APOBECs and Herpesviruses

Abstract

The APOBEC family of DNA cytosine deaminases provides a broad and overlapping defense against viral infections. Successful viral pathogens, by definition, have evolved strategies to escape restriction by the APOBEC enzymes of their hosts. HIV-1 and related retroviruses are thought to be the predominant natural substrates of APOBEC enzymes due to obligate single-stranded DNA replication intermediates, abundant evidence for cDNA strand C-to-U editing (genomic strand G-to-A hypermutation), and a potent APOBEC degradation mechanism. In contrast, much lower mutation rates are observed in double-stranded DNA herpesviruses and the evidence for APOBEC mutation has been less compelling. However, recent work has revealed that Epstein-Barr virus (EBV), Kaposi's sarcoma herpesvirus (KSHV), and herpes simplex virus-1 (HSV-1) are potential substrates for cellular APOBEC enzymes. To prevent APOBEC-mediated restriction these viruses have repurposed their ribonucleotide reductase (RNR) large subunits to directly bind, inhibit, and relocalize at least two distinct APOBEC enzymes - APOBEC3B and APOBEC3A. The importance of this interaction is evidenced by genetic inactivation of the EBV RNR (BORF2), which results in lower viral infectivity and higher levels of C/G-to-T/A hypermutation. This RNR-mediated mechanism therefore likely functions to protect lytic phase viral DNA replication intermediates from APOBEC-catalyzed DNA C-to-U deamination. The RNR-APOBEC interaction defines a new host-pathogen conflict that the virus must win in real-time for transmission and pathogenesis. However, partial losses over evolutionary time may also benefit the virus by providing mutational fuel for adaptation.

Keywords: APOBEC; DNA cytosine deamination; DNA editing; evolution; herpesvirus; innate antiviral immunity; mutation; restriction factors; ribonucleotide reductase.

Figure 2

Human APOBEC3 enzymes. (A) The human A3 locus is comprised of seven tandemly arranged genes flanked by CBX6 and CBX7 on chromosome 22. Each A3 enzyme has a characteristic subcellular localization (e.g., representative images of U2OS cells expressing the indicated A3-mCherry constructs, colored green to be consistent with Figure 3; A.Z.C. and R.S.H., unpublished). (B) Model of the wild-type human A3B catalytic domain bound to ssDNA [43] based on crystal structures of an A3B catalytic domain variant and A3A bound to optimal ssDNA substrates [41]. The deamination preference of A3B for 5’-TC ssDNA substrates (orange and green) is governed by loop 7 residues (purple). A single zinc ion (blue) is located in the catalytic pocket adjacent to the target cytosine nucleobase (green). (C) Schematic of the ssDNA cytosine to uracil deamination mechanism, in which A3 enzymes accelerate the hydrolytic replacement of the amine group with oxygen from water.

Figure 3

Conservation of herpesvirus RNR activities. (A) Phylogram of human herpesviruses based on RNR large subunit amino acid sequences. Gamma-, alpha, and beta-herpesvirus groups are highlighted in red, green, and orange, respectively (EBV BORF2 YP_001129452.1; KSHV ORF61 YP_001129418.1; VZV ORF19 NP_040142.1; HSV-1 ICP6 YP_009137114.1; HSV-2 UL39 YP_009137191.1; HCMV UL45 YP_081503.1; HHV-7 U28 YP_073768.1; HHV-6B U28 NP_050209.1; HHV-6A U28 NP_042921.1). The sequences were aligned using MUSCLE [51] and subsequent phylogeny was generated using neighbor joining tree without distance corrections [52]. (B) Schematics of each human herpesvirus RNR large subunit (scale bar = 100 amino acids). Colored regions represent the conserved RNR core domain, and gray regions represent unique N- and C-terminal extensions. The cysteine (C) and tyrosine (Y) residues required for RNR catalytic activity are labeled, with at least one cysteine and one tyrosine lacking from each beta-herpesvirus RNR large subunit. The nucleotide binding domain is also shown with the three essential glycine (G) residues (GxGxxG). (C) Representative immunofluorescent microscopy images of HeLa cells expressing BORF2-mCherry or HSV-1 UL39-mCherry (stably) and A3A-eGFP or A3B-eGFP (transiently). The nuclei are stained blue with Hoechst to help distinguish A3A/B relocalization events, which are yellow in the merged images (S.N.M. and R.S.H., unpublished images representative of published work [49]).

Figure 1

General schematic of herpesvirus replication. A prototypic herpesvirus consists of an outer lipid bilayer with an envelope protein that binds to cellular receptors (e.g., CD21 and MHC-II for EBV) and mediates entry into cells, a tegument layer of viral and host proteins and RNAs used for processes immediately following entry, and an inner icosahedral capsid necessary for nuclear entry and sheltering of the linear double-stranded DNA genome. Several factors, including those listed, determine whether the viral genome enters a latent or lytic replication mode. During latency, herpesviruses express a small subset of proteins and RNAs that combine to maintain the viral genome throughout normal cellular division (e.g., EBV EBNA proteins, LMP proteins, and EBV-encoded small RNAs [EBERs]). Under appropriate conditions, the virus is able to reactivate into the lytic cycle, which results in the production of numerous lytic proteins (e.g., EBV BZLF1, the viral transcriptional activator; BALF5, the viral DNA polymerase; BORF2, the viral ribonucleotide reductase large subunit; BMRF1, the viral processivity factor). New virus production requires viral DNA replication, encapsidation of the viral genome in the nucleus, acquisition of a lipid bilayer envelope in the cytoplasm, and subsequent release from the cellular membrane.

Figure 4

Mechanisms of APOBEC3 counteraction by viruses. Top—Schematic of the herpesviral RNR-mediated A3 counteraction mechanism reviewed here. The RNR large subunit (pink) directly binds each A3 enzyme (green/orange), inhibits ssDNA deaminase activity, and protects viral DNA during lytic replication (blue). RNR-A3 complexes accumulate in cytoplasmic aggregates that sometimes associate with the endoplasmic reticulum (ER) and may include as-yet-unidentified cellular factors (gray). Bottom—Schematics of additional A3 counteraction mechanisms (left to right): the nucleocapsid component of HTLV-1 Gag blocks A3G from packaging into viral particles, the foamy virus Bet protein promotes A3G aggregation, the murine mammary tumor virus (MMTV) reverse transcriptase synthesizes dsDNA rapidly and thereby limits murine A3 access to ssDNA intermediates, and the Vif proteins of lentiviruses, such as HIV-1, nucleate the formation of an E3 ubiquitin ligase complex that degrades cytoplasmic A3 enzymes (A3D, A3F, A3G, and A3H).