Introduction
The perennial trees Ilex paraguariensis St. Hill. and Ilex dumosa Reissek (Aquifoliaceae Bartl.) are native to tropical and subtropical South America, which are frequently found in southern Brazil, south-eastern Paraguay and north-eastern Argentina; only I. paraguariensis reaches Uruguay as patchy populations (Giberti, Reference Giberti1989, Reference Giberti2001, Reference Giberti2011). These species are dioecious trees, insect-pollinated, and have the same diploid chromosome number (2n= 40; Barral et al., Reference Barral, Poggio and Giberti1995; Greizerstein et al., Reference Greizerstein, Giberti and Poggio2004).
The aerial parts of I. paraguariensis, the ‘yerba mate’ tree, are commercialized in the southern countries of South America to prepare a popular beverage called ‘mate’, highly appreciated for its peculiar flavour and stimulating properties. Recently, countries in the Middle East have incorporated the custom of drinking ‘mate’, with Syria being the largest consumer. Its choleretic, hypocholesteraemic, antioxidant and hepatoprotective properties (Berté et al., Reference Berté, Beux, Spada, Salvador and Hoffmann-Ribani2011; Bracesco et al., Reference Bracesco, Sanchez, Contreras, Menini and Gugliucci2011) are attributed to the high concentration of xanthines, flavonoids, saponins, amino acids, minerals and vitamins (Heck and de Mejia, Reference Heck and de Mejia2007). I. dumosa, known as ‘yerba señorita’, has comparatively mild effects due to a significant lower content of caffeoyl derivatives and flavonoids (32 and 10 times, respectively), and methylxanthine content varying from little to none (Filip et al., Reference Filip, López, Coussio and Ferraro1998, Reference Filip, López, Giberti, Coussio and Ferraro2001). For a long time, I. dumosa has been used as an adulterant (Giberti, Reference Giberti1989); however, nowadays, it is legally marketed as a low-caffeine alternative.
Notably, the commercial production of ‘yerba mate’ is maintained by intensive clonal propagation of a small number of genotypes, restricting its genetic background and its response to pests and diseases. I. dumosa constitutes a natural reservoir of agronomically interesting genes, and is a suitable potential donor for introgression into the commercial lines of ‘yerba mate’, as it shows increased resistance to drought conditions and to the main pest (Gyropsylla spegazziana, Psyllidae) of ‘yerba mate’ (Prat Kricun, Reference Prat Kricun2009).
Although the species have great socio-economic importance, the scarcity of basic genetic data halts the implementation of modern approaches to genetic improvement. Formal molecular microevolutionary studies concur in highlighting that the natural populations of ‘yerba mate’ still available in the wild, have a narrow genetic base (Gauer and Cavalli-Molina, Reference Gauer and Cavalli-Molina2000; Neumann-Wendt, Reference Neumann-Wendt2005; Cascales, Reference Cascales2013). So far, only one study has characterized I. dumosa using molecular markers (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2011). In this context, the generation of genetic information about I. paraguariensis and I. dumosa will aid ongoing traditional improvement activities by enhancing the development of more modern programmes. To contribute to their knowledge, we initiated an exploration for specific distinctive features through a comparative characterization, using relatively low-cost methodological molecular approaches. We implemented a subtractive hybridization technique, such as the representational difference analysis (RDA; Lisitsyn et al., Reference Lisitsyn, Lisitsyn and Wigler1993), coupled with single-strand conformation polymorphism (SSCP; Orita et al., 1989) analysis of amplicons, bioinformatic characterization of nucleotide sequence data, and a series of reverse and direct dot-blot filter hybridization assays. Additionally, we estimated the genome sizes of the two species using flow cytometry. Therefore, we herein report for the first time the quantitative and qualitative differences between the genomes of I. paraguariensis and I. dumosa. Our approach could encourage research in other non-model plants, for which cutting-edge technologies (namely Next Generation Sequencing (NGS)) are still economically and methodologically unfeasible.
Materials and methods
Plant materials
Leaf samples of I. paraguariensis and I. dumosa were provided by the ‘Banco de Germoplasma de Yerba Mate y Té’ at the ‘Estación Experimental INTA Cerro Azul (EEINTA-CA)’, Misiones, Argentina. A total of two allopatric accessions were selected: I. paraguariensis EEINTA-CA no. CA1-74 (Misiones, Argentina), which is catalogued as the best leaf producer (Belingheri and Prat Kricum, 2000), and I. dumosa var. dumosa EEINTA-CA no. 20 (Santa Catarina, Brazil), which represents the wild phenotype (G. C. Giberti, pers. commun.). For DNA extraction and subsequent RDA assays, four individual plants per accession were used.
To prepare genomic probes for dot-blot assays, we used leaf samples of ‘yerba mate’ from accessions EEINTA-CA no. CA1-74 and CA4-74 and from six materials (namely SI-16♂, SI-19♀, SI-49♂, SI-67♀, G-18♀ and Y-383♂; Misiones, Argentina), kindly provided by ‘Establecimiento Las Marías’ (Gdor. Virasoro, Corrientes, Argentina). For I. dumosa, we used the leaf samples from the accession EEINTA-CA no. 20 and Argentinean accessions EEINTA-CA no. 7, 49, 48 and 222, and BACP no. 606, 607, 608, 609, 610 and 611 (Herbario del Centro de Estudios Farmacológicos y Botánicos – CEFYBO, held at “Museo de Farmacobotánica de la Facultad de Farmacia y Bioquímica”, Universidad de Buenos Aires, Argentina).
For flow cytometry analysis, we obtained fresh leaf samples from seven plantlets (‘yerba mate’: SI-19, SI-49, SI-67, YM1 and YM sn; ‘yerba señorita’: CRV7/4, CRV48/10, YS1 and YS sn), kindly donated by Dr P. A. Sansberro (‘Instituto de Botánica del Nordeste IBONE’, Universidad Nacional del Nordeste, Corrientes, Argentina).
DNA extraction
Young leaves preserved in silica gel were used for the extraction of total genomic DNA with the DNeasy Plant kit (Qiagen Inc., Valencia, California, USA), according to the manufacturer's instructions. Quality control and quantification were carried out by agarose gel (0.8%, w/v) electrophoresis and by comparison with a DNA molecular size standard (Lambda EcoRI/HindIII; Promega Corp., Madison, Wisconsin, USA). Gels were stained with ethidium bromide and photographed under UV light.
Representational difference analysis
The RDA can be performed without prior sequence information, and implies a subtractive hybridization technique applied to a representation of the genome (Zoldos et al., Reference Zoldos, Siljak-Yakovlev, Papes, Sarr and Panaud2001). This technique relies on the occurrence of missing restriction sites and/or genome rearrangements, such as deletions, insertions, duplications and translocations (Chen et al., Reference Chen, Phillips and Rines1998). We performed two reciprocal RDA assays. In the first assay, pooled genomic DNA from I. paraguariensis was employed as the tester (i.e. used in limited amounts) and that from I. dumosa was used as the driver to block the sequences shared by the two species, hereafter referred to as the RDA-Pa/Du assay. In the second assay, I. dumosa was used as the tester and I. paraguariensis as the driver, hereafter referred to as the RDA-Du/Pa assay. The RDA was performed, following the procedure proposed by Panaud et al. (Reference Panaud, Vitte, Hivert, Muzlak, Talag, Brar and Sarr2002), using approximately 400 ng of each pooled genomic DNA. The first oligonucleotide adapter set (R-set) was ligated to restriction fragments. Amplicons were obtained via polymerase chain reaction (PCR) as described previously by Panaud et al. (Reference Panaud, Vitte, Hivert, Muzlak, Talag, Brar and Sarr2002). The subtraction step was initiated by removing the R-adaptors from tester amplicons and then ligating the second set of adaptors (N-set). The N-ligated tester was mixed with the R-ligated driver in a 1:100 ratio, and ethanol-precipitated overnight ( − 20°C). The following steps (resuspension, hybridization, PCR amplification and nuclease treatment) were performed as described previously by Panaud et al. (Reference Panaud, Vitte, Hivert, Muzlak, Talag, Brar and Sarr2002). A second PCR was run for 20 cycles. Prior to the second round of subtractive hybridization, the N-adaptors were removed from the tester and replaced by the J-adaptor set. In this round, tester-J/driver-R ratios of 1:100, 1:500 and 1:1000 were assayed. In the third round, the J-adaptors were removed from the tester and replaced by the N-adaptor set. In this cycle, tester-N/driver-R ratios of 1:100, 1:500 and 1:1000 were used. Provisional results were checked at each step by electrophoresis in 2% (w/v) agarose gels, as described previously. The tester-N/driver-R amplicons derived from the third round (1:1000 ratio) were used for cloning with the pGEM-T Easy Vector Cloning Kit (Promega Corporation, Madison, Wisconsin, USA), following the manufacturer's instructions. Isolated white colonies were only selected for purification of the plasmid vector plus the insert, which was achieved using the QIAprep Spin Miniprep kit (Qiagen Inc., Valencia, California, USA).
Nucleotide screening of clones
The nucleotide variation in isolated RDA clones was initially screened via SSCP analyses (Orita et al., 1989), following the procedures described by Gottlieb and Poggio (Reference Gottlieb and Poggio2010). PCR amplicons were obtained from each purified clone by using universal primers T7 and SP6. Air-dried polyacrylamide gels were scanned and visually analysed. Clones were clustered according to their banding patterns.
Sequence characterization
From each SSCP cluster, at least one clone was selected for its characterization. Nucleotide sequences were obtained with universal primers T7 and SP6 at the ‘Servicio Interno de Genotipificación y Secuenciación de ADN–INTA’ (Castelar, Buenos Aires, Argentina), using an Applied Biosystems (Thermo Fischer Scientific, Waltham, Massachusetts, USA) automated 3130 XL DNA sequencer. Chromatograms were proofread and sequences manually edited in BioEdit (Hall, Reference Hall1999). Boundaries of the plasmid vector and of the oligonucleotide adaptors were determined and eliminated. All sequences were compared with each other, and the G+C content was calculated. Blast searches were performed using the BlastN, TBlastX and BlastX algorithms of the NCBI (http://www.ncbi.nlm.nih.gov/Blast) and E-value threshold (probability of random match) ≤ 10− 10. Additional searches were conducted against the Gypsy Database (Lloréns et al., Reference Lloréns, Futami, Bezemer and Moya2008, Reference Lloréns, Futami, Covelli, Dominguez-Escribá, Viu, Tamarit, Aguilar-Rodríguez, Vicente-Ripolles, Fuster, Bernet, Maumus, Munoz-Pomer, Sempere, Latorre and Moya2011) and the Conserved Domain Database (CDD) of the NCBI (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) using a threshold ≤ 10− 6. The RepeatMasker program (version open-4.0.1; Smit et al., 1996–Reference Smit, Hubley and Green2010) was used for the screening of interspersed repeats and low-complexity DNA sequences, assuming queries to be Arabidopsis thaliana. The sequences obtained herein were deposited in GenBank (HN280454–HN280499).
Dot-blot filter hybridization assays
Species-specificity assessment of the RDA-derived fragments and the estimation of their copy number within the genome of each species were performed through reverse dot-blot assays in triple- to quintuplicate, following the protocol described by Sabot et al. (Reference Sabot, Sourdille and Bernard2004). The digoxigenin labelling and detection kit (Roche, Pacheco, Buenos Aires, Argentina) was employed, according to the manufacturer's protocols, to generate non-radioactive species-specific genomic probes. We pooled equal amounts of total DNA from eight I. paraguariensis plants, on the one hand, and from 11 I. dumosa plants, on the other hand. A minimum of 3 μg from each bulked genomic DNA were labelled independently. The selected RDA-derived fragments were blotted onto a positively charged nylon filter membrane (Genescreen Plus, Perkin Elmer Inc., Boston, Massachusetts, USA), in aliquots (1–2 μl) of known amounts and fixed under UV light, according to the manufacturer's instructions. As controls for hybridization reactions and internal references for quantification of the signals, we blotted aliquots (1 μl) from serial dilutions (up to 156 pg/μl) of unlabelled bulked genomic DNA with respect to the probe used.
The molecular affinity between the genomes of the two species was assessed through direct dot-blot assays, in quadruplicates. Serially diluted DNA of each species (up to 110 pg/μl) was consecutively blotted, in 1 μl aliquots onto the same hybridization filter membrane. Each row functioned either as the control for hybridization reactions (intraspecific controls) and as the reference for quantification of the signals, or as the hybridization targets (i.e. for interspecific hybridizations). A control for the staining procedure, provided by the kit, was included in each assay. For each hybridization temperature assayed, two such membranes were prepared. At the established temperature, each membrane was hybridized with a single digoxigenin-labelled genomic probe. Optimal hybridization temperatures were calculated using the formula suggested by the kit which considers: fragment length, G+C content and average melting temperature subtracting 20°C or 25°C. The resulting optimal temperatures were 40 and 50°C. Post-hybridization washes comprised two 5 min washes in 2 × saline–sodium citrate (SSC) and 0.1% sodium dodecyl sulphate (SDS) at room temperature, followed by two 15 min washes in 0.5 × SSC and 0.1% SDS at the corresponding hybridization temperature. The experimental conditions used are regarded as stringent, allowing up to 18% mismatches between the probe and the targets. Staining was performed according to the manufacturer's protocol. Hybridization signals were quantified with Adobe Photoshop using the rapid selection tool and histogram function, and comparing the intensity value of each query dot with dot intensities of known amounts of the corresponding genomic DNA. Following Linares Miquel (Reference Linares Miquel1997), we assumed that differences in the intensity of hybridization signals were due to the differences in copy number. For estimation of copy number, we considered the quantity and the length of the blank DNA blotted, and the quantity of the reference DNA blotted and its genome size (Linares Miquel, Reference Linares Miquel1997). For I. paraguariensis and I. dumosa, we used the genome sizes estimated here (in pg; see below) multiplied by 0.978 × 10− 9 to obtain the DNA content in base pairs (bp) (Doležel et al., Reference Doležel, Bartoš, Voglmayr and Greilhuber2003).
Flow cytometry
Nuclear suspensions were prepared by simultaneously chopping 1 cm2 of fully expanded leaves (maintained either in darkness for 24 h or under natural conditions) or roots of Ilex spp., and 1 cm2 of the internal standard leaf (i.e. Zea mays cultivar CE-777, 2C = 5.43 pg; Lysák and Doležel, Reference Lysák and Doležel1998) using a razor blade and 2.5 ml of LB01 buffer (Doležel et al., Reference Doležel, Binarová and Lucretti1989) supplemented with 0.4% (w/v) of polyvinylpyrrolidone 10 and bovine serum albumin. The crude suspension was treated with 10 μl RNase (20 mg/ml), filtered through a nylon mesh (50 μm) and incubated for approximately 30 min at room temperature. Then, 500 μl propidium iodide (50 mg/ml) were added, and staining was allowed to develop for either 30 or 5 min in darkness. The measurements were made on a CyFlow Ploidy Analyser (Partec GmbH, Münster, Germany) at the facility of the ‘Instituto de Floricultura – INTA’ (Castelar, Buenos Aires, Argentina). Histograms were evaluated using the Flowing software 2.5.0 (P. Terho, Turku Centre for Biotechnology, University of Turku, Finland). For each sample, at least 5000 particles were analysed. Estimates of nuclear DNA content were considered only when the coefficients of variation were below 5%. The genome size was calculated by multiplying the DNA content of the standard (in pg) by the ratio of the mean fluorescent intensity of Ilex spp. to that of the standard. The seeds of the internal reference were kindly provided by Dr J. Doležel (Laboratory of Molecular Cytogenetics and Cytometry, Institute of Experimental Botany, Czech Republic). Estimation of genome sizes (in bp) was done as explained above.
Results
The RDA clones isolated from both assays (208 in total) were grouped according to their SSCP banding patterns in 51 clusters (data not shown). One to ten representatives of each cluster (67 in total) were sequenced and characterized with bioinformatic tools and similarity searches; in total, 7318 bp were generated for I. paraguariensis and 7420 bp for I. dumosa (4124 and 5879 bp, respectively, when redundancy is discounted). Based on the characterization, nucleotide data were arranged in 46 sequence types and, additionally, in three major groups (see Tables S1 and S2, available online). The first group involved sequences showing either spurious affinities or no similarity with any sequence deposited in the databases surveyed (group A, N= 91; Fig. 1(a)). The second group showed significant blast hits with plant entries at either the nucleotide or protein level (group B; N= 67). Eleven sequence types were recognized as conserved domains of retrotransposons (namely aspartic peptidase, chromodomain, integrase, polyprotein precursor, RNaseH and retrotranscriptase), which belonged to seven Ty3/Gypsy and Copia lineages (Fig. 1(b)). Moreover, the RepeatMasker program applied on group B detected an integrase domain of the Atlantys3I long terminal repeat (LTR)/Gypsy retrotransposon from Arabidopsis (sequence type 2.9; 63% identity; 72% query coverage). In addition, the CDD search detected conserved domains of metabolic enzymes (i.e. calcium-binding motif, tyrosine kinase catalytic domain and nucleoside diphosphate kinase active site). For sequence types 1.7, 2.13, 2.14 and 2.15, no additional information was obtained, and thus remained unidentified. The third group of sequences matched plant ribosomal, mitochondrial or plastid DNA (group C; N= 50).
(a) Schematic representation of the results from bioinformatic characterization and similarity searches. The description of groups A–C is detailed in the Results section. ‘Total’ refers to the total number of sequences per group. ‘Nucleotides’ refer to the number of sequences showing significant blast hits with entries from the databases surveyed at the nucleotide level. ‘Hypothetical proteins’ refer to the number of sequences showing significant blast hits to protein targets. ‘RTN’ refers to the number of sequences showing significant blast hits to retrotransposons. (b) Pie chart showing the retroelement composition in the representational difference analysis-derived sequences.
The dot-blot assays performed with isolated RDA-derived fragments showed differences in hybridization signals, at both temperatures (40 and 50°C) and with both genomic probes (Fig. 2(a)). The estimation of copy number for the unidentified sequences (group A) indicated that I. paraguariensis had 1.20 × 105 copies on average (range 3.95 × 104–3.22 × 105) and that I. dumosa had 6.45 × 104 copies (range 1.74 × 103–2.66 × 105). Most fragments appeared more abundant in the genome of I. paraguariensis (range 1.21- to 39.62-fold; Table 1). For retrotransposon sequences, the genomes of both species showed a similar average content; I. paraguariensis had 1.14 × 105 copies (range 1.85 × 103–2.88 × 105), whereas I. dumosa contained 1.08 × 105 copies (range 2.03 × 103–2.63 × 105). However, most fragments appeared more abundant in the genome of I. dumosa (range 1.41- to 35.77-fold; Table 1).
(a) Examples of filter hybridizations of representational difference analysis-derived fragments by means of reverse dot-blot assays. The membrane showing dot blots from fragments of groups A and B, hybridized with a digoxigenin (DIG)-labelled genomic probe from Ilex paraguariensis. The amount (in ng) of unlabelled genomic DNA from I. paraguariensis is indicated in bold font. The dot blots are named as in Table 1. A staining control (Co) dot is included. (b) The membrane showing dot blots hybridized with a DIG-labelled genomic probe from Ilex dumosa. The amount (in ng) of unlabelled I. dumosa genomic DNA is indicated in bold font. (c) Examples of filter hybridizations of genomic DNA by means of direct dot-blot assays. Hybridization with a DIG-labelled genomic probe from I. paraguariensis is shown. The dot blots of the upper row correspond to the Co and to the serial dilutions of unlabelled total DNA from I. dumosa; the dot blots of the lower row correspond to the serial dilutions of unlabelled total DNA from I. paraguariensis (i.e. intraspecific control). Here, 3.4 ng of genomic DNA from I. dumosa have an intensity equal to 0.11 ng of genomic DNA from I. paraguariensis. (d) Hybridization with a DIG-labelled genomic probe from I. dumosa is shown. Here, the dot blots of the upper row correspond to the Co and to the intraspecific control (i.e. serial dilutions of unlabelled total DNA from I. dumosa); the dot blots of the lower row correspond to the serial dilutions of unlabelled total DNA from I. paraguariensis. Here, 7 ng of genomic DNA from I. dumosa are equivalent to 4.5 ng of genomic DNA from I. paraguariensis.
Estimation of copy number for selected sequence types of groups A and B
RDA, representational difference analysis.
a For GenBank accession numbers, see Tables S1 and S2 (available online).
b Average of triple to quintuple replicates. The numbers in bold and italic fonts represent the maximum and minimum values per group of sequences, respectively.
c Ratio of the average copy number estimate of I. paraguariensis to I. dumosa; the values given in parentheses indicate the inverse ratio.
The dot-blot assays for testing interspecific genomic affinity showed that, at a more stringent condition (50°C), only intraspecific hybridization signals were detected. At the lower stringency (40°C), both intra- and interspecific signals were evident, but I. paraguariensis was always detected more efficiently, even with genomic probes from I. dumosa. The quantification of signal intensities indicated that I. dumosa required, on average, 4.16–22 times higher amounts of DNA to produce dot intensities equivalent to those yielded by I. paraguariensis (Fig. 2(b)).
The estimation of genome size by means of flow cytometry suggested that, under the standard conditions (30 min nuclei staining), the cytosolic compounds of I. paraguariensis severely interfered with the dying process, affecting even the internal standard and impeding any measurement (data not shown). Depending on the samples surveyed from I. dumosa, we detected two opposite performances; the cytosolic crude extracts cause either a severe interference of nuclei staining or only a mild effect. When the nuclei staining procedure was developed for 5 min, it yielded clear peaks for all the samples, tissues (leaves or roots) and growth conditions (light and darkness) (Fig. 3). Thus, for I. paraguariensis, the estimated genome size value was 2C = 1.71 ± 0.08 pg ( = 1670 Mbp) and for I. dumosa the value was 2C = 1.89 ± 0.07 pg ( = 1848 Mbp).
Overlay flow cytometry histograms [nuclei counts vs. relative fluorescence (propidium iodide, PI) intensity]. The shaded area shows the fluorescence peak derived from nuclei preparations of Zea mays (cultivar CE-777, 2C = 5.43 pg) when measured alone. The white area shows the fluorescence peaks derived from the simultaneous analysis of nuclei preparations from the leaves of Ilex spp. and leaves of maize (as the internal standard), showing the impact that the cytosolic leaf components of Ilex has on the fluorescence of the standard. (a) Overlay histogram of Ilex paraguariensis (YM sn) plus the internal standard (maize), and that of maize alone (shaded area). (b) Overlay histogram of Ilex dumosa (YS sn) plus the standard, and that of maize alone (shaded area).
Discussion
The trees I. paraguariensis (‘yerba mate’) and I. dumosa (‘yerba señorita’) possess peculiar phytochemical properties valued for human consumption, for which they have been largely studied (for a review, see Heck and de Mejia, Reference Heck and de Mejia2007). However, information about the basic genetic aspects of these species is scarce. At present, only the natural populations of I. paraguariensis have been formally investigated under a microevolutionary framework (Gauer and Cavalli-Molina, Reference Gauer and Cavalli-Molina2000; Neumann-Wendt, Reference Neumann-Wendt2005; Cascales, Reference Cascales2013; Pereira et al., 2013). Those studies agree in showing evidence for low population divergence, explained by the occurrence of gene flow mediated by birds and/or by a recent fragmentation of habitats. Although such population data are lacking for I. dumosa, the phylogenetic and genetic relationships for these species have been assessed (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2005, Reference Gottlieb, Giberti and Poggio2011). Still, the paucity of molecular information can be evidenced by a simple GenBank search. Currently, in GenBank, there are 77 entries for I. paraguariensis, most of which are microsatellite sequences (41), followed by chloroplast sequences (22) and by nuclear entries (14) corresponding to nine different regions. For I. dumosa, the data available are even fewer; there are only nine nuclear entries for three regions and four plastid sequences. In this study, we have characterized 67 genomic regions with bioinformatic tools, thus increasing the availability of non-redundant sequence data by 18% for I. paraguariensis and by 114% for I. dumosa.
Moreover, the information provided here demonstrates that the two species differ quantitatively and qualitatively in highly repetitive DNA, and in their genome sizes. The repeated recovery of some genomic regions indicates that the SSCP, while helped detecting sequence redundancies, overestimated the actual nucleotide variation of some fragments; at the same time, it suggests that isolated regions mostly involve repetitive DNA. A quantitative difference detected between the species relies on the group of non-coding sequences (group A), which appear on average 8.5 times more abundantly in the genome of I. paraguariensis than in the genome of I. dumosa. The high copy numbers estimated for this group clearly support their classification as highly repeated DNA. It is known that hybridization signals revealed with total DNA probes, such as those used here, are significant only if blotted sequences are highly repeated (Sabot et al., Reference Sabot, Sourdille and Bernard2004). The fact that this group shows no significant similarity with any other sequence data deposited in public repositories inclines us to follow the criteria applied by Nagarajan et al. (Reference Nagarajan, Navajas-Pérez, Pop, Alam, Ming, Paterson and Salzberg2008) and to consider these as genomic regions specific for the genus Ilex. These repetitive sequences will be used as molecular probes in fluorescence in situ hybridization assays, to examine their distribution pattern over the chromosomal complement. Also, they have the potentiality for development of random and/or region-specific markers, with specificity for these particular species and with potentiality for future exploration of breeding populations.
Our outcomes also indicate that genomes of both species show a similar average copy number of the mobile elements (group B) isolated herein. We are aware that our estimates may be biased since we cannot distinguish full-length elements (i.e. autonomous) from truncated retrotransposons (i.e. non-autonomous, possibly nested, defective or inactive), though similar estimates have been reported for other plants. The genome of cocoa tree, Theobroma cacao, was estimated to contain approximately 6.75 × 104 copies of transposable elements, although this estimate was regarded as a rough underestimation (Argout et al., Reference Argout, Salse, Aury, Guiltinan, Droc, Gouzy, Allegre, Chaparro, Legavre, Maximova, Abrouk, Murat, Fouet, Poulain, Ruiz, Roguet, Rodier-Goud, Barbosa-Neto, Sabot, Kudrna, Ammiraju, Schuster, Carlson, Sallet, Schiex, Dievart, Kramer, Gelley, Shi, Bérard, Viot, Boccara, Risterucci, Guignon, Sabau, Axtell, Ma, Zhang, Brown, Bourge, Golser, Song, Clement, Rivallan, Tahi, Akaza, Pitollat, Gramacho, D'Hont, Brunel, Infante, Kebe, Costet, Wing, McCombie, Guiderdoni, Quetier, Panaud, Wincker, Bocs and Lanaud2011). The genome of Populus trichocarpa has over 5000 copies of retrotransposons with a prevalence of Gypsy-like elements (Tuskan et al., Reference Tuskan, DiFazio, Jansson, Bohlmann, Grigoriev, Hellsten, Putnam, Ralph, Rombauts, Salamov, Schein, Sterck, Aerts, Bhalerao, Bhalerao, Blaudez, Boerjan, Brun, Brunner, Busov, Campbell, Carlson, Chalot, Chapman, Chen, Cooper, Coutinho, Couturier, Covert, Cronk, Cunningham, Davis, Degroeve, Déjardin, Depamphilis, Detter, Dirks, Dubchak, Duplessis, Ehlting, Ellis, Gendler, Goodstein, Gribskov, Grimwood, Groover, Gunter, Hamberger, Heinze, Helariutta, Henrissat, Holligan, Holt, Huang, Islam-Faridi, Jones, Jones-Rhoades, Jorgensen, Joshi, Kangasjärvi, Karlsson, Kelleher, Kirkpatrick, Kirst, Kohler, Kalluri, Larimer, Leebens-Mack, Leplé, Locascio, Lou, Lucas, Martin, Montanini, Napoli, Nelson, Nelson, Nieminen, Nilsson, Pereda, Peter, Philippe, Pilate, Poliakov, Razumovskaya, Richardson, Rinaldi, Ritland, Rouzé, Ryaboy, Schmutz, Schrader, Segerman, Shin, Siddiqui, Sterky, Terry, Tsai, Uberbacher, Unneberg, Vahala, Wall, Wessler, Yang, Yin, Douglas, Marra, Sandberg, Van de Peer and Rokhsar2006); recently, 1479 full-length LTR retrotransposons were identified in poplar (Cossu et al., Reference Cossu, Buti, Giodani, Natali and Cavallini2012). In Eucalyptus, 24–226 copies of transcriptionally active Copia-like LTR retrotransposons have been reported (Marcon et al., Reference Marcon, Domingues and Marino2011). A preliminary analysis of repetitive DNA from pilot bacterial artificial chromosome (BAC)-end sequences indicates that over 98% of the genome of the tea crop, Camellia sinensis, could be repetitive (Lin et al., Reference Lin, Kudrna and Wing2011). We detected conserved domains associated with the replication and integration of mobile elements from the Ty3/Gypsy Del lineage, Chromoviridae branch (i.e. Del, Retrosat-2 and Peabody), the Athila/Tat lineage (i.e. Calypso5-1, Tat4-1 and B1147A04.5) and the Ty1/Copia Retrofit clade (i.e. Vitico1-2) according to the current classification of retrotransposons (Lloréns et al., Reference Lloréns, Futami, Bezemer and Moya2008, Reference Lloréns, Futami, Covelli, Dominguez-Escribá, Viu, Tamarit, Aguilar-Rodríguez, Vicente-Ripolles, Fuster, Bernet, Maumus, Munoz-Pomer, Sempere, Latorre and Moya2011). We found that the Ilex paraguariensis retrotransposon element (IPRE) retroelement from I. paraguariensis (Gottlieb and Poggio, Reference Gottlieb and Poggio2010) is also present in I. dumosa (sequence type 2.16). This is somehow expectable because the Chromoviridae branch is ancient and ubiquitous (Lloréns et al., Reference Lloréns, Futami, Bezemer and Moya2008, Reference Lloréns, Futami, Covelli, Dominguez-Escribá, Viu, Tamarit, Aguilar-Rodríguez, Vicente-Ripolles, Fuster, Bernet, Maumus, Munoz-Pomer, Sempere, Latorre and Moya2011); yet, this is the first report of retroelements isolated from I. dumosa. As in many plant species, it seems plausible that a vast proportion of the genomes of ‘yerba mate’ and ‘yerba señorita’ are represented by repetitive DNA. The pervasiveness of the repetitive DNA, a lack of sufficient driver competition during the hybridization step, a linear increase in contaminating heteroduplexes, an inefficient annealing in the complex mixture after the subtraction step, and/or a trapping of a fraction of the tester molecules in heteroduplexes (Kuvachieva and Goffinet, Reference Kuvachieva and Goffinet2002) would explain the absence of species-specific fragments among the products isolated. Regarding the sequences identified as organellar or ribosomal DNA (group C), albeit they represent a minor fraction of the isolated products (19–31%), they are non-differential products that could have been generated due to the ubiquity of organellar DNA in preparations from leaf tissue (Gottlieb and Poggio, Reference Gottlieb and Poggio2010; Shi et al., Reference Shi, Hu, Huang, Gao, Zhao and Gao2012).
In addition, we detected protein domains from enzymes involved in plants' primary and secondary metabolism and signal transduction and regulation (Grabarek, Reference Grabarek2006; Hammargren et al., Reference Hammargren, Sundström, Johansson, Bergman and Knorpp2007; Tonfack et al., Reference Tonfack, Moummou, Latché, Youmbi, Benichou, Pech and Van der Rest2011; Bienert et al., Reference Bienert, Delannoy, Navarre and Boutry2012; Moummou et al., Reference Moummou, Kallberg, Tonfack, Persson and van der Rest2012). Particularly, two fragments show similarity to a flavonoid-reductase involved in lignin biosynthesis, and to a serine carboxypeptidase that may contribute to defensive chemistry towards insects (Mugford et al., Reference Mugford, Qi, Bakht, Hill, Wegel, Hughes, Papadopoulou, Melton, Philo, Sainsbury, Lomonossoff, Roy, Goss and Osbourn2009), as was observed in wild Solanaceae (Li and Steffens, Reference Li and Steffens2000). These sequences deserve a close examination, for instance by comparing their expression levels and investigating any relationship with psilid resistance.
It is worth remarking that Ilex spp. studied here are not closely related, nor are their sister taxa. Their phylogenetic relationship has not been clearly established yet because I. paraguariensis shows an unstable phylogenetic position (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2005). Moreover, I. dumosa shows either a relationship with the Asian species Ilex crenata and Ilex mutchagara (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2005) or with Ilex argentina (Manen et al., Reference Manen, Boulter and Naciri-Graven2002), but not with ‘yerba mate’. Our hybridization outcomes suggest that the genomes of both species have less than 80% similarity. This value is a little lower than the value reported previously, on the basis of amplified fragment length polymorphism analysis, which points to an approximately 23% difference (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2011). That study has also shown that, in comparison with other southern South American Ilex spp. (i.e. I. argentina, Ilex brasiliensis, Ilex brevicuspis, Ilex theezans and Ilex integerrima), I. paraguariensis shows a range of interspecific genetic differences of 22–24.5%, whereas I. dumosa shows a range of higher interspecific differences (24–27%) (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2005). In addition, the genetic distance between the two species duplicates and quadruplicates their respective intraspecific value (Gottlieb et al., Reference Gottlieb, Giberti and Poggio2005). Thus, the overall genetic difference between I. paraguariensis and I. dumosa obtained here is concurrent with previous results. Additionally, we detected contrasting performances through filter hybridizations that could be explained by the occurrence of a class (or classes) of homologous but slightly divergent repetitive DNA, highly amplified in the ‘yerba mate’ genome but not in ‘yerba señorita’. In support of our proposal, the repetitive sequences detected in I. paraguariensis outnumber (by up to 40-fold) those found in I. dumosa. These repetitive DNA sequences could include any of the telomeric or centromeric sequences, long or short repeats, micro- or minisatellites, satellite DNA, ribosomal tandem repeats, and transposons or retrotransposons (Sharma and Raina, Reference Sharma and Raina2005). This deserves a careful verification through physical mapping.
As mentioned previously, the leaves of ‘yerba mate’ have abundant caffeine and chlorogenic acids (Filip et al., Reference Filip, López, Coussio and Ferraro1999, Reference Filip, López, Giberti, Coussio and Ferraro2001; Reginatto et al., Reference Reginatto, Athayde, Gosmann and Schenkel1999). The effect of cytosolic chlorogenic acids on flow cytometry fluorescence was demonstrated by Noirot et al. (Reference Noirot, Barre, Duperray, Louarn and Hamon2003, Reference Noirot, Barre, Duperray, Hamon and De Kochko2005) in Coffea. The presence of a complex mixture of secondary metabolites in cytosolic preparations of I. paraguariensis could account for the failure in measuring its nuclear content under conventional procedures. By using a highly reducing isolation buffer and short incubation times, we were able to estimate the genome size via flow cytometry for both species, for the first time. For I. paraguariensis, it has been shown that the estimate value results in 1.33 times lower than the one previously inferred by microdensitometry (2C = 2200 Mbp; Barral et al., Reference Barral, Poggio and Giberti1995). Notwithstanding, the value obtained herein represents a more precise estimate, being an average of 24,119 nuclei, compared with the 384 nuclei measured by Barral et al. (Reference Barral, Poggio and Giberti1995). In I. dumosa, we observed a variable performance that may be due to an intraspecific variation in the abundance of cytosolic compounds; a fact not previously reported for this species. The genome size reported herein for I. dumosa constitutes a novel contribution. The difference in the DNA content between I. paraguariensis and I. dumosa (approximately 10%) does not appear to be related to the chromosome number since both species have 2n= 40 chromosomes (Barral et al., Reference Barral, Poggio and Giberti1995; Greizerstein et al., Reference Greizerstein, Giberti and Poggio2004); a correlation with other cytogenetic parameters remains to be investigated.
The genetic information gathered herein contributes to the scientific knowledge of the two species considered, offers a basis for developing suitable probes for physical mapping and genome organization studies, and certainly will facilitate the design of future genomic projects.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1479262114000756
Acknowledgements
The authors are indebted to S. D. Prat Kricun, L. D. Belingheri and R. M. Mayol from EE INTA Cerro Azul (Misiones, Argentina), to the Establecimiento Las Marías (Corrientes, Argentina) and to P. A. Sansberro from IBONE-UNNE (Corrientes, Argentina), who kindly provided specimens from living collections. We are also grateful to G. C. Giberti who identified the plants and to M. S. Rodriguero and anonymous referees who helped improve the present and earlier versions of the manuscript. The authors gratefully acknowledge the grants from the Consejo Nacional de Investigaciones Científicas y Técnicas (PIP 12-200801-00342), the Universidad de Buenos Aires (UBA 20020100100859) and the Agencia Nacional de Investigaciones Científicas y Técnicas (PICT 2010-1665 and PICT 2011-2420).
