RNA-binding activity of 33N-term hAPE1. (A) Primary structure and biophysical parameters of 33N-term hAPE1. (B, left) 33N-term hAPE1 shows RNA-binding ability. EMSA experiments were performed as described in ‘Materials and Methods’ section with purified recombinant hAPE1 or a synthetic 1–33 hAPE1 peptide (33N-term hAPE1). Competition experiments were performed with the indicated amount of peptide co-incubated with the hAPE1 during the binding reaction. Mixtures were separated on a native 10% (w/v) polyacrylamide gels at 120 V for 2 h. F indicates the free 34FRNA oligonucleotide ³²P-labelled probe. The image is from a representative gel out of three independent experiments, (B, right) 33N-term hAPE1 competes with hAPE1 for 28S rRNA binding. GST pull-down experiments were performed with total RNA extracted from HeLa cells as prey, a 10-fold molar excess of the 33N-term hAPE1 synthetic peptide as competitor and GST, GST–hAPE1 or GST-hAPE1NΔ33 as baits (see ‘Materials and Methods’ section for details). Histograms represent average values with SDs of three independent experiments.

Mapping of the RNA-binding domain of hAPE1. (A) Schematic representation of the hAPE1 deletion mutants used. (B–D) An amount of 120 pmol of each bait protein was incubated with 10 µg of total RNA purified from HeLa cells, as described in ‘Materials and Methods’ section. (B, right) After GST pull-down, samples underwent either western blot analysis using an anti-GST antibody or RNA extraction and quantification by reverse transcription and Q-PCR analysis using 28S rRNA specific primers. The amounts of 28S rRNA recovered were then normalized versus the amount of bait proteins and the resulting values are plotted in the histograms in which average values with SDs of three independent experiments are plotted (*P < 0.05) (B, left). (C) The same procedure as in (B) was used in the presence of 10-fold molar excess competitor peptides: N1-9 hAPE1 spanning the first 9N-terminal aminoacids of APE1; 1–9ac hAPE1 indicates the same peptide bearing acetylation on residues 6 and 7; N23–33 hAPE1 spanning the region 23–33 of hAPE1, and 33N-term hAPE1 spanning the first 33N-terminal amino acids of the protein. Histograms represent the normalized amount of recovered 28S rRNA and the bars represent SDs of three independent experiments (*P < 0.05).

Lysine residues in the 23–33 basic region are responsible for hAPE1 binding to RNA. (A, left) Lysine to alanine mutants of hAPE1 were generated through site-directed mutagenesis. Recombinant proteins were expressed in E. coli, purified to homogeneity and used for further assays. The correct molecular mass of each recombinant protein was assessed by LC–MS analysis. Representative Coomassie-stained 10% SDS–PAGE of 1 µg of each recombinant protein is shown. (A, right) Purified hAPE1, hAPE1-K3 and hAPE1-K5 proteins were assayed for their ability to bind to 34FRNA by EMSA. A representative EMSA gel image of three independent experiments is shown. (B) GST pull-down experiments were performed, as described in ‘Materials and Methods’ section, with total RNA extracted from HeLa cells as prey to test the RNA-binding ability of the hAPE1-K3 and hAPE1-K5 mutant proteins in comparison with the wild-type hAPE1 and the hAPE1NΔ33. After pull-down, samples underwent either western blot analysis (using an anti-GST antibody) to confirm the amount of bait protein present in each sample, or 28S rRNA quantification. The amounts of 28S rRNA recovered after pull-downs were then normalized versus the amount of bait proteins. The resulting histogram shows the average values with SD of three independent experiments (*P < 0.05).

Lysine residues in the 23–33 basic region control the catalytic activity of hAPE1. (A) For each different hAPE1 proteins, dose-response experiments were performed in order to evaluate the most suitable concentration of enzyme to be used for the calculation of catalytic constants. An enzyme concentration capable of cleaving ∼7 nM 34FDNA over the time of reaction, indicated by the red spots, was selected for time-lapse experiments. (B) Effect of K⁺ concentration on the endonuclease activities of hAPE1 and of its mutants, on abasic DNA. Reaction rates, obtained as in (A), were quantitatively evaluated and the calculated values are plotted. The resulting histogram shows the average values with SD of three independent experiments performed with different batches of purified proteins.

Mapping of the amino acids involved in hAPE1 binding to NPM1. (A, left) Overlay of SPR sensorgrams relative to the binding of NPM1 to immobilized hAPE1 (see ‘Materials and Methods’ section for experimental details). (A, middle) Plot of RU_max from each binding versus NPM1 concentrations (micro molar); data were fitted by non-linear regression analysis. (A, right) Overlay of SPR sensorgrams relative to the same experiment carried out on immobilized hAPE1NΔ33. (B and C) GST pull-down experiments with different hAPE1 deletion (B) or with K to A (C) mutants as baits and a recombinant NPM1 as prey. Pull-down samples underwent western blot analysis using either an anti-GST antibody or an anti-NPM1 antibody. The NPM1 to GST signal intensity ratio for each sample is plotted in the histogram. (D) K24, K25, K27, K31 and K32 residues of hAPE1 are necessary for protein nucleolar localization and binding to NPM1. HeLa cells were transfected with pcDNA5.1 expression vector containing the cDNA of Flag-tagged hAPE1 or that of hAPE1NΔ33, hAPE1K3 or hAPE1K5 mutants. Confocal microscopy (left) of HeLa cells fixed and stained with antibodies against NPM1 (red) and ectopic Flag-tagged hAPE1 (green). Total cell extracts from transiently transfected HeLa cells with empty vector, hAPE1, hAPE1NΔ33, hAPE1K3 or hAPE1K5 mutants were co-immunoprecipitated. Co-immunoprecipitated material was separated onto 10% SDS–PAGE and analysed by western blotting to evaluate the levels of NPM1 binding by using the specific antibody. (E) hAPE1 binding to 28S rRNA is inhibited by NPM1. GST pull-down experiments with purified hAPE1 were carried out using 100 pmol of each bait protein incubated with the indicated amount of total RNA from HeLa cells. Pull-down samples were analysed as described in Figure 2B, The amounts of 28S rRNA recovered after pull-downs were then normalized versus the amount of bait proteins. The resulting histogram shows the average values with SD of three independent experiments. (F) Effect of K⁺ concentration on the stimulation of the endonuclease activities of hAPE1 and its mutants induced by NPM1. For each recombinant hAPE1 protein, 20 pM of each enzyme were incubated with or without 0.4 nM NPM1 and reacted, in either 40 or 150 mM KCl, with the 34FDNA as described in ‘Materials and Methods’ section. Then, reaction mixtures were separated in denaturing gels, which were then scanned for quantification. Histograms, representing the fold of increase in the reaction rates of each hAPE1 recombinant proteins, are the mean of two independent experiments.

Lysine residues in the 23–33 basic region are acetylated in vivo. (A) Tandem mass spectrometry of some acetylated peptides present in the endoprotease AspN digest of hAPE1 purified from HeLa cells (see ‘Materials and Methods’ section for details). The fragmentation mass spectrum of the quadruply charged ion at m/z 568.5 associated with the peptide (15–33)Ac3 is shown. (B) The fragmentation mass spectrum of the triply charged ion at m/z 852.5 associated with the peptide (15–35)Ac4 in shown. Spectral assignment demonstrated that hAPE1 is acetylated at K27, K31, K32 and K35, as highlighted in the peptide sequence associated to each panel.

Evolutionary analysis of APE1N-terminal sequence. (A) Evolutionary tree for vertebrate APE1 sequences based on the analysis of amino acid sequences aligned with the first 60 amino acids in hAPE1. Numbers indicate the bootstrap values obtained during the construction of the tree. Mammalian and non-mammalian vertebrate APE1 sequences are found in distinct branches with C. elegans representing the sequence most evolutionarily distant from the human one. HUMAN, H. sapiens; PANTRO, P. troglodytes; MACACA, M. mulatta; HORSE, E. caballus; PIG, S. scrofa; COW, B. taurus; DOG, C. lupus familiaris; ORNITHO, O. anatinus; CELEGANS, C. elegans; URCHIN, S. purpuratus; MOUSE, M. musculus; RAT, R. norvegicus; XENLAE, X. laevis; XENTRO, X. tropicalis; SALMON, S. salar; ZEBRAFISH, D. rerio; STICKLEBACK, G. aculeatus. (B) Multiple sequence alignment of N-term APE1 sequences. Multiple alignment was performed using the software MAFFT (v6.531b) (35), available at the EMBL-EBI server (www.ebi.ac.uk/Tools/sequence.html). Arrows indicate K24, K25, K27, K31 and K32, respectively. (C, left) zAPE1 has lower affinity for RNA than hAPE1, as determined by EMSA. Reactions were performed as described in ‘Materials and Methods’ section with homogeneously purified recombinant proteins. The mixtures were separated on a native (non-denaturing) 8% (w/v) polyacrylamide gel at 120 V for 2 h. In each reaction, 2.5 pmol of ssFRNA were used. F indicates the free 34FRNA oligonucleotide ³²P-labelled probe. A representative EMSA gel image of three independent experiments, using the abasic ssFRNA sequence, is shown. (C, right) zAPE1 is catalytically more active than hAPE1 over abasic DNA. AP endonuclease activity of different mutant APE1 enzymes was tested by performing a dose-response experiment, as described in the ‘Materials and Methods’ section.

Proposed mechanism of regulation of hAPE1 functions and compartmentalization through acetylation of critic lysine residues. Within cells, APE1 is present with different degrees of acetylation on residues 27, 31, 32, 35 (Ac-APE1) and may also be involved in interaction with NPM1 and/or different protein partners (indicated here as ‘X’ and ‘Y’). Genotoxic stress may shift the equilibrium between the non-acetylated and the Ac-APE1 forms toward these latter and more active form, possibly by its reduced affinity for the incised DNA product. Acetylation, in turn, may inhibit binding to RNA substrates thus, possibly, this PTM may change the equilibrium of the hAPE1 pool from the RNA-bound to the DNA-bound substrates. Moreover, acetylation may also control compartmentalization of hAPE1 within specific subnuclear structures (such as nucleoli), where the enzyme could be stored once bound to interacting partners as NPM1. Whether NPM1 may specifically participate in DNA repair inside of nucleoli, or even outside of the subnuclear condensed area, is currently unknown. In the first case, the genome containing rRNA gene may be a favoured substrate for hAPE1 in vivo. On the other hand, if the latter is the case, stimuli triggering NPM1 to get out of nucleoli may affect the cell response in terms of hAPE1 impact on genome stability. Further experiments are currently underway to address these issues.