Indeed, loop regions in HIV-1 cDNA have been shown to be preferred sites for A3G-mediated DNA deamination57, and a recent study of 560 breast cancer genomes reported multiple hotspots for APOBEC signature mutations within loop regions of predicated stem-loop structures58. vertebrates also have APOBEC1, which edits cytosine nucleobases in RNA and ssDNA and functions in regulating the transcriptome and likely also in blocking the spread of endogenous and exogenous mobile elements including viruses5, 6. The APOBEC3 subfamily of enzymes is specific to mammals, subject to extreme copy number variation, elicits strong preferences for ssDNA, and provides innate immune protection against a wide variety of DNA-based parasites including common retrotransposons L1 and Alu and retroviruses such as HIV-12, 7, 8. Human cells have the potential to produce up to 7 distinct APOBEC3 enzymes, APOBEC3A through APOBEC3H (A3A-A3H, excluding A3E), Rabbit Polyclonal to CDC25C (phospho-Ser198) though most cells express subsets due to differential gene regulation912. The local substrate preference of each of these enzymes is an intrinsic property that has helped to elucidate multiple biological and pathological functions including the elucidation of AID as an antibody gene DNA deaminase3, 4, the delineation of the subset of APOBEC3 enzymes responsible for HIV-1 hypermutation (A3D, A3F, A3G, and A3H)2, 7, 8and, recently, the implication of at least one APOBEC family member in mutagenesis in a wide variety of cancers1315. The nucleobase immediately 5 of the target cytosine (1 relative to the target cytosine at position 0) is the most important determinant of each enzymes intrinsic substrate preference1619. AID preferentially deaminates single-stranded AM 103 DNA cytosine bases preceded by an adenine or guanine (5-RC), corresponding to cytosine mutation spectra in immunoglobulin gene variable and switch regions. A3G uniquely targets cytosine bases preceded by another cytosine (5-CC), a pattern that is evident in patient-derived HIV-1 sequences. APOBEC1 and the remaining APOBEC3 enzymes elicit preferences for cytosine bases preceded by a thymine (5-TC). The 1 nucleobase AM 103 preference is governed largely by a loop adjacent to the AM 103 AM 103 active site (loop 7), such that loop exchanges can convert one enzymes intrinsic preference into that of another (e. g., A3G with loop 7 from A3A becomes a 5-TCpreferring enzyme17). The 2 and +1 nucleobases relative to the target cytosine are also likely to be involved in local target selection but are less influential. The APOBEC mutation signature in cancer has been defined as C-to-T and C-to-G base substitution mutations within 5-TCdinucleotide motifs and most commonly within 5-TCA and 5-TCT trinucleotide contexts (also 5-TCG if accounting for the genomic under-representation of this motif12, 20). Unlike AID in antibody gene diversification and A3G in HIV-1 hypermutation, absolute assignments of cause and effect in cancer cannot be made based on intrinsic ssDNA deamination preferences alone because most APOBEC family members prefer substrates with 5-TCmotifs and the potential for additional specificity-conferring enzyme-nucleobase contacts is poorly understood. Thus, corroborating data sets including expression profiles, biochemical activities, global mutation analysis, animal experiments, and clinical information are being combined to deduce which subset of enzymes is actually responsible for the observed APOBEC signature mutations in cancer. The leading candidates to explain APOBEC mutagenesis in cancer are A3A2123, A3B12, 20, 24, and A3H20. However , to date, only A3B has met all experimental criteria including clinical correlations between high expression levels and poor outcomes for multiple cancer types2529. Despite the importance of APOBEC3-mediated mutations in a variety of biological and pathological processes, the molecular mechanisms underlying global ssDNA binding activity and local target selectivity are poorly understood. In particular, prior structural studies have yielded numerous apo-enzyme structurese. g., 3040as well as two structures with non-substrate nucleotides37, 41, but complexes with relevant ssDNA substrates have proven elusive. Here we solve crystal structures of human A3A and a variant of the human A3B catalytic domain in complex with optimal and commonly mutated ssDNA substrates,.