Quali–Quantitative Characterization of Volatile and Non-Volatile Compounds in Protium heptaphyllum (Aubl.) Marchand Resin by GC–MS Validated Method, GC–FID and HPLC–HRMS2

Protium heptaphyllum (Aubl.) Marchand (PH) trees are endemic to the tropical region of South America, mostly Brazil. Antibacterial, antinociceptive, anti-inflammatory, anxiolytic, antidepressant and anti-hyperlipidemic/anti-hypercholesterolemic effects were reported for its resinous exudate Protium heptaphyllum resin (PHR). This work aims to provide a qualitative and quantitative consistent chemical profiling of the major constituents of this resin and two extracts enriched in acid (acidic triterpene concentrated extract, ATCE) and neutral triterpenes (α and β-amyrin concentrated extract, AMCE). GC–MS/GC–FID was used for volatile terpene fraction, a validated GC–MS method was developed for quantification of neutral α and β-amyrin and HPLC–APCI HRMS2 was used for acidic triterpenes analysis. The chemical investigation reported 29 molecules, including 14 volatile terpenes, 6 neutral triterpenes and 11 acid triterpenes. The most abundant compounds were α-amyrin (251.28 g kg−1, 123.98 g kg−1 and 556.82 g kg−1 in PHR, ATCE and AMCE, respectively), β-amyrin (172.66 g kg−1, 95.39 g kg−1 and 385.58 g kg−1 in PHR, ATCE and AMCE, respectively), 3-oxo-tirucalla-7,24-dien-21-oic acid (80.64 g kg−1, 157.10 g kg−1 and 15.31 g kg−1 in PHR, ATCE and AMCE, respectively) and 3α-hydroxy-tirucalla-8,24-dien-21-oic acid (77.71 g kg−1, 130.40 g kg−1 and 11.64 g kg−1 in PHR, ATCE and AMCE, respectively). Results showed specific enrichment of acidic and neutral triterpenoids in the two respective extracts.


Introduction
The genus Protium is the most heterogeneous and abundant of the Burseraceae family, and the 135 known species that spread throughout the Amazon region, mostly in Brazil, make this genus the second most hyperdominant genus in the Amazon. Since ancient times, the oleum resin of Protium heptaphyllum (Aubl.) Marchand that is exudated from resiniferous ducts has been collected by the Amazon local population of Quilimbola. Oleum resin of Protium heptaphyllum is called by different popular names: Breu, Breu Branco, Olíbano Brasileiro, Resina de Almecéga, Almecegueira/Almesca Resin and Amazonic White Pitch [1][2][3][4]. The P. heptaphyllum oleum resin is a rich source of biologically active volatile and nonvolatile terpenes. Both these groups of secondary metabolites were recognized as critical factors in plant defense, since they can mediate the plant-herbivore and oxygenated monoterpenes (1,8-cineole, trans-verbenol, terpinen-4-ol, p-cymen-8-ol), as well as the sesquiterpene β-E-caryophyllene. The most abundant VOCs in PHR were αpinene 1.26% w/w, p-cymene 1.11% w/w, δ-3-carene 0.70% w/w, β-phellandrene 0.67% w/w and α-phellandrene 0.42% w/w. The total content of VOCs in PHR amounts to 5.07% w/w. Table 1. Volatile organic compound (VOC) identification and quantification in Protium heptaphyllum resin (PHR), acidic triterpene concentrated extract (ATCE) and α and β-amyrin concentrated extract (AMCE) samples.  1 Values are expressed as g kg −1 of dry weight and are the mean of at least three replicates ± standard deviation. 2 Calculated Linear Retention Index (LRI C ) with a C 7 -C 30 n-alkanes mixture and the LRI reported in literature (LRI L ) [16].

Method Validation for Amyrins Quantification
Instrumental quantitation of neutral triterpenes in the oleum resin of P. heptaphyllum and the two extracts (ATCE and AMCE) was performed upon diethyl ether solvent extraction and by gas chromatography coupled to mass spectrometry.
The separation of the two isomers, αand β-amyrin, occurs with complete resolution in the isothermal zone at 320 • C ( Figure 1). The two isomers belong to nonacidic ursane and oleanane triterpenes groups, respectively. Therefore, both isomers show similar fragmentation spectra with 218 m/z as base peak and 203 and 189 m/z as main fragment ions. The only difference between α and β-amyrin spectra resides in 203 m/z peak abundance, which is slightly superior in β-amyrin (see supplementary materials, Figure S2). Quantitation method for α and β-amyrins was validated according to FDA and EMA guidelines [17,18]. The concentration range of validation was between 0.1 to 10 mg L −1 . Validation parameters investigated were linearity (DIFF%), selectivity (SEL%), sensitivity (limit of detection (LOD), limit of quantitation (LOQ)), percent standard deviation (RSD%) and precision (ERR%). Eight-point calibration curves were analyzed (0.1, 0.2, 0.5, 1, 2, 6, 8, 10 mg L −1 ), and three replicates performed in three different days were compared to ensure the best repeatability, selectivity and precision of the analytical method (Table 2) (formulas are reported in supplementary materials, Figure S1). The reported values are R 2 for each curve, the linearity and repeatability (DIFF%), selectivity (SEL%), sensitivity (limit of detection (LOD), limit of quantitation (LOQ)), percent standard deviation (RSD%) and precision (ERR%) for the unweighted calibration method; values reported in brackets are relative to 1/Y weighted calibration. Parameters are calculated as reported in supplementary materials. Quantitation method for α and β-amyrins was validated according to FDA and EMA guidelines [17,18]. The concentration range of validation was between 0.1 to 10 mg L −1 . Validation parameters investigated were linearity (DIFF%), selectivity (SEL%), sensitivity (limit of detection (LOD), limit of quantitation (LOQ)), percent standard deviation (RSD%) and precision (ERR%). Eight-point calibration curves were analyzed (0.1, 0.2, 0.5, 1, 2, 6, 8, 10 mg L −1 ), and three replicates performed in three different days were compared to ensure the best repeatability, selectivity and precision of the analytical method (Table 2) (formulas are reported in supplementary materials, Figure S1).
The recovery was observed to be quantitative for all the molecules: results were 85 ±13%, 101 ± 3% and 98 ± 1%, respectively, in PHR, ATCE and AMCE extracts. In Table S1 (supplementary materials), calculation errors of three calibration curves are reported according to EMA validation parameters [18]. Validation requires that the back calculated concentrations of the calibration standards should be within ± 15% of the nominal value, except for the LLOQ, which should be within ± 20% and at least 75% of the calibration, with a minimum of six calibration standard levels; recalculation errors are reported in Table S1. The reported data showed that the ERR% fit into the acceptability range only above 0.5 mg mL −1 . To obtain an acceptable error at the lowest calibration points, a weighted regression system model (1/Y) was used, increasing the sensitivity of method ( Table 2). In supplementary Table S2, the recalculation errors ERR% are reported also in the case of 1/Y weighted calibration. The reported values are R 2 for each curve, the linearity and repeatability (DIFF%), selectivity (SEL%), sensitivity (limit of detection (LOD), limit of quantitation (LOQ)), percent standard deviation (RSD%) and precision (ERR%) for the unweighted calibration method; values reported in brackets are relative to 1/Y weighted calibration. Parameters are calculated as reported in supplementary materials.

Quantification of Neutral Triterpens
The developed and validated GC-MS method was used to quantify αand β-amyrins and the occurring lipophilic triterpenoids in raw material, ATCE and AMCE extract samples. The quantitative data were summarized in Table 3. The presence of amyrones, the oxidized forms of amyrins, in position 3 due the occurrence of carbonyl group (C3=O) instead of hydroxyl (3-OH), previously reported by [11,13], was confirmed by NIST identification and, furthermore, by similar fragmentation pathway and molecular peaks decreased by two mass to charge units (424 instead of 426 m/z) with respect to amyrins. Additionally, two further triterpenes were found: brein (3b,16b-dihydroxy-olean-12-ene) and maniladiol. Their quantification was obtained with α-amyrin calibration curve for brein (ion 234 m/z) (oleanolic family) and β-amyrin calibration curve for maniladiol (ion 234 m/z) (ursanic family). In Figure S3, all found neutral triterpenes structures are proposed. Molecules identified and quantified belong to the ursanic (α-amyrin, α-amyron and Brein) and oleanic (β-amyrin, β-amyron and Maniladiol) triterpenes classes; their presence has been reported in all the samples except for the amyrones in AMCE extract. Neutral triterpenes made up the 48% (w/w) of PHR, the starting raw material, with predominance of α and β-amyrin, 25.1 and 17.2% w/w, respectively; α and β amyron registered concentrations of, respectively 1.84 and 1.43% (w/w), while brein and maniladiol were 1.49 and 1.22% (w/w), respectively. The extraction process for the production of AMCE and ATCE lowered the total concentration of neutral triterpenes in ATCE to 27% (w/w), which lead to their enrichment in AMCE up to 98% (w/w). In both crude extracts, the most abundant neutral triterpenes were α and β amyrins. However, in AMCE α and β amyrins reached the 55 and 38% (w/w), respectively, whereas their concentration dropped to 12 and 9.5% (w/w) in ATCE extract. The oxidated amyrins, α and β amyrons, were found only in ATCE in concentration of 2.71 and 1.27% (w/w), while maniladiol and brein were the least abundant molecules among the three samples, respectively 1.6 and 2.1% w/w in AMCE and 0.94 and 1.28% in ATCE.

Identification and Quantification of Acidic Triterpenes
An HPLC-APCI HRMS 2 method was tailor-created to annotate the most abundant features among the acidic triterpenes. Molecules were identified in the instrumental raw data by their accurate mass-to-charge ratio, by their typical fragmentation pattern and using the specific analytical standards, when available. A total of 11 molecules were identified: 6 of them are tetracyclic triterpenes, whereas 5 are pentacyclic triterpenes. Identification was performed using specific analytical standards for oleanolic, ursolic and 3α-hydroxy-tirucalla-8,24-dien-21-oic acid (elemolic acid), while putative identification by exact mass and analysis of fragmentation pattern was carried out for the other triterpenes. In case of putatively annotated molecules, annotation level is the second, according to MSI [19]. Table 4 shows the identified or annotated compounds and their quantification, while Figure 2 shows the HPLC-HRMS chromatogram and molecular structures.
Qualitative analyses show that, in the samples, the most abundant compounds were always 3-oxo-tirucalla-8,24-dien-21-oic acid, 3α-hydroxy-tirucalla-8,24-dien-21-oic acid and 3β-hydroxy-tirucalla-8,24-dien-21-oic acid (respectively, elemonic acid, αand β-elemolic acid). However, the extraction process affected the detection of less abundant compounds in AMCE: gypsogenin, siaresinol, 3α-acetyl-tirucalla-8,24-dien-21-oic acid and 3β-acetyltirucalla-8,24-dien-21-oic acid were not detected in this sample. Quantitation data were obtained by using moronic acid as analytical standard to build an external calibration curve. This acid triterpene, from the oleanane group, was selected due to its similarity with the identified molecules in the samples, for the high purity grade of the available standard and for its absence in PHR matrices, making it possible to also use it as an internal standard. Qualitative and quantitative differences were reported for the assayed samples. The total content of acid triterpenes was 23.4% w/w in the raw material, the P. heptaphyllum oleum resin. However, the extraction process led to an enrichment of acidic triterpenes in the ATCE sample, since their total content reached 40.6% w/w, whereas, in AMCE, we observed a drastic decrease of their content (up to 3.0% w/w). Table 4. Acidic triterpene features annotation and quantitation.

ID#
Putative Name Formula MSI Level

VOC Qualitative and Quantitative Characterization
P. heptaphyllum oleum resin is rich in volatile terpenes, mainly monoterpenes. The chemical profiling of VOCs in oleum resin samples from different geographical origins [2] and aging stages [20] are available in literature studies. Therefore, the VOCs profiles are intended to be the most rugged chemical parameter to carry out taxonomical evaluation of plant material. However, VOC fraction exhibits a chemical composition that can be strongly altered with time; in particular, the increase of oxidated monoterpenes was observed [14]. In the present study, the sample of oleum resin contains (limited to the VOC fraction) the 95.7% of monoterpenes with predominance of p-cymene (21.85%), δ-3-carene (13.90%) and β-phellandrene (13.33%), whereas β-(E)-caryophyllene was the only sesquiterpene quantified.
The reported composition of volatile fraction matches P. heptaphyllum profile, although differences in chemical composition and abundance of VOC profile may be related to different factors, such as method of extraction (solvent extraction vs. essential oil), seasonal variations, plant collection site and environmental conditions (phenotypic plasticity) and collection strategy from naturally exudates or tapped trunks [2].

Triterpenes Identification and Quantitation
Neutral triterpenes were analyzed by GC-MS technique, which takes advantage of the electronic ionization for the detection and quantification of this poorly ionizable class of molecules, whereas the acidic triterpene molecules in the samples were investigated by HPLC coupled to APCI HRMS 2 . The APCI ionization source is capable of ensuring a higher sensitivity for lower volatile, poorly chargeable and highly lipophilic molecules with single acidic moieties [24].

Method Validation for Amyrin Quantification
The reported data show a validated quantitative method for α and β-amyrin that was developed according to FDA and EMA guidelines [18]. The gas chromatography validated method that we reported for the analysis of oleum resin and its extracts was fine-tuned in the concentration range with which they occur naturally to ensure a selective, repeatable and accurate quantification of these neutral bioactive triterpenoids. The proposed method can reach LOQ values ranging from 670 ppb for α-amyrin and 790 ppb for its β isomer up to 390 ppb and 350 ppb, respectively, using the weighted calibration. The use of a weighted calibration method reduced both LOD and LOQ by decreasing the ERR% parameter at 0.1 mg L −1 , thus enhancing the sensitivity of the quantitative method.
In literature data, only one validated quantitative method for α-and β-amyrin was reported for quantitation of amyrins in rat plasma [25]. In the reported case, the method was developed to achieve great sensitivity and selectivity using a SIM (Selected Ion Monitoring) scan and allowed for reaching very low LOQ and LOD values in the concentration order of 0.001 mg L −1 -1 ppb, as required by clinical analysis approaches that involve complex sample preparation, including purification and reconcentration steps.
Several literature studies report triterpenes quantitation methods in different food and botanical samples with similar validated quantitative approaches. In 2013, a GC-MS quantitative method for derivatized ursolic acid was reported in food supplement analysis with a LOQ of 10 mg kg −1 (10 ppm) [26]. Another work reports a HPLC-MS/MS validated method created for acidic triterpene analysis in botanicals and proposes LOQ values of 175 ng mL −1 (175 ppb), 108 ng mL −1 (108 ppb) and 85 ng mL −1 (85 ppb) for oleanolic acid, ursolic acid and betulinic acid, respectively. The lower quantitation limits in this latter case are related mostly to the higher sensitivity of the HPLC tandem mass spectrometry method proposed, made possible by the acidic nature of the analyzed molecules. A further study reports the validated method for the quantitation of triterpenes and sterols in the aerial part of Justicia anselliana. The method was validated for the lupeol triterpene; the authors chose to use GC-MS instrumentation for the identification and a GC-FID for the quantitation. The calibration curve was obtained with a weighted 1/X regression, and the method provided a LOQ of 5 mg L −1 (5 ppm). [27]. Therefore, the method proposed in this work has lower LOQ to the ones already reported with gas-chromatographic techniques, establishing a reliable and accurate quantitative method for the investigated matrices. Therefore, this is the first validated method for the quality control analysis of these specific phytochemical markers in plant raw material and extracts that can be easily adopted by dietary supplement industries.

Neutral Triterpenes Quantitation
Amyrins and amyrones isomers, brein and maniladiol profiling in PHR has been thoroughly investigated in several publications, mostly by GC-FID [8,11]. However, there are no literature data proposing a complete quantitation approach with a MS detector. The α and β-amyrin quantitation results in PHR are in line with previously published data obtained by Silva et al. [11], which compared seven Protium species by GC-FID. In P. heptaphyllum, the authors reported a concentration of 40.9% among α, β-amyrin and α-amyrone. Unfortunately, the authors did not record β-amyrone, brein or maniladiol, and limited information was reported regarding the quantification method that could allow an accurate comparison with our data. Other literature data have shown quantitative differences with respect to our data [8]. In detail, Siani et al. observed α-amyrin concentrations ranging from 30 to 43%; this difference may be linked to the different quantitative approaches that the authors used. Siani et al. reported the use of the relative peak area by a GC-FID, while our work employs a specific calibration curve performed on αand β-amyrin. Due to the relevant differences in the quantitative methods that were reported in literature, the abundance ratio between ursanic and oleanic triterpenes could be used to compare our results with data obtained by previous studies. In our work, the ursanic and oleanic triterpenes ratio is 1.3 in all analyzed samples (PHR, ATCE and AMCE). The observed value is slightly higher than Silva et al.  [12]. Different ursanic-oleanic abundance ratios found by Neves et al. in their GC-MS quantitation experiment could be also explained by using α-amyrin as a unique calibration standard also applied to β-amyrin quantitation. In fact, data in our present work show that α-amyrin MS response factor (218 m/z quantifier ion) is substantially higher (1.2-fold) than its oleanic isomer. The observed difference in response factor could be responsible for β-amyrin underestimation and change in ursanic and oleanic triterpenes ratio if titrated by α-amyrin equivalents. Furthermore, data comparison confirms that significant natural variability between PHR samples can occur also in αand β-amyrin content related to different factors, such as seasonal variations, plant collection site and environmental conditions (phenotypic plasticity) [28]. Therefore, the reported data highlight the importance of using a validated quantitative method to perform accurate comparisons and the relevance of the described GC-MS approach.

Acidic Triterpenic Fraction Identification and Quantification
The proprietary extraction process for ATCE and AMCE production allows for the concentrate of acidic triterpenes in ATCE extract and the amyrins and other neutral oleanic and ursanic triterpenes in AMCE, respectively. The obtained data are the first quantitative report of the acidic triterpenes in P. heptaphyllum oleum resin and its derivates using a HPLC-APCI HRMS mass spectrometry approach. The use of a HPLC-based technique has a great advantage for this class of molecules, since does not need any derivatization or sample pretreatment. In contrast, gas-chromatographic approaches for the analysis of more polar analytes or ionized groups, such as carboxyl, need derivatization steps by formation of methyl esters or trimethyl siloxanes in order to lower their boiling point and reduce possible interactions with stationary phase [8,28].
Interestingly, the peaks #4 and #6 (Figure 2a) show similar fragmentation pattern with respect to peak #5. We attempted a fragmentation-led recognition also referring to elder publication data [8]. In their paper, Siani et al. reported the presence of 3α-hydroxytirucalla-7,24-dien-21-oic acid, which is a B ring isomer of 3α-hydroxy-tirucalla-8,24-dien-21-oic acid with the double bond shifted from position C8 to position C7, and an epimer of 3α-hydroxy-tirucalla-8,24-dien-21-oic acid. Those molecules were also elucidated by crystal structure analysis [30,31]. The main difference in the spectral fragmentation of peak #4 with respect to peaks #5 and #6 is the presence of 339 m/z ion instead of 441 m/z. This 2 Dalton shift can suggest the different ring rearrangement due to a different position of the double bond. In particular, we assumed that the presence of the double bond in position 7 was more likely to form an aromatic rearrangement in ring B upon fragmentation, while the double bond in position 8 formed a less stable fragment with a nonaromatic ring. To further confirm this hypothesis, we found that 339 m/z ion in peak #4 had a higher abundance than 441 m/z ion in peaks #5 and #6. Finally, peak #6 showed the same fragmentation pattern of peak #5 and could be attributed to its C-3 epimer. In addition, 3-oxo -tirucalla-8,24-dien-21-oic acid, a tirucalladienoic acid oxidized derivative was identified by the use of its analytical standard and mass fragmentation interpretation. With respect to the work of Siani et al. [8], we found different relative abundances among these molecules. This can be related to 0the plasticity of metabolic profiles, which is typical in plants.

Reported Terpenes Biologic Activities
p-Cymene is the major constituent in VOC fraction in both PHR and ATCE. It is found in more than 100 plants belong to Lamiaceae, Mirtaceae, Burseraceae (such as Protium) and Asteraceae genus. This secondary metabolite has relevant antioxidant, anti-inflammatory, antinociceptive, anxiolytic and antimicrobial activities [34,35]. B-Phellandrene, 3-carene and α-pinene have reported antibacterial and antifungal activities by disrupting the pathogen membrane structure and leading to energy metabolism dysfunctions in pathogen microorganisms [36][37][38]. β-(E)-caryophyllene is naturally present in several plant species and marketed as extracts due to its relevant commercial value. β-(E)-caryophyllene is defined as a dietary cannabinoid, due to its selective agonist activity against CB 2 receptor on the CP55,940 binding site. Agonist ligands of CB 2 receptors have been shown to inhibit inflammation and edema formation, to exhibit analgesic effects and to play a protective role in hepatic ischemia-reperfusion injury [39,40].
Triterpenoids have found to carry several beneficial health effects in vitro and in vivo [41]. The α and β-amyrins from P. heptaphyllum oleum resin showed antinociception in mice, in which orofacial pain was induced by capsaicin and formalin subcutaneous injection [42,43] or visceral pain, which was induced by cyclophosphamide and mustard oil (colonic inflammation) [44]. Putative interaction between α and β-amyrins and GABA-A receptor was also investigated in vivo, leading to the hypothesis of antidepressant effects carried out by PHR oral administration [6]. Anti-inflammatory effects were investigated in gastric ulcer model by oral administration in mice of oleum resin. The PHR, at 200 and 400 mg kg −1 b.w., significantly attenuated the gastric damage induced by both ethanol and acidified ethanol [45]. In periodontitis-induced inflammation, αand β-amyrins pretreatment at 5 mg kg −1 significant reduced TNF-alpha, the gingival myeloperoxidase and thiobarbituric acid-reactive substances, retarding acute inflammation in rat model [46]. Moreover, several publications reported the impact of PHR on regulation of lipidemic and cholesterolemic levels. PHR administration showed a body weight reduction in HFD (High Fat Diet)-treated mice with a sibutramine-comparable efficiency; in addition, ghrelin decrease and leptin increase (typical in HFD control mice) were successfully modulated [45,47].
Several acidic triterpenoids have been found to carry beneficial health effects in human and animal models, such as anti-inflammatory, anti-microbial and antiviral agents. In HIV-1 infected H9 lymphocytes, it was reported as an in vitro HIV-protease inhibition by maslinic acid [48]. Moreover, oleanolic, ursolic and tirucallic acids have been long-recognized to have anti-inflammatory and antihyperlipidemic properties in vivo test [49,50], as well as 3-oxo -tirucalla-8,24-dien-21-oic acid, which also reports anti-proliferative activities [51]. In a recent study, oleanolic and ursolic acid have also shown in silico putative activity against the main protease of the SARS-CoV-2 responsible for coronavirus disease (COVID-19), suggesting further investigations for their potential antiviral activity [52][53][54].

Samples of Protium Heptaphyllum
Samples that were used in this study are listed and described in Table 5. The samples of oleum resin and crude extracts were from Protium heptaphyllum (Aubl.) Marchand was provided by Abel Nutraceuticals (Turin, Italy). The oleum resin was collected in Brazil in April 2019. The samples of crude extracts were obtained by Abel Nutraceuticals (Turin, Italy) (Batches # P75-2-0 and P75-2-1 used under the commercial brand Hepamyr ® ) through a patent pending hydroalcoholic extraction process and supplied in a powder form.

Extraction of VOC and Neutral Triterpenes
Neutral triterpenes and volatile organic compounds (VOC) extraction from PHR, ATCE and AMCE was carried out with DEE. This solvent was chosen to include the widest range of molecule polarity. An aliquot of 20 mg of powdered samples was spiked with the internal standard (M-TEST), according to Ching et al. [25], to obtain a final concentration of 3 mg L −1 of M-TEST in the analyzed solution. Extraction was carried out with 20 mL DEE. The solution was bath-sonicated for 20 min, and the undissolved particles were precipitated by centrifugation at 8000× g for 5 min. The solution was then diluted in DCM or HEX, 1:50 or 1:10 depending on the matrix (1:50 for oleum resin and extract, 1:10 for GC-FID quantitation of VOC fraction) before analysis.

Extraction of Acidic Triterpenes
Extraction of acidic triterpene from PHR, ATCE and AMCE was performed with EtOH as follows: 50 mg of powdered sample were extracted with 10 mL of EtOH, then the solution was bath-sonicated for 20 min and precipitated by centrifugation at 8000× g for 5 min and finally diluted with MeOH.

VOC Profiling and Quantitation
Volatile terpenes were analyzed with a GC-FID/MS instrument (GCMS-QP2010 SE-Shimadzu, Japan). The VOCs were firstly identified with the MS detector and then quantified with the FID. The analyses were carried out with the same chromatographic condition for GC-MS and GC-FID. Samples (1 µL) were injected in an injection port kept at 280 • C in split mode (GC-FID) (Split Ratio 1:5) and splitless mode (GC-MS). Split flow was restored after injection at 3 mL min −1 . The instrument used helium as carrier gas (column flow 1 mL min −1 ) and was equipped with a capillary column: Restek RXI 5-MS (30 m × 0.25 mm × 25 µm film thickness). The thermal gradient was set as follows: 0 50 • C, 30 140 • C, 45 320 • C, 55 320 • C. Mass spectrometry analysis was performed with an EI ionization source set at 70 eV, source temp: 200 • C, quadrupole temp: 150 • C, and the detector scan ranged between 50 and 650 m/z. GC-MS and GC-FID data were handled and managed with LabSolution (Shimadzu, Japan). MS analytes identification was achieved by NIST17 and FFNSC3 Shimadzu Mass Spectra libraries comparison and also by comparison of calculated linear retention indexes (LRIc) obtained with a C 7 -C 30 n-alkanes mix with literature LRI (LRIL) [16]. Retention times and LRI were then used to identify the peaks in GC-FID chromatograms for quantitation. Quantitation was performed by external calibration with a standard mixture of VOC terpenes (Mix Cannabis Terpenes A-B, Sigma-Aldrich, see chemicals). Terpenes in sample which were not present in the standard mix were quantified with the nearest-eluting molecules in the standard.

Neutral Triterpenes GC-MS Detection and Quantification
The quantification of neutral triterpenes was carried out by a GC-MS. The instrument was composed by a TRACE 1310 and a TSQ Quantum Ultra QqQ (Thermo Scientific). Analysis was carried out by a temperature gradient, and detection was made in full-mass mode (50-

Quantitation Method Validation
Method validation was performed in order to assess at first its repeatability and overtime stability; parameters studied were overtime linearity and repeatability (DIFF%), selectivity (SEL%), sensitivity in terms of LOD (limit of detection) and LOQ (limit of quantitation), percent standard deviation (RSD%) and imprecision (ERR%). Recovery was calculated respectively to the recovery of internal standard added in sample. Calculation equations used for validation are reported in supplementary materials ( Figure S1).

Conclusions
The numerous reports describing and demonstrating the biological activities of triterpenes from Protium heptaphyllum and other plant sources are driving a growing interest in exploiting the biologic potential of triterpene in this botanical source. However, natural chemical variability of oleum resin, that can be correlated to seasonal variations, plant collection site and environmental conditions, impairs the investigation and comparison of in vitro and in vivo mechanisms of action and the correlation dose/effects, as well as assessments of safety and bioavailability.
A combined analytical approach based on different chromatographic techniques was proposed to face the issue of chemical identification and quantification, as well as standardization in P. heptaphyllum oleum resin extracts. GC with FID or MS detectors was used for the identification and the quantification of VOCs and neutral triterpenes. A faster and more reliable solution has been developed and validated for bioactive neutral triterpenes in P. heptaphyllum oleum resin and two fractionated extracts enriched in acid (ATCE) and neutral triterpenes (AMCE) that can be of valuable interest for food supplements applications. Moreover, for the first time, an untargeted method was tailor-created to annotate, identify and quantify the most abundant compounds among acid triterpenes in P. heptaphyllum oleum resin and the two crude extracts by HPLC-HRMS 2 .
A total of 29 molecules were identified and quantified in both resin and extracts, covering the 70% by weight of chemical characterization in P. heptaphyllum oleum resin without need for any derivatization protocols for any of the identified molecules. The methods that allowed for the qualitative and qualitative chemical profile obtained in this work represent a solid starting point for the varietal chemical characterization of other Protium species. Furthermore, the application of the developed methods and the use of accurate data on resin composition in future studies will allow a clearer correlation on observed biological activities.

Patents
Patent related to part of this work is pending, application number 102020000015598.
Supplementary Materials: The following are available online, Table S1: Calibration curves and relative recalculation errors title, Table S2: Weighted calibration curves recalculation errors, Figure S1: Calculating validation parameters, Figure

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available because a patent related to a part of this work is pending (application number 102020000015598).

Acknowledgments:
The authors are grateful to Regione Piemonte for supporting the activity of the Ph.D. Program in Apprenticeship of Alberto Asteggiano. Moreover, the University of Turin is thankful to Abel Nutraceuticals S.r.l. for providing sample material free of charge. The authors are grateful to Alessandra Porcu for support in the preparation and traceability of crude extracts.

Conflicts of Interest:
A.O. and A.C. were employed by Abel Nutraceuticals S.r.l.; A.A. was a Ph.D. student of the Ph.D. Program in Chemical and Material Sciences of the University of Turin (XXXVI cycle) of the University of Turin and in apprenticeship in Biosfered S.r.l. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
Sample Availability: Not available.