Direct and Auger Electron-Induced, Single- and Double-Strand Breaks on Plasmid DNA Caused by 99mTc-Labeled Pyrene Derivatives and the Effect of Bonding Distance

It is evident that 99mTc causes radical-mediated DNA damage due to Auger electrons, which were emitted simultaneously with the known γ-emission of 99mTc. We have synthesized a series of new 99mTc-labeled pyrene derivatives with varied distances between the pyrene moiety and the radionuclide. The pyrene motif is a common DNA intercalator and allowed us to test the influence of the radionuclide distance on damages of the DNA helix. In general, pUC 19 plasmid DNA enables the investigation of the unprotected interactions between the radiotracers and DNA that results in single-strand breaks (SSB) or double-strand breaks (DSB). The resulting DNA fragments were separated by gel electrophoresis and quantified by fluorescent staining. Direct DNA damage and radical-induced indirect DNA damage by radiolysis products of water were evaluated in the presence or absence of the radical scavenger DMSO. We demonstrated that Auger electrons directly induced both SSB and DSB in high efficiency when 99mTc was tightly bound to the plasmid DNA and this damage could not be completely prevented by DMSO, a free radical scavenger. For the first time, we were able to minimize this effect by increasing the carbon chain lengths between the pyrene moiety and the 99mTc nuclide. However, a critical distance between the 99mTc atom and the DNA helix could not be determined due to the significantly lowered DSB generation resulting from the interaction which is dependent on the type of the 99mTc binding motif. The effect of variable DNA damage caused by the different chain length between the pyrene residue and the Tc-core as well as the possible conformations of the applied Tc-complexes was supplemented with molecular dynamics (MD) calculations. The effectiveness of the DNA-binding 99mTc-labeled pyrene derivatives was demonstrated by comparison to non-DNA-binding 99mTcO4–, since nearly all DNA damage caused by 99mTcO4– was prevented by incubating with DMSO.

Tc is commonly used as a γ-emitter for imaging purposes in nuclear medicine. Due to its decay properties (γ-radiation at 141 keV, t 1/2 = 6.0 h) it can be detected using SPECT (single photon emission computer tomography) imaging techniques. In addition to the imageable 141 keV γ-emission, 99m Tc also converts by internal conversion resulting in emission of low energy Auger and conversion electrons (see Table A in S1 File). However, the resulting effects due to these Auger electron emissions are still controversial [1,2].
More than 50% of all radionuclides decay by internal conversion or/and electron capture causing the emission of a cascade of very low energy electrons. This phenomenon is known as the Auger effect. The main advantage of Auger electrons is that they behave like high LET (linear energy transfer) radiation despite their low emission energy. Therefore, Auger electrons can induce various types of DNA damage such as multiple double-strand breaks (DSB) when they are localized in close proximity to the DNA [3][4][5][6]. The diagnostic and therapeutic use of Auger emitters, especially 125 I ( 123 I) or 111 In, has been widely discussed [7][8][9][10]. In addition to these prominent Auger emitters, 99m Tc emits almost five low energy electrons with a range of less than 200 nm [1,11] and initial tests demonstrated that it might be possible to use 99m Tc for Auger electron therapy [12,13]. Cell experiments proved that Auger electrons emitted by 99m Tc can initiate cell death [14] but the damage was not clearly quantified. The use of double-stranded plasmid DNA as a biological model enables the investigation of unprotected interactions between molecules and supercoiled DNA (SC) allowing the distinction between single-strand breaks (SSB) and double-strand breaks. The appearance of open circle DNA (OC) is commonly used as an indicator of SSB and linear DNA (L) is the result of DSB.
Previous studies on the therapeutic effect of Auger electrons from 99m Tc provided information about the DNA damage caused by 99m Tc interaction with plasmid DNA [15] [16].To better understand the biophysical mechanisms underlying the induction of DNA single-and doublestrand breaks caused by the emitted Auger electrons, we translated Balagurumoorthy's studies regarding to the effect of distance between 125 I-labeled Hoechst derivatives and plasmid DNA to 99m Tc. They were able to demonstrate the damage of DNA provoked by Auger electron emitters especially of 125 I in dependance on the distance between the radionuclide and the DNA strand [3,17].
In most cases, radioiodine was covalently bound to several intercalating molecules [3,18]. Conversely, 99m Tc is exclusively bound to the molecules via complexation using chelating compounds; a direct covalent bond to the Tc is not feasible. Therefore, a tridentate chelating unit to bind the 99m Tc-(or 186,188 Re)-tricarbonyl moiety has become a well established system for the complexation of Tc and Re radionuclides [19][20][21]. Additionally, pyrene is known to be a DNA intercalating molecule [12]. For these reasons, we designed and synthesized four new pyrene derivatives with different-sized linkers using a tridentate chelating unit for complexation of 99m Tc. The primary aim of our study was to learn more about the properties of Auger electrons emitted by 99m Tc. To achieve this aim, we prepared a 99m Tc-complex with an optimal distance between the 99m Tc(CO) 3 -core and the pyrene residue to position the 99m Tc in close proximity to the plasmid DNA to induce direct SSB and DSB. By increasing the distance between the DNA-intercalating moiety and the bonding moiety for 99m Tc, we expected to observe decrease of direct DNA damages. This distance dependance has not been reported for 99m Tc until now. Additionally, we sought to differentiate between indirect effects of radicals caused by the radiolysis of water and direct destruction caused by Auger electrons to DNA. Therefore, DMSO, a commonly known radical scavenger, was incubated in a second set of experiments to illustrate the effect on DNA damage [22]. Finally, we compared the effects of the DNA-binding 99m Tc pyrene derivatives with that of non-DNA binding 99m TcO 4 -.

Materials and Methods
Pyrene 9 was synthesized according to Hafliger et al. [12], compound 6a was prepared according to Pretze and Mamat [23] and 6b was prepared according to Adres-Guisot et al. [24]. All syntheses of the pyrene chelators as well as the radiolabeling procedures can be found in detail in the Supporting Information (S1 File).

pUC 19 plasmid DNA
The pUC 19 plasmid (2686 base pairs) was purchased from New England Biolabs (Ipswich, UK). The DNA stock solution was diluted in TE-Buffer (10 mM TRIS-HCl, 1 mM EDTA, pH 7.5) to obtain a final concentration of 0.1 μg/μL (aliquots were stored at -20°C). Only samples containing more than 95% of supercoiled plasmid DNA were used for our studies. To obtain linear plasmid DNA, pUC 19 was incubated with BamHI (Invitrogen, Karlsruhe, Germany).

Incubation procedure and gel electrophoresis
The total volume of each incubation was brought up to 20 μL with purified water. For characterizing the DNA damage, a second set of experiments was performed as described above except the samples were incubated with DMSO (Roth, Karlsruhe, Germany) instead of water with a final concentration of 0.2 M DMSO.
After irradiation, 1.25 μL of blue juice gel loading buffer (Invitrogen, Karlsruhe, Germany) was added to 10 μL of each sample. The samples were then pipetted in wells of a 1.4% agarosegel in Tris-Acetate-EDTA buffer (TAE buffer, Sigma Aldrich). Gel electrophoresis (Sub Cell GT, BioRad, Munich, Germany) was run at 120 V for 100 min. After electrophoresis, the gel was stained with an ethidium bromide solution (0.5 μg/mL) for 30 min and three different conformations of pUC 19 plasmid DNA were identifiable (SC, OC and L). Each fraction resulted in different DNA bands and were separated by electrophoresis due to their different sizes and loadings.
The different conformation forms of pUC 19 were visualized by using an UV transilluminator (Diana III Digital Imaging System, Raytest, Straubenhardt, Germany). The DNA bands were imaged by a charge-coupled device (CCD) camera. For the interpretation of our results, we assumed that the sum of SC, OC and L plasmid fraction equaled 100%. The SC plasmid band was the native one. If there was any DNA damage, two other conformations appeared. The OC plasmid DNA corresponds to the fraction of induced SSB, while the L plasmid fraction corresponds to the fraction of induced DSB. The DNA damage was quantified by integrating the fluorescing intensities of the three possible DNA conformations. The yields of each conformation form of pUC 19 were quantified by integrating the corresponding fluorescing intensities by using the open-source platform software Fiji [25].

Calculations of SSB and DSB
The SSB and DSB calculations were based on the assumption that the binding of 99m Tc-labeled pyrene derivatives to DNA and thus the strand breaks follow a Poisson distribution. The average of SSB (X SSB ) and DSB (X DSB ) was calculated from the fraction of linear DNA (F L ) and the remaining supercoiled DNA fraction (F SC ) determined for each sample with a defined amount of activity.
Assuming that 200 ng of pUC 19 plasmid DNA contain 6.88Á10 10 plasmid molecules, the calculation of SSB and DSB per plasmid are as follows: Incubating an activity of 15 MBq of 99m Tc-labeled pyrene derivatives or 99m TcO 4 -, respectively, within an incubation time of 24 h resulted in a total amount of 4.38Á10 11 decays. Refering the total number of decays within 24 h to the total number of plasmid DNA molecules leads to an amount of 6.37 decays per plasmid molecule. Combining all information, it is possible to calculate the yield of SSB (Y SSB ) and DSB (Y DSB ) per plasmid molecule per decay as follows. Results

Preparation of the chelators
To study the dependance on the distance of the Tc(CO) 3 -core to the DNA, three pyrene derivatives were chosen and compared regarding the influence of DNA damage and association to different chain lengths between the pyrene moiety and the radionuclide. In this case, the dipicolylamine (DPA) residue acts as tridentate ligand, which is known to be an excellent chelator for the Tc(CO) 3 as well as for the Re(CO) 3 core [26]. The first pyrene derivative 4 was prepared according to a modified synthesis procedure [27] from compound 3, which was reacted with 2-(chloromethyl)pyridine to yield the desired pyrene derivative 4 in 70% yield. The preparation of the pyrene derivatives 8a and 8b with longer alkyl chains was realized using two azides 7a,b which were obtained from the reaction of 5 with azidoalkyl tosylates 6a and 6b, respectively, in yields of 59% and 55%. In the next step, compounds 7a,b were first treated with triphenyl phosphane at 60°C for 3 h. Subsequently, 1-pyrenebutyric acid N-hydroxysuccinimide ester was added at ambient temperature to this mixture and the desired pyrenes 8a,b were obtained in yields of 57% and 61%, respectively. The reaction path is pointed out in Fig 1. Further details can be found in the Supporting Information (S1 File).

99m
In addition to the three pyrene derivatives 4, 8a and 8b, a literature known pyrene 9 with an aliphatic tridentate binding moiety for the Tc(CO) 3 -core was synthesized [12].
All radiolabeled derivatives were analyzed and purified by HPLC. It was possible to separate the radiolabeled product from the initial pyrene derivative to ensure a maximum specific activity of the product. The radiochemical purity of the isolated final products was greater than 95%, a representative chromatogram is shown in Fig 2.

Increase of open circular and linear plasmid conformation
The increase of induced SSB and DSB as a function of added activity is illustrated in Fig 5 as noted by the increased fractions of OC and L of the pUC 19 plasmid DNA with increased activities. It was observed that 99m TcO 4 induced the highest quantity of SSB and even a small . To identify whether the induced SSB and DSB by the [ 99m Tc]pyrenes were caused directly or indirectly (e.g. by radiolysis of water) we performed the same experiments and incubated 0.2 M DMSO additionally as a free radical scavenger. All resulting data is illustrated in Fig 6. The results in Fig 6 clearly [2].
The proximity between the Tc-core and the pyrene residue is not only dependent on the chain length but also on the conformation of the derivative itself. Thus, we predicted that both the structural change as well as the conformation could alter the distance between the Tc-core and the pyrene moiety; effectively placing the Tc in close proximity to the pyrene even with longer spacers. With this in mind, 100-ns molecular dynamics (MD) simulations were performed to get conformational sampling of the Tc-pyrene derivatives Ia, Ib, II, and III in the absence and presence of DNA. Details of the methodology and analysis of results for these simulations are included in the Supporting Information (S1 File). In broad terms, distances between pyrene moiety and Tc-core for Ia and Ib took short values (as expected) and II and III were in extended and twisted conformations. It is noteworthy, that Ib was very rigid due to the combination of DPA characteristics with a very short spacer (Tc-pyrene distances are around 7.5 Å in the absence and presence of DNA). Conversely, Ia adopted two sets of conformations when DNA is present due to the higher mobility of diethylenetriamine chelator exhibiting Tc-pyrene distances around 5.5 and 6.8 Å. Furthermore, the extended conformations were preferred over the twisted ones for II and III when DNA is present, see Fig 7. Comparing the proximities of the Tc-core and DNA as calculated by MD simulations, a closer proximity was calculated for Ia, followed by Ib, with II and III being further away.
99m TcO 4 behaves as a non-DNA-bonding molecule. Concerning the plasmid DNA damage caused by 99m TcO 4 -(3-15 MBq), we found that most DNA damages could be avoided by the additional incubation of 0.2 M DMSO, because it functions as a scavenger of reactive, radical oxygen species, especially hydroxyl radicals. Therefore, almost all of SSB and DSB by 99m TcO 4 were shown to be indirect DNA damage caused by radiolysis of water. For the tested 99m Tc-labeled pyrene derivatives, we observed that the type of DNA damage was strongly dependent on the distance between the DNA and the decaying radionuclide 99m Tc. By using pyrenes as a basis for intercalating molecules, we expected a specific bonding of our 99m Tc-labeled pyrenes to the minor grooves of the DNA, but it must be noted that all the pyrene derivatives used also acted as free radical scavengers and prevented indirect DNA damage themselves ( Ib was also dependent on the selected chelating unit for the 99m Tc(CO) 3 -core. MD simulations gave also some evidences of the effects of these chelators on distances between Tc and pyrene for Ia and Ib, because Ia has a shorter distance than Ib. However, Ia has also conformations with a distance similar to Ib, but Ib is more rigid than Ia. Recently, Chung et al. reported similar effects for [ 99m Tc]Ia with slightly differences in the relation of DSB vs. ESB [2]. In contrast to our recent studies they reported that the majority of DNA damages is caused by radiolysis of water. The chelating unit of [ 99m Tc]Ia consists of an aryl amine moiety whereas the chelating moiety of [ 99m Tc]Ib consists of two connected pyridine residues. Thus, there might be the possibility of a scavenging effect induced by the DPA unit.
The plasmids were incubated with 99m Tc-labeled pyrenes of high specific activity without an excess of unlabeled chelators. The maximum incubated 99m Tc activity of 15 MBq resulted in  4.7Á10 11 atoms of 99m Tc and approx. 10Á10 11 atoms of long-lived 99 Tc. An amount of 200 ng of pUC 19 plasmid DNA corresponds to 6.88Á10 10 plasmid molecules. Normally, pyrene molecules bind to double-stranded DNA in the minor groove. The distance between two minor grooves is determined to be 10 base pairs. Therefore, we assumed that there are 270 binding sites per plasmid for 99m Tc-labeled pyrene derivatives. For our labeling procedure, approx. seven 99m Tc-labeled pyrene molecules (plus approx. fourteen 99 Tc-labeled pyrene molecules) per plasmid corresponded to 15 MBq of activity in one assay. However, this result demonstrated that every available pyrene molecule could be bound to plasmid DNA.
This activity dependent DNA damage was shown to be caused by the emitted Auger electrons and not by the long-distance γ-rays of 99m Tc. This result indicated that the impact of the Auger events of [ 99m Tc]Ia and [ 99m Tc]Ib was occurred either direct or indirect with an extremely high local energy transfer and could not be prevented by DMSO or other radical scavengers like the pyrenes themselves. SSB are usually assumed to be radical-mediated. However, we demonstrated that the induction of SSB by DNA-bound pyrenes [ 99m Tc]Ia and [ 99m Tc]Ib might occur via another mechanism. It is possible that these SSB were also caused by direct interactions of the low-energy electrons emitted and plasmid-bound 99m Tc-pyrene derivatives could merely damage a single strand as assumed by Kotzerke et al. [16]. On average, 4.9 electrons of 99m Tc are emitted per decay. Those events create a dramatic charge transfer in the electron shell, which can cause a Coulomb explosion of the molecule, as discussed by Pomplun and Sutmann [28].
By comparing our studies with Balagurumoorthy's results, several differences emerge. They synthesized a series of 125 I-labeled Hoechst derivatives with a minor groove-binding motif and an increasing distance between the radiolabeled molecule and the binding position (ranging from 10.5 Å to 13.9 Å) [3]. The DSB yield decreased with distance from 0.50 DSB/decay to 0.00 DSB/decay, demonstrating the crucial influence of relatively small distances. Although DMSO could not prevent the DSB induced damage at the closest distance, their occurrence was reduced (approx. 10-fold) at higher distances. Comparing these results with our results in Table 3, a lower yield of DSB is found in our experiments. The difference (0.044 vs. 0.50 DSB per plasmid and decay) is explainable by comparing the different characteristics of the emitted Auger electrons of 99m Tc and 125 I. A commonly used DNA structure model assumes that the diameter of the DNA helix is 2 nm. Therefore, we concluded that the difference arises from Auger electrons which have a range lower than 2 nm. Comparing the spectra of 99m Tc and 125 I, there are two relevant Auger electrons per decay for 99m Tc but 18 for 125 I [1,29]. Therefore, it is conclusive that 125 I can cause direct DSB in a higher efficiency than 99m Tc. Regarding the maximum distance for an effect caused by Auger electrons, we found a value of 1.3 nm for 99m Tc, which is similar to that found by Balagurumoorthy [3].

Conclusion
In this study, four new 99m Tc-labeled pyrene derivatives were developed and their direct DNA damages were quantified for both DSB and SSB using pUC 19 plasmid DNA as a model. It is evident that 99m Tc can develop radical-mediated DNA damage and pyrene is a known DNA intercalating molecule. We demonstrated that direct DSB as well as direct SSB to plasmid DNA occurred in high efficiency due to the emitted Auger electrons of 99m Tc when the radionuclide is located in close proximity ([ 99m Tc]Ia and [ 99m Tc]Ib). The result showed that this effect could not be prevented by DMSO, a free radical scavenger, providing further proof of the direct action of Auger electrons on DNA. For the first time, we were able to extinguish this effect for 99m Tc by increasing the distance of the pyrene moiety and the chelating unit for 99m Tc by using linkers with different carbon chain lengths. In contrast to the DNA binding 99m Tc-labeled pyrene molecules, 99m TcO 4 was applied as a non-DNA binding molecule in our experiments, since nearly the whole DNA damage caused by 99m TcO 4 is preventable by incubating DMSO.
In summary, it was shown that Auger electrons from 99m Tc can destroy DNA by direct interaction, however, only if the isotope is placed directly at the DNA. Clinical relevant Auger electron tumor therapy is hampered by the prerequisite of DNA binding which is hindered by cell and nucleus membranes. Vice versa this shield prevents critical damage in diagnostic application in nuclear medicine.
Supporting Information S1 File. Syntheses of the pyrene chelators and details of the computational molecular dynamics calculations.