Extraction of Cecropia Pachystachya Leaves by Supercritical Carbon Dioxide: Kinetics, Phytochemical Characterization, Antibacterial and Antioxidant Activities

Cecropia pachystachya Trécul is popularly known in Brazil as embaúva or embaúba. Pharmacological activities were reported in extant studies including the development of anxiolytic, antidepressant, hypoglycemic, antioxidant, cardiotonic, sedative, anti-inflammatory, antimicrobial and leishmanicidal properties for aqueous and alcoholic extracts prepared by infusion or maceration of C. pachystachya leaves. However, there is a paucity of studies examining extracts of leaves of C. pachystachya obtained by supercritical fluid technology. This study evaluated the extraction of Cecropia pachystachya oil using supercritical carbon dioxide (CO2). Extractions were performed at 35 °C, 45 °C, and 55 °C and 14 MPa, 17 MPa, and 20 MPa. The extraction yield, characteristics of the oil, antioxidant and antibacterial potentials were evaluated. The highest extraction yield obtained from C. pachystachya leaves was 1.48% using supercritical CO2 at 55 °C and 20 MPa. The main identified compounds include terpenoids (ursolic and pomolic acids, squalene and β-sitosterol). The most promising results for antibacterial activity relative to Staphylococcus Journal of Agricultural Studies ISSN 2166-0379 2020, Vol. 8, No. 4 http://jas.macrothink.org 572 aureus and antioxidant potentials were observed on the extract obtained at 55 °C and 14 MPa. The mathematical Sovová model satisfactorily represented the experimental data.

The quality of the vegetal extracts is extremely influenced by the methodology of extraction employed. However, non-optimized processes limit the extractive potential and consequently interfere in the pharmacological activity. Extracts rich in phenolic compounds including flavonoids and tannins are obtained by various solvent extraction techniques and more recently by supercritical fluid technology (Souza-Moreira et al., 2010).
Extraction using supercritical CO2 fluid technology is attractive due to selectivity, solubility, and mass transfer rates. The selectivity of the extraction process mainly depends on the density of the supercritical CO2, and this can be altered by varying the process conditions (Subroto et al., 2017). The low latent heat of evaporation in the process and the high volatility of the solvents make it possible to obtain extracts free of toxic residues and to preserve thermally degradable products, thereby obtaining extracts with high quality when compared to the products obtained by conventional techniques (Souza-Moreira et al., 2010;Martinez-Correa et al., 2017). Furthermore, this technique is considered green technology because the solvent is completely removed after the extraction process (Martinez-Correa et al., 2017). However, there is a paucity of studies examining extracts of leaves of C. pachystachya obtained by supercritical fluid technology. The objective of the present study is to evaluate the kinetics of supercritical CO2 extraction, effects of temperature and pressure on yield and antimicrobial and antioxidant activities, and to perform a preliminary chemical characterization of the compounds present in the extracts.

Plant Material Preparation and Characterization
The plant material was collected at Quatro Pontes, Paraná state, Brazil (24°35'40.56''S; 53°57'46.60''W). The species were identified and a voucher (no. 3000) was deposited at the Herbarium of the West Paraná State University. The material was dried in an oven at 40 °C for 3 days and was ground by a knife grinder. The average size of the particles (720 μm) of C. pachystachya leaves was determined by using a Master-Sizer particle analyzer (Malvern Instrument Ltda). The moisture content of the leaves (2.94 ± 0.03 wt%) was obtained by using the gravimetric method to dry sample at 105 ºC until the mass stabilized. The density (1.44 g· cm −3 ) of the investigated sample was determined by pycnometer by using helium gas (Micromeritics, AccuPyc model 1330) (Santos et al., 2015).

Cecropia pachystachya Oil Extractions
The extractions were performed in a laboratory-scale unit. The equipment consists of a stainless steel extractor with a capacity of 58 cm 3 (with a height of 19.4 cm and a diameter of 1.95 cm), a solvent reservoir, a Syringe pump (Isco, 500D model), and two thermostatic baths. The extractions were performed with pure carbon dioxide (CO2).
A full factorial experimental 2 2 design with center points was used to analyze the influence of the independent variables pressure, and temperature, on the extraction yield. The experiments were performed at pressures of 14 MPa, 17 MPa, and 20 MPa and temperatures of 35 ºC , 45 º C, and 55 ºC . The selection of pressure and temperature conditions allowed an increase in the solvating capacity of the solvent and was based on the equipment limits (Santos et al., 2016).
In each extraction, the vessel was completely loaded with the material. After pressurization, contact between the solvent and material was maintained for 30 minutes, and then the micrometric valve was opened for a solvent mass flow rate of 1.96 g min −1 . The oil was collected in an amber vessel, and its mass (%, w/w) was determined at time intervals of 10 min (up to 150 min of extraction). The yields were calculated as the ratio of the extracted oil mass to the initial C. pachystachya leaves mass. The experimental data were statistically analyzed using Statistica TM , version 7.0 software (Statsoft) at the 95% confidence level.

Kinetics of the C. pachystachya Oil Extraction
The kinetic curves were described by a second-order model proposed by Sovová (StatSoft, 2007). The model considers two mass transfer mechanisms and three different extraction stages. In the first stage, the extraction of directly exposed oil (bioactive compounds) in the fluid phase occurs. This step depends on the solubility in the fluid phase and is characterized by a linear curve with a slope close to the solubility value of the oil in the solvent (Silva et al., 2015). Subsequently, a decrease in the extraction rate occurs at which the fraction of oil on the surface of the cells shows indications of being exhausted. Additionally, the extraction of a low amount of accessible oil occurs, and this is controlled by an internal diffusion mechanism (diffusion mass transfer mechanisms in conjunction with the convection). In the third stage, the extraction curve becomes almost linear at an extraction rate that is significantly lower than that in the first period (Nelder & Mead, 1965;Liu et al., 2013;Silva et al., 2015;Lopes et al., 2020). The solubility was calculated from the slope of the linear part of the general extraction curves. The parameters r, Z, and W, were adjusted as previously reported (Santos et al., 2016).

GC-MS Analysis
Chemical identification was performed on GC-MS QP2010 SE (Shimadzu). Specifically, 10 μL of the samples were diluted with 990 μL of dichloromethane (Anidrol) before injecting the same into a SH-RTx-5MS column (Shimadzu, 5% phenyl-methylsiloxane, 30 m × 0.25 mm id, 0.25 μm) by using an auto-sampler (Shimadzu AOC-20i). Helium was used as the carrier gas at a flow rate of 1.0 ml min −1 with a split ratio of 2:1, and the volume of the injected sample was 1 μL. The column temperature was initially programmed at 40 °C and heated at 6 °C min −1 to reach the final temperature of 300 °C. It remained at this temperature for 10 min. The total time of GC-MS analysis was 53.3 min. The injector and the GC-MS interface temperatures were maintained at 250 ºC . Mass spectra were recorded at 70 eV with a mass range from m/z 35 to 550 amu.

Total Phenolic Content (TPC)
The TPC of the oils was determined by using the Folin-Ciocalteu methods described previously (Singlenton & Rossi, 1965). The standard calibration (between 500 and 15.625 μg mL −1 ) curve was plotted by using gallic acid given the following equation for the line y = 0.0064x + 0.1135 with a correlation coefficient of R 2 = 0.9932. All determinations were performed thrice, and the total phenolic content was expressed as μg gallic acid equivalent/g of oil (µg GAE goil −1 ).

DPPH Radical Assay
The DPPH method is based on the quantification of the oil, or extract, concentrations necessary to reduce the initial concentration of the radical DPPH by 50% by the hydrogen and electron donation (Miguel, 2010). The radical scavenging capacities of each oil in different concentrations (100-1000 µg mL −1 in ethanol) were estimated according to the method of Choi et al. (2002). Negative control was prepared by mixing 1 mL of DPPH 0.3 mM with 2.5 mL of ethanol. The mixtures were shaken vigorously for 1 min and left to stand at room temperature (37 o C) in the dark for 30 min. Thereafter, the absorbance for the sample was measured at 517 nm. Essential oils and plant extracts are considered to have antioxidant activity when their IC50 value is less than 250 µg mL -1 (Czaikoski et al., 2015).

FRAP-Ferric Reducing Antioxidant Analysis
The FRAP analysis was performed using the methodology of Santos et al. (2016) at 595 nm using the FRAP reagent as a blank. A standard calibration (100-2000 µM) curve was plotted by using ferrous sulfate. All determinations were performed in triplicate, and the antioxidant capacity was expressed as µmol of Fe 2+ per g of oil.

ABTS •+ Free Radical Scavenging Assay
The ABTS activity was carried out based on a method developed by Santos et al. (2016). The antioxidant activity of the samples was calculated through the range of the dose-response curve of Trolox reagent (100-2000 µM) and presented as µmol of Trolox per g of oil. All reactions were carried out in triplicate.

Antibacterial Activity
Standardized bacterial strains of Staphylococcus aureus (ATCC 25922), Escherichia coli (ATCC 25923), and Pseudomonas aeruginosa (ATCC 27853) were transferred to sterile 0.9% sodium chloride until a transmittance of 25% was obtained at a wavelength of 580 nm by using a colorimeter. An inoculum with 1.0% in agar was prepared from this suspension by adding 1 mL of suspension into 100 mL of sterile Muller-Hinton agar at 46-48 °C (USP, 2011).
The antibacterial activity was assessed by an agar well diffusion method (CLSI, 2009). This consists of the preparation of plates containing one layer (20 mL) of sterile solid medium and another layer (5 mL) of culture medium that is previously inoculated with sensitive microorganisms. After solidification, five holes were created on the culture medium with a sterile perforator in which two of the holes are impregnated with the C. pachystachya oils extracted at a concentration of 30 mg mL −1 , and two of the holes are impregnated with antibiotics Amoxicillin and Neomycin Sulfate at a concentration of 10 mg mL −1 . The fifth hole was used a negative control (sterile 0.9% sodium chloride). The plates were incubated in a bacteriological oven at 35±2 º C for 24 h (USP, 2011), and the plates were read with a caliper (ANVISA, 2010). All tests were performed thrice.

Statistical Analysis
The significant differences between means of the experimental results were submitted to variance analysis (ANOVA) and compared by performing Tukey's test (p < 0.05) by using Statistica 7.0 software (Statsoft, USA).

Extraction Yield
The experimental conditions used for the oil extraction of C. pachystachya leaves, CO2 density, oil solubility (S), and yield obtained for each condition are shown in Table 1. The Ys values were calculated from the linear part of the extraction curves of the C. pachystachya leaves. Concerning the temperature and pressure conditions investigated in the experimental design, the pressure was the only variable that exhibited a significant effect (P < 0.05). It obtained a maximum increase of approximately 127% (0.65-1.48 wt%) for the yield with an increase in the extraction pressure from 14 MPa to 20 MPa at 55 o C. This was observed in previous studies (Ribas et al., 2014;Santos et al., 2016;Santos et al., 2017). Thus, an increase in pressure leads to increases in the solvating power of CO2 and results in a higher extraction yield (Palvic et al., 2017). The temperature did not exhibit a statistically significant influence on the extraction yield at the highest pressure that was used. However, the results suggest a positive effect on this variable. At low pressure (14 MPa), an increase in temperature decreases the yield (0.95 at 35 °C and 0.65 at 55 °C, almost 30% of decrease). Whereas at higher pressure (20 MPa), an increase in temperature slightly increases the yields (1.35 at 35 °C and 1.48 at 55 °C). Given the opposing effects of vapor pressure and solvent density and on solubility, the effect of temperature is a complicated parameter. This phenomenon is due to the retrograde solubility usually encountered with supercritical CO2, which is due to opposite effects of CO2 density and solute vapor pressure, where the temperature increases with decreases in the solvent density and on solubility (Hatami et al., 2014). A similar effect was reported for the extraction of Lupinus albescens Hook. & Arn. roots (Conforti et al., 2017), Humulus lupulus L. ISSN 2166-0379 2020 hops (Kupski et al., 2017), and Salvia officinalis L. (Palvic et al., 2017).

Journal of Agricultural Studies
The chemical compounds were identified in the oil of C. pachystachya extracted with CO2 (Appendix). The samples presented qualitative similarities in their compositions although the changes in temperature and pressure altered the percentage of the components (Table 2).  (Yang et al., 2020). Among the triterpenes, squalene it is the simplest representative, being acyclic (Shanmugam et al., 2012). The terpenes β-sitosterol (Rivera-Mondragón et al., 2017), ursolic and pomolic acids, and α-amyrin (Hikawczuk et al., 1998;Rivera-Mondragón et al., 2017), previously found in extracts of C. pachystachya leaves, have been identified in a significant ISSN 2166-0379 2020 concentration in our oils. The high concentration of these compounds can be justified by their non-polar characteristics, similar to supercritical CO2 that is a non-polar solvent.

Journal of Agricultural Studies
The aqueous and alcoholic extracts obtained from C. pachystachya are widely known for their high composition of flavonoids that include catechin (Costa et al., 2011;Cruz et al., 2013), isoorientin and isovitexin (Costa et al., 2011;Cruz et al., 2013;Brago-Vanegas et al., 2014). Additionally, quantification of the major phenolic compounds in the leaves of C. pachystachya showed that chlorogenic acid is the major compound in the aqueous extract of this species (Brago-Vanegas et al., 2014). Despite high content of flavonoids in C. pachystachya, the content in oils indicates that only a small part of flavonoids was extracted and much larger part remained in residuum. So the polarity of this solvent does not favor the extraction of flavonoids.

Kinetics and Mathematic Modeling of the Extraction
The apparent (0.775 g cm −3 ) and real (1.447 g cm −3 ) density values obtained allowed the determination of the porosity of the bed (0.812). The characteristics necessary for the implementation of the mathematical Sovová model were as follows: Initial concentration of oil in the inert solid = 0.015 g g −1 ; Bed density = 0.238 g cm −3 , and solid mass in an oil free base = 11.528 g. The experimental conditions, solubility, and adjustable parameters are shown in Table 3.  ISSN 2166-0379 2020 The parameter r (0.69) indicates the existence of a high oil level that is easily accessed by the solvent and provided by the milling process. Its value is constant since the raw material was subjected to the same treatment before the extractions. The Z and W coefficients are proportional to the mass transfer parameters in the solvent phase (kFa) and the solid phase (ksa), respectively. Values were observed in the same order of magnitude for kFa and ksa in the extraction of Candeia oil by using pressurized CO2 (Ribas et al., 2014). As observed, the kFa values exceed the ksa values, and thus demonstrate the higher yield during easy-to-reach oil extraction (Silva et al., 2015). According to the model, the extraction kinetics are divided into three periods with mass transfer. The first period (tCER) is fast (2.76-34.90 min) because it refers to the oil with the most easily accessed to the extraction. In the intermediate period (tFER), the extraction rate decreases due to the exhaustion of the oil that is readily available until the beginning of the third period in which the extraction rate is limited by the internal diffusion mechanism  with low mass transfer coefficients in the solid phase ksa (1.01 10 −3 to 5.33 10 −3 min −1 ) due to the difficulty in extracting the oil contained within the intact cells.   ISSN 2166-0379 2020

Total Phenolic Content, Antibacterial and Antioxidant Activity
The low TPC in extracts (61.07 -72.37 µgGAE goil −1 ) proves that only a small part of flavonoids was extracted (Table 4). In general, extracts obtained with supercritical CO2 present low extraction of TPC due to the non-polar capacity of this solvent that does not favor the extraction of phenolic compounds, a substance of a generally polar character, and responsible for the antioxidant activity (Dias et al., 2017). The oils extracted from C. pachystachya leaves were evaluated for their antioxidant capacity. The DPPH method revealed that the temperature and pressure employed in the extraction ISSN 2166-0379 2020 influenced the antioxidant capacity, which ranged from 196.39 µg mL -1 (55 º C x 14 MPa) to 359.23 µg mL -1 (45 º C x 17 MPa). These results are significant when compared to other studies evaluating the DPPH method from Eremanthus erythropappus supercritical CO2 extract that found IC50 of 716 µg mL -1 , and spent coffee extract who reported IC50 values of 478 to 2369 µg mL -1 (Andrade et al., 2012).

Journal of Agricultural Studies
In the tests using the FRAP method, the antioxidant potentials of extracts were estimated for their ability to reduce the TPTZ-Fe (III) complex to a TPTZ-Fe (II) complex. The FRAP reducing values were in the range of 191.14-381.14 µmolFe2+ goil −1 . Again, the most effective condition for the recovery of reducing substances in the FRAP assay was observed at 55 º C and 14 MPa. In contrast, reducing capacity of 470 µmolFe2+ goil −1 by FRAP method was observed for the supercritical CO2 extract from Populus nigra (Kús et al., 2018). So, it can be showed lower antioxidant activity by FRAP method for the extracted oil.
According to ABTS method, based on the capacity for reducing the chromophore radical ABTS •+ , the antioxidant potential varies from 3122.22 (45 º C x 17 MPa) to 5288.88 (55 º C x 14 MPa) µmolTrolox goil −1 . These results showed greater antioxidant potential by ABTS method compared to the results reported to candeia wood oil (232.11 to 405.48 µmolTrolox goil −1 ) obtained by supercritical CO2 (Santos et al., 2016). The best results for TPC and antioxidant potential were observed for "C" extract obtained at high temperature (55 º C) and low pressure (14 MPa), thus although the temperature does not show statistically significant on the extraction yield, the biological results obtained for "C" show the significant effect of temperature.
This study involved an evaluation of the in vitro antibacterial activity against P. aeruginosa, E. coli, and S. aureus, but only the last species exhibited inhibition to the concentration used (30 mg mL −1 ). There is no consensus on the acceptable level of inhibition for natural products when compared to antibiotic standards since the material is a mixture of compounds and not a single compound in isolation (Almeida et al., 2014;Waltrich et al., 2015).
A study evaluating the antimicrobial activity of the aqueous and alcoholic extracts of C. pachystachya leaves against S. aureus showed an inhibition zone of 7 mm for the alcoholic extract (at 100 mg mL -1 ) (Costa & Hoscheid, 2018). This inhibitory action was attributed to flavonoids and TPC, since they are well established as antimicrobial agents (Babii et al., 2016;Tsou et al., 2016;Echeverrí a et al., 2017). However, when compared with this work, we observed the formation of a greater inhibition zone (9 mm) in a lower concentration (30 mg mL -1 ) by the same method (agar diffusion method). This fact makes us understand that the main antimicrobial activity found in "C" extract (55 º C x 14 MPa) is not due to the unique presence of flavonoids, which showed low content by GC-MS and TPC.

Conclusion
The results of the study indicated that the oil extraction of C. pachystachya with supercritical CO2 provided satisfactory extract yields in 150 min. The best results were obtained by applying a higher pressure and temperature (20 MPa and 55 º C) investigated in the study with a yield of 1.48 wt%. The examined experimental conditions revealed that only one pressure exhibited a statistically significant effect without extraction yield, however, the biological results obtained show the significant effect of temperature. The oil obtained by applying different conditions did not reveal significant differences concerning its qualitative composition. The major identified compounds were ursolic and pomolic acids. The best results for TPC, antibacterial activity relative to Staphylococcus aureus, and antioxidant potentials were observed on the extract obtained at high temperature (55 º C). For extraction kinetics, the Sovová model provided an adjustment with experimental data.