Introduction

Floral fragrance compounds are secondary metabolites released by plant flowers (Fan et al. 2019). They mainly consist of abundant low molecular weight volatiles, including terpenes, benzenoid aromatics, and fatty acid derivatives (Dudareva et al. 2013; Shi et al. 2019). The ubiquity and significance of floral volatile organic compounds (VOCs) has been vocalized by phytochemists and entomologist (Pichersky and Gershenzon 2002; Dudareva et al. 2013; Yang et al. 2018). Undoubtedly, the diversity of flora VOCs suggests that plant flowers are subtle communicators interacting with their environment. Apart from attracting specific pollinators to ensure reproductive success (Raguso 2008; Schiestl 2015; Borghi et al. 2017), many flower VOCs can serve protective functions against colonizing microorganisms (Huang et al. 2011; Ortega et al. 2016; Shi et al. 2019) and florivores (Junker et al. 2011). Moreover, VOCs emitted from pollinated flowers can deter pollinators and direct them to as yet unpollinated counterparts (Schiestl and Ayasse 2001).

Generally, bee activity, though recorded occasionally in field, is not regarded as the main polling momentum for olive (Olea europaea L.) trees (Morettini 1972; Griggs et al. 1975). However, recent studies indicate that the honey bee is the active and efficient pollen collector of olive inflorescence in both wild and orchard (Canale and Loni 2010; Aronne et al. 2012; Giovanetti 2018; Carlos et al. 2019). Yet, as an important ornamental trait (Shi et al. 2019), flower scent is rarely reported in anemophilous species. So, do the volatile metabolites in olive flowers include pollinator attractants? The literatures concerning composition, biosynthesis and function of olive VOCs are many, but they focus mostly on olive oils and few on leaves (Kiritsakis 1988; Campeol et al. 2001; Flamin et al. 2003). Thus, no phytochemical information is available regarding the compounds of olive inflorescence scent; and the question above still remains unanswered.

As a star tree species in Mediterranean countries, olive was first introduced and propagated by Chinese government in large-scale at 1964 (Xu and Han 1965; Xu 1981). Now, the prosperous olive industries can be seen in Gansu, Sichuan, Yunnan and Shaanxi provinces (Rao et al. 2019). A few domestic olive varieties (O. europaea L. cv. Ezhi-8, O. europaea L. cv. Chenggu-32, etc.) were bred during the half century domestication, and ideally, Chinese olive varieties are less sensitive to abiotic and biotic stresses. Previous comparative studies between domestic and foreign olive cultivar mainly focused on phenology, fruit and oil quality, etc. (Kong et al. 2016; Zhang et al. 2016; Han et al. 2017). The flower fragrance and VOCs have not yet been mentioned.

Investigations of flower aroma were essentially encouraged by the rapid development of modern analytical method and equipment (Surburg et al. 1993). This present study was aimed to identify and discriminate floral VOCs in two Chinese and two Mediterranean olive varieties, using head-space solid-phase micro-extraction and gas chromatography-mass spectrometry (HS-SPME-GC-MS). Key VOCs in four taxa were screened through the multivariate statistical analysis, which was usually used in omics. In this paper, two basic questions are addressed: (1) What are the fractions of olive floral VOCs? (2) What is the difference between Chinese and Mediterranean varieties? The obtained results will fill the research gap in olive inflorescence VOCs, as well as may explain the bee activities during blossom. Further, this study could provide further knowledge for understanding the biodiversity, origin and cultivar certification in olives.

Materials and Methods

Plant materials

Two Chinese [O. europaea L. cv. Ezhi-8 (Ezhi), O. europaea L. cv. Chenggu-32 (Chenggu)] and two Mediterranean olive varieties [O. europaea L. cv. Hojiblanca (Hojiblanca), O. europaea L. cv. Koroneiki (Koroneiki)] were studied in the garden of Longnan Xiangyu Olive Development Co., Ltd, in Longnan, Gansu, China (105°00'48"~105°00'54"E, 33°20'09"~33°20'15"N). The site has a dry-warm valley climate (annual average temperature, 14.9°C; sunshine duration, 1872 h; annual rainfall, 400~900 mm; frost-free period, 120~284 days; data from www.weather.com.cn).

Experimental design

This study was conducted with one-factor (olive taxon) experimental design. From May 5^th~10^th, each taxon was sampled by three different trees with an approximately same height and crown size. Ten branches with flowering inflorescences attached were randomly collected per tree at 9:00~10:00 am. The branch cuts were covered by wet towels and sealed in plastic bags separately. Within 6 h, the fresh samples were transported by flight in a dry cooler (4±1°C) to the laboratory in Nanjing, Jiangsu, China.

HS-SPME extraction

Branches with inflorescences from four taxa were placed into deionized water separately once the sample arrived at the laboratory. And HS-SPME extraction was carried out on the following day (at 8:00~9:00 am). The 65 µm polydimethylsiloxane/divinylbenzene (PDMS/DVB) SPME portable sampler (Supelco, Bellefonte, Pennsylvania, USA). A 3.0 g of newly bloomed inflorescences were randomly collected and placed into a 20 mL clean capped vial. To absorb VOCs of olive flowers, the SPME sampler was inserted into the vial headspace. The extraction was equilibrated for 30 min at 60±1°C. After the extraction, the portable SPME samplers were stored in dry ice before GC-MS analysis. All measurements were triplicated.

GC-MS analysis

In a random fashion, flower VOCs in the SPME sampler were injected into a TRACE DSQ GC-MS (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). The GC injector port temperature was set at 250°C in non-split mode for 1 min. An Agilent DB-5ms silica capillary column (30 m×0.33 mm×0.25 µm, 5% phenylmethyl siloxane; Agilent Technologies Inc., Santa Clara, California, USA) was used to separate different volatile compounds. The carrier gas was helium at a constant flow rate of 1 mL/min. The GC oven temperature program was set as follows: (1) initially 50°C for 1 min, (2) gradually increasing at a rate of 6°C/min to 120°C, which was held for 1 min, (3) increase at a rate of 4°C/min to 140°C, with no hold, (4) final increase at a rate of 12°C/min to 250°C, and hold for 3 min. The temperature of the transfer line and ion source was both set as 250°C. Electron ionization was carried out with 70 eV of ionization energy. The mass spectra of eluted compounds were recorded over an m/z range of 33 to 450. The linear retention indices (LRI) of the volatile compounds were calculated using an alkane series standard (C7~C30) (Sigma, St. Louis, Missouri, USA) under the same conditions. The mass spectra peaks were compared using X-caliber 3.1 (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA) with reference to the National Institute of Standards and Technology (NIST) 12th database. The identification the floral VOCs was then made basing on the spectra similarity and LRI. An ion current peak area normalization method was used to calculate the relative amount of each component.

Statistical analysis

The relative amount of floral VOCs was expressed as the mean ± standard deviation (%) of three independent experiments. Among four olive taxa, one-way analysis of variance (ANOVA) was performed using SPSS 19.0 (IBM Corp., Armonk, NY, USA) followed by Duncan’s multiple range test. The principal component analysis (PCA) and partial least squares-discriminate analysis (PLS-DA) was performed by SIMCA 13.0 software (Umetrics, Umeå, Sweden). Differential accumulated metabolites (DAMs) between each two olive taxa were identified if variable important in projection (VIP)>1.0, t-test p<0.05, and fold change (FC) >1.50 or <0.67. Origin 2017 (OriginLab Corp., Northampton, Massachusetts, USA) was used to process PCA biplot, Venn diagram, histogram and heatmap with hierarchical cluster analysis (HCA), Pearson’s correlation coefficient analysis. The odor characteristic of each volatile compound was obtained from Pherobase (www.pherobase.com) and the “Good Scents” company network database (www.thegoodscentscompany.com).

Results

Identification and comparison of VOCs 

After the GC-MS data was acquired, parallelism of the biological replications in groups was checked using Pearson correlation analysis (Fig. S1). Identification (Table 1; Fig. S2) and classification (Fig. 1) of the floral VOCs were reported. The compounds released from flowers were significantly affected by olive taxa (p<0.001). A total of 52 VOCs were identified from four varieties, and only 28 common compounds were found, which exhibited a great diversity in olive floral VOCs. Categorical analysis indicated that the main fraction of flower VOCs were different (Fig. 1).

Fig. 1: Categorical discrimination of the volatile organic compounds (VOCs) present in the flowers of two Chinese (Ezhi and Chenggu) and two Mediterranean (Hojiblanca and Koroneiki) olive cultivars. Compounds of different chemical classes among taxa were shown

Concretely, the most abundant chemical component shared by four taxa was the odorless (Z)-8-heptadecene (constitutes 42.0, 43.6, 41.5 and 50.7% for Ezhi, Chenggu, Hojiblanca and Koroneiki, respectively). Other primary aliphatics also included heptadecane, pentadecane, 9-Octadecyne, (Z)-7-hexadecene and (Z)-4-Tetradecene (13.9% of total for Ezhi, 15.1% for Chenggu, 27.4% for Hojiblanca, 31.9% for Koroneiki). The rest volatiles were not identical among taxa. For example, cis-3-hexenyl acetate, cis-3-hexen-1-ol and (E)-butanoic acid, 3-hexenyl ester with fruity and green aroma were largely released in Chinese (constituting 16.8% and 17.2% for Ezhi and Chenggu, respectively) than in Mediterranean (1.2% and 0.7% for Hojiblanca and Koroneiki, respectively) olive flowers. Furthermore, as we can see in Table 1, a few olive floral volatiles (classified as terpenes and aliphatics) were found to process roles as insect semiochemicals in previous studies.

Data (Fig. 1) showed that Chinese olive flowers emitted abundant esters, alcohols and alkanes (total sum of 78.2 and 77.7% for Ezhi and Chenggu respectively). For Mediterranean taxa, the major metabolites were alkanes and alkenes (78.1 and 85.3% for Hojiblanca and Koroneiki respectively). Interestingly, benzenoids were predominantly emitted in Chenggu (5.8%), while no benzenoid was detected in Ezhi.

PCA and HCA based on the GC-MS data

To determine the contribution of 52 compounds identified to the floral scent variation, PCA was performed (Fig. 2a). The first principal component (PC 1) and the second principal component (PC 2) contribution reached 78.3% in total. All taxa were distributed clearly in four quadrants. Cis-3-hexenyl acetate, heptadecane, pentadecane and cis-jasmon had high contributions on PC 1, while (Z)-8-heptadecene, benzene, 1,3,5-trimethoxy-, (E)-4,8-dimethylnona-1,3,7-triene, and cis-jasmon contributed significantly to PC 2. The contents of cis-3-hexenyl acetate and benzene, 1,3,5-trimethoxy- in floral VOCs positively corresponded to the identity of Chenggu. 4-penten-1-ol, 3-methyl-, cis-3-hexen-1-ol, cis-3-hexenyl 2-methylbutanoate and butanoic acid, (E)-3-hexenyl ester, were highly correlated to with Ezhi. Compounds with highly negative scores on PC 1 (cis-jasmon) and PC 2 ((E)-4,8-dimethylnona-1,3,7-triene) were related to the discrimination of Hojiblanca. Koroneiki’s identity associated greatly to compounds of (Z)-8-heptadecene, heptadecane and pentadecane.

HCA was conducted to further interpret the data. The biological replications from two Chinese and two Mediterranean olive varieties were clearly clustered accordingly (Fig. 2b), which showed the good stability and reliability of the GC-MS results. Generally, most esters and some alcohols were clustered together under the Chinese taxa (Ezhi and Chenggu), whereas terpenes and aliphathics were grouped under the two Mediterranean olive varieties (Hojiblanca and Koroneiki).

Determination of key VOCs through DAMs analysis

Fig. 2: Multivariate statistical analysis of the volatile organic compounds (VOCs) present in the flowers of two Chinese (Ezhi and Chenggu) and two Mediterranean (Hojiblanca and Koroneiki) olive cultivars. (a) PCA biplot based on the volatile compounds of four olive taxa. Numbers correspond to those in Table 1. (b) Clustering of the 52 VOCs detected in four olive taxa. Light green indicates low expression metabolite and dark green indicates high expression metabolite

Fig. 3: Determination of key VOCs present in the flowers of two Chinese (Ezhi and Chenggu) and two Mediterranean (Hojiblanca and Koroneiki) olive cultivars using differential accumulated metabolites (DAMs) analysis. (a) The number of up-regulated (red bars) and down-regulated (green bars) DAMs. (b) Venn diagram indicating the similarities and differences in VOCS among different taxa. The numbers in the overlapping areas indicate the compounds shared by the corresponding taxa. (c) 18 key VOCs identified through differential metabolites analysis under partial least squares-discriminate analysis (PLS-DA) model. Fold changes (FCs) were marked in box. #indicate the relative content of benzene, 1,3,5-trimethoxy- was 0.00% in Ezhi. Red colour represents key VOCs identified to discriminate all four olive taxa. Yellow background highlights key VOCs to discriminate Chinese and Mediterranean olive cultivars. E-Ezhi, C-Chenggu, H-Hojiblanca, K-Koroneiki.

Under the PLS-DA model, DAMs between each two olive taxa were identified as key VOCs if variable important in projection (VIP)>1.0 with p-value<0.05 in the t-test and fold change (FC) >1.5 or <0.6. Compared with Ezhi, a total of 12 and 10 DAMs were identified for Hojiblanca and Koroneiki, respectively (Fig. 3a). There were 10 DAMs identified between Chenggu and Ezhi. Relative to Chenggu, a total of 10 and 11 DAMs were identified for Hojiblanca and Koroneiki, respectively, while there were 3 DAMs (all down-regulated) identified for Koroneiki compared with Hojiblanca. In general, more up-regulated DAMs were noted in Mediterranean than in Chinese taxa. For Venn diagram, there were 6 common DAMs identified among taxa (Fig. 3b). Eighteen DAMs were identified through differential metabolites analysis under PLS-DA model (Fig. 3c). Cis-jasmon (jasmine-like scent) was a key VOC to discriminate all four olive taxa. Six VOCs (cis-jasmon, cis-3-hexen-1-ol, 3-methyl-4-penten-1-ol, heptadecane, pentadecane and 9-octadecyne) were crucial in distinguishing Chinese olives from Mediterranean ones.

Discussion

Fig. 4: Proportions of odor compounds in two Chinese (Ezhi and Chenggu) and two Mediterranean (Hojiblanca and Koroneiki) olive cultivars

Although O. europaea flowers attract honey bees (Canale and Loni 2010; Aronne et al. 2012; Giovanetti 2018; Carlos et al. 2019), previous study barely discussed olive floral VOCs. During the pre-investigation in olive groves of China, the light fragrance could be sensed by sniffing close to the inflorescence. So in this paper, the question of whether olive flowers emit scent is firstly asked. Further, this study is the first to compare the differences in floral VOCs between two Chinese and two Mediterranean olive cultivars. Moreover, as a mainstream method, chromatogram combined with solid-phase micro-extraction and headspace analysis have been widely performed to determine the VOCs of some aromatic plants (Dötterl et al. 2005; Rusanov and Kovacheva 2011; Cai et al. 2014; Li et al. 2016). The PLS-DA has been always used in many omics researches to identify differential metabolites, proteins or genes. Also, the key VOCs to discriminate different plant taxa could be identified by PLS-DA in many plant species (Shi et al. 2019), such as in teas (Zhu et al. 2017), apples (Giannetti et al. 2017), and Lycoris (Shi et al. 2019). Our study successfully identified the VOC profiles of four olive taxa and distinguished domestic varieties from Mediterranean ones.

Floral perfume is a satisfactory feature for ornamental and olfactory markets. The scent emitted by flowers mainly depends on the contents and proportions from a range of VOCs and their mixtures, which is largely influenced by the genetic characteristics of the taxa (Shi et al. 2019). Fragrant compounds of different plants species varied, even different varieties of Syringa oblata had distinct floral VOCs (Li et al. 2006). But the major VOCs of aromatic flowers are mostly terpenoids (e.g. ocimene in Lycoris (Shi et al. 2019) or benzenoids (e.g. benzyl alcohol and phenylethanol in Malus (Li 2012; Fan et al. 2019)). Most VOCs emitted from olive flowers are odorless. Our results indicated odorless (Z)-8-heptadecene was the most abundant (41.5~50.7%) VOC in all four olive taxa. However, fragrant benzenoids (0.05~5.81%) and terpenoids (2.3~7.2%) were also presented (Table 1 and Fig. 1). In fact, 63.0~66.9% of the total identified floral VOCs were odorless (Fig. 4), which explained the light fragrance in olive flowers. The odor characteristics of other aromatic volatiles of the olive flowers were also illustrated in the pie chart (Fig. 4). The fragrance types of in all the taxa were different. VOCs with green and herbal odor were the more abundant in Chinese olive varieties (16.5~16.7%) than in Mediterranean ones (1.1~1.5%). Grass odor comes from C6/C9 aldehydes and C6 alcohols (Qian et al. 2019). However, in olive they were mostly attributed to the cis-3-hexen-1-ol and its acetate. Except odorless VOCs, scent compounds with unpleasant fusel-like odor (predominately alkanes) were the most abundant in Mediterranean olive flowers, reaching up to 20.0~22.8% (Fig. 4). Yet, VOCs with refreshing odor, including floral and citrus (mostly attributed to jasmone and some mono-terpenes), were found much higher in Mediterranean olives than in Chinese counterparts. However, the overall scent of olive flower cannot be described only by odor characteristic fractions. Sensory evaluation integrated with E-nose analysis could be further performed.

As confirmed by statistics, the four cultivars can be distinguished from each other on the basis of their floral volatile compositions. Apparent chromatogram peaks at RT 14.22 (Butanoic acid, 3-hexenyl ester, (E)-, RC 3.34%) and 15.52 (cis-3-Hexenyl 2-methylbutanoate, RC 3.12%) were seen in Ezhi only (Fig. S2). Moreover, the two corresponding substances were identified as DAMs of Ezhi VOCs compared with Chenggu, Hojiblanca and Koroneiki in PLS-DA (Fig. 3c). Therefore, they could be the key VOCs to discriminate Ezhi from other three Table 1: Volatile organic compounds (VOCs) in the flowers of two Chinese and two Mediterranean olive cultivars

Note: The odor characteristic of each VOC was obtained from “Pherobase” (www.pherobase.com) and “Good Scents” company network database (www.thegoodscentscompany.com). Different low-case letters within the same VOC item represent significance among taxa with p < 0.05, and upper-case represent significance with p<0.01. * represents attractants (verified or potential) for insects reported in the literatures

olive taxa. Using the same analytic method, 1,3,5-trimethoxy-benzene (RT 20.53, RC 5.81%) was uniquely emitted in Chenggu; ocimene (RT 10.76, RC 2.79%) and (E)-4,8-dimethylnona-1,3,7-triene (RT 12.43, RC 7.34%) were the specific VOCs emitted in Hojiblanca compared with other taxa. Evidences showed that ocimene may be described as a general pollinator attractant (Farré-Armengol et al. 2013). In addition, the terpenoids have strong antioxidant activity (Xu and Yu 2015). They protect plants against peroxidation and also play a key role in plant defense mechanisms (Loreto and Velikova 2001). (E)-4,8-dimethylnona-1,3,7-triene is formed by oxidative degradation of the sesquiterpene (E)-nerolidol (Ament et al. 2006). Interestingly, Hojiblanca emitted more cis-jasmon than other olive taxa (Table 1). Since cis-jasmon (related to plant defense) induces the production of volatile compounds, including the monoterpene ocimene (Birkett et al. 2000) and (E)-4,8-dimethylnona-1,3,7-triene (Ament et al. 2006), it is likely that the related metabolic pathway is up-regulated in Hojiblanca compared with other three olive taxa.

Among four olive taxa, the Chinese varieties were clustered together, while the two Mediterranean varieties were clustered together (Fig. 2b). These findings were likely. What is worth noticing is that the contents of cis-3-hexen-1-ol and 3-methyl-4-penten-1-ol were relatively high in Chinese taxa, while Mediterranean varieties emitted more cis-jasmon (Fig. 3c). The three key VOCs are all fatty acid derivatives, which arise from linoleic or linolenic (Dudareva et al. 2013) and share the same lipoxygenase (LOX) pathway; 13-hydroperoxy intermediate. Jasmonates (methyl jasmonate, jasmonic acid, jasmon, etc.) are formed through allene oxide cyclase, allene oxide synthase, and β-oxidation pathway (Song et al. 2005; Dudareva et al. 2013), while in contrast, C6 and C9 aldehydes/alcohols (including cis-3-hexen-1-ol) are formed via another LOX branch (through the alcohol dehydrogenases pathway) (Gigot et al. 2010; Dudareva et al. 2013). Since the relative content of jasmon and ocimene involving plant defense of Hojiblanca inflorescence were the most prominent, further research questions could be asked. For example: (1) Might Hojiblanca flowers receive more visits from honey bees, which would cause the chain of reactions in plant defense? (2) If yes, what the mechanism would be?

To answer the questions above, we may first discuss the role of (Z)-8-heptadecene, which is an uncommon floral VOC and is usually identified as a minor VOC if present (Azuma et al. 1997; Peter et al. 2003). It was detected as the major VOC in Litsea cubeba (Asakawa et al. 2017), and in Trimenia moorei (Peter et al. 2003). Yet, the biological and ecological function remained unknown. Interestingly, Bohman et al. (2020) identified 8-heptadecene and pentadecane (which were majorly detected in Mediterranean olive flowers than in Chinese ones) as the electrophysiologically active compounds in both Ophrys insectifera and its Argogorytes pollinators. Moreover, long-chain hydrocarbons in other Ophrys species were identified as pollinator attractants (Schlüter et al. 2011; Sedeek et al. 2016). And these specific semiochemicals may be formed from the same carboxylic acid precursors, after elongation and decarbonylation, or possibly by decarbonylation without elongation (Bohman et al. 2020). Yet, the putative role of sex pheromone mimicry of (Z)-8-heptadecene should be tested further in the field. Actually, 20 compounds (74.6~87.3% of total VOCs) identified in olive flowers were identified or potential insect attractants, and most of them were terpenes, alkanes and alkenes, which again match the description of pollinators (mostly honey bees) attraction in fields (Canale and Loni 2010; Aronne et al. 2012; Giovanetti 2018; Carlos et al. 2019) and our own observation in olive groves. As most members of the Oleaceae are entomophilous, perfumes for attraction could be a relic character.

Although the odor characteristics of olive VOCs were provided, the question of “how olive flowers smell like” still could not be answered. The sense of plant aroma is pretty objective (Fan et al. 2018). So, an accurate approach to evaluate plant scent is a problem for researchers (Dudareva et al. 2004). The sensory evaluation and gas chromatography analysis could be combined to conduct the olive floral research (e.g. in Malus taxa (Fan et al. 2019) and some commercial cut flowers (Aros et al. 2020)). This study also has limitations in determining methods. First, the VOCs extraction was performed on the in vitro flowers, the VOCs proportion and content may be affected. Second, no chemical compound standard was used to perform accurate quantitative analysis. However, since all experimental conditions (i.e. sampling, equilibrium temperature, scent extraction and GC-MS procedures, etc.), and since GC injection orders of samples were random, the present study still can provide the peer researchers with data of some reference value.

Conclusion

For the first time, olive floral VOCs were explored in this study. Abundant 52 volatiles were identified by HS-SPME-GC-MS in four olive varieties. Among them, less than 40% compounds were found to process fragrant odor, which explained the light scent of the small white flowers. Moreover, 20 compounds of olive VOCs were reported by previous researchers as insect attractants. All the studied taxa were distinguished based on their floral VOCs via PCA, PLS-DA, and olive cultivars from the same origin were clustered together through HCA. As the products from different LOX pathway chains, green leave volatiles (e.g. 3-hexen-1-ol) with fresh/grassy odor were much more in Chinese olive flowers than in Mediterranean ones, while the two Mediterranean taxa contained more fractions of hydrocarbon and jasmone with floral/fruit/fusel-like scent. Our findings support that olive is not an anemophilous species, which makes a further step on bringing a wider vision on the origin, biodiversity and cultivar certification in O. europaea.

Acknowledgments

This research was supported by the Fundamental Research Funds for the Central Non-profit Research Institution of CAF, grant number CAFYBB2017ZA004-5 and the Fundamental Research Funds for the Central Non-profit Research Institution of CAF, grant number CAFYBB2019SY010. The authors would like to thank Mr. Xiaoyue Ji from Advanced Analysis Testing Center (AATC), Nanjing Forestry University for the GC-MS technical assistance. We would like to acknowledge Dr. Qikui Wu for the manuscript language proof.

Author Contributions

Conceptualization, JZ; methodology, ZZ and MZ; software, MZ and LZ; validation, ZW and JZ; formal analysis, ZZ; investigation, MZ and LZ; resources, JZ; data curation, ZW, MZ and LZ; writing—original draft preparation, ZZ; writing—review and editing, ZZ, ZW and JZ; visualization, ZZ; supervision, JZ and ZW; project administration, JZ; funding acquisition, ZW and ZZ.

References

Ament K, CCV Schie, HJ Bouwmeester, MA Haring, RC Schuurink (2006). Induction of a leaf specific geranylgeranyl pyrophosphate synthase and emission of (E,E)-4,8,12-trimethyltrideca-1,3,7,11-tetraene in tomato are dependent on both jasmonic acid and salicylic acid signaling pathways. Planta 224:1197‒1208

Aronne G, M Giovanetti, MR Guarracino, VD Micco (2012). Foraging rules of flower selection applied by colonies of Apis mellifera: Ranking and associations of floral sources. Funct Ecol 26:1186‒1196

Aros D, N Garrido, C Rivas, M Medel, C Müller, H Rogers, C Úbeda (2020). Floral scent evaluation of three cut flowers through sensorial and gas chromatography analysis. Agronomy 10:131

Asakawa Y, K Tomiyama, K Sakurai, Y Kawakami, Y Yaguchi (2017). Volatile compounds from the different organs of Houttuynia cordata and Litsea cubeba (L. citriodora). J Oleo Sci 66:889‒895

Azuma H, M Toyota, Y Asakawa, R Yamaoka, JG Garcia-Franco, G Dieringer, LB Thien, S Kawano (1997). Chemical divergence in floral scents of Magnolia and allied genera (Magnoliaceae). Plant Species Biol 12:69‒83

Benítez E, D Paredes, E Rodríguez, D Aldana, M González, R Nogales, M Campos, B Moreno (2017). Bottom-up effects on herbivore-induced plant defences: A case study based on compositional patterns of rhizosphere microbial communities. Sci Rep 7; Article 6251

Birkett MA, CA Campbell, K Chamberlain, E Guerrieri, AJ Hick, JL Martin, M Matthes, JA Napier, J Pettersson, JA Pickett, GM Poppy, EM Pow, BJ Pye, LE Smart, GH Wadhams, LJ Wadhams, CM Woodcock (2000). New roles for cis-jasmone as an insect semiochemical and in plant defense. Proc Natl Acad Sci USA 97:9329‒9334

Bohman B, AM Weinstein, R Mozuraitis, GR Flematti, A Borg-Karlson (2020). Identification of (Z)-8-heptadecene and n-pentadecane as electrophysiologically active compounds in Ophrys insectifera and its Argogorytes pollinator. Intl J Mol Sci 21; Article pii: E620

Borghi M, AR Fernie, FP Schiestl, HJ Bouwmeester (2017). The sexual advantage of looking, smelling, and tasting good: The metabolic network that produces signals for pollinators. Trends Plant Sci 22:338‒350

Cai X, RZ Mai, JJ Zou, HY Zhang, XL Zeng, RR Zheng, C Wang (2014). Analysis of aroma-active compounds in three sweet osmanthus (Osmanthus fragrans) cultivars by GC-olfactometry and GC-MS. J Zhejiang Univ Sci 15:638‒648

Carlos MN, M Antonio, I Jorge, T Rubén, C Gemma, MP Luis, R Teresa, R Carlos, G José, R Pedro (2019). Low-intensity management benefits solitary bees in olive groves. J Appl Ecol 57:111‒120

Campeol E, G Flamin, S Chericoni, S Catalano (2001). Volatile compounds from three cultivars of Olea europaea from Italy. J Agric Food Chem 49:5409‒5411

Canale A, A Loni (2010). Insects visiting olive flowers (Olea europaea L.) in a Tuscan olive grove. J Zool 92:95‒98

Dötterl S, LM Wolfe, A Jürgens (2005). Qualitative and quantitative analyses of flower scent in Silene latifolia. Phytochemistry 66:203‒213

Dudareva N, A Klempien, JK Muhlemann, I Kaplan (2013). Biosynthesis, function and metabolic engineering of plant volatile organic compounds. New Phytol 198:16‒32

Dudareva N, E Pichersky, J Gershenzon (2004). Biochemistry of plant volatiles. Plant Physiol 135:1893‒1902

Fan J, W Zhang, D Zhang, G Wang, F Cao (2019). Flowering stage and daytime affect scent emission of Malus ioensis “Prairie Rose”. Molecules 24; Article 2356

Fan J, W Zhang, T Zhou, D Zhang, D Zhang, L Zhang, G Wang, F Cao (2018). Discrimination of Malus taxa with different scent intensities using electronic nose and gas chromatography-mass spectrometry. Sensors 18; Article 3429

Farré-Armengol G, I Filella, J Llusià, J Peñuelas (2013). Floral volatile organic compounds: Between attraction and deterrence of visitors under global change. Perspect Plant Ecol 15:56‒67

Flamin G, PL Cioni, I Morelli (2003). Volatiles from leaves, fruits, and virgin oil from Olea europaea cv. Olivastra Seggianese from Italy. J Agric Food Chem 51:1382‒1386

Flint HM, SS Salter, S Walters (1979). Caryophyllene: An attractant for the green lacewing. Environ Entomol 8:1123‒1125

Giannetti V, MB Mariani, P Mannino, F Marini (2017). Volatile fraction analysis by HS-SPME/GC-MS and chemometric modeling for traceability of apples cultivated in the Northeast Italy. Food Cont 78:215‒221

Gigot C, M Ongena, ML Fauconnier, JP Wathelet, Pd Jardin, P Thonart (2010). The lipoxygenase metabolic pathway in plants: Potential for industrial production of natural green leaf volatiles. Biotechnol Agron Soc Environ 14:451‒460

Giovanetti M (2018). Do bees like olive? A preliminary analysis of honey bee behaviour on flowers of the wind-pollinated species Olea europaea. Acta Hortic 1199:121‒126

Griggs WH, HT Hartmann, MV Bradley, BT Iwakiri, JE Whisler (1975). Olive Pollination in California, Vol. 869, p:49. California Agricultural Experiment Station Bulletin, University of California, Berkeley, USA

Han R, W Xing, W Kong, W Bai, X Bai (2017). Evaluation of fruit quality and characters of five leading olive varieties from Wudu district of Gansu province. Chin Oil Fat 42:140‒144

Huang M, AM Sanchez-Moreiras, C Abel, R Sohrabi, S Lee, J Gershenzon, D Tholl (2011). The major volatile organic compound emitted from Arabidopsis thaliana flowers, the sesquiterpene (E)-b-caryophyllene, is a defense against a bacterial pathogen. New Phytol 193:997‒1008

James DG (2003). Synthetic herbivore-induced plant volatiles as field attractants for beneficial insects. Environ Entomol 32:977‒982

Junker RR, J Gershenzon, SB Unsicker, (2011). Floral odor bouquet loses its ant repellent properties after inhibition of terpene biosynthesis. J Chem Ecol 37:1323‒1331

Kiritsakis AK (1988). Flavor components of olive oils, a review. J Amer Oil Chem Soc 75:673‒681

Kong W, W Li, W Xing, M Dong,R Han, W Zhao, W Bai, X Bai, J Zhang (2016). Research on the fruit quality of the main cultivated olive varieties in Wudu district. J Chin Cereal Oil Assoc 31:87‒92

Li W (2012). HS-SPME-GC-MS analysis of volatile constituents from the flowers and leaves of Malus baccata (Linn.) Borkh. Nat Prod Res Dev 24:490‒493

Li Y, H Ma, Y Wan, T Li, X Liu, Z Sun, Z Li (2016). Volatile organic compounds emissions from Luculia pinceana flower and its changes at different stages of flower development. Molecules 21; Article 531

Li ZG, MR Lee, DL Shen (2006). Analysis of volatile compounds emitted from fresh Syringa oblata flowers in different florescence by headspace solid-phase microextraction—gas chromatography–mass spectrometry. Anal Chim Acta 576:43‒49

Loreto F, V Velikova (2001). Isoprene produced by leaves protects the photosynthetic apparatus against ozone damage, quenches ozone products, and reduces lipid peroxindation of cellular membranes. Plant Physiol 127:1781‒1787

McDonough LM, JB Bailey, MP Hoffmann, BA Leonhardt, DF Brown, CL Smithhisler, K Olsen (1986). Western avocado leafroller, Amorbia cuneana (Walsingham), (Lepidoptera: Tortricidae): Discovery of populations utilizing different ratios of sex pheromone components. J Chem Ecol 12:2107‒2116

Mendki MJ, K Ganesan, S Prakash, MV Suryanarayana, RC Malhotra, K Rao, R Vaidyanathaswamy (2000). Heneicosane: An oviposition-attractant pheromone of larval origin in Aedes aegypti mosquito. Curr Sci 78:1295‒1296

Millar JG, RE Rice, Q Wang (1997). Sex pheromone of the mirid bug Phytocoris relativus. J Chem Ecol 23:1743‒1754

Miller DR, RJ Rabaglia (2009). Ethanol and α-Pinene: Attractant kairomones for bark and ambrosia beetles in the Southeastern US. J Chem Ecol 35:435‒448

Mitra S, N Sarkar, A Barik (2017). Long-chain alkanes and fatty acids from Ludwigia octovalvis weed leaf surface waxes as short-range attractant and ovipositional stimulant to Altica cyanea (Weber) (Coleoptera: Chrysomelidae). Bull Entomol Res 107:391‒400

Morettini A (1972). Olivicoltura, pp:522. Ramo Editoriale degli Agricoltori, Rome, Italy

Mozuraitis R, M Stranden, MI Ramirez, AK Borg-Karlson, H Mustaparta (2002). (-)-Germacrene D increases attraction and oviposition by the tobacco budworm moth Heliothis virescens. Chem Sen 27:505‒509

Murali-Baskaran RK, KC Sharma, P Kaushal, J Kumar, P Parthiban, S Senthil-Nathan, RW Mankin (2018). Role of kairomone in biological control of crop pests-a review. Physiol Mol Plant Pathol 101:3‒15

Ortega RA, A Mahnert, H Müller, C Berg (2016). The plant is crucial: Specific composition and function of the phyllosphere microbiome of indoor ornamentals. FEMS Microbiol Ecol 92; Article pii: fiw173

Pennanech M, DB Pho, JM Pechine, JM Jallon (1995). Insect hydrocarbons: Analysis, structures and functions. Eppo Bull 25:343‒348

Peter B, S Tammy, W Peter, A Hiroshi, L Mathew, LB Thien, J Bruhl (2003). The pollination of Trimenia moorei (Trimeniaceae): Floral volatiles, insect/wind pollen vectors and stigmatic self‐incompatibility in a basal angiosperm. Ann Bot 92:445‒458

Pichersky E, J Gershenzon (2002). The formation and function of plant volatiles: Perfumes for pollinator attraction and defense. Curr Opin Plant Biol 5:237‒243

Qian C, W Quan, Z Xiang, C Li (2019). Characterization of volatile compounds in four different Rhododendron flowers by GC×GC-QTOFMS. Molecules 24; Article 3327

Raguso RA (2008). Wake up and smell the roses: The ecology and evolution of floral scent. Annu Rev Ecol Evol Syst 39:549‒569

Rao G, J Zhang, X Liu, Y Luo (2019). Identification of putative genes for polyphenol biosynthesis in olive fruits and leaves using full-length transcriptome sequencing. Food Chem 300:125246

Rusanov K, N Kovacheva (2011). Traditional Rosa damascena flower harvesting practices evaluated through GC/MS metabolite profiling of flower volatiles. Food Chem 129:1851‒1859

Ruther J, A Reinecke, T Tolasch, M Hilker (2002). Phenol—another cockchafer attractant shared by Melolontha hippocastani Fabr. and M. melolontha L. Z Naturforsch 57c:910‒913

Schiestl FP (2015) Ecology and evolution of floral volatile-mediated information transfer in plants. New Phytol 206:571‒577

Schiestl FP, M Ayasse (2001). Post-pollination emission of a repellent compound in a sexually deceptive orchid: A new mechanism for maximising reproductive success? Oecologia 126:531‒534

Schlüter PM, S Xu, V Gagliardini, E Whittle, J Shanklin, U Grossniklaus, FP Schiestl (2011). Stearoyl-acyl carrier protein desaturases are associated with floral isolation in sexually deceptive orchids. Proc Natl Acad Sci USA 108:5696‒5701

Sedeek KEM, E Whittle, D Guthörl, U Grossniklaus, J Shanklin, PM Schlüter (2016). Amino acid change in an orchid desaturase enables mimicry of the pollinator’s sex pheromone. Curr Biol 26:1505‒1511

Shi T, Y Yue, M Shi, M Chen, X Yang, L Wang (2019). Exploration of floral volatile organic compounds in six typical Lycoris taxa by GC-MS. Plants 8; Article 422

Silva WD, JM Bento, LM Hanks, JG Millar (2018). (Z)-7-Hexadecene is an aggregation-sex pheromone produced by males of the South American Cerambycid beetle Susuacanga octoguttata. J Chem Ecol 44:1115‒1119

Song M, D Kim, S Lee (2005). Isolation and characterization of a jasmonic acid carboxyl methyltransferase gene from hot pepper (Capsicum annuum L.). J Plant Biol 48:292‒297

Surburg H, M Guentert, H Harde (1993). Volatile compounds from flowers analytical and olfactory aspects. In: Bioactive Volatile Compounds from Plants, ACS Symposium Series, Vol. 525, pp:160‒186. Teranishi R, S Kint (eds.). American Chemical Society, Washington DC, USA

Sutherland OR, RF Hutchins (1972). α-Farnesene, a natural attractant for codling moth larvae. Nature 239:170‒170

Witte V, LM Abrell, AB Attygalle, X Wu, J Meinwald (2007). Structure and function of dufour gland pheromones from the crazy ant Paratrechina longicornis. Chemoecology 17:63‒69

Xu W (1981). Introduction of olive trees in China. Sci Silv 17:78‒83. (In Chinese)

Xu W, Y Han (1965). Study on Albanian olive and its introduction in China. Sci Silv 10:86‒99. (In Chinese)

Xu L, F Yu (2015). Corolla structure and fragrance components in Styrax tonkinensis. Trees 29:1127‒1134

Yang L, P Cheng, X-F Zhu, Z-g Hang, H Li (2018). Composition and dynamic variations of the natural volatiles of Prunus armeniaca. Intl J Agric Biol 20:169‒182

Zhang D, Y Xue, L Zhu, J Chai (2016). Polyphenols composition of virgin olive oils in Longnan district in China. Chin Oil Fat 41:37‒40

Zhu Y, CY Shao, HP Lv, Y Zhang, WD Dai, L Guo, JF Tan, QH Peng, Z Lin (2017). Enantiomeric and quantitative analysis of volatile terpenoids in different teas (Camellia sinensis). J Chromatogr A 1490:177‒190