Ethylene: a unique plant hormone

Ethylene, the simplest unsaturated hydrocarbon (two carbons with a double bond) is a gaseous plant hormone which regulates essentially all physiological processes during the plant’s life cycle. It is responsible for signaling changes in: seed dormancy and germination, root growth and nodulation, shoot and leaf formation, flower and fruit development, different organs senescence and abscission, plant defense mechanisms, and a number of interactions with other plant hormones (Kępczyński and Kępczyńska 1997; Wang et al. 2002; Chaves and Mello-Farias 2006). Although, ethylene is undoubtedly essential for proper plant growth, development, and survival, it may also be deleterious to plants in some instances. Increased ethylene levels in plants exposed to various types of stress including chilling, heat, nutrient deprivation, anaerobiosis, wounding, and pathogen infection with increased damage to plant growth and health as the result has been reported (Stearns and Glick 2003). There is considerable commercial interest in modifying the amount of ethylene produced under ripening, senescing or stress conditions and thereby creating plants with more robust and/or desirable traits (Kępczyński and Kępczyńska 2005; Handa et al. 2010). The practical implications of ethylene biosynthesis regulation for plant improvement have supported the continuous basic research on dissecting the structure of genes encoding ethylene biosynthetic enzymes, their differential expression patterns and mechanisms underlying their transcriptional activity.

Ethylene biosynthesis begins from S-adenosylmethionine (SAM) and includes two key steps (Fig. 1) as reviewed by Yang and Hoffman (1984) and more recently by Pech et al. (2010). First, 1-aminocyclopropane-1-carboxylate synthase (ACS) [EC 4.4.1.14] catalyses the cyclization of SAM to 1-aminocyclopropane-1-carboxylic acid (ACC), which is often considered the rate-limiting reaction in the pathway. ACS also produces 5′-methylthioadenosine (MTA) which is recycled to regenerate methionine. The final step, oxygen-dependent conversion of ACC to ethylene, is catalyzed by ACC oxidase (ACO) [EC 1.14.17.4]. ACC is converted to ethylene by a modification of carbons C-2 and C-3 of ACC, while C-1 is converted to cyanide and the carboxyl group converted into carbon dioxide.

Fig. 1
figure 1

The key enzymes of ethylene biosynthetic pathway. ACC 1-aminocyclopropane-1-carboxylic acid, ATP adenosine triphosphate, HCN hydrogen cyanide, SAM S-adenosylmethionine (Yang and Hoffman 1984)

Full size image

In this study, we have focused on a specific involvement of ACO genes in the regulation of ethylene synthesis. We aimed to give a comprehensive and up-to-date coverage on their structure, patterns of expression and some regulatory mechanisms underlying their transcriptional activity. We have also discussed the possibilities of exploiting ACO genes in manipulating the level of ethylene production and subsequently controlling plant performance.

Characterization of ACO enzyme

ACC oxidase is a member of ferrous-dependent nonheme oxygenases, most of which utilize 2-oxoglutarate (2OG) as a co-substrate (Mirica and Klinman 2008). In its sequence, two distinct domains can be distinguished; an N-terminally located highly conservative non-heme dioxygenase DIOX_N region and a C-terminally located 2OG-FeII_Oxy region (Fig. 2; Punta et al. 2012). The Fe2OG dioxygenases catalyze a variety of two-electron oxidations including hydroxylations, desaturations, and oxidative ring closures (Costas et al. 2004). Although the sequence homology among these enzymes is not high, all their active sites contain a single ferrous ion bound in a tridentate ligand arrangement referred to as a “2-His-1-carboxylate facial triad” (Mirica and Klinman 2008). Studies by Yoo et al. (2006) revealed that the positive-charged surface in the C-terminal helix of the ACO, spanning the exclusively conserved motif (Lys296–Glu301), could be a major stabilizer in the spatial arrangement of reactants and that the positive-charge network between the active site and C-terminus is critical for ACO activity.

Fig. 2
figure 2

An exemplary domain organization of ACO (AT-ACCO4) polypeptide with DIOX_N (green box) and 2OG-FeII_Oxy (red box) architecture. Pfam-predicted active sites, residues which are directly involved in catalysis of a reaction, for ACO—three iron binding residues (H, D, H) are visualized as lollipops (Punta et al. 2012) (color figure online)

Full size image

Although numerous attempts were made to isolate the ACO, it was not purified to homogeneity because of its low abundance until molecular cloning and heterologous expression were employed as reviewed by Alexander and Grierson (2002). ACO can be eluted as a single, active, hydrophobic protein with molecular mass of about 35–40 kDa after partial purification, which correlates with the length of its polypeptide (about 300–330 AA) and known mRNA sequences with coding sequence in the range of 918–1,002 bp.

The subcellular localization of ACO is still under debate. The lack of a signal peptide classically required for the protein to cross the plasma membrane suggested a cytosolic targeting for ACO (Pech et al. 2010). However, the subcellular localization of ACO is still not clearly defined. Some experiments showed that, in tomato and apple fruit, it is mainly located in the apoplast (Ramassamy et al. 1998) whereas other immunolocalization studies indicated the cytosol as ACO final localization (Chung et al. 2002).

Structure of ACO gene

ACO is encoded by a small multigene family comprised usually of three to four members (Pech et al. 2010; Table S1). The most numerous ACO gene family has been identified in Populus trichocarpa in which genome seven homologs are present. There are many instances that only one ACO gene is known for a certain species so far, however, further studies should inevitably lead to the discovery of some other homologs.

The common ACO gene structure in many different plant species consists of four exons interspersed by three introns (Fig. 3a). The occurrence of such organization has been observed in over 80 % of genes with known sequence of their genomic DNA (Table S1). The coding sequence is divided unequally in exons: the first ~100 bp are located in exon 1 downstream of UTR5′ of varying length, the next ~200 and ~300 bp in exon 2 and 3, respectively. The last section of CDS, also nearly 300 bp, together with UTR3′ region comprises exon 4. The lengths of first three sections of ACO CDS are rather uniform among different species, the forth one is more variable ranging from 232 bp in AT-ACO5 to 327 bp in MD-ACO2. The common structure composed of four exons, however, is not always preserved among all members of the gene family in certain species. There are some exceptions to this rule, in which ACO genes are composed of only three exons interrupted by two introns. Such condensed structure was found, for instance, in two members of melon gene family (CM-ACO2,3) (Lasserre et al. 1996), one gene from peach (PP-ACO2) (Ruperti et al. 2001) and also in Arabidopsis thaliana AT-ACO1, Dianthus caryophyllus DC-ACO, Musa acuminata MA-ACO3, Pinus taeda PT-ACO3, Zea mays ZM-ACO35 (Fig. 3b, c). If we could consider the presence of four exons as a primary feature, the deletion of either third or second intron might have led to the derivative structures composed of only three exons with distinctive CDS length extension in subsequent last or middle exon, respectively.

Fig. 3
figure 3

A schematic presentation of different ACO gene structures. The most common structure composed of four exons (a). Alternative structures comprising three exons (b, c). Denotations: exons are filled blocks, white UTR regions, gray coding sequences; bold lines are introns. 1median value of CDS length; 2mean value of CDS length. AT, Arabidopsis thaliana; CM, Cucumis melo; DC, Dianthus caryophyllus; MA, Musa acuminate; PP, Prunus persica; PT, Pinus taeda; ZM, Zea mays

Full size image

A phylogenetic tree constructed on the basis of ACO amino acid sequences from different seed plant species (gymnosperms, dicots, monocots) illustrates general homology of ACO isoforms (Fig. 4). There is a clear distinction between ACO enzymes and other 2-oxoglutarate dioxygenases which constitute separate divisions. The ACO division can be further divided into two main subdivisions, each comprising several groups and subgroups with the sequences from phylogenetically distant species. The homology analysis reveals that the ACO sequence evolution most probably started with a gene duplication which should have taken place at least before the divergence of monocotyledonous and dicotyledonous species. Several further duplication events together with the speciation processes comprising the accumulation of mutations, created multigenic families of paralogs present in most if not all contemporary seed plants. It is also quite evident that deletions leading to the change in gene structure from four-exonic to three-exinic must have occurred independently several times during speciation.

Fig. 4
figure 4

Phylogenetic tree constructed from a multiple alignment of ACO full amino acid sequences from different species obtained from UniProt or NCBI databases. The analysis was performed on the Phylogeny.fr platform (Dereeper et al. 2008) and comprised the following steps: input sequences alignment with MUSCLE (v3.7) configured for highest accuracy (MUSCLE with default settings); curation of ambiguous regions with Gblocks (v0.91b) (parameters: minimum length of a block after gap cleaning: 10, no gap positions allowed in the final alignment, all segments with contiguous nonconserved positions bigger than eight rejected, minimum number of sequences for a flank position: 85 %); construction of the phylogenetic tree using the maximum likelihood method implemented in the PhyML program (v3.0 aLRT); graphical representation and edition of the phylogenetic tree with TreeDyn (v198.3)

Full size image

Solanaceae

Two species of Solanaceae family, Lycopersicon esculentum and Petunia hybrida, were objects of the pioneering experimental works concerning ACO, better known at that time as the ethylene forming enzyme (EFE). So it is not surprising that their ACO genes are often exploited as references for new isolated orthologs. One of the best characterized ACO gene families is from L. esculentum (tomato). It is one of the largest, comprising five different paralogs: LE-ACO1 (Köck et al. 1991), LE-ACO2 (Holdsworth et al. 1987), LE-ACO3 (Barry et al. 1996), LE-ACO4 (Nakatsuka et al. 1998), LE-ACO5 (Sell and Hehl 2005). The coding regions of LE-ACO1, LE-ACO2, and LE-ACO3 are interrupted by three introns. LE-ACO1 and LE-ACO3 display the highest nucleotide identity (99 %) in exon 3, while LE-ACO2 has lower degree of conservation to these two genes (Barry et al. 1996). Amino acid analysis shows that LE-ACO5 has the highest homology with LE-ACO4 (Sell and Hehl 2005). However, the tomato ACO genes have unique 5′- and 3′-untranslated regions. It was used to analyze gene-specific ACO expression in various tissues at different stages of plant development (Barry et al. 1996). All of the tomato ACO genes are transcriptionally active.

Also in P. hybrida ACO genes comprise a multigene family of four genes PH-ACO1, PH-ACO2, PH-ACO3, and PH-ACO4 (Tang et al. 1993; Table S1). These genes were identified in two tandemly arranged regions. The first region contains PH-ACO1 gene and a pseudogene PH-ACO2, which are separated from each other by a 1.2-kb intergenic region. The genes PH-ACO3 and PH-ACO4 are a part of the second region. In contrast, these two tandem genes are separated by a longer, 6 kb sequence (Tang et al. 1993). The structural organization of all of the characterized petunia ACO genes is identical to that observed in tomato genes (Barry et al. 1996). There are four coding regions interrupted by three introns. The proteins encoded by ACO genes share more than 90 % identity with one another and more than 70 % identity with ACOs from other plant species. However, the highest homology in introns and 3′-untranslated region was noticed only between PH-ACO1 and PH-ACO2. The studies have indicated that three members of petunia ACO gene family are transcriptionally active—PH-ACO1,3,4 (Tang et al. 1993). The fourth PH-ACO2 is postulated to be a pseudogene as it is characterized by specific features—a conversion of G to T disrupting the translation initiation codon and a 2 bp deletion at the beginning of the second exon disrupting the reading frame of this mRNA.

Brassicaceae

The family includes many economically important edible and industrial oilseed, vegetable, condiment, and fodder crop, such as Brassica oleracea (broccoli, cabbage, and cauliflower), Brassica rapa (turnip, Chinese cabbage), Brassica napus (rapeseed), or Sinapis alba (white mustard). It also includes the model organism A. thaliana that has revolutionized our knowledge in almost every field of modern plant biology.

In the Arabidopsis genome, six genes coding for ACOs were identified in silico by Babula et al. (2006). Until now, only EAT1 (AT-ACO4) and AT-ACO2 gene structures, gene expression patterns and protein activities have been investigated (Gόmez-Lim et al. 1993; Raz and Ecker 1999; Ramonell et al. 2002; Raghavan et al. 2006; Linkies et al. 2009). However, based on the data from the UniProt database, there is substantial transcriptional evidence for three more ACO paralogs (AT-ACO1,3,5). A query in TAIR database for genes with 1-aminocyclopropane-1-carboxylate oxidase activity resulted in 13 loci (Table S1), some of them representing putative ACOs or putative 2OG dioxygenases. The homology analysis between these available A. thaliana ACO coding sequences allows for a strict distinction of five AT-ACO genes from their homologs or putative dioxygenases (Fig. 4; Table S1). It is interesting that four out of five AT-ACO genes are located on the same chromosome 1 and have a common structure (four exons interspersed by three introns) with ACOs from other species. The exception is AT-ACO1 located on chromosome 2, which has only three exons similarly to all Arabidopsis putative dioxygenase genes.

Poaceae

Monocotyledonous plants are less studied in respect to ACO gene structural and functional analysis. Still, there are several grass species for which ACO genes were isolated and characterized (Table S1).

Although browsing Rice Genome Annotation Project for ACO genes has resulted in 23 relevant loci, there are only three paralogs of ACO gene in rice isolated and characterized so far with distinct annotations in GenBank database (Table S1; Ouyang et al. 2007). The first, OS-ACO1, was isolated by Mekhedov and Kende (1996). Two others, OS-ACO2 and OS-ACO3, were characterized by Chae et al. (2000). The overall nucleotide sequence homology between OS-ACO2 and OS-ACO3 was reported as 72.6 %. The coding regions of OS-ACO2 and OS-ACO3 share 92 and 72 % sequence identity at the amino acid level with OS-ACO1, indicating that the OS-ACO3 gene is a more divergent member of the gene family. In contrast to ACO genes from other species, most OS-ACOs predicted in rice genome, do not show the typical four exon structure (Ouyang et al. 2007).

Gallie and Young (2004) revealed that ACO gene family in maize is represented by four genes, designated ZM-ACO15, ZM-ACO20, ZM-ACO31, and ZM-ACO35, which can be grouped into two subfamilies. One of them comprises ZM-ACO20 and ZM-ACO35 which are highly similar in amino acid sequence (91 % amino acid identity) and in the number and position of their introns. The other subfamily consists of ZM-ACO15 and ZM-ACO31 which are also highly similar to each other (96 % amino acid identity). These two subfamilies have diverged considerably so that a greater degree of divergence is found among members of the maize ACO gene family than that reported for other species such as petunia and tomato (Tang et al. 1993; Barry et al. 1996). ZM-ACO15 and ZM-ACO31 are more similar to ACOs from Arabidopsis and rice than they are to ZM-ACO20 and ZM-ACO35, suggesting that gene duplication and divergence of the two subfamilies, occurred prior to the divergence of monocotyledonous and dicotyledonous species. ZM-ACO20 possesses all twelve amino acid residues that are conserved among iron- and ascorbate-dependent dioxygenases, whereas ZM-ACO35, ZM-ACO15, and ZM-ACO31 contain most but not all of the conserved residues.

Pinaceae

In comparison to other higher plants, either dicots or monocots, our knowledge of ACO genes in Gymnospermae is rather limited. Quite recently an ACO gene family in loblolly pine comprising five paralogs PT-ACO1, PT-ACO2, PT-ACO3, PT-ACO4, and PT-ACO5 has been characterized (Yuan et al. 2010). Alignment of the deduced PT-ACO1 amino acid sequence with the sequences from other plant ACOs showed the encoded protein to contain all the conserved residues characteristic of ACOs. PT-ACO2 gene showed 98 % identity to PT-ACO1 in the nucleotide coding sequence, leading to its classification as a “nearly identical paralog” (NIP) which is a very unique feature among ACO genes. In contrast, PT-ACO3 has shown to be quite divergent from the first two paralogs in overall gene structure in containing only two introns instead of three.

Expression activity of ACO genes

Typically in higher plants, ethylene production is regulated by both developmental and external signals. During plant development, higher rates of ethylene evolution occur from young (vegetative, floral, and fruit) tissues that display a high rate of cell division. During the subsequent cell expansion and elongation stages, ethylene production declines. Finally it increases during fruit ripening, and leaf and floral senescence (Wang et al. 2002).

The ACO genes appear to be expressed in all vegetative (leaves, root, and epicotyl) and reproductive (flower and fruit) plant tissues and originally they were thought to be expressed constitutively. Later studies, however, revealed that there are differences in accumulation of specific ACO transcripts observed in various physiological processes. This suggests a different role played by ACO homologs during different developmental stages. The conversion of ACC into ethylene by ACO enzyme well coincides with the site of ethylene evolution so the determination of ACO transcript abundance may serve as an ethylene biosynthesis indicator.

For example, the ST-ACO1 and ST-ACO2 genes of potato are strongly expressed in leaves, but their expression in roots and tubers remains at low levels (Nie et al. 2002). In apple (Malus domestica cv. Royal Gala), three identified isoforms of ACO gene are differently expressed in fruit and leaf tissues (Binnie and McManus 2009). MD-ACO1 is restricted to fruit tissues, with optimal expression in ripening fruit, MD-ACO2 expression occurs more predominantly in younger fruit tissue, with some expression in young leaf tissue, while MD-ACO3 is expressed predominantly in young and mature leaf tissue, with less expression in young fruit tissue and least expression in ripening fruit. Also Momonoi et al. (2007) observed a differential regulation of five ACO paralogs in tulip among the vegetative tissues and during flower senescence.

Seed dormancy and germination

Ethylene is implicated in the seed formation and germination of many plant species (reviewed by Kępczyński and Kępczyńska 1997; Matilla and Matilla-Vázquez 2008; Linkies and Leubner-Metzger 2012).

Not much is known about ACO expressional activity throughout zygotic embryogenesis, a morphogenetic program which proceeds seed maturation. Rodríguez-Gacio et al. (2004) reported that one of ACO homologs, BR-ACO1 is highly expressed during the first phases of zygotic embryo formation in Brassica rapa, to decline afterwards and become undetectable during desiccation. It has been suggested that the maximum ethylene synthesis found at the beginning of embryogenesis is dependent on ACO expression and activity. On the other hand, in Sysimbrium seeds, SO-ACO2 transcripts rose during embryo development and further increased with the seed desiccation (Iglesias-Fernández and Matilla 2009).

In some species, endosperm undergoes programmed cell death late in its development which onset is also known to be under control of ethylene. Such a situation appears during seed development in cereals. In maize kernel, ACO genes are nearly exclusively expressed in the endosperm, and ZM-ACO20 is largely responsible for this transcript accumulation (Gallie and Young 2004). Expression of ACO genes during late endosperm development is likely to occur in those regions of the endosperm that remain viable at this stage, e.g., the subaleurone, aleurone, or pedicel-proximal starchy endosperm. The ACO gene family, and ZM-ACO20 in particular, represents the only component of the ethylene biosynthetic and perception machinery examined whose expression increases during late endosperm development, perhaps to insure continued ethylene production in those regions of the endosperm that remain alive during this period.

The seeds of most angiosperms are dormant at maturity (Matilla and Matilla-Vázquez 2008). Seed dormancy is developed during the last period of embryogenesis and prevents germination under unfavorable conditions. It has been proposed that ethylene may contribute to the breaking of dormancy in some species but its involvement in the removal of dormancy is still debated, especially in relation to ethylene-GA synergism (Kępczyński and Kępczyńska 1997; Matilla and Matilla-Vázquez 2008). It was proposed for example that transcriptional regulation of the FS-ACO1 gene plays a major role in regulating ethylene production required to break dormancy in beechnuts (Calvo et al. 2004).

Seed germination starts with imbibition of a dry seed and is considered as completed with the emergence of the radicle. Accumulation of ACOs and increased ethylene evolution accompany seed germination in several species (Munoz de Rueda et al. 1995; Petruzzelli et al. 2000; Hermann et al. 2007; Linkies et al. 2009; Iglesias-Fernández and Matilla 2010). In many instances it is ACO that controls the ethylene evolution during seed germination. As shown in pea seeds, ethylene may regulate its own biosynthesis through a positive feedback loop that enhances ACO expression (Petruzzelli et al. 2000). Though, in contrast with pea seeds, no such autocatalytic enhancement was observed for ACO in germinating sugar beet (Hermann et al. 2007), or the ACO2 orthologs of Sysimbrium officinale (Iglesias-Fernández and Matilla 2009) and Lepidium (Linkies et al. 2009). The levels of ACO transcripts have been shown to be regulated not only by ethylene itself but also other phytohormones (Lin et al. 2009). It is evident that during seed germination upon hormone treatment (ABA or GA) of A. thaliana, the regulation of ethylene synthesis is mediated by ACO (Linkies and Leubner-Metzger 2012).

Rodríguez-Gacio et al. (2004) observed that the transcription of the BR-ACO1 gene throughout the germination process of turnip tops remains at a low level during the imbibition phase, however, the transcript accumulation increases during embryonic axis protrusion and occurs predominantly in the growing organs of the seed (embryonic axis and radicle), being scarce in cotyledons. Similarly, PS-ACO1 mRNA, ACO activity, and ethylene production were almost exclusively detected in the embryonic axis of germinating pea seeds (Petruzzelli et al. 2000). It is postulated that in the embryonic axis, ethylene could act by promoting radial cell expansion or by decreasing the seed base water potential (Linkies and Leubner-Metzger 2012).

In seeds, where endosperm acts as a mechanical barrier, its weakening is essential for the radical protrusion. It was shown that ethylene, counteracting the inhibitory action of abscisic acid (ABA), promotes endosperm cap weakening and endosperm rapture (Linkies et al. 2009). Prior to the onset of endosperm rapture, there has been an accumulation of ACO2 transcripts observed (Iglesias-Fernández and Matilla 2009; Linkies et al. 2009). The collective evidence from three different species of Brassicaceae suggests that ACO2 orthologs may play a key regulatory role in this specific step of seed germination.

Vegetative phase of development

The vegetative development of plants is strongly dependent on the action of ethylene which has a close and reciprocal relation with auxin, another major determinant of plant architecture (Vandenbussche et al. 2012). During vegetative development, differential ACO gene expression has been observed in mung bean (Kim and Yang 1994), tomato (Barry et al. 1996; Moeder et al. 2002), sunflower (Liu et al. 1997), potato (Nie et al. 2002), and tulip (Momonoi et al. 2007) tissues.

It is generally accepted that all plant organs are capable of synthesizing ethylene from ACC, however, some data clearly imply that the contribution of different cell types within an organ or tissue is not equal (Vandenbussche et al. 2012). In the root of Arabidopsis for example, at least one member of the ACS family is expressed in each cell type, whereas the expression of the known ACO genes does not cover all root cell types (Dugardeyn et al. 2008; Vandenbussche et al. 2012). Similarly, ZM-ACO expression was restricted to the root cap, protophloem, and the companion cells associated with metaphloem sieve elements (Gallie et al. 2009). Such a differential expression pattern highly probably reflects variable enzymatic activity and may indicate the exact ethylene production site which cannot be assessed using currently available analytical methods.

Ethylene exerts multiple effects on plants during the vegetative stage of development (Vandenbussche et al. 2012). In general, primary and lateral root elongation is inhibited, whereas root hair development is stimulated. By contrast in hypocotyls of light-grown seedlings, as in stems of some semi-aquatic species, and during shade avoidance, ethylene can stimulate elongation. Leaf emergence is enhanced by ethylene, leaf expansion, however, is mostly inhibited. It is also known that keeping ethylene levels low during vegetative development is important in order to allow an optimal expansion of the root and the shoot, which can be achieved through negative feedback regulation of ethylene biosynthesis genes. Despite this, and the fact that there are specific spatial and temporal expression patterns of ACO genes in vegetative tissues, the exact role of ACO as a possible regulatory point in ethylene production still needs to be elucidated.

There are some examples of direct evidence for ACO involvement in the regulation of ethylene biosynthesis. An expression study in white clover indicated that ACO plays a regulatory role in leaf ontogeny (Hunter et al. 1999). In the apical hook of A. thaliana, AT-ACO2 expression is positively regulated by ethylene and restricted to the time period that the cells are ethylene sensitive (Raz and Ecker 1999). The AT-ACO2 transcripts have exhibited differential localization in the hook, accumulating in its outer cells. Auxin antagonists promoted primary root elongation, simultaneously reducing root diameter and root hair development in Linum usitatissimum affecting the expression of ethylene biosynthesis genes, both ACS and ACO transcripts levels were inhibited (Zhao and Hasenstein 2009). However, Takahashi et al. (2010) observed that among the differentially regulated ACO genes, LS-ACO2 plays a key role in ethylene production during low-pH-induced root hair formation in lettuce. A predominant ACO gene, OS-ACO1, has been isolated from submerged internodes of deepwater rice, and its high transcript levels are associated with internode elongation during submergence (Mekhedov and Kende 1996) and also at the heading stage (Iwamoto et al. 2010). OS-ACO1 gene expression remains under phytochrome control, especially phytochromes A and B (Iwamoto et al. 2011).

Generative phase of development

The studies in ornamental plants, have shown that flower opening, pollination, petal senescence, and abscission are often associated with transcriptional regulation of the ACS and ACO genes which in turn control ethylene synthesis (Bui and O’Neill 1998; Jones 2003; Fernández-Otero et al. 2006; Roeder et al. 2009). Trivellini et al. (2011) concluded that at the early flower development stage of Hibiscus rosa-sinensis, both ACS and ACO genes are likely to be involved in the regulation of ethylene biosynthesis transcriptionally, but during flower senescence, ethylene biosynthesis is regulated at ACO levels solely. Also in Alstroemeria, ethylene synthesis is regulated through ACO gene expression, rather than ACS, at the last stages of development (Wagstaff et al. 2005). Different isoforms of ACO gene are also involved in ethylene production during fruit development and ripening as shown in melon (Lasserre et al. 1996), tomato (Barry et al. 1996), peach (Mathooko et al. 2004), Coffea arabica (Pereira et al. 2005), Ficus carica (Owino et al. 2006), and banana (Choudhury et al. 2008a, b).

It is widely accepted that ethylene biosynthesis in ripening fruits and senescing flowers is subject to both positive and negative feedback regulation (Kende 1993). A positive feedback regulation, through an increase in ethylene production triggered by exogenous ethylene, with the activation of ACS and/or ACO has been reported for petunia (Tang and Woodson 1996), carnation (Jones and Woodson 1997), and orchid (O’Neill et al. 1993). A negative feedback regulation has been shown for transgenic petunia flowers (Wilkinson et al. 1997) and ripening tomato and banana (Nakatsuka et al. 1998; Inaba et al. 2007). However, Trivellini et al. (2011) have shown that ethylene biosynthesis can be under both positive and negative feedback regulatory mechanisms in a tissue-temporal-specific manner during flower development and senescence depending on differential expression of ACS and ACO genes. Similarly, the studies in tomato and banana have shown a differential feedback regulation of ethylene biosynthesis enzymes between pre-climacteric and ripening fruits and also between different tissues (Nakatsuka et al. 1998; Inaba et al. 2007).

Stress responses

Autocatalytic or feed-forward effects on ethylene production are associated not only with developmental cues, such as seed germination, ripening, shedding, wilting, and senescence, but also are observed in stress situations (Linkies and Leubner-Metzger 2012; Vandenbussche et al. 2012). Substantial evidence implies that these effects can be achieved through the elevation of transcript levels of key biosynthetic enzymes. The expression of genes encoding ACOs differs not only with tissue, stage of development but also with environmental conditions. Studies concerning different abiotic (temperature, wounding, and flooding) and biotic (pathological infection) treatments showed many instances of ACO gene differential upregulation. It was proposed that a biphasic mode of ACO transcript accumulation, similar to ACS may be a feature of stress-induced ethylene synthesis in tomato (Moeder et al. 2002). In potato, highly increased expression of all ACO genes, in both low and high temperature was observed (Nie et al. 2002). Wounding stimulated principally the transcription of some ACO paralogs: ST-ACO1 (Nie et al. 2002), NG-ACO3 (Kim et al. 1998), PP-ACO1 (Ruperti et al. 2001), MA-ACO1 (Pan and Lou 2008). In contrast, Choudhury et al. (2008b) observed that in preclimacteric banana fruit, wounding down-regulated the expression of both MA-ACS1 and MA-ACO1 genes and thus inhibited ethylene production.

The effect of flooding in potato induced the expression of ST-ACO1 and ST-ACO2; nevertheless ST-ACO1 appeared to be more sensitive to this stress. Other reports showed that also expression of RP-ACO1 responded to these conditions (Vriezen et al. 1999). In many species, including maize, the root undergoes re-modeling in response to low oxygen that occurs during flooding through the selective death of cortical cells which generates aerenchyma (Evans 2003). Hypoxia-induced cortical cell death is mediated by ethylene and is accompanied by increases in ACS and ACO expression in several species (Wang and Arteca 1992; Zhou et al. 2001; Gallie et al. 2009). The cell death induced by hypoxia is specific to the cortex, suggesting differential expression of the ethylene biosynthetic machinery in the cortex or sensitivity of this cell type to ethylene.

Just like the abiotic stresses, biotic treatments played the crucial role in synthesis of ACOs. The virus invasion is correlated with higher ethylene synthesis; it is suggesting that ACO genes are involved in the response against pathogen attack. High accumulation of ST-ACO1 transcripts in potato was induced by PV A (Potato virus A) infection (Nie et al. 2002), and NA-ACO1 and NA-ACO3 in tobacco by TMV (Kim et al. 1998). In Nicotiana benthamiana leaves infected with the fungal pathogen, Colleotrichum orbiculare, highly increased levels of NB-ACO1 were detected (Shan and Goodwin 2006). Similarly, an accumulation of ACO gene transcripts has been noted in other pathogenic fungi–plant interactions (Chen et al. 2003; Lahey et al. 2004; Iwai et al. 2006). The studies on NB-ACO1-silenced plants demonstrated that ACO gene is not only inducible by a pathogenic attack but also plays an important role in the appearance of necrotrophy during the development of a fungal infection (Shan and Goodwin 2006).

Transcriptional regulation of ACO activity

Although transcriptional regulation of ACO in response to various developmental and environmental factors, has been studied extensively in many plant species, the knowledge on the corresponding transcriptional regulators is still scarce. The identification of transcription factors (TFs) specifically involved in ACO gene activation or repression has proved difficult despite the presence of numerous putative cis-elements in the promoter sequences of ACO genes (Lin et al. 2009). For example, the SL-ACO1 promoter contains ethylene-responsive elements (EREs), core motifs of TFs, such as GT-1 and Dof1 that have been shown to be important for light-regulated genes, and putative homeobox protein-binding sites. These elements can be also found in promoter regions of several other ACOs, namely DC-ACO, ME-ACO3, or VR-ACO1.

Biochemical analysis showed that one of tomato ethylene responsive factors LeERF2/TERF2 interacted with GCC box in the promoter of NT-ACS3 and with dehydration-responsive element in the promoter of LE-ACO3, resulting in transcriptional activation of the genes for ethylene biosynthesis in tomato and tobacco, which was suggested as a novel regulatory function of ERF proteins in plant ethylene biosynthesis (Zhang et al. 2009). The 5′ upstream region of OS-ACO1 shown by Iwamoto et al. (2010) to have a promoter activity, also contains a GCC-box element (AGCCGCC) to which ERE-binding proteins/factors (EREBPs/ERFs) may bind. Further, it has been shown that phytochrome-regulated EBL1 (OsEREBP1-LIKE1) gene contributes to the regulation of OS-ACO1 gene expression (Iwamoto and Takano 2011). Also Choudhury et al. (2008a, b) identified a GCC-box motif in the promoters of two ACO homologs in banana and also found a 25-kDa nuclear protein that was able to bind specifically to GCC-box DNA in the climacteric fruits. Furthermore, Huang et al. (2010) isolated two orthologs of EIN3 transcription factor which is necessary and sufficient for the activation of many ethylene-responsive genes in Arabidopsis (Solano et al. 1998) and demonstrated that they act as transcriptional activators of the MC-ACO1 gene in the melon ripening fruit.

Quite recently it has been reported that ACO is transcriptionally regulated by a homeotic protein. Lin et al. (2008) identified a leucine zipper homeobox protein, LeHB-1, that transcriptionally modulates the expression of tomato LE-ACO1. The LeHB-1 protein contains 285 amino acid residues with a conserved homeodomain (HD) and a leucine zipper domain (Zip), belonging to the class-I HD–Zip proteins. In vitro gel retardation assays showed that LeHB-1 binds to the LE-ACO1 promoter sequences that contain a 9 bp AATA(A)TATT or 10 bp AATA(AA)TATT sequence with dyad symmetry and similarity to the putative homeobox protein-binding sites. Using virus-induced gene silencing (VIGS) to down-regulate LeHB-1 in tomato resulted in reduced expression of LE-ACO1 mRNA and delayed ripening. These results indicated that LeHB-1 is a transcriptional activator that controls LE-ACO1 expression in vivo (Lin et al. 2008).

In silico analysis of some ACO gene promoters revealed the presence of TAAAATAT, which has been identified as a core element in LeCp (tomato Cys protease) binding cis-element of LeAcs2 gene (Matarasso et al. 2005). LeCp has been characterized to be a dual-function protein, having enzymatic activity and, upon elicitor signaling, exhibiting transcriptional factor activity that induces LeAcs2 expression. It can be hypothesized that it might act similarly as a regulator of ACO genes.

Altering ethylene levels in plants via ACO

Genetic transformation is a powerful method for breeding crop plants with desired traits. Modifying the amount of ethylene produced under ripening, senescing, or stress conditions is the goal of a wide array of transgenic strategies (reviewed by Stearns and Glick 2003; Kępczyński and Kępczyńska 2005; Handa et al. 2010). Most of the research in this area has been directed toward the control of fruit ripening and flower senescence through genetic manipulation of the two key genes involved in ethylene biosynthesis, ACS and ACO. There are several examples of successful application of genetic engineering of ACO, which include tomato (Hamilton et al. 1990; Xiong et al. 2005), cantaloupe melon (Ayub et al. 1996), muskmelon (Nuñez-Palenius et al. 2006), or papaya (Lόpez-Gόmez et al. 2009) fruits with longer shelf-life as well as carnation (Savin et al. 1995), torenia (Aida et al. 1998), and petunia (Huang et al. 2007) flowers with increased longevity. The studies conducted by Huang et al. (2007) have even shown that using the antisense BO-ACO1 gene from B. oleracea can be more effective than the antisense BO-ACS1 gene in reducing ethylene biosynthesis and delaying flower senescence of P. hybrida.

First genetic engineering approaches worked out for ethylene manipulation through ACO gene were based on the antisense gene technology (Hamilton et al. 1990). An introduction of the ACO gene in antisense orientation has been proved as an efficient method for ethylene suppression and is still in use (Nuñez-Palenius et al. 2006; Kinouchi et al. 2006; Huang et al. 2007; Wang et al. 2009). Other studies have demonstrated that the introduction of truncated ACO gene constructs in the sense orientation can result in co-suppression of homologous host genes, and modulate ethylene biosynthesis (Knoester et al. 1997; Aida et al. 1998; Lόpez-Gόmez et al. 2009). The discovery of double-stranded RNA (dsRNA)-mediated gene silencing referred to as RNA interference (RNAi) has led to the development of new strategies to block gene function. For genetic improvement of crop plants, RNAi has advantages over antisense-mediated gene silencing and co-suppression, in terms of its efficiency, stability, and specificity. Transgenic tomato fruits with prolonged shelf life of more than 120 days have been produced by using RNAi technology (Xiong et al. 2005).

Regarding the transgenic strategies that face strong concern in public acceptance, new approaches for ethylene biosynthesis regulation through ACO gene are worth mentioning. Over the past several years, consumer and environmental groups have expressed concerns regarding the use of antibiotic- and herbicide-resistance genes in ecological and food safety perspectives. A new transformation system overcoming such concerns has been established by Hao et al. (2011). A marker- and vector-free antisense ACO construct was successfully transformed into melon via the pollen-tube pathway generating transgenic fruits with superior characteristics of greatly reduced respiration rates and endogenous ethylene production thus promising a significantly prolonged shelf-life compared to wild-type ones. A different strategy, totally avoiding plant transformation, was proposed by Dahmani-Mardas et al. (2010). They were also able to engineer melon plants with improved fruit shelf-life by means of targeting induced local lesions IN genome (TILLING) which combines advantages of random chemical mutagenesis and high throughput mutation discovery methods.

Conclusions and perspectives

Even though ACO gene homologs have been isolated from a number of species, the biochemical properties, in vivo stability, and/or physiological functions of the gene products are in many cases not fully understood. Furthermore, it is often unclear whether a single gene is expressed in response to a variety of stimuli or a specific ACO gene responds to a specific stimulus. Thus, further study is needed to verify the role played by different ACO genes in the regulation of ethylene production. Different expression levels of ACO genes induced by environmental and developmental signals, could explain function of identified isoforms in each species. Similarly important would be to reveal the structural organization of promoter regions of ACO genes, as well as characterization of cis- and trans-acting regulatory elements. Better understanding of both structure and function of genes involved in ethylene biosynthesis, and also the mechanisms responsible for their activation or repression, make possible the regulation of their activity during different developmental processes of plants. All this has a profound practical importance as it develops the tools for crop and ornamental plants improvement.

Author contribution

This review paper was accomplished with the collaboration of all authors under the supervision of Prof. J. Kępczyński. The research topic and framework were defined by J. Kępczyński who also revised the final version of the manuscript. The literature search and the first manuscript was written by M. Sasiak and J. Kępczyński. Further literature exploration and the manuscript revision was performed by I. Ruduś who was also responsible for the preparation of figures, revision of the table, checking the references and formatting the text. All authors have contributed their efforts to this work.