Aromas from Quebec. V. Essential oils from the fruits and stems of Heracleum maximum Bartram and their unsaturated aliphaticacetates

Abstract Both the fruits and stems of Heracleum maximum Bartram, a common Apiaceae found throughout northern North America, have been traditionally consumed as spice and food by American First Nations and settlers. As both parts of the plant bear a distinctive scent, they have been submitted to volatile constituent extraction and their essential oils were studied by GC–FID and GC–MS. The fruits’ oil was found to be mostly constituted of aliphatic esters, with octyl acetate (65.6%) as main constituent, while the stems yielded a generally terpenic oil, dominated by limonene (45.2%). These results were in line with data for other Heracleum species, and constitute the first study of the volatile constituents of H. maximum. The observation of several octadecenyl acetates as trace compounds in the stems’ oil allowed for the comprehensive identification of these uncommon and closely related isomers.


Introduction
Heracleum maximum Bartram (formerly Heracleum lanatum Michaux), a member of the Apiaceae, is a large herbaceous plant found across North America, and is commonly known as cow parsnip. Its sturdy, hollow stem reaches up to 3 meters, bearing a flattened inflorescence up to 20 cm wide. The fruits are winged and ridged schizocarps. The plant's leaves bear three maple-shaped leaflets and their petioles are wrapped around the stalk (1)(2)(3). The species has had dietary uses: many native American populations consumed large quantities of the young and tender stalks as raw vegetables, hence another common name, Indian celery, used in soups and desserts (2,4); the fruits have been also used as spices (1); parts of the stem have also served as a salt substitute (1,4); and the Meskwaki people used the roots as food (1).
Previous studies have also found it to possess antibacterial and antifungal (5,6), immunostimulant (2) and antidiabetic (7) properties. As many other Apiaceae, cow parsnip is known to produce a wide range of phototoxic furanocoumarins (2,(8)(9)(10), some of which are responsible for the antimycobacterial activity along with (3R,8S)-falcarindiol (11). The furanocoumarins are partly removed by peeling the stalks, a process largely applied by the natives to avoid undesirable toxicity (5).
The stem and fruits both possess a distinct scent (1,2,4), which likely contributed to their dietary uses. As the volatile compounds of H. maximum have not been studied before, specimens from Saguenay (Quebec, Canada) were picked in order to analyze the composition of the essential oil obtained from the stems and the fruits.
Although hydrodiffusion of an essential oil is not the best way to appreciate the amounts of furanocoumarins, the presence of three of them (isobergapten, pimpinellin, and imperatorin) in trace amounts is not surprising. Pimpinellin and isobergapten were previously found in H. lanatum leaves (8) and H. maximum roots (11). Along with other furanocoumarins, they were too extracted with petroleum ether from H. crenatifolium (17). Angelicin and bergapten were also observed in the fruit oil of H. paphlagonicum (12).
One has to note the presence of very small percentages (<1%) of mono-and sesquiterpenes such as limonene (0.8%), germacrene D (0.3%) and amorpha-4,11-diene (0.3%), with only myrcene reaching a content of 1.6%. Terpinolene (10.8%) is one of the most important measured directly in the collecting burette. The oils were stored at 4°C until GC analysis.

GC-FID and GC-MS analysis
GC-FID analyses were carried out on an Agilent 6890N GC equipped with a split/splitless injector as well as two FID detectors. Compounds were identified from their retention indexes as calculated from even-numbered C 8 to C 36 alkane standards and/or from MS databases (NIST08 (30), Adams (40), MassFinder 3 (41), and custom libraries built from pure compounds). Quantification comes from the FID detector response on the DB-5 column without any correction factor. An estimated experimental relative standard deviation on concentrations obtained by this method is ±1% for compounds representing at least 10% of the oil, ±3% for compounds between 1.0 and 10%, and ±6% for compounds between 0.1% and 1.0%.

Chemicals
Oleyl acetate was bought from TCI chemicals (Portland, OR 97203) 11-cis-vaccenyl acetate from Cayman Chemical (Ann Arbor, MI 48108), and (Z)-4-decenol and 9-decenol from Sigma-Aldrich (St-Louis, MO 63103). A drop of each of the two alcohols was also incubated in 500 μL acetone with 200 μL triethylamine and 100 μL acetic anhydride for 24 hours in order to perform acetylation prior to injection. compounds in the fruit oil of H. antasiaticum (18). The portrait of the fruit essential oil of H. maximum is completed by the presence of the polyyne (Z)-falcarinol (0.6%), as well as of several aliphatic alcohols and aldehydes. Octanol (6.2%) appears in the oils of several Heracleum species. Generally, its concentration remains below 4%, as it is the case in H. sphondyllium ssp. ternatum oil (12,15).
The presence of (Z)-falcarinol − [(Z)-heptadeca-1,9diene-4,6-diyn-3-ol] − (3.3%) must be noted. This compound too was observed in other members of the Apiaceae family such as Daucus carota L. (25) and in low percentage in the aerial parts of the H. transcaucasicum oil (0.1%) (26) and in methanol extracts of the leaves and roots of H. moellendorffii (27). We have previously identified this compound in Anthriscus sylvestris root essential oil (28).
Three unknown compounds have been observed on both columns in the stems' oil, with Unknown I also being a noticeable constituent among the volatile compounds from the fruits. From its MS spectra, Unknown I is not of terpenic nature. It more likely belongs to the aliphatic esters class. This assumption is further backed by the retention index difference between the polar and   The derivatization of alkenes into methylsulfide adducts following the method of Buser et al. (29) confirmed these two assignations and lead to the identification of three more compounds from their specific fragments (given in parentheses): (Z)-5-decen-1-yl acetate (m/z = 115, 117, 292), (Z)-6-decen-1-yl acetate (m/z = 103, 129, 292) and (E)-4-decen-1-yl acetate, which has the same sulfide ions as its (Z)-counterpart (m/z = 101, 131, 292). Only the first ester adduct (m/z = 101, 103, 292) could not be rationalized into a structure, possibly due to rearrangements.
At least five octadecenyl acetates are observed in H. maximum stem essential oil. An attempt to produce the sulfide adducts proved fruitless given the small amount of oil available and the very small concentrations of these esters in the sample. Although several mass spectra are available from literature, their correct identification is not an easy task. For example, at least two dozens (including   non-polar columns, which is consistent with an ester moiety. Unknowns II and II more reasonably belong to the oxygenated sesquiterpenes class. Unknown II's molecular mass seems to be missing, as the large retention index difference indicates that it probably bears several oxygenated functions, while Unknown III is probably a sesquiterpenic alcohol, given its molecular mass and retention indexes.

Disambiguation of unsaturated esters identification
The MS of the six decen-1-yl acetates encountered in H. maximum fruits' essential oil are so similar that any identification of their right structure is not obvious. The RI values available from literature and those measured in the present work also preclude a confident identification. Only (Z)-4-decen-1-yl acetate and 9-decen-1-yl acetate could be formally identified by coinjection of pure compounds.   To our knowledge, the RI values on the DB-5 column for the n-C 18 monoene acetates have not been published altogether, contrarily to their C 12 , C 14 and C 16 counterparts (32) ( Table 3). The RI values measured for each of the five observed compounds (Figure 2) shed some light. For example, the saturated n-C 14 and n-C 16 acetates elute from the DB-5 column before the two last monoene compounds listed in Table 3. On this basis, one can eliminate the presence of the n-C 15 and n-C 16 and probably the n-C 17 isomers. Moreover, the RI values seem to increase as the double bond position x increases beyond 7. Thus, we suggest that the observed n-octadecenyl acetates were those indicated in the last (Z) and (E) isomers) of analogues are included in the NIST database (30). Unfortunately, these MS are very similar, and subtle differences introduced by the analytical methods used make it difficult to ascertain the double bond position of the compounds, x, on the basis of their sole MS spectra. From the intensity of the peak m/z = 68, one can eliminate the presence of 2-, 3-, 15-and 16-octadecenyl acetates. The best fits with the NIST database for the most intense peak (RI(DB-5) = 2184) are: x = 13, 935; x = 12, 926; x = 9, 924 and x = 11, 920. The observed MS for this peak is almost identical to that of the synthetic cis-vaccenyl acetate [(Z)-11-octadecenyl acetate] reported in literature (31). Table 3. retention indexes values of (Z)-monoene acetates on non-polar columns. notes: *tentative identification (see discussion). **obtained by interpolation, ±20 units.

Disclosure statement
No potential conflict of interest was reported by the authors.
Another study gives the relative retention time of the octadecenyl acetates to eicosane on a HP-1 column in isotherm conditions (33). There is a clear effect of the double bond position. As soon as the x value is higher than 8, higher x values result in longer retention times. Simultaneously, the vapor pressure of the compound becomes lower. Of course, this discussion takes into account the (Z) isomers and obliterates the (E) isomers. These cis compounds have similar MS and RI values (32).
These natural n-octadecenyl acetates are not very common. Several of them are observed in animals as pheromones or venoms. In the Plant kingdom, they generally are individually observed. Among the tentatively identified acetates, only the (E)-10 and the 11-isomers are reported in one member of the Brassicaceae (34) and the Asteraceae (35) family, respectively. On the other hand, the (Z)-3, (Z)-6-, (Z)-9 and seventeen isomers are observed in Apiaceae (36), Rubiaceae (37), Caesalpiniaceae (38) and Araliaceae (39) families.
In conclusion, examination of the chromatograms obtained for the essential oils of the fruits and stems of Heracleum maximum leads to the identification of 114 and 126 compounds, respectively. The fruits' essential oil is rich in aliphatic esters, mainly octyl acetate, in line with previously reported data for other Heracleum species around the world. Its pungent scent might very well explain the historical use of the fruits as a spice. The stems, on their part, yielded a mostly terpenic essential oil, with an interesting series of octadecenyl acetate traces. These uncommon compounds were identified by interpolation from shorter alkenyl derivatives along with coinjection of some compounds, and are reported for the first time altogether with consistent RI data. The pleasant lemon-scent of this oil likely contributed to the popularity of Heracleum stalks as a vegetable among American First Nations. In this study, the presence of coumarins, falcarinol isomers and daucane-type sesquiterpenes was detected, all of which were consistently reported in other Apiaceae species previously.