Systematic analysis of the sequence-function relationship underlying differentiation and dedifferentiation activity of SOX17

Abstract

SOX17 has long been regarded as a critical transcription factor for endoderm formation in embryonic development. Recent studies have revealed that SOX17 is also important for human primordial germ cell (PGC) development from the mesendoderm. It has been suggested that OCT4 and SOX17 cooperate to facilitate early PGC induction; however, the specifics regarding the molecular mechanisms underlying their cooperation on the compressed motif for PGC specification have yet to be elucidated. More effective tools used to study early in vitro gametogenesis are therefore needed. Previously, it was discovered that OCT4-SOX17 interactions can be altered by engineering SOX17. A change in binding preference from a compressed DNA motif on which OCT4-SOX17 dimers form, to a canonical DNA motif on which OCT4-SOX2 dimers form, allow engineered SOX17 to induce pluripotency reprogramming. The aim of this project is thus to pre-select for SOX17 high mobility group (HMG) box variants favouring the compressed motif by eliminating those competent for pluripotency reprogramming. To evaluate strategies for the identification of new SOX17 variants and to decode the sequence-function relationships of SOX17, the well-established mouse pluripotency reprogramming system was used. Drug-inducible lentiviral vectors containing pooled mutagenesis libraries were screened in mouse embryonic fibroblasts (MEFs). The results showed that the combinatorial SOX17 HMG box mutagenesis libraries were highly competent for pluripotency reprogramming of MEFs, and implicated position 57 as being key for altering SOX17 activity. Subsequent deep mutational scanning of the full 79 amino acid-long SOX17 HMG box in MEFs did not reveal any other potential candidate positions. A deep dive into the implications of mutating position 57 revealed that all non-acidic and non-proline residues were able to fully reprogram MEFs and adult human dermal fibroblasts. Despite being unable to reprogram fibroblasts, SOX17E57D and SOX17E57P still preferred binding the canonical over the compressed motif with OCT4. Interestingly, SOX17E57D was found to be capable of activating endodermal gene expression in both MEFs and mESCs, the former of which was not observed for wild-type SOX17. It was therefore concluded that having a glutamic acid at position 57 was critical for ensuring that SOX17 would still favour the compressed motif. Thus, future library screening of SOX17 would do well to adopt a variation on the deep mutational scanning approach by removing SOX17E57 variants from the associated library. In ensuring base levels of preference for the compressed motif, engineered SOX17 variants that work well with OCT4 on the compressed motif to facilitate PGC development would be more likely identified. By evaluating how changes in the SOX17 HMG box shape preferences for PGC development, research into later-stage PGC development would be facilitated. Engineered SOX17 variants could therefore lead to the faithful and robust specification of the human germline, eventually enabling in vitro generation of mature human gametes.