*2.1. Preliminary Examinations*

Several liquid phase extraction techniques, such as DLLME, UA-DLLME in water, 10% NaCl, NADES, and 1% β-CD media, SA-LLE, and SULLE, were performed in order to select the procedure providing the better quali-quantitative multicomponent profile of phenolic compounds. The extractions were carried out as described in the experimental section. Figure 1 shows that the best results were achieved in the case of DLLME in 10% NaCl and 10% NADES media and UA-DLLME in 1% β-CD. In UA-DLLME, the phase separation was observed only with β-CD, whereas no phase separation was observed using the other additives. The notable increasing of extraction recovery using UA-DLLME with β-CD could be explained because β-CD was able to better dissolve the metabolites in the extraction solvent, contributing to an increased inclusion in its cavity of a higher amount of phenolic compounds. Therefore, DLLME in NaCl and NADES, UA-DLLME in β-CD media were selected for optimization.

**Figure 1.** Selection of microextraction procedure. # TCPC—Total concentration of phenolic compounds. Values expressed are means ± S.D. of three measurements. All the values were statistically significant (*p* < 0.001). Raw data regarding the statistical analyses were reported in *Supplementary Materials* section S1.

#### *2.2. Optimization of DLLME and UA-DLLME*

Several parameters that could influence the extraction efficiency, such as solid:liquid ratio, extraction and dispersive solvent types, and volume, were selected for optimization. For UA-DLLME, ultrasonication time was also optimized.

#### 2.2.1. Optimization of Extraction Medium Concentration

The extraction medium can significantly affect the extraction yields; therefore, a series of experiments were carried out by adding 5–15% NaCl or NADES solution, or 0.5–1.5% β-CD solution into the vessel containing 10 mg of the dry herbal material. For β-CD, the concentration was lower due to their low water solubility. With 10% NaCl, 15% NADES, and 1% β-CD, the best extraction recoveries were achieved (Figure 2a). Thus, these conditions were applied in further experiments.

**Figure 2.** Optimization of DLLME, UA-DLLME, and MAE. (**a**) Effect of medium concentration; (**b**) Effect of solid:liquid ratio; (**c**) Selection of extraction solvent; (**d**) Effect of extraction solvent volume; (**e**) Effect of dispersive solvent volume. # TCPC—Total concentration of phenolic compounds. Values expressed are means ± S.D. of three measurements. All the values were statistically significant (*p* < 0.001), unless otherwise indicated as n.s. (not statistically significant), \*\* (statistically significant at *p* < 0.01), or \* (statistically significant at *p* < 0.05). Raw data regarding the statistical analyses were reported in *Supplementary Materials* section S1.

#### 2.2.2. Optimization of Solid:Liquid Ratio

Three solid:liquid ratios, expressed as mg/mL (5:1.4, 10:1.4, 15:1.4 *w*:*v*), were examined for their impact on the extraction efficiency. The experimental results showed that the tendency for NaCl and NADES was similar, and the maximum of the extraction recovery was reached with the ratio 10:1.4. For β-CD, with the ratios 10:1.4 and 15:1.4 ( *w*:*v*), no significant differences were observed. Therefore, the optimal solid:liquid ratio was established as 10:1.4 ( *w*:*v*) (Figure 2b). In fact, in the analytical chemistry workflow, if two different systems show similar data, the lower ratio is generally used because it can ge<sup>t</sup> the same analytical performances using a lower amount of solvents, raw material, chemicals, etc.

#### 2.2.3. Selection of Extraction Solvent Type and Volume

*n*-Hexane, ethyl acetate, chloroform, and diethyl ether were tested as potential extractants. The experiments revealed that a higher amount of phenolic compounds was extracted using ethyl acetate (Figure 2c). This could be explained by the different polarities of the extraction solvents and by the interaction with polar phenolic compounds. For instance, with *n*-hexane, a nonpolar solvent, the phenolic compounds were poorly extracted. Diethyl ether and chloroform showed similar extraction efficiency with NaCl and NADES additives, whereas the addition of β-CD did not provide an exhaustive extraction. Taking into account the high volatility of diethyl ether, it was easier to work with ethyl acetate. Therefore, ethyl acetate was selected as appropriate solvent for all samples.

To determine the optimal volume of the extraction solvent, 200, 300, 400, and 500 μL were examined. When the volume is less than 300 μL, the phase separation was not achieved in DLLME and UA-DLLME, while phase separation was reached with 300 μL or more of NaCl and NADES. In order to apply this volume amount to other solvent media, the extraction procedure was modified as follows: the extraction solvent was added in two steps, firstly 200 μL were added in order to achieve an emulsion, then an additional 100 μL of ethyl acetate were rapidly injected. The phase separation was achieved after 5 min in the rest. Applying this procedure, no phase separation in UA-DLLME in β-CD media was observed; therefore, the UA-DLLME was carried out with 400 and 500 μL of ethyl acetate. It was found that with the increase of volume of ethyl acetate, the extraction of total content of phenolics decreased. Therefore, 400 μL was selected for further study on solid samples (Figure 2d).

#### 2.2.4. Selection of Dispersive Solvent Volume

Commonly, ethanol, methanol, and acetonitrile are reported as dispersive solvents in DLLME. In this study, ethanol was selected as dispersive solvent because some food supplements, not considered in this study, of *Galium* are in ethanolic solution. Therefore, the effect of its volume (100–500 μL) on the extraction yields was tested. The results showed that the extraction efficiency was enhanced with the increase of the ethanol volume in the solution until 300 μL, while for higher volumes, phase separation was not achieved (Figure 2e).

#### 2.2.5. Optimization of Ultrasonication Time in UA-DLLME

The cyclodextrins ( α, β, γ) show amphiphilic characteristics related to a hydrophilic shell and a hydrophobic cavity and could be usefully used as emulsifiers in order to enhance the extraction recovery for the target analytes. Their capacity to improve the extraction efficiency is related to their ability to reduce the interfacial tension between the two phases by an organic solvent/cyclodextrin complex located in the liquid–liquid interface. In this way, an increased contact area between the two phases was observed [30–33]. The aid of ultrasonication was generally required in order to enhance the solubility, as discussed by Saokham et al. [34] in a recent review paper. Different times have been investigated in the range of 2 to 10 min. Since 5 and 10 min showed similar responses, 5 min was selected as optimal in order to reduce the time of analysis.

#### *2.3. Reference Method: Microwave-Assisted Extraction*

In order to evaluate the performances of the proposed procedure, as in the comparison method, MAE was selected and carried out in the same media as LPME procedures at different concentration levels of the solvents (5–15% solution of NaCl and NADES, and 0.5–1.5% solution of β-CD). Figure 2a shows that the extraction efficiencies obtained in 10% NaCl and 15% NADES were comparable to DLLME and UA-DLLME. Therefore, the recovery of total phenolics, using LPME and MAE, was also comparable.

Following our experimental data, the optimized DLLME conditions found were: 10 mg of the sample, 10% NaCl, 15% NADES or 1% β-cyclodextrin, extraction solvent—400 μL of ethyl acetate, dispersive solvent—300 μL of ethanol, vortex time—30 s, extraction time—1 min, centrifugation at 12000× *g* for 5 min. In the case of UA-DLLME, 5 min of ultrasonication was required.

#### *2.4. Total Phenolic and Flavonoid Content, Antioxidant Capacity, and Tyrosinase Inhibitory Activity*

#### 2.4.1. Total Phenolic Content (TPC) by Spectrophotometric Assay

The Folin–Ciocâlteau assay was employed to determine the TPC of *Galium* extracts. The maximum TPC was registered in the ethanolic extract of *G. purpureum* (10.3 ± 0.8 mg GAE/g extract), whereas the lowest concentration was present in the ethanolic extract of *G. rivale* (1.3 ± 0.2 mg GAE/g extract). A recent study by Laki´c et al. showed similar results regarding the low phenolic content of *G. verum* (2.4–5.2 mg GAE/g extract), using different extraction solvents [7].

#### 2.4.2. Total Flavonoid Content (TFC) by Spectrophotometric Assay

Results of the total flavonoid content (TFC) of the different plant materials are presented in Table 1. The highest amount for the TFC was obtained for *G. verum* extract, with a value of 8.60 ± 0.07 mg QE/g d.w., comparable with the value obtained for *G. purpureum* extract, containing 8.50 ± 0.04 mg QE/g (d.w.). According to the results of the present study, Vlase et al. reported a TPC of 5.2 ± 0.2 g/100 g for a *G. verum* extract [12] and, additionally, Laki´c et al. reported values of 6.4–17.9 mg QE/g (d.w.), for *G.·verum* extracts, using different solvents and extraction times [7] confirming the results herein presented.

## 2.4.3. Antioxidant Potential Assays

The ferric reducing antioxidant power (FRAP), scavenging of DPPH, and ABTS free radical assays were used to evaluate the antioxidant capacity of *Galium* species (Table 1). These methods are simple and widely used for the evaluation of antioxidant capacity of herbal extracts/pure compounds. Moreover, the values regarding the total phenolic content (TPC) and total flavonoid content (TFC) are in accordance with antioxidant capacity values of the extracts. In the DPPH assay, *G. purpureum* (6.3 ± 0.7 mg TE/g extract) exhibited a higher DPPH scavenging capacity than any other considered species (0.4–1.9 mg TE/g extract). The ABTS value for *G. purpureum* (16.7 ± 0.8 mg TE/g extract) was higher in comparison with the values obtained for the other considered species, which ranged from 4.5 to 7.6 mg TE/g extract.

#### 2.4.4. Tyrosinase Inhibitory Activity

*Galium* extracts had good tyrosinase inhibitory activities (4.66–70.98% at 8 mg/mL), as reported in Table 1. The extract of *G. album* presented the highest tyrosinase inhibition, with a value of 70.98%. Despite the highest concentration of rutin and chlorogenic acid, the ethanolic extract of *G. rivale* showed no inhibitory effect against tyrosinase. This shows that the synergic effect of the compounds from the tested *Galium* samples have no or low inhibitory effects in some cases, although it was demonstrated by many studies that phenolic and flavonoid compounds are, in general, good inhibitors of tyrosinase [19]. The modest tyrosinase inhibitory activity for *Galium* species is confirmed by other studies as well. For example, Chiocchio et al. reported no tyrosinase inhibitory activity for *G. album* [35]. The low inhibitory activity of these extracts can be explained by the presence of other nondetected compounds, which might block or interfere with the enzyme.

#### *2.5. Quantitative Analysis of Galium Species*

The dry extracts were analyzed to establish the fingerprint of phenolic compounds in five *Galium* species. Table 2 summarizes the results obtained by means of a validated HPLC-PDA method for phenolics determination. All measurements were performed in triplicate in order to obtain standard deviation. It can be observed that the amount of the phenolic compounds for all *Galium* species is in the range from 2526.2–11345.1 μg g<sup>−</sup>1. The major biologically active compounds are chlorogenic acid and rutin. The highest number of the detected phenolic compounds was found in *G. rivale* (11345.1 μg g<sup>−</sup>1), where the main compound was chlorogenic acid (10192 ± 34 μg g<sup>−</sup>1), but the fingerprint was poorer in comparison with other species. The richest multicomponent pattern was observed in *G. pseudoaristatum*, but the quantity of phenolic compounds was the lowest (2526.2 μg g<sup>−</sup><sup>1</sup> ± 46.21). Chromatograms for each *Galium* sp. were reported in Supplementary Materials section S2.

*p*-OH benzoic acid, vanillic acid, epicatechin, syringic acid, 3-OH-4-MeO benzaldehyde, *p*-coumaric acid, *t*-ferulic acid, naringin, 2.3-diMeO benzoic acid, benzoic acid, *o*-coumaric acid, harpagoside, *t*-cinnamic acid, and naringenin were not reported into the table because they were not detected by the HPLC-PDA method.



Antioxidant Power; GAE = Gallic Acid Equivalents; QE = Quercetin Equivalents; TE = Trolox Equivalents; KAE = Kojic Acid Equivalents; KA = Kojic Acid. Raw data regarding the

statistical analyses were reported in *Supplementary Materials* section S1. Data marked with different letters indicates significant difference (*p* < 0.05).


*Molecules* **2019**, *24*, 1226

#### **3. Materials and Methods**
