Location of chlorogenic acid biosynthesis pathway and polyphenol oxidase genes in a new interspecific anchored linkage map of eggplant

Eggplant (Solanum melongena L., Solanaceae; 2n = 2× = 24) ranks third in the genus Solanum, after potato and tomato, in total production and economic importance and is the most important Solanaceae crop native to the Old World [1]. The most nutritionally important bioactive constituents of the eggplant fruit are phenolics, which are responsible of the high antioxidant activity of eggplant [2]-[6]. The most abundant phenolics of eggplant are hydroxycinnamic acid (HCA) conjugates, which are synthesized by converting phenylalanine to cinnamic acid. Among HCA conjugates, chlorogenic acid (5-O-caffeoyl-quinic acid; CGA) constitutes between 70% to over 95% of the total phenolics content [7]-[10]. Growing interest in this compound is due to its many beneficial properties for the treatment for various metabolic and cardiovascular diseases and ailments. Several in vitro and in vivo experiments have shown that CGA has anti-oxidant, anti-inflammatory, analgesic, antipyretic, neuroprotective, cardioprotective, anti-carcinogenic, anti-microbial, hypotensive, anti-obesity and anti-diabetic activity [5],[11],[12]. Moreover, CGA is highly stable at high temperatures, and its bioavailability in eggplant increases, as compared to the raw product, after cooking [3].

Great diversity in the content of total phenolics and CGA has been observed in eggplant, due both to genetic and environmental factors [2],[6],[7],[9],[10],[13]. Some close wild relatives of cultivated eggplant, such as S. incanum [14],[15], have high levels of CGA [7],[9],[16]. Solanum incanum is native to northern Africa and the Middle East to Pakistan [15], and is a cross-compatible with S. melongena [1],[7]. Therefore, S. incanum shows promise for use in breeding programs for developing new eggplant varieties with increased phenolic content [5].

Raising the total phenolics content, however, may cause a negative effect on apparent quality of the fruit. When eggplant fruit flesh is cut, phenolics, mostly stored in vacuoles, become available to polyphenol oxidase enzymes (PPOs), which are present in chloroplasts. PPOs catalyse the oxidation of phenolics to quinones, which in turn, react non-enzymatically with oxygen in the air to give brown compounds, thus causing browning of fruit flesh [17]. Several authors have found differences in PPO activity between varieties of eggplant, which can lead to differences in the degree of browning in fruit flesh between varieties with similar content of total phenolics [18]-[20]. Molecular breeding for high CGA content and low PPO activity could contribute to developing improved cultivars with higher bioactive properties through a combination of high antioxidant activity and presenting a low degree of browning. For this purpose, a candidate gene approach shows promise, given that the genes involved in the CGA synthesis pathway, which include phenylalanine ammonia lyase, PAL; cinnamate 4-hydroxilase, C4H; 4-hydroxycinnamoyl-CoA ligase, 4CL; hydroxycinnamoyl-coA shikimate/quinate hydroxycinnamoil transferase, HCT; p-coumaroyl ester 3’-hydroxilase, C3’H; and, hydroxycinnamoyl CoA quinate hydroxycinnamoyl transferase, HQT, (Figure 1), in addition to the PPO genes, are known [20]-[24].

Understanding of eggplant genome organization, which is of great relevance for molecular breeding, has lagged behind that for other solanaceous crops such as potato, tomato and pepper. Several linkage maps for eggplant have been developed. Nunome et al. [25] developed a first intraspecific linkage map in eggplant using RAPD and AFLP markers. Two improved versions of the Nunome et al. [25] map were developed by adding SSR markers [26],[27]. Doganlar et al. [28] developed the first interspecific map using RFLP markers resulting from crossing S. melongena and S. linnaeanum. The resolution of this map was further improved by adding COSII and AFLP markers [29],[30]. Barchi et al. [31] also developed an intraspecific mostly AFLP and SSR marker. Finally, an intraspecific saturated integrated map of S. melongena was developed by Fukuoka et al. [32] from two F₂ populations in which SSR and SNP markers were mapped. Of the markers used by Fukuoka et al. [32], many were obtained from Solanum orthologous (SOL) gene sets from a multiple alignment between the unigenes of eggplant, tomato and potato.

Here we report the development of a new interspecific S. melongena × S. incanum linkage map with the aim of locating, and in the future introgressing, S. incanum alleles involved in the biosynthesis of CGA in the genetic background of S. melongena. In order to devise molecular tools for minimizing browning associated with high CGA levels, PPO genes were also targeted. This new map is anchored to the tomato genetic map and previous eggplant maps, which will facilitate molecular breeding in eggplant for high CGA content and reduced browning as well as other morphological traits of importance in eggplant breeding.

Although notable efforts have been made recently to better understand the structure and organization of the eggplant genome, the available genomic information is still very limited when compared to other major Solanaceae crops such as tomato, potato and pepper. Our interspecific map spread 1085 cM, and the 15 linkage groups we have found could be traced back to 12 chromosomes through the use of markers shared with other previously published maps. The SMIBC was the first map that connects all the most informative previously published genetic maps for eggplant, allowing us to exploit the information generated by these maps (Figures 3 and 4). A total of 130 anchoring points connect SMIBC and Tomato EXPEN-2000, many of which are COSII, enabling us to establish synteny between the two genomes (Figures 3 and 4, Additional file 1: Figure S1, Additional file 2: Table S1). This allowed us to confirm the position of all candidate genes according to orthologous gene location in tomato. The syntenic relationships between these two maps are in agreement with those observed by Doganlar et al. [28], Wu et al. [29], and Fukuoka et al. [32] and provide information on the genome evolution of both crops.

A general problem in the construction of genetic maps is that different types of molecular markers tend to map in a specific regions of the genome [27],[32],[39],[40]. Our use of a wide variety of molecular markers including those used in previous eggplant genetic maps [27],[29],[32],[33] allowed us to achieve better representation of the eggplant genome, which is important for effective molecular breeding and synteny studies in the future.

Distorted markers occurred in almost all linkage groups, except for E01 and E04. Segregation distortion in mapping populations is a well-documented phenomenon in different crops [41],[42]. Similar distortion levels to those we observed have been reported by Doganlar et al. [28] in their interspecific map between S. melongena and S. linnaeanum, with 16% of markers being skewed towards the one or the other parent.

CGA is the major phenolic compound present in eggplant, which is one of the crops with highest content in hydroxycinnamic acids [5],[8]-[10]. Until now, attempts to improve the CGA content in eggplant have been conducted in a classical way through hybridization and selection of materials with high CGA content [4],[5],[7]. The results of these programs have been positive, but due to moderate heritability (H ² ) of the CGA content trait and total phenolics in general, genetic advances have been limited [7],[43],[44]. Although these studies indicate that phenotypic selection to obtain materials with higher contents of CGA is possible, it would be desirable to apply marker-assisted selection (MAS) to improve the efficiency of selection for this trait.

The candidate gene (CG) approach has been successful in both animal and plant genetics [45]. The involvement of CGs in CGA biosynthesis pathway has been studied in plant species. The PAL gene catalyses one of the first steps of the phenylpropanoid pathway, that produces hundreds of specialized metabolic products including lignins, flavonoids, alkaloids and many other important phenolics in plants [46]. In Populus trichocarpa, five PAL genes have been described that are involved in wood formation [47], and in raspberry (Rubus idaeus) two PAL genes have been identified: RiPAL1 associated with early fruit ripening events while RiPAL2 involved with larger flower and fruit development [48]. Candidate genes involved in phenylpropanoid pathway have also been mapped in other plants like apple [49] and artichoke [50]-[52]. These are examples showing that homologs can have different functions, and paralogs can play largely different roles in CGA or lignin synthesis. Even within the Solanaceae, different genes in the pathway are strongly associated with the abundance of CGA. HQT seems to be the most important contributor to CGA synthesis in tomato and Nicotiana benthamiana [24], while HCT and C3’H are the most relevant in potato [53],[54].

In this work we were able to locate all the genes involved in the core CGA pathway in eggplant. The positions of the genes are in agreement with those expected based on synteny with tomato. With the exception of HCT and 4CL, which are co-situated at a genetic distance of 1.3 cM in the linkage group E03, all of the rest of the genes of the biosynthesis pathway of CGA are situated in separate linkage groups. This has important implications for breeding and for strategies based on pyramiding of favourable alleles introgressed from S. incanum in the genetic background of eggplant.

A potential problem for developing materials with higher CGA content is their oxidation mediated by polyphenol oxidase (PPO), which leads to the browning of the fruit flesh, reducing apparent fruit quality [5],[18]. A positive phenotypic linear correlation was found in eggplant between the degree of fruit flesh browning and total content in phenolic and CGA content [4],[43].

Several studies in other crops associated quantitative trait loci (QTL) for enzymatic browning with PPO alleles and these alleles were then used as markers for this trait [55],[56]. On the other hand, silencing of PPO genes result in a reduced browning reaction as has been shown in potato and apple [57],[58].

The studies of PPO genes suggest that in plants they form a gene family with high homology, usually have no introns and normally are present in several copies. For instance, tomato has seven PPOs, potato has six PPOs and soybean has eleven PPOs [56],[59]-[61]. These features have complicated the search of polymorphisms between PPOs because amplification of a single isoform is difficult. In this study we were able to detect polymorphism in all PPO genes with the exception of PPO6 (amplified in S. melongena but not in S. incanum) leading us to speculate that in S. incanum this PPO gene is absent or presents substantial changes with respect to S. melongena. As in tomato [60], eggplant PPOs were mapped very closely to each other in linkage group E08; this confirms the existence of a cluster also in eggplant (Figure 2), although a correct assignment using synteny was not possible between eggplant and tomato orthologs, due to high sequence similarity among the PPO genes of each species. In eggplant PPO genes identity varying between 72% and 95% at nucleotide level and between 62% and 92% at amino acid level [20].

Since both consumers and industry prefer eggplant varieties with white flesh and a low degree of browning, future varieties with high phenolic content should also be bred to have limited browning [43]. Accordingly, the search for allelic variants for increasing CGA content should be conducted in concert with a search for allelic variants decreasing PPOs. In eggplant, several authors have found differences in PPO activity between varieties [18]-[20] which suggest mining the biodiversity of this crop and its relatives can lead to discovering desirable alleles useful to produce varieties with low browning in fruit flesh regardless of the total phenolic abundance. In addition, the study of selection sweeps at these loci due to the culturally distinctive organoleptic preferences, would be of interest to understand the history of evolution of eggplant. A reduction in browning is quite probably a significant domestication trait in eggplant and many other species.

In the present study an interspecific genetic linkage map was developed to better understanding the genetics of CGA content and fruit flesh browning in eggplant through the mapping of candidate genes involved in these processes. This was assisted by the use of synteny of the orthologous genes in tomato. The use of S. incanum in this work and in breeding programs is not accidental, as it is a close relative of eggplant [14],[15] and produces fully fertile hybrids with regular meiosis [7],[14]. Solanum incanum has characteristics of great interest for plant breeders, such as bacterial and fungal resistance, including to Fusarium oxysporum, and tolerance to various abiotic stresses, especially drought [7],[62]. Recent discoveries have shown that S. incanum is a powerful source of variation for phenolic content, as its total phenolic content is several times higher than that of S. melongena [7],[9]. Solanum incanum also exhibits absence of anthocyanins, presence of prickles and subovoid green fruits [7].

Building an interspecific genetic map from the backcross generation of two species that differ in many traits has two important advantages. First of all, in this manner the problem of low intraspecific polymorphism in eggplant [28],[29],[32],[63] is partially overcome. Secondly, this BC1 population can be used as the starting point to develop a set of introgression lines of S. melongena, with S. incanum as the donor parent. Once obtained, these lines will be useful to understand and dissect candidate gene roles in CGA accumulation and browning, to detect QTLs and genes involved in the synthesis of other polyphenols present in S. incanum (mainly N-(E)-caffeoylputrescine, 3-O-Malonyl-5-O-(E)-caffeoylquinic acid and 5-O-Malonyl-4-O-(E)-caffeoylquinic acid) [7], as well as many other morphological and physiological traits of interest for breeding or for crop evolution and domestication studies. Candidate genes for traits thought to be important in domestication such as fruit shape and PRICKLINESS are mapped in SMIBC in agreement with their positions based on previous QTL studies in eggplant and in syntenic regions of tomato [64],[65], also highlighting the common occurrence of parallelism in loci causing domestication phenotypes between species.

Despite its economic importance few genomic and biotechnological resources exist for eggplant. Eggplant contains high quantities of phenolics, in particular of CGA, a secondary metabolite with many properties beneficial for human-health and which in the last years is attracting a lot of attention from the scientific community. We report the development of a new interspecific map in eggplant using different types of markers. We have mapped the six genes involved in CGA synthesis pathway on five linkage groups, and determined that the PPO genes cluster together on LG8. This new map is anchored to the tomato genetic map as well as to four previous genetic maps of eggplant, and therefore constitutes an important tool for exploiting synteny and for molecular breeding in eggplant. The molecular tools we have developed will be of interest for eggplant breeding programmes, in particular those aimed at using marker assisted selection for improved CGA content combined with low browning effects. These tools and results, together with the recently published eggplant genome sequence [66] will greatly facilitate the development of a new generation of eggplant varieties with dramatically enhanced bioactive properties that can significantly contribute to better human nutrition.