Introduction

The combination of electrospray ionization mass spectrometry (ESI-MS) with collision-induced dissociation (CID) is commonly used to study the properties of gas-phase ions and mechanisms of gas-phase reactions [1,2,3,4,5,6]. However, the interpretation of CID spectra is not always straightforward due to various rearrangements that occur during fragmentation [7,8,9,10,11,12,13,14]. The reported rearrangements in the gas-phase included benzyl cation transfer, sulfonyl cation transfer, and halogen transfer, and have been a subject of many mechanistic studies [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22].

Sulfur transfer reactions have been increasingly used in organic synthesis [23,24,25,26,27]. Among this class of rearrangement reactions, 1,2-sulfur transfer has been widely investigated and extensively used in the synthesis of heterocycles [23, 28, 29] and in carbohydrate chemistry [30, 31]. The most common pathway of 1,2-sulfur transfer proceeds through a key thiiranium intermediate, which can either undergo elimination to give allyl thioethers or substitution to generate formally transposed substitution products [25]. Other types of sulfur transfer, such as 1,3-sulfur transfer [24] and 1,4-sulfur transfer [32], are rarely described. To the best our knowledge, however, no report has been found on the gas-phase intramolecular sulfur transfer reaction, which deserves further investigation.

Phosphorothioates bearing an S=P bond display important chemical and biological properties that afford utility in various fields, including organic synthesis, medicinal chemistry, molecular biology, and agrochemistry [33, 34]. Previous studies mainly focused on the determination and quantification of phosphorothioates, while few studies focused on their gas-phase fragmentation [35,36,37,38,39]. As an example, a thiono-thiolo rearrangement (Newman-Kwart rearrangement), where the S=P–OR group rearranges to the O=P–SR group via R- transfer, can occur under electron ionization or tandem MS [36, 37]. A curious R-PhS+ type of ion was observed in the fragmentation of protonated fenthion [36, 40, 41], but no detailed mechanism has been documented to our knowledge. In this work, an intriguing rearrangement reaction to the formation of R-PhS+ ion via sulfur transfer has been investigated in the ESI-MS analysis of phosphorothioates. The mechanism of this reaction was examined in detail by a combination of experimental and theoretical calculation approaches.

Experimental Section

Chemicals and Material

Methanol HPLC grade was purchased from Sigma-Aldrich (St. Louis, MO, USA). Fenthion, parathion-methyl (compound 2), fenitrothion (compound 3), tolclofos-methyl (compound 4), and chlorpyrifos-methyl (compound 5) were purchased from J&K Scientific Ltd. (Shanghai, China) with a purity >99%.

Mass Spectrometry

The samples were analyzed on an LTQ-XL advantage IT-MS (Theromo Scientific, San Jose, CA, USA) and an Orbitrap-XL mass spectrometer (Theromo Scientific, San Jose, CA, USA) using a home-made ESI interface in the positive ion mode. Every diluted solution (1 μg mL−1 in methanol) was infused into the source chamber at a flow rate of 3 μL min−1. The optimized ESI source conditions were as follows: the ion-spray voltage, 3 kV; the nebulizing gas (N2), 25 arbitrary units (a.u.); the capillary temperature, 150 °C, the capillary voltage in 80 V; the tube lens in 100 V. Other parameters were automatically optimized by the system. The ion trap pressure of approximately 1 × 10−5 Torr was maintained with a Turbo pump and pure helium (99.99%) was used as the collision gas. The instrument was operated at a high resolution up to 100,000. The CID-MS experiments were performed by using an excitation AC voltage to the end caps of the ion trap to include collisions of the isolated ions (isolation width at 1 m/z) for a period of 30 ms and variable excitation amplitudes. The CID-MS spectra of the protonated molecules were obtained by activation of the precursor ions at the normalized collision energy of 15%~30%.

Theoretical Calculations

Theoretical calculations were performed using the Gaussian 09 program [42]. The geometries of reactants, transition states, intermediates, and products were optimized using the density functional theory (DFT) method at the B3LYP/6-31+G(d,p) level. All reactants, intermediates, and products were identified as true minima in energy by the absence of imaginary frequencies. Transition state (TS) structures were obtained through relaxed PES scans utilizing DFT method at the B3LYP/6-31+G(d,p) level to generate initial structures for the TSs, in which a bond length was scanned to find a first-order saddle point, and subsequently optimizing the corresponding transition state. Then, the relevant TS structures are searched and optimized using either TS or QST2 procedures. QST2 uses a quadratic synchronous transit approach to get closer to the quadratic region of the TS and then uses a quasi-Newton algorithm to complete the optimization. All transition states were confirmed by the presence of a single imaginary vibrational frequency and the reasonable vibrational mode. Intrinsic reaction coordinate (IRC) calculations at the same level of theory were performed on each transition state to further confirm that the optimized TS structures were actually connected to the correct reactants and products by a steepest descend path. Vibrational frequencies of all the key species were calculated at the same level of theory. Full structural details and energies of all structures involved are available in the supplementary material. The energies discussed here are the sum of electronic and thermal free energy.

Result and Discussion

Fragmentation Behavior of Protonated Fenthion

The gas phase sulfur transfer rearrangement reaction was explored by investigating the MS fragmentation behavior of protonated fenthion derivatives. Fenthion (compound 1) was selected as a model to perform a detailed investigation. The tandem mass spectrum of protonated fenthion (the mass-isolated m/z 279) shown in Figure 1 reveals the formation of a dominant fragment ion at m/z 247, corresponding to methoxyl O-(3-methyl-4-methylsulfanyl-phenyl) phosphonium thioate, a3 (Supplementary Material Scheme S1) via a neutral loss of 32 Da (methanol). Three minor fragment ions are observed at m/z 231 (b2), m/z 169 (a8), and m/z 137 (b1), corresponding to the neutral losses of 48 Da, 110 Da, and 142 Da, respectively (Figure 1(a)). The neutral loss of 48 Da likely arises from elimination of methanethiol. The fragment ion at m/z 137 is assigned as 3-methyl-4-methylsulfanyl-benzene cation (Scheme S1), which can be attributed to the loss of O,O′-dimethyl thiophosphate from the precursor ion at m/z 279. The fragment ion at m/z 169 is attributed to 3-methyl-4-methylsulfanyl-benzenesulfenylium cation (Scheme S1) or its isomer, resulted from the elimination of C2H7O3P of the precursor ion, which will be discussed in detailed in the following sections. The elemental compositions of these products were confirmed by accurate mass measurements performed on a high-resolution Orbitrap-XL mass spectrometer (Supplementary Material Figure S1 and Table S1).

Figure 1
figure 1

Collision-induced dissociation mass spectra of [1 + H]+ at the normalized collision energy of 16%. (a) MS2 spectrum of [1 + H]+ (m/z 279 →), (b) MS3 spectrum of [1 + H]+ (m/z 279 → m/z 247 →), (c) MS2 spectrum of [1-34S + H]+ (m/z 281 →), (d) MS3 spectrum of [1-34S + H]+ (m/z 281 → m/z 249 →)

Full size image

Fragmentation Pathways to R-benzenesulfenylium Cation

The characteristic fragment ion at m/z 169 can only be interpreted as a result of the C2H7O3P elimination via sulfur transfer. To interpret the structure of the ion at m/z 169, the MS3 experiments were performed. As shown in Figure 1(b), the MS3 spectrum of protonated fenthion (m/z 279 → m/z 247 →) shows a base peak ion at m/z 169 via successive eliminations of CH3OH (32 Da) and CH3OP=O (78 Da). However, there is no relevant moiety of CH3OP=O in the structure of a3. Thus, the generation of the ion at m/z 169 originated from dissociation of a3 (m/z 247) via skeletal rearrangement.

Two potential pathways for the generation of m/z 169 from a3 were proposed in Scheme 1. In route 1, a direct transfer of 1-methyl-2-methylsulfanyl-benzene group to S1 atom leads to an intermediate a4 (methoxyl S-(3-methyl-4-methylsulfanyl-phenyl) phosphonium thiolate), which undergoes the subsequent dissociation to form a8 at m/z 169. An analogous intramolecular O- to S-benzene migration has also been reported in the gas phase fragmentation diphenyl phosphorochloridothioate using electron impact MS by Cooks [37]. In route 2, the ortho-carbon atom of the phenyl ring firstly undergoes a nucleophilic attack on the positively charged P=S group, which leads to a bicyclic intermediate a6. Then, the S1 atom in a a6 undergo a 1,2-migration to form a spiro intermediate a7, which subsequently undergoes the dissociation to give a8 by lose of CH3O–P=O (route 2-A), or the H atom in C7 undergo a 1,2-migration to form a bicyclic intermediate a9, which subsequently undergoes the dissociation to generate a10 by elimination of CH3O–P=O (route 2-B).

Scheme 1
scheme 1

Proposed pathways for the generation of the ions at m/z 247 and m/z 169

Full size image

The potential pathways in Scheme 1 lead to the product ion at m/z 169 with the structure 3-methyl-4-methylsulfanyl-benzenesulfenylium cation or 2-methyl-3-methylsulfanyl-benzenesulfenylium cation. Benzene-sulfenylium cations have been generated in high abundance by ionization of different precursors containing a thiophenyl group and their gas-phase reactivity has also been reported [14, 43]. Two approaches have been proposed to prepare sulfonium cations. One is unimolecular sulfur-heteroatom bond fission of “cationoid” complexes or “carriers” of sulfonated compounds, a process usually attempted in the presence of strong Lewis acids [44]. The other approach involves the single-electron oxidation of disulfides [45]. The arylthio group (ArS) is of intrinsic interest and has long been incorporated into drug molecules and peptides, which can exhibit highly activities such as antiplasmodial activity, and antiviral activity [46,47,48]. Thus, it is of considerable interest to investigate the mechanistic formation of the benzenesulfenylium cation (m/z 169).

Native 34S Isotope and Isotope Labeling Experiments

The postulated decomposition reactions in Scheme 1 were confirmed by the MS/MS analysis on the native 34S isotopic ion (Figure 1(c), (d)) [49]. The sulfur element has two isotopes, 32S and 34S in nature, with the relative abundance at 100% and 4.4%, respectively. As shown in Scheme S1, decomposition of the mono isotope ion of a1 (MH+) at m/z 279 produces the fragment ion b2 at m/z 231 by lose of CH3SH. The first 34S isotope ion at m/z 281, however, contains two isomeric structures (a1-I1 and a1-I2 in Scheme S2), due to the different position of 34S. As expected, fragmentation of isomer a1-I1 gives the product ion b2-I1 at m/z 233, through the neutral loss of CH332SH; whereas dissociation of isomer a1-I2 results in the product ion b2-I2 at m/z 231 via the neutral loss of CH334SH. The almost identical abundance of the two product ions is attributed to the equal distribution of the 34S atom in nature. Interestingly, elimination of S=P(OH)(OCH3)2 from the isotopic ion at m/z 281 shows similar fragmentation behavior, with nearly equivalent abundance of the isotopic fragment ions at m/z 137 and m/z 139. The fragment b2-I1 (m/z 137) is generated by the dissociation of isomer a1-I1 through lose of 34S=P(OH)(OCH3)2, while the product ion b2-I2 (m/z 139) is formed via S=P(OH)(OCH3)2 elimination from a1-I2. The product ion of a3m/z 247 (or a8/a10 at m/z 169), however, has two sulfur atoms in the chemical formula, and thus both appear as the mono isotopic ion peak with an increasing mass shift of 2 Da in the CID spectrum. Similarly, as shown in Figure 1(d), an increasing mass shift of 2 Da was observed for a8/a10 (from m/z 169 to m/z 171) in the MS3 spectrum of protonated 34S isotopologue (m/z 281 → m/z 249 →).

The proposed dissociation pathways in Scheme 1 were also supported by the CID-MS analysis of the deuterium-labeling ion (Figure 2). As shown in Scheme 1, there is no external proton in the product ion a3, and thereby dissociation of [1 + H]+ and [1 + D]+ theoretically resulted in a3 with the same mass (247 Da). However, both the ions at m/z 248 and m/z 247 were observed in the CID mass spectrum of [1 + D]+, corresponding to the loss of CH3OH and CH3OD, respectively. The existence of the ion at m/z 248 implies that an H/D exchange in the fragmentation process, e.g. exchange of the external deuteron to the ortho-positions of phenyl ring via a six-membered ring. After the deuteron transfers to phenyl ring, the proton or deuteron may migrate back to the methoxyl oxygen competitively, which results in subsequent losses of CH3OH and CH3OD, respectively. The subsequent transfer of a proton or a deuteron on a1-1 via TS2 for the reaction to proceed was marked by a considerable kinetic isotope effect, kH/kD. The intensity ratio of elimination of CH3OH (m/z 248) to elimination of CH3OD (m/z 247) is about 5:1, which represents the majority of their kH/kD value. Our results are consistent with reports of kinetic isotope effect in the range of kH/kD = 5 during the interannular proton transfer steps of benzylbenzenium ions and 1,4-diphenyl-but-2-yne ions [50].

Figure 2
figure 2

Collision-induced dissociation mass spectra of [1 + D]+. (a) MS2 spectrum of [1 + D]+ (m/z 280 →), (b) MS3 spectrum of [1 + D]+ (m/z 280 → m/z 248 →)

Full size image

Density Functional Theory Calculations

To further investigate the mechanisms associated with the sulfur and benzene migration reactions, density functional theory (DFT) calculations were carried out at the B3LYP/6-31+G(d,p) level of theory. A lone pair of electrons of a heteroatom is much easier to capture proton [35]. Thus, there are multiple potential protonation sites for fenthion, including S1 of the thiophosphoryl, O3/O4 of the methoxy, O5 of the phenoxy, and S6 of the methyl sulfide (Table 1). The structures with different protonation sites of fenthion were optimized at the same level (B3LYP/6-31+G(d,p)) and the relative energies of these structures are summarized in Table 1. Overall, the calculation results indicates that S1 atom is the most thermodynamically favorable protonation site, which is 41.3 kJ/mol, 86.4 kJ/mol, 79.7 kJ/mol, and 45.0 kJ/mol lower than protonation site on S6 atom, O3/O4 atom, O5 atom, and C7 atom, respectively. It has been extensively accepted that the ionizing proton can transfer to other less favored sites during the subsequent fragmentation process [17, 19].

Table 1 Relative energies of [1 + H]+ ions with different protonation sites
Full size table

Figure 3 shows a schematic potential energy surface plot for the generation of a3 (m/z 247). Firstly, the external proton in a1 undergoes a 1,5-migration to the ortho-carbon atom of the phenyl ring via a six-membered ring transition state (TS1) to afford an isomer a1-1. This process only needs to surmount an energy barrier of 59.9 kJ mol−1. Then, the activated proton at the ortho-carbon atom of the phenyl ring undergoes 1,5-migration back to the oxygen atom (O3/O4) of the methoxy group to afford an isomer a2 via a six-membered ring transition state (TS2), which surmounts an energy barrier of 111.0 kJ mol−1. The oxygen atom of the methoxy group in a2 is charged trivalent oxygen atom. When a2 ion is formed, theoretical calculations indicates that breakage of the P2–O3/O4 bond occurs spontaneously, and the two formed fragments are still held together electrostatically as a stable ion-neutral complex (INC) [a3/methanol] with a stabilization energy of 49.2 kJ mol−1. A direct decomposition of [a3/methanol] results in the formation of a3.

Figure 3
figure 3

Potential energy diagram for the generation of m/z 247

Full size image

The two potential routes to the ion at m/z 169 in the subsequent fragmentations of a3 were compared by theoretical calculations (Figure 4), and details of the corresponding structures are available in the Supplementary Material. In route 1, the phenyl ring in a3 is transferred from O5 to S1 through a four-membered ring transition state (TS3), leading to the formation of a4 ion with a new carbon-sulfur bond. This process needs to surmount an energy barrier of 168.3 kJ mol−1. The rearrangement occurs with a concerted process with cleavage of C–O bond and formation of C–S bond, which can be viewed as an electrophilic substitution of the phenyl ring (Figure 5). The conversion of O-aryl carbamothioates to S-aryl carbamothioates in the solution-phase is called Newman-Kwart rearrangement, which is an efficient method for the straightforward preparation of thiophenol from the corresponding phenols [51]. Migration of phenyl from oxygen to sulfur has also been observed in spectra of sulphonyl derivatives and dimethylthiocarbamates in the gas phase [52]. The free energy of a4 ion is 8.8 kJ mol−1 lower than that of a3 ion. Then, the formed a4 continues to undergo the cleavage of the P–S bond induced by the positive charge in P2 atom, and gives rise to an INC intermediate a5 [3-methyl-4-methylsulfanyl-benzenesulfenylium cation/metaphosphorous acid methyl ester] with a small energy barrier of 15.2 kJ mol−1 (TS4). The free energy of intermediate a5 is 20.2 kJ mol−1 lower than that of a3 ion. The sum free energy of the separated ion a8 and metaphosphorous acid methyl ester is higher than that of a5 by 46.0 kJ mol−1; this indicates that a5 seems relatively stable from the view of stabilization energy.

Figure 4
figure 4

Potential energy diagram for the generation of m/z 169. (red line: route 1, black line: route 2-A, green line: route 2-B, d: metaphosphorous acid methyl ester)

Full size image
Figure 5
figure 5

The optimized structures of key intermediates. The bond lengths are given in Å

Full size image

In route 2, the ortho-carbon atom of the phenyl ring firstly undergoes a nucleophilic attack on the positively charged P=S group via a five-membered ring transition state (TS5 in Figure 5), and affords a bicyclic intermediate a6. This process only needs to surmount the energy barrier of 88.1 kJ mol−1. As shown in Figure 5, the length of P–S bond is increased to 2.055 Å in TS5, which appreciably longer than a P=S bond (1.887 Å) and shorter than a covalent P–S bond (2.209 Å). The reason for the elongation of P=S bond in a3 is the formation of five-membered ring by nucleophilic attack of the ring double bond on the positively charged P=S group. The free energy of a6 ion is 18.0 kJ mol−1 lower than that of a3 ion, indicating a more stable structure.

Interestingly, a subsequent 1,2-sulfur transfer in a6 occurs through a three-membered ring transition state (TS6) to afford a7, with a small energy barrier of 59.9 kJ mol−1 (in route 2-A). This sulfur scrambling process is similar to proton scrambling on the phenyl ring [20]. The lengths of two C–S bonds involving sulfur scrambling in TS6 are 2.169 Å and 2.063 Å, respectively. Both are longer than that of a covalent C–S bond (1.738 Å, Figure 5). The four-membered ring structure of a7 seems less stable than the five-membered ring structure of a6. Thus, the ion a7 subsequently undergoes the loss of metaphosphorous acid methyl ester via simultaneous cleavage of the P–S bond and C–O bond (Figure 5). This step is the key step in route 2-A, which surmounts an energy barrier of 113.3 kJ mol−1 (TS7). The rearrangement mechanism in route 2-A can be viewed as a successive stepwise of 1,4-sulfur transfer and 1,2-sulfur transfer, then forming a stable 3-methyl-4-methylsulfanyl-benzenesulfenylium cation (a8) via open-ring reaction accompanied by the cleavage of the P–S bond and C–O bond.

Alternatively, a subsequent 1,2-proton transfer in a6 occurs through a three-membered ring transition state (TS8) to afford a9, with a small energy barrier of 55.6 kJ mol−1 (in route 2-B). Then, the ion a9 subsequently undergoes the elimination of metaphosphorous acid methyl ester via direct cleavage of the P–S bond and C–O bond (TS9). This step is the key step in route 2-B, which surmounts an energy barrier of 98.0 kJ mol−1 (TS9). Thus, the rearrangement mechanism of route 2-B can be viewed as a successive stepwise of 1,4-sulfur transfer and 1,2-proton transfer, then forming a stable 2-methyl-3-methylsulfanyl-benzenesulfenylium cation (a10) via open-ring reaction accompanied by the cleavage of the P–S bond and C–O bond.

The key energy barriers of route 2-A (113.3 kJ mol−1) and route 2-B (98.0 kJ mol−1) are much lower than that of route 1 (phenyl migration, 168.3 kJ mol−1), indicating that sulfur migration is a kinetically more favored process than phenyl migration in the formation of ion at a8 (m/z 169). For route 2-A and route 2-B, the minimum activation energy for the cleavage of the P–S bond and C–O bond process viaTS7 and TS9 is 73.0 kJ mol−1 and 134.4 kJ mol−1, respectively, indicating a kinetically more favorable process of route 2-A. Additionally, the sum free energy of the separated ion a8 and metaphosphorous acid methyl ester is 72.5 kJ mol−1 less than that of the separated ion a10 and metaphosphorous acid methyl ester; this indicates that formation of a8 is a thermodynamically favored process. Thus, the formation of a8 in route 2-A is the major path.

The energy requirements for the formation of a8 and metaphosphorous acid methyl ester are 25.8 kJ mol−1 higher than that of a3. However, the free energy of a7-1 is 50.2 kJ mol−1 lower than that of the separated ion a8 and metaphosphorous acid methyl ester; this indicates that a7-1 seems relatively stable from the view of stabilization energy. In addition, the minimum internal excess energy of a7-1 is 137.7 kJ mol−1 or at least 87.5 kJ mol−1 above the separation energy. Thus, direct separation of a7-1 easily occurs in terms of energy, which generates an abundance of ion at m/z 169 (a8).

The Universality of the Gas-Phase Sulfur Transfer

To better delineate the universality of this gas-phase sulfur migration reaction, NO2-substituted (compounds 2 and 3) and Cl-substituted (compounds 4 and 5) derivatives of fenthion were also investigated by tandem MS experiments, and the tandem MS data (Supplementary Material Figures S2, S3, and S4) were summarized in Table 2. All of these compounds show similar fragmentation behaviors in the MS/MS experiments. Noteworthily, the intensive fragment ions of the corresponding a8 (m/z 154, m/z 168, m/z 191, and m/z 212 for compounds 2, 3, 4, and 5, respectively) were observed for all compounds.

Table 2 The collision-induced dissociation mass spectra data of [M + H – CH3OH]+ from protonated fenthion derivatives (1–5)
Full size table

To further investigate the electronic effects of substituents of the sulfur transfer reaction, compounds 2 and 4 were selected as models for comparison of their CID-MS behaviors. The two potential routes (sulfur transfer versus phenyl ring transfer) to the ions at m/z 154 and m/z 191 in the subsequent fragmentations of corresponding ion a3 (m/z 232 and m/z 269) were compared by theoretical calculations (Supplementary Material Figures S5 and S6). According to our calculation results, it could be found that the variation trends of key steps on the potential energy surface diagrams for the generation of m/z 154 and m/z 191 are in accordance with those of m/z 169. As shown in Figures S5 and S6, the key steps in route 1 (phenyl migration) and route 2 (sulfur migration) are TS3 and TS5, respectively. For the formation of ions at m/z 154 and m/z 191, the energy barriers of TS3 and TS5 follows the order: 193.4 kJ mol−1 (TS3 in route 1) > 155.4 kJ mol−1 (TS5 in route 2) and 214.5 kJ mol−1 (TS3 in route 1) > 166.0 kJ mol−1 (TS5 in route 2), respectively, indicating that sulfur migration is a dynamically more favored process in formation of ions at m/z 154 and m/z 191. These experimental and calculation results of fenthion derivatives indicate the universality and the facility of the sulfur transfer reaction in the dissociation process.

Interestingly, for compounds 2 and 3 with a nitro group on the benzene ring, the intensities of the corresponding a8 product declined significantly in the CID-MS, indicating the presence of a nitro group, will inhibit the process of sulfur transfer. Thus, the presence of an NO2 substituent on the phenyl ring inhibits the sulfur transfer reaction, whereas an -SCH3 substituent on the phenyl ring promotes this reaction pathway.

Conclusion

In summary, protonated fenthion derivatives firstly dissociates via the elimination of CH3OH to generate the predominant fragment ion a2 (R-phenoxyl, O-methyl phosphonium thioate) upon collisional activation. Then, a2 further dissociates via the loss of CH3O–P=O to form the arenesulfenylium cations, R-PhS+. On the basis of the mass spectra data together with isotope labeling experiments and theoretical calculations, an intriguing mechanism via intramolecular stepwise sulfur transfer has been proposed and validated for this fragmentation reaction. Further research is needed to address the gas-phase reactivity of PhS+ and related gas-phase ions.