The location of Spelt1 and Spelt52 at chromosomal ends and low copy number in the T. aestivum genome were the advantages enabling Spelt1 and Spelt52 to be used as probes for selecting appropriate subtelomeric BAC clones. We screened the T. aestivum (cv. Renan) BAC genomic libraries, representing 8× genome equivalents, with 1 032 192 clones. The first step consisted in PCR screening of pools of BAC clones. We used the universal Spelt1(F/R) and Spelt52(F/R) primers (see Materials and Methods). The PCR pattern corresponding to a multiband amplification, characteristic of tandemly organized DNA sequences, was obtained for five BAC clones with Spelt52(F/R) primers. These five BAC clones were used for further characterization (Table 1). We failed to identify Spelt1-containing BAC using this PCR screening strategy. Therefore, we hybridized part of the library (290 000 BAC clones) with Spelt1 probe and identified four BAC clones for further analysis (Table 1).

The BAC clones identified as containing Spelt1 and Spelt52 repeats were further characterized as follows:

(1) Determination of insert lengths;

(2) Estimation of the content of repetitive sequence in cloned DNA fragments, including various repeats occurring in the subtelomeric regions in wheat--Spelt1 14, Spelt52 26, pSc119.2 27, pAs1 28, and the telomeric repeat TTTAGGG 6; and

(3) Hybridization of BAC clones to mitotic metaphase chromosomes of the common wheat cvs. Chinese Spring (CS) and Renan.

The results were summarized in Table 1. The size of the inserted fragments varied from 87 to 168 kb. The comparison of the chosen BAC clones based on the length of fragments obtained by NotI, HindIII, BamHI, and EcoRI hydrolysis showed that the cloned DNA could partly overlap in the following BACs: (1) 2322J20 and 2424P01, (2) 110O07 and 1487N20. High content of repetitive DNA sequences was a characteristic feature of all the chosen clones. According to Southern hybridization of HindIII, BamHI, EcoRI fragments of BAC clones to common wheat total DNA, the highest content of repeats was 80% and the lowest, 25%. The results of in situ hybridization were consistent, confirming a high content of repeats (mainly of mobile elements) in BAC clones. The hybridization signals covered the entire length of T. aestivum chromosomes for eight of the nine BAC clones (Figure 1a); of these eight clones, two BAC clones (2383A24 and 112D20) hybridized to a limited set of chromosomes. Predominantly subtelomeric location on the common wheat chromosomes was demonstrated for BAC_2050O8 only; in this clone, the repeat content was lower then in other clones (Figure 1c, Table 1). The distribution pattern of BAC_2050O8 on the chromosomes was different. Some of the chromosomes, such as 5A, gave a distinct hybridization signal only at the chromosomal end, whereas in other chromosomes, such as 4AL, 5BL, and 7BL, hybridization signal displayed multiple locations on their distal third. Weak signals in the interstitial regions of chromosomes indicated that part of the sequences present in this BAC clone were present not only in telomeres, but also in other chromosome regions. The Spelt1 and Spelt52 contents in the selected BAC clones were estimated by dot hybridization. The BAC clones isolated by PCR using Spelt52(F/R) primers contained 23-28 copies of Spelt52. No Spelt1 repeats were found in these clones (Table 1). The BAC clones selected by the hybridization to the Spelt1 probe contained up to seven copies of this repeat. Among them, Spelt1 was detectable only in single copies in clone 112D20. The Spelt52 sequences were also identified in clone BAC_112D20, which was originally selected using the Spelt1 probe (Table 1). To confirm that Spelt1 and Spelt52 were organized in tandem arrays in the corresponding BAC clones, we assayed the corresponding BAC clones using two methods: PCR with the primers specific to these two repeats and RFLP analysis of partially digested BAC clones (HaeIII for Spelt1 and EcoRI for Spelt52). The arrays of tandemly arranged Spelt1 sequence were found in BAC clones 2322J20, 2383A24, and 2424P01 and arrays of Spelt52, in BAC clones 110B21, 2050O8, 478L20, 110O07, and 1487N20 (data not shown). The BAC clones were also examined for the presence of the tandem repeats pSc119.2 and pAs1, used for identification of T. aestivum chromosomes. Part of these repeats was located in the subtelomeric regions in the common wheat chromosomes 29. Southern hybridization showed that pSc119.2 repeats were absent in these BAC clones, whereas pAs1 repeats were present in BAC_478L20 only, giving a rather strong hybridization signal (Table 1). No telomeric repeats (TTTAGGG)n were identified in all BAC clones when using telomeric repeat synthetic probe (see Materials and Methods).

It is very difficult to accurately localize wheat BAC clones on chromosomes using in situ hybridization, because they are very rich in repetitive DNA sequences, mainly transposons, present everywhere on the chromosomes 30. Thus, we attempted to localize these BAC clones through their fragmentation and in situ hybridization of the derived fragments on the chromosomes of common wheat. The procedure consisted in (1) choosing and hybridization of BAC fragments covering more than 10 kb and free of repeats or repeat-poor and (2) hybridization of the fragments containing Spelt52 sequences. The repeat-poor BAC fragments were chosen based on hybridization of BAC restriction fragments with total wheat DNA and with the Spelt1 and Spelt52 probes. The DNA fragments of 10 kb and longer were used as probes for in situ hybridization. Consequently, we succeeded in isolating fragments of about 30 kb (2322J20/30) and 40 kb (478L20/40) from BAC clones 2322J20 and 478L20, respectively. These two fragments were localized to the subtelomeric chromosome regions; the result of in situ hybridization of 2322J20/30 is shown in Figure 1b. A fragment of about 21 kb was isolated from BAC_2050O8. In situ hybridization of this fragment confirmed its subtelomeric localization; however, this approach failed to identify precise chromosome localizations of BAC_2050O8. For the remaining BAC clones, the attempts to identify repeat-poor fragments were unsuccessful, presumably, because of the abundance of repetitive sequences (Table 1). The presence of Spelt52 tandem repeats with the total length of 8500-10 000 bp (the monomeric unit of 380-390 bp) in the BAC clones of cv. Renan made this repeat advantageous as a probe for localizing BACs (Table 1). With this in mind, we performed in situ hybridization of a PCR fragment of BAC_2050O8 (Spelt52_2050O8), obtained using Spelt52(F/R). Spelt52_2050O8 hybridized to the ends of three chromosome arms: 1BS, 4BL, and 7BL (Figure 1d).

For a more detailed analysis of the structure of subtelomeric DNA, the BAC_2050O8 with a predominant subtelomeric in situ localization was completely sequenced according to Chantret et al. 31. The obtained sequence was annotated according to the following strategy. First, the repetitive sequences were identified in the Triticeae Repetitive Sequence Database (TREP) 32 using the BLASTN and BLASTX algorithms 33 and in the RepBase 34 using the CENSOR program 35. Then, we predicted the genes using the GeneMark.hmm 36 and FGENESH programs 37. The structure of the ~120 kb BAC clone is shown in Figure 2. The transposable elements identified in 2050O8 were listed in Table 2. All the repetitive sequences (including transposons and tandem repeats) in this BAC clone constituted 48.9%. Interestingly, DNA transposons were predominant, representing 24.6% of the entire BAC clone length, whereas retrotransposons constituted only 8.4%. The main DNA transposon subclass in subtelomeric BAC_2050O8 was CACTA, constituting 53.1% of all transposable elements and 17.5% of the BAC clone length. CACTA was represented by Caspar and Isaac elements with complete ends and partial Jorge and Fergat elements (Table 2). Caspar_2050O8 was 11 666 bp in length; in addition, it was the only full-length element, which contained two ORFs (open reading frames) corresponding to the transposase and CTG-2 genes. The Spelt52 sequence represented 8.2% of the total BAC clone length (Table 2). Sequence analysis confirmed that aside from Spelt52, no other tandem repeats were identified. Five hypothetical genes were identified and accounted for 6.1% of the total BAC clone length. Analysis of the predicted ORFs, which were not the parts of transposable elements, was detailed in Table 3. The first two genes were homologous to the hypothetical rice genes and contained conserved xyloglucan endotransglycosylase and riboflavin deaminase-reductase domains; the third was a putative gene for peroxisomal membrane protein PEX11-1. The remaining two ORFs were hypothetical genes: one contained the armadillo/beta-catenin-like repeats, and the other was unknown, predicted by FGENESH 37 and GeneMark.Hmm 36. The genes were separated by LINE retrotransposon insertion and unassigned DNA sequences. BLAST alignments of the BAC_2050O8 sequence and the contigs containing mapped wheat ESTs (expressed sequence tags) from GrainGenes database 38 identified two contigs, 1802 and 11876, with a high homology to 2050O8 sequence. EST contig 1802 was located on chromosomes 6AS and 6BL [http://wheat.pw.usda.gov/cgi-bin/westsql/contig.cgi, NSFT03P2_Contig1802]. The level of homology between EST consensus 1802 (length 489 bp, location on 6AS and 6BL) and BAC 2050O8 sequence (length 501 bp, position 98604-99104 bp) was 86%. The corresponding region of 2050O8 was annotated as unassigned DNA sequences. The homology of the other EST contig 11876 [http://wheat.pw.usda.gov/cgi-bin/westsql/contig.cgi, NSFT03P2_Contig11876] (length 826 bp, location on 5AL, 4BL, and 4DL) to BAC_2050O8 was higher (97%) in a 805 bp stretch, and the region of homology was in the region of the putative peroxisomal membrane protein PEX11-1 (Table 3). It was of interest that comparison of the exon-intron structure between PEX11-1_2050O8 and EST contig revealed a complete correspondence of the EST to the coding part of the gene. Comparison of restriction sites in BAC_2050O8 with the mapping data for EST contig 11876 http://wheat.pw.usda.gov/cgi-bin/westsql/map_image.cgi suggested that BAC_2050O8 was localized to the telomeric part of chromosome 4BL (bin 4BL-5 0.86-1.0). BAC_2050O8 could be more precisely localized according to the homology with EST sequence BE638121 from contig 11876, which was mapped to the telomeric bin 4BL-10 0.95-1.0 39.

We performed the BLAST search for the predicted genes in BAC_2050O8 and their protein products in the Rice Genome Annotation Project Database 40. We found that the four predicted wheat genes, displaying a significant homology to four putative rice genes, belonged to clone OSJNBa0056G13 pseudomolecule (virtual contig), located at the distal end of rice chromosome 3S [http://rice.plantbiology.msu.edu/cgi-bin/gbrowse/rice/?name=OSJNBa0056G13, GenBank:AC134236] (Table 3). The rice genes were located close to each other, and their order was the same as that of the corresponding genes in BAC_2050O8; however, one of the genes, namely, Os03g02610, was in the opposite orientation as compared with BAC_2050O8. All wheat genes showed a significant homology at the amino acid and nucleotide levels to the putative expressed rice genes. At the amino acid level, gene_2050O8-3 matched best to the rice putative protein Os03g02590 (peroxisomal membrane protein PEX11-1), displaying 87% homology over the entire length. Both these genes (rice and wheat) had six introns of about 100 bp. The identity of coding sequences is 81%; 71% of the nucleotide changes were synonymous. The identity of introns is 37-47%.

Other genes also demonstrated significant homology to rice (E < e-80 for protein): at the amino acid level, gene_2050O8-1 was homologous to Os03g02610 (xyloglucan endotransglycosylase/hydrolase protein 5 precursor), gene_2050O8-2 to Os03g02600 (riboflavin biosynthesis protein ribD), and gene_2050O8-4 to Os03g02580 (armadillo/beta-catenin-like repeat family protein) (Table 3).

The Spelt52 sequences in BAC_2050O8 were organized as two arrays of repetitive units with a "head-to-tail" orientation. The first array consisted of seven complete units and one truncated fragment at its 3' end. The second array was located 1375 bp from the first and consisted of 18 complete and two outermost truncated units.

To estimate the degree of Spelt52 sequence conservation in BAC_2050O8, we aligned the nucleotide sequences of 25 complete units using the Multalin program 41 and calculated for each monomer the percentage of identity to the consensus sequence obtained from the Multalin alignment [see Additional file 1]. The mean identity to the consensus was 93% (within the range of 88-96%). There were no specific differences between the elements from the first and second arrays.

To investigate the evolution of Spelt52 repeat in Triticeae species, all the publicly available Spelt52 sequences were used. Five Spelt52 sequences [GenBank: AY117400, AY117401, AY082346, AY082347, Z21644] were found, all derived from Ae. speltoides genome. It was earlier demonstrated that the Spelt52 repeat family consisted of two types of monomeric units: Spelt52.1 and Spelt52.2, which shared a conserved 280 bp region and had short (110 and 96 bp) variable regions 26. AY117400 and AY117401 were the Spelt52.1 and Spelt52.2 monomers, respectively, whereas clones AY082346, AY082347, and Z21644 consisted of two Spelt52.1 and Spelt52.2 monomeric units.

We performed phylogenetic analysis of all the repeat monomers that contained conserved and variable regions. For Spelt52 sequences from BAC_2050O8, we took the consensus of 25 complete monomeric units. We considered the consensus as an adequate representative of the bulk of Spelt52_2050O8 sequences, because the monomer sequences displayed no specific differences. The phylogenetic tree constructed using MEGA4 42 contained two major branches formed by Spelt52.1 and Spelt52.2 sequences. The T. aestivum consensus clustered with the Spelt52.2 sequences of Ae. speltoides, indicating that BAC_2050O8 contained only the monomers of Spelt52.2 subfamily (Figure 3).

The gaps remaining in the telomeric regions pose hindrances to complete sequencing of eukaryotic genomes 22. New genomic libraries were constructed using various plasmid vectors for closing gaps and capturing the telomeric sequences 2343. The diversity of DNA that marks the telomeric/subtelomeric regions facilitates the task of sequencing regions at the chromosomal ends. This is quite timely in view of the wheat genome sequencing initiative 44. Our relevant finding was that Spelt1 and Spelt52 sequences were specific markers of the chromosome ends of Ae. speltoides and several polyploid wheat species. Hybridization of Spelt52_2050O8 probe to the wheat cv. Renan chromosomes detected three sites with a weak signal per haploid genome at chromosome ends (Figure 1d). Spelt1 was localized on two chromosome ends per haploid genome of cv. Renan; one site gave a rather strong hybridization signal (data not shown). Use of Spelt1 and Spelt52 for screening the BAC library of cv. Renan enabled us to choose nine clones presumably associated with the subtelomeric regions (Table 1). BAC_2050O8 was localized by in situ hybridization (at the levels of both the clone and subclones) predominantly to the subtelomeric chromosome regions (Figures 1c, d). Comparing the in situ hybridization data for the Spelt52_2050O8 and the BLAST search against the wheat mapped EST contigs 38, we localized 2050O8 to the end of 4BL. A subtelomeric localization of two additional BAC clones, 2322J20 and 478L20, was demonstrated by in situ hybridization of their subclones 2322J20/30 and 478L20/40, respectively (Figure 1b). Since the Spelt52 probe hybridizes only with the subtelomeric regions of chromosomes 1BS, 4BL, or 7BL, we can assume that the BAC clones containing over 23 copies (8740 bp) of the Spelt52 repeats were from the subtelomeric regions of chromosomes 1BS, 4BL, or 7BL (Table 1). Of the Spelt1-containing BAC clones, 2424P01, partly overlapping with clone 2322J20, can be also ascribed to subtelomeric regions (Figure 1b). The in situ hybridization on the chromosomes of cv. Renan detected two terminal blocks per haploid genome. However, the screening of BAC library with Spelt1 probe found no clones containing Spelt1 regions with a length exceeding 5000 bp (the region of repeated DNA detected by in situ hybridization) (Table 1). Two reasons can explain the absence of extended Spelt1 regions in the selected clones, namely:

(1) The terminal chromosome regions, containing longer Spelt1 repetitive sequences, escaped cloning when obtaining the genomic BAC library of T. aestivum cv. Renan, as they could be cut off by the restriction endonucleases EcoRI, HindIII, and BamHI, used for hydrolysis 24; and

(2) A high recombination activity in the region of Spelt1 repeats could result in discharge of part of these repeats during construction of this library. Such effects were observed earlier when we attempted to clone the tandem Spelt1 repeats with a length of 2000 bp and more (unpublished data). A high recombination level leading to deletions of Spelt1 repeats in the progeny of individual self-pollinated plants was also observed 45.

A subtelomeric localization of two clones, 2383A24 and 112D20, was questionable because of a low Spelt1 and Spelt52 copy number and a failure of in situ hybridization approach in determining their precise localization. Thus, the example of Spelt1 and Spelt52 repeats demonstrated that satellite DNA sequences from the subtelomeric regions of diploid wheat progenitor can be used for selecting the BAC clones from the corresponding regions of common wheat chromosomes.

The Spelt52 repeats were undetectable in polyploid Emmer wheats (BBAA and BBAADD genomes) by in situ hybridization 1225. On the other hand, both blot hybridization and PCR indicated the presence of individual sequences belonging to this family in some studied accessions of the polyploid wheats with the BBAA and BBAADD genomes 46. The failure of in situ hybridization in detecting these sequences can result not only from that detection of the DNA sequences shorter than 5000 bp by this method is hindered, but also with the divergence between the probe used for hybridization and the homologous sequences within the studied species. In particular, it has been shown that two probes with a homology of 87.3% give quite different hybridization patterns of the tandem repeat family pAs1, found in the genomes of Ae. tauschii and barley 47.

Use of the primers for Spelt52 repeat allowed us to detect a tandem organization of several BAC clones already when screening the BAC library. The amplification of Spelt52 fragments using the same primers and BAC_2050O8 as a template and produced probe for hybridization allowed us for the first time to localize Spelt52 repeats in the chromosomes of common wheat by in situ hybridization to the ends of chromosome arms 1BS, 4BL, and 7BL of wheat cv. Renan. Note that the PCR fragment of Spelt52 sequence was successfully mapped earlier to the end of the genetic map of 4BL chromosomes 39.

The 25 Spelt52 sequences detected in BAC_2050O8 displayed a high level of homology (on the average, 93%). Although BAC_2050O8 contains two regions with Spelt52 tandems, the clustering of repeated units into two groups according to the primary structure was not observed. It is impossible to compare it with extended Spelt52 tandem regions of other accessions and species as the corresponding tandem regions are absent in the relevant databases. Spelt52 family contained two types of repeated units -- Spelt52.1 and Spelt52.2; note that these monomers in the majority of the studied accessions Ae. speltoides alternate 26. The tandems formed by Spelt52.2 sequences have been found only in one Ae. speltoides accession, TS01. In the hexaploid wheat genome, in BAC_2050O8, we detected only Spelt52.2 tandem sequences (Figure 3). Taking into account that Ae. speltoides is the most likely donor of the B/G genomes of polyploid wheats, it would be useful to include the presence of different variants of the tandem arrangement of Spelt52 in the phylogenetic studies of Sitopsis and Triticum species.

A characteristic feature of the tandem repeats is their ability to amplify during the evolution, thereby increasing their copy numbers and creating a high level of intraspecies polymorphism in the copy number 48. An illustrative example of such variation is the Spelt1 family with the repetitive unit of 178 bp. The number of sites where this repeat is localized varies in a wide range in both Ae. speltoides, for which it is genome-specific, and in polyploid species, to whose formation Ae. speltoides has contributed 12. Both individual plants of the same accession and the progeny of the same plant can differ in the copy number of this repeat 45. The level of Spelt52 polymorphism is also rather high 12.

The amplification of the same tandem subtelomeric repetitive DNA sequences in evolutionary remote cereal species has been described, whereas such amplification is absent in closely related species. For example, the repeat family 350 bp/pSc200/pSc74 is characteristic of Secale cereale, S. montanum, Dasypyrum villosum, D. breviaristatum, and Agropyron cristatum; however, it has not been detected by hybridization in other rye species 49. Spelt52 repeats have been localized to chromosomes of Ae. speltoides and other two from four Sitopsis species (Ae. longissima and Ae. sharonensis) 1225. The diversity on the presence of tandem repeats in the subtelomeric chromosome regions of various cereal species suggested the existence of a common pool of sequences able to amplify during evolution, forming a subtelomeric repeat family. Cloning and PCR analysis of the amplification product obtained from the rice (belonging to other cereal tribe) genome using the primers specific to Spelt1 and Spelt52, we detected no homology to either Spelt1 or Spelt52 repeats except for the primer sequences (EA Salina, unpublished data). We have not also found any degenerate sequences or sequences homologous to Spelt52 repeat when comparing wheat syntenic BAC_2050O8 and rice OSJNBa0056G13 pseudomolecule. Presumably, the pool of the common sequences involved in amplification events during evolution exists yet is stringently confined to the frame of a taxonomic unit. Most likely, such unit for cereals is a tribe.

According to the data on distribution of EST DNA sequences, the wheat group 4 chromosomes display the most pronounced similarity to the short arm of rice chromosome 3 395051. Individual deviations from collinearity were observed in the centromeric regions and distal part of the consensus chromosome 4BL map. The distal region of wheat chromosome 4BL has been compared in detail with rice pseudochromosome 3S 39. The most pronounced distinctions were recorded in the terminal telomeric bin 4BL-10 (0.95-1.0), which displayed a higher recombination level. BAC_2050O8, which contains a sequence with a 90% homology to EST BE638121, is located at a distance of 17 mapped ESTs from the 4BL chromosome end (about two-thirds of the entire terminal bin) 39. The four putative rice genes, belonging to OSJNBa0056G13 pseudomolecule and displaying a significant homology to four predicted wheat gene in BAC_2050O8, are located at a distance of about 870 kb from the end of rice chromosome 3S. The terminal regions of wheat (BAC_2050O8, chromosome end) and rice (OSJNBa0056G13, chromosome end) displayed the following differences: (1) 10 of the 17 wheat ESTs located in this region have no homologous sequences in rice, and (2) the DNA sequences covering about 500 kbp from the end of chromosome 3S have no homologs on chromosome 4BL (according to BLAST results on February 20, 2009). In addition, comparison of the order of wheat ESTs and rice pseudomolecules suggested that at least two inversion events have occurred in this chromosome region besides duplications (data not shown). The changes in orientation of the first of the four genes predicted in BAC_2050O8 (encoding a hypothetical protein with xyloglucan endotransglycosylase domain) relative to the corresponding rice genes can result from such inversions. Note that the rice region (in OSJNBa0056G13) containing these four genes is of 9556 bp versus 29 612 bp of the corresponding wheat region in BAC_2050O8 (Table 3). This difference results from an enlarged intergenic region in BAC_2050O8 (totally amounting to 23737 bp versus 3641 bp in rice). The intergenic regions in BAC_2050O8 contains two degenerated LINE elements (2941 and 2815 bp), one MITE element (117 bp), and an unassigned sequence. At the nucleotide level, the intergenic regions of rice and wheat showed no similarity to each other, and the regions flanking this block of genes in rice showed no similarity to the overall BAC_2050O8 sequence.

The wheat regions containing the four predicted genes in BAC_2050O8 is more than threefold longer that the corresponding region from rice OSJNBa0056G13, whereas the overall wheat genome exceeds the rice genome more than 40-fold and the distal region of chromosome 4BL is 30-fold longer than the corresponding homologous region in rice 39. Thus, it is most likely that the wheat genome increased in the regions free of genes. In the case of BAC_2050O8, these are the regions of retrotransposons, CACTA transposons, Spelt52 tandem repeats, and unassigned sequences.

The BAC clones carrying subtelomeric tandem repeats Spelt1 and Spelt52 26 were obtained from the genomic BAC library of T. aestivum cv. Renan 24 by hybridization with Spelt1 probe and PCR screening using the primers specific to Spelt52.

BAC DNA preparation was obtained with a modified alkaline lysis procedure 55. To estimate the insert length of BAC clones, BAC plasmid DNA (2 μg) were digested with NotI to release the insert from the vector. Restriction fragments were separated in 1% (w/v) agarose gel (14°C, 15 h, 150 V, ramped from 1 to 15 seconds) using a contour-clamped homogeneous electric field (CHEF) 56. The electrophoresis buffer was 0.5 × TBE (45 mM Tris-borate pH 8.0 and 1 mM EDTA). Lambda concatemer was used as a molecular size standard. The lengths of restriction fragments were measured using the Gel analysis v1.0 program (Lytech, Russia)

The specific primers were used for Spelt1 (Sp1F, 5'-tccaa-accat-ccccg-tcaag-cg-3' and Sp1R, 5'-aagtt-cttct-ggccg-tgcca-ta-3') and Spelt52 (Sp52F, 5'-gcaca-caaac-ccgga-gaaag-t-3'and Sp52R, 5'-tcccc-gtttc-ttctc-tagcc-t-3') 26 repeats. PCR analysis of BAC clones was performed in an Eppendorf Mastercycler according to the following mode: 30 cycles of 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C followed by a final stage of 15 min at 72°C. PCR products were separated by electrophoresis in 1% agarose gel.

Plasmids with inserted telomere-associated tandem sequences Spelt1 and Spelt52, isolated from Ae. speltoides; pSc119.2 from S. cereale; and pAs1 from Ae. tauschii were used as probes 14262728. Telomeric repeat synthetic probe was obtained by PCR according to the protocol originally described by 57. Spelt1, Spelt52, pSc119.2, and pAs1 probes were labeled by the random hexamer method with α-³²P-dATP (Amersham Pharmacia Biotech, United Kingdom) using Klenow fragment 58. BAC DNA (2 μg) were digested by HindIII, BamHI and EcoRI restriction endonucleases, or partially digested with HaeIII (for Spelt1) and EcoRI (for Spelt52), separated in 1-1.2% agarose gel, and transferred to a Hybond-N+ membrane (Amersham Pharmacia Biotech, United Kingdom). Filters were first moistened by floating in 2 × SSC. Pre-hybridization was carried out at 65°C for 4 h in 6 × SSC, 5 × Denhardt's solution, 0.5% SDS, and denatured salmon sperm DNA (100 mg/ml). Hybridization was conducted at 65°C for 16 h in the same solution with denatured labeled probe (1 ng/ml). After hybridization, filters were washed at a room temperature in 2 × SSC, 0.1% SDS; 0.5 × SSC, 0.1% SDS and 0.1 × SSC, 0.1% SDS for 15 min each. The membranes were exposed with Kodak X-ray film for a few hours to 3-7 days at -70°C depending on signal intensity.

The copy number of Spelt1 and Spelt52 sequences in BAC clones was estimated by dot blot assay 59. The serial dilutions (1, 0.5, 0.1, and 0.05 μg) of each BAC DNA were spotted onto a Hybond N+ membrane. The plasmid DNA with Spelt1 and Spelt52 inserts were used as a standard with a copy number of 10⁸ to 10¹¹. BAC and plasmid DNA concentrations were determined by spectrophotometry and gel electrophoresis. The DNA samples were denatured with 0.3 M NaOH for 20 min before spotting, neutralized after spotting with 2 × SSC for 5 min, cross-linked to membrane using UV light, dried, and used for dot blot hybridization. The conditions of pre-hybridization and hybridization were as described above for Southern blot hybridization. The signal intensity was counted for each spot by a MiniBeta (LKB). The mean number of counts of the α-³²P-dATP-labeled Spelt1 and Spelt52 probes was calculated for each BAC. The copy number of specific sequences in BAC DNA was estimated by comparing the hybridization intensity with the standard.

For fluorescence in situ hybridization, we used BAC clones, recombinant plasmids, synthetic telomeric probes, or large specific DNA fragments. Large specific DNA fragments (more than 10 kbp long and containing no repetitive sequences) were isolated from agarose gel according to 55. BAC DNA was treated with restriction endonucleases BamHI, HindIII, and EcoRI, ran through agarose gel until fragments separated well, transferred to Hybond N+ membrane, and hybridized with α-³²P-dATP-labeled total DNA of T. aestivum cv. Chinese Spring. The bands that showed no hybridization signals and were longer than 10 kbp were isolated and labeled. DNA probe (1 μg) was labeled by nick translation to incorporate biotin-16-dATP (Life Technologies, Gibco BRL) or digoxigenin-11-dUTP (Boehringer Mannheim) according to the manufacturers' protocol. Telomeric probe was labeled by PCR to incorporate biotin-16-dATP (Life Technologies, Gibco BRL). Metaphase chromosomes were prepared and FISH was performed according to 12 with minor modifications. To identify chromosomes carrying signals, we used the probe combinations pSc119.2 + BAC and pAs1 + BAC. The chromosomes were identified according to standard chromosome nomenclature 6061.

BAC shotgun sequencing and sequence assembly were carried out at the National Center of Sequencing (Evry, France) as described 31. Sequence annotation was done according to the Guidelines for Annotating Wheat Genomic Sequences from the International Genome Wheat Sequencing Consortium 44. The procedure was two-step, including identification of repetitive elements and identification of gene structure. For identification of DNA repetitive elements, we screened the Triticeae Repetitive Elements Database at GrainGenes (TREP) 32, Rice Genome Annotation Project Database 40, NCBI non-redundant nucleotide databases 62 using the BLASTN algorithm 33 with a cutoff value of 1e^-5, and RepBase 34 using CENSOR program 35. Hypothetical transposable element proteins were identified by BLASTX algorithm against TREP hypothetical protein and NCBI non-redundant protein databases. Gene structures in BAC_2050O8 were identified by integrating the results of predictor programs GeneMark.hmm 36 and FGENESH 37. The putative functions were assigned through BLASTN searches against NCBI EST database (dbEST), BLASTP, and BLASTX searches against NCBI non-redundant protein and Rice Genome Annotation Project database 40 according to criteria described in the Guidelines for Annotating Wheat Genomic Sequences 44. The DNA sequences that were not assigned to transposable elements or genes were considered as unassigned DNA. The alignments of Spelt52 nucleotide sequences were performed by Multalin program 41; phylogenetic trees were constructed using a neighbor-joining method by MEGA4 software package 42. The nucleotide sequence reported in this paper is deposited in the GenBank database under accession number GQ165812.