Use of Repetitive Sequences for Molecular and Cytogenetic Characterization of Avena Species from Portugal

Genomic diversity of Portuguese accessions of Avena species—diploid A. strigosa and hexaploids A. sativa and A. sterilis—was evaluated through molecular and cytological analysis of 45S rDNA, and other repetitive sequences previously studied in cereal species—rye subtelomeric sequence (pSc200) and cereal centromeric sequence (CCS1). Additionally, retrotransposons and microsatellites targeting methodologies—IRAP (inter-retrotransposon amplified polymorphism) and REMAP (retrotransposon-microsatellite amplified polymorphism)—were performed. A very high homology was detected for ribosomal internal transcribed sequences (ITS1 and ITS2) between the species analyzed, although nucleolar organizing regions (NOR) fluorescent in situ hybridization (FISH) analysis revealed distinct number of Nor loci between diploid and hexaploid species. Moreover, morphological diversity, evidenced by FISH signals with different sizes, was observed between distinct accessions within each species. pSc200 sequences were for the first time isolated from Avena species but proven to be highly similar in all genotypes analyzed. The use of primers designed for CCS1 unraveled a sequence homologous to the Ty3/gypsy retrotransposon Cereba, that was mapped to centromeric regions of diploid and hexaploid species, being however restricted to the more related A and D haplomes. Retrotransposon-based methodologies disclosed species- and accessions-specific bands essential for the accurate discrimination of all genotypes studied. Centromeric, IRAP and REMAP profiles therefore allowed accurate assessment of inter and intraspecific variability, demonstrating the potential of these molecular markers on future oat breeding programs.


Introduction
Avena genus (2n = 2x = 14) comprises diploid (A and C genomes), tetraploid (AB and AC genomes) and hexaploid (ACD) species. Hexaploid species appear to have arisen by two rounds of hybridization followed by chromosome duplication whereas tetraploid species with AB genomes were suggested to result from an unrelated event involving an A diploid species autopolyploidization (for a review see [1]). Structurally different A and C genomes have variants and B and D genomes are considered to be derived from an A genome [1]. Only four of 27 Avena species [2] are cultivated and used for food, feed or forage, namely: diploid A. strigosa (A), tetraploid A. abyssinica (AB) and hexaploids A. byzantine and A. sativa (ACD) [1]. A. strigosa and A. sativa are the major Avena crops in Portugal, which is considered the center of diversity of the former one [1]. On the other hand, the most relevant Avena wild species in Portugal is A. sterilis [3] with a high potential in oat breeding programs due to important agronomic traits increased grain yield and protein content and resistance to abiotic and biotic stresses [4]. In fact, natural hybrids between A. sativa and A. sterilis occur frequently and are widespread in Portugal [3].
In this work, the characterization of distinct Portuguese accessions of A. strigosa, A. sativa and A. sterilis was performed through the analysis of distinct repetitive genome fractions, including Nor loci (ITS1 and ITS2) as well as sequences previously mapped in other cereals at centromeric (CCS1) and subtelomeric (pSc200) chromosome domains. Sequences dispersed throughout the genome such as retrotransposons and microsatellites were also analyzed through IRAP and REMAP methodologies. Besides interspecific variability, the results obtained also disclosed intraspecific genetic diversity in Portuguese Avena species allowing the establishment of molecular markers that may be useful to reliably identify valuable genotypes.

Results
To assess Avena diversity we analyzed distinct repetitive sequences such as 45S rDNA ITS1 and ITS2 and sequences previously mapped on subtelomeric (pSc200) and centromeric (CCS1) domains in other cereal species. Retrotransposons and microsatellites flancking sequences were also analyzed through IRAP (Inter Retrotransposons Amplified Polymorphism) and REMAP (Retrotransposons Microsatellite Amplified Polymorphism) methodologies. All molecular markers used revealed unique and consistent banding profiles in at least three distinct individuals of each genotype of A. strigosa, A. sativa and A. sterilis demonstrating the inexistence of intravarietal variability. The total number of bands per molecular marker used and genotype analyzed are summarized in Table 1. 45S rDNA ITS1 and ITS2 and pSc200 amplification yielded a single band in all Avena genotypes that was further isolated, purified and cloned for sequences analysis (Accession Numbers in Table S1).

45S rDNA Internal Transcribed Sequences Similarity and Nor Loci Diversity
Primers targeting 45S rDNA sequences [29] produced single bands with the expected sizes (~300 and ~380 bp, respectively) in all species/accessions analyzed as well as in pTa71 plasmid used as control. ITS1 and ITS2 sequences included in those amplified products present 219 and 216 bp, respectively, as already described for other Avena species [11]. Moreover, ITS sequence analysis using original parameters of Clustal W2 [30] revealed a high level of similarity-between 98% and 100% for ITS1 and between 99% and 100% for ITS2. ITS1 sequences alignment presented six polymorphic nucleotides and 100% identity was detected between A. sativa "S. Eulália" and A. sterilis. In ITS2 sequences only one polymorphic nucleotide was detected between A. strigosa and A. sativa "Madeira Island" from A. sativa Kyto, S. Eulália and A. sterilis.
Nor loci evaluation performed in root tip cells of A. strigosa landraces revealed four FISH positive NORs using pTa71 probe (Figure 1), presenting however some intraspecific morphological variability. In fact, in "Madeira Island" landrace two pairs of NORs with distinct sizes are observed ( Figure 2a) and in "Elvas" landrace the four NORs exhibit similar sizes (Figure 2b). In hexaploid species A. sativa and A. sterilis, 45S rDNA FISH revealed three pair of signals ( Figure 2). However, in A. sativa distinct landraces revealed Nor loci diversity in relation to NORs size. While in "S. Eulália" all three NOR pairs exhibit similar sizes (Figure 2a), "Madeira Island" landrace presented FISH signals with different sizes corresponding to one pair of large NORs, one pair of medium size NORs and one heteromorphic pair composed by one medium size and one minor NOR (Figure 2b). Furthermore, in A. sterilis prometaphase cells with distended chromosomes, In hexaploid species A. sativa and A. sterilis, 45S rDNA FISH revealed three pair of signals ( Figure 2). However, in A. sativa distinct landraces revealed Nor loci diversity in relation to NORs size. While in "S. Eulália" all three NOR pairs exhibit similar sizes (Figure 2a), "Madeira Island" landrace presented FISH signals with different sizes corresponding to one pair of large NORs, one pair of medium size NORs and one heteromorphic pair composed by one medium size and one minor NOR ( Figure 2b). Furthermore, in A. sterilis prometaphase cells with distended chromosomes, the three NOR pairs are distinguished through their different topological organization since one pair presents a single FISH domain whereas the other two pairs show more than one labeled domain intercalated by an unlabeled region that corresponds most probably to transcriptionally active uncondensed rDNA units. In one pair of NORs the two domains have similar dimensions, contrasting with the other pair where both domains have different dimensions (Figure 2c). The numbers of rDNA FISH signals and their described organization patterns were observed in at least 10 metaphase or prometaphase cells and are summarized in Table 2.  (Figure 2c). The numbers of rDNA FISH signals and their described organization patterns were observed in at least 10 metaphase or prometaphase cells and are summarized in Table 2.

Novel Repetitive Sequences Were Identified Contributing for Avena Accessions Discrimination
Genomic diversity evaluation in Avena was extended through the analysis of repetitive sequences previously described in subtelomeric and centromeric domains in other cereal species [26]. The amplification products obtained with primers designed to pSc200 sequence are presented in Figure 3. A 444 bp fragment is obtained, as a result of the amplification between the forward and reverse primers in subsequent units, as already described in [31], including in all species/accessions analyzed a single complete unit with 379 bp as described in rye [32], and where a significantly more intense band is observed when rye genomic DNA is used as control. Those bands were isolated, cloned and sequenced (Accession Numbers in Table S1) and their BLAST (Basic Local Alignment Search Tool) analysis in NCBI database revealed between 99%-100% similarity with rye subtelomeric sequence (Accession Number Z50039, [32] and 90% homology with a A. sativa sequence (Genbank Accession Number DQ481575, unpublished data). Moreover, the homology level between pSc200-like sequences isolated from the distinct Avena genotypes analyzed (alignment in Figure S1) is extremely high (~99%).

Novel Repetitive Sequences Were Identified Contributing for Avena Accessions Discrimination
Genomic diversity evaluation in Avena was extended through the analysis of repetitive sequences previously described in subtelomeric and centromeric domains in other cereal species [26]. The amplification products obtained with primers designed to pSc200 sequence are presented in Figure 3. A 444 bp fragment is obtained, as a result of the amplification between the forward and reverse primers in subsequent units, as already described in [31], including in all species/accessions analyzed a single complete unit with 379 bp as described in rye [32], and where a significantly more intense band is observed when rye genomic DNA is used as control. Those bands were isolated, cloned and sequenced (Accession Numbers in Table S1) and their BLAST (Basic Local Alignment Search Tool) analysis in NCBI database revealed between 99%-100% similarity with rye subtelomeric sequence (Accession Number Z50039, [32] and 90% homology with a A. sativa sequence (Genbank Accession Number DQ481575, unpublished data). Moreover, the homology level between pSc200-like sequences isolated from the distinct Avena genotypes analyzed (alignment in Figure S1) is extremely high (~99%). FISH analysis on the Avena genotypes studied was performed using pSc200 probe and rDNA pTa71 probe as positive control. Contrary to the rDNA probe that was detected in all metaphase spreads, the pSc200 probe was not detected in any Avena accession.
Another repetitive sequence hitherto mapped in cereal centromeric regions (CCS1) was studied in Avena through PCR amplification and FISH evaluation. The PCR profiles obtained are distinct in the three Avena species analyzed and contain bands ranging in size from 350 to 2500 bp (schematic representation in Figure 4). Two bands are common to all genotypes analyzed, one with a size identical (950 bp) to the published maize centromeric sequence [28] and a novel one with 700 bp. Both A. strigosa landraces exhibit identical profiles, containing besides the two common bands, three other bands, two A. strigosa-specific (~380 and ~1700 bp) and one common to A. sativa (~1650 bp). A. sativa and A. sterilis banding profiles present one band characteristic of Avena hexaploid species with ~2300 bp as well as a A. sativa-specific band with ~2500 bp, which distinguishes the hexaploid species studied. Also, A. sativa landraces present very similar profiles composed of five A. sativa common bands previously described, plus "Kyto" or "Madeira Island" specific bands (~900 and ~1100 bp, respectively). In A. sterilis, besides the two Avena characteristic bands and the hexaploid-specific band, a fourth band with ~800 bp is observed similar to the one present in A. sativa landrace from Madeira Island. The Avena characteristic novel band with ~700 bp was isolated, cloned and sequenced (Accession Numbers in Table S1) and the similarity values between genotypes range from 89% to 98% (Alignment in Figure S2). No similarity with other Avena centromeric sequences already published in Genbank was found. However, it yielded significant alignment (~72%) with retrotransposon-related centromeric sequences of Triticum (Acession Number HF541873.1) [33], Secale (Accession Number JQ963524.1) [34] and Hordeum (Accession Number AY040833.1) [35], FISH analysis on the Avena genotypes studied was performed using pSc200 probe and rDNA pTa71 probe as positive control. Contrary to the rDNA probe that was detected in all metaphase spreads, the pSc200 probe was not detected in any Avena accession.
Another repetitive sequence hitherto mapped in cereal centromeric regions (CCS1) was studied in Avena through PCR amplification and FISH evaluation. The PCR profiles obtained are distinct in the three Avena species analyzed and contain bands ranging in size from 350 to 2500 bp (schematic representation in Figure 4). Two bands are common to all genotypes analyzed, one with a size identical (950 bp) to the published maize centromeric sequence [28] and a novel one with 700 bp. Both A. strigosa landraces exhibit identical profiles, containing besides the two common bands, three other bands, two A. strigosa-specific (~380 and~1700 bp) and one common to A. sativa (~1650 bp). A. sativa and A. sterilis banding profiles present one band characteristic of Avena hexaploid species with~2300 bp as well as a A. sativa-specific band with~2500 bp, which distinguishes the hexaploid species studied. Also, A. sativa landraces present very similar profiles composed of five A. sativa common bands previously described, plus "Kyto" or "Madeira Island" specific bands (~900 and~1100 bp, respectively). In A. sterilis, besides the two Avena characteristic bands and the hexaploid-specific band, a fourth band with~800 bp is observed similar to the one present in A. sativa landrace from Madeira Island. FISH analysis on the Avena genotypes studied was performed using pSc200 probe and rDNA pTa71 probe as positive control. Contrary to the rDNA probe that was detected in all metaphase spreads, the pSc200 probe was not detected in any Avena accession.
Another repetitive sequence hitherto mapped in cereal centromeric regions (CCS1) was studied in Avena through PCR amplification and FISH evaluation. The PCR profiles obtained are distinct in the three Avena species analyzed and contain bands ranging in size from 350 to 2500 bp (schematic representation in Figure 4). Two bands are common to all genotypes analyzed, one with a size identical (950 bp) to the published maize centromeric sequence [28] and a novel one with 700 bp. Both A. strigosa landraces exhibit identical profiles, containing besides the two common bands, three other bands, two A. strigosa-specific (~380 and ~1700 bp) and one common to A. sativa (~1650 bp). A. sativa and A. sterilis banding profiles present one band characteristic of Avena hexaploid species with ~2300 bp as well as a A. sativa-specific band with ~2500 bp, which distinguishes the hexaploid species studied. Also, A. sativa landraces present very similar profiles composed of five A. sativa common bands previously described, plus "Kyto" or "Madeira Island" specific bands (~900 and ~1100 bp, respectively). In A. sterilis, besides the two Avena characteristic bands and the hexaploid-specific band, a fourth band with ~800 bp is observed similar to the one present in A. sativa landrace from Madeira Island. The Avena characteristic novel band with ~700 bp was isolated, cloned and sequenced (Accession Numbers in Table S1) and the similarity values between genotypes range from 89% to 98% (Alignment in Figure S2). No similarity with other Avena centromeric sequences already published in Genbank was found. However, it yielded significant alignment (~72%) with retrotransposon-related centromeric sequences of Triticum (Acession Number HF541873.1) [33], Secale (Accession Number JQ963524.1) [34] and Hordeum (Accession Number AY040833.1) [35], The Avena characteristic novel band with~700 bp was isolated, cloned and sequenced (Accession Numbers in Table S1) and the similarity values between genotypes range from 89% to 98% (Alignment in Figure S2). No similarity with other Avena centromeric sequences already published in Genbank was found. However, it yielded significant alignment (~72%) with retrotransposon-related centromeric sequences of Triticum (Acession Number HF541873.1) [33], Secale (Accession Number JQ963524.1) [34] and Hordeum (Accession Number AY040833.1) [35], targeting the long terminal region of the mobile element Ty3/gypsy retrotransposon Cereba firstly characterized in H. vulgare [35].
The Avena novel sequence amplified from A. sativa was labeled with biotin-dUTP and used as a probe for FISH evaluation in spreads of the species studied ( Figure 5). In A. strigosa metaphase cells all centromeric domains are FISH-positive using that centromeric probe (Figure 5a-c) and although hexaploid species also revealed discrete signals, only 28 from the 42 chromosomes were FISH labeled (Figure 5d-f). Thus, 14 chromosomes, probably composing one Avena haplome, do not show centromeric labeling.
Conclusively, the comparative analysis of amplification products obtained with CCS1 primers allows a clear discrimination of the three Avena species in the study. Moreover, A. sativa accessions exhibit distinct banding profiles since two landrace-specific bands were identified. The two A. strigosa accessions are not distinguishable, thus other molecular markers were further used to seek for a complete discrimination of all Avena genotypes analyzed. targeting the long terminal region of the mobile element Ty3/gypsy retrotransposon Cereba firstly characterized in H. vulgare [35].
The Avena novel sequence amplified from A. sativa was labeled with biotin-dUTP and used as a probe for FISH evaluation in spreads of the species studied ( Figure 5). In A. strigosa metaphase cells all centromeric domains are FISH-positive using that centromeric probe (Figure 5a-c) and although hexaploid species also revealed discrete signals, only 28 from the 42 chromosomes were FISH labeled (Figure 5d-f). Thus, 14 chromosomes, probably composing one Avena haplome, do not show centromeric labeling.
Conclusively, the comparative analysis of amplification products obtained with CCS1 primers allows a clear discrimination of the three Avena species in the study. Moreover, A. sativa accessions exhibit distinct banding profiles since two landrace-specific bands were identified. The two A. strigosa accessions are not distinguishable, thus other molecular markers were further used to seek for a complete discrimination of all Avena genotypes analyzed.

Retrotransposon and Microsatellite Flanking Sequences Untangled Avena Accessions Discrimination
The amplification of sequences flanking Nikita retrotransposon LTRs is presented in Figure 6. Banding profiles analysis allowed the identification of three bands common to all genotypes with ~400, ~650 and ~1300 bp (schematic representation in Figure 6). Both A. strigosa accessions showed identical banding patterns with two species-specific bands (~600 and ~1700 bp) besides the three common ones. In A. sativa two landrace-specific bands were detected in "Kyto" and "S. Eulália" (~1400 and ~2400 bp, respectively) and A. sativa from Madeira Island lacks the ~1250 bp band present in all other genotypes analyzed. Finally, the A. sterilis Pt banding profile includes a specific band

Retrotransposon and Microsatellite Flanking Sequences Untangled Avena Accessions Discrimination
The amplification of sequences flanking Nikita retrotransposon LTRs is presented in Figure 6. Banding profiles analysis allowed the identification of three bands common to all genotypes with~400, 650 and~1300 bp (schematic representation in Figure 6). Both A. strigosa accessions showed identical banding patterns with two species-specific bands (~600 and~1700 bp) besides the three common ones. In A. sativa two landrace-specific bands were detected in "Kyto" and "S. Eulália" (~1400 and~2400 bp, respectively) and A. sativa from Madeira Island lacks the~1250 bp band present in all other genotypes analyzed. Finally, the A. sterilis Pt banding profile includes a specific band with~1100 bp ( Figure 6). Thus, using the IRAP marker was not yet possible to distinguish the two A. strigosa landraces. with ~1100 bp ( Figure 6). Thus, using the IRAP marker was not yet possible to distinguish the two A. strigosa landraces. REMAP analysis performed using the same retrotransposon primer combined with the (CA)9G microsatellite anchored primer yielded a banding profile (Figure 7 with schematic representation) that includes four bands common to all Avena genotypes analyzed with ~250, ~380, ~550 and ~750 bp, respectively. A. strigosa banding profile contains also four bands characteristic of this species (~280, ~700, ~950 and ~1000 bp) and the landrace from Madeira Island presents a specific one (~200 bp). Hexaploid species profiles, besides containing the bands common to all genotypes, include two common bands (with ~600 and ~1150 bp). Furthermore, both in A. sativa and A. sterilis two species-specific bands with ~520 and ~780 bp were identified. Also, two landrace-specific bands were also identified in A. sativa "Kyto" (~100 and ~900 bp) and one in "S. Eulália" (~240 bp). Thus, the banding patterns obtained with this REMAP Nikita/(CA)9G marker allowed the clear distinction between all Avena genotypes evaluated. Avena species/accessions genomic diversity evaluation was performed using Cluster analysis of IRAP, REMAP and centromeric banding profiles concatenated by NTSYSpc software, and the REMAP analysis performed using the same retrotransposon primer combined with the (CA)9G microsatellite anchored primer yielded a banding profile (Figure 7 with schematic representation) that includes four bands common to all Avena genotypes analyzed with~250,~380,~550 and 750 bp, respectively. A. strigosa banding profile contains also four bands characteristic of this species (~280,~700,~950 and~1000 bp) and the landrace from Madeira Island presents a specific one (~200 bp). Hexaploid species profiles, besides containing the bands common to all genotypes, include two common bands (with~600 and~1150 bp). Furthermore, both in A. sativa and A. sterilis two species-specific bands with~520 and~780 bp were identified. Also, two landrace-specific bands were also identified in A. sativa "Kyto" (~100 and~900 bp) and one in "S. Eulália" (~240 bp). Thus, the banding patterns obtained with this REMAP Nikita/(CA)9G marker allowed the clear distinction between all Avena genotypes evaluated. with ~1100 bp ( Figure 6). Thus, using the IRAP marker was not yet possible to distinguish the two A. strigosa landraces. REMAP analysis performed using the same retrotransposon primer combined with the (CA)9G microsatellite anchored primer yielded a banding profile (Figure 7 with schematic representation) that includes four bands common to all Avena genotypes analyzed with ~250, ~380, ~550 and ~750 bp, respectively. A. strigosa banding profile contains also four bands characteristic of this species (~280, ~700, ~950 and ~1000 bp) and the landrace from Madeira Island presents a specific one (~200 bp). Hexaploid species profiles, besides containing the bands common to all genotypes, include two common bands (with ~600 and ~1150 bp). Furthermore, both in A. sativa and A. sterilis two species-specific bands with ~520 and ~780 bp were identified. Also, two landrace-specific bands were also identified in A. sativa "Kyto" (~100 and ~900 bp) and one in "S. Eulália" (~240 bp). Thus, the banding patterns obtained with this REMAP Nikita/(CA)9G marker allowed the clear distinction between all Avena genotypes evaluated. Avena species/accessions genomic diversity evaluation was performed using Cluster analysis of IRAP, REMAP and centromeric banding profiles concatenated by NTSYSpc software, and the Avena species/accessions genomic diversity evaluation was performed using Cluster analysis of IRAP, REMAP and centromeric banding profiles concatenated by NTSYSpc software, and the dendrogram produced is presented in Figure 8. As expected, A. strigosa landraces revealed to be sister taxa forming a monophyletic group and the hexaploid genotypes analyzed are clustered in another group. Moreover, the three A. sativa genotypes evaluated are gathered in a sub-group with a similarity coefficient of~0.74. A. sativa "Kyto" and A. sativa from Madeira Island present higher similarity (with similarity coefficient~0.82). Concordantly, the bootstrap value of the A. sativa node is the lowest one demonstrating the great similarity between the A. sativa genotypes analyzed. A. sterilis integrates the hexaploid Avena species branch although clearly separated from A. sativa cluster with the lowest similarity coefficient (0.67). dendrogram produced is presented in Figure 8. As expected, A. strigosa landraces revealed to be sister taxa forming a monophyletic group and the hexaploid genotypes analyzed are clustered in another group. Moreover, the three A. sativa genotypes evaluated are gathered in a sub-group with a similarity coefficient of ~0.74. A. sativa "Kyto" and A. sativa from Madeira Island present higher similarity (with similarity coefficient ~0.82). Concordantly, the bootstrap value of the A. sativa node is the lowest one demonstrating the great similarity between the A. sativa genotypes analyzed.
A. sterilis integrates the hexaploid Avena species branch although clearly separated from A. sativa cluster with the lowest similarity coefficient (0.67).

Discussion
In this work we aimed to portray the genomic diversity of Avena species cultivated in Portugal-A. strigosa and A. sativa, as well as the most representative Avena wild relative-A. sterilis, through molecular and cytogenetic analysis of distinct coding and non-coding repetitive sequences. The analysis of 45S rDNA internal transcribed sequences revealed high similarity between A. strigosa, A. sativa and A. sterilis-98% and 100% in ITS1 and ITS2, respectively, corroborating the great homology previously reported for these sequences in several phylogenetic studies of Avena genus [8,10,11]. However, 45S rDNA assessment through fluorescent in situ hybridization (FISH) revealed some Nor loci topological diversity. In both A. strigosa landraces, two pairs of NORs were identified, as formerly described for other A. strigosa lines [37,38], although only one accession presents pairs of NORs with distinct sizes as previously described in A. strigosa accession PI 258729 [37]. In A. sativa six FISH signals were observed in accordance with the number of Nor loci previously described [37,[39][40][41][42]. Though, while in "S. Eulália" all NORs exhibit similar sizes as previously reported in other A. sativa lines [43,44], A. sativa landrace from Madeira Island shows NOR pairs with distinct dimensions as well as one heteromorphic Nor pair. In A. sterilis six 45S rDNA FISH signals were also detected as formerly reported for this species [37,[39][40][41]. Moreover, distinct physical organization patterns for each pair of NORs were also described, similar to the organization inferred for A. byzantina NORs as revealed in C-banded chromosomes [42]. Thus, in contrast to the constancy of rDNA molecular analysis, Nor loci cytological evaluation allowed the identification of distinct ribosomal chromatin organization patterns.
Portuguese Avena species diversity evaluation was also studied through the analysis of other repetitive sequences previously assessed in cereals, such as pSc200 sequence-a major family of sequences isolated from rye and mapped in subtelomeric domains [32,45]. In this work we amplified in Avena a 444 bp fragment with primers previously used in rye and wheat lines with rye chromatin introgression [31] and sequence analysis revealed a high level of similarity (99%-100%) with rye original pSc200 sequence (Accession Number Z50039) [26]. pSc200 sequences were also detected by Southern blot and PCR as well as by FISH analysis in other Triticeae species from the genera Secale, Agropyron and Dasypyrum [26]. However, none of the Avena genotypes studied presented any

Discussion
In this work we aimed to portray the genomic diversity of Avena species cultivated in Portugal-A. strigosa and A. sativa, as well as the most representative Avena wild relative-A. sterilis, through molecular and cytogenetic analysis of distinct coding and non-coding repetitive sequences. The analysis of 45S rDNA internal transcribed sequences revealed high similarity between A. strigosa, A. sativa and A. sterilis-98% and 100% in ITS1 and ITS2, respectively, corroborating the great homology previously reported for these sequences in several phylogenetic studies of Avena genus [8,10,11]. However, 45S rDNA assessment through fluorescent in situ hybridization (FISH) revealed some Nor loci topological diversity. In both A. strigosa landraces, two pairs of NORs were identified, as formerly described for other A. strigosa lines [37,38], although only one accession presents pairs of NORs with distinct sizes as previously described in A. strigosa accession PI 258729 [37]. In A. sativa six FISH signals were observed in accordance with the number of Nor loci previously described [37,[39][40][41][42]. Though, while in "S. Eulália" all NORs exhibit similar sizes as previously reported in other A. sativa lines [43,44], A. sativa landrace from Madeira Island shows NOR pairs with distinct dimensions as well as one heteromorphic Nor pair. In A. sterilis six 45S rDNA FISH signals were also detected as formerly reported for this species [37,[39][40][41]. Moreover, distinct physical organization patterns for each pair of NORs were also described, similar to the organization inferred for A. byzantina NORs as revealed in C-banded chromosomes [42]. Thus, in contrast to the constancy of rDNA molecular analysis, Nor loci cytological evaluation allowed the identification of distinct ribosomal chromatin organization patterns.
Portuguese Avena species diversity evaluation was also studied through the analysis of other repetitive sequences previously assessed in cereals, such as pSc200 sequence-a major family of sequences isolated from rye and mapped in subtelomeric domains [32,45]. In this work we amplified in Avena a 444 bp fragment with primers previously used in rye and wheat lines with rye chromatin introgression [31] and sequence analysis revealed a high level of similarity (99%-100%) with rye original pSc200 sequence (Accession Number Z50039) [26]. pSc200 sequences were also detected by Southern blot and PCR as well as by FISH analysis in other Triticeae species from the genera Secale, Agropyron and Dasypyrum [26]. However, none of the Avena genotypes studied presented any detectable cytological signal, probably due to the reduced number of pSc200-like copies, substantiating the idea that the copy number of this repetitive sequences is extremely variable in distinct cereal taxa, as proposed by [26].
Contrasting with the molecular analysis of ITS and pSc200 sequences, cereal centromeric sequence (CCS1) strongly contributed to assess Avena genotypes diversity, as previously used for other cereals such as H. vulgare, S. cereale and T. aestivum [27,28]. PCR with CCS1 primers yielded consistent banding profiles that strengthen our analysis of Avena genotypes allowing the discrimination of most genotypes analyzed with the exception of the two A. strigosa landraces. CCS1 banding profiles obtained include a novel Avena band with~700 bp with high similarity in all accessions studied. The BLAST analysis of this sequence targeted retrotransposon-related centromeric sequences isolated from cereals [33,34], showing a high similarity with the LTR of Cereba retrotransposon firstly identified in Hordeum vulgare [35], supporting therefore previous works on retrotransposon-like sequences richness in grasses centromeres [46]. FISH mapping of the CCS1 Avena-700 sequence was restricted to centromeric domains in the Avena species here studied as formerly reported for similar sequences in Triticum, Secale and Hordeum chromosomes [47]. Furthermore, our FISH analysis in diploid and hexaploid Avena species also reinforces some phylogenetic aspects in Avena genus, due to its presence in all A. strigosa chromosomes but only in 2/3 of the chromosome complement from A. sativa and A. sterilis. This differential labeling suggests a distinct origin for C genome chromosomes as already proposed by Langdon and co-workrs [46] due to evidence on the absence of conserved retrotransposon centromeric sequences from several cereal species in the C genome putative progenitors of hexaploid oats such as A. eriantha and A. ventricosa. Moreover, the same authors also reported a reverse transcriptase probe derived from A. strigosa, the predicted A/D genome ancestor, which only labels A and D centromeric domains in A. sativa chromosomes, reinforcing therefore the suggested higher affinity of A and D genomes on hexaploid Avena (ACD) species [48].
Avena species/accessions discrimination was considerably increased by polymorphisms detected through the amplification of retrotransposons and microsatellites flanking sequences using IRAP and REMAP analysis [17]. Indeed, the accurate distinction of all Avena genotypes considered in this study was achieved through species and accession-specific bands identification, paving the way for the establishment of fast and simple molecular approaches to clearly identify Avena genotypes. Moreover, the dendogram constructed based on concatenated centromeric, IRAP and REMAP banding profiles undoubtedly allows a clear assessment of Avena diversity clustering all genotypes in two clades, one with A. strigosa and the other corresponding to hexaploid species, although separating A. sativa and A. sterilis accessions. Considering that REMAP technique was previously used to assess hexaploid oat species phylogenetic relationships [25] and proved to be related with oat important traits such as height [24] and cadmium accumulation [49] our work reinforces the high relevance of such molecular markers [50] also in the assessment of Avena biodiversity pool regarding both cultivated and wild relatives.

Plant Material
In this work we have used two Portuguese landraces of A. strigosa (2n = 14, AsAs); two Portuguese landraces and one line from Finland extensively used for comparative analysis of A. sativa (2n = 6x = AACCDD); and one wild Portuguese accession of A. sterilis. All lines used are described in detail in Table 3.
Seeds were germinated in petri dishes and three plants of each accession were maintained in soil pots at 16 h/light (25˝C) and 8 h/dark (18˝C). For DNA extraction fresh and healthy young leaves of 1 month-old plants were frozen and kept at´80˝C until further use. For cytogenetic analyses root tips were collected, cold treated for 24 h, fixed in ethanol/acetic acid (3:1 v/v) and stored at´20˝C until use.

Genomic Analysis through PCR
DNA extractions were performed using cetyltrimethylammonium bromide CTAB method [51]. PCR analysis was performed using different primers designed for 45S rDNA internal transcribed spacers (ITS), rye subtelomeric sequence (pSc200, Accession Number Z50039) [26], cereal centromeric sequences (CCS1, Accession Number, U52217; D primer) [28], retrotransposon Nikita and microsatellites (CA)9G. Primers used are described in detail in Table 4. PCR reactions were performed in a total volume of 20 microliter with 1ˆPCR buffer, 1.5 mM MgCl 2 , 0.25 mM dNTP's, 1 mM each primer, 0.5 U Taq polymerase, 50 ng DNA template. PCR amplification reactions of centromeric, subtelomeric and 45S rDNA ITS used the following program: 5 min 94˝C, 30 cycles of 45 s 94˝C, 45 s 60˝C (54˝C for subtelomeric sequence) and 1 min of 72˝C, final extension of 10 min at 72˝C. IRAP and REMAP were performed accordingly to [17]. PCR products were run on 1.7% agarose gels using as molecular weight marker 1 kb Plus DNA Ladder (Invitrogen, Carlsbad, CA, USA). Amplification products were detected with ethidium bromide and photographed using a Bio-Rad GEL DOC 2000 (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Each PCR banding pattern was considered reproducible after being obtained in at least three technical replicates for each PCR experiment. For each sequence and each line studied at least three individual plants were analyzed. For 45S rDNA internal transcribed spacers pTa71 plasmid, a 9 kb fragment from T. aestivum containing the 18S-5.8S-25S rDNA and intergenic spacers cloned in pUC18 [52], was also used as control template. Similarly, for pSc200 subtelomeric sequence rye genomic DNA was used as control template. Nikita LTR Retrotransposon 5 1 -CGCTCCAGCGGTACTGCC [53] (CA)9G Microsatellite 5 1 -CACACACACACACACACAG -Selected bands were gel isolated, purified using High Pure PCR Product Purification Kit (Roche, Basel, Switzerland), cloned using TOPO TA Cloning Kit in pCR2.1 vector (Invitrogen, Carlsbad, CA, USA) and sequenced. At least three distinct clone per sequence/line were analyzed.

Sequence Similarity Analysis
Sequences from isolated bands were verified and aligned using BioEdit version 7.1.3.0 [54] and analyzed along with sequences obtained from DDBJ/EMBL/GenBank International Nucleotide Sequence Database. For sequence similarity analyses Clustal W2 [30] was used.
Inter Retrotransposons Amplified Polymorphism (IRAP), Retrotransposons Microsatellite Amplified Polymorphism (REMAP) and centromeric sequences PCR banding profiles obtained were transformed in binary matrices, where 1 corresponds to band presence and 0 to absence. NTSYSpc software [55] was used to infer genetic similarity between all accessions and species using Dice coefficient [36]. Dendrograms were obtained through UPGMA (Unweighted Pair Group Method with Arithmetic Mean) method for a concatenated binary of CCS1, IRAP and REMAP profiles obtained using Concatenator Software [56]. Cophenetic correlation coefficient (r) was estimated to verify the adjustment between similarity matrix and respective dendrogram-derived matrix. The determination of confidence limits by bootstrap analysis was done using the WINBOOT software.

Cytogenetics Analysis through Fluorescent in Situ Hybridization (FISH)
FISH analysis was performed as described in [45] in meristematic cells spreads obtained from root tips fixed in ethanol/acetic acid (3:1) and digested with pectinase/cellulose. Pre-hybridization treatments to remove proteins and RNA were performed with 6.67 µg/mL pepsin and 100 ng/µL RNAse, respectively. pTa71 probe, a plasmid that includes a 9 kb EcoRI fragment of the rDNA unit from wheat (Triticum aestivum) [52] was labeled with digoxigenin-dUTP or biotin-dUTP (Roche, Basel, Switzerland) through Nick Translation. Centromeric sequence (Accession Number KM948610) was amplified from A. sativa and labeled with biotin-dUTP (Roche) through PCR using CCS1 primers. The subtelomeric probe pSc200 was amplified from A. sativa DNA and labeled with digoxigenin-dUTP through PCR. For each slide 50 ng of pTa71, 100 ng of centromeric sequence and 100 ng of pSc200 were diluted in a hybridization mixture composed by 50% formamide, 2ˆSSC (saline-sodium citrate buffer), 0.17% SDS (sodium dodecyl sulfate), 1 µg/µL salmon sperm, 10% dextrane sulphate (stringency of 77%). Post-hybridization washes were done with a stringency of 85% and bovine serum albumin 5% (w/v) was used as blocking reagent before probe detection with 2 µg/mL of Anti-digoxigenin Fluorescein (FITC) and 5 µg/mL Streptavidin-Cy3. Cells were counterstained with 4 1 ,6-diamidino-2-phenylindole hydrochloride (DAPI) in Citifluor antifade mounting medium (AF1; Agar Scientific, Stansted, Essex, UK). Samples were examined using a Zeiss AxioImager Z1 epifluorescence microscope. Images were obtained using a Zeiss AxioCam HRm digital camera (Zeiss, Oberkochen, Germany) and the selected ones were processed using Photoshop (Adobe Systems, San Jose, CA, USA). The described FISH signals illustrated in Figures 1, 2 and 5 were observed in at least 10 metaphase or prometaphase cells. 45S rDNA labeling was analyzed through the comparison of FISH signals sizes within each cell, considering that all chromosomes in the same cell assume a similar condensation level.

Conclusions
Genomic diversity and molecular discrimination of the most relevant crop and weed Avena species from Portugal were assessed through molecular and cytological evaluation of distinct repetitive sequences. FISH analysis of 45S rDNA revealed inter and intra-specific variability though the ITS high similarity observed. pSc200-like sequences were identified and revealed high homology in the distinct genotypes analyzed. Conversely, the analysis of centromeric banding patterns along with profiles produced by retrotransposons and microsatellites targeting methodologies (IRAP and REMAP) allowed an accurate assessment of the variability, providing novel prospective molecular markers for future oat breeding programs.