Excited to share a story about mutations in the IRF4 TF DNA binding domain that alter its binding specificity to cause a dominant form of immunodeficiency. Great work led by @TurveyLab https://www.science.org/doi/10.1126/sciimmunol.ade7953 1/12
The work reflects a virtuoso bioinformatics effort by @ofornes that started for us with an email about a single amino acid change in IRF4 that was observed in a child with immune problems 2/12
IRF4 DNA binding is complex as it partners with many different TFs for binding. On its own IRF4 binds to ISRE motifs, but in partnership targets composite sites - EICEs (ETS partnered) and AICEs (AP1 partnered) 3/12
After structural modeling, @ofornes suggested that the introduction of a positively charged amino acid at the position could cause increased DNA affinity, which would alter TF-DNA interactions 4/12
In an amazing effort spanning many countries and disciplines, more patients were discovered and extensive profiling of mutant IRF4 binding was generated – so many partners stepped up 5/12
to show that the mutations were indeed causing IRF4 to stick to a much broader range of DNA sequences 6/12
The binding data showed in particular that the mutation was allowing the TF to stick to novel target sequences, most strikingly those having a GATA instead of a GATA in the motif 7/12
ExplaiNN provided great insight into the wide range of IRF4 binding motifs, returning Importance scores that clearly highlight the novel motifs targeted by IRF4 9/12
I think this is a dramatic example of how disrupting the binding of a TF can have widespread impact on regulatory programs in cells and human phenotypes 10/12
It is also a great display of bioinformatics being well applied at the intersection of rare disease and gene regulation. Very pleased to have contributed 11/12
With deep appreciation to all of the participating scientists and clinicians in the vast IRF4 International Consortium. And most importantly to the patients and families for engaging in the research process. 12/12
I have focused on the DNA binding and bioinformatics, but read the paper for detailed characterization of the disease phenotype, transgenic mouse models and much more 13/12
And the mandatory typo --- GATA instead of GAAA
Pulling together the diverse types of binding data compiled, @ofornes used our ExplaiNN system (with input from @NovakovskyG ) to explore the relative contribution of the diverse IRF4 target sites observed … 8/12
https://www.biorxiv.org/content/10.1101/2022.05.20.492818v3