AAllmmaa MMaatteerr SSttuuddiioorruumm –– UUnniivveerrssiittàà ddii BBoollooggnnaa .. Dipartimento di Colture Arboree DOTTORATO DI RICERCA IN BIOTECNOLOGIE CELLULARI E MOLECOLARI Ciclo XXIII Settore scientifico-disciplinare: AGR/03, Arboricoltura Generale e Coltivazioni Arboree MMEETTAABBOOLLIICC AANNDD MMOOLLEECCUULLAARR AASSPPEECCTTSS OOFF CCYYAANNOOGGEENNEESSIISS IINN AAPPRRIICCOOTT SSEEEEDD Dissertazione presentata dalla Dott.ssa Claudia Cervellati Coordinatore Dottorato Relatore Prof. Santi Spampinato Prof. Andrea Masia Co-relatori Prof. Bernd Schneider Dr. Luca Dondini Esame finale anno 2011 Index Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.1 Cyanogenic glicosides . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.1.2 Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.1.3 Key enzymes for the anabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.1.4 Key enzymes for the catabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.1.5 Cyanogenesis in the Prunus genus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.2 The apricot system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 1.2.1 Systematic placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 1.2.2 Genetic improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.2.3 Early selection and molecular markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 1.2.4 Genetic maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 1.2.5 Quantitative trait loci analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 1.2.6 Fruit quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 1.2.7 Seed quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 1.3 Fine phenotyping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 1.3.1 Quantitative NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 1.4 Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 2. Aims of the work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.1 Starting material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.1.1 Apricot populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.1.2 Lito’s BAC library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.2 DNA extraction and quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3.3 Amplifications, electrophoresis and screening for markers . . . . . . . . . . . . . . . . . . . . 54 3.3.1 Primer design and PCR optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.3.2 Agarose gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 3.3.3 Polyacrylamide gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 3.4 Enzymatic digestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.5 BAC screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.5.1 Positive clones picking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.5.2 Plasmid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.5.3 Primer walking and sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.6 Quantification and localization of CNGs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.6.1 Colorimetric method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 3.6.2 Quantitative NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 3.6.3 Raman imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 3.7 QTL analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.1 Identification of apricot candidate genes involved in the CNGs metabolism . . . . . 73 4.1.1 PCR optimization and first sequencing step . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.1.2 BAC screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.1.3 Primer walking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.2 Implementation of the L×B and H×R maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 4.2.1 Functional markers development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 4.2.2 Markers mapping in L×B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 4.2.3 Markers mapping in H×R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.3 Amygdalin phenotyping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.3.1 Colorimetric method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.3.2 Quantitative NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 4.4 Amygdalin in situ localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 4.5 Quantitative trait loci analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 4.5.1 Genetic regions identified in the L×B population . . . . . . . . . . . . . . . . . . . . . . 99 4.5.2 Genetic regions identified in the H×R population . . . . . . . . . . . . . . . . . . . . . 103 5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Abstract Abstract An holistic investigation on the physiological and genetic factors influencing the apricot fruit quality must consider the characteristic of the whole fruit, seed included. In fact, usually only the edible part of the fruit is taken in consideration when quality is defined. However, especially in stone fruits, the seed represents an important portion of the fruit itself and the major by-product in fruit processing. Nowadays, huge amounts of seeds are discarded yearly when processing apricot fruits. This not only wastes a potentially valuable resource but also aggravates an already serious disposal problem (Femenia et al., 1995; Schieber et al., 2001). Besides, apricot seeds could be a very interesting source of food, thanks to their significant content in dietary protein, oil and fibre (Gòmez et al., 1998). Unfortunately, seeds from several apricot varieties contain also a remarkable amount of the cyanogenic glycoside amygdalin. This represents a big constraint to their use for human or animal nutrition, both for its bitter taste and, moreover, for the toxicity of the hydrogen cyanide (HCN) produced by its dissociation. So, development and cultivation of high quality cultivars with sweet kernel may represent a quantum change for the apricot production system, as far as it would decrease health hazards and in parallel increase the marketability of this by-product. Cyanogenic glycosides (CNGs) are β-glucosides of α-hydroxynitriles derived from amino acids and have been found in more than 2,650 plant species distributed among 130 families in pteridophytes, gymnosperms and angiosperms. Such a widespread occurrence among many different taxonomic groups implies that in plants the ability to synthesize CNGs is at least 300 million years old (Bak et al., 2006). The most known role of CNGs is linked with plant defences. In fact, secondary metabolites play an important role in these responses and CNGs represent an important group of phytoanticipins: constitutive defence compounds produced independently from the pathogen attack (Zagrobelny et al., 2004). These natural products are instantly activated upon tissue damage or pathogen attack. Briefly, CNGs are stored in the vacuoles (Vetter, 2000) and, when the cell integrity is lost, they are brought into contact with the specific β- glucosidases and α-hydroxynitrile lyases that hydrolyze them, thereby causing a release 1 Abstract . of toxic hydrogen cyanide. This binary system provides plants with an immediate defence against intruding herbivores and pathogens that cause tissue damage (Møller and Seigler, 1999; Møller, 2010). Thus, cyanogenesis requires the presence of two biochemical pathways: the former controlling the synthesis of the cyanogenic glycoside and the latter controlling the production of the specific degradative enzymes. The balance of these routes determines the cyanogenic potential of the plant. In addition, accumulation of cyanogenic glucosides in certain angiosperm seeds may provide a storage deposit of sugars and nitrogen for the developing seedlings (Swain et al., 1992b; Swain and Poulton, 1994). Other evidences demonstrate that cyanogenic compounds also have a function in the metabolism and transport of nitrogen (Gleadow and Woodrow, 2000; Busk and Møller, 2002), in the legume-rhizobium interaction and in reducing the damages caused by excessive light (Møller, 2010). These roles do not, however, need to be seen as an alternative to the herbivore defence: just as some plants store nitrogen as inactive Rubisco, cyanogenic plants may store nitrogen in a toxic form for the dual purpose of defence (Gleadow and Woodrow, 2002). Apricot and other species belonging to the Rosaceae family are among the plants producing the highest levels of CNGs (Swain and Poulton, 1994). CNGs inheritance has been studied in various Prunus species. In peach, the sweet kernel behaves as a recessive trait, controlled by a single gene (sk), linked to the fuzzless skin (nectarine) trait and therefore assigned to the linkage group 5 of a genetic map based on an interspecific almond×peach cross (Bliss et al., 2002). In almond, sweet has been reported as a dominant trait inherited as a simple Mendelian factor (Dicenta et al., 2007; Sánchez-Pérez et al., 2008). In this species, as well as in apricot, all the seeds produced by a single tree show the same phenotype (bitter, slightly-bitter or sweet); thus, the presence of CNGs is a seed-parent trait (Negri et al., 2008; Sánchez-Pérez et al., 2010). Regarding the apricot, different hypotheses on the inheritance of the bitter/sweet trait have been proposed (Kostina, 1977; Bassi and Negri, 1991) but no definitive model has been demonstrated. Negri et al. (2008), given that both the bitter and the sweet phenotypes were represented in populations from bitter×bitter and sweet×sweet crosses as well as from self-pollination of either bitter- or sweet-seeded trees, elaborated a model in which five non-linked genes (three for the anabolic and two for the catabolic way) were involved in the determination of this quantitative character. A very efficient approach to study quantitative trait loci (QTLs) implies the use of Candidate Genes (CGs, Pflieger et al., 2001; Etienne et al., 2002). This approach is based 2 Abstract on the a priori choice of one or more genes that seem to be functionally connected to the examined character: a correlation between the phenotypic data and an allelic polymorphism validates the hypothesis that the chosen gene controls, at least partly, the studied trait (Causse et al., 2004). On these CGs it would then be possible the development of functional markers, based either on direct sequence polymorphisms (SCAR: Sequenced Characterized Amplified Region) or on differences in restriction sites (CAPS: Cleaved Amplified Polymorphic Sequence) or on a different number of microsatellite repetitions (SSR: Short Sequence Repeat). These particular kind of molecular markers would eventually provide the breeders with a valuable method to carry on an early selection of the plants with the desired traits. In fact, being a direct analysis of the plant genotype, the MAS (Marker Assisted Selection) makes possible the recognition of the individuals carrying certain characteristics way before these would be showed in the field. To support these selection programmes, the availability of a densely populated linkage map is a prerequisite. Four apricot maps were realized for the apricot accessions Lito, BO81604311 (San Castrese×Reale di Imola; Dondini et al., 2007), Harcot and Reale (Dondini et al., 2010). Beyond functional markers and genetic linkage maps, the identification of the DNA regions and of the candidate genes that underline the multifactorial trait bitterness presupposes a fine phenotypization. Only a very precise analysis, in fact, could provide the characterization of the various QTLs accountable for the variation of the phenotypic trait, so telling us about the genetic architecture of this character. The nuclear magnetic resonance (NMR) is considered to be one of the most robust and precise analytical methods in research, and it’s slowly becoming more appreciated for the phenotyping purpose (Moing et al., 2004; Terskikh et al., 2005; Pereira et al., 2006). Under precise “quantitative conditions” (Pauli et al., 2005 and 2007) NMR enables a unique and quantitative determination of the relative amount of molecular groups, so being a real powerful tool to quantify entire molecular structures even in mixtures. Moreover, quantitative NMR (qNMR) analyses are time-saving, thanks to the ability of quantifying a single compound in complex mixtures without requiring fractionation or isolation procedures. Thus, to investigate the metabolic and molecular aspects of the cyanogenesis in apricot, a candidate gene approach coupled with a precision phenotyping via qNMR was used on two F1 apricot population with the aims of: (i) identify and sequence some 3 Abstract . apricot genes involved in the cyanogenesis; (ii) develop functional markers linked to those genes and implement them in four existent maps; (iii) achieve a fine phenotyping of the amygdalin content of the seeds collected from the seedlings of both progenies; (iv) identify the genetic loci involved in the determination of the bitterness; (v) localize the amygdalin directly in situ with Raman imaging. Screening the Lito’s BAC library with the heterologous primers designed on sequences available in public databases resulted in the achievement of the apricot sequences for: UDPG-glucosyl transferase, amygdalin and prunasin hydrolases, mandelonitrile lyase. The primer walking on those sequences is not yet completed, but the length achieved by now is sufficient enough to give high homology results with the other available Prunus sequences. Seven new functional markers developed on three candidate genes involved in the catabolism of CGs were added to the maps of the two apricot accessions Lito and BO81604311. Four of them were added in Lito: an SSR for an MDL gene in L1, an SSR for an isoform of AH in L7, two loci of a CAPS marker for a PH gene in L2 and L7. The remaining three were positioned in the map of BO81604311: a CAPS marker for an MDL isoform in B1, another linked to a PH gene in B6 and an SSR for an AH in B7. Other eight markers, seven developed on two catabolic enzyme sequences and one on an anabolic one, were instead mapped on the cultivars Harcot and Reale di Imola. Five were mapped in Harcot: an SSR for UGT in H3; an SSR and two CAPS for a PH gene in H1 and H6, respectively; and an SSR for an isoform of AH in H7. The other three were located in the map of Reale di Imola: a CAPS marker for an AHL isoform in R6A, another linked to a PH gene in R7 and an SSR for an AH always in the LG7. Analogously to almond (Sánchez-Pérez et al., 2010), none of the CGs mapped in this work was found to be in a QTL region. However, other isoforms of these genes could still be involved in the determination of the bitter phenotype. The major QTLs for bitterness, already identified in almond and peach (Joobeur et al., 1998, Dirlewanger et al., 2004, Sánchez-Pérez et al., 2007), was found in the linkage group 5 mainly in Harcot (30.63 of LOD), but also in Reale di Imola, even if with a relative lower effect. Another QTL has been identified in the bottom part of R4. As for Lito and BO81604311, putative QTLs were found in L1, L4 and L6, as well as in B1, B4 and B6, seeming to be located in the same position in both parents. In Lito another significant locus is found in L3. The Lito×BO81604311 cross produces a progeny without 4