Insights into the Structures and Multimeric Status of APOBEC Proteins Involved in Viral Restriction and Other Cellular Functions

Abstract

Apolipoprotein B mRNA editing catalytic polypeptide-like (APOBEC) proteins belong to a family of deaminase proteins that can catalyze the deamination of cytosine to uracil on single-stranded DNA or/and RNA. APOBEC proteins are involved in diverse biological functions, including adaptive and innate immunity, which are critical for restricting viral infection and endogenous retroelements. Dysregulation of their functions can cause undesired genomic mutations and RNA modification, leading to various associated diseases, such as hyper-IgM syndrome and cancer. This review focuses on the structural and biochemical data on the multimerization status of individual APOBECs and the associated functional implications. Many APOBECs form various multimeric complexes, and multimerization is an important way to regulate functions for some of these proteins at several levels, such as deaminase activity, protein stability, subcellular localization, protein storage and activation, virion packaging, and antiviral activity. The multimerization of some APOBECs is more complicated than others, due to the associated complex RNA binding modes.

Keywords: innate and acquired immunity; multimerization or oligomerization; mutation and cancer; structure; viral restriction.

Figure 1

The apolipoprotein B mRNA editing catalytic polypeptide-like (APOBEC) family members in humans. (A) Members of the APOBEC family contain either one or two cytidine deaminase (CD) domains. Even though all CD domains contain the canonical Zn-center motif of H-[P/A/V]-E-X_[23–28]-P-C-X2-4-C (x is any amino acid), some are catalytically active (indicated by a red star), and some are catalytically inactive (yellow star). Summary of the characterized biological functions of different members are shown on the right side. (B) The APOBEC2 (A2) core structure (PDBid 2nyt, containing residue 40–224)) showing the typical APOBEC core CD fold, which contains a five-stranded beta-sheet and six alpha-helices. Inset shows the highly conserved 3D arrangement of the Zn-center among all APOBECs.

Figure 2

The structure of A1 monomer (A), dimer (B), and the surface charge feature of a dimer (C) (PDBid 6X91). A1 has a unique C-terminal domain A1HD (A1 hydrophobic domain) that mediates the protein-protein dimer formation. A long strip of positively charged surface is formed through dimerization, which may be used for recruiting nucleic acids or other protein partners.

Figure 3

Human A2 crystal structure. (A) The crystal structure of A2-core tetramer, dimer, and monomer (PDBid 2nyt, residues 40–224). (B) Molecular weights from size-exclusion chromatography coupled with multi-angle light dynamic scattering (SEC-MALS) data for the A2 (full-length, 26 kD for a monomer) at a protein concentration of 1 mg/mL. Four peaks were observed, corresponding to the approximate molecular weight of a monomer (24 kD), dimer (56 kD), trimer (85 kD), and tetramer (95 kD). The dimer species is the predominant form. (C) Time course of glutaraldehyde cross-linking with the A2 protein. Reactions with A2 (2 µg total protein) were performed for the indicated time, with 0.25% glutaraldehyde at room temperature, 25 mM HEPES (pH 7.0), 160 mM NaCl, and 10% glycerol; they were then quenched with 1 M Tris, pH 8.5, with 2X SDS loading buffer and run on a 12% SDS-PAGE gel for 70 min, at 200 V. Proteins were visualized by Coomassie staining. Cross-linking reveals four bands, corresponding to a monomer, dimer, trimer, and tetramer band, respectively. Monomer and dimer appear to be the predominant forms under this cross-linking condition. Both full-length (residues 1–224) and the N-terminal truncated A2-core (residues 40–224) showed similar results. Note: Panels B and C are to-be-published data.

Figure 4

Surface charge of APOBECs CD domains. The general orientation for all panels is shown as the ribbon structure. Among all catalytic CD domains (not including CD1 of A3s), only A3A and A3H have their Zn-center embedded within positively charged surfaces (PCS, in blue). A2 and A3Gcd2 have the least PCS among single-domain and double-domain APOBECs, respectively. It is worth noting that A2, A3Gcd2, and A3A are the only three proteins that do not form large RNA-bound multimeric forms in cell lysates (data not shown). Note: A3Fcd1 structure is to be published structure. PDBids for other structures are A1, 6X91; A2-core, 2nyt; AID, 5W1C; A3A, 5KEG; A3C, 3VOW; A3H, 5W3V; A3Bcd1, 5TKM; A3Bcd2, 5CQI; A3Fcd2, 3WUS; A3Gcd1, 5K83; A3Gcd2, 3IQS.

Figure 5

The structure of A3A and the two packed A3C molecules in the crystal asymmetric unit. (A) The dimer structure of A3A without binding ssDNA (PDBid: 4XXO). The dimerization is mediated mainly through the N-terminal arm exchange between two subunits (colored in brown and green), which is further cemented by two Zn-atoms (Zn5, Zn6) coordinated by residues from two subunits. The Zn1 and Zn2 are active site Zn-atoms. (B) The SEC assay of wildtype A3A purified from E. coli expression system on Superdex 75 at a protein concentration of 2 mg/mL, showing monomer and dimer peaks in a buffer containing 20 mM Tris-HCl, 250 mM NaCl (pH 8.0), 1 mM DTT, and 1 mM EDTA (to be published data). The small peak at the left side is void volume (aggregated A3A protein). No RNA can be detected in any of the peaks. (C) The structure of ssDNA-bound A3A as monomer (DPBid: 5KEG, 5SWW) that is superimposed on to the two subunits of the apo dimer structure shown in (A). This superimposition does not show additional interaction of the bound ssDNA at the two Zn-center of each monomer with any residues from the other monomer, suggesting the short ssDNA bound to the active center does not need a dimer. However, these structural data alone cannot exclude the possibility that longer ssDNA binding to one Zn-center can bind to the other monomer to enhance the binding. (D) The crystal structure of A3C (PDBid: 3VOW), showing two A3C molecules packed into one asymmetric unit, which has the largest intermolecular contact interface in the crystal packing. The positions of the S188I variant (S188 side chain) and N115K (N115 side chain) are shown in sticks.

Figure 6

The crystal structures of A3H and dimer formation through dsRNA binding. (A) The A3H dimer (5W3V) with two subunits being connected through a dsRNA bound in between. No protein-protein contact exists between the two subunits within a dimer. (B) Superimposition of three available A3H dimers from three different organisms (5W3V, 6BBO, 5Z98), showing highly conserved dimerization mechanisms. (C) Superimposition of apo-A3H monomer structure with one subunit from the A3H dimers (5W45, 5W3V, and 5Z98), showing that conformational differences only for loops 1, 3, and 7, which are directly involved in dsRNA binding. (D) Superimposition of apo-A3H (5W45) and apo-A3C (3VOW) reveal that A3H has two extra alpha-helical turns for h6, compared to A3C and all other known APOBEC structures.

Figure 7

Structure of full-length rA3G. (A) The rA3G dimer structure (6P3X). Dimerization is mediated through CD1-CD1 contacts centered around helix 6 (h6) of CD1. (B) The positively charged residues centered around R24 on both sides of the dimer junction. Dimerization brings these positively charged residues from each subunit to close proximity. (C) The closely positioned charged residues around the dimer junction result in significantly enhanced positively charged electrostatic potentials (PEP or diPEP), as shown by the blue surface (+8.3 kT/e). (D) The surface charge potentials of a monomer at the same area around R24, revealing much less positively charged electrostatic potentials (+1.9 kT/e), as compared to those of a dimer.