Human Tribbles 3 protects nuclear DNA from cytidine deamination by APOBEC3A

Abstract

The human polydeoxynucleotide cytidine deaminases APOBEC3A, APOBEC3C, and APOBEC3H are capable of mutating viral DNA in the nucleus, whereas APOBEC3A alone efficiently edits nuclear DNA. Deamination is rapidly followed by excision of uracil residues and can lead to double-stranded breaks. It is not known to which protein networks these DNA mutators belong. Using a yeast two-hybrid screen, we identified the human homolog of Drosophila Tribbles 3, TRIB3, as an interactor for APOBEC3A and APOBEC3C. The interaction was confirmed by co-affinity purification. Co-transfection of APOBEC3A with a TRIB3 expression vector reduced nuclear DNA editing whereas siRNA knockdown of TRIB3 increased the levels of nuclear DNA editing, indicating that TRIB3 functioned as a repressor of A3A. It also repressed A3A-associated γH2AX positive double-stranded breaks. The interaction results in degradation of A3A in a proteasome-independent manner. TRIB3 has been linked to cancer and via its own interactors and links the A3A DNA mutators to the Rb-BRCA1-ATM network. TRIB3 emerges as an important guardian of genome integrity.

FIGURE 1.

APOBEC3A and 3C interact with human TRIB3 protein. A, GST pulldowns were followed by Western blotting with antibodies to 3×FLAG. Below are the TRIB3 and APOBEC3 controls. B, interaction was specific for TRIB3 compared with the paralogous human proteins TRIB1 and TRIB2 (see supplemental Fig. S1A for sequence comparisons). C, APOBEC3C but not APOBEC3A can form multimers.

FIGURE 2.

TRIB3 protects whereas TRIB3 siRNA knockdown increases A3A deamination of nuDNA. A, no apparent deamination of CMYC DNA by A3A or A3C ± TRIB3 in HeLa cells was due to rate-limiting UNG degradation. B, in HEK-293T-UGI cells TRIB3 protects nuDNA from editing by A3A in a dose-response manner. DNA concentrations were 2 μg throughout and were compensated for by plasmid vector. C, co-transfection of A3A with 1 μg of TRIB3 siRNA knockdown only results in recovery of hyperedited CMYC DNA in HeLa cells. D, transduction of IFN IαA stimulated U937 cells with TRIB3 shRNA or control shRNA lentiviral particles. E, co-transfection of A3A with 1 μg of TRIB3 siRNA knockdown only results in recovery of enhanced hyperediting of CMYC DNA in 293T-UGI cells. F, a collection of hyperedited CMYC sequences was recovered from samples C1, D1, and D2 (C and E). Only differences are shown for 100 bases of a total of 241 bp. To the right are the number mutations per sequence and the percentages of Cs or Gs (Cs on opposite strand) edited.

FIGURE 3.

Cell localization of the A3A and TRIB3. A–C, confocal microscopy yielding scant evidence of cytoplasmic and slightly nuclear A3A-V5 co-localizing with a nuclear 3×FLAG-TRIB3 in HeLa cells 36 h after transfection. D–F, A3C, TRIB3, and merge. G–I, addition of the SV40 NLS (residues PPKKKRKV) to the carboxyl terminus of A3A co-localization with TRIB3 is readily observed.

FIGURE 4.

The A3A interaction involves the amino terminus of TRIB3. A, immunofluorescence of TRIB3, its cytoplasmic ΔNLS derivative, and the murine homolog, mTRIB3. B, CMYC editing by A3A, A3A-NLS, and modulation by an amino-terminal 36-residue deletion of hTRIB3. C, GST pulldown analyses for the A3A+TRIB3, +TRIB3ΔNLS, and +mTRIB3 constructs. WB, Western blot.

FIGURE 5.

TRIB3 interaction with A3A results in lower steady-state levels. A, 293T cells were transfected by A3A, and increasing concentrations of TRIB3 and steady-state levels of A3A were determined by Western blotting at 24 h. The β-actin loading controls are shown below. B and C, TRIB1 and TRIB2 do not degrade A3A. D, the A3A-NLS construct is similarly degraded. E, mTRIB3 is capable of degrading hA3A. F, by contrast, deletion of the amino-terminal 36 residues of TRIB3 which include a NLS abrogates A3A degradation. G, epoxomycin treatment of 293T cells fails to abrogate A3A degradation by TRIB3.

FIGURE 6.

A3A-induced DSBs can be countered by TRIB3. A, superposition of FACS analysis of γH2AX-positive HeLa cells gated on the V5-tagged A3A with and without TRIB3 at two different A3A:TRIB3 ratios. B, means and standard deviations γH2AX-positive cell frequencies for quadruplicate transfections, DNA concentrations were always 2 μg and compensated for by plasmid vector. * indicates a statistically significant difference between two observed percentages (p < 0.05).

FIGURE 7.

Transcriptome analysis of A3A, A3C, and TRIB3 in hematopoietic cells and peritumoral cirrhotic tissues. A, TaqMan analysis of PBMC from two healthy donors at 20, 48, and 120 h after stimulation. B, TaqMan analysis of positively selected CD4⁺ T cells from two healthy donors at days 2 and 5. C, ABI chip analysis of A3A, A3C, TRIB1–3 in peritumoral tissues from patients with alcoholic, HBV-, HBV+hepatitic C virus (HCV)-, and HCV-associated cirrhosis. The inset gives paired A3A and TRIB3 levels from the same samples. **, statistically significant difference compared with healthy liver (p < 0.01). Expression levels are all normalized to RPL13A levels.