CBDA isolation and purification represent a prerogative of our latest laboratory activities. In particular, hemp pollen (HP), which is a resin produced by isolating the trichomes from the plant inflorescences by sieving process, was selected as CBDA source. HP was previously observed as a rich source of cannabinoids and flavonol glycosides [61], but GC-FID and GC-MS analyses, while highlighting the occurrence of 16 cannabinoids, were not able to analytically discriminate acidic from not-acidic compounds. Herein, hemp pollen, provided by Hemp Farm Lab farmers (Caserta, Italy), underwent ultrasound accelerated maceration (UAM) (Branson UltrasonicsTM Bransonic^TM M3800-E, Danbury, CT, USA). Exploiting the alternation of pressure and cavitation as useful means for cell decomposition and, therefore, for the release of the intracellular metabolic content, represented the starting point of our investigation. On the basis of the use of chloroform/n-hexane (1:1, v/v) as extracting solution, hemp pollen (10 g; drug/solvent ratio 1:5) underwent three extraction cycles (30 min each). A mixture enriched with cannabinoids was obtained (yield 26.5%).

The scheme of applied extraction and fractionation steps is depicted in Figure 4. The compound was unequivocally identified through spectrometric and spectroscopic analyses, able to strongly differentiate it from THCA-A constitutional isomer, equally purified through preparative thin-layer chromatography (PLC). Indeed, isolated compound was first analyzed by HPLC-UV-DAD.

The chromatographic profile is shown in Figure 5A, together with the DAD spectrum, which is comparable to that reported in the literature [62]. In order to calculate the molar extinction coefficients at the three maxima absorption (λ_max) exhibited by the compound, UV-Vis spectrophotometric analysis was carried out. The coefficients ε were in accordance with those previously reported in the literature (Figure 5B).

¹H-NMR and ¹³C-NMR spectra were acquired and compared with literature data [63].

The ¹H-NMR spectrum showed the pentyl chain H-1”-H-5” protons in the spectral region of 0.90–2.91 ppm (Figure 6). In particular, the triplet at δ_H 0.90, integrable for three protons, was attributable to H-5" protons, whereas the three multiplets at δ_H 1.32, 1.33, and 1.55 ppm were consistent with the methylene protons H-3", H-4", and H-2", respectively. H-1" benzyl protons were at δ_H 2.91. In the aliphatic region of the spectrum, the protons of the monoterpene core of CBDA were also distinguishable. In particular, the diastereotopic protons H-5 and H-4 resonated as multiplets at 1.29/1.74 ppm and 2.00/2.20 ppm, respectively, while the H-2 olefin proton was observed as a singlet at δ_H 5.22. The singlet at δ_H 1.63, integrable for three protons, was attributable to H-10 methyl protons of the propenyl chain, whereas the two olefin protons H-9a and H-9b were at δ_H 4.41 and 4.47, respectively. The de-shielding effect of the nearby aromatic ring and of the olefinic function C-2‒C-3 brought H-1 methine proton to resonate as a doublet at δ_H 3.97, whereas the neighboring H-6 proton was detected at δ_H 3.01. Finally, H-7 methyl protons were detectable at δ_H 1.66. In the aromatic region, H-5’ proton, which correlated with the carbon at δ_H 110.6 in the HSQC experiment (spectrum not shown), was evident at δ_H 6.09.

The presence of the carboxylic function was highlighted in the ¹³C-NMR experiment, which showed the relative carbon resonating at δ_C 177.2. The ¹³C-NMR data were in accordance with an aromatic nucleus, binding the CBDA monoterpene core, stabilized by the formation of a stable intramolecular hydrogen bond, which massively influenced the C-1’ resonance, bringing it to δ_C 116.0.

Electrospray ionization (ESI)-QqTOF-MS/MS analysis appeared to be another tool able to discriminate the two constitutional isomers. It was recently stated that Δ⁹-THCA and CBDA exhibit MS/MS spectra similar to their respective neutral compounds, with additional neutral losses of CO₂ and H₂O because of decarboxylation and dehydration [64]. Indeed, some our analyses could further detail the MS/MS technique value. Both the molecules showed the [M-H]^- ion at m/z 357.2071, in accordance with the molecular formula C₂₂H₃₀O₄. The TOF-MS/MS fragment ion at m/z 245.1548(59) as base peak, owing to −112 Da neutral loss. This latter could be the result of concurrent neutral losses of CO₂ (−44 Da) and isoprene moiety (−68 Da). Dehydration and decarboxylation of the deprotonated molecular ion were hypothesized as fragmentation mechanisms that arise from the structural proximity of the phenolic functions with the carboxyl group. The relative abundance of the ion at m/z 191.1082(7) appeared to be one of the key fragments for discrimination purposes. In fact, this ion showed an intensity of 90.5% in THCA-A, whereas it was 24.6% in CBDA TOF-MS/MS spectrum (Figure 7 and Figure 8). Another diagnostic tool could be the ratio of the intensity of the ions deriving from dehydration and decarboxylation. In particular, the ratio between the ions at m/z 313.22 and 339.20 was 0.75 and 3.12 for CBDA and THCA-A, respectively. CBDA TOF-MS/MS spectrum also showed an abundant ion at m/z 227.1441, almost undetectable in the THCA-A TOF-MS/MS spectrum, which was owing to the loss of 130 Da (−68 + 62(44 + 18)) from the deprotonated molecular ion.

The fragment ion at m/z 271.1341 (43.2%), which was recently suggested to derive from retro Diels–Alder reaction from the [M-H-H₂O]^- ion, could also be considered a discriminant ion, and corresponded to the breakdown of the monocyclic monoterpene nucleus. The acquired data allow us to state that tandem mass spectrometry is useful for differentiating the two isomers, contrary to what was recently observed by Citti et al. [65], who claim that the distinction between CBDA and THCA can only be made by considering the different retention time. Moreover, to further strengthen this concept and to highlight its relativity in optimizing mass spectrometric analysis, some parameters were modified, among which collision energy and its spread, as well as declustering potential, which is the voltage applied to the orifice that helps to prevent the ions from clustering together. It appeared clear that parameters modification made Δ⁹-THCA-A more suitable to decarboxylation, whereas the characteristic loss of −112 Da was the main CBDA feature for establishing its most stable anion (Figure 9). Thus, no absolute assumption could be carried out. It is certain that the concurrent loss of CO₂ and isoprene more favorably occur in CBDA than in Δ⁹-THCA-A, and that for qualification and quantification purposes, the ions at m/z 245, 179, and 107 should be considered for CBDA, whereas the ions at m/z 313 and 191 better recognize Δ⁹-THCA-A.

Fourier transform infrared (FT-IR) spectroscopy was also used for characterizing CBDA [62], although a clear description of FT-IR data was not provided. Herein, FT-IR spectrum, acquired for the pure compound from hemp pollen, is reported (Figure 10). The 3427 cm⁻¹ broad band was attributable to the stretching of the O–H bond of the carboxylic function, whereas the 3073 cm⁻¹ band was ascribable to the stretching of the C–H bonds of the alkenyl and aromatic groups. The stretching of the carbonyl bond of the carboxylic function, which is strongly affected by the establishment of an intramolecular hydrogen bond with the vicinal hydroxy group, could be masked by the broad band at 1616 cm⁻¹, attributed to the latter by the stretch C–C in the aromatic ring. The presence of the carboxylic function was also defined by the stretching of the C–O bond at 1259 cm⁻¹.

The bending vibrations of the O–H bond at 1435 cm⁻¹ seemed to be distinguished from the bending modes of the C–H bonds present in the same spectral region. The bending mode in the plane of the phenolic group was located at 1393 cm⁻¹, while the bending vibration outside the plane was at 619 cm^-1. The weak bands of overtones between 2200 and 1620 cm⁻¹ and the bands ascribable to the C–C stretching vibrations in the aromatic ring were detectable at 1616, 1578, and 1493 cm⁻¹. The bending vibrations in the plane of the aromatic C–C bonds were at 1111, 1072, and 1038 cm⁻¹. The bending vibration outside the plane of the C–H bond at 887 cm⁻¹ appeared strongly diagnostic as it could be attributable to the C=C vinylidene bond. Out-of-plane bending modes are typically informative of the location and geometry of the double bond, being able to discriminate terminal and median bonds. Finally, according to previous findings [66], when optical rotations were acquired using a PerkinElmer polarimeter (series 343; PerkinElmer Life and Analytical Sciences, Shelton, CT, USA), an [α]_d²⁰ –59.05 (c 0.127, MeOH) was found. Thus, isolated CBDA showed negative optical rotation, as the other CBD-type cannabinoids reported in literature, as well as those with different lengths of the alkyl chain. In particular, it was reported that absolute configuration of CBD-type cannabinoids is (−)-trans-(1R,6R) [66].