Supercritical Carbon Dioxide Antisolvent Fractionation for the Sustainable Concentration of Lavandula luisieri (Rozeira) Riv.- Mart Antimicrobial and Antioxidant Compounds and Comparison with Its Conventional Extracts

Lavandula stoechas subsp. luisieri is a Spanish subspecies from the Lamiaceae family. Its essential oil has been traditionally used for several medical applications though little is known about other extracts. Similar to many other studies aiming to obtain traditional plant extracts to be used in different applications, this work evaluated the antioxidant and antimicrobial activities of Lavandula luisieri extracts and the correlation with their composition. Traditional hydrodistillation and ethanolic maceration were used to obtain the essential oil and the maceration extract, respectively. A green and sustainable methodology was applied to the maceration extract that was under a Supercritical Antisolvent Fractionation process to obtain a fine solid enriched in rosmarinic acid and the terpenes oleanolic and ursolic acids. Antimicrobial activities of all extracts and pure identified compounds (rosmarinic and ursolic acids) were evaluated against five bacterial strains; Listeria monocytogenes, Enterococcus faecium, Staphylococcus aureus, Salmonella Typhimurium and Escherichia coli and were compared with the pure compounds identified, rosmarinic and ursolic acids. All strains were sensitive against L. luisieri essential oil. The solid product obtained from the supercritical process was concentrated in the identified actives compared to the maceration extract, which resulted in higher antimicrobial and DPPH scavenging activities. The supercritical sustainable process provided L. luisieri compounds, with retention of their antimicrobial and antioxidant activities, in a powder exemptof organic solvents with potential application in the clinical, food or cosmetic fields.


Introduction
Plants have been used since ancient times for their perfume, flavor, and preservative properties in a variety of products and applications with medicinal and cosmetic uses [1] because of their secondary metabolites with diverse bioactivities [2]. Nowadays, population concern about healthier and more natural habits has promoted the study of plant extracts in different research fields, specifically

Plant Material
Lavandula luisieri plant material was collected from Cariñena, Zaragoza (Spain) and provided by the Agrifood Research and Technology Centre of Aragon (CITA, Spain). This plant was adapted to the Cariñena experimental field in 2008 from a Toledo (Spain) wild population. It was cultivated and collected as reported by Julio et al., [34]. Plant material was dried at room temperature, pulverised and sieved. Its mean diameter was adjusted to 0.330 mm (ASAEA S319.3 from the American National Standards Institute).

Essential Oil
The essential oil was obtained by hydrodistillation in a Clevenger-type apparatus from Lavandula luisieri previously described. The essential oil was provided by the Agrifood Research and Technology Centre of Aragón (CITA), and preserved in an amber vial under refrigeration. Its chemical composition has been reported in a previous study [18].

Maceration Extract
The polar fraction of L. luisieri was obtained by submitting the pulverized plant material to two serial macerations: first with hexane, to eliminate the apolar compounds such as cuticular wax, and second with ethanol, to obtain the polar and active compounds. Macerations were carried out at room temperature for 48 hours, with a plant material:solvent ratio of 1:10 (w/w). Both solvents were removed with a rotary evaporator (Büchi R-200) to obtain the dry extract. Only the ethanolic extract (ME) was considered to perform the tests.
The yield of the maceration extract was determined according to Equation (1).
Y ME (wt.%) = mass plant extract (g) where mass plant extract is the mass of the extract obtained (dry extract, with the solvent removed), and mass plant material was the initial mass of dried and pulverised plant submitted to the extraction process.

Supercritical Antisolvent Fractionation
Part of the ethanolic maceration extract was dissolved in ethanol 3% (w/w), and filtered through a 0.45 µm pore size to constitute the feed solution (FS) of the SAF process. SAF was performed using a laboratory-scale apparatus (Green Chemistry Laboratory, I3A Researching Institute at University of Zaragoza) equipped with a CO 2 pump, an extract solution pump, a 0.5 L precipitation vessel and a downstream collector as main components ( Figure 1).
The FS was pumped towards the precipitation vessel with a flow rate of 0.5 mL/min. The SC-CO 2 flow rate was 30 g/min and the final pressure was 130 bar, i.e., conditions selected for a higher mass recovery according to previous SAF optimization ( Figure 1). The insoluble compounds in the mixture ethanol-supercritical CO 2 precipitate in this vessel constitute the precipitation vessel fraction (PV). Those compounds that were still soluble in the mixture were collected, along with the solvent ethanol, as the downstream vessel fraction (DV). The mass recovery yields in PV and DV were determined according to Equation (2).
where i is the location of PV or DV collection. The FS was pumped towards the precipitation vessel with a flow rate of 0.5 mL/min. The SC-CO2 flow rate was 30 g/min and the final pressure was 130 bar, i.e., conditions selected for a higher mass recovery according to previous SAF optimization ( Figure 1). The insoluble compounds in the mixture ethanol-supercritical CO2 precipitate in this vessel constitute the precipitation vessel fraction (PV). Those compounds that were still soluble in the mixture were collected, along with the solvent ethanol, as the downstream vessel fraction (DV). The mass recovery yields in PV and DV were determined according to Equation 2.
where i is the location of PV or DV collection

Chemical Composition Analysis
Lavandula luisieri maceration extract and its supercritical fractions were analysed with the following chromatographic procedure. The equipment used was an High Performance Liquid Chromatography (HPLC) Waters ® Alliance 2695 with a PDA Waters ® 2998 detector provided with the column CORTECS ® C18 2.7 µm (4.6 × 150 mm) and a CORTECS ® Pre-column VanGuard C18 2.7 µm (2.1 × 5 mm) (Barcelona, Spain). For the chromatographic analysis, an isocratic mobile phase of 88:12, methanol (MeOH): H3PO4 0.5% in Milli-Q water, was pumped for 10 min at 0.8 mL/min. Extracts were dissolved in ethanol to 100 ppm (approximately), filtered through a GH Polypropylene membrane (ACRODISC 13 mm, 0.2 µm, Waters, Barcelona, Spain) and finally 10 µL was injected. In order to quantify the maceration extract, PV and DV the identified rosmarinic, oleanolic and ursolic acids, standards were used to build the calibration curves. Rosmarinic acid was

Chemical Composition Analysis
Lavandula luisieri maceration extract and its supercritical fractions were analysed with the following chromatographic procedure. The equipment used was an High Performance Liquid Chromatography (HPLC) Waters ® Alliance 2695 with a PDA Waters ® 2998 detector provided with the column CORTECS ® C18 2.7 µm (4.6 × 150 mm) and a CORTECS ® Pre-column VanGuard C18 2.7 µm (2.1 × 5 mm) (Barcelona, Spain). For the chromatographic analysis, an isocratic mobile phase of 88:12, methanol (MeOH): H 3 PO 4 0.5% in Milli-Q water, was pumped for 10 min at 0.8 mL/min. Extracts were dissolved in ethanol to 100 ppm (approximately), filtered through a GH Polypropylene membrane (ACRODISC 13 mm, 0.2 µm, Waters, Barcelona, Spain) and finally 10 µL was injected. In order to quantify the maceration extract, PV and DV the identified rosmarinic, oleanolic and ursolic acids, standards were used to build the calibration curves. Rosmarinic acid was measured at 330 nm, whereas ursolic and oleanolic acids were measured at 210 nm. All analyses were performed in triplicate. The actives content was expressed as a percentage of the dry extract or SAF fraction analyzed.

Microorganisms and Growth Conditions
The bacterial strains assayed in this study were obtained from the Spanish Collection of Type Cultures (CECT) included and maintained frozen at −80 • C in cryovials until the sensitivity tests. Three gram-positive bacteria, Listeria monocytogenes (CECT 911), Enterococcus faecium (CECT 410) and Staphylococcus aureus (CECT 435), and two gram-negative bacteria, Salmonella Typhimurium (CECT 443) and Escherichia coli (CECT 516), were selected for the assays.
Broth subcultures were prepared by inoculating, with one single colony from a Tryptic Soy Agar (TSA, Oxoid, Madrid, Spain) plate, a test tube containing 10 mL of sterile Tryptic Soy Broth (TSB, Oxoid, Madrid, Spain). The inoculated tubes were incubated overnight (16 hours) at 37 • C. Then, the bacterial concentration was adjusted to an absorbance between 0.08 to 0.1 using a spectrophotometer (Jenway 3600, Tirana, Albania) with a wavelength of 620 nm which corresponds to 1 × 10 8 UFC/mL according to McFarland Turbidity scale (Standart N1 0.5, Becton Dickinson and Company, Madrid, Spain). Additionally, inocula concentration was confirmed by colony counting in agar plates after performing 1:10 dilutions in peptone water (Buffered peptone water, Oxoid, Madrid, Spain).

Disk Diffusion Method
Antimicrobial activity of the essential oil was tested against the five bacterial strains using the agar disk diffusion technique. Filter paper disks (Whatman 6 mm diameter, Sigma-Aldrich, Madrid, Spain) containing 15 µL of essential oil were placed on the surface of agar plates of Mueller-Hinton (Merck, Spain) that were previously seeded by spreading one sterile swab impregnated in strain culture of 1 × 10 8 UFC/mL. Ampicillin disks (10 µg, Oxoid, Madrid, Spain) were used as a positive control. The plates were incubated at 37 • C for 24 hours, and the diameter resulting from each inhibition zone (diameter of inhibition zone plus diameter of the disk) was measured in triplicate in millimeters. The scale of measurement was the following: ≥ 20 mm is strongly inhibitory, < 20-12 mm moderately inhibitory, and < 12 mm is non-inhibitory [35]. An average and standard deviation of the inhibition zone from three replicates were calculated.

Determination of Minimum Inhibitory Concentration (MIC) and Minimum Bactericide Concentration (MBC)
Antimicrobial activity of Lavandula luisieri extracts (essential oil, maceration extract, PV and DV) was quantified against the five working strains by the determination of the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC). These values were assessed for Lavandula luisieri essential oil by the macrodilution broth technique, whereas for ME, PV and DV, the microdilution broth technique was applied. The final bacterial working suspension for both assays was adjusted to 5 × 10 5 -1 × 10 6 CFU/mL by dilution from the measured 10 8 UFC/mL overnight culture. For both procedures, macrodilution and microdilution, the MIC was the lowest concentration of extract at which bacteria failed to grow, so no visible changes were detected in the broth medium, and the MBC was defined as the concentration at which bacteria were reduced by 99.9%.
The essential oil activity was tested using the macrodilution method adapted from Clinical and Laboratory Standards Institute (CLSI, M07-A10, 2018). The assays were performed in 10 mL of TSB (ethanol 3%), and the tested concentrations were obtained by adding suitable amounts of essential oil to a final working range of 0.5-30 µL/mL. Positive controls contained TSB with microorganisms plus 3% ethanol. Negative controls contained TSB plus 3% ethanol and 5 µL/mL of Satureja montana essential oil, whose activity has been widely studied and proved [35,36]. After a 24 h incubation at 37 • C in a shaking thermostatic bath (Bunsen, mod. BTG), the MIC was read as the concentration with no visible growth. In order to evaluate MBC, 100 µL of each case in which microbial growth was not observed was spread plated in TSA. Plates were incubated at 37 • C for 24 h. The evaluation of MIC and MBC values was carried out in triplicate.
Lavandula luisieri ME, PV and DV extracts were tested against the same bacterial strains with the microdilution broth method [37]. The test was performed in 96-well sterile microplates. All wells received Mueller Hinton Broth (MHB) supplemented with 10% glucose and 1% phenol red broth (Merck, Madrid, Spain). Extract working solutions were dissolved in water with DMSO 5% with a final well highest concentration of 2.5% [38]. The solutions were sterilized by filtration with a 0.2 µm pore membrane filter and added to the first column of wells in the microplate. The final extract concentration assayed ranged from 2000 to 2 µg/mL and was obtained by twofold serial dilution from the first column.
Finally, inoculum suspension was added to all wells. The growth controls constituted medium with extract (negative control) and medium with bacterial inoculum (positive control).
Each microplate was incubated for 24 h at 37 • C. A change of color from red to yellow was interpreted as positive growth. For MBC determination, 10 µL from each well presenting no visible growth was inoculated on Mueller Hinton agar plates and incubated at 37 • C for 24 h. Each analysis was performed in triplicate.
The standards rosmarinic and ursolic acids were also tested following the same microdilution procedure with final concentrations of 125 to 0.06 µg/mL and 1000 to 0.5 µg/mL, respectively, according to its proportion in the L. luisieri studied extracts. For the results of antimicrobial activity, the following were considered: significantly active when MIC < 100 µg/mL, moderately active 100 < MIC > 625 µg/mL and weakly active MIC > 625 µg/mL [37].

Antioxidant Activity
The capacity of the L. luisieri maceration extract and its supercritical fractions to scavenge DPPH free radicals was measured by an adaptation of the Brand-Williams, Cuvelier and Berset (1995) [39] spectrophotometric method. The extract solutions were mixed 1:1 (v/v) with a DPPH ethanol solution of 40 µg/mL. The DPPH solution was also confronted with pure RA and trolox (97% ACROS Organics) as a positive control and with ethanol as a negative control. The final well concentrations of ME, PV, DV and the positive controls rosmarinic acid and trolox ranged from 300 -0.1 µg/mL. The absorbance was measured at 520 nm after 30 min of reaction at room temperature with a microplate photometer (Multiskan EX mod. 355, Thermo Labsystems, Zaragoza, Spain). To determine the scavenging capacity the following equation was applied, Equation (3): Radical scavenging activity (wt%) = Abs control − Abs sample /(Abs control ) × 100 (3) where Abs control is the measured absorbance of the DPPH solution and Abs sample is the measured absorbance after the reaction between the extracts or control vs. DPPH. The antioxidant activity of plant extracts was expressed as IC 50 , which is defined as the concentration of extracts (in µg/mL) required to inhibit 50% of DPPH radicals. IC 50 values were estimated by a nonlinear regression (GraphPad Prism version 4.0). A lower IC 50 value indicates higher antioxidant activity. The results are given as the mean ± standard deviation (SD) of experiments performed in triplicate.

Chemical Composition of Lavandula luisieri Extracts
The non-volatile and polar fraction of this plant was obtained by ethanolic maceration. The final extraction yield obtained with this maceration was 12.2%. Regarding this same Lavandula luisieri population, other authors [18] obtained similar yield extraction (12.5%), although it was obtained with ethanol in a Soxhlet apparatus. This method applies heat for a long time and enhances the extraction of actives but it also can cause the degradation of the thermolabile compounds; it does not seem to provide better extraction yield. In this work, the extraction was performed at room temperature and in amber bottles to prevent degradation from heat and light.
The ethanolic maceration extract, ME, was fractionated through SAF into two fractions.It was recovered 50.9% (w/w) in the PV fraction and a 32.9% (w/w) in the DV fraction. According to these results, a high quantity of compounds was insoluble in the mixture scCO 2 -EtOH under these experimental conditions of pressure and CO 2 flow rate, since a 50.9% of the initial mass precipitated when the ethanolic solution encountered the scCO 2 in the PV.
The composition of ME, PV and DV was studied by means of HPLC and only the polyphenol rosmarinic acid and the two triterpenes ursolic acid and oleanolic acid were identified ( Figure 2). The content of these actives was previously reported [8]. The concentration of rosmarinic, oleanolic and ursolic acids was quantified in each fraction and compared to the initial ME (Table 1). seem to provide better extraction yield. In this work, the extraction was performed at room temperature and in amber bottles to prevent degradation from heat and light. The ethanolic maceration extract, ME, was fractionated through SAF into two fractions.It was recovered 50.9% (w/w) in the PV fraction and a 32.9% (w/w) in the DV fraction. According to these results, a high quantity of compounds was insoluble in the mixture scCO2-EtOH under these experimental conditions of pressure and CO2 flow rate, since a 50.9% of the initial mass precipitated when the ethanolic solution encountered the scCO2 in the PV.
The composition of ME, PV and DV was studied by means of HPLC and only the polyphenol rosmarinic acid and the two triterpenes ursolic acid and oleanolic acid were identified (Figure 2). The content of these actives was previously reported [8]. The concentration of rosmarinic, oleanolic and ursolic acids was quantified in each fraction and compared to the initial ME (Table 1).  The rosmarinic acid concentration was 51.4 mg/g of the initial maceration extract. After the supercritical fractionation process, rosmarinic acid was entirely retained in the PV fraction, at a final concentration of 93.5 mg/g of PV. Although ursolic acid and oleanolic acid were distributed between both fractions, ursolic acid also precipitated mostly in the first fraction, reaching a concentration of 118.6 mg/g of PV; meanwhile, oleanolic acid seem to be dragged to the downstream vessel. The total  The rosmarinic acid concentration was 51.4 mg/g of the initial maceration extract. After the supercritical fractionation process, rosmarinic acid was entirely retained in the PV fraction, at a final concentration of 93.5 mg/g of PV. Although ursolic acid and oleanolic acid were distributed between both fractions, ursolic acid also precipitated mostly in the first fraction, reaching a concentration of 118.6 mg/g of PV; meanwhile, oleanolic acid seem to be dragged to the downstream vessel. The total quantity of the studied compounds in the PV was 238.5 mg/g, 1.9 times more concentrated compared to the initial ME.
Other authors have previously performed this process of supercritical concentration for natural compounds. The application of SAF to rosemary extracts in order to concentrate its polyphenols, rosmarinic acid among them, produced a fraction with a higher antiproliferative and antioxidant activity. Sánchez-Camargo et al., [32] and Visentin et al., [33]. Bernatoniene et al., [40] also studied different techniques for the extraction of these three active constituents from Rosmarinus officinalis. The highest mass of active per gram of extract was achieved with ultrasound-assisted ethanolic extraction for ursolic acid (15.8 ± 0.2 mg/g) and rosmarinic acid (15.4 ± 0.1 mg/g), and with ethanolic maceration for oleanolic acid (12.2 ± 0.1 mg/g).
In our work, the ethanolic extraction revealed that L. luisieri is a rich source of RA, OA and UA, and the supercritical antisolvent fractionation process is a useful technique for the co-precipitation of the followed actives into a solid enriched product, with 93.5 mg/g, 118.5 mg/g and 26.4 mg/g of RA, OA and OA, respectively.

Essential Oil Antimicrobial Activity
Bacterial susceptibility of L. luisiseri essential oil was first tested with the paper disk agar diffusion method and then quantified with the broth dilution method. The results indicated in Table 2 represent bacterial sensitivity according to the classification previously described [35].
Regarding MIC and MBC values, L. luisieri essential oil had bacteriostatic activity against all bacteria tested, showing a wide antibacterial spectrum inhibiting both, gram-positive and gram-negative bacteria ( Table 2). The highest bacteriostatic activity obtained was against gram-positive tested (MIC = 0.5 µL/mL), whereas higher MIC values were acquired against gram-negative ones: 5 µL essential oil /mL against S. Typhimurium and 30 µL essential oil /mL vs E. coli. The evaluation of MBC revealed that L. luisieri essential oil was bactericidal against four of the five bacteria at the assayed concentrations. The bactericidal MBC values ranged from 0.5 µL/mL (S. aureus) to 5µL/mL (S. Typhimurium). E. coli seems to be the most resistant strain to this oil composition, since it showed bacteriostatic but not bactericidal activity, at the highest concentration tested, 30 µL. This behavior was also observed by Baldovini et al., [6], by applying a dilution agar method with a chemically defined necrodyl-rich essential oil from L. luisieri against different bacterial and yeast strains. Although the activity was tested with a different method, it was also observed that gram-positive strains were more sensitive to the essential oil than the gram-negative ones. Some E. coli strains had sensitivity to L. luisieri essential oil, but also at the highest concentration tested. L. luisieri essential oil has a higher antimicrobial activity compared to other Lavandula species, which only had activity against S. aureus such as L. lavandulifolia, L. angustifolia, Lavandin Super and Lavandin Abrialis [41]. Nevertheless, other species such as L. latifolia and Lavandin Grosso had similar MIC and MBC results to L. luisieri against several strains such as S. Typhimurium, E coli and L. monocytogenes. Generally, Lavandula species essential oil, as other plant families, has different sensitivity among gram-positive and gram-negative strains, with the gram-negative E. coli as the most resistant one. Nevertheless, the chemical composition of this other antimicrobial species differs with L. luisieri, and the activity was related to their content in 1,8-cineol, linalool and camphor [42].
So far, only Baldovini et al. [1] Roller et al. [16] and Zuzarte et al. [17] have tested L. luisieri essential oil antimicrobial activity. They obtained good results against all strains tested, especially against gram-positive bacterial and fungal strains, such as Candida albicans and Aspergillus species, and related these results to L. luisieiri essential oil necrodane terpenoid content. Roller, Ernest and Buckle [16] compared Lavandula species antimicrobial activity against methicillin resistant and sensitive Staphylococcus aureus with the disk diffusion method. L. luisieri essential oil, highly concentrated in necrodane monoterpenoids, with 34.5% relative content in α-necrodyl acetate, was the most active at lower volumes. In addition, its combination (50:50) with L. stoechas and L. angustifolia essential oils produced inhibitory zones twice the diameter than the zones obtained with both single oils. A possible synergy among necrodane terpenoids and 8-cineole, fenchone, and camphor from L. stoechas, or linalool and linalyl acetate from L. angustifolia, is suggested. Similarly, Zuzarte et al., [17] also related L. luisieri essential oil antimicrobial activity to its particular content in necrodane-type compounds when studied in two different populations from central and southern regions of Portugal against different Aspergillus strains. The activity results were considered relevant since this strain is usually less sensitive to essential oils.
In our work, the obtained essential oil also had a good antimicrobial activity. Its composition was reported in a previous study [18]. Its main compounds were camphor (60.3%) and 2,3,4,4-tetramethyl-5-methylidenecyclopent-2-en-1-one (8.5%), a necrodane-type compound, along with other substances such as fenchone (2.9%) and 1,8-cineol (2.0%). Other authors studying L. luisieri essential oil composition observed that the differences among populations appeared to be only quantitative [6,8,9,18,43]. The antimicrobial activity results obtained in this work could be a consequence of L. luisieri essential oil high content in camphor, which is a well-known antimicrobial compound, obtained mainly from Cinnamomum camphora but widely distributed in the plant kingdom [44]. According to other authors' results, besides camphor, the L. luisieri content in atypical necrodane compounds and the combination with other volatiles could also be contributing to this activity.

Antimicrobial Activity of SAF Fractions
The antimicrobial activity of the ethanolic extract and its supercritical fractions PV and DV was tested. To our knowledge, this is the first work to study the antimicrobial activity of polar non-volatile fractions of this plant. For this type of extract, the disk diffusion method could not be applied as a screening method because of the poor diffusion from the cellulose disks to the agar. The antimicrobial activity of ME, PV and DV, as well as the pure compounds rosmarinic and ursolic acids, was tested and quantified with the microdilution broth procedure. From all studied strains assayed, only L. monocytogenes, E. faecium and S. aureus showed sensitivity to the L. luisieri maceration extract and its supercritical fractions. However, Salmonella Typhimurium and Escherichia coli did not show sensitivity to the extracts at the assayed concentrations (Table 3). Table 3. Minimum inhibitory and bactericidal concentrations (µL/mL) of usolic acid, L. luisieri ethanolic maceration extract (ME), and the supercritical fractions PV and DV against L. monocytogenes, E. faecium and S. aureus. Rosmarinic acid did not show any antimicrobial activity in the studied range (0.06-125 µg/mL). Although other authors have confirmed this lack of activity against food bacteria [45], antimicrobial activity against fitopathogenic bacteria was reported [46]. This compound is produced and released by some plants only as a natural defense in challenging environments.

Extract or
Pure ursolic acid had inhibitory and bactericidal activity against the three gram-positive bacteria, L. monocytogenes, E. faecium and S. aureus; however, while L. monocytogenes and E. faecium MIC and MBC values ranged from 33 to 66 µg/mL, S. aureus was more resistant, with MIC and MBC values up to 263 µg/mL. L. luisieri maceration extracts, PV and DV, were also active in inhibiting these bacteria. The lack of activity against gram-negative strains could be a consequence of the restricted penetration because of their different and more complex cell walls, since they have a second external phospholipid bilayer designed to reduce permeability to all compounds [47].
After the fractionation process, the PV fraction had better antimicrobial results than ME results. The precipitated solid showed activity against the three gram-positive bacteria. The MIC and MBC values against L. monocytogenes (242-83 µg/mL) and E. faecium (242-242 µg/mL) were lower than those obtained with the maceration extract and also had inhibitory properties against S. aureus (1933 µg/mL). The corresponding ursolic acid concentration in the PV inhibitory and bactericidal values ranged from 28.6 to 56.9 µg/mL. In this case, the final terpene concentration in PV MIC (UA 28.7-57.3 µg/mL) was lower compared with the inhibitory and biocidal concentrations of the pure active (33-66 µg/mL). Although this compound is concentrated in PV regarding the initial maceration extract, the activity cannot entirely be attributed to it. Nevertheless, oleanolic acid, the other triterpene in this work, with a very similar molecular structure, may contribute to the final activity by addition or synergistic effect. Indeed, several studies have reported the antimicrobial activity of oleanolic and ursolic acids. There is an inhibitory capacity against both gram-negative and gram-positive and a synergetic effect between ursolic acid and aminoglucoside antibiotics [48] when tested against 12 bacterial strains. Indeed, oleanolic and ursolic acids have antimicrobial activity which has been related to the peptidoglycan structure, bacterial gene expression and biofilm formation [23].
These differences between gram-positive and gram-negative bacteria was observed by Lai et al., [15] who performed a methanolic extraction of the aerial parts of L. luisieri. In this work, the extract was dissolved in DMSO 100%, which may affect the strain sensitivity to the extracts [38].
Even though rosmarinic acid did not show activity on its own, its combination with other non-identified compounds could enhance it. In any case, the concentrations required to inhibit and kill bacterial strains by L. luisieri maceration and supercritical extracts were higher than those obtained with the essential oil. This decreased antimicrobial activity could be a consequence of the diverse compounds that constitute these extracts. Their penetration into bacterial cells and action mechanism to inactivate them could be different. Some authors studying new treatments for food preservation have observed a synergistic effect between heat treatment in combination with the addition of natural plant extracts, allowing the reduction of the heat conditions [27].

Antioxidant Activity
The antioxidant activity was evaluated against the free radical DPPH. Results represented in Figure 3 showed that for every compound and extract, the activity was concentration dependent. L. luisieri ethanolic maceration extract and its supercritical fractions showed antioxidant activity, but always lower than the positive controls trolox (IC 50 3.5 ± 0.3 µg/mL) and rosmarinic acid (IC 50 1.7 ± 0.1 µg/mL). The antioxidant measure of L. luisieri maceration extract IC 50 was 30.66 ± 1.9 µg/mL. This result is similar to other published results about different Lavandula species and extracts. The DPPH scavenging activity of the methanolic extract of L. stoechas was also studied previously [49]; its IC 50 value was 34.2 ± 3.1 µg/mL, which was related to its phenolic acid content, 25.2 ± 0.4 mg GAE/g.
After the supercritical processing of maceration extract, the antioxidant activity showed an increase in PV, IC 50 to 16.17 ± 0.7 µg/mL. The DV fraction, however, was the less active, not inhibiting the 50% of DPPH free radicals in the studied range of concentrations. These results can be graphically observed in Figure 3, where the PV and DV antioxidant curves are in the left and right, respectively, of the maceration extract curve.  [49][50][51]. In contrast, other authors [52] reported for the L. stoechas methanolic extract an IC50 of 300 ± 10 µg/mL. The decreased antioxidant activity reported could be a consequence of the heat application in the extraction process and the different proportion of extract: DPPH applied. After the supercritical processing of maceration extract, the antioxidant activity showed an increase in PV, IC50 to 16.17 ± 0.7 µg/mL. The DV fraction, however, was the less active, not inhibiting the 50% of DPPH free radicals in the studied range of concentrations. These results can be graphically observed in Figure 3, where the PV and DV antioxidant curves are in the left and right, respectively, of the maceration extract curve. According to these results and the chemical analysis, after the supercritical fluid process, the antioxidant compounds extracted in maceration extract were mainly concentrated into the solid of PV fraction. The increased scavenging activity of PV regarding the initial maceration extract and the lack of activity in DV may be a consequence of rosmarinic acid, which completely precipitates during the supercritical process when converging with CO2, increasing its proportion in this fraction According to these results and the chemical analysis, after the supercritical fluid process, the antioxidant compounds extracted in maceration extract were mainly concentrated into the solid of PV fraction. The increased scavenging activity of PV regarding the initial maceration extract and the lack of activity in DV may be a consequence of rosmarinic acid, which completely precipitates during the supercritical process when converging with CO 2 , increasing its proportion in this fraction from 51.5 mg/g of maceration extract to 93.5 mg/g of PV. The IC 50 of pure rosmarinic acid was 1.7 µg/mL, which correlates with the rosmarinic acid concentration in the maceration extract and PV IC 50 , 1.5 µg/mL.
On the other hand, oleanolic and ursolic acids did not seem to scavenge DPPH free radicals since the DPPH inhibition (%) did not reach 50% at the highest concentration tested, even though their concentration in the second fraction was 48.7 mg/g and 56.1 mg/g, respectively. This result did not correspond to other authors, who reported that ursolic acid IC 50 was 59.70 µg/mL [48].
L. luisieri scavenging activity seems to be related to its content in rosmarinic and not in oleanolic and ursolic acids. Other Lavandula species extracts have also been reported to display several antioxidant mechanisms such as reductive potential, organic, cation and superoxide free radical scavenging, electron or metal cation chelation [52][53][54][55] and has always been related to their content in phenolic compounds such protocatechuic acid, caffeic acid and rosmarinic acid.

Conclusions
L. luisieri extracts showed antioxidant activity against the free radical DPPH and antimicrobial activity against Listeria monocytogenes, Enterococcus faecium and Staphylococcus aureus, and two gramnegative bacteria, Salmonella Typhimurium and Escherichia coli. Although the essential oil showed a strong antimicrobial activity against the studied bacterial strains, its application could be limited because of its organoleptic properties, which should be further evaluated. The application of the Supercritical antisolvent green technology to the non-volatile fraction of L. luisier, not only allowed the concentration of bioactive compounds in a final solid product but also the enhancement of the studied bioactivities in the precipitated solid fraction. The increased antimicrobial activity seems to be a consequence of the ursolic acid enrichment and the antioxidant activity because of rosmarinic acid. Nevertheless, a thorough analysis by HPLC-MS to know the full chemical composition of the ethanolic extract and its supercritical fraction could be interesting for identifying other bioactive compounds. This improved final product could have potential applications in food, cosmetic and pharmaceutical industries as a preservative or nutraceutical. However, it is recommended to perform additional studies in order to assess the preservative effect after its application in the final product as well as its safety for consumers. Funding: Authors give thanks to EFA188/16/SPAGYRIA (Este proyecto está cofinanciado por el Fondo Europeo de Desarrollo Regional (FEDER)), MINECO-FEDER funds [project number CTQ2015-64049-C3-2-R], Gobierno de Aragón-FSE-FEDER "Construyendo Europa desde Aragón" (Group E39_17R) and reference group A06_17R (Gobierno de Aragón-FEDER) for the financial support.