Physicochemical characterization and cosmetic applications of Passiflora nitida Kunth leaf extract

Passiflora nitida Kunth, an Amazonian Passiflora species, is little studied, although the specie’s high biological potential. Herein the plant’s pharmacognostic characterization, extract production, antioxidant potential evaluation, and application of this extract in cosmetic products is reported. The physical chemical parameters analyzed were particle size by sieve analysis, loss through drying, extractive yield, total ash content, laser granulometry, specific surface area and pore diameter (SBET), differential scanning calorimetry, thermogravimetry (TG), and wave dispersive X-Ray fluorescence (WDXRF). Total phenol/flavonoid content, LC-MS/MS analysis, DPPH and ABTS antioxidant radical assays, cytotoxicity, melanin, and tyrosinase inhibition in melanocytes test provided evidence to determine the content of the major constituent. P. nitida dry extract provided a fine powder with mesopores determined by SBET, with the TG curve showing five stages of mass loss. The antioxidant potential ranged between 23.5-31.5 mg∙mL -1 and tyrosinase inhibition between 400–654 μg∙mL -1 . The species presented an antimelanogenic effect and an inhibitory activity of cellular tyrosinase (26.6%) at 25 µg/mL. The LC-MS/ MS analysis of the spray-dried extract displayed the main and minor phenolic compounds constituting this sample. The results indicate that P. nitida extract has promising features for the development of cosmetic formulations.


INTRODUCTION
The Amazon Rainforest has an enormous diversity of plants that remains primarily underused and which has a great potential for industry and research. The search for new assets for cosmetic use has aroused great interest Priscilla T. Ribeiro, Tatiana N. Pedrosa, Francisco C. M. Chaves, Lucindo J. Quintans-Júnior, Adriano A. S. Araújo, Marne C. Vasconcellos, Silvya S. Maria-Engler, Cláudia C. Silva, Felipe M. A. Silva, Héctor H. F. Koolen, Emerson S. Lima, Ádley A. N. Lima coumarins in the methanolic extract (Carvalho et al., 2010). Another study demonstrated the antioxidant and antimicrobial activity of P. nitida, which seems to have a relationship with the phenolic content (Bendini et al., 2006). Furthermore, the plant extract is used in traditional medicine to treat gastrointestinal disorders (Carvalho et al., 2010).
Other studies carried out with the Passif lora genus demonstrated their potential in cosmetic use, presenting antioxidant activities, sun protection ability, skin whitening by inhibition of the tyrosinase enzyme, and collagen synthesis (Lourith, Kanlayavattanakul, 2013;Maruki-Uchida et al., 2013). Despite the relevant evidence of the use of Passiflora in skin abnormalities, few studies on this use were found for P. nitida in the literature. To our knowledge, no studies have been done on characterizing the plant's raw material, nor on obtaining the dry extract.
This work aimed at characterizing the species leaves' raw material, dry extract production, the study of the antioxidant potential, and application of this extract for cosmetic formulations.

Material
P. nitida Kunth leaves were collected from EMBR APA (Empresa Brasileira de Pesquisa Agropecuária). The voucher specimen was deposited in the INPA (National Institute of Amazonian Research) herbarium under No. 209547. Access to botanical material was registered in the SISGEN (National System of Management of Genetic Heritage and Associated Traditional Knowledge) under the number A3965C3. All solvents and chemical reagents used in the experiments were of analytical purity.

Granulometric analysis through sieving
Approximately 25.0 g of dry leaves were crushed using a knife mill, weighed and passed through sieves with mesh openings of 2.00, 1.00, 0.850, 0.710, 0.600, 0.500, 0.425, 0.355, 0.250 mm and a collector. The sifting was performed through sieves at 60 vibrations per second for 15 min, following the methodology described in the Brazilian Pharmacopoeia 5 th edition (Anvisa, 2010). The fractions retained in the sieves and collector were weighed and analyzed through a graphical method. Histogram distribution was constructed in order to obtain the distribution of the particles. By means of passage and retention curves, the average particle size was calculated. The results represent the mean of triplicates (Figueiredo et al., 2014).

Determination of loss through drying
This was determined following gravimetric method described in the Brazilian Pharmacopoeia 5 th edition (Anvisa, 2010). The assay was performed in triplicate.

Extractive content
About 1.0 g of ground dry leaves was transferred to an Erlenmeyer flask with 100 mL of distilled water. The mixture was weighed and subjected to extraction by decoction for 10 minutes. After cooling, it was weighed again and reconstituted to the initial mass with distilled water and filtered discarding the first 20 mL of filtrate. About 20 g of the filtrate was weighed and placed on weight filter that had been previously weighed, and evaporated to dryness in a water bath, with occasional stirring. After evaporation, the flat-bottomed dish containing the sample was placed in an oven at 105 ºC ± 2 ºC until reaching a constant weight, and the solids were then calculated. The result was expressed in triplicate and the extractive yield was calculated according to the equation below (Bundesvereinigung, 1986): (1) where: EC = extractive content (%, m/m); g = mass (g) of dry residue, FD = constant, equal to 5; m = mass of the initial sample (g); Ld = Loss through drying of the sample (m/m).

Total ashes analysis
Physicochemical characterization and cosmetic applications of Passiflora nitida Kunth leaf extract About 3 g of sample, accurately weighed, were transferred and distributed evenly in porcelain crucibles that were previously tared. The samples were incinerated in a muffle furnace at 500 ºC until all the organic matter was eliminated, then the crucibles were cooled in a desiccator cabinet and weighed, following the methodology described in the Brazilian Pharmacopoeia 5 th edition (Anvisa, 2010). We calculated the percentage of ash in relation to the dried drug. The analysis was performed in triplicate.

Extractive solution preparation
To prepare the extraction solution, we used P. nitida leaves that had been previously dried in a circulating air oven and then ground in a knife mill. The extraction solution of 5%, 7.5% and 10% (w/v) was obtained using, respectively, 50 g, 75 g and 100 g, to 1000 mL of hydroethanolic solvent at 20% (v/v). The extractive solutions were macerated for 48 h under agitation and at room temperature. Subsequently, they were subjected to both regular and vacuum filtering.

Dry extract through spray drying (DEPN)
The volume of 1,000 mL of extractive solution 5 %, 7.5 % and 10 % (w/v) was subjected to a spray dryer (LM MSD 1.0, Labmaq of Brazil Ltda), under the following conditions: inlet and outlet temperature 140/100°C, 85 % flow, flow rate 1.2 mL•min -1 and 0.07 mm nozzle. Thus, three kinds of DEPN were prepared.

Specific surface area and pore diameter (S BET )
Specific surface area and mean pore diameter of P. nitida extracts were determined through the nitrogen adsorption technique by applying the model proposed by Brunauer, Emmett and Teller [BET], following the methodology described by (Maximiano, Costa, Souza, 2010) and USP 30 th edition (USP, 2008). A sample of about 200 mg of the drug was transferred to the sample tube and was degassed for 72 hours at 120 °C to remove any material adsorbed within the pores and on the surface of the material. An ASAP 2420 accelerated surface area and porosimetry system was used, which was equipped with software for determining the surface area (SBET) and porosity.

Particle size distribution through light scattering (laser granulometry)
The sample was dispersed in a solution of surfactant Triton X-100 at 0.02 % and then subjected to ultrasound for 3 minutes. In order to analyze the particle size distribution, a Microtrac S3500 laser diffraction particle size analyzer was used, with a measurement time of 10 seconds, flow at 70% by volume, and distribution using the Fraunhofer method.

Differential scanning calorimetry (DSC)
The analysis of differential scanning calorimetry was performed using a calorimeter (DSC-50, Shimadzu, Tokyo Japan), operating at a temperature of 25-500 °C. Approximately 3 mg of the sample were placed into flat aluminum caps and then subjected to a heating rate of 10 °C•min -1 under a synthetic air atmosphere (50 mL•min -1 ). The calorimeter was calibrated with the reference standards, indium and zinc. Heating-coolingheating cycles were also performed in order to confirm the thermal phenomena (Araújo et al., 2006).

Thermogravimetry analysis
The thermoanalytical characterization was performed in duplicate using a thermobalance (model TGA 50, Shimadzu, Tokyo Japan), under nitrogen atmosphere flow at 50 mL•min -1 , and the mass of the sample of about 5 mg (± 0.4) was heated in an aluminum crucible over a temperature range of 30-600°C, at a heating rate of 10 °C•min -1 . Before the tests, instrument calibration was performed using the reference standards aluminum and zinc (Araújo et al., 2006).

Wave dispersive X-ray fluorescence (WDXRF) analysis
The samples were impregnated in a filter paper, suitable for analysis. Data was collected from a sample of white filter paper for reference during the analysis process. The samples had their data collected under the same conditions as the filter paper. The water and alcohol used in the extract preparation, as well as cotton samples and filter paper used in the filtration of the extracts were also analyzed.
For all analyses, the wave dispersive x-ray fluorescence (WDXRF) method was used, using the equipment Rigaku ® Supermini, with a palladium tube and 200 s of exposure time and 200 W of power. The conditions were adjusted, taking into account the sample matrix, sampler and sample holder. All elements were identified by their Kα and/or Kβ energy (Janssens, 2003).
Geological references, such as GBW 3125, 7105 and 7113, were used for equipment calibration. Through mathematical software, the emission peaks were related to the respective concentrations of a particular element. Each element found was quantified using an external standard of salts of a known purity, diluted boric acid, also with known purity of at least six predetermined concentrations, which were subjected to the same conditions as the analysis samples.

Total phenol content
The determination of total phenolic content present in the DEPN was performed through spectrophotometry using the Folin-Ciocalteau reagent, according to Singleton, Rossi (1965) with modifications. First, 1 mg•mL -1 solutions were prepared from the dry extracts of 5%, 7.5% or 10% (w/v). In 96-well plates, 50 μL of Folin-Ciocaulteau reagent (Sigma) was added to 10 μL of extract/gallic acid/ water. To eliminate the reading of the color of extracts and gallic acid, 10 μL of the extract or gallic acid was added to 50 mL of water. The microplate was incubated at 37 °C for 8 min and the first reading using a plate reader (DTX 800, Beckman Coulter) was performed at 715 nm. After this reading, 240 μL of sodium carbonate at 0.4% was added. The plate was incubated for three min and a second reading was performed. The test was performed in triplicate. The polyphenol content (PC) in mgEAG by mg of extract was calculated from the formula generated by the curve of gallic acid and converted to be expressed as a percentage using the formula below: (2)

Total flavonoid content
The total flavonoids were determined according to the methodology used by Chang et al. (2002), with modifications. First, the extracts were diluted with ethanol at a concentration of 1 mg•mL -1 . Then, 30 μL of the extract and/or quercetin standard (diluted in 80% ethanol at a concentration of 1 mg•mL -1 ) and 90 μL of ethanol were added to each well of the microplate. Afterwards, 6 μL of aluminum chloride 10% and 6 μL potassium acetate (1 mol•L -1 ) were added to both and diluted in 168 μL of distilled water. In addition, for the blank, 30 μL of the extract and/or standard quercetin were used, as well as 270 μL of ethanol. The plate was then incubated for 30 min at room temperature and the absorbance assessment was performed at a wavelength of 405 nm in an microplate reader (DTX 800, Backman Counter). The analysis was performed in triplicate, and the results were obtained using, first, the difference between the absorbance of the sample and the standard absorbance by the respective blank. Subsequently, the percentage concentration of flavonoids was calculated: (3)

DPPH scavenging activity
The antioxidant activity was determined by evaluating the ability of the antioxidants present in the samples to scavenge stable radical DPPH, according to the methodology by Molyneux (2003), with modifications. A total of 270 μL of DPPH (1 mg•mL -1 ) and 30 μL of sample solution and/or standard quercetin were first added to each well of the microplate at a concentration of 100 mg•mL -1 , with their respective dilutions. For control, 270 μL of DPPH and 30 μL of ethanol were used. The plate was incubated for 30 min at room temperature in the dark and the reading performed at 517 nm. The calculations of the inhibition were made using the following formula: (4)

ABTS radical scavenging activity
The determination of antiradical activity was carried out following the methodology described by Re et al. (1999) with minor modifications. The solution was initially prepared by the reaction of ABTS: 0.7 mM of ABTS radical dissolved in 5 mL of deionized water with 5 mL of 5 mM potassium persulfate. The reaction mixture was incubated at room temperature and in the dark for 12 h. The solution obtained an oxidized greenish-blue tint. Afterwards, the solution was diluted in a proportion of 1:7 solution/deionized water, which was then adjusted for an absorbance of 1.000 ± 0.1 at a wavelength of 630 nm. A 30-μL aliquot of the sample solution and 270 µL of the ABTS solution were added to microplates, in triplicate. In the control group, the same volume of diluent and ABTS was added. The samples and controls were incubated for 15 min in the dark, after reading was performed in a microplate reader. The standard used in this determination was quercetin. The calculations of the inhibition were made using the following formula: (5)

Tyrosinase inhibition
The inhibition of tyrosinase by the extracts was evaluated following the method described by Hearing Jr. (1987). The extracts were prepared at a concentration of 10 mg•mL -1 and tested as follows: 20μL of the inhibitor was added to the microplate [extract, diluent standard (control) or (ac. Kojic)]. Afterwards, 80 μL of the enzyme tyrosinase solution was added and left to incubate for 5 min at 37 °C. Subsequently, 100 μL of DOPA color reagent was added, which was subjected to the immediate reading times: 5, 10, 15 and 20 min, with incubation of the plate at 37 °C. The plates were read in an ELISA plate reader and tyrosinase inhibition was calculated using the formula below: (6) Tyrosinase activity and melanin content assay B16F10 cells were stimulated with IBMX (Isobutylmethylxantine) 25 µM for 24 h and then treated with 25 µg/mL of P. nitida for 48 h. Subsequently, the cells were placed in two microtubes (one for melanin content with 1 X 10 6 cells/microtube) and lysed with Triton-X 100 in PBS with 1mM PMSF (phenylmethanesulfonyl fluoride). Next, this solution was centrifuged at 10,000 g at 4 °C for 10 min to obtain the supernatant and the pellet for the measurement of tyrosinase activity and melanin content, respectively.
Tyrosinase activity was estimated by measuring the rate of L-DOPA oxidation, using the method reported by Tomita & Tagami (1992) with modifications. Protein concentrations were determined by the Bradford method, using bovine serum albumin (BSA) as standard. A total of 100 µL of supernatant containing 100 µg of total proteins were added to each well in a 96-well plate, and then mixed with 100 µL of L-DOPA at 3 mg/mL. After incubation at 37°C for 90 min, the dopachrome was monitored by measuring the absorbance at 475 nm.
Melanin content was measured as described by Hosoi et al. (1985) with modifications. After the lysis, the pellets of the B16F10 cells (1 X 10 6 ) were mixed with 250 μl of 1 N NaOH containing 10% DMSO for 1 h at 95 °C. The absorbance at 405 nm was measured using a microplate reader. The melanin content was determined from a standard curve prepared from a standard of synthetic melanin (Sigma, USA).

Phytochemical analysis
An LC-MS system (6,550 iFunnel, Agilent Technologies, Santa Clara, CA, USA) consisted of a liquid chromatography system coupled to a quadrupole time-offlight mass spectrometer equipped with an electrospray ionization (ESI) source was used to explore the phenolic composition of extracts using an untargeted metabolomic approach (Koolen et al., 2013). Chromatographic separation was performed in a Poroshell column (120 EC-  (Bataglion et al., 2015). Results were expressed by means of µg/g of spray-dried extract (SDE).

Granulometric analysis through sieving
From the analysis, it was noted that most of the particles have a particle size range of 0.25 to 0.71 mm. The powder is classified as a coarse powder in accordance with the classification of the Brazilian Pharmacopoeia 5 th edition, which refers to particles that can pass in their entirety through a 1.70-mm sieve mesh and a maximum of 40% powder through 0.35-mm sieve mesh (Anvisa, 2010). The average particle size of the pulverized leaves dry was 0.506 (± 0.087) mm.
The granulometric evaluation is an important parameter because it has a direct influence on the efficiency of the extraction procedure (Migliato et al., 2007). According to Sharapin, powders of a coarser nature are highly recommended for most botanical material because very fine powders can compromise the extraction (Fonseca, Silva, Leal, 2010;Sharapin, 2000). Therefore, the powder from P. nitida leaves showed good results regarding their particle size, since the range of 1.70 to 0.35 mm is considered coarse powder and, as such, favors the extraction of active compounds.

Determination of loss through drying
The average loss through drying (9.08 ± 0.08) and the coefficient of variation (0.9 %) is in accordance with the specification limit for botanical material humidity, i.e., 8 to 14%, as established by the Brazilian Pharmacopoeia 5 th edition (Anvisa, 2010). The moisture content is an important parameter to be evaluated, since excess moisture may allow the development of fungi and bacteria in the botanical material, besides favoring the action of enzymes whose activity can lead to degradation of the chemical constituents (Fonseca, Silva, Leal, 2010). The presence of excess moisture is directly related to the correct storage of the raw material (Migliato et al., 2007). Thus, the results are within the established parameters for minimizing microbial contamination and chemical degradation and demonstrate the correct storage of the material.

Extractive content
The average of extractive content was 17.16 ± 0.05 and the coefficient of variation was 0.32%. This study was employed exclusively in order to assist in testing physicochemical characterization of the botanical material, but it is also related to water-soluble compounds present in the raw material, such as sugars, amino acids, glycosides and flavonoids, among others (Barni, Cechinel Filho, Couto, 2009).

Total ash
Determination of ash content allows the verification of non-volatile inorganic impurities that may be present as contaminants (Barni, Cechinel Filho, Couto,2009). Using the Brazilian Pharmacopoeia 5 th edition as a reference (Anvisa, 2010), which sets the maximum limit of 10% for total ash in two other species, the Passiflorae dulcis folium and Passiflorae acetum folium, the result of the ash content in P. nitida was 5.43 ± 0.11 and the coefficient of variation was 1.99%, suggesting a low amount of nonvolatile inorganic contaminants present in the sample.

Specific surface area and pore diameter
Allen, Popovich, Ansel (2007) reported that a small particle size leads to an increase in the specific surface of the powder. Correlating the results of the specific surface area with a pore diameter shown in Table I and the results of the laser granulometry that indicated a very fine powder, the three kinds of DEPN presented a high specific surface area and the pores are classified as mesopores.
The porosity of a material determines important physicochemical properties such as adsorption, mechanical strength, dissolution characteristics and wettability (Porte, Leão, Porte, 2011;Schoonman et al., 2001). The DEPN showed good results in relation to porosity. It is useful for a dry extract in pre-formulation or formulation that the porosity be in the range 20-500 Å, which ranks as mesopores, and confers increased specific surface area, greater cohesion between the particles and wettability. All these characteristics are important and should be considered in the formulation of the development phase.

Differential scanning calorimetry (DSC)
The DSC curve of the DEPN from the 5% (w/v) extractive solution showed four thermal events ( Figure  1). The first event corresponds to its loss of water, the second and third events are related to thermal decomposition and the fourth event is associated with the removal of carbonaceous material. The DEPN from the 10% (w/v) extractive solution had well-defined endothermic events between 25 °C and 200 °C. These results were confirmed by the TG/DTG curves (Araújo et al., 2006).
In the DSC curve of the DEPN from the 7.5% (w/v) extractive solution, two endothermic events were observed: the first between 25 °C and 100 °C, and the other above 400 °C. However, the TG/DTG curves of the DEPN from the 7.5 % (w/v) extractive solution had four weight loss events. Thus, the two intermediate events in the DSC curve of the DEPN from the 7.5 % (w/v) extractive solution are a junction of endothermic and exothermic events, making it impossible to clearly observe any changes in the baseline.

Analysis of the particle size distribution
The distribution range of the particle size, the retention rate and the average particle size of the dried extract from the 5% (w/v), 7.5% (w/v) and 10% (w/v) extractive solutions are displayed in Table I. The average size of particles found in the DEPN classify it as a very fine powder, in accordance with the classification of the post by the Brazilian Pharmacopoeia 5 th edition, which passes in its entirety through a sieve with a 125-μm nominal aperture (Anvisa, 2010). A comparison between the different kinds of DEPN demonstrated that 7.5% had the smallest average particle size (25.06 μm), and 95% of the particles have a size of ≤ 59.74 μm. The smaller particle size is desirable because it influences many factors such as rate of dissolution, content uniformity and stability. It also reduces the gritty sensation in formulations for topical use (Allen, Popovich, Ansel, 2007).

Thermogravimetry (TG)
The TG/DTG curves of the three types of DEPN studied (Figure 2) showed four events of mass loss, of which the first event occurred between 25 and 150°C and is related to loss of surface water; this result is consistent with endothermic events that appear on the DSC curve. It is possible to observe three thermal events in the TG/ DTG curve after dehydration. The first, between 150 and 400 °C, is associated with the thermal decomposition of organic compounds. The second stage of thermal decomposition, with the formation of carbonaceous material, occurs between 400 and 600 °C. The latter event happens above 600 °C, which is related to the loss of carbonaceous material formed in the previous step. Although there is a difference in the curve of the DEPN from the 10% (w/v) extractive solution, it has a derivative tending to zero because of the continuous loss of decomposition products (Araújo et al., 2006;Melo et al., 2013). FIGURE 1 -Differential scanning calorimetry (DSC) of the DEPN 5%, 7.5% and 10% extractive solutions in a synthetic air atmosphere (50 mL•min -1 ) at a heating rate of 10 °C min -1 .
Physicochemical characterization and cosmetic applications of Passiflora nitida Kunth leaf extract

Wave dispersive X-ray fluorescence (WDXRF)
Many nutrients found in plants are absorbed from the soil and, among these, heavy metals or other elements may exist that cause disease or toxicity. Thus, to ensure the safe use of a plant raw material, the identification and quantification of the elements is of great importance. Various elements can be quickly found using the WDXRF technique (Rüdiger, Silva, Veiga-Júnior, 2009).
In all three types of the DEPN examined, the following chemical elements were found: Ca, K, Cl, S, P, Si, Mg, Na and Br, which probably originate from the soil from which the plant was removed. It was expected that the concentrations of the elements would increase with the increase in the concentration of the extracts, but it can be clearly seen that the 7.5% extract has a higher concentration of the elements observed, as shown the Table II. It is believed that this fact is related to the extraction process, where the concentration of 7.5% can be considered optimal for better extraction. Other data presented in this work, such as laser particle size, phenol and flavonoid content, among others, show the same standard of behavior, confirming the WDXRF data.  Table III shows the results of total phenolic content and total flavonoids for the DEPN. P. nitida presents a phenol content that is higher than that found in the species P. edulis and P. alata (Zhang et al., 2013).

Determination of total phenolic content and total flavonoids
It is also interesting to note that increasing botanical material concentration does not necessarily mean that we will have a dry extract with a higher quantity of phenols and flavonoids. The results found in the DEPN from the 10% (w/v) extractive solution demonstrate this point since the levels were lower compared with the extracts from the 5% (w/v) and 7.5% (w/v) extractive solution. This occurs probably due to the saturation of the medium during the extraction process.  Table III shows the results of antioxidant activity of the DEPN. The analysis shows that there is antioxidant activity, though there are no significant differences in the results with the change in the types of extracts. These results, when compared to the values found for ascorbic acid, which was used as a standard, proved to be ± 10 times lower for DPPH and ± 5 times lower for ABTS. However, the antioxidant activity found in the DEPN was very good, especially compared to what has been reported for other species of Passiflora. For example, Zhang et al. (2013) reported that P. edulis presents DPPH values between 875 and 1,100 mg•mL -1 and P. incarnata between 76 and 282 mg•mL -1 . In a previous study, Bendini et al. (2006) demonstrated excellent antioxidant activity of P. nitida through DPPH and ABTS assays.
Phenolic compounds are usually directly correlated with antioxidant activity, and their ability to donate hydrogens, which allows scanning of free radicals (Yilmaz et al., 2013). Flavonoids are polyhydroxylated phenolic compounds that have antioxidant properties, which include free radical termination and chelating action of metals, besides inhibiting several enzymes. These multiple functions are related to their distinct chelating action, which occurs through their hydroxyl groups (Ziaullah, Warnakulasuriya, Rupasinghe, 2013). Previous studies have demonstrated the antioxidant activity of some flavonoids, such as orientin, isoorientin, vitexin and isovitexin, using HPLC in combination with the analytical results of DPPH (Demirkiran et al., 2013;Yao et al., 2012).
The antioxidant activity of P. nitida is related to phenols and flavonoids present in its composition. Teixeira et al. (2014) found four major peaks by using HPLC in the analysis of flavonoids from P. nitida leaves. Three flavonoids were identified, namely vitexin, swertisina, orientin, and one peak could not be identified when compared to the most common flavonoids.

Tyrosinase inhibition
In evaluating the inhibition of tyrosinase, the IC50 showed that the DEPN 5, 7.5 and 10% were respectively 654.4, 436.9 and 400.1 µg•mL -1 . The results of the standard, which is kojic acid (KA), was 1.0 ± 0.1 µg•mL -1 ( Figure 3B). Despite the difference between the standard and the extract, the DEPN inhibits tyrosinase. This inhibiting activity is related to the flavonoids present and their metal chelating activity. The mode of inhibition of flavonoids is typically a competitive inhibition of the oxidation of L-DOPA by tyrosinase, and a key role is attributed to the chelation of copper. Flavonoids are the most investigated secondary metabolites of plants that show inhibition of tyrosinase; and quercetin, kaempferol, myricetin and morin act as competitive inhibitors of tyrosinase. Studies have shown that vitexin and isovitexin inhibit tyrosinase (Chang, 2009;Ziaullah, Warnakulasuriya, Rupasinghe, 2013;Teixeira et al., 2014;Yao et al., 2011). It has also been demonstrated that DEPN is a promising natural skin-whitening agent and could be used by the cosmetic industry for smoothing benign pigments (Lourith, Kanlayavattanakul, 2013).
Physicochemical characterization and cosmetic applications of Passiflora nitida Kunth leaf extract

Tyrosinase activity and melanin content
The inhibitory effect of P. nitida Kunth on the tyrosinase activity of B16F10 cells treated with 25 μM IBMX ( Figure 3C) inhibited tyrosinase activity by 26.6 % when compared with cells treated with IBMX alone. When compared with the standard, rutin, the cellular tyrosinase activity was not statistically significant ( Figure 3A). Melanin synthesis was effectively inhibited ( Figure 3D). The P. nitida Kunth dried extracts showed a statistically significant decrease in melanin content (27.1%), thus indicating depigmentation activity.
These results demonstrate that the depigmentation activity in P. nitida is due to inhibition of tyrosinase. Kojic acid is reported as an excellent inhibitor of mushroom tyrosinase and studies reported low activity in B16F10 cells and the need for high concentrations and combination with other depigmenting substance in order to have a strong FIGURE 3 -(A) Inhibitory activity of tyrosinase in P. nitida (IC50: 0.44 ± 0.2) and (B) standard kojic acid (IC50: 1.0 ± 0.01). (C) Inhibitory effect P. nitida Kunth on tyrosinase in B16F10 cells. Cells were stimulated for 24 h in the presence of 25 μM IBMX and, subsequently, B16F10 cells were treated with P. nitida Kunth and the standard, rutin (RUT), at 25 µg.mL -1 for 48h. The difference in cellular tyrosinase activity between P. nitida and RUT was not statistically significant. (D) Effects of P. nitida on cellular melanin content. Melanin contents were determined in B16F10 cells treated with P. nitida and the standard, RUT, at 25 µg.mL -1 for 48h after stimulation for 24 h with 25 µM of IBMX. Unstimulated cells with IBMX (WS -without stimulation) were also quantified for comparison. Melanin levels were measured at 405 nm. (C), (D) Values represent mean ± SD (n=3). ***p < 0.001 compared to cells treated with 25 μM IBMX. effect in vivo (Ahn et al., 2011;Cho et al., 2012). Because of this, rutin was used as standard on cells in tests. In addition, rutin has been reported as a metabolite that is present in fruits and plants with great depigmentation potential (Drewa et al., 1998;Si et al., 2012).
The data show that the DEPN has greater cell penetration activities and that these are higher than in KA, which suggests a greater ability of the extract to penetrate the cells and inhibit tyrosinase. This activity is possibly related to the flavonoid compounds found in P. nitida leaves.

Phytochemical analysis
The qualitative chemical analysis by LC-MS/MS of the spray-dried extracts from P. nitida (Figure 4) showed that the phenolic composition is dominated by two main flavonoid glucosides, tentatively identified as vitexin (1, m/z 431.0962) (Zucolotto et al., 2012) and kaempferol-3-O-galactosyl-rhamnosyl-glucoside (2, m/z 755.2071). Compound 1 was previously identified in the leaves of P. edulis, P. tripartita, and P. manicata (Zucolotto et al., 2012), while compound 2 has not been previously identified in any Passiflora species, however, it has been described in oolong tea extracts (Dou et al., 2007) and Camelia sinensis (Fraser et al., 2013). The crude hydroethanolic extract from the leaves of P. nitida displayed a similar composition of a previous study (Teixeira et al., 2014), being identified as C-glucoside flavonoids isoorientin, vitexin, and swertinin. Since the metabolite composition of the spray-dried extract displayed just two phenolic compounds, a targeted quantitative analysis was performed to investigate the minor compound of the obtained sample.

CONCLUSIONS
Dried leaf extracts of P. nitida can be used as plant raw material for developing new products with antioxidant or depigmenting properties. This characterization allowed for the determination of parameters for quality control within pharmacopoeia specifications and for the use of new techniques in characterizing the material with improved understanding of the new raw material. Moreover, the dry leaf extract of P. nitida shows moderate antioxidant and tyrosinase inhibitory activities, although further studies of efficacy and safety are required. Among the three DEPN studied, this was the best choice of extract for future use in developing formulations from the dry extract of P. nitida 7.5% (w/v). Therefore, the results of this study can be used as the basis for developing potential dermocosmetics containing P. nitida Kunth leaves.