Production of Similar Fragments from the Glycine, Alanine, and Methionine Amino Acid Molecules Under Low-Energy Electron Impact L.G

Vol. 128 (2015) ACTA PHYSICA POLONICA A No. 1

Production of Similar Fragments from the Glycine, Alanine, and Methionine Amino Acid Molecules under Low-Energy Electron Impact L.G. Romanovaa, J. Tamulieneb, V.S. Vuksticha, T.A. Snegurskayac, A.V. Pappa and A.V. Snegurskya,* aInstitute of Electron Physics, Ukrainian National Academy of Sciences, 21 Universitetska str., 88017 Uzhgorod, Ukraine bVilnius University, Institute of Theoretical Physics and Astronomy, 12 A. Go²tauto str., 01108, Vilnius, Lithuania cUzhgorod National University, Department of Physics, 46 Pidhirna str., 88000 Uzhgorod, Ukraine (Received March 30, 2015; in nal form April 24, 2015) Production of the ionic fragments of the same chemical composition due to a low-energy electron impact on the glycine, alanine, and methionine amino acid molecules has been studied both experimentally and theoretically. The mass-spectrometric technique was applied to detect the above ionic fragments within the 1720 a.m.u. mass range with the ±0.25 a.m.u. resolution, while the density functional theory based theoretical approach allowed the relevant species to be identied and the mechanisms of their production to be claried. A special attention has been paid to analysis of the geometrical structures of the molecules under study as well to nding the appearance potentials of their fragments with the accuracy of ±0.1 eV. DOI: 10.12693/APhysPolA.128.15 PACS: 34.50.Lf, 34.90.+q

1. Introduction electrons are of a primary interest from the viewpoint of tracing the consequences of malignant transformations in Amino acids being biologically relevant organic sub- living cells under the inuence of ionizing radiation and stances involved in the live organisms include gener- also provide useful radiation therapy eect on tumors in ally the amine (NH2) and the carboxylic acid (COOH) human beings [5]. functional groups. Their generic formula looks like The studies of the basic mechanisms of the amino acid H2NCHRCOOH (R being an organic substituent, the so- molecule structural changes caused by low-energy elec- called side-chain [1]). Amino acids also serve as build- tron impact are far from being complete despite their ing blocks of proteins and intermediates in metabolism. undoubted signicance. According to the National Insti- The chemical properties of these molecules determine the tute of Standards (NIST) database [6], the available rele- biological activity of the proteins. In addition, proteins vant data are a bit disputable, while those on the parent contain the necessary information needed to determine molecule ionization thresholds and fragment appearance the structure and the stability of the live organism. This energies are extremely scarce. is an important eld of scientic research, and today it is still one of the most important tasks of modern biological and chemical science [2]. Regarding the possible degradation of amino acids under the external impact, it is a well-known fact that ionizing radiation causes in live tissue the irreversible eects at the genetic level [3]. The mechanism of such degradation does not include the direct action of high-energy Fig. 1. Structural formulae of the molecules under study. ionizing radiation only, because the inuence of the processes related to the low-energy secondary electrons appears to be signicant [4]. Secondary products of ion- In our previous papers (see, e.g. [79]), we have stud- izing radiation in living tissues are capable of damaging ied both experimentally and theoretically fragmenta- the structural units of nucleic acids and proteins leading, tion of the glycine, alanine, and methionine molecules in particular, to dissociative ionization with both posi- by low-energy (below 150 eV) electrons and noticed tive and negative-ion production. Thus, the low-energy yield of some fragments of the above three molecules having the same chemical composition not comparing their production mechanisms. Therefore, the aim of the present paper was to study the mechanisms of their *corresponding author; e-mail: [email protected] production and compare these mechanisms with each other. The amino acid glycine, alanine, and methionine

(15) 16 L.G. Romanova et al.

molecules (see schematic view in Fig. 1) have dier- electron gun provided an electron current Ie = 30−50 µA ent substituents R: H, CH3 and C2H4SCH3. One within a wide (0150 eV) energy range with the energy of them (methionine) involves not only the main con- spread < 0.3 eV (full width at the half-maximum of the stituents such as carbon and hydrogen, but also the sul- beam energy distribution) [10]. The experimental ap- fur atom. The above molecules present good example pearance potentials were determined by means of a t- for predicting the inuence of substituents on the pro- ting technique based on the least-squares method approx- cess of fragmentation of the core part (H2NCHCOOH) imation using the MarquardtLevenberg algorithm de- of the amino acid. To our knowledge, no similar data are scribed in detail in our previous studies (see. e.g. [710]). present in literature. The accuracy of the appearance potential determination was not worse than ±0.1 eV. The energy dependences 2. Experimental of ionization and dissociative ionization cross-sections to be measured in the incident electron energy range from In our study, we used a conventional magnetic mass- the threshold up to 150 eV. The ions produced in the spectrometer MI-1201 (see, e.g. [10]) as the mass- ion source and extracted by the electric eld entered a separator unit enabling the fragments of electron magnetic ion separator and were detected by means of molecule interaction to be identied. Figure 2 shows the an electrometer. schematic diagram of the experimental setup. The ex- The data acquisition and processing system was con- perimental apparatus developed and applied is based, as trolled by a PC. Special measures were taken to stabi- already mentioned above, on the magnetic mass spectro- lize the mass analyzer transmission, thus, making the meter capable of operating within the 1720 a.m.u. mass of the fragment under study to be reliably xed. mass range. High sensitivity (≈10−16 A) and resolu- An electron energy scale was calibrated with respect to tion (±0.25 a.m.u.) of the mass analyzer enabled the known ionization thresholds for argon atom and nitrogen fragments of the target molecule to be reliably separated molecule (see below) with the accuracy not worse than and detected even at low levels of ion currents reaching ±0.1 eV. The molecular beam M (see Fig. 2) was directed the ion detector (see Fig. 2). to the ionization chamber normally to the electron beam (the beams intersect in the gure plane). The experimental procedure was as follows. First both electron and molecular beam sources were put into operation and after reaching the optimal conditions of beam generation a mass spectrum of the molecules under study was measured. The masses of the fragments produced were then xed and the energy dependences of the relevant ion yields were measured. Elec- tron energy was varied with an energy step of 0.10.3 eV enabling the threshold areas of the dissociative ionization functions for all the fragments under study to be measured. In this case the problem of the incident electron energy scale calibration becomes very signicant. To calibrate the above energy scale we have measured the threshold areas of the ionization functions for the two Fig. 2. Schematic diagram of the experimental setup test gases Ar and N2. The experimental ionization and supporting electrical circuits. thresholds were determined by means of a tting technique is based on a least-squares method approximation The primary molecular beam ( in Fig. 2) was formed M using the MarquardtLevenberg algorithm suggested by by means of an eusion source with a resistive oven Maerk's group (see, e.g., our earlier paper [11]). providing target molecule concentrations not less than 1010 molecule/cm3. The operating temperature of the 3. Theoretical molecular beam source was varied up to 200 ◦C being controlled by a thermocouple allowing the temperature The structures of the molecules and their fragments dependences of the fragment yields to be determined. were studied using the generalized gradient approxima- As it has been shown in our previous papers [79], the tion for the exchange-correlation potential in the den- linear behaviour of the above temperature dependences sity functional theory (DFT) as it is described by the (plotted in the semi-logarithmic scale) within the temper- Becke three-parameter hybrid functional, applying the ature range of 50120 ◦C indicated unambiguously that non-local correlation provided by Lee, Yang, and Parr. production of the above fragments due to the thermal The DFT method is commonly referred to as B3LYP [12], degradation of the parent molecule was negligible. i.e. as a representative standard one described in more The temperature and pressure conditions of the molec- detail below. The cc-pVTZ basis set was used as well [13]. ular beam formation excluded the possibility of molecular The structures of the molecule isomers/conformers and cluster formation. A specially designed three-electrode their fragments under study were optimized globally Production of Similar Fragments from the Glycine, Alanine. . . 17 without any symmetry constraint. The bond order and the bond length of the molecule conformers were investigated to nd the weaker bonds possible to be destructed. Additionally, the vibration spectrum was evaluated to predict the possible elongation of bonds and the change of the angle aiming to analyze the most probable fragments produced due to electron impact. On the other hand, the results on the vibration modes were analyzed to be sure that the equilibrium point of the molecular systems was found. In order to model the fragmentation processes, the possible fragment anions, cations and fragments with a zero charge were evaluated. Dissociation energies were calculated as the dierence between the total energy of the molecule and the sum of the energies of the fragments predicted. We assumed that in our experimental conditions the structure of the fragments formed could be changed inuencing, thus, the dissociation energy. To evaluate the above inuence, the dissociation energy was calculated for the following two cases: (i) the single point energy calculation of the fragments was performed taking into account the geometry of a certain part of the molecules under study (in these cases the energy of fragment formation is not the lowest one); (ii) the structure of the fragments was optimized, i.e. the fragments were allowed to reach their equilibrium geometry and the obtained energy (the lowest energy of the fragments) was used to calculate the dissociation energy.

4. Results and discussion The amino acid molecules studied by us, despite their dierent composition, produce a series of neutral and ionic fragments with the same mass. Therefore, comparison of possible mechanisms of their production is of certain interest when studying the processes of the initial molecule decomposition. Obviously, it is intriguing to trace their production in this process as well as to study the possible mechanisms of their formation. In this sec- tion, we will summarize some results of our studies. Figure 3 shows the mass spectra of the three molecules under study measured using the technique described Fig. 3. Initial areas of the mass spectra of the above. We would like to draw attention that the glycine, molecules under study: (a) glycine molecule, inset alanine, and methionine molecules could be presented as that in the vicinity of the parent molecule peak, (b) α- alanine molecule, (c) methionine molecule. the COOHCHRNH2 compounds, where R is H, CH3 and C2H4SCH3 in case of glycine, alanine, and methionine, respectively. This allows one to expect that several fragments with equal mass could be formed un- 4.1. Detachment of amino acid molecules der the low-energy electron impact. Formally, such frag- functional groups ments could be NH2 (m = 16 a.m.u.), CH2N, CO (m = 4.1.1. Deamination 28 a.m.u.), CH3N, COH (m = 29 a.m.u.), CH4N, CH2O The mass spectra of the above amino acids reveal a ( a.m.u.), C H N( a.m.u.), CO ,C H N small-scale peak of the NH+ ( a.m.u.) fragment, m = 30 2 4 m = 42 2 2 6 2 m = 16 (m = 44 a.m.u.), CHO2 (m = 45 a.m.u.) and C2HO2, because the peaks observed are due to fragmentation C2H3NO (m = 57 a.m.u.) and they may be charged of the molecule by a preferred cleavage of those bonds, both positively and negatively or be neutral. It should which lead to energetically more favored, i.e. best stabi- be emphasized that we have no possibility to measure the lized, ions. Indeed, NH+ stability is very low as compared 2 mass spectra of negatively charged ions as well as the ap- to that of the neutral and negatively charged ones (see pearance energies for such fragments, thus, some of the Table I). results presented below are based on the theoretical con- The positively charged amino group detachment from siderations only. the parent molecule is the rare event, despite the rela- 18 L.G. Romanova et al.

TABLE I of amino acids aects considerably the dissociative pro- Binding energy per atom (eV) for the dierently cess (1). While the hydrogen atom (R in the glycine charged fragments produced from the above molecules. molecule) substitution by the methyl group does not inuence the character of the main fragmentation channel, Fragment charge Fragment the presence of the sulfur atom complicates signicantly 1 0 1 this process. NH 3.43 3.38 0.83 2 In case of glycine after the CC bond rupture two COOH 4.13 4.45 1.31 α complementary fragments CHO (m = 45 a.m.u.) and CH N 3.94 4.12 3.06 2 4 CH4N(m = 30 a.m.u.) appear. The calculated bind- CH N 3.94 4.24 2.63 + 3 ing energy per atom shows that CH4N is three times CH2N 4.10 4.39 2.27 more stable than COOH+ (see Table I). Hence, forma- CHO 4.02 3.63 0.93 + − tion of CH4N (m = 30 a.m.u.) and COOH (m = CH6N 4.20 4.32 3.71 45 a.m.u.) is more probable because of the energetically C4H10NS 4.24 4.28 3.94 more favored ion formation, and this leads to the small C2H4NO2 4.71 4.63 3.82 (nearly 3% relative intensity) peak at m = 45 a.m.u. in C2HO2 4.78 4.34 2.14 the glycine mass spectrum. C2H3NO 4.53 4.73 3.60 In the mass-spectra of the alanine and methionine molecules, the fragment with the m = 45 a.m.u. mass assigned to COOH could be observed as well with the in- tively low values of calculated energies required to break tensities ≈3% and ≈14%, respectively. In case of the ala- the CN bond but our theoretical investigations pre- nine molecule, the theoretical results prove that reaction dict that, under a low-energy electron impact, all the C H NO + e → C H N+ + CHO− + e (2) molecules investigated could produce the a.m.u. 3 7 2 2 6 2 m = 16 could be more probable due to the stability of the frag- mass fragment assigned as NH−. In addition, the amine 2 ments produced, however, the comparison of the calcu- fragment could be obtained as the other product when lated and measured appearance energies indicates that it has negative or zero charge. In the case of glycine the reaction NH−/0/+ fragment could be formed with very low possi- 2 + (3) bility in the appearance energy range 12.8525.58 eV [7] C3H7NO2 + e → C2H6N + CHO2 + 2e is more energetically probable, i.e. C H N+ and CHO0 (the other fragment formed is positively charged). For 2 6 2 formation is more expected, while CHO+ could be formed the alanine molecule this calculated energy is equal to 2 12.0424.77 eV depending on the structural dierences only in the case when the hydrogen bond O···H in the between the isomers and intramolecular hydrogen bond- alanine molecule occur favoring this fragment stability. ing [8], and for the methionine molecule, the NH−/0/+ It is interesting that in case of methionine the frag- 2 fragment can be formed as well at 10.4223.14 eV [9]. ment with the m = 45 a.m.u. mass is produced with higher intensity as compared with other amino acids un- 4.1.2. Carboxyl group detachment der study. This peak increase may be due to the iso- According to the early (that has become classical) + + baric fragments, such as CHS and C2H7N that can work [14], fragmentation of amino acid molecules is re- contribute to the peak intensity. We assign the peak at lated to the removal of one electron from the nitrogen m = 45 a.m.u. to the COOH+ fragment because in order lone pair resulting in the charge localization on the ni- to produce the above fragment the rupture of only one trogen atom and on the adjacent α-carbon atom. Such CC bond is required, while formation of other isobaric amine type ionization seems to dominate over the other + + fragments, such as CHS and C2H7N , require a break possible ionization channels with the nitrogen atom pro- of several bonds and, in some cases, the rearrangement of ducing the iminium ion structure typical in the dissocia- the hydrogen atoms. In addition, our experimental ap- tive ionization of amines pearance energy values obtained for the fragments with a.m.u. ( eV) and a.m.u. . (1) m = 45 13.5 ± 0.1 m = 104 (9.7 ± 0.1 eV) are close to the calculated ones. According Here R is an amino acid side chain. The one-electron to our data, the intensity of the peak at m = 45 a.m.u. is shift ( ) from both the ionized nitrogen atom and the somewhat lower than that of the complementary peak at

α-carbon atom results in the double bond formation ac- m = 104 a.m.u., thus, after the CCα bond rupture the companied by the energy release. cation center is mainly displaced to the C4H10NS fragment. The calculated appearance energies for CHO and This CCα bond rupture and the carboxyl group de- 2 tachment is a dominant channel of the glycine and ala- for the complementary fragments of the glycine, alanine, nine molecule fragmentation. But in case of methionine and methionine molecules are presented in Table II. this process is much less ecient as compared to the dis- The last row of this table shows the energy of the CCα sociation channel occurring via the CβCγ bond rupture bond rupture calculated not taking into account the that is, evidently, due to the inuence of the sulfur atom ionization energies and electron anities of the comple- that changes the HOMO character. So the side chain mentary fragments produced. Certainly, the CCα bond Production of Similar Fragments from the Glycine, Alanine. . . 19

TABLE II + fact the CH4N ion peak intensity in its mass-spectrum is very small and lower than those of the CH N+ and Calculated appearance energies (in eV) for the CHO2 and 3 + complementary fragments for glycine, alanine, and me- CH2N ion peaks, production of which requires less num- thionine molecules. ber of regrouping. However, in case of the methionine molecule, for- CHO Appearance energy 2 Complementary mation of the CH N+ ion can proceed due to the ( a.m.u.) 4 m = 45 fragment charge rearrangement of one hydrogen atom upon fragmenta- fragment charge Glycine Alanine Methionine tion from dierent parts of the molecule. The fragment 1 1 8.57 7.58 7.35 with a.m.u. could be produced dierently via 0 1 9.99 9.12 8.7 m = 30 the following pathways: 1 -1 16.11 13.16 12.34 1 0 14.99 12.08 11.76 C5H11NO2S + e → 0 0 3.51 3.14  + −  (CH3N + H) + (COOH + C3H6S) + e, (5)  + −  (CH3N + H) + (COO + C3H7S) + e, (6) + 0 dissociation energy and the appearance energy for the  (CH3N + H) + (COO + C3H7S) + 2e, (7)  + 0 above ions depend on the side-chain type in the amino  (CH3N + H) + (COOH + C3H6S) + 2e. (8) acid molecule and decrease with its increase. Summarizing the results discussed above related to the In the above cases the (CH3N + H) compound be- functional group detachment, one may predict that the comes the most stable when the H atom is joined with NH−/0/+ and COOH−/0/+ fragments are formed when CH N+, i.e. when the lowest-energy geometry is reached 2 3 + the glycine, alanine, or methionine molecules undergo the and the CH4N fragment is formed. The hydrogen atom + low-energy electron impact. The comparison of the cal- attachment to CH3N is followed by release of the 6.12 culated appearance energies for these fragments of three or 7.85 eV energy, when H originates from the C3H7S amino acids shows that the decrease of the complemen- or COOH fragment, respectively. Thus, the hydrogen tary fragment mass reduces the calculated appearance atom from the hydroxyl group preferably takes part in + energy for these fragments. the CH4N fragment formation. In this case, the lowest energy is required to divide the methionine molecule into 4.2. Production of the ionic fragments the (COO + C H S)− anion and the CH N+ cation. of the same chemical composition 3 7 4 According to our measurements, the appearance en- 4.2.1. Fragments with the , 29 and 30 a.m.u. + m = 28 ergy for the CH4N fragment production from the me- masses thionine molecule is 11.0 ± 0.1 eV. Taking into account The fragment peak with m = 30 a.m.u. in the glycine, the energy released during the H atom attachment, the alanine, and methionine mass spectra is accompanied by calculated appearance energies for this fragment accord- the satellite peaks with the m = 28 and m = 29 a.m.u. ing to pathways (5) and (6) are 12.13 eV and 9.30 eV, masses. When analyzing the appearance of the fragments respectively. with the masses m = 28 and 29 a.m.u., we would like to Another assignment of the peak with the m = draw a special attention to the formation of the fragment + 30 a.m.u. mass is CH2O . But its appearance from with the m = 30 a.m.u. mass due to the assumption that the amino acid molecules requires essential energy, and, the fragments with the m = 28 and 29 a.m.u. masses may therefore, formation of this ion is less probable than that be related to this one. of the CH N+ ion. The fragment with the a.m.u. mass (i.e. the 4 m = 30 The fragment with the m = 29 a.m.u. mass may be + CH4N ion) is the most prominent peak in the glycine + + assigned to the COH or CH3N ion. According to the mass-spectrum (Fig. 3a). According to our results re- results of theoretical investigation, one may state that in + ported in [15], the appearance energy for the CH4N all cases the COH fragment formation is a complicated fragment of the glycine molecule is 10.1 ± 0.1 eV being process because the probability of the simultaneous CO close to the calculated value and this fragment is pro- and CC bonds rupture is very low. In addition, the duced via the following pathway: stability of the CHO+ fragment is lower than that of the 0 + (4) + C2H5NO2 + e → CHO2 + CH4N + 2e, CH3N ion and this fact explains why the peak with m = i.e. the fragment is formed due to a simple rupture of 29 a.m.u. is not assigned as CHO+. Hence, the fragment with a.m.u. is preferably the CH N+ cation. the CCα bond and the carboxyl group detachment ac- m = 29 3 + cording to pathway (1). For the glycine molecule, the CH3N fragment for- Formation of the fragment with m = 30 a.m.u. at mation occurs with the minimal energy consumption in 0 the dissociative ionization of the alanine and methionine case of the process C2H5NO2 + e → (CHO2 + H) + + molecules requires additional energy consumption for the CH3N + 2e. Our analysis of the charge distribution for carboxyl group detachment and one more skeleton bond the (CHO2 + H) fragment group shows that the minimal rupture with simultaneous hydrogen atom displacement. energy corresponds to the CH2O2 compound production. Of the molecules under study, the H-atom migration Thus, the break of the CC bond accompanied by the process is most complicated in alanine, and due to this H atom migration from the amino group to oxygen of 20 L.G. Romanova et al.

+ the carbonyl group is the most probable channel of the is more probable than that for the CH3N fragment. + CH3N fragment formation for glycine. The m = 28 a.m.u. ion may have the following gross formulae: CH2N, C2H4 and CO. The comparison of the stability of the C H+, CH N+ and CO+ ions allows us 2 4 2 to conclude that C H+ is more stable than both CO+ 2 4 + and CH2N . In case of alanine and methionine [17], we found that formation of the CO+ ion is energetically less probable than that of the C H+ and CH N+ ions. 2 4 2

Fig. 5. The CH N+ cation isomers before (top) and Fig. 4. The area of the DL α-alanine mass spectrum 2 in the 26.529.5 a.m.u. mass region. after (bottom) the geometry optimization. + The structure of the CH2N ion depends on the par- In the case of the alanine molecule, we have found an ent or intermediate ion bonds being broken. Four pos- interesting phenomenon not revealed in the mass spectra sible relevant isomers are shown in Fig. 5. It is impor- + of other amino acids under study. A double-headed peaks tant that in cases when dierent bonds of the CH4N or + of the fragments with the m = 27−29 a.m.u. masses are CH3N cation are broken the fragments become planar due to the two constituent parts of each peak (see Fig. 4). after the geometry optimization. Calculations [16] have A double-headed peak with the m = 29 a.m.u. mass is shown that the HCNH structure is the most stable of due to the two components and consists of the CHO+ those presented in Fig. 5. According to our calculations, + and NH2CH ions with some preference being given when the geometry optimization is performed with the to the CHO+ ion. The isobaric fragments with the B3LYP cc-pVTZ approach application, the trans- and m = 29 a.m.u. mass could be formed according to the cis-isomers (III and IV in Fig. 5) transit to the most following pathways: stable structure with the linear conguration (the point  group ). Note that the C and N atoms in this case CHO+ + C H NO + 2e, (9) Cs  2 6 undergo the -hybridization.  + − sp C3H7NO2 + e → CH3N + C2H4O2 + e, (10) In the case of the glycine molecule the fragment with  CH N+ + C H O + 2e. (11) + 3 2 4 2 m = 28 a.m.u. is the positively charged CH2N ion because the experiment for the deuterated d5- and d3- It should be noted that deuteration of the α-alanine molecule does not assist in choosing between these two glycine [16] has unambiguously shown that this peak be- + + possible assignments [16] and our results show that both longs to the CH2N (CD2N ) ion. ions are formed in the collision event. In the alanine molecule mass spectrum, the ion with For the methionine molecule this fragment, as men- the m = 28 a.m.u. mass is the second intensity-related tioned above, could be formed according to pathways peak and may have gross formulae: CH2N, C2H4 and CO. (5)(8) in the case when it is not joined with the H atom. Jochims et al. [16], Ipolyi et al. [18] and Bari et al. [19] According to our calculations, the lowest appearance en- assigned it to the HCNH+ ion exceptionally, but Lago + et al. [20] identied this peak as consisting of the HCNH+ ergy of CH3N is required when the methionine molecule is divided into the ion pair according to the following and CO+ ions. In our spectrum, only two distinct peaks pathway (when the ion geometry is the same as in the arise in the vicinity of the 28 a.m.u. mass (see Fig. 4), core molecule, i.e. the equilibrium point of ion is reached so at least two ions may contribute to this peak. very slowly): According to the qualitative mass-spectrometry theory, where the direction of fragmentation of the molecu- C5H11NO2S + e → lar ion is dened by the stability of the fragments produced, formation of the CH N+ ion in the case of alanine CH N+ + (COOH + C H S)− + e. (12) 2 3 3 7 molecule seems to be most probable according to the fol- It should be noted that the calculated appearance energy + lowing pathways: for the CH3N fragment is at least 3.48 eV higher than + + + • that for the (CH3N + H) fragment. Based on these re- C3H7NO2 → CH2N + (CH3 + CO2 + 2H) → sults, we can predict that the electron-impact fragmenta- + • + CH2N + (CH + CO2 + H2) (13) tion of methionine producing the (CH3N + H) fragment 3 Production of Similar Fragments from the Glycine, Alanine. . . 21 and from sp3 to sp2. However, based on our results of studies, + + • it is not possible to nd formation of what fragments, i.e. C3H7NO → CH2N +(CH3+CO+(OH+H)) 2 CH N+ or C H+, is more probable. (14) 2 2 4 & As for the calculated appearance energies for the H O. CH N+ and C H+ fragments for alanine (see Ta- 2 2 2 4 Thus, the fragments that might be produced are the sta- bles III, IV), it should be noted that they are a bit dif- ble molecules and the radical. But our calculations of the ferent. + appearance energy for the CH2N fragment according to the pathways (13, 14) show that the system does not TABLE IV reach its equilibrium state, i.e. these processes are not Calculated appearance energies (in eV) for the C2H4 completely realized during the alanine molecule electron- and (NH2 + COOH) fragments formed from the alanine + molecule. impact dissociation. So, we have calculated the CH2N ion production from the alanine molecule with dierent C2H4 (m = 28 a.m.u.) NH2 + COOH Appearance bonds being ruptured. If the equilibrium geometry struc- fragment charge fragment charge energy ture of the alanine molecule fragments is taken into ac- 1 13.44 count, the smallest appearance energies were obtained by −1 1 0 11.80 + us when the CH2N ion was formed according to the following general pathways: Thus, in case of the C H+ ion, the appearance energy 2 4 C3H7NO2 + e → is smaller by more than 4 eV if the alanine molecule ge-  + −  CH2N + (COOH + H + CH3) + e, (15) ometry was not changed. It implies that when the frag-  + CH2N + (COOH + H + CH3) + 2e, (16) ment energy is minimal and the level of its excitation is  + + below the energy barrier for dissociation, the appearance  CH2N + (COOH + H + CH3) + 3e. (17) of the C H+ ion is more probable. When the positively 2 4 The calculated energies required to produce the above charged ion and the neutral fragment are formed, the fragments are listed in Table III. appearance energy values become very close, so both the C H+ and CH N+ ions might arise. Thus, we may con- TABLE III 2 4 2 clude that two isobaric ions peaks with the a.m.u. The calculated appearance energies (in eV) for the CH N m = 28 2 mass in the experimental spectra belong to the C H+ and and (COOH + H + CH ) fragments formed from the 2 4 3 CH N+ ions and the channel of the C H+ ion formation alanine molecule. 2 2 4 is more ecient at the 70 eV collision energy than that CNH of the CH N+ ion. 2 COOH + H + CH Appearance 2 (m = 28 a.m.u.) 3 The C H+ fragment of the methionine molecule could fragment charge energy 2 4 fragment charge be formed as follows: 1 1 9.98 C5H11NO2S + e → 1 0 12.41 1 1 19.51 + −/0/+ (19) C2H4 + (C2H4O2N + CH3S) + ne. Here n = 1, 2, 3. + The experimental appearance energies for the CH2N fragment are most adequately described by path- TABLE V ways (15) and (16). Calculated appearance energies (in eV) for the C2H4 Two ways of formation of the CH2N fragment from the and (C2H4O2N + CH3S) fragments of the methionine methionine molecule were investigated because this frag- molecule. ment could be produced when two CC and CH or NH bonds of the initial molecule are ruptured. It implies that C2H4 (m = 28 a.m.u.) (C2H4O2N + CH3S) Appearance the CH2N fragment could have structures just like those fragment charge fragment charge energy presented in Fig. 5. The analysis of the appearance ener- 1 −1 9.52 + gies for the CH2N fragment produced from the neutral 1 0 10.98 and ionized molecule shows that the calculated appear- 1 1 22.2 ance energy 11.83 eV according to the pathway + + C5H11NO2S → CH2N +(CHO2+H+C3H7S) (18) In this case much attention should be paid to the en- coincides with the measured value 11.4 ± 0.1 eV. ergy released due to a rapid change in the geometrical The fragment with m = 28 a.m.u. in the cases of structure of the C2H4 fragment. This energy is equal alanine and methionine molecules can be not the pos- to 2.44 eV. Hence, if the released energy is taken into ac- itively charged CH N+ ion only, but the C H+ ion as count, the C H+ ion is produced, most probably, accord- 2 2 4 2 4 well. The stable structure for this C H+ fragment corre- ing to a pathway with ion pair production (see Table V). 2 4 + + sponds to that of the ethylene molecule, and its formation The positively charged CH3N and CH2N ions could + is accompanied by the carbon atom hybridization change result from the secondary dissociation of the CH4N ion, 22 L.G. Romanova et al.

+ while the CH2N ion could also appear due to deproto- ture for this fragment, not mentioned among the possible + nation of the CH3N ion. The above secondary dissocia- structures calculated for this stoichiometry, presented in tion is more possible in case when the energy transferred Table VI. To indicate the stabilities of the structures un- to the initial molecule increases. Our experimental mass der study the energy dierence is presented as well when spectra of the glycine and methionine molecules reveal a the lowest energy is assumed to be zero. A comparison diuse peak at the mass of about m∗ ≈ 26.1 a.m.u. [7, 9] of the values of the binding energy per atom indicates + that corresponds to the 30 → 28 transition with the that the CH3CNH fragment is the most stable one, but + detachment of a neutral fragment with m = 2 a.m.u., our calculations show that the C2H4N ion formed in the + i.e. the secondary fragmentation of the CH4N ion oc- case of the alanine and methionine molecules dissociation curs. Due to this dehydration, when the stable hydrogen has another structure. + molecule is formed as the dissociation product, the frag- The appearance energies for the C2H4N ions cal- ment with the mass of m = 29 a.m.u. seems less probable culated in our recent paper [9] according to the path- as well. However, this dissociation channel was not ob- way + + − are C3H7NO2 → CH3CNH + (H2 + CHO2) served by us for α-alanine. Thus, the dehydration process closer to the experimental value than the ab initio re- according to the pathway + 0 + is not CH4N → H2 + CH2N sult (10.87 eV) presented in [18], where it was stated + realized in the electron-impact-induced alanine molecule that the CH3CNH ion is formed via the following reac- fragmentation. The probable cause of this phenomenon tion process: + + − . C3H7NO2 → CH3CNH + HCO + H2O + + is the marginal pathway of the CH4N formation in the Note that in both pathways the C2H4N ion appears in electron impact dissociation of the alanine molecule as a case of an ion pair formation. Thus, the mechanism + compared with glycine and methionine. of the C2H4N ion production includes the detachment 4.2.2. Fragments with the m = 42, 44 and 57 a.m.u. of a carboxyl group from the initial molecule accompa- masses nied by the molecular hydrogen formation rather than In the mass range 4044 a.m.u., a perceptible peak at the disintegration process of the COH group and the water molecule elimination as suggested in [18]. m = 42 a.m.u. was observed experimentally in the methionine and alanine mass spectra. This peak may be + attributed to the C2H4N ion. In case of glycine, to pro- + duce the C2H4N fragment (m = 42 a.m.u.) the C = O and the COH bonds must be broken. It is known that the average bond energy for the C = O and the CO bonds is ≈8.28 eV and ≈3.7 eV, respectively, while for CC it is about 3.61 eV [21]. The simplest observation al- + lows one to predict that reaction producing the C2H4N fragment in the case of glycine is energy consuming and, thus, should be deemed not possible or having very low probability to occur. + Fig. 6. Views of the CH3CHN fragment when alanine is ruptured (left) and at the equilibrium point (right). TABLE VI

+ Total (internal) energies of the C2H4N fragment with It is interesting to note that optimization results indi- dierent geometrical structure. + cate changeability of the initial CH3CHN ion geomet- + Total energy Total energy rical structure. The CH3CHN fragment transforms to Fragment structure + [a.u.] dierence [eV] the CH3NCH one at its equilibrium point. We started our optimization from the CH CNH+ structure and - NH CH C 132.88 2.84 3 2 2 nally obtained the CH N=CH+ structure. The geomet- CH CH=N 132.97 0.54 3 3 rical structures of the C H N+ ion both before and after CH C NH 132.99 0 2 4 3 ≡ optimization are shown in Fig. 6. CH =CHNH 132.91 2.15 2 The mechanism of the structural change during forma- CH =CNH 132.97 0.44 2 2 tion of the above ion could take place via the intermediate cyclic structure of the dehydrated ethylene imine accord- For the alanine molecule, the C H N+ ion formation 2 4 ing the pathway given below is energetically more probable, because in this case the weaker CC and two NH bonds could be broken. It im- + (20) plies that the mechanism of the C2H4N ion production . includes the detachment of a carboxyl group from the initial molecule being accompanied by the molecular hy- This intermediate structure can be obtained when ana- drogen formation as well as by the H atoms migration. lyzing the optimization process. The bond order analy- + Jochims et al. [16] and Ipolyi et al. [18] identied this sis indicates that in the C2H4N positive ion structure ·+ + fragment as NH2CH2 = C and CH3C ≡ NH , respec- the double bonds are formed between both the N and tively. On the other hand, we suggest the another struc- C atoms, while the bond order between the CC atoms Production of Similar Fragments from the Glycine, Alanine. . . 23 is 0.846. Thus, this bond is the weakest one in the inter- appearance energy is approximately twice larger than + mediate cyclic structure, and the above bond rupture is that for the isobaric C2H6N fragment. Summarizing possible. In the obtained structure, the C atom under- previously mentioned, it is possible to conclude that the goes, obviously, the sp-hybridization. peak at m = 44 a.m.u. in the experimental mass spec- + tra of the alanine and methionine molecules could be at- In case of the methionine molecule, C2H4N fragment + could be formed via the following pathway: tributed to the formation of the C2H6N fragment. For + the alanine molecule the appearance of the C2H6N frag- C5H11NO2S + e → ment is described above as the result of the CCα bond + − rupture and the carboxyl group detachment. However, C2H4N + (C2H5S + H + CHO2) + e, (21) where the C H N+ ion has the following structure: we failed to nd the channel of this fragment formation 2 4 from the methionine molecule: the calculated appear- CH2=CHNH. Formation of the initial structure was predicted based on studying the bond order. On the other ance energy of this fragment is approximately 5 eV less hand, we have checked the channel of formation of the than the measured one or 4 eV higher, when this peak is assigned as CS+. fragment with the CH2=CNH2 structure that is more stable than CH =CHNH, but less stable than CH We also predicted the formation of the m = 57 a.m.u. 2 3 fragment that can be attributed to the C HO+ or C≡NH. Notably, the most stable CH C≡NH fragment 2 2 3 + cannot be formed in the case of methionine. It should C2H3NO ions, but the relevant peak with small inten- be mentioned that the measured appearance energy for sity (≈12%) is present in the methionine mass-spectrum + only. For the glycine and alanine molecules the inten- the C2H4N fragment is 12.9 eV, while calculated one is 9.37 eV or 6.95 eV. Taking into account that the frag- sity of the peak at m = 57 a.m.u. is small (≈0.20.4%) + ments are not the most stable ones (see Table VI), one despite of the high stability of the C2H3NO fragment may predict the H atom migration that requires the and the small-scale calculated appearance energy for this 2.15 eV or 0.44 eV energy. fragment (9.38 eV and 11.23 eV for glycine and alanine, + respectively). Hence, the C2H4N ion in the case of the alanine and methionine molecules has a dierent structure. It is interesting, in our opinion, to note that in the glycine and methionine mass spectra (but not in the case The peak at m = 44 a.m.u. observed in the experi- of alanine) a weak peak located at a.m.u. is mental mass spectra can be attributed to the CO+ for m = 28.5 2 observed experimentally [7, 9]. Obviously, this peak is as- + all three molecules under study, C2H6N in case of ala- 2+ signed to the doubly charged C2H3NO ion. Formation nine and methionine but in the latter case it may also of this ion requires the water molecule elimination after + be CS . the two-electron loss from the rst two highest occupied In the glycine mass spectrum, the peak at m = molecular orbitals of the parent molecule. The calcula- a.m.u. can be attributed to the CO+ ion only and 44 2 tion of the energy required to produce the doubly charged the relative intensity of this peak is 5.8% [6]. For the 2+ C2H3NO ion from the neutral glycine molecule with CO2 fragment, the relevant complementary fragment cor- multiplicity 1 shows that the hydrogen atom detachment responds to CH3NH2. This pair of fragments is produced from the carbon atom proceeds with a higher probability due to the hydrogen atom migration from the hydroxyl than that from the nitrogen atom. As for the methion- group to the carbon atom via the 4-term transient state. 2+ ine molecule, the C2H3NO ion occurs after the doubly Two alternative decay channels are possible here diering charged parent molecule ion skeleton CC bond disso- by both the reaction rate and the nal charge localization ciation accompanied by the water molecule elimination. Taking into account that the hydroxyl group and the hydrogen atom required for the water molecule formation . (22) are located in the opposite parts of a parent molecule, one may conclude that H2O is eliminated simultaneously or slightly before the skeleton bond dissociation. The intensities of the corresponding peaks in the mass- spectrum may characterize the eciencies of the above 5. Conclusions reaction channels. According to our data, the intensity of the m = 44 a.m.u. peak is higher than that We have studied both experimentally (applying the of the m = 31 a.m.u. peak, thus, in case of the C mass-spectrometric technique with the ±0.25 a.m.u. C bond dissociation accompanied by the hydroxyl group mass resolution) and theoretically (using the DFT-based H atom migration; the cation center is mainly displaced theoretical approach) the mass spectra of the amino acid to the CO2 fragment and the neutral CH3NH2 fragment molecules (glycine, alanine, methionine) and have iden- is eliminated. tied the main components having the same chemical We have checked the possibility of CO+ formation at composition/masses. Their absolute appearance poten- 2 the dissociative ionization of the alanine and methionine tials have been determined. The analysis of the obtained molecules. The calculated results indicate that produc- results using the newly developed DFT approach has al- tion of this ion is energetically non-favorable because its lowed the principal mechanisms of the initial amino acid 24 L.G. Romanova et al. molecules dissociation/fragmentation to be established [3] S. Cristoni, L.R. Bernardi, Mass Spectrosc. Rev. 22, with the allowance made for the charges of the ionic frag- 369 (2003). ments produced. [4] B.D. Michael, P.A. O'Neill, Science 287, 1603 (2000). We have shown that the dierent substituents inuence [5] A. Brack, The Molecular Origins of Life, Cambridge the amino acid molecules fragmentation, i.e. only few University Press, Cambridge 1998, p. 417. amino acids produce the same fragments despite their [6] National Institute of Standards, Standard Reference core part similarities. Database: Chemistry Webbook, http://webbook. The hydrogen atom substitution in glycine by the nist.gov. methyl group in alanine does not inuence the main frag- [7] J. Tamuliene, L.G. Romanova, V.S. Vukstich, mentation channel when the molecules undergo the low- A.V. Snegursky, Chem. Phys. 404, 36 (2012). [ ] J. Tamuliene, L.G. Romanova, V.S. Vukstich, energy electron impact: the CCα bond rupture and the 8 carboxyl group detachment, but the presence of the sul- A.V. Snegursky, Lith. J. Phys. 53, 195 (2013). fur atom due to the inuence on the HOMO character [9] J. Tamuliene, L.G. Romanova, V.S. Vukstich, A.V. Snegursky, Chem. Phys. 404, 74 (2012). leads to the dissociation occurred via the CβCγ bond rupture. [10] V.S. Vukstich, A.I. Imre, A.V. Snegursky, Instr. Exp. Our calculations have shown that the side chain of Tech. 54, 207 (2011). amino acid inuences both the atomic composition of [11] G. Hanel, B. Gstir, T. Fiegele, F. Hagelberg, fragments with the same mass and their structure. For K. Becker, P. Scheier, A. Snegursky, T. Märk, + J. Chem. Phys. 116, 2456 (2002). example, the C2H4N ion in the case of the alanine and methionine molecules has dierent structure. [12] A.D. Becke, J. Chem. Phys. 98, 5648 (1993). One may conclude, based on our data of both ex- [13] R.A. Kendall, T.H. Dunning Jr., R.J.J. Harrison, perimental and theoretical investigations, that for all J. Chem. Phys. 96, 6796 (1992). + [ ] G.A. Junk, H.J. Svec, J. Am. Chem. Soc. 85, 839 molecules studied the positively charged CH3N and 14 + (1963). CH2N ions could result from the secondary dissocia- + [ ] V.S. Vukstich, A.I. Imre, L.G. Romanova, A.V. Sne- tion due to the deprotonation of the CH4N ions, while 15 + gursky, J. Phys. B 43, 185208 (2010). the CH2N ion also appear due to the deprotonation of + [ ] H.-W. Jochims, M. Schwell, J.-L. Chotin, M. Clem- the CH3N ion. The observations of that kind could be 16 of great importance for studying the malignant trans- bino, F. Dulieu, H. Baumgärtel, S. Leach, Chem. formations in living cells under the inuence of ionizing Phys. 298, 279 (2004). radiation and also provide useful radiation therapy eect [17] J. Tamuliene, L.G. Romanova, V.S. Vukstich, on tumors in human beings. A.V. Snegursky, Nucl. Instrum. Methods Phys. Rev. B 279, 128 (2012). [18] P. Ipolyi, S. Cicman, V. Deni, K. Matej£ik, P. Mach, Acknowledgments J. Urban, P. Scheier, T.D. Märk, . Matej£ik, Int. J. Mass Spectrosc. 252, 228 (2006). The authors would like to thank Prof. M. Cegla [19] S. Bari, P. Sobocinski, J. Postma, F. Alvarado, (Jagiellonian University, Kraków, Poland) and our col- R. Hoekstra, V. Bernigaud, B. Manil, J. Rangama, league V. Patasiene (Vilnius University, Vilnius, Lithua- B. Huber, T. Schlathölter, J. Chem. Phys. 128, nia) for their technical support and assistance in carry- 074306 (2008). ing out this study. Special thanks go to the InSpire and [20] A.F. Lago, L.H. Coutinho, R.R.T. Marinho, A. Naves NGI.LT projects for the resources and support of the- de Brito, G.G.B. de Souza, Chem. Phys. 307, 9 oretical calculations provided as well as to the COST (2004). MP0802 activity. [21] Mr. Kent's Chemistry Page: Bond Enthalpy, www.kentchemistry.com. References

[1] U. Meierhenrich, Amino Acids and the Asymmetry of Life, Springer-Verlag, Berlin 2008. [2] L. Delaye, A. Lazcano, Phys. Life Rev. 2, 47 (2005).