Characterization of Larix decidua Mill. (Pinaceae) oleoresin’s essential oils composition using GC-MS

Introduction Larch oleoresin has been described regarding several biological activities and medicinal applications, such as wound healing and treatment of ulcers, but little is known about its chemical composition. Material and methods Eight oleoresins from Larix decidua Mill. obtained from four companies and one adulterated control were therefore investigated to determine their content of essential oils and to verify possible differences in their composition in relation to the harvest and manufacturing processes. Essential oils (EOs) were isolated by distillation and the yield was analysed. Results and discussion The yield of EO varied among all samples. The yield of the pure larch samples covered a range of 7.8% to 15.5%. A higher yield (19.0%) was observed for adulterated control, which contained oleoresins from different Pinaceae trees. Age of samples had no impact on yield. However, there was a significant statistical variation (p<0.05) in the yields of the mid-summer oleoresins (>10%) compared to early or late summer (<10%), emphasising the importance of the time of collection. Samples were subsequently analysed by GC-MS. EO samples confirmed the presence of various chemical classes, such as monoterpenes, sesquiterpenes, and diterpenes. α-pinene was the compound with the highest concentrations (>50%), followed by β-pinene (>6%), D-limonene (>2.5%), α-terpineol (>0.9%), β-myrcene (>0.2%), and 3-carene (>0.05%). Samples were grouped using multivariate data analysis (MVDA) with respect to the chemical variation between the oleoresins’ EOs. The resulting four clusters were named low (low yield obtained for the samples), mixed (mixed oleoresin from different Pinaceae species, adulteration control), old (old oleoresin kept in the institute), and normal (other oleoresins) samples, each presenting distinct chemical biomarkers. There were considerable differences between site and time of collection. Essential oil yield did not always meet requirements as defined by the German Homeopathic Pharmacopoeia. In addition, adulterated or aged samples could be identified as compared to pure and fresh larch oleoresins. Conclusion We conclude that larch oleoresin used for pharmaceutical applications has to be carefully analysed and standardised to guarantee reproducible product quality.


Introduction
Natural products have long been an essential source of new drugs against various diseases, leading to the discovery of several natural antibiotics, and are equally employed successfully as cancer therapeutics (Atanasov et al., 2021;Dzobo, 2022).Due to the complex mixture of different compounds, there is a potential for synergistic therapeutic effects (Atanasov et al., 2021).Essential oils (EOs) are aromatic and volatile substances extracted from various parts of plants, such as leaves, flowers, fruits, and even bark (Harrewijn et al., 2000;Tongnuanchan and Benjakul, 2014) and have been used for centuries for their therapeutic, cosmetic, and culinary properties (Figueiredo et al., 2008;Sharmeen et al., 2021).Among the diverse sources of EOs, conifers stand out as a remarkable botanical group (Franz and Novak, 2015).
Larix decidua Mill.(Pinaceae), commonly known as the European larch (Tropicos, 2023), is a deciduous (The_World_Flora, 2023), coniferous tree with delicate foliage, which occurs in the central (Alps) and eastern mountains of Europe (Da Ronch et al., 2016).Beyond its striking appearance and ecological significance, this tree holds a valuable secret within its resinous sap, known as oleoresin, commonly referred to as turpentine (Dietemann et al., 2019).This oleoresin, rich in useful compounds, has been employed in various industries, from traditional medicine to perfumery and beyond (Lagoni, 2012).Our recent review showed that a lot is known about the species' phytochemical composition, but only little is known about the oleoresin's composition and biological properties (Batista et al., 2022).
Plant extracts are complex mixtures of compounds and need to be analysed and identified to monitor the quality of the sample and its identity.In this work, we investigated the yield and differences in composition of the EOs in eight different larch oleoresin batches and in one adulterated sample used as control.Gas chromatography tandem mass spectrometry enabled the separation and identification of volatile organic compounds, which were grouped using multivariate data analysis to identify preparations, which comply with Pharmacopeia and product quality requirements.

Hydrodistillation
The essential oils (EOs) were obtained following the German Homeopathic Pharmacopeia method (HAB, 2014).The oleoresins were submitted to hydrodistillation in a 500 mL round flask with 200 mL of distilled water and boiling pebbles (sort A, ROTH) for 2 h, and a Clevenger-type apparatus was used (as in Ph.Eur.2.8.12), with distillation at a rate of 3 mL/min.The procedure was performed using xylol (0.5 mL) in the graduation of the apparatus.The EO yield was determined by sight using the graduation of the apparatus as a subtraction of the final volume and the initial one (0.5 mL).The distillates were stored in a glasssealed vial at 4°C until chromatographic analysis.

GC-MS analyses
The GC-MS analysis was performed in a Shimadzu GCMS-QP2010 SE (Shimadzu, Ireland).A capillary column ZB-5Plus of 30 m x 0.25 mm x 0.25 µm was used for the separation.The GC parameters were as follows: the carrier gas was highly pure helium with a 1 mL/min flow rate.The inlet temperature was 250°C with a split ratio of 20:1 and the pressure was 49.7 kPa.The column oven temperature was initially set at 40°C for 1 min, and then ramped to 290°C at 5°C/min, and kept at 290°C for 5 min.
MS parameters were as follows: data were acquired in the electron impact (EI) mode, using the full scan mode from m/z 40 to 750.The ion source and interface temperatures were 200°C and 300°C, respectively.The identification of the volatile compounds was based on a comparison of their GC retention time and mass spectra with the retention index of n-saturated alkanes and the reference spectra from the US National Institute of Standards and Technology (NIST, 2023).Data was analysed by Shimadzu LabSolution Postrun software.

Statistical analysis
The GC-MS data of the volatiles was analysed using multivariate data analysis (MVDA) to group samples with respect to the chemical variation between the nine oleoresins' essential oils.Relative abundance (% area) of each compound were calculated based on the ratio between the peak area of each compound and the sum of all integrated compounds.The data of the MVDA was exported to Metaboanalyst 5.0 web server to observe how the samples are clustered.Firstly, unsupervised analysis was done by hierarchical cluster analysis (HCA) evaluated by Euclidian distance dissimilarity using the aggregation criterion of Ward's method and by principal component analysis (PCA).Afterwards, a supervised analysis was done by the partial least squares discriminant analysis (PLS-DA) to examine the separation between the groups and to better comprehend the variables responsible for classification (Melo et al., 2022).Analysis of variance (ANOVA) followed by Tukey's post hoc test was performed, using the same web server for the boxplot analysis.Differences were considered significant when p<0.05.

Distillation process
The distillation process using a Clevenger-type apparatus allowed not only to obtain the EOs but also provided the yield of EO, an important information for the oleoresin's quality control.Table 2 shows the essential oils yield obtained for each of the nine samples of oleoresin.
Other larch parts were investigated for their VOC (Kubeczka and Schultze, 1987;Weissmann and Reck, 1987;Holm and Hiltunen, 1997;Wajs et al., 2007;Garcia et al., 2017;Mofikoya et al., 2020;Visan et al., 2021).Weissmann and Reck (1987) obtained a similar composition of oleoresin VOC as we measured in our samples, with a-pinene (2) (76.4%), b-pinene (7) (7.7%), bmyrcene (8) (5.4%), 3-carene (10) (4.3%) as the major compounds.Although our results were similar to Weissmann and Reck (1987), the advantage of our study is a higher number of samples analysed, describing small variabilities within the same species.Kubeczka and Schultze (1987) examined three tree parts (needles, wood, bark) and observed a variability of their VOCs: a-pinene (2) was more concentrated in the wood (44.72%), followed by the bark (38.34%) and the needles (28.57%).On the contrary, 3-carene (10) was observed in higher concentration in the needles (19.19%), followed by the bark (4.80%) and the wood (2.78%).3carene (10) was observed to be in higher concentration in the needles, as described by Kubeczka and Schultze (1987) (19.19%),Holm and Hiltunen (1997) (5.8-21.6%),and (Weissmann and Reck, 1987) (24.2%), and therefore could be used as a chemical marker for the VOCs from L. decidua's needles.Holm and Hiltunen (1997) described that the monoterpenes composition of leaf oils could be used as marker for genetic research and to edit issues related to population genetics of Larix species, which are characterised by high contents of a-pinene (2) and 3-carene (10).In conclusion, there are considerable differences with respect to VOC composition in different plant parts.Therefore, L. decidua oleoresins cannot be substituted by other plant parts to isolate their EOs.

Profiles of multivariate analyses
Unsupervised and supervised multivariate analyses were conducted to group and classify the differences between the analysed samples.After pre-processing and data normalisation, the final dataset used for the multivariate analysis consisted of 27 EOs samples x 74 features (relative abundance as % area).
Firstly, hierarchical cluster analysis (HCA) revealed four clearly defined groups (Figure 2).Group 1 was named low (low yield samples), group 2 mixed (mixed species oleoresin), group 3 old (the old sample kept in the institute), and group normal (oleoresins samples within specification).The corresponding colour-coding is defined and used in Figure 2 and Table 3. Sample definitions are provided in sections 2.1 and 3.1.We then carried out a more detailed investigation with principal component analysis (PCA) to analyse the chemical pattern of these different groups.
Unsupervised PCA was applied to assess the composition of 9 oleoresin EOs and to identify a possible correlation between the various samples.The PCA score plot shows that principal component 1 (32.5%),principal component 2 (20.9%), and principal component 3 (12.4%)explained 65.8% of the data variance, which can reflect most of the information of the original data of the sample.The results of the PCA revealed four distinct groups (Supplementary Figure 2), corroborated by HCA analysis.The score plot demonstrates that the "mixed" group is composed only of one oleoresin EO, BUVT, the oleoresin named Venetian Turpentine composed of Larix decidua, Abies alba, Pinus pinaster and Picea excelsa's oleoresins.The cluster "old" is formed only by one oleoresin EO (Ha), the oldest oleoresin studied, collected in the 1990s.The "low" group is characterised by two oleoresins' EO, BU1 and BU3, which present the lowest EO yield and collection in the same year, 2020.The "normal" group comprises BU2, BU4, H, S, and R3, young oleoresins (collected 2018-2021) within the range of accepted EO yield.We concluded that PCA allows for a meaningful grouping of oleoresin samples based on their chemical fingerprint.
To understand the primary chemical compounds responsible for the initial separation observed in the PCA, a discriminant analysis (PLS-DA) was conducted to identify the main chemical constituents correlated to the clustering pattern observed in the scores plot through the 25 more critical variables in the projection (VIP).In Figure 3, for the old group (Ha), considering the VIP values, a-pinene (2) (VIP 1.69) and verbenone (29) (VIP 1.5) were the compounds with higher intensity in this group, followed by isopinocarveol (20).Concerning the normal group, the most important compounds for its differentiation are a-cubebene (41) (VIP 1.89) and g-elemene (50) (VIP 1.82).The mixed group, composed of BUVT only, presented high intensity of the humulene (51) (VIP 2.05), copaene (45) (VIP 1.43), caryophyllene oxide (64) (VIP 1.26) and a-terpinolene (16) (VIP < 1.2).Lastly, the low group presented 17 high-intensity compounds among the 25 main VIPs in which the d-elemene (38) (VIP 2.09) and b-elemene (46) (VIP 1.94) were the most important.We conclude that these most important compounds represent principal components, which can be used for grouping.This information is summarised in Figure 3. Consequently, monitoring of the nine most important compounds with VIP scores higher than 1.6 is sufficient to reliably group oleoresins and to detect, for example, adulterated or old samples.
The question arises if, besides the statistical PCA approach, additional grouping markers can be defined based on biochemical considerations.In addition to a-pinene (2), verbenone (29) and isopinocarveol (20), which were more intense in the old group (Supplementary Figure 3), trans-carveol (30) could be used as a marker for the ageing or degradation of this species' oleoresin EO (Figure 4).Supporting the idea of the degradation process, verbenone (29) and isopinocarveol (20) are degradation products of a-pinene (2) (Schrader et al., 2001), and D-limonene ( 14) was obtained in a low concentration in the old group, which is oxidised to trans-carveol (30) (Bouwmeester et al., 1998), found in higher amounts for this group (Figure 4), an indication of a degradation reaction (Figure 4).These oxidation reactions in the oleoresin may be related to daylight radiation and/or temperature influence in the storage process (Schrader et al., 2001).
For the mixed group (BUVT), 1,4-cineole (11) and alongipinene (42) were present only in this group, the last previously described for Abies alba oleoresin (Zeneli et al., 2001).Although more samples should be considered to prove this idea, these compounds could be classed as adulterants of L. decidua oleoresin EOs since they were not found in pure samples nor described in the literature.b-pinene (7) was present in a higher concentration than the other groups (Figure 4), which the influence of different species can explain.Oleoresins from Abies alba presented similar proportions for aand b-pinene (Zeneli et al., 2001) and equivalent amounts of b-pinene (7) (17.53-18.91%)were found for Pinus pinaster (Arrabal et al., 2002).Therefore, a higher concentration of b-pinene (7) for the mixed oleoresin is explained by the influence of the other species' oleoresins in the sample (Figure 4).As previously discussed (section 3.2), 3-carene (10) is a vital chemical marker for the Larix oleoresin EOs.The PCA verified that a low concentration in the mixed group was obtained (Figure 4), resulting from an absence of this compound in Abies alba and Pinus pinaster oleoresins (Zeneli et al., 2001;Arrabal et al., 2002;Arrabal et al., 2005).
The compound in higher concentration in all samples was apinene (2) (VIP 1.69) but presented differences between the groups.Data (Figure 4) shows that its concentration decreased from the old, normal, mixed to the low group, statistically significant (p<0.05)except for the normal and mixed groups.a-pinene (2), a bicyclic monoterpene, is generated by the cyclization of geranyl pyrophosphate (GPP) by monoterpene synthases, specifically pinene synthases (I, II, III), responsible for the different stereochemistry (Loza-Tavera, 1999).It is found primarily in pine trees (coniferous) EOs and is the main secondary metabolite in many conifer-derived EOs, the one responsible for the characteristic smell of pine trees (Salehi et al., 2019;Allenspach and Steuer, 2021;Nyamwihura and Ogungbe, 2022).VOCs, such as a-pinene (2) and b-pinene (7), possess influence on plants defences, working as plant-to-plant signalling, leading to a systematic acquired resistance Dendrogram (hierarchical cluster analysis) representing the relationship of 27 oleoresins' essential oils obtained by Euclidian distance dissimilarity using the aggregation criterion of Ward's method.Four main groups of samples (normal: normal composition according to specification; old: old sample; mixed: mixed species origin; low: low yield samples) were defined and colour-coded (represented by coloured arrows).The last digit of the sample code denotes the sample replicate number (1-3).(Riedlmeier et al., 2017) and helping plants to communicate and to fight against parasites, such as fungi and bacteria (Nyamwihura and Ogungbe, 2022).Their medical properties are described for several purposes since they possess therapeutic potential as anticoagulant, antitumoral, gastroprotective, anxiolytic, neuroprotective, antimicrobial, antimalarial, insecticidal and larvicidal, antifungal, anti-inflammatory, analgesic products, among others (Salehi et al., 2019;Allenspach and Steuer, 2021;Nyamwihura and Ogungbe, 2022).
Other VOCs that also appear to bear some importance in the analysed oleoresins are D-limonene ( 14), a-terpineol (27), bmyrcene (8), and 3-carene (10).They are not only important for the typical conifer fragrance, but are involved in intraspecific hostfinding pheromones communication, play a major defensive role against insects and pathogens, and are important for cultures due to economic reasons and pharmacological properties (Langenheim, 2003).An interesting review compared the activity of D-limonene (14) and perillyl alcohol, a hydroxylated analogue of D-limonene (14), on breast cancer in human trials.They concluded among 5 studies that D-limonene ( 14) possessed better tolerability and chemopreventive properties than the perillyl alcohol, but further well-designed studies should be carried out (Chebet et al., 2021).A recent study described the potential of D-limonene ( 14) as anti-SARS-CoV-2 candidate, since it possesses similarities in structure with the thymidine of SARS-CoV-2 genome and low cytotoxic effects in MRC-5 (fibroblast) and HaCaT (keratinocyte) cell lines (Correa et al., 2023).Although several in vivo studies have described the potential of D-limonene (14), limited data exists for its tolerability and safety in humans (Anandakumar et al., 2021).Khaleel et al. (2018) described in a review several biological properties of a-terpineol (27), such as antihypertensive, antiproliferative, antiulcer and insecticidal.The most important activity correlated to a-terpineol (27) is its anti-nociceptive activity, with highly analgesic effects in mice, mainly due to inhibition of pro-inflammatory molecules release.Oil from Eucalyptus globulus as well as a-terpineol (27) demonstrated anti-parasitic effects against Pediculus humanus capitis, an ectoparasite confined in human scalp and hair (Yang et al., 2004).In an in vitro study, aterpineol ( 27) inhibited the growth and induced cell death in tumour cells via inhibition of NF-kB activity, among other mechanisms (Hassan et al., 2010).A recent publication described that the biological properties of b-myrcene (8) are coupled with its non-allergic, non-toxic and antimutagenic activities.It has anxiolytic and sedative effect; it acts as an antioxidant agent, which is accountable for prevention of ageing and degenerative diseases; its powerful anti-inflammatory activity in vitro lies mainly through PGE-2; the analgesic effects are central and peripheral (Surendran et al., 2021).In addition, McDougall and McKenna (2022) demonstrated in vivo the reduction of joint pain and inflammation in rats, suggesting the potential of b-myrcene (8) to reduce chronic arthritis pain and inflammation.Lastly, 3-carene (10) was proven to be the most prominent agent against Variables important projection (VIPs) from PLS-DA of L. decidua oleoresins' essential oils.The concentration was standardised at a range scale before analysis.Alpha-pinene is the most abundant volatile organic compound.Monitoring of the nine most important compounds (VIP scores > 1.6) is recommended to reliably group oleoresins.Groups of samples: normal: normal composition according to specification; old: old sample; mixed: mixed species origin; low: low yield samples (see Figure 2).dermatophytes and could be used as an antifungal compound (Cavaleiro et al., 2006).In addition, another study showed that 3carene (10), among other compounds, possessed the broadest spectrum of activity against fungi and gram-positive bacteria, which could be used as antimicrobial agent and to prevent aflatoxin contamination in foods (Cosentino et al., 2003).
Keeping the oleoresins in closed packages at room temperature for up to three years prevented degradation.Thus, storage under these conditions does not seem to influence the quality of the samples.Further studies should be performed to verify when degradation starts as a function of storage methods.It remains to be elucidated to which degree these metabolites change in relation to the time of collection, geographical location, and seasonal variation.We conclude that further analysis is necessary to decide whether they are, with exception of a-pinene (2), reliable markers for grouping, such as evaluated, into "normal" (normal composition according to specification), "old" (old sample), "mixed" (mixed species origin), and "low" (low yield samples).

Conclusion
Chemical fingerprinting based on GC-MS analysis is a prerequisite to group oleoresins and to detect preparations, which are not suitable for pharmaceutical applications.In the present work, strategies are provided to carry out this task.Chemical variances in essential oils observed in nine samples of oleoresins obtained from four companies and collection sites show the importance of standardisation and storage to guarantee reproducible chemical composition in production batches.Information on geographic location and collection date is mandatory.Care should be taken to avoid preparations adulterated by addition of volatile organic compounds from preparations other than oleoresins or other plant species.In addition, we identified for the first time a-pinene-oxide, aphellandren-8-ol, d-elemene, cyclosative, ylangene, and butylated hydroxytoluene in Larix oleoresin and suggested possible adulterants (1,4-cineole and a-longipinene) and compounds related to ageing (trans-carveol).The question arises if alternative analytical technologies could be used to increase the number of detected metabolites.With this respect, headspace solid-phase microextraction would avoid potential loss of compound during hydrodistillation and accelerate the analytical procedure.

FIGURE 4
FIGURE 4Selected chemical markers which can be used for grouping (normal: normal composition according to specification; old: old sample; mixed: mixed species origin; low: low yield samples, see Figure2) based on biochemical considerations.Box plot representation of normalised concentrations of the respective compound and each group.Box-plot values represent median, 25% and 75% quartiles, and whiskers for minimum and maximum values.*p<0.05one-way ANOVA with Tukey's post-hoc analysis.

TABLE 1
List of oleoresins analysed and their collection information.
Ha is an old oleoresin sample kept in our institute, dated 1994.BUVT, called Venice Turpentine, is a mixture of oleoresins from several species, such as Larix decidua, Abies alba, Pinus pinaster and Picea excelsa.nd = not declared / unknown.

TABLE 2
Yield of essential oil in each sample of oleoresin after distillation (mean ± SD, n=3).

TABLE 3
Chemical composition in percentage (%) of the investigated Larix decidua oleoresin essential oils (as well as those from the adulteration control BUVT) obtained by gas chromatography coupled with mass spectrometry (mean±SD; n=3).

TABLE 3 Continued
a Odour descriptions were from Flavornet (www.flavornet.org).N.s.Not smelled.bCompoundsidentification:RI and mass spectra mass spectra (MS) data compared against commercially available MS library NIST 23. tr, trace (<0.05%).cRI:retentionindex experimentally determined on a ZB-5Plus column relative to the Rt of n-alkanes (C7-C40); compounds are listed in order of elution.Columns are colour coded according to group assignments defined in Figure2.Dark-blue represents normal samples, light-blue represents the old sample, red represents low yield samples, green represents the mixed species sample (group definition see section 3.3).