A Novel Regulator of Activation-Induced Cytidine Deaminase/APOBECs in Immunity and Cancer: Schrödinger's CATalytic Pocket

Abstract

Activation-induced cytidine deaminase (AID) and its relative APOBEC3 cytidine deaminases boost immune response by mutating immune or viral genes. Because of their genome-mutating activities, AID/APOBECs are also drivers of tumorigenesis. Due to highly charged surfaces, extensive non-specific protein-protein/nucleic acid interactions, formation of polydisperse oligomers, and general insolubility, structure elucidation of these proteins by X-ray crystallography and NMR has been challenging. Hence, almost all available AID/APOBEC structures are of mutated and/or truncated versions. In 2015, we reported a functional structure for AID using a combined computational-biochemical approach. In so doing, we described a new regulatory mechanism that is a first for human DNA/RNA-editing enzymes. This mechanism involves dynamic closure of the catalytic pocket. Subsequent X-ray and NMR studies confirmed our discovery by showing that other APOBEC3s also close their catalytic pockets. Here, we highlight catalytic pocket closure as an emerging and important regulatory mechanism of AID/APOBEC3s. We focus on three sub-topics: first, we propose that variable pocket closure rates across AID/APOBEC3s underlie differential activity in immunity and cancer and review supporting evidence. Second, we discuss dynamic pocket closure as an ever-present internal regulator, in contrast to other proposed regulatory mechanisms that involve extrinsic binding partners. Third, we compare the merits of classical approaches of X-ray and NMR, with that of emerging computational-biochemical approaches, for structural elucidation specifically for AID/APOBEC3s.

Keywords: DNA mutations; HIV infections; antibodies; enzymes and coenzymes; leukocytes; lymphoma.

Figure 1

Core architecture, catalytic pocket occlusion, and computational/biochemical approaches to solving activation-induced cytidine deaminase (AID)/APOBEC structures. (A) Representative ribbon structure (left) and surface topology (right) of AID modeled from APOBEC templates. In the ribbon structure, N- to C-termini progression is shown from blue to red and the gray sphere depicts active site zinc. In the surface topology, positive, negative, and neutral residues have blue, red, and white surfaces, respectively. The Zn-coordinating residues and catalytic glutamic acid surface are colored purple. A distinct feature of AID among the APOBECs is its high positive charge at neutral pH, concentrated along two single-stranded DNA (ssDNA)-binding grooves that pass over the catalytic pocket. (B) Ribbon structures of A3A (transparent) and AID (non-transparent) were superimposed. In each protein structure, the secondary catalytic loops 2, 4, 6, and 8 are colored red, orange, green, and blue, respectively. (C) Catalytically accessible (left), partially occluded (middle), and catalytically restricted (right) conformations of A3A (top), A3B-CTD (center), and AID (bottom). The surface of secondary catalytic loop 2, 4, 6, and 8 were colored red, orange, green, and blue, respectively, to correspond with the ribbon structure shown in panel B. Catalytically accessible conformations are shown with bound dC in the catalytic pocket. Conformations were deemed catalytically accessible if they bound dC in a deamination-feasible configuration in the catalytic pocket via molecular docking [AutoDock VINA (59)]. In catalytically restricted conformations, the secondary catalytic loops adopt a configuration that block the pocket. (D) Proportion of catalytically restricted (red), partially occluded (purple), and catalytically accessible (green) conformations in A3A (top), A3B-CTD (center), and AID (bottom). A3A showed a dramatically higher proportion of catalytically accessible conformations in comparison to A3B-CTD and AID. NMR conformations of A3A (PDB: 2M65), A3B (PDB: 2NBQ), and previously reported structures of AID (58) were used. (E) Combinatorial computational/biochemical approach for solution of functional and native enzyme structures. A library of thousands of predicted structures is generated through homology modeling with a range of suitable template structures, generating multiple low energy conformations. The resulting conformational ensemble is then evaluated mathematically (e.g., Ramachandran and other means of evaluating model quality). Models are also checked for concordance with known biochemical properties of the enzyme. Molecular docking can be used to determine the substrate binding regions of the active site and surrounding regions. Concurrently, specific hypotheses are formed based on the highest confidence predicted conformations and their interaction with substrate. To test these hypotheses and to validate the positions and relative attitudes of specific core or surface residues, a large variant library ought to be constructed and tested in functional enzyme assays. This library can include point mutants, multiple mutants, orthologous and chimeric versions of the enzyme. For key residues involved in catalysis regulation, several point mutations spanning the range of synonymous to severe are more informative. Functional evaluation of this variant library ought to be used to confirm the involvement of key residues/motifs in specific biochemical aspects such as substrate binding, catalysis, and structural stability. Collectively, information from functional testing of the variant library is used to refine and validate the predicted enzyme structure and its interactions with substrate, to yield a functional and native structure.