The NI mass spectra of all compounds examined showed peaks of anions [ArN(C₆H₄R²)-CH=C(COOR³)COO]⁻ ([A]⁻) (Ar = R²indolyl or Me₂NPh) and products of their dissociation: [A − CO₂]⁻ (except 5) and [ArN(C₆H₄R²)]⁻. Protonated and cationated dimers, [A₂H]⁻ and [A₂Na]⁻, trimers [A₃Н_nNa_m]^- (n + m = 2) and tetramers [A₄Н_nNa_m]⁻ (n + m = 3) were also recorded (Table 1 and Figure 2(a)).

Besides the above ion peaks, the NI-ESI mass spectra of compounds 1 and 2 displayed small intensity peaks of ions at m/z 259 and 273, respectively. It was reasonable to assign them to the ions formed due to the elimination of the CO₂ and R³OH molecules from [A]^- ([A]^- and [A²]^-, respectively). Moreover, the

^aThe Table lists the averaged over scans values.

corresponding peaks at 273 and 263 m/z were absent in the spectra of compounds 3, 4 and 5. Since 3 and 4 lack the hydrogen atoms at the nitrogen one of the indole ring and 5 hasn’t this ring at all, the first impulse was to consider that for 1 and 2 the elimination process discussed involved the hydrogen atom at the indolyl nitrogen atom. However, it was rather surprising to find that, while in the case of compound 4, the MS² spectrum of the similar ion [A⁴]^- lacked a peak at m/z 273 which would correspond to ion [A⁴ − CO₂ − EtOH]^-, the MS² spectrum of [A³]^- in the case of compound 3 showed an abundant peak of the [A³ − CO₂ − MeOH]^- ion. It may be caused by the point that, though the compound lacks the hydrogen atom at the indole nitrogen one, but under the more drastic conditions of MS² CID in comparison with those of in-source CID, the fragmentation pathway with the participation of another and less mobile hydrogen atom occurred. Nevertheless, this MS² spectrum of 3 cased some doubts on the validity of the above suggestion that the hydrogen atoms at the indole nitrogen atoms were involved in these fragmentations of ions [A]^- and [A²]^-.

To solve this problem and for the elucidation of other reactions peculiarities, experiments with compounds 1 and 4 were performed in the presence of heavy water. Four modes were employed; in the first variant, D₂O was introduced in the ion source as a solvent for the analytes (method A). Protocols C and D involved additional introduction in the ion source of heavy water and water via a syringe pump, respectively. By application of method B, the analytes were introduced in the source in water solutions, while heavy water was admitted through the syringe pump. In the case of compound 1, ion [A]^- and its fragment ions save one contained deuterium-labeled components thus indicating hydrogen-deuterium exchange to have occurred (Table 2). The comparison of the data for method A with those for method B demonstrates that the H,D-exchange proceeded inside the ion source to a noticeable extent. With that, however, the percentage of deuterium containing ions was greater in each deuterated entry of this ionic group when method A was employed. It can be explained by the fact that under the protocol A conditions, the exchange occurred in pure D₂O, while a mixture of D₂O and H₂O was present in the ion source when method B was employed. This finds support in the points that the percentages of deuterated components in the ions increased or decreased when methods C or D were used, respectively. It reveals especially clear by the example of ions [A]^- with m/z 335 and 336 (Table 2). However, the results obtained can’t exclude the possibility that the exchange could partly proceed in the solution of the analyte in D₂O before its introduction into the ion source.

Since ion [A − CO₂ − MeOH]^- in the case of compound 1 and all ions listed in Table 2 for compound 4 did not contain deuterium atoms, these experiments strictly supported two points: 1) the hydrogen-deuterium exchange in compound 1 occurred with the hydrogen atom located at the nitrogen atom of the indole ring; 2) the fragmentation of [A]^- to give the above mentioned first ion involved this ion exactly. The latter is apparently valid for ion [A²]^- of compound 2.

In the cases of compounds 1, 2, and 4 the dissociation patterns of [A]^-, [A²]^-, and [A⁴]^- ions, respectively, occurred due to in-source CID turned out to be virtually the same as in MS² experiments differing in the intensities of the corresponding peaks only (cf. Figure 2(a) and Figure 2(b)). At the same time, besides

^aThe contributions from ions containing ¹³C and the other natural heavy isotopes were removed as described afore (see Section 2).

the coincided ions (305 and 221 Da), other fragment ions of [A³]^- were present in its MS² spectrum (331, 323, 317, 291, and 273 Da). This difference again seems to be caused by the fact that the collision energies were different in these CID modes.

Ions [A₂Na]^- were formed most likely due to interaction of ions [A]^- with neutral molecules of analytes (reaction 1).

Two possible ways for the formation of ions [A₂H]^- could be anticipated. The first path involved hydrolysis of ions [A₂Na]^- (reaction 2). The second implied the hydrolysis of neutral ANa or protonation in the ion source of [A]^- by H₂O and even CH₃CN (a solvent in many experiments). AH formed then added [A]^- (reaction 3)¹.

Trimeric and tetrameric ions appear to be formed by the analogous schemes involving addition the corresponding dimeric or trimeric ions to neutral molecules and hydrolysis processes.

Peaks of the protonated and cationated molecules [M + H]⁺ and [M + Na]⁺of compounds 1 - 5 (M = ANa) were present in their PI-ESI mass spectra. Besides them, the mass spectra showed peaks of fragment ions of the former ions, ions [AH + H]⁺ (with significant peak abundances of such ion and its fragments for 1 only), and also ions of dimers and trimers (Table 3, Figure 3(a)).

With that, the peak abundances of protonated ions [M + H]⁺ significantly exceeded those of cationated ions [M + Na]⁺, while the relative abundances of ion peaks of dimers at m/z 695, 751, 723, 751 and 703 for compounds 1 - 5, respectively, were great, being even 100% for 1 - 3, and 5. This suggests [MAH + H]⁺ structure of these ions and the reaction 4 for their formation.

Two other theoretically possible schemes involving the participations of [AH + H]⁺ or [M + Na]⁺ ions, and AH neutral molecules seem less likely for compounds 2 - 5. Actually, the peak abundances of [AH + H]⁺ were extremely small in their spectra (see Table 3). This means that the amounts of the neutral AH molecules were also small. Further, as we underlined above, the peak abundances

^aSee footnote “a” to Table 1.

of [M + Na]⁺ ions were small as compared with those of [M + H]⁺ ones that speaks in favor of the fact that the reaction mostly occurred via the latter ions.

As relates to compound 1, the peak of [AH + H]⁺ possessed a pronounced abundance in the mass spectrum. Thus, the reactions 5 could contribute to the process of the [MAH + H]⁺ cation formation.

The participation of [M + Na]⁺ is unlikely in this case by the same reason as above.

The origin of fragment ions of [M + H]⁺ with m/z 319, 347, 333, 347, and 323 in the spectra of 1 - 5, respectively, is worth discussion. The ions can be described as [M + H ? NaOH]⁺, and this fragmentation seems quite reasonable. However, the MS² spectra of the precursor ions showed these products ions either to be present with negligible abundances, or to be absent. An alternative fragmentation process is [AH + H]⁺ − H₂O (see reaction 5 for a possible way of formation of this precursor ion; it could also be generated via hydrolysis of the [M + H]⁺ ion.) The above MS² argument that would indirectly speak in favor of the latter process has a hint of doubt for compounds 2 - 4, since the abundances of the precursor-product pair were rather small in these cases, while ions [M + H]⁺ were abundant. Moreover, ion [A⁵H + H]⁺ was virtually absent on the spectrum of 5, whereas a small abundance ion with m/z 323 was present and again

[M⁵ + H]⁺ was abundant. Thus, we cannot exclude possibility of contributions of [M + H − NaOH]⁺ to the abundances of the ion discussed in the cases of species 2 ? 5. At the same time, the peak abundances of the ions of the corresponding pair, 337 and 319 Da, were relatively great in the case of 1 (Table 3). Hence, the aforementioned MS² data for the protonated molecule of 1 indirectly support the fact that the 319 Da ion was formed mostly owing to elimination of water from the [AH + H]⁺ cation.

Another fragmentation of [M + H]⁺ cations is of interest. It involved the loss of two water molecules, was expressed for compounds 1, 3, and 5, while minor for 2 and 4 (Table 3). It seemed rather unexpected, but was supported by MS² spectra in the cases of the first above group of the compounds, these product ions being dominant in the spectra (see, e.g., Figure 3(b)).

The 297 Da ions in the spectra of 2 and 4 can be described as either [M + H − EtOH − CO₂]⁺ or [M + H − CO₂ − EtOH]⁺. Though both ions [M + H − EtOH]⁺ and [M + H − CO₂]⁺ were absent in the spectra, the MS² spectrum of [M⁴ + H]⁺ ion of 4 showed the former ion rather than the latter. Also, analogous ion [M⁵ + H − MeOH]⁺ and not [M⁵ + H − CO₂]⁺ was present in the MS² spectrum of [M⁵ + H]⁺ in the case of compound 5. Based on all this, we believe that the 297 Da ions discussed were formed due to the sequential elimination of the alcohol and carbon dioxide from the protonated molecules of 2 and 4. We also believe that this path took place in the cases of compounds 1, 3, and 5 producing ions 283, 297, and 287 Da, respectively.

An alternative fragmentation that could also have provided the above m/z ions in the mass spectra of these compounds seemed to be [M + H]⁺ − H₂O − H₂O − NaOH. However, the mass spectra of 2 and 4, and the MS² spectrum of ion [M⁴ + H]⁺ of 4 lacked the corresponding product ions (311 Da). It allowed us to exclude such fragmentation from the consideration.

When 1 was introduced in D₂O solution, the treatment of the mass spectrum as mentioned in section 2 showed ions of analyte 1 formed owing to addition of a proton or deuteron to contain between zero and two deuterium atoms [359 Da (12%), 360 Da (43%) and 361 Da (45%)]. This means that the deuterium atoms were introduced into the analyte molecule by both H,D-exchange and the addition of a deuteron. At the same time, the corresponding ions of compound 4 obtained and registered under the same conditions contained maximum 1 deuterium atom [387 Da (25%) and 388 Da (75%)]. Contrary to 1, this compound has no hydrogen atom at the nitrogen one of the indole ring. Thus, all of this is consisted with the conclusion drawn before when negative ions were examined that the H,D-ex- change in compound 1 and its ions involved just this hydrogen atom. It is essential to note, that protonation (or deuteron addition) could hardly proceed in the solutions of analytes before introduction in the ion source, since the analytes are salts of weak acids and a strong base and thus provide basic mediums. Hence, the addition of proton or deuteron most likely occurred immediately in the ion source.

As we mentioned above, the abundant peaks of fragment ions [A − CO₂]^- and [ArN(C₆H₄R²)]^- are present in the NI-ESI mass spectra of compounds 1 - 4 (Table 1). At the same time, peaks of ions [A − CO₂ − R³OH]^-, though of low intensities, were detected reliably for compounds 1 and 2. Their formation from ions [A]^- ([A]^- and [A²]^-, respectively) was supported by the corresponding MS² spectra. Quantum-chemical calculation performed in the case of compound 1 provided three possible structures of [A − CO₂ − MeOH]⁻ (Figure 4). However, the ion with structure 1 turned out to be the most stable and even more stable than that with a cumulene fragment (such cumulene motif was reported, e.g., as propadienone [22] ). The formation of this ion seems to be kinetically rather than thermodynamically controlled. However, during the period when this ion was present in the ion source, isomerization could occur for it to acquire the most stable structure 1. Moreover, according to the extended Hammond-Leffler postulate, [23] the more the reaction is endothermic the more the transition state for generating the [A − CO₂ − MeOH]⁻ ion should be shifted to the reaction products

along the reaction coordinate. If the process is exothermic, the shift has to occur in the opposite direction, to the reagents. In both cases, the activation barrier is minimal most likely for the most stable product. From all of this, anion [A − CO₂ − MeOH]⁻ appears to have structure 1.

Its formation apparently included the nucleophilic attack of the C(2) atom of the indole ring upon the C-atom of a carbonyl group followed by elimination of MeOH involving abstraction of the hydrogen atom from the indole nitrogen one. Then 1,3-hydrogen shift occurred to give ion 1. The other formally possible structures that could be suggested a priori proved to be either impossible sterically, or the optimization of their geometry resulted in structure 2.

When a compound lacked the hydrogen atom at the indole nitrogen one and the collision energy was high enough, an ion with the structure like 3 (e.g., with N-Me bond instead N-H one) might be formed as a result of the abstraction of the hydrogen located at the α-carbon atom of the indole ring. It seems to be the case of compound 3, for which fragmentation [A³]^- − CO₂ − MeOH was observed under MS² conditions.

As indicated above, molecules of 1 - 5 added a proton or sodium cation under conditions of PI-ESI (reactions 6 an 7, respectively). The abundances of ions [M + H]⁺ were always higher than those of the [M + Na]⁺ ions (see Table 3). As above, in the framework of the extended Hammond-Leffler postulate, this is consistent with the calculated values of proton affinity (PA) and the affinity to cation Na⁺ for compound 1, the former being 3.5 times greater than the later.

It is worthy of noting that the calculated PA of molecule 1 is very high prevailing that for ammonia (852.6 kJ∙mol⁻¹) [24] .

There is little likelihood that the reaction of hydrolysis of [M + Na]⁺ ion (reaction 8) could to any noticeable extent contribute to the process of the formation of [M + H]⁺ ion, since it is highly endothermic.

A priori, eight atoms in molecule 1 are amenable to protonation. Those are C(2), C(3), N(10), C(18), O(20), O(22), O(23), and O(25) (see Figure 5). The calculations demonstrated that there are two positions in the molecule with the lowest and relatively close PA, the protonation leading to structures A and B. However, the PA value for structure A is nevertheless greater, and protonation seems to predominantly occur at O(25) position. As aforementioned, ions [M₂ + H]⁺ were formed most probably by interaction of molecules M with their protonoted forms (the first scheme of reaction 4). The heat effect of this interaction was calculated to be −261.8 kJ∙mol⁻¹ for compound 1. An alternative path could

comprise the interaction of the molecules with their cationated forms and the hydrolysis of the dimers obtained (reaction 9).

However, the abundances of [M + Na]⁺ ions in the spectra were considerably less than those of [M + H]⁺ ions, the ΔН of the interaction between M and [M + H]⁺ was less than that of M and [M + Na]⁺ exemplified by the case of 1. Besides, the formation of [M₂ + H]⁺ ions involving the latter interaction would be a two-step process, the second step being highly endothermic. All this supports that the [M₂ + H]⁺ ions were formed predominantly according reaction 4.