Desirability Function Approach for the Optimization of Hydroalcoholic Solvent Extraction Conditions for Antioxidant Compounds from Olive Leaves

Abstract

The goal of this work was to obtain a rich olive leaf extract with high antioxidant activity due to its content of oleuropein (OLE), hydroxytyrosol (HT) and other phenolics, which have a synergistic effect on total antioxidant activity (TAA). The extraction parameters in the solvent extraction were investigated using response surface methodology (RSM) to produce the best results of HT and TAA in olive leaf extracts. A Box Wilson-CCD design was applied, the multi-objective optimization (MOO) was computed with Pareto solutions, and the desirability function (DF) was employed to define the best process variables. The optimized conditions (solvent concentration, temperature, time, solvent:solid ratio) were as follows: 63.30 %, 36 °C, 62 min, 11.80 mL/g for MeOH:H20 extracts and 43.80 %, 52 °C, 58 min, 9.40 mL/g for EtOH:H2 O extracts. Under these conditions, the highest results of HT were reached to 0.809±0.110 mg/g dw and 0.175±0.004 mg/g dw for MeOH:H2 O and EtOH:H2 O extracts, respectively. Similarly, the best results for TAA were obtained at higher concentrations in MeOH:H2 O extracts (451±2.32 mM Trolox) than in EtOH:H2O extracts (297±0.817 mM Trolox). Overall, the synergistic effect of OLE, HT, and flavonoids could make the olive leaf extract a potential cheap source for high value-added products.

Key words
Total antioxidant activity (TAA); Hydroxytyrosol (HT); Desirability function; optimization; method validation

INTRODUCTION

Olea europaea L. is an evergreen tree that is commonly considered to have strong antioxidant activity through its fruits, oil, and leaves (Jemai et al. 2009). Olive leaf (OLL) is a significant byproduct of both olive tree agriculture and the olive-processing industry that accumulates in vast quantities, producing economic and environmental problems (Herrero et al. 2011, Abaza et al. 2015, Clodoveo et al. 2022). Researchers are interested in the leaves because they are rich source of natural bioactive compounds with the potential to promote health. As a result, OLL have a great deal of potential for use in the food, medicine, and cosmetic industries (Quirantes-Piné et al. 2013, Rahmanian et al. 2015, Clodoveo et al. 2022).

Secoiridoids (oleuropein-OLE), simple phenolic compounds (hydroxytyrosol-HT and tyrosol), flavonoids (apigenin-7-O-glucoside, luteolin-7-O-glucoside, quercetin-7-O-rutinoside), and phenolic acids (verbascoside and caffeic acid) are the most active phenolics in OLL (Rahmanian et al. 2015, Talhaoui et al. 2015). These bioactive compounds have anti-cancer, anti-obesity, antioxidant, antiviral, anti-inflammatory, anti-diabetes, anti-microbial, anti-atherogenic properties, as well as cardioprotective, gastroprotective, and hepatoprotective properties (Lama-Muñoz et al. 2019, Taamalli et al. 2012, Hassen et al. 2015, Markovic et al. 2019).

OLE and HT are the most frequent phenolic compounds found in olive leaves. When the phenolic content of olive leaves was investigated, it was discovered that the major phenolic compound was OLE, followed by HT. OLE and HT content in olive leaf extract was determined to be 24.54% and 1.46% with antioxidant capabilities of 0.88±0.09 and 1.57 ± 0.12 mM (ABTS), respectively (Benavente-Garcia et al. 2000). HT attracts particular attention with its high antioxidant activity. It originates from the degradation of OLE and has higher antioxidant activity than OLE (Paradiso et al. 2016, Yao et al. 2019). The health advantages of HT have also been thoroughly investigated. It has been demonstrated to be an effective peroxyl radical scavenger. In numerous human systems, it crosses cell membranes and counteracts the reactive oxygen species’ harmful effects (D’Angelo et al. 2005). Moreover, it is known to have inhibiting proliferation, antimicrobial properties, anti-atherogenic capacity, cardioprotective, antidiabetic, lipid regulating, anti-obesity effects, retino-protective activity, and skin-related effects (D’Angelo et al. 2005, Granados-Principal et al. 2010, Haloui et al. 2011, Wani et al. 2018, Markovic et al. 2019, Gallardo-Fernández et al. 2022). In recent years, the antiviral, anti-inflammatory, and antithrombotic characteristics of HT have been investigated against COVID-19 disease (Takeda et al. 2021, Crudele et al. 2022, Abdelgawad et al. 2022), and it has been suggested that because of its various qualities, it might be a useful natural resource for the treatment of COVID-19 infection (Abdelgawad et al. 2022).

While the amount of HT in olive leaves is lower than in other olive processing wastes, it adds significantly to the total antioxidant efficacy of the leaves (Taamalli et al. 2012, Dias et al. 2019). It was discovered that HT in leaves has a higher antioxidant capacity than OLE and other phenolics (Benavente-Garcia et al. 2000, Erbay & Icier 2010). Additionally, Benavente-Garcia et al. (2000) found that olive leaf extract had a higher potential for antioxidant protection than either pure hydroxytyrosol or the vitamins C and E. In general, it has previously been shown that the flavonoids included in olive leaves may considerably add to the extracts’ antioxidant capacity. As a result, the importance and effect of the flavonoids present should not be underestimated (Goulas et al. 2010, Herrero et al. 2011). This is mostly due to the synergy of all phenolics found in olive leaves, including flavonoids, oleuropeosides, and substituted phenols (Benavente-Garcia et al. 2000).

HT is a remarkable bioactive compound that has numerous health benefits (Wani et al. 2018). Although the antioxidant potential of HT increases its value, extracting pure HT from natural sources is difficult. It is an easily oxidized chemical that is difficult and expensive to synthesize (Erbay & Icier 2010). As a result, it has been attempted to extract it from natural sources, such as olive leaves. HT can also be produced from OLE, but it takes additional procedures (Bouaziz & Sayadi 2005, Erbay & Icier 2010, Rahmanian et al. 2015). Because the properties of a bioactive individual component may differ in the presence of other compounds present in the extracts due to synergy between flavonoids, oleuropeosids, and phenols, olive leaf extract may be more beneficial than individual components (Benavente-Garcia et al. 2000, Şahin et al. 2015). The use of phenolic-rich olive leaf extracts as an alternative functional source to expensive purified biomolecules like oleuropein has the benefit of being a low-cost technique of processing since it avoids the requirement for the purification stage while being considerably effective. It has been stated that the most potent olive leaf products available on pharmacy shelves are made from leaf extracts rich in organic biophenols, which interact naturally to optimize the health advantages of the plant (Şahin & Bilgin 2018, Medina et al. 2019, Clodoveo et al. 2022). As a result, we tried to emphasize the significance of the extraction conditions of olive leaf extract, including phenolics with synergistic activities such as HT, OLE, and other active polyphenols in olive leaf. By effectively extracting phenolic substances from olive leaf, a low-cost raw material, high-value products can be obtained (Rahmanian et al. 2015). However, various factors, including olive cultivar, geographical origin, extraction circumstances, and solvent type, have a important effect on the phenolic content of olive leaves (Şahin & Bilgin 2018, Dias et al. 2019).

Extraction is a crucial step in the detection and quantification of these valuable phenolics in olive leaves. Although novel methods such as microwave-assisted extraction and superheated liquid extraction have been investigated (Rahmanian et al. 2015), in this study, an effective, easier and more economical extraction method that does not require special equipment was proposed in order to find the ideal conditions for the extraction of antioxidant phenolics from olive leaves. The solvent extraction (SE) technique, a traditional method, has been used extensively for the recovery of biotherapeutic substances (Rahmanian et al. 2015). Aqueous solvents of MeOH and EtOH have been employed for the SE of polyphenols from OLLs (Abaza et al. 2015, Tsakona et al. 2012). Due to their diverse solubilities and polarity, OLE, HT, flavonoids, and phenolic acids in olive leaves have been reported to be extracted using various solvents (Mohamed & Khan 2013, Herrero et al. 2011, Lee & Lee 2010, Quirantes-Piné et al. 2013). In this investigation, the extracts were made using aqueous MeOH and EtOH since the choice of solvent affects the concentration of phenolic components and, consequently, the antioxidant activity of olive leaf extracts. Studies on the effective recovery of phenols by identifying the ideal extraction conditions should be increased in considering the effective factors.

In order to get the best results for HT and TAA of MeOH aqueous and EtOH aqueous olive leaf extracts, the parameters of solvent concentration, solvent solid ratio, temperature, and time were examined. RSM, a multivariate statistical technique, was used to gather information on the relationship between the variables (Habibi et al. 2018, Zuorro et al. 2019, Zuorro 2020). The real problem in a process often has more than one quality feature. The desirability function (DF) approach is the most commonly employed approach for simultaneous quality feature improvement. The major challenge of in-process optimization is determining the optimal operating conditions that best represent a process’s multi-objective features. This is a MOO issue in RSM and an important field of research in experimental design (DOE) (Chen et al. 2012, Karande et al. 2013, Algan Cavuldak et al. 2019). Harrington (1965) developed the DF approach, which has been widely employed in industry to deal with MOO problems.

The goals of this work are as follows: a) To create phenolic-rich antioxidant extracts from olive leaves utilizing hydroalcoholic EtOH and MeOH as solvents, and to analyze the phenolic content of the OLL extracts using advanced characterisation methods. b) The goal of a MOO issue handled by the Pareto-optimal front and Harrington’s DF was to maximize HT and TAA at the same time. c) Then, these compounds -oleuropein as secoiridoids, tyrosol, and hydroxytyrosol as substituted phenols, luteolin, apigenin-7-glucoside, and luteolin-7-glucoside as flavones- were determined and quantified by HPLC. d) A simple and reliable method for separating and measuring six bioactive components was designed and validated. The ICH approach was used to build and validate in-house validation procedures.

MATERIALS AND METHODS

Material

Olea europaea L. was collected from an olive grove in Ayvalık-Turkey region at locations 39°16ʹ40.55K and 26°42ʹ47.77D. The sampling location was about 270 meters above sea level. Sampling was done in October 2019. Initial steps such as transportation, drying, reduction in size, and storage before analysis were done according to Vural et al. (2020).

Standards and chemicals

Analytical standards (Gallic acid (GA), 99%, Sigma); (hydroxytyrosol (HT), 98%, Sigma); (tyrosol (Ty), 99.5%, Fluka); (Luteolin-7-glucoside (L7G), 98%, Fluka); (oleuropein (OLE), 98.6%, Extrasynthese); (apigenin-7-glucoside (A7G), 97%, Fluka); (luteolin (Lut), 97%, Fluka) were HPLC grade. Chemicals (MeOH, 99.9%; EtOH, 99.5%; sodium carbonate (Na₂CO₃), 99.5%; Trolox, 97%; potassium persulfate, 99%; and acetic acid, 90.8-100.5%) were purchased from Sigma (USA). Deionized water was purified (18.2MΩ) with Destup (Ankara, Turkey). The TOC content of the ultrapure water was 3.27 µg/L.

Solvent extraction of bioactive compounds

Extraction was performed according to Vural et al. (2020) with different concentrations of MeOH/H₂O and EtOH/H₂O solvent systems using 1 g OLL sample (Table I). Hydroalcoholic solvent extraction treatment was performed with a glass reactor in a water bath, with a cooling system at the base and around the glass reactor jacket. Cooling the reactor jacket using a coolant (ethylene glycol-water) allowed for temperature control of the reaction.

Table I
Table I. Experimental values and coded levels of independent variables used in CCD/small factorial design in MeOH/H₂O and EtOH/H₂O solvent extraction systems.

Coded levels						Coded levels
Independent variables	-α	-1	0	+1	+α	Independent variables	- α	-1	0	+1	+α
X1: MeOH conc (% v/v)	12	25	45	65	79	X’1: EtOH conc (% v/v)	12	25	45	65	79
X2: Temperature (°C)	20	30	45	60	70	X’2: Temperature (°C)	20	30	45	60	70
X3: Time (min)	8	25	50	75	92	X’3: Time (min)	8	25	50	75	92
X4:Solvent:solid ratio (mL/g)	2	5	9.5	14	17	X’4: Solvent:solid ratio (mL/g)	2	5	9.5	14	17

• (α=±1.68; Non-center points:16; Center points:6).

A Box Wilson-CCD/small factorial DOE

A Box Wilson-CCD/small factorial DOE was made to determine the optimum extraction conditions and 22 experimental points were determined. To investigate the relationships between the constituents and observed outcomes and to improve the operating circumstances, a small/central composite design (CCD) with four elements, sixteen randomized trials, and six duplicates of the central point was utilized (Table I). In addition, the run sequence was randomized to reduce the influence of uncontrollable factors. Table IV(a,b) shows the coded and real components of the experimental design.

Table IV
Experimental design and observed responses for MeOH/H₂O and EtOH/H₂O hydroalcoholic solvent extraction.

	Std	x1	x2	x3	x4	Y 1	Y 2	Ty (mg/g dw)	L7G (mg/g dw)	OLE (mg/g dw)	A7G (mg/g dw)	Lut (mg/g dw)
		MeOH conc (%v/v)	Temp. ( °C)	Time (min)	Solvent: solid ratio (mL/g leaf)	HT (mg/g dw)	TAA (mM Trolox)
	1	65(+1)	60(+1)	75(+1)	5(-1)	0.083±0.001	472±1.99	0.81±0.080	0.005±0.001	12.56±0.31	0.21±0.025	0.08±0.009
	2	65(+1)	60(+1)	25(-1)	5(-1)	0.983±0.121	485±1.89	0.39±0.002	0.016±0.002	6.54±0.22	0.18±0.017	0.08±0.010
	3	65(+1)	30(-1)	75(+1)	14(+1)	1.069±0.125	490±1.75	0.12±0.010	0.024±0.005	4.08±0.18	0.05±0.006	0.04±0.003
	4	25(-1)	60(+1)	25(-1)	14(+1)	0.04±0.001	284±0.85	0.06±0.002	0.012±0.007	1.94±0.08	0.04±0.007	0.01±0.000
	5	65(+1)	30(-1)	25(-1)	14(+1)	0.098±0.002	411±1.01	0.05±0.002	0.012±0.008	3.96±0.11	0.04±0.005	0.01±0.000
	6	25(-1)	30(-1)	75(+1)	5(-1)	0.049±0.001	385±0.98	0.32±0.018	0.031±0.007	6.11±0.18	0.25±0.022	0.04±0.002
	7	25(-1)	60(+1)	75(+1)	14(+1)	0.09±0.002	345±0.87	0.10±0.011	0.001±0.000	2.33±0.09	0.10±0.011	0.04±0.002
	8	25(-1)	30(-1)	25(-1)	5(-1)	0.035±0.001	321±0.76	1.48±0.115	0.093±0.008	4.49±0.16	0.13±0.012	0.01±0.001
	9	12(-1.68)	45(0)	50(0)	9.50 (0)	0.001±0.001	210±0.65	0.14±0.011	0.015±0.002	1.22±0.05	0.04±0.006	0 (nd)
	10	79(1.68)	45(0)	50(0)	9.50 (0)	1.145±0.121	498±0.98	0.21±0.012	0.006±0.001	5.55±0.12	0.08±0.007	0.03±0.002
	11	45(0)	20(-1.67)	50(0)	9.50 (0)	0.11±0.003	397±0.75	0.06±0.003	0.015±0.002	2.15±0.08	0.05±0.004	0 (nd)
	12	45(0)	70(+1.67)	50(0)	9.50 (0)	0.29±0.0025	405±0.98	0.24±0.012	0.003±0.000	4.16±0.14	0.11±0.012	0.06±0.008
	13	45(0)	45(0)	8(-1.68)	9.50 (0)	0.005±0.001	195±0.57	0.001±0.00	0 (nd)	1.02±0.08	0.004±0.000	0 (nd)
	14	45(0)	45(0)	92(+1.68)	9.50 (0)	0.297±0.020	398±0.75	0.15±0.012	0.01±0.003	5.47±0.21	0.16±0.014	0.07±0.005
	15	45(0)	45(0)	50(0)	2(-1.67)	0.074±0.003	359±0.77	0.11±0.008	0.004±0.000	3.51±0.18	0.08±0.005	0.02±0.001
	16	45(0)	45(0)	50(0)	17.00(+1.67)	0.223±0.019	309±0.71	0.05±0.001	0.005±0.000	2.35±0.11	0.07±0.006	0.01±0.000
	17-22	45(0)	45(0)	50(0)	9.50 (0)	0.175±0.012	345±0.74	0.16±0.02	0.013±0.010	3.31±0.14	0.11±0.021	0.04±0.010
Desirability	Opt(Obs)	63.30	36.00	62.00	11.80	0.809±0.110	451±2.32	0.09±0.003	0.018±0.006	9.25±0.23	0.11±0.017	0.05±0.009
	Opt(Pred)	63.30	36.00	62.00	11.80	0.721	442
	Std	x’1	x’2	x’3	x’4	Y’ 1	Y’ 2	Ty (mg/g dw)	L7G (mg/g dw)	OLE (mg/g dw)	A7G (mg/g dw)	Lut (mg/g dw)
		EtOH conc (%v/v)	Temp.( °C)	Time (min)	Solvent:solid ratio (mL/g leaf)	HT (mg/g dw)	TAA (mM Trolox)
	1	65(+1)	60(+1)	75(+1)	5(-1)	0.01±0.000	293±1.11	1.42±0.02	0.68±0.05	7.90±0.11	0.35±0.02	0.11±0.030
	2	65(+1)	60(+1)	25(-1)	5(-1)	0.02±0.001	281±0.98	1.00±0.01	0.21±0.02	4.54±0.07	0.18±0.01	0.091±0.010
	3	65(+1)	30(-1)	75(+1)	14(+1)	0.098±0.009	272±0.77	2.89±0.06	0.419±0.04	2.64±0.03	0.047±0.00	0.032±0.002
	4	25(-1)	60(+1)	25(-1)	14(+1)	0.01±0.001	129±0.23	0.07±0.01	0.11±0.01	0.90±0.01	0.05±0.00	0.01±0.001
	5	65(+1)	30(-1)	25(-1)	14(+1)	0.07±0.002	222±0.39	1.996±0.03	0.321±0.02	1.30±0.02	0.029±0.00	0.02±0.002
	6	25(-1)	30(-1)	75(+1)	5(-1)	0.03±0.000	121±0.21	0.15±0.02	0.137±0.01	0.70±0.01	0.073±0.01	0.03±0.004
	7	25(-1)	60(+1)	75(+1)	14(+1)	0.19±0.010	178±0.18	0.02±0.00	0.199±0.02	2.60±0.03	0.088±0.01	0.023±0.003
	8	25(-1)	30(-1)	25(-1)	5(-1)	0.06±0.002	156±0.17	0.057±0.01	0.115±0.01	0.98±0.01	0.11±0.01	0.028±0.003
	9	12(-1.68)	45(0)	50(0)	9.50 (0)	0.05±0.003	130±0.11	0.001±0.00	0.005±0.00	1.18±0.02	0.07±0.00	0.006±0.00
	10	79(1.68)	45(0)	50(0)	9.50 (0)	0.021±0.001	292±0.41	1.01±0.04	0.385±0.02	3.92±0.05	0.067±0.00	0.04±0.001
	11	45(0)	20(-1.68)	50(0)	9.50 (0)	0.05±0.002	180±0.12	1.00±0.03	0.311±0.03	1.75±0.02	0.05±0.00	0.011±0.00
	12	45(0)	70(+1.68)	50(0)	9.50 (0)	0.25±0.020	298±0.38	1.24±0.04	0.452±0.05	3.79±0.06	0.12±0.01	0.06±0.002
	13	45(0)	45(0)	8(-1.68)	9.50 (0)	0.001±0.000	198±0.13	1.01±0.02	0.332±0.04	0.80±0.01	0.003±0.00	0.002±0.00
	14	45(0)	45(0)	92(+1.68)	9.50 (0)	0.05±0.002	235±0.14	1.11±0.02	0.348±0.04	5.20±0.08	0.09±0.01	0.049±0.001
	15	45(0)	45(0)	50(0)	2(-1.68)	0.14±0.015	288±0.13	0.66±0.01	0.033±0.004	3.40±0.05	0.06±0.01	0.025±0.002
	16	45(0)	45(0)	50(0)	17.00(+1.68)	0.04±0.002	225±0.12	1.01±0.05	0.335±0.05	1.90±0.02	0.044±0.00	0.019±0.001
	17-22	45(0)	45(0)	50(0)	9.50 (0)	0.14±0.01	285±1.05	1.25±0.05	0.442±0.08	3.10±0.05	0.10±0.01	0.041±0.005
Desirability	Opt(Obs)	43.80	52.00	58.00	9.40	0.175±0.004	297±0.817	1.28±0.04	0.475±0.06	4.89±0.07	0.11±0.01	0.07±0.006
	Opt(Pred)	43.80	52.00	58.00	9.40	0.172	298

• Analytical results are the average of triplicates (mean ± sd).

TAA (mM Trolox) and the amount of HT (mg/g dw) were determined as the dependent variables (Table I). The ranges for the independent variables (the different ratios of hydroalcoholic solvent mixtures, temperature, time and solvent:solid ratio) were found as 12-79% (v/v), 20-70°C, 8-92 min, 2-17 (mL/g leaf), taking into account the values found in the literature and early research (Vural et al. 2020). Stat-Ease Design-Expert 10.0 (Minneapolis, USA) was utilized for chemometric techniques in this work.

Desirability function (DF) approach

Finding an input variable set for which all output variables (responses) fall to the desired values or are as near to them as feasible is the goal of the DF optimization strategy. The DF is used to convert each of the m predicted responses ​​​ˆ y​​ 1​​​, ​​​ˆ y​​ 2​​​, …, ​​​ˆ y​​ m​​​ from the m potentially various models to an individual DF di, where 0 ≤ di ≤ 1, for a given set of factor values (Harrington 1965).

The DFs are defined by the answers to be optimized, which are divided into three categories: (a) the nominal-the-best (NTB) type response, (b) the larger-the-better (LTB) type response, and (c) the smaller-the-better (STB) type response (Karande et al. 2013, Harrington 1965, Wu 2004).

The DF from Eq. (1) is written as follows:

for the LTB-type response,

$d_i=\left\{\begin{array}{c}\begin{array}{c}0for{\hat{\mathrm{y}}}_{\mathrm{i}}\le {\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}},\\ {\left(\frac{{\hat{\mathrm{y}}}_{\mathrm{i}}-{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}}{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)}^{r}for{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}<{\hat{\mathrm{y}}}_{\mathrm{i}}<{\mathrm{y}}_{\mathrm{i}}^{max},\\ \end{array}\\ 1for{\hat{\mathrm{y}}}_{\mathrm{i}}\ge {\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}},\end{array}\right.$ (1)

for the STB-type response,

$d_i=\left\{\begin{array}{c}\begin{array}{c}1for{\hat{\mathrm{y}}}_{\mathrm{i}}\le {\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}},\\ {\left(\frac{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-{\hat{\mathrm{y}}}_{\mathrm{i}}}{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)}^{r}for{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}<{\hat{\mathrm{y}}}_{\mathrm{i}}<{\mathrm{y}}_{\mathrm{i}}^{max},\\ \end{array}\\ 0for{\hat{\mathrm{y}}}_{\mathrm{i}}\ge {\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}},\end{array}\right.$ (2)

for the NTB-type response,

$d_i=\left\{\begin{array}{c}\begin{array}{c}{\left(\frac{{\hat{\mathrm{y}}}_{\mathrm{i}}-{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}}{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)}^{r1}for{\mathrm{y}}_{\mathrm{i}}^{min}\le {\hat{\mathrm{y}}}_{\mathrm{i}}\le T_i,\\ {\left(\frac{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-{\hat{\mathrm{y}}}_{\mathrm{i}}}{{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}-T_i}\right)}^{r2}for{T}_i<{\hat{\mathrm{y}}}_{\mathrm{i}}<{\mathrm{y}}_{\mathrm{i}}^{max},\\ \end{array}\\ 0for{\hat{\mathrm{y}}}_{\mathrm{i}}<{\mathrm{y}}_{\mathrm{i}}^{min}or{\hat{\mathrm{y}}}_{\mathrm{i}}>{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}},\end{array}\right.$ (3)

where$y_i^{min}$ and $y_i^{max}$ are the lower and the upper bound on the on the i ^th response, respectively, r1 > 0 and r2 > 0 are the two shape parameters, and ​​$T_i$ is the target value of the i ^th response which is a NTB-type one.

Different values of r, r1, and r2 will result in different desirability shapes, and small values of r1 and r2 should be chosen. Both shape parameters in such DFs can be set to different values, and the range type conversion is frequently selected since it is easy. The DF can be given as follows:

$d_i=\left\{\begin{array}{c}1for{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}<{\hat{\mathrm{y}}}_{\mathrm{i}}<{\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}\\ 0for{\hat{\mathrm{y}}}_{\mathrm{i}}\le {\mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{i}\mathrm{n}}{\mathrm{a}\mathrm{n}\mathrm{d}{\hat{\mathrm{y}}}_{\mathrm{i}}\ge \mathrm{y}}_{\mathrm{i}}^{\mathrm{m}\mathrm{a}\mathrm{x}}\end{array}\right.$ (4)

All individual DFs are then integrated into a general DF after choosing a workable individual DF for each predicted response (Harrington 1965) using the geometric mean as in Eq (5):

$$DF=(d_1{d}_2···d_m{)}^{1/m}$$ (5)

Total antioxidant activity (TAA) assay

TAA was determined for the extracts using the 2,2’-azino-bis (3-ethyl benzothiazoline-6-sulfonic acid) (ABTS) technique reported by Re et al. (1999), with minor modifications. The method was preferred due to being used for aqueous and lipophilic systems. 2.45 mM potassium persulfate was mixed with 7 mM ABTS solution to make the ABTS+ radical solution. After that, the mixture was maintained at room temperature in the dark for 12 to 24 hours. It was diluted with phosphate buffer (PBS) before use to attain an absorbance of 0.700±0.02 at 734 nm. 50 µL of each sample was mixed into 950 µL of ABTS+ solution. At 0 and 6 minutes of reaction time, the absorbance was measured. Each sample was examined three times. Standard Trolox solutions (0-700 M) were utilized to generate a calibration curve. The results are given in mM Trolox equivalent g^-1 of dw.

Analysis of bioactive compounds

The Shimadzu HPLC apparatus and LC Solution software were employed. In the analysis, a PDA detector was used in the wavelength range 190-550 nm, and the quantitative measurements were made at 280 nm wavelength. HPLC analyses of HT, Ty, L7G, OLE, A7G and Lut in OLL extracts were performed. It was used by altering the gradient technique utilized by Vural et al. (2020). Two solvents (A: H₂O +%1 Acetic Acid, B: MeOH) were used in the gradient procedure. The column oven’s temperature was programmed at 30°C. The flow rate was 1.0 mL/min, and the following elution procedure was used: 0-10 minutes from 85 to 60% A; 10-15 minutes from 60 to 30% A; and 15-30 minutes from 30 to 85% eluent A. The elute was analyzed for bioactive chemicals at 280 nm using an Inertsil ODS-4 (250x4mm; 5m) column and a 25 µL injection volume.

Analytical method validation

Stock solutions of the standards (HT, Ty, L7G, OLE, A7G and Lut) were freshly prepared (200 mg L^-1) by dissolving 6 authentic compounds in MeOH:H₂O (80:20) except for OLE. OLE stock solution was prepared within the range of 2000 mg L^-1. The concentration range of the standard solutions was chosen based on the expected analyte levels in the samples (Table II). The calibration curve was constructed by graphing the standard concentration (mg mL^-1) vs the peak area for each standard. Standard solutions were then filtered using a 0.22 μm syringe membrane filter, injected in triplicate in quantities of 25 μL, and evaluated under chromatographic conditions. For each identified analyte, the limits of detection (LOD) and quantification (LOQ) were computed.

Table II
Assay validation parameters of the proposed HPLC method for determination of bioactive compounds.

Parameter	HT	Ty	L7G	OLE	A7G	Lut
Accuracya (mean recovery % ±SD)	97.32±0.51	98.85±1.02	97.98±0.88	95.85±1.45	98.33±0.79	98.43±0.92
Precision
Repeatabilityb	±1.01	±0.70	±0.55	±1.70	±0.70	±0.35
Intermediate precisionc	±2.00	±2.25	±1.99	±4.25	±2.25	±1.80
Linearity
Slope	191262	258476	3.107	65974	295601	634194
Intercept	+123121	+114873	-87408	+1.106	196496	300382
Correlation cofficient (r)	0.9979	0.9968	0.9993	0.9952	0.9983	0.9921
Range (mg L−1)	0.5-100	0.01-20	0.01-20	2.5-1750	0.25-50	0.010-20
LODd (mg L−1)	0.012	0.005	0.008	0.4.10-3	0.011	0.6.10-3
LOQd (mg L−1)	0.042	0.015	0.026	1.7.10-3	0.037	2.0.10-3

• ^aThe accuracy average of (n = 3). Analytical results are the average of triplicates (mean ± sd).
  ^bThe intraday (n = 3), an average of three concentrations (0.5, 5 and 10 mg L−¹) Ty, L7G, Lut and (5, 10 and 25 mg L−¹) HT, A7G and (50, 500 and 1000 mg L−¹) OLE, for compounds repeated three times within the day.
  ^cThe inter-day (n = 3), an average of three concentrations (0.5, 5 and 10 mg L−¹) Ty, L7G, Lut and (5, 10 and 25 mg L−¹) HT, A7G and (50, 500 and 1000 mg L−¹) OLE, for compounds repeated three times in 3 days.
  ^dDetermined via calculations, LOD = 3.3 (SD of the response/slope), LOQ = 10 (SD of the response/slope).

Six standards were identified and quantified by comparing retention times and peak areas to those of the standard and by co-injection with the sample (spike test), respectively. The standard solutions were applied to the samples in three different quantities, and recovery was evaluated in triplicate. Accuracy was evaluated with the mean percentage recoveries method. For comparison, an unspiked sample was concurrently prepared and analyzed. Repeatability was assessed through the relative standard deviation (RSD) values. Precision was evaluated by the performance of intraday (repeatability) by three replicated injections of the same solution, same analyst within the same day and inter-day (reproducibility) was determined by analyzing the same solution on three different days (three injections a day). Satisfactory results were obtained for the used method. The validation was performed according to International Conference on Harmonization (ICH) guidelines (ICH 2005), and the calculated validation parameters were shown in Table II.

RESULTS AND DISCUSSION

Fitting the model

In this study, the DOE was employed in which the independent variables were chosen as solvent concentration (SolC), temperature (Temp), time (T), solvent:solid ratio (SSR) (X₁, X₂, X₃, X₄, X’₁, X’₂, X’₃, X’₄) for two different hydroalcoholic solvent systems and the response variables were HT and TAA (Y₁, Y₂, Y’₁, Y’₂). MeOH and EtOH mixtures with water were used to optimize the extraction terms to have the maximum yield of HT and TAA of bioactives in the OLL. After the determination of optimum extraction points, the prominent bioactives in OLL -Ty, L7G, OLE, A7G and Lut- were measured.

Table III(a) demonstrates the ANOVA results for HT and TAA values acquired by MeOH:H₂O extraction. Both models were highly significant (p <0.0001) and had high F values of 107.22 and 16.26 for HT and TAA, respectively. The R² of the predicted models were 0.9877 and 0.7928, respectively. The models for (MeOH:H₂O) SE were adequate to describe the experimental results and were described by the following equations:

$$Y_{1HT}=+0.17+0.34X_1+0.054X_2+0.046X_3+0.029X_4+0.061X_1{X}_4-0.23X_2{X}_3+0.084X_2{X}_4+0.24X_3{X}_4+0.14X_1^2$$ (6)

$$Y_{2TAA}=+345.81+73.48X_1-0.57X_2+38.99X_3+31.28X_2^2$$ (7)

Table IIIa
ANOVA results and second order polynomial equation for HT and TAA for MeOH/H₂O solvent system (Backward Elimination Regression).

	Y1:HT				Y2: TAA
	Hierarchial Terms added after Backward Elimination Regression: X 4 , X 1 2 , X 2 2 , X 3 2 , X 5 2 , X 3 X 5 , X 2 X 4 , X 4 X 5 , X 1 X 3 , X 3 X 4 , X 2 X 5 , X 1 X 4				Hierarchial Terms added after Backward Elimination Regression: X 4 , X 1 2 , X 2 2 , X 3 2 , X 5 2 , X 3 X 5 , X 2 X 3 , X 1 X 3 , X 2 X 4 ,X 1 X 4 ,X 4 X 5 ,X 3 X 4 ,X 2 X 5
	SSE	df	F value	p value	SSE	df	F value	p value
Model	2.41	9	107.22	< 0.0001	1.099 105	4	16.26	< 0.0001
X1 (MeOH con.)	0.65	1	262.11	< 0.0001	74409.20	1	44.04	< 0.0001
X2 (Temperature)	0.016	1	6.49	0.0256	4.34	1	2.566 10-3	0.9602
X 3 (Time)	0.029	1	11.49	0.0054	20745.38	1	12.28	0.0027
X4 (solvent:solid ratio)	0.012	1	4.62	0.0527
X1X4	0.012	1	4.87	0.0476
X2X3	0.42	1	168.59	< 0.0001
X2X4	0.024	1	9.51	0.0095
X3X4	0.45	1	182.08	< 0.0001
X1 2	0.31	1	125.39	< 0.0001
X2 2					14759.24	1	8.74	0.0089
Residual	0.030	12			28721.66	17
Lack of Fit	0.022	7	1.96	0.2386	22948.33	12		0.3013
Pure Error	8.009.10-3	5			5773.33	5
Cor Total	2.44	21			1.386 105	21
R2	0.9877				0.7928
R2 Adj	0.9785				0.7441
R2 Pred	0.9467				0.5957
PRESS	0.13				56047.29
Significant Model Terms	X 1 , X 2 , X 3 , X 1 X 4 , X 2 X 3 , X 2 X 4 , X 3 X 4 , X 1 2				X 1 , X 3 , X 2 2
Second order polynomial equation by CCD-Optimal Responce Surface-after Backward Elimination	Y1HT =+0.17+0.34X1+0.054X2 +0.046X3 +0.029X4+0.061X1X4 -0.23X2X3 +0.084X2X4+0.24X3X4 +0.14X1 2				Y2TAA=+345.81+73.48X1-0.57X2 +38.99X3 +31.28X2 2

Table IIIb ANOVA results and second order polynomial equation for HT and TAA for EtOH/H₂O solvent system (Backward Elimination Regression).

	Y’1:HT				Y’2: TAA
	Hierarchial Terms added after Backward Elimination Regression: X’ 2 X’ 4				Hierarchial Terms added after Backward Elimination Regression: ----
	SSE	df	F value	p value	SSE	df	F value	p value
Model	0.090	13	88.26	< 0.0001	78865.70	14	1729.93	< 0.0001
X ‘1 (EtOH con.)	1.45 10-3	1	18.43	0.0036	13122.00	1	4029.65	< 0.0001
X’ 2 (Temperature)	0.020	1	254.37	< 0.0001	6962.00	1	2137.97	< 0.0001
X’ 3 (Time)	4.59 10-3	1	58.41	0.0001	1398.93	1	429.60	< 0.0001
X’ 4 (solvent:solid ratio)	5.00 10-3	1	63.59	< 0.0001	1984.50	1	609.42	< 0.0001
X’ 1 X’ 2	0.012	1	155.17	< 0.0001	524.66	1	161.12	< 0.0001
X’ 1 X’ 3	2.18 10-3	1	27.70	0.0012	288.00	1	88.44	< 0.0001
X’ 1 X’ 4	0.013	1	168.15	< 0.0001	1536.80	1	471.94	< 0.0001
X’ 2 X’ 3	3.70 10-3	1	47.03	0.0002	264.50	1	81.23	<0.0001
X’ 2 X’ 4					554.34	1	170.23	< 0.0001
X’ 3 X’ 4	7.69 10-3	1	97.78	< 0.0001	1860.50	1	571.34	< 0.0001
X’ 1 2	0.016	1	208.15	< 0.0001	10495.98	1	3223.22	< 0.0001
X’2 2	8.06 10-4	1	10.25	0.0150	4213.67	1	1293.98	< 0.0001
X’ 3 2	0.020	1	256.76	< 0.0001	9110.50	1	2797.75	< 0.0001
X’ 4 2	2.91 10-3	1	37.02	0.0005	1672.56	1	513.63	< 0.0001
Residual	5.50 10-4	8			19.54	7
Lack of Fit	4.30 10-4	3	4.78	0.0824	14.74	2	6.14	0.0604
Pure Error	1.20 10-4	5			4.80	5
Cor Total	0.091	21			78885.24	21
R2	0.9939				0.9998
R2 Adj	0.9827				0.9992
R2 Pred	0.7540				0.9839
PRESS	0.022				1272.51
Significant Model Terms	X’ 1 ,X’ 2 , X’ 3 , X’ 4 , X’ 1 X’ 2 , X’ 1 X’ 3 , X’ 1 X’ 4 , X’ 2 X’ 3 , X’ 3 X’ 4 , X’ 1 2 , X’ 2 2 , X’ 3 2 , X’ 4 2				X’ 1 ,X’ 2 , X’ 3 , X’ 4 , X’ 1 X’ 2 , X’ 1 X’ 3 , X’ 1 X’ 4 , X’ 2 X’ 3 , X’ 2 X’ 4 ,X’ 3 X’ 4 , X’ 1 2 , X’ 2 2 , X’ 3 2 , X’ 4 2
Second order polynomial equation by CCD-Optimal Responce Surface-after Backward Elimination	Y’ 1 HT = +0.13- 0.01X’ 1 + 0.06X’ 2 +0.018X’ 3 -0.03X’ 4 -0.06X’ 1 X’ 2 – 0.017X’ 1 X’ 3 + 0.063X’ 1 X’ 4 + 0.021X’ 2 X’ 3 +0.031X’ 3 X’ 4 – 0.033X’ 1 2 + 7.45 10 -3 X’ 2 2 -0.037X’ 3 2 -0.014X’ 4 2				Y’ 2 TAA =+285.55 + 47.65X’ 1 + 35.40X’ 2 + 10.13X’ 3 -18.90X’ 4 -12.65X’ 1 X’ 2 +6.00X’ 1 X’ 3 + 21.65X’ 1 X’ 4 +5.75X’ 2 X’ 3 -12.85X 2 X’ 4 + 15.25X’ 3 X’ 4 – 26.05X’ 1 2 – 17.03X’ 2 2 -24.737X’ 3 2 -10.73X’ 4 2

ANOVA results for HT and TAA amounts obtained by (EtOH:H₂O) SE can be seen in Table III(b) as well. According to the ANOVA results of the models for the HT and TAA, the model performance was good, with correlation coefficients of 0.9939 and 0.9998, respectively. The models were demonstrated as significant (p<0.0001) and high F-values of 88.26 and 1729.93, respectively. The models for (EtOH:H₂O) SE were illustrated by the subsequent equations:

$${Y^{\prime }}_{1HT}=+0.13-0.01{X^{\prime }}_1+0.06{X^{\prime }}_2+0.018{X^{\prime }}_3-0.03{X^{\prime }}_4-0.06{X^{\prime }}_1{X^{\prime }}_2–0.017{X^{\prime }}_1{X^{\prime }}_3+0.063{X^{\prime }}_1{X^{\prime }}_4+0.021{X^{\prime }}_2{X^{\prime }}_3+0.031{X^{\prime }}_3{X^{\prime }}_4–0.033{X^{\prime }}_1^2+7.45{10}^{-}3{X^{\prime }}_2^2-0.037{X^{\prime }}_3^2-0.014{X^{\prime }}_4^2$$ (8)

$${Y^{\prime }}_{2TAA}=+285.55+47.65{X^{\prime }}_1+35.40{X^{\prime }}_2+10.13{X^{\prime }}_3-18.90{X^{\prime }}_4-12.65{X^{\prime }}_1{X^{\prime }}_2+6.00{X^{\prime }}_1{X^{\prime }}_3+21.65{X^{\prime }}_1{X^{\prime }}_4+5.75{X^{\prime }}_2{X^{\prime }}_3-12.85X_2{X^{\prime }}_4+15.25{X^{\prime }}_3{X^{\prime }}_4–26.05{X^{\prime }}_1^2–17.03{X^{\prime }}_2^2-24.737{X^{\prime }}_3^2-10.73{X^{\prime }}_4^2$$ (9)

Analysis of response surfaces

The extraction variables; SolC, Temp, T, and SSR were used to find optimum extraction conditions in SE. Generally, the linear terms of MeOH concentration (X₁) showed the largest effect (p <0.0001) for HT and TAA in the MeOH solvent system. For HT, MeOH concentration (X₁) was followed by time (X₃) and Temp (X₂), whereas it was followed by time (X₃) for TAA. For the EtOH solvent system, the linear terms of all the variables (X’₁, X’₂, X’₃, X’₄) had the greatest influence on TAA, however, extraction Temp (X’₂) and SSR (X’₄) showed the largest effects (p<0.0001), followed by time (X’₃) (p=0.0001), and EtOH concentration (X’₁) (p=0.0036) for HT. Furthermore, 3D response surface plots of HT and TAA, including interaction effects of independent variables, are presented in Figs. 1 and 2 for the MeOH solvent system, and Figs. 3 and 4 for the EtOH solvent system, respectively.

Figure 1
Response surface plots of HT of olive leaves extracts in MeOH/H₂O hydroalcoholic solvent extraction as affected by (a) temperature and solvent:solid ratio (b) MeOH conc. and solvent:solid ratio (c) temperature and time (d) time and solvent:solid ratio.

Figure 2
Response surface plots of TAA of olive leaves extracts in MeOH/H₂O hydroalcoholic solvent extraction as affected by MeOH conc. and temperature.

Figure 3
Response surface plots of HT of olive leaves extracts in EtOH/H₂O hydroalcoholic solvent extraction as affected by (a) EtOH conc. and temperature (b) EtOH conc. and time (c) EtOH conc. and solvent:solid ratio (d) temperature and time (e) temperature and solvent:solid ratio (f) time and solvent:solid ratio.

Figure 4
Response surface plots of TAA of olive leaves extracts in EtOH/H₂O hydroalcoholic solvent extraction as affected by (a) EtOH conc. and temperature (b) EtOH conc. and time (c) EtOH conc. and solvent:solid ratio (d) temperature and time (e) temperature and solvent:solid ratio (f) time and solvent:solid ratio.

Solvent (MeOH:H₂O and EtOH:H₂O) concentration (SolC)

SE is a traditional procedure that has been used for many years. The solvent type is one of the most critical factors impacting the performance of traditional SE of phenolics from olive by-products (Nakilcioğlu-Taş & Ötleş 2019). All compounds cannot be removed by a single solvent because of their various solubilities and polarities (Mohamed & Khan 2013, Sifaoui et al. 2016). Plants contain a variety of phenolic chemicals with varying chemical characteristics and polarity. Choosing the solvent is very important, as it affects the amount and type of phenols extracted (Nakilcioğlu-Taş & Ötleş 2019, Wissam et al. 2016, Chan et al. 2009). Additionally, the difference between the type and composition of bioactive compounds affects the TAA of the plant. The solubility and extraction efficiency of bioactive polyphenols may be affected by differences in solvent polarity (Nakilcioğlu-Taş & Ötleş 2019). This might be owing to the polarity of the phytochemicals in OLLs, which allows them to be extracted using solvents of varying polarities (Wissam et al. 2016).

Depending on the target analytes and sample matrix, several ratios of MeOH:H₂O and EtOH:H₂O were used for phenolic compound extraction (Pérez-Serradilla et al. 2007, Habibi et al. 2018). Alcohols have intermediate polarity rather than more or less polar solvents (Wissam et al. 2016, Galanakis et al. 2013). In this study, a range of 25–65% MeOH:H₂O mixtures was used. It was shown that the change in hydroalcoholic composition influenced both HT and TAA significantly (p<0.0001) (Table IIIa). Fig. 1(b) shows the effect of MeOH:H₂O concentration with SSR on HT, which is shown (p=0.0476) in Table III(a) also. However there was no significant relationship between Temp and MeOH:H₂O concentration for HT and TAA in the ANOVA results which are shown in Fig. 2(a). The quadratic term of MeOH:H₂O concentration (X₁ ²) was significant for HT (p<0.0001).

Table III(b) shows the ANOVA results of optimization conditions of EtOH:H₂O (25-65%) solvent system for HT and TAA. Similar to MeOH results, HT was affected by changes in EtOH:H₂O concentration (p=0.0036) and TAA (p<0.0001) (Table IIIb). For both HT and TAA, the binary relationship of EtOH:H₂O concentration with Temp (p<0.0001, p<0.0001), with T (p=0.0012, p<0.0001) and with SSR (p<0.0001, p<0.0001) were significant which can be seen in Figs. 3(a), (b), (c) and Figs. 4(a), (b), (c) respectively. Quadratic term of EtOH:H₂O concentration (X’₁ ²) was significant for both HT and TAA (p<0.0001)

It was determined that the hydroalcoholic solvent had a positive linear effect on both responses of HT and TAA. Sifaoui et al. (2016) showed that MeOH was an acceptable solvent for extracting bioactive compounds from OLLs with highest extraction yield whereas EtOH provided lower concentrations of phenolics. Similarly, Nakilcioğlu-Taş & Ötleş (2019) specified that obtaining high yields of TAA and polyphenols including HT was easier in MeOH, implying that these compounds were more polar than EtOH. Contrary to these findings, using EtOH:H₂O mixtures as an inexpensive and non-toxic extraction solvent, the optimal ratio for the simultaneous SE of OLE and HT in olive pomace was determined to be 60% EtOH:40% H₂O (Habibi et al. 2018). In a similar study, compared to MEOH, EtOH was chosen as the ideal solvent for the SE from OLLs to acquire extracts with high TPC and TAA due to being a food-grade solvent and being classified as generally recognized as safe (GRAS) (Wissam et al. 2016). In most investigations, a 40% EtOH:60% H₂O solution was determined to get a high phenolic yield (Wissam et al. 2016, Şahin et al. 2015, Thoo et al. 2010) in which they observed a drop as EtOH concentration in the solvent increased. Studies on the SE of bioactive compounds derived from OLLs have indicated an EtOH:H₂O ratio of 40-80% v/v (Mylonaki et al. 2008, Japón-Luján et al. 2006). Stamatopoulos et al. (2014) demonstrated the effect of EtOH:H₂O on TPC, with the highest value reached at a concentration of 70% EtOH:H₂O. SolC showed also a significant influence (p<0.05) on TAA (Irakli et al. 2018) similar to our results. The greatest TAA values were attained at concentrations of 50% and 70% concentration, however, the yield of HT was maximal at the lowest organic solvent concentration and declined dramatically as the solvent concentration grew, which may explain why HT has a more polar structure. In several studies on HT and TAA, the maximal HT and TAA were obtained by H₂O, 80% MeOH, and 44% EtOH using ultrasound-assisted extraction (UAE), respectively; however, for TAA (DPPH), the opposite was true (Yao et al. 2019).

Temperature (Temp)

Temp ranges of 30-60 °C were chosen for both MeOH and EtOH solvent systems in this research. In MeOH solvent system, the linear term of Temp was significant (p=0.0256) for HT. For TAA, the linear term of Temp was not significant (p=0.9602) while the quadratic term of Temp was significant (p=0.0089). The interaction term of Temp-time (X₂X₃) (p<0.0001) and Temp-SSR (X₂X₄) (p=0.0095) were significant on HT. While Fig. 2 (a) shows the change in TAA value as a function of Temp and MeOH concentration, there is no p-value in the ANOVA results.

In the EtOH solvent system, the linear term of extraction Temp was both significant for HT and TAA (p< 0.0001). Similarly, the quadratic term of Temp was significant for both HT (p= 0.0150) and TAA (p<0.0001). The binary relationship of Temp with EtOH:H₂O concentration (p<0.0001, p<0.0001), Temp and time (p=0.0002, p<0.0001) were significant for both HT (Fig. 3(a) and Fig. 3(d)) and TAA (Fig. 4(a) and Fig. 4(d)). The interaction term of Temp and SSR was significant for TAA (p<0.0001), as shown in Fig. 4(e). Also, Fig. 3(e) shows the surface plot for the relationship of Temp and SSR for HT however it was not shown in the ANOVA table (Table IIIb). Similar positive effect of Temp on HT was found in the study of Irakli et al. (2018) who found that TPC rose as the Temp increased from 25 to 60°C in the UAE for phenolic compounds from OLLs. It was found that 60°C was the most effective temperature for acquiring phenolic compounds, including HT. However, no significant difference was found for TAA with increasing Temp (Irakli et al. 2018).

Similar results about HT, which acts in a Temp-dependent manner, were also obtained by Yao et al. (2019), who chose a temp of 60°C for the study of the optimization of UAE of total flavonoids and HT from OLLs. Furthermore, Stamatopoulos et al. (2014) found that high amounts of HT appeared when Temp particularly ≥ 60 °C. Similar to our results, an interaction term of Temp and time generated a significant effect on HT content. In terms of TAA, independent of time, a reduction in Temp resulted in a greater value of DPPH and FRAP (Nakilcioğlu-Taş & Ötleş 2019). Similarly, as the temperature climbed from 40 ̊C to 60 ̊C, the IC₅₀ reduced by 2.3±0.6 times (Wissam et al. 2016). Because of the high sensitivity of phenolics to heat, it is critical to define an upper limit to prevent the destruction of these thermosensitive bioactive compounds (Wissam et al. 2016).

Time (T)

Phenolic compounds generally rise steadily over time, resulting in increased extraction efficiency (Şahin et al. 2015). However, after a crucial period, the degree of chemical reactions, particularly phenolic oxidation, may increase, resulting in a fall in TAA level (Wissam et al. 2016, Şahin et al. 2015, Candrawinata et al. 2014). Time must thus be carefully considered and adjusted in order to protect polyphenols and antioxidants. The use of a modest extraction time has additional benefits in that it reduces solvent loss through vaporization as well as the process’s energy requirements (Nakilcioğlu-Taş & Ötleş 2019, Wissam et al. 2016).

OLL bioactive substances were extracted using hydroalcoholic solutions for 25 to 75 minutes. The important linear effect of time was acquired for HT (p=0.0054), and TAA (p=0.0027) in methanolic aqueous extract. The binary relationship of time and Temp (p<0.0001), time and SSR (p<0.0001) were significant for HT, whereas these interactions were not significant for TAA. Fig. 1(c) and Fig. 1(d) show the effect of time with Temp and time with SSR for HT, respectively.

In hydroethanolic extraction, the effects of the extraction time on HT (p=0.0001) and TAA (p<0.0001) were found statistically significant. The interaction effect of time and EtOH:H₂O concentration (p=0.0012, p<0.0001), time and Temp (p= 0.0002, p<0.0001), time and SSR (p<0.0001, p<0.0001) were significant for both HT and TAA, respectively. The response surface plots given in Figs. 3(b), 3(d), 3(f) and Figs. 4(b), 4(d) and 4(f) show these interactions, respectively. Also, the quadratic term of time was significant for both HT (p< 0.0001) and TAA (p< 0.0001).

The linear term of time and the interaction term of time and Temp on HT extraction were found to be statistically significant, which is consistent with our findings. It was reported that the content of HT increased by 9.80 % when the period at 40 °C extended from 30 to 60 minutes. Contrary to our findings, time did not significantly influence the change in TAA (Nakilcioğlu-Taş & Ötleş 2019). Yao et al. (2019) reported that the HT efficiency was highest in conditions where the ultrasonic effect was greater than 50 min. In another study, TAA was found to be constant between 10 and 30 min but decreased further as time increased. However, yields of other phenolic compounds, including HT, did not change after 10 min of UAE (Irakli et al. 2018).

Solvent:solid ratio (SSR)

In (MeOH:H₂O) SE, the linear term of SSR had no statistically significant effect on both HT and TAA (p>0.05) (Table IIIa). Although there was no significant relationship of SSR with the other variables for TAA; SSR and MeOH concentration (p=0.0476), SSR and Temp (p=0.0095) (Fig. 1a), SSR and time (p< 0.0001) were found significant for HT. The surface plot for the interaction between SSR and MeOH concentration is shown in Fig. 1(b), whereas the surface plot for the interaction between SSR and time on HT is shown in Fig. 1(d). Contrary to (MeOH:H₂O) SE, the linear term of SSR was significant on both HT and TAA (p<0.0001) in (EtOH:H₂O) SE. Similarly, the quadratic term of SSR was significant for both HT (p=0.0005) and TAA (p<0.0001), respectively. The interaction terms of SSR and EtOH concentration (X’₁ X’₄) (p<0.0001) and the SSR and time (X’₃ X’₄) were significant (p<0.0001) on HT and TAA. The changes in HT and TAA as a function of these variables were demonstrated by 3D response surface plots in Fig. 3(c), Fig. 3(f) and Fig. 4(c) and Fig. 4(f), respectively. Additionally, the interaction terms of SSR and Temp for TAA were significant (p<0.0001) which is shown in Fig. 4(e). Also, Fig. 3(e) shows the interaction between SSR and Temp, however, no significant value was found for HT in Table IIIb.

SSR and extraction temperature had a strong interactive impact. It should be noted that extremely high extraction temperatures may result in higher solvent loss owing to vaporization. The solvent evaporation changes SSR according to the solvent’s boiling point (Wissam et al. 2016, Khemakhem et al. 2017, Irakli et al. 2018). The determination of the optimum SSR is also important because it has a direct effect on the extraction process’s cost due to solvent consumption (Stamatopoulos et al. 2014). In the results of Stamatopoulos et al. (2014), who improved the multistage extraction approach in which OLLs were previously steam blanched, a solvent-solid ratio of 7:1 was proposed. Similarly, Bilek (2010) reported that the optimal point in the SE of phenolics from OLLs was seven times the SSR. In the study of Goldsmith et al. (2014), an SSR of 1:60 g/mL was suggested since less extraction solvent was consumed.

Optimization

The primary goal of the optimization study was to find the hydroalcoholic SE conditions that gave the maximum extraction yield for HT amount and TAA. To establish the ideal parameter values for the MOO process, the maximizing total DF model was created. The experiments were performed in random order using a CCD with six center points, the mean values of the triplicate trials were calculated, and the design and results are given in Table IV(a,b). A quadratic polynomial model for each response was done.

$yi={\beta }_0+{\sum }_{j=1}^4{\beta }_j{x}_j+{\sum }_{j’=2}^4{\sum }_{j=1}^{j\prime -1}{\beta }_{jj’}{x}_j{x}_{j\prime }+{\sum }_{j=1}^4{\beta }_{jj}{x_J^2}_{}+\epsilon i,i=\mathrm{1,2}$ (10)

The experimental region, after coding, was xj = [−α +α], j = 1, 2, 3, 4. Our aim in this study was to maximize the responses​ y1​ and ​y2​ and minimize the standard errors, under the constraints that following [​ymin​ , ​ymax​ ] and r=0.2 imposed for all the two individual DFs in Eq (1). The restrictions applied were given below:

for MeOH:H₂O solvent system​:

$y1\ge 0.001y2\ge 195$

for EtOH:H₂O solvent system:

$y1\ge 0.001y2\ge 121$

In the DF described by the geometric mean (Eq (5)), the predicted values were confirmed using least-squares predictions. For MeOH:H₂O hydroalcoholic SE, the optimal solution was (X₁, X₂, X₃, X₄)= (63.30 36.00 62.00 11.80) with DF=0.722 while for EtOH:H₂O hydroalcoholic SE, it was (X’₁, X’₂, X’₃, X’₄)= (43.80 52.00 58.00 9.40) with DF=0.891.

A maximum value of DF (0.722) for MeOH:H₂O solvent system is reached for Y ₁(HT) = 0.820 and Y ₂(TAA) = 455. The point DF (0.722) was chosen as the optimal solution for experimental validation. The results of the confirmation run are presented in Tables II and V. The optimum conditions in terms of controllable variables were [63.30 36.00 62.00 11.80] (X= [X₁,X₂,X₃,X₄]). Under such conditions, the responses confirmed by experiment were Y ₁(HT) = 0.809±0.110 mg/g dw and Y ₂(TAA) = 451±2.32 mM Trolox. Similarly, the maximum value of DF (0.891) for EtOH:H₂O solvent system was attained for Y’₁(HT) = 0.179 and Y’₂(TAA) = 298. This point was the optimal solution for experimental validation and the results of the confirmation run are given (Table II and Table V). The optimum conditions for controllable variables were as follows: [43.80 52.00 58.00 9.40] (X= [X’₁, X’₂, X’₃, X’₄]). The responses confirmed by experiment under such conditions were Y’₁(HT) = 0.175±0.004 mg/g dw and Y’₂(TAA)= 297±0.817 mM Trolox. The optimum point determined by the MeOH:H₂O solvent system was more appropriate for the enhanced extraction of a higher amount of HT and TAA.

Table V
Best compromise solution with observed and predicted response values for HT and TAA and comparison of Desirability.

	Method	Real values of independent variables X= [X1,X2,X3,X4]	Observed response values HT TAA	Predicted response value HT TAA
			(mg/g dw) (mM Trolox)	(mg/g dw) (mM Trolox)
MeOH/H 2 O hydroalcoholic solvent extraction	Desirability (0.722)	[63.30 36.00 62.00 11.80]	0.809±0.110 451±2.32	0.820 455
EtOH/H 2 O hydroalcoholic solvent extraction	Desirability (0.891)	[43.80 52.00 58.00 9.40]	0.175±0.004 297±0.817	0.179 298

• X₁: MeOH or EtOH conc; X₂: Temperature; X₃:Time; X₄: solvent:solid ratio, Analytical results are the average of triplicates (mean ± sd).

Figs. 5 and 6 indicate a close link between the experimental and estimated values, with no significant (p> 0.05) difference between them.

Figure 5
In MeOH solvent system comparison of experimental and predicted values of (a) HT (b) TAA.

Figure 6
In EtOH solvent system comparison of experimental and predicted values of (a) HT (b) TAA.

Method development and validation

HPLC method development and validation were performed for six important bioactive compounds (HT, Ty, L7G, OLE, A7G and Lut) of the OLL MeOH and EtOH aqueous extracts. The proposed HPLC method was evaluated in terms of accuracy, precision, linearity, range, limit of detection (LOD) and limit of quantitation (LOQ) (Singh 2013, Julia et al. 2011, Khan et al. 2012). Table II summarizes the method’s performance characteristics.

The accuracy was determined by comparing the measured and added concentrations (ICH 2005, Gonzalez et al. 2009, Singh 2013). After a spike was introduced to a blank sample, the percentage of analyte recovered was calculated. Accuracy was achieved at three concentrations covering the method range. In the evaluation of the mean recovery, the rule of being within 100%±5.0 over the entire studied range was valid (Al-Rimawi 2014, Green 1996, Winslow & Meyer 1997). The mean recovery and the RSD for each compound were calculated. It was determined that the validated method has good recovery for HT (97.32%±0.51), Ty (98.85%±1.02), L7G (97.98%±0.88), OLE (95.85%±1.45), A7G (98.33%±0.79), and Lut (98.43%±0.92) with a low RSD value (Table II).

Precision, defined as a measure of repeatability, was described as repeatability and intermediate precision in this study (Al-Rimawi 2014). Reproducibility is also known as intra-assay precision, defined as three replicates of each concentration and extract, and it is expressed as the RSD of the replicate (Singh 2013). RSD for repeat injections of the standard solutions with three concentrations were (0.5, 5 and 10 mg L^-1) for Ty, L7G, Lut (5, 10 and 25 mg L^-1) for HT, A7G and (50, 500 and 1000 mg L^-1) for OLE. They were determined to be 1.01, 0.70, 0.55, 1.70, 0.70 and, 0.35 for the analyzed compounds, respectively. The RSD of replicates was not higher than 1.5 (Al-Rimawi 2014, Huber 1998) demonstrating that the method is repeatable. Intermediate precision (three replicates of each concentration and extract, 3 days), called inter-day precision, measures the reproducibility of the result performed by the same method, on the same sample, in the same laboratory, but by different operators and on different days, to confirm that the method will produce the same results in the same laboratory after development (Al-Rimawi 2014). The intermediate precision of the method was assessed by determining the % recovery of the analyzed compounds by another analyst on a different day at three concentration levels: (0.5, 5 and 10 mg L^-1) for Ty, L7G, Lut; (5, 10 and 25 mg L^-1) for HT, A7G and (50, 500 and 1000 mg L^-1) for OLE. The RSDs were discovered to be 2.00, 2.25, 1.99, 4.25, 2.25, 1.80 for the analyzed compounds, respectively. Table II displays the inter-day and intra-day precisions (RSDs) for the bioactive chemicals studied.

In this work, the calibration equations’ correlation coefficients (r) for six analytes were more than 0.9921 (Table II). The significance of the deviation of tthe calibration line’s intercept was statistically evaluated by setting limits of confidence for the intercept, usually at the 95% level and the lack of fit was found to be insignificant. A one-way ANOVA with post test Tukey was employed to evaluate this set of data (Singh 2013, Miller & Miller 2005, Shabir et al. 2007).

In the validation of HPLC method to determine the quantification of HT, Ty, L7G, OLE, A7G, and Lut in OLL extract, the linearity was demonstrated over the range of 0.5 to 100 mg L^-1, 0.01 to 20 mg L^-1, 0.01 to 20 mg L^-1, 2.5 to 1750 mg L^-1, 0.25-50 mg L^-1, 0.010-20 mg L^-1, respectively. The estimation of LOD and LOQ was calculated using the formulas LOD = 3.3 Sa/b and LOQ = 10 Sa/b, based on signal-to-noise ratios of 3 and 10, respectively, where Sa is the standard deviation of the response and b is the slope of the calibration curve. The LOD and LOQ for six bioactive compounds of OLL extract ranged between 0.4.10^-3- 0.012 mg L^-1 and 1.7.10^-3-0.042 mg L^-1, respectively (Table II). As a consequence, the observed low LOD and LOQ values demonstrated that the approach is suitable for the detection and quantification of substances at low concentrations. The suggested HPLC technique may be utilized to determine valuable compounds such as HT, Ty, L7G, OLE, A7G and Lut in OLL extracts, as shown in Table II.

Individual phenolic amounts in OLLs

The leading bioactives (Ty, L7G, OLE, A7G, and Lut) in OLL were also determined by HPLC analysis in the present study. The DOE was studied over a wide range of intervals, and in these intervals these phenolic compounds of OLL extracts were determined. Ty, L7G, OLE, A7G, and Lut were determined in (MeOH:H₂O) SE in the range of 0.001-1.48 mg/g dw, 0-0.093 mg/g dw, 1.02-12.56 mg/g dw, 0.004-0.25 mg/g dw, 0-0.08 mg/g dw, whereas in (EtOH:H₂O) SE, the phenolics were determined as 0.001-2.89 mg/g dw, 0.005-0.68 mg/g dw, 0.70-7.90 mg/g dw, 0.003-0.35 mg/g dw, 0.002-0.11 mg/g dw, respectively Table IV (a,b). Additionally, Ty, L7G, OLE, A7G, and Lut were found at the optimum extraction points for both solvent systems. They were 0.09±0.003, 0.018±0.006, 9.25±0.23, 0.11±0.017, 0.05±0.009 in the optimum (MeOH:H₂O) SE points, whereas they were 1.28±0.04, 0.475±0.06, 4.89±0.07, 0.11±0.01, 0.07±0.006 mg/g dw in the optimum (EtOH:H₂O) SE points.

The findings of the studies on bioactive chemicals in OLLs were compared to the literature. Water, acetone, methanol, ethanol, and aqueous alcohol mixtures have been reported to be the solvents typically utilized in the extraction of polyphenols from olive leaves (Abaza et al. 2015). Our results are generally in agreement with previous findings, which indicated OLE to be the main phenolic component in both MeOH and EtOH hydroalcoholic extracts. Hayes et al. (2011) found six major phenolic compounds including OLE (1151.5±57.2 µg/ml), L7G (25.6±0.6µg/ml), A7G (15.9±0.7µg/ml), HT (10.2±0.1µg/ml) and Ty (15.6 ±0.1 µg/ml) in OLL-MeOH extract obtained by conventional extraction. OLE was the major component with 24.5% and followed by HT (1.5%), L7G (1.4%), verbascoside (1.1%), Ty (0.7%), A7G (1.4%) in olive leaf extract in the results of Benavente-Garcia et al. (2000). Similarly, the dominant phenolic and flavonoid components in OLL extracts were found to be OLE and L7G. Their levels were 65.57±0.70 g/kg and 1.32±0.03 g/kg for Soxhlet extraction, and 69.91±1.53 g/kg and 1.82±0.04 g/kg for UAE extraction, respectively. Wang et al. (2018) discovered several flavonoids in OLLs extracted by the UAE at optimum conditions using MeOH. A7G and Lut were found in concentrations ranging from 1.00±0.02–2.06±0.04 mg/g and 0.07±0.001–0.60±0.006 mg/g, respectively. In another study, Lut was determined as 1.42±0.04 mg/g in the OLL by SE using 50% MeOH (Haghi & Hatami 2010). Under ideal circumstances, OLE, A7G, and L7G were determined to be 2610±632 mg/kg, 1072±38 mg/kg, 970±43 mg/kg, respectively, employing a dynamic ultrasound-assisted method using an EtOH-H₂O combination as an extractant (Japón-Luján et al. 2006). Similarly, Xie et al. (2015) suggested that a 75% EtOH combination may be an appropriate solvent for extracting OLE. Under ideal conditions, it was recovered at 7.08 ± 0.05% by ultrasound-assisted and reduced-pressure extraction (URPE) from OLL. Unlike in other searches, OLE was not determined to be the dominant form; instead, the ideally obtained extract had substantial amounts of luteolin 7-O-glucoside and apigenin 7-O-rutinoside. This discrepancy has been attributed to varietal diversity and environmental factors (Mylonaki et al. 2008). Another factor influencing the extraction of phenols is the solvent’s variable polarity and solubility (Mohamed & Khan 2013). When EtOH extracts are compared to water extracts, the flavonoid concentration of ethanol extracts is shown to be greater (Quirantes-Piné et al. 2013, Herrero et al. 2011). Water, methanol, and ethanol were tested as solvents by Sifaoui et al. (2016) who found that the methanolic extract had the greatest extraction yield while the aqueous extract had the lowest. EtOH provided a lower concentration of antioxidant phenolics than methanol. In our study, high amount of HT and OLE were quantified in MeOH aqueous extract compared to EtOH aqueous extract which were also in accordance with TAA results higher in MeOH:H₂O extracts. However, more phenolic compounds (secoiridoids, simple phenols, phenolic acids, and flavonoids) should be identified in order to assess the components’ synergistic effects on the extract’s total antioxidant activity.

CONCLUSIONS

Olive leaf is a byproduct of olive processing that is widely known for its numerous health advantages, including its high antioxidant activity. The extraction procedure was modeled and optimized using chemometric approaches to acquire the highest amount of HT and TAA in olive leaf extract. Under RSM, a Box-Wilson-CCD/small factorial design was employed for the extraction’s DOE. MOO was solved using Pareto solutions, and DF was used to find the ideal input variable values. Under optimal conditions with MeOH:H₂O and EtOH:H2O solvent systems, the maximum quantity of HT and TAA was extracted from OLLs. The MeOH:H₂O solvent solution determined higher HT and TAA levels at the optimal point. Furthermore, OLE was shown to be the major component in both extracts, with a greater concentration in MEOH:H₂O extracts. In addition, a simple and reliable HPLC method for the isolation and quantification of Ty, L7G, OLE, A7G, and Lut was developed and validated. The obtained results could be a promising alternative for the valorization of olive leaves, which are a low-cost source of natural antioxidants. By carefully selecting the process conditions, it is possible to produce olive leaf extracts with strong antioxidant capacities for large-scale applications in the food, pharmaceutical, and cosmetics industries. It is critical to support the literature with experiments utilizing various extraction methods on various olive varietals in order to produce efficient, simple and low-cost procedures.

REFERENCES

• ABAZA L, TAAMALLI A, NSIR H & ZARROUK M. 2015. Olive tree (Olea europeae L.) leaves: Importanceand advances in the analysis of phenolic compounds. Antioxidants 4(4): 682-698.
• ABDELGAWAD SM, HASSAB MAE, ABOUREHAB MAS, ELKAEED EB & ELDEHNA WM. 2022. Olive leaves as a potential phytotherapy in the treatment of COVID-19 disease; A mini-review. Front Pharmacol 13: 879118.
• ALGAN CAVULDAK Ö, VURAL N, AKAY MA & ANLI RE. 2019. Optimization of ultrasound-assisted water extraction conditions for the extraction of phenolic compounds from black mulberry leaves (Morus nigra L.). J Food Process Eng 42(5): e13132.
• AL-RIMAWI F. 2014. Development and validation of a simple reversed phase HPLC-UV method for determination of oleuropein in olive leaves. J Food Drug Anal 22: 285-289.
• BENAVENTE-GARCIA O, CASTILLO J, LORENTE J, ORTUNO A & DEL RIO JA. 2000. Antioxidant activity of phenolics extracted from Olea europaea L. leaves. Food Chem 68: 457-462.
• BILEK SE. 2010. The effects of time, temperature, solvent: solid ratıo and solvent composition on extraction of total phenolic compound from dried olive (Olea europaea l.) leaves. The Journal of Food 35(6): 411-416.
• BOUAZIZ M & SAYADI S. 2005. Isolation and evaluation of antioxidants from leaves of a Tunisian cultivar olive tree. Eur J Lipid Sci Technol 107: 497-504.
• CANDRAWINATA VI, GOLDING JB, ROACH PD & STATHOPOULOS CE. 2014. Total phenolic content and antioxidant activity of apple pomace aqueous extract: effect of time, temperature and water to pomace ratio. Int Food Res J 21(6): 2337-2344.
• CHAN SW, LEE CY, YAP CF, WAN AWM & HO CW. 2009. Optimisation of extraction conditions for phenolic compounds from limau purut (Citrus hystrix) peels. Int Food Res J 16(2): 203-213.
• CHEN HW, WONG WK & XU H. 2012. An augmented approach to the desirability function. J Appl Stat 39(3): 599-613.
• CLODOVEO ML, CRUPI P, ANNUNZIATO A & CORBO F. 2022. Innovative extraction technologies for development of functional ingredients based on polyphenols from olive leaves. Foods 11: 103.
• CRUDELE A ET AL. 2022. Hydroxytyrosol recovers SARS-CoV-2-PLpro-dependent impairment of interferon Rrelated genes in polarized human airway, intestinal and liver epithelial cells. Antioxidants 11: 1466.
• D’ANGELO S, INGROSSO D, MIGLIARDI V, SORRENTINO A, DONNARUMMA G, BARONIC A, MASELLA L, TUFANO MA, ZAPPIA M & GALLETTI P. 2005 Hydroxytyrosol, a natural antioxidant from olive oil, prevents protein damage induced by long-wave ultraviolet radiation in melanoma cells. Free Radic Biol Med 38: 908-919.
• DIAS MC, FIGUEIREDO C, PINTO DCGA, FREITAS H, SANTOS C & SILVA AMS. 2019. Heat shock and UV-B episodes modulate olive leaves lipophilic and phenolic metabolite profiles. Ind Crops Prod 133: 269-275.
• ERBAY Z & ICIER F. 2010. The importance and potential uses of olive leaves. Food Rev Int 26 (4): 319-334.
• GALANAKIS CM, GOULAS V, TSAKONA S, MANGANARIS GA & GEKAS V. 2013. A knowledge base for the recovery of natural phenols with different solvents. Int J Food Prop 16(2): 382-396.
• GALLARDO-FERNÁNDEZ M, GONZALEZ-RAMIREZ M, CEREZO AB, TRONCOSO AM & GARCIA-PARRILLA MC. 2022. Hydroxytyrosol in Foods: Analysis, Food Sources, EU Dietary Intake, and Potential Uses. Foods 11: 2355.
• GOLDSMITH CD, VUONG QV, STATHOPOULOS CE, ROACH PD & SCARLETT CJ. 2014. Optimization of the aqueous extraction of phenolic compounds from olive leaves. Antioxidants 3: 700-712.
• GONZALEZ O, IRIARTE G, FERREIRÓS N, MAGUREGUI MI, ALONSO RM & JIMÉNEZ RM. 2009. Optimization and validation of a SPE-HPLCPDA-fluorescence method for the simultaneous determination of drugs used in combined cardiovascular therapy in human plasma. J Pharm Biomed Anal 50(4): 630-639.
• GOULAS V, PAPOTI VT, EXARCHOU V, TSIMIDOU MZ & GEROTHANASSIS IP. 2010. Contribution of flavonoids to the overall radical scavenging activity of olive (Olea europaea L.) leaf polar extracts. J Agric Food Chem 58 (6): 3303-3308.
• GRANADOS-PRINCIPAL S, QUILES JL, RAMIREZ-TORTOSA CL, SANCHEZ-ROVIRA P & RAMIREZ-TORTOSA MC. 2010. Hydroxytyrosol: from laboratory investigations to future clinical trials. Nutr Rev 68(4): 191-206.
• GREEN JM. 1996. Peer reviewed: A practical guide to analytical method validation. Anal Chem 68(9): 305A-309A.
• HABIBI H, MOHAMMADI A, FARHOODI M & JAZAERI S. 2018. Application and optimization of microwave-assisted extraction and dispersive liquid–liquid microextraction followed by high-performance liquid chromatography for the determination of oleuropein and hydroxytyrosol in olive pomace. Food Anal Method 11: 3078-3088.
• HAGHI G & HATAMI A. 2010. Simultaneous quantification of flavonoids and phenolic acids in plant materials by a newly developed isocratic high-performance liquid chromatography approach. J Agric Food Chem 58: 10812-10816.
• HALOUI E, MARZOUK B, MARZOUK Z, BOURAOUI A & FENINA N. 2011. Hydroxytyrosol and oleuropein from olive leaves: Potent anti-inflammatory and analgesic activities. J Food Agric Environ 9: 128-133.
• HARRINGTON EC. 1965. The desirability function. Ind Qual Control 21: 494-498.
• HASSEN I, CASABIANCA H & HOSNI K. 2015. Biological activities of the natural antioxidant oleuropein: Exceeding the expectation-A mini-review. J Funct Foods 18: 926-940.
• HAYES JE, ALLEN P, BRUNTON N, O’GRADY MN & KERRY JP. 2011. Phenolic composition and in vitro antioxidant capacity of four commercial phytochemical products: Olive leaf extract (Olea europaea L.), lutein, sesamol and ellagic acid. Food Chem 126(3): 948-955.
• HERRERO M, TEMIKHON N, TEMIRZODA TN, SEGURA-CARRETERO A, QUIRANTES R, PLAZA M & IBANEZ E. 2011. New possibilities for the valorization of olive oil by-products. Journal of Chromatogr A 1218: 7511-7520.
• HUBER L. 1998. Validation of analytical methods, Validation and qualification in the analytical laboratories Buffalo Grove, IL: Interpharm Press.
• ICH. 2005. Validation of analytical procedures: Text and methodology Q2 (R1). International Conference on Harmonization, Geneva, Switzerland, 11-12.
• IRAKLI M, CHATZOPOULOU P & EKATERINIADOU L. 2018. Optimization of ultrasound-assisted extraction of phenolic compounds: Oleuropein, phenolic acids, phenolic alcohols and flavonoids from olive leaves and evaluation of its antioxidant activities. Ind Crop Prod 124: 382-388.
• JAPÓN-LUJÁN R, LUQUE-RODRIGUEZ JM & LUQUE DE CASTRO MD. 2006. Dynamic ultrasound-assisted extraction of oleuropein and related biophenols from olive leaves. J Chromatogr A 1108: 76-82.
• JEMAI H, EL FEKI A & SAYADI S. 2009. Antidiabetic and antioxidant effects of hydroxytyrosol and oleuropein from olive leaves in alloxan-diabetic rats. J Agric Food Chem 57: 8798-8804.
• JULIA T, MENA AJ, AUCOIN MG & KAMEN AA. 2011. Development and validation of a HPLC method for the quantification of baculovirus particles. J Chromatogr B 879(1): 61-68.
• KARANDE P, GAURI SK & CHAKRABORTY S. 2013. Applications of utility concept and desirability function for materials selection. Mater Des 45: 349-358.
• KHAN MC, REDDY NK, RAVINDRA G, REDDY KVSRK & DUBEY PK. 2012. Development and validation of a stability indicating HPLC method for simultaneous determination of four novel fluoroquinolone dimers as potential antibacterial agents. J Pharm Biomed Anal 59: 162-166.
• KHEMAKHEM I, AHMAD-QASEM MH, CATALÁN EB, MICOL V, GARCIA-PÉREZ JV, AYADI MA & BOUAZIZ M. 2017. Kinetic improvement of olive leaves’ bioactive compounds extraction by using power ultrasound in a wide temperature range. Ultrason Sonochem 34: 466-473.
• LAMA-MUÑOZ A, DEL MAR CONTRERAS M, ESPINOLA F, MOYA M, ROMERO I & CASTRO E. 2019. Optimization of oleuropein and luteolin-7-O-glucoside extraction from olive leaves by ultrasound-assisted technology. Energies 12: 2486.
• LEE OH & LEE BY. 2010. Antioxidant and antimicrobial activities of individual and combined phenolics in Olea europaea leaf extract. Bioresour Technol 101(10): 3751-3754.
• MARKOVIĆ AK, TORIĆ J, BARBARIĆ M & JAKOBUŠIĆ BRALA C. 2019. Hydroxytyrosol, tyrosol and derivatives and their potential effects on human health. Molecules 24(10): 2001.
• MEDINA E, ROMERO C, GARCÍA P & BRENES M. 2019. Characterization of bioactive compounds in commercial olive leaf extracts, and olive leaves and their infusions. Food Funct 10: 4716-4724.
• MILLER JN & MILLER JC. 2005. Statistics and chemometrics for analytical chemistry Harlow: Pearson Prentice Hall.
• MOHAMED SA & KHAN JA. 2013. Antioxidant capacity of chewing stick miswak Salvadora persica. BMC Complement Altern Med 13: 40.
• MYLONAKI S, KIASSOS E, MAKRIS DP & KEFALAS P. 2008. Optimisation of the extraction of olive (Olea europaea) leaf phenolics using water/ethanol-based solvent systems and response surface methodology. Anal Bioanal Chem 392: 977-985.
• NAKILCIOĞLU-TAŞ E & ÖTLEŞ S. 2019. The optimization of solid–liquid extraction of polyphenols from olive stone by response surface methodology. J Food Meas Charact 13: 1497-1507.
• PARADISO VM, DI MATTIA C, GIARNETTI M, CHIARINI M, ANDRICH L & CAPONIO F. 2016. Antioxidant behavior of olive phenolics in oil-in-water emulsions. J Agric Food Chem 64: 5877-5886.
• PÉREZ-SERRADILLA J, DYNAMIC-LUJAN R & DE CASTRO ML. 2007. Simultaneous microwave-assisted solid–liquid extraction of polar and nonpolar compounds from alperujo. Anal Chim Acta 602: 82-88.
• QUIRANTES-PINÉ R, LOZANO-SÁNCHEZ J, HERRERO M, IBÁÑEZ E, SEGURA-CARRETERO A & FERNÁNDEZ-GUTIÉRREZA A. 2013. HPLC–ESI–QTOF–MS as a powerful analytical tool for characterising phenolic compounds in olive-leaf extracts. Phytochem Anal 24: 213-223.
• RAHMANIAN N, JAFARI SM & WANI TA. 2015. Bioactive profile, dehydration, extraction and application of the bioactive components of olive leaves. Trends Food Sci Tech 42(2): 150-172.
• RE R, PELLEGRINI N, PROTEGGENTE A, PANNALA A, YANG M & RICE-EVANS C. 1999. Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radical Bio Med 26(9/10): 1231-1237.
• ŞAHIN S & BILGIN M. 2018. Olive tree (Olea europaea L.) leaf as a waste by-product of table olive and olive oil industry: A review. J Sci Food Agric 98: 1271-1279.
• ŞAHIN S, ILBAY Z & KIRBAŞLAR Şİ. 2015. Study on optimum extraction conditions for olive leaf extracts rich in polyphenol and flavonoid. Sep Sci Technol 50(8): 1181-1189.
• SHABIR GA, LOUGH WJ, ARAIN SA & BRADSHAW TK. 2007. Evaluation and application of best practice in analytical method validation. J Liq Chromatogr RT 30 (3): 311-333.
• SIFAOUI I, MECHA E, SILVA A, CHAMMEM N, MEJRI M, ABDERABBA M & BRONZE MR. 2016. Optimized extraction of antioxidants from olive leaves using augmented simplex centroid design. Anal Lett 49(9): 1323-1333.
• SINGH R. 2013. HPLC method development and validation- an overview. J Pharm Educ Res 4(1): 26-33.
• STAMATOPOULOS K, CHATZILAZAROU A & KATSOYANNOS E. 2014. Optimization of multistage extraction of olive leaves for recovery of phenolic compounds at moderated temperatures and short extraction times. Foods 3: 66-81.
• TAAMALLI A, ARRÁEZ-ROMÁN D, IBAÑEZ E, ZARROUK M, SEGURA-CARRETERO A & FERNÁNDEZ-GUTIÉRREZ A. 2012. Optimization of microwave-assisted extraction for the characterization of olive leaf phenolic compounds by using HPLC-ESI-TOF-MS/IT-MS2. J Agric Food Chem 60: 791-798.
• TAKEDA Y, JAMSRANSUREN D, MATSUDA S, CREA R & OGAWA H. 2021. The SARS-CoV-2-inactivating activity of hydroxytyrosol-rich aqueous olive pulp extract (HIDROX®) and its use as a virucidal cream for topical application. Viruses 13: 232.
• TALHAOUI N, TAAMALLI A, GÓMEZ-CARAVACA AM, FERNÁNDEZ GUTIÉRREZ A & SEGURA-CARRETERO A. 2015. Phenolic compounds in olive leaves: Analytical determination, biotic and abiotic influence, and health benefits. Food Res Int 77: 92-108.
• THOO YY, HO SK, LIANG JY, HO CW & TAN CP. 2010. Effects of binary solvent extraction system, extraction time and extraction temperature on phenolic antioxidants and antioxidant capacity from mengkudu (Morinda citrifolia). Food Chem 120: 290-295.
• TSAKONA S, GALANAKIS CM & GEKAS V. 2012. Hydroethanolic mixtures for the recovery of phenols from mediterranean plant materials. Food Bioproc Tech 5: 1384-1393.
• VURAL N, ALGAN CAVULDAK Ö, KENAR A & AKAY MA. 2020. Green alcoholic solvent and UAE extraction of oleuropein from the Olea europaea L. leaves: Experimental design, optimization, and comparison with Pharmacopoeia method. Sep Sci Technol 55(10): 1813-1828.
• WANI TA, MASOODI FA, GANI A, BABA WN, RAHMANIAN N, AKHTER R, WANİ IA & AHMAD M. 2018. Olive oil and its principal bioactive compound: Hydroxytyrosol – A review of the recent literature. Trends Food Sci Technol 77: 77-90.
• WANG B, QU J, LUO S, FENG S, LI T, YUAN M, HUANG Y, LIAO J, YANG R & DING C. 2018. Optimization of ultrasound-assisted extraction of flavonoids from olive (Olea europaea) leaves, and evaluation of their antioxidant and anticancer activities. Molecules 23: 2513.
• WINSLOW PA & MEYER RF. 1997. Defining a master plan for the validation of analytical methods. Computer Science J Valid Technol 14: 361-367.
• WISSAM Z, ALI A & RAMA H. 2016. Optimization of extraction conditions for the recovery of phenolic compounds and antioxidants from Syrian olive leaves. J Pharmacogn Phytochem 5(5): 390-394.
• WU FC. 2004. Optimization of correlated multiple quality characteristics using Desirability Function. Qual Eng 17(1): 119-126.
• XIE P, HUANG L, ZHANG C, YOU F & ZHANG Y. 2015. Reduced pressure extraction of oleuropein from olive leaves (Olea europaea L.) with ultrasound assistance. Food Bioprod Process 93: 29-38.
• YAO Q, SHEN Y, BU L, YANG P, XU Z & GUO X. 2019. Ultrasound-assisted aqueous extraction of total flavonoids and hydroxytyrosol from olive leaves optimized by response surface methodology. Prep Biochem Biotechno 49(9): 837-845.
• ZUORRO A. 2020. Enhanced lycopene extraction from tomato peels by optimized mixed-polarity solvent mixtures. Molecules 25: 2038.
• ZUORRO A, LAVECCHIA R, GONZÁLEZ-DELGADO AD, GARCÍA-MARTINEZ JB & L’ABBATE P. 2019. Optimization of enzyme-assisted extraction of flavonoids from corn husks. Processes 7: 804.

Publication Dates

• Publication in this collection
  27 May 2024
• Date of issue
  2024

History

• Received
  26 May 2023
• Accepted
  28 June 2023