Student Theses and Dissertations


Tatsuya Araki

Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

RU Laboratory

Victora Laboratory


Because high-affinity antibodies produced by germinal centers (GCs) are critical in protecting against the ever-increasing array of pathogens, understanding how GC response allows our immune system to recognize a variety of pathogens and develop specific and effective immunity against each pathogen encounters are important. We gained insight into how GCs modulate clonal diversity in recent years, including a discovery that individual B cells clones can occasionally undergo “clonal bursts” that eliminate clonal diversity through massive proliferation. However, it remains unclear what are the factors driving such bursts, or whether GCs respond reproducibly to similar magnitude increases in affinity, mainly because answering such a question would require an experiment that repeats the same GC responses many times to disentangle affinity-based GC selection events from stochastic “noise” influenced by local circumstances. In this thesis, I investigated how GCs might respond to given increases in affinity by exploring how stochasticity and determinism work together in GCs to shape the course of affinity maturation and modulate the resulting clonal diversity. In the first part of this thesis, I demonstrate that a B cell clone (clone 2.1) specific for the model antigen chicken IgY (chIgY) that previously triggered a clonal burst has potential to increase its affinity much further than the 7-fold jump that triggered the burst itself. Affinity maturation in vivo can generate up to a ~90-fold increase in affinity over 10 weeks of immune response, whereas a ~380-fold increase in affinity was achieved by artificially engineering the combination of mutations that has previously occurred in GCs. In the second part of this thesis, I take a reductionist approach to generate reproducible GC responses by starting GCs with the same chIgY-specific B cells so that many GCs with the same starting condition can be “replayed” and analyzed repeatedly for their outcomes in terms of selection. In the third part of this thesis, I demonstrate that phylogenetic trees structures have low reproducibility based on statistical measures that we employed to describe tree shapes. The outcome of replay GCs were found to be as diverse as polyclonal responses. In the fourth part of this thesis, I interrogate the impact of all possible individual amino acid (aa) mutations on antibody binding and expression of clone 2.1 through deep mutational scanning (DMS) and cryo-electron microscopy (CryoEM). DMS elucidated that chIgY B cells must avoid ~13 deleterious mutations in order to make one affinityincreasing mutations. CryoEM revealed that central paratope of clone 2.1 is largely optimal in its ability to bind chIgY, but it can be further improved by changing the periphery of its contact residues or leverage allosteric fine-tuning to optimize antibody geometry. In the fifth part of this thesis, I investigate how GCs respond to affinity improvements by combining DMS data to replay trees. In contrast to the broad diversity of outcomes observed at the phylogenetic level, GCs were found to reproducibly increase their median affinity. However, clonal bursting was not restricted to B cells with the highest affinity increases, and successful affinity maturation was not favored by specific tree shapes, suggesting that clonal bursts are not a primary driver of affinity increases in GCs. Instead, affinity maturation of clone 2.1 is driven primarily by elimination of B cells that largely lost affinity and marginally stronger expansion of B cells with small improvements in affinity, especially at earlier time points.


A Thesis Presented to the Faculty of The Rockefeller University in Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy

Available for download on Tuesday, October 17, 2023

Included in

Life Sciences Commons