Traces of isotopic reactive species produced from a non-thermal plasma jet in bio-molecules

Heavy water (D2O) is introduced into a non-thermal plasma jet (NTPJ) device to generate deuterium monoxide (OD) radicals instead of hydroxyl (OH) radicals. An NTPJ generated from a vapor mixture of N2/H2O and N2/D2O is applied to a cell membrane component and its effects are analyzed by means of 1H NMR, GC-FID and TOF-SIMS spectroscopies. The results show that OH and OD radical species induce similar levels of oxidative breakage of lipid molecules. In addition, the 2H NMR spectra show that deuteriums are incorporated into the lipid oxidative products. In order to trace these effects in vivo, E. coli bacteria are treated with an NTPJ and analyzed using NanoSIMS. Deuterium is observed in both the cytoplasm and membrane, which are colocalized well with nitrogen and phosphorus atoms. The high colocalization of D atoms inside E. coli provides the first direct and visual evidence of the role of OD radicals, which may be utilized to visualize OH radical interactions inside cells.


Introduction
Recently, the non-thermal plasma jet (NTPJ) has been widely extended to applications in biological and medical devices [1][2][3]. Among various radical species, such as OH, O 3 , H 2 O 2 , NO, and ONOO, the amount of OH radical and H 2 O 2 generated from a NTPJ are an important parameter when attempting to induce biological responses [4][5][6][7][8][9][10][11][12][13]. A group in Germany explored that H 2 O 2 had the strongest influence on plasma-treated media [6], and a Japanese group introduced OH radicals as an important factor in their plasma treatment method [7]. Simulations showed the oxidation of DNAs and lipids in a cell by OH radicals and the detailed information on their exact interaction mechanisms [8][9][10]. However, few experimental clues have been introduced due to the extremely short life time of OH radicals. Therefore, we investigate the possibility of the use of an isotope as a new method for the detection of OH radical interactions with biomolecules. OD radicals instead of OH radicals are generated through a plasma jet and their incorporation in a lipid molecule is detected using NMR, GC-FID and TOF-SIMS spectroscopes. The high intensity of a deuterium atom in a specific position of a lipid molecule is detected by 2 H NMR spectra. In addition, E. coli are treated with a deuterated NTPJ and the location of deuterium is determined by means of NanoSIMS. The NanoSIMS image shows that deuterium is incorporated in both the cytoplasm and membrane, which are colocalized well with nitrogen and phosphorus atoms in E.coli. This is the first direct and visual evidence of OH radical effects on cell components comparing to other previous reports using NanoSIMS [13][14][15]. The new method involving the use of OD radicals instead of OH radicals through a plasma jet can be used to elucidate the inter-cellular mechanisms of OH radicals.

Experimental
2.1. NTPJ with N 2 and N 2 /H 2 O and N 2 /D 2 O gases Figure 1 shows the schematics of the NTPJ device with porous alumina. Plasma is generated in the porous alumina between the electrodes, which are separated by a gap of 1 mm [16]. The percentage of the distilled water or heavy water vapor in the gas mixtures is controlled by a bubbling system with a mass flow controller (GMC 1200, Atovec). The flowing volume of gas in the NTPJ device is maintained at 1 SLM during all of the experiments. High voltage power is supplied to the NTPJ by a commercial transformer, resulting in an output frequency of 10 kHz and output voltage of 3 kV. The electrical power is measured with an oscilloscope (TDS2002C, Tektronix) and an AC current probe (P6021, Tektronix), and the optical emission spectra of the discharge in the region of 200-1000 nm is recorded by a spectrometer (HR4000, Ocean optics) with a resolution of 0.7 nm.

Measurement of the pH, ion concentration and radical concentrations
An amount of 500 μl of de-ionized (DI) water in a 24-well plate is placed 6 mm below the outer electrode and treated by a NTPJ with gases of N 2 /H 2 O (1.16%) or N 2 /D 2 O (0.75%) for 1, 3, and 5 min. Immediately after the plasma treatment, the pH and ion concentration of each liquid are measured using a pH meter (Eutech Instruments, Singapore) and an ion chromatograph (ICS-3000, Dionex, USA), respectively. The liquid is filtered with a 0.2 μm syringe filter before the ion chromatography measurement. OH radical generation in the liquid is evaluated using terephthalic acid (TA, Sigma-Aldrich), which reacts with OH radicals to generate 2hydroxyterephthalic acid (HTA, Sigma-Aldrich) with stable fluorescence. 10 mM of a TA solution is made in a 7.5 mM NaOH aqueous solution. The concentrations of hydrogen peroxide and nitric oxide are determined with a Quantichrom peroxide assay kit (DIOX-250, BioAssay Systems) and a Quantichrom nitric oxide assay kit (D2NO-100, BioAssay Systems), respectively. Optical emission spectra of NTPJ with feeding gases of N 2 , and N 2 with 1.16% H 2 O, and N 2 with 0.75% D 2 O in the ranges of (c) 200-500 nm and (d) 305-315 nm. OH and OD radicals are detected with the addition of water vapor, and their slightly shifted emission spectra imply their slight energy difference.

Plasma treatment and lipid sample preparation
Oleic acid (O1008, Sigma) is dissolved in CHCl 3 to a final concentration of 25 mg/ml. 5 μl of the solution is loaded on a 10×10 mm 2 silicon wafer which is cleaned in a piranha solution (H 2 SO 4 : H 2 O 2 = 4:1 (v:v)) and dried at 80°C overnight. The oleic acid thin film is dried under N 2 gas flow to remove the solvent fully. The oleic acid thin film is located 6 mm away from the outer electrode and treated with a NTPJ for 5 min with N 2 /H 2 O (1.16%) or N 2 /D 2 O (0.75%). For the TOF-SIMS analysis, the oleic acid thin film is directly analyzed, and for the liquid NMR and GC-FID analyses, the plasma-treated oleic acid is recovered in a CHCl 3 and CH 3 OH mixture (4:1 (v:v)).

NMR analysis
For the 1 H NMR measurement, the plasma-treated liquid samples are dried and recovered with CDCl 3 , and the 1 H-NMR spectra are recorded by a Bruker AVANCE 600 MHz (14.1 T) spectrometer using a 5 mm TXI cryoprobe. For the 2 H-NMR analysis, the samples are dried and recovered with CHCl 3 and the spectra are recorded at 92.1 MHz. The NMR spectra are assigned with reference to the NMR lipid database.

GC-FID analysis
For the GC-FID analysis, the plasma-treated liquid samples are dried and recovered with a methylation mixture consisting of methanol, benzene, 2,2-dimethoxypropane and H 2 SO 4 at volume ratios of 39, 20, 5, and 2. Heptanes are added to the samples and the samples are digested at 80°C for 2hr. Transmethylation of the lipids takes place in a single phase. After cooling at room temperature two phases are formed; the upper one contains the fatty acid methyl esters (FAMES) ready for the GC-FID analysis [17]. The length of the fatty acids is analyzed by gas chromatography (Agilent 7890A, Agilent).

TOF-SIMS analysis
TOF-SIMS experiments are performed with a Bi + primary ion beam with an intensity of 1 pA and a density of 1.50×1013 ions cm −2 (TOFSIMS 5, ION-TOF GmbH, Germany). The pressure in the chamber is less than 10-9 mbar. For an oleic acid analysis, negative ions from areas in size 50×50 μm 2 are acquired in the high current bunched mode for 60 s. The spectra were acquired three times at different locations of each sample. The spectra of the mass-to-charge ratios (m/z (amu)) are analyzed with SurfaceLab 6.1 software.

Preparation of bacteria for NanoSIMS
Escherichia coli (ATCC 11775) are cultured in Luria-Bertani (LB) culture media (244610, BD Difco) until they reach the logarithmic growth phase. The suspension of E. coli is diluted into 5×104 cells ml −1 and 0.5 ml of this solution is treated with the plasma jet in a 24-well plate (30024, SPL). The distance between the outer electrode and the solution is maintained 6 mm during the exposure. The plasma-treated E.coli is transferred onto LB agar and covered with a 1% low-melting agar solution (15510-019, Invitrogen) to avoid any loss of bacteria during the washing steps. A bacteria-containing agar block is fixed with Karnovsky's fixative for 6 h, post-fixed with 1% osmium tetroxide for 1 h, dehydrated with an ethanol series, and then embedded in epoxy resin (18010, Eponate 12 kit, Ted Pella, Inc.). Once the resin is polymerized, the bacteria-containing region is cut by ultramicrotomy to a thickness of 100 nm (MTX Ultracut, UCT). The thin sections are placed on a formvar-coated Cu grid for TEM observation or placed on alumina holder for the NanoSIMS analysis.

NanoSIMS analysis
NanoSIMS imaging is performed using the Cameca Nano-SIMS 50 instrument at the Korea Basic Science Institute (KBSI). Primary Cs + ions are used as the primary ion source with the current of 0.4 pA and an impact energy of 16 keV. The ion beam is focused to a ∼100 nm diameter spot size with a raster size of 10×10 μm 2 over the sample areas. The images were constructed with five replicate scans at a resolution of 512×512 pixels with a dwell time of 1 ms/pixel. The secondary total ions, 14 N − , 31 P − , and 2 H − are collected in the combined analysis mode at a mass resolution of ca. 6000 (M/ΔM), which is enough to resolve the potential molecular interference. Prior to the data collection step, each sample is exposed to pre-sputtering with a high Cs + current in order to remove surface contamination and to ensure equilibrium during the sputtering process. Data are analyzed using the L'IMAGE® software developed by KBSI. The NanoSIMS images are analyzed to determine the colocalization of deuterium with other atoms such as nitrogen or phosphorus in E.coli. The value obtained from the equation is used for the colocalization of two ions [18].

Physical properties of NTPJ with N 2 and N 2 /H 2 O and N 2 /D 2 O gases
The addition of water vapor into N 2 gas can change the discharge voltages as well as other physical and chemical parameters of plasma. At first, the length of the plasma plume is measured while increasing the ratio of the water vapor. Figure 1( Secondly, we analyze the radical species generated from the NTPJ with the addition of water vapor. Figure 1(c) shows the optical emission spectra of NTPJs with N 2 , N 2 /H 2 O (1.16%) and N 2 /D 2 O (0.75%) gases in the region of 200-500 nm. As shown in the figure, the emission spectrum of the pure nitrogen plasma contains NO radicals and a N 2 second-positive system (SPS) only. With the addition of water vapor, the OH related spectra become more dominant in both the H 2 O and the D 2 O cases. The expanded spectra in the range of 305-315 nm show the slight difference between the OH and the OD radicals in figure 1(d). The discharge with water vapor produces significantly enhanced UV radiation stemming from the transitions of the OH band at 307.1 nm and 309.6 nm and of the OD band at 306.9 nm and 308.9 nm. This enhancement was expected from the mathematical calculation of OH radical concentration with mixtures of N 2 gas and water vapor. The metastable state of N 2 (A 3 u S + ) can generate OH radicals easily according to N 2 (A 3 u S + )+H 2 O→OH+H+N 2 with a high dissociation coefficient of α OH =5×10 −14 cm 3 s −1 [19]. Although the lengths of the plasma plume are short with this vapor ratio, the peaks of OH/OD are the highest, as shown in a previous study [14]. Therefore, gases with a ratio of 1.16% H 2 O in N 2 and 0.75% D 2 O in N 2 are used to increase the OH and OD radical effects in all of the following experiments.

Chemical effects of NTPJ on DI water with N 2 and N 2 /H 2 O and N 2 /D 2 O gases
The effects of additions of 1.16% H 2 O and 0.75% D 2 O to N 2 gas are compared by examining the chemical properties of de-ionized (DI) water after NTPJ treatment. Figure 2(a) shows the pH of DI water before and after the plasma treatment. Every measurement was taken immediately after the plasma treatment. The pH of the water changes from 6 to 10 within 1 min of the plasma treatment at which point the value becomes saturated. The high pH of the water appears to be a consequence of the formation of an ammonium base based on the result of an ion analysis of the plasma-treated water, indicating an enormous increase of the NH 4 + concentration ( figure 2(b)). Approximately 50 ppm (∼3 mM) ammonium ions are generated in the water after 5 min of plasma treatment. In detail, the actual pH value decreases slightly after 1 min; this may be related to the increase in the number of nitrate ions from nitrite ions with time.
In addition to the changes in the pH and ionic concentrations, the radical species generated inside the water after the plasma treatment are compared. It is found that the concentration of the nitrates, hydroxyl radicals and hydrogen peroxides in the water increases as the exposure time increases (figures 2(d)-(f)). After 5 min of NTPJ treatment, approximately 150 μM NO x species, 100 μM hydroxyl radicals, and 250 μM hydrogen peroxide are measured inside the water upon vapor additions of either case of both N 2 /H 2 O and N 2 /D 2 O. The 100 μM concentration can be calculated as 2.01×10 14 OH molecules/mL/s. The neutral density in the plasma jet above the water surface is 2.6×10 19 /cm 3 , and the approximate density of OH radicals is about a few times of 10 15 /cm 3 [20]. Considering these values, more than 10% OH radical species may be bombarded into the water. In addition, the concentration of H 2 O 2 before 5 min was found to be higher in the N 2 /D 2 O than in the N 2 /H 2 O plasma treatments. We assume that the reactions of OD and OH (H 2 O 2 and D 2 O 2 ) are nearly identical, as reported by Mededovic et al [21]; however, the exact isotopic effect on the yields of the radicals inside the water should be studied in detail in more specific ways in the future. In the following experiments, we exposed biological samples to N 2 /H 2 O and N 2 /D 2 O plasma for 5 min while considering their similar chemical effects. To understand in more detail the structural changes of plasma-treated oleic acid, GC-FID and TOF-SIMS analyses are performed. The GC-FID analysis provides the length of the fatty acid chains remaining after the plasma treatment. The peak area percentages of the compositions are shown in figure 4(a). The standard hydrocarbon chains from C4 are utilized to find the peak position. The peaks between each standard fatty acid peak are grouped such that they are close to the standard peak in an effort to determine the carbon chain length distribution of the remaining molecules. As expected from the 1 H NMR data, the double bond of most oleic acids (C18:1) is broken to become stearic acids (C18:0) and subsequently the chains are broken and become shorter. Heptanoic acid (C7:0) is formed predominantly and hydrocarbon chains from C8 to C16 are generated similarly. This result implies the highly reactive potential of plasma, which can react with hydrogen at any position along the carbon chain as well as with hydrogen near a double bond. One end of the chain would consist of an aldehyde group which appears at 9.69 ppm in 1 H NMR spectrum. The reduced solubility of oleic acid in chloroform after the plasma treatment supports the chemical changes of oleic acid from carboxyl acid to aldehyde (data not shown). Figure 4 ) nearly disappeared, the peaks between 100 and 200 amu show low intensity levels, and the peaks in the lower mass range show high intensity levels (figures 4(d) and (e)). The data explain the severe breakage of the lipid chains, which can be understood as the local measurement of the TOF-SIMS analysis. The peaks are broadly distributed and generally shifted by about 1∼2 amu from the main peaks of the control sample. The N 2 /H 2 O plasma-treated samples have higher intensity peaks at m/z=57, 59, and 89 than the control, and the N 2 /D 2 O treated samples have higher intensity peaks at m/z=60, 75, 149, and 223 as compared to the other two cases. These mass shifts should be ascribed to isotopic D atom substitution in the oleic acid molecule. These three measurements taken together show that

Trace of isotopic reactive species produced from non-thermal plasma jet in oleic acid
The isotopic D atom coming from the NTPJ can be detected in high-precision mass spectroscopy and by means of a 2 H NMR spectra analysis. Figure 5(a) shows that the TOF-SIMS spectra ranged from 2.005 to 2.022 m/z and from 17.988 to 18.023 m/z, indicating a slight mass difference between D − and H 2 or H 2 O − and OD − molecules, respectively. Greater incorporation of OD ions by nearly tenfold is found in the N 2 /D 2 O treated samples as compared to the others. A much greater amount of D − atoms is found in the N 2 /D 2 O treated samples. Although these peaks show only fractional information, we can confirm that the deuterium-containing molecules are not only in the form of D 2 O from which they originate. The 2 H NMR spectra can give more information about these molecules. Figure 5(b) shows the overlap of the 1 H NMR and 2 H NMR spectra, as displayed by the solid line and the dotted line, respectively. Except for the solvent peak, only one 2 H NMR peak at 2.14 ppm is observed in the N 2 /D 2 O-treated sample. This peak can be attributed to methyl ketone, RCOCH 3 , where the protons are displaced with deuterium. Despite the fact that it does not precisely match the 1 H NMR peak at 2.08 ppm, this type of slight shift may be attributed to the physical differences between a proton and a deuterium. Our NTPJ treatment may induce the creation of methyl ketones mainly in the oleic acid molecule among the general products of lipid oxidation including aldehydes, acids and ketones [20]. Thus, these results assure us that the tracing of the 2 H atom or deuterated molecules makes it possible to trace radicals from an NTPJ directly in target molecules.

Colocalization of an isotopic D atom with cellular components in E. coli
As the next tracing target, we choose a live bacterium, Escherichia coli (E. coli). After the N 2 /D 2 O plasma treatment, we visualize the molecular images of E. coli using NanoSIMS. Figure 6 shows the NanoSIMS images of 12 C 14 N − , 31 P − , and 12 CD − respectively from bacteria exposed to N 2 /D 2 O plasma inside water. The 12 C 14 N − ions provide the bacterial features, as they mainly come from the proteins. The 31 P − ions are mainly located at the cell membrane whose dominant molecules are phospholipids. Occasionally there are very strong 31 P − ion signals inside the bacteria, which may come from the DNA whose backbone is full of phosphorus. 12 CDions are found in high amounts inside the bacteria, confirming that the D atoms from the plasma are incorporated into the cellular molecules. Here, the control is bacteria exposed to a N 2 /D 2 O gas flow without discharge for 10 min The other two D − molecular images from bacteria treated for 5 and 10 min with plasma show a timedependent increase in the ratio of D atoms to H atoms (CD/CH) inside the bacteria.
For a clear comparison of the images of D − and the other two molecules, the colocalization between 12 C 14 N and 12 CD and 31 P and CD molecules is visualized. Images are normalized and multiplied by each other as expressed in the experimental section. In general, there is a strong correlation between C 14 N and CD, indicating much interaction between the radicals and the intracellular proteins. The correlation between 31 P and CD is not as strong as that of CN and CD. They are overlapped generally in the cell membrane; however, there is occasional strong colocalization between the 31 P and D atoms in the bacteria treated with plasma for 10 min (arrow). Their positions in the cytoplasm imply that they are from the nucleic acids. This is an important result supporting the direct interaction of radicals from an NTPJ with DNA inside cells with a long time exposure.

Conclusion
In summary, we directly visualize evidence of OH radical interaction with biological molecules by utilizing OD radicals instead of OH radicals. We confirm that OD radicals generated by means of a N 2 /D 2 O discharge have physical or chemical properties similar to those of OH radicals generated via a N 2 /H 2 O discharge. OH and OD radicals induce similar chemical changes in DI water and similar oxidative breakage in lipid molecules. Due to its isotopic property, the direct interaction between OD radicals and lipid molecules is visually apparent in their chemical products. Most interestingly, D atoms are found inside bacteria in different positions depending on the plasma exposure times. Proteins appear to be vulnerable to OD radicals; however, a longer exposure time induces very strong interaction between OD radicals and DNA in the cytoplasm. This technique is shown to be very promising in an analysis of the interaction mechanism between plasma-generated OD (or OH) radicals and cell components.