The Ecology and Evolution of Fruit Odor: Implications for Primate Seed Dispersal
- 428 Downloads
Primates are now known to possess a keen sense of smell that serves them in various contexts, including feeding. Many primate species are frugivorous and provide essential seed dispersal services to a variety of plants. Studies of pollination ecology, and recently seed dispersal ecology, indicate that animal mutualist behavior exerts selection pressures that drive changes in flower and fruit traits. As a result, the use of olfaction in in primate feeding ecology may have affected the evolution of fruit odor in species that rely on primate seed dispersal. However, this hypothesis is seldom tested. Here, we summarize the available information on how primates may have affected the evolution of fruit odor. We ask what the chemistry of primate fruit odor may look like, what information fruit odor may convey, whether there are geographical differences in fruit odor, and what other factors may affect the odor of fruits consumed by primates. We identify many gaps in the available data and offer research questions, hypotheses, and predictions for future studies. Finally, to facilitate standardization in the field, we discuss methodological issues in the process of odor sampling and analysis.
KeywordsCoevolution Fruit aroma Fruit secondary metabolites Olfaction Sensory ecology
Primates were long considered primarily visual mammals in which the sense of smell is reduced to almost negligible levels (Fobes and King 1982). This view changed after it became clear that primates possess olfactory abilities that are often comparable to those of traditionally macrosmatic mammals such as rodents or dogs (Laska et al. 2000). Primates are now known to use their sense of smell in multiple domains: reproduction and kin recognition (Boulet et al. 2009), species recognition (delBarco-Trillo and Drea 2014), and food selection (Hiramatsu et al. 2009; Nevo and Heymann 2015).
The use of olfaction when feeding on fruit raises the possibility of evolutionary and ecological interactions between the primate sense of smell and the odors of the fruits they consume. Primates constitute a substantial part of the frugivore biomass in tropical systems and provide seed dispersal services to many plant species (Chapman and Russo 2007; Wright et al. 2005), a pattern of interaction that appears to date back to the Paleocene (Eriksson 2014). This implies that if primates use their sense of smell for food selection, plants that rely on their seed dispersal services may experience selective pressure to increase the detectability and attractiveness of their fruit and thus promote seed dispersal (Nevo et al. 2016). Such evolution of fruit odor in response to primate feeding behavior would be in line with the well-established notion that floral scent in many species has evolved to mediate communication with the animals that pollinate the flowers (Raguso 2008; Schiestl 2015; Valenta et al. 2017). The causal link between insect behavior and floral scent evolution has recently been demonstrated in experiments in which pollination by different insects over several generations yielded real-time evolution of floral scent (Gervasi and Schiestl 2017).
The study of fruit odor, i.e., the bouquet of volatile organic compounds (VOCs) emitted by fruit, has lagged behind the study of floral scent (Valenta et al. 2017). This is probably because the less specialized interaction networks in frugivory (Blüthgen et al. 2007) are more likely to be characterized by selection pressures by multiple frugivores, and thus frugivore–fruit mutualisms are predicted to result in diffuse coevolution, relative to pollinator–flower interactions. Moreover, fruits are developmentally simpler than flowers, as in most cases they develop from a single floral structure (Giovannoni 2004), and fruit traits, including VOCs, have been hypothesized to convey no adaptive advantage (Eriksson and Ehrlén 1998). Yet fruit secondary metabolites have been shown to fulfill many functions (e.g., Cipollini et al. 2004; Whitehead and Bowers 2013), and studies in the past decade support the hypothesis that a major function of fruit odor is to attract animal seed dispersers (Borges et al. 2008; Hodgkison et al. 2007, 2013; Lomáscolo et al. 2010; Nevo et al. 2015, 2016).
The facts that primates are important seed dispersers that employ their sense of smell during fruit foraging and selection, and that fruit odor may be malleable to selection pressures imply that primates may have influenced the evolution of fruit odor. However, this hypothesis remained untested until recently and only a few studies have quantified the odors of fruits consumed by primates (Nevo et al. 2016; Valenta et al. 2013, 2015b, 2016b). These studies show that fruit odor may be important for primate feeding ecology and that primates may act as selective forces shaping the way fruits smell, and thus that an important aspect of primate behavioral ecology remains little understood. The goal of this review is to summarize the data currently available relevant to the question of what odors fruits consumed by primates emit. More specifically, we first examine what the primate olfactory system is capable of and what chemicals characterize fruits consumed by primates. We then ask what information primates may be getting from fruit odor, whether there are any regional differences in this information, and what other factors may affect the odors of fruits consumed by primates. Given the paucity of studies, the open questions far outnumber the answers. Therefore, in addition to summarizing what we know, we include research questions, hypotheses, and predictions that we hope will be addressed in future research. Finally, we provide a discussion of methodological issues in the study of fruit odor, and then end with conclusions.
Primate Olfactory Capabilities: What can they Smell?
Olfactory perception begins on the surface of the olfactory epithelium in the nose, on which thousands of olfactory sensory neurons express a variety of olfactory receptors (ORs), each of which can be activated only by a number of VOCs (Lledo et al. 2005). By extension, at least in theory, a species that possesses more OR genes should be able to detect more VOCs and better discriminate between individual compounds and VOC mixtures (Niimura 2012). Primates of different lineages vary in the number of potentially functional OR genes (i.e., those that are not disrupted by mutations, insertions, deletions, and frame shifts, and can thus code for a functional olfactory receptor) they possess, with ca. 300–400 in haplorrhines and ca. 640 in strepsirrhines (Matsui et al. 2010; Niimura et al. 2014). These figures are lower than those of some other mammals such as rodents (ca. 1000 potentially functional genes) and African elephants (Loxodonta africana, ca. 2000: Niimura et al. 2014) but comparable with those of birds of several lineages and higher than those of the main group of avian frugivores, the passerines (Steiger et al. 2008). However, the assumption that a larger number of functional ORs directly translates into higher olfactory discrimination capacity is somewhat naïve. It is unclear how strong this relationship is (Nevo and Heymann 2015), and even relatively small numbers of functional ORs may allow detection and discrimination of up to 1012 different odor mixtures (Bushdid et al. 2014; Weiss 2014).
Physiological tests that quantified phenotypes rather than genetic proxies for olfactory capabilities show that both platyrrhines and catarrhines have keen sensitivity (the minimum required concentration for detection of an odorant) to and discrimination capacity (the ability to discriminate between two odorants) of various odorant groups. These include various classes of compound, some of which are very common in fruit odor: aliphatic esters (Hernandez Salazar et al. 2003; Laska and Freyer 1997; Laska and Seibt 2002a; Laska et al. 2000), alcohols (Laska and Seibt 2002b; Laska et al. 2000, 2006b), aldehydes (Laska et al. 2000, 2006b), aliphatic acids (Laska et al. 2000, 2004), terpenoids (Laska et al. 2005, 2006a), nitrogen- and sulfur-containing compounds (Laska et al. 2007), and mixtures of odorants (Laska and Hudson 1993), particularly those that mimic fruit odor (Hübener and Laska 1998; Nevo et al. 2015). Detection thresholds for these compounds are typically around or below 1 ppm and sometimes even as low as 1 ppb (Laska and Seibt 2002a; Laska et al. 2000, 2006a, b). As a result, it is likely that primates are capable of effectively detecting and discriminating a great diversity of plant VOCs. This implies that if fruits are under selection to emit odors that are accessible to primates, they can readily use a substantial proportion of the existing repertoire of VOCs emitted by other plant organs, like the flowers from which fruits develop (Dobson 2006; Knudsen et al. 2006). The available data do not tell us whether the primate sense of smell tends to be particularly tuned to any particular odorant class. Thus, if primates do exert a selection pressure for detectable fruit odor, there is no reason to predict that this selection would favor any particular VOC or VOC class. Further, despite the variation in olfactory-related neuroanatomical structures and OR gene repertoires, there is so far no evidence for systematic and predictable differences in the olfactory capacities of different primate lineages (Nevo and Heymann 2015).
The Chemistry of Primate-Consumed Fruit Odor
Plants synthesize a very large number of VOCs. Studies of floral scent have identified thousands of different odorants in flower headspace (the air surrounding the flower), which is probably an underestimate (Knudsen et al. 2006). Floral scent can contain both common and unique VOCs, which are often used as private communication channels in highly specialized pollination interactions (Raguso 2008; Schiestl 2015). Generalist flowers, which are pollinated by a diverse community of pollinators, tend to use different mixtures of common plant volatiles, namely terpenoids, aromatic compounds, and fatty acid derivatives (Dobson 2006). Flowers have evolved to secondarily use biosynthetic machinery that originally evolved as a chemical defense system (Pellmyr and Thien 1986), and fruit odor can contain any of those many compounds. Given that fruit–mutualist interaction networks tend to be more generalized than those of flowers (Blüthgen et al. 2007), it is likely that fruit odor would resemble the odor of more generalist flowers and primarily contain more common odor compounds, rather than unique VOCs that characterize specialized pollination interactions.
Fruit odor has been studied primarily in domesticated cultivars, in which both the proportions and overall emission intensity of VOCs are likely to have been changed in the process of cultivation (Borges 2015; Rodríguez et al. 2013). However, the compositions of their odors are probably limited by the availability of odorants in their wild ancestors. Not surprisingly, odors of domesticated fruits are dominated by common plant VOCs. Ripe tomatoes are mainly rich in green leaf volatiles (GLVs) (Goff and Klee 2006), compounds that function in leaf defense, among other things (Kessler 2015). Tomato odor is also rich in other fatty acid derivatives such as alcohols, aldehydes, and ketones (Goff and Klee 2006). Citrus fruits tend to be highly rich in monoterpenes and sesquiterpenes and their derivatives, e.g., monoterpene alcohols (Rodríguez et al. 2011, 2013). Other domesticated fruits are rich in aldehydes, aromatic compounds, and aliphatic esters (Rodríguez et al. 2013). Aliphatic esters dominate the odors of fruits such as bananas and apples and are responsible for odors commonly described as “fruity” (Schwab et al. 2008).
Research on the odors of wild fruits is much scarcer. Studies have focused mainly on various species of figs (Ficus spp.) and found that ripe fig odor is dominated by common plant VOCs such as terpenoids, aromatic compounds, and fatty acid derivatives (Borges et al. 2008, 2013; Hodgkison et al. 2007, 2013). This pattern is also true for a species of epiphytic cactus (Rhipsalis juengeri: Schlumpberger et al. 2006) and for two species of bird-dispersed fruits in Peruvian Amazonia (Nevo et al. 2016), although the latter emitted very low amounts of odor. However, some fruits primarily emit less common VOCs. For example, the odor of the durian fruit (Durio zibethinus) is characterized by sulfur-containing compounds (Teh et al. 2017).
Although there are few published data, we predict that the odor of primate-consumed fruits is likely to be dominated by “generic” odor mixtures that contain primarily common compounds such as terpenoids, aromatic compounds, and fatty acid derivatives. This is based on available data for a few dozen species in three different locations: Madagascar, continental Africa, and Amazonia, and mainly on the fact that these chemical classes dominate the odors of domesticated fruits and flowers that, like most fruits, exhibit a generalist pollination syndrome. Compounds belonging to these classes are exactly those to which several primate species showed high sensitivity and discrimination capacity (Laska and Seibt 2002a, b; Laska et al. 2000, 2005, 2006a). It is, of course, possible that rarer chemicals are present in fruits and that primates can readily perceive them. However, from an evolutionary perspective, the fact that primates can readily perceive common odor compounds and that fruits tend to belong to rather generalist dispersal syndrome makes it likely that plants that have been selected to attract primates through olfactory signals have done so using the existing biosynthetic pathways plants use to produce common plant VOCs.
What Information can Fruit Odor Convey?
Presence and location of fruiting trees: Fruit odor may allow primates to detect whether a species currently provides ripe fruits. This would be in the interest of both animal and plant: Plants may benefit if major seed dispersal vectors are aware that their fruits are ripe, and animals can increase the efficiency of their foraging efforts if they can remotely sense the availability of ripe fruits. In addition to foraging, primate movement in their habitats can be driven by many factors, such as the maintenance of territories (Langergraber et al. 2017). Primates possess at least some spatial knowledge about their habitat (Janson and Byrne 2007) and travel directly toward known feeding trees to inspect them (Asensio et al. 2011; Janmaat et al. 2013). Thus, the knowledge that a species provides ripe fruits is a benefit, as it allows navigation to ripe fruits, given knowledge of where trees of this species are located. There is no evidence that primates use fruit odor in this way, although one study found that chimpanzees (Pan troglodytes verus) are especially drawn to large trees that tend to emit strong odors when ripening (Janmaat et al. 2013). One way to test the hypothesis that primates are attracted to odorous fruiting trees over long distances would be a controlled release of odors of natural fruits in times in which they are not naturally ripe. The hypothesis would be supported if primates change their ranging be behavior to increase inspections of individuals of the given plant species. However, while reproduction of fruit odor in the lab is feasible (see Nevo et al. 2015), controlled emissions in sufficiently large amounts in natural settings may prove more challenging.
In addition to phenological information, strong odors can help a group of animals to detect the individual tree and thus promote seed dispersal. In this case, there is no need for prior knowledge of the tree’s location. In interactions in which the seed disperser can scent-track an individual tree, a strong odor that can attract seed dispersers over long distances is more likely to be selected than in the previous scenario because it would increase an individual’s own fitness rather than risk attracting seed dispersers to other conspecifics. Long-distance attraction and direction of pollination agents through floral scent is common in flowers (Valenta et al. 2017). This, however, requires an ability to scent-track the odor’s source, and there is so far no evidence that primates use long-distance olfactory cues to locate feeding trees (Nevo and Heymann 2015). Thus, while the absence of evidence may derive from the difficulty of testing this hypothesis, it is unlikely that fruit odor attracts primates over long distances.
Ripeness of an individual fruit within a tree: Over shorter distances, fruit odor may allow primates to determine whether an individual fruit in a patch is ripe or not. Several species of primates sniff individual fruits (Dominy et al. 2016; Hiramatsu et al. 2009; Melin et al. 2009) and the use of olfaction to detect the quality of individual fruits seems to be ubiquitous among primates (Nevo and Heymann 2015). The main prediction of this hypothesis is that the selection pressure exerted by primate feeding behavior would select for ripe fruits to emit an odor that is not necessarily strong or pleasant but is substantially different from the odor of conspecific unripe fruits (Nevo et al. 2016). For example, in a community of Malagasy species of which many are consumed by lemurs, ripe and unripe fruits were found to emit comparable amounts of odor (Valenta et al. 2016b). The hypothesis that odor change in ripeness is an evolved signal that allows primates to identify ripe fruits has been tested in two studies, which showed that the odor profiles of two primate-dispersed fruits from the Peruvian Amazon change on ripening (Nevo et al. 2016) and that spider monkeys (Ateles geoffroyi) can discriminate between odor profiles of ripe and unripe fruits and use this information to reliably choose ripe fruits (Nevo et al. 2015). Odor changes in relation to ripeness were not observed in two bird-dispersed fruiting species, suggesting that a change in odor profile is not ubiquitous in ripe fruits and may be unique to fruits whose main seed dispersal vector is likely to use it (Nevo et al. 2016). Thus, although we need more studies of both behavior and fruit chemistry, the available evidence supports the hypothesis that fruits consumed by primates provide olfactory signals for ripeness that may be the result of a primate-generated selection pressure.
Fruit nutritional content: In addition to phenological information, location, and ripeness, fruit odor may provide reliable information regarding fruit quality. In other words, fruit odor may give information about fruit nutrient content. Reliable odor signals would likely require a direct biochemical association with specific nutrient traits. There is evidence for the evolution of reliable odor signals of flower quality (Knauer and Schiestl 2015), and the same has been found for fruit color (Schaefer et al. 2014), even in the absence of biochemical links between signal and trait.
The most prominent fruit odorant that has been suggested to convey reliable information of fruit quality is ethanol (Dudley 2000, 2002, 2004). Ethanol, or ethyl alcohol, is the product of sugar fermentation by microbes, and it has been suggested that ethanol levels are proportional to sugar content (Dudley 2000, 2002, 2004; Peris et al. 2017), thus providing a reliable indication of fruit nutritional quality. A positive correlation between sugar and ethanol levels has been found in some fruits (Dominy 2004; Sánchez et al. 2004, 2006), and a recent behavioral study found that two species of nectrarivorous primates can discriminate between varying concentrations of ethanol and prefer higher ethanol concentrations (Gochman et al. 2016). Moreover, a recent genetic study demonstrated that the human ability to effectively digest ethanol was acquired by 10 mybp and is shared with all African great apes (Carrigan et al. 2015), indicating that the capacity to metabolize alcohol lies deep in our evolutionary roots. However, despite growing evidence for preference for ethanol in some contexts, evidence for its attraction, i.e., its use as a foraging or food selection cue, is scarce. For example, fruit bats are not attracted to low ethanol levels and are deterred by higher concentrations, which may allow them to avoid overripe fruits (Sánchez et al. 2004, 2006). We are not familiar with any study that has shown that primates are attracted to ethanol in fruits.
In addition to ethanol, a few common plant and fruit VOCs may provide reliable signals for fruit quality. A prime candidate would be ethyl esters: VOCs that are synthesized by a reaction of ethanol and an aliphatic acid (Beekwilder et al. 2004). Flowers have been suggested to upregulate ester synthesis to increase the volatility of some compounds and thus odor profile (Dudareva et al. 2004). Crucially, some fruits increase ester emission when they are ripe and the limiting factor in ester synthesis is the amount of alcohol available (Beekwilder et al. 2004). As a result, the amount of ethyl esters in fruit odor bouquets may be tightly linked to ethanol production and hence sugar levels. For example, in oranges, a cultivated fruit, yeast infection increases ethanol and ester levels and increases the fruit’s attractiveness to animals (Peris et al. 2017). Esters are common in many fruits consumed by primates in Madagascar and Uganda (Fig. 1), which would be excellent model systems to test this hypothesis.
Terpenoids may also be somewhat correlated with sugar levels. Terpenoids are the most diverse group of plant secondary compounds (Gershenzon and Dudareva 2007). The main building block of terpenoids is isoprene (McGarvey and Croteau 1995), a compound whose emission from leaves correlates positively with photosynthetic activity (Lerdau and Throop 2000). In addition, terpenoid biosynthesis is relatively expensive (Gershenzon 1994). Thus, terpenoid amounts in fruits have the potential to provide a costly and reliable signal of fruit quality and sugar levels. Terpenoids are ubiquitous in fruits and have been found in fruits consumed by bats (Hodgkison et al. 2013) and primates (Fig. 1; Nevo et al. 2016). Thus, there are multiple model plant species on which this hypothesis can be tested.
Another hypothesis involving terpenoids reflects the fact that some of these compounds are synthesized through degradation of carotenoids (Knudsen et al. 2006), a group of chemicals that convey a variety of health benefits in humans (Johnson 2002). In flowers of the mustard family (Brassica spp.), up- and downregulation of a single gene drives conversion of carotenoids to terpene VOCs (Zhang et al. 2015). Thus, terpene concentration in fruit odor may be directly linked to the amount of carotenoids in the fruit.
A final class of compounds that may provide reliable signals for fruit quality is nitrogen- and sulfur-containing compounds. These VOCs are far less common in fruits and many are probably defensive compounds (Farmer 2014). Nitrogen- and sulfur-containing compounds are a product of protein metabolism (Knudsen et al. 2006) and therefore their presence in fruit odor may be directly related to protein levels. Protein is a limiting factor in Madagascar (Donati et al. 2017; Ganzhorn et al. 2009), and indeed, fruits there tend to be richer in nitrogen- and sulfur-containing compounds than fruits in Uganda (Fig. 1).
In conclusion, the hypothesis that fruit odor conveys reliable information regarding fruit quality remains untested. However, the emission rates of several VOC classes that are common in primate-consumed fruits may be correlated with the amount of desired macronutrients in fruits. This generates a set of straightforward predictions that will hopefully be addressed in the coming years.
Regional Differences in the Odor of Primate-Consumed Fruits
So far, we have considered primates to be a more or less homogeneous group whose interactions with fruits are similar across systems. Physiological studies that quantified olfactory sensitivity and discrimination capacities found no substantial differences between catarrhines and platyrrhines (e.g., Laska et al. 2000, 2005). However, these studies did not include strepsirrhines, which possess relatively larger olfactory-related neuroanatomic structures (Baron et al. 1983; Stephan et al. 1981) and a larger number of functional OR genes (Matsui et al. 2010). Further, OR genes have experienced a rapid birth-and-death process since the last common primate ancestor, which means that lineages have acquired OR gene repertoires that may allow them to detect or discriminate among different VOCs (Dong et al. 2009; Gilad et al. 2005). Thus, there may be significant variation in olfactory ability among primates, which may translate into differences in the degree of use of olfaction and also the selection pressures on fruits.
Another factor that may drive differences in the selection pressures primates might have exerted on fruit traits is their tendency to rely on visual cues. First, primates vary in their ability to detect color: Strepsirrhines tend to be monochromatic, dichromatic, or show a population-level polymorphism in which some females and males are dichromats whereas a minority of females are trichromats. A polymorphism in color vision phenotypes is almost ubiquitous among platyrrhines, with the exceptions of the monochromatic (and nocturnal) owl monkeys (Aotus spp.) and the routinely trichromatic howlers (Alouatta spp.). Finally, all catarrhines are fully trichromatic (Jacobs 2009). In addition, many strepsirrhines are nocturnal and are less likely to rely on visual cues and signals than diurnal species (Valenta et al. 2016a). It is possible that the more species tend to rely on visual cues such as fruit color (chroma: hue, saturation), the less they use olfaction (Gilad et al. 2004; Nevo and Heymann 2015). For example, within species, dichromat capuchins (Cebus capucinus) are less efficient foragers and tend to rely more on olfaction than trichromats (Melin et al. 2009, 2017). If this pattern also occurs between species, it raises the possibility that plant communities that interact with less visual frugivores would be under stronger selection to signal ripeness through olfactory cues. For example, in Madagascar plants rely on the dispersal services of lemurs (Wright et al. 2005). Since both a nocturnal activity pattern and routine dichromacy are more common in lemurs, Madagascan fruits may have been under strong selection to favor olfactory signals than have plant species that interact with monkeys and apes. However, intraspecific differences in reliance on olfaction between di- and trichromats is not ubiquitous (e.g., Hiramatsu et al. 2009) and more broadly, the (in)famous tradeoff between olfaction and vision in primates has been based on questionable genetic and neuroanatomical proxies (Laska et al. 2005; Nevo and Heymann 2015; Niimura 2012). In particular, the hypothesis that trichromacy is directly linked to lower olfactory capabilities (Gilad et al. 2004) has been refuted (Matsui et al. 2010) and diurnal trichromats also rely on olfactory cues (Dominy et al. 2016).
In addition, regardless of sensory capacities, although frugivorous primates tend to be similar in the sense that they eat a wide range of fruits, species may have slightly different needs. For example, protein is a stronger limiting factor in Madagascar than it is in the Neotropics (Donati et al. 2017; Ganzhorn et al. 2009), which may drive lemurs to be especially attracted to protein-rich fruits and hence exert selection pressure for emission of VOCs indicating the presence of protein.
If regional differences in primate foraging behavior and sensory anatomy result in significant differences in the selection pressures experienced by fruits, primates may generate observable differences in fruit odor across systems. The nature of the differences is likely to be tightly connected to the function fruit odor fulfills in the feeding ecology of different primates. That is, a higher importance of fruit odor in Madagascar may lead to stronger odors if the main function of fruit odor is to serve as a phenological cue or attractant over longer distances, to a larger difference between odors of conspecific ripe and unripe fruits if the main function is ripeness signaling, and to higher emission rates of nitrogen- or sulfur-containing compounds if fruits signal protein content.
Unfortunately, at this point, this hypothesis remains untested, as few studies have characterized odors of fruits across primate habitats. Our own preliminary data indicate some qualitative differences between fruits in continental Africa and Madagascar, namely a stronger emphasis on aliphatic esters and nitrogen- or sulfur-containing compounds in Madagascar (Fig. 1). This raises the possibility that fruit VOCs do signal sugar or protein levels in Madagascar. A good test of this hypothesis would require the combination of these data with nutritional analysis and behavioral experiments to test whether lemurs draw this information from fruit odor.
Confounding Factors in the Evolution of Fruit Odor
Fruit odor is a complex trait affected by a multitude of selection pressures and constraints. Thus, just like any other plant or fruit trait, there is likely to be a measurable gap between how we predict a fruit will smell based on primate sensory and feeding ecology and how the fruits on which they feed actually smell. Among others, three main confounding factors are 1) other functions that ripe fruit VOCs fulfill; 2) developmental constraints; and 3) the interaction of fruits with nonprimate seed dispersers.
First, fruit odor may fulfill functions unrelated to attracting seed dispersers. Secondary metabolites in ripe fruit have been proposed to play a role in the defense of fruits against predators and abiotic stress, help regulate seed germination, and regulate intake by frugivores and their gut retention time to maximize seed establishment probability (Cipollini and Levey 1997). These hypotheses are supported and may apply to some fruit VOCs (Cipollini and Levey 1997; Nevo et al. 2016). To further complicate matters, some VOCs serve as attractants as well as fulfilling other functions, e.g., seed predator deterrence (Nevo et al. 2017; Pellmyr and Thien 1986). Moreover, different VOCs can fulfill different functions in different species, tissues, or contexts. For example, terpenoids play an important role in leaf defense (Farmer 2014), but upregulation of terpene emission in orange fruits renders them more susceptible to attack (Rodríguez et al. 2011).
Second, the presence of some VOCs may be due to developmental constraints rather than adaptation. In other words, they may be a remnant from unripe fruits or flowers that developed into the ripe fruits, or a “leak” from leaves (Eriksson and Ehrlén 1998). The hypothesis that these are the main factors determining fruit secondary metabolites has been refuted as it became increasingly clear that ripe fruit secondary metabolite profiles differ both quantitatively and qualitatively from other plant organs (Cipollini et al. 2004; Rodríguez et al. 2013; Whitehead and Bowers 2013). However, such constraints are still likely to have some effect on ripe fruit VOC profile and consequently affect the way fruits smell.
Finally, some of the fruits that primates consume are also consumed and dispersed by other frugivore guilds such as bats, birds, and other arboreal mammals. The latter may share many traits with primates and hence exert similar selection pressures. However, other frugivores possess substantially different sensory capacities and employ different foraging and food selection strategies. For example, birds tend to have excellent color vision based on four different opsin genes, as opposed to the three present in the best color-discriminating primates (Vorobyev et al. 1998). There is substantial evidence that fruits that rely on birds for seed dispersal advertise ripeness using color (Lomáscolo and Schaefer 2010). Thus, fruits dispersed by primates that also rely on the dispersal services of birds are more likely to invest in color signals, which may be accompanied with a reduction in the olfactory signal (Valenta et al. 2015a). For example, in continental Africa primates tend to disperse seeds of fruits that are also dispersed by other animals, primarily birds, whereas in the Neotropics primates and birds form more discrete dispersal syndromes (cf. Flörchinger et al. 2010; Gautier-Hion et al. 1985; Janson 1983; Link and Stevenson 2004). This predicts that plants consumed by primates in continental Africa invest in visual cues such as color contrast against the background leaves to signal to birds, and that if primates can rely on these signals (Lomáscolo and Schaefer 2010), these plants would experience a relaxation of selection pressures to signal through odor. As a result, differences in the nonprimate frugivore community may drive intersystem differences in the odor of primate-consumed fruits.
Methodological Issues in the Study of Fruit Odor
What to sample: We can use more than one fruit to enhance the signal and increase the probability of recording minor VOCs during fruit odorant sampling. However, the results should be corrected for either the number of individual fruits or their overall surface area, depending on the question to be addressed. It would be helpful if all studies report both the number of fruits and the surface area sampled so that future studies can calculate standardized measures across studies.
- 2.Sampling protocol: The two main procedures in headspace sampling are dynamic headspace systems, in which constant airflow carries VOCs from their source to the VOC trap, and static sampling, in which the system is closed and VOCs emitted are slowly absorbed by the probe (Tholl et al. 2006). Applying these two different procedures to the same sample may yield different results, both qualitatively and quantitatively. Our experience indicates that a nearly static procedure in which odor builds up for ca. 2 h and is then pumped onto the probe for 10 min retrieves most VOCs while minimizing the introduction of contaminants that results from dynamic headspace sampling (Fig. 2). In addition, it allows researchers to sample more than a single sample at a time using a single pump, which may be useful in field conditions.
VOC trap: There are many different traps and adsorbent media on the market, each suitable for different conditions. For example, solid-phase microextraction (SPME) is commonly used because it is reusable and allows easy static headspace sampling and immediate analysis (Tholl et al. 2006). Its biggest disadvantage for researchers working at remote field sites is that a single probe is very costly, and thus, if analysis in the field is impossible, it becomes difficult to collect large samples. Recently, portable GC-MS systems have become commercially available, which may allow sampling and immediate analysis in the field. However, when compared in field conditions, these systems appear to be substantially less sensitive and provide poorer discrimination owing to the short column length compared to established systems (Kücklich et al. 2017).
We use the chromatoprobe system (Fig. 2) (Dötterl and Jürgens 2005; Dötterl et al. 2005). These probes are inexpensive because they can be made in house and comparative studies show that they are often superior to commercially available probes (Dormont et al. 2013) or other techniques, including a portable GC-MS (Kücklich et al. 2017). Other probes are likely to produce comparable results. However, this is true only for probes that can be analyzed using thermal desorption technique, in which the entire sample is injected and analyzed, as opposed to probes that require extraction with an organic solvent before analysis (Tholl et al. 2006). The latter tend to dilute the sample and, in addition to the loss of lower-concentration compounds, further complicate quantitative interstudy comparisons.
GC-MS analysis: While each GC-MS system is different and yields slightly different outputs, a few adjustments can ensure that results from different studies are comparable. First, the choice of capillary column installed in the GC can affect results. Most plant volatiles can be readily analyzed using a nonpolar column, e.g., DB-5 (Adams 2007), which is our column of choice. Further, GC-MS system outputs quantify the amounts of VOCs in area units that are fully internal and incomparable to other systems. Professional chemists use various methods to convert these outputs to absolute amounts (e.g., Lavagnini and Magno 2007), which is an impractical procedure for biologists dealing with hundreds of different VOCs in their samples. However, the raw results can be semistandardized by analyzing a known amount of a standard compound and dividing the raw output results of each VOC by that of the standard. This will reduce some of the noise originating from differences in the sensitivity of GC-MS systems.
Primates are now known to possess a reasonably good sense of smell, which plays a role in their feeding ecology. They provide essential dispersal services to many plants. Both flowers and fruits have evolved in response to selection pressures exerted on them by their main pollination and seed dispersal vectors. Thus, the use of olfaction in feeding ecology is likely to have affected the evolution of fruit odor. Although this hypothesis has received growing interest and support in recent years, only a few studies have attempted to test it and even fewer conducted chemical profiling of fruits consumed by primates. As a result, many of the questions raised here cannot yet be answered. Nonetheless, both theoretical considerations and the available data indicate that fruit odor is a promising domain whose study can shed new light on the ecology and evolution of both plants and primates. We hope that this review provokes future studies that will address these and other questions and help establish the field of primate chemical ecology.
We thank Hiroki Sato, Laurence Culot, Yamato Tsuji, and Onja Razafindratsima for inviting us to contribute to this special issue. We also thank Joanna Setchell, Onja Razafindratsima, and three anonymous reviewers for helpful comments on a previous draft of this manuscript. O. Nevo was funded by a German Science Foundation grant (NE 2156/1-1) while working on this manuscript.
- Adams, R. P. (2007). Identification of essential oil components by gas chromatography/mass spectrometry (4th ed.). Carol Streams, IL: Allured.Google Scholar
- Beekwilder, J., Alvarez-Huerta, M., Neef, E., Verstappen, F. W. A., Bouwmeester, H. J., & Aharoni, A. (2004). Functional characterization of enzymes forming volatile esters from strawberry and banana. Plant Physiology, 135, 1865–1878. https://doi.org/10.1104/pp.104.042580.PubMedPubMedCentralCrossRefGoogle Scholar
- Borges, R. M. (2015). Fruit and seed volatiles: Multiple stage settings, actors and props in an evolutionary play. Journal of the Indian Institute of Science, 95, 93–104.Google Scholar
- Carrigan, M. A., Uryasev, O., Frye, C. B., Eckman, B. L., Myers, C. R., et al (2015). Hominids adapted to metabolize ethanol long before human-directed fermentation. Proceedings of the National Academy of Sciences of the USA, 112, 458–463. https://doi.org/10.1073/pnas.1404167111.PubMedCrossRefGoogle Scholar
- Chapman, C. A., & Russo, S. E. (2007). Linking behavioral ecology with forest community structure. In C. J. Campbell, A. Fuentes, K. C. MacKinnon, M. Panger, & S. K. Bearder (Eds.), Primates in perspective (pp. 510–525). New York: Oxford University Press.Google Scholar
- Cipollini, M. L., Paulk, E., Mink, K., Vaughn, K., & Fischer, T. (2004). Defense tradeoffs in fleshy fruits: Effects of resource variation on growth, reproduction, and fruit secondary chemistry in Solanum Carolinense. Journal of Chemical Ecology, 30(1), 1–17. https://doi.org/10.1023/B:JOEC.0000013179.45661.68.PubMedCrossRefGoogle Scholar
- Dormont, L., Bessière, J.-M., McKey, D., & Cohuet, A. (2013). New methods for field collection of human skin volatiles and perspectives for their application in the chemical ecology of human–pathogen–vector interactions. The Journal of Experimental Biology, 216, 2783–2788. https://doi.org/10.1242/jeb.085936.PubMedCrossRefGoogle Scholar
- Farmer, E. E. (2014). Leaf defence. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199671441.001.0001.CrossRefGoogle Scholar
- Fobes, J. L., & King, J. E. (1982). Vision: The dominant primate modality. In J. L. Fobes & J. E. King (Eds.), Primate behavior (pp. 219–243). New York: Academic Press.Google Scholar
- Hodgkison, R., Ayasse, M., Kalko, E. K. V., Häberlein, C., Schulz, S., et al (2007). Chemical ecology of fruit bat foraging behavior in relation to the fruit odors of two species of Paleotropical bat-dispersed figs (Ficus hispida and Ficus scortechinii). Journal of Chemical Ecology, 33, 2097–2110. https://doi.org/10.1007/s10886-007-9367-1.PubMedCrossRefGoogle Scholar
- Hodgkison, R., Ayasse, M., Häberlein, C., Schulz, S., Zubaid, A., et al (2013). Fruit bats and bat fruits: The evolution of fruit scent in relation to the foraging behaviour of bats in the new and old world tropics. Functional Ecology, 27, 1075–1084. https://doi.org/10.1111/1365-2435.12101.CrossRefGoogle Scholar
- Knudsen, J. T., Eriksson, R., Gershenzon, J., & Ståhl, B. (2006). Diversity and distribution of floral scent. The Botanical Review, 72, 1–120. https://doi.org/10.1663/0006-8101(2006)72[1:DADOFS]2.0.CO;2.Google Scholar
- Laska, M., Genzel, D., & Wieser, A. (2005). The number of functional olfactory receptor genes and the relative size of olfactory brain structures are poor predictors of olfactory discrimination performance with enantiomers. Chemical Senses, 30, 171–175. https://doi.org/10.1093/chemse/bji013.PubMedCrossRefGoogle Scholar
- Laska, M., Höfelmann, D., Huber, D., & Schumacher, M. (2006a). The frequency of occurrence of acyclic monoterpene alcohols in the chemical environment does not determine olfactory sensitivity in nonhuman primates. Journal of Chemical Ecology, 32, 1317–1331. https://doi.org/10.1007/s10886-006-9090-3.PubMedCrossRefGoogle Scholar
- Laska, M., Bautista, R. M. R., Höfelmann, D., Sterlemann, V., & Hernandez Salazar, L. T. (2007). Olfactory sensitivity for putrefaction-associated thiols and indols in three species of non-human primate. The Journal of Experimental Biology, 210, 4169–4178. https://doi.org/10.1242/jeb.012237.PubMedCrossRefGoogle Scholar
- Lerdau, M., & Throop, H. L. (2000). Sources of variability in isoprene emission and photosynthesis in two species of tropical wet forest trees. Biotropica, 32, 670–676. https://doi.org/10.1646/0006-3606(2000)032[0670:SOVIIE]2.0.CO;2.Google Scholar
- Niimura, Y., Matsui, A., & Touhara, K. (2014). Extreme expansion of the olfactory receptor gene repertoire in African elephants and evolutionary dynamics of orthologous gene groups in 13 placental mammals. Genome Research, 24, 1485–1496. https://doi.org/10.1101/gr.169532.113.PubMedPubMedCentralCrossRefGoogle Scholar
- Raguso, R. A. (2008). Wake up and smell the roses: The ecology and evolution of floral scent. Annual Review of Ecology, Evolution, and Systematics, 39, 549–569. https://doi.org/10.1146/annurev.ecolsys.38.091206.095601.CrossRefGoogle Scholar
- Rodríguez, A., San Andrés, V., Cervera, M., Redondo, A., Alquézar, B., et al (2011). Terpene down-regulation in orange reveals the role of fruit aromas in mediating interactions with insect herbivores and pathogens. Plant Physiology, 156, 793–802. https://doi.org/10.1104/pp.111.176545.PubMedPubMedCentralCrossRefGoogle Scholar
- Steiger, S. S., Fidler, A. E., Valcu, M., & Kempenaers, B. (2008). Avian olfactory receptor gene repertoires: Evidence for a well-developed sense of smell in birds? Proceedings of the Royal Society of London B: Biological Science, 275, 2309–2317. https://doi.org/10.1098/rspb.2008.0607.CrossRefGoogle Scholar
- Valenta, K., Brown, K. A., Rafaliarison, R. R., Styler, S. A., Jackson, D., et al (2015b). Sensory integration during foraging: The importance of fruit hardness, colour, and odour to brown lemurs. Behavioral Ecology and Sociobiology. https://doi.org/10.1007/s00265-015-1998-6.
- Valenta, K., Miller, C. N., Monckton, S. K., Melin, A. D., Lehman, S. M., et al (2016b). Fruit ripening signals and cues in a Madagascan dry forest: Haptic indicators reliably indicate fruit ripeness to dichromatic lemurs. Evolutionary Biology, 43, 344–355. https://doi.org/10.1007/s11692-016-9374-7.CrossRefGoogle Scholar