members29,91,180,183,194,202. In addition, both canonical and non-canonical SIgA-microbiota interactions can have either pro-colonizing effects or counteracting pathogen challenges25,29,180-183,194,203. Defective or impaired IgA responses in the gut can lead to global shifts in microbial composition, associated with inflammatory responses and higher disease 19 susceptibility9,67,74,111. Thus, both SIgA-mediated pathogen clearance as well as SIgA- microbiota interactions are critical in maintaining intestinal integrity and regulating the diversity of beneficial microbial members in the gut27,29,91.

Figure 5: Effector functions of SIgA in the gut. (a) SIgA-coating of luminal pathogens enhances M cell mediated sampling of antigens/bacteria into GALT e.g. Peyer’s patches. (b) The endocytosed dIgA-J chain and pIgR complex mediates intracellular neutralization of epithelial cell invading pathogens such as viruses. Dimeric IgA in the lamina propria (LP) mediates removal of pathogens and antigens from the mucosal LP by pIgR mediated excretion of antigen-dIgA complexes into the gut lumen. (c) SIgA in the gut lumen binds to pathogen-or microbial-(derived) antigens leading to clump formation to prevent or inhibit attachment to or invasion of epithelial cells. (d) SIgA-binding to flagellated bacteria can directly impair their motility and fitness and thus impacts virulence factors of invading pathogens by reducing the attachment to epithelial cells. Also, SIgA-binding can lead to the enchainment of dividing bacteria that prevents bacteria separation and results in growth-dependent clump formation that can be expelled from the lumen. (e) Binding of SIgA to capsular polysaccharides can modulate bacterial gene expression leading to the downregulation of epitopes thereby stabilizing niche occupancy. (f) Mucus embedded SIgA-binding to bacteria anchors bacteria within the mucus layer, thereby stabilizing bacterial communities in a distinct intestinal niche (Figure adapted from12,155).

1.11 T cell dependent and T cell independent generation of intestinal IgA Both, T cell dependent and T cell independent pathways have been described for the generation of IgA producing PCs in the gut11,18,20,27,37,95,204-206. In the gut, IgA responses directed against pathogens as well as some microbial members are generated via T cell dependent pathways, resulting in affinity-matured and specific antibodies to cognate antigen67,111,136,174,205,207. Specific IgA responses occur when conventional bone marrow derived immature naïve B cells from circulation enter gut immune-inductive compartments, where they encounter cognate antigen that leads to a first activating signal induced by the antigen-dependent cross-linking of the BCR. BCR cross-linking upon cognate antigen binding leads to BCR-associated antigen internalization and processing for antigen peptide display on MHCII127 to CD4+ T cells at lymphoid follicle borders99,134. The engagement with T helper cells provides co-stimulatory signals via CD40-CD40L interactions and cytokines required for T cell dependent GC reactions, where activated B cells undergo affinity-maturation and thereafter differentiate into long-lived IgA antibody secreting PCs or IgA committed memory B cells. T 20 cell dependent affinity-matured IgA responses require T cell “help” and GC engagement of primed B cells. Affinity-matured antibody producing PCs in the gut typically carry high numbers of somatic mutations and are class switched27,93,106,205,208-210. It is therefore suggested that in adult humans and likely also in mice the majority of IgA-microbiota interactions rely on affinity- matured and antigen specific responses to distinct microbial members27,91,106,136,168,209. However, in the gut controversy comes from T cell independently generated IgA responses with microbiota binding capacity18,20,27,190,205-207,211,212. It is still debated how and where such IgA responses are generated and whether they constitute functional relevance in maintaining host- microbiota homeostasis. T cell independent responses do not require GC formation and produced antibodies are typically germ-line or close to germ-line encoded, hence do not carry somatic mutations and do not show hallmarks of affinity-maturation11,20,91,204,205. T cell and GC independent pathways usually result in un-switched IgM antibody responses occurring early during adaptive immune responses that are less-specific and of relatively low affinity11,133,213. However, in the gut T cell and GC independent pathways can curiously elicit class switched IgA antibodies with microbiota reactivity11,37,95,133,204-207,214. Moreover, T cell independent microbiota reactive IgA responses are often associated with polyreactivity - that is promiscuous binding to structurally unrelated antigens11,213,215. Indeed, recent studies demonstrated the presence of such T cell independent microbiota-reactive IgA responses in young mice and human fetuses, where the microbiota-binding capacity of IgA did not rely on somatic mutations and affinity-maturation, but was instead associated with polyreactivity/autoreactivity11,20,137,216. It has been proposed that in the gut, T cell independent IgA generation may occur in both organized lymphoid tissues and non-lymphoid tissues206,214,217. GALT sites such as ILFs but also mucosal effector sites such as the SI LP have been suggested to support T cell independent IgA generation95,206,217,218, whereas IgA class switch recombination (CSR) in the SI LP is highly debated67,205,207. Albeit, there is no direct evidence for IgA CSR in the gut lamina propria, this discrepancy has been difficult to solve due to the numerous ILF structures that contain both naïve and IgA+ B cells, which likely obscure LP preparations93,95. Furthermore, in the context of T cell independent IgA responses, TLR co-stimulation in addition to BCR cross- linking or B cell activating factor (BAFF) and proliferation inducing ligand (APRIL) expressed by DCs and likely other cytokines derived from ILCs may be sufficient for alternative mechanisms that promote extra-follicular CSR to IgA24,102,120,130,205,219,220. BAFF and APRIL are soluble B cell stimulating factors that are structurally and functionally related to CD40L and interact with transmembrane activator and calcium-modulating cyclophilin-ligand interactor (TACI) receptor and BAFF-receptor (BR3) expressed on B cells214,220,221.

Alternative generation of intestinal IgA responses As we have outlined before, for high-affinity and specific IgA responses T cell dependent CD40-CD40L signaling is required64,207. However, in both mice and humans neither CD40 nor CD40L deficiency completely abrogates intestinal IgA levels205,207, suggesting that T cell independent pathways of IgA induction may partly compensate for the lack of T cell dependent IgA responses11,12. DC mediated BAFF and APRIL signaling may potentially compensate for the lack of CD40-CD40L interactions, as IgA CSR is impaired in APRIL- and APRIL-receptor deficient (TACI) mice220,222. In mice, B1 B cells typically do not undergo SHM or affinity- 21 maturation and are believed to secrete “primitive” antibodies18,212,223,224. B1 B cells derive from a self-renewing fetal liver stem cell population and are typically found in the peritoneal and pleural cavity. IgA derived from B1 B cell is described as the “innate” form of IgA, due to the lack of somatic mutations and affinity-maturation18,212,225. The precise location and factors required for T cell independent IgA CSR in B1 B cells are still debated, but a role for the DC derived cytokines BAFF and APRIL, RA and ILC involvement have been suggested18,225,226. Also, BCR-TLR signaling in synergy has been proposed as alternative mechanism for AID expression enabling IgA CSR or to boost specific antibody responses120,129,130. Typically, B1 B cells are locally activated independently of T cell help but they may still home to the intestine upon TLR activation. TLR ligands present in the serum have been shown to initiate the recruitment of B1 B cells from the peritoneum to the murine gut224,227. B1 B cell derived IgA can also target commensal bacteria and potentially may act as a compensatory mechanism for delayed T cell dependent antibody responses by targeting a broader bacterial range albeit with lower affinity15,18,20,212,225,228. Such low affinity antibody responses may mediate immune exclusion of commensals, but only provide limited protection against invading pathogens155,227,229,230. Although, combined TLR-mediated and BCR-induced immune responses may contribute to microbial recognition and antibody production, additional signals yet seem to be required for robust antibody responses to the microbiota27,136. Unlike in mice, there is little evidence for B1 B cells in the human gut231. Although, humans seem to lack B1 B cells, T cell independent low-affinity IgA responses with microbiota binding capacity have been reported83,214,232,233. Evidence for T cell independent IgA responses in humans comes from patients with severe HIV associated CD4+ T cell deficiency or patients lacking CD40 that still show class switching to IgA, however, with less diverse IgA repertoires214. In humans, other B cell subsets might be involved in T cell independent IgA responses, possibly including IgM+ memory B cells and transitional B cells that express germ-line encoded polyreactive antibodies227,233-235. Although, both mechanisms have been described to result in IgA antibodies with microbiota binding capacity, T cell independent IgA antibody responses likely have functionally different properties and affinities to the microbiota12,27,67,92,111,206. Moreover, the importance and relative contribution of T cell dependent and T cell independent IgA responses contributing to host- microbiota homeostasis is not fully clear, but both mechanism potentially act in concert to facilitate effective microbiota control depending on the context and system looked at25- 27,91,180,192,211.

1.12 Generation of affinity-matured IgA responses in Peyer’s patches In the following paragraphs we will in more detail outline the steps required for the generation of T cell dependent IgA responses in PPs. Intestinal IgA responses generated in PP GCs elicit high-affinity and antigen-specific responses directed against toxins, pathogens and also members of the microbiota. This process relies on canonical - T cell dependent pathways and requires additional signals from DCs and FDCs. Notably, IgA+ B cells found in PPs clonally overlap with IgA+ PCs in the SI LP105,236, indicating that PPs, in addition to the MLNs and ILFs, indeed constitute major inductive sites for the generation of affinity-matured intestinal IgA+ PCs15,64,65,93,102,237. 22

1.13 How does luminal antigen reach follicles in Peyer’s patches? In order to mount intestinal adaptive immune responses, luminal antigens first need to be delivered to inductive gut mucosal compartments. The surface of PPs is covered by the FAE, which acts as an interface between the gut lumen and the underlying SED of PPs. The FAE is in contrast to the villous epithelium mainly composed of enterocytes, interspersed with M cells, and a limited number of goblet cells1,64,71,72,238,239. While several sampling mechanisms have been described for the transport of luminal antigens across the FAE, antigen uptake through M cells is considered to be critical for the on-set of mucosal immune responses29,239-241. In fact, M cell facilitated sampling of luminal antigens derived from bacteria, viruses, fungi, toxins, inert particles or immune complexes is required for antigen-specific intestinal IgA responses71,238,239,241,242. M cells take up luminal antigen by phago-or pinocytosis at their apical cell surface241,243, which results in the release of antigens to immune cell containing “pockets” on their basolateral membrane. This close contact with antigen allows antigen presenting cells (APCs) such as macrophages and DCs, and potentially also B cells240, to directly sample antigen from M cells, followed by antigen presentation to both T and B cells and antigen transport to FDCs within B cell follicles64,102,241. Luminal sampling by M cells is facilitated by the expression of certain receptors, such as specific lectins, Dectin-1 and GP2, that can bind bacteria or soluble antigens. Dectin-1 has been implicated in the up-take of glycosylated SIgA- antigen complexes and GP2 expression allows for the recognition of adhesins (e.g. FIMH), which are virulence factors predominantly expressed on gram-negative bacteria.173,189,238 In contrast to enterocytes, M cells do not express pIgR but are able to bind SIgA and SIgA-antigen complexes via Dectin-171,189,239, which mediates the transcytosis of SIgA-coated bacteria into the SED29,189. In addition to M cells, goblet cells may also provide a transit for antigens with subsequent delivery to APCs244. Also, intestinal DCs and macrophages themselves have been suggested to directly sample luminal antigens via dendrites extended between epithelial cells245. While the relative contribution of these various antigen sampling pathways is currently not known, a large proportion of continuously sampled antigens constitute microbial proteins and polysaccharides that in turn keep initiating B cell priming in PPs, thereby fostering the continuous differentiation of IgA producing PCs and memory B cells directed against gut antigens and the microbiota.

1.14 B cell encounter with antigen in Peyer’s patches B cells can recognize both membrane-bound and soluble antigens via their surface-expressed BCR, but membrane-bound antigens seem to be more effective in inducing B cell activation as compared to soluble antigens246,247. However, the exact process of B cell encounter with and uptake of antigen in PPs is still somewhat mysterious. Different types of antigen in addition to opsonized antigens can be bound and displayed by FAE associated specialized macrophages or DCs that can then be captured by cognate B cells248-252. Alternatively, complement receptors on non-cognate B cells might facilitate the uptake of opsonized antigen complexes64,251. Another, and perhaps more likely mechanism, is the transport of DC-sampled antigens into interfollicular regions with subsequent display to B cells99,102,250. It was moreover recently shown that B cells in PPs can indeed directly sample antigen from M cells in the SED independent of DCs240. B cells may also directly encounter free soluble antigen while migrating 23 through medullary regions253,254, and antigen encounter can be promoted by follicular dendritic cells (FDCs)64,252,255,256. FDCs (that are not related to conventional DCs) are specialized non- hematopoietic stromal cells that reside in lymphoid follicles and GCs256,257. FDCs characteristically carry intact antigen on their surface and have a crucial role for B cell immune responses as they serve as long-term reservoirs to display opsonized antigens252,256,258. There may be multiple mechanisms and pathways in PPs that enable B cell encounter with and uptake of antigen, which however depend on the location of B cells in PPs, the route antigen reaches the tissue and the physical properties of the antigen itself.

1.15 SHM and CSR in germinal centers require AID activity In PPs, the centers of B cell follicles constantly host chronically active GCs, which contain recently antigen-activated B cells as well as antigen-experienced memory B cells interspersed with a network of FDCs and TfH cells64,97,98. The initiation that directs B cells into GC reactions requires the coordination of different cell types, chemokines, cytokines and co-stimulatory signals64,97,259. In GCs, antigen-experienced B cells diversify their antibody repertoire and specificity through somatic hypermutation (SHM) and change their antibody isotype/effector function through class switch recombination (CSR), which results in IgA+ PCs that produce specific and high-affinity antibodies98,260 (Figure 6). B cells undergoing SHM with subsequent affinity-based clonal selection, leads to affinity-matured antibody responses over time, a process referred to as affinity-maturation64,97,98,261. GC reactions thus foster both the process of SHM and CSR entailing the generation of B cell clones with a broad range of affinities for the initial antigen, resulting in affinity-matured and class switched B cells that express highly selective and specific antibody repertoires with distinct effector functions64,97,98,136. Importantly, affinitiy-maturation and CSR are temporally but not mechanistically related processes. However, both of these genetic alterations (CSR and SHM) absolutely require the activity of activation-induced cytidine deaminase (AID), which is a DNA editing enzyme that is transiently expressed in activated (memory) B cells112,262. AID encoded by the AICDA gene belongs to the inducible (apolipoprotein B mRNA editing enzyme, catalytic component 1) APOBEC family of cytidine deaminases that introduce DNA mutations at high rates during cell division112,263-265. The activity of AID is regulated transcriptionally (e.g. through HoxC4 and NFkB), but also post- transcriptionally and post-translationally112. The AID dependent process of SHM introduces point mutations into the V(D)J exons, which constitute the antibody variable region of both the H and L chains, thereby stochastically enabling the molecular basis for high-affinity selection of antibody variants to specific antigens112,266. In contrast, the genetic alteration of antibody molecules by CSR enables the switch from the initial antibody IgM isotype to any other isotype, although in PPs, the majority of CSR results in IgA committed B cells (Figure 6). Thus, CSR involves the exchange of the initially expressed antibody constant regions μ (Cµ) and Cδ (encoding IgM and IgD isotypes in naïve B cells) to one of the other downstream CH exons266,267. The immunoglobulin locus contains multiple CH genes (Cγ, Cε or Cα) that encode antibody proteins with different effector functions and each CH gene is preceded (except Cδ) by so called switch (S) regions. S regions are highly repetitive sequences upstream of each CH gene providing substrates for AID activity112,119. Each S region is preceded by a short intronic exon and a promoter that initiates germ-line CH gene transcription upon exposure to 24 activating stimuli such as cytokines119. For example, TGFβ prominently activates CSR from IgM to IgA, whereas IL-4 leads to CSR to IgG1119. CSR involves a recombination event between two S regions, with the intervening DNA sequences being looped out and deleted. Mechanistically, CSR is a deletional recombination reaction, in which introduced dsDNA breaks at S regions are fused again and subsequently expressed with the rearranged V(D)J exons119 (Figure 6). In conclusion, SHM affects the specificity and affinity of the antibody Fab, whereas CSR allows the expression of antibodies that have the same antigen specificity but express secondary isotypes (IgG, IgE or IgA), which determine antibody effector functions143,267 and their characteristic location in the body15,83,145,148. While the processes of SHM and CSR are distinct in the antibody moiety they target, both of these processes only occur in antigen-stimulated B cells, require AID activity and B cell proliferation112,119 (Figure 6). Albeit in PPs, GC derived B cells have almost entirely undergone switching from IgM to IgA93,111 and carry numerous somatic mutations106,208,268, it is not completely understood why the environment in PPs particularly promotes CSR to IgA.

Figure 6: Genetic alterations and diversification of antibodies schematically displayed for a human antibody. Developing B cells undergo RAG1/2 induced antigen-independent V(D)J recombination. In the bone marrow B cells rearrange the gene loci (exons) encoding for the immunoglobulin (Ig) heavy (H) and light (L) chain variable (V) regions from variable (V), diversity (D) and joining (J) segments. Assembly of the recombined V(D)J segments in combination with the μ constant region leads to the expression of unique membrane bound isotype-specific IgM-BCR. In SLOs mature naïve B cell can undergo further antigen-driven Ig-gene diversification through somatic hypermutation (SHM) and class switch recombination (CSR) that both require the activity of the enzyme activation-induced cytidine deaminase (AID). Secondary isotypes are produced by CSR, a process exchanging the constant region of the heavy chains (CH) (except for Cδ) to another downstream constant-region gene set (white boxes) (CSR to IgA2 is shown). Cytokines stimulate CSR and determine 25 the isotype a B cell will switch to. The process of SHM introduces point mutation in the V region of both the H and L chains. Mutations in the V regions increase the genetic diversity of BCRs and determine the specificity and affinity of the antibody binding part (Fab). VH and VL regions contain complementary regions (CDR) that are hypervariable regions prone for replacement mutations providing the substrate for high-affinity antibody selection119,269 (Figure adapted from119,139).

1.16 Mechanisms leading to IgA CSR in Peyer’s patches Antigens captured by APCs in the SED are presented to CD4+ T cells in the interfollicular regions, thereby inducing effector T cell differentiation (Figure 7). Some of these T cells migrate to the border of the B cell zones, interact with primed B cells and differentiate into specialized T follicular helper (TfH) cells that release cytokines and provide co-stimulatory signals supporting CSR to IgA. Upon mature naïve B cell entry into PPs and BCR-antigen encounter within the follicles, B cells receive first activation signals99,247,248. In order to become fully activated, B cells migrate to the interface between the B cell and T cell zone, where they can present peptides from captured antigen on MHCII to cognate CD4+ T cells67,127,260. This interaction provides further activation and co-stimulatory signals via T cell secreted cytokines and CD40 and CD40L interactions (Figure 7). Long-lasting interactions between B cells and CD4+ T cells lead to their differentiation into TfH cells that can migrate into the B cell zone and initiate a GC reaction97,100,260,270. Thus, GC reactions rely on cognate interactions between antigen-activated B cells expressing CD40 and CD4+ T cells that express CD40 ligand (CD40L). T cell signals and co-stimulation (help) lead to the upregulation of AID in antigen- activated B cells required for CSR112,267. Here, several subsets of CD4+ T cells have been described to facilitate GC-dependent IgA CSR, including different Treg cell subsets and Th17 cells22,83,271. In addition to CD40-CD40L interactions, the combined signaling of IL-5, IL-6, IL- 10, TGFβ1 released from activated TfH cells are crucial for IgA CSR, but also retinoic acid (RA) and nitric oxide (NO) derived from gut DCs seem to contribute to IgA CSR64,102,272,273. In particular, the crucial role for TGFβ to initiate IgA CSR has been demonstrated in TGFβ deficient mice and similarly in mice deficient for TGFβ-receptor on B cells that virtually lacked the production of IgA+ B cells altogether138,273. Notably, key sources of TGFβ are B cells themselves64,138,271. The combined signaling of co-stimulatory signals and cytokines may enhance the production of TGFβ by B cells, triggering IgA CSR via an autocrine pathway273. Also, T cells, FDCs and DCs may produce TGFβ in PPs, respectively102,206. Although, T cell dependent CSR to IgA has been appreciated to primarily occur in PP GCs, the SED has recently been implicated to constitute a site for IgA CSR24,240,261 (Figure 7). Here, activated B cells that express CCR6 gain access to the SED, enabling the engagement with SED located DCs and TfH cells207,240,261. In the SED located DCs may express integrin αVβ8 that activates the latent form of TGFβ1 and potentially along with TfH cell signaling may promote pre-GC stage IgA CSR of activated B cells102,240,261,262. While DCs and/or TfH cells present in the SED may support GC-independent IgA CSR240,261, AID expression in SED located B cells was however markedly lower as compared to AID expression in GC-engaged B cells, suggesting that GC-dependent IgA CSR is more effective24,205,207,274. Accordingly, mouse models with impaired PP GC structures showed significantly reduced numbers of intestinal IgA+ PCs24,274. It has however been suggested that the SED might be an important site for the initial expansion of activated, CCR6-expressing B cells261. SED located TfH cells 26 potentially promote B cell expansion without eliciting clonal B cell competition261. Instead, clonal competition of B cells takes place when entering GCs, which depends on their relative BCR-affinity to a given antigen. Only B cells with high-affinity BCR specificities enter GC reactions to undergo SHM and affinity-maturation209,261,275. While GC reactions are required for the generation of high-affinity, long-lived memory B cells imminent for circulation or PCs ready for homing to effector sites, the presence of mutated IgA+ memory B cells and IgA+ plasma blasts in the SED may indicate migration between the SED and GCs240,261.

1.17 Mechanism promoting increased affinity of antibody responses Upon antigen encounter and subsequent T cell dependent GC formation, B cells upregulate AID and the transcription factor BcL6. Upregulation of the transcription factor BcL6 is essential for the initiation of GC reactions, as it allows B cells to enter follicles. In addition, BcL6 upregulates the expression of CXCR4, a chemokine receptor expressed by GC dark zone B cells97,270. Also, BcL6 downregulates sphingosine-1-phosphate receptor type 1 (S1PR1) and upregulates S1PR2 genes that in turn control the positioning of B cells in the follicle and GCs22,97,276. In contrast to S1PR1 that mediates B cell trafficking out of follicles into efferent lymph, the migratory inhibitory receptor S1PR2 promotes B cell confinement in GCs276. Moreover, a phenotypic hallmark for GC-engaged B cell is the expression of GL7105,236. The mature GC is spatially separated into the dark and light zone (Figure 7), which is largely organized by the expression of the chemokine receptors CXCR4 and CXCR5. High expression of CXCR4 is characteristic for dark zone B cells and low expression of CXCR4 for light zone B cells99,256,270. In contrast to the dark zone, which primarily contains proliferating B cells, the light zone of GCs harbors additional CXCR5 expressing TfH cells, FDCs and macrophages98,256,277. In the GC dark zone, B cells rapidly proliferate and undergo SHM leading to a genetically modified BCR within an interconnected network of reticular cells that characteristically express the CXC-chemokine ligand 1298,259,278. Following SHM in the GC dark zone, B cells migrate to the light zone, where newly generated high affinity BCRs are subjected to clonal selection, before undergoing further iterative rounds of SHM in the dark zone (Figure 7). Selection of high affinity over low affinity or autoreactive clones is the critical step in high affinity antibody generation and is facilitated by TfH cells and FDCs, the latter probing improved binding to an immunizing antigen98. Within GCs the role of the BCR is primarily to capture and internalize antigen for subsequent MHCII presentation to T cells97. GC patrolling TfH cells differentiate from CD4+ T cells initially engaged with B cells and form extended contacts with B cells capable of presenting large densities of antigen in MHCII complexes100. Given that these B cell - T cell interactions are yet transient, TfH cells are able to probe many B cells within in the light zone, thereby guiding the selection of BCRs with high affinities within a GC B cell population (Figure 7). B cells that present high densities of peptide-MHCII complexes of captured antigen acquire more T cell help, suggesting that competition for T cell help is the driving factor for positive high-affinity clonal selection98. Upon contact between TfH cells and GC B cells, TfH cells increase IL-4 and IL-21 expression, which stimulate B cell proliferation (Figure 7). Moreover, prolonged interactions are supported by the upregulation of ICOS (TfH cells) and ICOS ligand (B cells), which in turn stimulates upregulation of CD40L, providing enhanced CD40-CD40L signaling100,259,279. Along with the crucial role of TfH cells for the 27 selection of high-affinity B cell clones, FDCs provide additional selection pressure to probe the affinity of B cells. FDCs residing in the GC light zone store antigen on their surface for long- term display to B cells, thereby contributing to the optimal selection of antigen-specific B cells (Figure 7). Recently, it has been suggested that FDCs fine tune affinity selection of B cells by limiting the access to antigens deposited on FDCs. This process involves masking of FDC- presented antigens with antibodies that were generated during earlier GC reactions. This likely enables only those GC B cells that express BCRs with high affinities to successfully gain competitive access to antigen on FDCs280, which facilitates the iterative generation of progressively selected B cell clones with high-affinity BCRs (Figure 7). Antigen masking on FDCs and inter-GC patrolling by TfH cells has also been proposed to foster inter-GC competition and selection of B cells during recirculation between different GCs105,236,280. Thus, the integration of several signals by TfH cells and FDCs provide crucial signals to B cells, which facilitate their selection and instruct high affinity B cell clones to recirculate between light and dark zone to undergo multiple rounds of divisions and SHM. Positive selection of high-affinity B cell clones promotes re-entry to and division within the dark zone that is controlled by TfH cells located in the GC light zone98. Re-circulation between the GC dark and light zones mediates efficient rounds of affinity-maturation, in order to support expansion of the highest- affinity B cell clones in a relatively short time (Figure 7). Additionally, cyclic re-entry between dark and light zones maintains the GC reaction over time, assuming constant and sufficient antigen availability64,105,236. Although, new waves of IgA+ PCs and memory B cells are constantly generated in response to the microbiota, the mechanisms regulating the decision to either cease re-entry into GCs or to commit to a memory B cell or PC fate are not fully understood.

Figure 7: Schematic overview of IgA induction in Peyer’s patches. (1) Mature naïve IgM+/IgD+ B cell enter PPs from blood circulation and are able to encounter and recognize luminal-sampled antigen via their BCR typically in interfollicular T cell zones or in the subepithelial dome (SED). (2) B cells recognize both soluble and processed membrane-bound antigen displayed by follicular dendritic cells (DCs), which leads to BCR-associated antigen internalization, processing and display of antigen-derived peptides on MHC II. Activated B cells either migrate to the border of lymphoid follicles where antigen- presentation on MHC II generates interactions with primed CD4+ T cells (3a) or the presence of sufficient alternative co-stimulatory signals likely provided by T follicular helper (TfH) cells and DCs may lead to IgA CSR at a pre-GC stage in the SED (3b). Cognate interactions of B and T cells can either lead to the extra-follicular generation of short-lived IgM producing plasma cells (PCs) and memory B cells (4) or cognate B and T cells form long-term interactions supported by co-stimulatory CD40-CD40L interactions and cytokines enabling the migration into germinal centers (GCs) (5). The GC is spatially divided into the dark zone (DZ) constituting proliferating B cells undergoing somatic hypermuation (SHM) that genetically alters the affinity af the BCR (6). New BCR configurations are probed in the GC light zone (LZ), where antigen displayed on the surface of FDCs provides the basis for affinity selection (7). Affinity- selected B cell clones receive further activation and proliferation signals from TfH cells, encompassing the expression of a secondary antibody isotype via class switch recombination (CSR) (8). Controlled by LZ TfH cells, positively selected B cell clones with improved antigen binding can re-engage in DZ clonal expansion and SHM (9), whereas unfavorable BCR configurations are not supported and lead to apoptosis. (10) Affinity-matured B cell clones exit the GC reaction as IgA committed plasma blasts or memory B cells. Blasma plasts further differentiate into long-lived, high-affinity IgA-producing PCs (Figure adapted from83).

1.18 Plasma blast homing to the intestinal lamina propria Following the process of secondary antigen-dependent BCR diversification (SHM and CSR) in PPs, IgA committed plasma blasts recirculate through lymph and blood to eventually home to effector sites. The final differentiation process of B cells into PCs requires cytokines, such as IL-21 provided by TfH cells, and is further characterized by the up-regulation of the transcription factors IRF4 and Blimp-1 that are required for PC formation and antibody secretion281,282. Moreover, PP resident DCs are crucial for promoting the effector phase of mucosal humoral responses by releasing RA that upregulates the expression of gut homing factors, CCR9 and α4β7-integrin on IgA-switched plasma blasts237,272. IgA-committed plasma blasts exit PPs through efferent lymphatic vessels, re-enter the bloodstream and subsequently home either to the small intestine directed by the gut homing receptor CCR9 or to the large intestine via CCR10145,237,272,274,283. Upon circulation the interaction between α4β7, CCR9 and CCR10 and their ligands, that are mucosal vascular cell adhesion molecules, such as MAdCAM-1, CCL25, and CCL27 and CCL28 direct migration into the intestinal LP. During this migration process, plasma blasts fully differentiate into matured IgA+ PCs, a process driven by cytokines released from lymphoid cells, DCs as well as stromal-and epithelial cells. Once PCs have successfully colonized the intestinal LP, PC survival is facilitated by factors such as BAFF (secreted by gut DCs) or APRIL (produced by both DCs and the intestinal epithelium) that engage with B cell maturation antigen (BCMA) expressed by PCs22,284. The intestinal LP primarily contains IgA- secreting PCs that characteristically express the cell surface markers CD138, CD38, CD27 and the receptors BCMA and TACI but are intermediate or low for CD45 and low/negative for MHCII, CD19 and B220 (in mice). Notably, the genetic reprogramming into antibody producing PCs coincides with a drastic expansion of the endoplasmic reticulum (ER) and increased metabolic requirements to support the production of large amounts of antibodies285,286. In order to accommodate these high biosynthetic capacities, PCs have adopted intrinsic coping mechanisms. One such mechanism to regulate cellular stress responses is autophagy, which is a universal catabolic process enabling homeostatic maintenance that in addition to other gut-environmental factors ensure the survival of antibody secreting PCs at mucosal effector sites286,287.

2 Aims Our aim was to elucidate the binding capacity and specificity of human intestinal antibodies to members of the gut microbiota. The underlying mechanisms of antibody-microbiota interactions as well as functional consequences of such interactions are still controversial. We further sought to investigate how somatic mutations shape intestinal IgA responses to the microbiota. Moreover, it is not fully understood where in the gut and to what extent IgA- producing plasma cells are generated under homeostatic conditions. Although, Peyer’s patches have been proposed to constitute the main inductive GALT for IgA responses, additional sites such as ILFs or other gut immune-inductive compartments such as the MLNs may also play a role for gut IgA responses. Therefore, we sought to investigate the kinetics of IgA+ PC generation in distinct GALT sites and to examine the contribution of newly generated IgA+ PCs to the prevailing PC pool in the gut under homeostatic conditions.

3 Materials and Methods Part I 3.1 Animals Rag2-deficient mice and C57BL/6J wild type (WT) mice were bred and reared at RWTH Aachen University under specific pathogen-free (SPF) conditions. Germ-free C57BL/6J mice were bred and reared at the Laboratory for Animal Sciences at Hannover Medical School. RAG2-deficient oligoMM12 and WT oligoMM12 mice were bred in reared in gnotobiotic isolators at the Laboratory for Animal Sciences at Hannover Medical School. All performed experiments were approved by the North Rheine-Westphalia State agency for nature, environment and consumer protection (Landesamt für Natur, Umwelt und verbrauerschutz Nordrhein-Westfalen, LANUV) and in accordance with animal welfare guidelines of the German law for animal protection welfare (Tierschutzgestz).

3.2 Human fecal material The human biomaterials were provided by the centralized biomaterial bank of the Medical Faculty of RWTH Aachen University (RWTH cBMB) and were used in accordance with the regulations of the RWTH cBMB and the Ethics Vote 206/09 of the Ethics Committee of the Medical Faculty of RWTH Aachen University.

3.3 Preparation of human tissue samples 3.3.1 Human tissue samples for monoclonal antibody generation Generation of monoclonal antibodies from three healthy donors (HD) undergoing right-sided hemicolectomy has been described before106. None of the donors had a history of intestinal inflammation, and the samples showed no signs of inflammation as determined by macroscopic evaluation and histopathologic examination of the adjacent mucosa106. A second collection of antibodies was generated from surgical samples of three Crohn’s disease (CD) patients undergoing ileocecal resection (CD1: male, 25 years; CD2: male, 47 years; CD3: male, 22 years) and were obtained after signed informed consent in accordance with protocols reviewed and approved by the Charité University Hospital institutional review board (EA1/257/12).

3.3.2 Flow cytometry analysis and single B cell sorting Intestinal plasma blasts/ plasma cells were isolated in the laboratory of Professor Hedda Wardemann affiliated to the German Cancer Research Center, Division of B cell Immunology in Heidelberg and the method previously described106. In brief, lamina mucosa and lamina propria were dissected from the lamina muscularis mucosae by blunt preparation and 2-3 mm tissue pieces were digested using 0.1% DNase and 0.1% collagenase followed by discontinuous Ficoll density gradient centrifugation (GE Healthcare). Purified lamina propria lymphocytes were stained on ice with fluorochrome-coupled anti-human CD38 fluorescein- isothiocyanate (FITC), anti-human CD27 phycoerythrin (PE), anti-human CD19 PECy7, anti- human IgG APC (all from BD Bioscience) or anti-human IgA APC (Jackson Laboratories). Single CD38+CD27+IgA+ or CD38+CD27+IgG+ plasma cells were sorted into 96-well PCR

32 plates using a FACSVantage cell sorter with Diva configuration (BD Bioscience), snap frozen on dry ice and stored at -80°C until further processing.

3.3.3 PCR amplification and expression vector cloning Single cell cDNA synthesis and nested PCR amplification of IGG or IGA and IGK or IGL genes were performed in the laboratory of Professor Hedda Wardemann affiliated to the German Cancer Research Center, Division of B cell Immunology in Heidelberg and described before106. All PCR products were sequenced before and after cloning into previously described eukaryotic expression vectors288. In brief, PCR products of IGH V(D)J and IGK and IGL VJ variable regions, were cloned into separate expression vectors encoding for the constant regions of the human IgG1, IGK and IGL light chain to allow for expression of all antibodies as fully human Fc-IgG1 antibodies. Ig gene sequence analysis including Ig gene usage, clonal relationships, IgG subclass, somatic mutations, IGH CDR3 length and positive charged amino acids, was performed using IgBlast (http://www.ncbi.nlm.nih.gov/igblast/).

3.4 Recombinant antibody production and purification We kindly received the IGH and IGL sequence containing plasmids cloned from human intestinal plasma blasts/ plasma cells from Professor Hedda Wardemann affiliated to the German Cancer Research Center, Division of B cell Immunology in Heidelberg, Germany.

3.4.1 Re-transformation of plasmids Plasmids were re-transformed into competent 5-alpha E. coli cells (New England Biolabs (NEB). 5-alpha E. coli cells were thawed on ice for 5 min and 10 µl of cells mixed with 1 µl of respective plasmid (300 - 500 ng/µl) in 2 ml Eppendorf tubes, gently mixed and incubated on ice for 20 min. Transformation was performed by heat-shock treatment at 42°C for 45 sec. Cells were immediately after the heat-shock treatment transferred onto ice and added with 190 µl Lysogeny Broth (LB) Medium and incubated at 37°C, 150 rpm for 1 h. Finally, 1.8 ml of LB medium was added to the samples and 50 µl of cells plated onto LB-ampicillin (100 µl/ml) containing agar plates and incubated at 37°C, overnight (ON). Successful transformation of plasmids was indicated by the growth of ampicillin resistant plasmid containing cells. Single colonies from the ON cultures were picked and grown in 50 ml LB-ampicillin (100 µg/ml) containing Erlenmeyer flasks and incubated at 37°C, 150 rpm for 16 h.

3.4.2 Plasmid purification and Ig gene sequence analysis After re-transformation of IGH and IGL sequence containing plasmids, plasmids were purified using the QIAGEN Plasmid Midi Kit (QIAGEN®) according to the manufacturer’s instructions and IGH, IGK, and IGL chain genes sequenced using the 5’Ab sense Plasmid Insert check oligo primer 5’GCTTCGTTAGAACGCGGCTAC3’ (SIGMA) and analyzed for product similarity using the GENTle free software.

3.4.3 Expression and Protein A based purification of recombinant antibodies For transfection ~0.7 million human embryonic kidney 239T (HEK) cells with 80% cell confluence seeded in standard medium; DMEM supplemented with 10% heat-inactivated FCS (Invitrogen), 1 mM sodiumpyruvate, 100 μg/ml streptomycin, 100 U/ml penicillin-G, 20 mM Hepes (all from Gibco) were transiently transfected with equal amounts of the corresponding IGH and IGL chain plasmids adjusted to 0.92 µg/plasmid using a calcium phosphate kit and transfection was performed according to the manufacturer’s instructions (Invitrogen transfection Kit: K2780-01). Transfected HEK cells were thereafter incubated for 6 h at 37°C,

5% CO2 and after incubation washed with phosphate buffered saline (PBS) and media changed to Pro 293a (Lonza) supplemented with 1 mM sodium pyruvate, 2 mM L-Glutamin and 100 U/ml Penicillin-G and 100 µg/ml streptomycin (all from Gibco). Supernatants were carefully collected after 60 h and 72 h after transfection and filtered through a sterile filter (0.22 µm) to remove cell debris and stored at 4°C or Igs were purified from supernatants using Protein A based affinity chromatography (ÄktaTM Start, GE Healthcare Life Sciences). Protein A HiTrapTM columns (GE Healthcare, Life Sciences) were loaded with Ig containing supernatant, washed with 20 mM sodium-phosphate buffer and antibodies eluted using 0.1 M citric acid (pH 3.0) and pH neutralized with 1 M Tris-HCL (pH 9.0). Purified Igs were concentrated to ~ 1 - 2 mg/ml in roughly 150 – 200 µl using a 30K MWCO Microsep filter (Microsep centrifugal devices) according to the manufacturer’s instructions. Samples were washed and antibodies collected in PBS. Recombinant antibody concentrations in supernatant or after Protein A based purification were determined by enzyme linked immunosorbent assay (ELISA).

3.5 Enzyme linked Immunosorbent Assay (ELISA) 3.5.1 Anti-human IgG1 ELISA Concentrations of recombinant antibodies were determined by ELISA. ELISA 96 plates (NUNC Maxi Sorp) were coated with 1.25 ng/ml of anti-human IgG Fc (AffiniPure goat anti-human IgG, Fcγ - specific; Jackson Immuno Research, 109-005-098) in PBS. Plates were washed with PBS/0.05% Tween-20 (Sigma) and blocked with PBS/2% BSA for 1 h at 37°C. Human serum

IgG1 standard (Sigma) and recombinant antibodies diluted in PBS/0.05% Tween-20 at two- fold dilutions were incubated for 1.5 h at 37°C and detected with horseradish peroxidase (HRP) conjugated anti-IgG1 antibody (Peroxidase AffiniPure goat anti-human IgG, Fcy – specific, Jackson Immuno Research, 109-035-098) diluted in PBS/0.05% Tween-20 at a concentration of 0.8 µg/ml (Jackson Immuno Research Laboratories) according to the manufacturer’s instructions. Optical densities (OD) were measured at 450 nm and antibody concentrations were determined using the SpectraMax microplate reader software (Molecular Devices) and Excel (Microsoft Excel 2010).

3.5.2 Polyreactivity ELISA of recombinant antibodies Polyreactivity assays were performed as described in106. In brief, ELISA plates were coated with 50 µl of the respective antigens (10 µg/ml calf thymus DNA, 5 µg/ml human recombinant insulin, 10 µg/ml lipopolysaccharide (LPS), 5 µg/ml flagellin, 10 µg/ml cardiolipin, 10 µg/ml

34 human albumin, 10 µg/ml keyhole limpet hemocyanin (KLH) in PBS, overnight at 4°C. ELISA plates were washed 3 times with PBS and incubated with 50 µl of recombinant antibody at 1µg/ml, 0.25 µg/ml, 0.06 µg/ml and 0.015 µg/ml in PBS for 1 h. Recombinant human monoclonal antibodies (mAb) mGO53233 (non-polyreactive) and ED38289 (high-polyreactive) were used for comparison and controls on each plate. ELISA plates were developed with HRP- labeled goat anti-human IgG Fcy specific antibody (Jackson Immuno Research, 109-035-098) at a concentration of 0.8 µg/ml with 2 mM EDTA and 0.05% Tween-20 followed by the HRP chromogenic substrate (ABTS, Pierce). OD450 was measured using a microplate reader (Molecular Devices) and Soft-Max software (Molecular Devices). The cut-off value for each plate was determined based on the highest OD450 value at a concentration of 1 µg/ml minus background of the non-polyreactive control monoclonal antibody (mGO53). Antibodies were considered polyreactive if they showed binding to two or more unrelated antigens tested.

3.6 Deglycosylation of immunoglobulins 3.6.1 Deglycosylation of recombinant Fc-IgG1 antibodies To remove human Ig-N-glycan moieties by deglycosylation, GLycINATOR (EndoS2) an endoglycosidase for hydroxylation of β-1,4 linkage of core GlcNAc residues was used according to the manufacturer’s instructions (GlycINATOR R, Genovis). Deglycosylation was performed with 1 unit of GlycINATOR/1 μg IgG and incubated for 30 min at 37°C.

3.7 Generation of germ-line variants 3.7.1 Ig sequence analysis of recombinant antibodies Antibody sequences were analyzed by IgBLAST comparison with GenBank (http://www.ncbi.nlm.nih.gov/igblast/) and germ-line V(D)J DNA segments with the highest probability determined. Exchange of initial mutated sequences with the germ-line sequence was performed using the GENTle free software. Additional specific restriction sites were introduced for subsequent vector cloning (Table 1). The reverted sequences were generated as gBlock® double stranded DNA fragments from IDT (integrated technologies) with included flanking primer and restriction sites (Table 9).

3.7.2 Cloning of germ-line variants Germ-line (gBlock®) fragments (100 ng/µl) and original vector plasmids containing IGH and IGL sequences were digested with respective restriction enzymes (all from New England Biolabs (NEB)) (Table 1). Vector backbone was purified by gel electrophoresis on a 1.8% agarose gel for 45 min at 120 V. gBlocks were purified by a PCR purification Kit (QIAGEN®) according to the manufacturer’s description. Ligation was performed with T4 DNA ligase (NEB) for 2 h at RT or at 4°C, ON and products transformed into competent 5-alpha E. Coli cells (NEB). After plasmid purification IGH, IGK, and IGL chain genes were re-sequenced using 5’Ab sense Plasmid Insert check oligo primer 5’GCTTCGTTAGAACGCGGCTAC3’ (SIGMA) and analyzed for product similarity. For each antibody germ-line variant, several clones were analyzed and the germ-line and original insert sequences confirmed with IgBLAST comparison

35 with GenBank and Gentle free software (GENtle). Production of germ-line (GL) antibodies was performed as described for the production of recombinant mAbs.

Table 1: List of selected HD and CD derived recombinant antibodies with annotated isotypes and cloning restriction sites. mAb Isotype Restriction sites IGH Restriction sites IGL HD2

HD2a7 IgA1 AgeI/SalI EcoRI/(BsiWI)SpII

HD2a88 IgA2 AgeI/SalI AgeI/PaeR7I HD3

HD3a14 IgA1 AgeI/SalI AgeI/ (BsiWI)SpII

HD3a75 IgA1 AgeI/SalI AgeI/ (BsiWI)SpII

HD3a103 IgA2 AgeI/SalI AgeI/ (BsiWI)SpII

HD3a147 IgA1 AgeI/SalI AgeI/PaeR7I CD1

CD1a293 IgA2 AgeI/SalI AgeI/ (BsiWI)SpII CD2

CD2a61 IgA1 AgeI/SalI AgeI/PaeR7I

CD2a70 IgA1 AgeI/SalI AgeI/PaeR7I

CD2a127 IgA2 AgeI/SalI AgeI/ (BsiWI)SpII

CD2a146 IgA2 AgeI/SalI AgeI/ (BsiWI)SpII

CD2a148 IgA1 AgeI/SalI AgeI/PaeR7I CD3

CD3a32 IgA2 AgeI/SalI AgeI/ (BsiWI)SpII

CD3a549 IgAnd AgeI/SalI AgeI/PaeR7I

CD3a565 IgAnd AgeI/SalI AgeI/PaeR7I

3.8 Western blot 3.8.1 Western blot analysis of germ-line variants To verify the accurate generation of monoclonal antibodies both for germ-line (GL) variants and original mAbs, the molecular weight of GL variants and mAbs was validated by western blot of human heavy and κ or λ light chains. In brief, heat denatured samples (95°C) were separated on a 12% polyacrylamide/SDS gel (custom made) and transferred onto a nitrocellulose membrane (0.45 µm, Bio Rad, 1620115). Membranes were blocked with 5% milk powder in PBS/0.1% Tween ON at 4°C and incubated with HRP – labeled detection antibodies for 1 h. IGH and IGL proteins were detected using goat anti-human IgG, Fcy – HRP conjugated antibody (Jackson Immuno Research, 109-035-098), anti-human kappa light chain-HRP conjugated antibody (1:5000) (ThermoFischer, A18853) or anti-human lambda light chain-HRP conjugated antibody (1:10.000) (Abcam, ab99811) and Western ECL used as substrate (Bio- Rad).

3.8.2 Western blot analysis of deglycosylated recombinant antibodies Deglycosylation efficiency of human Fc-IgG1 antibodies was analyzed by western blot. Heat denatured samples were separated on a 12% polyacrylamide gel (Biorad) and transferred onto 36 a nitrocellulose membrane (0.45 µm, Bio Rad, 1620115). For detection of IgG1 heavy chains, membranes were blocked with 5% milk powder in PBS/0.1% Tween-20 ON at 4°C and incubated with anti-IgG1-HRP (Jackson Immuno Research, 109-035-098) for 1 h. To detect Ig-glycan moieties, membranes were blocked with Hepes 2% Tween-20 buffer, washed twice with Hepes and incubated with Concanavalin A conjugated with horseradish peroxidase (Con-

A-HRP) for 16 h at RT (0.2 µg/ml in Hepes 0.05% Tween-20 and 1 mM CaCl2, MgCl2, MnCl2, Concanavalin A Sigma-Aldrich, L6397). HRP activity was detected with ECL substrate (Bio- Rad).

3.9 Mass spectrometry (MS) 3.9.1 Mass spectrometry of deglycosylated recombinant antibodies Deglycosylation was additionally confirmed by MS. Samples were reduced with TCEP prior to ESI-MS analyses and - desalted using a C4 ZipTip (Millipore, USA) and analyzed in MeOH:2- PrOH:0.2% FA (30:20:50). The solutions were infused through a fused silica capillary (ID75um) at a flow rate of 1 µL/min and sprayed through a PicoTips (ID30um, New Objective Woburn, MA). Nano ESI-MS analyses of the samples were performed on a Synapt G2_Si mass spectrometer and the data were recorded with the MassLynx 4.2 Software (both Waters, UK). Mass spectra were acquired in the positive-ion mode by scanning an m/z range from 100 to 5000 Da with a scan duration of 1 s and an interscan delay of 0.1s. The spray voltage was set to 3 kV, source temperature 80°C and the cone voltage to 50 V. The recorded m/z data were then deconvoluted into mass spectra by applying the maximum entropy algorithm MaxEnt1 (MaxLynx) with a resolution of the output mass 0.5 Da/channel and Uniform Gaussian Damage Model at the half height of 0.7 Da. Mass spectrometry analysis was performed at the Protein Analysis Group (Dr. Chia-wei Lin) coordinated by Prof. Emma Slack, Functional Genomics Center, Zürich.

3.10 Fecal material collection and preparation 3.10.1 Preparation of fecal material Fecal pellets from RAG2-deficient or WT-SPF mice were collected as fresh pellets directly from the mice. For experiments performed on aliquots of pooled fecal material, fecal pellets were collected as descried above and completely homogenized on dry ice using a scalpel. Aliquots of ~ 5 mg fecal material were prepared and fecal pellets or aliquots stored at - 80°C until further processing. Fecal material from RAG2-deficient oligoMM12 and WT oligoMM12 mice were kindly provided by Anna Smoczek from the Laboratory of Animal Sciences of Hannover Medical University. Fecal material from germ-free mice was kindly collected and provided by Susan Jennings from the functional Microbiome Research Group at the Institute of Medical Microbiology, RWTH Aachen University. Obtained human fecal material (5 healthy donors (HD) and 5 ulcerative colitis (UC) patients) was stored at - 80°C unitl further processing.

3.11 Flow cytometry of antibody-stained bacteria 3.11.1 Bacterial flow cytometry and bacterial sort purification (FACS) Single cell suspensions of murine fecal pellets, from aliquots of pooled murine fecal material or from aliquots of human fecal material (5 healthy donors and 5 ulcerative colitis (UC) patients) were prepared by homogenization in 1.5 ml tubes in 1 ml sterile-filtered (0.2 µm) Hank’s Balanced Salt Solution (HBSS, Gibco). Homogenized samples were centrifuged at 500 rpm for 1 min and this process repeated until the supernatant was clear of debris. Clear supernatant was collected and centrifuged at 8000 rpm (VWR Microstar 17R) for 8 min and the supernatant discarded. The remaining bacterial pellet was resuspended in 100 µl of sterile-filtered HBSS/2% normal goat serum (Sigma Aldrich) and passed through a sterile cell strainer (40

µm). The OD600 was determined in a 1:100 dilution and suspensions adjusted to OD600 of 0.13 ± 0.01 (5x107/ml bacteria). Antibody concentrations were adjusted to 1 µg/ml in HBSS/2% normal goat serum. 2 µl of bacterial suspension were stained with 100 µl of diluted antibody, incubated on ice for 20 min and washed in 1 ml HBSS/2% goat serum at 13,000 rpm for 8 min at 4°C. The supernatant was discarded and pellets resuspended in 100 µl anti-human IgG1- AF647 (1:200) (0.5 mg/ml, Alexa Fluor 647-conjugated AffiniPure Goat, anti-human IgG (H + L), Jackson Immuno Research) and incubated for 20 min on ice. 50 µl of Syto9 dye (SytoTM 9 green fluorescent nucleic acid stain, 5 mM in DMSO, Thermo Fischer Scientific) in HBSS was added and incubated for 10 min on ice. Samples were washed in 1 ml HBSS and centrifuged at 13,000 rpm for 8 min at 4°C. The supernatant was discarded and pellets resuspended in 100 µl HBSS and analysed by flow cytometry on an LSR Fortessa (BD). For bacterial suspensions isolated from human fecal material, 2 µl of bacterial suspensions were stained with 5 µg/ml AF647-directly conjugated mAbs for 20 min on ice followed by Syto9 staining and washing as described for murine samples. A minimum of 100,000 Syto9-positive events were acquired for each sample. For FACS purification, samples were prepared as described above in replicates, pooled and sorted on a FacsAria II (BD). Sorted samples were centrifuged at 13,000 rpm for 8 min at 4°C and pellets stored at - 80°C. Staining antibodies and dyes are listed in Table 2. Sort purification of mAb-stained bacteria isolated from oligoMM12 fecal material was performed by Lena Küsgens, a former MSc student at the Institute of Molecular Medicine, AG Pabst at the RWTH Aachen University Clinic.

Table 2: List of fluorescently labeled antibodies and dyes used for flow cytometry and cell sorting, listed with specificity, concentration and company. Antibody target & Clone Concentration Company Fluorochrome conjugate

α-human-IgG1 (H+L)-AF647® 109-605-088 1:200 Jackson α-IgA-PE mA-6E1 1:100 eBioscience α-IgA-Isotype CTRL-PE eB149/10H5 1:100 eBioscience Dyes Cat. No . Concentration Company Syto9 DNA dye S-34854 1:1000 Thermo Fischer

3.12 16S rRNA gene amplicon sequencing and analysis 3.12.1 Metagenomic bacterial DNA isolation Bacterial DNA from human fecal samples was extracted via glass bead preparation and further isolated using the QIAamp DNA Stool Kit (QIAGEN®) according to the manufacturer’s protocol. Metagenomic bacterial DNA of sort purified bacterial samples of mAb positive, mAb negative fractions and input material was extracted using the Qiagen bacterial DNA isolation kit (QIAamp Fast DNA Stool MiniKit, QIAGEN®) according to the manufacturer’s instructions.

3.12.2 PCR amplification of bacterial DNA The V3/V4 regions of the 16S rRNA gene were amplified in one PCR reaction (35 cycles) using V3/V4 16S rRNA bacterial/archeal specific primers - 515F and 816R that was tagged with a 12-nt-long barcode sequence290,291 (https://earthmicrobiome.org/protocols-and- standards/16s/). PCR reactions were carried out using the AccuPrime Kit (Thermo Fisher) in duplicates for each sample and pooled after the PCR reaction (Table 3). 515F (forward primer) PCR primer sequence: 5’AATGATACGGCGACCACCGAGATCTACACTATGGTAATTGTGTGCCAGCMGCCGCGG TAA-3’ 806R (reverse primer) PCR primer sequence: 5’-CAAGCAGAAGACGGCATACGAGAT - [Golay Barcode] - AGTCAGTCAGCCGGACTACHVGGGTWTCTAAT-3’ The [12bp Golay Barcode] assigns a multiplex identifier sequence that is unique for each sample290,291 (https://earthmicrobiome.org/protocols-and-standards/16s/).

Table 3: Reagents and PCR program for V3/V4 16S rRNA gene amplification. Reagent Volume PCR program

ddH2O 19.4 µl 1. 94°C for 3 minutes Buffer II (Accu Prime Kit) 2.5 µl 2. 94°C for 45 seconds 515 Forward primer, 10 µM 0.5 µl 3. 50°C for 60 seconds

816R Reverse primer, 10 µM 0.5 µl 4. 72°C for 90 seconds Template DNA 2 µl 5. Steps 2 – 4 repeated 35 x Taq Polymerase (Accu Prime Kit) 0.1 µl 6. 72°C for 10 minutes Volume per sample 25 µl 7. 4°C hold

PCR amplicons were purified by agarose gel (1.5%) electrophoresis followed by a QIAquick Gel Extraction Kit (QIAGEN®). Empty samples were included in the overall sample processing

(FACS buffer collected during cell sorting) and analyzed along with H2O and buffer controls used during library preparation to identify potential artifact sequences and spurious operational taxonomic units (OTU). DNA amplicon concentrations were determined using the Quant-iTTM PicoGreen® dsDNA Kit according to the manufacturer’s instruction (Invitrogen) and PCR products were stored at - 20°C until final Illumina MiSeq sequencing.

3.12.3 16S rRNA gene amplicon Illumina sequencing Samples were prepared according to the Illumina amplicon library preparation protocol (manufactures instructions for Illumina Miseq). Individual samples for the DNA library pool preparation were normalized to a final concentration of 4 nM and 5 µl of each sample dilution used for the 16S DNA library pool. The 16S rRNA gene amplicon libraries were sequenced with single barcodes in paired end mode (2 x 300 nt) using a MiSeq sequencer (Illumina Miseq). Christina Petrick (RWTH Aachen University Clinic) contributed substantially to the sample preparations for 16S rRNA sequencing.

3.12.4 16S rRNA gene amplicon analysis Initial raw reads after 16S rRNA sequencing were processed using a custom-made trimming Perl-Script and a demultiplexing Perl-Script available at the IMNGS292 platform (www.imngs.org). Raw reads were further processed using the IMNGS platform for custom study analysis based on UPARSE approach293. After demultiplexing, all sequences were trimmed to the first base with a quality score < 3. Operational taxonomic units (OTUs) were clustered at a sequence similarity of 97% and only OTUs occurring at a relative abundance ≥ 0.25% in at least one sample were used for further analyses. Taxonomic assignment was based on the RDP (Ribosomal Database Project) classifier version 2.11 with a confidence cut off of 80% and dendrogram clustering of OTUs (phylum to family) was based on genomic sequence similarity obtained by the RDP taxonomic classification294. The lineage and taxonomic identity (closest species with a valid name and corresponding 16S rRNA gene sequence identity) of relevant OTUs were obtained using EZbiocloud295. Final downstream analysis was performed using the Rhea pipeline, consisting of a set of R scrips (Ri386 3.3.2). Measures for α and β-diversity and bacterial composition analyses were performed using Rhea version 2.0 in R296. α-diversity measures were used to assess microbial diversity within samples and were assessed as species richness and Shannon effective count. β-diversity measurements that compare microbial profiles between samples were analyzed by PCoA, based on computed generalized UniFrac distances297.

3.13 Microbiology 3.13.1 In vitro cultivation of oligoMM12 bacteria The cultivation of cryo-preserved MM12 bacteria was kindly aided by Theresa Streidl and David Wylensek, both from the microbiology department at RWTH Aachen University Clinic. In brief: cryostocks of individual bacteria of the oligoMM12 bacterial consortium were initially inoculated in sterile and anaerobic Hungate-tubes supplemented with the appropriate cultivation media best suited for growth of each bacterial species and incubated at 37°C. Bacteria were re-cultivated 3 to 4 times until stable growth and morphology was acquired. If required, bacteria were cultivated on specific blood agar plates at 37°C under strict anaerobic conditions. All cultivation steps on agar plates were performed in an anaerobic box. After cultivation, bacteria were washed, centrifuged and prepared for flow cytometry analysis.

3.14 Data analysis 3.14.1 Formulas Bacterial specificity of individual monoclonal antibodies (mAbs) was determined as enrichment index. The enrichment index for each mAb was calculated as the ratio of relative operational taxonomic unit (OTU) abundance in the mAb positive fraction and negative fraction, taking into account the overall relative abundance of a given OTU in the positive fraction. mAb+ denotes the relative abundance value of a given OTU that is equivalent to a bacterial species in the mAb targeted fraction. mAb- denotes the value of the relative abundance of a given OTU that was not targeted by single mAbs.

The enrichment index was calculated according to the following formula: (mAb+ - mAb-) / (mAb+ + mAb-) * IgA+

3.14.2 Statistical analysis Statistical analysis for FACS generated data was performed as indicated in the figure legends using the GraphPad Prism software package (GraphPad Prism 7). p-values ≤ 0.05 were considered significant. Statistics for the 16S rRNA gene amplicon analysis were obtained by the Rhea296 implemented R scripts (https://lagkouvardos.github.io/Rhea/). In brief, α-diversity was calculated as effective diversity (Shannon diversity) within a given sample. β-diversity was determined by calculation of a phylogenetic distance matrix using the generalized UniFrac approach297. Generated distances are displayed using principal coordinate analysis (PCoA). Significant separation of samples was tested using a PERMANOVA test between pairs and as a whole296.

4 Materials and Methods Part II 4.1 Animals AID-Cre-ERT2 x Rosa26-loxP-eYFP and AID-Cre-ERT2 x Villin-Cre x Rosa26-loxP-eYFP mice all on C57BL/6J background were bred and reared at the animal facility of RWTH Aachen University under specific pathogen-free (SPF) conditions. AID-Cre-ERT2 x Villin-CRE x Rosa26-loxP-eYFP mice were generated at the Institute of molecular medicine group at RWTH Aachen University by crossing the parental lines AID-Cre-ERT2 x Rosa26-loxP-eYFP mice and Villin-Cre-ERT2 x Hif1a x bcat until the desired genotype AID-Cre-ERT2 x Villin-Cre x Rosa26-loxP-eYFP was established. All performed experiments were approved by the North Rhine-Westphalia State agency for nature, environment and consumer protection (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, LANUV) and in accordance with animal welfare guidelines of the German law for animal protection welfare (Tierschutzgestz).

4.2 Intraperitoneal application of Tamoxifen Conditional loxP transgenic AID-Cre-ERT2 x Rosa26-loxP-eYFP or AID-Cre-ERT2 x Villin-Cre x Rosa26-loxP-eYFP mice were intraperitoneally (ip) injected with tamoxifen (3 mg/mouse) diluted in laboratory corn oil (both from Sigma Aldrich) using a 25G x 1 needle (Smiths medical) at day 0 of respective experiments. Mice were analyzed by flow cytometry or tissues kept for histology.

4.3 FTY720 administration Assessment of eYFP+ B cell generation in inductive GALT Fingolimod (FTY720) is a functional antagonist of S1PR1, a molecule expressed on lymphocyte enabling the egress of secondary lymphoid organs (SLO)298. The administration of FTY720 allows lymphocytes to enter SLOs from blood but effectively prevents the egress via efferent lymphatic vessels, leading to an accumulation of lymphocytes in SLOs. FTY720 was administered in drinking water (2.5 µg/ml) at day - 1 of respective experiments and maintained throughout the experimental time period. At day 0 mice were intraperitoneal (ip) injected with tamoxifen. FTY720 administration together with ip tamoxifen injection will allow for assessing potential inductive sites of eYFP+ B cell generation and may demonstrate the iterative increase of eYFP+ B cells in respective inductive compartments. At respective experimental time points, mice were analyzed by flow cytometry and/or tissues kept for histological analyzes.

Assessment of 4-hydroxy tamoxifen diffusion after Peyer’s patch injections In this experimental set-up, administration of FTY720 allows to analyze whether injections with 4-hydroxy tamoxifen (4-OHT) are locally confined to the injected Peyer’s patch (PP) or if 4- OHT diffusion would lead to the generation of eYFP+ B cells outside injected PPs, indicated by eYFP+ B cells within other compartments but injected PPs.

4.4 Surgical procedures 4.4.1 4-OHT micro-injections of single Peyer’s patches Before surgical procedures were performed mice were anesthetized by intraperitoneal injection with Ketamine (100 mg/kg) and Xylazine (10 mg/kg) in PBS. When no response to reflex testing was detected, mice were stably mounted on their back exposing the ventral plane and kept on a heating mat throughout surgery. The lower abdomen of the mouse was moisturized with 70% ethanol followed by a small midline incision of ~ 1.5 cm into the abdominal skin (Figure 8 A). For laparotomy, the skin was gently separated from the peritoneum by blunt preparation and the peritoneum was cut open along the linea alba by a small incision (~ 1 cm). After opening the skin and peritoneal cavity, the incision sites were covered with phosphate buffer saline (PBS) soaked cotton pads and the small intestine gently removed with PBS soaked cotton swabs from the peritoneal cavity and placed on the soaked pads. PPs were localized using a binocular microscope. PPs were fixed in place with sterile plastic tweezers and thereafter micro-injected with 5 µg/patch (~ 1µl) 4-OHT in DMSO (Cayman Chemicals) using a fine glass capillary (Figure 8 B).

A B

Figure 8: Surgical procedure of Peyer’s patch 4-OHT micro-injections. (A) Anesthetized mice mounted on their back exposing the ventral plane. The abdominal skin was opened by a small incision and the peritoneum opened along the linea alba. (B) The small intestine was gently removed from the peritoneal cavity and placed on PBS soaked cotton pads. PPs were localized using a binocular microscope, fixed in place and microinjected for 0.1 - 0.3 s with 10 - 28 psi. with 5 µg/patch (~ 1 µl) 4-OHT using beveled glass capillaries.

Custom made glass capillaries for micro-injections were pulled using glass capillaries (1.5 OD x 1.17 x 100 L mm, GC150TF-10, HARVARD Apparatus) and a micropipette puller (Sutter Instrument, heat 513°C, pull 70, velocity 70, time 155 ms, pressure 500). Capillaries were filled with ethanol and capillary tips with a diameter of ~ 30 µm subsequently beveled to 60° using a grind stone (EG-44, NARISHIGE). Tips were checked for accurate beveling using a microscope (Leica, DM750). Single Peyer’s patches were injected multiple times to not disrupt the tissue structure. During the entire procedure the intestine and surrounding skin was kept moist with PBS. After microinjections, the small intestine was gently placed back into the peritoneum and the peritoneal incision surgically sutured with absorbable surgical thread (Marlin violet, HR17, catgut). The abdominal skin was closed using surgical metal clips (9 mm, fine science tools). Mice were thereafter transferred to heat regulated cages to prevent hypothermia until they were waking up. Afterwards, mice received the analgesic Novaminsulfon (Ratiopharm®) ad libitum (0.8 mg/500 ml) in drinking water for the next three days. Mice were checked and scored daily for their well-being. In longer experiments, clips were removed after 14 days.

4.5 Lymphocyte isolation 4.5.1 Tissue preparation and cell isolation

Mice were sacrificed as indicated for respective experiments by CO2 inhalation followed by cervical dislocation. Required organs for analysis were surgically removed and kept on ice in PBS/3% fetal calf serum (FCS). Depending on the organ, different protocols were applied to prepare single cell suspensions for subsequent analysis.

4.5.2 Cell isolation from blood

To withdraw blood, mice were sacrificed by CO2 inhalation and immediately after death mounted on a Styrofoam tray and kept tapered at a 30° angle. The abdominal skin and peritoneum were opened and organs gently moved aside to expose the aorta abdominals. Using an insulin needle and syringe (29G x 12.7 mm, BD Micro-Fine™) filled with 100 µl 1X citrate buffer (10X citrate buffer, 11.2 g NaH2PO4, 0.65 g tri-sodium citrate dehydrate in 100 ml ddH2O, pH 4.5) ~ 500 µl blood was withdrawn from the aorta. After blood withdrawal, blood was transferred into 1.5 ml Eppendorf tubes containing 400 µl of citrate buffer and tubes inverted to prevent blood from clotting. Samples were thereafter lysed in 10 ml erythrocyte lysis buffer (0.17 M NH4CL, 10 mM KHCO3, 0.1 mM EDTA) for 10 minutes at RT and afterwards washed with PBS/3% FCS at 400 g, for 8 minutes at 4°C and supernatants discarded. The erythrocyte lysis step was repeated twice. After final washing, supernatants were discarded and cell pellets resuspended in 200 µl PBS/3% FCS and transferred into a 96 well V-bottom plate.

4.5.3 Cell isolation from lymphoid organs Single PPs were carefully dissected from the small intestine using an illuminated magnifier. Mesenteric lymph nodes (MLNs) were dissected from the mesentery and the fat tissue removed. Spleen was carefully removed from the abdominal cavity. PPs, MLNs and spleen were kept on ice in PBS/3% FCS. Single PPs and MLNs were thoroughly mashed and the homogenized organs filtered through a 48 µm nitex gauze. The gauze was rinsed with PBS/3% FCS to obtain single cells in suspension in ~ 200 µl PBS/3% FCS. Cells isolated from PPs and MLNs were then centrifuged at 400 g for 8 minutes at 4°C. Remaining erythrocytes from mashed and filtered spleens were lysed in 10 ml erythrocyte lysis buffer (0.17 M NH4CL, 10 mM KHCO3, 0.1 mM EDTA) in 15 ml falcon tubes and incubated for 10 minutes at RT. After incubation, falcon tubes were filled up to 15 ml with PBS/3% FCS and centrifuged at 400 g, for 8 minutes at 4°C. Supernatants were discarded and cell pellets resuspended in 500 µl BPS/3% FCS and kept on ice.

4.5.4 Lymphocyte isolation from small intestinal lamina propria To isolate lymphocytes from the small intestinal lamina propria (SI LP), mice were sacrificed as described above, the entire small intestine (SI) removed and placed in a petri dish containing 10 ml PBS/3% FCS on ice. All PPs were carefully removed and further processed as described above. Adjacent fat was carefully removed and the SI cut open longitudinally to remove SI content by washing the tissue in PBS/3% FCS. The SI was cut into 2 to 3 pieces and again washed in 10 ml ice cold PBS/3% FCS in 50 ml Falcon tubes by vigorous shaking. Thereafter, SI pieces were transferred into a 50 ml Falcon tube containing 10 ml of HBSS/10% FCS/2mM EDTA and incubated for 15 minutes at 37°C at 150 rpm on a platform shaker. After incubation, supernatants were discarded and SI pieces briefly washed in PBS/3% FCS (step repeated 3x). After the third EDTA wash and a final PBS/3% FCS washing step, the SI pieces were incubated in 10 ml of pre-warmed RPMI/10% FCS supplemented with collagenase A (0.24 mg/ml) (Sigma Aldrich) and 50µg/ml DNase I (Roche) for 30 to 40 minutes at 37°C at 150 rpm on a platform shaker. Every 5 to 10 minutes the tissue was examined and vigorously shaken by hand, making sure to not fully digest the tissue. The tissue was subsequently filtered through a 40 µm cell strainer (Corning™, Fisher Scientific) into a 50 ml Falcon tube. The tube was topped up with PBS/3% FCS and centrifuged at 400 g, for 10 minutes at 4°C. Supernatants were discarded and the cell pellets resuspended in PBS/3% FCS and cell suspensions kept on ice until further cell staining for flow cytometry.

4.6 Epithelial cell isolation from the small intestine For epithelial cell isolation from the small intestine (SI) either the entire small intestine was prepared or single gut pieces (~ 1 – 1.5 cm) were prepared. As described above, the entire SI was removed from the mouse abdominal cavity, the fat tissue and PPs carefully removed and placed in a petri dish containing 10 ml of ice cold PBS/3% FCS. Single gut pieces were placed in 200 µl ice cold PBS/3% FCS. For cell isolation from the entire SI, the SI was cut into 2 to 3 pieces and pieces cut open longitudinally to remove SI content by washing in ice cold PBS/3% FCS. Subsequently, the SI pieces were placed in a 50 ml Falcon tube containing ice cold PBS/10% FCS and were vigorously shaken by hand. The supernatant was removed and gut 45 pieces transferred in 10 ml of pre-warmed RPMI/5% FCS/ 2 mM EDTA and incubated at 37°C, 15 minutes at 150 rpm. After incubation the tissue and supernatant were filtered through a 40 µm sterile cell strainer (Corning™, Fisher Scientific) into a fresh 50 ml Falcon tube and the filtered supernatant containing the epithelial cells kept on ice. The procedure was repeated three times and the supernatants containing gut SI epithelial cells pooled for subsequent analysis. Cells were washed in 50 ml RPMI/5% FCS at 450 g, 10 min at 4°C. The supernatant was discarded and the cell pellets resuspended in PBS/3% FCS and kept on ice until further staining for flow cytometry. For isolation of epithelial from single gut pieces (1 – 1.5 cm) the tissue pieces were incubated twice in 5 ml pre-warmed RPMI/5% FCS/2 mM EDTA at 37°C, 15 minutes at 150 rpm as described above. Filtered supernatants were collected in 15 ml Falcon tubes and cell pellets were resuspended in 150 µl PBS/3% FCS and kept on ice until further processing for flow cytometry staining.

4.7 Cell stainings for flow cytometry analysis 4.7.1 Staining of B cells and plasma cells For surface and live/dead cell staining, 200 µl of cell suspensions from PPs, MLNs, blood, spleen or SI LP cells were transferred in a 96 well V-bottom plate (Sarstedt) and supplemented with 10% rat serum and incubated for 5 – 10 minutes on ice (Sigma Aldrich). Cells were centrifuged (400 g, 8 min, 4°C) and surface stained in 100 µl PBS/3% FCS containing the respective antibody master mix for 30 – 60 minutes on ice, light protected. Subsequently, 50 µl of live/dead dye diluted in PBS/3% FCS was added to the cells and incubated again for 10 – 20 minutes on ice, protected from light. Cells were washed with PBS/3% FCS (centrifuged at 400 g, 8 min, at 4°C). The pellets were resuspended in 100 µl PBS/3% FCS and cells filtered through 48 µm nitex gauze into flow cytometry tubes. Antibodies and dyes used for flow cytometry are listed in Table 4.

4.7.2 Staining of small intestinal epithelial cells Single cell suspensions of gut epithelial cells were supplemented with 10% rat serum and incubated on ice 5 – 10 minutes. Cells were transferred into a 96 V-bottom well plate (Sarstedt) and centrifuged at 450 g, for 10 minutes at 4°C. Cells were subsequently stained in 100 µl PBS/3% FCS containing the antibody master mix and incubated for 30 – 60 minutes, light protected on ice. 50 µl of live/dead dye dilute in PBS/3% FCS was added to the cells and incubated for another 10 – 20 minutes on ice, light protected. Stained cells were washed (centrifuged at 450 g, 10 min, 4°C) with PBS/3% FCS, supernatants were thereafter discarded and pellets resuspended in ~ 100 µl PBS/3% FCS and filtered through 48 µm nitex gauze into flow cytometry tubes. Samples were analyzed using the LSR FortessaTM flow cytometer (BD Bioscience) and further analyzed using the FlowJo software (FLowJo LLC). Antibodies and live/dead dyes used for cell stainings are listed in Table 4.

Table 4: Fluorescently labeled antibodies and reagents used for flow cytometry, listed with specificity, concentration and company. Antibody target & Clone Concentration Company Fluorochrome conjugate CD45-BV510 30-F11 1:200 BioLegend CD3-BV650 17A2 1:100 BioLegend CD4-BV650 RM4-5 1:200 BioLegend CD19-BV786 6D5 1:200 BioLegend CD138-APC 281-2 1:100 BioLegend B220-BV711 RA3-6B2 1:200 BioLegend EpCam (CD326)-PE G8.8 1:2000 ebioscience IgA-PE mA-6E1 1:200 ebioscience GL7-Pacific blue GL7 1:200 BioLegend IgM APC-Cy7 RMM-1 1:100 BioLegend eYFP/FITC transgene Live/Dead dye Cat. No . Concentration Company 7AAD 4204404 1:50 BioLegend DAPI D9542 1:10000 Sigma-Aldrich

4.8 Histology 4.8.1 Tissue preparation for histological analysis Small intestinal gut pieces containing PPs were transversally cut into ~ 0.5 cm long tubes, washed in ice cold PBS and subsequently fixed in 4% paraformaldehyde (PFA) for 1 - 2 hours at RT. Samples were thereafter incubated in ddH2O/30% sucrose ON at 4°C. After incubation in ddH2O/30% sucrose, gut pieces were slowly frozen in O.C.T™ (embedding medium for frozen tissue specimens, Tissue-Tek®) on dry ice and kept at - 80°C until further processing (Figure 9). After ddH2O/30% sucrose incubation tissue samples for whole mount preparations were transferred into PBS and kept at 4°C.

4.8.2 Cryotome tissue sections Frozen O.C.T™ tissue blocks were trimmed and afterwards mounted on the cryotome cutting block (Figure 9). Tissues were cut transversally into 8 - 9 µm sections using a LEICA CM3050 S Cryotome. Depending on the size of the tissue block, 5 to 10 sections were mounted on one object slide (adhesive, Klinipath) and set aside to let dry ON at RT, protected from light. Object slides with sections were either stored at – 20°C until further processing or ON dried sections were rehydrated (2 x 5 min) in TBST (1X tris buffered saline/0.1% Tween20) and rinsed in ddH2O. Object slides were embedded with Mowiol® (Roth) and mounted with cover slips and again set aside to let dry at RT, ON, protected from light. When the Mowiol® was thoroughly dried, sections were ready for confocal microscopy analysis.

Figure 9: Preparation of small intestinal tissue samples for Cryotome sectioning. Small intestinal gut pieces containing Peyer’s patches were fixed and mounted for transversal sectioning.

4.9 Confocal microscopy Fixed whole mount Peyer’s patches or tissue sections of fixed Peyer’s patches and adjacent small intestine were histologically analyzed using a confocal microscope (LSM 710, AxioObserver). eYFP signal of whole mount preparations and of tissue sections was acquired using a 519 - 583 nm detection channel (green light) and laser settings of 40% laser power with laser wavelength of 514 nm excitation and emission at 551 nm. To discriminate true eYFP signal from tissue autofluorescence, emission was additionally detected using the 583 - 696 nm channel (red light), with laser settings of 2.6% laser power with laser wavelength of 561 nm excitation and emission at 640 nm. For whole mount tissue preparations, the bright field (BF) channel was also acquired. Sabrina Ernst (RWTH Aachen University Clinic) provided technical support with confocal microscopy.

4.10 Data Analysis 4.10.1 Statistical analysis Statistical analyses of flow cytometry data were performed as indicated in the figure legends using the GraphPad Prism Software package (GraphPad Prism 7). p-values ≤ 0.05 were considered significant. Graphical outputs were generated using GraphPad Prism Software package (GraphPad Prism 7), FlowJo (FlowJo LLC) and finalized in Adobe Illustrator (Adobe).

5 Results Part I 5.1 Microbiota-reactive IgA and IgG plasma cell derived antibodies are prevalent in the adult human small intestine To investigate the binding capacity and specificity of Fab-dependent canonical binding of human intestinal antibodies to members of the microbiota under both homeostatic-and chronic inflammatory conditions, two collections of recombinant monoclonal antibodies originally derived from intestinal plasma blasts/ plasma cells were generated. The first collection of 162 antibodies was derived from the terminal ileum of three healthy donors (HD) and described previously106. The second collection of 118 antibodies was obtained from the terminal ileum of three Crohn’s disease (CD) patients. From both HD and CD collections, IgA – and IgG antibodies were cloned from single small intestinal IgA+CD38+CD27+ or IgG+CD38+CD27+ plasma blasts/ plasma cells. Both IgA – and IgG derived antibodies were cloned and expressed as human Fc-IgG1 hybrid proteins in mammalian HEK-T293 cells106,288 (Figure 10). The isolation of human plasma cells and the original production of recombinant antibodies was performed in the laboratory and group of Professor Hedda Wardemann, affiliated to the German Cancer Research Center, Division of B cell Immunology in Heidelberg, Germany. In total, 280 recombinant monoclonal antibodies (mAbs) were generated and initially screened for their reactivity to murine microbiota (IgA mAbs: 105 HD derived, 84 CD derived; IgG: 57 HD derived and 34 CD derived).

Figure 10: Workflow depicting the isolation of small intestinal lamina propria plasma cells from three healthy donors (HD) and three Crohn’s disease (CD) patients. IgA and IgG derived antibodies were cloned from single IgA+CD38+CD27+ or IgG+CD38+CD27+ small intestinal lamina propria (SI LP) plasma cells. IgA and IgG monoclonal antibodies were cloned and expressed as Fc-IgG1 hybrid proteins in human HEK 293 cells. Protein concentrations of recombinant antibodies were determined via ELISA. Antibody containing supernatant or Protein A based purified antibodies were used for FACS analysis. The isolation of human plasma cells and the original production of recombinant antibodies was performed in collaboration with the laboratory and group of Professor Hedda Wardemann and the methods previously described106,288.

The common Fc-IgG1 part of recombinant IgA and IgG mAbs allowed for direct comparison and unified detection of microbiota reactivity and specificity (Figure 11 A). Potential differences in binding capacity and specificity of individual mAbs to gut microbial members can therefore be directly related to differences in Fab-dependent canonical antigen binding. To determine the binding capacity of mAbs, isolated fecal bacteria were stained with individual mAbs and the frequency of bound bacteria detected by flow cytometry. Identification of microbial cells by flow cytometry was based on the scatter profiles of isolated bacteria and staining properties with the DNA binding dye Syto-9 (Figure 11 A). In vitro cultivated GFP-expressing E. coli and fecal material isolated from germ-free mice allowed for establishing precise gating of bacteria and was used thereafter for gating of bacteria isolated from SPF or oligoMM12 feces (Figure 11 B). As proof of methodological concept, bacteria isolated from SPF or oligoMM12 feces were stained with a high microbiota-reactive control antibody (ED38) and binding detected with anti-human Fc-IgG1-AF647 fluorescently labeled detection antibody (Figure 11 B).

Figure 11: Identification of fecal bacteria by flow cytometry. (A) Representative workflow of fecal bacteria staining and analysis. Single recombinant monoclonal antibodies (mAbs) were used to stain bacteria isolated from murine fecal material and screened for their microbiota binding capacity using flow cytometry. The DNA binding dye Syto-9 (green) stains live and dead bacteria and mAb-binding to bacteria was detected by AF647-conjugated anti-human Fc-IgG1 antibody (red). (B) Gating of fecal bacteria was set according to in vitro cultured GFP-expressing E. coli, germ-free fecal material and fecal material from oligoMM12 RAG2-deficient mice, SPF RAG2-deficient mice or SPF WT mice (top row). GFP signal or Syto-9 nucleic acid dye were used for proper bacterial identification (middle row). Anti- human Fc-IgG1 detection antibody was used to determine mAb positive bacteria. Representative binding is displayed for the polyreactive high-microbiota reactive control mAb (ED38) on different SPF and oligoMM12 derived fecal material (bottom row). 52

A large proportion of the commensal gut microbiota has been shown to be coated with luminal endogenous IgA in mice and humans11,20,170,176,299. Therefore, to avoid potential competition for epitope binding between endogenous IgA and our recombinant mAbs, bacteria used for mAb staining were isolated from fecal samples of RAG2-deficient mice, which lack T and B cells and therefore soluble Igs. Bacteria stained with the negative control antibody (mGO53)289 and antibody-free cell culture supernatant (MedCTRL) showed low but variable background staining (MedCTRL 0.52 ± 0.31, mean ± SD, n = 36). These low bacteria binding frequencies were comparable to binding frequencies of a representative low microbiota-reactive mAb (HD1a377) (Figure 12 A). The high-microbiota polyreactive control antibody (ED38)233 showed similar high binding frequencies comparable to a representative high microbiota-reactive IgA mAb (HD3a75) (Figure 12 A). To exclude potential false positive binding to bacteria, we considered binding to bacteria ≥ 1% of total bacterial cells as the threshold of mAb microbiota- reactivity. Screening of 280 mAbs for their binding capacity to microbial members revealed that both IgA and IgG mAbs showed a heterogeneous spectrum of microbiota-reactivity. We found IgA and IgG mAbs with no microbiota reactivity but also antibodies derived from both HD and CD donors binding to more than 5% of fecal bacteria (Figure 12 B). To find an objective and qualitative measure for microbiota-reactivity, we categorized antibodies as having either, no/low (0 - 1% binding to bacteria), intermediate (1 - 5% binding to bacteria) or high (> 5% binding to bacteria) binding capacity for gut bacteria (Figure 12 C). Notably, mAbs with high microbiota-binding capacities were found for both original isotypes (IgA and IgG) and were observed in both sets of HD and CD derived antibodies. Importantly, while CD derived mAbs almost exclusively did not contain mAbs with no/low binding capacity, intermediate microbiota- reactivity however was more pronounced in CD derived mAbs from both isotypes as compared to HD derived mAbs (Figure 12 B, C). Our results strongly suggest that a pronounced microbiota-binding capacity is a distinguishing property of human intestinal IgA as well as IgG antibodies under both homeostatic and inflammatory conditions.

Figure 12: Microbiota-reactive IgA and IgG mAbs are present in the adult human gut. (A) IgA-free microbes were isolated from feces of SPF RAG2-deficient mice. Anti-human Fc-IgG1 detection antibody was used to determine mAb positive staining of fecal bacteria. Representative staining of bacteria isolated from feces of RAG2-deficient SPF mice are depicted for supernatant from non-transfected HEK- cells (Med CTRL), low (mGO53) and high (ED38) polyreactive control mAbs and representative high and low microbiota-reactive IgA mAbs. (B, C) mAbs generated from IgA- or IgG expressing plasma cells from healthy donors (HD) or Crohn’s disease (CD) patients were FACS screened for their reactivity to bacteria isolated from feces of RAG2-deficient SPF mice. (B) Symbols represent individual mAbs (open symbols HD derived, filled symbols CD derived). Binding of mAbs to ≤ 1% of bacteria was considered background (dashed line). Statistical significance between groups was determined by Kruskal-Wallis test, followed by Mann-Whitney test (** p ≤ 0.001; *** p ≤ 0.0001). (C) Microbiota reactivity of mAbs was categorized according to their binding capacity as: no/low (0 - 1%), intermediate (1 - 5%), and high (> 5%) bacteria binding.

5.2 Microbiota-reactive IgA mAbs consistently bind a major fraction of the microbiota After screening our collections of mAbs for their microbiota reactivity, we selected 15 high microbiota-reactive IgA mAbs (6 HD derived and 9 CD derived) for further analysis. Representative flow cytometry plots of the selected HD and CD derived IgA mAbs are displayed in Figure 13 A. Notably, two of the IgA mAbs derived from the HD mAb collection are clonally related (HD3a14 and HD3a75), potentially suggesting that both antibodies might display similar binding properties and specificities to gut bacteria. To assess the reproducibility of high microbiota-reactivity, selected mAbs were tested multiple times on murine donor feces obtained from animals housed in different cages and collected at different times. All selected high microbiota-reactive IgA mAbs showed consistent high microbiota-binding capacities (> 5%) to gut bacteria across independent experiments (Figure 13 A, B). However, across independent experiments, tested IgA mAbs showed variability in their binding capacities to gut bacteria (Figure 13 B). These variations in frequencies of mAb-bound bacteria point to differences in microbial composition between donor feces. To directly test for the contribution of individual gut microbiota compositions to the binding capacity of mAbs, we assessed binding to gut bacteria on aliquots of pooled fecal material in direct comparison to binding of gut bacteria isolated from individual fecal material across multiple independent experiments. Performing experiments on bacteria isolated from pooled fecal material reduced the previously observed variability of mAb binding capacities to gut bacteria, suggesting that differences in microbial diversity and composition of individual fecal samples affect the microbiota binding capacity of mAbs (Figure 13 C). The data set obtained on individual and pooled fecal material was performed by Lena Küsgens during her masters in our institute of molecular medicine.

Figure 13: High microbiota-reactive IgA mAbs consistently bind a major fraction of the intestinal microbiota. (A) Representative FACS plots depicting the staining of bacteria (gated on FSC/SSC profile and Syto-9-positive events) with selected high microbiota-reactive HD (left) and CD derived (right) mAbs. Numbers indicate mean percentage of bacteria binding ± SD of at least four independent experiments. (B) Microbiota-reactivity of selected HD (open circles) and CD (black circles) mAbs over a large set of independent experiments performed on unrelated fecal material. (C) Binding capacity of selected HD mAbs is displayed to bacteria isolated from unrelated fecal material (white bars) or to bacteria isolated from aliquots of pooled identical fecal material (grey bars). (B, C) Single dots represent independent experiments; bars represent min to max and mean percent (%) of mAb-microbiota reactivity. Dashed line represents the cut-off of 1% bound bacteria. The data set obtained on individual and pooled fecal material was performed by Lena Küsgens.

5.3 Donor-dependent variability of IgA mAb binding capacities to human gut bacteria Despite the limitations posed by working with human fecal material, such as prominent inter- individual differences in microbial complexity and composition, we yet sought to recapitulate our findings found for mAb binding capacities on RAG-/- feces using bacteria isolated from human fecal material. In order to capture and cover different human microbial configurations, we screened our selected high-microbiota HD and CD derived mAbs on bacteria isolated from 5 healthy (HF) and 5 ulcerative colitis (UC) human fecal samples. Antibody screening on human fecal samples required direct fluorochrome conjugation of selected mAbs, due to the high background binding to human fecal bacteria of the anti-human Fc-IgG1 secondary antibody. In addition, we performed 16S rRNA sequence analysis of the HF and UC derived human fecal samples. Consistent with previous studies, we found markedly different microbial diversity and composition of UC derived microbial members in contrast to healthy donors indicated by lower α-diversity measures (richness and Shannon effective count) of UC samples (Figure 14 A, B). α-diversity determines the microbial diversity within samples. However, also microbial diversity within HF samples showed heterogeneity. In addition, β-diversity analysis demonstrated that bacterial compositions from UC samples were substantially distinct from HF samples (Figure 14 C). In contrast to α-diversity measures, the β-diversity estimates differences between samples based on their microbial complexity and composition. Similar to the observed variability of mAb binding capacities when using unrelated RAG-/- fecal samples, we also found marked variability of mAb binding profiles and capacities to human gut bacteria (Figure 14 D, E). Seven of eight high microbiota-reactive mAbs nonetheless showed high reactivity to bacteria isolated from individual human fecal samples. However, high microbiota- reactive mAbs showed donor dependent differences of binding capacities to microbiota from both HF donors and UC patients (Figure 14 D, E). The observed variability of binding profiles and capacities to gut microbial members from individual human fecal samples most likely reflect global differences of microbial species composition and diversity between human samples. One might further speculate that additional factors may affect antibody binding capacities to gut bacteria that explain differences in binding capacity, such as transcriptomic states of gut microbiota and gut-environmental modification of epitopes. While potentially not capturing the full scope of antibody reactivity to microbial members present in murine fecal samples, our data nevertheless demonstrate that high microbiota-reactivity is a feature shared across a broad range of fecal samples irrespective of the host-origin. However, in light of the variability of binding profiles and capacities of mAbs to bacteria from different human donors and the uncertain impact of human endogenous Ig coating the microbiota, we decided to further characterize mAb binding and specificities to murine gut bacteria.

Figure 14: High microbiota-reactive intestinal IgA mAbs show donor-dependent variability of binding capacities to human gut bacteria. Bacteria were isolated from fecal samples of 5 healthy individuals (HF) and 5 ulcerative colitis (UC) patients and stained with directly AF647-conjugated mAbs. (A, B) α-diversity parameters determined for species richness (A) and Shannon effective count (B) of 5 HF and 5 UC derived fecal samples. (C) β-diversity displayed as principal coordinates analysis (PCoA) of generalized Unifrac distances. PCoA plots show comparison of bacterial composition between HF and UC fecal samples. (D) Representative FACS plots of mAb microbiota-reactivity to human gut bacteria. Microbes were stained with Syto-9 nucleic acid dye and microbiota reactivity of AF647-directly conjugated mAbs assessed by flow cytometry. Data are representative of two independent experiments and numbers in plots show mean ± SD. (E) HF and UC derived fecal samples were characterized by 16S rRNA amplicon sequence analysis and organized as a dendrogram based on their bacterial genomic sequence composition. Heat map depicts mean binding capacity of single mAbs to bacteria isolated from HF and UC samples.

5.4 IgA-microbiota interactions are Fab-mediated and independent of glycosylation A prominent feature of luminal SIgA is the strong glycosylation of the Fc-region, J chain and hinge region147,195,196. However, not only both antibody chains and the J chain are glycosylated, but also the secretory component (SC) is glycosylated. Luminal SIgA is therefore heavily glycosylated. SIgA associated glycans have been reported to mediate binding to members of the gut microbiota, in particular to gram-positive bacteria25,192,195,196. In this respect, both canonical Fab-dependent and non-canonical glycan-mediated binding can confer antibody- microbiota interactions. Although, the monoclonal recombinant antibodies used in this study, were expressed as human Fc-IgG1 hybrid proteins independently of their original isotype and subclass, we nevertheless wanted to determine the potential contribution of glycan-moieties of mAbs to their high microbiota-reactive interactions. To directly discern whether the high- microbiota reactivity of mAbs rely on non-canonical glycan-mediated interactions, we deglycosylated our collection of high microbiota-reactive mAbs (Figure 15). To rule out that deglycosylation did not affect the overall antibody protein structure, we confirmed Ig protein mass and deglycosylation efficiency by western blot (Figure 15 A). Deglycosylation efficiency was on average > 80%, which was in accordance with additional MS analysis and confirmed efficient deglycosylation (Table 5). The MS analysis was performed in collaboration with the group of Professor Emma Slack at the ETH in Zurich. Notably, deglycosylation of mAbs did not affect their overall high-binding capacity to members of the gut microbiota, which was evident across several independent experiments (Figure 15 B, C). Moreover, as microbiota binding capacity of mAbs was assessed by flow cytometry using an anti-human Fc-IgG1 detection antibody, it is unlikely that the binding patterns described here rely on Fc-mediated and/or glycan-mediated, non-canonical antibody-microbiota interactions. Together, these results show that tested IgA mAbs bind to gut microbial members by canonical Fab-dependent interactions.

Table 5: Efficiency of Fc-Ig mAb deglycosylation by EndoS2 endonuclease. mAb sample % deglycosylated ED38 HD2a7 HD2a88 HD3a14 HD3a75 HD3a103 HD3a147 #1 88.3 94.5 84.9 79.9 89.9 90.4 81.5 #2 88.6 95.3 86 77.9 83 79 91.9 The relative amount of deglycosylated antibody fragments to non-deglycosylated samples with equal protein concentrations was calculated as the ratio of deglycosylated sample to the reference sample in two independent assays (#1 and #2).

Figure 15: High microbiota reactivity of IgA mAbs is not abrogated by deglycosylation. (A) Representative western blots showing Ig protein detection and mass of high microbiota-reactive glycosylated (mAb) and deglycosylated (dg) HD IgA mAbs (upper plot) and deglycosylation efficiency using concanavalin-A (Con A), (lower plot). (B) Representative FACS plots showing staining of bacteria (gated on FSC/SSC profile and Syto-9-positive events) with control mAb (ED38) and selected high microbiota-reactive HD glycosylated mAbs or deglycosylated mAbs. Numbers indicate percentage of gated events. (C) Comparison of microbiota-reactivity of glycosylated (white bars) and deglycosylated (grey bars) mAbs. Symbols denote independent experiments, bars indicate means and lines connect data points of respective experiments. Significance was tested by paired Mann-Whitney test, ns – not significant. Deglycosylation efficiency (in %) is shown in Table 5. Technical assistance for the deglycosylation assays and Western blot analysis was provided by Sabine von Oy.

5.5 Intestinal microbiota-reactive IgA antibodies show broad VH gene usage Having shown that high microbiota-reactivity of mAbs is Fab-mediated, we further went on to interrogate in more detail the features of IgA mAbs that might explain high microbiota-reactivity. We chose to focus on IgA mAbs only, as IgA is the dominant isotype in the gut lumen with vast microbiota targeting properties. A recent study pointed to the interesting properties of specific microbial taxa that exhibit “super antigen” like surface structures300. Unlike typical B cell super antigens, these “super antigen” like structures are expressed by microbial members belonging to the family of Lachnospiraceae300. The study reported that B cells with particular VH3 gene 60 family usage are more prone to be induced by these “super antigen” like structures. Subsequent antibody responses showed enhanced binding to members of the family Lachnospiraceae resulting in high IgA-coating of these gut bacteria. As VH3 gene family usage is commonly overrepresented among human B cells, we investigated whether VH3 gene family usage is overly expressed among our high-microbiota reactive mAbs. In accordance with previous studies216,233,300, also in our HD and CD IgA mAb collections VH3 family encoded antibodies were dominant (Figure 16 A, B). However, we did not find any enrichment of VH3 gene usage among high microbiota-reactive IgA mAbs as compared to the overall collection of antibodies irrespective of their microbiota binding capacity (Figure 16 C, D). We can therefore rule out that the herein described high microbiota-binding patterns rely on the biased usage of a distinct VH gene family.

Figure 16: Intestinal microbiota-reactive IgA mAbs show broad VH gene usage. Relative distribution of VH gene usage of all tested HD derived IgA mAbs (A) and CD derived IgA mAbs (B). HD IgA mAbs (C) and CD IgA mAbs (D) were ranked according to microbiota binding capacity as no/low, intermediate or high microbiota-reactive and assigned to their VH gene segment usage.

5.6 Intestinal IgA shows broad but distinct binding to members of the gut microbiota To characterize the microbiota binding profile of intestinal antibodies, we selected 15 high microbiota-reactive mAbs for detailed analysis of targeted fecal bacteria. An established tool for the detailed analysis of IgA-bound and un-bound microbial members is high throughput 16S rRNA sequencing of PCR amplified bacterial V3/V4 gene segments that code for 16S ribosomal DNA (Figure 17).

Figure 17: Representative workflow of bacterial DNA isolation and 16S rRNA gene amplicon generation. Bacteria were isolated from murine fecal material and isolated bacteria stained with recombinant antibodies (mAbs). Metagenomics DNA of sort purified stained and unstained bacteria were used for V3/V4 16S rRNA gene amplification. Purified PCR products were used for 16S rRNA library preparation followed by high throughput community 16S rRNA illumina sequencing290,291. Technical assistance for the 16S sample preparations was provided by Christina Petrick.

After bacterial DNA amplicons were subjected to 16S rRNA sequencing, raw reads were processed using the bioinformatics database (IMNGS)292. Processed reads for each sample were evaluated, statistically analyzed and visualized using a set of R scripts296 (Figure 18).

Figure 18: Workflow illustrating the analysis of 16S rRNA sequencing data. 1. After Illumine Miseq sequencing, raw reads in Fastq format are assigned to their corresponding samples via demultiplexing and each sample denoted with a unique pair of barcodes. Demultiplexing was performed using a custom-made Perl script. This step generates input files for further analysis on the IMNGS (www.imngs.org)292 platform. 2. After demultiplexing, all sequences were trimmed to the first base with a quality score < 3. Data were further analyzed based on the UPARSE approach293 USEARCH was used for pairing, quality filtering and operational taxonomic (OTU) clustering. UCHIME was used for chimera filtering and taxonomic classification by RDP (Ribosomal Database Project294). OTUs were clustered at a sequence similarity of 97%, and only sequences with a relative abundance ≥ 0.25% were included in the analysis. Sequences were aligned by MUSCLE301 and treeing was performed by Fasttree302. 3. Downstream analyses of intermediated IMNGS files was performed using Rhea in the R programming environment292,296. Samples were normalized and α-diversity parameters (richness and Shannon effective count) computed to assess within sample diversity of microbial profiles. To compare microbial profiles between samples β-diversity measurements, computed as PCoA analyses based on generalized UniFrac distances, were used297.

To investigate which gut bacteria isolated from RAG2-deficient mice were bound and not bound by the selected high microbiota-reactive mAbs, we FACS sort-purified both fractions in addition to the fecal input material and performed 16S rRNA amplicon sequencing (Figure 19 A). In order to gain a robust description of bacterial members that are bound by IgA mAbs independent of differences in microbial compositions, we assessed mAb binding profiles in two separate sets of experiments using feces from non-cohoused mice. Next, we examined the α- diversity of both mAb positive and mAb negative samples and input material, measured as richness and Shannon effective count. Both α-diversity measures determine the microbial complexity within a given sample based on the relative abundance of detected operational taxonomic units (OTUs) that is equivalent to the composition and diversity of bacterial species. While richness merely estimates the bacterial complexity based on the overall detected OTUs within each sample, the more elaborated Shannon effective count also integrates the evenness/numeric abundance of bacterial species (OTUs) and can therefore be considered as a proxy for numbers of dominant species within a given sample. For both measures, higher values indicate more diverse microbial communities (Figure 19 B). The richness values showed no striking differences between samples, indicating that each mAb bound multiple bacterial taxa (Figure 19 B). However, the Shannon effective count of mAb positive bacteria in particular, showed a notable variability suggesting a heterogeneity of mAb binding profiles to bacteria. This heterogeneous distribution among mAb positive samples in comparison to the negative fractions, however, points to an overall broad bacterial binding profile of individual mAbs. The variation of Shannon effective counts among mAb coated bacterial fractions may further indicate that each individual mAb binds multiple dominant OTUs, but also indicates potential differences of mAb binding specificities and profiles between mAbs. Notably, this observation was evident for IgA mAbs derived from both HD and CD donors. Interestingly, there was no correlation between the percentage of bound bacteria (microbiota reactivity) and the overall diversity of bound species (Shannon effective count) amongst the high-microbiota reactive mAbs (Figure 19 C). These results further suggest that the overall frequency of bound bacteria does not necessarily correlate with increased or decreased binding to diverse bacteria and that mAbs with relatively high microbiota binding capacities (> 5%) enrich for similar complex bacterial members as compared to mAbs with binding capacities > 20% of total bacteria. However, α-diversity measures cannot describe true differences between samples, as these parameters only estimate the OTU composition within individual samples. As α- diversity measures showed heterogeneity in microbial diversity of HD and CD derived mAbs positive fractions, we next compared differences between samples of HD and CD mAbs employing β-diversity. The measure of β-diversity accounts for dissimilarities between samples based on their microbial composition. To illustrate differences between samples, we employed principal coordinate analysis (PCoA) based on the phylogenetic composition within samples, computed as generalized UniFrac distances296,297. Surprisingly, positive samples from both HD and CD donor derived antibodies (i.e. healthy and inflamed intestine) showed largely overlapping binding profiles to gut microbiota (Figure 19 D, left panel). While several studies demonstrated that shifts in microbial community structure and diversity are associated with gut inflammation46,110,151, broad binding profiles of mAbs from healthy and inflamed gut were yet similar. This may indicate, that despite shifts in the microbiota composition under inflammatory 64 conditions, IgA specificity for particular members of the microbiota is not fundamentally affected. In contrast, β-diversity analysis revealed major differences between enriched bacteria in mAb positive and mAb negative samples for both HD and CD derived antibodies, illustrated by clear cluster formation of positive and negative samples (Figure 19 D, middle and right panel). In summary, both α- and β-diversity measures show that single high microbiota-reactive intestinal IgA antibodies derived from both healthy and inflamed gut do not only enrich for individual gut bacteria, but instead show specific broad binding to multiple bacterial taxa. Notably, microbial profiles found for mAb positive fractions were distinct from mAb negative fractions and input material.

Figure 19: Intestinal high microbiota-reactive IgA mAbs bind diverse groups of gut bacteria. (A) Representative schematic of cell sort purification of mAb positve (mAb pos bac) and mAb negative (mAb neg bac) bacteria. Analysis of diversity and composition of sorted bacteria was obtained by 16S rRNA amplicon sequencing. (B) α-diversity parameters (determined for species richness and Shannon effective count) of sorted mAb positive (pos) and mAb negative (neg) bacteria displayed for two separate experiments (#1 and #2). Groups were compared by one-way ANOVA and paired Wilcoxon test (* p ≤ 0.05, ns-not significant). (C) Pearson’s correlation (two-tailed) analysis of α-diversity (Shannon effective count) and mean microbiota reactivity in percent (%) of healthy donors (HDs) (filled symbols) and Crohn’s disease (CDs) patients (open symbols) derived mAbs. Symbols represent the mean of two independent experiments. (D) β-diversity displayed as principal coordinate analysis (PCoA) of generalized Unifrac distances. PCoA plots show comparison of bacterial composition between HD and CD mAb positive fractions from (left plot) and between HD mAb positive (open symbols) and mAb negative (filled symbols) fractions (middle plot) and CD mAb positive (open symbols) and mAb negative (filled symbols) fractions (right plot).

5.7 High microbiota-reactive intestinal IgA mAbs are cross-species reactive In order to comprehensively characterize bacterial taxa targeted by HD and CD derived IgA mAbs and to identify bacterial taxa that explain the observed differences in β-diversity, we examined the relative abundance of detected OTUs in mAb positive and mAb negative samples (Figure 20 A). The relative abundance of detected OTUs in positive and negative samples of HD and CD derived antibodies is illustrated as log10 transformed relative abundance values for each detected OTU. We only considered relative abundances of detected OTUs in positive samples ≥ 0.5% prevalence (Figure 20 A). Consistent with α-and β- diversity measures, high microbiota-reactive mAbs showed broad binding to different bacteria that could be assigned to OTUs belonging to phylogenetically distant taxonomic groups. However, detected OTUs in mAb positive fractions were typically also detectable in mAb negative fractions (Figure 20 A). Therefore, using the relative abundance of detected OTUs alone to assess specific enrichment for distinct OTUs in mAb positive fractions did not seem applicable. To identify OTUs that were enriched in the antibody-bound fraction, we employed an enrichment index for each mAb as the ratio of relative OTU abundance in the mAb-positive fraction and negative fraction plus taking into account the overall relative abundance in the positive fraction (see Materials and Methods Part I). For all tested mAbs, the index revealed enrichment of several OTUs with distant phylogeny. 16 OTUs were targeted by more than three individual mAbs, indicating that these OTUs are preferential targets of tested mAbs (Figure 20 B, Table 6). Highlighted OTUs were defined by two-fold enrichment in the mAb positive fractions of at least three individual mAbs and were referred hereafter as “commonly targeted OTUs”. Taxonomic assignment of these commonly targeted OTUs included gut microbial members of the main abundant phyla found in the murine gut, Actinobacteria, Firmicutes, Proteobacteria, Bacteroidetes and Deferribacteres (Figure 20 A, B, Table 6). Commonly targeted OTUs included the following species: Enterocloster bolteae (OTU 16), Mucispirillium schaedleri (OTU 40), Alcaligenes faecalis (OTU 41), Helicobacter typhlonius (OTU 6), and several species identity-matched to members of the family Muribaculaceae (OTU 7, 15, 18, 19, 20, 31, and 32). These taxa had previously been reported to be coated by endogenous IgA107,151,170,299. In contrast, some OTUs showed particularly strong enrichment by only single mAbs and included OTUs with an overall low relative abundance, e.g. Moraxella osloensis OTU 45 (HD2a7), Bacteroidetes acidifaciens OTU 5757 (HD3a103), Muribaculum sp. OTU 4 (CD1a293) and Lactobacillus gasseri OTU 39 (CD2a146). Importantly, these unique enrichment patterns were observed for mAbs derived from both HD donors as well as CD patients (Figure 20 A, B). This indicates that intestinal IgA antibodies not only target dominant members of the microbiota but also target underrepresented populations within microbial communities. Our data further demonstrate that individual intestinal antibodies with high microbiota-reactivity can bind a broad range of microbial species while exhibiting unique binding profiles. Importantly, we observed phylogenetically unrelated taxa among commonly and selectively targeted OTUs, suggesting that single monoclonal human IgA antibodies can bind to phylogenetically unrelated groups of gut bacteria. In the following we will refer to this mode of binding as cross-species reactive binding91. We further conclude that specific cross- species reactivity is one principle property of intestinal plasma cell-derived antibodies in the adult human healthy and inflamed gut. 67

Figure 20: High microbiota-reactive intestinal IgA is cross-species reactive. (A) Detected operational taxonomic Units (OTUs) are organized as a taxonomic dendrogram (phylum to family) based on the RDP taxonomic classification294. OTUs that are marked by asterisks denote commonly or selectively targeted OTUs. Heat map shows Log10 transformed relative abundance for OTUs with ≥ 0.5% relative abundance. (B) OTUs are organized as phylogenic dendrogram in the center of the plot (phylum to family) based on the RDP taxonomic classification. Only OTUs with ≥ 0.5% relative abundance in the mAb positive fraction are displayed. mAb specificity to a given OTU is shown as enrichment index (as defined at the bottom of the figure). 16 commonly targeted OTUs, i.e. OTUs enriched 2-fold by at least three mAbs are marked by asterisks, indicated in color and listed below the figure. Selectively targeted OTUs, i.e. OTUs enriched by single mAbs are marked by asterisks and listed. Non-listed phyla: Deinococcus-Thermus (green), Cyanobacteria (brown), Deferribacteres (yellow). (A, B) The lineage and taxonomic identity (closest species with a valid name and corresponding 16S rRNA gene sequence identity) of marked OTUs were obtained using EZbiocloud295, taxonomic classifications of all OTUs are listed (Table 6). Dr. Tom Hitch and Professor Thomas Clavel from the institute of medical microbiology at the RWTH Aachen University Clinic substantially contributed to the 16S rRNA sequence analysis and Dr. Tom Hitch performed the dendrogram taxonomic clustering.

5.8 IgA cross-species reactivity is not conferred by polyreactivity Next, we sought to examine the underlying mechanisms that potentially account for the property of intestinal IgA antibodies to show cross-species reactivity. Recently, one mechanism explaining cross-species reactivity was demonstrated in mice, suggesting a correlation of broad microbial binding and polyreactivity (i.e. binding to multiple, structurally dissimilar antigens) of murine intestinal antibodies20. To examine whether polyreactivity is a crucial feature, or at least contributes to cross-species reactivity of human intestinal antibodies, we tested the HD and CD collections of IgA and IgG derived antibodies for binding to a panel of seven unrelated antigens commonly used to determine polyreactivity. Both, IgG and IgA mAbs were considered polyreactive when binding to at least two unrelated antigens. Based on this criterion, we did not find an enrichment of polyreactive mAbs among high microbiota- reactive IgA or IgG mAbs derived from HDs or CD patients (Figure 21).

Figure 21: High microbiota-reactive mAbs are not enriched among polyreactive antibodies. IgA and IgG mAbs from healthy donors (HD) and Crohn’s disease (CD) patients were ranked according to their microbiota binding capacity (top row) as no/low reactive antibodies, intermediate or high microbiota- reactive antibodies (% binding is depicted above plots). Polyreactivity of intestinal IgA and IgG mAbs derived from HDs or CD patients were tested for binding to a set of seven non-related antigens (insulin, cardiolipin, flagellin, KLH, LPS, albumin, calf-thymus DNA) by ELISA (OD405). mAbs were considered polyreactive when binding to more than two structurally unrelated antigens above cut-off and classified as either non-polyreactive (white) or polyreactive (black) as described for HD derived mAbs106. 70

In line with reported frequencies of polyreactivity among intestinal plasma cells106, we consistently found 23% of HD IgA mAbs and 13% of CD IgA mAbs that showed polyreactivity among all screened IgA antibodies (Figure 22 A). To assess if polyreactivity might be a critical mechanism for high microbiota-binding of IgA, we determined the frequencies of polyreactive mAbs within our previously described categories for microbiota binding capacities (Figure 22 B). We found similar frequencies of polyreactive IgA mAbs of HD and CD origin among no/low, intermediate and high microbiota-reactive antibodies. Consequently, polyreactive mAbs were not overrepresented among mAbs with high microbial binding capacity (Figure 22 B). These findings were confirmed by correlation analysis, which did not reveal any positive relationship between polyreactivity and percentage of mAb-bound bacteria (Figure 22 C). While dimeric SIgA is the prevalent isotype in the gut lumen, IgG is also found83,108,110. However, trans- epithelial gut luminal transport of IgG is less well understood. Hence, IgG at gut mucosal surfaces might differ in functionality and specificity. We therefore applied the same criteria as described for IgA mAbs to test the collection of IgG mAbs from HDs and CD patients for polyreactivity. We found similar frequencies of polyreactive antibodies among the overall panel of screened HD (23% among HD IgG mAbs) and CD (18% among CD IgG mAbs) derived IgG mAbs compared to IgA mAbs (Figure 22 A, D). Notably, polyreactive IgG antibodies derived from HDs were observed at similar frequencies among no/low, intermediate and high microbiota reactive antibodies, and correlation analysis did not show any positive relationship between microbiota binding capacity of mAbs and polyreactivity (Figure 22 E, F). Consequently, polyreactive HD derived IgG mAbs were not overrepresented among antibodies with high microbiota-reactivity. In contrast, IgG mAbs derived from CD donors showed a slight positive correlation between microbiota binding capacity and polyreactivity (Figure 22 E, F). However, it needs to be noted that IgG mAbs derived from CD patients did not contain antibodies with no/low microbiota binding capacity and only contained few numbers of high- microbiota reactive mAbs. The analysis of a greater number of high microbiota-reactive CD derived IgG mAbs might give more insight into whether polyreactivity is indeed a distinguishing mechanism for IgG-microbiota interactions in the inflamed gut.

Figure 22: High microbiota reactivity of intestinal IgA does not correlate with polyreactivity. (A, D) Pie charts depict the frequency of non-polyreactive (white) and polyreactive (black) mAbs among all screened IgA (A) and IgG (D) mAbs derived from healthy donors (HD) or Crohn’s disease (CD) patients. Number of tested mAbs is shown in the pie chart center, and other numbers indicate frequencies of polyreactive and non-polyreactive mAbs. (B, E) Bar charts show the relative distribution of polyreactive and non-polyreactive mAbs among no/low, intermediate and high microbiota-reactive HD and CD derived IgA (B) and IgG (E) mAbs. (C, F) Correlation analysis between polyreactivity and microbiota reactivity was determined by linear regression. p-values are displayed in the plots.

5.9 Microbiota-reactive intestinal antibodies carry high numbers of somatic mutations Our data thus far demonstrate, that non-clonally related mAbs derived from HDs or CD patients show broad but specific binding to phylogenetically unrelated bacteria, which we refer to as cross species-reactivity. Since we did not find an overrepresentation of polyreactive antibodies among high microbiota-reactive mAbs and no impact of antibody glycosylation to microbial binding, we further investigated the possibility that the high capacity for microbiota binding in human intestinal antibodies might rely on accumulated somatic mutations. As had been demonstrated before in adult individuals, intestinal IgA typically contained high numbers of somatic mutations in both heavy and light chain V gene sequences106,208,210,268. Screening the HD and CD mAb collections used in this study, we did not detect any major differences in the number of somatic mutations in either heavy or light chain V genes when comparing HD IgA mAbs with CD IgA mAbs (Figure 23 A). Also, mAbs with high microbiota-binding capacity showed on average no differences in the number of accumulated mutations compared to the number of somatic mutations found in the overall collection of IgA antibodies (Figure 23 A, B). In addition, we performed a correlation analysis to investigate whether the number of mutations in either heavy or light chain V genes correlate with the overall binding capacity to gut bacteria. We found no correlation between the number of mutations in heavy or light chains of HD or CD derived IgA mAbs and the capacity to bind gut bacteria (Figure 23 C, D).

Figure 23: High microbiota-reactive intestinal IgA+ plasma cells have high numbers of somatic mutations. (A) Number (No.) of somatic mutations in the heavy and light chain V genes of all screened IgA mAbs from HDs and CD patients and for selected high microbiota-reactive IgA mAbs (B). Significance was tested by Mann-Whitney test, ns-not significant. (C, D) Correlation between microbiota binding capacity (displayed as percentage of bound microbiota for mAbs showing binding ≥ 1% to bacteria) and number of somatic mutations in heavy and light chain V genes of HD derived IgA mAbs (C) and CD derived IgA mAbs (D). Mean microbiota binding capacity of selected high microbiota-reactive mAbs is indicted in red. p-values were obtained by linear regression and are displayed in the plots.

Conversely to intestinal IgA, IgG antibodies are usually not actively pIgR-transported into the gut lumen, but are yet present in the human gut83,108,110. Based on our results, IgG mAbs also have the capacitiy for microbiota binding (Figure 12 and Figure 22). Thus, we further investigated whether the number of somatic mutations might correlate with the microbiota- binding properties of IgG mAbs, e.g. higher numbers of mutations leading to increased microbiota-binding capacity. Similar to IgA mAbs (Figure 23), IgG mAbs from both HDs and CD patients showed the typical pattern of highly mutated intestinal antibodies in both heavy and light chain V gene sequences (Figure 24 A). Albeit, IgG mAbs were derived from healthy or inflamed gut, there was no difference in the mean number of mutations in either heavy or light chain V genes when comparing the HD and CD derived IgG antibody collections. Next, correlation analysis was used to address a potential interrelation between the number of mutations of IgG mAbs and the capacity to bind to gut bacteria. The overall number of mutations in either heavy or light chains of IgG mAbs did not correlate with their capacity to 74 bind microbiota, however, with the exception of a positive correlation between number of mutations in HD derived IgG light chains and microbial binding (Figure 24 B, C). Collectively, IgA as well as IgG derived mAbs from both healthy donors and Crohn’s disease patients show the typical pattern of highly mutated intestinal antibodies, strongly suggesting that they had undergone affinity-maturation.

Figure 24: Intestinal IgG+ plasma cells carry high numbers of somatic mutations. (A) Number (No.) of somatic mutations in the heavy and light chain V genes of all screened IgG mAbs from HDs and CD patients. Significance was tested by Mann-Whitney test, ns-not significant. (B, C) Correlation between microbiota binding capacity (displayed as percentage of bound microbiota for mAbs showing binding ≥ 1% to bacteria) and number of somatic mutations in heavy and light chain V genes of HD derived IgG mAbs (B) and CD derived IgG mAbs (C). p-values were obtained by linear regression and are displayed in the plots.

5.10 High microbiota reactivity of human intestinal IgA requires somatic mutations While germ-line (GL) encoded BCRs represent the genetic make-up of newly generated naïve B cells that are imminent to leave the bone marrow to enter circulation, somatic mutations on the other hand, demonstrate the status of affinity-matured non-naïve B cells and intestinal PCs (Figure 25). Affinity maturation is typically characterized by accumulated somatic mutations of the BCR resulting in increased affinity and specificity of secreted antibodies to a given antigen. If accumulated somatic mutations of mAbs indeed contribute to their binding capacity and specificity to microbial members, one would expect a marked decrease in microbial-binding capacity of mAbs in their GL configuration. To directly address the contribution of somatic mutations to the microbiota-binding capacity and specificity of mAbs to gut bacteria, we generated predicted GL variants of selected IgA antibodies from both sets of HD and CD derived mAbs (Figure 25). In order to generate mAb GL variants, we first acquired the gene sequences of Ig heavy and light chains of mutated mAbs and determined the respective GL sequence with highest homology and identity. Ig heavy and light chain GL sequences were subsequently cloned into expression vectors and GL mAbs generated in mammalian HEK 293 cells. Flow cytometry was used to assess and directly compare the microbiota-reactivity of GL- reverted mAbs and their mutated counterparts (Figure 26 A).

Figure 25: Schematic depicting the generation of predicted mAb germ-line variants from their mutated counterparts. Germ-line (GL) sequences with highest homology and identity analyzed by IgBLAST and comparison with GenBank (http://www.ncbi.nlm.nih.gov/igblast/) were used to generate double stranded DNA fragments (gBLocks) containing restriction sites for vector cloning and subsequent expression as recombinant GL mAbs in mammalian HEK 293 cells. Ig protein concentrations were determined by ELISA. Mutated mAbs and their corresponding GL variants were tested for their microbiota-reactivity using flow cytometry (FACS).

Flow cytometry analysis of microbiota-binding capacity showed that all GL reverted IgA mAbs derived from HDs significantly lost their capacity to bind bacteria compared to their mutated counterparts (Figure 26 A, B). Similarly, the majority of GL reverted mAbs derived from CD patients showed substantial loss of microbiota binding. Notably, a small fraction of CD derived GL mAbs (2 mAbs) showed an altered binding profile, and another two GL mAbs had unchanged binding capacities as compared to their mutated counterparts (Figure 26 A, B). To ascertain that the observed loss of microbiota-binding capacity after GL reversion was in fact due to the germ-line reversion itself, correct Ig protein mass was verified by western blot. Western blot analysis confirmed the correct mass of Ig heavy and light chains of both mutated mAbs and their GL reverted counterparts (Figure 26 C). Collectively, these data show that in the majority of cases GL reversion significantly abrogated high microbiota-binding capacity or resulted in altered binding patterns to gut microbiota. We therefore conclude that in the human gut, somatic mutations crucially contribute to high microbiota-binding capacity and cross-species reactivity of intestinal IgA antibodies.

Figure 26: Somatic mutations confer high microbiota-binding capacity of intestinal IgA mAbs. (A) Representative FACS plots of fecal bacteria stained (gated on FSC/SSC profile and Syto-9 positive events) with IgA mAbs carrying somatic mutations or their respective germ-line (GL) variants. Numbers indicate percentage of gated events. (B) Pairwise comparison of microbiota-binding capacity of IgA mAbs (filled symbols) and their respective germ-line variant (GL) (open symbols) derived from HDs and CD patients. Symbols represent the mean value of two independent experiments. Significance was tested by paired Mann-Whitney test, ** p ≤ 0.01. (C) Representative Western blot to illustrate correct protein mass of heavy and light chains of mAbs before and after GL reversion.

5.11 Germ-line reversion of intestinal IgA mAbs does not cause polyreactivity Our data demonstrate that somatic mutations significantly contribute to the high-microbiota binding capacity and cross-species reactivity of intestinal IgA mAbs. However, several studies reported the presence of polyreactive “natural or primitive” IgA contributing to the pool of intestinal antibodies able to coat gut bacteria18,20,216,225. Hence, one might speculate that, while the majority of high microbiota-reactive IgA mAbs are not polyreactive, a fraction of intestinal IgA antibodies may have originally derived from PCs that were polyreactive in their GL configuration. To address whether GL reversion renders antibodies polyreactive, we tested our collection of GL reverted HD and CD derived IgA mAbs and their mutated counterparts for polyreactivity using the standard panel of 7 non-related antigens (Figure 27 A). Our results demonstrate that GL reversion of antibodies did not render antibodies polyreactive. In contrast, mAbs that were polyreactive in their mutated state (HD2a7, CD3a32 and CD3a565) lost polyreactivity after GL reversion, indicating that these antibodies might have become polyreactive during the course of affinity-maturation (Figure 27 A, B). Collectively, our data demonstrate that microbiota-binding of mAbs is unrelated to polyreactivity and that the majority of cross-species reactive IgA mAbs did not originally derive from polyreactive antibodies.

Figure 27: Germ-line reversion does not lead to polyreactivity of intestinal IgA mAbs. (A) Polyreactivity ELISA (OD450) showing the binding of selected high microbiota-reactive HD and CD derived IgA mAbs and their corresponding germ-line (GL) variants to the standard set of seven unrelated antigens (insulin, cardiolipin, flagellin, KLH, LPS, albumin, calf-thymus DNA). mAbs were considered polyreactive when binding to more than two structurally unrelated antigens above cut-off (dashed red line is OD450 of mGO53 at 1 µg/ml minus background). Green lines show mAbs carrying somatic mutations, blue lines GL variants, black lines show the polyreactive positive control antibody (ED38) and red lines show the negative control antibody (mGO53). (B) Heat map representation of polyreactivity ELISA and microbiota-reactivity. Polyreactive or microbiota-reactive mAbs and GL variants are indicated in black.

5.12 Germ-line variants with sustained microbiota binding do not show relevant changes in their specificities to gut bacteria While the majority of CD derived IgA mAbs significantly lost their binding capacity to gut bacteria after GL reversion, four CD-derived GL mAbs retained a sufficient microbiota binding capacity to allow for sort-based purification of antibody-bound bacteria. Two of these GL reverted antibodies (CD2a148 and CD3a549) maintained high binding capacity to bacteria and two other mAbs (CD1a293 and CD2a127) showed overall reduced but also altered binding patterns (Figure 26 A). To investigate whether these GL antibodies would show differences in specificities to gut bacteria as compared to their mutated counterparts, we sort-purified bound and un-bound fractions of mutated mAbs and their GL variants across two independent experiments and subjected them to 16S rRNA gene sequencing (Figure 28). α-diversity measures indicated no significant differences in microbial diversity of bound bacteria by mutated mAbs or their GL reverted counterparts (Figure 28 A). Similarly to mutated mAbs, 80 their GL variants also showed binding to a broad array of gut bacteria (Figure 28 A). To determine differences between positive samples of mAbs and their GL variants (β-diversity), we employed PCoA analysis of bound bacteria. β-diversity analysis revealed overall close clustering of OTUs enriched by mAbs CD2a148 and CD3a549 in both their GL and mutated configuration (Figure 28 B). In contrast, the mutated mAbs CD1a293, CD2a127 and the GL variant of CD2a127 segregated from the rest of mAbs and GL variants, indicating overall differences of microbial-binding specificities. Curiously, the GL variant of mAb CD1a293 showed overlapping binding profiles with the majority of mAbs and GL variants, but the GL variant of mAb CD2a127 was more similar to its mutated counterpart (Figure 28 B). This indicates that the microbial specificity of mAb CD2a127 was maintained in its GL configuration but was different from the binding profiles of the other mAbs and GL variants (Figure 28 B). Thus, some mAbs (CD2a148 and CD3a549) showed similar microbial binding patterns regardless of their GL reversion, whereas for other mAbs (CD1a293 and CD2a127), GL reversion resulted in variable relative abundance of enriched OTUs (Figure 28 C). For mAb CD1a293, the OTUs 3, 4, 6, 9 and 263 showed a marked reduction in the positive sample of its GL configuration, which may explain the observed differences in β-diversity. Similarly, we found a reduction of OTU 3 and OTU 9 in the GL positive fraction of antibody CD2a127 as compared to its mutated counterpart. Nonetheless, both the mutated and the GL configuration of these mAbs bound overall similar bacterial species, although with differing efficiency (Figure 28 C, D). This in turn indicates that even though germ-line reversion of CD1a293 and CD2a127 retained some microbiota-binding capacity of these mAbs, somatic mutations in these two antibodies may contribute to their microbiota specificity (Figure 28 C). However, comparing shared and selectively enriched OTUs showed that mAbs and their GL variants had largely overlapping binding profiles, regardless of their binding efficiency (Figure 28 D).

Figure 28: Somatic mutations contribute to the microbiota binding profile of human intestinal IgA. Fecal bacteria bound by mutated mAbs and their germ-line (GL) variants were obtained by cell sorting and analyzed by 16S rRNA sequencing in two sets of experiments (#1 and #2). (A) α-diveristy (richness and Shannon effective count) are displayed for fecal input material, mAb-positive (filled symbols) and GL-positive (open symbols) fractions. (B) β-diversity displayed as PCoA plots (based on generalized Unifrac distances) show comparison between mAb-positive (filled symbols) and corresponding GL- positive (open symbols) samples. (C) Dendrogram clustering of OTUs (phylum annotations shown in color) based on the RDP taxonomic classification294. Only OTUs with a relative abundance ≥ 0.5% in any positive fraction are displayed. Heat map shows Log10 transformed relative abundances of OTUs. (D) Venn diagram of shared and selectively enriched OTUs in mAb-positive and GL-positive samples. Selectively enriched OTUs are marked by asterisks, EZbiocloud295 identified and listed.

5.13 Intestinal IgA shows reduced binding to oligoMM12 fecal bacteria During our investigations of IgA cross-species reactivity, we observed that individual antibodies showed incomplete binding to abundant gut bacteria; i.e. highly abundant OTUs in mAb positive fractions were also prevalent in negative fractions. This incomplete or partial binding to members of the gut microbiota potentially highlights inherent factors of a complex microbial system found in SPF mice. While this incomplete binding may simply reflect technical limitations, it may also indicate a hitherto unsuspected genetic or growth-state heterogeneity within abundant OTUs, impacting SIgA-microbiota interactions. To further investigate microbial properties contributing to SIgA-interactions and underlying mechanisms of SIgA-microbiota interactions, we employed a mouse system with a defined microbial structure, yet replicating the situation of a complex microbiota. OligoMM12 mice are colonized with a defined set of 12 highly conserved microbial strains303 (Table 7). Hence, this system allows for the analysis of antibody specificity and mechanisms of SIgA-microbiota interactions in a highly defined but less diverse microbial context. We employed flow cytometry screening and sort purification of mAb-bound and mAb-unbound bacteria with subsequent 16S rRNA gene amplicon sequencing to assess the binding capacity as well as the binding profile of mAbs to microbial members of the oligoMM12 consortium. First, we asked whether the high microbiota-binding capacity of selected mAbs found for bacteria isolated from SPF RAG2-deficient mice can be recapitulated using fecal bacteria from oligoMM12 mice. To rule out any other effect but the microbial composition, we tested our collections of high microbiota-reactive HD and CD derived IgA mAbs on bacteria isolated from SPF RAG2-deficient mice and oligoMM12 RAG2- deficient mice and compared their binding capacity by flow cytometry (Figure 29 A, B). This approach enabled direct comparison of mAb-binding capacity to bacteria from SPF and oligoMM12 mice. To account for correct staining of bacteria with mAbs, we included the afore mentioned high and low microbiota-reactive control antibodies (ED38 and mGO53) and Ig free supernatant (MedCTRL) (Figure 29 C). While the polyreactive positive control mAb (ED38) showed similarly high binding-capacities to bacteria isolated from either SPF RAG2-deficient or oligoMM12 RAG2-deficient mice, binding capacity of both HD and CD derived IgA mAbs was markedly reduced to oligoMM12 bacteria as compared to SPF gut bacteria (Figure 29 D, E). These data demonstrate that the high binding-capacity found for a complex SPF microbiota was significantly decreased for oligoMM12 bacteria (Figure 29 F). Therefore, we next examined if this decreased binding capacity to oligoMM12 bacteria may be related to fewer targeted species or overall reduced binding efficiency to bacteria.

Figure 29: Reduced microbiota binding capacity of IgA mAbs to oligoMM12 fecal bacteria. (A, B) Representative FACS plots of microbiota binding capacity of mAb HD2a7 to fecal bacteria from SPF RAG2-deficient mice (A) and oligoMM12 RAG2-deficient mice (B). (C) Binding of antibody-free supernatant (MedCTRL), (mGO53) low reactive and (ED38) high reactive control mAb to fecal bacteria from SPF RAG2-deficient mice (white) and oligoMM12 RAG2-deficient mice (grey). (D) Binding of HD mAbs to fecal bacteria from SPF RAG2-deficient mice (white) and oligoMM12 RAG2-defcient mice (grey). (E) Binding of CD mAbs to fecal bacteria from SPF RAG2-deficient mice (white) and oligoMM12 RAG2-deficient mice (grey). Bars show mean ± SD of at least 2 independent experiments. (F) Mean binding capacity of HD and CD mAbs to fecal bacteria from SPF RAG2-deficient mice (white) and oligoMM12 RAG2-defcient mice (grey). Significance was tested by paired Mann-Whitney test (* p ≤ 0.05; ** p ≤ 0.001).

5.14 Incomplete binding to in vitro cultivated oligoMM12 bacteria To further investigate the reduced binding capacity to oligoMM12 gut bacteria, we examined the binding of mAbs to single in vitro cultivated oligoMM12 bacteria (Figure 30). One might speculate that the combined binding frequencies of mAbs to single in vitro cultivated bacteria would altogether recapitulate the total binding frequency to fecal oligoMM12 bacteria (Figure 30 E). Moreover, we presumed that using single cultivated bacteria would allow to determine the extent of mAb-binding capacity to one particular bacterial species. We therefore cultivated all bacterial species of the oligoMM12 consortium under strict anaerobic conditions and used the selected HD and CD derived mAbs to stain individual bacteria (Figure 30 A - E). However, the majority of tested mAbs did not show binding to single bacteria above background (≥ 1%) (Figure 30 C, D). Although, none of the tested HD or CD mAbs showed appreciable binding to in vitro cultivated bacteria, we nevertheless found that the polyreactive mAb HD2a7 showed relatively high binding to multiple bacterial species as compared to the other HD mAbs, including Acutalibacter muris, Flavonifractor plautii and Muribaculum intestinale, as well as intermediate binding to Bacteroides caecimuris, Enterococcus faecalis, Lactobacillus reuteri and Clostridium clostridioforme (Figure 30 C, E). Antibody HD3a103 showed moderate selective binding to Blautia coccoides (Figure 30 C, E). Furthermore, three CD derived mAbs showed either moderate selective binding to individual bacteria or targeted multiple single bacteria (Figure 30 D, E). Antibody CD2a146 showed selective but moderate binding to Enterococcus faecalis and mAb CD3a549 to Muribaculum intestinale. In contrast, polyreactive mAb CD3a32 showed moderate binding to Bacteroides caecimuris, Bifidobacterium longum, Flavonifractor plautii and Lactobacillus reuteri (Figure 30 D, E). These results are in contrast to our expectations and point to important additional factors such as genetic heterogeneity or different growth-states within bacterial species that need consideration when employing in vitro approaches to assess antibody-binding capacity to bacteria.

Figure 30: Decreased binding of IgA mAbs to in vitro cultivated oligoMM12 bacteria. (A, B) Representative FACS plots of mAb HD2a7 binding to two single in vitro cultivated oligoMM12 bacteria. (C) Mean values of binding capacity of HD mAbs to in vitro cultivated oligoMM12 bacteria. (D) Mean values of binding capacity of CD mAbs to in vitro cultivated oligoMM12 bacteria. (C, D) Binding of ≤ 1% was considered background (dashed line). The 12 bacterial species are shown in colored bars and annotated below the plots. (E) Heat map depicts binding capacity (% bacteria binding) of HD and CD mAbs to microbiota from oligoMM12 RAG2-deficient mice and to single in vitro cultivated oligoMM12 bacteria.

5.15 High microbiota-reactive IgA targets distinct oligoMM12 bacteria In order to explain the observed reduced binding to oligoMM12 fecal bacteria and in vitro cultivated bacteria, we performed sort purification of mAb-bound and unbound bacteria to assess the binding profiles of mAbs to oligoMM12 fecal bacteria (Figure 31). However, due the afore mentioned technical limitations would not allow for reliable sort based purification of low frequency mAb-bound bacteria without facing the risk of insufficient sort purity and false positive microbial diversity within mAb-positive samples. Therefore, we were only able to perform bacterial sort purification with three mAbs that allowed for sufficient discrimination between mAb-positive and mAb-negative fractions (Figure 31 A, B). To minimize the variation in fecal microbial composition and to rule out potential cage differences, we used aliquots of pooled WT oligoMM12 fecal material. Notably, we did not find major differences in mAb-binding capacity to bacteria isolated from RAG2-deficient oligoMM12 mice, that lack soluble Ig, as compared to bacteria isolated from WT oligoMM12 mice that have endogenous Igs (Figure 31 A, B). We thus made use of endogenous IgA-coating in these mice as a proxy for IgA specificity to oligoMM12 bacteria in vivo. In addition, we used the polyreactive positive control antibody (ED38) enabling to discriminate between binding profiles of our cross-species reactive mAbs and binding profiles mediated by polyreactive microbiota-interactions. α-diversity measures showed that mAb-positive samples displayed variation in their richness parameters, whereas endogenous IgA-positive samples showed lower richness values compared to IgA-negative samples (Figure 31 C). In contrast, ED38 mAb-positive samples showed no difference in richness measures as compared to ED38 mAb-negative samples (Figure 31 C). However, when analyzing the Shannon effective count that takes the evenness of abundant species into account, positive samples of mAbs, endogenous IgA and ED38 showed reduced species diversity as compared to their negative samples (Figure 31 D). Lower species diversity in positive samples indicate, in light of the defined microbial context used here, that mAbs and endogenous IgA show specific and most importantly selective enrichment of distinct bacterial species (OTUs). Accordingly, β-diversity analysis showed distinct clustering between mAb- positive and negative samples (Figure 31 E). Positive samples of endogenous IgA clustered with mAb-positive samples and negative samples of endogenous IgA were overlapping with mAb-negative fractions. In contrast, ED38 mAb-positive samples showed closer clustering to the input material (Figure 31 E). This indicates that bacteria targeted by the three selected HD mAbs were distinct from non-targeted bacteria and input material as demonstrated before for binding specificity to SPF fecal bacteria. Interestingly, while mAb-bound bacteria showed overlapping specificity with endogenous IgA, they were distinct from bacteria bound by polyreactive control ED38 mAb (Figure 31 E). To examine which OTUs were targeted by mAbs and endogenous IgA, we analyzed the relative abundance of detected OTUs in both positive and negative fractions (Figure 31 F). Importantly, only 9 OTUs out of the initial oligoMM12 consortium could overall be detected in the input material (Figure 31 F). This however, is in line with reports suggesting that the 12 bacterial species constituting the oligoMM12 consortium colonize the murine gut with different efficiencies. In our hands the most dominant species (OTUs) detected were Akkermansia muciniphila, Bacteroides caecimuris and Clostridium clostridioforme (Figure 31 F). However, conversely to our data obtained for a complex SPF microbiota (Figure 20), we found a very prominent enrichment of Akkermansia 87 muciniphila in the input material as well as positive fractions of mAbs (Figure 31 F). Here, both mAbs and endogenous IgA showed close to complete binding of Akkermansia muciniphila (Figure 31 F). In addition, Bacteroides caecimuris and Clostridium clostridioforme also showed a high prevalence in the positive fraction of mAbs and endogenous IgA (Figure 31 F). However, as compared to the almost complete binding of Akkermansia muciniphila, the enrichment of Bacteroides caecimuris and Closridium clostridiforme was less pronounced, despite a higher relative abundance of these species in the input material (Figure 31 F). Importantly, mAbs and endogenous IgA showed also strong enrichment for Clostridium clostridiforme but not for Clostridium innocuum, suggesting highly and selective specificity for individual bacterial species (Figure 31 F). As described before, we employed an enrichment index for detecting significant OTU enrichment in positive fractions only (Figure 31 G). Based on the enrichment index we found that Akkermansia muciniphila was the dominant OTU (OTU 3) commonly enriched in positive fractions of all mAbs and also in positive samples of endogenous IgA and ED38 polyreactive control mAb (Figure 31 G). These results were in accordance with the high relative abundance of OTU 3 in mAb and IgA positive fractions, respectively (Figure 31 F). Notably, albeit endogenous IgA-coated bacteria were highly similar to the binding profiles of mAbs, endogenous IgA showed additional selective enrichment for Turicimonas muris (OTU 5). The polyreactive mAb ED38 also showed enrichment for Clostridium clostridioforme (OTU 1) and Flavinofractor plautii (OTU 8) (Figure 31 G). In line with our results obtained from a complex SPF microbiota (Figure 20), we also found some selective binding of individual mAbs to single OTUs from the oligoMM12 consortium. For example, Flavonifractor plautii (OTU 8) that was absent in positive fractions of other mAbs and endogenous IgA, was highly enriched in the positive fraction of mAb HD3a147 (Figure 31 G). Curiously, while bacterial members of the family Muribaculaceae were highly targeted by mAbs when derived from SPF feces (Figure 20), they were not enriched by mAbs when isolated from oligoMM12 fecal material (Figure 31 F, G). Although, bacterial specificity of endogenous IgA and our recombinant human mAbs were to some degree overlapping, endogenous IgA also showed additional bacterial targeting that was distinct from the binding profiles of mAbs. This in turn shows that endogenous IgA- coating may serve as a proxy for microbial-targeting profiles of intestinal antibodies, but also emphasizes the unique binding profiles of our collection of human IgA mAbs. These data demonstrate that in the context of the oligoMM12 consortium, IgA showed selective and nearly complete binding of distinct species (e.g. Akkermansia muciniphila) as compared to our results obtained for SPF microbiota.

Figure 31: High microbiota-reactive IgA mAbs show selective binding to oligoMM12 bacteria. (A) Representative FACS plots showing the binding capacity of HD mAbs to fecal bacteria from RAG2- deficient oligoMM12 and WT oligoMM12 mice and mean binding shown in (B). (C - G) Analysis of α- diversity and β-diversity and microbial composition of sorted WT oligoMM12 bacteria after 16S rRNA gene amplicon sequencing. (C, D) α-diversity parameters shown as species richness (C) and Shannon effective count (D) of sorted positive (pos) and negative (neg) bacteria of HD mAbs, ED38 polyreactive control mAb and endogenous IgA. Groups were compared by one-way ANOVA (** p ≤ 0.01, **** p ≤ 0.0001). (E) β-diversity PCoA plot (based on generalized Unifrac distances) shows comparison between HD mAb-positive and negative fractions, ED38-positive and negative fractions and endogenous IgA- positive and negative samples. (F) Detected OTUs are assigned to phylum level (indicted in colors) and listed to species level (Table 8). Heat map shows Log10 transformed relative abundance of OTUs occurring ≥ 0.1%. (G) Enrichment index (as defined at the bottom of the figure) of OTUs for endogenous IgA, HD mAbs and ED38 control mAb. (F, G) Data is shown as mean of three independent experiments performed on aliquots of pooled input material. Staining and sort purification of IgA/mAb-bound and unbound oligoMM12 fecal bacteria was performed in collaboration with Lena Küsgens during her masters at the institute of molecular medicine, RWTH Aachen University Clinic. 89

Conclusion In conclusion, our data demonstrate that high microbiota-binding and cross-species reactivity of intestinal IgA to fecal microbiota is not conferred by polyreactivity, but crucially depends on the accumulation of somatic mutations. Accordingly, the majority of IgA germ-line variants showed significantly decreased microbiota-binding capacities, strongly indicating that somatic- hypermutation is a crucial mechanism for high microbiota and cross-species reactive SIgA- microbiota interactions in the healthy and inflamed human gut. Furthermore, our data obtained using the low diversity oligoMM12 microbial consortium suggest that the extent of the coating of bacterial species by IgA can be affected by the complexitiy of the microbiota. Thus, SIgA- microbiota interactions may depend on additional host and environmental factors that need consideration when defining IgA specificity.

6 Results Part II In the first part of this thesis we have demonstrated that in the adult human gut individual intestinal IgA antibodies with high microbiota-binding capacity show specific cross-species reactivity to phylogenetically distant bacteria of the microbiota. Moreover, cross-species reactivity relies on Fab-dependent interactions. We thus propose that IgA cross-species reactivity relies on affinity-matured IgA responses to the microbiota that are most likely derived from T cell dependent pathways. In the following part of the thesis our aim was to examine the generation of IgA producing plasma cells (PCs) under steady state conditions and to identify dominant gut mucosal- immune inductive compartments for the priming of IgA committed B cells in vivo. Under homeostatic conditions, without neo-antigen challenge, activated B cells are constantly generated in gut immune-inductive compartments such as Peyer’s patches (PPs) and characteristically express AID. Activated B cells may differentiate into IgA-committed plasma blasts that will ultimately home as IgA-producing PCs to mucosal effector sites or differentiate into circulating IgA+ memory B cells. However, it is not fully clear to what extent newly activated B cells contribute to the prevailing PC pool in the small intestinal lamina propria (SI LP) under steady state conditions. Hence, this part of the thesis will largely focus on the methodological approach for the spatial-temporal fate tracking of activated B cells. We employed the transgenic mouse “AIDCreERT2Rosa26loxPeYFP” described before304 (hereafter referred to as AIDcre) that enables the conditional activation of an enhanced yellow fluorescent protein (eYFP) reporter in activated B cells (Figure 32), which allows for the long-term fate tracking of activated B cells and all their progeny. In this mouse model, Cre-recombinase is expressed as a fusion product with the estrogen receptor (ERT2). The Cre-ERT2 fusion protein is generated under the control of the activation-induced cytidine deaminase (AID) encoding gene promoter and is only active upon ligand administration for ERT2, that is tamoxifen or its hydroxylated form 4-Hydroxytamoxifen (4-OHT). AID is exclusively, but transiently, expressed in antigen- activated B cells. Cre-activation leads to the excision of the loxP-flanked stop cassette of the eYFP reporter protein gene that is transcribed under the control of the ubiquitous Rosa26 promoter. Tamoxifen or 4-OHT administration to AID expressing B cells thus leads to the constitutive expression of eYFP. Conversely to other systems, the expression of eYFP is not abrogated by cell division or differentiation (Figure 32). Therefore, this system is an excellent model to monitor the kinetics of IgA+ B cell generation and differentiation in distinct gut- inductive compartments and allows for the long-term fate tracking of activated B cells and homing properties of their progeny to gut effector compartments.

Figure 32: Transgenic mouse model enabling the spatial-temporal tracking of activated B cells and their progeny. This transgenic mouse model employs a conditional Cre-loxP system The ERT2- Cre-recombinase fusion product is expressed under the AID (Activation-Induced Cytidine Deaminase) promoter in activated B cells. Cre-recombinase is only active in mice receiving the ERT2-ligand tamoxifen or tamoxifen analogue 4-OHT. The activated Cre-recombinase cuts out the loxP-flanked stop cassettes of the yellow fluorescent reporter protein (eYFP), which leads to the constitutive expression of eYFP under the Rosa26 promoter. Consequently, tamoxifen or 4-OHT administration will lead to a stable eYFP expression in AID+ B cells and all their progeny, including terminally differentiated plasma cells (PCs). Local 4-OHT micro-injection into inductive compartments such as PPs allows for eYFP labeling of activated AID+ B cells in a locally confined and controlled manner. Consequently, all memory B cells and PCs in the small intestinal lamina propria (SI LP) originating from eYFP+ B cells induced in a specific PP will be distinguishable from the prevailing PC pool.

6.1 Systemic tamoxifen administration leads to stable eYFP expression in B cells Our aim was to use the AIDcre mouse system to elucidate the dynamics of B cell generation in distinct mucosal immune-inductive compartments as well as the homing kinetics of their progeny under homeostatic conditions in vivo. We assessed the efficiency and reliability of the system by employing systemic administration of tamoxifen, with subsequent analysis of different gut-inductive compartments. To ascertain the specificity of the system we examined in addition, negative control mice without systemic tamoxifen administration. As a first read out, we determined the frequency of eYFP expressing cells in Peyer’s patches (PPs), mesenteric lymph nodes (MLN) and spleen (SPL) at day 2, 3 and 5 after systemic tamoxifen application in direct comparison to non-tamoxifen treated mice by flow cytometry (Figure 33 A, B). As exemplified in Figure 33 B, eYFP expression of live, CD45+ cells of tamoxifen treated mice showed a distinct population, in contrast to non-treated mice. To discern whether eYFP expression is restricted to AID+ B cells and not expressed by other lymphoid or myeloid derived cells, we further characterized the phenotype of live, CD45+eYFP+ cells based on cell surface marker expression (Figure 33 B). Double positive CD45+eYFP+ cells were also positive for the characteristic early B cell markers CD19 and B220, indicating that exclusively B cells expressed eYFP. Importantly, we did not observe any eYFP expression among live, CD45+ 92 cells of non-tamoxifen treated AIDcre mice (Figure 33 B), which validates the activation of Cre- recombinase only upon tamoxifen administration. Based on the displayed gating strategy (Figure 33 B), we determined the frequency of eYFP+ B cells in different immune-inductive sites i.e. PPs, MLN and SPL at indicated time points. At all investigated time points we were able to identify eYFP+ B cells (Figure 33 C, D). Notably, the mean fluorescent intensity (MFI) of eYFP in PP, MLN and SPL B cells showed an increase over time (Figure 33 C). While we did not find a significant increase of eYFP+ B cell frequencies over time in any of the investigated organs, we however, found the highest frequency of eYFP+ B cells in PPs, and slightly lower frequencies of eYFP+ B cells in MLN compared to the overall low frequencies of eYFP+ B cells in SPL (Figure 33 D). These results are in line with the common view that PPs constitute major gut immune-inductive sites for B cell priming and intestinal PC generation65,93,111,136,236. However, the use of systemic tamoxifen administration does not allow to precisely assess and discriminate the contribution of single gut-inductive compartments to the generation of newly activated B cells and PCs. Because, eYFP+ B cells will de novo be generated in any compartment hosting AID-expressing B cells at the time of tamoxifen administration that potentially had homed to the organ under investigation.

Figure 33: Generation of eYFP+ B cells under steady state conditions. (A) AIDcre mice were intraperitoneal (ip) injected with tamoxifen or non-treated. PPs, MLN and SPL of mice were analyzed by flow cytometry for the presence of eYFP+ cells at indicated time points. (B) Representative gating strategy for eYFP+ B cells (gated as live, CD45+) from PPs at day 3 after systemic tamoxifen administration or of AIDcre control mice without tamoxifen administration. (C) Representative FACS plots of B cell eYFP expression in PPs, MLN and SPL at indicated time points of tamoxifen ip treated mice (+ Tam ip) or non-treated mice (Ø Tam ip). (D) Graphs show the mean percentage ± SD of eYFP+ B cells in PPs, MLN and SPL at indicated time points. In PPs symbols represent data from pooled PPs of individual mice, in MLN and SPL symbols represent individual mice. Groups were compared by Kruskal-Wallis test with post hoc Dunn’s multiple comparison test (ns – not significant). 94

6.2 Histological validation of eYFP expression in Peyer’s patches As an alternative method to confirm eYFP+ cells in PPs that we observed using flow cytometry analysis (FACS), we used confocal microscopy. However, the detection of eYFP by microscopy can be quite challenging due to the high autofluorescence inherent to gut tissue. Therefore, we used non-tamoxifen treated AIDcre mice to assess autofluorescence of small intestinal gut sections and PPs in order to distinguish true eYFP signal from autofluorescent background. Non-treated control mice showed no unspecific activation of Cre-ERT2 i.e. no eYFP detection in tissue sections (Figure 34 A), which was in line with obtained FACS data (Figure 33 B, C, D). Based on the sections from non-treated control mice, we adjusted laser and confocal-microscopic settings accordingly for the detection of eYFP signal of systemically tamoxifen treated mice (Figure 34 B, C). We first examined fixed whole mount preparations of single PPs (top view) and adjacent mucosal tissue (Figure 34 B). True eYFP signal (depicted as green signal) was clearly distinguishable from autofluorescent signal (depicted as yellow signal) i.e. equal fluorescence signal detection in all acquired channels. Moreover, in PPs detected eYFP signals were mostly found as organized cell clusters, indicative of B cell follicles (Figure 34 B). Next, we examined eYFP signals on cryo-sections of fixed PPs from systemically-tamoxifen treated AIDcre mice (Figure 34 C). Although fewer as compared to whole mount PP preparations, we again found eYFP+ cells in clusters (Figure 34 C). Notably, we used histology in addition to FACS analysis to merely verify the presence or absence of eYFP+ cells in PPs. Together, the results of FACS analysis in conjunction with confocal- microscopy confirmed the validity of the AIDcre transgenic mouse model enabling the sufficient and specific eYFP labeling of B cells.

Figure 34: Histological validation of eYFP expression in Peyer’s patches by confocal microscopy. Intracellular eYFP signal of activated B cells in PPs was confirmed by confocal microscopy. (A) PFA/sucrose fixed PP cryo- sections of non-treated AIDcre control mice. (B) Confocal microscopy of PFA/sucrose fixed whole mount PP preparations (top view) at day 3 after ip tamoxifen administration of AIDcre mice. (C) Tissue cryo-sections of PFA/sucrose fixed PPs of ip tamoxifen treated AIDcre mice. (B, C) White rectangles indicate areas which are displayed enlarged in the bottom row. (A, C) To distinguish real eYFP signal from autofluorescence, only green and red channels and merged channels are shown. Merged channels (left), green channel (middle) and red channel (right). PP structures in whole mount preparations or tissue sections are indicated and outlined by dashed lines, examples of YFP+ cells are indicated by white asterisks. Scale bars are 100 µm.

6.3 Frequencies of eYFP+ B cells increase over time in inductive gut compartments Our results thus far have demonstrated that the AIDcre transgenic mouse model allows for effective eYFP labeling of B cells in different mucosal immune-inductive compartments upon systemic tamoxifen administration. We next went on to assess the kinetics of eYFP+ B cell accumulation and thus the capacity of the AIDcre system. Therefore, mice were in addition to tamoxifen administration treated with FTY720. FTY720 is a functional antagonist of S1PR1, a prominent surface molecule enabling the egress of lymphocytes from secondary lymphoid organs (SLOs) such as PPs or MLNs298. The administration of FTY720 allows lymphocytes to enter SLOs from blood but effectively prevents their egress via efferent lymphatic vessels

96 leading to the trapping of lymphocytes in respective SLOs. In our experimental set-up mice received FTY720 in drinking water throughout the entire experimental time period and were in addition systemically treated with tamoxifen (Figure 35 A). This approach can thus point to dominant inductive sites of eYFP+ B cell generation. Furthermore, we could determine the potential incremental increase of eYFP+ B cells in different compartments over time. The frequencies of eYFP+ B cells were analyzed by flow cytometry at indicated time points in PPs, MLN, SPL and blood. Effective FTY720 administration was indicated by pronounced lymphopenia in blood (data not shown). FTY720 admission in combination with systemic tamoxifen administration resulted in a significant increase of eYFP+ B cell frequencies over time in PPs (Figure 35 B). Notably, we also observed a significant increase of eYFP+ B cells in MLN but overall, eYFP+ cell frequencies were 10-fold lower in MLN as compared to PPs (Figure 35 B, C). Conversely, the overall low frequencies of eYFP+ cells found in SPL indicate that SPL might not support eYFP+ B cell generation and proliferation (Figure 35 D). In agreement with blocked lymphocyte egress from SLOs and therefore abrogated re-circulation of lymphocytes, we consistently detected negligible frequencies of eYFP+ B cells in blood (Figure 35 E). As tamoxifen was administered systemically and lymphocyte egress and re- circulation blocked from SLOs, the increased frequencies of eYFP+ B cells in PPs and MLN must result from both local generation and proliferation in these compartments. While these results indicate that next to PPs, also the MLN acts as an immune-inductive compartment, PPs nevertheless appear to be the dominant inductive sites for B cell generation and proliferation in the gut. Therefore, in the following we will put our focus on PPs supporting the generation and differentiation of eYFP+ B cells.

Figure 35: Increase of eYFP+ B cell frequencies in PPs and MLN after FTY720 administration. (A) FTY720 was administered in drinking water throughout the entire experimental period in addition to tamoxifen ip injections of AIDcre mice were. (B) Frequency of eYFP+ B cells (gated as live, CD3-/CD4- , CD45+, CD19+) in PPs (symbols represent single PPs) obtained by flow cytometry. (C, D, E) Frequency of eYFP+ B cells (gated as live, CD3-/CD4-, CD45+, CD19+) in MLN (C) SPL (D) and blood (E) analyzed by flow cytometry at indicted time points (symbols represent individual animals from at least two experiments). Bars represent mean ± SD. All groups were compared by Kruskal-Wallis test (** p ≤ 0.01, *** p ≤ 0.001, ns – not significant).

6.4 Single 4-OHT Peyer’s patch injections lead to local eYFP labeling of B cells In accordance with other studies93,136,236,279, we identified PPs as major inductive compartments for B cell priming and proliferation in the intestinal mucosa. Additionally, we could show that eYFP+ B cells in PPs increase in frequency over time after systemic tamoxifen and FTY720 administration. In order to monitor the differentiation kinetics of newly activated B cells originating from a single gut immune-inductive compartment, we established the method of local 4-Hydroxytamoxifen (4-OHT) injections into single PPs (Figure 36 A, B and Materials and Methods Part II, Figure 8). Injection of 4-OHT into single PPs allows local labeling of all AID- expressing B cells by inducing constitutive expression of eYFP. This enables subsequent fate tracking of AID+ B cells and their progeny originating from one single PP. To our knowledge, this is the first method enabling spatial-temporal fate tracking of B cells and their progeny originating from a single inductive compartment. However, a prerequisite before using this B cell fate-tracking model was to assess the potential “leakiness” of local 4-OHT micro-injections. We used FTY720 administration to determine whether 4-OHT would be locally confined to single injected PPs, or if active or passive diffusion of 4-OHT into surrounding tissue or blood would lead to the generation of eYFP+ B cells within other compartments but injected PPs. Thus, eYFP+ cells found in any other compartments but injected PPs must have been generated elsewhere. In this set-up, mice received FTY720 prior to 4-OHT PP micro-injections

98 and were maintained on FTY720 throughout the experimental time period (Figure 36 A, B). At day three after 4-OHT PP injections, single PPs, MLN, SPL and blood were analyzed by flow cytometry. We chose to include systemically tamoxifen treated litter-mate control mice for the purpose of gaining a relative read-out as comparison to results from 4-OHT single PP injected mice (Figure 36 C, D). In contrast to systemically treated mice, eYFP+ B cells after 4-OHT PP injections were mainly confined to that single injected PP, albeit at lower frequencies than after systemic tamoxifen treatment (Figure 36 C, E, F). Interestingly, we observed variation in eYFP+ B cell frequencies between individual PPs of systemically-treated mice (Figure 36 C, E), potentially indicating differences in immune-activation status between PPs. Similarly, we observed variability of eYFP+ B cell frequencies between 4-OHT injected PPs among individual mice (Figure 36 C, F). Our data demonstrate that local micro-injections lead to the efficient generation of eYFP+ B cells in single injected PPs, despite differences of eYFP+ B cell frequencies between PPs. Conversely to systemically tamoxifen treated mice, in 4-OHT single PP injected mice we observed little to no eYFP+ B cells in non-injected PPs, and negligible frequencies of eYFP+ B cells in MLN or SPL and blood (Figure 36 D, E, F). These results indicate that 4-OHT after single PP injections stays largely confined to the tissue of injected PPs, and thus results in the local generation of eYFP+ B cells.

100

Figure 36: Single 4-OHT Peyer’s patch injections locally eYFP label B cells. (A) Experimental schematic of single 4-OHT PP injections and systemically tamoxifen treated control mice. AIDcre mice were treated with FTY720, single PPs injected with 4-OHT or mice ip treated with tamoxifen and PPs, MLN, SPL and blood analyzed after 3 days. (B) Surgical procedure of single PP injections. The small intestine was removed from the abdomen and single PPs micro-injected with 4-OHT using a fine glass capillary. Single PPs are indicated by arrows (magnification is indicated by dotted line). (C) Representative flow cytometry plots of eYFP+ B cells (gated as live, CD3-/4-, CD45+, B220+, CD19+) in single PPs of AIDcre mice either ip injected with tamoxifen (upper panel) or of 4-OHT single PP injected mice (lower panel, injected PPs are marked in blue and dashed outline). (D) Representative flow cytometry plots of eYFP+ B cells (gated as live, CD3-/4-, CD45+, B220+, CD19+) in MLN, SPL and blood of tamoxifen ip treated mice (upper panels) or of 4-OHT single PP injected mice (lower panels). (E, F) Summary plots of eYFP+ B cell frequencies in PPs, MLN, SPL and blood of FTY720 treated AIDcre mice either ip injected with tamoxifen (E) or of single 4-OHT PP injected mice (F). (E, F). Bars show mean ± SD. (E) For PPs symbols represent individual PPs of four individual mice. Symbols in MLN and SPL represent individual mice. Data shown are pooled from four independent experiments. (F) Symbols represent individual PPs of 18 individual mice. Symbols shown for MLN and SPL represent data of individual mice. Data shown are pooled from 18 individual mice from six experiments.

101

6.5 Examination of 4-OHT dispersion after single Peyer’s patch injections To assess potential dissemination of 4-OHT after PP injections more accurately, we employed an additional transgenic mouse model (Figure 37). This double transgenic mouse model combines the conditional labeling of activated AID+ B cells in immune-inductive compartments as well as the conditional labeling of gut epithelial cells (IECs). The mouse was created by crossing the AIDcreERTxRosa26loxPeYFP (AIDcre) mice with VillincreERT2 mice. In the resultant mouse (hereafter referred to as AIDcreVillcre), eYFP expression is driven by both the AID promoter in activated B cells and the Villin promoter in gut epithelial cells. Administration of CreERT2-ligand tamoxifen or 4-OHT will thus result in eYFP labeled B cells as well as eYFP labeled gut epithelial cells (Figure 37). Consequently, this mouse would allow to precisely determine the “leakiness” of 4-OHT after single PP micro-injections.

Figure 37: Double transgenic mouse model enabling the eYFP marking of AID+ B cells and Villin+ epithelial cells. In this double transgenic mouse Cre-recombinase is expressed under the control of both AID-promoter and Villin-promoter. AID is expressed in activated B cells, while Villin is expressed in gut epithelial cells. In both cases the Cre-recombinase is expressed as a Cre-ERT2 fusion product 102 and can only be activated after tamoxifen or tamoxifen analogue 4-OHT administration. Tamoxifen administration will activate Cre-recombinase that in turn cuts out the loxP-flanked stop cassettes of a yellow fluorescent reporter protein (eYFP), which is expressed under the control of the Rosa26 promoter. PP injection with 4-OHT irreversibly labels local AID+ B cells, but also shows the extent of 4- OHT “leakiness” into the surrounding tissue. Leakiness of 4-OHT can be assessed by eYFP labeling of gut epithelial cells.

We first determined whether this double transgenic mouse model would enable the efficient labeling of both AID+ B cells and Villin+ gut epithelial cells. Therefore, we treated AIDcreVillcre mice systemically with tamoxifen and analyzed pooled PPs and gut epithelial cells at day three after tamoxifen administration (Figure 38 A). We found that both B cells isolated from PPs (gated as shown) and gut epithelial cells (gated as shown) isolated from the same mouse showed eYFP expression (Figure 38 B, C). Notably, the frequencies of eYFP+ B cells in PPs were similar to what we have observed using single transgenic AIDcre mice, indicating similar AID-driven Cre-activity in both mouse models.

Figure 38: eYFP expression in B cells and gut epithelial cells after systemic tamoxifen administration. (A) AIDcreVillcre mice were injected ip with tamoxifen and eYFP expression of B cells isolated from PPs and gut epithelial cells isolated from whole SI were analyzed by flow cytometry. (B, C) Representative gating strategy for eYFP+ B cells (gated as live, CD3-/4-, CD45+, B220+, CD19+, eYFP+) isolated from PPs (B) or epithelial cells (gated as live, CD45-, EpCam+, eYFP+) isolated from whole SI (C). 103

In order to rule out potential unspecific Cre-recombinase induced eYFP expression in both compartments, we examined AIDcreVillcre mice that were either systemically treated with tamoxifen, non-treated AIDcreVillcre mice or tamoxifen treated WT mice. Here, we included WT mice to assess whether tamoxifen itself would have any unintended side effects and to rule out that any fluorescence observed in non-treated AIDcreVillcre mice is due to background fluorescent and not induced by spontaneous Cre-activation. Mice were analyzed both by flow cytometry and histology at day three post tamoxifen administration (Figure 39). Flow cytometry analysis of systemically tamoxifen treated AIDcreVillcre mice showed clear eYFP+ cell populations in both compartments, PPs and gut epithelium, in contrast to non-treated AIDcreVillcre mice or of tamoxifen treated WT mice (Figure 39 B). In addition to flow cytometry analysis, eYFP expression in PPs and gut epithelium was determined by confocal microscopy of fixed whole mount tissue preparations and cryo-sections (Figure 39 C - G). On fixed whole mount preparations (PP top view) of tamoxifen-treated mice, we detected eYFP+ cells both in PPs as well as within the crypt-epithelial layer as compared to non-treated AIDcreVillcre mice. (Figure 39 C, D). Similarly, we observed distinct eYFP signal in frozen sections in both SI epithelial cells and as organized clusters in PPs (Figure 39 E). In contrast, no eYFP+ cells were detected in PPs or SI epithelial cells of either non-treated AIDcreVillcre (Figure 39 F) or tamoxifen treated WT mice (Figure 39 G). Altogether, these results demonstrate that our double transgenic “AIDcreVillcre” (AIDcreVillcreRosaLoxPeYFP) mouse can be efficiently induced to express eYFP in both activated B cells and gut SI epithelial cells that can be readily detected by flow cytometry and confocal microscopy. Notably, we did not observe any activation of Cre-recombinase without ERT2-ligand (Tamoxifen) application.

104

105

106

Figure 39: Lack of unspecific eYFP expression in AIDcreVillcre mice in the absence of tamoxifen. (A, B) To assess unspecific Cre-recombinase activation and eYFP expression in B cells and gut epithelial cells, AIDcreVillcre mice and WT mice were ip injected with tamoxifen or control AIDcreVillcre mice were left untreated. Mice were analyzed by flow cytometry for eYFP expression in PPs and SI epithelial cells three days after tamoxifen administration. (C-G) Confocal images of PFA/sucrose fixed whole mount PP preparations or PFA/sucrose fixed cryo-sections of AIDcreVillcre and WT mice ip injected with tamoxifen or non-treated AIDcreVillcre control mice. (C, D) Fixed whole mount PP preparations (top view) of AIDcreVillcre mice ip injected with tamoxifen (C) or untreated AIDcreVillcre mice (D). PP structures are outlined by dashed lines and eYFP signal is indicated by an asterisks in PPs and an arrow in LP. Merged channels (left), green channel only (middle), and red channel only (right). Scale bars are 100 µm. (E, F, G) Confocal images of PFA/sucrose fixed PP cryo-sections of AIDcreVillcre mice ip injected with tamoxifen (E), non-treated (F) or WT mice ip injected with tamoxifen (G). PPs in images are outlined by dashed lines and magnifications are indicated by white box and/or arrows. Merged channels (left), green channel only (middle), red channel only (right). (E) Scale bars are 200 µm, 100 µm and 50 µm. (F) Scale bars are 200 µm (upper panels) and 100 µm (lower panels). (G) Scale bars are 100 µm.

107

6.6 Spatial distribution of Cre-activity throughout the small intestine To investigate the overall activity of Cre-recombinase throughout the gut epithelium after systemic tamoxifen administration, we sectioned the SI into approximately 1 cm pieces for analysis of eYFP expression throughout the SI. We included non-treated AIDcreVillcre control mice to rule out unspecific Cre-activation. For analysis single PPs and corresponding/adjacent SI pieces were analyzed by flow cytometry (Figure 40 A). In line with our data obtained from AIDcre mice we also observed similar variation in eYFP+ B cell frequencies between individual PPs in AIDcreVillcre mice (Figure 40 B, C). Likewise, we also observed minor variations of eYFP expression in epithelial cells between single SI gut pieces (Figure 40 B, C). Non-treated AIDcreVillcre control mice did not show eYFP expression in either compartment (PPs, SI) (Figure 40 D, E). These results verify the specific induction of Cre-recombinase in both compartments, but also demonstrate that Cre-recombinase acts uniformly throughout the SI gut epithelium.

108

Figure 40: Spatial distribution of eYFP expression in Peyer’s patches and small intestinal epithelial cells in tamoxifen treated AIDcreVillcre mice. (A) Mice were treated or non-treated with tamoxifen ip and the frequency of eYFP+ B cells in single PPs or eYFP+ epithelial cells in single SI gut pieces determined by flow cytometry. (B) Representative flow cytometry plots of eYFP+ B cells in single PPs (upper panel) and eYFP expression in epithelial cells isolated from single SI pieces (lower panel) of tamoxifen treated mice. (C) Graphs show frequency of eYFP+ B cells (left axis) and eYFP+ SI epithelial cells (right axis) of tamoxifen treated mice. Lines show mean ± SD of three mice. Numbers on the X-axis indicate individual PPs and corresponding SI sections. (#1 distal, #7 proximal). (D) Summary of eYFP expression in epithelial SI cells of tamoxifen treated or non-treated mice. Symbols represent single SI gut pieces of four mice. Groups were compared by unpaired Mann-Whitney test (**** p ≤ 0.0001). (E) Representative flow cytometry plots of single PPs (upper panel) and SI epithelial cells (lower panel) of non-treated AIDcreVillcre control mice.

6.7 4-OHT after single Peyer’s patch injections is confined to the injection site Based on the consistent Cre-activity throughout gut intestinal epithelial cells (IECs) in AIDcreVillcre mice, we could now investigate whether 4-OHT after single PP micro-injections would diffuse into surrounding tissue. We micro-injected both multiple or one single PP with 4- OHT and analyzed single PPs and respective PP surrounding SI pieces (Figure 41 A). 4-OHT injections of multiple PPs resulted in efficient eYFP labeling of B cells in the respective PPs, whereas we also observed low frequencies of eYFP+ B cells in non-injected PPs located in the immediate vicinity (Figure 41 B). Similarly, eYFP+ epithelial cells were detected in the SI surrounding the immediate vicinity of 4-OHT injected PPs, but also at low frequencies in distal 109

SI parts (Figure 41 B, E) potentially due to minimal diffusion of 4-OHT after multiple PP injections. However, 4-OHT injection of one single PP resulted on the hand in marked frequencies of eYFP+ B cells in the injected PP as compared to non-injected PPs, but in contrast to multiple PP injections showed only minimal expression of eYFP in IECs immediately surrounding the injected PP. Luminal dissemination of 4-OHT was therefore negligible indicated by a 10-fold lower eYFP expression in IECs surrounding non-injected PPs (Figure 41 C, D). Importantly, the overall eYFP+ cell frequencies of both IECs immediately surrounding injected PPs and distal IECs were significantly lower after single PP injections as compared to multiple PP injections (Figure 41 D, E) or systemically tamoxifen treated controls (Figure 40 B, C, D). This indicates minimal Cre-activation in IECs immediate to both injected PPs and non- injected PPs after single PP 4-OHT injection (Figure 41 E).

110

111

Figure 41: Single Peyer’s patch injections do not lead to 4-OHT dissemination. (A, B) Either multiple (B) or one single PP (C) of AIDcreVillcre mice were injected with 4-OHT and analyzed after 3 days. (B, C) Representative FACS plots show the frequency of eYFP+ B cells in PPs (upper panel) and the frequency of eYFP+ epithelial cells (IECs) isolated from surrounding SI tissue pieces (lower panel). Injected PPs are marked in blue and outlined by dashed frames and corresponding SI tissue pieces are marked in orange and are indicated by dashed frames. (D) Frequency of eYFP+ B cells isolated from single PPs (dashed lines) and eYFP+ gut IECs (solid lines) of single 4-OHT PP injected AIDcreVillcre mice. Arrows indicate injected PPs of individual mice. Lines in the same color represent PPs (dashed lines) and epithelial cells (solid lines) of individual mice. Numbers on the X-axis indicate individual PPs and corresponding SI sections (#1 distal, #8 proximal). (E) Bar graph shows the frequency of eYFP+ IECs isolated from single SI gut pieces of multiple PP injected mice or from single PP injected mice and corresponding IECs of non-injected PPs. Symbols represent data from individual gut SI pieces of individual mice. Groups were compared by Kruskal-Wallis test with post hoc Dunn’s multiple comparison test (* p ≤ 0.05, ns-not significant).

To validate the results obtained by flow cytometry, we additionally examined the potential “leakiness” of 4-OHT after single PP micro-injections by confocal microscopy. Single PPs of AIDcreVillcre mice injected with 4-OHT and adjacent non-injected PPs were prepared for histology and eYFP signals in PPs and SI gut epithelium determined on cryo-sections of fixed tissue samples (Figure 42 A, B). In contrast to non-injected PPs, we readily found eYFP+ cells in injected PPs, but only sparse eYFP+ IECs in the immediate vicinity, which is in accordance with our flow cytometry data demonstrating minor, if any, luminal dissemination of 4-OHT into immediate gut tissues (Figure 41 and Figure 42). Collectively, these results demonstrate that single PP micro-injections efficiently eYFP label B cells within a single PP and that 4-OHT is locally confined after single PP injections.

112

Figure 42: 4-OHT after single Peyer’s patch injections is locally confined. (A, B) Confocal images of 4-OHT micro-injected PP (A) or non-injected PP of AIDcreVillcre mice (B). (A) Confocal images of PFA/sucrose fixed cryo-sections of a single 4-OHT injected PP. PP structures in images are outlined by dashed lines and magnification is indicated by white box and arrows. eYFP signal in PPs is indicted by asterisks and in epithelium by arrows. (B) PFA/sucrose fixed cryo-section of a non-injected PP. PP structure is outlined by dashed line. (A, B) Merged channels (left), green channel only (middle), red channel only (right). Scale bars are 100 µm (A) and 200 µm (B).

113

6.8 eYFP+ B cells activated in a single PP home to the SI LP in steady state The afore mentioned control experiments demonstrated that our methodological approach of local 4-OHT micro-injections efficiently eYFP label AID+ B cells in a single PP without luminal dispersion of 4-OHT. Based on these results, we next wanted to examine the contribution of newly differentiated plasma cells (PCs) originating from a single PP to the PC pool in the SI LP in a long-term experimental set-up. To probe our system of 4-OHT micro-injections to monitor the fate and migration kinetics of B cells activated in a single PP under steady state conditions, we established a surface marker panel to discriminate between activated B cells and differentiated plasma blasts/plasma cells.

Phenotypic characterization of eYFP+ cells in PPs and SI LP We phenotypically characterized eYFP+ cells in PPs and SI LP after both systemic tamoxifen treatment and single 4-OHT PP injection (Figure 43 A) using surface staining for CD3, CD19, B220, GL7, CD138 and IgA (Figure 43 B). After both treatments, live, CD45+eYFP+ cells detected in PPs were as expected negative for CD3/CD4 and the vast majority of PP eYFP+ cells were double positive for CD19 and B220, with minor single B220+ or CD19+ cell populations (Figure 43 B). Moreover, for both treatments the majority of PP eYFP+ cells were positive for GL7, a marker characteristic of GC B cells. While, systemic administration of tamoxifen resulted in marked frequencies of IgA+CD138- and low frequencies of IgA-CD138+ or IgA+CD138+ PP eYFP+ cells (Figure 43 B, upper panel), PP eYFP+ cells after single 4- OHT PP injections were largely negative for either CD138 or IgA (Figure 43 B, lower panel). These results indicate, that for both treatments, the vast majority of eYFP+ cells found in PPs acquired the typical phenotype of activated B cells likely engaged in GC reactions, but also contained minor populations showing a more mature differentiation status. To compare the differentiation and activation status of eYFP+CD19+ cells with eYFP-CD19+ B cells, we examined the expression of GL7, IgA and CD138 in respective PPs (Figure 43 C). After systemic tamoxifen administration we found that the majority of eYFP-CD19+ cells were negative or low for GL7, and negative for IgA and CD138 as compared to eYFP+CD19+ B cells that were prominently GL7+, heterogeneous for IgA and low for CD138 expression. Similarly, after 4-OHT PP injection, the vast majority of eYFP-CD19+ were negative for GL7, IgA and CD138 as compared to eYFP+CD19+ (Figure 43 C). Thus, eYFP expression reflects an overall frequency of activated AID+ B cells likely engaged in GC reactions, albeit lower after single PP injections. We next interrogated whether locally generated PP eYFP+ B cells would further differentiate enabling to home to the SI LP. Indeed, we detected eYFP+ cells in the SI LP after systemic tamoxifen treatment and amazingly also after 4-OHT injection of one single PP (Figure 43 D). However, eYFP+ cell frequencies in the SI LP after systemic tamoxifen administration were notably higher as compared to SI LP eYFP+ cell frequencies observed after single 4-OHT PP injections (Figure 43 D). To investigate if eYFP+ cells found in the SI LP reflect a plasma cell phenotype, we compared the phenotype of eYFP+ B cells in PPs and eYFP+ cells found in the corresponding SI LP after both treatments (Figure 43 D). In contrast to the activated B cell phenotype of eYFP+ cells in PPs (Figure 43 B, C, D), live, CD45+eYFP+ cells in SI LP showed a plasma blast/plasma cell phenotype indicated by the marked upregulation of IgA and CD138, downregulation of CD19 and the absence of GL7 expression. 114

Notably, the plasma cell phenotype of eYFP+ cells in the SI LP was found after both systemic tamoxifen administration and 4-OHT PP injections (Figure 43 D), demonstrating that after single 4-OHT injections eYFP+ B cell differentiation kinetics are similar to those after systemic tamoxifen application. These data demonstrate that single 4-OHT PP injections are sufficient to locally generate eYFP+ B cells that have the capacity to transition into plasma blasts/plasma cells able to home to the SI LP in steady state.

115

116

Figure 43: Phenotypic characterization of eYFP+ cells in Peyer’ patches and in the small intestinal lamina propria. AIDcre mice were either tamoxifen ip treated or a single Peyer’s patch (PP) micro- injected with 4-OHT. (A) Representative flow cytometry plots of eYFP+ B cells in individual PPs (gated as live, CD3-/4-, CD45+, B220+, CD19+) of ip tamoxifen treated mice (upper panel) or of single 4-OHT PP injected mice (lower panel) analyzed at day 21. Single 4-OHT injected PP is indicated in blue and dashed frame and eYFP+ cells outlined in orange. (B) Characterization of eYFP+ cells in PPs (gated as live, CD45+) of ip tamoxifen treated mice (upper panel) or 4-OHT single PP injected mice (lower panel). (C) Representative characterization of eYFP- and eYFP+ (gates as live, CD3-/4-, CD45+, B220+, CD19+) cells in PPs after tamoxifen ip treatment (upper panel) or single 4-OHT PP injection (lower panel) analyzed at day 21. (D) Representative characterization of eYFP+ B cells in PPs (gated as live, CD3-/CD4-, CD45+, B220+, CD19+) and eYFP+ cells in the small intestinal lamina propria (SI LP) (gated as live, CD45+) of ip tamoxifen treated mice (upper panel) or 4-OHT single PP injected mice (lower panel).

117

Differentiation and migration kinetics of eYFP+ B cells in steady state Although, we have demonstrated the applicability of single PP 4-OHT micro-injections to locally generate eYFP+ B cells, we yet included systemically tamoxifen treated control mice in this experimental set-up (Figure 44 A). To our knowledge, spatial-temporal fate tracking data obtained after single PP 4-OHT injections have not been reported before in any long-term experimental context. Including systemic controls would provide some sort of “reference” data to compare B cell migration patterns after both treatments. We therefore in parallel micro- injected single PPs or systemically treated AIDcre mice with tamoxifen and analyzed eYFP+ cell frequencies in PPs, MLN, SPL and SI LP after 7, 14, 21 and 90 days (Figure 44 A). At indicated time points eYFP+ cells were analyzed according to the representative gating depicted in Figure 43 B. After systemic tamoxifen administration, we found eYFP+ B cells in all PPs analyzed at all investigated time-points. (Figure 44 B). As previously observed for shorter time periods, we found variability in eYFP+ B cell frequency between PPs up to 21 days (Figure 44 B). Similarly, eYFP+ B cells were readily detectable in injected PPs (indicated in orange) at all time-points examined, surprisingly up to 90 days after injection (Figure 44 C). Notably, we also found eYFP+ B cells in adjacent and distal non-injected PPs that was most pronounced at day 14 after 4-OHT PP injection, suggesting the re-circulation of eYFP+ B cells between GCs of different PPs and might reflect the time course of B cell activation and GC reactions in vivo (Figure 44 B). Overall, we found the peak frequency of eYFP+ B cells in PPs at day 14 after both systemic treatment and single PP injections and thereafter decreasing frequencies of eYFP+ B cells (Figure 44 D). In addition, eYFP+ B cell frequencies after systemic tamoxifen administration were consistently higher at all time-points as compared to 4-OHT PP injections (Figure 44 D). Based on our data (Figure 33), we speculated that activated B cells may disseminate from PPs to other secondary lymphoid compartments most likely to the MLN until finally homing to the SI LP. To further explore migration patterns into other immune-inductive compartments, we examined the eYFP+ B cell frequencies in MLN and SPL over time. Despite a small trend towards an increase of eYFP+ B cell frequency in MLN at day 21 after systemic tamoxifen administration no overall significant increase of eYFP+ B cell frequencies over time was observed. A similar picture was observed in SPL (Figure 44 F). Overall, eYFP+ B cell frequencies were consistently at least 10-fold lower in MLN and SPL as compared to PPs (Figure 44 D, E, F, upper panels). Although, the results from systemically tamoxifen treated mice suggest that PPs constitute the primary inductive compartment of B cells, one cannot deduce migration kinetics of B cells from these data, because systemic administration of tamoxifen also results in eYFP+ B cell generation in MLN and SPL, irrespectively of the overall lower extent. However, making use of our established method to locally generate eYFP+ B cells at a given time point in one distinct PP, would circumvent the limitations to interpret B cell migration kinetics after systemic tamoxifen administration, but instead enable to monitor B cell migrations patterns coherently. Albeit, eYFP+ B cell frequencies did not significantly increase towards a given time point in PPs, we could yet observe a marked increase of eYFP+ B cell frequency in PPs at day 14 after 4-OHT PP injections. Notably, similar to the results obtained from systemically tamoxifen treated mice, we observed markedly lower frequencies of eYFP+ B cell in MLN and SPL at all investigated time-points as compared to PPs (Figure 44 D, E, F, 118 lower panels). These results suggest that activated B cells generated in a single PP may circulate through other secondary lymphoid organs such as MLNs, but our data indicate a relatively small contribution of MLNs supporting further B cell proliferation and differentiation. Furthermore, the low frequencies of eYFP+ B cells found in the SPL largely exclude any necessity of activated B cells to home to the SPL for further differentiation before homing to the SI LP. Next, we examined to what extent and at which time-points eYFP+ B cells after induction in PPs would be detectable in the SI LP, excluding the unlikely possibility of bona fide induction of eYFP+ B cells in the SI LP. Thus, we analyzed eYFP+ plasma cell frequencies in the SI LP of tamoxifen treated or 4-OHT PP injected mice over time (Figure 44 G). We observed the highest frequencies of SI LP eYFP+ plasma cells at day 14 and day 21 after both treatments, in accordance with highest frequencies of eYFP+ B cells observed in PPs at these time-points, albeit overall lower eYFP+ cell frequencies in PPs and the SI LP after single 4- OHT PP injections (Figure 44 D, G). Notably, eYFP+ plasma cells were still detectable 90 days after single PP injections (Figure 44 G, lower panel). Together, these results suggest that PPs constitute the primary inductive compartments for intestinal B cells differentiation and proliferation. Moreover, our data from single PP injections further suggest particular migration dynamics of activated B cells through other PPs before homing to the SI LP, with only transient homing to the MLN. Furthermore, when analyzing SI LP CD45+eYFP+ cells, we noticed in some samples, a marked population of eYFP+ lymphocytes that showed the typical phenotype of early activated B cells (Figure 44 H, orange arrows). A common source of B cell “contamination” that are usually absent from the SI LP may stem from incomplete removal of PPs or from tertiary lymphoid structures such as isolated lymphoid follicles (ILFs). Using typical cell surface markers, we were able to dissect “contaminations” by B cells in SI LP preparations (Figure 44 H). Notably, PP or ILF derived B cells showed lower mean fluorescent intensity (MFI) eYFP expression and were typically positive for CD19, B220 and GL7 as compared to eYFP+ plasma blasts/plasma cells (Figure 44 H). Thus, being aware of possible contaminants, we next investigated whether newly arriving plasma blasts/plasma cells to the SI LP would show distinct changes in their phenotypical properties when analyzed over time (Figure 34 I, J). We assessed the expression of CD19, B220, GL7, CD138 and IgA on eYFP+ cells from PPs and corresponding SI LP at day 7, 14, 21 and 90 after induction, according to the staining shown in Figure 43 B. We first focused our phenotypical evaluation on data from systemically tamoxifen treated mice (Figure 44 I). The detection of CD19+B220+ cells among SI LP eYFP+ plasma cells at early time-points (day 7) may indicate a less mature state of plasma blast/plasma cells shortly after entry into the SI LP, whereas the detection of GL7+ cells most likely indicates B cell “contamination” from PPs or ILFs. We observed a gradual decrease of CD19 and B220 expression after day 7 coinciding with the upregulation of IgA and CD138 on SI LP eYFP+ plasma cells over time (Figure 44 I). Although, SI LP eYFP+ plasma cells showed slightly decreased levels of IgA and CD138 by day 90, these results nevertheless demonstrate that a substantial proportion of newly generated B cells transitioned into plasma blasts/plasma cells and stably added to the prevailing pool of SI LP plasma cells. However, some eYFP+ plasma cells may eventually become apoptotic, suggested by overall decreased eYFP+ plasma cell frequencies at day 90 (Figure 44 G, I). In PPs, the vast majority of eYFP+ B cells were consistently CD19+B220+ up to day 90 after induction, but frequencies of GL7+eYFP+ 119 cells showed a slight decrease by day 90. Moreover, in PPs, eYFP+ cells showed low surface expression of IgA and CD138 throughout, however, with slightly increased frequencies at day 21 and curiously at day 90 (Figure 44 I). Next, we investigated the phenotype of PP and SI LP eYFP+ cells over time after 4-OHT PP injections. Similarely to the data observed for systemic tamoxifen administration, after PP injections, eYFP+ cells in PPs were consistently positive for CD19 and B220. Conversely, SI LP eYFP+ cells showed heterogeneous expression of CD19 and B220 that was most pronounced at day 7, but progressively decreased thereafter. In addition, SI LP eYFP+ cells showed a progressive expression of the typical gut plasma cells markers IgA and CD138 that was however less pronounced by day 90, in line with systemic tamoxifen treated mice (Figure 44 J). These data show that single 4-OHT PP injection leads to the generation of eYFP+ plasma cells. When analyzed over time, eYFP+ plasma cells showed similar phenotypes as compared to those after systemic tamoxifen treatment (Figure 44 J). Therefore, these results clearly demonstrate that eYFP labeled B cells originating from a single PP have the capacity to differentiate into plasma blasts/plasma cells with subsequent homing to the SI LP and can persist in the SI LP up to 90 days after induction. Regardless of the limitations posed by our model considering the relatively low frequencies of eYFP+ cells in PPs and SI LP after single PP 4-OHT injection, we nevertheless were able to demonstrate the applicability to monitor the generation and migration properties of locally induced B cells under homeostatic conditions. B cell fate-tracking after local PP injections moreover may help to circumvent caveats posed by data acquired after systemic tamoxifen administrations, such as ruling out the potential contribution of eYFP+ cells generated in other inductive compartments to the organ under investigation. We suggest for future work, to use in addition to single PP injections, a model of increased B cell induction enabling a more feasible analysis of phenotypical and genetic properties of B cells and their progeny originating from a single inductive compartment.

120

121

122

123

Figure 44: Plasma cells originating from a single Peyer’s patch persist up to 90 days in the small intestine under homeostatic conditions. (A) AIDcre mice were either tamoxifen ip treated or single PPs were micro-injected with 4-OHT and analyzed at indicated time points. (B, C) Heat maps showing the frequency of eYFP+ B cells in single PPs of ip tamoxifen treated mice (B) or 4-OHT single PP injected mice (C) (stomach to caecum orientation) at indicated time points. Individual mice are indicated by # and frequencies are displayed as grey scale. Individually injected PPs are outlined in orange. (D - G) Frequency of eYFP+ B cells in PPs (D), MLN (E), and SPL (F) or eYFP+ plasma cells in SI LP (G) of tamoxifen ip treated mice (+ Tam ip) or of single 4-OHT PP injected mice (4-OHT PP inj). Groups were compared by Kruskal-Wallis test (**** p ≤ 0.0001, ns – not significant). (H) Representative assessment of eYFP+ B cells from PP or ILF contamination (gated as live, CD45+) in SI LP of ip tamoxifen treated mice analyzed at indicated time points. (I) Phenotypic characterization of eYFP+ cells isolated from PPs and corresponding SI LP from ip tamoxifen treated mice analyzed at indicated time points. Bars represent mean ± SD. White bars show data of eYFP+ cells isolated from PPs and grey bars show data of eYFP+ cells isolated from SI LP. (J) Representative surface marker expression (CD19, B220, GL7, IgA, CD138) of eYFP+ cells (gated as live, CD45+) isolated from single PPs and corresponding SI LP of either tamoxifen ip injected mice (+ Tam ip) or 4-OHT single PP injected mice (4-OHT PP inj) analyzed at indicated time points. eYFP+ cells from the same mouse isolated from PPs are displayed in orange and eYFP+ cells isolated from SI LP are displayed in green.

124

7 Discussion 7.1 General discussion and main findings Intestinal SIgA-microbiota interactions became a central interest for different scientific fields over the last years. However, the mechanisms underlying homeostatic SIgA-microbiota interactions are still not fully defined. Additionally, the role of SIgA in inflammatory context is still debated. Our data describe a new facet of SIgA-microbiota interactions in the adult human healthy and inflamed gut. We show that a high frequency of intestinal IgA and IgG antibodies cloned from single intestinal plasma cells bind to a substantial proportion of fecal gut bacteria. Notably, high microbiota-binding capacity was associated with binding to phylogenetically distant gut bacteria. In the adult human gut, specific cross-species reactive binding to phylogenetically diverse gut bacteria relied on the accumulation of somatic mutations and was unrelated to polyreactivity. Furthermore, we found that cross-species reactivity is a prevalent feature of both HD and CD derived intestinal antibodies. However, microbial antigens associated with the initial activation and subsequent affinity maturation of cross-species reactive antibodies are still elusive. We propose that either similar bacterial strains present in different hosts or structurally similar epitopes shared between the microbiota might provide a mechanistic explanation for the generation of cross-species reactivity. Our data demonstrate that somatic mutations crucially contribute to antibody cross-species reactivity, but mechanistically it is not clear how somatic mutations shape the generation of cross-species reactive IgA antibodies. In an attempt to understand the principles of microbiota-targeting, cross-species reactive IgA, we used an in vivo transgenic mouse model. In line with other studies100,105,236,268, we suggest that re-circulation of activated- or memory B cells between different PP GCs accounts for affinity-matured, cross-species reactive IgA. Moreover, we found that constantly generated new waves of IgA+ PCs under homeostatic conditions without neo-antigen challenge contribute to the prevailing PC pool in the mouse gut. While cross- species reactivity likely contributes to regulating the gut microbiota, IgA-microbiota targeting never extended to all bacteria of an enriched species. This “partial” binding suggests genetic heterogeneity or different growth-states within commensal species. We therefore suggest that in addition to antibody-specificity, other host and environmental factors need to be considered when investigating antibody-microbiota interactions.

7.2 Cross-species reactive antibodies – defining the term The phenomenon of intestinal IgA to bind multiple microbial species has recently been described in in both mice and humans20,168,305,306. Trying to find a consistent terminology, it was recently suggested to introduce the term “cross-species reactivity” to describe this mode of IgA-microbiota binding91. Cross-species reactivity seems to occur in different species but does not make any assumptions about the mode of binding or mechanistic inferences. Thus, cross- species reactivity of IgA merely describes the capacity of single monoclonal intestinal antibodies to bind multiple but defined phylogenetically-distant bacteria.

125

7.3 Different binding modes lead to cross-species reactivity of intestinal antibodies Several modes of binding may lead to cross-species reactivity. Both canonical and non- canonical binding can lead to microbiota cross-species reactivity of SIgA. Here, canonical interactions are mediated by Fab-dependent binding of SIgA to cognate antigen. This mode of binding relies on specific genetic configurations of the complementary determining (CDR) regions and frame work (FR) regions. In particular, the CDR3 regions seem to determine antigen specificity of canonical interactions209,269. The CDR regions of antibodies are “hot- spots” for antigen and T cell dependent somatic hypermutation (SHM)209,269,307, resulting in high-affinity antibody responses. The accumulation of somatic mutations (SM) is a characteristic indicator of T cell dependent affinity-matured antibody responses136,208-210,274. While classical T cell dependent responses lead to increased specificity and affinity to the initial cognate antigen, we found in our set of high-microbiota reactive mAbs that the accumulation of somatic mutations leads to the “broadening” of bacterial specificity, enabling cross-species- reactivity. Our data thus demonstrate a novel facet of SIgA-microbiota interactions. We suggest that affinity-matured intestinal IgA antibodies show broad but specific cross-species reactivity to common but also select members of the microbiota. Canonical Fab-mediated interactions also include polyreactive binding. Polyreactivity describes the capacity of antibodies to bind a broad range of structurally unrelated antigens such as LPS, CpG, insulin, flagellin, and KLH, commonly used to assay polyreactivity213,308,309. Typically, polyreactive antibodies bind with low affinity to antigens and are germ-line or close to germ-line encoded20,213,216,308. The induction of polyreactive antibodies usually does not rely on T cell help and is not associated with accumulated somatic mutations20,174,205,216. Thus, polyreactive antibodies usually carry no or only few somatic mutations and are believed to constitute the “natural pool” of antibodies as they are also present in germ-free (GF) animals and newborns18,216,225. However, little is known about the induction of such antibodies in a seemingly “antigen-free” environment found in GF mice and newborns. These “innate like” antibodies have been found to react with low affinity to self- and foreign antigens the host never encountered before215,216,308-311. Notably, a lot of these antibodies also display cross-species reactivity and react at low levels to a variety of unrelated bacteria and viruses215,216,229,233,309,310. For as long as those antibodies have been known, there have been conflicting ideas about their function and mechanisms of polyreactiviy. However, identifying the molecular structure of antigen binding pockets, as well as the concept of V(D)J recombination facilitating the enormous diversity of BCR transcripts, were major steps forward in finding a mechanistic explanation for polyreactivity. Accordingly, one hypothesis suggests that the antigen-binding pockets of germ-line encoded polyreactive antibodies are more flexible thereby accommodating multiple antigens that can bind to different residues compared to antigen binding pockets of somatically mutated high-affinity antibodies213,308. Conversely to T cell independent polyreactive antibody responses, polyreactive responses have also been described for affinity-matured antibodies arising from inefficient checkpoint regulations during B cell development and differentiation233,312,313. In contrast to canonical interactions, non- canonical interactions are not modified by SHM. Thus, non-canonical binding of SIgA involves any other structure of the antibody but the Fab, including Fc and glycan-mediated binding. Glycan-moieties can be found at different parts of the antibody, such as O-glycans decorating 126 the hinge region and N-glycans are largely present at the Fc-portion, J-chain and the SC of SIgA25,147,195,196,314. Glycan-mediated binding modes have been demonstrated for gram- positive bacteria as well as viruses25,76,195-197. SIgA glycan-mediated microbial binding may serve different purposes. Such interactions may be useful in the context of competing for bacterial adhesins, thereby preventing the attachment of bacteria to the epithelium25,195,315, altering bacterial gene expression or to generate inter-or intra species interactions in a distinct niche in the gut34,76,194,196,201. Glycans decorating SIgA seem to provide an ample scaffold for non-canonical microbial interactions in the gut, however the glycosylation status of antibodies might be subjected to variations caused by inflammation, overall IgA production and generated antibody clones, but nevertheless poses an intriguing further concept of SIgA-microbiota interactions25,195. Thus, the status and type of glycosylation seem to impact non-canonical antibody binding to gut bacteria, but may also contribute to canonical SIgA-microbiota interactions potentially by adding to the overall avidity of such interactions. In addition, a mode of binding has been recently described that depends on “superantigen”-like microbial proteins that are expressed by bacteria of the family Lachnospiraceae300. These “superantigens” have the capacity to exclusively bind and stimulate B cells that express human VH3 or murine VH5/6/7 variable regions independent of CDR3 configurations300. Despite important differences between these binding modalities, such different mechanisms readily exist in the mammalian gut and seem to be evolutionary preserved. While there is evidence for a coordinated response of canonical IgA-microbiota interactions, the relevance and contribution of low-affinity/polyreactive responses and non-canonical binding to the microbiota are not well defined. Potentially, these different binding modes act in concert to efficiently regulate and shape the microbiota25,27,91. There might be different “demands” in the host for the prevalence of different binding modes, influenced by the overall microbial diversity, niche occupancy and healthy or inflammatory settings.

7.4 Cross-species reactivity in the adult human gut requires somatic mutations and does not correlate with polyreactivity We investigated the requirement of somatic mutations for cross-species reactivity in our high microbiota-reactive mAb collection by creating mAb germ-line variants that in turn showed substantial loss of microbiota binding capacity. In addition, one might imagine that high- microbiota reactive and cross-species reactivity may have derived from initially polyreactive germ-line or close to germ-line encoded antibodies before undergoing affinity maturation. To test this hypothesis, we screened our high-microbiota reactive mAbs for polyreactivity after germ-line reversion. In fact, for the majority of mAbs we did not find that germ-line reversion caused polyreactivity, suggesting that cross-species reactive mAbs did not derive from originally polyreactive mAbs. Moreover, we were only partially able to assess non-canonical binding properties potentially attributing to high-microbiota and cross-species reactivity of endogenous SIgA, because our human derived mAbs were cloned and expressed as human IgG1-Fc fusion proteins independent of their original isotype and subclass. Although, it seems unlikely that the demonstrated high-microbiota and cross-species reactivity of our tested mAbs rely on non-canonical binding, SIgA glycan-mediated microbial interactions have been descried76,196,201. In the scope of this study we did not find glycosylation of tested mAbs relevant 127 for high microbial binding patterns and believe that the observed differences in binding capacities and specificities between single mAbs reflect differences in canonical Fab- dependent antigen binding. As high microbiota-binding capacity and cross-species reactivity of our tested mAbs was not related to polyreactivity and glycosylation, these results indicate Fab-dependent selection for cross-species reactivity of high microbiota-reactive mAbs. Here, the accumulation of somatic mutations may in turn favor or even enhance the selection for cross-species reactivity. We therefore conclude that affinity-matured intestinal mAbs require somatic mutations to show high-microbiota and cross-species reactive binding properties and likely derive from T cell dependent interactions.

7.5 Cross-species reactive antibody responses in mice and humans Taking a closer look at the host and age Cross-species reactivity appears to be a prevalent feature of both mice and human intestinal antibodies. A recent study in the field of IgA-microbiota interactions found that cross-species reactive IgA in the mouse intestine was mostly polyreactive. Notably, their observations relied on young SPF mice and they found relatively few somatic mutations in isolated PCs used for the generation of mAbs20. When reverting these antibodies to their germ-line configurations, polyreactive cross-species reactivity was preserved. Therefore, the authors suggested that mechanistically, polyreactivity might be the dominant mode of binding for cross-species microbiota targeting IgA in the (mouse) intestine. We in contrast demonstrate that microbiota cross-species reactivity of human intestinal IgA was unrelated to polyreactivity but substantially relied on the accumulation of somatic mutations. Accordingly, in the adult human gut, germ- line transcribed PCs are virtually absent and the vast majority of intestinal IgA+ PCs show substantial numbers of somatic mutations106,208,210,268. In line with our results306, numerous other studies consistently demonstrate little contribution of germ-line like IgA-antibody responses in adult humans106,208,210,268. Usually, the emergence of affinity-matured SIgA- antibody responses in the gut is associated with the presence of microbiota21,27,72,74,89,169,202. But how then to explain the presence of SIgA in germ-free (GF) mice that also show a gradual increase of fecal IgA levels and intestinal IgA+ PC frequencies in a seemingly microbial-derived antigen free environment20,27. Curiously, it has been shown that SIgA in GF mice has the capacity for diverse microbiota binding20. The situation in GF mice might be comparable to very young humans that lack mature and fully developed immune compartments. In fact, a recent study showed the presence of highly abundant autoreactive/polyreactive germ-line or close to germ-line encoded mature naïve B cells isolated from “sterile” human fetuses with the capacity to bind to different members of the microbiota216. Thus, the generation of IgA with the capacity to bind microbiota appears to be evolutionary conserved and may already be present in prenatal naïve B cell repertoires216. The authors suggested that microbiota reactive germ- line encoded prenatal B cells may act as anticipatory precursor B cells of commensal targeting IgA+ PCs enabling the colonization of beneficial microbial taxa and the development of homeostatic host-microbiota interactions after birth216. In addition, it was further suggested that autoreactive/polyreactive B cells might be beneficial early in life to clear apoptotic cells during development, thereby constituting a first line defense against pathogens216,308,309. One hypothesis for the prevalence of such polyreactive/cross-species reactive antibody producing 128

B cells in human fetuses suggests not yet established B cell tolerance checkpoints for the counterselection and elimination of autoreactive/polyreactive B cells216,233, a situation that may be sustained in GF mice. Conversely, and in line with our results, the frequency of polyreactive mature naïve B cells with broad microbiota binding capacity was significantly reduced in adult- derived mature naïve B cell clones216,306. Although, polyreactive antibodies are still prevalent at low frequencies in adult humans106,216,306,312, affinity-matured, cross-species reactive antibodies seem to represent the distinguishing type of IgA-microbiota responses during adulthood168,306. In the context of a highly complex and dynamic microbial setting, the establishment of an affinity-matured antibody repertoire appears to be more efficient in order to adapt and refine antigen specificities during adulthood. We thus propose that the selection for cross-species reactive antibodies depends on the accumulation of somatic mutations and is a hallmark of IgA-microbiota interactions in the adult human gut. However, T cell dependent antibody responses require at least a few days to develop, which is an important notion for fast replicating pathogens. Conversely, T cell independent low-affinity antibody responses are in turn rapidly produced. Low-affinity, polyreactive bacterial targeting might therefore be maintained during adulthood to compensate for the time delay of specific antibody responses15,206,212,225, thus providing broad commensal targeting and fast - albeit limited protection against pathogens27,229,316. Moreover, an important point to be made are the potential differences between the hosts used in such studies. Polyreactivity may indeed be prevalent in young laboratory mice, which have short life spans and are exposed to far fewer environmental antigens than adult humans. However, instead of considering one or the other mechanism as dominant, we think that those two findings may likely reveal fundamental aspects of SIgA-immune responses depending on the host’s microbial configuration and age and may not be mutually exclusive. Therefore, it might be likely that the time and kinetics it takes for intestinal IgA+ PCs to accumulate sufficient amounts of somatic mutations is not given in young SPF mice housed under strictly controlled laboratory conditions. We therefore hypothesize that “innate-like” polyreactive antibodies with the capacity to bind diverse microbial taxa early in life become largely supplanted by non- polyreactive but affinity-matured microbiota cross-species reactive antibodies later in life (Figure 45). This hypothesis seems especially attractive in light of few germinal centers and low numbers of affinity-matured PCs in young individuals20,64,97,216. At the outset, newborns are largely devoid of a complex microbiota and are protected by maternally transmitted antibodies that in turn delay immune response but also shape first immune responses and microbial composition36,38,40,161, which may support the prevalence of low-affinity polyreactive-cross- species reactive antibodies in very young individuals. Those “innate-like” low-affinity antibody responses might act in conjunction with maternally transferred IgA to establish early life intestinal protection and represent a meaningful mechanism for shaping the microbiota while the immune system initially develops. While polyreactive cross-species reactivity may constitute the dominant mechanisms of the microbiota-targeting antibody repertoire in young mice and very young human individuals, affinity-matured antibody repertoires with cross- species reactive binding capacities yet emerge later in life20,168,216,306. We thus suggest that affinity-matured, cross-species reactive antibodies described in our study are the distinguishing type of high-microbiota reactive IgA antibodies in the adult human gut and 129 gradually supplant low-affinity polyreactive antibodies during aging (Figure 45). Both binding mechanisms of cross-species reactivity may be important and may act in different contexts to protect against pathogens, shape the microbiota and provide adequate microbiota-targeting antibody responses throughout life. Prospectively, it would be intriguing to monitor the nature of IgA-responses during aging offering more insight into the proposed transition of antibody repertoires from polyreactivity to affinity-matured, cross-species reactivity.

Figure 45: Putative transition of the prevalent mechanism of microbiota-binding from polyreactive IgA in young individuals to affinity-matured, cross-species reactive IgA in adults. (A) Age dependent shift from polyreactive, cross-species reactive antibodies to affinity-matured, somatic mutation-dependent cross-species reactive antibodies in the adult human gut. (B) Cross-species reactivity early in life is likely driven by polyreactive germ-line or germ-line like antibodies lacking somatic mutations (SM), with low but broad affinity to diverse microbial members (1), whereas later in life these early polyreactive antibodies may become supplanted by affinity-matured antibodies (2) that increase their broad but specific microbial reactivity through continuous accumulation of SMs (red arrow), thereby enhancing the selection for cross-species reactive antibodies (3).

130

7.6 Somatic mutations shape cross-species reactivity of intestinal antibodies Role of structurally similar epitopes in the generation of cross-species reactivity We propose that in a highly complex and dynamic microbial setting found in adult humans, the establishment of affinity-matured antibodies with cross-species reactivity reflects a fundamental mechanism regulating the gut microbiota during adulthood. The enrichment of defined bacterial taxa by our collection of high-microbiota and cross-species reactive IgA mAbs implies antibody specificity. However, the factors that determine which bacterial taxa are targeted are still largely elusive. While early polyreactive antibody responses leading to microbial cross-species binding mostly rely on low-affinity binding to structurally unrelated epitopes present on different bacterial species213,308, affinity-matured intestinal antibodies in the adult gut are likely the result of iterative selection for high-affinity binding24,105,236,268. The generation of high-affinity and cross-species reactive antibody responses may rely on shared epitopes between different bacterial taxa such as highly conserved glycan or peptide structures209,300,305. Accordingly, when testing our mAbs with high-microbiota reactivity to murine fecal bacteria on different microbial configurations from unrelated human donors, high- microbiota reactivity was maintained. These results suggest that either closely related species are uniformly targeted by single mAbs, or that mAbs indeed target structurally similar epitopes shared between divergent bacterial species, independent of the host. This would however indicate that cross-species reactivity is not necessarily associated with promiscuous binding to antigens with distinct molecular structures. Cross-species reactivity observed for our tested mAbs, may be achieved via binding to similar glycan structures shared between divergent microbial species. While SIgA glycan-mediated microbiota binding has mostly been described for non-canonical interactions, evidence for T cell dependent generation of high-affinity IgA responses to glycans has been demonstrated in the setting of whole bacteria encounter182,305. Moreover, distinct glycan patterns identified for different Klebsiella pneumoniae serotypes seem to be shared between unrelated bacteria that facilitate high-affinity cross-species binding and protect against pathogens expressing almost identical glycan structures305,317. We, however, assume that the co-recognition of structurally similar glycans in addition to highly-conserved peptide structures drives the generating of affinity-matured, cross-species reactive intestinal antibodies in the human gut. To further comprehend the phenomenon of specific cross-species reactive antibody responses, the identification of putative epitopes shared between microbial members that are recognized by cross-species reactive IgA is the crucial next step.

Potential mechanisms entailing cross-species reactivity during affinity maturation In contrast to early polyreactive antibody responses216,233, we think that during aging the consecutive accumulation of somatic mutations during affinity-maturation likely enables or even enhances the capacity for microbial cross-species reactivity. One might imagine that the continuous exposure to structurally similar antigens present on different bacteria may favor the selection of B cells with cross-species reactivity, thereby enabling the “broadening” of bacterial specificities during affinity-maturation. Moreover, the generation of those highly diversified SIgA responses seem to require a progressive process that mediates the proliferation of activated B cells undergoing repeated rounds of selection in GCs. In fact, we and others have 131 shown that activated B cells (and memory B cells) show frequent re-entry into pre-existing GCs hosted in different PPs105,209,236. Continuous selection of activated B cells or memory B cells upon re-entry into GCs, however, presumes the sufficient availability of antigen. In contrast to other peripheral SLOs, PPs are a site of continual GC activity prompted by continuous luminal antigen sampling, thereby supporting the selection and proliferation of high-affinity antibody responses to potentially different - but structurally similar microbial antigens209,305. Thus, non- identical but structurally similar antigens from different microbial members presented in different GCs might drive B cell selection to become cross-species reactive. Consequently, the refinement of already existing memory B cells or activated B cells to microbial changes likely enables to continuously shape and refine the prevalent antibody repertoire (Figure 46). Thus, once fully initiated, chronically active GCs allow for the continuous provision of diverse antigens, hereby providing an environment that entails the generation of highly mutated IgA+ PCs typically found in the adult intestine106,136,208,210,306. The observation of clonally related B cells in different PPs and SI LP strongly indicates that the generation of SIgA responses is not restricted to one PP GC, but rather emerges from the re-entry of activated B cells or circulating memory B cells into pre-existing GCs undergoing repeated rounds of clonal selection and expansion in different PPs105,236,274. Sustained SHM in different PP GCs likely contributes to the generation of cross-species reactive antibodies, assuming that affinity-maturation selects for increased affinity for the prevalent antigen but specificity100,105. But for such a mechanism to work requires an initial affinity of the primary responding B cell318. However, defining the initial B cell activating antigen and subsequent antigens facilitating selection during affinity- maturation still poses a great challenge. Our data supports the hypothesis of B cell re-circulation through different PP GCs, as we found eYFP+ B cells in neighboring PPs of single injected PPs. Consequently, the re-entry of activated B cells into different PP GCs with subsequent further expansion and differentiation may indicate a principal concept of IgA-repertoire diversification under homeostatic conditions (Figure 46). Interestingly, we found this mechanism to apply in SPF mice, where we found a continuous generation of GC engaged AID/eYFP expressing B cells that re-circulated through different PPs up to 21 days after induction. Notably, eYFP+ cells derived from a single PP were readily found as IgA+ PCs in the SI LP, which is in line with other reports showing clonal overlap of PP derived B cells and PCs in the SI LP93,105,236,268. Quite strikingly, we found a continuous generation of PP derived IgA+ PC waves without neo-antigen challenge in SPF mice under homeostatic conditions. At all investigated time points we found GC engaged eYFP+GL7+ B cells throughout PPs, and to a lesser extend in MLN. While we did not find strong evidence for major contributions of the MLN to B cell priming and differentiation as compared to PPs under homeostatic conditions, the MLN might become important upon systemic infections. Crucially, the presence of eYFP+GL7+ B cells strongly indicates that GC derived B cells significantly contribute to the intestinal PC pool under homeostatic conditions in young laboratory mice. Importantly, the maintenance of chronically active GCs requires T cell engagement with activated B cells64,166,209. Therefore, eYFP expression observed in activated B cells and intestinal PCs most likely relies on T cell interactions and subsequent GC formation. This observation largely excludes the notion that microbiota targeting antibodies in mice are mainly the result from “innate-like/natural” antibody responses independent of T cell help and affinity- 132 maturation. In line with this, a recently published study conducted in mice demonstrates that a high proportion of microbiota-targeting IgA rely on T cell engagement and GC formation136. Accordingly, the high rate of GC engaged eYFP+ B cells observed in our system strongly indicates that the mechanism of T cell dependent IgA responses to the microbiota is prevalent in mice. Furthermore, it is likely that also in mice, factors such as age and complex microbial colonization, might play a crucial role for the generation of affinity-matured antibody responses found in adult humans. Our observation of IgA+ PC generation under laboratory SPF conditions likely indicates an incomplete picture of IgA repertoire kinetics as compared to the natural situation. Nonetheless, our established model to fate-track activated B cells may prospectively allow us to monitor the age-dependent generation of different types of IgA- microbiota responses in more detail, using microbial configurations resembling the natural situation. While our data demonstrate that somatic mutations crucially contribute to antibody cross-species reactivity, future work lies ahead in identifying the initial B cell activating antigen, as well as subsequently encountered antigens, contributing to the affinity-maturation of cross- species reactive intestinal antibodies. We suggest that colonization models in combination with single cell sequencing and antibody cloning will give more mechanistic insight into affinity- matured, cross-species reactive antibody responses to the microbiota.

Figure 46: Model for B cell re-entering into pre-existing germinal centers of different Peyer’s patches. The re-entry into different Peyer’s patch (PP) germinal centers (GCs) leads to a continuous accumulation of somatic mutations entailing the selection for (mem) B cells with cross-species reactivity to structurally similar antigens or shared epitopes between divergent bacteria. PP#1 and PP#2 represent different PPs found in the small intestine. Individual PPs might display distinct antigen environments created by the uptake of select microbial members (a, b). Antigen experienced B cells (yellow) primed in a single PP (1) may transiently migrate to the MLN (2) thereafter entering the blood stream (3) to re- enter pre-existing GCs of other PPs (4). Activated B cells or memory B cells undergoing repeated rounds of selection to potentially different microbial antigens (a, b) present in different GCs via migration (displayed by the black dotted line) may entail the generation of affinity-matured and cross-species reactive IgA+ PCs homing to the SI LP (5) adding to the prevailing IgA secreting PC pool (6) (Figure adapted from91).

133

7.7 Role of T cells in affinity-matured cross-species reactive IgA responses While studies show the capacity of early polyreactive intestinal antibodies to coat gut commensals in young mice and very young children20,216, the generation of high-affinity antibody responses is typically acquired via T cell help and requires GC engagement of activated B cells64. The important role of T cells to establish potent SIgA functions has been demonstrated in mice capable of CSR but not SHM that showed dysbiosis and compromised mucosal integrity166,176. Similarly, T cell transfer into T cell deficient mice resulted in an IgA- dependent shift of microbial composition and diversity67,111,174. These studies strongly suggest that although IgA responses are maintained in T cell deficient settings, T cells are nevertheless required to mediate antigen dependent SHM to mount potent and high-affinity antibody responses27,136,205,209. Thus, T cell mediated affinity-maturation is critical to establish fully adequate IgA responses. While cross-species reactivity can be achieved through T cell- independent mechanisms (e.g. through polyreactivity), the hallmarks of affinity-maturation observed in our collection of cross-species reactive mAbs strongly suggest T cell dependent induction. However, we are only starting to understand the complex processes involved in T cell dependent GC derived antibody responses, how chronic GC responses are sustained and how such diversification mechanisms operate in the context of maintaining a stable and homeostatic microbiota configuration. One approach to mechanistically understand how affinity-matured cross-species reactive antibodies are generated and contribute to the overall microbiota-targeting antibody repertoire may come from a deeper understating of how the accumulation of somatic mutations shapes the generation of such antibody responses. Thus, future work should aim to examine the contribution of accumulated somatic mutations to affinity-matured, cross-species IgA-microbiota responses. We think that one intriguing approach to interrogate somatic mutations shaping antibody cross-species reactivity would be to gradually revert single mutations of highly mutated antibodies. Employing 16S sequencing of targeted bacterial species after step-wise reversion would elucidate the kinetics of acquired somatic mutations to the “broadening” of IgA-specificities to the microbiota.

7.8 What determines the extent of SIgA-bacteria binding? Our data shows that high-microbiota and cross-species reactive mAbs maintain high binding- capacities to different microbiota configurations. Although cross-species reactive mAbs only enrich for select microbial taxa, mAb-binding never affected all members of any particular taxon. This may suggest genetic or phenotypic heterogeneity among individual bacterial species, which might explain “partial” species binding by cross-species reactive mAbs.

Modulations of SIgA by the microbiota We meanwhile acknowledge that the host is not the only source controlling IgA-microbiota interactions, as more seminal work addresses the contribution of the microbiota itself to modulate and interact with intestinal IgA76,194,201,319. Accordingly, members of the microbiota have evolved ways to either exploit or escape IgA targeting. Some murine-derived bacterial species have the capacity to actively enhance or decrease intestinal IgA levels9,79,319. Bacteria targeting by IgA has been shown to be altered in IBD patients109,170, undernourished children172 and in diet-dependent murine settings76. An inflamed intestinal environment is moreover 134 associated with the overgrowth and infiltration of facultative anaerobes and opportunistic pathogens such as Proteobacteria9,46,74,109. However, the circumstances for distinct epitope expression of bacterial antigens that are targeted by IgA are still largely elusive. Only recently, data emerged trying to unveil potential regulatory systems that promote or limit microbial recognition by SIgA76,194,201,319. The understanding of IgA-microbiota interactions still contains major gaps, regarding the precise host-immune as well microbial factors leading to altered or increased IgA binding during both healthy situations and dysbiosis. Upon pathology, the expression of microbial epitopes recognized by cross-species reactive IgA may potentially be impacted by changes in the host, which in turn facilitates the overgrowth of pathogens. However, we speculate that microbial epitope expression and SIgA-targeting is not only affected in pathology. Commensals “exploiting” SIgA-binding may also act to facilitate their retention within a specific host gut-niche. In line with this, some commensals have the capacity to “invite” their binding by SIgA194,201. For instance, some bacteria may use SIgA-linked glycans as a carbon source198-200 or as a stimulus regulating epitope expression25,180,187,201. IgA glycan- mediated interactions with B. thetaiotaomicron modulates the expression of polysaccharide utilization loci, which engenders gut symbiosis between B. thetaiotaomicron and mucus- associated bacteria from the Firmicutes phylum. In Bacteroides fragilis a sensor/regulatory system shapes the surface architecture to “invite” its binding by SIgA leading to robust mucus anchoring, thereby providing colonization resistance against competitors and stabilizing an overall beneficial bacterial community25,180,186,194,201,320. However, understanding the intestine as a dynamic ecosystem, it is easy to picture that depending on the niche and nutrient availability bacteria display different growth-states and different metabolic activity. Dividing bacteria, spore formation or dying bacteria may not express the same epitopes as their viable forms, resulting in altered targeting by IgA.

Influence of bacterial strain level variation on IgA-binding The gut microbiota as well as microbial derived metabolites have the capacity to influence immune cell populations10,74,321, thereby influencing and shaping host immune functions. However, a lot of immune-modulatory functions seem to be bacterial strain specific. As such, bacterial strains that play a causal immune-modulatory role in one host may not correlate with that phenotype in others60,319,322. Given the high amount of strain level variation within bacterial populations, it is complicated to determine if the host itself or variations in effector functions excerted by different strains, impact the IgA-coating status of bacterial species. It is moreover possible that IgA-microbiota responses may be host-tailored to target particular sub-strains or morphological subgroups within bacterial species. The question thus remains if one can assume that similar strains are always equally targeted by IgA when present in different microbiota configurations or hosts.

Low microbial diversity model to investigate partial IgA-binding to bacteria To investigate the incomplete bacterial species binding and to address potential strain level variations, we employed a restricted microbial diversity model offered by the oligoMM12 consortium. Notably, when screening our high microbiota-reactive mAbs on this defined microbial configuration we observed overall decreased binding capacities as compared to 135 binding capacities to SPF bacteria. Nonetheless, 16S sequencing of bound bacteria derived from oligoMM12 fecal material revealed almost complete binding to Akkermansia muciniphila in contrast to A. muciniphila binding when present in SPF fecal samples or in vitro cultivated. Conversely, we found high enrichment of Muribaculum intestinale in mAb positive samples when isolated from SPF feces but not when present in oligoMM12 feces. Similarly, we found this to be true for several Clostridia and Bacteroidia species that were highly enriched in mAb positive samples when present in SPF feces but not enriched in the context of oligoMM12 feces. Thus, depending on the ecological context (oligoMM12 or complex SPF) some bacteria may establish different genetically driven phenotypical states resulting in varied coating by IgA. Partial binding found for SPF fecal bacteria may be the result of either strain level differences within bacterial species or different growth-states. We assumed that the overall abundance of bacterial species present in a given microbiota sample might affect their binding frequency by IgA. To test this, we used an in vitro approach, where confounding factors such as discrepancies in relative species abundances or strain level differences would be circumvented. However, we found that in vitro cultivated bacteria were significantly less targeted by IgA mAbs and in the case of few target species these were largely not in accordance to the 16S data obtained from oligoMM12 fecal samples. This observation was most striking for A. muciniphila that was almost completely targeted by mAbs in the context of oligoMM12 feces but not at all when in vitro cultivated. Distinct bacterial epitopes recognized by our collections of high-microbiota, cross-species reactive mAbs may only be present in a complex microbial setting in vivo. Our data suggest that other aspects but IgA-specificity, such as growth-or metabolic states and genetic variations within bacterial species require consideration when investigating IgA-microbiota interactions. As demonstrated recently, nutritional availability and the presence of a complex microbiota have major implications for the expression and availability of bacterial epitopes34,76,319. Thus, merely putting bacteria in culture does not seem to resemble and reflect the situation in vivo. While we think that using a defined low diversity microbial model may be a good approach to explain the impact of strain level variations to partial IgA binding, prospectively one may want to exploit elaborated in vitro cultivating systems that allow for controlled manipulation of distinct variables. This may help to identify conditions and bacterial networks promoting or limiting bacterial epitope expression. Differently cultivated bacteria could then be used for subsequent antibody screening with known specificities in combination with proteomic and metabolomics analysis.

7.9 Functional consequences of cross-species reactive IgA in the healthy gut Polyreactive, non-mutated intestinal antibodies might be relevant in early life to provide broad targeting of the microbiota and may act as a first line defense, whereas the acquisition of affinity-matured, cross-species reactive SIgA-microbiota responses seem to be a hallmark in the adult gut to efficiently control the microbiota. We meanwhile accept the concept that gut microbiota drives the production of intestinal IgA 21,27,72,136,202, with the majority of the produced IgA displaying specific binding to particular microbial antigens136,166,319. In humans, microbial targeting by SIgA is usually the result of T cell dependent high-affinity Fab-mediated interactions168,306,319. Intestinal high-affinity IgA typically targets commensals that colonize the mucus or are closely associated with the gut 136 epithelium74,79,109,141,323,324. A common feature of these taxa is their potential for opportunistic pathology. Evidence for the generation of T cell dependent high-affinity IgA responses to such microbial members have been shown for Mucispirilum sp., Akkermansia muciniphila, SFB and beneficial bacteria from the Firmicutes phylum9,79,109,141,166, which is in accordance with our observations306. While IgA-binding has many effects on the microbiota, such as motility alterations, modulating bacterial phenotypes and commensal growth rates25,91,180,193, IgA- binding to bacteria also promotes adhesion and colonization of commensals in distinct gut niches25,34,186,194,201. Intestinal antibodies thus seem to have a context-dependent role of either excluding microbes from - or including microbes to distinct niches29,180,183,186,194,203,315. Although, IgA in the intestine is exceedingly produced on a daily basis, the question remains how the host is able to sustain the production of relevant IgA levels to affect the microbiota, considering luminal flow rates, nutrient availability and a protease rich environment in the intestine. In the setting of this highly complex and dynamic microbial ecosystem, it seems quite challenging to produce and maintain sufficient amounts of single-species targeting IgA. The generation of affinity-matured, cross-species reactive antibodies under homeostatic conditions, might circumvent this challenge. One of the best characterized mechanisms of SIgA-bacteria interactions describes immune exclusion, a process regulating bacterial access to intestinal epithelial cells29,180,184,185. However, in the context of a highly diluted intestinal environment and high niche competition, some bacteria may be too sparse to allow for effective immune exclusion by monospecific antibodies. Instead, coating of several low abundant species by cross-species reactive mAbs potentially allows for more effective immune exclusion of opportunistic pathogens. Another hypothesis in favor of the sustained generation of affinity- matured, cross-species reactive antibodies comes from the recently proposed concept of enchained growth, which describes antibody-mediated cross-linking and agglutination of dividing bacteria described for Salmonella specific antibodies182. Under homeostatic conditions, however, cross-species reactive antibodies may employ this binding mode to cross-link planktonic members of the microbiota. In this scenario, cross-linking might be especially relevant for the retention of beneficial but low-abundant microbial members promoting bacterial colonization to intestinal niches, thereby maintaining bacterial diversity in the gut25,142,183,194,201,315. The generation of cross-species reactive IgA may enable to maintain efficient antibody responses to the microbiota and may promote a fast adaptation to microbial alterations in the gut.

7.10 Cross-species reactive antibodies are maintained in the inflamed gut Notably, we found cross-species reactive mAbs with overlapping binding profiles between HD and CD high microbiota-reactive mAbs. Both HD and CD IgA binding profiles comprised selective broad binding to commonly targeted bacteria in addition to distinct microbial members that were uniquely enriching by single mAbs. Commonly targeted bacteria were similarly bound by HD and CD derived mAbs, indicating that IgA cross-species-reactive specificities are largely maintained independently of a healthy or inflamed situation. The presence of such antibody responses under healthy and inflammatory conditions suggests a functional importance for cross-species reactive antibodies in both settings. Moreover, these results further indicate that microbial shifts associated with IBD diseases states do not necessarily result in rapid changes 137 of the prevailing IgA repertoire and specificities, which is consistent with previous observations210,268. Although we propose that affinity-matured, cross-species reactive IgA antibodies are the distinguishing type of relevant antibody responses in the healthy and inflamed gut, we cannot rule out additional functional consequences of antibodies generated under inflammatory conditions.

Role of intestinal IgG to mucosal homeostasis and inflammation While cross-species reactive IgA with high microbiota-binding capacities are highly prevalent in both healthy and inflamed gut, we also found high microbiota-binding for IgG mAbs. Although the role of IgG in contributing to mucosal homeostasis became recently more appreciated, we did not further investigate their binding profiles within the scope of this work. Nonetheless, IgG specific to commensals can be detected in the mucosa and peripheral blood of healthy humans, albeit usually with non-overlapping specificities to the gut microbiota as compared to SIgA35,83,325. It has been proposed that mucosal IgG149,164 may control both invasive and non- invasive mucosal bacteria in addition to specific IgA responses78,165,324. This suggests that IgG may contribute to mucosal homeostasis by controlling microbial members with penetrant or mucus-degrading properties78,108,110. Further evidence for a protective role of IgG in the gut comes from studies demonstrating the requirement for specific IgG, but not IgA, in clearing the attaching/effacing murine pathogen Citrobacter rodentium326. IgG induced in response to disseminating microbial members was shown to provide protection against Escherichia coli or Salmonella typhi infections150. While IgG appears to synergize with SIgA for effective microbiota control as well as clearance of pathogens, increased IgG responses are typically associated with mucosal inflammation46,78,108,110,140. Accordingly, gut inflammation is associated with increased accumulation of intestinal IgG-expressing PCs83,140,149. Depending on the subclass, mucosal IgG responses can be non-or pro-inflammatory83,104,140,148. During gut inflammation, pro-inflammatory IgG subclasses exhibit a high complement-activating and FCγ- receptor I (FcγRI)/FcyRIII signaling function140,148,327. The exacerbating inflammatory function of IgG found in IBD patients is driven by increased TH17 cell expansion and neutrophil recruitment as well as by stimulation of pathogenic gut macrophages via FcγR signaling140,148,327. In line with this, we found a complete lack of no/low microbiota-reactive CD IgG mAbs and a higher proportion of intermediate microbiota-reactive CD derived IgG mAbs, in contrast to HD derived IgG mAbs. However, it is difficult to assess the specificities and targeting profiles of these CD derived IgG mAbs due to the technical limitation of sorting bacteria below 5% binding for subsequent 16S sequencing. Although we focused our work on investigating the binding properties of high microbiota-reactive IgA in the healthy and inflamed gut, the role of intestinal IgG contributing to both homeostasis and gut inflammation needs to be further addressed in the future.

Functions of intestinal cross-species reactive IgA during inflammation During gut inflammation, cross-species reactive IgA may in fact contribute to efficient agglutination of bacteria leading to enhanced luminal clearance of pathogens. Mechanistically, cross-species reactivity may act by using the endogenous microbiota as scaffolding to more efficiently target planktonic pathogens182,183. Regardless of such beneficial effects of cross- 138 species reactivity during inflammation, pro-inflammatory outcomes of SIgA-microbiota targeting may also occur. While IgA-induced inflammation is important to efficiently clear pathogens, the excessive activation of pro-inflammatory pathways is associated with chronic inflammation and sustained pathology. In humans, IgA interaction with the Fc alpha receptor (FcαRI) can initiate immune activating functions and cytokine production328. The expression of the FcαRI is restricted to myeloid immune cells such as neutrophils, monocytes, eosinophils, macrophages and DC subsets recruited to the intestine upon infection148,329,330. Under homeostatic conditions only few FcαRI expressing cells are present in the SI LP and intestinal macrophages lack FCαRI expression altogether148,330, instead CD103+ DCs in the intestine are the main FcαRI expressing cells for the initial recognition of IgA-immune-complexes148,328. Individual stimulation of the FcαRI is not sufficient for IgA-induced pro-inflammatory functions and cytokine production, but requires the collaboration with pattern recognition receptors (PRRs) to either amplify or inhibit specific cytokine production. Therefore, a tightly regulated balance to control tolerance and inflammatory responses is crucial, considering the extremely vast abundance of commensals and microbial products in the gut. The intestinal immune system typically requires a second signal to differentiate homeostatic conditions from inflammation. As it seems, the key switch from FcαRI induced suppression of pro-inflammatory responses under homeostatic conditions to promoting inflammation depends on whether IgA binds to FcαRI in soluble or aggregated form. Upon infection, invading bacteria are opsonized by local dimeric IgA in the SI LP and form aggregated immune complexes. The co-activation of PRRs and FCαRI by IgA-opsonized bacteria drives inflammatory responses in myeloid cells86,148,328,331. In the situation of impaired gut integrity and chronic inflammation typical for IBD, cross-species reactive SIgA might add to exacerbated inflammatory responses to IgA- coated microbial members. An impaired intestinal barrier promotes the massive influx and accumulation of IgA-microbiota immune complexes in the SI LP. Consequently, gut inflammation will affect the overall immune cell composition in the intestine. Here, pro- inflammatory myeloid cells will encounter IgA-coated intestinal bacteria, which further enhances inflammation via chronically activated FcαRI, resulting in the release of pro- inflammatory cytokines such as TNF and IL-1ß and IL-23148. Thus, SIgA-coating of bacteria might entail different outcomes emphasizing the capacity of IgA to modulate ensuing immune responses. Depending on the context, interaction with the FcαRI can initiate both pro- or anti-inflammatory responses328.

139

8 Conclusion In conclusion, our data demonstrate a new facet of IgA-microbiota interactions in the adult human intestine. High-microbiota binding and cross-species reactivity of intestinal IgA relies on the accumulation of somatic mutations in both healthy and inflamed gut (Figure 45). Cross- species reactive antibodies in the human gut show the typical hallmarks of classical T cell dependent affinity-matured antibody responses. The selection for cross-species reactivity likely follows a progressive process of accumulating somatic mutations through continuous affinity-maturation in different PP GCs. This process may enable broad but specific antibody responses, enabling a fast adaptation to microbial alterations as well as infection (Figure 46). We propose that cross-species reactivity is one of the principal mechanisms allowing the host to efficiently interact with the intestinal microbiota in order to maintain host-microbiota homeostasis as well as to counteract pathogen challenges under inflammatory conditions.

140

9 Appendix 9.1 List of abbreviations AID activation induced cytidine deaminase BCR B cell receptor BM bone marrow bp base pairs CD cluster of differentiation cm centimeter Cre tyrosine recombinase CSR class switch recombination D diversity or D gene segment DC dendritic cell ERT estrogen receptor FDC follicular dendritic cell g gravity GC germinal center GF germ-free GL germ-line h hours IBD inflammatory bowel disease Ig immunoglobulin IgA immunoglobulin A IL interleukin ip intraperitoneal LP lamina propria mAb monoclonal antibody mg milligram min minutes ml milliliter MLNs mesenteric lymph nodes ON over night OTU operational taxonomic unit PC plasma cell PCR polymerase chain reaction PFA paraformaldehyde PP Peyer’s patch PRR pattern recognition receptor RAG recombinantion-activating gene rmp revolutions per minute RT room temperature s seconds SHM somatic hypermutation SI small intestine SLOs secondary lymphoid organs SM somatic mutations

141

SN supernatant SPF specific pathogen free TfH cell T follicular helper cell UC ulcerative colitis WT wild-type µg microgram µl microliter

142

9.2 OTU tables Table 6: Taxonomic classification of OTUs based on 16S rRNA gene amplicon sequence identity. OTU Phylum Class Order Family Genus Species 96.12% Olsenella OTU_59* Actinobacteria Coriobacteriia Coriobacteriales Atopobiaceae Olsenella profusa OTU_91 Actinobacteria Actinobacteria Actinomycetales Micrococcaceae Rothia

OTU_108 Actinobacteria Actinobacteria Actinomycetales Micrococcaceae Rothia

OTU_86 Actinobacteria Actinobacteria Actinomycetales Corynebacteriaceae Corynebacterium

OTU_77 Actinobacteria Actinobacteria Actinomycetales Micrococcaceae Micrococcus

OTU_100 Actinobacteria Actinobacteria Actinomycetales Micrococcaceae Kocuria

OTU_106 Actinobacteria Actinobacteria Actinomycetales Pseudonocardiaceae Pseudonocardia

OTU_557 Actinobacteria Actinobacteria Actinomycetales Corynebacteriaceae Corynebacterium

OTU_114 Actinobacteria Actinobacteria Actinomycetales Actinomycetaceae Actinomyces

OTU_128 Actinobacteria Actinobacteria Actinomycetales Corynebacteriaceae Corynebacterium

OTU_1659 Actinobacteria Actinobacteria Actinomycetales Micrococcaceae Rothia 87.93% Muribaculum OTU_152* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 97.42% Bacteroides OTU_5757* Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides acidifaciens 91.38% Muribaculum OTU_19* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale OTU_32* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Duncaniella 100% Duncaniella muris 99.57% Bacteroides OTU_3* Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides vulgatus 100% Muribaculum OTU_2* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 98.71% Parabacteroides OTU_9* Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae Parabacteroides distasonis 96.12% Muribaculum OTU_4* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 93.53% Muribaculum OTU_20* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 90.52% Duncaniella OTU_18* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Duncaniella muris Muribaculum 91.38% Muribaculum OTU_5* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae intestinale OTU_8* Bacteroidetes Bacteroidia Bacteroidales AC160630 PAC002482 100% EF603735 87.5% Paramuribaculum OTU_1* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Paramuribaculum intestinale OTU_51 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae 100% Bacteroides OTU_11* Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides caecimuris 95.26% Muribaculum OTU_15* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 94.4% Muribaculum OTU_29* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 92.24% Duncaniella OTU_30* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Duncaniella muris OTU_143 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_47 Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae 89.66% Muribaculum OTU_7* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale 95.26% Muribaculum OTU_31* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale

143

OTU_44 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae Parabacteroides 100% Paramuribaculum OTU_28* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Paramuribaculum intestinale OTU_64 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae 93.53% Muribaculum OTU_23* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum intestinale OTU_61 Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum 93.1% Paramuribaculum OTU_22* Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Paramuribaculum intestinale OTU_65 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_611 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU_55 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae 99.57% Bacteroides OTU_21* Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides uniformis OTU_533 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU_555 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_3119 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_381 Bacteroidetes Bacteroidia Bacteroidales

OTU_2037 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_444 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_3266 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_685 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_437 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_72 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_3683 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_2077 Bacteroidetes Bacteroidia Bacteroidales

OTU_2049 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU_70 Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Paramuribaculum

OTU_1370 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_500 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_78 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_6055 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU_369 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU_440 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_428 Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Muribaculum

OTU_1754 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae

OTU_140 Bacteroidetes Bacteroidia Bacteroidales Prevotellaceae

OTU_98 Bacteroidetes Bacteroidia Bacteroidales Prevotellaceae Prevotella

OTU_117 Bacteroidetes Bacteroidia Bacteroidales Porphyromonadaceae Odoribacter

OTU_2112 Bacteroidetes Bacteroidia Bacteroidales Muribaculaceae Paramuribaculum

OTU_103 Cyanobacteria Chroobacteria Oscillatoriales Planktothrix Trichodesmium

OTU_146 Cyanobacteria Chroobacteria Oscillatoriales Pseudanabaenaceae Leptolyngbya_g2 100% Mucispirillum OTU_40* Deferribacteres Deferribacteres Deferribacterales Deferribacteraceae Mucispirillum schaedleri Deinococcus- OTU_144 Deinococci Deinococcales Deinococcaceae Deinococcus Thermus Deinococcus- OTU_151 Deinococci Thermales Thermaceae Meiothermus Thermus

144

99.57% Robinsoniella OTU_14* Firmicutes Clostridia Clostridiales Lachnospiraceae Robinsoniella peoriensis OTU_10* Firmicutes Clostridia Clostridiales Lachnospiraceae Blautia 100% Blautia hansenii 96.93% Eubacterium OTU_26* Firmicutes Clostridia Clostridiales Lachnospiraceae Eubacterium_g17 xylanophilum OTU_53 Firmicutes Clostridia Clostridiales Oscillospiraceae Pseudoflavonifractor 100% Enterocloster OTU_16* Firmicutes Clostridia Clostridiales Lachnospiraceae Enterocloster bolteae OTU_43 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_75 Firmicutes Clostridia Clostridiales Ruminococcaceae 100% Lactobacillus OTU_34* Firmicutes Bacilli Lactobacillales Lactobacillaceae Lactobacillus reuteri OTU_69 Firmicutes Bacilli Lactobacillales Lactobacillaceae Lactobacillus

OTU_37 Firmicutes Clostridia Clostridiales Lachnospiraceae 100% Lactobacillus OTU_39* Firmicutes Bacilli Lactobacillales Lactobacillaceae Lactobacillus gasseri OTU_71 Firmicutes Bacilli Lactobacillales Streptococcaceae Streptococcus

OTU_92 Firmicutes Bacilli Lactobacillales Streptococcaceae Streptococcus

OTU_1275 Firmicutes Clostridia Clostridiales Lachnospiraceae Clostridium XlVa

OTU_7685 Firmicutes Clostridia Clostridiales Lachnospiraceae 94.42% Anaerotignum OTU_24* Firmicutes Clostridia Clostridiales Lachnospiraceae Anaerotignum lactatifermentans OTU_42 Firmicutes Clostridia Clostridiales Lachnospiraceae 93.99% Clostridium OTU_13* Firmicutes Clostridia Clostridiales Lachnospiraceae Lacrimispora methoxybenzovorans OTU_17* Firmicutes Clostridia Clostridiales Lachnospiraceae Kineothrix 97% Kineothrix alysoides

OTU_111 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_76 Firmicutes Bacilli Bacillales Staphylococcaceae Staphylococcus

OTU_54 Firmicutes Clostridia Clostridiales Ruminococcaceae Flavonifractor

OTU_4806 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_549 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_130 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_79 Firmicutes Clostridia Clostridiales Ruminococcaceae

OTU_189 Firmicutes Clostridia Clostridiales Lachnospiraceae Anaerotignum

OTU_83 Firmicutes Clostridia Clostridiales Ruminococcaceae Anaerotruncus 93.13% Stomatobaculum OTU_33 Firmicutes Clostridia Clostridiales Lachnospiraceae Moryella longum OTU_2803 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_5173 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_68 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_38 Firmicutes Clostridia Clostridiales Lachnospiraceae Blautia 94.85% Blautia faecicola 93.99% Butyrivibrio OTU_805 Firmicutes Clostridia Clostridiales Lachnospiraceae Butyrivibrio crossotus OTU_60 Firmicutes Clostridia Clostridiales Lachnospiraceae 94.4% Kineothrix OTU_5341 Firmicutes Clostridia Clostridiales Lachnospiraceae Kineothrix alysoides OTU_48 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_57 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_84 Firmicutes Clostridia Clostridiales Ruminococcaceae Clostridium IV

OTU_1198 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_1204 Firmicutes Clostridia Clostridiales Lachnospiraceae 145

OTU_2621 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_627 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_970 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_73 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_3535 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_707 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_570 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_366 Firmicutes Bacilli Lactobacillales Aerococcaceae Granulicatella

OTU_80 Firmicutes Negativicutes Selenomonadales Veillonellaceae Veillonella

OTU_87 Firmicutes Bacilli Lactobacillales Carnobacteriaceae Dolosigranulum

OTU_115 Firmicutes Bacilli Bacillales Bacillaceae Halolactibacillus Bacillales_Incertae OTU_107 Firmicutes Bacilli Bacillales Gemella Sedis XI OTU_124 Firmicutes Clostridia Clostridiales Ruminococcaceae Clostridiales_Incertae OTU_95 Firmicutes Clostridia Clostridiales Finegoldia Sedis XI Clostridiales_Incertae OTU_82 Firmicutes Clostridia Clostridiales Peptoniphilus Sedis XI OTU_350 Firmicutes Bacilli Bacillales Staphylococcaceae Macrococcus

OTU_142 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_123 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_131 Firmicutes Clostridia Clostridiales Ruminococcaceae

OTU_129 Firmicutes Clostridia Clostridiales Lachnospiraceae

OTU_134 Firmicutes Clostridia Clostridiales Ruminococcaceae Epsilonproteoba Campylobacterale 100% Helicobacter OTU_6* Proteobacteria Helicobacteraceae Helicobacter cteria s typhlonius Gammaproteoba 100% Pseudomonas OTU_12* Proteobacteria Pseudomonadales Pseudomonadaceae Pseudomonas cteria aeruginosa Deltaproteobact OTU_46 Proteobacteria eria Alphaproteobact OTU_36 Proteobacteria Kiloniellales Kiloniellaceae Kiloniella eria Betaproteobacte OTU_27* Proteobacteria Burkholderiales Sutterellaceae Turicimonas 100% Turicimonas muris ria Betaproteobacte OTU_58 Proteobacteria Burkholderiales Comamonadaceae ria Gammaproteoba OTU_35 Proteobacteria Pseudomonadales Pseudomonadaceae Pseudomonas cteria OTU_25* Proteobacteria Betaproteobacte Burkholderiales Sutterellaceae Parasutterella 93.99% Parasutterella ria excrementihominis Betaproteobacte 100% Alcaligenes OTU_41* Proteobacteria Burkholderiales Alcaligenaceae Alcaligenes ria faecalis subsp. faecalis Gammaproteoba OTU_6772 Proteobacteria Enterobacteriales Enterobacteriaceae Escherichia/Shigella cteria Deltaproteobact OTU_56 Proteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio eria Gammaproteoba OTU_67 Proteobacteria Enterobacteriales Enterobacteriaceae cteria Alphaproteobact OTU_63 Proteobacteria Rhizobiales Phyllobacteriaceae Aquamicrobium eria Deltaproteobact OTU_62 Proteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio eria Epsilonproteoba Campylobacterale OTU_1764 Proteobacteria Helicobacteraceae Helicobacter cteria s

146

Gammaproteoba OTU_74 Proteobacteria Xanthomonadales Xanthomonadaceae Stenotrophomonas cteria Alphaproteobact OTU_94 Proteobacteria Rhodobacterales Rhodobacteraceae Paracoccus eria Gammaproteoba OTU_50 Proteobacteria Enterobacteriales Enterobacteriaceae cteria Gammaproteoba 100% Moraxella OTU_45* Proteobacteria Pseudomonadales Moraxellaceae Moraxella cteria osloensis Betaproteobacte Burkholderiales_incert OTU_127 Proteobacteria Burkholderiales Tepidimonas ria ae_sedis Betaproteobacte OTU_113 Proteobacteria Burkholderiales Comamonadaceae Curvibacter ria Gammaproteoba OTU_133 Proteobacteria Xanthomonadales Xanthomonadaceae cteria Epsilonproteoba Campylobacterale OTU_90 Proteobacteria Helicobacteraceae Helicobacter cteria s Gammaproteoba OTU_101 Proteobacteria Pseudomonadales Moraxellaceae Acinetobacter cteria Gammaproteoba OTU_66 Proteobacteria Pasteurellales Pasteurellaceae cteria Alphaproteobact OTU_104 Proteobacteria Rhodobacterales Rhodobacteraceae Paracoccus eria Alphaproteobact Sphingomonadale OTU_132 Proteobacteria Sphingomonadaceae Sphingomonas eria s Alphaproteobact OTU_97 Proteobacteria Rhizobiales Hyphomicrobiaceae Pedomicrobium eria Gammaproteoba OTU_49 Proteobacteria Pseudomonadales Moraxellaceae Moraxella cteria Gammaproteoba OTU_157 Proteobacteria Pasteurellales Pasteurellaceae Mannheimia cteria Betaproteobacte OTU_135 Proteobacteria Burkholderiales Alcaligenaceae ria Alphaproteobact OTU_120 Proteobacteria Rhodobacterales Rhodobacteraceae Amaricoccus eria Taxonomic assignment of phylum to family level was based on the RDP classifier294 and with * marked OTUs are lineage and identity matched (closest species with a valid name and corresponding 16S rRNA gene sequence identity using EZbiocloud295; OTU_8 could not be classified to a species name with a valid name, but was identified as reported uncultured clone EF603735.

Table 7: Bacterial species constituting the oligoMM12 microbial consortium303. Phylum Order Family Genus Species Strain Gram

Firmicutes Clostridiales Ruminococcaceae Acutalibacter Acutalibacter muris KB18 + Akkermansia Verrucomicrobia Verrucomicrobiales Akkermansiaceae Akkermansia YL44 Ø muciniphila Bacteroides Bacteroidetes Bacteroidales Bacteroidaceae Bacteroides I48 Ø caecimuris Bifidobacterium Actinobacteria Bifidobacteriales Bifidobacteriaceae Bifidobacterium YL2 + longum animalis Firmicutes Clostridiales Lachnospiraceae Blautia Blautia coccoides YL58 + Clostridium Firmicutes Clostridiales Lachnospiraceae Clostridium_g24 YL32 + clostridioforme Clostridium Firmicutes Erysipelotrichales Erysipelotrichaceae Longicatena I46 + innocuum Enterococcus Firmicutes Lactobacillales Enterococcaceae Enterococcus KB1 + faecalis Flavonifractor Firmicutes Clostridiales Ruminococcaceae Pseudoflavonifractor YL31 + plautii

147

Lactobacillus Firmicutes Lactobacillales Lactobacillaceae Lactobacillus I49 + reuteri Muribaculum Bacteroidetes Bacteroidales Muribaculaceae Muribaculum YL27 Ø intesinale Proteobacteria Burkholderiales Sutterellaceae Turicimonas Turicimonas muris YL45 Ø

Table 8: Taxonomic assignment of OTUs from sort purified oligoMM12 feces. OTU Phylum Species/Strain OTU_3 Verrucomicrobiales Akkermansia muciniphila- YL44 OTU_2 Bacteroidales Bacteroides caecimuris- I48 OTU_6 Bacteroidales Muribaculum intestinale- YL27 OTU_4 Firmicutes Blautia coccoides- YL58 OTU_1 Firmicutes Clostridium clostridiforme- YL32 OTU_10 Firmicutes Clostridium innocuum- I46 OTU_8 Firmicutes Flavonifractor plautii- YL31 OTU_7 Firmicutes Lactobacillus reuteri- I49 OTU_5 Proteobacteria Turicimonas muris- YL45

148

9.3 mAb and germ-line sequences Table 9: IGH and IGL sequences of HD and CD derived mutated mAbs and their reverted germ- line (GL) configuration (restriction sites are annotated in blue and orange). HD mAbs: ------HD2a7 (H_chain) BLAST Query: CAGGTGCAGCTGCAGGAGTCGGGCCCAGGCATGCTAAAGTCTTCGGAGACCCTGTCCCTCACTTGCACTGTCACGGGA GGCCCCATCACTAGGCACTACTGGGCCTGGATTTGCCAGGCCCCCGGGAAGGAACCTGAGTGGATTGGATACATCTCC TCTGATGGGGCCGCCAAGTCCAGTCCTTCCCTCAGGGATCGAGTCGCCTTGTCATTAGACACGTCCAAGAACCAGGTG TCTTTGACGGTGAATTCTGTGACCGCGGCGGACACGGCCATGTACTTCTGTGCGACGGCCCCTTACTTAGCGACAGCT GACACCCTCTGGTTTGGCACCTGGGGCCAGGGGATCACGGTCACCGTCTCCTCAG(367bp) HD2a7_H_chain_GL CTGATAACCGGTGTACATTCCCAGGTGCAGCTGCAGGAGTCGGGCCCAGGACTGGTGAAGCCTTCGGAGACCCTGTCC CTCACCTGCACTGTCTCTGGTGGCTCCATCAGTAGTTACTACTGGAGCTGGATCCGGCAGCCCCCAGGGAAGGGACTG GAGTGGATTGGGTATATCTATTACAGTGGGAGCACCAACTACAACCCCTCCCTCAAGAGTCGAGTCACCATATCAGTAG ACACGTCCAAGAACCAGTTCTCCCTGAAGCTGAGCTCTGTGACCGCTGCGGACACGGCCGTGTATTACTGTGCGACGG CCCCTTACTTAGCGACAGCTGACACCCTCTGGTTTGGCACCTGGGGCCAGGGGATCACGGTCACCGTCTCCTCAGCGT CGACTGAATGTAGA HD2a7 (L_chain; κ) BLAST Query: TGACCCAGTCTCCAGACACCCTGTCTGTGTCTCCAGGGCAGAGAGTCACTCTCTCCTGCAGGGCCAGCGGGACTGTTC GGAGAACCTATGTAGCCTGGTACCAACTTTCACCTGGCCGGCCTCCCAGACTCCTCCTTTATGGTGTGTCCAGTAGGGC CGTTGGCGTCCCAGACAGGTTCAGTGGCAGTGGGTCTGGGACCTACTTCACTCTCACCATTAACCCACTGGAGCCTGA AGATTTTGGAATGTATTATTGTCAGCAATATGGTGCGTCACCGGTCACTTTCGGCGGGGGGACCAAGGTGGAGATCAAA C(315bp) HD2a7_L_chain_GL TCGATTGAATTCCACCATGGGATGGTCATGTATCATCCTTTTTCTAGTAGCAACTGCAACCGGTGTACATTGTGCCATCC GGATGACGCAGTCTCCAGGCACCCTGTCTTTGTCTCCAGGGGAAAGAGCCACCCTCTCCTGCAGGGCCAGTCAGAGTG TTAGCAGCAGCTACTTAGCCTGGTACCAGCAGAAACCTGGCCAGGCTCCCAGGCTCCTCATCTATGGTGCATCCAGCA GGGCCACTGGCATCCCAGACAGGTTCAGTGGCAGTGGGTCTGGGACAGACTTCACTCTCACCATCAGCAGACTGGAGC CTGAAGATTTTGCAGTGTATTACTGTCAGCAGTATGGTAGCTCACCGGTCACTTTCGGCGGGGGGACCAAGGTGGAGAT CAAACGTACGGTGGCT ------HD2a88 (H_chain) BLAST Query: CAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGGGGGTCCCTGAGACTCTCCTGTGCAGCGTCTGG ATTCATTTTCAAAGACTATGGCATCCACTGGGTGCGCCAGGCTCCAGGCAAGGGGCTCGATTGGGTGACATATATATGG CGTGATGGAAATAATAAATTCTATACAGACTCCGTGAAGGGCCGATTCACTGTCTCCAGAGACAACTCCAGGAACACGTT GTATCTGCAAATGAACAGCCTGAGACCTGAGGACACGGCTGTGTATTATTGCGCGAAAGCAGGGGCCACTGGCTTGGA CGTCTGGGGCAGAGGGACCACGGTCACCGTCTCCTCAG(352bp) HD2a88_H_chain_GL CTGATAACCGGTGTACATTCTCAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGGGGGTCCCTGAG ACTCTCCTGTGCAGCGTCTGGATTCACCTTCAGTAGCTATGGCATGCACTGGGTCCGCCAGGCTCCAGGCAAGGGGCT GGAGTGGGTGGCATTTATACGGTATGATGGAAGTAATAAATACTATGCAGACTCCGTGAAGGGCCGATTCACCATCTCC AGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCTGAGGACACGGCTGTGTATTACTGTGCGA AAGCAGGGGCCACTGGCTTGGACGTCTGGGGCAGAGGGACCACGGTCACCGTCTCCTCAGCGTCGACTGAATGTAGA HD2a88 (L_chain; λ) BLAST Query: TGCTGACTCAGCCGCCCTCAGTGTCTGCGGCCCCAGGACAGAAGGTCACCATCTCCTGCTCTGGAAGCACCTCCAACA TTGGGAATAATGATGTCACCTGGTACCAGCAGCTCCCAGGAACAGCCCCCAAATTCCTCATTTATGAAAATAATAAGCGA CCCTCAGGGATTCCTGACCGATTCTCTGGCTCCAAGTCTGGCACGTCAGCCACCCTGGACATCACCGGACTCCAGACT GGGGACGAGGCCGATTATTACTGCGGAACATGGGATAACAGTATGAGTGCTGAAGTGTTCGGCGGAGGGACCAAGGTG ACCGTCCTAG(324bp) HD2a88_L_chain_GL CTGATAACCGGTTCTTGGGCCAATTTTATGTTGACGCAGCCGCCCTCAGTGTCTGCGGCCCCAGGACAGAAGGTCACC ATCTCCTGCTCTGGAAGCAGCTCCAACATTGGGAATAATTATGTATCCTGGTACCAGCAGCTCCCAGGAACAGCCCCCA AACTCCTCATCTATGAAAATAATAAGCGACCCTCAGGGATTCCTGACCGATTCTCTGGCTCCAAGTCTGGCACGTCAGC CACCCTGGGCATCACCGGACTCCAGACTGGGGACGAGGCCGATTATTACTGCGGAACATGGGATAGCAGCCTGAGTGC TGAAGTGTTCGGCGGAGGGACCAAGGTGACCGTCCTAGGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCCACC CTCGAGTGAATGTAGAGCCG HD3a14 (H_chain) BLAST Query: CAGGTCCAGCTGGTACAGTCTGGGGCTGAGGTGAAGGAGCCTGGGGCCTCAGTGAAGGTCTCCTGCAAGGCTTCTGG ATACACCTTCACCAATTATGATATCAACTGGGTGCGACAGGCCACTGGACAAGGGCTTGAGTGGATGGGATGGATGAAC CCTAACAGTGCTAACACAGGCTATGCACAGAGGTTCCAGGGCAGAGTCACCATGACCAGAGACACCTCCATAAGTACAG CCTACATGGAGCTGAGCAGCCTGAGATCTGAGGACACGGCCGTGTATTACTGTACGAGAAGCGGGTTTGGGGCACCTG TGAACTTTGACCATTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCA(360bp) HD3a14_H_chain_GL CTGATAACCGGTGTACATTCCCAGGTGCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAAG GTCTCCTGCAAGGCTTCTGGATACACCTTCACCAGTTATGATATCAACTGGGTGCGACAGGCCACTGGACAAGGGCTTG AGTGGATGGGATGGATGAACCCTAACAGTGGTAACACAGGCTATGCACAGAAGTTCCAGGGCAGAGTCACCATGACCA GGAACACCTCCATAAGCACAGCCTACATGGAGCTGAGCAGCCTGAGATCTGAGGACACGGCCGTGTATTACTGTGCGA GAAGCGGGTTTGGGGCACCTGTGAACTTTGACCATTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGTCGACTG AATGTAGA 149

HD3a14 (L_chain; κ) BLAST Query: GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAAGGTCACCATCACTTGCCGGACAAGTC AGAGCATTAACAACTATTTGAATTGGTATCAGCAGAAACCAGGGAAAGCCCCCAAACTCCTGATCTATACTGCATCCAGT TTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATCTGGGACAGATTTCACTCTCACCATCAGCAGTCTGCAAC CTGAAGATCTTGCAACTTACTACTGTCAACAGAGTTTCGAAACCCGGGCGTTCGGCCCAGGGACCAAGGTGGAAATCAA AC(319bp) HD3a14_L_chain_GL CTGATAACCGGTGTACATTCTGACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAGAGTCA CCATCACTTGCCGGGCAAGTCAGAGCATTAGCAGCTATTTAAATTGGTATCAGCAGAAACCAGGGAAAGCCCCTAAGCT CCTGATCTATGCTGCATCCAGTTTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATCTGGGACAGATTTCACT CTCACCATCAGCAGTCTGCAACCTGAAGATTTTGCAACTTACTACTGTCAACAGAGTTACAGTACCCGGGCGTTCGGCC CAGGGACCAAGGTGGAAATCAAACGTACGTGAATGTAGAGC ------HD3a75(H_chain) BLAST Query: CAGGTCCAGCTGGTACAGTCTGGGGCTGAGGTGAAGGAGCCTGGGGCCTCAGTGAAGGTCTCCTGCAAGGCTTCTGG ATACACCTTCACCAATTATGATATCAACTGGGTGCGACAGGCCACTGGACAAGGGCTTGAGTGGATGGGATGGATGAAC CCTAACAGTGCTAACACAGGCTATGCACAGGGGTTCCAGGGCAGAGTCACCATGACCAGAGACACCTCCATAAGTACA GCCTACATGGAGCTGAGCAGCCTGAGATCTGAGGACACGGCCGTGTATTACTGTACGAGAAGCGGGTTTGGGGCACCT GTGAACTTTGACCATTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCA(360bp) HD3a75_H_chain_GL CTGATAACCGGTGTACATTCCCAGGTGCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAAG GTCTCCTGCAAGGCTTCTGGATACACCTTCACCAGTTATGATATCAACTGGGTGCGACAGGCCACTGGACAAGGGCTTG AGTGGATGGGATGGATGAACCCTAACAGTGGTAACACAGGCTATGCACAGAAGTTCCAGGGCAGAGTCACCATGACCA GGAACACCTCCATAAGCACAGCCTACATGGAGCTGAGCAGCCTGAGATCTGAGGACACGGCCGTGTATTACTGTGCGA GAAGCGGGTTTGGGGCACCTGTGAACTTTGACCATTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGTCGACTG AATGTAGA HD3a75 (L_chain; κ) BLAST Query: GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAAGGTCACCATCACTTGCCGGACAAGTC AGAGCATTAACAACTATTTGAATTGGTATCAGCAGAAACCAGGGAAAGCCCCCAAACTCCTGATCTATACTGCATCCAGT TTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATCTGGGACAGATTTCACTCTCACCATCAGCAGTCTGCAAC CTGAAGATCTTGCAACTTACTACTGTCAACAGAGTTTCGAAACCCGGGCGTTCGGCCCAGGGACCAAGGTGGAAATCAA AC(319bp) HD3a75_L_chain_GL CTGATAACCGGTGTACATTCTGACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAGAGTCA CCATCACTTGCCGGGCAAGTCAGAGCATTAGCAGCTATTTAAATTGGTATCAGCAGAAACCAGGGAAAGCCCCTAAGCT CCTGATCTATGCTGCATCCAGTTTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATCTGGGACAGATTTCACT CTCACCATCAGCAGTCTGCAACCTGAAGATTTTGCAACTTACTACTGTCAACAGAGTTACAGTACCCGGGCGTTCGGCC CAGGGACCAAGGTGGAAATCAAACGTACGTGAATGTAGAGC ------HD3a103 (H_chain) BLAST Query: GAAGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTGAAGCCTGGGGGGTCCCTTAGACTCTCCTGTGAAGTCTCTGG ATTAATTTTCAGTGACGCCTGGGTGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTTGGCCGTATTAA AAGTAAAGGTTCTGGTGGGGCAATAGACTACGCTGCACCCGTGAGAGGCAGATTCACCATCTCAAGAGATGATTCAAAA AGCACGCTGTTCCTGCAAATGGACAGCCTGAAAACCGAGGACACAGCCATGTATTACTGTGCCCACCACCCCGGCGGA TCTTGTCCTGGTGCCACCTGTCTTCGCTCCTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAG(376bp) HD3a103_H_chain_GL CTGATAACCGGTGTACATTCTGAGGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTAAAGCCTGGGGGGTCCCTTAGA CTCTCCTGTGCAGCCTCTGGATTCACTTTCAGTAACGCCTGGATGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCTG GAGTGGGTTGGCCGTATTAAAAGCAAAACTGATGGTGGGACAACAGACTACGCTGCACCCGTGAAAGGCAGATTCACC ATCTCAAGAGATGATTCAAAAAACACGCTGTATCTGCAAATGAACAGCCTGAAAACCGAGGACACAGCCGTGTATTACTG TACCCACCACCCCGGCGGATCTTGTCCTGGTGCCACCTGTCTTCGCTCCTGGGGCCAGGGAACCCTGGTCACCGTCTC CTCAGCGTCGACTGAATGTAGA HD3a103 (L_chain; κ) BLAST Query: GAAATTGTGTTGACACAGTCTCCAGGCACCCTGTCTTTGTCTCCAGGGGAAAGAGCCACCCTCTCCTGCAGGGCCAGTC AGAGTGTGAGCGGCGACTACTTAACCTGGTACCAGCAGAAACCTGGCCAGGCTCCCAGGCTCCTCATCTATGGTACATA CACCAGGGCCACTGGCATCCCAGACAGGTTCAGTGGCAGTGGGTCTGGGACAGACTTCACTCTCACCATCAGCAGACT GGAGCCTGAAGATTTTGCAGTGTATTACTGTCAGCACTATGGTAGCTCACCTCCCCAACTCACTTTCGGCCCTGGGACC AAAGTGGATATCAAAC(331bp) HD3a103_L_chain_GL CTGATAACCGGTGTACATTCAGAAATTGTGTTGACGCAGTCTCCAGGCACCCTGTCTTTGTCTCCAGGGGAAAGAGCCA CCCTCTCCTGCAGGGCCAGTCAGAGTGTTAGCAGCAGCTACTTAGCCTGGTACCAGCAGAAACCTGGCCAGGCTCCCA GGCTCCTCATCTATGGTGCATCCAGCAGGGCCACTGGCATCCCAGACAGGTTCAGTGGCAGTGGGTCTGGGACAGACT TCACTCTCACCATCAGCAGACTGGAGCCTGAAGATTTTGCAGTGTATTACTGTCAGCAGTATGGTAGCTCACCTCCCCAA CTCACTTTCGGCCCTGGGACCAAAGTGGATATCAAACGTACGTGAATGTAGAGC ------HD3a147 (H_chain) BLAST Query: GAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTGCAGCCTGGGGGGTCCCTGAGACTCTCCTGTGCAGCCTCTGG ATTCACCTTTAGTAGTTATGCCGCGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTCTCCGGTATTAG TGGTAGTGGTACTAGCACCTACTACGCAGACTCCGTGAGGGGCCGGTTCACCATCTCCAGAGACATTTCCAAGAACACG TTGTATCTGCAAATGAACAGCCTGAGAGACGAGGACACGGCCATATATTACTGTGCGAAAGATCTGTGGAGGCGCGATG GCGGGGGGCTTGGGCCGTTTGACTACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAG(373bp) 150

HD3a147_H_chain_GL ACTGCAACCGGTGTACATTCTGAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAG ACTCTCCTGTGCAGCCTCTGGATTCACCTTTAGCAGTTATGCCATGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCT GGAGTGGGTCTCAGCTATTAGTGGTAGTGGTGGTAGCACATACTACGCAGACTCCGTGAAGGGCCGGTTCACCATCTC CAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCCGAGGACACGGCCGTATATTACTGTGCG AAAGATCTGTGGAGGCGCGATGGCGGGGGGCTTGGGCCGTTTGACTACTGGGGCCAGGGAACCCTGGTCACCGTCTC CTCAGCGTCGACCAAGGG HD3a147 (L_chain; λ) BLAST Query: CTGACTCAGCCTGCCTCCGTGTCTGGGTCTCCTGGACAGTCGATCACCATCTCCTGCACTGGAGCCAGCAGTGACGTT GTTTATTATAACTATGTCTCCTGGTACCAACAGCACCCAGGCAAAGCCCCCAAACTCATGATTTATGATGTCAGTGCTCG GCCCTCAGGGGTTTCTCATCGCTTCTCTGGCTCCAAGTCTGGCAACACGGCCTCCCTGACCATCTCTGGGCTCCAGGC TGAGGACGAGGCTGATTATTACTGTATGTCATCTACAAGCAGTTCCACTGTGCTATTCGGCGGAGGGACCAAGGTGACC GTCCTAG(322bp) HD3a147_L_chain_GL ACTGCAACCGGTTCTTGGGCCAATTTTATGCTGACTCAGCCTGCCTCCGTGTCTGGGTCTCCTGGACAGTCGATCACCA TCTCCTGCACTGGAACCAGCAGTGACGTTGGTGGTTATAACTATGTCTCCTGGTACCAACAGCACCCAGGCAAAGCCCC CAAACTCATGATTTATGAGGTCAGTAATCGGCCCTCAGGGGTTTCTAATCGCTTCTCTGGCTCCAAGTCTGGCAACACG GCCTCCCTGACCATCTCTGGGCTCCAGGCTGAGGACGAGGCTGATTATTACTGCAGCTCATATACAAGCAGCAGCACTG TGCTATTCGGCGGAGGGACCAAGGTGACCGTCCTAGGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCCACCCT CGAGTGAGGA ------CD mAbs: ------CD1a293 (H_chain) BLAST Query: TGCAGCTGGTGGAGTCGGGCCCAGGACTGGTGAAGCCTTCGGAGACCCTGTCCCTCACCTGCAGTGTCTCTGGTGGCT CCATCGGCACTAGTACTACCCTCTGGGATTGGATCCGCCAGACCCCAGGGAAGGGGCTGGAGTGGATTGGGAGTATCT CCTATAGTGGGAGCGCCTACTACAACCCGTCCCTCAAAAGTCGAGTCACCATATCCGCTGACACGTCCAAGAATTACTT CTCCCTGAGGTTGAGTTCTGTGACCGCCGCAGACACGGCTGTGTATTACTGTGCGAGACATCGGTACGAAAGTAGTCGT TATCACTTTGACTACTGGGGCCAGGGAACCACGGTCACCGTTTCCTCAG(363bp) CD1293_H_chain_GL ACTGCAACCGGTGTACATTCTGAAGTGCAGCTGCAGGAGTCGGGCCCAGGACTGGTGAAGCCTTCGGAGACCCTGTCC CTCACCTGCACTGTCTCTGGTGGCTCCATCAGCAGTAGTAGTTACTACTGGGGCTGGATCCGCCAGCCCCCAGGGAAG GGGCTGGAGTGGATTGGGAGTATCTATTATAGTGGGAGCACCTACTACAACCCGTCCCTCAAGAGTCGAGTCACCATAT CCGTAGACACGTCCAAGAACCACTTCTCCCTGAAGCTGAGCTCTGTGACCGCCGCAGACACGGCTGTGTATTACTGTGC GAGACATCGGTACGAAAGTAGTGGTTATCACTTTGACTACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGTC GACCAAGGG CD1a293 (L_chain; κ) BLAST Query: CATCCGGATGACCCAGTCTCCAGACTCCCTGGCAGTGTCTCTGGGCGAGAGGGCCACCATCAACTGCAAGTCCAGCCA GAGTGTTTTATTGAGCTCCAACAATAAGAACCTCTTAGCTTGGTACCAGCAGAAACCAGGACAGCCTCCTAAGGTGCTCA TTTACTGGGCATCTACCCGGGAATCCGGAGTCCCTGACCGATTCAGTGGCAGCGGGTCTGGGGCAGATTTCACTCTCA CCATCAGCAGCCTGCAGGCTGAAGATGTGGCACTTTATTACTGTCAGCAATATTATAGTCCTCCTCTCACTTTCGGCGGA GGGACCAAGGTGGAAATCAAAC(338bp) CD1a293_L_chain_GL ACTGCAACCGGTGTACATTCTGCCATCGTGATGACCCAGTCTCCAGACTCCCTGGCTGTGTCTCTGGGCGAGAGGGCC ACCATCAACTGCAAGTCCAGCCAGAGTGTTTTATACAGCTCCAACAATAAGAACTACTTAGCTTGGTACCAGCAGAAACC AGGACAGCCTCCTAAGCTGCTCATTTACTGGGCATCTACCCGGGAATCCGGGGTCCCTGACCGATTCAGTGGCAGCGG GTCTGGGACAGATTTCACTCTCACCATCAGCAGCCTGCAGGCTGAAGATGTGGCAGTTTATTACTGTCAGCAATATTATA GTACTCCTCTCACTTTCGGCGGAGGGACCAAGGTGGAGATCAAACGTACGGTGGCT ------CD2a61 (H_chain) BLAST Query: CAGGTGCAGCTGGTGGAGTCGGGGGGAGGCGTGATCCAGCCTGGGACGTCCCTGAGACTCTCCTGTGCAGCCTCTGG ATTCGTCTTCAGTGACTATGCCATGCACTGGGTCCGCCAGGTTCCAGGCAAGGGGCTGGAGTGGGTGGCATTTATTTCA TATGATGGAACTAGACAACAATATGCAGACTCCGTGAAGGGCCGATTCATCATCTCCAGAGCCACTTCCACGGGCACGG TTCATCTACAGATGAACAGCCTCAGAGGTGAAGACACGGCTGTTTATTACTGTATCAAAGAAGGCGAATATAGTAGGACC TGGTACCCATTTGAGTACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAG(367bp) CD2a61_H_chain_GL ACTGCAACCGGTGTACATTCTCAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGGAGGTCCCTGAG ACTCTCCTGTGCAGCCTCTGGATTCACCTTCAGTAGCTATGGCATGCACTGGGTCCGCCAGGCTCCAGGCAAGGGGCT GGAGTGGGTGGCAGTTATATCATATGATGGAAGTAATAAATACTATGCAGACTCCGTGAAGGGCCGATTCACCATCTCC AGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCTGAGGACACGGCTGTGTATTACTGTGCGA AAGAAGGCGAATATAGTAGGACCTGGTACCCATTTGACTACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGT CGACCAAGGG CD2a61 (L_chain; λ) BLAST Query: CAGTCTGTGCTGACTCAGCCGCCCTCAGTGTCTGGGGCCCCAGGGCAGAGGGTCACCATCTCCTGCTCTGGGAGCAG CTCCAACATCGGGGCAGATTATCATGTACACTGGTACCTCCAGGTTCCAGGAACAGCCCCCAAACTCCTCATCTATGGT TCCAACAATCGGCCCTCAGGGGTCCCTGACCGATTCTCTGGCTCCAAGTCTGGCACCTCAGCCTCCCTGGCCATCACT GGGCTCCAGGCTGAGGATGAGGCTGATTATTACTGCCAGTCCTATGACATCAGTCTGGATGCCTGGGTGTTCGGCGGA GGGACCAAGTTGACCGTCCTA(333bp) CD2a61_ L_chain_GL ACTGCAACCGGTTCCTGGGCCCAGTCTGTGCTGACGCAGCCGCCCTCAGTGTCTGGGGCCCCAGGGCAGAGGGTCAC CATCTCCTGCACTGGGAGCAGCTCCAACATCGGGGCAGGTTATGATGTACACTGGTACCAGCAGCTTCCAGGAACAGC CCCCAAACTCCTCATCTATGGTAACAGCAATCGGCCCTCAGGGGTCCCTGACCGATTCTCTGGCTCCAAGTCTGGCACC 151

TCAGCCTCCCTGGCCATCACTGGGCTCCAGGCTGAGGATGAGGCTGATTATTACTGCCAGTCCTATGACAGCAGCCTG GATGCCTGGGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTAAGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTC CCACCCTCGAGTGAGGA ------CD2a70 (H_chain) BLAST Query: CAGGTCCAGCTGGTACAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAAGGTCTCCTGCAAAGTTTCTGGA TACACCTTCAAGGACACTTTTGTGCACTGGCTGCGACGGGCCCCTGGGCAAGGACTTGAGTGGATGGGACAAATCTAC ACTAAAGATGGTGTCACATACTATGCACGGAAGTTTCAGGGCAGGGTCACCGTGACCAGGGACACGTCCATCACCACAT CCTACATGGAGCTGAGCAGCCTGACATCTGACGACACGGCCGTCTACTATTGTGCGAGAGAGTCTTTTACTAACACCTG GACGAAGGCTTTTGATCTTTGGGGCCAAGGGACAATGGTCACCGTCTCTTCAG(367bp) CD2a70 _H_chain_GL ACTGCAACCGGTGTACATTCCCAGGTGCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAA GGTCTCCTGCAAGGCTTCTGGATACACCTTCACCGGCTACTATATGCACTGGGTGCGACAGGCCCCTGGACAAGGGCT TGAGTGGATGGGATGGATCAACCCTAACAGTGGTGGCACAAACTATGCACAGAAGTTTCAGGGCAGGGTCACCATGAC CAGGGACACGTCCATCAGCACAGCCTACATGGAGCTGAGCAGGCTGAGATCTGACGACACGGTCGTGTATTACTGTGC GAGAGAGTCTTTTACTAACACCTGGACGAAGGCTTTTGATGTCTGGGGCCAAGGGACAATGGTCACCGTCTCTTCAGCG TCGACCAAGGG CD2a70 (L_chain; λ) BLAST Query: CTGACTCAGCCTCCCTCCGCGTCCGGGTCTCCTGGACAGTCAGTCACCATCTCCTGCACTGGAACCAGCGGTGACGTT GGTGGTTATAACTATGTCTCCTGGTACCAACAACACCCAGGCAAAGCCCCCAAACTCATAATTTATGAGGTCACTAAGCG GCCCTCGGGGGTCCCTGATCGCTTCTCTGGCTCCAAGTCTGGCAACACGGCCTCCCTGACCGTCTCTGGGCTCCAGGC TGTGGATGAGGCTGATTATTTCTGCAGCTCCTCTGCAGGCGACAACCCTTATGTCTTCGGAACTGGTACCAAGGTCACC GTCCT(320bp) CD2a70_L_chain_GL ACTGCAACCGGTTCTTGGGCCAATTTTATGCTGACTCAGCCTCCCTCCGCGTCCGGGTCTCCTGGACAGTCAGTCACCA TCTCCTGCACTGGAACCAGCAGTGACGTTGGTGGTTATAACTATGTCTCCTGGTACCAACAGCACCCAGGCAAAGCCCC CAAACTCATGATTTATGAGGTCAGTAAGCGGCCCTCAGGGGTCCCTGATCGCTTCTCTGGCTCCAAGTCTGGCAACACG GCCTCCCTGACCGTCTCTGGGCTCCAGGCTGAGGATGAGGCTGATTATTACTGCAGCTCATATGCAGGCAGCAACCCTT ATGTCTTCGGAACTGGGACCAAGGTCACCGTCCTGAGTCAGCCCAAGGCCAACCCCACTGTCACTCTGTTCCCACCCTC GAGTGAGGA CD2a127 (H_chain) BLAST Query: GAAGTGCAGCTGGTGGAGTCTGGGGGAGGCCTGGTCAAGCCTGGGGGGTCCCTGAGACTCTCCTGTGCAGGCTCTGG ATTCATCTTCAGTTCTTATAGCATGAACTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTCTCTTCCATTGAT ACTACTAGTGCTTTCATATTCTACGCAGACTCAATGAAGGGCCGCTTCACCATCTCCAGAGACAACGCCAAGAACTCACT GTATCTGCAAATGGACAGCCTGAGAGCCGAGGACACGGCTGTGTATTACTGTGCGAGATCGGTCCCCTCCGATGCTTTT GATATCTGGGGCCAAGGGACAATGGTCACCGTCTCTTCAG(355bp) CD2a127_H_chain_GL ACTGCAACCGGTGTACATTCTGAGGTGCAGCTGGTGGAGTCTGGGGGAGGCCTGGTCAAGCCTGGGGGGTCCCTGAG ACTCTCCTGTGCAGCCTCTGGATTCACCTTCAGTAGCTATAGCATGAACTGGGTCCGCCAGGCTCCAGGGAAGGGGCT GGAGTGGGTCTCATCCATTAGTAGTAGTAGTAGTTACATATACTACGCAGACTCAGTGAAGGGCCGATTCACCATCTCCA GAGACAACGCCAAGAACTCACTGTATCTGCAAATGAACAGCCTGAGAGCCGAGGACACGGCTGTGTATTACTGTGCGA GATCGGTCCCCTCCGATGCTTTTGATGTCTGGGGCCAAGGGACAATGGTCACCGTCTCTTCAGCGTCGACCAAGGG CD2a127 (L_chain; κ) BLAST Query: GAAATAGTGATGACGCAGTCTCCAGCCACCCTGTCTGTGTCCCCAGGGGAAAGAGTCACCCTCTCCTGCAGGGCCAGT CAGAGCGTTAGCAGCAACTTAGCCTGGTACCAACAGAAACCTGGACAGGCTCCCAGGCTCCTCATCCGTGGTGCATCC ACCAGGGCCACTGGTATCCCGGCCAGGTTCAGTGGCAGTGGGTCTGGGACAGAGTTCACTCTCACCATCAGCAGCCTC CAATCTGAAGATGTTGCAGTTTATTACTGTCAGCAGTGTATTAAGTGGCCTAAGACGTTCGGCCAAGGGACCAAGGTGG AAATCAAAC(322bp) CD2a127_L_chain_GL ACTGCAACCGGTGTACATTCAGAAATAGTGATGACGCAGTCTCCAGCCACCCTGTCTGTGTCTCCAGGGGAAAGAGCCA CCCTCTCCTGCAGGGCCAGTCAGAGTGTTAGCAGCAACTTAGCCTGGTACCAGCAGAAACCTGGCCAGGCTCCCAGGC TCCTCATCTATGGTGCATCCACCAGGGCCACTGGTATCCCAGCCAGGTTCAGTGGCAGTGGGTCTGGGACAGAGTTCA CTCTCACCATCAGCAGCCTGCAGTCTGAAGATTTTGCAGTTTATTACTGTCAGCAGTATAATAACTGGCCTAAGACGTTC GGCCAAGGGACCAAGGTGGAGATCAAACGTACGGTGGCT ------CD2a146 (H_chain) BLAST Query: GAAGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTAAATCCTGCGGGGTCCCTTAGACTCTCCTGTGCAGCCTCTGGA TTCAGTTTCAGTGGCGCCTGGTTGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTTGGCCGTATTAAA AGCAAAGTTGATGGTGAAACAACAGACTACGGTGGTCCCGTGAAAGGCAGATTCACCATCTCAAGAGATGATTCAAAAA ACATACTGTATCTGCAAATGAACAGCCTGATAACCGAGGACACAGCCGTGTATTATTGTACCACAGTGCTGCTGTGGGG GGCGGATTGTGGGACCTGCGGTATCCCATACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAG(379bp) CD2a146_H_chain_GL ACTGCAACCGGTGTACATTCTGAGGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTAAAGCCTGGGGGGTCCCTTAGA CTCTCCTGTGCAGCCTCTGGATTCACTTTCAGTAACGCCTGGATGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCTG GAGTGGGTTGGCCGTATTAAAAGCAAAACTGATGGTGGGACAACAGACTACGCTGCACCCGTGAAAGGCAGATTCACC ATCTCAAGAGATGATTCAAAAAACACGCTGTATCTGCAAATGAACAGCCTGAAAACCGAGGACACAGCCGTGTATTACTG TACCACAGTGCTGCTGTGGGGGGCGGATTGTGGGACCTGCGGTATCCCATACTGGGGCCAGGGAACCCTGGTCACCG TCTCCTCAGCGTCGACCAAGGG CD2a146 (L_chain; κ) BLAST Query: GAAATAGTGATGACGCAGTCTCCAGCCACCCTGTCTGTGTCTCCAGGGGAAAGAGCCACCCTCTCCTGCAGGGCCAGT CAGAGTGTTAGCAGCAAATTAGCCTGGTACCAGCAGAAACCTGGGCAGGCTCCCAGGCTCCTCATGTATGGTGCATCC 152

ACCAGGGCCACTGGTATCCCAGCCAGGTTCAGTGGCAGTGGGTCTGGGACAGAATTCACTCTCACCATCAGCAGCCTG CAGTCTGAAGATTTTGCAGTTTATTACTGTCAGCAGTATAGTAACTGGTGGACGTTCGGCCAAGGGACCAAGGTGGAGA TCAAAC(319bp) CD2a146_L_chain_GL ACTGCAACCGGTGTACATTCAGAAATAGTGATGACGCAGTCTCCAGCCACCCTGTCTGTGTCTCCAGGGGAAAGAGCCA CCCTCTCCTGCAGGGCCAGTCAGAGTGTTAGCAGCAACTTAGCCTGGTACCAGCAGAAACCTGGCCAGGCTCCCAGGC TCCTCATCTATGGTGCATCCACCAGGGCCACTGGTATCCCAGCCAGGTTCAGTGGCAGTGGGTCTGGGACAGAGTTCA CTCTCACCATCAGCAGCCTGCAGTCTGAAGATTTTGCAGTTTATTACTGTCAGCAGTATAATAACTGGTGGACGTTCGGC CAAGGGACCAAGGTGGAAATCAAACGTACGGTGGCT ------CD2a148 (H_chain) MUT BLAST Query: CAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGAGGTCCAGCCTGGGACGTCCCTGAGACTCTCCTGTGCAGCCTCTGG ATTCACCTTCAGTGATTTGGCTTTCCACTGGGTCCGCCAGGCTCCAGGCAGGGGGCTCCAGTGGGTGGCATTTATAACA TATGACGGAACGAGGCAACAATATGCAGACTCCGTGAGGGGCCGATTCGCCATCTCCAGAGACGACTCCAAGAAGACG CTATATCTCCTAATGAATAGCCTGAGACCTGAGGACACGGCTGTCTATTACTGTGTGAGAGAGGGAAAATACTATGAAAG TGGTGATTATTTTTACCCTCTTGACTACTGGGGCCAGGGAACAATGGTCACCGTCTCTTCAG(376bp) CD2a148_H_chain_GL ACTGCAACCGGTGTACATTCTCAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGGAGGTCCCTGAG ACTCTCCTGTGCAGCCTCTGGATTCACCTTCAGTAGCTATGCTATGCACTGGGTCCGCCAGGCTCCAGGCAAGGGGCT GGAGTGGGTGGCAGTTATATCATATGATGGAAGTAATAAATACTACGCAGACTCCGTGAAGGGCCGATTCGCCATCTCC AGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCTGAGGACACGGCTGTGTATTACTGTGCGA GAGAGGGAAAATACTATGAAAGTGGTTATTATTTTTACCCTCTTGACTACTGGGGCCAAGGGACAATGGTCACCGTCTCC TCAGCGTCGACCAAGGG CD2a148 (L_chain; λ) BLAST Query: CAGGCTGTGGTGACCCAGCCGCCCTCAGTGTCTGGGGCCCCAGGGCAGAGGGTCACCATCTCCTGCATTGGGAGCAG CTCCAACATCGGGGCAGGCTATGATGTACACTGGTACCAGCACCATCCAGGAACAGCCCCCAAACTCCTCATCTTTGGT AACACCAATCGGCCCTCAGGGGTCCCTGACCGATTCTCTGGCTCCAACTCCAAGGCTGGCACCTCAGCCTCCCTGGCC ATCACTGGGCTCCGGGCTGAGGATGAGGCTGATTATTACTGCCAGTCCTATGACAGCAGCCTGGGTGGTTGGGTGTTC GGCGGAGGGACCAAGCTGACCGTCCTA(339bp) CD2a148_L_chain_GL ACTGCAACCGGTTCCAATTCTCAGTCTGTGCTGACGCAGCCGCCCTCAGTGTCTGGGGCCCCAGGGCAGAGGGTCACC ATCTCCTGCACTGGGAGCAGCTCCAACATCGGGGCAGGTTATGATGTACACTGGTACCAGCAGCTTCCAGGAACAGCC CCCAAACTCCTCATCTATGGTAACAGCAATCGGCCCTCAGGGGTCCCTGACCGATTCTCTGGCTCCAACTCCAAGTCTG GCACCTCAGCCTCCCTGGCCATCACTGGGCTCCAGGCTGAGGATGAGGCTGATTATTACTGCCAGTCCTATGACAGCA GCCTGAGTGGTTGGGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTACGTCAGCCCAAGGCTGCCCCCTCGGTCACT CTGTTCCCGCCCTCGAGTGAGGA ------CD3a32 (H_chain) BLAST Query: CAGCTGCAGCTGCAGGAGTCGGGCCCAGGACTGGTGAAGCCTTCGGAGACCCTGTCCCTCGCCTGCAGTGTCTCTGG TGACTCCGTCAGCAGCACCAATTACTACTGGGGTTGGATCCGCCAGCCCCCAGGGAAGACACTGGAGTGGATTGGGGG ATTTCATAACAGAGGGTACACCTACGACAACCCGTCCCTCAAGAGTCGAGTCACCATATCCAAAGACACGTCCAAGAAT CAGTTCTCCCTGAAATTGAACCCTGTGACAGCCGCAGACACGGCTGTGTATTATTGTACGGCTGGCTACGATTGGGCTA AGGCGGGCTACTGGAGCCAGGGTACCCTGGTCACCGTCTCCTCAG(358bp) CD3a32_H_chain_GL ACTGCAACCGGTGTACATTCCCAGCTGCAGCTGCAGGAGTCGGGCCCAGGACTGGTGAAGCCTTCGGAGACCCTGTCC CTCACCTGCACTGTCTCTGGTGGCTCCATCAGCAGTAGTAGTTACTACTGGGGCTGGATCCGCCAGCCCCCAGGGAAG GGGCTGGAGTGGATTGGGAGTATCTATTATAGTGGGAGCACCTACTACAACCCGTCCCTCAAGAGTCGAGTCACCATAT CCGTAGACACGTCCAAGAACCAGTTCTCCCTGAAGCTGAGCTCTGTGACCGCCGCAGACACGGCTGTGTATTACTGTG CGGCTGGCTACGATTGGGCTAAGGCGGGCTACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGTCGACCAAG GG CD3a32 (L_chain; κ) BLAST Query: GACATCCAGATGACCCAGTCTCCTTCCACGCTGTCTGCATCTGTAGGAGACAGAGTCACCATCACTTGTCGGGCCAGTC AGAGTATTGGTAACTGGTTGGCCTGGTATCAGCACAAACCAGGGAAAGCCCCTAAACTCCTGATCTTTAAGTCGTCTAGT TTAGAAAGTGGGGTCCCATCAAGGTTCAGCGGCAGTGGATCTGGGACAGAATTCACTTTCACCATCAGCAGCCTGCAGC CTGATGATTTTGCAACTTATTTCTGCCAACAGTATCATGCTTATCCTTGCACTTTTGGCCAGGGGACCAAGCTGGAGATC AAAC(322bp) CD3a32_L_chain_GL ACTGCAACCGGTGTACATTCTGACATCCAGATGACCCAGTCTCCTTCCACCCTGTCTGCATCTGTAGGAGACAGAGTCA CCATCACTTGCCGGGCCAGTCAGAGTATTAGTAGCTGGTTGGCCTGGTATCAGCAGAAACCAGGGAAAGCCCCTAAGC TCCTGATCTATAAGGCGTCTAGTTTAGAAAGTGGGGTCCCATCAAGGTTCAGCGGCAGTGGATCTGGGACAGAATTCAC TCTCACCATCAGCAGCCTGCAGCCTGATGATTTTGCAACTTATTACTGCCAACAGTATAATAGTTATTCTTGCAGTTTTGG CCAGGGGACCAAGCTGGAGATCAAACCGTACGGTGGCT ------CD3a549 (H_chain) CAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGAGAGTCCCTGAGACTCTCCTGTGCAGCGTCTGG ATTCACCTTCGGAAGCTATGGCATGCATTGGGTCCGCCAGGCTCCAGGCAAGGGGCTGGAGTGGGTGGCTTTTATTTG GTATGACGGAAGTAAAAAAGACTATGCAGACTCCATGAAGGGCCGAATCACCATCTCCAGAGACAATTCCAAGAACACG GTGTCTCTGCAAATGAACAGTCTGAGATCCGACGACACGGCTGTGTATTACTGTGCGAAAGACGGGCAGCGGTGGGCC CCGTACGGTTTGGACGTCTGGGGCCAAGGGACCACGGTCACCGTCTCCTCAG(364bp) CD3a549_H_chain_GL ACTGCAACCGGTGTACATTCTCAGGTGCAGCTGGTGGAGTCTGGGGGAGGCGTGGTCCAGCCTGGGGGGTCCCTGAG ACTCTCCTGTGCAGCGTCTGGATTCACCTTCAGTAGCTATGGCATGCACTGGGTCCGCCAGGCTCCAGGCAAGGGGCT GGAGTGGGTGGCATTTATACGGTATGATGGAAGTAATAAATACTATGCAGACTCCGTGAAGGGCCGATTCACCATCTCC 153

AGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCTGAGGACACGGCTGTGTATTACTGTGCGA AAGACGGGCAGCGGTGGGCCCCGTACGGTATGGACGTCTGGGGCCAAGGGACCACGGTCACCGTCTCCTCAGCGTCG ACCAAGGG CD3a549 (L_chain; λ) BLAST Query: CTGACTCAGCCTGCCTCCGTGTCTGGGTCTCCTGGACAGTCGATCACCATCTCCTGCACTGGAACCAGCAGTGACATTG GTGCTTATAACCATGTCTCCTGGTACCAACAACACCCAGGCAAAGCCCCCAATCTCTTGCTTTATGAGGTCAGTAATCGG CCCTCAGGGGTTTCTACTCGCTTCTCTGGCTCCAAGTCTGGCAACACGGCCTCCCTGACCATCTCTGGGCTCCAGGCT GATGACGAGTCTGATTATTACTGCCACTCATTTACAACCAGTGGTGTTCGGGTGTTCGGCGGAGGGACCAAGCTGACCG TCCTA(321bp) CD3a549_L_chain_GL ACTGCAACCGGTTCTGTGACCTCCTATGAGCTGACTCAGCCTGCCTCCGTGTCTGGGTCTCCTGGACAGTCGATCACCA TCTCCTGCACTGGAACCAGCAGTGACGTTGGTGGTTATAACTATGTCTCCTGGTACCAACAGCACCCAGGCAAAGCCCC CAAACTCATGATTTATGAGGTCAGTAATCGGCCCTCAGGGGTTTCTAATCGCTTCTCTGGCTCCAAGTCTGGCAACACG GCCTCCCTGACCATCTCTGGGCTCCAGGCTGAGGACGAGGCTGATTATTACTGCAGCTCATATACAAGCAGTGGTGTTC GGGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTACGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCACCCTC GAGTGAGGA ------CD3a565 (H_chain) BLAST Query: GAGGTGCAGCTGTTGGAGTCTGGTGGAGGCTTTACACAGCCTGGGGGGTCCCTGAGACTCTCCTGTGCAGCCTCTGGA TTCACCTTTAGCGACTATGCCATGAGCTGGGTCCGCCAGGCTCCAGGGGGGGGACTGGAGTGGGTCTCAAGTCTTAGT GGTGGTGGTGATAAAACATTCTACGCAGACTCCGTGAGGGGCCGGTTCTCCATCTCCAGAGACAATTCCAAGAACACAG TTTCTCTGCAGATGAACAGCCTGAGAGTCGAGGACACGGCCGTATATTACTGTGCGGCAGATCGGGAGGGCGATCCCA GTGCTTATTACCCCCTAGGCCGGGGAACCCTGGTCACCGTCTCCTCA G(361bp) CD3a565_H_chain_GL ACTGCAACCGGTGTACATTCTGAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAG ACTCTCCTGTGCAGCCTCTGGATTCACCTTTAGCAGCTATGCCATGAGCTGGGTCCGCCAGGCTCCAGGGAAGGGGCT GGAGTGGGTCTCAGCTATTAGTGGTAGTGGTGGTAGCACATACTACGCAGACTCCGTGAAGGGCCGGTTCACCATCTC CAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCCGAGGACACGGCCGTATATTACTGTGCG AAAGATCGGGAGGGCGATCCCAGTGCTTATTACCCCCTAGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCGTCGACC AAGGG CD3a565 (L_chain; λ) BLAST Query: CAGACTGTGGTGACTCAGCCTCGCTCAGTGTCCGGGTCTCCTGGACAGTCAGTCACCATCTCCTGCAATGGAACCAGC AGTGATGTTGGTGTTTATAACTATGTCTCCTGGTACCAACAGCACCCGGGCAAAGCCCCCACAGTCATGATTTATGATGT CAATCAGCGGCCCTCAGGGGTCCCTGATCGCTTCTCTGGCTCCAAGTCTGGCAAGACGGCCTCCCTGACCATCTCTGG GCTCCAGGGTGACGATGAGGCTGATTATTATTGCTGCTCATATGCAGGCAGTTACCGTTTTGTCTTCGGAACTGGGACC AAGGTCACCGTCCTAG(331bp) CD3a565_L_chain_GL ACTGCAACCGGTTCCAATTCTCAGTCTGCCCTGACTCAGCCTCGCTCAGTGTCCGGGTCTCCTGGACAGTCAGTCACCA TCTCCTGCACTGGAACCAGCAGTGATGTTGGTGGTTATAACTATGTCTCCTGGTACCAACAGCACCCAGGCAAAGCCCC CAAACTCATGATTTATGATGTCAGTAAGCGGCCCTCAGGGGTCCCTGATCGCTTCTCTGGCTCCAAGTCTGGCAACACG GCCTCCCTGACCATCTCTGGGCTCCAGGCTGAGGATGAGGCTGATTATTACTGCTGCTCATATGCAGGCAGCTACACTT TTGTCTTCGGAACTGGGACCAAGCTGACCGTCCTAGGTCAGCCCAAGGCCAACCCCACTGTCACTCTGTTCCCACCCTC GAGTGAGGA

154

9.4 List of Figures Figure 1: Distribution of bacterial load and organization of the gut immune system throughout the intestine...... 10 Figure 2: Membrane bound BCR and secreted antibodies...... 12 Figure 3: Molecular structure of antibody molecules and antibody isotypes...... 14 Figure 4: Polymeric immunoglobulin receptor (pIgR) mediated secretion of dimeric IgA (dIgA) into the gut lumen as SIgA...... 16 Figure 5: Effector functions of SIgA in the gut...... 20 Figure 6: Genetic alterations and diversification of antibodies schematically displayed for a human antibody...... 25 Figure 7: Schematic overview of IgA induction in Peyer’s patches...... 29 Figure 8: Surgical procedure of Peyer’s patch 4-OHT micro-injections...... 43 Figure 9: Preparation of small intestinal tissue samples for Cryotome sectioning...... 48 Figure 10: Workflow depicting the isolation of small intestinal lamina propria plasma cells from three healthy donors (HD) and three Crohn’s disease (CD) patients...... 50 Figure 11: Identification of fecal bacteria by flow cytometry...... 52 Figure 12: Microbiota-reactive IgA and IgG mAbs are present in the adult human gut. 54 Figure 13: High microbiota-reactive IgA mAbs consistently bind a major fraction of the intestinal microbiota...... 56 Figure 14: High microbiota-reactive intestinal IgA mAbs show donor-dependent variability of binding capacities to human gut bacteria...... 58 Figure 15: High microbiota reactivity of IgA mAbs is not abrogated by deglycosylation...... 60 Figure 16: Intestinal microbiota-reactive IgA mAbs show broad VH gene usage...... 61 Figure 17: Representative workflow of bacterial DNA isolation and 16S rRNA gene amplicon generation...... 62 Figure 18: Workflow illustrating the analysis of 16S rRNA sequencing data...... 63 Figure 19: Intestinal high microbiota-reactive IgA mAbs bind diverse groups of gut bacteria...... 66 Figure 20: High microbiota-reactive intestinal IgA is cross-species reactive...... 70 Figure 21: High microbiota-reactive mAbs are not enriched among polyreactive antibodies...... 70 Figure 22: High microbiota reactivity of intestinal IgA does not correlate with polyreactivity...... 72 Figure 23: High microbiota-reactive intestinal IgA+ plasma cells have high numbers of somatic mutations...... 74 Figure 24: Intestinal IgG+ plasma cells carry high numbers of somatic mutations...... 75 Figure 25: Schematic depicting the generation of predicted mAb germ-line variants from their mutated counterparts...... 76 Figure 26: Somatic mutations confer high microbiota-binding capacity of intestinal IgA mAbs...... 78 Figure 27: Germ-line reversion does not lead to polyreactivity of intestinal IgA mAbs. 80 Figure 28: Somatic mutations contribute to the microbiota binding profile of human intestinal IgA...... 82 Figure 29: Reduced microbiota binding capacity of IgA mAbs to oligoMM12 fecal bacteria...... 84 Figure 30: Decreased binding of IgA mAbs to in vitro cultivated oligoMM12 bacteria...86

155

Figure 31: High microbiota-reactive IgA mAbs show selective binding to oligoMM12 bacteria...... 89 Figure 32: Transgenic mouse model enabling the spatial-temporal tracking of activated B cells and their progeny...... 92 Figure 33: Generation of eYFP+ B cells under steady state conditions...... 94 Figure 34: Histological validation of eYFP expression in Peyer’s patches by confocal microscopy...... 96 Figure 35: Increase of eYFP+ B cell frequencies in PPs and MLN after FTY720 administration...... 98 Figure 36: Single 4-OHT Peyer’s patch injections locally eYFP label B cells...... 101 Figure 37: Double transgenic mouse model enabling the eYFP marking of AID+ B cells and Villin+ epithelial cells...... 102 Figure 38: eYFP expression in B cells and gut epithelial cells after systemic tamoxifen administration...... 103 Figure 39: Lack of unspecific eYFP expression in AIDcreVillcre mice in the absence of tamoxifen...... 107 Figure 40: Spatial distribution of eYFP expression in Peyer’s patches and small intestinal epithelial cells in tamoxifen treated AIDcreVillcre mice...... 109 Figure 41: Single Peyer’s patch injections do not lead to 4-OHT dissemination...... 112 Figure 42: 4-OHT after single Peyer’s patch injections is locally confined...... 113 Figure 43: Phenotypic characterization of eYFP+ cells in Peyer’ patches and in the small intestinal lamina propria...... 117 Figure 44: Plasma cells originating from a single Peyer’s patch persist up to 90 days in the small intestine under homeostatic conditions...... 124 Figure 45: Putative transition of the prevalent mechanism of microbiota-binding from polyreactive IgA in young individuals to affinity-matured, cross-species reactive IgA in adults...... 130 Figure 46: Model for B cell re-entering into pre-existing germinal centers of different Peyer’s patches...... 133

156

9.5 List of Tables Table 1: List of selected HD and CD derived recombinant antibodies with annotated isotypes and cloning restriction sites...... 36 Table 2: List of fluorescently labeled antibodies and dyes used for flow cytometry and cell sorting, listed with specificity, concentration and company...... 38 Table 3: Reagents and PCR program for V3/V4 16S rRNA gene amplification...... 39 Table 4: Fluorescently labeled antibodies and reagents used for flow cytometry, listed with specificity, concentration and company...... 47 Table 5: Efficiency of Fc-Ig mAb deglycosylation by EndoS2 endonuclease...... 59 Table 6: Taxonomic classification of OTUs based on 16S rRNA gene amplicon sequence identity...... 143 Table 7: Bacterial species constituting the oligoMM12 microbial consortium303...... 147 Table 8: Taxonomic assignment of OTUs from sort purified oligoMM12 feces...... 148 Table 9: IGH and IGL sequences of HD and CD derived mutated mAbs and their reverted germ-line (GL) configuration (restriction sites are annotated in blue and orange). .... 149

157

10 Acknowledgements First, I wish to thank my supervisors Prof. Oliver Pabst and Prof. Lars Blank for the supervision of this thesis and the advice throughout. In particular, I would like to express my gratitude to you, Oliver, for giving me the opportunity to work and thrive on this exciting PhD project. I would like to thank you for your immense scientific insight, inspirational ideas and guidance throughout my PhD and not less for your continuous patience for me as a PhD student as “work in progress”. But you also taught me to be autonomous and to deal with all the mistakes that came with it. I am moreover thankful for the encouragement and the many given opportunities to present our data at conferences that most certainly would have been mentally more challenging without your reassurance. At last, your expertise and support led me to accomplish this thesis and we even managed to pull out a paper. Moreover, I would like to thank you, Dr. Vuhuk, for your blunt honesty, critical comments (I mean this in a good way), endless times spent explaining and reading, discussing, reassuring and continuous motivation. (However, I slightly have the feeling that perhaps I “sometimes” posed a challenge to your patience and sanity). Your incredible enthusiasm and general curiosity for science that you so easily manage to share makes wanting to know more about science “contagious”, at least for me. But foremost, I really appreciate that our office college relationship turned into a dear friendship. Thank you Vuk for all your mental, scientific and yes, also emotional support! Also, I would like to thank my “colleague friend” Dr. Annchen for your natural optimism, help with sometimes excruciating long experiments, looking after my mice and your constant support in so many ways that would already fill the page to mention only a few. Annchen, thank you for always finding something positive to say, for cheering me up, laughing, crying and occasionally sipping the odd Sambuca together. You made me look forward to coming to the office every day. I am grateful that I could spent my PhD time with such an awesome person that you are. In addition, I would like to thank Dr. Milas Ugur, from whose scientific knowledge and expertise I greatly profited. Milas, you introduced me to the “doomed to failure” method of Peyer’s patch micro-injections and encouraged me to keep going and get it to work, eventually. Thank you for your enormous patience and great explanations (even though I should have known this), enlightening discussions, your natural thoughtfulness, critique and your friendship. I furthermore wish to thank all my enchanting colleagues, who created an atmosphere that made it so enjoyable working in this group. So, let me say thank you to all of you: Joёl, Lydia, Andrea, Britta, Katharina, Nathalie, Hormoz, Girmay, Denise, Ana, Heike, Necla, Hilde, Sabine and Tina. Thank you all for your support not just in the lab, but also for making days so cheerful and for celebrating my 29th birthday for 4 years in a row . Especially Hilde, Sabine and Tina you truly are the heart of our lab, without your support and experience I would not be where I stand today. You guys were vital and a constant “source” of so much help with my experiments. I am very happy I got to spent my time in Aachen with such unique and lovely people that never hesitated singing along to Helene Fischer in the lab. I will always remember our discussions about, well, basically all topics one can think of, our lively lunch breaks, superbly organized social events, birthdays, Christmas parties (thank you Andrea) and just your nice company for 158 the last (almost) five years. Andrealein, I really enjoyed our “twerking attempts” but most of all, I always appreciated your open ear and encouragement. Joel, despite the odd little argument in the car, thank you for providing the squash rides (“wij komen squashe”), taking us for delicious Dutch “Bierchens” afterwards and just taking our minds off work for a little while. Moreover, I would like to express my appreciation for the great help especially with the 16S taxonomy analysis, but also for the constructive and vital input I received throughout my PhD from Prof. Thomas Clavel and Tom Hitch. Without the two of you my first published manuscript would not have gone so “smoothly”. Janine, although immunology is a “black box” for you, I could always count on your encouragement for my scientific endeavours. Thank you for tolerating my social abstinence and just appreciating the time we do get to spend together. I am grateful that you and your family are part of my life and for our close friendship that is already lasting for over 20 years.

Most importantly, I would like to thank my family, my sister Sophie and my parents Elke and Robert, who substantially contributed to making me into the person that I am today. You fostered my natural curiosity about the world and my enthusiasm for science. I thank you so much for always supporting me and accepting the decisions I made in life even though you sometimes disapproved, but thereby granting me the freedom I appreciate so much in my life. You always encouraged and pushed me to pursue my goals but also when I failed, you were there to comfort me. What I have accomplished thus far, would not have been possible without your overwhelming support, care and sometimes critical assessment. Sophie, you are not only my sister but much more, you are a close friend, a consultant, a mediator and a critic I actually listen to. Thank you for your great attitude, optimism, your encouragement and reassurance. You have the admirable aptitude to always share your unconditional support and love. I am not sure there is anything more powerful than knowing that you will always be there for me. Also, thank you Anna for being an awesome person and making Sophie so happy. Thank you all for being proud of me and making me feel proud of myself. Danke liebe Eltern und Schwesti für alles, ich hab’euch sehr lieb.

Finally, I want to express my gratitude to nature. All the enigmas “you” came up with, providing the science people something to struggle with, but from time to time allowing a tiny glimpse of truth.

159

11 References 1 Mowat, A. M. & Agace, W. W. Regional specialization within the intestinal immune system. Nature reviews. Immunology 14, 667-685, doi:10.1038/nri3738 (2014). 2 McGhee, J. R. & Fujihashi, K. Inside the mucosal immune system. PLoS biology 10, e1001397, doi:10.1371/journal.pbio.1001397 (2012). 3 Kiyono, H. & Azegami, T. The mucosal immune system: From dentistry to vaccine development. Proc Jpn Acad Ser B Phys Biol Sci 91, 423-439, doi:10.2183/pjab.91.423 (2015). 4 Salzman, N. H. et al. Enteric defensins are essential regulators of intestinal microbial ecology. Nature immunology 11, 76-83, doi:10.1038/ni.1825 (2010). 5 Cheroutre, H., Lambolez, F. & Mucida, D. The light and dark sides of intestinal intraepithelial lymphocytes. Nature reviews. Immunology 11, 445-456, doi:10.1038/nri3007 (2011). 6 Mowat, A. M. Anatomical basis of tolerance and immunity to intestinal antigens. Nature reviews. Immunology 3, 331-341, doi:10.1038/nri1057 (2003). 7 N. Aguilera Montilla, M. P. B., M. López Santalla, J.M. Martín Villa. Mucosal immune system: A brief review. 8 Willinger, T. Metabolic Control of Innate Lymphoid Cell Migration. Frontiers in immunology 10, 2010-2010, doi:10.3389/fimmu.2019.02010 (2019). 9 Zheng, D., Liwinski, T. & Elinav, E. Interaction between microbiota and immunity in health and disease. Cell Research 30, 492-506, doi:10.1038/s41422-020-0332-7 (2020). 10 Belkaid, Y. & Harrison, O. J. Homeostatic Immunity and the Microbiota. Immunity 46, 562-576, doi:10.1016/j.immuni.2017.04.008 (2017). 11 Bunker, J. J. et al. Innate and Adaptive Humoral Responses Coat Distinct Commensal Bacteria with Immunoglobulin A. Immunity 43, 541-553, doi:10.1016/j.immuni.2015.08.007 (2015). 12 Bunker, J. J. & Bendelac, A. IgA Responses to Microbiota. Immunity 49, 211-224, doi:10.1016/j.immuni.2018.08.011 (2018). 13 Brandtzaeg, P. The human intestinal immune system: basic cellular and humoral mechanisms. Baillière's Clinical Rheumatology 10, 1-24, doi:https://doi.org/10.1016/S0950-3579(96)80003-3 (1996). 14 Brandtzaeg, P. Role of secretory antibodies in the defence against infections. International Journal of Medical Microbiology 293, 3-15, doi:https://doi.org/10.1078/1438-4221-00241 (2003). 15 Macpherson, A. J., McCoy, K. D., Johansen, F. E. & Brandtzaeg, P. The immune geography of IgA induction and function. Mucosal immunology 1, 11-22, doi:10.1038/mi.2007.6 (2008). 16 Brandtzaeg, P. Induction of secretory immunity and memory at mucosal surfaces. Vaccine 25, 5467-5484, doi:https://doi.org/10.1016/j.vaccine.2006.12.001 (2007). 17 Mestecky, J., Russell, M. W. & Elson, C. O. Intestinal IgA: novel views on its function in the defence of the largest mucosal surface. Gut 44, 2, doi:10.1136/gut.44.1.2 (1999). 18 Macpherson, A. J. et al. A primitive T cell-independent mechanism of intestinal mucosal IgA responses to commensal bacteria. Science 288, 2222-2226, doi:10.1126/science.288.5474.2222 (2000). 19 Smith, K., McCoy, K. D. & Macpherson, A. J. Use of axenic animals in studying the adaptation of mammals to their commensal intestinal microbiota. Semin Immunol 19, 59-69, doi:10.1016/j.smim.2006.10.002 (2007). 20 Bunker, J. J. et al. Natural polyreactive IgA antibodies coat the intestinal microbiota. Science 358, doi:10.1126/science.aan6619 (2017). 21 Hapfelmeier, S. et al. Reversible microbial colonization of germ-free mice reveals the dynamics of IgA immune responses. Science 328, 1705-1709, doi:10.1126/science.1188454 (2010). 22 Gommerman, J. L., Rojas, O. L. & Fritz, J. H. Re-thinking the functions of IgA(+) plasma cells. Gut Microbes 5, 652-662, doi:10.4161/19490976.2014.969977 (2014).

160

23 Brandtzaeg, P. Secretory IgA: Designed for Anti-Microbial Defense. Frontiers in immunology 4, 222, doi:10.3389/fimmu.2013.00222 (2013). 24 Lycke, N. Y. & Bemark, M. The regulation of gut mucosal IgA B-cell responses: recent developments. Mucosal immunology 10, 1361-1374, doi:10.1038/mi.2017.62 (2017). 25 Mestecky, J. & Russell, M. W. Specific antibody activity, glycan heterogeneity and polyreactivity contribute to the protective activity of S-IgA at mucosal surfaces. Immunology letters 124, 57-62, doi:10.1016/j.imlet.2009.03.013 (2009). 26 Slack, E. et al. Innate and adaptive immunity cooperate flexibly to maintain host- microbiota mutualism. Science 325, 617-620, doi:10.1126/science.1172747 (2009). 27 Slack, E., Balmer, M. L., Fritz, J. H. & Hapfelmeier, S. Functional flexibility of intestinal IgA - broadening the fine line. Frontiers in immunology 3, 100, doi:10.3389/fimmu.2012.00100 (2012). 28 Uchimura, Y. et al. Antibodies Set Boundaries Limiting Microbial Metabolite Penetration and the Resultant Mammalian Host Response. Immunity 49, 545-559 e545, doi:10.1016/j.immuni.2018.08.004 (2018). 29 Corthesy, B. Multi-faceted functions of secretory IgA at mucosal surfaces. Frontiers in immunology 4, 185, doi:10.3389/fimmu.2013.00185 (2013). 30 Savage, D. C. Microbial ecology of the gastrointestinal tract. Annual review of microbiology 31, 107-133, doi:10.1146/annurev.mi.31.100177.000543 (1977). 31 Nicholson, J. K. et al. Host-Gut Microbiota Metabolic Interactions. Science 336, 1262, doi:10.1126/science.1223813 (2012). 32 Rinninella, E., Raoul, P. & Cintoni, M. What is the Healthy Gut Microbiota Composition? A Changing Ecosystem across Age, Environment, Diet, and Diseases. 7, doi:10.3390/microorganisms7010014 (2019). 33 Ley, R. E., Peterson, D. A. & Gordon, J. I. Ecological and evolutionary forces shaping microbial diversity in the human intestine. Cell 124, 837-848, doi:10.1016/j.cell.2006.02.017 (2006). 34 Donaldson, G. P., Lee, S. M. & Mazmanian, S. K. Gut biogeography of the bacterial microbiota. Nature Reviews Microbiology 14, 20-32, doi:10.1038/nrmicro3552 (2016). 35 Janzon, A. et al. Interactions between the Gut Microbiome and Mucosal Immunoglobulins A, M, and G in the Developing Infant Gut. mSystems 4, doi:10.1128/mSystems.00612-19 (2019). 36 Zheng, W. et al. Microbiota-targeted maternal antibodies protect neonates from enteric infection. Nature 577, 543-548, doi:10.1038/s41586-019-1898-4 (2020). 37 Sutherland, D. B., Suzuki, K. & Fagarasan, S. Fostering of advanced mutualism with gut microbiota by Immunoglobulin A. Immunological reviews 270, 20-31, doi:10.1111/imr.12384 (2016). 38 Koch, M. A. et al. Maternal IgG and IgA Antibodies Dampen Mucosal T Helper Cell Responses in Early Life. Cell 165, 827-841, doi:10.1016/j.cell.2016.04.055 (2016). 39 Gopalakrishna, K. P. et al. Maternal IgA protects against the development of necrotizing enterocolitis in preterm infants. 25, 1110-1115, doi:10.1038/s41591-019- 0480-9 (2019). 40 Hanson, L. A. & Korotkova, M. The role of breastfeeding in prevention of neonatal infection. Seminars in neonatology : SN 7, 275-281, doi:10.1016/s1084- 2756(02)90124-7 (2002). 41 Hanson, L. A. Breastfeeding Provides Passive and Likely Long-Lasting Active Immunity. Annals of Allergy, Asthma & Immunology 81, 523-537, doi:https://doi.org/10.1016/S1081-1206(10)62704-4 (1998). 42 Mueller, S. et al. Differences in fecal microbiota in different European study populations in relation to age, gender, and country: a cross-sectional study. Applied and environmental microbiology 72, 1027-1033, doi:10.1128/AEM.72.2.1027-1033.2006 (2006). 43 Biasucci, G., Benenati, B., Morelli, L., Bessi, E. & Boehm, G. Cesarean delivery may affect the early biodiversity of intestinal bacteria. The Journal of nutrition 138, 1796s- 1800s, doi:10.1093/jn/138.9.1796S (2008).

161

44 Yatsunenko, T. et al. Human gut microbiome viewed across age and geography. Nature 486, 222-227, doi:10.1038/nature11053 (2012). 45 Combellick, J. L. et al. Differences in the fecal microbiota of neonates born at home or in the hospital. Scientific reports 8, 15660, doi:10.1038/s41598-018-33995-7 (2018). 46 Caruso, R., Lo, B. C. & Núñez, G. Host-microbiota interactions in inflammatory bowel disease. 20, 411-426, doi:10.1038/s41577-019-0268-7 (2020). 47 Renz, H., Brandtzaeg, P. & Hornef, M. The impact of perinatal immune development on mucosal homeostasis and chronic inflammation. Nature reviews. Immunology 12, 9-23, doi:10.1038/nri3112 (2011). 48 Bäckhed, F. et al. Dynamics and Stabilization of the Human Gut Microbiome during the First Year of Life. Cell host & microbe 17, 690-703, doi:10.1016/j.chom.2015.04.004 (2015). 49 James, K. R. & Gomes, T. Distinct microbial and immune niches of the human colon. 21, 343-353, doi:10.1038/s41590-020-0602-z (2020). 50 Li, H. et al. The outer mucus layer hosts a distinct intestinal microbial niche. Nature communications 6, 8292, doi:10.1038/ncomms9292 (2015). 51 Maslowski, K. M. & Mackay, C. R. Diet, gut microbiota and immune responses. Nature immunology 12, 5-9, doi:10.1038/ni0111-5 (2011). 52 Arumugam, M. et al. Enterotypes of the human gut microbiome. Nature 473, 174-180, doi:10.1038/nature09944 (2011). 53 Ley, R. E. et al. Evolution of mammals and their gut microbes. Science 320, 1647- 1651, doi:10.1126/science.1155725 (2008). 54 Falony, G. et al. Population-level analysis of gut microbiome variation. Science 352, 560-564, doi:10.1126/science.aad3503 (2016). 55 Gensollen, T., Iyer, S. S., Kasper, D. L. & Blumberg, R. S. How colonization by microbiota in early life shapes the immune system. Science 352, 539-544, doi:10.1126/science.aad9378 (2016). 56 Greenhalgh, K., Meyer, K. M., Aagaard, K. M. & Wilmes, P. The human gut microbiome in health: establishment and resilience of microbiota over a lifetime. Environmental Microbiology 18, 2103-2116, doi:10.1111/1462-2920.13318 (2016). 57 Turnbaugh, P. J. et al. A core gut microbiome in obese and lean twins. Nature 457, 480-484, doi:10.1038/nature07540 (2009). 58 Schloissnig, S. et al. Genomic variation landscape of the human gut microbiome. Nature 493, 45-50, doi:10.1038/nature11711 (2013). 59 Tidjani Alou, M., Lagier, J.-C. & Raoult, D. Diet influence on the gut microbiota and dysbiosis related to nutritional disorders. Human Microbiome Journal 1, 3-11, doi:https://doi.org/10.1016/j.humic.2016.09.001 (2016). 60 Ahern, P. P., Faith, J. J. & Gordon, J. I. Mining the human gut microbiota for effector strains that shape the immune system. Immunity 40, 815-823, doi:10.1016/j.immuni.2014.05.012 (2014). 61 Qin, J. et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature 464, 59-65, doi:10.1038/nature08821 (2010). 62 Brestoff, J. R. & Artis, D. Commensal bacteria at the interface of host metabolism and the immune system. Nature immunology 14, 676-684, doi:10.1038/ni.2640 (2013). 63 Artis, D. Epithelial-cell recognition of commensal bacteria and maintenance of immune homeostasis in the gut. Nature reviews. Immunology 8, 411-420, doi:10.1038/nri2316 (2008). 64 Reboldi, A. & Cyster, J. G. Peyer's patches: organizing B-cell responses at the intestinal frontier. Immunological reviews 271, 230-245, doi:10.1111/imr.12400 (2016). 65 Makala, L. H., Suzuki, N. & Nagasawa, H. Peyer's patches: organized lymphoid structures for the induction of mucosal immune responses in the intestine. Pathobiology : journal of immunopathology, molecular and cellular biology 70, 55-68, doi:10.1159/000067305 (2002). 66 van de Pavert, S. A. & Mebius, R. E. New insights into the development of lymphoid tissues. Nature reviews. Immunology 10, 664-674, doi:10.1038/nri2832 (2010).

162

67 Kato, L. M., Kawamoto, S., Maruya, M. & Fagarasan, S. Gut TFH and IgA: key players for regulation of bacterial communities and immune homeostasis. Immunology and cell biology 92, 49-56, doi:10.1038/icb.2013.54 (2014). 68 Fung, T. C. et al. Lymphoid-Tissue-Resident Commensal Bacteria Promote Members of the IL-10 Cytokine Family to Establish Mutualism. Immunity 44, 634-646, doi:10.1016/j.immuni.2016.02.019 (2016). 69 Arpaia, N. et al. Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature 504, 451-455, doi:10.1038/nature12726 (2013). 70 Sun, M. et al. Microbiota-derived short-chain fatty acids promote Th1 cell IL-10 production to maintain intestinal homeostasis. Nature communications 9, 3555, doi:10.1038/s41467-018-05901-2 (2018). 71 Mabbott, N. A., Donaldson, D. S., Ohno, H., Williams, I. R. & Mahajan, A. Microfold (M) cells: important immunosurveillance posts in the intestinal epithelium. Mucosal immunology 6, 666-677, doi:10.1038/mi.2013.30 (2013). 72 Yamanaka, T. et al. Microbial colonization drives lymphocyte accumulation and differentiation in the follicle-associated epithelium of Peyer's patches. Journal of immunology 170, 816-822, doi:10.4049/jimmunol.170.2.816 (2003). 73 Bauer, E., Williams, B. A., Smidt, H., Verstegen, M. W. & Mosenthin, R. Influence of the gastrointestinal microbiota on development of the immune system in young animals. Current issues in intestinal microbiology 7, 35-51 (2006). 74 Round, J. L. & Mazmanian, S. K. The gut microbiota shapes intestinal immune responses during health and disease. Nature reviews. Immunology 9, 313-323, doi:10.1038/nri2515 (2009). 75 Hooper, L. V., Littman, D. R. & Macpherson, A. J. Interactions Between the Microbiota and the Immune System. Science 336, 1268-1273, doi:10.1126/science.1223490 (2012). 76 Huus, K. E. et al. Commensal Bacteria Modulate Immunoglobulin A Binding in Response to Host Nutrition. Cell host & microbe 27, 909-921 e905, doi:10.1016/j.chom.2020.03.012 (2020). 77 Schnupf, P., Gaboriau-Routhiau, V., Sansonetti, P. J. & Cerf-Bensussan, N. Segmented filamentous bacteria, Th17 inducers and helpers in a hostile world. Current opinion in microbiology 35, 100-109, doi:10.1016/j.mib.2017.03.004 (2017). 78 Ansaldo, E. & Slayden, L. C. Akkermansia muciniphila induces intestinal adaptive immune responses during homeostasis. 364, 1179-1184, doi:10.1126/science.aaw7479 (2019). 79 Lécuyer, E. et al. Segmented filamentous bacterium uses secondary and tertiary lymphoid tissues to induce gut IgA and specific T helper 17 cell responses. Immunity 40, 608-620, doi:10.1016/j.immuni.2014.03.009 (2014). 80 Rios-Covian, D., Salazar, N., Gueimonde, M. & de Los Reyes-Gavilan, C. G. Shaping the Metabolism of Intestinal Bacteroides Population through Diet to Improve Human Health. Frontiers in microbiology 8, 376-376, doi:10.3389/fmicb.2017.00376 (2017). 81 Sun, M., Wu, W., Liu, Z. & Cong, Y. Microbiota metabolite short chain fatty acids, GPCR, and inflammatory bowel diseases. Journal of gastroenterology 52, 1-8, doi:10.1007/s00535-016-1242-9 (2017). 82 Smith, P. M. et al. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science 341, 569-573, doi:10.1126/science.1241165 (2013). 83 Chen, K., Magri, G., Grasset, E. K. & Cerutti, A. Rethinking mucosal antibody responses: IgM, IgG and IgD join IgA. Nature reviews. Immunology 20, 427-441, doi:10.1038/s41577-019-0261-1 (2020). 84 Bach Knudsen, K. E. et al. Impact of Diet-Modulated Butyrate Production on Intestinal Barrier Function and Inflammation. Nutrients 10, 1499, doi:10.3390/nu10101499 (2018). 85 Kim, H., Jeong, Y., Kang, S., You, H. J. & Ji, G. E. Co-Culture with Bifidobacterium catenulatum Improves the Growth, Gut Colonization, and Butyrate Production of Faecalibacterium prausnitzii: In Vitro and In Vivo Studies. Microorganisms 8, 788, doi:10.3390/microorganisms8050788 (2020). 163

86 Aparicio-Domingo, P. et al. Type 3 innate lymphoid cells maintain intestinal epithelial stem cells after tissue damage. The Journal of experimental medicine 212, 1783-1791, doi:10.1084/jem.20150318 (2015). 87 Ley, R. E. et al. Obesity alters gut microbial ecology. Proceedings of the National Academy of Sciences of the United States of America 102, 11070-11075, doi:10.1073/pnas.0504978102 (2005). 88 Macpherson, A. J., Koller, Y. & McCoy, K. D. The bilateral responsiveness between intestinal microbes and IgA. Trends in immunology 36, 460-470, doi:10.1016/j.it.2015.06.006 (2015). 89 Macpherson, A. J., Hunziker, L., McCoy, K. & Lamarre, A. IgA responses in the intestinal mucosa against pathogenic and non-pathogenic microorganisms. Microbes and infection / Institut Pasteur 3, 1021-1035, doi:10.1016/s1286-4579(01)01460-5 (2001). 90 Pabst, O. New concepts in the generation and functions of IgA. Nature reviews. Immunology 12, 821-832, doi:10.1038/nri3322 (2012). 91 Pabst, O. & Slack, E. IgA and the intestinal microbiota: the importance of being specific. Mucosal immunology 13, 12-21, doi:10.1038/s41385-019-0227-4 (2020). 92 Suzuki, K., Kawamoto, S., Maruya, M. & Fagarasan, S. GALT: organization and dynamics leading to IgA synthesis. Advances in immunology 107, 153-185, doi:10.1016/B978-0-12-381300-8.00006-X (2010). 93 Fenton, T. M. et al. Immune Profiling of Human Gut-Associated Lymphoid Tissue Identifies a Role for Isolated Lymphoid Follicles in Priming of Region-Specific Immunity. Immunity 52, 557-570 e556, doi:10.1016/j.immuni.2020.02.001 (2020). 94 Brandtzaeg, P. Mucosal immunity: induction, dissemination, and effector functions. Scandinavian journal of immunology 70, 505-515, doi:10.1111/j.1365- 3083.2009.02319.x (2009). 95 Tsuji, M. et al. Requirement for lymphoid tissue-inducer cells in isolated follicle formation and T cell-independent immunoglobulin A generation in the gut. Immunity 29, 261-271, doi:10.1016/j.immuni.2008.05.014 (2008). 96 Buettner, M. & Lochner, M. Development and Function of Secondary and Tertiary Lymphoid Organs in the Small Intestine and the Colon. Frontiers in immunology 7, 342, doi:10.3389/fimmu.2016.00342 (2016). 97 De Silva, N. S. & Klein, U. Dynamics of B cells in germinal centres. Nature reviews. Immunology 15, 137-148, doi:10.1038/nri3804 (2015). 98 Victora, G. D. & Nussenzweig, M. C. Germinal Centers. Annual review of immunology 30, 429-457, doi:10.1146/annurev-immunol-020711-075032 (2012). 99 Cyster, J. G. & Allen, C. D. C. B Cell Responses: Cell Interaction Dynamics and Decisions. Cell 177, 524-540, doi:10.1016/j.cell.2019.03.016 (2019). 100 Shulman, Z. et al. T follicular helper cell dynamics in germinal centers. Science 341, 673-677, doi:10.1126/science.1241680 (2013). 101 Brandtzaeg, P. & Johansen, F. E. Mucosal B cells: phenotypic characteristics, transcriptional regulation, and homing properties. Immunological reviews 206, 32-63, doi:10.1111/j.0105-2896.2005.00283.x (2005). 102 Reboldi, A. et al. IgA production requires B cell interaction with subepithelial dendritic cells in Peyer's patches. Science 352, aaf4822, doi:10.1126/science.aaf4822 (2016). 103 Neutra, M. R., Mantis, N. J. & Kraehenbuhl, J.-P. Collaboration of epithelial cells with organized mucosal lymphoid tissues. Nature immunology 2, 1004-1009, doi:10.1038/ni1101-1004 (2001). 104 Brandtzaeg, P. et al. The B-cell system of human mucosae and exocrine glands. Immunological reviews 171, 45-87, doi:https://doi.org/10.1111/j.1600- 065X.1999.tb01342.x (1999). 105 Bergqvist, P. et al. Re-utilization of germinal centers in multiple Peyer's patches results in highly synchronized, oligoclonal, and affinity-matured gut IgA responses. Mucosal immunology 6, 122-135, doi:10.1038/mi.2012.56 (2013).

164

106 Benckert, J. et al. The majority of intestinal IgA+ and IgG+ plasmablasts in the human gut are antigen-specific. The Journal of clinical investigation 121, 1946-1955, doi:10.1172/JCI44447 (2011). 107 Magri, G. et al. Human Secretory IgM Emerges from Plasma Cells Clonally Related to Gut Memory B Cells and Targets Highly Diverse Commensals. Immunity 47, 118-134 e118, doi:10.1016/j.immuni.2017.06.013 (2017). 108 Fadlallah, J. et al. Synergistic convergence of microbiota-specific systemic IgG and secretory IgA. J Allergy Clin Immunol 143, 1575-1585 e1574, doi:10.1016/j.jaci.2018.09.036 (2019). 109 Kubinak, J. L. & Round, J. L. Do antibodies select a healthy microbiota? Nature reviews. Immunology 16, 767-774, doi:10.1038/nri.2016.114 (2016). 110 Armstrong, H. et al. Host immunoglobulin G selectively identifies pathobionts in pediatric inflammatory bowel diseases. Microbiome 7, 1-1, doi:10.1186/s40168-018- 0604-3 (2019). 111 Kawamoto, S. et al. Foxp3(+) T cells regulate immunoglobulin a selection and facilitate diversification of bacterial species responsible for immune homeostasis. Immunity 41, 152-165, doi:10.1016/j.immuni.2014.05.016 (2014). 112 Keim, C., Kazadi, D., Rothschild, G. & Basu, U. Regulation of AID, the B-cell genome mutator. Genes & development 27, 1-17, doi:10.1101/gad.200014.112 (2013). 113 Hardy, R. R. & Hayakawa, K. B cell development pathways. Annu Rev Immunol 19, 595-621, doi:10.1146/annurev.immunol.19.1.595 (2001). 114 Pelanda, R. & Torres, R. M. Central B-cell tolerance: where selection begins. Cold Spring Harbor perspectives in biology 4, a007146, doi:10.1101/cshperspect.a007146 (2012). 115 Xu, Y. et al. No receptor stands alone: IgG B-cell receptor intrinsic and extrinsic mechanisms contribute to antibody memory. Cell Research 24, 651-664, doi:10.1038/cr.2014.65 (2014). 116 Schlissel, M. S. Regulating antigen-receptor gene assembly. Nature reviews. Immunology 3, 890-899, doi:10.1038/nri1225 (2003). 117 Fritz, J. H. et al. Acquisition of a multifunctional IgA+ plasma cell phenotype in the gut. Nature 481, 199-203, doi:10.1038/nature10698 (2012). 118 Guo, C. et al. CTCF-binding elements mediate control of V(D)J recombination. Nature 477, 424-430, doi:10.1038/nature10495 (2011). 119 Chaudhuri, J. & Alt, F. W. Class-switch recombination: interplay of transcription, DNA deamination and DNA repair. Nature reviews. Immunology 4, 541-552, doi:10.1038/nri1395 (2004). 120 Pone, E. J. et al. BCR-signalling synergizes with TLR-signalling for induction of AID and immunoglobulin class-switching through the non-canonical NF-kappaB pathway. Nature communications 3, 767, doi:10.1038/ncomms1769 (2012). 121 Pinto, D. et al. A functional BCR in human IgA and IgM plasma cells. Blood 121, 4110- 4114, doi:10.1182/blood-2012-09-459289 (2013). 122 Yam-Puc, J. C., Zhang, L., Zhang, Y. & Toellner, K.-M. Role of B-cell receptors for B- cell development and antigen-induced differentiation. F1000Res 7, 429-429, doi:10.12688/f1000research.13567.1 (2018). 123 Kraus, M., Alimzhanov, M. B., Rajewsky, N. & Rajewsky, K. Survival of resting mature B lymphocytes depends on BCR signaling via the Igalpha/beta heterodimer. Cell 117, 787-800, doi:10.1016/j.cell.2004.05.014 (2004). 124 Samitas, K., Lotvall, J. & Bossios, A. B cells: from early development to regulating allergic diseases. Archivum immunologiae et therapiae experimentalis 58, 209-225, doi:10.1007/s00005-010-0073-2 (2010). 125 DeFranco, A. L. The complexity of signaling pathways activated by the BCR. Current opinion in immunology 9, 296-308, doi:10.1016/s0952-7915(97)80074-x (1997). 126 Brummer, T., Elis, W., Reth, M. & Huber, M. B-cell signal transduction: tyrosine phosphorylation, kinase activity, and calcium mobilization. Methods in molecular biology (Clifton, N.J.) 271, 189-212, doi:10.1385/1-59259-796-3:189 (2004).

165

127 Amigorena, S., Drake, J. R., Webster, P. & Mellman, I. Transient accumulation of new class II MHC molecules in a novel endocytic compartment in B lymphocytes. Nature 369, 113-120, doi:10.1038/369113a0 (1994). 128 West, A. P., Koblansky, A. A. & Ghosh, S. Recognition and signaling by toll-like receptors. Annual review of cell and developmental biology 22, 409-437, doi:10.1146/annurev.cellbio.21.122303.115827 (2006). 129 Schweighoffer, E. & Nys, J. TLR4 signals in B lymphocytes are transduced via the B cell antigen receptor and SYK. 214, 1269-1280, doi:10.1084/jem.20161117 (2017). 130 Pone, E. J., Xu, Z., White, C. A., Zan, H. & Casali, P. B cell TLRs and induction of immunoglobulin class-switch DNA recombination. Front Biosci (Landmark Ed) 17, 2594-2615, doi:10.2741/4073 (2012). 131 Kawasaki, T. & Kawai, T. Toll-like receptor signaling pathways. Frontiers in immunology 5, 461, doi:10.3389/fimmu.2014.00461 (2014). 132 MacLennan, I. C. et al. Extrafollicular antibody responses. Immunological reviews 194, 8-18, doi:10.1034/j.1600-065x.2003.00058.x (2003). 133 Taylor, J. J., Pape, K. A. & Jenkins, M. K. A germinal center-independent pathway generates unswitched memory B cells early in the primary response. The Journal of experimental medicine 209, 597-606, doi:10.1084/jem.20111696 (2012). 134 McHeyzer-Williams, L. J. & McHeyzer-Williams, M. G. Antigen-specific memory B cell development. Annu Rev Immunol 23, 487-513, doi:10.1146/annurev.immunol.23.021704.115732 (2005). 135 McHeyzer-Williams, M., Okitsu, S., Wang, N. & McHeyzer-Williams, L. Molecular programming of B cell memory. Nature reviews. Immunology 12, 24-34, doi:10.1038/nri3128 (2011). 136 Nowosad, C. R. et al. Tunable dynamics of B cell selection in gut germinal centres. Nature, doi:10.1038/s41586-020-2865-9 (2020). 137 Cerutti, A. The regulation of IgA class switching. Nature reviews. Immunology 8, 421- 434, doi:10.1038/nri2322 (2008). 138 Cazac, B. B. & Roes, J. TGF-beta receptor controls B cell responsiveness and induction of IgA in vivo. Immunity 13, 443-451, doi:10.1016/s1074-7613(00)00044-3 (2000). 139 Boyd, S. D. & Joshi, S. A. in Antibodies for Infectious Diseases (American Society of Microbiology, 2015). 140 Vidarsson, G., Dekkers, G. & Rispens, T. IgG Subclasses and Allotypes: From Structure to Effector Functions. Frontiers in immunology 5, doi:10.3389/fimmu.2014.00520 (2014). 141 Mantis, N. J. & Forbes, S. J. Secretory IgA: Arresting Microbial Pathogens at Epithelial Borders. Immunological Investigations 39, 383-406, doi:10.3109/08820131003622635 (2010). 142 Mantis, N. J., Rol, N. & Corthésy, B. Secretory IgA's complex roles in immunity and mucosal homeostasis in the gut. Mucosal immunology 4, 603-611, doi:10.1038/mi.2011.41 (2011). 143 Steffen, U. et al. IgA subclasses have different effector functions associated with distinct glycosylation profiles. Nature communications 11, 120, doi:10.1038/s41467- 019-13992-8 (2020). 144 Cymer, F., Beck, H., Rohde, A. & Reusch, D. Therapeutic monoclonal antibody N- glycosylation - Structure, function and therapeutic potential. Biologicals : journal of the International Association of Biological Standardization 52, 1-11, doi:10.1016/j.biologicals.2017.11.001 (2018). 145 Pakkanen, S. H. et al. Expression of homing receptors on IgA1 and IgA2 plasmablasts in blood reflects differential distribution of IgA1 and IgA2 in various body fluids. Clinical and vaccine immunology : CVI 17, 393-401, doi:10.1128/CVI.00475-09 (2010). 146 Woof, J. M. & Russell, M. W. Structure and function relationships in IgA. Mucosal immunology 4, 590-597, doi:10.1038/mi.2011.39 (2011).

166

147 Plomp, R. et al. Comparative Glycomics of Immunoglobulin A and G From Saliva and Plasma Reveals Biomarker Potential. Frontiers in immunology 9, 2436, doi:10.3389/fimmu.2018.02436 (2018). 148 Hansen, I. S., Baeten, D. L. P. & den Dunnen, J. The inflammatory function of human IgA. Cellular and molecular life sciences : CMLS 76, 1041-1055, doi:10.1007/s00018- 018-2976-8 (2019). 149 Yoshida, M. et al. IgG transport across mucosal barriers by neonatal Fc receptor for IgG and mucosal immunity. Springer seminars in immunopathology 28, 397-403, doi:10.1007/s00281-006-0054-z (2006). 150 Zeng, M. Y. et al. Gut Microbiota-Induced Immunoglobulin G Controls Systemic Infection by Symbiotic Bacteria and Pathogens. Immunity 44, 647-658, doi:10.1016/j.immuni.2016.02.006 (2016). 151 Fadlallah, J. et al. Microbial ecology perturbation in human IgA deficiency. Sci Transl Med 10, doi:10.1126/scitranslmed.aan1217 (2018). 152 Andreas, B., Hartmut, K., Klaus, E. & Norbert, H. Intra- and Interchain Disulfide Bridges of the Human J Chain in Secretory Immunoglobulin A. Biological Chemistry 373, 1255- 1264, doi:https://doi.org/10.1515/bchm3.1992.373.2.1255 (1992). 153 Brandtzaeg, P. & Prydz, H. Direct evidence for an integrated function of J chain and secretory component in epithelial transport of immunoglobulins. Nature 311, 71-73 (1984). 154 Johansen, F. E. & Kaetzel, C. S. Regulation of the polymeric immunoglobulin receptor and IgA transport: new advances in environmental factors that stimulate pIgR expression and its role in mucosal immunity. Mucosal immunology 4, 598-602, doi:10.1038/mi.2011.37 (2011). 155 Strugnell, R. A. & Wijburg, O. L. The role of secretory antibodies in infection immunity. Nat Rev Microbiol 8, 656-667, doi:10.1038/nrmicro2384 (2010). 156 Van Der Feltz, M. J. M. et al. Lymphocyte Homing and Ig Secretion in the Murine Mammary Gland. Scandinavian journal of immunology 54, 292-300, doi:https://doi.org/10.1046/j.1365-3083.2001.00933.x (2001). 157 Shulzhenko, N. et al. Crosstalk between B lymphocytes, microbiota and the intestinal epithelium governs immunity versus metabolism in the gut. Nature medicine 17, 1585- 1593, doi:10.1038/nm.2505 (2011). 158 Brown, W. R. & Kloppel, T. M. The liver and IgA: immunological, cell biological and clinical implications. Hepatology 9, 763-784 (1989). 159 Brown, W. R. & Kloppel, T. M. The role of the liver in translocation of IgA into the gastrointestinal tract. Immunol Invest 18, 269-285 (1989). 160 Krajci, P. et al. Molecular cloning of the human transmembrane secretory component (poly-Ig receptor) and its mRNA expression in human tissues. Biochemical and Biophysical Research Communications 158, 783-789, doi:https://doi.org/10.1016/0006-291X(89)92790-3 (1989). 161 Pabst, O., Cerovic, V. & Hornef, M. Secretory IgA in the Coordination of Establishment and Maintenance of the Microbiota. Trends in immunology 37, 287-296, doi:10.1016/j.it.2016.03.002 (2016). 162 Moro-Sibilot, L. et al. Mouse and Human Liver Contain Immunoglobulin A-Secreting Cells Originating From Peyer's Patches and Directed Against Intestinal Antigens. Gastroenterology 151, 311-323, doi:10.1053/j.gastro.2016.04.014 (2016). 163 Dickinson, B. L. et al. Bidirectional FcRn-dependent IgG transport in a polarized human intestinal epithelial cell line. The Journal of clinical investigation 104, 903-911, doi:10.1172/JCI6968 (1999). 164 Yoshida, M. et al. Human neonatal Fc receptor mediates transport of IgG into luminal secretions for delivery of antigens to mucosal dendritic cells. Immunity 20, 769-783, doi:10.1016/j.immuni.2004.05.007 (2004). 165 Rath, T., Baker, K., Pyzik, M. & Blumberg, R. S. Regulation of immune responses by the neonatal fc receptor and its therapeutic implications. Frontiers in immunology 5, 664, doi:10.3389/fimmu.2014.00664 (2014).

167

166 Kubinak, J. L. et al. MyD88 signaling in T cells directs IgA-mediated control of the microbiota to promote health. Cell host & microbe 17, 153-163, doi:10.1016/j.chom.2014.12.009 (2015). 167 van der Waaij, L. A., Limburg, P. C., Mesander, G. & van der Waaij, D. In vivo IgA coating of anaerobic bacteria in human faeces. Gut 38, 348-354, doi:10.1136/gut.38.3.348 (1996). 168 Sterlin, D. et al. Correction: Human IgA bind a diverse array of commensal bacteria. The Journal of experimental medicine 217, doi:10.1084/jem.2018163501152020c (2020). 169 Planer, J. D. et al. Development of the gut microbiota and mucosal IgA responses in twins and gnotobiotic mice. Nature 534, 263-266, doi:10.1038/nature17940 (2016). 170 Palm, N. W. et al. Immunoglobulin A coating identifies colitogenic bacteria in inflammatory bowel disease. Cell 158, 1000-1010, doi:10.1016/j.cell.2014.08.006 (2014). 171 Blanton, L. V. et al. Gut bacteria that prevent growth impairments transmitted by microbiota from malnourished children. Science 351, doi:10.1126/science.aad3311 (2016). 172 Kau, A. L. et al. Functional characterization of IgA-targeted bacterial taxa from undernourished Malawian children that produce diet-dependent enteropathy. Sci Transl Med 7, 276ra224, doi:10.1126/scitranslmed.aaa4877 (2015). 173 Fransen, F. et al. BALB/c and C57BL/6 Mice Differ in Polyreactive IgA Abundance, which Impacts the Generation of Antigen-Specific IgA and Microbiota Diversity. Immunity 43, 527-540, doi:10.1016/j.immuni.2015.08.011 (2015). 174 Kawamoto, S. et al. The inhibitory receptor PD-1 regulates IgA selection and bacterial composition in the gut. Science 336, 485-489, doi:10.1126/science.1217718 (2012). 175 Okai, S., Usui, F., Ohta, M., Mori, H. & Kurokawa, K. Intestinal IgA as a modulator of the gut microbiota. 8, 486-492, doi:10.1080/19490976.2017.1310357 (2017). 176 Wei, M. et al. Mice carrying a knock-in mutation of Aicda resulting in a defect in somatic hypermutation have impaired gut homeostasis and compromised mucosal defense. Nature immunology 12, 264-270, doi:10.1038/ni.1991 (2011). 177 Reikvam, D. H. et al. Epithelial-microbial crosstalk in polymeric Ig receptor deficient mice. European journal of immunology 42, 2959-2970, doi:10.1002/eji.201242543 (2012). 178 Johansson, M. E. et al. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and patients with ulcerative colitis. Gut 63, 281-291, doi:10.1136/gutjnl-2012-303207 (2014). 179 Lycke, N., Erlandsson, L., Ekman, L., Schon, K. & Leanderson, T. Lack of J chain inhibits the transport of gut IgA and abrogates the development of intestinal antitoxic protection. Journal of immunology 163, 913-919 (1999). 180 Mantis, N. J., Rol, N. & Corthesy, B. Secretory IgA's complex roles in immunity and mucosal homeostasis in the gut. Mucosal immunology 4, 603-611, doi:10.1038/mi.2011.41 (2011). 181 Phalipon, A. et al. Secretory component: a new role in secretory IgA-mediated immune exclusion in vivo. Immunity 17, 107-115, doi:10.1016/s1074-7613(02)00341-2 (2002). 182 Moor, K. et al. High-avidity IgA protects the intestine by enchaining growing bacteria. Nature 544, 498-502, doi:10.1038/nature22058 (2017). 183 Bansept, F. et al. Enchained growth and cluster dislocation: A possible mechanism for microbiota homeostasis. PLoS Comput Biol 15, e1006986, doi:10.1371/journal.pcbi.1006986 (2019). 184 Boullier, S. et al. Secretory IgA-mediated neutralization of Shigella flexneri prevents intestinal tissue destruction by down-regulating inflammatory circuits. Journal of immunology 183, 5879-5885, doi:10.4049/jimmunol.0901838 (2009). 185 Forbes, S. J., Bumpus, T., McCarthy, E. A., Corthésy, B. & Mantis, N. J. Transient suppression of Shigella flexneri type 3 secretion by a protective O-antigen-specific monoclonal IgA. mBio 2, e00042-00011, doi:10.1128/mBio.00042-11 (2011).

168

186 Peterson, D. A., McNulty, N. P., Guruge, J. L. & Gordon, J. I. IgA response to symbiotic bacteria as a mediator of gut homeostasis. Cell host & microbe 2, 328-339, doi:10.1016/j.chom.2007.09.013 (2007). 187 Peterson, D. A. et al. Characterizing the interactions between a naturally primed immunoglobulin A and its conserved Bacteroides thetaiotaomicron species-specific epitope in gnotobiotic mice. J Biol Chem 290, 12630-12649, doi:10.1074/jbc.M114.633800 (2015). 188 Robinson, J. K., Blanchard, T. G., Levine, A. D., Emancipator, S. N. & Lamm, M. E. A mucosal IgA-mediated excretory immune system in vivo. Journal of immunology 166, 3688-3692, doi:10.4049/jimmunol.166.6.3688 (2001). 189 Rochereau, N. et al. Dectin-1 Is Essential for Reverse Transcytosis of Glycosylated SIgA-Antigen Complexes by Intestinal M Cells. PLoS biology 11, e1001658, doi:10.1371/journal.pbio.1001658 (2013). 190 Mathias, A., Pais, B., Favre, L., Benyacoub, J. & Corthésy, B. Role of secretory IgA in the mucosal sensing of commensal bacteria. Gut microbes 5, 688-695, doi:10.4161/19490976.2014.983763 (2014). 191 Bollinger, R. R. et al. Human secretory immunoglobulin A may contribute to biofilm formation in the gut. Immunology 109, 580-587, doi:10.1046/j.1365-2567.2003.01700.x (2003). 192 Mathias, A., Pais, B., Favre, L., Benyacoub, J. & Corthesy, B. Role of secretory IgA in the mucosal sensing of commensal bacteria. Gut Microbes 5, 688-695, doi:10.4161/19490976.2014.983763 (2014). 193 Cullender, T. C. et al. Innate and adaptive immunity interact to quench microbiome flagellar motility in the gut. Cell host & microbe 14, 571-581, doi:10.1016/j.chom.2013.10.009 (2013). 194 Donaldson, G. P. & Ladinsky, M. S. Gut microbiota utilize immunoglobulin A for mucosal colonization. 360, 795-800, doi:10.1126/science.aaq0926 (2018). 195 Royle, L. et al. Secretory IgA N- and O-glycans provide a link between the innate and adaptive immune systems. J Biol Chem 278, 20140-20153, doi:10.1074/jbc.M301436200 (2003). 196 Mathias, A. & Corthesy, B. N-Glycans on secretory component: mediators of the interaction between secretory IgA and gram-positive commensals sustaining intestinal homeostasis. Gut Microbes 2, 287-293, doi:10.4161/gmic.2.5.18269 (2011). 197 Maurer, M. A. et al. Glycosylation of Human IgA Directly Inhibits Influenza A and Other Sialic-Acid-Binding Viruses. Cell reports 23, 90-99, doi:10.1016/j.celrep.2018.03.027 (2018). 198 Briliūtė, J. & Urbanowicz, P. A. Complex N-glycan breakdown by gut Bacteroides involves an extensive enzymatic apparatus encoded by multiple co-regulated genetic loci. 4, 1571-1581, doi:10.1038/s41564-019-0466-x (2019). 199 Cao, Y., Rocha, E. R. & Smith, C. J. Efficient utilization of complex N-linked glycans is a selective advantage for Bacteroides fragilis in extraintestinal infections. Proceedings of the National Academy of Sciences of the United States of America 111, 12901- 12906, doi:10.1073/pnas.1407344111 (2014). 200 Eilam, O. et al. Glycan degradation (GlyDeR) analysis predicts mammalian gut microbiota abundance and host diet-specific adaptations. MBio 5, doi:10.1128/mBio.01526-14 (2014). 201 Nakajima, A. et al. IgA regulates the composition and metabolic function of gut microbiota by promoting symbiosis between bacteria. The Journal of experimental medicine 215, 2019-2034, doi:10.1084/jem.20180427 (2018). 202 Macpherson, A. J., Yilmaz, B., Limenitakis, J. P. & Ganal-Vonarburg, S. C. IgA Function in Relation to the Intestinal Microbiota. Annu Rev Immunol 36, 359-381, doi:10.1146/annurev-immunol-042617-053238 (2018). 203 Corthesy, B. Role of secretory immunoglobulin A and secretory component in the protection of mucosal surfaces. Future Microbiol 5, 817-829, doi:10.2217/fmb.10.39 (2010).

169

204 Macpherson, A. J. & McCoy, K. D. Independence Day for IgA. Immunity 43, 416-418, doi:10.1016/j.immuni.2015.08.024 (2015). 205 Bergqvist, P., Stensson, A., Lycke, N. Y. & Bemark, M. T cell-independent IgA class switch recombination is restricted to the GALT and occurs prior to manifest germinal center formation. Journal of immunology 184, 3545-3553, doi:10.4049/jimmunol.0901895 (2010). 206 Fagarasan, S., Kawamoto, S., Kanagawa, O. & Suzuki, K. Adaptive immune regulation in the gut: T cell-dependent and T cell-independent IgA synthesis. Annu Rev Immunol 28, 243-273, doi:10.1146/annurev-immunol-030409-101314 (2010). 207 Bergqvist, P., Gardby, E., Stensson, A., Bemark, M. & Lycke, N. Y. Gut IgA class switch recombination in the absence of CD40 does not occur in the lamina propria and is independent of germinal centers. Journal of immunology 177, 7772-7783 (2006). 208 Barone, F. et al. IgA-producing plasma cells originate from germinal centers that are induced by B-cell receptor engagement in humans. Gastroenterology 140, 947-956, doi:10.1053/j.gastro.2010.12.005 (2011). 209 Chen, H. et al. BCR selection and affinity maturation in Peyer's patch germinal centres. Nature 582, 421-425, doi:10.1038/s41586-020-2262-4 (2020). 210 Landsverk, O. J. et al. Antibody-secreting plasma cells persist for decades in human intestine. The Journal of experimental medicine 214, 309-317, doi:10.1084/jem.20161590 (2017). 211 Bemark, M., Boysen, P. & Lycke, N. Y. Induction of gut IgA production through T cell- dependent and T cell-independent pathways. Annals of the New York Academy of Sciences 1247, 97-116, doi:10.1111/j.1749-6632.2011.06378.x (2012). 212 Fagarasan, S. & Honjo, T. T-Independent immune response: new aspects of B cell biology. Science 290, 89-92, doi:10.1126/science.290.5489.89 (2000). 213 Notkins, A. L. Polyreactivity of antibody molecules. Trends in immunology 25, 174-179, doi:https://doi.org/10.1016/j.it.2004.02.004 (2004). 214 He, B. et al. Intestinal Bacteria Trigger T Cell-Independent Immunoglobulin A2 Class Switching by Inducing Epithelial-Cell Secretion of the Cytokine APRIL. Immunity 26, 812-826, doi:https://doi.org/10.1016/j.immuni.2007.04.014 (2007). 215 Zhou, Z. H. et al. The broad antibacterial activity of the natural antibody repertoire is due to polyreactive antibodies. Cell host & microbe 1, 51-61, doi:10.1016/j.chom.2007.01.002 (2007). 216 Chen, J. W. et al. Autoreactivity in naïve human fetal B cells is associated with commensal bacteria recognition. Science 369, 320, doi:10.1126/science.aay9733 (2020). 217 Suzuki, K. & Fagarasan, S. Diverse regulatory pathways for IgA synthesis in the gut. Mucosal immunology 2, 468-471, doi:10.1038/mi.2009.107 (2009). 218 Yuvaraj, S. et al. Evidence for Local Expansion of IgA Plasma Cell Precursors in Human Ileum. The Journal of Immunology 183, 4871, doi:10.4049/jimmunol.0901315 (2009). 219 McCarthy, D. D., Chiu, S., Gao, Y., Summers-deLuca, L. E. & Gommerman, J. L. BAFF induces a hyper-IgA syndrome in the intestinal lamina propria concomitant with IgA deposition in the kidney independent of LIGHT. Cellular immunology 241, 85-94, doi:10.1016/j.cellimm.2006.08.002 (2006). 220 Castigli, E. et al. TACI and BAFF-R mediate isotype switching in B cells. The Journal of experimental medicine 201, 35-39, doi:10.1084/jem.20032000 (2005). 221 Schneider, P. The role of APRIL and BAFF in lymphocyte activation. Current opinion in immunology 17, 282-289, doi:10.1016/j.coi.2005.04.005 (2005). 222 Castigli, E. et al. Impaired IgA class switching in APRIL-deficient mice. Proceedings of the National Academy of Sciences of the United States of America 101, 3903, doi:10.1073/pnas.0307348101 (2004). 223 Kroese, F. G., de Waard, R. & Bos, N. A. B-1 cells and their reactivity with the murine intestinal microflora. Semin Immunol 8, 11-18, doi:10.1006/smim.1996.0003 (1996).

170

224 Ha, S.-a. et al. Regulation of B1 cell migration by signals through Toll-like receptors. The Journal of experimental medicine 203, 2541-2550, doi:10.1084/jem.20061041 (2006). 225 Suzuki, K., Ha, S. A., Tsuji, M. & Fagarasan, S. Intestinal IgA synthesis: a primitive form of adaptive immunity that regulates microbial communities in the gut. Semin Immunol 19, 127-135, doi:10.1016/j.smim.2006.10.001 (2007). 226 Marks, E. et al. Retinoic Acid Signaling in B Cells Is Required for the Generation of an Effective T-Independent Immune Response. Frontiers in immunology 7, 643-643, doi:10.3389/fimmu.2016.00643 (2016). 227 Rothstein, T. L., Griffin, D. O., Holodick, N. E., Quach, T. D. & Kaku, H. Human B-1 cells take the stage. Annals of the New York Academy of Sciences 1285, 97-114, doi:10.1111/nyas.12137 (2013). 228 Fagarasan, S. & Honjo, T. Intestinal IgA synthesis: regulation of front-line body defences. Nature reviews. Immunology 3, 63-72, doi:10.1038/nri982 (2003). 229 Wijburg, O. L. et al. Innate secretory antibodies protect against natural Salmonella typhimurium infection. The Journal of experimental medicine 203, 21-26, doi:10.1084/jem.20052093 (2006). 230 Macpherson, A. J. & Harris, N. L. Interactions between commensal intestinal bacteria and the immune system. Nature reviews. Immunology 4, 478-485, doi:10.1038/nri1373 (2004). 231 Spencer, J., Klavinskis, L. S. & Fraser, L. D. The human intestinal IgA response; burning questions. Frontiers in immunology 3, 108, doi:10.3389/fimmu.2012.00108 (2012). 232 Suzuki, K. & Fagarasan, S. How host-bacterial interactions lead to IgA synthesis in the gut. Trends in immunology 29, 523-531, doi:10.1016/j.it.2008.08.001 (2008). 233 Wardemann, H. et al. Predominant autoantibody production by early human B cell precursors. Science 301, 1374-1377, doi:10.1126/science.1086907 (2003). 234 Weller, S. et al. Human blood IgM "memory" B cells are circulating splenic marginal zone B cells harboring a prediversified immunoglobulin repertoire. Blood 104, 3647- 3654, doi:10.1182/blood-2004-01-0346 (2004). 235 Sims, G. P. et al. Identification and characterization of circulating human transitional B cells. Blood 105, 4390-4398, doi:10.1182/blood-2004-11-4284 (2005). 236 Lycke, N. Y. & Bemark, M. The role of Peyer's patches in synchronizing gut IgA responses. Frontiers in immunology 3, 329, doi:10.3389/fimmu.2012.00329 (2012). 237 Masahata, K. et al. Generation of colonic IgA-secreting cells in the caecal patch. Nature communications 5, 3704, doi:10.1038/ncomms4704 (2014). 238 Hase, K. et al. Uptake through glycoprotein 2 of FimH+ bacteria by M cells initiates mucosal immune response. Nature 462, 226-230, doi:10.1038/nature08529 (2009). 239 Nakamura, Y., Kimura, S. & Hase, K. M cell-dependent antigen uptake on follicle- associated epithelium for mucosal immune surveillance. Inflamm Regen 38, 15-15, doi:10.1186/s41232-018-0072-y (2018). 240 Komban, R. J. et al. Activated Peyer's patch B cells sample antigen directly from M cells in the subepithelial dome. Nature communications 10, 2423, doi:10.1038/s41467- 019-10144-w (2019). 241 Rios, D. et al. Antigen sampling by intestinal M cells is the principal pathway initiating mucosal IgA production to commensal enteric bacteria. Mucosal immunology 9, 907- 916, doi:10.1038/mi.2015.121 (2016). 242 Owen, R. L. Uptake and transport of intestinal macromolecules and microorganisms by M cells in Peyer's patches— a personal and historical perspective. Seminars in immunology 11, 157-163, doi:https://doi.org/10.1006/smim.1999.0171 (1999). 243 Kimura, S. Molecular insights into the mechanisms of M-cell differentiation and transcytosis in the mucosa-associated lymphoid tissues. Anatomical Science International 93, 23-34, doi:10.1007/s12565-017-0418-6 (2018). 244 McDole, J. R. et al. Goblet cells deliver luminal antigen to CD103+ dendritic cells in the small intestine. Nature 483, 345-349, doi:10.1038/nature10863 (2012).

171

245 Farache, J. et al. Luminal bacteria recruit CD103+ dendritic cells into the intestinal epithelium to sample bacterial antigens for presentation. Immunity 38, 581-595, doi:10.1016/j.immuni.2013.01.009 (2013). 246 Batista, F. D. & Neuberger, M. S. Affinity dependence of the B cell response to antigen: a threshold, a ceiling, and the importance of off-rate. Immunity 8, 751-759, doi:10.1016/s1074-7613(00)80580-4 (1998). 247 Batista, F. D., Iber, D. & Neuberger, M. S. B cells acquire antigen from target cells after synapse formation. Nature 411, 489-494, doi:10.1038/35078099 (2001). 248 Carrasco, Y. R. & Batista, F. D. B cells acquire particulate antigen in a macrophage- rich area at the boundary between the follicle and the subcapsular sinus of the lymph node. Immunity 27, 160-171, doi:10.1016/j.immuni.2007.06.007 (2007). 249 Junt, T. et al. Subcapsular sinus macrophages in lymph nodes clear lymph-borne viruses and present them to antiviral B cells. Nature 450, 110-114, doi:10.1038/nature06287 (2007). 250 Qi, H., Egen, J. G., Huang, A. Y. & Germain, R. N. Extrafollicular activation of lymph node B cells by antigen-bearing dendritic cells. Science 312, 1672-1676, doi:10.1126/science.1125703 (2006). 251 Phan, T. G., Grigorova, I., Okada, T. & Cyster, J. G. Subcapsular encounter and complement-dependent transport of immune complexes by lymph node B cells. Nature immunology 8, 992-1000, doi:10.1038/ni1494 (2007). 252 Suzuki, K., Grigorova, I., Phan, T. G., Kelly, L. M. & Cyster, J. G. Visualizing B cell capture of cognate antigen from follicular dendritic cells. Journal of Experimental Medicine 206, 1485-1493, doi:10.1084/jem.20090209 (2009). 253 Pape, K. A., Catron, D. M., Itano, A. A. & Jenkins, M. K. The humoral immune response is initiated in lymph nodes by B cells that acquire soluble antigen directly in the follicles. Immunity 26, 491-502, doi:10.1016/j.immuni.2007.02.011 (2007). 254 Roozendaal, R. et al. Conduits mediate transport of low-molecular-weight antigen to lymph node follicles. Immunity 30, 264-276, doi:10.1016/j.immuni.2008.12.014 (2009). 255 Suzuki, K. et al. The Sensing of Environmental Stimuli by Follicular Dendritic Cells Promotes Immunoglobulin A Generation in the Gut. Immunity 33, 71-83, doi:https://doi.org/10.1016/j.immuni.2010.07.003 (2010). 256 Allen, C. D. & Cyster, J. G. Follicular dendritic cell networks of primary follicles and germinal centers: phenotype and function. Semin Immunol 20, 14-25, doi:10.1016/j.smim.2007.12.001 (2008). 257 Szakal, A. K., Kosco, M. H. & Tew, J. G. Microanatomy of lymphoid tissue during humoral immune responses: structure function relationships. Annu Rev Immunol 7, 91- 109, doi:10.1146/annurev.iy.07.040189.000515 (1989). 258 Wu, Y. et al. Immune complex-bearing follicular dendritic cells deliver a late antigenic signal that promotes somatic hypermutation. Journal of immunology 180, 281-290, doi:10.4049/jimmunol.180.1.281 (2008). 259 Mesin, L., Ersching, J. & Victora, G. D. Germinal Center B Cell Dynamics. Immunity 45, 471-482, doi:10.1016/j.immuni.2016.09.001 (2016). 260 Zhang, Y., Garcia-Ibanez, L. & Toellner, K.-M. Regulation of germinal center B-cell differentiation. Immunological reviews 270, 8-19, doi:10.1111/imr.12396 (2016). 261 Biram, A. et al. BCR affinity differentially regulates colonization of the subepithelial dome and infiltration into germinal centers within Peyer's patches. 20, 482-492, doi:10.1038/s41590-019-0325-1 (2019). 262 Muramatsu, M. et al. Class switch recombination and hypermutation require activation- induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102, 553-563, doi:10.1016/s0092-8674(00)00078-7 (2000). 263 Xu, Z. et al. Regulation of aicda expression and AID activity: relevance to somatic hypermutation and class switch DNA recombination. Crit Rev Immunol 27, 367-397, doi:10.1615/critrevimmunol.v27.i4.60 (2007). 264 Chaudhuri, J. et al. Transcription-targeted DNA deamination by the AID antibody diversification enzyme. Nature 422, 726-730, doi:10.1038/nature01574 (2003).

172

265 Liu, M. & Schatz, D. G. Balancing AID and DNA repair during somatic hypermutation. Trends in immunology 30, 173-181, doi:10.1016/j.it.2009.01.007 (2009). 266 Chaudhuri, J., Khuong, C. & Alt, F. W. Replication protein A interacts with AID to promote deamination of somatic hypermutation targets. Nature 430, 992-998, doi:10.1038/nature02821 (2004). 267 Stavnezer, J., Guikema, J. E. & Schrader, C. E. Mechanism and regulation of class switch recombination. Annu Rev Immunol 26, 261-292, doi:10.1146/annurev.immunol.26.021607.090248 (2008). 268 Lindner, C. et al. Diversification of memory B cells drives the continuous adaptation of secretory antibodies to gut microbiota. Nature immunology 16, 880-888, doi:10.1038/ni.3213 (2015). 269 Schramm, C. A. & Douek, D. C. Beyond Hot Spots: Biases in Antibody Somatic Hypermutation and Implications for Vaccine Design. Frontiers in immunology 9, 1876, doi:10.3389/fimmu.2018.01876 (2018). 270 Allen, C. D. C., Okada, T. & Cyster, J. G. Germinal-Center Organization and Cellular Dynamics. Immunity 27, 190-202, doi:https://doi.org/10.1016/j.immuni.2007.07.009 (2007). 271 Cerutti, A. & Rescigno, M. The biology of intestinal immunoglobulin A responses. Immunity 28, 740-750, doi:10.1016/j.immuni.2008.05.001 (2008). 272 Mora, J. R. et al. Generation of gut-homing IgA-secreting B cells by intestinal dendritic cells. Science 314, 1157-1160, doi:10.1126/science.1132742 (2006). 273 Cerutti, A. et al. Regulation of mucosal IgA responses: lessons from primary immunodeficiencies. Annals of the New York Academy of Sciences 1238, 132-144, doi:10.1111/j.1749-6632.2011.06266.x (2011). 274 Bemark, M. et al. Limited clonal relatedness between gut IgA plasma cells and memory B cells after oral immunization. Nature communications 7, 12698, doi:10.1038/ncomms12698 (2016). 275 Schwickert, T. A. et al. A dynamic T cell-limited checkpoint regulates affinity-dependent B cell entry into the germinal center. The Journal of experimental medicine 208, 1243- 1252, doi:10.1084/jem.20102477 (2011). 276 Green, J. A. & Cyster, J. G. S1PR2 links germinal center confinement and growth regulation. Immunological reviews 247, 36-51, doi:10.1111/j.1600-065X.2012.01114.x (2012). 277 Tas, J. M. et al. Visualizing antibody affinity maturation in germinal centers. Science 351, 1048-1054, doi:10.1126/science.aad3439 (2016). 278 Gatto, D. & Brink, R. The germinal center reaction. J Allergy Clin Immunol 126, 898- 907; quiz 908-899, doi:10.1016/j.jaci.2010.09.007 (2010). 279 Biram, A. et al. BCR affinity differentially regulates colonization of the subepithelial dome and infiltration into germinal centers within Peyer's patches. Nature immunology 20, 482-492, doi:10.1038/s41590-019-0325-1 (2019). 280 Zhang, Y. et al. Germinal center B cells govern their own fate via antibody feedback. The Journal of experimental medicine 210, 457-464, doi:10.1084/jem.20120150 (2013). 281 Nutt, S. L., Taubenheim, N., Hasbold, J., Corcoran, L. M. & Hodgkin, P. D. The genetic network controlling plasma cell differentiation. Semin Immunol 23, 341-349, doi:10.1016/j.smim.2011.08.010 (2011). 282 Ochiai, K. et al. Transcriptional regulation of germinal center B and plasma cell fates by dynamical control of IRF4. Immunity 38, 918-929, doi:10.1016/j.immuni.2013.04.009 (2013). 283 Kunkel, E. J. & Butcher, E. C. Plasma-cell homing. Nature reviews. Immunology 3, 822- 829, doi:10.1038/nri1203 (2003). 284 O'Connor, B. P. et al. BCMA is essential for the survival of long-lived bone marrow plasma cells. The Journal of experimental medicine 199, 91-98, doi:10.1084/jem.20031330 (2004).

173

285 Ma, Y., Galluzzi, L., Zitvogel, L. & Kroemer, G. Autophagy and Cellular Immune Responses. Immunity 39, 211-227, doi:https://doi.org/10.1016/j.immuni.2013.07.017 (2013). 286 Pengo, N. et al. Plasma cells require autophagy for sustainable immunoglobulin production. Nature immunology 14, 298-305, doi:10.1038/ni.2524 (2013). 287 Kroemer, G., Mariño, G. & Levine, B. Autophagy and the integrated stress response. Molecular cell 40, 280-293, doi:10.1016/j.molcel.2010.09.023 (2010). 288 Tiller, T. et al. Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR and expression vector cloning. J Immunol Methods 329, 112- 124, doi:10.1016/j.jim.2007.09.017 (2008). 289 Meffre, E. et al. Surrogate light chain expressing human peripheral B cells produce self- reactive antibodies. The Journal of experimental medicine 199, 145-150, doi:10.1084/jem.20031550 (2004). 290 Caporaso, J. G. et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences 108, 4516, doi:10.1073/pnas.1000080107 (2011). 291 Caporaso, J. G. et al. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. The ISME journal 6, 1621-1624, doi:10.1038/ismej.2012.8 (2012). 292 Lagkouvardos, I. et al. IMNGS: A comprehensive open resource of processed 16S rRNA microbial profiles for ecology and diversity studies. Scientific reports 6, 33721, doi:10.1038/srep33721 (2016). 293 Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods 10, 996-998, doi:10.1038/nmeth.2604 (2013). 294 Wang, Q., Garrity, G. M., Tiedje, J. M. & Cole, J. R. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and environmental microbiology 73, 5261-5267, doi:10.1128/AEM.00062-07 (2007). 295 Yoon, S. H. et al. Introducing EzBioCloud: a taxonomically united database of 16S rRNA gene sequences and whole-genome assemblies. Int J Syst Evol Microbiol 67, 1613-1617, doi:10.1099/ijsem.0.001755 (2017). 296 Lagkouvardos, I., Fischer, S., Kumar, N. & Clavel, T. Rhea: a transparent and modular R pipeline for microbial profiling based on 16S rRNA gene amplicons. PeerJ 5, e2836, doi:10.7717/peerj.2836 (2017). 297 Chen, J. et al. Associating microbiome composition with environmental covariates using generalized UniFrac distances. Bioinformatics (Oxford, England) 28, 2106-2113, doi:10.1093/bioinformatics/bts342 (2012). 298 Cyster, J. G. & Schwab, S. R. Sphingosine-1-phosphate and lymphocyte egress from lymphoid organs. Annu Rev Immunol 30, 69-94, doi:10.1146/annurev-immunol- 020711-075011 (2012). 299 Sterlin, D. et al. Human IgA binds a diverse array of commensal bacteria. The Journal of experimental medicine 217, doi:10.1084/jem.20181635 (2020). 300 Bunker, J. J. et al. B cell superantigens in the human intestinal microbiota. Sci Transl Med 11, doi:10.1126/scitranslmed.aau9356 (2019). 301 Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic acids research 32, 1792-1797, doi:10.1093/nar/gkh340 (2004). 302 Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree 2--approximately maximum-likelihood trees for large alignments. PloS one 5, e9490, doi:10.1371/journal.pone.0009490 (2010). 303 Eberl, C. et al. Reproducible Colonization of Germ-Free Mice With the Oligo-Mouse- Microbiota in Different Animal Facilities. Frontiers in Microbiology 10, doi:10.3389/fmicb.2019.02999 (2020). 304 Dogan, I. et al. Multiple layers of B cell memory with different effector functions. Nature immunology 10, 1292-1299, doi:10.1038/ni.1814 (2009). 305 Rollenske, T. et al. Cross-specificity of protective human antibodies against Klebsiella pneumoniae LPS O-antigen. Nature immunology 19, 617-624, doi:10.1038/s41590- 018-0106-2 (2018). 174

306 Kabbert, J. et al. High microbiota reactivity of adult human intestinal IgA requires somatic mutations. Journal of Experimental Medicine 217, doi:10.1084/jem.20200275 (2020). 307 Sheng, Z. et al. Gene-Specific Substitution Profiles Describe the Types and Frequencies of Amino Acid Changes during Antibody Somatic Hypermutation. Frontiers in immunology 8, doi:10.3389/fimmu.2017.00537 (2017). 308 Gunti, S. & Notkins, A. L. Polyreactive Antibodies: Function and Quantification. The Journal of infectious diseases 212 Suppl 1, S42-46, doi:10.1093/infdis/jiu512 (2015). 309 Zhou, Z.-H., Tzioufas, A. G. & Notkins, A. L. Properties and function of polyreactive antibodies and polyreactive antigen-binding B cells. J Autoimmun 29, 219-228, doi:10.1016/j.jaut.2007.07.015 (2007). 310 Michael, J. G., Whitby, J. L. & Landy, M. Studies on natural antibodies to gram-negative bacteria. The Journal of experimental medicine 115, 131-146, doi:10.1084/jem.115.1.131 (1962). 311 Chai, J. N., Zhou, Y. W. & Hsieh, C. S. T cells and intestinal commensal bacteria-- ignorance, rejection, and acceptance. FEBS letters 588, 4167-4175, doi:10.1016/j.febslet.2014.06.040 (2014). 312 Tiller, T. et al. Autoreactivity in human IgG+ memory B cells. Immunity 26, 205-213, doi:10.1016/j.immuni.2007.01.009 (2007). 313 Meffre, E. & Wardemann, H. B-cell tolerance checkpoints in health and autoimmunity. Current opinion in immunology 20, 632-638, doi:https://doi.org/10.1016/j.coi.2008.09.001 (2008). 314 Hughes, G. J., Reason, A. J., Savoy, L., Jaton, J. & Frutiger-Hughes, S. Carbohydrate moieties in human secretory component. Biochim Biophys Acta 1434, 86-93 (1999). 315 McLoughlin, K., Schluter, J., Rakoff-Nahoum, S., Smith, A. L. & Foster, K. R. Host Selection of Microbiota via Differential Adhesion. Cell host & microbe 19, 550-559, doi:10.1016/j.chom.2016.02.021 (2016). 316 Macpherson, A. J. & Uhr, T. Induction of protective IgA by intestinal dendritic cells carrying commensal bacteria. Science 303, 1662-1665, doi:10.1126/science.1091334 (2004). 317 Tsui, F. P. et al. Determination of the structure of the Escherichia coli K100 capsular polysaccharide, cross-reactive with the capsule from type b Haemophilus influenzae. Carbohydr Res 173, 65-74 (1988). 318 Chen, Y. et al. Microbial symbionts regulate the primary Ig repertoire. The Journal of experimental medicine 215, 1397-1415, doi:10.1084/jem.20171761 (2018). 319 Yang, C. et al. Fecal IgA Levels Are Determined by Strain-Level Differences in Bacteroides ovatus and Are Modifiable by Gut Microbiota Manipulation. Cell host & microbe 27, 467-475 e466, doi:10.1016/j.chom.2020.01.016 (2020). 320 Okai, S. et al. High-affinity monoclonal IgA regulates gut microbiota and prevents colitis in mice. Nat Microbiol 1, 16103, doi:10.1038/nmicrobiol.2016.103 (2016). 321 Ivanov, I. I. et al. Induction of intestinal Th17 cells by segmented filamentous bacteria. Cell 139, 485-498, doi:10.1016/j.cell.2009.09.033 (2009). 322 Dijkshoorn, L., Ursing, B. M. & Ursing, J. B. Strain, clone and species: comments on three basic concepts of bacteriology. Journal of medical microbiology 49, 397-401, doi:10.1099/0022-1317-49-5-397 (2000). 323 Suzuki, K. et al. Aberrant expansion of segmented filamentous bacteria in IgA-deficient gut. Proceedings of the National Academy of Sciences of the United States of America 101, 1981-1986, doi:10.1073/pnas.0307317101 (2004). 324 Gaboriau-Routhiau, V. et al. The key role of segmented filamentous bacteria in the coordinated maturation of gut helper T cell responses. Immunity 31, 677-689, doi:10.1016/j.immuni.2009.08.020 (2009). 325 Fadlallah, J. et al. Synergistic convergence of microbiota-specific systemic IgG and secretory IgA. J Allergy Clin Immunol 143, 1575-1585.e1574, doi:10.1016/j.jaci.2018.09.036 (2019). 326 Bry, L. & Brenner, M. B. Critical Role of T Cell-Dependent Serum Antibody, but Not the Gut-Associated Lymphoid Tissue, for Surviving Acute Mucosal Infection with 175

<em>Citrobacter rodentium</em>, an Attaching and Effacing Pathogen. The Journal of Immunology 172, 433, doi:10.4049/jimmunol.172.1.433 (2004). 327 Castro-Dopico, T. & Clatworthy, M. R. IgG and Fcγ Receptors in Intestinal Immunity and Inflammation. Frontiers in immunology 10, 805-805, doi:10.3389/fimmu.2019.00805 (2019). 328 Bakema, J. E. & van Egmond, M. The human immunoglobulin A Fc receptor FcalphaRI: a multifaceted regulator of mucosal immunity. Mucosal immunology 4, 612-624, doi:10.1038/mi.2011.36 (2011). 329 Monteiro, R. C. & Van De Winkel, J. G. IgA Fc receptors. Annu Rev Immunol 21, 177- 204, doi:10.1146/annurev.immunol.21.120601.141011 (2003). 330 Smith, P. D. et al. Intestinal macrophages lack CD14 and CD89 and consequently are down-regulated for LPS- and IgA-mediated activities. Journal of immunology 167, 2651-2656, doi:10.4049/jimmunol.167.5.2651 (2001). 331 Bain, C. C. et al. Resident and pro-inflammatory macrophages in the colon represent alternative context-dependent fates of the same Ly6Chi monocyte precursors. Mucosal immunology 6, 498-510, doi:10.1038/mi.2012.89 (2013).

176

Eidesstattliche Erklärung/Declaration of academic honesty Ich, Johanna Kabbert, erkläre hiermit, dass diese Dissertation und die darin dargelegten Inhalte die eigenen sind und selbstständig, als Ergebnis der eigenen originären Forschung, generiert wurden. I, Johanna Kabbert, hereby declare that this thesis and the presented work in it is original and has been generated by myself as the result of my own original research.

Hiermit erkläre ich an Eides statt:

1. Diese Arbeit wurde vollständig oder größtenteils in der Phase als Doktorand dieser Fakultät und Universität angefertigt;

2. Sofern irgendein Bestandteil dieser Dissertation zuvor für einen akademischen Abschluss oder eine andere Qualifikation an dieser oder einer anderen Institution verwendet wurde, wurde dies klar angezeigt;

3. Wenn immer andere eigene- oder Veröffentlichungen Dritter herangezogen wurden, wurden diese klar benannt;

4. Wenn aus anderen eigenen- oder Veröffentlichungen Dritter zitiert wurde, wurde stets die Quelle hierfür angegeben. Diese Dissertation ist vollständig meine eigene Arbeit, mit der Ausnahme solcher Zitate;

5. Alle wesentlichen Quellen von Unterstützung wurden benannt;

6. Wenn immer ein Teil dieser Dissertation auf der Zusammenarbeit mit anderen basiert, wurde von mir klar gekennzeichnet, was von anderen und was von mir selbst erarbeitet wurde;

7. Ein Teil dieser Arbeit wurden zuvor veröffentlicht in: Kabbert J, Benckert J, Rollenske T, Hitch C.A., Clavel T, Cerovic V, Wardemann H, Pabst O. High micorbiota reactivity of adult intestinal IgA requires somatic mutations. J Exp Med (2020) 217 (11): e20200275.

Datum:

Unterschrift:

177