SYSTEMATIC REVIEW published: 31 July 2020 doi: 10.3389/fpls.2020.01154 Deciphering the Biotic and Climatic Factors That Influence Floral Scents: A Systematic Review of Floral Volatile Emissions Gerard Farré-Armengol 1,2,3*, Marcos Fernández-Martı́nez 4, Iolanda Filella 2,3, Robert R. Junker 1,5 and Josep Peñuelas 2,3 1 Department of Biosciences, University of Salzburg, Salzburg, Austria, 2 CSIC, Global Ecology Unit CREAF-CSIC-UAB, Barcelona, Spain, 3 CREAF, Barcelona, Spain, 4 PLECO (Plants and Ecosystems), Department of Biology, University of Antwerp, Wilrijk, Belgium, 5 Evolutionary Ecology of Plants, Department of Biology, Philipps-University Marburg, Marburg, Germany Currently, a global analysis of the information available on the relative composition of the floral scents of a very diverse variety of plant species is missing. Such analysis may reveal general patterns on the distribution and dominance of the volatile compounds that form these Edited by: Aleš Svatoš, mixtures, and may also allow measuring the effects of factors such as the phylogeny, Max Planck Institute for Chemical pollination vectors, and climatic conditions on the floral scents of the species. To fill this gap, Ecology, Germany we compiled published data on the relative compositions and emission rates of volatile Reviewed by: organic compounds (VOCs) in the floral scents of 305 plant species from 66 families. We also Islam S. Sobhy, Keele University, United Kingdom gathered information on the groups of pollinators that visited the flowers and the climatic John A. Morgan, conditions in the areas of distribution of these species. This information allowed us to Purdue University, United States characterize the occurrence and relative abundances of individual volatiles in floral scents *Correspondence: Gerard Farré-Armengol and the effects of biotic and climatic factors on floral scent. The monoterpenes trans-b- gerardfa1987@gmail.com ocimene and linalool and the benzenoid benzaldehyde were the most abundant floral VOCs, in both ubiquity and predominance in the floral blends. Floral VOC richness and relative Specialty section: This article was submitted to composition were moderately preserved traits across the phylogeny. The reliance on Plant Metabolism and Chemodiversity, different pollinator groups and the climate also had important effects on floral VOC a section of the journal richness, composition, and emission rates of the species. Our results support the Frontiers in Plant Science hypothesis that key compounds or compounds originating from specific biosynthetic Received: 26 April 2020 Accepted: 15 July 2020 pathways mediate the attraction of the main pollinators. Our results also indicate a Published: 31 July 2020 prevalence of monoterpenes in the floral blends of plants that grow in drier conditions, Citation: which could link with the fact that monoterpene emissions protect plants against oxidative Farré-Armengol G, Fernández-Martı´nez M, Filella I, stresses throughout drought periods and their emissions are enhanced under moderate Junker RR and Peñuelas J (2020) drought stress. Sesquiterpenes, in turn, were positively correlated with mean annual Deciphering the Biotic and temperature, supporting that sesquiterpene emissions are dominated mainly by ambient Climatic Factors That Influence Floral Scents: A Systematic Review temperature. This study is the first to quantitatively summarise data on floral-scent emissions of Floral Volatile Emissions. and provides new insights into the biotic and climatic factors that influence floral scents. Front. Plant Sci. 11:1154. doi: 10.3389/fpls.2020.01154 Keywords: climate, floral volatiles, phylogeny, pollination syndromes, terpenoids, VOC composition, VOC richness Frontiers in Plant Science | www.frontiersin.org 1 July 2020 | Volume 11 | Article 1154 Farré-Armengol et al. Systematic Review of Floral Scents INTRODUCTION Plant emissions of volatile organic compounds (VOCs), including floral scents, can be affected by climatic variables Floral scent is an important trait of flowering plants and plays major such as temperature and humidity and by other environmental roles in the interactions of plants with other organisms, including abiotic variables such as light, CO2 concentration, wind speed, or the attraction of pollinators (Raguso, 2004; Schiestl, 2010; Farré- the concentration of diverse oxidative pollutants such as ozone Armengol et al., 2013; Junker and Parachnowitsch, 2015; Kantsa and nitrogen oxides. The effects of all these environmental et al., 2018). Effective pollinators (those that carry pollen from the abiotic variables and stresses on foliar VOC emissions and on anthers to the stigmas of conspecific plants) are either specialist VOC emissions from vegetation at a global scale are well floral visitors of a limited spectrum of plant species or generalist characterized (Kesselmeier and Staudt, 1999; Peñuelas and floral visitors with a short-term specialization known as flower Llusià, 2001; Owen et al., 2002; Niinemets et al., 2004; Duhl constancy (Chittka et al., 1999). Both pollinators with specialized et al., 2008; Niinemets et al., 2010; Holopainen and Gershenzon, innate flower preferences and those temporarily specialized via 2010; Niinemets, 2010; Peñuelas and Staudt, 2010), as well as associative learning depend on cues or signals to distinguish those of endogenous variables that are partially controlled by the amongst plant species (Chittka and Thomson, 2001; Kunze and environment, such as plant nutrient contents (Ferná ndez- Gumbert, 2001; Chittka and Raine, 2006; Majetic et al., 2008; Burger Martı́nez et al., 2018 and references therein). Few studies, et al., 2010; Leonard et al., 2011a; Leonard et al., 2011b). Floral though, have addressed the effects of climatic variables on volatiles are key floral traits that mediate flower–visitor interactions floral-scent emissions. Some of these studies have shed some by attracting pollinators, structuring flower–visitor communities, light on the responses of floral volatile emissions to temperature and defending against plant and flower antagonists (Junker and (Jacobsen and Olsen, 1994; Sagae et al., 2008; Hu et al., 2013; Blüthgen, 2010; Junker et al., 2010; Galen et al., 2011; Schiestl et al., Farré -Armengol et al., 2014; Farré -Armengol et al., 2015a), 2014; Junker and Parachnowitsch, 2015). In addition to pollinator drought (Burkle and Runyon, 2016; Glenny et al., 2018), light attraction, floral scents play major roles in the interactions with (Jacobsen and Olsen, 1994; Hu et al., 2013), and pollution herbivores, parasitoids, and floral larcenists (Junker and Blüthgen, (Girling et al., 2013; Lusebrink et al., 2015; Farré -Armengol 2008; Raguso, 2008a; Kessler et al., 2008; Junker and Blüthgen, et al., 2016b; Saunier and Blande, 2019). Environmental 2010; Galen et al., 2011; Farré-Armengol et al., 2013; Junker, 2016), variables have such effects on floral-scent emissions, so we and they also have important effects on the growth and composition hypothesize that climate can potentially select floral scents with of floral microbial communities (Heil, 2011; Junker et al., 2011; properties that are most suited to the environmental conditions Huang et al., 2012; Junker and Tholl, 2013; Farré -Armengol that plants and their flowers experience. et al., 2016a). A global analysis of the currently available information on the Pollinators play a major role in the reproduction of most floral scents of various species from many families is needed to angiosperms (Ollerton et al., 2011) and exert important selection shed light on how factors such as phylogeny, pollinators, and pressures on plant and floral phenotypes, including floral scents climate determine floral VOC emissions of the species. Previous (Wright and Schiestl, 2009; Parachnowitsch et al., 2012; studies by Knudsen et al. (1993) and Knudsen et al. (2006) Parachnowitsch et al., 2013; Schiestl and Johnson, 2013). The qualitatively described the occurrence of >1,700 compounds in pollination syndrome hypothesis postulates that the floral traits the flowers of 991 species and discussed whether the occurrence of unrelated plants pollinated by the same pollinators tend to and richness of particular volatiles had phylogenetic signals and converge, including advertising signals (Faegri and van der Pijl, whether the compounds depended on the pollination biology of 1979; Fenster et al., 2004). Researchers have long discussed the species, i.e. their main pollinator type. The available data on pollination syndromes, arguing in favour or against their the quantitative compositions and emission rates of floral VOCs, reliability as effective classifiers of floral phenotypes that can be however, have not yet been compiled and analyzed. We aimed to used to predict the plant’s most efficient pollinators (Herrera, fill this gap by searching published studies for data on the 1996; Ollerton, 1996; Waser et al., 1996; Armbruster et al., 2000; complete composition of floral scents and the emission rates Fenster et al., 2004; Lá zaro et al., 2008; Raguso, 2008b; Ollerton of each compound or alternatively describing the relative et al., 2009; Rosas-Guerrero et al., 2014; Ollerton et al., 2015). percentage of contribution of each compound to the blend. Many studies have described cases of floral-trait convergence by We aimed to identify the effects of biological (pollinators) and mono- and polyphyletic groups of plant species that share their climatic factors on the floral scents of the species by combining main pollinators (Thomson et al., 2000; Stuurman et al., 2004; the data on the floral scents with the available data on the Wilson et al., 2004; Rosas-Guerrero et al., 2014). Some studies pollinators and climatic conditions in the regions where the plant have reported convergent evolution in floral-scent composition species were sampled. driven by a shared reliance on the same pollinator group We compiled the available information on floral-VOC (Knudsen and Tollsten, 1993; Knudsen and Tollsten, 1995; emission rates and/or relative VOC compositions for 305 plant Miyake et al., 1998; Andersson et al., 2002; Knudsen et al., species from 66 families. The database we compiled contained 2004). Notable evidence also suggests that pollinators have >800 compounds classified into nine groups: fatty acid derivatives, strong evolutionary impacts on the intensity and composition amino acid derivatives, benzenoids, monoterpenes, sesquiterpenes, of floral scents emitted by plants (Parachnowitsch et al., 2012; irregular terpenes, nitrogen-containing compounds, sulphur- Parachnowitsch et al., 2013). containing compounds, and miscellaneous compounds. We also Frontiers in Plant Science | www.frontiersin.org 2 July 2020 | Volume 11 | Article 1154 Farré-Armengol et al. Systematic Review of Floral Scents obtained information about the location and the climatic index of aridity using the climatic data as: Gaussen index = conditions where the populations from the different species that annual precipitation/(2 * mean annual temperature). We further were measured grew and about the type of pollinators that visited obtained information on the pollinators that visited the flowers the flowers, as described by the original studies. The information of each plant species as described by the source studies. contained in our database allowed us to identify the most We classified floral VOCs based on their biosynthetic ubiquitous and dominant VOCs in floral scents (those that pathways, which was the predominant classification in all our more frequently had the highest relative abundances in floral data sources and enabled comparisons with phylogenetic VOC blends). We further determined whether phylogeny, reliance hypotheses. We thus divided floral VOCs into the nine major on different types of pollinators, and climate were correlated with classes: fatty acid derivatives, amino acid derivatives, benzenoids, floral VOC richness, scent composition, and rate of emission. We monoterpenes, sesquiterpenes, irregular terpenes, nitrogen- hypothesized that the compositions and rates of emission of floral containing compounds, sulphur-containing compounds, and scents have been preserved throughout evolutionary history, and miscellaneous compounds. We entered the data on the floral we aimed to differentiate between the effects of phylogeny on floral scent of each species as “presence/absence” (1/0), “relative scents and the effects of biotic and climatic factors. We expected percentages of the total blend” (%), and “emission rates” for all that the pollination syndrome would be correlated with VOC individual compounds and for the nine major groups of floral richness, composition, and emission rate of floral scents. Finally, volatiles identified above. The source studies did not always we hypothesized that climate would exert some selective pressures provide emission rates, but all studies provided percentages of on the production and emission of floral scents, thus positively or the total floral VOC blend. Emission rates were provided in negatively stimulating the emissions of all or some compounds various units (mg h-1 flower-1, mg h-1 inflorescence-1, mg h-1 g DW-1, under the environmental conditions where each species grew and mg h-1 g FW-1), depending on the methods used in each and flowered. study. Some studies specified the isomer(s) of various isomeric compounds in the floral scent of the species, but others did not. Several compounds thus appear in our database as repeated variables with and without a specified isomer. METHODS Search Criteria Classification of the Plant Species Into We exhaustively searched the Web of Knowledge and Google Pollination Syndromes Scholar for studies of floral scent using combinations of the We obtained information from the source studies of the keywords “floral”/“flower” and “volatiles”/”VOCs”/“scent”. We pollinators that visited each plant species. Several species were chose studies that provided complete data on the emission rates pollinated or visited by different pollinator groups, so we created and/or relative percentages of all floral VOCs emitted. We several binary variables in our database to indicate whether a discarded studies that did not report the complete bouquet of particular plant species was or was not pollinated by a particular VOCs of the floral scents but focused only on particular pollinator group (wind, animals, insects, Lepidoptera, Coleoptera, compounds, thus omitting other compounds that were emitted Diptera, Hymenoptera, bats, and birds). The description of but were not the focus of the study. We finally selected 58 studies pollinators differed across classification levels amongst the that provided information on the complete compositions of floral studies, so we created different binary variables to characterize scents of one or more species. The references for all the studies the spectrum of pollinators, some of which were included in from which we used data to make our database can be found in other groups (e.g. Lepidoptera