*2.1. MD Modeling*

Molecular dynamics (MD) modeling of the complex of multidomain human PARP-1 with DNA and 7-MG (Figure S1) has been performed for the first time. 7-MG was docked into its putative binding region (binding site of the NAD<sup>+</sup> nicotinamide group), and the obtained complex was subjected to MD simulation in explicit solvent. The system contained 703 amino acid residues, 52 nucleotides, 2 Zn2<sup>+</sup> ions, 45 Na<sup>+</sup> ions, and 73,302 water molecules (232,862 atoms). Analysis of the 20-ns equilibrium simulation trajectory revealed the following important intermolecular interactions. The lactam group and 2-amino group of 7-MG form hydrogen bonds with the Gly863 residue (Figure 1 and Table 1). 7-Methyl group forms a hydrophobic contact with the Ala898 side chain, and purine rings stack with the Tyr907 side chain. An additional hydrogen bond is formed between the 7-MG lactam group and the Ser904 side chain. This interaction is characterized by an increased mean distance between atoms (Table 1) because Ser904 periodically forms an alternative hydrogen bond with the Trp861 backbone and thus can be observed in two possible conformations (Figure S2a,b). However, the hydrogen bond of Ser904 with 7-MG is more prevalent, with the occupancy of about 60% (Figure S2c). As demonstrated in Figures S3 and S4, the position of 7-MG and its hydrogen bonds with protein were stable during the simulation. Noticeably, we did not observe significant changes in the PARP-1 multidomain organization upon 7-MG binding (compared with the PARP-1–DNA crystal structure); thus the inhibitor apparently does not affect the interaction between PARP-1 and DNA in the case of double-stranded oligonucleotides.

*Int. J. Mol. Sci.* **2020**, *21*, 2159 3 of 11

**Figure 1.** Interactions of 7‐MG molecule in the PARP‐1 active site revealed by MD simulation: hydrogen bonds with Gly863 and Ser904, *π*‐stacking of purine rings with Tyr907, and hydrophobic contact between the 7‐MG methyl group and Ala898. **Figure 1.** Interactions of 7-MG molecule in the PARP-1 active site revealed by MD simulation: hydrogen bonds with Gly863 and Ser904, π-stacking of purine rings with Tyr907, and hydrophobic contact between the 7-MG methyl group and Ala898.

**Table 1.** Distances and angles describing 7‐MG position in the PARP‐1 active site determined by the 20‐ns MD simulation. Mean values are presented with standard deviations. **Table 1.** Distances and angles describing 7-MG position in the PARP-1 active site determined by the 20-ns MD simulation. Mean values are presented with standard deviations.


7‐MG:NH2:N ∙∙∙ 7‐MG:NH2:H ∙∙∙ Gly863:O 138 ± 10 <sup>1</sup> Distance between the geometric center of 7‐MG fused rings and the center of the Tyr907 benzene ring. <sup>1</sup> Distance between the geometric center of 7-MG fused rings and the center of the Tyr907 benzene ring.

#### *2.2. Fluorescence Anisotropy Analysis 2.2. Fluorescence Anisotropy Analysis*

PARP‐1 inhibition by 7‐MG was studied using analysis of the fluorescence anisotropy of PARP‐ 1 complexes with a labeled double‐stranded oligonucleotide. Upon PARP‐1 binding to DNA, the anisotropy level is increased due to a decrease in the fluorophore mobility. The addition of the NAD+ substrate enables synthesis of negatively charged PAR and automodification of the enzyme, leading to the dissociation of the PARP‐1–DNA complexes accompanied by a decrease in anisotropy. The observed dissociation rate is proportional to the PARP‐1‐catalyzed reaction rate [27]. PARP-1 inhibition by 7-MG was studied using analysis of the fluorescence anisotropy of PARP-1 complexes with a labeled double-stranded oligonucleotide. Upon PARP-1 binding to DNA, the anisotropy level is increased due to a decrease in the fluorophore mobility. The addition of the NAD<sup>+</sup> substrate enables synthesis of negatively charged PAR and automodification of the enzyme, leading to the dissociation of the PARP-1–DNA complexes accompanied by a decrease in anisotropy. The observed dissociation rate is proportional to the PARP-1-catalyzed reaction rate [27].

We have found that 7‐MG does not affect the oligonucleotide structure by itself and does not significantly interfere with the binding of PARP‐1 to DNA (Figure S5), but inhibits the dissociation of the PARP‐1–DNA complex (due to suppression of PARP‐1 catalytic activity). Figure 2 shows a typical plot of the reaction rate as a function of 7‐MG concentration. The absolute IC50 [28] was found to be 162 ± 4 μM at 100 μM NAD+ concentration. To determine the type of inhibition, we calculated apparent *K*<sup>M</sup> and *V*max values at various 7‐MG concentrations (Figure 3). The 7‐MG addition altered only the *K*Mapp leaving the *V*max value the same, which displays the competitive enzyme inhibition; the corresponding *K*<sup>i</sup> value was found to be 61 ± 9 μM. We have found that 7-MG does not affect the oligonucleotide structure by itself and does not significantly interfere with the binding of PARP-1 to DNA (Figure S5), but inhibits the dissociation of the PARP-1–DNA complex (due to suppression of PARP-1 catalytic activity). Figure 2 shows a typical plot of the reaction rate as a function of 7-MG concentration. The absolute IC<sup>50</sup> [28] was found to be 162 <sup>±</sup> 4 <sup>µ</sup>M at 100 <sup>µ</sup>M NAD<sup>+</sup> concentration. To determine the type of inhibition, we calculated apparent *K*<sup>M</sup> and *V*max values at various 7-MG concentrations (Figure 3). The 7-MG addition altered only the *K*<sup>M</sup> app leaving the *V*max value the same, which displays the competitive enzyme inhibition; the corresponding *K*<sup>i</sup> value was found to be 61 ± 9 µM.

*Int. J. Mol. Sci.* **2020**, *21*, 2159 4 of 11

*Int. J. Mol. Sci.* **2020**, *21*, 2159 4 of 11

**Figure 2.** Dependence of the PARP‐1‐catalyzed reaction rate on the concentration of 7‐MG inhibitor determined by fluorescence anisotropy (100 μM NAD+ concentration). **Figure 2.** Dependence of the PARP-1-catalyzed reaction rate on the concentration of 7-MG inhibitor determined by fluorescence anisotropy (100 µM NAD<sup>+</sup> concentration). **Figure 2.** Dependence of the PARP‐1‐catalyzed reaction rate on the concentration of 7‐MG inhibitor by fluorescence anisotropy concentration).

**Figure 3.** Dependence of the PARP‐1‐catalyzed reaction rate on the NAD+ concentration at different concentrations of 7‐MG added to the reaction mixture. Insert: calculated *K*Mapp values increase with increasing 7‐MG concentrations, thus demonstrating the competitive inhibition mechanism. **Figure 3.** Dependence of the PARP‐1‐catalyzed reaction rate on the NAD+ concentration at different concentrations of 7‐MG added to the reaction mixture. Insert: calculated *K*Mapp values increase with increasing 7‐MG concentrations, thus demonstrating the competitive inhibition mechanism. **Figure 3.** Dependence of the PARP-1-catalyzed reaction rate on the NAD<sup>+</sup> concentration at different concentrations of 7-MG added to the reaction mixture. Insert: calculated *K*<sup>M</sup> app values increase with increasing 7-MG concentrations, thus demonstrating the competitive inhibition mechanism.

#### *2.3. spFRET Analysis 2.3. spFRET Analysis 2.3. spFRET Analysis*

To model the inhibitor activity in the chromatin environment, we have studied effects of 7‐MG on the PARP‐1 complexes with fluorescently labeled nucleosomes P147 and P167 using single‐ particle Förster resonance energy transfer (spFRET) microscopy. Cy3 and Cy5 labels were introduced in the neighboring DNA gyres and served to probe DNA conformation near the entrance of DNA into the nucleosome by measuring FRET efficiency between the labels. To model the inhibitor activity in the chromatin environment, we have studied effects of 7‐MG on the PARP‐1 complexes with fluorescently labeled nucleosomes P147 and P167 using single‐ particle Förster resonance energy transfer (spFRET) microscopy. Cy3 and Cy5 labels were introduced in the neighboring DNA gyres and served to probe DNA conformation near the entrance of DNA into the nucleosome by measuring FRET efficiency between the labels. To model the inhibitor activity in the chromatin environment, we have studied effects of 7-MG on the PARP-1 complexes with fluorescently labeled nucleosomes P147 and P167 using single-particle Förster resonance energy transfer (spFRET) microscopy. Cy3 and Cy5 labels were introduced in the neighboring DNA gyres and served to probe DNA conformation near the entrance of DNA into the nucleosome by measuring FRET efficiency between the labels.

spFRET microscopy revealed the presence of two subpopulations for both P147 and P167 nucleosomes in solution, which differed in FRET efficiency (presented as the proximity ratio EPR [29] in Figure 4). These subpopulations were observed in the calculated EPR profiles of nucleosomes (i.e., frequency distributions of nucleosomes by EPR) as two peaks. In agreement with our previous work [30], the major peaks (EPR ≈ 0.79 for P147 and EPR ≈ 0.52 for P167) can be assigned to nucleosome spFRET microscopy revealed the presence of two subpopulations for both P147 and P167 nucleosomes in solution, which differed in FRET efficiency (presented as the proximity ratio EPR [29] in Figure 4). These subpopulations were observed in the calculated EPR profiles of nucleosomes (i.e., frequency distributions of nucleosomes by EPR) as two peaks. In agreement with our previous work [30], the major peaks (EPR ≈ 0.79 for P147 and EPR ≈ 0.52 for P167) can be assigned to nucleosome spFRET microscopy revealed the presence of two subpopulations for both P147 and P167 nucleosomes in solution, which differed in FRET efficiency (presented as the proximity ratio EPR [29] in Figure 4). These subpopulations were observed in the calculated EPR profiles of nucleosomes (i.e., frequency distributions of nucleosomes by EPR) as two peaks. In agreement with our previous work [30], the major peaks (EPR ≈ 0.79 for P147 and EPR ≈ 0.52 for P167) can be assigned to nucleosome

subpopulations with tightly wrapped nucleosomal DNA, while the minor peaks (EPR ≈ 0.03 for P147 and P167) can be related to nucleosomes with partially unwrapped DNA and/or free DNA.

subpopulations with tightly wrapped nucleosomal DNA, while the minor peaks (EPR ≈ 0.03 for P147 and P167) can be related to nucleosomes with partially unwrapped DNA and/or free DNA. subpopulations with tightly wrapped nucleosomal DNA, while the minor peaks (EPR ≈ 0.03 for P147 and P167) can be related to nucleosomes with partially unwrapped DNA and/or free DNA. Differences in the EPR profiles of P147 and P167 nucleosomes are related with the presence of the 20-bp linker DNA arm in the P167 that seems to affect both wrapping of DNA on the surface of the histone octamer and DNA "breathing" (temporal spontaneous partial DNA unwrapping) near the entrance of DNA into nucleosome. *Int. J. Mol. Sci.* **2020**, *21*, 2159 5 of 11 bp linker DNA arm in the P167 that seems to affect both wrapping of DNA on the surface of the histone octamer and DNA "breathing" (temporal spontaneous partial DNA unwrapping) near the entrance of DNA into nucleosome.

**Figure 4.** spFRET studies of 7‐MG effects at the nucleosome level. Typical frequency distributions of P147 (**a**,**b**) and P167 (**c**,**d**) nucleosomes by EPR in different mixtures are shown. Conditions: 1 nM nucleosomes, 10 nM or 20 nM PARP‐1, 100 μM NAD+, 450 μM 7‐MG. Inserts: schemes of nucleosome structure and positions of fluorescent labels (asterisks). **Figure 4.** spFRET studies of 7-MG effects at the nucleosome level. Typical frequency distributions of P147 (**a**,**b**) and P167 (**c**,**d**) nucleosomes by EPR in different mixtures are shown. Conditions: 1 nM nucleosomes, 10 nM or 20 nM PARP-1, 100 µM NAD+, 450 µM 7-MG. Inserts: schemes of nucleosome structure and positions of fluorescent labels (asterisks).

The formation of the complexes between nucleosomes (P147 or P167) and PARP‐1 resulted in the appearance of a new peak characterized by EPR ≈ 0.38–0.41 in the EPR profiles (Figure 4). Relative intensity of this peak indicates that only a part of P147 nucleosomes formed complexes at the studied PARP‐1 concentrations (10–20 nM), and this subpopulation was increased after the increase in the PARP‐1 concentration. In contrast, nearly all P167 nucleosomes formed the complexes even at 10 nM concentration of PARP‐1, indicating higher affinity of PARP‐1 to nucleosomes with a linker DNA arm. The PARP‐1‐induced shift of the main EPR peak to lower values is related to structural changes in nucleosomal DNA, which are accompanied by an increase in the distance between neighboring DNA gyres in the region of Cy3 and Cy5 label position, i.e., at the H2A–H2B interface (+13 bp position) and the H4–H2B interface (+91 bp position) [30]. Although 7‐MG alone did not affect the nucleosome structure (Figure S6), its addition to P147 together with PARP‐1 resulted in a considerable increase of the fraction of nucleosomes with the EPR distribution having maximum at 0.42–0.47 (Figure 4a,b), thus indicating that (i) 7‐MG is involved in the formation of complexes of nucleosomes with PARP‐1, (ii) 7‐MG only weakly disturbs the structure of nucleosomal DNA in the PARP‐1–nucleosome complexes, (iii) 7‐MG facilitates PARP‐1 binding to nucleosomes. Similarly, The formation of the complexes between nucleosomes (P147 or P167) and PARP-1 resulted inthe appearance of a new peak characterized by EPR <sup>≈</sup> 0.38–0.41 in the EPR profiles (Figure 4). Relative intensity of this peak indicates that only a part of P147 nucleosomes formed complexes at the studiedPARP-1 concentrations (10–20 nM), and this subpopulation was increased after the increase in the PARP-1 concentration. In contrast, nearly all P167 nucleosomes formed the complexes even at 10 nM concentration of PARP-1, indicating higher affinity of PARP-1 to nucleosomes with a linker DNA arm. The PARP-1-induced shift of the main EPR peak to lower values is related to structural changes in nucleosomal DNA, which are accompanied by an increase in the distance between neighboring DNA gyres in the region of Cy3 and Cy5 label position, i.e., at the H2A–H2B interface (+13 bp position) and the H4–H2B interface (+91 bp position) [30]. Although 7-MG alone did not affect the nucleosome structure (Figure S6), its addition to P147 together with PARP-1 resulted in a considerable increase ofthe fraction of nucleosomes with the EPR distribution having maximum at 0.42–0.47 (Figure 4a,b), thus indicating that (i) 7-MG is involved in the formation of complexes of nucleosomes with PARP-1, (ii) 7-MG only weakly disturbs the structure of nucleosomal DNA in the PARP-1–nucleosome complexes, (iii) 7-MG facilitates PARP-1 binding to nucleosomes. Similarly, spFRET analysis demonstrated

however, the effect was less pronounced than in the case of P147 nucleosome (Figure 4c,d).

spFRET analysis demonstrated the ability of 7‐MG to promote PARP‐1 binding to P167 nucleosome;

Incubation of PARP‐1–nucleosome complexes with NAD+ resulted in the changes of the EPR

the ability of 7-MG to promote PARP-1 binding to P167 nucleosome; however, the effect was less pronounced than in the case of P147 nucleosome (Figure 4c,d).

Incubation of PARP-1–nucleosome complexes with NAD<sup>+</sup> resulted in the changes of the EPR profiles characterized by disappearance of the peak at EPR ≈0.38–0.41 (a signature of nucleosome-PARP-1 complexes) and appearance of the peak in the region of EPR ≈ 0.52 (for P167) or EPR ≈ 0.71 (for P147), which was characteristic for intact nucleosomes (Figure 4). The data indicate that, as expected, PARP-1 automodification resulted in its dissociation from nucleosomal DNA and almost complete recovery of the intact nucleosome structure. The presence of 7-MG has blocked the described changes in the EPR profiles of enzyme–nucleosome complexes induced by addition of NAD+: only a small shift of the peak at EPR ≈ 0.38–0.47 was observed without appearance of the subpopulation of free nucleosomes (no peaks with a maximum at EPR ≈ 0.52 for P167 or EPR ≈ 0.71–0.79 for P147). This effect of 7-MG is likely related to inhibition of PARP-1 enzymatic activity that prevents the automodification of PARP-1 and its dissociation from the nucleosomes.

### **3. Discussion**

In our earlier study, we have modeled a complex of PARP-1 with 7-MG [23]; however, the preliminary data were obtained using an isolated catalytic domain of chicken PARP-1 and might be incomplete. Here we report the MD model of human multidomain PARP-1 complex with DNA fragment and 7-MG that most accurately describes interactions between the inhibitor and PARP-1 bound to a DNA double-strand break. In the refined model, 7-MG formed polar and hydrophobic interactions, in particular hydrogen bonds with Gly863, similar to the NAD<sup>+</sup> substrate and known effective PARP-1 inhibitors [15,25,26]. One important difference from chicken PARP-1 model was the formation of an additional hydrogen bond between 7-MG and Ser904 which apparently contributes to the binding of the inhibitor with human PARP-1. Revealed interactions of 7-MG in the PARP-1 active site and the stability of the enzyme–inhibitor complex during MD simulation suggest that 7-MG occupies the binding site of the NAD<sup>+</sup> nicotinamide group. Kinetics analysis of the ability of 7-MG to suppress PAR synthesis was performed using a recently developed fluorescent method for the real-time measurement of PARP-1 activity [27]; it corroborated conclusions of MD modeling that 7-MG is a competitive PARP-1 inhibitor.

To confirm the inhibitory properties at a more complex level, 7-MG effects have been investigated using spFRET microscopy, an advanced technique allowing analysis of structurally different subpopulations of nucleosomes in heterogeneous samples [31–33]. Mononucleosomes used in our study represented a convenient model of DNA double-strand break in the chromatin environment. Although 7-MG exerted no significant effect on the PARP-1 interaction with isolated double-stranded oligonucleotides (as shown using MD modeling and fluorescence anisotropy experiments), spFRET data showed that it promotes PARP-1 binding to nucleosomes. It seems that the inhibitor stabilizes the interaction between PARP-1 and nucleosomal DNA. We propose that the relative orientation of PARP-1 domains undergoes changes to fit the nucleosome structure, and bound 7-MG stabilizes the new PARP-1 conformation. The formed enzyme–nucleosome complexes are nonproductive because of the concomitant 7-MG inhibition of PARP-1 catalytic activity. PARP-1 molecules fail to regulate their dissociation from DNA via PAR synthesis and get trapped on DNA. The trapped PARP-1 complexes are considered to be even more deleterious for cancer cells than unrepaired DNA strand breaks, because PARP-1 protein tightly bound to DNA interferes with transcription, replication, and DNA repair [34–36]. Noticeably, the most effective PARP-1 inhibitors, including olaparib, display the strong ability to trap PARP [37,38].

In conclusion, the molecular mechanisms of a promising anticancer compound, 7-MG, can be outlined as follows. (1) 7-MG forms substrate-specific interactions in the PARP-1 active site and inhibits the synthesis of PAR, a signal polymer that induces the reorganization of chromatin structure and recruits DNA repair proteins to eliminate the damage. (2) 7-MG inhibits the dissociation of PARP-1 from the DNA damage site in the context of nucleosome and likely prevents further steps in DNA repair, as well as DNA replication and transcription, inducing cancer cell death. Despite the fact that 7-MG is a weaker inhibitor compared to some synthetic PARP-1 inhibitors, we believe that this natural compound has more favorable profile of pharmacokinetics and toxicity, and therefore can be considered as a promising new component of chemotherapy.
