The innate antiretroviral factor APOBEC3G does not affect human LINE-1 retrotransposition in a cell culture assay

APOBEC3G (apolipoprotein B mRNA-editing enzyme catalytic polypeptide-like 3G) is an innate intracellular antiretroviral factor that can inhibit viral retroelements such as retroviruses and hepadnaviruses. However, it is unknown whether it can act on non-viral substrates. Retrotransposons are transposable elements that cumulatively account for about one third of the human genome. They are commonly classified in long terminal repeat (LTR) retrotransposons, which are strongly homologous to retroviruses, and non-LTR retrotransposons also known as L1 elements or LINE-1 (long interspersed nucleotide element-1) elements. Most of the L1 elements are defective and only a small number are very active in vivo, but they are responsible for nearby all of the retrotransposition in the human population. The cloning of active human L1 elements has allowed the development of tissue culture-based assays for measuring their retrotransposition potential. We used such an assay to demonstrate that APOBEC3G, which impairs the replication of exogenous retroelements, does not affect the replication of endogenous L1 retrotransposons.


INTRODUCTION
Human APOBEC3G (Apolipoprotein B mRNA-editing enzyme catalytic polypeptidelike 3G), also known as CEM15, is a cytidine deaminase that can block the replication of a wide array of retroelements (1)(2)(3). Packaged in retroviral particles, APOBEC3G (APO3G) associates with the reverse transcription complex where it deaminates cytosine residues to uracil in the growing minus-strand viral DNA. These dU-rich transcripts are either degraded or yield G-to-A hypermutated hence largely non-functional proviruses (1,2,4). APOBEC3G is countered by the Vif (virion infectivity factor) protein of lentiviruses, which associates with the cellular enzyme to prevent its virion incorporation and trigger its proteasomal degradation (1,2,(5)(6)(7)(8). In the absence of Vif, the innate antiviral protein can inhibit human and simian immunodeficiency viruses (HIV and SIV, respectively), equine infectious anemia virus (EIAV), and the gammaretrovirus murine leukemia virus (MLV). In addition, APOBEC3G can block the replication of hepatitis B virus (HBV), a hepadnavirus whose life cycle also includes reverse transcription (3). In this case, however, one observes an inhibition of viral pregenomic RNA packaging, and thus of viral DNA synthesis, rather than the lethal editing of viral reverse transcripts.
Retrotransposons are commonly classified as long terminal repeat (LTR) and non-LTR retrotransposons (9). The Ty1/3 elements and copia elements are LTR retrotransposons of yeast and Drosophila melanogaster, respectively, and actively participate in the genome remodeling of these organisms. In higher species, LTR retrotransposons, also known as endogenous retroviruses, are most commonly defective and usually cannot spread from cell to cell. Evidence exists for ongoing LTR retrotransposition in the mouse, but no active LTR retrotransposon has so far been isolated in humans, even though human endogenous retroviruses (HERVs) make up roughly 7% of the genome. Yet the presence of human-specific HERV proviruses and their absence in corresponding loci of other primates indicate that a subset of these LTR retrotransposons has been active during the recent evolution of hominoids. L1 elements occupy close to 20% of the human genome (10). Their vast majority is defective due to truncations or rearrangements, but three to four thousand human L1s are full-length, amongst which an estimated 80-100 in the average human are active (11). By comparison, the mouse genome contains up to three thousand active L1 elements. L1 retrotransposition accounts for several genetic disorders in humans, including cases of haemophilia A, Duchene muscular dystrophy, type-2 retinitis pigmentosa, ß-thalassaemia and chronic granulomatous disease (12)(13)(14)(15)(16). Active human L1 retrotransposons have been cloned and tissue-culture-based assay has been developed to compare their

RESULTS AND DISCUSSION
L1 retrotransposons have been cloned and in vitro assays have been developed to measure their capacity to retrotranspose. Here we used the pL1 RP -EGFP plasmid which contains a full-length active human L1 element with a retrotransposition detector cassette inserted in its 3' untranslated region (Fig. 1). This cassette comprises a CMV promoter driving expression of an EGFP reporter separated by a γ-globin intron in the antisense orientation. EGFP expression can only occur when the intron is removed by splicing during a retrotransposition event. A puromycin resistance gene further allows the selection of cells transfected with this plasmid. Construct pL1-ac002980 -EGFP has a similar organization, being derived from a second human L1 element previously found to have higher levels of retrotransposition activity than pL1 RP -EGFP (11). The pL1 RP -(JM111)-EGFP encodes an L1 element containing missense mutations that were previously shown to abolish retrotransposition (18).
We took advantage of this system to ask whether APOBEC3G could interfere with human L1 retrotransposition. The 143BTK-cell line is a substrate very conducive to L1 retrotransposition, but we first needed to ascertain that it could reveal APOBEC3G action. For this, we established 143BTK-derivatives that stably express the cytidine deaminase owing to transduction by an APOBEC3G-containing retroviral vector ( Fig.   2A). Vif-defective HIV-1 produced by transfection from these but not control cells exhibited levels of infectivity that were about fifty-fold lower than wild-type, very similar to those measured for ∆Vif virions produced in these cells when an APOBEC3Gexpressing plasmid is co-transfected together with the viral DNA (Fig. 2B). We thus transfected the APOBEC3G-expressing and control 143BTK-cells with pL1 RP -EGFP, pL1-ac002980 -EGFP and pL1 RP -(JM111)-EGFP, the latter serving as a negative control.
Puromycin selection was then applied. Expression of APOBEC3G-HA was verified in by guest on March 24, 2020 http://www.jbc.org/ Downloaded from these doubly selected cells by western blotting (Fig. 3A) and the percentage of GFPpositive cells was monitored over time by FACS analysis (Fig. 3B). Cells transfected with the retrotransposition-incompetent pL1 RP -(JM111)-EGFP plasmid remained negative throughout the study. With pL1 RP -EGFP, 6% of the cells were GFP-positive at day 7, and this fraction increased to 9% at day 21, reflecting ongoing retrotransposition.
As previously described (11), pL1-ac002980 -EGFP was significantly more active, with 8 and 14% GFP-positive cells at days 7 and 21, respectively. However, in no case were these results affected by the presence of APOBEC3G. Accordingly, L1 elements which have lost the intron during retrotransposition have been specifically cloned and sequenced but no G-to-A mutations were found (data not shown). Because expression of the GFP marker from the pL1 RP -EGFP and pL1-ac002980 -EGFP constructs requires transcription, splicing, reverse transcription and integration of the GFP cassette together with the full preservation of its coding sequence, we conclude that APOBEC3G does not interfere with any of these steps. Accordingly, mechanisms put into play by the cellular enzyme in the control of exogenous retroelements such as lentiviruses, gammaretroviruses and hepatitis B virus do not affect human L1 retrotransposition.
Specific features of L1 elements may explain their resistance to APOBEC3G action.
Indeed, L1 retrotransposition is based on target-primed reverse transcription, a process in which the endonuclease-cleaved target DNA serves as a primer from which complementary DNA is made using the polyadenylated L1 RNA as a template (23,24).
Synthesis of the L1 DNA thus takes place in the nucleus, whereas APOBEC3G is a predominantly cytoplasmic protein. Nevertheless, the L1 ribonucleoprotein complex thought to act as an intermediate in retrotransposition is assembled in the cytoplasm, therefore one can assume that it does not carry the determinants that recruit the cytidine deaminase in its retroviral and perhaps hepadnaviral counterparts.   Antibodies against HA and PCNA were used (B) The retrotransposition rate was monitored with time by FACS analysis, in the above described neomycin and puromycin selected cells.