Abstract
Fruits result from complex biological processes that begin soon after fertilization. Among these processes are cell division and expansion, accumulation of secondary metabolites, and an increase in carbohydrate biosynthesis. Later fruit ripening is accomplished by chlorophyll degradation and cell wall lysis. Fruit maturation is an essential step to optimize seed dispersal, and is controlled by a complex network of transcription factors and genetic regulators that are strongly influenced by phytohormones. Abscisic acid (ABA) and ethylene are the major regulators of ripening and senescence in both dry and fleshy fruits, as demonstrated by numerous ripening-defective mutants, effects of exogenous hormone application, and transcriptome analyses. While ethylene is the best characterized player in the final step of a fruit’s life, ABA also has a key regulatory role, promoting ethylene production and acting as a stress-related hormone in response to drought and pathogen attack. In this review, we focus on the role of ABA and ethylene in relation to the interconnected biotic and abiotic phenomena that affect ripening and senescence. We integrate and discuss the most recent data available regarding these biological processes, which are crucial for post-harvest fruit conservation and for food safety.
Introduction
Seeds and fruits communicate
Angiosperms are the most successful land plants: more than 400 000 species are included in the division Magnoliophyta (Pimm and Joppa, 2015). Fruits are one of the innovations that explain the quick and sudden colonization of the earth by flowering plants (Knapp and Litt, 2013).
Fruit formation is triggered by signals, most probably produced by the female gametophyte, that communicate to the plant that fertilization has occurred (Vivian-Smith et al., 2001). Consequently, ovules turn into seeds and pistils reactivate their growth to host the developing seeds harbouring the new generation (van Doorn and Woltering, 2008). Additionally, it is known that developing seeds promote cell division and expansion within the fruit by the production of hormones (Vivian-Smith et al., 2001). In normal conditions, the successful completion of pollination and fertilization is a pivotal process for fruit-set determination and initiation of fruit growth. In the absence of fertilization, pistils undergo senescence (Carbonell-Bejerano et al., 2010, 2011).
Fruits protect the developing seeds and ensure seed dispersal, and to this end they have evolved many mechanisms to optimize dissemination of seeds. Dry dehiscent fruits mechanically disperse the seeds (Spence et al., 1996; Pabón-Mora and Litt, 2011; Seymour et al., 2013). Fleshy fruits develop tasty tissues that induce fauna to eat them, and consequently disperse the seeds.
Ripening initiates after the conclusion of the seed maturation process and is a developmental feature unique to fruit. During fruit ripening, metabolites are converted into sugars and acids, whilst in senescing leaves metabolites are mobilized and delivered to the fruit (Gillaspy et al., 1993). Indeed, fleshy fruit tissues undergo changes in organoleptic characteristics such as colour, texture, and flavour that made them appealing to frugivorous animals. These events lay the foundations for the mutualism between Magnoliophyta species that produce fleshy fruits and the animals that eat them and contribute to their seed dispersal (Tiffney, 2004; Seymour et al., 2013; Duan et al., 2014).
The crosstalk between seeds and fruits is important in early development as well as during maturation phases that are modulated by seeds. This crosstalk can be deduced from the comparison of seedless and seeded fruits (Mazzucato et al., 1998; Acciarri et al., 2002; Hershkovitz et al., 2011). At early stages, the number of developing seeds influences the final size and weight of the fruit, because the developing embryos control the rate of cell division and promote cell expansion in the surrounding fruit tissues (Gillaspy et al., 1993; Gouthu and Deluc, 2015).
Interestingly, a recent study implicates an important seed–fruit signalling pathway in the opposite direction, from mother plant to progeny, during late developmental stages. Seeds of flowering locus t (ft) mutants display altered seed coat flavonoid content and seed dormancy. FT is expressed in the silique, and its expression, sharply controlled by temperature, measures the seasonal fluctuations. It has been proposed that FT can act as a messenger able to record environmental conditions and transfer such information to the seeds (Chen et al., 2014).
Carpel patterning anticipates fruit architecture
Fruits derive mostly from the fertilized mature gynoecium, which is the female reproductive part of a flower located in the innermost whorl of flowers and composed of one or more pistils. The pistil is the female reproductive unit and is comprised of one or more carpels, which enclose and protect the ovules. However, especially in fleshy fruits, additional floral components are frequently recruited to form the fruit (Esau, 1960; Fait et al., 2008). Fruit morphology and function depend to a great extent on gynoecium patterning, and this is especially true for dry fruits (Seymour et al., 2013). Carpel identity is determined by the product of class C homeotic genes. The Arabidopsis class C gene is AGAMOUS (AG; Yanofsky et al., 1990; Becker and Theissen, 2003), while in tomato there are two AG-like genes, TAG and TAGL1 (TOMATO AG and TOMATO AG-LIKE 1; Pnueli, 1994a,, b; Itkin et al., 2009; Vrebalov et al., 2009; Giménez et al., 2010). TAGL1 silencing does not affect floral organ specification but alters ripening. However, TAGL1 overexpression determines the swelling of the sepals. In the Arlequin (Alq) mutant, TAGL1 is ectopically expressed as a consequence of a gain-of-function mutation. Consequently, sepals are converted into pistil-like structures able to turn into fleshy organs, thus confirming that TAGL1 performs a class C function (Vrebalov et al., 2009; Giménez et al., 2010; Pan et al., 2010; Zhao et al., 2018).
A pistil may consist of a single carpel or of several fused carpels. The main functional modules in the pistil are as follows: (i) the stigma, formed by specialized cells for pollen reception and germination; (ii) the style, a narrow extension of the ovary, connecting it to the stigmatic papillae; sometimes it is missing, defining a sessile stigma; and, finally (iii) the ovary, a chamber that contains the ovules (Roeder and Yanofsky, 2006). In a transverse section of the ovary, several features are detectable depending on the number of carpels and the type of placentation. In dry dehiscent fruits, dehiscence zones differentiate after fertilization from carpel margins (less frequently in other positions). The dehiscence zones open when fruits are ready to release the seeds (Dong et al., 2014).
From a molecular point of view, the regulatory network for carpel patterning has been studied in detail in Arabidopsis thaliana. Briefly, YABBY (YAB) genes, expressed in the lateral domains of the developing gynoecium, up-regulate the MADS-box genes FRUITFULL (FUL) and SHATTERPROOF1 and 2 (SHP1/2; Dinneny, 2005; Colombo et al., 2010). The activity of SHPs is confined to the valve margins, where it specifies the dehiscence zone (Liljegren et al., 2004).
In contrast to Arabidopsis, studies on carpel and fruit patterning in other species are still scarce. Fleshy fruits, for which tomato is considered the reference species, lack the distinct organization in the dehiscence zone and valves. Tomato fruits consist of two or more fused carpels forming locules separated by fleshy septa, with seeds protruding into the locules from a central placenta. The carpel walls form the pericarp during fruit development and grow through cell division, followed by cell expansion (Gillaspy et al., 1993).
Hormonal control of fruit ripening
Processes underlying the formation and the progression of fruits life are the subject of intense study, since fruit maturation, and the consequent seed dispersal, is the ultimate developmental objective of a plant. Moreover, the comprehension of ripening is as yet a relevant unreached goal in science, for improving post-harvest conditions faced along the entire food chain, from the field to the customers. Fruits are important food sources, and the reduction of their spoilage is a big challenge to prevent food waste, to ensure safer food, and to strive for environmental sustainability.
The molecular network controlling fruit maturation in Arabidopsis is largely unexplored, and to date few genes involved in the regulation of silique senescence have been identified. Recently, our group published the transcriptome of developing Arabidopsis silique valves to shed light on the pathways controlling fruit growth and maturation (Mizzotti et al., 2018). Previously it had been demonstrated that AtNAP (NAC-LIKE, ACTIVATED BY AP3/PI, NAC029) is a transcription factor belonging to the NAC family (NAC stands for NAM/ATAF/CUC), that controls the progression of silique senescence (Kou et al., 2012). The tomato AtNAP orthologue is NON-RIPENING (NOR; Guo and Gan, 2006; Kou et al., 2012). NOR also represses fruit ripening (Tigchelaar et al., 1973), suggesting a conserved function for this gene among dry and fleshy fruits (Gómez et al., 2014). The transcriptional regulation of tomato ripening has been better clarified than that of Arabidopsis and it involves several players, among them MADS-box (MADS stands for MCM1/AGAMOUS/DEFICIENS/SRF), HD-Zip (Homeodomain-leucine zipper), and AP2/ERF (APETALA2/Ethylene Response Factor) transcription factors that modulate an intricate regulatory network. Beside NOR, two other transcription factors, Colorless Non-Ripening (CNR, a SQUAMOSA promoter-binding type protein) and Ripening Inhibitor (RIN, a MADS-box transcription factor), act early in fruit development and orchestrate the expression of genes involved in ethylene production (Vrebalov et al., 2002; Giovannoni, 2004; Manning et al., 2006).
Fruit life is also notably affected by phytohormones, such as auxins [indole-3-acetic acid (IAA)], cytokinins (CKs), jasmonic acid (JA), abscisic acid (ABA), brassinosteroids (BRs), and ethylene (reviewed by McAtee et al., 2013; Kumar et al., 2014). Hormone molecules regulate fruit set, development, maturation, and ripening, and each step is generally modulated by two or three hormones simultaneously. The combined action of auxins, gibberellins (GAs), and CKs is the major regulator of fruit set (Dorcey et al., 2009; Mariotti et al., 2011; Ruan et al., 2012). Auxins and CKs modulate fruit development (Yang et al., 2002; de Jong et al., 2011; Kumar et al., 2011; Pattison and Catalá, 2012), although auxins also trigger fruit maturation (Davey and Van Staden, 1978; Sorefan et al., 2009; Devoghalaere et al., 2012; Kumar et al., 2012). ABA and ethylene are the main ripening regulators (Fedoroff, 2002; Giovannoni, 2004; Setha, 2012; McAtee et al., 2013; Kumar et al., 2014). More information is available about ethylene, since it plays a pivotal role in fruit ripening and it has been considered for several years the master regulator of fruit maturation (Bapat et al., 2010).
Recently it has been demonstrated that ABA is an important ripening-associated hormone, and its action is transversal since it accumulates, in both fleshy and dry fruits, preceding the ethylene peak in the ripening phase (Buesa et al., 1994; Kojima et al., 1995; Kondo and Inoue, 1997; Kanno et al., 2010; Sun et al., 2012a; Leng et al., 2014). The molecular regulation of ripening in dry and fleshy fruits highlights strong similarities in both fruit types, suggesting its conservation throughout the angiosperms (Seymour et al., 2013; Kumar et al., 2014). In dry and fleshy fruits, ripening relies mostly on hormones such as ABA and ethylene (McAtee et al., 2013). In the next sections, the role of these two hormones in the different fruit typologies (dry and fleshy, climacteric and non-climacteric fruits) will be examined in more depth (Fig. 1).
Fruit development and ripening are modulated by phytohormones and are integrated with cell wall re-organization and pathogen susceptibility. The progression of the life of a fruit is represented using the Arabidopsis thaliana silique (above) as reference for dry fruits and Solanum lycopersicum (tomato) for fleshy fruits (below). The pattern of changes in key hormones during the fruit’s life is described. While development requires hormones such as auxins, cytokinins and gibberellins, ripening and senescence principally rely on abscisic acid and ethylene. Their fluctuations are similar for both dry, fleshy climacteric and fleshy non-climacteric fruits. Cell wall remodelling is a common process of all ripening fruits, since it guarantees the tissue softening and the seed dispersal. Similarly, the susceptibility to pathogen attacks increases after maturation and therefore influences the post-harvest conservation of the products.
Dry and fleshy fruits: differences and common aspects
Dry fruits grow after fertilization and, once development is completed, they activate a senescence programme. In dry dehiscent fruits, seeds are dispersed after the differentiation of the dehiscence zone and the progression of cell separation (Spence et al., 1996; Pabón-Mora and Litt, 2011; Seymour et al., 2013; Gómez et al., 2014). Many fleshy fruits lignify the endocarp (the innermost epidermal layer; Karlova et al., 2014) but the rest of the pericarp expands. The pericarp accumulates sugars, after the conversion of complex carbohydrates, and secondary metabolites such as carotenoids and anthocyanins. Fleshy fruits also achieve colour change through chlorophyll degradation. All of these changes aim to attract animals that will promote the biotic dispersion of seeds (McAtee et al., 2013; Seymour et al., 2013).
Both fruit types trigger the hydrolysis of specific cell walls (Brummell, 2006; Klee and Giovannoni, 2011; Seymour et al., 2013). In dry fruits, cell wall metabolism causes the formation of dehiscence zones and the fruit splits, while, in fleshy fruits, tissues become softer and less resistant.
Paleo-botanical reconstructions suggest that fleshy fruit-producing species most probably evolved from dry fruit-producing species, since fleshiness as we mean it today appeared later in the history of angiosperms (Eriksson et al., 2000; Friis et al., 2010). This hypothesis is further supported by phylogenetic analyses. In the Rosaceae and Solanaceae, fleshy fruit-producing species evolved from species that produced dry fruits (Knapp, 2002; Xiang et al., 2017). Also in the Campanulidae, dry, dehiscent, multiseeded fruits, or capsules, are the ancestral form, although this occurs relatively rarely in the group (Beaulieu and Donoghue, 2013).
The molecular network controlling pistil and fruit patterning has been well elucidated in Arabidopsis, which, although considered most representative of the Brassicaceae family, emerged as the reference plant for dry fruits (Gómez et al., 2014; Łangowski et al., 2016; Provart et al., 2016). Comparative studies demonstrate that it is possible to transfer information from Arabidopsis silique development to other species whose fruits are also dry. For instance, in the genus Medicago, some species have coiled pods with increased valve margin lignification, which correlates with a change in the protein sequence of SHP orthologues (Fourquin et al., 2013). In soybean, pod dehiscence resistance is modulated by the NAC protein SHATTERING1-5 (SHAT1-5; Dong et al., 2014) which is highly expressed and causes increased secondary cell wall thickening in the fibre cap cells. SHAT1-5 is homologous to Arabidopsis NAC SECONDARY WALL THICKENING PROMOTING FACTOR1/2 (AtNST1/2). The Arabidopsis nst1 mutant fails to lignify the valve margins (Mitsuda and Ohme-Takagi, 2008).
Similarly, we will refer to tomato as the model organism for fleshy fruits (Karlova et al., 2014) although several lines of evidence imply that molecular programmes, cellular modifications, and epigenetic marks are conserved between dry and fleshy fruits (Gómez et al., 2014; Lü et al., 2018).
ABA and ethylene in dry fruit ripening
In Arabidopsis siliques, ripening and senescence are tightly bound to each other, and many authors consider them synonymous (Gapper et al., 2013). Recently, transcriptomic analyses of senescing siliques have underlined genes differentially expressed in this process (Wagstaff et al., 2009; Carbonell-Bejerano et al., 2010; Jaradat et al., 2014; Mizzotti et al., 2018) and have led to the identification of several pathways involved in fruit maturation. This work has pinpointed a pivotal role for gene products that contribute to macromolecule catabolism, chloroplast degradation, and seed protein storage. Moreover, genes related to ABA and ethylene metabolism are over-represented in ageing leaves and siliques. For example, ABSCISIC ACID INSENSITIVE 4 (ABI4), an ethylene-responsive factor (ERF) that also acts in response to ABA, is more highly expressed in older stages. ABI4 participates in plastid-to-nucleus and mitochondrion-to-nucleus retrograde signals (Giraud et al., 2009; León et al., 2013), although this aspect is still controversial (Kacprzak et al., 2019), and modulates ethylene production (Dong et al., 2016). ETHYLENE INSENSITIVE 3 (EIN3), which is abundantly transcribed in senescing siliques, triggers ethylene signal and regulates ABI4 expression (Kou et al., 2012).
Several genes associated with ethylene biosynthesis increase their transcription during Arabidopsis fruit maturation (Wagstaff et al., 2009; Jaradat et al., 2014; Mizzotti et al., 2018), preceded by an accumulation of ABA (Kanno et al., 2010); therefore, for fruit dehiscence, it is nowadays well accepted that maturation is mediated by the two hormones (Child et al., 1998; Kou et al., 2012). ABA also accumulates in leaf and petal tissues, to contrast the drought stress that senescing plants usually have to face (Jaradat et al., 2014). Valves evolved from leaves, and several modules participating in leaf senescence are also conserved during fruit ripening (Wagstaff et al., 2009; Koyama, 2018). For instance, members of the NAC transcription factor family, involved in leaf senescence in response to biotic and abiotic stresses and hormone signal transduction (Shao et al., 2015), are also over-represented in the transcripts of Arabidopsis fruits (Wagstaff et al., 2009; Mizzotti et al., 2018). In contrast, genes involved in CK and GA signal transduction were shown to be down-regulated, confirming the role of ABA and ethylene as senescence-associated hormones in dry fruits.
Although ethylene plays a strong role in Arabidopsis silique ripening, it is not yet clear whether they are classified as climacteric or non-climacteric fruits (see next section; Kou et al., 2012). However, as described below, both types of fruits require the action of ethylene and ABA in order to complete the ripening process.
Fleshy fruits are classified as climacteric or non-climacteric
Ripening of fleshy fruits has always been the focus of intense study because of its relevance in determining the nutritional features that define the overall quality of the fruit (Giovannoni, 2004; Carrari and Fernie, 2006). Depending on the respiration pattern displayed, fleshy fruits can be divided into climacteric fruits, that show an increase in respiration rate with a concomitant ethylene burst during ripening, and non-climacteric fruits, in which there is no increase in the respiration rate and no accumulation of ethylene (reviewed in Cherian et al., 2014).
The first group includes fruits such as tomato, apple, pear, peach, banana, mango, and kiwi (Abdul Shukor et al., 1990; Buesa et al., 1994; White, 2002; Hiwasa et al., 2003; Xu et al., 2008; Kondo et al., 2009; Atkinson et al., 2011; Zaharah et al., 2013), while the second group includes grape, strawberry, cherry, and orange (Kondo and Inoue, 1997; Rodrigo et al., 2003; Trainotti et al., 2005; Deytieux et al., 2007; Koyama et al., 2010). Both climacteric and non-climacteric fruits display the same upstream components of the ethylene signal transduction pathway (Liu et al., 2015) and accumulate ABA at the beginning of ripening (Leng et al., 2014). ABA accumulation precedes and thus modulates ethylene production in climacteric fruits, and triggers maturation in non-climacteric fruits. Very recently, the pivotal role of ABA in non-climacteric fruit was demonstrated in Fragaria ananassa (Liao et al., 2018).
Liu et al. (2015) have hypothesized that, since upstream elements are conserved, the myriad ethylene-related pathways during ripening could be explained by the huge diversity represented by the downstream ERF elements. Indeed, ERF proteins belong to one of the biggest families of transcription factors that could confer specific and variable responses to this hormone. Moreover, Leng et al. (2014) proposed that additional ethylene-independent regulatory factors might co-operate to control ripening in both fruit types, acting upstream of the ethylene signalling pathway. In tomato, ABA and ethylene crosstalk is not yet clearly understood. Transcriptomic analysis (Mou et al., 2016) suggested that ABA triggers ethylene production and response, but on the other hand ethylene itself is needed to maintain ABA production. Additionally, some transcription factor genes involved in ethylene synthesis and sensitivity (e.g. MADS-RIN, TAGL1, CNR, and NOR) are ABA responsive.
ABA, ethylene, and fleshy fruit ripening
As previously stated, ethylene plays a pivotal role in fleshy fruit ripening, and its involvement has been known for decades (Burg and Burg, 1962). Perturbations of ethylene production, perception, or signalling altering ripening have been widely documented (Hamilton et al., 1990; Oeller et al., 1991; Lanahan et al., 1994; Tieman et al., 2001; Lee et al., 2012; Liu et al., 2014). According to the currently accepted model (Liu et al., 2015), ethylene is sensed by specific receptors that trigger a signalling cascade that terminates with the transcription of ERFs. Such responsive factors regulate the progression of senescence-associated processes, leading to the acquisition of the traits typical of mature fruits (Solano and Ecker, 1998; Ju et al., 2012; Chang et al., 2013).
Two different ethylene biosynthetic systems operate in fleshy fruits (McMurchie et al., 1972; Lelièvre et al., 1997). System 1 keeps the synthesis at the basal level and it is present in both climacteric and non-climacteric fruit types. System 1 is autoinhibitory, since the perception of ethylene blocks its synthesis (Barry and Giovannoni, 2007). In fact, the aminocyclopropane-1-carboxylic acid (ACC) synthase 1A (ACS1A) and ACS6 enzymes that produce ethylene are inhibited once ethylene accumulates to a basal level (Liu et al., 2015). In contrast, system 2, active during ripening in climacteric fruits, is autocatalytic and relies on ACS2 and ACS4, which are both regulated by positive feedback of ethylene, as well as on ACC oxidase1 (ACO1) and ACO4 (Nakatsuka et al., 1998; Barry et al., 2000; Van de Poel et al., 2012). In climacteric plants, ethylene is thought to be involved in a crosstalk with IAA, since they accumulate at the same time in numerous fleshy fruits, such as tomato and peach, and auxins up-regulate the transcription of genes whose products mediate ethylene biosynthesis and signalling (Gillaspy et al., 1993; Jones et al., 2002; Trainotti et al., 2007). The IAA might originate from the seeds, which accumulate high concentrations of this hormone, which is then degraded by ripening-associated genes (Kumar et al., 2014). Ethylene plays a role in de-greening, since manipulation of genes related to its biosynthesis and signalling influences the pigmentation of tomato (Karlova et al., 2011; Lee et al., 2012) as well as tissue softening (Xiong et al., 2005; Nishiyama et al., 2007; López-Gómez et al., 2009).
Although ethylene plays a pivotal role, ABA accumulates in both climacteric (Buesa et al., 1994) and non-climacteric fruits, and several authors suggest that ABA might be the major controller of ripening and senescence not only in leaves but also in fruit (Kojima et al., 1995; Kondo and Inoue, 1997; Setha, 2012). Ethylene biosynthesis can be triggered by exogenous application of ABA (Jiang et al., 2000; Sun et al., 2012a), while low ABA concentration delays fruit ripening and precedes the release of ethylene in climacteric fruits (Zhang et al., 2009b), evidence further confirmed by the transient silencing of SlNCED1 (9-cis-epoxycarotenoid dioxygenase) by virus-induced gene silencing in developing fruits (Ji et al., 2014). Moreover, transcription factors, such as RIN, NR, and CNR that trigger ethylene production, are also up-regulated in response to ABA (Mou et al., 2016). Exogenous application of ABA also promotes the production of metabolites associated with senescence, such as anthocyanins, and decreases organic acids (Ban et al., 2003; Cakir et al., 2003; Jeong et al., 2004; Giribaldi et al., 2010), making the fruit more attractive and palatable for frugivorous animals. Further confirmation comes from tomato ABA-deficient mutants which do not display normal growth and ripening (Taylor et al., 2000; Galpaz et al., 2008), and from orange ABA-deficient mutants with delayed peel tissue de-greening (Rodrigo et al., 2003). ABA is thought to be involved in sugar accumulation, an essential process that ensures the palatability of fleshy fruits for seed dispersal and the human diet. In fact, application of ABA causes an increase in sugar uptake into vacuoles in apples (Yamaki and Asakura, 1991), and in the sugar content of citrus (Kojima et al., 1995) and grape (Deluc et al., 2007), and promotes starch hydrolysis in melon (Sun et al., 2012b). Moreover, ABA seems to influence the colour change during ripening, as demonstrated by overpigmentation of tomato mutants, in which ABA levels are lower compared with wild-type fruits (Galpaz et al., 2008). In tomato, the silencing of SlNCED1, whose gene product participates in ABA metabolism, triggers carotenoid accumulation (Sun et al., 2012b), although this is probably due to lower accumulation of ethylene, since pigmentation changes are caused by blocking ethylene production (Chervin et al., 2004).
In strawberry, ABA homeostasis is strictly controlled through the modulation of FveCYP707A4a (cytochrome P450 monooxygenase) expression. ABA accumulates during ripening because FveNCED is enhanced and FveCYP707A4a, which catalyses ABA catabolism, is repressed. Accordingly, the alteration of FveCYP707A4a expression changed the endogenous ABA levels and FveNCED expression (Liao et al., 2018).
Finally, ABA and ethylene promote fruit softness, as demonstrated, for example, in banana (Lohani et al., 2004) and in tomato fruits (Sun et al., 2012a,, b). In banana, ABA treatments sharpen the softening of the fruit in the presence or absence of ethylene, while the ethylene itself is involved in the regulation of cell wall hydrolases (Lohani et al., 2004). In tomato fruits, ABA application induces the production of ethylene, resulting in a softer fruit (Zhang et al., 2009b). Conversely, in cases where ABA is reduced, for instance in SlNCED1-RNAi plants, the decrease in ABA determines an up-regulation of the genes involved in ethylene biosynthesis and perception, and resulting in a final increase in ethylene content (Sun et al., 2012b).
Taken together, these works highlight the essential role of ethylene and ABA in ripening and senescence of fleshy fruits, strengthening the important role of the latter in climacteric fruits (Setha, 2012; McAtee et al., 2013; Kumar et al., 2014; Leng et al., 2014; Shen and Rose, 2014).
Fleshy fruits and cell wall modifications
During ripening, softening and textural changes are caused by fruit cell wall modifications, that impact fruit cell shape, turgor, and size (Fig. 1; Harker et al., 1997).
The cell wall is composed of polysaccharide networks (cellulose microfibrils) formed by the assembly of β-1,4-linked glucans. Microfibrils are rigid elements that interact, via H-bonds, with hemicellulose polysaccharides (linear, neutral sugar-rich polysaccharide backbones with simple lateral groups) and generate the cell wall resistance to applied stress. Cell walls are further stiffened by hemicellulose polysaccharides interacting with two or more microfibrils. The plant cell wall also contains a matrix of pectic polysaccharides, which include homogalacturonan and rhamnogalacturonan, and many proteins and glycoproteins, including enzymes and structural proteins.
During ripening, the matrix of glycans is depolymerized. Such depolymerization has been described in several fruits, such as in strawberry (Posé et al., 2011), tomato (Brummell et al., 1999), hot pepper (Ghosh et al., 2011), melon (Rose et al., 1998), kiwifruit (Wilson et al., 2001), avocado (Huber and O’Donoghue, 1993), persimmon (Cutillas-Iturralde et al., 1994), and peach (Ghiani et al., 2011).
During ripening, some cell wall modifications are species specific. For instance, in plum and cucumber, galactose (Gal) losses are not observed, but, in apple, plum, and apricot, arabinose (Ara) degradation occurs (Gross and Sams, 1984). During fruit ripening in kiwi, tomato, and plum, pectins are depolymerized, a process that is absent in apple and watermelon (Karakurt and Huber, 2002). In strawberry, banana, and apple, the depolymerization of ionically bound pectins does not occur (Airianah et al., 2016), it is very limited in melon (Rose et al., 1998) and massive in avocado and watermelon (Karakurt and Huber, 2002).
Softening and textural changes are catalysed by a multitude of cell wall-localized enzymes. In tomato, fruit softening involves the actions of ripening-related expansins (Tsuchiya et al., 2015) and β-galactosidase (Smith et al., 2002), whereas the solubilization and depolymerization of pectin mediated by endo-polygalacturonase (endo-PG) has little effect on firmness (Goulao and Oliveira, 2008), as demonstrated by endo-PG silencing that favours fruit integrity and longer shelf life (Langley et al., 1994). Recent findings have suggested that the expression of cell wall modification-related genes could be induced by members of the GRAS family of transcription factors, such as SlFSR, whose expression is in turn regulated by ethylene during ripening (Zhang et al., 2018). In bell pepper, pectin depolymerization is undetectable during ripening, but there are important differences between wild and domesticated accessions (Ahmed et al., 2011). Wild accessions soften quickly; domesticated accessions develop firm fruits since they do not produce endo-PG. Attenuated expression of endo-PG and corresponding firmness changes are also reported in peach. Initially, peach softening is quite slow but then accelerates (melting) as a result of increases in soluble pectins and pectin depolymerization (Zhu et al., 2017). In non-melting flesh peaches, the final melting phase is absent, thus fruit remain relatively firm when fully ripe (Porter et al., 2000).
Endo-PG accumulates only in ripening melting varieties just before the melting phase (Orr and Brady, 1993; Paniagua et al., 2014). In non-melting peaches, endo-PG is not detected by specific antibodies as a consequence of genomic deletions or production of truncated transcription products (Lester et al., 1994, 1996; Callahan et al., 2004).
The importance of the cell wall environment is emphasized by the presence of membrane-spanning sensors, wall-associated and receptor-like kinases, WAKs and RLKs (Decreux and Messiaen, 2005; Hématy et al., 2009; Kohorn and Kohorn, 2012), positioned to monitor the wall’s chemical and physical status. It has been shown that leucine-rich repeat (LRR) RLK receptors participate in hormone homeostasis modulation. For instance, the strawberry LRR-RLK Red-Initial Protein Kinase 1 (FaRIPK1) can physically interact with the ABA receptor (ABAR; Hou et al., 2018), also known as the magnesium-chelatase subunit H protein (CHLH; Shen et al., 2006). ABA binds at the C-terminal domain (Wu et al., 2009), but not the other components of the Mg-chelatase complex (Du et al., 2012). Virus-induced gene silencing of FaRIPK1 and FaABAR indicated that both genes promote ripening in a synergistic way.
The role of ABA and ethylene in cell wall modifications
Morphological modifications of the cell wall depend on the activity of several enzymes that change the properties (physical and chemical) and the structure of the cell wall components. Some of these enzymes have been associated with cell wall modifications during ripening and are responsible for the softening of fruit pulp (Tucker et al., 2017). Because cell wall-modifying enzymes are sometimes regulated by ABA or ethylene, these hormones may act during fruit ripening through modification of the cell wall components (Fig. 2). The application of small molecules to plants reversibly perturbs the normal physiological homeostasis and helps shed light on to the molecular mechanisms faster than conventional genetic approaches. Chemical genetics is rapidly advancing our understanding of the role of plant hormones and is also contributing to the identification of novel compounds for commercial applications based on phytohormone agonists and antagonists (Rigal et al., 2014).
Exogenous application of phytohormones is a useful tool to shed light on the pathways they trigger. In fleshy fruits, one of the principal feature of the ripening phase is the cell wall remodelling; cell wall modifications make the fruit palatable for animals and favour seed dispersion. Here we have summarized the behaviour of the most studied cell wall-related genes after the application of the major ripening hormones, ABA and ethylene. Both ABA and ethylene induce a transcriptional increase in cell wall-degrading genes, but, intriguingly, different species selectively up-regulate only some genes. In bilberry, most of the genes involved in cell wall modification increase their transcription after a treatment with ABA and/or ethylene, except for pectin esterases, polygalacturonases, and xylosidases. In the figure, the bilberry appears twice both in the pectin esterases and in the pectin lyases, since pectin lyases are up-regulated whilst pectin esterases are down-regulated by hormone application. Tomato β-galactosidases 5 and 6 (see the section ‘Ethylene application: climacteric fruit cell wall modifications’) also decrease their transcription after the application of ethylene. In peach, according to the stage after full bloom, polygalacturonases and pectin esterases are regulated differently. In the figure, peach no. 1 is 102 days after blooming, while peach no. 2 is between 115 and 118 days after full bloom. Mango fruits also show different expression patterns of polygalacturonases, pectin esterases, and pectin lyases, while the endoglucanases do not show any change in the expression levels after the treatment with ABA. In raspberry, β-galactosidase transcription levels are not influenced by ethylene application.
ABA application: non-climacteric fruit cell wall modifications
ABA application to non-climacteric fruit induces the expression of genes involved in cell wall modification. For instance, ABA application to berries of the red wine grape variety Cabernet Sauvignon triggers the transcription of the xyloglucan endotransglycosylase gene (XET), whose product modifies the cell wall (Giribaldi et al., 2010). In the non-climacteric bilberry fruit (Vaccinium myrtillus L.), ABA treatments enhance the expression of expansins, pectate lyases, rhamnogalacturonate lyases, β-galactosidases, xyloglucan endotransglycosylases/hydrolases, and endo-β-1,4-glucanases (EGs), whilst ABA represses the transcription of pectin esterases, polygalacturonases (PGs), and β-xylosidases (Karppinen et al., 2018). In strawberry fruits, transcriptomic analysis of receptacles under different hormonal and ripening conditions demonstrated that several genes involved in cell wall modification such as pectate lyase B, PGs, EGs, rhamnogalacturonate lyases, and expansins are activated by ABA and repressed by auxins in the receptacle (Medina-Puche et al., 2016).
ABA application: climacteric fruit cell wall modifications
For a long time it was thought that in climacteric fruits ripening is controlled only by ethylene; however, in recent years a pivotal role for ABA in fruit softening has been demonstrated by studies of tomato, peach, melon, and mango (Zhang et al., 2009a, b; Sun et al., 2013; Zaharah et al., 2013). For instance, applications of ABA or an ABA biosynthesis inhibitor (NDGA) to mango fruits demonstrated that ABA stimulates endo-PG activity but represses the pectin esterase activity (Zaharah et al., 2013). Indeed ABA-treated fruits showed lower pectin esterase activity, while NDGA-treated fruits had higher pectin esterase activity. In contrast, endo-PG activity was increased by ABA and reduced by its inhibitor. However, these treatments did not affect exo-polygalacturonase or endo-1,4-β-d-glucanase activity (Zaharah et al., 2013). Instead, in peach, ABA applications modulate cell wall enzymes depending on the stage of the application. For instance, 5 days after the applications at mid-S3 stage [102 days after full bloom (dAFB)], peach fruits exhibited drastically reduced levels of endo-PG, pectin methylesterase (PME) inhibitor, and expansins (Soto et al., 2013). Later during fruit development, ABA application at S3/S4 and S4 fruit stage (115–118 dAFB) induced a significant increase in endo-PG and PME inhibitor levels just 1 day after treatment (Soto et al., 2013).
Further confirmation of the role of ABA in climacteric fruit ripening came from molecular studies of tomato fruit. SlNCED1-RNAi fruits had reduced levels of ABA and an extended shelf life, 2- to 4-fold relative to controls, probably because most of the cell wall catabolic enzymes were poorly transcribed in the transgenic fruits. In particular, the expression levels of expansin (SlEXP1), polygalacturonase (SlPG1), pectin methylesterase (SlPME), β-galactosidase precursor mRNA (SlTBG), endo-1,4-β-cellulose (SlCels), and xyloglucan endotransglycosylase (SlXET16) are reduced during tomato ripening (Sun et al., 2012a; Ji et al., 2014). Moreover, in SlNCED1-silenced lines, the amount of ethylene increased, suggesting control exerted by ABA on fruit ripening and ethylene production. In contrast, silencing of SlCYP707A2, which encodes a protein with ABA 8'-hydroxylase activity, involved in ABA catabolism, causes an up-regulation of cell wall catabolic enzymes. In SlCYP707A2-RNAi fruits, SlNCED1 was up-regulated, as was ABA production, promoting the ripening process. In these lines, SlEXP1, SlPG1, and SlXET16 transcripts accumulate more than in control fruits (Ji et al., 2014).
Ethylene application: non-climacteric fruit cell wall modifications
The clearest differences between climacteric and non-climacteric fruits are determined by the presence or absence of the autocatalytic ethylene system (see above) and the lack of uniformity in the response of non-climacteric fruits to ethylene application. In non-climacteric fruits, several ripening-related indicators respond to the application whilst others do not (Goldschmidt, 1998). Moreover, some fruits, such as guava, melon, Japanese plum, Asian pear, and pepper, behave as climacteric or non-climacteric depending on the cultivar or genotype (Paul et al., 2012).
In grape, ethylene application increases the expression of PGs, xyloglucan endotransglucosylases, PMEs, cellulose synthases, and expansins (Chervin et al., 2008). Expansin expression is complex as transcription varies in different tissues, and is influenced by the treatment duration.
In strawberries, the application of an ethylene perception inhibitor, 1-MCP (1-methylcyclopropene), reduces PG expression and activity (Villarreal et al., 2009). PG1 is also down-regulated in FaCTR1-RNAi fruits, since the ethylene cascade is affected (Sun et al., 2013).
Application of ethylene to raspberry (Rubus idaeus) fruits confirmed that it enhances the activity of PGs, PME, and Cx-cellulase enzymes but does not affect β-galactosidase (Iannetta et al., 1999).
Ethylene application: climacteric fruit cell wall modifications
Transcriptomic analysis of papaya fruits treated with ethylene revealed that the expression of cell wall-related genes (PGs, β-galactosidase, pectate lyase, PME, β-glucosidase, xyloglucan endotransglucosylase, endoglucanase 8-like, endoxylanase, β-d-xylosidase 5, and expansin A) is higher with respect to untreated control samples (Shen et al., 2017).
In avocado fruits treated with 1-MCP, the pectin metylesterase activity was maintained at high levels for a longer time in comparison with untreated control fruits (Jeong and Huber, 2004). Also EGase activity and α- and β-galactosidase are affected: the increase in EGase activity is delayed in the treated samples, while the typical decline in α- and β-galactosidase activity is delayed in the 1-MCP-treated samples (Jeong and Huber, 2004).
The role of ethylene on cell wall enzyme activity in tomato has been studied for decades. For instance, the effect of ethylene treatment on PG activity in tomato was defined in 1983 by Grierson and Tucker (1983) who demonstrated that exogenous ethylene application stimulated the synthesis of PGs, while an environment with low levels of ethylene caused a delay in PGs synthesis (Grierson and Tucker, 1983). In tomato and banana, ethylene application activated expansin1 (EXP1; Rose et al., 1997; Trivedi and Nath, 2004) while in apple and tomato, ethylene increased xyloglucan endotransglucosylase/hydrolase (XTH) activity. Ethylene induces this surge by enhancing the level of expression of 15 different SlXTH and three MdXTH genes and, among these genes, SlXTH5 and SlXTH8 in tomato and MdXTH10 in apple are ripening associated (Muñoz-Bertomeu et al., 2013). The relationship between tomato β-galactosidases (TBGs) and ethylene has been demonstrated by ethylene treatment of fruit at 35 days after pollination at the mature green stage in the wild type and three ripening-impaired mutants: rin, nor, and Never ripe (Nr; Moctezuma et al., 2003). While the level of TBG4 mRNA increased in ethylene-treated fruit, TBG5 and TBG6 levels decreased after the application. The same trend was also recorded in ripening-impaired mutants: TBG4 transcription increased in treated rin and Nr mutant fruits, but was not affected in the nor mutant. In contrast, TBG5 and TBG6 transcription was decreased in all the ripening-impaired mutants. Further studies, using different time points for ethylene exposure, revealed that the up-regulation observed in TBG4 upon the treatment was an indirect response to the hormone application, rather than a primary or direct response (Moctezuma et al., 2003).
Fruit and pathogens
Ripe fleshy fruits are more susceptible to disease and decomposition than unripe green fruits (Fig. 1; Prusky, 1996). The increased susceptibility of ripe fruits to opportunistic pathogens in nature facilitates the dispersal of mature seeds (Gillaspy et al., 1993), but causes important fruit losses when the fruits have the highest economic value, and chemical control strategies are strictly limited. An understanding of the specific ripening events associated with this susceptibility has a relevant economic impact on fruit production and commercialization, facilitating the development of commodities that ripen acceptably, with extended shelf life and less prone to pathogen infections.
The plant cell wall is an important barrier to be circumvented by pathogens, and the breaching of the cell wall triggers plant responses to counteract the pathogen infection (Cantu et al., 2008).
More than 200 plant species can be attacked by Botrytis cinerea, an opportunistic aggressive ascomycete that causes grey mould rot on different organs (fruits, stems, flowers, and leaves). Botrytis can infect many crops such as tomato, berries, chickpeas, French beans, and grapes, as well as cut flowers. Like many other fungal pathogens, B. cinerea secretes a large set of extracellular enzymes to degrade plant cell wall polymers to infect the host organs. In the B. cinerea secretome there are PGs, PMEs, proteases, and laccases (ten Have et al., 1998, 2001; Kars et al., 2005). Nevertheless, B. cinerea cannot diffuse when disassembly of the endogenous fruit cell wall is impaired (Cantu et al., 2008). Key evidence has been obtained with tomato: the silencing of LeExp1 or of LePG does not prevent B. cinerea infection, but the simultaneous down-regulation of both LeExp1 and LePG causes a reduced susceptibility to the pathogen.
Ripening in tomato fruit is regulated by ethylene and transcription factors, including NOR, RIN, and CNR (Vrebalov et al., 2002; Manning et al., 2006). The disruption of these genes affects fruit ripening, delaying the maturation in a manner similar to what occurs when the ethylene receptor LeETR3 is abolished, causing the NEVER RIPE phenotype (Chang and Shockey, 1999). All these mutants fail to produce significant amounts of ethylene, and therefore the fruits maintain a stronger firmness and accumulate fewer carotenoids. However, despite the fact that nor and rin fruits have similar features, nor fruits are more susceptible than rin to B. cinerea. In rin fruits, B. cinerea is able to trigger the transcription of LeExp1 and LePG, but not in nor fruits. In agreement with these observations, the application of 1-MCP, an ethylene perception inhibitor, prevents fruit ripening but treated fruits are still susceptible to B. cinerea (Díaz et al., 2002).
It is interesting to observe that in tomato leaves some mechanisms have developed to counteract B. cinerea, using the hormonal response networks such as those of ethylene, salicylic acid, and ABA (Ferrari et al., 2003; Glazebrook, 2005; AbuQamar et al., 2006; Asselbergh et al., 2007). These pathways are also triggered by B. cinerea in ripe and unripe fruits, and their activation accelerates fruit ripening. Recently, Sun and co-workers (Sun et al., 2018) showed that the ethylene response factor gene SlPti4 is involved in the response to B. cinerea through the regulation of ABA levels in fruit and seeds, thus influencing both ripening and germination (Sun et al., 2018).
β-glucosidase (BG) hydrolyses ABA-glucose ester and releases active ABA, thus participating actively in ABA homeostasis. BG genes are expressed in ripening fruits, and in strawberry the down-regulation of FaBG3 delays maturation. The delay occurs because the transcription of genes whose products are involved in cell wall catabolism, anthocyanin synthesis pathway, aroma-related genes, and sugar metabolism is not triggered (Molina-Hidalgo et al., 2013). Transgenic fruits are also less susceptible to B. cinerea attacks, most probably because the cell wall integrity is preserved.
PMEs catalyse the demethylesterification of homogalacturonans and produce acidic pectins and methanol (Pelloux et al., 2007). PMEs cross-link pectins by calcium bridges, causing wall stiffening. However, PMEs are also responsible for cell wall loosening (Micheli, 2001). Four PME genes (FaPE1–FaPE4) are present in strawberry. FaPE1 is fruit specific (Castillejo et al., 2004) and it is triggered by auxin at the onset of fruit ripening and suppressed by ethylene during fruit senescence (Castillejo et al., 2004). Overexpression of FaPE1 caused a 20% reduction in the methyl esterification of soluble and chelated pectins (Osorio et al., 2008). The transgenic fruits displayed enhanced resistance to B. cinerea since a pathogenesis-related gene involved in the salicylic acid pathway is constituously expressed as a consequence of the lower degree of methyl esterification of oligogalacturonides (Osorio et al., 2008), small pectins responsible for several cellular responses, including fruit ripening (Dumville and Fry, 2000).
Conclusions
Fruit ripening maximizes seed dispersal through meticulous co-ordination of a network of genetic and biophysical processes. In this review, we summarized present knowledge about the mechanisms modulating these complex developmental processes. Ripening, once started, is an irreversible process that can only be delayed.
Ripening is accomplished by cellular modifications; here we have focused our attention on cell wall modification. A deeper knowledge of ripening mechanisms, as well as the associated cell wall modifications, will help the improvement of post-harvest protocols and prevention of pathogen infections.
Acknowledgements
We thank Sonia Balestri for the graphical help with the figures, and Edward Kiegle for critical reading of the manuscript. The work has been supported by Ministero dell’Istruzione, dell’Università e della Ricerca (PRIN ISIDE; grant no. 2015BPM9H3_005) to SM. We also acknowledge the support of the SEB for making it possible for the authors to attend the symposium and generate this article. Finally, we apologize to all the researchers whose work could not be cited due to space limitations.
Comments