Ubiquitination of chromatin by modification of histone H2A is a critical step in both regulation of DNA repair and regulation of cell fate. These very different outcomes depend on the selective modification of distinct lysine residues in H2A, each by a specific E3 ligase. While polycomb PRC1 complexes modify K119, resulting in gene silencing, the E3 ligase RNF168 modifies K13/15, which is a key event in the response to DNA double-strand breaks. The molecular origin of ubiquitination site specificity by these related E3 enzymes is one of the open questions in the field. Using a combination of NMR spectroscopy, crosslinking mass-spectrometry, mutagenesis and data-driven modelling, here we show that RNF168 binds the acidic patch on the nucleosome surface, directing the E2 to the target lysine. The structural model highlights the role of E3 and nucleosome in promoting ubiquitination and provides a basis for understanding and engineering of chromatin ubiquitination specificity.

The nucleosome surface contains an area with negative electrostatic potential known as the acidic patch, which functions as a binding platform for various proteins to regulate chromatin biology. The dense clustering of acidic residues may impact their effective pKa and thus the electronegativity of the acidic patch, which in turn could influence nucleosome-protein interactions. We here set out to determine the pKa values of residues in and around the acidic patch in the free H2A-H2B dimer using NMR spectroscopy. We present a refined solution structure of the H2A-H2B dimer based on intermolecular distance restraints, displaying a well-defined histone-fold core. We show that the conserved histidines H2B H46 and H106 that line the acidic patch have pKa of 5.9 and 6.5, respectively, and that most acidic patch carboxyl groups have pKa values well below 5.0. For H2A D89 we find strong evidence for an elevated pKa of 5.3. Our data establish that the acidic patch is highly negatively charged at physiological pH, while protonation of H2B H106 and H2B H46 at slightly acidic pH will reduce electronegativity. These results will be valuable to understand the impact of pH changes on nucleosome-protein interactions in vitro, in silico or in vivo.