Low-Energy Electron Interaction with Melatonin and Related

Apr 10, 2017 - Boccalandro , H. E.; Gonzalez , C. V.; Wunderlin , D. A.; Silva , M. F. Melatonin Levels, Determined by LC-ESI-MS/MS, Fluctuate during ...
0 downloads 7 Views 2MB Size
Article pubs.acs.org/JPCB

Low-Energy Electron Interaction with Melatonin and Related Compounds Stanislav A. Pshenichnyuk,*,†,‡ Alberto Modelli,§,∥ Derek Jones,⊥ Eleonora F. Lazneva,‡ and Alexei S. Komolov‡ †

Institute of Molecule and Crystal Physics, Ufa Research Centre, Russian Academy of Sciences, Prospeκt Oktyabrya 151, 450075 Ufa, Russia ‡ St. Petersburg State University, Universitetskaya nab. 7/9, 199034 St. Petersburg, Russia § Dipartimento di Chimica “G. Ciamician″, Università di Bologna, via Selmi 2, 40126 Bologna, Italy ∥ Centro Interdipartimentale di Ricerca in Scienze Ambientali, via S. Alberto 163, 48123 Ravenna, Italy ⊥ ISOF, Istituto per la Sintesi Organica e la Fotoreattività, C.N.R., via Gobetti 101, 40129 Bologna, Italy ABSTRACT: The electron attaching properties and fragmentation of temporary negative ions of melatonin and its biosynthetic precursor tryptophan are studied in vacuo using dissociative electron attachment (DEA) spectroscopy. The experimental findings are interpreted in silico with the support of Hartree−Fock and density functional theory calculations of empty orbital energies and symmetries, and evaluation of the electron affinities of the indolic molecules under investigation. The only fragment anions formed by DEA to melatonin at incident electron energies below 2 eV are associated with the elimination of a hydrogen atom (energetically favored from the NH site of the pyrrole ring, leaving the ring intact) or a CH3· radical from the temporary molecular negative ion. Opening of the pyrrole ring of melatonin is not detected over the whole electron energy range of 0−14 eV. The DEA spectra of L- and D-tryptophan are almost identical under the present experimental conditions. The adiabatic electron affinity of melatonin is predicted to be −0.49 eV at the B3LYP/6-31+G(d) level, indicating that the DEA mechanism in melatonin is likely to be present in most life forms given the availability of low energy electrons in living systems in both plant and animal kingdoms. In particular, H atom donation usually associated with free-radical scavenging activity can be stimulated by electron attachment and N−H bond cleavage at electron energies around 1 eV. scavenging activity,12,13 MLT is much more effective due to its cascade effect,14,15 distinguishing it from other classical antioxidants, where together with its secondary and tertiary metabolites, it can be 1 order of magnitude more effective. In humans the role of MLT has been studied in antiaging,16−18 immune system regulation,7 oncostasis9 among its many others. A review of its wide range of clinical uses in humans is available.19 The presence of MLT in plants was discovered in 199520,21 and its role in plants has been investigated only relatively recently.22 Its concentration in plants can be more than double that found in vertebrates,23−25 due to the additional presence of MLT-generating chloroplasts26 which are absent in animals. It regulates the growth of roots, shoots, and explants, activates seed germination and rhizogenesis, delays leaf senescence, increases photosynthesis efficiency, and serves as a biocide against fungi and bacteria.27,28 There is now great interest in phytomelatonin15 and modern analytical methods have

1. INTRODUCTION A century has passed since McCord and Allen1 discovered evidence of an “active substance” capable of inducing pigmentation changes, irrespective of environmental conditions, in tadpoles fed with pineal gland. Forty years later Lerner et al. isolated this “active substance” and determined its structure,2−4 suggesting it be called melatonin (MLT). Since then, this molecule, especially in its biological roles, has been extensively studied, leading to over 30 thousand publications in the scientific literature involving MLT. It is present in all living organisms studied, in all three domains of life (archaea, bacteria, eukarya), along the whole evolutionary chain. Its amphiphilic nature allows it to diffuse quite freely5,6 to effect its pleiotropic7 activities throughout the organism. Most MLT studies involve its multifarious roles in the plant and animal kingdoms. In vertebrates MLT is not only secreted by the pineal gland from which it reaches the bloodstream but is also synthesized by, and has been detected in, most other organs, tissues, and biological fluids.8,9 MLT is known to regulate circadian rhythms,10,11 and has also been widely studied as an antioxidant and radical scavenger. Although tryptophan (TRP) also shows antioxidant and free-radical © 2017 American Chemical Society

Received: February 13, 2017 Revised: March 31, 2017 Published: April 10, 2017 3965

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B

Chart 1. Structures of Melatonin (Acetamide, N-[2-(5-Methoxy-1H-indol-3-yl)ethyl]-) and Tryptophan (2-Amino-3-(1H-indol3-yl) Propanoic Acid) Showing the Three Synthetic Paths to Melatonina

a

Continuous line: in animals and dinoflagellates; dotted line: in plants; dashed line: recently proposed alternative pathway33).

measured high concentrations of MLT in Chinese herbs29 and even allowed monitoring of MLT levels in grapes30 to optimize harvesting time. Recent work has also shown that increases in MLT levels in plants are induced when under stress from extreme cold, heat, solar irradiation and pollution.14,15 The structure of MLT and of its precursor L-TRP are shown in Chart 1. Its synthesis in vertebrates, increasing with darkness,31 starts with the 5-hydroxylation of L-TRP, followed by decarboxylation to serotonin, side-chain N-acetylation, and finally O-methylation of the 5-OH group. In plants, however, LTRP is first decarboxylated before being 5-hydroxylated to produce serotonin, MLT synthesis then continuing along the vertebrate pathway.32 Recently, a further synthetic route has been proposed,33 where serotonin is methoxylated before sidechain N-acetylation, and which may predominate in some organisms. Cellular MLT distribution favors the membrane and mitochondrion with respect to the nucleus and cytosol,34 but there are significant differences in its concentration in different organs which may be related to specific functions in the cell. For example, MLT concentrations in cerebral cortex mitochondria are much higher (mitochondria > cell membranes > nuclei > cytosol) than in the liver (cell membranes > mitochondria > nuclei > cytosol) perhaps due to the brain’s need for more protection as it uses much more oxygen.35 In fact, the primary role of MLT is that of an antioxidant and it has even been described as a “molecule which makes oxygen metabolically tolerable”.36 MLT, by scavenging reactive species in mitochondria, lowers protein and mitochondrial DNA damage and improves electron transport chain (ETC) activity, also contributing to mitochondrial homeostasis37 and may even control currents through the mitochondrial permeability transition (MPT) pore opening which leads to apoptosis.38 MLT also modulates mitochondrial respiratory activity by increasing or decreasing effective electron transport through the inner mitochondrial membrane.39 For TRP free-radical scavenging, two major pathways have been found: sequential proton loss electron transfer (SPLET) and sequential double proton loss electron transfer (SdPLET).13

The use of gas-phase free molecule-electron interactions as an aid to understanding in vivo processes began with the pioneering work of Lovelock40−42 and Gregory,43 since then leading to increasing numbers of important studies in this field.44−52 Molecular species can, in fact, be studied using gas phase electron transmission and dissociative attachment spectroscopies (ETS, DEAS) allowing an experimental determination of empty level structures through electron attachment and the subsequent destiny of their resulting temporary anion states.53,54 ETS and DEAS studies, together with theoretical calculations, have allowed this research group to elucidate some of the in vivo mechanisms of common pharmaceuticals, such as antipyretics,55 the antimalarial artemisinin, and its derivatives,56,57 biologically active compounds including flavonoids,58,59 Vitamin C,60 Anthralin,61 xenobiotics,62 Resveratrol,63 plant hormones,64 as well as indole and related molecules,65 the latter compound being the allimportant nucleus of the MLT and TRP molecules studied here. Thus, the present paper reports on the gas-phase interaction of low-energy electrons with MLT and TRP (in its L- and D-forms; see Chart 1) and their resulting temporary negative ions by means of DEAS and quantum-chemical calculations, to continue our investigations into the electron accepting properties of indole-moiety-based molecules.

2. EXPERIMENTAL AND COMPUTATIONAL PROCEDURES Low-energy electron attachment to the compounds under investigation is characterized using DEAS.54,66 Our magneticsector mass spectrometer has been described previously.67 Briefly, a magnetically collimated electron beam of defined energy is passed through a collision cell containing the sample in gas phase, under single-collision conditions. The resulting mass-selected negative ion signal is recorded as a function of the incident electron energy in the 0−14 eV energy range. The electron-energy scale is calibrated with the SF6− signal at zero energy, generated by the attachment of thermal electrons to SF6. The full-width at half-maximum (fwhm) of the electronenergy distribution is 0.4 eV, and the accuracy of the measured peak positions within 0.1 eV. The samples used are 3966

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B Chart 2. B3LYP/6-31+G(d) Relative Stabilities.a

a

(a) D-TRP; Conformers II and III are 0.017 and 0.042 eV less stable than I, respectively. (b) L-TRP; Conformers V and VI are 0.016 and 0.068 eV less stable than IV, respectively. (c) MLT; Conformers VIII and IX are 0.089 and 0.152 eV less stable than VII, respectively.

commercially available (Aldrich # M5250, T0254, T9753) and were used without additional purification. Acceptable anion current intensity for MLT was obtained by heating the collision cell to 115 °C, close to its melting point (118 °C). Both tryptophans required DEAS collision cell temperatures of up to 190 °C. Calculations were carried out with the Gaussian 09 set of programs.68 Evaluation of the virtual orbital energies (VOEs) of the neutral molecules was performed with the 6-31G(d) basis set at the Hartree−Fock (HF), with MP2 optimized geometries, and B3LYP levels of theory.69 The B3LYP hybrid functional with the standard 6-31+G(d) basis set, which includes the minimum addition of diffuse functions (s and p type diffuse functions to the non-hydrogen atoms), was used to evaluate the thermodynamic energy thresholds for production of negative and radical fragments, obtained as the total energy of the (geometrically relaxed) ground-state fragments relative to the neutral ground-state molecule. Zero-point vibrational energy corrections were also calculated. An approach adequate for the description of unstable anion states involves difficulties not encountered for neutral or cation states.70−73 The most correct method is, in principle, the calculation of the total scattering cross section with the use of continuum functions, but complications arise from the lack of an accurate description of the electron-molecule interaction.74 A proper description of the spatially diffuse electron distributions of anions normally requires a basis set with diffuse functions.75,76 However, calculated anion state energies decrease as the basis set is expanded, so the choice of an appropriate basis set is a delicate task.70,71,77,78 The Koopmans’ theorem (KT) approximation79 neglects correlation and relaxation effects. However, it has been demonstrated70,72 that good linear correlations exist between the energies of vertical attachment (VAEs) to π*CC MOs measured in a large number of alkenes and benzenoid hydrocarbons and the corresponding VOEs of the neutral molecules obtained with simple HF calculations, using basis sets which do not include diffuse functions. More recently it has been shown77 that similar linear correlations are also found with the π* VOEs obtained with B3LYP/6-31G(d) calculations. Here we use two linear correlations to scale the π* VOEs of D-TRP and MLT. The first (VAE = 0.64795 VOE − 1.429872) was derived from HF/

6-31G(d) VOEs obtained for the geometries optimized at the MP2/6-31G(d) level and the second (VAE = 0.8065 VOE + 0.9194)80 from the B3LYP/6-31G(d) π* VOEs of a series of alternating phenyl and ethynyl groups. In a recent ETS/DEAS study65 the scaled VOEs (SVOEs) obtained with both equations nicely reproduced the π* VAEs measured in the ET spectrum of the reference molecule indene. Additionally, we have previously found a good correspondence between scaled VOEs of small conjugated molecules and energies of peaks in the density of vacant electronic states of adsorbed ultrathin films of these molecules81,82 where the molecular anions constitute the main structural units of the interfacial molecular layer.83,84 Such scaled VOEs are also relevant in intermolecular charge transfer with charge generation in organic photovoltaics85 as well as for charge transport in nanojunctions.86,87 A review of the dynamics of temporary negative ions formed via resonance electron attachment under electron-molecule collisions was recently reported.88

3. RESULTS AND DISCUSSION 3.1. Empty-Level Structure. TRP and MLT possess five empty π* MOs, deriving from the three benzene π* MOs, a conjugated ethene π* MO, and a remote carbonyl π* MO. Both molecules can be viewed as derivatives of the reference bicyclic π-system of indole, with a substituent bearing a carbonyl group. One of the most suitable means for measuring the energies of formation of gas-phase temporary anions is the ETS technique devised by Sanche and Schulz.53 Unfortunately, at the maximum temperature (110 °C) of the collision chamber attainable with our ETS apparatus, the vapor pressure of the three compounds was not sufficient to obtain a suitable attenuation of the electron beam. Chart 2 represents the most stable conformers of L-TRP, DTRP, and MLT and their relative energies, as obtained with B3LYP/6-31+G(d) calculations. Similar results are supplied by B3LYP and HF//MP2 calculations using the 6-31G(d) basis set. The π* VAEs (0.99, 1.88, 2.78, and 4.7 eV) measured in the ET spectrum of indene65 and the π*CO VAE of acetone (1.31 eV)89 suggest that in the present compounds the five anion states are expected to lie below 6 eV, and that even the lowest3967

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B

mainly π*CO and benzene π* character, respectively, as found in MLT, while HF//MP2 gives the opposite energy sequence. The lowest-lying anion state of D-TRP is predicted to be unstable by about 0.7 eV, i.e., 0.3 eV more stable than that measured in the ET spectrum of indene, in line with the presence of an electron-withdrawing carbonyl group. In MLT this effect is balanced by the destabilization caused by mixing of the oxygen lone pair of the methoxy substituent with the adjacent π* MOs.89 The next three π* anion states of D-TRP are predicted to lie at 1.1, 1.5, and 2.3 eV, the latter possessing mainly π*CC character. A similar energy trend is predicted in MLT, all the π* SVOEs being 0.3−0.5 eV higher than the corresponding SVOEs of TRP. 3.2. Fragmentation by Low-Energy Electrons. As a general comment for all three compounds, even at the highest permissible evaporation temperatures (avoiding thermal decomposition) the DEA spectra show a low signal-to-noise ratio. Mass-selected anion currents as a function of incident electron energy are reported in Figures 1, 2, and 3. Likely

lying anion state is unstable (negative electron affinity). The results supplied by both scaling procedures (see Table 1) Table 1. HF/6-31G(d)//MP2/6-31G(d) and B3LYP/631G(d) VOEs, and Corresponding Scaled Values (SVOEs)a HF//MP2 orbital

VOE

π*O σ*NH2 σ*OH σ*NH ring π*CC π*A π*CO πS* + π*CC

10.257 6.867 6.370 5.969 5.803 4.496 4.946 3.286

π*O π*CC σ*NH ring σ*NH am π*S π*CO π*A+ π*CC

11.037 6.456 6.281 5.751 5.388 5.099 3.745

B3LYP

SVOE

VOE

SVOE

4.948 2.966 2.124 2.037 1.651 0.722 0.235 −0.318

4.91

D-tryptophan

a

5.22

2.33 1.48 1.77 0.70 melatonin 5.72 2.75

2.06 1.87 1.00

5.495 2.078 2.322 1.888 1.387 0.831 0.104

2.25 1.50 1.11 0.66 5.35 2.59

2.04 1.59 1.00

All values are in eV.

confirm this expectation. The first column of Table 1 indicates the main character of the empty MOs. Schematic representations of the lowest three empty MOs for D-TRP and MLT are reported in Chart 3. The symmetric and antisymmetric Chart 3. B3LYP/6-31G(d) Representations of the Lowest Three Vacant MOs for (a) D-TRP and (b) MLT

components of the benzene e2u (π*) LUMO are labeled π*S and π*A, respectively, while the higher-lying totally antibonding benzene MO is labeled π*O. The π* MOs with mainly ethene and carbonyl character are denoted as π*CC and π*CO, respectively. Systematic ETS studies90 have demonstrated that hydrocarbons without third-row or heavier heteroatom substituents do not generally give rise to distinct low-energy resonances associated with empty MOs of (local) σ symmetry (due to their shorter lifetimes), so that empirical equations to scale the σ* VOEs of the present compounds are not available. The two sets (HF//MP2 and B3LYP) of scaled π* VOEs are in good agreement (see Table 1). The main difference lies in the localization properties of the second and third anion states of D-TRP. According to the B3LYP calculations they possess

Figure 1. DEA spectra of D-tryptophan.

structures of the negative fragments, peak energies and relative intensities are listed in Table 2, while the calculated thermodynamic energy thresholds for fragment formation are given in Table 3. Under the present experimental conditions, DEA spectra of L- and D-tryptophans (Figure 1 and Figure 2) are almost identical. Small differences are seen in the relative intensities whereas peak positions are the same within the reported accuracy (see Table 2). As expected for compounds containing the indole moiety with alkyl substituents65 the most intense anion fragment in the DEA spectra of TRP is the dehydrogenated molecular anion 3968

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B

Figure 2. DEA spectra of L-tryptophan. Figure 3. DEA spectra of melatonin. Thick green line indicates isotopic contribution from the m/e = 216 to m/e = 217 current.

[TRP−H]− (m/e = 203) observed at an incident electron energy of 1.1 eV. Hydrogen loss is therefore likely associated with electron addition to the second vacant π* MO of TRP, predicted by the B3LYP calculations to possess mainly π*CO character, but having mainly benzene character according to the HF calculations with a significant localization on the adjacent nitrogen atom (see Table 1). In fact, according to the B3LYP/ 6-31+G(d) thresholds (see Table 3), at this energy abstraction of a H atom is possible only from the NH group of the indole moiety. Given the structure of the chain substituent the intense signal observed at m/e = 71 could be ascribed to elimination of negatively charged carbonocyanidic acid NCCOOH− from the temporary negative molecular ions of both L- and D-TRP. According to the calculations (see Table 3), this negative fragment can be formed below 0.5 eV together with the closedshell neutral molecules H2 and 3-methylindole, a rather unusual decay which produces three fragments. The observed signals peaking at 0.3−0.4 eV are thus associated with the low-energy portion of the first shape resonance (predicted VAE = 0.7 eV). The m/e = 74 signal is likely due to the glycinate anion CHNH2COOH−. Its relative intensity is low, even though its formation requires only the simple cleavage of a C−C single bond. This signal peaks at about 1.5 eV, in agreement with a calculated energy threshold of 0.9 eV (see Table 3), and is thus ascribed to electron capture into the third π* MO (predicted VAE = 1.5 eV). The carboxyl group of TRP is also involved in two more dissociative decay channels stimulated by electron attachment over a broad energy range, as observed for m/e = 142 and 158. These weak signals could be associated with elimination of a neutral COOH fragment along with an NH3 group or a H atom, respectively. Provided that closed-shell H2 and CO2

molecules are formed as neutral counterparts of the m/e = 158 negative fragment, the reaction is found to be possible on energetic grounds close to 1 eV (see Table 3). The energy threshold for elimination of a neutral NH2 group from the molecular anion to form the [TRP−NH2]− species (m/e = 188) is predicted to be 1.28 eV, in line with the observed peak energies of 1.4−1.5 eV. The complementary dissociative channel leading to generation of NH2− (m/e = 16) is energetically favored with respect to O− formation (see Table 3). The energy thresholds calculated for both these decays are below the observed m/e = 16 peak (5.5 eV), but formation of NH2− requires only the cleavage of one single C−N bond and is therefore preferred to O− elimination. It should be mentioned that the relative intensity of the NH2− anion is the most sensitive to the structure (L- or D-isomer) of the target molecule. Namely, the m/e = 16 anion current is three times more intense in the biologically inactive D-TRP than in the Lisomer, so that the C-NH2 bond should be stronger in the latter. The observed peak energies for the m/e = 16 currents (5.5−5.6 eV) lie above the highest π*O MO (predicted VAE = 4.91 eV, see Table 1), and can thus be associated with coreexcited resonances. Despite that loss of CN− from the molecular negative ion accounting for the observed m/e = 26 current is accompanied by strong rearrangements and cleavage of several covalent bonds, this decay is found to be possible below 1 eV provided that the indole moiety of the neutral counterpart rearranges to form an aromatic benzene ring (structure 3 in Chart 4). In fact, this signal is observed with high intensity at incident electron energy of 1.5−1.6 eV (see Table 2). Formation of OH− (m/e = 3969

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B Table 2. Likely Structure of Fragment Anions Formed by DEA, Peak Energies (eV) and Relative Intensities Taken from Peak Maximaa m/e

anion structure

peak energy

Table 3. B3LYP/6-31+G(d) Energies (eV) Relative to the Neutral Ground Statesa m/e

relative intensity

203 203 188 158 158 142 142

D-tryptophan/L-tryptophan

203 188

[TRP−H]− [TRP−NH2]−

158

[TRP−H2−CO2]−

142

[TRP−NH3−COOH]−



74

CH(NH2)COOH

71 26 17 16

CNCOOH− CN− OH− NH2− or O−

231

[MLT−H]−

217 216

[MLT−CH3]− [MLT−CH4]− or [MLT−NH2]− or [MLT−O]−

202 189

[MLT−NO]− [MLT−C(O)CH3]−

172 145 144 58

[MLT−HNC(O)CH3 − H2]− [MLT−CH3CH2N(H)C(O)CH3]− [MLT− N(H)C(O)CH3 − OCH2]− HNC(O)CH3−

1.1/1.1 1.5/1.4 5.0/5.1 1.2/1.7 5.5 sh./5.5 sh. 7.4/7.4 1.2/1.0 5.0/5.0 7.7/7.8 sh. 1.6/1.5 5.7 sh./5.9 sh. 7.9/8.0 0.4/0.3 1.6/1.5 5.0/5.4 5.5/5.6

100/100 1.4/1.5 0.1/0.1 0.3/0.2 0.7/0.6

1.3 4.8 8.2 1.5 6.5 sh. 8.9 8.2 5.2 9.1 8.3 9.7 9.5 8.8

100 4.2 4.6 0.3

74

0.2/0.2 0.3/0.6 0.3/− 0.7/0.6 0.8/0.4

71 26 26 26 26 17 16 16

21/45 11/6.7 1.0/0.5 0.3/0.1

231 231 217 217 216 216 216 202 189 172

melatonin

a

5.4 0.2 0.9 0.3 0.9 0.3 0.6 2.2

145 144

sh: stands for shoulder.

58

relative energy

fragment species (anion + neutral) D-tryptophan [TRP−H]− + H· (from NH ring) [TRP−H]− + H· (from NH2) [TRP−NH2]− + NH2· [TRP−HC(O)OH]− + HC(O)OH [TRP−H2−CO2]− + CO2 + H2 [TRP−NH3−COOH]− + NH3 + COOH· [TRP−NH2−HCOOH]− + NH2· + HC(O) OH CH(NH2)COOH− + [TRP−CH(NH2) COOH]· NC−COOH− + 3-methylindole + H2 CN− + C6H5CHCH2CH(NH2)COOH· CN− + 3-methylindole + H2 + COOH· CN− + 3-methylindole + HCOOH + H· HCCH− + 3-COOH-indole + NH3 OH− + [TRP−OH]· O− + [TRP−O] NH2− + [TRP−NH2]· melatonin [MLT−H]− + H· (from NH ring) [MLT−H]− + H· (from NH amidic) [MLT−CH3]− + CH3· (from OCH3) [MLT−CH3]− + CH3· (from CO) [MLT−CH4]− + CH4 [MLT−NH2]− + NH2· [MLT − O]− + ·O· (from OCH3) [MLT−NO]− + NO· [MLT−C(O)CH3]− + C(O)CH3· [MLT−HNC(O)CH3−H2]− + HNC(O)CH3· + H2 [MLT−CH3CH2N(H)C(O)CH3]− + CH3CH2N(H)C(O)CH3 [MLT−N(H)C(O)CH3−OCH2]− + N(H) C(O)CH3· + CH2O HNC(O)CH3− + [MLT−HNC(O)CH3]·

(1) (1) (2) (2)

1.48 3.13 1.66 1.43 1.51 3.99 4.26

(1.09) (2.67) (1.28) (1.16) (0.90) (3.52) (3.73)

1.20 (0.93)

(3)

(4)

(5) (6) (7)

(8)

1.07 1.15 2.65 2.84 3.59 2.97 3.50 2.74

(0.47) (0.89) (1.94) (2.22) (3.21) (2.70) (3.34) (2.44)

1.69 2.07 0.69 2.98 0.91 2.43 4.47 4.62 3.40 5.13

(1.29) (1.68) (0.34) (2.61) (0.62) (2.06) (4.29) (4.27) (3.07) (4.41)

3.03 (2.80) 5.71 (5.16) 1.65 (1.36)

a

The values in parentheses include zero-point vibrational energy corrections. Italic numbers refer to the structures reported in Chart 4.

17) is observed as a broad signal centered at 5.0−5.4 eV, i.e., deriving from dissociation of core-excited states. Because the observed anion currents in TRP are relatively small even at the highest evaporation temperatures we ascribe the m/e = 17, 26, 74 peaks at zero energy and the m/e = 142 signal below 2 eV to artifacts originating from trace impurities. The assignment of the TRP signals is in line with the calculated threshold energies reported in Table 2. In contrast with a previous study91 the m/e = 45 anion current is not observed above background. Additionally, the m/e = 71 signal detected around 0.4 eV, in line with the calculated energy threshold, is ascribed to dissociation of the first resonance to form the NCCOOH− negative fragment. This signal, relatively intense in the present investigation, was absent in the previous study cited above.91 A likely explanation for this discrepancy could arise from the possibility that the m/e = 71 fragment is formed with high internal energy so that it dissociates during its flight toward the detection system. Since the flight time through a quadrupole mass filter91 is expected to be significantly higher than that through a magnetic spectrometer, where anions are accelerated by kilovolts,67 the m/e = 71 negative species could decay before detection with formation of CN− or COOH− fragments. The low relative intensity of the m/e = 203 fragment detected by

means of a quadrupole mass analyzer91 could also be attributed to a longer detection time since the dehydrogenated anion can detach the extra electron. Another reason can be associated with the lower electron energy resolution in the present study and an intensity decrease in near zero-energy signals. The dominant decay in the DEA spectrum of MLT, given its similar molecular structure, is the formation of its dehydrogenated species [MLT−H]− (m/e = 231) which is observed at slightly higher electron energy (1.3 eV) compared to TRP, in perfect agreement with the calculated thermodynamic energy thresholds. This energy matches well the predicted positions (1.0 and 1.6 eV) of the first two π* shape resonances (see Table 1), and the observed signal could be associated with unresolved contributions from both states. Calculated thresholds (see Table 3) imply that the most intense [MLT−H]− signal at 1.3 eV is almost totally due to H atom abstraction from the NH group of the indole moiety, whereas cleavage of the amidic N−H bond is predicted to require about 1.7 eV. This decay is energetically possible for assigning two small peaks observed at 4.8 and 8.2 eV in the m/e = 231 current. Although the former could correspond to the π*O MO 3970

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B Chart 4. Structure of the Fragments Formed by DEA to TRPa

a (1) m/e = 158, (2) m/e = 142, (3) and (4) neutral counterparts of m/e = 26 and 16, respectively; MLT: (5) and (6) m/e = 216, (7) m/e = 202, (8) m/e = 145.

involvement of the methoxy group to form the m/e = 144 anion and two neutral fragments, one of them being associated with a rearrangement (one H atom attaches back to the ring) with elimination of an OCH2 fragment from the six-membered cycle. The acetyl group is involved in the generation of two almost complementary fragments observed with relatively high intensity at m/e = 58 and 172 in a very broad spectral feature with maxima in the 8.3−8.8 eV range. The core-excited resonances above 8 eV could decay along these two channels leading to production of the HNC(O)CH3− species (m/e = 58) whereas the complementary fragment without H2 gives rise to formation of the m/e = 172 negative species (see Table 2). Although the thermodynamic energy threshold is much higher in the latter case (4.41 eV; Table 3), the observed signals are accounted for by the present calculations. Finally, rearrangement of the temporary anions to structure 3 (Chart 4) allows for loss of neutral nitrogen monoxide slightly above 4 eV to form the m/e = 202 anion, but the observed signal peaks at much higher energy (>8 eV). 3.3. DEA Properties and Biochemistry of Melatonin. Despite a vast amount of observations concerning the beneficial physiological effects of MLT in plants and mammalians the molecular mechanisms of its activity are not completely understood.19,22,96 The most important aspect of the data presented here is the predominance of [MLT−H]− formation in the MLT DEAS spectra, confirming facile proton removal from MLT, as is also observed in solution under UV irradiation.97 The ease of this proton removal at very low energies undoubtedly plays a major role in the efficient scavenging of hydroxyl radicals by MLT in solution as shown by pulse radiolysis98 and confirmed using DFT calculations99 at nearly diffusion controlled rates (about 1010 M−1 s−1) although hydroxyl radicals do tend to react instantly with almost anything in their vicinity.5 According to the present MLT DEA results, production of the [MLT−H]− species occurs through H abstraction from the NH group of the pentacyclic ring. In fact, donation of a hydrogen atom by an antioxidant molecule can be considered as the first step in its radical-scavenging activity as occurs, for example, in polyphenolic antioxidants.100 Given also that in many cases the toxic activity of a molecular species depends upon its ability to produce a free radical by dissociative electron attachment together with the ability of that free radical to abstract a hydrogen atom from cellular material,45 the role of MLT with its ability to produce copious

predicted to lie at 5.35 eV, the latter more likely derives from core-excited states. Formation of [MLT−H]− is associated with σ bond cleavage and can occur either via direct formation of a σ* anion state92 (although distinct σ* resonances in the 0−5 eV range in the ET spectrum of pyrrole were not observed93) or through π*/σ* mixing caused by out-of-plane vibrations which overcome symmetry restrictions, in line with the interpretation given for DEA to the reference compound pyrrole.94 In contrast to tryptophans [MLT−H]− anions are the only fragment species observed at incident electron energies below 4 eV with the exception of the much lower intensity m/e = 217 anion current, due to loss of the methyl group from the OCH3 substituent (see Figure 3). In fact, according to the calculated thresholds (see Table 3) the m/e= 217 current observed below 2 eV can only be ascribed to CH3 loss from the methoxy group, the threshold for elimination from the acetyl group being >2.5 eV. In fact, cleavage of the alkyl-oxygen bond in the anisole anion leading to demethylation is known to occur readily in most solvents.95 Abstraction of one additional hydrogen atom from the neighboring ring position (structure 5 in Chart 4) leads to generation of the m/e = 216 fragment, the energetically most favored (calculated energy threshold = 0.62 eV) when formed with a methane molecule as a neutral counterpart. Through the rupture of a single bond, the other likely pathway to form the m/e = 216 negative ion, i.e., elimination of a neutral NH2 fragment, is predicted to require much more energy (>2 eV; Table 3), and formation of a rearranged structure of the [MLT−O]− anion is calculated to be possible only above 4 eV. An intense m/e = 216 signal is indeed observed at incident electron energies >6 eV. Elimination of a neutral acetyl group from the substituent chain in MLT could be responsible for formation of the m/e = 189 species observed at 5.2 eV matching well with the position of the shape resonance predicted at 5.35 eV (see Table 1) although the calculated threshold is much lower (3.10 eV). Formation of the m/e = 145 negative ions would correspond to loss of the whole substituent chain along with two additional hydrogen atoms from the indole moiety, so that a triple bond is formed (structure 5 in Chart 4) and a closed-shell Nacetylethylamine molecule is produced as a neutral counterpart. This dissociative decay is observed as a broad peak at electron energies above 9.5 eV, although the thermodynamic threshold is predicted to be only about 3 eV. A signal very similar in peak shape but twice as intense could derive from a dissociation with 3971

DOI: 10.1021/acs.jpcb.7b01408 J. Phys. Chem. B 2017, 121, 3965−3974

Article

The Journal of Physical Chemistry B ORCID

amounts of hydrogen atoms through interaction with low energy electrons ceded, for example, by enzymes in the biological medium, helps to explain its positive role against damaging free-radical mechanisms in that medium. It is also known that negative ion states produced by electron attachment to isolated molecules can occur at about 1 eV lower energies in condensed media45 and thus their formation may even become exoergic in biological systems. This would again confirm a positive role of MLT against cellular damage in biological systems, given that the present data show a negative gas-phase electron affinity (the B3LYP/6-31+G(d) value is −0.49 eV), and the LUMO of MLT is predicted to lie at 1 eV in agreement with an earlier estimation.101 The estimated electron affinity of MLT is not large enough to compete with cellular oxygen for quasi-free electrons.102 However, should electron transfer occur from an excited state of MLT (as in the case of plant tissues under natural UVirradiation) the efficiency of [MLT−H]− formation by DEArelated mechanisms should increase. We therefore propose that under these conditions MLT can be “activated” by electron attachment into the LUMO, followed by H atom donation, the whole concerted process being responsible for MLT action as an antistress molecule in plants.22 It should also be mentioned that the extremely low toxicity of MLT103 is in accordance with the present data and with conclusions reported elsewhere,45 namely DEA to MLT at thermal energy of incident electrons is not associated with the generation of active radical species, only H release from the pentacycle NH group being observed. As a final comment, melatonin acts as a protector from oxidative stress conditions associated with reactive oxygen species, which in turn are associated with increasing electron leakage from the ETC. In this case the probability of interaction of these electrons with neighboring MLT molecules in mitochondria is also increased, thus promoting dehydrogenation of cellular melatonin through a DEA-related pathway.

Stanislav A. Pshenichnyuk: 0000-0001-5318-3638 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors are grateful to the Russian Foundation for Basic Research (grants #15-29-05786 and #15-02-02809), and the Italian Ministero dell’Istruzione, dell’Università e della Ricerca for financial support.



(1) McCord, C. P.; Allen, F. P. Evidences Associating Pineal Gland Function with Alterations in Pigmentation. J. Exp. Zool. 1917, 23, 207−224. (2) Lerner, A. B.; Case, J. D.; Takahashi, Y.; Lee, T. H.; Mori, W. Isolation of Melatonin, The Pineal Factor that Lightens Melanocytes. J. Am. Chem. Soc. 1958, 80, 2587−2592. (3) Lerner, A. B.; Case, J. D.; Heinzelman, R. V. Structure of Melatonin. J. Am. Chem. Soc. 1959, 81, 6084−6085. (4) Lerner, A. B.; Case, J. D.; Takahashi, Y. Isolation of Melatonin and 5-Methoxyindole-3-acetic Acid from Bovine Pineal Glands. J. Biol. Chem. 1960, 235, 1992−1997. (5) Allegra, M.; Reiter, R. J.; Tan, D. X.; Gentile, C.; Tesoriere, L.; Livrea, M. A. The Chemistry of Melatonin’s Interaction with Reactive Species. J. Pineal Res. 2003, 34, 1−10. (6) Leon, J.; Acuña-Castroviejo, D.; Sainz, R. M.; Mayo, J. C.; Tan, D. X.; Reiter, R. J. Melatonin and Mitochondrial Function. Life Sci. 2004, 75, 765−790. (7) Hardeland, R.; Cardinali, D. P.; Srinivasan, V.; Spence, D. W.; Brown, G. M.; Pandi-Perumal, S. R. Melatonin − A Pleiotropic, Orchestrating Regulator Molecule. Prog. Neurobiol. 2011, 93, 350− 384. (8) Acuna-Castroviejo, D.; Escames, G.; Venegas, C.; Diaz-Casado, M. E.; Lima-Cabello, E.; Lopez, L. C.; Rosales-Corral, S.; Tan, D. X.; Reiter, R. J. Extrapineal Melatonin: Sources, Regulation, and Potential Functions. Cell. Mol. Life Sci. 2014, 71, 2997−3025. (9) Pandi-Perumal, S. R.; Trakht, I.; Srinivasan, V.; Spence, D. W.; Maestroni, G. J. M.; Zisapel, N.; Cardinali, D. P. Physiological Effects of Melatonin: Role of Melatonin Receptors and Signal Transduction Pathways. Prog. Neurobiol. 2008, 85, 335−353. (10) Acuna-Castroviejo, D.; Lowenstein, P. R.; Rosenstein, R.; Cardinali, D. P. Diurnal Variations of Benzodiapine Binding in Rat Cerebral Cortex: Disruption by Pinealectomy. J. Pineal Res. 1986, 3, 101−109. (11) Reiter, R. J. The melatonin Rhythm: Both a Clock and a Calendar. Experientia 1993, 49, 654−664. (12) Peyrot, F.; Ducrocq, C. Potential Role of Tryptophan Derivatives in Stress Responses Characterized by the Generation of Reactive Oxygen and Nitrogen Species. J. Pineal Res. 2008, 45, 235− 246. (13) Pérez-González, A.; Alvarez-Idaboy, J. R.; Galano, A. Freeradical Scavenging by Tryptophan and Its Metabolites Through Electron Transfer Based Processes. J. Mol. Model. 2015, 21, 213/1−11. (14) Tan, D. X.; Manchester, L. C.; Esteban-Zubero, E.; Zhou, Z.; Reiter, R. J. Melatonin as a Potent and Inducible Endogenous Antioxidant: Synthesis and Metabolism. Molecules 2015, 20, 18886− 18906. (15) Reiter, R. J.; Tan, D. X.; Zhou, Z.; Cruz, M. H. C.; FuentesBroto, L.; Galano, A. Phytomelatonin: Assisting Plants to Survive and Thrive. Molecules 2015, 20, 7396−7437. (16) Hardeland, R. Melatonin and the Theories of Aging: A Critical Appraisal of Melatonin’s Role in Antiaging Mechanisms. J. Pineal Res. 2013, 55, 325−356. (17) Acuna-Castroviejo, D.; Martın, M.; Macias, M.; Escames, G.; Leon, G.; Khaldy, H.; Reiter, R. J. Melatonin, Mitochondria, and Cellular Bioenergetics. J. Pineal Res. 2001, 30, 65−74.

4. CONCLUSIONS Interaction of biologically active melatonin and tryptophan molecules with low-energy (0−14 eV) electrons and their fragmentation via resonance dissociative electron attachment was studied under gas-phase conditions using DEA spectroscopy. The experimental findings were assigned on the basis of quantum-chemical calculations of localization properties and energies of empty molecular orbitals. The present calculations predict a negative value of adiabatic electron affinity for melatonin and tryptophan, in agreement with the absence of their molecular negative ions in the DEA spectra. Due to the relatively low value of its negative electron affinity (