Premna odorata: Seasonal Metabolic Variation in the Essential Oil Composition of Its Leaf and Verification of Its Anti-Ageing Potential via In Vitro Assays and Molecular Modelling

The metabolic variation in the essential oil composition of Premna odorata leaves obtained from different seasons was quantitatively and qualitatively determined employing GC/MS (Gas Chromatography coupled with Mass Spectrometry) and GC/FID (Gas chromatography equipped with flame ionization detector) techniques. It displayed the existence of 97 constituents accounting for 94.19%, 92.27%, 91.95% and 92.63% for POS (spring), POM (summer), POA (autumn) and POW (winter) whole essential oils. β-Caryophyllene constituting the main metabolite in the oil in the different seasons. To better visualize the differences between them, GC data were exposed to chemometric analysis. A PCA (principal component analysis) score plot revealed the closeness of POS and POW. Molecular modelling on collagenase, elastase and hyaluronidase enzymes active centres shows that different compounds existing in the essential oil of Premna odorata leaves shows binding to the active sites with variable degrees that suggested its anti-ageing potential. Palmitic acid displayed the highest fitting for both the collagenase and elastase active centres in both pH-based and rule-based ionization methods with ∆G equals −78.27 and −44.77 kcal/mol, respectively; meanwhile, heptacosane showed the highest fitting score in the hyaluronidase centre with ∆G = −43.78 kcal/mol. In vitro assays consolidates the obtained modelling studies in which essential oil shows considerable anti-elastase and anti-hyaluronidase potential as evidenced by their IC50 values being 49.3 and 37.7 μg/mL, respectively; meanwhile, the essential oil of Premna odorata leaves displayed mild anti-collagenase potential. Thus, it can be concluded that Premna odorata could serve as a promising anti-ageing naturally occurring drug that could be effectively incorporated by pharmaceutical industries in cosmetics combating ageing and skin wrinkling.


Introduction
Ageing is primarily caused by the pronounced disturbance in the balance between the regenerative and the degenerative potential of the skin causing wrinkling and thinning of its epidermis. This is subsequently accompanied by the appearance of crevices, furrows, lines and creases in the lines of facial expression in particular. Basically, ageing appears due to the extensive morphological alteration in the dermal layer triggered by repeated exposure to UV radiations and free radicles as well. This alteration is accompanied by extensive loss in collagen as well as elastin fibres with reduced circulatory perfusion [1].

Plant Material
The leaves of Premna odorata Blanco (Family Verbenaceae) were collected from Zoo botanical garden during the four seasons spring, summer, autumn and winter 2018. They were kindly authenticated and confirmed morphologically by Mrs. Terease Labib, Consultant of Plant Taxonomy at Ministry of Agriculture, El-Orman Botanical Garden and National Gene Bank, Giza, Egypt. Voucher materials representing leaves from spring, summer, autumn and winter were kept as a herbarium in Pharmacognosy Department, Faculty of Pharmacy, Ain Shams University under the codes (PHG-P-PO-426 (POS); PHG-P-PO-427(POM); PHG-P-PO-428 (POA) and PHG-P-PO-429 (POW)), respectively.

Preparation of Essential Oil Samples
The fresh leaves of Premna odorata obtained from spring, summer, autumn and winter was air dried in the lab (at temperature not exceeding 35 • C). The essential oil samples were prepared via the hydrodistillation of the air-dried leaves of Premna odorata obtained from spring, summer, autumn and winter using Clevenger-type apparatus for 6 h. Essential oils POS, POM, POA and POW were dehydrated using anhydrous Na 2 SO 4 for spring, summer, autumn and winter samples to yield 0.031, 0.025, 0.018 and 0.022% v/w of dry weight, respectively. The different oil samples were analyzed directly by GC/FID and GC/MS and then kept in dark-coloured stoppered glasses at −30 • C for further biological analyses [4].

GC/FID and GC/MS Analyses
Shimadzu GC-17A gas chromatograph (Shimadzu Corporation, Kyoto, Japan) with a DB-5 fused-bonded cap column (Phenomenex; 29 m × 0.25 mm i.d., film thickness 0.25 µm; Torrance, CA, USA) and a FID detector was used for the quantitative analysis of the samples. Shimadzu GC-2010 plus a gas chromatograph (Shimadzu Corporation, Kyoto, Japan) were employed for GC analyses and equipped with Rtx-5MS (Restek, Bellefonte, PA, USA) and a quadrupole mass spectrometer for identification of the volatile oil components. Instrumental settings were adjusted according to what was previously reported [5]. Three independent runs were used to calculate AUP (areas under the peaks) using Class GC 10 ® software (Shimadzu Corporation, Kyoto, Japan) in which the total area is considered 100%. Meanwhile, for qualitative interpretation of the chromatograms, GC solution ® software ver. 2.4 (Shimadzu Corporation, Kyoto, Japan) was used. For comparing and identifying the constituents of the different essential oil samples, Wiley Registry of Mass Spectral Data 8th edition, NIST MassSpectral Library (December 2011), and previously reported techniques were employed [6].

Chemometric Analysis
To better visualize the differences between the compositions of essential oils obtained from the leaves of Premna odorata collected at different seasons, the GC data were exposed to unsupervised pattern recognition chemometric analysis represented by principal component analysis (PCA). This was done using Unscrambler ® 9.7 (CAMO SA, Oslo, Norway) as previously described [7,8]. Additionally, a heat map was also done using GC data that was processed using Hierarchical Clustering Explorer 3.5 software (Human computer interaction laboratory, University of Maryland, College Park, MD, USA) [7,8].

In Silico Virtual Studies
In silico virtual screening was performed inside the active centres of three crucial enzymes implicated in the occurrence and progression of ageing process viz. collagenase (PDB ID: 2D1N; 2.37 Å), elastase (PDB ID: 1BRU; 2.30 Å) and hyaluronidase (PDB ID: 1FCV; 2.65 Å) using Discovery Studio 4.5 (Accelrys Inc., San Diego, CA, USA) applying the C-Docker protocol. The X-ray crystal structures of the utilized enzymes were downloaded from a protein data bank (www.pdb.org) in PDB format. Each enzyme structure was prepared using the default protocol for protein preparation of Discovery Studio 4.5 (Accelrys Inc., San Diego, CA, USA) where water molecules were removed, hydrogen atoms were added followed by the cleanliness of the protein structure from any undesirable interactions. CHARMm was chosen as the forcefield meanwhile MMFF94 was used as a method for partial charge calculation followed by minimization of the added hydrogen in about 2000 steps. The binding centre was determined based upon the reported data near the enzyme catalytic domain. All the compounds 2D structures were drawn using ChemDraw 15.0 and saved as PDB files. The structures were prepared using the default protocol for ligand preparation of Discovery Studio 4.5 (Accelrys Inc., San Diego, CA, USA) using both pH-based and rule-based ionization methods. Then the prepared compounds were docked within the active sites of the energy-minimized protein applying C-Docker protocol leaving its parameters as its default parameters. CHARMm force field was assigned and the binding energies were calculated using a distance dependent dielectric implicit solvation model for the selected docking poses. However, the free binding energies were calculated in kcal/mol using the following equation: [9,10].
∆G binding = E complex − (E protein + E ligand ) where; ∆G binding : The ligand-protein interaction binding energy, E complex : The potential energy for the complex of protein bound with the ligand, E protein: The potential energy of protein alone and E ligand : The potential energy for the ligand alone 2.6. Assessment of the Anti-Ageing Potential of Premna Odorata Leaves Essential Oil

Anti-Collagenase Evaluation
Assessment of the collagenase inhibitory activity was measured spectrophotometrically as previously described [11]. Briefly, collagenase from Clostridium histolyticum (ChC-EC.3.4.23.3) was solubilized in 50 mM Tricine buffer (pH 7.5 with 400 m MNaCl and 10 mM CaCl 2 ) to form 0.8 units/mL as an initial concentration in accordance to data given with 0.8 units/mL. The prepared enzyme in the buffer was added to the essential oil samples in a range of concentrations (1000-7.81 µg/mL) for 15 min, then 2 mM of the synthetic substrate N-[3-(2-furyl) acryloyl]-Leu-Gly-Pro-Ala (FALGPA) in Tricine buffer was added to the samples. Absorbance was continuously measured for 20 min after the immediate addition of the substrate at 335 nm by a Microplate reader. Negative control was prepared using water; meanwhile, EGCG (epigallocatechin gallate) was employed as a positive control. The IC 50 is the concentration of sample used to inhibit 50% of collagenase under the assay conditions.
The percentage of collagenase inhibition (%) = 1 − S C × 100; S: is the corrected absorbance of the samples; C: is the corrected absorbance of the control.

Anti-Elastase Evaluation
Estimation of elastase inhibitory activity was determined spectrophotometrically as previously reported [12]. Pancreatic elastase was solubilized in sterile water to prepare a stock solution of 3.33 mg/mL. N-Succinyl-Ala-Ala-Ala-p-nitroanilide was prepared by its dissolution in mM Tris-HCL buffer with pH equals 8.0 to reach 1.6 mM. The substrate was added to the samples in the range of (1000-7.81 µg/mL) after being incubated with the enzyme for 15 min. Absorbance was continuously measured for 20 min after the immediate addition of the substrate at 381-402 nm by a Microplate reader. Negative control was prepared using water; meanwhile, EGCG (epigallocatechin gallate) was employed as a positive control. The IC 50 is the concentration of sample used to inhibit 50% of elastase under the assay conditions.
The percentage of elastase inhibition (%) = 1 − S C × 100; S: is the corrected absorbance of the samples; C: is the corrected absorbance of the control.

Anti-Hyaluronidase Evaluation
Determination of hyaluronidase inhibitory potential was determined spectrophotometrically as previously reported using Type-1-S bovine hyaluronidase (Sigma Aldrich, St. Louis, MO, USA) [13]. Absorbance was measured at 585 nm by a Microplate reader. Negative control was prepared using water; meanwhile, EGCG (epigallocatechin gallate) was employed as a positive control. The IC 50 is the concentration of sample used to inhibit 50% of hyaluronidase under the assay conditions.
The percentage of elastase inhibition (%) = 1 − S C × 100; S: is the corrected absorbance of the samples; C: is the corrected absorbance of the control.

Statistical Analysis
Data is represented in the form of mean ± SD meanwhile the IC 50 values were computed employing Microsoft Excel 2007 package in which the level of significance is p < 0.05.

Chemometric Analysis
There are multiple differences between the qualitative and quantitative compositions of essential oils obtained from the leaves of Premna odorata collected at different seasons that cannot be easily detected by the naked eye. Thus, to better visualize the differences between them, the data resulted from GC analyses were exposed to chemometric analysis using unsupervised pattern recognition chemometric analysis represented by Principal component analysis (PCA). The score plots obtained from PCA were illustrated in Figure 3A in which PCA considerably differentiated the examined essential oil samples into four clusters significantly discriminated along the first component (PC1) and second component (PC2) that account for 55% and 29%, respectively representing 84% of the total variances existing along the obtained data. PCA score plot revealed the closeness between the essential oil collected of Premna odorata leaves collected in spring and that in winter evidenced by the presence of their clusters grouped together in the lower quadrant at the left

Chemometric Analysis
There are multiple differences between the qualitative and quantitative compositions of essential oils obtained from the leaves of Premna odorata collected at different seasons that cannot be easily detected by the naked eye. Thus, to better visualize the differences between them, the data resulted from GC analyses were exposed to chemometric analysis using unsupervised pattern recognition chemometric analysis represented by Principal component analysis (PCA). The score plots obtained from PCA were illustrated in Figure 3A in which PCA considerably differentiated the examined essential oil samples into four clusters significantly discriminated along the first component (PC1) and second component (PC2) that account for 55% and 29%, respectively representing 84% of the total variances existing along the obtained data. PCA score plot revealed the closeness between the essential oil collected of Premna odorata leaves collected in spring and that in winter evidenced by the presence of their clusters grouped together in the lower quadrant at the left hand side of the plot. Meanwhile, both PC1 and PC2 effectively discriminated between POS and POW, which were located in the lower quadrant at the left hand side that show negative values of both PC1 and PC2 and POA that is placed in the upper quadrant at the right hand side of the plot and displayed positive values for both components. However, PC2 could be perfectly differentiated between the clusters of POS and POW that show negative values of PC2 and that of POM that is located upper quadrant at the left hand side of the plot that shows the positive value of PC2. Furthermore, PC1 successfully differentiated between POM (negative value of PC1) and POA (positive value of PC1). The loading plot represented in Figure 3B showed that the main discriminatory markers are caryophyllene, limonene, linoleic and linolenic acid. In addition, a heatmap was constructed from GC data, the amounts decrease gradually by moving from dark blue (highest amounts), light blue, white, light red till they become undetectable as represented by small dark red squares (pixels) and this promptly gives immediate visualization of information for easier understanding of the complexity of data (Figure 4).
Biomolecules 2020, 10, x FOR PEER REVIEW 10 of 17 hand side of the plot. Meanwhile, both PC1 and PC2 effectively discriminated between POS and POW, which were located in the lower quadrant at the left hand side that show negative values of both PC1 and PC2 and POA that is placed in the upper quadrant at the right hand side of the plot and displayed positive values for both components. However, PC2 could be perfectly differentiated between the clusters of POS and POW that show negative values of PC2 and that of POM that is located upper quadrant at the left hand side of the plot that shows the positive value of PC2. Furthermore, PC1 successfully differentiated between POM (negative value of PC1) and POA (positive value of PC1). The loading plot represented in Figure 3B showed that the main discriminatory markers are caryophyllene, limonene, linoleic and linolenic acid. In addition, a heatmap was constructed from GC data, the amounts decrease gradually by moving from dark blue (highest amounts), light blue, white, light red till they become undetectable as represented by small dark red squares (pixels) and this promptly gives immediate visualization of information for easier understanding of the complexity of data ( Figure 4).

In Silico Virtual Studies
For determination of the anti-ageing potential of Premna odorata leaves essential oil, molecular modelling for its major identified components was performed inside the active pockets of three crucial enzymes implicated in the occurrence and progression of the ageing process, namely collagenase, elastase and hyaluronidase enzymes ( Table 2). Collagen represents one of the main building units of the skin and is the major constituents of the nails, hairs and connective tissue as well. It plays a pivotal role in maintaining the elasticity, flexibility and strength of the dermal tissue as well. Meanwhile, it is degraded by collagenase enzymes, so prohibition of the collagenase enzyme perfectly retards pre-collagen fibres formation with subsequent inhibition to the wrinkling process [14,15]. In addition, elastin constitutes a protein that exists in connective tissue and carries the responsibility of keeping the skin elastic that is degraded by intracellular elastase enzymes that upon elevation, either by age or continual exposure to UV-radiation, leads to ageing. In addition, hyaluronic acid performs a crucial role in keeping the elasticity, moisture and structure of the skin via facilitating the transportation of nutrients as well as waste products. It is also incorporated in fast tissue division, renewal and repair [14,15]. Loss of dermal strength and elasticity due to destruction of collagen, elastin as well as hyaluronic acid results in the appearance of wrinkle that is a characteristic feature of ageing, thus inhibition of collagenase, elastase and hyaluronidase as well as the promotion of collagen, elastin as well as hyaluronic acid production have been recently adopted as therapeutic approaches in combating ageing. Naturally occurring substances derived from the plant kingdom have been widely employed to prohibit these enzymes and stimulate collagen, elastin and hyaluronic acid formation [14,15]. Regarding the collagenase enzyme, some of the identified constituents revealed tight fitting with the active sites as evidenced by the free binding energy (∆G) in which palmitic acid displayed the highest fitting in both pH-based and rule-based ionization methods with ∆G equals −78.27 kcal/mol showing a superior binding score comparable to the positive standard EGCG (∆G = −59.27 kcal/mol). Palmitic acid is followed by heptacosane (−63.64 kcal/mol), linoleic acid (−59.31 kcal/mol), oleic acid (−57.81 kcal/mol), linolenic acid (−29.85 kcal/mol), trans-phytol (−28.44 kcal/mol), p-vinylanisole (−20.06 kcal/mol), caryophyllene oxide (−7.19 kcal/mol) and β-linalool (−3.84 kcal/mol), meanwhile others, which show positive values of ∆G, revealed unfavorable interactions ( Table 2). The tight binding of palmitic acid at the active sites can be interpreted in virtue of formation of a π-bond with Phe 252, a metal acceptor bond with the Zn 270 present at the active site in addition to the formation of multiple hydrophobic interactions and its size that fits the active site. Meanwhile, EGCG forms four hydrogen bonds with Ile 243, Phe 241, Pro 242 and Glu 223, π-π interaction with His 222, π-alkyl interaction with Leu 185, Val 219, Van der Waals interaction with Leu 184 and Pro 242 and a metal acceptor bond with the Zn 270; thus, it is clear that the latter is crucial for the interaction ( Figure 5).  Table 2). The tight binding of palmitic acid at the active sites can be interpreted in virtue of formation of a π-bond with Phe 252, a metal acceptor bond with the Zn 270 present at the active site in addition to the formation of multiple hydrophobic interactions and its size that fits the active site. Meanwhile, EGCG forms four hydrogen bonds with Ile 243, Phe 241, Pro 242 and Glu 223, π-π interaction with His 222, π-alkyl interaction with Leu 185, Val 219, Van der Waals interaction with Leu 184 and Pro 242 and a metal acceptor bond with the Zn 270; thus, it is clear that the latter is crucial for the interaction ( Figure 5). Concerning the elastase enzyme, palmitic acid also exhibited the highest fitting score in both pH-based and rule-based ionization methods with ∆G equals −44.77 kcal/mol approaching that of EGCG (∆G = −49.35 kcal/mol). This is followed by heptacosane (−40.48 kcal/mol), linoleic acid (−31.51 kcal/mol), oleic acid (−31.50 kcal/mol), trans-phytol (−16.62 kcal/mol), p-vinylanisole (−15.36 kcal/mol) and linolenic acid (−1.43 kcal/mol) ( Table 2). The strict binding of palmitic acid to the active centres can be attributed to the formation of H-bond between its carbonyl group and Arg 143, a salt bridge between the carboxylic acid group and Arg 143 in addition to the formation of multiple hydrophobic interactions and its size that fits the active site; however, EGCG forms six H-bonds with Gly 193, Cys 42, Thr 41, His 57, Ser 195 and Gly 216, Van der Waals interaction with Cys 191, Ser 195 and Ser 217, π-π interaction with His 57, π-alkyl interaction Cys 58 ( Figure 6). Concerning the elastase enzyme, palmitic acid also exhibited the highest fitting score in both pH-based and rule-based ionization methods with ∆G equals −44.77 kcal/mol approaching that of EGCG (∆G = −49.35 kcal/mol). This is followed by heptacosane (−40.48 kcal/mol), linoleic acid (−31.51 kcal/mol), oleic acid (−31.50 kcal/mol), trans-phytol (−16.62 kcal/mol), p-vinylanisole (−15.36 kcal/mol) and linolenic acid (−1.43 kcal/mol) ( Table 2). The strict binding of palmitic acid to the active centres can be attributed to the formation of H-bond between its carbonyl group and Arg 143, a salt bridge between the carboxylic acid group and Arg 143 in addition to the formation of multiple hydrophobic interactions and its size that fits the active site; however, EGCG forms six H-bonds with Gly 193, Cys 42, Thr 41, His 57, Ser 195 and Gly 216, Van der Waals interaction with Cys 191, Ser 195 and Ser 217, π-π interaction with His 57, π-alkyl interaction Cys 58 ( Figure 6). Regarding hyaluronidase, heptacosane also displayed the highest fitting score in both pH-based and rule-based ionization methods with ∆G equals −43.78 kcal/mol approaching that of EGCG (∆G = −43.70 kcal/mol). This is followed by palmitic acid (−40.85 kcal/mol), oleic acid (−31.59 kcal/mol), linoleic acid (−30.48 kcal/mol), p-vinylanisole (−11.44 kcal/mol) and trans-phytol (−4.10 kcal/mol) ( Table 2). Heptacosane forms two alkyl interactions with Arg 47 and Lys 45 in addition to the formation of multiple hydrophobic interactions and its size that fits the active site. Meanwhile, four H-bonds with Asp 305, Glu 113 and Ser 304, π-π interaction with Trp 303, Phe 46, π-δ interaction with Tyr 55 in addition to π-alkyl interaction with Pro 18 (Figure 7). Regarding hyaluronidase, heptacosane also displayed the highest fitting score in both pH-based and rule-based ionization methods with ∆G equals −43.78 kcal/mol approaching that of EGCG (∆G = −43.70 kcal/mol). This is followed by palmitic acid (−40.85 kcal/mol), oleic acid (−31.59 kcal/mol), linoleic acid (−30.48 kcal/mol), p-vinylanisole (−11.44 kcal/mol) and trans-phytol (−4.10 kcal/mol) ( Table 2). Heptacosane forms two alkyl interactions with Arg 47 and Lys 45 in addition to the formation of multiple hydrophobic interactions and its size that fits the active site. Meanwhile, four H-bonds with Asp 305, Glu 113 and Ser 304, π-π interaction with Trp 303, Phe 46, π-δ interaction with Tyr 55 in addition to π-alkyl interaction with Pro 18 (Figure 7).

In Vitro Assessment of the Anti-Ageing Potential of Premna odorata Leaves Essential Oil
Molecular modelling study shows that different compounds existing in the essential oil of Premna odorata leaves revealed promising fitting scores so the different samples were mixed in equal volumes (PO) and subjected to in vitro assays namely, anti-collagenase, anti-elastase and anti-hyaluronidase assays. The essential oil shows considerable anti-elastase and anti-hyaluronidase potential as evidenced by their IC50 values being 49.3 and 37.7 μg/mL, respectively, approaching that of EGCG, the standard anti-ageing drug, that shows IC50 values equal to 18.2 and 15.5 μg/mL for elastase and hyaluronidase, respectively. Meanwhile, the essential oil of Premna odorata leaves displayed mild anti-collagenase with IC50 value equals to 241.9 μg/mL whereas EGCG shows IC50 value equals 24.7 μg/mL ( Figure 8). This, in turn, further consolidates the results obtained from molecular modelling and supported the previously reported results that palmitic, oleic and linoleic acid exerted potent anti-ageing potential. In addition to their inhibitory potential to ageing enzymes, and moisture loss prevention, oleic acid showed asignificant antioxidant behaviour [16]. It is noteworthy to highlight that β-caryophyllene, the major constituent in the oil, has previously showed a significant anti-ageing potential in C. elegans models evidenced by increasing the longevity in its lifespan by more than 22% and was found to be attributed to its antioxidant activity where it reduced the levels of the free radicals level [17]. Thus, the oil exhibited anti-ageing potential may be attributed to the synergistic action of all metabolites as each may act with a different mechanism.

In Vitro Assessment of the Anti-Ageing Potential of Premna odorata Leaves Essential Oil
Molecular modelling study shows that different compounds existing in the essential oil of Premna odorata leaves revealed promising fitting scores so the different samples were mixed in equal volumes (PO) and subjected to in vitro assays namely, anti-collagenase, anti-elastase and anti-hyaluronidase assays. The essential oil shows considerable anti-elastase and anti-hyaluronidase potential as evidenced by their IC 50 values being 49.3 and 37.7 µg/mL, respectively, approaching that of EGCG, the standard anti-ageing drug, that shows IC 50 values equal to 18.2 and 15.5 µg/mL for elastase and hyaluronidase, respectively. Meanwhile, the essential oil of Premna odorata leaves displayed mild anti-collagenase with IC 50 value equals to 241.9 µg/mL whereas EGCG shows IC 50 value equals 24.7 µg/mL (Figure 8). This, in turn, further consolidates the results obtained from molecular modelling and supported the previously reported results that palmitic, oleic and linoleic acid exerted potent anti-ageing potential. In addition to their inhibitory potential to ageing enzymes, and moisture loss prevention, oleic acid showed asignificant antioxidant behaviour [16]. It is noteworthy to highlight that β-caryophyllene, the major constituent in the oil, has previously showed a significant anti-ageing potential in C. elegans models evidenced by increasing the longevity in its lifespan by more than 22% and was found to be attributed to its antioxidant activity where it reduced the levels of the free radicals level [17]. Thus, the oil exhibited anti-ageing potential may be attributed to the synergistic action of all metabolites as each may act with a different mechanism.

Conclusion
The essential oil composition of Premna odorata leaves obtained from different seasons showed quantitatively and qualitatively metabolic variation as determined employing GC/MS as well as GC/FID techniques. The picture of this metabolic variation was clear when GC data was coupled with PCA as an unsupervised technique of chemometric analysis. Furthermore, molecular modelling on collagenase, elastase and hyaluronidase enzymes active centres shows that different compounds existing in the essential oil of Premna odorata leaves shows binding to the active sites with variable degrees that suggested its anti-ageing potential. This was further consolidated by in vitro assays that confirmed the obtained results. Thus, it can be concluded that Premna odorata could serve as a promising anti-ageing naturally occurring drug that could be effectively incorporated by pharmaceutical industries in cosmetics combating ageing and skin wrinkling. Further studies are recommended to be done to make comparative studies between the anti-ageing potential of each of the oils sample alone collected in each season relative to the activity of their mixture.

Conclusions
The essential oil composition of Premna odorata leaves obtained from different seasons showed quantitatively and qualitatively metabolic variation as determined employing GC/MS as well as GC/FID techniques. The picture of this metabolic variation was clear when GC data was coupled with PCA as an unsupervised technique of chemometric analysis. Furthermore, molecular modelling on collagenase, elastase and hyaluronidase enzymes active centres shows that different compounds existing in the essential oil of Premna odorata leaves shows binding to the active sites with variable degrees that suggested its anti-ageing potential. This was further consolidated by in vitro assays that confirmed the obtained results. Thus, it can be concluded that Premna odorata could serve as a promising anti-ageing naturally occurring drug that could be effectively incorporated by pharmaceutical industries in cosmetics combating ageing and skin wrinkling. Further studies are recommended to be done to make comparative studies between the anti-ageing potential of each of the oils sample alone collected in each season relative to the activity of their mixture.