1 Introduction

ESI-MS/MS has assumed increasing importance as an invaluable tool for the study of the gas phase behaviors of biomolecules [13]. In the past few decades, continuous efforts have been devoted to the investigation of the dissociation reactions of charged compounds derived from protonation [48] or deprotonation [913]. On the other hand, the mass spectrometric fragmentations of metal cationized, especially lithiated biomolecules, have attracted much attention because of the desire to measure intrinsic metal ion affinity and identify the binding sites [14, 15], and because the interesting dissociation pathways exhibited by these compounds are usually different from the protonated and deprotonated analogues [16, 17]. Recently, the fragmentation behaviors of lithiated species have been reported for peptides, fatty acids, phospholipids, polytetrahydrofuran, cardiolipins, thioesters, cholesteryl esters, and disaccharides [1826].

The amide bond forms the fundamental linkage between monomeric amino acid residues in peptide polymers, and it is important to understand its gas-phase chemistry, which has so far been studied mainly through the decomposition reactions of protonated amides and peptides [2731]. The mobile proton model [3237] was proposed and applied to the flexible peptides and proteins besides the rigid molecules to describe the mobility of the proton across the protonated species. In addition, the studies of the gas-phase interactions between Li+ and peptides also have drawn a great deal of attention. Several groups [3843] have examined the gas-phase decomposition reactions of the adduct ions in fast atom bombardment and electrospray ionization tandem mass spectrometry. However, little attention so far has been paid to the fragmentation pattern of the complexes of Li+ with other amide-containing molecules.

One class of amides that are of interest are pesticide precursors, such as 2-(4, 6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide and its derivatives [44, 45]. Here we report our studies of the gas phase chemistry of Li+ adducts of these compounds upon ESI-MS/MS, and among the fragment ions we observe are those reflecting loss of LiOH. Product ions that reflect the Smiles rearrangement, which has been well documented in the solution phase [46, 47] as summarized in Scheme 1, and other fragmentation reactions are also observed. To further examine the elimination of LiOH from Li+ adducts of this class of amide-containing compounds, we have also studied spectra obtained from CID of [M + Li]+ ions of the simpler compounds N-phenylbenzamide and N-phenylcinnamide. The reaction mechanisms involved in formation of product ions were elucidated, and the effects of substituents on the loss of LiOH were characterized. In addition, theoretical calculations based on DFT method were performed to evaluate the proposed mechanisms.

Scheme 1
scheme 1

Classic example of Smiles rearrangement

2 Experimental

2.1 Mass Spectrometry

The samples were first analyzed in positive ion mode on a Varian 500-MS ion trap mass spectrometer equipped with an electrospray ionization (ESI) source, with data acquisition using the Varian MS Workstation (Varian, Palo Alto, CA, USA). The compounds were dissolved in methanol in normal collision-induced dissociation experiment or dissolved in methanol-d4 in the D-labeling experiment, and then lithium chloride solution was added to the sample solutions. The samples were infused into the source chamber at a flow rate of 10 μL min–1 with the following parameters: spray chamber temperature, 50 °C; needle voltage, 5000 V; spray shield voltage, 600 V; capillary voltage, 80 V; rf loading, 80 %; scan mode, standard; drying gas temperature, 300 °C. Nitrogen was used as the drying gas at a pressure of 15 psi and the nebulizing gas at a pressure of 35 psi. Tandem mass spectra were obtained by CID with helium as the collision gas after isolation of the desired precursor ion and the isolation window was set as 1.0 m/z unit. The excitation amplitude (resonance mode) was set to give suitable energy for the dissociation of all compounds.

All accurate mass spectrometric experiments were carried out on an Apex III (7.0 Tesla) FTICR mass spectrometer (Bruker, Billerica, MA, USA). XMASS software version 6.1.1 (Bruker) was used for instrument control, data acquisition, and processing. Sodium trifluoroacetate was used as an external calibration compound. Solutions were infused from the ESI source at 3 μL min–1 with parameters: capillary, –4269 V; end plate, –3803 V; skimmer 1, 12.00 V; skimmer 2, 6.61 V; offset, 0.98 V; rf amplitude, 582.5 Hz; drying gas temperature, 150 °C. Nitrogen was used as the nebulizing and drying gas, and argon was used as the collision gas. MS/MS analysis was performed through isolation of the desired precursor ion using a correlated sweep. Sustained off-resonance irradiation (SORI) was applied for CID. The pulse parameters were set as follows: correlation sweep pulse length, 1000 μs; correlation sweep attenuation, 12 dB; ejection safety belt, 1000 Hz; ion activation pulse length, 250,000 μs; ion activation attenuation, 43 dB; frequency offset from activation mass, 600 Hz; user delay length, 3 s.

2.2 Theoretical Calculations

All theoretical calculations were carried out by using the DFT method (a quantum mechanical modeling method used in physics and chemistry to investigate the electronic structure (principally the ground state) of many-body systems, in particular atoms, molecules, and ions) at the B3LYP/6-31++G(d,p) level of theory in the Gaussian 03 program [48]. The candidate structures of the reactants, products, intermediates, and transition states were optimized by calculating the force constants. No symmetry constrains were imposed in the optimizations. The reaction pathways were traced forward and backward by the intrinsic reaction coordinate (IRC) method. All optimized structures were subjected to vibrational frequency analysis for zero-point energy (ZPE) correction. The sum of electronic and thermal energies of the optimized structures was discussed.

3 Results and Discussion

3.1 Investigation of the Fragmentation of the Lithiated 2-(4, 6-Dimethoxypyrimidin-2-Ylsulfanyl)-N-Phenylbenzamide and Its Derivatives

The structures of 2-(4, 6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide and its derivatives are listed in Table 1. Generally, all the compounds gave similar fragmentation, whereas the nature of the substituents merely affected the relative abundances (RAs) of the product ions.

Table 1 Relative abundances of product ions in the CID spectra of [M + Li]+ ions of 2-(4, 6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide and its derivatives (Excitation amplitude was 1.20 V)

Compound 1 is selected as a model to demonstrate the possible fragmentation pathways. Theoretical calculations indicate that the carbonyl oxygen is the most thermodynamically favored site for Li+ complexation because that molecular ion has the lowest energy (Table S1). The full scan mass spectrum of Compound 1 shows an abundant lithiated molecule at m/z 374. Upon collisional activation, it yields four product ions at m/z 139, 238, 256, and 350, as shown in Figure 1a. The formation of these fragment ions can be rationalized by the reaction mechanism described in Scheme 2. The base peak (ion b at m/z 238) was produced by the Smiles rearrangement [49] reaction (the nucleophilic attack of the amide nitrogen atom to the pyrimidine carbon atom). Ion c is 18 Da more than ion b, and it was confirmed by FTICR-MS/MS that ion c resulted from the solvation [24] of the ion b (Table S2). Ion d at m/z 139, 4,6-dimethoxypyrimidin ion, was produced by the direct cleavage of C–S bond. Another interesting fragmentation pathway involved the elimination of 24 Da, which was confirmed to be LiOH from the precursor ion. In 1990, Larrivee and Allison [50] examined gas phase fragmentation reactions of Li+ adducts of α,ω-bifunctional alkanes and observed that hydroxylated compounds could undergo loss of LiOH, but such losses from Li+ adducts of amides has not yet been reported to our knowledge.

Figure 1
figure 1

The CID mass spectra of [M + Li]+ ions of (a) Compound 1 and (b) deuterated Compound 1

Scheme 2
scheme 2

Proposed fragmentation mechanism of the [M + Li]+ ion of 2-(4, 6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide

Isotope labeling is always powerful in the studies of reaction mechanisms, and D-labeling experiments were carried out. The CID mass spectrum of the lithiated deuterated Compound 1 is shown as Figure 1b. Upon collisional activation after H/D exchange, interestingly, the ions at m/z 350 and 351 were both observed, and the formulas were confirmed to be C19H16O2N3S+ and C19H15O2N3SD+, respectively (Table S2). The existence of ion at m/z 351, which results from the loss of LiOH, implies that the deuteron migrates to the phenyl ring and exchanges with the proton on the phenyl ring. In fact, the proton transfer and H/D exchange processes of protonated aromatics have been studied in the past decades [5162]. After the deuteron transfers from the nitrogen to one of the ortho-positions of the phenyl ring, the proton may migrate back to the nitrogen while the deuteron retains on the phenyl ring [63, 64], which allows the interconversion of f-1 and f-2 via g. The subsequent loss of LiOD from f-1 or LiOH from f-2 results in the formation of the ion at m/z 350 or 351, respectively. The proposed mechanism is shown in Scheme 3.

Scheme 3
scheme 3

Proposed mechanism to explain the losses of LiOD and LiOH from [M + Li]+ ion of deuterated 2-(4,6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide

The fragmentation mechanisms could be further explored by studying the substituent effects on the distribution of the product ions and the spectra of a series of compounds with different substituents at para position of the N-linked phenyl ring were measured. A plot of the abundance ratio of these two ions, ln[(M + Li – C7H4OS)+/(M + Li – LiOH)+] (the ion at m/z 139 is not considered because the substituents are too distant to affect its formation) versus the substituent constants, σ +p , was obtained as shown in Figure 2. The trend clearly indicates that the electron-donating substituents expedited the loss of LiOH, whereas the electron-withdrawing groups retarded the loss. In the reaction of the LiOH loss, a positive charge is developing at the carbon atom connecting the substituted N-linked phenyl ring. An electron-donating substituent (e.g., OCH3 or NMe2) reduces the overall activation barrier of the transition state, and the product ion is stabilized by delocalization of the positive charge at the carbon atom over the substituted phenyl ring system. This favors elimination of LiOH from [M + Li]+ ion relative to Compound 1. In contrast, an electron-withdrawing substituent destabilized the product ion, and the overall activation barrier of the transition state is increased, which militates against elimination of LiOH. In addition, electron-withdrawing moieties have previously been demonstrated to favor the Smiles rearrangement reaction [49]. These two factors cause elimination of LiOH from [M + Li]+ ion to be less favorable for compounds with electron-withdrawing substituents.

Figure 2
figure 2

Plot of ln[(M + Li – C7H4OS)+/(M + Li – LiOH)+] vs the σ +p substituent constants for the collision-induced fragmentation of the lithiated para-monosubstituted 2-(4,6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamides. Excitation amplitude is 1.20 V (helium)

3.2 Investigation of the Fragmentation of the Lithiated N-Phenylbenzamide and Its Derivatives

To further explore the interaction of Li+ with amide-containing biomolecules and the elimination of LiOH from such adduct ions, we examined the fragmentation of simpler compounds, including N-phenylbenzamide and its derivatives. Effects of A- and B-ring substituents on this loss were also investigated. The structures of these compounds and their major product ions from CID of [M + Li]+ ions with their RAs are summarized in Table 2.

Table 2 Relative abundances of product ions in the CID spectra of [M + Li]+ ions of N-phenylbenzamide and its derivatives (excitation amplitude was 0.50 V)

CID of [M + Li]+ of N-phenylbenzamide (Compound 8) produced the sole fragment ion at m/z 180, as shown in Figure 3, by the putative mechanism proposed in Scheme 4. The proposed mechanism was examined further by obtaining CID spectrum after H/D exchange, as illustrated in Figure S1. As expected, fragment ions at m/z 180 and 181 were observed that reflected loss of LiOD and LiOH (Table S2), respectively, and their formation is proposed to occur by the mechanism illustrated in Scheme S1.

Figure 3
figure 3

The CID mass spectrum of [M + Li]+ ion of Compound 8

Scheme 4
scheme 4

Proposed mechanism to explain the loss of LiOH from [M + Li]+ ion of N-phenylbenzamide

To exclude the possibility that the H atom in the eliminated LiOH arises from direct transfer from the phenyl ring, fragmentation of [M + Li]+ of Compound 17 (Scheme 5) was studied (Figure S2a). In this case, the H atom bonded to the amide nitrogen atom is replaced with a methyl group. It proved to be difficult to induce fragmentation of the precursor ion even at high collision energies, and the loss of LiOH was not discovered. To examine this issue further, CID of [M + Li]+ of Compound 18 was performed after H/D exchange. In Compound 18, the H atoms at both ortho-positions of the N-linked phenyl ring are replaced by Cl. In this case, losses of LiOD and LiCl were observed, but loss of LiOH was not (Figure S2b). This implies that the hydrogen ring-walk [65] (it means that the deuteron migrated to the ortho-position will further exchange with the protons at the meta-positions and para-position) following the 1,3-H shift does not take place.

Scheme 5
scheme 5

Structures of Compounds 17, 18, and 26

Theoretical calculations were performed to evaluate the proposed mechanism and the possible involvement of a transition state and its intermediate in the loss of LiOH. A schematic potential energy surface for the proposed mechanism is shown in Figure S3 that describes the energy requirement of the reactions quantitatively.

Theoretical computations indicate that the carbonyl oxygen is the most favorable site for Li+ complexation to yield the MLi-a ion depicted in Scheme 4, which is in equilibrium with the tautomer MLi-b. The positive charge induces cleavage of the C–O bond via transition state a (TSa) and formation of an ion-neutral complex a (INC-a). Elimination of LiOH from INC-a results in product ion A (Scheme 4). The optimized structures, the lengths of important bonds (in Angstroms) and the calculated heats of formation of species involved in this reaction are illustrated in Figure 4.

Figure 4
figure 4

DFT optimized structures of the species in the reaction of losing LiOH of lithiated N-phenylbenzamide. Heats of formation are given kcal/mol and the lengths of the chemical bonds are given in Angstroms

Theoretical calculations were performed to evaluate substituent effects on the activation barrier and the stability of product ions. For compounds with electron-donating groups (e.g., Compound 9), which contains the electron-donating A-ring substituent OCH3, the energy of the TSa (51.6 kcal/mol) is lower than that of Compound 8. In addition, the product ion is stabilized by delocalization of the positive charge over the OCH3-substituented aromatic ring. These two factors cause the loss of LiOH to be more favorable for Compound 9 relative to Compound 8. These theoretical calculations yield predictions that are consistent with the experimental findings (Table 2). In contrast, compounds with electron-withdrawing groups [e.g., Compound 12 (which has an A-ring Br substituent)], exhibits an energy barrier of the TSa (52.8 kcal/mol) that is higher than that of Compound 8. Moreover, the product ion cannot readily be stabilized by delocalization of the positive charge. Consequently, loss of LiOH is less favorable for Compound 12 than for Compound 8. Similar results were obtained for compounds substituted on the B-ring (Table 2).

3.3 Investigation of the Fragmentation of the Lithiated N-Phenylcinnamide and Its Derivatives

The structures of N-phenylcinnamide and its derivatives and their major product ions with their RAs are summarized in Table 3. CID of [M + Li]+ of these compounds resulted in similar fragmentation pattern that included losses of LiOH and H2O. CID of [M + Li]+ of the prototype N-phenylcinnamide (Compound 19) produced two major fragment ions at m/z 206 and 212 (Figure 5), and their formation is rationalized by the competitive reactions illustrated in Scheme 6.

Table 3 Relative abundances of product ions in the CID spectra of [M + Li]+ ions of N-phenylcinnamide and its derivatives (excitation amplitude is 0.90 V)
Figure 5
figure 5

The CID mass spectrum of [M + Li]+ ion of Compound 19

Scheme 6
scheme 6

Proposed fragmentation mechanism of the [M + Li]+ ion of N-phenylcinnamide

Figure S4 illustrates the CID spectrum of [M + Li]+ of Compound 19 after labeling by H/D exchange, and it contains product ions at m/z 207 and 213 in addition to those at m/z 206 and 212 that arise from loss of LiOH and H2O, respectively (Table S2). This reflects H/D exchange between the H atom on the amide nitrogen atom and an H atom of the phenyl ring via a 1,3-H shift by the putative mechanism proposed in Scheme S2. In addition, CID of [M + Li]+ of Compound 26 [in which the H atom on the amide nitrogen is replaced by a methyl group (Scheme 5)] did not result in loss of either LiOH or H2O (Figure S5). This observation implies that the amide H atom is required for loss of LiOH and H2O, and the observation excludes the possibility that direct transfer of an H atom from the phenyl ring or the vinyl carbon to the amide carbonyl oxygen results in loss of LiOH.

Theoretical calculations yielded a schematic potential energy surface for these reactions illustrated in Figure 6, which describes the energy requirement quantitatively. As indicated above, the carbonyl oxygen is the most favorable site for Li+ complexation, which results in formation of the ion MLi-c in Scheme 6 that is in equilibrium with its tautomer MLi-d. As proposed in Route 1, the fragmentation pathway of elimination of LiOH via transition state 1 (TS1) (52.3 kcal/mol) is similar to that of N-phenylbenzamide. Figure S6 illustrates the optimized structures, the lengths of important bonds (in Angstroms) and the calculated heats of formation of species involved in the loss of LiOH. In Route 2, MLi-d rotates to generate a conformer MLi-e through transition state 2 (TS2) first and the energy barrier is 29.9 kcal/mol, then the transfer of the proton from the carbon adjacent to the carbonyl to the oxygen through a four-membered-ring transition state 3 (TS3), which is 57.7 kcal/mol, leads to the formation of an ion-neutral complex c (INC-c). The decomposition of the INC-c leads to the loss of H2O and the formation of ion C. This also implies that the proton transfer reaction is the rate determining step. Alternatively, the proton may transfer from the carbon which is adjacent to the carbonyl to the nitrogen, leading to the formation of MLi-f through a four-membered-ring transition state 4 (TS4) first. Then the proton migrates from the nitrogen to the carbonyl oxygen, leading to the cleavage of C–O bond through another four-membered-ring transition state 5 (TS5). The activation barriers of these two 1,3-H shift processes are 95.1 kcal/mol and 104.3 kcal/mol, respectively and they are much higher than these of TS2 and TS3. This indicates that the stepwise proton transfer reaction through TS4 and TS5 is less accessible in terms of energy. The optimized structures, the lengths of important bonds (in Angstroms) and the calculated heats of formation of species involved in this reaction are reported in Figure S7.

Figure 6
figure 6

Schematic potential energy surface (relative energies are given in kcal/mol) for the reaction of lithiated N-phenylcinnamide, calculated by DFT at the B3LYP/6-31 G++(d,p) level

The spectra of a series of compounds with different substituents at para position of the phenyl ring were measured to study the substituent effects on the fragmentation systematically. And a plot of the abundance ratio of these two ions, ln[(M + Li – H2O) +/(M + Li – LiOH)+] versus the substituent constants, σ +p , was obtained as shown in Figure S8. The trend clearly indicates that the electron-donating substituents were in favor of the elimination of LiOH, whereas the electron-attracting substituents favored the loss of H2O. As elucidated before, for compounds with electron-donating substituents, the energy of TS1 is reduced and the product ion is stabilized in Route 1, while the situation is opposite for compounds with electron-withdrawing substituents. In addition, the positive charge in the reaction of losing water mainly concentrates on the lithium cation and can not resonate to the substituted phenyl ring system, which means the energy of TS3 and the stability of product ion are less affected by the substituents. For example, Compound 19 (R = H) yielded [M + Li – LiOH]+ ion and [M + Li – H2O]+ ion in the ratio 100:23.7 (Figure 5); Compound 20 (R = N(CH3)2) yielded these two ions in the ratio 100:2.7 (Figure S9a) and in the case of Compound 25 (R = Br), the ratio of m/z 284 to 290 is 100:26.8 (Figure S9b).

4 Conclusions

The gas phase chemistry of Li+ with pesticides precursor 2-(4, 6-dimethoxypyrimidin-2-ylsulfanyl)-N-phenylbenzamide was studied by ESI-MS/MS and FTICR-MS/MS first, and then N-phenylbenzamide and N-phenylcinnamide which have simpler structures were investigated to get a better understanding. The compounds studied in each series showed similar fragmentation pathways and the loss of LiOH was observed. The H/D exchange between the H atom attached to amide nitrogen and an H atom of the phenyl ring via a 1,3-H shift was witnessed by the D-labeling experiments. The effects of substituent on the LiOH loss from the lithiated molecules were also investigated. The presence of electron-releasing groups expedited the loss of LiOH, whereas the electron-attracting groups retarded the loss. DFT theoretical calculations were also carried out to further investigate the reaction mechanisms.