Introduction

The protonation of molecules that contain multiple basic sites has been the focus of extensive studies [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19]. Prototropic isomers formed by protonation at different loci in a molecule are known as protomers. Results from many studies confirm that different protomers (or deprotomers) of the same molecule can coexist in the gas phase [20,21,22,23,24,25]. Recently, different ion-mobility-dispersion methods have been used to separate protomers, prior to subjecting them to mass spectrometric (MS) or infrared (IR) detection [5, 17, 26,27,28,29].

In a recent study, Warnke et al. [28] provided conclusive evidence that different protomeric populations of benzocaine, a widely used local anesthetic, can coexist in the gas phase. From their investigations, based on ion-mobility techniques combined with IR and MS, they concluded that the immediate environment of benzocaine has a profound influence on the ratio of the N- and O-protomers manifested in the gas phase. They attributed the presence of the N-protomer to kinetic trapping of the solution-phase structure during transfer into the experimental setup [28], and concluded that the electric properties of the surrounding medium are the main determinant for the preferred protonation site that enables the kinetic trapping.

Generally, it has been assumed that whatever the MS ionization technique may be, the initial structures of gas-phase ions formed and their abundance ratios remain intact until detection [26]. Very often, the experimentally determined gas-phase ratios have been reported as values specific to the particular compound under investigation [30]. Sometimes, the ratios have even been assumed to be even intrinsic [26]. However, Campbell et al. [31], using differential mobility spectrometry, reported that the N-protonated form of 4-aminobenzoic acid can undergo conversion to the O-protonated form during gas-phase hydrogen-deuterium exchange (HDX) reactions. Previously, we have shown that the gas-phase ratios of protomers [17] and deprotomers [32] change when ion-source parameters are adjusted. Herein, we describe that the ambient humidity inside the ion source plays a decisive role in transforming the O-protomer of benzocaine to its N-protomer before the ions are subjected to mass analysis (Scheme 1).

Scheme 1
scheme 1

Protonation of benzocaine to afford either N- or O-protonated species

Full size image

Experimental

Materials

All chemicals, including benzocaine and D2O, were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA) and used without further purification. Ultrapure water was obtained from a Milli-Q purification system (Millipore Corporation, Billerica, NJ, USA).

Mass Spectrometry

Spectra were acquired on a Synapt G2 HDMS instrument (Waters, Milford, MA, USA) equipped with a HePI source [33]. Mass calibration (m/z 10−1500) was performed using a solution of sodium formate (100 ppm). Mass spectra were acquired in the positive-ion mode over a range of m/z 10−1500. For ion-mobility (IM) experiments, ions of interest were mass-selected by the quadrupole, briefly accumulated in the Trap region, and then released to the TWIM cell for separation [34]. After the IM separation, the ions traverse the Transfer cell and undergo separation in the time-of-flight analyzer according to m/z ratios. Unless otherwise specified, all IM separation experiments were carried out under the following typical instrumental conditions: Trap collision energy 4 eV, Transfer collision energy 2 eV, IM wave velocity 1500 m/s, IM wave height 40.0 V, scroll-pump pressure 3.07 mbar, source pressure 1.38 × 10–3 mbar, helium-cell pressure >14.1 mbar, IM-cell pressure 4.56 mbar (N2), ToF analyzer pressure 8.39 × 10–7 mbar, Trap pressure 3.83 × 10–2 mbar (Ar), and Transfer pressure 4.20 × 10–2 mbar (Ar).

For helium-plasma ionization (HePI) [33], a stream of high-purity helium at a flow rate of about 30 mL/min was passed through a stainless steel capillary needle (100 μm i.d.) held at an electrical potential of +3.50 kV. Other instrument parameters for the HePI experiment were as follows: extraction cone 3.0 V, and Vernier-probe-adjuster position 5.92 mm. The source- and desolvation-gas temperatures were held at 80 and 300 °C, respectively. To investigate sampling-cone ion-activation effects, the voltage was varied between 10 and 50 V while other parameters were kept constant. For atmospheric-pressure ion generation, a few milligrams of neat benzocaine on a glass slide was placed near the cone entrance and the capillary tip within the ion-source chamber (Supplementary Figure S1a). The “desolvation-gas” at a flow rate of 300 L/h was used to heat the sample. For the experiment of water vapor adding, either the desolvation gas was bubbled though water in a Drexel bottle (Supplementary Figure S1b), or a moistened cotton swab was inserted into the ion source though the hole for the lockspray connection (Supplementary Figure S1c). For some experiments, a cotton wool plug moistened with H2O or D2O was placed in the “sample-heating” gas line. (Supplementary Figure S1d).

Computational Methods

All calculations were done using Gaussian 09 program package [35]. Geometries for all proposed structures were fully optimized by Becke’s three-parameter nonlocal-exchange functional B3LYP method [36], together with the nonlocal correlation functional of Lee et al. [37], using a large 6-311++G(2d,2p) basis set. Frequency calculations were performed at the same level to verify the nature of each stationary state on the potential energy surface. Reactants/products were associated with all positive vibrational frequencies, and transition states (TS) were associated with only one imaginary frequency, for which the normal vibrational mode corresponded to the expected bond formation/breaking movements in a specific reaction pathway.

Results and Discussion

For the initial studies on protomer populations, we selected helium-plasma ionization (HePI) as the ion-generation method because of its simplicity. For HePI, samples were introduced as dry deposits by placing a few milligrams of benzocaine near the cone entrance and capillary tip within the ion-source chamber (Supplementary Figure S1a). The m/z 166 ion generated for protonated benzocaine was mass-selected by the quadrupole analyzer and transferred to the traveling-wave ion-mobility (TWIM) cell for separation. The arrival-time distribution (ATD) profile recorded for the initial 0.5-min period, without any heating-gas flow to the enclosed ion source, showed essentially only one arrival-time peak at 8.85 ms (Figure 1, Inset a1). To characterize the protomer represented by the 8.85 ms peak, a fragmentation-ion spectrum was recorded by increasing the Transfer collision energy to 20 eV (Supplementary Figure S2). The recorded spectrum showed intense fragment-ion peaks at m/z 138 and 120 for consecutive ethylene and water losses from the precursor m/z 166 ion (Inset a1, Supplementary Figure S2). The water elimination from the m/z 138 ion has been reported to be suppressed when the proton resides on the amino group: thus, we concluded that the initial protomer generated under HePI conditions is the O-protomer [28]. This result was not surprising because of the solvent-free nature of HePI. Moreover, ab intio calculations have established that the O-protomer is the thermodynamically favored species in the gas phase [28]. Furthermore, it has been reported that under certain conditions, the thermodynamically unfavorable form of protonated benzocaine can be transferred under spray conditions to the gas phase by kinetic trapping of the solution-phase structure [28]. The presence of a thermodynamically unfavored isomer in the gas phase has been commonly attributed to kinetic trapping, which enables retention of the solution-based structures [18, 19, 25, 28, 46].

Figure 1
figure 1

An ion-intensity chronogram (m/z 10–1500) recorded from mass-selected m/z 166 ion for protonated benzocaine (a). The m/z 166 ion was generated by placing a few milligrams of benzocaine near the cone entrance and capillary tip within the ion-source chamber, and passing 30 mL/min of helium through the metal capillary tube held at 3.50 kV. The flow of humidified nitrogen to the enclosed ion source was switched back-and-forth from 0 (gray areas) to 300 (white areas) L/h at every 0.5 min. Insets a1–a6 show arrival-time distributions (ATDs), co-added for each 0.5 min segment, of the mass-selected m/z 166 ion. IM separation was carried out at a wave velocity of 1500 m/s and a wave height 40.0 V. Other experimental parameters were: Vernier-probe-adjuster setting 5.92 mm, sampling-cone voltage 50 V, extraction cone 3.0 V, source temperature 80 °C, “desolvation-gas” temperature 300 °C, trap collision energy 4 eV, and transfer collision energy 2 eV

Full size image

To verify if the O-protomer of benzocaine can be transformed to the N-protomer entirely in the gas phase, we added water vapor to the heated (300 oC) nitrogen supply of the ion source (by bubbling thorough water in a Drexel bottle; Supplementary Figure S1b). The changes were rapid and very dramatic: the intensity of the O-protomer peak began diminishing at once and a new prominent peak appeared at 11.27 ms for the N-protomer (Figure 1, Inset a2). The transformation of the initial ATD chronogram to one that was dominated by the N-protomer peak took only about 150 ms (Supplementary Figure S3). Moreover, as soon as the humidified N2-gas flow was stopped, the intensity of the peak for the so-called “kinetically trapped” N-protomer diminished rapidly, and within 120 ms the O-protomer peak regained its original dominance (Supplementary Figure S4).

A minor peak is seen at 7.53 ms in the arrival-time distributions (ATDs) depicted in Figures 1 and 2. This peak was due to an ion of m/z 138 that formed in the Trap-region at a low collision energy setting of 4 eV, by an ethylene loss from the O-protomer. When the mass-selected m/z 166 ion, generated at a conevoltage setting of 30 V (with water vapor added to the desolvation-gas line) was subjected to Trap fragmentation at a collision energy setting of 20 eV, several new peaks appeared in the arrival-time distribution profile (Supplementary Figure S5, Panel “a”). The increase of the intensity of the 7.53 ms peak, while that for the m/z 166 at 8.85 ms decreased, indicated that the m/z 138 ion represented by the 7.53 ms peak originated from the O-protomer. Analogously, the peak at 9.12 ms (Supplementary Figures S5 and S6) represents the m/z 138 ion that originated from the N-protomer.

Figure 2
figure 2

Arrival-time distributions (ATDs) recorded from the mass-selected m/z 166 ion generated by HePI from benzocaine without (a1–a5), and with (b1–b5) water vapor added to the desolvation-gas supply. The sampling-cone voltage settings were 10 V (a1 and b1), 20 V (a2 and b2), 30 V (a3 and b3), 40 V (a4 and b4), and 50 V (a5 and b5). The m/z 166 ion was generated by HePI from a few milligrams of benzocaine placed near the cone entrance and capillary tip within the ion-source chamber. The sample was warmed up by a flow of nitrogen (“desolvation gas”) humidified by bubbling it through water beforehand. The desolvation-gas temperature was set to 300 °C, and its flow to 300 L/h

Full size image

To further scrutinize the role of water vapor on the gas-phase protomer ratio of benzocaine, we conducted an experiment by inserting a cotton swab soaked with water into the HePI source region (Supplementary Figure S1c). As soon as the swab was inserted at 0.93 min, the relative intensity of the 11.27 ms peak for the N-protomer started to increase (Supplementary Figure S7). When the swab was withdrawn at 2.00 min, the relative intensity of the N-protomer diminished gradually while the overall signal intensity remained approximately the same. Clearly, a high water-vapor pressure in the source shifts the equilibrium ratio toward the N-protomer.

Another experiment was conducted with and without the addition of water vapor to the desolvation gas, while changing the sampling-cone voltage. The electrical potential difference maintained between the entrance nozzle (sampling-cone entrance) and an exit orifice (extraction cone) in the first vacuum region of an atmospheric-pressure ionization mass spectrometer is generally known as the sampling-cone voltage. The voltage bias maintained between these two orifices serves to control the ion transmission efficiency. It is known that in addition to increased transfer efficiency, the ions gain acceleration and endure vibrational activation due to collisions between themselves and the residual gaseous molecules in this region [38,39,40,41]. In our experiments, when no water vapor was deliberately added, the activation of the m/z 166 ion by increasing sampling-cone voltage showed little effect on the protomer ratio of benzocaine (Figure 2a1–a5). In stark contrast, when there was a sufficient amount of water vapor in the ion source, the gas-phase protomer ratio changed dramatically (Figure 2b1–b5; for non-normalized data see Supplementary Figure S8). For example, at a sampling-cone voltage setting of 10 V and regular nitrogen heating-gas flow, the ion-mobility profile recorded essentially showed only one peak for the first-to-arrive O-protomer (Figure 2a1). Under the same source conditions, when the source was engulfed with moistened nitrogen, the intensity ratio of the peaks for O- and N-protonated benzocaine immediately became 3:1 (Figure 2b1). Thereafter, as the sampling-cone voltage was increased, more and more N-protomer was produced; above 40 V, the N-protomer peak became the base peak (Figure 2b4–b5). Although the decrease of the O-protomer intensity may manifest some contribution due to the discriminatory fragmentation rates of the two protomers, the impact of in-source fragmentation to the protomer ratio is negligible at low cone voltages. For example, a peak for the N-protomer appeared even at a cone voltage of 10 V (Supplementary Figure S8).

Our results clearly demonstrate that the transformation of the O-protomer of benzocaine to its N-protomer is mediated by ambient water vapor. The high mobility of the proton in liquid water has been rationalized by Grotthuss long time ago [42,43,44,45]. We propose that a similar mechanism operates in the gas phase as well. In fact, proton migrations in the gas phase are expected to be even faster than those in the liquid phase. The mechanism can be envisioned as a water-mediated “relay” process. A similar mechanism has been proposed to rationalize the tautomerization of ions such as deprotonated 4-hydroxyl benzoic acid [20] and protonated 4-aminobenzoic acid [31].

With benzocaine, the initial gas-phase O-protomer generated by HePI is more favored than the N-protomer. The latter is formed preferentially in “humid” ion sources. Similar observations have been made by Warnke et al. [28]. However, in the humidified atmospheric region of the ion source, the energy difference between the two species is rendered smaller by interactions with gaseous water clusters (Figure 3). We considered several water-adduct combinations, and our ab initio calculations indicated that the energy-optimized structure with three water molecules hydrogen-bonded to the O-protonated center is the energetically most favorable form (Figure 3 and Supplementary Table S1). Similar results have been reported by Chang et al. [46] for 4-aminobenzoic acid methyl ester. They provided conclusive gas-phase infrared data to show that 4-aminobenzoic acid methyl ester when hydrated with one water molecule bears an O-protonated structure, whereas a cluster with three water molecule is N-protonated. We propose that upon ion activation the water in O-protonated benzocaine hydrated by three or four water molecules removes the charge-imparting proton from the carbonyl site, while transferring one of its own protons to the amino group to form the N-protomer. The presence of small water clusters under low-vacuum conditions has been demonstrated by infrared laser absorption spectroscopy [47]. For the proton transfer process to be effected, a water bridge should be formed between two sites (Scheme 2). Once the solvent bridge is formed, a proton could be transferred along the hydrogen-bonded water “wire” to the amino site by a mechanism similar to that described by Grotthuss [43, 45].

Figure 3
figure 3

Calculated zero-point-corrected electronic energies of the bridged, O- and N-protomer of (ethyl 4-aminobenzoate•H)+ (H2O)3 relative to those of respective solvated N-protonated structures [in kcal/mol by the density functional theory method B3LYP using a 6-311++G(2d,2p) basis set]

Full size image
Scheme 2
scheme 2

The solvent-bridged structure of (ethyl 4-aminobenzoate•H+)(H2O)3

Full size image

The energetics of three water molecules interacting with O- and N-protonated forms of benzocaine is illustrated in Figure 3. According to our calculations, the 4.56 kcal/mol energy difference between the products formed by the association of three water molecules with each protomer of benzocaine is smaller than the 11.69 kcal/mol value calculated for the unhydrated structures (Supplementary Table S1).

In the energy-optimized water-bridged structures obtained, the charge is localized more on the carbonyl-bound proton (19-H atom in Supplementary Figure S9) participating in the three-water-molecule bridge (Figure 4). It is known that ions endure vibrational activation due to collisions with residual gaseous molecules in the super-sonic expansion region when ions first encounter the vacuum [38,39,40,41]. In this sampling-cone ionactivation region, the charge-imparting proton on the carbonyl group can be shunted to the NH2 group via a water-bridged transition structure (Figure 4). The ion-activation process initially supplies the energy required to overcome the transformation barrier from the favored O-protomer to the N-protomer; then the adduced water molecules are stripped to generate the N-protonated species, which enter the high-vacuum region for ion-mobility separation and eventual detection. When the ion-source humidity is low, the water-bridge is not formed: consequently, the O-protomer is the primary ion that is transferred to the high-vacuum region under dry conditions. Moreover, our results show that in addition to humidity, high sampling-cone voltages enhance the N-protomer formation (Figure 2b1–5). The local activation (“heating”) afforded by the acceleration and collisions with the residual molecules while passing through the sampling-cone region provides the energy required for the transformation to the N-protomer. Similar ion activation and transformation in the sampling-cone region have been reported for 4-hydroxylbenzoic acid and aniline [17, 32].

Figure 4
figure 4

The geometry-optimized solvent-bridged structures of (ethyl 4-aminobenzoate+H)+ • (H2O)3 by the density functional theory method B3LYP using a 6-311++G(2d,2p) basis set

Full size image

Further support for the proposed water-bridge relay mechanism was obtained from experiments conducted with D2O vapor. Though the deuterium bond is known to be stronger than the hydrogen bond in neutral water clusters, the situation is reversed with protonated water clusters [48]. To prove our proposed mechanism, we added D2O to the desolvation gas line (Supplementary Figure S1d). Under these conditions, a complete H/D exchange was obtained. We mass-selected the m/z 169 ion for the d 3 -isotopologue of protonated- benzocaine, generated at a cone voltage of 40 V, and subjected it to IMS. The peak-intensity ratio between the O- and N-protomers with H2O added to the desolvation-gas line was about 0.42:1 (Figure 5a), whereas that of the O- and N-deuteromers was about 0.23:1 (Figure 5b). This significant increase in O/N- peak ratio suggests that the transfer of a deuteron is easier than the transfer of a proton by the water bridge mechanism.

Figure 5
figure 5

Arrival-time distributions (ATDs) recorded from mass-selected the m/z 166 (a) and 169 (b) ion generated by HePI at desolvation-gas temperature 150 °C, from benzocaine at a sampling-cone voltage setting of 40 V. The m/z 166 and 169 ions were generated by HePI by placing a few milligrams of benzocaine near the cone entrance and capillary tip within the ion-source chamber. The sample was warmed up by desolvation gas that passed over a cotton wool plug moistened with H2O (a) or D2O (b)

Full size image

Conclusions

An oft-repeated question asks whether electrospray ionization samples solution-based structures. After an extensive study, Schröder et al. [23] concluded that ESI-MS does not sample the solution structures but it does reflect the situation in solution to a considerable extent. In other words, the ions can still retain some “solution memory.” Herein, we have demonstrated that source ion activation conditions and moisture exert a decisive influence on the protomer population of benzocaine. When the water vapor pressure is low the population of the O-protomer increases. However, the so-called kinetically trapped N-species can be “untrapped,” by lowering the water vapor pressures in the ion source. Thus, it is most likely that even under electrospray conditions, the manifested protomer ratios are due to post-ionization transformations that occur in the gas phase during the ion transfer processes. In reality, the manifested protomer ratio is dictated by the amount of water and other solvent vapors present in the numerous intermediate vacuum zones along the entire ion pathway between the ion source and the mass detector, and the internal energy changes the ions undergo during transmission.