Quantitative and Qualitative HPLC Analysis of Mycosporine-Like Amino Acids Extracted in Distilled Water for Cosmetical Uses in Four Rhodophyta

Mycosporine-like amino acids (MAAs) have gained considerable attention as highly active photoprotective candidates for human sunscreens. However, more studies are necessary to evaluate the extraction efficiencies of these metabolites in cosmetic compatible solvents, as well as, their subsequent HPLC analysis. In the present study, MAA extraction using distilled water and 20% aqueous methanol in four Rhodophyta was investigated. Different re-dissolution solvents and a C8 and C18 columns were tested for the HPLC analysis. Porphyra-334, shinorine, palythine, palythine-serine, asterina-330, and palythinol were identified by HPLC/ESI-MS. The separation of these MAAs were improved employing the C8-column, and using methanol as re-dissolution solvent. Regarding total MAAs concentrations, no differences between the two solvents were found. The highest MAA amounts were observed injecting them directly in the HPLC. According to these results, distilled water could be an excellent extraction solvent for MAAs. Nevertheless, the re-dissolution in pure methanol after dryness would be the best option for the qualitative analysis of the most common MAAs in these red algae. Our results entail important implications regarding the use of red macroalgae as promising candidates as environment-friendly sources of natural sunscreens.


Introduction
Solar radiation exposes the intertidal marine macroalgae to elevated irradiances of UV-B (280-315 nm) and UV-A (315-400 nm). In fact, depletion of the ozone layer in the stratosphere during the past two decades has intensified the levels of solar UV-B radiation, which has reached unexpected levels that can be harmful for many biological processes [1][2][3]. The UV radiation can produce several detrimental effects on biologically important molecules such as lipids, proteins, or DNA [4][5][6].
The marine organisms that are exposed to UV radiation can accumulate natural UV-absorbing compounds [7]. Among others, mycosporine-like amino acids (MAAs) [8,9] have drawn special attention as molecules act as antioxidants and provide photoprotection [10,11]. The main producers of these molecules in the marine coast are the red algae [12][13][14].
MAAs are a family of intracellular compounds of low molecular weight (<400 Da), they are distributed in a wide range of marine organisms, are water-soluble, and have high molar extinction   Chromatograms of P. leucosticta have been selected using C8 column ( Figure 2) and C18 column ( Figure 3). Chromatograms for the other species are shown in Figures S5-S10 (Supplementary Materials). In general, the use of the C18 column resulted in a deficient MAA separation ( Figure 3). In three of the seven procedures, we could not identify different MAAs, so we observed a few mixed unidentifiable peaks. We had only been able to identify different MAAs using HPLC eluent or distilled water as a re-dissolution solvent, regardless of extraction solvent. The most used method for MAA analysis based on extract with 20% methanol and re-dissolute in 100% methanol (M4 protocol) resulted in a mix of MAAs using the C18 column. Regarding the MAA analysis using the C8 column, all protocols led to the separation of different MAAs. In fact, the protocols in which an improved MAA separation were M1, M4, and W4 for C8 column, while the separation was more efficient under M2, M3, W1, and W3 procedures for C18 column. However, none of them could separate all MAAs. According to these results, the incoming data presented in this work have been obtained applying the different procedures using the C8 column.
The qualitative distribution of MAAs for each species was determined by ESI-MS and up to six different MAAs were identified in the analyzed samples: Porphyra-334, shinorine, palythine, asterina-330, palythinol, and palythine-serine (Table 1). Also, the relative proportions expressed in percentage were calculated in each alga according to the total concentrations obtained applying the procedures M4 and W4 (Table 1). Porphyra-334, shinorine, and asterina-330 were common in all studied species. Palythine and porphyra-334 were the dominants MAAs in A. vermiculophyllum, shinorine in C. corneus, palythinol in G. longissima, and porphyra-334 in P. leucosticta. Asterina-330 was the minority MAAs in all studied species except P. leucosticta, and palythine-serine was only present in C. corneus. Palythinol was only identified in three of these species: G. longissima, C. corneus, and P. leucosticta.

Figure 1.
Extraction and re-dissolution protocols used to analyze mycosporine-like amino acids (MAAs); 20% aqueous methanol was used as extraction solvent in M1, M2, M3, and M4. Distilled water was utilized as extraction solvent in W1, W2, and W3. Number 1 indicates HPLC analysis done without re-dissolution. Number 2: Re-dissolution in distilled water. Number 3: Re-dissolution in HPLC eluent. Number 4: Re-dissolution in pure methanol.       In addition, the flavonoid myricetin was identified by ESI-MS in A. vermiculophyllum, but no chromatographic peak was found associated to this molecule in any protocols or columns.
Using the C8 column, all MAAs described by ESI-MS were detected in A. vermylophyllum under all procedures. In P. leucosticta, only the M1, M4, and W4 protocols allowed a good identification ( Table 2). In C. corneus and G. longissima, the protocols M4 and W4 offered similar results, i.e., only extraction in distilled water or 20% aqueous methanol and subsequent re-dissolution in 100% methanol must be accepted for the qualitative MAAs analysis. According to the obtained data, M4 and W4 would be the most appropriate procedures for the qualitative HPLC analysis of the following MAAs: Porphyra-334, shinorine, palythine, asterina-330, palythinol, and palythine-serine.
In general, total MAA concentrations did not differ between C18 and C8 columns (data not shown for C18). The concentration of MAAs calculated depended on the procedure in the four species ( Figure 4). For C8 column, the highest MAA concentrations were observed in M1 and W1 procedures in the four Rhodophytes (p < 0.05). And no significant differences were found using distilled water or 20% aqueous methanol as extraction solvents in any species (Figure 4). The average value of MAA concentration for M1 and W1 procedures was 9.6 ± 0.04 mg g −1 DW in P. leucosticta, followed by C. corneus (2.11 ± 0.1 mg g −1 DW), G. longissima (0.69 ± 0.09 mg g −1 DW), and A. vermiculophyllum (0.4 ± 0.09 mg g −1 DW). In general, M4 and W4 protocols offered the lowest MAA concentrations (p < 0.05) in the most studied algae. The average value of MAA concentration for these two procedures was 4.3 ± 0.26 mg g −1 DW in P. leucosticta, 1.19 ± 0.68 mg g −1 DW in C. corneus, 0.26 ± 0.07 mg g −1 DW in G. longissima, and 0.1 ± 0.04 mg g −1 DW in A. vermiculophyllum. We found a tendency of progressive increases in MAA concentration as the percentage of methanol in the re-dissolution solvent decreased ( Figure 4).
In the Table 3, the strengths and weaknesses of each protocol used in this work are summarized. Table 2. Mycosporine-like amino acids (MAAs) identified in each analyzed species and protocols for their identification by HPLC using the C8-Luna column.

Discussion and Conclusions
It is well known that MAAs have high bioactive activities, therefore, there is a high interest to include MAAs in cosmetic formulations [20]. The potential uses of MAAs as sunscreen and antioxidant additives in cosmeceutical products implies the use of extraction solvents with no certified toxic effects and that allow the preservation of the physico-chemical and absorption properties of these molecules. MAAs have high solubility in aqueous solution; therefore, distilled water could be a good solvent in terms of extraction efficiency.
In order to study the influence of distilled water, as a solvent of extraction in MAA analysis, four Rhodophytes were selected based on their medium-high MAA contents and their different MAA profiles [17,31,32]: Agarophyton vermiculophyllum, Crassiphycus corneus, Gracilariopsis longissima, and Pyropia leucosticta. We selected different procedures and two HPLC columns that are commonly used for MAA analysis in order to elucidate which procedure would be the best option for the isolation, identification, and quantification of MAAs for cosmetic purposes.
Different solvents had been used for MAA extraction, the most common are ethanol and methanol from 20% to 80% [14,17,30,33,34]. However, other solvents such as 2-octyl dodecanol, octyldodecyl ester of l-pyrrolidone carboxylic acid (OEL-PCA) (cosmetic emollient and moisturizer), and ethyl acetate revealed that UV absorption spectra of the extracts, number of elution chromatographic peaks, and their maxima absorption were affected by the nature of the solvents, being the extract in OEL-PCA, which showed the poorest UV absorption, and only palythenic acid could be identified in these extracts [35]. In our study, we compared the extraction in 20% aqueous methanol and in distilled water and no significant differences between them in terms of total MAA concentrations were found when these extracts were injected directly into HPLC. Furthermore, under this treatment, no hypsochromic shifts in UV-absorption maxima of the identified MAAs were observed.
Regarding qualitative and quantitative analysis of MAAs, HPLC-MS is the most used technique nowadays [36]. However, other techniques had been used such as capillary electrophoresis [33]. The most common columns used to separate and characterize MAAs by reverse-phase high-performance liquid chromatography (HPLC) are a monomeric octylsilica C8 column [29] and a low silanol-free group octadecylsilica C18 column [16,30,34]. The most common HPLC eluent contains a mixture of distilled water:methanol with a slow percentage of acetic acid run isocratically. Except for the technique of Carreto et al. [37], there is no isolation and identification MAA protocol able to completely isolate large MAAs mixtures on marine organisms. The method used by Carreto et al. [37] combines a C18 column with an acidic eluent. Although this approach is able to separate high mixtures of MAAs, it is a time-consuming analysis that we recommend only in case of samples with a complex mix of MAAs.
We have analyzed the 20% aqueous methanol and distilled water MAA extracts, using a Luna-C8 and a Poroshell-C18 columns. The characteristics of them are different; while the C8 have 5 µm particle size and a dimension of 4.06 mm x 250 mm, the C18 column have 4 µm particle size and a dimension of 3 mm x 250 mm. Therefore, besides the different material of each column, we have different particles sizes and thickness. Thus, this is why we had to decrease the flow using C18 from 0.5 mL/min to 0.3 mL/min. Recently, Rosic et al. [34] described a general method for the isolation and characterization of MAA compounds from red algae and symbiotic dinoflagellates using methanolic extracts and HPLC and LC-MS with electrospray ionization source interface. They used a C18 column (1.7 µm particle size) and concluded that porphyra-334, palythine, and mycosporine-glycine could not be separated, as these mixed MAAs appeared only in an elution peak at the beginning of the chromatogram. According to the results obtained in our work, the C18 column offered the worst results in terms of MAA separation using these seven procedures, although the particle sizes of the columns were different. However, Karsten et al. [28], who used three reverse-phase C8 and C18 HPLC columns, obtained better results than us using the Synergi C18 column with the same particle size and dimensions, but acquired from different companies, Phenomenex versus Agilent. They compared, similarly to this work, pure methanol, distilled water, and HPLC eluent as re-dissolution solvent for dried Prasiola and Porphyra extracts (Pyropia former Porphyra) [28]. On the contrary, Karsten et al. [28] found analytical problems using the C8 columns, this was not the case for this study; in fact, we used the same C8 Luna column, and we obtained better results than with the C18 column. One explanation to the mixed MAAs could be related to other highly polar, low molecular weight UV-absorbing substances unrelated to MAAs that are generally present in extracts of marine organisms. These molecules may interact with the eluent solvent and MAAs presents in the samples increasing their molecular weights and changing their polarity, and therefore their elution times [30]. Matrix effect is minimized in analysis of purified and partially purified MAA extracts [20,38]. Although, why this matrix effect could occur under some conditions and not others must be investigated.
According to our results, the highest MAA concentration was obtained by injecting the sample directly in HPLC after its extraction, regardless if it is extracted in distilled water or 20% methanol. The drying and subsequent re-dissolution of the pellets in the different studied solvents declined total MAA concentrations, as occurred in Karsten et al. [28]. With the independence of solvent of extraction, the MAA concentration when the sample was injected directly into the HPLC was approximately double compared to after drying the sample and resuspending it in 20% methanol. Consequently, methanol should be avoided as a re-dissolution solvent for the HPLC analysis in terms of estimation of total MAA concentration. However, these treatments (M4-W4), together with W1, led to improved peak separations. Although, the use of methanol as a re-dissolution solvent provokes double peaks with similar absorption maxima and, therefore, limits its use for qualitative MAA analysis. This result could also be related to matrix effect [39].
Other authors had studied MAA composition and concentration in the same species included in this work. Roleda et al. [40] found, in A. vermiculophyllum (former Gracilaria vermiculophylla), two MAAs but they were not identified. Lalegerie et al. [14] found eight different MAAs in the same species, some of them unidentified, but no total MAA concentration was shown. Barceló-Villalobos et al. [32] identified four MAAs, the same as in this work, porphyra-334, shinorine, palythine, and asterina-330. However, total MAA concentration of samples collected from the field was at least two times higher than our MAA concentration. Regarding Crassiphycus corneus (formerly Hydropuntia or Gracilaria cornea), Sinha et al. [41] identified shinorine and porphyra-334, but they did not show total MAA concentration.
Figueroa et al. [42] found shinorine, pophyra-334, and palythine. The concentration in this species was very similar to our work. Álvarez-Gómez et al. [31] found four MAAs, but the concentration was much lower than those presented in this work. In another article, similar concentration and composition to that in our work were found [43,44], around 1 mg g −1 , and porphyra-334, shinorine, palythine, asterina-330, and palythinol. MAAs in G. longissima were analyzed firstly by Torres et al. [45], but they identified asterina-330, palythinol, palythene, and usujirene. However, similar concentration and composition (porphyra-334, shinorine, palythine, asterina-330, and palythinol, and a concentration around 1 mg g −1 ) have been shown in this species by other authors [31,43,44]. Finally, concentration and composition of Porphyra species previously published did not differ from the results presented here. This species has porphyra-334 as the dominant MAA, and in minor concentrations, shinorine, palythine, asterina-330, and palythinol. It belongs to the Order Bangiales, which has the highest MAA concentration [17,46,47]. To our knowledge, it is the first time in the literature in which myricetin, a flavonoid found in several foods, has been detected in one species of the Family Gracilariaceae. Other flavonoids have been found in this taxonomic group, such as camptothecin and quercetin [48]. More studies are needed to quantify myricetin in this species and investigate the regulation of its synthesis. This flavonoid has shown anti-photoaging effects through its antioxidant and anti-inflammatory properties [49,50]. The beneficial effects have encouraged researchers to develop photo-protective products from natural sources [51]. It is known that onions, kale, lettuce, tomatoes, apples, grapes, berries, tea, and red wine are rich sources of flavonols [52]. However, it is not deeply studied in algae. A recent study has described the presence of myricetin, among other flavonoids, in Caulerpa spp. [53]. The last study revealed that Caulerpa spp. are a promising functional food ingredient and could be explored as daily dietary supplements. Higher content of flavonoids in seaweeds were reported [54]. In contrast, the flavonoid content of some of them was reported to be very low [55].
This study allows us to obtain a general procedure for MAA analysis in species with similar MAA profiles than those included in this study, i.e., those containing porphyra-334, shinorine, palythine, asterina-330, and palythinol. In fact, these are the most abundant MAAs in macroalgae. Therefore, for a routine analysis of those MAAs, we suggest the following procedures: the best protocol to optimize MAA characterization is the extraction of MAAs in distilled water, followed by the extract dryness, and the re-dissolution in pure methanol. However, in order to obtain a better estimation of the total MAA concentration, the extraction in distilled water and direct injection of the extract in HPLC should be used.
Our results entail important implications regarding the use of red macroalgae as promising candidates as environment-friendly sources of industrially important compounds, like MAAs, because of their photoautotrophic properties, which can convert solar energy and carbon dioxide into useful chemicals as it has been reported by Navarro et al. [56]. MAAs, in addition to protecting against erythema, can reduce other negative biological effects as photocarcinogenesis, immunosupression CHYS, photisomerization of urocanic acid, or photo-aging [11,25]. Additionally, the genus Gracilaria has high potential as a source of high-value compounds and extracts for several applications [57]. However, only agar is commercially exploited. A better use of the biomass from commercial cultivation of species from the Family Gracilariaceae may be an important alternative for exploiting the other bioactive components, such as MAAs and flavonoids. Further studies must be carried out. Algae were dried by silica gel and stored under controlled humidity. Dried algal material was crushed to powder prior to extraction to guarantee homogeneity.

Extraction and Identification of MAAs
Analysis of MAAs was assayed according to Korbee-Peinado et al. [17] with some modifications. For the MAA extraction were used two different solvents: 20% aqueous methanol (v/v) and distilled water. Samples of dried algal (20 mg DW) were extracted for 2 h in screw-capped centrifuge vials filled with 1 mL 20% aqueous methanol (v/v) or distilled water in a waterbath at 45 • C. After this, 700 µL of the supernatant was evaporated to dryness under vacuum (Jouan evaporator centrifuge, France), except two samples of each species whose extracts in 20% aqueous methanol or distilled water were passed through a 0.2 µm membrane filter to be analyzed directly by HPLC. All samples were extracted in triplicate. In Figure 1, the seven protocols used in this work are represented. Three replicates were analyzed directly by HPLC (protocols with number 1). The other extracts were dried and re-dissolved in 700 µL 100% aqueous methanol (protocols with number 4), HPLC eluent (1.5% methanol (v/v) plus 0.15% acetic acid (v/v) in distilled water) (number 3) or distilled water (number 2), and mixed for 30 s. After passing through a 0.2 µm membrane filter, samples were analyzed with a Waters 600 HPLC system (Waters Cromatografía, Barcelona, Spain). Sample volumes of 20 µL were injected into two columns: a reverse-phase Luna C8 column (5 µm particle size; 4.06 mm × 250 mm, Phenomenex, Aschaffenburg, Germany) with a pre-column (C8, Octyl, MOS; Phenomenex) and an Infinity Lab Poroshell 120 reverse-phase C18 column (4 µm particle size; 3 mm × 250 mm, Agilent, Santa Clara, CA, USA). The mobile phase used as eluent was 1.5% methanol (v/v) plus 0.15% acetic acid (v/v) in distilled water, run isocratically at 0.5 mL min −1 with the Luna column and 0.3 mL min −1 with the Agilent column. The pressure was very high with higher fluxes in the last column.
MAAs were detected online with a Waters Photodiode Array Detector 996 at 330 nm, and absorption spectra (290-400 nm) were recorded each second directly on the HPLC-separate peaks. Identification of MAAs was performed by comparison of the absorption spectra and retention times and by co-chromatography using high purified grade MAAs extracts (porphyra-334, shinorine, palythine y asterina-330) provided from the Unit of Photobiology of the Central Service for Research Support (SCAI, University of Málaga, Málaga, Spain). Quantification was made by using published extinction coefficients [58,59] and the formula found in Pliego-Cortes et al. [60]. When we obtained a mixed unidentifiable peak, we estimated the total MAA concentration using an average extinction coefficient for those MAAs present in the species [38].
The qualitative analysis was complemented by electrospray ionization mass spectrometry (ESI-MS) with a high-resolution mass spectrometer (model Orbitrap Q-Exactive, Thermo Scientific S.L., Bremen, Germany) provided with an electrospray ionization heated probe (HESI-II), in the premises of the Central Service for Research Support (SCAI, University of Málaga, Málaga, Spain). The mass spectrometer indicated 70-700 m/z in positive mode for MAA and negative mode for myricetin with a voltage of 3.5 kV

Statistical Analysis
Values of mean concentration and their standard deviations per treatment and column were calculated from three replicates. Total MAA concentration was tested by one-way analysis of variance (ANOVA) for each species. Cochran's test was used to check the homogeneity of variances. A post-hoc Student Newman-Keuls (SNK) multiple comparison test was applied and significant differences are indicated with different letters in Figure 4. Analysis were done using the program STATISTICA (8.0, Dell and StatSoft, CA, USA).

Conflicts of Interest:
The authors declare no conflict of interest.