Aromas and Flavours of Fruits Aromas and Flavours of Fruits

Aromas and flavours play an important role in horticultural crops’ quality, namely in fruits. Plant breeders have made considerable advances producing cultivars with higher yields, resistant to pests and diseases, or with high nutritional quality, without paying enough attention to flavour quality. Indeed, consumers have the perception that fruit aromas and flavours have declined in the last years. Attention is given nowadays not only to flavoured compounds but also to compounds with antioxidant activity such as phenolic compounds. Fruit flavour is a combination of aroma and taste sensations. Conjugation of sugars, acids, phenolics, and hundreds of volatile compounds contribute to the fruit flavour. However, flavour and aroma depend on the variety, edaphoclimatic conditions, agronomical practices and postharvest handling. This chapter reviews the aromas and flavours of the most important fruits and discusses the most recent advances in the genomics, biochemistry and biotechnology of aromas and flavours.


Introduction
Quality in horticulture can be defined as the traits of a given commodity, regardless of its yield [1]. Here, we not only include visual appearance, ability to endure postharvest processing but also chemical and nutritional composition and flavour. Great advances have been made in horticultural breeding, obtaining fruits with characteristics that are those that growers (e.g. yield, composition [10]. This limiting step has been studied and the formation of volatile compounds can be significantly increased, both qualitatively and quantitatively, if fruits are incubated in vitro with adequate metabolic precursors [11]. Some of the fruits with a higher amount of production and more commonly consumed worldwide are apples, bananas, cherries, oranges and grapes, which are shortly addressed here. In apples, over 300 volatile compounds were described [12], although they can be considered cultivar specific [13] and maturation dependent, from aldehydes to alcohols and esters [14]. The latter chemical class is predominant in ripe apples, and straight and branched esters can be found, namely ethyl, butyl and hexyl acetates, butanoates and hexanoates [15]. There is a clear increase of volatile compound production in apple skin, rather than in the internal tissues, due to a higher abundance of fatty acid substrates or increased metabolic activity [16]. The relative amount of each compound is, as referred earlier, linked to a specific cultivar and cannot only be used for cultivar discrimination but also to monitor ripening of fruits [17]. In apples, branched chain esters are produced from the breakdown of leucine, isoleucine and valine, while straight chain esters are synthesised from membrane lipids [18]. The hydroperoxides that result from these reactions are converted to aldehydes, then to alcohols and finally to esters. This sequence leads to the flavour of immature apples ('green notes') due to C6 aldehydes and alcohols to the 'fruity notes' given by the increased ester production [19]. For banana, about 250 volatile compounds have been described, although the really odorant are less than 40 [20]. Olfactometric methods have described several aromas and linked those to some compounds, namely 'banana' to 3-methylbutyl esters and acetate esters, 'grassy' to aldehydes and alcohols and 'spicy' to phenols [20,21]. Major volatile compounds that contribute to banana aroma are volatile esters, such as isoamyl acetate and isobutyl acetate [22] but also isoamyl alcohol, butyl acetate and elemicine [23]. As for other fruits, the ripening process changes the volatile profile, with increased concentration of acetates and butanoates [24] and is cultivar dependent [25]. Recently, Bugaud and Alter [26] have found that 3-methybutyl esters were the most abundant volatile compounds, with 2-methylpropyl butanoate and 3-methylbutyl butanoate linked to 'banana' note; the presence of 3-methyl acetate to 'fermented' and 'chemical' notes, while the presence of 'grassy' (freshly cut green grass) aroma decreased as the total amount of volatiles increased with ripening, namely esters. For cherries, over 100 volatile compounds have been identified, including free and glycosidically volatile compounds, belonging to the chemical classes of carbonyls, alcohols, acids, esters, terpenes and norisoprenoids [27]. Major compounds include hexanal, (E)-2 hexenal and benzaldehyde and are associated with green/grassy notes. For some cultivars, other minor compounds gain increased importance, due to their low odour detection threshold such as (Z)-3-hexenal, decanal, nonanal, (E,Z)-2,6-nonadienal and (E,E)-2,4-nonadienal in 'Lapins', 'Rainier', 'Stella', 'Hongdeng' and 'Zhifuhong' cultivars [28,29]. Some ketones have also been found in cherries, although they have relatively low importance in overall aroma [28], while alcohols, being the most abundant benzyl alcohol, 1-hexanol and (E)-2-hexen-1-ol, are responsible for green notes and the fresh green odour. Additionally, 1-Octen-3-ol has been described as one of the most predominant volatile compound in 'Hongdeng', 'Hongyan' and 'Rainier' sweet cherry cultivars [29]. The content of esters in cherries increases during ripening, but their relative abundance is low. The most common are ethyl acetate, butyl acetate, hexyl acetate, (Z)-2-hexenyl acetate and (E)-2-hexenyl acetate, with methyl benzoate described as among the most powerful volatiles in some cultivars, such as 'Rainier' [28]. Terpenoid compounds are also present in cherries at low levels, limonene, linalool and geranylacetone being the most common [30]. However, when analysing the glycosidically bound aroma compounds in three sweet cherry cultivars ('Hongdeng', 'Hongyan' and 'Rainier'), Wen et al. [29] show that terpenoids are the second major class, after alcohols. In oranges, more than 300 volatile compounds have been reported, the major ones being limonene, β-myrcene and linalool [31], but valencene can also be of great importance, depending on the cultivar [32]. However, these compounds, although representing the large majority of the volatiles, are not the ones more responsible for the aroma, as their contribution is limited due of high odour-detection thresholds. Other minor compounds, like aldehydes (octanal, decanal, undecanal, (Z)-3-hexenal and (E)-2-decenal), esters (ethyl butanoate, ethyl 2-methylbutanoate and ethyl isobutyrate) and other terpenes (β-sinensal, geranial and neral) are those with a significance for the overall flavour of oranges [31]. Most of the grape cultivars have no scent, although the wines obtained from them are full of aromas [33,34]. A great number of compounds have been recorded, including monoterpenes, C13 norisoprenoids, alcohols, esters and carbonyls [35,36]. If linalool and geraniol have been identified as major aroma compounds in both red and white grapes [37], the volatile profile can be useful for the discrimination of grape cultivars [36]. Major free volatile compounds are hexanal, (E)-2-hexenal [36] while glycosidically bound include terpene and benzenic glycosides [34]. In more aromatic grape cultivars, like Muscat, major free compounds include linalool, geraniol, citronellol, nerol, 3,7-dimethyl-1,5-octadien-3,7-diol and 3,7-dimethyl-1,7-octadien-3,6-diol while those glycosidically bound were geraniol, linalool, citral, nerol, citronellol, α-terpineol, diendiol I, diendiol II, trans-furan linalool oxide, cis-furan linalool oxide, benzyl alcohol and 2-phenylethanol. Other monoterpenes that can also add to Muscat aroma were rose oxide, citral, geraniol, nerol and citronellol [38]. As for the other fruits, the volatile profile of grapes changes during ripening, and apparently a greater number of volatile compounds exist pre-veraison than post-veraison, as recorded for Riesling and Cabernet Sauvignon grapes, that also recorded differences (esters and aldehydes were the major class of compounds from Riesling grapes and alcohols for Cabernet Sauvignon) at veraison ( Table 1) [39].
Although the flavour of fruits is the interaction of taste and aroma, the chemical composition of fruits (organic acids, sugars, amino acids, pro-vitamins, minerals and salts) can also influence aroma perception and ultimately, flavour. For sugars, glucose, sucrose and fructose are the most important sugars affecting the perception of sweetness (ranking fructose > sucrose > glucose) [40] and their proportion in a given fruit will change flavour. However, this relationship is not completely understood, as measurement of sugars as soluble solids, in orange, does correlate to sweetness but in mango does not [4]. The main organic acids in fruits are malic, citric and tartaric, citric being the most sour and tartaric the least [40]. Citric acid is linked to citrus fruits, tartaric to grapes and malic to apples, and they are responsible for the sour flavour detected on those fruits. Other fruits, like melon or banana, have reduced acidity [41]. The presence of minerals and salts can change the perception of acidity, by combining with organic acids, influencing the buffering capacity [40]. Many research studies on the flavour of fruits give us a good overview of this particular trait of these commodities. However, much is still to be done, since many cultivars are yet still less studied. Furthermore, the link between taste and aroma compounds and the consumer perception of those is still not well understood, and this should be the ultimately goal to achieve consumer-oriented commodities.

Nuts
Global consumption of nuts grew in the last years and it is expected to grow continuously on a yearly basis [48]. In 2015, almonds, cashews, walnuts and hazelnuts were the most preferred nuts by the consumers [49] but other nuts, such as pine nuts, pecans, chestnuts, Brazil nuts, macadamias and pistachios, are also an appreciated food, especially in the regions where they are regularly produced. They are generally consumed as whole nuts (fresh, roasted or salted) or used in a variety of commercial products and processed food [50]. Europe and North America are the largest nut consumer regions, accounting for almost 50% of the worldwide consumption [48]. Nuts have been a regular part of the human diet since pre-agricultural times [51] due to their nutritional value, sensory properties [49] and potential health properties [50,52], and their consumption can reduce cardiovascular disease risk, the incidence of cancer and type 2 diabetes mellitus [53], as well as obesity and ageing effects [54].
Nut quality related to consumer purchase decisions is based on nut appearance such as size, colour, cleanness and freedom from decay and defects [55] but textural properties [54,56] such as aroma and flavour also play an important role in consumer acceptability [57]. Sweetness, oiliness and roasted flavour are commonly associated with good overall nut sensory attributes [55], some compounds generated during the roasting process responsible for the typical nut flavour [58]. Roasting is a common practice used by the nut industry and involves several physical-chemical processes [59], which can modify the odour, flavour and quality of the final product [60], including negative effects, such as rancidity [61].
In general, nuts are characterised by their high content in unsaturated fatty acids [49,50,57] which make them highly sensitive to oxidation during the roasting process leading to the formation of harmful free radicals [61], which are responsible for undesirable odours and flavours [62]. As a result, the roasting process negatively affects the nutritional quality of nuts but also may influence both the formation of health-promoting components and those with potentially adverse health effects [63]. So, selecting the appropriate roasting conditions, mainly temperature and time, is crucial for achieving higher nut quality [55], which is also dependent on the genotype. For example, in walnuts, roasting treatments under 180°C, for 20 min, produced 17 times higher levels of compounds that indicated oxidation, when compared to raw walnuts [63]. In comparison, the compounds that indicated oxidation only increased by 1.8 times for hazelnuts and 2.5 times for pistachios [63]. According to the same authors [63], the roasting process at low/middle temperatures (120-160°C) preserves constitutional compounds and sensory properties of different nuts (macadamia nuts, hazelnuts, almonds, pistachios and walnuts). Nevertheless, as it occurs with other foods, the characteristic flavour of nuts is dependent on the volatile compounds.
During roasting and other heat processes, additional volatile compounds are formed from reactions among food compounds. In roasted nuts, a wide range of volatiles contribute to the typical and desirable roast flavour. According to Xiao et al. [64], in raw almonds, a total of 41 volatile compounds were identified, including aldehydes, ketones, alcohols, pyrazines and other volatile compounds. The benzaldehyde was the predominant volatile compound present in the raw samples and is associated with a marzipan-like flavour [64]. Roasting resulted in about a 90% decrease in the benzaldehyde level and in the formation of up to 17 new volatile compounds that were not found in raw almonds. Many of these compounds are typically generated during the complex and well-known Maillard (non-enzymatic browning) reaction that occurs during roasting. Volatile compounds like pyrazines, furans and pyrroles have been previously identified as key compounds of roasted almond aroma and concentration of many of these volatile compounds increased with roasting time [64]. It was theorised that one of the reasons for the uncertainty surrounding the characterisation of the 'nutty' term is that nuts have aroma qualities that may be typical to only their own species and that there is no common aroma quality present among all nuts [65]. In a research conducted by Clark and Nursten [66], over 200 aroma compounds were identified as having nutty aromas. This work indicated benzaldehyde, 3,4-methylenedioxybenzaldehyde and 4-methylbenzaldehyde as responsible for nutty aromas in almonds, while 2,4-octadienal and 4-pheynyl-4-pentenal were linked to the same attribute in walnuts and 2-ethyl-3-methylpyrazine in roasted peanuts. In the harvest year, edaphoclimatic conditions of orchards and storage conditions have also been mentioned as key factors determining overall nut quality. In order to evaluate the influence of time and temperature conditions on the oxidative degradation of hazelnuts, Ghirardello et al. [63] observed that storage of nuts at low temperatures reduced the effects of lipid oxidation during 8 months, but refrigeration was necessary to preserve high nut quality for up to 1 year.
In grapes, volatile aroma compounds are found both as 'free' and as 'bound' to a sugar moiety, if 'bound', they are not odour active, but, upon hydrolysis of the glycoside, they may then be volatilised [77]. The amount of 'free' volatile aroma compounds makes it possible to classify the grape cultivars into neutrals or aromatic [78,79]. The aromatic grape cultivars presented a varietal character resulting from higher concentration in the amount of 'free' volatile aroma compounds, namely terpenes, norisoprenoids and isoprenoids [80]. The importance of these 'free' volatile aroma compounds is related not only to their high concentrations but also to their lower perception thresholds. Therefore, grape flavour depends on the content and composition of several groups of compounds [81]. Among the compounds responsible for the aromatic quality are monoterpenes and C 13 -norisoprenoids. These compounds are indigenous from the grape and responsible for intense fruity and floral attributes in wines, contributing to the wine varietal aroma [82][83][84]. Other volatile compounds present in grapes are terpene hydrocarbons, pyrazines [38,[85][86][87] and some C 6 −aldehydes and alcohols [88].
During ripening, grapes develop a characteristic flavour and/or aroma by synthesising volatile compounds [89,90]. For example, linalool and geraniol have been shown to contribute to the aroma of 'Concord' grapes, closely resembling the aroma of methyl anthranilate [91,92]. The aroma compounds, which are secondary metabolites of the plant metabolism, are distributed between the pulp and skin of the grape berry, with the highest concentration in the grape skin [92,93]. Wu et al. [94] characterised the aromas of table grapes, and they found that in 20 grape cultivars, a total of 67 volatile compounds, 61 in the mesocarp and 64 in the skin and the total contents of volatiles of mesocarp and skin largely depended on the levels of esters and terpenes, respectively (Figure 1).
Vitis labrusca and Vitis rotundifolia cultivars have a distinct and pronounced odour; the foxy aroma of V. labrusca is attributed to methyl anthranilate [95]. Chemical compounds originated from several sources contribute to wine aroma. Grape volatile aroma compounds, such as monoterpenes, C 13 -norisoprenoids, methoxypyrazines and thiols, if present, are of major importance for the wine varietal character [96]. The volatile compounds found in wine presented different sensory attributes like fruits such as cherry, pear or passion fruit [97]. As an example, in Figure 2, different fruit flavours' attributes perceived in red and white wines are identified.

Genomics, biochemistry and biotechnology of aromas and flavour
As already mentioned, the flavour of fruits is a complex set of interactions between two main sensations: taste and aroma [2]. Taste is mainly a set of sweet and sour sensations linked to the presence of sugars and organic acids (although other minor compounds affect bitterness, astringency or saltiness). However, the aroma is usually the predominant sensation, surpassing taste [98]. Indeed, if taste sensations, detected in mouth, are recognised by six classes of receptors (sweet, sour, salty, bitter, umami and fat-taste), for flavour complexity, where the olfactory system is essential, 350 olfactory receptor genes are known in humans [1].
The known decrease in flavour of fruits is strongly connected to the pressure on the producers: they are usually paid depending on physical characteristics of fruits (size, shape and colour) but not to chemical traits, so the selection of cultivars is performed to enhance those qualities; the ripening of fruits is delayed as much as possible to make sure that they are able to withstand harvest, handling, storage and shipping without damages, but without a normal ripening, flavour sensations decreased [99]. Considering that flavour perception relies on the interaction of a considerable amount of compounds, it makes it one of the most challenging quality attributes to manipulate, which has led to a reduced attention given to this theme [40]. However, consumers' pressure is growing to bring back the typical flavour of old horticultural commodities, where the flavour sensations were almost instantly detected by odour, followed by the recognition of taste.
To achieve the goal of horticultural commodities of full flavour, some strategies can be followed, including changes in agricultural practices but also genetics tools, using the information on the known pathways of formation of those compounds linked to taste and aroma. Considering the first approach, one should cite the preharvest factors such as genome or growing conditions, harvest maturity or postharvest storage like those important in the final flavour of any horticultural commodity [40]. Some of them are somewhat easy to control (growers are able to choose the cultivar, cultural practices and postharvest procedures), while others, such as climate conditions, are outside human influence. The choice of the cultivar to grow and its link to flavour and how chemical components in the plant tissue are expressed are connected to genetic backgrounds [99]. Indeed, recent works comparing cultivars of sweet cherry [100], peach [101], gooseberry [102], fig [103] or pear [104], to cite a few, show how genetic backgrounds can influence chemical composition and ultimately flavour, recognised by sensory evaluation. However, although genetics have a major role when determining the flavour of freshly harvested fruits, the gene expression can be modified by pre-and postharvest factors [105] (Figure 3), as recently reported for peach [106]. Included in those preharvest factors are weather, soil preparation and cultivation, soil type, irrigation, fertilisation practices and crop loads, while for postharvest, it should be mentioned that storage temperature management, packaging under controlled or modified atmosphere, the use of edible coating, heat or physicochemical treatments are the factors [107]. The next step on flavour research was given when information on biosynthesis was obtained by using molecular and biochemical approaches. Knowing the metabolic pathways, namely the genes involved and the associated enzymes but also the regulatory elements (hormones and transcription factors) or which mechanisms are implicated in the storage or sequestration of volatile precursors, is key in allowing a biotechnological approach to their manipulation [108]. The genes that are linked to flavour can be mostly divided into two categories: those encoding for enzymes and those responsible for factors regulating pathway output [1]. If the knowledge for synthesis pathways and genes for those enzymes responsible has been increasing rapidly, the regulation of metabolic pathway output is not well understood, and the number of genes involved may be quite large, as found for strawberry, where 70 quantitative trait loci (QTLs) affecting volatiles and their precursors have been identified [109] or mandarin (206 QTLs) [110], for instance. As referred earlier, the compounds responsible for aroma can be divided in several classes, the most important being monoterpenes, sesquiterpenes, lipids-, sugars-and amino acid-derived compounds. Knowing how they are biosynthesized and what is involved, when and how are key steps to allow their manipulation. In fact, several steps of aroma volatile biosynthesis for which genes have been characterised and used as targets for genetic transformation are presented in Figure 4 (adapted from [108]). The large part of the available research on the manipulation of flavour has been conducted on tomato, as it is a plant easy to transform, with an associated high economic importance [111] and information regarding this fruit is readily available (e.g. [1]). However, some data regarding other horticultural commodities are available, and some are cited here. For instance, modulation of the soluble sugar content in strawberry has been achieved, by an antisense cDNA of ADP-glucose pyrophosphorylase (AGPase) small subunit (FagpS), a key regulatory enzyme for starch biosynthesis. The down-regulation of the AGPase gene led to an increase of the soluble sugar content, which primarily changed the taste sensation of strawberries but can ultimately also change aroma and flavour, as soluble sugars may be converted into volatile compounds [112]. For orange, the down-regulation of the D-limonene synthase (important as D-limonene is the most abundant volatile component  of all commercially grown citrus fruits) did not affect negatively fruit and juice intensity and discrimination but provided a clear insight to how the combined presence of several volatiles can influence fruit flavour [113]. Considering the compounds that are precursors of aromatic compounds, several strategies can be developed to modify their amounts, and will be shortly addressed, providing few but key examples. For fatty acids are derived into saturated and unsaturated short-chain alcohols, aldehydes and esters by the lipoxygenase (LOX) pathway. The first way to manipulate the volatile composition using fatty acids is to change their amounts present in plant organs. Recently, it was observed in pear that if incubated in vitro with metabolic precursors of volatile compounds, the formation of those was significantly increased, both qualitatively and quantitatively [11]. Some of the enzymes involved in the fatty acid conversion to volatiles can also be tuned to modify the final aroma. For desaturases, they have been identified in strawberry as being responsible for the production of lactones, a group of fatty acid-derived volatiles in peach, plum, pineapple and strawberry [114]. Another group of enzymes, phospholipases, are involved in the formation of polyunsaturated free fatty acids, the substrates for lipoxygenases [115]. The expression of phospholipases can be modified by the use of hexanal-based formulations [116] or by the application of chilling [117,118]. Hydroxyperoxide lyase (HPL) forms very unstable hemiacetals from hydroperoxides generated by LOX, leading to the formation of aldehydes. HPL silencing in potato plants have reduced the content of the C6 compounds in the leaves, while increasing that of C5 [119]. Alcohol dehydrogenase (ADH) catalyses the interconversion of aldehydes and their corresponding alcohols and is a key enzyme in volatile ester biosynthesis [13]. Recent works show that the overexpression of an alcohol dehydrogenase (ADH) from mango led to a change in alcohols and aldehydes related to flavour [120], with previous works also showing that overexpression of an ADH increased the level of alcohols [121]. Alcohol acyl-transferases (AAT) catalyse the transfer of an acyl-CoA to an alcohol, resulting in the synthesis of a wide range of esters [122]. The reduction of AAT expression in apples resulted in reduced levels of key esters in ripe fruit, altered ratios of biosynthetic precursor alcohols and aldehydes, changing in a perceptible way, by sensory analysis, the ripe fruit aroma [123], and recent works show that they may be linked to the volatile ester and phenylpropene production in many different fruits [124]. The volatile formation pathway from amino acids is mainly due to the decarboxylases activity, but few are known to date. The catabolism of melon amino acid aminotransferase and branched-chain amino acid aminotransferase (BCAT) is connected to the amino acid-derived aroma compound formation [125]. Terpenoids are structurally diverse and the most abundant plant secondary metabolites, being of great significance, as they have vast applications in the pharmaceutical, food and cosmetics industries [126], with information regarding volatile terpenoids having been recently reviewed [127]. For carotenoid-derived compounds, the major enzymes involved are carotenoid cleavage dioxygenases (CCD), and the suppression of one gene encoding for CCD leads to the reduction of the production of β-ionone, geranylacetone and pseudoionone [128,129]. Finally, for sugar-derived compounds, information is also available. One of the enzymes responsible for their conversion into volatiles is O-methyltransferase. This enzyme has been overexpressed in strawberries and a reduced expression of its encoding gene (FaOMT) changed furaneol to the 2,5-dimethyl-4-methoxy-3(2H)-furanone (DMMF) ratio, ultimately changing the aroma of the fruit [130]. Many other works have been looking to gain insights into this specific theme, providing important information on how to manipulate aroma and flavour components, and some of those can be found reviewed by Aragüez and Valpuesta Fernández [44] or Dudareva et al. [131].

Conclusions
The quality of horticultural commodities can be assessed in many ways, including by their aroma and flavour. This chapter overviews the large amount of information available regarding these characteristics in fruits. However, all this information is still not enough to fully understand the processes behind the formation of compounds, the interaction of those compounds with each other, but, more importantly, how they will finally influence the consumers' perception of aroma and flavour, and, ultimately, their tendency to buy such commodities. This is true to not only all fruits referred in this chapter but also to those not included here, and a continuous effort to identify volatile and non-volatile compounds for flavour and aroma in understudied species or cultivars must be undertaken. Furthermore, the improvement of flavour and aroma by adequate cultural practices must be achieved without a decline in other quality traits of crops. This must also be the goal of gene manipulation focused in metabolic and regulatory pathways of compound formation. The future appears to be bright concerning flavour in horticultural commodities, as we are likely to see multidisciplinary approaches, from genetic engineering to biochemical and metabolic characterisation, linked to sensory evaluations, which will result in flavour-rich and healthier fruits, with increased interest for both producers and consumers.