Comprehensive analytical and structural characteristics of methyl 3,3-dimethyl-2-(1-(pent-4-en-1-yl)-1H-indazole-3-carboxamido)butanoate (MDMB-4en-PINACA)

Acetonitrile and methanol (all LC–MS grade) were supplied by the Polish Chemical Plant POCh (Gliwice, Poland); formic acid from Sigma-Aldrich (Seelze, Germany); deuterated chloroform from Armar AG (Döttingen, Switzerland); the Milli-Q system from Millipore (Millipore, Bedford, MA, USA) to be applied for water purification. MDMB-4en-PINACA sample was donated by the forensic laboratory of the Provincial Police Department in Lublin. Working solutions were stored in stable conditions at − 20 °C.

The qualification of MDMB-4en-PINACA in the prepared solution was performed using GC–MS QP2010 (Shimadzu, Kyoto, Japan). GC–MS equipment details and measurement conditions are described in [17]. The retention index of the examined substance was calculated according to [18].

The UV–VIS spectrum for the tested compound was obtained using HPLC system equipped with a Thermo Scientific Dionex UltiMate DAD-3000 detector—1024 element diode array working with the frequency of collecting the spectrum equal 100 Hz. The UV–VIS spectrum for the MDMB-4en-PINACA was recorded in the range 190–600 nm. The LC–PDA equipment details and measurements conditions are described in [17].

The experiments were performed on an LC–MS(/MS) system consisting of a UHPLC chromatograph (UltiMate 3000; Dionex, Sunnyvale, CA, USA), a high-resolution (HR) linear ion trap quadrupole-Orbitrap mass spectrometer (LTQ-Orbitrap Velos; Thermo Fisher Scientific, San Jose, CA, USA) and an electrospray ionization (ESI) source. The LC–MS equipment details and measurements conditions are described in [17].

The solution of MDMB-4en-PINACA in methanol (5 µg/mL) was directly introduced to the above linear trap quadrupole-Orbitrap mass spectrometer at the rate 50 µL/min. The positive ionization mode with the ESI needle potential at 4.5 kV was used. The function of secondary ion fragmentation (MS²) was applied for the m/z 358 ion. The collision energy used in the experiment was 10, 15, 20, 25 and 30%.

The infrared (IR) spectrum of the examined compound was registered in the wavenumber range from 4000 to 400 cm⁻¹ using a Nicolet™iS50 FTIR spectrometer (Thermo Fisher Scientific) with the ATR crystal.

The examined compound (10 mg dissolved in 0.5 mL of CDCl₃) was analyzed by NMR using an ASCEND 600 apparatus (Bruker, Bremen, Germany). The following NMR spectra were acquired:

The chemical shifts were referenced to the tetramethylsilane (TMS) signal.

Conformer search for MDMB-4en-PINACA was performed using Pcmodel (Serena Software, Bloomington, IN, USA) utilizing MMFF94 force field [19]. The DFT method with B3LYP hybrid exchange correlation functional [20] and 6-31G basis set was employed for the initial geometry optimization of the obtained conformers. The geometries exhibiting the lowest energy were re-optimized on the B3LYP/6-311++G(d,p) level of theory. The final geometries were used for the calculation of vibrational frequencies. Described ab initio computations were performed using the NWChem package [21].

The GC–MS chromatogram of methanolic extract of the seized material and the electron ionization (EI) MS spectrum of MDMB-4en-PINACA are presented in Figs. 2a and 3a, respectively. It can be observed that the obtained EI spectrum is in a good agreement with that from the SWGDRUG database. The estimated retention index equaled 2537. The structure of molecular ion and fragment ions corresponding to the most intense peaks are shown in Fig. 3b. The following ions can be recognized:

The exemplary LC–MS chromatogram of MDMB-4en-PINACA methanolic solution is shown in Fig. 2b, whereas HR-MS data corresponding to this compound, for 30% collision energy (CE), are gathered in Table 1. The obtained results indicate that the molecular weight of the MDMB-4en-PINACA [M + H]⁺ ion and its elemental composition are as follows: 358.2132 Da and C₂₀H₂₈N₃O₃, respectively. The very low difference between the theoretical and the experimental mass (Δppm) for MDMB-4en-PINACA ions proved the correctness of the elemental analysis.

Figure 4 shows the influence of the applied fragmentation energy (CE) on the profile of ions forming during the HR-MS/MS analysis with direct injection. The presented GC–MS and LC–MS data indicate a similarity between ESI and EI fragmentation patterns (see Figs. 3b and 5).

As results from the LC–PDA analysis, two maxima (at 209 nm and at 301 nm) are observed on the UV–VIS spectrum of MDMB-4en-PINACA (see supplementary material Fig. S1).

¹H and ¹³C NMR spectra of the analyzed material dissolved in CDCl₃ are shown in Figs. S2 and S3. The 2D spectra, required for the assignment of resonance signals, are presented in Figs. S4-S7. Table 2 summarizes the data from 1 and 2D spectra. The ¹H spectrum contains overlapping signals in the regions of 7.38–7.42, 5.03–5.08 and 2.04–2.13 ppm. Their assignment was mainly conducted with the use of HSQC and HMBC spectra. It is worth mentioning that, on the ¹H MDMB-4en-PINACA spectra, signals corresponding to dimethyl formamide (¹H signals: 2.88 ppm, 2.95 ppm and 8.01 ppm) were also appearing, which indicated a probability of its usage as a solvent or a catalyst during the drug synthesis.

Powder X-ray diffraction was employed for the structural characterization of MDMB-4en-PINACA (see Fig. S8). All of the major peaks were observed in the range of 2θ below 35°, with the most intensive ones at the positions of 6.4, 16.4 and 19.1°. The calculated parameters of the crystal structure are as follows:

The ab initio characterization of MDMB-4en-PINACA was performed on DFT [23] B3LYP level in gas phase, where only R-enantiomer was considered. Conformational analysis was performed to obtain stable compound geometries. During the rotamers searching process, the molecular mechanics with MMFF94 force field revealed 73 conformers that were subsequently used for further ab initio computation. Initial geometry optimization was performed with the B3LYP functional and 6-31G basis set. To avoid high computational costs, the frequency calculation was not performed. The obtained results indicated the presence of 5 stable lowest-energy conformers. To prove that the minima of energy were reached, the obtained stable geometries were re-optimized and subjected to vibrational analysis using the B3LYP functional and 6-311++G(d,p) basis set. The structures of the re-optimized conformers and their relative populations are presented in Fig. S9. As results from the figure, the final conformers contained amide protons with hydrogen bonds towards the closest hydrogen acceptors: indazole 2H nitrogen and carbonyl oxygen. The H bond length of 2.3 and 2.5 Å for N–H and H–O pairs, respectively, indicated strong character of these interactions. Intramolecular hydrogen bonding seemed to be the interaction stabilizing the lowest-energy conformers. Moreover, the rigid position of the ester group, forced by intramolecular bonding, did not allow for conformational flexibility of methyl and tert-butyl fragments. These groups were similarly oriented in all the stable conformers. The alkyl chain was the only MDMB-4en-PINACA fragment with position differing between the obtained rotamers. The conformers d and e are the most stable energetically (conformer e had only slightly higher energy), which leads to their similar abundances—38 and 36%, respectively. The other conformations of pent-4-enyl chain only slightly increased the energy of the molecule—the differences in free energy between all of the conformers did not exceed 2 kcal/mol.

The IR spectrum of the examined material is presented in Fig. S10. Its interpretation was as follows:

To conduct the assignment of the most important absorption bands below 1600 cm⁻¹, the analysis of theoretically obtained harmonic frequencies was performed, using conformer d with highest abundance. The theoretical harmonic frequencies of normal modes computed on the B3LYP/6–311++G(d,p) level of theory were scaled using the least square methods on the basis of experimental wavenumbers of the most prevalent absorption bands below 1600 cm⁻¹. The obtained scaling factor was 0.980. The scaled frequencies and calculated intensities of rotamer d were subsequently used for the assignment of normal modes of the experimental signals. Table 3 presents the assigned bands. Several bands of the considered spectral range were assigned to the normal modes of indazole and vinyl groups. The region of 1600–1300 cm⁻¹ contains multiple singals from the stretching of indazole ring with aromatic C–H rocking. The out-of-plane modes of indazole were assigned to the signals from the range of 1000–700 cm⁻¹. For the vinyl moiety, 1007 and 925 cm⁻¹ were correlated with the out-of-plane bending modes of alkene C–H bonds. Moreover, the contribution of in-plane bending of vinyl was observed for the modes assigned to the positions of 1214 and 1328 cm⁻¹.

The obtained sample of MDMB-4en-PINACA had well-developed crystal line. The material was used for the powder X-ray diffraction. Unfortunately, the sample did not contain monocrystals suitable for the single crystal diffraction. Attempts for the recrystallization did not result in the growth of monocrystals, and it was decided to conduct only powder diffraction measurements. For the acquired powder diffractogram (Fig. S8), the parameters of elementary cell were determined, but the determination of crystal lattice structure was not possible. Although the crystallographic data for synthetic cannabinoids are scarce, it can be seen that most of these compounds form crystals of monoclinic system with density in the range of 1.1–1.4 g cm⁻³ [17, 24]. Analogous structural features were observed for the MDBM-4en-PINACA sample. For the powder diffraction measurements of organic compounds, majority of the peaks appear in the 2θ range below 50°. In the case of MDMB-4en-PINACA, most of the significant signals were observed within the mentioned region, with the most intense and characteristic peaks placed below 20°.

Most of MDMB-4en-PINACA proton signals were clearly assigned to the corresponding nuclei. The differentiation of the overlapping aromatic and pent-4-en-1-yl chain resonances was conducted using heteronuclear 2D techniques. These could be done due to the greater dispersion of ¹³C signals for mentioned fragments leading to the well-separated correlation peaks. This was especially useful for 7.38–7.42 ppm signals corresponding to proton H6 and H7 (Table 2, Fig. 1).

The acquired 2D NOESY spectrum contained multiple cross-peaks (see Table 2) corresponding to the through-space correlations between protons. In the case of tert-leucinate and indazole groups, observed nuclear Overhauser effect (NOE) signals led to the connectivities similar to those from other 2D spectra. The only novel conformational information was obtained from cross-peaks of pent-4-en-1-yl protons. Signals correlating nuclei H1a, H2a and H3a to the aromatic protons of 7.38–7.42 ppm (Table 2) indicated the presence of small intramolecular distances between considered atoms pairs. Although the 7.38–7.42 ppm region contained overlapping signals from protons H6 and H7, the NOE typically occurs, when the internuclei distance does not exceed 4.5 Å, which excludes the possibility of NOE for pairs of H6-aliphatic proton. Hence, the aromatic—pent-4-en-1-yl cross-peaks were assigned to the correlations H7-H1a, H7-H2a and H7-H3a (see Table 2). Considering the possible conformations of MDMB-4en-PINACA aliphatic chain, it could be noticed that protons H2a and H3a may have differing positions and distances with respect to nuclei H7 in particular rotamers. The steric hindrance between considered aliphatic protons may lead to the conformations with H2a being in close proximity to aromatic proton, and H3a with distance to H7 greater than 4 Å, and opposite situation with H3a oriented near H7 nucleus. The former conformation would lead to the NOE cross-peak occurring only between H7 and H2a, while the latter would give correlation signal for pair H7-H3a. As those correlations were observed in 2D NOESY, it indicates that MDMB-4en-PINACA rotamers occurring in solution contain both of the described H2a and H3a orientations. Moreover, very weak cross-peaks between vinyl and aromatic protons were also observed indicating long intramolecular distances between nuclei or small abundance of rotamers with these structural feature.

Although multitude of physicochemical techniques can be employed for structural characterization of synthetic cannabinoids, only the single crystal X-ray diffraction is able to determine precisely the atom positions and the resulting conformations. The other way is to perform the geometry optimization of the molecule by quantum chemical methods. The results obtained using B3LYP DFT [23] functional for MDMB-4en-PINACA indicate the presence of five stable conformers of the molecule (see Fig. S9 in supplementary material). For each of the obtained rotamers, the conformation of carboxamide moiety is fixed by intramolecular hydrogen bonding. The amide group, which is one of the synthetic cannabinoid parts responsible for the bonding to the cannabinoidical receptors, should have a substituent generating steric hindrance required for effective binding [25]. The position of the tert-leucinate group, a steric substituent in MDMB-4en-PINACA, does not change during conformational transitions. The only fragment of MDMB-4en-PINACA with conformational flexibility is the pentenyl chain. Considering the structures of the obtained rotamers and their populations, none of the determined chain orientations is highly favored over others (see abundances of rotamers at Fig. S9 caption). Although the presence of the alkene double bond may suggest the possibility of π–π stacking with the indazole ring, such interaction did not occur in the stable conformers. This fact is also confirmed by the experimental results. In the case of π–π stacking, the centers of interacting groups are in close proximity, which leads to multiple NOE signals between the protons of the interacting groups. 2D NOESY spectrum showed only a very small correlation signal of protons 5a and 7 (see Table 2, Fig. 1), which is consistent with the edge-to-edge orientation of the vinyl and indazole ring in conformers a and b (see Fig. S9). The lack of π–π interactions is related to the polarized electrostatic nature of the indazole ring. The nitrogen atoms of the indazole and carbonyl group attached to the ring cause a significant polarization of aromatic moiety leading to the change in its quadrupole moment. A significant negative partial charge, being a component of the quadrupole moment, is distributed above both aromatic ring faces in the aromatic rings without electron-withrawing motifs [25]. For the indazole part of MDMB-4en-PINACA, the negative partial charge is distributed in the regions close to the nitrogen atoms. On the other hand, in the subsequent region of the heteroaromatic group, the deficit of electron density is observed. This distribution of electrostatic potential favors π–π interactions to the π fragments with electron realizing groups. No such effect was observed in MDMB-4en-PINACA due to fact that only a simple vinyl group is present in the described analyte.

GC–MS and ATR-FTIR are most frequently used in routine identification of synthetic cannabinoids, including MDMB-4en-PINACA. In these cases, the substance identification is based on the comparison of EI mass spectra or infrared spectra with those in the database (e.g., NIST database or SWGDRUG mass spectral library). Although it ensures fast routine workflow, a more elaborate analysis of the EI and IR spectra should also be conducted. It is especially apparent for the vibrational spectra of carboxyamideindazole with regions below 1600 cm⁻¹ being the place of multiple signals (Fig. S10). Most of the bands appearing in these range cannot be simply assigned with the “rule of thumb” as little is known about the character of their normal modes, e.g. for MDMB-4en-PINACA. For this compound, it is difficult to differentiate between the bands of indazole and vinyl vibrations and those of alkyl and carbonyl fragments in the range of low wavenumbers. The theoretical vibrational analysis of MDMB-4en-PINACA allowed us to assign the positions of the multiple bands to the stretching and the out-of-plane bending modes of the indazole ring in the regions of 1600–1300 cm⁻¹ and of 1000–700 cm⁻¹, respectively. Yet, the indazole normal modes are influenced by the vibrations of the rest of the molecule. The normal modes of the ring stretching (1600–1300 cm⁻¹) involve the aromatic C–H bond rocking. Moreover, a small contribution of C–(C=O)–N bending is observed in the out-of-plane bending modes (1000–700 cm⁻¹). For the vinyl group, the out-of-plane vibrations were assigned to the signals in 1007 cm⁻¹ and 925 cm⁻¹ positions. Contrary to the enlisted modes of indazole, the normal modes of the vinyl group include insignificant contributions from other parts of the molecule.