Positive and negative ions of the amino acid histidine formed in low‐energy electron collisions

Abstract Histidine is an aromatic amino acid crucial for the biological functioning of proteins and enzymes. When biological matter is exposed to ionising radiation, highly energetic particles interact with the surrounding tissue which leads to efficient formation of low‐energy electrons. In the present study, the interaction of low‐energy electrons with gas‐phase histidine is studied at a molecular level in order to extend the knowledge of electron‐induced reactions with amino acids. We report both on the formation of positive ions formed by electron ionisation and negative ions induced by electron attachment. The experimental data were complemented by quantum chemical calculations. Specifically, the free energies for possible fragmentation reactions were derived for the τ and the π tautomer of histidine to get insight into the structures of the formed ions and the corresponding neutrals. We report the experimental ionisation energy of (8.48 ± 0.03) eV for histidine which is in good agreement with the calculated vertical ionisation energy. In the case of negative ions, the dehydrogenated parent anion is the anion with the highest mass observed upon dissociative electron attachment. The comparison of experimental and computational results was also performed in view of a possible thermal decomposition of histidine during the experiments, since the sample was sublimated in the experiment by resistive heating of an oven. Overall, the present study demonstrates the effects of electrons as secondary particles in the chemical degradation of histidine. The reactions induced by those electrons differ when comparing positive and negative ion formation. While for negative ions, simple bond cleav ages prevail, the observed fragment cations exhibit partly restructuring of the molecule during the dissociation process.


| INTRODUCTION
The ever-increasing amount of ionising radiation to which human beings are exposed to in modern society is understood to be the major cause of damage to living cells. 1,2 Electrons are one of the most abundant secondary species formed after primary radiation impact with yields of around 5 × 10 4 per MeV deposited energy of the incident radiation 3,4 and kinetic energies < 20 eV. Also in the case of ion beam tools for cancer therapy, it was shown that a large majority of secondary electrons produced have kinetic energies < 30 eV. 5,6 These electrons can react with molecular components of the cell to form positive and negative ions and radicals and can substantially contribute to the cell damage through sugar modifications, base release, and single strand and double strand breaks. 7,8 Thus, the underlying processes of the damage and the repair of biologically essential molecules, such as DNA and proteins, are of great importance and a subject of intense research. Especially the onset of the different processes induced by low-energy electrons leading to formation of positive and negative ions and radicals is crucial to provide more accurate predictions for ion beam cancer therapy, shielding of human space missions, prediction of the consequences of exposure to radiation, and for possible formation and detection of amino acids in extraterrestrial environments. 9 Histidine (His; see Scheme 1) is an essential aromatic amino acid that is crucial for the biological functioning of proteins and enzymes.
It is a precursor of the hormone histamine, an inflammatory agent in the immune response system, and a key residue in active sites of enzymes. 10,11 Due to the imidazole moiety, it can adopt two tautomeric conformations (Scheme 1) with H on either of the two N atoms referred to as the τ tautomer ( τ His) and the π tautomer ( π His), while it is suggested that the τ tautomer is preferred (80: 20) in the gas phase at 298 K. 12,13 While the electron ionisation mass spectra of most of the amino acids are available, 14 detailed knowledge on the ionisation energies (IE) of the molecules and the appearance energies (AE) of fragment ions formed upon ionisation are restricted to few amino acids investigated up to date. A literature survey reveals that the IEs and AEs were investigated experimentally for glycine, alanine, 15 and valine. [15][16][17] The vertical IE for His has been reported in the theoretical computational works of Huang et al 12 and Close,18 who examined the vertical IEs of all common α-amino acids. The formation of negative ions from amino acids upon resonant attachment of free electrons was previously studied as a function of the electron energy for glycine, proline, tryptophan, and methionine. [19][20][21][22][23][24][25][26][27][28] In contrast, for His, the electron energy dependence was neither studied nor discussed to the authors' knowledge. Voigt and Schmid studied negative ion mass spectroscopy of His in the gas phase, 29 where they used a low-temperature plasma source to generate low-energy electrons with approximate kinetic energies between 2 and 4 eV. They reported the relative abundances of fragment anions like [His-COOH] − in the negative ion mass spectrum and obtained the dehydrogenated parent anion [His-H] − as the most abundant fragment anion. 29 The electron energy dependence of negative ions formed via dissociative electron attachment (DEA) to His was only reported for the condensed phase by Abdoul-Carime and Sanche. 30 They measured the desorption of anions from thin films of His upon the impact of low-energy electrons with kinetic energies between 5 and 35 eV.
In the present gas-phase experiments, we investigate the formation of positive and negative ions of His in low-energy electron collisions. The measurement of temperature dependent mass spectra at the electron energy of 70 eV shows an increase of mass peaks with temperature. Such behaviour was previously assigned to a fractional thermal decomposition of the His sample during the sublimation process. Therefore, we assign the ion yield of the parent cation and the dehydrogenated parent anion to form unambiguously from the nondecomposed sample. For other fragment anions and fragment cations, the possibility of formation from possible thermal decomposition products was considered. As will be shown below, the comparison of experimental data and the results of quantum chemical calculations allow a feasible assignment of most obtained ion yields to the intact neutral His precursor.

| Chemicals
His with stated purity of 99% was purchased from Sigma-Aldrich and used as received. It appears as white powder under standard conditions. The vapour pressure of His required heating of the oven to about 160°C at around 10 −8 to 10 −7 mbar to allow the gas phase measurements with the utilised experimental set-up. The temperature of the ion source was with 90°C always lower than the oven temperature, excluding subsequent effects at a later stage. Wilson et al 31 performed VUV synchrotron ionisation studies of His and concluded that a fraction of His sample is already decomposed at the heating temperature of 100°C. Therefore, we also measured the temperature dependence of the electron ionisation mass spectrum at 70 eV in the course of the present experiments. The resulting mass spectra at three different temperatures of the sample are shown in Figure 1. The spectra are normalised to the ion yield of the parent ion at m/z 155. These data indicate that the most abundant ions above m/z 32 increase with oven temperature, as will be discussed in more details below. SCHEME 1 Molecular structures of two tautomeric forms of Lhistidine

| Mass spectrometry experiments
The present experimental data were recorded with a crossed electronmolecular beam set-up combined with a quadrupole mass spectrometer. It was described in detail elsewhere. 32 Briefly, electrons were emitted from a hairpin filament of the ion source and subsequently formed into an energetically and spatially focused beam by a hemispherical electron monochromator (HEM). A Faraday cup at the end of the HEM allows for checking stable electron beam conditions. The energy resolution for the present data was around 100 meV as determined from the full-width-half-maximum (FWHM) of the wellknown SF 6 − /SF 6 resonance. 33 The molecular beam originates from an oven filled with His that was mounted perpendicularly to the HEM. It was heated at abovementioned conditions to evaporate the His. The gaseous compound was guided through a capillary with 1 mm inner diameter leaving it as effusive beam and crossing the electron beam in the interaction region at single-collision conditions. Here, anions and cations were formed via dissociative electron attachment and electron ionisation, respectively. The ions were extracted by a weak electrostatic field into a quadrupole mass analyser. The mass selected ions were detected by a channel electron multiplier and processed by a discriminator and pulse counting unit.
For cations, ionisation and appearance energies were obtained by recording the ion efficiency curves of the mass-selected ions. The electron energies were varied between 5 and 17 eV. The electron energy range was adapted for each cation to include the threshold and a region of about 3 eV below and above the threshold. The energy scale was calibrated by measuring the well-known ionisation energy 34 of Ne at 21.56 eV.
For negative ions, the anion efficiency curve of the mass-selected products was obtained in the electron energy range of~0 to 17 eV.
The well-known resonance position at 0 eV of SF 6 − /SF 6 33 was deployed for calibration of the electron energy scale.

| Determinations of IEs/AEs
The behaviour of ionisation cross sections in the threshold region was first described by Wigner and later extended by Wannier. Wigner theoretically developed a simple power law, where the shape of the crosssection close to threshold depends on the number of outgoing electrons. 35 In case of single ionisation by impact of an electron, he predicted a linear behaviour of the cross section. Wannier extended the theory for electron ionisation processes leading to a three charged particles final state (two electrons and one ion), 36 where such developed model describes the cross-section behavior close to the threshold.
Including a Heaviside function θ, the cross-section σ can be expressed as follows: where b is a linear background, c is a scaling parameter, E is the electron energy, AE is the appearance energy, and n is the exponent.
Wannier calculated n only for hydrogen 36 and must otherwise thus be determined experimentally.
The experimental set-up entails a finite energy resolution in form of a Gaussian distribution with the standard deviation ρ representing the resolution. The crosssection function is convoluted with the Gaussian profile giving the prediction for an experimentally measured cross-section curve with resolution ρ where Γ is the gamma function and D n+1 is a parabolic cylinder function.
This function for the cross-section allows fitting the experimental data. As a result, the parameters of such function, especially the AE, can be extracted. For this purpose, a software tool was developed based on a previous version 37 written in Python, 38 using the SciPy, 39 NumPy, 40  Errors resulting from the fit can be extracted from the covariance matrix. A further uncertainty arises from the setting of the fitting range.
We conducted test series and found that the determined onset value remains stable within maximal deviations of 50 meV for fitting ranges ≥3 eV, which was then chosen as standard input value. Thus, the statistical uncertainty on an AE value consists of the uncertainty arising from the fit and the uncertainty caused by the choice of the fitting range. For the analysis, the square of their quadratic addition is stated.
Since the energy scale was calibrated with this analysis tool, the uncertainty on the AE of the neon ion yield transfers into a systematic uncertainty of all analysed cations and amounts to 10 meV. Since the value is equal for all cations, it will not be stated during the results part for reasons of clarity. It is important to note that the energy resolution of the HEM and the error on the AE value are two different parameters.
For channels suggesting two or more appearance energies, the number of thresholds is a further input parameter for the programme.
The first onset is fitted as described above, and subsequently, the fit is subtracted from the data and the same method is applied on the next threshold. The defined fitting ranges may each only include one threshold.

| Determination of peak onsets in anion efficiency curves
For data analysis of the DEA processes, (multiple) Gaussian fits were applied to the experimental data curves. The peak maxima x max,i of each i-th peak was taken from the fit together with the error. To obtain the onsets of the reaction, it has to be taken into account that the resolution of the set-up broadens the peak by introducing a natural tail. Earlier methods 42,43 applied linear fits to the steepest part of the Gaussian profile which resulted in large uncertainties in the onset.
Here, a new method is introduced. The onset is defined as follows: with x max,1 the peak maximum and σ the standard deviation of the first peak. This cuts the outbounding tail at a defined position, making the method both robust concerning uncertainties and reproducible. Two σ were chosen as the tail is cut at a comparable level to earlier methods.

| Quantum chemical calculations
The lowest energy structures of His tautomers, τ ( τ His) and π ( π His), were taken from the study of Stover et al. 13 Quantum chemical calculations employing M062x/aug-cc-pVTZ level of theory 44,45 and basis set 46,47 were carried out to calculate ionisation energies, adiabatic electron affinities, and the free energy of reactions, ΔG(298K), which is calculated for each fragmentation pathway as ΔG = ΣG (products) − ΣG (reactants). 48 Calculated frequencies confirmed that the structures are local minima on the potential energy surface and not transition states. We estimate an error of less than 0.09 eV for the reaction energies and 0.11 eV for ionisation potentials from the reported mean unsigned error for M062x thermochemistry and ionisation potentials, respectively. 44 We have carried out some computations with the DSD-PBEP86 double-hybrid DFT method, 49 which performs well for a wide range of chemical properties. 50 The results show that M062x compares favourably to this higher level method, which lends support to the use of M062x more generally for the present set of systems.

| The mass spectrum at~70 eV and electron ionisation close to threshold
The mass spectrum shown in Figure 1A reveals that the parent ion His˙+ at m/z 155 is a minor species among the other ions observed.
Therefore, one may conclude that strong fragmentation may occur  Figure 1) while the ion source temperature of 230°C was considerably higher than the presently used 90°C.
We note that electron 53  thermal decomposition was already operational at the His sample temperature of 100°C; however, we note that they produced gas-phase His by particle evaporation. Interestingly, they also observed a peak at m/z 111, which was about 1.5 times more abundant than m/z 110. These authors proposed that His loses CO 2 by thermal decompo- This assignment is further supported by the constancy of the ion signal at m/z 111 when the sample temperature is increased (see Figure 1C).
This means that the increase of the [His-COOH] + isotope and the decrease of the background signal, both relative to the signal of the parent ion, keep their balance with increasing temperature.   The presently observed ion yields of fragments may result through the following reactions: The experimental threshold for the product ion [His-COOH] + at m/z 110 is (8.52 ± 0.08) eV (Table 1). This fragment ion may be formed from His through a simple bond cleavage releasing CO 2 + H˙, see reaction (5) and Figure 3. The calculated bond dissociation energy is similar for the τ His and the π His tautomers (3.27 and 3.19 eV, respectively).
The calculated free energy of reaction (5)   signal by a factor of about 1.5 %, when the temperature is increased from 160°C to 190°C. We ascribe this effect rather to the increase of internal energy in the His precursor instead of the increase in thermal decomposition. The alteration of mass spectra due to the increase of molecular temperature was previously reported for various molecules. 60 We further note that neutral [His-COOH]˙is a radical, and one may expect that even electron species are formed by thermal decomposition.
The fragment ion [C 4 N 2 H 6 ]˙+ at m/z 82 has a measured appearance energy of (8.54 ± 0.04) eV, which is in fair agreement with the   9.67 eV for π His, which would be in good agreement with the experimental threshold of (9.56 ± 0.11) eV (see Table 1). Simple the conclusion from the mass spectra discussed above (m/z 82 is a fragment ion from ionised His), and therefore, we assign the experimentally found threshold of (9.56 ± 0.11) eV to dissociative ionisation of His.
The fragment ion at m/z 44 exhibits two clearly visible thresholds in the ion efficiency curve (see Figure 2). We found the first threshold at (10.23 ± 0.09) eV and the second at (13.8 ± 0.5) eV. It should be noted that the ion yield associated to the first threshold is much less abundant than the ion yield rising above the second threshold. The   The second threshold at (13.8 ± 0.5) eV is likely due to the ionisation of neutral CO 2 formed by thermal decomposition of His, reaction 8b.
The calculated IE value (13.94 eV) is only slightly higher than the experimental IE of 13.78 eV reported on the NIST homepage. 14 We also investigated computationally if CO 2˙+ may be formed upon dissociative ionisation of His. Since the free reaction energies for CO 2˙+ formation from His is 13.45 eV for τ His and 13.33 eV for π His, we can exclude this process. Further breaking of the neutral fragment would require more than 14.10 eV.
As noted above, from the experimental ionisation energy of His it is not obvious, whether τ His or π His is present in the beam. When comparing the calculated thresholds for the two tautomers (see Table 1

| Dissociative electron attachment
In ). Figure 4 shows the ion yields as a function of the incident electron energy together with cumulative multiple fits of the experimentally observed peak structures. The anion assignment and the composition of the corresponding neutral products for each reaction channel are summarised in Table 3 together with the measured peak positions. The respective structures of anions and neutral products are shown in Figure 5. The dehydrogenated parent anion is formed in the following DEA reaction upon electron attachment to His: where (His˙−) # denotes the temporary negative ion (TNI) formed by initial resonant capture of a free electron. In general, the TNI may undergo dissociation into various stable anions, where the excess charge is retained, and including at least one neutral (fragment) formed. However, auto-detachment is also in competition with DEA, where the excess charge is spontaneously emitted from the TNI. The ion yield of [His-H] − is shown in Figure 4A) and exhibits two peaks at 0.73 eV (with the experimental threshold of (0.4 ± 0.1) eV) and  The dehydrogenation reaction upon DEA to amino acids is often an abundant channel and was therefore observed in linear amino acids like for example glycine, 19 alanine, 64 methionine, 23 and aromatic amino acids like proline. 65,66 Early DEA studies with formic acid 67 as a model molecule for amino acids suggested that hydrogen loss from the carboxyl group proceeds through electron attachment into the π * orbital of the −COOH group. As the position of the observed resonances and the shape of the ion yields from the amino acid is identical to the one observed for formic acid HCOOH, this suggestion was tentatively used also for amino acids. 19 However, more recent experimental 68 and theoretical studies 69 favoured a direct mechanism with attachment of the excess electron into the σ * (OH) orbital. In all previous cases where the anion yield was measured as a function of the electron energy, the ion yield had a characteristic shape with an abrupt steep onset at~1 eV. In contrast, in the present case of His, the ion yield has a notably very different shape and the first peak observed is at lower energy, ie, 0.73 eV. Moreover, it is also not the most abundant anion among the observed fragment anions in contrast to other amino acids. Interestingly, in DEA to His the shape of the ion yield of [His-H] − shown in Figure 4A rather resembles the shape of the dehydrogenated anion formed in electron attachment to gas-phase imidazole, 70 but it is shifted here by about~1 eV towards lower energies. In the case of imidazole, it was suggested that the electron attachment takes place through the π * resonance that couples to a dissociative σ * state localised at the N−H bond. 70 Such is only possible if the nuclear wave packet survives long enough to allow the system adiabatic crossing between states. According to the quantum chemical calculations, the removal of H • from the N position of the imidazole moiety of His requires at least 0.45 eV. It is therefore indeed energetically possible that the observed peak structure results from H-loss from the imidazole. However, also H-loss from the COOH group may occur as well.
As mentioned in the Introduction section, Voigt and Schmidt studied negative ion mass spectroscopy of His in the gas phase. 29 They obtained  Note. The uncertainty of the peak positions amount to ≤0.05 eV resulting from the Gaussian fit and stepwidth set. The uncertainty stated for the experimentally derived threshold relates to the fit. a M062x/aug-cc-pVTZ calculated free energies of reactions ΔG(298K).
b Values require not just a simple bond cleavage but also rearrangement to form the most stable neutral shown in Figure 5.
shows that the imidazole moiety, which has released the H˙from the N position, retains preferentially the captured electron, which makes sense due to its particular high electron affinity (2.6 eV). 70 Figure 4C shows (see Figure 5). This structure has an AEA of 2.47 eV, which is even higher than the imidazole radical formed by simple C α −C β bond cleav-   Table 3 for exact positions). The OH − anion may be formed by simple C−OH bond cleavage. The corresponding neutral fragment is shown in Figure 5. The anion yield of OH − is depicted in Figure 4D. arises from another (so far unknown) process than DEA of a single electron to His. We note that a highly abundant peak at zero eV (exceeding features at higher energies by a factor of about 10) was observed for OH − in DEA to tryptophan. 20 The origin of this signal remained unclear. In contrast, for proline and aliphatic amino acids, features in the OH − ion yields close to zero eV were rather weak and lower in intensity compared with features at higher energies, and they were assigned to hot band transitions 65 or impurities. 71 The most prominent peak observed in the present study is located at the electron energy of 7 eV (see Figure 4D). In comparison, the ion yield of OH − desorbed from an electron bombarded condensed film of His showed a peak at the electron energy 30 of 7.7 eV noting also a shoulder of another faint peak. Therefore, we may assign these peaks observed in the gas phase and the condensed phase to the same TNI state of His, in view that anions formed with low kinetic energiesrepresenting the low-energy tail of a resonance-may be more strongly discriminated against in the desorption process than anions with higher kinetic energies. 72 The most abundant anion observed in the present experiments is found at m/z 16, assigned to O˙−/NH 2 − . The mass resolution of the used quadrupole mass spectrometer is not sufficient to separate these isobaric anions. The measured anion yield is shown in Figure 4E, and the corresponding neutrals formed are shown in Figure 5, if formed from His. The anion yield shows 5 peaks above the experimental threshold of (3.9 ± 0.1) eV (Table 3). Utilising a double focusing mass spectrometer, it was shown in previous DEA studies to the amino acids valine, 22 beta alanine 73 and glycine 74 that both isobaric ions were formed. These studies showed that NH 2 − is formed with a threshold of~4 to 5 eV dominating at electron energies of 5 to 8 eV, whereas O˙− was observed dominantly in a peak close to 4 eV and in the region of 7 to 15 eV. In the valine ion yield measured previously, 22 the O˙− peak close to 4.4 eV and a second peak at 8.2 eV resembled the characteristics 75,76 of O˙`− from CO 2 and were ascribed to thermally decomposed valine. Also in the present case, CO 2 is present in the molecular beam (see above), and therefore, the first peak at about 4.3 eV may be assigned to O˙−/ CO 2 . The second well-known peak 75,76 at 8.2 eV in the O˙− from CO 2 may be obscured due to stronger abundant peak at 9.2 eV. We further note that the latter peak may also arise from water impurities in the sample. This hypothesis is further supported by the presence of other peaks at about 7 and 11.7 eV which were also found for O˙− from H 2 O. 77 Therefore, one may conclude that only the peak at 5.74 eV found experimentally results from DEA to His, corresponding to formation of the NH 2 − fragment ion.
Our quantum chemical calculations predict that the thermodynamic threshold for removal of O˙− from C═O requires 5.56 eV for τ His and 5.72 eV for π His. However, removal of O˙− from the −COOH site with H transfer to C (see Figure 5) lowers the threshold

| CONCLUSIONS
In the present study, we investigated electron ionisation of histidine close to threshold of the different fragments and dissociative electron attachment in the electron energy range from about 0 up to 18 eV.
The experimental data were supported by quantum chemical calculations, which derived free energies of reactions and allowed the identifi- Such unusual order of ionisation and appearance energies is very scarce in electron ionisation and was previously reported for CHF 2 Cl. 78 In the case of negative ion formation, five fragment anions were obtained in the course of the present studies. Like for all other studied aromatic (as well as aliphatic) amino acids so far, no molecular anion was observed within the detection limit of the apparatus, which indicates that the temporary negative ion undergoes fast which is a gentle method to transfer neutral molecule into the gas phase. The only volatile thermal decomposition product of His detected in the present mass spectrum is carbon dioxide, which was straightforward to identify in the electron energy scans due its wellknown ionisation energy and resonance energies.