Apobec‐1 complementation factor (ACF) regulates mRNA stability and polysome distribution but has no role in apoB RNA editing

Abstract

The AU‐rich RNA binding protein ACF is thought to represent the RNA binding subunit of the apoB mRNA editosome. Known ACF targets include apoB and IL‐6; ACF binds and stabilizes IL‐6 mRNA playing a key role in liver growth. ACF exhibits nanomolar binding affinity to apoB RNA and recombinant ACF plus Apobec‐1 are both necessary and sufficient for C to U RNA editing of single stranded apoB RNA in vitro, specifically at the canonical site C6666. Mutations or deletions of any of 3 RNA binding motifs of ACF attenuate interaction with Apobec‐1, apoB RNA binding and eliminates in vitro C to U RNA editing. These findings led to the dogma that ACF is required for Apobec‐1 dependent C to U RNA editing. However an unequivocal demonstration of the in vivo role for ACF in RNA editing has heretofore been impossible owing to embryonic lethality following germline Acf deletion. Acf +/− mice are viable but show no alteration in C to U editing of apoB at C6666: (Acf +/−81%U vs WT 92%U). Antisense mediated ACF knockdown (ASO) in Acf +/− mice further reduced hepatic ACF to <5% WT levels, yet without alteration in apoB RNA editing (86%U). We then generated hepatocyte‐specific ACF transgenic mice (Acf +/Tg) in which hepatic ACF expression was ~3–4 fold higher than wild type littermates. Acf +/Tg mice exhibit spontaneous hepatic steatosis on a chow diet but again demonstrate no change in hepatic apoB RNA editing (82%U). Thus, over a >60 fold range of hepatic ACF expression, there was no alteration in apoB RNA editing in vivo. We then asked if ACF could constrain promiscuous RNA editing in vivo. We delivered Adenoviral Apobec‐1 (Ad‐A1) into wild type, ASO treated Acf +/− and Acf +/Tg mice and examined RNA editing across a 650 nt region flanking the canonical site. Ad‐A1 administration induced promiscuous hyperediting upstream and downstream of the C6666 in WT liver (at 65 sites between nucleotides 6563 and 7210). However, Ad‐A1 administration into Acf +/Tg or ASO treated Acf +/− mice resulted in only subtle variations in promiscuous editing sites and efficiency. These observations suggest that, contrary to prevailing dogma, ACF is dispensable for Apobec‐1 dependent editing of apoB RNA at the canonical site and does not constrain promiscuous RNA editing. We also sequenced all 22 physiological Apobec‐1 RNA targets (all in 3′UTR) from Acf +/Tg liver and found that 6 (Colec10, Mpeg1, Aldh6a1, B2m, Dcn, Tmem30a) exhibited 2–5 fold increased editing efficiency, 8 showed 2 to 5‐fold reduced editing efficiency and 8 showed no alteration in editing frequency. Finally, we found that Acf +/Tg mice exhibit spontaneous hepatic dysplasia and hepatocellular carcinoma, with significant shifts in polysomal distribution and increased translational activity of several targets including CD36, CIDEA and apoB and identified sox9 RNA within anti‐ACF immune complexes. This spontaneous tumor phenotype was completely replicated in Acf +/Tg mice crossed into the Apobec‐1 null background. Together these findings suggest that ACF is not required for Apobec‐1 dependent C to U apoB RNA editing in vivo, but plays a key role in posttranscriptional regulation of genes involved in liver growth and tumorigenesis. ACF likely functions in RNA binding, mRNA stability and as a chaperone regulating polysome distribution of AU‐rich targets.