Chemical Diversity and Potential Target Network of Woody Peony Flower Essential Oil from Eleven Representative Cultivars (Paeonia × suffruticosa Andr.)

Woody peony (Paeonia × suffruticosa Andr.) has many cultivars with genetic variances. The flower essential oil is valued in cosmetics and fragrances. This study was to investigate the chemical diversity of essential oils of eleven representative cultivars and their potential target network. Hydro-distillation afforded yields of 0.11–0.25%. Essential oils were analyzed by GC-MS and GC-FID which identified 105 compounds. Three clusters emerged from multivariate analysis, representative of phloroglucinol trimethyl ether (‘Caihui’), citronellol (‘Jingyu’, ‘Zhaofen’ and ‘Baiyuan Zhenghui’) and mixed (the rest of the cultivars) chemotypes. ‘Zhaofen’ and ‘Jingyu’ also exhibited low levels of other rose-related compounds. The main components were subjected to a target network approach. Drug-likeness screening gave 20 compounds with predictive blood–brain barrier permeation. Compound target network identified six key compounds, namely nerol, citronellol, geraniol, geranic acid, cis-3-hexen-1-ol and 1-hexanol. Top enriched terms in GO, KEGG and DisGeNET were mostly related to the central nervous system (CNS). Protein—protein interactions revealed a core network of 14 targets, 11 of which were CNS-related (targets for antidepressants, analgesics, antipsychotics, anti-Alzheimer’s and anti-Parkinson’s agents). This work provides useful information on the production of woody peony essential oils with specific chemotypes and reveals their potential importance in aromatherapy for alternative treatment of CNS disorders.


Introduction
Woody peony is a shrub in section Moutan DC., genus Paeonia L., family Paeoniaceae (synonymously subfamily Paeonioideae in family Ranunculaceae). The cultivated species of the woody peony mainly refers to Paeonia × suffruticosa Andr. The plant height ranges from 0.5 to 2 m. It has alternate leaves, with a mostly bi-ternate compound leaf. The flower is solitary, terminal and double. The petal is obovate with various colors such as white, pink, red, purplish red, purple, dark purple or green. The flowering period is from April to May, and the fruiting period is from August to September. This species has a long history of cultivation in China and is planted countrywide, especially in the Central Plains region. The number of cultivars is up to hundreds of, or nearly one thousand, cultivars. Its flower is known as the national flower of China. It is also cultivated worldwide, especially in Japan, France, Britain and America [1,2].
The flower of the woody peony has long been used in traditional Chinese medicine, mainly for the treatment of irregular menstruation and menstrual abdominal pain [3]. The essential oil of the woody peony flower is especially valuable, which is used in cosmetics and fragrances. Zhang et al. [4] analyzed P. × suffruticosa petal essential oil obtained by and fragrances. Zhang et al. [4] analyzed P. × suffruticosa petal essential oil obtained by microwave-assisted steam distillation, and found it contained tetracosane, hexacosane and docosane, and low levels of α-terpineol, geraniol and linalool. In a study using GC × GC coupled with TOF-MS, the petal essential oils of two woody peony cultivars were characterized by alcohols, aldehydes, alkanes, terpenes and acids [5]. Han and Bhat [6] found that P. × suffruticosa flower bud essential oil contained high levels of alkanes and low levels of alcohols, ketones and benzenoids, which showed an antimicrobial effect against food-borne bacteria. In another study, supercritical extraction from the petals of eleven P. × suffruticosa cultivars afforded (Z)-5-dodecenyl acetate, nonadecane, ethyl linoleate, (Z)-5-nonadecene, heneicosane and linalool oxide as the main components with varied contents. Moreover, the essential oils of some cultivars showed antioxidant activities [7]. Our previous studies mainly focused on the flower hydrolate extracts of some P. × suffruticosa cultivars, which revealed the predominance of benzene ethanol, phloroglucinol trimethyl ether, citronellol + geraniol or a balance of these types of compounds in different cultivars [8,9].
The chemical diversity of plant essential oils depends on geographical or ecological factors [10], genotypes [11,12], harvest, post-harvest and extraction methods [13,14]. The genotype is the major factor that determines the chemical type. The chemical type further determines specific pharmacological effects and aromatherapy applications. Active essential oil compounds work through interactions with related targets in the human body. Moreover, there might be a network of targets for the complex mixture of compounds such as essential oils [15][16][17].
Up to now, studies on the chemical diversity of woody peony flower essential oils, especially those utilizing multivariate analyses, are rather limited. Given that P. × suffruticosa has many cultivars with genetic variances, relevant studies are needed. Moreover, there are no reports on the potential target network of the woody peony flower essential oils. The network study of target proteins is important for giving insight into the in vivo mechanisms of essential oils. Therefore, the objectives of this work were to investigate the chemical diversity of the flower essential oils of eleven representative woody peony cultivars, and to further reveal the potential target network of the essential oil components.

Hydro-Distillation Yields of the Essential Oils of Woody Peony Flowers
The flowers of the eleven woody peony cultivars are shown in Figure 1. The flower colors varied considerably. Notably, the 'Jingyu' flowers presented white, and 'Wujin Yaohui' presented dark purplish red. 'Erqiao' presented purplish red and pink in one flower, or in some other cases, in one plant. Hydro-distillation of the fresh flowers afforded essential oil yields of 0.11-0.25% (w/w, on dry basis) from the eleven cultivars (Table 1). Results of analysis of variance (ANOVA) showed that the essential oil yield was influenced by the factor of woody peony Hydro-distillation of the fresh flowers afforded essential oil yields of 0.11-0.25% (w/w, on dry basis) from the eleven cultivars (Table 1). Results of analysis of variance (ANOVA) showed that the essential oil yield was influenced by the factor of woody peony cultivar (p < 0.05). The post-hoc test which utilized Tukey's method for multiple comparisons suggested that the yield values basically fell into two groups overlapping partially. The yield from 'Taohong Feicui' belonging to group A was relatively high, whereas those from 'Jingyu', 'Wujin Yaohui', 'Zhaofen', and 'Fugui Mantang' belonging to group B was relatively low.

Comparative Chemical Composition of the Flower Essential Oils
Typical gas chromatograms of the essential oils of the eleven cultivars are given in Figure 2. A total of 105 compounds were identified, which accounted for 98.7-99.7% of total composition (Table 2). For individual cultivars, the compound numbers varied from 47 ('Baiyuan Zhenghui', 'Zhaofen' and 'Manjiang Hong') to 63 ('Jingyu'). The essential oil components belong to alkanes, alkenes, terpenes, aliphatic alcohols, aliphatic aldehydes, benzoids, other oxygenated non-terpenes, terpene alcohols and other oxygenated terpenes ( Figure 3).
The alkanes included tricosane, nonadecane, pentadecane, heptadecane, pentacosane and heneicosane, which were present in all the essential oils with varied percentages. Particularly, 'Erqiao', 'Manjiang Hong' and 'Lan Baoshi' essential oils contained alkanes at 65.6-73.1%. Interestingly, almost all of the major alkanes possessed an odd number of carbons, whereas alkanes having an even number of carbons were minor components (<2.5%). Branched alkanes were also found in the present study. Interestingly, the methyl was almost always attached to position 2 of the n-alkane main chain with the even carbon numbers (2-methylhexadecane, for example), and position 3 of the n-alkane main chain with the odd carbon numbers (3-methyltricosane, for example).
Terpene hydrocarbons were basically at low percentages. Only 'Baiyuan Zhenghui' essential oil presented a medium content (18.2%), which was mainly attributed to germacrene D (10.5%) and trans-β-ocimene (6.7%). In contrast, the percentages of germacrene D and trans-β-ocimene were much lower or not detected in other cultivars. In a recent report, the essential oil of Pogostemon plectranthoides from India presents chemotype of trans-β-ocimene-germacrene D-trans-β-guaiene [20]. High levels of germacrene D are usually accompanied by cadinane and muurolane sesquiterpenoids [21]. Interestingly, δ-cadinene and τ-muurolol were uniquely present in 'Baiyuan Zhenghui' essential oil.  Table 2. Their identification was based on mass spectra and retention indices.  Table 2. Their identification was based on mass spectra and retention indices.       [8,9]; c Compounds identified by mass spectra and retention indices; d Code for cultivar name, see Table 1. Percentages of individual components are presented as mean value of triplicate samples (RSD basically below 10% for major components); tr, trace, percentages lower than 0.05%; -, not detected; * Compounds tentatively identified.  [8,9]; c Compounds identified by mass spectra and retention indices; d Code for cultivar name, see Table 1. Percentages of individual components are presented as mean value of triplicate samples (RSD basically below 10% for major components); tr, trace, percentages lower than 0.05%; -, not detected; * Compounds tentatively identified.

Multivariate Analysis Results
The hierarchical cluster analysis (HCA) dendrogram clearly assigned the essential oils into three clusters ( Figure 4). Cluster I only consisted of 'Caihui', which was distinct from others. Cluster II consisted of 'Baiyuan Zhenghui', 'Jingyu' and 'Zhaofen'. Cluster III consisted of the rest of the cultivars.   Table 1.
As for principal component analysis (PCA) (Figure 5), the chemical variance was presented with the first two principal components, which in sum explained 74.1% of the total variance (PC1 39.8%, PC2 34.3%). PCA was generally consistent with HCA in the sample grouping. It gave further information on the correlation of each cluster with corresponding characteristic compounds. Cluster I consisting of 'Caihui' was located at the lower left corner. It was highly correlated with phloroglucinol trimethyl ether. Cluster II consisting of 'Jingyu', 'Zhaofen' and 'Baiyuan Zhenghui' was located at the upper left corner. This cluster was correlated with citronellol. Interestingly, 'Zhaofen' and 'Jingyu' exhibited not only high levels of citronellol but also low levels of other rose-related aroma compounds (citronellyl esters, cis-rose oxide and trans-rose oxide). Therefore, they might serve as alternatives to rose-type essential oils. Cluster III consisting of the rest of cultivars was somewhat diverse. 'Fugui Mantang' located at the upper right corner was mainly correlated with geraniol, hydroquinone dimethyl ether and trans-8-heptadecene. 'Lan Baoshi' and 'Taohong Feicui' were correlated with phloroglucinol trimethyl ether, nonadecane and nerol. Others were correlated with pentadecane, 6,9-heptadecadiene, nonadecane and pentadecane.
Molecules 2022, 27, x FOR PEER REVIEW 9 of 21 As for principal component analysis (PCA) (Figure 5), the chemical variance was presented with the first two principal components, which in sum explained 74.1% of the total variance (PC1 39.8%, PC2 34.3%). PCA was generally consistent with HCA in the sample grouping. It gave further information on the correlation of each cluster with corresponding characteristic compounds. Cluster I consisting of 'Caihui' was located at the lower left corner. It was highly correlated with phloroglucinol trimethyl ether. Cluster II consisting of 'Jingyu', 'Zhaofen' and 'Baiyuan Zhenghui' was located at the upper left corner. This cluster was correlated with citronellol. Interestingly, 'Zhaofen' and 'Jingyu' exhibited not only high levels of citronellol but also low levels of other rose-related aroma compounds (citronellyl esters, cis-rose oxide and trans-rose oxide). Therefore, they might serve as alternatives to rose-type essential oils. Cluster III consisting of the rest of cultivars was somewhat diverse. 'Fugui Mantang' located at the upper right corner was mainly correlated with geraniol, hydroquinone dimethyl ether and trans-8-heptadecene. 'Lan Baoshi' and 'Taohong Feicui' were correlated with phloroglucinol trimethyl ether, nonadecane and nerol. Others were correlated with pentadecane, 6,9-heptadecadiene, nonadecane and pentadecane. Figure 5. Score and loading plots from PCA of flower essential oils of the eleven woody peony cultivars (Paeonia × suffruticosa Andr.). Plots are presented along the first two principal components (74.1% of total variance). For codes of cultivar names, see Table 1.
Target prediction for the 20 compounds afforded 190 potential targets (Table S2 in Supplementary Materials). The compound target network consisted of 210 nodes (com-pounds and targets) and 420 edges (compound target connections) ( Figure 6). The top ranked compounds were nerol, citronellol, geraniol, geranic acid, cis-3-hexen-1-ol and 1-hexanol, which were the most important in the network (Table 3). They also constituted the major components of the essential oils. Target prediction for the 20 compounds afforded 190 potential targets (Table S2 in Supplementary Materials). The compound target network consisted of 210 nodes (compounds and targets) and 420 edges (compound target connections) ( Figure 6). The top ranked compounds were nerol, citronellol, geraniol, geranic acid, cis-3-hexen-1-ol and 1hexanol, which were the most important in the network (Table 3). They also constituted the major components of the essential oils.  Table  2). Targets are represented by pink circles noted with standard gene names. cis-3-Hexen-1-ol 27 6 1-Hexanol 27 15 trans-β-Ocimene 21 35 Neral 21 47 Phloroglucinol trimethyl ether 21 17 cis-Linalool oxide furan 20 18 trans-Linalool oxide furan 20 26 Hydroquinone dimethyl ether  20  20  Nonanal  19  21  Linalool  19  37 Geranial 19 Figure 6. Compound target network of 20 essential oil compounds of woody peony flowers and 190 potential targets. Compounds are represented by green diamonds noted with numbers (see Table 2). Targets are represented by pink circles noted with standard gene names. cis-rose oxide 13 24 Benzeneethanol 9 a No., numbering of essential oil compounds, see Table 2; b Degree value of the compound in the compound target network as demonstrated in Figure 6.
Enrichment analyses included GO, KEGG and DisGeNET. GO molecular functions were mainly enriched in oxidoreductase and neurotransmitter receptor activities. GO cellular components were enriched in synaptic membrane and transmembrane transporter complex. GO biological processes were enriched in the regulation of hormone levels, synaptic signaling and cellular response to organic cyclic compounds (Figure 7). KEGG pathways were predominantly enriched in neuroactive ligand-receptor interaction, the log 10 P (−32.08) of which was much lower than others (Figure 8). Others included drug metabolism, nitrogen metabolism and steroid hormone biosynthesis. DisGeNET diseases were mainly enriched by chronic alcoholic intoxication, memory impairment and addictive behavior (Figure 9). cis-rose oxide 13 24 Benzeneethanol 9 a No., numbering of essential oil compounds, see Table 2; b Degree value of the compound in the compound target network as demonstrated in Figure 6.
Enrichment analyses included GO, KEGG and DisGeNET. GO molecular functions were mainly enriched in oxidoreductase and neurotransmitter receptor activities. GO cellular components were enriched in synaptic membrane and transmembrane transporter complex. GO biological processes were enriched in the regulation of hormone levels, synaptic signaling and cellular response to organic cyclic compounds (Figure 7). KEGG pathways were predominantly enriched in neuroactive ligand-receptor interaction, the log10P (−32.08) of which was much lower than others (Figure 8). Others included drug metabolism, nitrogen metabolism and steroid hormone biosynthesis. DisGeNET diseases were mainly enriched by chronic alcoholic intoxication, memory impairment and addictive behavior ( Figure 9).    . Enrichment analysis on KEGG pathways that demonstrates the represen riched terms. Count, the number of input genes that fall into this term; -Log P, the m in log base 10; InList/InTerm ratio, the ratio of the gene count in the list that hit the count of the total genome of this term. Figure 9. Enrichment analysis on DisGeNET that demonstrates the representative diseases. Count, the number of input genes that fall into this term; -Log P, the min log base 10; InList/InTerm ratio, the ratio of the gene count in the list that hit the t gene count of this term. The protein-protein interaction (PPI) network of the 190 targets identified 183 connected nodes (targets) and 992 edges (interactions) (Figure 10a; Table S3 in Supplementary Materials). The first filtering was applied to retain nodes with degree, closeness centrality and betweenness centrality, all above average (10.84, 0.3669 and 9.842 × 10 −3 , respectively), resulting in a subnetwork of 41 targets (Figure 10b). Topological parameters of this subnetwork were recalculated and the second filtering was applied to retain nodes with new degree, closeness centrality and betweenness centrality, all above average (10.83, 0.5435 and 2.208 × 10 −2 , respectively), which afforded 14 core targets (Figure 10c).

Discussion
The eleven cultivars of the woody peony (P. × suffruticosa) investigated in this study were representative ones that are widely cultivated in the Central Plains of China, especially the Luoyang district. They belong to different color series of petals due to varied contents of anthocyanins [24]. In the present study, they were grown under similar ecological and climatic conditions. Moreover, the harvest and extraction conditions were almost constantly controlled. Therefore, the variances in the essential oil yields were mainly attributed to genotypes. On the other hand, the yields obtained herein were lower than that of the dried flower buds of P. × suffruticosa (0.79%, cultivar not specified) [6]. They were also lower than the reported yields achieved by supercritical CO 2 extraction from the dried petals of eleven P. × suffruticosa cultivars (0.81-1.09%, including 'Jingyu' 0.93% and 'Erqiao' 0.95%, which were investigated in the present study as well) [7]. Meanwhile, they were comparable to or lower than those of the flower essential oils of eight other P. × suffruticosa cultivars investigated in our previous study (0.28-0.93%) [8]. In contrast, they were higher than the reported yields of the fresh petals of two woody peony cultivars (0.09% and 0.10%, possibly on fresh basis) [5] and dried petals of P. × suffruticosa (0.66 mg/g, cultivar not specified) [4]. The yield variances between the present study and literature data can be attributed to genotypes, ecological and climatic conditions [10,12], harvest or postharvest factors, and, more importantly, extraction methods [13,14]. Indeed, application of innovative techniques improves the extraction efficiency in terms of time and energy consumption, as well as production yield and quality of essential oils [4,14,23]. Further studies are needed for the evaluation and comparison of green and innovative techniques with the same woody peony flower materials.
As for the multivariate analysis of woody peony flower essential oils, literature data are scarce. Our previous study focused on hydrolate extracts of woody peony flowers. The PCA results revealed that the cultivar 'Linghua Zhanlu' was most distinct among the ten cultivars, due to the occurrence of geraniol, 6,9-heptadecadiene, 2-heptanol, pentadecane and cis-3-nonen-1-ol [9]. In another study involving eight other cultivars, the clustering of hydrolate samples was somewhat not typical [8]. Zhang et al. [7] reported four clusters from supercritical extracted petal essential oils of eleven P. × suffruticosa cultivars and three wild Paeonia species, which were represented by ethyl linoleate, (Z)-5-dodecenyl acetate, octadecanal and linalool oxide, respectively. In the present study, three clusters representative of phloroglucinol trimethyl ether, citronellol and mixed chemotypes emerged from the HCA and PCA, which were different from literature data. Even for the two cultivars 'Jingyu' and 'Erqiao', which were also reported on in the literature [7], their chemotypes in the present study were different. It is reported that the chemical type variances of essential oils are due to genotypes [11,12], geographical and climatic conditions [10], harvest and postharvest factors and extraction methods [13,14].
The 20 compounds that survived drug-likeness screening should be stressed in terms of potential activities of woody peony essential oils, particularly the six key compounds (nerol, citronellol, geraniol, geranic acid, cis-3-hexen-1-ol and 1-hexanol) summarized from the compound target network. They were mostly oxygenated compounds, which are recognized to be more important than hydrocarbons [8,25]. Literature reports suggest central nervous system (CNS) activities for most of the 20 compounds. Nerol possesses an anxiolytic effect in mice [26]. Geraniol shows neuroprotective effects against zinc oxide nanoparticles in rats [27]. Citronellol possesses anti-nociceptive, anti-anxiety and anticonvulsant effects with low toxicity [28]. Geranic acid is the sequential oxidative product of geraniol and geranial, which is less odor-active but acts as enhancer for geraniolrelated oxygenated terpenes [29]. The cis-3-hexen-1-ol and 1-hexanol possess fresh green leaf aroma [30]. A mixture of cis-3-hexen-1-ol and trans-2-hexenal attenuates behavioral and stress responses induced by olfactory and noxious stimuli in rats [31]. In addition, 1-hexanol, as well as other short chain alcohols, acts as a CNS depressant through the inhibition of NMDA (N-methyl-D-aspartate) gated currents [32]. Besides the six key compounds, others of the 20 compounds should not be disregarded. Linalool, as well as linalool abundant Lavandula angustifolia essential oil, counteracts social aversion induced by social defeat [33]. Linalool oxide furan has an anxiolytic effect upon inhalation [34]. Citronellyl acetate possesses an anti-nociceptive effect in mice, possibly involving TRPV1, TRPM8 and ASIC [35]. Rose oxide has an antidepressant activity, probably through the serotonergic pathway that involves the HTR1A receptor [36]. Citral, the mixture of geranial and neral, presents sedative and motor relaxant effects in mice [37]. In a recent report, citral exhibits anticonvulsant effect in zebrafish [38]. Phloroglucinol trimethyl ether, as well as similar oxygenated benzoids, constitutes typical flower volatiles of some varieties of Rosa odorata and R. chinensis. It presents a sedative effect and is used as perfume ingredients [39]. Benzeneethanol, a benzoid alcohol with pleasant aroma characteristic of rose flowers, has a sedative activity via inhalation [40]. The 4,7-dimethylbenzofuran also occurs in the volatiles of woody peony seed oil, which possesses aromas of grass, slight bitterness and fragrance [41].
Further enrichment analysis of the potential targets clearly indicated that the top enriched terms in GO, KEGG and DisGeNET were mostly related to the CNS, particularly neuroactive ligand-receptor interaction in KEGG pathways, and chronic alcoholic intoxication, memory impairment and addictive behavior in DisGeNET diseases. More importantly, PPI analysis indicated that 11 of the 14 core targets were CNS-related. Specifically, SLC6A4, HTR1A and MAOA are targets for antidepressants. CNR1 and OPRM1 are targets for analgesics. DRD2 and HTR2A are targets for antipsychotics. ACHE is the target for treatment of Alzheimer's disease. CHRNA4, MAOB and CHRM2 are targets for the treatment of Parkinson's disease [42,43]. Indeed, all the 20 compounds that survived drug-likeness screening showed predictive BBB permeation. Besides, when essential oil compounds pass through the BBB, they possibly affect the CNS [44]. Further, some essential oils and their main constituents are active on the CNS [33]. As for the other three core targets, NR3C1 and PTGS2 are targets for steroidal and non-steroidal anti-inflammatory agents, respectively. EGFR is the target for treatment of non-small-cell lung cancer with gene mutants [42,43].
The findings that came from this study are important and informative for developing the practical applications of the woody peony flower essential oils. The chemotype identification provides useful information on their specific uses in flavors, fragrances and cosmetics. For example, the 'Zhaofen' and 'Jingyu' essential oils may serve as alternatives to rose-type essential oils, due to high levels of citronellol as well as low levels of other rose-related aroma compounds. This study also gives instructions on the selection of specific chemo-cultivars of woody peony for cultivation and essential oil production. Moreover, the results of target network analysis suggest that the woody peony essential oils have a potential importance in aromatherapy, especially for alternative treatment of CNS disorders. Importantly, natural fragrant compounds possess relatively low neurotoxicity and are basically safe [44]. Besides, treatment for CNS disorders nowadays has more or less side effects [28]. Therefore, woody peony flower essential oils can be developed for aromatherapy applications. Nevertheless, further studies are needed on their in vivo effects and specific mechanisms.  Table 1). Authentication of plant materials was performed by Dr. Gaoming Lei, Department of Pharmaceutical Sciences, Henan University of Science and Technology. Voucher specimens numbering PS190412-01~PS190412-11 are deposited at Herbarium of Department of Pharmaceutical Sciences, Henan University of Science and Technology. Analytical standard of C7-C30 saturated alkanes with individual component of 1000 µg/mL in hexane (ID 49451-U, lot LRAC0353) was purchased from Sigma-Aldrich, Laramie, WY, USA. Purified water was produced by Wahaha Group Co., Ltd., Hangzhou, China. The n-hexane, sodium chloride, anhydrous sodium sulfate and other chemical reagents were of analytical grade.

Hydro-Distillation
Fresh flowers of woody peony were hydro-distilled in a Clevenger-type apparatus (Changsheng Apparatus for Science and Education, Hangzhou, China). The essential oil was trapped in n-hexane on top of a water layer inside the receiver of the apparatus. The hydro-distillation was performed in triplicate. The organic layer was collected and dried with anhydrous sodium sulfate. It was evaporated to afford the essential oil which was stored in 5 • C prior to chromatographic analysis. The yield was measured as the mass percent of the essential oil relative to the dried basis of fresh flowers. The water content of fresh flowers which was determined through distillation with toluene was used for calculation.

GC-MS and GC-FID Analysis
GC-MS was performed on 6890N GC with 5975 inert MSD (Agilent Technologies, Santa Clara, CA, USA) using HP-5ms column (length of 30 m, inner diameter of 0.25 mm, film thickness of 0.25 µm, Agilent Technologies, Santa Clara, CA, USA). The essential oil was diluted in hexane, 1.0 µL of which was injected to the chromatograph. The split mode was selected and split ratio was 20:1. The injector temperature was 250 • C. The carrier gas was helium and column flow was set at 1.0 mL/min. The initial temperature of the column was set at 50 • C. After injection, it was immediately raised to 200 • C at 3 • C/min, further to 240 • C at 10 • C/min, and kept thermostatically for 15 min. For the essential oil samples, there was a post run set at 250 • C for 2 min. For the C7-C30 saturated alkanes standard, after being kept at 240 • C for 15 min, the column temperature was further raised to 250 • C at 10 • C/min and then kept at 250 • C for 10 min. The MSD transfer line was set at 280 • C. Full scan mode was used at the range of 29-400 amu. The MS quadrupole was set at 150 • C and the ion source was 230 • C. The ion energy was 69.9 eV. GC-FID was performed on 6890N GC with FID (Agilent Technologies, Santa Clara, CA, USA) using HP-5ms column (30 m × 0.25 mm id, film thickness 0.25 µm, Agilent Technologies, Santa Clara, CA, USA). Analytical conditions were as follows: injection volume of 1.0 µL, carrier gas of helium, split mode, split ratio of 20:1, injector temperature of 250 • C, constant flow mode, column flow of 1.0 mL/min, FID temperature of 280 • C. The temperature program of the column was the same as that of GC-MS.
The essential oil components were identified taking into account mass spectra and retention indices. Experimental mass spectrum of each peak was searched against the reference mass spectra database NIST05a. The experimental retention index of each peak determined with C7-C30 saturated alkanes standard was compared with literature values [8,9,18,19]. The percentage of the individual components was calculated using the method of peak area integration based on FID response.

Multivariate Analysis
The total essential oil composition from the eleven woody peony cultivars was subjected to multivariate analysis of HCA and PCA. The data for analysis contained percentages of individual components of all the eleven cultivars. For HCA, Ward linkage and squared Euclidean distance was used. For PCA, the covariance matrix was used. The analysis was performed using R software [45].
Enrichment analysis on GO [55,56], KEGG [57] and DisGeNET [58] was performed on Metascape (https://metascape.org/gp/index.html#/main/step1, accessed on 18 June 2021) [59]. Parameters were: input as H. sapiens, minimum overlap 3, p-value cutoff 0.01 and minimum enrichment factor 1.5. The list of potential targets was also submitted to STRING (https://cn.string-db.org/, accessed on 4 June 2021) [60] to evaluate the PPI network. Parameters were: type of full network, active interaction sources as text mining, experiments, databases, co-expression, neighborhood, gene fusion and co-occurrence, minimum required interaction score as medium confidence at 0.40. The PPI network of connected nodes (the disconnected nodes were not considered in this step) was visualized and analyzed by Cytoscape software [54]. To identify the core targets, the entire network was refined in two consecutive steps, each of which involved retaining nodes with degree, closeness centrality and betweenness centrality all above average.

Other Statistics
Other statistics were calculated with R software [45]. ANOVA was performed to evaluate the statistical significance of data variation. Post-hoc test with Tukey's method was used for multiple comparisons. The significance level was set as 0.05 unless otherwise noted.

Conclusions
Eleven representative woody peony cultivars (Paeonia × suffruticosa Andr.) were investigated for the chemical diversity of flower essential oils and potential target network of the components. The chemical compositions of the essential oils exhibited variances, suggesting that they were affected by the genotypes. HCA and PCA analyses gave three distinct clusters representative of phloroglucinol trimethyl ether ('Caihui'), citronellol ('Jingyu', 'Zhaofen' and 'Baiyuan Zhenghui') and mixed (the rest of the cultivars) chemotypes. Interestingly, 'Zhaofen' and 'Jingyu' exhibited not only high levels of citronellol but also low levels of other rose-related aroma compounds, which allowed them to serve as alternatives to rose-type essential oils. Drug-likeness screening gave 20 essential oil compounds. Interestingly, they all exhibited predictive BBB permeation. The compound target network demonstrated six key compounds, namely, nerol, citronellol, geraniol, geranic acid, cis-3-hexen-1-ol and 1-hexanol, which should be especially stressed in terms of the potential activities of the woody peony essential oils. Enrichment analysis on GO, KEGG and DisGeNET indicated that the top enriched terms were mostly related to CNS. PPI analysis resulted in 14 core targets, 11 of which were CNS-related, including those for antidepressants, analgesics, antipsychotics, and anti-Alzheimer's and anti-Parkinson's agents. The findings that came from this work provide useful information for the selection of the specific chemo-cultivars of woody peony for cultivation and essential oil production. They also demonstrate the potential importance of woody peony flower essential oils in aromatherapy, especially for the alternative treatment of CNS disorders. Further studies are needed on their in vivo effects and specific mechanisms.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules27092829/s1, Table S1: Results of drug-likeness screening for the 45 essential oil compounds; Table S2: The list of the 190 predictive potential targets for the 20 essential oil compounds; Table S3: Main topological parameters for the total network of protein-protein interactions.