The criterion for bunch harvesting is the same for African palm oil and the ripe bunches were collected when they had 5-10 detached fruits that had fallen to the ground. After harvesting, a heat treatment was carried out on twelve bunches with boiling water for one hour. The fruits were detached and pressed using a hydraulic press consisting of a stainless-steel square box of 40×40×40 cm, specially developed for Eo fruits. The separation of oil and water was carried out by centrifugation. The oil was kept frozen until the analysis. The assays were performed on two different days at an interval of two months and the samples were identified as 1 (day 1) and 2 (day 2) and each sample was collected in triplicate, and the analyses were performed in triplicate or quadruplicate for each plastic bottle totaling nine or twelve replicates.

Free fatty acid (FFA), expressed as percentage of oleic acid, was determined according to method Ca 5a-40 (AOCS, 2009). The refractive index (RI) was determined by refractometry at 40 °C on a Bausch & Lomb Abbe refractometer, according to method Cc7-25 (AOCS, 2009). Relative Density was determined at 20 °C by a digital densimeter Anton Paar DMA-46.

The analysis of the fatty acid profile of oil is accredited to ISO/IEC 17025. The methyl esters were obtained according to Antoniassi et al. (2018b) and analyzed by gas chromatography on an Agilent 7890, equipped with a flame ionization detector operating at 280 °C and a HP FFAP capillary column (25 m x 0.2 mm x 0.30 µm). The oven temperature was programmed as follow: initial temperature of 150 °C for 1 min; from 150 to 180 °C at a rate of 30 °C/min; from 180 to 200 °C (20 °C/min); 200 to 230 °C (3 °C/min) and held at the final temperature of 230 °C for 10 min. A pressure ramp was used as described: initial pressure 15 psi for 10 min, from 15 to 25 psi with a 5 psi/min ramp rate and a final pressure of 25 psi for 11 min. The injector was set to 250 °C and operated in a 1:50 split mode and 1 μL of sample was injected. Identification was performed by comparing retention times with the standards of NU-CHEK PREP, Inc. (Elysian, MN, USA) and Supelco (Bellefonte, PA, USA) and quantification was performed by internal normalization.

The iodine value and saponification values were calculated based on the fatty acid composition.

The analyzes were performed simultaneously using a Waters Technologies Alliance 2695 liquid chromatograph (HPLC) (Milford, Massachusetts, USA), equipped with a quaternary pump system, on-line degasser, column furnace, automatic injection Rheodyne valve (Rheodyne LCC, Rohnert Park, USA), connected to Waters 2998 photodiode array (PDA) and Waters 2475 fluorescence detectors. The oil was dissolved with acetone. The compounds were separated on a YMC C30 reverse-phase analytical column (0.25 m × 3.0 mm I.D., 5.0 μm particle size) (Kyoto, JAPAN) conditioned to 35 °C. Samples were kept to 15 °C. The separation was performed using a gradient (flow 0.8 mL/min) with the mobile phase methanol/methyl tert-butyl ether (9/1, v/v) for 20 min, methanol/methyl tert-butyl ether (1/9, v/v) for 5 min, and methanol/methyl tert-butyl ether (9/1, v/v) for 3 min totaling 28 min. The identification of the compounds was made by comparing the sample retention time with the peak retention time of the carotenoid and tocochromanol standards and by the absorption spectra of the carotenes (Davis, 1976). Chromatograms were processed at 290 nm (excitation) and 330 nm (emission) at the fluorescence detector for tocopherols; and, between 200 and 800 nm, in scan mode, in the PDA detector and quantification at 450 nm. The quantification of tocopherols was done through external standardization. The concentration of each tocochromanol stock solution was calculated according to AOCS (2009). For total carotenes, quantification was performed by spectrophotometry considering the wavelengths, molar absorptivities and the solvent used in the dilution, according to Davies (1976). The calculation of Retinol Activity equivalent was calculated based on the conversion factors reported by Charrondière et al. (2012).

Analysis of variance and Tukey test were performed using Statgraphics (Statgraphics Technologies Inc) at a significance level of 0.05.

The main fatty acids were C18:1 (54.72 - 55.69%), C16:0 (23.72 - 24.23), C18:2 (15.47 - 16.95%) and C18:0 (1.91 - 1.97%), which accounted for about 97% of total fatty acids. The C18:1 comprised oleic (C18:1 cis-9) and cis-vaccenic (C18:1 cis-11). C18:3 was quantified up to 0.7% and minor fatty acids below 0.5% were C14:0, C16:1, C17:0, C17:1, C20:0 and C20:1. C22:0 and C24:0 were quantified as traces. There was no difference among the caiaué oil samples (p < 0.05) except for C18:0 (1.97 - 1.91%) and C20:0 (0.18 - 0.15%). The samples were obtained within a two-month interval from mature bunches available at the time and not necessarily from the same plants. However, the low variation observed among samples indicated low variability among different plants whose accession had not been identified at this time. (Table 1).

The fatty acid profile of pressed Caiaué oil was in the ranges reported for the oil obtained by solvent extraction from the dried mesocarp of five Brazilian accessions of Eo cultivated in Brazil (España et al., 2018). In addition, the results were consistent with the variation observed for Eo accessions from Panama, Suriname, Brazil, Ecuador, Costa Rica, Colombia grown in Costa Rica (Lieb et al., 2017) and for genotypes from Brazil, Peru and Colombia grown in Colombia (Chaves et al.; 2018), both by solvent extraction. The same was observed for the results reported by Mohd et al. (2000) for progenies of Eo from Colombia, Panama, Costa Rica and Honduras cultivated in Malaysia, although the method of extraction was not reported. The oil extraction with solvents such as hexane for industrial extraction or for laboratory purposes (e.g. Soxhlet extraction with petroleum ether) are very effective methods for oil recovery since the meal presents less than 1% oil. However, pressing is the usual process for oil extraction from palm fruits and the fiber from palm oil processing may contain 5-6% residual oil (Basiron, 2005). Therefore, the fatty acid profile of pressed oil is more reliable for nutritional purposes. As far as we know, this is the first report on the fatty acid from E. oleifera oil obtained by pressing.

The variation observed among studies for the fatty acid profile of Eo could be due to genetic factors, growing conditions and according to the extraction protocols employed. The fatty acid profile of Eo oil is expected to be more unsaturated than palm oil and the Eo×Eg hybrid according to Jalani et al. (1997), Choo et al. (1997) and Yap et al. (1991). According to Rios et al. (2011), the highest contents of oleic and linoleic acids and the lower content of palmitic acid of Eo mesocarp oil are useful characteristics to transmit to interspecific hybrids with Eg. In addition, Eo is useful for breeding due to its resistance to Bud rot-type (fatal yellowing) and red ring diseases, besides the smaller size palm tree.

The saturated fatty acids (SFA) comprised 26.4 and 26.7% and the SFA/UFA (unsaturated fatty acids) ratio was 0.4 and in the range of 0.3 to 0.5 reported by Lieb et al. (2017) for Eo genotypes and 1.2 for Eg. The high UFA concentration in Caiaué oil increases its fluidity and minimizes the formation of stearin, a high melting point fraction responsible for the semi-solid physical state of similar oils during storage that could be undesirable for some food applications.

The fatty acids of pressed Caiaué oil were in the range reported for “Palm oil with higher oleic acid” derived from the fleshy mesocarp of hybrid palm fruit (Eo × Eg) by Codex Alimentarius (FAO, 2022) which covered the genetic variation presented for different countries. However, the saturated and unsaturated fatty acids of pressed Caiaué oil were in the lower and higher limits of these reported ranges, respectively (Table 1). Palm olein and Palm super olein are commercial products from palm oil fractionation for obtaining liquid and clear oil with a lower content of saturated fatty acids than palm oil. Caiaué oil showed significantly lower content of saturated (C16:0 and C18:0), higher content of mono- and polyunsaturated fatty acids (C18:1 and C18:2) and higher Iodine value (77-79 g/100 g) than these palm oleins (Table 1). In addition, Caiaué oil is liquid at room temperature, indicating that it can meet the same food applications as these two products. The saponification value was in the range of palm oil and its oleins and Eo×Eg oils.

No differences wereas observed for Refractive index or Relative density between the two samples of pressed Eo oil (Table 2). The Refractive index depends on the fatty acid profile and the results were in the range described for palm oil with higher oleic acid by Codex Alimentarius (FAO, 2022), and higher than palm olein and palm superolein. The relative density was similar to palm olein. The usual temperature for liquid oils analysis is 40 ºC while 50 ºC is suitable for fats. In this way, the comparison is not possible.

The samples presented low free fatty acid (FFA) content (1.06 and 1.33%, as oleic acid) (Table 2). The acidity of Caiaué oil met the limit (2%) reported by Brazilian Regulation for edible oil obtained by pressing (Brasil, 2021). Oil acidity is related to the action of lipases that promote triacylglycerol hydrolysis. After harvest and due to the cell disruption of mesocarp by handling or processing, the enzymes are activated. Therefore the heat treatment is useful for palm fruits with high moisture to control the oil’s acidity (Basiron, 2005). Udeh and Obibuzor (2017) found a FFA range of 0.41 - 1.48% (as palmitic acid) for eight samples of Eo oil while a wider range was described for Eo×Eg hybrids and palm oil (2-2.2 and 3.2-3.6%, respectively) by Mozzon et al. (2013). The oil acidity imparts a flavor that may be undesirable in crude oils and in the oil refining process, the higher the acidity, the greater the loss of neutral oil, reducing the yield of the process. Then, Eo genotypes with low lipase activity associated with good practices of bunch harvest followed by heat treatment are desirable to produce oil with low acidity (Basiron, 2005).

The carotenes of pressed Caiaué oil showed significant differences between the two samples analyzed as reported in Table 3 (p < 0.05). Alpha-carotene and beta-carotene accounted for approximately 92% of the total carotenoids, with averages of 620-725 µg/g and 1,358- 1,403 µg/g of oil, respectively, while the total carotene content was 2,145-2,330 µg/g of oil. The differences observed between the two samples could be due to maturity at harvest, harvesting and post-harvest handling, processing, and storage. Ripening is a main factor for enhanced carotenogenesis and carotenoids increase markedly in number and quantity (Rodriguez-Amaya et al., 2008). The comparison between the two samples regarding carotenes, tocochromanols and acidity will be addressed in the next section.

The long-established function of carotenoids in terms of human health is the activity of provitamin A. In recent years, the focus has been on reduction of the risk of developing chronic degenerative diseases (Tan et al., 2021). The Recommended Dietary Allowance (RDA) of vitamin A for men and women is 900 and 700 μg retinol activity equivalents (RAE)/day, respectively (Institute of Medicine, 2001). Regarding the contribution of pressed Caiaué oil as a source of vitamin A, the consumption of 8 mL is equivalent to 1000 µg of RAE and meets the RDA for both men and women. The RAE of the two samples was 139 and 147 μg/g (Table 3). The composition of oils from native palm fruits makes them a new source of high-added-value phytochemicals. In addition, lipids are known to stimulate the absorption of carotenoids, so they may have the advantage of greater bioavailability (Rodriguez-Amaya et al., 2008; Santos et al., 2015).

The results were higher than all fruits, vegetables and Brazilian foods reported by Rodriguez-Amaya et al. (2008). The analysis of the carotenoid content in mesocarp oils from different Brazilian palms (Astrocaryum vulgare, Bactris gasipaes, Maximiliana maripa, Oenocarpus bacaba and Mauritia flexuosa) showed large variations ranging from 3 to 567 µg/g of beta-carotene and negligible amounts of alpha-carotene and beta-cryptoxanthin (extraction by Soxhlet with ethyl ether and the HPLC method) (Santos et al., 2015). Speranza et al. (2016) reported 781 µg/g of beta-carotene for Mauritia flexuosa (spectrophotometric method). Although there is potential for the application of Brazilian palm mesocarp oils, the pressed Caiaué oil is a remarkable source of pro-vitamin A.

The oil from the mesocarp of fruits of Elaeis palms is the main source of pro-vitamin A carotenoids. Trujillo-Quijano et al. (1990), Yap et al. (1991), Jalani et al. (1997) and Choo et al. (1997), compared the total amount and profile of carotenoids in oils and found a higher level of carotenoids (up to 4,600 µg/g of oil) for Eo than Eg and the hybrid Eo×Eg. However, the content of carotenoids depends on the extraction method, which was not informed by Jalani et al. (1997) and Choo et al. (1997), while Yap et al. (1991) reported extraction by Soxhlet (hexane). Trujillo-Quijano et al. (1990), reported a higher content of carotenoids (1577 µg/g) for the oil from Eo and up to 6-fold than Eg, although the extraction was performed by the Bligh and Dyer method (Bligh and Dyer, 1959). Silva et al. (2020) employed three different protocols for oil extraction and the highest content of beta-carotene (spectrophotometric method) was found for Eo compared to Eg and their interspecific hybrid Eo×Eg, but a significant difference was observed among the extraction methods. Then, the results of the carotene content in oil obtained by solvent extraction from mesocarp is useful for genotype comparison but the industrial processing of palm bunches requires sterilization (heat treatment) for enzyme inactivation and the temperature of screw pressing and centrifugation usually reach 90 ºC (Basiron, 2005). As far as we know, this is the first report on the carotenes (HPLC method) of pressed E. oleifera oil.

The comparing among Eo genotypes for five Brazilian accessions of E. oleifera cultivated in Brazil revealed that the carotene content varied from 1527 to 3344 µg/g (España et al., 2018). These results were obtained by spectrophotometric analysis from solvent-extracted oil.

The comparison among available results is difficult. Besides the differences in oil extraction protocols, there are different strategies to perform carotene quantification. The HPLC method provides separation, then the quantification of individual carotenoids is feasible. The spectrophotometric method requires the molar absorptivity of the predominant carotene as an estimate but the result should be expressed as total carotenes instead of content of beta-carotene.

The difference observed between the sum of alpha- and beta-carotene and total carotenes (Table 3 and Figure 1) is due to the isomerization of the original carotenes present in the mesocarp and these compounds showed no pro-vitamin A activity. The isomerization is consistent with the heat treatment of the bunches and it was observed for the screw-pressed Eo×Eg hybrid palm oil by Antoniassi et al. (2018a).

The two samples of Eo oil presented the same profile of tocochromanols, three forms of tocotrienols (alpha, delta and gamma) and two forms of tocopherols (alpha and gamma) (Table 4). Significant differences (p < 0.05) between the tocochromanol concentrations of the two samples were observed for alpha-tocotrienol, gamma-tocotrienol and total tocochromanol contents. Gamma-tocotrienol was the most abundant of the isomers (circa 80% of the total tocochromanols), followed by alpha-tocotrienol (14.0-17.6%). Total tocochromanols accounted for 979 and 1330 µg/g and gamma-tocotrienol corresponded to 799 and 1066 µg/g.

The screw-pressed oils from Costa Rica presented 599 and 211 µg/g of gamma-tocotrienol for Eo samples, while for Eg, it varied from 70 to 90 µg/g and EoxEg hybrid showed an intermediate amount of 146 µg/g, but lower than the results obtained in this work. Solvent extraction yielded 2 to 2.5 times more tocochromanols than the screw-press method (Irias-Mata et al., 2017).

Chaves et al. (2018) reported gamma-tocotrienol (376 to 827 µg/g) and alpha-tocotrienol (45-263 µg/g) for genotypes of Eo from Brazil, Peru and Colombia grown in Colombia but the oil was obtained by solvent extraction. Regarding the Elaeis species, Choo et al. (1997) reported a similar range of 700-1000 µg/g of tocochromanols. Jalani et al. (1997), quoted a lower range for Eg (600-1000 µg/g) and higher for Eo and Eo×Eg hydrid oils (600-1000 µg/g), although the extraction method was not mentioned.

Caiaué oil presented a higher content of gamma-tocotrienol and in the range of alpha-tocotrienol reported for Palm Olein, Palm Superolein and Eo×Eg hybrid oil by Codex Alimentarius (FAO, 2022) (Table 4). The comparison should be among crude oils because refining is a usual process for palm oil in order to obtain low acidity and light-colored oils. Refining destroys the carotenes and reduces tocochromanol content by 50%. This is an advantage of crude oils because tocopherols and tocotrienols are antioxidants and the synergism among tocochromanols, carotenes and the combination with the fatty acid profile provide high oxidative stability (Basiron, 2005). However, the acidity of Elaeis oleifera oil is expected to be lower than palm oil due to the low lipase activity and the result obtained below 1.3% allows its consumption as crude oil which is rich in pro-vitamin A and tocochomanols, and the oil is liquid at room temperature.

The sample 2 showed higher amounts for alpha-, gamma-tocotrienol, total tocochromanols, alpha-, beta-carotene and total carotenes and oil acidity as well, indicating that the bunches presented different maturity stages for the two assays (Rodriguez-Amaya et al., 2008). Surely, the interval after harvest until the heat treatment was higher for sample 2 because of its higher acidity as observed by España et al (2018). Enzymes promote the hydrolysis of the cell wall and release oil and bioactive compounds and increase oil acidity.

Pressed Brazilian Caiaué oil is a valuable source of tocotrienols. In vitro studies suggested that tocotrienols display stronger antioxidant, anti-inflammatory and chemopreventive activities when compared with tocopherols. In addition, while various studies have indicated that alpha-tocotrienol is neuroprotective while gamma-tocotrienol exhibited the greatest anticancer effects (Azzi, 2019). The tocochromanols are effective antioxidants in oils but the difference among them depends on the type of oil, fat, emulsion or food system evaluated. Alpha-tocopherol has been tested and shown to prevent vitamin E deficiency disease according to Azzi (2019) and there are no conversion factors for tocotrienols to Vitamin E. The contribution of Caiaué as a source of Vitamin E is negligible, but some authors have shown the sum of tocochromanols as total vitamin E content.