**1. Introduction**

Apurinic/apyrimidinic endonuclease 1 (APE1) is an essential protein in mammals with multiple functions in base excision DNA repair (BER), regulation of gene expression, RNA metabolism, and other specific cellular processes [1–3]. BER is the primary mechanism for correcting apurinic/apyrimidinic (AP) sites (created through N-glycosidic bond cleavage), modified bases, and single-strand breaks (SSBs) [4,5]. The major enzymatic function of APE1 in BER is the incision of AP sites, one of the most abundant types of oxidative DNA damage [6]. Additional activities of APE1, 30 -diesterase, and 30 -50 -exonuclease contribute to removing terminal blocking groups in BER DNA intermediates and to proofreading DNA mismatches introduced by DNA polymerase β (Polβ) [1,3]. APE1 is also capable of incising DNA at certain base lesions (nucleotide incision repair) and RNA at abasic site and specific regions, as well as of processing abasic and oxidized ribonucleotides embedded in the DNA [1–3,7,8]. A redox activity of APE1 is responsible for the regulation of DNA-binding activities of different transcription factors [1,2]. The DNA repair and redox activities of APE1 can

play a collaborative role in different processes [9,10]. Interactions with multiple protein partners that modulate diverse enzymatic activities of APE1 and mediate its regulatory function in transcription are highly interconnected and dynamically regulated through various post-translational modifications of APE1 [1,2,11–13]. Dysregulation of multifunctional activities of APE1 is associated with various human pathologies, making APE1 a potential therapeutic target [1,5,12,14]. partners that modulate diverse enzymatic activities of APE1 and mediate its regulatory function in transcription are highly interconnected and dynamically regulated through various post-translational modifications of APE1 [1,2,11–13]. Dysregulation of multifunctional activities of APE1 is associated with various human pathologies, making APE1 a potential therapeutic target [1,5,12,14].

of APE1 can play a collaborative role in different processes [9,10]. Interactions with multiple protein

The human APE1 is composed of two structural domains and a disordered N-terminal region [15], formed mostly by the sequence highly conserved in mammals (Figure 1). The catalytically active core responsible for the enzyme activities in BER is created by the universally conserved sequence 62–318 [16]. The region required for the redox activity of APE1 is in the N-terminal domain [1]. The unique disordered region comprises the nuclear localization signal and multiple post-translational modification sites. This region, enriched in lysine residues, contributes to APE1 interaction with various DNA/RNA structures and to acetylation-mediated modulation of the enzyme DNA repair activity in vitro and in vivo [17–21]. Ubiquitination and acetylation of specific Lys residues of the N-terminal region modulate the expression level and different functions of APE1 in vivo via proteolytic degradation or limited N-terminal proteolysis, respectively [22,23]. The conserved catalytic core and the N-terminal extension are both required for the APE1 function at telomeric DNA substrates and their protective protein complexes [24,25]. The N-terminal extension of APE1 contributes to a different extent to the interaction with proteins involved in RNA processing and ribosome biogenesis [26]; its interaction with nucleophosmin (NPM1) is very unstable to be detected in the absence of the remaining C-terminal portion [20]. The first 35 residues are critical for the physical interaction of APE1 with X-ray repair cross-complementing protein 1 (XRCC1) functioning as a scaffold protein of BER [27]. However, involvement of the N-terminal extension of APE1 in the interaction with other canonical BER proteins still remains unexplored. The human APE1 is composed of two structural domains and a disordered N-terminal region [15], formed mostly by the sequence highly conserved in mammals (Figure 1). The catalytically active core responsible for the enzyme activities in BER is created by the universally conserved sequence 62–318 [16]. The region required for the redox activity of APE1 is in the N-terminal domain [1]. The unique disordered region comprises the nuclear localization signal and multiple post-translational modification sites. This region, enriched in lysine residues, contributes to APE1 interaction with various DNA/RNA structures and to acetylation-mediated modulation of the enzyme DNA repair activity in vitro and in vivo [17–21]. Ubiquitination and acetylation of specific Lys residues of the N-terminal region modulate the expression level and different functions of APE1 in vivo via proteolytic degradation or limited N-terminal proteolysis, respectively [22,23]. The conserved catalytic core and the N-terminal extension are both required for the APE1 function at telomeric DNA substrates and their protective protein complexes [24,25]. The N-terminal extension of APE1 contributes to a different extent to the interaction with proteins involved in RNA processing and ribosome biogenesis [26]; its interaction with nucleophosmin (NPM1) is very unstable to be detected in the absence of the remaining C-terminal portion [20]. The first 35 residues are critical for the physical interaction of APE1 with X-ray repair cross-complementing protein 1 (XRCC1) functioning as a scaffold protein of BER [27]. However, involvement of the N-terminal extension of APE1 in the interaction with other canonical BER proteins still remains unexplored.

**Figure 1.** Structural organization of human apurinic/apyrimidinic endonuclease 1 (APE1). Protein is composed of two structural domains and disordered N-terminal region invisible in crystal structures [15]. Sequence 62–318 responsible for AP endonuclease activity is conserved in pro- and eukaryotes; first 35 residues of N-terminal eukaryote-specific extension (residues 1–61) are highly conserved in mammals [16]. Isoelectric points (calculated using ExPASy proteomics server) of full-length protein, its N-terminally truncated forms (APE1NΔ35, APE1NΔ61), the entire extension, and its mammalian-specific fragment are presented. **Figure 1.** Structural organization of human apurinic/apyrimidinic endonuclease 1 (APE1). Protein is composed of two structural domains and disordered N-terminal region invisible in crystal structures [15]. Sequence 62–318 responsible for AP endonuclease activity is conserved in pro- and eukaryotes; first 35 residues of N-terminal eukaryote-specific extension (residues 1–61) are highly conserved in mammals [16]. Isoelectric points (calculated using ExPASy proteomics server) of full-length protein, its N-terminally truncated forms (APE1N∆35, APE1N∆61), the entire extension, and its mammalian-specific fragment are presented.

We demonstrated the usefulness of fluorescence-based techniques for the detection and quantification of physical interactions between various BER proteins, and for the detection of modulation of protein–protein interactions by DNA intermediates [28]. In the present study, we used the same approaches to explore the role of the N-terminal extension of APE1 in coordination of mammalian BER process via protein–protein interactions. The relative binding affinities of the full-length and two N-terminally truncated forms of APE1 (APE1NΔ35 and APE1NΔ61) for three We demonstrated the usefulness of fluorescence-based techniques for the detection and quantification of physical interactions between various BER proteins, and for the detection of modulation of protein–protein interactions by DNA intermediates [28]. In the present study, we used the same approaches to explore the role of the N-terminal extension of APE1 in coordination of mammalian BER process via protein–protein interactions. The relative binding affinities of the full-length and two N-terminally truncated forms of APE1 (APE1N∆35 and APE1N∆61) for three functionally

and structurally different BER proteins, downstream enzyme Polβ, scaffold XRCC1 protein, and poly(ADP-ribose) polymerase 1 (PARP1) responsible for detecting SSBs and initiating their repair, were determined in the absence and presence of model DNA intermediates of BER, and the DNA-induced rearrangements of the protein–protein complexes were examined. Additionally, the influence of the N-terminal truncation on the functional cooperation between APE1 and Polβ in the absence and presence of accessory XRCC1 and PARP1 proteins was explored. *Int. J. Mol. Sci.* **2020**, *21*, x 3 of 19 functionally and structurally different BER proteins, downstream enzyme Polβ, scaffold XRCC1 protein, and poly(ADP)ribose polymerase 1 (PARP1) responsible for detecting SSBs and initiating their repair, were determined in the absence and presence of model DNA intermediates of BER, and the DNA-induced rearrangements of the protein–protein complexes were examined. Additionally, the influence of the N-terminal truncation on the functional cooperation between APE1 and Polβ in

#### **2. Results** the absence and presence of accessory XRCC1 and PARP1 proteins was explored.

#### *2.1. Contribution of N-terminal Extension of Human APE1 to the Interaction with Canonical BER Proteins* **2. Results**

To explore involvement of the N-terminal extension of human APE1 in the interaction with other BER proteins, and to quantify the contribution of the entire eukaryote-specific region and its disordered mammalian-conserved fragment to the binding affinities, fluorescence-titration experiments were performed using fluorescein-labelled full-length APE1 and its N-terminally truncated forms (APE1N∆35 and APE1N∆61). The proteins N-terminally labelled with 5(6)-carboxyfluorescein (FAM) were prepared as previously described [28] and are detailed in Supplementary Materials. The change in the fluorescence intensity of a FAM-labelled protein (FAM-APE1, FAM-APE1N∆35, or FAM-APE1N∆61) was monitored in the presence of unlabeled protein partners added at increasing concentrations (Figure 2). *2.1. Contribution of N-terminal Extension of Human APE1 to the Interaction with Canonical BER Proteins*  To explore involvement of the N-terminal extension of human APE1 in the interaction with other BER proteins, and to quantify the contribution of the entire eukaryote-specific region and its disordered mammalian-conserved fragment to the binding affinities, fluorescence-titration experiments were performed using fluorescein-labelled full-length APE1 and its N-terminally truncated forms (APE1NΔ35 and APE1NΔ61). The proteins N-terminally labelled with 5(6)-carboxyfluorescein (FAM) were prepared as previously described [28] and are detailed in Supplementary Materials. The change in the fluorescence intensity of a FAM-labelled protein (FAM-APE1, FAM-APE1NΔ35, or FAM-APE1NΔ61) was monitored in the presence of unlabeled protein partners added at increasing concentrations (Figure 2).

**Figure 2.** Fluorescence titration of 5(6)-carboxyfluorescein (FAM)-labelled APE1, APE1NΔ35, and APE1NΔ61 with (**A**) DNA polymerase β (Polβ) (**B**) X-ray repair cross-complementing (XRCC1) protein 1, and (**C**) poly(ADP)ribose polymerase 1 (PARP1). FAM-labelled protein (40 nM) excited at 482 nm in absence or presence of increasing concentrations of protein partner, and relative fluorescence intensities monitored at 530 nm. Curves show best fits (R2 values met or exceeded 0.98) **Figure 2.** Fluorescence titration of 5(6)-carboxyfluorescein (FAM)-labelled APE1, APE1N∆35, and APE1N∆61 with (**A**) DNA polymerase β (Polβ) (**B**) X-ray repair cross-complementing (XRCC1) protein 1, and (**C**) poly(ADP-ribose) polymerase 1 (PARP1). FAM-labelled protein (40 nM) excited at 482 nm in absence or presence of increasing concentrations of protein partner, and relative fluorescence intensities monitored at 530 nm. Curves show best fits (R<sup>2</sup> values met or exceeded 0.98) of four-parameter equation; EC<sup>50</sup> values derived from respective curves are presented. Data shown are representative of at least three independent experiments.

The fluorescence intensity of the FAM-labelled APE1 and its truncated forms increased in the presence of Polβ, XRCC1, or PARP1, indicating that the local environment of the fluorophore changed upon protein–protein association. Apparent equilibrium dissociation constants of the complexes determined by nonlinear regression analyses as effective concentrations (EC<sup>50</sup> values) of the protein partners at the half-maximal increase in fluorescence intensity [29] are presented in Table 1. Removal of the first 35 amino acid residues slightly decreased the affinity of APE1 for all three proteins (a 1.2–1.3-fold increase in EC50). Removal of the entire eukaryote-specific extension (residues 1–61) produced more significant effects: The binding affinity of APE1 for Polβ decreased 1.4-fold, and the affinity for PARP1 and XRCC1 decreased 1.7-fold.

**Table 1.** Binding parameters of protein–protein interactions determined for full-length and truncated forms of APE1 by fluorescence-based approaches.


<sup>a</sup> Titration experiments were performed at the constant 40 nM concentration of FAM-labelled protein. Underlined data for FAM-APE1 (previously published [28]) reproduced to compare binding parameters of full-length and truncated forms. <sup>b</sup> Parameters derived from titration curves by fitting to four-parameter equation, where EC<sup>50</sup> is half-maximal effective concentration of the protein partner, at which F – F<sup>0</sup> = (F<sup>∞</sup> – F0)/2, where F, the fluorescence intensity of a solution containing the FAM-labelled protein and the binding partner at a given concentration (C); F0, the fluorescence of a solution of the labelled protein alone; F∞, the fluorescence of the labelled protein saturated with the binding partner. Values are the mean (± SD) of at least three independent experiments. Values determined for complexes formed by truncated forms of APE1 with each protein partner statistically different from those of the full-length protein: *p* < 0.05 (\*), *p* < 0.01 (\*\*), *p* < 0.001 (\*\*\*); *t*-test, *n* = 3–4. <sup>c</sup> Effect of N-terminal truncation of APE1 on protein–protein affinity determined as the ratio of EC<sup>50</sup> values for complexes formed by the truncated APE1 and by the full-length APE1 with each binding partner. <sup>d</sup> Fluorescence resonance energy transfer (FRET) efficiency calculated from the fractional decrease of fluorescence intensity, E = 1 – Fda/Fd, where Fda and F<sup>d</sup> are fluorescence intensities of donor-labelled protein measured in the presence of the acceptor-labelled or the unlabeled protein partner, respectively. Values are the mean (± SD) of three independent experiments. Values determined for truncated forms of APE1, which were statistically different from those of the full-length protein in the respective complexes, are marked *p* < 0.05 (\*), *p* < 0.01 (\*\*).

To further explore influence of the N-terminal truncation on physical contacts between APE1 and BER proteins, we performed fluorescence resonance energy transfer (FRET) experiments. FRET is characterized by the efficiency of energy transfer (ET) from a fluorescent donor to an acceptor, which depends on the extent of the spectral overlap between donor emission and acceptor absorption, on the distance between the two fluorophores and on their relative orientation [30]. We chose fluorescein and tetramethylrhodamine as the donor–acceptor pair previously used to characterize interactions between BER proteins [28]. FAM-labelled APE1, APE1N∆35, or APE1N∆61 was titrated with Polβ, unlabeled or labelled with 5(6)-carboxytetramethylrhodamine (TMR) (Figure 3). The fluorescence intensity of the FAM-labelled protein increased less in the presence of TMR-Polβ than in the presence of Polβ, indicating participation of the donor- and acceptor-labelled proteins in FRET. FRET efficiencies determined for the FAM-APE1N∆35-TMR-Polβ and FAM-APE1-TMR-Polβ pairs were identical (Table 1). FRET efficiency determined for the FAM-APE1N∆61-TMR-Polβ pair was appreciably (6%) higher, indicating that, in this complex, the distance between two fluorophore probes and/or their relative orientation are more favorable for FRET. Analogous FRET measurements were performed for the other complexes. FRET signals detected for the XRCC1 complexes with the truncated forms of APE1 were lower than for the respective complex with the full-length APE1 (Table 1). On the other hand, efficiencies of FRET for the TMR-PARP1 pairs with FAM-APE1N35 and FAM-APE1N∆61 were higher than for the

respective pair with FAM-APE1. The differences in detected FRET signals evidently reflected changes in the localization of the FAM-labelled N-terminus within complexes formed by the full-length and N-terminally truncated forms of APE1. truncation on the binding affinity constants suggest that the interaction domain for proteins is basically formed by the conserved catalytic core of APE1, and the N-terminal extension mediates dynamic interactions.

in the interactions of APE1 with other BER proteins. Small effects produced by the N-terminal

*Int. J. Mol. Sci.* **2020**, *21*, x 5 of 19

in the presence of Polβ, indicating participation of the donor- and acceptor-labelled proteins in FRET. FRET efficiencies determined for the FAM-APE1NΔ35‒TMR-Polβ and FAM-APE1‒TMR-Polβ pairs were identical (Table 1). FRET efficiency determined for the FAM-APE1NΔ61‒TMR-Polβ pair was appreciably (6%) higher, indicating that, in this complex, the distance between two fluorophore probes and/or their relative orientation are more favorable for FRET. Analogous FRET measurements were performed for the other complexes. FRET signals detected for the XRCC1 complexes with the truncated forms of APE1 were lower than for the respective complex with the full-length APE1 (Table 1). On the other hand, efficiencies of FRET for the TMR-PARP1 pairs with FAM-APE1N35 and FAM-APE1NΔ61 were higher than for the respective pair with FAM-APE1. The differences in detected FRET signals evidently reflected changes in the localization of the FAM-labelled N-terminus within complexes formed by the full-length and N-terminally truncated

**Figure 3.** Characterization of Polβ interaction with APE1 and its truncated form by FRET. The FAM-labelled protein (40 nM FAM-APE1 or FAM-APE1NΔ61) excited at 482 nm in the absence or presence of increasing concentrations of unlabeled or TMR-labelled Polβ. FRET efficiency calculated from the fractional decrease (shown by arrow) of fluorescence intensity of the donor due to the presence of acceptor E = 1 – Fda/Fd, where Fda and F<sup>d</sup> are fluorescence intensities measured in the presence of identical subsaturating concentrations of TMR-Polβ or Polβ (open or filled symbols, respectively), and determined at saturation by fitting to four-parameter equation. EC<sup>50</sup> and E values determined for respective complexes are shown. **Figure 3.** Characterization of Polβ interaction with APE1 and its truncated form by FRET. The FAM-labelled protein (40 nM FAM-APE1 or FAM-APE1N∆61) excited at 482 nm in the absence or presence of increasing concentrations of unlabeled or TMR-labelled Polβ. FRET efficiency calculated from the fractional decrease (shown by arrow) of fluorescence intensity of the donor due to the presence of acceptor E = 1 – Fda/Fd, where Fda and F<sup>d</sup> are fluorescence intensities measured in the presence of identical subsaturating concentrations of TMR-Polβ or Polβ (open or filled symbols, respectively), and determined at saturation by fitting to four-parameter equation. EC<sup>50</sup> and E values determined for respective complexes are shown.

*2.2. Involvement of N-terminal Extension of APE in DNA-Dependent Modulation of Protein–Protein Interactions* Using fluorescence-based approaches, we showed in a previous study that DNA intermediates Taken together, these results indicate that the N-terminal extension of human APE1 is involved in the interactions of APE1 with other BER proteins. Small effects produced by the N-terminal truncation on the binding affinity constants suggest that the interaction domain for proteins is basically formed by the conserved catalytic core of APE1, and the N-terminal extension mediates dynamic interactions.

#### of BER induce rearrangements of various protein–protein complexes and modulate the strength of the interaction [28]. To explore the possible involvement of the N-terminal region of APE1 in this *2.2. Involvement of N-Terminal Extension of APE in DNA-Dependent Modulation of Protein–Protein Interactions*

modulation, fluorescence titration and FRET experiments with the full-length and truncated forms of APE1 were performed in the absence and presence of model DNA ligands (shown in Figure S1). A double-stranded DNA with a synthetic abasic site (a tetrahydrofuran residue, F), AP-DNA, is an initial BER substrate of APE1. A 1-nucleotide gapped DNA with a 5′-F group at the margin of the gap models a product of the APE1-catalyzed incision (AP-DNA inc) and a stable analog of the Polβ substrate not processed by the 5′-deoxyribose phosphate lyase activity. A 1-nucleotide-gapped DNA (gap-DNA) and AP-DNA inc are substrates of both APE1 (in the 3′-5′-exonuclease reaction) and Polβ Using fluorescence-based approaches, we showed in a previous study that DNA intermediates of BER induce rearrangements of various protein–protein complexes and modulate the strength of the interaction [28]. To explore the possible involvement of the N-terminal region of APE1 in this modulation, fluorescence titration and FRET experiments with the full-length and truncated forms of APE1 were performed in the absence and presence of model DNA ligands (shown in Figure S1). A double-stranded DNA with a synthetic abasic site (a tetrahydrofuran residue, F), AP-DNA, is an initial BER substrate of APE1. A 1-nucleotide gapped DNA with a 50 -F group at the margin of the gap models a product of the APE1-catalyzed incision (AP-DNA inc) and a stable analog of the Polβ substrate not processed by the 50 -deoxyribose phosphate lyase activity. A 1-nucleotide-gapped DNA (gap-DNA) and AP-DNA inc are substrates of both APE1 (in the 30 -50 -exonuclease reaction) and Polβ (in the DNA repair synthesis). APE1 and Polβ bind with the highest affinity AP-DNA and gap-DNA, respectively; their affinities for AP-DNA inc are relatively high and comparable, and those for the nonsubstrate double-stranded DNA are very low [31]. XRCC1 prefers 1-nt-gapped and nicked oligonucleotide duplexes, while there is no significant difference in the strength of PARP1 interaction with different duplex models of damaged DNA due to preferential binding to blunt ends [32–36]. Functional interactions between APE1, Polβ, XRCC1, and PARP1 on various DNA intermediates, contributing to regulation of BER, were demonstrated in previous studies [27,31,37–42] and are further detailed in Section 2.3.

42] and are further detailed in Section 2.3.

Binding of the truncated forms of APE1 with each protein partner in the absence and presence of a given DNA intermediate was explored in parallel with the full-length APE1 (Figure 4). From these experiments, the effects produced by the model DNA on the quantitative characteristics of protein–protein interactions, apparent binding affinity constant and FRET efficiency, were determined (Table 2). AP-DNA inc produced the highest effect on the binding affinity of FAM-APE1 for Polβ (two-fold decrease), while its effects detected for the respective complexes formed by FAM-APE1N∆35 and FAM-APE1N∆61 were not statistically significant. The presence of intact AP-DNA revealed no significant effect on the binding affinity of both the full-length and truncated forms of APE1 for Polβ. Gap-DNA produced a small effect (1.3-fold decrease detected as statistically significant) for the FAM-APE1 complex. Changes in the efficiency of FRET detected for complexes of the truncated forms of APE1 with TMR-Polβ in the presence of the various DNA intermediates (a 7–9% decrease) were indicative of the DNA-induced rearrangement of the complexes. The different DNAs produced significantly different changes in the FRET signals only in the case of the full-length protein complex (a 5% increase vs. a 6–11% decrease). These combined results suggest that conformational changes in the APE1–Polβ complex caused by DNA binding involve both the conserved catalytic core of APE1 and the N-terminal extension, with the last being important to control the strength of interaction between proteins in the complex with damaged DNA depending on its type. of a given DNA intermediate was explored in parallel with the full-length APE1 (Figure 4). From these experiments, the effects produced by the model DNA on the quantitative characteristics of protein–protein interactions, apparent binding affinity constant and FRET efficiency, were determined (Table 2). AP-DNA inc produced the highest effect on the binding affinity of FAM-APE1 for Polβ (two-fold decrease), while its effects detected for the respective complexes formed by FAM-APE1NΔ35 and FAM-APE1NΔ61 were not statistically significant. The presence of intact AP-DNA revealed no significant effect on the binding affinity of both the full-length and truncated forms of APE1 for Polβ. Gap-DNA produced a small effect (1.3-fold decrease detected as statistically significant) for the FAM-APE1 complex. Changes in the efficiency of FRET detected for complexes of the truncated forms of APE1 with TMR-Polβ in the presence of the various DNA intermediates (a 7– 9% decrease) were indicative of the DNA-induced rearrangement of the complexes. The different DNAs produced significantly different changes in the FRET signals only in the case of the full-length protein complex (a 5% increase vs. a 6–11% decrease). These combined results suggest that conformational changes in the APE1–Polβ complex caused by DNA binding involve both the conserved catalytic core of APE1 and the N-terminal extension, with the last being important to control the strength of interaction between proteins in the complex with damaged DNA depending on its type.

*Int. J. Mol. Sci.* **2020**, *21*, x 6 of 19

Binding of the truncated forms of APE1 with each protein partner in the absence and presence

(in the DNA repair synthesis). APE1 and Polβ bind with the highest affinity AP-DNA and gap-DNA, respectively; their affinities for AP-DNA inc are relatively high and comparable, and those for the nonsubstrate double-stranded DNA are very low [31]. XRCC1 prefers 1-nt-gapped and nicked oligonucleotide duplexes, while there is no significant difference in the strength of PARP1 interaction with different duplex models of damaged DNA due to preferential binding to blunt ends [32–36]. Functional interactions between APE1, Polβ, XRCC1, and PARP1 on various DNA

**Figure 4.** Influence of DNA intermediate of base excision DNA repair (BER) on the interaction of APE1 and its truncated form with Polβ. Fluorescence titration of FAM-labelled APE1 or APE1NΔ61 (40 nM) with Polβ was performed in the absence (open symbols) or presence (filled symbols) of incised AP-DNA (400 nM); higher values of F0 measured in DNA presence are indicative of protein–DNA binding. Curves show best fits of four-parameter equation with R2 values exceeding 0.98; EC50 values derived from respective curves are presented. **Figure 4.** Influence of DNA intermediate of base excision DNA repair (BER) on the interaction of APE1 and its truncated form with Polβ. Fluorescence titration of FAM-labelled APE1 or APE1N∆61 (40 nM) with Polβ was performed in the absence (open symbols) or presence (filled symbols) of incised AP-DNA (400 nM); higher values of F<sup>0</sup> measured in DNA presence are indicative of protein–DNA binding. Curves show best fits of four-parameter equation with R<sup>2</sup> values exceeding 0.98; EC<sup>50</sup> values derived from respective curves are presented.

Both the structure and stability of the FAM-APE1 complex with XRCC1 were detected as described previously to be modulated in the presence of gap-DNA [28]: The binding affinity and FRET efficiency increased 1.5-fold and by 8%, respectively (Table 2). However, the parameters of FAM-APE1 interaction with XRCC1 were not affected by the presence of intact AP-DNA. Here, we revealed that interaction of FAM-APE1NΔ35 and FAM-APE1NΔ61 with XRCC1 was appreciably influenced by both gap-DNA and AP-DNA. For each of the truncated forms, the effects produced by the different DNAs on the strength of interaction with XRCC1 and on the FRET signals were very Both the structure and stability of the FAM-APE1 complex with XRCC1 were detected as described previously to be modulated in the presence of gap-DNA [28]: The binding affinity and FRET efficiency increased 1.5-fold and by 8%, respectively (Table 2). However, the parameters of FAM-APE1 interaction with XRCC1 were not affected by the presence of intact AP-DNA. Here, we revealed that interaction of FAM-APE1N∆35 and FAM-APE1N∆61 with XRCC1 was appreciably influenced by both gap-DNA and AP-DNA. For each of the truncated forms, the effects produced by the different DNAs on the strength of interaction with XRCC1 and on the FRET signals were very similar. Clearly, integrity of the N-terminal portion of APE1 is important to control the mode of APE1–XRCC1 interaction during the processing of distinct DNA intermediates.

We compared the influence of AP-DNA and gap-DNA on the interaction of the full-length and truncated forms of APE1 with PARP1 (Table 2). Significant effects on the interaction of FAM-APE1 with PARP1 were detected only in the presence of AP-DNA: Binding affinity decreased 1.7 fold and FRET efficiency increased by 5% (Table 2). Both AP-DNA and gap-DNA produced similar positive effects

(a 1.4-fold increase) on the strength of interaction between FAM-APE1N∆61 and PARP1. In the case of FAM-APE1N∆35, no statistically significant effects of DNA intermediates on the protein interaction with PARP1 were detected. Interestingly, EC<sup>50</sup> values determined for the complexes of FAM-APE1 and FAM-APE1N∆61 with PARP1 were identical in the presence of gap-DNA, but the affinity of FAM-APE1 for PARP1 in the presence of AP-DNA was practically identical with that of FAM-APE1N∆61 in the absence of DNA (76 nM vs. 75 nM). Evidently, the N-terminal extension is responsible for the destabilizing effect of the preferred binding substrate of APE1 on APE1–PARP1 interaction.


**Table 2.** Effects of BER DNA intermediates on protein–protein interactions.

<sup>a</sup> Titration experiments performed at the constant concentrations of FAM-labelled protein (40 nM) and DNA (400 nM). Underlined data for FAM-APE1 (previously published [28]) reproduced to compare binding parameters of full-length and truncated forms. <sup>b</sup> Parameters derived from titration curves by fitting to four-parameter equation. Values are the mean (± SD) of at least three independent experiments. Values determined for each protein pair in the presence of DNA, which were statistically different from the respective value in the absence of DNA, are marked *p* < 0.05 (\*), *p* < 0.01 (\*\*), *p* < 0.001 (\*\*\*). <sup>c</sup> Effect of DNA on protein–protein affinity determined as the ratio of EC<sup>50</sup> values in the presence or absence of DNA. <sup>d</sup> Increase (+) or decrease (–) in FRET efficiency between FAM- and TMR (5(6)-carboxytetramethylrhodamine)-labelled proteins in the presence of DNA. Statistically significant changes in E values produced by DNA are marked *p* < 0.05 (\*), *p* < 0.01 (\*\*).
