Response Surface Methodology‐Based Optimized Ultrasonic‐Assisted Extraction and Characterization of Selected High‐Value Components from Gemlik Olive Fruit

This article presents an optimized ultrasound‐assisted ethanolic extraction (UAEE) and characterization of selected high‐value components from Gemlik olive fruit (GOF) harvested from Potohar region of Pakistan. Response surface methodology (RSM), involving central composite design (CCD), was applied to optimize the extraction variables i. e., temperature (25–65 °C), extraction time (15–45 min) and aqueous ethanol concentration (60–90 %) for optimal recovery of bioactives extract, total phenolic contents (TPC) and DPPH free radical scavengers. Under the optimized set of conditions such as 43 °C temperature, 32 min extraction time and 80 % aqueous ethanol, the best extract yield (218.82 mg/g), TPC (19.87 mg GAE/g) and DPPH scavenging activity (63.04 %) were recorded. A quadratic polynomial model was found to be reasonably fitted to the observed results for extract yield (p<0.0001 and R2=0.9941), TPC (p<0.0001 and R2=0.9891), and DPPH radical scavenging activity (p<0.0001 and R2=0.9692). Potent phenolic compounds were identified by GC/MS in GOF extract and considerable amount of essential fatty acids were also detected. The current findings support the use of UAEE as an effective green route for optimized recovery of high‐value components from GOF and hence its applications can be extended to functional food and nutra‐pharmaceutical developments.


Introduction
With the revival of interest in the use of plants as a source of food and medicine, now there is increasing interest in the exploration of plants-derived high-value components as natural health supplements and nutraceuticals. [1][2][3] Among the major classes of phytochemicals, the phenolics are especially valued as an important group of bioactives due to their strong medicinal and chemo-preventive roles. In fact, plant phenolics represent an important and a major class of natural antioxidants and exhibit multiple biological such as free radical scavenging, anticarcinogenic, anti-inflammatory and antimicrobial activities. [4] Based on the medicinal potential and antioxidant attributes of plant phenolics, functional food and therapeutic industry is keenly interested in the utilization of these compounds as natural preservatives and chemo-preventive agents. [5,6] The extraction and isolation of phenolics from various plant materials is a challenging task due to their variable concentration, bioactivities and structural features. [4,7] Various conventional techniques, mostly involving the toxic and non-biodegradable organic solvents, have been employed for the extraction of plant phenolics. However, currently, there is greater focus of researchers on the applications of eco-friendly green extractions for the same purposes. [4,5] In this regard, ultrasonic-assisted extraction (UAE) is gaining popularity as a quick, efficient, and inexpensive method compared to conventional extraction techniques. [4,6,8] However, the recovery rate and extraction yield of phenolic compounds depend upon optimization of various factors such as temperature, extraction time, solvent concentration, ultrasonic frequency, type of solvent and solvent sample ratio using response surface methodology (RSM). [4,7,9] Among the ancient crops of Mediterranean basin, the olive (Olea europaea L.) is famous as a nutrient-dense food plant with multiple traditional health benefits. [10,11] A native to the Eastern Mediterranean Basin, the Olive, due to its growing demand, is now widely cultivated in several other regions such as the coastal areas of the Southeastern Europe, North Africa, Western Asia and Northern Iran. [10] In Pakistan olive is gaining recognition as an emerging oil yielding and functional food crop with appreciable economic benefits for the farmers and community. The cultivation of olives in Pakistan has been extensively started by the last 10-15 years.
Pakistan is a promising country for cultivation of olives on commercial scale due to its favorable agro-climatic conditions and topographic diversity. Due to availability of suitable agroecological conditions of the land, wild olive subspecies are naturally distributed on large scale in several areas of the country, especially, in the provinces of Punjab, Khyber Pakhtunkhwa and Baluchistan that can be grafted to develop productive cultivars. With the financial and technical assistance of Federal and Provincial governments, and Barani Agriculture Research Institute (Chakwal, Punjab), around areas of 10 million acres have already been identified for olives cultivation across the country. Keeping in view the agro-climatic and topographic conditions suitable for olive cultivation, The Pothwar region in Punjab province of Pakistan is being declared as an "Olive Valley" with nine million olive plants already been planted. [12] In the Pothwar region numerous new olive cultivars are developed and/or are under development trials; however, their nutraceutical profiling, especially, with regard to composition of high-value bioactives such as phenolics, have not been yet explored. As with other fruit crops, specific agro-climatic and agro-ecological conditions may impart a unique bioactives profile to olive fruits in this particular region. So, there is dire need to explore the valuable components profile so as to establish scientific basis for the functional food and nutraceutical benefits of such newly developed local olive cultivars. The present research is mainly planned to unveil and explore the profiling of high-value components (mainly phenolics) in an important newly developed olive cultivar namely Gemlik Olive fruit (GOF), harvested from Pothwar region of Pakistan. An optimized ultrasound-assisted ethanolic extraction protocol has been devised for efficient/optimal recovery of phenolics from olive fruit using RSM. The important variables namely temperature, extraction time, and aqueous ethanol concentration (AEC) have been optimized involving central composite design for improved yield and antioxidant attributes of the bioactives extract. Moreover, the bioactive components have been analytically characterized using GC/MS technique.

Effect of Ripening Stage on Antioxidant Compounds
The ripening stages of olive fruit have a significant effect on the extract yield and composition of phenolic bioactives in the fruit. [12] The changes in phenolic content and composition during ripening are attributed to various factors such as changes in enzymatic activity, oxidative stress, and genetic variability of cultivars. [10,12] The TPC was found to be 17.76 � 0.54, 15.12 � 0.47 and 11.67 � 0.70 mg/g DW for unripen, semiripen and fully-ripen olives. Based on this, the unripen olives were selected for bioactives extraction.

Effect of Solid-to-Liquid Ratio
The mass transfer of bioactive compounds from solid materials to extraction solvents also depends on the sample-to-solvent ratio. [13] A proper sample-to-solvent ratio must be chosen in order to achieve the best extract yield and its antioxidant potential. The TPC value increased gradually and peaked at a sample-to-solvent ratio (w/v) of 1g/10 mL. Nevertheless, the extraction yield of phenolic components decreased as the sample-to-solvent ratio was raised than the given optimum ratio. Therefore, it was considered that 1/10 (w/v) was the ideal sample-to-solvent ratio offering best yield.

Effect of Solvent
The choice of solvent for the extraction of phenolic bioactives is based on various factors such as specific properties and chemical nature of the compounds of interest, and safety, cost, and efficiency of the extraction medium/solvent. It is also important to consider the potential impact of the solvent on the environment and human health. It can be noted that in the present study, pure water offered the least TPC, while 80 % aqueous ethanol represented the highest TPC, as shown in Figure 2A.

Effect of Temperature
Temperature has a significant effect on the extraction of phenolic bioactives from plant materials. The optimal temperature for extraction depends on various factors, including the type of phenolic compounds being extracted, the solvent used, and the nature of the plant material. In general, increasing the temperature can increase the solubility of phenolic compounds in the solvent, resulting in higher extraction yields. However, excessive high temperatures can also lead to the degradation or denaturation of the phenolic compounds thus decreasing their bioactivities. In the present study, although maximum TPC was recovered at 40°C, but no significant difference was noticed over 40°C to 60°C. Meanwhile, at 80°C, a significantly low TPC was extracted.

Effect of Extraction Time
The time of extraction is another important factor that can affect the extraction of phenolic compounds from a plant material. In general, increasing the extraction time can increase the yield of phenolic compounds, but up to a certain period/ level. Beyond that period, the extraction yield may not be affected or even may decrease, as prolonged extraction times may lead to the degradation or denaturation of the phenolic compounds. Longer extraction times may lead to the loss of some phenolic compounds due to their degradation or denaturation. It is important to note that the effect of extraction time on phenolic yield can vary depending on the type of plant material and the nature of specific phenolic compounds being targeted. In the current study trials, the maximum TPC was observed for 30 min extraction time. Wardhani et al. (2013) investigated the effect of various conditions during UAE of phenolic antioxidants from Eucheuma cottonii seaweed. The highest TPC (4.44 GAE mg/g DW) was found at 1 : 3 (w/v) solid-to-liquid ratio, 35 min of extraction time, and 55°C of extraction temperature. [14] In another study, Wang et al, (2013) investigated the optimum ethanol concentration (75.3 %), solid-to-liquid ratio (1 : 22.72 w/v), and extraction time (33.54 min) during UAE of phenolic bioactives from Sparganii rhizoma. [15] Similarly, Bouafia et al. (2021) appraised optimal extraction conditions to be 34.2 mint time, 55°C temperature, and 0.5 g/55 mL solid to liquid ratio for the best recovery of TPC and antioxidant potential of Centaurea sp. [16] Ameer et al. (2019) optimized the influence of UAE conditions on the extraction of phenolic bioactives from Stevia rebaudiana leaves via a hybrid RSMÀ ANNÀ GA approach. The maximum yield of bioactive compounds was obtained at 75 % ethanol, 43 min extraction time, and a 0.28 g/mL leaf-to-solvent ratio. [17] In the present study, maximum TPC was found in unripen Gemlik olives using 10 g of solute per 100 mL of 80 % aqueous ethanol using UAE at 40°C extraction temperature and 30 min extraction time. The results ( Figure 1) revealed that extraction at elevated temperatures (80°C) and longer extraction times disfavors the recovery of TPC from olive fruit. The investigated solid-liquid ratio is higher than that reported by Bouafia et al. (2021) [16] and lower than that found by Wang et al. (2012). [15] Such variation of optimal conditions among studies might be linked to the varying nature of extraction solvent and the material employed. [15,[27][28][29]

Fitness of the Multi-Response Model
Linear, quadratic and interaction terms of the selected variables were studied for statistical analysis (F-value and p-value) using response surface second-order model as illustrated in Table 1. The results obtained for the devised model reveal that it is significant (p < 0.05) and hence might be employed for optimization of extraction process. In the case of extraction yield, the model terms A, B, C, AC, BC, A 2 B 2, and C 2 exerted a significant effect. For TPC, the model terms B, C, AB, BC, A 2 B 2 and C 2 were significant. While for the DPPH radical scavenging A, B C, BC, CA, A 2 , B 2, and C 2 were the significant model terms.
The empirical regression equation for the extract yield (Y 1 ), TPC (Y 2 ), and DPPH radical scavenging (Y 3) as a function of temperature (A), extraction time (B), and ethanol concentration (C) are shown as, respectively The lack of fit F-value (3.02, 1.05, and 1.75 for extract yield, total phenolic contents, and DPPH radical scavenging, respectively) implies that it is insignificant compared to the pure error ( Table 2). The quadratic regression model displayed that the value of the determination coefficient (R 2 ) was 0.9941, 0.9891and 0.9609 for extract yield, total phenolic contents, and DPPH radical scavenging, respectively. It means that the model could elaborate on 99.41, 98.91, and 96.09 % of the variation for  (Table 2). For a decent statistical model, adj. R 2 should be approximate near to R 2 . The ''pred R 2 '' and ''adj R 2 '' for extract yield (0.9583 and 0.9888), TPC (0.9467 and 0.9793), and DPPH radical scavenging (0.8953 and 0.9582) are in reasonable agreement with each other. Additionally, relatively low values of coefficient of variation (0.73, 0.98, and 1.43 % for extract yield, total phenolic contents, and DPPH radical scavenging, accordingly) depicted good reliability of the experimental results. Adeq. precision measures the signal to noise ratio, and a ratio larger than four is required. Here in this model, the ratio of 37.886, 30.422, and 20.108 for the selected responses indicated high compatibility for the model.
In a previous study by Chakraborty et al. (2020), the Adeq. precision was 25.12 and 53.16 for antioxidant and TPC, respectively. [18] Similarly, in another study the "pred R 2 '' and ''adj R 2 '' were calculated, during RSM based optimization, for TPC (0.96 and 0.91), DPPH (0.7835 and 0.9258), and extract yield (0.94 and 0.88) from pomegranate peels. [19] The above data established that the models given by the respective regression equation in the present work can be successfully applied to navigate the design space.

Prediction of Multiple Response
Optimization of selected variables such as temperature (T), time (t), and AEC (%) was carried out at different levels to analyze the improved recovery of extractable compounds mainly the phenolic bioactive components from GOF. The process variables were optimized against the selected responses such as extract yield, total phenolic contents, and DPPH radical scavenging, as shown in Table 3. In the present RSM design, extract yield (222.06 mg/g dried weight) and total phenolic contents (20.52 GAE mg/g) were found to be highest at optimized UAE conditions of temperature 45°C, time 30 min and AEC 75 %. While, the highest DPPH percent scavenging (75.66 %), was examined in the extract obtained at temperature 45°C, time 30 min and 100 % ethanol. A close agreement was observed     responses (extract yield, TPC, and DPPH free radical scavenging). There existed a significant effect of temperature on extract yield and radical scavenging but was non-significant in the case of TPC. The extract yield and DPPH radical scavenging (%) reached at the maximum level around 45°C. After 45°C extract yield became constant, while DPPH radical scavenging (%) started decreasing. But in the case of TPC, the initial increase in temperature considerably increased the TPC value but later on, a slight increase was noticed. Higher temperature facilitates higher content to diffuse into the medium up to a certain level; further increase in temperature may disrupt the chemical nature of bioactive compounds, thereby reducing the extract yield. Similarly, extract yield, and TPC was considerably timedependent, and a parallel relation was observed in relation to time, but after half an hour, both of them (extract yield and TPC) became almost constant. DPPH radical scavenging (%) was directly dependent on time (required for extraction). Prolonged extraction may enhance the mass transfer from solute to the solvent. The effect of percent AEC was significant in the case of all three responses. Extract yield and TPC increased with an increase in ethanol concentration up to 75 %; however, after that the values of extract yield and TPC reduced with further increase in EC%. The extraction conditions for the DPPH radical scavenging (Figure 2) suggested that the prolonged extraction at elevated temperature lowered the DPPH radical scavenging potential of extract which might be resulted due to degradation of bioactive (responsible for this activity) due to longer exposer at elevated temperature.
Absolute ethanol concentration was ideal for DPPH radical scavenging. The increase in ethanol (%) may provide better interaction with phenolic bioactive components and promote more diffusion of polyphenol that might be linked to higher DPPH radical scavenging of the corresponding extract. However, to maximize the extract yield, it was noted that the increase in temperature up to certain extent was a more dominating factor as compared to ethanol proportion. The mutual interactive effect of selected variables on responses (extract yield, TPC, and DPPH free radical scavenging) in depicted in 3D plot (Figure 2). Figure 2, plot (A, D, and G) reveals that the time-temperature interactive effect (at 75 % AEC) on the extract yield and DPPH was non-significant but significant for TPC. Similarly, Figure 2, plot (B, E, and H) depicts the significant interactive effect between time and AEC (at constant temperature 45°C) for extract yield and TPC and nonsignificant for DPPH radical scavenging. However, there was a significant interaction between temperature and AEC in the case of extract yield but insignificant for TPC and DPPH, as shown in Figure 2, plot (C, F, and I). The observed optimum parameters were found to be 42.68°C, 32.09 min and 79.83 % for the selected responses. The experimental and predicted response values for extract yield, TPC, and DPPH radical scavenging activity are shown in Table 4.
These results are in line with studies done by various researchers. For example, Sun et al. (2015) investigated that the highest extract yield, from Beijing propolis, was observed at 75 % EC. [20] Similarly, according to Tabaraki et al. (2012), the optimal conditions for UAE extraction of phenolic bioactives from pomegranate peel was found to be 70 % AEC (extracting solvent), 30 min extraction time and 60°C extraction temperature. [19] Furthermore, Wu et al. (2017) reported 75.3 % ethanol and 43.5 min extraction time as the optimal UAE conditions for best recovery of TPC from Chrysanthemi flos. [21] Do et al. 2014 investigated the effect of various solvents on the antioxidant potential of Limnophila aromatica extracts and observed that 100 % ethanolic extract showed the highest DPPH radical scavenging potential. [22] The small variations in optimal extraction conditions in our study compared to literature reports may be due to difference in physico-chemical nature of material employed for extraction.
Moreover, the optimized UAEE extract yield, TPC and DPPH radical scavenging capacity were compared with those of the maceration process (controlled). It is evident (from Table 5) that optimized UAEE is more efficient and less time-consuming technique for extraction of bioactives in comparison with MEE. Using UAEE, we noted a significant (p > 0.05) increase in the extract yield, TPC and DPPH radical scavenging as compared with MEE.

Phytochemicals Characterization in GOF Extract using GC/MS
The UAEE based typical GOF extract (with high antioxidant potential as evident by the determinations of different assays) was further characterized by GC/MS for profiling of bioactives.  GC/MS analysis (Table 6) showed the separation and presence of different high-value compounds. The total ion chromatogram of GOF extract is displayed as Figure 3. The results given in Table 6 revealed the presence of different bioactives GC/MS Library Search Report and other supportive identification tools indicated that different classes of compounds such as phenolics, terpenoids, esters of fatty acids, tocopherol etc. were present in the GOF extract (Table 6). Major identified phenolic compounds were found to be vanillic acid ( . Furthermore, fatty acids such as; 9,12,15-octadecatrienoic acid (linolenic), 9,12-octadecadienoic acid (linoleic), stearic acid, hexadecanoic acid, undecanoic acid, 3-methyldecanoic acid, ricinoleic acid, 9-octadecenoic acid (9Z)-oxiranyl and 2-palmitoylglycerol were also detected in sizeable amounts. The linolenic acid, was the highest (11. 37 %) among all fatty acids. Moreover, significant quantity of methyl 2-formylbenzoate, an important phenolic lipid was also identified.
In the present study vanillic acid is detected as the major (2.45 %) while p-hydroxyphenol as the minor (0.35 %) phenolic bioactives among other characterized phenolics. This result is in agreement with that investigated by Uylaşer et al. (2015) who reported that vanillic acid was the highest (at un-ripened/green stage of olive) among others samples harvested from Mudanya (571.05 mg/kg FW), Umurbey (2071.37 mg/kg FW), Cagrisan (971.30 mg/kg FW) and Kumla (485.25 mg/kg FW) locations. [23] Furthermore, Benlarbi et al. (2018) reported that vanillic acid, in green olive fruits from Algerian cultivars, ranged from 27.55 to 211.01 mg/100 g with highest amount in Sigoise and the lowest in Dahbia cultivars. [24] The most prominent secoiridoids of olive fruits and leaves is oleuropein (responsible for bitterness), which is one of the main precursors of phenolic compounds found in olives. [25] Various processes such as hydrolysis, chemical and enzymatic activity may be responsible for degradation of oleuropein (as shown in Figure 4) to its basic components (hydroxytyrosol, glucose and elenolic acid) during the course of pre-, and post-harvest processing. [26,27] Functional bioactives, such as tocopherols, squalene, and sterols along with phenolic imparts organoleptic properties to olive oil. Pérez et al. (2019) observed that the virgin olive oils from most of the cultivars have αtocopherol but some cultivars also showed the presence of β-, and γ-forms of tocopherol with their quantities mainly cultivars dependent. [28]  In another work, it was observed that α-Amyrin and β-Amyrin were higher in the unripe fruits and diminished as ripening proceeded for Spanish cultivars, such as Picual, Arbequina and Hojiblanca. [29] Moreover, β-sitosterol is an important phytosterol which possesses broad spectrum biological activities such as anticancer, anti-diabetic, anti-inflammatory, antioxidant, antimi-crobial in combination with lipid lowering, sedative and wound healing effect. [30] Hannachi et al. (2013) found that sterolic fraction was comprised of almost 85 to 76 % of β-sitosterol in the wild olive of Tunisia. [31] These findings are in agreement with our present results in which a considerable quantity (7.65 %) of β-sitosterol is examined. Furthermore, an important  triterpene/hydrocarbon, namely, squalene, is also detected in a considerable amount as high as 4.65 %. In line with our present result, Mousavi et al. (2022) reported that the levels of squalene in olives may vary in the range of 2.75-4.35 %. [32] In human beings, squalene acts as cholesterol precursor and is known as an intermediate for biosynthesis of steroid hormones. [33] Other key components present in olive fruit are the fatty acids. Higher levels of unsaturated fatty acids are important for human health and play a significant role in the functional food and nutraceutical value of table olives. In our study linolenic acid (C18 : 3) was found in highest amount in green GOF. It has been reported by Bodoira et al. (2015) that, in olive fruit pulp, linolenic acid was the highest (23.1 %) at low maturity (green stage) compared to high maturity stage (3.32 %). [34] It is also established that the solubility of fatty acid in organic solvent gradually increases with increase in the degree of unsaturation. Since, during the current extraction ethanol was employed as extracting solvent; it might be a chance that relatively a higher content of C18 : 3 (11.37 %) has been extracted in to the extraction media as compared with linoleic acid (2.65 %) and while negligible amount of oleic acid is detected. In another study, C18 : 3 was found to be 2.54 % in Arbequina and 1.22 % in Picual olives. [35] In the present work, undecanoic acid ethyl ester is also detected in minor concentration. The presence of ethyl ester may be linked to the fermentation/degradation of olives [36] and its formation may result during sample treatment and processing.
Another compound, phytol is also delectated in significant concentration (7.11 %) and its formation may be due to degradation of chlorophyll. [37] Phenolic lipids (derivatives of mono and poly-phenol) are a diversified class of unique compounds. These compounds can integrate into erythrocytes and liposomal membranes because of their strong amphiphilic nature. [38] An important phenolic lipid, methyl-2-formyl benzoate, as detected in a considerable amount (7.79 %) in the present work, is recognized as a well-known bioactives precursor in the chemical synthesis of drugs due to its wide range of pharmacological effects such as, antiviral, antihypertensive, antifungal anticancer, antiulcer and antipsychotic. [39] The presence of this precursor of bioactive can also be linked to multiple medicinal benefits of olives.
The variations in the profile of high-value phytochemicals in the presently tested GOF compared with literature olives data may be related to the difference in the genetic makeup as well as varied morphological features based on different factors such as stage of olive fruit maturity, agro-climatic conditions of harvest, and method of extraction. Literature reports reveal that the consumption of diets containing such phenolic bioactives along with other functionally valuable compounds exert a beneficial effect on human health. [40,41] Presence of wide array of high-value bioactives in olive fruit support its vital role in promoting human health. [42,43]

Conclusions
In the present investigation an optimized ultrasound-assisted ethanolic extraction (UAEE) protocol was devised via RSM to improve the yield of bioactives extracts along with antioxidant attributes. The model p value, lack of fit F-value, pred R 2 , adj R 2 and coefficient of variation, all suggested the reliability of the designed model. Similarly, the good correspondent nature of experimental and predicted results revealed the trueness of optimized extraction conditions. This research concluded that RSM based optimal UAEE conditions such as 32.09 min extraction time, 42.68°C extraction temperature, and 79.83 % AEC were useful to maximize extract yield, TPC and DPPH radical scavenging potential of GOF extracts compared with conventional Maceration assisted extraction (MEE). Appreciable levels of phenolic bioactives mainly, vanillic acid, p-coumaric acid and hydroxy tyrosol along with other medicinally important compounds such as α and β-amyrin, β-sitosterol, squalene and α-tocopherol were detected in the tested GOF extract using GC/MS. Among the fatty acids detected, linolenic acid was dominant component followed by linoleic acid. A considerably high amount (7.79 %) of an important phenolic lipid such as methyl-2-formyl benzoate was also detected. The newly devised RSM optimized UAEE protocol in the present study might be applicable for the extraction of phytochemicals, especially, phenolic bioactives from different parts of olive as well as its application can be extended to other related fruit crops.

Sample Preparation
Gemlik olive fruits (~1 kg) were collected on 15 of August, October and December in the year 2020 (at unripen, semi-ripen and fullyripen stages (as shown in Figure 5) from Barani Agriculture Research Institute (BARI) of District Chakwal, Pothwar region of Punjab, Pakistan. Each time, samples were collected under the supervision of Dr. M. Ramzan Anser, senior Agronomist at BARI institute, Chakwal. After cleaning and washing the debris, the fruits were destoned using domestic cutter (shown in Figure 5 D) and dried under shade. The dried fruits (olives) were grinded using a grinder (Lab-scale, FM-909T, Hanil Electric Co., Seoul, Korea). Coarse powder was then filtered through a mesh sieve (1 mm-ASTM No 18) to obtain GOF powder with particle size of below 1 mm.

Effect of Ripening Stages of Olive Fruits
In olive fruits gernally there are three ripening stages namely: green stage (unripen), turning stage (semi-ripen) and black stage (fullyripen). The olive fruits of selected cultivar (at each ripening stages) were extracted for phenolic bioactives while the extraction was carried out at 40°C for 30 min taking 5g sample in 50 mL of 80 % aqueous ethanol.

Effect of the Solute Concentration
To optimize the effect of sample/solid ratio, the ratio was ranged from 5/100 to 12.5/100 (g/mL) of unripen olive using 80 % aqueousethanol keeping other variables constant (extraction temperature 40°C and extraction time 30 min).

Effect of Solvent
Water and ethanol are considered as green and eco-friendly solvents. Initially pure water, absolute ethanol, 40 % aqueousethanol and 80 % aqueous-ethanol were selected for extraction of phenolic bioactives while maintaining all other factors constant (extraction time:30 min and solid to liquid ratio: 10g/100 mL).

Effect of Temperature
For optimal extraction of phenolic compounds/components from unripen Gemlik olives, temperature was varied from 20 to 80°C, while extraction time (30°C), ethanol concentration (80 %) and sample-to-solvent ratio (10g/100 mL) were kept constant.

Effect of Time
For optimization of extraction time, extraction was performed for different duration (15, 30 45 and 60 min) keeping the other parameters constant (80 % AEC, 40°C and 10g/100 mL solute-tosolvent ratio).

Experimental Design
Full factorial design is true experimental design which controls threats to internal validity due to randomly assign subjects but it is difficult to analyze, complex and time consuming. Both central composite design (CCD) [44,45] and Box-Bhenken Design (BBD) [16,46] can be applied for extraction optimization. Although, BBD requires less time in comparison to CCD (due to a smaller number of experimental runs) and hence it is cost-effective [47] but CCD has shown slightly improved results in comparison to BBD. [48,49] That's why CCD was selected for optimization in the present experiments. After preliminary examination using single factor optimization; temperature, extraction time and aqueous ethanol concentration were selected (level of these variables are shown in Table 7) for further optimization keeping the solute to solvent ratio constant (1/ 10). RSM suggested 20 different combinations of conditions (runs) were executed for three responses. Using these set of conditions, sample was extracted (to find best conditions for optimal recovery of bioactives extract, TPC and DPPH free radical scavengers) by applying ultrasonic cleaner bath (LPAU-A10 Labtron Equipment Ltd. UK). After the extraction, residue was removed by filtration. The excess solvent from the filtrate was distilled off under reduced pressure using a rotary evaporator (Daihan Scientific, Japan) at 40°C to obtain the solvent free extract (the extract yield was expressed in mg/g dry weight). The recovered extracts were used to find TPC and DPPH radical scavenging potential. The different responses were modeled by a quadratic polynomial using Design Expert (ver. 8.0.6) software as follows: Where: b * -intercept; n-number of factors analyzed; b i -linear model coefficient; b ii -quadratic model coefficient; b ij -interactive model coefficient. While, X i and X j represent the level of independent variables, and Y is a response to be optimized. [50]

Conventional Extraction
The maceration ethanolic extraction (MEE) was performed according to the reported protocol. [51] Briefly, 10 g dried sample was soaked in 100 mL solvent (80.00 % aqueous ethanol as optimized) and stirred for 8 h using orbital shaker. After filtration solvent was evaporated under reduced pressure to get solvent free concentrated extract.

Determination of Total Phenolic Contents (TPC)
Total phenolic contents of GOF extract were calculated by slightly modifying the reported method. [52] In short, 50 μL of GOF extract (1 mg/mL solution) was mixed with 50 μL of FCR and diluted with 750 μL of distilled water. After 10 min incubation at room temperature, 150 μL of sodium carbonate (20 % w/v) was added to it. The resulting mixture was incubated in a water bath at 40°C for 20 min and then allowed to cool in an ice bath. In the end, the absorbance of the resulting mixture was taken at 755 nm using a spectrophotometer (UV-1800 spectrophotometer, Shimadzu, Kyoto, Japan). TPC were computed as Gallic Acid Equivalents (GAE) mg/g of dry extract after preparation of standard calibration curve.

DPPH Radical Scavenging Assay
The DPPH radical scavenging potential of the GOF extract was measured spectrophotometrically, using a previously reported protocol. [53] Briefly, in 50 μL of GOF extract (5.0 mg/mL), 5 mL of DPPH solution (0.004 % in methanol) was added. After 45 min incubation at 27°C, the absorbance of the resulting solution was recorded at 517 nm using a spectrophotometer. A similar procedure was repeated for the standard (using different concentrations) and blank. The radical scavenging (%) activity was calculated using the following formula: Where "D c " and "D s " represent absorbance for control and extract, respectively.

Derivatization and GC/MS analysis of the GOF Bioactive Extract
The dried optimized GOF extract was mixed with methanol (200 μL) and 20 μL of internal standard (o-methylbenzoic acid) in a 2 mL sample vial. The resulting mixture was dried by a stream of nitrogen gas. This residue was silylated by adding 50 μL of the derivatization reagent (1 % TMCS + BSTFA) and 50 μL of the pyridine followed by heating at 60°C for 30 min. The excess amount of derivatization reagent (1 % TMCS + BSTFA) was used to ensure complete derivatization. The anhydrous condition was maintained during the derivatization process. The sylated/derivatized GOF extract was diluted (1/100) with methanol and filtered by micro-filter, and 1 mL of this diluted sample was injected (split ratio 30 : 1) in to the column. The extracts bioactives analysis was carried out on GC (5977A) coupled with MS (7890A, Agilent USA). A DB-5 fused-silica capillary column with 30 m × 0.25 mm inner diameter and a 0.25 μm film thickness (J&W Scientific, CA, USA) was used for the separation purposes. High purity helium gas (99.9 %) was run as a continuous mobile phase with a flow rate of 1 mL/min. The operating temperature for injector, transfer line, and detector was maintained at 290°C while, column oven temperature (for 2 min) was programmed at 60°C and then increased by the rate of 3°C/ min to final temperature of 280°C. The mass spectrum was scanned (at a rate of 1.5 scans/s) from m/z 50 to 650. The bioactive compounds were identified by matching their retention times with those of pure standards as well as by comparing MS spectra with those available in mass spectrum library of the machine and quantified by the percent peak area of the signals for characterized compounds. [54]